REVIEW

DOI: 10.1002/adma.200502696

Protonated Titanates and TiO2

Nanostructured Materials: Synthesis,
Properties, and Applications**
By Dmitry V. Bavykin, Jens M. Friedrich,
and Frank C. Walsh*

Tubular and fibrous nanostructures of titanates have recently been

synthesized and characterized. Three general approaches (template as-
sisted, anodic oxidation, and alkaline hydrothermal) for the preparation of nanostructured tita-
nate and TiO2 are reviewed. The crystal structures, morphologies, and mechanism of formation
of nanostructured titanates produced by the alkaline hydrothermal method are critically dis-
cussed. The physicochemical properties of nanostructured titanates are highlighted and the
links between properties and applications are emphasized. Examples of early applications of
nanostructured titanates in catalysis, photocatalysis, electrocatalysis, lithium batteries, hydrogen
storage, and solar-cell technologies are reviewed. The stability of titanate nanotubes at elevated
temperatures and in acid media is considered.

1. Introduction
Titanium is the ninth most abundant element in the Earth’s
crust. TiO2, which is the most common compound of titanium,
is often used in many applications ranging from anticorrosion,
self-cleaning coatings, and paints to solar cells and photocata-
lysts.[1] Nanostructured TiO2 materials, with a typical dimen-
sion less than 100 nm, have recently emerged. Such materials
include spheroidal nanocrystallite and nanoparticles together
with (more recently synthesized) elongated nanotubes, nano-
sheets, and nanofibers.
Much of the current interest in nanotubular materials was
initiated by the discovery of carbon nanotubes, which are
promising for many applications, particularly in the field of
electronic materials. Although noncarbon nanotubes were
– properties and applications of conventional TiO2 nanoparti-
[*] Prof. F. C. Walsh, Dr. D. V. Bavykin, J. M. Friedrich
Electrochemical Engineering Group School of Engineering Sciences
University of Southampton
Highfield, Southampton SO17 1BJ (UK)
E-mail: F.C.Walsh@soton.ac.uk
[**] The authors are grateful to Dr. Alexei Lapkin (University of Bath)
and Prof. John Owen (University of Southampton) for stimulating
discussions.
identified in the early 1990s, relatively little research has been
carried out on their synthesis and characterization. Other
nanotubular materials represent a diverse chemistry, signifi-
cantly extend the area of possible applications, possess unique
combinations of physicochemical properties, and are often
easier to synthesize than carbon nanotubes. These properties
and relatively low synthesis costs can render noncarbon nano-
materials attractive for technological applications.
A timeline showing the development and application of
TiO2 and titanate nanotubes is shown in Figure 1. Three
methods for the preparation of TiO2 nanotubes were devel-
oped in the late 1990s. Alkaline hydrothermal synthesis can
produce protonated titanate nanotubes (rather than TiO2),
which can formally be considered as hydrated forms of TiO2.
Since the exact crystal structure is currently disputed, this
nanotubular material is referred to as both protonated tita-
nate and TiO2 nanomaterial. Titanate nanotubes combine the
cles (e.g., wide-bandgap semiconductor, photocatalysts) with
the properties of layered titanates (e.g., ion exchange). During
the last few years, many reports on physicochemical proper-
ties, targeting particular applications, have been published; it
is now timely to review the known properties and potential
application of these materials.

Adv. Mater. 2006, 18, 2807–2824 © 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 2807
 D. V. Bavykin et al./ Protonated Titanates and TiO2 Nanostructured Materials
REVIEW

In this review, the methods of preparation, possible crystal
structures, and mechanisms of formation of nanostructured
TiO2 and titanates are compared. The conditions of synthesis,
which are responsible for regulation of the morphology of ti-
tanate nanostructures, are considered. Analysis of the physi-
cochemical properties of titanates and recent advances in
their applications allow the identification of gaps in our
knowledge and highlight the need for critical studies in the
area of nanostructured titanates.
2. Synthesis of Nanostructured TiO2 and Titanates
There are three general approaches to the synthesis of TiO2
and titanate nanotubes, namely, chemical (template) synthe-
sis, electrochemical approaches (e.g., anodizing of Ti), and the
alkaline hydrothermal method.
2.1. Chemical (Template) Synthesis
The method of template synthesis of nanostructured materi-
als has become extremely popular during the last decade.[2]
The method utilizes the morphological properties of known
and characterized materials (templates) in order to construct
materials with a similar morphology by methods including re-
active deposition or dissolution. This method is very general;
by adjusting the morphology of the template material, it is
possible to prepare numerous new materials with a regular
and controlled morphology on the nano- and microscale. A

Prof. Frank Walsh holds the degrees of B.Sc. in Applied Chemistry (Portsmouth), M.Sc. in Mate-
rials Protection (UMIST/Loughborough), and a Ph.D. on electrodeposition of high-surface-area
(powder) materials from Loughborough University, UK. He is the author of over 200 papers
and three books in the areas of electrochemistry and electrochemical engineering. Frank is a
chartered and registered European Engineer and an international consultant; he leads a research
group in electrochemical engineering. Previous positions include Business Development Direc-
tor (Science) at the University of Portsmouth and Head of the Chemical Engineering Depart-
ment at the University of Bath. He is Professor in Electrochemical Engineering at the University
of Southampton and takes a particular interest in fuel cells, surface engineering, and nanomateri-
als as electrode structures.

Dr. Dmitry Bavykin received an M.Sc. (Chemistry) at Novosibirsk State University in 1995 and
a Ph.D. (Chemical Engineering) at Boreskov’s Institute of Catalysis, Novosibirsk in 1998. He
was awarded a NATO/Royal Society Fellowship in 2002 to study the preparation, characteriza-
tion, and application of TiO2 nanotubes in electrochemistry and catalysis at the University of
Bath. Currently, he is a Research Fellow in the Research Institute for Industry at the University
of Southampton. His major area of interest is the application of novel, nanostructured materials
to renewable energy problems.

Jens Friedrich received his M.Eng. at the University of Bath, UK in 2002. He is completing his
Ph.D. studies at the University of Southampton in the area of fuel-cell (and other high-surface-
area) electrode materials. His interests include porous, 3D carbon supports and electrocatalysts
for proton exchange membrane fuel cells as well as titanate nanotubes.

2808 www.advmat.de © 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Adv. Mater. 2006, 18, 2807–2824
D. V. Bavykin et al./ Protonated Titanates and TiO2 Nanostructured Materials
Figure 1. Simplified timeline describing the development of TiO2-related nanotubular structures.
disadvantage is that, in most cases, the template material is sa-
crificial and needs to be destroyed after synthesis, leading to
an increase in the cost of materials. As in the case of all sur-
face-finishing techniques, it is also important to maintain a
high level of surface cleanliness to ensure good adhesion be-
tween the substrate and the surface coating.
The preparation of TiO2 nanotubes by chemical templating
usually involves controlled sol–gel hydrolysis of solutions of
titanium-containing compounds in the presence of templating
agents, followed by polymerization of TiO2 in the self-as-
sembled template molecules or deposition of TiO2 onto the
surface of the template aggregates. The next stages are selec-
tive removal of the templating agent and calcination of the
sample. Template syntheses of TiO2 nanotubes can be separat-
ed into several groups according to the type of template mole-
cules used.
The group of self-assembled organic surfactant template
molecules is probably the largest one, since there is a wide di-
versity of organic molecules. Among the organic templates
used for TiO2 nanotube synthesis are the organogel of trans-
(1R,2R)-1,2-cyclohexanedi(11-aminocarbonylundecylpyridi-
nium)[3] and dibenzo-30-crown-10-appended cholesterol,[4] the
hydrogel of tripodal cholamide-based materials,[5] and lauryla-
mine hydrochloride surfactant.[6] Other examples of special-
ized templating agents include tobacco mosaic viruses[7] or
precipitated platinum salts.[8]
Porous alumina, produced by anodization of aluminum foil,
has been widely used as a template for the preparation of
TiO2 nanotubes. The internal surface of cylindrical pores of
anodic alumina is used for the deposition of TiO2 thin films
from various precursors.[9–11] After selective removal of alumi-
na, the external diameter of the TiO2 hollow fibers corre-
sponds to the diameter of the pores in the alumina. The inter-
nal diameter of the TiO2 nanotubes depends on the synthesis
conditions and the thickness of the wall (Table 1).
Adv. Mater. 2006, 18, 2807–2824 © 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
REVIEW
After calcination at 500 °C, the crystal structure of TiO2
nanotubes produced by templating is usually amorphous or
polycrystalline anatase. The tubes have different mean inter-
nal diameters depending on the nature of the templating
agent (Table 1). The specific surface area for very wide tubes
is not very high except for the cases where the structure of the
microtube wall contains small channels.[6] Nanotubes have
potential uses in photocatalytic removal of organic pollutants
or in solar cells,[12,13] although industrial applications may be
limited by the cost of materials, insufficient material charac-
terization, and concerns over long-term instability.
2.2. Electrochemical Synthesis (Anodization of Ti)
In 2001, Grimes and co-workers reported the preparation
of self-organized TiO2 nanotube arrays by direct anodization
of titanium foil in a H2O–HF electrolyte[14] at room tempera-
ture. The nanotubes were oriented in the same direction per-
pendicular to the surface of the electrode, forming a continu-
ous film. The thickness of this film (or the length of the tubes)
was only 200 nm (see Table 2). The average internal diameter
of the nanotubes exceeded 50 nm. One end of the nanotubes
was always open while the other end, which was in contact
with the electrode, was always closed. A number of papers
describing electrolytes for the preparation of TiO2 nanotubes
by anodization of titanium have been published. A number
of fluoride-ion-containing electrolytes, including NH4F–
(NH4)2SO4,[15] HF in dimethyl sulfoxide (DMSO)–ethanol
mixture,[16] phosphate,[17] acetate,[18] and non-acidic Na2SO4–
NaF,[19] have been used. The morphological properties of the
resultant nanotubes are summarized in Table 2. The internal
diameter of the tubes can be as small as 20 nm, while the
length of the tube can be less than 2.4 lm. As prepared, the
TiO2 nanotubes have an amorphous crystal structure; after
www.advmat.de 2809
 D. V. Bavykin et al./ Protonated Titanates and TiO2 Nanostructured Materials
REVIEW

Table 1. Morphological properties of TiO2 nanotubes produced by the sol–gel method in the ameter of 90 nm, length of 300 nm, and
presence of templating agents. Bu = n-butyl; iPr = isopropyl. wall thickness of 15 nm, the value of specif-
ic surface area can be estimated to be
Precursor Template Conditions Nanotube Ref.
ca. 20 m2 g–1, by assuming a density of
diameter
[nm] 3.9 g cm–3. In contrast to techniques using
alumina as a template, TiO2 nanotubes pre-
-
PF6
O
+
pared by direct anodization are not usually
C N
25 °C, ethanol, 50–300 [3] separated from each other in a regular man-
NH
Ti(OiPr)4 NH4OH ner and do not have well-developed cavities
NH between the tubes.
+
C N The advantage of TiO2 nanotubes pro-
O -
PF6
duced by anodization is that they have been
25 °C, effectively immobilized on a titanium sur-
Ti(OiPr)4 O O O O O NH 1-butanol, 500 [4] face during preparation. As a result, these
R
HN benzylamine
nanotubes have several possible applica-
R O O O O O
tions. It was recently found that the electro-
conductivity of a TiO2 nanotube film in-
OH
NH 25 °C, ethanol,
creases by several orders of magnitude in
Ti(OBu)4 = R CH3COOH, H2O 4–7 [5] the presence of gaseous hydrogen at
NR3 O
290 °C,[20] making this material promising
HO OH for hydrogen sensing. Similar TiO2 nano-
tubes have also demonstrated promise as
Ti(OBu)4 CH3(CH2)11NH2·HCl 25–40 °C, H2O 1800–6000 [6]
Ti(OiPr)4 Tobacco mosaic viruses 25 °C, ethanol 20 [7] photocatalytic, self-cleaning surfaces[21] or
Ti(OBu)4 [Pt(NH3)4](HCO3)2 25 °C, ethanol 100 [8] as photoanodes for water splitting,[22] where
Ti(OiPr)4 Porous alumina 25 °C, pressure 60–70 [9] the efficiency of the photoanodic response
impregnation depends on the nanotube wall thickness.
TiF4 Porous alumina 60 °C, HCl 2.5–5, [10]
70–100
Ti(OiPr)4 Porous alumina 25 °C, ethanol, 120–140 [11]
CH3COOH 2.3. The Alkaline Hydrothermal Method

In 1998, Kasuga et al.[23] first reported a

simple method for the preparation of TiO2
Table 2. Morphological properties of TiO2 nanotubes produced by anod- nanotubes, without the use of sacrificial templates, by treat-
izing Ti foil in various electrolytes at 25 °C.
ment of amorphous TiO2 with a concentrated solution of
NaOH (10 mol dm–3) in a polytetrafluoroethylene-lined batch
Electrolyte composition Electrode Internal Length Ref.
potential diameter [lm]
reactor at elevated temperatures. In a typical process, several
vs. SCE [nm] grams of TiO2 raw material can be converted to nanotubes,
[V] with close to 100 % efficiency, at temperatures in the range
0.5–3.5 wt % HF 3–20 25–65 0.2 [14] 110–150 °C, followed by washing with water and 0.1 mol dm–3
in H2O HCl. It has since been demonstrated that all polymorphs of
0.5 wt % NH4F 20 90–110 0.5–0.8 [15] TiO2 (anatase, rutile,[24,25] brookite,[26] or amorphous forms)
in 1 mol dm-3 (NH4)2SO4 can be transformed to the nanotubular or nanofibrous TiO2
4 wt % HF 20 60 2.3 [16]
under alkaline hydrothermal conditions. The detailed crystal
in 48 wt % DMSO, 48 wt %
ethanol structure and dependence of nanotube morphology on the
0.5 wt % NH4F 20 40–100 0.1–4 [17] synthesis conditions will be discussed in the next section.
in 1 mol dm-3 (NH4)H2PO4,
1 mol dm-3 H3PO4
0.5 wt % NH4F 10–120 20 0.1–0.5 [18]
in CH3COOH
3. Structure
0.1–1 wt % NaF 20 100 2.4 [19]
0.1–2 in 1 mol dm-3 Na2SO4 3.1. Crystallography

Nanostructured titanate materials produced by alkaline hy-

calcination at 500 °C, the crystal structure was reported[14] to drothermal treatment are not amorphous and have well-de-
correspond to an anatase and rutile mixture. The value of spe- fined X-ray diffraction (XRD) patterns. Our understanding of
cific surface area of the TiO2 nanotubes could be estimated crystal structure of nanotubular titanates is incomplete due to
from their geometry. For nanotubes with typical internal di- several difficulties. Firstly, there is a large number of crystal

2810 www.advmat.de © 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Adv. Mater. 2006, 18, 2807–2824
D. V. Bavykin et al./ Protonated Titanates and TiO2 Nanostructured Materials

REVIEW
modifications not only for pure TiO2 (ana-
tase, rutile, and brookite) but also for pro-
tonated forms of polytitanic acids, H2mTin-
I II
O2n + m. Secondly, the small size of crystals M’’ M’
leads to a small value of coherence area,
which results in broadening of the reflec-
tions in the XRD pattern. Thirdly, wrapping
along a certain crystallographic axis during c
the formation of nanotubes can result in
the widening of peaks of a given Miller in-
dex, making interpretation and assignment
of peaks more difficult. Another difficulty β
is that the nanotubular structure of tita-
nates is relatively unstable and can undergo a c
further phase transformation during heat-
ing,[27] acid treatment,[28] or other chemical M
treatments, during or after preparation of
nanotubes. All of these factors contribute III b IV
to controversy over the exact crystal struc-
(020)
ture of titanate nanotubes produced by al-
(211)
kaline hydrothermal treatment. Here, the
(001) (110)
possible crystal structures of titanate nano- (200)

tubes or other nanostructures currently

found in the literature are reviewed. b
In an early study, Kasuga et al.[23] charac-
terized their product as anatase. In a recent
paper, the crystal structure of nanotubular 10 20 30 40 50 60 70 80

titanates has still been interpreted as ana- a 2 Theta / degree

tase,[29] despite the fact that the crystal

structure of these nanotubes is known to be Figure 2. Crystal structure of monoclinic trititanic acid (H2Ti3O7) in three different projections
in octahedral presentations (I,II,III) and IV) typical X-ray diffraction (XRD) pattern of TiO2
more sophisticated. nanotubes together with reflections (bars) from monoclinic H2Ti3O7. The dashed line shows
Based on XRD, selected-area electron dif- the dimension of a unit cell. Atoms of hydrogen are not shown for clarity [30,49]. a, b, and c are
fraction (SAED), and high-resolution trans- crystallographic axes, b is the monoclinic angle. Connection of point M to point M′ results in
mission electron microscopy (HRTEM) data, the formation of nanotubes with chirality equal to zero.
Peng and co-workers[30,31] proposed that the
crystal structure of titanate nanotubes corresponded to the that rolling of the plane occurs around the [010] axis such that
layered trititanic acid (H2Ti3O7) with a monoclinic crystal struc- the axis of the nanotube is parallel to the b-axis of monoclinic
ture (Table 3). A schematic showing the crystal structure of H2Ti3O7. Recently, Wu et al.[32] have proposed that rolling of
monoclinic trititanic acid in a TiO6 edge-sharing octahedron rep- the (100) plane could occur around axis [001]. In both cases, the
resentation is shown in Figure 2; the three different projections walls of the nanotubes consist of several layers, typically separat-
corresponding to crystallographic axes. A nanotubular morphol- ed by 0.72 nm. The structure of each layer corresponds to the
ogy of layered trititanic acid can by obtained by rolling several structure of the (100) plane of monoclinic titanates, which is a set
(100) planes around axis [010] or [001]. It has been proposed[30] of closely packed TiO6, edge-sharing octahedra (Fig. 2 II).

Table 3. Comparison of the crystal structures of layered titanates.

Crystallographic Symmetry Lattice parameters [b] XRD reflection, 2
[°] Ref.

phase [a] a b c b
Nanotubes 10.5 24.4 28 34 38.5 44.5 48.2 61.5 [49]
H2Ti3O7 Monoclinic 1.602 0.375 0.919 101.5° 11 24.4 29 33 38 48.4 60 62 [31]
H2Ti2O4(OH)2 Orthorhombic 1.926 0.378 0.300 90° 9 24.3 28 34 38 48 62 [34]
H2Ti4O9·H2O Monoclinic 1.825 0.379 1.201 106.4° 10 24 28 48 [33]
HxTi2–x/4&x/4O4·H2O Orthorhombic 0.378 1.834 0.298 90° 9.5 24.5 28 48 62 [35]
TiO2-B Monoclinic 1.218 0.374 0.652 107.1° 15 25 29.5 44 48 57 62 [39]
H2Ti5O11·H2O Monoclinic 2.001 0.376 1.499 124.0° 10 14 36 43 46 [43]

[a] 䊐 indicates a vacancy. [b] Lattice parameters a,b and c given in [nm].

Adv. Mater. 2006, 18, 2807–2824 © 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.advmat.de 2811
 D. V. Bavykin et al./ Protonated Titanates and TiO2 Nanostructured Materials
REVIEW
Based on XRD and transmission electron microscopy
(TEM) data, Nakahira et al.[33] characterized a TiO2 nanotube
sample, produced using alkaline hydrothermal treatment, as a
tetratitanic acid H2Ti4O9·H2O. The presented XRD spectra,
however, have very broad reflections, which have similar posi-
tions to those of trititanic acid. The crystal structure of tetrati-
tanic acid is similar to trititanic acid, the former having four
edge-sharing octahedra in the unit cell rather than the three
edge-sharing octahedra of the latter.
Based on studies of the sodium content of titanate nano-
tubes at various pH values, during acid washing (in combina-
tion with XRD and TEM data) Jin and co-workers[34] pro-
posed the following crystal structure of titanate nanotubes:
H2Ti2O4(OH)2 with an orthorhombic unit cell. Both protons
of the bititanic acid could be ion exchanged with sodium ions.
The XRD pattern had similar reflections to the monoclinic
trititanic acid (Table 3).
The orthorhombic dititanic acid also has a layered structure
of walls, where each layer represents a (100) plane rolled
around the b-axis. The (100) plane is built up from edge-shar-
ing TiO6 octahedra, forming a zigzag structure.
Ma et al.[35,36] obtained XRD, Raman spectroscopy, X-ray
absorption fine structure, and electron-diffraction data and
suggested that lepidocrocite-type (H0.7Ti1.827&0.175O4.0·H2O,
where & indicates a vacancy) structures could be present in
the mixture of layered nanotube titanates. Detailed electron-
diffraction studies of selected zones on the nanotubes also
support this hypothesis.[37] Unlike trititanic acid, which has
three steps of corrugated layers, the lepidocrocite titanate
consists of a continuous and planar 2D array built up from
TiO6 edge-sharing octahedra.
It was demonstrated that bulk layered protonated titanates
could be transformed to the metastable monoclinic modifica-
tion of TiO2 (TiO2-B) under calcination.[38] This modification
of TiO2 has a lower density than anatase or rutile and has a
monoclinic unit cell. The structure of TiO2-B is characterized
by a combination of edge- and corner-sharing TiO6 octahedra
forming a structure with channels in which transport and ex-
change of small cations can occur. On the basis of XRD,
TEM, and Li+ exchange studies, Bruce and co-workers[39]
have suggested that the crystal structure of TiO2 nanotubes
produced by alkaline hydrothermal treatment, followed by
acid washing and calcination, corresponds to the structure of
TiO2-B. Indeed, washing of sodium titanate nanotubes with
acid or water results in the formation of the protonated form
of titanate nanotubes. Following drying at increased tempera-
tures, dehydration of solids and formation of TiO2-B nano-
tubes could occur, the product having a density in the range
3.64–3.76 g cm–3,[40] which is less than the density of anatase or
rutile (3.9 and 4.25 g cm–3, respectively). The XRD pattern of
TiO2-B crystals, however, is slightly different from that of
TiO2 nanotubes, especially at small 2h values (Table 3). As
our understanding of the exact structure of titanate nanotubes
continues to evolve, it is common to see both ‘titanate nano-
tubes’ and ‘TiO2 nanotubes’ used as terms to describe these
materials.
2812 www.advmat.de © 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Recent analysis of pre-edge structure in (Ti K-edge) X-ray
absorption near-edge structure (XANES) spectra of titanate
nanotubes demonstrated that almost 40 % of the Ti atoms are
undercoordinated rather than occupying symmetrical octahe-
dral positions.[41] These undercoordinated centers are allo-
cated on both the surfaces of nanotubes as well as between
the layers inside the walls of the nanotubes.
During alkaline hydrothermal treatment of TiO2 under spe-
cific conditions the formation of fibrous instead of tubular
nanostructures can occur. Usually, the crystal structure of
these materials corresponds to the structure of layered pro-
tonated polytitanates H2TinO2n + 1 (n = 5,[42,43] 6,[26] 8[42]) or
TiO2-B.[44] The calcination of titanate nanofibers results in
consecutive transformation from titanate to TiO2-B (at
400 °C), then to anatase (at 700 °C) and rutile (at 1000 °C).
Nanofibrous morphology disappears at 1000 °C.[45]
All of the proposed crystal structures above have the fol-
lowing properties in common. Firstly, this is a well-defined,
layered structure with a relatively large interlayer distance
(d100) ca. 0.7–0.8 nm resulting from the observation of a char-
acteristic reflection (200) in the XRD patterns at small 2h val-
ues of ca. 10°. Secondly, an atom of hydrogen situated in these
interlayer cavities could be exchanged with alkali metal ions.
Thirdly, the layers of the (100) plane consist of edge- and
corner-sharing TiO6 octahedra building up to zigzag struc-
tures. We propose that a typical sample of TiO2 nanotubes
produced using hydrothermal treatment of initial titania with
NaOH(aqueous), followed by proton exchange, could be a mix-
ture of polytitanic acids with trititanic acid H2Ti3O7 as the
main component. When protonated titanate nanotube sam-
ples were calcined at moderate temperatures (less than
450 °C) the appearance of the TiO2-B phase is then likely. The
phase transformation of titanate nanotubes and their stability
will be discussed later.
3.2. Texture and Morphology
It is unfortunate that the morphological forms of the mate-
rials have not been carefully defined and used in the litera-
ture; some confusion has resulted. In Figure 3, four different
morphologies of titanates, observed during synthesis of tita-
nate nanotubes by alkaline hydrothermal treatment, are de-
fined. The proposed morphologies are consistent with re-
cently suggested classifications.[46]
Nanotubes (Fig. 3 I) have a distinguished geometry, that is,
long cylinders with a hollow cavity lying though the centre
along their length. The aspect ratio (i.e., length divided by di-
ameter) of nanotubes is usually greater than ten and can
achieve several thousands. The walls of titanate nanotubes are
always multilayered and the number of layers varies from two
to ten. Structurally, nanotubes can be scrolled, “onion”, or
concentric in type. Sometimes, the single nanotube has a dif-
ferent number of layers in the two different walls, as seen by
analysis of axial cross sections of the tube obtained using
TEM imaging (Fig. 4a). The nanotubes are usually straight
Adv. Mater. 2006, 18, 2807–2824
D. V. Bavykin et al./ Protonated Titanates and TiO2 Nanostructured Materials
Figure 3. Four different morphologies observed during alkaline hydro-
thermal reaction of TiO2. Nanotubes (I), nanosheets (II), nanorods or
nanowires (III), and nanofibers, nanoribbons, or nanobelts (IV). a, b and
c are crystallographic axes.
with a relatively constant diameter. However, small amounts
of tubes with a variable internal diameter and closed ends are
also found (Fig. 4c).
Nanotubes can be produced by folding nanosheets
(Fig. 3 II). There are two types of nanosheets: single-layer
nanosheets, which are the isolated (100) planes of titanates or
Figure 4. Electron microscopy images of different morphologies observed during alka-
line hydrothermal treatment of TiO2 [49], a) nanotubes, b) nanosheets, c) a closed
nanotube, and d) nanofibers (arrow indicates the rectangular base). (a) and (b) repro-
duced with permission from [49]. Copyright 2004 Royal Society of Chemistry.
Adv. Mater. 2006, 18, 2807–2824 © 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
REVIEW
multilayer nanosheets, which are several conjugated (100)
planes of titanates. Both types of nanosheets are very thin and
could be found in planar or curved shapes. The typical dimen-
sions of nanosheets are less than 10 nm in thickness and more
than 100 nm in height and width. Nanosheets are usually ob-
served in the early stage of preparation or as a small impurity
in the final product (Fig. 4b).
Nanowires or nanorods (Fig. 3 III) are long, solid cylinders
with a circular base; nanowires are longer then nanorods.[46]
Both morphologies do not usually have internal layered struc-
tures and have a similar aspect ratio to nanotubes. Nanowires
can often be found in samples of nanotubes annealed at tem-
peratures higher than 400 °C.[27]
The long, solid, parallelepiped titanates are named nanorib-
bons, nanobelts, or nanofibers in the literature (Fig. 3 IV).
These structures tend to have good crystallinity and usually
the relation between the length of the edges corresponding to
each crystallographic axis is in the order l001 >> l100 > l010.[43]
The length of nanofibers (l001) can be several tens of microm-
eters, while the width of nanofibers (l001 or l010) is typically in
the range 10–100 nm. The aspect ratio can be as large as
several thousand. Nanofibers, which are usually produced dur-
ing alkaline hydrothermal reactions at high temperatures
(Fig. 4d), can be found in straight as well as in curved forms.
During hydrothermal treatment, individual morphological
forms of titanates tend to agglomerate into secondary parti-
cles. The resulting textures include nanotubular bundles,[49]
split nanofibers, hierarchical linked nanofibers,[47] and so
forth. Unfortunately, there are few reported systematic stud-
ies which allow an integrated treatment of the reasons for pro-
ducing a given texture. Below, reported studies on the effects
of synthesis conditions on the texture of titanate
nanostructures are presented. The average length
of nanotubes can be increased by an order of mag-
nitude under intensified mass-transport conditions
in the batch reactor during hydrothermal synthe-
sis.[48] The size of secondary particles (agglomer-
ates of nanotubes) can also be affected by the ratio
of TiO2(s) to NaOH(aqueous) during synthesis.[49] A
degree of control over the shape of the nanotubu-
lar agglomerates can be gained by using hydrogen
peroxide.[50]
3.3. The Mechanism of Hydrothermal Growth
The key to developing and exploiting new nano-
structured materials lies in an increased under-
standing of how synthesis conditions affect proper-
ties of nanostructured materials in order to tailor
materials to specific needs. In particular, a knowl-
edge of the mechanism of nanostructure formation
is important.
Since the discovery of the wet chemical method
of transformation of raw TiO2 to nanotubular tita-
nates, many attempts have been made to describe
www.advmat.de 2813
 D. V. Bavykin et al./ Protonated Titanates and TiO2 Nanostructured Materials
REVIEW

the mechanism and to rationalize these transforma-

tions. Current conceptions of the mechanism for the
formation of nanotubular titanates are reviewed
below. Originally Kasuga et al.[24] considered that
a nanotubular morphology occurred during post
hydrothermal acid washing. Some researchers still
support this suggestion[51] but it was clearly demon-
strated,[52,53] by washing of samples with ethanol or
acetone, that nanotubular sodium titanates are
formed during alkaline hydrothermal processing.
It has proved impossible to adapt the concept of
catalytic growth of hollow fibers on metal nanopar-
ticles of catalyst where the diameter of the fiber
corresponds to the diameter of the particle, like in
the preparation of carbon nanotubes. Indeed, dur-
ing the alkaline hydrothermal synthesis of titanate
nanotubes, there is no possible candidate to play
the role of a size-determining nanoparticle. Unlike
carbon nanotubes produced by catalytic pyrolysis
of hydrocarbons, titania nanotubes produced via
the alkaline hydrothermal method were never ob-
served in a single layered form. All reports of TiO2
nanotubes describe these samples as multilayered
wall, nanotubular structures.
Although the detailed sequence of events is still
unclear, it is clear that during transformation of
TiO2 (anatase, rutile, or amorphous) under alka-
line conditions, the observed intermediate single- Figure 5. The driving forces for bending titanates nanosheets under alkaline hydrother-
mal conditions. a) Asymmetrical chemical environment resulting in difference in sur-
layer and multilayered titanates nanosheets play a face tensions [55,56]; k1 and k2 are spring constants on each side of the nanosheet.
key role in the formation of tubular morphol- b) Imbalance in layer widths resulting in shifting of the layer and bending of nano-
ogy.[49,54] These nanosheets can scroll or fold into sheets. Reproduced with permission from [49]. Copyright 2004 Royal Society of Chemis-
a nanotubular morphology. The driving force for try. c) Stabilization of curved nanosheets due to the periodic potential of the lattice.
curving these nanotubes has been considered by
several groups.
Zhang et al.[55] considered that single surface layers experi- elasticity of the nanosheet and the imbalance in surface ten-
enced an asymmetrical chemical environment, due to the im- sions, respectively. The detailed analysis of structural changes
balance of H+ or Na+ ion concentration on two different sides of nanosheets and ab initio density functional theory (DFT)
of a nanosheet, giving rise to excess surface energy, resulting calculations of total energy of curved fragment of trititanate
in bending. The system could be presented as a plane with two nanosheets as a function of curvature radius[56] demonstrates
springs on each side parallel to this plane (crystallographic c- that the optimal radius of nanotubes is ca. 4 nm and the num-
and b-axis, see Fig. 5a). When both sides have a symmetrical ber of layers in the final titanate nanotube is four. This, how-
chemical environment, both spring constants have similar val- ever, contradicts the recently observed dependence of the na-
ues. As a result, all tensions are compensated and the plane is notube average diameter on the conditions of synthesis, for
straight. When the trititanate nanosheets have a proton-distri- example, temperature or the ratio of TiO2 to NaOH.[49,57]
bution asymmetry, then both sides have different values of Another reason for bending multilayered nanosheets,[49] is
free surface energy (spring constants) and, in order to com- that mechanical tensions arise during the process of dissolu-
pensate imbalance in surface tensions, the plane bends to- tion/crystallization in nanosheets. During spontaneous crystal-
wards the surface with a higher spring constant value. During lization and rapid growth of layers, it is possible that the width
the bending process, strain energy arises and prevents work of different layers varies. It is likely that the imbalance in the
against bending. In a simple approach, the excess energy of layer width creates the tendency of layers to move within the
the system (Elayer) can be expressed as the sum of two terms: multiwalled nanosheet in order to decrease the excess surface
energy (see Fig. 5b). This can result in the bending of multi-
a b
Elayer ˆ 2 (1) layered nanosheets. It was demonstrated that, during simulta-
R R
neous shift of the layer and bending of the nanosheet, the gain
where R is the radius of curvature of the curved nanosheet, a in surface energy is sufficient to compensate for mechanical
and b are the proportionality constants responsible for the tensions arising in the material during curving and wrapping

2814 www.advmat.de © 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Adv. Mater. 2006, 18, 2807–2824
D. V. Bavykin et al./ Protonated Titanates and TiO2 Nanostructured Materials
into nanotubes. The kinetic rate of curving the nanotubes
might control the diameter of the nanotubes produced, which
can be adjusted by varying the condition of synthesis. This ex-
planation of the reason for curving nanosheets is not applic-
able, however, to single-layered nanosheets, which remain un-
folded in aqueous solution.[54,58]
In the absence of factors which stabilize the bended form of
nanosheets, the reverse process of transformation of bended
nanosheets back to the planar form should occur. While it is
difficult to find stabilization factors for a curved single-layer
nanosheet (and unless an asymmetrical environment is main-
tained), the curved form of multilayered nanosheets could be
stabilized by periodic potentials in the crystal lattice. This ef-
fect can be illustrated using the example of two nanosheets
curved into a semicircle (Fig. 5c). Taking the internal diame-
ter of nanotubes as 3.5 nm, the distance between layers in the
wall as 0.72 nm, and the distance between Ti atoms along the
c-axis as 0.919/3 nm, the number of Ti atoms in the first layer
will be 18 and the number of Ti atoms in the second layer will
be 25. In this case, 8 from 18 atoms in the first layer will be
situated in the same radial lines with the other eight atoms
from the second layer. These lines are shown in Figure 5c for
clarity. The distance between these eight pairs of atoms is
minimal and all their positions are very close to the positions
in the crystal. Any small unbending of nanosheets back will
increase the energy of the system since these eight pairs start
to increase the distance between atoms. This means that cur-
ving of multilayered nanosheets to nanotubes passes through
a sequence of metastable states that are stabilized by a period-
ic potential in the nanosheet lattice.
It is remarkable that, starting from colloidal single-layered
isolated trititanate nanosheets, titanate nanotubes can be
readily produced in alkaline conditions at room tempera-
tures.[54] In contrast, starting from bulk layered sodium tritita-
nate Na2Ti3O7 in hydrothermal alkaline conditions the forma-
tion of nanotubes was not observed.[48] However, after several
days of hydrothermal treatment (without further additions of
NaOH) at temperatures of 140–170 °C the resultant titanate
nanotubes are characterized by very wide diameters (several
tens of nanometers).[59] Not only is the presence of layered ti-
tanates important but so is their morphology.
When the temperature during alkaline hydrothermal treat-
ment of TiO2 exceeds 170 °C, the formation of titanate nanofi-
bers occurs.[42,49] At the same time, the solubility of dissolved
TiIV in 10 mol dm–3 NaOH solution positively depends on tem-
perature, a typical value being 0.0047 mol dm–3 at 170 °C (the
enthalpy of dissolution is ca. 23 kJ mol–1).[49] Hydrothermal
treatment of TiO2 in the presence of a concentrated solution of
KOH always transforms titania to titanate nanofibers over a
wide range of temperatures.[60,61] The concentration of dis-
solved TiIV in 10 mol dm–3 KOH solution also positively de-
pends on temperature but is several times higher than in
NaOH solutions. The alkaline treatment of titanate nanotubes
at temperatures higher then 170 °C results in the formation of
titanate nanofibers[36] but there is no data about the transfor-
mation of nanofibers back to nanotubes at lower temperatures.
Adv. Mater. 2006, 18, 2807–2824 © 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
REVIEW
It is interesting to note that the axis of nanotubes (vector
[010]) does not always coincide with the axis of nanofibers
(vector [001]). This provides an insight into the mechanism of
titanate nanotube growth. The fact that nanofibers crystallize
preferably along the crystallographic c-axis suggests that the
rate of dissolution/crystallization along this axis is maximized.
Under certain conditions, an imbalance in nanosheet width
is expected along the c-axis (Fig. 5b). Thus, the bending of
nanosheets will occur around the b-axis. When the curved
nanosheet closes the loop (a nanoloop[48] or rather short
tubes), the direction of the fastest growth will disappear.
There will be only two directions for nanotube growth,
namely, the radial direction (along the crystallographic a-axis)
and the axial direction (along the crystallographic b-axis). Ku-
kovecz et al.[48] propose that the nanoloops provide seeds for
further growth of nanotubes and the preferable direction of
growth is the b-axis. If the nanosheet rollup to tubes of conical
shape occurs, then further growth will result in the formation
of closed-end nanotubes (Fig. 4c). At higher temperatures,
bending of nanosheets does not occur and the resulting nano-
fibers are long in the crystallographic direction c.
Overall, the process of transformation of raw TiO2 to nano-
tubular titanate can be considered to take place in several
stages: i) partial dissolution of raw TiO2 accompanied by
epoxial growth of layered nanosheets of sodium tritinatates,
ii) exfoliation of nanosheets, iii) crystallization of dissolved ti-
tanates on nanosheets resulting in mechanical tensions, which
induce the curving and wrapping nanosheets to nanotubes,
iv) growth of nanotubes along the length, and v) exchange of
sodium ions to protons.
4. Physicochemical Properties
Since the first report on the synthesis of titanate nano-
tubes,[23] many efforts have been made to characterize the parti-
cles. Most studies have focussed on the crystal structure and the
possible mechanism of nanotube formation. In order to deter-
mine possible applications of nanotubes, systematic studies of
other characteristics are necessary. Systematic studies of physi-
cochemical properties have only recently been reported.[27,49]
4.1. Electronic Structure and Optical Properties
Titanate nanotubes are wide-bandgap semiconductor mate-
rials. From studies of the optical properties of aqueous col-
loids of titanate nanotubes at room temperature, the bandgap
has been estimated to be ca. 3.87 eV.[62] This value is very
close to the bandgap of ca. 3.84 eV in 2D titanate nano-
sheets[63] but much higher then the value of ca. 3.2 eV in bulk
TiO2. The absorption spectrum of titanate nanotubes shows
several features resulting from the appearance of discrete lev-
els above the bandgap due to the size quantization. The ener-
gy of these absorption bands is shown in Table 4 as a photolu-
minescence excitation band.
www.advmat.de 2815
 D. V. Bavykin et al./ Protonated Titanates and TiO2 Nanostructured Materials
REVIEW
Table 4. Spectroscopy data on titanate nanotubes.
Spectroscopy
1
H MAS NMR, chemical –0.149 2.341 4.781 6.124
shift, d [ppm] [a]
PL excitation band [eV] 5.22 4.84 4.64 4.44
PL emission band [eV] 3.99 3.77 3.54 3.09
FTIR [cm–1] 900 1370 1630 3400
Raman [cm–1] 290 388 448 668
[a] MAS: magic-angle spinning.
The photoluminescence spectrum of titanate nanotubes
(Table 4) shows many bands in the short-wavelength region
(shallow traps transitions) as well as in the long-wavelength
region (deep traps transitions). The exact nature of these ra-
diative transitions has not been determined. Surprisingly, the
energy position for most of the transitions coincides with lu-
minescence transitions of single or multilayered titanate nano-
sheets[64] demonstrating apparent 2D behavior of 1D tubes at
room temperatures. This phenomenon has been related to the
large effective mass of electrons and holes in TiO2 semicon-
ductors (me = 9 m0 and mh = 3 m0) resulting in small quantum
size effects making it difficult to detect the difference in elec-
tronic structure between nanotubes and nanosheets at room
temperature. This is also consistent with the observation that
the energy position and intensity of radiative transition on ti-
tanate nanotubes measured at room temperature do not de-
pend on the average diameter of the nanotubes.[62]
4.2. Electrical Conductivity, Thermal Conductivity, and
Magnetic Susceptibility
For carbon nanotubes, the electrical conductivity strongly
depends on chirality and can vary over several orders of mag-
nitude. This can result in the properties of the nanotubes tend-
ing to change from those of a conductive metal to those of a
semiconductor. For titanate nanotubes, the existence of labile
protons suggests that this material should have pronounced
proton conductivity. Using AC impedance spectroscopy, Zhou
and co-workers[65] demonstrated that proton conductivity
dominated at temperatures lower than 130 °C in air in a
pressed powder pellet of titanate nanotubes. The value of con-
ductivity was ca. 5.5 × 10–6 S cm–1 at 30 °C and increased with
temperature to 1.5 × 10–5 S cm–1 at 130 °C. A further increase
in temperature resulted in removal of physiadsorbed water
from the pores of the nanotubes, which coincides with the dra-
matic fall in conductivity to a value of ca. 5.6 × 10–8 S cm–1.
This small conductivity represents the electron conductivity of
titanate nanotubes. It positively depended on temperature,
with an apparent activation energy of 0.57 eV and represents
a higher value than the conductivity of anatase or rutile nano-
particles. For instance, at a temperature of 225 °C, the conduc-
tivity for titanate nanotubes was ca. 7.9 × 10–7 S cm–1,[65]
whereas for anatase nanoparticles with a diameter of 6 nm,
the value was only ca. 10–9 S cm–1.[66]
2816 www.advmat.de © 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Transitions; lines in spectrum Ref.
7.954 12.054 [76]
[62]
2.94 2.51 2.38 2.16 2.08 1.99 [62]
[27]
827 917 [79, 80, 81]
Heat treatment of titanate nanotubes in vacuum at 100 °C or
more resulted in the appearance of a symmetrical electron spin
resonance signal with a g factor of 2.003.[67] Jin and co-workers
attributed this signal to the single-electron trapped oxygen va-
cancies. Indeed, heating of TiO2 in vacuum can result in slow
removal of lattice oxygen and formation of oxygen vacancies.
This process is also usually accompanied by the rise of absorp-
tion in the visible wavelength range of the absorption spectrum
due to the formation of Ti3+ centres, which can form Magneli
phases (Ti4O7 or Ti5O9), which were not considered in the
work of Zhang et al.[67] From a practical point of view, synthesis
of nanotubes containing significant Magneli phases would be
of great interest since it would dramatically increase the electri-
cal conductivity.[68] Besides, the reverse electrochemical reduc-
tion of Ti4+ to Ti3+ accompanied by the change in color of the
transparent nanotubular titanate film deposited on the electro-
conductive glass can be exploited in electrochromic devices.[69]
By analogy with carbon nanotubes, one would expect to ob-
serve interesting thermoconductive properties of low-dimen-
sional titanate nanotubes. Preliminary studies of the specific
heat of titanate nanotubes have demonstrated that, at temper-
atures lower than 50 K, low-dimensional behavior of acoustic
phonons dominates, resulting in a large enhancement of spe-
cific heat compared to bulk anatase or rutile.[70]
Recent interest in room-temperature ferromagetic semicon-
ductors and magnetic nanometer-sized materials with high as-
pect ratio (motivated by possible application in spin-base
semiconductor devices) have stimulated research in the syn-
thesis and characterization of TiO2 nanotube-based magnetic
materials. It was demonstrated that 4 mol % cobalt-doped ti-
tanate nanotubes prepared by Co2+ doping during hydrother-
mal systhesis[71] have distinctive ferromagnetic properties at
room temperature. The coercivity of this material is ca. 40 Oe
(1 Oe = 1000/4 p A m–1), which is very close to the coercivity
of cobalt(II)-doped titania nanoparticles prepared using a
sol–gel technique (55 Oe).[72] In contrast, the ion-exchanged
Co2+ titanate nanotubes with a ratio Co/Ti = 1:3.5 possess the
antiferomagnetic properties at room temperature, probably
due to the Co–Co interaction.[27]
4.3. Adsorption, Surface Area, and Porosity
Nanotubular materials are mesoporous adsorbents. Adsorp-
tion of nitrogen into the pores of titanate nanotubes has been
Adv. Mater. 2006, 18, 2807–2824
D. V. Bavykin et al./ Protonated Titanates and TiO2 Nanostructured Materials
thoroughly studied at –196 °C.[49,57] It was demonstrated that
molecules of nitrogen can not occupy interlayer spaces inside
the walls of nanotubes resulting in the absence of microporos-
ity. The typical function of pore size distribution determined
using the Barett–Joyner–Halenda (BJH) algorithm has a max-
imum in the range 5–20 nm, depending on the average diame-
ter of the nanotubes. A comparison of average pore diameter
obtained from nitrogen-adsorption data, with the average di-
ameter of nanotubes obtained from TEM data, has shown
that the method of nitrogen adsorption overestimates the
average diameter of the nanotubes, since the method accounts
for the pores not only inside the tubes but also for the pores
between the tubes. These external pores are usually bigger
than the internal pores and the size of these pores depends on
the way the nanotubes have been agglomerated into the bun-
dles, and as a result it can be altered by mechanical forces
such as ultrasound.
The typical value of the total specific pore volume of nano-
tubes is in the range 0.60–0.85 cm3 g–1, while the typical value
of specific surface area determined using the Brunauer–Em-
mett–Teller (BET) algorithm is in the range 150–300 m2 g–1.
Both values weakly depend on the average diameter of the
nanotubes. The adsorption of water into the surface of tita-
nate nanotubes at 25 °C is similar to that of nitrogen[73] and
water molecules do not intercalate between the layers into the
wall of the nanotubes at this temperature. The typical values
of BET specific surface area and BJH specific pore volume
are 293 m2 g–1 and 0.38 cm3 g–1, respectively. The pore size dis-
tribution has a maximum at a pore diameter of 4 nm. Adsorp-
tion of NO2 on the surface of nanotubes and nanofibers has
been studied using electron spin resonance (ESR) spectrosco-
py and it was demonstrated that physiadsorbed NO2 mole-
cules have a different ESR signal to nanotubes and nano-
fibers.[74] Adsorption of hydrogen, however, shows very
different behavior and will be discussed below.
The surface area and porosity of titanate nanofibers pro-
duced at temperatures higher than 170 °C is much lower then
that of nanotubes. The typical value of specific surface area is
ca. 20–25 m2 g–1 and a typical value of specific porosity is less
than 0.1 cm3 g–1. Due to their large aspect ratio, nanofibers
can be used to prepare novel electrorheological fluids.[75]
5. Chemical Properties
5.1. Ion-Exchange Properties
According to the crystal structure of titanate nanotubes,
protons occupy the cavities between the layers of TiO6 octa-
hedra. These protons of titanate nanotubes have very broad
multiple-peak signals in the 1H MAS NMR spectrum[76]
(MAS: magic-angle spinning) (Table 4) due to solid-state in-
homogeneous broadening. The open morphology of the nano-
tubes results in effective ion-exchange properties.[27] HRTEM
studies of cesium-saturated titanate nanotubes have revealed
that the alkaline ions in the titanate framework[77] are posi-
Adv. Mater. 2006, 18, 2807–2824 © 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
REVIEW
tioned inside the walls of the nanotubes between the layers.
The intercalation of alkaline ions (Li+, Na+, K+, Rb+, Cs+)[77]
and some transition-metal ions (Co2+, Ni2+, Cu2+, Zn2+,
Cd2+)[27] between the layers in the multilayered nanotubes
does not affect the interlayer distance, indicating that the ri-
gidity of the titanate framework is maintained during ion ex-
change.
The maximum reported ion-exchange capacity for sodium
ions, defined as a ratio between sodium and titanium atoms
in the nanotubes (Na/Ti ratio), varies from 0.67[78] to 1.1.[34]
The first number was obtained from analysis of titration
curves for suspended TiO2 nanotubes in aqueous solution by
0.05 mol dm–3 NaOH, where the nanotubes behave like a
weak acid. The other number was evaluated by chemical anal-
ysis of the amount of sodium in the sample of fresh titanate
nanotubes taken directly from a 10 mol dm–3 NaOH solution
and washed with ethanol. However, there is no evidence that
the last Na/Ti ratio corresponds to a reversible process since
washing with water up to pH 12 results in a smaller ratio
(0.81).[34] If it is assumed that the crystal structure of titanates
corresponds to the structure of trititanate[30] then the value
0.67 corresponds to the complete ion exchange of titanate
protons to the sodium ions in the reaction:
x Men+ + H2Ti3O7 > MexH2–xTi3O7x(n–1)+ + x H+ (2)
The fact that all protons in titanate nanotubes are exchange-
able to alkaline ions in solution means that the ion transport
in the interlayer space along the length of the nanotubes is
very effective. This is also confirmed by the fast establishment
of equilibrium between ions in the nanotubes and ions in solu-
tion.
The degree of ion exchange of precious metals to the pro-
tons of titanate nanotubes is less then that for alkaline metals.
The isotherm of cation adsorption in titanate nanotubes has
the Langmuir type (see Fig. 6a and b) and the maximum
Me/Ti ratio is ca. 0.069, 0.081, and 0.09 for Pd2+, [Au(en)2]3+,
and Ru3+ ions, respectively.[73]
5.2. The Chemistry of the Surface
The surface of titanate nanotubes depends on the method
of study. For alkaline ions in aqueous solution, all OH groups
on both sides (internal external) of the nanotubes, as well as
between the layers in multilayered nanotubes, are accessible
and all of the material presents a monolayer of surface. For
gaseous nitrogen, however, the interlayer cavities in the nano-
tube wall are not accessible and only the external and internal
side of the nanotubes provide the nanotube surface. In any
case, the titanate nanotube surface has functional OH groups,
which determine its chemistry.
Fourier transform IR spectroscopy usually detects both the
stretching at 3400 cm–1 and binding at 1630 cm–1 vibration of
OH bonds in titanate nanotubes.[27] Raman spectroscopy pro-
duces a set of oscillations (Table 4), the assignment of which
www.advmat.de 2817
 D. V. Bavykin et al./ Protonated Titanates and TiO2 Nanostructured Materials
REVIEW

Ion-exchange molar ratio / mol(M)⋅mol-1(Ti) acids after treatment with sulfuric acid and can catalyze the
reaction of esterification in organic solutions. The acidity of
0.06 7 sulfated titanate nanotubes was estimated to be –12.70 on the
Hammett scale.[82]

Metal loading / % wt.

0.04 4

5.3. The Electrochemistry of Nanostructured Titanates

0.02 2
Due to their moderate electrical conductivity, high surface
area, and good ion-exchange properties, titanate nanostruc-
0.00 0
0.0 0.5 1.0 1.5 2.0 tures have interesting electrochemical properties. For success-
3+ -1
Concentration of Ru / mmol⋅dm ful application of titanate nanostructures in electrochemical
a) applications, a method of immobilization of nanostructures on
the surface of conductive materials, which allows good electri-
Ion-exchange molar ratio / mol(Au)⋅mol-1(Ti)

cal contact, is needed. A suitable approach for controlled

0.08 16
deposition of nanostructured titanates on the surface of the
electrode is electrophoretic deposition of nanosheets[58] or
nanorods[83] from aqueous suspension. Owing to the dissocia-
Gold loading / %wt.

0.06 12

tion of titanate in an aqueous suspension the nanostructures

0.04 8 have a negative zeta potential (e.g., –43 mV at pH 8 for nano-
tubes[84]) and could be electrophoretically deposited on an an-
0.02 4 ode surface. Modification of the zeta potential of nanostruc-
tures in aqueous suspension, by adjusting pH or adding
0.00 0 surfactants, provides the possibility of electrophoretic deposi-
0 1 2 3 4 5 6 7 8 tion of titanate nanostructures on electrode surfaces. The
-1
Concentration of gold / mmol ⋅ dm combination of electrodeposition of metals with electropho-
b) retic deposition of titanate nanostructures would provide a
method for the preparation of novel nanoparticle composite
600
coatings.[85]
500 Composite electrodes containing titanate nanotubes dem-
onstrate very interesting behavior. Preliminary studies of
400
RuO2–TiO2 nanotube composite electrodes[86] show asym-
-1

metric supercapacitor properties even without optimization of

TOF / h

300
200
100
0 of titanate nanotubes demonstrate similar characteristics.[88]
0 1
Figure 6. a) Isotherm of RuIII adsorption onto a suspension of TiO2
nanotubes in water at 25 °C. b) Three isotherms of adsorption of AuIII
from different complexes: [Au(en)2]Cl3 (䊏), [Au(dien)Cl]Cl2 (䊉), H[AuCl4]
(&). c) The dependence of activity of a RuIII–TiO2 nanotube catalyst on
the ruthenium loading for selective oxidation of alcohols [73,78]. (a) and
(c) reproduced with permission from [78]. Copyright 2005 Elsevier. (b) re-
produced with permission from [73]. Copyright 2006 Springer.
is still under dispute.[36,79–81] These features could be inter-
preted as Ti–O–Ti crystal phonons (more likely for 448 cm–1
and 668 cm–1 peaks), surface Ti–O–Na vibrations (more likely
for 917 cm–1 peak), second-order harmonics, or radial
“breathing” oscillations, intrinsic to nanotubular structures.
Despite the fact that titanate nanotubes in aqueous suspen-
sion behave like weak acids, they become strong Broensted
the method of ruthenium deposition. The typical value of ca-
pacitance is greater than 1000 F g–1[87] which can be complete-
ly attributed to RuO2 rather then titanate nanotubes. The
mixed cobalt and nickel hydroxides deposited on the surface
2 3 4 5 6 7 8 9
In all these cases, nanotubes play the role of a conductive, in-
Catalyst loading / % wt.
c) ert support for metal-oxide nanoparticles.
Titanate nanotubular film electrodes also demonstrate
some advantages in immobilization of myoglobin and im-
provement in interfacial electron-transfer reaction rates com-
pared to TiO2 nanoparticulate electrodes.[89] The level of de-
naturing of myoglobin in a titanate nanotubular film is much
less than on the TiO2 nanoparticulate ones and a myoglobin
loading of 15 % can be achieved which may prove useful in
biocatalysts. Titanate nanostructures can bond not only to the
protein but also to cationic redox systems, providing new elec-
trode materials for electrocatalysis and bioelectrocatalysis.[84]
6. Applications
The unique physicochemical properties of nanostructured
titanates, together with their unusual morphology, render

2818 www.advmat.de © 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Adv. Mater. 2006, 18, 2807–2824
D. V. Bavykin et al./ Protonated Titanates and TiO2 Nanostructured Materials
these materials very promising for many applications. Table 5
summarizes potential applications and relates them to the es-
sential properties of nanostructured titanates.
6.1. Catalysis, Electrocatalysis, and Photocatalysis
Titanate nanotubes are of great interest for catalytic pro-
cesses since their high cation-exchange capacity provides the
possibility of achieving a high loading of active catalyst with
an even distribution and a high dispersion. The open mesopo-
rous morphology of nanotubes, the high specific surface area,
and the absence of micropores for reactant molecules facili-
tate transport of reagents to the active sites during the catalyt-
ic reaction. The semiconducting properties of TiO2 nanotubes
result in strong electronic interaction between the support
and a catalyst, improving catalytic performance in redox reac-
tions. The moderate electrical conductivity of nanotubes sti-
mulates the use of this material in electrocatalytic processes.
There are few reports of catalytic activity of titanate nano-
tubes as acid-based catalysts in the reaction of esterification[82]
and hydrolysis of 2-chlorethyl ethylsulfide.[90] Most of the cata-
lytic studies are focused on the use of this mesoporous semi-
conductor as a high-surface-area catalyst support. There are
several examples of TiO2 nanotubes as mesoporous catalyst
supports for different nanoparticles: CdS decorated TiO2
nanotubes[91,92] and Zn2+ doped nanotubes[93] in the reaction of
photocatalytic oxidation of dyes, Pd–TiO2 nanotube cata-
lysts[94] for anodic oxidation of methanol, Pt–TiO2 nanotube
photocatalysts for generation of H2,[95] Au–TiO2 nanotube cat-
alyst for the water-gas shift reaction[96] and CO oxidation,[97]
RuO2–TiO2 nanotubes as an electrocatalyst for reduction of
CO2,[86] and ruthenium(III) hydrated oxide deposited on TiO2
nanotubes for selective oxidation of alcohols[78] (Table 5).
The ability of titanate nanotubes to adsorb metal cations
from aqueous solution provides a strategy for the deposition
of active catalysts on the surface of nanotubes. During adsorp-
tion, the catalyst precursor should be in cationic form in aque-
ous solution. This will allow the achievement of a high value
of ion-exchange ratio and washing out of the impregnating so-
lution in order to avoid further formation of catalyst outside
the surface of the titanate nanotubes. In Figure 6b the effect
of the charge of catalyst precursor on its ion-exchange ratio is
Table 5. The possible applications of protonated titanate nanotubes.
Application Example
Catalyst TiO2, esterification and Hydrolysis
Catalyst support Au/TiO2, CO oxidation, water shift reaction;
Ru(III)/TiO2, selective oxidation of alcohol
Photocatalysis CdS/TiO2, Zn2+/TiO2, Pt/TiO2, oxidation reduction
Electrocatalysis Pd/TiO2, oxidation methanol, RuO2/TiO2, reduction CO2
H2 storage and separation Sorption of hydrogen in pores of TiO2 nanotubes
H2 sensing Array of TiO2 nanotubes produced by anodizing
Lithium batteries Negative electrode for rechargeable Li cells
Photovoltaic cells Dye-sensitized solar cells
Adv. Mater. 2006, 18, 2807–2824 © 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
REVIEW
demonstrated using as an example various complex com-
pounds of AuIII. The negatively charged HAuCl4 precursor
adsorbs weakly on the surface of the nanotubes, whereas posi-
tively charged ethylendiamine (en) and diethylentriamane
(dien) complexes adsorb strongly. In all reported works using
cationic reactants for deposition of nanoparticles on the sur-
face of titanate nanotubes good dispersions of nanoparticles
was observed.[73,91,92]
The combination of ion exchange followed by reactive fixa-
tion of catalyst allows a relatively high catalyst loading which
maintains specific catalytic activity[78] (Fig. 6c) as was demon-
strated in the example of ruthenium(III) hydrated oxide
nanoparticles on the surface of titanate nanotubes for the se-
lective oxidation of alcohols.[78] TEM data (Fig. 7) show that
an increase of ruthenium loading from 0.6 to 8.7 wt % re-
sulted in a higher density of deposited particles, while the
average particle size remained constant. The independence of
catalyst activity and loading is consistent with the indepen-
dence of ruthenium average particle size and loading.
The moderate electrical conductivity of nanotubes, com-
bined with the possibility of deposition of dispersed catalysts,
extends the area of application to electrocatalysis.[84,86,94] The
good electrical contact between nanoparticles and titanate
nanotubes can provide an efficient interfacial charge-transfer
region, which makes this binary system also suitable for
photocatalysis. In photocatalytic processes, the photosensibili-
zer could be either wide-bandgap titanate nanotubes[95] or
narrow-bandgap semiconductor nanoparticles.[91,92]
6.2. Hydrogen Sensing, Storage, and Separation
The ability of titanate nanotubes to reversibly accumulate
molecular hydrogen with a relatively high uptake, over a wide
range of temperatures from –196 to 125 °C,[76,98] opens up the
possibility of hydrogen storage and related applications. The
hydrogen-adsorption isotherm for the TiO2 nanotubes at
–196 °C is shown in Figure 8.[76] The hydrogen uptake is rela-
tively high, almost 1.5 hydrogen molecules per one Ti atom
(ca. 3.8 wt %) can be adsorbed at a 2 bar (1 bar = 100 000 Pa)
pressure of hydrogen. Further increase in the hydrogen pres-
sure does not significantly change the hydrogen uptake. Dur-
ing the desorption of hydrogen, a large hysteresis is observed;
Utilized properties Ref.
Broensted acidity OH groups [82, 90]
Ion-exchange capacity, strong metal-support interaction, [96, 97]
mesoporous morphology [78]
Photocatalytic activity, interfacial charge transfer [91, 92], [93, 95]
Mesoporous morphology, electrical conductivity [94, 86]
Slit pores 0.72 nm, OH groups inside [76, 98]
Electrical conductivity [20]
Effective transport of lithium, electrical conductivity [100–109]
Semiconductor, ion-exchange, electrical conductivity [110, 111]
www.advmat.de 2819
 D. V. Bavykin et al./ Protonated Titanates and TiO2 Nanostructured Materials
REVIEW

even at 0 bar pressure, the uptake of hydrogen achieves a 1.25

molar ratio (point B in Fig. 8). Heating the sample in vacuum
to 200 °C leads to a complete desorption of hydrogen, return-

Number of particles / a.u.

ing the weight of the sample to its initial value (point A in
Fig. 8). Such behavior indicates that adsorption of hydrogen is
a relatively reversible process that occurs without formation
of stable hydrides (this would lead to point B being above
point A) and in the absence of redox processes followed by
the elimination of lattice oxygen (this would lead to point B
0.0 0.5 1.0 1.5 2.0
Particle size / nm
2.5 3.0 being below point A).
a) Physical adsorption of hydrogen on the internal and the ex-
ternal surface of titanate nanotubes alone could not provide
such a high value of observed hydrogen uptake. Besides, the
kinetic rate of hydrogen sorption is relatively small with a
Number of particles / a.u.

characteristic time of sorption in the range of a few hours. All

of these observations suggest that hydrogen can occupy inter-
stitial cavities between layers in the wall of nanotubes without
chemical-bond formation. Indeed, the size of interstitial cav-
ities—zigzag slit pores formed between two (100) planes, is
ca. 0.72 nm (see Fig. 2 III), which is larger than the dynamic
0.0 0.5 1.0 1.5 2.0 2.5 3.0 diameter of hydrogen molecules (0.41 nm) and much larger
Particle size / nm
than the nuclear distance in the hydrogen molecule (0.07 nm).
These cavities can also accommodate relatively large cations
of Cs.[77] OH groups in the nanotube lattice could stabilize the
hydrogen molecules via weak van der Waals interactions, re-
sulting in the formation of TiO2 xH2 clathrates (c.f. the hydro-
Number of particles / a.u.

gen clathrate hydrate (32 + x)H2 136H2O).[99]

0.0
Figure 7. TEM images of ruthenium(III) hydrated oxide nanoparticles de-
posited onto TiO2 nanotubes and histograms of particle size distribution
for different metal loadings: a) 1.1 wt %, b) 3.4 wt %, and c) 8.7 wt %
[73]. Reproduced with permission from [73]. Copyright 2006 Springer.
3.5
Molar ratio / mol(H)⋅mol (TiO2)
The rate of hydrogen intercalation increased with tempera-
ture but steady-state hydrogen uptake decreased with temper-
ature. The apparent activation energy and enthalpy of clath-
rate formation are 44 and –30 kJ mol–1, respectively.[76] The
rate of hydrogen sorption does not depend on hydrogen pres-
0.5 1.0 1.5 2.0 2.5 3.0
Particle size / nm
sure in the range 0.4–1 bar. The limiting stage in the process
of hydrogen sorption is the diffusion of hydrogen inside the
wall of the nanotubes along their length.
In contrast to carbon nanotubes or metal-alloy hydrides, ti-
tanate nanotubes can also operate over a convenient range of
pressures and temperatures. Moreover, simple pressure and
temperature swings can be used to adsorb and desorb hydro-
4.4 gen from the solid-state, nanotubular titanates. Such a selec-

tiveness of titanate nanotubes for sorption of hydrogen can be

3.0 3.8
used in the design of the membranes for separation of hydro-
-1

Hydrogen uptake / % wt

2.5 B 3.1 gen from other gases. This could possibly find application in
various industrial processes, such as the water-shift reaction.
2.0 2.5
Another interesting property of TiO2 nanotubes produced
1.5 1.9 using anodization is that their electrical resistance decreases
1.0 1.3 by several orders of magnitude in the presence of sufficient
hydrogen.[20] If titanate nanotubes possess similar behavior,
0.5 0.6
A
then it might be possible to manufacture integrated systems,
0.0 0.0 which can self-control the amount of intercalated hydrogen.
0 1 2 3 4 5 6
H2 pressure / bar
6.3. Lithium Batteries
Figure 8. Isotherm for (䊏) hydrogen sorption into and (䊊) desorption
out of the pores of TiO2 nanotubes at –196 °C. The left axis represents
the amount of hydrogen per unit amount of TiO2 [76]. Reproduced with Nanostructured titanates attract attention as a possible neg-
permission from [76]. Copyright 2005 American Chemical Society. ative electrode for rechargeable lithium batteries owing to

2820 www.advmat.de © 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Adv. Mater. 2006, 18, 2807–2824
D. V. Bavykin et al./ Protonated Titanates and TiO2 Nanostructured Materials
their open, mesoporous structure, efficient transport of lithi-
um ions, and good ion-exchange ratio, which results in a high
value of charge/discharge capacity (< 300 mA h g–1), good ki-
netic characteristics, and very good robustness and safety.
These new electrodes can substitute commercial carbon nega-
tive electrode batteries, which suffer from safety concerns
(electroplating of lithium) and the formation of a solid-elec-
trolyte interface (SEI) layer consuming the charge. The elec-
trode reaction on nanostructured electrodes includes interca-
lation of lithium ions according to the reaction:
xLi+ + TiO2 + x e– > LixTiO2 (3)
where x is the lithium insertion coefficient. The simplified
product formula does not represent the crystal structure of
TiO2 or titanates but it is convenient to express the value of x.
The nature of LixTiO2 intercalate is unclear (e.g., oxidative
state of Ti of Li) but it is suggested that lithium atoms accom-
modate positions of ion-exchanged protons in titanate nano-
structures or channels in TiO2-B nanofibers.[100,101]
Recent studies[39,100,102] on the use of TiO2 nanotubular an-
odes in a lithium battery demonstrate an initial 282 mA h g–1
discharge capacity at a specific current density of 0.24 mA g–1.
This corresponds to x = 0.98 in Equation 3, which is larger
then the value of ca. 0.5[103] for an anatase electrode. In a typi-
cal experiment, the working electrode was a pressed pellet of
75–80 % nanotubular titanates, 10–15 % carbon conductive
materials, and 5–15 % binder. The electrolyte composition
was a 1 mol dm–3 solution of LiPF6 or LiN(CF3SO2)2 in a (1:1
volume) mixture of ethylene carbonate and dimethylcarbo-
nate. The counter and reference electrodes were lithium met-
al. The cyclic voltammogram usually shows a pair of wide
cathode/anode peaks at 1.69 and 2.08 V relative to a lithium
electrode at a scan rate 0.05 mV s–1. Electron microscopy
studies have revealed that morphology on titanate nanotubes
remains unchanged after 50 charge/discharge cycles.[104]
Nanofibrous titanate materials have been more extensively
studied than nanotubular ones for lithium-battery applica-
tions.[38,101,105] These materials include protonated titanate
nanofibers[106] produced by alkaline hydrothermal treatment
followed by acid washing, spinel Li4Ti5O12 nanofibers[107] pro-
duced from protonated titanate nanofibers by hydrothermal
ion exchange, nanofibers of TiO2-B,[39,44,108] and titanate nano-
fibers calcined at 400 °C.[109] The typical cyclic voltammogram
of titanate nanofibers shows several pairs of peaks in the
range of voltage 1.5 to 2.0 V relative to Li+/Li. The amplitude
of some of these peaks lineally depends on the potential
sweep rate, whereas the amplitude of others has a square-root
dependence on sweep rate,[38,105] indicating the pseudocapaci-
tative and external diffusion-controlled nature of processes.
The lithium insertion coefficient, x varies from 0.71[106] to
0.91[101] for nanofibrous structures.
The current tendency in research is to switch from nanotub-
ular to nanofibrous titanate nanostructures in lithium battery
applications, since the latter structure type retains a better ca-
pacity retention on cycling.[105,106] However, nanotubes still
Adv. Mater. 2006, 18, 2807–2824 © 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
REVIEW
have some advantages (see Fig. 9); they have higher surface
area (ca. 200 m2 g–1 against ca. 20 m2 g–1) and better ion-ex-
change properties (Na/Ti ratio is ca. 0.67 against ca. 0.41)
Lithium insertion, x / dimensionless
0.0 0.1 0.3 0.4 0.6 0.7 0.9 1.0
2.4
2.2
Potential vs. Li /Li / V
2.0
+
1.8
1.6
1.4
4
1.2
1.0
3 2 1
0.8
0 50 100 150 200 250 300 350
Discharge capacity / mA h g
-1
Figure 9. Discharge curve of electrodes prepared from different titanium
oxide nanostructured materials. 1) Titanate nanotubes, current 10 mA g–1
[39]; 2) TiO2-B nanofibers, current 10 mA g–1 [39]; 3) Li4Ti5O12 spinel fi-
bers, current 40 mA g–1 [107]. Degussa Ti-HV-2, current 8 mA g–1 [103].
then nanofibers. The dimension of nanotubes is ca. 5 nm di-
ameter and several hundreds of nanometers in length,
whereas for nanofibers it is several tenth of nanometers diam-
eter and several micrometers in length. This should affect the
efficiency of diffusion of lithium ions in the titanate nano-
structures. The mechanism of lithium-ion diffusion in solid ti-
tanate nanostructures is also unclear. Does it occur only in the
axial direction through the channels (TiO2-B) and interlayer
cavities (tubes) or also in the radial direction? The electrical
conductivity of nanotubes and nanofibers has not been sys-
tematically studied and the method of immobilization of
nanostructures has not been optimized.
6.4. Solar Cells
Nanotubular titanate films were also examined as an elec-
trode of dye-sensitized solar cells[110,111] and had the following
characteristics: open-current voltage of 0.704 V, short current
of 1.26 mA, energy conversion efficiency of 2.9 % (7.1 % in
the literature[111]) and fill factor of 0.66 for a 0.25 cm2 cell.
These characteristics are very similar to those from conven-
tional titania nanoparticles (P-25) used for solar-cell technolo-
gy and the authors considered this result as negative. At that
time, however, the following parameters had been chosen,
which minimize the benefit of titanate nanotubes. Firstly, the
negatively charged standard cis-di(thiocyanate)bis(2,2-bipyri-
dyl-4,4-di-carboxylate) ruthenium(II) complex was used as
the dye sensitizer deposited on the surface of titanate nano-
tubes from ethanol solution. In this condition, adsorption of
dye molecules is not optimal. The best results are expected
www.advmat.de 2821
 D. V. Bavykin et al./ Protonated Titanates and TiO2 Nanostructured Materials
REVIEW
using adsorption of positively charged molecules of dye from
aqueous solution. Secondly, the annealing of the acid form of
a titanate nanotube film at 450 °C could result in loss of nano-
tubular morphology and formation of nanorods, with have less
surface area and ion-exchange capacity. Thirdly, as the
authors indicated, the agglomeration of nanotubes into sec-
ondary particles needs to be avoided during the preparation
of nanotubes. It should now be possible to strategically im-
prove the performance of solar cells based on TiO2 nanostruc-
tures by tailoring synthesis conditions to produce optimum
materials.
6.5. The Stability of Titanate Nanostructures
In many applications, titanate nanostructures are exposed
to chemically aggressive media, which can affect their stabil-
ity. The successful application of nanostructured titanates re-
quires data on the range of operational conditions under
which nanotubes are stable. In this section, the stability and
phase transformation of titanate nanostructures at different
temperatures and pH of aqueous suspension is considered.
Titanate nanostructures are thermodynamically metastable
phases, which can undergo further transformation into a more
stable state in the absence of stabilizing factors. The low solu-
bility of titanates in water suspensions is an important factor
in kinetic stabilization during the process of ageing. Any in-
crease in the rate of dissolution at higher temperatures or ad-
dition of chemical agents can facilitate the process of ageing
and decrease the stability of nanotubes.
The generalized scheme of transformation of nanostruc-
tured titanate produced by alkaline hydrothermal treatment
of titania is shown in Figure 10. The thermal stability of tita-
Figure 10. Scheme of transformation of nanostructures.
2822 www.advmat.de © 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
nate nanofibers is very high. Calcination of titanate nanofi-
bers results in a change of the crystal structure, maintaining
nanofibrous morphology almost until 1000 °C.[45] Hydrother-
mal treatment of titanate nanofibers in water at 150 °C results
in the formation of fibrous anatase with a corroded surface
texture.[112]
In contrast to the behavior of nanofibers, nanotubes are
characterized by a worse temperature and acid stability. Un-
fortunately, there is some inconsistency in the literature re-
lated to the thermal stability of titanate nanotubes. This can
be attributed to various researchers analyzing titanate nano-
tubes, which have different diameter, sodium content, history
of preparation, and so forth. Here, we present general tenden-
cies in nanotubular titanate thermal and acid stability. There
is a need for careful, systematic work in this area, since many
applications demand well-characterized, time-stable materi-
als.
The sodium saturated titanate nanotubes are characterized
by better thermal stability[27] and at temperatures higher than
600 °C transform to the sodium titanate nanorods. It is inter-
esting that the diameter of the nanorods can be very close to
the external diameter of the initial nanotubes. This transfor-
mation results in disappearance of nanotube hollows accom-
panied by a decrease of porosity and surface area in the sam-
ple. Similar changes occur with protonated titanate nanotubes
but at lower temperatures. Annealing of the protonated form
of titanate nanotubes in air at temperatures higher than
400 °C results in transformation of the nanotubes and forma-
tion of anatase with a nanorod morphology.[52]
The recent preliminary studies of phase transformation of
titanate nanotubes at elevated temperatures during acid hy-
drothermal treatments have shown that both anatase and ru-
tile polymorphs can be formed in the presence of 1 mol dm–3
nitric acid at temperatures higher than 80 °C, after 48 h of
treatment.[28] The resulting anatase and rutile polymorphs
have nanowire or nanocrystalline morphology. Acidic hydro-
thermal treatment of titanate nanotubes at 175 °C results in
the formation of polycrystalline anatase nanorods.[113] More-
over, titanate nanotubes have poor stability even at room
temperature in diluted inorganic acids and slowly transform
to rutile nanoparticles within several months.[114] The rate of
this transformation depends on the nature of the inorganic
acid and can be correlated with the solubility of titanates in
the acid.
Mechanically, the titanate nanotubes are also fragile struc-
tures and could easily be broken by ultrasonic treatment of
the aqueous suspension resulting in a decrease of the average
length of the nanotubes.[49] All these data of stability of tita-
nate nanostructures narrows the range of conditions at which
they can be used in the long term. The use of alkaline, neutral
conditions, or organic solvents at temperatures less than
400 °C will provide stable conditions for titanate nanotubes.
For example, the stability of a catalyst prepared using titanate
nanotubes in toluene solvent at 120 °C was very good and no
degradation of protonated titanate nanostructures was de-
tected under these conditions.[78]
Adv. Mater. 2006, 18, 2807–2824
D. V. Bavykin et al./ Protonated Titanates and TiO2 Nanostructured Materials
7. Conclusions
The recently emerged nanostructured titanates produced by
the alkaline hydrothermal method provide a wide variety of
possible applications due to the unique combination of physi-
cal and chemical properties of these tiny materials. One of the
advantages of titanate nanostructures is that they can be
cheaply mass produced in a one-pot alkaline hydrothermal
process. There are two major directions for research into
nanotubular titanate materials, namely preparation and appli-
cations.
Preparation: although the mechanism of transformation
and detailed of synthesis have been thoroughly studied, there
are several unanswered questions. Is the mechanism of trans-
formation of layered nanosheets to nanotubes universal and
could the method be used to prepare non-titanate nanotubes
materials (e.g., alumosilicates)? Will an improved knowledge
of the mechanism allow us to control the morphology (diame-
ter, length, texture) of nanotubes and tailor them for specific
needs? There is currently a gap in our knowledge of immobili-
zation of nanotubular materials without loosing their proper-
ties for different applications. The preparation of films,
coatings, and composite materials are targets for further in-
vestigations.
Applications: in this paper, we have identified some of the
promising applications of nanostructured titanate materials.
Other applications, such as lithium titanate nanotubes as a
CO2 adsorbent or fuel-cell electrode, require urgent develop-
ment.
Received: December 16, 2005
Revised: March 31, 2006
Published online: October 10, 2006
– [35] R. Z. Ma, Y. Bando, T. Sasaki, Chem. Phys. Lett. 2003, 380, 577.
[1] A. Fujishima, K. Hashimoto, T. Watanabe, TiO2 Photocatalysis: Fun-
damentals and Applications, BKC, Tokyo 1999.
[2] J. C. Hulteen, C. R. Martin, J. Mater. Chem. 1997, 7, 1075.
[3] S. Kobayashi, K. Hanabusa, N. Hamasaki, M. Kimura, H. Shirai,
S. Shinkai, Chem. Mater. 2000, 12, 1523.
[4] J. H. Jung, H. Kobayashi, K. J. C. Bommel, S. Shinkai, T. Shimizu,
Chem. Mater. 2002, 14, 1445.
[5] G. Gundiah, S. Mukhopadhyay, U. G. Tumkurkar, A. Govindaraj,
U. Maitrab, C. N. R. Rao, J. Mater. Chem. 2003, 13, 2118.
[6] T. Peng, A. Hasegawa, J. Qiu, K. Hirao, Chem. Mater. 2003, 15, 2011.
[7] S. Fujikawa, T. Kunitake, Langmuir 2003, 19, 6545.
[8] C. Hippe, M. Wark, E. Lork, G. Schulz-Ekloff, Microporous Meso-
porous Mater. 1999, 31, 235.
[9] A. Michailowski, D. Al-Mawlawi, G. S. Cheng, M. Moskovits, Chem.
Phys. Lett. 2001, 349, 1.
[10] S. M. Liu, L. M. Gan, L. H. Liu, W. D. Zhang, H. C. Zeng, Chem.
Mater. 2002, 14, 1391.
[11] S.-Z. Chu, K. Wada, S. Inoue, S.-I. Todoroki, Chem. Mater. 2002, 14,
266.
[12] M. Adachi, Y. Murata, I. Okada, S. Yoshikawa, J. Electrochem. Soc.
2003, 150, G488.
[13] S. Ngamsinlapasathian, S. Sakulkhaemaruethai, S. Pavasupree,
A. Kitiyanan, T. Sreethawong, Y. Suzuki, S. Yoshikawa, J. Photo-
chem. Photobiol. A 2004, 164, 145.
[14] D. Gong, C. A. Grimes, O. K. Varghese, W. Hu, R. S. Singh,
Z. Chen, E. C. Dickey, J. Mater. Res. 2001, 16, 3331.
[15] L. V. Taveira, J. M. Macak, H. Tsuchiya, L. F. P. Dick, P. Schmuki, J.
Electrochem. Soc. 2005, 152, B405.
Adv. Mater. 2006, 18, 2807–2824 © 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
REVIEW
[16] C. Ruan, M. Paulose, O. K. Varghese, G. K. Mor, C. A. Grimes, J.
Phys. Chem. B 2005, 109, 15 754.
[17] A. Ghicov, H. Tsuchiya, J. M. Macak, P. Schmuki, Electrochem.
Commun. 2005, 7, 505.
[18] H. Tsuchiya, J. M. Macak, L. Taveira, E. Balaur, A. Ghicov, K. Sirot-
na, P. Schmuki, Electrochem. Commun. 2005, 7, 576.
[19] J. M. Macak, K. Sirotna, P. Schmuki, Electrochim. Acta 2005, 50,
3679.
[20] O. K. Varghese, D. Gong, M. Paulose, K. O. Ong, E. C. Dickey,
C. A. Grimes, Adv. Mater. 2003, 15, 624.
[21] G. K. Mor, M. A. Carvalho, O. K. Varghese, M. V. Pishko, C. A.
Grimes, J. Mater. Res. 2004, 19, 628.
[22] G. K. Mor, K. Shankar, M. Paulose, O. K. Varghese, C. A. Grimes,
Nano Lett. 2005, 5, 191.
[23] T. Kasuga, M. Hiramatsu, A. Hoson, T. Sekino, K. Niihara, Lang-
muir 1998, 14, 3160.
[24] T. Kasuga, M. Hiramatsu, A. Hoson, T. Sekino, K. Niihara, Adv. Ma-
ter. 1999, 11, 1307.
[25] Y. Lan, X. Gao, H. Zhu, Z. Zheng, T. Yan, F. Wu, S. P. Ringer,
D. Song, Adv. Funct. Mater. 2005, 15, 1310.
[26] X.-D. Meng, D.-Z. Wang, J.-H. Liu, S.-Y. Zhang, Mater. Res. Bull.
2004, 39, 2163.
[27] X. Sun, Y. Li, Chem. Eur. J. 2003, 9, 2229.
[28] H. Y. Zhu, Y. Lan, X. P. Gao, S. P. Ringer, Z. F. Zheng, D. Y. Song,
J. C. Zhao, J. Am. Chem. Soc. 2005, 127, 6730.
[29] W. Z. Wang, O. K. Varghese, M. Paulose, C. A. Grimes, Q. L. Wang,
E. C. Dickey, J. Mater. Res. 2004, 19, 417.
[30] G. H. Du, Q. Chen, R. C. Che, Z. Y. Yuan, L.-M. Peng, Appl. Phys.
Lett. 2001, 79, 3702.
[31] Q. Chen, G. H. Du, S. L.-M. Zhang Peng, Acta Crystallogr. Sect. B
2002, 58, 587.
[32] D. Wu, J. Liu, X. Zhao, A. Li, Y. Chen, N. Ming, Chem. Mater. 2006,
18, 547.
[33] A. Nakahira, W. Kato, M. Tamai, T. Isshiki, K. Nishio, H. Aritani, J.
Mater. Sci. 2004, 39, 4239.
[34] J. J. Yang, Z. S. Jin, X. D. Wang, W. Li, J. W. Zhang, S. L. Zhang,
X. Y. Guo, Z. J. Zhang, Dalton Trans. 2003, 3898.
[36] R. Z. Ma, K. Fukuda, T. Sasaki, M. Osada, Y. Bando, J. Phys. Chem.
B 2005, 109, 6210.
[37] Y. Kubota, H. Kurata, S. Isoda, Mol. Cryst. Liq. Cryst. 2006, 445, 107.
[38] R. Marchand, L. Brohan, M. Tournoux, Mater. Res. Bull. 1980, 15,
1129.
[39] G. Armstrong, A. R. Armstrong, J. Canales, P. G. Bruce, Chem.
Commun. 2005, 2454.
[40] M. Zukalova, M. Kalbac, L. Kavan, I. Exnar, M. Grätzel, Chem. Ma-
ter. 2005, 17, 1248.
[41] Z. V. Saponjic, N. M. Dimitrijevic, D. M. Tiede, A. J. Goshe, X. Zuo,
L. X. Chen, A. S. Barnard, P. Zapol, L. Curtiss, T. Rajh, Adv. Mater.
2005, 17, 965.
[42] Z.-Y. Yuan, B.-L. Su, Colloids Surf. A 2004, 241, 173.
[43] H. G. Yang, H. C. Zeng, J. Am. Chem. Soc. 2005, 127, 270.
[44] A. R. Armstrong, G. Armstrong, J. Canales, P. G. Bruce, Angew.
Chem. Int. Ed. 2004, 43, 2286.
[45] S. Pavasupree, Y. Suzuki, S. Yoshikawa, R. Kawahata, J. Solid State
Chem. 2005, 178, 3110.
[46] Y. Ding, Z. L. Wang, J. Phys. Chem. B 2004, 108, 12 280.
[47] Z.-Y. Yuan, W. Zhou, B.-L. Su, Chem. Commun. 2002, 1202.
[48] A. Kukovecz, M. Hodos, E. Horvath, G. Radnoczi, Z. Konya, I. Kir-
icsi, J. Phys. Chem. B 2005, 109, 17 781.
[49] D. V. Bavykin, V. N. Parmon, A. A. Lapkin, F. C. Walsh, J. Mater.
Chem. 2004, 14, 3370.
[50] Y. Mao, M. Kanungo, T. Hemraj-Benny, S. S. Wong, J. Phys. Chem.
B 2006, 110, 702.
[51] C.-C. Tsai, H. Teng, Chem. Mater. 2006, 18, 367.
[52] Q. Chen, W. Zhou, G. Du, L.-M. Peng, Adv. Mater. 2002, 14, 1208.
www.advmat.de 2823
 D. V. Bavykin et al./ Protonated Titanates and TiO2 Nanostructured Materials
REVIEW
[53] M. Zhang, Z. Jin, J. Zhang, X. Guo, J. Yang, W. Li, X. Wang,
Z. Zhang, J. Mol. Catal. A 2004, 217, 203.
[54] R. Ma, Y. Bando, T. Sasaki, J. Phys. Chem. B 2004, 108, 2115.
[55] S. Zhang, L.-M. Peng, Q. Chen, G. H. Du, G. Dawson, W. Z. Zhou,
Phys. Rev. Lett. 2003, 91, 256 103.
[56] S. Zhang, Q. Chen, L.-M. Peng, Phys. Rev. B 2005, 71, 014 104.
[57] C.-C. Tsai, H. Teng, Chem. Mater. 2004, 16, 4352.
[58] W. Sugimoto, O. Terabayashi, Y. Murakami, Y. Takasu, J. Mater.
Chem. 2002, 12, 3814.
[59] M. Wei, Y. Konishi, H. Zhou, H. Sugihara, H. Arakawa, Solid State
Commun. 2005, 133, 493.
[60] Z.-Y. Yuan, X.-B. Zhang, B.-L. Su, Appl. Phys. A 2004, 78, 1063.
[61] X. Sun, X. Chen, Y. Li, Inorg. Chem. 2002, 41, 4996.
[62] D. V. Bavykin, S. N. Gordeev, A. V. Moskalenko, A. A. Lapkin,
F. C. Walsh, J. Phys. Chem. B 2005, 109, 8565.
[63] N. Sakai, Y. Ebina, K. Takada, T. Sasaki, J. Am. Chem. Soc. 2004,
126, 5851.
[64] T. Sasaki, M. Watanabe, J. Phys. Chem. B 1997, 101, 10 159.
[65] A. Thorne, A. Kruth, D. Tunstall, J. T. S. Irvine, W. Zhou, J. Phys.
Chem. B 2005, 109, 5439.
[66] T. Dittrich, J. Weidmann, V. Y. Timoshenko, A. A. Petrov, F. Koch,
M. G. Lisachenko, E. Lebedev, Mater. Sci. Eng. B 2000, 69–70, 489.
[67] S. Zhang, W. Li, Z. Jin, J. Yang, J. Zhang, Z. Du, Z. Zhang, J. Solid
State Chem. 2004, 177, 1365.
[68] J. R. Smith, F. C. Walsh, R. L. Clarke, J. Appl. Electrochem. 1998,
28, 1021.
[69] H. Tokudome, M. Miyauchi, Angew. Chem. 2005, 117, 2010.
[70] C. Damesa, B. Poudel, W. Z. Wang, J. Y. Huang, Z. F. Ren, Y. Sun,
J. I. Oh, C. Opeil, M. J. Naughton G. Chen, Appl. Phys. Lett. 2005,
87, 031 901.
[71] D. Wu, Y. Chen, J. Liu, X. Zhao, A. Li, N. Ming, Appl. Phys. Lett.
2005, 87, 112 501.
[72] J. D. Bryan, S. M. Heald, S. A. Chambers, D. R. Gamelin, J. Am.
Chem. Soc. 2004, 126, 11 640.
[73] D. V. Bavykin, A. A. Lapkin, P. K. Plucinski, L. Torrente-Murciano,
J. M. Friedrich, F. C. Walsh, Top. Catal. 2006, doi: 10.1007/s11244-
006-0051-4.
[74] P. Umek, P. Cevc, A. Jesih, A. Gloter, C. P. Ewels, D. Arcon, Chem.
Mater. 2005, 17, 5945–5950.
[75] J. Yin, X. Zhao, Nanotechnology 2006, 17, 192.
[76] D. V. Bavykin, A. A. Lapkin, P. K. Plucinski, J. M. Friedrich, F. C.
Walsh, J. Phys. Chem. B 2005, 109, 19 422.
[77] R. Ma, T. Sasaki, Y. Bando, Chem. Commun. 2005, 948.
[78] D. V. Bavykin, A. A. Lapkin, P. K. Plucinski, J. M. Friedrich, F. C.
Walsh, J. Catal. 2005, 235, 10.
[79] M. Hodos, E. Horvath, H. Haspel, A. Kukovecz, Z. Konya, I. Kiricsi,
Chem. Phys. Lett. 2004, 399, 512.
[80] B. D. Yao, Y. F. Chan, X. Y. Zhang, W. F. Zhang, Z. Y. Yang,
N. Wang, Appl. Phys. Lett. 2003, 82, 281.
[81] L. Qian, Z. L. Du, S. Y. Yang, Z. S. Jin, J. Mol. Struct. 2005, 749, 103.
[82] C. H. Lin, S. H. Chien, J. H. Chao, C. Y. Sheu, Y. C. Cheng, Y. J.
Huang, C. H. Tsai, Catal. Lett. 2002, 80, 153.
[83] G. Cao, J. Phys. Chem. B 2004, 108, 19 921.
______________________
2824 www.advmat.de © 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
[84] D. V. Bavykin, E. V. Milsom, F. Marken, D. H. Kim, D. H. Marsha,
D. J. Riley, F. C. Walsh, K. H. El-Abiary, A. A. Lapkin, Electro-
chem. Commun. 2005, 7, 1050.
[85] C. T. J. Low, R. G. A. Wills, F. C. Walsh, Surf. Coat. Tech. 2006, 201,
371.
[86] Y.-G. Wang, X.-G. Zhang, Electrochim. Acta 2004, 49, 1957.
[87] Y.-G. Wang, Z.-D. Wang, Y.-Y. Xia, Electrochim. Acta 2005, 50, 5641.
[88] H. Kuanxin, Z. Xiaogang, L. Juan, Electrochim. Acta 2006, 51, 1289.
[89] A. Liu, M. Wei, I. Honma, H. Zhou, Anal. Chem. 2005, 77, 8068.
[90] A. Kleinhammes, G. W. Wagner, H. Kulkarni, Y. Jia, Q. Zhang,
L.-C. Qin, Y. Wu, Chem. Phys. Lett. 2005, 411, 81.
[91] J. Cao, J.-Z. Sun, H.-Y. Li, J. Hong, M. Wang, J. Mater. Chem. 2004,
14, 1203.
[92] M. Hodos, E. Horvath, H. Haspel, A. Kukovecz, Z. Konya, I. Kiricsi,
Chem. Phys. 2004, 399, 512.
[93] J.-C. Xu, M. Lu, X.-Y. Guo, H.-L. Lia, J. Mol. Catal. A 2005, 226,
123.
[94] M. Wang, D. J. Guo, H. L. Li, J. Solid State Chem. 2005, 178, 1996.
[95] C. H. Lin, C. H. Lee, J. H. Chao, C. Y. Kuo, Y. C. Cheng, W. N.
Huang, H. W. Chang, Y. M. Huang, M. K. Shih, Catal. Lett. 2004, 98,
61.
[96] V. Idakiev, Z. Y. Yuan, T. Tabakova, B. L. Su, Appl. Catal. A 2005,
281, 149.
[97] T. Akita, M. Okumura, K. Tanaka, K. Ohkuma, M. Kohyama,
T. Koyanagi, M. Date, S. Tsubota, M. Haruta, Surf. Interface Anal.
2005, 37, 265.
[98] S. H. Lim, J. Luo, Z. Zhong, W. Ji, J. Lin, Inorg. Chem. 2005, 44,
4124.
[99] K. A. Lokshin, Y. Zhao, D. He, W. L. Mao, H. K. Mao, R. J. Hem-
ley, M. V. Lobanov, M. Greenblatt, Phys. Rev. Lett. 2004, 93, 125 503.
[100] Y. K. Zhou, L. Cao, F. B. Zhang, B. L. He, H. L. Liz, J. Electrochem.
Soc. 2003, 150, A1246.
[101] A. R. Armstrong, G. Armstrong, J. Canales, P. G. Bruce, J. Power
Sources 2005, 146, 501.
[102] J. Li, Z. Tang, Z. Zhang, Electrochem. Commun. 2005, 7, 62.
[103] L. Kavan, M. Graetzel, J. Rathousky, A. Zukal, J. Electrochem. Soc.
1996, 143, 394.
[104] J. Li, Z. Tang, Z. Zhang, Chem. Phys. Lett. 2006, 418, 506.
[105] L. Kavan, M. Kalbac, M. Zukalova, I. Exnar, V. Lorenzen, R. Ne-
sper, M. Grätzel, Chem. Mater. 2004, 16, 477.
[106] J. Li, Z. Tang, Z. Zhang, Chem. Mater. 2005, 17, 5848.
[107] J. Li, Z. Tang, Z. Zhang, Electrochem. Commun. 2005, 7, 894.
[108] A. R. Armstrong, G. Armstrong, J. Canales, P. G. Bruce, Adv. Mater.
2005, 17, 862.
[109] X. Gao, H. Zhu, G. Pan, S. Ye, Y. Lan, F. Wu, D. Song, J. Phys.
Chem. B 2004, 108, 2868.
[110] S. Uchida, R. Chiba, M. Tomiha, N. Masaki, M. Shirai, Electrochem-
istry 2002, 70, 418.
[111] Y. Ohsaki, N. Masaki, T. Kitamura, Y. Wada, T. Okamoto, T. Sekino,
K. Niiharab, S. Yanagida, Phys. Chem. Chem. Phys. 2005, 7, 4157.
[112] H. Yu, J. Yu, B. Cheng, M. Zhou, J. Solid State Chem. 2006, 179, 349.
[113] J. N. Nian, H. Teng, J. Phys. Chem. B 2006, 110, 4193.
[114] D. V. Bavykin, J. M. Friedrich, A. A. Lapkin, F. C. Walsh, Chem.
Mater. 2006, 18, 1124.
Adv. Mater. 2006, 18, 2807–2824