Applied Catalysis A: General 481 (2014) 127–142

Contents lists available at ScienceDirect

Applied Catalysis A: General

journal homepage: www.elsevier.com/locate/apcata

Review

A critical review on the recent progress of synthesizing techniques

and fabrication of TiO2 -based nanotubes photocatalysts
Yean Ling Pang ∗ , Steven Lim, Hwai Chyuan Ong, Wen Tong Chong
Department of Mechanical Engineering, Faculty of Engineering, University of Malaya, 50603 Kuala Lumpur, Malaysia

a r t i c l e i n f o a b s t r a c t

Article history: One-dimensional titanium dioxide (TiO2 )-based nanotubes have attracted great interest to be used as
Received 28 February 2014 photocatalysts in the field of environmental applications. Three main approaches via template-assisted,
Received in revised form 13 April 2014 electrochemical anodic oxidation and alkaline hydrothermal treatment used to prepare the nanostruc-
Accepted 12 May 2014
tured TiO2 -based nanotubes are reviewed. The parameters that affect the formation of TiO2 -based
Available online 20 May 2014
nanotubes via hydrothermal method such as phases and particle sizes of starting materials, types and con-
centrations of alkaline solution, temperature and duration of hydrothermal treatment, ultrasonication-
Keywords:
and microwave-assisted hydrothermal synthesis, acid washing and calcination have been reviewed in
TiO2 nanotubes
Photocatalyst
details. This paper also discussed the possible crystal structure and formation mechanism of TiO2 -based
Hydrothermal synthesis nanotubes via alkaline hydrothermal treatment. In addition, the recent research progress on the struc-
Mechanism tural modification of TiO2 -based nanotubes to be used as photocatalysts is summarized in this review.
Modifications Modification strategies investigated include metal nanoparticles deposition, single- and co-doping of
metal ions/non-metal ions, coupled with other semiconductors to form binary composites and hybrid
with carbon nanomaterials.
© 2014 Elsevier B.V. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
2. TiO2 nanoparticles, randomized and oriented nanotubes for photocatalytic reactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
3. Synthesizing techniques of TiO2 nanotubes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
3.1. Template-assisted method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
3.2. Electrochemical anodization method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
3.3. Hydrothermal method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
4. Factors influencing the formation of titanate nanotubes via hydrothermal treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
4.1. Starting materials: phases and particle sizes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
4.2. Alkaline solution: types and concentrations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
4.3. Hydrothermal treatment: temperature and duration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
4.4. Assisted hydrothermal synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
4.5. Post-hydrothermal treatments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
4.5.1. Acid washing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
4.5.2. Calcination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
5. Crystal structure, formation mechanism and structural-activity of titanate nanotubes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
6. Modification of titanate nanotubes as photocatalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
6.1. Metal-loaded . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
6.2. Doping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
6.3. Coupled semiconductors/binary composites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
6.4. Hybrid with carbon nanomaterials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141

∗ Corresponding author. Tel.: +60 16 408 0412; fax: +60 3 7967 5317.
E-mail address: pangyeanling@hotmail.com (Y.L. Pang).

http://dx.doi.org/10.1016/j.apcata.2014.05.007
0926-860X/© 2014 Elsevier B.V. All rights reserved.
128 Y.L. Pang et al. / Applied Catalysis A: General 481 (2014) 127–142

7. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
Acknowledgement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141

1. Introduction nanotubes photocatalyts through hydrothermal treatment has

been emphasized in this review. The factors during the hydro-
Titanium dioxide (TiO2 ) is a very well-known photocatalysis thermal reaction which affect the formation of TiO2 -based
material due to its superior photo-reactivity, non-toxicity, long nanotubes, possible mechanism, modification of TiO2 -based nano-
term stability, high corrosion resistance and available in low cost tubes and their effects on the photocatalytic reactions have been
[1–3]. The key to photolysis of water (H2 O) molecule under illumi- reviewed in details.
nation of light is the positions of the conduction and valence bands
relative to the redox potential of water [4]. TiO2 fulfills the par- 2. TiO2 nanoparticles, randomized and oriented nanotubes
ticular requirement as compared to the other metal oxides, where for photocatalytic reactions
the conduction band edge lay higher than that of the species to
be reduced (hydrogen ion/hydrogen) and the valence band edge For better clarification, the mechanism of photocatalytic reac-
located lower than that of the species to be oxidized (H2 O/oxygen tion is briefly summarized in Eqs. (1)–(6) [13,15]. Photocatalytic
(O2 )). process occurs based on the sufficient energy obtained from
The drawbacks of TiO2 nanoparticles photocatalyst include absorption of light radiation which equals to or is greater than the
the requirement of large amount of catalysts, difficulty in cata- semiconductor band gap energy. This will lead to the excitation and
lyst recycling and problematic agglomeration of TiO2 nanocrystals transfer of electrons from the valence band to the conduction band,
into large particles [5]. Although immobilized TiO2 is a feasible leaving behind holes at the valence band. The electrons could react
option, the overall photocatalytic activity has been compromised with electron acceptors such as O2 adsorbed on the catalyst surface
due to the reduction of surface area and limitation in mass trans- or dissolved in water and reduce to superoxide radical anion (• O2 − ).
fer [6]. In order to enhance photocatalytic activity and facilitate Meanwhile, holes tend to oxidize directly on the organic molecules
catalyst separation after photocatalytic reaction, one-dimensional or react with hydroxyl group and oxidize to hydroxyl radical (• OH).
(1D) nanostructured materials such as nanotubes, nanowires and The resulting radicals are capable of promoting oxidation reactions
nanorods have been investigated. Besides, charge recombination to mineralize organic pollutant.
probability within TiO2 nanoparticles also increases due to the
presence of defects, surface states and grain boundaries which act TiO2 + energy(heatorlight) → hvb + + ecb − (1)
as electron trapping sites [7]. Meanwhile, 1D nanostructured mate- − •O −
O2(ads) + ecb → 2 (2)
rials are expected to have faster electron transport and lower charge
+ • OH +
recombination due to 1D channel for electron transportation and H2 O + hvb → +H (3)
decrement of inter-crystalline contacts, respectively. +
hvb + pollutant → • pollutant+ → degradationproducts (4)
The present review focuses on TiO2 -based nanotubes rather
•O − + pollutant → degradationproducts + CO2 (5)
than nanowires and nanorods due to the internal layer of the hollow 2
nanotubes provide larger surface area and easily functionalized by • OH + pollutant → degradationproducts + CO2 (6)
filling the hollow nanotubes [8]. However, TiO2 is a wide band gap
semiconductor with 3.2 eV for anatase [1,9], 3.0 eV for rutile [1,9] + −
where hvb and ecb are the positive hole in the valence band and
and nanotubes material is reported to have a mobility band gap the electron in the conduction band, respectively while subscripts
of about 3.1–3.87 eV [10,11]. The blue shift of the absorption edge ‘ads’ denote the adsorbed forms of the O2 molecule.
wavelength is attributed to the quantum size effect of TiO2 nano- In randomly packed TiO2 nanoparticles, the transportation of
tubes, which is due to the thin nanotube wall thickness of around photo-excited electron is limited due to low electron diffusion
1 nm [12]. This is the major drawback that limits the use of TiO2 coefficient and scattering effect of free electrons [7]. The electron
photocatalyts for large scale application. Therefore, intense effort trapped sites could occur at defects, surface states and grain bound-
focuses on the modification of TiO2 nanotubes by various strate- aries between nanoparticles [16]. Consequently, charge collection
gies such as metal nanoparticles deposition, single or co-doping efficiency was diminished due to the increment of charge recom-
metal ions/non-metal ions, coupling with other semiconductors bination at the trapping sites. Moreover, other drawbacks of TiO2
and hybrid with carbon nanomaterials. These efforts alter the elec- nanoparticles are often related to the separation and recycling of
tronic properties of TiO2 nanotubes by narrowing the band gap the particulate catalysts from the reaction media [5]. Filtration of
energy, resulting in higher reaction rates for the photocatalytic ultra-fine nanoparticles is a tedious and time consuming separation
degradation of organic pollutants. process.
The review begins with a brief overview of the synthesis In order to improve these behaviors, recent researches have
techniques including template-assisted, electrochemical anodiza- moved toward the use of 1D nanostructures of TiO2 such as nano-
tion and hydrothermal treatment with their research progress, tubes, nanorods and nanowires. The unique properties of high
advantages and disadvantages. Extensive applications of template- aspect ratio of TiO2 nanotubes include large surface area, high
assisted method often encounter difficulties like pore size cation exchangeability, high catalytic activity, easier separation
limitation of the prepared template and post-removal template and recyclability [8]. These advantages make nanotubes attrac-
usually results in impurities [13]. The method of electrochemi- tive for industrial scale applications. Ohsaki et al. [17] found that
cal anodization is based on the anodization of titanium foil to the use of TiO2 nanotubes, even in a randomly disordered system,
obtain nanoporous TiO2 film [14] and can be considered as an was able to suppress the possibility of charge recombination. At
expensive technique for large scale production of TiO2 nanotubes the same electron diffusion coefficient of TiO2 nanoparticles and
[3]. Besides, this method may lead to un-separatable tubes and randomly packed TiO2 nanotubes, TiO2 nanotubes with longer dif-
uneven length distribution of tubes over a large surface area [9,10]. fusion length and longer electron lifetime would exhibit excellent
Thus, the recent research progress of the formation of TiO2 -based electron transport. In addition, TiO2 nanotubes can provide strong
 Y.L. Pang et al. / Applied Catalysis A: General 481 (2014) 127–142 129

Fig. 1. A comparison of the electron pathways through (a) nanoparticles [16], (b) randomized [18] and (c) oriented nanotubular structured TiO2 [16].

light scattering effect and enhance light harvesting properties [16]. Table 1
Comparison of available methods for TiO2 nanotubes preparation [19].
The special features of internal and external surface area are avail-
able for the adsorption and chemical reaction of organic pollutant Fabrication method Characteristics
molecules. Template-assisted • Ordered arrays
It should be highlighted that TiO2 nanotubes can be synthesized method • Advantages:
into randomized powdery or immobilized nanotubes arrays. The The dimension of nanotubes can be controlled by the
immobilized nanotubes are self-organized, vertically oriented and size and type of applied templates
Uniform size of nanotubes can be formed
supported on a surface to form an integrated unit. Immobilization
• Disadvantages:
of nanocatalyst on a support can prevent catalyst agglomeration Nanotubes morphology may be destroyed during
and facilitate catalyst recycling. The arrangement of highly ordered post-removal of the templates
TiO2 nanotubes array is perpendicular to the titanium substrate, Dissolution of template may result in contamination of
nanotubes
providing channels for efficient charge transfer from the solution
to the collecting substrate [7]. This reduces the losses incurred by Electrochemical • Oriented arrays
charge trapping across the nanoparticles grain boundaries [16]. anodization method • Advantages:
The dimension of nanotubes can be controlled by
Fig. 1 shows the electron transporting pathways in nanoparti-
varying the voltage, electrolyte, pH and anodizing time
cles, randomized and oriented nanotubular structured TiO2 . The Ordered alignment of nanotubes with high aspect ratio
ordered TiO2 nanotubes arrays are expected to improve the charge can be formed
collection efficiency due to the fast electron transportation along 1D • Disadvantages:
The requirement of fabrication apparatus
channel with minimum charge recombination site. Consequently,
Length distribution and separation of nanotubes over a
a lot of research works had been concentrating on the formation of large surface area is not well-developed
ordered arrays of nanotubes especially in the application of pho-
Hydrothermal • Random alignment or can be aligned
toelectrochemical water splitting and dye sensitized solar cells
treatment • Advantages:
[7,9]. However, dispersed non-ordered nanotubes showed great Easy route to obtain nanotubes in relatively large
potential as heterogeneous catalyst to accelerate photocatalytic amount
degradation of organic pollutants in liquid medium [18]. Thus, the • Disadvantages:
research progresses on the formation of nanotubes via different Difficult in achieving uniform size of nanotubes

synthesizing techniques are discussed in details in subsequent sec-

tion.
3.1. Template-assisted method

3. Synthesizing techniques of TiO2 nanotubes
Three popular synthesizing techniques of TiO2 nanotubes have
been investigated in recent years including template-assisted,
electrochemical anodization and hydrothermal treatment. A com-
parison between unique features of these three fabrication
methods is summarized in Table 1.
The dimensions (diameter and length) of TiO2 nanotubes formed
through hard template approach membranes can be controlled eas-
ily by the dimension of templates [20]. Generally, TiO2 nanotubes
derived from the templates are obtained after removal of tem-
plates by selective chemical etching [21] or thermal decomposition
[22]. Template can be divided into two types: positive and nega-
tive templates as shown in Fig. 2 [20]. Positive template synthesis

Fig. 2. Illustration of two different types of hard templates: (a) positive template and (b) negative template [23].
130 Y.L. Pang et al. / Applied Catalysis A: General 481 (2014) 127–142

Table 2
Template-assisted method for preparation of TiO2 nanotubes.

Template categories Types of Preparation methods of Dimensions of nanotubes References

(i) precursors (i) template
(ii) template (fabrication time) (ii) TiO2 nanotubes

Positive template (i) Amorphous TiO2 (i) Two-step anodization Internal diameter (Di ) = 70–100 nm Hoyer [24]
(ii) Anodized aluminum oxide (AAO) (ii) Template-based electrochemical Outer diameter (Do ) = 140–180 nm
(>3 h) acted as template for PMMA rods deposition process Average length (L) = 8 ␮m
(i) TiO2 (i) Low-temperature aqueous Wall thickness = 50–120 nm for Lee et al. [25]
(ii) Zinc oxide (ZnO) nanorod template solution method deposition times of 3–10 h
(5 h) (ii) Template-based liquid phase
deposition
(i) TiO2 sol (i) Sol–gel followed by Average diameter (D) = 400–600 nm Qiu et al. [21]
(ii) ZnO nanorod template (no fabrication hydrothermal method L = 30 ␮m
time is reported) (ii) Layer-by-layer absorption and Shell thickness = 150 nm
reaction technique Do and L of TiO2 branches were 60 nm
and 1 ␮m, respectively
(i) TiO2 sol (i) Rapid thermal chemical vapor Do = 60 nm Kim et al. [22]
(ii) Multi-walled carbon nanotubes (>2 h) deposition Wall thickness = 15 nm
(ii) Sol–gel method

Negative template (i) Titanium (IV) iso-propoxide (i) Commercially available Do = 80–100 nm Bae et al. [26]
(ii) AAO (ii) Template-based gas phase L = 30 ␮m
atomic layer deposition Wall thickness = 2.5 nm
Wall thickness = 7 nm
Specific surface area = 400 m2 /g
(i) Titanium (IV) t-butoxide (i) Commercially available Do = 200 nm Yuan et al. [27]
(ii) AAO (ii) Template-based alcoholate L = 60 ␮m
hydrolysis Wall thickness = 10–45 nm at
precursors concentration of 0.1–2.0 M

is employed when oxide materials are coated on the outer sur-
face of the template [21]. Meanwhile, negative template synthesis
involved oxide materials which are deposited inside the template
channel space [23]. Table 2 shows several selected research works
on the production of TiO2 nanotubes through positive and negative
templates.
In 1996, Hoyer [24] was the first researcher reported the
preparation of TiO2 nanotubes via template-assisted method. TiO2
nanotubes were formed by electrochemical deposition method
into polymethyl methacrylate (PMMA) mold. Anodically grown
of porous anodized aluminum oxide (AAO) membrane consists
of an array of parallel straight nanopores with uniform diam-
eter and length. It served as negative template to produce a
positive type polymer mold. TiO2 nanotubes were obtained after
dissolution of the polymer mold using acetone. Recently, Kara-
man et al. [28] demonstrated that TiO2 nanotubes thin films
could be deposited over electrospun PMMA fibers by hot fila-
ment chemical vapor deposition method. Post-heat treatment of
the deposited materials leads to the decomposition of polymeric
inner layer and formation of randomly distributed anatase TiO2
nanotubes. Several types of soft template had been reported in the
literature such as core–shell–corona cylindrical polymer brushes,
self-assembled block copolymers and biological superstructures
[29]. Ku et al. [30] introduced silicon moiety into a block copolymer,
polystyrene-block-poly(4-(tert-butyldimethylsilyl)oxystyrene) to
enhance template properties such as thermal stability, dielectric-
ity and mechanical strength. They found that the wall thickness of
the TiO2 nanotubes can be easily tuned by adjusting the number of
atomic layer.
On the other hand, zinc oxide (ZnO) nanostructure was
employed as a template due to the low material cost and ease
of fabrication [31]. It has characteristic of amphoteric in nature,
which can be easily dissolved in mild acids or bases [31]. Lee et al.
[25] demonstrated that TiO2 nanotubes could be produced with-
out using chemical medium to dissolve ZnO nanorod template. The
removal of ZnO could be achieved by the reaction with hydrogen
ions during liquid phase deposition process. However, the sam-
ple could be contaminated easily by precipitation resulting from
the liquid phase deposition process of TiO2 . Qiu et al. [21] formed
three-dimensional (3D) branched TiO2 nanotube arrays on trans-
parent conducting oxide substrates using ZnO nanorods and ZnO
spacers as template. The ZnO nanorods template and sacrificial
spacers can be removed by selective chemical etching in dilute
aqueous solution of titanium tetrachloride, which might lead to
the collapse of TiO2 tubes.
Recently, carbon nanotubes have been considered to be an ideal
template to synthesis TiO2 nanotubes due to the template easily
being removed, small diameter of carbon nanotubes and can sup-
port tubular morphology [22]. A sintering process at an appropriate
temperature is performed after coating TiO2 on carbon nanotubes
to burned out the carbon nanotubes template and high crystallinity
of TiO2 nanotubes could be obtained [22]. Meng et al. [32] reported
crystalline anatase TiO2 nanotubes were prepared using atomic
layer deposition on carbon nanotubes and calcination at 250 ◦ C.
Meanwhile, Kim et al. [22] successfully synthesized anatase TiO2
nanotubes by performing a simply calcination process above 500 ◦ C
on TiO2 coated multi-walled carbon nanotubes composite materi-
als via sol–gel method.
The advantages of positive template are inner diameters and
length of the resulting nanotubes could be determined by the
outer diameters and length of the original template, whereas outer
diameters are determined by the thickness of deposited wall lay-
ers [20]. Meanwhile, the disadvantage of this type of template is
high difficulty in fabrication of uniform length and open ended
nanotubes. Thereby, the use of negative template has attracted
attention.
Nanoporous AAO, which prepared by electrochemical oxidation
of aluminum in acidic solution has been extensively used as a nega-
tive template. The attractive advantages of anodic AAO membranes
are large specific surface area, ordered pore geometry, adjustable
structured properties such as pore diameter and length and the
ease of dissolution with chemical solvents [23]. Assisted by AAO
templates, various methods have been developed to prepare TiO2
nanotubes such as electrochemical technique, sol–gel and atomic
layer deposition [27]. By controlling the conditions of anodization
process, 10–400 nm of pore diameter, 50–600 nm of inter-pore dis-
tance, pore aspect ratio of 10 to 5000 and 10 nm to 150 ␮m of
thickness of porous aluminum oxide layer can be prepared [23].
 Y.L. Pang et al. / Applied Catalysis A: General 481 (2014) 127–142 131

Table 3
Electrochemical anodization method for preparation of TiO2 nanotubes.

Anodization categories Types of Conditions anodization Dimensions of References

(i) titanium nanotubes
(ii) electrolytes

Hydrofluoric acid (HF) based (i) Titanium foil (thickness: n.d.) Voltages: 10–40 V D = 60 nm Gong et al. [33]
electrolytes (first generation) (ii) 0.5–3.5 wt.% HF aqueous solution Duration: 20 min to 6 h L = 250 nm
(i) Titanium foil (thickness: 0.2 mm) Voltages: 20 V D = 40–110 nm Şennik et al. [35]
(ii) 1 wt.% HF aqueous solution Duration: 45 min L = 1 ␮m

Water-based electrolyte containing (i) Titanium foil (thickness: 0.2 mm) Voltages: 20 V D = 90 nm Sreekantan et al. [34]
fluoride ion (second generation) (ii) 3 wt.% of ammonium fluoride (NH4 F) in 1 M Duration: 30 min to 3 h L = 0.7–2.5 ␮m
sodium sulfate (Na2 SO4 ) solution Wall thick-
ness = 20 nm
(i) Titanium foil (thickness: 0.5 mm) Voltages: 20 V Pore diame- Baram and Ein-Eli [36]
(ii) 0.5 wt.% sodium fluoride in 1 M Na2 SO4 Duration: 2 h ter = 75 nm
Surface
area = 94 cm2
on 4 cm2
sample

Organic solvent electrolyte containing (i) Titanium plate (thickness: n.d.) Voltages: 5–40 V Di = 15–53 nm Roman et al. [37]
fluoride ion (third generation) (ii) 0.55 wt.% NH4 F, 1% H2 O in ethylene glycerol Duration: 15–300 min Do = 45–67 nm
or 0.7 wt.% NH4 F, 9.3% H2 O in glycerol L = 1–3 ␮m
Wall thick-
ness = 7 nm
(i) Titanium foil (thickness: 0.13 mm) Voltages: 10–40 V D = 30–219 nm Nischk et al. [38]
(ii) 0.09 M NH4 F and 2% (v/v) H2 O in glycol or Duration: 1 h L = 0.5–5.9 ␮m
0.27 M NH4 F and 45% (v/v) H2 O in ethylene Wall thick-
glycol ness = 10–25 nm

Other approach (i) Titanium foil (thickness: 0.5 mm) Voltages: 10 V D = 7.53 ± 1.53 nm Antony et al. [14]
(ii) 0.1 M perchloric acid Duration: n.d. L = 25 ␮m
(i) Titanium foil (thickness: 100 ␮m) Voltages: 40 V D = 20–30 nm Jha et al. [39]
(ii) 0.05 M sodium perchlorate and 0.05 M Duration: 30 min L = >220 ␮m
sodium chloride in water–ethanol mixed followed 20 min
solution

Note: n.d. is defined as no data provided.

This makes the anodic AAO membrane as an excellent template
material for 1D of ordered arrays nanostructures.
Bae et al. [26] successfully fabricated multi-walled anatase TiO2
nanotubes using AAO as a negative template. Multi-walled nano-
tubes were produced by alternating TiO2 and aluminum oxide
nanolaminate structures onto porous aluminum oxide templates
with atomic layer deposition method, followed by wet etching of
the sacrificial aluminum oxide. The structural parameters including
diameter, length, wall thickness, inter-wall spacing and number of
wall layers of the multi-walled TiO2 nanotubes can be adjusted. In
addition, Yuan et al. [27] synthesized TiO2 nanotubes through an
AAO template-based titanium (IV) t-butoxide (Ti(OC4 H9 )4 ) hydrol-
ysis process. In their experiment, the wall thickness of the TiO2
nanotubes can be controlled by adjusting the concentration of
Ti(OC4 H9 )4 solutions.
In short, both types of template show the disadvantage of
obtaining smaller nanotubes due to the restriction of pore sizes
of the prepared template. Besides, it requires multi-step of pre-
fabrication template and post-removal of template which are time
consuming for the practical applications. In the case of the post-
removal for hard template, array of nanotubes may be collapsed,
contaminated or become bundles onto the substrate during the
wet-chemical etching in acidic or alkaline solution [30].
3.2. Electrochemical anodization method
TiO2 nanotubes with ordered arrangement and high aspect ratio
can be produced by using electrochemical anodization method.
Basically, TiO2 nanotubes grow on the surface of the anode of tita-
nium foils or thin sheets. The dimensions of nanotubes formed
in various electrolytes can be controlled by applying different
electrolyte composition, applied voltage, pH and anodizing time
[33,34]. The development of electrochemical anodization method
can be divided into three generations based on the length of TiO2
nanotubes or type of electrolyte as shown in Table 3.
In 1999, Zwilling et al. [40] had reported that the titanium
metal surface became porous when using anodization in chromic
acid electrolytes containing hydrofluoric acid. They recognized that
small addition of fluoride ions (F− ) to electrolyte was a crucial
ingredient in the formation of self-organized porous structures.
The latter work done by Gong et al. [33] demonstrated clearly self-
organized and highly uniform nanotubes on titanium substrate
using an aqueous dilute hydrofluoric solution as electrolyte. The
high chemical dissolution rate of titanium and TiO2 in hydroflu-
oric acid containing electrolytes and high H2 O content of the
electrolytes could result in limitation of length to few hundred
nanometers and non-ideal tubular shape of nanotubes [33,35].
Hence, the first generation of nanotubes length with approximately
500 nm or less could not meet the requirement of practical appli-
cations.
Subsequently, the use of buffered neutral electrolytes contain-
ing various fluoride salts such as sodium fluoride [36] and ammonia
fluoride [34] had been proposed to slow down the TiO2 dissolution
rate. For this second generation of nanotubes, longer nanotubes
with lengths up to several micrometers can be produced as the
ultimate concentration of F− and hydrofluoric acid are determined
by the solution pH. Taking into account the importance of the pH
during anodization, Sreekantan et al. [34] reported that the reduc-
tion of acidity could decrease the chemical dissolution of titanium.
This allowed the formation of nanotube arrays up to 2.5 ␮m long.
On the other hand, electrochemical anodization in non-aqueous
organic solvent electrolyte containing F− can be considered as third
generation of nanotubes [37]. This method appeared as a promis-
ing approach to form smooth and long nanotube arrays. Various
types of organic solvents such as glycerol [37,38], ethylene glycol
[37,38], formamide [41], N-methylformamide [41] and dimethyl
132 Y.L. Pang et al. / Applied Catalysis A: General 481 (2014) 127–142
sulfoxide [41] had been successfully employed for the formation
of TiO2 nanotube arrays. Paulose et al. [41] found that the key to
obtain very long nanotube arrays was to minimize the amount of
water to be less than 5 wt.% during anodization process. Addition of
small amount of water allowed sufficient titanium oxidation rate
since the oxidation process was difficult in pure organic electrolyte
[38,41].
According to the literature, the most widely accepted formation
mechanism of TiO2 nanotubes during electrochemical anodization
involves several stages [35,38,42]. At the initial stage of anodiza-
tion, a dense TiO2 layer was formed on the titanium surface by
field-assisted oxidation via reaction (7). Ion migration occurred
within metal/metal oxide interface under the applied electric field.
Titanium ions could dissolve and react with oxygen ion (O2− ) or
hydroxyl ion (OH− ) provided by water to form TiO2 or titanium
hydroxide, respectively. In the next stage, the titanium metal and
oxide layer, which were in contact with the electrolyte, would etch
by F− ions through chemical dissolution to form soluble hexaflu-
orotitanium complex (Eq. (8)). The localized dissolution of oxide
resulted in the formation of tiny pits on the oxide layer and gradu-
ally grew into bigger pores with increasing anodization time.
Ti + 2H2 O → TiO2 + 4H+ + 4e−
+ − 2−
TiO2 + 4H + 6F → [TiF6 ] + 2H2 O
It is believed that the competition between rate of oxide growth
at the metal/metal oxide interface (Eq. (7)) and the rate of oxide
dissolution at the metal oxide/electrolyte interface (Eq. (8)) is the
key factor in determining nanotube formation [35]. When the rate
of electrochemical etching equals to the rate of chemical dissolution
of the nanotubes, the tube length remains the same. Thus, the final
stage determines the length of the nanotubes during the reaction.
The crucial key factors to achieve high degree arrangement of
hexagonal close-packed nanotubes structured are anodic voltages
for a given electrolyte before dielectric breakdown [43], purity of
the material [43] and multi-step anodization [42]. By altering the
electric field during the growth of nanotubes, several types of tube
geometry can be achieved such as bamboo- [44], nanolace- [45] and
branched-type nanotubes [45]. For example, bamboo-type nano-
tubes were obtained by alternating voltage and holding time on
the microstructure [44], nanolace-type nanotubes were obtained
when extending the etching process of bamboo-type nanotubes
[45], while tube branching morphology could be obtained when
altering a higher to a lower anodization voltage [45]. Meanwhile, Ji
et al. [46] found that double-walled nanotubes with bamboo-like
morphologies could be formed if the anodization process was car-
ried out under alternating voltage conditions. The double-walled
structure could be further tuned by heat treatment.
In addition to the dilute fluoride electrolyte approach, alter-
native approach using electrolyte containing perchlorate [14],
chloride [39,47], bromide [47] or nitrate [36] ion is possible to
form bundles of TiO2 nanotubes on titanium substrate under very
rapid growth conditions. This method is known as rapid breakdown
anodization and micrometers long nanotubes can be obtained in a
few min [39]. Hence, it is not easy to tune the morphology of tube
bundles into highly ordered self-organized TiO2 nanotubes through
this ultra-fast anodization technique.
Production of TiO2 nanotubes via electrochemical anodiza-
tion method requires two/three electrode configurations with
anodization voltage and the help of a power supply. Besides, the
disadvantages of this method include the length distribution of
tubes is not well-defined [9], nanotubes are not separated from
each other in an organized manner and do not have well-developed
gaps between the tubes over a large surface area [10]. Hydrogen
gas is evolved at the cathode in the form of gas bubble nucleates,
which will accumulate at the electrode surfaces and increase the
electrical resistance of the cell [48]. As a result, more electric energy
must be used and consequently increases the cost of operating
electrochemical anodization.
3.3. Hydrothermal method
Among these preparation methods, hydrothermal method has
attracted more attention from the scientific community due to the
simple, cost-effective and environmental friendly method for large
scale production of nanotubes. This synthesis conditions could be
adjusted to fabricate alternative low dimensional of TiO2 nano-
structures [10]. These nanostructures included TiO2 nanotubes,
nanofibers, nanowires, nanoribbons and nanorods by controlling
various parameters during hydrothermal synthetic conditions. The
term nanotube was used for high aspect ratio materials with a
hollow inner channel and nanowire for high aspect ratio materi-
als without an inner void. Meanwhile, nanofibers are similar to
nanowires but with significantly larger diameter [49]. Nanorods
usually featured shorter aspect ratio as compared to nanowires.
Fig. 3 shows the HRTEM images for the morphology of nanotubes,
nanowires, nanofibers and nanorods, adapted from Yuan and Su
(7) [50].
In year 1998, Kasuga et al. [51] first reported the synthesis of
(8) TiO2 nanotubes by employing an alkaline hydrothermal treatment
of TiO2 nanoparticles. With respect to the described methods, the
alkaline hydrothermal route could produce nanotubes with tunable
diameters (∼8–200 nm) [52]. After further investigation, major-
ity of the researchers found that the so-called “TiO2 nanotubes”
obtained via hydrothermal method were actually titanate nano-
tubes [53,54]. Thus, the term of titanate nanotubes was used for
the subsequent review instead of TiO2 nanotubes.
Although this synthesis method had been carried out exten-
sively over the past decade, the crystalline structure, composition,
thermal stability and formation mechanism are still in controversy.
Several factors can influence the formation of titanate nanotubes
such as phases and particle sizes of the starting materials [55], types
and concentrations of the alkaline solution [56], hydrothermal tem-
perature and duration [57] and post-treatment processes (acid
washing and calcination) [58]. Table 4 shows several selected works
on the conventional- [59], ultrasonication- [57,60] and microwave-
hydrothermal [11,61] methods for synthesizing titanate nanotubes.
The details of these combination techniques will be further dis-
cussed in Section 4.4.
The main disadvantage of this nanotube fabrication method
is the requirement of long reaction time in concentrated sodium
hydroxide (NaOH) medium, which lead to the excessive intercala-
tion of sodium ions in the produced nanotubes. Hence, more efforts
have to be made in order to shorten the synthesize time and pro-
duce ordered alignment of TiO2 nanotubes apart from the random
alignment powder form [62].
4. Factors influencing the formation of titanate nanotubes
via hydrothermal treatment
The microstructure of titanate nanotubes is easily affected by
the starting materials, concentrations and types of the alkaline
solution, synthesize conditions (reaction time and synthesis tem-
perature) and post-hydrothermal treatments (acid washing and
calcination temperature).
4.1. Starting materials: phases and particle sizes
The effects of different phases and particle sizes of the start-
ing materials for titanate nanotubes are still debatable. Table 5
summarizes several selected results on the formation morphology
 Y.L. Pang et al. / Applied Catalysis A: General 481 (2014) 127–142
Fig. 3. HRTEM images of (a) overall view nanotubes, detailed view on (b) nanotubes, (c) nanowires, (d) nanofibers and (e) nanorods.
Adapted from Yuan and Su [50].
of titanate nanotubes using different types of starting materi-
als, alkaline solution and hydrothermal treatment. Morgan et al.
[55] observed that the dissolution process of anatase precursor
corresponded to zero-order kinetic, where the dissolution was
independent of precursor concentrations. Meanwhile, the disso-
lution kinetic for the rutile phase was second-order, indicated TiO2
rutile was affected by the diffusion of products and replenishment
of reactants to the surface. Anatase has often been observed to react
rapidly to produce nanotubes, whereas rutile has been observed to
react incompletely with remaining unreacted rutile in the product
or nanorods shaped after similar hydrothermal conditions. They
also observed that anatase with small crystalline and rutile phase
with big crystalline size could produce nanotubes with lengths up
to several hundred nanometers, but the former starting material
could produce nanotubes with smaller average external diameter.
Even though several authors reported that amorphous TiO2
could only produce nanofibers not nanotubes [50], Ylhäinen et al.
[68] claimed to have obtained nanotube products from amorphous
precursors. Hence, it seems that there were other parameters play-
ing a role apart from crystalline form of TiO2 . Yuan and Su [50]
reported that the effect of particle size of the starting material
toward the yield of the produced nanotubes was quite small in
comparison to the hydrothermal temperature and concentration
of alkaline solution. On the other hand, Dawson et al. [69] con-
cluded that the length of titanate nanotubes was influenced by
133
both phase and particle size of the starting material. Pure anatase
phase with a particle size of 300 nm could produce nanotubes with
1 ␮m in length, micrometer sized of TiO2 with mixed phases pro-
duced tubes with lengths between 200 and 500 nm, while P25
mixed phases with particle size of 30 nm produced lengths less than
100 nm. Viriya-empikul et al. [57] reported that interlayer spac-
ing of nanotubes was not significantly affected by the size of the
starting material. They verified that smaller crystal size of start-
ing material could lead to titanate nanotubes with a higher surface
area.
In addition, Morgado et al. [63] reported that TiO2 anatase raw
materials with smaller crystal size usually led to faster rate of
thermal transformations of TiO2 nanotubes. Anatase with smaller
crystalline size could produce larger external diameter as com-
pared to those produced from anatase phase with bigger crystalline
size. However, Papa et al. [70] found that anatase particles with
a grain size of less than 17 nm could not form nanotubes at
hydrothermal temperature of 150 ◦ C even with prolong treatment
duration from 36 h to 72 h. They concluded that too many small pri-
mary germs involved in the dissolution/recrystallization reactions
and it was hard to obtain super-saturation condition for the growth
of the nanotubes. Meanwhile, Miao et al. [71] used 14 nm diame-
ter anatase aerogel powders as the starting material and produced
nanotubes after treatment duration of 40 h at 150 ◦ C. These con-
flicted results reported by both of the studies might probably due
134 Y.L. Pang et al. / Applied Catalysis A: General 481 (2014) 127–142
Table 4
Hydrothermal method for preparation of titanate nanotubes.
Auxiliary methods Conditions hydrothermal reaction
Conventional hydrothermal • 5 g TiO2 powder in 10 M sodium hydroxide
(NaOH)
• 150 ◦ C for 20 h
• Acid washing with 0.1 M nitric acid (HNO3 )
Ultrasonication-assisted • 150 mg TiO2 powder in 30 mL of 10 M NaOH
hydrothermal • Mixture was sonicated at 280 W for 60 min
• 110 ◦ C for 4 h
• Acid washing with 0.1 M HNO3
• 0.5 g TiO2 powder in 50 ml 10 M NaOH
• Mixture was sonicated at 7.6, 38.1 W for 8 min
• 150 ◦ C for 3 days
• Acid washing with 0.1 M hydrochloric acid
(HCl)
Microwave-assisted • 0.5 g TiO2 rutile in 25 ml of 10 M NaOH
hydrothermal • Microwave power for 45 min at 200 ◦ C
• Acid washing with 0.1 M HNO3
• 0.6 g P25 TiO2 in 70 ml 10 M NaOH
• Microwave power of 70, 400, and 700 W at
130 ◦ C for 3 h
• Acid washing with 0.5 M HCl
to the different conditions used such as fabrication vessels, amount
of starting materials and concentrated alkaline solution.
4.2. Alkaline solution: types and concentrations
The alkaline type and their concentrations play important role
in hydrothermal process. The most favorable titanate nanotubes
were prepared via hydrothermal treatment of TiO2 powders in a
10 M NaOH solution [63,70]. There was only small proportion of
nanotubes formed when NaOH concentration was lower than 5 M
or higher than 18 M [56]. The concentration of dissolved Ti4+ in solu-
tion can determine the rate of crystallization of titanate nanosheets,
which in turn control the morphology of the final nanostructures
[10].
Bavykin et al. [72] found that an increment in the molar ratio
of TiO2 to NaOH generally resulted in higher average pore diam-
eter and a reduction in the specific surface area. Meanwhile, the
pore volume of nanotubes was independent of the molar ratio of
TiO2 to NaOH. Furthermore, Huang et al. [64] demonstrated that
nanosheets, nanotubes and nanowires were three kinetic products
from the reaction of TiO2 with NaOH when increasing alkaline con-
dition from 5 to 12 M. They observed that the nanowires had wider
diameter and longer length as compared to nanotubes morphol-
ogy. However, Sikhwivhilu et al. [56] reported that a small portion
of shorter nanotubes was produced when concentration of NaOH
solution increased to 18 M. Thus, the concentration of NaOH could
influence in the yield of tubular morphology.
Furthermore, Bavykin et al. [73] discovered that titanate nano-
tubes could be prepared using 10 M potassium hydroxide (KOH)
and low temperature of 56 ◦ C after 12 days. Thus, titanate nano-
tubes could also be produced using a NaOH/KOH binary aqueous
mixture at 100 ◦ C [74]. Recently, Sikhwivhilu et al. [56] found
that potassium titanate nanotubes could also be produced by
hydrothermal treatment in 10 M KOH at 120 and 150 ◦ C for
24 h. They reported that KOH was more reactive with TiO2 to
yield different shapes of nanostructured materials compared to
NaOH. Ammonium hydroxide solution showed no influence on the
microstructure of TiO2 . Meanwhile, similar spherical morphology
Dimensions of nanotubes References
Average Do = 9 nm Kitano et al. [59]
Average Di = 6 nm
Interlayer spacing = 0.7 nm (2–4 walls)
L = 0.3–3.0 ␮m
Specific surface area = 423 m2 /g
Average Do = 5 nm Zhu et al. [60]
Wall thickness = 1.3 nm
L = 200–300 nm
Di = 4–6 nm Viriya-empikul et al. [57]
Interlayer spacing = 0.8 nm (2–6 walls)
Average hydrodynamic size of 53, 490, and
1760 nm at sonication power of 0, 7.6, and 38.1 W
Specific surface area = 179, 258, and 248 m2 /g at
sonication power of 0, 7.6, and 38.1 W
Di = 3–5 nm Huang and Chien [11]
Do = 8–10 nm
L = 100–500 nm
Interlayer spacing = 1 nm (3–4 walls)
Specific surface area = 214 m2 /g
Average Do = 9.9 nm Ou et al. [61]
Average Di = 5.1 nm
Interlayer spacing = 0.82, 0.86, and 0.90 nm at
microwave power of 70, 400, and 700 W (4–5 walls)
Specific surface area = 323, 367, and 320 m2 /g at
microwave power of 70, 400, and 700 W
was observed in 10 M lithium hydroxide (LiOH) at the temperatures
of 120 and 150 ◦ C [56], whereas flower-like structure on the TiO2
surface could be obtained when treated in 10 M LiOH solution at
temperature of 80 ◦ C [75].
4.3. Hydrothermal treatment: temperature and duration
In general, the amount, length and crystallinity of titanate nano-
tubes increase with the hydrothermal temperature [76]. However,
there exists an optimal temperature, which usually reported in
the range of 130–150 ◦ C to obtain high purity of nanotubes with
high surface area [49,57]. During higher reaction temperatures
(>180 ◦ C), the main products were nanoribbons or nanowires like
morphologies [19]. Yuan and Su [50] reported that TiO2 nanotubes
could be formed in the reaction temperature ranged from 100
to 180 ◦ C when either crystalline anatase or rutile or commercial
P25 were used as the raw materials. Nanoribbons were produced
when increasing reaction temperature from 180 to 250 ◦ C, while
nanowires were obtained when reaction temperature ranged from
130 to 240 ◦ C if KOH were used as the alkaline solution.
Bavykin et al. [72] found that increasing temperature from 120
to 150 ◦ C for 22 h resulted in larger average diameter of titanate
nanotubes. Subsequent increment of temperature resulted in a
sharp reduction in pore volume of produced samples and wider
size distribution in the diameter of fibers morphology. However,
Sreekantan and Wei [76] observed that the average outer diame-
ters of the nanotubes produced at 110, 130 and 150 ◦ C for 24 h were
similar, at about 10 nm. Seo et al. [66] observed that increasing the
synthesis temperature up to 150 ◦ C would improve crystallinity of
the tube and increased tube lengths from few hundred nanometers
to few micrometers, but with the same diameters. Meanwhile, Lan
et al. [67] found that the outer diameter of the nanotubes increased
slightly when increasing hydrothermal temperature above 150 ◦ C.
This indicates that reaction temperature is a key factor in deter-
mining the overall morphology of the product. Hence, the effect
of hydrothermal temperature on the formation of the nanotubes
products is similar to that of the alkaline concentration.
 Y.L. Pang et al. / Applied Catalysis A: General 481 (2014) 127–142 135

Table 5
Effects of starting materials, alkaline solution and hydrothermal treatment on the morphology of titanate nanotubes.

Starting materials (phase, Alkaline solution Hydrothermal treatment Dimensions of nanotubes References
composition, particle size) (types, concentration) (temperature, duration)

(i) Anatase, >95%, 112 ± 1 nm
(ii) Rutile, >93%, 320 ± 10 nm
9 M NaOH (i) 160 ◦ C, 2–8 h (i) L = several hundred nanometers, average Morgan et al. [55]
(ii) 160 ◦ C, 4 h Di = 3.7 ± 0.7 nm, average Do = 9.4 ± 1.7 nm,
interlayer spacing = 0.74–0.75 nm (3–5 walls)
(ii) L = several hundred nanometers, average
Di = 3.6 ± 0.7 nm, average Do = 9.8 ± 1.5 nm,
interlayer spacing = n.d. (6–7 walls)

(i) Anatase, 220 nm
(ii) Anatase, 8 nm
10 M NaOH (i) 120 ◦ C, 30 h (i) L = several hundred nanometers, average Morgado et al. [63]
(ii) 120 ◦ C, 15 h Di = 3–5 nm, average Do = 6–10 nm, interlayer
spacing = n.d. (4–5 walls), pore size
distribution = 2–5 nm; 5–60 nm
(ii) L = several hundred nanometers, average
Di = 4–10 nm, average Do = 10–30 nm, interlayer
spacing = 0.74–0.75 nm (>5 walls), pore size
distribution = 2–10 nm

Anatase (i) 3 M NaOH 180 ◦ C, 24 h (i) Nanosheets Huang et al. [64]

(ii) 5 M NaOH (ii) Nanotubes, L = several hundred nanometers,
(iii) 10 M NaOH average Di = 2–4 nm, average Do = 7–10 nm,
(iv) 12 M NaOH interlayer spacing = n.d. (3–5 walls)
(v) 15 M NaOH (iii) Mixture of nanotubes and nanowires,
L = several tens of micrometers, average
D = 40–250 nm
(iv) Mixture of nanowires and amorphous particles

P25 Degussa (i) 5 M NaOH 140 ◦ C, 24 h (i) Fibrous structures, L = 100–200 nm, average Camposeco et al. [65]
(ii) 7 M NaOH D = 25–140 nm
(iii) 10 M NaOH (ii) Nanotubes structures, average Di = 3–4 nm,
(iv) 12 M NaOH average Do = 8–10 nm, wall thickness = 2.5 nm
(iii) Nanotubes structures, average Di = 4 nm,
average Do = 8 nm, wall thickness = 3.5 nm,
interlayer spacing = 0.7 nm (2–4 walls)
(iv) Nanowires structures, L = 300–500 nm, average
D = 15–20 nm

P25 Degussa, 25 nm 10 M NaOH (i) 70 ◦ C, 48 h (i) 2D nanosheets Seo et al. [66]

(ii) 90 ◦ C, 48 h
(iiii) 110 ◦ C, 48 h
(iv) 150 ◦ C, 48 h
(ii) Nanosheet and fiber-like mixed phase
(iii) 1D fiber/tube-like, L = few hundred
nanometers, average Do = 10–20 nm
(iv) 1D separated fibers/tube, L = few micrometers,
average Do = 10–20 nm

Rutile powder, 120–280 nm
10 M NaOH (i) 100 ◦ C, 48 h
(ii) 125 ◦ C, 48 h
(iii) 150 ◦ C, 48 h
(iv) 180 ◦ C, 48 h
(i) Nanotubes, average Do = 7–11 nm Lan et al. [67]
(ii) Nanotubes, average Do = 8–12 nm
(iii) Nanotubes, L = 50–200 nm, average
Di = 2–3 nm, average Do = 8–15 nm
(iv) Nanorods, L = several micrometers, average
D = 40–120 nm

Note: n.d. is defined as no data provided.

At relatively low reaction temperature around 90 ◦ C, nanosheets
[76], nanofibers [66] or nanotubes [57] could be formed after 24 h,
48 h, 72 h, respectively. This proved that the yield of nanotubes
could increase with duration treatment. However, prolonged reac-
tion duration (72 h to 1 week) might lead to the transformation
of nanotubes to nanofibers or nanowires [77]. The width of the
obtained nanofibers or nanowires was about 100–300 nm, which
was significantly larger than the diameter of nanotubes. Thus, an
increment in the average nanotubes diameter can be observed
at longer synthesis time. Hence, the treatment duration could be
reduced if higher reaction temperature was performed in order to
obtain nanotubular structures.
4.4. Assisted hydrothermal synthesis
The purposes of the application of sonication- or microwave-
assisted hydrothermal treatment are due to the advantages such
as shorter synthesis duration [11,60] and enhance the yield
and length of titanate nanotubes [57]. The power of sonication
and irradiation duration played an important role in fabricat-
ing titanate nanostructured materials. Sonication showed a strong
influence on the morphology and properties of titanate nanotubes
[57], apart from dispersing TiO2 uniformly in alkaline solution.
Meanwhile, microwave irradiation is able to deliver energy to
the reactants through molecular interactions and heat treatment
using microwave is an economic, rapid and homogeneous heating
method [11].
Zhu et al. [60] found that titanate whiskers and nanotubes
were obtained when sonication powers used were 580 W and
280 W, respectively and the synthesis duration was shortened from
20 to 4 h. Meanwhile, nanosheet-, nanofiber- and nanotube-like
morphologies were observed under sonication of 100, 280 and
380 W, respectively [78]. Appropriate sonication power and dura-
tion may help the reaction of sodium and hydroxyl species on
the de-agglomerated TiO2 particles and the dissolution/breaking
of Ti O Ti bonding. This resulted in smooth unhindered peeling
of large nanosheets and consequently rolled into longer titanate
nanotubes [57,78]. Ou et al. [61] reported that higher microwave
irradiation power was beneficial to the intercalation of Na atoms
into titanate nanotubes in order to enlarge the spacing between
interlayers. Titanate nanotubes could be obtained after 45 min
under microwave irradiation [11].
136 Y.L. Pang et al. / Applied Catalysis A: General 481 (2014) 127–142
Fig. 4. Schematic representation of the thermal transformations pathways as a function of sodium content in TiO2 nanotubes [53].
In addition, organic solvent-controlled hydrothermal synthe-
sis had been investigated to achieve different 1D nanostructures
morphology of TiO2 . Solvents with different physical and chemi-
cal properties have a strong effect on the solubility, reactivity and
diffusion behavior of the reactants, which are able to influence the
structure and morphological features [79]. It had been reported that
the samples prepared using ethanol [80], methanol [80], methyl
ethyl ketone [80] and ethylene glycol [81] exhibited nanotube
structures. However, Das et al. [81] found that nanosheets were
transformed to nanorods and nanowires in water–ethanol and
water–ethylene glycol mixed solvents, respectively. Meanwhile,
Santara and Giri [82] obtained nanoporous nanoribbons products in
water–ethylene glycol mixed solvents. On the other hand, titanate
nanotubes with ultrahigh crystallization can be obtained after
hydrogen peroxide treatment under reflux at 40 ◦ C for 4 h [83]. This
proves that titanate microstructures are more easily affected by the
preparation conditions, rather than the types of organic solvent.
4.5. Post-hydrothermal treatments
In order to achieve high catalytic activity of titanate nanotubes
and control the tubular structure of nanotubes, the investigation of
post-hydrothermal treatment has received more attention com-
pared to hydrothermal synthetic conditions. Post-hydrothermal
treatments include acid washing and calcination steps. Acid wash-
ing might affect the nanotubes formation [84], crystalline structure
or even chemical composition of the final nanotubes product [53].
Meanwhile, it is believed that calcination process can induce the
transformation of titanate to anatase phase, modify microstructure
based on the amount of sodium present in nanotubes, improve crys-
tallinity and reduce surface defects of the material [79]. From the
point of view of practical usage, it is an important issue to study the
structural stability of the titanate nanotubes and their correspond-
ing microstructure changes at various calcination temperatures.
4.5.1. Acid washing
Kasuga et al. [84] considered that the acid washing process was a
crucial step for the formation of titanate nanotubes. They proposed
that the purpose of acid washing was to remove electrostatic repul-
sion of residual charge, in which Ti O Na converted to Ti OH.
Then, the dehydration of Ti-OH bonds could result in the formation
of nanotubes from nanosheets.
Meanwhile, several researchers suggested that acid washing
was only applied for the ion-exchange process as they found that
titanate nanotubes were formed during hydrothermal reaction
[53,55,66]. When proton replaces sodium ion in the structure, acid
washed titanate nanotubes possesses stronger hydroxyl bonds,
results in decrement of interlayer spacing and increment in specific
surface area [53]. Papa et al. [70] found that acid washing turned
out to be necessary when nanoribbons structure was obtained after
hydrothermal treatment. Researchers often reported an optimum
concentration of hydrochloric acid (HCl) during acid washing in
order to prepare sodium-free nanotubular structure effectively.
Further increment of the concentration of HCl would destroy nan-
otubular structures [58].
The quantity of sodium ions plays a crucial role in the mor-
phological, structural stability and phase transformation of the
nanotubes [53]. The detailed discussions on the role of sodium
in stabilizing the nanotubular structure of titanate nanotubes
had been reported by Zhang et al. [85]. They demonstrated that
heat-treatment of sodium titanate nanotubes may induce the dehy-
dration of interlayered OH groups, whereas hydrogen titanate
nanotubes experienced dehydration of intralayered and interlay-
ered OH groups. These types of dehydrations did not only reduce
the interlayer distance and nanotube length, but also destroyed the
nanotube morphology.
4.5.2. Calcination
The level of sodium ions in titanate nanotubes had been iden-
tified as an important factor that can influence thermal stability
and phase transformation of the materials [53]. Hence, the ther-
mal behavior of titanate nanotubes could be categorized into three
groups: high, medium and low or non-sodium titanate nanotubes.
The structural changes in high, medium and low sodium titanate
nanotubes generally follow the sequence as shown in Fig. 4 [53].
High sodium titanate nanotubes without acid washing showed
the best thermal stability and it could remain its nanotubu-
lar structure when increasing temperature from 300 to 500 ◦ C
[85]. The coexistence of sodium hexatitanate (Ti6 O13 ) and rutile
appeared after calcined at 600–800 ◦ C. A few studies found that
the tubular structures of medium sodium titanate nanotubes can
be maintained up to 300–400 ◦ C [53]. Conversion of titanate nano-
tubes into anatase occurred at about 500 ◦ C. Meanwhile, low or
 Y.L. Pang et al. / Applied Catalysis A: General 481 (2014) 127–142
Fig. 5. The structure models of (a) anatase TiO2 [88], (b) trititanate H2 Ti3 O7 [87],
(c) dititanate H2 Ti2 O5 ·H2 O [88] and (d) lepidocrocite-type titanate (Hx Ti2−x/4 Vx/4 O4 ,
where x ∼ 0.7, V represents vacancy) [87].
non-sodium titanate nanotubes could only retained the tubu-
lar shape at 250–350 ◦ C. Further increment of temperature, they
transformed into the intermediate phases of hydrogen Ti6 O13 , fol-
lowed by monoclinic and metastable TiO2 (B) [58]. Anatase usually
appeared during calcination temperature above 450 ◦ C. This phase
transformation temperature might be different at varying synthetic
conditions.
In addition, annealing as-prepared titanate nanotubes at ele-
vated temperature in an oxygen-poor environment such as a
pure helium, nitrogen, argon gas atmosphere or vacuum condi-
tion would create oxygen vacancies [86]. Zhang et al. [85] reported
that single electron trapped oxygen vacancy and hydrogen vacancy
were produced during intralayered dehydration of titanate nano-
tubes. However, these as-formed oxygen vacancies on titanate
nanotubes will slowly disappeared while the material was exposed
to air, even at room temperature [86]. Hence, modification of
titanate nanotubes can be considered to stabilize the oxygen vacan-
cies.
5. Crystal structure, formation mechanism and
structural-activity of titanate nanotubes
A number of crystalline structures for as-synthesized titanate
nanotube materials including anatase TiO2 [84], lepidocrocite-
type titanate (Hx Ti2−x/4 Vx/4 O4 , where x ∼ 0.7, V represents
vacancy) [87], dititanate acid (Nax H2−x Ti2 O5 ·H2 O, where 0 < x < 2)
[54,88], trititanate acid (Nax H2-x Ti3 O7 ) [53] and tetratitanate
acid (H2 Ti4 O9 ·H2 O) [89] had been proposed. Fig. 5 shows the
structure models of anatase TiO2 [88], H2 Ti3 O7 [87], H2 Ti2 O5 ·H2 O
[88] and lepidocrocite-type titanate (Hx Ti2−x/4 Vx/4 O4 , where
x ∼ 0.7, V represents vacancy) [87]. The crystal system of anatase
is tetragonal, lepidocrocite-type titanate and H2 Ti2 O4 (OH)2 are
orthorhombic, while H2 Ti3 O7 and H2 Ti4 O9 ·H2 O are monoclinic
structures. Similar to anatase TiO2 , all titanate acids consist of
edge-shared TiO6 octahedra and zigzag ribbons configurations
[90]. The arrangement of layers of TiO6 octahedrons is diverse in
different crystal structures. Three edge-sharing octahedral is in
the unit cell of trititanate acid while four edge-sharing octahedra
is in tetratitanate acid. Particularly, titanate acid Tin O2n+1 (where
n = 3–5) consists of parallel corrugated layers of edge-sharing
TiO6 octahedra, where octahedra joins at the corners and stepped
137
Fig. 6. Schemes of three possible mechanisms for multi-walled titanate nanotubes
formation, include (a) the helical scrolling of a single-layer nanosheet, (b) the curv-
ing of several conjoined nanosheets and (c) direct production of a multi-layered
nanotube [72].
by every nth octahedral [77]. Lepidocrocite-type titanate lattice
is composed of two-dimensional (2D) lepidocrocite-type or flat
host layers sheets in which TiO6 octahedra are connected to each
other via edge-sharing [91]. This depicts that the preparation
conditions play an important role in determining morphologies
and microstructures of titanate nanostructures.
The mechanism of formation of the various types of nanotubes
has been the subject of intense research. According to the liter-
ature, acid washing after hydrothermal was essential for sodium
exchange purpose as well as formation of nanotubes [51,84]. Mean-
while, strong evidences state that the key step for the formation
of nanotubes is during hydrothermal reaction [53]. However, it
is widely accepted that the formation mechanism of nanotubes
follows the 3D → 2D → 1D model [92]. Chemical bonds in the start-
ing 3D TiO2 will break under hydrothermal reaction, in which 2D
layered entities will be formed and then convert into 1D nano-
tubes through sheet folding mechanism. Bavykin’s group [72] had
proposed three possible mechanisms for the formation of multi-
walled titanate nanotubes. These included (a) the helical scrolling
of a single-layer nanosheet, (b) the curving of several conjoined
nanosheets and (c) direct production of a multi-layered nanotube,
as shown in Fig. 6.
Among these mechanisms of formation, the wrapping of a sin-
gle nanosheet has been favored by a number of research groups
[53,93]. This would result in the inequality number of wall layers
on two sides of nanotubes and interlayer spacing of approximately
0.78 nm. Zhang et al. [93] indicated that the imbalance charge of
hydrogen or sodium ion concentration at two different sides of
nanosheets may introduce excess surface energy to strain driv-
ing the peeling process and resulting in wrapping of nanotubes.
This kind of hydrogen deficiency driving force in surface layers is
unable to bend multiple layers of nanosheets although it can be
done through mechanical stress. On the other hand, Kukovecz et al.
[94] argued that particle surfaces were cut up into small terraces
before curving up of the nanoloop resulted from nanosheets.
Meanwhile, Bavykin et al. [72] demonstrated the imbalance
width of different layers during crystallization and rapid growth
of layers which might lead to nanosheet with a single curved layer
138 Y.L. Pang et al. / Applied Catalysis A: General 481 (2014) 127–142
and then several conjoined curved layer in order to decrease the
excess energy. They suggested that formation of nanotubes under-
went several steps: (i) dissolution of TiO2 accompanied by epoxial
growth of layered nanosheets; (ii) dissolution–recrystallization of
nanosheets; (iii) curving of nanosheets and (iv) exchange of sodium
ions to proton during washing of nanotubes.
Morgado et al. [95] found that small crystal TiO2 possessed fast
dissolution rate which accelerated recrystallization of nanosheets.
When the rate of nanosheet crystallization was higher than the rate
of their curvature into nanotubes, the formation of thicker planar
entities with multiple layers would be favored before they could
be easily rolled up. Thus, small crystal TiO2 led to low yields of
nanotubes and non-hollowed layered particles with thicker walls
and larger external diameter. They also suggested that this was the
reason for those researchers who were not able to obtain nanotubu-
lar materials from freshly precipitated amorphous TiO2 due to the
fast dissolution rate under alkali solution.
Both titanate and anatase structures contain zigzag ribbons of
TiO6 octahedra that share four edges with each other [96], while
rutile structure is composed of TiO6 octahedrons that share two
edges with each other [97]. This can be surmised that anatase
easily converted into titanate nanotubes, while rutile requires a
higher temperature in order to convert into titanate nanotubes
due to the structural dissimilarity. Dawson et al. [69] reported
that pure anatase phase with large particle size able to allow the
dissolution–recrystallization processes and no other phase impu-
rity to impede greater growth for the nanotubes. The mixture of
phases in the starting material would inhibit the reaction of anatase
and thus produced shorter tubes than the pure phase. The rutile
structure cannot simply undergo the dissolution and exfoliation
processes, thereby, total structural destruction and recrystalliza-
tion become necessary [97].
Zhu et al. [96] revealed that the phase transition from titanate
to anatase was a topochemical reaction process, rather than dis-
solution of titanate and recrystallization of anatase. They found
that the anatase retained the morphology of the reactant titanate
because the large structural units remained relatively unchanged
during dehydration process. It is noted that more investigations,
evidences and discussions need to be done in order to support the
formation mechanism of titanate nanotubes.
In general, photocatalytic activity of titanate nanotubes was
determined by three structural factors which were specific surface
area, pore volume and amount of anatase phase in titanate nano-
tubes [98]. It was often reported that larger specific surface area
would allow more reactants to adsorb to the surface of the cata-
lyst, which was beneficial for the subsequent oxidation reaction.
Higher pore volume would result in more rapid diffusion of var-
ious compounds during the photocatalytic reaction. Moreover, it
is well known that more anatase phase can induce faster photo-
catalytic reaction. Turki et al. [58] demonstrated that large surface
area was not the only crucial factor to achieve the highest pho-
tocatalytic activity. A good crystalline quality of anatase phase and
coherent crystallographic domains of TiO2 must be fulfilled in order
to achieve high photo-degradation activity.
A straight forward comparison of the photocatalytic activity
results obtained in different studies is difficult because final prod-
ucts with different specific structural and optical properties can
be obtained. For example, in the literature, the amount of sodium
content present in the titanate nanotubes might lead to different
photocatalytic activities. According to Qamar et al. [99], titanate
nanotubes with high, medium or free sodium content without
calcination showed no photocatalytic activity for the degradation
of the amaranth dye. The photocatalytic activities of free sodium
titanate nanotubes calcined at 400 ◦ C and medium sodium titanate
nanotubes calcined at 500 ◦ C were less active than P25 though
they reached their maximum activity with increasing calcination
temperature. In contrast, Turki et al. [58] observed that photo-
catalytic activity of free sodium titanate nanotubes calcined at
400 ◦ C showed 4 times more active than P25 for the degradation
of formic acid. Meanwhile, Lee et al. [98] found that medium con-
tent of sodium titanate nanotubes possessed higher photocatalytic
activity as compared to low and high medium content of titanate
nanotubes. However, its activity was still lower than P25 in the pho-
tocatalytic oxidation of a basic dye. All these results indicated that
the role of sodium was not precisely evaluated while the relation-
ship between photocatalytic activity and morphology of titanate
nanotubes was not clearly defined.
6. Modification of titanate nanotubes as photocatalysts
It should be noted that the photocatalytic activity of TiO2 is
limited by the light absorption properties, electron–hole recom-
bination rate, reduction and oxidation reaction rates on the surface
by the electrons and holes [1,79]. Many efforts on the modifica-
tion of titanate nanotubes have been made to modify the electronic
band structure in order to promote the separation of electron–hole
pair and improve catalytic activity. Titanate nanotubes can be mod-
ified by various strategies such as deposition of metal nanoparticles,
single- and co-doping with two or more foreign ions, coupling
with other semiconductors and hybrid with carbon nanomateri-
als. Several recent research works on the modification of titanate
nanotubes classified according to different methods have been
summarized in Table 6.
6.1. Metal-loaded
Metal loading on titanate nanotubes surface is one of the
methods used to improve the catalytic activity. Various metal
nanoparticles such as Pt [100], Au [100] and Ag [1] have
been employed in titanate nanotubes through photo-deposition,
deposition–precipitation or ion-exchange. It is well known that
Schottky junction formed by the interface of metal–semiconductor
[100]. The noble-metal particles have the property of collective free
electron charge at the interface when particles excited by light,
known as the process of surface plasmon resonance [1]. When these
noble-metal particles load on to titanate nanotubes surface, the
plasmon-induced electrons are transported from the photo-excited
noble-metal particles to the conduction band of the titanate nano-
tubes [100]. This facilitates electron–hole separation, promotes the
interfacial electron transfer process and enhances photocatalytic
performance of this hybrid catalyst [1,2]. Besides, metal ions with a
standard redox potential is more positive than the conduction band
of TiO2 can be photo-reduced to the low-valent states and then be
deposited on the surface of TiO2 nanomaterials [15].
6.2. Doping
According to previous research, metal or non-metal ion doping
tends to narrow the band gap energy, increases separation of
electron–hole pairs and inhibits recombination photo-generated
electron–hole pair [79,112]. Thereby, the overall catalytic activ-
ity of the catalyst can be enhanced. Various methods have been
introduced to prepare doped titanate nanotubes through chemical
(sol–gel, co-hydrothermal) or physical method (physical mixing).
It had been reported that the dopant cation with almost similar
ionic radius was easier to substitute Ti4+ cation and more difficult
to replace an oxygen ion with other anion due to the difference
in charge states and ionic radius [79]. In principle, the formation
of oxygen vacancies on titanate nanotubes during doping process
leads to the creation of unpaired electrons or Ti3+ centers in order
to keep the charge balance [86]. The excess electrons located on the
Table 6
Several recent studies on modified-titanate nanotubes and their effects.
Modifier Preparation methods, modifier source
Metal deposition
Copper (Cu) Hydrothermal, photodeposition
Source: 0.5–2.0 wt.% copper (II) nitrate
pentahemihydrate
Gold (Au) and platinum (Pt) Hydrothermal, photodeposition
Source: 1 wt.%
chloro-platinic acid (H2 PtCl6 ·6H2 O) and auric
chloride acid
Silver (Ag) Hydrothermal, deposition–precipitation
Source: 1 at.% silver nitrate (AgNO3 )
Ag Hydrothermal, deposition–precipitation and
AgX photo-deposition
Ag–AgX (X = Cl, Br, I) Source: AgNO3 , hydrogen halide
Single- and co-doping
Carbon (C) and C–tungsten (W) Sol–gel, hydrothermal
Source: 1 at.% melamine borate (as C dopant)
and tungstic acid (C and W dopants)
Silver ion (Ag3+ ) Hydrothermal, photo-reduction
Source: 2.5 wt.% AgNO3
Aluminum ion (Al3+ )–iron ion (Fe3+ ) Co-hydrothermal
Source: 0.1–5.0 at.% aluminum chloride and
iron(III) nitrate
Pt–nitrogen (N) Hydrothermal, photo-deposition,
Source: 0.6 wt.% H2 PtCl6 and 0.5 wt.%
ammonium hydroxide
N Co-hydrothermal
Source: Weight ratio of TiO2 :urea was 1:1–1:10
Gadolinium ion (Gd3+ )–N Hydrothermal, ion-exchange
Source: Molar ratio of gadolinium (III)
nitrate:ammonium chloride were 1:1, 1:2 and
2:1
C–N–sulfur (S) Hydrothermal, physical mixing, calcination
Source: Molar ratio of Ti:thiourea were 1:1, 1:2
and 1:3
Coupled semiconductors/binary
composites
ZnO Hydrothermal, wet-impregnation
Source: 10–30 wt.% zinc acetate
Modification effect
Cu served as the electron mediator to prolong
the retention time of photo-generated radicals
Au or Pt formed a higher potential gradient of
Schottky junction than at the TiO2 /electrolyte
interface, which facilitated charge separation
Ag nanoparticles acted as additional separator
of the electron and hole pair
Ag–AgCl–TNTs hybrid enhanced visible light
absorption which induced by the surface
plasmonic effect
W acted as an electron and hole recombination
center. C-doped material exhibited high
surface area, enhanced optical properties and
charge separation
Ag nanoparticles acted as electron acceptor
and transferred to the adsorbed O2 effectively
Fe3+ acted as shallow trapping sites. Al–Fe
co-doping narrowed band gap energy and
promoted the generation of OH radicals
Pt doping used as electron trapper to prevent
charge recombination and increased stability
of N in catalyst
N doping narrowed the band gap of TiO2 and
served as electron trapper to inhibit the charge
recombination
Gd3+ acted as electron-trapped agent to inhibit
the charge recombination. Co-doping inhibited
the particle growth and resulted in higher
crystallinity
C acted as a photo-sensitizer to inject electron
to TNTs. N–S co-doping modified the electronic
band structure and narrowed the band gap
energy of TNTs
ZnO–TNTs facilitated charge separation,
increased the lifetime of the charge carriers
and interfacial charge transferred to adsorbed
substrates
(i) Application Reference
(ii) Important findings
(i) Photocatalytic degradation of bisphenol A (BPA) Doong et al.
(ii) The pseudo first-order rate constants for BPA [15]
photo-degradation by 1 wt.% Cu-deposited TiO2 /TNTs
were 1.8–5.2 and 4.3–12.7 times higher than those of
pure Degussa P25 and anatase TiO2
(i) Photocatalytic degradation of methyl orange Zhao et al.
(ii) The degradation rates of methyl orange solution [100]
were 96.1%, 95.1%, 88.9% and 82.6% in the presence of
Au–TNTs, Pt–TNTs, TNTs and P25 catalysts,
respectively after 40 min
(i) Photocatalytic degradation of caffeine Plodinec et al.
(ii) The hydrogenated TNTs@Ag sample showed two [1]
times photocatalytic activity higher than the
Y.L. Pang et al. / Applied Catalysis A: General 481 (2014) 127–142
hydrogenated TNTs
(i) Photocatalytic degradation of methylene blue Tang et al. [2]
(ii) Ag–AgCl–TNTs was found to possess the highest
methylene blue removal rate as compared to Ag–AgCl
and Ag–AgBr
(i) Photocatalytic degradation of 4-chlorophenol Neville et al.
(ii) The degradation rates of 4-chlorophenol were [18]
36.1%, 31.4%, 24.3% and 10% in the presence of C–TNTs,
C–W–TNTs, TNTs and P25, respectively after 1 h
(i) Photocatalytic degradation of ceftiofur sodium (CFS) Pugazhenthiran
(ii) Photocatalytic degradation of CFS was about 6 et al. [3]
times higher than bare TNTs
(i) Photocatalytic degradation of humic acids Yuan et al.
(ii) Al–Fe co-doping at 0.25: 0.75 at.% showed the [101]
highest degradation efficiency 79.4% as compared to
1 at.% Fe-doped (77.4%) and 1 at.% Al-doped (60.4%)
(i) Photocatalytic production of hydrogen from Slamet et al.
glycerol–water mixture [102]
(ii) Pt–N–TNTs produce H2 amounts approximately 13
times, 6 times, and 20% higher as compared with P25,
TNTs, and Pt–TNTs, respectively
(i) Photocatalytic degradation of rhodamine B Hu et al. [103]
(ii) N-doped TNTs at weight ratio of 1:2 exhibited the
highest photocatalytic activity than the undoped TNTs
and P25
(i) Photocatalytic degradation of rhodamine B Liu et al. [104]
(ii) The degradation rates were 81.4%, 61.5% and 37.1%
in the presence of 1:1 Gd–N–TNTs, Gd–TNTs and
N–TNTs, respectively after 2 h
(i) Photocatalytic degradation of toluene Dong et al.
(ii) The molar ratio of Ti:triourea of 1:1 and 1:2 showed [105]
the highest pseudo first-order reaction rate constants
under UV–vis light and visible light, respectively
(i) Photocatalytic degradation of rhodamine B Wang et al.
(ii) 20 wt.% of ZnO–TNTs nanocomposites exhibited the [106]
highest photocatalytic activity than pure TNTs, P25
and ZnO
139
140 Y.L. Pang et al. / Applied Catalysis A: General 481 (2014) 127–142

oxygen vacancy states could serve as the centers to capture elec-

Kim et al. [109]

Qianqian et al.
trons [86]. This facilitates the efficient adsorption of oxygen and

Li et al. [108]
Grandcolas
et al. [107]
water on titanate surface and reaction with electrons and holes,

Jiang et al.
Reference

leading to the formation of O2 − and OH− , respectively. However,

[111]
[110]
these oxygen vacancies can also act as recombination sites in the
presence of high amount of dopants [112]. Hence, an optimum
doping amount (with an acceptable amount of oxygen vacancies)

(ii) TiO2 –TNTs composites on the CNFs surface showed

within 4 h), this followed by 10 wt.% CdS–TNT and TNT

photocatalytic activity (more than 90% within 40 min)

could exhibit better visible light absorption and higher photocat-
(i) Photocatalytic degradation of diethylsulfide (DES)

highest degradation efficiency (94%) as compared to

(ii) 10 wt.% PbS–TNT showed the best activity (95%

the highest photocatalytic activity as compared to

(i) Photocatalytic degradation of gaseous CH3 CHO

(ii) CNTs–TNTs at a mass ratio of 0.10 showed the

alytic activity.
efficiency and resistance to sulfate deactivation
(i) Photocatalytic degradation of methyl orange

(i) Photocatalytic degradation of methyl orange

photocatalytic activity in terms of DES removal
(ii) 4 wt.% WO3 –TNTs composite enhanced the

Titanate nanotubes doped with metal ions such as transition

(i) Photocatalytic degradation of rhodamine B

pure TNTs (86%) and CNTs (21%) after 40 min

(ii) 5 wt.% graphene–TNTs possess the best

metals (Al [101] and Fe [101]), rare earth metals (Gd [104]) and
noble metals (Ag [3]) have been widely investigated to enhance
the photocatalytic activity of the titanate nanotubes. Among these
elements, Fe3+ has been found to be the most used dopant due to its
low band gap energy of 2.6 eV and the radius of Fe3+ (0.64 Å) is quite
TNTs–CNFs and TiO2 –CNFs

similar to that of Ti4+ (0.68 Å) [113]. A new energy level within the
(ii) Important findings

band gap of TiO2 would be introduced when Fe3+ ions substituted

Ti4+ sites, which enabled visible light absorption. This is attributed
to the charge transfer transition between the d-state electron of the
(i) Application

dopant and the conduction band of TiO2 or from valence band to

the new energy level [114].
Moreover, Fe3+ dopants could act as hole and electron trapped
sites and change the oxidation states to Fe4+ and Fe2+ , respectively
[113,114]. According to crystal field theory, the electronic config-
reduced charge recombination. Adsorptivity of

urations of Fe4+ and Fe2+ ions are relatively unstable as compared

light, facilitated the charge photo-generated

PbS and CdS acted as the sensitizers to light

The UV light absorption and surface area of

to Fe3+ ions (half-filled high spin d5) [114]. The electrons and holes
WO3 –TNTs increased absorption of visible

CNTs were good electronic conductor and

organic pollutant on CNTs was increased

absorption, large BET surface areas, and

are easily de-trapped to the adsorbed oxygen and surface hydroxyl

groups, respectively and restores to its half-filled electronic con-
graphene–TNTs and generated the
separation and prevented charge

Graphene acted as a sensitizer in

figuration. Therefore, the recombination rate of electron and hole

pairs can be retarded with the introduction of Fe3+ . Besides, the
increased separation charge

CNFs–TNTs were increased

photo-generated electrons

effect of Fe doping is highly dependable on the dopant amount as

the doping sites could also serve as charge recombination centers
Modification effect

when present in high concentration [113]. The additional benefit of

Fe doping is that the trace amount of Fe that might be leaching out
recombination

from the photocatalyst could undergo fenton-like reactions [115].

Cationic doping materials suffer from thermal instability
Carriers

and cationic doping could be localized in the d-state deep of

titanate band gap, which usually act as recombination center for
electron–hole pairs [103]. Hence, non-metal ion doping appears
Source: 1, 5 and 10 wt.% commercially available
Microwave-assisted hydrothermal, dip-coating

polyacrylonitrile solutions via electrospinning

as a promising way to avoid the disadvantages experienced by

vapor deposition of ethanol as carbon source
Source: CNTs were prepared using chemical

Note: TNTs are defined as titanate nanotubes produced through hydrothermal treatment.

the cationic doping materials. Titanate nanotubes doped with non-

metal elements show positive effect in catalytic performance.
Preparation methods, modifier source

Source: 2, 4 and 8 wt.% pentahydrated

Source: Lead(II) nitrate and Cd(NO3 )2

Doping titanate nanotubes with non-metal ions including N

Electrospinning, carbonization, and

Source: CNFs were prepared using

Hydrothermal, wet-impregnation

and exposed to H2 S atmosphere

[103], S [105] and C [18] have attracted considerable attentions

due to the increasing photocatalytic activity performed under
ammonium paratungstate

visible light irradiation. In 2001, Asahi et al. [116] found that

nitrogen-doped TiO2 nanomaterials exhibited superior photocat-
alytic activity among other non-metals (i.e. C, N, P, F and S) doped
Co-hydrothermal

Co-hydrothermal

TiO2 materials in various reactions. N element is the most inves-

hydrothermal

tigated dopant for narrowing the band gap energy of titanate

graphene

nanotubes as photocatalysts. N-doped TiO2 could be obtained by

adding different nitrogen sources such as ammonia, ammonium
salts or urea [103,116]. The reason for the improvement of cat-
alytic activity has been often related to the decrease of the band
gap energy of TiO2 , which is due to either N 2p states mixed with
Hybrid with carbon nanomaterials

O 2p states or the creation of an isolated impurity level above

Lead(II) sulfide (PbS) and CdS

valence band [79,116]. In general, substitutional type of N-doping

Carbon nano fibers (CNFs)

accompanied by oxygen vacancy is favored when the doping is car-

Carbon nanotubes (CNTs)
Tungsten trioxide (WO3 )

ried out under highly reducing condition; whereas interstitial type

N-doping takes place under oxygen-rich condition [103]. Substi-
tutional (N Ti O and Ti O N) and interstitial (characteristic NO)
Table 6 (Continued)

states created localized states above the valence band, resulting in

Graphene

visible light response and faster photocatalytic degradation.

Modifier

It should be bear in mind that the content of non-metal could

decease during annealing process, thereby, reducing its catalytic
activity. Hence, co-doping materials have been investigated to
 Y.L. Pang et al. / Applied Catalysis A: General 481 (2014) 127–142
compensate the drawbacks of each single doped titanate nano-
tubes. For example, co-doping titanate nanotubes such as Gd3+ N
[104], Pt N [102], and B N [117] are proposed to be more advan-
tageous over single nitrogen doping. Pt is used as electron trapper
[102], Gd3+ acted as a sensitizer to increase light absorption [104],
while B could minimize the oxygen vacancies produced by N doping
[117]. Consequently, photo-charge separation, electron transfer
ability and photocatalytic activity can be enhanced as well as main-
taining non-metal doping stability in the surface of photocatalysts.
6.3. Coupled semiconductors/binary composites
It has been reported that good matching of the conduction and
valence bands of two semiconductors could enable efficient charge
carrier transfer from one to another [106]. Wide band gap energy
of titanate nanotubes coupling with small band gap semiconductor
could result in the formation of heterojunctions [107]. This allows
electron to be injected from small band gap of the semiconductor
(sensitizer) to its neighbor counterpart (titanate nanotubes) and
improve separation of electron–hole pairs.
The modification of titanate nanotubes with quantum dots of
narrow band gap semiconducting materials such as CdS [108] and
PbS [108] is more promising as couple nanomaterials since it can
improve the photocatalytic activity effectively. CdS with a narrow
band gap of 2.4 eV allows CdS to act as photo-sensitizer to absorb
visible light irradiation [118]. The higher conduction band of CdS of
about 0.5 eV with respect to that of TiO2 could also enhance charge
separation [118]. Owning to the photo-generated electron transfer
from CdS to TiO2 , this charge separation restrains charge recom-
bination. Thus, the coupling of CdS/titanate nanotubes has been
applied to enhance the photocatalytic degradation organic pol-
lutants [108]. In addition, the incorporation of WO3 into titanate
nanotubes has also gained much attention due to smaller band
gap energy of WO3 (2.8 eV) and higher corrosion stability in aque-
ous solution [107]. Both the valence band and conduction band of
WO3 are lower than those of titanate nanotubes. Photo-generated
electrons would transfer from titanate nanotubes to that of WO3
and the holes transferred from WO3 to that of titanate nanotubes.
Besides, titanate nanotubes composited with wide band gap energy
of semiconductor such as ZnO (3.2 eV) [106] could suppress the
charge recombination for higher photocatalytic activity.
6.4. Hybrid with carbon nanomaterials
Recently, titanate nanotubes composites with carbon nano-
tubes, graphene and fullerene have attracted increasing attention.
Carbon nanotubes show excellent mechanical strength, thermal
conductivity, unique electronic properties and thermal stability
[110]. Meanwhile, graphene with zero band gap possesses excel-
lent electrical and optical properties, large specific surface area and
outstanding mechanical properties [111]. In addition, graphene can
act as photo-sensitizer to transfer electron to the titanate nano-
tubes surface due to the more positive Fermi level of graphene
[119]. Solid fullerenes in the form of C60 with band gap energy of
1.6–1.9 eV and a unique structure of delocalized ␲ electrons may act
as a sensitizer when attached to the surface of titanate nanotubes
[120]. These outstanding structural characters enable them to be
used as promising materials for environmental catalysts. In short,
these hybrid materials showed large surface area with high quality
of active sites and able to inhibit the recombination electron–hole
pairs.
7. Conclusions
A series of comprehensive review on TiO2 -based nanotubes are
discussed, which include three main routes used to synthesize TiO2
141
nanotubes, factors affecting the formation of titanate nanotubes by
hydrothermal treatment, crystal structure, mechanism, structural-
activity correlation and modification of TiO2 -based nanotubes.
Template-assisted method allows the fabrication of materials with
regular and controlled morphology by adjusting the morphology
of the applied template. However, it encounters difficulties such
as time consuming due to pre-fabrication and post-removal tem-
plate and contamination may occur during post-removal template.
Electrochemical anodization method can build highly ordered self-
organized TiO2 nanotubes immobilized on a titanium foil surface
with controllable pore size. However, this preparation method can
be considered as an expensive technique for large scale produc-
tion of TiO2 nanotubes and length distribution of tubes is not
well-defined over a large surface area. Meanwhile, TiO2 -based
nanotube prepared by alkaline hydrothermal method is a sim-
ple and environmental friendly synthesis technique as compared
to template-assisted and electrochemical anodization. Additional
advantages are more cost-effectiveness and larger amount of
TiO2 -based nanotubes can be produced in one-pot hydrothermal
process.
TiO2 -based catalyst has become popular to be used as photo-
catalyst in the water treatment industry owning to its low cost,
high safety and high photocatalytic activity. However, the lack of
visible light activity limits its practical applications. Hence, the
improvement in the photocatalytic activity of TiO2 -based nano-
tubes has been achieved via deposition with metal nanoparticles,
single- or co-doping with metal ions/non-metal ions, coupled with
other semiconductors and hybridized with carbon nanomaterials.
The modification on TiO2 nanotubes exerts a substantial influence
in modifying the optical, interface and bulk properties of the pure
TiO2 nanotubes, which then reflects to the charge carrier separa-
tion and transfer behavior. In short, modified TiO2 -based nanotubes
have higher potential to be utilized as photocatalyst in large scale
applications.
Acknowledgement
The authors would like to acknowledge The University of
Malaya, Kuala Lumpur, Malaysia for the financial support under
UM.C/HIR/MOHE/ENG/06 (D000006-16001).
References
[1] M. Plodinec, A. Gajović, G. Jakša, K. Žagar, M. Čeh, J. Alloys Compd. 591 (2014)
147–155.
[2] Y. Tang, Z. Jiang, Q. Tay, J. Deng, Y. Lai, D. Gong, Z. Dong, Z. Chen, RSC Adv. 2
(2012) 9406–9414.
[3] N. Pugazhenthiran, S. Murugesan, S. Anandan, J. Hazard. Mater. 263 (Pt 2)
(2013) 541–549.
[4] A. Ghicov, P. Schmuki, Chem. Commun. (2009) 2791–2808.
[5] L.L. Costa, A.G.S. Prado, J. Photochem. Photobiol. A 201 (2009) 45–49.
[6] H. Yu, J. Yu, B. Cheng, J. Lin, J. Hazard. Mater. 147 (2007) 581–587.
[7] P. Roy, D. Kim, K. Lee, E. Spiecker, P. Schmuki, Nanoscale 2 (2010) 45–59.
[8] A. El Ruby Mohamed, S. Rohani, Energy Environ. Sci. 4 (2011) 1065–1086.
[9] P. Roy, S. Berger, P. Schmuki, Angew. Chem. Int. Ed. 50 (2011) 2904–2939.
[10] D.V. Bavykin, J.M. Friedrich, F.C. Walsh, Adv. Mater. 18 (2006) 2807–2824.
[11] K.-C. Huang, S.-H. Chien, Appl. Catal. B 140/141 (2013) 283–288.
[12] T. Sekino, in: T. Kijima (Ed.), Inorganic and Metallic Nanotubular Materi-
als: Recent Technologies and Applications, Springer-Verlag, Berlin, Germany,
2010, pp. 17–32.
[13] J. Chen, H. Wang, X. Wei, L. Zhu, Mater. Res. Bull. 47 (2012) 3747–3752.
[14] R.P. Antony, T. Mathews, C. Ramesh, N. Murugesan, A. Dasgupta, S. Dhara, S.
Dash, A.K. Tyagi, Int. J. Hydrogen Energy 37 (2012) 8268–8276.
[15] R.-a. Doong, S.-m. Chang, C.-w. Tsai, Appl. Catal. B 129 (2013) 48–55.
[16] J. Yan, F. Zhou, J. Mater. Chem. 21 (2011) 9406–9418.
[17] Y. Ohsaki, N. Masaki, T. Kitamura, Y. Wada, T. Okamoto, T. Sekino, K. Niihara,
S. Yanagida, Phys. Chem. Chem. Phys. 7 (2005) 4157–4163.
[18] E.M. Neville, J.M.D. MacElroy, K.R. Thampi, J.A. Sullivan, J. Photochem. Photo-
biol. A 267 (2013) 17–24.
[19] H.-H. Ou, S.-L. Lo, Sep. Purif. Technol. 58 (2007) 179–191.
[20] C. Bae, H. Yoo, S. Kim, K. Lee, J. Kim, M.M. Sung, H. Shin, Chem. Mater. 20 (2008)
756–767.
142 Y.L. Pang et al. / Applied Catalysis A: General 481 (2014) 127–142
[21] J. Qiu, X. Li, X. Gao, X. Gan, B. Weng, L. Li, Z. Yuan, Z. Shi, Y.H. Hwang, J. Mater.
Chem. 22 (2012) 23411–23417.
[22] J.H. Kim, X.H. Zhang, J.D. Kim, H.M. Park, S.B. Lee, J.W. Yi, S.I. Jung, J. Solid State
Chem. 196 (2012) 435–440.
[23] A.M. Md Jani, D. Losic, N.H. Voelcker, Prog. Mater. Sci. 58 (2013) 636–704.
[24] P. Hoyer, Langmuir 12 (1996) 1411–1413.
[25] J.H. Lee, I.C. Leu, M.C. Hsu, Y.W. Chung, M.H. Hon, J. Phys. Chem. B 109 (2005)
13056–13059.
[26] C. Bae, Y. Yoon, H. Yoo, D. Han, J. Cho, B.H. Lee, M.M. Sung, M. Lee, J. Kim, H.
Shin, Chem. Mater. 21 (2009) 2574–2576.
[27] L. Yuan, S. Meng, Y. Zhou, Z. Yue, J. Mater. Chem. A 1 (2013) 2552–2557.
[28] M. Karaman, F. Sarıipek, Ö. Köysüren, H.B. Yıldız, Appl. Surf. Sci. 283 (2013)
993–998.
[29] M. Müllner, T. Lunkenbein, M. Schieder, A.H. Gröschel, N. Miyajima,
M. Förtsch, J. Breu, F. Caruso, A.H.E. Müller, Macromolecules 45 (2012)
6981–6988.
[30] S.J. Ku, G.C. Jo, C.H. Bak, S.M. Kim, Y.R. Shin, K.H. Kim, S.H. Kwon, J.B. Kim,
Nanotechnology 24 (2013) 085308.
[31] X. Hou, F. Li, G. He, J. Zhang, Ceram. Int. 40 (2014) 5811–5815.
[32] X. Meng, M.N. Banis, D. Geng, X. Li, Y. Zhang, R. Li, H. Abou-Rachid, X. Sun,
Appl. Surf. Sci. 266 (2013) 132–140.
[33] D. Gong, C.A. Grimes, O.K. Varghese, W. Hu, R.S. Singh, Z. Chen, E.C. Dickey, J.
Mater. Res. 16 (2001) 3331–3334.
[34] S. Sreekantan, Z. Lockman, R. Hazan, M. Tasbihi, L.K. Tong, A.R. Mohamed, J.
Alloys Compd. 485 (2009) 478–483.
[35] E. Şennik, Z. Çolak, N. Kılınç, Z.Z. Öztürk, Int. J. Hydrogen Energy 35 (2010)
4420–4427.
[36] N. Baram, Y. Ein-Eli, J. Phys. Chem. C 114 (2010) 9781–9790.
[37] I. Roman, R.D. Trusca, M.-L. Soare, C. Fratila, E. Krasicka-Cydzik, M.-S. Stan, A.
Dinischiotu, Mater. Sci. Eng. C 37 (2014) 374–382.
[38] M. Nischk, P. Mazierski, M. Gazda, A. Zaleska, Appl. Catal. B 144 (2014)
674–685.
[39] H. Jha, P. Roy, R. Hahn, I. Paramasivam, P. Schmuki, Electrochem. Commun.
13 (2011) 302–305.
[40] V. Zwilling, E. Darque-Ceretti, A. Boutry-Forveille, D. David, M.Y. Perrin, M.
Aucouturier, Surf. Interface Anal. 27 (1999) 629–637.
[41] M. Paulose, K. Shankar, S. Yoriya, H.E. Prakasam, O.K. Varghese, G.K. Mor,
T.A. Latempa, A. Fitzgerald, C.A. Grimes, J. Phys. Chem. B 110 (2006) 16179–
16184.
[42] T. Ruff, R. Hahn, P. Schmuki, Appl. Surf. Sci. 257 (2011) 8177–8181.
[43] J.M. Macak, S.P. Albu, P. Schmuki, Phys. Status Solidi RRL 1 (2007) 181–183.
[44] D. Kim, A. Ghicov, S.P. Albu, P. Schmuki, J. Am. Chem. Soc. 130 (2008)
16454–16455.
[45] S.P. Albu, D. Kim, P. Schmuki, Angew. Chem. Int. Ed. 47 (2008) 1916–1919.
[46] Y. Ji, K.-C. Lin, H. Zheng, J.-j. Zhu, A.C.S. Samia, Electrochem. Commun. 13
(2011) 1013–1015.
[47] Q.A. Nguyen, Y.V. Bhargava, T.M. Devine, Electrochem. Commun. 10 (2008)
471–475.
[48] M.Y.A. Mollah, P. Morkovsky, J.A.G. Gomes, M. Kesmez, J. Parga, D.L. Cocke, J.
Hazard. Mater. 114 (2004) 199–210.
[49] E. Horváth, Á. Kukovecz, Z. Kónya, I. Kiricsi, Chem. Mater. 19 (2007) 927–931.
[50] Z.-Y. Yuan, B.-L. Su, Colloids Surf. A 241 (2004) 173–183.
[51] T. Kasuga, M. Hiramatsu, A. Hoson, T. Sekino, K. Niihara, Langmuir 14 (1998)
3160–3163.
[52] H. Kochkar, N. Lakhdhar, G. Berhault, M. Bausach, G. Abdelhamid, J. Phys.
Chem. C 113 (2009) 1672–1679.
[53] E. Morgado Jr., M.A.S. de Abreu, O.R.C. Pravia, B.A. Marinkovic, P.M. Jardim,
F.C. Rizzo, A.S. Araújo, Solid State Sci. 8 (2006) 888–900.
[54] S. Preda, V.S. Teodorescu, A.M. Musuc, C. Andronescu, M. Zaharescu, J. Mater.
Res. 28 (2013) 294–303.
[55] D.L. Morgan, G. Triani, M.G. Blackford, N.A. Raftery, R.L. Frost, E.R. Waclawik,
J. Mater. Sci. 46 (2011) 548–557.
[56] L.M. Sikhwivhilu, S. Sinha Ray, N.J. Coville, Appl. Phys. A: Mater. Sci. Process.
94 (2009) 963–973.
[57] N. Viriya-empikul, T. Charinpanitkul, N. Sano, A. Soottitantawat, T. Kikuchi,
K. Faungnawakij, W. Tanthapanichakoon, Mater. Chem. Phys. 118 (2009)
254–258.
[58] A. Turki, H. Kochkar, C. Guillard, G. Berhault, A. Ghorbel, Appl. Catal. B 138/139
(2013) 401–415.
[59] M. Kitano, E. Wada, K. Nakajima, S. Hayashi, S. Miyazaki, H. Kobayashi, M.
Hara, Chem. Mater. 25 (2013) 385–393.
[60] Y. Zhu, H. Li, Y. Koltypin, Y.R. Hacohen, A. Gedanken, Chem. Commun. (2001)
2616–2617.
[61] H.H. Ou, C.H. Liao, Y.H. Liou, J.H. Hong, S.L. Lo, Environ. Sci. Technol. 42 (2008)
4507–4512.
[62] B. Liu, E.S. Aydil, J. Am. Chem. Soc. 131 (2009) 3985–3990.
[63] E. Morgado Jr., M.A.S. de Abreu, G.T. Moure, B.A. Marinkovic, P.M. Jardim, A.S.
Araujo, Mater. Res. Bull. 42 (2007) 1748–1760.
[64] J. Huang, Y. Cao, Z. Deng, H. Tong, J. Solid State Chem. 184 (2011) 712–719.
[65] R. Camposeco, S. Castillo, I. Mejia-Centeno, J. Navarrete, R. Gómez, Mater.
Charact. 90 (2014) 113–120.
[66] H.-K. Seo, G.-S. Kim, S.G. Ansari, Y.-S. Kim, H.-S. Shin, K.-H. Shim, E.-K. Suh,
Sol. Energy Mater. Sol. Cells 92 (2008) 1533–1539.
[67] Y. Lan, X. Gao, H. Zhu, Z. Zheng, T. Yan, F. Wu, S.P. Ringer, D. Song, Adv. Funct.
Mater. 15 (2005) 1310–1318.
[68] E.K. Ylhäinen, M.R. Nunes, A.J. Silvestre, O.C. Monteiro, J. Mater. Sci. 47 (2012)
4305–4312.
[69] G. Dawson, W. Chen, T. Zhang, Z. Chen, X. Cheng, Solid State Sci. 12 (2010)
2170.
[70] A.L. Papa, N. Millot, L. Saviot, R. Chassagnon, O. Heintz, J. Phys. Chem. C 113
(2009) 12682–12689.
[71] L. Miao, S. Tanemura, T. Jiang, M. Tanemura, K. Yoshida, N. Tanaka, G. Xu,
Superlattices Microstruct. 46 (2009) 357–364.
[72] D.V. Bavykin, V.N. Parmon, A.A. Lapkin, F.C. Walsh, J. Mater. Chem. 14 (2004)
3370–3377.
[73] D.V. Bavykin, B.A. Cressey, F.C. Walsh, Aust. J. Chem. 60 (2007) 95–98.
[74] D.V. Bavykin, A.N. Kulak, F.C. Walsh, Cryst. Growth Des. 10 (2010) 4421–4427.
[75] G. Hasegawa, K. Kanamori, Y. Sugawara, Y. Ikuhara, K. Nakanishi, J. Colloid
Interface Sci. 374 (2012) 291–296.
[76] S. Sreekantan, L.C. Wei, J. Alloys Compd. 490 (2010) 436–442.
[77] R. Ma, K. Fukuda, T. Sasaki, M. Osada, Y. Bando, J. Phys. Chem. B 109 (2005)
6210–6214.
[78] Y. Ma, Y. Lin, X. Xiao, X. Zhou, X. Li, Mater. Res. Bull. 41 (2006) 237–243.
[79] X. Chen, S.S. Mao, Chem. Rev. 107 (2007) 2891–2959.
[80] C.T. Nam, J.L. Falconer, L.M. Duc, W.-D. Yang, Mater. Res. Bull. 51 (2014) 49–55.
[81] K. Das, S.K. Panda, S. Chaudhuri, J. Cryst. Growth 310 (2008) 3792–3799.
[82] B. Santara, P.K. Giri, Mater. Chem. Phys. 137 (2013) 928–936.
[83] M.A. Khan, H.T. Jung, O.B. Yang, J. Phys. Chem. B 110 (2006) 6626–6630.
[84] T. Kasuga, M. Hiramatsu, A. Hoson, T. Sekino, K. Niihara, Adv. Mater. 11 (1999)
1307–1311.
[85] M. Zhang, Z. Jin, J. Zhang, X. Guo, J. Yang, W. Li, X. Wang, Z. Zhang, J. Mol. Catal.
A: Chem. 217 (2004) 203–210.
[86] X. Pan, M.Q. Yang, X. Fu, N. Zhang, Y.J. Xu, Nanoscale 5 (2013) 3601–3614.
[87] R. Ma, Y. Bando, T. Sasaki, Chem. Phys. Lett. 380 (2003) 577–582.
[88] J.N. Nian, H. Teng, J. Phys. Chem. B 110 (2006) 4193–4198.
[89] C. Huang, K. Zhu, M. Qi, Y. Zhuang, C. Cheng, J. Phys. Chem. Solids 73 (2012)
757–761.
[90] P.-T. Hsiao, K.-P. Wang, C.-W. Cheng, H. Teng, J. Photochem. Photobiol. A 188
(2007) 19–24.
[91] Y. Mao, S.S. Wong, J. Am. Chem. Soc. 128 (2006) 8217–8226.
[92] Y.Q. Wang, G.Q. Hu, X.F. Duan, H.L. Sun, Q.K. Xue, Chem. Phys. Lett. 365 (2002)
427–431.
[93] S. Zhang, L.M. Peng, Q. Chen, G.H. Du, G. Dawson, W.Z. Zhou, Phys. Rev. Lett.
91 (2003) 2561031–2561034.
[94] Á. Kukovecz, M. Hodos, E. Horváth, G. Radnóczi, Z. Kónya, I. Kiricsi, J. Phys.
Chem. B 109 (2005) 17781–17783.
[95] E. Morgado Jr., M.A.S. De Abreu, G.T. Moure, B.A. Marinkovic, P.M. Jardim, A.S.
Araujo, Chem. Mater. 19 (2007) 665–676.
[96] H.Y. Zhu, Y. Lan, X.P. Gao, S.P. Ringer, Z.F. Zheng, D.Y. Song, J.C. Zhao, J. Am.
Chem. Soc. 127 (2005) 6730–6736.
[97] M.J. Li, Z.Y. Chi, Y.C. Wu, J. Am. Chem. Soc. 95 (2012) 3297–3304.
[98] C.-K. Lee, C.-C. Wang, M.-D. Lyu, L.-C. Juang, S.-S. Liu, S.-H. Hung, J. Colloid
Interface Sci. 316 (2007) 562–569.
[99] M. Qamar, C.R. Yoon, H.J. Oh, N.H. Lee, K. Park, D.H. Kim, K.S. Lee, W.J. Lee, S.J.
Kim, Catal. Today 131 (2008) 3–14.
[100] Q. Zhao, M. Li, J. Chu, T. Jiang, H. Yin, Appl. Surf. Sci. 255 (2009) 3773–3778.
[101] R. Yuan, B. Zhou, D. Hua, C. Shi, J. Hazard. Mater. 262 (2013) 527–538.
[102] D. Slamet, Tristantini, M. Valentina, Ibadurrohman, Int. J. Energy Res. 37
(2013) 1372–1381.
[103] C.-C. Hu, T.-C. Hsu, S.-Y. Lu, Appl. Surf. Sci. 280 (2013) 171–178.
[104] H. Liu, G. Liu, G. Xie, M. Zhang, Z. Hou, Z. He, Appl. Surf. Sci. 257 (2011)
3728–3732.
[105] F. Dong, W. Zhao, Z. Wu, Nanotechnology 19 (2008) 365607.
[106] L.S. Wang, M.W. Xiao, X.J. Huang, Y.D. Wu, J. Hazard. Mater. 161 (2009) 49–54.
[107] M. Grandcolas, T. Cottineau, A. Louvet, N. Keller, V. Keller, Appl. Catal. B
138/139 (2013) 128–140.
[108] Q. Li, T. Kako, J. Ye, J. Mater. Chem. 20 (2010) 10187–10192.
[109] S. Kim, M. Kim, Y.K. Kim, S.-H. Hwang, S.K. Lim, Appl. Catal. B 148/149 (2014)
170–176.
[110] T. Jiang, L. Zhang, M. Ji, Q. Wang, Q. Zhao, X. Fu, H. Yin, Particuology 11 (2013)
737–742.
[111] Z. Qianqian, B. Tang, H. Guoxin, J. Hazard. Mater. 198 (2011) 78–86.
[112] R. Marschall, L. Wang, Catal. Today 225 (2014) 111–135.
[113] Y. Yacin, M. Kilic, Z. Cinar, Appl. Catal. B 99 (2010) 469–477.
[114] J. Yu, Q. Xiang, M. Zhou, Appl. Catal. B 90 (2009) 595–602.
[115] Y.L. Pang, A.Z. Abdullah, J. Hazard. Mater. 235/236 (2012) 326–335.
[116] R. Asahi, T. Morikawa, T. Ohwaki, K. Aoki, Y. Taga, Science 293 (2001) 269–271.
[117] P. Dong, Y. Wang, B. Liu, L. Guo, Y. Huang, S. Yin, J. Am. Ceram. Soc. 95 (2012)
82–84.
[118] R. Pan, Y. Wu, Z. Li, Z. Fang, Appl. Surf. Sci. 292 (2014) 886–891.
[119] G. Hu, B. Tang, Mater. Chem. Phys. 138 (2013) 608–614.
[120] M. Grandcolas, J. Ye, K. Miyazawa, Ceram. Int. 40 (2014) 1297–1302.