Overview

The goal of this thesis is to study the structural and morphological transformation of TNAs and TiO2 hybrid structure using electrochemical anodization techniques, and furthermore approach some photocatalytic applications of these materials. The important electrochemical parameters controlling the growth of the hybrid structure have been extensively studied and optimized. Also, appropriate post-fabrication processing conditions such as heat treatment conditions have been optimized for the high crystallinity TiO2. Chapter 2 offers a literature survey on the TiO2 materials, development of fabrication methods for TiO2 nanotubes, and annealing method for TiO2

crystallization. Chapter 3 covers the experimental method and instrumentation.

Chapter 4 discusses the microstructure and composition of TiO2 nanotube arrays fabricated with HF and NH4F electrolytes and their evolution during annealing. Chapter 5 discuses the structural and morphological transformation of TiO2 induced by excimer laser annealing. Chapter 6 describes the formation mechanism of TiO2 hybrid structure (TNWs/TNAs), and presents the photocatalysis applications. Finally, Chapter 7 provides a summary of the key results in this study.

Chapter 2 Literature Review

This chapter is a review of journals and references available in literature that are relevant to the research topic. Section 2.1 is an introduction of TiO2 materials. Section 2.2 is synthesis of TiO2 nanostructure, and section 2.3 is a discussion on anodic oxidization technique. Finally, the annealing method for TNAs crystallization is described in section 2.4.

2.1 Introduction of TiO2 materials

Titanium dioxide (TiO2) is the most commonly used compound of titanium. Since its commercial production in the early twentieth century, TiO2 has been widely used as a pigment in sunscreens, paints, ointments, and toothpaste. TiO2 powder is chemically inert, stable under sunlight, and is very opaque: This allows it to impart a pure and brilliant white color to the brown or gray chemicals that form the majority of household plastics. However, in 1972, Fujishima and Honda discovered the phenomenon of photocatalytic splitting of water on a TiO2 electrode under ultraviolet (UV) light. Since then, enormous efforts have been devoted to the research of TiO2 [36].

TiO2 can exist in three distinct crystalline polymorphs: anatase, rutile, or brookite crystalline phase, respectively shown in Fig. 2.1 [37]. From Fig. 2.1, all three crystal structures are made up of distorted octahedra, each one representing a TiO6 unit, where each Ti⁴⁺ is at the centre of the unit and coordinates six O^2- ions. The manner in which the octahedra assemble to form a TiO6 based chain is different and characteristic of each polymorph. In these three phases, rutile and anatase are the most commonly synthesized phases. Anatase and brookite are metastable phases and convert into rutile at high

temperature, usually above 600 °C [38]. Table 2.1 lists some of the key properties [37,39]

of the three TiO2 polymorphs. Both rutile and anatase have a tetragonal crystal structure, whereas brookite has an orthorhombic symmetry. Rutile is the densest phase and has the highest refractive index, while anatase is characterised by the widest band-gap (~3.2eV) [39]. The properties (density, band-gap and refractive index) of brookite fall between those of rutile and anatase.

2.2 Synthesis of TiO2 nanostructure

The wide ranging properties of TiO2 have generated a great deal of interest in many different fields. Many researchers have focused their studies on the production and processing of nanostructure TiO2. One of the most obvious advantages provided by nano-materials is that of the greatly increased surface area offered by small nanoparticles. For example, in DSSCs applications, the active electrode in the DSSC is composed of a high surface area TiO2 nanoparticle film that carries an anchored organic dye [2]. However, as the injected electrons, which excited from dye, diffuse through the TiO2 particle network to the collecting transparent conducting oxide (TCO) substrate, due to the randomly packed TiO2 nanoparticle (NP) films, it has been inferred that transport is limited by the residence time of electrons in traps. In recent years, in order to reduce random-walk effects and suppress potential recombination at grain boundaries, 1D morphologies such as nanowires, nanorods and nanotubes have been explored as an alternative to nanoparticle based films. Thus, various approaches such as template synthesis [40,41,42], chemical vapor deposition (CVD) [43], hydrothermal reactions [44,45], and anodic oxidization [28, 46 , 47 , 48 ] have been developed for preparing TiO2

nanostructure. Among these approaches, anodic oxidization is a relatively low cost process and represents a simple technique that can be easily automated for preparing TiO2

nanostructure. The following part of the chapter reviews the most relevant achievements regarding the research on anodic oxidization TiO2 nanotubes, along with a discussion of current ideas and understanding of the process.

2.3 Anodic oxidization technique

2.3.1 The developments of anodized TiO2 nanotubes First generation of TiO2 nanotubes

In 1999, Zwilling and co-workers anodized Ti and Ti–6Al–4V(TA6V) alloy in an electrolyte containing 0.5 mol/l chromic acid and 0.095 mol/l HF [28], while the first report on anodized TiO2 nanotubes (called first generation). A typical porous structure obtained in TA6V is shown in Fig. 2.3. It was clear that the nanoporous structure observed only formed when sufficient HF was added to the electrolyte mixture, as pure chromic acid (CA) was leading to the formation of a thin but stable oxide layer with no apparent pore structure. However, unlike anodized alumina, where tube length increases indefinitely with anodizing time, TiO2 nanotubes reach a steady state length when anodized. That is, after typically 10 to 20 minutes of anodization, the etching rate equals the dissolution rate so that the tube length does not show any further increase with additional anodizing time [28].

Second generation of TiO2 nanotubes

In subsequent work, Grimes and co-workers overcome this limitation since they used other fluorine salts (as fluorine ion source besides HF) and combined buffers, bases and milder acids to adjust the pH and fluorine ion content. Salts like KF, NH4F or NaF totally dissociate in aqueous solution and then hydrolyze with water to form HF [49,50].

Moreover, HF is a relatively mild acid and in acidic solutions (pH<3.45) more than 50%

of the fluorine exists in the form of HF. As a result pH and fluorine ion concentration are closely related (and solutions with KF, NaF or NH4F and no additional acid are basic.) The experiment worked found that they could grow nanotubes up to 4.4 m (Fig. 2.4) using a solution of 0.1M KF as fluorine source, 1M H2SO4 as acid, 0.2 M citric acid presumably serving as buffer and NaOH as base to be added until the desired pH of 4.5 was obtained [50]. Later in 2005 Grimes and coworkers reported even longer nanotubes of up to 6 m, over 17 to 20 h of anodization using the same electrolyte as before [51,52].

Third generation of TiO2 nanotubes

The third generation of nanotubes refers to smooth tubes (i.e. no ripples along the wall), prepared in organic electrolytes (some almost water-free), where the dissolution rate of the forming oxide is minimized. For clarity a comparison of SEM cross section images, taken from literature [30], of nanotubes grown respectively in an aqueous and in an organic environment are shown, respectively, in Fig.2.5. It can be observed that tubes obtained in water are much rougher an irregular (i.e. ripples along the wall) than the smooth tubes grown in organic solution. The reason was to use a viscous electrolyte, where ion diffusion is slower than in water, to increase the pH gradient between the bottom and the top of the tubes. This led to the formation of TiO2 nanotubes up to 7 m thick (compared to 0.5 m for first generation). They also attributed the smoothness and the regular morphology of the tube walls to the lower diffusion coefficient of the electrolyte which suppresses pH bursts at the pore bottom which occur when working in aqueous media. Over the last few years, TiO2 nanotube arrays with lengths of up to approximately 1000 m were achieved using a non-aqueous, polar organic electrolyte such as formamide, dimethyl sulfoxide, ethylene glycol or diethylene glycol [53,54,55].

In 2007, Grimes and co-workers published the synthesis of 0.36 mm long nanotubes [53],

practically demonstrating that the nanotube lengths was only limited by the initial titanium foil thickness. The following section will discuss the fundamental aspects and chemistry of the growth of TiO2 nanotubes by anodization.

2.3.2 The growth of TiO2 nanotubes: fundamental aspects Formation mechanism of TiO2 nanotube arrays

The formation mechanism of TNAs structures is similar to anodic aluminum oxide (AAO), which is the result of competition between field-assisted anodic oxidation, defined as the formation of the anodic layer under an applied electric field by Eqs.

(2.1)-(2.3) and chemical/field assisted dissolution of the forming oxide by Eq. (2.4), which can be regarded as dissolution promoted by the presence of fluoride ions (chemical dissolution) and by the electric field weakening the bond between Ti and O (field assisted dissolution) [56] :

Electrochemical reactions of anodic titanium oxide At Ti/Ti oxide interface:

Meanwhile, the formation mechanism of the TiO2 nanotubes at various stages is schematically illustrated by Figs. 2.6(a)-(d). Initially, field-enhanced oxidation occurs at the Ti/Ti oxide interface by Eqs. (2.5)-(2.7) when oxygen ions diffusion to the Ti layer as shown in Figs. 2.6(a). At the same time, competing field-enhanced oxide dissolution occurs at TiO2/solution interface illustrated by Fig. 2.6(b). Specifically, fluoride-containing electrolyte reacts with TiO2 to form TiF62- as described by Eqs. (2.4).

Moreover, small pores are generated and spread uniformly over the surface of the film under an electric field. When the pore to pore distance achieves a suitable value at

which the electric field of each pore would not affect to each other, the distribution of electric-field strength would change. As a consequent, increased in local field strength at the bottom of the pore in conjunction with low dissolution rate at sidewall, highly-order pore structures were formed as shown by Fig. 2.6(c). Finally, in the growth stage, fully developed TiO2 nanotube arrays are shown in Fig. 2.6(d). Field-enhanced dissolution developed the depth of pore and, therefore, the steady-state nanotubes morphology was created.

Key parameters for controlling the growth of the nanotubes

Summarizing these observations, the key parameters to be taken into account when growing anodized TiO2 nanotubes are the following:

Electrolyte

The electrolyte plays a crucial role in the growth of anodized TiO2 nanotubes, as previously discussed. The main distinction is between aqueous and organic-based, where the water content is the important rule to limit dissolution of the oxide. Moreover, the pH of the solution is also important, considering the higher dissolution rate of the oxide in an acidic environment. As shown schematically in Fig. 2.7. While the pore bottom is at a low pH, the pore mouth (top of the pores/tubes) remains under a protective environment (higher pH) by using chemical buffer species [NH4F/(NH4)2SO4]. The rapid rate of TiO2 dissolution in the first generation of nanotubes was reduced by replacing the HF acid with less aggressive solutions containing fluoride salts, raising the maximum thickness up to 2-3m [29]. This is one of the reasons why dissolution of the anodic oxide is at its lowest when using fluoride salts (some of them have basic hydrolysis) instead of hydrofluoric acid.

Mechanical stirring system

The mechanical stirring was sometimes used to accelerate the reaction rate for nanotube growth. In addition, mechanical stirring also affect the inner tube morphology of TNAs. Fig. 2.8 shows the schematic diagram of the formation of tube spatial periodicity under different conditions: (a) without stirring; (b) at medium stirring rate;

(c) at high stirring rate [57]. The smoothing effect of the tube inner surface and the acceleration of the growth are due to the redistribution of the F⁻ anions in the nanotubes.

Based on the experimental results, the mechanism has been discussed with the consideration of the local reactions and transport processes of the main reaction species.

Under this interpretation, the current oscillation and the morphology change in the pore are attributed to the redistribution of the ionic species by the fluctuation in the tube layer when there is no stirring. They can be significantly influenced by the convection above the tube layer and the slow transport process in the tubes with the existence of mechanical stirring.

Temperature

The temperature of the electrolyte affects the chemical dissolution and electrochemical etching rate in the growth of nanotube arrays via anodic oxidation of titanium. For example Grimes reported nanotube arrays were grown with a constant 10 V anode potential in an electrolyte of acetic acid plus 0.5% HF mixed in 1:7 ratio and kept at each of four different electrolyte bath temperatures: 5^oC, 25^oC, 35^oC and 50^oC.

Table 2.2 shows the variation in 10 V wall thicknesses and tube length as a function of anodization temperature. Results show that with decreasing anodization bath temperature, the length of the nanotubes increases from 120 nm at 50^oC to 224 nm at 5^oC. Also, with decreasing anodization temperature the wall thickness increases from 9 nm at 50^oC to 34 nm at 5^oC, confirming the trend of increasing nanotube wall thickness

with lower anodization temperature [58].

2.3.3 Applications of TiO₂ nanotubes arrays

The main motivation behind research on anodized TiO2 nanotubes is the possibility to investigate their impact on a wide range of technologies, including photocatalysis [59,60], chemical sensing [61,62,63] and photovoltaic devices such as DSSCs [64]. As a result, these applications are described in detail in the following sections.

Photocatalysis

Since the pioneering work by Honda and Fujishima [1], photocatalysis using various semiconductors has received much attention for their potential in the utilization of light energy. In particular, TiO2 photocatalysts have been extensively studied due to relative cheap, high chemical stability and high reactivity of photo-generated holes. Most work on TiO2 photocatalysts has been devoted to the study of reactions associated with the photodecomposition of H2O into H2 and O2. The principle of the photocatalysis is shown in Fig. 2.9. According to this simplified scheme, electron-hole pair generated upon UV excitation is trapped at the surface as spatially separated redox centers. The reactive electron reduces O2 from air initially to a superoxide and finally to hydrogen peroxide and an OH radical, whereas the reactive hole oxidizes the pollutant to its radical cation either directly or through a primarily formed OH radical produced by the oxidation of ubiquitous water. Thus, both the reductive and oxidative interfacial electron transfer processes lead to strong oxidizing agents which can induce mineralization of organic and inorganic pollutants and kill bacteria.

It has been known that TiO2 with high surface area and crystallinity can enhance catalytic abilities [65,66,67], because high surface area increases the adsorption of reactants more efficiently and crystallinity decreases the electron-hole recombination

sites. Therefore, many efforts have been directed to create high surface area structure with crystallinity, nanostructured TiO2. TiO2 nanotubes have been also studied for dehydrogenation of ethanol [59], decomposition of gaseous isopropanol into acetone and carbon dioxide [60].

Chemical sensing

Semiconducting metal-oxides such as ZnO, SnO2 and TiO2 have been widely utilized for applications in gas sensors. Their gas sensing properties are largely based on the surface reaction between the metal-oxides and adsorbed gas species. The charge transfer interactions on the surface of such metal oxides, i.e., the adsorption of negatively charged oxygen and the oxidative/reductive interaction between target gases and adsorbed oxygen, lead to the significant variation in electrical conductivity upon exposure to analyte gases.

To meet the requirements of environmental and air-quality monitoring, there have been significant efforts to enhance the sensitivity of gas sensors. Recently, many nano technological approaches have been employed to enlarge the surface area or improve the charge-collection efficiency. From this perspective, sensing materials with one-dimensional (1D) geometry are very promising for their high sensitivity and fast response speed. Thus far, various promising 1D materials such as carbon nanotubes and Si nanowires have been studied for the fabrication of high sensitivity chemical sensors.

For example, Comini et al. [61] used nanobelts for CO and NO2 sensing. The nanobelt sensors detected a few ppb levels of NO2, which was difficult to monitor with conventional sensors. Vargese et al. [62] used TiO2 nanotubes for high sensitivity sensors and the sensitivity of nanotube sensors reached up to 10³. Ryu et al. [63] created TiO2 nano-honeycomb structure by using photoelectron chemical etching and applied it as a H2

sensor. They reported that the nano-honeycomb structure enhanced the response time as well as the sensitivity.

Dye-Sensitized Solar Cells (DSSC)

Dye-sensitized solar cell (DSSC) is low cost alternative to inorganic semiconductor photovoltaic devices. Energy conversion in a DSSC is based on the injection of an electron from a photo-excited state of the sensitizer dye (typically a bipyridine metal complex) into the conduction band of semiconductor (TiO2 is by far the most employed semiconductor). The original state of the dye is subsequently restored by electron donation from the electrolyte, usually an iodide/triiodide redox couple system. The iodide is regenerated at the counter electrode by electrons passed through a load.

The large band gap semiconductor, TiO2, in DSSC is usually fabricated in the form of nanoporous structure in order to increase the contact of dye-TiO2 and simultaneously allow dye to directly contact with the redox electrolyte. Also, light absorption at the dye on nanoporous TiO2 is higher than that on flat TiO2 surface which only harvests a negligibly small fraction of the incoming light [68].

Lately, research has been directed toward synthesizing structures with a high degree of order than the random assembly of nanoparticles. A desirable morphology of the films would have the mesoporous channels or nanorods aligned parallel to each other and vertically with respect to the glass substrate. This would facilitate charge diffusion in the pores and the mesoporous film, giving easier access to the film surface, avoiding grain boundaries and allowing the junction to be formed under better control.

One approach to fabricate such oxide structures is based on anodized TiO2 nanotubes by by Frank and co-workers [64]. They reported that the nanotubes and the nanoparticles in conventional DSSCs have similar transport properties (i.e. the electron has to diffuse

through the oxide and reach the electrode to feed the circuit), however lower recombination occurs in the nanotubes because of a higher charge collection efficiency.

In addition, the nanotubes harvest the light more efficiently than conventional DSSCs, because of stronger light scattering effects, however their use in DSSCs is far from optimized.

2.3.4 Challenges of TiO2 nanotubes arrays

From the literature, it has know crystalline nanotube or nanowire based TiO2 , in contrast with the random transport path in nanoparticle, have been investigated to improve electron collection. But, there still have challenges need to overcome. One key challenge of using vertically aligned 1-D nanostructures is that, compared to mesoporous films, the 1-D nanostructures typically have a low internal surface area. For example, in DSSC application, the reported efficiency of TiO2 NT based DSSC is generally much lower than that of DSSCs based on nanoparticles and amounted to 0.61%–2.9% [10,11,27,69]. The possible reason is that the internal surface area of NT based photoanode is much smaller than that of NPs, with a lower dye loading and sunlight absorption.

It seems that the high surface area and the good electron transport cannot be satisfied simultaneously in DSSC based on simple TiO2 nanostructures of first generation. In order to satisfy both requirements we should design more complex nanostructures with a multiscale organization, in which small nanoparticles, nanowires or nanotubes are organized around long central cores connected directly to the electrode.

This is a next generation of TiO2 nanostructures that has a great potential to simultaneously provide high surface area and good electron transport properties.

2.4 Annealing of the nanotubes

Factors such as crystal structure, degree of crystallinity of the anodized TiO2

nanotubes have to be taken into account in order to optimize their use in any possible

nanotubes have to be taken into account in order to optimize their use in any possible