Structure Evolution and Tetragonal Phase Stabilization Mechanism of Electrospun Barium Titanate Nanofibers

Material Information

Structure Evolution and Tetragonal Phase Stabilization Mechanism of Electrospun Barium Titanate Nanofibers
YUH, JUNHAN ( Author, Primary )
Copyright Date:


Subjects / Keywords:
Annealing ( jstor )
Barium ( jstor )
Crystallites ( jstor )
Crystallization ( jstor )
Electric potential ( jstor )
Heat treatment ( jstor )
Perovskites ( jstor )
Polymers ( jstor )
Surface energy ( jstor )
Titanates ( jstor )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Junhan Yuh. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
Resource Identifier:
658230410 ( OCLC )


This item is only available as the following downloads:

Full Text




Copyright 2006 By Junhan Yuh


To my family.


iv ACKNOWLEDGMENTS Most of all, I am very grateful to my supervisory committee chair Dr. Wolfgang M. Sigmund and co-chair Dr. Juan C. Nino for all of their help and guidance. It is also my honor to have Dr. Hassan El-Sha ll, Dr. David Norton, and Dr. Arthur F. Hebard as my supervisory committee members. I must thank to the valuable help of my colleagues in Dr. Sigmund’s group and friends in materi als science and engineering department. Finally I sincerely and deeply appreciate the encouragement and support of my family.


v TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES............................................................................................................vii LIST OF FIGURES.........................................................................................................viii ABSTRACT......................................................................................................................xii CHAPTER 1 INTRODUCTION........................................................................................................1 2 SYNTHESIS OF MATERIALS WITH 1D STRUCTURE; ELECTROSPINNING..................................................................................................6 2.1 Introduction.............................................................................................................6 2.2 History of Electrospinning......................................................................................7 2.3 Experimental Setup...............................................................................................11 2.4 Electrospinning Theory........................................................................................14 3 SIZE EFFECTS IN BARIUM TITANATE...............................................................20 3.1 Barium Titanate....................................................................................................20 3.2 Size Effects in Barium Titanate............................................................................22 3.3 Previous Research: Theore tical Study on Size Effects.........................................25 3.3.1 Surface Energy...........................................................................................25 3.3.2 Surface Structure; Core Structure Model...................................................28 3.3.3 Stress Induced by Defective and Disordered Structure..............................29 3.4 Combined Effects of Stain Energy and Surface Free Energy in Constrained System.....................................................................................................................31 4 ELECTROSPINNING OF BARIUM TITANATE NANOFIBER............................34 4.1 Precursor Selection; Barium and Titanium Selection...........................................34 4.2 Precursor selection: Backbone Polymer...............................................................37 4.3 Experimental Procedure........................................................................................39 4.3.1 Precursor Mixture Synthesis.......................................................................39 4.3.2 Electrospinning of BaTiO3 Nanofibers......................................................40 4.4 Characterizations..................................................................................................41


vi 4.4.1 Thermogravitic / Differentia l Thermal Analyzer (TG/DTA).....................41 4.4.2 Scanning Probe Microscope (SPM)...........................................................42 4.4.3 Scanning Electron Microscope (SEM).......................................................42 4.4.4 X-Ray Diffraction (XRD)...........................................................................42 4.4.5 Transmission Electron Microscope (TEM)................................................42 4.4.6 Raman Spectroscopy..................................................................................43 4.4.7 X-Ray Photoelectron Spectroscopy (XPS).................................................43 4.4.8 Fourier Transform Infrar ed Spectroscopy (FTIR)......................................44 5 EFFECT OF HEAT TREATM ENT ON PHASE EVOLUTION..............................45 5.1 Thermoanalytical Investigations: Op timizing Heat Treatment Conditions..........45 5.2 Designing Heat Treatment Conditions: Two-step Annealing..............................46 5.3 Electrospun BaTiO3 Nanofibers...........................................................................48 5.4 Investigation on Perovskite BaTiO3 Phase Evolution..........................................52 5.5 Effect of Heat Treatment Time on BaTiO3 Phase Evolution on Electrospun Nanofibers...............................................................................................................54 5.6 Tetragonal Distortion and Grain Size of Electrospun BaTiO3 Nanofibers...........56 5.7 Morphology Change of Electrospun Nanofibers with Heat Treatment................65 5.8 Tetragonal Perovskite St ructure Evolution of BaTiO3.........................................66 5.9 Single Crystalline BaTiO3 Nanofibers via Electrospinning.................................69 5.10 Raman Spectroscopy of Electrospun BaTiO3 Nanofibers..................................72 5.11 Oxidation State Change of Cations with Heat Treatment; XPS Study...............79 5.12 Crystallization Reaction of Electrospun BaTiO3 Nanofibers.............................82 5.13 Effect of Electric Fi eld; Electrospinning............................................................85 6 TETRAGONAL STRUCTURE STABILIZING MECHANISM ON ELECTROSPUN BARIUM TITANATE NANOFIBERS........................................88 6.1 Stabilizing Mechanism Affecting the Cr itical Crystallite Size of Electrospun BaTiO3 Nanofibers..................................................................................................88 6.2 Hydroxyl Induced Strain Energy..........................................................................89 6.3 Depolarization Field Effect...................................................................................90 6.4 Surface and Strain Energy in Electrospun BaTiO3 Nanofibers............................91 6.5 Strain Energy Release: Strained BaTiO3 Crystallites...........................................98 6.6 Critical Crystallite Si ze of Electrospun BaTiO3 Nanofibers..............................101 7 CONCLUSIONS......................................................................................................103 7.1 Synthesis of BaTiO3 Nanofibers.........................................................................103 7.2 Future Work........................................................................................................106 7.2.1 Ferroelectric Characterization..................................................................106 7.2.2 Single Crystalline Electro spun Nanofibers Synthesis..............................107 7.2.3 Effect of Electric Field on Fiber Properties..............................................107 LIST OF REFERENCES.................................................................................................108


vii LIST OF TABLES Table page 1-1 Stable crystal structure of bulk bari um titanate with varying temperature................2 2-1 Binary ceramic nanofib ers via electrospinning..........................................................9 2-2 Complex oxide nanofibers via electrospinning and applications.............................10 3-1 Critical size for tetragonal to cubic transition of BaTiO3.........................................24 4-1 Barium and titanium precursors for sol-gel process and disadvantages..................35 5-1 Two step annealing conditions of electrospun BaTiO3 fibers..................................48 5-2 Average and standard deviation of fi ber diameters: as-synthesized and heat treated 750 oC (applied voltage: 10kV, working distance: 10 cm)..........................50 5-3 Comparison of 2 and R2 values on Lorentzian fitting, *DOF; degree of freedom.58 5-4 Comparison of 2 and R2 values on Lorentzian fitting, *DOF; degree of freedom.59 6-1 Obtained and calculated values for Williamson-Hall plot.......................................99


viii LIST OF FIGURES Figure page 2-1 The first patent on fiber spi nning process issued in 1934 [47]..................................8 2-2 Schematic diagram of electrospinning setup............................................................12 2-3 Schematic diagram of the forces acting on a droplet...............................................13 2-4 Schematic diagram of (a) electrospun nanofibers in axis symmetric instability, (b) non-axis symmetric inst ability; whipping mode................................................17 2-5 Scanning electron microscope image of electrospun BaTiO3 nanofibers deposited directly on Si wafer for 3 seconds. Deposited shape reveals whipping motion during fiber formation steps.........................................................................19 3-1 Cubic perovskite structure of BaTiO3......................................................................21 3-2 Schematic diagrams of Ti 4+ ion locations (a) above Curie point (T > Tc), (b) below Curie point (T < Tc). Ionic radii; Ba2+: 1.61 , O2-: 1.32 , and Ti4+ 0.68 [88].......................................................................................................................22 4-1 A flow chart of BaTiO3 nanofiber synthesis via electrospinning.............................41 5-1 TG / DTA profile of electrospun BaTiO3 nanofibers heated from room temperature to 750 oC (10 oC / min).........................................................................45 5-2 Schematic of two-step annealing process.................................................................47 5-3 SEM images of Ba-Ti-PVP composite na nofibers; (a) as-synthesized fibers, (b, c) dried at 120 oC for 1 h. (d) nanofiber diameter before polymer burn out (applied voltage: 10kV, working distance: 10 cm)..................................................49 5-4 Electrospun nanofibers (a) SEM images of BaTiO3 nanofibers after heat treated at 750 oC, 1h. (b) Magnified image shows multi-grain structure (applied voltage: 10kV, working distance: 10 cm)..............................................................................50 5-5 Electrospun nanofibers (a) SPM inte rmittent contact image of electrospun nanofibers deposited directly on go ld foil after heat treatment (750 oC, 1 h), (b) Three-dimensional image of electrospun nanofibers...............................................51


ix 5-6 XRD patterns of BaTiO3 nanofibers heat treated at various conditions; (a) 450 oC 3 h, (b) 550 oC 12 h, (c) 580 oC 12 h, (d) 580 oC 16 h. Highlighted on (c) and (d) indicates intermediate phases developed during the crystallization (applied voltage: 30kV, working distance: 15 cm)................................................................52 5-7 XRD patterns of BaTiO3 nanofibers heat treated at various conditions; (a) 600 oC 16 h (b) 650 oC 1 h, (c) 750 oC 1 h, (d) 750 oC 12 h, (e) 750 oC 16 h (applied voltage: 30kV, working distance: 15 cm)................................................................53 5-8 XRD patterns of BaTiO3 nanofibers heat treated for 24 h; (a) 580 oC, (b) 600 oC, (c) 650 oC, (d) 700 oC, Intermediate phases ~ 27o of 2 are highlighted on (a) and (b) (applied voltage: 30kV, working distance: 15 cm)......................................55 5-9 XRD patterns of BaTiO3 nanofibers heat treated for 48 h; (a) 580 oC, (b) 620 oC, (c) 650 oC, (d) 700 oC, Red box indicates removed intermediate phase with increased annealing time (applied volta ge: 30kV, working distance: 15 cm).........56 5-10 Lorentzian fitting of (a ) (002) and (200) peaks around 45o BaTiO3 nanofibers heat treated at 620 oC 48 h. (b) (002) and (200) BaTiO3 nanofibers heat treated at 700 oC 12 h (applied voltage: 20kV, working distance: 10 cm)..........................57 5-11 Lorentzian fitting of BaTiO3 nanofibers annealed at 750 oC, 16 h (a) (101) and (110), (b) (002) and (200 ), (c) (112) and (211)........................................................59 5-12 Change of tetragonality, c/a ratio, with annealing temperature and time.................60 5-13 An XRD pattern and tetragonality (a) Lo rentzian fitting of (002) and (200) peaks around 45o (650 oC 48 h annealing), Tetragonal distortion (c/a) of BaTiO3 nanofibers annealed at (b) 48 h, (c ) 24 h (Applied volta ge: 30kV, working distance: 15 cm).......................................................................................................61 5-14 Lattice parameters of electrospun BaTiO3 nanofibers; changes of unit cell parameters are plotted with temperature. (a: a-axis, c: c-axis, 3 at 2ct: cubic root of unit cell volume at subjected h eat treatment) (Applied voltage: 30kV, working distance: 15 cm).........................................................................................62 5-15 Crystallite size change with heat tr eatment temperature at various annealing time; (a)1-16 h heat treatment, (b) 24 h h eat treatment, (c) 48 h heat treatment (Applied voltage: 30kV, working distance: 15 cm).................................................64 5-16 SEM images of electrospun fi bers; (a) PVP burnout at 450 oC 3 h, (b) 550 oC 6 h, (c) 580 oC 12 h, (d) 600 oC 16 h, (e) 700 oC 12h, (f) 750 oC 1h (Applied voltage: 30kV, working distance: 15 cm)..............................................................................65 5-17 Polycrystalline fibers he at treated at different cond itions revealed grain growth with temperature; (a) 600 oC 16 h, (b) 1200 oC 1 h (Applied voltage: 10kV, working distance: 10 cm).........................................................................................66


x 5-18 TEM images of nanofibers (a) TEM image of Ba-Ti-O as-synthesized nanofibers. (b) High-resolu tion image of nanofiber. (c) Selected area diffraction pattern of amorphous nanofibers (Applie d voltage: 10kV, working distance: 10 cm)............................................................................................................................67 5-19 Bright field image of BaTiO3 nanofibers heat treated at 750 oC 1 h (Left) and magnified image which shows multi-g rains ranges between 25 – 50 nm. (Applied voltage: 30kV, working distance: 15 cm).................................................68 5-20 TEM images of nanofibers (a) A high resolution image obtained from the tip of barium titanate nanofibers. (b) A high resolution image that shows lattice fringes. (c) Selected area diffraction pattern obtained from the electrospun BaTiO3 nanofiber; zone axis [-134]. (Applied vo ltage: 30kV, working distance: 15 cm)....69 5-21 Single crystal nanofibers (a) SEM im age of single crystal nanofibers with polycrystalline nanofiber. (b ) Magnified image of singl e crystal nanofibers. (c) XRD pattern obtained from nanofiber mats containing single crystal fibers annealed at 750 oC 16 h. (d) Lorentzian fitting of (002) and (200) peaks. (Applied voltage: 30kV, working distance: 15 cm).................................................70 5-22 Single crystal nanofibers (a, b) TEM images of single-crystalline BaTiO3 nanofiber after heat treating at 450 oC for 3h followed by 750 oC 16h. (c) Selected area diffraction pattern with [3-10] zone axis. (d) A high resolution image with lattice fringes. (Applied vo ltage: 30kV, working distance: 15 cm)......71 5-23 Raman spectra of BaTiO3 nanofibers after heat treated at (a) 450 oC 3 h (polymer burnout), (b) 550 oC 12 h, (c) 580 oC 16 h, (d) 600 oC 6 h. Red circle indicates a characte ristic peak of CO3 2(applied voltage: 30kV, working distance: 15 cm).......................................................................................................74 5-24 Raman spectra of electrospun BaTiO3 nanofibers after heat treating at (a) 600 oC 16 h, (b) 750 oC 1 h, (c) 750 oC 16 h, Red circles on (a) and (b) indicate carbonate species. (Applied voltage: 30kV, working distance: 15 cm)...................76 5-25 Raman spectra of electrospun BaTiO3 nanofibers after 24 h heat treatment at various temperatures; (a) 580 oC, (b) 600 oC, (c) 680 oC, (d) 700 oC, a red circle at 850 cm-1 shows an intensity drop, which is an indicative of disordered and defective structure. (Applied volta ge: 30kV, working distance: 15 cm)..................77 5-26 Raman spectra of electrospun BaTiO3 nanofibers after 48 h heat treatment at various temperatures; (a) 580 oC, (b) 620 oC, (c) 650 oC. Intensity edge observed on 5-25 (a) was removed with temperatur e (5-25 (b)) and complete removal of carbonate was observed temperature 100 oC lower than 16 h annealing. (Applied voltage: 30kV, working distance: 15 cm)................................................................78 5-27 Ti 2p XPS spectrum of BaTiO3 nanofibers; (a) heat treated at 450 oC 3 h, (b) 550 oC 6 h, (c) 750 oC 16 h......................................................................................80


xi 5-28 Ba 3d XPS spectrum of BaTiO3 nanofibers; (a) heat treated at 450 oC 3 h, (b) 550 oC 6 h.................................................................................................................81 5-29 TEM images of BaTiO3 nanofibers after annealing at 600 oC, 1 h. A High resolution image of a nanofiber shows partially crystallized structure in amorphous matrix. (Obtained from hi ghlighted section of nanofiber)....................84 5-30 XRD patterns obtained from (a) BaTiO3 particles precipitated from Ba and Ti precursors mixed with PVP, (b) BaTiO3 particles precipitated from Ba and Ti precursors without PVP add ition. (Heat treated at 750 oC 16 h)..............................85 5-31 Raman spectra obtained from (a) BaTiO3 particles precipitated from Ba and Ti precursors mixed with PVP, (b) BaTiO3 particles precipitated from Ba and Ti precursors without PVP add ition. (Heat treated at 750 oC 16 h), A peak at 882 cm-1 (a red circle) indicates peroxide species (assignm ent is not definitive)...........86 6-1 FTIR analysis of the electrospun BaTiO3 nanofibers; (a) heat treated 600 oC 1 h, (b) 750 oC 16 h.........................................................................................................90 6-2 Schematic diagram of domain configuration with 90o and 180o domains...............91 6-3 TEM image of BaTiO3 nanofiber after h eat treatment at 750 oC for 1 h. Voids are highlighted by red circles...................................................................................97 6-4 A Williamson-Hall plot of electrospun BaTiO3 nanofibers heat treated at 750 oC, 16 h (Q= 2sin / ).................................................................................................100


xii Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy STRUCTURE EVOLUTION AND TETRAG ONAL PHASE STABILIZATION MECHANISM OF ELECTROSPUN BA RIUM TITANATE NANOFIBERS By Junhan Yuh December 2006 Chair: Wolfgang M. Sigmund Co-chair: Juan C. Nino Major Department: Materials Science and Engineering In this research, barium titanate (BaTiO3) nanofibers were synthesized for the first time via the electrospinni ng process utilizing a conve ntional sol-gel precursor. Electrospinning is a fiber synthesis method th at can generate nanofibers by applying a high electric field to a liqui d precursor mixture. Fiber diameters ranged between 180 to 320 nm before heat treatment with averag e diameter of 273 nm. After proper heat treatment fibers crystallized and the diameters reduced to a range of 80 to 190 nm. The average diameter was 116 nm. The length of th e heat treated nanofibers was in the order of a few hundred micrometers up to millimeters. X-ray analysis (Scherrer’s formula) revealed 30 nm ( 6 nm) of crystallite size and a tetragonal structure was confirmed via {200} and {110} peak separations. Tetragonal distortion (c/a ratio) of ~ 1.007 was obtained. Selected area electr on diffraction patterns confirme d a perovskite structure. Furthermore, the advancement in heat treatme nt of the electrospun fi bers yielded single


xiii crystalline BaTiO3 nanofibers that were 40 nm in diam eter and had lengths up to 0.6 m. This is the first report of single crystallin e electrospun nanofibers for a ternary oxide system. Raman spectra obtained after crystallization matched well with the spectrum of bulk tetragonal BaTiO3. An extra peak at around 640 cm-1 was generated from planar defects such as (111) twins th at can generate peaks located very similar to the hexagonal phase of barium titanate. An X-ray photoe lectron spectroscopy study found that barium and titanium ions existed with a different ox idation states. This indicates defective and disordered structures during crystallizati on and supports intermediate phase formation and decomposition followed by BaTiO3 crystallization. Moreover, thermodynamic analyses were made to investigate factors stabilizing tetragonal structure with small critical crystallite size (30 nm, 6 nm). Th e combined effect of strain and surface free energy was numerically calculated. Strain energy on crystallite composing BaTiO3 nanofibers was two orders of magnitude smaller (8.49 x 105 J/m2) than that observed from the cubic to tetragonal phase tr ansition of a metal oxide system (ZrO2). Fiber deformation during the growth and single crysta l structure of crystallites with negligible lattice misfits may contribute independently to lower strain energy. Williamson-Hall analysis confirmed strained BaTiO3 crystallites (by 0.255 %). Surface energy of gassolid interface expected to be between 0.00415 < tetra s g , < 0.01 (J/m2). It is concluded that low strain ener gy of electrospun BaTiO3 attributed to stabilize tetragonal structure. This is an indication that electrospinning is an effective method for synthesizing one dimensional nano scale tetragonal BaTiO3 with proper strain en ergy release mechanism.


1 CHAPTER 1 INTRODUCTION Over the last few decades there has been rapid progress in the degree of integration of electronic devices. Among them, scaling down is one of the most important issues in electronic device fabrication [1-3]. In dynamic random access memories (DRAMs) for instance, the number of transistors per squa re centimeter doubles every 12 months [4]. Although this trend of miniaturization has b een pursued intensely, industry and research laboratories still face limits of lithography that may prevent further progress of current techniques. In the industry, limitations of lithography and etching in defining device structure fabrication has rais ed costs exponentially of micr ofabrication below the 0.1 m scale [5]. Currently used DRAMs for example, genera lly require large capacitance capacitors [6]. Silicon dioxide has been widely used as a capacitor but it shows a low dielectric constant as the device scales down to below 100 nm in the fabrication process [7,8]. In order to keep dielectric properties of capaci tors while devices scale down, utilizing nano scale high dielectric constant nanostructural materials which are compatible for current device fabrication processes is inevitable. For the reason stated above, developing nanos tructure materials that maintain their own characteristics is a key i ssue to overcome limitations. Since the discovery of carbon nanotubes by Iijima [9], nanos cience and nanotechnology have r eceived a lot of attention over the last few years. Recently, one di mensional (1D) nanostructures, such as nanofibers, nanotubes, and nanorods have attr acted an increasing amount of interest due


2 to their fascinating and unique properties [10] . Various kinds of nanostructural materials [11] have been investigated with di fferent synthesis approaches [12-14]. Since the first observation of ferroelectric behavior from barium titanate (BaTiO3) in the 1940s [15-16], this material has been studied and used widely in both research laboratories and industry. Its applications include use in multilayer ceramic capacitors (MLCCs), and positive temperature coefficient resistors (PTCRs) [17-19]. Also, this material is competitive for replacing curre ntly used capacitors [20-25] for DRAMs because of its relatively high dielectric c onstant (k > 1000), high dielectric breakdown strength, and low leakage current [6,26-30]. BaTiO3 is categorized as a perovskite stru cture ceramic oxide that can transform into different structures with variations of temperature [ 17]. Crystal structures with different temperature range s are listed in Table 1. Table 1-1. Stable crystal stru cture of bulk barium titanat e with varying temperature -90oC

3 structure of barium titanate, reported limits differ from tens of nanometers to micrometers in diameter and differ from synthesis me thods and precursors used [34-38]. Even a structure, which has been known as unstable at room temperature i.e., the hexagonal, was suggested as a stable form of BaTiO3 with decreasing crystallite size [39]. Further study on the critical size effect of ferroelectric properties and physical properties would be required to enhance the performance of scaled down devices. Previous research conducted on nanoscale BaTiO3 mainly focused on particles [4041]. However, to adjust charac teristics of nanostructure BaTiO3 to next generation devices and/or currently applied processes, investigation of the materials should be performed with confined structures. Nanof ibers and nanotubes are good examples of confined geometry [42-43]. The goal of this dissertation is the synthesi s of one dimensional (1D) nano structure BaTiO3 and characterization of the tetragonal st ructure stabilization mechanism at room temperature. In this dissertation, barium titanate nanofibers were synthesized via electrospinning for the first time and struct ural evolution was inve stigated. Previous research on critical size and structure st abilizing mechanisms of nanoscale barium titanate has been reported [35, 38]. However, controversy ha s arisen on the critical size limit for phase transformation as well due to the complexity of phase stabilizing mechanisms. Therefore, this dissertati on consists of the following objectives: Synthesis of barium titanate nanofib ers via the electrospinning method; Describe the crystallization processes of electrospun BaTiO3 nanofibers with heat treatment conditions; tetragonal stru cture evolution with temperature


4 Thermodynamic modeling of the domina nt phase stabilizin g mechanisms for electrospun BaTiO3 nanofibers, and strain energy and surface free energy of crystallite composing electrospun nanofibers numerically calculated Among the various methods for synthesizi ng nanofibers, electrospinning is a remarkable straightforward way to fabricat e one dimensional nanostructures [44]. In other words, electrospinning is an “el ectrostatic fiber formation process.” Electrospinning was originally invented more than seventy years ago and has been used mainly for polymer fiber formation [45-47]. Recently, in combination with conventional sol-gel process, electrospinni ng has provided versatility for synthesizing nanofibers with various materials from polymers to ceramic na nofibers with different structures, tens of nanometers to microns in diameters and millimeter s in length. The versatility of material selection, controllable diameter, and conti nuous structure compared to other techniques have created interest in this process [47]. In chapter 2, the physics of electrospinning is introduced and modification of electrospinning for ceramic nanofibers is presented briefly. In Chapter 3, an overview of the material characteristics of BaTiO3 is presented. Previously suggested hypotheses and theories on room temperature cubic-tetragonal phase stabilizing mechanisms are introduce d. Combining previously suggested phase stabilizing mechanism, new hypotheses ar e made on effective phase stabilizing mechanisms affecting electrospun nanofibers. Details on precursor selection, precursor synthesis, and nanofibers synthesis, and characterization techniques are presented in chapter 4.


5 All details about the charact erization of nanofibers are explained in Chapter 5. The weight loss and reactions during crystallizati on of the electrospun fibe rs are investigated by thermogravitic / differential therma l analyzer (TG/DTA). Morphology and dimensional change with heat treatment is presented. Nanofibers formation and crystallization of the material is character ized via various micros cope techniques and Xray diffraction patterns (XRD), and step scan mode of X-ray diffraction patterns with Lorentzian profile fittings. Even if X-ra y diffraction patterns provi de information about barium titanate, this technique has drawbacks in investigating the nanosize structure. Instrumental broadening of peaks generally ha ppens when investiga ting nano crystallites [47], particularly with barium titanate that has a low c/a ratio (~ 1.01) and requires more precise techniques. Various approaches have been rendered and among them, Raman spectroscopy has been proven to be the best t ool for investigating th e subtle structural changes of perovskite polymorphs. The oxida tion states of Ba, and Ti ions, which can provide information during crystallizati on, were studied via X-ray photoelectron spectroscopy (XPS). Details of these characte rizations and experiments are discussed in Chapter 5. Nanoparticles were also obtained by precipitation of the same precursor for comparison, and the electric field effect on structural evolution was investigated. In Chapter 6, the phase stabilizi ng mechanism in electrospun nanofibers is demonstrated based on thermodynamic calcula tions. Strain energy and surface free energy were calculated numerically. Factors st abilizing crystallites that are exceptionally small are reasoned based on strain and surface free energy. The efficacy of early suggested theories is verified and the theoretical ly probable critical size of electrospun BaTiO3 crystallite is also presented.


6 CHAPTER 2 SYNTHESIS OF MATERIALS WITH 1D STRUCTURE; ELECTROSPINNING 2.1 Introduction Generally, it is known that nanostructure materials provide a favorable means to investigate the effects of size reduction on elec trical and thermal transport or mechanical properties [48]. Moreover, in electronic applications w ith miniaturization trends of device fabrication, existing top-dow n fabrications are appro aching their limit. As a consequence, both laboratories and industries are looking for alternat ive ways to realize ‘bottom-up’ processes for further miniat urization with enhanced performance characteristics [49]. Fortunately, recent technological achievements have provided various promising synthesis methods. Various approaches have been carried out to investigate the feasibility and compatibilit y of 1D nanostructure for next generation device applications [50]. Fabrication of th e prototype devices such as field-effect transistors (FETs), bipolar junction transistors (BJTs) , and p-n junctions are good examples [51-52]. In the area of information storage, ther e have been many active studies conducted to develop magnetic and optical storage com ponents with critical dimensions down to tens of nanometers as well [53-56]. Although reports have demonstrated possibilities on the laboratory level, there are still many draw backs that must be resolved to realize 1D nanostructure as functional units themselves. Despite the drawbacks, further research is ongoing.


7 Metal oxide nanostructures have been st udied since they can act as important components in both the scaling down of curre nt applications as well as for next generation devices [57]. Consequently, vari ous metal oxide nanofibers, nanowires, and nanotubes have been synthesized via various st rategies. Various methods that have been utilized to achieve 1D metal oxide structure can be categorized into two approaches: The first is downsizing the thin films by vari ous advanced nanolithography techniques, and the second is synthesizing 1D structural mate rial directly [58]. Although fabrication of nanostructure materials via nanolithogra phy dose not require much modification on currently used setups and expe riences, this approach is no t economically beneficial for mass production. In contrast, dire ct synthesis of 1D nano struct ures has its advantages in terms of versatility, high throughput, and economical efficiency. Consequently, techniques involving precursor decompositi on [59], chemical vapor deposition [60], solution based growth [61], template based electrodeposition [62], and laser ablations [63] have been widely employed. Since each method has its own merits and drawbacks, these synthesis techniques have been utilized in a complementary manner to achieve their objectives. Among the various approaches, electrospi nning has recently received a lot of attention since it appears to be the most stra ightforward and versatil e synthesis technique to achieve a 1D nanostructure, espe cially for metal oxide nanofibers. 2.2 History of Electrospinning Different from the conventional spinning process, electrospinning utilizes an electric field to form submicron fiber st ructures. Electrospinning produces non-woven fiber mats and has been shown to produce many 1D shapes including fibers, ribbons, and tubes [64]. All of these provi de micro/nanostructures with la rge specific su rface areas.


8 The diameter ranges from a few hundred micr ometers and down to tens of nanometers. Electrospinning is not new, but in fact has b een utilized to produce polymer fibers for 70 years. As presented in Figure. 2-1, Formhals patented his first invention for electrostatic spinning of fibers in 1934 [45]. Fig.2-1. The first patent on fiber spin ning process issued in 1934 [47] The invention did not claim electrospinni ng as a technique fo r the nanostructure formation process, and after more than 20 y ears after the first pate nt, the electrostatic fiber producing process itself did not gain much attention before a breakthrough made by Taylor [65]. In the 1960s, fundamental studies on the fiber forming process conducted by Taylor stirred attention for electrosta tic fiber synthesis as an alte rnative for scaling down [65]. Taylor investigated the shape of the initial liquid droplet formed at the end of the tube capillary under the influence of an applied field. Taylor’s study on the initial stage of


9 electrospinning proved the existence of the conical shape of a liquid droplet under the influence of an electric field for the first ti me. This was a very important discovery since these fibers are generated from the conical shape of the liquid precursor deformed under the influence of an electric field. Taylor also discovered that balance on the deformed liquid precursor, between surface tension and el ectrostatic force, results in a 49.3 degree angle. Various kinds of polymer fibers were synthesized based on Taylor’s research [6668]. In 1971, polymer microfibers were synthesi zed via electrospinning for the first time. Since then, the physics of the fiber formation process has been studied extensively and is presented briefly in Chapter 2.4 and 2.5. Table 2-1. Binary ceramic nanofibers via electrospinning Precursors Reference SiO2 Tetraethylorthosilicate [68] Al2O3 Aluminum di-sec-but oxide ethylacetate [69] TiO2 Titanium butoxide [70] CeO2 nitrate [71] Up until 2002, when electrospinning of ceramic nanofibers was reported for the first time [44], electrospinning of polymer nanofibers was of great interest. Recently, electrospinning of ceramic na nofibers has become one of th e rapidly growing research areas [72-73]. Table 2-1 gives examples of electrospun binary oxide nanofibers [68-71]. Electrospinning of ceramic nanofibers shows several advantages over other conventional processes introduced in Chapter 2.1. First of all, unlike other methods that produce relatively short nanorods and tubes, electrospinning provides long and continuous structures with diameters as low as tens of


10 nanometers. The continuous nature of elect rospun nanofibers offers the potential for alignment, direct writing, and spooling of nanofibers. Its versatility in materials selection is another advantage. Table 2-2. Complex oxide nanofibers vi a electrospinning and applications Ceramic Precursor Application Reference BaTiO3 Barium acetate Titanium isopropoxide MLCCs Capacitors for high densityDRAM [74] Pb(Zr0.52Ti0.48)O3 Lead acetate Zirconium acetate Titanium isopropoxide Ferroelectric random access memory (FRAM) [75] MgTiO3 Magnesium ethoxide Titanium isopropoxide Optical and microwave devices [76] Ba(Sr0.5Ti0.5)O3 Barium acetate Strontium acetate Titanium(diisopropoxide) bis(2,4-pentanedionate) Voltage controlled oscillator Delay line Phase shifter [77] NiTiO3 Nickel acetate Titanium isopropoxide High temperature application; tribological coating [78] NiFe2O4 Iron (II) ethylhexano isopropoxide Nickel ethylhexano isopropoxide Superparamagnetic materials [79] La2CuO4 Lanthanum nitrate Copper (II) nitrate Gas sensor [80] Theoretically, all ceramic materials which ha ve liquid precursors can be utilized for electrospinning. The ease of stoichio metry control owes its advantages to the conventional sol-gel process and is another merit of electrospinning [44]. Due to its versatility, the application of electrospinning expanded to more complex oxide nanofiber synthesis. Ternary oxide fibers with different compositions are good examples. Inexpensive and relatively simp le setup makes electrospinning an attractive process as


11 well. Ternary oxide nanofibers for ferro electric, piezoelectr ic, and gas sensing applications are summarized in Table 2-2. The most crucial factors in nanostructu re synthesis are the control over the dimension and compositional homogeneity. Judging from this point of view, electrospinning may be one of the most suit able solutions to resolve currently faced issues on nanostructure. With the expansion of electrospinning from polymers to ceramics, the applications of electrospun nanofibers ar e being vastly expanded. With the synthesis of electroceramic nanofibers, the building bloc ks for various applications are becoming more readily available. 2.3 Experimental Setup The electrospinning setup is amazingly simple and generally consists of three major parts: A high voltage power supply; a liquid precursor supplying system; and an electrically conductive fiber collector. A conductive plate and spool are most commonly used. In some cases, the electrospinning set up requires isolation in closed box for humidity, air flow, and temperature control [ 44, 74]. For the electrostatic fiber formation process, the electric field of several kilovolts is applied between a reservoir containing a complex fluid ceramic precursor and a grounde d collector. A schematic diagram for the experimental set up and a typical labora tory setup is shown in Figure. 2-2. The liquid precursor mixture is loaded into a plastic syringe that is equipped with a metallic needle. The feeding rate of the precursor mixture is generally controlled by a programmable syringe pump. The metallic n eedle tip is connected to a high voltage power source, either positive or negative.


12 Figure.2-2. Schematic diagram of electrospinning setup A liquid precursor droplet at the end of the needle tip becomes deformed as the applied voltage is increased. Forces acting on a liquid precursor dropl et are described in Figure. 2-3. By applying the high voltage , the precursor mixture droplet becomes electrified and induced charges are distributed over the surface droplet. As a result, two opposite types of electrostatic for ces have an effect on the drople t. One is an electrostatic repulsion of the surface charges while the Coul ombic force is exerted by an externally applied electric field [65-66]. In addition to the surface tension and gravity, additional distinct forces affect the shap e of the precursor mixture drople t, which is explained in the


13 following section. Under the in fluence of these electrostatic interactions, the precursor droplet is distorted to a conical shape, comm only referred as the Taylor cone [65, 81-82]. Fig.2-3. Schematic diagram of th e forces acting on a droplet Depending on the experimental condition, at some threshold value of electric field when the electrostatic force overcomes th e surface tension and viscosity, fiber jet emission is finally initiated. As fiber jets travel down to the collector, the emitted jet experiences whipping and bendi ng instability modes [81-82]. Most of jet thinning and elongation processes happen when the jet is passing through these instabilities. Accumulation of charges is responsible for initiating instabilities. The same charge repulsion on the initial fiber jet le ads fibers to split into smaller and thinner fibers so as to form nanofibers. The solvent in the precu rsor mixture also evaporates during the


14 electrospinning process. Guid ed by an electrostatic attractive force, charged fibers are deposited on the grounded colle ctor (aluminum plates, foil s, and spools are commonly used). Without any fiber aligning setup, fibers are randomly deposited and form nonwoven fiber mats. Numerous reports have demonstrated that after experiencing inst abilities, diameters can be reduced from hundreds of micrometers to as small as tens of nanometers. Final fiber diameters can also be affected by rheologi cal properties of the precursor mixture. It has been proved that terminal fiber diamet ers can also be affected by several process characteristics. The distance between the or ifice end and grounded collector, i.e. the travel distance, applied el ectric field, temperature, ai r flow, and humidity are good examples [44]. Details and the relations hip regarding critical factors will be demonstrated based on electro hydro dynami c theory in the subsequent section. 2.4 Electrospinning Theory Electrospinning looks like a very simple a nd straightforward process but there has been controversy on the physics that govern th e fiber forming procedure. For instance, spherical beads can be formed and this is one of the major issues of electrospinning. Undesired beads formation can lower the homoge neity of nanofiber mats and can affect properties of final products. Viscosity, char ge density on jet, and surface tension can independently influence the forming of bead s. However, insufficient understanding of physics prevented the rationalizing of the fi ber formation process in electrospinning. There are difficulties in building up establishe d theory because of the complicated and combined effects of individual factors. It took more than a couple of decades to reach consensus on the complicated physics of fiber formation processes.


15 Research groups around the world have b een working on developing a theory of electrospinning and electro hydrodynamic theory (EHD) is generally accepted as an explanation of the fiber forming phenomena by in stability modes. In this chapter, brief review of EHD is introduced [44, 66-67, 83-84], which was actually developed for electrospinning of polymers. EHD theory modified for electrospinning is used for ceramic systems. The main focus of the theoretical work presented here is on the prediction of final fiber diameters since th ey will strongly reduce the number of experiments needed to find the minimal possible diameters for specific ceramic systems and can assist in designing electrospinning solutions. Electrospun fiber synthesis can be divided in to three distinct sections (see Figure. 2-3, 2-4), including the Taylor cone formation, the thinning/ elongation of the fiber, and the drying of the fiber. Most of theoretical work normally concentrated on the first two stages where dimensional characteristics of fibers are set. The theoretical work published by Rutledge et al. [66-67, 82-83] on the first stage focused on the steady stretching of the precurs or droplet. Here th ey proposed a slender body theory that couples jet st retching, charge transport, and the electric field as described above. As presented in the experi mental setup (Chapter 2.3), a complex fluid jet is ejected from the end of a syringe and is accelerate d and elongated by electrostatic forces. Overall there is a force balance in th e cone where gravity a nd electric polarization stress tend to elongate the dr oplet from within, aided by ta ngential electri c stress on the surface. Viscous drag slows down the fiber jet ini tiation. In addition, there are two more forces active on the surface of the deformed precursor cone. These are the surface tension


16 that tries to minimize the surface area pulli ng the surface vertically in and an opposing force which is the normal electrical stress that tries to maximize the distance between electrical charges on the surface by enlarging the surface area (Figure. 2-3). When the electrostatic repulsive forces overcome the surf ace tension, a fine char ged jet is ejected. Generally, the diameter of the initial jet is typically 1/100 in diameter of the needle opening. After leaving the cone this jet then moves toward the counter electrode. While traveling, the jet will be come unstable and experi ence instabilities. In the second stage, there ar e a total of three possible inst abilities [84]. The first is referred to as Rayleigh instab ility where droplets are form ed. Rayleigh instability is predominant when the surface tension is large enough. The second possible instability occurs as the initial jet travels further down. In this mode, varicose structures are visible. This mode is referred to as the axis-symme tric instability mode. Finally, the last possibility involves the fiber synthesis mode coming up. This is referred to as a nonaxisymmetric mode since movement of the fi bers is no longer symmetrical. This last bending instability is the most important among the three modes that needs to be achieved, since it is responsible for the thinning of the fibers from about a micron into the nanometer range. The last instability is more commonly referred as the whipping mode. The entire second stage with all three instabilities wa s also modeled by Rutledge and colleagues [84]. Here, the polymer solu tion jet is modeled using a linear Maxwell equation. This electro hydrodynamic (EHD) inst ability theory predic ts that the non-axis symmetric instability can prevail over the Ra yleigh instability if the system is highly charged [66-679, 82-83]. EHD describes the e ffects in the unstable region based on surface tension (Rayleigh type) and electri cal forces (purely electrical competition


17 between free charges and the electric fields ). In the Rayleigh instability mode, insufficient electrical force pr events jet initiati on. Spraying of the liquid precursor is observed from this condition. Fig.2-4. Schematic diagram of (a) electrospun na nofibers in axis symmetric instability, (b) non-axis symmetric inst ability; whipping mode At high fields/charge densities, the Raylei gh instability can be compensated, and all observed instabilities are domina ted by electrical forces. Ther efore, surface tension does not affect the stability of a strongly forced electrical jet. The electrical or conducting type can be categorized into an axis symmetric conducting mode and a whipping conducting mode . While the Rayleigh instability is suppressed with an increasing applied elec tric field and surface charge density, the conducting modes are enhanced. The dominan t instability modes among the two is strongly dependent on the viscos ity/conductivity ratio of the jet and also the static charge density of the jet.


18 As with increasing static charge dens ity, the whipping mode tends to dominate because the same charge repulsive inte raction induced by a high surface charge simultaneously suppresses the axis symmet ric modes and enhances the non-axis symmetric whipping mode. EHD allows for the calculation of the thi nning rate of the jet in the bending instability regi on [84]. The simplified relationship between terminal fiber diameter and parameters are expressed as follows: I Q ht (2-1) Q represents liquid jet flow rate, I as the total current directed toward the lower electrode as the surface tension of liquid precursor mixture. It can be seen from the equation 2-1 that the terminal diameter of the whipping jet is therefore controlled by the fl ow rate, the electric current, and the surface tension of the complex fluid. Experimental and predicte d data by EHD theory matched for polymer fibers within a 20% error ra nge. Evidence of the whipping instability mode during the electrospinning was observed by direct deposition of BaTiO3 nanofibers on a silicon wafer (Figure. 2-5). Several assumptions we re required for the current EHD theory. The thinning rate of the jet in th e bending instability region is calculated assuming Newtonian flow and negligible elastic effects due to the drying of the jet. Furthermore, the electric conductivity through the bulk of the fiber is neglected since polymer systems are dominated by electrical effects on the jet surface.


19 Fig. 2-5. Scanning electron microsco pe image of electrospun BaTiO3 nanofibers deposited directly on Si wafer fo r 3 seconds. Deposited shape reveals whipping motion during fiber formation steps


20 CHAPTER 3 SIZE EFFECTS IN BARIUM TITANATE 3.1 Barium Titanate The structure of barium titanate is categori zed as perovskite since it is an isomorph of CaTiO3, the mineral perovskite [31]. The per ovskite structure is generally expressed as ABO3, where A denotes cations coordinated with 12 oxygen an ions while B represents different cations that are located in the octa hedral sites [85]. A schematic of a cubic perovskite structure is presented in Figure 3-1. Barium titanate can transform into five different structures according to temperature variations. Below -90 oC, its structure is rhombohe dral with a face diagonal polar axis. In the temperat ure range between -90 and 0 oC, it turns into an orthorhombic with a body diagonal polar axis [31]. From 0 oC to at about 132 oC, BaTiO3 turns into a structure of interest; a tetragonal perovskite. Cubic structure BaTiO3 is stable in the temperature range between 132 oC and 1460 oC. Above 1460 oC to the melting point (1625 oC), its hexagonal phase is stable [17]. The temperature at around 130 oC is designated as Curie point (Tc) above which the spontaneous polarization ceases [31]. From 0 to around 132 oC, spontaneous displacement of Ti4+ ions from its center position to face centered oxygen ions distort the crystal structure and yield the tetragonal form. The energy of the crystal can be lo wered by forming distortion through the displacement of Ti4+ ions in the temperature range between 0 oC and 132 oC [86].


21 Fig 3-1. Cubic perovskite structure of BaTiO3 Tetragonal distortion (c/a ratio) between cubic to tetragonal structures is very small (c/a = 1.01, 0.012 nm elongation). Among the fi ve different structures introduced, perovskite BaTiO3 with a tetragonal structure has genera ted a great deal of interest since the highest and most stable spontaneous polarization, which can be utilized for various applications, is observable [87]. Displacements of Ti4+ ions with temperatures below and above Tc are presented in Figur e. 3-2 (a) and (b). Above the Curie temperature, the structure of barium titanate is cubic, in which the Ti4+ ion is located at the center of the mass of the negative char ge (Figure 3-2(a)). On the other hand, between the Curie temperature and 0 oC (0 ~ 130 oC) the center Ti4+ ion displaces from its center symmetrical pos ition and the crystal is polarized by the separation of the center of mass of the negative a nd positive charges [87].


22 Fig 3-2. Schematic diagrams of Ti 4+ ion locations (a) above Curie point (T > Tc), (b) below Curie point (T < Tc). Ionic radii; Ba2+: 1.61 , O2-: 1.32 , and Ti4+ 0.68 [88]. As depicted in Figure 3-2( b), displacement of the Ti4+ ion elongates the cubic structure, which becomes tetragonal. Long-rang e interaction between ions in unit cells is involved to develop the permanent dipole mome nt throughout the lat tice. The energy of the crystal is lower when the Ti4+ ion in each unit cell is ali gned in c-direction compared to a-direction aligned state. 3.2 Size Effects in Barium Titanate Recently, with the ongoing drive towards de vice miniaturization, it has been crucial to synthesize controllable nanoscale ferro electric materials an d to understand their ferroelectric behavior [61, 89-94] . To adjust nanoscale BaTiO3 into various device applications, especially for electronic applic ation, thorough understand ing of the critical size limit that can retain tetragonal structure is required. It has been known that the room temperature stable structure of BaTiO3 varies with crystallite size. The room temperature


23 stable phase of BaTiO3, tetragonal, turns cubic as cr ystallite size decreases to submicrometers [38, 95-97]. The first observation of the size dependent structure change in ferroelectric materials was reported from ferroelectric salt, potassium dihydrogen phosphate (KH2PO4) [96]. For BaTiO3, Knzig published a theoretical background on the dependence of dielectric properties on part icle size in 1950 [99]. Knzig’s research predicted that cubic rath er than tetragonal BaTiO3 is expected below 10 nm of the critical size of fine BaTiO3 grains. In 1954, Anliker et al . [100] observed size dependent structure change of BaTiO3 powders for the first time and reported the decreased tetragonality of the particles smaller than 5 m. Since then, numerous reports have been conducted to reveal the critical size limitati on for the cubic to tetragonal phase transition of BaTiO3 synthesized by various methods. As with the progress of synthesis methods and characterization t echniques, recently reported critical sizes are down to the nanom eter scale. Recent reports on the phase transforming size of BaTiO3 obtained via various methods and measuring characterization tools is summarized in table 31. It is noted here th at critical sizes of tetragonal to cubic phase transformation di ffer from synthesis methods and starting materials. Size dependent phase change of BaTiO3 has been considered and as a consequence, several approaches have come up with plausibl e explanations. It is argued that various synthesis methods with differe nt starting materials can intr oduce defects which generate internal strain energy [35, 101-102], and it has been generally accepted that induced strain energy is responsibl e for stabilizing cubic BaTiO3 at room temperature. Others have suggested stress can be generated by clustering between nanoparticles during


24 crystallization. Clamped particles are subjected to isostatic pressure that results in cubic BaTiO3 by preventing tetragonal distortion at ro om temperature [103]. Large specific surface energy on nano structure materials has been suggested as a system that stabilizes high temperature phase at room temperature by introducing hydrostatic pressure to the materials and preventing tetragona l phase transformation [38, 42, 102]. Table 3-1. Critical size for tetra gonal to cubic transition of BaTiO3 Synthesis method Critical size of tetragonal to cubic transition (nm) Size characterization method(s) Characterization of the size effect on electrical properties Ref. 30 XRD N/A [35] 80 BET Dielectric constant [36] 100 TEM N/A [104] Sol-gel 50 TEM Dielectric constant [105] 62~68 XRD N/A [34] Barium titanyl oxalate tetrahydrate decomposition 30 BET N/A [106] 90 120 XRD TEM N/A [38] Hydrothermal method 90 XRD, TEM N/A [40] 150 N/A Dielectric constant [37] Ground powder 200 TEM Dielectric constant [41] It seems that defining the absolute critical size of tetragonal BaTiO3 is unlikely to be achieved since crystallization processes an d/or reactions may vary when a different synthesis method is applied. However, establ ishing a theoretical background for a sol-gel based electrospun nanofiber system would be be neficial when it is adjusted for device fabrication.


25 3.3 Previous Research: Theoretical Study on Size Effects Theoretical approaches have been re ndered to investigat e factors affecting tetragonal to cubic phase stabi lization at room temperature. Stabilization of the high temperature phase is attributed to surface e ffects, defective/disordered structure, and internal stress during crystallization. As introduced in this chapter, theoretical reviews on size limiting mechanisms are presented followed by a combined thermodynamic criterion for deciding critical size for phase tr ansformation at room temperature. 3.3.1 Surface Energy It has been proposed that physical prope rties in submicron size particles are generally affected by macroscopic effects relate d to the surface tensi on of fine particles. Surface tension, measured in Newtons per meter (N/m) is defined as the force along a line of unit length perpendicula r to the surface, or work done per unit area. Dimensional analysis revealed that the unit of surface tensi on (N/m) is equivalent to joules per square meter (J/m2). This means surface tension can be considered surface energy. Due to the relatively large surface area of nanos cale materials, nanoscale BaTiO3 is expected to have large specific surface energy. Uchino et al. [38] suggested surface energy affected by internal pressure originates from surface te nsion. Internal pressures which may act on nanoparticles were found to be proportiona l to the surface energy and are calculated according to the following equation [38]. R P 2 (3-1) P denotes internal pressu re required to lower Tc, Curie temperature, down to room temperature and R and are radius of the particle and surface tension (N/m), respectively.


26 Using experimentally obtained data for critical particle size (0.1m for BaTiO3) and published hydrostatic pre ssure (2 GPa for BaTiO3), = 50 N/m for BaTiO3 [26] was calculated. Compared to = 1 N/m measured from MgO, this suggests that the surface tension effect on the perovskite structure pha se transformation must be significant. It should be noted here that research perform ed by Uchino et al. did not consider surface energy related to gas-solid interface that should be included. Saegusa et al. [34] explained the effect of surface energy on defining critical size limit in terms of free energy difference between tetragonal and cubi c phases. Free energy difference of the tetragonal and cubic BaTiO3 can be stated as follows [107-109]: o T cubic o T tetra o T cubic o T tetra o T cubic o T tetra o TS S T H H G G G, , , , , , Tc T tetra P cubic P Tc T o Tc trans Tc T tetra p cubic P o Tc transT dT C C S T dT C dT C H ) (, , , , , , (3-2) (T; temperature, Go T: free energy difference at given temperature at 1 atm, So: entropy change at 1 atm, Ho: heat of transformation at 1 atm, Cp: heat capacity at constant pressure) The free energy difference between the tetra gonal and cubic at room temperature is Go 298 = 48 J/mol [34]. The excess free energy, Go, can be expressed as follows: A Go surface 2d n (3-3) ( : excess surface energy per unit area related to the volume work, A: particle surface area, n: number of particle s, d: particle diameter) Excess surface energy is now calculated as Go T = Go surface. This is based on nucleation and growth theory th at cubic to tetragonal transf ormation is not involved with nucleation from liquid state. Th is indicates that volume free energy contribution from nucleation can be excluded. Based on their e xperimental data, Saegusa et al. assumed dc,


27 to be a critical particle diameter for te tragonal to cubic transformation, 0.1m. The calculation of BaTiO3 was 0.02 J/m2 while that of PbTiO3 was found to be 0.35 J/m2. The amount of work required for tetragonal to cubic transition at room temperature was calculated from published data. The volume changes at the tetragonal to cubic transformation, Vt c, for BaTiO3 and PbTiO3 are 3.7x10-8 m3/mol and 16.2x10-8 m3/mol respectively. The work required on phas e transformation is 130 J/mol for BaTiO3 and 1000 J/mol for PbTiO3 [34, 110]. It is worthy to note that PbTiO3 of which the critical size is an order of magnitude smaller than BaTiO3, requires more work for phase transformation. Larger , which results in larger Go T, was observed from PbTiO3 as well. This suggests that the stabilization of the tetragona l structure at a temperature concerned is strongly affected by excess surface energy per unit area ( ), and the difference of the free energy ( Go T) between tetragonal and cubi c phases. The larger the , or the larger the free en ergy differences between the tetragonal and cubic at the temperature concerned, the smaller the critical crystallize would be. It should be noted that the stabilizatio n mechanism described above by Saegusa was based on the assumption that particles ar e an unconstrained system of free particles. Consideration on both stress and strain effect s should be included in order to expand surface energy effects to polycrystalline and/ or multi layer thin films. Also, surface energy contribution from both gas-solid inte rfaces together with solid-solid interface (surface energy related to the volume work) should be considered. However, Saegusa’s approach still provides valuable tips on obt aining surface energy related to the volume work.


28 3.3.2 Surface Structure; Core Structure Model The core shell structure of BaTiO3 was proposed by Niepce and McKinstry after observing BaTiO3 single crystal covered with a surf ace layer of low index of refraction with the thickness less than 10 nm [111]. Independently, Sh aikh et al. also suggested on the basis of electrical measurement that 300 nm particles be covered with low dielectric constant surface layer of 9 nm thick [112]. Based on the observations above, the coexistence of the tetragonal and cubic phases ha s been studied. Furt hermore, Takeuchi et al. [111] observed the coexiste nce of the cubic and tetragonal phase from XRD pattern analysis and proposed two possible models fo r the coexistence of the phases in BaTiO3 [35,113]: (1) One particle is composed of both tetragonal and cubic phases; (2) The particles of the cubic phase coex ist with those of the tetragonal phase. Experiments were performed to validate ea ch model. As with the decrease of crystallite size, the amount of te tragonal phase ([t]/[c]) decrea sed as well. This indicates that Model (2), which depends on the number of cubic and tetragonal particles, is irrelevant. Based on Model (1), the relativ e amount of the cubi c phase should increase with decreasing particle size. Model (1) matches well with experimental observation. The existence of a cubic surface layer of constant thickness was proposed as follows: 3 3 32 3 4 2 3 4 2 3 4 ] [ ] [ l d d l d c t (3-4) = 3 2 2 3 2 2 38 12 6 8 12 6 l l ld l d l ld d (3-4) (l: surface layer thickness, d: particle diameter)


29 The above equation can be reduced to l d c t 6 ] [ ] [, since d is much larger than l. A cubic layer thickness of l = 5 nm was calculated and the suggested thickness of the cubic layer is independent of the starting material s and synthesis condition, and therefore grain size. That is, in nanoparticles, surface layer with cubic phase becomes increasingly larger. However, the core-shell model did no t clarify the critical size of BaTiO3 nanoparticles of tens of nanometers in scal e. Since recent investigation of BaTiO3 particles and thin films less th an 100 nm in diameter revealed fully developed tetragonal structures [34-38, 40, 104-105], the core-she ll model must be modified or may be considered inappropriate for extremely fine particles. 3.3.3 Stress Induced by Defectiv e and Disordered Structure It has long been suggested that la ttice imperfections introduced during crystallization are responsible for preven ting tetragonal phase st abilization at room temperature [35, 40]. For solution based synt hesis, defective and disordered structures can be developed during the precursors and intermediate phase decomposition. Hydroxyls (OH-) located in oxygen sites and rand om cations and anion vacancies substitution are most probable cases [114]. BaTiO3 synthesized by the wet-chemical method generally contains hydroxyl ions. Chemically adsorbed hydroxyl was obser ved via infrared absorption spectra from the BaTiO3 particles that were ev en heat treated at 800 oC. The hydroxyl ions that reside in oxygen ion sites are followed by the creation of cation vacanc ies to maintain electrical neutrality. Lattice volume is en larged due to the adsorbed OHions in O2sites. Shrinkage of tetragonal unit ce lls was found to correlate wi th the releas e of water, determined to exist as chemically bound OHions [35, 40, 114-115]. It was observed


30 from previous research that adsorbed hydroxyl ions can be removed at around 900 oC. However heat treatment over 1000 oC is required for conversion to the tetragonal structure. The expected hydroxyl release process is as follows [115]: O xH BaTiO x OH O V Ti V Bax x x Ti x x ba x 2 3 3 6 / ' ' ' ' 6 / 1 6 / ' ' 6 / 13 ) 6 ( ) ( ) ( ) ( 6 (3-5) This was explained by considering the diffusion of the coefficient of the constituents. Oxygen vacancy s hows significant mobility around 650 oC. On the other hand, cation vacancy requires above 1050 oC heat treatment to achieve minimum mobility [115]. The difference in mobility re sults in points defects and as a consequence, point defects ultimately introduce lattice stra ins which hinder tetragonal transformation. The hydroxyl originated strain itself does not limit the cr itical size since defects can be removed by proper heat treatment. Howe ver, hydroxyl elimination with an ordered structure requires high temperature heat treatment that leads to grain growth. It is worth noting here that recently, tetragonal BaTiO3 synthesized via the wet chemical method was reported with highe r amounts of hydroxyls than commercially available BaTiO3. It was concluded that the hydroxyl s effect on cubic pha se stabilization is minimal [102]. Since an electrospinning BaTiO3 nanofiber is based on the sol-gel process, hydroxyls can be generated during organic pr ecursor decomposition and different types of reactions. In this study, the effect of hydroxyls is investigated experimentally via Fourier transform infrar ed spectroscopy (FTIR).


31 3.4 Combined Effects of Stain Energy and Surface Free Energy in Constrained System The effects of strain energy and surface energy on phase stability were introduced in an earlier section. However, proposed theories assumed free particles, not under constrained state such as sandwiched thin film s and agglomerate particles. On the other hand, crystallization and growth of particles in nanofibers and/or thin films are subjected to complex effects of strain energy and surface energy. To evaluate the phase stability of BaTiO3 in nanofibers, the combined effect of two critical thermodyna mic factors should be considered. It can be expected that complex stress may exert on BaTiO3 nanofibers when nanofibers are crystallized from the matrix of different crystal structures. The theory concerning the stress on particles in an isotropic matrix has been established. It was assumed that a high te mperature stable phase is crystallized in spherical shapes from a low temperature stable phase matrix. Since the thermal expansion coefficient of partic les and matrices are different, the stress will be introduced on spherical particles with st ructural changes. Weyl a nd Selsing have established a pressure subjected to a particle in the infinite matrix [116] as follows: 2 2 1 1/ ) 2 1 ( 2 / ) 1 ( E E T P (3-6) Difference in the expansion coefficient of two polymorphs is , T is the temperature difference between two polymorphs . The Poisson’s ratio of matrix and particle are 1 and 2 respectively. Young’s modulus of ma trix and particle are expressed as E1 and E2. Total strain energy, U, per unit vol ume is expressed as follows [117]: 2 3 2 1 2 3 2 2 2 11 2 1 P P P P P P E U (3-7)


32 P1, P2, and P3 are orthogonal tensile a nd compressive stresses. The strain energy in the matrix, U1, using –P1 = 2P2 = 2P3 = PR3/r3, is formatted as [117], dr r r R P E UR 2 6 6 2 1 1 14 2 3 2 1 (3-8) 3 4 2 ) 1 ( 33 1 1 2R E P The strain energy within the high temperature phase, U2, is expressed as [117], 3 4 2 2 1 33 2 2 2 2R E P U (3-9) Total strain energy, UT is obtained by considering both U1 and U2 [117]. 2 2 1 1 3 22 1 2 1 E E R P UT (3-10) By utilizing strain energy imposed to the particles in matrix, Gravie set up a thermodynamic criterion for the transforma tion for particles experiencing phase transformation [118]. c t c c t c c t V cU r r G r G 3 2 ) ( 33 4 4 3 4 (3-11) ( G: difference in free energy of polymorph, Gv(t-c): volume free energy change of a crystal with pha se transformation, rc: critical particle size, (t-c): surface energy per unit area, U(t-c): strain energy induced by phase tran sformation, subscripts t, c denote tetragonal and cubic phase respectively) Thermodynamic criterion for the phase tr ansformation can be achieved when free energy difference in free energy of polymorph is balanced by the change in surface free energy and strain energy [118].


33 0 3 4 4 3 43 2 ) ( 3 c t c c t c c t V cU r r G r G (3-12) From the equation above, it can be stated that the critical size, dc, can be calculated from equation (3-13) [118]. c t c t v c t cU G d ) (3 (3-13) It can be expected from equation (3-13) that the smaller the strain energy ( Ut-c), the smaller the critical crystallite size (dc). Small surface tension ( t-c) per unit area is preferable for stabilizing tetr agonal phase with smaller critic al crystallite size as well. Equations stated here do not consider piezoelectric effect that can be expected from perovskite tetragonal BaTiO3. However, degree of strained state will be estimated by Williamson-Hall analysis and it is presented on Chapter 6.5.


34 CHAPTER 4 ELECTROSPINNING OF BARIUM TITANATE NANOFIBER 4.1 Precursor Selection; Barium and Titanium Selection Due to its simplicity and cost effectiveness, BaTiO3 has been synthesized mostly by the solid state reaction betw een barium carbonate (BaCO3) and titanium dioxide (TiO2) [119-121]. This process is an economically beneficial for mass production. To obtain particles on a micrometer scale by solid state reaction, repeated ball-milling is generally followed. A high temperature h eat treatment process, over 1000 oC and up to 1200 oC, is required as well to achieve stoichiometric BaTiO3 [122]. However, BaTiO3 by a solid state reaction results in particles larger than 2 m and this prevents the solid state reaction process to be utilized to synthesize nanoscale BaTiO3 with high purity. The ball-milling process which is essential for scaling down the particle size can introduce impurities as well [123]. Recent progress in solution based synthesi s, such as the sol-gel method [32, 124128], oxalate method [129], and hydrothermal method [130] have achieved considerable attention. Among them, the sol-gel process shows several advantages over solid state reactions and other solution ba sed synthesis. First of a ll, quasi-atomic dispersion of materials in a liquid precursor can lead to the synthesis of nano crystals with a homogeneous size distribution. In addition, crystallization occurs at relatively lower temperatures (lower than at least 500 oC). A low temperature process can make the solgel process more compatible with device fabr ication on the industrial level and help to prevent grain growth during high temperature heat treatment. An advantage from sol-gel


35 methods is by simple modification of rheolo gical properties in wh ich the conventional sol-gel process can be adjusted to electrospinning without de grading precursor materials. By combining advantages of the sol-gel pr ocess with electrospi nning, various kinds of electroceramics have been synthesized. Various kinds of barium and titanium so l-gel precursors have been reported and used in both industry and rese arch laboratory levels under di fferent synthesis conditions. Among these, barium ethoxide, barium acetat e, barium hydroxide, barium methoxide, and barium metal are commonly used starting materials for barium precursor synthesis while titanium isopropoxide, tetra-n-butyl t itanate, and titanium methoxyethoxide are widely used as starting materials of the tita nium precursor. A comb ination of Ba and Ti precursors is listed in Table 4-1 [32, 123-126]. Table 4-1. Barium and titanium precursors for sol-gel process and disadvantages Ba and Ti precursors Disadvantages Reference Barium ethoxide Titanium isopropoxide Requires nitrogen (N2) atmosphere [124] Barium acetate Tetra-n-butyl-titanate Multiple steps for synthesizing final precursors [125] Barium hydroxide Titanium isopropoxide High temperature and refluxing [32] Barium metal Titanium isopropoxide Nitrogen blowing required for hydrolysis [32] Barium oxyhydrate Tetra-n-butyl-titanate Temperature control: low temperature (2 oC) process [123] Barium methoxyethoxide Titanium methoxyethoxide Temperature control: high temperature (120 oC) process [126]


36 Precursors insensitive to experimental conditions are favorable since they can enhance use of the process. As presented in Table 4-1, each precursor pair has its own merits, however drawbacks are revealed when considering feasibility and compatibility. Using barium ethoxide and titanium isopr opoxide as starting ma terials for BaTiO3 synthesis, a nitrogen atmosphere in the dry box is always required [124]. Tetra-n-butyltitanate is compatible to room temperature and ambient atmosphere conditions but needs additional processes to obtain the Ti precurs or [125]. Barium me tal, barium hydroxide, and barium oxyhydrate to obtain a barium pr ecursor, require a nitrogen atmosphere, temperature control, and refluxing [32, 123, 126]. Barium isopropoxide and titanium butoxide are another combina tion, however, barium isopropoxide generally reacts violently with moisture in the air and hydrolys is occurs very rapidl y when using titanium butoxide. As a consequence, stoichiometric barium titanate is hard to achieve. Additional setups required to maintain precu rsors in controlled conditions make sol-gel based electrospinning costly and time consuming. On the other hand, controlled synthesis c onditions are not required when barium acetate [Ba(CH3COO)2] and titanium isopropoxide [Ti((CH3)2CHO)4] are used for BaTiO3 synthesis [127]. These are compatible with room temperature and ambient atmosphere. Violent reaction with moisture in air and rapid hydrolys is can be avoided. Among the possible solvents of barium acet ate, glacial (wate r-free) acetic acid (AcOH) was selected in this research. Acetic acid not only serves to dissolve Ba-acetate but also controls the rapid r eaction of titanium ions. By preventing segregation of Ti-rich phase and Ba-rich phase, quasi-atomic mixing of precursors can be maintained [128].


37 To utilize precursor sol for electrospinni ng, the precursor mixture must meet the following requirements: (1) Moderate hydrolysis and condensation rate (2) Compatible to room temper ature and atmosphere pressure (3) Enough viscosity to give the bac kbone structure on nanofibers continuous structure Among the various Ba and Ti precursors pr esented, barium acetate and titanium isopropoxide meet these requirements. Based on the requirements fulfilled, barium acetate and titanium isopropoxide were selected as starting materials in this research. Viscosity can be easily contro lled by adding the proper polymer and its selection criteria will be presented on Chapter 4.2. 4.2 Precursor selection: Backbone Polymer Rheological properties of precursor mixtur es directly affect the geometry of nanofibers in electrospinning. Viscosity play s an important role in the fiber formation process since it forms the backbone of the el ectrospun fibers as we ll [44]. Even though homogeneously dispersed sol can be prepared easily by mixing metal alkoxide precursors, precise control of rheological properties must be considered. It has been observed that a precursor with low viscosity prevents formation of fibrous shapes. In this case, formation of the Taylor cone is hindered by dominati ng electrostatic forces and results in the spraying of liquid precursor (R ayleigh instability mode throughout the fiber synthesis). On the other hand, with high viscosity, high su rface tension leads to an increased electric field to surpass the force balance between electrostatic force and viscosity and surface tension. Higher viscosity increases the nanof iber diameter as well. Viscosity should not vary radically with time to achieve homoge neity of nanofiber dimensions [44, 65-67].


38 In general, polymers with high molecular weight (Mw) provide high viscosity. However, a small amount of polymers (polymer solution) is preferable [44]. Chances for generating structural defects and degree of random substitution of ions can be increased with both the amount and molecular weight of the polymer during the subsequent heat treatment. Recently, viscosity has been controlled and modified by adding polymers to the precursor sol. Poly (vinyl pyrrolidone, PVP) [74], Poly (vinyl acetate, PVA) [129], Polycarbonate (PC) [130], and Poly (ethylene oxide, PEO) [79] with different molecular weights have been widely used. Compatib ility between precursors and polymers should be considered since rapid reaction of salt or metal alkoxide with the dissolved polymer can occur. By changing the precursor or polymer, suitable additives can prevent rapid reaction and precipitation. Among the polymer s presented above, due to convenience in preparing a solution by co-di ssolving the polymer in alcohol and/or water, poly (vinyl pyrrolidone, PVP, Mw = 1,300,000) was selected as the backbone forming polymer. Previous research showed that PVP provides narrow distribution of fiber diameter under fixed conditions, while broad di ameter distributions were ob served from PVA, PC, and PEO [80]. Based on studies regarding compatibility of various barium and titanium precursors as well as different kind of polymers, it wa s proved that barium acetate and titanium isopropoxide precursors combined with PVP fo rms one of the most suitable and stable precursor mixtures for electrospinning of nanofibers at ambient conditions. Violent reaction between precursors, rapid hydrolys is, and condensation are expected to be prevented.


39 The stability of the precursor mixture was confirmed. The precursor mixture formed a transparent sol and kept its transparency for more than 5 days. 4.3 Experimental Procedure 4.3.1 Precursor Mixture Synthesis The BaTiO3 precursor was synthesized using a composition with a Ba:Ti molar ratio of 1:1, which is known to exhibit th e highest dielectric constant when crystal structure is tetragonal. Acetic acid was used as a solvent for barium acetate. Note that, as explained in earlier section, acetic acid controls the differe nt reaction rates of titanium and barium ions, that in it prevents the fo rmation of the Ti-rich and/or Ba-rich phase. First, 1.275 g of barium acetate Ba (CH3COO)2 was dissolved in 3ml of acetic acid and stirred for 0.5 hour in a capped bottle. After fully dissolving the barium acetate powder in acetic acid, an equimolar amount (1.475 ml) of titanium isopropoxide [Ti((CH3)2CHO)4] was added by drops via a pipette. The barium and titanium precursor mixture was kept stirring for 2 hours to achieve a complete di ssolution and mixing condition. To provide viscosity to precu rsors and to form a backbone structure of the fibers during the electrospinning pr ocess, a solution consisting of poly (vinyl pyrrolidone) (PVP, Mw = 1,300,000) dissolved in 2-methoxyeth anol was mixed with a precursor solution (PVP: 0.2 g and 2-methoxyethanol: 3 ml). Likewise acetic acid, 2-methoxy ethanol controls the rapid hydrolysis and c ondensation. PVP with different molecular weight, such as Mw = 850,000, can be used. However, a lower molecular weight polymer restricts other parameters of electrospinni ng. Limitations on spinning distance and applied voltage generally occur from the precursor mixture with low viscosity sol. In this regard, a high molecular wei ght polymer is preferable since the polymer forms the


40 backbone structure and will be burned out after subsequent heat treatment. The dissolved PVP solution was kept stirri ng for 2 hours in a capped bottle for complete mixing. A separately synthesized Ba and Ti precu rsor mixture and PVP solution was brought together to form a precursor-polymer mixtur e solution and was stirred in a capped bottle for 24 hours to form a homogeneous and transparent sol. 4.3.2 Electrospinning of BaTiO3 Nanofibers The precursors/PVP mixed sol was loaded in to a 5 ml plastic syringe (diameter: 12 mm, Luer hub, Fisher Science, Fairlawn , Ohio, USA) equipped with a metallic tip (diameter: 0.51 mm, chrome plated, Luer hub, Fisher Science, Fairlawn, Ohio, USA). A positive terminal was connected via an alligator c lip to a metallic tip. An aluminum plate (Aluminum 99.9 %, 1.0mm thick, 100 x 100 mm, Luer hub, Fisher Science, Fairlawn, Ohio, USA) worked as a ground counter electr ode. The distance betw een the tip end and the grounded collector were adjusted between 10 and 15 cm. Positive voltage between 15 to 30 kV was applied (EP 30-P, Gamma high Voltage Research Supply, Ormond Beach, Florida, USA) and electric fields be tween the tip end and the grounded collector were 1~3 kV/cm. The flow rate of the precursor mixture was controlled by a syringe pump, 1.27 cm3/h (BSP 99, Braintree Scientific Inc ., Braintree, Massachusetts, USA). Right after applying the voltage , as explained in the previ ous chapter, a liquid droplet deformed and at a critical value of the a pplied voltage, the fiber emission began. Fiber synthesis was conducted inside a fume hood, ambient atmosp here; temperature was kept at room temperature and humidity was at around 50 %. Fibers formed non-woven mats when they were collected on the grounded aluminum electrode. As-synthesized mats were dried in an oven (at 120 oC, 1 h) for further hydrolysis and water molecules removal. In order to obtain crystallized barium titanium oxide fibers, dried nanofiber


41 mats were subjected to heat treatment with different temperature ranges and time durations. A flow chart of the electrospi nning process is presented in Figure. 4-1. Fig.4-1. A flow chart of BaTiO3 nanofiber synthesis via electrospinning 4.4 Characterizations 4.4.1 Thermogravitic / Differential Thermal Analyzer (TG/DTA) Reaction processes during the heat tr eatment were studied by a ThermoGravimetric / Differential Thermal An alyzer (TG/DTA, STA449C, NETZSCH Instrument INC., Winchester, VA, USA). In formation obtained from TG/DTA was used as a base to design the a nnealing conditions. From pr evious annealing studies on BaTiO3, it was determined that the sol-gel based barium titanate crystallizes at temperatures between 600 and 700 oC [32, 123-127, 131-132]. Based on reference data, TG / DTA analysis was performed at temperatures up to 750 oC.


42 4.4.2 Scanning Probe Microscope (SPM) In order to achieve a more direct insight into the surface stru cture of electrospun nanofibers, a scanning probe microscope (SPM) (Dimension 3100, Veeco, Woodbury, NY.) was employed. SPM provides digital imag es of the surface of the nanofibers which allows observation on surface morphology from twoand three-dimensional stimulations. Measurements were performed with intermitte nt contact mode, 1.221 Hz of scan rate. 4.4.3 Scanning Electron Microscope (SEM) Scanning electron microscopy was empl oyed to study dimension and morphology change of the electrospun fibers. A field emission SEM JEOL-6335F was used in this study. Electrospun fiber mats were placed on the carbon tape and coated with carbon or gold-palladium. 4.4.4 X-Ray Diffraction (XRD) Phase evolution of electrospun barium tita nium oxide/PVP composite fibers was investigated by X-ray di ffraction patterns (Cu-K radiation = 1.5406 , scanning rate of 0.02 o/sec in 2 ranging from 20 to 80o) with continuous scan mode. The accelerating voltage and current was 40 kV and 20 mA, re spectively. Step mode was employed to study tetragonal distortion, lattice para meters (scanning rate of 0.01 ~ 0.005 o/sec, 5~10 sec/step). The information obtained from scan st ep mode was used to estimate crystallite size. Both Scherrer’s formula and Williamson-Hall plot were employed [133]. 4.4.5 Transmission Electron Microscope (TEM) A JEOL 2101F was utilized to further analyze the microstructure accelerating voltage: 200kV, camera length: 12~15 cm). Cr ystal structure of th e electrospun fibers was studied by selected area electron diffraction patterns (SAEDP).


43 4.4.6 Raman Spectroscopy Electrospun BaTiO3 nanofiber mats heat treated at various conditions were placed on microscope slides. Isopropanol was dripped on to the slide and fiber mats were placed and dried in air for 1 hour. Measurements were conducted with an inVia Reinshaw Raman system (Reinshaw, Hoffman Estates, IL, USA) equipped with a High Power Near Infrared (HPNIR) diode laser that had an excitation wave length of 785 nm. All measurements were performed at ambient atmo sphere. Nanofiber mats were attached on microscope slides. Calibration of the Rama n system was performed using silicon as a reference. The laser beam of a few microm eters spot size was positioned on the sample with the aid of an optical microscope. The Raman spectra were observed in the backscattering mode. The resolution was better than 2 cm-1, and the spectra acquisition consisted of three accumulations of 30 seconds. The Raman spectra were recorded in the range between 200 and 1200 cm-1. 4.4.7 X-Ray Photoelectron Spectroscopy (XPS) The X-ray photoelectron spectroscopy (XPS) was employed to examine the oxidation status of Ba a nd Ti ions from the BaTiO3 nanofibers heat treated at different conditions. XPS spectra we re obtained using a KRATOS XSAM 800 spectrometer (Kratos Analytical, Manchester, UK) with Mg K radiation. An ultrahigh vacuum is required to minimize or avoid surface contamin ation since the technique is sensitive to the composition of only a few atomic layers. BaTiO3 nanofibers were electrospun directly on silicon wafer pieces (approximate ly 1.5 cm by 1.5 cm) for 10 seconds. The silicon wafers were cleaned with acetone in an ultrasonication ba th for 30 seconds and rinsed with deionized water.


44 4.4.8 Fourier Transform Infrared Spectroscopy (FTIR) FTIR emission spectra from barium tita nium oxide composite nanofibers after polymer burnout (450 oC 3 h) and BaTiO3 nanofibers heat treated at 750 oC 16 hours were collected using a Thermo Electron Magna 760 (Thermo Electron Corporation, Waltham, MA. USA) in transmission mode. The BaTiO3 nanofibers were mixed with KBr in a 5:100 weight ratio. Each sample was analyzed in duplicate and sixteen transmittance mode scans were taken for each sample at a resolution of 4 cm-1 in the midinfrared range of 4000 – 400 cm-1.


45 CHAPTER 5 EFFECT OF HEAT TREATMENT ON PHASE EVOLUTION 5.1 Thermoanalytical Investigations: Op timizing Heat Treatment Conditions Reaction processes during the heat tr eatment were studied by a ThermoGravimetric / Differential Thermal Analyzer (TG / DTA). Information obtained from TG / DTA was used to figure out the proper anneali ng conditions. Fig.5-1. TG / DTA profile of electrospun BaTiO3 nanofibers heated from room temperature to 750 oC (10 oC / min) Based on reference data, TG / DTA analysis was performed at a temperature up to 750 oC. Non-woven fiber mats were brought into the alumina crucible and heated in air


46 at a rate of 5 oC/min and 10 oC/min. Aluminum crucible served as reference. Figure 5-1 shows characteristic TGA and DTA curves of the BaTiO3 nanofibers. TG/DTA curves showed multiple steps of we ight loss. The first weight loss of 17.4 % was observed with an endot hermic reaction at around 91 oC. The endothermic reaction corresponded to the loss of water molecules and (CH3COOH). Weight loss of 16.6 % was found for the temperat ure range of 300 to 420 oC, accompanied with a strong exothermic reaction at 329 oC. This is attributed to the burnout of poly (vinyl pyrrolidone, PVP) in air. Two more exothe rmic reactions were observed at 421.97 and 581.97 oC. Organic groups such as CH3COO-, expected to chelate Ba and Ti, decomposed at 421.97 oC, exhibiting an exothermic reac tion. An exothermic reaction at 421.97 oC is assigned to the decomposition of or ganics (carbonates and/or oxycarbonate systems). The main source of the carbonate phase is thermal decomposition of acetic acid or an acetate compound. The third we ight loss of 5.71 % was accompanied with an exothermic reaction at 581.97 oC. This reaction at 581.97 oC is consistent with the decomposition of intermediate phases. The last exothermic peak with 4.62 % of weight loss at 746.54 oC is indicative of the onset in crystallization of the perovskite structure [134]. This peak and weight losses were at tributed to the reaction between decomposed intermediate phases to form BaTiO3. Further proof will be shown in the following XRD analysis. An overall 44% of weight lo ss was observed after heat treatment. 5.2 Designing Heat Treatment Co nditions: Two-step Annealing Heat treatment conditions were designe d based on information obtained from thermoanalytical investigation (TG/DTA). Th e polymers that served as the backbones of the electrospun nanofibers should be removed to form desired ceramic nanofibers. Our study revealed that a PVP burnout process was completed at around 400 oC. To provide


47 enough thermal energy and time for complete decomposition of the polymers, a two-step heat treatment process was introduced. As presented in Figure. 5-2, synthesized and dried BaTiO3 nanofiber mats were subjected to a polymer burnout process at 450 oC for 3 hours. The temperature incr easing rate was fixed at 5 oC/min. Fig.5-2. Schematic of tw o-step annealing process Heat treatment conditions were focu sed on temperature between 550 and 750 oC. Higher temperature annealing (up to 1200 oC) was performed as a re ference to investigate temperature effects on structure evolution, morphology, and crystallite size. Heat treatment conditions of as-s ynthesized and dried BaTiO3 nanofibers are summarized in Table 5-1. Polymer burnout and crystalliz ation were performed under an ambient atmosphere in a conventional box furnace. Fiber mats were cooled down to room temperature (5 oC/min).


48 Table 5-1. Two step annealing conditions of electrospun BaTiO3 fibers Pre-heat treatment temperature Subsequent heat treatment temperature Duration of heat treatment *450 oC 3 hours 550 oC 12 hours 6 hours 12 hours 16 hours 24 hours 580 oC 48 hours 6 hours 12 hours 16 hours 600 oC 24 hours 620 oC 48 hours 24 hours 650 oC 48 hours 680 oC 24 hours 12 hours 24 hours 700 oC 48 hours 1 hours 12 hours 16 hours (3 oC/m) 450oC 3 hours (5 oC/m) (Subjected to all nanofibers) 750 oC 16 hours (5 oC/m) (* no further heat treatment after polymer burnout) 5.3 Electrospun BaTiO3 Nanofibers SEM images of Ba-Ti-PVP composite fibers synthesized and dried at 120 oC for 1 hour are presented in Figure. 5-3. The una ligned structure of nanofibers was observed from as-synthesized mat (Figure. 5-3(a)). It can be seen in Figure. 5-3 that the surface of as-synthesized and dried composite fibers ar e smoother due to the amorphous nature of


49 the barium titanium PVP composite. Figur e. 5-3(b-d) shows fiber morphology after drying at 120 oC. No dimensional change was obser ved from the dried fibers composite. Fig.5-3. SEM images of Ba-Ti-PVP composite nanofibers; (a) as-synthesized fibers, (b, c) dried at 120 oC for 1 h. (d) nanofiber diameter before polymer burn out (applied voltage: 10kV, working distance: 10 cm) Diameters of at least 25 fibers were m easured by dimensional analysis software equipped with SEM for both as-synthesized and heat treated fibers. Average diameters of as-synthesized fibers are 273 nm with 57 nm of standard deviation. After annealing at 750 oC for an hour, the PVP is fully decomposed and multi-grain structure fibers were obtained. About a 60 % reduction of the av erage diameter was attributed to PVP decomposition and organic molecule burnout (Table 5-2).


50 Table 5-2. Average and standard deviation of fiber diameters: as -synthesized and heat treated 750 oC (applied voltage: 10kV, working distance: 10 cm) Average fiber diameter (nm) Standard deviation (nm) As-synthesized 273 57 Fibers annealed at 750 oC, 1h116 30 SEM images of crystallized BaTiO3 nanofibers are presented in Figure. 5-4. Surface roughness increased due to multigrain st ructure development. Fiber continuity was maintained after heat treated at 750 oC. Typical lengths of th e annealed fibers were in the order of a few hundred micrometers up to a few millimeters. Fig.5-4. Electrospun nanofibers( a) SEM images of BaTiO3 nanofibers after heat treated at 750 oC, 1h. (b) Magnified image shows multi-grain structure (applied voltage: 10kV, working distance: 10 cm) Even though the diameter of the electros pun nanofibers was reduced (Figure. 5-4 (a)), their continuous microstructure was main tained. Crystallized fibers are made of grains between 25 and 50 nm in diameter (measured by TEM) with fiber diameter between 80 and 190 nm from the employed fi ber synthesis condition. An average fiber


51 diameter after crystallization is 116 nm. The length of the heat treated nanofibers is in the order of a few hundred micrometers up to millimeters. In order to get a better understanding of the surface structure of electrospun nanofibers, a SPM was employed. SPM provi des digital images of the surface of the nanofibers which allows observation on su rface morphology from twoand threedimensional stimulations. Fig.5-5. Electrospun nanofibers (a) SPM in termittent contact image of electrospun nanofibers deposited directly on go ld foil after heat treatment (750 oC, 1 h), (b) Three-dimensional image of electrospun nanofibers Nanofibers were directly deposited on gold foil (Au 99%) for 10 seconds followed by drying and heat treatment at 750 oC for 1 hour at an ambient condition. The SPM was operated in tapping mode to record height and amplitude images of the electrospun nanofibers. Scan size and rate were 1 m by 1 m and 1.221 Hz. Two and three dimensional image of electrospun nanofibers are presented in Figure. 5-5 (a) and (b) respectively. Both images clearly show nanofibers have multi-grain structure after heat treatment.


52 5.4 Investigation on Perovskite BaTiO3 Phase Evolution Phase evolution of electrospun barium tita nium oxide/PVP composite fibers was investigated by X-ray diffraction patterns. The XRD pattern obtained from nanofibers after PVP burnout revealed amorphous nature of the fibers (F igure. 5-6(a)). Fig.5-6. XRD patterns of BaTiO3 nanofibers heat treated at various condi tions; (a) 450 oC 3 h, (b) 550 oC 12 h, (c) 580 oC 12 h, (d) 580 oC 16 h. Highlighted on (c) and (d) indicates intermediate phases developed during the crystallization (applied voltage: 30kV, worki ng distance: 15 cm) The onset of crystallization was observed after annealing at 550 oC for 12 hours. A peak at around 31o of 2 which could be assigned to the {110} family, clearly developed (Figure. 5-6(b)). However, the intensities of other peaks did not show significant concentrations and it was difficult to verify their locations (2 ). With further heat treatment at a higher temperature and time (580 oC, 12 hours), the two peaks that did not


53 correspond to perovskite BaTiO3 were observed at 27o and 43o of 2 (Figure. 56(c)). Amorphous background intensity increa sed as well. These peaks with amorphous background were attenuated with prolonged duration. Fig.5-7. XRD patterns of BaTiO3 nanofibers heat treated at various condi tions; (a) 600 oC 16 h (b) 650 oC 1 h, (c) 750 oC 1 h, (d) 750 oC 12 h, (e) 750 oC 16 h (applied voltage: 30kV, worki ng distance: 15 cm) Considering exothermic reaction and weight loss at 581 oC from the TG /DTA profile, it is suggested that intermediate phases were formed below 600 oC as a part of perovskite BaTiO3 evolution. The unidentified peaks can be assigned to both BaTi2O4 and BaCO3. However, considering all other peak s which can be assigned to perovskite BaTiO3, it is plausible to say that the unident ified peaks may be caused by a non-reacted intermediate phase (BaTi2O4) at this annealing condition (600 oC).


54 But the assignment is not definitive since no exact match was found. The assignment will be further demonstrated with the Raman spectra study. The perovskite BaTiO3 XRD patterns developed with increasing temperature and time are presented in Figure. 5-7. Unidentified peaks were still observed from the nanofibers heat treated at lower temperatures (Figure. 5-7 (a-c)). Well defined perovskite BaTiO3 patterns appeared after 750 oC for 16 hours of annealing. This also matches well with the exothermic reaction peak that is due to the perovskite ph ase formation observed from the DTA at 730 oC. It is suggested here that unknown patterns observed at 27 and 43o with increased amorphous background at temperature range up to 650 oC are intermediate phases formed as a result of precursors decomposition. Relatively well defined perovskite BaTiO3 patterns were observed at over 750 oC. Increased time duration enhanced the intensit ies of the peaks. This crystallization temperature (750 oC) matches well with data obtai ned from the thermoanalytical investigation presented in Chapter 5.1. 5.5 Effect of Heat Treatment Time on BaTiO3 Phase Evolution on Electrospun Nanofibers Further investigations were performed to study the effect of heat treatment on phase evolution and morphology of electrospun BaTiO3 nanofibers at longer heat treatment durations. Since well-developed perovskite BaTiO3 was observed at 750 oC and 16 hours of heat treatment, it was proposed that longer annealing time could provide enough energy for crystallization and resulti ng perovskite evolution lower than 750 oC. Based on the assumption above, temperature ranges between 580 to 700 oC and 24 to 48 hours of heat treatment time were selected in this research.


55 Fig. 5-8. XRD patterns of BaTiO3 nanofibers heat treated for 24 h; (a) 580 oC, (b) 600 oC, (c) 650 oC, (d) 700 oC, Intermediate phases ~ 27o of 2 are highlighted on (a) and (b) (applied voltage: 30kV , working distance: 15 cm) The significant effect of prolonged heat treatment is that th e perovskite BaTiO3 phase was formed at a lower temperature. This may be due to the enough thermal energy for intermediate phase to decompose provided by longer annealing time. It is theorized that Ba2+ and Ti4+ could have enough energy to locate their sites in unit cells. While nanofibers subjected to 24 hours of anneal ing at a lower temperature showed high amorphous background with an intermediate phase (Figure. 5-8(a), (b)), XRD results of 48 hour of annealing nanofibers showed that peaks corresponded to perovskite BaTiO3 although intensities were relativ ely low (Figure. 5-9 (a-c)).


56 Fig 5-9. XRD patterns of BaTiO3 nanofibers heat treated for 48 h; (a) 580 oC, (b) 620 oC, (c) 650 oC, (d) 700 oC, Red box indicates removed intermediate phase with increased annealing time (applied vol tage: 30kV, working distance: 15 cm) Compared to BaTiO3 particles synthesized by a solid state reaction, the perovskite structure of electrospun BaTiO3 nanofibers can be achieved at temperature 300 oC to 500 oC lower than at a solid state reaction (1000 – 1200oC) [119-122]. 5.6 Tetragonal Distortion and Grai n Size of Electrospun BaTiO3 Nanofibers To further investigate te tragonal structure evolution, Lorentzian peak profile fitting routines were conducted on the XRD profiles of samples subjected to different heat treatment conditions. The average partic le size and the lattice di stortion (c/a ratio) were determined from the XRD. The room temperature XRD pattern of mainly the (002) and (200) reflection of electrospun BaTiO3 nanofibers was carried out to investigate tetragonal distortion. A split of (002) and ( 200) peaks is a charac teristic feature of


57 tetragonal perovskite. Scan step increments of 0.005o of 2 with a step time between 1 to 5 seconds were employed to acquire enough counts and to minimize noise. Automated powder diffraction profiling fitting (PM 1877) was employed. Fig.5-10. Lorentzian fitting of (a) (002) and (200) peaks around 45o BaTiO3 nanofibers heat treated at 620 oC 48 h. (b) (002) and (200) BaTiO3 nanofibers heat treated at 700 oC 12 h (applied voltage: 20kV, working distance: 10 cm) Profile fitting also re vealed that nanofibers heat treated below 580 oC and less than 24 hours did not show separation of (002) and (200) peaks. Separation of these peaks is characteristic of tetra gonal perovskite BaTiO3. Profiling fitting resu lts matched well with the ~2 scan since intermediate phases were obser ved and in that case it is unlikely to form a BaTiO3 phase. The split of the {200} family was observed from temperatures above 620 oC with 48 hours of heat tr eatment (Figure. 5-10). Table 5-3 shows the peak positions with 95 % confidence limits and goodness of the fit (Chi square and correla tion coefficient) when the obser ved intensity is assigned to a single peak compared with assign ing the reflection to two peaks.


58 Table 5-3. Comparison of 2 and R2 values on Lorentzian fitting, *DOF; degree of freedom Annealing condition Peak index Peak position (2 , degree) 2/DOF R2 (200) 45.253 (.0022) 711.821 0.8466 (002); split 45.038 (.0165) 620 oC 48 h (200); split 45.293 (.0072) 104.467 0.9744 (200) 45.218 (.0030) 119.97 0.9300 (002); split 44.956 (.0677) 700 oC 48 h (200); split 45.243 (.0214) 102.46 0.9508 (110) 31.581 (.0009) 658.71 0.9901 (101); split 31.521 (.0016) 750 oC 12 h (110); split 31.622 (.0053) 393.06 0.9942 Lattice distortion of nanofibers after heat treating at 700 oC at 12 hours was observed from the split of the {200} family as well (Figure. 5-10). The location of both {200} family corresponded well with reference da ta [135]. Peak splits of nanofibers annealed at 750 oC, 16 hours were studied with {110}, {200}, and {112} families since this condition resulted th e best defined pattern ( ~ 2 scan) among the subjected conditions. Scan step increment of 0.01o of 2 with a step time of 10 seconds was employed. Lorentzian peak profile fitting rou tines were conducted in the XRD profiles. XRD patterns after profile fittings are pr esented on Figure 5-11. Peak splits were observed from all {110}, {200}, and {112} families. These indicate that distorted structure was developed from this annealing condition. Table 5-4 presents the peak position w ith 98 % confidence limits and goodness of the fit (Chi square and correla tion coefficient) when the obser ved patterns are assigned to a single peak compared with two peaks assignment.


59 Fig. 5-11. Lorentzian fitting of BaTiO3 nanofibers annealed at 750 oC, 16 h (a) (101) and (110), (b) (002) and (200 ), (c) (112) and (211) Table 5-4. Comparison of 2 and R2 values on Lorentzian fitting, *DOF; degree of freedom Annealing condition Peak index Peak position (2 , degree) 2/DOF R2 (110) 31.470 (.0011) 56288.63 0.9849 (101); split 31.385 (.0106) (110); split 31.520 (.0042) 23355.43 0.9913 (200) 45.148 (.0016) 6537.44 0.9873 (002); split 45.012 (.0136) (200); split 45.231 (.0063) 4363.51 0.9916 (211) 56.115 (.0010) 13232.24 0.9848 (112); split 56.028 (.0103) 750 oC, 16 h (211); split 56.198 (.0127) 2418.34 0.9972


60 The unit cell dimension, a, of the BaTiO3 was changed with the heat treatment condition (Figure 5-12). Af ter 1 hour annealing at 750 oC, the c-axis developed as a lattice structure tran sformed from cubic to tetragonal. Fig. 5-12. Change of tetragonality, c/a ra tio, with annealing temperature and time c/a ratio showed increasing tendenc y with temperature and heat treatment time (except for 750 oC 1h nanofibers). Although obtained values are somewhat lower than the theoretical value (c/a = 1.01) at room temperature, it is reasonable and in agreement with reference data [136]. The effect of annealing time on lattice distortion was signifi cant for electrospun BaTiO3 nanofibers. The peak split of the {200} family was observed at a lower temperature, 650 oC for 24 hour annealed nanofibers and 620 oC for 48 hour annealed nanofibers (Figure. 5-13 (a)).


61 Fig.5-13. An XRD pattern and tetragonality (a) Lorentzian fitting of (002) and (200) peaks around 45o (650 oC 48 h annealing), Tetragonal distortion (c/a) of BaTiO3 nanofibers annealed at (b) 48 h, (c) 24 h (Applied voltage: 30kV, working distance: 15 cm) As presented in Figure 5-13 (b) and (c), although sudden decrease was observed at 680 oC for 24 hour nanofibers, the tetragonality of both annealing conditions resulted in an increasing tendency of tetragona l distortion with temperature. These is further proof of the increased time-effect that indicates that intermediate phase can be decomposed and as a consequence, free constituents may have enough thermal energy and time to locate their sites to form perovskite struct ures. A further view of the lattice parameters is plotted below.


62 Fig.5-14. Lattice parameters of electrospun BaTiO3 nanofibers; changes of unit cell parameters are plotted with temperature. (a: a-axis, c: c-axis, 3 at 2ct: cubic root of unit cell volume at subjected heat treatment) (Applied voltage: 30kV, working distance: 15 cm) The lattice constant cont inuously decreased to 750 oC, revealed that the size of the unit cell (3 at 2ct) increased with temperature (Figure. 5-14). (3 at 2ct) values observed from BaTiO3 nanofibers that are somewhat smalle r (less than 4.00) compared to the previous research, which reported over ~ 4.00 [103]. The sizes of the individual BaTiO3 crystallites composi ng electrospun nanofibers were calculated from XRD data using Scherrer’s formula. Individual peaks were fit to define their full width at half maximum inte nsity [137]. Scherrer’s formula is expressed as follows: BB d cos 89 . 0 (5-1)


63 Where d is the crystallite size and is the wavelength of CuK radiation (1.5406 for 1). B measures peak broadening, correct ed for instrumental effects with single crystalline Si (004) sample, Binst = 0.07 [137]. 1802 1 B (5-2) 1 and 2 are angle locations of full wi dth at half maximum (FWHM). The calculated peak width fo r the (110) and (111) BaTiO3 were used in this analysis. Crystallite size varia tions with heat treatment cond itions are presented in Figure. 5-15. The calculated crystall ite sizes in electrospun BaTiO3 nanofibers were 28 ~ 40 nm. Increasing trends of crys tallite size relative to heat treatment temperature were evident (Figure 5-15 (a)). Crystallite sizes obtained from the XRD analysis matches with TEM study (25 ~ 50 nm, Chapter 5.8). The crys tallite sizes after 24 hours annealing were 29 ~ 36 nm ( 20 % error) (Figure 5-15 (b)). Increasing tendency of crystallite size to heat treatment temperature was observed as well. Tetragonal struct ure crystallite sizes were larger than 30 nm (above 650 oC) for 24 hours annealing. Figure. 5-15(c) shows tetragonal crystallite sizes obtained from na nofibers after 48 hours annealing. Relatively uniform size distribution (28 ~ 30 nm, 20 % error) was observed throughout the temperature range. The general range of the calcu lated crystallite size is of interest, especially with regard to the crystal system of the BaTiO3. Previous research proposed different size limits of BaTiO3, generally considering particle size less than 100 nm to restrict crystallization to the cubic stru cture [36-40]. In this resear ch, it is worth noting that the 30 nm ( 4 nm) BaTiO3 crystallite size with tetragonal structur e (c/a = 1.007) was observed (Figure 5-12).


64 Fig. 5-15. Crystallite size change with heat treatment temperature at various annealing time; (a)1-16 h heat treatment, (b) 24 h h eat treatment, (c) 48 h heat treatment (Applied voltage: 30kV, working distance: 15 cm) This exceptionally small crystallite size of BaTiO3 with tetragonal structure suggests that in electrospun BaTiO3 nanofibers, the effect of strain energy and surface free energy, and depolarization energy may work in different manners compared to other 1-D or 2-D types of BaTiO3, which eventually stabilize c ubic structures at this size range. This will be further proved by Raman spectroscopy analysis in following chapters.


65 5.7 Morphology Change of Electrospun Nanofibers with Heat Treatment The surface morphology of the electrospun na nofibers were investigated by SEM. Fig. 5-16. SEM images of electro spun fibers; (a) PVP burnout at 450 oC 3 h, (b) 550 oC 6 h, (c) 580 oC 12 h, (d) 600 oC 16 h, (e) 700 oC 12h, (f) 750 oC 1h (Applied voltage: 30kV, worki ng distance: 15 cm) Even after the polymer burnout and organic (CH3COO-) decomposition, the fibers with smooth morphology was mainta ined a temperature up to 580 oC (Figure. 5-16(a-c)). The surface began to roughen at 600 oC 16 hours. The polycrystalline nanofibers together with the smooth surface fibers were presented in (Figure 5-16(d)). Above 700 oC BaTiO3 nanofibers crystallization is complete d and with increased surface roughness.


66 It can be seen that the BaTiO3 nanofibers subjected to higher temperature heat treatment have larger grain size (Figure. 5-17 (e) and (f)). The structure of the crystallized nanofibers is polycrystalline wi th porosity. The porous structure of the electrospun nanofibers provides a surface larger th an that of continuous films. Figure. 517 clearly shows the formation of particulates with temperature. The fibers heat treated at 1200 oC that barely sustain continuity with in creased grain size are presented in Figure. 5-17(b). Fig.5-17. Polycrystalline fibers heat treated at different conditions revealed grain growth with temperature; (a) 600 oC 16 h, (b) 1200 oC 1 h (Applied voltage: 10kV, working distance: 10 cm) 5.8 Tetragonal Perovskite Stru cture Evolution of BaTiO3 The development of electrospun nanofibe rs with tetragonal perovskite BaTiO3 was investigated by utilizing HR-TEM. The representative tr ansmission electron microscope images of nanofibers (polymer burned out at 450 oC) are shown in Figure. 518. An analysis of high resolution images revealed that the nanofibers are amorphous in nature and this was confirmed by a diffu sed selected area diffraction pattern.


67 Fig.5-18.TEM images of nanofibers (a) TEM image of Ba-Ti-O as-synthesized nanofibers. (b) High-resolution imag e of nanofiber. (c) Selected area diffraction pattern of amorphous nanofi bers (Applied voltage: 10kV, working distance: 10 cm). TEM images collected after nanofibers annealing provide additional crystal structures and microstructure details are presented in Figure. 5-19. The bright field TEM image confirms that the fibers consist of barium titanate grains of approximately 25 – 50 nm.


68 Fig.5-19. Bright field image of BaTiO3 nanofibers heat treated at 750 oC 1 h (Left) and magnified image which shows multi-g rains ranges between 25 – 50 nm. (Applied voltage: 30kV, working distance: 15 cm) High resolution TEM images were taken from the electrospun nanofiber and are presented in Figure. 5-20. A selected area electron diffraction pattern (SAEDP) of the individual grain collected from the fiber is pr esented as well. The lattice fringes show the crystal quality of the nanofibers. The observed pattern is consistent with a [-134] zone axis of perovskite BaTiO3 as confirmed by pattern simulation. The limited tilting condition imposed by the nanofiber dimensions made it difficult to collect data for the lower index zone axis. The orientation of the individual grain is limited as additional tilting results in diffraction patterns collected from more than one grain. However the SAEDP is further confirmation of the crystallin ity of the electrospun BaTiO3 nanofibers.


69 Fig.5-20. TEM images of nanofibers (a) A hi gh resolution image obtained from the tip of barium titanate nanofibers. (b) A high resolution image that shows lattice fringes. (c) Selected area diffraction pattern obtained from the electrospun BaTiO3 nanofiber; zone axis [-134]. (Applied voltage: 30kV, working distance: 15 cm) 5.9 Single Crystalline BaTiO3 Nanofibers via Electrospinning The chosen annealing procedure with a separation of polymer burnout and nucleation and growth proved to be successf ul in the formation of single crystal nanofibers. Nanofibers annealed at 750 oC for 16 hours showed single crystalline


70 BaTiO3 together with polycrystalline nanofibers. SEM images of single crystalline fibers with polycrystalline nanofibers are presented in Figur e 5-21(a, b). It can clearly be seen that a large amount of fibe rs with smooth morphology exists. The X-ray diffraction pattern corresponds well with perovskite BaTiO3 Figure. 5-21 (c). Fig.5-21.Single crystal nanofibers (a) SEM image of single crystal nanofibers with polycrystalline nanofiber. (b) Magnified image of si ngle crystal nanofibers. (c) XRD pattern obtained from nanofibe r mats containing single crystal fibers annealed at 750 oC 16 h. (d) Lorentzian fitting of (002) and (200) peaks. (Applied voltage: 30kV, working distance: 15 cm) The tetragonality of single crystal nanofi bers was investigated. Automated powder diffraction profile fitting (PM 1877) revealed separation of ( 002) and (200) peaks, which is characteristic of tetragonal perovskite BaTiO3 (Figure. 5-21(d)). The location of (002) and (200) peaks matches well with the referenced data [135].


71 Fig.5-22. Single crystal nanofibers (a, b) TEM images of single-crystalline BaTiO3 nanofiber after heat treating at 450 oC for 3h followed by 750 oC 16h. (c) Selected area diffraction pattern with [3-10] zone axis. (d) A high resolution image with lattice fringes. (Applied voltage: 30kV, worki ng distance: 15 cm) TEM analysis of the BaTiO3 nanofibers after heat treating at 750 oC for 16 hours revealed single crystalline nanofibers with diameters no larger than 40 nm and lengths up to 1m along polycrystalline nanofibers as pr esented in Figure.5-22 (a, b). The bright field image shows a single cr ystal nanofiber with thickness fringes. A selected area electron diffraction pattern collect ed from the fiber is presented in Figure. 5-22 (c). The


72 observed pattern is consistent with a [3-10] zone axis of pseudo cubic perovskite as confirmed by pattern simulation. Simila r to SAEDP obtained from polycrystalline BaTiO3 nanofibers, the limited tilting conditi on imposed by the nanofiber dimensions made it difficult to collect a lower index zone axis. A high resolution TEM image was taken from the single crystal nanofiber and is presented in Figure. 5-22 (d). The lattice fringes confirm the crystallinity of the nanofiber . It should be menti oned here that single crystalline BaTiO3 nanofibers synthesized via electros pinning are being reported for the first time. XRD analysis combined with TEM st udies proved crystallites which consist of electrospun BaTiO3 nanofibers were perovskite. Crys tallite sizes of ~ 30 nm ( 6 nm) can be synthesized with a chosen heat tr eatment condition with reasonable tetragonal distortion (~ 1.007). Splitting of peaks such as (200) is typically understood to indicate a tetragonal unit cell. However, although the XRD pattern presented so far is consistent with tetragonal BaTiO3, further proof of the tetragonality of the na nofibers is required. 5.10 Raman Spectroscopy of Electrospun BaTiO3 Nanofibers Studies and discussions ha ve focused on the structure change of nanostructure BaTiO3. The most widely adapted method for structure investigation is XRD, as performed in this research. However, it is difficult to reveal the structure of BaTiO3 from XRD analysis because of the line broadening of XRD peaks, which usually originate from the combined effects of the crystall ite size, non-uniform strain, and instrumental broadening. There are various methods which are be tter suited for investigation of local structure and subtle symmetry changes. Among those, Raman spectroscopy has been known as the most powerful tool to probe nanostructure BaTiO3 [138]. Unlike infrared


73 spectroscopy, Raman spectroscopy is not sens itive to adsorbed water molecules, but highly sensitive to the presen ce of both titanium oxide and carbon oxide systems. Information on cation coordination, bonding, and the molecular stru cture of electrospun fibers can be investigated. These characteristics of Raman spectroscopy are advantageous for the characterization of th e evolution of bonding a nd structure in the precursors and unknown phases [139]. In this research, investigation of subtle structural changes between the cubic and tetragonal phase of BaTiO3 at room temperature of the Raman spectra was collected at the different heat treatment stages. The Raman spectra were recorded in the range between 200 and 1200 cm-1. The Raman spectra of BaTiO3 nanofibers at the different heat treatment stages were collected for further proof of the tetragonality and phase evolution via intermediate phase decomposition. It has been known from group th eory and previous research that cubic BaTiO3 has no Raman active modes, while tetragonal BaTiO3 (space group P4mm) has 8 Raman active modes: 4E(TO+LO) + 3A1(TO+LO) + 1B1(TO+LO). The characteristic bands of tetragonal BaTiO3 at about 717 cm-1 were assigned to E(4LO)+A1(3LO) modes, while those around 520 cm-1 were assigned to E(4TO)+A1(3TO) modes, and those around 307 cm-1 were assigned to (E(3TO)+E(2LO)+B1) [140]. The spectra presented in Figure 5-23 (a) was obtained from BaTiO3 after polymer burnout (450 oC 3 h). Broad bands observed at around 500 cm-1 was consistent with the intermediate Ba-Ti-oxycarbonate system. A band observed at 1059 cm-1 is a characteristic peak for CO3 2[141] ions and this confirms that this spectrum represents the oxycarbonate system. It can be suggested that non-stoichiometric intermediate phases


74 were developed as a result of reactions between the Ba-Ti precursor and organic molecules from solvents and the polymer. Fig.5-23. Raman spectra of BaTiO3 nanofibers after heat treated at (a) 450 oC 3 h (polymer burnout), (b) 550 oC 12 h, (c) 580 oC 16 h, (d) 600 oC 6 h. Red circle indicates a characte ristic peak of CO3 2(applied voltage: 30kV, working distance: 15 cm). The broad shape of the spectrum is an i ndication of a disordered structure. The band observed at 1059 cm-1 is indicative of carbonate phase. Nanofibers heat treated at 550 oC and 580 oC are presented in Figure 5-23 (b, c). The significant differences observed from both conditions were broad bands followed by an intensity edge at around 850 cm-1, and increased intensity 1060 cm-1 as well as the


75 symmetric stretching of the CO3 2[141]. Evidenced by a band at 684 cm-1, both spectra suggest the existence of carbonate phases. Br oad bands followed by an intensity edge at around 850 cm-1 were observed from samples heat treated at 580 oC for 16, 24 and 48 hours. This intensity edge was recently explained by Kumar et al. [141]. They hypothesized that a sharp intens ity edge followed by a broad continuum is a result of a highly disordered and defective structure. Random substitution of cations and anion vacancies during the intermediate phase decomposition may introduce translational disorder and lead to the breakdown of the Raman selection rule. For BaTiO3, the highly polar Ti ions are responsible for broad bands with an inte nsity edge, which leads to a highly disordered and defective structure [141] . Furthermore, the exothermic reaction at around 580 oC (Chapter. 5.1) combined with XRD patterns for 580 oC samples support this explanation. Existence of a locally ordered structure was proven by presence of carbonates (CO3 2-). Oxidation status of cations is presented in the following chapter with XPS analysis. The solid state reaction between BaCO3 and TiO2, all of which as intermediate phases, has been proposed in sol-gel based BaTiO3. However, obtained Raman spectra revealed that there is no evidence of both BaCO3 and TiO2 as intermediate phases. The Ti and O band at around 634 cm-1 was unlikely to be assigned as a TiO2 phase and even broadened with increasing temperatur e and time indicates that the TiO2 phase was not formed during the precursor decomposition processes. The Raman band at 306 cm-1 appeared from 600 oC 6 hour heat treated nanofibers. Bands at 520, 639, and 729 cm-1 together with carbonate phases were observed as well. It was noted that the intensit y of the carbonate band (1060 cm-1) decreased significantly.


76 This is indicative of the removal of disordered structures with temperature, in other words, enhanced decomposition of intermedia te phases with temperature. XRD analysis did not show tetragonal distortion at 600 oC of heat treatment. High resolution XRD analysis did not determine ( 200) peak split either. The existence of bands which are assigned to ch aracteristic tetragonal BaTiO3 suggests the coexistence of tetragonal and cubic phases together with unreacted carbonates. Fig.5-24. Raman spectra of electrospun BaTiO3 nanofibers after heat treating at (a) 600 oC 16 h, (b) 750 oC 1 h, (c) 750 oC 16 h, Red circles on (a) and (b) indicate carbonate species. (Applied voltage : 30kV, working distance: 15 cm)


77 Raman spectra for higher temperatures, 600 and 750 oC for 1 to 16 hours annealed nanofibers are presented in Figure. 5-24. The intensity drop at 850 cm-1 was completely removed while the intensity of bands at 717, 520 and 307 cm-1, increased with temperature and annealing time. This is indicat ive of an ordered structure evolution. The appearance of these peaks is evidence of the presence of asymmetrical TiO6 octahedra and provides further confirmation that th e fibers are in fact tetragonal BaTiO3. Fig.5-25. Raman spectra of electrospun BaTiO3 nanofibers after 24 h heat treatment at various temperatures; (a) 580 oC, (b) 600 oC, (c) 680 oC, (d) 700 oC, a red circle at 850 cm-1 shows an intensity drop, which is an indicative of disordered and defective structure. (Applied vo ltage: 30kV, working distance: 15 cm) It should be noted that there were ad ditional bands observed at around 640 and 1059 cm-1. The band at around 640 cm-1 was attributed to the strong scattering from high


78 temperature hexagonal BaTiO3. The band at around 1059 cm-1 was attributed to a carbonate that always forms with the hexagona l phase [39]. These bands were observed up to 750 oC for 1 hour of heat treatment (Figure. 5-24 (a), (b)). Fig.5-26. Raman spectra of electrospun BaTiO3 nanofibers after 48 h heat treatment at various temperatures; (a) 580 oC, (b) 620 oC, (c) 650 oC. Intensity edge observed on 5-25 (a) was removed with temperature (5-25 (b)) and complete removal of carbonate was observed temperature 100 oC lower than 16 h annealing. (Applied voltage: 30kV, working distance: 15 cm)


79 However, previous rese arch proved that BaTiO3 nano particles synthesized from polymeric precursors contain (111) planar de fects originating from the reducing condition during organic decomposition. Under reducing condition, Ti3+ substitutes Ti4+ and intermediate phases are formed. Intermediate phases disappear after forming BaTiO3 at higher temperatures but remain as planar de fects and generate sate llite peaks at around 640 cm-1 which is very similar to the hexagonal phase [138, 143-144]. Since the XRD and TEM investigation did not find any existence of a hexagonal phase, a plausible explanation is that electrospun BaTiO3 nanofibers are tetragonal with extended planar defects such as (111) tw ins. Especially, XRD pattern of hexagonal BaTiO3 shows significant difference fr om cubic and tetragonal BaTiO3. The disappearance of th e carbonate band (1059 cm-1) which always forms with the hexagonal phase confirms the suggestion that decomposition of the remaining Ba-Tioxycarbonate intermediate phases as BaTiO3 phase forms. Raman spectra obtained at different heat treatment stages showing disappearance of carbonate bands are present in Figure. 5-25 and 5-26. In both conditions, a highly defective and disordered structure was observed at 580 oC (Figure. 5-25(a), Figure. 526(a)) with intensity edges at around 850 cm-1. It should be mentioned here that a highly ordered structure, perovs kite tetragonal BaTiO3 can be achieved at least 400 oC lower than the solid state reaction (T > 1000 oC). 5.11 Oxidation State Change of Cations with Heat Treatment; XPS Study It was proposed in Chapter 5.7 that metal i ons with different oxidation state may be generated during the decomposition of interm ediate phases. The X-ray photoelectron spectroscopy (XPS) was employed to examine th e oxidation status of Ba and Ti ions from the BaTiO3 nanofibers heat treated at different conditions. BaTiO3 nanofiber


80 synthesis conditions described on chapter 4 we re applied for consis tency. All binding energies were referenced to C 1s at 2 81.0 eV. Multiregional spectra of Ti 2p photoelectron peaks were obtained and peaks were fitted using Lorentzian curves due to contributions of different chemical states or coordination. Fi gure. 5-27 represents examples of Ti 2p (Ti 2p3/2, Ti 2p1/2) photoelectrons obtained from BaTiO3 nanofibers after heat treatment. Fig. 5-27. Ti 2p XPS spectrum of BaTiO3 nanofibers; (a) heat treated at 450 oC 3 h, (b) 550 oC 6 h, (c) 750 oC 16 h The analysis of the Ti 2p spectra varia tions with annealing temperature yielded information on the oxidation status of titanium during crystallization. Figure. 5-27(a) shows the XPS spectrum of nanofib ers after heat treatment at 450 oC for 3 hours. It was


81 found that those Ti binding en ergies of 459.0 and 465.1 eV, were consistent with the reported Ti4+. On the other hand, additional peaks were observed with binding energies 458.2 and 464.2 eV. These additional states were assigned to Ti2+ [144]. The Ti 2p2/3 and 2p1/2 peaks corresponded to the BaTiO3 observed after higher temperature annealing (Figure 5-27(b)). It was noticed here that only one peak (464.5 eV) showed 2p1/2 and the lower oxidation status Ti peak completely disappeared. The intensity of Ti 2p2/3 was enhanced as well [138]. Th e spectra obtained from complete crystallization revealed no exis tence of Ti species with lo wer oxidation state (Figure 5-27 (c)). Only Ti4+ peaks of perovskite BaTiO3 were observed. The peak at 458.09 eV was assigned to BaTiO3 and was consistent with precious research. Fig.5-28. Ba 3d XPS spectrum of BaTiO3 nanofibers; (a) heat treated at 450 oC 3 h, (b) 550 oC 6 h The peak located at 458.9 eV did not matc h any of previously reported Ti of the BaTiO3 system. But its binding en ergy corresponds to SrTiO3 which has a perovskite structure [145]. This indicates that the peak located at 458.9 eV represents binding energy of unreported perovskite BaTiO3.


82 Figure. 5-28 shows the oxidation status of Ba ions. The binding energy of 779.3 eV obtained from nanofibers annealed at 450 oC (Figure 5-27 (a)) co rresponds well with barium oxide [138, 144-145]. Further inve stigation at a higher temperature (550 oC, 6 h) revealed Ba ions and bi nding energy 779.5 eV combin ed with hydroxyl ions (OH-) and is presented in Figure. 5-28(b). Observations of non-stoichiometric barium oxide and barium ions bound with hydroxyls are indica tive of intermediate phase formation followed by disordered and defective st ructures due to the decomposition and crystallization of perovskite BaTiO3 [138, 145-148]. The existence of the Ti species with lowe r oxidation status in the nanofibers heat treated at a lower temperature, further s upports the suggestion pr ovided in an earlier section. As shown in Figure. 5-28, the amount of BaTiO3 formation increased with a decrease of Ti3+ ions. It can be suggested here again that, intermediate Ba-Tioxycarbonate, substituting Ti3+ or Ti4+ are formed in the electrospun BaTiO3 nanofibers crystallization process. The highly defec tive and disordered structure of electrospun BaTiO3 can be introduced during the random substitution of Ti3+ and Ti4+. Oxygen vacancies can be introduced to maintain char ge balance as well. These are in positive agreement with the observed Raman spectra discussed in an earlier chapter. The Ti3+ ions disappeared as perovskite BaTiO3 crystallized but the disord ered structures remained as defects and generated a Raman band (640 cm-1) similar to hexagonal BaTiO3. 5.12 Crystallization Reaction of Electrospun BaTiO3 Nanofibers Theoretically, the amorphous mixture of Ba-O -Ti can be turned into various types of oxides and carbonates; BaTiO3, TiO2, BaCO3, BaTi2O4, BaO, BaTiO11, and numerous oxycarbonate systems are possible.


83 BaTiO3 crystallization th rough the solid state reaction between TiO2 and BaCO3 has been claimed since BaCO3 has been proven to be the most stable phase in/among ambient conditions. BaCO3 + TiO2 BaTiO3 +CO2 (5-3) If an assumption is made that BaCO3 forms prior to BaTiO3 crystallization in air, then BaTiO3 crystallization can concei vably crystallize above 172 oC through a solid state reaction. However, the XRD investigation did not determine any existence of BaCO3 and the 450 oC annealed sample revealed the amor phous nature of composite nanofibers. Unidentified peaks at ~27 and ~43 of 2 can be assigned for not only BaCO3 but all intermediate Ba-Ti-O systems as well. The maximum intensity of BaCO3 isomorphs appears ~ 25 of 2 and this value is over the averag e error. The observation from the Raman study did not determine the TiO2 phase that supports the solid state reaction and it is unlikely to happen in the electrosp un Ba-Ti-O composite to form BaTiO3 nanofibers. This was evidenced by defective and highly disordered structures observed via the Raman spectra. It has been reported that ~16 % weight loss was observed when BaTiO3 was crystallized through solid state reactions [141]. The observed weight loss from electrospun composite fibers was about one fort h. Therefore, combined results suggest that solid state reactions s hould be excluded for phase formation of electrospun BaTiO3 nanofibers. From the observation above, different re action processes via intermediate phase formation and decomposition in electrospun BaTiO3 nanofibers is proposed here. (1) A hydroxyl titanyl acyl ate precursor is synthesi zed via a reaction between acetic acid and titanium is opropoxide. When titanium isopropoxide is mixed with


84 the acetic acid, the titanium precursor experiences the following reaction to form hydroxyl titanyl acylate precu rsor. The possible crys tallization process of electrospun nanofibers are as follows. [(CH3)2CH4]4Ti + CH3COOH (acetic acid) Ti[(OAc)X(OC3H7)y] (5-4) (2) After forming a hydroxyl titanyl acylate precursor, the reaction with barium acetate follows and forms intermediate phase, Ba-Ti-oxycarbonate phases. Ti[(OAc)X(OC3H7)y] + Ba(Ac)2 Ba2Ti2O5CO3 + CO2 +H2O (l) (5-5) (3) The BaTiO3 nuclei are finally crystallized in a highly disordered composite matrix of nanofibers after experienci ng oxycarbonate phase decomposition at a higher temperature. (4) Crystallization of amorphous matrix by growth of BaTiO3 nuclei. Fig.5-29. TEM images of BaTiO3 nanofibers after annealing at 600 oC, 1 h. A High resolution image of a nanofiber shows partially crystallized structure in amorphous matrix. (Obtained from hi ghlighted section of nanofiber) Figure. 5-29 demonstrates the microstr ucture evolution of nanofibers after annealing at 600 oC 1 hour. Partial crystallized structur es that could clearly be seen were


85 embedded in an amorphous matrix. This is supportive evidence that crystallization of electrospun BaTiO3 nanofibers occurs by nucleation a nd a growth process as opposed to solid state reaction between TiO2 and BaCO3. The fact that no TiO2 was ever observed throughout this work supports this explanation as well. 5.13 Effect of Electric Fi eld; Electrospinning Effect of the synthesis method, electrosp inning was studied by precipitated BaTiO3 particles from the precursor and PVP mixture. As demonstrated in chapter 2, it is well known that nanofibers experience hi gh electric fields. However, the effect of the electric field on ceramic nanofibers has not proved. (e .g. chances for trapped charge effect on nanofibers crystallization). In this section, structural analysis for precipitated particles is presented. Fig. 5-30. XRD patterns obtained from (a) BaTiO3 particles precipitated from Ba and Ti precursors mixed with PVP, (b) BaTiO3 particles precipitated from Ba and Ti precursors without PVP add ition. (Heat treated at 750 oC 16 h) Particles were obtained from the precursor mixture used in the electrospinning process demonstrated in Chapter 4.3. Two di fferent types of mixtures were synthesized; the precursor only (barium acetate and titan ium isopropoxide dissolved in acetic acid)


86 and the precursor solutio n mixed with PVP. Both precurs or mixtures were stirred in capped bottles for 5 days to complete hydrolys is and condensation. Precipitated particles were collected and dried at 120 oC for 1 hour followed by a 2-step annealing process (750 oC, 16 h). XRD patterns obtained from precursors only solution and precursor/PVP mixture are presented on Figure. 5-30. Pa rticles obtained from different precursors showed a highly amorphous background as we ll as Ba-Ti-oxycarbona te peaks at around 27o of 2 which was not observed from electrospun BaTiO3 nanofibers heat treated at the same condition. Further inve stigation was carried out by Raman spectroscopy (Figure. 531). Fig.5-31. Raman spectra obtained from (a) BaTiO3 particles precipitated from Ba and Ti precursors mixed with PVP, (b) BaTiO3 particles precipitated from Ba and Ti precursors without PVP add ition. (Heat treated at 750 oC 16 h), A peak at 882 cm-1 (a red circle) indicates peroxide species (assignm ent is not definitive)

PAGE 100

87 All bands correspond to the BaTiO3 were observed from particles obtained from both conditions. On the other hand, ba nds which were not assigned to BaTiO3 were observed as well (805, 883~832 cm-1). Bands at 805 cm-1 were attributed to BaO2 while the bands at 883~832 cm-1 were probably the result from peroxide species, but the assignment is not definitive. It is somewhat ambiguous to determine that the disordered structures or non-reacted phases were due to the lack of an electric field, in other words, electrospinning. Then again, it also implie s the process itself may af fect crystallization. Unfortunately, a lack of published background on electric field effects prevented further research of the electric field effect and this remains a future project.

PAGE 101

88 CHAPTER 6 TETRAGONAL STRUCTURE STABILIZ ING MECHANISM ON ELECTROSPUN BARIUM TITANATE NANOFIBERS 6.1 Stabilizing Mechanism Affecting the Critical Crystallite Size of Electrospun BaTiO3 Nanofibers It is suggested here that electrospun nanof ibers are on the border of a constrained and unconstrained system since electrospun fibers have defined geometry but are expected to release induced strain energy by deformation. As a result, the phase stabilizing mechanism (i.e. factors affecting critical crystallite size) working on electrospun nanofibers are expected to s how a unique behavior different from both unconstrained (e.g. free particles) and constrai ned (e.g. agglomerate particles and/or thin film) states. The surface fr ee energy of electrosp un nanofibers should be lower compared to free particles. This can be reasoned from fiber dimensionality. On the other hand, due to the polycrystalline structure, strain energi es originated from different factors will be introduced to the electrospun nanof ibers, such as clamping. Expected origins of strain energy on electrospun BaTiO3 nanofibers are as follows: (1) Hydroxyl defects introduced duri ng the reaction of precursors and decomposition of intermediate phases; (2) Point defects introduced by random s ubstitution of cations and anion vacancies during the BaTiO3 crystallization; (3) Strain energy originated during th e crystallization of tetragonal BaTiO3 in a fiber matrix. The structure of the matrix can be either cubic BaTiO3 or amorphous barium titanium oxide composite fibers.

PAGE 102

89 (4) Since electrospun nanofibers are polycry stalline in nature after experiencing heat treatment, strain energy may be induced by clamping between particles composing electrospun nanofibers. As mentioned earlier, in its qua si state, strain energy stor ed in the nanofibers can be released with different manners of constr ained and unconstrained systems. In this research, phase stabilizing mechanisms wo rking on electrospun na nofibers are studied. Suggested strain energy introducing mechanisms were investigated based on experiments and thermodynamics. The crystallite size and structure of electrospun BaTiO3 nanofibers were presented in Chapter 5 that the structur e is tetragonal with reasonable tetragonality (c/a ~ 1.007). As introduced in Chapter 3, various factor s affect the structur e of nanosize BaTiO3. The dominant stabilizing mechanism may vary with starting precursors, synthesis conditions, and the dimensions of final product. In th is chapter, the phase stabilizing mechanisms that led electrospun BaTiO3 nanofibers as tetragonal are investigated. 6.2 Hydroxyl Induced Strain Energy Hydroxyls (OH-) that reside in oxygen sites ha ve been suggested to introduce defective and highly disordered structur e [103]. Since electrospinning utilizes a conventional so-gel process, hydroxyls formati on is expected as a result of precursors and solvent decomposition, and evapor ation during heat treatment. Fourier transform infrared (FTIR) emi ssion spectra from barium titanium oxide composite nanofibers after 600 oC with 1 hour heat treatment and BaTiO3 nanofibers heat treated at 750 oC 16 hours were collected. Figure. 6-1 shows the FTIR spectra for the electrospun BaTiO3 nanofibers heat treated at different conditions. The presence of both carbonate and hydroxyl ions may be

PAGE 103

90 clearly seen with the band at a wave number of 1420-1300 cm-1 being characteristic of carbonate, while the other band at 3500-2900 cm-1 corresponds to hydroxyl ions [102]. Fig.6-1. FTIR analysis of the electrospun BaTiO3 nanofibers; (a) heat treated 600 oC 1 h, (b) 750 oC 16 h It can be seen that with an increase in the heat treatment temperature, the intensity of the hydroxyl peaks gradually decreased a nd several bands assi gned to hydroxyl groups disappeared. The center position of the absorpti on band shifts gradually toward a smaller wave number due to an increase of the Ti-O stretching. The decr ease of the Ti-O bond distance is due to an increase of the Ti-O bond strength [102]. 6.3 Depolarization Field Effect The depolarization field in ferroelectric ma terials is the field de veloped as a result of spontaneous polarization. When surf ace charge density appears as a result of spontaneous polarization, a depolarization field (Ed) is accompanied in the opposite direction of the polarization. The depolarizing field is defined as follows [87]. P N Ed d 0 (6-1) (Ed: depolarization field, Nd: depolarization factor, Nd ~1/3 for spherical geometry, P: polarization, 0: absolute permittivity)

PAGE 104

91 Size effects such as a decrease in Curie temperature (Tc), can be strongly affected by failure of relieving depolarization energy. Minimizing depolarization energy can be achieved by breaking up a domain into 180o domains while induced strains can be minimized by forming 90o domains (Figure 6-2). If the BaTiO3 crystallites composing electrospun nanofibers are unable to break up into domains, tetr agonal structures will turn cubic, and that will cease ferroelectricity. Fig 6-2 Schematic diagram of domain configuration with 90o and 180o domains Since various structural characteri zations (Raman, and XRD) confirmed tetragonality of electrospun BaTiO3 nanofibers, simplification was introduced by setting the negligible contribution of the depolarization field on tetragonal phase stabilizing mechanism. 6.4 Surface and Strain Energy in Electrospun BaTiO3 Nanofibers The effect of surface free energy and stra in energy on phase stabilization with critical size was explained in Chapter 3. It is generally accepted that nanoparticles have

PAGE 105

92 large specific surface areas. On the other hand, due to the dimensional differences, 1D nanostructure (wires and fibers) is expected to show a relatively smaller specific surface area. As shown in equation (3-13), the in fluence of surface free energy on critical crystallite may be significant. By adjus ting previously developed theories, the surface energy related to the volume work of th e electrospun fibers was calculated [34]. In equation (3-2), the difference of the fr ee energy of the tetragonal and cubic state was, o T cubic o T tetra o T cubic o T tetra o T cubic o T tetra o c t V TS S T H H G G G, , , , , , ) ( , Tc T tetra P cubic P Tc T o Tc trans Tc T tetra p cubic P o Tc transT dT C C S T dT C dT C H ) (, , , , , ,(3-2) While Cp is expressed as follows. 3 5 310 10 T c T b a Cp (6-2) Previously published thermodynamic data used here are a= 125.85, b=5.52, and c= -13.35 for both cubic and tetragonal phases. The standard molar entropy change at phase transition, So trans,Tc was 0.51 J/K/mol and the sta ndard enthalpy change of the reaction at phase transition temperature ( Ho trans,Tc) was -0.2 KJ/mol [34, 110]. Putting in the values presented, the free en ergy difference from cubic to tetragonal phase transition at room temperature of Go 298,V (t c) = 48 J/mol was obtained for BaTiO3. The assumption that was made in Chapter 3 was that Go surface, affects tetragonal and cubic phase transition critical size is expressed as follows: A Go surface 2d n (3-3)

PAGE 106

93 The excess surface energy related to the vol ume work can be estimated by inputting Go 298,V(t c) = Go surface. This is based on assumption th at free energy change related to the volume work is only affected by the surf ace energy change between polymorphs. In electrospun nanofibers, there is an assump tion that fiber is composed of 30 nm crystallites. This is based on electron mi croscopy observation from nanofibers heat treated at 750 oC for 16 hours and XRD analysis. Volume (V) and area (A) of a crystallite were calculated. 3 24 3 9 310 137 . 14 10 15 3 4 3 4 m r V (6-3) 2 15 210 87 . 2 4 m r A (6-4) The number of BaTiO3 particles per mol was also obtained from the density and molecular weight of BaTiO3. First, the molecular weight 233.19 g/mol was divided by density of the BaTiO3, 6.02 g/cm3. mol m mol cm cm g mol g MBaTiO w 3 6 3 3 310 73 . 38 73 . 38 02 . 6 19 . 233 (6-5) Then this value was divided by volume of a crystallite to obtain number of BaTiO3 per 1 mol. mol m mol m / 10 739 . 2 10 137 . 14 10 73 . 3818 3 24 3 6 (6-6) The surface energy related to volume work, c t,V can be obtained as follows [34]: 2 298 ,r n Go V t c (6-7)

PAGE 107

94 c t and V denote cubic to tetragonal tran sition and volume work respectively. 2 3 2 15 18 ,10 199 . 6 10 827 . 2 1 10 739 . 2 48 m J m mol mol JV t c (6-8) Another surface energy contribution, the solid-gas interface of tetragonal BaTiO3 should be considered in order to estimate the critical crystall ite size and strain energy. From the Raman studies, the tetragonal st ructure of the crystallite was confirmed (Chapter. 5). This strongly suggests th at 30 nm (error 20 %) tetragonal BaTiO3 crystallites were broken up into domains so as to minimize depolarization energy. This indicates that the domain wall area may account for a large portion of the surface area of the 30 nm crystallites. The average surface energy density of BaTiO3 domain walls has been estimated to 0.001 ~ 0.01 J/m2 [149-150]. Value differs depending on the sample preparation and experimental conditions. Consequently , the surface energy of the domain wall may be predominant on the su rface energy of the domain. The validity of this assumption will be confirmed by following numerical calculations. The overall amount of surface energy ( ct) can be estimated from a summation of surface energy of volume work ( ct,V) and solid-gas interface contribution of tetragonal BaTiO3 crystallites ( g-s tetra). tetra s g V t c t c , , (6-9) The surface energy term in equation (3-13) that determines critical crystallite size should be revised as follows: U G dt c V tetra s g V t c c ) ( , ,) ( 3 (6-10)

PAGE 108

95 It is demonstrated that total strain energy when one phase is grown as particles in the matrix of the other phase as [117], 2 2 1 1 3 22 1 2 1 E E R P UT (3-10) (1, and 2 denotes matrix a nd particles respectively) In this dissertation, strain energy calcul ation was tried. Unfortunately, direct calculation was unavailable due to the lack of fundamental mechanical and thermodynamic data. The thermal expansion coefficient ( ), Poisson’s ratio ( ), and elastic modulus (E) for tetragonal and cubic phases are not published separately (bulk BaTiO3; E: 50 ~ 100 GPa, : 14 x 10-6 /K, : 0.3). However, strain energy in crystallites composing electrospun nanofibers can be obt ained by calculating back from given and calculated values by rear ranging the equation (6-9) as follows [117]: ) ( , ,) ( 3t c V tetra s g V t cG d U (6-11) Unit conversion was performed on Gv(c t) of 48 J/mol by multiplying mol per unit volume of BaTiO3. 3 6 3 3 ) ( ,10 23 . 1 10 81 . 25 48 m J m mol mol J Gt c V (6-12) Based on the surface free energy of vol ume work (6-8) and surface free energy density of the BaTiO3 domain wall, the overall surface energy (-0.00619 + 0.001) and (0.00619 + 0.01) J/m2 and 30 nm of the crystallite diam eter were selected for possible strain energy calculation and pres ented on equation (6-13) and (6-14). (1)tetra s g V t c, , : (-0.00619 + 0.001) J/m2, Gv(ct): -1.23 x 106 J/m3, d = 30 x 10-9 m

PAGE 109

96 3 6 3 6 9 210 75 . 1 ) 10 23 . 1 ( 10 30 ) 001 . 0 00619 . 0 ( 3 m J m J m m J U (6-13) (2) tetra s g V t c , , : (-0.00619 + 0.01) J/m2, Gv(ct): -1.23 x 106 J/m3, d = 30 x 10-9 m 3 5 3 6 9 210 49 . 8 ) 10 23 . 1 ( 10 30 ) 01 . 0 00619 . 0 ( 3 m J m J m m J U (6-14) Previous study on zirconium dioxide (ZrO2) revealed strain energy of the tetragonal phase in the cubic matrix 4.6 x 107 J/m3 [118]. This is up to two orders of magnitude greater than that obtain ed from electrospun BaTiO3 (8.49 x 105 J/m3). It is suggested here that low strain energy predominantly cont ributed to the stabilization of the tetragonal structure of electrospun fibers . However, if the gas-solid surface energy of tetragonal BaTiO3 ( g-s, tetra) is smaller than 0.00619 J/m2, increase in critical si ze is expected as with a decrease in strain energy. This is very unlikely to happen since large strain energy prevents the sustaining of the tetragonal phase and increases th e phase transition of critical size [38, 118]. Here, the availabl e surface energy of the gas-solid interface of BaTiO3 crystallites can be defined as 0.00619 < tetra s g, < 0.01 (J/m2). Within the defined range of surface energy, the increasi ng tendency of the critical crystallite size with the increasing strain energy was mainta ined and was in agreement with previous studies [103, 118]. The smallest probable st rain energy in crystallites of electrospun nanofibers was found to be as low as 8.49 x 105 J/m3. The assumption made here is based on previ ous research that indicates strain energy in electrospun BaTiO3 nanofibers introduced by clampi ng between particles can be

PAGE 110

97 released by various methods. Crystallites ma y exert stress to their neighbors during growth. BaTiO3 thin films sandwiched between layers are restricted for deformation and as a result, strain energy is stored. On th e other hand, electrospun nanofibers are easier to deform. Fig.6-3. TEM image of BaTiO3 nanofiber after heat treatment at 750 oC for 1 h. Voids are highlighted by red circles Experimentally, grains that were grown out in different directions were observed. Crack formation is also an effective strain energy releasing mechanism. Cracks less then 10 nm are good examples. Cracks of nanofiber s after heat treatment and crystallites grown out from the fiber support this assump tion. Nanofibers with voids and roughened

PAGE 111

98 morphology are presented in Figure. 6-3. Void s are highlighted with red circles. By releasing stored strain energy via deformati on, relatively low stored strain energy in electrospun BaTiO3 nanofibers is expected. Another important source of strain ener gy is misfit during growth. Crystallites composing BaTiO3 nanofibers of 30 nm may have singl e crystal structures. It can be expected that in a single crys tal, the strain energy that or iginates from lattice mismatch should be negligible. The boundary of probabl e strain energy stored in a 30 nm BaTiO3 crystallite was estimated by divi ding obtained strain energy by vol ume term of a grain, (1 / (14.137 x 10-24 m3)). U= 8.49 x 105 J/m3; J m V m J U17 24 5 3 310 200 . 1 10 137 . 14 1 10 49 . 8 1 1 (6-15) From the calculations above, minimum stored strain energy in a crystallite composing electrospun nanofibers is 1.200 x 10-17 J. Considering that strain energy of fullerene (C60, 1.9 nm) of 4.99 x 10-17, stored strain ener gy of electrospun BaTiO3 crystallite is realisti c and reasonable [151]. 6.5 Strain Energy Release: Strained BaTiO3 Crystallites Scherrer’s equation was modifi ed with instrumental broadening, however, the line broadening effect from strain was not consid ered. Investigation on strain and size of electrospun BaTiO3 nanofibers was conducted by Williamson-Hall method. Merits of this method are estimation of strain and crystall ites size by considering all broadening effects (instrumental, size induced, and strain induced broadening) [133].

PAGE 112

99 Williamson-Hall equation is expressed as follows, sin 2 2 1 cos d (6-16) ( : integral breadth, d: volume averaged crystallite size, : microstrain) The extrapolation plot of Williamson-Hall equation gives the value of strain ( ) from the gradient and crystallite size from the ordinate intercept. By fitting the individual peak by a Voigt f unction, strain and crystallite size was obtained. (110), (200), and (211) patterns were chosen. XRD patterns were obtained with step-scan mode (0.01 step size, 10 s / step). Automated fitting (Philips profile fitting 1.0) was employed. Individual peak was fitted by a Voigt function. Table 6-1. Obtained and calculated values for Williamson-Hall plot (hkl) 2 FWMH (radian) Standard (radian) Structural (110) 31.490 0.355 0.008426 0.001256 0.007170 (200) 45.159 0.485 0.010467 0.001919 0.008548 (211) 56.148 0.502 0.012560 0.002442 0.010118 Fig. 6-4 displays Williamson-Hall plot of the resulting integral peak width as a function of scattering vector for the (110), (200), and (211) peaks for electrospun BaTiO3 nanofibers heat treated at 750 oC, 16 hours. The points in Figure 6-4 are all fall on a gr adient and an intercept which represent strain and crystallite size respectively. The gradient of Williamson-Hall plot indicated crystallites composing BaTiO3 nanofibers were strained and strain was found to be 0.255 %. The grain size of 37 nm was obtained. This matches well wi th crystallite sizes

PAGE 113

100 obtained by Scherrer’s formula (~ 30 nm) and is close to experimentally observed value (TEM, 25 ~ 50 nm). Tetragonal distortions were calculated from the data, {110}, {200}, and {211} families that used for Williamson-Hall plot. Obtained c/a ratio of {110}, {200}, and {211} families were 1.009, 1.004, and 1.007, respectively. c/a ratio further confirmed tetragonality of BaTiO3 nanofibers. Fig. 6-4. A Williamson-Hall plot of electrospun BaTiO3 nanofibers heat treated at 750 oC, 16 h (Q= 2sin / ) It should be noted here th at crystallites were strain ed. On Chapter 3, it was demonstrated that stored strain energy forces BaTiO3 to have cubic structure. The Williamson-Hall plot showed stored strain energy in BaTiO3 crystallites was released by strain and consequently contri buted to stabilize tetragonal phase at room temperature. The plot also confirmed crystallite size and its tetragonal distortion.

PAGE 114

101 6.6 Critical Crystallite Si ze of Electrospun BaTiO3 Nanofibers If crystallites were perfect, free particles, it can be assumed that no strain energy would be stored in electrospun BaTiO3 ( U = 0). In this case, it would be theoretically possible to estimate the crystallite size for tetr agonal to cubic transition. The equation (610) was modified for a strain energy fr ee state and presented as follows [123]: v tetra s g V t c cG d ) ( 3, , (6-17) The overall surface energy of tetragonal BaTiO3 ( g-s tetra) can range between 0.001 and 0.01 J/m2. However, when taking into acc ount that surface energy of volume work ( ct,V = -0.00619 J/m2), g-s tetra must be larger than 0.00619 J/m2. It should be noted that a lower limit can be calculated num erically. However, the lowest limit does not provide valuable information. For example, if (g-s, tetra) is slightly larger than the critical value, dc ~1.2 can be obtained but this is ev en smaller than the lattice constant (~4 ) of a BaTiO3 unit cell. In this research, the cri tical crystallite size with the highest surface energy ( (g-s, tetra)=0.01 J/m2) was obtained by utilizing equation (6-17). nm m m J m J dc29 . 9 10 2 . 14 10 23 . 1 ) 01 . 0 00619 . 0 ( 39 3 6 2 (6-18) Recalling that the XRD analysis (Scherre r’s equation) revealed particle size less than 30 nm as cubic BaTiO3, the calculated critical size without strain energy is close to what was experimentally observed. This i ndicates that crystallite s in electrospun BaTiO3 nanofibers have less strain energy compared to those synthesized via different routes. In conclusion, the stored strain energy of 8.49 x 105 J/m2 contributed to the achievement of 30 nm tetragonal BaTiO3 crystallites. Low strain energy was attributed

PAGE 115

102 by various strain energy releasing mechanisms such as deformation of nanofibers and a single domain structure of compos ing grains. The calculated critical size of 9.2 nm is a third of the experimentally observed crysta llite size in this dissertation. However, crystallite size of tetragonal BaTiO3 is still smaller than most of previous reports (Table 3-1) and crystallite size is reliable si nce it was cro ss-checked by various methods (Scherrer’s equation, TEM, and Williamson-Hall plot). Numerous reports have been carried out to explain the size effect and critical crystallite size of phase transiti on. However, most studies di d not consider all factors or even ended up with ambiguous guesses that render ed the explanations in complete. In this research, phase stabilization mechanism under the effect of a complex state was presented for the first time. It is concluded here that critical crystall ite sizes that had been suggested by numerous reports need further confirmation since most of them consider only one mechanism or without proving each mechanism numerically. Furthermore, it is proved in this dissertation that depending on th e strain energy releas ed and the amount of free energy, more scaling down of BaTiO3 can be achieved.

PAGE 116

103 CHAPTER 7 CONCLUSIONS 7.1 Synthesis of BaTiO3 Nanofibers In this dissertation, barium titanate (BaTiO3) nanofibers were synthesized via electrospinning for the first time ever. It was proved that combined with conventional sol-gel process, electrospinni ng can provide a straightforw ard way to synthesize a 1D nanostructure of a complex ceramic oxide. Th e characteristics of perovskite structure evolution and tetragonal pha se stabilizing mechanisms of electrospun BaTiO3 nanofibers were investigated and the follo wing conclusions are drawn: 1. BaTiO3 nanofibers were synthesized via el ectrospinning and a bout 60% of the diameter reduction was observed after heat tr eatment. Diameter reduction occurs when the polymer and organic molecules burnout. The average diameter after heat treatment was around 120 nm. Continuous fiber structur es up to a few hundred micrometers long were sustained after heat treatment and nanof ibers became polycrystalline made of grains between 25 and 50 nm. 2. A TG/DTA study showed crystallized BaTiO3 formed through reactions that can be attributed to intermediate phase formation and decomposition. An exothermic reaction at around 746 oC was due to the perovskite phase formation. Observation from XRD patterns supports formation of interm ediate phases during crystallization. Unknown phases were observed from the samples heat treated up to 600 oC. These disappeared at temp eratures above 650 oC and all patterns matched with perovskite BaTiO3. Well-defined perovskite peaks with higher intensities and no detectable

PAGE 117

104 secondary phases were observed from 750 oC 16 hour heat treatment. Information obtained from the TG/DTA study was applied to set up heat treatment conditions. Time dependant perovskite phase evolution was conduc ted. The onset of perovskite formation was observed at a lower temperature compar ed to the short heat treatment time (650 oC for 24 h, and 580 oC for 48 h heat). Lattice distortion (c/a) was investigated from the split of {200} and {110} families which are charact eristic for a tetragonal structure. The c/a ratio of electrospun nanofibers ranged betw een 1.003 and 1.009. An increasing tendency of c/a ratio was observed with increasing temp erature and time. The distortion was small compared to the theoretical value (c/a = 1.01) , however it was in agreement with previous reports. The size of individual crys tallite composing tetragonal BaTiO3 nanofibers ranged from ~ 30 nm (Scherrer’s formula). The size of the tetragonal crystallite measured by Williamson-Hall plot was 37 nm while TEM study showed 25~50 nm size distribution. The perovskite structure of nanofibers was further confirmed by selected area electron diffraction patterns taken for individual grains of nanofibers. 3. One of the key controve rsies of electrospinning is the potential of single crystalline nanostructure synthe sis. There has been a report on the formation of single crystal fibers via electrospinning, however, it involves a binary oxide system (WO3) [152]. In this research, BaTiO3 single crystalline nanofibers with 40 nm in diameter and 1m in length were reported for the first tim e. The availability of synthesizing single crystalline complex oxide (ternary) nanofibers will enhance the compatibility of electrospinning for various applications. 4. The stabilization of the high temperatur e hexagonal phase suggested by previous reports was verified. Raman bands at 640 and 1059 cm-1, which were assigned to the

PAGE 118

105 hexagonal structure BaTiO3 were observed. However, a carbonate band at 1059 cm-1 that always forms with hexagonal BaTiO3 disappeared at 750 oC 1 hour heat treatment while the 640 cm-1 remained. This corresponds we ll with previously showed BaTiO3 with planar defects. It is suggested that the orig in of the defects stem from the defective and disordered structure. The XPS spectra obtai ned from fibers subject ed to various heat treatment conditions revealed Ti (3+ and 4+) and Ba ions with differe nt oxidation statuses. The defective structure was introduced dur ing the random substitution of cations. Annealing at proper condition recovered a char ge balance but left planar defects. Since the structural character izations conducted by different techniques did not determine any evidence of hexagonal BaTiO3, it can be concluded here that electrospun nanofibers are tetragonal with extended planar defects. XRD pa tterns with no peak of hexagonal phase BaTiO3 further supports this. 5. The BaTiO3 crystallization reaction was investig ated in this research. The solid state reaction was excluded based on XRD, Raman, and XPS st udies. The reaction between hydroxyl titanyl acylate resulted in barium titanium oxycarbonate followed by nucleation and growth of BaTiO3. 6. Crystallite size of the tetragonal phas e is defined by strain energy and surface free energy. Depolarization ener gy, which can lead to BaTiO3 becoming cubic, was excluded for consideration since tetragonality was proved and that strongly suggested a minimum effect from depolari zation energy. The surface energy and strain energy of 30 nm crystallites composing electrospun nanofiber s were calculated. The surface energy of volume work (cubic tetragonal) and gas-solid surface energy of tetragonal BaTiO3 contributed to the overall surface free ener gy. The surface energy of the volume work

PAGE 119

106 was -0.00619 J/m2 while the gas-solid surface free energy of tetragonal BaTiO3 was between 0.00619 < g-s tetrag < 0.01 J/m2. The minimum probable strain energy was 8.49 x 105 J/m3 and this is two orders of magnitude smaller than the reported value. A significant portion of strain energy may be re leased by deformation and crack formation. Small misfits also contributed to lower strain energy in crystallites composing electrospun BaTiO3 nanofibers. Strained state of BaTiO3 crystallites (0.2255 %) was proved by Williamson-Hall plot and supported strain energy release. The minimum stored strain energy in crys tallite composing electrospun nanofibers was revealed as 1.200 x 10-18 J. Theoretically probable critical size of tetragonal to cubic structure was 9.29 nm. 7.2 Future Work This research mainly focused on the st ructural characteri zation of electrospun BaTiO3 nanofibers, particularly tetragonal pha se stabilizing mechanisms. However, many questions still remain as to the ferroel ectric properties of na nofibers, the single crystalline fiber growth mechanism, and the effects of an electric field during fiber synthesis. Thus, further st udies should be carried out. 7.2.1 Ferroelectric Characterization Direct measurement of ferroelectric prope rties from electrospun nanofibers is still challenging due to the difficulties in locating electrostatic force microscopy (EFM) on the nanofiber surface. Even though this disserta tion showed the tetragonal structure of BaTiO3 at 30 nm (error 20 %), its ferroelectr icity will remain controversial without further electrical characterization. Neve rtheless, the research conducted in this dissertation suggests that if proper measurement conditions are achieved, ferroelectric

PAGE 120

107 characterization of single nanof ibers (not bulk fibers or fiber mats) can be accomplished since the tetragonal BaTiO3 may show spontaneous polar ization and its reversal. 7.2.2 Single Crystalline Electro spun Nanofibers Synthesis This dissertation presented single crys talline of BaTiO3 nanofibers via the electrospinning process for the first time. Single crystal na nofiber growth was intentionally conducted through ad justment of heat treatment conditions. Precise control of heating rate (1~3 oC/min) implies a possibility to regulate single cr ystal nanofiber growth through temperature alone by limited nucleation and growth. Further studies on both thermodynamics a nd growth condition optimization should be carried out. If the grow th mechanism of electrospun sing le crystalline fiber is known thoroughly, this process may be a breakth rough in the nano-devices realization. 7.2.3 Effect of Electric Fi eld on Fiber Properties In this dissertation, it was revealed that low strain energy is a predominant factor for tetragonal phase stabilization at room temp erature. Low strain energy together with various strain releasing mechanisms were sugge sted for low stored strain energy. As an extension of this work, it seems feasible to i nvestigate the effects of an electric field for electrospun nanofiber char acteristics, such as effect of charge on phase evolution. This may enhance understanding of not only ferroel ectric nanofibers but of synthesis of electrospun nanofibers for vari ous applications as well.

PAGE 121

108 LIST OF REFERENCES 1. A.I. Kington, J. Maria, and S.K. Stre iffer: Alternative Dielectrics to Silicon Dioxide for Memory and Logic Devices. Nature. 406, 1032 (2000). 2. Z. Yao, H.W.C. Postma, L. Bale nts, and C. Dekker: Carbon Nanotube Intramolecular Junctions. Nature. 402, 273 (1999). 3. Y. Cui, and C.M. Lieber: Functiona l Nanoscale Electronic Devices Assembled Using Silicon Nanowire Building Blocks. Science. 291, 851 (2001). 4. G.E. Moore: Electronics Magazine. 38, 114 (1965). 5. P.S. Peercy: The Drive to Miniaturization. Nature. 406, 1023 (2000). 6. N. Kim, P. Ko, Y. Seo, and W. L ee: Chemical Mechanical Polishing of BTO Thin Film for Vertical Sidewall Patter ning of High-density Memory Capacitor. Thin Solid Films. 504, 261 (2006). 7. Y. Seo, and W.S. Lee: Chem ical Mechanical Polishing of Ba0.6Sr0.4TiO3 Film Prepared by Sol–gel Method. Microelectron. Eng. 75, 149 (2004). 8. P. Vitanov, A. Harizanova, T.Ivanova, D. Velkov, and Z. Raytcheva: Deposition, Structure Evolution and Diel ectric Properties of BaTiO3 and BaxSr1 xTiO3 Thin Films Prepared by the Sol–gel Method. Vacuum. 69, 371 (2001). 9. S. Iijima: Nature Helical Microtubules of Graphitic. Carbon. 354, 56 (1991). 10. K.S. Shankar, A.K. Raychauhuri: Fabr ication of Nanowires of Multicomponent Oxides: Review of Recent Advances. Mater. Sci. Eng. C. 25, 738 (2005). 11. P.G. Collins, A. Zettl, H. Bando, A. Thess, and R.E. Smalley: Nanotube Nanodevice. Science. 278, 100 (1997). 12. H.R. Shea, R. Martel, H. Hertel, T. Schmidt, and P. Avouris: Manipulation of Carbon Nanotubes and Properties of Nanotube Field-effect Transistors and Rings. Microelectron. Eng. 46, 101, (1999). 13. P. Avouris: Carbon nanotube electronics. Chem. Phys. 281, 429 (2002). 14.P. Avouris, R. Martel, V. Derycke, and J. Appenzeller: Carbon Nanotube Transistors and Logic Circuits. Physica B 323, 6 (2002). 15. W.J. Merz: The Dielect ric Behavior of BaTiO3 Single-Domain. Crystals, Phys. Rev. 75, 687 (1949).

PAGE 122

109 16. W.J. Merz: The Electric a nd Optical Behavior of BaTiO3 SingleDomain.Crystals, Phys. Rev. 76, 1221 (1949). 17. M. Yashima, T. Hoshina, D. Ishimura, S. Kobayasho, W. Nakamura, T. Tsurumi, and S. Wada: Size Effect on the Crystal structure of Barium Titanate Nanoparticles. J. Appl. Phys. 98. 014313 (2005). 18. H.A. Suer, and J.R. Fisher: Proce ssing of Positive Temperature Coefficient Thermistors. J. Am. Ceram. Soc. 43, 297 (1960). 19. W. Heywang: Semiconducting Barium Titanate. J. Mater. Sci. 6, 1214 (1971). 20. S. Schalg, H.F. Eicke: Size Driven Phase Transition in Na nocrystalline BaTiO3. Solid State Comm. 91, 883 (1994). 21. M.E. Lines, A.M. Glass: Principles and Applications of Ferroelectrics and Related Materials. (Clare ndon Press, Oxford, 1977). 22. F. Cardarelli: Materials Handboo k. (Springer, New York, 2001). 23. S. O’Brien, L. Brus, and C. B. Murray: Synthesis of Monodisperse Nanoparticles of Barium Titanate: Towa rd a Generalized Strategy of Oxide Nanoparticle Synthesis. J. Am. Chem. Soc. 123, 12085 (2001). 24. C. Pithan, D. Hennings, and R. Wa ser: Progress in the Synthesis of Nanocrystalline BaTiO3 Powders for MLCC. Int. J. Appl. Ceram. Technol. 2, 1 (2005). 25. Y.J. Song, H.J. Joo, S.K. Kang, H.H. Kim, J.H. Park, Y.M. Kang, E.Y. Kang, S.Y. Lee, and K. Kim: Electrical Proper ties of Highly Reliable 32 Mb FRAM with Advanced Capacitor Technology. Microelectron. Reilab. 45,1150 (2004). 26. T. Takeuchi, M. Tabuchi, K. Ado, K. Honjo, O. Nakamura, H. Kageyama, Y. Sutama, N.Ohtori, M. Nagasawa: Grain Size Dependence of Di electric Properties of Ultrafine BaTiO3 Prepared by a Sol–crystal Method. J. Mater. Sci. 32, 4053 (1997). 27. R. Bacsa, P Ravindranathan, a nd J.P. Dougherty: Electrochemical, Hydrothermal, and Electroc hemical–hydrothermal Synthe sis of Barium Titanate Thin Films on Titanium Substrates. J. Mater. Res. 7, 423 (1992). 28. E. Shi, C.R. Cho, M.S. Jang, S.Y. Jeong, and H.J. Kim: The Formation Mechanism of Barium Titanate Thin F ilm under Hydrothermal Conditions. J. Mater. Res. 9, 2914 (1994).

PAGE 123

110 29. A.T. Chien, J.S. Speck, F.F. Lange, A.C. Daykin, and C.G. Levi: Low Temperature/Low Pressure Hydrothermal Synthesis of Barium Titanate: Powder and Heteroepitaxial Thin Films. J. Mater. Res. 10, 1784 (1995). 30. C.R. Cho, E. Shi, M.S. Jang, S.Y. Je ong, and S.C Kim: Structural and Electrical Properties of BaTiO3 Thin Films on Si(100) Substrat e by Hydrothermal Synthesis. Jpn. J. Appl. Phys. 33, 4984 (1994). 31. A.J. Moulson, and J.M. Herbert: El ectroceramics. (Wiley and Sons Ltd, West Sussex 2003). 32. H.P. Beck, W. Eiser, and R. Haberkorn: Pitfalls in the Synthesis of Nanoscaled Perovskite Type Compounds. Part I: Influence of Diffe rent Sol–gel Preparation Methods and Characteriza tion of Nanoscaled BaTiO3. J. Eur. Ceram. Soc. 21, 687 (2001). 33. U Hwang, H. Park, and K. Koo: Low-Temperature Synthesis of Fully Crystallized Spherical BaTiO3 Particles by the Gel–Sol Method. J. Am. Ceram. Soc. 87, 2168 (2004). 34. K. Saegusa, W.E. Rhines, and H.K.Bo wen: Effect of Co mposition and Size of Crystallite on Crystal Phase in Lead Barium Titanate. J. Am. Ceram. Soc. 76, 1505 (1993). 35. S. Naka, F. Nakakita, Y. Suwa, and M. Inagaki: Change from Metastable Cubic to Stable Tetragonal Form of Submicron Barium Titanate. Bull. Chem. Soc. Jpn. 47, 1168 (1974). 36. T. Fujita, and I.J. Lin: Dielectric Fluid Preparation by Dispersing Ultrafine Barium Titanate Particles in Kerosene. Powder Technol. 68, 235 (1991). 37. K. Okazaki, and H. Maiwa: Space Charge Effects on Ferroelectric Ceramic Particle Surfaces. Jpn. J. Appl. Phys. 31, 3113 (1992). 38. K. Uchino, E. Sadanaga, and T. Hiro se: Dependence of the Crystal Structure on Particle Size in Barium Titanate. J. Am. Ceram. Soc. 72, 1555 (1989). 39. N.G. Eror, T.M. Loehr, B.C. Co rnilsen: Low Temperature Hexagonal BaTiO3 Polymorph and Carbon Adsorption. Ferroelectrics. 28, 321 (1980). 40. R.Vivekanandan,T.R.N. Kutty: Characterization of Barium Ttitanate Fine Powders Formed from Hydrothermal Crystallization, Powder Technol. 57, 181 (1989). 41. A. Herczog: Microcrystalline BaTiO3 by Crystallization from Glass. J. Am. Ceram. Soc. 47, 107 (1964).

PAGE 124

111 42. C.N.R. Rao, A. Govindaraj, G. G undiah and S.RC. Vivekchand: Nanotubes and Nanowires. Chem. Eng. Sci. 59, 4665 (2004). 43. P. Kim, and C.M. Lieber: Nanot ube Nanotweezers. Science. 286, 2148 (1999). 44. W. Sigmund, J. Yuh, H. Park, V. Man eeratana, G. Pyrgiotakis, A. Daga, J. Taylor, and J.C. Nino: Pro cessing and Structure Relations hips in Electrospinning of Ceramic Fiber Systems. J. Am. Ceram. Soc. 89, 395 (2006). 45. A. Formahals: Process and Apparatus for Preparing Artificial Threads. U.S. Patent 1975504, (1934). 46. D.H. Reneker, and I Chun: Nanometre Diameter Fibres of Polymer. Produced by Electrospinning, Nanotechnology. 7, 216 (1996). 47. T. Subbiah, G.S. Bhat, R.W. Tock S. Pararneswaran, and S.S. Ramkumar: Electrospinning of Nanofibers. J. Appl. Polym. Sci. 96, 557 (2005). 48. Y. Xia, P. Yang, Y. Sun, Y. Wu, B. Ma yer, B. Gates, Y. Yin, F. Kim, and H. Yan: One-Dimensional Nanostructures: Synthesis, Characterization, and Applications. Adv. Mater. 15, 353 (2003). 49. J.T. Sanders, R.M.V. Alwin, and C. Dekker: Room-temperature Transistor Based on a Single Carbon Nanotubes. Nature. 393, 49 (1998). 50. R.H. Baughman, A.A. Zakhidov, and W.A. de Heer: Carbon Nanotubes--the Route Toward Applications. Science. 297, 787 (2002). 51. X. Duan, Y. Huang, and C. Lieb er: Nonvolatile Memory and Programmable Logic from Molecule-Gated Nanowires. Nano Lett. 2, 487 (2002). 52. C. Thelander, H.A. Nilsson, L.E. Jensen, and L. Samuelson: Nanowire SingleElectron Memory. Nano Lett. 5, 635 (2005). 53. K. Tsukagoshi, N. Yoneya, S. Uryu, Y. Aoyagi, A. Kanda, Y. Ootuka, and B.W. Alphenaar: Carbon Nanotube Devices for Nanoelectronics. Physica B. 323, 107 (2002). 54. L. Roschier, R. Tarkiainen, M. Ahlskog, M. Paalanen, and P. Hakonen: Manufacture of Single Electron Tran sistors Using AFM Manipulation on Multiwalled Carbon Nanotubes. Microelec. Eng. 61, 687 (2002). 55. Y.V. Nastushev, T. Ga vrilova, M. Kachanova, L. Nenasheva, V. Kolosanov, O.V. Naumova, V.P. Popov, A.L. As eev: 20-nm Resolution of Electron Lithography for the Nano-devices on Ultr athin SOI Film. Mater. Sci. Eng. C. 19, 189 (2002).

PAGE 125

112 56. V. Derycke, R. Martel, J. Appen zeller, and Ph. Avouris: Carbon Nanotube Interand Intramolecular Logic Gates.Nano Lett. 1, 453 (2001). 57 C.N.R. Rao, F.L. Deepak, G. Gundiah, and A. Govindaraj: Inorganic Nanowires. Prog. Sol. State Chem. 31, 5 (2003). 58. T. Maruyama, M. Saitoh, I. Sakai, T. Hidaka, Y. Yano, and T. Noguchi: Growth and Characterization of 10-nm -thick c-axis Oriented Epitaxial PbZr0.25Ti0.75O3 Thin Films on (100) Si S ubstrate. Appl. Phys. Lett. 73, 3524 (1998). 59. T.J. Trentler, K.M. Hickman, S.C. Geol, A.M. Viano, P.C. Gibbons, and W.E. Buhro: Solution-Liquid-Solid Growth of Crystalline III-V Semiconductors: An Analogy to Vapor-Liquid-Solid Growth. Science. 270, 1791 (1995). 60. C.R. Martin: Nanomaterials: A Membra ne-Based Synthetic Approach. Science. 266, 1961 (1994). 61. W.S. Yun, J.J. Urban, Q. Gu, and H. Park: Ferroelectric Prope rties of Individual Barium Titanate Nanowires Investigated by Scanned Probe Microscopy. Nano Lett. 2, 447 (2002). 62. Y. Luo, I. Szafraniak, N.D. Zakhar ov, V. Nagarajan, M. Steinhart, R.B. Wehrspohn, J.H. Wendorff, R. Ramesh, and M. Alexe: Nanoshell Tubes of Ferroelectric Lead Zirconate Titanate a nd Barium Titanate. Appl. Phys. Lett. 83, 440 (2003). 63. A.M. Morales, and C.M. Lieber: A La ser Ablation Method for the Synthesis of Crystalline Semiconductor Nanowires. Science. 279, 208 (1998). 64. D. Li, and Y. Xia: Direct Fabric ation of Composite and Ceramic Hollow Nanofibers by Electrospinning. Nano Lett. 4, 933 (2004). 65. G.I. Taylor: Electrically Driven Jets. Proc. R. Soc. London, Ser A 258, 383 (1964). 66. M.M. Hohman, M. Shin, G. Rutledge, and M.P. Brenner: Electrospinning and Electrically Forced Jets. I. St ability Theory. Phys. Fluids. 13, 2201 (2001) 67. M.M. Hohman, M. Shin, G. Rutledge, and M.P. Brenner: Electrospinning and Electrically Forced Jets. II. Applications. Phys. Fluids. 13, 2221 (2001). 68. G. Zhang, W. Kataphinan, R. Teye-Men sah, P. Katta, L. Khatri, E.A. Evans, G.G. Chase, R.D. Ramsier, and D.H. Reneker: Electrospun Nanofibers for Potential Space-based Applications. Mat. Sci. Eng. B-Solid .116, 353 (2005).

PAGE 126

113 69. G. Larsen, R. Velarde-Ortiz, K. Minc how, A. Barrero, and I.G. Loscertales: A Method for Making Inorganic and Hybrid (O rganic/Inorganic) Fibers and Vesicles with Diameters in the Submicrometer and Micrometer Range via Sol-Gel Chemistry and Electrically Forced Liquid Jets. J. Am. Chem. Soc. 125, 1154 (2003). 70. S. Madhugiri, B. Sun, P.G. Smirniotis, J.P. Ferraris, and K.J. Balkus: Electrospun Mesoporous Titanium Dioxide Fibers. Micropor. Mesopor. Mat. 69, 77 (2004). 71. X.G. Yang, C.L. Shao, Y.C. Liu, R. X. Mu, and H.Y. Guan: Nanofibers of CeO2 via an Electrospinning Technique. Thin Solid Films. 478, 228 (2005). 72. D. Li, and Y. Xia: Electrospinning of Polymeric and Ceramic Nanofibers as Uniaxially Aligned Arrays. Nano Lett. 3, 1167 (2003). 73. G.E. Wnek, M.E. Carr, D.G. Simps on, and G.L. Brown, Electrospinning of Nanofiber Fibrinogen Structures. Nano Lett. 3, 213 (2003). 74. J. Yuh, W. Sigmund, and J.C. Nino: Synthesis of Barium Titanate Nanofibers via Electrospinning. Mater. Lett. 59, 3645 (2005). 75. Y. Wang and J.J. Santiago-Aviles: Synthesis of Lead Zirconate Titanate Nanofibres and the Fourier-Transform Infr ared Characterizati on of their Metalloorganic Decomposition Pr ocess. Nanotechnology. 15, 32 (2004). 76. N. Dharmaraj, H.C. Park, B.M. Lee, P. Viswanathamurthi, H.Y. Kim, and D.R. Lee: Preparation and Morphology of Magnesium Titanate Nanofibres via Electrospinning. Inorg. Chem. Commun. 7, 431 (2004). 77. S. Maensiri, W. Nuansing, J. Klinkaewnarong, P. Laokul, J. Khemprasit: Nanofibers of Barium Strontium Tita nate (BST) by Sol–gel Processing and Electrospinning. J. collid.Sci. Inter. 297, 578 (2006). 78. N. Dharmaraj, H.C. Park, C.K. Kim, H. Y. Kim, and D.R. Lee: Nickel Titanate Nanofibers by Electrospinni ng. Mater Chem Phys. 87, 5-9 (2004). 79. D. Li, T. Herricks, and Y.N. Xia: Magnetic Nanofibers of Nickel Ferrite Prepared by Electrospinning. Appl. Phys. Lett. 83, 4586 (2003). 80. H. Park: Ph.D. Thesis, Fabrication of Lanthanum Copper Oxide Nanofibers by Electrospinning. University of Florida, Gainesville. 2005. 81. Y.M. Shin, M.M. Hohman, M.P. Brenner, and G.C. Rutledge: Electrospinning: A Whipping Fluid Jet generates Submicr on Polymer Fibers. Appl. Phys. Lett. 78, 1149 (2001).

PAGE 127

114 82. D.H. Reneker, A.L. Yarin, H. F ong, and S. Koombhongse: Bending Instability of Electrically Charged Liquid jets of Po lymer Solutions in Electrospinning. J. Appl. Phys. 87, 4531 (2000). 83. Y.M. Shin, M.M. Hohman, M.P. Brenner, and G.C. Rutledge: Experimental Characterization of Electrospinning: the El ectrically Forced Jet and Instabilities. Polymer. 42, 9955 (2001). 84. S.V. Fridrikh, J.H. Yu, M.P. Brenne r, and G.C. Rutledge: Controlling the Fiber Diameter during Electrospi nning. Phys. Rev. Lett. 90, 144502 (2003). 85. F.S. Galasso: Structure, Propertie s and Preparation of Pervoskite-type Compounds. (Pergamon Press, Oxford, 1969). 86. R.W. Schwartz: Chemical Solution Depos ition of Perovskite Thin Films. Chem. Mater. 9, 2325 (1997). 87. S.O. Kassap: Principles of Electro nic Materials and Devices. 3rd edition (McGraw-Hill, New York, 2006). 88. M.A.A. Issa, A.M. Hassib, and Z.H. Dughaish: A Study of the BaTiO3:CeO2 Crystal Structure by EPR. J. Phys. D: Appl. Phys. 17, 2037 (1984). 89. J.J.Urban, J.E. Spanier, L. Ouyang, W.S. Yoon, and H. Park: Single-Crystalline Barium Titanate Nanowires. Adv. Mater. 15, 423 (2003). 90. J. Hu, T.W. Odom, and C.M. Lieber: Chemistry and Physics in One Dimension: Synthesis and Properties of Nanowires and Nanotubes. Acc. Chem. Res. 32, 435 (1999). 91. A.M. Morales, and C.M. Lieber: A La ser Ablation Method for the Synthesis of Crystalline Semiconductor Nanowires. Science. 279, 208 (1998). 92. Y. Cui, and C.M. Lieber: Functional Nanoscale Electronic Devices Assembled Using Silicon Nanowire Building Blocks. Science. 291, 851, (2001). 93. X. Duan, Y. Huang, Y. Cui, J. Wang, C.M. Lieber: Indium Phosphide Nanowires as Building Blocks for Na noscale Electronic and Optoelectronic Devices. Nature. 409, 66 (2001). 94. M.H. Huang, S. Mao, H. Feick, H.Q. Yan, Y.Y. Wu, H. Kind, E. Webber, R. Russo, and P.D. Yang: Room-Temperature Ultraviolet Nanowire Nanolasers. Science. 292, 1897 (2001). 95. G. Arlt, D. Hennings, and G. de With : Dielectric Properties of Fine-grained Barium Titanate Ceramics. J. Appl. Phys. 58, 1619 (1985).

PAGE 128

115 96. D. Hennings: Barium Tita nate Based Ceramic Materials for Dielectric Use. Int. J. High Technol. Ceram. 3, 91 (1987). 97. K. Fukai, K. Hidaka, M. Aoki, and K. Abe: Preparation and Properties of Uniform Fine Perovskite Powders by H ydrothermal Synthesis. Ceram. Int. 16, 285 (1990). 98. C. Jaccard, W. Knzig, and M. Peter: Das Verhaken Von Kolloidalen Seignetteelektrika I, KH2PO4. Helv. Phys. Acta. 26, 521 (1953). 99. W. Knzig: R ntgenunter Suchungen ber die Seigenetteelek Trizitt von Barium Titanat. Helv. Phys. Acta. 24, 175 (1951). 100. M. Anliker, H.R. Brugger, and W. Knzig: Das Verhalten von Kolloidalea Seigenetteelek Trika, BaTiO3. Helv. Phys. Acta. 27, 99 (1954). 101. R. Vivekanandan, and T.R.N. Kutty: Hydrothermal Synthesis of Ba(Ti, Sn)O3 Fine Powders and Dielectric Properties of the Corresponding Ceramics. Ceram. Int. 14, 207 (1988). 102. B.D. Begg, E.R. Vance, and J. Nowot ny: Effect of Particle Size on the RoomTemperature Crystal Structure of Barium Titanate. J.Am Ceram. Soc. 77, 3186 (1994). 103. F.S. Yen, H. Hsiang, and Y. Chang: Cubic to Tetragonal Phase Transformation of Ultrafine BaTiO3 Crystallites at Room Te mperature. Jpn. J. Appl. Phys. 34, 6149 (1995). 104. K. Kiss, J. Manger, M.S. Vukasovich , and R.J. Lockhart: Ferroelectrics of Ultrafine Particle Size: I, Synthesis of Titanate Powders of Ultrafine Particle Size. J. Am. Ceram. Soc. 49, 291 (1966). 105. M. Kuwabara, K. Miki, S. Takahash i, and H. Shimooka: Preparation of Monolithic Barium Titanate Xerogels by Sol–gel Processing and the Dielectric Properties of their Sintered Bodies. Appl. Phys. Lett. 66, 1704 (1995). 106. F.S. Yen, C.T. Chang, and Y.H. Ch ang: Characterization of Barium Titanyl Oxalate Tetrahydrate. J. Am. Ceram. Soc. 73, 3422 (1990). 107. D.D. Wagman, W.H. Evans, V.B.Pa rker, R.H. Schumm, I. Halow, S.M. Bailey, K.L. Churney, and R.L. Nuttal: The NBS Tables of Chemical Thermodynamic Properties, Selected Values for Inorganic and C1 and C2 Organic Substances in SI Units. J. Phys. Chem. Ref. Data, 2 Suppl. 2(1982). 108. I. Barin, O. Knacke, and O. Ku baschewski: Thermodynamic Properties of Inorganic Substances in SI Units. (Verlag Stahleisen, Berlin, FRG 1977).

PAGE 129

116 109. I. Barin: Thermodynamic Data of Pure Substances. (VCH, Weinheim, FRG, 1989). 110. T. Mitsui, and S. Nomura: Ferroele ctrics and Related Substances: Oxides. (Springer-Verlag, Berlin, FRG,1981). 111. T. Takeushi, K. Ado, T. Asai, H. Ka geyama, Y. Saito, C. Masquelier, and O. Nakamura: Thickness of Cubic Phase on Bari um Titanate Single-Crystalline Grains. J. Am. Ceram. Soc. 77, 1665 (1994). 112. A.S. Shaikh, R.W.Vest, and G.M. Ves t: Dielectric Proper ties of Ultrafine Grained BaTiO3. IEEE Trans. Ultrason. Ferroelectr. Freq. Control 36, 407 (1989). 113. K. Fukai, K. Hidaka, M. Aoki, and K. Abe: Preparation and Properties of Uniform Fine Perovskite Powders by H hydrothermal Synthesis. Ceram. Int. 16, 285 (1990). 114. R. Vivekanandan, S. Philip, and T.R.N. Kutty: Hydrothermal Preparation of Ba(Ti, Zr)O3 Fine Powders. Mater. Res. Bull. 22, 99 (1987). 115. D. Hennings, and S. Schreinemach er: Characterization of Hydrothermal Barium Titanate. J. Eur. Ceram. Soc. 9, 41 (1992). 116. J. Selsing: Internal Stresses in Ceramics. J. Am. Ceram. Soc. 44, 419 (1961). 117. R.W. Davidge, and T.J. Green: The Strength of Two-phase Ceramic/Glass Materials. J. Mater. Sci. 3, 629 (1968). 118. R.C. Garvie: Stabilizati on of the Tetragonal Structur e in Zirconia Micrcrystals. J.Phys. Chem. 82, 218 (1978). 119. J.P. Gramatico, and J.M.P. Lopez: Reaction Sequences in the Systems Ti(OiPr)4 + Ba(CH3COO)2 and TiO2 + Ba(CH3COO)2. J. Mater. Sci. Mater. Electron. 3, 82 (1992). 120. M. Drofenik, A. Popovic, L. Irmancnik, D. Kolar, and V. Krasevec: Release of Oxygen During the Sintering of Doped BaTiO3 Ceramics. J. Am. Ceram. Soc. 65, C-203 (1982). 121. R. Waser: Solubility of Hydrog en Defects in Doped and Undoped BaTiO3. J. Am. Ceram. Soc. 71, 58 (1988). 122. L.K. Templeton, and J.A. Pask: Formation of BaTiO3 from BaCO3 and TiO2 in Air and in CO2. J. Am. Ceram. Soc. 42, 212 (1959).

PAGE 130

117 123. W. Luan, and L. Gao: Influence of pH Value on Properties of Nanocrystalline BaTiO3 Powder. Ceram. Int. 27, 645 (2001). 124. H. Shimooka, and M. Kuwabura: Preparation of Dense BaTiO3 Ceramics from Sol-Gel-Derived Monolithic Ge ls. J. Am. Ceram. Soc. 78, 2849 (1995) 125. B. Li, X. Wang, and L. Li: Synthe sis and Sintering Behavior of BaTiO3 Prepared by Different Chemical Methods. Mater. Chem. Phys. 78, 292 (2002). 126. J. Yang, J.M.F. Ferreira, W. Weng, and Y. Tang: Materials Lett. Synthesis, Gelation Behaviour and Evolution of O ligomeric Structure of BaTi Double Alkoxide, Materials Lett. 42, 257 (2000). 127. H.B. Sharma, and A. Mansingh: Sol-gel Processed Barium Titanate Ceramics and Thin Films. J. Mater. Sci. 33, 4455 (1998). 128. S. van der Gijp, M.H.J. Edmond, A.J.A. Winnubst, and H. Verweij: Preparation of BaTiO3 by Homogeneous Precipitation. J. Eur. Ceram. Soc. 19, 1683 (1990). 129. H.Q. Dai, J. Gong, H. Kim, and D. Lee: A Novel Method for Preparing Ultrafine Alumina-borate Oxide Fibres via an Electrospinning Technique. Nanotechnology. 13, 674 (2002). 130. P. Viswanathamurthi, N. Bhattarai, H.Y. Kim, D.I. Cha, and D.R. Lee: Preparation and Morphology of Palladium Oxide Fibers via Electrospinning. Materials Lett. 58, 3368 (2004). 131. P. Durn, D. Gutierrez, J. Tartaj, M.A. Baares, and C. Moure: On the Formation of an Oxycarbonate Intermedia te Phase in the Yynthesis of BaTiO3 from (Ba,Ti)-polymeric Organic Precursors. J. Eur. Ceram. Soc. 22, 797 (2002). 132. M.H. Frey, and D.A. Payne: Synthesi s and Processing of Barium Titanate Ceramics from Alkoxide Solutions and Monolithic Gels. Chem. Mater. 7, 123 (1995). 133. X. Xing, J. Deng, J. Chen, and G. Li u: Phase Evolution of Barium Titanate from Alkoxide Gel-derived Precursor. Alloy and Comp. 384, 312 (2004). 134. P. Mukherjee, A. Sarkar, P. Barat, S.K. Bandyopadhyay, P. Sen, S.K. Chattopadhyay, P. Chatterjee, S.K. Chatterjee, and M.K. Mitra: Deformation Characteristics of Rolled Zirconium A lloys: a Study by X-ray Diffraction Line Profile Analysis. Acta Materialia. 52, 5687 (2004). 135. JCPDS Powder Diffraction File No 5-626. (1988).

PAGE 131

118 136. B.C. Frazer, H.R. Danner, and R. Pe pinsky: Single-Crysta l Neutron Analysis of Tetragonal BaTiO3. Phys. Rev.100, 745 (1955). 137. B.D. Cullity: Element of X-ray Di ffraction. (Addison-Wesley Publishing Company INC. Reading, 1978). 138. W. Cho: Structural Evoluti on and Characterization of BaTiO3 Nanoparticles Synthesized from Polymeric Pr ecursor. J. Phys. Chem. Solids 59, 659 (1998). 139. J. Javadpour, and N.G. Eror: Raman Spectroscopy of Higher Titanate Phases in the BaTiO3-TiO2 System. J. Am. Ceram. Soc. 71, 206 (1988). 140. T. Noma, S. Wada, M. Yano, and T. Suzuki: Analysis of Lattice Vibration in Fine Particles of Barium Titanate Si ngle Crystal Including the Lattice Hydroxyl Group. J. Appl. Phys. 80, 5223 (1996). 141. S. Kumar, G.L. Messing, and W. B. White: Metal Orga nic Resin Derived Barium Titanate: I, Formation of Barium Titanium Oxycarbonate Intermediate. J. Am. Ceram. Soc. 76, 617 (1993). 142. W. Cho, and E. Hamada: Synthesis of Ultrafine BaTiO3 Particles from Polymeric Precursor: Their Structure a nd Surface Property. J. Alloy Compd. 266, 118 (1998). 143. W. Cho, and E. Hamada: Planar Defects and Luminescence of BaTiO3 Particles Synthesized by a Polymerized Complex Method. J. Alloy Compd. 268, 78 (1998). 144. S.A. Nasser: X-ray Photoelectro n Spectroscopy Study on the Composition and Structure of BaTiO3 Thin Films Deposited on Silicon. Appl. Surf. Sci. 157, 14 (2000). 145. R. Ayouchi, F. Martin, J.R. Ramos-Barrado, and D. Leinen: Compositional, Structural and Electrical Char acterization of Barium Titanate Thin Films Prepared on Fused Silica and Si(111) by Spray Ppyrolysis. Surf. Interface Anal. 30, 565 (2000). 146. P.A.W. Heide, Q.D. Jiang, Y.S. Kim, and J.W. Rabalais: X-ray Photoelectron Spectroscopic and Ion Scattering Study of the SrTiO3(001) Surface. Surface Science. 473, 59 (2001). 147. K.E. Smith, and V.E Henrich: Interaction of SO2 and CO with the Ti2O3 Surface. Phys. Rev. B. 32, 5384 (1985).

PAGE 132

119 148. C. Miot, E. Husson, C. Proust, R. E rre, and J.P. Countres: X-ray Photoelectron Spectroscopy Characterization of Barium T itanate Ceramics Prepared by the Citric Route. Residual Carbon Study. J.Mater. Res. 12, 2388 (1997). 149. A.N. Morozovska, and E.A. Elis eev: Screening and Size Effects on the Nanodomain Tailoring in Ferroelectrics Semiconductors. Phys. Rev. B, 73 104440 (2006). 150. M.E. Lines, and A.M. Glass: Principl es and Application of Ferroelectrics and Related Materials. (Clarendon, Oxford 1977). 151. A.S. Barnard, S.P. Russo, and I.K. Snook: Size Dependent Phase Stability of Carbon nanoparticles: Nanodiamond vers us Fullerenes. J. Chem. Phys. 118, 5094 (2003). 152. K.M. Sawicka, A.K. Prasad, and I. Gouma: Metal Oxide Nanowires for Use in Chemical Sensing Appli cations. Sensor Lett. 3, 1 (2005).

PAGE 133

120 BIOGRAPHICAL SKETCH Junhan Yuh was born in Seoul, South Korea. In academic aspect, he earned his B.A. degree in Materials science and engineeri ng in Hanyang University. After earning B.A. degree, he joined Hanyang University for his M.S. degree in Materials Science and Engineering. He graduated from University of Florida in December 2006 with Ph. D.