This item is only available as the following downloads:
1 DEFECT ENGINEERING OF ELECTROCERAMICS: BISMUTH TRIIODIDE AND BARIUM TITANATE By HYUKSU HAN A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DE GREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2014
2 2014 HyukSu Han
3 To m y f uture s elf
4 ACKNOWLEDGMENTS At the very first, I sincerely appreciate my advisor Dr. Juan C. Nino for his support and g uidance. His knowledge, care, optimism, diligence, and aspirations inspired me to work hard and expand my potential. In addition, I am also deeply grateful to my other committee members (Dr. Kelly Jordan, Dr. James Baciak, Dr. Yong Yang, and Dr. Jason We aver) for their guidance. I would like to acknowledge all the collaborators who significantly improved my research : Dr. Susan B. Sinnott and Dr. Minki Hong for their great help on all the computational work Dr. Dipankar Ghosh, Dr. Jacob L. Jon e s Dr. Soph ie Guillemet Fritsch, and Dr. Christophe Voisin for their contributions on the synthesis of dielectric materials and analysis of dielectric properties. I also want to acknowledge Mr. Rudy Strohschein from Southern Scientific, Inc. for his glass ware work I would like to thank the people who were involved and helped me through the entire project, Dr. Valentin Craciun, Dr. Tanner, Dr. Burger, Dr. Holland Smith, and Evan Thatcher. I deeply thank the current and previous NRG group members: Dr. Wei Qui, Dr. Beverly Hinojosa, Dr. Wei Zhou, Roberto Esquivel Elizondo, Donald Moore, Luping Li, Robert Kasse, Chris Turner, Trey Davis, George Baure, Mehrad Mehr, Sasmit Gokhale, and etc., for providing me with an excellent research environment and being helpful. Last but not least, I want to give my special thanks to my parents and my friends (KeKis) for their loving encouragement. Their love ha s encouraged me to endure the hard time s I cannot imagine myself having any of my progress without their strong suppo rt behind me.
5 TABLE OF CONTENTS page ACKNOWLE DGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 8 LIST OF FIGURES ................................ ................................ ................................ .......... 9 LIST OF ABBREVIATIO NS ................................ ................................ ........................... 12 ABSTRACT ................................ ................................ ................................ ................... 13 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 15 1.1 Statement of Problem and Motivation ................................ ............................... 15 1.2 Scientific Approach ................................ ................................ ........................... 16 1.3 Organization of Dissertation ................................ ................................ .............. 19 1.4 Contributions to the Field ................................ ................................ .................. 20 2 BACKGROUND ................................ ................................ ................................ ...... 23 2.1 Defect Chemistry ................................ ................................ .............................. 23 2.1 1 The Krger Vink N otation ................................ ................................ ..... 23 2.1 2 Intrinsic P ont D efect R eactions in S toichiometric C ompounds: S chottky and Frenkel ................................ ................................ .................... 24 2.1 3 Diffusion and E lectrical C onductivity ................................ ........................ 30 2.2 Dielectric Polarizations in Solids ................................ ................................ ....... 34 2.2.1 Polarization M echanisms ................................ ................................ ......... 35 2.2.2 Space C harge P olarization ................................ ................................ ...... 38 126.96.36.199 Hopping polarization ................................ ................................ ...... 39 188.8.131.52 Interfacial polarization ................................ ................................ .... 40 2.2.3 The Universal D ielectric R esponse ................................ .......................... 41 2.3 Crystal Structures of BiI 3 and BaTiO 3 ................................ ............................... 45 2.3.1 BiI 3 ................................ ................................ ................................ ........... 45 2.3.2 BaTiO 3 ................................ ................................ ................................ ..... 47 3 EXPERIMENTAL PROCEDURES AND PROCESSING ................................ ......... 52 3.1 Sample Preparation ................................ ................................ .......................... 52 3.1.1 BiI 3 S ingle C rystals ................................ ................................ .................. 52 184.108.40.206 PVT powder synthesis technique ................................ ................... 52 220.127.116.11 TMZ purification technique ................................ ............................. 53 18.104.22.168 Single crystal growth of Te and Sb doped BiI 3 ............................... 54 3.1.2 BaTiO 3 C eramics ................................ ................................ ..................... 55
6 22.214.171.124 Microwave sintering ................................ ................................ ....... 55 126.96.36.199 Spark plasma sintering ................................ ................................ ... 55 188.8.131.52 Conventional pressureless sintering ................................ .............. 56 3.2 Characterization ................................ ................................ ................................ 57 3.2.1 X ray D iffraction ................................ ................................ ....................... 57 3.2.2 Scanning E lectron M icroscopy ................................ ................................ 57 3.2.3 X ray P hotoelectron S pectroscopy ................................ .......................... 58 3.2.4 Inductively C oupled P lasma A tomic E mission S pectroscopy ............... 58 3.2.5 Dielectric C haracterization ................................ ................................ ....... 58 3.2.6 Electrical P roperty C haracterization ................................ ........................ 59 3.2.7 Radiation R esponse M easurement ................................ ......................... 60 3.2.8 Computational M ethods ................................ ................................ ........... 60 4 ENHANCED ELECTRICAL PROPERTIES AND RADIATION RESPONSE OF ULTRA PURE BISMUTH TRI IODIDE SINGLE CRYSTAL DETECTORS ............. 61 4.1 Introduction ................................ ................................ ................................ ....... 61 4.2 Ultra High Pure BiI 3 Powder Synthesis by PVT and TMZ Techniques .............. 61 4.3 Electrical Properties and Radiation Response of Ultra Pure BiI 3 Single Crystal Detectors ................................ ................................ ................................ 62 4.4 Conclusion ................................ ................................ ................................ ........ 65 5 DEFECT MODELING AND ENGINEERING OF BISMUTH TRI IODIDE SINGLE CRYSTALS: ENHANCED ELECTRICAL AND RADIATION DETECTION PERFORMANCE ................................ ................................ ................................ .... 67 5.1 Introdu ction ................................ ................................ ................................ ....... 67 5.2 Defect Modeling of BiI 3 Single Crystals: Donor (Te) Doped BiI 3 ....................... 67 5.3 Defect Engineering of BiI 3 Single Crystals: S b doped BiI 3 Single Crystal .......... 72 5.4 Enhanced Electrical and Radiation Performance of Sb doped BiI 3 Single Crystal Detectors for Room Temperature Gamma Ray Detection ....................... 81 5.5 Conclusion ................................ ................................ ................................ ........ 86 6 COLOSSAL PERMITTIVITY IN FAST FIRED BARIUM TITANATE CERAMICS ... 88 6.1 I ntroduction ................................ ................................ ................................ ....... 88 6.2 Microstructure, Density, and Phase Purity of Pressureless, Spark Plasma, and Microwave Sintered BaTiO 3 ................................ ................................ ......... 89 6.3 Dielectric Properties of Pressureless, Spark Plasma, and Microwave Sintered BaTiO 3 ................................ ................................ ................................ ... 93 6.4 Effect of Annealing on Dielectric Properties of Microwave Sintered BaTiO 3 Ceramics ................................ ................................ ................................ ............. 97 6.5 Conclusion ................................ ................................ ................................ ...... 102 7 DIELECTRIC POLARIZAION MECHANISMS AND VARIABLE HOPPING CONDUCTION IN FAST FIRED BARIUM TITANATE CERAMICS ...................... 103 7.1 Introduction ................................ ................................ ................................ ..... 103
7 7.2 Microstructure, Density, and XRD of Spark Plasma Sintered BaTiO 3 Using Co Precipitated Nanocrystalline Powder ................................ ........................... 103 7.3 Broadband Dielectric Spectroscopy and Polarization Mechanism Investigation on Colossal Permittivity of Barium Titanate Ceramics .................. 106 7.4 Variable Range Hopping Conduction in Barium Titanate Ceramics Exhibiting Colossal Permittivity ................................ ................................ .......... 118 7. 5 Tailoring Contributions of Each Polarization Mechanism to Colossal Permittivity of Barium Titanate Ceramics ................................ ........................... 127 7. 6 Conclusion ................................ ................................ ................................ ...... 131 8 SUMMARY AND FU TURE WORK ................................ ................................ ....... 133 8.1 Summary ................................ ................................ ................................ ........ 133 8.1.1 Defects and E lectrical P roperty R elationships in BiI 3 ............................. 133 8.1.2 Colossal P ermittivity i n F ast f ired BaTiO 3 ................................ .............. 134 8.2 Future Work ................................ ................................ ................................ .... 135 8.2.1 Defect E ngineering of BiI 3 : C ontrolling A tmospheric P ressure .............. 135 8.2.2 Polarization M echanism in (Nb+In) C o d oped TiO 2 C eramics E xhibiting T emperature f requency I ndependent C olossal P ermittivity ........ 137 LIST OF REFERENCE ................................ ................................ ................................ 139 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 149
8 LIST OF TABLES Table page 2 1 Experimentally determined Schottky defect formation enthalpy in Alkali halide compounds. ................................ ................................ ................................ ........ 28 2 2 Examples of Frenkel defect formation enthalpy. ................................ ................. 30 2 3 Defect characterization tools. ................................ ................................ ............. 34 2 4 Examples for numerical values of the exponent "s" for different materials. ........ 4 5 2 5 Crystal structure information for BiI 3 ................................ ................................ .. 47 2 6 Crystal structure information for tetragonal BT. ................................ .................. 51 4 1 Impurity concentrations o f PVT and TMZ synthesized BiI 3 powder measured by using ICP AES. ................................ ................................ .............................. 62 5 1 Te 4+ substitutional donor dopant for Bi 3+ ................................ ............................ 67 5 2 Summar y of the dominant defect type and concentration for and the slope in the Log( T) vs. 1/T plot. ................................ ................................ ...................... 69 5 3 ICP AES analysis on (a) BiI 3 and (b) SBI single crystals. ................................ ... 75 5 4 Sb I and Bi I bond lengths in Sb incorporated BiI 3 on Bi site. ............................. 80 6 1 Summary of sintering conditions, densities, and dielectric properties. ................ 92 6 2 Dielectric properties and bulk resistivity of as sintered MWS, SPS, and PS BT ceramics at room temperature and 1 kHz. ................................ .................... 97 6 3 Effect of annealing on dielectric of MWS BT. ................................ .................... 101 6 4 Dielectric properties of the annealed MWS BT (950C for 12 hours) with various electrodes (Au, Ag, Ni, and Al) at room temperature and 1 kHz. ......... 101 7 1 Summary of chemical composition, grain size and relative density of the sintered ceramics. ................................ ................................ ............................ 105 7 2 Dielectric property of S PS BT 0.95 and 1.00 at room temperature and 1, 10, and 100 kHz. ................................ ................................ ................................ .... 108 7 3 model. ................................ ................................ ................................ ............... 129
9 LI ST OF FIGURES Figure page 2 1 Representative defects in MO oxide ................................ ................................ .. 26 2 2 Schematics of four main polarization mechani sms. ................................ ............ 37 2 3 Contributions of each polarization mechanism to the total relative permittivity as a function of frequency. ................................ ................................ ................. 38 2 4 H opping polarization due to the hopping of the charged particel over the barrier from one site to the other. ................................ ................................ ....... 39 2 5 A dielectric material comprising of two different components. ............................ 40 2 6 Examples of dielectric behaviors departed from the Debye theory. .................... 43 2 7 Schematic representations of the two types of dielectric respons e. ................... 44 2 8 Polyhedral schematic for corner sharing octahedra in BiI 3 and empty octahedral sites in one I Bi I layer. ................................ ................................ ..... 47 2 9 Cryst al structure of BT . ................................ ................................ ...................... 48 2 10 Phase transformation and relative permittivity change of BT ............................. 50 2 11 Approximate ion displacements in tetragonal BT. ................................ ............... 51 4 1 IV characteristics of the PVT and commercial BiI 3 detectors. ............................. 63 4 2 Leakage current comparison between PVT and commercial BiI 3 detectors. ...... 64 4 3 241 particle spectrum recorded from the ultra pure PVT BiI 3 detector at room temperature. ................................ ................................ .............................. 65 5 1 Anticipated slopes in Log( T) vs. Temp plot for donor doped BiI 3 ................... 70 5 2 Te doped BiI 3 single crystal. ................................ ................................ ............... 70 5 3 ICP analysis for Te concentrations along the distance from the tip. ................... 71 5 4 Ionic conductivity of Te doped BiI 3 single crystals as a function of temperature. ................................ ................................ ................................ ....... 72 5 5 XRD patterns for BiI 3 and SBI single crystals. ................................ .................... 74 5 6 Calculated formation energies as a function of Fermi energy. ............................ 77
10 5 7 Crystal structures of SBI and total density of states (DOS) for SBI and pure BiI 3 ................................ ................................ ................................ ..................... 79 5 8 Electrical properties of BiI 3 and SBI single crystal d etectors. .............................. 82 5 9 241 Am alpha source radiation response tests for single crystal detectors under bias at different time periods ................................ ................................ .. 84 5 10 Electron rise time of the SBI detector from alpha particles. ................................ 86 6 1 SEM micrographs of starting nano powder and fractured surfaces. ................... 90 6 2 MWS BT annealed in air at different temperatures and times (1050C for 12 h, 1100C for 24 h,1250C for 24 h from left to right). ................................ ........ 91 6 3 XRD patterns of nano powder, as sin tered PS, MWS, SPS, and annealed MWS BT. ................................ ................................ ................................ ............ 93 6 4 Dielectric properties of as sintered BT at 1, 10, 100, 500 kHz, and 1 MHz. ....... 95 6 5 Effect of annealing on the dielectric properties of MWS BT (at 1 kHz and room temperature). ................................ ................................ ............................. 98 6 6 Dielectric properties (at 1 kHz) of MWS BT with different annealing temperatures (1050C, 1 000C and 900C) for 12 hrs. ................................ .. 100 6 7 Dielectric properties of MWS BT annealed at 950C for 12 hrs as a function of frequency in the temperature range from room temperature (27C) to ( 233C) ................................ ................................ ................................ ............. 101 7 1 SEM images for SPS 0.95 and SPS 1.00 sintered ceramics. .......................... 104 7 2 The X ray diffraction patterns for the starting powde rs (BT 1.00 and 0.95) and the sintered ceramics (SPS BT 1.00 and 0.95). ................................ ................ 105 7 3 Dielectric property of SPS BT as a function of temperature (20 ~ 300 K). ........ 107 7 4 Activation energy of thermally activated relaxations. ................................ ........ 109 7 5 Log 10 r f) vs. log 10 f plot for the SPS (1150C) BT at different temperatures (30 K ~ 300 K). ................................ ................................ ................................ 112 7 6 Dielectric properties of SPS BT 0.95 sample as a function of frequency (40 Hz 1 MHz) at different temperatures (40 K 300 K). ................................ ..... 113 7 7 Dielectric properties of SPS BT 0.95. ................................ ............................... 115 7 8 T he imaginary part of the relative permittivity changes for SPS BT 0.95 as a function of frequencies (40 Hz ~ 1 MHz) at 300 K ~ 100 K. .............................. 117
11 7 9 Activation energy for polaron hopping polarization at 300 K ~ 200 K. .............. 118 7 10 Dielectric properties and conductivity of SPS BT as a function of frequency (1 kHz ~ 4 MHz) at different temperatures (110 K ~ 220 K). ................................ 121 7 11 Temperature dependence of dc conductivity ( ) with respect to 1/T and 1/T 1/2 Black lines are the linear fits of the data. ................................ ............... 125 7 12 Activation energies and hopping distance as a function of temperature for VRH conduction in SPS BT. ................................ ................................ ............. 127 7 13 Experimental data with fitting data of the real and imaginaly part of permittivity for SPS BT. ................................ ................................ .................... 129 7 14 Interfacial and hopping polarization model for the colossal permittivity of BT ceramics. ................................ ................................ ................................ .......... 131 8 1 Kroger Vink diagram for iodine and bismuth vacancies with respect to iodine partial pressure. ................................ ................................ ................................ 136
12 LIST OF ABBREVIATIONS DFE Defect Formation Energy DFT Density Functional Theory DOS Density of State EIS Electrochemical Impedance Spectroscopy FWHM Full With at Half Maximum IBLC Inter Barrier Layer Capacitance ICP AES Inductively Coupled Plasma Atomic Emission Spectroscop y LDA Local Density Approximation MCA Multi Channel Analyzer NIM Nuclear Instrumentation Module PAW Projector Augmented Wave PVT Physical Vapor Transportation SEM Scanning Electron Microscope SCCM Standard Cubic Centimeters per Minute SCFH Standard Cubic F eet per Hour T c Currie Temperature TMZ Traveling Molten Zone UDR Universal Dielectric Response XRD X ray Diffraction
13 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requ irements for the Degree of Doctor of Philosophy DEFECT ENGINEERING OF ELECTROCERAMICS: BISMUTH TRIIODIDE AND BARIUM TITANATE By HyukSu Han M ay 2014 Chair: Juan C. Nino Major: Materials Science and Engineering Defect engineering has a significant potent properties, and thus much attention has been focused on the defect functionality relationships in the fields of materials science and engineering. Point defects in compounds have great impact on electrical properties o f the material and the fabricated devices as well. The migration and concentration of point defects are two key parameters being attributed to functional properties of the compound. These parameters can be optimized by intentionally introducing extrinsic defects (i.e. dopants) or changing processing conditions (i.e. temperature and time). In this work, the defect electrical property relationships were investigated for two different materials (BiI 3 and BaTiO 3 ). BiI 3 is a promising material as room tempe rature gamma ray detector due to its high atomic number, high density, and wide bandgap. However, the high concentration and mobility of iodine vacancies in the compound increases the leakage current in the device and degrades detector performance. Thus, defect engineering has a potential to resolve this problem and improve the electrical properties of the detector. BaTiO 3 is a one of the best dielectric material in the world. Recently, it has been reported that fast
14 fried BaTiO 3 shows abnormally high r elative permittivity due to induced defect dipoles. However, the precise defect dipole polarization mechanisms acting in fast fired BaTiO 3 ceramics have not been understood, and therefore here a comprehensive investigation w as performed to reveal the pola rization mechanisms associated with defect dipoles in fast fired BaTiO 3
15 CHAPTER 1 INTRODUCTION 1.1 Statement of Problem and Motivation In the modern microelectronics industry, attention has been focused on modifying the functionality of materials in order to optimize their electrical, optical, magnetic, and mechanical properties. 1 3 These functional properties are significantly influenced by various dimensional de fects in the material such as vacancies, impurities, grain boundaries, dislocations, and pores, etc. Thus, in the field of microelectronics, a tremendous amount of studies have been conducted on the nature, concentration, and arrangement of defects, espec ially in semiconducting materials such as silicon (Si) and gallium arsenide (GaAs), which have resulted in remarkable developments in modern electronic devices. 4 11 Recently, electroceramic compounds are receiving much attention as functional electronic materials in the fields of sensors and actuators, energy conversion and storage devices, and high K dielectrics and ferroelectrics, given their high susceptibility with excellent thermal and mechanical stabilities. 12 19 In electroceramics, optimization of many functional properties relies on precise control of the zero dim ensional defects in the material (i.e. point defects). 20 27 However, only a few studies have been completed for electroceramic compounds on understating how point defects can affect the functionality of the material. Thus, the key objective of this dissertation is to investigate relationships between point defects and electrical properties of electroceramics studied here, bismuth tri iodide (BiI 3 ) and barium titana te (BaTiO 3 BT), and thus to provide defect engineering strategies for the compounds to further improve their functionalities.
16 1. 2 Scientific Approach BiI 3 is a promising material as room temperature semiconductor for gamma ray radiation detectors given its high detection efficiency. 28 30 However, to date the potential of BiI 3 as a semiconductor radiation detector compound has not been realized mainly due to it s high i onic conductivity that can lead to degrad ation in detecti o n performance over time 28 A t low and room temperature t he ionic conduction of BiI 3 can be attributed to the high concentration of intrinsic defects, mainly iodine vacancies resulting from the high volatility of iodine. The large number of intrinsic and/or extrinsic defects reduces the dark resistivity of BiI 3 and increase s the leakage current and polarization effect s, all of which consequently degrade radiation de tection performance of the device. 31 35 Based on this, it is clear that limiting the formation and migration of the in trinsic and ex trinsic defects is essential to enhance the electrical property of BiI 3 and realize its potential as a radiation detection material. Thus, d efect engineering has the potential to mitigate the obstacles currently associated with the material. In order to understand and control intrinsic defects in BiI 3 a prevalent ionic conduction defect model of BiI 3 will be established by donor ( t ellurium, Te) doping. To reduce the formation and migration of iodine vacancies, antimony (Sb) dopants will be doped into BiI 3 based on crystallographic conc epts Detail s of the scientific approach for dopant selection are given in Chapter 5.2 and 5.3, respectively. Table 1 1 summarizes the main intrinsic defect reactions in BiI 3 and extrinsic defect reactions by ex ternal dopants that will be used to extract thermodynamic parameters for intrinsic defects of BiI 3 and to control iodine vacancy formation.
17 Table 1 1 M ain intrinsic and extrinsic defect reactions in BiI 3 Intrinsic defect reactio n Extrinsic defect reaction Shottky defect Donor doping Iodine volatilization Interstitia l doping *D: Dopant In order to prove this defect engineering strategy, pure BiI 3 and doped BiI 3 s ingle crystals were grown by using a modified vertical Bridgman method with a previously optimized temperature profile (growth rate of 0.5 mm/h and temperature gradient of 10 C/cm). 29 Then, the grown single crystals were characterized by various techniques. The crystal structure and phase purity was determined using X ray diffraction (XRD). The chemical composition and impurity concentration of the crystal were characterized by inductively coupled plasma atomic e mission spectroscopy (ICP AES). Ionic conductivity was measured using two point AC electrochemical impedance spectroscopy (EIS). I V characterization was performed to determine bulk resistivity of the crystal. Leakage current and radiation response meas urements were also performed on the fabricated single crystal detector. D ensity functional theory (DFT) calculations were employed to provide insight into the most energetically favorable site where Sb might be incorporated in BiI 3 lattice and how the dop ants bond with surrounding ions. In addition, efforts were made to reduce extrinsic impurity defects in the crystal by purifying BiI 3 starting powder through physical vapor transportation (PVT) and travelling molten zone (TMZ) techniques. As a second pa rt of this dissertation, dielectric properties associated with point defects such as space charge were also investigated for BT compounds. Recently, an
18 unusual dielectric response has been reported for BT ceramics when synthesized through certain fast fir ing processing techniques. 36 These studies indicate that abnormally high relative permittivity ( = 10 4 ~ 10 5 ) with low dielectri c loss ( < 0.05) can be achieved in BT ceramics, and this extremely high relative permittivity has been referred to as colossal permittivity (CP). 37 38 At first, it was originally thought that CP might be related to a new intrinsic polarization mechanism due to crystal structure and/or electrical charge ordering. However, re cent investigations have revealed that CP can be possibly attributed to a well known extrinsic Maxwell Wagner (MW) polarization associated with induced point defects inside the material. 36 39 40 However there is an ongoing debate regarding the exact mechanisms and explanations for the origin of CP in BT ceramics. In addition, while using fast fired BT ceramics in conventional capacitive applications has been investigated, the strong temperature and frequency dependenc e of CP limits the use of fast fired BT ceramics in actual applications 41 48 It is widely accepted that the temperature and frequency dependenc e of CP is the result of extrinsic effects such as interfacial polarization and hopping conduction contributing to the bulk conductivity of the material. 49 51 T hus, understand ing the fundamental mechanisms of the bulk conduction in fast fired BT ceramics is essential for tailo r ing the temperature and frequency dependent dielectric response and fully realiz ing their potential. I n order to address th ese issue s the first step of this work is to synthesize BT ceramics exhibiting CP through fast firing techniques such as spark pl asma sintering (SPS) and microwave sintering (MWS). Processing variables, such as sintering and annealing conditions, were optimized to improve the dielectric properties of fast fired BT
19 ceramics. For comparison purposes, BT ceramics were also sintered b y a conventional pressureless sintering (PS) process Then, the second step is to characterize dielectric properties of synthesized BT ceramics at different temperatures and frequencies via broadband dielectric spectroscopy. Measured dielectric data was analyzed following polarization and conduction mechanisms in fast fired BT ceramics. In addition, the contributions of each polarization mechanism to the relative permitt ivity and the conductivity of fast fired B T ceramics we re estimated based on corresponding analytical models. 1.3 Organization of Dissertation Chapter 2 provides a brief introduction and background information on the defect chemistry in electroceramics. The correlations between the point defects and electrical properties are discussed. Chapter 2 also introduces the dielectric polarization mechanisms linked with point defects. A brief crystallographic overview and a summary of relevant physical properti es of BiI 3 and BT are presented. Chapter 3 describes the main experimental procedures and characterization techniques used in this study. Chapter 4 reports the effect of extrinsic impurities on the electrical properties and radiation response of BiI 3 det ectors. PVT and TMZ techniques are utilized to reduce the concentration of extrinsic impurities. The electrical properties and radiation response of BiI 3 detectors fabricated using commercial and ultra pure BiI 3 powders are compared. Chapter 5 presents the defect engineering strategies to further improve the electrical properties and radiation response of BiI 3 detectors. Defect models are theoretically established for BiI 3 system by donor doping. Defect engineering through
20 Sb doping into BiI 3 lattice is proposed in order to reduce formation and migration of iodine vacancy related with current obstacles for iodine compound semiconductor detectors. DFT calculations are used to determine most probable doping site of Sb in BiI 3 lattice and to investigate how Sb bonds with surrounding atoms. The electrical properties and radiation response of Sb doped BiI 3 and pure BiI 3 detectors are compared. Chapter 6 covers fast firing synthesis methods to induce CP in BT ceramics. Dielectric properties of fast fired B T synthesized through SPS and MWS are discussed with comparison to those of PS sintered BT ceramics. Moreover, the effect of annealing on dielectric properties of MWS BT ceramics is present. Chapter 7 identifies the dielectric polarization mechanisms in fast fired BT ceramics exhibiting CP. Broadband dielectric spectroscopy (40 Hz ~ 4 MHz, 300 K ~ 20 K) is performed in order to reveal dielectric polarization mechanisms and also bulk conduction mechanisms in fast fired BT ceramics exhibiting CP. In addit ion, the contributions of the each polarization mechanisms to the relative permittivity were estimated Finally in Chapter 8 a summary of the dissertation is presented and the future work in the relevant research areas is discussed. 1.4 Contributions to the Field The main contributions of this dissertation to the development of BiI 3 single crystal as a room temperature gamma ray detector and fast fired BT as a high K dielectrics Improved electrical properties, such as low leakage current and high resist ivity, were achieved in ultra pure BiI 3 detectors (total impurity level below 10 ppm) which were
21 fabricated by using PVT synthesized ultra pure BiI 3 powder. This demonstrated that extrinsic impurities have significant influence on the functionality of the BiI 3 detector. This work was submitted to the Journal of instrumentation Growth, fabrication and testing of bismuth tri iodide semiconductor radiation detectors author was responsible for all the sample synthesis and preparation s, data measurements for impurity analysis, and writing part of the manuscript. A defect engineering strategy was developed to mitigate obstacles associated with BiI 3 detectors (i.e. high ionic conductivity due to high concentration of iodine vacancy). To accomplish this defect engineering in BiI 3 Sb was determined as an optimal extrinsic dopant with a concentration of 0.5 at%. Low leakage current, high resistivity, and less polarization effect were achieved in defect engineered BiI 3 single crystal detec tors. This work was published in the Journal of Physical Chemistry C with Enhanced electrical and radiation detecting properties of Sb doped BiI 3 single crystal as a room temperature gamma ray detecto r esponsible for lead ing the manuscript, all the sample synthesis and preparations, and the data measurements and analysis except for computational works, leakage current, and alpha spectrum. CP was induced in BT ceramics synthesized by MWS techniques for the first time. D ielectric properties of MWS BT ceramics were optimized through various annealing conditions. This work was published in the Journal of American Ceramic Society Colossal p ermittivity in m icrowave s intered b arium t itanate and e ffect of a n nealing on d ielectric p roperties leading the
22 manuscript, all the data measurements and analysis, and all the sample synthesis and preparation except for SPS BT ceramics. The origin of CP in fast fired BT ceramics was dete rmined to be a result of a hopping polarization within semiconducting grains in combination with interfacial polarization at the insulating grain boundary. The relative contributions of ea ch polarization to CP of BT ceramics sintered by SPS technique were calculated as ~5 7 % hopping polarization ~2 8 % interfacial polarization, and ~15 % electrode effect. This work was published in the Journal of Applied Physics Origin of c olossal p ermittivity in BaTiO 3 via b roadband d ielectric s pectros copy responsible for leading the manuscript, all the data measurements except for ICP AES data, and all the data analysis except for ICP AES results. A nalysis of the temperature dependen t bulk dc conductivity reveal s that the bulk conduct ion in fast fired BT exhibiting CP is the result of variable range hopping (VRH) rather than nearest neighboring hopping (NNH) This work was accepted by the Journal of Physical Chemistry C Variable range hopping conduction in BaTiO 3 ce ramics exhibiting colossal permittivity leading the manuscript, all the data measurements and analysis.
23 CHAPTER 2 BACKGROUND This chapter summarizes the theoretical background required for understanding the rese arch work covered in the following chapters. 2.1 Defect Chemistry Th e logic and flow of this section is chiefly based on the 52 53 2.1 1 The Krger Vink N otation Point defects in the crystal lattice indicate a deviation from the long range periodicity. The physical and chemical properties of materials are significantly affected by various point defects. Thus, a simple and consistent way to describe point defects in the material is necessary. In ceramic compounds the notation is widely used to describe point defects. 54 The notation allows to apply chemical thermodynamics to defect equilibria by incorporating defect formation into chemical equations. In the notation, vacancies are indicated by the symbol V, and the absent atom at a normally occupie d site is represented as a subscript by the chemical symbol for the element. For example, V O represents an oxygen atom vacancy in the lattice. The position of a defect substituting for atom in the crystal is specified by the subscript chemical symbol of the originally occupied atom at the site. The impurity is written in its chemical symbol, and hence, if Mg atom substitutes Ni site in NiO, the notation for this defect would be Mg Ni Interstitial positions are note by one or more lattice defects can associate with one another to form defect clusters. In this case, parentheses enclosing the components are used to
24 indicate the defect cluster. For instance, (V M V X ) indicates a defect cluster of a metal and non metal vacan cies pair. The notation can also define the effective charges (q e ) of the defects. The effective charge of a defect is the relative charge of the defect with respect to that of the originally occupied atom at the site in a perfect crystal. For atomic or ionic species, where z d and z s are the real charges on the defect species and to denote each unit of effective negative and positive charges, respectively. In an imperfect crystal, free electrons exist in the lattice, which is denoted by the symbol of and the 2.1 2 Intrinsic P o i nt D efect R eactions in S toichiometric C ompounds: Schottky and Frenkel Traditional chemical equations can be modified in order to describe defect reactions in the crystal. Thi s allows to applying chemical thermodynamics into the defect system, and thus possible to establish defect chemistry in which the defects are regarded as the chemical species in traditional chemistry. The defect reaction can be considered as the formation of defects in a perfect crystal with dopant. The rules for writing defect reaction are similar to those of traditional chemistry, however, defects have to be quantified with respect to crystallographic sites rather than molecules or moles. The following rules have to be followed:
25 Mass balance: mass cannot be created or destroyed. Vacancies have zero mass. Charge balance: charges cannot be created or destroyed. Site balance: the ratio between the numbers of regular cation and anion sites must remain const ant and equal to the ratio of the parent lattice. In stoichiometric compounds which the composition of the crystal is fixed, the point defects should be formed maintaining the composition of the material. There are two representative intrinsic point defec ts in stoichiometric compounds which called Schottky and Frenkel defects, respectively. Schottky defects are defines as the defects arising from balanced populations of cation and anion vacancies in any crystal. For example, in MO oxide, one Schottky def ect is equal to one pair of cation vacancy ( ) and anion vacancy ( ) in the crystal, and each vacancy does not necessarily to be near each other. In addition, if only Schottky defect exists in the crystal, the total number of anion vacancies should be con sistent with that of cation vacancies in order to maintain the composition and charge neutrality. Thus, the total number of Schottky defect should be equal to one half of the total number of vacancies. The defect reaction for MO oxide can be written as, (2 1) For Frenkel defect, one atom or ion is displaced from the original lattice and moves into an interstitial site in the crystal. The original lattice left as a vacancy site. Thus, a Frenkel defect consists of one interstitial ion and one vacan cy in the lattice, which means the number of interstitials is equal to that of vacancies forming a Frenkel defect. The defect reaction for a cation Frenkel defect in MO oxide can be represented as,
26 (2 2) Figure 2 1 depicts graphical images of Schottky and cation Frenkel defects in MO oxide. A B Figure 2 1. Representative defects in MO oxide A ) Schottky and B ) Frenkel defects The equilibrium of defect formation can also be treated as a chemical equilibrium. In the defect formation, the concent ration of defects replaces the activity of species in chemical equilibrium. Therefore, in the case of vacancy formation,
27 (2 3 ) and, at equilibrium the change in Gibbs free energy ( G v ) can be given by, (2 4 ) K v is the equilibrium constant for a single vacancy formation and k is Boltzmann constant Thus, K v can be expressed as, (2 5 ) n v is a concentration of vacancy at equilibrium in N lattices. Thus, n v can be written as (2 6 ) I f neglect entropy term, n v can be expressed as follow. (2 7 ) where is the enthalpy change for vacancy formation. Equilibrium of Schottky defect in a crystal can be considered in a same manner described above. For a cry stal of composition MX, Schottky defect formation can be written as, cation vacancy + anion vacancy (2 8 ) (2 9 ) where V M and V X represent vacancies on cation and anion sites, respectively. The law of mass action can yield the equilibrium constant for the formation of Schottky defects, K S
28 (2 1 0 ) where n cv and n av denote the number of cation vacancies and the number of anion vacancies, and N is the number of cation sites, which is equal to the number of anion sites in the crystal. is the mola r Gibbs free energy of the formation of the Schottky defect, and R represents the gas constant. Since n cv is equal to n av for the Schottky defect, the number of Schottky defects (n s ) can be described as, (2 1 1 ) n s indicates the number of Schottky defe cts per unit volume in the crystal. By neglecting entropy terms, (2 1 2 ) where is the enthalpy required to form 1 mol of Schottky defects in the crystal. Table 2 1 summarizes the enthalpy of Schottky defects for some of Alkali halide compounds (MX) w hich have sodium chloride (NaCl) structure. Table 2 1 Experimentally determined Schottky defect formation enthalpy in Alkali halide compounds. Compound LiF 225.2 2.33 LiCl 204.1 2.12 LiBr 173.4 1.80 LiI 102.4 1.06 NaF 233.1 2.42 NaCl 225.8 2. 34 NaBr 203.0 2.10 NaI 140.9 1.46 KF 262.0 2.72 KCl 244.5 2.53 KBr 224.6 2.33 KI 153.0 1.59
29 For Frenkel defects on the cation sublattice, the chemical equilibrium can be expressed by, cation vacancy + cation interstitial (2 1 3 ) (2 1 4 ) wher e V M and M i are a vacancy on a cation site and a cation interstitial, respectively. Hence, the equilibrium constant for the formation of Frenkel defects, K cF can be written as, (2 15 ) here, n cv represents the number of cation vacancies and n ci is the number of cation interstitials. In addition, N and N i indicate the number of cation and interstitial sites, respectively. is the molar Gibbs free energy for the formation of cation Frenkel defects. In the case of Frenkel defects, n cv is equal to both of n ci and n F (the number of Frenkel defects), and thus following equation can be established for the Frenkel defects in the crystal. (2 16 ) (2 17 ) where n cF and are the number of cation Frenkel defects per unit volume and the enthalpy of Frenkel defect formation in the crystal. The same approach can be applied for deriving defect equilibria equations for Frenkel defects on the anion sublattice, and the following equations are obtained.
30 (2 18 ) (2 19 ) Experimentally determined Frenkel defect s formation energies are given in Table 2 2 for some compounds. Table 2 2 Examples of Frenkel defect formation enthalpy. Compound AgCl 139.7 1.45 AgBr 109.0 1.13 AgI 57.8 0.60 CaF 2 261.4 2.71 SrF 2 167.4 1.74 BaF 2 184.3 1.91 2.1 3 Diffu sion and E lectrical C onductivity Diffusivity and electrical conductivity of ceramics are significantly influenced by the presence of point defects. Point defects increase the number of available jumping sites for the diffusion in the material. In additio n, since point defects have effective charges, it can be migrated under an electric potential gradient which affects electrical conductivity of the material. Theoretical relationships between point defects and diffusion or electrical conductivity will be given in the following discussions. The diffusivity (D) of an atom or ion is a measure of the ease and frequency with which that atom or ion jumps around in a crystal lattice. It is widely accepted that D is thermally activated and can be written as, (2 2 0 )
31 where Q is the temperature independent activation energy and D o is the temperature dependent constant for the diffusion. In addition to temperature, it also has been understood that diffusivity of the material is strongly dependent on the stoichiome try and purity level of the material. The following equation (2 2 1 ) can be used to fundamentally relate the diffusion coefficient D to the atomistic diffusion process in a solid. (2 2 1 ) where is the number of successful jumps per second, is the jump ing distance, and is a geometric parameter which depends on the crystal structure. For cubic lattices, where is the coordination of the vacancy. Furthermore, is the product of ( ) and to have empty adjacent site for diffusion ( ). (2 2 2 ) According to the Boltzman distribution law, the probability (P) of a particle to have an energy ( ) or greater is defined by, (2 2 3 ) Thus, if an adjacent site is vacant, can be expresse d as, (2 24 ) where is the vibrations of the atoms. Moreover, is significantly lower than 1 since most of the sites are surrounded by other atoms, and can be assumed to equal with
32 the concentration of vacancies ( ) in a crystal. Thus, considering the vibrational entropy, the diffusion coefficient can be driven by, (2 25 ) where is given by, (2 26 ) The vibration entropy of is associated with the frequencies of the ions in ground ( ) and activated ( ) states by, (2 27 ) Putting all togeth er, a final expression for the diffusion coefficient can be obtained as, (2 28 ) In the presence of electric field, an ion is diffusing under an electrical potential gradient of where is the electric potential in volts. Ions diffuse from higher to lower electric potentials, and the diffusion process is no longer random however is biased in the direction of decreasing free energy. In such situations, the driving force for the diffusion can be expressed by, (2 29 ) where z i is the net charge on the moving ion. The ionic flux, is related with the current density, by (2 3 0 )
33 Substituting equation (2 29 ) into (2 3 0 ), (2 3 1 ) by comparing equation (2 3 1 ) can be given as, (2 3 2 ) This equation which is known as the Nernst Einstein equation relates the diffusion coefficient to the ionic conductivity. Furthermore, for the mobile ions drifting with an average velocity ( ), the current density can be given by, (2 3 3 ) where c m is the concentration of t he mobile ions (number of carriers per cubic meter). In addition, the electric mobility of the carrier, is defined as the average drift velocity per electric field. (2 34 ) By substituting equation (2 3 3 ) and (2 34 tionship can be driven, (2 35 ) If there are more than one type of mobile charged species, the total ionic conductivity should be,
34 (2 36 ) Moreover, by comparing equation (2 35 ) with (2 32 ), one can expect that the electric mobility of the charged mobi le ion can be written as, (2 37 ) This equation implies that the mobility of a charged species is directly related to the diffusivity of the species. Several experimental techniques and computational tools can be utilized to extract the key thermodynamic and kinetic parameters related with defect formation and migration. Some of the most useful techniques for defect characterizations are summarized in Table 2 3. 25 Table 2 3. Defect characterization tools. Technique Thermodynamic parameters related with defects Impedance spectroscopy 55 57 Diffusivity measurements via cond uctivity relaxation Defect concentrations by measuring chemical capacitance Separation of bulk and interfacial contributions to the electrical p properties by measuring frequency dependent complex impedance Thermogravimetric analysis (TGA) / coulometr ic titration 56 58 Defect concentration by measuring mass change or electrochemically induced stoichiometry cha nges Secondary ion mass spectroscopy 59 Chemical composit ion profiles and tracer diffusion measurements X ray and neutron diffraction 60 61 Defect ordering and concent ration by investigating atomic occupancy factors and lattice strain Optical absorption and emission 62 63 E lectronic structure and defect energy states 2.2 Dielectric Polarizations in Solids K. C. 64 65
35 2.2.1 Polarization M echanisms Polarization in dielectrics is defined as short range movement or a limited rearrangement of ch arge carriers under an applied electric field ( E ). The polarization (P) is given by, (2 38 ) where N q and represent the number of the dipole moments per unit volume, the charge, and the distance between the charges, respectively. The dielectric pe rmittivity ( ) of the material is related with P by, (2 39 ) where is the vacuum permittivity (8.854 x 10 12 F/m). Furthermore, the polarizability ( ) of an atom or ion is described by, (2 4 0 ) where E loc is the local electric field, and is determi ned for a cubic lattice as, (2 4 1 ) by substituting equation (2 4 0 ) and (2 4 1 ) into (2 39 ), (2 4 2 ) which can be rearranged as, (2 4 3 )
36 This equation, which is known as the Clausius Mossotti equation, links the microscopic property ( ) of the mater ial with its macroscopic property ( ). It is important to note that four main polarization mechanisms exist in dielectric materials: space charge polarization, dipolar polarization, ionic polarization, and electronic (atomic) polarization. Electronic (at omic) polarization indicates the displacement of the electron cloud relative to the nucleus under the applied electric field. This polarization can respond up to frequencies of ~10 15 Hz. Ionic polarization is the displacements of cation and anion toward the negative and positive bias, respectively. The displacements of both positive and negative ions result in ionic dipoles, which can respond to an applied electric field up to ~10 13 Hz. Dipolar polarization is attributed to the alignment of randomly ori ented dipoles under an external field, and it can respond at relatively lower frequencies (up to ~10 8 Hz). Lastly, space charge polarization is due to an inhomogeneous spatial distribution of charge centers over the microstructure. Space charge polarizat ion can respond up to only ~10 6 Hz. Figure 2 2 depicts schematics of four main polarization mechanisms, and Figure 2 3 represents the contributions of each polarization mechanism to the total relative permittivity as a function of frequency. Since charg ed point defects can be significantly attributed to space charge polarization at interfaces (grain boundaries) and in grains through interfacial polarization and hopping polarization, respectively, more details about space charge polarization is given in t he following section.
37 Unpolarized Polarized --------> Electric Field Space Polarization Dipolar Polarization Ionic Polarization Atomic Polarization Figure 2 2. Schematics of four main polarization mechanisms.
38 Figure 2 3. Contributions of each polarization mechanism to the total relative permittivity as a function of frequency. 2.2.2 Space C harge P olarization Dielectric polarization can be associated with localized electronic defects, such as electrons and holes, and ioni c defects, such as vacancies and impurities. These charge carriers can move under the applied electric filed, and build inhomogeneous spatial charge distribution though the material, which is generally referred to as space charge. Moreover, the opposite charges may form electrical dipoles, which can be polarized under a.c. field, and thus contribute to the dielectric properties of the material. This polarization phenomenon is defined as space charge polarization, which can be considered following two pos sible ways.
39 184.108.40.206 Hopping p olarization Localized charge carriers in a dielectric material can hop together by jumping from one site to the neighboring site, which results in forming hopping dipoles. Under applied electric field, these hopping dipoles ca n affect dielectric response of the material in the same manner with dipolar polarization, and is commonly referred to as hopping polarization. In hopping polarization, charge carriers should overcome a potential barrier between neighboring sites to hop, and thus the hooping process is normally temperature activated process. Therefore, in thermal equilibrium, the probability (p o ) for a charged particle to form a hopping dipole can be written as, (2 44 ) which is the case for the negatively charged parti cle that hop from site A to B, leaving behind a positive charge at site A and creating a negative charge at site B. Here, C and E A are a constant and the activation energy for hopping process, respectively ( Figure 2 4 ) Figure 2 4 Hopping polarizat ion due to the hopping of the charged particel over the barrier from one site to the other.
40 220.127.116.11 Interfacial polarization The interfacial polarization is due to the positive and negative space charges at the interfaces between different materials or m icrostructures such as grain boundaries (Figure 2 2). In this case, equivalent circuit can be established with two components in series, and each component is composed of two parallel elements, i.e., a capacitance and a conductance (Figure 2 5). This mod el is referred to as internal barrier layer capacitance (IBLC) model, and widely used to describe dielectric response of the material being attributed to interfacial polarization. Figure 2 5 A dielectric material comprising of two different compone nts. In IBLC model, one component is highly conductive while the other component is highly capacitive, and conductive phases are separated by very thin layers of capacitive phases. Thus, space charges in conductive component can be attributed to dielectr ic polarization while not conducting through the material. Then, the effective relative permittivity ( ) according to IBLC model can be represented by, (2 45 )
41 where, and d 1 are the relati ve permittivity and thickness of conductive component, and d 2 represents the thickness of capacitive thin layer. 2.2.3 The U niversal D ielectric R esponse The frequency dependence of P can be given by, (2 46 ) where f is the circular frequency in Hz, an d is the complex permittivity, which can be written as, (2 47 ) here The components of polarization in phase and out of phase with the external field are respectively associated with the real part ( ) and the imaginary part ( ) of The component of imaginary part is known as the dielectric loss, and is directly related to the energy lost per radian given by, (2 48 ) where E o is a peak amplitude of applied electric field. Since the power lost due to phase difference is defined by the a.c. co nductivity of dielectrics can be driven by, (2 49 ) This equation reveals the relationship between the dielectric loss and the a.c. conductivity of the material. The responses of electronic and ionic polarizations are so rapid that they can contribute to a purely real value of the permittivity ( ) below GHz frequencies. However, permanent dipoles such as dipolar ionic defects and hopping charge carriers
42 have much slower response such that the total permittivity of a medium can be written as, (2 5 0 ) wh ere index indicates the various polarization mechanisms. thermal excitation between two preferred sites separated by a potential barrier. The Debye susceptibility can be expres sed by, (2 5 1 ) where is the relaxation time which is equal to the inverse of the relaxation frequency ( ). is also identical to the loss peak frequency, and is normally thermally activated, (2 5 2 ) where and E A are the jump frequency and the activation energy, respectively. It is important to note that in the Debye model, the loss peak given by the imaginary part in equation (2 5 1 ) is symmetric with a width of 1.144 decades at the half max height on a log scale. However, it is often observe d that the dielectric loss in real materials strongly deviates from the Debye response. Figure 2 6 shows some examples of dielectric behaviors departed from the Debye theory. For the comparison purpose, the typical Debye response is also presented.
43 Figure 2 6 Examples of dielectric behaviors departed from the Debye theory. A. K. Jonsher firstly pointed out that the asymmetric dielectric loss behavior of the various materials can be characterized by the relation, with 0 < s < 1 (2 5 3 ) which is dielectric response of solids can be represented by the diagrams shown in Figure 2 7 The dielectric response dominated by the dipoles may be represented as Figure 2 7 (A ), while the di agram of Figure 2 7 (B ) represents when the mobile carriers dominates low frequency response. However, for both cases, the UDR law (equation (2 53 )) obeys the responses at the extended range of frequencies. Furthermore, the Kramers Kroniq transformation of gives rise to the real part of susceptibility, which has
44 the same functional form but with a constant factor of Thus, the ratio of the real and the imaginary parts of susceptibility can be represented as, (2 54 ) which is completely contras t with the Debye theory where the ratio is equal to A B Figure 2 7 Schematic representations of the two types of dielectric response, A ) response dominated by dipoles and B ) re s ponse dominated b y mobile
45 carriers at low frequencies. By substit uting equation (2 5 3 ) into (2 49 ), it is clear that the frequency dependence of the electrical conductivity of the materials follows the expression, (2 55 ) where and s are constants, and is the direct current conductivity. It should be noted that t he exponent s represents the degree of localization related charge dipoles, and s value closer to 1 indicates that the charge carriers are more highly localized. In addition, the response with s values close to unity is associated with the intrinsic latti ce response, whereas the response with the s values between 0.5 < s < 0.9 are strongly related with the extrinsic carrier such as hopping electrons, induced defects, and impurities. Table 2 4 shows some examples of the numerical s values for different mate rials. Table 2 4 Examples for numerical values of the exponent "s" for different materials. Material Temperature (K) Exponent s Si 3 ~ 12 0.79 ~ 0.74 Al 2 O 3 77 ~ 87 0.75 ~ 0.70 Sb x As 2 x S 3 293 1.0 As 2 Se 3 300 1.0 Se 300 0.95 As 2 S 3 300 0.92 As 2 Se 3 300 0.92 P2O5 FeO CaO glass 300 0.85 ~ 0.89 SiO x 211 ~ 297 0.6 ~ 0.7 2.3 Crystal Structures of BiI 3 and BaTiO 3 2.3.1 BiI 3 Effective defect engineering relies on the understanding of the crystal structure of the compound. Therefore, it is important t o recall that the crystal structure of BiI 3 is a
46 rhombohedral structure ( space group No. 148) with the lattice parameters of a = 7.516 and c = 20.718 , respectively. In the structure, Bi ions reside on the 6c Wyckoff positions with coordinates (0, 0 0.1667) and I ions occupy the 18f positions with coordinates (0.3415, 0.3395, 0.0805). 66 67 BiI 3 is comprise d of layers of corner sharing BiI 6 octahedra stacked in the  with weak van der Waals forces between the layers ( Figure 2 8 ( A ) ) While I Bi I layers are bonded by weak van der Waals forces, bismuth and iodine are bonded by strong ionic bonds Thus, this anisotropic crystallography leads to (001) cleavage planes in BiI 3 crystals and the preferred crystal growth direction along (001) planes. In addition, BiI 3 shows various anisotropic physical properties, such as optical, thermal and electrical proper ties due to crystallographic anisotropy. 68 70 Three close packed layers stack in the sequence of I Bi I in each I Bi I layer. A polyhedral schematic for one I Bi I layer with empty octahedral sites is presented in Figure 2 8 ( B ) The red (lighter on the top and darker below the bismuth layer ) and blue atoms represent iodine anions and bismuth cations, respectively Bismuth ions can occupy three different sit es within the iodine layers. It can be seen in Figure 2 8 ( A ) that two thirds of the bismuth sites (A and B) are occupied by bismuth ions and one third of the C sites remain empty octahedral sites, which follow the sequence ABCABCABC... In the crystal st ructure of BiI 3 it is hard to expect perfect stacking of the layers, however stacking faults usually exist in BiI 3 crystals leading to unique electrical properties such as stacking fault excitons. 71 Table 2 5 tabulates crystal structure information for BiI 3
47 A B Figure 2 8 Polyhedral schematic for A ) corner sharing octahedra in BiI 3 and B ) empty octahedral sites in one I Bi I layer. Table 2 5 C rystal structure information for BiI 3 Space Group: ( ) Atom Coordination Wyckoff position Site symmetry B i (0, 0, 0.1667) 6c 3 I (0.3415, 0.3395, 0.0805) 18f 1 Lattice parameters ( ) a = 7.516 c = 20.718 2.3.2 BaTiO 3 At ambient temperature, BT crystallizes in a tetragonal structure of which is identical with the mineral perovskite (CaTiO 3 3 perovskite structure, it can be visualized that A ions, which are cubic close packed, coordinate with surro unding twelve O ions, and B ions reside in the octahedral interstitial sites ( Figure 2 9 ).
48 A B Figure 2 9 Crystal structure of BT. A ) Unit cell of tetragonal BT B ) Ti polyhedral schematic. It can be easily seen that for the perovskite struc ture, the ionic radii of each ions should follow, (2 56 )
49 However, this equation does not hold exactly true for the most perovskite compounds due to small variations in the ionic radii of A and B ions. Therefore, one should invoke a t 56 ), 65 which can be written as, (2 57 ) t small lattice distortions, which is and this distortion gives rise to an excellent dielectric properties into perovskite compounds. The unit cell of BT is centrosym metric above Currie temperature (T c 120 C) however below the Currie temperature, the structure is slightly distorted along c axis leading to non centrosymmetric tetragonal structure with a dipole moment. In addition, BT exhibit s two more distinguishable temperature dependent phase transitions: (i) tetragonal to orthorhombic near 0C and (ii) orthorhombic to rhombohedral near 90 65 Each crystal structure, tetragonal, orthorhombic, and rhombohedral, ca n be thought as elongated ABO 3 cubic structure along , , and , respectively. Net charge displace ment between cations with respect to the oxygen anion octahedr on exist s in the BT crystal structures except centrosymmetric cubic crystal struct ure. Figure 2 10 illustrates (a) the transformations with (b) the corresponding lattice parameters changes, and (c) the relative permittivity of BT as a function of temperature.
50 A B C Figure 2 10 Phase transformation and relative permittivity change of BT. A ) The phase transformations B ) L attice parameters change C ) Relative permittivity change The displacements of ions due to the cubic tetragonal transformation can demonstrate how the spontaneous polarization can be coupled within unit cells. X ray studies have revealed that the ions in the tetragonal unit cell are slightly (about 3 to 10 pm) displaced against the four central oxygen ions, which is depicted in Figure 2 11 It is obvious that in this figure, Ti ion at the central positi on is displaced toward one of the
51 oxygen ion marked as A, which consequently leads Ti ion on the opposite side of A to be displaced distantly from that oxygen ion. Hence, all the Ti ions in the same column are displaced in the same direction, and thus Ti ions in the next column are coupled with the displaced Ti ions resulting in ferroelectricity in BT. Table 2 5 summarizes crystal strucuture information for tetragonal BT. 72 Tetragonal BT has space group of P4mm ( C 1 4V ) with the lattice parameters of 3.9909 for a and 4.0352 for c Each atoms, Ba (0, 0, 0), Ti (0.5, 0.5, 0.5224), O1 (0.5, 0.5, 0.0244), and O2 (0.5, 0, 0.4895), occupy Wyckoff position s of 1a 1b 1b and 2c in the crystal structure, respectively. Thermal displacement and site symmetries for each site are also listed in Table 2 6 Figu re 2 11 Approximate ion displacements in tetragonal BT. Table 2 6 Crystal structure information for tetragonal BT. Space Group: P4mm (C 1 4V ) Atom Coordination Wyckoff position Thermal displacement ( 2 x100 0) Site symmetry Ba (0, 0, 0) 1a U 11 =U 22 =2.90, U 33 =3.50 4mm Ti (0.5, 0.5, 0.5224) 1b U 11 =U 22 =0.02, U 33 =5.40 4mm O1 (0.5, 0.5, 0.0244) 1b U 11 =U 22 = 0.30, U 33 =4.80 4mm O2 (0.5, 0, 0.4895) 2c U 11 =9.80, U 22 =5.40, U 33 =10.50 2mm Lattice parameters ( ) a = 3.9909 c = 4.0352
52 CHAPTER 3 EXPERIMENTAL PROCE DURES AND PROCESSING 3.1 Sample Preparation 3.1.1 BiI 3 S ingle C rystals 3.1. 1 .1 PVT p owder synthesis technique Ultra pure (total impurity level below 10 ppm) BiI 3 powder was synthesized by the PVT technique. Commercial bismuth polycrystalline lump (99.999 %) and iodine lump (99.999%) from Alfa Aesar were used as the starting materials. The reaction between bismuth and iodine vapors was conducted in a quartz tube with diameter of 2 and length of 36 Bismuth and iodine lumps were weighed with a mole ratio of Bi : I 2 = 1 : 5 and loaded into a quartz boat separately The extra iodine is added to compensate the loss of iodine due to high volatility of iodine. The boat with iodine lump was placed at one end of the quartz tube and kept at room temperature. T he boat with bismuth lump was placed at center of the tube and heated up to 400 o C. A Teflon sheet was placed at the other end of the quartz tube at room temperature to collect the synthesized BiI 3 powder. During the reaction, Ar gas (ultra high purity) w as infused from the iodine side to facilitate the reaction between iodine and bismuth vapor, and the resulting BiI 3 powder was deposited on the Teflon sheet at lower temperature. A gas flow meter (Digital High Performance Mass Flow, Smart Track 100 Serie s from SIERRA) was used to consistently regulate the Ar flow Optimal Ar flow rates were investigated between 80 and 160 standard cubic centimeter per minute (SCCM) Maximum yield was achieved at a 120 SCCM gas flow rate Figure 3 1 shows the experimental set up for BiI 3 powder synthesis using the PVT technique.
53 Figure 3 1 E xperimental set up for BiI 3 powder synthesis using the PVT technique 3.1. 1 2 TMZ purification technique The setup was built b y Sasmit Gockhale under the guidance of Prof. Juan C. Nino. TMZ technique was employed for purifying BiI 3 starting powder. A band heater (MPP51101, TEMPCO Electric Heater corp.) was utilized to form a narrow molten zone in the ampoule. The maximum tempe rature of the heater was approximately 760C. The heating zone was narrowed to 0.125 inch by using alumina insulation rings. A temperature controller (ATHENA Series 16C) combined with a thermocouple was used to control the temperature of band heater. Th e powder was sealed in a Pyrex glass ampoule of 0.75 inch inner diameter. The ampule was mounted on a clamp, and the heater was moved about 90 zone refining passes at a rate of 4.4 cm / h. Figure 3 2 illustrates the experimental set up for the TMZ technique. Figure 3 2 E xperimental set up for the TMZ technique
54 18.104.22.168 Single crystal growth of Te and Sb doped BiI 3 BiI 3 single crystal s were grow n by the modified vertical Bridgman method Pyrex g lass ampoule was selected as the growth chamber due to relatively low annealing (= 560C) and softening point (= 815C), however those temperatures are still higher than the maximum temperature (= 440C) for the BiI 3 single crystal growth process. Customi zed Pyrex glass ampoules (inner diameter = 0.75 in, tip length 5 cm, and tip angle 70) were purchased from Southern Scientific, Inc. The Pyrex ampoule was cleaned with glassware cleaning solution (Decon Contr ad 70) to remove residual organics. The ampoule was soaked with the solution for overnight under the fume food, and then it was washed ten times with DI water ( 17.0 ) followed by baking at drying oven (120C) for overnight. The amount of 20 g of BiI 3 powder with appropriate amount of TeI 4 and SbI 3 based on the doping concentration were weighed and loaded in the ampoule. TeI 4 and SbI 3 were initially added to the ampoule and loaded at the e nd of the tip. Before vacuuming the ampoule, a neck area with a diameter of approximately 0.25 inch was preformed at the middle of the ampoule using propane torch in order to facilitate sealing process. The ampoule was firstly vacuumed to a pressure of 4 .0 10 2 mbar by using mechanical roughing pump, and then molecular turbo pump was turned on to finally evacuate the ampoule to 1.0 10 4 mbar. The vacuumed ampoule was then sealed by using a propane hand torch. The sealed ampoule was vertically mounted on a standing steel frame using galvanized steel wires, and placed in programmable 24 or 12 multi heating zone furnaces (EDG 13 and EDG 11, Mellen Company). The previously optimized growth condition, such as tem perature
55 gradient of 10C / cm and cooling rate of 0.5 mm / h, were utilized for all the single crystal growth process. 28 29 3.1.2 BaTiO 3 C eramics 3.1. 2 .1 Microwave sintering For MWS processing, commercially available nanocrystalline BT powder was purchased from Alfa Aesar (99+% metals basis 50 to 70 nm average particle size ). Before sintering, the starting powder was first ball milled for 24 h in ethanol (70 ml) and dried in an oven at 120C for 24 h. To make green pellets, the dried powder was mixed with 1 wt % of PVA binder (Celvol 103), ground in a mortar and pestle for approximately 10 min, and then si eved Gr een pellets (diameter 7 mm and thickness 1 mm) were formed initially in a uniaxial press of 170 MPa using a steel die and then isostatically pressed at a pressure of 200 MPa. A Thermwave TW 1.3 ( Microwave Systems, Inc. USA) apparatus was used for MWS process S usceptors were used to heat the samples during MWS process The following heating and cooling schedules were used during MWS; (i) first pellets were heated from room temperature to the sintering temperature of 1320C at a rate of 65 C / min (ii) then held at the sintering temperature for 30 m in and (iii) finally cooled to room temperature in 2 h Post sintering annealing treatments were performed at different temperatures rang ing between 9 0 0C and 1200C, and for 12 to 24 h in an air atmosphere at 10 standard cubic feet per hour ( SCFH ) flowing rate. 22.214.171.124 Spark plasma sintering [This process was performed by Christophe Voisin, at Universit Paul Sabatier in Toulouse France under the guidance of Dr. Sophie Guillemet Fritsch] The starting
56 nanocrystalline BT powder was synthesized by an oxalate route. Briefly, BaCl 2 .2H 2 O and TiCl 3 were used as precursors to synthesize the starting powders. In order to control the powder stoichiometry, the precursors were weighted in appropriate proportions. The precursors were then dissolved in water and added to an ethanolic oxalic acid solution. The solution was stirred and aged approximately for 5 h, and it is centrifuged and then dried overnight at 80C. After calcination at 85 0C for 4 h in static air, the oxide powders were obtained. SPS was performed on the synthesized powder by means of a Dr. Sinter 2080 from Sumimoto Coal Mining (SPS Syntx Inc., Dr. Sinter 2080). Approximately 0.9 g of powder without binder was placed in a 8 mm diameter graphite die and then sintered at different sintering temperatures (900 C 1050 C and 1150C ) with a dwell time of 3 minutes under 50 MPa of mechanical stress and electric current up to 350 A. A heating rate of 25C / min was used to reac h the sintering temperature from 600C. After 3 min dwelling, the sample was cooled down from the sintering temperature to room temperature at a rate of 100C / min. A thin carbon layer was present at the surface of the as sintered ceramic due to graphit e contamination from the die. However, it was easily removed by polishing the surface of the ceramic. For annealing process, a furnace was preheated to 850C and then the sample was placed into the furnace for 15 min dwell in air followed by air quenchin g. 126.96.36.199 Conventional pressureless sintering Experimental procedures for preparing green pellets are same as MWS sintering (Chapter 188.8.131.52). T he green pellets were then (i) first heated from room temperature to 400 C at a rate of 3C / min and (ii) hel d for 2 h with the aim to burn out the added organic binder, and then (iii) heated from 400 C to the sintering temperature of 1350C
57 at a rate of 3C / min, ( iv ) held at the sintering temperature for 2 h and ( v ) finally cooled to room temperature at a rat e of 3C / min. No annealing treatment was performed on conventionally sintered BT ceramics. 3.2 Characterization 3.2.1 X ray D iffraction Philips APD 3720 and Inel CPS X ray diffractometers were used for structural characterization of BiI 3 single cryst als and BT ceramics, respectively. The XRD data was measured using CuK radiation, and the operation voltage and current were set as 40 kV and 20 mA for APD 3720, and 30 kV and 30 mA for Inel CPS, respectively. The CPS detector is one dimensional detector that measures data over a wide 2 range simultaneously and therefore allows for rapid acquisition of diffraction data. For BT samples, a ll the XRD patterns were measured for a period of 5 min using the CPS detector, while APD 3720 w as utilized for measuring XRD data of BiI 3 samples over the 3.2.2 Scanning E lectron M icroscopy Scanning electron microscopy (SEM, JEOL 6335F FEG SEM) was performed to reveal the particle size and morph ology of the BT starting powders, and microstructure of the sintered BT ceramics The operating voltage for the SEM was 15 kV with a probe current of 8 BT ceramics were fractured using a pestle and hammer, and subsequently sonicat ed in DI water for about 30 min. The sonicated samples were dried at drying oven (120C) for approximately 1 h. The samples were mounted on SEM holder using copper (Cu) tape, and then it was coated by gold (Au) film with thickness of approximately 60 nm using sputter system (Cressington Scientific Inc., Cressington 108 Auto)
58 3.2.3 X ray P hotoelectron S pectroscopy [This process was performed by Eric Lambers, at Major Analytical and Particle Analysis Instrumentation Centers (MAIC & PAIC) in University o f Florida]. X ray photoelectron spectroscopy (XPS) was performed on finely ground BT powders using a P erkin E lmer 5100 XPS system for investigation of Ti oxidation states. An Al monochromatic X ray source (1486.6 eV) was used for c ollecting the XPS spectra. The take off angle of 45 and the pass energy of 35.75 eV were used. Step size and integration time were set as 0.1 eV and 50 ms, respectively. The pressure inside the chamber was maintained below 10 9 torr during measurements 3.2.4 Inductively C oupled P lasma A t omic E mission S pectroscopy The chemical compositions of BiI 3 starting powders and grown single crystals were determined by ICP AES. ICP AES equipment of Perkin Elmer Optima 3200 RL was used for the measurements. T he starting powder and crystals were dissolved in 10 % trace metal grade hydrochloric acid ( HCl ) solution with a concentration of 10 mg / ml. The elemental impurities were analyzed based on the reported impurity elements from the vendor. ICP AES stand ard solutions (Quality Control standard 7 and 21, SPEX Certiprep. Inc.) were used for calibrating measurements. Standard solutions with the concentrations of 1 ppm, 0.1 ppm, and 0.01 ppm were used for the impurity analysis of synthesized BiI 3 powders, whi le 1 ppm, 10 ppm, and 100 ppm were exploited for the dopant concentration analysis. Error analysis was performed on the measured data by using standard deviation with the upper and lower 95 % confidence limits. 3.2.5 Dielectric C haracterization S intered ceramic s were coated with thin Au electrodes (thickness of ~30 nm) by sputtering (Cressington Scientific Inc., Cressington 108 Auto). The coated samples
59 were placed in a closed cycle cryonic workstation (CTI Cryogenics, Model 22), and dielectric proper ties as a function of temperature (20 K ~ 300 K) were measured at the different frequencies (40 Hz 100 kHz) through an Agilent 4284 LCR meter. D ielectric measurements from room to high temperature up to 473 K were performed in a temperature chamber and controller from Delta Design, Inc. Precision impedance analyzer ( PIA, Agilent Ltd., Agilent 4294A) combined with cryonic workstation was used in order to perform broadband dielectric spectroscopy analysis (40 Hz 1 MHz) in the temperature range from 20 K to 300 K. Error analysis was performed on the measured dielectric data by using standard deviation with the upper and lower 95 % confidence limits 3.2.6 Electrical P roperty C haracterization Planar electrodes were deposited onto the fresh surfaces of BT ceramics and BiI 3 single crystals by sputter coating (KJL CMS 18 Multi Source). For BT samples, various electrodes such as Al, Ni, and Ag were deposited using different targets in order to investigate the electrode effect on dielectric properties. For Bi I 3 samples, b ased on previous electrode compatibility work, Au was selected as an electrode material with a thickness of approximately 100 nm. 73 I V characterizatio n was performed on both BT ceramics and BiI 3 single crystals by using a microprobe station (Micromanipulator 450PM Test Station) and a semiconductor parameter analyzer (Agilent 4156C) Leakage current measurements were performed on fabricated BiI 3 single crystal detectors by an electrometer ( Keithley 6517B, Cleveland, Ohio, USA ) under an electric field of 100 V/cm.
60 3.2.7 Radiation R esponse M easurement This process was performed by Sasmit Gockhale under the guidance of Prof. Juan C. Ni no and Prof. Kelly A. Jordan. The radiation response of BiI 3 single crystal detectors was investigated using Amptek A250 charge sensitive pre amplifiers (Bedford, MA, USA) and standard nuclear instrumentation module (NIM) electronics. Spectrum acquisitio n was carried out by a multi channel analyzer (MCA) and the spectrum acquisition software Maestro 32. 3.2.8 Computational M ethods This process was performed by Dr. Minki Hong under the guidance of Prof. Juan C. Nino and Prof. Susan B. Sinnott. DFT calcula tions were performed to explore the Sb incorporation in the BiI 3 lattice. Projector augmented wave (PAW) 74 75 pseudopotentials method with LDA 76 was used as implemented in the VASP software 77 and the cutoff energy for plan e waves was set to 400 eV. To minimize the artificial interaction of the dopant, 221 supercell containing 96 atoms was used for all calculations and a 444 Monkhorst Pack 78 k point mesh was used. The Bi 6s 2 6p 3 I 5s 2 5p 5 and Sb 5s 2 5p 3 were treated as valence electrons. The equilibrium lattice parameters of undoped BiI 3 were found to be a = 7.408 and c = 20.035 , which agree well with values of a = 7.519 and c = 20.721 determined by X RD 30 Four different sites Bi site I site C site of Bi layer (C site) and interstitial site in between I layers ( I I I site) were considered a s the possible Sb incorporation site s in the BiI 3 supercell. For a better description of the band structure of Sb doped system, spin orbit coupling was considered. In the next chapter, the effect of extrinsic impurities on electrical properties of BiI 3 si ngle crystal detectors will be discussed.
61 CHAPTER 4 ENHANCED ELECTRICAL PROPERTIES AND RADIATION RESPONSE OF ULTRA PURE BISMUTH TRI IODIDE SINGLE CRYSTAL DETECTORS 4.1 Introduction It is well known that extrinsic defect s, such as impurities, ha ve signifi cant impact on functional properties of semiconducting material s, and thus controlling impurity level is a key to achieve high performance of fabricated devices. Thus, in this chapter, the effect of impurity level on the electrical properties and radiatio n performance of BiI 3 single crystal detectors will be discussed. 4.2 Ultra High Pure BiI 3 Powder Synthesis by PVT and TMZ Techniques Several techniques such as PVT and TMZ, have been employed to synthesize high pure BiI 3 powder Impurity analysis was performed on PVT and TMZ synthesized BiI 3 powder by using ICP AES, and the results are tabulated in Table 4 1 with impurity levels of commercial BiI 3 powder included for the comparison. Impurity e lements which have concentrations below the detection limit (~1 ppm) of the equipment were marked as ND ( n ot d etect ed ) in the table. The total impurity level s were measured as 7.9 ppm, 26.3 ppm, and 116.9 ppm for the PVT TMZ, and commercial BiI 3 powders, respectively. It is clear that PVT and TMZ are effective techniques to decrease the total impurity level of synthesized BiI 3 powder. Major metallic impurities in commercial BiI 3 powder were Ni (63.73 ppm) and Fe (12.83 ppm). However, these elements were not detected in PVT BiI 3 while detected in TMZ BiI 3 pow der with a very low concentration (1.6 ppm). Boron was observed as a major impurity for both PVT and TMZ BiI 3 powder possibly coming from quartz glass boats and Pyrex glass ampoule.
62 Although the impurity level of TMZ BiI 3 powder is not as low as PVT BiI 3 powder, the TMZ technique has advantages in terms of a much higher yield and lower cost compared to the PVT technique. Specifically, the TMZ technique is able to purify 100 g of powder in a week, while the PVT technique can yield 1 g of purified BiI 3 powd er in three days. Table 4 1 Impurity concentrations of PVT and TMZ synthesized BiI 3 powder measured by using ICP AES Element PVT (ppm) TMZ (ppm) Com. a) (ppm) Element PVT (ppm) TMZ (ppm) Com. (ppm) Cu 0.3 ND b) ND Cr ND ND ND Ag ND 4.2 4.47 Mn ND ND N D Pb ND ND ND Co ND 0.1 1.57 Ni ND 1.6 63.73 Zn ND ND ND Fe ND ND 12.83 As 1.3 2.7 ND Na ND ND ND Se ND ND ND Si ND ND ND Sr ND ND ND Li ND ND ND Mo 0.1 ND 0.75 Be ND ND ND Cd 0.1 0.1 0.1 Mg 1.1 ND 0.58 Sb ND ND ND Ca ND ND 6.59 Tl ND ND ND K ND ND 22.18 B 4.3 17.5 2.36 Ti ND ND 0.73 Al ND ND 1.03 V ND ND ND Ba 0.7 0.1 ND a): commercial, b): not detected 4.3 Electrical Properties and Radiation Response of Ultra Pure BiI 3 Single Crystal Detectors BiI 3 single crystal detectors were fabricated u sing PVT BiI 3 powder, and electrical properties, such as resistivity and leakage current, and radiation response were characterized. It is important to note that the detectors fabricated by using TMZ BiI 3 powder did not show any improvement in terms of el ectrical properties and radiation response such that further discussion will be focused solely on PVT BiI 3 single crystal detectors.
63 In order to determine the resistivity of the fabricated detector, I V characterization was performed on BiI 3 detectors wi th Au electrode (~ 100 nm) deposited by using sputtering, and the result is represented in Figure 4 1 The detectors showed Ohmic IV characteristic, and the resistivity of PVT BiI 3 detector was determined as 2.27 x 10 10 which is much higher (two orders of magnitude) than that of BiI 3 detector fabricated from commercial BiI 3 powder (2.55 x 10 8 ). This clearly demonstrates that extrinsic defects (i.e., metallic impurities) can significantly affect electrical properties of the device. Higher resistivity is favored for gamma ray semiconductor detectors, since it ensures lower dark current flowing through the device and therefore high signal / noisy ratio can be expected. Figure 4 1 IV characteristics of the PVT and commercial BiI 3 detectors. Leaka ge current was measured for the detectors to determine the variation of dark current flowing through the device under a constant bias for prolonged times. Figure 4 2 shows the results for leakage current measurements of the detectors fabricated form PVT a nd commercial BiI 3 crystals. It is shown in Figure 4 2 that the
64 leakage current of the PVT BiI 3 detector is orders of magnitude lower compared to that of the commercial BiI 3 detector. In addition, it is important to note that the ultra pure PVT BiI 3 dete ctor performed consistently during subsequent tests and had low leakage current compared to the commercial BiI 3 detector. Figure 4 2 Leakage current comparison between PVT and commercial BiI 3 detectors. Radiation responses of the ultra pure PVT BiI 3 detector were recorded at room temperature using an Americium 241 ( 241 Am) particle source (Figure 4 3 ). The full width at half maximum (FWHM) for the spectrum was calculated as 1.46 MeV, which is 34 % improved energy resolution than previously reported value for commercial BiI 3 detectors. 29 However, the detector stopped working after approximately 1 ~ 2 h possibly due to background noisy and the leakage current in the detector overwhelmed the output signal preventing any true radiation sig nal from being recorded. This might be due to polarization effect of intrinsic defects in BiI 3 The polarized intrinsic defects may build internal electric field in opposite direction to the external field such that prevents radiation generated electron hole pairs from being collected through the
65 circuit. The possible solution via defect engineering to overcome this problem will be further investigated in the following chapter. Figure 4 3 241 particle spectrum recorded from the ultra pure PVT BiI 3 detector at room temperature. 4.4 Conclusion The effect of impurity level on electrical properties and radiation response was investigated for BiI 3 semiconductor detectors. Material purificatio n techniques such as PVT and TMZ were applied to synthesize high pure BiI 3 starting materials. Ultra pure BiI 3 powder was successfully synthesized through PVT method, which has total impurity level less than 10 ppm. BiI 3 single crystals were grown by usi ng ultra high pure PVT BiI 3 powder. The fabricated detectors show resistivity up to 10 10 orders of magnitude higher than that of BiI 3 detectors fabricated from commercially purchased powder (total impurity level of approximately 200 ppm). In addition, the leakage current of ultra high pure BiI 3 detectors was significantly lower (two orders of magnitude) than that of commercial BiI 3 detectors. Furthermore, the energy resolution
66 for radiation detection of PVT detector (1.4 MeV) was improved up to ~ 34 % compared to commercial BiI 3 detector (2.2 MeV), while PVT BiI 3 detector showed polarization effect after a few hours under bias possibly due to high concentration of intrinsic defects and migration of those charged defects in the material.
67 CHAPTER 5 DEFECT MODELING AND ENGINEERING OF BISMUT H TRI IODIDE SINGLE CRYSTALS: ENHANCED ELECTRICAL AND RADIATION DETECTION PERFORMANCE 1 5.1 Introduction In this chapter, firstly, theoretical defect modeling for BiI 3 system will be discussed via extrinsic donor doping approach. Then, defect engineering strategies will be proposed in order to reduce high polarization effect in BiI 3 detector owing to high concentration and migration of iodine vacancies, and finally to achieve enhanced electrical properties and radiation detection performance of BiI 3 compo und as a promising radiation detector material. 5.2 Defect Modeling of BiI 3 Single Crystals: Donor (Te) Doped BiI 3 Tuller and Bishop et. al. has proved that donor doping can complete theoretical defect modeling for certain material. 24 79 Thus, i n order to understand and establish an intrinsic defect model for BiI 3 Te has been selected as a donor dopant, based o n Humm Rothery rules ( Table 5 1 ) Table 5 1 Te 4+ substitutional donor dopant for Bi 3+ Coordination number a) r ion (%) b) E neg Te 4+ 6 5.13 0.08 a) r ion : I onic radius difference, and b) E neg : electronega t ivity differences between Te 4+ and Bi 3+ Te wi ll introduce compensating less mobile intrinsic defects (bismuth vacancies) by substituting for bismuth atoms ( equations (5 1 ) and (5 2)) 1 Reprinted with the permission from Journal of Physical Chemistry C in press. Unpublish ed work copyright 2014 American Chemical Society. T Engineering of BiI 3 Single Crystal s: Enhanced Electrical and Radiation Performance for Room Temperature Gamma Ray Detect i o HyukSu Han, Minki Hong, Sasmit S. Gokhale, Susan B. Sinnott, K elly Jordan, James E. Baciak, and Juan C. Nino It is reprinted with permission from Wiley Blackwell doi: 10.1021/jp411201k
68 (5 1) For the overall charge neutrality, equation (5 2) should be satisfied. (5 2) Th us extrinsic Te dopants will increase the concentration of intrinsic bismuth vacancies in BiI 3 Furthermore, t otal ionic conductivity of BiI 3 where Schottky defect s dominate defect equilibria can be written as, (5 3) and the equilibrium constant for Schottky defect reaction in BiI 3 can be given as, (5 4) where is the enthalpy for Schottky defect formation in BiI 3 lattice and, and represent the concentrations of Bi and I lattices, respectively. In a dono r doped BiI 3 system, two separate regions ( intermediate and low temperature) a t the extrinsic region of the Log( T) vs. 1/T plot will appear due to increased bismuth vacancies introduced by Te dopants. For the extrinsic low temperatur e region, the slope in Log( T) vs. 1/T plot should be equal to since the concentration of bismuth vacancies is much larger than iodine vacancies neglect ing the second term in equation (5 3). Also, the concentr ation of bismuth vacancies is equal to one third of the concentration of tellurium dopants by equation (5 2) (5 5)
69 However, at the extrinsic intermediate temperature region, even though the concentration of bismuth vacancies is lar ger than iodine vacancies, iodine vacancies can dominate the total ionic conductivity of BiI 3 owing to higher mobility at the elevated temperature when compared to bismuth vacancies As such, the concentration of iodine vacancies can be calculated by subs tituting equation (5 5) into (5 4), (5 6) Thus, the slope of Log( T) vs. 1/T plot at the extrinsic intermediate temperature region will be equal to by substituting equation (5 6) into (5 3), neglect ing the f irst term in equation (5 3) Combined with the information from the intrinsic high temperature region where intrinsic Schottky defect s dominate defect formation all the necessary variables for establishing a defect model of BiI 3 ( i.e. Schottky defect for mation energy, migration energies, concentrations, and mobilit ies for bismuth and iodine vacancies) can be extracted from the Log( T) vs. 1/T plot. Table 5 2 summarizes the dominant defect types and concentrations for BiI 3 at separate d extrinsic regions due to Te doping. Figure 5 1 illustrates anticipated slopes in the Log( T) vs. 1/T plot for donor (Te) doped BiI 3 Table 5 2 Summary of the dominant defect type and concentration for and the slope in the Log( T) vs. 1/T plot high temp region i ntermediate temp r egion low temp region Dominant d efect Concentration Slope i n Log( T) vs.
70 Figure 5 1 Anticipated slopes in Log( T) vs. Temp plot for donor doped BiI 3 Based on the hypothesis above, efforts were made on single crystal growth of Te doped BiI 3 Tellurium (IV) iodide (99.9 % Te I 4 ) was purchased from Strem Chemicals, Inc. and doped into commercial BiI 3 powder with the concentration s of 40 and 120 ppm. Th ese amounts w ere determined by following H. I. Tuller s work on the defect modeling for TlBr single crystal. 79 Figure 5 2 shows a successfully grown Te doped BiI 3 single crystal extracted from the as grown polycrystalline boule Figure 5 2 Te doped BiI 3 single crystal. Photo (s) courtesy of HyukSu Han. In order to investigate Te concentrations in donor doped BiI 3 the grown poly crystalline bulk s w ere cut i n to several 10 mm segments and dissolved in HCl to perform
71 ICP AES measurements Figure 5 3 shows ICP AES results for the 40 ppm and 120 ppm Te doped BiI 3 single crystals (TBI 40 and TBI 120), respectively. It was observed that for TBI 40, Te was homogen eously distributed with the average concentration of 31.3 4.27 ppm along the crystal. However, relatively inhomogeneous dopant distribution was found in TBI 120, and dopants were segregated at the ends of the crystal while the middle of the crystal has an initially intended dopant concentration (120 ppm). Thus, all the samples were extracted from the middle of the crystal for the ionic conductivity measurements Figure 5 3 ICP analysis for Te concentrations along the distan ce from the tip. Impedance analyses were performed on TBI crystals in order to extract formation and migration energies of intrinsic defects in BiI 3 and Figure 5 4 shows Arrhenius plots for measured grain ionic conductivities from TBI 40 and TBI 120 resp ectively. Activation energy transition was observed at ~80C due to activated Schottky defect formation at elevated temperature. However activation energy transition at
72 intermediate temperature region was not presented for the ionic conductivity data of TBI 40 and TBI 120, which was expected based on the defect model ing for donor doped BiI 3 This might be due to the low concentration of the compensating defects (bismuth vacancy for this case) introduced by extrinsic dopants (Te) and thus ionic conduc tion at low temperature may still be dominated by iodine vacancies with a higher mobility and concentration compared to bismuth vacancy. Therefore, BiI 3 single crystals dope with higher donor dopant concentration should be grown and tested for impedance m easurements in order to investigate the characteristics of native defects in BiI 3 Figure 5 4 Ionic conductivity of Te doped BiI 3 single crystals as a function of temperature. 5.3 Defect Engineering of BiI 3 Single Crystal s: Sb doped BiI 3 Single Crys tal Although all X I 3 ( X = As, Sb, and Bi) compounds are isomorphic (space group ), it has been shown that the molecular characteristic s of the X I 3 units increase from
73 BiI 3 to SbI 3 and to AsI 3 As a result, AsI 3 can have three fold c oordination and the molecular geometry is almost the same as that of gas phase AsI 3 while BiI 3 exhibits nearly perfect six fold coordination of the metal and the molecular character is lost in BiI 3 The molecular characteristic s of SbI 3 are in between t h ose of BiI 3 and AsI 3 66 67 This trend can be explained by the different types of bonding available for the me tal atom. Namely, t he electronic configuration of group V metals ( ns 2 np 3 ) enables X I 3 compounds to have either ionic or covalent bonding characteristics. For the perfect ionic case, only three p electrons participate in the bonding and donate to the six iodine ions is neighboring octahedral sites However, in the pure covalent case, the metal atoms tend to have complete sp 3 hybridization and form three covalent bonds with iodine atoms which leaves one lone pair of electrons belonging to the metal atom. As a consequence, although isomorphic in the solid state X I 3 compounds exhibit distinct molecular characteristic s. Taking advantage of this, the dopant of group V metal ( i.e Sb) was dop ed into the BiI 3 lattice to create covalent bonds with iodine atoms and thus successfully inhibit the formation and the migration of iodine vacancies resulting in reduced ionic conductivity in BiI 3 As a result improved electrical properties as a high energy radiation detector, such as higher resistivity lower leakage cu rrent, and less polarization effect can be expected in Sb doped BiI 3 (SBI) system. A series of doping levels such as 0.5 at%, 1.0 at%, and 5.0 at%, were doped into BiI 3 and SBI single crystals were grown using previously optimized growth conditions. 28 29 The experimental XRD patterns of BiI 3 and S B I single crystals with the theoretical pattern are shown i n Figure 5 5 Given the crystal habit, t he XRD patterns only consist of reflections parallel to the (001) plane such as (003), (006), (009), and
74 (0012) and t he observed peak positions for the BiI 3 single crystal were in good agre ement with the published powder diffraction data for a rhombohedra l crystal structure (JCPDS PDF# 48 1795). Peak shifts and intensity decrease s were found in the XRD pattern of SBI single crystal possibly due to the lattice parameter change when Sb substi tutes Bi atoms and the difference of atomic scattering factors between Bi and Sb atoms, respectively ( Figure 5 5 ) In addition, the XRD pattern verified that no secondary phase exists in the SBI single crystal. Figure 5 5 XRD patterns for BiI 3 and SBI single crystals. T he c hemical composition of grown BiI 3 and SBI single crystals w as investigated by ICP AES measurements Table 5 3 shows ICP AES results for BiI 3 and SBI (Sb = 5 at%) single crystals. Total impurity lev els of 137.5 ppm and 74.3 ppm were observed for BiI 3 and SBI single crystals, respectively. The m ajor impurity for BiI 3 was Cu with a concentration of 130.4 ppm, while Ca with a concentration of 21.6 ppm was the major impurity for SBI. Furthermore, ~5 % of Sb was detected in the SBI single crystal while Sb was not detectable in the BiI 3 single crystal. This demonstrates that Sb is successfully doped into the BiI 3 single crystal during the crystal growth process. It is
75 important to note that although the starting materials were the same, different impurity levels were observed for BiI 3 and SBI crystals since each impurity element has different solubility in BiI 3 and SBI lattices, respectively Table 5 3 ICP AES analysis on (a) BiI 3 and (b) SBI single cr ystals (a) BiI 3 element c. a) (ppm) element c. (ppm) element c. (ppm) element c. (ppm) Cu 130.4 Mg 1.7 Cr ND Mo ND Pb ND b) Ca ND Ti 0.4 Zn ND Ni 0.4 Fe 4.2 Co 0.4 Sb ND (b) SBI element c. (ppm) element c. (ppm) element c. (ppm) element c. (ppm) Cu 1 8.6 Mg ND Cr ND Mo ND Pb ND Ca 21.6 Ti 4.3 Zn 11.7 Ni 9.3 Fe 8.8 Co ND Sb 5 (%) a) Concentration, b) Not detectable (impurity level under 1 ppm ) To further support the defect engineering strategy, i t is necessary to determine the precise incorporation si te of the Sb dopants and investigate the bonding nature between Sb and neighboring ions. It has been proposed by Motsnyi et al. that dopants might reside mainly between the layers of BiI 3 single crystal forming covalent bonds between neighboring layers, w hich gives rise to covalent bridges 80 However there has not been any detailed fundamental research on the incorporation of dopants in the BiI 3 lattice. To this end the SBI system was modeled by DFT and the defect formation energy (DFE) of Sb at four possible incorporation sites w as calculated The DFE of a point defect or impurity ( ) wi th charge q E f q ) is defined a s follows: ( 5 7 )
76 where E total q ) is the total energy of the system with a defect E total (BiI 3 bulk) is the total energy of the system in the absence of a defect, and i is the chemical potential of species i which is added (n > 0) or removed (n < 0) from the system. E v and E f indicate the valence band maximum (VBM) of the perfect supercell and the Fermi energy referenced to the VBM, respectively. Lastly, is the correction of the VBM deviation in the defective supercell. The resulting DFEs are de picted in Figure 5 6 Bi rich conditions, where the chemical potential of Bi reaches the maximum value ( Bi Bi,metal ), were considered given that iodine compounds are normally iodine deficient during growth due to the high vap or pressure of iodine and BiI 3 dissociation into Bi and I 2 at 250 ~ 300 C. 81 83 The DFE of is predicted to have the lowest energy and thu s, Sb will preferably reside in the Bi site ( Figure 5 6 ( B ) ). However, BiI 3 is normally under iodine deficient conditions that impose high concentration s of iodine vacancies in the system. Figure 5 6 ( A ) shows that bismuth vacancy formation is considerab ly suppressed over a wide range of E f under iodine deficient conditions resulting in the significantly higher concentration of V I than V Bi in the BiI 3 lattice. It is important to note that Bercha et al. have shown that, in theory, Sb atoms may not be inc orporated in cation vacancy in the layered BiI 3 compound. 84 Interestingly, the DFE calculations in this wor k indicate that , and have nearly the same DFEs and are in compet ition with each other for Fermi energ ies of 0.4 eV of less However, for Fermi energ ies above this value is predicted to have the lowest DFEs compared to those of other charged defects except for the ( Figure 5
77 6 ( B ) ). It can thus be hypothesized that under Bi rich conditions the Bi site and I site become the preferred Sb incor poration sites within the BiI 3 lattice. A B Figure 5 6 Calculated formation energies as a function of Fermi energy for A ) iodine and bismuth vacancies and B ) Sb Bi Sb I Sb C and Sb I I in BiI 3 under Bi rich conditions. DFT calculation results for c ases when Sb incorporates in the Bi site and the I site are represented in Figure 5 7 ( A ) and ( B ), respectively. The results predict that the resulting atomic configuration is almost identical to the pure BiI 3 system when Sb resides on the Bi site while Sb forms a dimer with a neighboring iodine when it incorporates in the I site. Since Sb 3+ has a smaller ionic radius (r sb = 90 pm) than Bi 3+ (r Bi = 117 pm), it can thus be expected that the shorter and stronger Sb I bonds, compared to the Bi I bonds, shou ld be formed when Sb incorporates in Bi site. 85 Table 5 indicates the bond lengths of the Sb and Bi ions with the surrounding I ions, and clearly demonstrates that the above statement holds true for this system In addition, Sb I dimer formation was predicted by the DFT calculations with a shorter bond length
78 of ~2.8 compared to that of Bi I at equilibrium (~3.06 ) when Sb incorporates in the I site ( Figure 5 6 ( B ) ). Figure 5 7 ( C ) represents the total density of state (DOS) for pure BiI 3 SBI (Sb incorporated in Bi site), and SBI (Sb incorporated in I site), respectively. The band gap of ~1.3 eV was predicted for pure BiI 3 which is a slight underestimation (~15 %) relative to previous calculations; this is not un expected as the local density approximation (LDA) pseudopotentials were used in these calculations 86 88 It is worth mentioning that spin orbit coupling is required in order to establish the proper distribution of energy states near the band gap. The total DOS of Sb doped BiI 3 was predicted to be nearly identical to that of u ndoped BiI 3 when Sb incorporates in the Bi site given that no additional energy states were created. In contrast, the total DOS for Sb doped BiI 3 at I site shows that energy states were created just ab ove the valance band and just below the conduction band; the energy states near the valance band are occupied by electrons, while the energy states near the conduction band are empty. The partial DOS analysis also confirmed that the energy states just abo ve the valance band are due to Sb and the surrounding Bi and I ions while the states near the conduction band originated purely from the Sb I dimer. Moreover, it was determined that a clear hybridization of Sb 5p and I 5p exists in the energy state of Sb I dimer implying the bonding nature of Sb I dimer is highly covalent ( Figure 5 7 ( C ) ). T he difference between the states formed by Sb I dimer and valance band (~1.0 eV) is almost equal to the energy of the observed peak from the photoluminescence spectra of Sb doped BiI 3 single crystal. 80
79 A B C Figure 5 7 Crystal structures of A ) SBI (Sb at Bi site) B ) SBI (Sb a t I site) and C ) total density of states (DOS) for SBI (Sb at Bi site), SBI (Sb at I site), and pure BiI 3
80 Table 5 4 Sb I and Bi I bond lengths in Sb incorporated BiI 3 on Bi site. bond length () bond length () Sb 1 3.012 Bi 7 3.049 Sb 2 3.011 Bi 3 3.047 Sb 3 3.011 Bi 8 3.052 Sb 4 2.993 Bi 9 3.064 Sb 5 2.994 Bi 10 3.068 Sb 6 2.994 Bi 5 3.062 Recalling, the i onic conductivity ( ) of the material is related to the concentration of the vacancies and/or interstitials through e quation ( 2 3 6 ), and t he mobility of the charge carrier i can be defined by, ( 5 8 ) where and are the pre exponential factor and the migration energy for the charge carrier i Given that Schottky defect s are the dominant defect complexes and iodine vacancy dominates ionic conduction in BiI 3 ionic conductivity of BiI 3 can be written as, 32 34 35 ( 5 9 ) Moreover, the defect reaction when Sb is incorporated in the Bi site (most likely substitutional) and the I site (most likely iodine vacancy) can be respectively desc ribed by, (5 10 ) ( 5 11 )
81 As demonstrated by the DFT calculations, substituted Sb atoms may be able to increase the formation energy of iodine vacancies by creating stronger and shorter bonds between Sb and neighboring I ions when Sb incorporates in th e Bi site ( equation (5 10) ). It has also been demonstrated that for CdTeZn (CZT) a certain amount of Zn dopants in CdTe will increase defect formation energies by forming less ionic Zn Te bonds. Furthermore, according to equation (5 11) Sb will effecti vely decrease the number of jumping sites for iodine ions as they occupy V I sites when Sb resides on the I site. Through these defect engineering strategies, the ionic conductivity and dark current in BiI 3 are expected to be significantly reduced. 5.4 Enh anced Electrical and Radiation Performance of Sb doped BiI 3 Single Crystal Detectors for Room Temperature Gamma Ray Detect i o n To validate this defect engineering approach, t he resistivity of BiI 3 and SBI single crystals were measured using a microprobe sta tion and the results are presented in Figure 5 8 ( A ) The measured resistivity of the 5.0 at% and 0.5 at% doped SBI single crystals w as 2.63 10 9 and 2.44 10 9 respecti vely, while 1.0 at% doped SBI crystal shows almost identical resistivity value with undoped BiI 3 ( 1.45 10 8 ). However, a s expected, SBI (5.0 and 0.5 at%) has a higher resistivity (about one magnitude order) compared to that of BiI 3 s ingle crystal. Given the promising detection performance previously shown in undoped BiI 3 S B I can be seen as an enhanced candidate for radiation detect ion since higher resistivity leads to lower dark currents of the device resulting in sensitive sensor response. It is also important to understand the behavior of single crystal detector under a constant bias for a prolonged period of time. As such, leakage current was measured in order to investigate the variation of leakage current flowing through BiI 3 and SBI single
82 crystal detectors. Figure 5 8 ( B ) shows the comparison of the leakage current in BiI 3 and SBI single crystal (5.0, 1.0, and 0.5 at%) detectors. As illustrated in Figure 5 8 ( B ) the leakage current in SBI single crystal detector regardle ss of doping concentration is orders of magnitude lower (about a factor of 10 4 ) than that in the undoped BiI 3 single crystal detector. In addition, it was observed that leakage current of SBI single crystal detector was initially increased up to 10 1 A/ cm 2 and then stabilized to 10 2 A/cm 2 after 150 min. In contrast, leakage current of BiI 3 single crystal detector kep t increasing above 10 A/cm 2 after 150 min ( Figure 5 8 ( B ) ). This result, combined with the enhanced dark resistivity of SBI, demonstrat es that V I migration can be effectively controlled by defect engineering and in particular by Sb dop ing strategies. A detector with the higher resistivity and the lower leakage current should have a better response to radiation since a lower dark current and a higher signal to noise ratio can be achieved, and this is critically important for radiation spectroscopy measurements. A B Figure 5 8. Electrical properties of BiI 3 and SBI single crystal detectors A ) IV characterization B ) leakage curr ent measurements.
83 Figure 5 9 ( B ) shows the radiation response of SBI single crystal detector to a 241 Am source at room temperature. In addition, i t is also worth noting that altho ugh the SBI crystals with different doping concentrations (5.0, 1.0, and 0.5 at%) produced distinct particle spectrums, SBI ( 5.0 at% ) has been selected as the material discussed here for radiation response since it demonstrated superior radiation response in terms of long term detector stability. Here, particle instead of gamma rays has been utilized to characterize detector performances, since particle has a well defined range and, thus the electrons are generated at the same depth within the detector. Therefore, using particle allows for a more direct measurement of the mobility and lifetime of electrons. A distinct spectrum was observed after 5 minutes under electric field of 532 V/cm ( Figure 5 9 ( B ) ). In our previous work, it was reported that spectrum for undoped BiI 3 single crystal detector disappeared onl y after 90 minutes under bias due to the high leakage current and the polarization effect in the detector at prolonged time ( Figure 5 9 ( A ) ). 28 However, in the case of SBI single crystal detector, while the count rate decreased po ssibly due to field variations in the detector, a clear spectrum was attained after 8 hours under electric field of 532 V/cm ( Figure 5 9 ( B ) ). It is a well known phenomenon that the polarization effect is the main reason for the observed degradation of the detectors over time under steady state operating con ditions. This result thus clearly demonstrates that the polarization effect in SBI single crystal detector is significantly lower with respect to that in undoped BiI 3 single crystal detector resulting in enhanced detector stability of SBI. A relative p ea k shift was observed in the spectrum measured after 8 hours which is commonly observed after biasing a compound semiconductor detector for the first time spectrum measurements. 89
84 Resolutio n of the SBI detector was calculated as 62.66 % and 32.96 % for 5 min and 8 hours after biasing, respectively Interestingly, resolution was improved after time periods possibly due to reduced and stabilized leakage currents in the detector. Al though res olution of SBI detector is not as good as that of commercial CZT detectors, it is improved compared to previously reported value for undoped BiI 3 detectors (~40.15 %) 29 and shows the potential of the material to function as a radiation detec tor. A B Figure 5 9 241 Am alpha source radiation response tests for A ) BiI 3 and B ) SBI (5.0 at%) single crystal detectors under bias at different time periods. Ch.cnt and FWHM indicate channel number at the centroid and full width half max, r espectively. Finally, t he electron mobility ( e ) of the SBI detector was estimated by measuring the drift time (t drift ) for electrons from an interaction near the cathode since the electrons will drift the entire thickness of the detector. The e can be calculated by ( 5 12 ) D and E are the detector thickness ( ~ 0.4 cm) and the applied electric field (532.5 V/cm) respectively 241 Am alpha source was used to irradiate the cathode of the
85 detector i n order to estimate the drift time and the electron r ise time was recorded via output pulses. Figure 5 10 illustrates the recorded out pulse of the SBI detector using 241 Am alpha source. Several measurements (~10 times) of rise time were performed in order to enhance reliability of the data, and an average value of 785 163 ns was calculated It should be noted that the electron rise time is almost equal to the electron drift time since depth penetration of the particle is very small. Therefore, the electron mobility of the SBI detector was estimated as 1000 200 cm 2 /Vs. This value of electron mobility is on par with that of CZT, the leading room temperature detector material for and X rays. 90 Furthermore, defect engineering strategies in SBI have achieved the material with enhanced electron mobility (~70 %) compared to that of undo ped BiI 3 detector (~600 cm 2 /Vs) 91 Here, the hole mobility was not measured due to the fact that wide bandgap iodine compounds experience severe hole trapping. Ho wever, a potential low hole mobility would not degrade the detector performance of SBI since single polarity charge sensing technique can be employed to overcome this problem (i.e. hole trapping). 92 93 The defect engineering strategies proposed in this Chapter demonstrate s the ability to tune enhanced electrical properties and radiation response of the material, thou gh additional factors such as surface treatments and purification of starting materials are known to affect radiation detection properties of semiconductor detectors. The results of these experiments imply that controlling detrimental defects for radiatio n response should be considered for the search for the semiconductor materials with the superior detection properties. In particular, the present search for superior iodine compounds in
86 novel semiconductor detectors should consider iodine vacancy control as a dominant factor to realize enhanced electrical and radiation detection properties. Figure 5 10 Electron rise time of the SBI detector from alpha particles. 5.5 Conclusion Theoretical defect models for BiI 3 was established for BiI 3 system throu gh donor doping approach, and efforts were made on extracting thermodynamical parameters for the native defects in BiI 3 such as formation energy, migration energy, and mobility, by measuring ionic conductivities of TBI crystals as a function of temperatur e. Based on defect engineering strategies the e lectrical properties, such as high resistivity and low leakage current, and radiation response were significantly improved in SBI single crystal detectors compared to undoped BiI 3 single crystal detector s. Leakage current measurements and particle radiation response demonstrate that time dependent detect ion properties were also remarkably enhanced in S BI single crystal detector. In addition, the electron mobility of S BI detector was significantly enhanced (~70 %) compared to that of undo ped BiI 3 detector. Density functional theory calculations and defect modeling verified that Sb dopants in BiI 3 can effectively reduce the formation and
87 the migration of iodine vacancies which is consistent with the enhanced electrical and radiation detect ing properties in the SBI single crystal detector. Therefore SBI single crystal s can now be considered realizable material s for gamma ray detector s surpassing undoped BiI 3 based on enhanced electrical properties and radiation response of the detector.
88 CHAPTER 6 COLOSSAL PERMITTIVITY IN FAST FIRED BARIUM TITANATE CERAMICS 2 6.1 Introduction In this chapter, different processing techniques, such as SPS and MWS, will be investigated in order to induce CP in BT ceramics. Processing variables, such a s sintering and annealing conditions, will be optimized to improve dielectric properties of fast fired BT ceramics, i.e. high relative permittivity with low dielectric loss. SPS and MWS are nonconventional fast firing techniques that has been used to sint er several electronic ceramics such as Pb(Zr x Ti 1 x )O 3 (PZT) and Sr doped PZT, and YBa 2 Cu 3 O 7 x 94 96 SPS and MWS also have several advantages over conventional si ntering such as rapid heating rate and short sintering time. 97 These unique aspects of fast firing techniques gives rise to severe reduction atmosphere during sintering, which results in high concentration of point defects, such as oxygen vacancies, electrons, and Ti 3+ in the sintered ceramics. Thus, the induced point defects in the fast fired ceramic significantly affect dielectric properties of the material compared to the conventionally sintered ceramic. In this study, we demonstrate that CP can be achieved in BT ceramics when sintered by SPS and MWS. In addition, post sintering annealing conditions were optimized with the aim to reduce the dielectric loss, but to still maintain the high relative permittivity values in BT ceramics sintered by MWS technique. 2 Reprinted with the permission from Wiley Blackwell Journal of American Ceramic Society, 96  485 490 (2013) al Permittivity in Microwave Sintered J. L. Jones, and J.C. Nino.
89 6.2 Microstr ucture, Density, and Phase Purity of Pressureless, Spark Plasma, and Microwave Sintered BaTiO 3 Figure 6 1 shows SEM images of the starting nano meter sized BT powder and of the fractured surfaces of the as sintered BT ceramics synthesized by PS, SPS, and MW S. Figure 6 1 ( A ) confirms that the particle size of starting BT powder was below 100 nm. Average grain sizes of 0.21 0.01 m, 0.67 0.11 m, 1.29 0.08 m, and 10.30 1.39 m were estimated on as sintered SPS (900C) SPS (1050C) MW S (1320 C ), and PS (13 5 0C) BT samples respectively (Figure 6 1) For the grain size measurements, the mean lineal intercept method by ASTM E112 standard, even though this method can only be used for planar grains, was applied on the SEM images from the fractured surfaces as the SPS and MWS samples were too brittle to investigate nice planar microstructures. 98 Overall, the SEM micrographs of the fractured surfaces of the sintered BT ceramics revealed that, as expected, the BT ceramic sintered by SPS at 900C has the smallest grain size whereas the largest grain size BT ceramic was ob tained when sintered by PS at 1350C. The grain size was observed to increase from SPS BT (900C) to S PS BT (1050C) to MWS BT (1320C) to PS BT (1350C). Table 6 1 shows that the density of as sintered SPS BT ceramics vary from 93.89% to 98.88% of theore tical value (6.02 g/cm 3 ) for sintering temperatures of 900C and 1050C, respectively. The a verage d ensity of the as sintered MW S samples was 94.15 % and for the PS BT, 97.16% of theoretical density was achieved.
90 Figure 6 1. SEM micrographs of A ) star ting nano BT powder and fractured surfaces B ) SPS 900C C ) SPS 1050C D ) MWS 1320C E ) PS 1350C.
91 It was observed that all the as sintered MWS BT (1320 C ) ceramic s as well as the SPS BT (1050C) ceramic were dark blue wh ereas the as sintered SPS BT (900 C) and PS BT (1350 C ) ceramics were slightly yellowish or white in color. This color change in the as sintered MWS and SPS BT (dark blue) ceramics compared to the starting BT powder (white) can be attributed to the reduction in BT during sintering. 38 As mentioned previously, post sintering annealing treatments were conducted for the MW S samples in the temperature range of 9 0 0C ~ 1250C in air. It was obse rved that color of the BT ceramics started to change gradually upon annealing and the final color after annealing was dependent on the annealing temperature and time. As the annealing temperature increased from 9 0 0C to 1250C, color of the BT samples gra dually changed from dark blue to cream color to finally almost white at 1250C when annealed for 24 hours. Figure 6 2 illustrates the gradual change in color of the MWS BT ceramics as a result of annealing process. In addition, s intered density of the MW S BT ceramics increased upon post sintering annealing and the maximum density achieved after annealing at 1200 C for 24 hours was 96.30% of theoretical value ( Table 6 1 ). These gradual color change and increased density upon annealing confirm that the reo xidation process with substantial elimination of oxygen vacancies has occurred during annealing process. Figure 6 2. MWS BT annealed in air at different temperatures and times (1050C for 12 h, 1100C for 24 h,1250C for 24 h from left to right).
92 Ta ble 6 1 Summary of sintering conditions densities, and dielectric properties (secondary phase (5 ~ 8 %) of Ba 6 Ti 17 O 40 x was included for d ensity calculation of as sintered MWS. Sintering technique Sintering temp (C) / time Heating rate Density (%) Gra in size ( m) MWS 1320 / 30 min 45C / min 94.15 (as sintered) 96.30 (annealed) 1.29 0.08 SPS 900 / 5 m in 50C / min 93.89 (as sintered) 0.21 0.01 1050 / 5 min 98.88 (as sintered) 0.67 0.11 PS 1350 / 2 hrs 3C / min 97 16 (as sintered) 10.30 1.39 Figure 6 3 shows the XRD patterns of the starting BT powder and BT ceramics (before and after annealing) sintered by MWS, PS and SPS. The XRD pattern of the BT powder was typical of cubic structur e (JCPDS 74 1968), however, a small amount of barium carbonate (BaCO 3 JCPDS 45 1471) was observed in the starting powder, which is commonly used for synthesis of BT pow d ers. The XRD patterns of the as sintered ceramics showed a mixture phase of cubic and tetragonal (JCPDS 05 0626) perovskite peak diffraction intensities. Although the XRD pattern of as sintered PS BT did not reveal any secondary phase, as sintered SPS and MWS BT ceramics showed a secondary phase (5 ~ 8 phase %) of Ba 6 Ti 17 O 40 x (JCPDS 43 0 559) This phase was still present in the MWS BT when annealed at 1050C for 12 h our s, but was completely removed when annealed at 1250C for 24 h our s. The absence of Ba 6 Ti 17 O 40 x phase when annealed at 1250C for 24 hours c ould be attributed to the reox idation of BT ceramic during annealing process. This is consistent with the observed density and color changes.
93 Figure 6 3. XRD patterns of nano powder, as sintered PS, MWS, SPS, and annealed MWS BT. 6.3 Dielectric Properties of Pressureless, Spark Pl asma, and Microwave Sintered BaTiO 3 The dielectric responses of as sintered BT ceramics synthesized by MWS, SPS, and PS, measured in the temperature ranges of room (27C) to 200C and frequency ranges of 1 kHz 1 MHz, are presented in Figure 6 4. BT cerami cs synthesized by PS at 1350C showed maximum relative permittivity value of 6,810 which is typical value of BT ceramics (Figure 6 4 ( A )). It is well known that BT ceramics exhibit a phase transition from cubic to tetragonal near 130C, which is known as Curie temperature (Tc). 99 In the permi ttivity measurements of conventional BT ceramics, this phase transition is evidenced by sharp increase in the permittivity values near the transition temperature. The sharp peak at the transition temperature is clearly observed in the PS samples (see Figu re 6 4 ( A )). Figure 6 4 ( B ) shows dielectric properties of SPS BT
94 (900C) ceramics. It was observed that the maximum relative permittivity of 6,350 was achieved near the phase transition temperature. Although permittivity values were similar to PS BT (1 350C), much higher dielectric losses ( ~ 0.483) were observed above room temperature for SPS BT (900C). The higher dielectric loss in the SPS BT (900C) can be attributed to oxygen vacancies as a result of reduced sintering atmosph ere (vacuum) and rapid heating rate in SPS. 100 Furthermore, it has been shown that the phase transitions in BT ceramics become diffuse in nature when the grain size approaches submicrometer range or lower, 101 and thus the observed diffused peak at the phase transition temperature (Figure 6 4 ( B )) might be due to submicrometer grain size of SPS BT (900C) ceramic. Even though none of PS BT (1350C) and SPS BT (900C) samples show CP, considering the size effect on permittivity of sintered BT ceramics, 101 slightly higher permittivity of the PS BT than the SPS BT (900C) at the phase transition temperature could be attributed to the higher grain size in the former than that in the latter. On the other hand, dielectric measurements of as sintered BT ceramic synthesized by SPS at 1050C revealed CP of the order of 10 5 at room (27C) to 200C (Figure 6 4 ( C )). No discernible peak at the transition tempera ture was observed (Figure 6 4 ( C )), and the increase in the sintering temperature from 900C to 1050C resulted in a significant increase in the dielectric loss. Table 6 2 summarizes relative permittivity and dielectric loss for as sintered MWS, SPS, and PS BT ceramics.
95 Figure 6 4. Dielectric properties of as sintered A ) PS (1350C) B ) SPS (900C) C ) SPS (1050C) D ) MWS (1320C) BT at 1, 10, 100, 500 kHz, and 1 MHz. Barium titanate ceramics synthesized using SPS readily lose oxygen due to reducing atm osphere because of high vacuum during sintering (see Eq. (6 1)). The charge imbalance due to large oxygen deficiency can be compensated by the reduction in Ti 4+ to Ti 3+ (see Eq. (6 2)). 38 (6 1) (6 2)
96 During sintering, several factors such as sintering temperature, atmosphere, and heating and cooling rates can contribute to the extent of oxygen loss in the as sin tered BT ceramics. In the current study, comparing SPS BT (1050C) with SPS BT (900C), it can be suggested that a large amount of charge carriers such as oxygen vacancies, Ti 3+ and electrons in the sintered BT ceramic were readily induced due to higher sintering temperature (1050C), and the observed CP resulted from polarization of the large amount of charge carriers under ac field (Figure 6 4 ( C )). Similar to the SPS BT (1050C), as sintered MWS BT (1320C) also exhibited CP with huge dielectric loss ( ~ 6.901), and the maximum permittivity value achieved during the dielectric measurement was 301,484 at 27C (Figure 6 4 ( D )). In addition, the CP of MWS BT (1320C) was dropped about one order of magnitude at 1 MHz. The resistivit y of the as sintered ceramics (Table 6 2) revealed that samples showing CP (SPS BT (1050C) and MWS BT (1320C)) are semiconductive while PS BT (1350C) and SPS BT (900C) are quite insulating. This also suggests that the induced large amount of charge ca rriers during sintering process may play significant role in achieving CP in BT ceramics. Although SPS sintering was performed under vacuum, MWS was conducted in air atmosphere, but at a significantly higher temperature than SPS. Interestingly, while the BT ceramics were synthesized by PS (1350C) and MWS (1320C) at comparable sintering temperatures and under similar atmosphere (air), colossal permittivity was only exhibited in the latter. Thus, it can be stated that apart from sintering atmosphere (red ucing), high sintering temperature, and rapid heating rate play crucial roles in synthesizing BT ceramics with CP. While Valdez Nava et al. showed CP for nanocrystalline BT ceramics (~300 nm grain size), the current investigation revealed
97 that CP can also be observed in microcrystalline BT ceramics (~1.3 ) synthesized by MWS. 38 Table 6 2 Dielectric properties and bulk resistivity of as sintered MWS, SPS, and PS BT ceramics at room temperature and 1 kHz. Sintering technique Relative Permittivity Dielectric Loss Resistivity (MOhm cm) MWS (1320 C ) 301,484 6.901 0.56 SPS (900 C ) 4 110 0.483 52.9 SPS (1050 C ) 202 000 1.860 0.35 PS (1350 C ) 2 359 0.030 >1,000 6.4 Effect of Annealing on Dielectric Properties of Microwave Sintered BaTiO 3 Ceramics It was observed that the as sintered MWS BT has very high value due to its conductive characteristic (Figure 6 4 ( D )). Th erefore, thermal annealing processes were performed in flowing air (10 SCFH) with the aim to reduce the conductivity and to recover insulating characteristics, although still maintaining high permittivity. Figure 6 5 shows the variation in relative permit tivity and dielectric loss of the MWS BT ceramics as a function of annealing temperature. A significant change in the dielectric properties was observed upon annealing treatment. It can be seen that at the annealing temperature (900C), a significant dec rease in both permittivity and dielectric loss was observed relative to the as sintered MWS BT ceramics (Figure 6 5). Although the as sintered MWS BT showed relative permittivity of ~301,000 and dielectric loss of ~6.9 at 1 kHz and room temperature (27C) a relative permittivity of ~60,500 with dielectric loss of ~0.123 was obtained when annealed at 900C for 12 h (Table 6 3). Although a relative permittivity was decreased from ~60,500 to ~36,000, a low dielectric loss below
98 0.05 was achieved at further increased annealing temperature from 900C to 950C. As the annealing temperature increased from 950C to 1100C, the relative permittivity continued to decrease from ~36,000 to ~15,000 while keeping low dielectric loss values between 0.030 and 0.045 (Fig ure 6 5). Finally, when annealed at 1250C for 24 h, almost fully annealed BT ceramic was obtained with the relative permittivity and dielectric loss of 1,257 and 0.031 (Table 6 3), respectively. Figure 6 5. Effect of annealing on the dielectric prop erties of MWS BT (at 1 kHz and room temperature). As mentioned previously, charge imbalance of reduced BT can be compensated by partial reduction in titanium (Eqs. (6 1) and (6 2)). Therefore, as BT reduced more, it has higher concentration of charge car ries and Ti 3+ / Ti 4+ ratio should be higher than less reduced BT. XPS results showed that the ratio of Ti 3+ / Ti 4+ were 0.023 and 0.016 for MWS BT samples annealed at 950C and 1250C, respectively. Thus, higher Ti 3+ / Ti 4+ ratio of MWS BT sample anneale d at 950C may indicate higher concentration of charge carriers and can explain observed high permittivity. Moreover, resistivity of annealed MWS BT samples was increased from 0.62 to 27 as annealing
99 temperature increased from 900C to 1100C. Resistivity up to 1 was achieved when annealed at 1250C for 24 h (Table 6 3). Thus, it can be stated that post sintering annealing process helps recover insulating properties of reduced semiconductive MWS BT ceramics t o achieve high permittivity with low loss. The dielectric data and bulk resistivity for the annealed MWS B T are summarized in Table 6 3. According to Heywang et. al. noble metals tend to form Schottky junction with the surfaces of semiconducting barium t itanate, and permittivity of the material can be increased via the extrinsic Schottky junction effect. 102 Therefore, to test the effect of Schottky contact on high permittivity of annealed MW BT sample, I V characteristic and dielectric property with various electrodes materials, including Au, Ag, Ni, and A l, were tested. The I V curves showed that Au, Ag, and Ni formed Schottky junction while Al formed Ohmic contact with MWS BT sample annealed at 950C for 12 h. Dielectric properties of the annealed MWS BT with various electrodes are listed in Table 6 4. The Extrinsic component could be estimated for the observed high permittivity of the annealed MWS BT by subtracting permittivity measured with Al electrode ( = 30,665) from measured permittivity with Au electrode ( = 36,055). The calculated result ( = 5,390) implies that ~ 15 % of high permittivity of the annealed MWS BT ( = 36,055) can be attributed to the extrinsic effect of Schottky junction between a metal contact (Au) and the surface of the semiconducting BT. Figure 6 6 shows the temperature dependence of dielectric properties of the MWS BT ceramics for different annealing temperatures. Compared to the as sintered MWS BT ceramics (Figure 6 4 ( D )), sharp changes in relative permittivity values were
100 observed at Tc (~ 130C) of BT ceramics, which is commonly observed in conventionally sintered (PS) BT ceramics. 99 Maximum relatively permittivity of 110,000 was achieved at 129.1C in the MWS BT annealed at 950C, which is 10 times higher than La doped BT ( ~ 10,000). 103 Dielectric losses of annealed BT were below 0.1 between room temperature (27C) to 130C (Figure 6 6). Figure 6 6 Diel ectric properties (at 1 kHz) of MWS BT with different annealing temperatures (1050 C 1000 C and 900 C ) for 12 hrs. Frequency dependence of dielectric properties of annealed MWS BT sample (950C for 12 h) in the temperature range from room temperature ( 27C) to (233C) is presented in Figure 6 7, where it can be seen that high dielectric constant was maintained up to 1 MHz at room temperature. However, it substantially decreased when frequency increased above 1 MHz. In addition, as temperature decrease d, dielectric constant was dropped at lower frequencies and corresponding dielectric loss peaks were observed at the frequencies, which can be explained by Debye relaxation theory. 57 This is consistent with the origin of the high dielectric constant of the annealed MWS sample being the result of the interfacial polarization of charge carriers within
101 electrically inhomogeneous grains and grain boundaries. The precise mechanism of the origin of the CP in fast fired BT ceramics will be discussed n In the next chapter. Figure 6 7. Di electric properties of MWS BT annealed at 950 C for 12 hrs as a function of frequency in the temperature range from room temperature (27 C ) to ( 233 C). Table 6 3 Effect of annealing on dielectric o f MWS BT. Annealing Temp (C) / Time (hrs) Permittivity / tan room temperature Resistivity (MOhm cm) B efore Annealing After Annealing 900 / 12 301 484 / 6.901 60,584 / 0.123 0.62 950 / 12 36 055 / 0.045 1.75 1000 / 12 24 457 / 0.043 4.27 1050 / 12 22 990 / 0.04 5 7.62 110 0 / 24 15 012 / 0.039 27.0 1250 / 24 1 257 / 0.031 >1,000 Table 6 4. Dielectric properties of the annealed MWS BT (950 C for 12 hours) with various electrodes (Au, Ag, Ni, and Al) at room temperature and 1 kHz Electrodes Permittivity tan Au 36 ,055 0.04 Ag 8,189 0.07 Ni 27,543 0.10 Al 30,665 0.06
102 6.5 Conclusion BT ceramics were synthesized using conventional pressureless sintering, SPS sintering, and MWS sintering techniques from nanocrystalline BT powder. For the first time, CP was obser ved in the MWS sintered BT ceramic sintered at 1320 C for 30 min. High sintering temperature and high heating rate of MWS were able to induce colossal permittivity in the sintered ceramics. There was no phase transition effect on permittivity change of a s sintered BT through MWS sintering Post sintering annealing treatments were performed to reoxidize the BT ceramics sintered by MWS w ith the aim to improve insulating characteristic while still maintaining h igh permittivity The best dielectric properti es ( > 10 4 and ~ 0.04) of MWS sintered BT ceramics were achieved after annealing process at temperature range (950 C ~ 1100 C ) for 12 ~ 24 hours
103 CHAPTER 7 DIELECTRIC POLARIZAION MECHAN ISMS AND VARIABLE HOPPING CONDUCTION IN FAST FIRED BARIUM TITANATE CERAMICS 3 7.1 Introduction In this chapter, broadband dielectric spectroscopy analysis was performed on SPS BT ceramics in order to better understand and clarify the polarization mechanism s associated with CP of BT ceramic. Data analyses using Debye relaxation and In addition, bulk conduction mechanism in fast fired barium titanate ceramics was determined based on various hopping conduction models. Fur thermore, the contributions of each polarization mechanism to the colossal permittivity in SPS BT, such as a hopping polarization, IBLC effect and electrode effect, were estimated by using experimental results. 7.2 Microstructure, Density, and XRD of Spa rk Plasma Sintered BaTiO 3 Using Co Precipitated Nanocrystalline Powder ICP AES results show that starting BT powders synthesized by BaCl 2 .2H 2 O and TiCl 3 or BaCl 2 .2H 2 O and TiOCl 2 have Ba / Ti ratios of 0.95 (BT 0.95) and 1.00 (BT 1.00), respectively. Synth esized powders were sintered at 1150C for 3 min. The sintered ceramics, SPS BT 0.95 and SPS BT 1.00 using BT 0.95 and 1.00 powders, were dark blue in color due to the reducing sintering atmosphere (vacuum) during SPS process. 37 The SEM images in F igure 7 1 shows the microstructures of fre shly fractured surfaces of SPS BT 0.95 and SPS BT 1.00 ceramics with a respective average grain size of 5 6 6 nm and 7 2 8 nm Grain sizes were measured by using the mean linear intercept 3 Reprinted with the permission from AIP Publishing LLC., Journal of Applied Physics, 113 024102 (2 013) 3 via Broadband Fritsch, P. Dufour, C. Tenailleau, C. Tuner, and J.C. Nino.
104 method of ASTM E112 standard. 98 As expected due to the fast firing nature of SPS technique, no significant grain growth is observed after sintering and the grain sizes of the sintered ceramics are almost in the same size as the starting powders. Density measurement by Archimedes m ethod reveals that the both sintered ceramics have theoretical densities up to 98 % (Table 7 1). A B Figure 7 1. SEM images for A ) SPS 0.95, and B ) SPS 1.00 sintered ceramics The XRD patterns of the starting powders (BT 1.00 and BT 0.95) and the sintered ceramics (SPS BT 1.00 and SPS BT 0.95) are shown in Figure 7 2. No secondary phases are formed in BT 1.00 and SPS BT 1.00. However, secondary phases of BaTi 2 O 5 (~ 5 vol % relative amount) and Ba 4 Ti 12 O 27 (~ 7 vol % relative amount) are observ ed for the Ti rich BT 0.95 powder and SPS 0.95 ceramic, respectively. It is well known that Ba 4 Ti 12 O 27 phase is formed when Ti rich BT powder is obtained under reduced (low P O2 ) conditions; here the reduction of a part of Ti 4+ to Ti 3+ give rise to the for mation of Ba 4 Ti 2 3+ Ti 10 4+ O 27 104 Besides, both sintered ceramics of SPS BT 0.95 and 1.00 crystallized in a mixture of a cubic and a tetragonal perovskite
105 phases whi ch can be demonstrated by the broadening of the (200) peak. On the contrary, all of the starting powders consisted of a single cubic phase (Figure 7 2). This result is also well in good agreement with previous results for SPS and MWS sintered ceramics in Chapter 6. Chemical composition, particle and grain sizes, theoretical density, crystal lattice, and phase purity for the starting powders and sintered ceramics are summarized in Table 7 1. Figure 7 2. The X ray diffraction patterns for the starting powders (BT 1.00 and 0.95) and the sintered ceramics (SPS BT 1.00 and 0.95). Table 7 1. Summary of chemical composition, grain size and relative density of the sintered ceramics Samples Ba/Ti ratio Grain Size (nm) Density (%) Structure 2 nd Phase
106 SPS BT 0.95 0.95 5 6 6 98 a ) C + b) T Ba 4 Ti 12 O 27 (~ 7 vol %) SPS BT 1.00 1.00 7 2 8 98 C+ T None a) C: cubic perovskite, b) T: tetragonal perovskite. 7.3 Broadband Dielectric Spectroscopy and Polarization Mechanism Investigation on Colossa l Permittivity of Barium Titanate Ceramics Relative permittivity and dielectric loss of SPS BT 0.95 and SPS BT 1.00 were measured as a function of temperature (25 K ~ 300 K) at different frequencies (40 Hz ~ 100 kHz), and the results are shown in Figure 7 3. Extremely high permittivity of up to 10 5 with low loss ( ~ 0.05) is observed at room temperature in both samples. Table 7 2 shows dielectric properties of SPS BT 0.95 and SPS BT 1.00 at certain frequencies (1, 10, 100 kHz). The CP of the sintered ceramics decreases in a step like shape as the temperature decreases, and the relative permittivity becomes independent of temperature below 50 K. This behavior is also found in a well known high permittivity material CaCu 3 Ti 4 O 12 (CCTO) 105 The decrease of relative permittivity is accompanied by the dielectric loss peak at the given temperature, which indicates a temperature activated dielectric relaxation has occurred: the dielectric los s peaks appear at lower temperatures as frequencies decrease (Figure 7 3).
107 A B Figure 7 3. Dielectric property of A ) SPS 0.95 and B ) SPS 1.00 as a function of temperature (20 ~ 300 K).
108 Table 7 2 Dielectric property of SPS BT 0.95 and 1.00 at roo m temperature and 1, 10, and 100 kHz. 1 kHz 10 kHz 100 kHz SPS BT 0.95 667,941 0.04 638,737 0.05 587,040 0.17 SPS BT 1.00 114,505 0.06 103,387 0.09 88,869 0.20 Generally, the Debye model can be used to describe the dielectric relaxation, and the relaxation frequency can be represented by, (7 1 ) o k B and E A represent the pre expo nential factor, the Boltzmann constant, and the activation energy for relaxation, respectively. 57 The imaginary part of the dielectric response ( ) is proportional to and the maxima of occurs when where is the dielectric relaxation time. Thus, the relaxation temperatures at different frequencies can be extracted from the maximums of and E A can be calculated from the Arrhenius plot for Eq. (7 1). As such, the maxima for SPS BT 0.95 and 1.00 are determined from Figure 7 3 and the Arrhenius plots for the both samples are shown in Figure 7 4. The change in slope of the fitt ed curves for SPS BT 0.95 clearly indicates that two different thermally activated polarization mechanisms exist (Figure 7 4 ( A )). By contrast, it is hard to distinguish a slope change for SPS BT 1.00 sample (Figure 7 4 ( B )). The activation energies of 0 .054 eV, 0.078 eV and the jump frequencies of 4.5 10 5 Hz, 1.89 10 6 Hz are obtained for SPS BT 0.95 and 1.00, respectively, in the low temperature region (30 K ~ 100 K), while 0.093 eV, 0.108 eV and 3.57 10 7 Hz, 4.2 5 10 7 Hz are calculated in
109 the high temperature region (100 K ~ 300 K) (Figure 7 4). Higher activation energy of the high temperature region is consistent with the co existence of two polarization mechanisms. A B Figure 7 4. Activat ion energy of thermally activated relaxations for A ) SPS BT 0.95 and B ) SPS BT 1.00.
110 As mentioned in Chapter 6, it has been widely investigated that reducing sintering atmosphere (vacuum) of SPS technique can cause high concentration of Ti 3+ and in the sintered ceramics. 37 Induced charged defects and electrons can be the localized charge carriers in the grain or at the grain boundary under the applied ac electric field. The origin of CP of barium titanate has been widely explained by the inte rfacial polarization effect by oxygen vacancies in the vicinity of gain boundaries and mobile electrons in the grains. 37 106 However, it is also well known that the localized charge carriers can induce hopping dipoles in the material by polaron hopping such that can affect dielectric response of the material. 107 Thus, the second polarization mechanism of the CP in barium titanate could be attributed to a polaron hopping process in the grains. If that is the case, then the large change of activation energies in SPS BT 0.95 compared to SPS BT 1.00 (Figure 7 4 ( A )) could be attributed to the higher concentration of polarons due to non stoichiometric Ba / Ti ratio of SPS BT 0.95 sample. To further investigate this hypothesis, the UDR model can be applied in order to investigate dielectric response due to localized charge carriers in the material. 108 111 can be represented as a power law of resulting from the Kramers Kronig transformation for a power law of the ac conductivity, (Eq. 3 76). 111 Therefore, relative permittivity ( ) can be written as,
111 (7 2) where and s represent the temperature dependent constants, and and f denote the permittivity of free space and experimental frequency ( ), respectively. Eq (7 2) also can be written as, (7 3) where and s is the constant value between 0 ~ 1. Thus, a straight line should appear in plot at the given temperature, and the slope of the line indicates the value of s. Figure 7 5 represent s the plot for SPS BT 0.95 at different temperatures (30 K ~ 300 K). A straight line does appear at high temperatures and low frequencies. However, the straight line is deviated from the slope as frequency increases due to relaxati on, and consequently it decreases in a step like shape and forms another straight line at the high frequency region. As temperature decreases, deviations from the slope are gradually occurred at lower frequencies as relaxation frequencies shift to lower f requencies at lower temperature (Figure 7 5). The values of s are found to be 0.98 and 0.95 when obtained from the slopes at high and low indicates polarization ch arges that are more strictly localized. 112
112 Figure 7 5. Log 10 r f) vs. log 10 f plot for the SPS (1150C) BT at different temperatures (30 K ~ 300 K). In the standard hopping polarization model, the value of s tends to increase and erature decreases since the hopping dipoles freeze at low temperatures. 112 Interestingly however, it is shown here that for the CP of BT, the value of s decreases as temperature decreases, which means charge carriers for polarization are less localized at low tem peratures when compared to high temperatures. This indicates that, while hopping polarization is becoming inactive at low temperatures due to insufficient energy to overcome energy barrier for polarization, interfacial polarization associated with mobile electrons in the grains and oxygen vacancies at the vicinity of the grain boundaries are still active as a polarization mechanism. Thus, the lower value of s may be due to the fact that the electrons for interfacial polarization are free to move
113 in the gr ains compared to hopped electrons with oxygen vacancies for hopping polarization. Figure 7 6 shows the permittivity and dielectric loss change of the SPS BT 0.95 sample as a function of frequency (40 Hz 1 MHz) at different temperatures (40 K 300 K). At high temperatures, both polarization mechanisms (interfacial and hopping polarizations) are able to contribute to the CP of BT and form the upper plateau (Figure 7 6). As frequency increases, the permittivity decreases significantly with a correlated dielectric loss peak. This is consistent with dielectric relaxation as described by Debye theory, 57 where relaxation frequencies are also shifted to lower frequencies as temperature decreases. Figure 7 6. Dielectric properties of SPS BT 0.95 sample as a function of frequency (40 Hz 1 MHz) at different temperatures (40 K 300 K).
114 To further depict this behavior, the relative permittivity and dielectric loss of SPS BT 0.95 between 40 Hz ~ 1 MHz are separately shown in Figure 7 7 fo r representative temperatures across the range investigated (i.e. 300 K 160 K, 100 K, and 30 K). It is observed that an extremely high permittivity of up to 2.4 10 5 and dielectric loss below 0.1 is achiev ed at 300 K between 40 Hz ~ 10 kHz. Here, the CP is maintained ( > 10 5 ) up to 100 kHz, however, it drops significantly as frequency increases from 100 kHz to 1 MHz, and correspondingly dielectric loss shows sharp increases to ~ 2.8 (Figure 7 7 ( A )); this physically meaningless value is rather a clear indication of the onset of conductivity at those higher frequencies. By contrast, at low temperature, no CP is observed and for example, at 30 K, the highest perm ittivity is ~1,000. Moreover, the decrease in relative permittivity as the measuring frequency increases is markedly different when low and high temperature responses are compared. At high temperature (Figure 7 7 ( A )), it is clear that the drop in permit tivity undergoes two relaxation mechanisms with different characteristic frequencies, while at low temperature, only one relaxation mechanism with very low characteristic frequency is active (Figure 7 7 ( D )). This further confirms that interfacial and pol aron hopping polarizations contribute to the CP at high temperatures (300 K), however only interfacial polarization is active at low temperatures (30 K), which is consistent with the UDR analysis.
115 A B C D Figure 7 7. Dielectric properties of SP S BT 0.95 at A) 300 K B) 160 K, C) 100 K, and D) 30 K in the frequency range of 40 Hz ~ 1 MHz. Furthermore, to gain insight into the physical characteristics of the processes driving these relaxations, it is important to recall that in the thermally acti vated polaron hopping model, the maximum of the imaginary part of the relative permittivity, is related to the number of hopping polarons by, ( 7 4 )
116 where N and represent the number of hopping polarons and the hopping dipole moment, respectively. N is exponentially dependent on the temperature, which can be written as, (7 5) where N o is the pre exponential factor and E A is the activation energy associated with d ielectric relaxation of hopping dipoles. 113 114 Thus, substituting Eq. (7 5) into Eq. (7 4) results in, (7 6) The imaginary part of the relative permittivity for SPS BT 0.95 is plotted in Figure 7 8 as a function of frequency (40 Hz ~ 1 MHz) between 300 K ~ 100 K. The obtained for each temperature was then plotted in the Arrhenius form vs. 1/T to calculate the activation energy for hopping polarization (Figure 7 9). It is clear that two linear slopes appear with a transition temperature around 180 K. Accordingly, activation energ ies of 0.03 5 eV and 0.018 eV are calculated from the fitted line for high (300 K ~ 180 K) and low temperature (160 K ~ 100 K) regions, respectively Not surprisingly increases as temperature increases, which further confirms that thermally act ivated polarization, associated with polaronic dipoles, is a contributing polarization mechanism to the CP of BT.
117 Figure 7 8. The imaginary part of the relative permittivity changes for SPS BT 0.95 as a function of frequencies (40 Hz ~ 1 MHz) at 300 K ~ 100 K. More importantly, comparison between the activation energy for the high temperature (E AH ) and low temperature (E AL ) regions can reveal the activation energy for the different polarization mechanisms (similar to the analysis of the data in Figure 7 4) as follows: (7 7) (7 8) where E AP and E AI are the activation energies for hopping polarization and interfacial polarization in the CP of BT, respectively. From Figure 7 4 ( A ) E AH and E AL were calculate d as 0.094 eV and 0.054 eV; using these values in Equation (7 7) and (7 8), E AP of 0.039 eV and E AI of 0.054 eV are obtained. These results are in good agreement with the calculated activation energy (E AP = 0.03 5 eV) for high temperature region by
118 using t he thermally activated polaronic model (Figure 7 9). All these results demonstrates that in addition to interfacial polarization mechanism, which has been widely accepted for the origin of CP in BT ceramics, a hopping polaron mechanism co exists as an add itional polarization mechanism contributing to the CP of BT ceramics. Figure 7 9. Activation energy for polaron hopping polarization at 300 K ~ 200 K. 7.4 V ariable R ange Hopping Conduction in Barium Titanate Ceramics Exhibiting Colossal Permittivity Se veral techniques have been investigated to discover novel CP materials to be utilized in microelectronics, high energy density storage applications, and high performance dielectric devices. 41 48 H owever, a strong temperature and frequency dependency is commonly observed in dielectric properties of the CP materials, which limits the use of CP materials in conventional capacitive applications. 115 It is widely accepted that in CP materials the temperature and frequency de pendency of dielectric
119 properties is the result of extrinsic effects such as interfacial polarization and hopping conduction contributing to the bulk conductivity of the material. 49 51 Thus, understand ing the fundamental mechanisms of the bulk conduction i n CP materials is essential to tailor the temperature and frequency dependent dielectric response and thus fully realize their potential Fundamental a nalys e s of this kind ha ve been attempted in the past for some CP materials such as CCTO and Ni 0.5 Zn 0.5 Fe 2 O 4 but those results cannot be generalized to BT ceramics exhibiting CP 49 51 In this chapter a detailed analysis of the electrical conducti vity of BT ceramics exhibiting CP is presented and a corresponding analytical model elucidating the conduction mechanism s is in troduced In addition, the relative contributions of the respective polarization mechanism s to the dc conductivity and the relative permittivity a re estimated. Figure 7 10 shows d ielectric properties and the real part of ac conductivity ( ) for BT sinte red by SPS technique (SPS BT) as a function of frequency (1 kHz 4 MHz) at different temperature (110 K 220 K). In Figure 7 10 ( A ), i t is observed that CP up to 10 5 is maintained across a frequency range of 1 kHz 10 0 KHz at 220 K. As frequency incre ases, a Debye type dielectric relaxation occurs, which is naturally accompanied by a dielectric loss peak at the relaxation frequency as expected ( Figure 7 10 ( B )) The relaxation frequency shifts to lower frequencies as temperature decreases. It is wort h mentioning that a relative permittivity of 650 ~ 1150 was observed at 10 MHz in the measured temperature range presented here (110 K 220 K). Also, the dielectric response at temperatures above 220 K to room temperature did not vary substantially
120 and w as almost identical to that at 220 K. By contrast, the relative permittivity was significantly decreased to below 10 4 at temperatures lower than 110 K. Towards a systematic analysis of this dielectric behavior, it is important to recall that a ccording to UDR model, the conductivity of the material can be described as equation (7 9) 116 ( 7 9 ) where and s are temperature dependent constants, f represent s the experimental frequency ( ), and is the dc conductivity. Figure 7 1 0 ( C ) shows the experimental conductivity data for SPS BT with fitted curves using equation (7 9) It can be observed that for a given temperature, t wo different types of response with characteristic frequency ranges occur. O ne range shows a strong decre ase of the conductivity at low frequencies (below ~10 4 Hz) 49 however in the high frequency range ( above ~10 4 Hz) the conductivity of SPS BT shows low frequency dependence that can be well fitted by using the UDR model ( equation (7 9) ) Moreover, it should be noted that t obtain ed from the fitted curves is in around 0.50 ~ 0.64 Typic ally, it is understood that an s free to conduct through the material 116 while an s etween 0.5 and 0.8 is observed in materials w ith more localized charge carriers affect ing not only conductivity but also the dielectric polarization of the material. 116 117 In addition, can be extracted at a given temperature from the fitted parameters using equation (7 9) which gives valuable information to understand conduction mechanisms in the materia l.
121 F igure 7 10 Dielectric properties and conductivity of SPS BT as a function of frequency (1 kHz ~ 4 MHz) at different temperatures (110 K ~ 220 K) A ) The relative
122 permittivity B ) D ielectric loss C ) C onductivity. In trying to understan d the origin of these localized charges, it is worth remembering that materials sintered by fast firing methods such as SPS and MWS, are significantly reduced during the process. 38 118 119 As such, a high concen tration of oxygen vacancies, Ti 3+ and free electrons can be induced in the sintered ceramic T hese anticipated charge carriers can then contribute to the conductivity of the material via long range hopping conduction under applied electric field. In hopping conduction, the dc conductivity of the material ( ) can be well described by the Arrhenius equation described below ( 7 10 ) where is a constant and E 1 is the activation energy for hopping conduction. 120 In th is model, hopping conduction always occurs through the neighborin g site s and hence the hopping range and activation energy are independent of temperature. This model is commonly referred to as the nearest neighbor hopping (NNH) conduction. 120 121 In the NNH model, equation (7 10) can be applied to describe the temperature dependent where the concentration of c harge carriers for hopping conduction is temperature dependent It is obvious from equation (7 10) that should have a linear relationship with 1/T for materials with a conduction mechanism follow ing the NNH model. To test this model, va lues extracted from each of the fitting curves in Figure 7 10 ( C ) were plotted as a function of inverse temperature (1/T), and are presented in Figure 7 1 1 It is clearly observed that increases as temperature increases
123 indicating that thermally activated bulk conduc tion exists in SPS BT. An a ctivation energ y of 112 5 12.9 m eV is obtained from the vs. 1/T plot. However, contrary to the NNH model assumptions, Mott et.al. have pointed out that hopping conduction in certain materials would not process through the nearest site at low temperature s According the theory proposed by Mott the hopping range may vary as temperature decreases and become larger than the distance between neighboring sites due to the lower activation energy involved at lower temperature s range hopping (VRH) conduction. that follows ( 7 11 ) ( 7 12 ) where and T o are constants, and N(E F ) are the decay length of the localized wave function and the density of localized states at the Fermi level respectively The exponent p can be 1/4, 1/3, and1/2 for different materials 121 122 According to equation (7 11) the linear relationship between and 1/T p exists for the material following the VRH conduction mechanism. I n order to equation (7 11) one should invoke the activation energy defining equation which is given b y, 123 ( 7 13 ) Most importantly, this equation also enables to investigate the activation energy of the bulk conducti vity (E A ) regardless of a particular c onduction mechanism The inset in Figure 7 1 1 depicts the activation energies calculated by using equation (7 13) as a
124 function of temperatur e It is clearly shown that in Figure 7 1 1 E A is temperature dependent with E A decreas ing as temperature decr eases. This result is in contradiction to the NNH conduction model, however consistent with the VRH theory, in which E A is supposed to be variable as a function of temperature. Furthermore, it can be sho wn by comparing equation (7 13) and equation (7 11) that 1 p should be equal to the slope of plot. A s can be seen in the inset of Figure 7 1 1 the slope of ~ 0.5 is obtained from the linear fit of the plot. Thus, it can be inferre d that the most probable is 1/2 T he temperat ure dependence of with respect to 1/T 1/2 for SPS BT is represented in Figure 7 1 1 It is observed that the linear fit of against 1/T 1/2 matches very well the experimental data over a whole range of temperature from 110 K to 220 K It is important to note that compared to the NNH model, the linear relationship of and temperature following the VRH model can more satisfactorily describe the experimental results in the whole range of temperature s, especially for the experimental data at low temperature region s All these results support that the conduction mechanism in SPS BT follows VRH model rather than NNH model with t he value of p as 1/2, which is similar to the case of well known CP materials, such as CCTO. 49
125 F igure 7 11 Temperature dependence of dc conductivity ( ) with respect to 1/T and 1/T 1/2 Black lines are the linear fits of the dat a. In the VRH model, the activation energy (W) can be written as the following equation by substituting equation (7 11) into equation (7 13) ( 7 14 ) From the fitt ed value of T o (4.48 10 4 K), W can be calculated as a function of temperature, and the result ing values are shown in Figure 7 1 2 It is observed that W varies from 95.7 meV to 135.3 meV as temperature increases from 110 K to 220 K, which is well consistent with the activation energies calculated by using the defining equation equation (7 1 3) (E A = 98.0 ~ 125.7 meV at 110 K ~ 220 K). This also apparently suggests that the conduction mechanism in SPS BT follows the VRH model. Furthermore, in the VRH model it is possible to investigate the most probable hopping range (R) at a certain temper ature by using the equation below 49
126 ( 7 15 ) for SPS BT, can be assumed as the distance between nearest Ti ions which in the case of BT is approximately 0.40 nm Using these values for T o and the density of localized states at the Fermi level can be estimated as N(E F ) 3.11 10 22 (eV 1 cm 3 ) which is three orders of magnitude higher than that of CCTO ( N(E F ) 3. 25 10 19 (eV 1 cm 3 ) ). 49 Then, the hopping range of R can be calculated for SPS BT as a function of temperature, which is also presented in Figure 7 1 2 It is found that R increases from 0.41 nm to 0.48 nm as temperature decreases from 220 K to 110 K. T h e hopping distance at 220 K for SPS BT is almost identical with the distance between nearest Ti ions and it varies approximately 0.07 nm as temperature changes from 220 K to 110 K. It should be denoted that the variation of hopping distance for SPS BT is much less than that of other CP materials such as CCTO which also follows the VRH conduction mechanism. For example, Lei Zhang et. al. have demonstrated that the hopping range of CCTO increases from 2.32 nm to 2.76 nm as temperature decreases from 180 K to 90 K. 49 The relatively high number of N(E F ) and less variation of hopping range for SPS BT, c ompared to other CP materials, might be due to the high concentration of Ti 3+ and oxygen vacanc ies existing in SPS BT since the sample was sintered under severe reduction conditions (i.e., vacuum and fast heating rate ). Moreover the activation energy for dielectric polarization (9 3 m eV) 115 of SPS BT calculat ed by using the Debye model is consistent with the activation e nergy of hopping conduction (W 9 0 ~ 1 2 0 m eV) for the temperature range of 200 K ~ 100 K. This point to the possibility that an inherent correlation between bulk conducti on and
127 polarization exists in SPS BT ceramics, which will be further demonstrated in the following discussions. F igure 7 12. Activation energies and hopping distance as a function of temperature for VRH conduction in SPS BT. 7. 5 Tailoring Contribut ions of Each Polarization Mechanism to Colossal Permittivity of Barium Titanate Ceramics Recognizing that various polarization mechanisms are contributing to the observed CP, it is important to estimate their relative contribution. The dielectric response of CP materials can be well described by the model 124 the grain and the grain boundary are supposed to be in series, and each component is composed of three parallel elements: a capacitance, a conductance, and a constant phase element (CPE). The analytical equation for this equivalent circuit can be driven as, 50
128 (7 16) where represent frequency independent part of the universal conductivity for grains (g) and grain bo undaries (gb) i and are the permittivity of the free space the imaginary unit, the ratio of thickness of grain boundary to thickness of grain and angular frequency ( ) respectively. Also, is the exponent of angular frequency for the grain s and the grain boundaries, and is equal to where is a prefactor. It is worth noting that in the model, the value of p g(gb) should be in between 0 and 1, and p g(gb) become closer a more capacitive respons e results in the p g(gb) The e xperimental data with theoretical fitted curves using equation (7 16) for the relative permittivity of SPS BT is shown in Figure 7 13 Additionally, Table 7 3 summarizes the fitting parameters used in eq uation (7 16) It is noticeable that the fitted curves are significantly in good agreement with the measured experimental data over the entire frequency range. Furthermore, is orders of magnitudes higher than which is in accordance with the IBLC mod el ( Table 7 3 ). As expected for the IBLC model the value of p g ( 0.91) is much gb ( 0.03) which demonstrates that conductive grains and capacitive grain boundaries are present in SPS BT.
1 29 F igure 7 13 Experimental data with fitting data of the real and imaginaly part of permittivity for SPS BT. T able 7 3. model. Related parameters SPS c g 0.32 c gb 1.27 10 8 p g 0.90971 p gb 0.02853 g,dc (Scm 1 ) 1.39 10 3 gb,dc (Scm 1 ) 4.54 10 10 b / t g ) 0.0100 BT, 494.03477 T he contribution of interfacial polarization to the CP can be calculated by using the IBLC model, which can be presented by, (7 1 7 ) where , t g and t b are the effective relative permittivity, relative permittivity of the material, grain size, and thickness of grain boundary, respectively. From the fitting
130 parameter in Table 7 3, r elative permittivity ( ~ 494) and t g /t b ( ~ 100) are used to calculate the effective permittivity of SPS BT. An effective permittivity of 4.94 10 4 ( equivalent to ~2 8 % of the experimentally observed CP ) is calculated for the SPS B T It is important to note that this result is well agreeable with the experimentally calculated value. Relative permittivity of barium titanate ( ~ 2,400) 125 grain size ( 5 6 6 nm from Figure 7 1 (A) ), and thickness of grain boundary ( ~ 1 nm) 126 were experimentally determined for SPS BT, and the contribution of interfacial polarization to the CP was calculated as 20.12 %. Furthermore, as describ ed in Chapter 6, the electrode effect due to Schottky junction between metal electrode and semiconductive surface of BT ceramics was estimated by subtracting permittivity measured with Al electrode from measured permittivity with Au. The calculation showe d that about 15 % of the CP of SPS BT ceramic s at low frequencies can be attributed to the extrinsic electrode effect. Moreover, the contribution of hopping polarization can also be obtained simply by subtracting the interfacial contribution from the tota l intrinsic permittivity resulting in ~ 57 % of total relative permittivity It is interesting to note that the low loss ( ~ 0.05) of SPS BT samples can be attributed to a thin re oxidation layer on the surfaces associated with the s hort annealing treatment (850C for 15 min). This is further confirmed as the measured dielectric loss of the samples is extensively increased ( > 1.00) after polishing the surfaces. The surface of the samples become insulating afte r the short annealing process while the interior are still reduced (semi conductive). This configuration is
131 comparable with a barrier layer capacitance (BLC) effect, and is able to effectively lower the dielectric loss while maintaining the high permittiv ity of sintered ceramics. Thus, it can be stated that the origin of CP in BT ceramics is due to the combination effects of BLC (thanks to an insulating surface), interfacial polarization at the interior insulating grain boundaries, and hopping polarizati ons in semiconducting grains by a large number of induced charge carriers. These different contributions to the observed CP are schematically depicted in Figure 7 1 4 F igure 7 1 4 Interfacial and hopping polarization model for the colossal permittivit y of BT ceramics. 7. 6 Conclusion BT ceramics with different Ba / Ti ratios (0.95 and 1.00) were synthesized by co precipitation method and SPS sintering technique. Microstructure, chemical stoichiometry, and phase purity analyses were followed on sintere d ceramics. Broadband dielectric spectroscopy was performed in order to reveal the polarization mechanisms in SPS BT ceramics. A ccording to dielectric data analyses by Debye
132 i n addition to interfacial polarization at insulating grain boundary, polarization due to polaron hopping was proposed as a co existing polarization mechanism in the CP of BT ceramics. A nalysis of the temperature dependen t bulk dc conductivity reveal s that the bulk conduction in fast fired barium titanate is the result of VRH model rather than NNH model and the hopping distance was calculated as a function of temperature Moreover, relative contributions of ea ch polarization to the CP of BT ceramics were calculated as ~ 57 % hopping polarization, ~2 8 % interfacial polarization, and ~15 % electrode effect respectively
133 CHAPTER 8 SUMMARY AND FU TURE WORK 8.1 Summary 8.1 .1 Defects and E lectrical P roperty R elationships in BiI 3 Purification of the BiI 3 powder to a high level (6N) resulted in the ove rall improvement of properties of the BiI 3 detectors. Ultrapure powder with an impurity level of less than 10 ppm was synthesized by the PVT technique. The resistivity of the single crystals grown from the ultra pure powder was found to be on the order of 10 10 cm which is greater than the resistivity of crystals grown from commercial powder. An alpha particle spectrum (using a 241 Am source) was recorded at room temperature, using a pro totype detector fabricated by using ultra pure PVT powder While the detect or showed polarization effects, it constitutes as proof that room temperature detection using BiI 3 is possible and the improved resolution was observed in PVT BiI 3 detector compared to commercial BiI 3 detector Electron mobility was estimated to be 433 7 9 cm 2 /Vs, hole mobility could not be estimated due to lack of signal generated by holes. In order for improving the e lectrical properties of BiI 3 detectors such as high resistivity low leakage current, and radiation response defect engineering was empl oyed to reduce detrimental effects of iodine vacancies, based on proposed defect models for BiI 3 DFT calculations and defect modeling verified that Sb dopants in BiI 3 can effectively reduce the formation and the migration of iodine vacancies by forming h ighly covalent bondings with surrounding ions or reducing jumping sites for iodine vacancies. E lectrical and radiation detecting properties in the Sb doped BiI 3 single crystal detector were significantly improved compared to undoped BiI 3 single crystal
134 de tector s Leakage current measurements and particle radiation response demonstrate that time dependent detect ion properties were also remarkably enhanced in Sb doped BiI 3 single crystal detector. In addition, the electron mobility of Sb doped BiI 3 detec tor was significantly enhanced (~70 %) compared to that of undoped BiI 3 detecto r. Therefore Sb doped BiI 3 single crystal s can now be considered realizable material s for gamma ray detector s surpassing undoped BiI 3 based on enhanced electrical properties a nd long term stability of the detector 8. 1. 2 Colossal P ermittivity in F ast f ired BaTiO 3 In the current work, BT ceramics were synthesized using conventional PS SPS and MWS sintering techniques from nanocrystalline BT powder. CP was observed in the S PS and MWS BT ceramics, sintered at 1050 C for 3 min and at 1320 C for 30 min respectively High sintering temperature and high heating rate of MWS were able to induce CP in the sintered ceramics. P ost sintering annealing treatments were performed to re oxidize SPS and MWS BT ceramics w ith the aim to improve insulating characteristic while still maintaining h igh permittivity The best dielectric properties ( > 10 4 and tan 0.04) of SPS and MWS ceramics were achieved after annealing process at temperature 850C for 15 min and 950 C ~ 1100 C for 12 ~ 24 hours respectively Efforts were made on synthesizing SPS BT ceramics with specific Ba/Ti ratios (0.95 and 1.00) Dielectric spectroscopy demonstrated that sintered ceramics have CP up to ~10 5 at room temperature and 1 kHz This CP c an be maintained up to 100 kHz at room temperature. Onset of d ielectric relaxation following Debye theor y occurred as temperature decrease d Activation energy changes for relaxation indicate that at least
135 two different relaxational polarization mechanisms may be contributing the CP in fast fired BT ceramics A ccording to DR model, i n addition t o interfacial polarization at insulating grain boundary, hopping polarization due to induced defect dipoles was proposed as a co existing polarization mechanism in the CP of BT ceramics. In addition relative contributions of ea ch polarization to the CP w ere calculated: ~ 57 % hopping polarization, ~2 8 % interfacial polarization, and ~15 % electrode effect. Furthermore, in depth analysis for revealing the conduction mechanism was conducted for the fast fired BT ceramics since defect dipoles are also attr ibuted to bulk conductivity of the material which degrades its functional properties as dielectric material The conduction mechanism in the fast fired BT ceramics followed the variable range hopping (VRH) conduction mechanism rather than the nearest neig hboring and the relative permittivity of the fast fired BT were deduced for each polarization mechanism 8.2 Future Work 8.2.1 Defect E ngineering of BiI 3 : C ontrolling A tmospheric P ressure It is well known that changing the pressure of atmosphere will affect the defect formation, such as oxidation/reduction control in the oxide ceramics. Thus, in bismuth tri iodide, iodine vacancy forma tion can be controlled by regulating iodine partial pressure through equation (8 1) (8 1)
136 T he equilibrium constant for this reaction can be represented as, (8 2) since is equal to n b y the electro neutrality equation, (8 3) B ismuth vacancy formation can be related with partial pressure of iodine by, (8 4) T he equilibrium constant for bismuth vacancy formation can be established as, (8 5) needs to be same as 3 for electroneutrallity, (8 6) T hus, the Kroger Vink diagram for idodine and bismuth vacancies with respect to iodine pa rtial pressure can be represented as Figure 8 1 Kroger Vink diagram for iodine and bismuth vacancies with respect to iodine partial pressure
137 It is clearly shown in Figure 8 1 that the concentration of iodine vacanc ies can be controlled by changing th e partial pressure of iodine vapor. The concentration of iodine vacanc ies are decreasing at high pressure region. Although the concentration of bismuth vacanc ies are increased a t the expense of a decrease in iodine vacanc ies provided that iodine vacanc i es are more mobile than bismuth vacanc ies total conductivity could be reduced in terms of reduced iodine vacancies. At the melting point of BiI 3 (408 C), the maximum vapor pressure (0.09 atm) of iodine is calculated through equation (8 7) 82 (8 7) Thus, the atmospheric pressure will be controlled between 0.09 atm to 1 atm by injecting iodine vapor into the single crystal growth chamber, and the effect of atmospheric pressure on the electrical properties o f BiI 3 single crystals will be investigated. 8.2.2 Polarization M echanism in (Nb+In) C o d oped TiO 2 C eramics E xhibiting T emperature f requency I ndependent C olossal P ermittivity Recently, Hu et. al. demonstrate d that through novel defect engineering strateg ies quasi intrinsic CP (nearly temperature frequency independent) can be achieved in (Nb + In) co doped TiO 2 due to highly localized defect clusters 127 Hopping di stance of (Nb + In) co doped TiO 2 remains constant at different temperatures due to highly localized defect dipoles, which consequently gives rise to the excellent temperature frequency stable CP. While this approach suggest ed a promising strategy for the development of new high performance CP materials precise polarization mechanisms have not been determined on that material. Thus, detailed broadband dielectric spectroscopy analyzes following this work will be performed in order to reveal
138 polarization m echanisms in (Nb+In) co doped TiO 2 ceramics and deduce CP in this material into respective intrinsic or extrinsic effects
139 LIST OF REFERENCE 1. S.R. Phadke, C.R. Bowers, E.D. Wachsman, and J.C. Nino, "Proton C onduction in A cceptor D o ped SnP 2 O 7 ," Solid State Ionics 183  26 31 (2011). 2. S. Omar, E.D. Wachsman, and J.C. Nino, "Higher C onductivity Sm 3+ and Nd 3+ C o doped C eria based E lectrolyte M aterials," Solid State Ionics 179 [33 34] 1951 1951 (2008). 3. S. Omar, E.D. Wachsman, an d J.C. Nino, "Higher I onic C onductive C eria based E lectrolytes for S olid O xide F uel C ells," Appl Phys Lett 91  (2007). 4. P. Caroff, K.A. Dick, J. Johansson, M.E. Messing, K. Deppert, and L. Samuelson, "Controlled P olytypic and T win plane S uperlatt ices in III V N anowires," Nat Nanotechnol 4  50 55 (2009). 5. A.A. Bonapasta and P. Giannozzi, "Defect E ngineering in III V T ernary A lloys: Effects of S train and L ocal C harge on the F ormation of N ative D eep D efects," Nucl Instrum Meth B 186 229 2 33 (2002). 6. A.A. Bonapasta and P. Giannozzi, "Defect E ngineering in III V T ernary A lloys: E ffects of S train and L ocal C harge on the F ormation of S ubstitutional and I nterstitial N ative D efects," Physica B 308 846 849 (2001). 7. P.N.K. Deenapanray, A. Mar tin, and C. Jagadish, "Defect E ngineering in A nnealed n type GaAs E pilayers U sing SiO 2 /Si 3 N 4 S tacking L ayers," Appl Phys Lett 79  2561 2563 (2001). 8. S. Ostapenko, "Defect E ngineering in P olycrystalline S ilicon U sing U ltrasound," Solid State Pheno m 67 8 485 495 (1999). 9. R.M. Cohen, G. Li, C. Jagadish, P.T. Burke, and M. Gal, "Native D efect E ngineering of I nterdiffusion U sing T hermally G rown O xides of GaAs," Appl Phys Lett 73  803 805 (1998). 10. P. Kightley, R.I. Taylor, A.J. Moseley, P. D. Augustus, A.C. Marshall, and R.J.M. Griffiths, "Dislocation Engineering in Advanced I II V Device Structures," Inst Phys Conf Ser  187 192 (1989). 11. J. Lagowski and H.C. Gatos, "Defect Engineering in Ga A s for Device Processing," J Electroche m Soc 134  C576 C576 (1987). 12. V.Q. Nguyen, J.K. Kang, H.S. Han, H.Y. Lee, S.J. Jeong, C.W. Ahn, I.W. Kim, and J.S. Lee, "Bi based L ead free C eramic M ultilayer A ctuators U sing AgPd (Na 0.51 K 0.47 Li 0.02 )(Nb 0.8 Ta 0.2 )O 3 C omposite I nner E lectrodes," Sen sor Actuat a Phys 200 107 113 (2013).
140 13. J.G. Chen, G.X. Liu, X.T. Li, Z.J. Chen, and S.X. Dong, "High Temperature Actuation Performance of BiScO3 PbTiO3 Ceramics and Their Multilayer Configuration," I EEE T Ultrason Ferr 60  446 450 (2013). 14. J. Wang, G. Elghoul, and S. Peters, "Lead Zirconium Titanate Alternatives for Nanoactuators," I EEE T Ultrason Ferr 60  256 260 (2013). 15. C.Y. Wu, X.P. Li, W.S. Li, B. Li, Y.Q. Wang, Y.T. Wang, M.Q. Xu, and L.D. Xing, "Fe 2 O 3 N anorods/ C arbon N anofib ers C omposite: Preparation and P erformance A s A node of H igh R ate L ithium I on B attery," J Power Sources 251 85 91 (2014). 16. Q. Wang, J. Zhao, W.F. Shan, X.B. Xia, L.L. Xing, and X.Y. Xue, "CuO N anorods/ G raphene N anocomposites for H igh performance L ithi um ion B attery A nodes," J Alloy Compd 590 424 427 (2014). 17. B.H. Qu, L.L. Hu, Q.H. Li, Y.G. Wang, L.B. Chen, and T.H. Wang, "High Performance Lithium Ion Battery Anode by Direct Growth of Hierarchical ZnCo 2 O 4 Nanostructures on Current Collectors," Ac s Appl Mater Inter 6  731 736 (2014). 18. V.S. Puli, D.K. Pradhan, B.C. Riggs, D.B. Chrisey, and R.S. Katiyar, "Investigations on S tructure, F erroelectric, P iezoelectric and E nergy S torage P roperties of B arium C alcium T itanate (BCT) C eramics," J Al loy Compd 584 369 373 (2014). 19. L.F. Zhu, B.P. Zhang, W.G. Yang, N. Ma, X.K. Zhao, and L. Zhao, "High P iezoelectric P roperties of (Ba,Ca)TiO 3 0.04LiF C eramics S intered at a L ow T emperature," J Electroceram 30 [1 2] 24 29 (2013). 20. M. Kuhn, S.R. B ishop, G. Ciampi, H.L. Tuller, W. Higgins, and K.S. Shah, "Ionic C onduction in TlBr Impact of Te A cceptor D opant E xsolution," Solid State Ionics 241 30 35 (2013). 21. H.L. Tuller, S.R. Bishop, D. Chen, Y. Kuru, J.J. Kim, and T.S. Stefanik, "Praseodymium D oped C eria: Model M ixed I onic E lectronic C onductor with C oupled E lectrical, O ptical, M echanical and C hemical P roperties," Solid State Ionics 225 194 197 (2012). 22. S.R. Bishop, G. Ciampi, C.D. Lee, M. Kuhn, H.L. Tuller, W. Higgins, and K.S. Shah, "Ioni c Conductivity of TlBr 1 x I x (x=0, 0.2, 1): Candidate Gamma Ray Detector," I EEE T Nucl Sci 59  2424 2427 (2012). 23. S.R. Bishop, T.S. Stefanik, and H.L. Tuller, "Defects and T ransport in Pr x Ce 1 x O 2 delta : Composition T rends," J Mater Res 27  2009 2016 (2012). 24. S.R. Bishop, H.L. Tuller, G. Ciampi, W. Higgins, J. Engel, A. Churilov, and K.S. Shah, "The D efect and T ransport P roperties of A cceptor D oped TlBr: R ole of
141 D opant E xsolution and A ssociation," Phys Chem Chem Phys 14  10160 10 167 (2012). 25. H.L. Tuller and S.R. Bishop, "Point Defects in Oxides: Tailoring Materials Through Defect Engineering," Annu Rev Mater Res 41 369 398 (2011). 26. S.R. Bishop, T.S. Stefanik, and H.L. Tuller, "Electrical C onductivity and D efect E quilibr ia of Pr 0.1 Ce 0.9 O 2 delta ," Phys Chem Chem Phys 13  10165 10173 (2011). 27. H.L. Tuller and S.R. Bishop, "Tailoring Material Properties through Defect Engineering," Chem Lett 39  1226 1231 (2010). 28. A.T. Lintereur, W. Qiu, J.C. Nino, and J Baciak, "Characterization of B ismuth T ri iodide S ingle C rystals for W ide B and gap S emiconductor R adiation D etectors," Nucl Instrum Meth A 652  166 169 (2011). 29. M. Matsumoto, K. Hitomi, T. Shoji, and Y. Hiratate, "Bismuth T ri iodide C rystal for N uclear R adiation D etectors," I EEE T Nucl Sci 49  2517 2520 (2002). 30. D. Nason and L. Keller, "The Growth and Crystallography of Bismuth Tri Iodide Crystals Grown by Vapor Transport," J Cryst Growth 156  221 226 (1995). 31. F. Ayres, L.V.C. Assali, W.V.M. Machado, and J.F. Justo, "Role of I ntrinsic D efects in the E lectronic and O ptical P roperties of A lpha HgI 2 ," Appl Phys Lett 88  (2006). 32. F. Ayres, W.V.M. Machado, J.F. Justo, and L.V.C. Assali, "Defects in M ercuric I odide: an APW I nvestigation," Physica B 340 918 922 (2003). 33. K. Biswas and M.H. Du, "First P rinciples S tudy of N ative D efects in InI," J Appl Phys 109  (2011). 34. T.A. Kuku, "Ionic Conductivity of Sni 2 ," Solid State Ionics 20  217 222 (1986). 35. Schoon ma.J, "Role of Anion Vacancy Migration in Lead Halide Photolysis," Solid State Commun 13  673 676 (1973). 36. T.B. Adams, D.C. Sinclair, and A.R. West, "Giant B arrier L ayer C apacitance E ffects in CaC u 3 Ti 4 O 12 C eramics," Adv Mater 14  1321 + (200 2). 37. S. Guillemet Fritsch, Z. Valdez Nava, C. Tenailleau, T. Lebey, B. Durand, and J.Y. Chane Ching, "Colossal P ermittivity in U ltrafine G rain S ize BaTiO 3 x and Ba 0.95 La 0.05 TiO 3 x M aterials," Adv Mater 20  551 + (2008). 38. Z. Valdez Nava, S. Guil lemet Fritsch, C. Tenailleau, T. Lebey, B. Durand, and J.Y. Chane Ching, "Colossal D ielectric P ermittivity of BaTiO 3 based
142 N anocrystalline C eramics S intered by S park P lasma S intering," J Electroceram 22 [1 3] 238 244 (2009). 39. M. Li, A. Feteira, and D .C. Sinclair, "Origin of the H igh P ermittivity in (La 0.4 Ba 0.4 Ca 0.2 )(Mn 0.4 Ti 0.6 )O 3 C eramics," J Appl Phys 98  (2005). 40. D.C. Sinclair, T.B. Adams, F.D. Morrison, and A.R. West, "CaCu 3 Ti 4 O 12 : One step I nternal B arrier L ayer C apacitor," Appl Phys Lett 80  2153 2155 (2002). 41. H.F. Hess, K. Deconde, T.F. Rosenbaum, and G.A. Thomas, "Giant Dielectric Constants at the Approach to the Insulator Metal Transition," Phys Rev B 25  5578 5580 (1982). 42. C.C. Homes, T. Vogt, S.M. Shapiro, S. Wa kimoto, and A.P. Ramirez, "Optical R esponse of H igh dielectric constant P erovskite related O xide," Science 293  673 676 (2001). 43. N. Ikeda, H. Ohsumi, K. Ohwada, K. Ishii, T. Inami, K. Kakurai, Y. Murakami, K. Yoshii, S. Mori, Y. Horibe, and H. Ki to, "Ferroelectricity from I ron VA lence O rdering in the C harge frustrated S ystem LuFe 2 O 4 ," Nature 436  1136 1138 (2005). 44. B.G. Kim, S.M. Cho, T.Y. Kim, and H.M. Jang, "Giant D ielectric P ermittivity O bserved in Pb based P erovskite F erroelectrics," Phys Rev Lett 86  3404 3406 (2001). 45. S. Krohns, P. Lunkenheimer, C. Kant, A.V. Pronin, H.B. Brom, A.A. Nugroho, M. Diantoro, and A. Loidl, "Colossal D ielectric C onstant up to G igahertz at R oom T emperature," Appl Phys Lett 94  (2009). 46 M. Shu Nan and M. Maki, "Dielectric Anisotropy in the Charge Density Wave State of K 0.3 MoO 3 ," J Phys Soc Jpn 80  (2011). 47. C.C. Wang and L.W. Zhang, "Surface layer E ffect in CaCu 3 Ti 4 O 12 ," Appl Phys Lett 88  (2006). 48. J.B. Wu, C.W. Nan, Y.H. Lin, and Y. Deng, "Giant D ielectric P ermittivity O bserved in Li and Ti doped NiO," Phys Rev Lett 89  (2002). 49. L. Zhang and Z.J. Tang, "Polaron R elaxation and V ariable range hopping C onductivity in the G iant dielectric constant M aterial CaC u 3 Ti 4 O 12 ," Phys Rev B 70  (2004). 50. H. Zheng, L. Li, Z.J. Xu, W.J. Weng, G.R. Han, N. Ma, and P.Y. Du, "Ferroelectric/ferromagnetic C eramic C omposite and I ts H ybrid P ermittivity S temming from H opping C harge and C onductivity I nhomogeneity," J Appl Phys 113  (2013).
143 51. H. Zheng, W.J. Weng, G.R. Han, and P.Y. Du, "Colossal Permittivity and Variable Range Hopping Conduction of Polarons in Ni 0.5 Zn 0.5 Fe 2 O 4 Ceramic," J Phys Chem C 117  12966 12972 (2013). 52. W. Barsoum, Fundamentals of Ce ramics, Taylor & Francis, 2002. 53. R.J.D. Tilley, Defects in Solids, Wiley, 2008. 54. F.A. Kroger and H.J. Vink, "Relations B etween the Concentrations of Imperfections in Solids," J Phys Chem Solids 5  208 223 (1958). 55. W.C. Chueh and S.M. Haile, "Electrochemical S tudies of C apacitance in C erium O xide T hin F ilms and I ts R elationship to A nionic and E lectronic D efect D ensities," Phys Chem Chem Phys 11  8144 8148 (2009). 56. W. Lai and S.M. Haile, "Impedance S pectroscopy as a T ool for C hemic al and E lectrochemical A nalysis of M ixed C onductors: A C ase S tudy of C eria," J Am Ceram Soc 88  2979 2997 (2005). 57. E. Barsoukov and J.R. Macdonald, Impedance Spectroscopy: Theory, Experiment, and Applications, Wiley, 2005. 58. T.S. Stefanik and H.L. Tuller, "Nonstoichiometry and D efect C hemistry in P raseodymium cerium O xide," J Electroceram 13 [1 3] 799 803 (2004). 59. R.A. De Souza and M. Martin, "Probing Diffusion Kinetics with Secondary Ion Mass Spectrometry," Mrs Bull 34  907 914 ( 2009). 60. M. Yashima and T. Kamioka, "Neutron D iffraction S tudy of the P erovskite type L anthanum C obaltite La 0.6 Sr 0.4 Co 0.8 Fe 0.2 O 3 delta at 1260 C and 394 C ," Solid State Ionics 178 [39 40] 1939 1943 (2008). 61. Y. Kuru, S.R. Bishop, J.J. Kim, B. Yildiz, and H.L. Tuller, "Chemomechanical P roperties and M icrostructural S tability of N anocrystalline Pr doped C eria: An in S itu X ray D iffraction I nvestigation," Solid State Ionics 193  1 4 (2011). 62. E.D. Wachsman, N. Jiang, C.W. Frank, D.M. Mason, and D.A. Stevenson, "Spectroscopic Investigation of Oxygen Vacancies in Solid Oxide Electrolytes," Appl Phys a Mater 50  545 549 (1990). 63. S.R. Rotman, H.L. Tuller, and C. Warde, "Defect Property Correlations in Garnet Crystals .6. The Electrical Conducti vity, Defect Structure, and Optical Properties of Luminescent Calcium and Cerium Doped Yttrium Aluminum Garnet," J Appl Phys 71  1209 1214 (1992). 64. K.C. Kao, Dielectric Phenomena in Solids, Elsevier Science, 2004.
144 65. A.J. Moulson, J. Binner, R. Morrell, and I.o.C.B.S. Section, Electroceramics, Institute of Ceramics, 1988. 66. H.C. Hsueh, R.K. Chen, H. Vass, S.J. Clark, G.J. Ackland, W.C.K. Poon, and J. Crain, "Compression M echanisms in Q uasimolecular XI 3 (X = As, Sb, Bi) S olids," Phys Rev B 58  14812 14822 (1998). 67. J. Trotter and T. Zobel, "Crystal Structure of Sb I 3 and Bi I 3 ," Z Kristallogr Krist 123  67 & (1966). 68. V.B. Timofeev and Vashchen.Vi, "Optical Properties of Bii 3 Single Crystals near Long Wavelength Edge of Fundamenta l Absorption," Opt Spectrosc Ussr 24  396 & (1968). 69. Y. Kaifu and T. Komatsu, "Optical Properties of Bismuth Tri Iodide Single Crystals .2. Intrinsic Absorption Edge," J Phys Soc Jpn 40  1377 1382 (1976). 70. T. Komatsu and Y. Kaifu, "Opti cal Properties of Bismuth Tri Iodide Single Crystals .1. Interband Transitions," J Phys Soc Jpn 40  1062 1068 (1976). 71. T. Komatsu, Y. Kaifu, S. Takeyama, and N. Miura, "Effects of High Magnetic Fields on Cationic Exciton Lines in Bii 3 ," Phys Re v Lett 58  2259 2262 (1987). 72. G.H. Kwei, A.C. Lawson, S.J.L. Billinge, and S.W. Cheong, "Structures of the Ferroelectric Phases of Barium Titanate," J Phys Chem Us 97  2368 2377 (1993). 73. W. Qiu, G.J. Dudder, X.Y. Zhao, S.S. Perry, and J.C. Nino, "Interfacial Reactivity of Au, Pd, and Pt on BiI 3 (001): Implications for Electrode Selection," Acs Appl Mater Inter 3  1910 1917 (2011). 74. P.E. Blochl, "Projector Augmented Wave Method," Phys Rev B 50  17953 17979 (1994). 75. G. Kresse and D. Joubert, "From U ltrasoft P seudopotentials to the P rojector A ugmented wave method," Phys Rev B 59  1758 1775 (1999). 76. W. Kohn and L.J. Sham, "Self Consistent Equations Including Exchange and Correlation Effects," Phys Rev 140 [4A] 1133 & (1965). 77. G. Kresse and J. Furthmuller, "Efficient I terative S chemes for ab I nitio T otal energy C alculations U sing a P lane wave B asis S et," Phys Rev B 54  11169 11186 (1996). 78. H.J. Monkhorst and J.D. Pack, "Special Points for Brillouin Zone Integrations," Phys Rev B 13  5188 5192 (1976).
145 79. S.R. Bishop, W. Higgins, G. Ciampi, A. Churilov, K.S. Shah, and L. Tuller, "The Defect and Transport Properties of Donor Doped Single Crystal TlBr," J Electrochem Soc 158  J47 J51 (2011 ). 80. F.V. Motsnyi, S.S. Ishchenko, S.M. Okulov, and D.I. Bletskan, "Nature of Infrared Photoluminescence and of Doping of Layer Bismuth Iodide Single Crystals," Sov Phys Semicond+ 11  617 620 (1977). 81. J.F. Condeles, R.C.Z. Lofrano, J.M. Rosolen and M. Mulato, "Stoichiometry, S urface and S tructural C haracterization of L ead I odide T hin F ilms," Braz J Phys 36 [2A] 320 323 (2006). 82. D. Cubicciotti and F.J. Keneshea, "The Vapor Pressures of Bi I 3 over Liquid Bi Bi I 3 Solutions," J Phys Chem Us 63  295 297 (1959). 83. Z.H. Li, W.T. Li, J. Liu, B.J. Zhao, S.F. Zhu, H.P. Xu, and H. Yuan, "Improved Technique for Hg I 2 Purification," J Cryst Growth 156 [1 2] 86 90 (1995). 84. D.M. Bercha, I.I. Nebola, S.I. Radautsan, N.N. Syrbu, V.K. Kiosev, and V.E. Tezlevan, "Photoelectric Properties of Ni Cd P 2 Schottky Barriers," Sov Phys Semicond+ 8  1343 1347 (1975). 85. R.D. Shannon, "Revised Effective Ionic Radii and Systematic Studies of Interatomic Distances in Halides and Chalcogenides," Act a Crystallogr A 32 [Sep1] 751 767 (1976). 86. N.J. Podraza, W. Qiu, B.B. Hinojosa, H.X. Xu, M.A. Motyka, S.R. Phillpot, J.E. Baciak, S. Trolier McKinstry, and J.C. Nino, "Band G ap and S tructure of S ingle C rystal BiI3: Resolving D iscrepancies in L iterature," J Appl Phys 114  (2013). 87. D.J. Singh, "Structure and O ptical P roperties of H igh L ight O utput H alide S cintillators," Phys Rev B 82  (2010). 88. H. Yorikawa and S. Muramatsu, "Theoretical S tudy of C rystal and E lectronic S tructure s of BiI 3 ," J Phys Condens Mat 20  (2008). 89. J.E. Baciak and Z. He, "Long term S tability of 1 cm thick P ixelated HgI 2 G amma ray S pectrometers O perating at R oom T emperature," I EEE T Nucl Sci 51  1886 1894 (2004). 90. Y. Eisen and A. Shor, CdTe and CdZnTe M aterials for R oom temperature X ray and G amma R ay D etectors," J Cryst Growth 184 1302 1312 (1998). 91. A. Owens and A. Peacock, "Compound S emiconductor R adiation D etectors," Nucl Instrum Meth A 531 [1 2] 18 37 (2004).
146 92. J.E. Baci ak and Z. He, "Spectroscopy on T hick HgI 2 D etectors: A C omparison B etween P lanar and P ixelated E lectrodes," I EEE T Nucl Sci 50  1220 1224 (2003). 93. Z. He, G.F. Knoll, and D.K. Wehe, "Direct M easurement of P roduct of the E lectron M obility and M ean F ree D rift T ime of CdZnTe S emiconductors U sing P osition S ensitive S ingle P olarity C harge S ensing D etectors," J Appl Phys 84  5566 5569 (1998). 94. J.G.P. Binner and I.A.H. Al Dawery, "Bulk YBCO H igh Tc S uperconductors with U niform and F ull O xygen C ontent via M icrowave P rocessing," Supercond Sci Tech 11  449 457 (1998). 95. A. Goldstein and M. Kravchik, "Sintering of PZT P owders in MW F urnace at 2.45 GHz," J Eur Ceram Soc 19 [6 7] 989 992 (1999). 96. Y.J. Wu, S.H. Su, J.P. Cheng, and X.M. Chen, "Spark Plasma Sintering of Barium Zirconate Titanate/Carbon Nanotube Composites with Colossal Dielectric Constant and Low Dielectric Loss," J Am Ceram Soc 94  663 665 (2011). 97. J.D. Katz, "Microwave Sintering of Ceramics," Annu Rev Mater Sci 22 153 170 (1992). 98. H. Abrams, "Grain S ize M easurement by the I ntercept M ethod," Metallography 4  59 78 (1971). 99. A. von Hippel, "Ferroelectricity, Domain Structure, and Phase Transitions of Barium Titanate," Reviews of Modern Physics 22  221 237 (1950). 100. N.H. Chan and D.M. Smyth, "Defect Chemistry of Donor Doped Ba T i O 3 ," J Am Ceram Soc 67  285 288 (1984). 101. K. Kinoshita and A. Yamaji, "Grain Size Effects on Dielectric Properties in Barium Titanate Ceramics," J Appl Phy s 47  371 373 (1976). 102. W. Heywang, "Semiconducting Barium Titanate," J Mater Sci 6  1214 & (1971). 103. D.Y. Lu, M. Toda, and M. Sugano, "High permittivity D ouble R are earth doped B arium T itanate C eramics with D iffuse P hase T ransition," J Am Ceram Soc 89  3112 3123 (2006). 104. S. Lee, C.A. Randall, and Z.K. Liu, "Comprehensive L inkage of D efect and P hase E quilibria through F erroelectric T ransition Behavior in BaTiO 3 based D ielectrics: Part 2. Defect M odeling U nder L ow O xygen P artia l P ressure C onditions," J Am Ceram Soc 91  1753 1761 (2008).
147 105. M.A. Subramanian, D. Li, N. Duan, B.A. Reisner, and A.W. Sleight, "High D ielectric C onstant in ACu 3 Ti 4 O 12 and ACu 3 Ti 3 FeO 12 phases," J Solid State Chem 151  323 325 (2000). 106. M. Maglione, "Dielectric R elaxation of S pace C harges and P olarons in F erroelectric P erovskites," Ferroelectrics 254 [1 4] 151 158 (2001). 107. Y.H. Li, L. Fang, L.J. Liu, Y.M. Huang, and C.Z. Hu, "Giant D ielectric R esponse and C harge C ompensation of Li and Co doped NiO C eramics," Mater Sci Eng B Adv 177  673 677 (2012). 108. A. Isnin and A.K. Jonscher, "Dielectric R esponse of S ome F erroelectrics at "low" F requencies I," Ferroelectrics 210 [1 4] 47 65 (1998). 109. A.K. Jonscher, "The Universal Di electric Response and Its Physical Significance," I EEE T Electr Insul 27  407 423 (1992). 110. K.L. Ngai, A.K. Jonscher, and C.T. White, "Origin of the Universal Dielectric Response in Condensed Matter," Nature 277  185 189 (1979). 111. A.K. Jonscher, "Universal Dielectric Response," Nature 267  673 679 (1977). 112. A.S. Nowick and B.S. Lim, "Electrical R elaxations: Simple V ersus C omplex I onic S ystems," Phys Rev B 63  (2001). 113. L.L. Hench and J.K. West, Principles of E lectroni c C eramics, Wiley, 1990. 114. S. Komine and E. Iguchi, "Relation of C onduction C hanges and M agnetic P roperties in Nd 1 x Pb x MnO 3 (0.10 <= x <= 0.25)," J Phys Condens Mat 16  1061 1073 (2004). 115. H. Han, C. Voisin, S. Guillemet Fritsch, P. Dufour, C. Tenailleau, C. Turner, and J.C. Nino, "Origin of C olossal P ermittivity in BaTiO 3 via B roadband D ielectric S pectroscopy," J Appl Phys 113  (2013). 116. A.K. Jonscher, "Dielectric R elaxation in S olids," J Phys D Appl Phys 32  R57 R70 (1999). 117. S.R. Elliott, "A C Conduction in Amorphous Chalcogenide and Pnictide Semiconductors," Adv Phys 36  135 218 (1987). 118. D. Ghosh, H. Han, J.C. Nino, G. Subhash, and J.L. Jones, "Synthesis of BaTiO 3 20wt%CoFe 2 O 4 Nanocomposites via Spark Plasma Si ntering," J Am Ceram Soc 95  2504 2509 (2012).
148 119. H. Han, D. Ghosh, J.L. Jones, and J.C. Nino, "Colossal Permittivity in Microwave Sintered Barium Titanate and Effect of Annealing on Dielectric Properties," J Am Ceram Soc 96  485 490 (201 3). 120. G.H. Jonker, "Analysis of the Semiconducting Properties of Cobalt Ferrite," J Phys Chem Solids 9  165 175 (1959). 121. N.F. Mott and E.A. Davis, Electronic Processes in Non Crystalline Materials, OUP Oxford, 2012. 122. A.L. Efros and B.I. S hklovskii, "Coulomb Gap and Low Temperature Conductivity of Disordered Systems," J Phys C Solid State 8  L49 L51 (1975). 123. S.S.N. Bharadwaja, C. Venkatasubramanian, N. Fieldhouse, S. Ashok, M.W. Horn, and T.N. Jackson, "Low T emperature C harge C arr ier H opping T ransport M echanism in V anadium O xide T hin F ilms G rown U sing P ulsed dc S puttering," Appl Phys Lett 94  (2009). 124. C.G. Koops, "On the Dispersion of Resistivity and Dielectric Constant of Some Semiconductors at Audiofrequencies," Phys Rev 83  121 124 (1951). 125. G. Arlt, D. Hennings, and G. Dewith, "Dielectric Properties of Fine Grained Barium Titanate Ceramics," J Appl Phys 58  1619 1625 (1985). 126. Z. Valdez Nava, C. Tenailleau, S. Guillemet Fritsch, N. El Horr, T. Lebe y, P. Dufour, B. Durand, and J.Y. Chane Ching, "Structural C haracterization of D ense R educed BaTiO 3 and Ba 0.95 La 0.05 TiO 3 N anoceramics S howing C olossal D ielectric V alues," J Phys Chem Solids 72  17 23 (2011). 127. W.B. Hu, Y. Liu, R.L. Withers, T.J. Frankcombe, L. Noren, A. Snashall, M. Kitchin, P. Smith, B. Gong, H. Chen, J. Schiemer, F. Brink, and J. Wong Leung, "Electron pinned D efect dipoles for H igh performance C olossal P ermittivity M aterials," Nat Mater 12  821 826 (2013).
149 BIOGRAPH ICAL SKETCH HyukSu Han was born in 1982 in Cheongju, Republic of Korea. He obtained his B.S. degree in materials science and engineering at Hanyang University, Seoul, gro up at Tokyo Institute of Technology as an exchange student. In 2010, he began his