<%BANNER%>

Growth and Characterization of Bismuth Tri-Iodide Single Crystals by Modified Vertical Bridgman Method

Permanent Link: http://ufdc.ufl.edu/UFE0042085/00001

Material Information

Title: Growth and Characterization of Bismuth Tri-Iodide Single Crystals by Modified Vertical Bridgman Method
Physical Description: 1 online resource (160 p.)
Language: english
Creator: Qiu, Wei
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: bismuth, crystal, modified, single
Materials Science and Engineering -- Dissertations, Academic -- UF
Genre: Materials Science and Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Bismuth tri-iodide (BiI3) is a wide band gap semiconductor material with potential for room temperature gamma-ray detection applications. The inability to produce high quality, pure, and large single crystals is one of the major challenges for this material preventing its use in gamma-ray detection. This work presents the growth and characterization of BiI3 single crystals by a modified vertical Bridgman (MVB) method. The growth parameters to produce BiI3 single crystals were explored by adjusting growth rate and temperature gradient at the solid-liquid interface. Single crystals of BiI3 have been successfully grown in Pyrex glass ampoule at different growth conditions. Through etch pit density (EPD) and X-ray rocking curve measurement, the crystal with the best quality was determined to be obtained at 0.5 mm/h growth rate and 10oC/cm temperature gradient. Single crystal (001) oriented slabs up to 18 x 13 x 5 mm3 (the largest BiI3 single crystal ever reported) were obtained under this condition. Impurity characterization on the starting BiI3 powder and the grown crystals indicated a relatively high total impurity concentration up to 1017 cm-3, and the crystal growth using Pyrex glass does not introduce additional impurities to the crystal. The energy band gap is an important parameter for materials used as a room temperature gamma-ray detector, as it determines the energy to produce an electron-hole pair which is the source of the signal for the detector. Remarkably, very different band gap characteristics and values of BiI3 have been reported in literature. This study characterized the band gap of BiI3 through Ultraviolet-visible spectroscopy (UV-Vis), which yielded an indirect band gap of 1.68 plus or minus 0.09 eV at room temperature. Impurity and temperature effects, as commonly observed in many semiconductors, have been investigated as the major extrinsic factors that influence the band gap value of BiI3. Three different metals, Au, Pd, and Pt were sputtered on the surface of BiI3 single crystal for fabrication of the electrode. The metal-BiI3 interface was characterized by X-ray photoelectron spectroscopy (XPS). Chemical reactions were observed for the surface sputtered with Pd and Pt electrode, while no reaction was observed on the surface with Au electrode. It was proposed that Au should be used for the contact material for BiI3. Current-voltage characterization indicated that BiI3 crystal with Au electrode exhibited the highest resistivity with values in the order of magnitude of 109 ? cm, which is sufficiently large for the application of a room temperature gamma-ray detector. Based on the work presented in this dissertation, a series of avenues for future work are presented to be able to take this material into actual devices.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Wei Qiu.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Nino, Juan C.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0042085:00001

Permanent Link: http://ufdc.ufl.edu/UFE0042085/00001

Material Information

Title: Growth and Characterization of Bismuth Tri-Iodide Single Crystals by Modified Vertical Bridgman Method
Physical Description: 1 online resource (160 p.)
Language: english
Creator: Qiu, Wei
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: bismuth, crystal, modified, single
Materials Science and Engineering -- Dissertations, Academic -- UF
Genre: Materials Science and Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Bismuth tri-iodide (BiI3) is a wide band gap semiconductor material with potential for room temperature gamma-ray detection applications. The inability to produce high quality, pure, and large single crystals is one of the major challenges for this material preventing its use in gamma-ray detection. This work presents the growth and characterization of BiI3 single crystals by a modified vertical Bridgman (MVB) method. The growth parameters to produce BiI3 single crystals were explored by adjusting growth rate and temperature gradient at the solid-liquid interface. Single crystals of BiI3 have been successfully grown in Pyrex glass ampoule at different growth conditions. Through etch pit density (EPD) and X-ray rocking curve measurement, the crystal with the best quality was determined to be obtained at 0.5 mm/h growth rate and 10oC/cm temperature gradient. Single crystal (001) oriented slabs up to 18 x 13 x 5 mm3 (the largest BiI3 single crystal ever reported) were obtained under this condition. Impurity characterization on the starting BiI3 powder and the grown crystals indicated a relatively high total impurity concentration up to 1017 cm-3, and the crystal growth using Pyrex glass does not introduce additional impurities to the crystal. The energy band gap is an important parameter for materials used as a room temperature gamma-ray detector, as it determines the energy to produce an electron-hole pair which is the source of the signal for the detector. Remarkably, very different band gap characteristics and values of BiI3 have been reported in literature. This study characterized the band gap of BiI3 through Ultraviolet-visible spectroscopy (UV-Vis), which yielded an indirect band gap of 1.68 plus or minus 0.09 eV at room temperature. Impurity and temperature effects, as commonly observed in many semiconductors, have been investigated as the major extrinsic factors that influence the band gap value of BiI3. Three different metals, Au, Pd, and Pt were sputtered on the surface of BiI3 single crystal for fabrication of the electrode. The metal-BiI3 interface was characterized by X-ray photoelectron spectroscopy (XPS). Chemical reactions were observed for the surface sputtered with Pd and Pt electrode, while no reaction was observed on the surface with Au electrode. It was proposed that Au should be used for the contact material for BiI3. Current-voltage characterization indicated that BiI3 crystal with Au electrode exhibited the highest resistivity with values in the order of magnitude of 109 ? cm, which is sufficiently large for the application of a room temperature gamma-ray detector. Based on the work presented in this dissertation, a series of avenues for future work are presented to be able to take this material into actual devices.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Wei Qiu.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Nino, Juan C.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0042085:00001


This item has the following downloads:


Full Text

PAGE 1

1 GROWTH AND CHARACTERIZATION OF BISMUTH TRI -IODIDE SINGLE CRYSTALS BY MODIFIED VERTICAL BRIDGMAN METHOD By WEI QIU A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 20 10

PAGE 2

2 20 10 Wei Qiu

PAGE 3

3 With my deepest love to my entire family

PAGE 4

4 ACKNOWLEDGMENTS At the very first I want to express my deepest gratitude to my advisor, Dr. Juan C. N ino, for his support and guidance. His knowledge, care, optimism, diligence and aspirations inspired me to work hard and expand my potential. I am also extremely grateful to my other committee members (Dr. Gerald Bourne, Dr. Jacob L. Jones, Dr. Scott S. Perry, and Dr James E. Baciak ) for their time and guidance. I want to acknowledge Dr. Valentin Craciun for his help and guidance on the X -ray rocking curve characterizations I want to thank Dr. Jiangeng Xue and Dr. Franky So for letting me use their vac uum pumps. I want to give my special thanks to Mr. Michael Braddock in department of mechanical and aerospace engineering at UF, Mr. Joe Caruso in department of chemistry at UF, and Mr. Rudy Strohschein from Analytical Research Systems, Inc. in Gainesvill e for their teaching and help on the glass work. I also want to acknowledge the people who were actively involved and helped me through the entire project Dr. Ying Zheng, Dr. Haixuan Xu, Dr. Xueying Zhao, and Gregory Dudder I thank the group members: Sh obit Omar Peng Xu, Samantha Yates, Laurel Wucherer, Lu Cai, Satyajit Phadke, Donald Moore, Adam Wilk Alex Arias, and etc., for providing me excellent research environment and being helpful. Last but not least, I am deeply indebted to my parents and my wi fe for their loving encouragement. Their loves have encouraged me to go through the hard time. I cannot imagine myself having any of my progress without my familys strong support behind me.

PAGE 5

5 TABLE OF CONTENTS ACKN OWLEDGMENTS ...................................................................................................... 4 page LIST OF TABLES ................................................................................................................ 9 LIST OF FIGURES ............................................................................................................ 10 ABSTRACT ........................................................................................................................ 14 CHAPTER 1 INTRODUCTION ........................................................................................................ 16 1.1 Statement of Problem and Motivation .................................................................. 16 1.2 Scientific Approach ............................................................................................... 17 1.3 Organization of Dissertation ................................................................................. 19 1.4 Contributions to the Field ..................................................................................... 20 2 BACKGROUND .......................................................................................................... 22 2.1 Types of Gamma-ray Detectors ........................................................................... 22 2.1.1 Gas -Filled Detectors ................................................................................... 22 2.1.2 Scintillation Counters .................................................................................. 24 2.1.3 Semiconductor Detectors ........................................................................... 26 2.2 Mater ials for Room Temperature Semiconductor Detector ................................ 29 2.3 Physical Properties of BiI3 .................................................................................... 37 2.4 Thermodynamics of Crystal Growth from Melt .................................................... 41 2.4.1 Nucleation ................................................................................................... 42 2.4.1.1 Homogeneous Nucleation ................................................................ 42 2 .4.1.2 Heterogeneous Nucleation ............................................................... 43 2.4.2 Nucleation Rate .......................................................................................... 46 2.4.3 The Crystal Growth Face ............................................................................ 46 2.5 BiI3 as a Radiation Detector ................................................................................. 48 3 BISMUTH IODIDE SINGLE CRYSTAL GROWTH AND CHARACTERIZATION .... 52 3.1 Introduction ........................................................................................................... 52 3.2 Crystal Growth Techniques of BiI3 ....................................................................... 52 3.2.1 Vertical Bridgman Technique ..................................................................... 52 3.2.2 Physical Vapor Transportation Technique ................................................. 56 3.3 Experimental Procedures ..................................................................................... 58 3.3.1 Modified Vertical Bridgman Growth ............................................................ 58 3.3.2 Physical Vapor Transport Growth .............................................................. 60 3.3.3 Characterization .......................................................................................... 61 3.3.3.1 X-ray Diffraction ................................................................................ 61

PAGE 6

6 3.3.3.2 Vickers Hardness .............................................................................. 61 3.3.3.3 Etch Pit Density ................................................................................. 61 3.3.3.4 X-ray Rocking Curve ......................................................................... 62 3.3.3.5 Scanning Electron Microscopy (SEM) .............................................. 63 3.3.3.6 Atomic Force Microscopy (AFM) ...................................................... 63 3.3.3.7 X-ray Photoelectron Spectroscopy (XPS) ........................................ 64 3.3.3.8 X-ray Fluorescence Spectros copy (XRF) ......................................... 64 3.4 Results and Discussion ........................................................................................ 64 3.4.1 MVB Crystal Growth ................................................................................... 64 3.4.1.1 Determination of Temperature Profile .............................................. 64 3.4.1.2 Crystal Growth Result ....................................................................... 67 3.4.2 PVT Crystal Growth .................................................................................... 70 3.4.3 Crystal Structure and Phase Purity ............................................................ 70 3.4.4 Vickers Hardness ........................................................................................ 71 3.4.5 Etch Pit Density Measurement ................................................................... 73 3.4.6 X-ray Rocking Curve ................................................................................... 77 3.4.7 SEM Characterization ................................................................................. 80 3.4.8 AFM Characterization ................................................................................. 81 3.4.9 XPS Characterization ................................................................................. 82 3.4.10 XRF Characterization ............................................................................... 83 3.5 Summary and Conclusions .................................................................................. 84 4 BAND STRUCTURE CHARACTERIZATION OF BISMUTH IODIDE SINGLE CRYSTAL .................................................................................................................... 86 4.1 Introduction ........................................................................................................... 86 4.2 Experimental Procedures ..................................................................................... 87 4.2.1 UV -Vis Spectra ........................................................................................... 87 4.2.2 Inductively Coupled Plasma Atomic Emission Spectroscopy ................. 88 4.3 Results .................................................................................................................. 89 4.3.1 B and Gap Determined by Optical Spectra at Room Temperature ........... 89 4.3.2 Effect of Temperature on Band Gap .......................................................... 91 4.3.3 Effects of Impurities on Band Gap ............................................................. 94 4.4 Discussion ............................................................................................................. 96 4.5 Conclusions .......................................................................................................... 99 5 ELECTR ICAL CONTACTS AND RESISTIVITY CHARACTERIZATION ............... 101 5.1 Introduction ......................................................................................................... 101 5.2 Experiment Procedures ...................................................................................... 101 5.2.1 XPS ........................................................................................................... 101 5.2.1.1 Principles of XPS ............................................................................ 101 5.2.1.2 Sample Preparation for XPS .......................................................... 102 5.2.2 Atomic Force Microscopy (AFM) ............................................................... 103 5.2.3 I -V Characterization .................................................................................. 103 5 .3 Results and Discussion ...................................................................................... 103 5.3.1 The XPS Spectra ...................................................................................... 103

PAGE 7

7 5.3.1.1 Au Sample ....................................................................................... 107 5.3.1.2 Pd Sample ....................................................................................... 108 5.3.1.3 Pt Sample ........................................................................................ 111 5.3.2 AFM Analysis ............................................................................................ 113 5.3.3 I -V Characterization .................................................................................. 115 5.4 Conclusions ........................................................................................................ 116 6 SUMMARY AND FUTURE WORK .......................................................................... 117 6.1 Summary ............................................................................................................. 117 6.2 Future work ......................................................................................................... 120 6.2.1 Tilted Growth ............................................................................................. 120 6.2.2 Synthesis of High Purity BiI3 Powder ....................................................... 121 6.2.3 Sublimation of BiI3 Powder ....................................................................... 122 6.2.4 Post growth Annealing ............................................................................. 123 6.2.5 Encapsulation of the Crystal ..................................................................... 123 6.2.6 Surface Preparation .................................................................................. 123 APPENDIX A PURIFICATION OF BISMUTH IODIDE POWDER ................................................. 125 A.1 Introduction ......................................................................................................... 125 A.2 Experiment Procedures ..................................................................................... 125 A.3 Results and Discussion ...................................................................................... 125 A.3.1 Sublimation ............................................................................................... 125 A.3.2 Vertic al Zone -melting ............................................................................... 126 A.4 Summary and Conclusions ................................................................................ 127 B GROWTH OF BISMUTH IODIDE CRYSTAL WITH DIFFERENT DIAMETERS ... 129 B.1 Introduction ......................................................................................................... 129 B.2 Experimental Procedures ................................................................................... 129 B.3 Results and Discuss ion ...................................................................................... 130 B.4 Summary and Conclusions ................................................................................ 131 C TEM SAMPLE PREPARATION BY FOCUSED ION BEAM ................................... 132 C.1 Introduction ......................................................................................................... 132 C.2 Experimental Procedures and Results .............................................................. 132 C.3 Summary and Conclusions ................................................................................ 136 D NKN TAPE FOR PIEZOELECTRIC TRANSFORMER ........................................... 137 D.1 Introduction ......................................................................................................... 137 D.2 Experimental Procedures .................................................................................. 137 D.3 Results and Discussion ..................................................................................... 139 D.3.1 X -ray Diffraction ........................................................................................ 139

PAGE 8

8 D.3.2 Transformer Test ...................................................................................... 140 D.3.3 Tape Casting ............................................................................................ 142 D.4 Summary and Conclusions ................................................................................ 142 LIST OF REFERENCES ................................................................................................. 143 BIOGRAPHICAL SKETCH .............................................................................................. 160

PAGE 9

9 LIST OF TABLES Table page 2 -1 Properties of selected compound semiconductor materials for room temperature gamma-ray spectroscopy. ................................................................. 30 3 -1 Growth conditions for BiI3 single crystals in 0.75 diameter ampoules. ............... 67 3 -2 ................................................................................ 71 3 -3 Vickers hardness of (001) surface of BiI3 single crystals grown by the MVB method. ................................................................................................................... 73 3 -4 Etch pit density of single crystals grown by the MVB method. ............................. 77 3 -5 FWHM of X -ray rocking curves for BiI3 single crystals. ........................................ 78 3 -6 Peak properties of Bi 4f and I 3d for a freshly cleaved BiI3 crystal surface. ........ 83 3 -7 Quantitative r esults of XRF spectrum of BiI3 single crystal. ................................. 84 4 -1 Reported band gap of BiI3 single crystal. .............................................................. 87 4 -2 Temperature coefficient o f BiI3 calculated based on the fitting parameters. ........ 94 4 -3 Impurity concentrations of BiI3 powder and crystal (ppm). ................................... 95 5 -1 Obser ved binding energies of the elements Bi, I, C, O, Au, Pd, and Pt from XPS spectra. ......................................................................................................... 106 5 -2 Binding energies (eV) of the elements Bi, I, C, O, Au, Pd, and Pt in different compounds. .......................................................................................................... 107 5 -3 Peak properties of Bi 4f, Pd 3d, Pd 3p, C 1s, and O 1s observed in the Pd sample. ................................................................................................................. 110 5 -4 Peak properties of Bi 4f, I 3d, C 1s, and O 1s observed in the Pt sample. ........ 113 5 -5 Sample information and resistivity of BiI3 single crystals for I -V characterization. ................................................................................................... 116 A-1 Impurity concentrations of commercial BiI3 powder before and after sublimation (ppm). ................................................................................................ 126 B-1 Growth conditions for growing BiI3 crystals with different diameters. ................. 130 D -1 Batch formulation of tape casting slurry of NKN ................................................. 138

PAGE 10

10 LIST OF FIGURES Figure page 1 -1 Bi -I phase diagram. (re plotted after Rosztocz and Cubiccio.)4 .......................... 18 2 -1 General schematic of a gas -filled detector. Diagram after Ahmed.16 ................... 23 2 -2 Scheme of a scintillation counter. The scintillator and PMT are often optically linked by a light pipe. Diagram after Hastings and Marcia.17............................... 24 2 -3 Basic operating configuration of a planar detector. .............................................. 27 2 -4 Hexagonal representation of BiI3 crystal structure, a) unit cell consists of three I -Bi I layers; b) projection of one I -Bi -I layer indicating missing Bi sites. Smaller Bi3 + ions appearing in blue, while the larger I ions are presented in red. The dashed box represents one unit cell. ..................................................... 38 2 -5 The absorption spectrum near the indirect edge and its stacking disord er effect.110 .................................................................................................................. 41 2 -6 The nucleation of a spherical solid cap at a liquid-substrate interface. ................ 44 2 -7 Variation of free energy of interface with fraction of sites filled.125 ....................... 48 2 -8 Intrinsic photopeak efficiency as a function of energy for 1 cm3 detectors located 1 cm from a surface source. Replotted after Lintereur e t al. .127 ............ 49 3 -1 A typical Bridgman system. (a) Schematic diagram of the furnace; (b) longitudinal temperature profile at the center of the furnace. Diagram after Brice.124 ................................................................................................................... 53 3 -2 Schematic diagram of physical vapor transport (PVT) system. Diagram after Isshiki and Wang.144 ............................................................................................... 5 7 3 -3 XPert system for X -ray rocking curv e measurement. .......................................... 62 3 -4 Illust -scan. ...................................................................... 63 3 -5 -scan. The shadowed bands numbered 1 and 2 represent the X -ray incident area for each scan. ................... 63 3 -6 Growth recipe of the crystal growth at 10oC/cm and 0.5 mm/h. ........................... 67 3 -7 Single crystals produced by MVB method, (a) as grown polycrystall ine rod; (b) crystal grown at 10oC/cm, 0.5 mm/h; (c) crystal grown at 10oC/cm, 1 mm/h; (d) crystal grown at 15oC/cm, 0.5 mm/h; (e) crystal grown at 15oC/cm, 1 mm/h; (f) polycrystalline cross section obtained at rest conditions. .................. 68

PAGE 11

11 3 -8 BiI3 single crystal with dimension of 18 x 13 x 5 mm3, the largest crystal size ever reported. ......................................................................................................... 69 3 -9 BiI3 single crystals grown by the PVT method. ..................................................... 70 3 -10 A typical X -ray diffraction pattern of BiI3 single crystals. ....................................... 71 3 -11 Illustration of Vickers hardness test. ...................................................................... 72 3 -12 BiI3 etched in different concentration of KI solutions. The unit of weight loss is mg/cm2. ............................................................................................................... 73 3 -13 Optical images of BiI3 single cry stal surface after etched for time of (a) 10 sec; (b) 20 sec; (c) 30 sec; (d) 40 sec; (e) 50 sec; (f) 60 sec. .............................. 75 3 -14 Etch pits of BiI3 single crystals grown by MVB and PVT methods. (a) (c): MVB crystals etched for 10 sec, the crystal was grown at (a) 10oC/cm, 0.5 mm/h; (b) 10oC/cm, 1 mm/h; (c) 15oC/cm, 0.5 mm/h; (d) PVT crystal etched for 10 sec; (e) PVT crystal etched for 60 sec. ....................................................... 76 3 -15 X -scans of the single crystals. (a) (c): single crystals grown by MVB method at a) 10oC/cm and 1 mm/h; b) 10oC/cm and 0.5 mm/h; c) 15oC/cm and 0.5 mm/h; (d) single crystal grown by PVT method. Solid line linear intensity; circles log scale intensity. ...................................... 78 3 -16 -3 single crystals indicated the wide distribution of d -spacing. ........................................................................................................... 79 3 -17 X -ray rocking curves of BiI3 (0012) planes by Nason and Keller. A and B represent the rocking curves obtained at different sample locations.10 ............... 80 3 -18 Deformation of BiI3 caused by cutti ng as in circled area. ..................................... 80 3 -19 SEM images of BiI3 (001) surfaces show (a) deformation induced by cutting; (b) layered structure of (001) planes. .................................................................... 81 3 -20 AFM images of BiI3 (001) cleaved surfaces (a) topographic morphology; (b) 3D view. .................................................................................................................. 81 3 -21 XPS spectrum of freshly cleaved BiI3 single crystal. ............................................. 82 3 -22 XPS spectra of Bi 4f and I 3d for a freshly cleaved BiI3 crystal surface. .............. 83 3 -23 XRF spectrum of BiI3 on a freshly cleaved crystal surface. .................................. 84 4 -1 Illustration of sample cutting for ICP -AES characterization. ................................. 89 4 -2 UV -Vis reflection and transmission spectra of BiI3 sing le crystal. ........................ 90

PAGE 12

12 4 -3 1/2 2 as a function of photon energy. ........................... 90 4 -4 UV -Vis transmission spectra of BiI3 single crystal at different temperatures. ...... 91 4 -5 1/2 2 on the photon energy. ................................................................................................. 92 4 -6 Impurity distribution along the BiI3 crystal grown in Pyrex amp oule. .................... 96 4 -8 spectral range from 1.88 to 5.15 eV and transmission and reflection below 1.8 eV. Spectroscopic elli 1, while transmission and reflection measurements for a thick, bulk sample are only 1. As spectroscopic ellipsometry is a reflection-based technique relying on the complex amplitude reflectance r atios, it retains accurate values when the material thickness is greater than the penetration 185 ............... 98 5 -1 XPS spectra of Bi 4f and I 3d for the freshly cleaved BiI3 (001) surface and surface with Au, Pd, and Pt electrodes. .............................................................. 104 5 -2 X -ra y photoelectron spectra of Au 4f, Pd 3d, and Pt 4f on the surface of BiI3 single crystals. ...................................................................................................... 105 5 -3 X -ray photoelectron spectra of C 1s and O 1s for all the samples. .................... 106 5 -4 Deconvolution of XPS spectra for the Au sample peaks (a) C 1s and (b) O 1s. ......................................................................................................................... 108 5 -5 Deconvolution of XPS spectra of the Pd sample for (a) Bi 4f, (b ) Pd 3d, (c) C 1s, and (d) O 1s. ................................................................................................... 109 5 -6 Deconvolution of XPS spectra of the Pt sample for (a) Bi 4f, (b) I 3d, (c) C 1s, and (d) O 1s. ......................................................................................................... 112 5 -7 AFM image and 3D view of topography for the (a) Au sample, (b) Pd sample, and (c) Pt sample. ................................................................................................ 114 5 -8 I -V curves for BiI3 single crystals with Au, Pd, and Pt electrodes. ...................... 115 6 -1 Illustration of the crystal in a tipped ampoule. ..................................................... 121 6 -2 Illustration for the sublimation growth of BiI3. (After Curtis and Brunner213) ....... 122 A-1 BiI3 powder in Pyrex ampoule after sublimation. ................................................ 126 A-2 Impurity concentrations of BiI3 powder after vertical z onemelting refinement. 127

PAGE 13

13 B-1 BiI3 single crystal with 0.5 diameter. (a) the as grown crystal rod, and (b) cleaved single crystal (001) surface. ................................................................... 130 C -1 Pt deposition. ........................................................................................................ 133 C -2 Carbon layer removal. .......................................................................................... 133 C -3 Formation of trench. ............................................................................................. 134 C -4 ........................................................................................ 134 C -5 ..................................................................................... 135 C -6 Undercut. .............................................................................................................. 135 C -7 Final thinning down to 100 nm. ............................................................................ 136 D -1 Rosen type configuration of prototype NKN transformer. ................................... 138 D -2 XRD patt ern of calcined and sintered NKN. ........................................................ 139 D -3 Equivalent circuits of a piezoelectric transformer: (a) general model, (b) after reflecting the output capacitance and load resistance to the primary, and (c) after parallel to series transformation. (After Ivensky et al ..223) ......................... 141 D -4 NKN tape casted on a Mylar sheet. ..................................................................... 142

PAGE 14

14 Abstract of Dissertation Present ed to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy GROWTH AND CHARACTERIZATION OF BISMUTH TRI -IODIDE SINGLE CRYSTALS BY MODIFIED VERTICAL BRIDGMAN METHOD By W ei Q iu August 20 10 Chair: Juan C. Nino Major: Materials Science and Engineering Bismuth tri -iodide (BiI3) is a wide band gap semiconductor material with potential for room temperature gamma-ray detection applications. The inability to produce high quality, pure, and large single crystals is one of the major challenges for this material preventing its use in gamma-ray detection. This work presents the growth and characterization of BiI3 single crystals by a modified vertical Bridgman (MVB) method. The growth parameters to produce BiI3 single crystals were explored by adjusting growth rate and temperature gradient at the solid-liquid interface. Single crystals of BiI3 have been successfully grown in Pyrex glass ampoule at different growth conditions Through etch pit density (EPD) and X -ray rocking curve measurement, the crystal with the best quality was determined to be obtained at 0.5 mm/h growth rate and 10oC/cm temperature gradient. Single crystal (001) oriented slabs up to 18 x 13 x 5 mm3 (the largest BiI3 single crystal ever reported) were obtained under this condition. Impurity characterization on the starting BiI3 powder and the grown crystals indicated a relatively high total impurity concentration up to 1017 cm3, and the crystal growth using Pyrex glass does not introduce additional impurities to the crystal.

PAGE 15

15 The energy band gap is an important parameter for materials used as a room temperature gamma-ray detector, as it determines the energy to produce an electronhole pair which is the s ource of the signal for the detector. Remarkably, very different band gap characteristics and values of BiI3 have been reported in literature. This study characterized the band gap of BiI3 through Ultraviolet visible spectroscopy (UV -Vis), which yielded an indirect band gap of 1.68 0.09 eV at room temperature. Impurity and temperature effects as commonly observed in many semiconductors, have been investigated as the major extrinsic factors that influence the band gap value of BiI3. Three different met als, Au, Pd, and Pt were sputtered on the surface of BiI3 single crystal for fabrication of the electrode. The metal -BiI3 interface was characterized by X ray photoelectron spectroscopy (XPS). Chemical reactions were observed for the surface sputtered wi th Pd and Pt electrode, while no reaction was observed on the surface with Au electrode. It was proposed that Au should be used for the contact material for BiI3. Current voltage characterization indicated that BiI3 crystal with Au electrode exhibited th e highest resistivity with values in the order of magnitude of 109 cm, which is sufficiently large for the application of a room temperature gamma-ray detector. Based on the work presented in this dissertation a series of avenues for future work are pr esented to be able to take this material into actual devices

PAGE 16

16 CHAPTER 1 INTRODUCTION 1.1 Statement of Problem and Motivation Gamma -ray detectors have become increasingly important for the applications of medical imaging, environmental safety, and homeland security. Most of these applications desire a portable device with high gamma-ray energy resolution and detection efficiency, compact size, room temperature operation, and low cost. The most widely used gamma-ray detectors include gas -filled detectors scintillation counters, and semiconductor detectors, among which semiconductor detectors received most of the attention due to their much better energy resolutions. The great success of high purity Ge and Si detectors encouraged the discovery and devel opment of new semiconductor materials for radiation detection.1 The operation of high purity Ge and Si detectors requires liquid N2 cooling due to their small band gaps, resulting in a bulky and heavy system that limits their use. Compound semiconductor materials with large band gap overcome this limit and become the most promising candidates as an alternative to high purity Ge and Si. In the past few decades, many compound semiconductor materials hav e been extensively studied. Although not all of these materials are readily available for gamma -ray detection, several compound semiconductors including CdTe, CdZnTe, and HgI2 have demonstrated significant energy resolution and detection efficiency.2 Even though CdTe and CdZnTe have been commercialized for almost 25 years, the high cost of fabrication and limited supply of high quality crystals prevent the large scale production.3 The discovery and development of new materials is still the main focus of current research for room temperature gamma ray detect ors.

PAGE 17

17 BiI3 is a wide band gap semiconductor material that has a potential for room temperature gamma ray detection and spectroscopy. This material has a much higher effective atomic number than Ge and Si, thus has higher gamma -ray detection efficiency. Unfortunately, the growth of large volume and high quality crystals remains as the major challenge for BiI3 to be commercialized. The present work focuses on the growth and evaluation of BiI3 single crystals which are grown by a modified vertical Bridgman m ethod. The general material properties of BiI3 (hardness, band gap, electrical resistivity, etc.) have also been investigated. 1.2 Scientific Approach BiI3 has a low melting point of 408oC. Unlike HgI2 which experiences a phase transformation from -HgI2 to -HgI2 below melting point, BiI3 is stable within the range from the melting point to room temperature, as shown in the Bi -I phase diagram ( Figure 1 -1 ).4 BiI3 was reported to have a vapor pressure of around 90 mbar at the melting point and 1 13 mbar at temperature s of 250oC to 300oC.5 6 Therefore BiI3 single crystals can be grown from both liquid and vapor phases. Single crystals of BiI3 have been successfully grown by physical vapor transportation (PVT) method7 10 and vertical Bridgman technique.11 12 For the application of gamma -ray de tection, single crystal with large volume (~1 cm3) is desired as higher interaction volume increases the det ection efficiency. While PVT method generally produces crystals with small volumes up to several mm3, vertical Bridgman technique has the potential to grow much larger crystals up to tens of cm3. Therefore, the present work chooses vertical Bridgman technique to grow BiI3 single crystals with some modifications.

PAGE 18

18 0 20 40 60 80 100 0 100 200 300 400 500 Temperature (oC)Atom Percent Two LiquidsBiI BiI3BiI+Liquid BiI3+Liquid Bi+BiI BiI+BiI3BiI3+Liquid BiI3+I2Bi 1/2 I2 Figure 1 1. Bi I phase diagram. (re p l otted after Rosztocz and Cubiccio.)4 As crystal growth is a time-consuming process, the temperature profile has been carefully designed and tested. Due to the unavailability of the thermal properties of BiI3, the growth conditions have been determined on an educated initial guess basis. T he grown single crystals were characterized by various techniques and research tools. The crystal structure was identified by X -ray diffraction (XRD). The crystal quality was characterized by etch pit density (EPD) measurement and X ray rocking curve -s cans. The impurity concentration in the crystal was characterized by inductively coupled plasma atomic emission spectroscopy (ICP AES). The surface morphology was investigated by scanning electron microscopy (SEM) and atomic force microscopy (AFM). Th e elemental composition of the crystal was characterized by X -ray photoelectron spectroscopy (XPS). The band gap of BiI3 has been reported with different values and characteristics. This work investigated the band gap of BiI3 through UV -Vis transmission and reflection spectra. By comparing and analyzing the results obtained from ellipsometry and theoretical calculation using density functional theory (DFT), the band gap of BiI3 was

PAGE 19

19 determined, and the factors that affect the band gap were analyzed to prov ide possible solution to the discrepancy. Palladium and gold have been used as the electrical contact materials for BiI3.1213 This work investigated the interface between BiI3 single crystal and different contact materials of Au, Pd, and Pt using XPS. The morphology of the electroded surface was characterized by AFM. The resistivity of BiI3 with different electrode was determined by measuring the I -V characteristic. The compatibility of these contact materials w ith BiI3 was determined by the above characterizations. Finally, based on the study of this work, suggestions for future work are proposed for better control of crystal growth and post -growth treatment. 1.3 Organization of Dissertation Chapter 2 provides a background that covers the fundamentals of gamma-ray detectors and some material properties of BiI3. The compound semiconductors that have been studied for room temperature gamma-ray detector are discussed in detail. The purpose of this is to help the r eader understand the working principles and the important material properties of a gamma-ray detector. The crystal structure of BiI3 and structure related optical properties are discussed following the detector description. Then the fundamental thermodynamics of crystal growth including nucleation and the crystal growth interface are discussed. Finally, the previous research of developing BiI3 towards a room temperature gamma-ray detector is summarized to provide the reader the background of this work. C hapter 3 discusses the crystal growth by a modified vertical Bridgman (MVB) and PVT method. An in depth discussion on the thermodynamics of vertical Bridgman and PVT growth is presented. The detailed growth procedures and crystal characterizations

PAGE 20

20 are de scribed and results are presented. The relationship between the growth conditions and the crystal quality is discussed. Chapter 4 discusses the band gap characterization of BiI3 single crystals. The previous work with different band gap values and charac teristics is reviewed. The experimental results on UV -Vis spectra and impurity characterization are presented. The possible reasons caus ing the discrepancy in literature about the band gap are discussed. Chapter 5 discusses the BiI3electrode interface c haracterized by XPS. The elemental compositions on BiI3 surface with Au, Pd, and Pt electrode are presented and discussed. The resistivity of BiI3 with different electrodes is compared. Recommendation for the electrode material of BiI3 is given based on the experimental results. Finally in Chapter 6, a summary of the dissertation is presented and the future work in the relevant research areas is discussed. At the end of the dissertation, four appendices are provided. Appendix A presents the purification results of BiI3 powder using sublimation and vertical zone melting method; Appendix B presents the attempts to grow BiI3 single crystal with different diameters using MVB method; Appendix C presents the sample preparation of BiI3 single crystal for TEM ch aracterization using focused ion beam (FIB) technique; Appendix D presents the synthesis of sodium potassium niobate (Na0.5K0.5NbO5, NKN) for the application of piezoelectric transformer. 1.4 Contributions to the Field The main contributions of this disser tation to the development of BiI3 single crystal as a room temperature gamma -ray detector are summarized as follows:

PAGE 21

21 The complete temperature profile of growing BiI3 single crystals using MVB method was established. The parameters including heating rate, soaking temperature and time, ending temperature of crystal growth, and post growth cooling rate were determined. The growth conditions of producing large volume BiI3 single crystals were determined for MVB method, r esulting in the largest grown BiI3 singl e crystal ever reported (18 x 13 x 5 mm3). The procedures to measure the etch pit density of BiI3 single crystals were established. The etchant concentration and etching time were determined and could serve as a standard for BiI3 single crystal. The poss ible solutions to the discrepancy of band gap values and characteristics for BiI3 were proposed. The proper electrode material for BiI3 single crystal was determined to be Au. The result suggested that using Pd as the electrode for iodide compound may not be appropriate. This will lead to the reevaluation of the electrode materials for all the iodine based compound semiconductors such as HgI2 and PbI2.

PAGE 22

22 CHAPTER 2 BACKGROUND 2.1 Types of Gamma -ray Detectors Gamma -rays interact with materials primarily through three mechanisms:14 photoelectric absorption, Compton scattering, and pair production. In the photoelectric absorption process, the photon transfers all its energy to an orbital electron. This photoelectron then interacts with the lattice of the detector material and creates many electron -hole pairs until it completely loses its kinetic energy. In the Compton scattering process, the photo n only transfers partial of its total energy to an orbital electron leaving the material with a degraded photon. In the pair production process, a photon with energy higher than 2mec2 interacts in the vicinity of a nucleus and produces an electron and a p ositron pair. Based on which interaction mechanism the gamma-ray detector utilizes, there are two broad classes of gamma-ray detectors. The first is generally called gamma -ray spectrometer which records the gamma -ray energy by collecting photoelectrons. These types of detectors include gas -filled detectors, scintillation counters, and semiconductor detectors. The second class of detectors includes those used in gamma-ray telescopes and imaging systems. Detectors of this type register the direction and energy of the incoming gamma-rays through Compton scattering or pair production. Here the discussion will be focused on gamma-ray spectrometers 2.1.1 GasFilled Detectors Gas -filled detectors consist of a metal vessel filled with gas (argon or helium), a gamma -ray transparent window (such as Be or mylar), and a center wire functioning as an anode.15 A high voltage is maintained between the metal vessel and the anode (Figure 2-1) When high energy rays or particles collide with an inert gas atom, they

PAGE 23

23 ionize the atom and transfer a massive amount of energy to an outer shell electron, kn own as photoelectron. The photoelectron releases its excess kinetic energy by subsequently colliding with other outer shell electrons. As a result, a cascade of accelerated electrons are collected by the anode and detected as an electrical pulse. Figure 2 1 General schematic of a gas filled detector. Diagram after Ahmed. 16 If the voltage between the electrodes is increased, the electrons will obtain enough kinetic energy to cause further ionizations. By controlling the gas pressure and the operation voltage, a proportional counter can be tailored and constructed such that the output signal is proportional to the deposited gamma -ray energy. Proportional counters have been used for gamma-ray spectroscopy with low energy detection (< 100 keV).17 If the operation voltage is further increased, charge multiplication increases (avalanches) until the space charge produced by the residual ions inhibits further ionization. As a result, the collected charge pulse is no longer proportional to the energy deposited in the gas. This type of detector is known as the Geiger Mueller (GM) detector.15 A GM tube detector does not differentiate the energies of the particles it + w ire anode inert gas cathode insulator signal high energy rays

PAGE 24

24 detects; it only counts the number of particles entering the detector. This type of detector is widely used for dosimetery in health physics.17 Gas -filled detectors account for only a small portion of gamma-ray spectrometers. The scintillation and semiconductor detectors are much more desirable for obtaining the spectroscopic detail of gamma-rays in the energy range from 100 to 1000 keV. 2.1.2 Scintillation Counters A scintillation counter consists of two main components: a scintillator and a photomultiplier tube (PMT). A scintillator is a material that emits low energy photons (usually in t he visible range) when absorbing radiation. The light produced from the scintillation process is then converted to electrical signal by a PMT. A PMT is normally composed of a photocathode, an electron -multiplier section, and an anode.17 The photocathode has the unique characteristic of producing electrons when light strikes on its surface. These electrons are accelerated through a series of secondary emission electrodes (known as dynodes), knocking out a number of new electrons subsequently, multiplying their number in each step. After multiplication, the electrons are collected by an anode to form an electronic pulse ( Figure 22 ). In a typical PMT, the multiplication factor is on the order of 107.14 Figure 2 2 Scheme of a scintillation counter. The scintillator and PMT are often optically linked by a light pipe. Diagram after Hastings and Marcia.17 Scintillator Anode L ight pipe Photocathode Dynodes Photomultiplier tube

PAGE 25

25 Scintillators can be made of organic or inorganic materials, depending on the intended applications. Inorganic scintillators are more widely used due to their stability and higher density which leads to higher detection efficiency. Examples of organic scintillators are organic crystals, plastics, and liquids.14 Organic crystal scintillators are composed of ar omatic hydrocarbons such as anthracene (C14H10).18 They are very durable but hard to be machined. However, they are not often used due to the lack of success to grow large single crystals. Plastic scintillators are solutions of fluorescent organic compounds dissolved in a solidified polymer matrix. Plastic scintillators give fast signal and relatively high light output. They can also be easily shaped and machined to the forms desired in detectors. Due to their low intrinsic s cintillation efficiency, plastic scintillators are not suitable for detection of lower energy gamma-rays and cannot be used for gamma-ray spectroscopy. However, they are useful for gross counting gamma-rays with energies higher than 100 keV.18 Their relatively low cost and availability as large size sheets make them the most widely used scintillators. Liquid scintillators are fluid solutions with similar fluorescent organic compounds as in plastic scintillators. They can be eas ily loaded with other additives such as wavelength shifters to match the spectral sensitivity range of a particular PMT Liquid scintillators are normally required to be sealed to avoid the degradation caused by dissolved oxygen. Inorganic scintillators a re usually made of alkali halides (i.e., NaI, CsI) in which the high Z component enhances the photoelectric interaction efficiency.14 Bismuth ge rmanate (Bi4Ge3O12, BGO) is another popular scintillation material which has a very high gamma -ray counting efficiency, but low energy resolution.19 When gamma-rays interact with scintillation materials, electrons can jum p up to higher energy levels

PAGE 26

26 through excitation, then relax back to a lower energy state by emitting photons of light. In a pure inorganic scintillation material, the relaxation with the emission of a photon has very low efficiency. Furthermore, the emit ted photons usually possess higher energy to which the PMT is sensitive. Small amount of impurities (activators) are often added to the scintillator to enhance the emission of visible photons Thallium and sodium are often used for this purpose. For sci ntillation detector described as NaI(Tl), it means that it is a sodium iodide crystal with a thallium activator Inorganic scintillators have been widely used in process industry, nuclear medicine, life sciences, and many spacebased missions to observe s ources of cosmic gamma -radiation .17 2.1.3 Semiconductor Detectors In semiconductor detectors, the charge produced by the photon interactions is directly collected. Gamma -rays incident to the detector excite electron hole pairs in pr oportion to the energy deposited in the detector ( Figure 2-3 ). By applying a voltage across contacts on opposite sides, these free charges will drift through the detector and create an induced charge on the electrodes proportional to the distance traveled by the electrons and holes, as described by Shockley Ramo theorem,2021 dx L qN dQ0 (2 -1 ) where Q is the induced charge, N0 is the initial number of electron hole pairs, and L is the detector thickness. These charges are integrated by an external circuit and a pulse height spectrum is generated with the peak positions in the spectrum proportional to the energies of the incident photons. If the charge t rapping can be ignored, the total induced charge is equal to the initial created charge, Q = qN0. However, for most compound semiconductors some extent of charge traps are always present and the loss

PAGE 27

27 from charge carrier trapping has to be taken into accou nt. The charge collection efficiency (CCE) describes the fraction of charges that reach the electrodes, and is given by the Hecht equation:22 ) exp( 1 ) ) ( exp( 10 h h h h e e e e v x L v v x L L v Q Q CCE (2 -2 ) where L is the detector thickness, x is the distance from the cathode to the location of the radiation interaction, ve and vh are the drift velocity of electrons and holes, and e and h are the electron and hole lifetimes. Equation ( 2 -2 ) indicates that the CCE depends not only on the mean free paths (v ) of electrons and holes, but also on the location where the charges are created. This provides ways to improve the detector energy resolution for a specific material by varying the detector shape, contact geometry, and detector configurations.14,20,23 24 Figure 2 3 Basic operating configuration of a planar detector. The energy resolution of semiconductor detectors is dramatically higher than that of a scintillation or gas -filled detector, which makes it an excellent candidate for gamma ray spectroscopy. Lithium dr ifted germanium (Ge(Li)) detectors were developed in the early 1960s25 and remained the most popular design for gamma -ray spectroscopy for two decades. Lithium is drifted through the material to compensate th e residual V + V gamma ray h + e x L

PAGE 28

28 acceptor impurities after the crystal has been grown. The drifted region can easily be depleted over its entire thickness which allows charges (electrons and holes) to move freely during the charge collection process. In the early 1980s, high-purity Ge (HPGe) became widely available and rapidly replaced Ge(Li).14 Whereas Ge(Li) detectors must be continuously maintained at liquid nitro gen temperature (77 K) to keep the lithium drifting at a frozen stage, HPGe detectors can be allowed to warm to room temperature between uses. This provides a much greater operational convenience and simplifies the fabrication of detectors. In general, G e(Li) and HPGe detectors have essentially identical detection efficiency and energy resolution. Besides Ge, lithium drifted Si (Si(Li)) has also been successfully fabricated for gamma -ray spectrometer.14 A disadvantage for single element semiconductor detectors is that they must be operated at low temperatures to reduce the thermally generated leakage current, resulting in a bulky system that lim its their use in the field. In principle, compound semiconductors with a wider bandgap (greater than 1.5 eV) could reduce this thermally generated leakage current so that operation at room temperature is possible. However, since Van Heerden first demonst rated that silver chloride (AgCl) crystal was able to detect gamma -ray and alpha particles in 1945,26 compou nd semiconductors have experienced slow development over the years, largely due to the difficulties of obtaining large volumes, impurities and defects free crystals. So far the most widely explored compound semiconductors for room temperature gamma-ray de tection are CdTe, Cd1 xZnxTe, and HgI2,14 which will be discussed in detail in the following section. Photoelectric absorption is the favored gam ma -ray interaction for a spectrometer since only this process leads to the total absorption of the incident energy. The

PAGE 29

29 interaction cross -section for photoelectric absorption varies as NZ5, where N is the number of atoms per unit volume and Z is the effec tive atomic number. Thus, high Z material should be chosen for efficient gamma -ray spectroscopy. This makes another advantage of compound semiconductor detectors over single element detectors by incorporating high Z elements in the compounds. For semiconductor compounds, the photoelectric absorption is most influenced by the element with the highest atomic number, instead of the average Z. For example, based on the XCOM Photon Cross Sections Database from NIST, while both GaAs and InP have the same aver age Z, InP has two to three times higher absorption coefficients in the hard X -ray region due to the higher Z of In (49) compared to Ga (31) and As (33). Therefore, the much higher atomic numbers of semiconductor materials, such as CdTe (ZTe = 52) and HgI2 (ZHg = 80), determine the significantly higher photoelectric absorption than Ge (Z = 32) and Si (Z = 14). The detection efficiency per unit volume will therefore be higher. 2.2 Materials for Room Temperature Semiconductor Detector The general material r equirements for a room temperature gamma-ray detector include high density and atomic number s which lead to superior energy resolution and detection efficiency; moderate bandgap between 1.5 eV and 3 eV which prevents thermally generated leakage current at room temperature and allows as many electronhole pairs to be excited and high resistivity, which improves the signal to noise ratio. Additionally, high quality material in large volume is required for high gamma-ray interaction efficiency. Various co mpound semiconductor materials have been investigated for the potential use as a detector for room temperature gamma-ray spectroscopy. Table 21 shows the properties of a selection of these materials: gallium arsenide (GaAs), cadmium telluride (CdTe), cad mium zinc telluride (Cd 1 x Zn x Te),

PAGE 30

30 mercuric iodide (HgI2), lead iodide (PbI2), indium phosphate (InP), thallium bromide (TlBr), and bismuth iodide ( BiI3) Table 2 1 Properties of selected compound semiconductor materials for room temperature gamma-ray spe ctroscopy. Detector Material Atomic Numbers Density (g/cm3) Bandgap (eV) Work Function (eV/e h) References GaAs 31, 33 5.4 1.4 4.2 27 28 CdTe 48, 52 6.1 1.5 4.4 27 28 CdZn Te 48, 30, 52 6.1 1. 6 4.7 28 HgI 2 80, 53 6.4 2.1 4. 4 27 28 PbI 2 82, 53 6.2 2. 6 4.9 28 InP 49, 15 4.8 1.4 4.2 28 TlBr 81, 53 7.6 2.7 6.5 28 BiI 3 83, 53 5.8 1.7 5.5 27 GaAs has been studied as a radiation detector since the early 1960s29 and was the first compound semiconductor material that clearly demonstrated good gamma ray energy resolut ion at room temperature.30 GaAs crystallizes in a cubic, zincblende structure (space group m F 3 4 No. 216) with four formula units per unit cell. GaAs has high charge carrier mobilities o f 9000 cm2/Vs for electrons and 500 cm2/Vs for holes.27 The resistivity of GaAs crystal is on the order of 107 108 -cm, which is relatively low for a high resolution gamma-ray spectrometer. To reduce the leakage current, detector designs often employ Schottky barrier blocking contact s.30 31 GaAs has close atomic numbers to Ge, therefore the photoelectric absorption efficiency should be similar to Ge. High quality GaAs single crystals have been grown by liquid phase epitaxial (LPE) technique, but relatively small in volume and re peatability issues have prevented their commercial application. Attempts have been made to produce larger volume GaAs single crystals by bulk growth techniques such as vertical zone melt (VZM)32 and vertical gradient freeze (VGF).33 Unfortunately, charge trapping effects resulting from

PAGE 31

31 the high impurity concentration have prevented the commercial realization of bulk GaAs as radiation spectrometers.34 CdTe has b een studied for room temperature gamma-ray detector since the mid1960s.35 37 CdTe has the same crystal structure as GaAs (space group m F 3 4 No. 216) with four formula units per unit cell. The c harge carrier mobilities are 1000 cm2/Vs for electrons and 80 cm2/Vs for holes.38 The resistivity of CdTe crystal is on the order of 108 109 -cm.39 The photoelectric absorption efficiency is roughly a factor of 4 to 5 times higher than in Ge, and 100 times higher than in Si for typical gamma -ray energies. CdTe crystals have been grown using various techniques. Floating-zone melt growth has yielded crystals with low etch pit densities (EPD) on the order of 102 cm2.40 T he crystal rod has a moderate diameter of approximately 5 mm, primarily limited by the small surface tension and high density of the melt. The resistivity of the grown crystal was 106 to108 -cm. Variations of Bridgman growth methods (including vertical, horizontal and high pressure) produced large ingots of CdTe.41 43 T h e resistivity was on the same order of magnitude as in the crystal grown by floatingzone. Crystals grown by these techniques are generally les s suitable for gamma -ray spectrometers due to the inclusions of background impurities during the growth process.2 The traveling heater method (THM) is more commonly used to produce detector grade CdTe for several reasons including lower growth temperatures, reduced impurity concentrations, and lower defect concentrations than typical Bridgman methods.38 THM CdTe is often Cl doped to compensate background impurities and native defects, resulting in p type material with high resistivity up to 109 -cm.44 45 Due to the poor hole mobility, energy resolution of CdTe detectors is generally not comparable with that of Si

PAGE 32

32 or Ge. T h e best energy resolution of CdTe detectors has been reported to be approximately 1.2% full width at half maximum (FWHM, 7.9 keV) for a 137Cs gamma -ray source (662 keV),46 while HPGe detectors can achieve 0.1% FWHM at 662 keV.47 Commercially available CdTe detectors have diameter ranging from 1 mm to over 1 cm, but charge trapping and nonun iformities in the electrical properties restrict detector thickness to a few millimeters or less. This makes CdTe detectors have less detection efficiency and poor spectroscopic performance for highenergy gamma -rays. Another drawback of CdTe detectors i s polarization effect, which leads to a time-dependent decrease in the counting rate and charge collection efficiency.48 This polariza tion is largely related to the capture of electrons by acceptor type impurities within the material. The resulting space charge buildup counteracts the charge collection and decreases the depletion region of the detector. The use of Cd1 xZnxTe (CZT) cryst als as gamma-ray detectors started around 1991.49 CZT has the same crystal structure as CdTe, with Zn substituting a fraction x of the Cd atoms, where x ranges from 0.04 to 0.2. The band gap energy at room temperature varies with Zn concentration, and is given by Olego et al .50 eV x x x E2 0139 0 606 0 510 1 ) ( 2 -3 in which the band gap energy of Cd0.96Zn0.04Te and Cd0.8Zn0.2Te are around 1.534 eV and 1.637 eV, respectively. The photoelectric absorption efficiency is similar to that of CdTe. The resistivity of CZT crystal is on the order of 1010 1011 -cm.51 The increased band gap reduces the intrinsic charge carrier concentrations. Furthermore, the operating leakage current i s generally below that of CdTe devices due to the higher resistivity. The measured charge carrier mobilities are 1350 cm2/Vs for electrons and 120 cm2/Vs for holes.52 Similar to CdTe, the reported hole life time is much shorter than

PAGE 33

33 the electron life time, which produce s a condition in which position dependent charge collection can degrade energy resolution.53 High resistivity CZT is generally grown through the unseeded highpressure vertical Bridgman (HPVB) technique.54 HPVB CZT ingots generally have high defect densities with reported EPD on the order of 104 cm2.55 The defects can be metal inclusions twins, grain boundaries, and in some cases severe polycrystallinity. Consequently, detector grade samples have to be painstakingly identified and separated from the bulk ingots, which increases the overall cost of the detectors. Despite the difficulties of obtaining high quality crystal samples, CZT detectors have shown significant improvement over CdTe in both size and spectrometric performance.49,56 T h e best energy resolution of CZT detectors has been reported to be of approximately 0.82% FWHM for a 137Cs gamma -ray source (662 keV).57 Unlike CdTe detectors, CZT detectors do not appear the phenomenon of polarization. CZT has become a very promising material for room temperature gamma ray spectrometer and efforts to improve the crystal quality and detector performance are still ongoing. HgI2 has been investigated as a room temperature gamma -ray detector since the early 1970s .5860 HgI2 crystallizes in a tetragonal structure (space group P42/nmc No. 137) at room temperature, appearing red color ( phase) with a band gap energy of 2.13 eV. The large band gap means this material has hi gher resistivity (1012 1013 cm) compared to CdTe or CZT. -HgI2 has a layered structure with cleavage planes parallel to the (001) planes. It undergoes a phase transformation to an orthorhombic structure (space group Cmc21 No. 36) at 127oC, appearing a yellow color ( phase) with a band gap of 2.5 eV.61 Because of this, HgI2 detectors are usually operated in

PAGE 34

34 environments with temperature below 127oC, and the detector material is always HgI2. The charge carrier mobilities of -HgI2 are 100 cm2/Vs for electrons and 4 cm2/Vs for holes. The photoelectric absorption efficiency is much higher than CdTe and CZT detectors due to the higher atomic number of Hg (80), which means HgI2 detector can be relatively smaller for a required gamma-ray detection efficiency. For example, for 140 keV gamma -rays a 2 mm thick HgI2 detectors have comparable photoelectric absorption efficiency to a 4.2 mm thick CdTe or CZT detectors.2 Due to the phase transformation, -HgI2 single crystals cannot be grown from melt. HgI2 crystals were initially produced by solution growth technique.62 Radiation detectors fabricated from the solution grown material demonstrated some gamma-ray spectroscopic performance. However, crystals produced by this method are relatively small; impurity and solvent inclusions are commonly found in the grown crystal. Vapor phase processes including vertical and horizontal vapor growth are generally preferred for the growth of HgI2 c rystals. Vapor phase methods allow the growth of large HgI2 crystals with high purity and low dislocation densities.63 65 HgI2 single crystals are generally collected from bulk crystals by cleaving samples perpendicular to the c axis. For detector fabrication, samples are etched in a KI solution before the contacts are applied. The most commonly used contact materials are colloidal carbon (Aquadag) and Pd. T h e best energy resolution of HgI2 detectors has been reported to be approximately 0.93% FWHM for a 137Cs gamma -ray source.66 HgI2 has a high vapor pressure (5 x 105 torr) at room temperature; this will lead to a sublimation in air at a rate of 3 x 105 g/cm2sec. 67 Therefore the material must be encapsulated with parylene68 or a similar material to reduce the deterioration of the crystals over time. So far the development of

PAGE 35

35 detectors from HgI2 largely remains at the research st age, mainly due to the problems associated with low hole mobilities, short charge carrier mean free path, space charge polarization, and surface degradation. Despite these operation difficulties, HgI2 is still a promising material for gamma-ray detectors because of the high detection efficiency, high density and high resistivity. PbI2 remains interesting as an alternative semiconductor material for gamma-ray spectroscopy.69 71 The material crystallizes in a layered, hexagonal close packed (HCP) structure (space group 1 3 m P No. 164) at room temperature.7 2 The reported charge carrier mobilities are 8 cm2/Vs for electrons and 2 cm2/Vs for holes.73 The low mobilities result in low energy resolution and detection effi ciency, which prevent it from becoming a commercially viable semiconductor spectrometer. The relatively large band gap (2.55 eV) implies a high electrical resistivity (up to 1013 -cm)74 and th e device fabricated from this material can operate with low leakage current even at elevated temperatures. PbI2 crystals have been grown by vertical Bridgman,75 traveli ng molten zone (TMZ),76 and physical vapor deposition techniques.77 Similar to HgI2, single crystals of PbI2 can be cut from the bul k crystals along their cleavage planes (001). Crystal surfaces are polished and etched with KI solution before applying the contacts. The commonly used contact materials are Au, Pd, and Aquadag. Detectors fabricated from PbI2 was reported to have an energy resolution of 3% FWHM (1.83 keV) for 59.5 keV gammarays.78 InP has been investigated as a possible gamma -ray spectrometer because its structural and electronic properties are similar to Si and GaAs, yet it has a higher photoelectric absorption coefficient. The crystal structure of InP is identical to GaAs.

PAGE 36

36 The theoretical charge carrier mobilities are 4600 cm2/Vs for electrons and 150 cm2/Vs for holes.79 Single crystals of InP have been successfully grown by various techniques, including l iquid e ncapsulated Czochralski (LEC) 80 horizontal and vertical Bridgman,81 82 vertical gradient freeze (VGF),83 and synthesis solute diffusion (SSD) method.84 However, the quality and size of grow n InP crystals are still far below the requirement for a good gamma-ray spectrometer. Charge trapping remains a persistent problem for this material. Development work still continues in the hope of producing better quality crystals. TlBr crystals have dr awn great interest as an alternative material for room temperature gamma-ray spectrometers due to their high atomic number, high density, and large band gap.85 86 These favorable properties potentially yield a gamm a -ray detector with compact size and high detection efficiency. The crystal structure of TlBr is simple cubic CsCl -type (space group m Pm 3 No. 221). Since all thallium compounds are toxic, this material requires extremely precautions d uring handling and use. TlBr has poor charge carrier mobilities of 30 cm2/Vs for electrons and 4 cm2/Vs for holes,28 which limit it from being commercialized. Single crystals of TlBr have mainly been grown by eith er the vertical Bridgman87 or Bridgman-Stockbarger technique.8 8 89 Similar to other currently researched materials, poor crystal quality and charge trapping issues require further study of this material before commercialization. BiI3, which is the main focus of this work, has been studied as a room temperature gamm a -ray detector since the mid-1990s.10 13 The development of this material is rather slow, mainly due to the difficulties of producing large and high quality crystals. With recent progress on material purification, BiI3 single crystal platelets have been

PAGE 37

37 grown by physical vapor deposition (PVD). They have shown response to X -rays, yet the crystal was small in volume therefore no spectra was collected.9,13 The theoretical ch arge carrier mobilities are 680 cm2/Vs for electrons and 20 cm2/Vs for holes.90 Similar to -HgI2, BiI3 has a layered structure with cleavage planes perpendicular to [001] direction. With similar high atomic number, high density, and large band gap to HgI2, BiI3 should have comparable radiation performance to HgI2 if detector grade material can be produced. Other semiconductor materials have been studied with limited success, including CdSe,91 GaSe,92 cubic boron phosphide (BP),93 AlSb,94 CdS,95 ZnSe,96 and SiC.97 Most of these materia ls are limited to the high concentrations of impurities and traps, and they are generally not commercially available. 2.3 Physical Properties of BiI3 BiI3 crystallizes in a rhombohedral structure (space group 3 R No. 148) with six form ula units per unit cell. The lattice parameters of BiI3 are a = 7.519 and c = 20.720 .10 Within the lattice, the Bi3+ ions occupy the 6c Wyckoff positions with coordinates (0, 0, 0.1693), while the I ions reside in the 18f positions with coordinates (0.3322, 0.3146, 0.0797) .98 A schematic representation of the BiI3 crystal structure is presented in Figure. 2 -4. It can be seen from Figure 2 4 a that BiI3 adopts a layered structure wit h Bi3+ ions establishing a six -fold coordination with Iions Also, the Iions establish a non-linear two-fold coordination with the I Bi -I angle close to 90o. Each unit cell consists of three I -Bi -I layers stacked along [00l] direction, and within eac h I -Bi -I layer three close-packed atomic sheets stack in the sequence of I -Bi -I. The sites in I sheets are fully occupied, while only two thirds of the sites in the Bi sheets are occupied as shown

PAGE 38

38 in Figure 2 -4 b. There are three different ways of Bi sit es occupation in the close packed array. In the normal stacking, the three different layers are successively stacked along the c axis in the sequence of A B C where and denote the three possible positions of the close -packed I -sheet, and A, B and C the three Bi -sheets. Stacking of the layers is rarely perfect, therefore stacking faults are commonly found in BiI3 crystal. Furthermore, b ased on the electronic structure of BiI3, the Bi -I bonds are highly ionic with the 3 pelectrons of bismuth lost to iodine. The I Bi -I layers are held together by weak van der Waals forces Thus BiI3 crystal can be easily cleaved along [00l] direction. On the other hand, the weak van der Waals bonding mak es this material soft and difficult to handle The anisotropic crystallography of this material indicates the anisotropy of various physical properties such as electrical and thermal conductivity. This special crystal structure also determines a preferred growth along the (00l) planes during crystal growth. (a) (b) Figure 2 -4 Hexagonal representation of BiI3 crystal structure, a) unit cell consists of three I -Bi -I layers; b) projection of one I -Bi I layer indicating missing Bi sites. Smaller Bi3+ ions appearing in blue, while the larger I ions are presented in red. The dashed box represents one unit cell.

PAGE 39

39 The layered structure of BiI3 leads to a quasi two d imensional characteristic of excitons in BiI3, which makes it an attractive material for nonlinear optics.99 101 The behaviors of excitons in BiI3 have been extensively studied since late 1960s.102113 An exciton is a quasiparticle consisting of an electron and a virtual particle called an electron hole with a bound state in insulators and semiconductors An exciton is form ed when a photon enters a semiconductor, exciting an electron from the valence band into the conduction band, leaving a localized hole of opposite electric charge behind, to which the electron is attracted by the Coulomb force. Exciton states in BiI3 are observed near the fundamental absorption edge and c haracterized by a small Bohr radius (~ 6 )110 of about the unit cell size. Excitonic transitions have been described by a model on the basis of the excitation from s to p electronic states in the cation Bi.105 T h e model has been confirmed by the studies of the magne to optical spectra under strong magnetic field.114 The exciton motions in BiI3 have been well studied near the band edge transitions by optical techniques. As shown in Figure 25 ,110 t he optical absorption edge of Bi I3 starts as an indirect exciton transition assis ted by three different momentum conserving phonons, named A, B, and C. The indirect exciton transition (Eigx = 2.0081 eV at 2 K, gx = exciton ground state) occurs at slightly lower energy to the direct exciton energy (E d gx = 2.072 eV) and subsequently exh ibits an extraordinarily large oscillator strength. In addition, a series of sharp sample dependent absorption lines can be observed below the indirect exciton lines at liquid helium temperature. They are labeled as P, Q, R, S, and T from the higher ene rgy side The P line appears very close to the indirect exciton energy and it does not always appear together with Q, R, S, and T lines, which

PAGE 40

40 suggests that the origin of the P line is different from the others. It has been suggested by Kaifu et al that the P line is associated with the polytypic stacking disorder .107 The Q, R, S, and T lines were initially ascribed to the formation of a bielectron (or bihole) by Gross and his co workers.115 However, since the appearance and the intensities of these lin es vary with the respective samples, they should be attributed to some defects in origin; also various difficulties in the bielectron model have been pointed out.116 119 The appearance of R, S, and T lines was late r attributed to excitonic transitions perturbed by a stacking fault that occurred between unit layers.114 The stacking fault distorts the unit cell symmetry and creates new exciton states near the stacking fault plane. On the stacking fault plane, a new translational symmetry is constructed, which allows a quasi two -dimensional motion for the excitons. These excitons bind at the stacking fault and are called the stacking fault exc itons (SFE). The model based on the stacking fault has been confirmed by the magnetooptical studies, resonant Raman scattering, and space-resolved optical measurements.109,114,120 The absorption coefficients of t he R, S, and T lines vary from sample to sample, but the intensity ratios among them are nearly constant and are about 4:2:1, which has been explained as reflecting the degeneracy of the SFE levels.114 Exciton propagation along the stacking fault plane has very high efficiency and coherent characters since they have small radius and are less scattered by phonons. There is no significant dissociation reported for BiI3 at temperature below 600oC. However, a study by Cubicciotti and Keneshea suggested that BiI3 tends toward dissociation into Bi and I2 at 250 300oC,5 although only by a few percent. This

PAGE 41

41 tendency will lead to a distortion of stoichiometry and result in the deteriorate properti es of the material.13 The other properties of BiI3 will be discussed in th e following chapters. Chapter 3 will cover the chemical and mechanical properties including stability, hardness, and resistance to etchant. Chapter 4 will discuss the optical properties and band gap of BiI3. Chapter 5 will discuss the electrical propert ies of BiI3. Figure 2 5 The absorption spectrum near the indirect edge and its stacking disorder effect.110 2.4 Thermodynamics of Crystal Growth from Melt This section discusses the general thermodynamics of crystal growth. The whole discussion is based on chapter 7 in Solidification by Dantzig and Rappaz .121 An in depth discussion on the vertical Bridgman crystal growth will be given in chapter 3 in this dissertation.

PAGE 42

42 2.4.1 Nucleation The crystal growth process begins with the creation of small cluster s of atoms with crystalline structure when the temperature is below the material s melting point. Most of the clusters will repeatedly come together and fall apart. Some clusters may by chance grow larger than a critical size. Once the cluster exceeds this critical siz e, it can continue to grow and eventually form large crystals. The process to create a sufficiently large and stable cluster is called nucleation. T h ere are two classes of nucleation processes: homogeneous and heterogeneous nucleation. Homogeneous nucleation involves the spontaneous formation and subsequent growth of small particles of the new phase. Heterogeneous nucleation begins on a foreign material such as a particle or a surface. Heterogeneous nucleation has a smaller energy barrier to overcome, therefore most nucleation processes in the real world are heterogeneous. The driving force for the nucleation is the Gibbs free energy difference of the two phases, which is caused by the undercooling of the melt. 2.4.1.1 Homogeneous Nucleation Consider a simple case of formation of solid nuclei in the melt of a liquid phase, the change in Gibbs free energy upon the formation of a cluster of atoms can be described by two contributions, one from the decreased free energy associated with the formation of the new phase, and the other from the surface tension of the small cluster. Assuming that the cluster has a sphere shape with radius r, the change in Gibbs free energy can be written as r G r Gv o 2 3 hom4 3 4 (2 -4 )

PAGE 43

43 where Gv is the difference in free energy per unit volume between the liquid and solid phase, and is the specific surface free energy. The critical size r* of the cluster can be determined by 0 8 42 hom r G r r Gv o (2 -5 ) which gives the critical radius vG r 2 (2 -6 ) A cluster that is below the critical size is called an embryo. If it reaches critical size, it becomes a nucleus The Gibbs free energy to form a nucleus is then given by 2 3 hom3 16v oG G (2 -7 ) Equation ( 2 7 ) gives the height of the energy barrier which should be ove rcome for homogeneous nucleation to occur. For a liquid that is supercooled below its melting point, we can write m vT T L G (2 -8 ) where T is the undercooling, L is the latent heat, and Tm i s the melting point. Substituting equation (2 8 ) in equation (2 7 ) yields 2 22 3 hom3 16 T L T Gm o (2 -9 ) Equation ( 2 -9 ) indicates that the energy barrier to formation of a critical nucleus decreases rapidly with increasing undercooling T. A large undercooling is therefore essential for the homogeneous nucleation. 2.4.1.2 Heterogeneous Nucleation As previously stated, most nucleation processes are not homogeneous. Nucleation will prefer to occur on impurity particles, ions or foreign surfaces first due to

PAGE 44

44 the smaller energy barrier. Considering the case of nucleation on a foreign surface as shown in Figure 2 6, the embryo is assumed to take the shape of a spherical cap. The wetting angle is determined by the balance of the three s urface energies between the solid, liquid, and foreign surface. Figure 2 6 The nucleation of a spherical solid cap at a liquid substrate interface. The equilibrium condition can be expressed by the well known Laplace Youn g equation sl fs flcos (2 -10 ) where the two subscripts on the surface energies indicate the two substances in contact, namely s for solid and l for liquid, and f represents the foreign substrat e. The change of the Gibbs free energy upon formation of the embryo is given by v hetG V G (2 -11 ) where V and are the volume and surface energy of the embryo, respectively. They are given by 4 ) cos 2 ( ) cos 1( 3 42 3 r V (2 -12 ) ) ( sin ) cos 1 ( 22 2 2 fl fs sl r r (2 -13 ) by solving equation (2 10 ) through ( 2 -13 ), we get Foreign substrate fl r sl fs

PAGE 45

45 ) ( ) ( ) 4 3 4 (hom 2 3 f G f r G r Go sl v het (2 -14 ) 4 ) cos 2 ( ) cos 1 ( 3 / 4 ) (2 3 r V f (2 -15 ) where f( ) is a geometric factor given by the ratio of the volume of the spherical cap and a full sphere of identical radius. Equation ( 2 -14 ) reveals that the Gibbs free energy change for heterogeneous nucleation has the same form as for homogeneous one, but is multiplied by f( ). Following the same procedure as before, the critical radius gives exactly the same result as for the homogeneous nucleation T L T G rm v 2 2 (2 -16 ) The energy barrier for heterogeneous nucleation is then determined to be ) ( ) ( 3 16 ) ( 3 16hom 2 2 2 3 2 3 f G f T L T f G Go m v het (2 -17 ) T h e nature of the heterogeneous nucleation is determined by the function f( ). When = f( ) = 1, which corresponds to a complete nonwetting of the solid on the substrate. This case is equivalent to homogeneous nucleation. When = /2, f( ) = 1/2, and In this case, the cap forms a hemisphere. Although there is no surface energy reduction when the solid embryo forms on the substrate, the presence of the substrate reduces the nucleation energy barrier required to form a critical nucleus by half. When = 0, f( ) = 0, corresponding to a perfect wetting of the solid phase on t he substrate. An example is when the solid itself is the substrate, there is clearly no nucleation barrier and solidification is limited only by growth. This case is equivalent to crystal growth on a seed. In summary, whenever easily wetted foreign part icles are present in the melt, they induce heterogeneous nucleation under normal cooling conditions. The required undercooling for heterogeneous nucleation is much smaller than that for homogeneous nucleation.

PAGE 46

46 2.4.2 Nucleation Rate A nucleus with critical size r* will grow if it manages to add one more atom. The rate at which this occurs is proportional to the atomic vibration frequency 0 and the probability of capturing an atom at the surface pc.121 Thus, the nucleation rate is given by )) ( 3 16 exp( ) exp(2 2 2 3 0 0 0 0 0 f kT T L T n p kT G n p n p Im c c i c (2 -18 ) where ni* is the number of critical nuclei per unit volume, G* is the energy barrier to nucleation, n0 is the density of atoms in the liquid for homogeneous nucleation, or density of the heterogeneous nucleation sites in the liquid for heterogeneous nucleation. For homogeneous nucleation, f( ) = 1. The nucleation rate strongly depends on temperature through two competing ph enomena as denominated in the exponential in equation (2 18) A large undercooling T tends to increase the nucleation rate, while the low temperature decreases the mobility of atoms thus decreases the nucleation rate. The competition between these two phenomena results in the appearance of the maximum nucleation rate at a compromised undercooling. 2.4.3 The Crystal Growth Face Once the nucleation is complete, it is immediately followed by growth stage in which at oms in the melt attach themselves to the solid-liquid interface. There are two different types of interface structures: atomically rough or smooth interface. In the atomically rough interface, the transition from liquid to solid takes place over a number of atomic layers. With increasing distance into the interface, the thermodynamic properties of the atoms within the layers vary continuously from those of the liquid to

PAGE 47

47 those of the solid. In the atomically smooth interface, the transition from liquid t o solid ideally takes place across a single atomic layer. The earliest theoretical predictions on the interface structures were made by Burton et. al. ,122 who demonstrated that the interface would be rough at the growth temperature for most metal crystals. Jackson pointed out that the equilibrium interface structure is in equilibrium with both the crystal and the melt.123 He assumes that atoms are randomly placed on a surface that was initially atomically smooth, and calculates the change in surface free energy as follows: where R is the gas constant, Tm is the melting temperature, x is the occupation fraction of the surface sites, and is called the Jackson alpha factor, which is defined as: where L is the heat of fusion, k is the Boltzmann constant, u and w are the nearest neighbors of an atom on the surface and inside the solid respectively. The fraction u/w gives a measure of the binding force. In particular, if a cut is made through a crystal leaving a surface described by u and w, the free energy of the surfac e is proportional to (1 -u/w).124 Equation ( 2 -19 ) gives a relationship between the change of free energy G, the fraction of sites filled x, and the Jackson alpha factor. Figure 2 -7 shows the results of Jackson s calculation. For materials with < 2, there is a broad minimum at x = 0.5, indicating that the interface energetically prefers to be atomically r ough. On the other hand, in a material with > 2, there are two minima at x 0 and x 1. These minima approach the limiting x values (0 and 1) and become sharper as increases. The two )] 1 ln( ) 1 ( ln ) 1 ( [ x x x x x x RT Gm (2 -19 ) w u kT L m (2 -20 )

PAGE 48

48 minima indicate that either very few sites are filled on a compl ete surface or almost all sites are filled. Both conditions describe a smooth interface. Figure 2 7 Variation of free energy of interface with fraction of sites filled. 125 T he factors are different for different materials. Metal crystals generally have < 2; semiconductor crystals have 2 < < 3; and oxides, silicates, and polymer crystals, which have complicated crystal structures, have > 3. 126 Due to the presence of u/w, the factors will also be different in different crystallographic directions in the same crystal. 2.5 BiI3 as a Radiation Detector BiI3 may allow one to efficient ly detect and discriminate medium and high energy gamma -rays. The high effective atomic number yields a detector that has a relatively high photoelectric cross -section for moderate and high energy gamma-rays. Theoretical modeling of intrinsic photopeak e fficiency has been performed to compare BiI3 to HgI2, PbI2 and CZT by Azaree et al .127 The modeling was conducted by using the Los Alamos developed Monte Carlo N -Particle (MCNP) code. MCNP generates

PAGE 49

4 9 results based on the probability of physical interactions occurring. The probability of interaction occurrence is determined through the use of data libraries that are built into the software p ackage. The geometry for the simulations with each of the materials was kept the same. The data was generated by modeling a one cubic centimeter detector located a distance of one centimeter from a square centimeter surface source. Ten million photons w ere started, which negated the need for any variance reduction to be performed. The number of photons that yielded a total energy deposition w as obtained for energies ranging from the zero to 3 MeV at intervals of 5 keV. It is evidenced in Figure 28 tha t the peak efficiency for BiI3 is comparable to HgI2 and PbI2, and a factor of two higher than CdZnTe. For example, the intrinsic photopeak efficiency for BiI3 is approximately 20%, while CdZnTe is 10% for a 662 keV gamma -ray. The wide bandgap of BiI3 wi ll allow for room temperature operation and harsher temperature environments may be possible. 0.0 0.5 1.0 1.5 2.0 2.5 3.0 1 10 100 Intrinsic Photopeak EfficiencyPhoton Energy (MeV) HgI2 PbI2 BiI3 CdZnTe Figure 2 8 Intrinsic photopeak efficiency as a function of energy for 1 cm 3 detectors located 1 cm from a surface source. Re plott ed after Lintereur et al. .127 Although BiI3 has been studied for over a century, the attempt to use it as a radiation detector only started in 1990s. Nason and Kellor were the first researchers to

PAGE 50

50 test the radiation response of BiI3 single crystal.10 The single crystals were grown by the PVT method with a seed obtained via the sublimation of commercial BiI3 powder. The largest single crystal they obtained has a dimension of 1.2 x 1.2 x 0.4 cm3. The detectors were fabricated by applying colloidal graphite electrodes and connected with Pd wire. The resistivity was reported to be (2 5) x 109 cm However, they were unable to collect any signal from 60 keV gamma-ray or 5 MeV particle irradiation, mainly due to the poor crystal quality.10 Following their work, Dmitriev et al. grew BiI3 single crystal by the vertical Bridgman method direc tly from commercially available powder (99.999% purity).11 The detector s were fabricated by applying vacuum evaporated Pd electrode and attached with Pd wire using silver epoxy. The resistivities of the crystals were measured to be (0.5 2 ) x 109 cm By applying AC coupling on the detector, they successfully observed the detector response to pulse X -ray excitation. However, spectrum relationship could not be established due to polarization effects, which were related to the defect structure in the crystal. In 2002, Matsumoto et al fabricated a better detector from BiI3 crystals grown by the vertical Bridgman method.12 Electrodes with 1 mm in diameter were deposited by vacuum evaporation of Pd on the surface, and Pd wires were bonded to the Pd electrodes using silver epoxy. The resistivity of a 82 m detector was measured at 1.7 x 1010 cm Alth ough this detector exhibited a clear peak response to a 241Am particle source (5.48 MeV), the energy resolution was rather poor with 40% FWHM (2.2 MeV). The detector was not able to obtain any spectra from X -ray and gamma -ray sources.

PAGE 51

51 Different from t he approach using bulk crystals, Fornaro et al focused their work on BiI3 platelets13 and thin film128 with the application towards X -ray detection. The platelets were grown by the PVT method with an average dimension of (100 200) mm2 x (50 80) m. Both Pd and Au were tested as electrodes for the detector. The Au electrode gave the higher resistivity val ues about 0.6 x 1012 cm higher than that of detector with Pd electrode. The resistivit ies, as high as 2 x 1012 cm were obtained and were 2 orders higher than that of the bulk crystal However, the detectors only showed response to a 241Am particle source and no X -ray spectra were collected. BiI3 polycrystalline film was grown by the PVT method on a 1 x 1 glass substrate. The glass substrate was initially deposited with Pd electrode as rear contact by thermal evaporation. After film growth, Pd electrode with area of 4 mm2 was deposited as the front contact. The measured resistivity of the polycrystalline film was relatively high with values up to 6 x 1012 cm Detectors fabricated from the films showed significant response to a 241Am source, and the current density was observed to increase linearly with increasing exposure time under an X ray beam of several kilovoltages. Previous efforts of pushing Bi I3 as a room temperature radiation detector were very encouraging. Similar to the development of CZT and HgI2 detectors, the success of the detector is largely determined by the crystal quality.3,129 Therefore, th is study investigated the growth conditions of BiI3 single crystal using a modified vertical Bridgman method. The dependence of crystal quality on the growth conditions has been investigated using various characterization techniques, which will be discuss ed in detail in chapter 3. The preliminary results will provide a guideline for a future crystal grower to have a better control on the qualities of BiI3 single crystals.

PAGE 52

52 CHAPTER 3 BISMUTH IODIDE SINGL E CRYSTAL GROWTH AND CHARACTERIZATION 3.1 Introducti on As discussed in Chapter 2, the crystal growth techniques for BiI3 can be summarized as vertical Bridgman and PVT. While the PVT method can produce high quality but small volume crystals, the vertical Bridgman technique can grow crystals with larger vol ume but the crystal quality is not as good as those grown from PVT.11 13,130 This chapter discusses the growth of BiI3 single crystals using the modified vertical Bridgman (MVB) technique. The quality of the cryst als grown under different conditions was evaluated and compared to that of crystals grown by PVT method. 3.2 Crystal Growth Techniques of BiI3 3.2.1 Vertical Bridgman Technique As the name says, this technique is named after P. W. Bridgman who is the fir st one using this method to grow a series of metal single crystals.131 The principle of this technique involves lowering a crucible through a furnace so that the freezing starts at the lowest point in the crucible and the solidification interface slowly moves up along the crucible. The movement r ates for such processes range from around 0.1 to 200 mm/h.124 A typical vertical Bridgman system is shown in Figure 31 The movement of the crucible is controlled by a dropping motor. A longitudinal temperature profile is established at the center of the furnace with specific temperature gradient near the melting point of the material. The hole in the lid should be small and the lid should fit well with the furnace body to prevent thermal disturbance.

PAGE 53

53 Longitudinal PositionTemperature (oC) Melting point (a) (b) Figure 3 1 A typical Bridgman system. (a) Schematic diagram of the furnace; (b) longitudinal tempera ture profile at the center of the furnace. Diagram after Brice.124 The crystal growth with vertical Bridgman method can use a seed or no seed. A non-seeded Bridgman method relies o n random nucleation to produce a single crystal which then propagates to form one large crystal through the whole frozen charge. The formation of just one nucleus requires a small nucleation volume. Thus the crucibles are normally shaped with sharp tips to meet the nucleation criteria, as illustrated in Figure 3 -1 (a). For a seeded Bridgman growth, the crucible shape is not that important. Normally a flattened bottom is made to accommodate the seed material. The most obvious requirement for a crucible i s that it should not react with the crystals within the temperature range of crystal growth. It also should have a smaller coefficient of thermal expansion than the crystal to avoid the induced strain and stress during cooling. Lid Melt Crystal Dropping motor

PAGE 54

54 For vertical Bridgman crys tal growth, the quality of the single crystal is largely determined by the control of heat flow during crystallization. T h e thermal conditions of a practical crystal growth system are rather complicated, thus analytical solutions are not possible without simplifications and assumptions.132 In the following discussion, it is assumed that the melt is a single component material, the density of the melt and its crystal are the same, temperature gradients within the crystal transverse to the growth direction are negligible, and the solid-liquid interface is planar. During crystal growth, a heat balance is set up by the conservation of heat flux at the interface, which can be written as133 L L s s sG K G K LR (3 -21 ) where Ks and KL are the ther mal conductivity of solid and liquid (cal/(cm oC s)), Gs and GL are the temperature gradient in solid and liquid at the solidliquid interface (oC/cm), s is the density of solid (g/cm3), L is the latent heat of melt (cal/g), and R is the growth rate (cm/s). The left side in Equation ( 3 1 ) indicates the heat generated at the interface due to solidification. The first term on the right side is the heat flux going out of the solid-liquid interface. The second term on the right side refers to the heat flux coming into the solid-liquid interface. Equation (3 -1 ) indicates th at the growth rate R can be controlled by independently adjusting Gs and GL. This is important as growing good single crystals requires that the temperature gradient be high and growth rate be low. The temperature gradient in liquid (GL) is directly prop ortional to the heat flux in the liquid at the solid -liquid interface. The temperature gradient in solid (Gs) is normally determined empirically or through heat flow calculations. In a cylindrical system, the heat conduction in solid can be described as124

PAGE 55

55 z T R t T c z T K zs s s (3 -22 ) where cs is the specific heat of solid (cal/(goC)), and z is the coordinate parallel to the growth direction. Physically, the term on the left represents the net rate of heat conduction away from the solid per unit volume. The first term on the right side represents the decrease rate of internal energy per unit volume. The second term on the right si de is the rate of energy released per unit volume due to interface moving. Some simplifications need to be made to solve equation (3 2 ). For example, in the case of floating zone (crucibleless) growth, assuming t he crystal has a cylindrical shape, temperature gradients along the transverse direction are negligible, and heat transfer from the crystal to surrounding is by convection. At steady state, the temperature gradient in solid at the solid-liquid interface i s given by133 s s s m s R rK h R T T G 2 2 ) 2 ( ) (2 0 (3 -23 ) where T0 is the ambient temperature (oC), Tm is the melting point of the crystal, s is the thermal diffusivity of the solid crystal (Ks/ scs, cm2/s), h is the heat transfer coefficient from solid to ambient (cal/(cm2oC s)), and r is the radius of crystal (cm). Since the growth rate R is usually small, when R/2s << 1, equation (3 3 ) becomes 2 / 1 0) 2 )( (s m srK h T T G (3 -24 ) For crystals with high melting point, (Tm T0) is large and the heat transfer coefficient h is increased by radiation heat t ransfer. The attainable Gs can be as high as 100oC/cm or more. For low melting point crystals, external cooling has to be applied to achieve steep gradients. It should be noted that the analytical solutions cannot completely explain and predict the cry stal growth in a practical vertical Bridgman system. This is mainly

PAGE 56

56 because the simplifying assumptions usually tend to discard some important details which control crystal growth.134 135 Accordingly, numerical st udies have been performed to gain more understandings in the vertical Bridgman growth process.82,132,136 142 Studies have been focused on controlling the solidmelt interface shape,82,137 models of the heat and mass transfer,138139,142 and the influence of the thermal environment and process parameters.140141 Accurate material thermal p ropert ies and operating environment are required for successful numerical predictions. 3.2.2 Physical Vapor Transportation Technique The crystal growth by the PVT method is typically conducted in closed or semi closed systems. If the vapor pressure of a material at a desired growth temperature exceeds 102 torr, it may be efficiently grown by the PVT method.143 In the PVT process, an ampoule containing a polycrystalline source of the desired material is heated to a temperature at which the material starts to sublime at a rate leading to crystal growth. The ampoule is typically placed in a furnace that provides a temperature gradient over the length of the ampoule, so that the polycrystalline source material sublimes at the higher temperature and deposits at the end of the ampoule where the temperature is lower. This temperature difference causes the different vapor pressure at two ends of the ampoule and forces the material transport from the source to the cooler end. T h e PVT system can adopt horizontal or vertical set up. A typical horizontal PVT system is illustrated in Figure 32 For crystal growth by PVT method, the growth temperature T2 should exceed about one third of the melting temperature of the crystal.124 The temperature T1 is fixed so that the crystal grows at an acceptable rate which can be deduced from simple theory as follow. The vapor pressure can be expressed in the form of

PAGE 57

57 ) / exp(0RT H P Ps (3 -25 ) then the rate of deposition in an ideal case with an evacuated ampoule is124 RT M RT H RT H P fs s/ ) 2 / )]( / exp( ) / [exp(2 / 1 2 2 2 1 1 0 (3 -26 ) Tx Figure 3 2 Schematic diagram of physical vapor transport (PVT) system. Diagram after Isshiki and Wang.144 where s is the latent heat associated with the vaporization of a solid below its melting point (cal/g), is the density of the crystal (g/cm3), M is the molecular weight of the vapor (g/mole), and R is the gas constant (cal/(K mole)). Equation (3 -6 ) indicates that the growth rate of PVT method can be controlled by adjusting the temperature difference at two ends of the ampoule. Normally inert gases (N2, Ar) are infused into the ampoule after vacuum treatment to assis t the nucleation of PVT crystal growth.9,129,145 Since PVT growth is carried out at temperature much lower than the melting point of the material, it has multiple benefits of:126 (a) reduced diffusion of impurities from containers, (b) reduced defects such as voids and dislocation densities, and (c) effective separation of impurities with low vapor pressure during transport. Th e major drawbacks for PVT methods are small crystal volume and uncontrolled Source materials T1 T 2 Furnace Single crystals

PAGE 58

58 nucleation. In practice, unless there is a large temperature gradient near the cooler end of the ampoule, multiple nucleations occur and the products at the cooler end are needles platelets, or polycrystalline formed by many small crystals. 3.3 Experimental Procedures 3.3.1 Modified Vertical Bridgman Growth In this study, BiI3 single crystals were mainly grown by a modified vertical Bridgman method. Pyrex glass ampoules were sele cted as the crystal growth containers since the annealing and softening points of the glass are 560oC and 815oC respectively, which are much higher than the maximum temperature (440oC) used in the growth of BiI3 single crystal. The Pyrex glass tube was so aked in Decon Contrad 70 for 12 hours to effectively remove the organics rinsed 10 times with high purity DI water -cm) and baked at 120oC for 12 hours. One end of the cleaned glass tube was melted and shaped to a sharp tip using a propane h and torch. BiI3 powder (99.999%, MV Laboratories) was weighed and loaded in this tipped Pyrex glass ampoule. The ampoule needs to be evacuated and sealed to get rid of the water and oxygen as BiI3 will react with them at elevated temperature. To do that the ampoule was first vacuumed to a pressure of approximately 4 x 105 mbar. Heat was then gently applied on the ampoule wall using a propane hand torch to help evaporate the water. This step has to be done very carefully to avoid heating up the BiI3 p owder. Ar gas was then back filled in the ampoule to ~700 mbar. The purpose of infusing Ar gas is to make sure any residual gas after vacuum is inert. The cycle of pumping down, heating ampoule wall, and backfilling Ar gas was repeated three times, the final evacuation took around another 2-3 hours to take the pressure down to 106 mbar. The ampoule was then sealed using a propane hand torch. To help seal the ampoule, a neck area was

PAGE 59

59 preformed in air before the ampoule was pumped down. As the soft gla ss will collapse inwards due to the inner vacuum, the neck area will help the glass collapse into the opposing wall without breaking the seal. After sealing, t he ampoule was vertically fixed in a standing frame and placed in a programmable 24 heating zone furnace (EDG 13, Mellen Company) The length of each heating zone is one inch and the temperature of each heating zone can be independently controlled by a computer program, which allows a precisely controlled temperature profile along the vertical axis of the furnace. The powder was first heated and kept at 440oC for 24 hours then cooled to 420oC at a rate of 1 0oC/h. The temperature profile was programmed to move up along the vertical axis with a fixed rate. T his simulate d the downwards movement of th e ampoule in the conventional vertical Bridgman method. The set up of this modified Bridgman system provides two advantages over the conventional one. First, possible vibration during crystal growth is eliminated since the ampoule is kept static in the n ew system. Second, a hole is no longer needed at the top of the furnace to provide the pass way for the hanging wire. Thus, less thermal disturbance is expected. As a result, the modified Bridgman system can provide a more stable crystal growth environm ent leading to better crystal qualities. As discussed previously, temperature gradient and growth rate are two critical parameters for crystal growth by vertical Bridgman technique. The best way to start the crystal growth is to assess the relationship between the growth rate and temperature gradient at solidliquid interface established in equation ( 3 1 ). However, it is difficult to do that in the case of BiI3 due to the unavailability of its thermal conductivity values. This is probably because BiI3 will dissociate to Bi and I2 at 250oC 300oC,5 and I2 is

PAGE 60

60 very corrosive with extremely high reactivity to metals, which could destroy the instrument for the thermal conductivity measurement. Therefore, in this study, differ ent growth rates and temperature gradients for growing crystals with 0.75 inch diameter were investigated on an educated initial guess basis. Previous studies reported that BiI3 single crystals have been grown at the conditions of 1 mm/h growth rate with 15oC/cm temperature gradient11 and 1 mm/h growth rate with 5.7oC/cm temp erature gradient .12 As equation (3 4 ) in dicates that the temperature gradient depends on the thermal properties of the furnace, BiI3 single crystal grown in different furnaces could have very different temperature gradient. The highest temperature gradient allowed in the EDG 24 zone furnace is around 18oC/cm. The temperature variation in each zone is controlled within 0.5oC, which corresponds to a gradient variation of 0.2oC/cm. The investigated temperature gradients were then selected as 5, 10, and 15oC/cm to make a distinct difference betw een them. For each temperature gradient the crystal growth has been conducted at a speed of 0.5, 1, and 2 mm/h. 3.3.2 Physical Vapor Transport Growth For the PVT crystal growth, the Pyrex glass ampoule was cleaned using the same procedure described for the MVB growth. A charge of 5 g BiI3 powder was loaded into a cleaned ampoule with diameter of 0.75 inches. The work charge was first vacuumed to around 4 x 105 mbar and cleaned with ultra high purity grade Ar gas three times, then finally sealed with ultra high purity grade Ar at 700 mbar. Following the crystal growth conditions used by Cu a et al. ,9 the ampoule end with BiI3 powder was placed at 310oC and the other end was placed at room temperature. T h e growth time was 7 days for 5 g of powder. After the growth, the furnace was c ooled down at a rate of 1oC/min.

PAGE 61

61 3.3.3 Characterization A polycrystalline ingot was obtained after the crystal growth by the MVB method Single crystal samples were cut from the ingot along the cleavage plane (001) with various cross sectional areas and t hicknesses. For crystal grown by the PVT method, single crystal platelets with various dimensions were carefully collected. Single crystal samples obtained from both methods were characterized by the following techniques. 3.3.3.1 X -ray Diffraction X -ray the crystallinity and phase purity of BiI3 single crystals. The power of the X -ray generator was set to 40 kV and 20 mA. The 2 angle was scanned from 10o to 60o with a step size o f 0.02o. 3.3.3.2 Vickers Hardness The Vickers hardness of the grown single crystals was measured by a micro hardness tester (Buehler MicroMet II). Each sample was measured for 5 indentations and the applying force was 50 g. 3.3.3.3 Etch Pit Density The di slocation density of the crystal surface was examined by EPD measurement using a KI solution as the etchant. To determine the ideal etchant concentration for EPD measurement, single crystal samples were etched in 5%, 10%, 15%, and 20% weight percent of KI solution respectively, and the etching time varied from 10 sec to 150 sec. After etching, the sample was rinsed with acetone and dried at 120oC for 10 min. The etched sample was weighed to determine the weight loss per unit surface area. The time depen dence of the weight loss was calculated for each etchant concentration. The ideal etchant concentration was determined at which the etchant

PAGE 62

62 can effectively remove the mass of BiI3 single crystal. After that, BiI3 single crystals grown by the MVB and PVT methods were etched for a specific time. The density of pits was examined by an optical microscope (Nikon Eclipse LV100). For each sample, five randomly selected areas were studied to count the EPD. 3.3.3.4 X -ray Rocking Curve The quality of the grown cr ystals was inspected by X -ray rocking curve on a Philips analytical MRD X'Pert System as shown in Figure 3 3 Figure 3 3 X Pert system for X ray rocking curve measurement. The presence of defects such as dislocations and small angle grain bound aries will cause local variation of lattice spacing or plane orientation. For X -ray diffraction in the vicinity of a defect, Bragg s law cannot be simultaneously satisfied by the perfect lattice and the distorted one, which causes the broadening of X -ray diffraction peak. By performing X-ray rocking curve analysis, the peak broadening of a specific lattice plane can be measured by scanning the incident angle at a fixed receiving angle 2 as illustrated in Figure 34 The full width at half maximum (FWHM) of this peak is generally used to evaluate the quality of a single crystal.146 148 The rocking curve scan was conducted at a rate of 0.002o/s with a step size of 0.001o. Each sample was X ray generation tube Sample stage Detector

PAGE 63

63 scanned twice by rotating the sample 90o ( angle in Figure 34 ) relative to the previous scan, as illustrated in Figure 3-5 Figure 3 4 Illustrati on of rocking curve -scan. Figure 3 5 Illustration of the sampling for rocking curve scan. The shadowed bands numbered 1 and 2 represent the X -ray incident area for each scan. 3.3.3.5 Scanning Electron Microscopy (SEM ) The surface microstructure of the as grown BiI3 single crystals was characterized using SEM (JEOL 6335F). The freshly cleaved surface was sputtered with carbon film with thickness of around 20 nm. The SEM was operated with an accelerating voltage of 15 kV, a probe current of 10 3.3.3.6 Atomic Force Microscopy (AFM) To study the surface morphology of the metal contacts, the sample surfaces were examined by AFM ( Asylum Research AFM ). The AFM tip from Veeco has a triangular X -ray band 1 2 Sample ~1 mm 2 Incident X ray Diffracted X ray Sample

PAGE 64

64 cantilever and a nominal tip radius of 20 nm. The normal stiffness of the tip cantilever is 0.58 N/m. The topography images were acquired in contact mode at a fixed applied load of several nano N ewton. 3.3.3.7 X -ray Photoelectron Spectroscopy (XPS) The stoichiometry of the grown BiI3 single crystal is determined by XPS spectrometer using an Omicron XPS system. The XPS spectra were collected from a freshly cleaved BiI3 crystal surface -ray (1486.7 eV). The s pect ra were recorded at a sampling area of 1.3 mm2 and a take off angle of 5 5o with pass energy of 22 eV (step size 0.05 eV, integration time 1 sec). The chamber pressure was kept below 5 x 1010 mbar d uring measurement. The instrument resolution was around 0.1 eV. The peak energies were calibrated using a Ag 3d5/2 peak at 368 eV. The composition of bismuth and iodine was estimated by comparing the photoelectron peak intensities of the Bi 4f with the I 3d core level peaks. 3.3.3.8 X -ray Fluorescence Spectroscopy (XRF) The stoichiometry of the grown single crystal was also determined by XRF using an Orbis Micro XRF s pectrometer (EDAX Inc.). The XRF spectrometer was operated using a Rh tube at 30 kV. T he spot size of the X -ray was 30 m. The composition of bismuth and iodine was estimated by comparing the X -ray intensities of the Bi M -line with the I Lline in the measured spectrum. 3.4 Results and Discussion 3.4.1 MVB Crystal Growth 3.4.1.1 Determinat ion of Temperature Profile The concept of vertical Bridgman growth is quite simple. The powder is first heated to the melting temperature and soaked for a period of time. Then the crystal

PAGE 65

65 growth is conducted under certain temperature gradient and moving rate. Finally the entire growth system is cooled back to room temperature While the temperature gradient and growth rate are vital for producing high quality single crystals, the heat up and cool down cycles are also important for a successful growth run. As the furnace thermocouples are located at the furnace wall while the growth ampoule is in the center, the temperature difference is presented between the zone temperature reading and the actual temperature of the ampoule. To determine this difference, the central temperature was recorded by a K -type thermocouple placed along the central axis of the furnace. It was observed that the central temperature was around 20oC higher than the temperature reading at the furnace wall. To ensure the complete mel ting of BiI3, the highest temperature for the soaking period was set as 440oC. This is based on the fact that ideally BiI3 will melt at 408oC and the actual soaking temperature will be around 460oC. The heating rate is not a major concern for BiI3 as the processing temperature is in the medium range and Pyrex ampoule has a very low thermal expansion coefficient (3.2 x 106/oC). The only concern is to make sure that the furnace temperature does not go up too fast or too slow, as both conditions require additional time to reach thermal equilibrium. Experimentally, the heating rate has been tested at 1, 3, and 5oC/min. It was found that at 3 and 5oC/min heating rates the furnace temperature could not follow the set temperature. The temperature lags were of around 10 and 30oC for the heating rate of 3 and 5oC/min respectively. When the heating rate was lowered to 1oC/min, the furnace temperature completely followed the set point with 1oC variations. Based on the test result, the heating rate was set at 1oC/min for all the crystal growth runs.

PAGE 66

66 After the furnace reached the maximum temperature the powder in the growth ampoule may still be lower than this point since heat transfer takes time. A soak period has to be set up to allow every component inside th e furnace reach the thermal equilibrium. A K type thermal couple was attached at the outer wall of the ampoule to test the equilibrium time. It was observed that the temperature was stabilized in the first 30 minutes after reaching the maximum temperatur e. However, the BiI3 melt inside the ampoule may need longer time to reach thermal equilibrium. A soaking time of 12 hours will be safe enough to thermally stabilize the whole growth system. After soaking, the furnace was cooled down to 410oC at a rate of 1oC/min. The corresponding central temperature was around 430oC. This is to make sure the overall furnace temperature is still above the melting point of BiI3. Starting from this point, the crystal began to grow at a specific growth rate and temperature gradient described in section 3.3.1. The temperature at which the crystal growth ends is another important parameter needed to be determined. If the crystal growth ends at a high temperature, lattice disorder may happen due to the fast cooling of the atoms with high kinetic energies. On the other hand, ending the crystal growth at a very low temperature is time wasting. It is noted that the crystal growth of BiI3 by PVT method happens at ~250oC,130 which implies that atoms at this temperature barely have enough kinetic energies to move around. Therefore, in this st udy, the crystal growth was set to end at 200oC, which corresponding to a furnace temperature of 180oC. The crystal after growth was cooled to room temperature at a rate of 1oC/min. A temperature profile was created based on the above discussion, as illustrated in Figure 36 The total growth

PAGE 67

67 time varies with the temperature gradient and growth rate. For example, the growth time for a crystal grown at 10oC/cm and 0.5 mm/h is around 20 days. 0 100 200 300 400 500 0 100 200 300 400 500 Temperature (oC)Time (hr) Figure 3 6 Growth recipe of the crystal growth at 10 o C/cm and 0.5 mm/h. 3.4.1.2 Crystal Growth Result For crystal growth in 0.75 inch diameter ampoules it was determined that only three different combinations of growth rate s and temperature gradient s produced large single crystals: 10oC/cm and 0.5 mm/h; 10oC/cm and 1 mm/h; and 15oC/cm and 0.5 mm/h ( Table 3 -1 ). The typical polycrystalline rods and cleavage planes of crystals obtained at different growth conditions are shown in Figure 3 7 Table 3 1 Growth c onditions for BiI3 s ingle c r ystals in 0.75 d iameter a mpoules. Temperature Gradient (oC/cm) 5 10 15 Growth Rate (mm/h) 0.5 P olycrystalline Single crystal Single crystal 1 P olycrystalline Single crystal P olycrystalline 2 P olycrystalline P olycrystalline P olycrystalline It i s noticed that single crystals were obtained at the same temperature gradient (10oC/cm) but different growth rates (0.5 and 1 mm/h) or the same growth rate (0.5 mm/h) but different temperature gradients (10 and 15oC/cm). It is therefore clear that equatio n (3 1 ) cannot explain these experimental results. However, one has to keep in

PAGE 68

68 mind that equation (3 -1 ) is established on the steady state condition. A real crystal growt h always happens at a non-steady state, although it may be very close to a steady state. T h erefore, the requirement for crystal growth becomes LR G K G Ks L L s s (3 -27 ) which suggests as long as the overall heat flux flowing away from the growth interface exceeds the rate of latent heat evolution, the growth of single crystal could happen.124 This explains why BiI3 single crystals were obtained at the three conditions listed in Table 3 1 (a) (b) (c) (d) (e) (f) Figure 3 7 Single crystals produced by MVB method, (a) as grown polycrystalline rod; (b) crystal grown at 10oC/cm, 0.5 mm/h; (c) crystal grown at 10o C/cm, 1 mm/h; (d) crystal grown at 15oC/cm, 0.5 mm/h; (e) crystal grown at 15oC/cm, 1 mm/h; (f) polycrystalline cross section obtained at rest conditions. Based on the above discussion, single crystal should also be grown at 1 5oC/cm and 1 mm/h given that single crystal was obtained at 1 0oC/cm and 1 mm/h Actually the

PAGE 69

69 solidification at this condition was characterized by parallel growth of multiple crystals, which had a non-smooth growth interface, as shown in Figure 37 e. A possible reason for this kind of interface morph ology was the thermal fluctuations at the growth interface, a phenomenon that has been commonly observed for other single crystal growths.149 151 The growth furnace was observed to have a temperature fluctuation of around 1oC at a temperature gradient of 1 5oC/cm compared to 0.5oC at a temperature gradient of 1 0oC/cm A higher growth rate would cause more fluctuations and non-smooth interface compared to a lower growth rate. Therefore, crystal grown at 1 5oC/cm a nd 1 mm/h was under a condition with high thermal fluctuations, which caused the formation of a non-smooth interface. By decreasing the temperature gradient to 10oC/cm or the growth rate to 0.5 mm/h, the thermal fluctuations were suppressed and smooth growth face was obtained. Based on the experimental results, the critical conditions to grow BiI3 single crystals with a diameter of 0.75 inch can be concluded as a temperature gradient between 10oC/cm and 15oC/cm, and a growth rate between 0.5 mm/h and 1 mm /h. A single crystal with a dimension of 18 x 13 x 5 mm3 has been successfully grown at the temperature gradient of 10oC/cm and growth rate of 0.5 mm/h, as shown in Figure 38 Figure 3 8 BiI 3 single crystal with dimension of 18 x 13 x 5 mm 3 the l argest crystal size ever reported.

PAGE 70

70 This is the largest crystal size ever reported for BiI3. The previously reported largest BiI3 single crystal was prepared by Nason and Keller with a dimension of 1 2 x 1 2 x 4 mm3.10 3.4.2 PVT Cry stal Growth The growth ampoule and a single crystal platelet obtained after PVT growth are shown in Figure 3 -9 Many tiny crystals, needles, and platelets were found to stack up together in the ampoule where the temperature was measured to be around 230oC which was determined to be the growth temperature of the single crystal platelets. The platelets were carefully separated from the polycrystalline mixture by using tweezers and a razor blade. The typical platelet has a surface area of 5 20 mm2 and a thickness of 100 200 m. Figure 3 9 BiI 3 single crystals grown by the PVT method. 3.4.3 Crystal Structure and Phase Purity X -ray diffraction was used to determine the crystal structure and phase purity of single crystals grown by the two techniques described before. All the diffraction patterns were almost identical and a typical XRD pattern is shown in Figure 310 Only 003, 006, 009, and 0012 reflections parallel to the (001) plane were observed, and the peak positions matched well with the publ ished BiI3 powder diffraction data for a rhombohedral crystal structure (JCPDS PDF# 48-1795 ).10

PAGE 71

71 The lattice parameters of the single crystal grown by the MVB method were determined by performing 2 s on several selected planes. The scan was conducted by rotating and tilting the single crystal to satisfy the Bragg s condition for a specific plane that is not parallel to the crystal surface. The planes and their corresponding d in Table 32 The lattice parameters were calculated by least squares fit with values of a = 7.5160.002 and c = 20.6970.003 which match ed well with the published powder diffraction data of a = 7.519 and c = 20.721 .10 The XRD results verified that BiI3 single crystals have been successfully grown by the MVB and PVT methods. Table 3 2 3.4.4 Vickers Hardness Vicke rs hardness test is a method to determine the hardness of a material by measuring the size of an indentation produced under force by a pyramid shape diamond 10 20 30 40 50 60 0 5000 10000 15000 20000 25000 30000 35000 40000 (0012) (009) (006) JCPDS PDF# 48-1795Intensity (counts)2 (degrees) (003) Figure 3 10 A typical X ray diffraction pattern of BiI 3 single crystal s Planes (hkl) 003 104 107 116 018 12.790 21.920 33.258 35.205 37.371

PAGE 72

72 indenter.152 The angle between the two opposite faces of the pyramid is 136o. A fter removal of the load, t he two diagonals (d1, d2) of the indentation on the surface of the material are measured under a microscope and their average is calculated ( Figure 31 1 ). The area of the sloping surface of the indentation is calculated by 8544 1 ) 2 / 136 sin( 22 2d d Ao (3 -28 ) where d is the mean diagonal of the indentation with a unit of mm. The Vickers hardness (HV number) i s then determined by 28544 1 d F A F HV (3 -29 ) where F is the applied force with a unit of Kg. The Vickers hardness on the (001) surface of single crystals grown by the MVB method was measured and is shown in Table 3 -3 Single crystals grown by the PVT method were not subjected to the hardness test as the crystals are too thin. The average HV numbers of the single crystals grown at different conditions showed no significant difference. The results indicated that BiI3 single crystal is soft with a HV number between 12 and 15. As a comparison, the HV number of a HgI2 (001) surface is ~22,153 a 316L stainless steel is 140.154 Figure 3 11 Illustration of Vickers hardness test. d1 d2 Force 136 o

PAGE 73

73 Table 3 -3 Vickers hardness of (001) surface of BiI3 single crystals grown by the MVB method. 10 o C/cm, 0.5 mm/h 10 o C/cm, 1 mm/h 15 o C/cm, 0.5 mm/h Meas. d1 d2 HV # d1 d2 HV # d1 d2 HV # 1 0.088 0.089 12 0.082 0.078 14 0.077 0.079 15 2 0.073 0.088 14 0.093 0.085 12 0.083 0.080 14 3 0.082 0.089 13 0.088 0.085 12 0.086 0.079 14 4 0.091 0.088 12 0.082 0.080 14 0.075 0.080 15 5 0.086 0.087 12 0.082 0.087 13 0.084 0.080 14 Aver. 13 0.15 13 0.16 14 0.18 3.4.5 Etch Pit Density Measurement The EPD measurement is generally utilized to determine the quality of semiconductor single crystals.155 157 An etching solution is applied on the surface of a crystal and the etching rate increases at dislocation sites resulting in the formation of pits. Therefore, EPD is often used to evaluate the dislocation density of singl e crystals.158 In this study, the etchant was KI solution as it was routinely used to reveal dislocations in HgI2 129 and PbI2 2. The time dependence of the weight loss for different etchant co ncentrations is shown in Figure 3 12. 0 20 40 60 80 100 120 140 160 0 10 20 30 5% KI 10% KI 15% KI 20% KIWeight Loss (mg/cm2)Time (sec) Figure 3 12 BiI 3 etched in different concentration of KI solutions. The unit of weight loss is mg/cm2.

PAGE 74

74 It was observed that the initial etching rate increased with the etchant concen tration. The weight loss reached a maximum after etching approximately 2 minutes for all solutions. The maximum weight loss was around 27 mg/cm2, which corresponded to an etched depth of approximately 50 m for BiI3. The weight loss curve showed almost no difference once the etchant concentration was higher than 15%. Therefore, 15% KI solution was selected as the etchant for the EPD study. The optical images of BiI3 single crystal surfaces after etched i n 15% KI solution are shown in Figure 3 13. The pits had a curved triangular shape, which represented the atomic arrangement of (00l) plane as described in Figure 24b. The size of the pits was around 60 m (the side length of the triangle) after etched for 10 sec. It was observed that the size of the pits increased with the increasing etching time. After etched for 30 sec, the pits started to overlap with each other and the number of pits became uncountable. Therefore the etching time for EPD measurem ent s was determined to be 10 and 20 sec. The typical images of etch pits for BiI3 single crystals are shown in Figure 3-14 No obvious etched pits were observed for crystals grown by the PVT method even when the etching time was extended from 10 to 60 s ec. This indicated that crystals grown by the PVT method had much lower dislocation density than crystals grown by the MVB method. W h en counting the number of the pits, if the pit crosses the corner of the area, it is counted as 1/4 pit; if the pit crosse s the side, it is counted as 1/2 pit. The pit number was counted in an area of 6.3 x 103 cm2. The results of EPD measurements are summarized in Table 3 4

PAGE 75

75 (a) (b) (c) (d) (e) (f) Figure 3 -13 Optical images of BiI3 single crystal surface after etched for time of (a) 10 sec; (b) 20 sec; (c) 30 sec; (d) 40 sec; (e) 50 sec; (f) 60 sec.

PAGE 76

76 (a) (b) (c) (d) (e) Figure 3 14 Etch pits of BiI 3 single crystals grown by MVB and PVT methods. (a) (c): MVB crystals etched for 10 sec, the crystal was grown at (a) 10oC/cm, 0.5 mm/h; (b) 10oC/cm, 1 mm/h; (c) 15oC/cm, 0.5 mm/h; (d) PVT crystal etched for 10 sec; (e) PVT crystal etched for 60 sec. It was observed that the number of pits varied from one area to another, which indicated the inhomogeneous distribution of dislocations in the MVB crystal. The EPD didn t show an apparent increase for a longer etching time (from 10 sec to 20 sec) and the extended time mainly increased the size of the pits, which is more obviously shown in Figure 3-13 The EPD of BiI3 single crystals gr own by the MVB method was on the

PAGE 77

77 order of 104 cm2 and the crystal grown at 10oC/cm and 0.5 mm/h showed the smallest EPD. The EPD value of BiI3 is comparable with other semiconductor crystals such as HgI2 (~103 105 cm2)159 161 and CZT (~104 105 cm2).3,43 Table 3 4 Etch pit density of single crystals grown by the MVB method. 3.4.6 X-ray Rocking Curve X -ray rocking curve characterization is a nondestructive analysis method compared to the EPD measurement. In this study, t he X -ray rocking curves of BiI3 (006) plane were measured as this plane showed the highest intensity in X -ray diffraction (Figure 3 10). The X -ray rocking curve -scans of BiI3 single crystals grown by MVB and PVT methods were shown in Figure 3-15 Features of extra peaks and shoulders were observed in all the crystals, which suggested that crystal planes in all these samples were not perfectly aligned. Defects such as dislocations and twin planes may account for these features. Multiple small crystals with orientation close to [001] may also exist in these crystals. The plot with log scale can reveal the low intensity features more clearly. The rocking curves scanned at two perpendicular directions for the same sample showed different shape and peak w idth. This indicated that the distribution of the defects was inhomogeneous in these crystals. T h e FWHM of the rocking curves for each sample is summarized in Table 3 5 The crystals grown by MVB method showed similar rocking curve FWHMs thus the Growth condition Etching time (s) Number of pits Average pit # EPD (104 cm 2) A rea 1 A rea 2 A rea 3 A rea 4 A rea 5 10 o C/cm 0.5mm/h 10 65.50 60.25 81.50 77.00 68.50 70.55 1.1 0.1 20 83.00 72.50 78.00 76.50 82.25 78.45 1.2 0.1 10 o C/cm 1mm/h 10 87.25 73.50 105.50 88.00 95.50 89.95 1.4 0.2 20 91.50 82.50 87.00 97.25 85.50 88.75 1.4 0.1 15 o C/c m 0.5mm/h 10 85.00 80.25 95.50 71.25 83.50 83.1 1.3 0.1 20 102.50 86.00 93.50 88.00 77.50 89.5 1.4 0.1

PAGE 78

78 cryst al quality could not be differentiated by this technique. On the other hand, the crystal grown by PVT method showed overall smoother peaks and smaller peak width compared to the crystals grown by MVB method. This suggested that crystals grown by PVT meth od have better quality, which was in agreement with the EPD result. Table 3 5 FWHM of X -ray rocking curves for BiI3 single crystals. MVB Crystals PVT Crystal 10 o C/cm and 0.5 mm/h 10 o C/cm and 1 mm/h 1 5 o C/cm and 0.5 mm/h 0 o scan 0.1818 0.3438 0.2292 0.0415 90 o scan 0.1085 0.1670 0.1839 0.1181 12.5 13.0 13.5 20k 40k 60k 80k Intensity (log) 0 degree 90 degreesIntensity (a) 10 100 1k 10k (degree) 12.5 13.0 13.5 10k 20k 30k 40k 50k 0 degree 90 degreesIntensity (log) (degree)Intensity10 100 1k 10k (b) 12.5 13.0 13.5 5k 10k 15k 20k 0 degree 90 degreesIntensity (c) 10 100 1k 10k Intensity (log) (degree) 12.5 13.0 13.5 5k 10k 15k 20k 25k 30k 0 degree 90 degreesIntensity (log) (degree)Intensity (d) 10 100 1k 10k Figure 3 15 X ray rocking curve -scans of the single crystals (a) (c): single crystals grown by MVB method at a) 10oC/cm and 1 mm/h; b) 10oC/cm and 0.5 mm/h; c) 15oC/cm and 0.5 mm/h ; (d) single crystal grown by PVT method. Solid line linear intensity; circles log scale intensit y. The rocking curve -2 scan was performed on the MVB and PVT crystals to check the distribution of the (006) d-spacing, as shown in Figure 3-16 It was observed

PAGE 79

79 that the d-spacing distribution for both crystals had wide ranges, which could partially a ccount for the broadening of the rocking curves. 25.6 25.7 25.8 25.9 26.0 0 20k 40k 60k Intensity2 (degrees) MVB crystal grown at 10oC/cm, 0.5 mm/h PVT crystal Figure 3 16 Rocking curve 2 scan of BiI 3 single crystals indicated the wide distribution of d-spacing. The poor crystalline perfection observed in BiI3 single crystals was also commonly observed in other iodine compound crystals such as HgI2 and PbI2.162 This was possibly due to the softness of all these materials. Strains induced from the thermal stress during the cooling of the crystal could easily cause the lattice distortion thus the broadening of the rocking curves. For BiI3 single crystal, simila r FWHM values w ere also reported by Nason and Keller who grew the crystal s on a seed using the PVT method. Their rocking curve measurement on (0012) plane showed FWHM values of 600 and 900 sec of arc, corresponding to 0.17o and 0.25o respectively ( Figure 3 17 ).10 The rocking curves of BiI3 have exceedingly large FWHM values compared to other single crystals with similar EPD values. For example, a typical CZT crystal with ~104 cm2 EPD value exhibited a rocking curve FWHM of ~0.01o.3 A high quality GaAs crystal with EPD less than 103 cm -2 showed a rocking curve FWHM of ~0.001o.163 It should be noted that due to the softness of BiI3, the crystal could also be deformed

PAGE 80

80 during cutting and handling, as evidenced in Figure. 3 18 It is possible that the peak broadening was exaggerated by thes e additional strains. Figure 3 17 X ray rocking curves of BiI 3 (0012) planes by Nason and Keller. A and B represent the rocking curves obtained at different sample locations.10 Figure 3 18 Deformation of BiI 3 cause d by cutting as in circled area. 3.4.7 SEM Characterization The (001) surface morphologies of BiI3 single crystal were observed by SEM. The surfaces were clean, smooth, and free of cracks, as shown in Figure 3 -19 The deformation caused by cutting was c learly observed at the edge of the single crystal (Figure 3 -19a). The layered structure of (001) planes was revealed in Figure 3 -19 b.

PAGE 81

81 The grey square box observed in Figure 319 b was the evidence of surface damag e caused by the strong electron beams, which illustrated the vulnerability of BiI3 to electrons. Figure 3 19. SEM images of BiI3 (001) surfaces show (a) deformation induced by cutting; (b) layered structure of (001) planes. 3.4.8 AFM Character ization The topographic morphologies of the cleaved BiI3 surface were characterized by AFM. The surface was observed to be atomically rough with an average surface roughness of 0. 4 84 nm ( Figure 3-20 ). Figure 3 20 AFM images of BiI 3 (001) cleaved surfaces (a) topographic morphology; (b) 3D view. (a) (b )

PAGE 82

82 The atomic layers were revealed in the topographic images with irregular edges, which was possibly due to the weak van der waal s forces betwee n the layers along [001] direction. 3.4.9 XPS Characterization The elemental composition near the surface of BiI3 single crystal was characterized by XPS. The full scanned XPS spectra are shown in Figure 3-21 1400 1200 1000 800 600 400 200 0 0 5k 10k 15k 20k Fresh surface one month in airC 1sBi 4d5/2Bi 4f7/2I 3p1/2I 3d5/2 Binding Energy (eV)I 4d I 3d3/2I 3p3/2 Bi 5d Bi 4f5/2Bi 4d3/2 IntensityO 1s Figure 3 21 XP S spectrum of freshly cleaved BiI 3 single crystal. On the freshly cleaved surface (1 day exposure in air), only carbon contamination was observed. For the surface with two weeks exposure in atmosphere, O 1s peak was clearly identified. The result sugges ted that BiI3 was stable in air for a short time, possibly several days. It will be eventually oxidized after a long time exposure in air. Therefore, BiI3 single crystals were stored in inert gas environment to avoid the oxidization. The Bi 4f and I 3d s pectra of the freshly cleaved surface were collected with higher resolution, as shown in Figure 3 22. The properties of these peaks are listed in Table 3-

PAGE 83

83 6 The I/Bi atomic ratio was calculated by the intensity ratio between the I 3d5/2 peak the Bi 4f peaks. The calculation indicated an iodine deficiency with I/Bi ratio of 2.78. 170 165 160 155 0 5k 10k 15k 20k 4f5/2Binding Energy (eV)Bi 4f4f7/2Intensity 640 635 630 625 620 615 0 10k 20k 30k 40k 50k 60k 3d5/2Binding Energy (eV)I 3d3d3/2Intensity Figure 3 22 XPS spectra of Bi 4f and I 3d for a freshly cleaved BiI 3 crystal surface. Table 3 6 Peak properties of Bi 4f and I 3d for a freshly cleaved BiI3 crystal surface. 3.4.10 XR F Characterization The XRF spectrum of BiI3 is shown in Figure 3 -23 Quantitative analysis was performed by comparing the intensities of Bi M -line to I L-line, as summarized in Table 3 -7 The result indicated an I/Bi ratio of 2.54. Considering XRF is no t a very sensitive analytical method, an uncertainty of ~10% is generally expected for the quantitative analysis.164165 The stoichiometry determined by XRF was consistent with the XPS characterization. Since XRF has a much higher sampling depth than XPS, the iodine deficiency is considered as the bulk defect of the single crystal. Element Line Peak (eV) Area FWHM Atomic Sensitivity Factors Norm Area Bi 4f7/2 1 59 .32 8612 0.93 7. 4 1163.7838 4f5/2 16 4 .64 6932 0.94 7. 4 936.7568 I 3d5/2 6 19 .47 3 5 03 7 1.14 6 5839.3333

PAGE 84

84 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 0 5k 10k 15k 20k 25k 30k 35k CountsEnergy (keV)Si K Bi M I L I L Figure 3 23 XRF spectrum of BiI3 on a freshly cleaved crystal surface. The stoichiometric deviation of BiI3 single crystal was also observed by Matsumoto et al. who reported an I/Bi ratio of ~2.9 using XRF analysis.12 T h e iodine deficiency issue has been commonly observed in other iodide semiconductor compound such as HgI2 166 and PbI2.167 It was als o reported that BiI3 tended toward dissociation into Bi and I2 at 250 300oC,5 which implied that this stoichiometric deviation may come from the evaporation of iodine after the crystal growth. Table 3 7 Quantitative results of XRF spectrum of BiI3 single cryst al. Element Intensity Atomic % Si K 35.29 1.15 Bi M 10297.84 27.96 I L 11465.62 70.89 3.5 Summary and Conclusions BiI3 single crystals have been successfully grown by the MVB and PVT methods. The phase purity of the grown single crystals was verified by XRD. The largest BiI3 single crystal was obtained with a dimension of 18 x 13 x 5 mm3. Similar to other iodide semiconductor compound, BiI3 single crystals are soft with measured Vickers hardness between 12 and 15.

PAGE 85

85 Single crystal s grown by the MVB method showed high EPD values of ~104 cm2. While the single crystals grown by the PVT method did not show apparent etch pits in 15% KI solution. This suggested that crystals grown by PVT method had much lower dislocation densities, which was verified by X -ray rocking curve characterizations. The rocking curve -scans for BiI3 single crystals grown by MVB method showed similar characteristics. Poor crystalline perfection was evidenced by the multiple rocking curve peaks and shoulders. Since MVB is a confined growth method and BiI3 is soft, the poor crystalline perfection was mainly attributed to the residual strains induced during the cooling of the crystal. The deformation during handling and cutting was also accounted for the broad rocking curves. The surface of BiI3 single crystal was observed to be crack -free and macroscopically smooth by SEM. AFM revealed an atomically rough surfa ce with average roughness of 0. 4 86 nm. XPS spectra revealed that BiI3 will be oxidized if exposed in air over two weeks. The stoichiometry of the grown single crystals was de termined to be iodine deficiency by XPS and XRF. The deficiency of iodine was also observed in HgI2 and PbI2, which was attributed to the evaporation of iodine. Among the crystals grown by the MVB method, those grown at a 0.5 mm/hour grow th velocity and 1 0C/cm temperature gradient showed the smallest EPD values and fair rocking curve FWHM Thus, single crystals grown at this condition were selected for the further evaluations. It has been observed that single crystals grown by MVB method exhibited vari ous flaws and defects, potential avenues to overcome th ese limitations will be addressed in the future work

PAGE 86

86 CHAPTER 4 BAND STRUCTURE CHARA CTERIZATION OF BISMU TH IODIDE SINGLE CRY STAL 4.1 Introduction BiI3 has been reported with different band gap charac teristics and values. The experimentally and computationally determined band gap value of BiI3 was ranging from 1.59 eV to 2. 2 eV as summarized in Table 4-1 For instance, Vashchenko and Timofeev measured the absorption spectra of BiI3 at room temperatur e and reported an indirect band gap of 1.73 0 0.00 5 eV.102,168 Jellison et al used twomodulator generalized ellipsometry to determine ordinary and extraordinary band gaps of 1.991 0.0 0 5 eV and 1.997 0.002 eV at room temperature.169 Density functional theory (DFT) calculations by Schlter et al yielded a direct band gap of ~2.2 eV.170 However, more DFT calculations by Yorikawa and Muramatsu suggested that BiI3 has an indirect band gap of approximately 1. 59 eV with a rhombohedral crystal structure.171 The much smaller band gap value was ascribed to the well known, endemic underestimation of band gaps by local density approximation (LDA) calculations There thus appears to be a discrepancy in the electronic structure and band gap values reported for BiI3 in the literature. This chapter explores the ban d gap characteristic of BiI3 through optical UV Vis spectra. The possible extrinsic factors (temperature and impurities) that could affect the band gap were discussed. The impurity concentrations were determined by inductively coupled plasma atomic emiss ion spectroscopy (ICP AES). The results were compared and discussed with the previously reported band gap characterizations.

PAGE 87

87 *value obtained as peak energy at fund ament al absorption edge 4.2 Experimental Procedure s 4.2.1 UV -Vis Spectra In this study, the band gap at room temperature was characterized by measuring optical transmission and reflection spectra at the wavelength of 300 to 800 nm. The sample is a freshl y cleaved single crystal with thickness of 0.9 mm. The light source was a Newport ARC lamp housing 66902. A parallel single wavelength light beam was generated through a CornerstoneTM 260, Newport monochromator and incident on the crystal surface at an angle of 3 to the crystal optic axis. The intensity of the incident, reflected and transmitted light was measured separately by a UV -silicon detector (Newport). The temperature dependence of the band gap was determined by optical transmission spectra from 298 to 330 K using a Varian 50 Bio UV -Vis Table 4 1 Reported b and gap of BiI 3 single crystal. Crystal growth method C haracterizati on method Band gap nature Band gap ( eV) T (K) Reference Theory EPM Direct 2. 2 0 Schlter et al 170 Theory NCPP -LDA Indirect 1.59 0 Yorikawa and Muramatsu 172 Vertical Bridgman Optical absorption Indirect 1.730 293 Vashchenko and Timofeev102,168 1.922 85 2.029 20.4 Vertical Bridgman Optical absorption Direct 1.72 298 Dmitriev et al .173 Vapor transportation Optical transmission 1.73 298 Schieber et al .8 Vapor transportation Ellipsometry 1.991 298 Jellison et al .169 Vapor transportation Optical reflection Indirect 1.97 286 Komatsu and Kaifu105 2.072 77 2.08 6

PAGE 88

88 spectrophotometer. The crystal was attached to the flat surface of a quartz vial which was imbedded in an aluminum block. Holes were drilled through the top and bottom half of the aluminum block which was connec ted to a water bath (NESLAB RTE -140). The aluminum block was heated up by the hot water running through the holes. As a result, the quartz vial and the crystal were heated by the surrounding hot aluminum. The temperature of the sample was measured by a k type thermocouple attached to the sample surface. To effectively absorb the heat, the crystal was cut to 5 x 5 mm2 with 0.1 mm thickness. 4.2.2 I nductively C oupled Plasma Atomic Emission Spectroscopy The impurity levels of the crystals were determined by inductively coupled plasma atomic emission spectroscopy (ICP -AES, Perkin Elmer Plasma 3200). The analysis was conducted on an as grown crystal ingot to determine the distribution of the impurity elements. T he polycrystalline ingot was cut into segmen ts every 5 mm starting from the tip, as illustrated in Figure 41 A small amount of BiI3 crystal ( tens of mg) was taken from the central portion of each segment and dissolved in 10% tracemetal grade hydrochloric acid ( HCl ) to form a 10 mg/mL solution. The elemental impurities analyzed in this study were Cu, Ag, Pb, Ni, Fe, and Si, which are the same impurity elements ( except for Si ) reported by the vendor. Although Si i s not reported to be an impurity element in BiI3 powder, it i s important to determin e the Si concentration in the crystal to check if there is any contamination coming from the Pyrex glass ampoule. To make a comparison, s ingle crystals were also grown in a quartz ampoule as quartz is generally accepted as a non-contamination container fo r crystal growth. The impurity concentration in the crystal grown in quartz was measured and compared to those in crystals grown in the Pyrex glass.

PAGE 89

89 Figure 4 1 Illustration of sample cutting for ICP AES characterization. 4.3 Results 4.3.1 Band Gap Determined by Optical Spectra at Room Temperature The UV -Vis transmission and reflection spectra were measured from 300 nm to 800 nm at room temperature, and are shown in Figure 42 was determined from the experimental spectra by solving the following equation:174 d d e R e R T2 2 21 ) 1 ( =-(4 -30 ) where T is transmittance, R is reflectance, and d is the sample thickness. The relation between the absorption coefficient can be written as:175 ) ( ) (/ 1 g nE hv A hv (4 -31 ) where incident photon, and Eg is the optical band gap. The exponent n depends on the type of optical transition. For direct transition, n = 1/2; for indirect transition, n = 2. In order to determine the nature of the band gap, both direct and indirect band gap relations were examined. 0 5 10 15 20 25 mm

PAGE 90

90 According to Equation (4 2 ), by plotting ( 2 1/2 versus hv and 2 1/2 = 0, the optical band gap was determined, as shown in Figure 4-3 1/2 versus hv plot, indicating an indirect band characteristic of BiI3. The band gap value was determined by the average 1/2 = 0, which gave an indirect band gap of 1.68 0.09 300 400 500 600 700 800 25 30 35 40 45 0 10 20 30 40 50 60 Transmission (%) Reflection (%)Wavelength (nm) Figure 4 2 UV Vis reflection and transmission spectra of BiI 3 single crystal. 1.65 1.70 1.75 1.80 0 10 20 0 1x1052x105 (hv )2 (eV cm-1)2 ( hv)1/2 (eV cm-1)1/2Photon Energy (eV) Figure 4 3 1/2 2 as a function of photon ene rgy.

PAGE 91

91 eV at room temperature. This value is in agreement with experimental value reported in reference s 7 106, 166 and 171, but smaller than the values reported in reference s 104 and 167, as listed in Table 4 1 4.3.2 Effect of Temperature on Band Gap Th e effect of temperature on the band gap was determined by measuring UV -Vis transmission spectra from 298 to 330 K; the results are shown in Figure 4 -4 It was observed that the absorption edge shifts to the lower energy side when the temperature of the transmitted light was solely due to the sample absorption, ) ln( 1 = T d (4 -32 ) 1/2 2 on the photon energy is plotted in Figure 4-5 600 700 800 0 10 20 30 40 50 Transmittance T (%)Wavelength (nm) 298 K 310 K 320 K 330 K Figure 4 4 UV Vis transmission spectra of BiI 3 single crystal at differ ent temperatures. 1/2 versus hv plot, which indicate that the gap is not indirect band. However, it is important to note that compared with the optical spectra discussed in sec tion 4.3.1, here the

PAGE 92

92 reflectance was not taken into account for the calculation of absorption coefficient. This is particularly relevant because equation (4 3 ) is only valid for a sample with negligible reflectivity, which is not the case here given the mirror surface of the BiI3 single crystals investigated here. Therefore, this simplified solution of absorption coefficient will not truly represent the band gap charact eristics of BiI3 as key features may be lost in the 1/2 versus hv plot when approaches zero. 1.60 1.65 1.70 1.75 1.80 1.85 1.90 0 5 10 15 20 25 30 35 hv )1/2 (eVcm-1)1/2Photon Energy (eV) 298 K 310 K 320 K 330 K a) 1.60 1.65 1.70 1.75 1.80 1.85 1.90 0 5x1051x106 (hv )2 (eVcm-1)2Photon Energy (eV) 298 K 310 K 320 K 330 K b) Figure 4 5 1/2 2 on the photon energy.

PAGE 93

93 Nevertheless, it is clearly observed that in the UV -Vis transmission spectra the absorption edge of BiI3 red-shifts with increasing temperature ( Figure 44 ). The temperature coefficient between 298 and 330 K was calculated based on the plot in Fi gure 45 a and the value is about -1.18 x 103 eV/K. This temperature effect has also been reported by other researchers, as indicated in Table 41 This red-shifting of the band gap with temperature has been generally found in many semiconductor material s.176178 An empirical fit has been experimentally determined to describe the band gap as a function of temperature:179 T T E T Eg g+) 0 ( = ) (2(4 -33 ) where Eg F itting the data obtained from Figure 4-5 a g ives Eg( 0 ) = 2.0 6 3 eV/K and = 52 K. The band gap at a bsolute 0 K is then determined to be 2.06 eV, which is about 0.2 eV smaller than the theoretical value. coefficient of the band gap, this coefficient increases with increasing temperature. Different temperature coeffic ient s with close values have been reported for BiI3. For example, Komatsu and Kaifu obtained temperature coefficient of -1.8 x 104 eV/K between 6 and 77 K, and -3.4 x 104 eV/K between 77 and 286 K Vashchenko and Timofeev reported a coefficient of -9.2 x 104 eV/K between 85 and 293 K. As a comparison, temperature coefficients of BiI3 were calculated using above fitted parameters and listed in Table 4 2 It is shown that temperature coefficients obtained in this study have slightly higher values, yet t he range of the values have good agreement with those reported in literature.

PAGE 94

94 Table 4 2 Temperature coefficient of BiI3 calculated based on the fitting parameters. 4.3.3 Effects of Impurities on Band Gap Band gap narrowing is an effect commonly observed in heavily doped semiconductor materials .180182 The physical mechanisms for band gap na rrowing fall into two categories: dopant atoms interact ing with the semiconductor and modify ing its band structure, and strain introduced by the doping process.183 The band gap narrowing effect has a general form of 3 / 1AN Eg where A is constant, and N is the dopant concentration.183 For Si, the band gap narrowing is about 0.02 eV with 1017 cm3 dopant concentration.184 If the impurities in BiI3 single crystal are considered as an extrinsic dopant and the concentration is high, a similar band gap narrowing effect is expected. The impurity levels of the BiI3 powder along with the crystals grown in Pyrex glass and quartz ampoules are reported in Table 43 The concentration was calculated based on the mole ratio between the impurity element and BiI3. The concentration of Pb in the powder and crystals was below the detection limits of ICP AES. It was observed that the impurity concentrations of the grown crystal maintained the same level as BiI3 powder The crystals g rown in Pyrex and quartz ampoules had similar impurity concentrations, which suggested that crystal growth in Pyrex glass ampoule was sufficient to maintain the impurity level of the starting material. This is a great processing advantage given the easier ampoule vacuum sealing of Pyrex when compared to quartz. Temperature (K) 6 77 85 286 293 Temperature Coefficient (eV/K) -1.4 x 104 -8.1 x 104 -8 .4 x 104 -1.2 x 103 -1.2 x 103

PAGE 95

95 Impurity concentrations of crystal grown in Pyr ex glass ampoule Distance from the tip (mm) Ag Cu Fe Ni Pb Si 0 2.20.57 0.30.01 5.50.05 0.20.01 <1 4.80.38 5 4.60.55 0.30.01 6.10.08 0.10.01 <1 6.50.46 10 2.40.10 0.50.01 6.00.24 0.20.01 <1 5.90.78 15 2.90.13 0.20.01 6.60.12 0.10.01 <1 6.30.53 20 1.30.03 0.20.01 4.30.18 0.10.01 <1 6.50.38 25 2.60.06 0.10.01 5.30.08 0.10.01 <1 4.60.34 Impurity concentrations of crystal grown in quartz ampoule. 0 2.4 0.11 0.2 0.01 5.6 0.21 0.3 0.01 <1 4.2 0.15 5 2.0 0.17 0.2 0.01 4.2 0. 11 0.3 0.02 <1 3.5 0.23 10 3.1 0.08 0.2 0.01 4.9 0.15 0.1 0.01 <1 4.7 0.36 The sum of the impurity concentrations for all the analytes were around 20 ppm in the crystal. Given that BiI3 has an atomic concentration of 2.37 x 1022 cm3, 20 ppm impurity c oncentration was converted into an atomic concentration of 4.74 x 101 7 cm3. This was considerably high doping concentration if all the impurity elements were doped into the crystal. It was therefore reasonable to expect that the band gap of BiI3 cryst al was narrowed by the impurities. The impurity distribution in the crystal grown in Pyrex glass ampoule was uniform along the axis of the crystal ingot, as shown in Figure 4 -6 which suggested that the impurities was hardly segregated during MVB crystal g rowth. Table 4 3 Impurity concentrations of BiI 3 powder and crystal (ppm). Ag Cu Fe Ni Pb Si Raw powder 3.40.60 0.80.02 7.11.86 0.10.01 <1 7.20.71

PAGE 96

96 0 5 10 15 20 25 0 1 2 3 4 5 6 7 PyrexImpurity Concentration (ppm)Distance from tip (mm) Ag Cu Fe Ni Si Figure 4 6 Impurity distribution along the BiI3 crystal grown in Pyrex ampoule. 4.4 Discussion The band gap measured at room temperature showed close value to most of the literature listed in Table 41 except for refe rence 95 and 158. Regarding to reference 95, one should be noted that the values were actually taken from the peak energy of the fundamental absorption edge, which should be higher than the intrinsic band gap as the value is taken at the maximum absorption coefficient instead of at approaching to zero. In other words, the values reported in reference 95 are related to band gap but with higher energy. They are not the true band gap values, but their change with temperature represents the temperature dependence feat ure of band gap. The band gap in reference 158 was measured by ellipsometry. A comparison of spectroscopic ellipsometry and transmission and reflection spectroscopy measurements indicates the relative strengths and weaknesses of the two techniques. Spect roscopic ellipsometry has proven to be useful in extracting the material microstructure in the form of surface features, the index of refraction below the band gap, and the critical point features above the direct gap. This technique, however, lacks the s ensitivity when the

PAGE 97

97 absorption coefficient is lower than 104 cm1. When the absorption coefficient is extracted using transmission and reflection spectroscopy, sensitivity is retained to ~ 102 cm1. Transmission and reflection measurements generally lack sensitivity to the material microstructure, which can lead to erroneous assignment of the above-gap critical point transitions. A combination of these two techniques allows for the accurate determination of the material microstructure and of the opt ical properties across the given spectral range. Figure 47 shows the absorption coefficient extracted by a combination of spectroscopic ellipsometry and transmission and reflection measurements. It is clearly apparent that when the band gap characterist ics are above 104 cm1 either technique is useful for extraction of the direct or indirect gap, but when the characteristics are lower than ~ 104 cm1 transmission and reflection measurements should be used. Thus the different band gap values obtain ed by UV Vis and ellipsometry can be attributed to the different range of absorption coefficient used for the data fitting, which is determined by the nature of the techniques. In the case of BiI3, UV -Vis is more accurate to determine the band gap as its band characteristics are lower than ~ 104 cm1. The decrease of band gap value due to increasing temperature and high impurity concentration has been observed in BiI3. However, the effects are rather small such that they cannot match the difference between experimental and theoretical va lues. Theoretical DFT calculation gave an indirect band gap value of 2.28 eV.185 The experimentally determined value was 2.06 eV at zero temperature by fitting the data obtained from 298 to 330 K. The band gap narrowing caused by impurit ies has been theoretically studied using DFT calculation, which ga ve a 0.3 eV decrease of band gap

PAGE 98

98 for Cu impurity with concentration of ~1020 cm3.185 The theoretical study also showed that not every impurity element will cause the band gap narrowing of BiI3. For instance, Si was observed to have no effect on the band structure modification. The band gap narrowing in BiI3 should be far less than 0.3 eV, since the impurity concentration in the crystal (~1017 cm3) is much smaller than the value used in DFT calculation. Figure 4 8 a spectr al range from 1.88 to 5.15 eV and transmission and reflection below 1, while transmission and reflection measurements for a thick, bulk sample are only 1. As spectrosco pic ellipsometry is a reflectionbased technique relying on the complex amplitude reflectance ratios, it retains sensitivity to provide accurate values when the material thickness is greater than the energy.185

PAGE 99

99 Although the actual value can not be determined in this study, the band gap narrowing effect should be less than 0.05 eV, as observed in other compound semiconductor materials with similar impurity concentration.182 Another possib le source that could cause the decrease of band gap is the strain within the crystal. Strain could be introduced during crystal growth by the formation of defects such as twin planes, stacking faults and incorporation of impurities. The deviation from st oichiometry as evidenced by XPS analysis could also create strains in the lattice. Strain induced band gap narrowing has been reported for various semiconductor materials.186190 For example, the induced strain in heavily phosphorous -doped (7 x 1020 cm3) silicon was reported to cause ~0.05 eV band gap reduction.188 Germanium grown on silicon substrate with a tensile strain as large as 0.2% was reported to have a band gap reduction of ~0.03 eV.190 However, even with the combination effects of temperature impurity, and strain, the band gap value obtained by experiment is still smaller than the theoretical value. Given that there is no other possible sources that could decrease the band gap, it is concluded that DFT calculation over estimated the band gap by ~0.1 eV. 4.5 Conclusions The band gap of BiI3 at room temperature was determined to have indirect character with value of 1.68 0.09 eV by UV -Vis transmission and reflection spectra. The band gap was observed to decrease with increasing temperature. An empirical model has been constructed to fit t he temperature dependence of band gap. The fitting result indicated a band gap of 2.06 eV at the absolute zero degree, which is around 0.2 eV lower than the theoretically calculated band gap. The different band gap values obtained by UV -Vis and ellipsometry were attributed to the different absorption coefficient used to extract the band gap. W hen the band gap characteristics appear

PAGE 100

100 at < 104 cm1, which is the case of BiI3, transmission and reflection measurements should be used to determine the band gap. The impurity concentration in BiI3 single crystal has been determined to be ~1017 cm3 by ICP -AES. Band gap narrowing was likely to happen at such a high impurity concentration. By comparing the band gap value determined by experiments and DFT calc ulation, it was concluded that t heoretical DFT calculation overestimated the band gap values by about 0.1 eV. In conclusion, the controversy in reported band gap values of BiI3 has been resolved. Temperature and impurities have been determined to be the major factors that affect t he band gap of BiI3. In addition, the experimentally determined band gap values also depend on how the data was fitted. For example, the ellipsometry tends to overestimate the band gap value of BiI3 as this technique is only se nsitive when the absorption coefficients are higher than 1000 cm 1, while the band characteristics of BiI3 appear at lower than 100 cm 1.

PAGE 101

101 CHAPTER 5 ELECTRICAL CONTACTS AND RESISTIVITY CHAR ACTERIZATION 5.1 Introduction For detector fabrication, a metal electrode has to be deposited on the s urface of BiI3. The proper choice of the electrode materials is extremely important for the success of the detector. Au and Pd have both been used as electrode materials for BiI3 detector s .1113 However, the prop erties of the electrical contacts have never been reported. It is important to understand how the metal contact interacts with the substrate. The reactivity of the metal -semiconductor interface should be used as a guide for the selection of contact mater ial. X -ray photoelectron spectroscopy (XPS) has been widely used to explore the metal -semiconductor interfaces with great success at explaining and predicting the reactivity of the interfaces.191192 In this study different electrode materials of Au, Pd, and Pt have been sputter coated on the surface s of BiI3 single crystal s The metal BiI3 interface has been studied by XPS and AFM. The resistivity of BiI3 single crystal s with different contact materials have been measured and compared. 5.2 Experiment Procedures 5.2.1 XPS 5.2.1.1 Principles of XPS XPS is an electron spectroscopic method that uses monoenergetic soft X -rays to eject electrons from the inner shell orbital s of a sample and analyzes the energy of the detected electrons.193 In this study, the X -ray source is a monoenergetic Al K (1486.7 eV) X ray. The kinetic energy of an ejected electron is given by s b k E hv E (5 -34 )

PAGE 102

102 where hv is the energy of the photon, Eb is the electron binding energy which is the energy differ ence between the Fermi level and the energy of the atomic orbital from which the electron originates, and s is the work function of the spectrometer. Since an atom has multiple orbitals, the corresponding photoelectron spectrum has a variety of peaks that are named according to the orbital (l = 0, 1, 2, 3 denoted as s, p, d, f ) and spin (s = 1/2) quantum numbers. The total momentum of the photoelectrons (J = l + s) is included in the labeling of a measured photoelectron peak (e.g., Bi 4f7/2 where l + s = 3 + 1/2 = 7/2). The electron binding energy of an element varies with different compounds due to the change of electronic environment. Therefore XPS can be used to identify the chemical state of the materials being analyzed. Since the energy of the X -ray source is low, the penetration depth of the photons is on the order of micrometers. Only those electrons that originate within tens of angstroms below the solid surface can escape the surface without appreciable energy loss, which makes XPS majorly used as a surface analytical technique. 5.2.1.2 Sample Preparation for XPS The (001) surface of BiI3 single crystal was cleaved in air by peeling off the top layer of the crystal using Scotch tape. The cleaved surface was sputtered with Au, Pd, and Pt metals with thickness of around 1 nm using Kurt J. Lesker CMS 18 multi target sputter system. The sputtering was conducted in a vacuum chamber with a pressure of 108 torr at a deposition speed of around 0.5 nm/s at room temperature. A freshly cleaved surfac e of a BiI3 single crystal was used as a reference. X -ray photoelectron spectra were collected by an Omicron XPS system -ray (1486.7 eV). The s pect ra were recorded at a sampling area of 1.3 mm2 and a takeoff

PAGE 103

103 angle of 5 5o with pass energy of 22 eV and step size 0.05 eV. The chamber pressure was kept below 5 x 1010 mbar during measurement. Th e instrument resolution was < 0.1 eV. The peak energies were calibrated using a Ag 3d5/2 peak at 368 eV. The studied elements included bismuth, iodine, gold, palladium, platinum, carbon, and oxygen. 5.2.2 Atomic Force Microscopy (AFM) To study the surface morphology of the metal contacts, the sample surfaces were examined by AFM with an Asylum MFP -3DTM Research AFM. The AFM tip from Veeco has a triangular cantilever and a nominal tip radius of 20 nm. The normal force constant of the tip cantilever is 0.5 8 N/m. The topography images were acquired in contact mode at a fixed applied load of several nN 5.2.3 I -V Characterization T he resistivity of BiI3 single crystal s with different electrodes was studied by I V characterization using an Agilent 4155B semic onductor parameter analyzer. Samples were prepared by sputtering Au, Pd, and Pt on the freshly cleaved surface resulting in an electrode thickness of ~ 200 nm. T h e sputtering was conducted at the same conditions as described for the XPS samples. Pt wires were attached to the electrodes using silver paste (SPI Supplies) for electric al connection. 5.3 Results and Discussion 5.3.1 The XPS Spectra The XPS core-level spectra of all the elements detected are shown from Figure 51 to 5 3 The observed elemental binding energies are summarized in Table 5 -1 The binding energies of each element in different compounds are given i n Table 5 -2

PAGE 104

104 170 165 160 4f5/2 IntensityBinding Energy (eV) BiI3 Au Pd Pt Bi 4f 4f7/2 640 635 630 625 620 615 3d3/2 IntensityBinding Energy (eV) BiI3 Au Pd Pt I 3d 3d5/2 Figure 5 -1 XPS spectra of Bi 4f and I 3d for the freshly cleave d BiI3 (001) surface and surface with Au, Pd, and Pt electrode s In the following discussion, the sample without deposited metal is called BiI3 sample, and the samples with deposited metallic contacts are named Au, Pd, and Pt sample s respectively. A quic k survey of the XPS spectra reveals that the Au sample maintains the chemical environment of BiI3, while Pd and Pt show interactions with the substrate. The detailed discussion for each sample is given in the following sections.

PAGE 105

105 95 90 85 80 0 1000 2000 3000 4000 4f7/24f5/2IntensityBinding Energy (eV) Au 4f 345 340 335 330 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 3d5/23d3/2IntensityBinding Energy (eV)Pd 3d 80 75 70 65 0 2000 4000 6000 8000 10000 12000 14000 16000 18000 4f5/2 IntensityBinding Energy (eV) Pt 4f4f7/2 Figure 5 2 X ray photoelectron spectra of Au 4f, Pd 3d, and Pt 4f on the surface of BiI 3 single crystal s

PAGE 106

106 290 285 280 IntensityBinding Energy (eV) BiI3 Au Pd Pt C 1s 540 535 530 525 IntensityBinding Energy (eV) BiI3 Au Pd Pt O 1s Figure 5 3 X ray photoelectron spectra of C 1s and O 1s for all the samples. Table 5 1 Observed binding energies of the elements Bi, I, C, O, Au, Pd, and Pt from XPS spectra. Binding energy (eV) Peak BiI 3 sample Au sample Pd sample Pt sample Bi 4f7/2 159.2 159.1 158.6 157.1 158.6 157.3 I 3d5/2 619. 4 619.1 619. 3 619 .0 C 1s 284.7 284.6 284.0 284.3 O 1s 532.2 529.9 533.0 532.1 Au 4f7/2 84.0 Pd 3d5/2 335.0 336.5 Pt 4f7/2 71.1

PAGE 107

107 Table 5 2 Binding energies (eV) of the elements Bi, I, C, O, Au, Pd, and Pt in differ ent compounds. Compounds Bi 4f7/2 I 3d5/2 C 1s O 1s Au 4f7/2 Pd 3d5/2 Pt 4f7/2 Ref. Bi 156.9 157.1 193 194 BiI 3 159.3 193 Bi2O3 158.6 159.8 529.8 530.22 193,195 197 I 2 619.9 193 HgI 2 619.4 193 Amorphous carbon 284.2 284.8 198 200 CO 285.3 285.9 532.1 199 201 CO 2 291.9 202 Au 83.9 84.2 193 Au 2 O 3 530.2 85.9 193,203 Pd 335.0 335.3 193 PdI 2 336 .4 193,204 PdO 529.3 335.2 337.1 193,205 206 PdO2 337.7 338.2 207 208 Pt 71 71.2 193 PtO 73.8 74.2 193 PtO2 531.4 531.9 74.6 75.0 193 5.3.1.1 Au S ample The XPS spectra of the Au sample show peak positions of Bi 4f and I 3d very close to those of the BiI3 sample; and the Au 4f spectrum exhibits the same peak

PAGE 108

108 positions with that of a pure gold. The results suggest that Bi, I, and Au maintain their respective chemical state and there is no chemical interaction between Au and BiI3. The C 1s and O 1s spectra are shown in Figure 5 4 295 290 285 280 1400 1600 1800 2000 2200 IntensityBinding Energy (eV) C 1s in Au sample Deconvoluted curves Background Fitted curve 284.5 eV 285.7 eV (a) 540 535 530 525 1800 2100 2400 2700 3000 3300 3600 IntensityBinding Energy (eV) O 1s in Au sample Deconvoluted curves Background Fitted curve (b) 529.9 eV 532.2 eV Figure 5 4 Deconvolution of XPS spectra for the Au sample peaks (a) C 1s and (b) O 1s. The C 1s spectrum of the Au sample can be deconvoluted in to two peaks with the peak at 284.5 eV assigned to amorphous carbon198 and the peak at 285.7 eV assigned to carbon species bound to oxygen199 201 (Figure 5 -4 a). Two distinct peak s are observed in O 1s spectrum ( Figure 5-4 b). The peak at 532.2 eV is attributed to oxygen atoms bound to carbon.199 The peak at 529.9 eV could be assigned to oxygen atoms in Au2O3 (530.2 eV)7 or atomic O a d sorbed at Au surface (530.2 eV).199 Since there is no indication of any Au -O bond in the Au 4f spectrum, this peak is assigned to atomic O adsorbed at the Au surface. 5.3.1.2 Pd Sample Peak shifting of Bi 4f and Pd 3d is observed, which indicates the existence of an interfacial reacti vity. The iodine peaks remain the same as the BiI3 sample. The d econvolution of Bi 4f, Pd 3d, C 1s, and O 1s spectra are shown in Figure 5 5 The properties of the fitted peaks and their assignments are summarized in Table 5 3

PAGE 109

109 170 165 160 155 1000 1500 IntensityBinding Energy (eV) Bi 4f in Pd sample Deconvoluted curves Background Fitted curve (a) 157.1 eV 158.5 eV 159.2 eV 162.4 eV 163.8 eV 164.4 eV 345 340 335 330 2000 4000 6000 8000 IntensityBinding Energy (eV) Pd 3d Deconvoluted curves Background Fitted curves (b) 335.0 eV 336.5 eV 340.3 eV 341.7 eV 295 290 285 280 800 1000 1200 1400 1600 IntensityBinding Energy (eV) C 1s in Pd sample Deconvoluted curves Background Fitted curve (c) 283.9 eV 285.6 eV 540 535 530 525 1500 2000 2500 3000 3500 4000 O 1s in Pd sample Deconvoluted curves Background Fitted curveIntensityBinding Energy (eV)(d) 533.1 eV 532.1 eV Figure 5 5 Deconvolution of XPS spectra of the Pd sample for (a) Bi 4f, (b) Pd 3d, (c) C 1s, and (d) O 1s. In the Bi 4f7/2 region, three peaks are identified at 157. 1, 158.5, and 159.2 eV, which are consistent with the literature values assigned to bismuth atoms in metallic bismuth,194 Bi2O3,195 and BiI3 193 respectively. However the assignment of Bi2O3 is not plausible as the corresponding amount of oxygen is not observed in the O 1s spectrum. The peak at 158.5 eV is possibly due to bismuth atoms in a reduced iodine environment (BiI3 x with x < 3). The Pd 3d5/2 peak deconvolut es in to two peaks at 335.0 and 336.5 eV, which are assigned to palladium atoms in metallic palladium193 and PdI2 12 respectively.

PAGE 110

110 Table 5 -3 Peak propert i es of Bi 4f, Pd 3d, Pd 3p, C 1s, and O 1s observed in the Pd sample. Spectrum Position (eV) Element FWHM (eV) Area ASF Corrected Area Assignment Bi 4f 157.1 Bi 4f7/2 0.87 716 7.4 97 Bi Bi 4f 158.5 Bi 4f7/2 0 .87 460 7.4 62 BiI 3 x Bi 4f 159.2 Bi 4f7/2 0.87 152 7.4 21 BiI 3 Bi 4f 162.4 Bi 4f5/2 0.87 537 7.4 73 Bi Bi 4f 163.8 Bi 4f5/2 0.87 345 7.4 47 BiI 3 x Bi 4f 164.4 Bi 4f5/2 0.87 114 7.4 15 BiI 3 Pd 3d 335.0 Pd 3d5/2 1.03 4512 4.6 981 Pd Pd 3d 336.5 Pd 3d5 /2 1.03 6796 4.6 1477 PdI 2 Pd 3d 340.3 Pd 3d3/2 1.03 2978 4.6 647 Pd Pd 3d 341.7 Pd 3d3/2 1.03 4486 4.6 975 PdI 2 C 1s 283.9 C 1s 1.38 871 0.25 3484 amorphous carbon C 1s 285.6 C 1s 1.38 143 0.25 572 C bound to O O 1s 533.1 Pd 3p3/2 3.69 5862 1.43 4099 Pd O 1s 532.1 O 1s 1.80 378 0.66 573 O bound to C The C spectrum of the Pd sample has two peaks with the one at 283.9 eV assigned to amorphous carbon and the one at 285.6 eV assigned to carbon species bound to oxygen .193 The O 1s spectrum is dominated by Pd 3p3/2 peak at 533.1 eV.193 The intensity of this peak (area under the peak) is consistent with the peak intensity in Pd 3d spectrum, which validates this peak assignment. The identification of oxygen peak in O 1s spectrum is difficult as the majority peak intensity is contributed by Pd atoms. The peak is fitted at 532.1 eV with a fixed 1:1 intensity ratio to the corresponding C 1s peak at 285.6 eV in C 1s spectrum. XPS analysis indicates the existence of metallic Bi and PdI2 at the Pd -BiI3 interface, which suggests that Pd reacts with BiI3. The majority of the bismuth elements on the surface exist in the form of metallic Bi (~54%) and BiI3 x (~35%), with a small portion (~11%) that remains as BiI3. The reactivity between Pd and iodine-rich crystals

PAGE 111

111 has been previously reported.209 However, Pd is still widely used as the electrode contacts for iodide compound materials such as HgI2 and PbI2.75,129,210 There is no discussion on how this Palladium iodine reactivity will affect the detector performance. This suggests that a careful reevaluation of the electrode contacts for these materials is necessary. 5.3.1.3 Pt Sample Similar to the Pd sample, peak shifting is also observed in the Bi 4f spectrum in the Pt sample. The Pt 4f peaks exhibit an asymmetric feature, indicative multiple chemical states for the Pt atoms found in the interfacial region. However, this asymmetry is not due to reactivity there fore no attempt is made to deconvolute the peaks. The similar peak feature has been previously observed in Pt nanoparticles on pyrolytic graphite surfaces .211 Since Pt was reported to have different electronic configurations for surface and bulk atoms,212 the asymmetric peak was explained due to the surface Pt atoms in the 5d96s1 configuration, bulk Pt ato ms in the 5d86s2 configuration, and surface Pt absorbed oxygen.211 The d econvolution of Bi 4f, I 3d, C 1s, and O 1s spectra of the Pt sample are shown in Figure 56 The peak properties and their assignment s are summarized in Table 5 4 The assignment for Bi 4f spectrum is similar to that in Pd sample. The Bi 4f7/2 peaks at 157.1, 158.3, and 158.9 eV are assigned to bismuth atoms in metallic bismuth, BiI3 x, and BiI3 respectively. The existence of I2 is evidenced by the peak at 619.4 eV in I 3d spectrum.193 The two peaks in C 1s spectrum are assigned to amorphous carbon at 284.0 eV and carbon species bound to oxygen at 285.9 eV. The O 1s spectrum deconvolutes into two peaks with the one at 529.5 eV assigned to oxygen bound to Pt

PAGE 112

112 and the one at 532.0 eV assigned to oxygen bound to carbon. The uphill tail in the low energy side of O 1s spectrum is the trail of Pt 4p3/2 peak at 520 eV.193 170 165 160 155 2000 3000 4000 IntensityBinding Energy (eV) Bi 4f in Pt sample Deconvoluted curves Background Fitted curve (a) 157.1 eV 158.3 eV 158.9 eV 162.4 eV 163.6 eV 164.2 eV 640 635 630 625 620 615 2000 4000 6000 8000 10000 12000 14000 IntensityBinding Energy (eV) I 3d in Pt sample Deconvoluted curves Background Fitted curve (b) 618.6 eV 619.4 eV 630.1 eV 630.9 eV 295 290 285 280 1400 1600 1800 2000 2200 IntensityBinding Energy (eV) C 1s in Pt sample Deconvoluted curves Background Fitted curve (c) 284.0 eV 285.9 eV 540 535 530 525 2800 3000 3200 3400 IntensityBinding Energy (eV) O 1s in Pt sample Deconvoluted curves Background Fitted curve (d) 529.5 eV 532.0 eV Figure 5 6 Deconvolution of XPS spectra of the Pt sample for (a) Bi 4f, (b) I 3d, (c) C 1s, and (d) O 1s. The XPS results indicate that Pt is not good as the contact material for BiI3 either. The effect of Pt on the surface of BiI3 is different from that of Pd. Here the presence of Pt causes the reduction of BiI3. But unlike Pd, Pt does not involve in the reaction. The function of Pt in this case is similar to that of a cat alyst.

PAGE 113

113 Table 5 4 Peak properties of Bi 4f, I 3d, C 1s, and O 1s observed in the Pt sample. Spectrum Position (eV) Element FWHM (eV) Area ASF Corrected Area Assignment Bi 4f 157.1 Bi 4f7/2 0.86 2303 7.4 311 Bi Bi 4f 158.3 Bi 4f7/2 0.86 1501 7.4 2 03 BiI 3 x Bi 4f 158.9 Bi 4f7/2 0.86 891 7.4 120 BiI 3 Bi 4f 162.4 Bi 4f5/2 0.86 1727 7.4 233 Bi Bi 4f 163.6 Bi 4f5/2 0.86 1125 7.4 152 BiI 3 x Bi 4f 164.2 Bi 4f5/2 0.86 669 7.4 90 BiI 3 I 3d 618.6 I 3d5/2 1.12 10385 6 1731 BiI 3 I 3d 619.4 I 3d5/2 1.12 4 365 6 728 I 2 I 3d 630.1 I 3d3/2 1.12 7555 N/A BiI 3 I 3d 630.9 I 3d3/2 1.12 2550 N/A I 2 C 1s 284.0 C 1s 1.46 884 0.25 3536 amorphous carbon C 1s 285.9 C 1s 1.46 259 0.25 1036 C bound to O O 1s 529.5 O 1s 1.59 185 0.66 280 O bound to Pt O 1s 532.0 O 1s 1.59 624 0.66 945 O bound to C 5.3.2 AFM Analysis AFM characterization has been used to check the distribution of the sputtered metals on the surface of BiI3 ( Figure 5-7 ). It is observed that Au fully cover s the BiI3 surface while Pd and Pt form islands, which indicates a completely wetting of Au on the BiI3 surface while nonwetting characteristic of Pd and Pt. The islands on the Pt sample surface are more sparsely distributed with higher height compared to the Pd sample. To verify if the islands a re formed by the deposited metals or by other species produced in the metal -BiI3 interaction, the volume of the islands is calculated using an image processing software (Scanning Probe Image Processing, SPIPTM) by analyzing the scanned histogram of the sur face. The calculated island volume for Pd sample is 4.6 x 106 nm3 on a 4 m2 area, which is corresponding to a thickness of 1.15 nm if the Pd island volumes were evenly spread on the surface. This matches well with the

PAGE 114

114 experimentally deposited thickness of 1 nm and suggests that the islands are mainly formed by Pd. The island volume on Pt sample is calculated to be 7.7 x 108 nm3 on a 100 m2 area, which is corresponding to a deposited thickness of 7.7 nm. This result has a large deviation from the experiment (1 nm), which suggests that species generated by the reaction, i.e. I2, may deposit on the surface. (a) (b) (c) Figure 5 7 AFM image and 3D vie w of topography for the (a) Au sample, (b) Pd sample, and (c) Pt sample.

PAGE 115

115 5.3.3 I -V Characterization The resistivity of BiI3 single crystal s has been measured with Au, Pd, and Pt electrodes. The crystal samples were cut from adjacent positions o f a large single crystal to minimize the sample variation. The I -V curves for crystals with different electrodes are shown in Figure 58 It is observed that the crystal with Au electrodes has the highest resistivity, and the one with Pd electrodes shows the lowes t, as summarized in Table 5 5 -400 -200 0 200 400 -2x10-6-1x10-60 1x10-62x10-6 Au Pd PtCurrent Density (A/cm2)Electric Field (V/cm) Figure 5 8 I -V curves for BiI3 single crystal s with Au, Pd, and Pt electrode s The similar results were also reported by Fornaro et al on BiI3 crystals grown by PVT method.13 Given that Bi3+ ions are reduced to metallic Bi at the Pd-BiI3 interface, the Bi atoms c ould drift across the crystal body under an external electric field. This will decrease the effective thickness between the electrodes. Consequently, the measured resistivity is smaller than the actual value. B ased on the XPS results Au is expected to be physically contacted with BiI3, the measured resistivity thus represents a more accurate value of this material. The

PAGE 116

116 sample with Pt electrodes shows resistivity in between which is possibly due to the drift of Bi3+ ions resulting from less binding of iodine. Table 5 5 Sample information and resistivity of BiI3 single crystals for I -V characterization. Sample # Electrode Surface Area (cm 2 ) Thickness (cm) Resistivity 10 9 ( ) 1 Au 0.260 0.075 1.85 2 Pd 0.167 0.075 0.28 3 Pt 0.251 0.076 0.61 5.4 Conclusions XPS characterization indicates that sputtered Pd and Pt have chemical interactions with BiI3, while Au is observed to be unreactive with BiI3. Pd reacts with BiI3 and forms metallic Bi and PdI2. Pt does not react with BiI3 but functions as a catalyst and causes the reduction of Bi3+ ions. The I -V characterization collaborates with the XPS result, and the highest resistivity is obtained from sample with Au electrodes. It is concluded that Au should be used as the electrode contact for BiI3, while Pd and Pt are not recommended. Since the reactivity is observed to be between Pd and iodine, and Pd has long been used as electrode material for HgI2 and PbI2, this sugg est that an evaluation of the interface between Pd and these material is necessary for a better understanding of this interface issue.

PAGE 117

117 CHAPTER 6 SUMMARY AND FUTURE W ORK 6.1 Summary The increasing demand for room temperature gamma -ray detectors encourages the discovery and development of high quality compound semiconductor materials. The required material properties include high density and atomic number s for high gammaray detection efficiency; moderate bandgap between 1.5 eV and 3 eV for room temperatur e operation and generation of large number of electron -hole pairs and high resistivity for improved signal to noise ratio. BiI3 meets all these criteria thus has the potential to be a successful material for room temperature gamma -ray d ete ction applicati ons Crystal growth is the most challenging and time-consuming stage of developing a new detector material. This work presented the efforts of growing large volume BiI3 single crystals using MVB method. The whole growth process is composed of different s tages including heating up from room temperature to the maximum process temperature, soaking for a period of time, cooling down to growth temperature, crystal growth, and post -growth cooling down to room temperature. Each stage has been carefully designed and tested to ensure the smooth growth run. The crystal growth was conducted at a total of nine growth conditions (variation of growth rate and temperature gradient at solid-liquid interface), among which only three conditions produced single crystals. The phase purity of the grown single crystals was verified by XRD. The quality of the single crystals was characterized by EPD measurement and X -ray rocking curve -scans. For EPD measurement, the etching recipe was determined to be 10 sec immersion in 1 5% KI solution. The observed EPD values for all the crystals were in

PAGE 118

118 the order of magnitude of ~104 cm2, which was comparable to other compound semiconductor crystals such as HgI2 and lab produced CZT, but one order of magnitude higher than the commercial CZT crystals. The high EPD values indicated the high dislocation densities in the crystals. Among the crystals grown by MVB method, the one grown at 10oC/cm and 0.5 mm/h showed the smallest average EPD values. The rocking curve -scan peaks for all the crystals showed similar FWHM and peak characteristics. Peak broadening with multiple peaks and shoulders were observed. The poor crystalline perfection was commonly observed among the iodide compound crystals such as HgI2 and PbI2, which implied that the imperfections were possibly caused by the residual strains induced during the cooling of the crystals due to the softness of these materials. In addition, the deformation during handling and cutting could also contribute to the broadening of the rocking curves. Although the BiI3 single crystals grown at three different conditions did not show distinct differences of EPD values and rocking curve characteristics, the crystals grown at 10oC/cm and 0.5 mm/h were selected for further studies due to their slightly better qualities. A single crystal with a dimension of 18 x 13 x 5 mm3, the largest crystal size ever reported for BiI3, has been obtained at this growth condition. The cleaved surface of BiI3 was characterized by SEM and AFM respectively. The SEM characterization indicated a crack -free and macroscopically smooth surface. The AFM topographic image revealed an atomically rough (001) surface with average roughness of 0. 4 86 nm. The irregular edges of the atomic layers were att ributed to the weak van der waal s forces between the layers. The chemical composition at the surface of BiI3 was characterized to be of BiI2.78 by XPS. The iodine deficiency was

PAGE 119

119 attributed to the dissociation of BiI3 into Bi and I2 during cooling of the crystal, and the evaporation of iodine when exposed in air. The surface oxidation was also observed for BiI3 crystals exposed in air for over two weeks, which suggested that a proper sealing or inert gas environment is necessary for the crystal storage. Through UV -Vis transmission and reflection spectra, the band gap of BiI3 was characterized to be of 1.68 0.09 eV with indirect transition characteristic. The band gap value of BiI3 was observed to decrease with increasing temperature, which is commonly observed in semiconductor materials. The impurity concentrations in commercial powder and BiI3 crystals were observed to be as high as 1017 cm3, at which the band gap narrowing effect was likely to happen. T h e band gap value determined in this work was consistent with previous studies using UV Vis spectra. The indirect band transition was confirmed by the most recent theoretical calculation on the band structure of BiI3 using DFT. The large difference between the band gaps determined by UV -Vis spectra and ellipsometry was attributed to the different absorption coefficient used to extract the band gap. For BiI3, band gap determined by UV -Vis spectra was more accurate. The metal -semiconductor interface was characterized for BiI3 (001) surface by XPS. The results indicated that the commonly used Pd electrode reacted with B iI3 and produced Bi metal and PdI2 on the surface. As another commonly used electrode material, Au was unreactive with BiI3. The third contact material, Pt, was observed to cause the reduction of Bi3+ ions, although it did not react with BiI3. I -V chara cterization indicated that BiI3 crystal with Au electrode exhibited the highest resistivity, which more accurately represented the resistivity of BiI3 since only Au was physically contacted with

PAGE 120

120 the crystal surface. Based on the above analysis, Au was rec ommended for the contact material for BiI3, while Pd and Pt should not be used without prior surface modification or passivation steps The observed reactivity between Pd and BiI3 could be extended to other iodine based semiconductor materials such as HgI2 and PbI2. This suggested that the re evaluations of the contact metals for all these materials are necessary Through the characterizations on BiI3 single crystals grown in this work, it was found that the crystals have high impurity concentrations and dislocation densities. The crystal quality is not comparable to the commercial CdTe and CZT crystals. The future work should focus on resolving these problems and developing BiI3 towards a room temperature gamma-ray spectrometer. 6.2 Future work 6.2.1 Ti lted Growth For vertical Bridgman crystal growth, the effect of gravity force could be significant. The mechanical stress due to the gravity force can be expressed as A gV / (6 -35 ) where is the density of the crystal, g is the gravitational constant, V is the volume of the crystal and A is the cross area at where the stress is measured. For BiI3, = 5.8 g/cm3, g is 9.81 N/kg, the stress can be simplified as ] [ ) / ( 0 569 ] [ cm A V Pa (6 -36 ) The ratio V/A varies with the location on the crystal. Since tipped ampoule was used for the crystal growth, the stress at the tip area could be very large. For example, for the crystal with 1 cm long tip of 3 cm in total length, the stress at the tip area with 1 mm diameter could be as high as 120 kPa ( Figure 61 ).

PAGE 121

121 Figure 6 1 Illustration of the crystal in a tipped ampoule. The critical resolved shear stress (CRSS) reported for HgI2 was around 0.1 MPa on (001) face.129 Considering BiI3 is softer than HgI2, appreciable slip may occur at such a high stress, which could lead to the observed high dislocation density and poor crystal perfection. The mechanical stress induced by gravity could be alleviated by tilting the ampoule for crystal growth. On one hand, the cross -section area will be increased by the tilting thus decreasing the stress applied on it. On the other hand, partial of the gravity force will be taken by the ampoule therefore the actual gravity force applied on the crystal will be greatly decreased. T h e angle of tilting needs to be experimentally determined for the crystal growth 6.2.2 Synthesis of High Purity BiI3 Powder The ICP -AES measurement indicated that even the commercial BiI3 powder have high impurity concentration. One way to solve this problem is to synthesize BiI3 in lab using high purity (99.999% and up) Bi and I2 particles. The synthesis could be done by sublimation method in a gas flowing ampoule, as illustrated in Figure 6-2 In this setup, Argon gas at a flow rate of 40 cc/min is passed over I2 at 25oC and the gas mixture reacts with Bi at 310oC when passing over it. The produced BiI3 vapor passes along with the Argon gas and deposits at 210oC for the crystal growth. The 3 cm 1 cm

PAGE 122

122 excess I2 condenses at the coolest part of the apparatus. The boat that holds bismuth and BiI3 crystals in the illustration is necessary to provide a removable secti on in the growth region so that the crystals could be easily separated from the condensed iodine. The collected high purity crystal platelets can be used for the MVB crystal growth. Although Pyrex glass ampoule was used in the original study,213 quartz or fused silica tube was recommended in other studies to largely reduce the possible contaminations.10,13 A horizontal tube furnace with precise temperatur e control is required for this process. Figure 6 2 Illustration for the sublimation growth of BiI 3 (After Curtis and Brunner 213 ) 6.2.3 Sublimation of BiI3 Powder Another way to improve the purity of the crystal could be done by multiple sublimation of the commercial powder. The experimental setup can use the same one shown in Figure 6 -2 Here the bismuth in Figure 6 -2 will be replaced by BiI3 powder. Since BiI3 tends to dissociate to Bi and I2 at 250 300oC, keep supplying iodine vapor could help suppress this reaction. One would immediately realize that the crystal platelets produced by the sublimation method could also be further purified by this method, which should be included in the future work.

PAGE 123

123 6.2.4 Post -growth Annealing High dislocation density was observed in the single crystals grown by MVB method. The dislocation density could be decreased by performing thermal annealing on the grown crystals. This method has been generally used for the post growth treatment of single crystals to improve the crystal qualities.214217 The dislocation density after annealing has been reported to be reduced by up to one order of magnitude for HgCdTe and C dZnTe.215,217 In addition, the residual stress and strain in the crystal were also observed to be greatly reduced by proper annealing process.218 219 Thus, better crystalli nity is expected after annealing. For BiI3, the temperature and time for annealing should be determined by characterizing the annealed crystal with EPD and X -ray rocking curve as there is no reported information available. The annealing atmosphere, eithe r under vacuum, Argon gas, or I2 vapor should also be experimentally determined. 6.2.5 Encapsulation of the Crystal As indicated by XPS characterization, the surface of BiI3 will be oxidized after a long time exposure in air. A proper encapsulation of the crystal and detector is therefore necessary to maintain the stability. HgI2 is normally encapsulated in a Parylene or amorphous carbon film to prevent the sublimation in air.68 The same material could be applied for BiI3 and should be tested in future work. 6.2.6 Surface Preparation In this work, the surface of BiI3 was prepared by cleaving the (001) planes, which was done by peeling off the top surface layers using Scotch tape. This will potentially cause the artifact of the surface morphologies and create variations from sample to sample. Post growth surface treatment has long been established for HgI2 and CZT

PAGE 124

124 crystals during the detector fabrication to improve the electrode-material interface and reduce the leakage current. Various surface treatments including passivation of the surface,220 polishing,221 and chemical etching222 have been tested and evaluated before th e finalized procedure was determined. Thus, a standard surface preparation procedure should be proposed and tested for BiI3 surface in future. The work should start with a similar surface treatment to HgI2.

PAGE 125

125 APPENDIX A PURIFICATION OF BISM UTH IODIDE POW DER A.1 Introduction As discussed in chapter 4, the commercial powder was found to have high impurity concentration up to 1017 cm3. T h e impurities are deteriorative to the performance of the detector as they work as charge trapping center. The impuritie s also reduce the resistivity of the crystal. This appendix discusses the efforts of purifying the commercial BiI3 powder using sublimation and vertical zone melting method respectively. A.2 Experiment Procedures For sublimation process, 20 g commercial BiI3 powder was vacuum sealed in a 16 long Pyrex glass ampoule. The ampoule was pre -cleaned as described in chapter 3. The sublimation was conducted in a 3zone horizontal furnace with only two zones utilized. The hot zone was set at 330oC and the col d zone at 250oC For vertical zone melting process, 40 g commercial BiI3 powder was vacuum sealed in a Pyrex glass ampoule and placed in the vertical furnace. The zone refinement was conducted at 460oC with a speed of 2 cm/h down the length of the ingot. Ten passes were repeated for the purification process. A.3 Results and Discussion A.3.1 Sublimation The powder was completely sublimed after 24 hours. I t was observed that BiI3 polycrystalline slabs evenly deposited through a 12 long range inside the ampoule (Figure A 1 ). The sublimed crystals were divided into three segments with 4 long each. They were named as front, middle, and end segment with increasing distance from the

PAGE 126

126 source. A small portion of the crystals were taken from each segment and s ubjected to ICP -AES for impurity analysis. Figure A 1 BiI 3 powder in Pyrex ampoule after sublimation. The measured impurity concentrations are summarized in Table A -1. It was observed that Cu and Fe were purified by sublimation, while Ag, Ni, and Si did not show appreciable concentration variation across the whole sublimation length thus they are hard to be purified by sublimation. T h e front segment showed the highest Cu and Fe concentration, which suggested that Cu and Fe tended to condense at hi gher temperature. Table A -1 Impurity concentrations of commercial BiI3 powder before and after sublimation (ppm). Ag Cu Fe Ni Pb Si Before sublimation 3.40.60 0.80.02 7.11.86 0.10.01 <1 7.20.71 Purified (front) 4.50.72 0.90.04 10.21.21 0.30. 02 <1 6.70.83 Purified (middle) 3.20.70 0.60.02 6.50.85 0.20.02 <1 7.10.52 Purified (end) 4.20.43 0.50.02 8.31.05 0.20.01 <1 6.80.61 A.3.2 Vertical Zone-melting The ingot after zonemelting was cut to several 10 mm segments. A small amount of BiI3 crystal (several mg) was taken from the central portion of each segment and dissolved in 10% tracemetal grade hydrochloric acid (HCl) to form a 10 mg/mL solution for ICP -AES measurement. The measured concentration for each impurity element is sho wn in Figure A -2. It was observed that Ni and Pb were not purified by zone melting. The Fe impurity was pushed from the top of the ingot to the bottom. The other impurity elements

PAGE 127

127 including Cu, Ag, and Si had the reverse concentration distribution. They had the highest impurity concentration at the top and the lowest at the bottom. The result suggested that the segregation coefficient (Cs/CL) for Fe is less than 1, while Cu, Ag, and Si have segregation coefficient higher than 1. In other words, Fe has higher solubility in liquid BiI3 while Cu, Ag, and Si have higher solubility in solid BiI3. This indicated that the impurities cannot be purified simultaneously by just performing zone melting process. However, a combination of this method with sublimati on could potentially achieve a high level purification. Zonemelting could be used to purify Cu, Ag, and Si first. Then the purified ingot is sublimed to purify the Fe impurity. The sublimation should be conducted more than once to achieve better purifi cation. 0 10 20 30 40 0 2 4 6 8 10 12 14 Concentration (ppm)Distance from Tip (mm) Cu Ag Pb Ni Fe Si Figure A 2 Impurity concentrations of BiI 3 powder after vertical zone melting refinement. A.4 Summary and Conclusions Through sublimation and vertical zonemelting processes, it was observed that sublimation is usefu l to purify Cu and Fe, and vertical zone-melting is useful for Cu, Fe, Ag, and Si. However, Fe cannot be purified simultaneously with Cu, Ag, and Si for zone -refinement due to their different segregation coefficients. A combination of

PAGE 128

128 sublimation and ver tical zonemelting will be the best solution to purify BiI3 single crystal.

PAGE 129

129 APPENDIX B GROWTH OF BISMUTH IO DIDE CRYSTAL WITH DI FFERENT DIAMETERS B.1 Introduction The growth of BiI3 single crystal with 0.75 diameter by MVB method has been discussed in c hapter 3. As large volume crystal is desired for the application of gamma ray detector, this appendix presents the attempts to grow BiI3 single crystals of 1 and 1.25 diameters using MVB method. The single crystal with 0.5 diameter was also grown usin g the same methodology. B.2 Experimental Procedures The relationship between the temperature gradient and crystal diameter has been described in equation (34), which indicates that the temperature gradient Gs is proportional to the square root of (1/r), w here r is the crystal radius. It is also shown in equation (31) that at equilibrium, the growth rate is proportional to the temperature gradient if the radial gradient is ignored. In this study, the growth conditions for growing 0.75 diameter crystal w ere selected as the reference. The detailed temperature gradient and growth rate were selected as 10oC/cm and 0.5 mm/h. The initial growth conditions for other diameters were determined based on equation (3-1) and (34), as shown in Table B 1. It is sho wn that for larger diameter crystal growth, both temperature gradient and growth rate should be lower. Physically this means that at larger diameter, the temperature gradient is lowered to reduce the amount of the heat that need to be conducted away. Th e growth rate is also decreased to give more time for the crystal to conduct away the increased amount of heat due to the crystallization of more crystals.

PAGE 130

130 Table B-1 Growth conditions for growing BiI3 crystals with different diameters. Crystal Diameter (inch ) Growth Conditions Temperature gradient (o C) Growth rate (mm/h) 0.75 10 0.5 0.5 12.2 0.61 1 8.7 0.44 1.25 7.7 0.39 B.3 Results and Discussion It was observed that 0.5 diameter single crystal was successfully grown. The grown crystal rod and cleaved crystal (001) surface are shown in Figure B -1. However, the crystal growth for 1 and 1.25 diameters ended up with polycrystalline. (a) (b) Figure B 1 BiI 3 single crystal with 0.5 diameter. (a) the as grown crystal rod, and (b) cleaved single crystal (001) surface. It is noticed that equation (3 -4) overestimates the temperature gradient. In addition, the radial gradient may not be ignored anymore at larger crystal diameter. Equation (3 -1) should be corrected as r r L L s s sG K G K G K LR (B-1 ) where Kr and Gr is the radial thermal conductivity and temperature gradient respectively. Thus the growth rate could be higher than the initial value. Based on the above discussion, the temperature gradient for growing 1 diameter crystal was reduced to 8oC/cm, and the growth rate was increased to 0.5 mm/h. The growth still ended up with polycrystalline. Due to the unavailability of the thermal properties of BiI3, the growth conditions are hard to be est imated and tested. It is impossible to try out every growth

PAGE 131

131 condition for the sake of time. Beside growth condition, the stress induced by gravity force could also cause the failure of the crystal growth. As discussed in 6.2.1, the stress at the tip are a is already very large for the 0.75 diameter crystal. When the crystal diameter increases, the stress could increase even more and cause serious slip at the growth face. A tilted growth should be set up to grow the large diameter crystal and verify thi s assumption. B.4 Summary and Conclusions Through theoretical deduction, the growth conditions for growing BiI3 single crystals at different diameter have been determined. Single crystal was only successfully grown for a smaller diameter of 0.5. The cry stal growth was failed for the larger diameters of 1 and 1. 2 5 It was proposed that the growth conditions may not be correctly determined. Stress induced by gravity force could also account for the failure of the crystal growth.

PAGE 132

132 APPENDIX C TEM SAMPLE PREPARATION BY FOCUSED ION BEAM C.1 Introduction I t is of interest to characterize the microstructure of the grown single crystals by transmission electron spectroscopy (TEM). The general sample preparation can be done through mechanical polishing, chemi cal etching, and focused ion beam (FIB). As BiI3 reacts with H2O, mechanical polishing and chemical etching are not suitable for BiI3. Thus FIB becomes the only way to prepare TEM sample of BiI3. This appendix discusses the efforts of preparing TEM samp le using FIB technique. C.2 Experimental Procedures and Results As BiI3 is highly sensitive to ion beams, the standard auto FIB procedure is not applicable for it. Manual operation has to be used to avoid the damage to the sample. After a series of tries a standard procedure has been determined for BiI3. The FIB was always conducted at an ion beam voltage of 30 kV. The sample was sputter coated with a thin film of carbon before performing FIB. The detailed procedure is described as follows: The sample was tilted to 53o. A Pt layer with dimension of 20 x 3 x 1 m (x, y, z) was deposited on top of the sample to protect it from ion beam damage. The ion beam current for deposition is 300 pA. The surface after Pt deposition is shown in Figure C -1.

PAGE 133

133 Figure C 1 Pt deposition. Surface carbon layer was removed at ion beam current of 3000 pA with z = 0.1 m. Figure C 2 Carbon layer removal. Trenches were cut on both sides of the Pt layer at ion beam current of 1000 pA with z = 0.3 m.

PAGE 134

134 Figure C 3 Formation of trench. The sample slab was thinning down to 1 m at ion beam current of 300 pA. Figure C 4 Thinning down to 1 m. The sample slab was further thinning down to 0.3 m at a smaller ion beam current of 50 pA.

PAGE 135

135 Figure C 5 Thinning down to 0.3 m. The sample was tilted to 7o to perform undercut. The ion beam current was 50 pA and z = 0.1 m. Figure C 6 Undercut. Final thinning down the slab to 100 nm at ion beam current of 10 pA.

PAGE 136

136 Figure C 7 Final thinning down to 100 nm. I t was observed that the sample slab crumbled/curved at either step 6 or 7, this always happened when BiI3 was thinning down. It is possibly due to the v ulnerability of BiI3 to the ion beams. C.3 Summary and Conclusions The TEM sample preparation for BiI3 was not successful The sample was always destroyed at the last several steps due to the highly ion beam sensitivity of BiI3.

PAGE 137

137 APPENDIX D NKN TAPE FOR PIEZOEL ECTRIC TRANSFORMER D.1 Introduction As a side project from BiI3 single crystal growth, initial efforts have been made to develop a piezoelectric transformer This appendix discusses the experiment and initial results of synthesizing sodium potassium niobate (Na0.5K0.5NbO5, NKN) powder and ceramic tape for the application of piezoelectric transformer. A prototype of transformer made from the NKN slab has been tested. D.2 Experimental Procedures Sodium potassium niobate (Na0.5K0.5NbO5, NKN) powder was synthesized by conventional solid state method. The starting materials were reagent grade K2CO3 (99%, Alfa Aesar), Na2CO3 (99.5%, Alfa Aesar) and Nb2O5 (99.5%, Alfa Aesar). The raw powders were dried at 120oC for 24h to remove the moisture then weighed according to the stoichiometr y. The powders were mixed and ball milled for 24h in anhydrous ethanol The milled slurry was dried at 120oC overnight then ground by mortar and pestle followed by sieving through a 212 m mesh. The powder was calcined at 850oC for 10h. The calcined powder was ground and sieved again then uniaxially pressed to a rectangular shape slab with dimension of 19 x 6 x 3 mm3. The slab was sintered at 1100oC for 2h. The sintered slab was polish ed and applied Au electrode by sputter coating. The prototype transformer adopted a Rosen type configuration, as illustrated in Figure D 1 The slab was first poled along the longitud inal direction then transverse direction, at 150oC under electric field of 1.7 kV/mm for 30 min. To test the performance of the transformer, it was connected to an Agilent 33220A wave generator at the input side

PAGE 138

138 and a multimeter at the output side. The wave generator can provide an alternative signal with frequency from 1 Hz to 20 MHz and amplitude up to 10 V. Figure D 1 Rosen type configuration of prototype NKN transformer Attempts have been made to produce NKN thin slides using tape casting technique. The tape casting slurry was made ac cording to the following procedures The calcined NKN powder obtained from solid state process was dried at 120oC for 24h. Fish oil (2 wt%) was weighed and dissolve d in the toluene (12 wt%), and the mixture was added to the ball mill. Anhydrous ethanol (12 wt%) was weighed and added to the ball mill. The dried NKN powder (59 wt%) was weighed and added to the mill. The mixture was ball milled for 24h. Butylbenzyl phthalate (5 wt%), polyvinyl butyral (5 wt%), and polyethylene glycol (5 wt%) were weighed and added to the ball mill. The final mixture was ball milled for an additional 24h. The slurry was poured to a beaker and deaired at 0.8 atm vacuum for 10 min. The materials and their compositions are listed in Table D 1 Table D -1. Batch formulation of tape casting slurry of NKN Component Composition (wt%) Function NKN powder 59 Menhaden fish oil, blown Z 3 2 dispersant Toluene 12 Solvent Anhydrous ethanol 12 Solvent Butylbenzyl phthalate, S 160 5 P lasticizer Polyvinyl butyral, B 98 5 Binder Polyethylene glycol, 8000 5 Plasticizer, binder V in V out t 2L

PAGE 139

139 The slurry was tape-casted on Mylar sheet using a doctor blade. The tape was covered and dried in air over night. To sinter the tape, it was heated to 600oC with a ramp rate of 0.3oC/min then hold for 6h During this stage, the polymer component in the tape was burned out. The slow ramp rate will prevent the formation of crack due to the removal of water and polymer component. After burning out, it was heated to 1100oC with a ramp rate of 4oC/min then hold for 3h. The sintered tape was cooled to room temperature with a ramp rate of 4oC/min. D.3 Results and Discussion D.3.1 X-ray Diffraction The XRD pattern of calcined and sintered NKN is shown in Figure D 2 which indicates that phase pure NKN was form ed after calcination and the sintered slab was phase pure also. The sintered NKN slab has a dimension of 16.34 x 5.38 x 2.68 mm3 with density of 4.26 g/cm3, which is 95.4% of the theoretical density of 4.46 g/cm3. 20 30 40 50 60 slab sintered at 1100 oC, 2h Othorhombic KNbO3 32-0822 powder calcined at 850 oC, 10h 022 311 131 112 221 130 002 220 021 001 110 Intensity2 (degrees)111 200 020 Figure D 2 XRD pattern of calcined and sintered NKN.

PAGE 140

140 D.3.2 Transformer Test The equivalent circuit for a Rosen type piezoelectric transformer is shown in Figure D -3 .223 The resonant frequency is strongly dependent on the load. The output to input voltage ratio is also a complicated function and dependent on the load. For a very large load resistance (open circuit), this ratio is approximately proportional to the mechanical quality factor Qm and electromechanical coupling factors :224 t L Q k k V Vm in out/ /33 31 (D -1) where k31 and k33 are the transverse and longitudinal electromechanical coupling factor s, L and t are the length and thickness of the transformer, respectively The NKN transformer tested in this study has a dimension of 14.01 x 4.72 x 2.12 mm3. The operating frequency was calculated to be F = c/4L = 12153 Hz according to Fuji & Co. Piezo Science, where c is velocity of sound. The input voltage was 10 V peak to peak at 12153 Hz. The output voltage was measured with or without load. The tested loads have resistance of 100 1.5 k and 5.6 M At 12153 Hz operating frequency, 5 V out put voltage was measured without load. The output voltage with loads was measured to be 5 V, 3.2 V and 0 with corresponding load of 5.6 M 1.5 k and 100 The operating frequency was also changed from 1 kHz to 30 0 kHz with step of 10 kHz. The output voltage kept at 5 V without load. Clearly there is problem with our prototype transformer since the output voltage is not amplified at all. Possible reason is that the working frequency is not even close to the resonant frequency, which was not calculated in this study due to lack of the information of the piezoelectric properties. Another possible reason is that the material was not completely poled since the electric field for poling NKN is normally reported to be 3 kV/mm and above,224226

PAGE 141

141 while in this study 1.7 kV/mm was used to pole the material due to the limitation of the equipment. Figure D 3 Equivalent circuits of a piezoelectric transformer: (a) general mode l, (b) after reflecting the output capacitance and load resistance to the primary, and (c) after parallel to series transformation. (After Ivensky et al .. 223)

PAGE 142

142 D.3.3 Tape Casting The casted tape has a thickness of around 0.2 mm. It has been successfully sintered without crack and bending. Figure D 4 shows a pre-sintered tape which is casted on a Mylar sheet. Figure D 4 NKN tape casted on a Myl ar sheet. D.4 Summary and Conclusions Phase pure NKN powder has been successfully prepared by solid state processing. The calcinations was conducted at 850oC for 10 h. NKN slab was sintered at 1100oC for 2 h and determined to be phase pure by XRD. The initial test on the prototype piezoelectric transformer was not successful It is probably due to the wrong working frequency or the poor poling. NKN tape with thickness of around 0.2 mm has been successfully prepared and sintered at a very careful temperature program.

PAGE 143

143 LIST OF REFERENCES 1. W.J. Price, "Nuclear Radiation Detection", 2nd ed., McGraw -Hill, New York, 1964. 2. D.S. McGregor and H. Hermon, "Room -temperature Compound Semiconductor Radiation Detectors," Nucl. Instrum. Meth ods Phys. Res., Sect. A 395 [1] 101-124 (1997). 3. T.E. Schlesinger, J.E. Toney, H. Yoon, E.Y. Lee, B.A. Brunett, L. Franks, and R.B. James, "Cadmium Zinc Telluride and Its Use as a Nuclear Radiation Detector Material," Mater. Sci. Eng., R 32 [4 5] 1031 89 (2001). 4. F.E. Rosztocz and D. Cubiccio, "The Bismuth Iodide-Iodine Phase Diagram," J. Phys. Chem. 69 [1] 124 -127 (1965). 5. D. Cubicciotti and F.J. Keneshea, "The Vapor Pressures of BiI3 over Liquid Bi -BiI3 Solutions," J. Phys. Chem. 63 [2] 295297 (1959). 6. J.H. Kim and S. Blairs, "Sublimation Study of BiI3," J. Chem. Thermodyn. 22 [8] 803-814 (1990). 7. B.J. Curtis and H.R. Brunner, "Crystal -Growth of Bismuth Iodide," Mater. Res. Bull. 9 [5] 715720 (1974). 8. M. Schieber, T.J. Davies, W. Schnepple, P.T. Randtke, and R.C. Carlston, "Crystal -Growth and Electrooptical Characterization of Bismuth Tri -Iodide," J. Appl. Phys. 45 [12] 53715372 (1974). 9. A. Cuna, A. Noguera, E. Saucedo, and L. Fornaro, "Growth of Bismuth Tri -iodide Platelets by the P hysical Vapor Deposition Method," Cryst. Res. Technol. 39 [10] 912-919 (2004). 10. D. Nason and L. Keller, "The Growth and Crystallography of Bismuth Tri -Iodide Crystals Grown by Vapor Transport," J. Cryst. Growth 156 [3] 221-226 (1995). 11. Y.N. Dmitrie v, P.R. Bennett, L.J. Cirignano, M. Klugerman, and K.S. Shah, "Bismuth Iodide Crystals as a Detector Material: Some Optical and Electrical Properties," Proc. SPIEInt. Soc. Opt. Eng., 3768 521529 (1999). 12. M. Matsumoto, K. Hitomi, T. Shoji, and Y. Hirat ate, "Bismuth Tri iodide Crystal for Nuclear Radiation Detectors," IEEE Trans. Nucl. Sci. 49 [5] 25172520 (2002). 13. L. Fornaro, A. Cuna, A. Noguera, M. Perez, and L. Mussio, "Growth of Bismuth Tri -iodine Platelets for Room Temperature X -ray Detection," IEEE Trans. Nucl. Sci. 51 [5] 2461-2465 (2004). 14. G.F. Knoll, "Radiation Detection and Measurement", Third Edition, John Wiley & Sons, Inc., New York, 1999.

PAGE 144

144 15. J.T. Bushberg, J.A. Seibert, E.M.J. Leidholdt, and J.M. Boone, "The Essential Physics of Me dical Imaging", Edwards Brothers, Philadelphia, 2002. 16. S.N. Ahmed, "Physics and Engineering of Radiation Detection ", Academic Press, San Diego, 2007. 17. S.J. Hastings and L. Marcia, "Passive Nondestructive Assay of Nuclear Materials", Edited by R. Doug, E. Norber, S.J. Hastings, and S. Kreiner, U.S. Government Printing Office, Washington, 1991. 18. K. Maher, "Basic Physics of Nuclear Medicine", Wikibooks Contributors, 2004. 19. C.E. Moss, E.J. Dowdy, and M.C. Lucas, "Bismuth -Germanate Scintillators Applications in Nuclear Safeguards and HealthPhysics," Nucl. Instrum. Methods Phys. Res., Sect. A 242 [3] 480 -483 (1986). 20. W. Shockley, "Currents to Conductors Induced by a Moving Point Charge," J. Appl. Phys. 9 [10] 635636 (1938). 21. S. Ramo, "Curr ents Induced by Electron Motion," Proc. Inst. Rad. Engi., 27 584585 (1939). 22. K. Hecht, "Primiirsnotnes in Isolierenden Kristallen," Zeitschrift fur Physik (Berlin) 77 235245 (1932). 23. N.B. Strokan, V.K. Yeryomin, S.A. Lomashev, and N.I. Tisnek, "Fi nal Results on Nature of Pulse-Height Spectrum from Semiconductor Detectors," IEEE Trans. Nucl. Sci. Ns19 [3] 365379 (1972). 24. E. Sakai, "Charge Collection in Coaxial Ge(Li) Detectors," IEEE Trans. Nucl. Sci. Ns15 [3] 310-320 (1968). 25. A.J. Tavendal e, "Semiconductor Lithium -Ion Drift Diodes as High-Resolution Gamma -Ray Pair Spectrometers," IEEE Trans. Nucl. Sci. Ns11 [3] 191200 (1964). 26. P.J. Van Heerden, "The Crystal Counter", Dissertation, Utrecht University, 1945. 27. V.M. Zaletin, "Developmen t of Semiconductor Detectors Based on Wide gap Materials," Atomic Energy 97 [5] 773780 (2004). 28. A. Owens and A. Peacock, "Compound Semiconductor Radiation Detectors," Nucl. Instrum. Methods Phys. Res., Sect. A 531 [1 2] 18 -37 (2004). 29. W.R. Harding C. Hilsum, M.E. Moncaster, D.C. Northrop, and O. Simpson, "Gallium Arsenide for Gamma-Ray Spectroscopy," Nature 187 [4735] 405405 (1960).

PAGE 145

145 30. J.E. Eberhardt, R.D. Ryan, and A.J. Tavendal, "Evaluation of Epitaxial N -GaAs for Nuclear Radiation Detection, Nucl. Instrum. Methods 94 [3] 463-476 (1971). 31. T. Kobayash and T. Sugita, "Gallium -Arsenide Surface Barrier Detectors," Nucl. Instrum. Methods 98 [1] 179180 (1972). 32. S.E. King, H.B. Dietrich, R.L. Henry, D.S. Katzer, W.J. Moore, G.W. Phillips, a nd R.C. Mania, "Development and Characterization of Zone Melt Growth GaAs for Gamma -ray Detectors," IEEE Trans. Nucl. Sci. 43 [3] 13761380 (1996). 33. B. Birkmann, M. Rasp, J. Stenzenberger, and G. Muller, "Growth of 3 '' and 4 '' Gallium Arsenide Crysta ls by the Vertical Gradient Freeze (VGF) Method," J. Cryst. Growth 211 [1 4] 157-162 (2000). 34. D.S. Mcgregor, R.A. Rojeski, G.F. Knoll, F.L. Terry, J. East, and Y. Eisen, "Present Status of Undoped Semiinsulating LEC Bulk GaAs as a Radiation Spectromete r," Nucl. Instrum. Methods Phys. Res., Sect. A 343 [2 3] 527538 (1994). 35. W. Akutagaw, K. Zanio, and J.W. Mayer, "CdTe as a Gamma Detector," Nucl. Instrum. Methods 55 [2] 383385 (1967). 36. J.W. Mayer, "Evalutation of CdTe by Nuclear Particle Measurements," J. Appl. Phys. 38 [1] 296301 (1967). 37. A. Cornet, P. Siffert, A. Coche, and Triboule.R, "Cadmium Telluride Surface Barrier Detectors," Appl. Phys. Lett. 17 [10] 432436 (1970). 38. M. Hage -Ali and P. Siffert, "Semiconductors for Room Temperature Nuclear Detector Applications: Semiconductors and Semimetals", Academic Press, San Diego, 1995. 39. M. Cuzin, "Some New Developments in the Field of High Atomic -Number Materials," Nucl. Instrum. Methods Phys. Res., Sect. A 253 [3] 407417 (1987). 40. W .M. Chang, W.R. Wilcox, and L. Regel, "Floating-Zone Melting of CdTe," Mater. Sci. Eng., B 16 [1 -3] 23 28 (1993). 41. K. Zanio, Krajenbr.F, and H. Montano, "CdTe Detectors from Indium Doped Tellurium -Rich Solutions," IEEE Trans. Nucl. Sci. Ns21 [1] 315 -3 21 (1974). 42. E. Raiskin and J.F. Butler, "CdTe Low Level Gamma-Detectors Based on a New Crystal -Growth Method," IEEE Trans. Nucl. Sci. 35 [1] 81 84 (1988). 43. P. Cheuvart, U. Elhanani, D. Schneider, and R. Triboulet, "CdTe and CdZnTe Crystal -Growth by Horizontal Bridgman Technique," J. Cryst. Growth 101 [1 4] 270-274 (1990).

PAGE 146

146 44. R.O. Bell and F.V. Wald, "Recent Advances in Preparation of CdTe for Nuclear Detectors," IEEE Trans. Nucl. Sci. Ns19 [3] 334338 (1972). 45. F.V. Wald and G. Entine, "Crystal -Growth of CdTe for Gamma-Ray Detectors," Nucl. Instrum. Methods 150 [1] 13 -23 (1978). 46. S. Watanabe, T. Takahashi, Y. Okada, G. Sato, M. Kouda, T. Mitani, Y. Kobayashi, K. Nakazawa, Y. Kuroda, and M. Onishi, "CdTe Stacked Detectors for Gamma -ray Detecti on," IEEE Trans. Nucl. Sci., 49 [3] 12921296 (2002). 47. E. Sakai, Y. Murakami, and H. Nakatani, "Performance of a High-Purity Ge Gamma -Ray Spectrometer System Using a Closed-Cycle Cryogenic Refrigerator," IEEE Trans. Nucl. Sci. 29 [1] 760 -763 (1982). 48. P. Siffert, J. Berger, C. Scharager, A. Cornet, R. Stuck, R.O. Bell, H.B. Serreze, and F.V. Wald, "Polarization in Cadmium Telluride Nuclear Radiation Detectors," IEEE Trans. Nucl. Sci. 23 [1] 159 -170 (1976). 49. J.F. Butler, C.L. Lingren, and F.P. Doty "Cd1 XZnxTe Gamma -Ray Detectors," IEEE Trans. Nucl. Sci. 39 [4] 605 -609 (1992). 50. D.J. Olego, J.P. Faurie, S. Sivananthan, and P.M. Raccah, "Optoelectronic Properties of Cd1 XZnxTe Films Grown by Molecular -Beam Epitaxy on GaAs Substrates," Appl. Phys. Lett. 47 [11] 11721174 (1985). 51. Y. Eisen, "Current Stateof -the art Industrial and Research Applications Using Room -temperature CdTe and CdZnTe Solid State Detectors," Nucl. Instrum. Methods Phys. Res., Sect. A 380 [1 2] 431439 (1996). 52. Z. Bursh tein, H.N. Jayatirtha, A. Burger, J.F. Butler, B. Apotovsky, and F.P. Doty, "Charge -Carrier Mobilities in Cd0.8Zn0.2Te Single -Crystals Used as Nuclear Radiation Detectors," Appl. Phys. Lett. 63 [1] 102 -104 (1993). 53. W. Akutagaw and K. Zanio, "Gamma Resp onse of Semi -Insulating Material in Presence of Trapping and Detrapping," J. Appl. Phys. 40 [9] 38383854 (1969). 54. J.F. Butler, F.P. Doty, B. Apotovsky, J. Lajzerowicz, and L. Verger, "Gamma-Ray and X -Ray Detectors Manufactured from Cd1 XZnx Te Grown b y a High-Pressure Bridgman Method," Mater. Sci. Eng., B 16 [1 3] 291-295 (1993). 55. F.P. Doty, J.F. Butler, J.F. Schetzina, and K.A. Bowers, "Properties of CdZnTe Crystals Grown by a High-Pressure Bridgman Method," J. Vac. Sci. Technol., B 10 [4] 1418-1 422 (1992). 56. Y. Eisen, A. Shor, and I. Mardor, "CdTe and CdZnTe Gamma Ray Detectors for Medical and Industrial Imaging Systems," Nucl. Instrum. Methods Phys. Res., Sect. A 428 [1] 158-170 (1999).

PAGE 147

147 57. Y. Ramachers and D.Y. Stewart, "Energy resolution im provement in room temperature CZT detectors," J Instrum 2 (2007). 58. S.P. Swierkow, G.A. Armantro, and R. Wichner, "Recent Advances with HgI2 X Ray Detectors," IEEE Trans. Nucl. Sci. Ns21 [1] 302 -304 (1974). 59. J. Llacer, M.M.K. Watt, M. Schieber, R. Carlston, and Schneppl.W, "Preliminary Studies of Charge Carrier Transport in Mercuric Iodide Radiation Detectors," IEEE Trans. Nucl. Sci. Ns21 [1] 305-314 (1974). 60. J.P. Ponpon, R. Stuck, P. Siffert, and C. Schwab, "Preliminary Results on Mercuric Iod ide Nuclear Radiation Detectors," Nucl. Instrum. Methods 119 [1] 197-198 (1974). 61. H. Hermon, M. Roth, M. Schieber, and J. Shamir, "On the Phase-Diagram of Mercuric Iodide near the Stoichiometric Composition," Mater. Res. Bull. 28 [3] 229-234 (1993). 6 2. I.F. Nicolau and J.P. Joly, "Solution Growth of Sparingly Soluble Single -Crystals from Soluble Complexes .3. Growth of Alpha-HgI2 Single -Crystals from Dimethylsulfoxide Complexes," J. Cryst. Growth 48 [1] 61 -73 (1980). 63. M. Schieber, W.F. Schnepple, and L. Vandenberg, "Vapor Growth of HgI2 by Periodic Source or Crystal Temperature Oscillation," J. Cryst. Growth 33 [1] 125135 (1976). 64. B. Coupat, J.P. Fournier, M. Silga, and J. Omaly, "X and Gamma-Detection with HgI2 Single -Crystals Obtained by the Forced Flux Method," Nucl. Instrum. Methods Phys. Res., Sect. A 269 [1] 213 -218 (1988). 65. Y.F. Nicolau, "Technology of High -Purity HgI2," Nucl. Instrum. Methods Phys. Res., Sect. A 213 [1] 1318 (1983). 66. J.E. Baciak,Ph.D Thesis, University of Michi gan, 2004. 67. R.S. Scott and G.E. Fredericks, "Evaporative Degradation of HgI2 X Ray Detectors," Appl. Phys. Lett. 27 [2] 99 100 (1975). 68. J.S. Iwanczyk, Y.J. Wang, J.G. Bradley, A.L. Albee, and W.F. Schnepple, "Advances in the Development of Encapsulants for Mercuric Iodide X -Ray Detectors," IEEE Trans. Nucl. Sci. 37 [6] 2214 2218 (1990). 69. S. Roth and W.R. Willig, "Lead Iodide Nuclear Particle Detectors," Appl. Phys. Lett. 18 [8] 328 -330 (1971). 70. J.C. Lund, K.S. Shah, M.R. Squillante, L.P. Moy, F. Sinclair, and G. Entine, "Properties of Lead Iodide Semiconductor Radiation Detectors," Nucl. Instrum. Methods Phys. Res., Sect. A 283 [2] 299 -302 (1989).

PAGE 148

148 71. V. Deich and M. Roth, "Improved Performance of Lead Iodide Nuclear Radiation Detectors," Nuc l. Instrum. Methods Phys. Res., Sect. A 380 [1 2] 169-172 (1996). 72. R.J.M. Konings, E.H.P. Cordfunke, and R.R. Vanderlaan, "Enthalpy Increment Measurements of PbI2 Evidence for a Reversible Polytypic Transition," J. Alloys Compd. 230 [2] 85 -88 (1995) 73. C. Manfredotti, R. Murri, A. Quirini, and L. Vasanelli, "PbI2 as Nuclear -Particle Detector," IEEE Trans. Nucl. Sci., 24 [1] 126-128 (1977). 74. M. Matuchova, K. Zdansky, M. Svatuska, J. Zavadil, and O. Prochazkova, "Electrical Resistivity and Photolu minescence of Lead Iodide Crystals," Chem. Pap. 61 [1] 36-40 (2007). 75. X.H. Zhu, Z.R. Wei, Y.R. Jin, and A.P. Xiang, "Growth and Characterization of a PbI2 Single Crystal Used for Gamma Ray Detectors," Cryst. Res. Technol. 42 [5] 456-459 (2007). 76. T. Shoji, K. Hitomi, T. Tiba, T. Suehiro, and Y. Hiratate, "Fabrication of a Nuclear Radiation Detector Using the PbI2 Crystal and It Rsponse Caracteristics for Gmma -rays," IEEE Trans. Nucl. Sci. 45 [3] 581-584 (1998). 77. L. Fornaro, E. Saucedo, L. Mussio, A. Gancharov, F. Guimaraes, and A. Hernandes, "Lead Iodide Platelets: Correlation between Surface, Optical, and Electrical Properties with X and Gamma -ray Spectrometric Performance," IEEE Trans. Nucl. Sci. 49 [6] 3300-3305 (2002). 78. V. Deich and M. Roth, "Improved performance lead iodide nuclear radiation detectors," Nucl Instrum Meth A 380 [1 -2] 169172 (1996). 79. S.M. Sze, "Physics of Semiconductor Devices", Wiley, New York, 1981. 80. R. Coquille, Y. Toudic, M. Gauneau, G. Grandpierre, and J.C. Par is, "Synthesis, Crystal -Growth and Characterization of InP," J. Cryst. Growth 64 [1] 2331 (1983). 81. A. Shimizu, J. Nishizawa, Y. Oyama, and K. Suto, "InP Single Crystal Growth by the Horizontal Bridgman Method Under Controlled Phosphorus Vapor Pressure," J. Cryst. Growth 229 [1] 119123 (2001). 82. P. Rudolph, F. Matsumoto, and T. Fukuda, "Studies on Interface Curvature during Vertical Bridgman Growth of InP in a Flat -bottom Container," J. Cryst. Growth 158 [1 -2] 43 48 (1996). 83. M. Young, X. Liu, D. Zhang, M. Zhu, and X.Y. Hu, "Latest Developments in Vertical Gradient Freeze (VGF) Technology: GaAs, InP, and GaP," Mater. Sci. Eng., B 66 [1 3] 1-6 (1999).

PAGE 149

149 84. A. Yamamoto and C. Uemura, "InP Single-Crystal Growth by Synthesis, Solute Diffusion Method," Jpn. J. Appl. Phys. 17 [10] 18691870 (1978). 85. I. Rahman, W.A. Fisher, R. Hofstadter, and J. Shen, "Behavior of Thallium Bromide Conduction Counters," Nucl. Instrum. Methods Phys. Res., Sect. A 261 [3] 427439 (1987). 86. K.S. Shah, J.C. Lund, F. Ols chner, L. Moy, and M.R. Squillante, "Thallium Bromide Radiation Detectors," IEEE Trans. Nucl. Sci. 36 [1] 199 -201 (1989). 87. I.B. Oliveira, F.E. Costa, J.F.D. Chubaci, and M.M. Hamada, "Purification and Preparation of TlBr Crystals for Room Temperature R adiation Detector Applications," IEEE Trans. Nucl. Sci. 51 [3] 1224 1228 (2004). 88. M.S. Kouznetsov, I.S. Lisitsky, S.I. Zatoloka, and V.V. Gostilo, "Development of the Technology for Growing TlBr Detector Crystals," Nucl. Instrum. Methods Phys. Res., Se ct. A 531 [1 -2] 174-180 (2004). 89. V. Kozlov, H. Andersson, V. Gostilo, M. Kemell, P. Kostamo, M.S. Kouznetsov, M. Leskela, H. Lipsanen, I.S. Lisitsky, M. Shorohov, and H. Sipila, "Improvements and Problems of Bridgman -Stockbarger Method for Fabrication of TlBr Single Crystal Detectors, Nucl. Instrum. Methods Phys. Res., Sect. A 607 [1] 126128 (2009). 90. G.A. Armantrout, S.P. Swierkowski, J.W. Sherohman, and J.H. Yee, "What Can Be Expected from High-Z Semiconductor -Detectors," IEEE Trans. Nucl. Sci. 24 [1] 121125 (1977). 91. M. Rot h, "Advantages and Limitations of Cadmium Selenide Room -Temperature Gamma -Ray Detectors," Nucl. Instrum. Methods Phys. Res., Sect. A 283 [2] 291298 (1989). 92. H. Nakatani, E. Sakai, C. Tatsuyama, and F. Takeda, "GaSe Nuclear -Particle Detectors," Nucl. I nstrum. Methods Phys. Res., Sect. A 283 [2] 303309 (1989). 93. Y. Kumashiro, "Refractory Semiconductor of Boron Phosphide," J. Mater. Res. 5 [12] 2933 2947 (1990). 94. J.H. Yee, S.P. Swierkowski, and J.W. Sherohman, "AlSb as a HighEnergy Photon Detecto r," IEEE Trans. Nucl. Sci., 24 [4] 19621967 (1977). 95. R.T. Johnson, "Cadmium Sulfide Neutron Detectors Annealing of Radiation Damage," IEEE Trans. Nucl. Sci., Ns17 [4] 14 -19 (1970). 96. T.M. Bochkova and A.V. Kovalenko, "Detector of High -Energy Radiat ion on the Structure Bi4Ge3O12/ZnSexTe1 X/ZnSe," Nucl. Tracks Radiat. Meas. 21 [1] 191194 (1993).

PAGE 150

150 97. P.C. Canepa, R.B. Campbell, Malinari.P, and J. Ostroski, "High Temperature Nuclear Particle Detector," IEEE Trans. Nucl. Sci. Ns11 [3] 262-270 (1964). 98. J. Trotter and T. Zobel, "Crystal Structure of SbI3 and BiI3," Zeitschrift Fur Kristallographie Kristallgeometrie Kristallphysik Kristallchemie 123 [1] 67 -72 (1966). 99. I. Akai, T. Karasawa, T. Komatsu, S. Iwai, K. Edamatsu, T. Itoh, S. Yano, and T. Goto, "Nonlinear -Optical Spectra of the Confined Excitons in Stress -Induced Nanostructures in BiI3," Jpn. J. Appl. Phys., Part 2 34 113115 (1994). 100. H. Kondo, T. Kawai, and T. Karasawa, "Nonlinear -Optical Responses due to the Lateral Transport of Exci tons in a Quasi 2 -Dimensional System in BiI3 Crystals," Solid State Commun. 94 [2] 129-134 (1995). 101. T. Mishina, H. Chida, and Y. Masumoto, "Nonlinear Luminescence of Stacking Fault Excitons in BiI3 Induced by the Exciton -Exciton Auger Process," Phys. Rev. B, 48 [3] 1460-1463 (1993). 102. V.B. Timofeev and V.I. Vashchen, "Optical Properties of BiI3 Single Crystals near Long-Wavelength Edge of Fundamental Absorption," Opt. Spectrosc. 24 [5] 396 (1968). 103. R.I. Shekhmam, "Indirect Transitions in Spectr a of BiI3 Crystals at Low Temperatures," Phys. Status Solidi B 66 [1] 371375 (1974). 104. Y. Kaifu and T. Komatsu, "Optical Properties of Bismuth Tri -Iodide Single -Crystals .2. Intrinsic Absorption-Edge," J. Phys. Soc. Jpn. 40 [5] 1377 1382 (1976). 105. T. Komatsu and Y. Kaifu, "Optical Properties of Bismuth Tri -Iodide Single -Crystals .1. Interband-Transitions," J. Phys. Soc. Jpn. 40 [4] 1062-1068 (1976). 106. T. Komatsu, T. Karasawa, T. Iida, K. Miyata, and Y. Kaifu, "Secondary -Emission of Indirect Exc iton in BiI3," J. Lumin. 24 -5 [Nov] 679682 (1981). 107. T. Karasawa, T. Komatsu, and Y. Kaifu, "Zone-Folding Effects on Phonons and Excitons in Polytypic BiI3 Single -Crystals," Solid State Commun. 44 [3] 323-327 (1982). 108. T. Karasawa, T. Iida, M. Sak ai, T. Komatsu, and Y. Kaifu, "Secondary -Emission Due to Recombination of Excitons near Indirect Band Bottom .2. Experimental and Theoretical -Analysis in BiI3," J. Phys. C: Solid State Phys. 18 [20] 40434059 (1985). 109. K. Watanabe, T. Karasawa, T. Komatsu, and Y. Kaifu, "Optical Properties of Extrinsic Two -Dimensional Excitons in BiI3 Single -Crystals," J. Phys. Soc. Jpn. 55 [3] 897907 (1986).

PAGE 151

15 1 110. Y. Kaifu, "Excitons in Layered BiI3 Single Crystals," J. Lumin. 42 [2] 6181 (1988). 111. H. Kurisu, T. Komatsu, and T. Karasawa, "Hydrostatic -Pressure Effects on the Excitons and ExcitonPhonon Interaction in BiI3 Crystals," J. Phys. Soc. Jpn. 62 [3] 1048 -1056 (1993). 112. T. Karasawa, T. Kawai, H. Mino, I. Akai, and T. Komatsu, "Coherent and Incoherent Tr ansport Properties of Excitons at High Density in a Quasi -two dimensional Space of a Stacking Disorder Interface in a Layered Crystal BiI3," J. Lumin. 72 -4 876878 (1997). 113. A. Saitoh, "Raman Scattering, Luminescence, and Absorption Edge under Hydrostatic Pressures of Layered BiI3 and SbI3," J. Raman Spectrosc. 38 [5] 537542 (2007). 114. T. Komatsu, Y. Kaifu, S. Takeyama, and N. Miura, "Effects of High Magnetic Fields on Cationic Exciton Lines in BiI3," Phys. Rev. Lett. 58 [21] 2259 -2262 (1987). 115. E.F. Gross, N.V. Starosti, M.P. Shepilov, and R.I. Shekhmam, "Spectroscopy of Energy -Levels of Bielectron or Bichol in Bismuth Iodide Crystals," Izvestiya Akademii Nauk Sssr Seriya Fizicheskaya, 37 [4] 885 -890 (1973). 116. W. Czaja, G. Harbeke, Krausbau.L, E. Meier, B.J. Curtis, H. Brunner, and E. Tosatti, "Existence of Bielectrons in BiI3," Solid State Commun. 13 [9] 1445 -1450 (1973). 117. Y. Petroff, P.Y. Yu, and Y.R. Shen, "Absorption, Photoluminescence, and Resonant RamanScattering in BiI3," Phys. St atus Solidi B 61 [2] 419 -427 (1974). 118. M.P. Lisitsa, L.F. Gudymenk, F.V. Motsnyi, and D.I. Beletska, "Low -Temperature Absorption and Reflection Spectra of BiI3 Single -Crystals," Fizika Tverdogo Tela, 16 [7] 1965-1973 (1974). 119. A.M. Vainrub, A.V. Ili nskii, and B.V. Novikov, "Low -Temperature Spectra of Photoconductivity and Absorption in BiI3 Single -Crystals," Fizika Tverdogo Tela, 15 [2] 490492 (1973). 120. T. Kawai, I. Akai, and T. Karasawa, "Space -Resolved Spectra Due to Highly Mobile Exciton at th e Stacking -Fault in BiI3 Crystals," J. Phys. Soc. Jpn. 58 [3] 969-977 (1989). 121. J.A. Dantzig and M. Rappaz, "Solidification", CRC Press, Lausanne, 2009. 122. W.K. Burton, N. Cabrera, and F.C. Frank, "The Growth of Crystals and the Equilibrium Structure of Their Surfaces," Philos Tr R Soc S -A, 243 [866] 299358 (1951).

PAGE 152

152 123. K.A. Jackson, "On Surface Roughness," Acta Met. 7 [11] 747748 (1959). 124. J.C. Brice, "Crystal Growth Processes", p. 150, Blackie & Son Ltd., Glasgow and London, 1986. 125. K.A. Ja ckson, "Liquid Metals and Solidification", American Society for Metals, Cleveland, 1958. 126. I. Sunagawa, "Crystals: Growth, Morphology, and Perfection", Cambridge University Press, Cambridge, UK, 2005. 127. A.T. Lintereur, W. Qiu, J.C. Nino, and J.E. Bac iak, "Bismuth Tri -iodide Radiation Detector Development," Proceedings of the SPIE on Hard X Ray, Gamma -Ray, and Neutron Detector Physics 7449 74491M 74491M7 (2009). 128. L. Fornaro, E. Saucedo, L. Mussio, A. Gancharov, and A. Cuna, "Bismuth Tri iodide Polycrystalline Films for Digital X -ray Radiography Applications," IEEE Trans. Nucl. Sci. 51 [1] 96-100 (2004). 129. M. Piechotka, "Mercuric Iodide for Room Temperature Radiation Detectors. Synthesis, Purification, Crystal Growth and Defect Formation," Mater Sci. Eng., R 18 [1 2] 1-98 (1997). 130. L. Fornaro, I. Aguiar, A. Noguer, M. Perez, and M. Rodriguez, "Improvements on Bismuth Tri -iodide Platelets For Room Temperature X -ray Detection," IEEE Nuclear Science Symposium R04 -3 [2] 3616 3621 (2006). 131. P.W. Bridgman, "Certain Physical Properties of Single Crystals of Tungsten, Antimony, Bismuth, Tellurium, Cadmium, Zinc, and Tin.," Proc. Am. Aca. Arts Sci. 60 [1/14] 305-383 (1925). 132. F.M. Carlson, A.L. Fripp, and R.K. Crouch, "Thermal -Convection during Bridgman Crystal -Growth," J. Cryst. Growth 68 [3] 747756 (1984). 133. M.C. Flemings, "Solidification Processing", McGraw -Hill, Inc., New York, 1974. 134. P.C. Sukanek, "Deviation of Freezing Rate from Translation Rate in the BridgmanStockbarger Techni que .1. Very Low Translation Rates," J. Cryst. Growth 58 [1] 208-218 (1982). 135. P.C. Sukanek, "Deviation of Freezing Rate from Translation Rate in the BridgmanStockbarger Technique .2. Moderate Translation Rates," J. Cryst. Growth 58 [1] 219-228 (1982 ). 136. L.Y. Chin and F.M. Carlson, "Finite-Element Analysis of the Control of Interface Shape in Bridgman Crystal -Growth," J. Cryst. Growth 62 [3] 561567 (1983).

PAGE 153

153 137. S.H. Hahn and J.K. Yoon, "Numerical Analysis for the Application of Radiative Reflectors to Vertical Bridgman Growth Configurations," J. Cryst. Growth 177 [3 4] 296-302 (1997). 138. S. Kuppurao and J.J. Derby, "Designing Thermal Environments to Promote Convex Interface Shapes during the Vertical Bridgman Growth of Cadmium Zinc Telluride," J. Cryst. Growth 172 [3 4] 350360 (1997). 139. O.Y. Hong and S.Y. Wei, "Numerical simulation of CdTe vertical Bridgman growth," J. Cryst. Growth 173 [3 -4] 352366 (1997). 140. A. Bachran, P. Reinshaus, and W. Seifert, "Influence of Thermal Processing Pa rameters and Material Properties on Velocity Configurations in Semiconductor Melts during the Vertical Bridgman Growth Technique," Cryst. Res. Technol. 33 [1] 2736 (1998). 141. V.M. Lakeenkov, V.B. Ufimtsev, N.I. Shmatov, and Y.F. Schelkin, "Numeric simu lation of vertical Bridgman growth of Cd1xZnxTe melts," J. Cryst. Growth 197 [3] 443-448 (1999). 142. H.J. Lee and A.J. Pearlstein, "Simulation of radial solute segregation in vertical Bridgman growth of pyridine doped benzene, a surrogate for binary org anic nonlinear optical materials," J. Cryst. Growth 218 [2 -4] 334352 (2000). 143. F. Rosenberger, "Crystal Growth in Science and Technology", NATO ASI Series Edited by H. Arend and J. Hulliger, Plenum Press, New York, 1989. 144. P. Capper, "Bulk Crystal Growth of Electronic, Optical & Optoelectronic Materials", Materials for Electronic and Optoelectronic Applications Edited by P. Capper, S. Kasap, and A. Willoughby, John Wiley & Sons, Ltd, West Sussex, England, 2005. 145. R. Cadoret, "AlphaMercuric Iod ide Crystal -Growth by Physical Vapor Transport," J. Cryst. Growth 146 [1 4] 9-14 (1995). 146. S.V. Tsivinskii and L.A. Maslova, "Rocking Curve Half Width Obtained by Means of a Double X -Ray Crystal Spectrometer and Dislocation Density in SingleCrystals," Mater Sci Eng 30 [2] 147-153 (1977). 147. C.R. Wie, "Relaxation and Rocking-Curve Broadening of Strained (Ga,In) as Single Layers on (001) GaAs," J. Appl. Phys. 65 [6] 22672271 (1989). 148. S. Nam, Y.M. Yu, J. Rhee, B. O, Y.D. Choi, J. Lee, and Y.J. Ju ng, "Evaluation of Crystal quality of ZnTe/GaAs Epilayers by the Double Crystal Rocking Curve," Mater. Chem. Phys. 69 [1 3] 3035 (2001). 149. M. Ohyama, "Thermal Instability of Growth Interface in Horizontal Solidification of GaAs," J. Phys. Soc. Jpn. 2 9 [3] 706710 (1970).

PAGE 154

154 150. E. Kuroda, H. Kozuka, and Y. Takano, "The Effect of Temperature Oscillations at the Growth Interface on Crystal Perfection," J. Cryst. Growth 68 [2] 613 623 (1984). 151. Y. Wang, K. Kudo, Y. Inatomi, R.B. Ji, and T. Motegi, "Growth Interface of CdZnTe Grown from Te Solution with THM Technique under Static Magnetic Field," J. Cryst. Growth 284 [3 4] 406-411 (2005). 152. J.G. Wood and R.S. Marriner, "The Measurement of Vickers Hardness Indentations," Metallurgia 73 [440] 297-300 (1966). 153. K. Baskar, K. Thangaraj, and R. Gobinathan, "Hardness Anisotropy of Mercuric Iodide Single -Crystals," J. Mater. Sci. Lett., 10 [2] 8586 (1991). 154. W.F. Gale and T.C. Totemeier, "Smithells Metals Reference Book", 8th Edition, Elsevier Butter worth -Heinemann, New York, 2004. 155. E. Springer, "Relation between Dislocation Density and Etch Pit Density in Germanium," Z. Metallkd., 62 [4] 298 -300 (1971). 156. D.J. Stirland, G.J. Rees, and A. Ritson, "The Relationship between Etch Pit Density and D islocation Density for (001)GaAs," J. Cryst. Growth 79 [1 -3] 493 -502 (1986). 157. R. Hirano, "Growth of Low Etch Pit Density Homogeneous 2 '' InP Crystals Using a Newly Developed Thermal Baffle," Jpn. J. Appl. Phys., Part 1 38 [2B] 969971 (1999). 158. D Hull and D.J. Bacon, "Introduction to Dislocations", 3rd Edition, Pergamon Press, New York, 1984. 159. J.P. Ponpon, R. Stuck, P. Siffert, B. Meyer, and C. Schwab, "Properties of Vapor Phase Grown Mercuric Iodide Single-Crystal Detectors," IEEE Trans. Nuc l. Sci. Ns22 [1] 182-191 (1975). 160. P.T. Randtke and C. Ortale, "Effects of Dislocations and Crystal Defects on Energy Resolution and Response Uniformity of Mercuric Iodide Detectors," IEEE Trans. Nucl. Sci. 24 [1] 129134 (1977). 161. L. Fornaro, L. Luchini, M. Koncke, L. Mussio, E. Quagliata, K. Chattopadhyay, and A. Burger, "Growth of Mercuric Iodide Platelets for X -ray Room Temperature Detectors in the HgI2-HI -H2O system," J. Cryst. Growth 217 [3] 263270 (2000). 162. M. Schieber, R.B. James, J.C. Lund, D.S. McGregor, T.S. Gilbert, J.M. VanScyoc, R.W. Olsen, A.E. Pontau, T.S. Schlesinger, and J. Toney, "State of the art of widebandgap semiconductor nuclear radiation detectors," Nuovo Cimento A 109 [9] 12531260 (1996).

PAGE 155

155 163. J.M. Tartaglia, S.M. Cr ochiere, C.E. Kalnas, D.L. Farrington, J.A. Kronwasser, and P.J. Pearah, "A Study of Etch Pit Density and X -Ray Rocking Curves for GaAs Substrate Evaluation," J. Electron. Mater. 20 [5] 345352 (1991). 164. H. Yoon, C. Yoon, C.S. Park, T. Ko, N.S. Kim, an d K.N. Han, "Quantitative Determination of PGM using ICP -MS, ICP -AES, AAS and XRF," Miner. Metall. Process. 22 [1] 59 64 (2005). 165. Rashmi, U. Dhawan, and D.K. Suri, "Evaluation of uncertainty of measurement in quantitative XRF analysis: I. Scattered radiation method," X-Ray Spectrom. 28 [3] 157-162 (1999). 166. Z.H. Li, W.T. Li, J. Liu, B.J. Zhao, S.F. Zhu, H.P. Xu, and H. Yuan, "Improved Technique for Hgi2 Purification," J. Cryst. Growth 156 [1 2] 8690 (1995). 167. J.F. Condeles, R.C.Z. Lofrano, J.M Rosolen, and M. Mulato, "Stoichiometry, Surface and Structural Characterization of Lead Iodide Thin Films," Braz J Phys, 36 [2A] 320 -323 (2006). 168. Vashchen.Vi and V.B. Timofeev, "Absorption Spectra and Indirect Optical Transitions in Bi I3 Single Cryst als," Soviet Physics Solid State,Ussr 9 [6] 1242-& (1967). 169. G.E. Jellison, J.O. Ramey, and A.A. Boatner, "Optical Functions of BiI3 as Measured by Generalized Ellipsometry," Phys. Rev. B 59 [15] 97189721 (1999). 170. M. Schluter, M.L. Cohen, S.E. Kohn, and C.Y. Fong, "Electronic -Structure of BiI3," Phys. Status Solidi B 78 [2] 737747 (1976). 171. H. Yorikawa and S. Muramatsu, "Theoretical study of crystal and electronic structures of BiI3," Journal of Physics-Condensed Matter 20 [32] (2008). 172. H. Yorikawa and S. Muramatsu, "Theoretical Study of Crystal and Electronic Structures of BiI3," J. Phys.: Condens. Matter 20 [32] (2008). 173. Y. Dmitriev, P.R. Bennett, L.J. Cirignano, M. Kiugerman, and K.S. Shah, "Bismuth Iodide Crystals as a Detect or Material: Some Optical and Electrical Properties," SPIE Conference on Hard X -Ray, Gamma -Ray, and Neutron Detector Physics 3768 521529 (1999). 174. P.J. Dean and D.G. Thomas, "Intrinsic Absorption -Edge Spectrum of Gallium Phosphide," Physical Review 1 50 [2] 690-703 (1966). 175. J.I. Pankove, "Optical Processes in Semiconductors", Prentice Hall, Englewood Cliffs, 1971. 176. W.W. Piper, P.D. Johnson, and D.T.F. Marple, "Temperature Dependence of the Optical Band Gap in ZnS," J. Phys. Chem. Solids 8 457461 (1959).

PAGE 156

156 177. G.D. Mahan, "Temperature Dependence of Band Gap in CdTe," Journal of Physics and Chemistry of Solids 26 [4] 751-& (1965). 178. A.G. Dmitriev, "Temperature -Dependence of the Width of the GaAs Band-Gap," Semiconductors 29 [3] 227-229 (1995 ). 179. Y.P. Varshni, "Temperature Dependence of Energy Gap in Semiconductors," Physica 34 [1] 149 -154 (1967). 180. S.R. Dhariwal and V.N. Ojha, "Band -Gap Narrowing in Heavily Doped Silicon," Solid -State Electron. 25 [9] 909 -911 (1982). 181. R.A. Abram, G.N. Childs, and P.A. Saunderson, "Band-Gap Narrowing Due to Many -Body Effects in Silicon and Gallium -Arsenide," J. Phys. C: Solid State Phys. 17 [34] 61056125 (1984). 182. V. Palankovski, G. Kaiblinger -Grujin, and S. Selberherr, "Study of Dopant dependent Band Gap Narrowing in Compound Semiconductor Devices," Mater. Sci. Eng., B 66 [1 -3] 46 49 (1999). 183. W. Zagozdzonwosik and W. Kuzmicz, "Physical Mechanisms of BandgapNarrowing in Silicon," Electron. Lett. 19 [14] 515-516 (1983). 184. R.J. Vanoverst raeten and R.P. Mertens, "Heavy Doping Effects in Silicon," Solid State Electron. 30 [11] 1077 -1087 (1987). 185. W. Qiu, H.X. Xu, N.J. Podraza, S.R. Phillpot, J.E. Baciak, S. Trolier McKinstry, and J.C. Nino, "Analysis of the Band Structure of BiI3 Single Crystals via Optical and Computational Methods," In Preparation, 186. M. Albrecht and H.P. Strunk, "Strain induced electronic alterations of dislocation mobility GeSi layers," Physica Status Solidi B -Basic Research, 222 [1] 95 -100 (2000). 187. F. Yao, C. L. Xue, B.W. Cheng, and Q.M. Wang, "Band gap narrowing in heavily B doped Si1xGex strained layers," Acta Phys Sin -Ch Ed, 56 [11] 66546659 (2007). 188. R.B. Fair, "Effect of Strain -Induced Band-Gap Narrowing on High-Concentration Phosphorus Diffusion in S ilicon," Journal of Applied Physics 50 [2] 860 868 (1979). 189. S.C. Jain, T.J. Gosling, D.H.J. Totterdell, J. Poortmans, R.P. Mertens, and R. Vanoverstraeten, "The Combined Effects of Strain and Heavy Doping on the Indirect Band-Gap of Si and Gexsi1X Al loys," Solid State Electron, 34 [5] 445-451 (1991).

PAGE 157

157 190. Y. Ishikawa, K. Wada, J.F. Liu, D.D. Cannon, H.C. Luan, J. Michel, and L.C. Kimerling, "Straininduced enhancement of near -infrared absorption in Ge epitaxial layers grown on Si substrate," Journal o f Applied Physics, 98 [1] (2005). 191. I.M. Dharmadasa, "Recent Developments and Progress on Electrical Contacts to CdTe, CdS and ZnSe with Special Reference to Barrier Contacts to CdTe," Prog. Cryst. Growth Charact. Mater. 36 [4] 249 290 (1998). 192. C Marinelli, L. Sorba, B.H. Muller, D. Kumar, D. Orani, S. Rubini, A. Franciosi, S. De Franceschi, M. Lazzarino, and F. Beltram, "Ohmic versus Rectifying Contacts through Interfacial Dipoles: Al/InxGa1 xAs," J. Cryst. Growth 202 769772 (1999). 193. J.F. Moulder, W.F. Stickle, P.E. Sobol, and K.D. Bomben, "Handbook of X -ray Photoelectron Spectroscopy", ULVAC PHI, Inc., Chanhassen, 1995. 194. F.M. Ismail and Z.M. Hanafi, "Some Physicochemical Properties of Bismuth Chalcogenides X -Ray Photoelectron and Diffuse Reflectance Spectra," Z Phys Chem Leipzig, 267 [4] 667672 (1986). 195. Y. Schuhl, H. Baussart, R. Delobel, M. Lebras, J.M. Leroy, L. Gengembre, and J. Grimblot, "Study of Mixed-Oxide Catalysts Containing Bismuth, Vanadium and Antimony Preparation, Ph ase -Composition, Spectroscopic Characterization and Catalytic -Oxidation of Propene," J Chem Soc Farad T 1, 79 2055 -& (1983). 196. V.I. Nefedov, D. Gati, B.F. Dzhurinskii, N.P. Sergushin, and Y.V. Salyn, "X -Ray Electron Study of Oxides of Elements," Zh. Neo rg. Khim. 20 [9] 2307 -2314 (1975). 197. V.S. Dharmadhikari, S.R. Sainkar, S. Badrinarayan, and A. Goswami, "Characterization of Thin -Films of Bismuth Oxide by X -Ray Photo -Electron Spectroscopy," J. Electron Spectrosc. Relat. Phenom. 25 [2 3] 181-189 (198 2). 198. S. Leroy, F. Blanchard, R. Dedryvere, H. Martinez, B. Carre, D. Lemordant, and D. Gonbeau, "Surface film formation on a graphite electrode in Li ion batteries: AFM and XPS study," Surf. Interface Anal. 37 [10] 773-781 (2005). 199. Y. Takasu, K. Y asuda, Y. Matsuda, and I. Toyoshima, "Adsorption of CO on Gold and GoldModified Platinum," Chem. Phys. Lett. 152 [1] 105-108 (1988). 200. J. Nakamura, I. Toyoshima, and K. Tanaka, "Formation of Carbidic and Graphitic Carbon from CO on Polycrystalline Cob alt," Surf. Sci. 201 [1 2] 185194 (1988). 201. M.C. Wu, S.Z. Dong, and A.R. Zhu, "Studies of CO Dissociation Promoted by Na on Ag(110)," Surf. Sci. 216 [3] 420-428 (1989). 202. U. Gelius, P.F. Heden, J. Hedman, B.J. Lindberg, R. Marine, R. Nordberg, C. Nordling, and K. Siegbahn, "Molecular Spectroscopy by Means of ESCA," Phys. Scr. 2 70 -80 (1970).

PAGE 158

158 203. J.J. Pireaux, M. Liehr, P.A. Thiry, J.P. Delrue, and R. Caudano, "Electron Spectroscopic Characterization of Oxygen-Adsorption on Gold Surfaces .2. Produ ction of Gold Oxide in Oxygen Dc Reactive Sputtering," Surf. Sci. 141 [1] 221-232 (1984). 204. G. Kumar, Blackbur.Jr, R.G. Aldridge, W.E. Moddeman, and M.M. Jones, "Photoelectron Spectroscopy of Coordination-Compounds .2. Palladium Complexes," Inorg. Chem 11 [2] 296 -& (1972). 205. T.H. Fleisch, G.W. Zajac, J.O. Schreiner, and G.J. Mains, "An Xps Study of the UV Photoreduction of Transition and Noble Metal Oxides," Appl. Surf. Sci. 26 [4] 488 497 (1986). 206. J.M. Tura, P. Regull, L. Victori, and M.D. De castellar, "Xps and IR (ATR) Analysis of Pd Oxide Films Obtained by Electrochemical Methods," Surf. Interface Anal. 11 [8] 447449 (1988). 207. J. Kappler, N. Barsan, U. Weimar, A. Dieguez, J.L. Alay, A. RomanoRodriguez, J.R. Morante, and W. Gopel, "Corr elation between XPS, Raman and TEM measurements and the gas sensitivity of Pt and Pd doped SnO2 based gas sensors," Fresen J Anal Chem 361 [2] 110114 (1998). 208. M. Moroseac, T. Skala, K. Veltruska, V. Matolin, and I. Matolinova, "XPS and SSIMS Studies of Pd/SnOx System: Reduction and Oxidation in Hydrogen Containing Air," Surf. Sci. 566 1118 -1123 (2004). 209. M. Schieber, T.E. Schlesinger, and R.B. James, "Semiconductors for Room Temperature Nuclear Detector Applications," 43 561 (1995). 210. J.F. Con deles, R.A. Ando, and M. Mulato, "Optical and Structural Properties of PbI2 Thin Films," J. Mater. Sci. 43 [2] 525529 (2008). 211. G.X. Zhang, D.Q. Yang, and E. Sacher, "X -ray Photoelectron Spectroscopic Analysis of Pt Nanoparticles on Highly Oriented Py rolytic Graphite, Using Symmetric Component Line Shapes," J. Phys. Chem. C 111 [2] 565 570 (2007). 212. L. Xiao and L.C. Wang, "Structures of Platinum Clusters: Planar or Spherical?," J. Phys. Chem. A 108 [41] 8605 -8614 (2004). 213. B.J. Curtis and H.R. Brunner, "The Crystal Growth of Bismuth Iodide," Mater. Res. Bull. 9 [5] 715720 (1974). 214. D. Alexiev, K.S.A. Butcher, and A.A. Williams, "Gamma-Ray Detectors from Thermally Annealed Bridgman-Grown CdTe," J. Cryst. Growth 142 [3 -4] 303309 (1994).

PAGE 159

159 215. J.M. Arias, M. Zandian, S.H. Shin, W.V. Mclevige, J.G. Pasko, and R.E. Dewames, "Dislocation Density Reduction by Thermal Annealing of HgCdTe Epilayers Grown by Molecular -Beam Epitaxy on GaAs Substrates," J Vac Sci Technol B 9 [3] 1646-1650 (1991). 216. Y.J. Li, G.L. Ma, X.N. Zhan, and W.Q. Jie, "The Annealing of Cd1 xZnxTe in CdZn Vapors," J. Electron. Mater. 31 [8] 834840 (2002). 217. W.W. Li, W.B. Sang, B. Zhang, J.H. Min, and F. Yu, "Annealing High Resistivity CdZnTe Wafers under Controlled Cd/Zn P artial Pressures," J Mater Sci Technol 20 [6] 703706 (2004). 218. D.M. Zeng, W.Q. Jie, G.Q. Zha, T. Wang, and G. Yang, "Effect of Annealing on the Residual Stress and Strain Distribution in CdZnTe Wafers," J. Cryst. Growth 305 [1] 5054 (2007). 219. H. Kotake and S. Takasu, "The Effect of Annealing on Residual -Stress and Dislocation Propagation in Silicon Slices with Damaged Layer Induced by Scribing," J. Mater. Sci. 16 [3] 767774 (1981). 220. H. Chen, K. Chattopadhyay, K.T. Chen, A. Burger, M.A. Georg e, J.C. Gregory, P.K. Nag, J.J. Weimer, and R.B. James, "Passivation of CdZnTe Surfaces by Oxidation in Low Energy Atomic Oxygen," J Vac Sci Technol A, 17 [1] 97-101 (1999). 221. M.C. Duff, D.B. Hunter, A. Burger, M. Groza, V. Buliga, and D.R. Black, Effect of Surface Preparation Technique on the Radiation Detector Performance of CdZnTe," Appl. Surf. Sci. 254 [9] 2889 2892 (2008). 222. A. Levi, A. Burger, J. Nissenbaum, M. Schieber, and Z. Burshtein, "Search for Improved Surface -Treatment Procedures i n Fabrication of Hgi2 X Ray Spectrometers," Nucl Instrum Methods 213 [1] 35 38 (1983). 223. G. Ivensky, I. Zafrany, and S. BenYaakov, "Generic O perational C haracteristics of Piezoelectric T ransformers," Ieee T Power Electr 17 [6] 1049-1057 (2002). 224. M. Guo, K.H. Lam, D.M. Lin, S. Wang, K.W. Kwok, H.L.W. Chan, and X.Z. Zhao, "A Rosen -type P iezoelectric T ransformer E mploying L ead -free K0.5Na0.5NbO3 C eramics," J Mater Sci 43 [2] 709 -714 (2008). 225. M.S. Kim, S. Jeon, D.S. Lee, S.J. Jeong, and J.S. Song "Lead-free NKN -5LT Piezoelectric M aterials for M ultilayer C eramic Actuator," J Electroceramics 23 [2 4] 372-375 (2009). 226. Z.Y. Shen, Y. Xu, and J.F. Li, "Fabrication and Electromechanical P roperties of M icroscale 13 -type P iezoelectric C omposites U s ing (Na,K)NbO3based Pb-free Piezoceramics," J Appl Phys 105 [10] (2009).

PAGE 160

160 BIOGRAPHICAL SKETCH Wei Qiu was born in 19 75 in Chongqing China. He obtained his B.S. degree in materials science and engineering at Tsinghua U niversity Beijing, in 1999 After graduation, h e was offered a research position in the state key laboratory of fine ceramic and processing at Tsinghua University, where he was in charge of the development of a ceramic polymer frictional material for the application in the ultrasonic motor. In 200 0 he began his Master s study in Dr. Huey H. Hng s research group in department of materials science and engineering at Nanyang Technological University (NTU), Singapore. His study focused on the microstructural and piezoelectric properties of l ead z irconate t itanate (PZT ) near the morphotropic phase boundary In 2002, He received his Master s degree and joined Dr. Ooi K. Tan s group as a research assistant in department of electrical and electronic engineering at NTU. His research focused on the s ynthesi s of barium strontium titanate (BST) thin film by sol gel process and developing the BST thin film for bio -sensing application. In 2005, he began his Ph. D study in Dr. Juan C. Ninos research group. His Ph.D study focuses on the single crystal growth of BiI3 by modified vertical Bridgman technique for the application of room temperature gamma -ray detector After spending 5 years at UF in Gainesville he received his Ph.D from the University of Florida in the s ummer of 20 10.