<%BANNER%>

Ion Implantation of Gadolinium in Compound Semiconductor Materials and Potential Spintronic Device Applications

Permanent Link: http://ufdc.ufl.edu/UFE0041092/00001

Material Information

Title: Ion Implantation of Gadolinium in Compound Semiconductor Materials and Potential Spintronic Device Applications
Physical Description: 1 online resource (181 p.)
Language: english
Creator: Davies, Ryan
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: gaas, gan, gd, implantation, ion, magnetic, semiconductor, spintronics, zno
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: ION IMPLANTATION OF GADOLINIUM IN COMPOUND SEMICONDUCTOR MATERIALS AND POTENTIAL SPINTRONIC DEVICE APPLICATIONS As device dimensions have continued to shrink, atomistic scale fluctuations in material properties are beginning to limit continued improvements in device performance. Various technologies are being pursued to overcome this problem. Spin transport electronics, or spintronics, has been proposed as an attractive approach. This technology utilizes the spin of the electron, in addition to the charge of the electron, to transmit information through a device. The most promising materials for spintronic device applications are dilute magnetic semiconductors, which are formed when dilute amounts of magnetic atoms are incorporated into semiconductor materials. Recently, ion implantation has been studied as the incorporation method of magnetic ions into a host semiconductor material system for potential spintronic applications. This method provides excellent control over the quantity of the implanted ion and the resultant magnetic properties of the implanted material. For this study, the compound semiconductor materials GaN, ZnO, and GaAs are examined as target materials for Gd ion implantation. Before implantation, these materials exhibited ferromagnetic behavior without the known presence of magnetic impurities and with a dependence on the applied magnetic field/sample surface orientation. Measuring the magnetic properties of these materials with a perpendicular orientation between the applied field and the sample surface exhibited a larger magnetic signal than examining with a parallel orientation between the applied field and sample surface, described in this work as an anisotropic enhancement effect. Ferromagnetism was demonstrated in hysteresis loops visible at both low temperature (10 K) and room temperature. The ferromagnetic mechanism occurring in the non-implanted materials is speculated as being due to anion-related defects (vacancies and interstitials). Ferromagnetism was also demonstrated in the implanted compound semiconductor materials. Implanting Gd ions into GaN has resulted in this material exhibiting ferromagnetic behavior before any thermal annealing treatment. Co-implanting Si ions with Gd ions in GaN also shows room temperature ferromagnetism and a larger magnetic moment than the same GaN only implanted with Gd. Additional studies into the effects of Gd ion implantation on the magnetic properties of another wurtzite crystal structure compound semiconductor material (ZnO) and a small band gap semiconductor material (GaAs) provide additional insight into the ferromagnetic mechanism present in these materials. The mechanism occurring in the implanted materials is speculated as being due to interactions between the native defects and defects introduced during implantation with the implanted Gd. This interaction may be caused by long-range spin polarization that is demonstrated by the large magnetic moments observed in this work. Anion-related defects (vacancies and interstitials) appear to be the most likely defects to exhibit a spin orbit coupling with the implanted Gd atoms based on studies of implanted p-GaN and n-GaN. The effects of thermal annealing on the magnetic properties of the implanted thin films have also been investigated in implanted GaN and demonstrated that annealing does reduce the ferromagnetic ordering due to the decreased defect density as a result of the repaired lattice damage. Based on these results, ion implantation provides an exemplary method to control the amount of incorporated magnetic ions and results in desirable magnetic properties in the implanted materials for spintronic applications. The ferromagnetic mechanism occurring in the implanted materials appear to rely on the type and density of defects interacting with the magnetic impurity implanted into the compound semiconductor material. The next step would be to further develop a model predicting the ferromagnetic mechanism exhibited by the implanted materials and then leveraging that mechanism to select the optimal compound semiconductor material/implant species and dose combination to create a functional spintronic device.
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 Ryan Davies.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Abernathy, Cammy R.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0041092:00001

Permanent Link: http://ufdc.ufl.edu/UFE0041092/00001

Material Information

Title: Ion Implantation of Gadolinium in Compound Semiconductor Materials and Potential Spintronic Device Applications
Physical Description: 1 online resource (181 p.)
Language: english
Creator: Davies, Ryan
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: gaas, gan, gd, implantation, ion, magnetic, semiconductor, spintronics, zno
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: ION IMPLANTATION OF GADOLINIUM IN COMPOUND SEMICONDUCTOR MATERIALS AND POTENTIAL SPINTRONIC DEVICE APPLICATIONS As device dimensions have continued to shrink, atomistic scale fluctuations in material properties are beginning to limit continued improvements in device performance. Various technologies are being pursued to overcome this problem. Spin transport electronics, or spintronics, has been proposed as an attractive approach. This technology utilizes the spin of the electron, in addition to the charge of the electron, to transmit information through a device. The most promising materials for spintronic device applications are dilute magnetic semiconductors, which are formed when dilute amounts of magnetic atoms are incorporated into semiconductor materials. Recently, ion implantation has been studied as the incorporation method of magnetic ions into a host semiconductor material system for potential spintronic applications. This method provides excellent control over the quantity of the implanted ion and the resultant magnetic properties of the implanted material. For this study, the compound semiconductor materials GaN, ZnO, and GaAs are examined as target materials for Gd ion implantation. Before implantation, these materials exhibited ferromagnetic behavior without the known presence of magnetic impurities and with a dependence on the applied magnetic field/sample surface orientation. Measuring the magnetic properties of these materials with a perpendicular orientation between the applied field and the sample surface exhibited a larger magnetic signal than examining with a parallel orientation between the applied field and sample surface, described in this work as an anisotropic enhancement effect. Ferromagnetism was demonstrated in hysteresis loops visible at both low temperature (10 K) and room temperature. The ferromagnetic mechanism occurring in the non-implanted materials is speculated as being due to anion-related defects (vacancies and interstitials). Ferromagnetism was also demonstrated in the implanted compound semiconductor materials. Implanting Gd ions into GaN has resulted in this material exhibiting ferromagnetic behavior before any thermal annealing treatment. Co-implanting Si ions with Gd ions in GaN also shows room temperature ferromagnetism and a larger magnetic moment than the same GaN only implanted with Gd. Additional studies into the effects of Gd ion implantation on the magnetic properties of another wurtzite crystal structure compound semiconductor material (ZnO) and a small band gap semiconductor material (GaAs) provide additional insight into the ferromagnetic mechanism present in these materials. The mechanism occurring in the implanted materials is speculated as being due to interactions between the native defects and defects introduced during implantation with the implanted Gd. This interaction may be caused by long-range spin polarization that is demonstrated by the large magnetic moments observed in this work. Anion-related defects (vacancies and interstitials) appear to be the most likely defects to exhibit a spin orbit coupling with the implanted Gd atoms based on studies of implanted p-GaN and n-GaN. The effects of thermal annealing on the magnetic properties of the implanted thin films have also been investigated in implanted GaN and demonstrated that annealing does reduce the ferromagnetic ordering due to the decreased defect density as a result of the repaired lattice damage. Based on these results, ion implantation provides an exemplary method to control the amount of incorporated magnetic ions and results in desirable magnetic properties in the implanted materials for spintronic applications. The ferromagnetic mechanism occurring in the implanted materials appear to rely on the type and density of defects interacting with the magnetic impurity implanted into the compound semiconductor material. The next step would be to further develop a model predicting the ferromagnetic mechanism exhibited by the implanted materials and then leveraging that mechanism to select the optimal compound semiconductor material/implant species and dose combination to create a functional spintronic device.
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 Ryan Davies.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Abernathy, Cammy R.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0041092:00001


This item has the following downloads:


Full Text

PAGE 1

1 ION IMPLANTATION OF GADOLINIUM IN COMPOUND SEMICONDUCTOR MATERIALS AND POTENTIAL SPINTRONIC DEVICE APPLICATIONS By RYAN PATRICK DAVIES 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 2009

PAGE 2

2 2009 Ryan Patrick Davies

PAGE 3

3 To my wife, family, and friends

PAGE 4

4 ACKNOWLEDGMENTS First and foremost, I thank my Lord for the many blessings in my life and for allowing me to think Gods thoughts after Him, as the scientist Jo hannes Kepler so aptly stated. I am truly blessed to have parents who instilled in me early in life the importance of a good education and have fully supported me t hroughout all of my schooling. I am also immeasurably grateful for my wife, without whom I would have likely never returned to graduate school and completed this work She pr ovided valuable encouragement for me to pursue my life goals. I also want to thank all of my family for giving me their love, support, and great memories throughout my life. I greatly appreciate my advisor, Dr. Cammy Abernathy, who mentored me both in my research activities and professional development. She helped me devel op the skills necessary to excel in my research and academia. Al so, I extend my gratitude to Dr. Brent Gila for providing both his valuable expertise and helpful advice He assisted me countless times when I was at a standstill in both my research and the use of problematic equipment. I am grateful for Dr. Stephen Pearton, D r. David Norton and Dr. Amlan Biswas taking the time to serve on my committee and providing their insight on my research. I also want to acknowledge the assistance provided by Dr. Mark M eisel and Dr. Christopher Stanton. I also extend many thanks to my fellow Abernathy re search group members for their camaraderie and the help they provided with my projects. Finally I would like to thank any and all people I forgot to mention here P lease know I am indebted to each of you for your prayers and support.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ...................................................................................................... 4 LIST OF FIGURES .............................................................................................................. 7 ABSTRACT ........................................................................................................................ 14 CHAPTER 1 INTRODUCTION ........................................................................................................ 17 2 LITERATURE REVIEW OF MAGNETIC ATOM INCORPORATION IN COMPOUND SEMICONDUCTOR MATERIALS ...................................................... 19 Literature Review of Experimental Data .................................................................... 19 Materials Selec tion Criteria .................................................................................. 19 Eu Chalcogenides and II -VI Semiconductors ..................................................... 19 Transition Metals in III -V Semiconductors .......................................................... 20 Rare Earth Elements in III-V Semiconductors .................................................... 22 Literature Review of Theoretical Studies ................................................................... 24 Free Carrier Mediated Model ............................................................................... 24 Percolation Model ................................................................................................ 25 Extension of Models to Rare Earth Elements ..................................................... 26 Consideration of Defects in Models ..................................................................... 26 3 EXPERIMENTAL PROCED URES FOR SYNTHESIS, MAGNETIC ATOM INCORPORATION AND CHARACTERIZATION OF COMPOUND SEMICONDUCTOR MATERIALS .............................................................................. 32 Synthesis ..................................................................................................................... 32 Molecular Beam Epitaxy ...................................................................................... 32 Metal -Organic Chemical Vapor Deposition ......................................................... 34 Magnetic Impurity Incorporation ................................................................................. 35 Molecular Beam Epitaxy ...................................................................................... 35 Ion Implantation .................................................................................................... 36 Characterization .......................................................................................................... 36 Superconducting Quantum Interference Device Magnetometry ........................ 36 X -ray Diffraction ................................................................................................... 37 Photoluminescence .............................................................................................. 38 4 ION IMPLANTATION OF GADOLINIUM AND SILICON IN GALLIUM NITRIDE ..... 42 MOCVD u -GaN ........................................................................................................... 43 Non -Implanted ...................................................................................................... 44 Si -Implanted ......................................................................................................... 45

PAGE 6

6 Gd -Implanted ........................................................................................................ 46 Gd and Si -Co -Implanted ..................................................................................... 47 Mechanism of Ferromagnetism ........................................................................... 49 MOCVD p -GaN ........................................................................................................... 52 Non -Implanted ...................................................................................................... 52 Gd -Implanted ........................................................................................................ 53 Mechanism of Ferromagnetism ........................................................................... 54 MOCVD n -GaN ........................................................................................................... 56 Non -Implanted ...................................................................................................... 56 Gd -Implanted ........................................................................................................ 57 Mechanism of Ferromagnetism ........................................................................... 58 MBE Gd -Doped GaN .................................................................................................. 58 Non -Implanted ...................................................................................................... 59 Si -Implanted ......................................................................................................... 60 Mechanism of Ferromagnetism ........................................................................... 60 Summary ..................................................................................................................... 61 5 ION IMPLANTATION OF GADOLINIUM IN ZINC OXIDE ...................................... 127 Non -Implanted .......................................................................................................... 127 Gd -Implanted ............................................................................................................ 128 Mechanism of Ferromagnetism ................................................................................ 129 6 ION IMPLANTATION OF GADOLINIUM IN GALLIUM ARSENIDE ....................... 142 Non -Implanted .......................................................................................................... 142 Gd -Implanted ............................................................................................................ 143 Mechanism of Ferromagnetism ................................................................................ 144 7 THERMAL ANNEALING EFFECTS ON IMPLANTED GALLIUM NITRIDE ........... 157 8 POTENTIAL APPLICATIONS, SUMMARY, AND FUTURE WORK ....................... 164 Potential Applications ............................................................................................... 164 Summary ................................................................................................................... 165 Future Work .............................................................................................................. 169 LIST OF REFERENCES ................................................................................................. 176 BIOGRAPHICAL SKETCH .............................................................................................. 181

PAGE 7

7 LIST OF FIGURES Figure page 2 -1 Magnetic impurity ions in compound semiconductor host material ...................... 29 2 -2 Predicted Curie temperatures as a function of band gap for Group IV and III V semiconductor materials ..................................................................................... 30 2 -3 Predicted Curie temperatures as a function of lattice constant for Group IV and vari ous c ompound semiconductor materials .................................................. 31 3 -1 Diagram of Varian Gen II MBE system ................................................................. 40 3 -2 Diagram of Veeco/Emcore P75 MOCVD vertical rotating disk reactor MOCVD system ...................................................................................................... 41 4 -1 Implantation profile f or the combined Si doses in GaN ......................................... 63 4 -2 Implantation profile for the Gd dose in GaN .......................................................... 64 4 -3 Powder XRD scan for the non-implant ed MOCVD u -GaN reference sample ...... 65 4 -4 Diagram s of parallel and perpendicular applied magnetic field/sample surface orientations ............................................................................................................. 66 4 -5 Magnetization vs. applied field for non -implanted MOCVD u-GaN at 10 K for the parallel sample surface/applied field orientation ............................................. 67 4 -6 Magnetization vs. applied field for non -implanted MOCVD u-GaN at 300 K for the parallel sample surface/applied field orientation ............................................. 68 4 -7 Magnetization vs. applied field for non -implanted MOCVD u-GaN at 10 K for the perpendicular sample surface/applied field orientation .................................. 69 4 -8 Magnetization vs. applied field for non -implanted MOCVD u-GaN at 300 K for the perpendicular sample surface/applied field orientation .................................. 70 4 -9 Powder XRD scan comparing nonimplanted MOCVD u-GaN to Si implanted MOCVD u -GaN ....................................................................................................... 71 4 -10 Magnetization vs. applied field for Si -implanted MOCVD u-GaN at 10 K for the parallel sample surface/applied field orientation ............................................. 72 4 -11 Magnetization vs. applied field for Si -implanted MOCVD u-GaN at 300 K for the parallel sample surface/applied field orientation ............................................. 73 4 -12 Magnetization vs. applied field for Si -implanted MOCVD u-GaN at 10 K for the perpendicular sample surface/applied field orientation .................................. 74

PAGE 8

8 4 -13 Magnetization vs. applied field for Si -implanted MOCVD u-GaN at 300 K for the perpendicular sample surface/applied field or ientation .................................. 75 4 -14 Powder XRD scan comparing nonimplanted MOCVD u-GaN to Gd implanted MOCVD u-GaN ..................................................................................... 76 4 -15 Magnetization vs. applied field for Gdimplanted MOCVD u-GaN at 10 K for the parallel sample surface/appl ied field orientation ............................................. 77 4 -16 Magnetization vs. applied field for Gdimplanted MOCVD u-GaN at 300 K for the parallel sample surface/applied field orientation ............................................. 78 4 -17 Magnetization vs. applied field for Gdimplanted MOCVD u-GaN at 10 K for the perpendicular sample surface/applied field orientation .................................. 79 4 -18 Magnetization vs. applied field for Gdimplanted MOCVD u-GaN at 300 K for the perpendicular sample surface/applied field orientation .................................. 80 4 -19 Powder XRD scan comparing nonimplanted MOCVD u-GaN to Gd and Si co implanted MOCVD u-GaN ................................................................................ 81 4 -20 Magnetization vs. applied field for Gdand Si -co -implanted MOCVD u-GaN at 10 K for the parallel sample surf ace/applied field orientation ........................... 82 4 -21 Magnetization vs. applied field for Gdand Si -co -implanted MOCVD u-GaN at 300 K for the parallel sample surface/applied field orientation ......................... 83 4 -22 Magnetization vs. applied field for Gdand Si -co -implanted MOCVD u-GaN at 10 K for the perpendicular sample surface/applied field orientation ................ 84 4 -23 Magnetization vs. applied field for Gdand Si -co -implanted MOCVD u-GaN at 300 K for the perpendicular sample surface/applied field orientation .............. 85 4 -24 Magnetization vs. temperature curves for Gdan d Si -co -implanted MOCVD u -GaN ..................................................................................................................... 86 4 -25 Magnetization vs. applied field for Si -, Gd and Gdand Si -co -implanted MOCVD u -GaN samples at 10 K for the perpendicular sample surface/applied field orientation ............................................................................. 87 4 -26 Magnetization vs. applied field for Si -, Gd and Gdand Si -co -implanted MOCVD u -GaN samp les at 300 K for the perpendicular sample surface/applied field orientat ion ............................................................................. 88 4 -27 Gd and Si implantation profiles compared to the vacancy concentration profiles due to the Gd and Si i mplants .................................................................. 89

PAGE 9

9 4 -28 PL spectra at 17 K for Gd and Si -co implanted MOCVD u-Ga N before and after implantation .................................................................................................... 90 4 -29 PL spectra at 300 K for Gdand Si -co -implanted MOCVD u -Ga N before and after implantation .................................................................................................... 91 4 -30 Powder XRD scan for the non-implant ed MOCVD p -GaN reference sample ...... 92 4 -31 Magnetization vs. applied field for non -implanted MOCVD p-GaN at 10 K for the parallel sample surface/applied field orientation ............................................. 93 4 -32 Magnetization vs. applied field for non -implanted MOCVD p-GaN at 300 K for the parallel sample surface/applied field orientation ............................................. 94 4 -33 Magnetization vs. applied field for non -implanted MOCVD p-GaN at 10 K for the perpendicular sample surface/applied field orientation .................................. 95 4 -34 Magnetization vs. applied field for non -implanted MOCVD p-GaN at 300 K for the perpendicular sample surface/applied field orientation .................................. 96 4 -35 Powder XRD scan comparing nonimplanted MOCVD p-GaN to G d implanted MOCVD p-GaN ..................................................................................... 97 4 -36 Magnetization vs. applied field for Gdimplanted MOCVD p-GaN at 10 K for the parallel sample surface/applied field orientation ............................................. 98 4 -37 Magnetization vs. applied field for Gdimplanted MOCVD p-GaN at 300 K for the parallel sample surface/applied field orientation ............................................. 99 4 -38 Magnetization vs. applied field for Gdimplanted MOCVD p-GaN at 10 K for the perpendicular sample surface/applied field orientation ................................ 100 4 -39 Magnetization vs. applied field for Gdimplanted MOCVD p-GaN at 300 K for the perpendicular sample surface/applied field or ientation ................................ 101 4 -40 Energies of formation as a function of Fermi level for native point defects in GaN ....................................................................................................................... 102 4 -41 PL spectra at 17 K for Gd implanted MOCVD p-Ga N before and after implantation .......................................................................................................... 103 4 -42 PL spectra at 300 K for Gd-implanted MOCVD p-Ga N before and after implantation .......................................................................................................... 104 4 -43 Powder XRD scan for the non-implanted MOCVD n -GaN reference sample .... 105 4 -44 Magnetization vs. applied field for non -implanted MOCVD n-GaN at 10 K for the parallel sample surface/applied field orientation ........................................... 106

PAGE 10

10 4 -45 Magnetization vs. applied field for non -implanted MOCVD n-GaN at 300 K for the parallel sample surface/applied field orientation ........................................... 107 4 -46 Magnetization vs. applied field for non -implanted MOCVD n-GaN at 10 K for the perpendicular sample surface/applied field orientation ................................ 108 4 -47 Magnetization vs. applied field for non -implanted MOCVD n-GaN at 300 K for the perpendicular sample surface/applied fie ld orientation ................................ 109 4 -48 Powder XRD scan comparing nonimplanted MOCVD n-GaN to Gd implanted MOCVD n-GaN ................................................................................... 110 4 -49 Magneti zation vs. applied field for Gdimplanted MOCVD n-GaN at 10 K for the parallel sample surface/applied field orientation ........................................... 111 4 -50 Magnetization vs. applied field for Gdimplanted MOCVD n-GaN at 300 K for the parallel sample surface/applied field orientation ........................................... 112 4 -51 Magnetization vs. applied field for Gdimplanted MOCVD n-GaN at 10 K for the perpendicular sample surface/applied field orientation ................................ 113 4 -52 Magnetization vs. applied field for Gdimplanted MOCVD n-GaN at 300 K for the perpendicular sample surface/applied field orient ation ................................ 114 4 -53 Magnetization vs. applied field for non -implanted MBE GaGdN at 10 K for the parallel sample su rface/applied field orientation ................................................. 115 4 -54 Magnetization vs. applied field for non -implanted MBE GaGdN at 300 K for the parallel sample surface/applied field orientation ........................................... 116 4 -55 Magnetization vs. applied field for non -implanted MBE GaGdN at 10 K for the perpendicular sample surface/applied field orientation ....................................... 117 4 -56 Magnetization vs. applied field for non -implanted MBE GaGdN at 300 K for the perpendicular sample surface/applied field orientation ................................ 118 4 -57 Magnetization vs. applied field for Si -implanted MBE GaGdN at 10 K for the parallel sample su rface/applied field orientation ................................................. 119 4 -58 Magnetization vs. applied field for Si -implanted MBE GaGdN at 300 K for the parallel sample su rface/applied field orientation ................................................. 120 4 -59 Magnetization vs. applied field for Si -implanted MBE GaGdN at 10 K for the perpendicular sample surface/applied field orientation ....................................... 121 4 -60 Magnetization vs. applied field for Si -implanted MBE GaGdN at 300 K for the perpendicular sample surface/applied field orientation ....................................... 122

PAGE 11

11 4 -61 Magnetization vs. applied field for non -implanted undoped, ptype, and ntype GaN samples at 10 K for the perpendicular sample surface/applied field orientation ............................................................................................................. 123 4 -62 Magnetization vs. applied field for non -implanted undoped, ptype, and ntype GaN samples at 300 K for the perpendicular sample su rface/applied field orientation ..................................................................................................... 124 4 -63 Magnetization vs. applied field for Gdimplanted un doped, p-type, and ntype GaN samples at 10 K for the perpendicular sample surface/applied field orientation ............................................................................................................. 125 4 -64 Magnetization vs. applied field for Gdimplanted undoped, p-type, and ntype GaN samples at 300 K for the perpendicular sample surface/applied field orientation ............................................................................................................. 126 5 -1 Powder XRD scan for the non-implanted ZnO reference sample ...................... 131 5 -2 Magnetization vs. applied field for non -implanted ZnO at 10 K for the parallel sample surface/applied field orientation .............................................................. 132 5 -3 Magnetization vs. applied field for non -implanted ZnO at 300 K for the parallel sample su rface/applied field orientation ................................................. 133 5 -4 Magnetization vs. applied field for non -implanted ZnO at 10 K for the perpendicular sample surface/applied field orientation ....................................... 134 5 -5 Magnetization vs. applied field for non -implanted ZnO at 300 K for the perpendicular sample surface/applied field or ientation ....................................... 135 5 -6 Implantation profile for the Gd dose in ZnO ........................................................ 136 5 -7 Powder XRD scan comparing nonimplanted ZnO to Gd -implanted ZnO ......... 137 5 -8 Magnetization vs. applied field for Gdimplanted ZnO at 10 K for the parallel sample surface/applied field orientation .............................................................. 138 5 -9 Magnetization vs. applied field for Gdimplanted ZnO at 300 K for the parallel sample surface/applied field orientation .............................................................. 139 5 -10 Magnetization vs. applied field for Gdimplanted ZnO at 10 K for the perpendicular sample surface/applied field orientation ....................................... 140 5 -11 Magnetization vs. applied field for Gdimplanted ZnO at 300 K for the perpendicular sample s urface/applied field orientation ....................................... 141 6 -1 Powder XRD scan for the non-implanted GaAs reference sample .................... 146

PAGE 12

12 6 -2 Magnetization vs. applied field for non -implanted GaAs at 10 K for the parallel sample su rface/applied field orientation ................................................. 147 6 -3 Magnetization vs. applied field for non -implanted GaAs at 300 K for the parallel sample su rface/applied field orientation ................................................. 148 6 -4 Magnetization vs. applied field for non -implanted GaAs at 10 K for the perpendicular sample surface/applied field orientation ....................................... 149 6 -5 Magnetization vs. applied field for non -implanted GaAs at 300 K for the perpendicular sample surface/applied field orientation ....................................... 150 6 -6 Implantation profile for the Gd dose in GaAs ...................................................... 151 6 -7 Powder XRD scan comparing nonimpl anted GaAs to Gdimplanted GaAs ..... 152 6 -8 Magnetization vs. applied field for Gd implanted GaAs at 10 K for the parallel sample surface/applied field orientation .............................................................. 153 6 -9 Magnetization vs. applied field for Gd implanted GaAs at 300 K for the parallel sample su rface/applied field orientation ................................................. 154 6 -10 Magnetization vs. applied field for G d implanted GaAs at 10 K for the perpendicular sample surface/applied field orientation ....................................... 155 6 -11 Magnetization vs. applied field for Gd implanted GaAs at 300 K for the perpendicular sample surface/applied field orientation ....................................... 156 7 -1 Magnetization vs. applied field curves for as implanted and annealed Gdand Si -co -implanted u-GaN at 300 K for the perpendicular sample su rface/applied field orientation ..................................................................................................... 160 7 -2 Normalized magnetizations values at +1000 G for Gdan d Si -co -implanted MOCVD u -GaN ..................................................................................................... 161 7 -3 PL spectra at 17 K for Gd and Si -co implanted MOCVD u-GaN before and aft er implantation and after annealing ................................................................. 162 7 -4 PL spectra at 300 K for Gdand Si -co -im planted MOCVD u -GaN before and aft er implantation and after annealing ................................................................. 163 8 -1 Schematic of vertical mag netic tunnel junction ................................................... 171 8 -2 Magnetization vs. applied field for non -implanted GaN (undoped, p-type, and n -type), ZnO, and GaAs samples at 10 K for the perpend icular sample su rface/applied field orientation ........................................................................... 172

PAGE 13

13 8 -3 Magnetization vs. applied field for non -implanted GaN (undoped, p-type, and n -type), ZnO, and GaAs samples at 300 K for the perpendicular sample su rface/app lied field orientation ........................................................................... 173 8 -4 Magnetization vs. applied field for implanted GaN (undoped, ptype, and n type), ZnO, and GaAs sampl es at 10 K for the perpendicular sample su rface/applied field orientation ........................................................................... 174 8 -5 Magnetization vs. applied field for implanted GaN (undoped, p type, and n type), ZnO, and GaAs samples at 300 K for the perpendicular sample su rface/applied field orientation ........................................................................... 175

PAGE 14

14 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy ION IMPLANTATION OF GADOLINIUM IN COMPOUND SEMICONDUCTOR MATERIALS AND POTENTIAL SPINTRONIC DEVICE APPLICATIONS By Ryan Patrick Davies December 2009 Chair: Cammy Abernathy Major: Materials Science and Engineering As device dimensions have continued to shrink, atomistic scale fluctuations in material properties are beginning to limit continued improvements in device performance. Various technologies are being pursued to overcome this problem. Spin transport elect ron ics, or spintronics, has been proposed as an attractive approach This technology utiliz es the spin of the electron in addition to the charge of the electron, to transmit information through a device. The most promising materials for spintronic device app lications are dilute magnetic semi conductors which are formed when dilute amounts of magnetic atoms are incorporated into semiconductor materials. Recently, ion implantation has been studied as the incorporation method of magnetic ions into a host semicon ductor material system for potential spintronic applications. This method provides excellent control over the quantity of the implanted ion and the resultant magnetic properties of the implanted material. For this study, the compound semiconductor material s GaN, ZnO, and GaAs are examined as target materials for Gd ion implantation. Before implantation, these materials exhibited ferromagnetic behavior without the known presence of magnetic

PAGE 15

15 impurities and with a dependence on the applied magnetic field/sampl e surface orientation. Measuring the magnetic properties of these materials with a perpendicular orientation between the applied field and the sample surface exhibited a larger magnetic signal than examining with a parallel orientation between the applied field and sample surface, described in this work as an anisotropic enhancement effect. Ferromagnetism was demonstrated in hysteresis loops visible at both low temperature (10 K) and room temperature. The ferromagnet ic mechanism occurring in the nonimplant ed mat erials is speculated as being due to anion-related defects (vacancies and interstitials). Ferromagnetism was also demonstrated in the implanted compound semiconductor materials. Implanting Gd ions into GaN has resulted in this material exhibiting fer romagnetic behavior before any thermal annealing treatment. Co implanting Si ions with Gd ions in GaN also shows room temperature ferromagnetism and a larger magnetic moment than the same GaN only implanted with Gd. Additional studies into the effects of G d ion implantation on the magnetic properties of another wurtzite crystal structure compound semiconductor material (ZnO) and a small band gap semiconductor material (GaAs) provide additional insight into the ferromagnetic mechanism present in the se materi als. The mechanism occurring in the implanted materials is speculated as being due to interaction s between the native defects and defects introduced during implanta tion with the implanted Gd. This interaction may be caused by long-range spin polarization t hat is demonstrated by the large magnetic moments observed in this work. Anion-related defects (vacancies and interstitials) appear to be the most likely defects to exhibit a spin orbit coupling with the implanted Gd atoms based on studies of implanted p-G aN and n-GaN. The effects of thermal

PAGE 16

16 annealing on the magnetic properties of the implanted thin films have also been investigated in implanted GaN and demonstrated that annealing does reduce the ferromagnetic ordering due to the decreased defect density as a result of the repaired lattice damage. Based on these results, ion implantation provides an exemplary method to control the amount of incorporated magnetic ions and results in desirable magnetic properties in the implanted materials for spintronic ap plications. The ferromagnetic mechanism occurring in the implanted materials appear to rely on the type and density of d efects interacting with the magnetic impurity implanted into the compound semiconductor material. The next step would be to further develop a model predicting the ferromagnetic mechanism exhibited by the implanted materials and then leveraging that me chanism to select the optimal compound semiconductor material/implant species and dose combination to create a functional spintronic device.

PAGE 17

17 CHAPTER 1 INTRODUCTION Spintronics, or spin transport electronics, are a topic of interest due to the barriers inherent to continued device downscaling. Due to this limitation, scientists and engineers are researching other options in hopes of continuing t he fulfillment of Moores law. A spintronic device would utilize the electron spin and spin transport to achieve additional functionality compared to traditional electronics sole use of the electron charge. Electron spin is an intrinsic quantum mechanical characteristic of electrons that exists in either a spin up or spin down configuration. I deally spintronic devices would manipulate the electron spin for switching functions while consuming less power and achieving faster switching times then conventional electronic devices. Spintronic devices have also been theorized to be able to open the door for the development of rewritable devic es that are nonvolatile ( i.e. can easily retain their data storage even when turned off).1 The current landscape of spintronics applications has been dominated by materials based on magnetoresistance, or the change in electrical resistance exhibited under an applied magnetic field. For example, electron spin has been exploited in spin valves which are used as magnetic tunnel junctions in magnetic random access memory (MRAM), a platform that aims to supersede current memory technologies. Preliminary pursuits in spintronic development in semiconductors proved problematic due to spin injection difficulties from the highly conductive ferromagnetic metal to the comparatively less conductive semiconductor.2 4 Based on these results incorporating the ferromagnetic metal into the semiconductor material has arisen as the ideal solution to this quandary and birthed the dilute magnetic semiconductor (DMS) field. DMS

PAGE 18

18 materials consist of a base semiconductor material that has a low concentration of ferromagnetic metals incorporated into the material These materials would retain the ir semiconducting properties (i.e. be able to achieve gain and possess a tunable band gap) while also demonstrating ferromagnetic behavior within the magnetically incorporated regions. For opt imal device operation, these materials must possess ferromagnetic properties observable at room temperature and efficient spin injection, transport and detection. Compound semiconductor materials, such as GaN, have established themselves as an ideal material base for specific electron ic and optoelectronic devices. Spintronic devices fabricated from these materials could integrate magnetic functionality with the electrical and photonic functions that have already been studied and developed Stud ies of electron spin in GaN have laid the groundwork for the feasibility of III-nitride DMS materials. Spin coherence in GaN has been observed at room temperature with a life time in the picosecond range.5 Using compound semiconductor materials such as GaN as the base DMS material system also provides advantages such as the achievability of tunable band gaps and easier lattice matching with substrate materials which results in better thin film synthesis for device applications.

PAGE 19

19 CHAPTER 2 LITERATURE REVIEW OF MAGNETIC ATOM INCORPORATION IN COMPOUND SEMICONDUCTOR MATERIALS Literature Review of Exp erimental Data Materials Selection Criteria Similar to most electronic devices, the practical application of spintronic devices hinges on being operational at room temperature. The suitability of these devices then require s that the constituent materials possess a Curie temperature (TC) above room temperature. Additional beneficial distinctions for device fabrication and operation would include the ability to control the Fermi level via doping or ion implantation, high quality crystal growth from substrate latti ce matching, radiation hardness for specific applications and the availability of a current technology base. Figure 2 -1 shows a schematic of how doping a semiconductor with a transition metal (TM) or rare earth (RE) ion can lead to alignment of spins associated with that ion and produce ferromagnetism.6 In wide band gap dilute magnetic semiconductor (DMS) materials there is still considerable controversy as to the mechanism that leads to the observed ferromagnetism. One aspect of this controversy comes from the formation of difficult to detect magnetically active clusters or second phases that may dictate the magneti c properties of the entire material (bot tom of Figure 21).6 Eu Chalcogenides and IIVI Semiconductors Concerning material selection, europium chalcogenides and II -VI semiconductors were some of the first materials considered for magnetic semiconductors. The Euchalcogenides proved problematic due to this materials lattice mismatch with the Si and GaAs substrates used for crystal growth.7 The II -VI semiconductors were difficult to dope p or n-type and the resulting magnetization was not of a ferromagnetic nature.7 8

PAGE 20

20 Also, both of these material systems had TCs at approximately 100 K, well below room temperature.9 Even though these material systems had low TC values, Dietl et al. predicted the TC of GaN and ZnO doped with 5% Mn to be above room temperature.10 Figure 2 2 shows these predicted TC values for Mn -doped Group IV and III -V semiconductor materials a nd their correlation with band gap. Additional studies have revealed a stronger correlation, demonstrated as a reciprocal function, between the predicted TC values and lattice constant, which is shown in Figure 2 3 The wider band gap semiconductors tend to have smaller lattice cons tants, large pd hybridization, and small spin orbit interaction and are predicted t o have higher TC values. Transition Metals in III-V Semiconductors Computational data predicting above room temperature TC values for III -V semiconductor materials doped with transition metals sparked interest in realizing a suitable material for spintronic applications. Mn -doped InAs was the first III -V material to exhibit ferromagnetic ordering due to the carrier being holes This material had to be grown in low temperature regimes via molecular beam epitaxy (MBE) to impede the growth of MnAs clusters and resulted in single phase p-type InMnAs with a TC of 7.5 K.1 1 ,1 2 Due to the more extensive knowledge base concerning GaAs and the possibility of numerous device applications, this mat erial was strongly considered. Unfortunately, the highest TC reported for GaMnAs was approximately 110 K1 3 and the corresponding research abated due to t he discovery of higher TC values for other DMS III -V materials. Room temperature ferromagnetism had later been shown in Mndoped GaP1 4, GaN1 5 1 7, and AlN.1 8 Fueled by the computation data by Dietl et al.10, research in transition metal doped III-V material s sped forward and issues such as crystal quality, g rowth conditions and

PAGE 21

21 the ferromagnetic mechanism in these materials came to the forefront. When considering the III nitride materials as the host matrix, both GaN and AlN were favored due to the extensiv e knowledge base arising from their suitability in optical and electronic applications such as ultraviolet light emitting diodes (LED s ) and high frequency, high temperature transistors. Primarily, transition metal doped GaN was grown via MBE under non equilibrium conditions to suppress the formation of p recipitating second phases. These phases could form from the impurities aggregating into nanoclusters and were problematic because they could be the origin of the ferromagnet ic ordering in the materi al.1 9 One of the most studied transition metal III nitrides is Mn doped GaN, which has had TC values ranging from as low as 10 K to an estimated 940 K.1 6,17,202 2 Overall, above a certain Mn atomic concentration, this materials exhibits paramagnetic behavior.16,17 20 ,2 2 Also, the magnetic properties of Mndoped GaN are not thermally stable without co-doping the material with oxygen up to 10 atomic percent.23 Research in DMS materials progressed towards other transition metals such as Fe, Co and Cr, with Cr emerg ing as an excellent candidate. Cr -doped GaN has shown ferromagnetic ordering above room temperat ure24,2 5 and up to 900 K.2 6 The magnetic moment increases with Cr concentration until the substitutional sites in the lattice are saturated and the Cr inco rporates interstitially which leads to degradation of the magnetic properties of the material.2 6 Cr -doped GaN exhibited similar ferromagnetic ordering before and after being annealed at 700C for 1 minute2 7, allowing this material to withstand high temperature device processing. Thin films of Cr -doped GaN displaying B/Cr atom.28

PAGE 22

22 Rare Earth Elements in III-V Semiconductors The rare earth metals, or lanthanide series elements, are the latest magnetic dopants to receive strong consideration. Gd, Tm, Eu and Er have all been studied in GaN since their excited states lead to emission lines throughout the spectrum from the ultraviolet (UV) through the v isible to the infrared (IR).2 9 Utilizing these emission lines with the wide band gap of GaN has led to research in developing these materials into t unable LEDs Eu doped GaN has been shown to exhibit room tempera ture ferroma gnetic ordering.30 Also, a paramagnetic -like phase has been reported to coexis t with a ferromagne tic like phase in Eu-doped GaN which has been attributed to div alent and trivalent Eu ions.3 1 M. Hashimoto et al. considered the Eu2+ ions to be responsible for the ferromagnetic behavior at room temperature but estimate that only 1.5% of the total Eu ions are divalent due to the intraatomic f -f transitions shown by photoluminescence (PL) data.3 2 The coexistence of paramagnetic and ferromagnetic ordering has also been seen in Er doped G aN grown via gas source MBE.3 3 Er-doped GaN grown v ia metal organic chemical vapor deposition (MOCVD) exhibits room temperature ferromagnetism with an increasing saturation magnetization with increasing Er concentration.3 4 Gd has been studied as a magnetic impurity in GaN through the incorporation process of mainly MBE growth35,36 37 39 with more current studies focusing on incorporation via MOCVD growth4 0 or ion implantation.41 4 4 Since the preferred valency of Gd in GaN is tr ivalent, this rare earth metal is ideal to substitute for the trivalent Ga atoms.4 5 Magnetic measurements of Gd -doped GaN grown via MBE have shown above room temperature ferromagnetism with a Gd conce ntration below 1016 cm3 and a colossal magnetic moment of up to 4000 B/ Gd atom .35 ,3 6 Th e significance of this

PAGE 23

23 observation becomes more apparent when comparing this colossal magnetic moment to BB/Gd atom). Gd doped GaN thin films grown via MOCVD has exhibited magnetization strengths as high as 110 emu/ cm3 and 500 emu/cm3 for Si -doped and Mgdoped Ga0.98Gd0.02N, respectively, without the presence of secondary phases or strain in the films.40 The discovery of an order of magnitude larger magnetic moment per Gd atom for Gd-implanted GaN as compared to epitaxially grown Gddo ped GaN41 increased the interest in ion implantation as an incorporation process. Gd3+ ions implanted into single -crystal GaN epilayers at total doses of 3 to 6 1014 cm2 and annealed at 700C to 1000C showed the formation of second phases such as Gd3Ga2, GdN and Gd from x -ray di ffraction (XRD) measurements .42 Annealing the implanted samples at 900C proved to be the optimal temperature for total magnetization, even though the ferromagnetic ordering could be attributed to Gd precipitates.4 2 Another annealing study showed that for GaN implanted with Gd at a dose of 2.4 1011 cm2 and then annealed at 900C, saturation magnetization was reduced as compared to a similar sample that was not annealed .43 Since XRD measurements showed the presence of Ga and N i nterstitials in the implanted GaN epilayers and the density of these defects were reduced upon annealing, Khaderbad et al. speculated that Gd could be inducing magnetic moment in either or both of the interstitials and giving rise to an effective colossal magnetic moment.43 These findings bring into question the role of defects and the host material crystal structure with regards to the ferromagnetic ordering due to the larger defect formation from the bombardment and incorporation of the large Gd ion into the GaN host material as compared to the defect formation associated with the non equilibrium

PAGE 24

24 epitaxial growth of Gd doped GaN. Further st udies into the type of defects present and their interactions with the implanted speci es and the host material crystal lattice would greatly assist in the development of a complete ferromagnetic mechanism model for these materials Literature Review of Theor etical Studies Since DMS materials were first studied, n umerous models have been suggested to explain the ferromagnetic phenomena occurring in these materials. As the research concerning these materials expanded and grew, one model would be refined or replaced with another model to better explain the nature of the ferromagnetism present and the interactions of the specific components necessary f or ferromagnetism. With the range of magnetic impurities, incorporation methods and host materials addressed in the previous section taken into consideration, a case-by -case approach to applying a spe cific ferromagnetic mechanism becomes appropriate. This section will consist of a brief review of the prominent models proposed for DMS materials. Free Carrier Mediated Model One of the first models dealing with the relationship between ferromagnetism and transition metal dop ants was proposed by Zener, who applied a mean-field approximation that defined the DMS material as an alloy containing random metallic atoms sitting substitutionally in the host material lattice.46 This model explained that the proclivity of the ferromagn etic alignment of d electron spins resulted from spin coupling between the incomplete d shell and conduction electrons or as an exchange interaction between carriers and localized spins .46 Dietl et al. later noted that Zener did not take into consideration the peripatetic nature of the magnetic electrons and the quantum (Friedel) o scillations of the electron spin polarization around the localized spins.10

PAGE 25

25 Building on this model, Dietl et al. showed that, for semiconductors, the effect of the Friedel oscillations average out to zero due to the mean distance between carriers being greater than that between spins, and therefore, the Zener model becomes equivalent to the Ruderman -Kittel -Kasuya Yosida (RKKY) interaction model, with the DMS material having a large carrier density ( on the order of 1020 cm3).10 Taking the RKKY interaction model into consideration, large hole densities were shown to drive a paramagnetic -ferromagnetic phase transition in II VI DMS materials.47 The RKKY interaction model was quest ionable for transition metal doped GaN since the degenerate electrons or holes are barely accessible in wide band gap semiconductors.48 Percolation Model The percolation model was developed due to the fact that, in DMS materials with low carrier concentrat ions, the indirect exchange interaction becomes short ranged and the mean -field theory becomes invalid.49 This model conjectures that, when the carrier concentration is much smaller than that of the magnetic impurities exchange interactions between the tw o result in their mutual polarization.50 This interaction results in the creation of a bound magnetic polaron consisting of one localized hole and a large number of magnetic impurities around the hole localization center with an effective radius that gr ows as the temperature decreases .51 The basis for the ferromagnetic mechanism comes from when these polarons grow as the temperature decreases and form clusters of ferromagnetic regions already existing at TC and becoming an infinite magnetic cluster at temperatures below TC.50 This model is limited by the necessities of lower temperatures and a carrier concentration that is much smaller than the magnetic

PAGE 26

26 imp urity concentration, which may not apply to the ferromagn etic mechanism occurring in certain DMS mat erials. Extension of Models to Rare Earth Elements The models described in the previous sections were mainly concerned with the role of transition metals in DMS materials. Attempting to apply the previous models to the lanthanide series becomes difficult s ince the additional f shell interactions make for a complicated modeling. Magnetophotoluminescence measurements of Gd-doped GaN supports a percolation-like model or long-range spin polarization of the host material matrix by Gd atoms.35,36 From ab initio b and structure calculations and symmetry arguments an electron-stabilized ferromagnetic model has been developed that explains that the observed colossal magnetic moments are directly related to the polarization of donor electrons .52 According to this model, the introduction of Gd causes localized states below the conduction band due to coupling between the s and f suborbitals. Spin polarization occurs due to these new states being filled by donor electrons, which may arise from the la rge concentration of int rinsic oxygen in these materials. Due to the focus on carriers presented in this model additional computational work is necessary based on more recent DMS material research into magnetic atom incorporation methods and host material crystal structures. Consideration of Defects in Models The role of defects with respect to ferromagnetic behavior becomes paramount when considering the electron-stabilized model mentioned in the previous section. In this model, defects should decrease the ferromagnetic ordering because they act as competition sites for electrons. Therefore, extensive studies and modeling have been pursued to understand the role of defects with regards to the ferromagnetic mechanism.

PAGE 27

27 For Gd -implanted GaN, the Gd bombardmen t process might be inducing the magnetic moment of Ga and/or N interstitials due to the long-range spin polarization of these defects.43 For the same long-range spin polarization to occur in epitaxially grown Gddoped GaN, a large c oncentration of these de fects (up to 1019 cm3) would be required. The majority of the earlier reports of colossal magnetic moments in Gd-doped GaN were exhibited in films grown via MBE, possibly due to this synthesis technique being a nonequilibrium method that allows for the f ormation of such defects during growth. The formation of these defects could be attributed to the larger atomic size of Gd compared to the Ga atom that constitutes the host material. X -ray linear dichroism and x -ray magnetic circular dichroism (XMCD) measurements of Gd in GaN showed that about 85% of the Gd goes to substitutional Ga sites and a small XMCD signal is detectable for Gd, which emphasizes the role of the GaN host matrix in the overall magnetic ordering.53 Liu et al. observed that th e strain exer ted on GaN films by Gd results in the creation of a high defect density due to a lattice mismatch ed substrate and the ion radii differ ential between the large Gd ion and the relatively smaller Ga ion .54 Total energy electronic -structure calculations have shown that a ferromagnetic interaction exists between Ga vacancies in the pre sence of Gd atoms resulting in an increase in magnetic moment with increasing Ga vacancies.55 As noted above, these Ga vacanci es are most likely formed during Gd incorporation, rather than existing intrinsically in the host matrix, and the formation of Gd atom -Ga vacancies defect complexes is energetically favorable.55 Additionally Ga vacancies located between two Gd atoms separ ated by about 15 favor ferromagnetic interaction compared to scarcely any interaction shown by the Gd atoms in the absence of vacancies.56 Although, the formation energy for Ga

PAGE 28

28 v acancies in the neutral charge state, which carries the 3 B per vacancy is the largest for all native defects in GaN at about 9 eV.57 More recently, nitrogen interstitials in GaN have been proposed as the source of ferromagnetism due to the localized nature of the N nonbonded ( p ) orbitals around these defects.58 C. Mitra et al. speculate that these orbitals could give rise to magnetism because magnetic moments arise from intra atomic Coulomb interactions and that the defect states have a sufficiently long-range tail to interact ferromagnetically with other N interstitials further away from the Gd.58 As this theoretical review clearly shows, a computational consideration of all the species present in DMS materia ls, including the types and concentrations of carriers and defects the magnetic impurity incorporation method and host m aterial crystal structure, is necessary for a determination of the exact ferromagnetic mechanism present in these materials.

PAGE 29

29 Figure 21. The additional of certain transition metal (TM) or rare earth (RE) ions can lead to alignment of spins and ferromagnetism in the compound semiconductor host material (top right ). However, instead of a random distribution of these ions (bottom left), clusters (bottom center) or magnetically active second phases may form (bottom right).6 Ferromagnetic Undoped Embedded Mnrich clusters in lattice Enhanced magnetic moments Random TM distribution Long-range ferromagnetic coupling Ferromagnetic phases segregate Examples: Ga x Mn y Mn x N y

PAGE 30

30 Figure 22 Predi cted Curie temperatures as a function of band gap for Group IV and III-V semiconductor mat erials with 5 % Mn and a hole concentration of 3.5x1020/cm3.10

PAGE 31

31 Figure 23 Predicted Curie temperatures as a function of lattice constant for Group IV and various compound semiconductor materials.6

PAGE 32

32 CHAPTER 3 EXPERIMENTAL PROCEDURES FOR SYNTHESIS, MAGNETIC ATOM INCORPORATION AND CHARACTERIZATION OF COMPOUND SEMICONDUCTOR MATERIALS Synthesis The majority of compound semiconductor materials are synthesized via one of tw o methods : molecular beam epitaxy (MBE) or metal organic chemical vapor deposition (MOCVD) These epitaxial growth methods possess a large research base and each technique is embraced in industry for specific material system and device fabrication purpose s. The main function of these techniques is in the growth of films that can be processed into devices. The general benefits of MBE over MOCVD include greater thickness control and film uniformity Unfortunately, as compared to MOCVD, MBE has a lower throug hput and longer downtime for the tool when repairs and source material replenishing is necessary MOCVD provides higher growth rates, which leads to higher throughput; an advantage th at has been fervently employed in industry. Unfortunately, MOCVD does not provide the layer thickness control granted by MBE, making this technique unsuitable for the synthesis of certain device structures. Molecular Beam Epitaxy The base MBE system used for this research is a Varian Gen II system consisting of three vacuum chambers separated by gate valves, as seen in Figure 3 -1. The first vacuum chamber acts as a loadlock for sample introduction and also contains a sample heater to remove any adventitious water vapor. The second vacuum chamber acts as a buffer chamber between the loadlock and the growth chamber to isolate the growth chamber from water vapor and oxygen that are introduced into the loadlock during sample loading. Th e last vacuum chamber serves as the growth chamber where the

PAGE 33

33 epitaxial growth occurs. The samples are moved from the loadlock to the vacuum chamber on a trolley attached to trolley tracks that is manipulated externally through the use of magnets Each vacu um chamber possesses pumps that can achieve the necessary pressure regime from atmosphere to high vacuum (HV) to ultra high vacuum (UHV) required for such tasks as sample loading and epitaxial growth. The growth chamber also contains two cryoshrouds which are filled with liquid nitrogen (LN2) befor e film synthesis and provide lower vacuum levels and a cleaner substrate for epitaxial growth. One cryoshroud envelops the source flange and prevents cross talk between sources when they are heated. The second cry oshroud envelops the substrate holder and heater assembly and protects the outer chamber from the high temperatures achieved by the substrate heater and absorbs any unused source materials. The growth chamber also possesses a source flange that contains ports for up to eight sources and a reflection high energy electron diffraction (RHEED) system. This RHEED system consists of an electron gun and a phosphor coated window and provides beneficial in situ feedback of the film growth process. This technique uti lizes an electron gun to produce an electron beam aimed at the surface of the epitaxial film at a 1 to 2 grazing angle (depicted in Figure 3-1) which undergoes diffraction with the top few monolayers of the surface and is reflected on the phosphor coated window. For the research purposes of this work, this growth monitoring technique gives information about the surface roughness and growth mode. The purity level of the source materials is an import ant consideration in MBE growth. The Group III and magnetic impurity sources are 99.99999% (7N) Ga and 99.99% (4N) Gd, respectively. The lower purity of Gd results from difficulties in

PAGE 34

34 disassociating rare earth metals. The Group V source is 99.9999% (6N) bottled nitrogen. The nitrogen flow is regulated by a mass f low controller which provides a constant gas flow The nitrogen then proceeds through an Oxford radio frequency (RF) plasma head. The plasma head operates at 13.56 MHz and ionizes the nitrogen entering the growth chamber to provide the desirable species fo r epitaxial growth. Metal -Organic Chemical Vapor Deposition The MOCVD system used for this research is a Veeco/Emcore P75 MOCVD vertical rotating disk reactor, diagrammed in Figure 32. In this epitaxial growth method, reactant gases are transported to the surface of the substrate that can be rotated up to 1500 rpm and heated to over 1100 C. For the purposes of this research, the Group III source is trimethylgallium (TMG) and the Group V source is ammonia (NH3). The n-type and p -type dopant sources are sila ne (SiH4) and bis -cyclopentadienyl magnesium (Cp2Mg) respectively Nitrogen is used as the carrier gas to deliver all of these precursors to the substrate in the reactor. The substrates used for grow th are (0001) oriented sapphire (Al2O3). The optical head portion of the MOCVD system is comprised of an emissivity compensated pyrometer This tool provides in situ growth observation and analysis by measuring the substrates thermal emission using a pyrometer and the surface refl ectivity using a reflectometer and then calculating the emiss ivity and substrate temperature. The surface reflectivity is utilized to determine what regime the growth process is in (i.e. nucleation layer ). The r eflectivity and temperature profile s produced by this tool c an be used to compare the epitaxial films that result from specific growth conditions and assist in optimizing the desirable characteristics of materials grown via this technique.

PAGE 35

35 Magnetic Impurity Incorporation Molecular Beam Epitaxy Gd is incorporated into GaN films via MBE growth carried out in the Varian Gen II system described above. A GaN buffer layer is grown on (0001) oriented sapphire in the MOCVD system described in the pre vious section. T o remove the native oxide that forms on the GaN film, a pretreatment process is performed. This process consists of the following steps: 3 minute s of immersion in hydrochloric acid (HCl) 25 minutes ultraviolet ozone (UV -O3 ) exposure, and 5 minute s of immersion in a buffer oxide etch The GaN buffer proceeds to a deionized water rinsing and drying under nitrogen before being indium mounted on a molybdenum block. The indium provides for uniform heat transfer between the block and GaN buffer during the MBE growth pro cess. The GaN is loaded into the MBE system which is pumped down, and transported to the growth chamber later The Ga and Gd source s are stored in Knudsen effusion cell s t hat can each be heated to a specific te mperature to produce a corresponding flux of the ir respective Ga and Gd molecular beam s The cell s are covered by externally controlled shutter s that provides for the abrupt starting and stopping of Ga and Gd atoms impinging on the surface of the GaN substrate. Once the Ga and/or Gd arrive at the sur face, these atomic species can undergo immediate incorporation via absorption, desorption or migration followed by absorption or desorption depending on the source cell temperature, substrate temperature and nitrogen plasma conditions. Therefore, the growt h of the Gddoped GaN is controlled by setting the substrate temperature and the Ga and Gd cell temperatures (Ga and Ga molecular beam fluxes ), and the nitrogen plasma conditions.

PAGE 36

36 Ion Implantation For this study, i on impl antation was pursued as the primary magnetic impurity incorporation technique. This technique provides numerous advantages such as allowing for uniform, reproducible doping and incorporating the exact amount of the desired species in the host material. Ion implantation has been utilized in industry for shallow doping and selective doping with masks during material processing for device applications. For this research, t he Gd and Si ion implantation processes were completed in a Varian 350D ion implanter by Cuttingedge Ions. The G d source for the ion implanter wa s solid GdF and the Si source wa s gaseous silicon tetrafluoride ( Si F4). The ion beam was oriented 7 degrees off from perpendicular to the surface of the implanted materials during implantation and the process was performed at room temperature. Characterization Superconducting Quantum Interference Device Magnetometry All of the magnetic measurements in this research were completed using either a Quantum Device Magnetic Properties Measurement System model MPMS 5S or MPMS XL S uperconducting Quantum Interference Device (SQUID) magnetometer. For SQUID sample preparation, an approximately 7 mm2 sample is cleaved and all indium is removed from the back of the substrate if the sample was grown via MBE. The majority o f the magnetic m easurements involved taking magnetization versus applied magnetic field curves at a constant temperature. For these measurement s the sample was first centered in the analysis area under an applied magnetic field and then demagnetized under a decreasing os cillating magnetic field. The sample is then submitted to a sweeping magnetic field with multiple data points taken at interval field strengths for

PAGE 37

37 statistical averaging. The sample is finally submitted to an applied field increased up to 5 T (50,000 Oe) t o measure the diamagnetic contribution of the substrate. The diamagnetic contribution is subtracted from the data to allow for accurate comparison of the magneti c properties of the samples since they may possess different substrate materials When analyzin g the magnetization versus applied field curve, hysteresis indicates ferromagnetic ordering. If the sample has a hysteretic magnetization versus applied field curve, saturation magnetization (MS) and coercivity (HC) values can be determined, assuming a hig h enough magnetic field was used for complete saturation. To confirm ferromagnetic behavior, a magnetization versus temperature or Field Cooled/Zero Field Cooled (FC/ZFC) measurement may be taken. In this measurement, the sample is cooled down to 10 K under zero field and the samples magnetization is measured in 10 K intervals ( Z FC) Next, the sample is heat ed back up to room temperature under zero applied field and the process is repeated under an applied field (FC) with the samples magnetization measured in the same 10 K intervals. For both of these types of measurements, the sample can be oriented with the applied field either parallel or perpendicular to the sample surface. The different sample surface/applied field orientations can provide additional information about the anisotropy of the magnetic properties, especially with regards to specific crystal structures. X-r ay Diffraction Powder X -ray diffraction (XRD) was utilized to structurally analyze the samples for secondary phase formation due to the incorporation of Gd and/or Si during MBE growth or ion implantation. These measurements were completed with a Philips APD 3720 system that utilizes a copper (Cu) x -ray source. This source mainly emits Cu K 1 x -rays with a 1.54056 wavelength, but K 2 and K x -rays are emitted as well. In XRD, the

PAGE 38

38 incident x -rays undergo constructive and destructive interference when they interact with the repeating planes of the crystalline sample, following Braggs Law: = 2 Where d is the atomic plane sp acing and is the angle between the incident x ray beam and an atomic plane. The Philips system contains a photomultiplier tube that measures the intensity of the diffracted x rays as a function of 2 the angle between the incident and diffracted x -rays. The measurement provides a plot of diffracted x -ray intensity versus 2 which shows peaks that correspond to specific crystal planes in the sample. These peaks can be analyzed to determine if they come from the semiconductor host material, substrate or undesirable dopant induced phases. Photoluminescence Photoluminescence (PL) measurements were taken to determine t he presence and type of defects in the implanted materials. These measurements occur from the emission of light that results from the optical st imulation of a sample material More s pecifically, the sample absorbs light of the wavelength of the excitation source which, for a semiconductor material, promotes an electron from the valence band to the conduction band. The electron subsequently experiences non-radiative internal relaxation and moves to a more stable excited energy level. After a certain amount of time in the excited state, the electron moves to a ground state. If the sample material luminesces, some or all of the energy related to the move from an excited to the ground state is released as light This emitted light can be detected as photoluminescence and the intensity of the emitted light can be measured as a function of wavelength (energy) to determine the presence of specific species such as dopants and defects in the sample material.

PAGE 39

3 9 PL spectra were examined for some of the GaN samples both before and after implantation. GaN samples were also analyzed before and after annealing. For these measurements a HeCd laser of wavelength 325.13 nm (3.81 4 eV energy) was used as the excitation source which is above the band edge of GaN ( 3.39 eV). The photoluminescence from the samples passed through a grating monochromator and a photomultiplier was used for detection. A low temperature stage consisting of a He closed system coldfinger was used fo r PL measurements down to ~ 17 K.

PAGE 40

40 Figure 31 Diagram of Varian Gen II MBE system and component vacuum chambers.

PAGE 41

41 Figure 3 2 Diagram of Veeco/Emcore P75 MOCVD vertical rotating disk reactor MOCVD system

PAGE 42

42 CHAPTER 4 IO N IMPLANTATION OF GADOLINIUM AND SILICON IN GALLIUM NITRIDE Gd -doped GaN films grown via molecular beam epitaxy ( MBE ) hav e demonstrated ferromagnetic ordering with a TC above room temperature.35,36,37 39 Certain studies have shown a magnetic moment per Gd atom as high as 4000 B.35,36 Dalpain et al. postulated that the ferromagnetic phase can be stabilized by introducing electrons while the coupling between only Gd atoms is antiferromagnetic .52 Based on that conjecture, GaN has been co doped with both Gd and Si in MBE grown films to study the effects that a higher carrier density and the presence of a shallow donor have on ferromagnetism.59 61 The Gd concentration in the co-doped GaN films ranged from below the detection limit of secondary ion mass spectroscopy (on the order of < 1017 at./cm3)59, 60 to 8.9%61 with all co doped samples exhibiting room temperature ferromagnetism. The MS value of a Gd and Si -co doped GaN film (1046 emu/cm3) was seven times larger than that of a similar Gd doped GaN film with both films having the same Gd concentration.61 As a dopant in wurtzite GaN, Si behaves as a shallow donor with an ionization energy between 0.12 and 0.20 eV.6 2 Gd -focused ion beam -implanted GaN layers were also found to display ferromagnetism with an TC above room temperature, but also with an order of magnitude larger effective magnetic moment per Gd atom than that of epitaxial Gddoped GaN .41 Even though studies i n Gd-implanted GaN has been pursued no co implantation studies have been attem pted to determine the effects that additional carriers (and a dditional defect formation) have on the magnetic properties of the implanted samples. Theref ore, the focus of this study is on the effect s that the implant species: Si, Gd and both Gd and Si together in undoped and doped GaN ha ve on the

PAGE 43

43 ferromagnetic behavior of this material. GaN thin films grown via MBE and metal organic chemical vapor deposition (MOCVD) will be explored as the implant target materials. The goal of this work is to develop a mechanism of f erromagnetism that can be tested and exploited in the development of a spintronic device. MOCVD u -GaN film of undoped GaN was grown via MOCVD on a 2inch sapphire substrate wafer. The u-GaN film was then cleaved into quarters and one quarter was se t aside as a reference sample. T he quarters were then sent off for ion implantation. The ion beam was oriented 7 degrees off from perpendicular to the surface of the target materials during all implantations to avoid channeling and the process was performed at room temperature. One quart er was implanted with Si ions of energies 5 keV and 40 keV and respective implantation doses of 8.00 1011 and 3.60 1012 cm2. The second quarter sample was implanted with Gd ions of energy 155 keV and an implantation dose of 2.75 1010 cm2. The last quarter sample was implanted with both of the Si doses and the Gd dose. Figure 41 displays an implantation profile for the combined Si doses in GaN and Figure 42 shows the implantation profile for the Gd dose in GaN. The energies and doses used for the i mplantations were selected based on achieving a Si concentration of approximately 1.00 1018 cm3 and a Gd concentration of approximately 1.00 1016 cm3 at about 300 below the sample surface from simulations using the Transport of Ions in Matter (TRIM ) program.6 3 Using data obtained from this program to plot the predicted ion ranges, the Si atoms and the Gd atoms were mainly distributed about 95 nm and 80 nm from the sample surface, respectively. The peak Si concentration was determined to be 8.81 1017 cm3 at a depth of about 290 below the surface based on data provided by TRIM Similarly, t he peak Gd

PAGE 44

44 concen tration was determined to be 9.73 1015 cm3 at a depth of about 300 below the surface based on data provided by TRIM Non -Implanted To create a baseline to compare the implanted u-GaN samples to, structural and magnetic characterization of the nonimplanted MOCV D u -GaN reference sample w as completed first. Figure 4 -3 shows the XRD spectrum for the nonimplanted u-GaN sample with the labeled peaks corr esponding to either the GaN film or the sapphire substrate. Magnetic measurements were performed at both 10 K and 300 K for two different sample surface/magnetic field orientations. As diagrammed in Figure 4 4 t he field was applied either parallel to the sample surface ( perpendicular to the c axis of the GaN wurtzite crystal structure), henceforth known as the parallel orientation, or perpendicular to the sample surface ( parallel to the c axis of the wurtzite crystal structure), henceforth known as the perpendicular orientation. Before the magnetization of each sam ple was measured in a differe nt orientation and/ or temperature, the sample was demagnetized using a decreasing oscillating magn etic field For all magnetic measurements, t he diamagnetic contribution of the substrate was subtracted from the measured magnetization values and these values were then normalized by the area of the sample. Figure 4 -5 and Figure 4 -6 show the magnetization versus applied magnetic field curves for the nonimplanted MOCVD u-GaN in the parallel orientation at 10 K and 300 K, respectively. Surprisingly, this sample exhibited ferromagnetic ordering at both low and room temperatures without the presence of a know n magnetic impurity. There were no expected magnetic impurities in the sapphire substrates used for the GaN MOCVD growth and no known magnetic impurities present in the Ga and N precursors in the MOCVD growth system. Concerning the other orientation, Figur e 4 -7 and Figure

PAGE 45

45 4 -8 show the magnetization versus applied magnetic field curves for the non-implanted MOCVD u -GaN in the perpendicular orientation at 10 K and 300 K, respectively. The non-implanted u-GaN also shows ferromagnetic ordering at both temperatures in this orientation without any magnetic impurities. Comparing the approximate magnetization values at +1000 G for the sample in the two orientations shows about a 9 -times larger signal in the perpendicular orientation over the parallel orientation. Th is enhancement in magnetic response exhibited parallel to the c axis of the non-implanted MOCVD u -GaN will be henceforth known as the anisotropic enhancement effect (AEE) of the magnetic signal. Th e possible mechanisms for this observed ferromagnetism and the origin of the AEE will be further discussed at th e conclusion of this section. Si -Implante d After Si implantation, the quarter wafer was cleaved into smaller sizes suitable for structural and magnetic characterization techniques. The Si implanted MOCVD u -GaN sample was not subjected to any annealing treatments before the structural and magnetic measurements were taken. Figure 4 9 shows the powder XRD scan for the Si implanted MOCVD u-GaN sample compared to the XRD scan of the non -implanted MOCVD u -GaN reference sample. An increase in the background signal is exhibited in the plots, but there are no additional peaks seen for the Si implanted plot indicating the absence of complex or secondary phase formation (within the XRD detection limits). M agnetic measurements were performed at both 10 K and 300 K in both the parallel and perpendicular orientations. Figure 410 and Figure 4-11 show the magnetization versus applied magnetic field curves for the Si -implanted MOCVD uGaN in the parallel orientation at 10 K and 300 K, respectively. This sample exhibits ferromagnetism at 10 K, but not at 300 K. At room temperature, the sample shows diamagnetic behavior

PAGE 46

46 when analyzed in the parallel orientation. Figure 4-12 and Figure 413 show the magnetization versus appli ed magnetic field curves for the Si -implanted MOCVD u-GaN in the perpendicular orientation at 10 K and 300 K, respectively. The sample showed ferromagnetic ordering at both temperatures in this orientation. At 10 K, the magnetiz ation at +1000 G is about 6times larger for the perpendicular orientation as compared to the parallel orientation. The AEE of the signal is of the same order of magnitude as that exhibited by the non -implanted MOCVD u-GaN therefore the Si does not appear to be a major contributing factor to the ferromagnetism. Neither of these samples was expected to contain any magnetic impurity, which raises the question of the species present in these materials to justify the observed ferromagnetic ordering Further analysis into the Gd implantat ion and the Gd and Si co -implantation may help in elucidating the mechanism of ferromagnetism occurring in these materials. Gd -I mplante d The Gdimplanted MOCVD u-GaN sample was subjected to the same characterization d iscussed for the previous samples Powd er XRD scans of the Gd implanted MOCVD u-GaN and the non -implanted MOCVD u-GaN reference sample are shown together in Figure 4 -14 for comparison Even though the background increases for the Gd implanted u-GaN sam ple, there are no additional peaks in the G d implanted sample that can be attributed to any complexes or secondary phases (within the detection limits of the XRD ). Therefore, the ferromagnetism measured in these samples does n ot result from the creation of secondary phases due to implantation, but from within the system comprised of the implanted material, implant species and any defects previously present or introduced as part of the implantation process.

PAGE 47

47 Magnetic measurements of the Gdimplanted MOCVD u-GaN film were performed at both 10 K and 300 K for the two different sample surface/magnetic field orientations. For the parallel orientation, t he magnetization versus applied magnetic field curves at 10 K a nd 300 K are shown in Figure 415 and Figure 4-16 respectively. The sample exhibits above room temperature ferromagnetic ordering even though the data points for the hysteresis curve at 300 K show more uncertainty at higher applied field strengths. For the perpendicular orientation, the magnetization versus applied magnetic field curves at 10 K and 300 K are shown in Figure 4 -17 and Figure 4 -18, respectively. Similarly to the results for the parallel orientation, the Gd-implanted MOCVD u -GaN also exhibits low and room temperature ferromagnetic ordering, but with less unce rtainty in the hysteresis curve at 300 K When comparing the hysteresis curves for the parallel and perpendicular orientations at each temperature, the AEE is seen in the approximate 13-times larger and 36 -times larger magnetizations at 1000 G exhibited by the sample in the perpen dicular as compared to the parallel orientation for 10 K and 300 K, respectively. The larger AEE seen in the Gdimplanted u-GaN over the non -implanted and Si -implanted u-GaN may be attributed t o a coupling between the defects and Gd atoms introduced into t he material system due t o the ion implantation process. Gd and Si-Co I mplanted Similar structural and magnetic characterization was performed on the Gdand Si co implanted MOCVD u-GaN film Powder XRD scans of the Gd and Si -co implanted MOCVD u -GaN and the non-implanted MOCVD u-GaN reference sample are shown together in Figure 4 19 for comparison. Once again, although the background increases for the co -implanted u-GaN s ample, there are no additional peaks in this sample that

PAGE 48

48 can be attributed to any com plexes or secondary phases (within the detection limits of the XRD ). Concerning the magnetic measurements, f or the parallel orientation, the magnetization versus applied magnetic field curves at 10 K and 300 K are shown in Figure 420 and Figure 4-21 resp ectively. This sample shows ferromagnetic behavior at both temperatures, but the error in the data points at 300 K show more uncertainty For the perpendicular orientation, the magnetization versus applied magnetic field curves at 10 K and 300 K are shown in Figure 4-22 and Figure 423 respectively. In this orientation, the sample exhibits an above room temperature ferromagnetism with both hysteresis curves being nearly identical at 10 K and 300 K. Due to the strong similarity between the hysteres is curves at the two temperatures a magnetization vs. temperature plot was taken (Figure 2-24). The Curie temperature (TC) for this material is estimated to be at about 340 to 350 K since the Field Cooled (FC) and Zero Field Cooled (ZFC) lines intersect approximately in that temperature range. The AEE is even more pronounced in the co -implanted sample when compared to the Si -implanted and Gd -implanted samples as there is an approximate 130times and 80times larger magnetization at 1000 G in the perpendicular as compared to the parallel orientation for 10 K and 300 K respectively Figure 425 shows a comparison of the magne tization versus applied field curves for the Si Gd and Gdand Si -co implanted MOCVD u-Ga N samples at 10 K for the perpen dicular orientation. Figure 4-26 shows a similar comparison for the implanted samples at 300 K. When comparing the magnetizations at + 1000 G for the Gdimplanted to the co -implanted MOCVD u-GaN i n the perpendicular orientation, there is a n approximate 4 times larger value for the co-implanted sample

PAGE 49

49 over the Gdimplanted sample at both temperatures Therefore, some species associated w ith the addition al Si implants has a coupling effect with the Gd species as exhibited by the large magnetic signal in the co -implanted sample and warrants a deeper investigation into the species introduced during the implantation process. Mechanism of Ferr omagnetism Due to the non -implanted MOCVD u -GaN sample exhibiting ferromagnetism, a consideration of the native species present in this material must be explored to assess the origin of the magnetic ordering. Concerning the growth conditions of this sample, a GaN nucleation layer was grown on a sapphire substrate via MOCVD Based on the mismatch between the lat tice constant of sapphire (4.759 ) and the wurtzite GaN (3.189 ), a large threading dislocation density is expected to propagate along the c axis o f the film. The u -GaN layer was grown on a GaN nucleation layer at a growth rate of 1.80 m/hr for 250 minutes and the NH3 flow during growth was set to maintain a V/III ratio of ~ 6000 indicating growth under N -rich conditions In addition to the dislocations, defects such as vacancies and interstitials are expected to form in the GaN film due to the noneq uilibrium growth process Due to the N rich growth conditions, a high density of nitrogen interstitials (Ni) is expected to b e created in the GaN. Both gallium vacancies (VGa)55,6 4 and nitrogen interstitials (Ni)58 have been proposed as possible species that contribute to ferromagnetic ordering. The concentration of these native point defects in this sample are expected to be much lower than the defect concentrations formed during the implantation process for the implanted samples. Therefore, a consideration of the defect formation in the implanted samples will be discussed next for further insight into the ferromagnetic mechanis m.

PAGE 50

50 Due to the combined effect of incorporating the relatively large G d ions and the larger doses of Si ions the implantation process creates a profusion of defects. This process introduces point defects such as vacancies and interstitials into the host ma terial, with the vacancies forming mainly near the surface and the interstitials forming deeper into the implanted material at the end of range of the implanted ions Specifically, the size differential between the implanted Gd ion radius and the host Ga i on radius results in the existence of a large vacancy concentration.54 According to the TRIM modeling of defect formation during implantation: Displacements = Vacancies + Replacement Collisions Where Displacements are defined as how many target atoms were set in motion in the cascade with energies above their displacement energy during the implantation process. Repla cement Collisions are defined as the following: if a moving atom strikes a stationary target atom and transfer more than its displacement ener gy to it, and the initial atom, after the collision, does not have enough energy to move onwards, and is the same element as the atom it struck, then it just replaces the atom in the target and no vacancy is created. This replacement collision mechanism can reduce the total vacancies created by up to 30%. Also, according to TRIM: Vacancies = Interstitials + (Atoms which leave the target volume) Therefore, when a recoil atom stops, and it is not a replacement atom, than that atom becomes an interstitial. For the purpose of simplifying the model, the assumption will be made that for every vacancy created during implantation, an interstitial is created ( due to negligibl e antisite formation and a negligible loss of atoms from the target surface) Figure 427 shows the prevalence of vacancy creation near the surface region and

PAGE 51

51 compares the v acancy concentration profiles due to the Gd implant and the Si implants to the Gd and Si implant profiles The additional Si doses introduce over an order of magnitude more vacancies (and interstitials) than those cr eated by the Gd implant alone. Based on the hysteresis curves for the implanted samples, the approximate 4 times larger magnetic signal in the co implanted u-GaN compared to the Gdimplanted u -GaN demonstrates that there is a coupling factor between the Gd atoms and the additional defects introduced during implantation. This coupling factor can also be seen in the approximate 3-times to 10-times larger magnetic signal in the Gd -implanted uGaN compared to the non -implanted and Si -implanted u-GaN samples. The defects formed due to the Si implantation alone do indeed show ferromagnetic ordering in the perpend icular orientation, but a significantly larger magnetic signal is demonstrated when Gd is present for both the Gd-implanted and co-implanted samples. Further studies in the role of native defects in doped GaN films will be the next step to help develop the ferromagnetic mechanism. Photoluminescence (PL) measurements were utilized to determine the defect species present in the nonimplanted and co implanted GaN sample s For these measurements a HeCd laser of wavelength 325.13 nm (3.814 eV energy) was used as the excitation source, which is above the band edge of GaN (3.39 eV). For low temperature measurements down to ~17 K, a stage comprised of a He closed system coldfinger was used. Figure 428 and Figure 429 show the PL spectra at 17 K and 300 K, respectiv ely, for the co implanted MOCVD u-GaN sample before and after implantation. The non -implanted MOCVD u-GaN clearly exhibits both the yellow luminescence (YL) and blue luminescence (BL) bands. The YL band is typically

PAGE 52

52 associated with gallium vacancy (VGa)-re lated defects and the BL band is associated with transitions from shallow donors (low temperatures) or the conduction band (elevated temperatures) to deep acceptor levels (ionization energy of about 0.34 to 0.40 eV).65 After the implantation process, the l arge reduction in signal from the YL region indicates a decrease in luminescence from the VGa-related defects. This decrease in the YL band corresponds with the largest ferromagnetic signal of the implanted samples exhibited by the co -implanted MOCVD u-GaN Therefore, decreasing the concentration of the anion vacancy (VGa) appears to increase the ferromagnetic ordering in GaN. MOCVD p -GaN A previously grown film of MOCVD p GaN was also considered as a target material for the ion implantation study. The p-ty pe dopant source was bis cyclopentadienyl magnesium (Cp2Mg) during MOCVD growth. The p GaN film was cleaved into halves and one half was set aside as a reference sample. The other half was implanted with the Gd dose ( 2.75 1010 cm2 at 155 keV) at room temperature with the ion beam oriented 7 degrees off from perpendicular to the surface of the target material to avoid channeling. The implantation profile for this sample is expected to be the same as the one shown in Figure 42. Non -Implanted Similar to the study for the MOCVD u -GaN, structural and magnetic characterization was completed for the non -implanted MOCVD p -GaN to create a baseline for comparison to the Gdimplanted p -GaN Figure 4 30 shows the XRD spectrum for the non -implanted p -GaN sample with the labeled peaks corresponding to either the GaN thin film or the sapphire substrate. Magnetic measurements were performed at both 10 K and 300 K for two different sample surface/magnetic field

PAGE 53

53 orientations. Before the magnetization of each s am ple was measured in a differe nt orientation and/or temperature, the sample was demagnetized using a decreasing oscillating magnetic field The diamagnetic contribution of the substrate was subtracted from the measured magnetization values and these value s were then normalized by the area of the sample. Figure 4 -31 and Figure 432 show the magnetization versus applied magnetic field curves for the non-implanted MOCVD p -GaN in the parallel orientation at 10 K and 300 K, respectively. Due to the uncertainty in some of the magnetization values at specific applied fields, a definitive case for ferromagnetic ordering in this sample analyzed in the parallel orientation cannot be made. Figure 433 and Figure 4 34 show the magnetization versus applied magnetic fiel d curves for the non-implanted MOCVD p -GaN in the perpendicular orientation at 10 K and 300 K, respectively. The non-implanted p -GaN clearly demonstrates ferromagnetism at both temperatures and the AEE is seen by the approximate 99times larger magnetic signal at +1000 G in the sample analyzed in the perpendicular orientation compared to the signal for the parallel orientation at 10 K. The nonimplanted p -GaN exhibits an AEE much larger than that seen in the non -implanted u-GaN which indicates that the native defects in p -GaN may play an important role in the ferromagnetic ordering. Gd -Implanted The Gdimplanted MOCVD p -GaN sample was not subjected to any annealing treatments before the structural and magnetic meas urements were taken. Figure 4-35 shows the powder XRD scan for the Gdimplanted MOCVD p -GaN sample compared to the XRD scan of the nonimplanted MOCVD p -GaN reference sample. An increase in the background signal is exhibited in the Gd-implanted plot but there are no additional peaks seen in the Gd-implanted plot indicating the absence of complex or secondary

PAGE 54

54 phase formation (within the XRD detection limits). Once again, magnetic measurements were performed at both 10 K and 300 K in both the parallel and perpendicular orientations Figure 436 and Figure 4-37 show the magnetization versus applied magnetic field curves for the Gd-implanted MOCVD p -GaN in the parallel orientation at 10 K and 300 K, respectively. This sample displays ferromagnetism in this orientati on at both temperatures, but the hysteresis curve at 300 K shows some uncertainty which indicates that the TC of this material might be near room temperature. Figure 4-38 and Figure 439 show the magnetization versus applied magnetic field curves for the G d implanted MOCVD p -GaN in the perpendicular orientation at 10 K and 300 K, respectively. At both temperatures, MOCVD p -GaN demonstrates ferromagnetism. The AEE is demonstrated once again by the approximate 44 -times larger signal at +1000 G exhibited by the sample in the perpendicular vs. the parallel orientation. Mechanism of Ferromagnetism For MOCVD p -GaN, the Gdimplanted sample showed a magnetic signal about twice that seen in the non-implanted sample when analyzed in the perpendicular orientation. Simi lar to the case for the MOCVD u -GaN, the Gd implantation has an effect on the magnetic ordering. Even though, the non-implanted p -GaN exhibited a larger signal than the non implanted u-GaN in the perpendicular orientation. Based on this observation, a cons ideration of the native defects in p -GaN must be pursued. Figure 4 -40 shows a diagram of formation energies as a function of Fermi level for native point defects in GaN.57 Since the zero of the Fermi level corresponds to the valence band maximum, the most likely point defect to form in p -GaN is nitrogen vacancies (VN). Therefore, VN may be the source of the ferromagnetic ordering in p -GaN Nitrogen interstitials (Ni) should also be considered since the p -GaN film is grown under N -rich

PAGE 55

55 conditions via MOCVD. Theoretical models have shown that Ni in octahedral sites next to Gd are a likely source of defect -induced magnetism.58 S pin orbit coupling between these defects and the implanted Gd may result in the large magnetic signal exhibited by the Gdimplanted p -G aN. Previous work in Gd -doped GaN thin films grown via MOCVD has exhibited magnetization strengths as high as 500 emu/cm3 for Mg -doped Ga0.98Gd0.02N40, which correlates nicely with the approximate 30 emu/ cm3 signal (magnetization normalized by implanted v olume) for a peak Gd concentration of 9.73 1015 cm3 demonstrated by the Gdimplanted MOCVD p -GaN in this work Based on this study, an interaction between the defects in Gd -implanted MOCVD p GaN (most likely anion vacancies and interstitials) and the incorporated Gd results in enhanced ferromagnetic ordering compared to non-implanted MOCVD p -GaN PL measurements were utilized to determine the defect species present in the non-implanted and Gd-implanted p-GaN samples. Figure 4 41 and Figure 4-42 show the PL spectra at 17 K and 300 K, respectively, for Gd -implanted p-GaN before and after implantation. As expected, the red luminescence (RL) and ultraviolet luminescence (UVL) bands exhibited by the non-implanted MOCVD p -GaN are decreased due to the defect for mation and lattice damage done by the implantation process. The RL band is typically observed in heavily Mg -doped GaN and is associated with transitions involving deep donors (i.e. VN, Mg -VN complex VN-H complex Mgi, and MgN) and the MgGa acceptor, while the UVL band is typically related to the MgGa acceptor.65 The shoulder of the UVL band can be attributed to the blue luminescence (BL) band. This band is caused by a transition between a deep localized donor and a Mgrelated acceptor state.65

PAGE 56

56 MOCVD n -GaN A previously grown film of MOCVD n GaN was considered as another target material for the ion implantation study. The n-type dopant source was silane (SiH4) during MOCVD growth. The n -GaN film was cleaved into halves and one half was se t asi de as a reference sample. The other half was implanted with the Gd dose ( 2.75 1010 cm2 at 155 keV) at room temperature with the ion beam oriented 7 degrees off from perpendicular to the surface of the target material to avoid channeling The implantation profile for this sample is expected to be the same as the one shown in Figure 42 for Gd -implanted GaN Non -Implanted Similar to the experimentation for the MOCVD u-GaN and p -GaN, structural and magnetic characterization was completed to create a baselin e for comparison to the Gd imp lanted n-GaN sample. Figure 4 -43 shows the XRD spectrum for the non -implanted n -GaN sample with the labeled peaks corresponding to either the GaN thin film or the sapphire substrate. Magnetic measurements were performed at both 10 K and 300 K for both the parallel and perpendicular orientations. Before the magnetization of each sam ple was measured in a differe nt orientation and/or temperature, the sample was demagnetized using a decreasing oscillating magnetic field The diamagnetic contribution of the substrate was subtracted from the measured magnetization values and these values were then normalized by the area of the sample. Figure 4-44 and Figure 445 show the magnetization versus applied magnetic field curves for the nonimplanted MOCVD n-GaN in the parallel orientation at 10 K and 300 K, respectively. The nonimplanted n-GaN doe s not show ferromagnetism when analyzed in the parallel orientation but appears to show diamagnetic behavior based on th e hysteresis curves.

PAGE 57

57 Figure 446 and Figure 4-47 show the magnetization versus applied magnetic field curves for the non-implanted MOCVD n -GaN in the perpendicular orientation at 10 K and 300 K, r espectively. The non implanted n-GaN clearly demonstrates ferromagnetism at both temperatures and the AEE is exhibited by the ferromagnetic ordering demonstrated in the perpendicular orientation and diamagnetic behavior demonstrated in the parallel orientation. Also, in the perpendicular orientation, the nonimplanted n-GaN has a magnetization value of about 2 to 2.38 105 emu/cm2 at +1000 G which is very similar to the magnetization measured fo r non-implanted u-GaN (~ 2 to 2.73 105 emu/cm2), while both magnetizations are dwarfed by the value seen in nonimplanted p -GaN (~ 1.20 104 emu/cm2). Clearly, based on these results, the native defects present in p -GaN contribute more to the ferromagnetic ordering in this material compared to the native defects in u-GaN and n-GaN. Gd -Implanted The Gdimplanted MOCVD n-GaN sample was not subjected to any annealing treatments before the structural and magnetic measurements were taken. F igure 4-48 shows th e powder XRD scan for the Gdimplanted MOCVD n-GaN sample compared to the XRD scan of the nonimplanted MOCVD n-GaN reference sample. An increase in the background signal is exhibited in the Gd-implanted plot and even overshadows some of the sapphire and G aN peaks, but there are no additional peaks seen in the Gdimplanted plot which indicates the absence of complex or secondary phase formation (within the XRD detection limits). M agnetic measurements were performed at both 10 K and 300 K in both the parallel and perpendicular orientations Figure 4 -49 and Figure 450 show the magnetization versus applied magnetic field cur ves for the Gd -implanted MOCVD n -GaN in the parallel orientation at 10 K and 300 K, respectively. This sample

PAGE 58

58 displays ferromagnetism in this orientation at both t emperatures Figure 4-51 and Figure 4 -52 show the magnetization versus applied magnetic field curves for the Gdimplanted MOCVD n -GaN in the perpendicular orientation at 10 K and 300 K, respectivel y. At both temperatures, Gdimpla nted n-GaN demonstrates ferromagnetism in the perpendicular orientation. The AEE is much smaller for this sample relative to the implanted u-GaN and p -GaN since there is only an approximate 4-times larger signal at +1000 G exhibited by the sample in the pe rpendicular vs. the parallel orientation. Mechanism of Ferromagnetism When comparing the hysteresis curves of Gd implanted n -GaN to non-i mplanted n -GaN in the perpendicular orientation, they are almost identical for both 10 K and 300 K. The Gd implantation does not exhibit any spin orbit coupling with the native defects in this material to produce a larger magnetic moment in the implanted n -GaN as compared to non -implanted n -GaN Based on Figure 4-40 the native defect most likely to form in n -GaN is VGa. T herefore, these native defects are expected to contribute less in mediation of the fer romagnetic ordering in n-GaN when compared to VN in p -GaN This result is consistent with the computation work of C. Mitra et al. who explained that even though VGa beco me favorable in ntype material in the triple negative charge state, this charge state does not carry magnetic moment because the minority spin states would become filled .58 MBE Gd Doped GaN A previously grown film of MBE Gddoped GaN (henceforth known as GaGdN) was also considered for the implant study. This time, the Gd had already been incorporated during MBE growth and Si was the sole implant species. Two Si implants were performed at room temperature with the ion beam oriented 7 degrees off from

PAGE 59

59 perpendicular to the surface of the target material (to avoid channeling) one with a dose of 8.00 1011 cm2 at an energy of 5 keV and the other with a dose of 3.60 1012 cm2 at an energy of 40 keV. Figure 41 shows the expected Si implantation profile for these doses in GaN Non -Implanted Magnetic measurements were completed for the nonimplanted MBE GaGdN to provide a baseline for comparison to the Si -implanted MBE GaGdN. These measurements were performed at both 10 K and 300 K for both the parallel and perpendicular orientations. Before the magnetization of each sam ple was measured in a differe nt orientation and/or temperature, the sample was demagnetized using a decreasing oscillating magnetic field. The diamagnetic contribution of the substrate was subtracted from the measured magnetization values and these values were t hen normalized by the area of the sample. Figure 453 and Figure 4-54 show the magnetization versus applied magnetic field curves for the non-implanted MBE GaGdN in the parallel orientation at 10 K and 300 K, respectively. Consistent with previous measurements of this material in the parallel orientation, the non -implanted GaGdN demonstrated ferromagnetism and a magnetization at +1000 G of about 1.60 105 emu/cm2. Figure 4-55 and Fig ure 456 show the magnetization versus applied magnetic field curves for the nonimplanted MBE GaGdN in the perpendicular orientation at 10 K and 300 K, respectively. Both of these hysteresis curves display ferromagnetic behavior, with the curve at 300 K s howing a larger magnetic signal than the curve at 10 K. There does not appear to be a clear explanation for this behavior, so the plot at 300 K may be the result of an error in the SQUID measurements. Nonetheless, the plot at 10 K demonstrates the AEE as t here is an about 4-times larger magnetic signal at

PAGE 60

60 +1000 G for the sample in the perpendicular orientation compared to the signal in the parallel orientation. Si -Implanted The Si implanted MBE GaGdN sample was not subjected to any annealing treatments before the magnetic measurements were taken. Magnetic measurements were performed at both 10 K and 300 K in both the parallel and perpendicular orientations. Figure 45 7 and Figure 4-58 show the magnetization versus applied magnetic field curves for the Si -implanted MBE GaGdN in the parallel orientation at 10 K and 300 K, respectively. The Si -implanted GaGdN sample exhibits ferromagnetism in the hysteresis curve at 10 K, but the curve at 300 K shows that this sample is approaching or has approached th e Curie temperature. Figure 459 and Figure 4-60 show the magnetization versus applied magnetic field curves for the Si -implanted MBE GaGdN in the perpendicular orientation at 10 K and 300 K, respectively. Both of these hysteresis curves show clear ferromagnetism and the AEE is again demonstrated by an approximate 31 to 38-times larger magnetic signal at +1000 G when comparing the sample in the perpendicular to the parallel orientation at the two temperatures Mechanism of Ferromagnetism These measurements verify the spin orbit coupling occurring between the defects introduced by the Si implantation and the Gd atoms present in the host material due to the approximate 3-times larger magnetic sig nal for the Si -implanted GaGdN compared to the nonimplanted GaGdN in the perpendicular orientation at 10 K. The large concentration of defects introduced by the Si implant could be allowing for an enhanced long -range spin polarization in the presence of t he Gd atoms. This effect was also demonstrated in the co implanted MOCVD u-GaN since the additional defects

PAGE 61

61 introduced by the Si implantation may have interacted with the implanted Gd to enhance the magnetic signal in that material by a factor of about 20 over the magnetic signal demonstrated by the non-implanted MOCVD u -GaN in the perpendicular orientation. Summary Considering non -implanted GaN Figure 4 61 and Figure 4 -62 show comparisons of the hysteresis curves for the undoped, p-type, and n-type GaN s amples in the perpendicular orientation at 10 K and 300 K, respectively. In this orientation, the nonimplanted n-GaN has a magnetization value of about 2 to 2.38 105 emu/cm2 at +1000 G, which is very similar to the magnetization measured for non-implan ted u-GaN (~2 to 2.73 105 emu/cm2), while both magnetizations are miniaturized by the value seen in non-implanted p-GaN (~1.20 104 emu/cm2). These measurements clearly demonstrate that the native defects present in p-GaN contribute more to the ferrom agnetic ordering in this material compared to the native defects in u-GaN and nGaN. Based on the formation energies of native de fects in GaN seen in Figure 4-40 the most energetically favorable point defect to form in p-GaN is nitrogen vacancies (VN). Fo r n -GaN, the most energetically favorable point defect to form is gallium vacancies (VGa). Considering only the role of the native defects, the cation vacancy in GaN is not expected to allow for the large degree of ferromagnetism demonstrated by the anion vacancy in GaN. These results are consistent with the work of C. Mitra et al. who explained that even though VGa become favorable in n-type material in the triple negative charge state, this charge state does not carry magnetic moment because the minority spin states would become filled.58 Also the anion interstitial cannot be ruled out when considering defect -induced magnetism in GaN. Due to non equilibrium MOCVD growth under N -rich c onditions (V/III ratio of ~ 6000), a large concentration of nitrogen

PAGE 62

62 interstitials (Ni) are expected to form. Since these defects could also contribute to the ferromagnetism, further studies into anion -related defects in GaN and their effect on ferromagnetic ordering are necessary. Considering implanted GaN Figure 463 and Figure 4-64 show comparisons of the hysteresis curves for the undoped, p-type, and n-type GaN samples in the perpendicular orientation at 10 K and 300 K, respectively. For these materials, VN appears to show the strongest inte raction with the implanted Gd due to the large degree of ferromagnetic ordering. The native defects and defects introduced during implantation in MOCVD n-GaN did not appear to have an effect on the ferromagnetic ordering. For the MBE GaGdN sample, the Si implantation d oes increase the ferromagnetism of the sample due to defect formation. The defects introduced by the Gd and Si implants interacting with the implanted Gd resulted in the largest magnetic signal displayed by the co-implanted sample. Therefore, for GaN based materials, the largest ferromagnetic ordering is exhibited by a material with a large population of defects (particularly point defects such as VN and Ni) and the magnetic impurity Gd. These materials possess a form of spin orbit coupling between the defe cts and Gd that is most strongly manifested when the magnetic properties of the material are analyzed with the applied field oriented perpendicular to the sample surface (parallel to the c axis of the wurtzite GaN crystal structure).

PAGE 63

63 0 100 200 300 400 500 600 700 800 900 1000 0.0 2.0x10174.0x10176.0x10178.0x10171.0x10181.2x1018 Si doses: 5 keV 8.0E11 cm-240 keV 3.6E12 cm-2 Concentration (cm3)Depth (Angstroms) Figure 41 Implantation profile for the combined Si doses in GaN.

PAGE 64

64 0 100 200 300 400 500 600 700 800 900 1000 0.0 2.0x10154.0x10156.0x10158.0x10151.0x1016 Gd dose: 155 keV 2.75E10 cm-2 Concentration (cm3)Depth (Angstroms) Figure 42 Implantation profile for the Gd dose in GaN.

PAGE 65

65 10 20 30 40 50 60 70 80 90 100 101102103104105106107 1. 1.non-implanted MOCVD u-GaN2. 2. 1. 2. 2. 2. 2.1. sapphire 2. GaN1. 1. 1. 1. 1. Intensity (arb. units)2 Theta Figure 43 Powder XRD scan for the non-implanted MOCVD u-GaN reference sample. The peaks corresponding to either the GaN thin film or the sapphire substrate are labeled.

PAGE 66

66 Figure 44 (Top) Diagram of parallel orientation between the applied field (H) and the sample surface. (Bottom) Diagram of perpendicular orientation between the applied field (H) and the sample surface.

PAGE 67

67 -1000 -500 0 500 1000 -4.0x10-6-3.0x10-6-2.0x10-6-1.0x10-60.0 1.0x10-62.0x10-63.0x10-64.0x10-65.0x10-6 T = 10 K Parallel Orientation M (emu/cm2)H (Gauss) Figure 45 Magnetization vs. applied field for nonimplanted MOCVD u-GaN at 10 K for the parallel sample surface/applied field orientation.

PAGE 68

68 -1000 -500 0 500 1000 -6.0x10-6-4.0x10-6-2.0x10-60.0 2.0x10-64.0x10-66.0x10-68.0x10-6 T = 300 K Parallel Orientation M (emu/cm2)H (Gauss) Figure 46 Magnetization vs. applied field for nonimplanted MOCVD u-GaN at 300 K for the parallel sample surface/applied field orientation.

PAGE 69

69 -1000 -500 0 500 1000 -4.0x10-5-3.0x10-5-2.0x10-5-1.0x10-50.0 1.0x10-52.0x10-53.0x10-54.0x10-5 T = 10 K Perpendicular Orientation M (emu/cm2)H (Gauss) Figure 47 Magnetization vs. applied field for nonimplanted MOCVD u-GaN at 10 K for the perpendicular sample surface/applied field orientation.

PAGE 70

70 -1000 -500 0 500 1000 -4.0x10-5-3.0x10-5-2.0x10-5-1.0x10-50.0 1.0x10-52.0x10-53.0x10-54.0x10-5 T = 300 K Perpendicular OrientationM (emu/cm2)H (Gauss) Figure 48 Magnetization vs. applied field for nonimplanted MOCVD u-GaN at 300 K for the perpendicular sample surface/applied field orientation.

PAGE 71

71 10 20 30 40 50 60 70 80 90 100 101102103104105106107 non-implanted Si-implanted Intensity (arb. units)2 Theta Figure 49 Powder XRD scan comparing non-implanted MOCVD uGaN to Si implanted MOCVD u-GaN.

PAGE 72

72 -1000 -500 0 500 1000 -5.0x10-6-4.0x10-6-3.0x10-6-2.0x10-6-1.0x10-60.0 1.0x10-62.0x10-63.0x10-64.0x10-65.0x10-66.0x10-67.0x10-6 T = 10 K Parallel Orientation M (emu/cm2)H (Gauss) Figure 410. Magnetization vs. applied field for Si -implanted MOCVD u-GaN at 10 K for the parallel sample surface/applied field orientation.

PAGE 73

73 -1000 -500 0 500 1000 -6.0x10-6-4.0x10-6-2.0x10-60.0 2.0x10-64.0x10-66.0x10-6 T = 300 K Parallel Orientation M (emu/cm2)H (Gauss) Figure 4 11. Magnetization vs. applied field for Si -implanted MOCVD u-GaN at 300 K for the parallel sample surface/applied field orientation.

PAGE 74

74 -1000 -500 0 500 1000 -2.0x10-5-1.5x10-5-1.0x10-5-5.0x10-60.0 5.0x10-61.0x10-51.5x10-52.0x10-52.5x10-53.0x10-5 T = 10 K Perpendicular Orientation M (emu/cm2)H (Gauss) Figure 412. Magnetization vs. applied field for Si -im planted MOCVD u-GaN at 10 K for the perpendicular sample surface/applied field orientation.

PAGE 75

75 -1000 -500 0 500 1000 -2.0x10-5-1.5x10-5-1.0x10-5-5.0x10-60.0 5.0x10-61.0x10-51.5x10-52.0x10-5 T = 300 K Perpendicular Orientation M (emu/cm2)H (Gauss) Figure 413. Magnetization vs. applied field for Si -implanted MOCVD u-GaN at 300 K for the perpend icular sample surface/applied field orientation.

PAGE 76

76 10 20 30 40 50 60 70 80 90 100 102103104105106107 non-implanted Gd-implanted Intensity (arb. units)2 Theta Figure 414 Powder XRD scan comparing nonimplanted MOCVD u -GaN to Gd implanted MOCVD u-GaN.

PAGE 77

77 -1000 -500 0 500 1000 -1.0x10-5-5.0x10-60.0 5.0x10-61.0x10-51.5x10-5 T = 10 K Parallel Orientation M (emu/cm2)H (Gauss) Figure 415 Magnetization vs. applied field for Gd-implanted MOCVD u-GaN at 10 K for the parallel sample surface/applied field orientation.

PAGE 78

78 -1000 -500 0 500 1000 -6.0x10-6-4.0x10-6-2.0x10-60.0 2.0x10-64.0x10-66.0x10-68.0x10-6 T = 300 K Parallel Orientation M (emu/cm2)H (Gauss) Figure 416 Magnetiza tion vs. applied field for Gd-implanted MOCVD u-GaN at 300 K for the parallel sample surface/applied field orientation.

PAGE 79

79 -1000 -500 0 500 1000 -1.0x10-4-5.0x10-50.0 5.0x10-51.0x10-4 T = 10 K Perpendicular OrientationM (emu/cm2)H (Gauss) Figure 417 Magnetization vs. applied field for Gd-implanted MOCVD u-GaN at 10 K for the perpendicular sample surface/applied field orientation.

PAGE 80

80 -1000 -500 0 500 1000 -1.0x10-4-5.0x10-50.0 5.0x10-51.0x10-4 T = 300 K Perpendicular OrientationM (emu/cm2)H (Gauss) Figure 418 Magnetization vs. applied field for Gd-implanted MOCVD u-GaN at 300 K for the perpendicular sample surface/applied field orientation.

PAGE 81

81 10 20 30 40 50 60 70 80 90 100 102103104105106107 non-implanted Gdand Si-implanted Intensity (arb. units)2 Theta Figure 419 Powder XRD scan comparing nonimplanted MOCVD u -GaN to Gdand Si -co -implanted MOCVD u-GaN

PAGE 82

82 -1000 -500 0 500 1000 -8.0x10-6-6.0x10-6-4.0x10-6-2.0x10-60.0 2.0x10-64.0x10-66.0x10-68.0x10-61.0x10-51.2x10-5 T = 10 K Parallel Orientation M (emu/cm2)H (Gauss) Figure 420 Magnetization vs. applied field for Gdand Si -co -implanted MOCVD uGaN at 10 K for the parallel sample surface/applied field orientation.

PAGE 83

83 -1000 -500 0 500 1000 -1.5x10-5-1.0x10-5-5.0x10-60.0 5.0x10-61.0x10-51.5x10-52.0x10-52.5x10-5 T = 300 K Parallel Orientation M (emu/cm2)H (Gauss) Figure 421 Magnetization vs. applied field for Gdand Si -co -implanted MOCVD uGaN at 300 K for the parallel sample surface/applied field orientation.

PAGE 84

84 -1000 -500 0 500 1000 -4.0x10-4-2.0x10-40.0 2.0x10-44.0x10-4 T = 10 K Perpendicular OrientationM (emu/cm2)H (Gauss) Figure 422 Magnetization vs. applied field for Gd and Si -co -implanted MOCVD uGaN at 10 K for the perpendicular sample surface/applied field orientation.

PAGE 85

85 -1000 -500 0 500 1000 -4.0x10-4-2.0x10-40.0 2.0x10-44.0x10-4 T = 300 K Perpendicular OrientationM (emu/cm2)H (Gauss) Figure 423 Magnetization vs. applied field for Gdand Si -co -implanted MOCVD uG aN at 300 K for the perpendicular sample surface/applied field orientation.

PAGE 86

86 0 50 100 150 200 250 300 350 3.10x10-43.15x10-43.20x10-43.25x10-43.30x10-4 FC ZFC H = 200 Oe Perpendicular Orientation M (emu/cm2)T (K) Figure 424 Magnetization vs. temperature curves for Gd and Si -co -implanted MOCVD u -GaN at an applied field of 200 Oe in the perpendicular sample surface/applied field orientation. The field curve is traced by the squares while the zero field curve is traced by the circles.

PAGE 87

87 -1000 -500 0 500 1000 -5.0x10-4-4.0x10-4-3.0x10-4-2.0x10-4-1.0x10-40.0 1.0x10-42.0x10-43.0x10-44.0x10-45.0x10-4 T = 10 K Perpendicular OrientationImplant Species: Si Gd Gd & Si M (emu/cm2)H (Gauss) Figure 425. Magnetization vs. applied field for Si -, Gd and Gdand Si -co -implanted MOCVD u -GaN samples at 10 K for the perpendicular sample surface/applied field orientation.

PAGE 88

88 -1000 -500 0 500 1000 -5.0x10-4-4.0x10-4-3.0x10-4-2.0x10-4-1.0x10-40.0 1.0x10-42.0x10-43.0x10-44.0x10-45.0x10-4 T = 300 K Perpendicular OrientationImplant Species: Si Gd Gd & Si M (emu/cm2)H (Gauss) Figure 426 Magnetization vs. applied field for Si -, Gd and Gdand Si -co -implanted MOCVD u -GaN samples at 300 K for the perpendicular sample surface/applied field orientation.

PAGE 89

89 0 100 200 300 400 500 600 700 800 900 1000 101210131014101510161017101810191020102110221x1023 Gd atoms/cm3 Si atoms/cm3 vac./cm3 (Gd dose) vac./cm3 (Si doses) Concentration (cm-3)Depth (Angstroms) Figure 427 Gd and Si implantation profile s compared to the vacancy concentration profiles due to the Gd and Si implants near the surface region.

PAGE 90

90 1.5 2.0 2.5 3.0 3.5 0.1 1 BL YL T = 17 K non-implanted co-implanted annealedIntensity (arb. units)Energy (eV) Figure 428. PL spectra at 17 K for Gdand Si -co -implanted MOCVD u-GaN before and after implantation.

PAGE 91

91 1.5 2.0 2.5 3.0 3.5 0.01 BL YL T = 300 K non-implanted co-implantedIntensity (arb. units)Energy (eV) Figure 429. PL spectra at 300 K for Gdand Si -co -implanted MOCVD u-GaN before and after implantation.

PAGE 92

92 10 20 30 40 50 60 70 80 90 100 101102103104105106107 1. 2. 1. 1. 2. 2. 2. 2. 1. 1. 1. 1. 2. 2. 1. 1. 1.non-implanted MOCVD p+-GaN1. sapphire 2. GaN Intensity (arb. units)2 Theta F igure 430 Powder XRD scan for the nonimplanted MOCVD p -GaN reference sample. The peaks corresponding to either the GaN thin film or the sapphire substrate are labeled.

PAGE 93

93 -1000 -500 0 500 1000 -6.0x10-6-4.0x10-6-2.0x10-60.0 2.0x10-64.0x10-66.0x10-6 T = 10 K Parallel OrientationM (emu/cm2)H (Gauss) Figur e 431 Magnetization vs. applied field for non-implanted MOCVD p -GaN at 10 K for the parallel sample surface/applied field orientation.

PAGE 94

94 -1000 -500 0 500 1000 -6.0x10-6-4.0x10-6-2.0x10-60.0 2.0x10-64.0x10-66.0x10-6 T = 300 K Parallel OrientationM (emu/cm2)H (Gauss) Figure 43 2 Magnetization vs. applied field for non-implanted MOCVD p -GaN at 300 K for the parallel sample surface/applied field orientation.

PAGE 95

95 -1000 -500 0 500 1000 -1.0x10-4-5.0x10-50.0 5.0x10-51.0x10-41.5x10-4 T = 10 K Perpendicular OrientationM (emu/cm2)H (Gauss) Figure 43 3 Magnetization vs. applied field for non-implant ed MOCVD p -GaN at 10 K for the perpendicular sample surface/applied field orientation.

PAGE 96

96 -1000 -500 0 500 1000 -1.5x10-4-1.0x10-4-5.0x10-50.0 5.0x10-51.0x10-41.5x10-4 T = 300 K Perpendicular OrientationM (emu/cm2)H (Gauss) Figure 43 4 Magnetization vs. applied field for non-implanted MOCVD p -GaN at 300 K for the perpendicular sample surface/applied field orientation.

PAGE 97

97 10 20 30 40 50 60 70 80 90 100 101102103104105106107 non-implanted Gd-implantedIntensity (arb. units)2 Theta Figure 43 5 Powder XRD scan comparing nonimplanted MOCVD p -GaN to Gd implanted MOCVD p -GaN.

PAGE 98

98 -1000 -500 0 500 1000 -1.0x10-5-5.0x10-60.0 5.0x10-61.0x10-51.5x10-52.0x10-5 T = 10 K Parallel OrientationM (emu/cm2)H (Gauss) Figure 43 6 Magnetization vs. applied field for Gd-implanted MOCVD p -GaN at 10 K for the parallel sample surface/applied field orientation.

PAGE 99

99 -1000 -500 0 500 1000 -1.5x10-5-1.0x10-5-5.0x10-60.0 5.0x10-61.0x10-51.5x10-52.0x10-52.5x10-5 T = 300 K Parallel Orientation M (emu/cm2)H (Gauss) Figure 43 7 Magnetization vs. applied field for Gd-implanted MOCVD p -GaN at 300 K for the parallel sample surface/applied field orientation.

PAGE 100

100 -1000 -500 0 500 1000 -3.0x10-4-2.0x10-4-1.0x10-40.0 1.0x10-42.0x10-43.0x10-4 T = 10 K Perpendicular Orientation M (emu/cm2)H (Gauss) Figure 43 8 Magnetization vs. applied field for Gd-implanted MOCVD p -GaN at 1 0 K for the perpendicular sample surface/applied field orientation.

PAGE 101

101 -1000 -500 0 500 1000 -2.0x10-4-1.0x10-40.0 1.0x10-42.0x10-43.0x10-4 T = 300 K Perpendicular Orientation M (emu/cm2)H (Gauss) Figure 43 9 Magnetization vs. applied field for Gd-implanted MOCVD p -GaN at 300 K for the perpendicular sample surface/applied field orientation.

PAGE 102

102 Figure 440 Energies of formation as a function of Fermi level for native point defects in GaN. Note that Ga -rich conditions are assumed and the zero of Fermi level corresponds to the top of the valence band.57

PAGE 103

103 1.5 2.0 2.5 3.0 3.5 0.1 1 UVL RL T = 17 K non-implanted Gd-implantedIntensity (arb. units)Energy (eV) Figure 441. PL spectra at 17 K for Gdimplanted MOCVD p-GaN before and after implantation.

PAGE 104

104 1.5 2.0 2.5 3.0 3.5 0.01 UVL RLT = 300 K non-implanted Gd-implanted Intensity (arb. units)Energy (eV) Figure 442. PL spectra at 300 K for Gd-implanted MOCVD p-GaN before and after implantation.

PAGE 105

105 10 20 30 40 50 60 70 80 90 100 101102103104105106107 1. 1. 2. 2. 2. 2. 1. 1. 1. 2. 2. 1. 1. 1.non-implanted MOCVD n-GaN1. sapphire 2. GaN Intensity (arb. units)2 Theta Figure 443 Powder XRD scan for the nonimplanted MOCVD n-GaN reference sample. The peaks corresponding to either the GaN thin film or the sapphire substrate are labeled.

PAGE 106

106 -1000 -500 0 500 1000 -6.0x10-6-4.0x10-6-2.0x10-60.0 2.0x10-64.0x10-66.0x10-6 T = 10 K Parallel OrientationM (emu/cm2)H (Gauss) Figure 44 4 Magnetization vs. applied field for non -implanted MOCVD n-GaN at 10 K for the parallel sample surface/applied field orientation.

PAGE 107

107 -1000 -500 0 500 1000 -1.0x10-5-5.0x10-60.0 5.0x10-61.0x10-5 T = 300 K Parallel OrientationM (emu/cm2)H (Gauss) Figure 44 5 Magnet ization vs. applied field for non-implanted MOCVD n-GaN at 300 K for the parallel sample surface/applied field orientation.

PAGE 108

108 -1000 -500 0 500 1000 -4.0x10-5-3.0x10-5-2.0x10-5-1.0x10-50.0 1.0x10-52.0x10-53.0x10-54.0x10-5 T = 10 K Perpendicular Orientation M (emu/cm2)H (Gauss) Figure 44 6 Magnetization vs. applied field for non -implanted MOCVD n-GaN at 10 K for the perpendicular sample surface/applied field orientation.

PAGE 109

109 -1000 -500 0 500 1000 -4.0x10-5-3.0x10-5-2.0x10-5-1.0x10-50.0 1.0x10-52.0x10-53.0x10-54.0x10-5 T = 300 K Perpendicular OrientationM (emu/cm2)H (Gauss) Figure 44 7 Magnetization vs. applied field for non -implanted MOCVD n-GaN at 300 K for the perpendicular sample surface/applied field orientation.

PAGE 110

110 10 20 30 40 50 60 70 80 90 100 101102103104105106107 non-implanted Gd-implanted Intensity (arb. units)2 Theta Figure 44 8 Powder XRD scan comparing nonimplanted MOCVD n -GaN to Gd implanted MOCVD n-GaN.

PAGE 111

111 -1000 -500 0 500 1000 -1.0x10-5-5.0x10-60.0 5.0x10-61.0x10-5 T = 10 K Parallel OrientationM (emu/cm2)H (Gauss) Figure 44 9 Magnetization vs. applied field for Gd-implanted MOCVD n-GaN at 10 K for the parallel sample surface/applied field orientation.

PAGE 112

112 -1000 -500 0 500 1000 -1.5x10-5-1.0x10-5-5.0x10-60.0 5.0x10-61.0x10-51.5x10-5 T = 300 K Parallel Orientation M (emu/cm2)H (Gauss) Figure 450 Magnetization vs. applied field for Gd-implanted MOCVD n-GaN at 300 K for the parallel sample surface/applied field orientation.

PAGE 113

113 -1000 -500 0 500 1000 -3.0x10-5-2.0x10-5-1.0x10-50.0 1.0x10-52.0x10-53.0x10-5 T = 10 K Perpendicular OrientationM (emu/cm2)H (Gauss) Figure 451 Magnetization vs. applied field for Gd-implanted MOCVD n-GaN at 10 K for the perpendicular sample surface/applied field orientation.

PAGE 114

114 -1000 -500 0 500 1000 -3.0x10-5-2.0x10-5-1.0x10-50.0 1.0x10-52.0x10-53.0x10-5 T = 300 K Perpendicular OrientationM (emu/cm2)H (Gauss) Figure 452 Magnetization vs. applied field for Gd-implanted MOCVD n-GaN at 300 K for the perpendicular sample surface/applied field orientation.

PAGE 115

115 -1000 -500 0 500 1000 -2.5x10-5-2.0x10-5-1.5x10-5-1.0x10-5-5.0x10-60.0 5.0x10-61.0x10-51.5x10-52.0x10-52.5x10-5 T = 10 K Parallel Orientation M (emu/cm2)H (Gauss) Figure 453 Magnetization vs. applied field for non-implanted MBE GaGdN at 10 K for the parallel sample surface/applied field orientation.

PAGE 116

116 -1000 -500 0 500 1000 -2.0x10-5-1.5x10-5-1.0x10-5-5.0x10-60.0 5.0x10-61.0x10-51.5x10-52.0x10-52.5x10-5 T = 300 K Parallel Orientation M (emu/cm2)H (Gauss) Figure 45 4 Magnetization vs. applied field for non-implant ed MBE GaGdN at 300 K for the parallel sample surface/applied field orientation.

PAGE 117

117 -1000 -500 0 500 1000 -8.0x10-5-6.0x10-5-4.0x10-5-2.0x10-50.0 2.0x10-54.0x10-56.0x10-58.0x10-5 T = 10 K Perpendicular Orientation M (emu/cm2)H (Gauss) Figure 45 5 Magnetization vs. applied field for non-implanted MBE GaGdN at 10 K for the perpendicular sample surface/applied field orientation.

PAGE 118

118 -1000 -500 0 500 1000 -1.5x10-4-1.0x10-4-5.0x10-50.0 5.0x10-51.0x10-41.5x10-4 T = 300 K Perpendicular Orientation M (emu/cm2)H (Gauss) Figure 45 6 Magnetization vs. applied field for non-implanted MBE GaGdN at 300 K for the perpendicular sample surface/applied field orientation.

PAGE 119

119 -1000 -500 0 500 1000 -5.0x10-60.0 5.0x10-61.0x10-51.5x10-5 T = 10 K Parallel Orientation M (emu/cm2)H (Gauss) Figure 45 7 Magnetization vs. applied field for Si -implanted MBE GaGdN at 10 K for the parallel sample surface/applied field orientation.

PAGE 120

120 -1000 -500 0 500 1000 -1.0x10-5-5.0x10-60.0 5.0x10-61.0x10-51.5x10-5 T = 300 K Parallel Orientation M (emu/cm2)H (Gauss) Figure 45 8 Magnetization vs. applied field for Si -implanted MBE GaGdN at 300 K for the parallel sample surface/applied field orientation.

PAGE 121

121 -1000 -500 0 500 1000 -2.0x10-4-1.5x10-4-1.0x10-4-5.0x10-50.0 5.0x10-51.0x10-41.5x10-42.0x10-4 T = 10 K Perpendicular Orientation M (emu/cm2)H (Gauss) Figure 45 9 Magnetization vs. applied field for Si -implanted MBE GaGdN at 10 K for the perpendicular sample surface/applied field orientation.

PAGE 122

122 -1000 -500 0 500 1000 -1.5x10-4-1.0x10-4-5.0x10-50.0 5.0x10-51.0x10-41.5x10-4 T = 300 K Perpendicular Orientation M (emu/cm2)H (Gauss) Figure 460 Magnetization vs. applied field for Si -implanted MBE GaGdN at 300 K for the perpendicular sample surface/applied field orientation.

PAGE 123

123 -1000 -500 0 500 1000 -1.0x10-4-5.0x10-50.0 5.0x10-51.0x10-41.5x10-4 T = 10 K Perpendicular OrientationDopant Species: undoped p-type n-type M (emu/cm2)H (Gauss) Figure 461 Magnetization vs. applied field for non-implanted undoped, p-type, and ntype Ga N samples at 10 K for the perpendicular sample surface/applied field orientation.

PAGE 124

124 -1000 -500 0 500 1000 -1.5x10-4-1.0x10-4-5.0x10-50.0 5.0x10-51.0x10-41.5x10-4 T = 300 K Perpendicular OrientationDopant Species: undoped p-type n-type M (emu/cm2)H (Gauss) Figure 462 Magnetization vs. applied field for non-implanted undoped, p-type, and ntype GaN samples at 3 00 K for the perpendicular sample surface/applied field orientation.

PAGE 125

125 -1000 -500 0 500 1000 -3.0x10-4-2.5x10-4-2.0x10-4-1.5x10-4-1.0x10-4-5.0x10-50.0 5.0x10-51.0x10-41.5x10-42.0x10-42.5x10-43.0x10-4 T = 10 K Perpendicular OrientationGd-implanted GaN: undoped p-type n-type M (emu/cm2)H (Gauss) Figure 463 Magnetization vs. applied field for Gd-implanted undoped, p-type, and ntype GaN samples at 10 K for the perpendicular sample surface/applied field orientation.

PAGE 126

126 -1000 -500 0 500 1000 -2.5x10-4-2.0x10-4-1.5x10-4-1.0x10-4-5.0x10-50.0 5.0x10-51.0x10-41.5x10-42.0x10-42.5x10-4 T = 300 K Perpendicular OrientationGd-implanted GaN: undoped p-type n-type M (emu/cm2)H (Gauss) Figure 46 4 Magnetization vs. applied field for Gd-implanted undoped, p-type, and ntype GaN samples at 300 K for the perpendicular sample surface/applied field orientation.

PAGE 127

127 CHAPTER 5 ION IMPLANTATION OF GADOLINIUM IN ZINC O XIDE Research in ZnO based DMS materials began when Dietl et al. predicted the TC of ZnO doped with 5% Mn to be above room tem perature.10 Various groups have demonstrated room temperature ferromagnetism in TM doped ZnO.66 6 8 Recently, studies in Gd-implanted ZnO have shown room temperature ferromagnetic ordering.69 70 A Gd+ dose of 5 101 5 cm2 at an energy of 180 keV was implanted in a hydr othermally grown ZnO (0001) single crystal and this sample was found to possess ferromagnetic properties above room temperature after being annealed at 820 K for 15 minutes.6 9 Implanting a Gd dose of 1 101 5 cm2 at an energy of 300 keV showed significant distortion of the ZnO lattice from XRD measurements with a ferromagnetic -like signal seen at room temperature and being increased by low temperature HV annealing at 350 C for 15 minutes.70 Further analysis into Gd implantation in ZnO is necessary to help determine the ferromagnetic mechanism present in these materials. Non -Implanted A commercially available 1 cm2 square piece of a (0001) ZnO single crystal substrate was set aside as a reference sample for this implantation study. Figure 5-1 shows the XRD spectrum for the non -implanted ZnO sample. The peaks at 11.55 and 22.87 could not be identified with JCPDS files and may be attributable to artifacts associated with the polychromatic nature of the x -ray source. Magnetic measurements were performed at both 10 K and 300 K for both the parallel and perpendicular orientations. Before the magnetization of each sam ple was measured in a differe nt orientation and/or temperature, the sample was demagnetized using a decrea sing oscillating magnetic field The diamagnetic contribution of the bulk was subtracted from

PAGE 128

128 the measured magnetization values and these values were then normalized by the area of the sample. Figure 5 -2 and Figure 5 3 show the magnetization versus applied magnetic field curves for the non-implanted ZnO in the parallel orientation at 10 K and 300 K, respectively. Interestingly, this sample exhibits ferromagnetism at both low and room temperatures without the known presence of a magnetic impurity. Just like in the study of the nonimplanted GaN, native defects in ZnO may be behind the observed ferromagnetic ordering. Figure 54 and Figure 55 show the magnetization versus applied magnetic field curves for the non -implanted ZnO in the perpendicular orientation at 10 K and 300 K, respectively. Both of these hysteresis curves show ferromagnetic behavior and demonstrate the anisotropic enhancement effect (AEE). This effect is seen by the approximate 3-times larger magnetic signal at +1000 G for the nonimplanted Z nO in the perpendicular orientation compared to the signal in the parallel orientation. Since the AEE is demonstrated by both the GaN and ZnO samples, this dependence on orientation may be related to the wurtzite crystal structure of these materials. Gd -Im planted Another piece of the commercially available (0001) ZnO single crystal substrate was implanted with Gd ions of energy 155 keV and an implantation dose of 2.75 1010 cm2 at room temperature with the ion beam oriented 7 degrees off from perpendicular to the surface of the target material to avoid channeling Figure 56 shows the implantation profile for the Gd dose in wurtzite ZnO. Usi ng data obtained from the Transport of Ions in Matter (TRIM) program to plot the predicted i on ranges, the Gd atoms w ere mai nly distributed about 85 nm from the sample surface. T he peak Gd concen tration was determined to be 9.71 1015 cm3 at a depth of 320 below the

PAGE 129

129 surface based on data obtained from the TRI M program Figure 57 shows the powder XRD scan for the Gd -implanted ZnO sample compared to the XRD scan of the nonimplanted ZnO reference sample. Even though there are ZnO peaks present in the Gdimplanted spectrum that were not in the non -implanted spectrum, there are still additional peaks present in the Gd-implanted ZnO spectrum. These peaks most likely originate from a GdxOy complex that is not expected to exhibit ferromagnetic behavior. Magnetic measurements were taken in both the parallel and perpendicular sam ple surface/applied field orientations. For the parallel orientation, the magnetization versus applied magnetic field curves at 10 K and 300 K are shown in Figure 5-8 and Figure 59 respectively. Both curves exhibit ferromagnetic behavior and appear to be nearly identical. For the perpendicular orientation, the magnetization versus applied magnetic field curves at 10 K and 300 K are shown in Fig ure 5 -10 and Figure 5 -11, respectively. Similar to the case for all the previously measured samples, the AEE is s een for the Gdimplanted ZnO samples as the magnetization in t he perpendicular orientation exceeds the magnetization in the parallel orientation. More specifi cally, there is an approximate 4 -times larger magnetization at + 1000 G in the perpendicular as com pared to the parallel orien tation for both 10 K and 300 K. When comparing the magnetic signal at +1000 G for the Gdimplanted to the nonimplanted ZnO in the perpendicular orientation, there is about a 2 to 3 times larger signal in the Gd-implanted ZnO. Once again, the defects formed by the Gd implantation are speculated to interact with th e Gd to enhance the ferromagnetism in this material. Mechanism of Ferromagnetism Considering factors that are similar to both the GaN and ZnO material systems, both of these compound semiconductors possess wurtzite crystal structures. Similarly to

PAGE 130

130 the case for defect formation in GaN, the implantation of Gd in ZnO is expected to form a host of defects, with the vacancies forming mainly near the surface and the interstiti als forming deeper into the implanted material. There is once again a size differential between the implanted Gd ion radius and the host Zn ion radius resulting in the existence of large vacancy and interstitial concentrations after Gd implantation Ney e t al. postulated that the generation of electrons in the ZnO of the potential donor Gd, coupled with the trapping of those respective carriers by the implantation defects, may qualitatively explain the measured magnetic behavior.69 Calculations have shown that oxygen vacancies (VO) have the lowest formation energy among defects that form in Znrich conditions.7 1 Similar to the situation with the MOCVD p -GaN, the anion vacancy has the lowest formation energy and may be partly responsible for ferromagnetic or dering in non-implanted samples. These VO will also be formed during implantation and this increased concentration of defects may interact with the implanted Gd to enhance the ferromagnetism. Also, oxygen interstitials (Oi) must be considered based on the formation of this defect in non implanted ZnO due to nonequilibrium growth conditions and in Gd-implanted ZnO due to the implantation process. Similar to the case for GaN, the anion-related defects appear to play the largest role in the observed ferromagn etic ordering in ZnO.

PAGE 131

131 10 20 30 40 50 60 70 80 90 100 102103 non-implanted ZnO(004) (110) (102) (002) Intensity (arb. units)2 Theta Figure 51 Powder XRD scan for the non-implanted ZnO reference sample. The peaks correspond to the labeled planes. The peaks at 11.55 and 22.87 could not be identif ied with JCPDS files and may be attributable to artifacts associated with the polychromatic nature of the x -ray source.

PAGE 132

132 -1000 -500 0 500 1000 -2.0x10-5-1.5x10-5-1.0x10-5-5.0x10-60.0 5.0x10-61.0x10-51.5x10-52.0x10-52.5x10-5 T = 10 K Parallel OrientationM (emu/cm2)H (Gauss) Figure 52 Magnetization vs. applied field for nonimplanted ZnO at 10 K for the parallel sample surface/applied field orientation.

PAGE 133

133 -1000 -500 0 500 1000 -2.0x10-5-1.5x10-5-1.0x10-5-5.0x10-60.0 5.0x10-61.0x10-51.5x10-52.0x10-52.5x10-5 T = 300 K Parallel OrientationM (emu/cm2)H (Gauss) Figure 53 Magnetization vs. applied field for nonimplanted ZnO at 300 K for the parallel sample surface/applied field orie ntation.

PAGE 134

134 -1000 -500 0 500 1000 -6.0x10-5-4.0x10-5-2.0x10-50.0 2.0x10-54.0x10-56.0x10-58.0x10-5 T = 10 K Perpendicular OrientationM (emu/cm2)H (Gauss) Figure 54 Magnetization vs. applied field for nonimplanted ZnO at 10 K for the perpendicular sample surface/applied field orientation.

PAGE 135

135 -1000 -500 0 500 1000 -6.0x10-5-4.0x10-5-2.0x10-50.0 2.0x10-54.0x10-56.0x10-5 T = 300 K Perpendicular Orientation M (emu/cm2)H (Gauss) Figure 55 Magnetization vs. applied field for nonimplanted ZnO at 300 K for the perpendicular sample surface/applied field orientation.

PAGE 136

136 0 100 200 300 400 500 600 700 800 900 1000 0.0 2.0x10154.0x10156.0x10158.0x10151.0x1016 Gd dose: 155 keV 2.75E10 cm-2Concentration (cm-3)Depth (Angstroms) Figure 56 Implantation profile for the Gd dose in ZnO.

PAGE 137

137 10 20 30 40 50 60 70 80 90 100 102103104105106 2. 2. 2. 1. 1. 1. 1. 1. 1. 1.1. ZnO 2. GdxOy non-implanted Gd-implantedIntensity (arb. units)2 Theta Figure 57 Powder XRD scan comparing non-implanted ZnO to Gdimplanted ZnO.

PAGE 138

138 -1000 -500 0 500 1000 -3.0x10-5-2.0x10-5-1.0x10-50.0 1.0x10-52.0x10-53.0x10-54.0x10-5 T = 10 K Parallel OrientationM (emu/cm2)H (Gauss) Figure 58 Magnetization vs. applied field for Gdimplanted ZnO at 10 K for the parallel sample surface/applied field orientation.

PAGE 139

139 -1000 -500 0 500 1000 -3.0x10-5-2.0x10-5-1.0x10-50.0 1.0x10-52.0x10-53.0x10-54.0x10-55.0x10-5 T = 300 K Parallel OrientationM(emu/cm2)H (Gauss) Figure 59 Magnetization vs. applied field for Gdimplanted ZnO at 300 K for the parallel sample surface/applied field orientation.

PAGE 140

140 -1000 -500 0 500 1000 -1.5x10-4-1.0x10-4-5.0x10-50.0 5.0x10-51.0x10-41.5x10-4 T = 10 K Perpendicular OrientationM (emu/cm2)H (Gauss) Figure 510 Magnetization vs. applied field for Gd-implanted ZnO at 10 K for the perpendicular sample surface/applied field orientation.

PAGE 141

141 -1000 -500 0 500 1000 -1.5x10-4-1.0x10-4-5.0x10-50.0 5.0x10-51.0x10-41.5x10-4 T = 300 K Perpendicular OrientationM (emu/cm2)H (Gauss) Figure 511 Magnetization vs. applied field for Gd-implanted ZnO at 300 K for the perpendicular sample surface/applied field orientation

PAGE 142

142 CHAPTER 6 ION IMPLANTATION OF GADOLINIUM IN GALLIU M ARSENIDE Possibly due to the fact that the highest TC reported for Mn-doped GaAs was approximately 110 K13, there ha s not been a large emphasis and progression of resear ch into rare earth metal implantation in GaAs for spintronic applications. Studying the effects on the magnetic properties of Gd implantation in GaAs may provide another facet of the ferromagnetic mechanism in compound semiconductor materials. GaAs provide s an opportunity to study a material system with a smaller band gap (1.42 eV) than GaN (3.39 eV) and ZnO (3.37 eV), and also a different crystal structure (zinc blende) than GaN and ZnO (wurtzite). Non -Implanted A 2 inch wafer of a commercially available u ndoped GaAs single crys tal substrate was cleaved into halves and one half was set aside as a reference sample. Figure 6 1 shows the XRD spectrum for the non -implanted GaAs The peaks at 59.35 and 63.31 could not be identified with JCPDS files and may be attributable to artifacts associated with the polychromatic nature of the x -ray source. Magnetic measurements were performed at both 10 K and 300 K for both the parallel and perpendicular orientations. Before the magnetization of each sam ple was measured i n a differe nt orientation and/or temperature, the sample was demagnetized using a decreasing oscillating magnetic field The diamagnetic contribution of the bulk was subtracted from the measured magnetization values and these values were then normalized by the area of the sample. Figure 6 2 and Figure 6 3 show the magnetization versus applied magnetic field curves for the nonimplanted GaAs in the parallel orientation at 10 K and 300 K, respectively. The nonimplanted GaAs exhibits very minute to no ferromagnetic ordering and the

PAGE 143

143 hysteresis curves show diamagnetic behavior in the parallel orientation. Figure 6 -4 and Figure 6 5 show the magnetization versus applied magnetic field curves for the nonimplanted GaAs in the perpendicular orientation at 10 K and 300 K, respectively. In this orientation, both of the hysteresis curves show ferromagnetism but the curve at 300 K indicates that the Curie temperature for this material may be near room temperature. The anisotropic enhancement effect (AEE) of magnetic ordering is also seen in GaAs, even though the magnetic signal in the perpendicular orientation is relatively small. This observation would appear to preclude any orientation dependence on the wurtzite crystal structure seen in GaN and ZnO, since this GaAs sample has a zinc blende crystal structure. Gd -Implanted The other half of the commercially available undoped GaAs single crystal substrate was implanted with Gd ions of energy 155 keV and an implantation dose of 2.75 101 0 cm2 at room temperature with the ion beam oriented 7 degrees off from perpendicular to the surface of the target material to avoid channeling. Figure 66 shows the implantation profile for the Gd dose in zinc blende GaAs. Usi ng data obtained from the Tr ansport of Ions in Matter (TRIM) program to plot the predicted i on ranges, the Gd atoms were mai nly distributed about 100 nm from the sample surface. T he peak Gd concen tration was determined to be 6.95 1015 cm3 at a depth of 360 below the surface base d on data obtained from the TRIM program Figure 67 shows the powder XRD scan for the Gd -implanted GaAs sample compared to the XRD scan of the nonimplanted GaAs reference sample. There were no additional peaks seen in the Gdimplanted GaAs spectrum when compared the nonimplanted spectrum, indicating no formation of complexes of secondary phases due to implantation (within the detection

PAGE 144

144 limits of the XRD). M agnetic measurements were taken in both the parallel and perpendicular sample surf ac e/applied field orientations. For the parallel orientation, the magnetization versus applied magnetic field curves at 10 K and 300 K are shown in Figure 68 and Figure 69 respectively. As can be seen in these figures, in the parallel orientation it is debatable whether Gdimplanted GaAs truly exhibits ferromagnetic behavior as both curves appear to indicate diamagnetic behavior. For the perpendicular orientation, the magnetization versus applied magnetic field curves at 10 K and 300 K are shown in Figure 6 -10 and Figure 611, respectively. Similar to the Gd -implanted MOCVD u -GaN, coimplanted MOCVD u-GaN and Gdimplanted ZnO samples, the AEE is seen for the Gdimplanted GaAs samples as the magnetization in the pe rpendicular orientation exceeds the magnetiz ation in the parallel orientation. For the Gd -implanted GaAs t here is an approximately 3 times larger magnetization at + 1000 G in the perpendicular as compared to the parallel orientation for both 10 K and 300 K. Mechanism of Ferromagnetism Even though th e magnetic signals measured in the GaAs samples are relatively lower than the signals observed for the GaN and ZnO samples, this material system also demonstrates the AEE albeit to the smallest extent of any of the compound semiconductor materials This d iscovery hinders the explanation of the wurtzite crystal structure playing an important role in the ferromagnetic ordering. Whatever the ferromagnetic mechanism occurring in these materials involves, GaAs appears to possess the least amount of that species based on a comparison of the magnetic properties measured for the compound semiconductor materials. Under Ga -rich conditions, the arsenic vacancy (VAs) has one of the highest formation energies for native defects in the mid band gap range.7 2 Therefore f or GaAs, the anion vacancy is

PAGE 145

145 not as likely to form. For MOCVD p -GaN and ZnO, the anion vacancy has the lowest formation energy. Previous studies have shown that VAs and arsenic interstitials (Asi) are the dominant point defects in GaAs grown from As -rich melts.7 3 Ion channel experiments have been used to determine large Asi concentrations in GaAs grown at low temperatures via molecular beam epitaxy (MBE).7 4 Since Asi is a more prevalent species in GaAs, this point defect may be responsible for the ferromagnetic ordering seen in the non -implanted GaAs. Similar to the case for MOCVD n -GaN, when comparing the hysteresis curve of Gdimplanted GaAs to nonimplanted GaAs in the perpendicular orientation at 10 K, they are almost identic al. The implanted Gd does not exhibit any spin orbit coupling with the defects in GaAs to produce a larger magnetic moment in implanted GaAs compared to nonimplanted GaAs. Further investigation s into the specific defect or defect complexes behind the ferr omagnetism are required, but these preliminary results point to anion-related defects in GaAs possibly playing a large role in the magnetic ordering.

PAGE 146

146 10 20 30 40 50 60 70 80 90 100 101102103104105106 (400) (220) (200) (111) non-implanted u-GaAsIntensity (arb. units)2 Theta Figure 61 Powder XRD scan for the non-implanted GaAs reference sample. The peaks correspond to the labeled planes. The peaks at 59.35 and 63.31 could not be identified with JCPDS files and may be attributable to artifacts associated with the polychromatic nature o f the x -ray source.

PAGE 147

147 -1000 -500 0 500 1000 -1.5x10-5-1.0x10-5-5.0x10-60.0 5.0x10-61.0x10-5 T= 10 K Parallel Orientation M (emu/cm2)H (Gauss) Figure 62 Magnetization vs. applied field for nonimplanted GaAs at 10 K for the parallel sample surface/applied field orientation.

PAGE 148

148 -1000 -500 0 500 1000 -5.0x10-60.0 5.0x10-61.0x10-5 T= 300 K Parallel Orientation M (emu/cm2)H (Gauss) Figure 63 Magnetization vs. applied field for nonimplanted GaAs at 300 K for the parallel sample surface/applied field orientation.

PAGE 149

149 -1000 -500 0 500 1000 -3.0x10-5-2.0x10-5-1.0x10-50.0 1.0x10-52.0x10-53.0x10-54.0x10-5 T = 10 K Perpendicular Orientation M (emu/cm2)H (Gauss) Figure 64 Magnetization vs. applied field for nonimplanted GaAs at 10 K for the perpendicular sample surface/applied field orientation.

PAGE 150

150 -1000 -500 0 500 1000 -2.5x10-5-2.0x10-5-1.5x10-5-1.0x10-5-5.0x10-60.0 5.0x10-61.0x10-51.5x10-52.0x10-52.5x10-5 T = 300 K Perpendicular Orientation M (emu/cm2)H (Gauss) Figure 65 Magnetization vs. applied field for nonimplanted GaAs at 300 K for the perpendicular sample surface/applied field orientation.

PAGE 151

151 0 100 200 300 400 500 600 700 800 900 1000 0 1x10152x10153x10154x10155x10156x10157x10158x1015 Gd dose: 155 keV 2.75E10 cm-2Concentration (cm3)Depth (Angstroms) Figure 66 Implantation profile for the Gd dose in GaAs.

PAGE 152

152 10 20 30 40 50 60 70 80 90 100 101102103104105106 non-implanted Gd-implantedIntensity (arb. units)2 Theta Figure 67 Powder XRD scan comparing non-implanted GaAs to Gd -implanted GaAs

PAGE 153

153 -1000 -500 0 500 1000 -4.0x10-6-2.0x10-60.0 2.0x10-64.0x10-66.0x10-68.0x10-61.0x10-51.2x10-5 T = 10 K Parallel OrientationM (emu/cm2)H (Gauss) Figure 68 Magnetization vs. applied field for Gdimplanted GaAs at 10 K for the parallel sample surface/applied field orientation.

PAGE 154

154 -1000 -500 0 500 1000 -6.0x10-6-4.0x10-6-2.0x10-60.0 2.0x10-64.0x10-66.0x10-68.0x10-61.0x10-51.2x10-51.4x10-51.6x10-5 T = 300 K Parallel OrientationM (emu/cm2)H (Gauss) Figure 69 Magnetization vs. applied field for Gdimplanted GaAs at 300 K for the parallel sample surface/applied field orientation.

PAGE 155

155 -1000 -500 0 500 1000 -5.0x10-5-4.0x10-5-3.0x10-5-2.0x10-5-1.0x10-50.0 1.0x10-52.0x10-53.0x10-54.0x10-5 T = 10 K Perpendicular OrientationM (emu/cm2)H (Gauss) Figure 610 Magnetization vs. applied field for Gd-implanted GaAs at 10 K for the perpendicular sample surface/applied field orientation.

PAGE 156

156 -1000 -500 0 500 1000 -3.0x10-5-2.0x10-5-1.0x10-50.0 1.0x10-52.0x10-53.0x10-54.0x10-5 T = 300 K Perpendicular OrientationM (emu/cm2)H (Gauss) Figure 611 Magnetization vs. applied field for Gd-implanted GaAs at 300 K for the perpendicular sample surface/applied field orientation.

PAGE 157

157 CHAPTER 7 THERMAL ANNEALING EF FECTS ON IMPLANTED GALLIUM NI TRIDE When processing compound semiconductor materials for device applications, thermal annealing treatments at temperatures ranging from 300 C to 900 C are necessary for such objectives as activating implant species or ohmic contacts. Since the majority of this work focused on the ion implantation of compound semiconductor materials and the resultant magnetic properties, the next logical step would be to determine the magnetic properties after annealing the implanted materials. This stu dy would provide additional insight into the mechanism of ferromagnetism present in these materials. Previous work on the annealing of Gd focused ion beam implanted GaN samples (Gd3+ dose of 2.4 1011 cm2 at energy 300 keV) showed an approximate 60% decr ease in the magnetization after annealing at 800 C and 900 C.43 These results reveal that the magnetic ordering of implanted GaN can be diminished due to annealing. The Gdand Si -co implanted MOCVD u-GaN sample was chosen for the anneal study because thi s material demonstrated the largest magnetic signal of any combination of implant species/compound semiconductor material. Due to this large magnetic signal, if the magnetic ordering of this material is reduced as expected due to annealing, this sample provides this experiment with the best chance of still detecting a magnetic signal. The thermal anneal treatments were done in a Solaris 150 Rapid Thermal Processing (RTP) system. The co-implanted u -GaN was annealed under ambient nitrogen for 1 minute per each anneal. The sample was annealed at 300 C and then analyzed in the Quantum Device Magnetic Properties Measurement System model MPMS 5S Superconducting Quantum Interference Device (SQUID) magnetometer at 300 K with the applied magnetic field oriented perpendicular to the sample surface

PAGE 158

158 (perpendicular orientation) This same process was completed for annealing temperatures of 400 C, 500 C, 600 C, 700 C, and 800 C Figure 71 shows the hysteresis cu rves for the co implanted u-GaN. The as implanted curve represents the sample before any thermal anneal treatments. Figure 72 shows the magnetizations values at +1000 G at eac h anneal temperature normalized by the as implanted magnetization value at +1000 G ( 1.74 104 emu/cm2, de picted at zero anneal temperature). A n approximately 7% decrease in magnetic signal is seen after the 300 C anneal and then the largest decrease in magnetization (about 80% compared to the as implanted value) is seen after the 400 C anneal. The magnetic or dering remains at approximately the same level for the remainder of the annealing treatments. Even though the magnetic signal decreases with annealing, this material could still be explored for spintronic device applications since ferromagnetic ordering is still exhibited by the annealed material. Photoluminescence (PL) measurements were taken to assist in determining the point defects present in the materials before and after implantation and annealing. For these measurements a HeCd laser of wavelength 325.13 nm (3.814 eV energy) was used as the excitation source, which is above the band edge of GaN (3.39 eV). For low temperature measurements down to ~17 K, a stage comprised of a He closed system coldfinger was used. Figure 73 and Figure 74 show the PL sp ectra at 17 K and 300 K, respectively, for the co -implanted MOCVD u-GaN sample before and after implantation and after the thermal annealing treatments from 300C to 800C. The non-implanted MOCVD u -GaN clearly exhibits both the yellow luminescence (YL) and blue luminescence (BL) bands. The YL band is typically associated with gallium vacancy

PAGE 159

159 (VGa)-related defects and the BL band is associated with transitions from shallow donors (low temperatures) or the conduction band (elevated temperatures) to deep acceptor levels (ionization energy of about 0.34 to 0.40 eV).65 After the implantation process, the large reduction in signal from the YL region indicates a decrease in luminescence from the VGa-related defects. After the thermal annealing treatments from 300C to 800C, the PL spectrum shows a significant decrease in luminescence from the BL region. This reduction in the BL band corresponds to the decreased magnetic ordering demonstrated by the co implanted MOCVD u-GaN after being annealed at 800C. Furthe r investigations are necessary, but the species responsible for the BL band may be responsible for the ferromagnetic ordering in this material. These measurements are consistent with the speculation that the defects are involved in the ferromagnetic orderi ng in this material since annealing the sample is expected to reduce the defect density by providing the vacancies and interstitials with enough energy to incorporate into the available lattice positions vacated due to the implantation process As the anne al temperature increases, a corresponding increase in carriers should be activated due to the large Si implant dose. Notice that the magnetic properties of the sample do not recover to the as implanted value, indicating that the activated carriers do not play as great a part in the ferromagnetic ordering as that of the defects.

PAGE 160

160 -1000 -500 0 500 1000 -2.0x10-4-1.5x10-4-1.0x10-4-5.0x10-50.0 5.0x10-51.0x10-41.5x10-42.0x10-4 T = 300 K Perpendicular OrientationAnneal Temp.: as implanted 300C 400C 500C 600C 700C 800C M (emu/cm2)H (Gauss) Figure 71. Magnetization vs. applied field curves for as implanted and annealed Gd and Si -co -implanted u-GaN at 30 0 K for the perpendicular sample surface/applied field orientation.

PAGE 161

161 0 100 200 300 400 500 600 700 800 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 T = 300 K Perpendicular Orientation Normalized M+1000 GAnneal Temperature (oC) Figure 72 Normalized magnetizations values at +1000 G for Gdand Si -co -implanted MOCVD u -GaN. The magnetization values at +1000 G were calculated by averaging the two values of M at +1000 G from hysteresis curves analyzed in the perpendicular applied field/sample surface orientation at 300 K The normalizing magnetization value was 1.74 104 emu/cm2.

PAGE 162

162 1.5 2.0 2.5 3.0 3.5 0.1 1 BL YL T = 17 K non-implanted co-implanted annealedIntensity (arb. units)Energy (eV) Figure 73. PL spectra at 17 K for Gdand Si -co -implanted MOCVD u-GaN before and after implantation and after thermal annealing treatments up to 800 C.

PAGE 163

163 1.5 2.0 2.5 3.0 3.5 0.01 BL YL T = 300 K non-implanted co-implanted annealedIntensity (arb. units)Energy (eV) Figure 74. PL spectra at 300 K for Gdand Si -co -implanted MOCVD u-GaN before and after implantation and after thermal annealing treatments up to 800 C.

PAGE 164

164 CHAPTER 8 POTENTIAL APPLICATIO NS, SUMMARY AND FUTURE WORK Potential Applications Utilizing defects as the mediating factor for ferromagnetism in a device structure at first appearance seems to be problematic. Typically, compound semiconductor materials are synthesized and processed to allow for the least possible defect densities for o ptimal device operation. The additional processing steps involved to provide the enhanced magnetic ordering described in this work would not be difficult for the current process technology. A very dilute amount of the magnetic impurity Gd was incorporated in GaN below the concentration levels commonly used for p-type or n -type dopi ng via ion implantation (peak Gd concentration of 9.73 1015 cm3). A larger amount of Si (peak Si concentration of 8.81 1017 cm3) was also incorporated in GaN and provided th e necessary defect formation to cause a spin orbit coupling between the defects and Gd atoms that greatly enhanced the magnetic ordering provided by the Gd incorporation alone. Both of these implantation steps can be easily added to the current GaN based m aterial s device processing techniques, providing a relatively smooth integration Based on the results of this work, an obstacle to overcome is utilizing the implanted material in a vertical orientation rather than the more standard horizontal orientation used by GaN -based high electron mobility transistors (HEMTs) and other GaN -based devices. This vertical orientation would take full advantage of the observed larger degree of ferromagnetic ordering when the applied field is perpendicular to the sample surf ace. A solution to this dilemma would be to use the ferromagnetic material as a vertical magnetic tunnel junction depicted in Figure 8-1 This device structure could operate as a spin valve, allowing spin transport when the ferromagnetic layers are

PAGE 165

165 aligned and allowing virtually no spin transport when the ferromagnetic layers are unaligned. Such a device structure could prove beneficial for switching applications with low power requirements since the spins can be quickly and easily manipulated. This device structure could also allow for higher switching speeds based on preliminary results showing that spin coherence in GaN possesses a life time in the picosecond range at room temperature.5 Summary In summary dilute magnetic semiconductor (DMS) materials continue to be studied for potential spintronic applications due to the fact that these materials exhibit above room temperature ferromagnetism and retain their semiconducting properties. The rare earth metal Gd has replaced the transition metals previously studied as the magnetic impurity in GaN due to the colossal magnetic moment seen in Gddoped GaN films. Ion implantation has more recently been pursued as the ideal incorporation method for Gd in GaN due to the exact control over the dopant inc orporation and higher magnetic signal exhibited by these materials. M etal organic chemical vapor deposition (MOCVD) grown undoped GaN thin films have been implanted with either Si, Gd or both Gd and Si ions to probe the effects of implantation on the magn etic properti es of the implanted materials. Nonimplanted, Si implanted, Gd -implanted, and Gdand Si -co implanted MOCVD u-GaN films exhibit ed ferromagnetic ordering at room temperature. C o -implanted MOCVD u -GaN exhibited an approximately 4times larger ma gnetic signal compared to that of Gd implanted MOCVD u -GaN. This enhancement effect is demonstrated when the implanted samples are analyzed with the magnetic field applied perpendicular to the sample surface (parallel to c axis of the sample, perpendicular orientation), signifying an anisotropic

PAGE 166

166 component to the magnetic properties of these materials, and described in this work as an anisotropic enhancement effect (AEE). The relatively large Gd ion radius and the large Si doses implanted into the GaN introd uced a multitude of defects along the c axis of the samples: vacancies near the surface region and interstiti als further into the material. The defects introduced by the Gd and Si implants interacting with the implanted Gd resulted in the largest magnetic signal of the implanted GaN samples being displayed by the co implanted sample. For non-implanted MOCVD p-GaN, the native defect VN has the lowest formation energy and is speculated as being a possible mechanism for ferromagnetism based on the magnetic measurements of this material. Also, the anion interstitial cannot be ruled out when considering defect -induced magnetism in GaN. Due to non equilibrium MOCVD growth under N -rich c onditions (V/III ratio of ~ 6000), a large concentration of nitrogen interstitials (Ni) are created. Both anion -related defects may play a role in the ferromagnetic ordering. For Gdimplanted MOCVD p -GaN, these defects appear to show a strong interaction wit h the implanted Gd and exhibit ed a large degree of ferromagnetic ordering. A similar interaction between t he native defects and defects introduced during implantation and the implanted Gd was not seen in Gd-im planted MOCVD n-GaN based on the almost identical hysteresis curves of the non -implanted and Gd-implanted n GaN in the perpendicular orientation at 10 K and 300 K For the MBE GaGdN sample, the Si implantation does increase the ferromagnetism of the sample due to defect formation, but not to the degree exhibited by the Gdand Si -co implantation of MOCVD u-GaN Therefore, for GaN based materials, the largest ferromagnetic ordering is exhibited by a material possessing a large population of defects (particul arly point defects such as VN

PAGE 167

167 and Ni) and the magnetic impurity Gd. These materials display a form of spin orbit coupling between the defects and Gd that is most strongly manifested when the magnetic properties of the material are analyzed with the applied field oriented perpendicular to the sample surface (parallel to the c axis of the wurtzite GaN crystal structure). ZnO and GaAs have also been considered as target materials for Gd ion implantation. N onand Gdimplanted ZnO samples showed ferromagnetism in both the parallel an d perpendicular applied field/sample surface orientations. The anion vacancy (VO) was speculated to be the defect responsible for the ferromagnetic ordering in the non-implanted ZnO due to the low formation energy of this defect. Als o, oxygen interstitials (Oi) may play a role in t he ferromagnetic ordering based on the presence of this defect due to nonequilibrium growth conditions. Once again, when Gd is incorporated, a large concentration of point defects (VO and Oi) is created. Due to the Gd implantation, t here is an enhancement in magnetic ordering, which is speculated to be a result of the defects and implanted Gd experiencing spin orbit coupling. The nonand Gdimplanted GaAs samples showed t he smallest magnetic signal of the compound semiconductor materials when analyzed in the perpendicul ar orientation. For GaAs the anion vacancy (VAs) has a relatively larger formation energy compared to the other point defects, which could explain the small magnetic signal exhibited by thi s material The arsenic interstitial (Asi) is the more prevalent point defect in GaAs and further studies are necessary to determine the effect this defect has on the measured ferromagnetism.

PAGE 168

168 For a visual comparison of the magnetic properties of the non-im planted compound s emiconductor materials, Figure 82 and Figure 83 show the hysteresis curves for the non-implanted GaN (undoped, ptype, and ntype), ZnO, and GaAs in the perpendicular orientation at 10 K and 300 K, respectively. MOCVD pGaN clearly demonstrates the largest magnetic signal of any of the nonimplanted materials and anion -related defects (VN and Ni) are speculated as being the species responsible for the observed ferromagnetism For a visual comparison of the magnetic properties of the impl anted compound s emiconductor materials, Figure 8-4 and Figure 8 -5 show the hysteresis curves for the implanted GaN (undoped, p-type, and n-type), ZnO, and GaAs in the perpendicular orientation at 10 K and 300 K, respectively. For the implanted materials, G d and Si -co implanted u -GaN clearly exhibits the largest deg ree of ferromagnetic ordering, with spin orbit coupling between the implanted Gd and the large concentration of defects postulated as being responsible for the ferromagnetism. Concerning a model for the ferromagnetic mechanism in these compound semiconductor materials, anion-related defects app ear to be the most likely defect s involved in the ferromagnetic ordering of the non implanted and implanted materials, based on the relationship between the magnetic properties and the presence of these defects in the materials studied for this work. This relationship is most clearly seen when the magnetic field is applied perpendicular to the sample surface, signifying there is an anisotropic factor involved in the magnetic ordering. An increased ferromagnetic signal is exhibited by materials containing the magnetic impurity Gd in addition to the defects due to possible spin orbit coupling interactions between these species.

PAGE 169

169 Future Work To fully develop the ferromagnetic mechanism in these compound semiconductor materials a deeper examination o f the defects present in the host material before and after implantation must be pursued. Focusing on GaN, a different orientation such as GaN grown on a different plane of sapphire, such as r plane, could be examined to determine how this plane changes the dislocation and defect densities and the magnetic properties of Gd -implanted and Gdand Si -co implanted GaN fi lms grown on this material. Along the same lines, epitaxial layer overgrowth (ELO) GaN can be studied to determine the effects a decreased dislocation and defect density can have on the magnetic properties Concerning the implant species, elements with large atomic radii that are ferromag netic can be explored to find the optimal element, dose, target material combination to maximize the ferromagnetism of the implanted material. In GaN, experimentation in Ga implantation and exploring various implant doses should be pursued to more fully understand the role defects play in the ferromagnetic ordering for materials not containing any magnetic impurities. To further probe the importance of anion -related defects in the ferromagnetic ordering, implantation studies could be performed on GaP. Since studies have been completed for GaN and GaAs, the magnetic properties of non-implanted and implanted GaP could further explain the exact ferromagnetic mechanism in these compound semiconductor materials. This material would be expected to exhibit magnetic properties similar to GaAs due to GaP also possessing a smaller band gap (2.26 eV) and a cubic crystal structure (zinc blende). The ult imate goal of this research would be to fully develop a fer romagnetic mechanism model that takes into account the specif ic host material, magnetic impurity type and concentration, and types and densities of defects to accurately predict the

PAGE 170

170 resultant ferromagnetic properties of dilute magnetic semiconductor materials The modeling would determine the specific magnetic properties needed for certain device applications and assist in determining the necessary method s of material synthesis magnetic and nonmagnetic dopant incorporation, and device fabrication to achieve the desired magnetic properties Based on that ferrom agnetic mechanism model, spintronic device s could be designed and fabricated to take full advantage of the spin transport provided by materials optimized for such applications.

PAGE 171

171 Figure 81. Schematic of vertical magnetic tunnel junction. The insulator layer acts as a tunnel barrier to allow spin transport based on whether the spins of the ferromagnetic layers are aligned or unaligned.

PAGE 172

172 -1000 -500 0 500 1000 -1.0x10-4-5.0x10-50.0 5.0x10-51.0x10-41.5x10-4 T = 10 K Perpendicular OrientationNon-implanted: u-GaN p-GaN n-GaN ZnO GaAs M (emu/cm2)H (Gauss) Figure 82. Magnetization vs. applied field for nonimplanted GaN (undoped, ptype, and n -type), ZnO, and GaAs samples at 10 K for the perpendicular sample surface/applied field orientation.

PAGE 173

173 -1000 -500 0 500 1000 -1.0x10-4-5.0x10-50.0 5.0x10-51.0x10-41.5x10-4 T = 300 K Perpendicular OrientationNon-implanted: u-GaN p-GaN n-GaN ZnO GaAs M (emu/cm2)H (Gauss) Figure 83 Magnetization vs. applied field for nonimplanted GaN (undoped, ptype, and n -type), ZnO, and GaAs samples at 300 K for the perpendicular sample surface/applied field orientation.

PAGE 174

174 -1000 -500 0 500 1000 -5.0x10-4-4.0x10-4-3.0x10-4-2.0x10-4-1.0x10-40.0 1.0x10-42.0x10-43.0x10-44.0x10-45.0x10-4 T = 10 K Perpendicular OrientationGd-implanted: u-GaN u-GaN (with Si) p-GaN n-GaN ZnO GaAs M (emu/cm2)H (Gauss) Figu re 84 Magnetization vs. applied field for implanted GaN (undoped, p-type, and ntype), ZnO, and GaAs samples at 10 K for the perpendicular sample surface/applied field orientation.

PAGE 175

175 -1000 -500 0 500 1000 -5.0x10-4-4.0x10-4-3.0x10-4-2.0x10-4-1.0x10-40.0 1.0x10-42.0x10-43.0x10-44.0x10-45.0x10-4 T = 300 K Perpendicular OrientationGd-implanted: u-GaN u-GaN (with Si) p-GaN n-GaN ZnO GaAs M (emu/cm2)H (Gauss) Figure 85 Magnetization vs. applied field for implanted GaN (undoped, p-type, and ntype), ZnO, and GaAs samples at 300 K for the perpendicular sample surface/applied field orientation.

PAGE 176

176 LIST OF REFERENCES 1 D. D. Awschalom, M. E. Flatte and N. Samarth, Sci. Am. 286, 66 (2002). 2 J. D. Albrecht and D. L. Smith Phys. Rev. B 68 035340 (2003). 3 P. R. Hammar, B. R. Bennett, M J. Yang and Mark Johnson, Phys. Rev. Lett. 83 203 (1999). 4 F. G. Monzon, H. X. Tang and M. L. Roukes, Ph ys. Rev. Lett. 84 5022 (2000). 5 B. Beschoten, E. Johnston-Halperin, D. K. Young, M. Poggio, J. E. Grimaldi, S. Keller, S. P. DenBaars, U. K. Mishra, E. L. Hu, and D. D. Awschalom, Phys. Rev. B 63 121202(R) (2001). 6 R. P. Davies; C. R. Abernathy; S. J. Peart on; D. P. Norton; M. P. Ivill and F. Ren Chemical Engineering Communications Volume 196, Issue 9, pages 1030 1053 (2009). 7 H. Ohno, J. Magnet. Mag. Mater. 200, 110 (1999). 8 H. Akinaga and H. Ohno, IEEE Trans. Nanotech. 1 9 (2002). 9 T. Dietl and H. Ohno, Physica E 9 185 (2001). 10. T. Dietl, H. Ohno, F. Matsukura, J. Cibert and D. Ferrand, Science 287, 1019 (2000). 11. H. Munekata, H. Ohno, S. von Molnar, A. Segmeuller, L. L. Chang and L. Esaki, Phys. Rev. Lett. 63, 18491852 (1989). 12. H. Ohno, H. Munekata, S von Molnar and L. L. Chang, J. Appl. Phys. 69 6103 (1991). 13. H. Ohno, Science 281, 951 (1998). 14. M. E. Overberg, B. P. Gila, G. T. Thaler, C. R. Abernathy, S. J. Pearton, N. Theodoropoulou, K. T. McCarthy, S. Arneson, A. F. Hebard, S. N. G. Chu, R. G. Wilson, J. M. Zavada, and Y. D. Park, J. Vac. Sci. Tech. B 20, 969 (2002). 15. M. Hashimoto, Y. K. Zhou, H. Tampo, M. Kanamura, and H. Asahi, J. Crystal Growth 252 499 (2003). 16. G. Thaler, R. Frazier, B. Gila, J. Stapleton, M. Davidson, C. R. Abernathy, S. J. Pearton, and C. Segre, Appl. Phys. Lett. 84, 1314 (2004).

PAGE 177

177 17. G. Thaler, R. Frazier, B. Gila, J. Stapleton, M. Davidson, C. R. Abernathy, S. J. Pearton, and C. Segre, Appl. Phys. Lett. 84, 2578 (2004). 18. R. Frazier, G. Thaler, M. Overberg, B. Gila, C. R. Abernat hy and S. J. Pearton, Appl. Phys. Lett. 83 1758 (2003). 19. M. Van Scilfgaarde and O. N. Myrasov, Phys. Rev. B 63, 233205 (2001). 20. M. E. Overberg, C. R. Abernathy, S. J. Pearton, N. A. Theodoropoulou, K. T. McCarthy and A. F. Hebard, Appl. Phys. Lett. 79 13 12 (2001). 21. S. Sonoda, S. Shimizu, T. Sasaki, Y. Yamamoto, and H. Hori, J. Crystal Growth 237, 1358 (2002). 22. M. Hashimoto, Y. K. Zhou, H. Tampo, M. Kanamura, and H. Asahi, J. Crystal Growth 252 499 (2003). 23. G. T. Thaler, Development of gallium nitride based dilute magnetic semiconductors for magnetooptical applications, http://purl.fcla.edu/fcla/etd/UFE0006460, Gainesville, FL (2004). 24. Y. -K. Zhou, M. Hashimoto, M. Kanamura, and H. Asahi, J. Supercond. 16, 37 (2003). 25. G. T. Thaler, R. M. Frazie r, C. R. Abernathy and S. J. Pearton, Appl. Phys. Lett. 86, 131901 (2005). 26. H. X. Liu, S. Y. Wu, R. K. Singh, L. Gu, D. J. Smith, N. Newman, N. R. Dilley, L. Montes and M. B. Simmonds, Appl. Phys. Lett. 85 4076 (2004). 27. G. T. Thaler, R. M. Frazier, C. R. Abernathy and S. J. Pearton, Appl. Phys. Lett. 86, 131901 (2005). 28. L. Gu, S. Y. Wu, H. X. Liu, R. K. Singh, N. Newman and D. J. Smith, J. Magnet. Mag. Mater. 290-291 1395 (2005). 29. J. Sol, An Introduction to the Optical Spectroscopy of Inorganic Solids (W iley, New York, 2005), p. 202. 30. J. Hite, G. T. Thaler, R. Khanna, C. R. Abernathy, S. J. Pearton, J. H. Park, A. J. Steckl and J. M. Zavada, Appl. Phys. Lett. 89 132119 (2006). 31. H. Asahi, Y. K. Zhou, M. Hashimoto, M. S. Kim, X. J. Li, S. Emura and S. Hasegawa, J. Phys.: Condens. Matter 16 S5555 (2004). 32. M. Hashimoto, A. Yanase, R. Asano, H. Tanaka, H. Bang, K.Akimoto, and H.Asahi, Jpn. J. Appl. Phys. 42 L1112 (2003).

PAGE 178

178 33. H. Bang, J. Sawahata, M. Tsunemi, J. Seo, H. Yanagihara, E. Kita, and K. Akimoto, "Struct ural and magnetic properties of Er doped GaN," Compound Semiconductors, 2003. International Symposium on 114 (2003). 34. J. M. Zavada, N. Nepal, C. Ugolini, J. Y. Lin, H. X. Jiang, R. Davies, J. Hite, C. R. Abernathy and S. J. Pearton, Appl. Phys. Lett. 91 054106 (2007). 35. S. Dhar, O. Brandt, M. Ramsteiner, V. F. Sapega, and K. H. Ploog, Phys. Rev. Lett. 94. 037205 (2005). 36. S. Dhar, L. Perez, O. Brandt, A. Trampert, K. H. Ploog, J. Keller and B. Beschoten, Phys. Rev. B 72 245203 (2005). 37. N. Teraguchi, A. Suzuk i, Y. Nanishi, Y. -K. Zhou, M. Hashimoto and H. Asahi, Solid State Commun. 122, 651 (2002). 38. Y. K. Zhou, S. W. Choi, S. Kimura, S. Emura, S. Hasegawa and H. Asahi, J. Supercond. Novel Magn. 20, 429 (2007). 39. J. K. Hite, R. M. Frazier, R. Davies, G. T. Thaler C. R. Abernathy and S. J. Pearton, Appl. Phys. Lett. 89 092119 (2006). 40. S. Gupta, A. Melton, E. Malguth, W. E. Fenwick, T. Zaidi, H. Yu and I. T. Ferguson, Mat. Res. Soc. Symp. Proc. 1111 D0304 71 (2009). 41. S. Dhar, T. Kammermeier, A. Ney, L. Perez, K. H. Ploog, A. Melnikov and A. D. Wieck, Appl. Phys. Lett. 89 062503 (2006). 42. S. Y. Han, J. Hite, G. T. Thaler, R. M. Frazier, C. R. Abernathy, S. J. Pearton, H. K. Choi, W. O. Lee, Y. D. Park, J. M. Zavada and R. Gwilliam, Appl. Phys. Lett. 88, 042102 (2006). 43. M. A. Khaderbad, S. Dhar, L. Prez, K. H. Ploog, A. Melnikov and A. D. Wieck, Appl. Phys. Lett. 91 072514 (2007). 44. F.Y. Lo, A. Melnikov, D. Reuter, A. D. Wieck, V. Ney, T. Kammermeier, A. Ney, J. Schrmann, S. Potthast, D. J. As and K. Lischka, Appl. Phys. Lett. 90 262505 (2007). 45. A. Svane, N. E. Christensen, L. Petit, Z. Szotek and W. M. Temmerman, Phys. Rev. B 74 165204 (2006). 46. C. Zener, Phys. Rev. 81 440 (1951). 47. T. Dietl, A. Haury and Y. Merle d'Aubign, Phys. Rev. B 55, R3347 (1997). 48. V. I. Litvinov, Phys. Rev. B 72 195209 (2005). 49. V. I. Litvinov and V. K. Dugaev, Phys. Rev. Lett. 86, 5593 (2001).

PAGE 179

179 50. A. Kaminski and S. Das Sarma, Phys. Rev. B 68 235210 (2003). 51. A. Kaminski and S. Das Sarma, Phys. Rev. Lett. 88 2472021 (2002 ). 52. G. M. Dalpian and S. H. Wei, Phys. Rev. B 72 115201 (2005). 53. A. Ney, T. Kammermeier, E. Manuel, V. Ney, S. Dhar, K. H. Ploog, F. Wilhelm and A. Rogalev, Appl. Phys. Lett. 90 252515 (2007). 54. L. Liu, P. Y. Yu, Z. Ma and S. S. Mao, Phys. Rev. Lett. 100 127 203 (2008). 55. Y. Gohda and A. Oshiyama, Phys. Rev. B 78, 161201 (R) (2008). 56. C. Mitra and W. Lambrecht, as presented at Symp. D of the Mat. Res. Soc. Fall Meeting 2008 on December 2, 2008. 57. S. Limpijumnong and C. G. Van de Walle, Phys. Rev. B 69, 035207 (2004). 58. C. Mitra and W. R. L. Lambrecht, Phys. Rev. B 80 081202(R) (2009). 59. J. K. Hite, R. M. Frazier, R. Davies, G. T. Thaler, C. R. Abernathy and S. J. Pearton, Appl. Phys. Lett. 89 092119 (2006). 60. J. K. Hite, R. M. Frazier, R. P. Davies, G. T. Thaler, C. R. A bernathy, S. J. Pearton, J. M. Zavada, E. Brown and U. Hmmerich, J. Electron. Mater. 36 391 (2007). 61. Y. K. Zhou, S. W. Choi, S. Emura, S. Hasegawa and H. Asahi, Appl. Phys. Lett. 92, 062505 (2008). 62. V. Bougr ov M. E. Levinshtein, S. L. Rumyantsev, and A. Zubrilov in Properties of Advanced Semiconductor Materials GaN, AlN, InN, BN, SiC, SiGe. Eds. M. E. Levinshtein, S. L. Rumyantsev and M. S. Shur John Wiley & Sons, Inc., New York, 1 -30, (2001). 63. J. F. Ziegler www.srim.org 64. Z. Xiong, L. Luo, J. Peng, and G. Liu, Journal of Physics and Chemistry of Solids 70 1223 -1225 (2009). 65. M. A. Reshchikov and H. Morko, J. Appl. Phys. 97 061301 (2005). 66. M. Venkatesan, C. B. Fitzgerald, J. G. Lunney and J. M. D. Coey, Phys. Rev. Lett. 93, 177206 (2004) 67. A. Dinia, G. Schmer ber, C. Meny, V. Pierron-Bohnes and E. Beaurepaire, J. Appl. Phys. 97 123908 (2005)

PAGE 180

180 68. A. J. Behan, A. Mokhtari, D. Score, X. -H. Xu, J. R. Neal, A. M. Fox, and G. A. Gehring, Phys. Rev. Lett. 100, 047206 (2008) 69. K. Potzger, S. Zhou, F. Eichhorn, M Helm, W. Skorupa, A. Mcklich, and J. Fassbender, J. Appl. Phys. 99 063906 (2006) 70. V. Ney, S. Ye, T. Kammermeier, A. Ney, H. Zhou, J. Fallert, H. Kalt, F.Y. Lo, A. Melnikov and A. D. Wieck, J. Appl. Phys. 104, 083904 (2008). 71. C.G. Van de Walle, Physic a B 308 310, pp. 899 903 (2001). 72. K. M Lukeny and R A Morrowz Semicond. Sci. Technol. 11 1156 1158 (1996) 73. D. T. J. Hurle, J. Appl. Phys. 85 6957 (1999) 74. K. M. Yu, M. Kaminska, and Z. Liliental -Weber, J. Appl. Phys. 72, 2850 (1992)

PAGE 181

181 BIOGRAPHICAL SKETCH Ryan P. Davies was born in Tampa, Fl orida in 1981. He graduated from H. B. Plant High School in 1999 and went on to att end the Honors Program at the University of Florida. Based on his enjoy ment of math and science subjec ts (especially physics) in high school, he decided to pursue a bachelors degree in the Department of Materials Science and Engineering. After graduating cum laude with a Bachelor of Science degree, Ryan returned to Tampa, looking unsuccessfully for a local engineering position. After working for a couple of years at a financial institut ion, Ryan rejoined Dr. Abernathys research group in the fall of 2005. He credits hi s enjoyment of research and learning about new technologies and materials as major motivating fa ctors in his return to school. His work focused on dilute magne tic semiconductor materials for spintronic applications and the design and asse mbly of a XPS/UPS system for in situ surface analysis of semiconductor mate rials. He received a Master of Science degree in 2007. Post-graduation, Ryan plans on pursuing postdoctoral opportunities at both the University of Florida and other research instit utions. His ultimate goal is to one day join the faculty of a research university explor ing the next innovative step in semiconductor materials and research.