Diluted Magnetic Semiconductor Cobalt and Europium Implanted Zno Thin Film

The University of New South Wales

Faculty of Science

School of Materials Science and Engineering

Diluted Magnetic Semiconductor Cobalt and Europium Implanted ZnO Thin Film

A Thesis in

Martials Science and Engineering

Pat Photongkam

Submitted in Partial Fulfilment

of the Requirement

For the Degree of

Doctor of Philosophy

August 2011

PLEASE TYPE THE UNIVERSITY OF NEW SOUTH WALES Thesis/Dissertation Sheet

Surname or Family name: Photongkam

First name: Pat Other name/s:

Abbreviation for degree as given in the University calendar: Ph.D.

School: School of Materials Science and Engineering Faculty: Science

Title: Diluted Magnetic Semiconductor Cobalt and Europium Implanted ZnO Thin Film

Abstract 350 words maximum: (PLEASE TYPE)

vacuum vapour arc (MEVVA) source ion implantation of cobalt and europium into ZnO/c-Al2O3 (0001) epitaxial thin films. The ion implantation is an effective technique for introducing dopants of heavy elements into thin film. The depth profile of as-prepared sample as well as dopants concentration was studied by ion beam analysis and transport of ions in matter (TRIM) calculation. It was found that the total magnet moment of Co doped ZnO was improved by

additional Eu doping. The correlation between the properties of Zn1-xEuxO and Zn1-xCox-yEuyO system and local coordination chemical environment as well as the underlying mechanism was investigated in details.

polarization of localized Eu atoms observed near surface of the Zn1-xEuxO thin films. The XMCD results also suggest Eu implanting to ZnO:Co system has suppressed Co metallic clustering. X-ray absorption fine structure (XAFS) confirms that Eu3+ had substituted for Zn2+ and resided in tetrahedral geometry without changing the wurtzite structure of ZnO 2+ host lattice in Zn1-xEuxO; whereas substitutional Co and Co metallic clusters are coexisting in Zn1-xCoxO. But the Co

clustering fraction can be significantly decreased by adding Eu into Zn1-xCoxO through ion implantation. The experimental and theoretical studies suggest a short range interaction between the substitutional Eu3+ is anti-

ferromagnetic in Zn1-xEuxO. The experimental results provide guidance to develop the new materials to enhance the intrinsic ferromagnetic properties of ZnO based DMSs via rare earth element implantation.

Declaration relating to disposition of project thesis/dissertation

I hereby grant to the University of New South Wales or its agents the right to archive and to make available my thesis or dissertation in whole or in part in the University libraries in all forms of media, now or here after known, subject to the provisions of the Copyright Act 1968. I retain all property rights, such as patent rights. I also retain the right to use in future works (such as articles or books) all or part of this thesis or dissertation.

I also authorise University Microfilms to use the 350 word abstract of my thesis in Dissertation Abstracts International (this is applicable to doctoral theses only).

…………………………………………………………… ……………………………………..……………… ……….……………………...…….… Signature Witness Date

The University recognises that there may be exceptional circumstances requiring restrictions on copying or conditions on use. Requests for restriction for a period of up to 2 years must be made in writing. Requests for a longer period of restriction may be considered in exceptional circumstances and require the approval of the Dean of Graduate Research.

FOR OFFICE USE ONLY Date of completion of requirements for Award:

THIS SHEET IS TO BE GLUED TO THE INSIDE FRONT COVER OF THE THESIS

Certificate of Originality

‘I hereby declare that this submission is my own work and to the best of my knowledge it contains no materials previously published or written by another person, or substantial proportions of material which have been accepted for the award of any other degree or diploma at UNSW or any other educational institution, except where due acknowledgement is made in the thesis. Any contribution made to the research by others, with whom I have worked at UNSW or elsewhere, is explicitly acknowledged in the thesis. I also declare that the intellectual content of this thesis is the product of my own work, except to the extent that assistance from others in the project's design and conception or in style, presentation and linguistic expression is acknowledged.’

Signed ……………………………………………......

Date ……………………………………………......

i i

‘I hereby grant the University of New South Wales or its agents the right to archive and to make available my thesis or dissertation in whole or part in the University libraries in all forms of media, now or here after known, subject to the provisions of the Copyright Act 1968. I retain all proprietary rights, such as patent rights. I also retain the right to use in future works (such as articles or books) all or part of this thesis or dissertation. I also authorize University Microfilms to use the 350 word abstract of my thesis in Dissertation Abstract International. I have either used no substantial portions of copyright material in my thesis or I have obtained permission to use copyright material; where permission has not been granted I have applied/will apply for a partial restriction of the digital copy of my thesis or dissertation.

Signed ……………………………………………......

Date ……………………………………………......

Authenticity Statement

‘I certify that the Library deposit digital copy is a direct equivalent of the final officially approved version of my thesis. No emendation of content has occurred and if there are any minor variations in formatting, they are the result of the conversion to digital format.’

Signed ……………………………………………......

Date ……………………………………………......

Acknowledgement

First, I would like to show my gratitude to my supervisor, Prof. Sean Li, whose constancy guidance, support, encouragement, and assistance me to complete my dissertation. It has been an honors and privilege to work with him.

I would also like to show my gratitude to following persons who have helped and supported me in different aspects to my work:

• Dr. Mihail Ionescu for his assistance on experimental at Australia Nuclear Research Technology and Organization (ANSTO) without his kindness which is impossible for me completing my work. • Dr. Dehong Yu for his guidance in my experiments with Taiwanese Synchrotron facilities. Without his support, it would not be able to secure the Synchrotron beam-time. • A/Prof. Alexey V. Pan for supporting on magnetic measurement at University of Wollongong • Dr. Yuebin Zhang for providing useful suggestion toward on my work. • Dr. Thiam Tech Tan for helping me on many aspects.

My appreciation goes to former and current staffs of School of Materials Science and Engineering UNSW and my colleagues.

Last but not least, I w ould like to show gratitude to my family for their support, understanding, patience and encouragement.

iii

Abstract

Diluted magnetic semiconductors (DMSs) have initiated enormous scientific interests because of their potential for multifunctional spintronics devices. ZnO based semiconductors have been identified to be the promising room temperature ferromagnetic materials with a wide band-gap. However, the intrinsic room temperature ferromagnetic spintronics materials are still far to be optimized. In this dissertation, the samples were prepared by using metal vacuum vapour arc (MEVVA) source ion implantation of cobalt and europium into ZnO/c-Al2O3 (0001) epitaxial thin films. The ion implantation is an effective technique for introducing dopants of heavy elements into thin film. The depth profile of as-prepared sample as well as dopants concentration was studied by ion beam analysis and transport of ions in matter (TRIM) calculation. It was found that the total magnet moment of Co doped ZnO was improved by additional

Eu doping. The correlation between the properties of Zn1-xEuxO and Zn1-xCox-yEuyO system and local coordination chemical environment as w ell as t he underlying mechanism was investigated in details.

The superconducting quantum interface device (SQUID) magnetometer shows all as- prepared samples are ferromagnetic at room temperature. However, it is unclear whether such a phenomenon is an intrinsic property or caused by t he Co metallic clusters. The X-ray magnetic circular dichroism (XMCD) shows that the strong spin polarization of localized Eu atoms observed near surface of the Zn1-xEuxO thin films. The XMCD results also suggest Eu implanting to ZnO:Co system has suppressed Co metallic clustering. X-ray absorption fine structure (XAFS) confirms that Eu3+ had substituted for Zn2+ and resided in tetrahedral geometry without changing the wurtzite 2+ structure of ZnO host lattice in Zn1-xEuxO; whereas substitutional Co and Co metallic clusters are coexisting in Zn1-xCoxO. But the Co clustering fraction can be significantly decreased by adding Eu into Zn1-xCoxO through ion implantation. The experimental and theoretical studies suggest a short range interaction between the substitutional Eu3+ is anti-ferromagnetic in Zn1-xEuxO. The experimental results provide guidance to develop the new materials to enhance the intrinsic ferromagnetic properties of ZnO based DMSs via rare earth element implantation.

List of Publication

1. P. Photongkam, Y. B. Zhang, M.H.N. Assadi, S. Li, D. Yu, M. Ionescu. A. V. Pan (2010) "Enhancement of Co substitution induced by Eu codoping in ZnO- based diluted magnetic semiconducting thin films." Journal of Applied Physics 107, 033909-033904. 2. M. Ionescu, P. Photongkam, D. Yu, R. Siegele, S. Li, D. D. Cohen (2010) “Doping of ZnO thin film with Eu using ion beams” Materials Science Forum, 638-642, 2962-2969 3. M.H.N Assadi, Y. B. Zhang, P. Photongkam, S. Li (2011) “Intrinsic ambient ferromagnetism in ZnO: Co induced by Eu codoping” Journal of Applied Physics 109, 4. M. H. N. Assadi, Y. B. Zhang, M. Ionescu, P. Photongkam, S Li (2009) “Intrinsic Ferromagnetism in Eu doped ZnO” Published in the proceedings of “16th Conference on Nuclear and Complementary Techniques of Analysis, AINSE, Lucas Heights, NSW” (arXiv:1006.3856).

v v

List of Abbreviation

-24 −1 µB Bohr magneton moment = 9.274 x 10 J·T

at% Atomic percentage

ALD Atomic Layer Deposition

CIP Current In Plane

CPP Current Perpendicular to Plane

CSD Charge State Distribution

CVD Chemical Vapour Deposition

DFT Density Function Theory

DMS(s) Diluted Magnetic Semiconductor(s)

EF Fermi Energy / Fermi Level

Eg Energy Band Gap

EL Electroluminescence

ERDA Elastic Recoil Detection Analysis

ERDA Heavy ion elastic recoil detection analysis

FC Field Cool

GMR Giant Magneto Resistance

HDD Hard Disk Drive

LDA Local Density Approximation

LED(s) Light Emitting Diode(s)

MBE Molecular Beam Epitaxy

MEVVA Metal Vacuum Vapour Arc

ML Monolayer = 1015 atom/cm2

MPMS Magnetic Property Measurement System

MRAM Magnetic Random Access Memory

MS Saturated Magnetization

MTJ(s) Magnetic Tunnel Junction(s)

PL Photoluminescence

PLD Pulsed Laser Deposition

RBS Rutherford Backscattering Spectrometry

RKKY Ruderman-Kittel-Kasuya-Yosida

SFET Spin Field Effect Transistor

SQUID Superconducting Quantum Interface Device

TC Curie Temperature

TEY Total-Electron-Yield

TFY Total-Fluorescence-Yield

TMR Tunnel Magneto Resistance

ToF Time-of-Flight

TRIM Transport in Matter

VSM Vibrating Sample Magnetometer

VUV Vacuum Ultraviolet

XAFS X-ray Absorption Fine Structure

XAS X-ray Absorption Spectroscopy

XMCD X-ray Magnetic Circular Dichroism

XRD X-ray diffraction

ZFC Zero Field Cool

vii

Table of Content

Certificate of Originality ...... i

Authenticity Statement ...... ii

Acknowledgement ...... iii

Abstract ...... iv

List of Publication ...... v

List of Abbreviation ...... vi

Table of Content ...... viii

Chapter 1 List of Figure ...... xii

List of Table ...... xix

Chapter 1 Introduction ...... 1

1.1 Background ...... 1

1.2 Research objective ...... 2

1.3 Outline ...... 4

Chapter 2 Literature Review ...... 6

2.1 Spintronics ...... 6

2.2 Application of spintronics ...... 6

2.2.1 Giant magneto-resistance (GMR) ...... 6

2.2.2 Tunnel magneto-resistance (TMR) ...... 8

2.2.3 Spin field effect transistor (SFET) ...... 10

2.2.4 Spin light emitting diodes (Spin LEDs) ...... 11

2.3 Diluted Magnetic Semiconductors ...... 12

2.4 Ferromagnetic Model in DMSs ...... 16

viii

2.4.1 sp-d exchange ...... 16

2.4.2 Super-exchange ...... 16

2.4.3 Double exchange ...... 18

2.4.4 Bound magnetic polarons...... 19

2.4.5 d0 Magnetism ...... 20

2.5 Current research efforts ...... 21

2.6 Properties of ZnO ...... 22

2.7 Defects in ZnO ...... 23

2.8 ZnO based DMSs ...... 24

2.9 Fundamental of X-ray spectroscopy ...... 26

2.10 X-ray Absorption Fine Structure (XAFS) ...... 28

2.10.1 Principle of XAFS ...... 28

2.10.2 Origin of the EXAFS oscillations ...... 29

2.10.3 The EXAFS equation ...... 30

2.10.4 XAFS data modelling ...... 31

2.11 X-ray magnetic circular dichroism (XMCD) ...... 32

2.11.1 Principle of XMCD ...... 32

2.11.2 XMCD sum rules ...... 34

2.12 Experimental efforts on XAFS and XMCD characterizations ...... 36

2.13 Summary ...... 38

Chapter 3 Experimental and Characterization Techniques ...... 39

3.1 Ion implantation ...... 39

3.2 Rutherford backscattering spectrometry (RBS) ...... 41

3.3 Heavy ion elastic recoil detection analysis (ERDA) ...... 42

3.4 Thin film X-ray diffraction ...... 43

3.5 Magnetic measurement with MPMS ...... 45

3.6 The band-gap energy (Eg) measurement ...... 48

3.7 X-ray absorption spectroscopy (XAS) ...... 50

3.7.1 X-ray absorption measurement ...... 51

3.7.2 X-ray absorption fine-structure (XAFS) ...... 53

3.7.3 X-ray magnetic circular dichroism (XMCD) ...... 54

3.8 Annealing ...... 56

Chapter 4 Depth Distribution Profiles of the Implanted Co and Eu ions in ZnO:Co and ZnO:Eu Thin Films ...... 57

4.1 Introduction ...... 57

4.2 Experimental details ...... 58

4.3 Result and discussion ...... 58

4.4 Summary ...... 68

Chapter 5 Eu Ion Implantation in ZnO Thin Films ...... 70

5.1 Introduction ...... 70

5.2 Experimental details ...... 70

5.3 Results and discussion ...... 72

5.4 Summary ...... 79

Chapter 6 Magnetic Properties of Eu implanted ZnO Thin Films ...... 80

6.1 Introduction ...... 80

6.2 Experimental details ...... 81

6.3 Result and discussion ...... 82

6.4 Summary ...... 90

Chapter 7 Enhancement of Co substitution induced by Eu codoping in ZnO-based Diluted Magnetic Semiconducting Thin Films ...... 91

7.1 Introduction ...... 91

x x

7.2 Experiment details ...... 92

7.3 Results and discussion ...... 93

7.4 Summary ...... 100

Chapter 8 Coordination Structure of Zn1-xCoxO, Zn1-xEuxO and Zn1-xCox-yEuyO Thin Films studies using XAFS ...... 101

8.1 Introduction ...... 101

8.2 Experimental details ...... 101

8.3 Result and discussion ...... 102

8.4 Summary ...... 116

Chapter 9 Conclusion ...... 117

9.1 Summary ...... 117

9.2 Future work ...... 119

9.3 Closing statement ...... 120

Reference ...... 121

List of Figure

Figure 2.1 This illustration shows a GMR device with two ferromagnetic (FM) layers separated by a n onmagnetic (NM) layer; when the magnetic fields in the two ferromagnetic layers are aligned, as shown in the diagram on t he left, half the electrons experience relatively little scattering, leading to a substantial reduction in resistance...... 7

Figure 2.2 Schematic representations of (a) CPP-GMR and (b) MTJ ...... 8

Figure 2.3 Schematic illustration of magnetic tunnel junctions (MTJs): (a) parallel and (b) anti-parallel orientation of magnetizations with the corresponding spin resolved density of the d states in ferromagnetic metals that have exchange spin splitting ∆ex. Arrows in the two ferromagnetic regions are determined by the majority spin sub- band. Dashed lines depict spin conserved tunnelling...... 9

Figure 2.4 Scheme of Datta-Das spin field effect transistor (SFET) ...... 10

Figure 2.5 Electrical spin injection in an epitaxially grown ferromagnetic semiconductor hetero-structure, based on G aAs. Spontaneous

magnetization develops below the Curie temperature TC in the ferromagnetic p-type semiconductor (Ga,Mn)As, depicted by the black arrows in the green layer. Under forward bias, spin-polarized holes from (Ga,Mn)As and unpolarised electrons from the n-type GaAs substrate are injected into the (In,Ga)As quantum well (QW, hatched region), through a spacer layer with thickness d, producing polarized electroluminescence...... 12

Figure 2.6 Some of the operative mechanisms for magnetic ordering in DMS materials...... 13

Figure 2.7 Computed value if the curie temperature TC for various p-type semiconductors containing 5% of Mn and 3.5 x 1020 holes/cm3...... 14

xii

Figure 2.8 Schematic of the super-exchange mechanism in MnO and Mn2O3 ...... 17

Figure 2.9 Double-exchange mechanism in mixed-valence manganites ...... 18

Figure 2.10 Schematic band structure of an oxide with 3d impurities and a spin split donor impurity band: (a) The position of the 3d level for low

Curie temperature TC, when the splitting of the impurity band is small, (b) and (c) show positions of the minority or majority-spin

3d bands, respectively, which lead to high TC...... 19

Figure 2.11 Stick and ball representation of ZnO crystal structures: (a) cubic rocksalt, (b) cubic zinc-blende and (c) hexagonal wurtzite. The shaded grey and black spheres denote Zn and O atoms, respectively ..... 22

Figure 2.12 Systematic diagram showing the X-ray absorption of core electron: the photoelectron excitation from core level to either (a) unoccupied state or (b) continuum. Both of these processes will create a core hole...... 27

Figure 2.13 Decay of excited state: x-ray fluorescence (left) and the Auger effect (right). In both case, the probability of emission (x-ray or electron) is directly proportional to absorption probability...... 27

Figure 2.14 XAS K-edge shows 2 major regions of XAS data ...... 28

Figure 2.15 (a) Schematic illustration of the photoelectric effect in terms of excitation of the different orbital (top) or different energy levels (bottom), (b) illustration of neighbouring shells of atoms about the absorber (top) and constructive (in phase) or destructive (out of phase) interference between outgoing photoelectron wave and backscattered off this shell of neighbouring atoms (bottom) ...... 30

Figure 2.16 Schematic diagram of X-ray magnetic circular dichroism (XMCD):

(a) L2,3 edge XAS spectra using left and right circularly polarization incident X-ray with magnetic field B parallel to - + incident X-ray direction. I and I representing µ(E) when helicity

xiii

of X-ray is anti-parallel and parallel to magnetic field respectively;

(b) Average µ(E) of L2,3 edge XAS spectra; (c) XMCD spectra; (d) Transition of 23pd→ absorption with right and left circularly polarization X-ray; and (e) Experimental set up for XMCD measurements...... 34

Figure 3.1 Schematic diagram of an ion implantation system which included an ion beam generator, beam manipulation and a process chamber ...... 40

Figure 3.2 Schematic representation of the experimental setup for Rutherford backscattering analysis ...... 41

Figure 3.3 Schematic representation of the experimental setup for Elastic recoil detection analysis ...... 43

Figure 3.4 Philips X'pert MRD Four circle, x-y-z translation, point or line focus ...... 44

Figure 3.5 Quantum Design MPMS-XL system components ...... 45

Figure 3.6 Basic scheme of MPMS Quantum Design: when a magnetic sample is displaced through a superconducting coils, a current I is induced which flows without losses to a t ransformer; the second circuit generate a flux in at third coil which will be detected by SQUID detector. Inset of this figure shows basic scheme of a SQUID thin film sensor, the two Josephson junctions, e.g.1 nm thick insulating alumina barriers...... 46

Figure 3.7 Superconducting solenoid internal detail showing: the configuration and location of the second-order gradiometer superconducting detection coil (inset). The coil sits outside of the sample space within the liquid helium bath. Basic scheme of MPMS Quantum Design...... 47

Figure 3.8 Schematic of optical system in PerkinElmer LAMBDA 950 ...... 49

xiv

Figure 3.9 The common methods for the measurement of X-ray absorption: (a) transmission geometry, (b) electron yield detection and (c) Florescence geometry. In the last column, it shows typical absorption spectra for each method...... 52

Figure 3.10 Schematic experimental layout of BL-17C at NSRRC ...... 53

Figure 3.11 Schematic optical layout of BL-11A at NSRRC ...... 54

Figure 3.12 Schematic structure shows the layout inside of the measurement chamber. Applied magnetic field axis has 30 degree angle to incident x-ray direction...... 55

Figure 4.1 The implanted ions distribution in ZnO target calculated by TRIM code with various mono-energetic from 10keV to 150keV: a) cobalt ions distribution in ZnO target and b) europium ions distribution in ZnO target...... 59

Figure 4.2 Stopping range (nm) of cobalt and europium in ZnO target as function of ions energy...... 60

Figure 4.3 The RBS spectra of Si1-xCox with various current integrator reading ..... 62

Figure 4.4 The RBS spectra of Si1-xEux with various current integrator reading...... 63

Figure 4.5 The dose calibration result for cobalt cathode and europium cathode: implanted dose for cobalt and europium in silicon (0001) wafer as function of current integrator...... 64

Figure 4.6 The RBS Spectra of ZnO implanted by c obalt ions (a) and europium ions (b) with implantation dose of 1.04 x 1016 atoms/cm2 using MEVVA ion implanter. The implantation voltage is 24 kV for cobalt ions and 45 kV for europium ions...... 66

xv xv

Figure 4.7 Ion depth distribution of all elements in ZnO implanted by cobalt (a) and europium (b) via RBS analysis ...... 67

Figure 4.8 Comparison between experimental and theoretical studies of implanted ions depth distribution in ZnO:Co 2.5at% and ZnO:Eu 2.5at% ...... 68

Figure 5.1 Schematic diagram of ERDA experiment showing the sample geometry and component of the Energy recoil ToF detector unit ...... 71

Figure 5.2 ERDA-ToF depth profile of ZnO/Al2O3 thin film ...... 73

Figure 5.3 ERDA ToF depth profile of the Eu-implanted ZnO/Al2O3 thin film ...... 73

Figure 5.4 X-ray diffraction patterns, 2θ-ω scan, of ZnO:Eu 1 at% - 4 at% ...... 74

Figure 5.5 Closed look of ZnO (002) of ZnO (0001) thin flim and Eu implanted ZnO film (bottom) ...... 75

Figure 5.6 Emission photoluminescence spectra at room temperature with 141.5mm excitation for ZnO and ZnO:Eu 4at% ...... 76

Figure 5.7 Transmission spectra for ZnO (0001) film and Eu implanted ZnO

film, using c-Al2O3 substrate as reference ...... 77

Figure 5.8 Plot of (αhν)2 vs. photo energy for ZnO (0001) flim and Eu implanted ZnO thin film. The optical bans gap of the films is determined according to Tauc Model ...... 78

Figure 6.1 Phase identification: XRD θ-2θ scan of ZnO/Al2O3 and Z n1-

xEuxO/Al2O3 (x = 0.01-0.04) ...... 82

Figure 6.2 Zero Field Cooled (ZFC)/Field Cooled (FC) DC magnetization

measurement of Zn0.96Eu0.04O/Al2O3 thin films using SQUID magnetometer with temperature changing rate of 10K/min...... 83

xvi

Figure 6.3 M-H curve of Eu implanted ZnO film at 300K measured by SQUID magnetometer, after the subtraction of diamagnetic

contribution of Al2O3 substrate. The external applied field was applied parallel to the film surface...... 84

Figure 6.4 Illustration showing crystallographic cross section of Eu implanted ZnO structure in configuration (a) Eu ions were separated by only one oxygen ion and (b) Eu ions were separated by a chain of -O- Zn-O- ions ...... 85

Figure 6.5 The x-ray absorption coefficient µ ()E of Zn0.96Eu0.04O at Eu M4,5-

edge measured at 77K and applied field of 1 Tesla in TFY mode. The inset of this figure displays M-H curve measured at same temperature using SQUID magnetometer ...... 87

Figure 6.6 The XAS and XMCD spectra of Zn0.96Eu0.04O at Eu M4;5-edges: left side is measured in TEY and right side is measured in TFY mode, the top spectra is XAS and bottom spectra is XMCD ...... 88

Figure 7.1 Distribution of Co and Eu ions as a function of depth. X axis is depth (nanometres) from the top surface and Y axis is Co or Eu dose at a particular depth...... 93

Figure 7.2 θ-2θ XRD patterns of virgin ZnO, Zn0.96Co0.04O, and

Zn0.92Co0.04Eu0.04O thin films. Inset shows the normalized (0002) peak...... 94

Figure 7.3 Magnetization curves of Zn0.92Co0.04Eu0.04O and Zn0.96Co0.04O thin films measured by a S QUID magnetometer at room temperature,

after the subtraction of diamagnetic contribution of Al2O3 substrate. The external applied field was applied parallel to the film surface...... 96

Figure 7.4 XAS spectra (top) and their derived XMCD spectra (bottom) of

Zn0.92Co0.04Eu0.04O measured at the Co L2,3 absorption edges at room temperature and 77 K...... 97

xvii

Figure 7.5 XMCD spectra of Zn0.92Co0.04Eu0.04O and Zn0.96Co0.04O measured

at the Co L2,3 absorption edges at 77 K...... 98

Figure 8.1 θ-2θ x-ray diffraction spectra for virgin ZnO, Zn0.96Co0.04O,

Zn0.96Eu0.04O and Zn0.96Co0.04Eu0.04O thin films ...... 102

Figure 8.2: The X-ray absorption coefficient µ(E) of epitaxial ZnO (0001) thin films at Zn K-edge (9659 eV)...... 103

Figure 8.3 Zn K-edge EXAFS spectra kχ(k) of epitaxial ZnO (0001) thin film as a function of the photoelectron wave vector k...... 104

Figure 8.4 The Fourier Transformation of k2χ(k) EXAFS spectra in R-space: (a) Magnetitude, (b) real part and (c) imaginary part ...... 106

Figure 8.5 Zn K-edge XAFS spectra comparison between ZnO and implanted ZnO thin film. (a) XANSE and (b) kχ(k) EXAFS ...... 108

Figure 8.6 The Fourier Transformation of k2χ(k) Zn K-edge EXAFS data (symbols) and fitting (line) for ZnO and implanted ZnO thin film...... 109

Figure 8.7 The Fourier Transformation of k2χ(k) EXAFS spectra for calculated

Zn L3-edge (line) and measured Eu L3-edge (symbols) for Eu2O3

powder, Zn0.96Eu0.04O and Zn0.96Co0.04Eu0.04O thin film as well as their EXAFS fitting (line) ...... 110

Figure 8.8 Eu L3-edge XANES of Eu2O3 powder, Zn0.96Eu0.04O thin film and

Zn0.96Co0.04Eu0.04O thin film ...... 111

Figure 8.9: Co K-edge XANES spectra for Zn0.96Co0.04Eu0.04O thin film and

Zn0.96Co0.04O, in comparison to reference Co metal and oxide compound of cobalt ...... 113

Figure 8.10: The Fourier Transformation of k2χ(k) EXAFS spectra data

(symbols) and fitting (line) for Zn0.96Co0.04Eu0.04O and

Zn0.96Co0.04O thin films ...... 114

xviii

List of Table

Table 5.1 The optical band-gap estimation values of sample by opt ical transmission photo spectrometer and estimation by gr ound state band structure in the Γ direction of the Brillouin zone ...... 79

Table 6.1 The total energy of Zn16Eu2O16 for parallel (FM) and anti-parallel (AFM) Eu-Eu coupling calculated by CAPTEP code and atomic distance for Eu-O and Eu-Eu before and after geometry relaxation ...... 86

Table 6.2 The sum rule analysis for orbital moment (µorb), spin moment

(µspin) and total spin (µtotal = µorb+ µspin) of europium ions in

Zn0.96Eu0.04O thin film measured in TFY and TEY mode ...... 87

Table 8.1 EXAFS fitting result for refinement of Zn K-edge for ZnO thin film ... 107

Table 8.2 ZnO, and implanted ZnO atomic distance and lattice parameter obtain by EXAFS analysis...... 109

Table 8.3 EXAFS fitting summary of Eu L3-edge for Zn0.96Eu0.04O thin film

and Zn0.96Co0.04Eu0.04O thin film ...... 112

Table 8.4: EXAFS fitting summary of Co K-edge for Zn0.96Co0.04O thin film

and Zn0.96Co0.04Eu0.04O thin film ...... 115

xix

Chapter 1 Introduction

1.1 Background

The silicon based devices have been evolving since 1950s [1]. The semiconductor industries have distinguished themselves both by r apid pace of performance improvement in their products, and by s teady path of constantly shrinking device geometries and increased functionality. The developing approaches of nowadays technologies have been following strategies of (1) reducing of technology scaling; (2) decreasing power consumption and (3) increasing operating frequency. Advance in optical lithography have allowed manufacturing of on chip structure with increasingly higher resolution. The area, power and speed characteristics of transistors with planar structure improve with decreasing in the lateral dimensions of the devices. In 1965, Moore predicted that complexity of integrated circuits has approximately doubled every year since their introduction [2] but this prediction was revised in 1975 suggesting a new slower rate of growth: “The number of transistors incorporated in a c hip will approximately double every 24 months” [3]. As the integration density (e.g. transistors and other components) is increased, devices have required more energy and have also generated more heat. These problems have become tremendous interests for both scientists and engineers to search for new alternative materials that may greatly increase the computing power and flexibility of future electronic devices while dramatically reducing their power consumption. These have become intensive researches and innovations during the past decades for alternative materials as a substitution for current silicon-based semiconductors.

The new materials are significant step toward the field of spin-based electronics or “Spintronics”. Conventional electronic devices use charge state of electrons to carry and transport information by varying the on/off state of electrical current. In contrast, spintronics devices exploit spin state of electrons to carry, manipulate and store information which considerably requires less power and consequently produces less heat. The technology has already been employed for storing information, such as storage hard drives and magnetic random access memories (MRAM), but using electron

1 1

spin states to process information through circuits would be a dramatic advance in computing. According to concept of spin field effect transistor (more detail in Chapter 2) proposed by Datta and Das (1990) [4], it requires a component which generates spin- polarized electron comprising either up or down state. One of solutions is to incorporate ferromagnetism into semiconductors to polarize electron’s spin; achieved by dopi ng magnetic ions into the diamagnetic host semiconductor in diluted limit, from which the concept of “diluted magnetic semiconductors” (DMSs) was raised. Ferromagnetism in DMSs is a cooperative phenomenon in which the localized spins of the dopants interacts among themselves mediated by carriers in the semiconductor. For spin-based devices to be functional at ambient condition, it is also crucial for developing DMSs materials to possess room temperature ferromagnetism.

1.2 Research objective

The experimental challenge in DMSs development is to achieve intrinsic ferromagnetic ordering with large spin polarization at room temperature. ZnO has become a potential candidate since it has been predicted to preserve room temperature ferromagnetism with certain dopants species, and has been already prospected in current optoelectronic application owing to its direct band gaps and its large excitation binding energy. Among of transition metal doped ZnO, cobalt doped ZnO has been reported as it has ferromagnetism with Curie temperature (TC) above room temperature however the origin of magnetism is still controvertible. It seems that strong localized spin polarization can be only observed when the extrinsic ferromagnetism, such as co balt clustering and cobalt secondary phase, was presented whereas intrinsic ferromagnetism, Co substituting Zn in lattice, is relative weak. The research objective is to develop intrinsic ferromagnetic ZnO based DMSs thin films which has strong localized spin polarization by doping two different elements, also known as codoping. Recent studies show that codoping technique has improved ferromagnetic properties in certain ZnO based semiconductor such as Zn(Mn,Cu)O [5] and Zn(Fe,Cu)O [6].

2 2

The aim of work is to enhance ferromagnetism of exciting ferromagnetic Co doped ZnO with addition Eu ion doping. Europium is of interest as dopant in ZnO due to its strong 4f orbital moment and its optical properties. The main challenge is to understand of magnetic properties and local structure around cobalt and europium ions in host structure. Therefore, it is involved lot of experimental work and related fundamental acknowledgement. In this thesis, it has been focused on f abrication of room intrinsic ferromagnetic Zn(Co,Eu)O semiconductor thin film using ion implantation technique, and their properties. The specific objectives of the thesis can be highlight as following

1. To achieve intrinsic ferromagnetic DMSs materials which have TC above room temperature with strong localized spin polarization. In this work, it focuses on developing the advanced DMSs materials through cobalt and europium codoped ZnO thin film prepared by ion implantation.

2. To study Zn1-xEuxO as a potential for room temperature ferromagnetic DMSs. In contrast with the transition metal doped ZnO, little attention has been paid to rare earth metal doped ZnO with respecting to ferromagnetic properties. In addition, Rare earth doped ZnO is promising for use in optoelectronic applications because of the defect luminescence.

3. To study local structure and localized spin polarization of Zn1-xCox-yEuyO in

comparing to Zn1-xCoxO and Zn1-xEuxO, and understand mechanism of its magnetic behaviour.

4. To identify the coordinated chemical environment of the specific atoms at atomic scale with X-ray absorption fine structure (XAFS)

5. To reveal the mechanism of magnetic enchantment with X-ray magnetic circular dichroism (XMCD)

3 3

1.3 Outline

In Chapter 1, the technological emergence for spintronics is emphasized. By taking major criteria of diluted magnetic semiconductor (DMSs) for practical application and experimental challenges into consideration, this dissertation aims of this dissertation is to realize a high performance room temperature intrinsic ferromagnetic DMSs.

Chapter 2 provides broad background information related to the field of spintronics as well as introduction on fundamental on X-ray absorption fine structure (XAFS) and X- ray magnetic circular dichroism (XMCD). The major criteria of DMSs for practical application are highlighted and a comprehensive survey of experimental studies on ZnO based DMS are discussed.

Chapter 3 b egins with the introduction of Metal Vacuum Vapour Arc (MEVVA) ion implanter which is used to prepare Zn1-xCox-yEuyO thin film. The fundamental principles of experimental technique employed in characterizing the as-prepared samples are also discussed. Finally, the synchrotron beamlines configuration for XAFS and XMCD experiment are layout.

In Chapter 4, the ion depth distribution profile of the implanted cobalt and europium ions are theoretically and experimentally investigated. The outcome of this study is to optimize the implanting condition for ZnO thin films and to understand the nature of dopants distribution profile, thus providing the critical information for experimental investigation.

In Chapter 5, the physical properties and optical properties of the europium implanted ZnO thin film are characterized. Elastic Recoil Detection Analysis (ERDA) is employed for study depth distribution of elements in the samples. The europium concentration dependence of energy band gap (Eg) is also investigated combined with ground state band energy calculation

4 4

In Chapter 6, the magnetic properties of Zn1-xEuxO are discussed with the combination of experimental results obtained by bot h magnetic measurement by S QUID magnetometer and the localized spin polarization measurement by X-ray magnetic circular dichroism (XMCD). Such investigation aims to understand the underlying mechanism of magnetic behaviours in Zn1-xEuxO.

In Chapter 7, the magnetic of ZnO:Co with Eu implanting addition were studied with X- ray diffraction, SQUID magnetometer and XMCD They provide crucial information for understanding the localized spin polarization nature of cobalt in ZnO. The experimental results suggest that the saturated magnetization of Co substitution in ZnO host can be enhanced significantly by addition of Eu through doping technique

In Chapter 8, the coordination local structure around specific atoms and their valence states in Zn1-xCox-yEuyO, Zn1-xCoxO and Zn1-xEuxO, are revealed by X-ray absorption fine structure (XAFS). The experimental results are used for understanding the origin of ferromagnetism in the fabricated materials.

Chapter 9 summarizes all the experimental finding and mechanisms of ferromagnetism in the as-prepared Zn1-xCox-yEuyO, Zn1-xCoxO and Zn1-xEuxO. Future experimental works, which are essential for further understanding and enhancement magnetism, are also proposed.

5 5

Chapter 2 Literature Review

2.1 Spintronics

“Spintronics” (spin transport electronics), also known as magneto-electronics, is defined as the technology which utilizes the spin of an electron besides its charge state as a carrier of information in electronics. It emerged from discoveries in the 1980s of spin dependent electron transport phenomena in solid state physics, e.g. polarized electron injection from a ferromagnetic metal [7] and discovery of giant magnetoresistance (GMR) [8, 9] Electrons are spin 1/2 fermions, and therefore, constitute a two- state system with up-spin and down-spin. Spin orientation of electron survive relative long time (nanoseconds) comparing to electron momentum decays (ten of femtosecond [10-15]). Due to this unique advantage of spin degree of freedom over charge degree of freedom, it is believed that spintronics will deliver future generations of lower power electronics, while providing the most feasible route to the realization of a quantum computer.

2.2 Application of spintronics

2.2.1 Giant magneto-resistance (GMR)

The 2007 Nobel Prize in physics was award to two physicists, A. Fert and P. Grünberg, who discovered GMR effect. In typical GMR device, two or more ferromagnetic layers are separated by a thin layer of non-magnetic spacer. At a particular spacer thickness, it makes magnetization vectors energetically preferable to align in anti-parallel which can be attributed to the quantum spin-spin coupling of the Ruderman-Kittel-Kasuya-Yosida (RKKY) interaction [10]. The magneto-resistive effect is observed as significant change in electrical conductivity depending on w hether the magnetization vectors of ferromagnetic layers are parallel or anti-parallel. The resistance is relatively larger in the anti-parallel alignment than in the parallel alignment due to differential spin scattering event of electrons. For current in plain (CIP) geometry, where current flows parallel to layers, GMR effects can be only observed for layers thinner than the mean free path of electrons, practically for thicknesses in the nm range. In the current perpendicular to

6 6

plane (CPP) geometry, the GMR is not only definitely higher than in CIP but also subsists in multi-layers with relatively thick layers, up to the micron range.

FM FMNM FM FMNM

Spin Spin

ρ ρ ρ ρ Spin ↑↑ ↑↑ Spin ↑↑ ↑↓

ρ↓↑ ρ↓↑ ρ↓↑ ρ↓↓

Figure 2.1 This illustration shows a GMR device with two ferromagnetic (FM) layers separated by a nonmagnetic (NM) layer; when the magnetic fields in the two ferromagnetic layers are aligned, as shown in the diagram on the left, half the electrons experience relatively little scattering, leading to a substantial reduction in resistance.

GMR can be qualitatively understood in term of Mott’s model [11] assuming that the electron conductivity can be described in terms of two largely independent conducting channels, corresponding to the up-spin and down-spin electrons. The principle is depicted in Figure 2.1 for the simplest case of a san dwich structure ferromagnetic | nonmagnetic metal | ferromagnetic. The spin scattering is strong for elections with spin anti-parallel to the magnetization direction and is weak for elections with spin anti- parallel to the magnetization direction. For parallel aligned magnetization vectors, the

7 7

up-spin electrons pass through the structure almost without scattering, because their spin is parallel to the magnetization of the layers. On the contrary, the down-spin electrons are scattered strongly within both ferromagnetic layers, because their spin is anti- parallel to the magnetization of the layers. The total resistivity appears to be low. For the anti-parallel aligned magnetization vectors, both up-spin and down-spin electrons are scattered strongly in one ferromagnetic layer or the other. Therefore, in this case the total resistivity of the multilayer is high. The mechanism can be also explained by spin scattering is the density of states theory of Edwards, et al. [12]

2.2.2 Tunnel magneto-resistance (TMR)

Tunnel magneto-resistance (TMR) is a m agneto-resistive effect in magnetic tunnel junctions (MTJs). In contrast with CPP-GMR structure, it c onsists of two layer ferromagnetic metals separated by thin insulating barrier, typically a few nanometres of aluminium oxide, showing in Figure 2.2. In quantum mechanics point of view, the wave property of electron can tunnel through the barrier if a bias voltage is applied between two ferromagnetic metal layers. This tunnelling current depends exponentially on height and thickness of barrier but in MTJ, also depends on t he relative orientation of magnetizations of the two ferromagnetic metal layers.

Figure 2.2 Schematic representations of (a) CPP-GMR and (b) MTJ

In other word, TMR is a dramatic change of the tunnelling current in MTJ when relative magnetizations of the two ferromagnetic layers change their alignment. TMR can be

8 8

understood in term of Jullière's model assuming that tunnelling of up- and down-spin electrons are two independent processes, so the conductance occurs in the two independent spin channels as shown in Figure 2.3 If magnetization vectors of two ferromagnetic layers are parallel, the minority spin (down-spins) tunnel to the minority (down) states and the majority spins (up-spins) tunnel to the majority (up) states. On the other hand, if magnetization vectors of two ferromagnetic layers are anti-parallel then the identity of the majority- and minority-spin electrons is reversed. Therefore, the majority spins (up spins) of the first layer tunnel to the minority (up) states in the second film and vice versa.

Figure 2.3 Schematic illustration of magnetic tunnel junctions (MTJs): (a) parallel and (b) anti-parallel orientation of magnetizations with the corresponding spin resolved density of the d s tates in ferromagnetic metals that have exchange spin splitting ∆ex. Arrows in the two ferromagnetic regions are determined by the majority spin sub-band. Dashed lines depict spin conserved tunnelling. [13]

Nowadays MTJs that are based on transition-metal ferromagnets and Al2O3 barriers can be fabricated with reproducible characteristics and with TMR values up to 50% at room temperature. Recently large values of TMR observed in crystalline MTJs with MgO

9 9

barriers further boosted interest in spin dependent tunnelling. MTJs are nowadays used in magnetic random access memories (MRAM).

2.2.3 Spin field effect transistor (SFET)

The spin field effect transistor (SFET) was first proposed by Datta and Das [4]. The principle of spin transistor can be illustrated in Figure 2.4. The device involves a ferromagnetic metal or ferromagnetic semiconductor source (spin injector) and drain (spin detector) with parallel magnetic moments. Spin–polarized electrons can be injected into the channel, with linear momentum along k direction; whereas the electron spin can process along the procession vector Ω, the magnitude of which can be controlled by the gate voltage VG that in terms determines the procession time. If the procession time is smaller than the time of electron flight, the electrons arrive at the drain will point at the same direction as in the source, with give rise to a large current; otherwise, the current is small.

Figure 2.4 Scheme of Datta-Das spin field effect transistor (SFET) [13]

The potential benefit of this device is that it is conceivably possible to perform this operation at much lower currents and at higher speeds than traditional FETs. This could have a g reat impact on the overall development of spin-based devices. T he crucial component in a functioning spin transistor is the source which could spin-polarize the electrons. One solution is to use ferromagnetic material which has a large net magnetic moment. However it is much more desirable to fabricate spin transistors based on

semiconductors that can be readily incorporated into current device fabrication technologies. Whereas ferromagnetsemiconductors junctions typically suffer interfacial dissimilarity which leads to overall decrease in spin–injection efficiency, the emerging field of diluted magnetic semiconductors (DMS) becomes a potential solution to overcome the problems and eventually lead to a functional device.

2.2.4 Spin light emitting diodes (Spin LEDs)

In traditional light emitting diodes (LEDs), photon emission is created by recombination of charge carriers flowing across p-n junction, when higher conduction electrons fall into lower valence band holes. The wavelength of photon depends on band-gap energy

(Eg) of direct band-gap semiconductors used to produce. Normal LEDs emit non- polarized photon due to statistic nature of carriers spin state. If we were able to injected spin polarized carriers into the device then it would be possible to emit circularly polarized light

Ohno et al. [14] successfully fabricated spin LEDs using p-type ferromagnetic as spin injector. A schematic diagram of device structure is displayed in Figure 2.5. A 300 nm

Ga0.955Mn0.045As was grown GaAs spacer/ 10 nm In0.13Ga0.87As quantum well (QW). The hysteresis is maintained up t o 52K in the electroluminescence measurement. Photoluminescence measurements excited through the substrate show no corresponding hysteresis, proving the emitted light polarization is indeed due to spin injection from the

Ga1-xMnxAs layer. GaAs spacers of up to 220 nm were used without appreciable loss in the observed spin polarization of the device. Unfortunately, due to the low TC of the device, spin injection is only observable at cryogenic temperatures. I n addition, although there is clear hysteresis observable in these samples, the degree of spin polarization is only 1% at 6K, owing to the fast relaxation time of the holes in the quantum well. However, this work demonstrates that polarized light emission is possible in the absence of a magnetic field with a suitable magnetic spin injector.

Figure 2.5 Electrical spin injection in an epitaxially grown ferromagnetic semiconductor hetero-structure, based on G aAs. Spontaneous

magnetization develops below the Curie temperature TC in the ferromagnetic p-type semiconductor (Ga,Mn)As, depicted by the black arrows in the green layer. Under forward bias, spin- polarized holes from (Ga,Mn)As and unpolarised electrons from the n-type GaAs substrate are injected into the (In,Ga)As quantum well (QW, hatched region), through a spacer layer with thickness d, producing polarized electroluminescence. [14]

2.3 Diluted Magnetic Semiconductors

Most of semiconductor based spintronics materials involve doping magnetic ions into non-magnetic semiconductor to actualize ferromagnetism. The magnetic properties of the ferromagnetic semiconductors are dependent to carrier concentrations in many cases [15, 16]. Whether the magnetic impurities in the semiconductor couple with each other ferromagnetically or anti-ferromagnetically depends on the carrier concentration. Therefore, it is possible to control magnetism in semiconductors by varying the carrier concentration either electrically or optically. Moreover, due to the fact that ferromagnetic ordering in semiconductor arises from short-range interactions between the dopants, it is possible to develop dense magnetic storage devices via spintronics technology. This not only offers insights into the development of new generation optical semiconducting devices, but also provides solutions to the problems of high power consumption in current semiconductors.

Diluted magnetic semiconductors (DMSs) are nonmagnetic semiconductors doped by magnetic ions, such as 3d transition metal and 4f lanthanides, dilutely and are expected to have coexisting of ferromagnetism and semiconducting properties [17]. DMSs have been interested research area because of their unique electrical and magnetic properties for “spintronics” application [18-20]. Such of their properties allow the control of electron’s spin as well as charge transportation in DMSs. Possible applications of DMS based spintronics devices include spin field effect transistors with faster switching time and lower power consumption with application in solid state quantum computing and spin-polarized light emitting diodes (Spin-LEDs) [21, 22]

Magnetic ions in DMSs can be referred as net of non-zero localized magnetic moment. According to model proposed by Zener based on mean field theory [23, 24], DMSs is more or less like a random alloy and ferromagnetism occur through interactions between the localized magnetic moment, mediated by host carriers (electrons or holes). This principle is illustrated in Figure 2.6.

Figure 2.6 Some of the operative mechanisms for magnetic ordering in DMS materials – full details discussed in Ref [23].

For example, In (Ga,Mn)As, Mn2+ become an accepter when it replace one of lattice constituents and occupy one lattice suite. At high Mn2+ concentration (Figure 2.6A), all Mn2+ are anti-ferromagnetic coupled. It is however in the diluted limit with large holes (carriers) concentration, the holes can mediate the parallel alignments of the magnetic 2+ moment of Mn separated large distances. The ferromagnetic ordering temperature TC is thus derived from the overall balance between ferromagnetic and anti-ferromagnetic coupling strength (Figure 2.6B). In the low carrier concentration regime, below a certain temperature, a percolation network is formed among the Mn ions; two possible mechanisms could be applied to explain the ferromagnetism. The first one suggested the ferromagnetic coupling is mediated by the delocalized free holes (Figure 2.6C); the second one, at the percolation limits, the ferromagnetic coupling is stabilized by electrons constrained within the percolation ring (Figure 2.6D). This shows that the magnetic state of the magnetic ions in DMS depends on the carrier concentration which gives rise to the well-known properties of “carrier–induced ferromagnetism” in DMSs. These properties allow electron to be manipulate for both spin and charge degree of freedom.

Figure 2.7 Computed value if the curie temperature TC for various p-type semiconductors containing 5% of Mn and 3.5 x 1020 holes/cm3.[18]

Since the idea of DMSs has been proposed, various DMSs materials have been investigated. Those of most research interests at the earlier stage includes (Ga,Mn)As [25-38], (In,Mn)As [39-44], Mn2+ doped S, Se, Te [45] based compound. The main challenge in this field is to achieve DMSs which have a transition/curie temperature TC above room temperature for practical application.

In 2000, Dielt et al. published a theoretical work based on Zener model which predicted that GaN and ZnO could achieve ferromagnetism well above room-temperature with p- type doping [18], as shown in Figure 2.7, further drives the research into GaN and ZnO as potential DMS materials. For those early stage materials such as (Ga,Mn)-BV, room temperature ferromagnetism has been reported. For examples, ferromagnetic behaviour of molecular beam epitaxy (MBE) derived (Ga,Mn)N film with 9 at.% Mn up to 750 K has been reported [46, 47]; whereas (Ga,Mn)P shows to persist ferromagnetism up to 330 K when the material is epitaxially derived, and 300 K when it is derived from ion implantation [15, 48]. As a consequence of this, (Ga,Mn)-BV type materials make very ideal candidates for spintronics materials. However, one major disadvantage of (Ga,Mn)-BV materials is that they occur naturally as indirect band-gap materials, although direct band-gap could be achieved by material modification and band-gap engineering, it will more or less alter the property of the original material which is less desired. Other semiconductor materials that had been investigated for potential DMS application include (Cd,Mn)GaP2 [49], (Zn,Mn)GeP2 [50], ZnSnAs2 [51], (Co,Ti)O2 [52] and so on.

2.4 Ferromagnetic Model in DMSs

2.4.1 sp-d exchange

The original idea has been provided by Zener [24] which describe interaction between s-like conduction band or p-like valence band and localized d shell of magnetic dopants

[53]. For EF locating near top of valence band, d electron shell – say spin down – is deep in the valence band and spin out hole above Fermi level. Hybridization (level repulsion of like-spin state) pushes energy of spin down p-like valence band up relative to spin up as resulting in anti-ferromagnetic coupling between holes carriers and dopants magnetic moment. The interaction is called p-d exchange interaction. In case EF locating near bottom of conduction band, hybridization pulls energy of spin down s-like conduction band down. The interaction is called s-d exchange interaction. The exchange constants are caused by direct exchange interaction between sp band and 3d electrons and hybridization between the sp band and 3d orbital [45] The s-d exchange interaction Nα is smaller than the p-d exchange interaction Nβ, where N is the number of cations per unit volume, and is almost independent of host material with tetrahedral symmetry. It is likely that the sp-d exchange interaction is long ranged because the carriers are itinerant.

2.4.2 Super-exchange

Super-exchange (also known as Kramers-Anderson super-exchange) was first proposed by Kramers in 1934 [54] and the theory was later developed by Anderson in 1950 [55]. The super-exchange interaction is usually strong anti-ferromagnetic coupling between two next-to-nearest neighbour cations through a non-magnetic anion. According to Anderson’s description, the super-exchange interaction acts as the conditions are met

[55]. The anti-ferromagnetic super-exchange of MnO and Mn2O3 are illustrated in Figure 2.8. The O2- has filled 2p6 orbital, acting as p-ligand, and Mn2+ has half-filled 5 5 3d orbital. The overlapping of 3d and p-ligand, due to bonding, leads to anti- ferromagnetic coupling between Mn2+, as shown in Figure 2.8a: electron spin in p- 2+ 5 ligand need to anti-parallel to Mn 3d for lowing kinetic energy. For Mn2O3, pairs of empty Mn3+ 3d state coupled by p-ligand are also anti-ferromagnetic as shown in Figure

2.8b. In this case, p-ligand spin align parallel to Mn3+ orbital moment, donating electron to empty 3d state. The actual electron transfer over the participating orbitals has been confirmed by measurements of the hyperfine interaction of the ligand nuclear spin with the electronic spins of the magnetic ion. [56]

a Mn2+ O2- Mn2+

b Mn3+ O2- Mn3+

Figure 2.8 Schematic of the super-exchange mechanism in MnO and Mn2O3

However the case o f europium chalcogenides (e.g. EuS) and chromium spinel (e.g.

ZnCr2Se4) implies that the super-exchange is not always anti-ferromagnetic and that ferromagnetism is not always related to presence of free carriers despite the large magnetic ion concentration. In the case of rock-salt Eu compound, there appears to be competition between anti-ferromagnetic cation-anions-cation and ferromagnetic cation- cation super-exchange. The latter can be traced back to the ferromagnetic s-f coupling, and the presence of s-f hybridization, which is actually stronger than p-f hybridization due to symmetry reason [57]. In such a situation, the lowering of the conduction band associated with ferromagnetic order enhances the energy gain due to hybridization. The Cr-spinels represents the case, in which, the d orbitals of the two cations are not coupled to the same p-ligand [58-60], resulting – in agreement with the Goodenough-Kanamori rules [61-63] – in a net ferromagnetic super-exchange.

2.4.3 Double exchange

In some oxide compound, ferromagnetic exchange interaction is due to mixed valence number of magnetic ions. The basis has been proposed by Z ener for ferromagnetic compounds of manganese with perrovskite structure [64] such as La1−xCaxMnO3 [65] and La1−xSrxMnO3 [66]. The oxidation of La is +3 but that of dopants ions: Ca and Sr are +2 which imply mix valance number of Mn3+ (d4) and Mn4+ (d3) at faction of 1-x and x respectively. Both Sr2+ and Ca2+ doping introduce holes in part of Mn 3d state. The d electron is hopping from Mn3+ to the adjacent Mn4+ via central O2- is considered as Mn3+ electron transferring to central O2- simultaneously with O2- electron transferring to Mn4+. The mechanism is called “double exchange” as shown in Figure 2.9. Combined with the on site Hund’s coupling, the double exchange interaction leads to a ferromagnetic alignment of the local magnetic moments [67]. Parallel spin alignment is favoured because it increases the hopping probability and therefore decreases the kinetic energy of spin-polarized electrons. The double exchange interaction gains energy through a d-electron hopping and occurs when EF crosses d band. It is likely that double exchange interaction is short ranged because carriers have d-band character.

Figure 2.9 Double-exchange mechanism in mixed-valence manganites [68]

In contrast with super-exchange, (1) the electrons do move between two cations ion via the intermediate ligand (e.g. oxygen) in double exchange but they do not in super- exchange. (2) In super-exchange, the ferromagnetic or anti-ferromagnetic alignment occur between two cations with same valence state; while double-exchange interaction occurs only when valence state of cation atom is different to that of another mixed valence compound.

2.4.4 Bound magnetic polarons

The bound polarons model describes ferromagnetism in low carrier concentration DMSs is due to interaction between localized spin of magnetic dopants and charge carriers which leads formation of polarons. This model is applicable to both p- and n-type host semiconductor [69] and is inherently attractive for low carrier density systems such as many of the electronic oxides.[70]. The expansion of these polarons is increased as the temperature is lowed and transition temperature occurs essentially when the polarons size is the same as that of the sample. The overlap of the individual polarons produces long range interactions and energetically it is favourable for the spin polarization to develop. Kaminski and Sarma [71] have developed an analytic polaron percolation theory for DMS’s in the limit of low carrier density and have obtained reasonable agreement with experimental results for high TC.

Figure 2.10 Schematic band structure of an oxide with 3d impurities and a spin split donor impurity band: (a) The position of the 3d level for

low Curie temperature TC, when the splitting of the impurity band is small, (b) and (c) show positions of the minority or majority-

spin 3d bands, respectively, which lead to high TC. [70]

Coey et al. [70] has proposed impurity band exchange explaining bound magnetic polarons in ZnO based DMSs. The carrier tend to localize around the magnetic formation of an impurity band as shown in Figure 2.10a. If minority or majority of dopants 3d state lie in the impurity band, as respectively shown in Figure 2.10b and c, then interaction between hydrogenic electrons in impurity band interact and magnetic dopants form the polarons, resulting in high TC. The magnetization of Zn0.95TM0.05O (TM = Sc - Cu) thin films prepared by pulsed laser deposition suggest 3d state do not overlaps to impurity band for Cr, Mn and Cu doped ZnO

2.4.5 d0 Magnetism

The term of d0 magnetism were used to describe ferromagnetic behaviour in semiconductor and insulator which free of transition metal or rare earth elements [72, 73]. For example, graphite possess weak ferromagnetism with magnetic moment of 10-5

µB/C [74] and The room temperature ferromagnetism was also observed in undoped 0 HfO2 [75]. One widely discussed scenario for functionalizing oxides as d dilute magnetic semiconductors involves the magnetic polarization of valence states by t he substitution of oxygen with sp-type impurities of lower valency [76]. d0 magnetism is induced by s pin-polarized defects or H impurity acting as l ocal spin rather than unpaired f- or d-shell electrons [77]. In general defects and impurities in semiconductors introduce levels in the band-gap and electrons that fill these levels form states with total spin of 0 or 1/2. Particularly in systems containing N defects such as ZnO:N [78], The long-range magnetic coupling of N-doped ZnO can be attributed to a p-d exchange-like p-p coupling interaction involving holes, which is derived from the similar symmetry and wave function between the impurity (p-like t2) and valence (p) states.

2.5 Current research efforts

The theoretical and experimental processes in spintronics have been achieved significantly over the last two decades. The research approaches in spintroics might be divided to two categories based on t ype of material: semiconductor and metal. Many device concepts have been proposed within semiconductor-based spintronics, yet fundamental problems remain in the development of room temperature ferromagnetic semiconductors and reliable transfer of spin-polarized electrons to and from other ferromagnetic materials. In contrast, room temperature metal-based spintronics is now well established which has proved to be of enormous economic significance with its application to disk drive sensor technology. In all application of spintronics, two common requirements are, first, a system that can generate a current of spin-polarized electrons comprising of one spin specie more than the other, and, second, a system that sensitive to the spin polarization of electrons. The researches in semiconductor-based spintronics are, therefore, focused on developing materials to sever these two requirements and also applicable for room temperature application.

The spin source and spin injector are critical requirements for the development of spintronics. To generate a spin-polarized current in semiconductor has become critical problem in semiconductor-based spintronics devices. In metal-based it can be supplely generated by passing current through a ferromagnetic metal and the spin-polarized electron injection from a ferromagnetic metal to a metal was successful since 1985. But, it is not yet possible for ferromagnetic semiconductor. It required materials in which the spins of the conduction electrons are highly polarized, preferably 100% with high injection especially at room temperature. As a consequence, most of current are focused on diluted magnetic semiconductors (DMSs). Among of them, ZnO based DMSs was known as candidate for room temperature ferromagnetic semiconductor. It is however the properties of present ZnO based DMSs have not been satisfactory to the major criteria for the applications of spin based devices. A part of current research effort has focused on enhancing the properties of existed material and seeking new candidates for the virtual applications. In this dissertation, Zn(Co,Eu)O semiconductor was exploited for potential room temperature ferromagnetic DMSs with strong localized spin polarization.

2.6 Properties of ZnO

Like most of the group II-VI binary compound semiconductor, ZnO crystallize in either cubic zinc-blende or hexagonal wurzite structure where each ion is surround by f our counter-ions at the corner of a tetrahedron [15]. At ambient conditions the thermodynamically stable phase is phase is wurzite while zinc-blende structure can be stable only when growth on cubic substrate. However, ZnO also can exist in rocksalt (NaCl) structure under high pressure. The crystal structures shared by Z nO are diagrammatically illustrated in Figure 2.11.

4 The wurzite structure has a hexagonal unit cell belongs to space group C6v or P63mc with ideal arrangement of ca= 83 and u = 38. For stable ZnO wurzite structure, the lattice parameter deviates from ideal arrangement. The lattice constants mostly range from 3.2475 to 3.2501 Å for a, from 5.2042 to 5.2075 Å for c, and from 0.3817 to 0.3856 Å for u. Most important, Zn- and O- plane stack alternatively along c-axis which gives the crystal a defined polarity, i.e. the basal planes are polar; whereas the side panels of the hexagonal unit cell contains both Zn and O atoms hence they are electrically neutral.

Electrically, ZnO occurs naturally as an n-type semiconductor due to the presence of defects [15, 79]. It is a direct bandgap material (Eg=3.5 eV) and can be increases to ~4.0 eV by substituting Zn sites with Mg. It also has a very large exciton binding energy of ~ 60 MeV. (Compare with GaN, which is an indirect band-gap material with lower exciton binding energy of ~25 MeV). This makes ZnO ideal for small wavelength optical electronic applications; especially where it i s readily available in bulk single crystal form; and improved epitaxial growth techniques can grow thin transparent film of ZnO. Hence ZnO has attracted vast amount of interests for its application to UV light emitter, varisistor, transparent high power electronics, surface acoustic devices, piezoelectric transducers and gas sensors apart from its potential applications in DMS.

2.7 Defects in ZnO

There are four major types of point defects that could occur in ZnO which deviate its composition from stoichiometry, which are zinc vacancies (VZn), zinc interstitials (Zni), oxygen vacancies (VO) and presence of hydrogen in the lattice [15]. VZn often occurs in

O–rich conditions, and acting as an acceptor; on the other hand, VO occurs in Zn–rich conditions, and acting as an n–type donor as well as Zni. Hydrogen in ZnO, unlike in other semiconductors, always occur in positive state and acting as donors in ZnO. Therefore the n-type nature of ZnO arises due to majority of defects in ZnO are n-type but which defect provide more dominant contribution to the n–type conductivity in ZnO is still controversial. The presence of various n-type defects makes n–type doping relatively easy compared to p–type doping, because p-type dopants would be compensated by n–type defects. In addition, p-type doping in ZnO is far more difficult because of very low solubility and forming deep acceptor levels in ZnO. The defects in ZnO can be determined with low-temperature photoluminescence (PL). The strong evidence was presented in favour of the oxygen vacancy (VO) as the defect responsible for the well-known green luminescence band at about 2.5 eV [80].

2.8 ZnO based DMSs

ZnO has recently come into research focus as a D MS based materials, especially after theoretical predictions on room temperature ferromagnetism were published by and Dietl [18], and Sato and Yoshida [81]. In among of the possible transition metal, Mn was common magnet dopants for ZnO based DMSs [82]. This is due to the fact that Mn2+ has a relative large magnetic moment (spin S= 5/2 and angular momentum L=0) with the characteristic of a half-filled d shell and can be incorporated in sizable amounts into ZnO host without affecting crystallographic quality of the DMS [83].

At early experimental state of ZnO based DMSs, the magnetic properties of Zn1-xNixO thin film were reported [84]. The Zn1-xNixO, with x = 3 – 25, prepared by pulse laser deposition (PLD) show ferromagnetic behaviour at 2 K but super-paramagnetic behaviour was observed above 30 K. Various PLD growth Zn1-xMnxO with Mn concentration well exceeding the equilibrium concentration shows no s ignal of ferromagnetism on elevated temperatures. Kim et al. [85] prepared Zn1-xMnxO film by sol–gel method and concluded that the observed ferromagnetic behaviour was due to the presence of magnetic Mn3O4 precipitates at the film/substrate interface. Sharma et al. [86] suggested ferromagnetism in Mn–doped ZnO was highly related to their processing conditions.

The room temperature ferromagnetic cobalt doped ZnO is first reported in 2001. Ueda et al. [87] observes ferromagnetism with TC above 280K in Zn1-xCoxO film prepared by

PLD. It is however the origin of the appearance of ferromagnetism in Zn1-xCoxO has not yet been clarified up to now. Jin et al. [88] reported no indication of ferromagnetism in

Zn1-xTMxO (where TM = Cr to Cu) growth by M BE while Kim et al. [89] suggests interaction between Co atom in Zn1-xCoxO is anti-ferromagnetic coupling. The origin of ferromagnetism in cobalt doped ZnO has still discussed controversially. The is partly cause by the problem that the experimental groups have published incongruous results ranging from room temperature ferromagnetism [89-97] to no ferromagnetism even at low temperatures [88, 98, 99]. A similar situation is applicable to ZnO doped with other transition metals such as Mn, Ni, Cr, Fe, etc. [6, 16, 82, 84, 86, 87, 100-105]

Many of groups have been trying to optimize ferromagnetic properties of ZnO based DMSs. A number of groups have expended a deal of efforts to realize p–type ZnO using nitrogen (N) as a possible shallow dopant [106-110]. A donor–acceptor co-doping method has also been investigated in order to increase the solubility of nitrogen in ZnO, which involves co–doping the nitrogen sources with metals such as Ga [111, 112], In [113] and Al [114]. However, only a few papers had demonstrated successful fabrication on p-type ZnO films, and the reproducibility of these methods are yet to be improved. Due to carrier induced ferromagnetic model, the ferromagnetic moment depends on carrier concentration. Norton et al. [115] had achieved TC > 250 K in Mn implanted ZnO:Sn due to that fact that Sn severs as a doubly ionized donor introducing deep states in the energy gap.

Rare earth elements are also of interest as dopants in ZnO because of their optical properties. Reports of Curie temperatures (TC) above 400 K and extremely large effective magnetic moments much higher than the atomic moment in Gd-doped GaN [116]. Defect emission in RE doped ZnO can be tailored to some extent, depending on the dopant. Reports of green luminescence from Tb-doped ZnO [117, 118] and red luminescence from Eu-doped ZnO [119-122] suggest that these materials may prove useful in optoelectronic applications. [123, 124]. The recent experiment shows Gd- implanted ZnO has ferromagnetic properties at room temperature [125, 126] with a saturated magnetization of 1.8 µB/Gd [127] which suggest that these materials may also be of use in spintronics applications.

2.9 Fundamental of X-ray spectroscopy

In general, X-ray strikers an atom and excited its core electron which can either be promoted to unoccupied state or ejected from an atom as shown in Figure 2.12. The X- ray or photon energy is absorbed by the core electron through the photo-electric effect. Any energy in excess of core electron’s binding energy is transferred to photoelectron. The photoelectron is either excited to higher energy unoccupied state or into an unbounded state called the continuum. The transition of electron from one quantum state to another within atom does follow quantum mechanical selection rules therefore photoelectron can be only promoted to certain unoccupied states.

Following an absorption process, an atom is now in excited atomic state with a core electron level left empty (called core hole), and a p hotoelectron. The excited atomic state will eventually decay typically in scale of 10-15 second later; however it does not affect the x-ray absorption process. During this decay, there are 2 main mechanisms. The first of these is x-ray fluorescence (Figure 2.13), in which a higher energy electron core-level electron fills the deeper core hole, ejecting an x-ray of well-defined energy. The fluorescence energies emitted in this way are characteristic of the atom, and can be used to identify the atoms in a system, and to quantify their concentrations. The second process for de-excitation of the core hole is the Auger effect, in which an electron drops from a higher electron level and a second electron is emitted into the continuum (and possibly even out of the sample). Either of these processes can be used to measure the absorption coefficient μ, though the use of fluorescence is somewhat more common.

Continuum Continuum

Photoelectron

Photoelectron N N

M M

L L

X-ray X-ray

K K Core hole Core hole

Continuum Continuum

N N

M M

L L

Kα Kβ

K K

2.10 X-ray Absorption Fine Structure (XAFS)

2.10.1 Principle of XAFS

X-ray absorption spectroscopy (XAS) is known as non-destructive technique broadly used to investigate local structure as well as electronic state of materials. Since each element on t he periodic table has a set of unique absorption edges corresponding to different binding energies of its electrons, the technique is therefore elemental selectivity. The XAS spectra displays energy absorption coefficient μ(E) by photoelectron around x-ray absorption edge of a selected element versus incident x-ray energy. The overall shape consists of a steep rise at electron binding energy and attenuates gradually with the energy as shown in Figure 2.14. The XAS spectra can be separated into 2 major regions, (1) X-ray absorption near edge structure (XANES) and (2) Extend X-ray absorption fine structure. The XANES is more sensitive to the electronic structure and the symmetry, while the EXAFS gives more information on bond distances, coordination numbers and local disorder. The combination of XANES and EXAFS is also known as X-ray absorption fine-structure (XAFS)

Figure 2.14 XAS K-edge shows 2 major regions of XAS data

XANES spectra cover pre-edge, rising-edge and near-edge regions. The absorption spectra of transition metal near K-edge usually shows pre-edge featuring of which 13sd→ dipole forbidden transition gains intensity a quadrupole mechanism and/or through 4p mixing into the final state. At rising-edge, photoelectron is excited to valence unoccupied state above Fermi level (EF) through electric dipole transition: e.g. 14sp→ at K-edge for 3d transition metal. The pre-edge and rising-edge contain information about the oxidation state of absorption atom. Near-edge is where photoelectron leaves an atom with low kinetic energy. The range is varying depending to absorption edge and atom, approximately 50-150eV above rising edge. Finally, the spectra beyond near-edge region are belonging to EXAFS energy range.

2.10.2 Origin of the EXAFS oscillations

At the rising edge, the incidental photon energy is defined to be equal to E0, threshold energy or core electron blinding energy. For any energy above this the kinetic energy of photoelectron escaping from atom is given by

Ek = hv − E0 (1.1)

The ejected photoelectron with kinetic energy Ek now propagates as wave and is scattered by nearby atoms as demonstrated in Figure 2.15. The expression for the wave vector k of the photoelectron can thus be written as

2 k=2 m() hv − E0 h (1.2)

At near-edge, photoelectron escapes into the continuum with low Ek But in EXAFS, ejected photoelectron has high Ek The oscillatory of µ(E) in EXAFS region, as well as near-edge region, is caused by interference between outgoing photoelectron wave and backscattered photoelectron wave from neighbouring atoms. There is however a difference in the physical phenomenon taking place between low and high Ek. In low Ek, the corresponding wavelength λ of outgoing photoelectron wave is large enough to be comparable to inter-atomic distance between absorber and its neighbours, and

corresponding amplitude of scattering wave is very large so that multiple scattering process dominant in near-edge region. In contrast, outgoing photoelectron wave of high

Ek has corresponding wavelength lower than inter-atomic distance and weak scattering amplitude so mainly single scattering of the photoelectron by the neighbouring atoms dominate.

2.10.3 The EXAFS equation

In EXAFS region, the absoption efficicient as function of photon energy is defined as

µtotal = µχ0 []1 + EX (1.3)

with µ0 representing atomic absorption background. χEX is called the EXAFS function, forming the oscillatory part in EXAFS region and carrying information about local structure around the absorber atom. Several authors have given derivation for the EXAFS theory [129-132]. The simplest theory is based on the single scattering plane- wave approximation. The χ function can be given as a summation over all interference patterns scattered of all neighbouring atoms:

2 NSF() k −−2/σλ2 Rk() χφ= jj0 jj − (1.4) ()k ∑ 2 e esin 2 kRjj() k j=1 kR j

2 where S0 is an amplitude reduction factor representing many-body effects such as central atom shake-up and shake-off due to relaxation processes after the photo ionisation event [131, 133]. The backscattering amplitude Fj(k) and phase-shift Φ(k) are depending on atomic number. λ(k) is mean-free-path, N is coordination number of neighbouring atom, σ2 is mean-square disorder of neighbour distance, and R is distance to neighbouring atom.

2.10.4 XAFS data modelling

From equation (2.4), if the scattering properties of neighbouring atom: F(k) and Φ(k), and λ(k) are known, it is possible to determine R, N and σ2. These properties can be theoretically calculated using available program code available these days such as FEFF [134] code. The FEFF is automated ab initio for multiple scattering calculations of X- ray absorption spectra. The algorithms used in FEFF were developed by R ehr and Albers [135], and Zabinsky et al. [136] in which FEFF was named after the effective scattering amplitude feff. The list of atomic (x,y,z) coordinated for a physical structure and selected central atom are required for FEFF code.

Many analysis programs use FEFF for model EXAFS data. One of them is IFEFFIT [137, 138] which is suite of interactive program for XAFS and is used in this thesis. The experimental date must be isolated into χ(k). As defined in equation (2.3), χ can be obtained once µ0 is known by calculating:

µµ− χ = total 0 (1.5) µ0

χ can be convert from energy space to wave vector k space easily using equation (2.2). EXAFS model and experimental data can be easily compared in Fourier transformation of χ(k). Sayers et al. [130] have shown Fourier transformation of χ(k) results in a radial distribution function The Fourier transformation is defined by

k 1 max n i2 kR χχ()R = ∫ k() k e dk (1.6) 2π kmin

The radial distribution function is defined in R-space (e.g., distance from the absorber atom). The distance found in the Fourier transformation is about 0.2–0.5A° shorter than the actual distance due to the energy dependence of the phase factors in the sine function (see equation(2.4)). To fit EXAFS model into experimental data, the EXAFS equation for each scattering path must be evaluated as computed by FEFF. The value of

2 2 the parametric terms in the EXAFS equation such as S0 , R, N, σ and energy E0-shift for each scattering paths of model. Using IFEFFIT, the fitting refinement is based in a minimization of χ2 function in R-space:

N N pts 22 χ2 =idp Re χχ ()RR−+ () Im χχ () RR − ()  (1.7) 2 ∑{( data iith )( data iith )} N ptsε i=1

2 The detail has been fully discussed in Ref [139]. The minimization of χ is performed by a Levenberg-Marquardt non-linear least square minimization. This method is explained well in Bevington's book [140], in Numerical Recipes [141], and in many other sources.

2.11 X-ray magnetic circular dichroism (XMCD)

2.11.1 Principle of XMCD

In optics, the term "dichroism" refers to changes in the absorption of polarized light on passing through a material in 2 di fferent directions. Using right hand and left hand

circular polarized X-ray in XAS, XMCD is defined as t he difference of absorption spectra when the helicity of the X-ray is parallel and anti-parallel to the magnetization of materials as sh own in Figure 2.16. The technique is sensitive to magnetic polarization, and therefore is suitable for study magnetic properties of particular orbital on each element properties of matter such as metal, thin films and multilayer nanostructures.

Magnetic properties of transition metal are mostly due to imbalance of spin up and spin down elections in their d-orbital electron. Therefore excitation of p-core electrons is a key for investigation. The origin of XMCD at L2,3 of 3d transition metal can be explained by the two-step model proposed by Stöhr and Wu [142, 143]. In the first step, circular polarized X-ray interacts to spin-orbital split level in 2p core shell, which acts as spin polarized excited photoelectron source. The spin polarization depends on orbital splitting level (or edge) and photon polarization (or helocity). In a nonmagnetic material the total (spin-up plus spin-down) transition intensities are the same for LCP and RCP light; as soon as there is an unbalance in the number of available empty spin up and down states (ferro-, para- or ferrimagnetic material) the absorption of the two polarizations will be different, with a d ifference which is opposite at the L2 and L3 edges. In the second step, the exchange split valence shell with unequal spin-up and spin-down acts as the detector for the spin polarized of the excited photoelectron because spin flips are forbidden in electric dipole transitions. In order to optimize detection and XMCD intensity, the valence shell spin quantization axis (the detector axis) has to be aligned with photon spin or photoelectron spin quantization axis by external magnetic field.

a d I - I + background

b I = [ I - + I + ] / 2 e I - B µ- I + L3 L2 µ+ µ- c OR B A  - k µ+ I

+ µ XMCD = I - - I + I -

Figure 2.16 Schematic diagram of X-ray magnetic circular dichroism

(XMCD): (a) L2,3 edge XAS spectra using left and right circularly polarization incident X-ray with magnetic field B parallel to - + incident X-ray direction. I and I representing µ(E) when helicity of X-ray is anti-parallel and parallel to magnetic field

respectively; (b) Average µ(E) of L 2,3 edge XAS spectra; (c) XMCD spectra; (d) Transition of 23pd→ absorption with right and left circularly polarization X-ray; and (e) Experimental set up for XMCD measurements.

2.11.2 XMCD sum rules

The XMCD measurement allows spin and orbital magnetic moment of absorbing atom to be measured. Three sum rules related the measurement intensity I , I , A and B, L3 L2 defined in figure xx, to the electronic and magnetic properties of the materials The first sum rule is related to the charge distribution and is given by [144-146]

I+=+ I CN() Nα (1.8) LL32α h Q where C is the square of pd→ radial transition matrix element and have value about 10 nB eV [147]. The index α specific the orientation of polarization E (linear polarization) or k (circular polarization) related to magnetization direction. The sum rule correlates the polarization dependent white line intensity with the total number of d

α holes Nh and a quadrupole term NQ which expresses the anisotropy of the charge

α density in atomic volume [144]. The sum rule expression NNh +=Q N eff can be written as a linear combination of N  and N ⊥ [143]. The spin sum rule originally derived by Carra et al. [148] is given by [144-146]

C α [A−=−+ 2] Bα () mmsD (1.9) µB

where ms is isotropic spin moment and µB is Bohr magneton. The term

α mTDB= 7/α µ h [143] is the intra-atomic magnetic dipole operator [148]. Finally, the orbital sum rule by Thole et al. [149] can be simplified to [144-146]

3C α −+[]ABα = mo (1.10) 2µB

α α α where the orbital moment mo along the α direction. The origin of NQ and mD are discussed more detail in Ref [145, 146]. These terms vanishes when an angular average

1 α =xyz ++ = 1 α = is performed, 3 ∑α NQ( NNN QQQ) /3 0 and 3 ∑α mD 0 . In this case, the isotropic sum rule is obtained.

2.12 Experimental efforts on XAFS and XMCD characterizations

XAFS is a powerful technique to investigate the local atomic geometry and the chemical state of the atoms of one specific element in almost any type of substance. In contrast, diffraction-based techniques required crystallized for studying atomic structure of matter. XMCD provides magnetic characterization for each individual element while conventional techniques such as v ibrating sample magnetometer (VSM) and superconducting quantum interference device (SQUID) magnetometer can only obtain the total magnetic moment of all atoms in matter.

In ZnO-based DMSs studies, XAFS have been used to characterize on the fate of dopants that is whether substituting into lattice matrix, clustering or second phase. The measurements on Zn1-xCoxO prepared by va rious fabricating technique commonly reveal that Co2+ ions have substituted for Zn2+ ions and incorporated into the ZnO lattice for high concentration of cobalt. Whereas, substitutional Co2+ is coexisted with another Co phases structure, such as metallic, and oxide compound for high concentration of cobalt [150-154]. There are two types of Co metallic found in Zn1-xCoxO, one phase is similar to Co metal [hexagonal close packed] and another phase is similar to CoEu inter-metallic. Kaspar et al. observed CoEu inter-metallic in Zn1-xCoxO deposited by

PLD after annealing in Zn Vapour [155, 156]. Using sol-gel preparing Zn1-xCoxO,

Co3O4 precipitated was found in its nano-composite [154] while Co clustering was found in its thin film [152]. Zn(Co,Ga)O and Zn(Co,Al)O prepared by molecular beam epitaxy shows absences of metallic clustering and all dopants are substituted Zn2+ in tetrahedral coordination. [157-159]. Similar results also found in (Zn,Mn)O[160], (Zn,V)O[161] and Zn(Co,Cu)O[162]. The defect can be investigated by XAFS as well.

Yan et al suggest room temperature intrinsic ferromagnetic in as deposited Zn0.98Co0.02O thin film is due to large amount of oxygen vacancy (VO). After annealing at high temperature, VO has been disappeared and the film has become paramagnetic [150]. Most of fabrication technique, Ney et al

Many studies show Co2+ substituted Zn2+ in ZnO matrix is paramagnetic and ferromagnetic behaviour in Zn1-xCoxO usually cause by i mpurity phase and structure

defect. It has been demonstrated that as l ong as phase separation or excessive defect format is absence, Zn1-xCoxO is paramagnetic [163]. The origin of ferromagnetic in ZnO based DMSs is still controversy. XMCD measurement can provide localized magnetic moment of each individual element. Xu et al.[164] shows that XMCD measurement at 30K confirm paramagnetic properties of the doped substitutional Co2+ in Co doped ZnO and Co, Al codoped ZnO. The similar results were also reported by Barla et al. [165] and Tietze et al.[166]. Ney et al. found Co dopants exclusively occupy Zn sites by XMCD and paramagnetic. They also revise that Co-O-Co is anti-ferromagnetic coupling and no s ign of intrinsic ferromagnetic interaction for isolated of paired Co dopant atoms in Co doped ZnO [167].

On the other hand, several studies shows intrinsic room temperature ferromagnetic is observed. For example, Singh et al.[168] prepared ZnO/Al2O3 by plasma-assisted molecular beam epitaxy (PAMBE) and consequence Co implanting using a source of negative ion by c aesium sputtering (SNICS). XANSE show absence of Co clustering but only Co2+ substituted Zn in tetrahedral symmetry with 10Dq = - 0.6eV. The XMCD results also confirm the substitution of Co at the Zn site in the ZnO matrix, which is responsible for room temperature ferromagnetism. Using control defect suggest V0 is not contributing in ferromagnetism but the contribution from the hybridization of empty 3d state with O 2p state cannot be rules out.

2.13 Summary

Spintronics is a promising candidate for next generation electronic devices where the spin of the carriers would play a crucial role in addition to or in place of the charge. However, the fundamental problems remain in the development of room temperature ferromagnetic semiconductor and reliable spin-polarized electrons transport properties. In contrast, the room temperature metal based spintronics have already been employed in current functional devices e.g. giant magneto-resistance (GMR) read head in high capacity hard disk drives (HDD) and magnetic tunnel junction (MTJ) in magnetic random access memory (MRAM). Great efforts have been put onto material synthesis and investigation of the mechanism behind ferromagnetism of diluted magnetic semiconductors, thus finding a way to raise the Curie temperature, TC, above room temperature. Although ZnO based DMSs have already played a c ritical role as a candidate for room temperature semiconductor based spintronics applications, there are still number of fundamental problems unclear. Up to date, the major efforts in ZnO based DMSs have been focusing on the transition metal doping ZnO to realize the p- type materials while the minor effort has put in rare earth doping ZnO with respecting to their magnetic properties. This dissertation studied the fabrications and properties of europium doped ZnO and cobalt doped ZnO with additional Eu doping by using the cutting-edge techniques.

Chapter 3 Experimental and Characterization Techniques

3.1 Ion implantation

Ion implantation is a materials engineering process by which ions of materials are accelerated in an electrical field and deposited into another solid, thereby changing chemical properties and physical properties of the implanted materials. This technology has been using in semiconductor and metal processing industries as well as various application in materials science research. The doping is the most common application of ion implantation. It is well known as method of producing p-n junction in solid state electronic devices, by introducing dopant ions such as boron, phosphorus or arsenic into semiconductors. It can create a change carrier in the semiconductor; hole (p-type doping) and electron for (n-type doping). SIMOX (Separation IMplant of OXygen) and Mesotaxy are also developed using ion implantation principle.

Currently, there are a number of techniques available for fabrication of the doped thin film such as ion implantation, pulsed laser deposition (PLD), molecular beam epitaxy (MBE), chemical vapour deposition (CVD) and atomic layer deposition (ALD). In this thesis, ion implantation was chosen to the fabricate transition and rare earth doped ZnO thin films as it offers several advantages over the other techniques:

• Ion implantation offers accurate and reliable measure of the total dopant ions penetrated to the required target thin film. Therefore, the material with the desired composition can be easily obtained and reproduced. • The ability to select doping area: by controlling ion beam size or masking, it is possible to partial dope a target or to have various doping species on one material. • The depth of implantation is determined by the energy of the ion beam. Thus, by controlling the energy of the incident ion beam, surface layers with precisely specified compositions can be produced. It is also possible to produce compositions that cannot be fabricated by conventional methods.

The following discussion briefly describes the implantation process and MEVVA ion implanter in Australian Nuclear Science and Technology Organisation (ANSTO).

HV supply Isolation transformer

HV Ion beam generator Terminal

DI Mass analysis Water magnet supply Control Acceleration console lens

Quadrupole magnet

Beam manupilation Magnetic scanning system

Process chamber

Figure 3.1 Schematic diagram of an ion implantation system which included an ion beam generator, beam manipulation and a process chamber

A basic ion implantation system comprises an ion source and a sample process chamber. In the source, plasma of the desired implant species is produced from which ions are extracted electro-statically to maximum energies of typically a few hundred keV. The resulting ion beam may then be passed through a magnet to select the ions with a particular mass to charge (m/e) ratio. The beam that emerges from the magnet can then be further accelerated before implanting on the target. An ion implanter incorporating

the latter feature is shown in Figure 3.1, presenting the schematic diagram of a typical system. The time taken to complete an implantation run depends on (1) the implantation concentration profiles required, (2) the target surface area to be implanted and (3) the ion beam current. However, several implantation parameters need to be optimized in order to obtain the desired composition and correct depth profile. The implantation dose calibration and theoretical calculation are required for this proposed. Detailed specifics will be discussed in Chapter 4.

3.2 Rutherford backscattering spectrometry (RBS)

RBS is an accelerator bound technique for quantitative composition analysis of thin layers or near surface regions of solids. The principle is based upon t he elastic two- particle scattering of energetic ions with sample atoms via the repulsive Coulomb force of the positively charged atomic nuclei. The energy distribution of backscattered ions presents the information on the profile of concentration versus depth range of typical few microns below the surface.

Target

Incident Ions

Scattering Angle Φ Detector

Scattered Particles

Figure 3.2 Schematic representation of the experimental setup for Rutherford backscattering analysis

Light ions such as H+, D+, He+ with energies in low MeV region are usually used for ion source. The mono-energetic ions beam penetrates to considerable distance below the surface of a material and loses their kinetic energy, mainly in collisions with electrons,

until they rest inside of matter. A small fraction of ion beam approach atomic nucleus closed enough can be scattered at large angles. Backscattered ions (< 0.1 MeV) can leave the sample and reach a particle spectrometer, where their energy is analysed. The energy of the backscattered particles depends both on the mass of the scattering nucleus and the depth of the collision in material. The composition of materials at a particular depth is determined from the measured energy loss of backscattered ions and the scattering cross sections of the atoms present. Independent chemical analyses may be required to identify the types of atoms present in a material if this is not already known.

For experimental setup in ANSTO, He+ ions were produced by the 3 MV Van de Graaff generator. Samples are mounted individually along a metal carrier which is then loaded into the sample chamber located at the end of beam line. The ions are directed by mean of large magnets, along each beam line which is evacuated to about 10-6 of Hg. The spectrometer consists of a si licon particle detector followed by a ch arge sensitive preamplifier, a shaping main amplifier and the multichannel analyser. It is noted that all analysed samples has been examined with 1.8 MeV He+ and detector placed at 170° to the incident beam. Samples are set normal to the incident beam as shown in Figure 3.2.

3.3 Heavy ion elastic recoil detection analysis (ERDA)

RBS is the most commonly used elemental analysis technique. However it requires to de-convolute the backscattering spectra originating from different target elements. This makes the detection of light element such as oxygen of carbon difficult. Heavy Ion Elastic Recoil Detection Analysis (ERDA), which is complementary to RBS, provide a way to profiles these light elements. In contrast with RBS, ERDA uses a beam of heavy particle to bombast the sample. The beam particle is usually heavier than the target atoms of interest so that the lighter target atoms can be knocked forward in the collision and recoil out of the target. The detection of these particles, after rejection of scattered beam, provides a method for the analysis and depth profiling of light elements. Using very heavy projectiles ERDA, is almost a universal technique, suitable to measure the whole range of element from H up to As and Ge or heavier. For example, it has been used to depth profile hydrogen in conducting oxide and synthetic diamond films. It has also been used to profile elements as heavy as zirconium.

Energy Dispersive Detectors

Recoil Ion Foil Range MeV Incident Ion Scattering Angle Sample

Figure 3.3 Schematic representation of the experimental setup for Elastic recoil detection analysis

Typical layout of ERDA is shown in Figure 3.3. When a sample is mounted at an angle with respect to an ion beam, the beam causes a minority of target atoms to be recoiled from the surface. The energy with which they recoil depends their mass and on the recoil angle. At known geometries, the energy dispersive detectors can thus identify and quantify the recoiled atoms. The depth profiles can be obtained by de-convoluting the energy lost as the ions travel into and exit from the sample surface.

3.4 Thin film X-ray diffraction

X-ray diffraction (XRD) is one of the most powerful non-destructive techniques to investigate the crystal structure of the thin films. Bragg’s Law for constructive interference states that

ndλθ= 2 sin (1.11)

Where λ is the x-ray wavelength, n is an integer, d is the crystal plane spacing, and θ is the x-ray incident angle. Based on this equation, the crystallographic information can be determined.

Figure 3.4 Philips X'pert MRD Four circle, x-y-z translation, point or line focus

XRD experiment has been performed using PANalytical X’Pert Pro MRD four circle high resolutions X-ray diffraction system showing in Figure 3.4. The main feature of this instrument is the vertical triple axis sample state holder, providing reliable and precise measurement for thin film. The incident X-ray Cu Kα emission is sourced by X- ray tube with copper anode. The setting configuration used in this thesis is medium resolution module with the X-ray mirror + collimator, and high resolution module with the X-ray mirror + Ge (220) monochromators for thin film analysis and phase identification.

3.5 Magnetic measurement with MPMS

The sample magnetometer is standard tool for study magnetic properties of materials such as magnetic superconductor, magnetic semiconductor and magnetic multilayer.

The magnetic properties of Zn1-xCox-yEuyO thin film are studied using the Quantum Design Magnetic Property Measurement System (MPMS-XL) as shown in Figure 3.5. Utilizing Superconducting Quantum Interference Device (SQUID) technology, it achieves superior measurement sensitivity, dynamic range, and reproducibility. The instrument is allowed magnetic field dependant measurement between -5 and +5 Tesla and temperature dependant measurement from 4.2K to 350K

Figure 3.5 Quantum Design MPMS-XL system components

SQUID is the most sensitive device available for measuring magnetic fields, although it does not directly detect the magnetic field from the sample. Instead, the sample moves through a system of superconducting detection coils which is connected to the SQUID by superconducting wires, allowing the current from the detection coils to inductively couple to the SQUID sensor as shown in Figure 3.6. The output voltage produced by

SQUID electronic is strictly proportional to input current flowing in the SQUID input coil. The thin film SQUID device sensor, illustrated in inset of Figure 3.6, is shielded inside the superconducting shield both from fluctuations in the ambient magnetic field of the laboratory and from the large magnetic fields produced by the superconducting magnet because of its extreme sensitivity to magnetic fluctuations. The superconducting detection coil is a single piece of superconducting wire wound in a set of three coils configured as a second-order (second-derivative) gradiometer shown in inset of Figure 3.7. The coils are positioned at the centre of the superconducting magnet outside the sample chamber such that the magnetic field from the sample couples inductively to the coils as the sample is moved through them. The gradiometer configuration is used to reduce noise in the detection circuit caused by fluctuations in the large magnetic field of the superconducting magnet and also minimized background drifts in the SQUID detection system caused by relaxation in the magnetic field of superconducting magnet.

Figure 3.6 Basic scheme of MPMS Quantum Design: when a magnetic sample is displaced through a superconducting coils, a current I is induced which flows without losses to a transformer; the second circuit generate a flux in at third coil which will be detected by SQUID detector. Inset of this figure shows basic scheme of a SQUID thin film sensor, the two Josephson junctions, e.g.1 nm thick insulating alumina barriers.

A measurement is performed in the MPSM by m oving sample though the superconducting detection coils locating outside the sample chamber and at the centre of the superconducting magnet as shown in Figure 3.7. As the sample moves through the coils, the magnetic moment of the sample induces an electric current in the detection coils. Because the detection coils, the connecting wires, and the SQUID input coil form a closed superconducting loop, any change of magnetic flux in the detection coils produces a change in the persistent current in the detection circuit, which is proportional to the change in magnetic flux. Since the SQUID functions as a highly linear current-to- voltage convertor, the variations in the current in the detection coils produce corresponding variations in the SQUID output voltage which are proportional to the magnetic moment of the sample.

3.6 The band-gap energy (Eg) measurement

For transparent wide band-gap semiconductors thin film such as ZnO, the band-gap energy (Eg) can be estimated by using optical transmittance measurement in ultraviolet - visible regions. In this thesis, transmittance measurement is performed by PerkinElmer LAMBDA 950 UV/Vis/NIR Spectrophotometer as shown in Figure 3.8. The absorption coefficient α can be calculated by

T= Ae ⋅ αd (1.12)

The optical band-gap of the films were determined by applying the Tauc model [169] and Davis and Mott [170] in the high absorbance region:

n αhv= B() hv − Eg (1.13)

,where

• T : the transmittance of thin film • d :the film thickness • hv :the photon energy

• Eg: the optical band-gap energy • A, B and n: constants

UV-VIS-NIR photo-spectroscopy refers to transmission or reflection spectroscopy in the ultraviolet-visible-near infrared spectral region. Routinely, it provides quantitative determination of solutions of transition metal ions highly conjugated organic compounds, and biological macromolecules. Another advantage of the UV-VIS-NIR photo-spectroscopy are film thickness determination of transparent films and layers, estimation of the electronic band-gap of semiconductor films, calculation of the optical properties of thin films, and measurement of the reflectance loss in photovoltaic cells.

Figure 3.8 Schematic of optical system in PerkinElmer LAMBDA 950

3.7 X-ray absorption spectroscopy (XAS)

XAS is a widely-used based technique for determining properties of matter at atomic scale. XAS is also element-specific and does not require crystalline samples unlike diffraction based techniques, which are its major advantages. An energy tuneable x-ray source, usually refer to synchrotron radiation light source, is required for the experiment. A synchrotron is a special type of charged particle accelerator circulating at speed of light in the ring-shaped chamber. This is done by creating strong magnetic field (to steers the particles) and electric fields (to accelerate the particles) in vacuum which minimizes collisions with air molecules and allows storage of the beam at high energy levels for many hours.

When the relativistic electrons in the storage ring are deflected by the bending magnets that keep them in a cl osed circular orbit, the electrons emit highly intense beams of linearly polarized x-rays in the plane of the electron orbit (bremsstrahlung). On the other hand, they emit circularly elliptically polarized light out of the plane. Currently, a number of alternative sources for circularly polarized synchrotron radiation are under development. The most notable ones are so-called insertion devices like helical wigglers [27] and crossed [28] undulator, which are complex arrays of magnets with which the electrons in a storage ring are made to oscillate in two directions that are perpendicular to their propagation direction, with the result that they emit circularly polarized light.

The energy extraction devices, e.g. wigglers and undulators, deflect particle beam inside a storage ring as result of synchrotron radiation emission. "Synchrotron radiation" refers to a continuous band of electromagnetic spectrum including infrared, visible light, ultraviolet, and X-rays. In this thesis, two XAS experiments: X-ray absorption fine- structure (XAFS) and (2) X-ray magnetic circular dichroism (XMCD) are performed at National Synchrotron Radiation Research Center (NSRRC), Hsinchu, Taiwan. The light source is equipped with 3.3 GeV and numbers of beamline station for different type of experiments. The details for both experiments will discuss in next session.

3.7.1 X-ray absorption measurement

XAS can be measured either in transmission geometries, fluorescence geometries or electron yield (EY) detection as shown in Figure 3.9. In transmission detection, the transmitted intensity significantly drops when incident x-ray energy channel opened at absorption edge, corresponding to loss of photons through core electron excitation to empty states. On the right side of for Figure 3.9, it shows typical measured transmission intensity (It), normalized to incident number of photons (I0). The transmission mode measurement is standard for hard x-rays, while for soft x-ray, they are difficult to perform because of the strong interaction of soft x-ray with the sample. For other two method, they were discussed by Stöhr [171]. Electron yield detection, also called total electron yield (TEY) detection, is particular often in the soft X-ray while fluorescence geometries, also called total fluorescence yield (TFY) detection is essential in the hard X-ray region. After X-ray excitation, the Auger emission is dominant over fluorescence decay in the soft X-ray regime, but X-ray fluorescence is more likely to occur than Auger process in hard X-ray regime. The energy dependence of absorption coefficient µ(E) either in transmission detection as

I µ ()E = ln 0 (1.13) It or in x-ray fluorescence (or Auger emission) as

I f I µ ()E ∝  or e (1.13) II00

where It, If and Ie are respectively the monitored intensity of transmission, fluorescence and electron emission associated with the absorption process.

(a) Transmission d

−µ = xd I0 It Ie0 e- hv core

Sample

(b) Electron yield X-ray e-

- e λe

IItx∝ 0µ - e - I0 e A

core (c) Florescence X-ray X-ray floresence

λe I0 e- - Kβ e

−µ xd Kα If = Ie0 core

3.7.2 X-ray absorption fine-structure (XAFS)

The XAS spectra, energy absorption coefficient μ(E) by photoelectron around x-ray absorption edge of a selected element versus the energy of incident x-ray, can be divided to two regions: XANES and EXAFS regions. X-ray absorption near edge structure (XANES) is more sensitive to electronic structure and symmetry of absorbing atom while extended x-ray absorption fine structure (EXAFS) provides information on bond distance, coordination numbers and local disorder. The combination of these two is referred to as XAFS.

Experiment Hutch

Photomultipliers Scatter slits Filter

Source Beamline Optics θ [Mirrors +monochromators + slits]

Ionization Chamber

Signal Amplifiers

Computer

Figure 3.10 Schematic experimental layout of BL-17C at NSRRC

The experiment is performed at Wiggler beamline BL17C1 at NSRRC. The typical setup is schematically shown in Figure 3.10. The beamline optics at this end station provide circular polarized x-ray with energy range of 6.5 -14 keV to the experimental hutch. The energy of incident x-ray is controlled by m onochromators inside of the beamline optics. The incident x-ray intensity (I0), sample transmitted intensity (It) and reference transmitted intensity (Ir) measured by ionization chambers detectors while sample fluorescence intensity (If) measured by photomultipliers. The sample, mounted on rotating sample holder, is aligned with incident angle of 45° to normal axis of the films

3.7.3 X-ray magnetic circular dichroism (XMCD)

S 2 RFM

VPM2 S1 VPM1 Top view G ES2 VFM ES1 HFM

6°

S2 S1 VPM1 HFM VFM RFM 5°

4° Side view G VPM2 ES1 ES2

0 3.2 6.0 9.99 11.99 13.1 14.9 15.98 17.5 21.3 22.3 m

Figure 3.11 Schematic optical layout of BL-11A at NSRRC

The experiment is performed at dragon beamline BL11A1 at NSRRC. Elliptically polarized bending magnet (EPBM) provides synchrotron radiation light with horizontal acceptance of 12 mrad. The light was reflected by horizontally and vertically focusing spherical mirrors (HFM and VFM) to spherical grating monochromators (SGM). The monochromatic light, up t o 1500 e V, was reflected by t oroidal refocusing mirror (RFM), and introduced to the measurement chamber is located at the end station. Two vertical plane mirrors (VPM) between the gratings and the exit slit to extend the lowest photon energy to 10 e V. In practical measurements, the photon energy was scanned using a grating which have 1200 lines/mm, covering the photon energy range 400 − 1200 eV and photon flux is 1 × 1010 sec-1 with the energy resolution E/ΔE = 10,000.

Top view Sample Position

Electromagnet Circular Polarized X-ray 30°

Magnetic Field

End Station Chamber

Figure 3.12 Schematic structure shows the layout inside of the measurement chamber. Applied magnetic field axis has 30 degree angle to incident x-ray direction.

In the XMCD measurement, the circular polarization of the incident photons was fixed and the direction of the applied was changed. The spectra were recorded at temperature of 77K and room temperature in an ultra-high vacuum with pressure below 10-10 Pa. The spectra were measured both in the total-electron-yield (TEY) and the total- fluorescence-yield (TFY) mode. In TEY mode, samples were connected to the ground via thought digital ampere meter and neutralization current is monitored. But, TFY mode monitors emitted photons which are created when soft x-ray hit the sample. The detection depth of TEY mode is about 5nm, while TFY mode is about 100 nm. Therefore surface and bulk properties of samples can be obtained by using TEY and TFY mode respectively.

3.8 Annealing

Annealing is a heat treatment process in which the microstructure of a material is modified, resulting in changes of its properties. The ion implantation causes damage to crystal structure of the target. Each individual ion may produce point defects in target thin films on impact such as vacancies and interstitials. These points defect can migrate and cluster each other, resulting dislocation loops and other defects. Therefore, ion implantation processing is often followed by a thermal annealing in order to recovery crystal structure of the target thin films. In present work, the implanted sample are introduced to the furnace at room temperature and heated up to 600oC at the rate of 10oC/min, the sample is annealed at 600oC for half an hour and cooling down to room temperature. During annealing process, Argon gas is purged with rate of 20ml/min at pressure 1 atm (or 1.013x105 Pa).

Chapter 4 Depth Distribution Profiles of the Implanted Co and Eu ions in ZnO:Co and ZnO:Eu Thin Films

4.1 Introduction

As aforementioned in chapter 3, the ion implantation is powerful technique widely used in semiconductor industry. This technique has also been used to fabricate some of DMSs materials such as rare earth doped ZnO single crystal and ZnO nanorod. In addition, it provides doping concentration with accurate amount. The ion implantation results from the entrapment of incident ions in a solid accompanied by t ransfer of energy to the electrons and nuclei of the target. Collisions which involve elastic scattering produce collision cascades in which both sputtering that is ejection of target atoms from the surface and atomic mixing occur. An ion beam current integrator has been developed to obtain an accurate measure of the total implant ions delivered to target. Calibration of the integrator, for each cathode material of interest, was carried out by r eferencing number of integrator reading to corresponding implant dosed determined from RBS analysis. This provided a convenient means of reliably setting the required dose for subsequent implantation. Thus, the ion current integrator is a useful ancillary tool of MEVVA ion implantation.

The range of implanted ions can be estimated by va rious theoretical calculation methods. The implantation depth profile in often similar to a Gaussian distribution, the latter is frequently assumed in first order approximations. Other more appropriate distribution function have been used as simulation techniques based on Monte Carlo methods which permit the analysis of complex targets such as multi-layered or multi- element substrate. In this chapter, we combine theoretical and experimental studies to understand ion depth distribution profiles of Co implanted ZnO and Eu implanted ZnO.

4.2 Experimental details

In this chapter, the depth profile of implanted ions in ZnO:Co and ZnO:Eu are investigated theoretically and experimentally. The theoretical studies are implemented using TRIM calculation to estimate implantation range and ion depth distribution in the ZnO host. In our calculation, 1,000 Å of ZnO layer with density 95% of the density of ZnO single crystal, which is 5.642 g/cm3, is implanted by mono-energetic cobalt ions and europium ions. The incident projectile angle is parallel to normal axis of ZnO layer. The calculation is performed with various ions energies up t o 150 keV. In the experiment studies, ion implantation is performed using MEVVA ion source in ANSTO. Total ions delivered to target are measured by an ion beam current integrator. To calibrate the integrator for each cathode, commercial available Si (0001) wafer were implanted with various integrator reading for cobalt and europium under a vacuum pressure of 2x10-4 Pa (or 2x10-6 mbar). The ion accelerated voltages are 24 kV and 45 kV for cobalt and europium ions respectively with a beam current of 30 mA. The Rutherford backscattering spectroscopy (RBS) were used to determine the implanted dose, corresponding to count accumulation by current integrator. Finally, ZnO:Co and ZnO:Eu are prepared by cobalt and europium implanting into ZnO epitaxial thin films growth on c-Al2O3 (0001) substrates, using PLD with an approximate thickness of 100 nm, with integrator reading of 100,000. T he implantation depth profile is studies by RBS analysis.

4.3 Result and discussion

The ions depth distribution profile of ZnO:Co and ZnO:Eu were investigated. Ion implantation results from the entrapment of incident ions in matter accompanied by transfer of energy to electrons and nuclei of the target though elastic collision. The cobalt ions and europium ions depth distribution profile are shown in Figure 4.1. The profile distribution for implanted ions is closed to Gaussian, obviously for low energies, with the peak closer to surface of target. As energy of incident ions is rising, the range of implantation depth (x-axis) is larger and the peak of distribution is shifted deeper while distribution weight, f(x), is decreasing.

a) 0.12 10keV 20keV 0.10 30keV 40keV 50keV 0.08 60keV 70keV 80keV 0.06 90keV f(x) 100keV

110keV 0.04 120keV 130keV 140keV 0.02 150keV

0.00 0 20 40 60 80 100 Depth (nm) b) 0.20 10keV 20keV 30keV 40keV 0.15 50keV 60keV 70keV 80keV 0.10 90keV 100keV f(x)

110keV 120keV 130keV 0.05 140keV 150keV

0.00 0 20 40 60 80 100 Depth (nm)

The loss of energy from ions of medium energy can be described in terms of the nuclear stopping power, Sn, and the electronic stopping power, Se. The rate of energy loss from the incident ion beam as a function of distance underneath the surface of target material is given by:

dE −=NS[ () E + S () E] (1.14) dx ne where N is the number of target atoms per unit volume of the solid. The stopping powers are depended on s pecies of target and incident ions, and ions energies. According to the calculation, energy loss by electron stopping power and nuclear stopping power by europium ions is larger than cobalt within energy region of interest for ZnO target. It results stopping range of europium ions is shorter than cobalt ions. The mean of ion stopping depth of cobalt and europium as function of ion energy is summarized in Figure 4.2. The ion stopping depth is same value as peak of ions distribution in Figure 4.1.

70 Co Projected Range Eu Projected Range 60

Projected range (nm) 20

0 0 50 100 150 Ion Energy (keV)

Figure 4.2 Stopping range (nm) of cobalt and europium in ZnO target as function of ions energy.

Using MEVVA ion implanter, the ions energies is controlled by electrical voltage, up to 50 kV: ions are accelerated in electric field region generated by large electrical voltage. The kinetic energy of incident ions is thus product of ions charge (q) and electrical voltage (V). For most elements, this type of source produces ions in several charge state, for example Co1+, Co2+ and Co3+. All charge states are implanted resulting a broader distribution rather than achievement with a single charge state of the average value. The proportion of ions energies or ions charge through ionizing process is equal to charge state distribution (CSD) of vacuum arc plasmas for specific element. Anders has theoretically and experimentally studied ion charge state distribution (particle fractions) of many elements, summarized in ref [172-175], e.g. for cobalt: Co+1 = 34%; Co+2 = 59%; Co+3 = 7% and for europium: Eu+1 = 2%; Eu+2 = 86%; Eu+3 = 12%.

For the ideal Zn1-xCox-yEuyO, stoichiometric ratio between cobalt and europium should remain constant for entire depth. In the other word, cobalt ion depth distribution profile should be similar to europium as much as possible. In the same time, substrate contamination must be prevented. For high ions energies, ions can travel through ZnO into substrate layer, e.g. 150 ke V of cobalt ions in Figure 4.1a. During ions are penetrating into target, they can transfer significant energy to recoil atoms, and these can move long distances and create significant collision cascades. When a recoil atom crosses from one target layer to another, the second layer is contaminated as a result of recoil implantation. The theoretical studies suggest that the maximum ion energy of europium is 150 keV and of cobalt is 80 keV for the ZnO thin films with thickness of 100nm. In the experiment, europium ions were implanted at 45 kV and cobalt ions were implanted at 24 kV. The change of morphology in the implanted thin films is caused by cascade collision from both incident ions and recoil ions. Almost every time, an implanted ion collides an atom in the film, creating a vacancy (the atoms in film are knocked away from its lattice site). Hence, a signal recoil atom may cause a lot of vacancy. The level of morphological changes is increasing as t he transfer energy or stopping power is increasing. Comparing to europium, the rate of cobalt ion energy loss is lower, causing less change to the implanted ZnO, therefore cobalt is implanted then followed by europium for the codoped ZnO. Consequently the europium implantation, cobalt recoil ions would diffuse cross into deeper layer.

Figure 4.3 The RBS spectra of Si1-xCox with various current integrator reading

Figure 4.4 The RBS spectra of Si1-xEux with various current integrator reading.

Figure 4.3 and Figure 4.4, show RBS spectra of Si1-xCox and Si1-xEux wafer implanted with various number of current integrator, respectively. For cobalt implanted silicon, the spectra only responded to silicon and cobalt and non-implanted region contained only silicon, therefore fitting parameter is quite simple. This is similar to the europium implanted silicon. The spectra shows backscattering yield by a ll elements within the samples. The cobalt and europium penetrate into silicon target up to 600ML (1 ML = 1015 atoms/cm2) or 120 nm from surface. The amounts of cobalt and europium ions are analysed in all samples. Figure 4.5 shows the amount of cobalt is linearly dependent to the accumulated count of the integrator. Such a phenomenon is similar to the europium doped samples. Therefore, the current integrator reading of 100,000 is equivalent to ions dose of 1.04x1016 atom/cm2 for the cobalt or europium ions, equivalent average 2.5 at% in 100nm ZnO thin film.

Figure 4.5 The dose calibration result for cobalt cathode and europium cathode: implanted dose for cobalt and europium in silicon (0001) wafer as function of current integrator.

The RBS spectra of ZnO:Co 2.5 a t% and ZnO:Eu 2.5 a t%, in Figure 4.6, show the integral of backscattered contribution from dopants (cobalt and europium) and all elements in ZnO as w ell as i n Al2O3 substrates. According to the analysis, the backscattered contribution from cobalt ion overlap that from zinc at the same position as shown in Figure 4.6a while the europium at higher energy appears beside the zinc as shown in Figure 4.6b It must be noted that the data interpretation would be sometimes complicated, for example, in the cases of the sample with a large number of elements or several layer of different compounds. For the implanted ZnO/Al2O3 thin film, RBS analysis would be difficult to be implemented due to the following reasons: (1) both thin film and substrate contain oxygen which integrate to one single contribution; (2) The film growth may cause defect in ZnO, such as VO and Zni, and diffusion between thin films and substrate; and (3) overlap of cobalt and zinc backscattered contribution. In contrast, the silicon wafer is considered as layer of silicon atom which is analysed easier than ZnO/Al2O3 thin film and there is no backscattered contribution overlap of the silicon and cobalt (or europium) as shown in Figure 4.3 (or Figure 4.4). Therefore, dopants amount derived from implanted silicon wafer would be more reliable than that from the implanted ZnO thin film alone. The cobalt and europium implanted into silicon are prerequisite as “dose calibration” and reference for ZnO:Co and ZnO:Eu analysis.

The atomic concentration of all elements in ZnO:Co 2.5 at% and ZnO:Eu 2.5 at% which is analysed with RBS measurement is shown in Figure 4.6. The results suggest that ZnO film used for the europium implantation atomic thickness is slightly thicker than one used for the cobalt implantation. The amount of zinc and oxygen are almost identical in both films. Theoretical and experimental cobalt and europium ions depth distributions of ZnO:Co 2.5 at% and ZnO:Eu 2.5 at% are shown in Figure 4.7. It is discernible that the peak of Co depth profile in the Co implanted ZnO and the peak of Eu depth profile in the Eu implanted ZnO are located at 100-200 ML, approximately 12-24 nm underneath the film surfaces, similar to the depth estimated by TRIM calculation. The concentrations at the peak of distributions are almost identical in both samples. Such a result suggests that samples prepared with this configuration give similar element concentration distribution.

a) 1000 Experimental Simulated 800 O Al Zn ]

C Co

µ 600

cps/50 [ 400 Yield Yield 200

0 50 100 150 200 250 300 350 400 Channel No.

b) 1000 Experimental Simulated 800 O Al Zn

] Eu

C 600 µ cps/50

[ 400 Yield Yield 200

0 50 100 150 200 250 300 350 400 Channel No.

a) 1.0 O 0.9 Al 0.8 Zn Co 0.7 0.6 0.5 0.4 0.3

Concentration (x100 at%) 0.2 0.1 0.0 0 200 400 600 800 1000 1200 1400 Depth (ML)

b) 1.0 O 0.9 Al 0.8 Zn Eu 0.7 0.6

0.5 0.4 0.3

Concentration (x100 at%) 0.2 0.1 0.0 0 200 400 600 800 1000 1200 1400 Depth (ML)

Figure 4.7 Ion depth distribution of all elements in ZnO implanted by cobalt (a) and europium (b) via RBS analysis

Depth (nm) 0 20 40 60 80 100 120 0.06

Co [TRIM Calculation] 0.05 Eu [TRIM Calculation] Co [Experiment] Eu [Experiment] 0.04

0.03

0.02 Concentration (x100 at%) 0.01

0.00 0 200 400 600 800 1000 Depth (ML)

Figure 4.8 Comparison between experimental and theoretical studies of implanted ions depth distribution in ZnO:Co 2.5at% and ZnO:Eu 2.5at%

4.4 Summary

The theoretical calculation was performed to estimate depth distribution of implanted ions in ZnO thin films. The results suggest that the accelerated voltage for cobalt and europium should be 24 kV and 45 kV respectively using MEVVA ion implanter. The amount of incident ions were measured by ion current integrator and the calibration was performed on silicon (0001) wafer for accurate implanted dosed amount. The amount of implanted ions and their depth distribution profiles were measured by R BS. The implanted ions can occupy all crystallographic possible sites in which they can fit. Atomic vacancies in the host materials were created through the energy transfer from the incidental ions. The experimental analyses in the depth distribution of the implanted ions in ZnO:Co and ZnO:Eu provides discernible information for the investigation of

enhancement of Co substitution induced by Eu codoping in ZnO-based diluted magnetic semiconducting thin film.

Chapter 5 Eu Ion Implantation in ZnO Thin Films

5.1 Introduction

The large potential for use of ZnO thin films for electronic applications is due to its attractive optical and electronic properties, such as a large band-gap (~3.3 eV) and high 2 exciton binding energy (60 meV) [176] , large piezoelectric constant e33 (1.3 C/cm ) [177], large thermal conductivity (0.54W/cm K) [83], as w ell as the attractive other mechanical properties such as hardness. ZnO is an intrinsic n-type semiconductor and may become ferromagnetic at room temperature by s mall additions of magnetic ions, and therefore be a possible candidate for spintronics applications [17, 178]. Doping of rare earth ions into the structure of ZnO may change its optical properties, and it was shown that optically-active centers can be created in bulk ZnO by Tm, Tb and Er substituting at the Zn from the ZnO crystallographic lattice [179] or by Eu doping into the ZnO nano-crystals [180]. Reported magnetic and optic properties of doped ZnO are sensitive to the level and the dispersion of dopants. In Co-doped ZnO, it was believed that the ambient temperature of ferromagnetism is induced by t he Co cluster [181], while the materials would exhibit spin glass behaviour if the Co are distributed in the materials homogenously [89]. In the Eu-doped ZnO nano-powder, the photoluminescence (PL) intensity was found to be related to the state of defects, which are caused by t he doping and annealing processes [119]. However, introducing and homogenously distributing of the isolated transition metallic ion into the ZnO thin films is challenging, but a convenient way to attempt this is to use the non-equilibrium ion irradiation technique. Here we study the doping behaviours of the Eu doped ZnO thin films with Eu by low energy ion irradiation.

5.2 Experimental details

The epitaxial ZnO (0001) thin film growth on 10 x 10 x 0.5 mm c-Al2O3 with typical approximate thickness of 200nm, were implanted with the europium ions at low energies under vacuum pressure 2x10-4 Pa (or 2x10-6 mbar). The temperature is initially starting at room temperature and raising up to a couple hundred degree cellules at the end of the implantation. Europium ions were produced using a metal vapour ion source

and were ejected along (0001) direction of the film (normal to the film surface), using a nominal acceleration voltage 45kV with a beam current of 20mA. Charge state distribution (CSD) of europium during the ionization process was reported to be: Eu+1=2%; Eu+2=86%; Eu+3=12% [173]. T RIM simulations show that for the above charge distribution, an acceleration voltage of 45kV results in an implanted layer with skewed-Gaussian distribution implanted layer and a tail on the low energy side, deeper into the film. In order to optimize the process, a number of ZnO thin films were implanted with the europium ions at various doses between 4.16 x 1015 and 1.67 x 10 16 atom/cm2. The equivalent chemical compositions of the europium implanted ZnO film sample was Zn1-xEuxO with x=0.01-0.12.

IBA Sample Chamber Sample Mounted on 2 axis Goniometer

4 Way Slits Energy Recoil ToF Detection Unit Cl5+ Ion Beam 67.5° from Accelerator 45° 67.5° Rotatable Surface surface normal axis Barrier Dectector

Electrostatic Time Detectors

Ion - implanted Surface Barrier Detector

Figure 5.1 Schematic diagram of ERDA experiment showing the sample geometry and component of the Energy recoil ToF detector unit, illustrating from Ref [182]

The implantation depth and distribution was simulated using TRIM code and the depth distribution of elements in the virgin and implanted ZnO films was measured by Elastic Recoil Detection Analysis (ERDA), using 33MeV Cl+5 projectiles and a time-of-flight (ToF) detection system in a vacuum state < 1x10-6 Pa, as illustrated in Figure 5.1. The beam was shaped in a rectangular form, 3 mm high and 1mm wide, and directed onto the sample at an incident angle of 67.5°, between the beam incident direction and the sample normal. The energy of recoils was measured with a time-of-flight detector

placed at an exit angle of 67.5° relative to the sample normal, resulting in a scattering angle of 45°. The configuration of the ToF detector used in this experiment was made up by two electrostatic mirrors with the active foils made of diamond-like carbon (DLC) of 2 µg/cm2, placed 0.5 m apart, and the final rest energy of recoils was measured with a surface barrier detector, placed at the end of the flight path [182]. The time axis projection of the 2D coincidence scatter plot was converted into depth profile using first principle calculations as well as the published information on t he stopping power of ions [183]. In this experiment, the total charge collected from each sample during the measurement was 12 µC. The crystallographic structure was characterized with a Philip

PANalytical X’Pert Pro. Cu Kα1 and Kα2 radiation with a weighted wavelength of 1.540 Å and 1.544 Å was used. The room temperature photo luminescence was conducted on vacuum ultraviolet (VUV) beamline BL03A1 at NSRRC. The optical transmission spectra were performed at room temperature with Perkin Elmer LAMBRA 950 spectrophotometer.

5.3 Results and discussion

The orientation of thin film was confirmed by XRD θ-2θ to be c-axis oriented. The thickness of virgin ZnO thin films was measured by RBS with 1.8MeV He ions, and it was around 1800ML, equivalent to about 200nm, assuming the density of the films was 95% of the density of ZnO single crystal, which is 5.642 g/cm3. The ERDA result for the depth profile of elements present in the as-grown ZnO film grown on c-Al2O3 substrate is shown in Figure 5.2. It reveals, except for Zn, O and Al, small quantities of H and C, as well as some N are observed. It seems that H and C may only present on the surface of ZnO thin film, and are most likely due to the sample handling of the sample in air. However, N appears at both at the surface and interface between the film and the substrate. The present of nitrogen could come from the growth processing while the presence of Al throughout the ZnO film, at a level from 5 to 7 at%, is most likely caused by diffusion from the substrate during the high temperature film growth, rather than from an impure the PLD target. In addition, the stoichiometry of the Al2O3 substrate in the first couple of thousands of mono-layers under the ZnO film may be affected. It exhibits that Al is slightly depleted with the presence of some N.

0.7 H C 0.6 N O 0.5 Al Zn 0.4

0.3

0.2 Concentration [x100 at%] 0.1

0.0

0 600 1200 1800 2400 3000 3600 4200 4800 5400 6000 Depth [ML]

Figure 5.2 ERDA-ToF depth profile of ZnO/Al2O3 thin film

0.7 H C 0.6 N O 0.5 Al Zn 0.4 Eu

0.3

0.2 Concentration [x100 at%] 0.1

0.0

0 600 1200 1800 2400 3000 3600 4200 4800 5400 6000 Depth [ML]

Figure 5.3 ERDA ToF depth profile of the Eu-implanted ZnO/Al2O3 thin film

The ERDA depth profile of the Eu-doped ZnO film, with atomic stoichiometric ratio of

Zn0.96Eu0.04O shows that the implanted Eu is located in the top half of the film. This is in good a greement with the TRIM calculation. However depth distribution of other elements appears to be similar with their depth distribution in the virgin ZnO films, except for the top layer of substrate underneath the film which the variation of chemical stoichiometric ratio appears to be more pronounced. This is a typical technical problem with the growth of high quality ZnO.

ZnO ZnO:Eu - 1 at% (006) 3 ZnO ZnO (002)

O ZnO:Eu - 2 at% 2

Al ZnO:Eu - 3 at% ZnO:Eu - 4 at%

ZnO ZnO (004) Intensity (arb.)

20 30 40 50 60 70 80 2Theta (degree)

Figure 5.4 X-ray diffraction patterns, 2θ-ω scan, of ZnO:Eu 1 at% - 4 at%

Figure 5.4 shows the XRD patterns (strip out Kα2) for ZnO thin film and Eu doped ZnO thin film with maximum concentration up to 4 at%. The result shows Eu-doped ZnO thin film have single phase wurtzite-like structure with the typical (002) and (004) diffraction peak of ZnO. Diffraction peaks rather than ZnO (002), ZnO (004) and Al2O3 (006) are not observed. So there is no e vidence of secondary phase formation and metallic cluster founded. The depth distribution of europium ion in ZnO studied by RBS, illustrated in Figure 4.7b, shows that the thin film was spitted to doped ZnO layer and undoped ZnO layer. In the XRD spectra of the ZnO (002) peak of the Eu-doped

ZnO sample exhibits a peak shoulder as shown in Figure 5.5. The normalized (002) peaks of all samples in Figure 5.5 reveal that the peaks shift from 34.40º of virgin ZnO sample to 34.27º of the Eu implanted ZnO film. This indicates that c lattice constant expanding from 5.210 Å to 5.229 Å is causing by E u implantation. The result also shows the Eu doped ZnO with various Eu concentration up to 4 at% has similar c-axis constant. The relative intensity between main (002) ZnO and (006) Al2O3 is deceased as implantation concentration increase. It is most likely Eu ions did penetrate into the film and substitute the Zn in ZnO lattices. Due to the large atomic size of Eu, it creates a significant distortion in ZnO lattices, resulting in lattice expansion. In contrast, the clusters of Eu would cause ZnO crystallographic degradation rather than lattice expansion.

1.0 ZnO ZnO:Eu - 1 at% ZnO:Eu - 2 at% ZnO:Eu - 3 at% ZnO:Eu - 4 at%

0.5 Normalized Intensity (arb.)

0.0 33.6 33.8 34.0 34.2 34.4 34.6 34.8 35.0 35.2 35.4 2θ (degree)

Figure 5.5 Closed look of ZnO (002) of ZnO (0001) thin flim and Eu implanted ZnO film (bottom)

Figure 5.6 shows the photo-luminescence result measured at room temperature for ZnO:Eu and ZnO film with 141.5 nm excitation. For ZnO film, there are 3 emission bands. The emission band in the ultraviolet region can be observed with a peak at 330 and 380 nm. Another one locates at visible region, peaking at 430 nm. We note the

absence of the commonly observed green emission [15]. The green emission in ZnO represents oxygen vacancies (VO) but it is not found in both samples. The VO could be self-recovery due to implantation process. The 380 nm emission band is associated to energy band energy (Eg) of bulk ZnO while the 430 nm band is attributed to the deep- level emission [184], which is related to oxygen vacancy defects. The 330 nm emission band is due to the radiative recombination of excitons in surfaces associated with the oxygen 2p dangling-bond state. The photo luminescence of ZnO:Eu show the absence of the well documented weak peaks within the red visible range, usually observed in other Eu-doped systems [185]. The emission peak of 330 and 440nm can still be observed while 380nm emission peak is almost invisible which is possible associated with the decrease of the band gap.

3 ZnO ZnO:Eu

) 2 11 x 10 (

Intensity 1

0 300 400 500 600 700 800 λ (nm)

Figure 5.6 Emission photoluminescence spectra at room temperature with 141.5mm excitation for ZnO and ZnO:Eu 4at%

In order to explore this possibility, four ZnO films were doped with Eu by ion irradiation, to a level of 1at%, 2at%, 3at% and 4at%. The optical transmission spectra were performed for all these ZnO: Eu sample included original ZnO at room temperature. The spectra measured with c-Al2O3 substrate reference are showed in

Figure 5.6. UV-VIS transmittance shows ZnO has high UV absorption and high visible light transmission with transmission cut off at ~380 nm. Eu increase absorption intensity of the visible light in ZnO:Eu. These would be affect band-gap energy (Eg) of materials. In the high absorption region of the transmission spectra were Tauc model [169] is applicable. For ZnO system we can write

1 2 αhv= D() hv − Eg 1.15 where α is the absorption coefficient, h is Plank’s constant, υ is the radiation frequency, 2 Eg is the optic band gap and D is constant. The plot of (αhv) as function of energy is shown in Figure 5.8 exhibiting that ZnO film has band-gap energy of 3.296 eV and it is decreased dynamically when Eu concentration in ZnO:Eu is increased from 1 at% to 4 at%. It is clearly demonstrated that band-gap energy of ZnO:Eu is dependant to europium concentration.

100

Transmission (%) Transmission ZnO ZnO:Eu - 1 at% 20 ZnO:Eu - 2 at% ZnO:Eu - 3 at% ZnO:Eu - 4 at% 0 300 400 500 600 700 800 Wavelength (nm)

Figure 5.7 Transmission spectra for ZnO (0001) film and Eu implanted ZnO

film, using c-Al2O3 substrate as reference

ZnO ZnO:Eu - 1 at% ZnO:Eu - 2 at% ZnO:Eu - 3 at% ZnO:Eu - 4 at% 2 ) hv α (

2.4 2.6 2.8 3.0 3.2 3.4 3.6 3.8 4.0 Energy (eV)

Figure 5.8 Plot of (αhν)2 vs. photo energy for ZnO (0001) flim and Eu implanted ZnO thin film. The optical bans gap of the films is determined according to Tauc Model

Ground state band structure calculation were performed on ZnO and ZnO:Eu of which Eu2+ randomly substituted Zn2+, in the crystallographic group P63mc, at proportion equivalent to 1at%, 2at%, 3at% and 4at%. The band structure calculations were performed with the DFT code using CASTEP, in the local density approximation (LDA) with a screen exchange correlation. It is well documented that the DFT theory is reproducing well the shape of the bands, but it under-estimates the band gap by 20-40% as compared with angle-resolved photo-emission spectroscopy [186]. Nevertheless, the band structure calculation results, listed in Table 5.1, show that band gap energy in the Γ direction of the Brillouin zone is 3.2eV for pure ZnO, and it decreases down to 2.4eV as 4 at% of Zn was replaced by Eu. These results, together with the estimated optical band gap Eg suggest that Eu is replacing Zn in the in the ion irradiated ZnO thin film samples.

Table 5.1 The optical band-gap estimation values of sample by opt ical transmission photo spectrometer and estimation by gr ound state band structure in the Γ direction of the Brillouin zone

Sample Experimental Theoretical ZnO 3.297 3.2 ZnO:Eu 1 at% 3.284 3.1 ZnO:Eu 2 at% 3.264 2.9 ZnO:Eu 3 at% 3.244 2.7 ZnO:Eu 4 at% 3.238 2.4

5.4 Summary

The depth profile of elements present in ZnO thin films grown by PLD on c-Al2O3 and implanted with Eu was measured by heavy ion ERDA. The presence of Al and N was detected in the virgin and Eu-doped ZnO films, suggesting a reason for the broad red emission around 1.8eV observed in PL spectra of the virgin ZnO film. The absence of green PL emission in virgin ZnO and Eu-doped ZnO suggests there is no oxyge n vacancy in both films. The expansion of ZnO c-axis is most likely caused by substitional Eu into ZnO. Eu doping between 1at% and 4at%, leads to a decrease of the optic band gap, and by c omparing this result with band structure calculation in the ground state we showed the possibility that Eu is substituted Zn in the ZnO structure of the film.

Chapter 6 Magnetic Properties of Eu implanted ZnO Thin Films

6.1 Introduction

Diluted magnetic semiconductors (DMSs) have attracted enormous interests because of their potential for the innovative spintronics application [17]. In order to achieve Curie temperatures (TC) above room temperature, DMSs are normally fabricated by incorporating transition metal or rare earth ions into a nonmagnetic semiconductor host lattice [187]. A mong many candidates, ZnO-based DMSs with unique physical properties have recently triggered an intense research efforts to develop a n ovel materials [188]. In such a material, the net magnetization in DMS materials should not arise from ferromagnetic inclusions (secondary phases). It should be generated from the localized magnetic moments of the separated ions, being distributed uniformly in the host materials and ferromagnetically aligned via an indirect magnetic coupling. Due to the highest magnetic moment that can be generate from the 7 unpaired electrons in 4f orbital, Eu is considered as a good candidate for a specific dopant in ZnO to achieve a high magnetization in DMS. However, the solubility of Eu in ZnO is strongly limited by the lattice distortion associated with its larger radius. The magnetic interactions between the localized impurities as well as lattice distortions caused by these atoms play an important role in determining atomic arrangements in such materials.

In this chapter, we report on ferromagnetic properties at room temperature of Eu- implanted ZnO epitaxial thin films using ion beam techniques. Then by u sing theoretical ab initio techniques, we reveal the arrangement of Eu dopants in the host lattice as well as the ferromagnetic interaction among Eu ions to ZnO’s native defects.

6.2 Experimental details

The Zn1-xEuxO thin film with x=0.01, 0.02, 0.03 a nd 0.04 w ere prepared using low energy ion implantation. The 100 n m epitaxial ZnO (0001) thin films for ion implantation were grown on 10 x 10 x 0.5 mm c-Al2O3 substrates by Pulse Laser Deposition (PLD) with typical thickness 100 nm. The implantation condition, such as projectile direction, vacuum pressure and ion beam current; have been detailed ones in Chapter 5. The crystallographic structure was analysed using Philip PANalytical X’Pert

MRD Pro using Cu Kα1 and Kα2 radiation with a weighted wavelength of 1.540 Å and 1.544 Å respectively. The films were later cut into approximately area of 5 x 5 mm. By using magnet property measurement system (Quantum design MPMS-XL), zero-field cooled (ZFC) and field cooled (FC) magnetization measurements were carried out in 500 Oe over a wide temperature range of 10K–300 K. In ZFC process, the sample is cooled from room temperature to 10K in absence of magnet field. At 10K, sample has been applying by magnetic field and heating up to 300K while its magnetization is measured as a f unction of increasing temperature with a constant field. In FC process, sample is cooled from room temperature to 10K in the constant applied field and its magnetization is measured as a function of decreasing temperature. The field dependent measurement is also carried out at room temperature. The films were position such that surface of the films is parallel to applied magnetic field. XMCD measurements were performed at the dragon beamline BL11A in National Synchrotron Radiation Research Center (NSRRC), Taiwan. XAS spectra were measured in the total-electron-yield (TEY) mode and the total-fluorescence-yield (TFY) mode.

In our theoretical study, total energy ( Et ) calculations were performed using plane- wave pseudopotential approach of density functional theory as implemented in CASTEP code [189] within the framework of generalized-gradient approximation. Ultrasoft pseudopotentials were represented in reciprocal space and Eu’s 4f were treated as valence electrons. An energy cut-off of 800 eV for plane-wave basis set and a 3×3×1 Γ-centred k-point grid for integration over reciprocal space were used. To model the doped ZnO, a 32-atomic 2 × 2 × 2 ZnO supercell was adopted for calculations with two Eu ions substituting Zn sites along c-axis.

6.3 Result and discussion

The XRD spectra of the Eu implanted ZnO thin films in Figure 6.1 shows all samples have single phase wurtzite-like structure with typical (002) and (004) ZnO diffraction peak of ZnO. There is no t race of europium secondary phase such EuO and Eu2O3 detected in XRD spectra. The soluble Eu ions in ZnO lattice causes peak shifting of ZnO (002) and (004) to lower angle by 0.13º indicating of c-axis expansion from 5.210 Å to 5.229 Å . This suggests Eu ions substitute Zn ions in host lattice causing lattice distortion that is associated with the large difference between atomic radius of Eu and Zn ions.

Zn0.96Eu0.04O

Zn0.97Eu0.03O

Zn0.98Eu0.02O

Zn0.99Eu0.01O Normalized Intensity (arb.)

ZnO ZnO (002) Al2O3 (006) ZnO (004) 20 30 40 50 60 70 80 2Theta (degree)

Figure 6.1 Phase identification: XRD θ-2θ scan of ZnO/Al2O3 and

Zn1-xEuxO/Al2O3 (x = 0.01-0.04)

7 FC 6

) 5 ZFC emu -5

4 x10 ( 3 M

1 H = 500 Oe 0 0 50 100 150 200 250 300 Temperature (K)

Figure 6.2 Zero Field Cooled (ZFC)/Field Cooled (FC) DC magnetization

measurement of Zn0.96Eu0.04O/Al2O3 thin films using SQUID magnetometer with temperature changing rate of 10K/min. (10-3 emu/cm3 = 1 A/m and 104 Oe = 1 Tesla)

The DC magnetization measurement is carried out in a field of 500 Oe. The ZFC/FC measurement, in Figure 6.2, shows there is no paramagnetic transition appeared between 10K and 300K. It suggests the transition temperature would be larger than 300K. The FC curves lie above the ZFC curve with small difference between FC and

ZFC curve indicating Zn1-xEuxO is ferromagnetism with Curie temperature (TC) above room temperature. The irreversible FC and ZFC magnetic behaviours is one of the characteristic feature of spin glass [190, 191].

Figure 6.3 shows isothermal magnetization of the Zn1-xEuxO samples with four different concentrations as a function of the external applied field at room temperature (~300K). The external field is applied parallel to the film surface. Such a configuration can minimize the diamagnetic contribution from c-Al2O3. The sample magnetization was carefully subtracted with undoped ZnO thin films and normalized with the number of

the implanted dose. The result demonstrates that all samples possess ferromagnetism at 300K with magnetic hysteresis and coercive fields of ~50 Oe. As the resulted signals are slightly above the background, a clear loop is difficult to obtain. But, again, compared to the measurements of the diamagnetic substrate, the remanence and hysteresis are clearly discernible. For x ≤ 2, the saturated magnetization (MS) increases with increasing concentration of europium content while MS of x= 3 and x=4 are almost identical, having a magnetization approximate 2 µB/Eu. These suggest higher Eu concentration lead to enhanced anti-ferromagnetic coupling, resulting in a lower magnetic moment per Eu. On the other word, Eu ions form anti-ferromagnetic coupling when their distance is shorter, as the concentration increasing.

20 T = 300K 15

5 ) /Eu

B 0 µ (

M -5

-10 ZnO:Eu - 1 at% ZnO:Eu - 2 at% -15 ZnO:Eu - 3 at% ZnO:Eu - 4 at% -20 -5000 -4000 -3000 -2000 -1000 0 1000 2000 3000 4000 5000 H (Oe)

Figure 6.3 M-H curve of Eu implanted ZnO film at 300K measured by SQUID magnetometer, after the subtraction of diamagnetic

contribution of Al2O3 substrate. The external applied field was applied parallel to the film surface.(104 Oe = 1 Tesla)

It is well known that magnetism in ZnO could be induced by non-magnetic defects and non-magnetic dopants, e.g. room temperature ferromagnetism in ZnO:C [192-194] and

ZnO:N [195], high spin state of zinc vacancies (VZn) [196], and oxygen vacancies (VO)

[197]. Due to technical problem in growth ZnO thin film without non-magnetic defects and non-magnetic dopants, therefore, the magnetism in ZnO:Eu could be induced by those defects as well. It is difficult to discriminate what makes a major contribution to the observed ferromagnetism. In our studies, the magnetism induced by above defects prior to implantation can be negotiated. It is because the measured magnetism has been subtracted by magnetism of virgin ZnO thin film. Thus, the observed magnetism showed in and was caused by magnetic induction by Eu ions and subsequently defects causing by ion implantation. In addition, the vacancies’ moments provide few order of magnetism which is insufficient to allow ferromagnetism. The ion irradiation on semiconductors typically generates vacancy-interstitial pair that compensates each order, and thus partially occupied defect-induced bands are not formed. [75].

To verify this finding, total energy ( Et ) of were calculated with a normalized atomic stoichiometric ratio Zn16Eu2O16 system in two spatial arrangements by using CASTEP code: (a) Eu ions were separated by only one oxygen ion and (b) Eu ions were separated by a chain of -O-Zn-O- ions, as shown in Figure 6.4 For both configurations geometry relaxation was performed allowing internal coordinates to relax until the Cartesian components of atomic forces acting on all ions in the supercell was < 0.05 eV/Å and simultaneously the energy converged to 10-5 eV per step, per atom. The calculation result is shown in Table 6.1

Table 6.1 The total energy of Zn16Eu2O16 for parallel (FM) and anti-parallel (AFM) Eu-Eu coupling calculated by CAPTEP code and atomic distance for Eu-O and Eu-Eu before and after geometry relaxation

t t DEu-O(Å) DEu-O(Å) .Configuration E AFM (eV) ΔE FM (eV) DEu-Eu(Å) c-axis ab plane

2.306 2.273 a 0.000 0.040 3.753 (1.991) (1.975)

2.263 2.210 b -0.044 0.036 4.460 (1.991) (1.975)

The result shows Eu ions anti-ferromagnetic coupling is preferable for both t t configurations. The E AFM is lower than E FM which make anti-parallel spins more stable. Due to the large difference of atomic radius, incorporation of Eu ions in ZnO introduces local lattice distortion. After the geometry relaxation, the expansion of Eu-O bonding length can be observed. By analysing the difference between the Eu-O bond length in unrelaxed and relaxed structures of, the bond length is increasing by approximately 15% and 13% for configuration 1 and 2 respectively. This indicates the expansion strain caused by substitutional Eu ions reduces as inter-atomic distance between Eu ions is larger.

1.0 ) 0.5 emu -4 0.0 x 10 ( -0.5 M -1.0 -6000 -4000 -2000 0 2000 4000 6000 H (Oe) Intensity (arb.)

1110 1120 1130 1140 1150 1160 1170 1180 1190 Energy (eV)

Figure 6.5 The x-ray absorption coefficient µ ()E of Zn0.96Eu0.04O at Eu

M4,5-edge measured at 77K and applied field of 1 Tesla in TFY mode. The inset of this figure displays M-H curve measured at same temperature using SQUID magnetometer

Table 6.2 The sum rule analysis for orbital moment (µorb), spin moment

(µspin) and total spin (µtotal = µorb+ µspin) of europium ions in

Zn0.96Eu0.04O thin film measured in TFY and TEY mode

Measurement Mode µspin (µB/Eu) µorb (µB/Eu) µtotal (µB/Eu)

TEY 7.141 -2.324 4.817

TFY -0.464 0.141 -0.323

Figure 6.6 The XAS and XMCD spectra of Zn0.96Eu0.04O at Eu M4;5-edges: left side is measured in TEY and right side is measured in TFY mode, the top spectra is XAS and bottom spectra is XMCD

The XAS spectra in Figure 6.5 shows Eu M4,5-edge photon absorption of Zn0.96Eu0.04O at the applied field of 1 Tesla and temperature of 77K. The X-ray absorption coefficient is defined as µ()E =() µµ+− + /2, where µ + and µ − are the x-ray absorption coefficient measured with photon helicity parallel and anti-parallel to magnetization direction respectively. The absorption at M5-edge()34df5/2 → shows at least 2 peaks that may be associated with the multiplet splitting of Eu3+ ions. The positions of the multiplet peaks in M4;5 absorption edge is in good a greement with the theoretical analysis reported in the literature [198]. Figure 6.6 shows the XAS and XMCD at Eu M4,5-edge of Zn0.96Eu0.04O measured in TFY and TEY mode at 77K. XMCD is defined as

∆=−µ()E µµ−+ and the sample magnetization is induced by t he applied field of 1 Tesla. The hysteresis [inset in Figure 6.5] shows the sample magnetization is saturated at 0.5 Tesla. Because the XMCD spectra measured at the Eu M4,5-edges, shown in figure xx is associated with a 4f final state, it directly gives a measure of the magnetic polarization in the 4f shell of Eu ions. The sum rule analysis enables the estimation of spin and orbital magnetic moments of Eu from the measured intensities of XAS and

XMCD [148]. The sum rule analysis of the experimental Eu M4,5-edge XAS and XMCD 3+ in Figure 6.6, the Eu with the assumptions of 7 unoccupied states in the 4f shell (N4f =

7) has yield orbital moment (µorb), spin moment (µspin) and total spin (µtotal = µorb+ µspin) as shows in Table 6.2.

The XMCD spectra suggest magnetic moment of europium ions measured in TEY mode have opposite polarization to ones measured in TFY mode. TEY mode has short detection range, ~ 5 nm, therefore it can only provide surface properties while TFY has larger detection range, covering entire depth of Zn0.96Eu0.04O thin film. Near thin film surface, Eu µtotal have same the polarization as µsample with a magnitude of 4.817 µB/Eu but average Eu µtotal of entire sample have opposite polarization with a magnitude of

0.323 µB/Eu. These XMCD analyses are well agreed to the magnetic measurement result shown in Figure 6.3. According to the ion depth distribution investigation in Chapter 4, t he concentration of europium near surface is little therefore the surface properties of Zn0.96Eu0.04O would be similar to the undoped or diluted Eu implanted

ZnO. The bulk properties suggest that the decreasing of average Eu µtotal is caused by anti-ferromagnetic coupling enchantment as increasing Eu concentration. The difference

between average Eu µtotal from XMCD and MS from SQUID measurement also suggests magnetic moment do not cause by europium alone.

6.4 Summary

The room temperature ferromagnetic properties were observed in Zn1-xEuxO thin film prepared by ion implantation. The XRD spectra show there is no trace of secondary phases of europium ions in Zn1-xEuxO thin film. The diffraction peak and geometry relaxation suggests that Eu ions substitute Zn ions in the host lattice, thus causing the lattice distortion due to large difference in their atomic radius. The magnet measurement shows the diluted Zn1-xEuxO provides larger magnetization per amount of europium ions which is in agreement with XMCD analysis. The critical XMCD measurement results were not highlighted here. The anti-ferromagnetic coupling of Eu is a new finding and it provides a reference and guidance for the co-doped technology.

Chapter 7 Enhancement of Co substitution induced by Eu codoping in ZnO-based Diluted Magnetic Semiconducting Thin Films

7.1 Introduction

Diluted magnetic semiconductors (DMSs) have attracted enormous interests because of their potential for innovative spintronics application [17, 178]. Currently, the most common DMSs are II-VI and III-V compounds (such as CdTe, ZnSe, CdSe, ZnO, GaN, GaAs, etc., with transition metal ions e.g., Mn, Fe, or Co) substituting their original cations. Although values of Curie temperature above room temperature have been reported in Co-doped TiO2 [199], ZnO [181], and Mn-doped GaN [200], the possibility of an extrinsic origin of the ferromagnetism, such as ferromagnetic impurity segregation, could not be ruled out in most cases. It was reported that the nanoscale ferromagnetic phase in Co-doped ZnO thin films is due to the Co clusters [152]. This gives rise to an argument: homogeneous films of (Zn,Co)O exhibited a spin-glass behaviour while room temperature ferromagnetism was only found in inhomogeneous films attributed to the presence of Co clusters [89]. Such an argument on the origin of ferromagnetism in (Zn,Co)O is caused by the often-performed magnetization measurements with a magnetometer [e.g., superconducting quantum interference device (SQUID)], which is unable to distinguish an intrinsic ferromagnetic signal from an extrinsic one [201]. Recent progress for the investigation of the nature of ferromagnetism in (Zn,Co)O with other advanced materials characterization tools provides a better way to develop the intrinsic DMS materials.

Although ion implantation is a common and simple technique to fabricate semiconductor materials with the advantage of creating non-equilibrium doping suited for achieving desired dopant concentrations, it was reported that the occurrence of Co clustering is frequently observed in the (Zn,Co)O thin films prepared by the ion implantation technique [102]. This has impeded the application of ion implantation in the fabrication of DMSs. Therefore, the development of dilute magnetic-impurity-doped functional wide-band gap semiconductors, such as (Zn,Co)O, with high temperature

ferromagnetism is eagerly awaited if it is unambiguously established to be intrinsic in nature [201]. In this study, we report the experimental evidence of Co substitution in ZnO epitaxial films induced by C o and Eu codoping as well as t he results on the localized magnetic moment of Co in Zn0.92Co0.04Eu0.04O determined by x-ray absorption spectroscopy XAS and x-ray magnetic circular dichroism XMCD.

7.2 Experiment details

The Zn0.96Co0.04O and Zn0.92Co0.04Eu0.04O materials were prepared by ion implantation [202]. The ZnO (0001) epitaxial thin film, ~100 nm thick grown on c -sapphire, was implanted with Co and Eu ion beams perpendicular to its surface under a vacuum pressure of 2x10-4 Pa (or 2x10−6 mbar). The TRIM calculation [203] was performed to optimize implanting parameters for the ZnO epitaxial thin film. For the codoped sample of Zn0.92Co0.04Eu0.04O, Co was implanted first followed by Eu. The dose concentrations and the ion distribution depth profiles were verified by Rutherford back scattering (RBS) measurements. At the above mentioned implantation parameters, both Co and Eu have an average concentration of 4%. The crystallographic structures of the as-prepared materials were characterized using X’Pert PRO Materials Research Diffractometer that is specially designed for x-ray diffraction (XRD) characterization of thin films with Cu

Kα1 radiation. The magnetic properties of doped ZnO films were determined using a SQUID magnetometer (Quantum Design MPMS-XL). Local electronic structure and magnetic moments of Co were studied by XAS and XMCD at BL11A dragon beamline of National Synchrotron Radiation Research Centre in Taiwan. XAS spectra at the Co

L2,3 edge were measured in total fluorescence yield mode to probe the properties throughout the entire thickness of the films at both room temperature and 77 K . The propagation direction of the incident circular polarized x-ray 80% polarization was 30° with respect to the surface of the samples. A magnetic field of 1 T was applied along the surface plane of the samples. XMCD spectra originated from the difference of the spin- resolved XAS corresponding to two opposite magnetic fields after proper background subtraction.

7.3 Results and discussion

Figure 7.1 shows the depth profiles of Co and Eu, which are analysed with RBS measurement, in the as-prepared thin films. It is discernible that the peak of Co depth profile in Zn0.96Co0.04O the sample with Co dopants only is located at ~15 nm underneath the film surface. The subsequent Eu implantation changed the implanted Co distribution profile of depth. Although the peak position of Co depth profile remains unchanged, Co distribution profile broadens, having more Co in the depth range of >25 nm by comparing to Zn0.96Co0.04O. In Zn0.96Co0.04Eu0.04O, the peak of Eu depth profile is positioned at ~26 nm, similar to the depth estimated by TRIM calculation. It is noted that in this work our strategy is to maintain the similarity of Co distribution profiles between the Co-doped ZnO film and the Co–Eu codoped ZnO film to investigate the subtle interplay between Co and Eu.

7 Co in Zn0.96Co0.04O Co in Zn Co Eu O 6 0.92 0.04 0.04 Eu in Zn0.92Co0.04Eu0.04O ) 2 5

atom/cm 4 13

x10 ( 3 Dose 2

0 0 20 40 60 80 100 Depth (nm) Figure 7.1 Distribution of Co and Eu ions as a function of depth. X axis is depth (nanometres) from the top surface and Y axis is Co or Eu dose at a particular depth.

(0006) 3 ZnO (0002) O 2 Al Zn0.92Co0.04Eu0.04O

Zn0.96Co0.04O Normalized Intensity

ZnO 30 40 50 60 70 80 2Theta (degree)

Figure 7.2 θ-2θ XRD patterns of virgin ZnO, Zn0.96Co0.04O, and

Zn0.92Co0.04Eu0.04O thin films. Inset shows the normalized (0002) peak.

XRD spectra of ZnO, Zn0.96Co0.04O, and Zn0.92Co0.04Eu0.04O thin films in Figure 7.2 show that the as-implanted films were single phase with a wurtzite like structure. No secondary phases including Co and Eu alloys were detected in the samples. The strong and sharp (0002) diffraction found in the as-grown ZnO film shows that the ZnO film was well crystallized. A small left shift of the (0002) peak appearing in Zn0.96Co0.04O demonstrates the expansion of c lattice parameter caused by Co dopants. The c was expanded from 5.2084 Å for the ZnO film to 5.2099 Å for the Zn0.96Co0.04O. It was reported that the substitution of Zn2+ ion radius of 0.60 Å with high-spin Co2+ ion radius of 0.58 Å in tetrahedral coordination results in the reduction in c with large compressive strain [97, 181]. Therefore, the expansion of c induced by Co suggests that Co ions do not fully substitute Zn in the wurtzite lattice points with +2 formal oxidation state. It implies that the Co ions may exist with the form of either Co metal nano-clustering or the interstices accommodated in octahedral interstitial sites of the wurtzite structure or both in Zn0.96Co0.04O. The lattice constant c of the Co and Eu codoped sample

Zn0.92Co0.04Eu0.04O is 5.2127 Å which is larger than that of Zn0.96Co0.04O. It is

understandable that the addition of large Eu ions into the Co-doped ZnO can result in a larger c than that of Zn0.96Co0.04O due to the substantial difference of ion radii between 2+ 3+ Zn and Eu, where the radii of Eu and Eu are 1.17 and 1.03 Å, respectively [97, 181]. The normalized (0002) peaks are shown in the inset of Figure 7.2. It is found that the peak of Zn0.96Co0.04O is asymmetrical and looks like consisting of two peaks in the first glance. However, if it were true, the other peaks of a secondary phase should be observed in the XRD spectra. So the observed phenomenon may be due to inhomogeneous distribution of Co in the Co-doped ZnO film. Furthermore, the full width at half maximum of the (0002) peak reduces remarkably for Zn0.92Co0.04Eu0.04O compared to that for Zn0.96Co0.04O. These strongly indicate that Eu and Co couples to reduce inhomogeneous strain, meaning the existence of a subtle interaction between Co and Eu in the codoped materials. The coupling of the compressive and tensile strains caused by the doped Co and Eu individually can minimize the energy of system. This may break the network of the Co metal nano-clusters, thus enhancing Co substitution at Zn lattice sites.

Figure 7.3 shows isothermal magnetization curves of Zn0.92Co0.04Eu0.04O and

Zn0.96Co0.04O as a function of the external magnetic field applied parallel to the film surface. It demonstrates that Zn0.92Co0.04Eu0.04O and Zn0.96Co0.04O possesses ferromagnetism at room temperature with a co ercive field of 60 Oe as shown in the 3 hysteresis loop. The saturation magnetization (Ms) of 65 emu/cm , equivalent to 2.09 3 µB/(Co+Eu), with the remanent magnetization of 5.2 e mu/cm and the Ms of 50 3 3 emu/cm , equivalent to 3.21 µ B/Co, with the remanent magnetization of 3.0 e mu/cm were determined for Zn0.96Co0.04Eu0.04O and Zn0.96Co0.04O, respectively. The presence of ferromagnetism in Zn0.96Co0.04O supports Co’s clustering in the film. The ferromagnetism in Zn0.92Co0.04Eu0.04O may be sourced from the substitution of the implanted Co and Eu. Our theoretical studies using ab initio calculations show that Eu dopants in the ZnO host lattice are preferably substitutional at the Zn site by examining three different possible cases: (i) the octahedral interstitial site of the perfect hexagonal ZnO lattice, (ii) the octahedral interstitial site where a neighbouring Zn vacancy exists, and (iii) the substitutional Zn site. Total energy calculations show that in

Zn0.875Co0.0625Eu0.0625O, the ferromagnetic interaction between Co and Eu is stronger than the ferri-magnetic one where the spin alignment of Eu and Co ions is anti-parallel.

4 Room Temperature 3

) 1 dopant / 0 B µ

(

M -1

-2

Zn0.92Co0.04Eu0.04O -3 Zn0.96Co0.04O -4 -3000 -2000 -1000 0 1000 2000 3000 H (Oe)

Figure 7.3 Magnetization curves of Zn0.92Co0.04Eu0.04O and Zn0.96Co0.04O thin films measured by a S QUID magnetometer at room temperature, after the subtraction of diamagnetic contribution of

Al2O3 substrate. The external applied field was applied parallel to

the film surface. (Note: dopants for Zn0.92Co0.04Eu0.04O and 4 Zn0.96Co0.04O refer to Co+Eu and Co, respectively, 10 Oe = 1 Tesla)

RT 77K (arb.)

XAS (arb.)

XMCD

770 780 790 800 Energy(eV)

Figure 7.4 XAS spectra (top) and their derived XMCD spectra (bottom) of

Zn0.92Co0.04Eu0.04O measured at the Co L2,3 absorption edges at room temperature and 77 K.

Figure 7.4 shows XAS and XMCD spectra of the Co L2,3-edge of Zn0.92Co0.04Eu0.04O measured at room temperature and 77 K . The absorption at the Co L2,3 edge is associated with Co 2p to 3d transition, which provides information of unoccupied density of Co 3d states by probing the projected 3d magnetism of Co [204]. It can be seen from the XAS and XMCD spectra that the absorption at Co L2 and L3-edge is different from those of Co metals. By comparing with the theoretical XAS results,[166] it is discernible that the observed tetrahedral multiple structures are consistent with the local electronic structure of Co substitution at the tetrahedral Zn site. This provides strong evidence of Co substitution induced by t he Eu codoping. From our comprehensive investigation with first-principles density functional calculations, it is interesting to find that the possibility of Eu2+ and Eu3+ coexistence can be ruled out and only Eu2+ can appear in the materials with the spinning state of 7. This detail work will be reported in elsewhere. The Co magnetic spin moment MS of Zn0.92Co0.04Eu0.04O derived from XMCD sum rules [205] is 0.082 ± 0.020 µB per Co atom at room

temperature and 0.096 ± 0.014 µB per Co atom at 77 K. It suggests that the temperature has very limit effects on the localized magnetic spin moment of Co in

Zn0.92Co0.04Eu0.04O. XMCD spectra of the Co L2,3-edge of Zn0.92Co0.04Eu0.04O and

Zn0.96Co0.04O measured at 77 K are compared in Figure 7.5. The ms of Co in

Zn0.96Co0.04O is 0.191±0.016 µB per Co atom at 77 K, which is obviously larger than that for Zn0.92Co0.04Eu0.04O. The large magnetic spin moment from the Zn0.96Co0.04O may be contributed from Co metal nano-clusters presented in the sample. On the other hand, the smaller Co magnetic spin moment in Zn0.92Co0.04Eu0.04O may be caused by enhanced Co substitution in Zn0.92Co0.04Eu0.04O, indicating a strong interaction between Co and Eu in the codoped sample. This is in agreement with aforementioned magnetic properties which further support the notion that the Co substitution was facilitated by Eu codoping in ZnO. (arb.)

XMCD

Zn0.92Co0.04Eu0.04O

Zn0.96Co0.04O 770 780 790 800 Energy (eV)

Figure 7.5 XMCD spectra of Zn0.92Co0.04Eu0.04O and Zn0.96Co0.04O measured

at the Co L2,3 absorption edges at 77 K.

It is well known that the Co metal nano-clustering easily forms in vacuum processing environment. It is often observed in the samples prepared with ion implantation techniques,[206] thus resulting in the large localized spin moments.[102] Consequently this process may facilitate the replacement of Zn with Co in the wurtzite lattice.

Therefore, the Co magnetic contribution in Zn0.92Co0.04Eu0.04O thin films would arise from the intrinsic Co substituting Zn. The energy of Eu L edge is too high to be determined due to the facility measurement limit. Since the signal at M4,5 edge in our measurement is very weak with a strong background of noise, it leaves an open question on the measurement of magnetic spin and orbit moments of Eu for future investigation. It is believed that the Co–Co spin coupling in perfect ZnO lattice structure is anti- parallel,[188] similar to Eu–Eu spin coupling. However, total energies calculations show that in its most stable configuration of Zn0.875Co0.0625Eu0.0625O parallel Co-Eu spin coupling is stronger by 57 meV than anti-parallel. Therefore, the magnetic contribution in Zn0.92Co0.04Eu0.04O thin films would arise from the ferromagnetic interaction between substitutional Co and Eu. Atomic ratio of Co and Eu needs to be optimized to achieve the best performance. A systematic study on the detailed interaction between Co and Eu on crystal and electrical structures with XAS to analyse the extended x-ray absorption fine structure is currently in progress. Furthermore, the room temperature photoluminescence was used to determine the structural disordering induced by the implantation process and its results show no artefact is induced during the materials preparation.

7.4 Summary

We have studied the structural and magnetic properties of Zn0.92Co0.04Eu0.04O thin films prepared by i on implantation technique. Room temperature ferromagnetism was observed with a saturation magnetization of 65 emu/cm3, a coercivity of 60 Oe and a 3 remanent magnetization of 5.2 emu/cm . XAS spectra of Co L2,3-edge show that the Co ions reside in the tetrahedral coordination, demonstrating its substitution of Zn in the wurtzite structure. XMCD spectra of Co L2.3-edge also reveal that Co metal nano- clustering may not form in Co and Eu co-implanted ZnO as evidenced by Co tetrahedral multiplet structures. It suggests that Co substitution at the Zn sites may be facilitated by Eu codoping and thus the Co metal clustering is suppressed significantly

100

Chapter 8 Coordination Structure of Zn1-xCoxO, Zn1-xEuxO

and Zn1-xCox-yEuyO Thin Films studies using XAFS

8.1 Introduction

In Chapter 5 and 6, w e have studied the properties of europium implanted ZnO thin films. The experimental results suggest that the europium ions substitute zinc ions in the host lattice. The large difference between the atomic radius of zinc and europium ions results in the lattice distortion and axis expansion. For the europium implanted ZnO thin film, higher europium concentration enhances anti-ferromagnetic coupling between the europium ions and then reduces the band-gap energy (Eg) of the system. Surprisingly, the europium implantation remarkably enhances the ferromagnetic moment in the cobalt doped ZnO, which has been discussed in Chapter 7. In this chapter, the local structure of europium implanted ZnO thin films as well as cobalt and europium co-implanted ZnO thin films are experimentally investigated by X-ray Absorption Fine structure Spectroscopy (XAFS) with Synchrotron Radiation.

8.2 Experimental details

The europium implanted ZnO as well as the cobalt and europium co-implanted ZnO thin film were prepared using low energy ion implantation. The epitaxial ZnO (0001) thin films for ion implantation were grown on 10 x 10 x 0.5 mm c-Al2O3 substrates by Pulse Laser Deposition (PLD) with typical thickness 100 nm. The temperature is increased from ambient temperature to a couple hundred degree cellules at the end of implantation. The Cobalt and europium ions were produced using a metal vapour ion source and were ejected along (0001) direction of the film (normal to the film surface), using a nominal acceleration voltage 24kV and 45kV for cobalt and europium respectively with a beam current of 20mA. The equivalent implanted dose for each implantation is 1.67x1016 atom/cm2. The corresponding chemical composition for the implanted ZnO films is Zn0.96Eu0.04O and Zn0.92Co0.04Eu0.04O. Both samples are annealed in Argon atmosphere at 600oC and 1 atm (or 1.013x105 Pa) for 0.5 hour as

101

described in Chapter 3. The crystallographic structure was analysed using Philip PANalytical X’Pert MRD Pro for phase identification. The x-ray absorption spectroscopy (XAS) technique is a powerful tool to characterize the local structure of DMSs. The local structure of both samples is studied by x-ray absorption near edge structure (XANES) and extended x-ray absorption fine structure (EXAFS) at Wiggler beamline BL17C1 of National Synchrotron Radiation Research Centre, Taiwan. The measurement is performed using circularly hard x-ray at room temperature with photon incident angle 45o to thin film normal axis. The sample is mounted on t he rotating sample holder. The x-ray absorption coefficient at Zn K-edge, Co K-edge and Eu L3- edge were recorded in total fluorescence mode using photomultiplier.

8.3 Result and discussion

Zn0.92Co0.04Eu0.04O

Zn0.96Eu0.04O

Intensity (arb.) Zn0.96Co0.04O

ZnO ZnO (002) Al2O3 (006) ZnO (004) 20 30 40 50 60 70 80 2Theta (degree)

Figure 8.1 θ-2θ x-ray diffraction spectra for virgin ZnO, Zn0.96Co0.04O,

Zn0.96Eu0.04O and Zn0.96Co0.04Eu0.04O thin films

102

Figure 8.1 shows θ-2θ x-ray diffraction spectra of Zn0.96Eu0.04O and Zn0.92Co0.04Eu0.04O thin films. It demonstrates that all samples have single phase wurtzite-like structure with typical (002) and (004) ZnO diffraction peaks, exhibiting the epitaxial growing patter with c-axis orientation. There is not signal of cobalt’s or europium’s secondary phase detected.

2.0

1.5 (E)

µ 1.0

Normalized 0.5

0.0

9600 9800 10000 10200 10400 Energy (eV)

Figure 8.2: The X-ray absorption coefficient µ(E) of epitaxial ZnO (0001) thin films at Zn K-edge (9659 eV).

It is believed that the x-ray absorption spectroscopy (XAS) is a powerful tool to characterize the local chemical environment of material structure. The x-ray absorption coefficient µ(E) near Zn K-edge (9659 eV) of the epitaxial ZnO (001) thin films is shown in Figure 8.2. The pre-edge and position of the absorption edge contains information on va lence states and electron configurations of the probed atom. The absence of the pre-edge suggests that Zn has filled 3d or bital, verifying that Zn ions have oxidation state of +2 with electron configuration [Ar]4s03d10. The structure around the probed atom can be described by studying the fine structure above the absorption edge [207]. EXAFS analysis was performed using IFEFFIT which is a su ite of

103

interactive programs, including Athena and Artemis, for XAFS analysis [137, 138]. Athena addresses all aspects of data processing including conversion of raw data to µ(E) , background subtraction, Fourier transform, and data plotting, while the data fitting and theoretical analysis were handle by Artemis. ARTEMIS works within the framework of FEFF’s multiple-scattering path expansion [134-136]. Figure 8.3 s how the EXAFS spectra kχ(k) of ZnO (001) epitaxial film as a function of the photoelectron wave vector k. The oscillation pattern is caused by constructive and destructive interference of multi- scattering wave. To minimize the error, only the data between the k ranges of 2-12 Å were used in further analysis.

0.8

0.4 (k)

0.0 χ k

-0.4

-0.8

0 2 4 6 8 10 12 14 k

Figure 8.3 Zn K-edge EXAFS spectra kχ(k) of epitaxial ZnO (0001) thin film as a function of the photoelectron wave vector k.

104

Figure 8.4(a) shows the radial distribution function [RDF], which is the Fourier transformation amplitude of EXAFS data at Zn K-edge of the epitaxial ZnO (001) thin film. T he peaks are corresponding to the positions of the probed atoms which contribute to the scattering wave construction. The first and second main peaks in the radial distribution function are associated with the nearest oxygen and Zn atoms from the central Zn atom respectively. The spectra were fitted in R-space with k-weight of 1, 2 and 3 i n the Fourier transform of the data. The fitted model is ZnO with wurtzite structure in a cell cluster of 77 atoms (up to 6Å from central atom). It is known that ZnO has three well-known polymorphs: (i) wurtzite (P63mc space group); (ii) zinc blende (F-43m) and (iii) rocksalt (Fm3m). Among these three structures, wurtzite structure is more thermodynamic stable at ambient conditions [15, 208]. The electron single and multi-scattering paths were taken into account for the EXAFS fitting. To optimize the models, each scattering paths length is written as math expression of lattice parameters: a, c and u. The magnitude, real part and imaginary part of the Fourier transform of the fit and data for a k-weight of 2 is shown in Figure 8.3. This model is well descried the experimental data. The inter-atomic distance/scatting path between the absorber and its neighbours, and σ2 were listed in Table 8.1. The value of the amplitude

2 reduction factor S0 has been found to lie around 0.94 ± 0.09 Å, and the energy shift E0 are 3.05 ± 1.71 eV and 6.38 ± 0.99 eV for Zn and O respectively. The fit values for the lattice constants a, c and u are 3.264 ± 0.039 Å, 5.176 ± 0.070 Å and 0.3825 ± 0.019 Å respectively.

105

2.8 Experimental 2.4 Fitting 2.0 1.6 1.2 | (kw =2) ) R ( 0.8 χ | 0.4 0.0 0 1 2 3 4 5 6 R (Å) 2 Experimental Fitting 1

0 (kw = 2) )] R ( χ [ -1 Re

-2 0 1 2 3 4 5 6 R (Å) 2 Experimental Fitting 1

0 (kw = 2) )] R

( -1 χ [

Im -2

0 1 2 3 4 5 6 R (Å)

Figure 8.4 The Fourier Transformation of k2χ(k) EXAFS spectra in R-space: (a) Magnetitude, (b) real part and (c) imaginary part

106

Table 8.1 EXAFS fitting result for refinement of Zn K-edge for ZnO thin film

Path Degeneracy R (Å) R expression σ2 (x10-3)

Zn-O1 1 1.980 r1 = uc × 5.3

1 2 2 Zn-O2 3 1.980 a 2 5.3 ru2 =+−()0.5 3

1 222 Zn-Zn1 6 3.201 ac 7.6 r3 = + 34

Zn-O3 1 3.196 r4 =−×()1 uc 5.6

Zn-Zn2 6 3.264 ra5 = 7.6

rrr++ Zn-O1-O2 6 3.581 123 13.5 2

rrr123++ Zn-Zn1-O1 6 3.581 13.5 2

rrr123++ Zn-Zn1-O2 6 3.581 13.5 2

r Zn-O2-O2 6 3.621 2 + r 13.5 2 5

r Zn-Zn2-O2 12 3.621 2 + r 13.5 2 5

1 2 142 Zn-O4 3 3.818 r= − uc ⋅+22 a 7.9 6  23

1 Zn-O5 6 3.818 2 222 7.9 r7 =() a + uc

107

Zn0.92Co0.04Eu0.04O

Zn0.96Eu0.04O Zn0.96Eu0.04O (E) µ (k) Zn Co O χ

0.96 0.04 k Zn0.96Co0.04O Normalized

ZnO

9600 9650 9700 9750 9800 0 2 4 6 8 10 12 Energy (eV) k

Figure 8.5 Zn K-edge XAFS spectra comparison between ZnO and implanted ZnO thin film. (a) XANSE and (b) kχ(k) EXAFS

Figure 8.5 shows Zn K-edge XANES spectra and EXAFS spectra of the implanted ZnO thin films. They are similar to the raw ZnO thin film. This suggests that the implanted ZnO thin films have same crystallographic phase structure as the raw ZnO thin film. In the Zn K-edge XANES spectra, the pre-edge peak is disappeared, which is a sign of 4s03d10 configuration of Zn2+. The magnitude of the Fourier transformation of the experimental data for k w eight of 2, as shown in Figure 8.6, shows that the local neighbour of Zn absorber in all the implanted films are similar to that in the raw ZnO thin film. The results were fitted with the same model and parameters previously used for ZnO thin film. The inter-atomic distance between the Zn absorber and the first two

108

nearest neighbours, and lattice constant of those thin films are listed in Table 8.2. It is clear that the inter-atomic distance between the absorber and its neighbours was not affected by the implanted ions strongly. The lattice constants are only changed slightly.

Zn0.92Co0.04Eu0.04O

Zn0.96Co0.04O

| (k-weight |= 2) (k-weight Zn Eu O ) 0.96 0.04 R ( χ |

ZnO

0 1 2 3 4 5 6 R (Å)

Figure 8.6 The Fourier Transformation of k2χ(k) Zn K-edge EXAFS data (symbols) and fitting (line) for ZnO and implanted ZnO thin film.

Table 8.2 ZnO, and implanted ZnO atomic distance and lattice parameter obtain by EXAFS analysis.

DZn-O1 (Å) DZn-Zn1 (Å) a-axis (Å) c-axis (Å)

ZnO 1.980 3.201 3.264 ± 0.039 5.176 ± 0.070

Zn0.96Co0.04O 1.981 3.200 3.262 ± 0.022 5.171 ± 0.049

Zn0.96Eu0.04O 1.979 3.199 3.261 ± 0.023 5.172 ± 0.051

Zn0.92Co0.04Eu0.04O 1.979 3.197 3.265 ± 0.039 5.166 ± 0.039

109

Eu L3-edge XANES spectra of Zn0.96Eu0.04O and Zn0.96Co0.04Eu0.04O thin films are shown in Figure 8.7 with the standard Eu2O3 powder as a r eference. The Eu L3-edge data obtained from both films resembles that from Eu2O3 with intense edge resonance at ~6981 eV. It possesses the characteristic of Eu3+, which has the descriptive ground state electronic configuration of [Xe]4f65d06s0. If Eu2+ appears inside the films, the additional edge resonance would be observed at ~6972 eV, 8eV below the Eu3+ resonance.

Zn L3-edge: ZnO

Eu L3-edge: Eu2O3 F.T. Intensity (a.u.) kw = 2

Eu L3-edge: Zn0.92Co0.04Eu0.04O

Eu L3-edge: Zn0.96Eu0.04O 0 1 2 3 4 5 6 R (Å)

Figure 8.7 The Fourier Transformation of k2χ(k) EXAFS spectra for

calculated Zn L3-edge (line) and measured Eu L3-edge (symbols)

for Eu2O3 powder, Zn0.96Eu0.04O and Zn0.96Co0.04Eu0.04O thin film as well as their EXAFS fitting (line)

110

Figure 8.8 compares the Fourier transformation magnitude with k-weight = 2 of the Zn

L3-edge and Eu L3-edge. Zn L3-edge for ZnO is calculated using experiment result in

Table 8.1 while Eu L3-edge for Eu2O3 powder, Zn0.96Eu0.04O and Zn0.96Co0.04Eu0.04O thin film are obtained from the measurement. Similar RDF from Eu L3-edge EXAFS spectra for both films and the calculated Zn L3-edge for ZnO reveals that the Eu ions incorporated into a Zn-atom position substitutionally rather than being aggregate or interstitial. The main peak in Eu L3-edge corresponds to single scattering between the nearest oxygen neighbours and the central Eu atom. The second main peak of Eu L3- edge corresponding to next Zn nearest neighbours is missing due to disordering by Eu dopants.

Zn0.92Co0.04Eu0.04O (Ε)

Zn0.96Eu0.04O Normalized Eu2O3

6920 6960 7000 7040 7080 Energy (eV)

Figure 8.8 Eu L3-edge XANES of Eu2O3 powder, Zn0.96Eu0.04O thin film and

Zn0.96Co0.04Eu0.04O thin film

111

Table 8.3 EXAFS fitting summary of Eu L3-edge for Zn0.96Eu0.04O thin film

and Zn0.96Co0.04Eu0.04O thin film

Zn0.96Eu0.04O Zn0.92Co0.04Eu0.04O Path N R (Å) σ2 (x10-3) N R (Å) σ2 (x10-3)

Eu-O1 4 2.309±0.008 5.2±1.4 4 2.301±0.008 6.2±1.2

Eu-Zn1 6 3.434±0.042 14.6±6.8 6 3.457±0.045 14.2±4.7

Eu-O3 1 3.207±0.063 5.3±1.2 1 3.222±0.070 6.3±1.3

Eu-Zn2 6 3.626±0.067 14.6±6.8 6 3.670±0.070 14.2±4.7

Eu-O1-O1 12 4.309±0.029 17.6±8.8 12 4.295±0.031 19.1±5.1

Eu-O1-Zn1 12 3.862±0.025 17.6±8.8 12 3.869±0.027 19.1±5.1

Eu-O1-Zn2 12 3.958±0.025 17.6±8.8 12 3.876±0.027 19.1±5.1

Eu-O3 9 3.818±0.028 7.9±3.1 9 3.818±0.028 7.9±3.1

Eu-Zn3 6 4.572±0.045 8.0±3.9 5 4.572±0.045 8.0±3.9

Eu-Co - - - 1 4.363±0.093 3.4±2.0

Eu-O4 6 4.568±0.034 7.9±3.1 6 4.568±0.034 7.9±3.1

The XANES is highly sensitive to the presence of transition clustering in host oxide.

For comparison, Figure 8.9 shows the Co K-edge XANES spectra of Zn0.96Co0.04O and

Zn0.92Co0.04Eu0.04O thin film as well as the standard cobalt metal and cobalt oxide compound. The Co K-edge spectra of both samples show 13sd→ pre-edge feature around 7709 eV. It is noticed that their pre-edge and energy edge are similar to those of CoO rather than metallic Co which show significant shoulder around 7712eV (as arrowed). However, the absorption edge of both films is broaden as the metallic Co. The XANES measurements, therefore, demonstrate that most of the Co atoms in both films are in the +2 oxidation state with some present of Co clustering. EXAFS is also adopted to clarify the local structure around Co atoms. Figure 8.10 plots the radial distribution function (RDF), the Fourier transformation amplitude of EXAFS for Zn0.96Co0.04O and

Zn0.96Co0.04O thin film at Co K-edge with k-weight of 2. The RDFs of Zn0.96Co0.04O reveals that Co ions are not properly substituted in the ZnO matrix. Some of them is in

112

form of Co clustering and the rest is substituted Zn2+. The first and second highest peaks represent Co-O of Co2+ substituted Zn2+ and Co-Co of Co clustering respectively. On the other hand, the RDFs of Zn0.92Co0.04Eu0.04O is similar to that of the Zn K-edge spectra shown in Figure 8.6, implying that the most of Co2+ ions have similar local structure to the Zn ions in host structure.

Co CoO Co3O4 LiCoO2 Zn0.96Co0.04O Zn0.92Co0.04Eu0.04O (Ε)

Normalized Normalized

7680 7700 7720 7740 7760 Energy (eV)

Figure 8.9: Co K-edge XANES spectra for Zn0.96Co0.04Eu0.04O thin film and

Zn0.96Co0.04O, in comparison to reference Co metal and oxide compound of cobalt

113

The EXAFS spectra suggest that the (most or part of) implanted ions do replace Zn ions in ZnO host matrix. The Co K-edge and Eu L3-edge EXAFS fitting were performed to study the local structure near Co and Eu, including the inter-atomic distance between the implanted ions and their neighbours. The central atom of ZnO wurtzite structure with a cell cluster of 77 atoms is placed by Co and Eu ion. The energy shift E0 of oxygen and zinc were control to exactly the same as Zn K-edge fitting. The fit results were listed in Table 8.3and Table 8.4

Zn0.92Co0.04Eu0.04O

Zn Co O F.T. Intensity (a.u.) kw = 2 0.96 0.04

0 1 2 3 4 5 6 R (Å)

Figure 8.10: The Fourier Transformation of k2χ(k) EXAFS spectra data

(symbols) and fitting (line) for Zn0.96Co0.04Eu0.04O and

Zn0.96Co0.04O thin films

114

Table 8.4: EXAFS fitting summary of Co K-edge for Zn0.96Co0.04O thin film

and Zn0.96Co0.04Eu0.04O thin film

Zn0.96Co0.04O Zn0.96Co0.04Eu0.04O Path N R (Å) σ2 (x10-3) N R (Å) σ2 (x10-3)

Co-O1 4 1.959±0.007 1.6±0.9 4 1.961±0.008 2.4±1.3

Co-Zn1 6 3.528±0.039 6.4±3.1 6 3.159±0.059 7.1±2.9

Co-O2 1 3.375±0.043 2.4±2.4 1 3.194±0.065 4.2±3.0

Co-Zn2 6 3.658±0.046 6.9±3.1 6 3.258±0.064 7.3±2.9

Co-O1-O1 12 3.656±0.027 8.1±4.2 12 3.541±0.038 8.5±4.3

Co-O1-Zn1 12 3.734±0.028 11.3±3.2 12 3.550±0.039 10.6±4.5

Co-O1-Zn2 12 3.799±0.030 12.7±4.2 12 3.590±0.040 10.7±4.7

Co-O3 9 3.804±0.043 7.3±2.1 9 3.811±0.048 5.2±2.7

Co clusters 0.47 0.08 faction

Co-Co 12 2.533±0.049 10.0±2.7 12 2.487±0.056 8.9±3.1

For Zn0.96Co0.04O, the EXAFS fitting suggests that is 43% of Co were found in form of Co clustering and the rest is substituted Zn2+ in ZnO matrix, while there is only 8% of

Co existence as clustering in Zn0.96Co0.04Eu0.04O. It evidences that Eu doping improve Co substitution in ZnO. By replacing Zn2+ with Eu3+, the distance to the first nearest oxygen expands from 1.980 Å to 2.309 Å while replacing by C o2+, the distance is shorten to 1.970 Å. The ion radius of Co2+ (0.74 Å) is similar to Zn2+ (0.72 Å) while Eu3+ is much larger. Therefore, the lattice structure was not affected when small quantity of Co2+ ions replacing Zn2+. On the other hand, replacing by Eu3+ causes large distortion in ZnO structure. For the Eu L3-edge EXFAS, RDF of Zn0.92Co0.04Eu0.04O is similar to that of Zn0.96Eu0.04O except the peak around 4 Å, as shown in Figure 8.7. This might cause by the Co2+ ion. The EXAFS fitting suggests that the Co2+ should replace Zn2+ at 4.363 Å from Eu3+ ion. It would be confirmed by Co K-edge fitting with remodel Zn0.92Co0.04Eu0.04O. Unfortunately, the Co K-edge (7709eV) is located between

115

Eu L3-edge (7617eV) and Eu L2-edge (8052eV). The measurement range of the beamline facility has reached to the limit and could not produce the reliable fitting for any distance longer than 4Å.

8.4 Summary

Our XAFS study has verified that the epitaxial ZnO (001) thin films has wurtzite phase structure with space groups P63mc. XANSE and EXAFS spectra at Zn K-edge for the implanted ZnO thin films suggest that their structural features were not affected by the ion implantation at this level. The inter-atomic distance from Zn2+ ions to the nearest O2- and next-to-nearest Zn2+ are almost identical to those in non-implanted ZnO thin film. The Co K-edge XANSE measurement reveals that oxidation state of Co ions is +2 while the Eu L3-edge XANSE shows that the oxidation state of Eu ions is +3 for both 2+ Zn0.96Eu0.04O and Zn0.96Co0.04Eu0.04O. The EXAFS data and fitting suggest that the Co 3+ 2+ and Eu substituted Zn in Zn0.96Co0.04O and Zn0.96Eu0.04O. Similarity found in

Zn0.96Co0.04Eu0.04O, EXAFS fitting suggest that both implanted ions spices substituted Zn2+ with distance 4.363Å from each other

116

Chapter 9 Conclusion

9.1 Summary

ZnO based DMSs have attracted attention over the last decade as a promising candidate for room temperature ferromagnetic semiconductor for the applications of spintronics devices. It is, however, the properties of present ZnO based DMSs have not yet meet the materials requirements for the applicable spin based devices: e.g. strong localized magnetic moment and high spin injection efficiency for spin field effect transistors etc. Therefore, part of current research effort has focused on enhancing the properties of existed materials and seeking new candidates. In this work, cobalt and europium codoped technique is proposed as an alternative method for realizing room temperature intrinsic ferromagnetism in ZnO based semiconductor.

Prior to fabricate Zn1-xCox-yEuyO thin films, the ions depth profile distribution of Co and Eu implanted ions were investigated by transport of ions in matter analysis (known as TRIM code) and Rutherford backscattering spectroscopy (RBS) in Chapter 4. T he

ZnO (0001) thin films were epitaxial growth on c-Al2O3 substrate by pulsed laser deposition (PLD). The Co and Eu ions were implanted into ZnO thin films using mix positive valence states by metal vacuum vapour arc (MEVVA) source. The studies show that the depth profile for implanted ions is closed to Gaussian distribution. As the energy of incident ions is rising, the range of implantation depth is widened and the stopping range of europium ions is shorter than cobalt ions with the same ion energy. This is because energy loss by electron stopping power and nuclear stopping power of europium ions is larger than those of cobalt. The total amount of implanted dose is controlled by current integrator. The results suggest that the suitable ions accelerated voltage for cobalt and europium should be 24 kV and 45 kV respectively using MEVVA ion implanter, resulting in similar implantation depth profile of both ions.

X-ray diffraction shows the crystallographic structure of Zn1-xEuxO thin films is wurtzite like phase, similar to the virgin ZnO with expansion of c-axis from 5.210 Å to 5.229 Å. There is no evidence of secondary phase formation including metallic clusters. The results suggest that Eu atom should substitute Zn in lattice structure with local

117

distortion. The room temperature photo luminescence spectra show that the absence of 380 nm emission band in ZnO:Eu. This might be associated with the decrease of band gaps energy (Eg). Eg of the virgin ZnO film is 3.296 eV and it is decreased dynamically when Eu concentration in ZnO:Eu is increased from 1 at% to 4 at%. The experimental result agrees well with the ground state band structure calculation using DFT in Eu substituted tetrahedral Zn sites of the ZnO wurtzite structure.

The room temperature ferromagnetic behaviors have been observed in ZnO:Eu thin films using SQUID magnetometer. The average magnetic moment per Eu atom is sharply decreased with Eu concentration above 2 a t% which suggests increase of paramagnetism. XAS spectra shows Eu atomic in Zn1-xEuxO are mostly trivalent state. Whereas, XMCD reveals that the localized magnetic moment of Eu atom is different. On surface region, Eu atoms have strong spin polarization while average Eu atoms of the entire film have weak opposite polarization. DFT calculation shows local structure disorder around the substitutional Eu atom to Zn lattice site with tetrahedral symmetry in Zn14Eu2O16 system, demonstrating that the anti-ferromagnetic coupling of Eu ions is preferable.

The structural and magnetic properties of Zn0.92Co0.04Eu0.04O thin films were prepared by ion implantation technique. Room temperature ferromagnetism was observed with a saturation magnetization of 65 emu/cm3, a coercivity of 60 Oe and a remnant 3 magnetization of 5.2 e mu/cm . XAS spectra of Co L2,3 edge shows that the Co ions reside in the tetrahedral coordination, indicating its substitution of Zn in the wurtzite structure. XMCD spectra of Co L2.3 edge also reveal that Co metal nano-clustering may not form in Co and Eu co-implanted ZnO as evidenced by Co tetrahedral multiplet structures. It suggests that Co substitution at the Zn sites may be facilitated by Eu codoping and thus the Co metal clustering is suppressed significantly.

The local structure around the specific atoms and their valence states in

Zn0.92Co0.04Eu0.04O, Zn0.96Co0.04O and Zn0.96Eu0.04O, are studied by X -ray absorption fine structure. The XANSE spectra verify that most of Co and Eu atom are in +2 and +3 oxidation state. The results confirmed every sample has wurtzite phase structure. In

Zn0.96Co0.04O, Co atoms are coexisting in both Co hcp metallic clustering and

118

substitutional Co2+ at the Zn2+ in tetrahedral symmetry coordination. In

Zn0.92Co0.04Eu0.04O, the fraction of Co clustering is significantly decreased from 0.47 to 0.08 by additional Eu implantation. EXAFS analysis shows low level of distortion for substitutional Co2+ while substitutional Eu3+ causes large level of distortion in ZnO structure. The Eu-O distance is 2.309 Å which is much larger than Zn-O, 1.980 Å . These are caused by t he difference of atomic radius among Zn, Co and Eu. EXAFS fitting also suggests localized coordination of Co2+ and Eu3+ forming long range order of 4.363 Å at position of the sixth nearest neighbour.

9.2 Future work

The current work might be further extended in two approaches. First, it is enhancement of spin polarization with optimizing the stoichiometric atomic ratio of Eu and Co in Zn(Co,Eu)O. The current state of this study shows that the additional Eu doping enhances Co substitution in ZnO-based DMSs with Eu substitution of Zn in tetrahedral symmetry. However, it has not yet investigated on concentration dependent effect, also electronic and magneto-transport properties on this material. Second, it is to optimize fabrication technique. In many circumstances, the dopant atoms in the host material were often observed in two (or more) different phases: e.g. coexistence of substitutional 2+ Co and metallic Co clusters in Zn1-xCoxO. To determine the localized moment of both Co2+ and the metallic Co individually is extremely difficult. In this case, substitutional Co2+ from two different systems could not be directly compared unless Co dopant atoms are only presented in single structure for both systems. Controlling of dopants’ distribution and defect concentration is another challenge in this approach.

119

9.3 Closing statement

In brief, the present work demonstrates that the rare earth doped ZnO based DMSs can be considered as a potential candidate for intrinsic ferromagnetic semiconductors due to the observation of strong spin polarization of localized Eu3+ ions near surface of thin films. Additional Eu implanting to ZnO:Co materials also shows significantly suppression of Co metallic clustering. X-ray spectroscopy study suggests substitutional Co2+ and substitutional Eu3+ are intrinsic ferromagnetic. This work provides a guidance to develop the novel ZnO based DMSs.

120

Reference

1. Deal, B.E. and J.M. Early, The Evolution of Silicon Semiconductor Technology: 1952--1977. Journal of The Electrochemical Society, 1979. 126(1): p. 20C-32C. 2. Moore, G.E., Cramming more components onto integrated circuits, Reprinted from Electronics, volume 38, number 8, April 19, 1965, pp. 114 ff. Solid-State Circuits Newsletter, IEEE, 2006. 20(3): p. 33-35. 3. Moore, G.E. Progress in digital integrated electronics. in Electron Devices Meeting, 1975 International. 1975. 4. Datta, S. and B. Das, Electronic analog of the electro-optic modulator. Applied Physics Letters, 1990. 56(7): p. 665-667. 5. Hong, N.H., V. Brize, and J. Sakai, Mn-doped ZnO and (Mn, Cu)-doped ZnO thin films: Does the Cu doping indeed play a key role in tuning the ferromagnetism? Applied Physics Letters, 2005. 86(8): p. 082505-3. 6. Han, S.J., et al., A key to room-temperature ferromagnetism in Fe-doped ZnO: Cu. Applied Physics Letters, 2002. 81(22): p. 4212-4214. 7. Johnson, M. and R.H. Silsbee, Interfacial charge-spin coupling: Injection and detection of spin magnetization in metals. Physical Review Letters, 1985. 55(17): p. 1790. 8. Baibich, M.N., et al., Giant Magnetoresistance of (001)Fe/(001)Cr Magnetic Superlattices. Physical Review Letters, 1988. 61(21): p. 2472. 9. Binasch, G., et al., Enhanced magnetoresistance in layered magnetic structures with antiferromagnetic interlayer exchange. Physical Review B, 1989. 39(7): p. 4828. 10. Bruno, P., Theory of interlayer exchange interactions in magnetic multilayers. Journal of Physics: Condensed Matter, 1999. 11(48): p. 9403. 11. Mott, N., The electrical conductivity of transition metals. Proceedings of the Royal Society of London. Series A, Mathematical and Physical Sciences, 1936. 153(880): p. 699-717. 12. Edwards, D.M., J. Mathon, and R.B. Muniz, A resistor network theory of the giant magnetoresistance in magnetic superlattices. Magnetics, IEEE Transactions on, 1991. 27(4): p. 3548-3552. 13. Zutic, I., J. Fabian, and S. Sarma, Spintronics: Fundamentals and applications. Arxiv preprint cond-mat/0405528, 2004. 14. Ohno, Y., et al., Electrical spin injection in a ferromagnetic semiconductor heterostructure. Nature, 1999. 402(6763): p. 790-792. 15. Özgür, Ü., et al., A comprehensive review of ZnO materials and devices. Journal of Applied Physics, 2005. 98: p. 041301. 16. Pearton, S.J. and et al., Dilute magnetic semiconducting oxides. Semiconductor Science and Technology, 2004. 19(10): p. R59.

121

17. Ohno, H., Making nonmagnetic semiconductors ferromagnetic. Science, 1998. 281(5379): p. 951. 18. Dietl, T., et al., Zener model description of ferromagnetism in zinc-blende magnetic semiconductors. Science, 2000. 287(5455): p. 1019. 19. Dietl, T. and H. Ohno, Ferromagnetic III–V and II–VI Semiconductors. Mrs Bulletin, 2003. 28: p. 714. 20. Pearton, S., et al., Wide band gap ferromagnetic semiconductors and oxides. Journal of Applied Physics, 2003. 93: p. 1. 21. Wolf, S., et al., Spintronics: A spin-based electronics vision for the future. Science, 2001. 294(5546): p. 1488. 22. Loss, D. and D.P. DiVincenzo, Quantum computation with quantum dots. Physical Review A, 1998. 57(1): p. 120. 23. Pearton, S.J., et al., Advances in wide bandgap materials for semiconductor spintronics. Materials Science and Engineering: R: Reports, 2003. 40(4): p. 137- 168. 24. Zener, C., Interaction Between the d Shells in the Transition Metals. Physical Review, 1951. 81(3): p. 440. 25. Beschoten, B., et al., Magnetic Circular Dichroism Studies of Carrier-Induced Ferromagnetism in (Ga1-xMnx)As. Physical Review Letters, 1999. 83(15): p. 3073.

26. Shioda, R., et al., Local structures of III-V diluted magnetic semiconductors Ga1- xMnxAs studied using extended x-ray-absorption fine structure. Physical Review B, 1998. 58(3): p. 1100. 27. Hayashi, T., et al., (GaMn)As: GaAs-based III-V diluted magnetic semiconductors grown by molecular beam epitaxy. Journal of Crystal Growth, 1997. 175-176(Part 2): p. 1063-1068. 28. Van Esch, A., et al., Interplay between the magnetic and transport properties in the III-V diluted magnetic semiconductor Ga1-xMnxAs. Physical Review B, 1997. 56(20): p. 13103-13112. 29. Tanaka, M. Epitaxial growth and properties of III--V magnetic semiconductor (GaMn)As and its heterostructures. 1998. Salt Lake City, Utah (USA): AVS. 30. Sadowski, J., et al., Structural and magnetic properties of GaMnAs layers with high Mn-content grown by migration-enhanced epitaxy on GaAs(100) substrates. Applied Physics Letters, 2001. 78(21): p. 3271-3273. 31. Shimizu, H., et al., Magnetic and transport properties of III--V based magnetic semiconductor (GaMn)As: Growth condition dependence. Applied Physics Letters, 1999. 74(3): p. 398-400. 32. Grandidier, B., et al., Atomic-scale study of GaMnAs/GaAs layers. Applied Physics Letters, 2000. 77(24): p. 4001-4003. 33. Kawakami, R.K., et al., (Ga,Mn)As as a digital ferromagnetic heterostructure. Applied Physics Letters, 2000. 77(15): p. 2379-2381.

122

34. Chiba, D., et al., Magnetoresistance effect and interlayer coupling of (Ga, Mn)As trilayer structures. Applied Physics Letters, 2000. 77(12): p. 1873-1875. 35. Ohno, H., et al., Spin-dependent tunneling and properties of ferromagnetic (Ga,Mn)As (invited). Journal of Applied Physics, 1999. 85(8): p. 4277-4282. 36. Hayashi, T., et al., Magnetic and magnetotransport properties of new III-V diluted magnetic semiconductors: GaMnAs. Journal of Applied Physics, 1997. 81(8): p. 4865-4867. 37. Hayashi, T., et al., III--V based magnetic(GaMnAs)/nonmagnetic(AlAs) semiconductor superlattices. Applied Physics Letters, 1997. 71(13): p. 1825- 1827. 38. Hayashi, T., M. Tanaka, and A. Asamitsu, Tunneling magnetoresistance of a GaMnAs-based double barrier ferromagnetic tunnel junction. Journal of Applied Physics, 2000. 87(9): p. 4673-4675. 39. Munekata, H., et al., Diluted magnetic III-V semiconductors. Physical Review Letters, 1989. 63(17): p. 1849. 40. Akai, H., Ferromagnetism and Its Stability in the Diluted Magnetic Semiconductor (In, Mn)As. Physical Review Letters, 1998. 81(14): p. 3002. 41. Munekata, H., et al., Preparation of (In,Mn)As/(Ga,Al)Sb magnetic semiconductor heterostructures and their ferromagnetic characteristics. Applied Physics Letters, 1993. 63(21): p. 2929-2931. 42. Koshihara, S., et al., Ferromagnetic Order Induced by Photogenerated Carriers in Magnetic III-V Semiconductor Heterostructures of (In,Mn)As/GaSb. Physical Review Letters, 1997. 78(24): p. 4617. 43. Soo, Y.L., et al., III-V diluted magnetic semiconductor: Substitutional doping of Mn in InAs. Physical Review B, 1996. 53(8): p. 4905. 44. Oiwa, A., T. Slupinski, and H. Munekata, Control of magnetization reversal process by light illumination in ferromagnetic semiconductor heterostructure p- (In, Mn)As/GaSb. Applied Physics Letters, 2001. 78(4): p. 518-520. 45. Furdyna, J.K., Diluted magnetic semiconductors. Journal of Applied Physics, 1988. 64(4): p. R29-R64. 46. Sonoda, S., et al., Molecular beam epitaxy of wurtzite (Ga,Mn)N films on sapphire(0001) showing the ferromagnetic behaviour at room temperature. Journal of Crystal Growth, 2002. 237-239(Part 2): p. 1358-1362. 47. Sasaki, T., et al., Magnetic and transport characteristics on high Curie temperature ferromagnet of Mn-doped GaN. Journal of Applied Physics, 2002. 91(10): p. 7911-7913. 48. Overberg, M.E., et al., Room temperature magnetism in GaMnP produced by both ion implantation and molecular-beam epitaxy. Journal of Vacuum Science & Technology B: Microelectronics and Nanometer Structures, 2002. 20(3): p. 969-973.

123

49. Medvedkin, G.A., et al., Room temperature ferromagnetism in novel diluted magnetic semiconductor Cd1- xMnxGeP2. Japanese Journal of Applied Physics, 2000. 39(Part 2, No. 10A): p. 3. 50. Medvedkin, G.A., et al., New magnetic materials in ZnGeP2-Mn chalcopyrite system. Journal of Crystal Growth, 2002. 236(4): p. 609-612. 51. Choi, S., et al., Room-temperature ferromagnetism in chalcopyrite Mn-doped ZnSnAs2 single crystals. Solid State Communications, 2002. 122(3-4): p. 165- 167. 52. Matsumoto, Y., et al., Room-Temperature Ferromagnetism in Transparent Transition Metal-Doped Titanium Dioxide. Science, 2001. 291(5505): p. 854- 856.

53. Stern, N.P., et al., Confinement engineering of s-d exchange interactions in Ga1- xMnxAs/ AlyGa1-yAs quantum wells. Physical Review B, 2007. 75(4): p. 045329. 54. Kramers, H.A., L'interaction Entre les Atomes Magnétogènes dans un Cristal Paramagnétique. Physica, 1934. 1(1-6): p. 182-192. 55. Anderson, P.W., Antiferromagnetism. Theory of Superexchange Interaction. Physical Review, 1950. 79(2): p. 350. 56. Shulman, R.G. and V. Jaccarino, Nuclear Magnetic Resonance in Paramagnetic MnF2. Physical Review, 1957. 108(5): p. 1219. 57. Wachter, P., Chapter 19 Europium chalcogenides: EuO, EuS, EuSe and EuTe, in Handbook on the Physics and Chemistry of Rare Earths, Karl A. Gschneidner, Jr. and E. LeRoy, Editors. 1979, Elsevier. p. 507-574. 58. Blazey, K.W., Ferromagnetic exchange coupling in the spinel lattice. Solid State Communications, 1966. 4(10): p. 541-544. 59. Lee, W.-L., et al., Dissipationless Anomalous Hall Current in the Ferromagnetic Spinel CuCr2Se4-xBrx. Science, 2004. 303(5664): p. 1647-1649. 60. Dietl, T., Spin Order Manipulations in Nanostructures of II—VI Ferromagnetic Semiconductors. ChemInform, 2004. 35(41): p. no-no. 61. Kanamori, J., Superexchange interaction and symmetry properties of electron orbitals. Journal of Physics and Chemistry of Solids, 1959. 10(2-3): p. 87-98. 62. Goodenough, J.B., Theory of the Role of Covalence in the Perovskite-Type Manganites [La, M(II)]MnO3. Physical Review, 1955. 100(2): p. 564. 63. Goodenough, J.B., An interpretation of the magnetic properties of the perovskite-type mixed crystals La1-xSrxCoO3-λ. Journal of Physics and Chemistry of Solids, 1958. 6(2-3): p. 287-297. 64. Zener, C., Interaction between the d-Shells in the Transition Metals. II. Ferromagnetic Compounds of Manganese with Perovskite Structure. Physical Review, 1951. 82(3): p. 403. 65. Roy, M., et al., Doping-induced transition from double exchange to charge order in La1-xCaxMnO3 near x=0.50. Physical Review B, 1998. 58(9): p. 5185.

124

66. Anane, A., et al., Transport properties and magnetic behaviour of La1-xSrxMnO3 single crystals. Journal of Physics: Condensed Matter, 1995. 7(35): p. 7015. 67. de Gennes, P.G., Effects of Double Exchange in Magnetic Crystals. Physical Review, 1960. 118(1): p. 141. 68. Bechlars, B., et al., High-spin ground states via electron delocalization in mixed-valence imidazolate-bridged divanadium complexes. Nat Chem, 2010. 2(5): p. 362-368. 69. Wang, Z., et al., Ferromagnetism and transport properties of Fe-doped reduced- rutile TiO2 - thin films. Journal of Applied Physics, 2003. 93(10): p. 7870-7872. 70. Coey, J., M. Venkatesan, and C. Fitzgerald, Donor impurity band exchange in dilute ferromagnetic oxides. Nature materials, 2005. 4(2): p. 173-179. 71. Kaminski, A. and S. Das Sarma, Polaron percolation in diluted magnetic semiconductors. Physical Review Letters, 2002. 88(24): p. 247202. 72. Coey, J.M.D., d0 ferromagnetism. Solid State Sciences, 2005. 7(6): p. 660-667. 73. Bouzerar, G. and T. Ziman, Model for Vacancy-Induced d Ferromagnetism in Oxide Compounds. Physical Review Letters, 2006. 96(20): p. 207602. 74. Esquinazi, P., et al., Ferromagnetism in oriented graphite samples. Physical Review B, 2002. 66(2): p. 024429. 75. Volnianska, O. and P. Boguslawski, Magnetism of solids resulting from spin polarization of p orbitals. Journal of Physics: Condensed Matter, 2010. 22(7): p. 073202. 76. Wu, H., et al., Magnetism in C- or N-doped MgO and ZnO: A Density- Functional Study of Impurity Pairs. Physical Review Letters, 2010. 105(26): p. 267203. 77. Zheng, W.T. and C.Q. Sun, Electronic process of nitriding: Mechanism and applications. Progress in Solid State Chemistry, 2006. 34(1): p. 1-20. 78. Shen, L., et al., Mechanism of ferromagnetism in nitrogen-doped ZnO: First- principle calculations. Physical Review B, 2008. 78(7): p. 073306. 79. Pearton, S.J., et al., Recent progress in processing and properties of ZnO. Progress in Materials Science, 2005. 50(3): p. 293-340. 80. Van de Walle, C.G., Defect analysis and engineering in ZnO. Physica B: Condensed Matter, 2001. 308-310: p. 899-903. 81. Sato, K. and H. Katayama-Yoshida, Material design for transparent ferromagnets with ZnO-based magnetic semiconductors. JAPANESE JOURNAL OF APPLIED PHYSICS PART 2 LETTERS, 2000. 39(6B): p. 555- 558. 82. Fukumura, T., et al., Magnetic properties of Mn-doped ZnO. Applied Physics Letters, 2001. 78(7): p. 958-960. 83. Özgür, Ü., et al., Thermal conductivity of bulk ZnO after different thermal treatments. Journal of Electronic Materials, 2006. 35(4): p. 550-555.

125

84. Wakano, T., et al., Magnetic and magneto-transport properties of ZnO:Ni films. Physica E: Low-dimensional Systems and Nanostructures, 2001. 10(1-3): p. 260-264.

85. Kim, Y.M., et al., Synthesis and magnetic properties of Zn1-xMnxO films prepared by the sol-gel method. Solid State Communications, 2004. 129(3): p. 175-178. 86. Sharma, P., et al., Ferromagnetism above room temperature in bulk and transparent thin films of Mn-doped ZnO. Nature materials, 2003. 2(10): p. 673- 677. 87. Ueda, K., H. Tabata, and T. Kawai, Magnetic and electric properties of transition-metal-doped ZnO films. Applied Physics Letters, 2001. 79: p. 988. 88. Jin, Z., et al., High throughput fabrication of transition-metal-doped epitaxial ZnO thin films: A series of oxide-diluted magnetic semiconductors and their properties. Applied Physics Letters, 2001. 78(24): p. 3824-3826. 89. Kim, J., et al., Magnetic properties of epitaxially grown semiconducting ZnCoO thin films by pulsed laser deposition. Journal of Applied Physics, 2002. 92: p. 6066. 90. Venkatesan, M., et al., Anisotropic Ferromagnetism in Substituted Zinc Oxide. Physical Review Letters, 2004. 93(17): p. 177206. 91. Yan, S.-S., et al., Ferromagnetism and magnetoresistance of Co-ZnO inhomogeneous magnetic semiconductors. Applied Physics Letters, 2004. 84(13): p. 2376-2378. 92. Rode, K., et al., Magnetic semiconductors based on cobalt substituted ZnO. Journal of Applied Physics, 2003. 93: p. 7676. 93. Antony, J., et al., Room temperature ferromagnetic and ultraviolet optical properties of Co-doped ZnO nanocluster films. Journal of Applied Physics, 2005. 97(10): p. 10D307-3. 94. Lee, H.-J., et al., Study of diluted magnetic semiconductor: Co-doped ZnO. Applied Physics Letters, 2002. 81(21): p. 4020-4022. 95. Tuan, A.C., et al., Epitaxial growth and properties of cobalt-doped ZnO on α- Al2 O3 single-crystal substrates. Physical Review B, 2004. 70(5): p. 054424. 96. Song, C., et al., Giant magnetic moment in an anomalous ferromagnetic insulator: Co-doped ZnO. Physical Review B, 2006. 73(2): p. 24405. 97. Jin, Z., et al., Combinatorial laser MBE synthesis of 3d ion doped epitaxial ZnO thin films. Journal of Crystal Growth, 2000. 214-215: p. 55-58. 98. Kolesnik, S., B. Dabrowski, and J. Mais, Structural and magnetic properties of transition metal substituted ZnO. Journal of Applied Physics, 2004. 95(5): p. 2582-2586. 99. Lawes, G., et al., Absence of ferromagnetism in Co and Mn substituted polycrystalline ZnO. Physical Review B, 2005. 71(4): p. 045201.

126

100. Fukumura, T., et al., An oxide-diluted magnetic semiconductor: Mn-doped ZnO. Applied Physics Letters, 1999. 75(21): p. 3366-3368.

101. Jung, S.W., et al., Ferromagnetic properties of Zn1-xMnxO epitaxial thin films. Applied Physics Letters, 2002. 80(24): p. 4561-4563. 102. Norton, D., et al., Ferromagnetism in cobalt-implanted ZnO. Applied Physics Letters, 2003. 83: p. 5488. 103. Park, M.S. and B.I. Min, Ferromagnetism in ZnO codoped with transition metals: Zn1-x(FeCo)xO and Zn1-x(FeCu)xO. Physical Review B, 2003. 68(22): p. 224436. 104. Han, S.J., et al., Magnetism in Mn-doped ZnO bulk samples prepared by solid state reaction. Applied Physics Letters, 2003. 83(5): p. 920-922. 105. Cheng, X.M. and C.L. Chien, Magnetic properties of epitaxial Mn-doped ZnO thin films. Journal of Applied Physics, 2003. 93(10): p. 7876-7878. 106. Iwata, K., et al., Nitrogen-induced defects in ZnO : N grown on sapphire substrate by gas source MBE. Journal of Crystal Growth, 2000. 209(2-3): p. 526-531. 107. Look, D.C., et al., Characterization of homoepitaxial p-type ZnO grown by molecular beam epitaxy. Applied Physics Letters, 2002. 81(10): p. 1830-1832. 108. Li, X., et al. Chemical vapor deposition-formed p-type ZnO thin films. 2003. Denver, Colorado (USA): AVS. 109. Guo, X.-L., H. Tabata, and T. Kawai, Epitaxial growth and optoelectronic properties of nitrogen-doped ZnO films on Al2O3 substrate. Journal of Crystal Growth, 2002. 237-239(Part 1): p. 544-547. 110. Ye, Z.-Z., et al., Preparation and characteristics of p-type ZnO films by DC reactive magnetron sputtering. Journal of Crystal Growth, 2003. 253(1-4): p. 258-264. 111. Nakahara, K., et al., Interactions between gallium and nitrogen dopants in ZnO films grown by radical-source molecular-beam epitaxy. Applied Physics Letters, 2001. 79(25): p. 4139-4141. 112. Ohshima, T., et al., Synthesis of p-type ZnO thin films using co-doping techniques based on KrF excimer laser deposition. Thin Solid Films, 2003. 435(1-2): p. 49-55. 113. Bian, J.M., et al., Deposition and electrical properties of N - In codoped p-type ZnO films by ultrasonic spray pyrolysis. Applied Physics Letters, 2004. 84(4): p. 541-543. 114. Ye, Z.-Z., et al., Preparation of p-type ZnO films by Al+N-codoping method. Journal of Crystal Growth, 2004. 265(1-2): p. 127-132. 115. Norton, D.P., et al., Ferromagnetism in Mn-implanted ZnO:Sn single crystals. Applied Physics Letters, 2003. 82(2): p. 239-241. 116. Dhar, S., et al., Colossal Magnetic Moment of Gd in GaN. Physical Review Letters, 2005. 94(3): p. 037205.

127

117. Çetin, A., et al., Optical properties of Tb implantation into ZnO. Surface and Coatings Technology, 2007. 201(19-20): p. 8534-8538. 118. Cetin, A., et al., Luminescence properties of Tb implanted ZnO. Physica B: Condensed Matter, 2009. 404(20): p. 3379-3385. 119. Armelao, L., et al., X-ray Excited Optical Luminescence Studies of ZnO and Eu- Doped ZnO Nanostructures. The Journal of Physical Chemistry C, 2007. 111(28): p. 10194-10200. 120. Gao, S., et al., Engineering white light-emitting Eu-doped ZnO urchins by biopolymer-assisted hydrothermal method. Applied Physics Letters, 2006. 89(12): p. 123125-3. 121. Ishizumi, A. and Y. Kanemitsu, Structural and luminescence properties of Eu- doped ZnO nanorods fabricated by a microemulsion method. Applied Physics Letters, 2005. 86(25): p. 253106-3. 122. Ishizumi, A., et al., Luminescence properties of ZnO and Eu3+-doped ZnO nanorods. Thin Solid Films, 2005. 486(1-2): p. 50-52. 123. Monteiro, T., et al., Optical active centres in ZnO samples. Journal of Non- Crystalline Solids, 2006. 352(9-20): p. 1453-1456. 124. Choi, S.W., et al., Magnetic, optical and electrical properties of GaN and AlN doped with rare-earth element Gd. physica status solidi (c), 2006. 3(6): p. 2250- 2253. 125. Lin, W., et al., Structural, electrical and optical properties of Gd doped and undoped ZnO:Al (ZAO) thin films prepared by RF magnetron sputtering. Applied Surface Science, 2007. 253(11): p. 5179-5183. 126. Ney, V., et al., Structural, magnetic, and optical properties of Co- and Gd- implanted ZnO(0001) substrates. Journal of Applied Physics, 2008. 104(8): p. 083904-6. 127. Potzger, K., et al., Ferromagnetic Gd-implanted ZnO single crystals. Journal of Applied Physics, 2006. 99: p. 063906. 128. Koningsberger, D.C., et al., XAFS spectroscopy; fundamental principles and data analysis. Topics in Catalysis, 2000. 10(3): p. 143-155. 129. Lee, P.A. and J.B. Pendry, Theory of the extended x-ray absorption fine structure. Physical Review B, 1975. 11(8): p. 2795. 130. Sayers, D.E., E.A. Stern, and F.W. Lytle, New Technique for Investigating Noncrystalline Structures: Fourier Analysis of the Extended X-Ray: Absorption Fine Structure. Physical Review Letters, 1971. 27(18): p. 1204. 131. Teo, B.K., EXAFS: Basic principles and data analysis1986: Springer-Verlag Berlin. 132. Gurman, S.J. and R.F. Pettifer, EXAFS study of the structure of amorphous and crystalline arsenic oxide. Philosophical Magazine Part B, 1979. 40(5): p. 345- 359.

128

133. Koningsberger, D. and R. Prins, X-ray absorption: principles, applications, techniques of EXAFS, SEXAFS and XANES. 1987. 134. Rehr, J.J. and R.C. Albers, Theoretical approaches to X-ray absorption fine structure. Reviews of Modern Physics, 2000. 72(3): p. 621. 135. Rehr, J.J. and R.C. Albers, Scattering-matrix formulation of curved-wave multiple-scattering theory: Application to x-ray-absorption fine structure. Physical Review B, 1990. 41(12): p. 8139. 136. Zabinsky, S.I., et al., Multiple-scattering calculations of X-ray-absorption spectra. Physical Review B, 1995. 52(4): p. 2995. 137. Newville, M., IFEFFIT : interactive XAFS analysis and FEFF fitting. Journal of Synchrotron Radiation, 2001. 8(2): p. 322-324. 138. Ravel, B. and M. Newville, ATHENA, ARTEMIS, HEPHAESTUS: data analysis for X-ray absorption spectroscopy using IFEFFIT. Journal of Synchrotron Radiation, 2005. 12(4): p. 537-541. 139. Ravel, B. EXAFS Analysis with FEFF and FEFFIT. 2000; Available from: http://cars9.uchicago.edu/~ravel/talks/course/. 140. Bevington, P.R. and D.K. Robinson, Data reduction and error analysis for the physical sciences. Vol. 336. 1969: McGraw-Hill New York. 141. Press, W.H., et al., Numerical recipes in Fortran; the art of scientific computing1993: Cambridge university press. 142. Stöhr, J. and Y. Wu, New directions in research with third-generation soft x-ray synchrotron radiation sources, ed. A.S. Schlachter and F. Wuilleumier1994: Springer. 143. Stöhr, J., X-ray magnetic circular dichroism spectroscopy of transition metal thin films. Journal of Electron Spectroscopy and Related Phenomena, 1995. 75: p. 253-272. 144. Stöhr, J. and H. König, Determination of Spin- and Orbital-Moment Anisotropies in Transition Metals by Angle-Dependent X-Ray Magnetic Circular Dichroism. Physical Review Letters, 1995. 75(20): p. 3748. 145. Stöhr, J., et al., Magnetism: from fundamentals to nanoscale dynamics2006: Springer Berlin, Germany. 146. Stöhr, J., Exploring the microscopic origin of magnetic anisotropies with X-ray magnetic circular dichroism (XMCD) spectroscopy. Journal of Magnetism and Magnetic Materials, 1999. 200(1-3): p. 470-497. 147. Schlachter, A.S. and F. Wuilleumier, New directions in research with third- generation soft x-ray synchrotron radiation sources1994: Springer. 148. Carra, P., et al., X-ray circular dichroism and local magnetic fields. Physical Review Letters, 1993. 70(5): p. 694. 149. Thole, B.T., et al., X-ray circular dichroism as a probe of orbital magnetization. Physical Review Letters, 1992. 68(12): p. 1943.

129

150. Yan, W., et al., Determination of the role of O vacancy in Co:ZnO magnetic film. Journal of Applied Physics, 2010. 108(1): p. 013901-4. 151. Rode, K., et al., Magnetism of (Zn,Co)O thin films probed by X-ray absorption spectroscopies. Applied Physics Letters, 2008. 92(1): p. 012509-3. 152. Park, J.H., et al., Co-metal clustering as the origin of ferromagnetism in Co- doped ZnO thin films. Applied Physics Letters, 2004. 84(8): p. 1338-1340. 153. Wei, H., et al., Role of Co clusters in wurtzite Co:ZnO dilute magnetic semiconductor thin films. Journal of Applied Physics, 2009. 105(4): p. 043903- 6.

154. Shi, T., et al., Structures and magnetic properties of wurtzite Zn1-xCoxO dilute magnetic semiconductor nanocomposites. Applied Physics Letters, 2007. 90(10): p. 102108-3. 155. Kaspar, T.C., et al., Hidden ferromagnetic secondary phases in cobalt-doped ZnO epitaxial thin films. Physical Review B, 2008. 77(20): p. 201303. 156. Heald, S.M., et al., X-ray absorption fine structure and magnetization characterization of the metallic Co component in Co-doped ZnO thin films. Physical Review B, 2009. 79(7): p. 075202. 157. Lu, Z., et al., Carrier-mediated ferromagnetism in single crystalline (Co, Ga)- codoped ZnO films. Applied Physics Letters, 2009. 94(15): p. 152507-3. 158. Lu, Z.L., et al., Enhanced ferromagnetism in single crystalline Co-doped ZnO thin films by Al codoping. Journal of Alloys and Compounds, 2010. 494(1-2): p. 392-395. 159. Lu, Z.L., et al., The origins of ferromagnetism in Co-doped ZnO single crystalline films: From bound magnetic polaron to free carrier-mediated exchange interaction. Applied Physics Letters, 2009. 95(10): p. 102501-3. 160. Guo, J.H. and et al., X-ray spectroscopic study of the charge state and local ordering of room-temperature ferromagnetic Mn-doped ZnO. Journal of Physics: Condensed Matter, 2007. 19(17): p. 172202. 161. Saeki, H. and et al., Transparent magnetic semiconductors based on ZnO. Journal of Physics: Condensed Matter, 2004. 16(48): p. S5533. 162. Hu, F., et al., Cu and Co codoping effects on room-temperature ferromagnetism of (Co,Cu):ZnO dilute magnetic semiconductors. Journal of Applied Physics, 2011. 109(10): p. 103705-4. 163. Ney, A. and et al., Advanced spectroscopic synchrotron techniques to unravel the intrinsic properties of dilute magnetic oxides: the case of Co:ZnO. New Journal of Physics, 2010. 12(1): p. 013020. 164. Xu, Q. and et al., Paramagnetism in Co-doped ZnO films. Journal of Physics D: Applied Physics, 2009. 42(8): p. 085001.

165. Barla, A., et al., Paramagnetism of the Co sublattice in ferromagnetic Zn1-xCoxO films. Physical Review B, 2007. 76(12): p. 125201.

130

166. Thomas, T. and et al., XMCD studies on Co and Li doped ZnO magnetic semiconductors. New Journal of Physics, 2008. 10(5): p. 055009. 167. Ney, A., et al., Absence of Intrinsic Ferromagnetic Interactions of Isolated and Paired Co Dopant Atoms in Zn1-xCoxO with High Structural Perfection. Physical Review Letters, 2008. 100(15): p. 157201. 168. Singh, A.P. and et al., NEXAFS and XMCD studies of single-phase Co doped ZnO thin films. Journal of Physics: Condensed Matter, 2009. 21(18): p. 185005. 169. Tauc, J., Amorphous and liquid semiconductors1974, New York: Plenum. 441. 170. Davis, E.A. and N.F. Mott, Conduction in non-crystalline systems V. Conductivity, optical absorption and photoconductivity in amorphous semiconductors. Philosophical Magazine, 1970. 22(179): p. 903-922. 171. Stöhr, J., NEXAFS spectroscopy1992: Springer. 172. Anders, A. Plasma fluctuations, local partial Saha equilibrium, and the broadening of vacuum-arc ion charge state distributions. in 18th International Symposium on Discharges and Electrical Insulation in Vacuum. 1998. Proceedings ISDEIV. . 173. Anders, A., Atomic scale heating in cathodic arc plasma deposition. Applied Physics Letters, 2002. 80(6): p. 1100-1102. 174. Anders, A., Ion charge state distributions of vacuum arc plasmas: The origin of species. Physical Review E, 1997. 55(1): p. 969. 175. Anders, A. and T. Schulke. Predicting ion charge state distributions of vacuum arc plasmas. in 17th International Symposium on Discharges and Electrical Insulation in Vacuum. 1996. Proceedings. ISDEIV. 176. Tsukazaki, A., et al., Repeated temperature modulation epitaxy for p-type doping and light-emitting diode based on ZnO. Nat Mater, 2005. 4(1): p. 42-46. 177. Gopal, P. and N. Spaldin, Polarization, piezoelectric constants, and elastic constants of ZnO, MgO, and CdO. Journal of Electronic Materials, 2006. 35(4): p. 538-542. 178. Matsukura, F., et al., Transport properties and origin of ferromagnetism in (Ga,Mn)As. Physical Review B, 1998. 57(4): p. R2037. 179. Monteiro, T., et al., Optical and structural analysis of bulk ZnO samples undoped and rare earth doped by ion implantation. Superlattices and Microstructures, 2006. 39(1-4): p. 202-210. 180. Yang, Y., et al., Morphology–luminescence correlations in europium-doped ZnO nanomaterials. Journal of Nanoparticle Research, 2010. 12(1): p. 217-225. 181. Risbud, A.S., et al., Magnetism in polycrystalline cobalt-substituted zinc oxide. Physical Review B, 2003. 68(20): p. 205202. 182. Martin, J.W., et al., Materials characterisation using heavy ion elastic recoil time of flight spectrometry. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms, 1994. 94(3): p. 277-290.

131

183. Ziegler, J., J. Biersack, and M. Ziegler, The Stopping and Ranges of Ions in Matter. Vol. 4. 1997, Elmsford, New York: Pergamon Press. 184. Feng, Y., et al., Photoluminescence spectra of nano-structured ZnO thin films. Journal of Luminescence, 2006. 119-120: p. 233-236. 185. Pauporte, T. and T. Yoshida, Hybrid layers of ZnO/lanthanide complexes with high visible luminescences. Journal of Materials Chemistry, 2006. 16(46): p. 4529-4534. 186. Fuchs, F., et al., Quasiparticle band structure based on a generalized Kohn- Sham scheme. Physical Review B, 2007. 76(11): p. 115109. 187. Sato, K. and H. Katayama-Yoshida, First principles materials design for semiconductor spintronics. Semiconductor Science and Technology, 2002. 17(4): p. 367. 188. Pan, F., et al., Ferromagnetism and possible application in spintronics of transition-metal-doped ZnO films. Materials Science and Engineering: R: Reports, 2008. 62(1): p. 1-35. 189. Segall, M.D. and et al., First-principles simulation: ideas, illustrations and the CASTEP code. Journal of Physics: Condensed Matter, 2002. 14(11): p. 2717. 190. Binder, K. and A.P. Young, Spin glasses: Experimental facts, theoretical concepts, and open questions. Reviews of Modern Physics, 1986. 58(4): p. 801. 191. Mydosh, J.A. and I. ebrary, Spin glasses: an experimental introduction1993: Taylor & Francis. 192. Pan, H., et al., Room-Temperature Ferromagnetism in Carbon-Doped ZnO. Physical Review Letters, 2007. 99(12): p. 127201. 193. Herng, T., et al., Stable ferromagnetism in p-type carbon-doped ZnO nanoneedles. Applied Physics Letters, 2009. 95: p. 133103. 194. Herng, T., et al., Magnetotransport properties of p-type carbon-doped ZnO thin films. Applied Physics Letters, 2009. 95: p. 012505. 195. Chang-Feng, Y. and et al., Origin of ferromagnetism in nitrogen embedded ZnO : N thin films. Journal of Physics D: Applied Physics, 2007. 40(21): p. 6497. 196. Evans, S., et al., Further characterization of oxygen vacancies and zinc vacancies in electron-irradiated ZnO. Journal of Applied Physics, 2008. 103: p. 043710. 197. Banerjee, S., et al., Enhancement of ferromagnetism upon thermal annealing in pure ZnO. Applied Physics Letters, 2007. 91: p. 182501. 198. Thole, B.T., et al., 3d x-ray-absorption lines and the 3d94fn+1 multiplets of the lanthanides. Physical Review B, 1985. 32(8): p. 5107. 199. Kim, J., et al., Ferromagnetism Induced by Clustered Co in Co-Doped Anatase TiO2 Thin Films. Physical Review Letters, 2003. 90(1): p. 17401. 200. Reed, M.L., et al., Room temperature ferromagnetic properties of (Ga, Mn)N. Applied Physics Letters, 2001. 79(21): p. 3473-3475.

132

201. Samarth, N., Ferromagnetic semiconductors: Ruled by a magnetic-rich minority. Nat Mater, 2007. 6(6): p. 403-404. 202. Evans, P.J., G.C. Watt, and J.T. Noorman, Metal vapor vacuum arc ion source research at Ansto. Review of Scientific Instruments, 1994. 65(10): p. 3082- 3087. 203. Ziegler, J.F., SRIM-2003. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms, 2004. 219-220: p. 1027-1036. 204. Choi, H., et al., Valencies of Eu in CaS:Eu, La. Journal of Solid State Chemistry, 1998. 138(1): p. 149-153. 205. Kobayashi, M., et al., Characterization of magnetic components in the diluted magnetic semiconductor Zn1-xCoxO by x-ray magnetic circular dichroism. Physical Review B, 2005. 72(20): p. 201201. 206. Chen, C.T., et al., Experimental Confirmation of the X-Ray Magnetic Circular Dichroism Sum Rules for Iron and Cobalt. Physical Review Letters, 1995. 75(1): p. 152. 207. Stern, E.A., Theory of the extended X-ray-absorption fine structure. Physical Review B, 1974. 10(8): p. 3027. 208. Chowdhury, R., S. Adhikari, and P. Rees, Optical properties of silicon doped ZnO. Physica B: Condensed Matter, 2010. 405(23): p. 4763-4767.

133