electrical and structural properties of high-κ barium .../67531/metadc84233/m2/1/high_res… ·...
TRANSCRIPT
APPROVED: Nigel Shepherd, Major Professor Rick Reidy, Committee Member Jincheng Du, Committee Member Mohammed El Bouanani, Committee Member Thomas Scharf, Committee Member Narendra Dahotre, Chair of the Department of
Materials Science and Engineering Costas Tsatsoulis, Dean of the College of
Engineering James D. Meernik, Acting Dean of the
Toulouse Graduate School
ELECTRICAL AND STRUCTURAL PROPERTIES OF HIGH-κ BARIUM TANTALITE
AND ALUMINUM OXIDE INTERFACES WITH ZINC OXIDE FOR APPLICATIONS
IN TRANSPARENT THIN FILM TRANSISTORS
Fang-Ling Kuo, B.S., M.S.
Dissertation Prepared for the Degree of
DOCTOR OF PHILOSOPHY
UNIVERSITY OF NORTH TEXAS
August 2011
Kuo, Fang-Ling. Electrical and structural properties of high-κ barium tantalite and
aluminum oxide interfaces with zinc oxide for applications in transparent thin film transistors.
Doctor of Philosophy (Materials Science and Engineering), August 2011, 149 pp., 15 tables, 88
figures, references, 155 titles.
ZnO has generated interest for flexible electronics/optoelectronic applications including
transparent thin film transistors (TFTs). For this application, low temperature processes that
simultaneously yield good electrical conductivity and optical transparency and that are
compatible with flexible substrates such as plastic, are of paramount significance. Further, gate
oxides are a critical component of TFTs, and must exhibit low leakage currents and self-healing
breakdown in order to ensure optimal TFTs switching performance and reliability. Thus, the
objective of this work was twofold: (1) develop an understanding of the processing-structure-
property relationships of ZnO and high-κ BaTa2O6 and Al2O3 (2) understand the electronic
defect structure of BaTa2O6 /ZnO and Al2O3/ZnO interfaces and develop insight to how such
interfaces may impact the switching characteristics (speed and switching power) of TFTs
featuring these materials.
Of the ZnO films grown by atomic layer deposition (ALD), pulsed laser deposition
(PLD) and magnetron sputtering at 100-200ºC, the latter method exhibited the best combination
of n-type electrical conductivity and optical transparency. These determinations were made using
a combination of photoluminescence, photoluminescence excitation, absorption edge and Hall
measurements. Metal-insulator-semiconductor devices were then fabricated with sputtered ZnO
and high-κ BaTa2O6 and Al2O3 and the interfaces of high-κ BaTa2O6 and Al2O3 with ZnO were
analyzed using frequency dependent C-V and G-V measurements. The insulator films were
deposited at room temperature by magnetron sputtering using optimized processing conditions.
Although the Al2O3 films exhibited a lower breakdown strength and catastrophic
breakdown behavior compared to BaTa2O6/ZnO interface, the Al2O3/ZnO interface was
characterized by more than an order of magnitude smaller density of interface traps and interface
trapped charge. The BaTa2O6 films in addition were characterized by a significantly higher
concentration of fixed oxide charge. The transition from accumulation to inversion in the Al2O3
MIS structure was considerably sharper, and occurred at less than one tenth of the voltage
required for the same transition in the BaTa2O6 case. The frequency dispersion effects were also
noticeably more severe in the BaTa2O6 structures. XPS results suggest that acceptor-like
structural defects associated with oxygen vacancies in the non-stoichiometric BaTa2O6 films are
responsible for the extensive electrical trapping and poor high frequency response. The Al2O3
films were essentially stoichiometric. The results indicate that amorphous Al2O3 is better suited
than BaTa2O6 as a gate oxide for transparent thin film transistor applications where low
temperature processing is a prerequisite, assuming of course that the operation voltage of such
devices is lower than the breakdown voltage. Also, the operation power for the devices with
amorphous Al2O3 is lower than the case for devices with BaTa2O6 due to the smaller fixed oxide
charges and interface trap density.
ii
Copyright 2011
by
Fang-Ling Kuo
iii
ACKNOWLEDGEMENTS
I want to gratefully and sincerely thank my advisor, Dr. Nigel Shepherd for his invaluable
guidance for my graduate study at University of North Texas. He gave me ample freedom to
express my ideas and encouraged me not only to grow as an experimentalist but also as an
instructor and independent thinker.
I am thankful to have other four professors, Dr. Rick Reidy, Dr. Mohammed El Bouanani,
Dr. Jincheng Du and Dr. Thomas Scharf as my committee members and also thanks for their
help when I had questions for my studying.
I would like to thank my colleagues, Ming-Te Lin, Mohammad E. Maneshian, Minghang
Li, for their help, understanding, encouragement and friendship. I also would like to thank my
classmates, Eric Osei-Yisdom and Benedict Mensah. They are so friendly and helpful during my
graduate study. It is so great thankful for the help from Craig Collins who is an Engineer Lab
technician and has excellent technology knowledge, great patient and understanding. I would
like to thank the great help from Nancy Bunce who is a research scientist in Center for Advanced
Research and Technology. Thanks for all of professors, stuff and students in Material Science &
Engineering Department
Also, I would like to thank my good friends, Yu Tsen Cheng, Wan. Ju Huang, Daniel Pan
and Ryan Chan for their encouragements and caring over the time of my studying here. I am
thankful that Huntern Su who always encouraged me to come to USA for my PhD degree. Last
but not the least; I would like to thank my family who give me endless love and support. I
cannot complete this work and get this degree without their great support.
iv
TABLE OF CONTENTS
ACKNOWLEDGMENTS ............................................................................................................. iii
LIST OF TABLES ....................................................................................................................... viii
LIST OF FIGURES ....................................................................................................................... ix CHAPTERS
1. INTRODUCTION .......................................................................................................................1
1.1 Motivation ............................................................................................................................1
1.2 Contributions of the Dissertation .........................................................................................3
1.3 Organization of the Dissertation ..........................................................................................5
1.4 References ............................................................................................................................5
2. LITERATURE REVIEW ...........................................................................................................8
2.1 Transparent Thin Film Transistors and Oxide Semiconductors ..........................................8
2.1.1 Transparent Thin Film Transistors Overview .............................................................8
2.1.2 Transparent Conducting Oxide Semiconductor Overview .......................................12
2.1.3 Introduction to Zinc Oxide........................................................................................12
2.1.4 High- κ Materials for Gate Oxides............................................................................18
2.1.5 Introduction to BaTa2O6 and Al2O3 ..........................................................................21
2.2 Metal-Insulator-Semiconductor Structures ........................................................................26
2.2.1 The Importance of Metal-Insulator-Semiconductor Structures to Transistor Technology ...............................................................................................................26
2.2.2 Factors Influencing MIS Performance and Characteristics: The Influence of Interfacial and Bulk Defects on Switching Speed and Power Required for Switching...................................................................................................................................30
2.2.3 The Current Status of MIS Technology ...................................................................33
2.3 Electrical Characterization Techniques .............................................................................34
v
2.3.1 Capacitance-Voltage (C-V) .....................................................................................35
2.3.2 Conductance-Voltage (G-V) ....................................................................................39
2.3.3 Frequency Dependent C-V and G-V Methods..........................................................43
2.4 References ..........................................................................................................................44
3. EXPERIMENTAL PROCEDURES ..........................................................................................52
3.1 Introduction ........................................................................................................................52
3.2 Substrate Preparation .........................................................................................................54
3.2.1 Procedures for Substrate Cleaning ............................................................................54
3.3 Semiconductor Thin Films Growth ...................................................................................54
3.3.1 Pulsed Laser Deposition ...........................................................................................55
3.3.2 Atomic Layer Deposition ......................................................................................... 56
3.3.3 RF Magnetron Sputter Deposition ............................................................................58
3.3.4 Post Deposition Annealing Treatments ....................................................................59
3.4 Insulator Thin Film Growth ...............................................................................................60
3.5 Metal-Insulator-Semiconductor Device Fabrication..........................................................61
3.6 Materials Characterization .................................................................................................64
3.6.1 Surface Profilometry .................................................................................................64
3.6.2 Scanning Electron Microscopy ................................................................................65
3.6.3 Structural Properties Using X-ray Diffraction .........................................................65
3.6.4 X-ray Photoelectron Spectrometer (XPS) ................................................................66
3.6.5 Optical Properties - Photoluminescence (PL), Photoluminescence Excitation (PLE) and Transmittance .....................................................................................................66
3.6.6 Optical Properties - Ellipsometry..............................................................................68
3.6.7 Electrical Properties - Hall Effect Measurements and Current-Voltage (I-V) ........69
3.7 Characterization of MIS Devices .......................................................................................73
3.7.1 Capacitance-Voltage (C-V) Measurements ..............................................................73
3.7.2 Conductance-Voltage (G-V) Measurements.............................................................74
vi
3.7.3 Hysteresis Curve from C-V Measurements ..............................................................74
3.7.4 Frequency Dispersion from C-V Measurements ......................................................74
3.8 References ..........................................................................................................................75
4. RESULTS: THE PROPERTIES OF ZINC OXIDE, BARIUM TANTALATE AND ALUMINUM OXIDE ..............................................................................................................76
4.1 Properties of ZnO Thin Films by PLD, ALD and Sputtering ............................................76
4.1.1Structural Properties and Compositional Analysis ....................................................76
4.1.2 Electrical Properties ..................................................................................................88
4.1.3 Optical Properties......................................................................................................91
4.2 Properties of BaTa2O6 and Al2O3 ..................................................................................107
4.2.1 Structure of BaTa2O6 and Al2O3 .............................................................................107
4.2.2 Electrical Properties of BaTa2O6 and Al2O3 ...........................................................109
4.2.3 Optical Properties of BaTa2O6 and Al2O3 ...............................................................111
4.3 Conclusions: Structure and Electro-optical Properties of ZnO, BaTa2O6 and Al2O3 Thin Films ...............................................................................................................................117
4.4 References ........................................................................................................................118
5. RESULTS: QUANTIFICATION OF TRAP STATES IN THE HIGH-κ BULK AND AT INTERFACES WITH ZINC OXIDE ......................................................................................123
5.1 Fabrication of MIS Structures Employing High-k BaTa2O6 and Al2O3 .........................123
5.2 Quantitative Determination of Trap States in MIS Structures Featuring High-k BaTa2O6 and Al2O3 .........................................................................................................................123
5.2.1 C-V Measurements .................................................................................................124
5.2.2 G-V Measurements .................................................................................................126
5.2.3 C-V Measurements by Voltage Sweeps and Hysteresis .........................................131
5.2.4 Frequency Dependent C-V Measurements ............................................................134
5.2.5 XPS with Depth Profiling for Composition Determination ....................................136
5.2.6 Scanning Electron Microscopy ...............................................................................143
vii
5.3 Conclusions: Interfacial and Bulk Traps, Their Structural Models and The Implications for Transparent Thin Film Transistors Based on ZnO and High-κ BaTa2O6 and Al2O3 144
5.4 References ........................................................................................................................145
6. CONCLUSIONS AND FUTURE WORK .............................................................................144
6.1 Conclusions ......................................................................................................................144
6.2 Future Work .....................................................................................................................145
6.3 References ........................................................................................................................147
viii
LIST OF TABLES
Table Page
2-1 A characteristic comparison among different TFTs and their corresponding applications .......9
2-2 The calculated energies for the defect formation in ZnO. The Kroger-Vink defect notation is used throughout ......................................................................................................................15
2-3 Basic conduction process in insulators ....................................................................................18
3-1 Deposition conditions for ALD ZnO thin film growth............................................................58
4-1 Gain sizes of PLD ZnO as-deposited at 100 oC and after annealing in Ar, vacuum and air ...82
4-2 The Zn/O ratio determined by XPS for PLD ZnO films in the as-deposited condition, and after annealing in Ar, vacuum and air ....................................................................................82
4-3 Gain sizes of ALD ZnO as-deposited at 100 oC and annealed in Ar, vacuum and air ............85
4-4 The Zn/O ratio determined by XPS for ALD ZnO films in the as-deposited condition, and after annealing in Ar, vacuum and air ....................................................................................84
4-5 Gain sizes of sputtered ZnO as-deposited at 100 oC and annealed in Ar, vacuum and air at 400 oC .....................................................................................................................................89
4-6 The Zn/O ratio determined by XPS for sputtered ZnO films in the as-deposited condition, and after annealing in Ar, vacuum and air ...............................................................................89
4-7 Electrical properties of PLD, ALD and sputtered ZnO thin films grown at 100 oC and 200 oC by Hall Effect measurement ...................................................................................................91
4-8 Electrical properties of PLD, ALD and sputtered ZnO as-deposited and annealed ZnO thin films by Hall Effect measurement ..........................................................................................93
5-1 The interface traps density and trapped charge concentration of MIS structures with BaTa2O6 and Al2O3 ...............................................................................................................................134
5-2 The atomic concentration and ratio determined by XPS for Ba and Ta ................................140
5-3 The atomic concentration and ratio determined by XPS for Al and O ..................................143
ix
LIST OF FIGURES
Figure Page
2-1 (a) Schematic illustration of the central elements of TFT operation (b) the diagram of ID and VDS at different gate voltage ...................................................................................................11
2-2 The hexagonal wurtzite structure of ZnO. O atoms are shown as large spheres and Zn atoms are smaller ones........................................................................................................................13
2-3 The band structure of wurtzite ZnO ........................................................................................14
2-4 PL spectra obtained from as-grown samples prepared under the conditions of oxygen pressure of 50, 200, 300, 400 and 500 mTorr at the substrate temperature of 400oC .............17
2-5 Calculated defect’s level in ZnO films ....................................................................................17
2-6 Structure of ZnO thin film transistor with high-κ HfO2 ..........................................................21
2-7 Capacitor leakage currents dependences on annealing temperature for HfO2 dielectric ........21
2-8 Dielectric constants and loss factors of amorphous-BaTa2O6 with the different substrate temperature ..............................................................................................................................23
2-9 Characteristics for leakage current of amorphous-BaTa2O6 thin films with different substrate temperature ..............................................................................................................................23
2-10 Current density verse electric field strength of Al2O3 films prepared with 10% O2 of Ar/O2 mixture gas ............................................................................................................................24
2-11 X-ray diffraction patterns of the Al2O3 films deposited with different O2 contents in working gas of Ar/O2 ............................................................................................................25
2-12 Current density verse applied field of Al2O3 films with various oxygen contents in a Ar/O2 working gas mixture .............................................................................................................26
2-13 A simple Metal-Insulator-Semiconductor (MIS) capacitor ...................................................28
2-14 The energy band diagrams without bias for (a) MIS with n-type semiconductor and (b) MIS with p-type semiconductor .....................................................................................................29
2-15 Energy band diagrams of (a) Accumulation case for MIS with p-type semiconductor (top) and MIS with n-type semiconductor (bottom) (b) depletion case for MIS with p-type semiconductor (top) and MIS with n-type semiconductor (bottom) (c) Inversion case for MIS with p-type semiconductor (top) and MIS with n-type semiconductor (bottom) ........29
2-16 Terminology for the charge associated with MIS devices ....................................................33
2-17 Typical C-V measurement for (a) an ideal MIS with p-type semiconductor (b) lateral shift C-V curve (c) distorted C-V curve........................................................................................35
2-18 The simple circuit for MIS with p-type semiconductor ........................................................36
2-19 High-frequency (1MHz) capacitance-voltage and conductance-voltage characteristics of MIS capacitors using Ta2O5 films deposited on ZnO layers ................................................37
2-20 The inverted surface charge on p-type semiconductor surface and the simple circuit for the oxide capacitance and the maximum depletion layer capacitance in series .........................38
x
2-21 The hysteresis curves with a voltage gap of 2.8V for capacitance-voltage characteristics of the sample .............................................................................................................................39
2-22 (a) An equivalent circuit including interface trap effects, Cit and Rit (b) An equivalent circuit with a frequency –dependent capacitance, Cp and frequency-dependent conductance, Gp ..41
2-23 Conductance-voltage characteristics of SiONC/(p)Si MOS structure measured at 1 MHz ..42
2-24 The measured capacitance at various frequency for Au/SiO2/n-Si MIS ...............................43
3-1 The sheet resistivity of ZnO film as a function of substrate temperature ................................53
3-2 The layout for pulsed laser deposition. The distance between the target and the substrate is 5 cm ...........................................................................................................................................56
3-3 An image of the Savannah 100 ALD system (Source: Cambridge Nanotech Inc) ................58
3-4 A sketch diagram of sputtering chamber .................................................................................59
3-5 A sketch for annealing curve ...................................................................................................60
3-6 ( a) A mask on ITO glass substrate (b) ZnO thin film on ITO glass substrate (c) mask on ZnO thin film for insulator deposition ............................................................................................62
3-7 (a) a mask on the insulator/ZnO/ITO glass (b) a top view ......................................................63
3-8 The boat for the Ag deposition ................................................................................................63
3-9 The schematic deposition of the thermal evaporation for silver layer on samples ..................64
3-10 A schematic structure of MIS ................................................................................................64
3-11A schematic architecture of thin films with steps and the measuring arrangement with a stylus of the profilometer .......................................................................................................65
3-12 A schematic diagram of photoluminescence measurement ...................................................67
3-13 PL spectra at room temperature of ZnO thin films grown under different oxygen partial pressures. The emission wavelength was around 370 nm ....................................................67
3-14 A schematic setup of an ellipsometry experiment .................................................................69
3-15 (a) The common geometry for contacts for Hall effect measurements. The labels 1-4 are the contacts. (b) Schematic of current and voltage measurement...............................................71
3-16 A I-V measurement for BaTa2O6 thin films by sputtering deposition ...................................72
3-17 A typical C-V measurement of MIS with n-type semiconductor with the bias swept from inversion region to accumulation region ................................................................................73
4-1 X-ray diffraction spectra of PLD ZnO grown at 100 oC (bottom), and 200 oC (top). The peaks are labeled in brackets and their 2 θ angles given below .............................................79
4-2 X-ray diffraction spectra of PLD ZnO grown at 100 oC with laser power density of 1.04 (black), 1.24 (red), 1.54 (blue) and 1.94 (green) mJ/cm2. The peaks are labeled in brackets and their 2 θ angles given below ..............................................................................................80
4-3 XRD spectra of PLD ZnO as-deposited at 100oC (black) and annealed in Ar (red), vacuum (blue) and air (green) at 400 oC for 2 hours .............................................................................81
xi
4-4 X-ray diffraction spectra of ALD ZnO grown at 100 oC (bottom), and 200 oC (top). The peaks are labeled in brackets and their 2 θ angles given below ...............................................84
4-5 XRD spectra of ALD ZnO as-deposited at 100oC (black) and annealed in Ar (red), vacuum (blue) and air (green) at 400 oC for 2 hours .............................................................................84
4-6 X-ray diffraction spectra of sputtered ZnO grown at 100 oC (bottom), and 200 oC (top). The peaks are labeled in brackets and their 2 θ angles given below ...............................................86
4-7 X-ray diffraction spectra of sputtered ZnO grown at 100 oC with sputtering power density of 4.44 (black), 5.92 (red), 7.40 (blue) and 8.88 (green) W/cm2. The peaks are labeled in brackets and their 2 θ angles given below ...............................................................................87
4-8 XRD spectra of sputtered ZnO as-deposited at 100oC (black) and annealed in Ar (red), vacuum (blue) and air (green) at 400 oC for 2 hours ...............................................................88
4-9 The photoluminescence (solid) and photoluminescence excitation (dashed) spectra of the samples synthesized by PLD at 100 oC (top) and 200 oC (bottom) ........................................95
4-10 The transmittance of PLD ZnO grown at 100 oC (black) and 200 oC (red) ..........................95
4-11 The photoluminescence (solid) and photoluminescence excitation (dashed) spectra of the samples synthesized by PLD at 100 oC (black) annealed in three ambient of Ar (red), vacuum (green) and air (blue) ...............................................................................................97
4-12 The transmittance of the samples synthesized by PLD at 100 oC (black) annealed in three ambient of Ar (red), vacuum (green) and air (blue) ..............................................................97
4-13 The photoluminescence (solid) and photoluminescence excitation (dashed) spectra of the samples synthesized by ALD at 100 oC (top) and 200 oC (bottom) .....................................99
4-14 The transmittance of ALD ZnO grown at 100 oC (black) and 200 oC (red) .........................99
4-15 The photoluminescence (solid) and photoluminescence excitation (dashed) spectra of the samples synthesized by ALD at 100 oC (black) annealed in three ambient of Ar (red), vacuum (green) and air (blue) ............................................................................................102
4-16 The transmittance of the samples synthesized by ALD at 100 oC (black) annealed in three ambient of Ar (red), vacuum (green color) and air (blue) ..................................................102
4-17 The photoluminescence (solid) and photoluminescence excitation (dashed) spectra of the samples synthesized by sputtering at 100 oC (top) and 200 oC (bottom) ............................103
4-18 The transmittance of sputtered ZnO grown at 100 oC (black) and 200 oC (red) .................104
4-19 The photoluminescence (solid) and photoluminescence excitation (dashed) spectra of the samples synthesized by sputtering at 100 oC (black) annealed in three ambient of Ar (red), vacuum (green) and air (blue) .............................................................................................105
4-20 The transmittance of the samples synthesized by sputtering at 100 oC (black) annealed in three ambient of Ar (red), vacuum (green) and air (blue) ...................................................106
xii
4-21 (a) The experimental curve (red) and Cauchy model fitting curves (black) for PLD ZnO at 1.04 mJ/cm2 at 100oC ..........................................................................................................107
4-21 (b) The experimental curve (red) and Cauchy model fitting curves (black) for ALD ZnO at 100oC...................................................................................................................................108
4-21 (c) The experimental curve (red) and Cauchy model fitting curves (black) for ALD ZnO at 100oC...................................................................................................................................108
4-22 The refractive indexes as function of wavelength for PLD ZnO, ALD ZnO and sputtered ZnO .....................................................................................................................................109
4-23 X-ray diffraction patterns of BaTa2O6 (bottom) and Al2O3 (top) thin films deposited with RF magnetron sputtering power densities of 6.91 W/cm2 and 9.87 W/cm2, respectively ..110
4-24.Current density vs. electric field strength of BaTa2O6 films sputtered with powers of power: 4.93 W/cm2, 5.92 W/cm2, 6.91 W/cm2, 7.89 W/cm2 and 8.88 W/cm2 ..............................112
4-25. Current density vs. electric field strength of Al2O3 films sputtered with powers of powers of 5.92 W/cm2, 6.91 W/cm2, 7.89 W/cm2, 8.88 W/cm2 and 9.87 W/cm2 ..........................113
4-26 The transmittance of BaTa2O6 with sputtering power: 6.91 W/cm2 (blue) and 7.89 W/cm2 (green) ................................................................................................................................114
4-27 The transmittance of Al2O3 with sputtering power: 8.88 W/cm2 (purple pink) and 9.87 W/cm2 (black) .....................................................................................................................114
4-28 (a) The experimental curve (red) and Lorentz model fit (black) for BaTa2O6 grown at 6.91 W/cm2 .................................................................................................................................116
4-28 (b) The experimental curve (red) and Lorentz model fit (black) for BaTa2O6 grown at 7.89 W/cm2 .................................................................................................................................116
4-29 The refractive index and extinction coefficient as a function of wavelength for BaTa2O6 films grown 6.91 W/cm2 and 7.89W/cm2 ...........................................................................117
4-30 (a) The experimental curve (red) and Lorentz model fit (black) for Al2O3 grown at 8.88 W/cm2 .................................................................................................................................117
4-30 (b) The experimental curve (red curves) and Lorentz model fit (black) for Al2O3 grown at 9.87 W/cm2 .........................................................................................................................118
4-31 The refractive index and extinction coefficient as a function of wavelength for Al2O3 films grown at 8.88 W/cm2 and 9.87W/cm2 ................................................................................118
5-1 C-V at 1MHz of Ag/BaTa2O6/ZnO MIS diodes. The insulator was sputtered at 6.91W/cm2 (red) and 7.89W/cm2 (black) .................................................................................................125
5-2 C-V at 1MHz for Ag/ Al2O3/ZnO MIS diodes. The Al2O3 was sputtered at 8.88W/cm2 (red) and 9.87W/cm2 (black) .........................................................................................................126
5-3 G-V at 1MHz for Ag/BaTa2O6/ZnO MIS diodes. The insulator was sputtered at 6.91W/cm2 (red) and 7.89W/cm2 (black) ................................................................................................128
5-4 G-V characteristics at 1 MHz for Ag/Al2O3/ZnO MIS diodes. The Al2O3 was sputtered at 8.88W/cm2 (red) and 9.87W/cm2 (black) ............................................................................128
xiii
5-5 (a) The capacitance as a function of frequency (b) G/ω vs frequency plots at various bias from 0.1V to 0.9V ................................................................................................................130
5-6 C-V plots of the BaTa2O6 MIS structure grown at 6.91 W/cm2 (red curve) and 7.89 W/cm2 (black curve). Strong hysteresis is observed in sweeps from accumulation through depletion to inversion. The arrows indicate the sweep direction .........................................................132
5-7 C-V plots of the Al2O3MIS grown at 8.88 W/cm2 (red curve) and 7.89 W/cm2 (black curve) in sweeps from accumulation through depletion to inversion. The arrows indicate the sweep direction .................................................................................................................................133
5-8 The overlay of C-V measurements at 500 kHz, 1 MHz and 2 MHz for BaTa2O6 MIS structure grown at 6.91W/cm2. The inset figure shows the normalized capacitance for the device .....................................................................................................................................135
5-9 The overlay of C-V measurements at 500 kHz, 1 MHz and 2 MHz for Al2O3 MIS grown at 9.87W/cm2. The inset figure shows the normalized capacitance for the device ...................136
5-10 (a) XPS Spectra of O(1s) during the depth profile for different etching levels (time) (b) XPS Spectra of Zn (2p3/2) during the depth profile for different etching levels (time) (c) XPS Spectra of Ba (3d5/2) during the depth profile for different etching levels (time) .............137
5-11 XPS Spectra of Ta 4f7/2 and 4f5/2 during the depth profile for different etching levels .......139
5-12 The atomic concentration of tantalum, barium, and oxygen are essentially constant throughout the thickness ....................................................................................................140
5-13 (a) XPS Spectra of O(1s) during the depth profile for different etching levels (time) (b) XPS Spectra of Zn(2p3/2) during the depth profile for different etching levels (time) (c) XPS Spectra of Al(2p) during the depth profile for different etching levels (time) ...........141
5-14 The atomic concentration of tantalum, barium and oxygen were essentially constant throughout the thickness .....................................................................................................142
5-15 Cross-sectional SEM images for (a)Ag/BaTa2O6(6.91W/cm2 )/ ZnO/ITO glass (b) Ag/BaTa2O6 (7.89W/cm2 )/ ZnO/ITO glass (c) Ag/Al2O3(8.88W/cm2)/ZnO/ITO glass (d) Ag/ Al2O3(9.87W/cm2) /ZnO/ITO glass .............................................................................143
6-1 (a) a top view for the Transparent TFT (b) a side view for the transparent TFT ..................148
1
CHAPTER 1
INTRODUCTION
1.1 Motivation
Transparent electronic devices are an emerging technology and have attracted significant
attention for next generation optoelectronic devices. Currently, amorphous and polycrystalline
Si, are commonly used for thin film transistors (TFTs). However, Si suffers from drawbacks
such as light sensitivity and low field effect mobility [1]. Recently, semiconducting zinc oxide
has attracted significant interest for TFT applications because it combines high electrical
conductivity with high optical transparency in the visible range. The interest in ZnO extends to
its potential utility for a variety of applications including piezoelectric devices, transparent
electrodes in photovoltaic cells, photodetectors, thin film gas sensors, spintronic devices and
solid lubricant thin films [2-8]. The author of this dissertation is particularly interested in the
potential of ZnO for flexible electronics/optoelectronic applications including transparent
conducting windows for flexible solar cells, transparent thin film transistors and thin film
electroluminescent devices on flexible substrates [9]. For these applications it is desirable that
incident or generated light not to be dampened or absorbed by the window material, and
therefore understanding the various absorption and emission bands as a function of processing is
important. In this regard, low temperature processes that simultaneously yield good electrical
conductivity and optical transparency and that are compatible with flexible substrates such as
plastic, are of paramount significance.
The high concentration of free carriers in ZnO semiconductor is attributed to point
defects including oxygen vacancies, zinc interstitials, and hydrogen, which introduce shallow
donor levels below the conduction band edge. However, the specific types and concentration of
2
defects and their influence on electrical properties are strongly dependent on the growth methods
and parameters. Growth temperature is particularly important because it strongly impacts the
thermodynamics of the growth process, and further, it limits the types of substrates which are
compatible with the growth temperature. Thus, while semiconducting ZnO has been grown by
various techniques including molecular beam epitaxy (MBE) [10], metal-organic chemical vapor
deposition (MOCVD) [11], sputtering [12-14], atomic layer deposition (ALD) [15,16] and
pulsed laser deposition (PLD) [ 17 - 19 ], not all of these methods satisfy the simultaneous
requirements for low temperature processing, good electrical conductivity and high optical
transparency. Pulsed laser deposition, atomic layer and sputtering deposition are relatively
simple, inexpensive and environmentally benign thin film deposition processes. In this work the
structural, electrical and optical properties of PLD films, ALD films, and sputtering films were
studied, and the connection between point defects, electrical behavior and photoluminescence
mechanisms is discussed. In general, low temperature processes and materials are required to
advance the emerging areas of transparent and flexible electronics and optoelectronics.
With the perpetual miniaturization of devices, conventional SiO2 dielectrics are becoming
inadequate, especially for thinner SiO2 layers, where the leakage currents become unacceptably
large and thus limit the performance of devices. Gate oxides are a critical component of TFTs,
and must exhibit low leakage currents and self-healing breakdown in order to ensure optimal
TFTs switching performance and reliability. Therefore, the motivation for this work was
identifying high-κ materials that are compatible with ZnO films processed at low temperatures.
Due to the higher permittivity of high-κ materials, these gate oxides can be thinner while
simultaneously exhibiting required smaller leakage currents.
3
Overall, this work involves two main focuses: (1) to develop an understanding of the
processing-structure-property relationships of ZnO and high-κ material BaTa2O6 and Al2O3 (2) to
understand the electronic defect structure of BaTa2O6 /ZnO and Al2O3/ZnO interfaces and
develop insight to how such interfaces may impact the switching characteristics (speed and
switching power) of TFTs featuring these materials.
1.2 Contributions of the Dissertation
For the semiconducting ZnO there are two major contributions:
(1) This study indicates that the optical and electrical characteristics of the ZnO films
grown by ALD appear to be determined by the relative concentration of zinc interstitials and
oxygen vacancies, whereas Zn interstitials appear to be primarily responsible for the electrical
conductivity and photoluminescence of PLD and magnetron sputtered ZnO films. The
transmissivity of all of the films was in the 75%-90% range.
(2) ZnO films could be fabricated by PLD, ALD and sputtering at a substrate temperature
of only 100oC. For the PLD, ALD and sputtered ZnO films the best free electron concentrations
obtained were 1.11x1017 cm-3, 1.81x1016 cm-3, and 3.59x1018 cm-3, respectively. The
corresponding resistivities were 3.87 Ω.cm, 303 Ω.cm and 0.05 Ω.cm, and the mobilities were
14.6 cm2/V.s, 1.14 cm2/V.s and 35.21 cm2/V.s, respectively. The sputtering ZnO films grown at
100 oC exhibited significantly better electrical properties compared to PLD and ALD ZnO films
grown at the same temperature.
In summary, ALD, PLD and sputtering produced ZnO thin films with excellent
transmissivity and electrical properties at low deposition temperatures, which are compatible
4
with flexible, plastic substrates such as polyethylene naphthalate (PEN), and flexible
optoelectronic applications. Mechanisms that explain the measured electrical behavior and
photoluminescent properties are proposed.
For the study of high-κ gate oxides on ZnO the main contributions are as follows:
(1) Ag/BaTa2O6/ZnO/ITO and Ag/Al2O3/ZnO/ITO MIS structures were sputter
grown between room temperature and 100 oC, which are compatible with flexible plastic
substrates. The interfaces of high-κ BaTa2O6 and Al2O3 with ZnO were analyzed using
frequency dependent C-V (capacitance-voltage) and G-V (conductance-voltage) measurements
and the density of interface traps determined. Although the Al2O3 films exhibited a lower
breakdown strength and catastrophic breakdown behavior compared to BaTa2O6/ZnO interface,
the Al2O3/ZnO interface exhibited more than an order of magnitude smaller density of interface
traps and interface trapped charge. The frequency dispersion effects are also noticeably more
severe in the BaTa2O6 structures. A significantly higher voltage was required to transition the
BaTa2O6/ZnO structure from accumulation to depletion which indicates that higher switching
power would be required in these structures compared to the Al2O3/ZnO devices.
(2) A chemical analysis using XPS analysis suggests that acceptor-like structural
defects associated with oxygen vacancies in the non-stoichiometric BaTa2O6 films are
responsible for the extensive electrical trapping and poor high frequency response. The Al2O3
films were essentially stoichiometric, which is postulated as the structural reason for the
comparatively better electrical behavior observed in these structures.
In summary, the results indicate that amorphous Al2O3 is better suited than BaTa2O6 as a
gate oxide for transparent thin film transistor applications where low temperature processing is a
5
prerequisite, assuming that the operation voltage of such devices is lower than the breakdown
voltage.
1.3 Organization of the Dissertation
There are six chapters in this dissertation, which are organized as follows:
Chapter 1 is an Introduction and summarizes the contributions of this whole work. Chapter 2
presents the Background related to transparent electronic devices, a Literature Review and also
the Theories of MIS Structures. In Chapter 3 the Instrumentation used and Experimental
Procedures are discussed. Chapter 4 discusses the obtained Properties of Zinc Oxide, Barium
Tantalate and Aluminum Oxide, including their structural, electrical and optical properties.
Chapter 5 reports the results of the electrical investigation of BaTa2O6/ZnO and Al2O3/ZnO
interfaces using MIS structures and capacitance-voltage and conductance-voltage methods.
Finally, Chapter 6 presents Conclusions and Future Work.
1.4 References
[1] Y. Sun and J. A. Rogers, Adv. Mater. 19, 1897 (2007)
[2] S. H. Yoon and D. Kim, Mater. Res. Soc. Symp. Proc. Vol. 966, 0966 T07-07 (2007)
[3] L. Ion, C. Tazlaoanu, G. Socol, L. Magherusan, I. Enculescu, D. Bazavan, I. Mihailescu and
S. Antohe, Mater. Res. Soc. Symp. Proc. Vol. 1013, 1013Z04-09 (2007)
6
[4] A. Ievtushenkoa, G. Lashkareva, V. Lazorenkoa, V. Karpynaa, V. Sichkovskyia, L.
Kosyachenkob, V. Sklyarchukb, O. Sklyarchukb, V. Bosyc, F. Korzhinskic, A. Ulyashind, V.
Khranovskyye, and R. Yakimovae, Acta Physicapolonica A, Vol. 114, 1123-1129 (2008)
[5] P. P. Sahay, J. Mater. Sci. 40, 4383-4385 (2005).
[6] D. P. Norton, Y. Heo, L. C. Tien, M. P. Ivill, Y. Li, B. S. Kang, F. Ren, J. Kelly, A. F.
Hebard, and S. Pearton, Mater. Res. Soc. Symp. Proc. Vol. 829, B8.5.1-12 (2005)
[7] J. S. Zabinski, J. Corneille, S. V. Prasad, N. T. Mcdevitt, and J. B. Bultman, Synthesis,
Characterization and Tribological Behavior, J. Mater. Sci. 32, 5313-5319 (1997)
[8] J.J. Nainaparampil, and J.S. Zabinski, J. Mater. Res. Vol. 16, 3423-3429 (2001)
[9] P. F. Carcia, R. S. McLean, M.H. Reilly, I. Malajovich, K.G. Sharp, S. Agrawal, and G.
Nunes, Mater. Res. Soc. Symp. Proc. Vol. 769, H7.2.1-6 (2003)
[10] A. El-Shaer, A. C. Mofor, A. Bakin, M. Kreye, and A. Waag, Superlatt. Microstruct. 38,
265-271 (2005)
[11] X. Li, S.E. Asher, B.M. Keyes, H.R. Moutinho, and J. Luther, T.J. Coutts, NREL/CP-520-
37378 (2005)
[12] J.B. Leea, H. J. Kimb, S. G. Kimc, C. S. Hwangb, S. Hongb, Y. H. Shinc, and N. H. Lee,
Thin Solid Films 435, 179-185 (2003)
[13] S. Flickyngerova, K. Shtereva, V. Stenova, D. Hasko, I. Novotny, V.Tvarozek, P. Sutta, and
E. Vavrinsky, Appl. Surf. Sci. 254, 3643-3647 (2008)
7
[14] S.J. Lim, Soonju Kwon, and H. Kim, Thin Solid Films 516, 1523-1528 (2008)
[15] S. Kwon, S. Bang, S. Lee, S. Jeon, W. Jeong, Hyungchul Kim1, S. C. Gong, H. J. Chang, H.
Park, and H. Jeon, Semi-cond. Sci. Technol. 24, 035015, 6pp (2009)
[16] J. W. Elam, G. Xiong, C.Y. Han, H. H. Wang, J. P. Birrell, J. N. Hryn, M. J. Pellin J. F.
Poco and J. H. Stcher, Mater. Res. Soc. Symp. Proc. Vol. 876E, R12.6.1-5 (2005)
[17] M. Oh, S. Kim, S. Park, T. Seong,, Superlattices Microstruct. 39, 130–137 (2006)
[18] H. W. Lee, Y. G. Wang, S. P. Lau and B. K. Tay, Mat. Res. Soc. Symp. Proc. Vol. 763
(2003)
[19] C. Hirose, Y. Matsumoto, Y. Yamamoto, and H. Koinuma, Appl. Phys. A 79, 807-809
(2004)
8
CHAPTER 2
LITERATURE REVIEW
2.1 Transparent Thin Film Transistors and Oxide Semiconductors
This section provides an overview of transparent thin film transistors (TFTs) and oxide
semiconductors. The applications of transparent TFTs and materials for fabrication of MIS are
discussed here as well. An oxide semiconductor material, zinc oxide and gate oxide materials,
BaTa2O6 and Al2O3 are particularly discussed due to their extensive use in this work.
2.1.1 Transparent Thin Film Transistors Overview
Si-based materials have been commonly used for most conventional TFTs in display
technology. Silicon TFTs can be fabricated using either crystalline silicon and poly-silicon, or
amorphous silicon. Table 2-1 shows a characteristic comparison of different TFTs and their
corresponding applications [1]. Single crystalline silicon exhibits the highest mobility due to its
crystal structure. However, its high processing temperature limits the types of substrates that are
compatible with this material. Amorphous silicon and has the lowest mobility and is processed
at the lowest temperature. Poly-silicon can be formed in a range of temperatures that are
compatible with glass and quartz substrates. However, the low temperature end is around 500oC,
which is not suitable for flexible plastic substrates. The processing temperature of amorphous
silicon is only 300oC and it is widely used in LCD monitors. The primary application of TFTs is
in liquid crystal displays. However a-Si-based TFTs are semi-opaque and restrict the amount of
light that can be transmitted to the observer. The need to increase the amount of out coupled
light from LCDs has driven the interest in transparent TFTs for years [2-6]. As mentioned
earlier, both a-si TFTs and p-si TFTs require a relatively high processing temperature of between
9
300oC~500 oC [1,7], which limits the selection of compatible substrates. Transparent transistor
technology and associated circuits are of great interest for applications including flexible
displays, radio-frequency identification tags (RFID) and electronic paper [8], which all use
plastic materials as substrates. Therefore, the need for low temperature processing that is
compatible with plastic substrates is of significant technological importance.
Table 2-1 A characteristic comparison among different TFTs and their corresponding applications
Switching Devices Mobility (cm2/V sec)
Process temperature Major application
a-Si TFT 0.3-1 ~300 oC (glass) Laptop screens, PC monitors, LCD TVs
High-T poly-Si TFT 100-300 ~1000 oC (quartz)
Projection light valves, view finders
Low-T poly-Si TFT 10-200 ~500 oC (glass)
PDA and laptop screens, Projection light valves, viewfinders
Crystalline Si MOSFET 400 ~1100 oC (c-Si)
Projection light valves, view finders
Figure 2-1(a) illustrates the basic concept of TFT operation [6]. Here the voltage and
charge polarities for n channel device (which means the channel current carried by electrons) are
shown. A positive gate voltage (VGS) draws additional electrons into the channel, while a
negative gate voltage reduces the channel electron density. TFT operation is quantified in terms
of the drain current (ID), the voltage between source and drain (VDS) and the voltage between gate
and source (VGS) i.e., the polarity and magnitude of the gate bias. Basically, the drain current of
a transistor is defined as a function of VDS, VGS and VT [6] according to:
( ) ( )
10
where μ is the mobility of electrons in the channel, Cox is the gate oxide capacitance per unit area,
W/L is the ratio of channel width to channel length and VT is the threshold voltage of the
transistor. From the equation describing drain current, carrier mobility is an important parameter
in determining device performance. For high carrier mobility in silicon based TFTs, a high
processing temperature is required, as is shown in Table 2-1. The a-Si TFT only has a carrier
mobility of only 0.3-1 cm2/V.sec due to its amorphous structure. Although the carrier mobility
of a-Si TFTs is enough for LCDs (liquid crystal displays) switching functions, it limits the use of
a-Si TFTs with other applications. Figure 2-1 (b) shows a diagram of the drain current versus the
drain- to- source voltage at different gate voltages. The ratio of the highest drain current to the
lowest drain current (Ion/Ioff) is a key parameter in transistor operation. A detailed discussion of
the importance of these parameters in the evaluation of transistor performance is presented below.
(a)
11
(b)
Figure 2-1. (a) Schematic illustration of the central elements of TFT operation (b) the diagram of ID and VDS at different gate voltages [6]
In addition to the requirements of transparency and low processing temperatures for TFTs,
TFTs are aggressively being scaled down. The limits of scaling down are imposed by the
thickness of the gate oxide. Thinner SiO2 leads to current leakage due to direct tunneling of
electrons through the gate oxide. For example, when 1V of gate voltage is applied in
complementary metal oxide semiconductor (CMOS) applications, the leakage current in SiO2
changes from 10-6 A/cm2 at 3 nm thickness to 10 A/cm2 when the thickness is reduced to 1.5 nm
[9]. The magnitude of the current density increase is seven orders although the thickness
changed by only a factor of 2. To solve the issue of leakage currents, alternative materials have
been investigated recently. In particular, high permittivity dielectric (high-κ) gate oxides are
being widely investigated as alternative gate oxide materials in metal-oxide-semiconductor
devices [10,11].
Summarizing the discussion above, the three main requirements for next generation TFTs
are transparency, low processing temperatures and high-κ gate oxides.
12
2.1.2 Transparent Conducting Oxide Semiconductor Overview
Transparent conducting oxide semiconductors have been widely investigated for optical
and electronic devices, such as transparent p-n junction rectifiers [12,13], light-emitting diodes
(LEDs) [ 14 ] and as mention above, transparent thin film transistors (TFTs) [3, 15 , 16 ].
Transparent oxide semiconductors have many advantages over silicon semiconductors for some
applications. As their name suggests, transparent semiconductors are transparent in visible range.
They are also environmentally stable, and have high field effect mobility when compared to
silicon. Transparent oxide semiconductors such as zinc oxide (ZnO), indium tin oxide (ITO)
[17-19], tin oxide (SnO2) [20] and gallium doped zinc oxide (GZO) [21], have been investigated
as channel materials in TFTs. In particular, ZnO has attracted considerable attention due to its
excellent electrical and optical properties. ZnO is discussed in more detail in the next section.
2.1.3 Introduction to Zinc Oxide
Zinc oxide exhibits an interesting combination of optical, electrical, optoelectronic,
semiconducting and piezoelectric properties [22]. The device applications for ZnO include
sensors [23], photovoltaics [24], photodetectors [25] and TFTs [16-20, 26-28]. ZnO has a wide
bandgap of 3.4 eV and its large exciton binding energy of 60 meV at room temperature is very
useful for blue and ultra-violet light emitting devices. There are other favorable aspects of ZnO
including compliance with wet chemical etching, low threshold power for optical applications,
and biocompatibility [29]. These properties make it an ideal candidate for a variety of devices
ranging from sensors to ultra-violet laser diodes and transparent TFTs for displays.
ZnO crystallizes in the wurtzite (hexagonal) structure shown in Figure 2-2 [6]. The lattice
parameters of the hexagonal are a = 0.3296 and c = 0.52065 nm [6]. As two interpenetrating
13
sublattices of Zn2+ and O2-, each Zn ion is surrounded by a tetrahedron of O ions, and vice-versa.
This tetrahedral coordination gives polar symmetry along the hexagonal axis which is c-axis.
This polarity is responsible for the piezoelectricity of ZnO.
ZnO is a direct band semiconductor. The excitation binding energy is 60 meV at 330 K.
The electronic band structure of ZnO has been calculated, and is shown in Figure 2-3 [30,31].
According to W. J. Fan et al. [31], the calculated band-edge energy at the Γ point for wurtzite
ZnO is 3.4409 eV. It is transparent in the visible region of the optical spectrum, and thus is less
light sensitive than smaller bandgap materials.
Figure 2-2. The hexagonal wurtzite structure of ZnO. O atoms are shown as large spheres and Zn atoms are the smaller ones [6]
14
Figure 2-3. The band structure of wurtzite ZnO [31]
The formation of defects in ZnO is mainly responsible for its conductive properties [32-
34]. C. Richard et al. [35] calculated the formation energies of the defects in ZnO, which are
shown in Table 2-2. They observed that the electrical properties of ZnO are strongly related to
the types of defects including oxygen interstitials (Oi), oxygen vacancies (VO), zinc interstitials
(Zni) and zinc vacancies (VZn). The defect species act as donors or acceptors in ZnO by the
mechanisms shown in Table 2-2 [35].
15
Table 2-2 The calculated energies for the defect formation in ZnO. The Kroger-Vink defect notation is used throughout [35]
The reactions based on the Kroger-Vink notation [35] indicate electron contributions in
ZnO from defects such as
. Therefore, when
zinc interstitials exist in ZnO films, the material is expected to be n-type. In contrast, reactions
such as
( )
( )
, result in the formation of p-type
ZnO. Therefore both Zni and Vo are donor levels while VZn and Oi produce acceptor levels,
which in turn determines the conductivity type. However, the formation energies of defects must
be considered. The reactions for
( )
( )
have
formation energies of 6.129 and 10.345 eV, respectively. The reactions for
16
have formation energies of 2.183 and 3.563 eV respectively, and
are therefore more thermodynamically favorable than the formation of oxygen interstitials.
Egelhaaf et al. [36] reported that the photoluminescence from ZnO thin films are caused
by radiative transitions between shallow donors (Vo and Zni) and deep acceptor (Vzn). B. J. Jin et
al. [37] proposed a similar mechanism from ZnO thin films deposited on a sapphire (001) at
400oC with different oxygen pressures. The photoluminescence of ZnO thin films deposited on a
sapphire is shown in Figure 2-4 and includes a narrow UV and a broad green-yellow band. The
narrow peak near 375 nm (3.3 eV) is the band edge and the broad band around 490 (2.52 eV) ~
550 nm (2.25 eV) is a combination of emissions from defects which form deep levels [36]. The
photoluminescence spectra related to the defects in ZnO thin films have been reported
extensively [38-42]. H. Q. Wong et al. [43] discussed the energy level diagram of the defects in
the band gap of ZnO, as is shown in Figure 2-5. The blue emission at 2.9 eV (427 nm), is related
to the transition of electrons between the zinc interstitial defect level and the valence band. The
green emission at 2.38eV (521 nm) is the transition between the conduction band and the antisite
oxygen Ozn defect level. The orange emission at 2.06eV (601.9 nm), is from the zinc interstitial
defect level to the oxygen interstitial defect level. Clearly, the emission properties of ZnO thin
films depend critically on their microstructure which in turn depends on the deposition method.
High intensity emission from ZnO films is obtained from a variety of methods including metal-
organic chemical vapor deposition (MOCVD) [44,45], molecular beam epitaxy (MBE) [46,47],
sol–gel deposition [48,49], RF magnetron sputtering [50], pulsed laser deposition (PLD) [51,52]
and atomic layer deposition [53].
17
Figure 2-4. PL spectra obtained from as-grown samples prepared under oxygen pressures of 50, 200, 300, 400 and 500 mTorr at a substrate temperature of 400oC. [37]
Figure 2-5. Calculated defect levels in ZnO films [43]
18
2.1.4 High- κ Materials for Gate Oxides
High-κ materials refer to materials with a relative high dielectric constant κ as compared
to silicon dioxide. The dielectric constant of silicon dioxide is about 3.9, which has been used as
a gate oxide for many years. However, there are some limitations in the use of silicon dioxide.
When the sizes of electronic devices such as transistors decrease, the thickness of the SiO2 gate
oxide also decreases. As the thickness of silicon dioxide scales below 2 nm, leakage currents
increase drastically due to the tunneling effect [54]. Theoretically, in an ideal MIS (metal-
insulator-semiconductor) structure the conductance of the insulator layer is assumed to be zero
[55]. However, some current will conduct through the insulator layer when the applied electrical
field is sufficiently high. Table 2-3 shows the basic conduction mechanism in insulators [55].
Table 2-3 Basic Conduction Processes in Insulators [55]
19
In a metal-insulator-semiconductor structure, the most common conduction mechanism in
the insulator is tunneling effect at high electrical fields. Tunnel emission is a quantum
mechanical phenomena in which the electron wave function penetrates through a potential
barrier [55]. Thermionic emission is the heat-induced excitation of carriers from the
semiconductor over the potential barrier into a metal or from the metal over a potential barrier
into a semiconductor. This occurs because the carrier overcomes the potential barrier by
acquiring thermal energy, so the carrier (electron or hole) transport is dependent on the
temperature. Frenkel-Poole emission is the emission of trapped electrons into the conduction
band. The supply of electrons from the traps is thermal excitation. Charge carriers are excited
by thermal energy and hop from one isolated site to the next. Therefore, this mechanism will
yield an Ohmic characteristic depending on temperature or low electrical field [55]. Ionic
conduction mechanism can be described as a diffusion process. Basically, ions cannot be
injected into or extracted from the insulator. However, when a field is applied on MIS, positive
or negative space charges will accumulate near the surface of the metal and the semiconductor.
The semiconductor-insulator interfaces will then cause a distortion of the potential distribution.
After the applied field has been removed, the internal fields which are caused by the
accumulation of charge remain and cause displaced ions to flow back toward their equilibrium
positions. Because of this residue of charges, hysteresis curves may be observed in current-
voltage traces. A space-charge-limited current is due to carriers being injected into the insulator,
where compensating charge is not present. Conduction by the space-charge-limited process is
typically small, since the mobility of carriers in insulators is very low.
20
The leakage currents caused by the thinner silicon dioxide layers lead to power
consumption and reduce the reliability of electronic devices. Therefore, many efforts are
replacing silicon dioxide with high-κ materials and this has been ongoing for decades [54, 56-60].
The principle of replacing silicon dioxides with high-κ materials can be explained by the
equation for a parallel plate capacitor:
( )
where A is the capacitor area, κ is relative dielectric constant of the material, εo is the permittivity
of free space, t is the thickness of the capacitor oxide.
From the equation of the parallel plate capacitors, if a larger capacitance is needed, the
thickness of capacitor has to be reduced. However, the leakage currents limit further reduction
of thickness, thus an alternative method to increase the gate capacitance is to replace silicon
dioxide by high-κ materials. Because of the high-κ value, a high-κ material can form a thicker
gate layer, but keep the same capacitance. The thicker gate layer would reduce the leakage
currents flowing through the structure, as well as improve the reliability of the gate dielectric
layer. There are many things to be considered when replacing silicon dioxide with a high-κ
material. The bulk and interface properties of the high-κ material must be comparable to those
of silicon dioxide. Many high-κ materials have been reported to have the potential to replace
silicon dioxide, such as SiON, Si3N4, Ta2O5, Al2O3 and HfO2 [54, 56-61]. For example, H. J. H.
Chen et al. [61] analyzed the electrical characteristics of ZnO thin film transistors with HfO2 gate
dielectrics, as schematically shown in Figure 2-6 [61]. The capacitor structure is
ITO/ZnO/HfO2/ITO/glass. The leakage current of the HfO2 dielectric for this capacitor tested as
a function of temperature is shown as Figure 2-7 [61]. For the as-deposited sample, the leakage
21
current was relatively high. After annealing in ambient oxygen, the leakage current is reduced
due to the significantly improved stoichiometry of the samples. The results show that leakage
currents can be reduced by optimal annealing conditions.
Figure 2-6. Structure of a ZnO thin film transistor with high-k HfO2 [61]
Figure 2-7. Capacitor leakage current dependence on annealing temperature for HfO2 dielectric [61]
2.1.5 Introduction to BaTa2O6 and Al2O3
Barium Tantalate (BaTa2O6) is a high κ material with dielectric constant up to 30, and
exhibits a low leakage current in high electrical field [62]. Due to the excellent electrical and
22
optical properties such as high transparency, BaTa2O6 thin films have good potential as a gate
oxide layer for electronic devices [63,64]. BaTa2O6 is a ternary compound composed of BaO
and Ta2O5. Hexagonal Barium Tantalate could be obtained at high temperatures above 750oC
[65]; otherwise, amorphous Barium Tantalate is formed at lower processing temperatures. Y.S.
Kim et al. [62] presented the electrical properties of the sputtered BaTa2O6 as a function of
processing parameters, especially at lower temperatures. They prepared three types of samples
by varying the deposition temperature for amorphous-BaTa2O6 thin films. The dielectric
constant of amorphous BaTa2O6 thin films as a function of the substrate temperature is shown in
Figure 2-8 [62]. The dielectric constant was as high as 30 and was almost independent of
frequency from 0.3-100 kHz. The characteristic of leakage currents for amorphous BaTa2O6 is
shown in Figure 2-9 [62]. The leakage currents for the three types of BaTa2O6 thin films are on
the order of 10-6~10-7 A/cm2 at the applied field of 1 MV/cm which is comparable with silicon
dioxide [66]. These characteristics showed that of BaTa2O6 is potentially good for electronic
devices.
23
Figure 2-8. Dielectric constant and loss factor of amorphous-BaTa2O6 with the different substrate temperature [62]
Figure 2-9. Characteristic leakage current of amorphous-BaTa2O6 thin films with different substrate temperatures [66]
24
Aluminum oxide (Al2O3) has been studied as a potential replacement for silicon dioxide
as a gate dielectric for years [72-74]. The dielectric constant of Al2O3 is in the 7~10 range,
which is almost twice that of silicon dioxide [67]. The basic properties of Al2O3 films grown as
gate dielectrics are well understood [68-69], thus Al2O3 can be used as a reference to investigate
new materials for alternative gate oxides. M. Voigt et al. [70] fabricated Al2O3 thin films by RF
magnetron sputtering. They used an Al2O3 target with purity 99.99%, base pressure: 2 x 10-6
mbar, a mixture Ar/O2 as the working gas, and working pressure of 1.3x10-3~8.0 x10-3 mbar.
When 10% of O2 in the Ar/O2 was used as the working gas, the break down strength was 0.3
MV/cm. The current density versus applied electrical fields is shown in Figure 2-10 [70].
Figure 2-10. Current density verse electric field strength of Al2O3 films prepared with 10% O2 of Ar/O2 mixture gas [70]
25
J. Choi et al. [71] also deposited Al2O3 thin films by sputtering at room temperature in a
Ar/O2 mixture gas. All of the Al2O3 films were amorphous regardless of the oxygen content in
the working gas. This is shown in the XRD plot of Figure 2-11.[71] The current density versus
applied field for Al2O3 grown with various oxygen contents in the Ar/O2 working gas is shown in
Figure 2-12 [71]. The leakage current was in the range of 10-6~10-7 at the applied field 600
KV/cm.
Figure 2-11. X-ray diffraction patterns of the Al2O3 films deposited with different O2 content in a working gas of Ar/O2 [71]
26
Figure 2-12. Current density versus applied field for Al2O3 films with various oxygen content in a Ar/O2 working gas mixture [71]
2.2 Metal-Insulator-Semiconductor Structures
A metal-insulator-semiconductor is composed of an insulator sandwiched between a
metal and semiconductor layers. Metal-insulator-semiconductor structures are very useful for
evaluating the performance of electronic devices. The following sections present more details.
2.2.1 The Importance of Metal-Insulator-Semiconductor Structures to Transistor Technology
The Metal-Insulator-Semiconductor (MIS) structure is one of the most useful devices for
the study of the interface properties of semiconductor devices [55,72]. Most of the practical
problems for the stability and reliability of all semiconductor devices are strongly related to the
conditions at the interfaces. Therefore, understanding the interface properties with the help of
27
MIS capacitors is of great importance. The first MIS structure was made in 1960 and a
comprehensive study can be found in MOS physics and Technology by Nicollian and Brews [73].
The discussions here are applicable to all insulator-semiconductor systems, even though the
example generally used is the SiO2-Si system.
A typical MIS structure is shown in Figure 2-13 [55]. It is composed of a metal contact,
an insulator layer, a semiconductor layer and an Ohmic contact in series. The insulator thickness
is d. , The energy band diagrams of n-type and p-type MIS structures without bias are shown in
Figure 2-14 (a) and (b). ϕm, ϕn and ϕp are the work functions of the metal, n-type semiconductor
and p-type semiconductor, respectively [55]. Eg is the energy gap while Ec, Ei, EF and Ev are the
energy levels of the conduction band edge, the Fermi level of the intrinsic material, the Fermi
level of the doped material and the valence band edge, respectively. χ and χi are the electron
affinities of the semiconductor and the insulator. For an ideal MIS, the following conditions
apply:
i. There are no interface trapped charges or any kind of oxide charge.
ii. The resistivity of the insulator is infinite so there is no carrier transport through the insulator
under dc bias.
iii. There is no work function difference between the metal and the semiconductor.
When an ideal MIS is positively or negatively biased, the energy band bending is as
indicated in Figure 2-15(a)(b)(c) [55]. The top figures represent p-type semiconductors and the
bottom figures represent n-type semiconductors. Figure 2-15 (a) shows the accumulation case,
where the accumulation of majority carriers (holes for p-type, electrons for n-type) near the
semiconductor surface is caused by band bending when a negative voltage is applied to the metal
contact (V<0) for p-type and a positive voltage is applied to the metal contact (V>0) for n-type.
28
Figure 2-15 (b) shows the inversion case with depleted majority carriers due to downward band-
bending when a small positive voltage (V>0) is applied to the metal contact for p-type and a
small negative voltage (V<0) is applied to the metal contact for n-type. When the applied
positive or negative voltage to the p-type and n-type is larger, the bands bend even more severely
downward. The intrinsic level Ei at the surface will cross over the Fermi level EF and then the
minority carrier (electrons for p-type, holes for n-type) at the surface will be larger than majority
carrier, which is shown in Figure 2-15 (c). The surface is thus inverted and this is the inversion
case.
Figure 2-13. A simple Metal-Insulator-Semiconductor (MIS) capacitor [55]
29
Figure 2-14. The energy band diagrams without bias for (a) MIS with n-type semiconductor and (b) MIS with p-type semiconductor [55]
Figure 2-15. Energy band diagrams for (a) Accumulation case for MIS with p-type semiconductor (top) and MIS with n-type semiconductor (bottom) (b) depletion case for MIS with p-type semiconductor (top) and MIS with n-type semiconductor (bottom) (c) Inversion case for MIS with p-type semiconductor (top) and MIS with n-type semiconductor (bottom) [55]
(a) (b) (C)
30
However, defects and impurities exist in real MIS materials and structures and the ideal
description is only a theoretical construct. Defects can be formed, for an example, when the first
oxide layer on the Si is a partial monolayer, i.e., SiOx rather than SiO2. In such a situation
interface traps and oxide charges will exist between the Si and a subsequently grown SiO2 layer,
and will affect the metal-oxide-semiconductor characteristics. In the following sections, a more
detailed account of real MIS structures will be given.
2.2.2 Factors Influencing MIS Performance and Characteristics: The Influence of Interfacial
and Bulk Defects on Switching Speed and Power Required for Switching
As mentioned above, MIS structures are very useful tools to evaluate the performance of
thin film transistors. In this regard, the threshold voltage, VT, is an important concept. It
accounts for the shift from non-zero flat band voltage which is due to extra charges from defects
and the work function difference between the semiconductor and the gate material. The
threshold voltage can be calculated by the equation [55]:
√ ( )
( )
(VT: the threshold voltage, VFB: the flat band voltage, ψB: the voltage across the semiconductor, εs:
the dielectric constant of the semiconductor, NA: the doping concentration, Cox: the capacitor of
the gate oxide)
In an ideal MIS, there is no work function difference between the gate metal contact and
the semiconductor layer. However, in reality the work function difference between the metal
contact and the semiconductor layer is not zero. The work function difference will cause energy
band bending at the interface. Defects such as impurities and incompletely oxidized elements in
insulator layers will induce interface traps and oxide charges. To balance the energy band
31
bending caused by the extra charges from defects and the work function difference, an extra
voltage has to be applied. This extra voltage is known as a flat-band voltage (the voltage
required to remove the band bending and establish flat-band conditions). The switching power
of transistors can be then estimated by shifts of the flat band voltages in MIS devices. A large
shift in flat band voltages is an indication of larger required switching powers for transistor
operation. The performance of MIS also affects the sub-threshold region in transistor operation.
The sub-threshold region is the linear part of Figure 2-1 (b). When the gate bias is below the
threshold voltage and the semiconductor is in weak inversion or depletion, the corresponding
current is the sub-threshold current which indicates how fast the current drops with the gate bias.
A parameter to quantify how fast the transistor is switched on/off by the gate voltage is called the
sub-threshold swing (S) which is the inverse of the sub-threshold slope.
By the definition, the sub-threshold swing can be calculated according to [55]:
( )
( ) ( )
In the sub-threshold region, the drain current varies exponentially with the surface
potential of the semiconductor, ψs. [55]
( )
( )
( ) (
) ( )
where Cox is the capacitance of the gate oxide and CD is the capacitance of the depletion region in
the semiconductor.
If significant interface defects exist at the interface between the semiconductor and the
gate oxide, the equation accounts for the capacitance of interface defects, Cit, which is taken to
be in parallel with the depletion layer capacitance, CD. Therefore,
32
( ) (
) (
) ( )
From Equation 2-6, the properties of gate oxide also affect the performance of transistors
profoundly. If a large Cit exists, which means a large concentration of interface defects exist,
then the slope of the drain current is smaller, which means the switching speed is slower in
transistor operation.
In addition, the charge trapped inside the oxide and traps at the interface between the gate
oxide and the semiconductor also affect the performance of transistors. This charge is trapped by
impurities and/or crystalline defects. Impurities may be intentionally or unintentionally
incorporated during the films growth and the manufacturing processes of devices. There are
generally four types of trapped charge associated with MIS devices. These are shown in Figure
2-16 [55]. They include 1) fixed oxide charges, 2) mobile oxide charges, 3) oxide trapped
charges and 4) interface trapped charges.
1) Fixed oxide charge: This charge is related to the oxidation process of the silicon
oxide, and depends on the oxidation temperature, cooling conditions, and the
orientation of silicon wafer. Fixed oxide charges exist near the surface of the
oxide and are immobile under any bias applied.
2) Mobile oxide charge: This is due to ionic impurities such as sodium or other alkali
ions and possibly H+. They are mobile within oxide under high temperature and
high electrical field operation.
3) Oxide trapped charge: The oxide trapped charges are associated with defects in
the SiO2 and are distributed inside the oxide layer. Unlike fixed oxide charges,
oxide trapped charges may be removed by thermal annealing.
33
4) Interface trapped charge: Interface trapped charges are dependent on structural
defects, oxidation-induced defects, and metal impurities and could be caused by
radiation or bond breaking processes.
Although the descriptions above use SiO2, as mentioned earlier, they apply to other gate oxides
as well.
Figure 2-16. Terminology for the charge associated with MIS devices [55]
2.2.3 The Current Status of MIS Technology
Currently, SiO2 is the most widely used insulator layer in MIS devices. A conventional
MIS capacitor with SiO2 dielectric layer is fabricated using n-type or p-type Si (001), and a gate
SiO2 (~3nm) layer, which is grown at 850oC in ambient oxygen [74]. For MIS capacitors having
a thermally grown SiO2 on Si, some of the interface trapped charges can be neutralized by
annealing. The interface trapped charge concentration is normally in the 1010~1011 eV-1cm-2
range. SiO2 generally exhibits low defect density and high resistivity. However, with continuous
scaling down of MIS technology involving the gate oxide being decreased below 2nm [75],
conventional SiO2 dielectrics are becoming inadequate, because the leakage currents become
34
unacceptably large, and limit the performance of devices [75]. Therefore, considerable attention
has been focused on the search for other insulator films that can replace conventional materials.
For example, silicon oxynitride thin film is now investigated as a gate oxide material. K. F.
Albertin et al. [ 76 ] used SiOxNy films, and their results showed that a small nitrogen
concentration (8%) in the films is enough to decrease the leakage current and also improve the
Si/SIOxNy interface. They reported an interface trap density of 4.55 x 1010 eV-1cm-2 [76]. E. H.
Oulachgar et al. [77] also used silicon oxynitride as gate oxides and observed an interface
trapped density of 3.2x1010 - 4.5x1010 eV-1cm-2. However, when the thickness of SiO2 or SiON
was decreased below 2 nm, the leakage current and power consumption were unacceptable and
caused reliability issues. Therefore, high k materials have been investigated to replace Si-based
gate oxides. These include Ta2O5 [78,79], HfO2 [80,81], ZrO2 [82,83], and Al2O3 [84,85]. Some
ternary compounds have also been investigated as potential candidates for gate oxides. For
example, B. Mereu et al. [86] used La2Hf2O7 as an insulator layer in MIS structures. In his study,
the interface trapped density ranged from 3.4 x1010 eV-1cm-2 to 4.8 x1012ev-1cm-2 depending on
the processing parameters.
2.3 Electrical Characterization Techniques
Capacitance-Voltage (C-V) measurements are fundamental characterization methods for
MIS devices. Capacitance is measured with an AC signal probe as a DC bias voltage is swept on
the device. A typical C-V curve of a MIS structure with a p-type semiconductor is shown in
Figure 2-17 (a)(b)(c) [55]. Figure 2-17 (a) shows the capacitance-voltage dependence of an ideal
MIS device. When the DC bias is negative, the device is in accumulation. When the DC bias is
swept from negative to positive, the capacitance is decreased due to the fact that the majority
35
carriers are away from the interface between the insulator and the semiconductor. When the DC
becomes large and positive the structure is biased into inversion. However, in a real MIS device
the curve tends to be shifted due to the difference in work functions, as shown by the curve in
Figure 2-17 (b). In addition to the work function difference, the charge at the interface between
the insulator and the semiconductor also contributes to the curve distortion. This is shown in
Figure 2-17 (c).
Figure 2-17. Typical C-V dependence for (a) an ideal MIS with p-type semiconductor (b) the lateral shift in the C-V curve (c) distorted C-V curve [55]
2.3.1 Capacitance-Voltage (C-V)
As already mentioned, the procedure for taking C-V measurements involves sweeping a
DC bias voltage across the MIS structure, while making the measurements with an AC probe
signal. In the accumulation region, a strong DC bias causes the majority carriers to accumulate
near the semiconductor surface. Since the majority carriers cannot be transported through the
36
insulating layer, the capacitance is at maximum under this condition. Therefore, the maximum
capacitance (Cmax) of the MIS device is equal to the oxide capacitance, Cox. A simple circuit for
a MIS structure with a p-type semiconductor is shown in Figure 2-18 [55]. As the DC bias is
applied, incremental charge is added to the metal and the semiconductor. The incremental
charge is separated by the gate oxide. Therefore, from the relation, Cox = εoεox A/ d, the dielectric
constant or gate oxide thickness can be calculated if one of them is known (εo is the vacuum
permittivity, εox is the dielectric constant of the oxide layer, A is the contact area and d is the
thickness of the oxide.). Y. S. Noh et al. [78] reported on the C-V characteristics of a MIS
structure with Ta2O5 oxide films on semiconducting ZnO, which is shown in Figure 2-19. The
C-V measurement was performed a 1MHz AC probe signal, and the DC bias was swept from -4
V to 3V. The maximum capacitance in the accumulation region is around 2.55nF. The dielectric
constant of Ta2O5 was calculated to be 25 when the oxide thickness is 105 Å [78.]
Figure 2-18. The simple circuit of a MIS structure with p-type semiconductor [55]
37
Figure 2-19. High-frequency (1MHz) capacitance-voltage and conductance-voltage characteristics of MIS capacitors using Ta2O5 films deposited on ZnO layers [78]
When the device starts to deplete, the majority carriers are pushed away from the
semiconductor surface and minority carriers are attracted to the depletion region. In other words,
these minority carriers become dominant near the semiconductor surface. The capacitance of the
MIS structure now equals Cox plus the depletion layer capacitance in series. The semiconductor
surface charge now is inverted from positive to negative, as is shown in Figure 2-20. If the probe
frequency is high enough, the minority carriers cannot respond to changes in the gate voltage and
the capacitance becomes equal to the oxide capacitance and the maximum depletion layer
capacitance in series. A simple circuit for the oxide capacitance and the maximum depletion
layer capacitance in series is also shown in Figure 2-20 [55]. Under these conditions, the flat-
band capacitance is calculated by
)72(1
1
EqL
C
C
s
D
ox
FB
38
where LD is the extrinsic Debye length, √
(εs is the dielectric constant of
semiconductor, KB is the Boltzmann constant, q is the electron charge, and N is the carrier
concentration). Once the value of CFB is known, the value of VFB can be interpolated from the
closest voltage value from C-V curves.
Figure 2-20. The inverted surface charge on the p-type semiconductor surface and the simple circuit for the oxide capacitance and the maximum depletion layer capacitance in series [55]
Another simple method for MIS characterization is to measure its hysteresis curve. The
hysteresis in the capacitance-voltage curves is due to injected electrons being trapped by
interfacial traps between the oxide and the semiconductor. Therefore, capacitance hysteresis is
observed when the voltage is swept forward through accumulation, then reverse biased through
depletion and into inversion. E. K. Kim et al. [81] measured a C-V hysteresis with a voltage gap
of 2.8 V for a MIS with ZnO measured at room temperature, as is shown in Figure 2-21. Figure
39
2-21 indicates a counterclockwise hysteresis in the C-V curve, and suggests charging at the
interface due to interfacial trapping. The interface trapped charge can be calculated according to:
)82(
EqqA
VCN FBox
it
where Cox is the previously discussed oxide capacitance measured in accumulation, ΔVFB is the
flat band voltage shift as determined directly from the forward and reverse C-V sweeps, q is the
electronic charge and A is the area.
Figure 2-21. The hysteresis curves with a voltage gap of 2.8V for capacitance-voltage characteristics of the sample [81]
2.3.2 Conductance-Voltage (G-V)
When a DC bias is applied to a MIS device, the Fermi level moves upward or downward
(as shown in Figure 2-15) with respect to the interface trap levels and then a change in the charge
state of the interface traps occurs [55]. This change of charge affects the MIS capacitance and C-
40
V curves. Figure 2-22 (a) shows an equivalent circuit incorporating the interface traps effect. Ci
and Cd are the insulator capacitance and the semiconductor capacitance in the depletion region.
Cit and R are the capacitance and the resistance associated with the interface traps. Figure 2-22
(a) can be converted into a frequency dependent capacitance Cp in parallel with a frequency-
dependent conductance, Gp, as is shown in Figure2-22(b) [55]. From the circuits, either a
capacitance measurement or a conductance measurement can be used to evaluate the interface
trap density, because both the input capacitance and input conductance in the equivalent circuits
contain similar information about the interface traps. However, difficulty arises in the
capacitance measurement for evaluation of interface trap density. As seen in Figure 2-22 (a), the
capacitance measurement needed must be extracted from the measured capacitance which
consists of the oxide capacitance, depletion layer capacitance and interface trap capacitance, so
that greater inaccuracies arises when extracting the capacitance from the devices. This difficulty
does not apply to the measurement for conductance, which is directly related to the interface
traps. Therefore, the conductance measurement for evaluation of the interface trap density yields
more accurate results [55].
41
Figure 2-22. (a) An equivalent circuit including interface trap effects, Cit and Rit (b) An equivalent circuit with a frequency dependent capacitance, Cp and frequency dependent conductance, Gp [55]
Conductance-voltage characterization is used to investigate the traps density at the
interface between the insulator and the semiconductor. The interface trap density, Dit, was
calculated from the peak of G-V plots by Nicollian, Goetzberger [87]. The interface trap density
can be determined by using the equivalent parallel conductance Gp/ω. (Gp is the parallel
conductance and ω is the angular frequency)
)92(
])([ 222
2
EqCCG
GCG
moxm
moxp
and then the interface trap density can be calculated from [80]
)102(})]([{
2222
2
EqGCCGqA
GCD
mmoxm
moxit
42
where Gm is the peak value of measured conductance, Cox is the capacitance of the
insulator, Cm is the measured capacitance and q is the charge (q =1.602 x 10-19 C). An example
is the work by El H. Oulachgar et al. [77] who used conductance-voltage measurements to
evaluate interface trap densities. Figure 2-23 shows the conductance-voltage measurement of
SiONC/(p)Si MOS structure at 1 MHz [77]. The peaks appeared as the bias was swept through
the depletion region. The interface trap density can be extracted by using Eq 2-10 above and
were determined to be 3.24 x1010~1.26 x1011 eV-1cm-2.
Figure 2-23. Conductance-voltage characteristics of a SiONC/(p)Si MOS structure measured at 1 MHz [77]
43
2.3.3 Frequency Dependent C-V and G-V Methods
In ideal cases C-V curves are frequency independent. However, if a large number of
interface traps exist, this causes frequency dispersion in the C-V measurements. The interface
traps may lead to dispersion both in the depletion and the accumulation regions. At lower
frequencies, the interface traps can easily follow the AC signal and when biased into
accumulation the interface traps are filled, yielding an excess capacitance. When a sufficient
high frequency is applied and the interface traps cannot follow the ac signal, an excess
capacitance cannot be produced. Therefore, the values of the capacitance decrease with
increasing frequency [88]. Ilbilge et al. [88] investigated the C-V characteristics of Au/SiO2/n-Si
MIS as function of frequency in the 0.5kHz-10MHz range at room temperature, as is shown in
Figure 2-24. Each C-V curve has a distinct accumulation, depletion and inversion region, and
the capacitance decreases when the frequency is increased.
Figure 2-24. The measured capacitance at various frequency for Au/SiO2/n-Si MIS [88]
44
2.4 References
[1] W. den Boer, “Active Matrix Liquid Crystal Displays: Fundamentals and Applications”
Elsevier (2005)
[2] A. Yamagishi, S. Naka and H. Okada, J. J. Appl. Phys. 49, 048002 (2010)
[3 ]J. Jo, O. Seo, E. Jeong, H. Seo, B. Lee and Y. Choi, J. J. Appl. Phys. Vol. 46, No. 4B, pp.
2493-2495 (2007)
[4] D. Hong, N. L. Dehuff, R. E. Presley, C. L. Munsee, J. P. Bender1, C.-H. Park, J. F. Wager
and D. A. Keszler, Mat. Res. Soc. Symp. Proc. Vol. 796, V1.2.1-V1.2.6 (2004)
[5] E. Artukovic, M. Kaempgen, D. S. Hecht, S. Roth and G. Gru1ner, Nano Lett. Vol.5, No.4,
757-760 (2005)
[6] C. Jaqadish and S. J. Pearton, “Zinc oxide bulk, thin films and nanostructures: processing,
properties”,. Elsevier (2006)
[7] S. Wagner, H. Gleskova, J. C. Sturm, and Z. Suo, “Novel processing technology for
macroelectronics,” in Technology and Application of Amorphous Silicon, R. A. Street, Ed.
Berlin, Germany: Springer-Verlag, vol. 37, pp. 222–251 (2000)
[8 ] H. Iechi, Y. Watanabe, H. Yamauchi and K. Kudo, IEICE TRANS. ELECTRON., Vol. E91-
C, No. 12 (2008)
[9 ] S. K. Ray R. Mahapatra S. Maikap, J Mater Sci: Mater Electron 17:689–710 (2006)
[10 ] C. Chui, S. Ramanathan, B. B. Triplett, P. C. Mcintyre and K. C. Saraswat, IEEE Electron
Device Lett. Vo. 23, No. 8 (2002)
[11] M. Tapajna, K. Husekova, J. P. Espinos, L. Harmartha and K. Frohlich, Mater. Sci. Semi.
Pro. 9 pp. 969-974 (2006)
45
[12] [12] Kallol Mohanta, Sudip K. Batabyal, and Amlan J. Pal, Chem. Mater. 19, 3662-3666
(2007)
[13 ] L. Zhuang and H. L. Wong , Appl. Phys. A 87, 787–791 (2007)
[14] T. Kamiya, H. Hiramatsu, K. Nomura and H. Hosono, J Electroceram 17, 267–275 (2006)
[15] G. Lavareda a,b, C. Nunes de Carvalho E. Fortunato, A.R. Ramos d, E. Alves d, O. Conde,
A. Amaral, J. Non-Crystal. Solids 352, 2311–2314 (2006)
[16] Chaun Gi Choi, Seok-Jun Seo, and Byeong-Soo Bae Electrochemical and Solid-State
Letters, 11 1 H7-H9 (2008)
[17] R. L. Hoffman, B. J. Norris and J. F. Wager, Appl. Phys. Lett. Vol. 82. No. 5 (2003)
[18] S. N. Cha, J. E. Jang, Y. Choi, G. W. Ho, D. J. Kang, D. G. Hasko, M.E. Welland and G. A.
J. Amaratunga, Proceddings of ESSDERC, Grenoble, France, pp.217-219 (2005)
[19] K. Keem, J. Kang, C. Yoon, D. Yeom, D. Jeong, B. Moon, S. Kim, Microelectronic Eng. 84
pp.1622-1626 (2007)
[20] R. E. Presley, C. L. Munsee, C. Park, D. Hong, J. F. Wager and D. A. Keszler, J. Phys. D:
Appl. Phys. 37, pp. 2810–2813 (2004)
[21 ] E. Fortunato, P. Barquinha, A. Pimentel, A. Goncalves, A. Marques, L. Pereira, R. Martins,
Thin Solid Films 487, pp. 205-211 (2005)
[22] “Zinc Oxide Bulk, Thin Films and Nanostructure”, by Chennupati Jagadish, Stephen J.
Pearton, Elsevier (2006)
[23] P. P. Sahay, “Zinc Oxide Thin Film Gas Sensor for Detection of Acetone”, J. Mater. Sci. 40,
4383-4385 (2005).
46
[24 ] L. Ion, C. Tazlaoanu, G. Socol, L. Magherusan, I. Enculescu, D. Bazavan, I. Mihailescu
and S. Antohe, Mater. Res. Soc. Symp. Proc. Vol. 1013, 1013Z04-09 (2007).
[25] A. Ievtushenkoa, G. Lashkareva, V. Lazorenkoa, V. Karpynaa, V. Sichkovskyia, L.
Kosyachenkob, V. Sklyarchukb, O. Sklyarchukb, V. Bosyc, F. Korzhinskic, A. Ulyashind,
V. Khranovskyye, and R. Yakimovae, Acta Physicapolonica A, Vol. 114, 1123-1129
(2008).
[26] H. Cheng, C. Chen and C. Tsay, Appl. Phys. Lett. 90, 012113-1-3 (2007)
[27] Jin Hyung Jun, Byoungjun Park, Kyoungah Cho and Sangsig Kim Nanotechnology 20,
505201 (2009)
[28] M. Chen, T. Chang, S. Huang, K. Chang, H. Huang, S. Chen, J. Lu, D. Gan, N. Ho, T.
Young, G. Jhang, Y. Tai, Surface & Coatings Technology 204, pp. 1112–1115 (2009)
[29] “Zinc Oxide Bulk, Thin Films and Nanostructures” Chapter 1 V. A. Coleman, C. Jagadish,
C. Jagadish and S. Pearton (2006)
[30] C. Yang and K. S Dy, Soild State Commun. Vol. 88, No. 6, pp 491-494 (1993)
[31] W. J. Fan, J. B. Xia, P. A. Agus, S. T. Tan, S. F. Yu, and X. W. Sun, J. Appl. Phys., 99,
013702 (2006)
[32] F. D. Auret, W. E. Meyer, P. J. Janse van Rensburg, M. Hayes, J. M. Nel, H. von
Wenckstern, H. Hochmuth, G. Biehne, M. Lorenz, M. Grundmann Journal of Physics:
Conference Series 100, 042038 (2008)
[33] L. J. Brillson,a H. L. Mosbacker, b. M. J. Hetzer, Y. Strzhemechny, G. H. Jessen, D. C.
Look, G. Cantwell, J. Zhang, and J. J. Song, Appl. Phys. Lett. 90, 102116 (2007)
47
[34] M. Khalid, M. Ziese, A. Setzer, P. Esquinazi, M. Lorenz, H. Hochmuth, M. Grundmann, D.
Spemann, T. Butz, Phys. Rev B 80, 035331 (2009)
[35] C. Richard, A. C. Samuel, A. French, A. A. Sokol, A. A. Alsunaidi and S. M. Woodley, J.
Comput. Chem. pp. 2234-2249 (2008)
[36] H.J. Egelhaaf and D. Oelkrug, J. Cryst. Growth. 161, pp. 190 (1996)
[37] B.J. Jin, H.S. Woo, S. Im, S.H. Bae, S.Y. Lee, Appl. Surf. Sci. 169-170, 521-524 (2001)
[38] C. Chandrinou, N. Boukos, C. Stogios, and A. Travlos, Microelectronics Journal 40 296-
298 (2009)
[39] H. Song, J. Kim, and E. K. Kim, Microelectronic Journal 40, 313-315 (2009)
[40] T. Hur, G. S. Jeen, Y. Hwang and H. Kim J. Appl. Phys. Vol. 94, No. 9, 5787-5790 (2003)
[41] M. P. Bole and D. S. Patil, J. Phys. Chem. Solid. 70, 466-471 (2009)
[42] N. Mehan, V. Gupta, K. Sreenivas and A. Mansingh J. Appl. Phys. Vol. 96, No6. 3134-3139
(2004)
[43] HQWang, G Z Wang, L C Jia, C J Tang and G H Li, J. Phys. D: Appl. Phys. 40, 6549–6553
(2007)
[44] J. Sun, T. Yang, G. Du, H. Ling, J. Bian and L. Hu, Appl. Surf. Sci. 253, 2-66-2070 (2006)
[45] S. Muthukumar, H. Sheng, J. Zhong, Z. Zhang, N. W. Emanetoglu and Y. Lu, IEEE
TRANSACTIONS ON NANOTECHNOLOGY, VOL. 2, NO. 1, 50-54 MARCH (2003)
[46] A. El-Shaer, A. Che Mofor, A. Bakin, M. Kreye, A. Waag, Superlattices and
Microstructures 38, 265–271 (2005)
[47] K. Ogata, T. Kawanishi, K. Maejima, K. Sakurai, Sz. Fujita and Sg. Fujita, Journal of
Crystal Growth 237–239, 553–557 (2002)
48
[48]M. Caglar, S. Ilican, and Y. Calar Thin Solid Films 517, 5023-5028 (2009)
[49] G. Srinivasan and J. Kumar, Cryst. Res. Technol. 41, No. 9, 893-896 (2006)
[50] Q.P. Wang, D.H. Zhang, Z.Y. Xue and X.J. Zhang, Optical Materials 26, 23–26 (2004)
[51] X.M. Fan, J.S. Lian, Z.X. Guo and H.J. Lu, Applied Surface Science 239, 176–181 (2005)
[52] Sang Hyuck Bae, Sang Yeol Lee, Beom Jun Jin and Seongil Im, Applied Surface Science
169-170, 525-528 (2001)
[53] E. Prze´zdziecka1, Ł. Wachnicki, W. Paszkowicz, E. Łusakowska,T. Krajewski, G. Łuka, E.
Guziewicz and M. Godlewski, Semicond. Sci. Technol. 24, 105014, (2009)
[54] X. Garros, M. Casse, G. Reimbold, M. Rafik, F. Martin, F. Andrieu, V. Cosnier and F.
Boulanger, Microelectronic Engineering 86, 1609–1614 (2009)
[55] S. M. Sze and Kowk K. Ng, “Physics of Semiconductor Devices” 3rd Edition (2007)
[56] N. Ndiege, T. Wilhoite, V. Subramanian, M. Shannon and R. Masel, Mater. Res. Soc.
Symp. Proc. Vol. 929, 0929-II04-18-25 (2006)
[57] M. A. Shriver, A. M. Gabrys, T. K. Higman and S. A. Campbell, Mat. Res. Soc. Symp. Proc.
Vol. 670, K2.3.1-2.3.5 (2001)
[58] E. Atanassova and A. Paskaleva, Microelectronics Reliability 47, 913–923 (2007)
[59] Y. Zheng, C. Troadec, A. T. S. Wee, K. L. Pey, S. J. O’Shea and N. Chandrasekhar, J. Phys.
Conference Series 61, 1347–1351 (2007)
[60] Y. Suguyama, S. Pidin and Y. Morisaki, Fujitsu Sci. Tech. J., 39, 1 (2003)
[61] H. J. H. Chen and B. B. L. Yeh J. J. Appl. Phys. 48, 031103 (2009)
[62] Y.S. Kim, Y.H. Lee and M.Y. Sung, Solid-State Electron. 43, 1189-1193(1999)
49
[63] Y. Lee, Y. Kim, D. Kim, B. Ju and M. Oh, IEEE Trans. Electrondevice, Vol. 47, NO.
1(2000)
[64] N. Shepherd, D. C. Morton, E. W. Forsythe and D. Chiu Thin Solid Films 515, 2342–2346
(2006)
[65] S.C. Navale, V. Samuel, A.B. Gaikwad and V. Ravi Ceramics International 33, 297–299
(2007)
[66] S. Tsai, Y. Lu and M. Hon J. Phys.: Conference Series 100, 042030 (2008)
[67] Y. Y. Zheng, H. Yue, F. Qian, Z. J. Cheng, M. X. Hua and N. J. Yu, Sci China Ser E-Tech
Sci. vol. 52, no. 9 2762-2766 (2009)
[68] C. Wang, Y. Chiou, C. Chang, J. Tseng, L. Wu, C. Chen and T. Wu, J. Phys. D: Appl. Phys.
40, 1673–1677 (2007)
[69] M. Roodbari, M. Rezaee and N. Shahtahmasbi, World Appl. Sci. J., 7 (11): 1401-1403,
(2009)
[70] M. Voigt, M. Sokolowski, Materials Science and Engineering B 109, 99–103 (2004)
[71] J. Choi, J. Kim and T. Oh Mat. Res. Soc. Symp. Proc. Vol. 666 (2001)
[72] Q. Deen, M. Khosru, A. Nakajima and S. Yokoyama ICSE 2002 Proc. 402-406 (2002)
[73] E. H. Nicollian and J. R. Brews, “MOS physics and Technology”, Wiley, New York (1982)
[74] D. Y. Joh and W. W. Grannemann, Solid Surf. Electronics. Vol. pp 685-687 (1978)
[75] J. Song, H. Chan Oh, T. J. Park, C. S. Hwang,a, S. K. Park, S. M. Yoon, and C. Hwang J.
Electrochem. Soc., 155p 858-863 (2008)
50
[76] K.F. Albertain, I. Pereyra and M.I. Alayo Material Characterization 50, 149-154 (2003)
[77] E. H. Oulachgar, C. Aktik, M. Scarelte, S. Dostie, R. Sowerby and S. Gujrathi J. Appl.
Phys. 101, 084107 (2007)
[78] Y. S. Noh, S. Chatterjee, S. Nandi, S. K. Samanta, C. K. Maiti, S. Maikap and W. K. Choi
Microelectronic Engineering 66, 637-642 (2003)
[79] R. M. Fleming, D. V. Lang, C. D. W. Jones, M. L. Steigerwald, D. W. Murphy, G. B. Alers,
Y. H. Wong, R. B. van Dover, J. R. Kwo, A. M. sergent J. Appl. Phys. 88, 850 (2000)
[80] J. Siddiqui, E. Cagin, D. Chen and J. D. Philips Appl. Phys. Lette. 88, 212903 (2006)
[81] El H. Oulachgar, C. Aktik, M. Scarlete, S. Dostie, R. Sowerby, and S. Gujrathi J. Appl.
Phys 101, 084104 (2007)
[82] E.K. Kim, J.H. Kim H.K. Noh and Y.H. Kim J ELECTRONIC MATERIALS Vol. 35, No.4
512-515 (2006)
[83] R. C. Smith, T. Ma, N. Hoilien, L.y Y. Tsung, M. J. Bevan, L. Colombo, J. Roberts, S. A.
Campbell and W. L. Gladfelter, Adv. Mater. Opt. Electron. 10, 105-114 (2000)
[84] A. B. Selcuk, N. Tugluoglu, S. Karadeniz, S. Bilge and O. Physica B 400, 149-154 (2007)
[85] I. Yucedag, S. Altindal and A. Tataroglu, Microelectronic Engineering 84, 180-186 (2007)
[86] B. mereu, A. dimoulas, G. vellianitis, G. apostolopoulos, R. scholz, M. alexe, Appl. Phys. A
80, 253–257 (2005)
[87] E. H. Nicollian and A. Goetzberger, “The Si-SiO2 Interface-Electrical Properties as
Determined by MIS Conductance Techniqu”, Bell Syst. Tech. J., 46, 1055 (1976)
51
[88] I. Dokmea, S. Altındal and M. Gökçen, Physica B 393, 328–335 (2007)
52
CHAPTER 3
EXPERIMENTAL PROCEDURES
3.1 Introduction
As discussed in Chapter 2, zinc oxide has generated interest for use as the conducting
channel in transparent thin film transistors due to its optical transparency and compatibility with
low processing temperature. It was noted that the properties of zinc oxide are affected by its
processing parameters such as substrate temperature, deposition power, working gas, working
pressure, etc. J. N. Zeng et al. [1] grew ZnO thin films on Si (001) and quartz substrates by a
KrF excimer laser deposition. The wavelength of KrF excimer laser is 248 nm and the pulse
width is 23 ns. The laser power density was fixed at 100 mJ/cm2 and substrate temperatures
were changed from room temperature to 750 oC. The results of the sheet resistivity versus the
substrate temperature are shown in Figure 3-1[1]. The measurements showed an increase in the
sheet resistivity with increasing substrate temperature. J. N. Zeng suggested that the defect
electronic states such as zinc interstitials or oxygen vacancies are reduced when the films were
deposited at higher temperature. Here, ZnO thin films processed at low temperatures of between
100 oC and 200 oC to obtain enhanced electrical properties and also fit the requirements for low
temperature substrates (as mentioned in Chapter 2) are discussed.
53
Figure 3-1. The sheet resistivity of ZnO films as a function of substrate temperatures [1]
Transparent TFTs with ZnO as conductive channels can potentially be used as the drive
circuitry of organic light emitting diodes or other thin film electroluminescent devices. A critical
component of TFTs is the gate oxide, which must exhibit low leakage currents and self-healing
breakdown in order to ensure optimal switching performance and reliability of TFTs. As
mentioned in Chapter2, high-κ materials such as HfO2, ZrO2, Al2O3 and Ta2O5 [2-4] have been
intensively investigated as gate insulators for electronic devices. In general, low temperature
processes and materials are required to advance the emerging areas of transparent and flexible
electronics. However, replacing gate oxides with a high-κ material is still a challenge due to the
requirement that the material bulk and interface properties must be comparable to those of silicon
dioxide. As discussed in the previous chapter, the gate oxide is critical for switching
characteristics of a thin film transistor. If there are large trapped charge concentrations inside the
gate oxide, devices consume more power, in addition exhibit degraded switching performance.
54
Two insulator materials, namely, BaTa2O6 and Al2O3, are investigated in this study. The process
for each layer and fabrication of MIS devices are presented in this chapter.
3.2 Substrate Preparation
Corning 2974 glass was used for the ZnO thin films deposition by PLD, ALD and RF
magnetron sputtering. Glass/indium tin oxide (~20 Ω/square) substrates from Luminescence
Technology Corporation of Taiwan were used for the investigation of the properties of insulators
and MIS devices.
3.2.1 The Procedures for Substrate Cleaning
There are three steps for cleaning substrates. First, the substrates are ultrasonically
cleaned with acetone for 10 minutes and then rinsed by deionized water (DI water) for 30
seconds. Second, the substrates are ultrasonically cleaned with methanol for 10 minutes and then
rinsed by DI water for 30 seconds. Finally, the substrates are ultrasonically cleaned with DI
water for 10 minutes. After the ultrasonic cleaning, the substrates are blown by nitrogen gas and
then oven-dried for 15 minutes. The substrates were then loaded into a vacuum chamber for thin
films deposition.
3.3 Semiconductor Thin Films Growth
For the investigation of optimal electrical and optical properties of the ZnO thin films,
working pressure, power density, and substrate temperature were studied. The ZnO thin films
were grown by pulsed laser deposition, atomic layer deposition and RF magnetron sputtering.
55
3.3.1 Pulsed Laser Deposition
Pulse laser deposition (PLD) is a thin film growth technique which uses a high power
pulsed laser beam to ablate a target inside a vacuum chamber. The target material is vaporized
and expands and forms a plume containing many energetic species. Those species from the
target include atoms, molecules, electrons, ions, and clusters. The plume deposits onto a
substrate resulting in thin film growth. Figure 3-2 shows the layout of the pulsed laser
deposition system used in this work. The distance between the target and the substrate is 5cm.
In this study, the ZnO thin films are grown in a vacuum chamber, which can be evacuated
by a turbo-molecular pump pumping to a base pressure of 5.0×10-6 Torr. A KrF excimer laser
(Lambda Physik, COMPEX 201, laser wavelength of 248nm, pulse duration of 25ns, repetition
rate of 10 Hz) was used to ablate a ZnO target (diameter for 2 inches, purity 99.99%). The used
laser fluences were 1.06, 1.26, 1.54 and 1.98J/cm2, respectively. The films were grown at 10
mTorr of oxygen on Corning 2974 glass substrates. The substrate temperatures are 100 oC and
200 oC. The deposition time was varied as the laser power density at the ZnO target was
changed, in order to maintain a constant thickness of ZnO films. Before every thin film
deposition, the ZnO target was cleaned by the laser fluence for 1min 40 seconds (1000 shots).
After cleaning the surface of the ZnO target, the shutter was opened for deposition onto the
substrates.
56
Figure 3-2. The layout of the pulsed laser deposition system. The distance between the target and the substrate is 5cm
3.3.2 Atomic Layer Deposition
ZnO thin films are grown on Corning 2974 glass substrates by atomic layer deposition
(ALD) technique. In this study the ALD ZnO thin films growth were performed in cooperation
with Dr. Thomas W. Scharf’s research group, Department of Material Science and Engineering,
UNT.. The more detailed discussion of the synthesis and structural characteristics of ALD ZnO
can be found in Dr. Benedict A. Mensah’s dissertation [5], Department of Material Science and
Engineering, UNT. ALD thin film growth relies on two sequential surface reactions between
precursor molecules and a solid surface. The surface of the substrate is first exposed to a
reactant, which reacts with the initial surface sites. Then, after purging byproducts and excess of
this first previous reactant, the surface is exposed to the reactant. In this study, diethyl zinc, DEZ
57
[(C2H5)2Zn] and H2O, which are the respective precursors for zinc and an oxygen were
alternately pulsed into the ALD reactor using nitrogen as a carrier gas using separate inlet lines
and nozzles. A Cambridge NanoTech Savannah 100 ALD reactor was used in this work. An
image of the system is shown in Figure 3-3. In order to prevent gas-phase reactions that are
caused by intermixing of the precursors which are supplied via mass flow controllers with a
throughput of 20 sccm, nitrogen is also used to purge the reactant lines. The pulse times for DEZ
and H2O feed were the same at 15 msec, and the purge times were 8 sec, as shown in Table 3-1.
The substrate temperatures were 100 oC and 200 oC. The films were grown at a growth rate of
0.14 nm/cycle. The reaction sequence for ALD ZnO is as follows:
ZnOH* + Zn(CH2CH3)2 ZnOZn(CH2CH3)* + C2H6 (g)
ZnOZn(CH2CH3)* + H2O ZnOH* + C2H6 (g)
where * indicates surface species.
All of the settings such as the duration of the metal organic and the de-ionized water
pulses, the purge time and the deposition temperature are controlled by Lab view software. After
finishing the deposition, the ZnO films are cooled inside the chamber to room temperature.
58
Figure 3-3. An image of the Savannah 100 ALD system (Source: Cambridge Nanotech Inc)
Table 3-1. Deposition conditions for ALD ZnO thin film growth
Precursor Pulse time (s) Purge time (s) Deposition
temperature
Growth rate
(nm/cycle)
DEZ
DI H2O
0.015
0.015
8
8
100 oC and
200 oC 0.14
3.3.3 RF Magnetron Sputter Deposition
A typical sputtering system consists of a vacuum chamber with substrate holders and
magnetron sputtering guns, vacuum pumps, a gas supply system and RF power supplies. A
sketch for the sputtering chamber is shown in Figure 3-4. Sputtering is a process where atoms
are ejected from a target material due to bombardment by energetic particles, such as ionized Ar.
A detailed review of sputtering including the design of magnetron guns can be found in REF [6].
The sputtered ZnO thin films were prepared by RF magnetron sputtering technique on 2974
Corning glass. Prior to opening the substrate shutter for deposition, the surface of the ZnO target
59
was cleaned by exposure to the plasma for 10 minutes. The separation the substrate and the
target is 6 cm. The deposition pressure was maintained at 20 mTorr of Argon and the sputtering
power densities were varied at 4.44 W/cm2, 5.92W/cm2, 7.40W/cm2 and 8.88W/cm2. The
substrate temperatures here were 100oC and 200 oC as with the PLD and ALD grown films.
Figure 3-4. A sketch diagram of sputtering chamber
3.3.4 Post Deposition Annealing Treatment
Post deposition annealing was performed in controlled ambient conditions and specific
substrate temperature. During thermal annealing, thin films can react with the processing
ambient such as oxidizing or reducing gases. Crystallization and changes in grain size are
possible outcomes of post growth annealing. The ZnO thin films fabricated by PLD, ALD and
RF sputtering deposition are annealed in vacuum (5x10-5 Torr), Ar (100mtorr) and air,
respectively. The temperature was ramped from room temperature to 400oC at a rate of
60
10oC/min. After the temperature reached 400oC, the films are annealed for 2 hours. The
annealing curve is shown in Figure 3-5. The samples were allowed to cool down to room
temperature in the chamber before removal.
Figure 3-5. A sketch for annealing curve
3.4 Insulator Thin Film Growth
The insulator thin films were prepared by RF magnetron sputtering on ITO/glass
substrates. Two inch BaTa2O6 and Al2O3 targets with 99.99% purity are used for the insulator
thin film deposition. The separation from the substrate to the target was 6 cm. The base pressure
of the turbo pumped vacuum chamber was 5.0 x 10-6 Torr and prior to opening the substrate
shutter for deposition, the surface of the insulator targets are cleaned by exposure to the plasma
for 10 minutes. Deposition were performed at a pressure of 20 mTorr of an Argon/Oxygen (Ar:
O2 = 77%:23%) mixture and the substrates were at room temperature. The sputtering powers
used were 3.95W/cm2, 4.93W/cm2, 5.92W/cm2, 6.91W/cm2, 7.89 W/cm2and 8.88 W/cm2 for
BaTa2O6 thin films deposition. The sputtering power densities used for the Al2O3 thin films
deposition were 4.93W/cm2, 5.92W/cm2, 6.91W/cm2, 7.89 W/cm2, 8.88 W/cm2 and 9.87W/cm2.
61
3.5 Metal-Insulator-Semiconductor Device Fabrication
The ZnO and the insulators thin films that exhibited the best electrical/optical properties
from the conditions used above were incorporated into MIS structures. The fabrication steps of
MIS devices are as follows:
1. ITO glass substrates with 2 cm x 2 cm sizes were cleaned with the substrate cleaning
procedure described in section 3.2.1. A microscope slip cover was used to create a
step as shown in Figure 3-6 (a) for the thickness measurement by profilometry. ZnO
thin films were then deposited onto the ITO glass substrates, as shown in Fig 3-6 (b).
The electrical contact between the ZnO and ITO is Ohmic.
2. A slip cover was then put on the ZnO thin film and the insulator layer was deposited
on top of the ZnO thin films, as shown in Figure 3-6 (c).
3. After the insulator thin film was deposited on ZnO/ITO glass, the slip cover is
removed, which allows access to the ZnO underneath. A shadow mask is then used to
deposit metal contacts onto the insulator as Figure 3-7shows. The metal typically used
was thermally evaporated silver. The based material was silver shot with a particle
size of 1-3mm and 9.9999% purity. A boat used for Ag deposition is shown in Figure
3-8. The boat (R. D. Mathis Company) is connected to a 250A Sorensen DC power
supply that is controlled by a computer to two copper electrodes, at heated by passing
a high current through it (Joule heating). The deposition rate was 2A/s and the
thickness of the Ag was typically ~150nm for all depositions. A schematic of the
complete fabricated MIS structure is shown in Figure3-10.
62
Figure 3-6. (a) A mask on ITO glass substrate (b) ZnO thin film on ITO glass substrate (c) mask on ZnO thin film for insulator deposition
63
Figure 3-7. (a) a mask on the insulator/ZnO/ITO glass (b) a top view
Figure 3-8. The boat for the Ag deposition
64
Figure 3-9. The schematic deposition of the thermal evaporation for silver layer on samples
Figure 3-10. A schematic structure of MIS
3.6 Materials Characterization
3.6.1 Surface Profilometry
Topography and thickness of the thin films were measured with a Veeco Dektak 150
stylus profilometer. As described above, a slip cover was used to create edges for measuring the
topography and thickness. Figure 3-11 shows the architecture of a thin film and the measuring
arrangement with a stylus profilometer.
65
Figure 3-11. A schematic of the architecture of thin films with steps and the measuring arrangement with a stylus profilometer
3.6.2 Scanning Electron Microscopy (SEM)
A FEI Nova 200 Dual Beam system equipped with a 30kV SEM (scanning electron
microscope) column and a 30 kV FIB (focused ion beam) column with a gallium source was
used for SEM analysis. The samples were milled with an accelerated Ga+ ion beam through the
silver to the ZnO layer. The milled area was 10 μm x 4 μm and milling depths of around 600 nm
reached to the glass substrates. Images of interface between the insulator and ZnO were then
scanned by SEM.
3.6.3 Structure Properties Using X-ray Diffraction
X-ray diffraction (XRD) analysis of the films was performed by using the 1.54056 Å Cu
K1 line of a Rigaku Ultima III X-ray diffractometer. To prevent measurements of the substrate,
a glancing angle of 5 degrees was used for the ZnO thin films, and 0.5 degrees for the BaTa2O6
and Al2O3 films.
66
3.6.4 X-ray Photoelectron Spectrometer (XPS)
A PHI 5000 VersaProbeTM Scanning X-ray Microprobe Spectrometer was used to
analyze the chemical bond states and the composition of the materials. XPS uses a
monochromatic X-ray beam (Al Kα1, 1486.6eV) to probe the sample surface. The chemical
composition of the sample surface are analyzed by using the X-ray source of 10.9W, a pass
energy of 58.7 eV, a 45o take-off angle and a scan rate of 0.5 eV/step. To clean the surface of
samples, Ar+ sputtering is used to remove adventitious carbon on the surface. The depth profile
was facilitated by sputter removal of the material using Ar+ ions accelerated at 2 keV with a
beam current of 2 μA.
3.6.5 Optical Properties - Photoluminescence (PL) and Photoluminescence Excitation (PLE)
and Transmittance
The photoluminescence (PL) and photoluminescence excitation (PLE) characteristics of
the specimens are determined using a reconfigurable optical bench which includes a xenon lamp,
two ¼ meter monochromators with appropriate gratings, reflective optics, and photomultiplier
tube and photodiode detectors. Lock-in detection and a chopper is used to minimize background
signals, and the system is integrated via Lab View software. A schematic diagram of the setup is
shown in Figure 3-12.
For a photoluminescence measurement monochromator 2 allows selection of the required
excitation wavelength from the Xenon lamp and photoluminescence generated in the specimens
is collected and measured with monochromator 1 and the detector. For measuring
photoluminescence excitation (detecting the best excitation wavelength for the specimens),
detecting monochromator 1 is set to the peak photoluminescence wavelength and the excitation
67
wavelength is scanned with monochromator 2. For example, one of the known emission
wavelengths of undoped ZnO is around 370 nm, as shown in Figure 3-13 [7]. Monochromator 1
would be set to 370nm, and the excitation wavelength scanned with monochromator 2 to
determine the best optimal excitation wavelength. After the optimal excitation wavelength is
known, it is used to excite the specimen in order to obtain the strongest PL intensity.
Figure 3-12. A schematic diagram of photoluminescence measurement
68
Figure 3-13. PL spectra at room temperature of ZnO thin films grown under different oxygen partial pressures. The emission wavelength was around 370 nm [7]
The transmittance is defined by the equation:
, where Io is the intensity of
the incident light and I is the intensity of the light transmitted through the sample. In this study,
Io is the incident light through a blank glass (Corning 2974) substrate, and I is the light
transmitted through the glass substrate onto which the ZnO or insulator thin films were deposited.
3.6.6 Optical Properties-Ellipsometry
The refractive indices of the thin films are with a J.A. Woollam M-2000 Ellipsometer.
The fundamentals of ellipsometry are well understood. In summary, plane-polarized light
impinges a sample at a certain angle, and the change of polarization of the reflected beam is
analyzed. The optical constants are determined from the measured relative phase change, ∆ and
the relative amplitude change, ψ, introduced by reflection from the surface. More specifically, ψ
and ∆ provides the ratio between the s-polarized (rs) and p-polarized (rp) light reflected off the
surface of interest [8];
69
i
s
pe
r
r).tan( (1)
which directly relates to the complex index of refraction through Fresnel’s Relations;
tl
tlp
n
nr
coscoscoscos
tl
tls
n
nr
coscoscoscos
(2)
where rp and rs are the p and s components of the reflected polarized light, and n is the complex
index of refraction. Both Ψ and Δ are wavelength dependent. Once the real (n) and imaginary
components (k) of the complex index of refraction have been measured, the thickness
determination is rather straightforward since the amplitude of the electrical component of the
propagating light wave is:
c
dntid
c
kx
expexp0
(3)
where d is thickness and t is time.
To determine the refractive index of thin films, 55o, 60o and 65o angles of the incident
light were used. The wavelength range used was 400 nm to 700 nm with a step of 5 nm. The
system recorded the spectra of psi (ψ) and delta (Δ) as function of wavelength (λ), then
calculated the complex refractive index.
70
Figure 3-14. A schematic setup of an ellipsometry experiment.
3.6.7 Electrical Properties-Hall Effect Measurement and Current-Voltage (I-V)
The Hall Effect is used to determine the resistivity, the carrier type, the carrier
concentration and the mobility of a specimen [9]. If an electrical current flows through a
conductive specimen in a magnetic field, the magnetic field produces the Lorentz force:
where represents the electron charge, the velocity and the magnetic field. The Lorentz
force deflects the moving carriers and pushes them to one side of the conductive specimen. The
deflected direction is determined by the type of carriers and the cross product of . The
deflected carriers create a potential, known as the Hall voltage, VH [9]. The Hall voltage is
defined as
, where I is the electrical current, t is the thickness of the specimen, n is
the carrier concentration. A Hall coefficient is defined as
The sign of the Hall coefficient indicates the dominant carrier type. A negative Hall coefficient
indicates that electrons are the dominant carriers. A positive Hall coefficient represents that
holes are the dominant carriers. The Hall effect mobility, μH, can be obtained by
which
71
is related to the conductivity or resistivity of carriers. A Hall Effect measurement requires four
Ohmic contacts to the samples, which are usually in the Van der Pauw arrangement. These are
labeled 1, 2, 3, and 4 counterclockwise as shown in Figure 3-15 (a). The film resistivity, ρ, for
example, is first assessed using, ρ
, where V43 and V14 are defined as the
potential difference between V4-V3 and V1-V4, respectively. I12 and I23 indicate the current flow
in the specimen entering through contact 1 and leaving through contact 2, and through contact 2
to contact 3, respectively. A schematic of the configuration is shown in Figure 3-15 (b).
Figure 3-15. (a) The common geometry for contacts for Hall effect measurements. The labels 1-4 are the contacts. (b) Schematic of current and voltage measurement [9]
72
The equations describing the measurements can be expressed as
,
where
In this study, an Ecopia HM-S5000 Hall effect measurement system using a current of 1 μA, and
a magnetic field of 0.5 Tesla was employed for electrical characterization of the ZnO films.
For the dielectric strength of insulators, MIS structures were biased into dielectric
breakdown by applying a reverse electric field to the specimens. Figure 3-16 shows an example
of I-V measurements of BaTa2O6 [10]. The breakdown strength of BaTa2O6 thin films was
around 1 MV/cm.
Figure 3-16. A I-V measurement for BaTa2O6 thin films by sputtering deposition [10]
The breakdown characteristics of the BaTa2O6 and Al2O3 were determined using a Keithley
2420 source-measure unit integrated with a Micromanipulator probe station via LabView.
73
3.7 Characterization for MIS Devices
3.7.1 Capacitance-Voltage (C-V) Measurements
The use of capacitance-voltage (C-V) measurements for the characterization
semiconductor devices was discussed in some detail in the previous section. In this study,
frequency dependent C-V and conductance-voltage (G-V) measurements are performed using an
Agilent 4294A impedance analyzer integrated with a Micromanipulator probe station.
An ac oscillation (probe) voltage of 100 mV and frequency of 1MHz were use. The dc
bias range is determined by the inversion and accumulation regions for each MIS device and is
swept from a negative voltage to a positive voltage. Figure 3-17 [11] shows an example of a C-
V measurement where the dc bias was varied from the negative to the positive voltage for a MIS
with n-type semiconductor at 1 MHz. The capacitance of the oxide layer can be determined
from the region of strong accumulation which is around 350 pF in this example.
Figure 3-17. A typical C-V measurement of a MIS with n-type semiconductor with the bias swept from the inversion region to the accumulation region [11]
74
3.7.2 Conductance-Voltage (G-V) Measurements
As previously mentioned, both the capacitance and the conductance as functions of
voltage and frequency contain identical information about interface traps. However, the
conductance is directly related to the interface traps and yields more accurate and reliable results
[12]. The measurement setup was the same as the capacitance measurement (the ac oscillation
voltage was 100 mV and frequency was 1MHz). The dc bias range is determined by the
inversion and accumulation regions for each MIS device and swept from a negative voltage to a
positive voltage.
3.7.3 Hysteresis Curve from C-V Measurements
The interface charge can be estimated from the in a C-V measurement as discussed in
Chapter 2. For the hysteresis curve measurement, the C-V measurement was swept from the the
inversion region to the accumulation region, then from the accumulation region to the inversion
region. The ac oscillation voltage was set to 100 mV and the frequency used was 1MHz.
3.7.4 Frequency Dispersion from C-V Measurements
Due to the strong dependence of MIS device capacitance on frequency, C-V
measurements were performed at different frequencies. In this study, to estimate the influence
of frequency dispersion in the MIS devices, C-V measurements were performed at frequencies of
500 kHz, 1 MHz and 2 MHz with an ac voltage probe of 100mV.
75
3.8 References
[1] J. N. Zeng, J. K. Low, Z. M. Ren, T. Liew and Y. F. Lu, Appl.d Surf. Sci. Vol.197-198, pp.
362-367 (2002)
[2] H. K. Tyagi and P. J. George J. Mater. Sci.: Materials in Electronics Vol.19, No. 8-9, 902-
907 (2007)
[3] E. Gerritsen, N. Emonet, C. Caillat, N. Jourdan, M. Piazza, D. Fraboulet, B. Boeck, A.
Berthelot, S. Smith and P. Mazoyer, Solid-State Electronics 49, 1767–1775 (2005)
[4] Y.S. Noh , S. Chatterjee , S. Nandi , S. K. Samanta , C.K. Maiti , S. Maikap and W. K. Choi
Microelectronic Engineering 66, 637–642 (2003)
[5 ] B. Mensah, dissertation “Growth, Structure and Tribological Properties of Atomic Layer
Deposited Lubricious Oxide Nanolaminates”, University of North Texas (2010)
[6 ] M. H. Francombe and J. L. Vossen, “Thin Films for Advanced Electronic Devices”, Vol. 15,
p 152 (1991)
[7] Y. Ma, G. T. Du, S. R. Yang, Z. T. Li, B. J. Zhao, X. T. Yang, T. P. Yang, Y. T. Zhang, and
D. L. Liu J Appl. Phys. Vol.95 No.11, 6268-6271(2004)
[8 ] H. Fujiwara “Spectroscopic Ellipsometry: Principles and Applications” John Wiley & Sons,
(2007)
[9] E. H. Hall, American Journal of Mathematics Vol.2, p287-292 (1879)
[10] Y. Lee, Y. Kim, D. Kim, B. and M. Oh, IEEE Transactions on electron devices, vol. 47, No.
1(2000)
76
[11] S. M. Chore, G. N. Chaudhari, S. V. Manorama and A. Bath, Semicond. Sci.
Technol. 17, No. 11, 1141-1143(2002)
[12] S.M. Sez and Kwok K. Ng “Physicas of semiconductor devies” (2006)
77
CHAPTER 4
RESULTS: THE PROPERTIES OF ZINC OXIDE, BARIUM TANTALATE AND
ALUMINUM OXIDE1
4.1 Properties of ZnO Thin Films by PLD, ALD and Sputtering
As mentioned in Chapter 3, three techniques (pulsed laser deposition, atomic layer
deposition and RF magnetron sputtering deposition) were used to fabricate ZnO thin films in
order to evaluate which method resulted in the best combination of electrical and optical
properties. The structural, optical and electrical properties of the resultant ZnO thin films are
discussed in this chapter.
4.1.1 Structural Properties and Compositional Analysis
-Pulsed Laser Deposited ZnO Thin Films
Figure 4-1 is the XRD spectra of the as-deposited PLD ZnO films grown at 100°C and
200°C with a laser power density of 1.04 mJ/cm2 and 10 mTorr of oxygen. The oxygen
diffusion is depending on the growth temperature and inversely proportional to square root of
growth temperature. However, the growth temperatures here are relative low, so the oxygen
diffusion coefficients for 100oC and 200oC are considered similar which in the range of 10−15–
10−17 cm2/s. XRD confirmed that PLD ZnO thin films grown at 100ºC and 200°C are
polycrystalline (wurtzite crystal structure). PLD ZnO thin films exhibit strong (002) texturing
which is often called c-axis orientation. The c-axis orientation is commonly observed in ZnO
1 Results presented in this chapter have been published in Physica Status Solidi (a) Vol. 207, Issue 11 (2010) 2487 and are used with permission from Wiley-VCH Verlag GmbH & Co. KGaA.
78
thin films because the c-plane which is perpendicular to the substrate is the most densely packed
and thermodynamically preferred in the wurtzite structure [1]. In addition, there is another weak
peak appearing at 2θ = 62.2o as shown in Figure 4-1, which corresponds to ZnO (103). During
the growth, the kinetics is mainly determined by the substrate temperature when the laser power
density is fixed. At a relatively higher temperature, the atoms on the substrate surface can
diffuse quickly to find the highest energy sites, and form the low energy structure [2]. With the
assumption of homogeneous strain, the average grain size D of (002) peak can be estimated from
the well-known Debye-Scherrer formula [3], that is
D = kλ /Bcosθ---------Eq (4-1)
where is the x-ray wavelength (1.54056 Å for the Cu K1 line), B and are the FWHM (the
full width at half maximum) and the Bragg diffraction angle of the dominant XRD peak, and k is
a constant equal to 0.9. Accounting for instrument broadening, the grain size of the PLD ZnO
thin film grown at 100°C is 16.58 nm, which increases slightly to 18.20 nm when the substrate
temperature is 200°C. The crystallinity of the ZnO at 100°C and 200 °C are comparable and
because of the low temperature process requirement for compatibility with plastic substrates
mentioned in Chapter 2, the substrate temperature of 100°C deemed to be appropriate.
The laser power density was then varied with the substrate temperature being constant at
100°C. Figure 4-2 is the XRD spectra of the as-deposited PLD ZnO thin films grown at 100°C
with the laser power densities of 1.04, 1.24, 1.54 and 1.94 mJ/cm2 respectively. All of the PLD
ZnO thin films exhibit strong (002) texturing. Basal plane growth appears to be
thermodynamically favorable in the energetic (due to atomic peening) pulsed laser deposition
process. The orientation of PLD ZnO thin films are not affected by the change of laser power
79
density, while the intensity of (002) peak increased slightly with the increase of the laser power
density.
20 30 40 50 60 70 80 90
62.24 (103)
34.14 (002)
Inte
ns
ity
[a
.u.]
2 thelta [degrees]
As grown PLD ZnO ST 100
As grown PLD ZnO ST 200
Figure 4-1. X-ray diffraction spectra of PLD ZnO thin films grown at 100 oC (bottom), and 200 oC (top). The peaks are labeled in brackets and their 2θ angles given below
80
20 30 40 50 60 70 80 90
62.24 (103)
34.14 (002)
Inte
ns
ity
[a
.u.]
PLD 10 mtorr of O2
1.98 J/cm2
1.54 J/cm2
1.24 J/cm2
1.04 J/cm2
2 thelta [degrees]
Figure 4-2. X-ray diffraction spectra of PLD ZnO thin films grown at 100 oC with the laser power densities of 1.04 (black), 1.24 (red), 1.54 (blue) and 1.94 (green) mJ/cm2. The peaks are labeled in brackets and their 2θ angles given below
The PLD ZnO thin films deposited at 100°C and 1.04 mJ/cm2 were annealed in different
ambient conditions. Figure 4-3 presents the XRD spectra of the as-deposited PLD ZnO thin
films and after annealing at 400 °C in air, vacuum (5x10-5 Torr) and Ar (1 Torr). With the
assumption of homogeneous strain, a calculation of grain sizes for the as-deposited PLD ZnO
thin film and the films annealed in Ar, vacuum and Air is shown in Table 4-1. The average grain
size of the PLD samples synthesized at 100 oC grew from 16.58 nm to 25.93 nm after post
deposition annealing at 400 oC in 1 Torr of Ar, which represents a 56.39 % increase. The post
deposition annealing in vacuum resulted in a 38.42 % increase from 16.58 nm to 22.95 nm. The
grain size of PLD ZnO annealed in air is 26.66 nm, which is an increase of 60.80% from 16.58
nm.
81
Understanding the structure mentioned above and correlation with the composition would
provide deeper insight to the processing-structure-properties of the ZnO thin films. Therefore,
XPS (X-ray Photoelectron spectra) measurements were performed. Table 4-2 presents the Zn/O
ratio determined by XPS for the as-deposited PLD ZnO thin films, and after annealing in Ar,
vacuum and air. As Table 4-2 shows, all of the films are oxygen deficient, and while annealing
in Ar and vacuum slightly increases the Zn/O ratio, annealing in air has the opposite effect and
clearly increases the oxygen content. This information provides some insights and a tentative
correlation between composition, electrical conductivity and optical properties. We shall return
to this discussion below after the electrical properties and the photoluminescence data of all ZnO
thin films by PLD, ALD and sputtering have been presented.
20 30 40 50 60 70 80 90
62.24 (103)
Inte
ns
ity
[a
.u.]
2 thelta [degrees]
34.14 (002)
As grown
Annealing in Ar
Annealing in vacuum
Annealing in air
PLD ZnO
Figure 4-3. XRD spectra of PLD ZnO as-deposited at 100oC (black) and annealed in Ar (red), vacuum (blue) and air (green) at 400 oC for 2 hours
82
Table 4-1. Grain sizes of PLD ZnO as-deposited at 100 oC and annealed in Ar, vacuum and air
Grain Size (nm)
Grain size increasing
percentage (%)
As grown PLD ST
100 16.58 ---
Annealed PLD ST
100 in Ar 25.93 56.39%
PLD ST 100 in
vacuum 22.95 38.42%
PLD ST 100 in air 26.66 60.80%
Table 4-2. The Zn/O ratio determined by XPS for PLD ZnO films in the as-deposited condition, and after annealing in Ar, vacuum and air
XPS (Zn/O)
As deposited
Annealed in Ar
Annealed in vacuum
Annealed in air
PLD ZnO 1.17±0.1 1.21±0.1 1.18±0.1 1.10±0.1
- Atomic Layer Deposition for ZnO Thin Films
XRD spectra of the as-deposited ALD ZnO films grown at substrate temperatures of
100°C and 200°C are shown in Figure 4-4. The films grown at 100 and 200ºC are
polycrystalline (wurtzite crystal structure). The investigated angular region (2θ = 20o-90o)
include the obvious peaks corresponding to 31.8o (100), 34.4o (002), 36.2o (101) and 56.7o (110).
Both of the as-deposited films (at 100ºC and 200ºC) show predominantly prismatic (100)
texturing, or preferred orientation, of the ZnO grains. The intensity ratio (I100/I002) of the (100)
peak compared to the (002) is 2.05 and 2.38 for the films grown at 100 ºC and 200 ºC
respectively. When the deposition temperature is increased to 200 ºC, the ZnO grains become
83
more randomly orientated, since the (101) peak grew to approximately the intensity of the (100).
The grain size of the ALD ZnO (100) increased slightly from 20.24 nm to 24.70 nm when the
substrate temperature is increased from 100 to 200 ºC.
ALD ZnO thin films deposited at the substrate temperature 100 ºC were then studied as a
function of different annealing conditions. Figure 4-5 is the XRD spectra of the as-deposited
ALD ZnO thin films at 100 ºC, and annealed at 400 ºC in Ar (1 Torr), vacuum (5x10-5 Torr) and
air. The films deposited by ALD exhibit an invariance of 2θ in their respective X-ray spectra
before and after annealing, which suggests that possible processing induced micro-strain if
present is homogeneously distributed throughout the films. With the assumption of
homogeneous strain, the average grain size D can be estimated. Table 4-3 shows the grain size
of ALD ZnO thin films before and after annealing. After annealing in Ar of 1 Torr, the average
grain size of the ALD ZnO samples synthesized at 100oC grew from 20.24 nm to 32.02 nm,
which represents a 58.20% increase. Annealing in vacuum resulted in a 40.32 % increase in
grain size from 20.24 nm to 28.4 nm. The grain size of ALD ZnO thin film annealed in air is
29.98 nm, which presents a 48.12% increase from 20.24 nm.
Table 4-4 presents XPS measurements of the composition of the ALD ZnO thin films.
This shows that all of the films were oxygen deficient. The Zn/O ratio of the as-deposited ZnO
thin film is 1.17 and slightly increases to 1.18 and 1.19 when ALD ZnO was annealed in Ar and
vacuum, respectively. When ALD ZnO was annealed in air, the result was similar to the PLD
ZnO thin film annealed in air; that is the oxygen content increased.
84
20 30 40 50 60 70 80 90
56.7 (110)36.2(101)
34.4(002)
31.8(100)
Inte
ns
ity
[a
.u.]
2 thelta [degrees]
As grown ALD ZnO ST 100
As grown ALD ZnO ST 200
Figure 4-4. X-ray diffraction spectra of ALD ZnO grown at 100 oC (bottom), and 200 oC (top). The peaks are labeled in brackets and their 2θ angles given below
20 30 40 50 60 70 80 90
As grown
Annealing in Ar
Annealing in vacuum
Annealing in air
56.7 (110)36.2(101)
34.4(002)
31.8(100)
ALD ZnO
Inte
ns
ity
[a
.u.]
2 thelta [degrees]
Figure 4-5. XRD spectra of ALD ZnO as-deposited at 100oC (black) and annealed in Ar (red), vacuum (blue) and air (green) at 400 oC for 2 hours
85
Table 4-3. Grain sizes of ALD ZnO as-deposited at 100 oC and after annealing in Ar, vacuum and air
Grain Size
(nm)
Grain size increasing
percentage
As grown
ALD ST 100 20.24
ALD ST 100
in Ar 32.02 58.20%
ALD ST 100
in vacuum 28.4 40.32%
ALD ST 100
in air 29.98 48.12%
Table 4-4. The Zn/O ratio determined by XPS for ALD ZnO films in the as-deposited condition, and after annealing in Ar, vacuum and air
XPS (Zn/O)
AS deposited
Annealed in Ar
Annealed in vacuum
Annealed in air
ALD ZnO 1.17±0.1 1.18±0.1 1.19±0.1 1.15±0.1
-RF Magnetron Sputtering for ZnO
The sputtered ZnO films were prepared by RF magnetron sputtering from a ZnO target in
20 mtorr of Ar. Figure 4.5 shows the XRD spectra of the sputtered ZnO thin films grown at
100°C and 200°C with power density of 4.44 W/cm2. Both films show a polycrystalline
(wurtzite) crystal structure. The film sputtered at 100 °C exhibits strong (002) texturing, but the
film grown at 200°C is more randomly oriented. The grain size of the sputtered ZnO increased
slightly from 17.42 nm to 18.88 nm when the substrate growth temperature increased from 100
ºC to 200 ºC. The crystallinity of ZnO thin films at 100°C and 200 °C are comparable. However,
since films that were sputter grown at 100 °C showed a stronger (002) peak than the one
86
produced at 200°C, a substrate temperature of 100°C was used in subsequent process
optimization
With the substrate temperature fixed at 100 ºC, the sputtering power densities were varied.
Figure 4-6 is the XRD spectra of that were sputter grown at power densities of 4.44 W/cm2, 5.92
W/cm2, 7.40 W/cm2 and 8.88 W/cm2 respectively. All of the sputtered ZnO films exhibit strong
(002) texturing. As shown in Figure 4-6, the orientations of sputtered ZnO films were not
affected by the change of sputtering power densities.
20 30 40 50 60 70 80 90
ST 100
Inte
ns
ity
[a
.u.]
2 thelta [degrees]
Sputtering ZnO at 4.44 W/cm2
ST 200
31.8(100)
34.2(002)
36.1(101)
67.462.355.746.9
Figure 4-6. X-ray diffraction spectra of sputtered ZnO grown at 100 oC (bottom), and 200 oC (top). The peaks are labeled in brackets and their 2θ angles given below
87
20 30 40 50 60 70 80 90
Sputtering ZnO 20 mtorr of Ar
5.92W/cm2
7.40W/cm2
8.88W/cm2
Inte
ns
ity
[a
.u.]
2 thelta [degrees]
4.44W/cm2
34.2 (002)
Figure 4-7. X-ray diffraction spectra of sputtered ZnO grown at 100 oC with sputtering power density of 4.44 (black), 5.92 (red), 7.40 (blue) and 8.88 (green) W/cm2. The peaks are labeled in brackets and their 2θ angles given below
Figure 4-8 is the XRD spectra of sputtered ZnO thin films deposited at 100 ºC and
annealed at 400 ºC in Ar (1 Torr), vacuum (5x10-5 Torr) and air. With the assumption of
homogeneous strain, the average grain size D can be estimated. Table 4-5 shows the grain size
of the sputtered ZnO thin films before and after annealing. The average grain size of the
sputtered ZnO sample grown at 100oC grew from 17.42 nm to 28.57 nm after annealing in 1 Torr
of Ar, which represents a 64.01% increase. The sputtered ZnO thin film annealed in vacuum
resulted in a 59.53 % increase from 17.42 nm to 27.79 nm. The grain size of the sputtered ZnO
thin film annealed in air is 29.93 nm, which presents a 71.81% increase from 17.42 nm.
The composition of the sputtered ZnO thin films from XPS measurement is shown in
Table 4-6. Table 4-6 shows that the Zn/O ratios of the sputtered ZnO thin films have a similar
trend to the ZnO deposited by PLD and ALD. All of the films show oxygen deficiency, but
88
when they were annealed in Ar and vacuum there were slight increases in the Zn/O ratios.
Annealing in air had the opposite effect and clearly increased the oxygen content. The Zn/O
ratio for the as-deposited sputtered ZnO thin film is 1.23. After the ZnO thin films were
annealed in Ar and vacuum, the Zn/O ratios increased to1.30 for both of them. The Zn/O ratio
for the ZnO annealed in air decreased to 1.09.
20 30 40 50 60 70 80 90
Annealing in air
Annealing in vacuum
Annealing in Ar
As grown
Inte
ns
ity
[a
.u.]
2 thelta [degrees]
Sputtered ZnO with 4.44 W/cm2
34.2(002)
31.6(100)
Figure 4-8. XRD spectra of sputtered ZnO as-deposited at 100oC (black) and annealed in Ar (red), vacuum (blue) and air (green) at 400 oC for 2 hours
89
Table 4-5. Gain sizes of sputtered ZnO as-deposited at 100 oC and after annealing in Ar, vacuum and air at 400 oC
Grain Size (nm)
Grain size increasing
percentage
As grown ST 100 17.42
ST 100 in Ar 28.57 64.01%
ST 100 in vacuum 27.79 59.53%
ST 100 in air 29.93 71.81%
Table 4-6. The Zn/O ratio determined by XPS for sputtered ZnO films in the as-deposited condition, and after annealing in Ar, vacuum and air
XPS (Zn/O)
As deposited
Annealed in Ar
Annealed in vacuum
Annealed in air
Sputtered ZnO
1.23±0.1 1.30±0.1 1.30±0.1 1.09±0.1
4.1.2 Electrical Properties
Hall Effect measurements were carried out to investigate the electrical properties of the
ZnO thin films. An Ecopia HM-S5000 Hall Effect system measured each sample 10 times, from
which average values are calculated. The results are shown in Table 4-7 for all of the ZnO thin
films grown at 100 oC and 200 oC by PLD, ALD and RF sputtering. All of the signs of the
carrier concentrations are negative, which show that all of the ZnO thin films are n-type
semiconductors. The ALD ZnO thin film grown at 100 oC has the lowest carrier concentration
(Ne, cm-3), smallest mobility (μ, cm2/V.s ), and highest resistivity (ρ, Ω.cm). The specific Ne of
the ALD ZnO thin film grown at 100 oC was 1.81x1016 cm-3, which increased to 8.85x1017 cm-3
90
at 200 oC. The mobility of carriers of the ALD ZnO thin film increases from 1.14 to 16.51
cm2/V.s for the 100 oC and 200 oC growth temperatures. It has been proposed that trapped
hydroxyl groups and associated deep levels lead to relatively poor electrical properties in ALD
ZnO thin films processed at low temperatures. Hall effect measurements showed that the free
carrier concentration varied between 1019 and 1017 cm-3 while the conductivity remained at 10-1
Ω-1cm-[4]. The increased carrier concentrations and the electron mobility together mentioned
above suggest that higher substrate temperature leads to a reduction in non-radiative electronic
trap concentrations. This author proposes that the grain coarsening results in diminished grain
boundary scattering which also contributes to the improved electrical behavior observed at 200
oC compared to 100 oC. Similar results are obtained in PLD ZnO thin films. The carrier
concentration and resistivity of the PLD ZnO at 100 oC and 200 oC are 1.11x1017 cm-3, 3.85
Ω.cm and 7.61x1017 cm-3, 0.47 Ω.cm, respectively. The mobility of carriers in the PLD ZnO thin
film grown at 200 oC is 17.42 cm2/V.s, which is slightly higher than the 14.6 cm2/V.s measured
for PLD ZnO thin film fabricated at 100 oC. The electrical properties of ZnO films grown by
ALD and PLD at 200 oC are comparable. In comparison, the sputtered ZnO thin films grown at
100 oC and 200 oC are characterized by significantly larger Ne, μ and smaller ρ values. However,
there is no significant influence of the substrate temperature on electrical properties of sputtered
ZnO thin films. The carrier concentration Ne, mobility, μ and resistivity, ρ values for the
sputtered ZnO sample at 100 oC are 3.59x1018 cm-3, 35.21 cm2/V.s and 0.05 Ω.cm, respectively.
The carrier concentration, Ne, mobility, μ and resistivity, ρ values for the sputtered ZnO sample
at 200 oC were 4.24x1018 cm-3, 32.71 cm2/V.s and 0.048 Ω.cm, respectively. Thus, the electrical
properties of the ZnO thin films at sputtered 100oC and 200 oC are comparable.
91
Table 4-7. Electrical properties of PLD, ALD and sputtered ZnO thin films grown at 100 oC and 200 oC from Hall Effect measurement. The error percentage is 6% for the measurements.
Sample Carrier concentration, Ne [ cm-3]
Mobility, μ
[cm2/Vs]
Resistivity, ρ
[ Ωcm]
PLD 100 oC as-deposited -1.11x1017 14.60 3.850
PLD 200 oC as-deposited -7.61x1017 17.42 0.470 ALD 100 oC as-deposited -1.81x1016 1.14 303.000 ALD 200 oC as-deposited -8.85x1017 16.51 0.420 Sputtered 100 oC as-deposited -3.59x1018 35.21 0.050 Sputtered 200 oC as-deposited -4.24x1018 32.71 0.048
The electrical properties of as-deposited PLD, ALD and sputtered ZnO are compared
with their properties after annealing in Table 4-8. All of the samples exhibited n-type electrical
conduction, and overall, post deposition annealing resulted in an increase of free carrier
concentration (Ne, cm-3), and electron mobility (μ, cm2/V.s), except for the films that were
annealed in air which became resistive. The change of conductivity for annealed ZnO films
could be related to the oxygen molecular density during the processing. According to the defect
reactions mentioned in Chapter 2, the interstitial zinc can be described as
. The equilibrium constant can be expressed as
[ ][ ]
[ ]
(K: the
equilibrium constant, [ ] : the concentration of the interstitial zinc, [ ] the concentration of
electrons, p: the partial pressure of oxygen). The concentration of electrons [ ] is inversely
proportion to the oxygen pressure , so the conductivity is inversely proportion to the oxygen
pressure . The oxygen molecular density for the annealing temperature, 400oC, Ar (1 Torr),
vacuum (10-5 Torr) and air are 1.58E+15/L, 1.58E+15/L and 2.18E+21/L, respectively. The ZnO
92
films which were annealed in air could get the most of oxygen molecular, therefore the ZnO
films which are annealed in air become resistive as shown in Table 4-8. These results are
consistent with the literature, which reports carrier concentrations of 1015-1017cm-3and resistivity
of 10-2-103 Ω.cm for undoped ZnO thin films grown by ALD at ~100 oC [ 5 , 6 ]. The
corresponding reported values for PLD films grown at ~100 oC range between 1017-1019cm-3 and
10-2-10-3Ω.cm respectively [7,8]. For the sputtered ZnO thin films, they are in the range of 1017-
1019cm-3 and 10-2-10-3Ω.cm respectively [9,10]. The ALD films grown at 100 oC has a free
carrier concentration of 1.81x1016cm-3, which increased by more than an order of magnitude with
post deposition annealing in vacuum to 2.79x1017cm-3. The mobility also improved noticeably
from 1.14 to 15.75 cm2/V.s with annealing in vacuum, and the resistivity decreased from 303 to
1.61 Ω.cm. As Table 4-8 shows, annealing in Ar also enhanced the electrical characteristics of
the ALD ZnO as-deposited films, although to a lesser extent. In comparison as Table 4-8 shows,
the sputtered ZnO thin film grown at 100 oC is generally characterized by a larger Ne, μ and σ
values both in their as-deposited condition and after post deposition annealing in Ar and vacuum.
These measurements show that of the PLD, ALD and sputtered ZnO thin films grown at 100 oC,
the sputtered ZnO thin exhibits the best electrical properties.
The improved carrier concentrations and electron mobility together with the observed
grain coarsening suggest that post deposition annealing will lead to a reduction in non-radiative
electronic trap concentrations, and reduced incoherent scattering from grain boundaries and
ionized impurities. We shall return to this discussion below after the photoluminescence data is
presented to clarify the proposed conductive mechanism of the ZnO films.
93
Table 4-8. Electrical properties of as-deposited PLD, ALD and sputtered ZnO and annealed ZnO thin films by Hall Effect measurement. The error percentage is 6% for the measurements.
Sample
Carrier concentration, Ne Mobility, μ Resistivity, ρ
[ cm-3] [cm2/V.s] [ Ω.cm] PLD ZnO as-deposited -1.11x1017 14.6 0.387 PLD ZnO annealed in Ar -1.87x1018 19.80 0.169 PLD ZnO annealed in vacuum -1.86x1018 12.92 0.260 PLD ZnO annealed in air Films became resistive - - ALD ZnO as-deposited -1.81x1016 1.14 303 ALD ZnO annealed in Ar - 2.1x1016 2.92 101 ALD ZnO annealed in vacuum -2.79x1017 15.75 1.61 ALD ZnO annealed in air Films became resistive - - Sputtered ZnO as-deposited -3.59x1018 35.21 0.050 Sputtered ZnO annealed in Ar -3.69x1018 38.72 0.043 Sputtered ZnO annealed in vacuum -4.20x1018 38.80 0.045 Sputtered ZnO annealed in air Films became resistive - -
4.1.3 Optical Properties
Photoluminescence measurements provide insight to the band structure and defect levels
in the band gap which can be assigned to structural models. Transparency and refractive index
measurements are also discussed in this section.
-PLD ZnO Thin Films
Figure 4-9 represents the photoluminescence (solid) and photoluminescence excitation
(dashed) spectra of the samples synthesized by PLD at the substrate temperature 100 oC (top)
and 200 oC (bottom). The photoluminescence of the PLD films grown at 100 oC was
94
characterized by a relatively narrow, near-UV peak at 385 nm (3.22 eV) and a shoulder at 425
nm (2.92 eV). The photoluminescence excitation maximum was 330 nm (3.76 eV). Both
emissions are assigned to radiative relaxation from a shallow donor band associated with zinc
interstitials to the valence band [11-14]. A band of defects (E1-E4) with activation energies
ranging from 0.12 eV to 0.57 eV below the conduction band edge has been measured in n-type
ZnO by deep level transient spectroscopy [15]. The zinc interstitial reactions based on Kroger-
Vink notation:
and
have formation energies of
2.183 and 3.563 eV respectively, and are therefore more thermodynamically favorable than the
reaction which has formation energy of 4.722 eV [16]. Figure 4-9 also shows
that the near UV peak of the sample deposited at 100 oC is broader compared to the film grown
at 200 oC and further, the intensity ratio of the two peaks I385nm/ I425nm of the film deposited at
100 oC was smaller than the one grown at 200 oC. The transmittance of the PLD ZnO thin films
grown at 100 oC and 200 oC is in the 80-90% range as shown in Figure 4-10.
95
200 250 300 350 400 450 500 550
As grown PLD ZnO ST 200
Wavelength [nm]
200 250 300 350 400 450 500 550
As grown PLD ZnO ST 100
Inte
ns
ity
[a
.u.]
Figure 4-9. The photoluminescence (solid) and photoluminescence excitation (dashed) spectra of the samples synthesized by PLD at 100 oC (top) and 200 oC (bottom)
400 450 500 550 6000
10
20
30
40
50
60
70
80
90
100
T%
Wavelength [nm]
As grown PLD ZnO ST100
As grown PLD ZnO ST200
Figure 4-10. The transmittance of PLD ZnO grown at 100 oC (black) and 200 oC (red)
96
Fig 4-11 shows the photoluminescence (solid) and photoluminescence excitation (dashed)
spectra of the samples synthesized by PLD at 100 oC (black curve) and annealed in Ar (red
curve), vacuum (green curve) and air (blue curve). As seen, both the near-UV peak at 385 nm
(3.22 eV) and a shoulder at 425 nm (2.92eV) of the annealed thin films increased in comparison
to the as-deposited sample. The PLD ZnO annealed in vacuum has the larger PL intensity at 425
nm compared to the one annealed in air. Annealing in Argon and vacuum results in a loss of
oxygen and enhances non-stoichiometry, which increases the density of zinc interstitials and PL
intensity at 3.22 eV. It is also likely that post deposition annealing reduces the concentration of
deep levels (as opposed to shallow donor levels) which serve as electronic traps and non-
radiative pathways. Such mechanisms are consistent with the increases in free carrier
concentration and mobility that are obtained after annealing as shown in Table 4-8. The overall
electrical and photoluminescent characteristics are therefore determined by the relative zinc
interstitial and oxygen vacancy concentrations which vary with annealing treatment.
The chemical analysis by XPS mentioned previously confirms the loss of oxygen. The
atomic concentration of Zn/O for the PLD ZnO thin films grown at 100 ºC before and after
annealing are summarized in Table 4.2. It indicates that ZnO annealed in vacuum lost the most
of oxygen, so the peak (425 nm, 2.92eV) of interstitial zinc increased the most. The
transmissivity of the as-deposited and annealed thin films are shown in Figure 4-12. All of PLD
ZnO thin films exhibit transmissivity in the 80-90 % range which was essentially unaffected by
post deposition annealing.
97
200 250 300 350 400 450 500 550
As grown
Annealing in Ar
Annealing in vacuum
Annealing in air
Inte
ns
ity
[a
.u.]
Wavelength [nm]
PLD ZnO
Fig 4-11. The photoluminescence (solid) and photoluminescence excitation (dashed) spectra of the samples synthesized by PLD at 100 oC (black) and after annealing at 400 oC in Ar (red), vacuum (green) and air (blue)
450 500 550 6000
20
40
60
80
100
PLD as grown T%
annealing in vacuum T%
annealing in Ar T%
annealing in air T%
Wavelength [nm]
T%
PLD ZnO 1.04mJ/cm2, 10mtorr of oxygen
Figure 4-12. The transmittance of the samples synthesized by PLD at 100 oC (black) and after annealing at 400 oC in Ar (red), vacuum (green) and air (blue)
98
- ALD for ZnO
The photoluminescence (solid) and photoluminescence excitation (dashed) spectra of the
samples synthesized by ALD at 100 ºC (top) and 200 oC (bottom) are shown in Figure 4-13.
The PL of the samples deposited at 100 oC is dominated by a broad band peak around 605 nm
(2.05 eV) with a significantly weaker peak at 425 nm (2.92 eV). The photoluminescence
excitation maximum is 367 nm (3.38 eV). It has been proposed that the former luminescence is
due to an overlap of the red, yellow and green defect bands [17], which accounts for its broad,
unstructured character. However, it must be noted that the specific origins of these individual
emissions are still disputed. For example, the green PL has been attributed to oxygen vacancies
[18], a zinc vacancy acceptor [19], and a complex defect involving zinc interstitials, zinc-oxygen
anti-site defects and oxygen vacancies [20]. Thus, the decrease in PL intensity with annealing
could be related to a reduction in luminescent defect concentrations, but the broad width and
unstructured nature of the band precludes assignment to specific defects.
In contrast to the 100 oC film, the ALD sample grown at 200 oC, which is shown in
Figure 4-13 (bottom), exhibits predominantly 380 nm (3.26 eV) photoluminescence under
optimal, 330 nm (3.76 eV), excitation. This near ultra-violet (UV) emission is relatively narrow,
and a shoulder at 425 nm (2.92 eV) is also observed. As with the discussion of the PLD films
above, the both emissions are assigned to radiative relaxation from a shallow donor band
associated with zinc interstitials to the valence band. A series of annealing PL experiments were
also performed for ALD ZnO thin films annealed in different ambient, which in combination
with chemical analysis were expected to provide deeper insights to the active mechanisms. The
transmissivity of the films at deposited 100 ºC and 200 ºC is shown in Figure 4-14. The ALD
ZnO thin films exhibit transmissivity in the 80-90 % range which is essentially unaffected by
deposition substrate temperature.
99
200 250 300 350 400 450 500 550 600 650 700
As grown ALD ZnO ST200
Inte
ns
ity
[a
.u.]
Wavelength [nm]
200 250 300 350 400 450 500 550 600 650 700
As grown ALD ZnO ST100
Figure 4-13. The photoluminescence (solid) and photoluminescence excitation (dashed) spectra of the samples synthesized by ALD at 100 oC (top) and 200 oC (bottom)
400 450 500 550 6000
10
20
30
40
50
60
70
80
90
100
T%
Wavelength [nm]
As grown ALD ZnO ST 100
As grown ALD ZnO ST 200
Figure 4-14. The transmittance of ALD ZnO grown at 100 oC (black) and 200 oC (red)
Figure 4-15 shows the photoluminescence (solid) and photoluminescence excitation
(dashed) spectra of the samples deposited by ALD at 100 oC before and after annealing at 400oC
100
in Ar (red), vacuum (green) and air (blue). All of the samples are measured with identical
spectrometer settings. The PL of the 605 nm band diminishes with post deposition annealing
while the intensity of the 380 nm peak increases with annealing in air as shown in the bottom of
Figure 4-15. The photoluminescence excitation maximum is 367 nm (3.38 eV) for the 605 nm
peaks, and 330 nm for the 380 nm peaks. These results indicate that the origins of these
emissions are different. Weak shoulders at 425 nm are also observed for these annealed ZnO
films. Considered together, these observations suggest that a reduction in the concentration of
oxygen vacancies is a possible explanation for the observed electrical and photoluminescent
characteristics in this case.
However, the situation is more complex because annealing in Ar and vacuum also
reduced the intensity of the 605 nm PL band, but increased the free carrier concentration and
mobility as shown in Table 4-8. These anneals also increased the intensity of the 380nm PL
peak as well as the 425 nm shoulder. It must be noted that the PL intensity of these two peaks
was significantly less intense than that of the broad 605 nm band. Similar to PL from PLD ZnO,
both emissions are assigned to radiative relaxation from a shallow donor band associated with
zinc interstitials to the valence band. The ZnO thin films which were annealed in Argon and
vacuum resulted in loss of oxygen and enhanced non-stoichiometry, which increased the density
of zinc interstitials and PL intensity at 3.26 eV. It is also likely that post deposition annealing
reduces the concentration of deep levels (as opposed to shallow donor levels) which serve as
electronic traps and non-radiative pathways. Such mechanisms are consistent with the increases
in free carrier concentration and mobility that are obtained after annealing. According to
Richard et al. [16], oxygen vacancies could contribute to a red-orange photoluminescence (at 2.1
eV and below) which is close to our experimental peak at 2.05 eV. However, after annealing in
101
air at 400 oC, the intensity of PL peak (2.05 eV) decreased. This might be due to the
incorporation of oxygen into the film which was induced by annealing in air. Thus, the oxygen
vacancies and the intensity of the associated PL peak decreased. A comparison of ZnO films
annealed in vacuum with those treated in air indicates that the intensity of the 2.05 eV decreased
more for the one annealed in air. By considering the fact that molecular density of oxygen in air
is 8-9 orders of magnitude higher than in our vacuum annealing conditions (10-6 Torr), it is not
unreasonable to expect that the diffusion of oxygen from air into the ZnO films is more probable
than from a vacuum atmosphere. This hypothesis was supported by the XPS analysis, the results
of which are summarized in Table 4-4. The obtained Zn/O ratio in the ALD ZnO thin films
grown at 100 oC and after vacuum annealing were 1.17 and 1.19 respectively, while for the
sample annealed in air it was 1.15. By assuming that the change in the amount of Zn is
negligible, one can conclude that annealing in air increases the oxygen concentration. Therefore,
it can be concluded that oxygen vacancies contributed to the broad peak (605 nm, 2.05eV) of
ALD ZnO deposited at 100℃, and after annealing in Ar, vacuum and air. The transmissivity of
the as deposited and as-annealed films is shown in Figure 4-16. All of as-grown ALD and
annealed ZnO films exhibit transmissivity in the 80-90 % range.
102
350 400 450 500 550 600 650 700
As grown
Annealing in vacuum
Annealing in air
Annealing in Ar
Inte
ns
ity
[a
.u.]
Wavelength [nm]
200 250 300 350 400 450
ALD ZnO
Figure 4-15. The photoluminescence (solid) and photoluminescence excitation (dashed) spectra of the samples synthesized by ALD at 100 oC (black), and after annealing at 400 oC in Ar (red), vacuum (green) and air (blue)
400 450 500 550 6000
10
20
30
40
50
60
70
80
90
100
ALD as grown T%
annealing in vacuum T%
annealing in air T %
annealing in Ar T %
T%
Wavelength [nm]
Figure 4-16. The transmittance of the samples synthesized by ALD at 100 oC (black) and after annealing at 400 oC in Ar (red), vacuum (green) and air (blue)
103
-RF Magnetron Sputtering for ZnO Thin Films
Figure 4-17 represents the photoluminescence (solid) and photoluminescence excitation
(dashed) spectra of the ZnO samples synthesized by sputtering at the substrate temperature 100
oC (top) and 200 oC (bottom). The photoluminescence of both is characterized by a relatively
narrow, near-UV peak at 385 nm (3.22 eV) and a shoulder at 425 nm (2.92 eV). The
photoluminescence excitation maximum was 330 nm (3.76 eV). The both emissions are
assigned to radiative relaxation from a shallow donor band associated with zinc interstitials to the
valence band, which are similar with the emission from PLD ZnO thin films. Figure 4-17 also
shows that the intensity ratio of the two peaks I385nm/ I425nm of the film deposited at 100 oC was
slightly smaller than one at 200 oC. The transmittance of the sputtered ZnO thin films grown at
100 oC and 200 oC are in the range of 75-85%, as shown in Figure 4-18.
200 250 300 350 400 450 500 550
As grown Sputteirng ZnO ST 200
Wavelength [nm]
200 250 300 350 400 450 500 550
As grown Sputterig ZnO ST 100
Inte
ns
ity
[a
.u.]
Figure 4-17. The photoluminescence (solid) and photoluminescence excitation (dashed) spectra of the samples synthesized by sputtering at 100 oC (top) and 200 oC (bottom)
104
400 450 500 550 6000
10
20
30
40
50
60
70
80
90
100
T%
Wavelength [nm]
ST100
ST200
Sputtering ZnO
Figure 4-18. The transmittance of the sputtered ZnO grown at 100 oC (black) and 200 oC (red)
Fig 4-19 shows the photoluminescence (solid) and photoluminescence excitation (dashed)
spectra of the samples synthesized by sputtering at 100 oC (black color) and after annealing in Ar
(red), vacuum (green) and air (blue). As seen, a UV peak at 385 nm (3.22 eV) and a shoulder at
425 nm (2.92eV) are observed for all of the sputtered ZnO thin films. The film annealed in air
has the largest ratio of I385/I425 peaks compared to those annealed in Ar and vacuum. Similar to
the annealed PLD films, the sputtered samples that were annealed in Ar and vacuum resulted in
loss of oxygen and enhanced non-stoichiometry, which increased the density of zinc interstitials
and PL intensity at 425 nm (3.22 eV). The photoluminescent and electrical characteristics are
therefore determined by the relative zinc interstitial and oxygen vacancy concentrations which
vary with annealing treatments. The composition of the sputtered ZnO from XPS chemical
analysis confirmed the loss of oxygen as previously shown in Table 4-6. This data indicates that
the films annealed in Ar and vacuum lost oxygen, which increased the density of zinc interstitials,
so that the intensity of I385/I425 decrease after annealing in Ar and vacuum. The transmissivity of
105
the sputtered ZnO films is shown in Figure 4-20. All of as-grown and annealed ZnO films
exhibit transmissivity in the 75-85 % range which was essentially unaffected by post deposition
annealing.
200 250 300 350 400 450 500 550
Inte
ns
ity
[a
.u.]
Wavelength [nm]
As deposited
Annealed in vacuum
Annealed in air
Annealed in Ar
Sputtering ZnO (4.44 W/cm2)
Figure 4-19. The photoluminescence (solid) and photoluminescence excitation (dashed) spectra of the samples synthesized by sputtering at 100 oC (black) and after annealing at 400 oC in Ar (red), vacuum (green) and air (blue)
106
400 450 500 550 6000
10
20
30
40
50
60
70
80
90
100
As grown
Annealing in Ar
Annealing in vacuum
Annealing in air
T%
Wavelength [nm]
Sputtering ZnO
Figure 4-20. The transmittance of the samples synthesized by sputtering at 100 oC (black) and
after annealing at 400 oC in Ar (red), vacuum (green) and air (blue)
The refractive indices of ZnO thin films were measured with a J.A. Woollam M-2000
Ellipsometer. The Cauchy Model [21] was used to fit the data in order to determine the optical
constants of the samples. The following formulas are used to define the refractive index (n) and
from Cauchy Model:
( )
( ) [ (
)]
where A, B, C, are constants and the extinction coefficient amplitude , the exponent factor ,
and the band edge are fitting parameters for the best fit. is the wavelength of the light.
The incident angles of the light used were 55o, 60o and 65o. The wavelength ranges are
from 400 nm to 700 nm with a step of 5 nm. The system recorded the spectra of psi (ψ) and
delta (Δ) as function of wavelength (λ) and then converted to the data to the complex of
refractive index, . For ZnO thin films deposited at 100oC, the best curve fits with
107
Cauchy model are shown in Figure 4-21; (a) PLD ZnO (b) ALD ZnO (c) Sputtered ZnO. The
red curves show the experimental data and the black curves are the Cauchy model fit. The
refractive indices of the ZnO films deposited at 100oC are shown in Figure 4-22. The refractive
index of the PLD ALD and sputtered ZnO is 1.9 at 500 nm, which is consistent with the reported
literature [22,23]. However, the extinction coefficient could not be obtained because the
material is transparent at wavelengths larger than 400 nm [23].
400 500 600 700
1.4
1.6
1.8
2.0
2.2
2.4
2.6
refr
ac
tiv
e in
de
x <
n>
Wavelength [nm]
Model fit 55
Model fit 60
Model fit 65
Exp 55
Exp 60
Exp 65
(a) PLD ZnO 1.04mJ/cm2 at 100
oC
Figure 4-21. (a) The experimental data (red) and Cauchy model fit curves (black) for PLD ZnO grown at 1.04 mJ/cm2 and 100oC
108
400 500 600 700
1.4
1.6
1.8
2.0
2.2
2.4
2.6
2.8
Re
fra
cti
ve
in
de
x <
n>
Wavelength [nm]
Model fit 55
Model fit 60
Model fit 65
Exp 55
Exp 60
Exp 65
(b) ALD ZnO at 100o
C
Figure 4-21. (b) The experimental data (red) and Cauchy model fit (black) for ALD ZnO grown at 100oC
400 500 600 700
1.2
1.5
1.8
2.1
2.4
2.7
3.0
3.3 Model Fit 50
Model Fit 55
Model Fit 60
Exp E 50
Exp E 55
Exp E 60
<n
> w
ith
Ca
uc
hy
mo
de
l
Wavelength [nm]
(c) Sputtered ZnO 4.44W/cm2 at 100
oC
Figure 4-21. (c) The experimental data (red) and Cauchy model fit (black) for sputtered ZnO fabricated at 4.44 W/cm2 and 100oC
109
400 500 600 700
1.6
1.8
2.0
2.2
2.4
Re
fra
cti
ve
in
de
x
Wavelength[nm]
PLD ZnO
ALD ZnO
Sputtered ZnO
Figure 4-22. The refractive index as a function of wavelength for PLD, ALD, and sputtered ZnO
4.2 Properties of BaTa2O6 and Al2O3
In this section, the results of the investigation of the structural, electrical and optical
properties of BaTa2O6 and Al2O3 deposited by RF magnetron sputtering on ITO-coated glass
substrates are reported.
4.2.1 Structure of BaTa2O6 and Al2O3
Normally, barium tantalate ( BaTa2O6) has three types of structures with respect to their
polymorphs [24]. An orthorhombic structure is obtained when the processing temperature is
below 1150 °C. A tetragonal tungsten bronze (TTB)-type structure exists in the processing range
of 1150–1300 °C. A hexagonal structure (H-BaTa2O6) exists when the process temperature is
above 1300 °C. However, in this study BaTa2O6 films are grown at room temperature with
sputtering power densities of 4.93 W/cm2, 5.92 W/cm2, 6.91 W/cm2, 7.89 W/cm2and 8.88 W/cm2.
110
The diffraction patterns are featureless and indicate that the BaTa2O6 films as amorphous. For
Al2O3 deposition, the most common form of crystalline alumina is corundum. The oxygen ions
form a hexagonal close-packed structure with aluminum ions filling two-thirds of the octahedral
interstices. However, the Al2O3 films in this work were grown at room temperature, and all of
the films were amorphous as evidenced by their essentially featureless diffraction spectra. The
sputtering power densities used were 5.92W/cm2, 6.91W/cm2, 7.89 W/cm2, 8.88 W/cm2 and
9.87W/cm2. Figure 4-23 shows the featureless XRD spectra of BaTa2O6 and Al2O3 films, which
were grown at 6.91 W/cm2 and 8.88 W/cm2 respectively.
20 30 40 50 60 70 80 90
BaTa2O
6 with 6.91 W/cm
2 BaTa
2O
6
2 thelta [degree]
20 30 40 50 60 70 80 90
Al2O
3 with 8.88 W/cm
2 Al2O
3
Inte
ns
ity
[a
.u.]
Figure 4-23. X-ray diffraction patterns of BaTa2O6 (bottom) and Al2O3 (top) thin films deposited with RF magnetron sputtering power densities of 6.91 W/cm2 and 9.87 W/cm2, respectively
111
4.2. 2 Electrical Properties of BaTa2O6 and Al2O3
Figure 4-24 displays the current density (J) vs. electric field strength (E) curves for
BaTa2O6 films which were prepared at sputtering power densities of 4.93 W/cm2, 5.92 W/cm2,
6.91 W/cm2, 7.89 W/cm2 and 8.88 W/cm2. Dielectric breakdown is observed in all of the films
at some field level. For example, when the sputtering power was 6.91 W/cm2, the current
increased exponentially after 1MV/cm. When the sputtering power density increased from 4.93
W/cm2 to 7.89 W/cm2, the breakdown strength increased from 0.2 MV/cm to 1.1 MV/cm.
However, the breakdown strength decreases to 0.2 MV/cm when the sputtering power density
was increased to 8.88 W/cm2. This may be due to gas bubbles being trapped in the films which
reduce density and breakdown strength. This study shows that the deposition window for
BaTa2O6 at room temperature is in the range of 5.92 W/cm2 to 7.89 W/cm2.
For semiconductor device applications, the leakage current of insulators must be as small
as possible. As seen in Figure 4-24, except for the BaTa2O6 films prepared at sputtering power
densities of 4.93 W/cm2 and 8.88W/cm2, the leakage current of BaTa2O6 films are in the order of
10-6-10-7 A/cm2 at the applied field of 1MV/cm. Specifically, the leakage currents were 5.73x10-6
A/cm2, 1.19x10-6 A/cm2 and 9.16x10-7 A/cm2 when the sputtering power densities were 5.92
W/cm2, 6.91W/cm2 and 7.89 W/cm2, respectively.
112
-0.3 0.0 0.3 0.6 0.9 1.2 1.5
0.00
5.00x10-6
1.00x10-5
1.50x10-5
2.00x10-5
Cu
rre
nt
de
ns
ity
[A
/cm
2]
Electrical field [MV/cm]
4.93 W/cm2
5.92 W/cm2
6.91 W/cm2
7.89 W/cm2
8.88 W/cm2
BaTa2O
6
Figure 4-24. Current density vs. electric field strength of BaTa2O6 films sputtered with power densities of 4.93 W/cm2, 5.92 W/cm2, 6.91 W/cm2, 7.89 W/cm2 and 8.88 W/cm2
Figure 4-25 depicts the current density vs. electric field of Al2O3 films synthesized at
sputtering power densities of 5.92 W/cm2, 6.91 W/cm2, 7.89 W/cm2, 8.88 W/cm2 and 9.87
W/cm2. When the sputtering power density was 8.88W/cm2, the breakdown electrical field was
0.55 MV/cm. The same behavior was observed for Al2O3 films sputter grown at 9.87 W/cm2.
The highest breakdown strength of the Al2O3 films was in the range of 0.55-0.62 MV/cm in this
study, and the corresponding leakage currents were in the order of 10-7A/cm2 before breakdown.
From the electrical properties of the BaTa2O6 and Al2O3 that were measured in this work,
BaTa2O6 films grown at 6.91W/cm2 and 7.89W/cm2, and Al2O3 films grown at 8.88W/cm2 and
9.87W/cm2 will be used for the insulators in the MIS structures discussed below.
113
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
0.0
2.0x10-8
4.0x10-8
6.0x10-8
8.0x10-8
1.0x10-7
Cu
rre
nt
de
ns
ity
[A
/cm
2]
Electrical field [MV/cm]
5.92 W/cm2
6.91 W/cm2
7.89 W/cm2
8.88 W/cm2
9.87 W/cm2
Al2O
3
Figure 4-25. Current density vs. electric field strength of Al2O3 films sputtered with powers densities of 5.92 W/cm2, 6.91 W/cm2, 7.89 W/cm2, 8.88 W/cm2 and 9.87 W/cm2
4.2.3 Optical Properties of BaTa2O6 and Al2O3
Figure 4-26 and Figure 4-27 show the transmittance of the BaTa2O6 and Al2O3 films
grown with the sputtering powers described at the end of the previous section. The average
thickness of both materials is 100 nm -120nm. As shown in Figure 4-26 and Figure 4-27, the
transmittance of both are in the 80-90% range, which are not affected by the sputtering powers.
114
450 500 550 600 650 700
0
20
40
60
80
100
Tra
ns
pa
ren
cy
[%
]
Wavelength[nm]
6.91 W/cm2
7.89 W/cm2
BaTa2O
6
Figure 4-26. The transmittance of BaTa2O6 grown with sputtering powers of 6.91 W/cm2 (blue) and 7.89 W/cm2 (green)
450 500 550 600 650 700
0
20
40
60
80
100
Tra
ns
pa
ren
cy
[%
]
Wavelength[nm]
8.88 W/cm2
9.87 W/cm2
Al2O
3
Figure 4-27. The transmittance of Al2O3 grown with sputtering powers of 8.88 W/cm2 (purple pink) and 9.87 W/cm2 (black)
115
For the refractive index of insulator films, the best fit is obtained using a Lorentz Model
[25]. The complex index of refraction of the BaTa2O6 films is shown in Figure 4-28 (a) and
Figure 4-28 (b). The red curves show the experimental data and black curves show the Lorentz
Model fit. The refractive index and extinction coefficient as a function of wavelength for the
BaTa2O6 films are shown in Figure 4-29. The refractive indexes of the BaTa2O6 grown with
6.91W/cm2 and 7.89W/cm2 are around 2.0~2.1 at 500 nm while the BaTa2O6 film grown with a
higher power density of 7.89W/cm2 is slightly larger. The extinction coefficients of the BaTa2O6
films are only~ 0.00003, which means absorption is negligible in the visible region. Similar
results were obtained in the Al2O3 films. The complex of refractive index of Al2O3 grown at 8.88
W/cm2 and 9.87 W/cm2 is shown in Figure 4-30 (a) and Figure 30 (b). The corresponding
refractive index and extinction coefficients are shown in Figure 4-31. The refractive index of
Al2O3 film grown at 9.87 W/cm2 is around 1.9 at 500 nm, and 1.7 at 500 nm for the film grown
at with 8.88 W/cm2. The extinction coefficient of the Al2O3 films is ~0.005, which is close to the
literature reports [26].
116
400 500 600 700
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
5.5
6.0 Model Fit 50
Model Fit 55
Model Fit 60
Exp E 50
Exp E 55
Exp E 60<
n>
wit
h L
ore
ntz
mo
de
l
Wavelength [nm]
BaTa2O
6 6.91 W/cm
2
Figure 4-28. (a) The experimental data (red) and Lorentz model fit (black) for BaTa2O6 grown at 6.91 W/cm2
400 500 600 700
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0 Model Fit 50
Model Fit 55
Model Fit 60
Exp E 50
Exp E 55
Exp E 60
<n
> w
ith
Lo
ren
tz m
od
el
Wavelength [nm]
(b) BaTa2O6 7.89 W/cm2
Figure 4-28. (b) The experimental data (red) and Lorentz model fit (black) for BaTa2O6 grown at 7.89 W/cm2
117
400 500 600 700
1.6
1.8
2.0
2.2
2.4
6.91 W/cm2
7.89 W/cm2
6.91 W/cm2
7.89 W/cm2
A
Re
fra
cti
ve
in
de
x (
n)
BaTa2O6
0.0000
0.0001
0.0002
0.0003
Ex
tin
cti
on
co
eff
icie
nt
(k)
Figure 4-29. The refractive index and extinction coefficient as a function of wavelength for BaTa2O6 films grown 6.91 W/cm2 and 7.89W/cm2
400 500 600 700
0.5
1.0
1.5
2.0
2.5
3.0 Model fit 50
Model fit 55
Model fit 60
Exp 50
Exp 55
Exp 60
<n
> w
ith
Ca
uc
hy
mo
de
l
Wavelength [nm]
Al2O3 8.88 W/cm2
Figure 4-30. (a) The experimental data (red) and Lorentz model fit (black) for Al2O3 grown at 8.88 W/cm2
118
400 500 600 700
0.5
1.0
1.5
2.0
2.5
3.0 (b) Al2O3 9.87 W/cm
2
<n
> w
ith
Ca
uc
hy
mo
de
l
Wavelength [nm]
Model fit 50
Model fit 55
Model fit 60
Exp 50
Exp 55
Exp 60
Figure 4-30. (b) The experimental data (red) and Lorentz model fit (black) for Al2O3 grown at 9.87 W/cm2
400 450 500 550 600 650 700
0.5
1.0
1.5
2.0
2.5
8.88 W/cm2
9.87 W/cm2
8.88 W/cm2
9.87 W/cm2
Wavelength [nm]
Re
fra
cti
ve
in
de
x (
n)
0.000
0.005
0.010
0.015
0.020
Ex
tinc
tion
co
effic
ien
t (k)
Figure 4-31. The refractive index and extinction coefficient as a function of wavelength for Al2O3 films grown at 8.88 W/cm2 and 9.87W/cm2
119
4.3 Conclusions: Structure and Electro-optical Properties of ZnO, BaTa2O6 and Al2O3 Thin
Films Deposition
A comparison of the structural properties using XRD confirmed that both PLD and
sputtered ZnO thin films grown at 100ºC and 200°C were polycrystalline (wurtzite) crystal
structure. PLD and sputtered ZnO films exhibit strong (002) texturing, which is commonly
observed in ZnO films because the c-plane perpendicular to the substrate normal is the most
densely packed and thermodynamically preferred in the wurtzite structure. The ALD ZnO films
deposited films deposited at 100 ºC and 200 ºC show predominantly prismatic (100) texturing.
With regards to the electrical properties of ZnO films, the Hall Effect measurements show that
all of the ZnO films exhibited n-type conductivity. The ZnO sputtered at 100 oC and 200 oC
have the highest free electron concentration of 3.6 x 1018 cm-3 and 4.41x1018cm-3, the highest
mobility of 35.21 cm2/V.s and 38.10 cm2/V.s, the smallest resistivity of 0.05 Ω and 0.037,
respectively. The corresponding values for the PLD films were 1.1x1017 cm-3, 14.6 cm2/V.s, and
3.85 Ω, respectively for ZnO films grown by at 100 oC, and 1.81x1016 cm-3, 1.14 cm2/V.s, and
303 Ω for films grown at 100 oC by ALD. The conduction mechanisms proposed for the ZnO
films were supported by the photoluminescence and XPS analysis. Interstitial Zn contributes to
the conductivity in most of ZnO films. The transparency ranged between 75 and 90 % for all of
ZnO films. The refractive index of the ZnO films irrespective of deposition method was around
1.9 at 500 nm. The results show that sputtered ZnO films exhibit the best combination of
properties of the methods used to deposit films at 100 oC. Sputtered films were therefore used
for the MIS device studies.
The insulators were deposited by radio-frequency magnetron sputtering at room
temperature. BaTa2O6 films grown at sputtering power densities of 6.91W/cm2 and 7.89W/cm2
120
had breakdown strengths of around 1.1 MV/cm and 1.6 MV/cm. The leakage currents were in
the order of 10-6-10-7 A/cm2 at an applied field of 1MV/cm. Among the Al2O3 films studied,
those grown at sputtering power densities of 8.88 W/cm2 and 9.87 W/cm2 exhibited the highest
breakdown strengths of 0.58MV/cm and 0.62 MV/cm. The leakage currents were in the order of
10-7A/cm2 before breakdown. The transparency of the BaTa2O6 and Al2O3 films were all in the
of 80-90 % range. The refractive index of the BaTa2O6 films is~ 2.0-2.1 while the refractive
index of the Al2O3 films is approximately 1.7-1.9 at 500 nm. From the electrical properties and
optical properties of BaTa2O6 and Al2O3 that were observed in this work, BaTa2O6 films with
6.91W/cm2 and 7.89W/cm2, and Al2O3 films with 8.88 W/cm2 and 9.87 W/cm2 were employed
for the growth of the insulators in the MIS structures. They are discussed in Chapter 5.
4.4 References
[1] Ü.Özgür, Y. I. Alivov, C. Liu, A. Teke, M. A. Reshchikov, S. Doğan, V. Avrutin, S. J. Cho,
and H. Morko, J. Appl. Phys. 98, 041301(1-103) (2005)
[2] X. M. Fan, J. S Lian, Z. X. Guo, Appl. Syrf. Sci. 239 (2005)
[3] A. L. Patterson, “The Scherrer Formula for X-ray Particle Size Determination”, Phys. Rev.
56, 978-982 (1939)
[4] E. Guziewicz, I. A. Kowalik, M. Godlewski, K. Kopalko, V. Osinniy, A. Wójcik, S.
Yatsunenko, E. Łusakowska, W. Paszkowicz, and M. Guziewicz, “Extremely Low
Temperature Growth of ZnO by Atomic Layer Deposition”, J. Appl. Phys. 103, 033515(1-6)
(2008).
121
[5] S.J. Lim, Soonju Kwon, and H. Kim, “ZnO Thin Films Prepared by Atomic Layer Deposition
and RF Sputtering as an Active Layer for Thin Film Transistor”, Thin Solid Films 516, 1523-
1528 (2008).
[6 ]S. Kwon, S. Bang, S. Lee, S. Jeon, W. Jeong, Hyungchul Kim1, S. C. Gong, H. J. Chang, H.
Park, and H. Jeon, “Characteristics of the ZnO Thin Film Transistor by Atomic Layer
Deposition at Various Temperatures”, Semicond. Sci. Technol. 24, 035015(6pp) (2009).
[7] J. W. Elam, G. Xiong, C.Y. Han, H. H. Wang, J. P. Birrell, J. N. Hryn, M. J. Pellin J. F. Poco
and J. H. Stcher, “Atomic Layer Deposition for the Conformal Coating of Nanoporous
Materials”, Mater. Res. Soc. Symp. Proc. Vol. 876E, R12.6.1-5 (2005).
[8] M. Oh, S. Kim, S. Park, T. Seong, Superlattices Microstruct. 39, 130–137 (2006)
[9]Soon-Jin So, Choon-Bae Park Journal of Crystal Growth 285, 606–612 (2005)
[10] J. Yang, H. Kim, J. Lim, D. Hwang, J. Oh, and S. Park J. Electrochem. Soc., 153 _3_ G242-
G244 (2006)
[11] E. Guziewicz, I. A. Kowalik, M. Godlewski, K. Kopalko, V. Osinniy, A. Wójcik, S.
Yatsunenko, E. Łusakowska, W. Paszkowicz, and M. Guziewicz, J. Appl. Phys. 103,
033515(1-6) (2008)
[12] A.E. Rakhshani, A. Bumajdad, and J. Kokaj, Appl. Phys. A 89, 923-928 (2007)
[13] Y. Sun, J. B. Ketterson, and G. K. L. Wong, Appl. Phys. Lett. 77, 2322-2324 (2000)
[14] D. H. Zhang, Z. Y. Xue, and Q. P. Wang, J. Phys. D 35, 2837-2840 (2002)
[15] F. D. Auret, S. A. Goodman, M. J. Legodi, W. E. Meyer, and D. C. Look, Appl. Phys. Lett.
80, 1340-1342 (2002)
[16] C. Richard, A. Catlow, S. A. French, A. A. Sokol, A. A. Al-Sunaidi, and S. M. Woodley, J.
Comput. Chem. 13, 2234-2249 (2008)
122
[17] E. Guziewicz, I. A. Kowalik, M. Godlewski, K. Kopalko, V. Osinniy, A. Wójcik, S.
Yatsunenko, E. Łusakowska, W. Paszkowicz, and M. Guziewicz, J. Appl. Phys. 103,
033515(1-6) (2008)
[18] F. Leitera, H. Alvesa, D. Pfisterera, N.G. Romanovb, D.M. Hofmanna, and B.K. Meyer,
Phys. B 340, 201-204 (2003)
[19] B. Guo, Z. R. Qiu, and K. S. Wong, Appl. Phys. Lett. 82, 2290-2292 (2003).
[20] K. Vanheusden, C.H. Seagera, W.L. Warren, D.R. Tallant, J. Caruso, M.J. Hampden-
Smithb, T.T. Kodasb, J. Lumines. 75, 11-16 (1997)
[21] H.G. Tompkins, W.A. McGahan, Spectroscopic Ellipsometry and Reflectometry, Wiley,
New York, pp. 93–94 (1999)
[22] Y.C. Liua,, S.K. Tunga, J.H. Hsieh, Journal of Crystal Growth 287, 105–111 (2006)
[23] Q. Hua Li, D. Zhu, W. Liu,Yi Liu and X. Cui Ma, Appl. Surf. Sci. Vol.254, 10, 2922-2926
(2008)
[24] W. Zhang, N. Kumada, T. Takei, J. Yamanaka and N. Kinomura, Mater. Research Bulletin
40, 117-1186 (2005)
[25] H. Tompkins and E. A Haber, “Handbook of Ellipsometry (Materials Science and Process
Technology)”, William Andrew (2005)
[26] L. Pereira, P. Barquinha, G. Gonc¸A. Vil `a, A. Olziersky, J.Morante,, E. Fortunato, and R.
Martins, Phys. Status Solidi A 206, No. 9, 2149–2154 (2009)
123
CHAPTER 5
RESULTS: QUANTIFICATION OF TRAP STATES IN THE HIGH-K BULK AND AT
INTERFACE WITH ZINC OXIDE
5.1 Fabrication of MIS Structures Employing High-k BaTa2O6 and Al2O3
ZnO, BaTa2O6 and Al2O3 films with the processing techniques discussed in Chapter 4
were incorporated into MIS structures by sequential sputter deposition onto ITO/glass substrates.
ZnO was first deposited on ITO glasses by RF sputtering with a sputtering power density of 4.44
W/cm2, 20 mtorr of Ar, and a substrate temperature of 100 oC. BaTa2O6 was deposited at
sputtering powers of 6.91 W/cm2 and 7.89 W/cm2, 20 mtorr of Ar and the substrate at room
temperature. Al2O3 was deposited at room temperature using sputtering powers of 7.89 W/cm2
and 8.88 W/cm2 and 20 mtorr of Ar. A thermally evaporated 200 nm thick Ag layer on top of
the insulator completed the MIS structure. The two complete MIS structures were shown in
Figure 3-11.
5.2 Quantitative Determination of Trap States in MIS structures Employing High-k BaTa2O6
and Al2O3
The ZnO film in the MIS structure would be the channel material in complete thin film
transistors. The drain current in the transistor operation depends on the mobility of carriers in
the channel, the properties of gate oxide, VDS, VGS and VT. Sputtered ZnO was selected as the
processing method in this work because of its high mobility (35.2 cm2/V.s) compared to the
ALD and PLD films. Simultaneously, the contacts between the ZnO and ITO substrates were
124
Ohmic, therefore understanding the bulk and interface properties of the high k dielectric becomes
the focal issue.
5.2.1 C-V Measurements
Frequency dependent C-V measurements are performed using an Agilent 4294A
impedance analyzer integrated with a Micromanipulator probe station. Figure 5-1 is the C-V
characteristics of the Ag/ BaTa2O6/ ZnO/ITO/glass MIS structure measured at 1MHz. Although
the transition from accumulation to inversion is distinct, it is not sharp. Further, as can be
determined from Figure 5-1, the flatband voltages for the BaTa2O6 MIS structures are
characterized by a noticeable positive bias shift, which indicates that the as-deposited films
contain a high concentration of negative fixed oxide charge at the BaTa2O6/ZnO interface. In the
accumulation region the measured capacitance is the oxide capacitance (Cox) and for the
BaTa2O6 films sputtered at 6.91 W/cm2 and 7.89 W/cm2, oxide capacitances of 7.71 nF and
12.08 nF were measured respectively. For a parallel plate capacitor containing a dielectric that
completely fills the space between the plates, the dielectric constants can be calculated by
where C is the capacitance, d is the thickness of insulator layer, εo is the permittivity of vacuum
and A is the area of the electrodes. The dielectric constant (k) of the films sputtered at 6.91
W/cm2 and 7.89 W/cm2 are 33.8 and 37.5 respectively. It is proposed that the enhanced
energetic atomic peening facilitated by the larger deposition power leads to increased film
density, and as a consequence increased dielectric constant, capacitance and also breakdown
strength.(see Figure 4-22)
125
Figure 5-2 shows the results of the C-V measurement at 1 MHz for the Ag/Al2O3/ZnO/
ITO/glass MIS structure. Compared to the BaTa2O6 films, the capacitance of the Al2O3 is
significantly smaller at 3.54 nF and 3.75 nF. The corresponding dielectric constants (k) of the
films sputtered at 8.88 W/cm2 and 9.87 W/cm2 are 9.8 and 10.6, respectively. Similar to the
BaTa2O6 films, the higher dielectric constant, capacitance and breakdown strength (Figure 4-23),
obtained at the higher deposition power, show that the higher film density facilitated more
energetic atomic peening. The voltage shift values for the two oxides (see Figures 5-1 and 5-2)
show that the as-deposited Al2O3 films are characterized by a much smaller fixed oxide charge
concentration. The shift of the flat band voltages of the MIS structures with BaTa2O6 is almost
18 times larger than the structure with Al2O3. According to Eq 2-3, a lager shift of flat band
voltages means a larger required switching power. Thus, if transistors were made with BaTa2O6,
they would consume much more power than their equivalents made with Al2O3.
0 2 4 6 8
2
4
6
8
10
12 MIS with 6.91W/cm
2BaTa
2O
6
MIS with 7.89W/cm2BaTa
2O
6
Voltage [V]
Ca
pa
cit
an
ce
[n
F]
Figure 5-1. C-V at 1MHz for Ag/BaTa2O6/ZnO MIS diodes. The insulator was sputtered at 6.91W/cm2 (red) and 7.89W/cm2 (black)
126
-0.5 0.0 0.5 1.0
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
Ca
pa
cit
an
ce
[n
F]
Voltage [V]
MIS with 8.88W/cm2 Al
2O
3
MIS with 9.87W/cm2 Al
2O
3
Figure 5-2. C-V at 1MHz for Ag/ Al2O3/ZnO MIS diodes. The Al2O3 was sputtered at 8.88W/cm2 (red) and 9.87W/cm2 (black)
5.2.2 G-V Measurement
Figure 5-3 shows that conductance-voltage (G-V) characteristics at 1 MHz of the
Ag/BaTa2O6/ZnO MIS diodes. The trap density (Dit) at the interface between the insulator and
semiconductor, were determined by conductance-voltage measurements using the methodology
developed by Nicollian-Goetzberger [1] and presented in detail by Schroder [2]. The equivalent
parallel conductance was determined according to:
25])([ 222
max
max2
EqCCG
GCG
mox
oxp
which neglects leakage currents and the series resistance of the MIS structure. The interface trap
density was then calculated using the single frequency approximation [3,4] :
127
)35(
})]([{2
2max
22max
max2
EqGCCGqA
GCD
mox
ox
it
In the above equations, Gp is the parallel conductance, ω is the angular frequency, Gmax is the
peak value of the measured conductance, Cm is the capacitance that corresponds to the peak
conductance value, Cox is the capacitance of insulator (the capacitance measured in
accumulation), q is the electron charge (q =1.602 x 10-19 C), and A is the area of the junction
which is effectively determined by the area of the metal contact. Thus, the interface trap
densities for the BaTa2O6 films sputtered at 6.91 W/cm2 and 7.89 W/cm2 were determined to be
3.67x1011 eV-1cm-2 and 6.18x1011 eV-1cm-2 respectively.
Similarly, Figure 5-4 shows the G-V characteristics measured at 1 MHz of the
Ag/Al2O3/ZnO MIS diodes. The interface trap densities were determined to be 2.96 x1010 eV-1
cm-2 and 1.09 x1010 eV-1 cm-2 for the deposition powers of 8.88 W/cm2 and 9.87 W/cm2,
respectively.
128
0 2 4 6 8
0.002
0.004
0.006
0.008
0.010
0.012
0.014
0.016
Co
nd
uc
tan
ce
[S
]
Voltage [V]
MIS with 6.91W/cm2BaTa
2O
6
MIS with 7.89W/cm2BaTa
2O
6
(b)
Figure 5-3. G-V at 1MHz for Ag/BaTa2O6/ZnO MIS diodes. The insulator was sputtered at 6.91W/cm2 (red) and 7.89W/cm2 (black)
-0.5 0.0 0.5 1.0
0.0000
0.0001
0.0002
0.0003
0.0004
0.0005
0.0006
0.0007 MIS with 8.88W/cm
2 Al
2O
3
MIS with 9.87W/cm2 Al
2O
3
Co
nd
uc
tan
ce
[S
]
Voltage [V]
(b)
Figure 5-4. G-V characteristics at 1 MHz for Ag/Al2O3/ZnO MIS diodes. The Al2O3 was sputtered at 8.88W/cm2 (red) and 9.87W/cm2 (black)
129
Another method to calculate the interface trap density is by measuring G/ω as a function
of frequency. Figure 5-5 (a) shows the capacitance of the MIS structure with Al2O3 grown at
9.87W/cm2. The capacitance decreased with increasing frequency and when the frequency was
larger than 2 MHz, the capacitance became constant. Figure 5- 5 (b) shows the G/ω versus
frequency plots for the same structure with different bias voltages from 0.1 V to 0.9 V. The
higher values of capacitance and conductance are obtained at low frequency because the the
population and depopulation of interface traps (Dit) can follow the ac signal in this frequency
range. The maximum peak value of G/ω was measured at 9.26 KHz and gate voltage at 0.1V.
Dit can be estimated from [5]:
(
)
The calculated Dit value for 0.1V and 9.26 MHz is 5.16 x 109 eV-1 cm-2. Comparison with the Dit
value obtained from the single frequency approximation discussed in Section 5.2.2 shows that
the two values are close. Specifically, 1.09 x1010 eV-1 cm-2 and 5.16 x 109 eV-1 cm-2 respectively
were determined with the two methods. However, Dit of the Al2O3 calculated from Figure 5-5(b)
is smaller than the one calculated from Figure 5-4. This is due to the fact that the peak value of
conductance in Figure 5-4 was measured at 1 MHz, while the maximum peak in Figure 5-5 (b)
obtained at the lower frequency of 9.26 KHz. It should be noted that the peak value of
conductance from the single frequency method (Figure 5-4) was obtained at 0.3V, the peak value
of conductance from the frequency dependent method (Figure 5-5 (b)) was in the 0.1V to 0.3 V
range. Thus, the results for the conductance measurements are consistent with different methods.
130
0.0 2.0x106
4.0x106
6.0x106
8.0x106
1.0x107
0.0
0.2
0.4
0.6
0.8
Ca
pa
cita
nce
[n
F]
Frequency [Hz]
0.1V
0.3V
0.5V
0.7V
0.9V
OSC: 100mV
Frequency:100Hz~10MHz
Ag/ZnO/Al2O
3/ITO
(a)
1x106
2x106
3x106
4x106
5x106
1.00E-010
1.50E-010
2.00E-010
2.50E-010
3.00E-010
3.50E-010
G/w
Frequency [Hz]
0.1 V
0.3 V
0.5 V
0.7 V
0.9 V
(b)
Figure 5-5. (a) The capacitance as a function of frequency (b) G/ω vs frequency plots at various
bias from 0.1V to 0.9V.
131
5.2.3 C-V Measurements by Voltage Sweeps and Hysteresis
A hysteresis C-V measurement is used for evaluation of the interface trapped charge and
oxide charges, which govern how rapidly the structure transitions between accumulation and
inversion and therefore provides insight to potential switching applications. The oxide charges
include the fixed oxide charge, the mobile ionic charge and the oxide trapped charge. A
hysteresis in the C-V curve originates from the polarization associated with charging and
discharging during the bidirectional sweep, so that the interface charge and oxide charge can be
estimated from the hysteresis curves. These charges cause a parallel shift in the flat band
voltages. The influence of the charges can be explained qualitatively as an extra gate bias being
needed to achieve the original semiconductor band bending. The trapped charge is calculated
according to:
55
EqqA
VCN FBox
it
here Cox is the previously discussed oxide capacitance measured in the accumulation region,
ΔVFB is the flatband voltage shift as determined directly from the forward and reverse C-V
sweeps, q is the electronic charge and A is the device area which is 0.04 cm2 in this study.
Figure 5-6 shows that the C-V plots of the MIS with 6.91 W/cm2 BaTa2O6 (red curve)
and 7.89 W/cm2 BaTa2O6 (black curve) exhibit strong hysteresis in sweeps from accumulation
through depletion to inversion. For BaTa2O6 films grown at 6.91 W/cm2 and 7.89 W/cm2, the
ΔVFB values from Figure 5-6 are 8 V and 8.99 V respectively. Thus, the trapped charges are
determined to be 3.50 x 1011 cm-2 for the former growth condition and 5.82 x 1011 cm-2 for the
latter. The large ΔVFB observed for these films indicate poor structural quality which is not
132
surprising considering the fact that the depositions were performed without substrate heating,
and that the films had not been subjected to post deposition annealing.
-6 -4 -2 0 2 4 6 8 10
0.0
0.2
0.4
0.6
0.8
1.0
No
rma
iliz
ed
ca
pa
cit
an
ce
[C
/Co
x]
Voltage[V]
MIS with 6.91W/cm2 BaTa
2O
6
MIS with 7.89W/cm2 BaTa
2O
6
VFB
(c)
Figure 5-6. C-V plots of the BaTa2O6 MIS structure grown at 6.91 W/cm2 (red curve) and 7.89 W/cm2 (black curve). Strong hysteresis is observed in sweeps from accumulation through depletion to inversion. The arrows indicate the sweep direction
Figure 5-7 shows the C-V plots of the Al2O3MIS structure grown with 8.88 W/cm2 (red
curve) and 7.89 W/cm2 (black curve) in sweeps from accumulation through depletion to
inversion. The smaller hysteresis area enclosed by the Al2O3 C-V hysteresis curves as shown in
Figure 5-7 confirm a smaller trap concentration than in BaTa2O6. The ΔVFB values for the Al2O3
films are considerably smaller at 0.44 and 0.48 V for the 8.88 W/cm2 and 7.89 W/cm2 sputtering
powers respectively. It indicates that the structural quality of the room temperature deposited
Al2O3 is superior to similarly deposited BaTa2O6. Using the methodology outlined above, the
interface trapped charge for the Al2O3 films was calculated as 1.06 x1010 cm-2 and 1.04 x 1010
133
cm-2 for the Al2O films grown at 8.88 W/cm2 3 and 7.89 W/cm2 respectively. The data are
summarized in Table 5-1 together with the BaTa2O6 analysis for comparison. The results
indicate that overall the BaTa2O6 films contain significantly more defects both at the interface
and the bulk than their Al2O3 counterparts.
-1.0 -0.5 0.0 0.5 1.0
0.0
0.2
0.4
0.6
0.8
1.0
No
rma
iliz
ed
ca
pa
cit
an
ce
[C
/Co
x]
Voltage[V]
MIS with 8.88W/cm2 Al
2O
3
MIS with 9.87W/cm2 Al
2O
3
VFB
(c)
Figure 5-7. C-V plots of the Al2O3MIS grown at 8.88 W/cm2 (red curve) and 7.89 W/cm2 (black curve) in sweeps from accumulation through depletion to inversion. The arrows indicate the sweep direction.
134
Table 5-1. The interface trap density and trapped charge concentration of MIS structures with BaTa2O6 and Al2O3.
MIS with 6.91 W/cm
2
BaTa2O6
MIS with 7.89W/cm2
BaTa2O6
Interface trap density 3.67x1011 ev-1cm-2 6.18x1011 ev-1cm-2
Trapped charge
concentration 3.50 x 1011 cm-2 5.82 x 1011 cm-2
MIS with 8.85W/cm2
Al2O3
MIS with 9.87W/cm2
Al2O3
Interface trap density 2.96 x1010 eV-1 cm-2 1.09 x1010 eV-1 cm-2
Trapped charge
concentration 1.06 x1010 cm-2 1.04 x 1010 cm-2
5.2.4 Frequency Dependent C-V Measurement
High concentration of interface traps leads to frequency dispersion in the C-V behavior.
Frequency dispersion may occur both in the depletion and accumulation regions and provides
insights to how the upper limit of switching speeds that may be influenced by the gate oxide
selected. To evaluate the magnitude of these effects in the oxides studied here, C-V
measurements are performed at 500 kHz, 1 MHz and 2 MHz. Based on the trap concentration
data (see Table 5-1), the BaTa2O6 and Al2O3 grown at 6.91W/cm2 and 9.87W/cm2 respectively
were selected for this study. As Figure 5-8 shows, severe frequency dispersion is observed in the
MIS with BaTa2O6 both in the depletion and accumulation regions. The insert shows the
normalized C-V spectrum and shows how the flatband voltages shift with frequency more clearly.
When the MIS structure is biased into the accumulation region, interface traps are filled by
135
charge. At higher frequency, if population and de-population of the interface trapped charges
cannot follow the ac signal, the measured capacitance will decrease. As Figure 5-8 shows, the
accumulation capacitance of the MIS with BaTa2O6 decreased by a factor of 6 between
measurements at 500 kHz and 1 MHz. Figure 5-9 shows the overlay of C-V measurements at
500 kHz, 1 MHz and 2 MHz for the Al2O3MIS structure. The inserted figure shows the
normalized capacitance. For the MIS with Al2O3, the corresponding decrease was 1.4 times as
shown in Figure 5-9. Overall, the data indicates that not only are there fewer traps at the
Al2O3/ZnO interface compared to those in the BaTa2O6/ZnO structure, but in addition they are
shallow compared to the BaTa2O6.
-4 -2 0 2 4 6 8 10 12 14
0.00
1.25
2.50
3.75
5.00
6.25
7.50(a) Ag/ BaTa
2O
6/ZnO/ITO glass
500 KHz
1MHz
2 MHz
Ca
pa
cit
an
ce
[n
F]
Voltage [V]
-4 -2 0 2 4 6 8 10 12 14
0.0
0.5
1.0
500 KHz
1MHz
2 MHz
No
rmali
zed
Cap
ac
itan
ce
[C
/Co
x]
Voltage [V]
Figure 5-8. The overlay of C-V measurements at 500 kHz, 1 MHz and 2 MHz for BaTa2O6 MIS structure grown at 6.91W/cm2. The inset figure shows the normalized capacitance for the device.
136
-1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0
1.00
2.00
3.00
4.00
5.00
6.00
Ca
pa
cit
an
ce
[n
F]
Voltage [V]
500KHz
1MHz
2MHz
Ag/ Al2O
3/ZnO/ITO glass
-1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0
0.0
0.5
1.0 500KHz
1MHz
2MHz
No
rmali
zed
Cap
ac
itan
ce
[C
/Co
x]
Voltage [V]
Figure 5-9. The overlay of C-V measurements at 500 kHz, 1 MHz and 2 MHz for Al2O3 MIS grown at 9.87W/cm2. The inset figure shows the normalized capacitance for the device.
5.2.5 XPS with Depth Profiling for Composition Determination
To develop qualitative insight to the relationship between the measured electrical
properties and the structure of the materials, the stoichiometry of the BaTa2O6/ZnO and
Al2O3/ZnO stacks were investigated using XPS with depth profiling. Starting with the top
surface, measurements are taken at 2 minutes interval throughout the approximately 120 nm
thickness of the insulator, through the interface with the ZnO, and to a depth of roughly 80 nm
into the ZnO. As previously mentioned, Ar+ ions is accelerated at 2 keV with a beam current of
2 μA used for sputter removal of material. The depth profile measurements of the BaTa2O6/ZnO
bi-layer show that the O(1s), Zn(2p3/2) and Ba(3d5/2) peaks are constant at 530.2 eV, 779.8 eV
and 1021.8 eV, respectively, throughout the thickness analyzed. This is shown in Figure 5-10 (a)
(b) and (c), respectively. For brevity these results are only presented as five peaks for each
element.
137
525 530 535 540
O 1s First measurement of O
Second measurement of O
Third measurement of O
Forth measurement of O
Fifth measurement of O
Inte
ns
ity
[a
.u.]
Binding Energy [eV]
530.2eV
(a)
1020 1025 1030
First measurementof Zn
Second measurement of Zn
Third measurement Of Zn
Forth measurement of Zn
Fifth measurement of Zn
Inte
ns
ity
[a
.u.]
Binding Energy [eV]
Zn 2p3/2
1021.8eV
(b)
138
775 780 785 790
Inte
ns
ity
[a
.u.]
Binding energy [eV]
First measurement of Ba
Second measurement of Ba
Third measurement of Ba
Forth measurement of Ba
Fifth measurement of Ba
Ba 3d 5/2
779.8 eV
(c)
Figure 5-10. (a) XPS Spectra of O(1s) during the depth profile for different etching levels (time) (b) XPS Spectra of Zn (2p3/2) during the depth profile for different etching levels (time) (c) XPS Spectra of Ba (3d5/2) during the depth profile for different etching levels (time)
However, a constant peak position was not the case for tantalum. Representative XPS
spectra for this constituent starting with the topmost surface of the insulator, and penetrating into
the bulk are presented in Figure 5-11. Tantalum oxide peaks corresponding to Ta5+(4f7/2) and
Ta5+(4f5/2) at 25.8 eV and 27.6 eV [6], respectively, are observed at each measurement including
the surface. However, as film layers were removed and the depth profile progressed, two
additional peaks associated with metallic tantalum emerged and can be seen in Figure 5-11.
These metallic peaks corresponding to Ta0(4f7/2) at 21.6 eV and Ta0(4f5/2) at 23.3 eV [6] are
always present in the bulk and interface XPS signals, indicating an oxygen deficiency (Ta rich)
in both regions of the BaTa2O6 film. Figure 5-12 shows that the atomic concentration of
tantalum, barium and oxygen are essentially constant throughout the thickness of the insulator
and at the interface. The Ba/Ta ratio was 0.12-0.13. The results are tabulated in Table 5-2.
Based on these findings and the electrical properties measured and discussed above, it is
139
hypothesized that the oxygen deficiency creates acceptor-type traps that accommodate electrons
in the non-stoichiometric BaTa2Ox films. Y. Lee et. Al [7] have similarly proposed that oxygen
vacancies strongly influence the interfacial and bulk properties of BaTa2O6 films deposited at
low temperature.
20 25 30
0
5000
10000
15000
20000
25000
30000
35000
40000
4f5/2
4f7/2
4f5/2
4f7/2
Ta5+
Tao
Inte
ns
ity
[a
.u.]
Binding Energy [eV]
BaTa2O
6 surface
First measurement in the bulk
Second measurement in the bulk
Third measurement in the bulk
Fourth measurement in the bulk
Figure 5-11. XPS Spectra of Ta 4f7/2 and 4f5/2 during the depth profile for different etching levels
140
20 40 60 80 100 120 140
0
10
20
30
40
50
60
70
80
90
100BaTa
2O
6/ZnO
Ato
mic
co
nc
en
tra
tio
n [
%]
Sputtering level (time) [min]
O
Ta
Ba
Zn
Figure 5-12. The atomic concentration of tantalum, barium, and oxygen are essentially constant throughout the thickness
Table 5-2. The atomic concentration and ratio determined by XPS for Ba and Ta
XPS for BaTa2O6
BaTa2O6 near surface BaTa2O6 bulk
BaTa2O6 interface with ZnO
Ba 3.60±0.1 3.76±0.1 2.18±0.1 Ta 27.03±0.1 31.72±0.1 16.79±0.1
Ratio 0.13±0.1 0.12±0.1 0.13±0.1
With regard with the MIS structures with Al2O3 no shifts are observed for the O(1s) and
Al(2p) peaks at 530.2 eV and 74.9 eV respectively, or the Zn (2p3/2) peak at 1021.8 eV, as is
shown in Figure 5-13 (a) ,(b) and (c). Again, for brevity these results are only presented as five
peaks for each element. Figure 5-14 shows that the atomic concentration of aluminum, zinc and
oxygen were essentially constant throughout the thickness of the insulator and at the interface,
with an Al/O ratio of 0.55-0.60. Table 5-3 shows the atomic concentrations and ratio of Al and
141
O in the Al2O3/ZnO stack. The ratio of Al/O is 2.0/3.3 throughout the thickness of the film
which is close to the stoichiometric 2/3 ratio. The more stoichiometric Al2O3 is consistent with
the smaller density of interface traps and concentration of trapped charge compared to the
BaTa2O6 case.
525 530 535 540
O 1s
530.2 eV
First measurement of O
Second measurement of O
Third measurement of O
Forth measurement of O
Fifth measurement of O
Inte
ns
ity
[a
.u.]
Binding Energy [eV]
(a)
1020 1025 1030
Zn 2p3/2
1021.8eV
First measurementof Zn
Second measurement of Zn
Third measurement Of Zn
Forth measurement of Zn
Fifth measurement of Zn
Inte
ns
ity
[a
.u.]
Binding Energy [eV]
(b)
142
70 75 80 85
First measurement of Al
Second measurement of Al
Third measurement of Al
Forth measurement of Al
Fifth measurement of Al
Inte
ns
ity
[a
.u.]
Binding Energy [eV]
Al 2p
74.9 eV
(c)
Figure 5-13. (a) XPS Spectra of O(1s) during the depth profile for different etching levels (time) (b) XPS Spectra of Zn(2p3/2) during the depth profile for different etching levels (time) (c) XPS Spectra of Al(2p) during the depth profile for different etching levels (time)
20 40 60 80 100 120 140
0
10
20
30
40
50
60
70
80
90
100
Ato
mic
co
nc
en
tra
tio
n [
%]
Sputtering time [min]
O
Al
Zn
Al2O3/ZnO
Figure 5-14. The atomic concentration of tantalum, barium and oxygen were essentially constant throughout the thickness
143
Table 5-3. The atomic concentration and ratio determined by XPS for Al and O
XPS for Al2O3
Al2O3 near surface Al2O3 bulk
Al2O3 interface with ZnO
Al 35.52±0.01 37.52±0.01 17.65±0.01 O 64.63±0.01 62.37±0.01 50.81±0.01
Ratio 0.55±0.01 0.60±0.01 0.35±0.01
5.2.6 Scanning Electron Microscopy
The interface between insulators and semiconductor layers were evaluated with a FEI
Nova 200 NanoLab Dual Beam FIB/SEM. Cross-sectional samples were created by milling with
an accelerated Ga+ ion beam (the focused ion beam). Figure 5-15 shows that the interface
between the insulator and zinc oxide semiconductor; (a) Ag/BaTa2O6(6.91 W/cm2)/ZnO/ITO
glass, (b) Ag/BaTa2O6 (7.89 W/cm2)/ZnO/ITO glass, (c) Ag/Al2O3(8.88 W/cm2)/ZnO/ITO glass
and (d) Ag/ Al2O3(9.87 W/cm2) /ZnO/ITO glass. From the cross-sectional morphologies in
Figure 5-15 (a), (b), (c), and (d), the insulator films deposited at room temperature consist of pin-
hole-free layers. However, no significant observations can be related to the interface traps due to
poor resolution caused by charging associated with the glass substrates.
144
Figure 5-15. Cross-sectional SEM images for (a)Ag/BaTa2O6(6.91W/cm2 )/ ZnO/ITO glass (b) Ag/BaTa2O6 (7.89W/cm2 )/ ZnO/ITO glass (c) Ag/Al2O3(8.88W/cm2)/ZnO/ITO glass (d) Ag/ Al2O3(9.87W/cm2) /ZnO/ITO glass
5.3 Conclusions: Interfacial and Bulk Traps, Their Structural Models and The Implications
for Transparent Thin Film Transistors Based on ZnO and High-κ BaTa2O6 and Al2O3
The interfaces of high-k BaTa2O6 and Al2O3 with ZnO were analyzed using frequency
dependent C-V and G-V measurements in Ag/BaTa2O6/ZnO/ITO and Ag/Al2O3/ZnO/ITO metal-
insulator-semiconductor structures that were sputtering grown between room temperature
(insulators) and 100 oC (ZnO). Although the Al2O3 films exhibit a lower breakdown strength and
catastrophic breakdown behavior compared to BaTa2O6/ZnO interface the Al2O3/ZnO interface
is characterized by more than an order of magnitude small density of interface traps and interface
trapped charge. The BaTa2O6 films in addition are characterized by a significantly higher
concentration of fixed oxide charge. The transition from the accumulation region to the
inversion region in the Al2O3 MIS structure is considerably sharper, and occurred at less than one
tenth of the voltage required for the same transition in the BaTa2O6 case. The frequency
dispersion effects are also noticeably more severe in the BaTa2O6 structures. XPS results suggest
145
that acceptor-like structural defects associated with oxygen vacancies in the non-stoichiometric
BaTa2O6 films are responsible for the extensive electrical trapping and poor high frequency
response. The Al2O3 films were essentially stoichiometric. In summary, the results indicate that
amorphous Al2O3 is better suited than BaTa2O6 as a gate oxide for transparent thin film transistor
applications where low temperature processing is a prerequisite, assuming of course that the
operation voltage of such devices is lower than the breakdown voltage. Also, the operation
powers for the devices with amorphous Al2O3 is lower than the devices with BaTa2O6 due to the
lower fixed oxide charges and interface trap density.
5.4 References
[1] E. H. Nicollian and A. Goetzberger, The Si-SiO2 Interface-Electrical Properties as
Determined by MIS Conductance Technique, Bell Syst. Tech. J., 46, 1055 (1976)
[2] Dieter K Schroder, “Semiconductor material and device characterization”, Chapter 6, 3rd
Edition, John Wiley and Sons (2006).
[3] W. A. Hill and C.C. Coleman, Solid State Electron., Vol 23, pp.987 (1980)
[4] N. Abboud and R. Habchi, Mater. Res. Soc. Symp. Proc. Vol. 1195, 1195-B14-05 (2010)
[5] H. Yang, Y. Son, S. Choi and H. Hwang, Jap. J. Appl. Phy. Vol.44, No.8, pp L1460-L1462 (2005)
[6 ] J. Zhang, I. Boyd Appl. Surf. Sci. 168, pp. 234-238 (2000)
[7 ] Y. Lee, Y. Kim, D. Kim , B. Ju, M. Oh, IEEE Trans. Electrondevice, Vol. 47, NO. 1,pp. 71-
76 (2000)
146
CHAPTER 6
CONCLUSION AND FUTURE WORK
6.1 Conclusions
This dissertation investigated ZnO for potential utility as the channel, and high κ
BaTa2O6 and Al2O3 as the gate oxide for use in transparent thin film transistors. A key
consideration was low temperature processing that is compatible with plastic substrates and
flexible electronics and optoelectronics. A particular emphasis was the electrical properties of
the ZnO/insulator interfaces, which were evaluated with the help of MIS structures.
In the first part of this work, ZnO films were deposited by PLD, ALD and sputtering, and
their structural, electrical, and optical properties as well as chemical composition compared. The
sputtered ZnO films exhibited the best combination of structural, electrical and optical properties
at a low processing temperature, and were therefore selected for the subsequent studies that
investigated the semiconductor/insulator interfacial properties. The highest free carrier
concentrations obtained by sputtering at 100 oC and 200 oC were 3.59 x 1018 cm-3 and 4.24x1018
cm-3, the highest mobilities were of 35.21 cm2/V.s and 32.71 cm2/V.s, and the smallest
resistivities were 0.05 Ω and 0.048 Ω. The proposed conduction mechanisms in the ZnO films
were based on photoluminescence (PL) characterization and composition analysis using XPS.
Specifically, it is proposed that interstitial Zn is the dominant factor to determine the
conductivity in the majority of the ZnO films. The transparency range was between 75 and 90 %
for all of ZnO deposited by sputtering.
147
The BaTa2O6 and Al2O3 insulators were deposited by RF magnetron sputtering at room
temperature. The BaTa2O6 films grown 6.91W/cm2 and 7.89W/cm2, and Al2O3 films
synthesized at 8.88 W/cm2 and 9.87 W/cm2 exhibited the highest breakdown strengths and
lowest leakage currents, and were therefore selected for use in the MIS structures. The leakage
currents of the BaTa2O6 films were on the order of 10-6~10-7 A/cm2 at an applied field of
1MV/cm. The leakage currents of Al2O3 films were on the order of 10-7A/cm2 before breakdown.
Ag/BaTa2O6/ZnO/ITO and Ag/Al2O3/ZnO/ITO MIS structures were used to evaluate the
interfaces of high-κ BaTa2O6 and Al2O3 with ZnO using frequency dependent C-V and G-V
measurements. The BaTa2O6 films were characterized by a significantly higher concentration of
fixed oxide charge. The frequency dispersion effects are also noticeably more severe in the MIS
with BaTa2O6 structures. XPS results suggest that acceptor-like structural defects associated
with oxygen vacancies in the non-stoichiometric BaTa2O6 films are responsible for the extensive
electrical trapping and poor high frequency response. The Al2O3 films were essentially
stoichiometric. In summary, MIS incorporating Al2O3 as the dielectric showed a small interface
trap density and fixed oxide charge. The results indicate that amorphous Al2O3 is better suited
than BaTa2O6 as a gate oxide for transparent thin film transistor applications at low processing
temperatures.
6.2 Future Work
Complete transparent TFTs can be fabricated based on the results above. Transparent
transistors based on oxide semiconductors have recently been proposed using undoped ZnO as
the active channel material [1]. In this study, the ZnO obtained by RF sputtering at a low
148
processing temperature exhibited excellent electrical and optical properties. The mobility of the
sputtered ZnO obtained in this study is 35. 21cm2/V.s, which is comparable with the literature
reports of 20-70 cm2/V.s [2]. A sketch of a ZnO transparent TFT with top gate electrodes is
shown in Figure 6-1(a) and (b). Moreover, high-κ materials were evaluated as replacements for
SiO2 in this study. Besides the interface traps and trapped charge in the bulk of high k materials,
another challenge is a better understand of how the high-κ dielectrics affect the transistor drive
performance, which is directly linked to the carrier mobility in the channel [3]. Mechanisms
such as Coulomb scattering, phonon scattering, surface roughness scattering and charge trapping
contribute to the mobility degradation [4]. These effects are yet to be quantified and should be
performed in future work. These insights are expected to lead to the successful design and
fabrication of transparent TFT devices.
Figure 6-1 (a) a top view for the Transparent TFT (b) a side view for the transparent TFT
149
6.3 References
[1 ] Y. J. Li, Y. W. Kwon, M. Jones, Y. W. Heo, J. Zhou, S. C. Luo, P. H. Holloway, E. Douglas,
D. P. Norton, Z. Park and S.Li, Semicond. Sci. Technol. 20 No 8, 720-725 (2005)
[2] P. F. Carcia, R. S. Mclean, M. H. Reilly and G. Nunes, Jr., Appl. Phys. Lett. Vol.82 No. 7
(2003)
[3] G. Groeseneken, L. Pantisano, L. A. Ragnarsson, R. Degraeve, M. Houssa, T. Kauerauf, P.
Roussel, S. D. Gendt, M. Heyns, “Achievements and Challenges for the Electrical
Performance of MOSFET’s with High-K Dielectrics”, 11th IPFA, Taiwan (2004).
[4] M. A. Negara, K. Cherkaoui, P. Majhi, C. D. Young, W. Tsai, D. Bauzd, G. Ghibaudo, P. K.
Hurley, Microelectronic Engineering Vol. 84, Issues 9-10, 1874-1877 (2007)