Citation
Development of ZnO-Based Thin Film Transistors and Phosphorus-Doped ZnO and (Zn,Mg)O by Pulsed Laser Deposition

Material Information

Title:
Development of ZnO-Based Thin Film Transistors and Phosphorus-Doped ZnO and (Zn,Mg)O by Pulsed Laser Deposition
Copyright Date:
2008

Subjects

Subjects / Keywords:
Annealing ( jstor )
Doping ( jstor )
Electric potential ( jstor )
Electrical resistivity ( jstor )
Lasers ( jstor )
Oxygen ( jstor )
Room temperature ( jstor )
Sapphire ( jstor )
Thin films ( jstor )
X ray film ( jstor )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Embargo Date:
7/24/2006

Downloads

This item has the following downloads:


Full Text












DEVELOPMENT OF ZnO-BASED THIN FILM TRANSISTORS AND
PHOSPHORUS-DOPED ZnO AND (Zn,Mg)O
BY PULSED LASER DEPOSITION

















By

YUANJIE LI


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


2006

































Copyright 2006

by

Yuanjie Li

































To my family















ACKNOWLEDGMENTS

I would like first to express my sincere appreciation to my advisor and committee

chairman, Professor David P. Norton, for providing me all the opportunities, guidance

and motivations throughout my graduate studies. I am grateful for his knowledge and

support that have helped me to finish my dissertation work. I would also like to thank my

supervisory committee members, Professor Stephen J. Pearton, Professor Paul Holloway,

Professor Cammy R. Abernathy and Professor David Tanner, for their suggestions and

guidance. I also want to thank Professor Fan Ren for his discussions and advice on my

research work.

I would like to thank Professor Andrew Rinzler, Dr. Zhihong Chen and Dr. Xu Du

for their help and suggestions on lithography processing. I appreciate Jau-Jiun Chen, Dr.

Valentin Craciun, Dr. Brent Gila, the Major Analytical Instrumentation Center staff and

many other people for their collaborations and assistance. My gratitude also goes to my

group members including Mat Ivill, George Erie, Hyun-Sik Kim, Daniel Leu, Ryan Pate,

Li-Chia Tien, Seemant Rawal, Charlee Calender, and Patrick Sadik for their help during

my graduate research.

I would like to express my deepest appreciation and love to my family members

and friends for their unconditional love and inspiration. Especially, I want to thank my

parents and my husband, Shengbo Xu, for their support and encouragement that have

helped me to overcome many difficulties during my development and make it to this

point in life.
















TABLE OF CONTENTS



A C K N O W L E D G M E N T S ................................................................................................. iv

LIST OF TABLES ............................... .............. ................. vii

LIST OF FIGURES ...................................................... ................... viii

ABSTRACT .............. .......................................... xi

CHAPTER

1 INTRODUCTION ............... ..................................................... 1

2 BA CK GROUN D REV IEW .............................................................................. 5

Properties of Z nO ................................................................ 5
C ry stal S tru ctu re .................................................................................. 5
P hy sical P rop erties ......................................................................... .......... .. .. ... 6
ZnO growth methods .................. ................................................ ......... 8
Z nO Single C crystal .................. .................. .................... .. ........ .... 8
ZnO Thin Film ................................................ ....... .. ........ .... 10
D oping of Z nO ................................................................... 11
Intrinsic defects in undoped ZnO ............................................. ............... 12
N -type doping of ZnO ......................................................... ............... 14
P-type doping of ZnO ........................................... .. ... ...................15
N itro g en d o p in g ............ ........................................................ .. .... .. .. .. .. 16
Phosphorus doping ......................................... .......... .. .... ...... .. 17
A rsen ic d o p in g ....................................................................................... 1 8
N itrogen and group III codoping................................ ....... ............... 18

3 EXPERIMENTAL TECHNIQUES....................................... ......................... 20

Film Growth via Pulsed Laser D eposition............................................................... 20
Post-grow th A nnealing Process............................................................. ... ............ 23
Experimental Characterization Techniques......................................................24
X -ray D iffraction ............................................. .. ...... ................. 24
Scanning Electron M icroscopy....................................... ......................... 25
Energy-dispersive X-ray Spectroscopy .................................... ............... 25
A tom ic Force M icroscopy .............................................................. .. ............. 26


v









H all E effect M easurem ent...................................................................................26
Photolum inescence .................. ............................. .. ....... .. ........ .. 29
X-ray Photoelectron Spectroscopy ............................................................30
Current-voltage M easurem ent................................... .............................. ...... 31

4 DEVELOPMENT OF OXIDE-BASED THIN-FILM TRANSISTORS ...................32

Introdu action ......................................................................................... 32
Transparent Semiconducting Oxides for Thin-Film Transistors.........................34
ZnO-based Transparent Thin-film Transistors..................................................36
Deposition and Properties of Channel Materials...................................................... 38
Fabrication of ZnO -based TFTs .................................... .... ................................. 39
ZnO-based TFT Device Characterization................................47
ZnO-TFTs Using Undoped ZnO as Active Channel Layer...............................49
ZnO-TFTs Using P-doped ZnO and (Zn,Mg)O as Active Channel Layer..........50

5 GROWTH AND CHARACTERIZATION OF P-TYPE PHOSPHORUS-DOPED
(Zn,Mg)O BY PULSED LASER DEPOSITION..................................................53

In tro d u c tio n ................................................................................................................. 5 3
E x p e rim e n ta l ...............................................................................................................5 4
Results and D discussion .................................... ..... .. ...... .............. 55

6 SYNTHESIS AND CHARACTERIZATION OF (Zn,Mg)O:P/ZnO
HETEROSTRUCTURES AND Al-DOPED ZnO ............................................. 67

Intro du action .................................................................................................... 6 7
Experimental ............... ............................................ 68
R results and D iscu ssion .............................. ........................ .. ...... .... ...... ...... 69

7 GROWTH AMBIENT AND ANNEALING STUDY OF PHOSPHORUS-
D O P E D Z n O ...................................................... ................ 82

In tro d u ctio n ........................................................................................8 2
E x p e rim e n ta l .................................................................................................... 8 2
R results and D iscu ssion .............................. ........................ .. ...... .... ...... ...... 83

8 C O N C L U SIO N S ..................... .... ............................ ........... ...... ... ...... 98

L IST O F R E FE R E N C E S ........................................................................ ................... 102

BIOGRAPHICAL SKETCH ........................................................................112
















LIST OF TABLES


Table page

2-1 Physical properties of ZnO and GaN ............................................. ............... 7

2-2 Different epitaxy substrates for ZnO thin film growth. ....................................11

2-3 Valence and ionic radii of candidate dopant atoms.......................................... 17

4-1 Properties of transparent semiconducting oxides....... ........................................ 35

6-1 Growth conditions of 0.01 at. % Al-doped ZnO films via PLD. ..........................70

6-2 FWHM values of ZnO (0002) omega rocking curve for 0.01at.% Al-doped ZnO..74

6-3 Room temperature Hall measurement of 0.01at.% Al-doped ZnO films under
different grow th conditions. .......................................................... .....................76
















LIST OF FIGURES


Figure page

2-1 A schematic illustration of ZnO crystal structure. ................ ................ ..............6

2-2 Schematic of the hydrothermal growth system ...................................................10

2-3 Defect formation enthalpies in Zn-rich and O-rich conditions after LDA
correction s. .......................................................... ................ 14

3-1 Schematic illustration of a pulsed laser deposition system .....................................23

3-2 A schematic of Hall effect on an n-type, bar-shaped semiconductor. The sample
has a finite thickness of d. ............................................................. .....................28

4-1 Schematic illustration of passive and active matrix displays...............................33

4-2 Electrical properties ofundoped ZnO thin films grown on glass at 400C as a
function of oxygen pressure. ............................................. ........................... 40

4-3 X-ray diffraction pattern of the undoped ZnO film deposited at Po2=20mTorr on
g la ss su b state ...................................... ............................. ................ 4 1

4-4 An AFM image of the surface of the undoped ZnO thin film deposited at
Po2=20m T orr on glass substrate ................................................................... ......41

4-5 Schematic cross section view of a top-gate-type TFT structure. ..........................42

4-6 Schematic fabrication sequence of ZnO-based TFT structure..............................46

4-7 Top-view microscopy image of the ZnO-TFTs on glass substrate.......................47

4-8 Drain current as a function of drain voltage characteristics for the undoped ZnO-
TFT ................................ ... ......... ..........................................49

4-9 The output characteristics of the TFT with alternative active channel materials:
(a) P-doped ZnO as the channel; (b) P-doped (Mg,Zn)O as the channel................51

4-10 Transfer characteristics of ZnO-TFT with P-doped (Zn,Mg)O as the channel
layer at the drain voltage of 6V ..................................................... .....................52









5-1 ZnO (0002) omega rocking curves of P-doped (Zno.9Mgo.1)O samples with and
w without L T-Z nO buffer layer. ....................................................... .....................57

5-2 Carrier concentration and carrier type in P-doped (Zno.9Mgo.1)O films as a
function of oxygen partial pressure. ........................................ ....... ............... 58

5-3 Effect of oxygen partial pressure on carrier mobility of P-doped (Zno.9Mgo.i)O
film s .................. ......... .................................................58

5-4 Resistivity of P-doped (Zno.9Mgo.i)O films vs oxygen partial pressure ................59

5-5 X-ray photoelectron spectroscopy survey of P-doped (Zno.9Mgo.i)O films ............62

5-6 X-ray photoelectron spectroscopy multiplex ofP 2s peak for P-doped
(Zno.9Mgo.1)O films grown at 500 C, 150 mTorr oxygen pressure..........................62

5-7 X-ray diffraction of P-doped (Zno.9Mgo.i)O films grown under different oxygen
partial pressures................................... ................................ .........63

5-8 XRD p scans of P-doped (Zno.9Mgo.i)O film and sapphire substrate. ...................63

5-9 AFM images of the P-doped (Zno.9Mgo.1)O films grown at different oxygen
pressures: (a) 20; (b) 100; (c) 150; (d) 200mTorr. ............. .................................... 64

5-10 RMS roughness of the P-doped (Zno.9Mgo.i)O films as a function of oxygen
partial pressure. .......................................................................65

6-1 Schematic illustration of ZnMgO:P/ZnO heterostructure ...................................... 70

6-2 The I-V curve of Au and Ti/Au metal contacts on: (a) p-ZnMgO; (b) n-ZnO
film s, respectively. ........................................... ............... .... ..... .. 72

6-3 Current-voltage characteristics of the ZnMgO/ZnO heterostructure on: (a)
sapphire; (b) ZnO substrate. ...... ........................... ......................................73

6-4 X-ray diffraction of 0.01 at.% Al-doped ZnO films grown under different
conditions. ...........................................................................75

6-5 Omega rocking curve of ZnO (0002) peak for 0.01 at.% Al-doped ZnO films
grown under 300 m J laser energy. ........................................ ....................... 75

6-6 Laser energy effect on RT-PL for 0.01 at.% Al-doped ZnO films grown at (a)
700C ; (b) 800 C .................................................................. ..... .......... 79

6-7 Oxygen partial pressure effect on room temperature PL of 0.01 at.% Al-doped
Z nO film s. ...........................................................................80









6-8 AFM image of 0.01 at.% Al-doped ZnO films grown under different conditions.
The z-scale is 40 nm/div ..... ...... .. ...... ...................... ....... .. ............ 81

7-1 Room temperature resistivity as a function of growth temperature for ZnO: P0.002
film s grown in different gas ambient. ........................................... ............... 85

7-2 Carrier density of ZnO: Po.002 films as a function of growth temperature ...............87

7-3 Carrier mobility of ZnO: Po.002 films as a function of growth temperature.............87

7-4 Photoluminescence spectra of P-doped ZnO grown in: (a) 02/Ar/H2; (b) pure
oxygen; (c) ozone/oxygen. .............................................. ............................. 89

7-5 Resistivity of 0.2 at.% P-doped ZnO films annealed at different temperatures in
02. The films were grown in: (a) 02/Ar/H2; (b) pure oxygen; (c) ozone/oxygen....92

7-6 Carrier concentration of 0.2 at.% P-doped ZnO films annealed at different
temperatures in 02. .. ........... .. ........ ......................................... 93

7-7 Mobility of 0.2 at.% P-doped ZnO films annealed at different temperatures in 02.
The films were grown in: (a) 02/Ar/H2; (b) pure oxygen; (c) ozone/oxygen. .........95

7-8 RT-PL of 0.2 at.% P-doped ZnO films annealed at different temperatures in 02.
The films were grown at 800"C and in 02/Ar/H2 mixture. .....................................96

7-9 Surface morphology of 0.2 at.% P-doped ZnO films annealed at different
temperatures in 02. The films were grown at 700 C and in 60 mTorr 02.....................97















Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

DEVELOPMENT OF ZnO-BASED THIN FILM TRANSISTORS AND
PHOSPHORUS-DOPED ZnO AND (Zn,Mg)O
BY PULSED LASER DEPOSITION

By

Yuanjie Li

May 2006

Chair: David P. Norton
Major Department: Materials Science and Engineering

Top-gate type ZnO-based TFTs were fabricated on glass substrate via

photolithography and wet chemical etching processing. The ZnO layers were deposited

using pulsed laser deposition (PLD). N-channel depletion-mode operation was shown for

the undoped ZnO and P-doped ZnO thin film transistors. The current-voltage

measurements demonstrated an enhancement-mode device operation for the thin film

transistors with P-doped (Zn,Mg)O as the active channel layer.

P-type phosphorus-doped (Zno.9Mgo.i)O films have been realized via PLD without

post-annealing process. The conduction type of the films strongly depends on the oxygen

partial pressure during the deposition process. Increasing the oxygen partial pressure

from 20 to 200 mTorr yielded a carrier type conversion from n-type to p-type. The P-

doped (Zn,Mg)O films grown at 150 mTorr oxygen partial pressure were p-type and









exhibited a hole concentration of 2.7x1016 cm-3, a mobility of 8.2 cm2/Vs and a resistivity

of 35 0 -cm.

(Zno.9Mgo.i)O:P/ZnO heterostructures were fabricated on sapphire and ZnO

substrates via PLD with Au and Ti/Au served as Ohmic contacts. Both structures exhibit

rectifying electrical characteristics. The turn-on voltages were determined to be 1.36 V

and 1.15 V for the structure grown on sapphire and ZnO substrate, respectively.

The resistivity of Al-doped ZnO depends on growth temperature, laser energy and

oxygen pressure. The photoluminescence properties of the Al-doped ZnO films have

strong correlations to the electrical properties and crystallinity of the films. The

possibility of the non-radiative trapping through deep level defect states decreases with

increasing the electron density of Al-doped ZnO films. AFM results showed that the root-

mean-square roughness increases with growth temperature and oxygen partial pressure.

The resistivity of the as-deposited 0.2 at.% P-doped films grown in ozone/oxygen

ambient rapidly increased with growth temperature. The improvement in band edge

emission intensity for the films grown in 02/Ar/H2 mixture may reflect the passivation

effect of the deep acceptor-related levels by hydrogen, which also yields the passivation

of the deep level emission. Oxygen interstitials may contribute to the deep level emission

of RT-PL for annealed P-doped ZnO films.














CHAPTER 1
INTRODUCTION

Semiconductor devices have been exerting a critical influence on our life since the

first transistor was invented at Bell Labs in 1947. The development of advanced

semiconductor materials to obtain desirable properties is one of the essential

contributions to modern semiconductor devices. Silicon (Si) as a conventional

semiconductor material is approaching the theoretical limits by recent technology

advances. To overcome the high power and high temperature limits of Si-based electronic

devices, wide bandgap semiconductors such as silicon carbide (SiC), gallium nitride

(GaN) and diamond have been developed to be the better candidates. For semiconductor

photonic devices such as ultraviolet (UV)/blue light-emitting diodes (LEDs) and laser

diodes, wide bandgap group III nitrides have been the focus of intensive research due to

their specific properties.

II-VI compound semiconductor zinc oxide (ZnO) has attracted much attention

because of its unique combination of electrical, optical, piezoelectric and acoustical

properties for decades. With the development of recent technologies, the research

interests in ZnO are renewed in a broad range of applications from optoelectronics,

transparent thin film transistors (TFTs) and nanostructured materials to spintronics (spin

+ electronics). ZnO has a direct wide bandgap of -3.3 eV at room temperature with a

large saturation velocity and a high breakdown voltage. Compared with GaN, ZnO has

several important advantages making it ideal for UV LEDs and lasers [1]: (1) larger

exciton binging energy of 60 meV (vs. -25meV for GaN) enhancing the radiative









recombination efficiency as well as lowering turn-on voltage for laser emission; (2) the

availability of large single crystal ZnO substrate desirable for vertical device

development; (3) low-temperature epitaxial growth and (4) possible wet chemical etching

process leading to potential low-cost ZnO-based devices. Polycrystalline ZnO gains

much attention in transparent TFTs for the electronic flat panel display industry. Due to

its transparency in the visible spectrum, minimal light sensitivity, and process

temperatures compatible with glass/plastic technology, ZnO-based transparent TFTs

show possible solutions to the limitations of Si-based TFTs [2-4]. With a reduction in

crystallite size to nanometer scale, ZnO introduces novel electrical, mechanical, chemical

and optical properties with rich family of nanostructures such as nanorings, nanowires

and nanobelts [5]. These one dimensional (ID) materials can be used to demonstrate the

potential applications in novel nanodevices. ZnO doped with transition metals such as

manganese (Mn) and cobalt (Co) also shows potential in spintronic applications due to a

predicted Curie temperature above room temperature [6,7].

The motivation for ZnO-based transparent TFTs is to overcome the opacity of Si-

based TFTs for active matrix arrays. In addition, by using ZnO as active channel layer in

TFTs the channel mobility can be increased leading to faster device operation and higher

drive current [8]. Key challenges existing in ZnO-based TFTs include device structure

fabrication and realization of an enhancement-mode device operation by controlled

channel carrier densities. Undoped ZnO is intrinsic n-type with electron concentration in

the range of 1018 cm-3. High electron density causes the channel layer to be conductive

even in the absence of applied gate voltage. Previous results on annealed P-doped ZnO

films showed that phosphorus substitution may introduce an acceptor level that reduces









the native electron density in ZnO. Therefore, to deplete the channel electron carrier and

realize the enhancement-mode operation devices, annealed P-doped ZnO and (Zn,Mg)O

films were employed as the active channel materials in ZnO-based TFTs.

In order to realize the practical applications of ZnO in optoelectronic devices, both

n-type and p-type materials with high carrier concentration and low resistivity have to be

achieved. While n-type ZnO is easily realized via Al or Ga doping, it has shown that ZnO

has the particular difficulty in producing reliable p-type conduction. This critical issue

impedes the widespread development of the ZnO-based UV LEDs and lasers. Therefore,

achieving high-conductivity p-type ZnO has become one of the key challenges for ZnO-

based optoelectronics. Recent studies in p-type doping of ZnO have focused on group V

ions such as nitrogen (N), phosphorus (P) and arsenic (As) substituted on the oxygen site

[9-11]. Thus, the motivation of this part of the dissertation research was to synthesize and

characterize phosphorus-doped ZnO and (Zn,Mg)O thin films for optoelectronic

applications via using pulsed laser deposition (PLD).

This introduction chapter presents the challenges and motivations of this

dissertation work. Following, a second chapter reviews the related background, including

general properties of ZnO; growth methods for single crystal and thin films of ZnO;

current experimental and theoretical studies of n- and p-type doping of ZnO. Chapter 3

describes the film growth and characterization techniques employed in this work. In

chapter 4, the fabrication process and device characteristics of the top-gate type ZnO-

based TFTs on glass substrates are described. Chapter 5 discusses the effect of oxygen

partial pressure on the realization of p-type P-doped (Zn,Mg)O films grown on LT-ZnO

buffer layer. The development of (Zn,Mg)O:P/ZnO heterostructures for light emitting






4


applications is the focus in the following Chapter 6. The growth condition effect on the

electrical and optical properties of Al-doped ZnO is also included in Chapter 6. The

systematic studies of growth condition, post-annealing on the electrical and

photoluminescence as well as crystallinity and surface morphology of P-doped ZnO films

are discussed in Chapter 7. Finally, chapter 8 will give the conclusions of this dissertation.














CHAPTER 2
BACKGROUND REVIEW

This chapter introduces the general properties of ZnO, including its crystal

structure and physical parameters; growth methods for ZnO single crystal and thin films;

n-type and p-type doping of ZnO focused on recent experimental and theoretical studies.

Properties of ZnO

Crystal Structure

Hexagonal wurtzite is the thermodynamically stable crystal structure for ZnO.

There are two other phases known to exist. A zincblende phase is formed under some

specific growth conditions [12]. A rocksalt structure can be synthesized under high

pressure above 10 GPa at room temperature or above 6 GPa at 1200 K [13]. However,

this rocksalt structure is difficult to retain under ambient conditions.

In the hexagonal wurtzite structure, each Zn cation is surrounded by four oxygen

anions at the corners of a tetrahedron, and vice versa. In other words, ZnO crystal

structure is composed of alternating planes of Zn2+ and 02- ions stacking along the c-axis.

Figure 2-1 presents the schematic diagram of ZnO crystal structure. ZnO shows a highly

ionic character due to the large difference in electronegativity between Zn and O atoms.

Non-centrol symmetric tetrahedral coordination in ZnO results in piezoelectric properties

and crystallographic polarity. The oppositely charged zinc-terminated (0001) Zn-face and

oxygen-terminated (000 1) O-face produce spontaneous polarization along the c-axis [5].









Zn2+


[0001]










Figure 2-1. A schematic illustration of ZnO crystal structure.

Physical Properties

The physical properties of ZnO are compared to GaN in Table 2-1 [23-25]. ZnO

and GaN have the same wurtzite crystal structure with -1.9% lattice mismatch on the c-

plane. Therefore, ZnO is a promising substrate candidate for GaN epitaxy due to its

stacking order match and crystal lattice match [14, 15]. High quality, low planar defects

GaN epilayers have been grown on ZnO (0001) substrate via reactive molecular beam

epitaxy (MBE) [16].

ZnO has an exciton binding energy of 60 meV, which is much higher than the

thermal energy at room temperature (26 meV). In principle, the excitonic recombination

in semiconductors is a more efficient radiative process and can enhance low-turn-on

stimulated emission [17-19]. In order to realize efficient excitonic laser action at room

temperature or even higher temperature, it is essential to have an exciton binding energy

greater than thermal energy at room temperature. The first optical pumping laser action in

single crystal ZnO grown by vapor phase method was reported by Reynolds, Look and

Jogai [20]. The lasing occurred at a very low pump power (- 4 Wcm-2) and the as-grown

crystal planes act as the lasing cavity. D. M. Bagnall et al. [21] and P. Yu et al. [22]









reported UV laser emission in ZnO thin films grown on sapphire substrates at room

temperature. Although the electron Hall mobility in single crystal ZnO is lower than that

of GaN, ZnO has a higher saturation velocity allowing it to compete with GaN in

semiconductor devices applications. In addition, ZnO is highly resistant to radiation

damage making it suitable for space and other extreme operating conditions [1].

Table 2-1. Physical properties of ZnO and GaN.
Property ZnO GaN
Crystal structure Wurtzite Wurtzite
Lattice constant at 300K (nm) ao = 0.32495 ao = 0.3189
co =0.52069 co =0.5185
Density (g cm-3) 5.606 6.15
Thermal conductivity (W/cm K) 0.6 1.3
Linear expansion coefficient (K-) a = 6.5 x 10-6 a = 5.59 x 10-6
co = 3.0 x 10-6 co = 7.75 x 10-6
Melting point (C) 1977 2497
Refractive index 2.008, 2.029 2.9
Bandgap at 300K (eV) 3.3 3.39
Exciton binding energy (meV) 60 25
Saturation velocity (cm s-) 3.0 x 107 2.5 x 107
Breakdown voltage (V cm-) 5.0 x 106 5.0 x 106

Alloying ZnO films with CdO (Eg = 2.3 eV) and MgO (Eg= 7.8 eV) makes bandgap

engineering possible for realizing ZnO-based heterostructure devices [26-32]. For many

advanced semiconductor devices, heterostructure designs are one of the key structures to

improve the device performance via introducing band offsets and carrier confinement.

The ionic radii of Cd2+ (0.74A) and Mg2+ (0.57A) are close to Zn2+ (0.60A) [33]. According

to the phase diagram of MgO-ZnO system, the thermodynamic solid solubility of MgO in

ZnO is less than 4 mol% [34]. However, previous work has reported the solid solubility

of MgO in ZnO thin films to be up to 33 mol% while maintaining the ZnO wurtzite

structure via using pulsed laser deposition [29]. Thus, it is possible to synthesize Znj-

Mlu O metastable phases well above the thermodynamic solubility limit by using pulsed









laser deposition techniques. Cadmium (Cd) substitution on the Zn site leads to a

reduction in the bandgap to -3.0 eV [26] and substituting magnesium (Mg) can increase

the bandgap up to -4.0 eV [29]. A quantum-confinement effect by a blueshift in

photoluminescence (PL) spectra was observed in ZnO/Zno.sMgo.20 superlattices grown

by laser molecular beam epitaxy (L-MBE) [35]. S. Choopun et al. summarized the

bandgap relations in Zn,. lMy O as a function of composition via a virtual crystal

approximation [30]. For x = 0 to 0.33, the bandgaps of Zn. lMy O have a linear

dependence on Mg content and the films retain the hexagonal structure. For x = 0.33-0.35,

there is a discontinuity in the bandgap relation due to the structural transition from

wurtzite to cubic phase.

ZnO growth methods

ZnO Single Crystal

There are three important growth methods for bulk ZnO single crystal: pressurized

melt growth [36], seeded sublimation growth [37] and hydrothermal solution growth [38-

40].

Pressurized melt growth employs the use of modified Bridgman configuration

developed by Cermet, Inc. High quality ZnO wafers up to 2-inch in diameter are

commercially available [41]. This technology involves a high pressure induction melting

apparatus, where the melt is contained in a water-cooled crucible [36]. An overpressure

of the oxygen as the growth atmosphere can overcome the ZnO decomposition problems

during heating under normal melt growth pressures. Radio frequency energy is used as

the source to melt the material. Part of the molten phase is cooled by the cold crucible

wall with the same composition as the melt. This cold material prevents the molten

material from direct contacting with the cooling wall surface, which eliminates the









contamination possibilities from the reactive crucible. High crystal quality (linewidths as

low as 49 arcsec) and low defect density (104 cm-2) ZnO crystals can be produced in a

fast growth rate (1-5 mm h-1) by using this method [36].

Seeded sublimation growth or vapor-phase growth technique uses pure ZnO

powder formed by the reaction of Zn-vapor and oxygen as the source material. This ZnO

source is put at the hot end of a horizontal tube which maintains at a certain growth

temperature. Hydrogen is used as a carrier gas during the process to make the growth

reactions achievable. At the hot end of the tube the possible reaction follows ZnO(s) +

H2(g) Zn(g) + H20(g). A reverse reaction takes place at the cold end to form ZnO

assisted by a single crystal seed [37]. High quality ZnO crystal has been grown by Eagle-

Picher Technologies using this method. Room temperature mobility is about 225 cm2 Vs

and an electron concentration is in the range of mid-1016 cm-3 [37].

Several works have reported high quality bulk ZnO single crystals grown

hydrothermally. Hydrothermal autoclaves made of high strength steel were used for

crystal growth. During the growth, high purity ZnO nutrient mixed with the solvent (also

called as mineralizer) is sealed in a platinum (Pt) inner container [40]. The purpose of

using Pt inner container is to isolate the growth environment from the wall of the

autoclave. There are two zones in the Pt container: the crystal growth zone and the

dissolution zone. Figure 2-2 shows a schematic hydrothermal growth system. The major

parameters of this growth technique are the temperature of growth zone, the temperature

difference between the two zones, the pressure and the concentration of the mineralizer

[40]. Good quality ZnO single crystals have been achieved using this technology with a

growth rate about 0.2mm per day [39].










Convection
occurs
Autoclave

Pt inner cot ainer Distilled water
Crystal
growth
Seed crystal zone

Heater


Baffle plate
Dissolution
zone
ZnO material




Figure 2-2. Schematic of the hydrothermal growth system [40].

ZnO Thin Film

ZnO epitaxial thin films have been grown via numerous deposition techniques

including molecular beam epitaxy (MBE), pulsed laser deposition (PLD), radio-

frequency (RF) magnetron sputtering and metal-organic chemical vapor deposition

(MOCVD) etc. All the ZnO and (Zn,Mg)O films presented in this dissertation were

synthesized via using pulsed laser deposition. The detailed description of this technique is

included in the following chapter.

Different substrates for ZnO thin film epitaxy are listed in Table 2-2. The most

common substrate used for ZnO thin films is sapphire due to its hexagonal structure, low

cost and high crystal quality. (0001) c-plane sapphire has been used frequently for

growing epitaxial ZnO films due to the strong tendency of ZnO to grown in c orientation.

Other possible substrates include (11 2 0) a-plane sapphire, (1 1 02) r-plane sapphire and

SiC. Most of the substrates have a large lattice mismatch with ZnO. The lattice-matched









substrate for ZnO is (0001) ScAlMgO4 (SCAM) which consists of alternating stacks of

rock salt layers and wurtzite layers [47]. A number of studies have reported the growth of

epitaxial ZnO and ZnO-based device structures on (0001) SCAM substrate [47-49].

Polycrystalline ZnO films can also be grown on inexpensive substrates such as glass at

low temperatures [55, 56]. By using these characteristics it is possible to realize thin film

transistors by using ZnO as the active channel layer for displays and transparent

electronic devices.

Table 2-2. Different epitaxy substrates for ZnO thin film growth.
Substrate Crystal structure Orientation Mismatch (%) Reference
c-plane A1203 (0001) 18.3 [42-45]
a-plane A1203 hexagonal (11 20) ... [46]
r-plane A1203 (1 02) ... [47]
ScAlMgO4 hexagonal (0001) 0.09 [48-50]
GaN wurtzite (0001) 1.9 [51]
SiC wurtzite (0001) 5.5 [52]
Si diamond (100) ... [53,54]

Doping of ZnO

To realize ZnO-related materials in electronic and photonic applications, both high

quality, low resistivity n- and p-type ZnO have to be achieved. However, wide bandgap

semiconductors such as ZnSe, ZnS, ZnTe or CdS can be doped either n-type or p-type,

but not both [57]. ZnSe and GaN can be easily doped n-type, but p-type doping is

difficult to be realized. On the contrary, ZnTe is hard to dope n-type while p-type is

formed readily. This asymmetry in n-type versus p-type doping also exists in ZnO

[58,59]. While n-type ZnO is easily realized via Al, Ga or In doping, ZnO exhibits

significant resistance to the formation of shallow acceptor levels. To explore and realize

the p-type doping of ZnO, it is essential to understand intrinsic defects in undoped ZnO.

The following sections discuss the intrinsic donors and acceptors in undoped ZnO,









introduce the n-type doping of ZnO and largely focus on the recent experimental and

theoretical studies of p-type ZnO doped with nitrogen (N), phosphorus (P), arsenic (As)

as well as co-doped with N and group III elements.

Intrinsic defects in undoped ZnO

Undoped ZnO is an n-type semiconductor with various electron concentrations

from mid-1016 cm-3 in high quality single crystal [60] to -1018 cm-3 in epitaxial thin films

grown on sapphire substrates. The origins of the dominant donors in undoped ZnO have

been investigated both theoretically and experimentally [61-68]. In the study of native

defects in ZnO by Kohan et al. [61] the most dominant native defects are suggested to be

Zn and O vacancies depending on the Zn partial pressure. The calculations based on the

first-principles pseudopotential method indicated that in Zn-rich conditions oxygen

vacancy (Vo) has lower energy than zinc interstitial (Zni) for all Fermi-level positions.

For oxygen-rich conditions, zinc vacancy (Vzn) dominant over the whole range of Fermi

level range. Zhang et al. [62] calculated the formation enthalpies of the intrinsic defects

in ZnO such as Vo, Zni, Zn-on-O antisite (Zno), Vzn and oxygen interstitial (Oi) based on

the plane-wave pseudopotential total energy and force method. Figure 2-3 shows the

defect formation enthalpies after the local density approximation (LDA) corrections. It

suggests that both Zni and Vo have low formation enthalpies. In contrast, the native

acceptor defects Vzn and Oi have high formation enthalpies for Zn-rich conditions.

Therefore, undoped ZnO shows intrinsic n-type conduction and can not be doped p-type

by these intrinsic acceptor defects. The study, however, suggested that Zni is a shallower

donor than Vo, thus it is considered to be the dominant intrinsic donor in ZnO. Look et al.

[63] studied high-energy electron irradiation on both ZnO (0001) and (000 1) face









samples and found the donors and acceptors production rate is much higher for Zn-face

irradiation. The donor is assigned to a Zn-sublattice defect which has an activation energy

of about 30 meV suggesting that Zni (and not Vo) is the dominant native shallow donor in

ZnO. Recent work by Janotti et al. also indicated that Vo is too high in energy to play any

significant role in as-grown n-type material. However, it is expected to act as a

compensating center in p-type ZnO and it can also be formed during irradiation and ion

implantation [64].

Several experimental and theoretical studies have shown that there is another

candidate for the dominant donor defect in undoped ZnO [65,66]. First-principles

calculation based on density functional theory (DFT) found that high electron

concentration (1017 -1018 cm-3) in undoped ZnO may result from the presence of

hydrogen (H) as a shallow donor [65]. Hydrogen forms a strong bond with oxygen

providing a strong driving force to incorporate into ZnO. The incorporation of hydrogen

in ZnO results in significant relaxations of surrounding atoms and always occurs as a

positive charge. The formation energy of H+ is low to allow a large solubility of hydrogen

in ZnO. Since hydrogen is generally present in the films growth ambient and has a high

diffusion coefficient into ZnO, control of hydrogen exposure during film growth has to be

carefully carried out.

The native acceptors in undoped ZnO including zinc vacancy (Vzn) and oxygen

interstitial (Oi) have high formation enthalpies in Zn-rich conditions, so these defects are

not abundant [61]. Recent work by Tuomisto et al. [68] studied the dominant acceptor in

undoped ZnO by using temperature-dependent Hall measurements and positron

annihilation spectroscopy. It was shown that Vzn densities in as-grown and irradiated






14


ZnO have a good agreement with the total acceptor densities determined by temperature-

dependent Hall measurements. Thus, Vzn is identified as a possible dominant acceptor in

both as-grown and irradiated ZnO.


1 2 30 1
Ec E

Fermi Energy E (eV)


2 3


Figure 2-3. Defect formation enthalpies in Zn-rich and O-rich conditions after LDA
corrections [61].

N-type doping of ZnO

Group III elements such as Al, Ga and In can act as extrinsic donors for ZnO by

substitution on Zn site. High quality and high electron concentration n-type ZnO epitaxial

films have been successfully synthesized by using MBE, sputtering or MOCVD [69-72].

Highly conductive and transparent n-type ZnO films have been utilized as a potential

candidate to replace indium tin oxide (ITO) for displays and solar cells. Miyazaki et al.









[69] reported transparent conductive Ga-doped ZnO films deposited on glass substrate

having a minimum resistivity of 2.2 x 10-4 0 cm at substrate temperature of 250C.

Highly conductive Al-doped ZnO thin films with average transmittance over 91% in the

visible spectrum region has been deposited by using photo-MOCVD method [70]. Chang

et al. [72] also synthesized Al-doped ZnO thin films on silicon and glass substrates by RF

magnetron sputtering. The dependence of film properties on different growth parameters

such as RF power, substrate temperature, oxygen pressure and target composition was

examined systematically.

P-type doping of ZnO

As mentioned earlier, theoretical and experimental studies have shown the

asymmetry of n-type versus p-type doping of ZnO. There are several possible

mechanisms to explain these difficulties [73-75]. First, the acceptor dopants may have

low solubility in the host material to limit the accessible hole concentration. In ZnO, the

native donor defects have low formation enthalpies and thus compensate extrinsic

acceptors. Zunger [76] proposed that in ZnO, the p-type pinning energy EF() where the

native "hole killers" spontaneously form to compensate p-type doping is considerably

above the valence band maximum (VBM). The Fermi energy EF in extrinsic p-type

doping which moves downwards will encounter EF() first before encountering the VBM.

As a result, "hole killers" such as Zni and Vo will be generated spontaneously before any

significant doping commences. Moreover, acceptor dopants can form deep levels in ZnO

that are not easily activated at room temperature. The doping difficulty in p-type ZnO

may also result from other compensation mechanisms such as the formation of deep

defect AX centers through a double broken bond (DBB) mechanism [73]. The net result









of the formation of these defect complexes is releasing two electrons (equivalent to

capture of two holes). Recent research in p-type dopants has mainly focused on group V

ions such as N, P, As or Sb substituted on the O site. The following sections will present

these results.

Nitrogen doping

Table 2-3 lists the valence and ionic radii of candidate acceptor dopants. Among

the group V atoms, N has the closest ionic radius to that of O. Theoretical studies also

suggested that N has some advantages for p-type doping of ZnO such as the smallest

ionization energy and metastable N AX complexes [73]. There have been significant

activities focused on N doping by using different nitrogen sources including NH3, N2,

N20, NO and Zn3N2 [77-88]. Minegishi et al. [77] reported p-type ZnO doped with N by

simultaneous addition of NH3 in hydrogen and excess Zn in source ZnO powder. A

proposed model explained that the high resistivity of about 100 0 cm may be reduced via

thermal annealing. Look et al. [78] reported p-type ZnO grown by MBE using N2 RF

plasma source and Li-diffused semi-insulating ZnO substrates. The N-doped ZnO layers

showed a hole concentration of 9 x 1016 cm-3, a mobility of 2 cm2/Vs and a resistivity of

40 0 cm. In another case, p-type ZnO has been grown by PLD, in which a N20 plasma is

used for doping [80]. Iwata et al. [81] grew N-doped ZnO using MBE by introducing N2

and 02 through a RF radical source. However, type conversion from n-type to p-type did

not occur even though the N concentration was in the range of 1019 cm-3. It is suggested

that the formation of N-N related complexes introduce compensating defects. Thus,

dopant sources contain only one nitrogen atom such as NO and NO2 are considered to be

better choices due to the large dissociation energy of N-N in N2 (9.76 eV) [75]. Yan and









co-workers [82] proposed a theoretical model predicting that the defect formation energy

of N on O site (No) from NO is negative in Zn-rich conditions and lower than that from

N20. Li et al. [83] reported on the realization of p-type ZnO using NO gas as the dopant

source by MOCVD. The carrier concentration is in the range of 1015-101 cm-3 and

mobility is in the range of 10-1 cm2/Vs.

Table 2-3. Valence and ionic radii of candidate dopant atoms [75].
Atom Valence Radius (A)
Zn 2+ 0.60
Li 1+ 0.59
Ag 1+ 1.00
O 2- 1.38
N 3- 1.46
P 3- 2.12
As 3- 2.22
Sb 3- 2.45

Phosphorus doping

While there have been many activities focused on nitrogen doping, few reported

efforts have addressed phosphorus doping. Nevertheless, Aoki et al. reported p-n

junction-like behavior between an n-type ZnO substrate and a surface layer that was

heavily doped with phosphorus [89]. Zinc-phosphide (Zn3P2) used as the phosphorus

dopant source was decomposed into Zn and P atoms by excimer laser radiation in high

pressure oxygen or nitrogen ambient. Light emission also was observed by forward

current injection at 110 K indicating that a p-type ZnO was formed.

Work by K. Kim et al. on P-doped ZnO films grown by sputter deposition showed

that n-type as-grown films were converted to p-type by a thermal annealing process to

activate the P dopant [90]. In this case, phosphorus was doped in ZnO using the ZnO

target mixed with 1 wt% P205. XPS results of the p-type ZnO showed that the P2p peak

came from P205 in the film. The p-type films showed a low resistivity of 0.59-4.4 0 cm,









a hole concentration of 1 x107-1.7 x 1019 cm-3 and a mobility of 0.53-3.52 cm2 /Vs. These

results indicate that phosphorus is another promising acceptor dopant for p-type ZnO.

Our previous annealing studies on P-doped ZnO films grown on sapphire substrates

by PLD showed that post-growth annealing process yielded semi-insulating behavior

which is consistent with activation of a deep acceptor level [91].

Arsenic doping

Although density functional theory (DFT) predicts that Aso should have a very

deep acceptor level (-1.15 eV) based on the large ionic size difference between As and O,

Ryu et al. [92] have synthesized p-type ZnO by As diffusion from a GaAs substrate. The

hole concentration can be increased up to the mid- 1017cm3 range. Homostructural ZnO

p-n junctions based on As-doped ZnO has also been reported [93]. Based on first-

principles calculations, a model for large-size-mismatched group-V dopants such as As

and Sb in ZnO was proposed by Limpijumnong and co-workers [94]. These dopants do

not occupy the O sites, but on the Zn sites: each forms a complex with two spontaneously

induced Zn vacancies in a process that involves fivefold As coordination. The model also

predicted that p-type ZnO doped with As could be realized under O-rich growth or


annealing conditions.

Nitrogen and group III codoping

A codoping method refers to using N acceptors and group III donors such as Ga, Al

or In as the reactive codopant to increase the N incorporation in ZnO [95]. The theoretical

calculations showed that codoping wurtzite ZnO with N and group III elements, the

distance between two N acceptors decreases from 6.41 A to 4.57 A indicating the

enhancement of N incorporation. It also showed that the simultaneous codoping









decreases the Madelung energy of p-type codoped ZnO compared with p-type ZnO doped

only with N. Several research groups reported the realization of p-type ZnO using

codoping method. Joseph et al. [96] demonstrated p-type ZnO films by applying Ga and

N codoping method. A resistivity of 0.5 Qcm and a carrier concentration of 5 x 1019 cm-3

have been obtained in ZnO films on glass substrate. Singh et al. [97] also reported p-type

conduction in ZnO by using N and Ga codoping technique. Ye and co-workers [98,99]

reported p-type ZnO thin films realized by the N-A1 codoping method. Secondary ion

mass spectroscopy showed that the N incorporation was enhanced by the presence of Al

in ZnO. The lowest room temperature resistivity is 57.3 Q, with a Hall mobility of 0.43

cm2 /Vs and carrier concentration of 2.25 x 1017 cm-3 for the N-A1 codoped film grown on

glass substrate.














CHAPTER 3
EXPERIMENTAL TECHNIQUES

This chapter describes the details of the experimental techniques and procedures

applied to the deposition of phosphorus-doped ZnO and (Zn,Mg)O thin films, post-

growth annealing process and characterization measurements for structural, surface

morphology, transport, optical as well as chemical state properties of the films.

Film Growth via Pulsed Laser Deposition

Pulsed laser deposition has been used for epitaxial growth of thin films and

multilayer/superlattice of complex materials. It provides many important advantages for

oxide films with high melting points and complicated stoichiometry, some of which can

not be achieved by using other growth techniques such as sputtering, MBE or MOCVD

[100]. PLD was the first method used to successfully deposit high-temperature

superconducting thin films [101,102]. The main advantage of PLD derived from the laser

material removal mechanisms. A pulsed laser ablates the target with very high peak

energy to evaporate the target material without change in the target composition. The

evaporants, also called a plume, arrives at the heated sample substrate and then starts the

film growth. As a result, films with the desirable stoichiometry similar to the target can

be obtained by using PLD. Other advantages of PLD include the easy adaptation to

different operational modes with no constraints from the internal powered sources and the

ability to change the deposition gas pressure over a broad range as well as a high

deposition rate [100].









In principle, the film deposition process in PLD can be divided into three separate

stages: (1) the interaction of the laser beam with the target material; (2) dynamic

formation of plasma; (3) plasma isothermal expansion and deposition of thin films [103].

In the first stage, the target material is rapidly heated above its melting point by the high

peak energy laser pulses. The evaporation of the target material with the same

stoichiometry as the target occurs. The turn-on energy defined as the minimum laser

energy above which appreciable evaporation is observed depends on the laser wavelength,

pulse duration, plasma losses as well as the optical and thermal properties of target

material [103]. The interaction of the laser beam with the target involves several

mechanisms such as thermal, collisional, electronic, exfoliational and hydrodynamic

sputtering [100].

The evaporated mixture of energetic species including atoms, molecules, electrons,

ions, and micron sized particulates further interacts with the laser beam by absorbing the

laser energy to generate a high-temperature plasma. The particle density in the plasma

depends on the ionization degree, evaporation rate and the plasma expansion velocities.

Most of the evaporated species are deposited perpendicular to the laser spot. However,

the thickness variation existing in the films is larger than that from a conventional

thermal evaporation process [103].

The plasma expansion is isothermal in vacuum due to the termination of the laser

pulse. With the kinetic energy in the plasma, it retains its elliptical shape during the

deposition process. Nucleation and growth of film starts after a condensation region on

the substrate is formed. Substrate temperature and surface mobility of the deposited

species affect the quality of thin film. Two dimensional, i.e. layer-by-layer growth is









more preferable compared to the three dimensional (island) growth in order to obtain

high quality epitaxial thin films.

Compared to the laser-target interaction and film deposition mechanisms, the

system setup is much simpler. A PLD system is usually composed of an excimer laser,

optical elements to guide and focus the laser beam and a vacuum chamber. Figure 3-1

shows a schematic of a PLD system along with the laser and optic lenses. In this work, a

Lambda Physik (COMPEX 205) KrF excimer laser which delivers 248 nm wavelength

and maximum average power of 50 W was used. The laser influence is in the range of 1-3

J/cm2 and a laser repetition rate from 1 tol0 Hz was used for the film growth. Multiple

targets can be installed inside the chamber simultaneously to realize multilayer and

superlattice structures with a target to substrate distance of 4-5 cm. A quartz lamp heater

was used to heat sample substrates up to 950 C. The chamber pumping system consists of

an oil-free diaphragm roughing pump and Pfeiffer turbo pumps. The base pressure of the

growth chamber was in the range of 8 x 108 -2 x 10-7 Torr. By using a mass flow control

valve (MKS 600 series) it is possible to adjust input of different gases into the chamber in

a wide range of deposition pressures from high vacuum (-10-7) to 10-1 Torr.

There are several factors that play important roles in obtaining high quality thin

films grown via PLD. The laser energy density on the target has a significant effect on

particulate formation. In order to make laser ablation occur, the laser energy has to

exceed the turn-on energy of the target material. However, if the laser energy is much

greater than the turn-on energy, the density of the particulates will increase. Therefore, it

is important to use the optimal laser energy density to obtain high quality films. An

external energy meter was used to determine the actual laser energy reaching the target in









Laser Beam




focusing Lens

Target Carousel aser
Target







eater




Figure 3-1. Schematic illustration of a pulsed laser deposition system.

front of the chamber laser window. The laser energy can be adjusted by panel settings

based on the external energy meter values during every experimental growth. In addition,

pre-ablation of the target is necessary to have a stable deposition rate and decrease the

particulate formation. Pre-ablation of the targets with a high laser repetition rate of 10 Hz

was used for 1000-2000 laser shots before the films growth. Deposition gas pressure and

substrate temperature also influence the films growth rate, the kinetic energy distribution

of depositing species, as well as the films crystal structure [100]. Therefore, it is essential

to carry out the systematic studies of these growth condition effects on the film properties.

Post-growth Annealing Process

There are two different post-growth annealing processes conducted in this

dissertation work: in-situ chamber annealing and tube furnace annealing. In-situ chamber

annealing was carried out after film growth. The annealing gas pressure could be









increased to above 100 Torr. This annealing process overcomes the possible surface

contamination imposed by taking samples out the vacuum chamber for post-annealing.

However, the annealing gas pressure is limited to certain levels. Compared to chamber

annealing, tube furnace is more versatile in realizing high temperature (-1100C),

controlled gas pressure (up to 1 atmosphere) and different ambients annealing. An

alumina crucible containing the samples was located at the center of a quartz tube, where

the temperature has been calibrated. The annealing tube was purged with high purity

annealing gas for up to 8 hours before increasing the temperature. Desirable heating and

cooling rates can be realized by using a pre-programmed temperature controller.

Experimental Characterization Techniques

This section describes the different characterization techniques used to investigate

the structural, surface morphology, transport, optical and chemical bonding properties of

the as-grown and annealed films. The effects of growth pressure, temperature, ambient,

dopant concentration and annealing conditions on film properties were explored.

X-ray Diffraction

X-ray diffraction (XRD) is an important technique to analyze the crystallinity,

phase, strain, preferred orientation and defects of samples. A collimated beam of X rays

is incident on a sample and diffracted by the crystalline phases in the sample according to

Bragg's law such that

2 = 2dsin 0 (3-1)
where A is the wavelength of the incident X-ray, d is the spacing between atomic planes

in the crystalline phase, 0 is the angle between atomic planes and the incident X-ray

beam. The intensity of the diffracted X-rays is measured as a function of the diffraction

angle 20. This diffraction pattern is used to identify the sample crystalline phases [104].









A Philips APD 3720 X-ray diffractometer with Cu Ka (A =1.5406A) was used to

examine the films structure and crystallinity in this work. The 20 scan range is from 30

to 75 degree. Omega rocking curve was measured by using a four-circle Philips X'pert

X-ray diffractometer. The full-width-at-half-maximum of the rocking curve was used to

delineate the films crystallinity.

Scanning Electron Microscopy

Scanning Electron Microscopy (SEM) is a useful instrument to examine the sample

surface when the light microscope reaches its resolution limits. In the SEM, an electron

beam is focused into a fine probe and scanned over a small area on the sample. Different

signals such as secondary electrons, photon emissions as well as internal currents are

created by the electron beam and sample interactions. These emitted signals are collected

by detectors and subsequently an image is produces on a cathode ray tube (CRT) [104].

A JEOL 6400 scanning electron microscopy was used to investigate the sample

surface in low and high magnifications. The SEM was operated at 10-15 keV depending

on the conductivity of the sample surface. The sample surface was not coated with a

carbon layer and preserved in its original status for other characterization measurements.

Energy-dispersive X-ray Spectroscopy

Qualitative and semi-quantitative analysis of elements presenting in a sample can

be achieved by using energy-dispersive X-ray spectroscopy (EDS). When atoms are

ionized by a high-energy radiation they emit characteristic X-rays. A solid state detector

made from Si(Li) is used to convert these X-rays into signals which are then processed

into an X-ray energy spectrum. Most applications of EDS are in electron column

instruments like SEM, field emission-SEM, transmission electron microscope [104]. In









this dissertation work, EDS was used to determine the composition of the surface

particulates as well as the films. However, the relative accuracies are about 10% for the

elements with concentration less than 5 wt% [104].

Atomic Force Microscopy

In order to measure the surface roughness and topography with atomic resolution,

atomic force microscopy (AFM) is a valuable technique. Unlike electron microscopes,

the resolution of AFM is determined by the size of the tip instead of electron beam

diffraction effects. In addition, a wide range of samples including metals, polymers,

glasses, semiconductors, thin films and composites can be measured in air even in liquids.

AFM works by measuring attractive or repulsive van der Waals forces between the atoms

of a tip and sample surface [105]. The magnitude of the deflection resulting from these

forces depends on the tunneling current and the tip-to-sample distance. A highly position-

sensitive photodiode detects the tip cantilever deflection and converts the signal to an

image.

In this work, AFM Dimension 3100 (Digital Instrument, Inc.) was performed in

contact mode to obtain the surface topographic images and roughness of as-grown and

annealed samples under different conditions. Root-mean-square (RMS) roughness was

calculated based on 2 x 2 or 5 x 5 1- m2 scan area. The scan rate is in the range of 1-1.5

Hz.

Hall Effect Measurement

Since its discovery by Edwin H. Hall (1879) the Hall effect has become one of the

most important electrical characterization methods of materials. The Hall effect provides









a relative simple way of measuring the carrier density, electrical resistivity and the

mobility of carriers in semiconductors.

The principle underlining the Hall effect is Lorenz force which is defined as a force

exerted on a charged particle in an electromagnetic field. Figure 3-2 shows an n-type,

bar-shaped semiconductor. It is assumed that a constant current I flows along the x-axis

from left to right in the presence of a z-directed magnetic field. Under Lorenz force

electrons drift toward the negative y-axis and accumulate on the side of the sample to

produce an electrical surface charge. As a result, a potential drop across the sample called

Hall voltage is formed. The induced electric field increases until it counteracts to the

opposite Lorenz force. In this case,

eEy = evuB = -eBjx /ne (3-2)
where eEy is the induced electric field force, evB is the Lorenz force,jx=-nev, is the total

current density. The Hall coefficient RH is defined as


RH =- (3-3)
ne
The mobility is defined as the coefficient of proportionality between v and E and

measured as follows:

S= = =- RHor (3-4)
E neE
where a is the conductivity. For p-type semiconductors, a hole has a positive charge e.

Therefore, the Hall coefficient is positive in sign.

The Van der Pauw technique which requires no dimension measured for the

calculation of sheet resistance or sheet carrier density solves the potential problem in a

thin layer of arbitrary shape. Thus, this method has increased in popularity relative to the

Hall-bar configuration. The validity of the van der Pauw method requires that the sample










Lorentz Force Coordinate
Coordinate
F=-qv x B System
B

Y
v B
F l V=0










Figure 3-2. A schematic of Hall effect on an n-type, bar-shaped semiconductor. The


Figure 3-2. A schematic of Hall effect on an n-type, bar-shaped semiconductor. The
sample has a finite thickness of d [106].

be flat, homogenous, isotropic, a single connected domain and have line electrodes on the

periphery [107]. In addition, the contact size has to be sufficiently small to reduce the

measurement corrections.

In this dissertation work, a Lake Shore 7500/9500 Series Hall System was used to

perform Hall measurements on the samples at room temperature. The van der Pauw

method was applied for the measurements. The samples have square geometry with

contact size to sample periphery ratio as low as possible. In order to eliminate the

persistent photoconductivity (PPC) relaxation effect on the transport properties, the

samples were maintained in the dark for 12 h prior to performing Hall measurements.

PPC effect has been observed in many III-V and II-VI compound semiconductors and

can be explained by several theoretical models [108-110]. In many cases, PPC is

attributed to the existence of deep defects such as DX centers which form when shallow









donors undergo a large lattice relaxation and convert to deep donors [108]. The difference

in lattice relaxation between these two states results in a barrier that prevents the

recapture of carriers into the stable state, thus yielding the PPC effect [109]. Other

possible models include band-bending resulted from the interface leading to PPC effect

and random local-potential fluctuation model inducing the separation of photoexcited

carriers from traps and reduced recapture rate of carriers.

Photoluminescence

Photoluminescence (PL) refers to emission of light resulting from optical

stimulation. The detection and analysis of this emission is widely used as an analytical

tool due to its sensitivity, simplicity, and low cost [104]. When an electron increases

energy by absorbing light there is a transition from the ground state to an excited state of

an atom or molecule. This excited system does not have the lowest energy and has to

return to the ground state. In luminescence materials the released energy is in the form of

light, which is called as radiative transition. This emitted light is detected as

photoluminescence, and the spectral dependence of its intensity provides information

about the properties of the materials. Specially, photoluminescence of a semiconductor is

related to both intrinsic and extrinsic defects in the material which usually create discrete

electronic states in the bandgap and therefore influence the optical emission of the

material.

PL can be performed both at room temperature and low temperatures. At room

temperature, PL emission is thermally broadened. With decreasing temperature, PL peaks

tend to be much sharper and the emission intensity is also stronger due to fewer

nonradiative transitions. PL is normally useful for semiconductors which have direct









bandgap. However, at low temperatures, localized bound states and phonon assistance

allow certain PL transitions to occur even in materials with indirect bandgap [104].

A He-Cd (325 nm) laser was used as the excitation source in room temperature PL

measurements for the samples presented in this dissertation. The measurements were

taken in a wavelength range of 340 to 800nm. A NESLAB chiller cooled GaAs PMT

detector was used for UV up to -900 nm.

X-ray Photoelectron Spectroscopy

X-ray photoelectron spectroscopy (XPS) is the most broadly applicable surface

analysis technique today due to its surface sensitivity, quantitative and chemical state

analysis capabilities [104]. All elements except for hydrogen and helium can be detected

by using XPS. In XPS monoenergetic X-rays bombard a sample and cause photoelectrons

to be ejected. Einstein photoelectric law in Equation (3-3) defines the relationship of the

kinetic energy of the ejected photoelectrons and the binding energy of the particular

electron to the atom.

KE hv -BE (3-3)
where KE is the kinetic energy of the photoelectron, hv is the X-ray photon energy and

BE is the binding energy. By measuring the photoelectron kinetic energy the

characteristic binding energy of the electron in the atom can be determined. The depth of

the solid samples varies from the top 2 atomic layers to 15-20 layers.

Perkin-Elmer PHI 5100 ESCA system was used to examine the chemical

environment of P in P-doped ZnO and (Zn,Mg)O films. The base pressure of the XPS

chamber was about 1 x10-9 Torr after increasing the voltage to 15 kV and the power to

300 Watt. Both Mg and Al anodes were used to acquire XPS spectrums of different






31


samples. Argon gas sputtering to remove the sample surface contamination is necessary

to obtain more accurate information about the film composition.

Current-voltage Measurement

Current-voltage (I-V) measurements were carried out at room temperature to

characterize the performance of the devices including ZnO-based thin film transistors and

(Zn,Mg)O/ZnO heterostructures. The electrical properties of ohmic contacts for Hall

measurements were also examined. Semiconductor parameter analyzers Agilent 4155A

and HP 4145B connected to a probe station were used in this dissertation work.














CHAPTER 4
DEVELOPMENT OF OXIDE-BASED THIN-FILM TRANSISTORS

Introduction

Display technologies based on organic/polymer light emitting diodes

(OLED/PLED) are promising for providing lightweight, power efficient, flexible, high

brightness performance. One challenge facing polymer light emitting devices is the thin-

film transistors (TFTs) array control circuitry [111-116]. In active matrix displays, each

pixel is programmed to provide a desirable current during the entire frame time,

eliminating the issue of continuous increased current density encountered in the passive

matrix approach. Figure 4-1 shows a schematic illustration of passive and active matrix

displays. The light-emitting pixels in active matrix displays may be controlled by a thin

film transistor array for much better brightness and efficiency. However, current opaque

TFTs made with amorphous and poly silicon severely restrict the amount of light

detected by observer. Depending upon the design of the array and interconnects, only a

fraction of the emitted light is used by the observer of the information resulting in

significant energy loss. In addition, TFTs based on amorphous Si have other limitations

such as light sensitivity, light degradation and low field effect mobility ( 1 cm2/Vs) [11].

One method to overcome these problems is to utilize the recent progress in transparent

oxides that are semiconducting to near-metallic.

Transparent conducting oxides (TCOs) can be regarded as a specific group of

oxides exhibiting both high optical transmittance and electrical conductivity. There are

many applications for TCOs such as transparent electrodes for flat-panel display and










Pas siv-matrix

Column 4il etrod*e

i4 r








Active-matrix *.=- -...






-X -









Figure 4-1. Schematic illustration of passive and active matrix displays [117].

solar cell, transparent electronic devices and selective window coatings in architecture.

N-type TCOs whose dominant carriers are electrons include indium tin oxide (ITO),

SnO2, impurity-doped ZnO, CdO, as well as their related multicomponent oxides [117-

119]. As used for thin film transparent electrodes, TCOs should have a carrier

concentration of order of 1020 cm-3 or higher and bandgap above 3 eV [120]. With the

increasing need for transparent electrodes for optoelectronic devices, it is essential to

develop alternatives to ITO due to its expensive indium source. Ga- and Al-doped ZnO

with a resistivity on order of 10-5 0 cm are promising candidates for thin film transparent

electrode applications [121]. In addition to binary compounds, ternary compounds such









as Cd2SnO4, CdSnO3, Cdln204, Zn2SnO4, and CdSb206 have been developed to be used

as n-type TCOs [120]. On the other hand, p-type TCOs such as CuA102, CuGaO2, and

SrCu202 have been demonstrated recently [122-126], offering the promise of transparent

two- and three-terminal devices. In fact, p-n junctions from n and p-type TCOs have been

realized [124], establishing the feasibility of a transparent oxide bipolar transistor. With

this development, the TFT and apertures could be replaced by a transparent TFT (TTFT)

array with significant gain in the emitted light. This should significantly lower both the

voltage and drive current required by the PLEDs. The sunlight readability and extended

device lifetime can also be realized.

Transparent Semiconducting Oxides for Thin-Film Transistors

A distinction should be made between transparent semiconducting oxides (TSOs)

and TCOs based on the different carrier densities for channel and interconnect

performance. Table 4-1 lists the candidate materials, along with the range of electronic

and photonic properties. For the channel layer, significant modulation of the channel

conductance is needed to achieve field effect transistor (FET) switching. Thus, only

moderate carrier densities (semiconducting behavior) are needed for the channel layer.

On the other hand, TCOs' with high conductivity are needed for interconnects. Both of

these materials are required to be highly transparent in the visible range of

electromagnetic spectrum. There are two problems for TCOs or TSOs materials with

marginal bandgaps for visible light transparency. First, the polycrystalline nature of the

functional device introduces defects that extend the optical absorption into the bandgap.

This reduces the efficiency of the devices, particular at the blue end of the spectrum

where luminescence is weakest. Second, the optical absorption by the channel region can









lead to band-to-band excitations of carriers and subsequent shifting of transistor

characteristics.

Table 4-1. Properties of transparent semiconducting oxides.
Material Band gap Resistivity Carrier Mobility Carrier
(mQ-cm) density(cm3) (cm2V- sec1) type
ZnO 3.35 eV 0.5-100 1017- 1021 200 n
ITO 3.6 eV 0.3-20 101-1021 10-40 n
Ag(Ino.95Sno.o5)02 4.4 eV 10-106 1019 0.47 n
ZnGaO4 4.4 eV 30-105 .. ... n
CuGaO2 3.6 eV 105 1018 0.46 p
(Co,Ni)Ox 3.8 eV 1-106 1015-1020 0.5 p

Different channel materials can be considered for transparent TFT development.

Although most device applications of interest focus on inexpensive glass as the substrate

of choice, experiments have also included epitaxial (on single crystals) channel materials

in order to delineate the effects of crystallinity and grain boundaries on device

performance. A key issue is to identify transparent materials that are suitable for use as

channel materials in thin-film FET structures. Relevant factors include carrier density,

carrier mobility, crystallinity, surface morphology, optical absorption, and

photoconductivity over the visible spectrum. The latter is relevant to the stability of the

TTFTs when coupled to PLED emitters. For TFT operation characteristics, enhancement-

mode or "normally-off" is preferable to depletion-mode or "normally-on" due to lower

power supply and simpler circuit design. Large off-current and the normally-on

characteristics may originate from the fact that channel materials display high carrier

densities resulting in a channel that will be conducting in the absence of applied field.

Thus, the primary focus for TSOs has been on materials with higher electron mobility

and easily controlled carrier densities.









ZnO-based Transparent Thin-film Transistors

ZnO has many characteristics that make it attractive to be used as an active channel

layer in transparent TFTs. Due to its wide bandgap at room temperature ZnO is highly

transparent in the visible spectrum. Thus, ZnO-based TFTs can realize increased aperture

ratio of active matrix arrays and overcome the light sensitivity and degradation issues

encountered with Si-based TFTs. Another particular interest in ZnO exists in the fact that

good quality polycrystalline ZnO films with mobility ranging from 10 to 50 cm2/Vs can

be realized at low temperatures (<500C) on amorphous glass substrates or

plastic/flexible substrates. The growth of ZnO thin films has been demonstrated using a

number of deposition techniques, including sputtering, pulsed-laser deposition (PLD) and

molecular beam epitaxy. In addition, ZnO can be processed by wet chemical etching

making the device fabrication processing relative simple and low cost.

Transparent TFTs based on poly- and single crystalline ZnO films as active channel

layers have been reported by several groups recently [127-131]. Hoffman et al. [127]

fabricated an n-channel, enhancement-mode transparent TFT in which ZnO served as the

channel layer with aluminum-titanium oxide (ATO) as the gate oxide, and ITO was used

as the source, drain and gate. Rapid thermal anneal (RTA) process was used to increase

the ZnO channel crystallinity and resistivity. The transfer characteristics of TTFTs

indicated a maximum drain current on-to-off ratio of 107 and the effective channel

mobility ranged from 0.3 to 2.5 cm2/Vs. Masuda et al. [128] reported ZnO-based TFTs

with a double layer gate oxide consisting of Si02 and SiNx. The ZnO-TFT fabricated on

Si substrate showed enhancement-mode operation. The drain current on-to-off ratio was

more than 105. The transparent ZnO-TFTs fabricated on glass substrate showed









depletion-mode characteristics. It is considered that the ZnO in the transparent device had

rougher surface and more background electrons than the ZnO in the TFTs on Si

substrates. The field-effect mobility uiF of the TFT device was 0.031 cm2/Vs which is

much smaller than the Hall mobility of ZnO film. However, inserting a high carrier

concentration layer between the channel layer and the source/drain contacts can

increase/ E, to 0.97 cm2/Vs. This indicates that the high carrier concentration layer

reduced the contact resistances and had a good effect on the theoretical drain current

[128].

In plastic/flexible electronics, the processing temperatures are limited to be less

than 100"C. Although organic semiconductors can be the material choices, their low

mobility (<0.1 cm2/Vs ) and instability in ambient conditions impede the further

applications for TFTs. TFTs employing ZnO as a channel layer have been realized at

room temperature with higher mobility and current on/off ratio. Carcia et al. [129]

fabricated ZnO TFTs by rf magnetron sputtering on Si substrates near room temperature.

The best devices had /uFE of more than 2 cm2/Vs and an on/off ration > 106. Fortunato et

al. [130] also reported high performance ZnO TFTs fabricated at room temperature. The

devices operated in the enhancement mode with a saturation mobility of 27 cm2/Vs and

an on/off ratio of 3 x 105. The ZnO films with very high resistivity of about 108 0 cm

were deposited by using rf sputtering for the TFT devices. Highly conductively Ga-doped

ZnO was used as the source and drain electrodes.

In this dissertation work, the development of ZnO-based TFTs with transparent

conducting oxides as the source/drain is presented. Specifically, the fabrication process of

these top-gate type TFT devices on glass substrate is described. Three types of active









channel materials including undoped ZnO, P-doped ZnO and (Zn,Mg)O films were

employed in the TFTs. The primary interest is TFT devices that operate in enhancement

mode exhibiting a normally off (gate voltage Vg = 0) channel state. The output and

transfer characteristics of the TFT devices using different channel materials are also

discussed.

Deposition and Properties of Channel Materials

Pulsed laser deposition was used to deposit the ZnO films as the channel layer in

TFTs. A KrF excimer laser ( X =248 nm) was used as the ablation source. A laser

repetition rate of 1 Hz was applied, with a laser pulse energy density of 1-3 J/cm2. The

dependence of film properties on deposition conditions for undoped ZnO was

investigated for the films grown at 400 C in an oxygen pressure ranging from 2mTorr to

300mTorr. Corning 7059 glass substrate (1cm x 1cm) was used as the film growth

substrate. The thickness of the films is in the range of 400-500 nm. 2 at.% P-doped ZnO

and (Zno.9Mgo.i)O thin films were also utilized as the channel materials for TFTs. Post-

growth annealing process was carried out in chamber at 600 C and 100 Torr oxygen

pressure for 1 hr for all the films. Previous results showed that this annealing process for

P-doped ZnO and (Zn,Mg)O films can further reduce the native electron density by

introducing an acceptor level from phosphorus substitution [91].

The electrical properties (carrier concentration, Hall mobility and resistivity) of the

undoped polycrystalline ZnO films as a function of oxygen pressure are shown in Figure

4-2 [131]. Note that all of the films are deposited on bare glass and post-annealed at

600C in 100 Torr oxygen. For films deposited at P(02) = 20 mTorr, a Hall mobility of

26 cm2 V- s- was realized. With increasing oxygen pressure, the n-type carrier









concentration steadily decreases. This behavior reflects suppressed oxygen vacancies

and/or Zn interstitials which contribute to the intrinsic electrons in ZnO as growth

pressure increases. It is necessary to decrease the background electron density of the ZnO

films as the active channel layer and also desirable to maximize the mobility of the

channel layer. Therefore, 400C and 20 mTorr oxygen pressure were chosen to grow ZnO

films for TFTs.

Crystallinity is an important factor is determining the transport properties of the

channel material. One interesting aspect of ZnO is its strong tendency to maintain

uniaxial texture in polycrystalline films deposited on almost any substrate. Figure 4-3

shows an XRD pattern of an undoped ZnO film grown on a glass substrate at

Po2=20mTorr. The (002) and (004) ZnO diffraction peaks are predominant indicting c-

axis orientation in this polycrystalline film. However, this uniaxial texture does not

eliminate grain boundaries. Figure 4-4 shows an AFM image of the ZnO film growth

under the conditions described above. Grain size is on the order of 100-150 nm. In a

field-gated structure, the effect of grain boundary conductance and charge density has to

be considered in order to model the field-effect characteristics.

Fabrication of ZnO-based TFTs

Top-gate type TFTs using ZnO as the active channel layer were realized using

photolithography followed by wet chemical etching processing. A schematic cross

section view of a top-gate type TFT structure is shown in Figure 4-5. Amorphous gate

oxides (Ce,Tb)MgAllOi9 (CTMA) and HfO2 were grown by PLD and sputtering,

respectively. Tin-doped indium oxide (ITO) served as the source and drain electrodes due

to its low resistivity (-2 x 10-4 0 cm) and high transparency in the visible spectrum.






40



30
10 2 -o- Carrier Density (cm3)
S 2 25
,119. -*- Mobility (cm/vs)


S1018

1017. \ o


5
1015
0
1014 I I I
50 100 150 200 250 300
(a)
Oxygen pressure(mTorr)





101
AA





101





50 100 150 200 250 300
(b)
Oxygen Pressure(mTorr)


Figure 4-2. Electrical properties ofundoped ZnO thin films grown on glass at 400C as a
function of oxygen pressure. Plots showing (a) carrier density, mobility, and
(b) resistivity of the films.











1x10-

1x105-
lxloi-
1x105-
1x105-

9x104-
8x104-

7x104-

6x104-

5x104-
Avln4-


(002)


Growth temp. : 400C
Oxygen pressure: 20mTorr
Film thickness: 400nm





(004)


20 30 40 50 60 70 80
20 (degrees)


Figure 4-3. X-ray diffraction pattern of the undoped ZnO film deposited at Po2=20mTorr
on glass substrate.


500 nm








250


0 250


500 nm


Figure 4-4. An AFM image of the surface of the undoped ZnO thin film deposited at
Po2=20mTorr on glass substrate.


-r/\. iw









There are several reasons why buried-channel structure is chosen: (1) the buried

channel device is expected to have higher mobility because in bulk conduction the

surface scattering can be avoided [132]; (2) the active channel layer is protected from

ambient by the gate oxide layer, thus the channel layer can be made thinner in order to

reduce the potential parallel resistance; (3) the source and drain electrodes make direct

contact with the channel material at the gate oxide/channel interface eliminating the

potential series resistance [133].

Gate


Gate Oxide
Source FDrain



Glass Substrate




Figure 4-5. Schematic cross section view of a top-gate-type TFT structure.

Standard photolithography process had been used to fabricate the ZnO-based TFTs

in this dissertation work. Positive photoresist (Shipley 1813) was chosen due to the

requirements of the available photoresist masks. Uniform application of photoresist was

accomplished by spinning the resist on the sample substrates. The final thickness of the

photoresist film depends on the spin rate and spin time. In this work, a uniform

photoresist layer with thickness of about 800-1000 nm was obtained by setting spinning

speed at 5500 RPM for 30-40 sec. Samples with different photoresist thicknesses were

also investigated in order to obtain the optimal exposure and development conditions for









the best resolution of the patterns. Before being exposed to UV light the photoresist was

soft-baked at 90"C in air for 45 min.

The resist-coated samples were then contacted with a photoresist mask and exposed

to light. Karl Suss mask aligner which has a mercury (Hg) UV lamp was used to expose

photoresist and align the drain electrodes with those of source and drain. After the

exposure, the sample was treated with a developing solution. The exposure and

development time for certain thickness of photoresist depend on the exposure UV light

intensity which may decrease during the extended usage of the UV lamp. Therefore, the

actual intensity of UV light had to be measured to determine and adjust the exposure and

development conditions.

The photoresist pattern obtained by the development was used as a mask for the

etching of the underlying layers. In the ZnO-based TFTs fabrication scheme, wet

chemical etching was utilized for the mesa isolation and the pattern formation of source,

drain and gate. Compared to dry etching, wet etching has several advantages such as high

selectivity, less damage to the underlying material and cheap system setup. During wet

etching, the slowest step, called the rate limiting step, determines the etch rate. Generally,

there are two types of rate limiting step in wet etching: diffusion-limited and reaction-

limited. In diffusion-limited etching mode, the etch depth has a square root dependence

on etch time and the solution agitation significantly affects the etch rate. The solution has

to be agitated in some manner to assist in the movement of etchant to the surface and the

removal of the etch product. On the contrary, in reaction-limited etching, the etch depth

has a linear dependence on etch time and the etch rate is independent of solution agitation.









Since it is desirable to have a reproducible and well controlled etch rate for device

fabrication, the reaction-limited etching is more preferable to the diffusion-limited mode.

In this work, wet etching of the materials employed in the TFT structure has been

investigated by using different acid solutions including H2SO4, HC1, HNO3, H3P04, HF

as well as the mixtures of these acids. To realize this top-gate TFT structure high

selectivity of gate, gate oxide and channel layers over underlying source/drain layer is

required. It was found that among these etchants H3P04 has the best resolution for the

gate metal and the highest selectivity of gate oxide and channel layers over underlying

ITO. For the etching of ITO source and drain, HC1 acid was found to have the best

etching results. Stylus profilometer was used to measure the film thickness in order to

calculate the etch rate after the removal of photoresist mask in acetone. In the same time,

the etch bias resulting from the isotropic wet etching can be observed and minimized by

optimizing the temperature of acid solution and the etching time.

Figure 4-6 schematically illustrates the fabrication processing sequence for the

ZnO-based TFT structure. The initial substrate is commercial ITO-coated display glass

substrate. First, the ITO source and drain patterns were defined via using

photolithography process, with wet etching of the ITO performed in HC1 acid at around

35 C. The dependence of etch depth on etch time showed a linear relationship, indicating

the etching process was reaction-limited. After removing the source/drain photoresist

patterns in acetone, ZnO was then deposited at 400 C in 20 mTorr oxygen as the active

channel layer on the ITO patterned glass substrate. The thickness of channel layer was in

the range of 20-50nm. A thin layer of gate oxide (100-200 nm) was deposited on top of

the ZnO channel layer. Then aluminum as the gate metal (-100nm) was deposited by









using magnetron rf sputtering at room temperature. The gate contact was defined via

using lithography alignment followed by selective wet etching gate metal (Al), gate oxide

and channel layers with H3P04 acid down to the source and drain contacts. Aluminum

deposited by sputtering can be easily etched by H3P04 at -60 C with a well-controlled

profile. Etching of gate oxide and ZnO films without removing the underlying ITO layer

as well as the glass substrate was realized in H3P04 acid at moderate temperatures (35-

40OC). Thus, etching this top-gate TFT structure by using H3P04 acid with high

selectivity made the device fabrication process much simpler and more controllable.

Top-view microscopy images of the ZnO-TFT devices on glass substrate are shown

in Figure 4-7 (a) and (b). Top-gate type TFT structures with well-defined contact patterns

were realized by using photolithography and wet etching. Devices with different channel

widths and lengths were fabricated as shown in Figure 4-7 (a). For the device shown in

Figure 4-7 (b), the channel length and width are 50 1 m and 90 i m, respectively.

In order to field gate this structure, it is necessary to form either a Schottky barrier

or gate oxide. For ZnO, Schottky barriers are low and are apparently unsuitable for field

gated rectifying structures. For oxide gates, the gate dielectric must be chosen to have a

band offset with the channel material so as to avoid carrier injection into the conduction

band and/or valence band of the gate insulator. For wide bandgap semiconductors, such

as ZnO, this necessitates the use of the larger bandgap oxides and precludes the use of

many insulators being considered for field gated structures on other semiconductors.

(Ce,Tb)MgAllOi9 (CTMA) with wide bandgap (>5 eV) was used as the gate dielectric

for the TFT. Low leakage current about 10-7 A cm-2 was obtained for a 200 1- m diameter

Al-CTMA-ZnO capacitor [131].













A


1. Source & Drain Lithography
Photoresist
ITO


/1


Glass Substrate




2. ITO Wet Etching


5. Gate lithography
Gate metal
Gate oxide
ZnO


Glass Substrate


i


,


Photoresist


Glass Substrate


3. Channel & Gate Oxide Deposition

Gate oxide
l9" A kZr0=ww ZnO
ITO ITO
Glass Substrate




4. Gate Contact Metallization
M rr Gate metal
r Gate oxide
SZnO
ITO ITO
Glass Substrate


6. Wet Etching Gate Contact/Gate oxide/Channel

Photoresist Gate metal
r Gate oxide


ITO ZnO ITO
Glass Substrate




7. ZnO-TFT

Af s Gate metal
Gate oxide


ITO ZnO ITO
Glass Substrate


Figure 4-6. Schematic fabrication sequence of ZnO-based TFT structure.


/



















































Figure 4-7. Top-view microscopy image of the ZnO-TFTs on glass substrate.

ZnO-based TFT Device Characterization

We characterized the output and transfer performance of the TFT devices using an

Agilent 4155A Semiconductor Parameter Analyzer at room temperature. The transfer








characteristics of the devices include drain current Id and gate current Ig as a function of

gate voltage Vg at a fixed drain voltage Vd, drain current on-and-off ratio (IoN/IoFF) and

field-effect mobility. When the current carriers are confined within a narrow channel

layer additional scattering mechanisms have to be considered. The location of the carriers

at the oxide-semiconductor interface introduces additional scattering mechanisms like

Coulomb scattering from oxide charges and interface states, as well as surface roughness

scattering. Generally, field-effect mobility can be obtained both from the

transconductance value and from the saturation current. In field-effect transistors,

transconductance is defined by [134]

BIDS
ngm --s VD (4-1)
GS
The field-effect mobility 1- FE is given by
Lg ,
SFE W V (4-2)
WC OX DS
Where L is the channel length, W is the channel width, Cox is the capacitance of

gate oxide, VDS is the source-drain voltage. The turn-on voltage need not be known for

the determination of LtFE. Another way to calculate field-effect mobility is fitting straight

line to the plots of the square root of drain current vs gate-drain voltage, while the drain

current in the saturated region, Idsat is given by [128]

W
cit = )iFECi (VGS Vth) (4-3)

(h, > -
Where Ci is the capacitance per unit area of the gate oxide.






49

ZnO-TFTs Using Undoped ZnO as Active Channel Layer

The electrical characteristics of the TFT device with undoped ZnO as the channel

layer are shown in Figure 4-8. For this device structure, the ZnO film thickness was

20nm. This TFT device operated as an n-channel, normally-on device, as evident from

the fact that there was a source-drain current at the gate voltage of OV and a negative

voltage was required to deplete the carriers in the channel layer. Note that the device

drain currents IDS are large, of order of mA, in its "ON" state is due to the high carrier

concentration in the undoped ZnO channel layer. To further modulate and deplete the

channel conductance, low channel carrier density is necessary to achieve.


25

20- OV

15- -5V
-10V
S10- 15 V
\_-20V
S 5-

0-

0 1 2 3 4 5 6
DainWttage(V)


Figure 4-8. Drain current as a function of drain voltage characteristics for the undoped
ZnO-TFT.









ZnO-TFTs Using P-doped ZnO and (Zn,Mg)O as Active Channel Layer

For efficient TFT operation, an enhancement-mode device is preferable over

depletion-mode, thus avoiding the need to apply voltage in order to turn the device off.

Much less power dissipation is possible when normally-off, enhancement-mode devices

are employed. For this motivation, alternative channel materials have been investigated in

order to decrease carrier density. Phosphorus-doped ZnO and (Zn,Mg)O have been

deposited as the active channel layer for the TFT. Post-deposition oxygen annealing

processes were used to further decrease the electron density in these films. Figures 4-9 (a)

and (b) show the output characteristics of devices with P-doped ZnO and (Zn,Mg)O (50

nm for both) as the channel materials, respectively. For P-doped ZnO based TFT, the

device has the same depletion-mode operation as the undoped ZnO one. However, the

channel conductivity is lower than that of TFT with undoped ZnO as the channel at the

gate voltage of OV.

Enhancement-mode operation (Figure 4-9 (b)) was realized for P-doped (Zn,Mg)O

based TFTs with HfO2 serving as the gate dielectrics. In these devices conducting

channels were induced by applying positive gate voltages. Channel length and width

were 20 1 m and 90 1 m, respectively. A saturation of IDS (pinch-off) was observed for

small values of VDS. This pinch-off behavior indicates that the channel layer is

sufficiently depleted in this TFT. The field-effect (FE) mobility was derived to be about

5.32 cm2 V-s-1 from the transfer characteristics of the devices operated at 6V shown in

Figure 4-10. This value is comparable to that realized in undoped ZnO channels

indicating that acceptor doping did not have a detrimental impact on channel mobility.

We also determined the carrier density was as low as 3.9x1016 cm-3 for this channel






51


material, which is two orders lower than that in the undoped ZnO thin film. The on/off

current ratio is on the order of 103 at the gate voltage of 10V for these devices.


0 2 4 6 8 10
DranVdltage(V)


40

(b) 4V
30-


20- 3V


10- 2V


0


ain1 2 3 4
Drain Voltage (V)


5 6 7


Figure 4-9. The output characteristics of the TFT with alternative active channel
materials: (a) P-doped ZnO as the channel; (b) P-doped (Mg,Zn)O as the
channel.













4.5x106 -

4.0x106

3.5x106

< 3.0x106 -

S2.5x1 0-6

2.0x106

- 1.5x106

0 1.0x106

5.0x10 -

0.0
-4


-3 -2 -1 0 1 2

Gate voltage (V)


1.8x10-6

1.6x10-6

1.4x10-6

1.2x 1 0-6
1.2x10-6


1.0x10-7
8.0x10'7 3

6.0x107 -^

4.0x10-7

2.0x10-7

0.0


3 4 5


Figure 4-10. Transfer characteristics of ZnO-TFT with P-doped (Zn,Mg)O as the channel
layer at the drain voltage of 6V.


I~I~I~I~I~I~I~I



V,:6V














I.I.I.I.I.I.I.I














CHAPTER 5
GROWTH AND CHARACTERIZATION OF P-TYPE PHOSPHORUS-DOPED
(Zn,Mg)O BY PULSED LASER DEPOSITION

Introduction

One of the critical issues in developing ZnO-based UV LEDs and lasers is to

realize low resistivity, high carrier density p-type ZnO material. Phosphorus is a possible

acceptor dopant that can be used to synthesize p-type ZnO. Doping of ZnO with Mg

provides a means to increase the band gap further into the UV. The motivation for

examining phosphorus doping in Mg-doped ZnO is two-fold. First, p-type conductivity in

(Zn,Mg)O films will be necessary for LED heterostructures in which carrier confinement

for efficient electron-hole recombination is needed. Second, the addition of Mg shifts the

conduction band edge to higher energy, perhaps increasing the activation energy of the

defect donor states. Previous results on annealed phosphorus-doped (Zn,Mg)O device

structures, in particular C-V and I-V characteristics, indicate that phosphorus yields an

acceptor state and p-type behavior [135]. However, these materials did not yield an

unambiguous positive Hall voltage, presumably due to the low mobility and high carrier

compensation.

In this chapter, an unambiguous positive Hall coefficient in as-grown P-doped

(Zn,Mg)O films is presented. The effect of oxygen partial pressure on the transport

properties of P-doped (Zn,Mg)O films grown on low-temperature (LT) undoped ZnO

buffer layers is described. The chemical state of phosphorus in p-type (Zn,Mg)O:P films









is presented. The crystallinity and the surface morphology of these films are also

discussed.

Experimental

The phosphorus-doped (Zno.9Mgo.i)O epitaxial films were grown via pulsed laser

deposition (PLD) on c-plane sapphire substrates. The target was fabricated using high-

purity ZnO (99.9995%) and MgO (99.998%), mixing with P205 (99.998%) as the doping

agent. The melting point and boiling point for P205 are about 340C and 360OC,

respectively. P205 has a lower heat of formation (Hf = -360 Kcal/mole) compared with

that of SiO2 (Hf = -202.6 Kcal/mole) [136]. The phosphorus doping level in the target

was 2 at.%. A KrF excimer laser ( X =248 nm) was used as the ablation source. A laser

repetition rate of 1 Hz was applied, with a laser pulse energy density of 1-3 J/cm2. The

ZnO growth chamber has a base pressure of 10-6 Torr. An undoped ZnO buffer layer

(-50nm) was initially deposited at 400 C and 20mTorr oxygen partial pressure before the

growth of P-doped (Zno.9Mgo.i)O films. The undoped ZnO buffer layer was post-annealed

at 650C in flowing 02 for lh in order to decrease the electron conductivity. Semi-

insulating buffer layers are preferable in order to perform Hall measurements without

influence from buffer layer conduction. The phosphorus-doped (Zno.9Mgo.i)O films were

then deposited on the annealed undoped ZnO buffer layer at a substrate temperature of

500"C under oxygen partial pressure ranging from 20 mTorr to 200 mTorr. The total film

thickness ranged from 500 to 700 nm.

Four-point Van der Pauw Hall measurements were performed at room temperature

in order to examine the transport properties of the as-grown P-doped (Zno.9Mgo.i)O films.

The chemical state of phosphorus in the films was examined by using XPS with an Al









anode (photon energy = 1486.6 eV). The film crystallinity and surface morphology of the

P-doped (Zno.9Mgo.1)O films as a function of oxygen growth pressure were also

investigated by using XRD and AFM, respectively.

Results and Discussion

The large lattice and thermal mismatch between ZnO and sapphire substrate will

generate considerable stress in the epitaxial film, affecting the growth and quality of the

film. To optimize the crystallinity and properties of the epitaxial films, a well-created

buffer layer is necessary. Kaidashev et al. [137] reported that by inserting a thin ZnO

relaxation layer grown at lower temperature between sapphire substrate and high-

temperature ZnO film the crystallinity and mobility of the films have been improved. In

this work, a thin layer of LT-ZnO was grown at 400 C on sapphire substrate before the

growth of P-doped ZnMgO film. The effect of the LT-ZnO buffer layer on crystallinity of

P-doped (Zno.9Mgo.i)O films were investigated by measuring the omega rocking curve for

ZnO (0002) plane. Figure 5-1 shows the ZnO (0002) omega rocking curves of the P-

doped (Zno.9Mgo.i)O sample with and without a LT-ZnO buffer layer. These films were

grown at 500 C and under 20 mTorr oxygen pressure. An increase in diffraction intensity

is clearly shown for the film grown with LT-ZnO buffer layer. The full-widths at half

maximum (FWHM) values of the omega rocking curve are 0.830 and 1.020 for the films

grown with and without buffer layer, respectively, indicating that the crystallinity was

improved by introducing a thin LT-ZnO buffer layer. Thus, by applying a thin layer of

LT-ZnO buffer layer, the properties of the P-doped ZnMgO films were expected to be

optimized.









For many samples, a persistent n-type photoconductivity was observed that

complicated the transport characterization. To eliminate the photoconductivity relaxation

effect on the transport properties, the samples were maintained in the dark for 12 hours

before performing Hall measurements at room temperature. Using this procedure, the

carrier density and conduction type of as-grown P-doped (Zno.9Mgo.i)O films was

determined. These properties as a function of oxygen growth pressure are shown in

Figure 5-2. Each sample was measured a minimum of twenty times to obtain the average

results shown here. The error bars represent the maximum deviation from the average

values. Note that films deposited at the oxygen partial pressure lower than 100 mTorr

show n-type conductivity with electron concentration in the range of 1016-1017 cm-3

However, as oxygen pressure increases above 100 mTorr, the electron concentration

continuously decreases and the samples started showing indeterminate carrier type

indicating the coexisting of electrons and holes in the films.

Unambiguous conduction type was not observed for the samples grown at 100 and

120 mTorr due to near-equivalent concentrations of holes and electrons in the films.

When the oxygen partial pressure was increased to 150 mTorr during the film deposition,

the Hall-effect data showed consistent p-type carrier type with a hole concentration of 2.7

X 1016 cm-3. Also note that as the oxygen partial pressure was increased up to 200

mTorr, the films reverted to an indeterminate carrier type. Figure 5-3 shows the carrier

mobility carrier type films grown under 100, 120 and 200 mTorr oxygen pressure yield

large of P- doped (Zno.9Mgo.1)O films as a function of oxygen pressure. The

indeterminate standard deviations compared with those with unipolar conduction type

films. These results imply that for the samples which may contain both types of carriers,







57


more attention needs to be paid for the Hall measurements and data analysis. The p-type

films grown at 500C and 150 mTorr oxygen show an average hole mobility of 8.2


cm2/Vs at room temperature. This mobility value is reasonable compared with the ones

reported in the previous studies on N, Al-N and P-doped p-type ZnO.



1.4x10'
no buffer layer
1.2xl0 (FWHM=1.02)
--with buffer layer
(FWHM=0.83)
1.0x105

S&8.Ox104

m 6.0x104




CD I I
4.0x104

2.0x104


14 16 18 20 22
S(degree)




Figure 5-1. ZnO (0002) omega rocking curves of P-doped (Zno.9Mgo.1)O samples with
and without LT-ZnO buffer layer.

The resistivity of the P-doped (Zno.9Mgo.i)O films grown under different oxygen

pressures are also shown in Figure 5-4. The resistivity of the films increases from 1.0395

0 -cm to 350.5 0 -cm as oxygen growth pressure increases from 20 to 100 mTorr. The


increase in resistivity for the films grown under oxygen pressure between 20 to 100

mTorr results from the decreased carrier concentration in the films. The films grown at

100 mTorr exhibit the highest resistivity of about 300 0 -cm. With further increasing


oxygen pressure, resistivity rapidly decreases due to the increased hole carrier conduction







58


in the films. For the p-type P-doped (Zno.9Mgo.i)O films grown at 150 mTorr oxygen

pressure, the resistivity is about 35 0 -cm.


1E18-



1E17-



1E16-


20 40 60 80 100 120 140 160 180 200
Oxygen Partial Pressure (mTorr)


Figure 5-2. Carrier concentration and carrier type in P-doped (Zno.9Mgo.1)O films as a
function of oxygen partial pressure.


20 40 60 80 100 120 140 160 180 200
Oxygen Partial Pressure (mTorr)


Figure 5-3. Effect of oxygen partial pressure on carrier mobility of P-doped
(Zno.9Mgo.i)O films.


* p-type
n-type
* indeterminate
carrier type


p-type
I n-type







59




1000
p-type
n-type
U indeterminate
carrier type
E 100o-

o *
0
> :


1-
U,






20 40 60 80 100 120 140 160 180 200
Oxygen Partial Pressure (mTorr)



Figure 5-4. Resistivity ofP-doped (Zno.9Mgo.i)O films vs oxygen partial pressure.

The Hall effect results show that p-type doping in P-doped (Zno.9Mgo.i)O films is

strongly dependent on the oxidation conditions. It is important to note that the effect of

oxygen partial pressure on ZnO p-type conduction has been investigated previously by

both experiment and theory. Xiong et al. [138] reported evidence for p-type conduction

in undoped ZnO films grown at high oxygen partial pressure by reactive sputtering. A

change in the sign of charge carriers from electrons to holes was identified around 55%

oxygen in Ar/02 mixture. ZnO-basedp-n homojunction was also formed by controlling

oxygen partial pressure during sputtering. This realization of p-type conductivity in

undoped ZnO was consistent with the effect of higher chemical potential of atomic

oxygen on defect formation enthalpies. The increased oxygen chemical potential by

electronic excitation to a dissociated state raises the formation enthalpy of the intrinsic

donor Vo and lowers the formation energy of the acceptor Oi..









Zunger also suggested several theoretical practical rules for p-type doping of wide

bandgap materials to overcome the doping bottlenecks [76]. First, the p-type doping is

facilitated by alloying an element that leads to upward bowing of the VBM. One way to

shift the VBM upwards is to add a tetrahedrally bonded 3d element with active d states.

In addition, Zunger suggests that limitations to p-type doping can be overcome by

manipulating the growth conditions, e.g. the use of the host anion-rich growth conditions

to inhibit the formation of so-called "hole killer" defects. Calculations from chemical

potentials suggest that the enthalpy of forming anion vacancies decreases under cation-

rich (zinc-rich) conditions. In the present study, the growth condition necessary to obtain

p-type P-doped (Zno.9Mgo.i)O films is rather narrow (150 mTorr oxygen partial pressure).

For the samples grown in the oxygen pressures of 100, 120, and 200 mTorr, the Hall-

effect results do not show unambiguous p-type conductivity. The observation of

indeterminate carrier type for a growth pressure of 200 mTorr may be explained by the

"host anion poor" rule for the anion-substituting p-type dopants which conjectures that

anion poor conditions are more favorable for a high solubility of acceptors on anion sites

[76]. Consequently, the host anion condition has to be optimized in order to reach an

equilibrium state under which p-type doping of ZnO can be realized.

In order to confirm the incorporation of P205 as a P doping source in the p-type

films, X-ray photoelectron spectroscopy was performed to examine the chemical states of

phosphorus using Al anode as the X-ray source. Figure 5-5 shows the XPS survey of P-

doped ZnMgo.iO films grown at 500"C, 150 mTorr oxygen pressure. Zn 2p, Mg Is and O

Is peaks as well as their Auger peaks are shown in the spectrum. The multiplex ofP 2s is

shown is Figure 5-6. Only one peak with the binding energy of 192.2 eV is observed









from the spectrum, which is consistent with P 2s binding energy of 192.8 eV in P205. The

P 2s peaks regarding to Zn3P2 and element phosphorus state have the binding energy

values of 186.3 eV and 187.7 eV, respectively [139].

X-ray diffraction was used to examine the crystallinity of the P-doped

(Zno.9Mgo.1)O films grown under different oxygen partial pressures as shown in Figure 5-

7. The diffraction data shows only ZnO (0002) and (0004) and sapphire (0006) peaks

indicating that the films are oriented only with c-axis perpendicular to sapphire substrates.

Thus, the out-of-plane orientation is ZnO [0001] I sapphire [0001]. No impurity phases

were observed from the XRD results. This suggests that the solid solubility of phosphorus

and magnesium has not been exceeded in the films under these growth conditions. Also

note that there is no discernable change in the c-axis lattice parameter with increasing

oxygen partial pressure. As discussed earlier in this chapter, the crystallinity of the P-

doped (Zno.9Mgo.i)O films grown on LT-ZnO buffer layer was also characterized by four-

circle XRD. The FWHM of the omega rocking curve for ZnO (0002) peak is 0.83.

The p scan of P-doped ZnMgO film grown on sapphire substrate was used to

examine the epitaxy of the film and determine the in-plane orientation relationship

between the film and substrate. Figure 5-8 shows the p scans through ZnO (10 1 1) plane

and sapphire (11 26) plane of the sample grown at 500 C, 20 mTorr oxygen pressure. A

sixfold symmetry of the plane is shown indicating good epitaxy of the P-doped ZnMgO

film. More interestingly, it is shown that P-doped ZnMgO film perfectly aligns with

sapphire substrate. No 300 twisted orientation was observed in these films grown at

500"C. Therefore, the in-plane orientation relationship between the film and sapphire

substrate is determined as ZnO [10 10] I sapphire [10 10].













Mm 2780 Max 900767


Mg 1s





Z2s -




_____, ____


Zn 2p


o 0LL
0 LMM
-I I I---------- --- ---
I I I


MSL ZIn 3p
M G L Z n I


1350 1215 1080 945 810 675 540
Binding Energy (eV)


405 270 135 0


Figure 5-5 X-ray photoelectron spectroscopy survey ofP-doped (Zno.9Mgo.i)O films.

Mm 1100 Max 1251
[ [ I I pP 2s ...
--- --- --- --- --- --- --- --- --


Binding Energy (eV)


Figure 5-6 X-ray photoelectron spectroscopy multiplex of P 2s peak for P-doped

(Zno.9Mgo.i)O films grown at 500"C, 150 mTorr oxygen pressure.












1.2x106-

1.0x10 -
1.0x106-
8.OxlO-

6.0x10-

4.0x10-

2.0x10'-

0.0.


ZnO (0002)


Pq=100mTorr



Po=20mTorr

30 35 40 45 50 55 60 65 70 75
2 thelta (degree)


Figure 5-7. X-ray diffraction of P-doped (Zno.9Mgo.1)O films grown under different
oxygen partial pressures.


-- sapphire substrate
-- P-doped ZnMgO film


1000

C





100


-200 -150 -100 -50 0
Phi (degree)


50 100 150


Figure 5-8. XRD p scans of P-doped (Zno.9Mgo.1)O film and sapphire substrate.


.11..Ij










The surface morphology of the films grown at 500"C in various oxygen pressures is

shown in Figure 5-8. The scan area is 2 X 2 1- m2 with a scan rate of 1 Hz. As the oxygen

pressure increased, both grain size and surface roughness increased. Figure 5-9 shows the

Root-Mean-Square (RMS) roughness of the P-doped (Zno.9Mgo.i)O films as a function of

oxygen partial pressure. The RMS roughness for the films increased from 2.60 nm to

12.8 nm as the oxygen growth pressure increased from 20 mTorr to 200 mTorr. For the p-

type (Zno.9Mgo.i)O films deposited at 150 mTorr oxygen pressure, the presence of grain

boundaries can contribute to the low carrier mobility.


15nm 15n






0.5 0.5
1 1(a(b)
1.5 1.5
pm pm

15nm 15nm





0.5
0.5 1 () (d)

1.5 p I
pm

Figure 5-9. AFM images of the P-doped (Zno.9Mgo.1)O films grown at different oxygen
pressures: (a) 20; (b) 100; (c) 150; (d) 200mTorr.












16

14

12-
E
10
Cn
8
0
j) 6 -

4-

2-

0 20 40 60 80 100 120 140 160 180 200 220
Oxygen partial pressure (mTorr)



Figure 5-10. RMS roughness of the P-doped (Zno.9Mgo.i)O films as a function of oxygen
partial pressure.

In summary, p-type phosphorus-doped (Zno.9Mgo.i)O films have been realized via

pulsed laser deposition without post-annealing process. The conduction type of the films

strongly depends on the oxygen partial pressure during the deposition process. The films

grown at oxygen pressure lower than 100 mTorr are n-type. However, at oxygen pressure

of 150 mTorr, the films showed p-type carrier type conduction with a hole concentration

of 2.7X 1016 cm-3, a mobility of 8.2 cm2/Vs and a resistivity of 35 0 -cm. XPS


measurements confirmed the existence of P205 in the p-type P-doped ZnMgO film. XRD

results showed good crystallinity of P-doped ZnMgO films grown under different oxygen

pressures. The in-plane and out-of-plane orientation relationships are determined as ZnO


[10 10] I sapphire [10 1 0] and ZnO [0001] I sapphire [0001]. The RMS roughness for


the films increased from 2.60 nm to 12.8 nm as the oxygen growth pressure increased






66


from 20 mTorr to 200 mTorr. The presence of grain boundaries can contribute to the low

carrier mobility of p-type P-doped ZnMgO films.














CHAPTER 6
SYNTHESIS AND CHARACTERIZATION OF (Zn,Mg)O:P/ZnO
HETEROSTRUCTURES AND AL-DOPED ZnO

Introduction

Several studies have been reported regarding ZnO-based p-n junctions for LED

applications. Alivov et al.[140,141] reported LEDs from n-ZnO/p-AlGaN and n-ZnO/p-

GaN heterostructures grown expitaxially on SiC substrates using hybrid vapor-phrase

epitaxy combined with chemical vapor deposition. The UV LEDs emitted UV light at

389 nm and 430 nm at room temperature, respectively. Osinsky et al. [142] also have

reported electroluminescence (EL) from p-(Al)GaN/n-ZnO junctions. Tsukazaki et al.

[143] obtained violet EL from ZnO homojunction grown on lattice matched ScAlMgO4

substrates. Hwang et al. also reported on the diode and emission characteristics for a

heterostructure of p-ZnO/n-GaN fabricated via RF magnetic sputtering [144]. We have

previously reported ZnO-based p-n junctions deposited on undoped ZnO substrates using

ZnMgO:P/ZnO heterostructure system [145]. The use of a ZnO buffer on the lightly n-

type ZnO substrate was critical in achieving acceptable rectification in the junctions.

Without this buffer, the junctions showed high leakage current. In this prior work, p-type

conductivity was only obtained by post-growth annealing of the P-doped ZnMgO.

Studies in the previous chapters show that oxygen partial pressure plays a

significant role in converting n-type to p-type conductivity for 2 at.% P-doped

Zno.9Mgo.10 films. The P-doped (Zn,Mg)O films grown at 150 mTorr oxygen partial

pressure exhibited p-type conductance without post-growth annealing. In this chapter, the









development of ZnMgO:P/ZnO heterostructures on both sapphire and single crystal ZnO

substrates is described. No post-growth annealing process was carried out for these

structures. The characteristics of n- and p-side ohmic contact and the ZnMgO:P/ZnO

heterojunction are presented. In addition, to reduce the presence of series resistance such

as the current spread resistance in the n-type ZnO layer, 0.01 at. % Al-doped ZnO films

were grown on sapphire substrates with MgO buffer layer. Three different growth

conditions including growth temperature, oxygen partial pressure and laser energy were

examined to provide a systematic study of growth condition effect on the properties of

the Al-doped ZnO films. The dependence of crystallinity, electrical properties,

photoluminescence and surface morphology on these growth conditions is discussed.

Experimental

Schematic diagrams of the p-ZnMgO/n-ZnO heterostructure on both (0001) c-

sapphire and ZnO substrate are shown in Figure 6-1. The (0001) undoped grade I quality,

single crystal ZnO substrate is obtained from Cermet. The room temperature electron

concentration and mobility were 1017 cm-3 and 190 cm2/Vs, respectively. Pulsed laser

deposition was used for film growth. The 2 at.% phosphorus-doped Zno.9Mgo.10 target

was fabricated using high-purity ZnO (99.995%) with or without MgO (99.998%),

mixing with P205 (99.998%) as the doping agent. Sapphire and ZnO substrates were

ultrasonically cleaned with trichloroethylene (TCE), acetone and methanol for 5 min and

dried in N2 prior to loading into the growth chamber. The growth chamber base pressure

was 1-2 x 10-7 Torr. A KrF excimer laser with a wavelength of 248nm was used as the

ablation source. A laser repetition rate of 1Hz was used, with a target to substrate

distance of 4cm and a laser pulse energy density of 1-3 J/cm2. The n-ZnO layer 0.6 tam

thick with an electron concentration of 2.53x1018 cm-3 and mobility 36.55 cm2/Vs was









grown first at 800"C in an oxygen pressure of 100 mTorr, followed by a 0.4[pm thick p-

ZnMgO:P layer grown at 500"C, in 150 mTorr 02.

Electron-bean evaporated Au (100nm) and Ti/Au (20nm/80nm) were deposited on

the p-ZnMgO layer and n-ZnO patterned by lift-off process. In order to improve the

ohmic characteristics, the post-growth annealing at 500"C and 450"C in N2 for 2 min were

performed, respectively. The I-V characteristics were measured using an Agilent 4145B

parameter analyzer at room temperature.

A MgO buffer layer (-200nm) was initially deposited at 450C and 10-4 mTorr

oxygen pressure before the growth of 0.01 at.% Al-doped ZnO films on sapphire

substrates. The 0.01 at. % Al-doped ZnO target was fabricated using high-purity ZnO

(99.995%) mixing with A1203 (99.998%) as the doping agent. The purpose of the MgO

buffer layer is to reduce the micro-cracks resulting from the thermal expansion difference

between ZnMgO and sapphire substrate. Table 6-1 shows the different growth conditions

of the Al-doped ZnO via PLD. The film thickness of Al-doped ZnO is in the range of

0.75-1 1i m. The transport properties of the as-grown films were determined using four-

point Van der Pauw Hall measurements at room temperature. The photoluminescence

properties of the films were also measured at room temperature using a He-Cd laser

(325nm). The film crystallinity and surface morphology were investigated via using four-

circle X-ray Diffraction (XRD) and atomic force microscopy (AFM).

Results and Discussion

The current-voltage characteristics of the metal contacts to n-ZnO andp-ZnMgO

were measured to examine the formation of the ohmic contacts. Previous results show

that Au and Ti/Au can be used as ohmic contacts to p-ZnMgO and n-ZnO, respectively.









Low specific contact resistance was obtained for Au contact to p-ZnMgO after post-

growth annealing process [146]. Figure 6-2 (a) and (b) show the I-V curves of the Au and

Ti/Au metal contacts between two square pads (500 x 500 ym) at room temperature. The

I-V characteristics indicate that ohmic contacts are formed on both electrodes. These

results show that the rectifying behaviors shown in Figure 6-3 are due to the

heterojunction of the ZnMgO/ZnO structure and not to the semiconductor/metal contacts.

Au (100nm)

Ti/Au(20/80nm) p-ZnMgO:P.02
(400nm)

n-ZnO(600nm)


ZnO or Sapphire substrate



Figure 6-1. Schematic illustration of ZnMgO:P/ZnO heterostructure.

Table 6-1. Growth conditions of 0.01 at. % Al-doped ZnO films via PLD.
Same Tg () Oxygen pressure Laser energy B
Sample Tg (C) ( Buffer layer
(mTorr) (mJ)
A 700 20 350
B 800 20 350
C 700 20 300
D 800 20 300MgO
E 800 5 300
F 800 50 300

The I-V characteristics of the ZnMgO:P/ZnO heterostructure fabricated on sapphire

and ZnO substrates are shown in Figure 6-3 (a) and (b), respectively. The devices exhibit

clear rectifying electrical characteristics for both structures. Note that for the device

grown on sapphire substrate, the drain current is higher than that of the device grown on

ZnO substrate. This result might result from the difference in the resistivity of the









epitaxial layers grown on different substrates. The turn-on voltage VT can be obtained

from the intercept of the linear fitting in the forward bias range with the voltage-axis. For

the structure grown on sapphire substrate VT was determined to be about 1.36 V. For the

ZnMgO:P/ZnO heterostructure grown on single crystal ZnO substrate, the turn-on

voltages are 1.15 V and 2.26 V for the lateral and vertical device structure, respectively.

Similar results on the I-V characteristics of oxide-based p-n junctions have been reported

previously [147-149]. According to the current-voltage characteristics of the real diodes

equations [150],


I = Iexp -1 (6-1)

q dV
n = (6-2)
kT dlnI
where the pre-exponential factor Io is the reverse saturation current, V is the voltage at the

junction, Vi = kT/q is the thermal voltage, k is the Boltzmann constant, T is the absolute

temperature, and n is the junction ideality factor. From equation (6-2), the ideality factors

of the ZnMgO/ZnO heterostructures on sapphire and ZnO substrate can be extracted to be

about 7.6 and 11.8, respectively. These large ideality factors possibly result from the

defect-level assisted tunneling [151] and carrier recombination in the space-charge region

via a deep level near midgap in the ZnMgO. Further work should focus on increasing the

hole carrier concentration and mobility ofp-ZnMgO and the optimization of structure

synthesis in order to improve ZnMgO/ZnO p-n junction characteristics.

Four-circle X-ray diffraction was used to examine the crystallinity of the Al-doped

ZnO samples (1cm x 1cm) grown on sapphire substrates. Figure 6-4 shows the XRD 2

theta scans of the as-grown 0.01at.%Al-doped ZnO films grown under different

conditions, suggesting that the all the ZnO:Al films are oriented with the (0001) c-axis








72



uniformly parallel to the surface normal. Little change in diffraction intensity and lattice


spacing was observed for the films grown under different conditions. To further delineate


1.0x10 3-

8.0x10 4-

6.0x10 4-

4.0x1 04-

S2.0x10 4-

S 0.0

-4
S-2.0x10 4-

-4.0xl 04-

-6.0x1 04-

-8.0x1 04-

-1.0x10-3-








8.0x105-

6.0x105-

4.0x105-

2.0x105-

0.0-
-
S-2.0x105-

-4.0x105-

-6.0x105-

-8.0x10l-

-1.0x104-


-1.0 -0.5 0.0 0.5 1.0
Bias (V)


-1.0 -0.5 0.0 0.5 1.0
Bias (V)


Figure 6-2. The I-V curve of Au and Ti/Au metal contacts on: (a) p-ZnMgO; (b) n-ZnO
films, respectively.


Ti/Au (20nm/80nm) to n-ZnO
After RTA annealing at 450C







73




0.045
0040 p-ZnMgO/n-ZnO/sapphire

0.035 -

0.030 -

S0.025 -

S0.020 -
o- U
0.015 -

0.010

0.005

0.000

-0.005
-2 -1 0 1 2 3
Bias (V)

6.0x104

p-ZnMgO/n-ZnO/ bulk ZnO
5.0x104


4.0x10U4
Lateral =

S3.X10'U4

0 2.0X10-4 ml
O 2.0xlO4


1.0x1 ,U 4

oto Vertical
0.0

-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6
Bias (V) (b)




Figure 6-3: Current-voltage characteristics of the ZnMgO/ZnO heterostructure on: (a)
sapphire; (b) ZnO substrate.









the crystallinity of the films grown under different conditions, the omega rocking curve

through the (0002) plane of ZnO was investigated. Figure 6-5 shows the omega rocking

curve of ZnO:Al films grown under 300 mJ laser energy. The full-width-at-half-

maximum (FWHM) values of the Al-doped ZnO films grown under different conditions

are listed in Table 6-2. It is shown that for the films grown under 350 mJ laser energy,

with increasing growth temperature the FWHM decreases from 0.370 to 0.330 indicating

an improved crystallinity of the films. The film grown at 800"C, 50 mTorr oxygen

pressure has the narrowest omega rocking curve with the FWHM of about 0.260. The

correlations of crystallinity to the electrical and optical properties of the films will be

discussed later.

Table 6-2. FWHM values of ZnO (0002) omega rocking curve for 0.01 at. % Al-doped
ZnO.
Sample A B C D E F
FWHM
reM 0.3703 0.3308 0.2791 0.6366 0.3889 0.2636
(degree)

Electrical properties of the ZnO:Al films was investigated via using four-point Van

der Pauw Hall measurement at room temperature. Table 6-3 shows the Hall measurement

results of the as-grown Al-doped ZnO films. The growth conditions of sample A-F are

described in Table 6-1 previously. All the films show n-type conductivity with resistivity

in the range of 10-310-1 ohm-cm. The electron concentration and mobility for these films

are in the ranges of 10 8-1019 cm-3 and 40-60 cm2/Vs, respectively. The effect of different

growth conditions on the electrical properties of the films was delineated by analyzing

these Hall data. As growth temperature increases, resistivity decreases for the samples

grown under the same laser energy. The decrease of resistivity results from the increase

in carrier concentration, mobility or both. In addition, the resistivity exhibits an increase




































20 (degree)




Figure 6-4. X-ray diffraction of 0.01 at.% Al-doped ZnO films grown under different
conditions.


1.4x106


1.2x106


1.0x106


8.0x105


S6.0x105


4.0x105


2.0x105


15.5 16.0 16.5 17.0 17.5 18.0 18.5


Omega (degree)




Figure 6-5. Omega rocking curve ofZnO (0002) peak for 0.01 at.% Al-doped ZnO films
grown under 300 mJ laser energy.


0
0






o < o
N






Sample A
C

D b
E
F









with oxygen growth pressure for the Al-doped ZnO films grown under same laser energy

and growth temperature. Several experimental studies reported similar dependence of

resistivity on growth temperature and oxygen pressure for Al-doped ZnO films [152-154].

Kim et al [153] suggested that the resistivity of Al-doped ZnO films is related to the Al

doping concentration, oxygen vacancies, Al and Zn concentrations at interstitial sites,

grain boundaries and ionized impurity scattering. An increase in O/Zn ratio was observed

with increasing growth temperature by using Rutherford Backscattering Spectrometry

(RBS) [153]. Regarding to laser energy effect on the electrical properties of Al-doped

ZnO films, it is shown that with increasing laser energy the resistivity of the films

decreases for samples grown at 700 C (sample C vs. sample A). However, for the films

grown at 800C, resistivity does not show large increase with laser energy. Considering

the donor contributions in Al-doped ZnO discussed above, the electron density is

expected to increase with laser energy when the ZnO:Al target was bombarded by laser

beam to create more interstitial atoms or vacancies. The number of scattering centers

affecting the mobility of carriers was also increased with increasing laser energy and

deposition rate [155].

Table 6-3. Room temperature Hall measurement of 0.01 at.% Al-doped ZnO films under
different growth conditions.
Resistivity Carrier density Carrier Carrier
Sample (ohm-cm) (cm-3) mobility(cm2/Vs) type
A 1.2376 x 10-2 7.82 x 101 62.73 n
B 1.3656 x10-2 1.158 x 1019 39.48 n
C 1.142x 10-1 1.399x 101 39.12 n
D 9.375 x 10-3 1.32 x 1019 50.45 n
E 8.336 x 103 1.329 x 1019 56.36 n
F 1.1285 x 10-2 9.094 x 101 60.85 n

Room temperature PL (RT-PL) was performed to further investigate the effect of

growth conditions on the optical properties of the Al-doped ZnO films. Figure 6-6 (a) and









(b) show the laser energy on RT-PL for the Al-doped ZnO films grown at 700 C and 800

"C, respectively. The Al-doped ZnO films exhibit the band edge photoluminescence at

-377 nm with very low deep level emission. As laser energy increases from 300mJ to

350 mJ, the band edge emission increases for the films grown at both temperatures. Also

note that with decreasing growth temperature from 800 C to 700 C the deposition laser

energy shows more prominent effect on the band edge emission, which is consistent with

the laser energy dependence of resistivity discussed in the previous section. In addition,

as shown in Figure 6-6 (a) and (b) the band edge emission increases significantly with

growth temperature. These results suggest that the photoluminescence of the Al-doped

ZnO films has strong correlations to the electrical properties and crystallinity of the films.

It is known that there are two recombination processes, i.e. radiative and nonradiative

transition determining the light emission intensity. The photoluminescence efficiency can

be enhanced by increasing the radiative transition and decreasing the nonradiative

transition. For the Al-doped ZnO films, the intensity of band edge emission increases as

the resistivity decreases due to the increased laser energy and growth temperature. With

increasing electron density, the Fermi level will move up toward the conduction-band

edge resulting in more mid-gap defect states being filled up. Thus, the possibility of the

non-radiative trapping through those defect states will be decreased. For the films grown

at a higher temperature, the nonradiative defects can also be reduced by improving the

crystallinity of the films. Accordingly, the intensity of the band edge emission is

significantly enhanced by increasing growth temperature for Al-doped ZnO films.

The oxygen partial pressure effect on the photoluminescence of the ZnO:Al films is

shown in Figure 6-7. Note that the band edge emission intensity does not have linear









dependence on the oxygen partial pressure as resistivity does. As the oxygen pressure

decreases from 5 mTorr to 20 mTorr, the band edge emission decreases. However, with

further increasing oxygen pressure to 50 mTorr, the band edge emission markedly

increases. The film grown at 800"C, 50 mTorr oxygen pressure shows the highest band

edge emission of all the films. The above analysis is consistent with the XRD results

discussed earlier. With increasing growth temperature, the FWHM value of the omega

rocking curve of the films decreases. The film grown at 800 C, 50 mTorr oxygen

pressure has the narrowest rocking curve suggesting that crystallinity plays an important

role in the photoluminescence properties of the films.

The surface morphology of the films grown in different growth conditions were

also examined by performing AFM measurement in air. Figure 6-8 shows the AFM

images of the Al-doped ZnO films grown under different conditions. The growth

conditions of sample A-F are described in Table 6-1 previously. The scan area is 5 X

5 l m2 and the scan rate is 1 Hz. Interestingly note that growth temperature and oxygen

pressure greatly affect the surface morphology of the films. The films grown at 700C

have the smoothest surface with the root-mean-square (RMS) roughness in the range of

8-9 nm. However, as growth temperature increases to 800"C, roughness rapidly increases.

In addition, with increasing oxygen pressure the surface of the films also becomes

rougher. The sample grown at 800C, 50 mTorr and 300 mJ laser energy has the highest

RMS roughness about 38nm. Similar growth temperature and oxygen pressure

dependence of surface morphology was also observed for the P-doped ZnO and ZnMgO

films.






































350 400 450 500 550 600 650 700
Wavelength (nm)


1.8-

1.6-

1.4-

1.2-

'-
d 1.0o-

'5 0.8-

0.6-

0.4-

0.2-


T =800C, P2=20mTorr
g 02
/350 mJ






300 miJ











350 400 450 500 550 600 650 700
Wavelength (nm)


Figure 6-6. Laser energy effect on RT-PL for 0.01 at.% Al-doped ZnO films grown at (a)
700"C; (b) 800"C.







80




7
Laser energy=300 mJ
T =8000C
6- 50 mTorr g


5-


4-


3-

5 mTorr
2-


1 / 20 mTorr



350 400 450 500 550 600 650 700
Wavelength (nm)




Figure 6-7. Oxygen partial pressure effect on room temperature PL of 0.01 at.% Al-
doped ZnO films.






81



Sample A Sample B
RMS: 9.043nm RMS: 17.224nm











Sample D
Sample C RMS: 20.346nm
RMS: 7.952nm











Sample E Sample F
RMS: 13.553nm RMS: 37.837nm














Figure 6-8. AFM image of 0.01 at.% Al-doped ZnO films grown under different
conditions The z-scale is 40 nm/div














CHAPTER 7
GROWTH AMBIENT AND ANNEALING STUDY OF PHOSPHORUS-DOPED ZnO

Introduction

Systematic study of the effects of growth condition and post-annealing process is

essential to control and optimize the properties of P-doped ZnO films. It has been

suggested that Zn interstitials, O vacancies, and/or hydrogen complexes as compensation

donors in p-type ZnO. Thus, studies need to include understanding the role of oxidizing

species in yielding low native defect thin-film materials. The background impurity

density of p-type ZnO during growth also needs to be minimized so as to observe the

presence of acceptors in transport measurements. Previous studies [91] have focused on

the effects of annealing on the transport properties of 1-5 at.% P-doped ZnO films grown

by PLD. It showed that annealing significantly reduced the carrier density, yielding semi-

insulating behavior which is consistent with activation of a deep acceptor level. However,

no detailed studies have been carried out on the growth ambient effect on the properties

of P-doped ZnO films as well as the annealing effect on these films.

In this chapter, the effects of different oxidation ambients, growth temperature and

post-growth annealing on the properties of P-doped ZnO have been examined. The

transport, photoluminescence as well as surface morphology properties of the films were

discussed to elucidate these effects.

Experimental

Pulsed laser deposition was used to deposit phosphorus-doped ZnO epitaxial films

on c-plane sapphire substrates. The targets were fabricated using high-purity ZnO









(99.995%) mixing with P205 (99.998%) as the doping agent. The targets were pressed

and sintered at 1000 C for 12h in air. Considering the solubility limit of phosphorus in

ZnO, the phosphorus doping level in ZnO:P target was chosen to be 0.2 at.%. A KrF

excimer laser with a wavelength of 248nm was used as the ablation source. A laser

repetition rate of 1Hz was used, with a target to substrate distance of 4cm and a laser

pulse energy density of 1-3 J/cm2. Sapphire substrates were ultrasonically cleaned with

trichloroethylene (TCE), acetone and methanol for 5 min and dried in N2 prior to loading

into the growth chamber. The growth chamber base pressure was 2 x 10-7 Torr. Film

thickness was in the range of 400-500nm. P-doped ZnO films were deposited under

different oxidizing conditions including oxygen and 4% H2/Ar mixture, pure oxygen and

ozone/oxygen mixture. The partial pressure ratio of oxygen and Ar/H2 in the gas mixture

was 1:1. The nitrogen-free plasma discharge ozone generator yielded an 03/02 ratio on

the order of 1-3%. The same total growth pressure was maintained at 60mTorr for the

different ambients. The growth temperature ranged from 600"C to 800 C. Post-growth

annealing was carried out at temperatures ranging from 800"C to 1100"C in a flowing 02

ambient for Ihr. The transport properties of the as-grown and annealed films were

determined using four-point Van der Pauw Hall measurements at room temperature. The

photoluminescence properties of the films were also measured at room temperature using

a He-Cd laser (325nm). The surface morphology of the films was investigated using

atomic force microscopy (AFM).

Results and Discussion

Several studies discussed the passivation effect of hydrogen incorporation into ZnO

[156-158]. Ogata et al [156] reported that high resistive N-doped ZnSe films were









converted to p-type by thermal annealing in nitrogen atmosphere. A hydrogen passivation

model was proposed to explain the activation and passivation ofN acceptors resulting

from the dehydrogenation and hydrogenation. Accordingly, hydrogen passivation could

be very helpful to enhance the p-type doping of ZnO since it would prevent self

compensation during growth [158]. Thus, H2/Ar gas mixture was introduced during the

growth of P-doped ZnO to enhance the p-type doping of ZnO in this work.

Hall measurement was performed on these samples to examine the transport

properties at room temperature. All of the films show n-type conductivity which is

consistent with previous results on the as-grown P-doped ZnO films. Figure 7-1 shows

the resistivity as a function of deposition temperature for films grown under different

ambients. With increasing growth temperature, the difference in the resistivity of the

films grown under different ambients increases. At growth temperature of 600 "C the

resistivity is within an order of magnitude for the films grown under different ambient.

As growth temperature increases above 700"C, there is little change in resistivity for the

films grown in pure oxygen and 02/Ar/H2 mixture. In contrast, resistivity increases from

2.2 X 10-2 0 -cm to 3.5 0 -cm for the films grown in ozone/oxygen mixture.

Figure 7-2 shows the carrier density of ZnO: Po.002 films grown over a temperature

range of 600 C to 800 C. The doping of phosphorus in ZnO films exhibits a significant

increase in electron concentration which is in the range of 1019 to 1020 cm-3. However, for

the films grown in ozone/oxygen, the electron density rapidly decreases as increasing

temperature. Compared with the 02/Ar/H2 mixture and pure 02, the ozone/oxygen

ambient presents a stronger oxidizing species because of the weaker 0-0 bonds in 03.

According to the theoretical calculations from chemical potentials [76], the anion-rich











103
-- Ozone/Oxygen, PO2=60mTorr
-- OxWyen+Ar/H2,, Po,2=60mTorr
102 Oxyen,PO=60mTorr


10-


100



1 0 2 - - - - - -

101

600 650 700 750 800
Growth Temperature ("C)



Figure 7-1. Room temperature resistivity as a function of growth temperature for ZnO:
P0.002 films grown in different gas ambient.

(oxygen-rich) growth conditions could inhibit the formation of compensating defects in

p-type doping of ZnO. Therefore, the as-deposited ZnO:P films grown in ozone/oxygen

condition shows significantly lower electron density with increasing growth temperature.

The Hall mobility of P-doped ZnO samples is shown in Figure 7-3. With increasing

temperature, mobility continuously increases possibly due to the improved crystallinity of

the films. For the films grown at 800 C, in pure oxygen and 02/Ar/H2 mixture, the

mobility is about 35 cm2/Vs.

Consequently, the increase in resistivity for the films grown in ozone/oxygen is due

to the rapid decrease in carrier density even though there is a slight increase in mobility.

Note that the resistivity of the films grown in 02/Ar/H2 mixture is higher than those of the

films grown in pure oxygen at different temperatures resulting from a lower value of









electron density. To further investigate and understand the effect of growth condition on

the properties of P-doped ZnO films, photoluminescence needs to be measured. Figure 7-

4 (a)-(c) show the RT-PL spectrums measured for the as-deposited ZnO:P0.002 films

grown in 02/Ar/H2, pure oxygen and ozone/oxygen, respectively. A strong dependence

on different growth ambient and temperatures has been demonstrated. The film grown in

02/Ar/H2 mixture shows a stronger band edge emission than those of the films grown in

the other two ambients. With increasing growth temperature, this difference becomes

more prominent. At growth temperature of 800 C, the films grown in 02/Ar/H2 mixture

show a strong band edge emission at around 3.29 eV. The improvement in band edge

emission intensity for the films grown in 02/Ar/H2 mixture may reflect the passivation

effect of the deep acceptor-related levels by hydrogen, which also yields the passivation

of the deep level emission. Thus, the radiative transition efficiency through band-band

recombination is greatly enhanced.

Also note that for the films grown under 02/Ar/H2 mixture and pure oxygen, the

peak in the band edge emission shifts to slightly longer wavelength, i.e. lower energy

with increasing growth temperature. Regarding to the Hall results in Figure 7-2, the

carrier density of the films grown in 02/Ar/H2 and pure oxygen decreases with the growth

temperature. For both growth conditions, the carrier density is in the range of 1019 to 1020

cm-3. Therefore, this shift in PL could be explained by using the Moss-Burstein effect

which refers to an increase in the band gap due to the increase in Fermi level in highly

degenerate conditions [159].




































600 650 700 750 800

Growth Temperature (C)




Figure 7-2. Carrier density of ZnO: Po.oo2 films as a function of growth temperature.


-*- Ozone/Oxgen, Po=60mTorr
-m- OxygemAr/IH, P..=60mTorr
Oxgen,PO=60mTorr














____-------_---


600 650 700 750 800
Growth Temperature ("C)


Figure 7-3. Carrier mobility of ZnO: Po.oo2 films as a function of growth temperature.


-*- Ozone/Oxygen, POa=60mTorr
-- Oxygen+Ar/K2, Po2 =60mTorr
Oxygen,PO2=6OmTorr









' r "-------------


S ~ ~ ~ ~~ ~ -


11









The RT-PL properties of the films grown in ozone/oxygen are shown in Figure 7-4

(c). The intensity of the band edge emission is very low for the films grown in 03/02

mixture, almost quenched entirely with growth temperature as shown in the inset of

Figure 7-4 (c). The deep level luminescence shows much stronger intensity, which

increases greatly as growth temperature increases. Previous annealing studies of undoped

ZnO [160] have shown a similar dependence of PL on the annealing ambient, with a

decrease in band edge emission and an increase in visible defect-related luminescence as

ZnO is annealed in an oxidizing environment. However, annealing in the reducing

hydrogen ambient increases the band edge emission while subsequently decreasing the

deep level emission.

In addition to the growth ambient effect, the growth temperature also plays an

important role in the band edge emission. Increasing the deposition temperature improves

the UV band edge emission due to a reduction in the structural defects, although the

opposite effect is shown for the films grown in 03/02 mixture which is might due to the

more effective oxidization process as increasing temperature.

In an effort to reduce electron density and elucidate the phosphorus doping in ZnO,

the effect of annealing process on the properties of the ZnO:P0.002 films grown under

different ambients was examined. The as-grown films were annealed in flowing oxygen

for Ihr at temperatures ranging from 800 C to 1100C. Figure 7-5 (a)-(c) shows the

resistivity of the samples annealed at different temperatures and ambients as a function of

growth temperature. After the samples being annealed at 800"C in oxygen, the resistivity

significantly increases from 10-3 to several 0 -cm regardless of growth ambient. However,

with further increasing annealing temperature the resistivity starts to decrease. In the




Full Text

PAGE 1

DEVELOPMENT OF ZnO-BASED THIN FILM TRANSISTORS AND PHOSPHORUS-DOPED ZnO AND (Zn,Mg)O BY PULSED LASER DEPOSITION By YUANJIE LI A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2006

PAGE 2

Copyright 2006 by Yuanjie Li

PAGE 3

To my family

PAGE 4

iv ACKNOWLEDGMENTS I would like first to express my sincere appreciation to my advisor and committee chairman, Professor David P. Norton, for pr oviding me all the opportunities, guidance and motivations throughout my graduate studi es. I am grateful for his knowledge and support that have helped me to finish my disse rtation work. I would al so like to thank my supervisory committee members, Professor St ephen J. Pearton, Professor Paul Holloway, Professor Cammy R. Abernathy and Professo r David Tanner, for their suggestions and guidance. I also want to thank Professor Fa n Ren for his discussions and advice on my research work. I would like to thank Professor Andrew Ri nzler, Dr. Zhihong Chen and Dr. Xu Du for their help and suggestions on lithogra phy processing. I appreciate Jau-Jiun Chen, Dr. Valentin Craciun, Dr. Brent Gila, the Major An alytical Instrumenta tion Center staff and many other people for their colla borations and assistance. My gratitude also goes to my group members including Mat Ivill, George Er ie, Hyun-Sik Kim, Daniel Leu, Ryan Pate, Li-Chia Tien, Seemant Rawal, Charlee Calend er, and Patrick Sadik for their help during my graduate research. I would like to express my deepest appreci ation and love to my family members and friends for their unconditional love and inspiration. Especially, I want to thank my parents and my husband, Shengbo Xu, for thei r support and encouragement that have helped me to overcome many difficulties dur ing my development and make it to this point in life.

PAGE 5

v TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES............................................................................................................vii LIST OF FIGURES.........................................................................................................viii ABSTRACT....................................................................................................................... xi CHAPTER 1 INTRODUCTION........................................................................................................1 2 BACKGROUND REVIEW..........................................................................................5 Properties of ZnO.........................................................................................................5 Crystal Structure....................................................................................................5 Physical Properties................................................................................................6 ZnO growth methods....................................................................................................8 ZnO Single Crystal................................................................................................8 ZnO Thin Film.....................................................................................................10 Doping of ZnO............................................................................................................11 Intrinsic defects in undoped ZnO........................................................................12 N-type doping of ZnO.........................................................................................14 P-type doping of ZnO..........................................................................................15 Nitrogen doping............................................................................................16 Phosphorus doping.......................................................................................17 Arsenic doping.............................................................................................18 Nitrogen and group III codoping..................................................................18 3 EXPERIMENTAL TECHNIQUES............................................................................20 Film Growth via Pulsed Laser Deposition..................................................................20 Post-growth Annealing Process..................................................................................23 Experimental Characterization Techniques................................................................24 X-ray Diffraction.................................................................................................24 Scanning Electron Microscopy............................................................................25 Energy-dispersive X-ray Spectroscopy...............................................................25 Atomic Force Microscopy...................................................................................26

PAGE 6

vi Hall Effect Measurement.....................................................................................26 Photoluminescence..............................................................................................29 X-ray Photoelectron Spectroscopy......................................................................30 Current-voltage Measurement.............................................................................31 4 DEVELOPMENT OF OXIDE-BASED THIN-FILM TRANSISTORS...................32 Introduction.................................................................................................................32 Transparent Semiconducting Oxides for Thin-Film Transistors.........................34 ZnO-based Transparent Thin-film Transistors....................................................36 Deposition and Properties of Channel Materials........................................................38 Fabrication of ZnO-based TFTs.................................................................................39 ZnO-based TFT Device Characterization...................................................................47 ZnO-TFTs Using Undoped ZnO as Active Channel Layer.................................49 ZnO-TFTs Using P-doped ZnO and (Z n,Mg)O as Active Channel Layer..........50 5 GROWTH AND CHARACTERIZATION OF P-TYPE PHOSPHORUS-DOPED (Zn,Mg)O BY PULSED LASER DEPOSITION.......................................................53 Introduction.................................................................................................................53 Experimental...............................................................................................................54 Results and Discussion...............................................................................................55 6 SYNTHESIS AND CHARACTERIZ ATION OF (Zn,Mg)O:P/ZnO HETEROSTRUCTURES AND Al-DOPED ZnO.....................................................67 Introduction.................................................................................................................67 Experimental...............................................................................................................68 Results and Discussion...............................................................................................69 7 GROWTH AMBIENT AND ANNEALING STUDY OF PHOSPHORUSDOPED ZnO...............................................................................................................82 Introduction.................................................................................................................82 Experimental...............................................................................................................82 Results and Discussion...............................................................................................83 8 CONCLUSIONS........................................................................................................98 LIST OF REFERENCES.................................................................................................102 BIOGRAPHICAL SKETCH...........................................................................................112

PAGE 7

vii LIST OF TABLES Table page 2-1 Physical properties of ZnO and GaN.........................................................................7 2-2 Different epitaxy substrates for ZnO thin film growth............................................11 2-3 Valence and ionic radii of candidate dopant atoms..................................................17 4-1 Properties of transparent semiconducting oxides.....................................................35 6-1 Growth conditions of 0.01 at. % Al-doped ZnO films via PLD..............................70 6-2 FWHM values of ZnO (0002) omega rocking curve for 0.01at.% Al-doped ZnO..74 6-3 Room temperature Hall measuremen t of 0.01at.% Al-doped ZnO films under different growth conditions......................................................................................76

PAGE 8

viii LIST OF FIGURES Figure page 2-1 A schematic illustration of ZnO crystal structure......................................................6 2-2 Schematic of the hydrothermal growth system........................................................10 2-3 Defect formation enthalpies in Zn -rich and O-rich conditions after LDA corrections................................................................................................................14 3-1 Schematic illustration of a pulsed laser deposition system......................................23 3-2 A schematic of Hall effect on an n-type, bar-shaped semiconductor. The sample has a finite thickness of d.........................................................................................28 4-1 Schematic illustration of passi ve and active matrix displays...................................33 4-2 Electrical properties of undoped Zn O thin films grown on glass at 400 C as a function of oxygen pressure.....................................................................................40 4-3 X-ray diffraction pattern of the undoped ZnO film deposited at PO2=20mTorr on glass substrate...........................................................................................................41 4-4 An AFM image of the surface of th e undoped ZnO thin film deposited at PO2=20mTorr on glass substrate...............................................................................41 4-5 Schematic cross section view of a top-gate-type TFT structure..............................42 4-6 Schematic fabrication sequen ce of ZnO-based TFT structure.................................46 4-7 Top-view microscopy image of th e ZnO-TFTs on glass substrate..........................47 4-8 Drain current as a function of drain voltage characteris tics for the undoped ZnOTFT...........................................................................................................................4 9 4-9 The output characteristics of the TFT with alternative active channel materials: (a) P-doped ZnO as the channel; (b) P-doped (Mg,Zn)O as the channel.................51 4-10 Transfer characteristics of ZnO-TF T with P-doped (Zn,Mg) O as the channel layer at the drain voltage of 6V................................................................................52

PAGE 9

ix 5-1 ZnO (0002) omega rocking curves of P-doped (Zn0.9Mg0.1)O samples with and without LT-ZnO buffer layer...................................................................................57 5-2 Carrier concentration and carrier type in P-doped (Zn0.9Mg0.1)O films as a function of oxygen partial pressure..........................................................................58 5-3 Effect of oxygen partial pressure on carrier mobility of P-doped (Zn0.9Mg0.1)O films.......................................................................................................................... 58 5-4 Resistivity of P-doped (Zn0.9Mg0.1)O films vs oxygen partial pressure...................59 5-5 X-ray photoelectron spectr oscopy survey of P-doped (Zn0.9Mg0.1)O films.............62 5-6 X-ray photoelectron spectroscopy multiplex of P 2 s peak for P-doped (Zn0.9Mg0.1)O films grown at 500 150 mTorr oxygen pressure..........................62 5-7 X-ray diffraction of P-doped (Zn0.9Mg0.1)O films grown under different oxygen partial pressures........................................................................................................63 5-8 XRD scans of P-doped (Zn0.9Mg0.1)O film and sapphire substrate......................63 5-9 AFM images of the P-doped (Zn0.9Mg0.1)O films grown at different oxygen pressures: (a) 20; (b) 100 ; (c) 150; (d) 200mTorr....................................................64 5-10 RMS roughness of the P-doped (Zn0.9Mg0.1)O films as a function of oxygen partial pressure.........................................................................................................65 6-1 Schematic illustration of ZnMgO:P/ZnO heterostructure........................................70 6-2 The I-V curve of Au and Ti/Au meta l contacts on: (a) p-ZnMgO; (b) n-ZnO films, respectively....................................................................................................72 6-3 Current-voltage characteristics of the ZnMgO/ZnO heterostructure on: (a) sapphire; (b) ZnO substrate......................................................................................73 6-4 X-ray diffraction of 0.01 at.% Al-d oped ZnO films grown under different conditions.................................................................................................................75 6-5 Omega rocking curve of ZnO (0002) peak for 0.01 at.% Al-doped ZnO films grown under 300 mJ laser energy............................................................................75 6-6 Laser energy effect on RT-PL for 0 .01 at.% Al-doped ZnO films grown at (a) 700 ; (b) 800 ......................................................................................................79 6-7 Oxygen partial pressure effect on r oom temperature PL of 0.01 at.% Al-doped ZnO films.................................................................................................................80

PAGE 10

x 6-8 AFM image of 0.01 at.% Al-doped ZnO films grown under different conditions. The z-scale is 40 nm/div...........................................................................................81 7-1 Room temperature resistivity as a fu nction of growth temperature for ZnO: P0.002 films grown in different gas ambient.......................................................................85 7-2 Carrier density of ZnO: P0.002 films as a function of growth temperature................87 7-3 Carrier mobility of ZnO: P0.002 films as a function of growth temperature..............87 7-4 Photoluminescence spectra of P-doped ZnO grown in: (a) O2/Ar/H2; (b) pure oxygen; (c) ozone/oxygen........................................................................................89 7-5 Resistivity of 0.2 at.% P-doped ZnO films annealed at different temperatures in O2. The films were grown in: (a) O2/Ar/H2; (b) pure oxygen; (c) ozone/oxygen....92 7-6 Carrier concentration of 0.2 at.% P-doped ZnO film s annealed at different temperatures in O2....................................................................................................93 7-7 Mobility of 0.2 at.% P-doped ZnO films annealed at different temperatures in O2. The films were grown in: (a) O2/Ar/H2; (b) pure oxygen; (c) ozone/oxygen..........95 7-8 RT-PL of 0.2 at.% P-doped ZnO films a nnealed at different temperatures in O2. The films were grown at 800 and in O2/Ar/H2 mixture.......................................96 7-9 Surface morphology of 0.2 at.% P-dope d ZnO films annealed at different temperatures in O2. The films were grown at 700 and in 60 mTorr O2...............97

PAGE 11

xi Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy DEVELOPMENT OF ZnO-BASED THIN FILM TRANSISTORS AND PHOSPHORUS-DOPED ZnO AND (Zn,Mg)O BY PULSED LASER DEPOSITION By Yuanjie Li May 2006 Chair: David P. Norton Major Department: Materials Science and Engineering Top-gate type ZnO-based TFTs were fabricated on glass substrate via photolithography and wet chemical etching pr ocessing. The ZnO layers were deposited using pulsed laser deposition (PLD). N-cha nnel depletion-mode operation was shown for the undoped ZnO and P-doped ZnO thin film transistors. The current-voltage measurements demonstrated an enhancementmode device operation for the thin film transistors with P-doped (Zn,Mg)O as the active channel layer. P-type phosphorus-doped (Zn0.9Mg0.1)O films have been realized via PLD without post-annealing process. The conduction type of the films strongly depends on the oxygen partial pressure during the deposition pro cess. Increasing the oxyge n partial pressure from 20 to 200 mTorr yielded a carrier type conversion from n-type to p-type. The Pdoped (Zn,Mg)O films grown at 150 mTorr oxyge n partial pressure were p-type and

PAGE 12

xii exhibited a hole concentration of 2.71016 cm-3, a mobility of 8.2 cm2/Vs and a resistivity of 35 -cm. (Zn0.9Mg0.1)O:P/ZnO heterostructures were fabricated on sapphire and ZnO substrates via PLD with Au and Ti/Au served as Ohmic contacts. Both structures exhibit rectifying electrical characteris tics. The turn-on voltages were determined to be 1.36 V and 1.15 V for the structure grown on sapphi re and ZnO substrate, respectively. The resistivity of Al-doped ZnO depends on growth temperature, laser energy and oxygen pressure. The photoluminescence prope rties of the Al-doped ZnO films have strong correlations to the elec trical properties and crysta llinity of the films. The possibility of the non-radiative trapping through deep level defect states decreases with increasing the electron density of Al-doped ZnO films. AFM re sults showed that the rootmean-square roughness increases with growth temperature and oxyge n partial pressure. The resistivity of the as-deposited 0.2 at.% P-doped films grown in ozone/oxygen ambient rapidly increased with growth temperature. The improvement in band edge emission intensity for the films grown in O2/Ar/H2 mixture may reflect the passivation effect of the deep acceptor-re lated levels by hydrogen, which also yields the passivation of the deep level emission. O xygen interstitials may contribut e to the deep level emission of RT-PL for annealed P-doped ZnO films.

PAGE 13

1 CHAPTER 1 INTRODUCTION Semiconductor devices have been exerting a critical influence on our life since the first transistor was invented at Bell La bs in 1947. The development of advanced semiconductor materials to obtain desirabl e properties is one of the essential contributions to modern semiconductor de vices. Silicon (Si) as a conventional semiconductor material is approaching the theoretical limits by recent technology advances. To overcome the high power and high temperature limits of Si-based electronic devices, wide bandgap semiconductors such as silicon carbide (SiC ), gallium nitride (GaN) and diamond have been developed to be the better candidates. For semiconductor photonic devices such as ultraviolet (UV)/b lue light-emitting diodes (LEDs) and laser diodes, wide bandgap group III nitrides have been the focus of intensive research due to their specific properties. II-VI compound semiconductor zinc oxide (Z nO) has attracted much attention because of its unique combination of electri cal, optical, piezoelectric and acoustical properties for decades. With the developm ent of recent technologies, the research interests in ZnO are renewed in a broad ra nge of applications from optoelectronics, transparent thin film transistors (TFTs) and nanostructured materials to spintronics (spin + electronics). ZnO has a direct wide bandga p of ~3.3 eV at room temperature with a large saturation velocity and a high breakdow n voltage. Compared with GaN, ZnO has several important advantages making it ideal for UV LEDs and lasers [1]: (1) larger exciton binging energy of 60 meV (vs. ~2 5meV for GaN) enhancing the radiative

PAGE 14

2 recombination efficiency as well as lowering turn-on voltage for la ser emission; (2) the availability of large single crystal ZnO substrate desirable for vertical device development; (3) low-temperature epitaxial gr owth and (4) possible wet chemical etching process leading to potential low-cost Zn O-based devices. Polycrystalline ZnO gains much attention in transparent TFTs for the electronic flat panel di splay industry. Due to its transparency in the visible spectrum, minimal light sensitivity, and process temperatures compatible with glass/plas tic technology, ZnO-based transparent TFTs show possible solutions to the limitations of Si-based TFTs [2-4]. With a reduction in crystallite size to nanometer scale, ZnO introduces novel electrical, mechanical, chemical and optical properties with rich family of nanostructures such as nanorings, nanowires and nanobelts [5]. These one dimensional (1D) materials can be used to demonstrate the potential applications in nove l nanodevices. ZnO doped with transition metals such as manganese (Mn) and cobalt (Co) also shows pot ential in spintronic applications due to a predicted Curie temperature a bove room temperature [6,7]. The motivation for ZnO-based transparent TFTs is to overcome the opacity of Sibased TFTs for active matrix arrays. In addi tion, by using ZnO as ac tive channel layer in TFTs the channel mobility can be increased leading to faster devi ce operation and higher drive current [8]. Key challenges existing in ZnO-based TFTs include device structure fabrication and realization of an enhancement-mode device operation by controlled channel carrier densities. Undoped ZnO is intr insic n-type with electron concentration in the range of ~1018 cm-3. High electron density causes th e channel layer to be conductive even in the absence of applied gate voltage Previous results on annealed P-doped ZnO films showed that phosphorus substitution ma y introduce an acceptor level that reduces

PAGE 15

3 the native electron density in ZnO. Therefore, to deplete the channel electron carrier and realize the enhancement-mode operation de vices, annealed P-doped ZnO and (Zn,Mg)O films were employed as the active cha nnel materials in ZnO-based TFTs. In order to realize the pract ical applications of ZnO in optoelectronic devices, both n-type and p-type materials with high carrier concentration a nd low resistivity have to be achieved. While n-type ZnO is easily realized via Al or Ga doping, it has shown that ZnO has the particular difficulty in producing relia ble p-type conduction. This critical issue impedes the widespread development of th e ZnO-based UV LEDs and lasers. Therefore, achieving high-conductivity p-type ZnO has b ecome one of the key challenges for ZnObased optoelectronics. Recent studies in p-type doping of ZnO have focused on group V ions such as nitrogen (N), phosphorus (P) and arsenic (As) substituted on the oxygen site [9-11]. Thus, the motivation of this part of the dissertation research was to synthesize and characterize phosphorus-doped ZnO and (Zn,M g)O thin films for optoelectronic applications via using puls ed laser deposition (PLD). This introduction chapter presents th e challenges and motivations of this dissertation work. Following, a second chapte r reviews the related background, including general properties of ZnO; growth methods for single crystal and thin films of ZnO; current experimental and theoretical studies of nand p-type doping of ZnO. Chapter 3 describes the film growth and characterizat ion techniques employed in this work. In chapter 4, the fabrication process and device characteristics of the top-gate type ZnObased TFTs on glass substrates are describe d. Chapter 5 discusses the effect of oxygen partial pressure on the rea lization of p-type P-doped (Zn,Mg)O films grown on LT-ZnO buffer layer. The development of (Zn,Mg)O:P/ZnO heterostructures for light emitting

PAGE 16

4 applications is the focus in the following Chapter 6. The growth condition effect on the electrical and optical proper ties of Al-doped ZnO is also included in Chapter 6. The systematic studies of growth conditio n, post-annealing on th e electrical and photoluminescence as well as crystallinity and surface morphology of P-doped ZnO films are discussed in Chapter 7. Finall y, chapter 8 will give the conclu sions of this dissertation.

PAGE 17

5 CHAPTER 2 BACKGROUND REVIEW This chapter introduces the general pr operties of ZnO, including its crystal structure and physical parameters; growth met hods for ZnO single crystal and thin films; n-type and p-type doping of Zn O focused on recent experimental and theoretical studies. Properties of ZnO Crystal Structure Hexagonal wurtzite is the thermodynamically stable crystal structure for ZnO. There are two other phases known to exist. A zincblende phase is formed under some specific growth conditions [ 12]. A rocksalt structure can be synthesized under high pressure above 10 GPa at room temperatur e or above 6 GPa at 1200 K [13]. However, this rocksalt structure is difficult to retain under ambient conditions. In the hexagonal wurtzite structure, each Zn cation is surrounded by four oxygen anions at the corners of a tetrahedron, a nd vice versa. In other words, ZnO crystal structure is composed of alternating planes of Zn2+ and O2ions stacking along the c-axis. Figure 2-1 presents the schematic diagram of ZnO crystal structure. ZnO shows a highly ionic character due to the large difference in electronegativity betw een Zn and O atoms. Non-centrol symmetric tetrahed ral coordination in ZnO result s in piezoelectric properties and crystallographic polarity. The oppositely charged zinc-terminated (0001) Zn-face and oxygen-terminated (000 1) O-face produce spontaneous polarization along the c-axis [5].

PAGE 18

6 Figure 2-1. A schematic illustration of ZnO crystal structure. Physical Properties The physical properties of ZnO are compar ed to GaN in Table 2-1 [23-25]. ZnO and GaN have the same wurtzite crystal st ructure with ~1.9% la ttice mismatch on the cplane. Therefore, ZnO is a promising subs trate candidate for GaN epitaxy due to its stacking order match and crystal lattice matc h [14, 15]. High quality, low planar defects GaN epilayers have been grown on ZnO (0001) substrate via reac tive molecular beam epitaxy (MBE) [16]. ZnO has an exciton binding energy of 60 meV, which is much higher than the thermal energy at room temperature (26 meV). In principle, the excitonic recombination in semiconductors is a more efficient radi ative process and can enhance low-turn-on stimulated emission [17-19]. In order to rea lize efficient excitonic laser action at room temperature or even higher temperature, it is essential to have an exciton binding energy greater than thermal energy at room temperat ure. The first optical pumping laser action in single crystal ZnO grown by vapor phase method was reported by Reynolds, Look and Jogai [20]. The lasing occurred at a very low pump power (~ 4 Wcm-2) and the as-grown crystal planes act as the lasing cavity. D. M. Bagnall et al. [21] and P. Yu et al. [22] O2[0001] Zn2+

PAGE 19

7 reported UV laser emission in ZnO thin film s grown on sapphire substrates at room temperature. Although the electron Hall mobility in single crystal ZnO is lower than that of GaN, ZnO has a higher saturation velocity allowing it to compete with GaN in semiconductor devices applications. In addi tion, ZnO is highly resi stant to radiation damage making it suitable for space and ot her extreme operating conditions [1]. Table 2-1. Physical proper ties of ZnO and GaN. Property ZnO GaN Crystal structure Wurtzite Wurtzite Lattice constant at 300K (nm) a0 = 0.32495 c0 = 0.52069 a0 = 0.3189 c0 = 0.5185 Density (g cm-3) 5.606 6.15 Thermal conductivity (W/cm K) 0.6 1.3 Linear expansion coefficient (K-1) a0 = 6.5 x 10-6 c0 = 3.0 x 10-6 a0 = 5.59 x 10-6 c0 = 7.75 x 10-6 Melting point ( 1977 2497 Refractive index 2.008, 2.029 2.9 Bandgap at 300K (eV) 3.3 3.39 Exciton binding energy (meV) 60 25 Saturation velocity (cm s-1) 3.0 x 107 2.5 x 107 Breakdown voltage (V cm-1) 5.0 x 106 5.0 x 106 Alloying ZnO films with CdO (Eg = 2.3 eV) and MgO (Eg = 7.8 eV) makes bandgap engineering possible for real izing ZnO-based heterostruct ure devices [26-32]. For many advanced semiconductor devices, heterostructure designs are one of the key structures to improve the device performance via introduci ng band offsets and carrier confinement. The ionic radii of Cd2+ (0.74 ) and Mg2+ (0.57 ) are close to Zn2+ (0.60 ) [33]. According to the phase diagram of MgO-ZnO system, th e thermodynamic solid solubility of MgO in ZnO is less than 4 mol% [34]. However, prev ious work has reported the solid solubility of MgO in ZnO thin films to be up to 33 mol% while maintaining the ZnO wurtzite structure via using pulsed laser deposition [29]. Thus, it is possi ble to synthesize Zn1xMgxO metastable phases well above the thermo dynamic solubility limit by using pulsed

PAGE 20

8 laser deposition techniques. Cadmium (Cd) substitution on the Zn site leads to a reduction in the bandgap to ~3.0 eV [26] a nd substituting magnesium (Mg) can increase the bandgap up to ~4.0 eV [29]. A quantumconfinement effect by a blueshift in photoluminescence (PL) spectra was observed in ZnO/Zn0.8Mg0.2O superlattices grown by laser molecular beam ep itaxy (L-MBE) [35]. S. Choopun et al. summarized the bandgap relations in Zn1-xMgxO as a function of compos ition via a virtual crystal approximation [30]. For x = 0 to 0.33, the bandgaps of Zn1-xMgxO have a linear dependence on Mg content and the films reta in the hexagonal structure. For x = 0.33-0.35, there is a discontinuity in the bandgap relati on due to the structural transition from wurtzite to cubic phase. ZnO growth methods ZnO Single Crystal There are three important growth methods for bulk ZnO single crystal: pressurized melt growth [36], seeded sublimation growth [37] and hydrothermal solution growth [3840]. Pressurized melt growth employs the use of modified Bridgman configuration developed by Cermet, Inc. High quality Zn O wafers up to 2-inch in diameter are commercially available [41]. This technology involves a high pressure induction melting apparatus, where the melt is contained in a water-cooled crucible [36]. An overpressure of the oxygen as the growth atmosphere can overcome the ZnO decomposition problems during heating under normal melt growth pressu res. Radio frequency energy is used as the source to melt the material. Part of the molten phase is cooled by the cold crucible wall with the same composition as the melt. This cold material prevents the molten material from direct contacting with the cooling wall surface, which eliminates the

PAGE 21

9 contamination possibilities from the reactive crucible. High crystal quality (linewidths as low as 49 arcsec) and low defect density (104 cm-2 ) ZnO crystals can be produced in a fast growth rate (1-5 mm h-1 ) by using this method [36]. Seeded sublimation growth or vaporphase growth technique uses pure ZnO powder formed by the reaction of Zn-vapor and oxygen as the source material. This ZnO source is put at the hot end of a horizontal tube which maintains at a certain growth temperature. Hydrogen is used as a carrier gas during the process to make the growth reactions achievable. At the hot end of the tube the possible reaction follows ZnO(s) + H2(g) Zn(g) + H2O(g). A reverse reaction takes plac e at the cold end to form ZnO assisted by a single crystal seed [37]. High quality ZnO crystal has been grown by EaglePicher Technologies using this method. Ro om temperature mobility is about 225 cm2 /Vs and an electron concentration is in the range of mid-1016 cm-3 [37]. Several works have reported high qu ality bulk ZnO single crystals grown hydrothermally. Hydrothermal autoclaves made of high strength steel were used for crystal growth. During the growth, high purity ZnO nutrient mixed with the solvent (also called as mineralizer) is sealed in a plati num (Pt) inner container [40]. The purpose of using Pt inner container is to isolate th e growth environment from the wall of the autoclave. There are two zones in the Pt c ontainer: the crystal growth zone and the dissolution zone. Figure 2-2 shows a schematic hydrothermal growth system. The major parameters of this growth t echnique are the temperature of growth zone, the temperature difference between the two zone s, the pressure and the conc entration of the mineralizer [40]. Good quality ZnO single cr ystals have been achieved using this technology with a growth rate about 0.2mm per day [39].

PAGE 22

10 Figure 2-2. Schematic of the hydrot hermal growth system [40]. ZnO Thin Film ZnO epitaxial thin films have been grown via numerous deposition techniques including molecular beam epitaxy (MBE), pulsed laser deposition (PLD), radiofrequency (RF) magnetron sputtering and metal-organic chemical vapor deposition (MOCVD) etc. All the ZnO and (Zn,Mg)O film s presented in this dissertation were synthesized via using pulsed laser deposition. Th e detailed description of this technique is included in the following chapter. Different substrates for Zn O thin film epitaxy are listed in Table 2-2. The most common substrate used for ZnO thin films is sapphire due to its he xagonal structure, low cost and high crystal quality. (0001) c -plane sapphire has been used frequently for growing epitaxial ZnO films due to th e strong tendency of ZnO to grown in c orientation. Other possible substrates include (11 20) a -plane sapphire, (1 102) r -plane sapphire and SiC. Most of the substrates have a large la ttice mismatch with ZnO. The lattice-matched

PAGE 23

11 substrate for ZnO is (0001) ScAlMgO4 (SCAM) which consists of alternating stacks of rock salt layers and wurtzite layers [47]. A nu mber of studies have re ported the growth of epitaxial ZnO and ZnO-based device stru ctures on (0001) SCAM substrate [47-49]. Polycrystalline ZnO films can also be grown on inexpensive substrates such as glass at low temperatures [55, 56]. By using these char acteristics it is possible to realize thin film transistors by using ZnO as the active ch annel layer for displays and transparent electronic devices. Table 2-2. Different epitaxy substr ates for ZnO thin film growth. Substrate Crystal structure Orie ntation Mismatch (%) Reference c -plane Al2O3 (0001) 18.3 [42-45] a -plane Al2O3 (11 20) Â… [46] r -plane Al2O3 hexagonal (1 102) Â… [47] ScAlMgO4 hexagonal (0001) 0.09 [48-50] GaN wurtzite (0001) 1.9 [51] SiC wurtzite (0001) 5.5 [52] Si diamond (100) Â… [53,54] Doping of ZnO To realize ZnO-related materials in el ectronic and photonic applications, both high quality, low resistivity na nd p-type ZnO have to be ach ieved. However, wide bandgap semiconductors such as ZnSe, ZnS, ZnTe or CdS can be doped either n-type or p-type, but not both [57]. ZnSe and GaN can be eas ily doped n-type, but p-type doping is difficult to be realized. On the contrary, ZnTe is hard to dope n-t ype while p-type is formed readily. This asymmetry in n-type versus p-type doping also exists in ZnO [58,59]. While n-type ZnO is easily realized via Al, Ga or In doping, ZnO exhibits significant resistance to the formation of sh allow acceptor levels. To explore and realize the p-type doping of ZnO, it is essential to unde rstand intrinsic defects in undoped ZnO. The following sections discuss the intrin sic donors and acceptors in undoped ZnO,

PAGE 24

12 introduce the n-type doping of ZnO and larg ely focus on the recent experimental and theoretical studies of p-type ZnO doped with nitrogen (N), phosphorus (P), arsenic (As) as well as co-doped with N and group III elements. Intrinsic defects in undoped ZnO Undoped ZnO is an n-type semiconductor with various electron concentrations from mid-1016 cm-3 in high quality single crystal [60] to ~1018 cm-3 in epitaxial thin films grown on sapphire substrates. The origins of the dominant donors in undoped ZnO have been investigated both theoretically and e xperimentally [61-68]. In the study of native defects in ZnO by Kohan et al. [61] the most dominant native defects are suggested to be Zn and O vacancies depending on the Zn partial pressure. The calculations based on the first-principles pseudopotential method i ndicated that in Zn-rich conditions oxygen vacancy (VO) has lower energy than zinc interstitial (ZnI) for all Fermi-level positions. For oxygen-rich conditions, zinc vacancy (VZn) dominant over the whole range of Fermi level range. Zhang et al. [62] calculated the formation enthalpies of the intrinsic defects in ZnO such as VO, ZnI, Zn-on-O antisite (ZnO), VZn and oxygen interstitial (OI) based on the plane-wave pseudopotential total ener gy and force method. Figure 2-3 shows the defect formation enthalpies after the local density approximation (LDA) corrections. It suggests that both ZnI and VO have low formation enthalpi es. In contrast, the native acceptor defects VZn and OI have high formation enthalpies for Zn-rich conditions. Therefore, undoped ZnO shows intrinsic n-t ype conduction and can not be doped p-type by these intrinsic acceptor defects. Th e study, however, suggested that ZnI is a shallower donor than VO, thus it is considered to be the dominant intrinsic donor in ZnO. Look et al. [63] studied high-energy electron irra diation on both ZnO (0001) and (000 1) face

PAGE 25

13 samples and found the donors and acceptors prod uction rate is much higher for Zn-face irradiation. The donor is assi gned to a Zn-sublattice defect which has an activation energy of about 30 meV suggesting that ZnI (and not VO) is the dominant native shallow donor in ZnO. Recent work by Janotti et al. also indicated that VO is too high in energy to play any significant role in as-grown n -type material. However, it is expected to act as a compensating center in p-type ZnO and it can also be formed during irradiation and ion implantation [64]. Several experimental and th eoretical studies have show n that there is another candidate for the dominant donor defect in undoped ZnO [65,66]. First-principles calculation based on density functional theory (DFT) found that high electron concentration (1017 -1018 cm-3) in undoped ZnO may result from the presence of hydrogen (H) as a shallow donor [65]. Hydrogen forms a strong bond with oxygen providing a strong driving force to incorporate into ZnO. The incorporation of hydrogen in ZnO results in significant relaxations of surrounding atoms and always occurs as a positive charge. The formation energy of H+ is low to allow a larg e solubility of hydrogen in ZnO. Since hydrogen is generally present in the films growth ambient and has a high diffusion coefficient into ZnO, control of hydr ogen exposure during film growth has to be carefully carried out. The native acceptors in undoped ZnO including zinc vacancy (VZn) and oxygen interstitial (OI) have high formation enthalpies in Zn -rich conditions, so these defects are not abundant [61]. Recent work by Tuomisto et al. [68] studied the dominant acceptor in undoped ZnO by using temperature-dependent Hall measurements and positron annihilation spectroscopy. It was shown that VZn densities in as-grown and irradiated

PAGE 26

14 ZnO have a good agreement with the total acc eptor densities determ ined by temperaturedependent Hall measurements. Thus, VZn is identified as a possible dominant acceptor in both as-grown and irradiated ZnO. Figure 2-3. Defect formation enthalpies in Zn-rich and O-rich conditions after LDA corrections [61]. N-type doping of ZnO Group III elements such as Al, Ga and In can act as extrinsic donors for ZnO by substitution on Zn site. High quality and high electron concentration n-type ZnO epitaxial films have been successfully synthesized by using MBE, sputteri ng or MOCVD [69-72]. Highly conductive and transparen t n-type ZnO films have been utilized as a potential candidate to replace indium tin oxide (ITO) for displays and solar cells. Miyazaki et al.

PAGE 27

15 [69] reported transparent conductive Gadoped ZnO films deposited on glass substrate having a minimum resistivity of 2.2 x 10-4 cm at substrate temperature of 250 Highly conductive Al-doped ZnO thin films w ith average transmittance over 91% in the visible spectrum region has been deposite d by using photo-MOCVD method [70]. Chang et al. [72] also synthesized Al-doped ZnO thin films on silicon and glass substrates by RF magnetron sputtering. The dependence of film properties on different growth parameters such as RF power, substrate temperature, oxygen pressure and target composition was examined systematically. P-type doping of ZnO As mentioned earlier, theoretical and experimental studies have shown the asymmetry of n-type versus p-type dopi ng of ZnO. There are several possible mechanisms to explain these difficulties [ 73-75]. First, the acceptor dopants may have low solubility in the host material to limit th e accessible hole concentration. In ZnO, the native donor defects have low formation enth alpies and thus compensate extrinsic acceptors. Zunger [76] proposed that in ZnO, the p-type pinning energy EF (p) where the native “hole killers” spontane ously form to compensate ptype doping is considerably above the valence band maximum (VBM). The Fermi energy EF in extrinsic p-type doping which moves downwards will encounter EF (p) first before encountering the VBM. As a result, “hole killers” such as ZnI and VO will be generated spontaneously before any significant doping commences. Moreover, acceptor dopants can form deep levels in ZnO that are not easily activated at room temperature. The dopi ng difficulty in p-type ZnO may also result from other compensation mechanisms such as the formation of deep defect AX centers through a double broken bond (DBB) mechanism [73] The net result

PAGE 28

16 of the formation of these defect complexes is releasing two electrons (equivalent to capture of two holes). Recent research in p-type dopants has mainly focused on group V ions such as N, P, As or Sb substituted on the O site. The following sections will present these results. Nitrogen doping Table 2-3 lists the valence and ionic ra dii of candidate acceptor dopants. Among the group V atoms, N has the clos est ionic radius to that of O. Theoretical studies also suggested that N has some advantages for p-type doping of ZnO such as the smallest ionization energy and metastable N AX complexes [73]. There have been significant activities focused on N doping by using di fferent nitrogen sources including NH3, N2, N2O, NO and Zn3N2 [77-88]. Minegishi et al. [77] reported p-type ZnO doped with N by simultaneous addition of NH3 in hydrogen and excess Zn in source ZnO powder. A proposed model explained that th e high resistivity of about 100 cm may be reduced via thermal annealing. Look et al. [78] reported p-type ZnO grown by MBE using N2 RF plasma source and Li-diffused semi-insula ting ZnO substrates. The N-doped ZnO layers showed a hole concentration of 9 x 1016 cm-3, a mobility of 2 cm2/Vs and a resistivity of 40 cm. In another case, p-type ZnO has been grown by PLD, in which a N2O plasma is used for doping [80]. Iwata et al. [81] grew N-doped ZnO using MBE by introducing N2 and O2 through a RF radical source. However, type conversion from n-type to p-type did not occur even though the N concen tration was in the range of 1019 cm-3. It is suggested that the formation of N-N related comple xes introduce compensating defects. Thus, dopant sources contain only one nitrogen atom such as NO and NO2 are considered to be better choices due to the large di ssociation energy of N-N in N2 (9.76 eV) [75]. Yan and

PAGE 29

17 co-workers [82] proposed a theoretical model predicting that the defect formation energy of N on O site (NO) from NO is negative in Zn-rich c onditions and lower than that from N2O. Li et al. [83] reported on the realization of p-type ZnO using NO gas as the dopant source by MOCVD. The carrier concen tration is in the range of 1015-1018 cm-3 and mobility is in the range of 10-1 cm2/Vs. Table 2-3. Valence and ionic radi i of candidate dopant atoms [75]. Atom Valence Radius ( Zn 2+ 0.60 Li 1+ 0.59 Ag 1+ 1.00 O 21.38 N 31.46 P 32.12 As 32.22 Sb 32.45 Phosphorus doping While there have been many activities focused on nitrogen doping, few reported efforts have addressed phosphor us doping. Nevertheless, Aoki et al. reported p-n junction-like behavior between an n-type ZnO substrate and a surface layer that was heavily doped with phosphorus [89]. Zinc-phosphide (Zn3P2) used as the phosphorus dopant source was decomposed into Zn and P atoms by excimer laser radiation in high pressure oxygen or nitrogen ambient. Li ght emission also was observed by forward current injection at 110 K indicating that a p-type ZnO was formed. Work by K. Kim et al. on P-doped ZnO films grown by sputter deposition showed that n-type as-grown films we re converted to p-type by a thermal annealing process to activate the P dopant [90]. In this case, phosphorus was doped in ZnO using the ZnO target mixed with 1 wt% P2O5. XPS results of the p-type ZnO showed that the P2p peak came from P2O5 in the film. The p-type films showed a low resistivity of 0.59-4.4 cm,

PAGE 30

18 a hole concentration of 1 x1017-1.7 x 1019 cm-3 and a mobility of 0.53-3.52 cm2 /Vs. These results indicate that phosphorus is another promising acceptor dopant for p-type ZnO. Our previous annealing studies on P-doped ZnO films grown on sapphire substrates by PLD showed that post-growth annealing process yielded semi-i nsulating behavior which is consistent with activation of a deep acceptor level [91]. Arsenic doping Although density functional theo ry (DFT) predicts that AsO should have a very deep acceptor level (~1.15 eV) based on the large ionic size difference between As and O, Ryu et al. [92] have synthesized p-type ZnO by As diffusion from a GaAs substrate. The hole concentration can be increased up to the mid1017cm-3 range. Homostructural ZnO p-n junctions based on As-doped ZnO has also been reported [93]. Based on firstprinciples calculations, a m odel for large-size-mismatched group-V dopants such as As and Sb in ZnO was proposed by Limpijumnong and co-workers [94]. These dopants do not occupy the O sites, but on the Zn sites: each forms a complex w ith two spontaneously induced Zn vacancies in a process that involv es fivefold As coordination. The model also predicted that -type ZnO doped with As could be re alized under O-rich growth or annealing conditions. Nitrogen and group III codoping A codoping method refers to using N accepto rs and group III donors such as Ga, Al or In as the reactive codopant to increase th e N incorporation in Zn O [95]. The theoretical calculations showed that codoping wurtzite ZnO with N and group III elements, the distance between two N acceptors decreases from 6.41 to 4.57 indicating the enhancement of N incorporation. It also showed that the simultaneous codoping

PAGE 31

19 decreases the Madelung energy of p-type codoped ZnO compared with p-type ZnO doped only with N. Several research groups repor ted the realization of p-type ZnO using codoping method. Joseph et al. [96] demonstrated p-type ZnO films by applying Ga and N codoping method. A resistivity of 0.5 cm and a carrier concentration of 5 x 1019 cm-3 have been obtained in ZnO films on glass substrate. Singh et al. [97] also reported p-type conduction in ZnO by using N and Ga codopi ng technique. Ye and co-workers [98,99] reported p-type ZnO thin films realized by the N-Al codoping method. Secondary ion mass spectroscopy showed that the N incorpora tion was enhanced by the presence of Al in ZnO. The lowest room temperature resistivity is 57.3 with a Hall mobility of 0.43 cm2 /Vs and carrier conc entration of 2.25 x 1017 cm-3 for the N-Al codoped film grown on glass substrate.

PAGE 32

20 CHAPTER 3 EXPERIMENTAL TECHNIQUES This chapter describes the details of the experimental techniqu es and procedures applied to the deposition of phosphorusdoped ZnO and (Zn,Mg)O thin films, postgrowth annealing process and characterizat ion measurements for structural, surface morphology, transport, optical as well as ch emical state properties of the films. Film Growth via Pulsed Laser Deposition Pulsed laser deposition has been used fo r epitaxial growth of thin films and multilayer/superlattice of complex materials. It provides many important advantages for oxide films with high melting points and comp licated stoichiometry, some of which can not be achieved by using other growth tec hniques such as sputtering, MBE or MOCVD [100]. PLD was the first method used to successfully deposit high-temperature superconducting thin films [101,102]. The main advantage of PLD derived from the laser material removal mechanisms. A pulsed laser ablates the target with very high peak energy to evaporate the target material without change in the target composition. The evaporants, also called a plume, arrives at the heated sample substrate and then starts the film growth. As a result, films with the desi rable stoichiometry similar to the target can be obtained by using PLD. Other advantages of PLD include the easy adaptation to different operational modes with no constraint s from the internal powered sources and the ability to change the deposition gas pre ssure over a broad range as well as a high deposition rate [100].

PAGE 33

21 In principle, the film deposition process in PLD can be divided into three separate stages: (1) the interaction of the laser b eam with the target material; (2) dynamic formation of plasma; (3) plasma isothermal e xpansion and deposition of thin films [103]. In the first stage, the target material is ra pidly heated above its melting point by the high peak energy laser pulses. The evaporation of the target material with the same stoichiometry as the target occurs. The tu rn-on energy defined as the minimum laser energy above which appreciable evaporation is observed depends on the laser wavelength, pulse duration, plasma losses as well as th e optical and thermal properties of target material [103]. The interaction of the lase r beam with the targ et involves several mechanisms such as thermal, collisiona l, electronic, exfoliational and hydrodynamic sputtering [100]. The evaporated mixture of energetic speci es including atoms, molecules, electrons, ions, and micron sized particulat es further interacts with the laser beam by absorbing the laser energy to generate a high-temperature pl asma. The particle density in the plasma depends on the ionization degree, evaporation rate and the plasma expansion velocities. Most of the evaporated species are deposited perpendicular to the laser spot. However, the thickness variation existi ng in the films is larger th an that from a conventional thermal evaporation process [103]. The plasma expansion is isothermal in v acuum due to the termination of the laser pulse. With the kinetic energy in the plasma it retains its elliptical shape during the deposition process. Nucleation and growth of film starts after a condensation region on the substrate is formed. Substrate temperat ure and surface mobili ty of the deposited species affect the quality of thin film. Two dimensional, i.e. layer-by-layer growth is

PAGE 34

22 more preferable compared to the three dimens ional (island) growth in order to obtain high quality epitaxial thin films. Compared to the laser-target interaction and film deposition mechanisms, the system setup is much simpler. A PLD system is usually composed of an excimer laser, optical elements to guide and focus the la ser beam and a vacuum chamber. Figure 3-1 shows a schematic of a PLD system along with the laser and optic lenses. In this work, a Lambda Physik (COMPEX 205) KrF excimer laser which delivers 248 nm wavelength and maximum average power of 50 W was used. Th e laser influence is in the range of 1-3 J/cm2 and a laser repetition rate from 1 to10 Hz was used for the film growth. Multiple targets can be installed inside the chambe r simultaneously to realize multilayer and superlattice structures with a target to substrate distance of 4-5 cm. A quartz lamp heater was used to heat sample substrates up to 950 The chamber pumping system consists of an oil-free diaphragm roughing pump and Pfeiffer turbo pumps. The base pressure of the growth chamber was in the range of 8 x 10-8 -2 x 10-7 Torr. By using a mass flow control valve (MKS 600 series) it is possible to adjust input of different gases into the chamber in a wide range of deposition pressures from high vacuum (~10-7) to 10-1 Torr. There are several factors that play importa nt roles in obtaining high quality thin films grown via PLD. The laser energy density on the target has a significant effect on particulate formation. In order to make la ser ablation occur, th e laser energy has to exceed the turn-on energy of the target materi al. However, if the laser energy is much greater than the turn-on energy, the density of the particulates will in crease. Therefore, it is important to use the optimal laser ener gy density to obtain high quality films. An external energy meter was used to determine the actual laser energy reaching the target in

PAGE 35

23 Figure 3-1. Schematic illustration of a pulsed laser deposition system. front of the chamber laser window. The laser energy can be adjusted by panel settings based on the external energy meter values dur ing every experimental growth. In addition, pre-ablation of the target is necessary to ha ve a stable deposition rate and decrease the particulate formation. Pre-ablation of the target s with a high laser repe tition rate of 10 Hz was used for 1000-2000 laser shots before the films growth. Depositi on gas pressure and substrate temperature also infl uence the films growth rate, the kinetic energy distribution of depositing species, as well as the films crysta l structure [100]. Therefore, it is essential to carry out the systematic studi es of these growth condition e ffects on the film properties. Post-growth Annealing Process There are two different post-growth a nnealing processes conducted in this dissertation work: in-situ cham ber annealing and tube furna ce annealing. In-situ chamber annealing was carried out af ter film growth. The annealing gas pressure could be Laser Beam Target Carousel Target Focusing Lens Laser Window Substrate Heater

PAGE 36

24 increased to above 100 Torr. This annealing process overcomes the possible surface contamination imposed by taking samples out the vacuum chambe r for post-annealing. However, the annealing gas pressure is limite d to certain levels. Compared to chamber annealing, tube furnace is more versat ile in realizing hi gh temperature (~1100 ), controlled gas pressure (up to 1 atmosphe re) and different ambients annealing. An alumina crucible containing the samples was loca ted at the center of a quartz tube, where the temperature has been calibrated. The a nnealing tube was pur ged with high purity annealing gas for up to 8 hours before increa sing the temperature. Desirable heating and cooling rates can be reali zed by using a pre-programmed temperature controller. Experimental Characterization Techniques This section describes the different charac terization techniques used to investigate the structural, surface morphology, transport, optical and chemical bonding properties of the as-grown and annealed films. The effects of growth pressure, temperature, ambient, dopant concentration and ann ealing conditions on film pr operties were explored. X-ray Diffraction X-ray diffraction (XRD) is an important technique to analyze the crystallinity, phase, strain, preferred orientat ion and defects of samples. A collimated beam of X rays is incident on a sample and diffracted by the cr ystalline phases in the sample according to BraggÂ’s law such that sin 2 d (3-1) where is the wavelength of the incident X-ray, d is the spacing between atomic planes in the crystalline phase, is the angle between atomic planes and the incident X-ray beam. The intensity of the diffracted X-rays is measured as a function of the diffraction angle 2 This diffraction pattern is used to iden tify the sample crystalline phases [104].

PAGE 37

25 A Philips APD 3720 X-ray diffractometer with Cu K ( =1.5406 ) was used to examine the films structure and crystallinity in this work. The 2 scan range is from 30 to 75 degree. Omega rocking curve was measured by using a four-circle Philips XÂ’pert X-ray diffractometer. The full-width-at-half-ma ximum of the rocking curve was used to delineate the films crystallinity. Scanning Electron Microscopy Scanning Electron Microscopy (SEM) is a useful instrument to examine the sample surface when the light microscope reaches its resolution limits. In the SEM, an electron beam is focused into a fine probe and sca nned over a small area on the sample. Different signals such as secondary electrons, photon em issions as well as internal currents are created by the electron beam and sample inte ractions. These emitted signals are collected by detectors and subsequently an image is pr oduces on a cathode ray tube (CRT) [104]. A JEOL 6400 scanning electron microscopy was used to investigate the sample surface in low and high magnifications. The SEM was operated at 10-15 keV depending on the conductivity of the sample surface. The sample surface was not coated with a carbon layer and preserved in its original stat us for other characteri zation measurements. Energy-dispersive X-ray Spectroscopy Qualitative and semi-quantitative analysis of elements presenting in a sample can be achieved by using energy-dispersive X -ray spectroscopy (EDS). When atoms are ionized by a high-energy radia tion they emit characteristic X-rays. A solid state detector made from Si(Li) is used to convert these X-rays into signals wh ich are then processed into an X-ray energy spectrum. Most app lications of EDS are in electron column instruments like SEM, field emission-SEM, transmission electron microscope [104]. In

PAGE 38

26 this dissertation work, EDS was used to determine the composition of the surface particulates as well as the films. However, the relative accuracies are about 10% for the elements with concentrati on less than 5 wt% [104]. Atomic Force Microscopy In order to measure the surface roughne ss and topography with atomic resolution, atomic force microscopy (AFM) is a valuab le technique. Unlike electron microscopes, the resolution of AFM is determined by th e size of the tip instead of electron beam diffraction effects. In addition, a wide range of samples including metals, polymers, glasses, semiconductors, thin films and composite s can be measured in air even in liquids. AFM works by measuring attractive or repulsive van der Waals forces between the atoms of a tip and sample surface [105]. The magnit ude of the deflection resulting from these forces depends on the tunneling current and the tip-to-sample dist ance. A highly positionsensitive photodiode detects th e tip cantilever deflection a nd converts the signal to an image. In this work, AFM Dimension 3100 (Digita l Instrument, Inc.) was performed in contact mode to obtain the surface topographic images and roughness of as-grown and annealed samples under different conditi ons. Root-mean-square (RMS) roughness was calculated based on 2 x 2 or 5 x 5 m2 scan area. The scan rate is in the range of 1-1.5 Hz. Hall Effect Measurement Since its discovery by Edwin H. Hall (1879) the Hall effect has become one of the most important electrical characterization methods of materials. The Hall effect provides

PAGE 39

27 a relative simple way of measuring the carri er density, electrical resistivity and the mobility of carriers in semiconductors. The principle underlining the Ha ll effect is Lorenz force wh ich is defined as a force exerted on a charged particle in an electro magnetic field. Figure 3-2 shows an n-type, bar-shaped semiconductor. It is a ssumed that a constant current I flows along the x-axis from left to right in the presence of a z-directed magnetic field. Under Lorenz force electrons drift toward the negative y-axis a nd accumulate on the side of the sample to produce an electrical surface charge. As a result a potential drop across the sample called Hall voltage is formed. The induced electric field increases until it counteracts to the opposite Lorenz force. In this case, ne eBj B e eEx x y/ (3-2) where eEy is the induced electric field force, evxB is the Lorenz force, jx=-nevx is the total current density. The Hall coefficient RH is defined as ne RH1 (3-3) The mobility is defined as the coefficient of proportionality between v and E and measured as follows: H x x xR neE j E v (3-4) where is the conductivity. For p-type semic onductors, a hole has a positive charge e Therefore, the Hall coefficient is positive in sign. The Van der Pauw technique which re quires no dimension measured for the calculation of sheet resistance or sheet carrie r density solves the potential problem in a thin layer of arbitrary shape. Thus, this me thod has increased in popularity relative to the Hall-bar configuration. The validity of the va n der Pauw method requires that the sample

PAGE 40

28 Figure 3-2. A schematic of Hall effect on an n-type, bar-shaped semiconductor. The sample has a finite thickness of d [106]. be flat, homogenous, isotropic, a single conne cted domain and have line electrodes on the periphery [107]. In addition, the contact size has to be sufficiently small to reduce the measurement corrections. In this dissertation work, a Lake Shore 7500/9500 Series Hall System was used to perform Hall measurements on the samples at room temperature. The van der Pauw method was applied for the measurements. The samples have square geometry with contact size to sample periphery ratio as low as possible. In order to eliminate the persistent photoconductivity (PPC) relaxation effect on the trans port properties, the samples were maintained in the dark for 12 h prior to performing Hall measurements. PPC effect has been observed in many III -V and II-VI compound semiconductors and can be explained by severa l theoretical models [108110]. In many cases, PPC is attributed to the existence of deep defects such as DX centers which form when shallow

PAGE 41

29 donors undergo a large lattice relaxation and co nvert to deep donors [108]. The difference in lattice relaxation between these two states results in a barrier that prevents the recapture of carriers into the stable state, thus yielding the PPC effect [109]. Other possible models include band-bending resulted from the interface leading to PPC effect and random local-potential fl uctuation model inducing the separation of photoexcited carriers from traps and reduced recapture rate of carriers. Photoluminescence Photoluminescence (PL) refers to emi ssion of light resulting from optical stimulation. The detection and an alysis of this emission is widely used as an analytical tool due to its sensitivity, simplicity, and low cost [104]. When an electron increases energy by absorbing light there is a transition from the ground st ate to an excited state of an atom or molecule. This excited system does not have the lowest energy and has to return to the ground state. In luminescence mate rials the released ener gy is in the form of light, which is called as radiative trans ition. This emitted light is detected as photoluminescence, and the spectral dependen ce of its intensity provides information about the properties of the materials. Speci ally, photoluminescence of a semiconductor is related to both intrinsic and extrinsic defects in the material which usually create discrete electronic states in the bandgap and therefor e influence the optical emission of the material. PL can be performed both at room temp erature and low temperatures. At room temperature, PL emission is thermally broade ned. With decreasing temperature, PL peaks tend to be much sharper and the emission intensity is also stronger due to fewer nonradiative transitions. PL is normally usef ul for semiconductors which have direct

PAGE 42

30 bandgap. However, at low temperatures, localized bound states and phonon assistance allow certain PL transitions to occur even in materials with indi rect bandgap [104]. A He-Cd (325 nm) laser was used as the ex citation source in r oom temperature PL measurements for the samples presented in this dissertation. The measurements were taken in a wavelength range of 340 to 800n m. A NESLAB chiller cooled GaAs PMT detector was used for UV up to ~900 nm. X-ray Photoelectron Spectroscopy X-ray photoelectron spectroscopy (XPS) is the most broadly applicable surface analysis technique today due to its surface sensitivity, quantitative and chemical state analysis capabilities [104]. All elements except for hydr ogen and helium can be detected by using XPS. In XPS monoenergetic X-rays bombard a sample and cause photoelectrons to be ejected. Einstein photoelectric law in Equation (3-3) defines the relationship of the kinetic energy of the ejected photoelectrons and the binding energy of the particular electron to the atom. KE = h -BE (3-3) where KE is the kinetic energy of the photoelectron, h is the X-ray photon energy and BE is the binding energy. By measuring the photoelectron kinetic energy the characteristic binding energy of the electron in the atom can be determined. The depth of the solid samples varies from the top 2 atomic layers to 15-20 layers. Perkin-Elmer PHI 5100 ESCA system was used to examine the chemical environment of P in P-doped ZnO and (Zn,Mg)O films. The base pressure of the XPS chamber was about 1 x10-9 Torr after increasing the voltage to 15 kV and the power to 300 Watt. Both Mg and Al anodes were used to acquire XPS spectrums of different

PAGE 43

31 samples. Argon gas sputtering to remove the sample surface contamination is necessary to obtain more accurate information about the film composition. Current-voltage Measurement Current-voltage (I-V) measurements were carried out at room temperature to characterize the performance of the devices including ZnO-base d thin film transistors and (Zn,Mg)O/ZnO heterostructures. The electri cal properties of ohmic contacts for Hall measurements were also examined. Semi conductor parameter analyzers Agilent 4155A and HP 4145B connected to a probe station were used in this dissertation work.

PAGE 44

32 CHAPTER 4 DEVELOPMENT OF OXIDE-BASED THIN-FILM TRANSISTORS Introduction Display technologies based on orga nic/polymer light emitting diodes (OLED/PLED) are promising for providing light weight, power efficient, flexible, high brightness performance. One challenge facing polymer light emitting devices is the thinfilm transistors (TFTs) arra y control circuitry [111-116]. In active matrix displays, each pixel is programmed to provide a desira ble current during the entire frame time, eliminating the issue of con tinuous increased current density encountered in the passive matrix approach. Figure 4-1 shows a schematic illustration of passive and active matrix displays. The light-emitting pixels in active matrix displays may be controlled by a thin film transistor array for much better bright ness and efficiency. However, current opaque TFTs made with amorphous and poly silicon severely restrict the amount of light detected by observer. Depending upon the design of the array and inte rconnects, only a fraction of the emitted light is used by the observer of the information resulting in significant energy loss. In addi tion, TFTs based on amorphous Si have other limitations such as light sensitivity, light degrad ation and low field effect mobility (1 cm2/Vs) [11]. One method to overcome these problems is to utilize the recent pr ogress in transparent oxides that are semiconducti ng to near-metallic. Transparent conducting oxides (TCOs) can be regarded as a specific group of oxides exhibiting both high optical transmitta nce and electrical conductivity. There are many applications for TCOs such as transp arent electrodes for flat-panel display and

PAGE 45

33 Figure 4-1. Schematic illustration of pa ssive and active matrix displays [117]. solar cell, transparent electronic devices and selective window coatings in architecture. N-type TCOs whose dominant carriers are el ectrons include indium tin oxide (ITO), SnO2, impurity-doped ZnO, CdO, as well as their related multicomponent oxides [117119]. As used for thin film transparent electrodes, TCOs should have a carrier concentration of order of 1020 cm-3 or higher and bandgap above 3 eV [120]. With the increasing need for transparent electrodes fo r optoelectronic devices it is essential to develop alternatives to ITO due to its e xpensive indium source. Gaand Al-doped ZnO with a resistivity on order of 10-5 cm are promising candidates for thin film transparent electrode applications [121]. In addition to binary compounds, ternary compounds such

PAGE 46

34 as Cd2SnO4, CdSnO3, CdIn2O4, Zn2SnO4, and CdSb2O6 have been developed to be used as n-type TCOs [120]. On the othe r hand, p-type TCOs such as CuAlO2, CuGaO2, and SrCu2O2 have been demonstrated recently [122126], offering the promise of transparent twoand three-terminal devices. In fact, p-n junctions from n and p-type TCOs have been realized [124], establishing the feasibility of a transparent oxide bipolar transistor. With this development, the TFT and apertures could be replaced by a transparent TFT (TTFT) array with significant gain in the emitted li ght. This should significantly lower both the voltage and drive current requ ired by the PLEDs. The sun light readability and extended device lifetime can also be realized. Transparent Semiconducting Oxid es for Thin-Film Transistors A distinction should be made between tr ansparent semiconducting oxides (TSOs) and TCOs based on the different carrier densities for channe l and interconnect performance. Table 4-1 lists the candidate ma terials, along with th e range of electronic and photonic properties. For the channel laye r, significant modulation of the channel conductance is needed to achieve field eff ect transistor (FET) switching. Thus, only moderate carrier densities (semiconducting beha vior) are needed for the channel layer. On the other hand, TCOsÂ’ with high conductivity are needed for interconnects. Both of these materials are required to be high ly transparent in th e visible range of electromagnetic spectrum. There are two pr oblems for TCOs or TS Os materials with marginal bandgaps for visible light transparency First, the polycrys talline nature of the functional device introduces de fects that extend the optical absorption into the bandgap. This reduces the efficiency of the devices, particular at the blue end of the spectrum where luminescence is weakest. Second, the optical absorption by th e channel region can

PAGE 47

35 lead to band-to-band excitati ons of carriers and subseque nt shifting of transistor characteristics. Table 4-1. Properties of tr ansparent semiconducting oxides. Material Band gap Resistivity (m -cm) Carrier density(cm-3) Mobility (cm2V-1sec-1) Carrier type ZnO 3.35 eV 0.5-100 10171021 200 n ITO 3.6 eV 0.3-20 1018-1021 10-40 n Ag(In0.95Sn0.05)O2 4.4 eV 10-106 1019 0.47 n ZnGaO4 4.4 eV 30-105 … … n CuGaO2 3.6 eV 105 1018 0.46 p (Co,Ni)Ox 3.8 eV 1-106 1015-1020 0.5 p Different channel materials can be considered for transparent TFT development. Although most device applications of interest focus on inexpe nsive glass as the substrate of choice, experiments have al so included epitaxial (on single crystals) channel materials in order to delineate the effects of cr ystallinity and grain boundaries on device performance. A key issue is to identify transp arent materials that are suitable for use as channel materials in thin-film FET structures Relevant factors include carrier density, carrier mobility, crystallinity, surfa ce morphology, optical absorption, and photoconductivity over the visible spectrum. The la tter is relevant to the stability of the TTFTs when coupled to PLED emitters. For TFT operation characteristics, enhancementmode or “normally-off” is preferable to de pletion-mode or “normally-on” due to lower power supply and simpler circuit design. Large off-current and the normally-on characteristics may originate from the fact that channel material s display high carrier densities resulting in a channel that will be conducting in the absence of applied field. Thus, the primary focus for TSOs has been on materials with higher electron mobility and easily controlled carrier densities.

PAGE 48

36 ZnO-based Transparent Thin-film Transistors ZnO has many characteristics that make it attr active to be used as an active channel layer in transparent TFTs. Due to its wide bandgap at room temperature ZnO is highly transparent in the visible spectrum. Thus, Zn O-based TFTs can realize increased aperture ratio of active matrix arrays and overcome th e light sensitivity and degradation issues encountered with Si-based TFTs. Another particular interest in ZnO exists in the fact that good quality polycrystalline ZnO films w ith mobility ranging from 10 to 50 cm2/Vs can be realized at low temperatures (<500) on amorphous glass substrates or plastic/flexible substrates. Th e growth of ZnO thin films has been demonstrated using a number of deposition techniques, including sputtering, pulsed-laser deposition (PLD) and molecular beam epitaxy. In addition, ZnO can be processed by wet chemical etching making the device fabrication processi ng relative simple and low cost. Transparent TFTs based on polyand single crystalline ZnO films as active channel layers have been reported by seve ral groups recently [127-131]. Hoffman et al. [127] fabricated an n-channel, enhancement-mode transparent TFT in which ZnO served as the channel layer with aluminum-titanium oxide (ATO) as the gate oxide, and ITO was used as the source, drain and gate Rapid thermal anneal (RTA) process was used to increase the ZnO channel crystallinity and resistiv ity. The transfer characteristics of TTFTs indicated a maximum drain cu rrent on-to-off ratio of ~107 and the effective channel mobility ranged from 0.3 to 2.5 cm2/Vs. Masuda et al. [128] reported ZnO-based TFTs with a double layer gate oxide consisting of SiO2 and SiNx. The ZnO-TFT fabricated on Si substrate showed enhancement-mode opera tion. The drain current on-to-off ratio was more than 105. The transparent ZnO-TFTs fabricated on glass substrate showed

PAGE 49

37 depletion-mode characteristics. It is considered that the Zn O in the transparent device had rougher surface and more background electr ons than the ZnO in the TFTs on Si substrates. The field-effect mobility FE of the TFT device was 0.031 cm2/Vs which is much smaller than the Hall mobility of Zn O film. However, inserting a high carrier concentration layer between the channel layer and the source/drain contacts can increaseFE to 0.97 cm2/Vs. This indicates that the high carrier concentration layer reduced the contact resistances and had a good effect on the theoretical drain current [128]. In plastic/flexible electronics, the processing temperatures are limited to be less than 100. Although organic semiconductors can be the material choices, their low mobility (<0.1 cm2/Vs ) and instability in ambient conditions impede the further applications for TFTs. TFTs employing ZnO as a channel layer have been realized at room temperature with higher mobility and current on/off ratio. Carcia et al. [129] fabricated ZnO TFTs by rf magnetron sputteri ng on Si substrates near room temperature. The best devices had FE of more than 2 cm2/Vs and an on/off ration > 106. Fortunato et al. [130] also reported high performance ZnO TFTs fabricated at room temperature. The devices operated in the enhancement m ode with a saturation mobility of 27 cm2/Vs and an on/off ratio of 3 x 105. The ZnO films with very high resistivity of about 108 cm were deposited by using rf sputtering for the TFT devices. Highly conductively Ga-doped ZnO was used as the source and drain electrodes. In this dissertation work, the developmen t of ZnO-based TFTs with transparent conducting oxides as the source/dra in is presented. Specifically the fabrication process of these top-gate type TFT devi ces on glass substrate is described. Three types of active

PAGE 50

38 channel materials including undoped ZnO, P-doped ZnO and (Zn,Mg)O films were employed in the TFTs. The primary interest is TFT devices that operate in enhancement mode exhibiting a normally off (gate voltage Vg = 0) channel state. The output and transfer characteristics of the TFT devices using different channe l materials are also discussed. Deposition and Properties of Channel Materials Pulsed laser deposition was used to depos it the ZnO films as the channel layer in TFTs. A KrF excimer laser (=248 nm) was used as the ablation source. A laser repetition rate of 1 Hz was applied, with a laser pulse energy density of 1-3 J/cm2. The dependence of film properties on depo sition conditions for undoped ZnO was investigated for the films grown at 400C in an oxygen pressure ranging from 2mTorr to 300mTorr. Corning 7059 glass substrate (1cm x 1cm) was used as the film growth substrate. The thickness of the films is in the range of 400-500 nm. 2 at.% P-doped ZnO and (Zn0.9Mg0.1)O thin films were also utilized as the channel materials for TFTs. Postgrowth annealing process was carried out in chamber at 600C and 100 Torr oxygen pressure for 1 hr for all the films. Previous results showed that this annealing process for P-doped ZnO and (Zn,Mg)O films can furthe r reduce the native electron density by introducing an acceptor level from phosphorus substitution [91]. The electrical properties (carrier concentr ation, Hall mobility and resistivity) of the undoped polycrystalline ZnO films as a functi on of oxygen pressure are shown in Figure 4-2 [131]. Note that all of the films are de posited on bare glass and post-annealed at 600C in 100 Torr oxygen. For films deposited at P(O2) = 20 mTorr, a Hall mobility of 26 cm2 V-1s-1 was realized. With increasing oxyg en pressure, the n-type carrier

PAGE 51

39 concentration steadily decreases. This be havior reflects suppressed oxygen vacancies and/or Zn interstitials which contribute to the intrinsic electrons in ZnO as growth pressure increases. It is necessary to decrea se the background electron density of the ZnO films as the active channel layer and also desirable to maximize the mobility of the channel layer. Therefore, 400 and 20 mTorr oxygen pressure were chosen to grow ZnO films for TFTs. Crystallinity is an important factor is determining the transport properties of the channel material. One interes ting aspect of ZnO is its strong tendency to maintain uniaxial texture in polycrystalline films de posited on almost any substrate. Figure 4-3 shows an XRD pattern of an undoped ZnO film grown on a glass substrate at PO2=20mTorr. The (002) and (004) ZnO diffrac tion peaks are predominant indicting caxis orientation in this polycrystalline film. However, this uniaxial texture does not eliminate grain boundaries. Figure 4-4 shows an AFM image of the ZnO film growth under the conditions described above. Grain size is on the order of 100-150 nm. In a field-gated structure, the e ffect of grain boundary conductan ce and charge density has to be considered in order to model the field-effect characteristics. Fabrication of ZnO-based TFTs Top-gate type TFTs using ZnO as the act ive channel layer were realized using photolithography followed by wet chemical et ching processing. A schematic cross section view of a top-gate type TFT struct ure is shown in Figure 4-5. Amorphous gate oxides (Ce,Tb)MgAl11O19 (CTMA) and HfO2 were grown by PLD and sputtering, respectively. Tin-doped indium oxide (ITO) se rved as the source and drain electrodes due to its low resistivity (~2 x 10-4 cm) and high transparency in the visible spectrum.

PAGE 52

40 50100150200250300 1014101510161017101810191020 Oxygen pressure(mTorr)Carrier Density(cm-3) Carrier Density (cm-3) 0 5 10 15 20 25 30 Mobility(cm2/vs) Mobility (cm2/vs) 50100150200250300 10-1100101 Resistivity ( cm)Oxygen Pressure(mTorr) Mobility (cm2/Vs) Figure 4-2. Electrical properties of undope d ZnO thin films grown on glass at 400 C as a function of oxygen pressure. Plots showi ng (a) carrier density, mobility, and (b) resistivity of the films. (a) (b)

PAGE 53

41 20304050607080 4x1045x1046x1047x1048x1049x1041x1051x1051x1051x105 (004) (002)Growth temp. : 400oC Oxygen pressure: 20mTorr Film thickness: 400nmIntensity (a.u.)2 (degrees) (002) (004) Figure 4-3. X-ray diffrac tion pattern of the undoped ZnO film deposited at PO2=20mTorr on glass substrate. Figure 4-4. An AFM image of the surface of the undoped ZnO thin film deposited at PO2=20mTorr on glass substrate.

PAGE 54

42 There are several reasons w hy buried-channel structure is chosen: (1) the buried channel device is expected to have highe r mobility because in bulk conduction the surface scattering can be avoided [132]; (2) th e active channel layer is protected from ambient by the gate oxide layer, thus the channel layer can be made thinner in order to reduce the potential parallel re sistance; (3) the source and dr ain electrodes make direct contact with the channel material at the gate oxide/channel in terface eliminating the potential series resistance [133]. Channel Drain Source Gate Oxide GlassSubstrate Gate Figure 4-5. Schematic cross section view of a top-gate-type TFT structure. Standard photolithography process had been used to fabricate the ZnO-based TFTs in this dissertation work. Positive photores ist (Shipley 1813) was chosen due to the requirements of the available photoresist ma sks. Uniform applica tion of photoresist was accomplished by spinning the resist on the sample substrates. The final thickness of the photoresist film depends on the spin rate and spin time. In this work, a uniform photoresist layer with thickness of about 8001000 nm was obtained by setting spinning speed at 5500 RPM for 30-40 sec. Samples w ith different photores ist thicknesses were also investigated in order to obtain the op timal exposure and development conditions for

PAGE 55

43 the best resolution of the patterns. Before being exposed to UV light the photoresist was soft-baked at 90 in air for 45 min. The resist-coated samples were then contac ted with a photoresist mask and exposed to light. Karl Suss mask aligner which has a mercury (Hg) UV lamp was used to expose photoresist and align the drain electrodes with those of source and drain. After the exposure, the sample was treated with a developing solution. The exposure and development time for certain thickness of phot oresist depend on the exposure UV light intensity which may decrease during the exte nded usage of the UV lamp. Therefore, the actual intensity of UV light had to be measur ed to determine and adjust the exposure and development conditions. The photoresist pattern obtained by the development was used as a mask for the etching of the underlying la yers. In the ZnO-based TFTs fabrication scheme, wet chemical etching was utilized for the mesa is olation and the pattern formation of source, drain and gate. Compared to dry etching, wet etching has several adva ntages such as high selectivity, less damage to the underlying mate rial and cheap system setup. During wet etching, the slowest step, called the rate lim iting step, determines the etch rate. Generally, there are two types of rate limiting step in wet etching: diffusion-limited and reactionlimited. In diffusion-limited etching mode, the etch depth has a square root dependence on etch time and the solution ag itation significantly affects the etch rate. The solution has to be agitated in some manner to assist in the movement of etchant to the surface and the removal of the etch product. On the contrary in reaction-limited etching, the etch depth has a linear dependence on etch time and the etch rate is independent of solution agitation.

PAGE 56

44 Since it is desirable to have a reproducible and well controlled etch rate for device fabrication, the reaction-limited etching is mo re preferable to the diffusion-limited mode. In this work, wet etching of the material s employed in the TFT structure has been investigated by using different acid solutions including H2SO4, HCl, HNO3, H3PO4, HF as well as the mixtures of these acids. To realize this top-gate TFT structure high selectivity of gate, gate oxi de and channel layers over unde rlying source/drain layer is required. It was found that among these etchants H3PO4 has the best resolution for the gate metal and the highest selectivity of ga te oxide and channel layers over underlying ITO. For the etching of ITO source and dr ain, HCl acid was found to have the best etching results. Stylus profilometer was used to measure the film thickness in order to calculate the etch rate after the removal of photoresist mask in acetone. In the same time, the etch bias resulting from the isotropic wet etching can be observed and minimized by optimizing the temperature of aci d solution and the etching time. Figure 4-6 schematically illustrates the fabrication processing sequence for the ZnO-based TFT structure. The initial substr ate is commercial ITO-coated display glass substrate. First, the ITO source and drain patterns were defined via using photolithography process, with wet etching of the ITO performed in HCl acid at around 35. The dependence of etch depth on etch time showed a linear relationship, indicating the etching process was reaction-limited. After removing the source/drain photoresist patterns in acetone, ZnO wa s then deposited at 400 in 20 mTorr oxygen as the active channel layer on the ITO patterned glass substr ate. The thickness of channel layer was in the range of 20-50nm. A thin layer of gate oxide (100-200 nm) was deposited on top of the ZnO channel layer. Then aluminum as the gate metal (~100nm) was deposited by

PAGE 57

45 using magnetron rf sputtering at room temper ature. The gate cont act was defined via using lithography alignment followed by selectiv e wet etching gate metal (Al), gate oxide and channel layers with H3PO4 acid down to the source and drain contacts. Aluminum deposited by sputtering ca n be easily etched by H3PO4 at ~60 with a well-controlled profile. Etching of gate oxide and ZnO fi lms without removing the underlying ITO layer as well as the glass substrate was realized in H3PO4 acid at moderate temperatures (3540). Thus, etching this top-ga te TFT structure by using H3PO4 acid with high selectivity made the device fa brication process much simple r and more controllable. Top-view microscopy images of the ZnOTFT devices on glass substrate are shown in Figure 4-7 (a) and (b). Topgate type TFT structures with well-defined contact patterns were realized by using photol ithography and wet etching. Devi ces with different channel widths and lengths were fabri cated as shown in Figure 4-7 (a). For the device shown in Figure 4-7 (b), the channe l length and width are 50m and 90 m, respectively. In order to field gate this structure, it is necessary to form either a Schottky barrier or gate oxide. For ZnO, Schottky barriers ar e low and are apparently unsuitable for field gated rectifying structures. For oxi de gates, the gate dielectric must be chosen to have a band offset with the channel material so as to avoid carrier inject ion into the conduction band and/or valence band of the gate insula tor. For wide bandgap semiconductors, such as ZnO, this necessitates the use of the la rger bandgap oxides and precludes the use of many insulators being considered for field gated structures on other semiconductors. (Ce,Tb)MgAl11O19 (CTMA) with wide bandgap (>5 eV) was used as the gate dielectric for the TFT. Low leakage current about 10-7 A cm-2 was obtained for a 200m diameter Al-CTMA-ZnO capacitor [131].

PAGE 58

46 Glass Substrate ITO ITO Gate metal Gate oxide 7. ZnO-TFT ZnO Photoresist Glass Substrate ITO ITO 2. ITO Wet Etching 1. Source & Drain Lithography ITO Glass Substrate Photoresist 3. Channel & Gate Oxide Deposition ZnO Gate oxide Glass Substrate ITO ITO 4. Gate Contact Metallization Gate metal Glass Substrate ITO ITO ZnO Gate oxide 5. Gate lithography Glass Substrate ITO ITO Gate metal ZnO Gate oxide 6. Wet Etching Gate Contact/Gate oxide/Channel Glass Substrate ITO ITOZnO Gate metal Gate oxide Photoresist Figure 4-6. Schematic fabrication se quence of ZnO-based TFT structure.

PAGE 59

47 Figure 4-7. Top-view micr oscopy image of the ZnO-TFTs on glass substrate. ZnO-based TFT Device Characterization We characterized the output and transfer performance of the TFT devices using an Agilent 4155A Semiconductor Parameter Analy zer at room temperature. The transfer (a) (b)

PAGE 60

48 characteristics of the devices include drain current Id and gate current Ig as a function of gate voltage Vg at a fixed drain voltage Vd, drain current on-and-off ratio (ION/IOFF) and field-effect mobility. When the current carri ers are confined within a narrow channel layer additional scattering mechanisms have to be considered. The location of the carriers at the oxide-semiconductor interface introdu ces additional scattering mechanisms like Coulomb scattering from oxide charges and interface states, as well as surface roughness scattering. Generally, field-effect mob ility can be obtained both from the transconductance value and from the saturation current. In field-effect transistors, transconductance is defined by [134] DV GS DS mV I g (4-1) The field-effect mobility FE is given by DS OX m FEV WC Lg (4-2) Where L is the channel length, W is the channel width, COX is the capacitance of gate oxide, VDS is the source-drain voltage. The tu rn-on voltage need not be known for the determination of FE. Another way to calculate field-effect mobility is fitting straight line to the plots of the square root of drain current vs gate -drain voltage, while the drain current in the saturated region, Idsat is given by [128] 2) ( ) 2 (th GS i FE dV V C L W Isat (4-3) ) (th gs dsV V V Where Ci is the capacitance per unit area of the gate oxide.

PAGE 61

49 ZnO-TFTs Using Undoped ZnO as Active Channel Layer The electrical characteris tics of the TFT device with undoped ZnO as the channel layer are shown in Figure 4-8. For this de vice structure, the ZnO film thickness was 20nm. This TFT device operated as an n-cha nnel, normally-on device, as evident from the fact that there was a source-drain current at the gate voltage of 0V and a negative voltage was required to deplete the carriers in the channel layer. Note that the device drain currents IDS are large, of order of mA, in its “ON” state is due to the high carrier concentration in the undoped ZnO channel la yer. To further modulate and deplete the channel conductance, low channel carrier density is necessary to achieve. 0123456 0 5 10 15 20 25 20 V 15 V 10 V 5 V 0 V Drain Current (mA)Drain Voltage (V) Figure 4-8. Drain current as a function of drain voltage charact eristics for the undoped ZnO-TFT.

PAGE 62

50 ZnO-TFTs Using P-doped ZnO and (Z n,Mg)O as Active Channel Layer For efficient TFT operation, an enhancement-mode device is preferable over depletion-mode, thus avoiding the need to appl y voltage in order to turn the device off. Much less power dissipation is possible when normally-off, enhancement-mode devices are employed. For this motivation, alternative channel materials have been investigated in order to decrease ca rrier density. Phosphorus-doped ZnO and (Zn,Mg)O have been deposited as the active channel layer for the TFT. Post-deposition oxygen annealing processes were used to further decrease the el ectron density in these films. Figures 4-9 (a) and (b) show the output characteristics of devices with P-doped ZnO and (Zn,Mg)O (50 nm for both) as the channel materials, re spectively. For P-doped ZnO based TFT, the device has the same depletion-mode opera tion as the undoped ZnO one. However, the channel conductivity is lower than that of TFT with undoped ZnO as the channel at the gate voltage of 0V. Enhancement-mode operation (Figure 4-9 (b )) was realized for P-doped (Zn,Mg)O based TFTs with HfO2 serving as the gate dielectric s. In these devices conducting channels were induced by applying positiv e gate voltages. Channel length and width were 20m and 90m, respectively. A saturation of IDS (pinch-off) was observed for small values of VDS. This pinch-off beha vior indicates that the channel layer is sufficiently depleted in this TFT. The field-effect (FE) mobility was derived to be about 5.32 cm2 V-1s-1 from the transfer characteristics of the devices operated at 6V shown in Figure 4-10. This value is comparable to that realized in undoped ZnO channels indicating that acceptor doping did not have a detrimental impact on channel mobility. We also determined the carrier density was as low as 3.9x1016 cm-3 for this channel

PAGE 63

51 material, which is two orders lower than that in the undoped ZnO thin film. The on/off current ratio is on the order of 103 at the gate voltage of 10V for these devices. 0246810 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 -7V -1V 0V 1V 2V 3V 4VDrain Current(mA)Drain Voltage (V) 1234567 0 10 20 30 40 0V 1V 2V 3V 4VDrain Current(A)Drain Voltage (V)(b) (a) Figure 4-9. The output characteristics of the TFT with alternative active channel materials: (a) P-doped ZnO as the ch annel; (b) P-doped (Mg,Zn)O as the channel.

PAGE 64

52 0.0 5.0x10-71.0x10-61.5x10-62.0x10-62.5x10-63.0x10-63.5x10-64.0x10-64.5x10-6 -4-3-2-1012345 0.0 2.0x10-74.0x10-76.0x10-78.0x10-71.0x10-61.2x10-61.4x10-61.6x10-61.8x10-6 Vd : 6V Drain current (A)Gate voltage (V) Gate current (A) Figure 4-10. Transfer charac teristics of ZnO-TFT with Pdoped (Zn,Mg)O as the channel layer at the drain voltage of 6V.

PAGE 65

53 CHAPTER 5 GROWTH AND CHARACTERIZATION OF P-TYPE PHOSPHORUS-DOPED (Zn,Mg)O BY PULSED LASER DEPOSITION Introduction One of the critical issues in developing ZnO-based UV LEDs and lasers is to realize low resistivity, high car rier density p-type ZnO mate rial. Phosphorus is a possible acceptor dopant that can be used to synthesize p-type ZnO. Doping of ZnO with Mg provides a means to increase the band gap further into the UV. The motivation for examining phosphorus doping in Mg-doped ZnO is two-fold. First, ptype conductivity in (Zn,Mg)O films will be necessary for LED hete rostructures in which carrier confinement for efficient electron-hole recombination is n eeded. Second, the addition of Mg shifts the conduction band edge to higher energy, perhap s increasing the activ ation energy of the defect donor states. Previous results on annealed phosphorus-doped (Zn,Mg)O device structures, in particul ar C-V and I-V characteristics, in dicate that phosph orus yields an acceptor state and p-type behavior [135]. However, these materials did not yield an unambiguous positive Hall voltage, presumably due to the low mobility and high carrier compensation. In this chapter, an unambiguous positive Hall coefficient in as-grown P-doped (Zn,Mg)O films is presented. The effect of oxygen partial pressu re on the transport properties of P-doped (Zn,Mg)O films gr own on low-temperature (LT) undoped ZnO buffer layers is described. The chemical stat e of phosphorus in p-type (Zn,Mg)O:P films

PAGE 66

54 is presented. The crystallinity and the surface morphology of these films are also discussed. Experimental The phosphorus-doped (Zn0.9Mg0.1)O epitaxial films were grown via pulsed laser deposition (PLD) on c -plane sapphire substrates. The ta rget was fabricated using highpurity ZnO (99.9995%) and MgO (99.998%), mixing with P2O5 (99.998%) as the doping agent. The melting point and boiling point for P2O5 are about 340 and 360, respectively. P2O5 has a lower heat of formation (Hf = -360 Kcal/mole) compared with that of SiO2 (Hf = -202.6 Kcal/mole) [136]. The phos phorus doping level in the target was 2 at.%. A KrF excimer laser (=248 nm) was used as the ablation source. A laser repetition rate of 1 Hz was applied, with a laser pulse energy density of 1-3 J/cm2. The ZnO growth chamber has a base pressure of 10-6 Torr. An undoped ZnO buffer layer (~50nm) was initially deposited at 400 and 20mTorr oxygen partial pressure before the growth of P-doped (Zn0.9Mg0.1)O films. The undoped ZnO buffer layer was post-annealed at 650 in flowing O2 for 1h in order to decrease the electron conductivity. Semiinsulating buffer layers are preferable in order to perform Hall measurements without influence from buffer layer conduction. The phosphorus-doped (Zn0.9Mg0.1)O films were then deposited on the annealed undoped ZnO buf fer layer at a substr ate temperature of 500 under oxygen partial pressure ranging from 20 mTorr to 200 mTorr. The total film thickness ranged from 500 to 700 nm. Four-point Van der Pauw Hall measurements were performed at room temperature in order to examine the transport pr operties of the as-grown P-doped (Zn0.9Mg0.1)O films. The chemical state of phosphorus in the f ilms was examined by using XPS with an Al

PAGE 67

55 anode (photon energy = 1486.6 eV). The film crystallinity and surface morphology of the P-doped (Zn0.9Mg0.1)O films as a function of oxyge n growth pressure were also investigated by using XRD and AFM, respectively. Results and Discussion The large lattice and thermal mismatch between ZnO and sapphire substrate will generate considerable stress in the epitaxial film, affecting the growth and quality of the film. To optimize the crystallinity and propert ies of the epitaxial films, a well-created buffer layer is necessary. Kaidashev et al. [137] reported that by inserting a thin ZnO relaxation layer grown at lower temperat ure between sapphire substrate and hightemperature ZnO film the crystallinity and mobility of the films have been improved. In this work, a thin layer of LT-ZnO was grown at 400 on sapphire substrate before the growth of P-doped ZnMgO film. The effect of the LT-ZnO buffer layer on crystallinity of P-doped (Zn0.9Mg0.1)O films were investigated by m easuring the omega rocking curve for ZnO (0002) plane. Figure 5-1 shows the Zn O (0002) omega rocking curves of the Pdoped (Zn0.9Mg0.1)O sample with and without a LT-Zn O buffer layer. These films were grown at 500and under 20 mTorr oxygen pressure. An increase in diffraction intensity is clearly shown for the film grown with LT-ZnO buffer layer. The full-widths at half maximum (FWHM) values of the omega rocking curve are 0.83 and 1.02 for the films grown with and without buffer layer, respectively, indicati ng that the crystallinity was improved by introducing a thin LT-ZnO buffer la yer. Thus, by applying a thin layer of LT-ZnO buffer layer, the properties of the P-doped ZnMgO films were expected to be optimized.

PAGE 68

56 For many samples, a persistent n-ty pe photoconductivity was observed that complicated the transport characterization. To eliminate the photoconductivity relaxation effect on the transport properties, the sample s were maintained in the dark for 12 hours before performing Hall measurements at room temperature. Using this procedure, the carrier density and conduction type of as-grown P-doped (Zn0.9Mg0.1)O films was determined. These properties as a function of oxygen growth pre ssure are shown in Figure 5-2. Each sample was measured a mini mum of twenty times to obtain the average results shown here. The error bars represen t the maximum deviation from the average values. Note that films deposited at the oxygen partial pressure lower than 100 mTorr show n-type conductivity with electr on concentration in the range of 1016-1017 cm-3. However, as oxygen pressure increases above 100 mTorr, the el ectron concentration continuously decreases and the samples star ted showing indeterminate carrier type indicating the coexisting of electro ns and holes in the films. Unambiguous conduction type was not obser ved for the samples grown at 100 and 120 mTorr due to near-equivalent concentrations of holes a nd electrons in the films. When the oxygen partial pressure was increa sed to 150 mTorr during the film deposition, the Hall-effect data showed consistent p-type carrier type with a hol e concentration of 2.716cm-3. Also note that as the oxygen part ial pressure was increased up to 200 mTorr, the films reverted to an indetermin ate carrier type. Figure 5-3 shows the carrier mobility carrier type films grown under 100, 120 and 200 mTorr oxygen pressure yield large of Pdoped (Zn0.9Mg0.1)O films as a function of oxygen pressure. The indeterminate standard deviations compared with those with uni polar conduction type films. These results imply that for the sample s which may contain both types of carriers,

PAGE 69

57 more attention needs to be paid for the Hall measurements and data analysis. The p-type films grown at 500 and 150 mTorr oxygen show an average hole mobility of 8.2 cm2/Vs at room temperature. This mobility value is reasonable compared with the ones reported in the previous studies on N, Al-N and P-doped p-type ZnO. 1416182022 2.0x1044.0x1046.0x1048.0x1041.0x1051.2x1051.4x105 Intensity (a.u.) (degree) no buffer layer (FWHM=1.02) with buffer layer (FWHM=0.83) Figure 5-1. ZnO (0002) omega rocking curves of P-doped (Zn0.9Mg0.1)O samples with and without LT-ZnO buffer layer. The resistivity of the P-doped (Zn0.9Mg0.1)O films grown under different oxygen pressures are also shown in Figure 5-4. The re sistivity of the films increases from 1.0395 -cm to 350.5 -cm as oxygen growth pressure incr eases from 20 to 100 mTorr. The increase in resistivity for the films grown under oxygen pressure between 20 to 100 mTorr results from the decreased carrier con centration in the films. The films grown at 100 mTorr exhibit the highest resistivity of about 300 -cm. With further increasing oxygen pressure, resistivity rapi dly decreases due to the in creased hole carrier conduction

PAGE 70

58 in the films. For the p-type P-doped (Zn0.9Mg0.1)O films grown at 150 mTorr oxygen pressure, the resistivity is about 35 -cm. 20406080100120140160180200 1E15 1E16 1E17 1E18 Carrier Concentration (cm-3)Oxygen Partial Pressure (mTorr) p-type n-type indeterminate carrier type Figure 5-2. Carrier concentrati on and carrier type in P-doped (Zn0.9Mg0.1)O films as a function of oxygen partial pressure. 20406080100120140160180200 -12 -10 -8 -6 -4 -2 0 2 4 6 8 10 12 10 10 5 0 5n-type p-type Carrier Mobility (cm2/Vs)Oxygen Partial Pressure (mTorr) Figure 5-3. Effect of oxygen partial pressure on carrier mobility of P-doped (Zn0.9Mg0.1)O films.

PAGE 71

59 20406080100120140160180200 1 10 100 1000 Oxygen Partial Pressure (mTorr) p-type n-type indeterminate carrier typeResistivity (ohm cm) Figure 5-4. Resistivity of P-doped (Zn0.9Mg0.1)O films vs oxygen partial pressure. The Hall effect results show that p-type doping in P-doped (Zn0.9Mg0.1)O films is strongly dependent on the oxidati on conditions. It is important to note that the effect of oxygen partial pressure on ZnO p-type conducti on has been investigated previously by both experiment and theory. Xiong et al. [138] reported evidence for p-type conduction in undoped ZnO films grown at high oxygen pa rtial pressure by reactive sputtering. A change in the sign of charge carriers from electrons to holes was identified around 55% oxygen in Ar/O2 mixture. ZnO-based p-n homojunction was also formed by controlling oxygen partial pressure during sputtering. This realization of p-t ype conductivity in undoped ZnO was consistent with the effect of higher chemical potential of atomic oxygen on defect formation enthalpies. Th e increased oxygen chemical potential by electronic excitation to a dissoci ated state raises the forma tion enthalpy of the intrinsic donor VO and lowers the formation energy of the acceptor Oi..

PAGE 72

60 Zunger also suggested several theoretical practical rules for ptype doping of wide bandgap materials to overcome the doping bottle necks [76]. First, the p-type doping is facilitated by alloying an element that leads to upward bowing of the VBM. One way to shift the VBM upwards is to add a tetrahedrally bonded 3 d element with active d states. In addition, Zunger suggests that limitations to p-t ype doping can be overcome by manipulating the growth conditions, e.g. the us e of the host anion-rich growth conditions to inhibit the formation of so-called “hole ki ller” defects. Calculations from chemical potentials suggest that the enthalpy of fo rming anion vacancies decreases under cationrich (zinc-rich) conditions. In the present st udy, the growth condition necessary to obtain p-type P-doped (Zn0.9Mg0.1)O films is rather narrow (150 mTorr oxygen partial pressure). For the samples grown in the oxygen pressu res of 100, 120, and 200 mTorr, the Halleffect results do not show unambiguous p-type conductivity. The observation of indeterminate carrier type for a growth pr essure of 200 mTorr ma y be explained by the “host anion poor” rule for the anion-substituting p-type dopa nts which conjectures that anion poor conditions are more favorable for a high solubility of acceptors on anion sites [76]. Consequently, the host anion condition ha s to be optimized in order to reach an equilibrium state under which p-type doping of ZnO can be realized. In order to confirm th e incorporation of P2O5 as a P doping source in the p-type films, X-ray photoelectron spect roscopy was performed to examine the chemical states of phosphorus using Al anode as the X-ray sour ce. Figure 5-5 shows the XPS survey of Pdoped ZnMg0.1O films grown at 500, 150 mTorr oxygen pressure. Zn 2 p Mg 1 s and O 1 s peaks as well as their Auger peaks are s hown in the spectrum. The multiplex of P 2 s is shown is Figure 5-6. Only one peak with the binding energy of 192.2 eV is observed

PAGE 73

61 from the spectrum, which is consistent with P 2 s binding energy of 192.8 eV in P2O5. The P 2 s peaks regarding to Zn3P2 and element phosphorus state have the binding energy values of 186.3 eV and 187.7 eV, respectively [139]. X-ray diffraction was used to examine the crystallinity of the P-doped (Zn0.9Mg0.1)O films grown under different oxygen part ial pressures as shown in Figure 57. The diffraction data shows only ZnO ( 0002) and (0004) and sapphire (0006) peaks indicating that the films are oriented only with c -axis perpendicular to sapphire substrates. Thus, the out-of-plane or ientation is ZnO [0001]sapphire [0001]. No impurity phases were observed from the XRD results. This sugge sts that the solid solubility of phosphorus and magnesium has not been exceeded in the films under these growth conditions. Also note that there is no discernable change in the c-axis lattice parameter with increasing oxygen partial pressure. As discu ssed earlier in this chapter, the crystallin ity of the Pdoped (Zn0.9Mg0.1)O films grown on LT-ZnO buffer laye r was also characterized by fourcircle XRD. The FWHM of the omega rock ing curve for ZnO (0002) peak is 0.83. The scan of P-doped ZnMgO film grown on sapphire substrate was used to examine the epitaxy of the film and determine the in-plane orientation relationship between the film and substr ate. Figure 5-8 shows the scans through ZnO (10 11) plane and sapphire (11 26) plane of the sample grown at 500, 20 mTorr oxygen pressure. A sixfold symmetry of the plane is show n indicating good epitaxy of the P-doped ZnMgO film. More interestingly, it is shown that P-doped ZnMgO f ilm perfectly aligns with sapphire substrate. No 30 tw isted orientation was observe d in these films grown at 500. Therefore, the in-plane orientation re lationship between the film and sapphire substrate is determined as ZnO [10 10]sapphire [10 10].

PAGE 74

62 Binding Energy (eV) N(E) Min: 2780Max: 900767 13501215 1080 945 810 675 540 405 270 135 0 Zn 3p Zn 3s C 1s Mg KLL Zn LMM O 1s O KLL Zn 2p3 Zn 2s Mg 1s Figure 5-5 X-ray photoele ctron spectroscopy survey of P-doped (Zn0.9Mg0.1)O films. Binding Energy (eV) N(E) Min: 1100Max: 1251 200198 196 194 192 190 188 186 184 182 180 Figure 5-6 X-ray photoelectron spectroscopy multiplex of P 2 s peak for P-doped (Zn0.9Mg0.1)O films grown at 500, 150 mTorr oxygen pressure. P 2s

PAGE 75

63 30354045505560657075 0.0 2.0x1054.0x1056.0x1058.0x1051.0x1061.2x106 PO2=20mTorr PO2=100mTorr PO2=150mTorr PO2=200mTorrZnO (0004) Al2O3 (0006) ZnO (0002)Intensity (a.u.)2 thelta (degree) Figure 5-7. X-ray diffr action of P-doped (Zn0.9Mg0.1)O films grown under different oxygen partial pressures. -200-150-100-50050100150 100 1000 Intensity (a.u.)Phi (degree) sapphire substrate P-doped ZnMgO film Figure 5-8. XRD scans of P-doped (Zn0.9Mg0.1)O film and sapphire substrate.

PAGE 76

64 The surface morphology of the films grown at 500 in various oxygen pressures is shown in Figure 5-8. The scan area is 22 m2 with a scan rate of 1 Hz. As the oxygen pressure increased, both grain size and su rface roughness increased. Figure 5-9 shows the Root-Mean-Square (RMS) roughness of the P-doped (Zn0.9Mg0.1)O films as a function of oxygen partial pressure. The RMS roughness for the films increased from 2.60 nm to 12.8 nm as the oxygen growth pressure increas ed from 20 mTorr to 200 mTorr. For the ptype (Zn0.9Mg0.1)O films deposited at 150 mTorr oxygen pressure, the presence of grain boundaries can contribute to the low carrier mobility. 1 m 0.5 1.5 15nm 15nm 15nm 15nm 1 m 0.5 1.5 1 m 0.5 1.5 1 m 0.5 1.5(a) (b) (c) (d) 1 m 0.5 1.5 15nm 15nm 15nm 15nm 1 m 0.5 1.5 1 m 0.5 1.5 1 m 0.5 1.5 1 m 0.5 1.5 1 m 0.5 1.5 15nm 15nm 15nm 15nm 1 m 0.5 1.5 1 m 0.5 1.5 1 m 0.5 1.5 1 m 0.5 1.5 1 m 0.5 1.5 1 m 0.5 1.5(a) (b) (c) (d) Figure 5-9. AFM images of the P-doped (Zn0.9Mg0.1)O films grown at different oxygen pressures: (a) 20; (b) 100; (c) 150; (d) 200mTorr.

PAGE 77

65 020406080100120140160180200220 2 4 6 8 10 12 14 16 RMS roughness (nm)Oxygen partial pressure (mTorr) Figure 5-10. RMS roughness of the P-doped (Zn0.9Mg0.1)O films as a function of oxygen partial pressure. In summary, p-type phosphorus-doped (Zn0.9Mg0.1)O films have been realized via pulsed laser deposition without post-annealing process. The conduction type of the films strongly depends on the oxygen partial pressure during the deposition process. The films grown at oxygen pressure lowe r than 100 mTorr are n-type. However, at oxygen pressure of 150 mTorr, the films showed p-type carrier type conduction with a hole concentration of 2.716cm-3, a mobility of 8.2 cm2/Vs and a resistivity of 35 -cm. XPS measurements confirmed the existence of P2O5 in the p-type P-doped ZnMgO film. XRD results showed good crystallinity of P-dope d ZnMgO films grown under different oxygen pressures. The in-plane and out-of-plane orie ntation relationships are determined as ZnO [10 10]sapphire [10 10] and ZnO [0001]sapphire [0001]. The RMS roughness for the films increased from 2.60 nm to 12.8 nm as the oxygen growth pressure increased

PAGE 78

66 from 20 mTorr to 200 mTorr. The presence of grain boundaries can contribute to the low carrier mobility of p-type P-doped ZnMgO films.

PAGE 79

67 CHAPTER 6 SYNTHESIS AND CHARACTERIZATION OF (Zn,Mg)O:P/ZnO HETEROSTRUCTURES AND AL-DOPED ZnO Introduction Several studies have been reported rega rding ZnO-based p-n junctions for LED applications. Alivov et al.[140,141] reporte d LEDs from n-ZnO/p-AlGaN and n-ZnO/pGaN heterostructures grown expitaxially on SiC substrates usi ng hybrid vapor-phrase epitaxy combined with chemical vapor deposition. The UV LEDs emitted UV light at 389 nm and 430 nm at room temperature, resp ectively. Osinsky et al. [142] also have reported electroluminescence (EL) from p -(Al)GaN/n-ZnO junctions. Tsukazaki et al. [143] obtained violet EL from ZnO homoj unction grown on lattice matched ScAlMgO4 substrates. Hwang et al. also reported on the diode and emission characteristics for a heterostructure of p-ZnO/n-Ga N fabricated via RF magnetic sputtering [144]. We have previously reported ZnO-based p-n junctions deposited on undoped ZnO substrates using ZnMgO:P/ZnO heterostructure system [145]. The use of a ZnO buffer on the lightly ntype ZnO substrate was critical in achievi ng acceptable re ctification in the junctions. Without this buffer, the junctions showed high leakage current. In this prior work, p-type conductivity was only obtained by post-growt h annealing of the P-doped ZnMgO. Studies in the previous ch apters show that oxygen partial pressure plays a significant role in conver ting n-type to p-type conductivity for 2 at.% P-doped Zn0.9Mg0.1O films. The P-doped (Zn,Mg)O film s grown at 150 mTorr oxygen partial pressure exhibited p-type conductance without po st-growth annealing. In this chapter, the

PAGE 80

68 development of ZnMgO:P/ZnO heterostructures on both sapphire and single crystal ZnO substrates is described. No post-growth annealing proce ss was carried out for these structures. The characteristics of nand p-side ohmic contact and the ZnMgO:P/ZnO heterojunction are presented. In addition, to reduce the presence of series resistance such as the current spread resistance in the n-t ype ZnO layer, 0.01 at. % Al-doped ZnO films were grown on sapphire substrates with Mg O buffer layer. Three different growth conditions including growth temperature, oxyge n partial pressure and laser energy were examined to provide a systematic study of growth condition effect on the properties of the Al-doped ZnO films. The dependence of crystallinity, elect rical properties, photoluminescence and surface morphology on thes e growth conditions is discussed. Experimental Schematic diagrams of the p-ZnMgO/nZnO heterostructure on both (0001) csapphire and ZnO substrate are shown in Figur e 6-1. The (0001) undoped grade I quality, single crystal ZnO substrate is obtained from Cermet. The room temperature electron concentration and mobility were 1017 cm-3 and 190 cm2/Vs, respectively. Pulsed laser deposition was used for film gr owth. The 2 at.% phosphorus-doped Zn0.9Mg0.1O target was fabricated using high -purity ZnO (99.995%) with or without MgO (99.998%), mixing with P2O5 (99.998%) as the doping agent. Sa pphire and ZnO substrates were ultrasonically cleaned with trichloroethylen e (TCE), acetone and methanol for 5 min and dried in N2 prior to loading into the growth chambe r. The growth chamber base pressure was 1-2 x 10-7 Torr. A KrF excimer laser with a wavelength of 248nm was used as the ablation source. A laser repetition rate of 1Hz was used, with a target to substrate distance of 4cm and a laser pul se energy density of 1-3 J/cm2. The n-ZnO layer 0.6 m thick with an electron concentration of 2.53x1018 cm-3 and mobility 36.55 cm2/Vs was

PAGE 81

69 grown first at 800 in an oxygen pressure of 100 mTorr, followed by a 0.4m thick pZnMgO:P layer grown at 500, in 150 mTorr O2. Electron-bean evaporated Au (100nm) and Ti/Au (20nm/80nm) were deposited on the p-ZnMgO layer and n-ZnO patterned by li ft-off process. In order to improve the ohmic characteristics, the post-growth annealing at 500 and 450 in N2 for 2 min were performed, respectively. The I-V characteristi cs were measured using an Agilent 4145B parameter analyzer at room temperature. A MgO buffer layer (~200nm) was initially deposited at 450 and 10-4 mTorr oxygen pressure before the growth of 0.01 at.% Al-doped ZnO films on sapphire substrates. The 0.01 at. % Al-doped ZnO targ et was fabricated us ing high-purity ZnO (99.995%) mixing with Al2O3 (99.998%) as the doping agen t. The purpose of the MgO buffer layer is to reduce the micro-cracks resu lting from the thermal expansion difference between ZnMgO and sapphire substrate. Table 6-1 shows the different growth conditions of the Al-doped ZnO via PLD. The film thic kness of Al-doped ZnO is in the range of 0.75-1m. The transport properties of the as-gro wn films were determined using fourpoint Van der Pauw Hall measurements at room temperature. The photoluminescence properties of the films were also measured at room temperature using a He-Cd laser (325nm). The film crystallinity and surface mo rphology were investigated via using fourcircle X-ray Diffraction (XRD) and atomic force microscopy (AFM). Results and Discussion The current-voltage characteris tics of the metal contacts to n -ZnO and p -ZnMgO were measured to examine the formation of the ohmic contacts. Previous results show that Au and Ti/Au can be used as ohmic contacts to p -ZnMgO and n -ZnO, respectively.

PAGE 82

70 Low specific contact resistance wa s obtained for Au contact to p -ZnMgO after postgrowth annealing process [146]. Figure 6-2 (a) and (b) show the I-V cu rves of the Au and Ti/Au metal contacts between two square pads (500 x 500 m) at room temperature. The I-V characteristics indicate that ohmic cont acts are formed on both electrodes. These results show that the rectifying behavi ors shown in Figure 6-3 are due to the heterojunction of the ZnMgO/ZnO structure a nd not to the semiconductor/metal contacts. ZnO or Sapphire substrate p -ZnMgO:P0.02 (400nm) n -ZnO(600nm) Au (100nm) Ti/Au(20/80nm) ZnO or Sapphire substrate p -ZnMgO:P0.02 (400nm) n -ZnO(600nm) Au (100nm) Ti/Au(20/80nm) Figure 6-1. Schematic illustration of ZnMgO:P/ZnO heterostructure. Table 6-1. Growth conditions of 0.01 at. % Al-doped ZnO films via PLD. Sample Tg () Oxygen pressure (mTorr) Laser energy (mJ) Buffer layer A 700 20 350 B 800 20 350 C 700 20 300 D 800 20 300 E 800 5 300 F 800 50 300 MgO The I-V characteristics of the ZnMgO:P/Zn O heterostructure fabricated on sapphire and ZnO substrates are shown in Figure 6-3 (a) and (b), resp ectively. The devices exhibit clear rectifying electri cal characteristics for both struct ures. Note that for the device grown on sapphire substrate, the drain current is higher than that of the device grown on ZnO substrate. This result might result from the difference in the resistivity of the

PAGE 83

71 epitaxial layers grown on different substrates. The turn-on voltage VT can be obtained from the intercept of the linear fitting in the forward bias range with the voltage-axis. For the structure grown on sapphire substrate VT was determined to be about 1.36 V. For the ZnMgO:P/ZnO heterostructure grown on singl e crystal ZnO substrate, the turn-on voltages are 1.15 V and 2.26 V for the lateral an d vertical device structure, respectively. Similar results on the I-V characteristics of oxi de-based p-n junctions have been reported previously [147-149]. According to the current-voltage characte ristics of the real diodes equations [150], 1 exp0 inV V I I (6-1) I d dV kT q n ln (6-2) where the pre-exponential factor I0 is the reverse saturation current, V is the voltage at the junction, Vi = kT/q is the thermal voltage, k is th e Boltzmann constant, T is the absolute temperature, and n is the junction ideality f actor. From equation (6-2), the ideality factors of the ZnMgO/ZnO heterostructures on sapphire and ZnO substrate can be extracted to be about 7.6 and 11.8, respectively. These larg e ideality factors possibly result from the defect-level assisted tunneling [151] and carri er recombination in the space-charge region via a deep level near midgap in the ZnMgO. Further work should focus on increasing the hole carrier concentration and mobility of p -ZnMgO and the optimization of structure synthesis in order to improve ZnMgO/ZnO p-n junction characteristics. Four-circle X-ray diffraction was used to examine the crystallinity of the Al-doped ZnO samples (1cm x 1cm) grown on sapphire substrates. Figure 6-4 shows the XRD 2 theta scans of the as-grown 0.01at.%A l-doped ZnO films grown under different conditions, suggesting that the all the ZnO:Al films are orie nted with the (0001) c-axis

PAGE 84

72 uniformly parallel to the surface normal. Little change in diffraction intensity and lattice spacing was observed for the films grown under di fferent conditions. To further delineate -1.0-0.50.00.51.0 -1.0x10-3-8.0x10-4-6.0x10-4-4.0x10-4-2.0x10-40.0 2.0x10-44.0x10-46.0x10-48.0x10-41.0x10-3 Current (A)Bias (V)Ti/Au (20nm/80nm) to n-ZnO After RTA annealing at 450oC-1.0-0.50.00.51.0 -1.0x10-4-8.0x10-5-6.0x10-5-4.0x10-5-2.0x10-50.0 2.0x10-54.0x10-56.0x10-58.0x10-5 Current (A)Bias (V)Au contact (100nm) to p-ZnMgO after RTA annealing at 500oC Figure 6-2. The I-V curve of Au and Ti/A u metal contacts on: (a) p-ZnMgO; (b) n-ZnO films, respectively. (a) (b)

PAGE 85

73 -2-10123 -0.005 0.000 0.005 0.010 0.015 0.020 0.025 0.030 0.035 0.040 0.045 Current (A)Bias (V) p-ZnMgO/n-ZnO/sapphire -6-5-4-3-2-10123456 0.0 1.0x10-42.0x10-43.0x10-44.0x10-45.0x10-46.0x10-4 p-ZnMgO/n-ZnO/ bulk ZnO Lateral VerticalCurrent (A)Bias (V) ( a ) ( b ) Figure 6-3: Current-voltage characteristics of the ZnMgO/ZnO heterostructure on: (a) sapphire; (b) ZnO substrate.

PAGE 86

74 the crystallinity of the film s grown under different conditions, the omega rocking curve through the (0002) plane of Zn O was investigated. Figure 6-5 shows the omega rocking curve of ZnO:Al films grown under 300 mJ laser energy. The full-width-at-halfmaximum (FWHM) values of the Al-doped ZnO films grown under different conditions are listed in Table 6-2. It is shown that for the films gr own under 350 mJ laser energy, with increasing growth temperature the FWHM decreases from 0.37 to 0.33 indicating an improved crystallinity of the films. The film grown at 800, 50 mTorr oxygen pressure has the narrowest omega rocking curve with the FWHM of about 0.26. The correlations of crystallinity to the electrical and optical properties of the films will be discussed later. Table 6-2. FWHM values of ZnO (0002) om ega rocking curve for 0.01 at. % Al-doped ZnO. Sample A B C D E F FWHM (degree) 0.3703 0.3308 0.2791 0.6366 0.3889 0.2636 Electrical properties of the ZnO:Al films wa s investigated via using four-point Van der Pauw Hall measurement at room temperature. Table 6-3 shows the Hall measurement results of the as-grown Al-doped ZnO films. The growth conditions of sample A-F are described in Table 6-1 previously. All the f ilms show n-type conductivity with resistivity in the range of 10-3-10-1 ohm-cm. The electron concentration and mobility for these films are in the ranges of 1018-1019 cm-3 and 40-60 cm2/Vs, respectively. The effect of different growth conditions on the electr ical properties of the films was delineated by analyzing these Hall data. As growth temperature increases, resistivity decreases for the samples grown under the same laser energy. The decrease of resistivity results from the increase in carrier concentration, mobility or both. In ad dition, the resistivity exhibits an increase

PAGE 87

75 30405060700 200000 400000 600000 800000 1000000 1200000 1400000 1600000 1800000 2000000 F E D C B sample AZnO (004) Al2O3(006) MgO (111) ZnO (002)Intensity (a.u.)2 (degree) Figure 6-4. X-ray diffraction of 0.01 at.% Al-doped ZnO films grown under different conditions. 15.516.016.517.017.518.018.5 0.0 2.0x1054.0x1056.0x1058.0x1051.0x1061.2x1061.4x106 Sample E Sample C Sample D Sample F Intensity (a.u.)Omega (degree) Figure 6-5. Omega rocking curve of ZnO ( 0002) peak for 0.01 at.% Al-doped ZnO films grown under 300 mJ laser energy.

PAGE 88

76 with oxygen growth pressure for the Al-dope d ZnO films grown under same laser energy and growth temperature. Several experiment al studies reported similar dependence of resistivity on growth temp erature and oxygen pressure fo r Al-doped ZnO films [152-154]. Kim et al [153] suggested that the resistivity of Al-doped ZnO films is related to the Al doping concentration, oxygen vacancies, Al and Zn concentrations at interstitial sites, grain boundaries and ionized impurity scattering. An increase in O/Zn ratio was observed with increasing growth temperature by usi ng Rutherford Backscattering Spectrometry (RBS) [153]. Regarding to laser energy eff ect on the electrical pr operties of Al-doped ZnO films, it is shown that with increasi ng laser energy the resi stivity of the films decreases for samples grown at 700 (sample C vs. sample A). However, for the films grown at 800, resistivity does not show large incr ease with laser energy. Considering the donor contributions in Al-doped ZnO discussed above the electron density is expected to increase with laser energy when the ZnO:Al target was bombarded by laser beam to create more interstitial atoms or vacancies. The number of scattering centers affecting the mobility of carriers was also increased with increasing laser energy and deposition rate [155]. Table 6-3. Room temperature Hall measurement of 0.01 at.% Al-doped ZnO films under different growth conditions. Sample Resistivity (ohm-cm) Carrier density (cm-3 ) Carrier mobility(cm2/Vs) Carrier type A 1.2376 x 10-2 7.82 x 1018 62.73 n B 1.3656 x10-2 1.158 x 1019 39.48 n C 1.142 x 10-1 1.399 x 1018 39.12 n D 9.375 x 10-3 1.32 x 1019 50.45 n E 8.336 x 10-3 1.329 x 1019 56.36 n F 1.1285 x 10-2 9.094 x 1018 60.85 n Room temperature PL (RT-PL) was performe d to further investigate the effect of growth conditions on the optical properties of the Al-doped ZnO film s. Figure 6-6 (a) and

PAGE 89

77 (b) show the laser energy on RT-PL fo r the Al-doped ZnO films grown at 700 and 800, respectively. The Al-doped ZnO films e xhibit the band edge photoluminescence at ~377 nm with very low deep level emission. As laser energy increases from 300mJ to 350 mJ, the band edge emission increases for th e films grown at both temperatures. Also note that with decreasing growth temperature from 800 to 700 the deposition laser energy shows more prominent effect on the band edge emission, which is consistent with the laser energy dependence of resistivity discus sed in the previous section. In addition, as shown in Figure 6-6 (a) a nd (b) the band edge emission increases significantly with growth temperature. These results suggest that the photolumines cence of the Al-doped ZnO films has strong correlations to the electrical properties a nd crystallinity of the films. It is known that there are two recombinati on processes, i.e. ra diative and nonradiative transition determining the light emission inte nsity. The photoluminescence efficiency can be enhanced by increasing the radiative transition and decreasing the nonradiative transition. For the Al-doped ZnO films, the intensity of band edge emission increases as the resistivity decreases due to the increased laser energy and growth temperature. With increasing electron density, the Fermi leve l will move up toward the conduction-band edge resulting in more mid-gap defect states being filled up. Thus, the possibility of the non-radiative trapping through those defect stat es will be decreased. For the films grown at a higher temperature, th e nonradiative defects can also be reduced by improving the crystallinity of the films. Accordingly, the intensity of the band edge emission is significantly enhanced by increasing growth temperature for Al-doped ZnO films. The oxygen partial pressure effect on the photoluminescence of the ZnO:Al films is shown in Figure 6-7. Note that the band e dge emission intensity does not have linear

PAGE 90

78 dependence on the oxygen partial pressure as resistivity does. As the oxygen pressure decreases from 5 mTorr to 20 mTorr, the band edge emission decreases. However, with further increasing oxygen pressure to 50 mT orr, the band edge emission markedly increases. The film grown at 800, 50 mTorr oxygen pressure shows the highest band edge emission of all the films. The above an alysis is consistent with the XRD results discussed earlier. With incr easing growth temperature, the FWHM value of the omega rocking curve of the films decr eases. The film grown at 800, 50 mTorr oxygen pressure has the narrowest rocking curve sugge sting that crystallinity plays an important role in the photoluminescence properties of the films. The surface morphology of the films grown in different growth conditions were also examined by performing AFM measurem ent in air. Figure 6-8 shows the AFM images of the Al-doped ZnO films grow n under different conditions. The growth conditions of sample A-F are described in Table 6-1 previously. The scan area is 5 m2 and the scan rate is 1 Hz. Interestingl y note that growth temperature and oxygen pressure greatly affect the surface morphol ogy of the films. The films grown at 700 have the smoothest surface with the root-m ean-square (RMS) roughne ss in the range of 8-9 nm. However, as growth temperature increases to 800, roughness rapidly increases. In addition, with increasing oxygen pressure the surface of the films also becomes rougher. The sample grown at 800, 50 mTorr and 300 mJ lase r energy has the highest RMS roughness about 38nm. Similar growth temperature and oxygen pressure dependence of surface morphology was also observed for the P-doped ZnO and ZnMgO films.

PAGE 91

79 350400450500550600650700 0.05 0.10 0.15 0.20 0.25 Tg=700oC, PO2=20mTorr 300 mJ 350 mJ Intensity (a.u.)Wavelength (nm) 350400450500550600650700 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 300 mJ 350 mJ Intensity (a.u.)Wavelength (nm) Tg=800oC, PO2=20mTorr Figure 6-6. Laser energy e ffect on RT-PL for 0.01 at.% Al-doped ZnO films grown at (a) 700; (b) 800. (a) (b)

PAGE 92

80 350400450500550600650700 1 2 3 4 5 6 7 20 mTorr 5 mTorr 50 mTorr Intensity (a.u.)Wavelength (nm) Laser energy=300 mJ Tg=800oC Figure 6-7. Oxygen partial pr essure effect on room temper ature PL of 0.01 at.% Aldoped ZnO films.

PAGE 93

81 Figure 6-8. AFM image of 0.01 at.% Al-doped ZnO films grown under different conditions. The z-scale is 40 nm/div. Sample A RMS: 9.043nm Sample B RMS: 17.224nm Sample C RMS: 7.952n m Sample D RMS: 20.346n m Sample E RMS: 13.553nm Sample F RMS: 37.837nm

PAGE 94

82 CHAPTER 7 GROWTH AMBIENT AND ANNEALING ST UDY OF PHOSPHORUS-DOPED ZnO Introduction Systematic study of the effects of growth condition and post-annealing process is essential to control and optimize the prope rties of P-doped ZnO films. It has been suggested that Zn interstitia ls, O vacancies, and/or hydrogen complexes as compensation donors in p-type ZnO. Thus, studies need to include understanding the role of oxidizing species in yielding low native defect th in-film materials. The background impurity density of p-type ZnO during growth also n eeds to be minimized so as to observe the presence of acceptors in transport measurements. Previous studies [91] have focused on the effects of annealing on the transport pr operties of 1-5 at.% P-doped ZnO films grown by PLD. It showed that annealing significan tly reduced the carrier density, yielding semiinsulating behavior which is consistent with activation of a deep acceptor level. However, no detailed studies have been carried out on the growth ambient effect on the properties of P-doped ZnO films as well as the annealing effect on these films. In this chapter, the effects of different oxidation ambients, growth temperature and post-growth annealing on the properties of P-doped ZnO have been examined. The transport, photoluminescence as well as surface morphology properties of the films were discussed to elucidate these effects. Experimental Pulsed laser deposition was used to deposit phosphorus-doped ZnO epitaxial films on c-plane sapphire substrates. The targets were fabricated us ing high-purity ZnO

PAGE 95

83 (99.995%) mixing with P2O5 (99.998%) as the doping agent. The targets were pressed and sintered at 1000 for 12h in air. Considering the solubility limit of phosphorus in ZnO, the phosphorus doping level in ZnO:P ta rget was chosen to be 0.2 at.%. A KrF excimer laser with a wavelength of 248nm was used as the ablation source. A laser repetition rate of 1Hz was used, with a targ et to substrate distance of 4cm and a laser pulse energy density of 1-3 J/cm2. Sapphire substrates were ul trasonically cleaned with trichloroethylene (TCE), acetone and methanol for 5 min and dried in N2 prior to loading into the growth chamber. The growth chamber base pressure was 2 x 10-7 Torr. Film thickness was in the range of 400-500nm. P-doped ZnO films were deposited under different oxidizing conditions including oxygen and 4% H2/Ar mixture, pure oxygen and ozone/oxygen mixture. The partial pressure ratio of oxygen and Ar/H2 in the gas mixture was 1:1. The nitrogen-free plasma disc harge ozone generator yielded an O3/O2 ratio on the order of 1–3%. The same total growth pr essure was maintained at 60mTorr for the different ambients. The growth temperature ranged from 600 to 800. Post-growth annealing was carried out at temperatures ranging from 800 to 1100 in a flowing O2 ambient for 1hr. The transport properties of the as-grown and annealed films were determined using four-point Van der Pauw Ha ll measurements at room temperature. The photoluminescence properties of the films were also measured at room temperature using a He-Cd laser (325nm). The surface morphology of the films was investigated using atomic force microscopy (AFM). Results and Discussion Several studies discussed the passivation effect of hydrogen incorporation into ZnO [156-158]. Ogata et al [156] reported that high resistive N-doped ZnSe films were

PAGE 96

84 converted to p-type by thermal annealing in nitrogen atmosphere. A hydrogen passivation model was proposed to explai n the activation and passiva tion of N acceptors resulting from the dehydrogenation and hydrogenation. Accordingly, hydrogen passivation could be very helpful to enhance the p-type doping of ZnO since it would prevent self compensation during growth [158]. Thus, H2/Ar gas mixture was introduced during the growth of P-doped ZnO to enhance the ptype doping of ZnO in this work. Hall measurement was performed on these samples to examine the transport properties at room temperatur e. All of the films show n-type conductivity which is consistent with previous results on the as -grown P-doped ZnO films. Figure 7-1 shows the resistivity as a function of deposition temperature for films grown under different ambients. With increasing growth temperatur e, the difference in the resistivity of the films grown under different ambients incr eases. At growth temperature of 600 the resistivity is within an orde r of magnitude for the films grown under different ambient. As growth temperature increases above 700, there is little change in resistivity for the films grown in pure oxygen and O2/Ar/H2 mixture. In contrast, re sistivity increases from 2.210-2 -cm to 3.5-cm for the films grown in ozone/oxygen mixture. Figure 7-2 shows the carrier density of ZnO: P0.002 films grown over a temperature range of 600 to 800. The doping of phosphorus in ZnO films exhibits a significant increase in electron concentrati on which is in the range of 1019 to 1020 cm-3. However, for the films grown in ozone/oxygen, the electr on density rapidly decreases as increasing temperature. Compared with the O2/Ar/H2 mixture and pure O2, the ozone/oxygen ambient presents a stronger oxidizing speci es because of the weaker O-O bonds in O3. According to the theoretical calculations from chemical potentials [76], the anion-rich

PAGE 97

85 600650700750800 10-310-210-1100101102103 Resistivity (ohm-cm)Growth Temperature (oC) Ozone/Oxygen, PO3/O2=60mTorr Oxygen+Ar/H2, PO2+Ar/H2=60mTorr Oxygen,PO2=60mTorr 10-310-210-1100101102103 Figure 7-1. Room temperature resistivity as a function of growth temperature for ZnO: P0.002 films grown in different gas ambient. (oxygen-rich) growth conditions could inhibit the formation of compensating defects in p-type doping of ZnO. Theref ore, the as-deposited ZnO:P films grown in ozone/oxygen condition shows significantly lo wer electron density with incr easing growth temperature. The Hall mobility of P-doped ZnO samples is shown in Figure 7-3. With increasing temperature, mobility continuously increases pos sibly due to the improved crystallinity of the films. For the films grown at 800, in pure oxygen and O2/Ar/H2 mixture, the mobility is about 35 cm2/Vs. Consequently, the increase in resistivity for the films gr own in ozone/oxygen is due to the rapid decrease in carrier density even though there is a slight increase in mobility. Note that the resistivity of the films grown in O2/Ar/H2 mixture is higher than those of the films grown in pure oxygen at different temperatures resulting from a lower value of

PAGE 98

86 electron density. To further investigate and understand the effect of growth condition on the properties of P-doped ZnO films, photolum inescence needs to be measured. Figure 74 (a)-(c) show the RT-PL spectrums measured for the as-deposited ZnO:P0.002 films grown in O2/Ar/H2, pure oxygen and ozone/oxygen, respectively. A strong dependence on different growth ambient a nd temperatures has been demonstrated. The film grown in O2/Ar/H2 mixture shows a stronger band edge emis sion than those of the films grown in the other two ambients. With increasing growth temperature, this difference becomes more prominent. At growth temperature of 800, the films grown in O2/Ar/H2 mixture show a strong band edge emission at ar ound 3.29 eV. The improvement in band edge emission intensity for the films grown in O2/Ar/H2 mixture may reflect the passivation effect of the deep acceptor-rela ted levels by hydrogen, which also yields the passivation of the deep level emission. Thus, the radia tive transition efficiency through band-band recombination is greatly enhanced. Also note that for the films grown under O2/Ar/H2 mixture and pure oxygen, the peak in the band edge emission shifts to s lightly longer wavelength, i.e. lower energy with increasing growth temperature. Regardi ng to the Hall results in Figure 7-2, the carrier density of the films grown in O2/Ar/H2 and pure oxygen decreases with the growth temperature. For both growth conditions, th e carrier density is in the range of 1019 to 1020 cm-3. Therefore, this shift in PL could be explained by using the Moss-Burstein effect which refers to an increase in the band gap due to the increase in Fermi level in highly degenerate conditions [159].

PAGE 99

87 600650700750800 1016101710181019102010211022 Carrier Concentration (cm-3)Growth Temperature (oC) Ozone/Oxygen, PO3/O2=60mTorr Oxygen+Ar/H2, PO2+Ar/H2=60mTorr Oxygen,PO2=60mTorr 1016101710181019102010211022 Figure 7-2. Carrier density of ZnO: P0.002 films as a function of growth temperature. 600650700750800 5 10 15 20 25 30 35 40 45 50 55 Mobility (cm2/ Vs)Growth Temperature (oC) Ozone/Oxygen, PO3/O2=60mTorr Oxygen+Ar/H2, PO2+Ar/H2=60mTorr Oxygen,PO2=60mTorr 5 10 15 20 25 30 35 40 45 50 55 Figure 7-3. Carrier mobility of ZnO: P0.002 films as a function of growth temperature.

PAGE 100

88 The RT-PL properties of the films grown in ozone/oxygen are shown in Figure 7-4 (c). The intensity of the band edge emissi on is very low for the films grown in O3/O2 mixture, almost quenched entirely with grow th temperature as shown in the inset of Figure 7-4 (c). The deep level lumines cence shows much stronger intensity, which increases greatly as growth temperature incr eases. Previous annealing studies of undoped ZnO [160] have shown a similar dependence of PL on the annealing ambient, with a decrease in band edge emission and an increase in visible defect-relat ed luminescence as ZnO is annealed in an oxidizing environment. However, annealing in the reducing hydrogen ambient increases the band edge em ission while subsequently decreasing the deep level emission. In addition to the growth ambient effect, the growth temperature also plays an important role in the band edge emission. Increasing the deposition temperature improves the UV band edge emission due to a reduction in the structural defects, although the opposite effect is shown fo r the films grown in O3/O2 mixture which is might due to the more effective oxidization proces s as increasing temperature. In an effort to reduce electron density and elucidate the phos phorus doping in ZnO, the effect of annealing process on the properties of the ZnO:P0.002 films grown under different ambients was examined. The as-gro wn films were annealed in flowing oxygen for 1hr at temperatures ranging from 800 to 1100. Figure 7-5 (a)-(c) shows the resistivity of the samples annealed at different temperatures and ambients as a function of growth temperature. After the samples being annealed at 800 in oxygen, the resistivity significantly increases from 10-3 to several -cm regardless of growth ambient. However, with further increasing anneali ng temperature the resistivity st arts to decrease. In the

PAGE 101

89 350400450500550600650700 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 (a)PL Intensity (a.u.)Wavelength (nm)PO2+Ar/H2=60mTorr Tg=600oC Tg=700oC Tg=800oC Figure 7-4. Photoluminescence spect ra of P-doped ZnO grown in: (a) O2/Ar/H2; (b) pure oxygen; (c) ozone/oxygen. 350400450500550600650700 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 (b)PL Intensity (a.u.)Wavelength (nm)PO2=60mTorr Tg=600oC Tg=700oC Tg=800oC Figure 7-4. Continued.

PAGE 102

90 350400450500550600650700 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 (c)PL Intensity (a.u.)Wavelength (nm)PO3/O2=60mTorr Tg=600oC Tg=700oC Tg=800oC Figure 7-4. Continued. previous studies on the annealing effect of P-doped ZnO, phosphorus was activated as an acceptor in P-doped ZnO at around 800 [90]. It also showed that with increasing annealing temperature above 800, the hole concentration in th e films starts decreasing. These results may suggest that around 800 is the optimized annealing temperature to thermally activate phosphorus in P-doped ZnO. The carrier density and Hall mobility for annealed ZnO:P0.002 films over a temperature range of 800 to 1100 are shown in Figure 7-6 a nd 7-7, respectively. The Hall measurements indicate that annealed ZnO: P0.002 films are also n-type. It indicates that carrier density shows a similar dependen ce on annealing temperature. However, the mobility of annealed films generally increa ses as annealing temperature increases. 350360370380390400 0.000 0.005 0.010 0.015 0.020 0.025 0.030 0.035 0.040 0.045 0.050 PL Intensity (a.u.)Wavelength (nm)PO3/O2=60mTorr Tg=600oC Tg=700oC Tg=800oC

PAGE 103

91 Very strong deep level emission has been shown for the RT-PL results of annealed samples grown in different conditions. Figure 7-8 shows the RT-PL results for the films grown in O2/Ar/H2 gas mixture subjecting to differ ent annealing temperatures. With increasing annealing temperature, the intensity of the band edge emission decreases while the intensity of deep level emission tremendous ly increases. This yellow-orange emission has also been observed in ZnO elsewhere [ 161-163]. It has been suggested that the yellow emission which peaks at around 2.0 eV is due to the single negatively charged interstitial oxygen ions [164] The intensity of the yellow emission increased with the temperature and oxidation time for the ZnO fi lms under thermal oxidization [161]. In this work, the intensity of the yellow-orange deep level emission was observed to increase with annealing temperature, which coul d be explained by the increased oxygen interstitials in the P-doped ZnO films. Surface morphology of annealed P-doped ZnO films was also examined by using AFM. RMS roughness of as-grown samples increases significantly with growth temperature. These results are consistent w ith the previous ones for the Al-doped ZnO films. Figure 7-9 shows the annealing temp erature effect on the surface morphology of Pdoped ZnO films grown at 700. Note that with increasing annealing temperature, root mean square roughness decreases from 14.5 nm to 3.5 nm. As annealing temperature increases, the mobility of the surface atoms increases resulting in the decreased surface roughness. Thus, after being annealed at 1100, the sample shows atomically flat surface.

PAGE 104

92 600650700750800 10-310-210-1100101102103 Resistivity (ohm-cm)Growth Temperature (oC) Ta=1100oC Ta=1000oC Ta=900oC Ta=800oC as-grown Figure 7-5. Resistivity of 0.2 at.% P-doped Zn O films annealed at different temperatures in O2. The films were grown in: (a) O2/Ar/H2; (b) pure oxygen; (c) ozone/oxygen. 600650700750800 10-310-210-1100101102103 Resistivity (ohm-cm)Growth Temperature (oC) Ta=1100oC Ta=1000oC Ta=900oC Ta=800oC as-grown Figure 7-5. Continued. (a) (b)

PAGE 105

93 600650700750800 10-210-1100101102103 Resistivity (ohm-cm)Growth Temperature (oC) Ta=1100oC Ta=1000oC Ta=900oC Ta=800oC as-grown Figure 7-5. Continued. 600650700750800 101710181019102010211022 Carrier Concentration (cm-3)Growth Temperature (oC) Ta=1100oC Ta=1000oC Ta=900oC Ta=800oC as-grown Figure 7-6. Carrier concentra tion of 0.2 at.% P-doped ZnO films annealed at different temperatures in O2. The films were grown in: (a) O2/Ar/H2; (b) pure oxygen; (c) ozone/oxygen. (c) (a)

PAGE 106

94 600650700750800 1E17 1E18 1E19 1E20 1E21 1E22 Carrier Concentration (cm-3)Growth Temperature (oC) Ta=1100oC Ta=1000oC Ta=900oC Ta=800oC as-grown Figure 7-6. Continued. 600650700750800 10171018101910201021 Carrier Concentration (cm-3)Growth Temperature (oC) Ta=1100oC Ta=1000oC Ta=900oC Ta=800oC as-grown Figure 7-6. Continued. (b) (c)

PAGE 107

95 600650700750800 10 15 20 25 30 35 40 45 50 55 Carrier Mobility (cm2 / Vs)Substrate Temperature (degree C) Ta=1100oC Ta=1000oC Ta=900oC Ta=800oC as-grown Figure 7-7. Mobility of 0.2 at.% P-doped ZnO films annealed at different temperatures in O2. The films were grown in: (a) O2/Ar/H2; (b) pure oxygen; (c) ozone/oxygen. 600650700750800 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 Carrier Mobility (cm2 / Vs)Substrate Temperature (degree C) Ta=1100oC Ta=1000oC Ta=900oC Ta=800oC as-grown Figure 7-7. Continued. (a) (b)

PAGE 108

96 600650700750800 5 10 15 20 25 30 35 40 45 50 55 Carrier Mobility (cm2/ Vs)Substrate Temperature (degree C) Ta=1100oC Ta=1000oC Ta=900oC Ta=800oC as-grown Figure 7-7. Continued. 350400450500550600650700 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 PL Intensity (a.u.)Wavelength (nm) As-grown Ta=800oC Ta=900oC Ta=1000oC Ta=1100oC Figure 7-8. RT-PL of 0.2 at.% P-doped ZnO f ilms annealed at different temperatures in O2. The films were grown at 800 and in O2/Ar/H2 mixture. (c)

PAGE 109

97 Figure 7-9. Surface morphology of 0.2 at.% P-doped ZnO films annealed at different temperatures in O2. The films were grown at 700 and in 60 mTorr O2. As-grown Ta=800oC Ta=900oC Ta=1000oC Ta=1100oC

PAGE 110

98 CHAPTER 8 CONCLUSIONS This work has focused on the process development for ZnO-based thin film transistors on glass substrate and the gr owth and annealing co ndition effects on the properties of P-doped ZnO and (Zn,Mg)O films. Top-gate type ZnO-based thin film transist ors on glass substrates were fabricated via photolithography and wet chemical etching process. HCl acid solution was used to etch ITO source and drain pa ttern. It was found that H3PO4 has the best resolution for the gate metal and the highest selectivity of ga te oxide and channel layers over underlying ITO. Three types of active channel materials including undoped ZnO, P-doped ZnO and (Zn,Mg)O were deposited at 400 in 20 mTorr oxygen by PLD. Post-growth annealing treatments were carried out to decrease th e electron density in the channel layer. The output characteristics of undoped and P-doped ZnO TFTs show the depletion-mode operation. A preferable enhancement-mode operation is shown for the device with Pdoped (Zn, Mg)O as the channel layer. A saturation of IDS is observed for small values of VDS and the calculated FE mobility is about 5.32 cm2 V-1s-1 at the drain voltage of 6V. The on/off current ratio is on the order of 103 at the gate voltage of 10V is obtained. P-type phosphorus-doped (Zn0.9Mg0.1)O films have been real ized via pulsed laser deposition without post-annealing process. The conduction type of the films strongly depends on the oxygen partial pressure during th e deposition process. For films grown at 500 C, increasing the oxygen partial pressure fr om 20 to 200 mTorr yielded a carrier type conversion from n-type to p-type. At oxygen pressure of 150 mTorr, the films show p-

PAGE 111

99 type carrier type conduction with a hole concentration of 2.71016 cm-3, a mobility of 8.2 cm2/Vs and a resistivity of 35 -cm. These results indicate the importance of oxidation conditions in realizing p-type (Zn,Mg)O films. XPS measurements confirmed the existence of P2O5 in the p-type P-doped ZnMg O film. XRD results showed good crystallinity of P-doped ZnMgO films grown under different oxygen pressures. The inplane and out-of-plane or ientation relationships are determined as ZnO [10 10]sapphire [10 10] and ZnO [0001]sapphire [0001], respectively. The RMS roughness of the films increased from 2.60 nm to 12.8 nm as oxygen partial pressure increased from 20 mTorr to 200 mTorr. The pr esence of grain boundaries can contribute to the low carrier mobility of p-type P-doped ZnMgO films. (Zn0.9Mg0.1)O:P/ZnO heterostructures were fabricated on sapphire and ZnO substrates via pulsed laser deposition. Au and Ti/Au metal served as Ohmic contacts to pZnMgO:P and n-ZnO, respectively. The Ohmi c characteristics of the contacts were improved after post-growth RTA process in N2. The heterojunction de vices exhibit clear rectifying electrical characteristics for bot h structures. For the structure grown on sapphire substrate the turn-on voltage VT is determined to be about 1.36 V. For the ZnMgO:P/ZnO heterostructure grown on singl e crystal ZnO substrate, the threshold voltages are 1.15 V and 2.26 V for the lateral an d vertical device structure, respectively. The large ideality factors may result from several current transport processes being present in the non-ideal wide bandgap p-n j unctions. Further work should be focused on increasing the hole carrier con centration and mobility of p -ZnMgO and the optimization of structure synthesis in order to improve ZnMgO/ZnO p-n junction characteristics.

PAGE 112

100 Systematic studies of growth condition effect on the properties of 0.01 at.% Aldoped ZnO with MgO buffer layer have been focused. The dependence of resistivity on growth temperature, laser ener gy and oxygen growth pressure i ndicates that the resistivity of Al-doped ZnO is related to Al doping concentration, oxygen vacancies, Al and Zn concentrations at intersti tial sites. The Al-doped ZnO films exhibit the band edge photoluminescence at ~377 nm with very low de ep level emission. The intensity of the band edge emission of Aldoped ZnO films significantly increases with growth temperature and deposition laser energy. The photoluminescence properties of the Aldoped ZnO films have strong correlations to the electrical properties and crystallinity of the films. The possibility of the non-radiativ e trapping through deep level defect states decreases with increasing th e electron density of Al-doped ZnO films. AFM results showed that the root-mean-square roughness increases with growth temperature and oxygen partial pressure. The effect of growth ambient and post-g rowth annealing proce ss on electrical and optical properties as well as the surfa ce morphology of 0.2 at.% P-doped ZnO films was examined. The resistivity of the as-deposited samples grown in ozone/oxygen ambient rapidly increased with growth temperature. The oxidizing spices play an important role in yielding low native donor defects in P-doped ZnO. A strong dependence of RT-PL on different growth ambients and temperatures has been demonstrated. The improvement in band edge emission intensity for the films grown in O2/Ar/H2 mixture may reflect the passivation effect of the deep acceptor-related levels by hydroge n, which also yields the passivation of the deep level em ission. Annealing in oxygen at 800 increased the resistivity of P-doped ZnO films by three magnitudes. Further increasing annealing

PAGE 113

101 temperature decreased the resi stivity of the samples. Both as-grown and annealed Pdoped ZnO films show n-type conductivity. Th e intensity of the yellow-orange emission in RT-PL was enhanced by annealing th e samples in oxygen. The increased oxygen interstitials resulted from annealing process ma y contribute to this de ep level emission for the P-doped ZnO films.

PAGE 114

102 LIST OF REFERENCES 1. D. C. Look, Mater. Sci. Eng. B80, 383 (2001). 2. R. L. Hoffman, B. J. Norris, Appl. Phys. Lett. 82, 733 (2003). 3. R. L. Hoffman, J. Appl. Phys. 95, 5813 (2004). 4. K. Nomura, Nature, 432, 488 (2004). 5. Z. L. Wang, J. P: Condens. Matt. 16, R829 (2004). 6. T. Dietl, H. Ohno, Science, 287, 1019 (2000). 7. K. Ueda, H. Tabata, and T. Kawai, Appl. Phys. Lett. 79, 988 (2001). 8. E. M. C. Fourtunato, P. M. C. Barquinha A. C. M. B. G. Pimentel, A. M. F. Goncalves, A. J. S. Marques, R. F. P. Ma rtins, and L. M. N. Pereira, Appl. Phys. Lett. 85, 2541 (2004). 9. D. C. Look, D. C. Reynolds, C. W. Litt on and R. L. Jones, D. B. Eason and G. Cantwell, Appl. Phys. Lett. 81, 1830 (2002). 10. K. K. Kim, H. S. Kim, D. K. Hwang, J. H. Lim, and S. J. Park, Appl. Phys. Lett. 83, 63 (2003). 11. Y. R. Ryu, T. S. Lee, and H. W. White, Appl. Phys. Lett. 83, 87 (2003). 12. A. B. M. Almamun Ashrafi, A. Ueta, A. Avramescu, H. Kumano, I. Suemune, Y.W. Ok and T.Y. Seong, Appl. Phys. Lett. 76, 550 (2000). 13. F. Decremps, J. Z. Jiang and R. C. Liebermann, Europhys. Lett. 51, 268 (2000). 14. T. Matsuoka, N. Yoshimoto, T. Sasaki, and A. Katsui, J. Electron. Mater. 21, 157 (1992). 15. F. Hamdani, A. Botchkarev, W. Kim, H. Morkoc, M. Yeadon, J. M. Gibson, D. C. Reynolds, D. C. Look, K. Evans, C. W. Litton, W. C. Mitchel, and P. Hemenger, Appl. Phys. Lett. 70, 467 (1997). 16. F. Hamdani, M. Yeadon, D. J. Smith, H. Tang, W. Kim, A. Salvador, A. E. Botchkarev, J. M. Gibson, A. Y. Polya kov, M. Skowronski, and H. Morkoc, J. Appl. Phys. 83, 983 (1998).

PAGE 115

103 17. Y. Kayamura, Phys. Rev. B 38, 9797 (1988). 18. W. Wegscheider, L. N. Pfeiffer, M. M. Dignam, A. pinczuk, K.W. West, S. L. McCall, R. Hull, Phys. Rev. Lett. 71, 4071 (1993). 19. J. Ding, H. Jeon, T. Ishihara, M. Hagerott, and A. V. Nurmikko, Phys. Rev. Lett. 69, 1707 (1992). 20. D. C. Reynolds, D. C. Look, a nd B. Jogai, Solid State Commun. 99, 873 (1996). 21. D. M. Bagnall, Y. F. Chen, Z. Zhu, T. Yao, S. Koyama, M. Y. Shen, and T. Goto, Appl. Phys. Lett. 70, 2230 (1997). 22. P. Yu, Z. K. Tang, G. K. L. Wong, M. Kawasaki, A. Ohtomo, H. Koinuma, Y. Segawa, J. Cryst. Growth 184/185, 601 (1998). 23. D.W. Palmer, http://www.semiconductors.c o.uk/propiivi5410.htm January, 2006. 24. National Compound Semiconductor Roadma p/compound semiconductor materials, http://www.onr.navy.mil/sci_tech/informa tion/312_electronics/ncsr/properties.asp January, 2006. 25. J. D. Albrecht, P. P. Ruden, S. Limpij umnong, W. R. L. Lambrecht, and K. F. Brennan, J. Appl. Phys. 86, 6864 (1999). 26. T. Makino, Y. Sewaga, M. Kawasaki, A. Ohtomo, R. Shiraki, K. Tamura, T. Yasuda, and H. Koinuma, Appl. Phys. Lett. 78, 1237 (2001). 27. L. M. Kukreja, S. Barik, and P. Misra, J. Cryst. Growth 268, 531 (2004). 28. K. Sakurai, T. Takagi, T. Kubo, D. Kajita, T. Tanabe, H. Takasu, S. Fujita and S. Fujita J. Cryst. Growth 237-239, 514 (2002). 29. A. Ohtomo, M. Kawasaki, T. Koida, K. Masubuchi, and H. Koinuma, Y. Sakurai, Y. Yoshida, T. Yasuda, and Y. Segawa, Appl. Phys. Lett. 72, 2466 (1998). 30. S. Choopun, R. D. Vispute, W. Yang, R. P. Sharma, and T. Venkatesan, H. Shen, Appl. Phys. Lett. 80, 1529 (2002). 31. A. K. Sharma, J. Narayan, J. F. Muth, C. W. Teng, C. Jin, A. Kvit, R. M. Kolbas, and O. W. Holland, Appl. Phys. Lett. 75, 3327 (1999). 32. W. I. Park, G. C. Yi, and H. M. Jang, Appl. Phys. Lett. 79, 2022 (2001). 33. R. D. Shannon, Acta Crystallogr. A 32, 145 (1976). 34. J. F. Sarver, F. L. Katnack, and F. A. Hummel, J. Electrochem. Soc. 106, 960 (1959).

PAGE 116

104 35. A. Ohtomo, M. Kawasaki, I. Ohkubo, H. Koinuma, T. Yasuda, and Y. Segawa, Appl. Phys. Lett. 75, 980 (1999). 36. J. Nause and B. Nemeth, Semicond. Sci. and Technol. 20, S45 (2005). 37. D.C. Look, D.C. Reynolds, J.R. Sizelove, R. L. Jones, C.W. Litton, G. Cantwell and W.C. Harsch, Solid-State Comm. 105, 399 (1998). 38. M. Suscavage, M. Harris, D. Bliss, P. Yi p, S.-Q. Wang D. Schwall, L. Bouthillette, J. Bailey, M. Callahan, D.C. Look, D.C. Reynolds, R.L. Jones and C.W. Litton, MRS Internet J. Nitride Semicond. Res 4S1 G3.40 (1999). 39. E. Ohshima, H. Ogino, I. Niikura, K. Maed a, M. Sato, M. Ito, and T. Fukuda, J. of Cryst. Growth 260, 166 (2004). 40. K. Maeda, M. Sato, I. Niikura, and T. Fukuda, Semicond. Sci. Technol. 20, S49 (2005). 41. Cermet, Inc. company website: http://www.cermetinc.com/ January, 2006. 42. V. Srikant and D. R. Clarke, J. Appl. Phys. 81, 6357 (1997). 43. Y. R. Ryu, S. Zhu, J. M. Wrobel, H. M. Jeong, P. F. Miceli and H. W. White, J. Cryst. Growth 216, 326 (2000). 44. Y. Segawa, A. Ohtomo, M. Kawasaki, H. Koi numa, Z. K. Tang, P. Yu, and G. K. L. Wong, Phys. Status Solidi B 202, 669 (1997). 45. M. Kawasaki, A. Ohtomo, I. Ohkubo, H. Koinuma, Z. K. Tang, P. Yu, G. K. L. Wong, B. P. Zhang, Y. Segawa, Mater. Sci. Eng. B 56, 239 (1998). 46. M. H. Huang, S. Mao, H. Feick, H. Yan, Y. Wu, H. Kind, E. Weber, R. Russo, P. Yang, Science 292, 1897 (2001). 47. Young Jin Kim and Yoo Taek Kim, Hyung Kook Yang, Jong Chul Park, Yong Eui Lee and Hyeong Joon Kim, J Vac. Sci. Technol. A 15(3), 1103 (1997). 48. A. Ohtomo, K. Tamura, K. Saikusa, K. Takahashi, T. Makino, Y. Segawa, H. Koinuma, M. Kawasaki, Appl. Phys. Lett. 75, 2635 (1999). 49. K. Tamura, A. Ohtomo, K. Saikusa, Y. Os aka, T. Makino, M. Sumiya, S. Fuke, Y. Segawa, H. Koinuma, M. Kawasaki, J. Cryst. Growth 214/215, 59 (2000). 50. A. Tsukazaki, M. Kubota, A. Ohtomo, T. Onuma, K. Ohtani, H. Ohno, S. F. Chichibu and M. Kawasaki, Jpn. J. Appl. Phys. 44, L643 (2005). 51. T. P. Smith, H. A. Mclean, D. J. Smith, P. Q. Miraglia, A. M. Roskowski, and R. F. Davis, J. Electron. Mater. 33, 826 (2004).

PAGE 117

105 52. A. B. M. A. Ashrafi, B. P. Zhang, N. T. Binh, K. Wakatsuki, and Y. Segawa, Jpn. J. Appl. Phys. Part I 43, 1114 (2004). 53. V. Craciun, J. Elders, J. G. E. Garden iers, and I. W. Boyd, Appl. Phys. Lett. 65, 2963 (1994). 54. J. H. Choi, H. Tabata and H. Kawai, J. Cryst.Growth 226, 493 (2001). 55. T. Yamamoto, T. Shiosaki, and A. Kawabata, J. Appl. Phys. 51, 3113 (1980). 56. S. Hayamizu, H. Tabata, H. Tanaka, and T. Kawai, J. Appl. Phys. 80, 787 (1996). 57. D. B. Laks, C. G. Van der Walle, G. F. Neumark, and S. T. Pantilides, Phys. Rev. Lett. 66, 648 (1991). 58. T. Yamamoto and H. Katayama-Yoshida, Jpn, J. Appl. Phys., Part 2 38, L166 (1999). 59. Y. Li, G. S. Tompa, S. Liang, C. Gorla, Y. Lu, and J. Doyle, J. Vac. Sci. Technol. A 15(3), 1063 (1997). 60. D. C. Look and B. Clafin, Phys. Stat. Sol. (b) 241, 624 (2004). 61. A. F. Kohan, G. Ceder, D. Morgan, and C. G. Van de Walle, Phys. Rev. B 61, 15 019 (2000). 62. S. B. Zhang, S. H. Wei, and Alex Zunger, Phys. Rev. B 63, 075205 (2001). 63. D. C. Look, J. W. Hemsky, and J. R. Sizelove, Phys. Rev. Lett. 82, 2552 (1999). 64. A. Janotti and C. G. Van de Walle, Appl. Phys. Lett. 87, 122102 (2005). 65. C. G. Van de Walle, Phys. Rev. Lett. 85, 1012 (2000). 66. D. M. Hofmann, A. Hofstaetter, F. Leiter, H. Zhou, F. Henecker, B. K. Meyer, S. B. Orlinskii, J. Schmidt, and P. G. Baranov, Phys. Rev. Lett. 88, 045504 (2002). 67. C. H. Seager and S. M. Myers, J. Appl. Phys. 94, 2888 (2003). 68. F. Tuomisto, K. Saarinen, and D. C. Look, Phys. Rev. Lett. 91, 205502 (2003). 69. M. Miyazaki M. Sato, K. Mitsui and H. Nishimura, J. Un-Cryst Solids 218, 323 (1997). 70. S. Y. Myong, S. J. Baik, C. H. Lee, W. Y. Cho and K. S Lim, Jpn. J. Appl. Phys. Part II, 36, L1078 (1997). 71. H. J. Ko, Y. F. Chen, S. K. Hong, H. Wenisch, T. Yao, and D. C. Look, Appl. Phys. Lett. 77, 3761(2000).

PAGE 118

106 72. J. F. Chang, H. L. Wang, and M. H. Hon, J. Cryst. Growth, 211, 93 (2000). 73. C. H. Park, S. B. Zhang, and S. H. Wei, Phys. Rev. B 66, 073202 (2002). 74. D. J. Chadi, Phys. Rev. B. 59, 15181 (1999). 75. S. J. Pearton, D. P. Norton, K. Ip, Y. W. Heo, T. Steiner, Pro. Mater. Sci. 50, 293 (2005). 76. A. Zunger, Appl. Phys. Lett. 83, 57 (2003). 77. K. Minegishi, Y. Koiwai, Y. Kikuchi, K. Ya no, M. Dasuga and A. Shimizu, Jpn. J. Appl. Phys. 36, L1453 (1997). 78. D. C. Look, D. C. Reynolds, C. W. Litt on and R. L. Jones, D. B. Eason and G. Cantwell, Appl. Phys. Lett. 81, 1830 (2002). 79. M. Joseph, H. Tabata, and T. Kawai, Jpn. J. Appl. Phys. 38, L1205 (1999). 80. X. L. Guo, H. Tabata, T. Kawai, J. Cryst. Growth 223, 135 (2001). 81. K. Iwata, P. Fons, A. Yamada, K. Mats ubara, and S. Niki, J. Cryst. Growth 209, 526 (2000). 82. Yanfa Yan and S. B. Zhang, Phys. Rev. Lett. 86, 5723 (2001). 83. X. Li, Y. Yan, T. A. Gessert, C. L. Perkins, D. Young, C. DeHart, M. Young, and T. J. Coutts, J. Vac. Sci. Technol. A 21, 1342 (2003). 84. A. B. M. A. Ashrafi, I. Suemune, H. Ku mano, and S. Tanaka, Jpn. J. Appl. Phys. 41, L1281 (2002). 85. B. S. Li, Y. C. Liu, Z. Z. Zhi, D. Z. Shen Y. M. Lu, J. Y. Zhang, X. W. Fan, R. X. Mu, and D. O. Henderson, J. Mater. Res. 18, 8 (2003). 86. J. Huang, Z. Ye, H. Chen, B. Zhao, and L. Wang, J. Mater. Sci. Lett. 22, 249 (2003). 87. J. Lu, Y. Zhang, Z. Ye, L. Wang, B. Zhao, and J. Huang, Mater. Lett. 57, 3311 (2003). 88. J. Wang, G. Du, B. Zhao, X. Yang, Y. Zha ng, Y. Ma, D. Liu, Y. Chang, H. Wang, H. Yang, and S. Yang, J. Cryst. Growth 255, 293 (2003). 89. T. Aoki, D. C. Look, and Y. Hatanaka, Appl. Phys. Lett. 76, 3257 (2000). 90. K. K. Kim, H. S. Kim, D. K. Hwang, J. H. Lim, and S. J. Park, Appl. Phys. Lett. 83, 63 (2003).

PAGE 119

107 91. Y. W. Heo, S. J. Park, K. Ip, S. J. Pearton, D. P. Norton, Appl. Phys. Lett. 83, 1128 (2003). 92. Y. R. Ryu, T. S. Lee, and H. W. White, Appl. Phys. Lett. 83, 87 (2003). 93. Y. R. Ryu, T. S. Lee, J. H. Leem and H. W. White, Appl. Phys. Lett. 83, 4032 (2003). 94. S. Limpijumnong, S. B. Zhang, S. H. Wei, and C. H. Park, Phys. Rev. Lett. 92, 155504 (2004). 95. T. Yamamoto, and H. Katayama-Yoshida, Jpn. J. Appl. Phys. 38, L166 (1999). 96. M. Joseph, H. Tabata, H. Saeki, K. Ueda, and T. Kawai, Physica B 302/303, 140 (2001). 97. A. V. Singh, R. M. Mehra, A. Wakaha ra, and A. Yoshida, J. Appl. Phys. 93, 396 (2003). 98. F. Zhu-Ge, Z.-Z. Ye, L. Zhu, J. Lu, B. Zhao, J. Huang, Z. Zhang, L. Wang, and Z. Ji, J. Cryst. Growth, 268, 163 (2004). 99. J. G. Lu, Z. Z. Ye, F. Zhuge, Y. J. Zeng, B. H. Zhao, and L. P. Zhu, Appl. Phys. Lett. 85, 3134 (2004). 100. D. B. Chrisey and G. K. Hubler, Puls ed Laser Deposition of Thin Films, John Wiley & Sons, Inc., New York, 1994. 101. X. D. Wu, R. E. Muenchausen, S. Foltyn, R. C. Estler, R. C. Dye, C. Flamme, N.S. Nogar, A. R. Garcia, J. Martin, and J. Tesmer, Appl. Phys. Lett. 56,1481 (1990). 102. G. Koren, A. Gupta, R. J. Baseman, M. I. Lutwyche, and R. B. Laibowitz, Appl. Phys. Lett. 55, 2450 (1989). 103. R. K. Singh and J. Narayan, Phys. Rev. B, 41, 8843 (1990). 104. C. R. Brundle, C. A. Evans, Jr., and S. Wilson, Encyclopedia of Materials Characterization, Butterworth-He inemann, Stoneham, MA, 1992. 105. G. Binnig, C. F. Quate, and Ch. Gerber, Phys. Rev. Lett. 56, 930 (1986). 106. The Electronics and Electrical Engi neering Laboratory/The Semiconductor Electronics Division/Hall effect measurements, http://www.eeel.nist.gov/812/effe.htm#lore January, 2006. 107. D. C. Look, Electrical Char acterization of GaAs materi als and Devices, John Wiley & Sons, Inc., New York, 1989. 108. D. J. Chadi and K. J. Chang, Phys. Rev. Lett. 61, 873 (1988).

PAGE 120

108 109. D. V. Lang and R. A. Logan, Phys. Rev. Lett. 39, 635 (1977). 110. H. X. Jiang and J. Y. Lin, Phys. Rev. Lett. 64, 2547 (1990). 111. C. C.Wu, S. Theiss, M. H. Lu, J. C. Sturm, and S. Wagner, Proc. IEDM’96, 957 (1996). 112. G. W. Jones, SID’01 DIGEST (2001). 113. M. Stewart, R. S. Howell, L. Pires, and M. K. Hatalis, IEEE Trans. Electron Devices, 48, 845, (2001). 114. Y. He, R. Hattori, and J. Kanicki, IEEE Trans. Electron Devices 48, 1322 (2001). 115. W. E. Howard and O. F. Prache, IBM J. Res. & Dev. 45, 1 (2001). 116. J.A. Nichols and T.N. Jackson, M. H. Lu and M. Hack. SID’02 DIGEST (2002). 117. Cambridge Display Technology, http://www.cdtltd.co.uk/technology/39.asp January, 2006. 118. D. S. Ginley and C.Bright, Guest Editors, MRS Bulletin 25, 15 (2000). 119. T. Minami, MRS Bulletin 25, 38 (2000). 120. A.J. Freeman, K.R. Poeppelmeier, T. O. Mason, R.P.H.Chang and T.J.Marks, MRS Bulletin 25, 45 (2000). 121. A. Suzuki, T. Matsushita, N. Wada, Y. Sakamoto, and M. Okuda, Japan. J. Appl. Phys. 35, L56(1996). 122. H. Kawazoe, M. Yasukawa, H. Hyodo, M. Kurita, H. Yanagi and H. Hosono, Nature, 389, 939 (1997). 123. T. Minami, Semicond. Sci. Technol. 20, S35 (2005). 124. N. Duan, A.W. Sleight, J. K. Jaya raj and J.Tate, Appl. Phys. Lett. 77, 1325 (2000). 125. C.F. Windisch, K.F. Ferris and G.J. Exarhos, J. Vac. Sci. Technol. 19, 1647 (2001). 126. C. F. Windisch, G. J. Exarhos, K. F. Ferris, M. H. Engelhard and D. C. Stewart Thin Solid Films, 398–399, 45 (2001). 127. R.L. Hoffman, B.J.Norris, J.F.Wager, App. Phys.Lett. 82, 733 (2003). 128. S. Masuda, K. Kitamura, Y. Okumura, S. Miyatake, H. Tabata and T. Kawai, J. Appl. Phys. 93, 1624 (2003).

PAGE 121

109 129. P. F. Carcia, R. S. Mclean, M. H. Reilly, and G. Nunes, Jr. Appl. Phys. Lett. 82, 1117 (2003). 130. E. M. C. Fourunato, P. M. C. Barquinha, A. C. M. B. G. Pimentel, A. M. F. Goncalves, A. J. S. Marques, R. F. P. Ma rtins, and L. M. N. Pereira, Appl. Phys. Lett. 85, 2541 (2004). 131. 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, 720 (2005). 132. S. M. Sze, Physics of Semiconductor Devices, John Wiley & Sons, New York, 1981. 133. J. A. Misewich and A. G. Schrott, Mat. Res. Soc. Symp. Proc. 623, 3 (2000). 134. D. K. Schrooer, Semiconductor Material and Device Characterization, John Wiley & Son, New York, 1990. 135. Y. W. Heo, Y. Kwon, Y. Li, J. Peart on, and D. P. Norton, Appl. Phys. Lett. 84, 3474 (2004). 136. N. Takahashi, M. Shiota, Y. Zhu, M. Shim izu, D. Hirata, Y. Sakamoto, T. Sugino, and J. Shirafuji, Proc. 7th Int. Conf. on Indium Phosphide and related materials, 597-600 (1995). 137. E. M. Kaidashev, M. Lorenz, H. Von Wencksterin, A. Rahm, H.-C. Semmelhack, K. H. Han, G. Benndorf, C. Bundesmann, H. Hochmuth, and M. Grundmann, Appl. Phys. Lett. 82, 3901 (2003). 138. G. Xiong, J. Wilkinson, B. Mischuck, S. Tuzemen, K. B. Ucer, and R. T. Willams, Appl. Phys. Lett. 80, 1195 (2002). 139. C. D. Wagner, W. M. Riggs, L. E. Davis, J. F. Moulder, G. E. Muilenberg, Handbook of X-ray Photoelectron Spectrosc opy, Perkin-Elmer Corporation, 1978. 140. Ya. I. Alivov, J. E. Van Nostrand, D. C. Look, M. V. Chukichev, and B. M. Ataev, Appl. Phys. Lett 83, 2943, (2003). 141. Ya. I. Alivov, E. V. Kalinina, A. E. Cherenkov, D. C. Look, B. M. Ataev, A. K. Omaev, M. V. Chukichev and D. M. Bagnall, Appl. Phys. Lett. 83, 4719, (2003). 142. A. Osinsky, J. W. Dong, M. Z. Kauser, B. Hertog, A. M. Dabiran, P.P. Chow, S. J. Pearton, Ol Lopatiuk, and L. Chernyak, Appl. Phys. Lett. 85, 4272, (2004). 143. A. Tsukazaki, A. Ohtomo, T. Onuma, T. Makino, M. Sumiya, K. Ohtani, S. F. Chichibu, S. Fuke, Y. Segawa, H. Ohno, H. Koinuma, and M. Kawasaki, Nat. Mater. 4, 42, (2005).

PAGE 122

110 144. D. K. Hwang, S. H. Kang, J. H. Lim, E. J. Yang, J. Y. Oh, J. H. Yang, and S. J. Park, Appl. Phys. Lett 86, 222101, (2005). 145. K. Ip, Y. W. Heo, D. P. Norton, S. J. P earton, J. R. Laroche, and F. Ren, Appl. Phys. Lett. 85, 1169 (2004). 146. K. Ip, Y. Li, D. P. Norton, S. J. Pearton, F. Ren, Appl. Phys. Lett. 87, 071906 (2005). 147. H. Ohta, H. Mizoguchi, M. Hirano, S. Narushima, T. Kamiya, and H. Hosono, Appl. Phys. Lett. 82, 1029 (2003). 148. A. Tiwari, C. Jin, D. Kumar, J. Narayan, Appl. Phys. Lett. 83, 1773 (2003). 149. D. K. Hwang, K. H. Bang, M. C. Jeong, and J. M. Myoung, J. Crystal Growth 254, 449 (2003). 150. D. K. Schroeder, Semiconductor Ma terial and Device Characterization, John Wiley & Sons, New York, 1990. 151. M. Trivedi and K. Shenai, J. Appl. Phys. 85, 6889, (1999). 152. Su-Shia Lin, Jow-Lay Huang, P. Sajgalik Surf. Coat. Technol. 190, 39 (2005). 153. K. H. Kim, K. C. Park, and D. Y. Ma, J. Appl. Phys. 81, 7764 (1997). 154. W. J. Jeong, S. K. Kim, and G. C. Park, Thin Solid Film (in press). 155. Y. Igasaki and H. Saito, J Appl. Phys. 70, 3613 (1991). 156. K. Kamata, H. Misuharshi, and H. Fujita, Appl. Phys. Lett. 63, 3353 (1993). 157. K. Ogata, D. Kawaguchi, T. Kera, Sz. Fujita, and Sg. Fujita, J. Cryst. Growth 159, 312 (1996). 158. D. C. Look and B. Claflin, Phys. Stat. Sol. (b) 241, 624 (2004). 159. I. Hamberg and C. G. Granqvist, J. Appl. Phys. 60, R123 (1986). 160. Onwona Agyeman, Chao-Nan Xu, Wensheng Shi, Xu-Guan Zheng and Morio Suzuki, Jpn. J. Appl. Phys. 41, 666 (2002). 161. Y.G. Wang, S. P. Lau, X. H. Zhang, H. W. Lee, S. F. Yu, B. K. Tay, H. H. Hng. Chem. Phys. Lett. 375, 113 (2003). 162. N. Ohashi, T. Nakata, T. Sekiguchi, H. Hosono, M.Mizuguchi, T. Tsurumi, J. Tanaka, H. Haneda, Jpn. J.Appl. Phys. 38, L113 (1999).

PAGE 123

111 163. A. Ortiz, C. Falcony, J. Hernandez, M. Garcia, J.C. Alonso, Thin Solid Films 293, 103 (1997). 164. X.L. Wu, G.G. Siu, C.L. Fu, H.C. Ong, Appl. Phys. Lett. 78, 2285 (2001).

PAGE 124

112 BIOGRAPHICAL SKETCH Yuanjie Li was born on July 4, 1974, in PeopleÂ’s Republic of China. After graduating from high school in 1993, Yuanjie Li entered the Beijing University of Aeronautics and Astronautics where she spen t about 7 years to study materials science and engineering. In 1997, Yuanjie Li received a Bachelor of Engin eering in the specialty of metallic materials and heat treatment. In the same year, she became a research assistant in the Thin Film and Coatings Laboratory at Beijing University of Aeronautics and Astronautics and started her graduate study unde r the supervision of Professor Shengkai Gong. Her research field for her masterÂ’s degree was focused on the novel properties of Fe/Cu nano-scale multilayer materials grown by electron beam physical vapor deposition. In 2000, Yuanjie Li received her Master of Engineering in materials science. Yuanjie Li entered the Department of Ma terials Science and Engineering at the University of Florida to pursue a Ph.D. degr ee in Fall, 2001. After working as a teaching assistant for a year, she joined Professor Da vid NortonÂ’s research group in the summer of 2002 and has been working on the ZnO-based th in films for electrica l and optoelectronic devices since then.