IIIIIIIIIIIIBIIIIIIIIIIIIIP..." ,

CAN ST NIVERSITY LIBRARIES

llllllllllllllllllllll2lllllll‘llI

3 1293 00914 3086

 

 

 

 

 

 

 

 

 

 

 

 

This is to certify that the

Experimentafhfiggglggfigggmof Microwave Cavity

Plasma Reactors for Large Area and High Rate Diamond
Film Deposition

presented by

Jie Zhang

has been accepted towards fulfillment
ofthe requirements for

Ph.D. . Electrical Engineering
degreeln

 

 

kw

Major professor I\

Date 1/1477“

MSU u an Alfirmanw Action/"Equal Opportunity Institution 0 12771

 

   
    

   

Lunar
“Milan State
thuvmnflty

PLACE IN RETURN BOX to remove thle checkout from your record
TO AVOID FINES return on or before date due.

 

DAIEDUE DATEDUE DATEDUE

W7? ' g , - gage

l

 

 

AC5: ‘

 

ll~11 fig

c”.

 

'5: ~22 i
SH’Z3I9W

 

 

 

 

 

 

' W91?

y;il¥?fl®3

 

 

 

 

 

'JflNl 053“

 

 

 

 

 

 

 

MSU Is An Affirmative Action/Equal Opportunity Institution
c.

EXPERIMENTAL DEVELOPMENT OF
MICROWAVE CAVITY PLASMA REACTORS FOR
LARGE AREA AND HIGH RATE DIAMOND FILM DEPOSITION

By

Jie Zhang

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY

Department of Electrical Engineering

1993

ABSTRACT

EXPERIMENTAL DEVELOPMENT OF
MICROWAVE CAVITY PLASMA REACTORS FOR
LARGE AREA AND HIGH RATE DIAMOND FILM DEPOSITION

By

Jie Zhang

Diamond, with its unique mechanical, thermal, chemical, optical
and electrical properties. is an attractive material in applications ranging
from wear-resistant coatings for mechanical and optical components to
substrates for advanced semiconductor devices. This research concerns
the experimental investigation and development of microwave cavity
plasma reactors for large area and high rate diamond film deposition.

The research and development have lead to the successful creation
of a new improved prototype reactor. Diamond films with uniformities
better than 2% have been deposited on 3" and 4" silicon wafers with this
prototype reactor. The linear growth rates obtained are 0.89 urn/ hour on
2" silicon wafers, 0.67 um/hour on 3" silicon wafers and 0.43 um/hour
on 4" silicon wafers. The performance “figures of merit” were developed to
quantitatively compare diamond film deposition reactors in terms of
linear growth rate, weight gain rate, deposition area, energy efficiency,
gas flow efficiency and carbon conversion efficiency. Three generations of

microwave cavity plasma reactors have been investigated and / or

developed for diamond film deposition over larger substrate areas. The
new prototype reactor employs an end feed coupling concept which
enables the reactor to accept high microwave power input (from 1.5 kW
to 4.5 kW) and create discharges up to 12.5 cm in diameter. Three
generations of microwave cavity jet reactors have also been developed
and/ or designed. These reactors have the potential to deposit diamond
films with higher growth rates. A method to obtain the operational
characteristics of the microwave cavity plasma reactors was developed.
The reactor was characterized by a set of input. internal and output
experimental parameters. The relationship between the internal and
input experimental parameters was established to locate the required
experimental conditions when diamond films with desired properties.
such as film thickness. morphology. and uniformity, etc. are deposited.
The diamond films deposited in the prototype reactor were investigated
with respect to their growth rate and morphology as the input
experimental parameters were varied. The electric fields in the reactor
were also measured to develop a better understanding of the
electromagnetic field / plasma interactions during the diamond film
deposition process. It was found that the tangential component of the
electric field is the main discharge excitation field.

Capyrlght by
JIE ZHANG

1993

To my parents

Zu-Di Zhang and Xin-Xuan Wu

ACKNOWLEDGMENTS

The author wishes to thank Dr. Jes Asmussen, Jr. for his
encouragement, guidance and support throughout this research. Special
thanks is extended to Dr. Donnie Reinhard. Dr. Timothy Grotjohn and
Dr. Chow Ling Chang of Norton for their valuable discussions. Additional
thanks is given to Mr. J eron Campbell for his experimental laboratory
help and to Dr. Aslam for his comments on the manuscript.

This research was supported in part by grants from Norton, Corp..
Wavemat. Corp.. and Michigan Research Excellence fund.

TABLE OF CONTENTS

LIST OF TABLES .

LIST OF FIGURES

Chapter One
Introduction
1.1 Introduction .
1.2 Research Objectives .
1.3 Research History . .
1 .4 Dissertation Outline .
Chapter TWO

Diamond Film Deposition Reactors: A Review

2. 1 Introduction . .
2.1.1 A Generic Diamond Film Deposition Reactor
2.1.2 Reactor Categorization . .
2.1.3 “Figures of Merit” of Reactor Performance .
2.1.4 Notes on Literature Review .

2. 2 A Review of Diamond Film Deposition Reactors .
2. 2. 1 Thermally Activated CVD Reactor
2.2.2 Direct Current Discharge CVD Reactor

2.2.2. 1 Direct Current Discharge Reactor
2. 2. 2. 2 Direct Current Discharge Jet Reactor
2. 2. 3 Combustion Flame Reactor . . .
2. 2. 4 Radio Frequency (RF) Plasma CVD Reactors .
2.2.4.1 RF Glow Discharge Reactor

vii

xiv

QUIUIH

IO
IO
12
12
13
15
15
17
17
20
22
24
24

2.2.4.2 RF Induction Thermal Plasma Reactor .
2.2.5 Microwave Plasma Enhanced CVD Reactors

2.2.5. 1 Introduction .

2.2.5.2 Tubular Microwave Reactor

2.2.5.3 Magneto-microwave Reactor

2.2.5.4 Magneto-active Microwave Reactor

2. 2. 5. 5 Bell-jar Microwave Reactor

2. 2. 5. 6 Microwave Jet Reactor . .
2. 3 Comparison of Diamond Film Deposition Reactors .
2. 4 Bachmann C- H-O Phase Diagram

Chapter Three
The Microwave-Cavity Plasma Reactors (MCPR)

3. 1 Introduction . .
3. 2 Experimental Systems
3.2.1 Introduction .
3. 2. 2 Microwave Power Delivery Systems .
3. 2. 3 Flow Control and Vacuum Pump Systems
3. 2. 4 Computer Monitor System .
3. 3 The First Generation Seven Inch MCPR
3.3. 1 Introduction .
3. 3. 2 Reactor Geometry
3. 3. 3 Reactor Operation
3. 3. 4 Substrate Holder and Quartz Domes
3. 3. 5 Reactor Performance
3. 4 The Second Generation Seven Inch MCPR
3.4. 1 Introduction
3.4.2 Reactor Geometry .
3.4.3 ’I'Mmn Modes Excitation
3.4.4 End Feed Assembly .
3.4.5 Reactor Performance .
3. 5 Microwave Coupling Methods .
3.5.1 Introduction . .
3. 5. 2 Loop Coupling and Probe Coupling.
3. 5. 3 End Coupling and Side Coupling
3..54 ModeExcitation . . .
3. 6 The Third Generation Seven Inch MCPR
3.6. 1 Introduction

viii

26
28
28
29
3 l

35
37
4O
43

3383

49
53
57
57

32833

67
67
69
71
72
75
75
75
78
79
80
80

3.6.2 Reactor Configuration
3.6.2.1 Introduction . .
3.6.2.2 The Basic Configuration
3.6.2.3 The Forced Flow Configuration
3.6.2.4 The Jet Configuration
3.6.2.5 The Lower Cavity Resonance Configuration
3.6.2.6 The Down-stream Configuration .
3.6.3 Reactor Operation
3.6.4 Reactor Performance
3.6.4.1 Introduction . .
3.6.4.2 The Basic Configuration
3.6.4.3 The Forced Flow Configuration
3.6.4.4 The Jet Configuration

3.6.4.5 The Lower Cavity Resonance Configuration

3.6.4.6 The Down-stream Configuration .
3.7 Comparison of Seven Inch MCPRs
3. 8 Fourteen and Eighteen Inch MCPRs
3. 8. 1 Introduction
3. 8. 2 Fourteen Inch MCPR
3.8.3 Eighteen Inch MCPR

Chapter Four
The Microwave-Cavity Jet Reactors (MCJR)

4.1 Introduction . .
4. 2 Experimental Systems
4.2.1 Introduction
4. 2. 2 Microwave Power Delivery System .
4. 2. 3 Flow Control and Vacuum Pump Systems
4.3 The First Generation MCJR .
4.3. 1 Introduction .
4.3.2 Reactor Geometry
4.3.3 Reactor Operation
4.3.4 Reactor Performance .
4.4 The Second Generation MCJR
4.4. 1 Introduction
4.4.2 Reactor Geometry
4.4.3 Reactor Performance .
4.5 The Third Generation MCJR

80
80
80
84
87
87

90

92
92
93
95

888888

106
107
107
107
108
110
110
110
113
114
116
116
116
119
121

4.5. 1 Introduction
4.5.2 Reactor Geometry
4.5.3 Remarks

Chapter Five
Microwave Electric Fields in the MCPR

5. I Introduction .
5. 2 Experimental Measurement System
5.3 Measurement Results .
5.3.1 Electromagnetic Field Patterns
5. 3. 2 Power Balance and Discharge Power Density.
5. 3. 3 Electric Field Strength
5.3.4 Substrate Temperature .
5.4 Discussion
5.5 Summary

Chapter Six
Diamond Film Deposition in the MCPR

6.1 Introduction . .
6. 2 Experimental Methodology
6.2.1 Introduction
6. 2. 2 Independent. Input Experimental Parameters
6. 2. 3 Internal. Dependent Experimental Parameters
6.2.4 External, Output Experimental Parameters
6. 3 Experimental Operational Characteristics
6.3.1 Introduction . .
6. 3. 2 Experimental Configuration
6. 3. 3 Common Experimental Procedures
6. 3.3.1 Seeding Procedures .
6. 3. 3. 2 Start-up and Shut-down Procedures
6. 3. 4 Experimental Operational Characteristics .
6. 3. 5 Summary . .
6. 4 Diamond Film Deposition on 3" Silicon Wafers
6.4.1 Introduction . .
6. 4. 2 Film Growth Rate and Growth Emciency
6.4.2.1 Efl'ects of Substrate Temperature

121
121
124

125
127
133
133
137
141
I44
147
148

150
152
152
155
156
157
159
159
159
162
162
164
164
169
173
173
174
174

6.4.2.2 Effects of CH4 Concentration .
6.4.2.3 Effects of C02 Concentration .
6.4.2.4 Efi'ects of Total Flow Rate
6.4.3 Film Morphology.
6.4.3.1 Introduction .
6.4.3.2 Effects of Substrate Temperature
6.4.3.3 Effects of CH4 Concentration .
6.4.3.4 Effects of C02 Concentration .
6.4.3.5 Effects of Total Flow Rate
6.4.4 Typical Raman Spectrum
6. 4. 5 Typical Uniformity Profile
6. 4. 6 Summary

6. 5 Study of Experimental Set-ups with 4" Silicon Wafers .

6.5.1 Introduction . .

6. 5. 2 Cavity Shell Geometry

6. 5. 3 Quartz Dome Geometry

6.5.4 Substrate Location .

6.5.5 Cavity Mode Excitation .

6. 5. 6 Seeding Density

6. 5. 7 Summary . .

6. 6 Diamond Film Deposition on 4" Silicon Wafers

6.6.1 Introduction .

6. 6. 2 Film Growth Rate and Growth Efficiency
6. 6. 2. 1 Effects of Substrate Temperature
6.6.2.2 Effects of CH4 Concentration .
6.6.2.3 Effects of C02 Concentration .
6.6.2.4 Efi'ects of Total Flow Rate

6. 6. 3 Film Morphology.
6.6.3.1 Introduction
6. 6. 3. 2 Effects of Substrate Temperature
6. 6. 3. 3 Effects of CH4 Concentration .
6.6.3.4 Effects of C02 Concentration .
6.6.3.5 Efi'ects of Total Flow Rate

6.6.4 Typical Raman Spectrum

6.6.5 Typical Uniformity Profile

6.6.6 Summary .

183
187
190
194
194
194

. 200
. 207
. 207
. 215
. 215
. 222
. 223
. 223
. 224
. 225
. 226
. 228
. 230
. 231
. 232
. 232
. 233
. 233
. 235
. 238
. 241
. 241
. 241
. 244
. 244
. 250
. 250
. 256
. 256
. 256

Chapter Seven
Conclusions and Recommendations

7.1 Summary of Significant Results
'7. 1.1 Introduction .
7. 1. 2 MCPR Development . . .
7.1.3 Microwave Plasma Jet Reactor Development .
7.1.3.1 Microwave Cavity Jet Reactors
7.1.3.2 Jet Configuration in MCPR
7.1. 4 Reactor Operational Characteristics
7.1. 5 Diamond Film Deposition on 3" Silicon Wafers
7.1.6 Diamond Film Deposition on 4" Silicon Wafers
7.1.6.1 Study of Experimental Set-ups .
7. 1.6.2 Diamond Film Deposition on 4" Silicon Wafers
7.1.7 Reactor Performance “Figures of Merit" .
7.1.7.1 Performance “Figures of Merit” of MCPR
7. 1.7.2 Comparison with other Reactors .
7. 1.8 Measurement of Electric Fields in MCPR
7.2 Recommendations for Future Research

LIST OF REFERENCES

. 261
. 261
. 262
. 264
. 264
. 264
. 265
. 266
. 267
. 267
. 268
. 269
. 269
. 271
. 272
. 273

. 275

Table 2.1

Table 3.1

Table 5.1

Table 5.2

Table 6.1
Table 6.2
Table 6.3
Table 6.4
Table 6.5
Table 6.6
Table 6.7
Table 6.8
Table 6.9

Table 7.1

LIST OF TABLES

Summary of “Figures of Merit“ of
Diamond Reactors

Resonant Cavity Lengths of TM
Modes in 14“ Cavity .

Diagnostic Probe Power Reading
versus Input Power
Absorbed Power and Probe Power Ratio

Substrate Temperature Conditions (I)
Substrate Temperature Conditions (11)
Substrate Temperature Conditions (III) .
Comparison of Two Cavity Shells
Comparison of TWO Quartz Domes
Effects of Substrate Location (I)

Effects of Substrate Location (11)

Efi‘ects of Cavity Mode Excitation
Substrate Temperature Conditions (4")

Summary of “Figures of Merit" of Diamond
Film Deposition Experiments in MCPR .

41

104

130

130

175

178
180

. 224
. 226
. 226
. 228
. 229

. 270

Figure 1.1

Figure 2.1
Figure 2.2
Figure 2.3
Figure 2.4
Figure 2.5
Figure 2.6
Figure 2.7
figure 2.8
Figure 2.9
Figure 2.10

Figure 2.11
Figure 2.12

Figure 2.13

Figure 3.1
Figure 3.2
Figure 3.3
Figure 3.4
Figure 3.5
Figure 3.6
Figure 3.7
Figure 3.8

Figure 3.9
figure 3.10

Figure 3.11

LIST OF FIGURES

Diamond Properties and Its
Engineering Applications [3]

A Generic Diamond Film Deposition Reactor .
Thermally Activated CVD Reactor [18.19] .
DC Discharge Reactor [23,24]

DC Discharge Jet Reactor [25]

Combustion Flame Reactor [26]

RF Glow Discharge Reactor [28] .

RF Induction Thermal Plasma Reactor [29]
Tubular Microwave Reactor [30] .
Magneto-microwave Reactor [3 1]
Magneto-active Microwave Reactor [32]
Bell-jar Microwave Reactor [34.35]
Microwave Jet Reactor [37.38] .
Bachmann C-H-O Phase Diagram [36] .

Microwave Power Supply Circuit

Flow Control and Vacuum System for MCPRs
Computer Monitor System Block Diagram
Monitor Program Flow Chart . .

Schematic Drawing of MCPR7- 1 .

Mode Diagram of an Ideal 7" Cavity

Field Pattern of Discharge Loaded mm 1 Mode .

Raman Spectrum of a Diamond Film
Deposited in MCPR7-I . . .
Schematic Drawing of MCPR7-2 .
Field Patterns of Discharge Loaded
mom and “IT/I013 Modes

Raman Spectrum of a Diamond Film
Deposited in MCPR7-2 .

xiv

88888 888888888355: to

as

883

70

74

figure 3.12

Figure 3.13
Figure 3.14

Figure 3.15
Figure 3.16
Figure 3.17
Figure 3.18
Figure 3.19
Figure 3.20
Figure 3.21
Figure 3.22
Figure 3.23

Figure 3.24

Figure 3.25

Figure 4.1
Figure 4.2
Figure 4.3
figure 4.4

Figure 4.5
Figure 4.6

Figure 5.1
Figure 5.2
Figure 5.3
Figure 5.4
Figure 5.5

Figure 5.6
Figure 5.7
Figure 5.8

Figure 5.9

Field Pattern of mm 1 Mode

Schematic Drawing of a Loop Coupling System .

The Basic Experimental Configuration .

The Base-plate of MCPR7-3 .

The Forced Flow Experimental Configuration.
Flow Pattern Regulators . .
Jet Experimental Configuration

Lower Cavity Resonance Configuration .
Down-stream Experimental Configuration
Preliminary Jet Experimental Configuration
Schematic Drawing of a 3" Substrate Heater .
Substrate Supported by a

Quartz Tube in 14" Cavity

Substrate Supported by a

Metal Tube in 14" Cavity

Resonant Cavity Lengths of

TM Modes in 14" Cavity

Flow Control and Vacuum System for MCJRs
Schematic Drawing of MCJR- 1

Schematic Drawing of MCJR-2

Raman Spectrum of a Film

Deposited in MCJR- 2 .

Schematic Drawing of MCJR- 3 . .
Enlarged Drawing of the Nozzle Insert .

Electric Field Measurement System .
Horizontal Cross-sectional View . .
Measured Electric Field Strength Distribution
Measured Field Pattern in MCPR
Discharge Boundaries and Electric Fields
Coaxial Probe Power versus

Microwave Power and Flow Rate .
Coaxial Probe Power versus

Microwave Power and Pressure

Substrate Temperature vs.

Microwave Power and Flow Rate .
Substrate Temperature vs.

Microwave Power and Pressure

XV

76
77
8 l

85
86
88
89
9 1
94

100
101
103
109
1 1 1
1 17
120
122
123
128
129
134
136
138
142
143
145

146

at-

4.-
..w

Ida.

av.

xvi

Figure 6.1
Figure 6.2

Figure 6.3
Figure 6.4

Figure 6.5
Figure 6.6
Figure 6.7
Figure 6.8

Figure 6.9

Figure 6.10

Figure6.11

Figure 6.12

Figure 6.13

Figure 6.14

Figure 6.15

Figure 6.16

figure 6.17

I“igure 6.18

Reactor I/O Block Diagram .

Schematic Drawing of Cavity Shell (1)
Schematic Drawing of Cavity Shell (11) .
Substrate Temperature versus Pressure

and Absorbed Microwave Power .

Substrate Temperature versus

CH4 Concentration .

Substrate Temperature versus

C02 Concentration .

Substrate Temperature versus Total Flow Rate
and Absorbed Microwave Power . .
Linear Growth Rate and Weight Gain Versus
Substrate Temperature and Power Flux (I)

(a) Location of Substrate Temperature Conditions
on Energy Map and (b) Carbon Conversion
Efiiciency and Gas Flow Efiiciency versus
Substrate Temperature (1) . .
Linear Growth Rate and Weight Gain versus
Substrate Temperature and Power Flux (II)

(a) Location of Substrate Temperature Conditions
on Energy Map and (b) Carbon Conversion
Efficiency and Gas Flow Efficiency versus
Substrate Temperature (11) . .
Linear Growth Rate and Weight Gain versus
Substrate Temperature and Power Flux (III)
(a) Location of Substrate Temperature
Conditions on Energy Map and (b) Carbon
Conversion Emciency and Gas Flow Efficiency
versus Substrate Temperature (111)

Location of Gas Compositions on the C-H-O
Phase Diagram (1) .

Linear Growth Rate. Weight Gain and Carbon
Conversion Efficiency versus CH4
Concentration (1).

Linear Growth Rate, Weight Gain And Carbon
Conversion Efficiency versus CH4
Concentration (11) .
Location of Gas Compositions on the C- H-O
Phase Diagram (11) .

Linear Growth Rate, Weight Gain and Carbon

xvi

153
160
161
167
168
170
171

175
176
17.8
179
180
181
182
184

185

186

I-

Figure 6.19

Figure 6.20

Figure 6.21

Figure 6.22

Figure 6.23

Figure 6.24
Figure 6.25

Figure 6.26
Figure 6.27
Figure 6.28
Figure 6.29
Figure 6.30
Figure 6.31
Figure 6.32
Figure 6.33
Figure 6.34

Figure 6.35

l"‘lgure 6.36

Conversion Emciency versus C02
Concentration (1).

Linear Growth Rate, Weight Gain and Carbon
Conversion Efiiciency versus C02
Concentration (11)

Location of Gas Compositions on the C-H- O

Phase Diagram (III) . .
(a) Linear Growth Rate. Weight Gain and
(b) Carbon Conversion Efficiency versus Total
Flow Rate and Gas Residence Time (I)

(a) linear Growth Rate, Weight Gain and
(b) Carbon Conversion Efficiency versus Total
Flow Rate and Gas Residence Time (11)
Effects of Substrate Temperature on
Film Morphologr (I) . .
Effects of Substrate Temperature on
Film Morpholog (II) . .
Effects of CH4 Concentration on

Film Morphology (I)

Effects of CH4 Concentration on

Film Morphology (II) . .
Effects of C02 Concentration on

Film Morphology (I) . .
Effects of C02 Concentration on

Film Morphology (11)

Effects of Total Flow Rate on

Film Morphology (I)

Efi'ects of Total Flow Rate on

Film Morphology (II)

Raman Spectrum of a 3" Diamond Film
Deposited in MCPR7- 3 . .

Cross- sectional Thickness Profile of

a 3" Diamond Film . .

Three Dimensional Thickness Profile of
a 3" Diamond Film . .

Effects of Quartz Dome Geometry on
Film Morphologr.

Linear Growth Rate and Weight Gain versus

Substrate Temperature and Power Flux
Carbon Conversion Efficiency and Gas Flow

xvii

188

189

191

192

193

195

197

. 201

.204

.208

. 210

. 212

. 216

. 219

. 220

. 221

. 227

. 234

Figure 6.37

Figure 6.38

Figure 6.39

Figure 6.40

Figure 6.41

Figure 6.42
Figure 6.43
Figure 6.44
Figure 6.45
Figure 6.46

Figure 6.47

Figure 6.48

Efficiency versus Substrate Temperature
and Power Flux . . .
Linear Growth Rate. Weight Gain and
Carbon Conversion Efficiency versus
CH4 concentration

Location of Gas Compositions on the
C-H- 0 Phase Diagram (IV) . .
Linear Growth Rate. Weight Gain And
Carbon Conversion Efficiency versus
002 concentration

Location of Gas Compositions on the

C-H-O Phase Diagram (V) . .
(a) Linear Growth Rate. Weight Gain and
(b) Carbon Conversion Efficiency versus Total
Flow Rate and Gas Residence Time .
Effects of Substrate Temperature on

Film Morphology. . .

Effects of CH4 Concentration on

Film Morpholog'. . .

Effects of CO2 Concentration on

Film Morphology. . .

Effects of Total F‘Iow Rate on

Film Morphologr.

Raman Spectrum of a 4" Diamond Film

Deposited in MCPR7-3 .
Cross-sectional Thickness Profile of

a 4" Diamond Film . . .

Three Dimensional Thickness Profile of
a 4" Diamond Film

xviii

. 236

. 237

. 239

.240

. 242

. 243

. 245

. 247

. 251

.254

. 257

.258

. 259

..h.

n

-\

.5.

CHAPTER ONE

INTRODUCTION

1.1 Introduction

The beauty. hardness and rarity of diamond have made it an object
of fascination for nearly as long as recorded history. The mechanical,
electrical. optical, chemical. and thermal properties of diamond make it
attractive in applications ranging from wear-resistant coatings for
mechanical and optical components to substrates for advanced
semiconductor devices. Ever since Tennant discovered that diamond is
simply made of carbon in 1797(1), synthesis of diamond has long been a
goal of numerous research efforts of both individuals and organizations.

The diamond-cubic lattice consists of two interpenetrating face-
centered cubic lattices. displaced by one quarter of the cube diagonal.
Each carbon atom is tetrahedrally coordinated, making strong. directed 0
bonds to its neighbors using hybrid sp3 atomic orbitals. The lattice can
also be visualized as planes of six-membered saturated carbon stacked in
an ABCABC... sequence along <1 1 1> directionsfz’

Figure l . 1 summarizes the diamond properties and its engineering
applications.<3>

Diamond is most noted for its mechanical hardness which gives it
great promise for a wide range of mechanical coating applications where

 

 

 

DIAMOND ENGINEERING
PROPERTIES APPLICATIONS
Hard Abrasive Coatings
’ for Cutting Tools

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

High Thermal Conductivity ,
Heat Sinks for
. Electronic Devices
Electric Insulator ,
Heat Resistive
Microwave Power
Devices
Large Band Gap
I RF Electronic
| Low Dielectric Constant r Devices
| High Hole Mobility Iv High Speed
Electronic Devices
Electronic Devices for
Add Resistant Severe Enviroments
a such as in Space or
Radiation Resistant in Nuclear Reactors
I Electro-optical
’Ii’ansparent r Devices

 

 

 

Figure 1.1 Diamond Properties and Its Engineering Applications [3]

 

3

resistance to abrasive phenomenon is desired. Diamond coated cutting
tools will provide significant productivity improvements for machining
super-hard materials, nonferrous refractory metals and superalloys.

At room temperature and above, the thermal conductivity of
diamond is greater than that of any other known material, including
copper and silver. Thermal transport in diamond is accomplished by
phonon mechanisms. whereas in a metal, thermal conduction is by
electron transport. Diamond is unique in that no other known material is
simultaneously a superb thermal conductor, optically transparent and
electrically insulating. With this combination of properties, diamond will
be a excellent material for many engineering systems. Diamond's
relatively low thermal expansion coemcient makes it ideal for
microelectronic circuit boards and semiconductor substrates since large
thermal expansion substrates create micro-cracks in the device circuitry
resulting in open circuits in high-power. high-temperature electronic
applications.

Diamond ofi'ers significant promises for high power and high
temperature semiconductor device applications. The dielectric
breakdown electric field is 1x107volts /m. approximately 50 times that of
conventional semiconductors such as GaAs. The theoretical power
handling capability of diamond is a factor of 2500 greater than silicon.
The bandgap of diamond is 5.45 eV which is more than three times that
of conventional semiconductor such as Si and GaAs. Diamond's thermal
conductivity is approximately five times that of copper and nearly 20
times that of silicon. Diamond's dielectric constant is about half that of
GaAs. Diamond semiconductors demonstrate approximately half the
capacitive loading of other conventional semiconductors at any given

4

frequency. Diamond's saturated carrier velocity is 2.7 times that of GaAs.
The saturated carrier velocity of diamond is greater than the peak
velocity of GaAs.

Ultra-pure diamond probably has the widest electromagnetic band-
pass of any known solid material. Diamond is composed of a
homogeneous. tetrahedral network of covalently bonded carbon.
Diamond's cubic crystal structure makes diamond optically isotropic.
The fact that diamond is composed of a single element means that pure
diamond has no fundamental infi'ared absorption modes. Diamond
windows can be used in conjunction with optical detection devices for
detection of UV, visible and IR radiation.<3>

In 1955, Bundy and co-workers succeeded in the reproducible
synthesis of diamond“) with a molten transition metal solvent—catalyst
at pressures and temperatures where diamond is the thermodynamically
stable phase of carbon.

Diamond growth at low pressures where graphite is the stable
carbon phase can be traced back to W. G. Eversole<5>. Angus et al.<6’
and Deryaguin et a1.<7>, but the low growth rate (less than 0.1
micrometer per hour) was not of practical commercial importance, and
thus at that time it prevented worldwide interest in low pressure
diamond growth. The breakthrough in the synthesis of diamond at low
pressures came in the late 1970’s and early 1980’s, when a group of
Soviet researchers<8> and Japanese researchers<9> started publishing a
series of research papers on diamond film growth at higher growth rates
(several micrometers per hour) from hydrocarbon-hydrogen gas mixtures.
Since then, great interest and various techniques have been developed
for diamond film growth at low pressures.

6

applications. The success of this dissertation depended upon these
earlier experimental and theoretical studies of microwave discharges
inside microwave cavities. The theoretical and experimental knowledge
used to design. build and understand the diamond film deposition
reactors presented in this dissertation relied on knowledge obtained in
these earlier experiments.

The research described in this dissertation reflects co-operation
with Norton Company and Wavemat. Inc. The first microwave cavity
plasma reactor (MCPR) developed for diamond film deposition was
designed and built for Norton Co. at MSU by J. Asmussen in 1986. It was
experimentally evaluated at Norton Co. in Nov./Dec. of 1987 and
immediately diamond films were successfully deposited. This MSU owned
and patented technology was exclusively licenced to Wavemat. Inc. /
Norton Co. , and Wavemat. Inc. then developed a commercial version of
this reactor. One of these reactors was put into operation at MSU in
1988. Continued co-operation with Norton Company and Wavemat. Inc.
has contributed to the success of the research described in this
dissertation.

The research conducted for this dissertation experimentally
investigated and developed two difi'erent groups of advanced reactors that
produce microwave plasma discharges for diamond film growth. The first
group of reactors are the microwave cavity plasma reactors. The second
group of reactors are the microwave plasma jet reactors. The research
described here is primarily experimental. Since there is no theoretical
model for the plasma assisted deposition of diamond films, the
improvement of reactors was carried out through many experiments. The
final experimental designs were achieved through numerous

7

experimental iterations and by the investigator’s best intuition of the
deposition process. The results summarized in this dissertation are the
results of at least five hundred separate experiments and of thousands of
hours of actual experimental time.

This dissertation summarizes the experimental research and
development that has lead to the successful creation of a working
prototype of an improved microwave cavity plasma reactor concept. This
type of reactor is already being used to deposit diamond films over larger
surface areas at Norton Co.. It has made an impact on the commercial
production of low pressure diamond coated commercial products by
increasing the productivity by more than three times over earlier
microwave diamond film deposition reactors. The reactor developments
described in this dissertation were continuously monitored by Wavemat.
Inc. and Norton Co. and as new reactor concepts were developed and
tested at MSU. the successful concepts were almost immediately
incorporated into commercial diamond film deposition reactor designs.
Some of the research described in this dissertation has already been
published in scientific publications and international scientific
(”‘16) and in patent disclosures‘17> . However. much of the

results on diamond film deposition on three and four inch silicon wafers

conferences

using the advanced reactor concept has not been submitted for
publication. It is expected that in the near future, these results will also
be published.

Diamond films were deposited under various experimental
conditions and characterized with respect to the film quality, growth rate
and uniformity. Diamond Film quality was characterized by Scanning
Electron Microscopy and Raman Spectroscopy. Diamond film growth rate

8

was characterized by weight gain and Scanning Electron Microscopy.
Diamond film uniformity was characterized by Laser Interference /
Reflection Measurement. Electromagnetic field pattern measurements
were conducted to develop better understanding of the electromagnetic
field / plasma interactions.

1.4 Dissertation Outline

This dissertation is organized as follows. Chapter two presents a
generic low pressure diamond deposition reactor concept and briefly
reviews the state-of-the-art of other diamond film deposition reactors.
The performance of these reactors is compared with respect to the
diamond film quality. growth rate, growth efficiency, surface area, and
uniformity whenever possible. Chapter three describes a series of
microwave cavity plasma reactors that were developed and tested as part
of this dissertation research. Chapter four summaries the experimental
development of three generations of microwave plasma jet reactors for
diamond film growth at high growth rates. Chapter five presents an
investigation of the electromagnetic filed patterns in the microwave cavity
plasma reactor while operating under diamond film deposition
conditions. The purpose of this study was to develop a better
understanding of the electromagnetic field / plasma interactions during
film deposition process. Chapter six presents a new method that
experimentally characterizes the operational characteristics of prototype
reactors. Parametric studies of diamond film deposition on three and four
inch silicon wafers conducted in the new prototype reactor were also

described. Chapter seven presents the conclusions and some

9

speculations on future research and development of diamond film
deposition using microwave technology.

CHAPTER 'IWO

DIAMOND FILM DEPOSITION REACTORS: A REVIEW

2. 1 Introduction

2.1 . 1 A Generic Diamond Film Deposition Reactor

A generic reactor for low pressure diamond film growth by
chemical vapor deposition can be illustrated by the diagram displayed in
Figure 2. 1. Here. the reactive gas input is usually a mixture of hydrogen
(H2), hydrocarbon (CH4, C2H2, etc.) and oxygen or oxides (02, C0, C02,
etc.). etc. The energy input is provided by electrical (DC, low and / or high
frequency ac) energy and / or chemical reaction energy. With sumcient
energ input, the input reactive gas mixture is dissociated. creating a
mixture of dissociated charged species, electrons and neutral gases. A
substrate is placed close to the dissociated reactive gas species and its
temperature and/ or electrical bias may be controlled by external
temperature controller and / or dc or RF power supply. With proper choice
of input reactive gas mixture, energy input, substrate, substrate
temperature and/ or bias. etc.. a diamond film is deposited on the
substrate.

10

11

Mixture of dissociated

charged species,
electrons and neutral
gases.
Reactive gas
Energy input
input

 

Substrate

 

 

Substrate temperature
and/ or bias controller

Figure 2.1 A Generic Diamond Film Deposition Reactor

12

2. 1.2 Reactor Categorization

The reactors most frequently used in low pressure diamond film
deposition can be divided into five major categories:

(1) Thermally activated (filament) chemical vapor deposition
(CVD):

(2) Direct current (DC) discharge assisted CVD;

(3) Combustion flame;

(4) High frequency (RF or microwave) discharge assisted CVD:

(5) Other and hybrid reactors.

2. 1.3 “Figures of Merit“ of Reactor Performance

In order to quantitatively compare the performance of the reactors,
it is necessary to define a set of performance “figures of merit”. In this
literature review, the following performance “figures of merit” are defined:

(1) Linear growth rate (um/hour);

It is defined as the diamond film thickness (um) gain per unit
time (hour). .

(2) Deposition area Ad (cmz):

It is defined as the area over which diamond film is
deposited.

(3) Weight gain 11 (mg/hour);

It is defined as the diamond film weight gain (mg) per unit
time (hour). Since most literature just report linear growth rate. weight
gain is usually calculated fi'om the reported linear growth rate.
deposition area and the density of diamond. 3.5 l g/cm3.

l3

(4) Uniformity;
It is defined as the film thickness variation divided by the
average film thickness over the deposition area.
(5) Growth Efficiency
(i) Weight gain / power input (mg / kW-hr);

It is defined as the diamond film weight gain (mg) per
energy input (kW-hr).

(ii) Weight gain / total flow input (mg / liter);

It is defined as the diamond film weight gain (mg) per
total gas input (liters). The total flow input used here is the sum of all
input gas flow rates.

(iii) Carbon Conversion Emciency (%):

It is defined as the percentage of carbon atoms in the

input gases that are converted into the diamond film.

2. 1.4 Notes on Literature Review

The review described in the following sections is based on
experimental data that is published in the open literature. Since
diamond film deposition has been used in industrial applications, it is
possible that most of the state-of-art deposition information is not
available in the open literature.

The performance “figures of merit“ described in the following
sections are mostly mm and / or W from the best data
available in the literature. There are limiting factors on the accuracies of
the performance “figures of merit“. This is due to that limited data is
available in the literature concerning growth rate. deposition area and

14

uniformity. It is also ofien not clear how these data are obtained.

Film linear growth rate, for example, is measured difierenfly by
difi'erent research groups. Some groups use the crystal growth rate (1) as
the linear growth rate, i.e., the size of an individual crystal divided by
growth time. Some groups determine the film thickness from SEM
photographs (2) of the cross-sectional area of a film. The linear growth
rate is then obtained from dividing the film thickness by the deposition
time. Some groups use weight gain (3) of a substrate after a deposition
process along with the deposition area and the density of diamond,
3.51mg/cm3, to determine the film thickness. The linear growth rate is
then obtained from dividing the film thickness by the deposition time.

The author believes that the first method, which is frequently used
in earlier literature. is not an accurate measure of linear growth rate of a
film since it is often possible to grow large individual crystals without
forming a complete continuous film. The second method can be an
accurate method if the SEM instruments are calibrated and the specimen
are photographed directly onto the cross-section. The third method is a
simple method and can be an accurate method of evaluating the film
thickness provided the deposited film is uniform diamond film. The
methods of linear growth rate measurement are indicated whenever
available.

Finally in this review. diamond film deposition area is often
estimated since film uniformity data are mostly not available in the
reviewed literature.

15

2.2 A Review of Diamond Film Deposition Reactors

2.2. l Thermally Activated CVD Reactor

A typical thermally activated (hot filament) diamond film deposition
reactor is shown in Figure 2.2.‘18'19> As shown, in this reactor, the feed
gas. which is a mixture of hydrocarbon and hydrogen, dissociates when
it is heated by a hot filament. The hot filament is heated by an electric
current. The substrate is placed near the hot filament and diamond film
is formed on the substrate when the hot dissociated gas species react on
its surface.

Variations of this reactor have also been used for diamond film
growth. For example, substrate bias and / or substrate heater have been
utilized to enhance the growth rate and uniformity of the deposited
diamond film<2°>. Multiple filaments have been used to scale up the
deposition to 4 inch substrate.<2 1’

Diamond film deposition experiments have been conducted under
the following typical experimental conditions (22’ : substrate = Si, H2 flow
rate = 100 sccm. C2H2 flow rate = 0.5 sccm. CO flow rate = 12 sccm. gas
pressure = 50 Torr, total flow rate = 100 sccm, temperature of deposition
chamber = 890 °C. filament temperature = 2400 °C, filament voltage = 30
V, filament current = 28 A. heater voltage = 70 V, heater current = 6.5 A.

The “figures of merit“ available in the literature arez<22>

(1) linear growth rate (um/hour) = 0.6;

(2) Deposition area (cm2) ~ 6.4;

(3) Weight gain (mg/hour) ~ 1.35:

(3) Uniformity = n/a;

16

 

 

 

 

 

Gas input
Filament
/
A / Substrate
L I4— - — Silica holder
Silica I I
tube < Thermocouple

 

 

 

 

 

 

 

 

 

r

11

To pump

Figure 2.2 Thermally Activated CVD Reactor [18.19]

17

(4) Growth Efficiencies
(1) Growth rate / power input (mg / kW-hr) ~ 1.04:
(ii) Growth rate / gas input (mg/ liter) ~ 0.2;
(iii) Carbon conversion efficiency (%) ~0.32.
The thermally activated diamond film deposition reactor is easy to
construct, operate. and scale up. It faces the problem of filament erosion

and breakage during the diamond film deposition process.

2.2.2 Direct Current Discharge CVD Reactor

2.2.2. 1. Direct Current Discharge Reactor

A schematic drawing of a DC discharge diamond film deposition
reactor is shown in Figure 2.3. (23:24) As shown, in this reactor. the
input gas which is a mixture of hydrocarbon and hydrogen is dissociated
by a DC voltage across the anode and cathode. A DC power supply is
used to sustain the DC discharge. The substrate is mounted on the
anode and both the anode and cathode can be water cooled. Diamond
film is formed on the substrate when the dissociated gas species react on
its surface.

Typical experiments have been conducted in this reactor under the
following experimental conditions: substrate = Si and A1203. gas mixture
= H2 and CH4, CH4/H2 = 0.3 - 4%, pressure = over 200 Torr. total flow
rate = 20 sccm. anode-cathode separation = 2 cm. substrate temperature
= 800 - 850 °C, discharge voltage = 1 kV. discharge current = 4A/cm2.
substrate area = 0.25 cm2.

Notes on reactor operation: The initial stages of diamond particle

18

Water

Anode

 

 

 

 

   
  
   

WWW/”WW— —— Substrate

1— Power

<— Supply

 

..,'.. 1
- o
u‘ u.

.. _1._. 4
...~..

 

 

 

 

—
—

DC
discharge

Water

Cathode

Gas input

Figure 2.3 DC Discharge Reactor [23.24]

19

growth at a substrate temperature of 800 °C and CH4 concentration of
2% indicate that the initial nucleus density of diamond on the mirror
polished Si (1 1 1) surface is about 108/cm2. This nucleation density is as
large as that on the surface scratched by diamond polishing powder (0.5
pm in diameter). After 30-min of deposition, a continuous film can be
obtained in both cases. The growth rate of the films is about 20 um/hr.
Diamond thin films are also formed on A1203 substrate without surface
tr'eatInent.<23'24>
Diamond thin films can be obtained at a substrate temperature
higher than 600 °C. At a substrate temperature below 600 °C.
amorphous carbon films were obtained. If the substrate is mounted on
the cathode, the growth of diamond can not be detected and amorphous
or graphitic carbon is found over a wide range of experimental
conditions.
The “figures of merit“ available are:
(1) Linear growth rate (um/hour) ~ 20:
(2) Deposition area (cm2) ~ 0.25:
(3) Weight gain (mg/hour) ~ 1.76;
(3) Uniformity = n/a:
(4) Growth Efiiciencies
(i) Growth rate / power input (mg/kW-hr) ~ 0.56:
(ii) Growth rate / gas input (mg/ liter) ~ 1 .47 ;
(iii) Carbon conversion emciency (%) ~ 1 3.7 .

20

2.2.2.2. Direct Current Discharge Jet Reactor

The schematic drawing of a DC discharge jet diamond film reactor
is shown in Figure 2.4525) As shown, in this reactor, the input gas
which is a mixture of Ar, H2 and CH4 is dissociated by a DC voltage Vd
across the electrodes. A high temperature discharge jet is created and
sustained by a DC power supply. The substrate is mounted down stream
of the jet stream on a water cooled substrate stage. Diamond film is
formed on the substrate when the dissociated gas species react on its
surface. The bias voltage Vb is used to enhance the film growth rate.

The typical experimental conditions are: Ar flow rate = 30 l/min.
H2 flow rate = 10 l/min, CH4 flow rate = 1 l/min, pressure = 140 Torr,
discharge voltage = 70 - 76 V, discharge current = 133 - 150 A. bias
voltage = 0 - 500 V, bias current = 0 - 5 A. substrate = Mo plate of 20 mm
in diameter, distance between the substrate and nozzle = 57 - 102 mm.
substrate temperature = 700 - 1 100 °C. deposition time = 10 min,
substrate pretreatment = scratched with 5 - 10 um particle size diamond
paste for about 5 min.

Notes on reactor operation: The deposition rate was increased by
more than twofold when positive bias was applied to the substrate. The
deposition area was also increased but the uniformity of film thickness
did not improve. The maximum growth rate obtained was 900 um / hour.

“Figures of merit“ that are available are:

(1) Linear growth rate (um/hour) ~ 900:

(2) Deposition area (cmz) ~ 2;

(3) Weight gain (mg/hour) ~ 632:

(3) Uniformity = n/a:

21

 

 

 

 

.5
I l
\\\\\\\\\\\\\\\\\\\\\\\\\\\s

 

 

III
I
3!
1% 2-
D. U
I. 'C
in ‘(
3' :5
c‘, ,:
I) Q
I ‘ ‘0
I g ::
{I '3
R 3 i: : e, I t
t g C I
1 .0 a cam
3 g. S
E ‘I 1
0 ‘) D
' n
I ,.
. ,< .
D .‘ C
e -' g
i
s

Substrate

III. ‘— CH 4

   

   

 

Stage

 

 

 

71

‘—

1|

Cooling
water

Figure 2.4 DC Discharge Jet Reactor [25]

22

(4) Growth Efiiciencies
(i) Growth rate / power input (mg/kW-hr) ~ 57 :
' (ii) Growth rate / gas input (mg/ liter) ~ 0.26;
(iii) Carbon conversion emciency (%) ~ 1 .97 .

2.2.3 Combustion Flame Reactor

The schematic drawing of a typical combustion flame diamond film
deposition reactor is shown in Figure 2.5.<26> As shown, in this reactor,
the feed gas which is a mixture of acetylene (C2H2) and oxygen
dissociates by the chemical reaction energy of the gas mixture after a
flame is ignited. The combustion flame is self-sustained by the energr
released by the chemical reaction of the gas mixture at high temperature.
The substrate is placed in the C2H2 feather region (see insert in Figure
2.5) and diamond film is formed on the substrate when the dissociated
gas species react on its surface. *

Typical experimental conditions are as follows‘27>: substrate
location = in the C2H2 feather region of the flame, substrate temperature
= 650 - 1050 °C, gas flow ratio R = 02/C2H2 = 0.9 - 1.2, total flow rate =
2 slm (nozzle dependent). Under these experimental conditions, the
feather area is ~ 6 mm in diameter. the growth rate is ~ 60 um/ hour.

The “figures of merit“ that are available are:<27>

(1) Linear growth rate (um/hour) ~ 60:

(2) Deposition area (cm2) ~ 0.28:

(3) . Weight gain (mg/hour) ~ 5.9:

(3) Uniformity = N /A;

(4) Growth Emciencies

23

 

 

 

 

 

 

Inner cone W
C2H2 feather

Outer flame

 

 

 

 

 

 

 

 

 

Outer
flame

It
W224 Substrate holder l

 

 

 

Figure 2.5 Combustion Flame Reactor [26]

24

(1) Growth rate / power input (mg/kW-hr): N /A:
(ii) Growth rate / gas input (mg/ liter) ~ 0.049;
(iii) Carbon conversion efficiency (%) ~ 0.009

2.2.4 Radio Frequency (RF) Plasma CVD Reactors
2.2.4.1 RF Glow Discharge Reactor

Figure 2.6 displays the schematic drawing a RF glow discharge
diamond film deposition reactor<28>. As shown, in this reactor, the input
gas, which is a mixture of hydrocarbon and hydrogen, is dissociated by a
radio frequency (RF) energy matched into reaction chamber through an
impedance matching network. A 13.56 MHz RF power supply is used to
sustain the RF discharge. The substrate is placed in the middle of the
coil. Diamond film is formed on the substrate when the dissociated gas
species react on its surface.

Typical experimental conditions are as follows: substrate = Si or
Mo or SiOz, input gas mixture CH4/H2 = 0.2 - 1%. total flow rate ~ 50
sccm. pressure = 3.8 - 22.8 Torr. RF power = 0.5 ~ 1 kW, substrate
temperature ~ 950 °C, quartz tube geometry = 3 cm in diameter and 1 m
long. coil geometry = 4 cm in diameter and 13 cm long, coil material = 13
turns of 6.4 mm diameter copper tube,

Notes on the reactor operation: The substrate temperature could
not be determined accurately at low temperatures due to radiation from
the discharge. Since there is no external substrate heating, the
temperatures of the substrates were raised by inductive heating and
energy transfer from the discharge. It is necessary to use high RF power

25

 

 

 

 

 

 

 

 

 

 

Working coil To Manometer
0 0 0 0 0 0 0 I I
Feed gas Substrate *
—. ___
m 22_ ->
o o o o o T" I “m9
Quartz tube
— Water

 

it

 

 

 

 

 

RF generator

13.56 MHZ

 

 

Figure 2.6 RF Glow Discharge Reactor [28]

.v.

'1‘

.. ‘-

"\

26

to create the discharge and the tube wall may be sputtered at low
pressures in this reactor.

The “figures of merit“ are not obtainable from the literature.

2.2.4.2 RF Induction Thermal Plasma Reactor

The schematic drawing of a RF induction thermal plasma diamond
film reactor is displayed in Figure 2.7.‘29> As shown. in this reactor, the
input gas which is a mixture of Ar, H2 and CH4 is dissociated by a radio
frequency (RF) energy matched into reaction chamber through an
impedance matching network. A 4 MHz 60 kW RF power supply is used
to sustain the RF thermal plasma. The substrate is placed within or at
the end of the thermal plasma on a water cooled holder. Diamond film is
formed on the substrate when the dissociated gas species react on its
surface.

Typical experimental conditions are as follows: chamber inside
diameter ~ 45 mm, substrate = 20 mm diameter Mo plate, sheath (S) gas
flow rate = 35 l/m Ar + 12 l/m H2, plasma (P) gas flow rate = 17 l/m Ar,
carrier (C) gas flow rate = 81/m Ar, reactant gas flow rate = 0.1 - 1.2 l/m
CH4, pressure = 760 Torr, RF power = 60 kW, substrate temperature =
N /A, substrate pre-treatment = diamond paste polished.

Notes on reactor operation: Pyrometric measurement of substrate
temperature was not possible because of the high reflection of plasma
emission from the substrate.

The film thickness was not uniform. One of the films was 12 pm
thick near the edge and 6 pm thick at the center of the substrate. The
growth rate is 60 um/ hour for films with well defined crystal planes. The

27

Carrier + Reactant Gas

Plasma
Gas (P)

(Ar+CH4)

Ill

 

(Ar) ——>

 

 

(Ar+H9) —>:'

 

 

Sheath __l
Gas (S) <—

 

 

 

 

Substrate Holder

Substrate

 

 

 

 

—:—— Water
I'—

 

 

Figure 2.7 RF Induction Thermal Plasma Reactor [29]

28

essential condition for diamond growth in the present thermal plasma
CVD seems to be the cooling of a substrate.
The “figures of merit“ that are available are as follows:
(1) Linear growth rate (um/hour) ~ 60:
(2) Deposition area (cmz) ~ 3:
(3) Weight gain (mg/hour) ~ 63:
(3) Uniformity ~ 50%;
(4) Growth Efficiencies
(i) Growth rate / power input (mg / kW-hr) ~ 1 .05:
(ii) Growth rate / gas input (mg / liter) ~ 0.0 14;
(iii) Carbon conversion efficiency (%) ~ 0.2
This reactor has the advantage of high growth rate. It has the
following disadvantages. Due to the high gas temperature of the thermal
plasma. control of substrate temperature is dimcult. The films obtained
have poor adhesion to the substrates and the film thickness is not
uniform.

2.2.5 Microwave Plasma Enhanced CVD Reactors

2.2.5. 1 Introduction

Diamond film deposition using microwave plasma has been
achieved by several different reactors. Since this dissertation concerns
with the development of microwave plasma reactors for the deposition of
diamond film, variations of microwave reactors are reviewed in more
detail.

29

2.2.5.2 Tubular Microwave Reactor

A schematic drawing of a typical tubular micrOwave diamond film
deposition reactor is shown in Figure 2.8‘30> . As shown, in this reactor,
the input gas which is a mixture of hydrocarbon and hydrogen is
dissociated by the microwave energy coupled into the quartz tube
through a set of waveguides. power monitors and tuners. The plasma
was adjusted to be at the center of the quartz tube. Microwave power
generated by a magnetron is used to sustain the microwave discharge.
The substrate is place in the middle of the quartz tube. Diamond film is
formed on the substrate when the dissociated gas species react on its
surface.

Typical diamond film deposition conditions are as follows: quartz
tube diameter = 40 mm, substrate = silicon wafer, gas mixture CH4/H2 =
l - 3%, total flow rate = N /A, pressure = 7.6 - 60.8 Torr, microwave
power = 300 - 700 W, substrate temperature ~ 800 - 1000 °C.

The “figures of merit“ that are available are as follows:

(1) Linear growth rate (um/hour) ~ 3‘ um/hour:

(2) Deposition area (cm2) ~ 3 - 6;

(3) Weight gain (mg/hour) ~ 3 - 6:

(3) Uniformity = n/a:

(4) Growth Efficiencies

(i) Growth rate / power input (mg/kW-hr) ~ 4.3 - 8.6;
(ii) Growth rate / gas input (mg/liter) = n/ a:
(iii) Carbon conversion efficiency (%) = n / a.

‘ It is noted that, in the paper, the author showed the SEM

photograph of a film deposited in the reactor for 3 hours. The film

30

H2 + CH4

L/ Quartz tube

Wave guide

cl H4 |H-I-

Plunger

/l/Substrate

 

 

 

 

 

 

 

 

 

I Sleeve

Magnetron

 

 

Vacuum Pump

Figure 2.8 Tubular Microwave Reactor [30)

’1.

an...

31

showed discrete particles of ~ 2 um in diameter and it is not continuous.
It is not clear from the paper how the reported growth rate of 3 pm/hour
was arrived at and what measurement technique was used to determine

that film thickness.

2.2.5.3 Magneto-microwave Reactor

The schematic drawing of a magneto-microwave discharge
diamond film deposition reactor is shown in Figure 2.9531) As shown,
the input gas, which is a mixture of hydrocarbon and hydrogen, is
dissociated by the microwave energy coupled into the cylindrical
waveguide (discharge area) from a rectangular waveguide. The cylindrical
waveguide is excited with TEI 1 mode. The discharge is sustained by the
microwave power fiom a microwave power source. The magnetic field
generated by the Helmholtz-type magnetic coils is used to enhance
reactive gas dissociation through electron cyclotron resonance (875
Gauss for 2.45 GHz microwave). The substrate is placed near the
discharge and the substrate heater is used to control substrate
temperature. Diamond film is formed on the substrate when the
dissociated gas species react on its surface.

Typical experimental conditions are as follows: cylindrical
waveguide inside diameter = 160 mm, substrate = Si, substrate pre-
treatment = scratched by 3 mm size diamond paste, gas mixture CH4/H2
= 0.5 - 5%, total flow rate = 100 sccm, pressure = 0.1 - 50 Torr,
microwave power = 500 - 600 W, substrate temperature = 800 - 900 °C.

The “figures of merit“ are not available in the literature.

Sui

32

_+ K /

Coils

_Ll__

: Gas Inlet

 

 

 

 

  

 

 

 

 

Substrate _I: _
Discharge Rectangular
’* . waveguide
I Cylindrical ——L—
Heater -- waveguide H H

 

 

 

_

Figure 2.9 Magneto-microwave Reactor [3 1]

33

2.2.5.4 Magneto-active Microwave Reactor

The schematic drawing of a magneto-active microwave diamond
film deposition reactor is shown in Figure 2.10.‘32> As shown, in this
reactor, the input gas, which is a mixture of hydrocarbon, hydrogen and
oxygen, is dissociated by the microwave energy coupled into the
cylindrical process chamber through the quartz window and cylindrical
waveguide section. The discharge is sustained by the microwave power
from a microwave power source. Electron cyclotron resonance (ECR) is
used to enhance the gas dissociation. The ECR magnetic field is provided
by the permanent magnet. The substrate is located below the ECR zone
and heated during deposition by a tungsten filament located above the
substrate. Diamond film is formed on the substrate when the dissociated
gas species react on its surface.

Typical experimental conditions: cylindrical chamber diameter = 25
cm, ECR magnet = Nd-Fe-B permanent magnet (15 x 15 x 9 cm3) with
3250 Gauss at the lower magnet face, ECR zone (875 Gauss) = 8 cm from
the lower magnet face. substrate location = 6 cm below the ECR zone,
filament temperature = 2040 °C, filament-substrate separation = 0.6 cm,
substrate pre-treatment = scratched with l um diamond paste, pressure
= 2 - 20 Torr, microwave power = 950 W, substrate temperature = 540 -
640 oC, CH4 flow rate = 6 sccm, 02 flow rate = 2 sccm, H2 flow rate = 400
sccm.

Notes on reactor operation: The filament is electrically insulated
from the chamber wall to eliminate any electron emission current. The
substrate temperature is monitored by a thermocouple embedded in the
substrate support. The temperature depends mainly on the filament

34

Permanent

magnet

i

 

 

 

 

 

 

 

 

\— — Thermocouple

 

 

 

 

Substrate I

Figure 2.10 Magneto-active Microwave Reactor [32]

[WISE

'1

20 To:

Tarp

35

temperature, but there is some microwave heating, 250 - 290 °C, for
typical input microwave levels of 400 - 1000 W. The pressure range 2 —
20 Torr is low for diamond growth but substantially higher than the sub-
Torr pressures typically used with ECR plasmas.

The “figures of merit“ that are available are as follows:

(1) Linear growth rate (um/hour) ~ 2“;

(2) Deposition area (cm2) ~ 6.45;<33>

(3) Weight gain (mg/hour) ~ 4.5:

(3) Uniformity = n/a:

(4) Growth Emciencies

(i) Growth rate / power input (mg/kW-hr) ~ 4.7 ;
(ii) Growth rate / gas input (mg/ liter) ~ 0. 18:
(iii) Carbon conversion efiiciency (%) ~ 2.3.

” It is noted that, in the paper, the author showed the SEM
photographs of diamond films which consist of discrete particles of ~ 4.4
- 9.7 11m in diameter. These films are not continuous. It is not clear from
the paper how the growth rate of 2 um / hour mentioned in the paper was
calculated and what measurement technique was used to determine that

film thickness.

2.2.5.5 Bell-jar Microwave Reactor

The schematic drawing of a bell-jar microwave diamond film
deposition reactor is shown in Figure 2. l 1 534-35) As shown, in this
reactor, the input gas which is a mixture of hydrocarbon, hydrogen and
oxygen is dissociated by the microwave energy coupled into the
cylindrical cavity through the waveguide and antenna. A ball-shaped

36

 

 

 
 
  
  
  

 

 

 

 

 

Waveguide
Antenna

generator I ]
.._ Cylindrical cavity
_ Quartz bell jar
- Plasma

Substrate __
— Substrate holder
_ Quartz
cylinder
Gas mixtures -> E—

 

 

 

Vacuum pump

Temperature
controller / meter

 

Figure 2. 1 1 Bell-jar Microwave Reactor [34,35]

37.

plasma is formed inside the quartz bell jar and the substrate is placed
near the plasma. Diamond film is formed on the substrate when the
dissociated gas species react on its surface.

Typical experimental conditions: substrate = Si, pressure = 40 - 70
Torr, substrate temperature = 850 - 1030 °C, CH4/H2 = 0.5%, total flow
rate = 200 - 600 sccm.

Notes on reactor operation: Coatings of more than 65 mm (2.75
inch) diameter were deposited on silicon with a separate substrate
heater. Pressure variations between 40 and 70 Torr were found to have
little effect on the diamond growth rate in this reactor. The variation of
the total flow between 200 sccm and 600 sccm (while keeping the
methane concentration constant at 0.5%) did not afiect the growth rate.

The figures of merit that are available in the literature are:<36>

(1) Linear growth rate (um/hour) ~ 3.5:

(2) Deposition area (cmz) ~ 12.5:

(3) Weight gain (mg/hour) ~ 15.4:

(3) Uniformity ~ n/a;

(4) Growth Efiiciencies

(i) Growth rate / power input (mg/kW-hr) ~ 10.3:
(ii) Growth rate / gas input (mg/ liter) ~ 28.5;
(iii) Carbon conversion efficiency (%) ~ 4.2.

2.2.5.6 Microwave Jet Reactor

The schematic drawing of a microwave jet reactor is shown in
Figure 2.12.437"38> As shown, the input gas, which is a mixture of
hydrocarbon, hydrogen and oxygen, is dissociated near the jet nozzle by

38

Water in

l

_.H

Water out <— __

T—

Gas in —->:,

7

 

h—Quartz

Microwave
power

A

 

 

 

__<—Gasm

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Water in —> ) < -> Water out
_ t Substrate
To pump , To exhaust
<— _ —>

_1 r——

_—>Water out
_ .1—
Water in

Figure 2.12 Microwave Jet Reactor [37,38]

39

the microwave energy. Microwave energr is transmitted from the power
source to the jet nozzle through TEOI mode in the rectangular waveguide,
a transition unit and TEM mode in the coaxial waveguide. The plasma jet
is generated from the end of the center plasma flow stabilizer and blows
into the chamber where the substrate is placed on a water-cooled
substrate holder. Diamond film is formed on the substrate when the
dissociated gas species react on its surface.

Diamond films have been deposited under the following
experimental conditions: substrate = Si, Ar flow rate = 10 l/min, H2 flow
rate = 20 l/min, CH4 flow rate = 0.6 l/min, 02 flow rate = 0.15 l/min,
total pressure = 760 Torr, substrate temperature = 887 - 927 °C,
microwave power = 3.8 - 4.2 kW.

Notes on reactor operation: The diameters of the center and outer
conductor in the coaxial waveguide are 20 and 57.2 mm, respectively.
The conductors taper ofi‘ and play the roles of plasma flow stabilizers for
plasma generation. These stabilizers are made of water-cooled copper in
order to prevent thermal evaporation. The edge of the outer electrode
(plasma jet nozzle) is 22 mm in diameter, which must be designed
properly depending upon the plasma gas composition.

The “figures of merit” that are available are as follows:

(1) Linear growth rate (um/hour) ~ 12;

(2) Deposition area (cm2) ~ 2.5:

(3) Weight gain (mg/hour) ~ 10.53:

(3) Uniformity ~ n/a;

(4) Growth Emciencies

(i) Growth rate / power input (mg/kW-hr) ~ 2.6:
(ii) Growth rate / gas input (mg/ liter) ~ 0.0057 ;

40

(iii) Carbon conversion efiiciency (%) ~ 0.055.

2.3 Comparison of Diamond Film Deposition Reactors

A summary of performance “figures of merit“ of diamond film
deposition reactors is given in Table 2. l .

Since there are no electrodes present in the microwave plasma, the
problem of metallic contamination in the process of diamond deposition
does not inherently exist. Contamination from reactor walls can be
minimized with a proper reactor design. Compared to the erosion of
filaments in hot filament reactors. erosion of electrodes in direct current
reactors and nozzle erosion in combustion flame reactors, microwave
plasma diamond film deposition is a cleaner process. It is also easier to
control and optimize the deposition process which makes microwave
plasma reactors a promising technique for growing pure, uniform, and
high quality diamond films.

To successfully commercialize diamond synthesis at low pressures,
diamond growth at high rates and low deposition temperatures on large
area substrates is desirable. Each of the microwave reactors described
above has its advantages and disadvantages for diamond film deposition.
What their disadvantages have in common is that they can not be easily
scaled up for large area diamond film deposition.

The tubular reactor has two disadvantages, first, the substrate size
is limited by the by the inside diameter of the silica tube, which is 40
mm, and the system is not easily scalable for diamond film growth on a
larger surface since the diameter of the silica tube is limited by the size
of the rectangular waveguide: second, the plasma generated inside the

41

Table 2.1 Summary of "Figures of Merit" of Diamond Reactors

 

 

 

 

 

 

 

 

 

 

[37.38]

 

 

 

 

 

Gas
Linear . Dep. Weight Energy flow Carbon
growth area gain effic. e ffi c conv.
rate (cm2) (mg /hr) (mg/ (mg) emc.
Hot 0.6 6.4 1.35 1.04 0.2 0.32
filament
[221
DC 20 0.25 1.76 0.56 1.47 13.7
plasma
[23.24]
DC 900 2 632 57 0.26 1.97
jet
[25]
Flame 60 0.28 5.9 n/a 0.049 0.009
[27]
RF 60 3 63 1.05 0.014 0.2
thermal
[29]
T‘ubular 3 3 - 6 3 - 5 4.3 - 8.6 n/a n/a
MW
[30]
Mag.- 2 6.45 4.5 4.7 0.18 2.3
active
[32]
Bell-jar 3.5 12.5 15.4 10.3 28.5 4.2
MW
[36]
MW 12 2.5 10.53 2.6 0.0057 0.055
jet

 

 

 

 

3.,

{us

In

511]

'9'

u\!

"p

3“

‘P:
Jr:

1."
I 4

Li’

42

silica tube is very close to the inside walls, under the conditions suitable
for diamond deposition, erosion of the silica walls and hence
contamination of the diamond film are likely.

The magneto-microwave reactor uses a narrow, electron cyclotron
resonance region to generate the high density plasma. Non-uniformity in
the deposited film can be expected as an inherent result of the non-
uniform distribution of the resonant magnetic field across the substrate
surface, especially for diamond film deposition over a larger surface. The
reported surface area for diamond growth was 30 mm in diameter. It has
a narrow pressure region for diamond film growth, namely from 4 to 50
Torr. At pressures above 50 Torr, the discharge area becomes unstable
and below 4 Torr, the products contain graphite or $10 phase and in
extreme cases the products are only these phases. The magneto-active
microwave reactor uses the same principle as the magneto-microwave
reactor. Thus it is expected that it has the same disadvantages as the
magneto-microwave reactor.

The bell-jar reactor uses a plasma ball and the substrate is located
near the lower pole of the plasma ball. The reactive species distribution
over the substrate surface is inherently non-uniform. This is especially
true when the substrate surface extends further away from the lower
pole of the plasma ball. The reported coating surface area with the
plasma is 25 mm in diameter. A separate heater/ cooler is needed to coat
a surface of 65 mm in diameter. The location of the substrate and
substrate holder are fixed in order to generate the ball shaped plasma.
External tuning is needed in order to minimize the reflected power from
the reactor since there is only one internal adjustment (i.e. the antenna)
available for microwave coupling.

43

The jet reactor uses a small jet nozzle as an exit orifice for the
reactive species. It can be used to deposit diamond films with higher
growth rates over small areas. But it can not be used to deposit uniform
diamond film on large substrates since fine particles of amorphous
carbon were deposited on the substrate edge. This came from that the
substrate was not heated uniformly, hotter near the center and cooler

near the edge.

2.4 Bachmann C-H-O Phase Diagram

A C-H-O phase diagram was introduced by P. K. Bachmann et al.
‘36) to provide a common scheme for all major diamond chemical vapor
deposition (CVD) methods used to date. A schematic drawing of this
phase diagram is shown in Figure 2.13. Each side of the equilateral
triangle represents the atom fractions of the gas phase composition of
one of the binary systems C-H, H-0 and O-C, ranging from 0 to 1. Any
ternary C-H-O gas compositions are located inside this triangle. There
are three regions in this phase diagram, i.e., the non-diamond carbon
growth region, the diamond growth domain and the no growth region. In
chapter six, the gas compositions used in the deposition experiments are
mapped onto this phase diagram.

0.1 0.9

non-diamond

O 9.
12‘ carbon f5
\. X
9

growth region

9
/d\°

. «Toy
0.9 é 0.1

no growth region

H'Ol1'l‘l‘ljl'l'lfillolg'lo
' O/(O+H) '

 

Figure 2.13 Bachmann C-H-O Phase Diagram [36]

CHAPTER THREE

THE MICROWAVE-CAVITY PLASMA REACTORS (MCPR)

3. 1 Introduction

In order to deposit diamond films on large substrate surfaces.
diamond film deposition reactors that can create large area discharges
are needed. These reactors should be designed such that they are easily
controllable and the deposition processes can be easily repeated. In this
chapter, a series of microwave-cavity plasma reactors (MCPRs) are
investigated in order to deposit diamond films uniformly on large
substrate surfaces. Cylindrically symmetric cavity applicators which
utilize internal matching and which are excited in a single
electromagnetic mode are investigated for their ability to deposit diamond
films over large surface areas. Earlier work with this type of reactor
included the development of plasma ion sources and plasma etching and
oxidation reactors.< 12:13:33“) This type of microwave applicator has the
advantage of easy Operation and adaptability to difierent substrate sizes
and types. High quality diamond film can be produced and easily
reproduced in these reactors. The difi‘erence between the difi'erent reactor
concepts investigated here is the size of plasma that can be safely created
and the associated substrate area that can be uniformly covered with
diamond films.

45

46

There are a number of reactor geometry parameters in the plasma
reactor design. These parameters include: (1) cavity shape and size, (2)
power coupling method and (3) base-plate design, etc. All these reactor
geometry parameters have a strong influence on the reactor performance.
In the research described in this chapter, a number of these parameters
are investigated. They include: (1) 7", 14" and 18" inside diameter
cylindrical cavities, (2) probe side and end coupling. and loop side
coupling, (3) difi'erent size and shape quartz domes and their associated
base-plates. These reactors are described below in the order in which
they were experimentally investigated. The experimental microwave
power supply, reactive gas supply and vacuum pump systems used in
the investigation of these reactors are first reviewed in Section 3.2. The
computer monitor system used in the experiments described in chapter 6
is also described in section 3.2.

3.2 Experimental Systems

3.2.1 Introduction

All the microwave-cavity plasma reactors that are described in this
dissertation use microwave power as their energy source and a single
vacuum pump system was used for reactive gas supply and operating
pressure control. The microwave power supply, gas flow and vacuum
pump system used in the expeMental evaluation of MCPRs are
described below in sections 3.2.2 and 3.2.3. The computer monitor
system used in the experiments described in chapter 6 is described in
section 3.2.4.

47

3.2.2 Microwave Power Delivery Systems

The experimental microwave circuit used to deliver power into the
microwave applicators is similar to those described in earlier work
<12'13°39°4°>. Thus only a brief description is given here. A schematic
drawing of the microwave power supply circuit is shown in Figure 3. 1 . It
consists of a 2.45 GHz or 915 MHz, variable power source (1), circulator
(2) and matched dummy load (3), dual-directional coupler (4) and
associated incident and reflected power meters (5) and (6), and
microwave cavity applicator (7). The microwave power supplied by the
power source (1) propagates through the circulator (2) and the dual-
directional power coupler (4) and is incident on the microwave cavity
applicator (7). In the case of any mismatch between the impedances of
the waveguide and the cavity applicator (7), some of the incident power
will be reflected back from the cavity applicator. This reflected power
travels in the opposite direction from the incident power. It propagates
through the dual-directional coupler (4) and is directed by the circulator
(2) into the matched dummy load (3), where it is absorbed and dissipated
as thermal energy. The circulator (2) and its matched dummy load (3)
permit the propagation of the microwave energ into the cavity applicator
(7) and prevent the propagation of the reflected power back into the
power source (1) where it may cause damage to the power source. The
incident power P, is measured by the incident power meter (5) attached
to the incident power sampling port of the dual-directional coupler (4)
and the reflected power Pr is measured by the reflected power meter (6)
attached to the reflected power sampling port. The microwave power
absorbed by the cavity applicator (7) is given by Pt = P, - Pr.

48

 

(7) Microwave
Cavity
Applicator

 

 

 

 

 

 

 

I
If

a (6) Reflected Power Meter

 

(4) Dual-directional '
Power Coupler '

 

 

(2) Circulator

 

 

(l)

2 .45 GHz
or
9 1 5 MHz
Microwave
Power
Source

 

 

 

Figure 3.1 Microwave Power Supply Circuit

49

Four microwave power sources have been used during the research
described in this dissertation. The first one is the Chuang (Model No.
CA2450-1200CW) 2.45 GHz, 1.2 kW power source. It was used in the
experiments conducted in the first generation MCPR, the 2.45 GHz
experiments conducted in the 18 and 14 inch MCPRs and the low power
experiments conducted in the second generation MCPR. The second
power source used is the Chuang (Model No. MPS915-500) 915 MHz, 500
W power source. It was the only 9 15 MHz power source available in our
laboratory and was used in the 915 MHz experiments conducted in the
18 inch MCPR. Due to its low power capacity, it did not find much use in
diamond film deposition process. The third power source used is the
Gerling (Model No. GL1 19) 2.45 GHz, 3 kW power source. It was used in
the high power (1 . 1 - 1 .7 kW) experiments conducted in the second
generation MCPR and the low power (1 .5 - 3 kW) experiments conducted
in the third generation MCPR. The fourth power source used was the
Cober (Model No. S6F/ 4503) 2.45 GHz, 6 kW power source which was
used in the high power (3 - 4.5 kW) experiments conducted in the third
generation MCPR. Among the 2.45 GHz power sources available, the
Gerling 3 kW power source had the narrowest and most stable frequency

spectrum and thus was the “cleanest” power source.

3.2.3 Flow Control and Vacuum Pump Systems

A schematic drawing of the gas flow control and vacuum pump
systems is shown in Figure 3.2. As shown, the source gases, H2 (1), CH4
(2) and 002 (3), are supplied by cylinder tank gases of high purity
(99.99% purity or better). The gas flows are controlled by three MKS type

 

 

 

(4) Mass Flow
Controllers

[VI
KN

(5) Base-plate

 

 

 

 

 

% (10) Cooling
Water

 

(7) Microwave
Cavrty
Applicators
(8) Discharges

   

 

‘ 1——e

(9) Air
Cit/ling Port

__J

(6) Cooling
Water

 

 

(13) Process Chamber

 

l |

 

 

 

 

 

 

 

   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

(17) my
(15)L0w I (18) Manual I I I
Pressure Gauge Throttle Valve (22) Gate Valve
19) Automatic (23) Throttle
Throttle Valve
(12) TC1* -
(24) Baffle
(16) Isolation
Valves
(14) High
Pressure Gauge (21);?‘rgline (20) Diffusion Pump
,7: N 2
(as; 53
. (26) Syste Vent (ll) Mechanicai:
Roughing * TC] and TC; are Thermal
Pump Conductivity Vacuum Gauges
. Flow Meter (27) Exhaust

Figure 3.2 Flow Control and Vacuum System for MCPRs

51

1 159A mass flow controllers along with a MKS type 247C power supply/
digital readout/ set point source (4). The flow control ranges of the mass
flow controllers are 0 - 7.2 sccm CH4, 0 - 7.4 sccm C02, 0 - 500 and 0 -
10,000 sccm H2. The gases are mixed before they reach the base-plate
(5). Cooling water (6) was also supplied to the base-plate (5). The
microwave cavity applicator (7) sits on top of the base-plate (5).
Discharges (8) are confined in the lower section of the cavity applicator
(7) which is usually cooled by both air (9) and water (10)

The mechanical roughing pump (1 1) is the ALCATEL 2033 pump
which has an air pumping speed of 10.9 Us (23 cfrn). It has a base
pressure of 7.5 x 10'2 mTorr (1 x 10" mbar) when used with ALCATEL
200 pump oil which is a type of mineral oil distilled under vacuum. This
type of oil can be used to pump corrosive products and provides reduced
back-streaming. The pump base pressure is measured by a MKS type
286 thermal conductivity vacuum gauge TC] (12) which has a
meaningful pressure measurement range of 5 - 200 mTorr. The pressure
in the process chamber (13) is measured by a MKS Baratron type 122A
1000 Torr full scale high pressure gauge (14) for process pressure
measurement and a MKS Baratron type 127A 0. 1 Torr full scale low
pressure gauge (1 5) for the chamber base pressure measurement. These
two pressure gauges are connected to the chamber (13) through isolation
valves (16) so that they can be isolated fi'om the chamber whenever not
in use. The isolation valves (16) ensure that accurate pressure readings
can be obtained by stabilizing the zero point of the pressure gauges. MKS
type PDR-C— 1C and type PDR-D— 1 power supplies/ digital readouts are
connected to these two pressure gauges and display the pressure
readings. A MKS type 286 thermal conductivity vacuum gauge TC2 (1 7)

 

52

with meaningful pressure measurement range of 5 - 200 mTorr is used
for general indication of the pressure in the process chamber. It is quite
useful during the chamber initial pump down after the chamber is
vented to the atmospheric pressure for cleaning, sample loading and / or
unloading, etc.

The pressure in the process chamber (13) is controlled by throttle
valves between the chamber (13) and the mechanical pump (1 I). A
Varian manual throttle valve (18) was used to control the chamber
pressure in the experiments conducted in the first and second generation
MCPRs and some of the earlier experiments conducted in the third
generation MCPR. A MKS type 253A 20 mm sealing butterfly throttling
valve with MKS type 252 exhaust valve controller (19) was used to
control the chamber pressure in the later experiments conducted with
the third generation MCPR.

A lower pressure pumping system is also part of the vacuum
system. It was used to calibrate the zero point of the low pressure gauge
(15) and check vacuum tightness of the vacuum system in the
experiments described in this dissertation. A Varian VHS-6 (Model No.
0184) 6 inch diffusion pump (20) provides the lower pressure pumping
capacity to the vacuum system. This diffusion pump has an optimum
operating pressure range of 1 x 10'3 to 1 x 10"9 Torr with an air pumping
speed of 2400 Us. The fore-line valve (2 1) connects or isolates the
difi'usion pump (20) from the mechanical pump (1 1). A Varian 8 inch
viton-sealed swing gate valve (22) provides the isolation or connection of
the process chamber (13) to the difi'usion pump (20). The process
chamber pressure is controlled by a MKS 6 inch throttle valve (23). A
Varian low profile water cooled baffle (24) is used to prevent the diffusion

53

pump oil from back streaming into the process chamber (13). By closing
the manual valve (18) and/or the 20 mm throttle valve (19) on the
chamber direct roughing line, and opening the gate valve (22) and fore-
line valve (21), the lower pressure pumping system can be used. By
closing the gate valve (22) and the fore-line valve (2 1), the lower pressure
pumping system is isolated from the rest of the vacuum system.

Nitrogen (25) is used for system vent (26) and exhaust (27) purge,
because flammable gases H2 (1) and CH4 (2) are used in the experiments
described in this dissertation. Nitrogen with flow rate of twenty (or more)
times of the total flow rates of flammable gases is used to dilute the
exhaust gas mixture in the exhaust (27) pipe so that the resultant
exhaust gas mixture is no longer flammable. Nitrogen is also used to vent
the process chamber (13) during cleaning, sample loading and
unloading, etc. It helped to minimize the out-gassing from the chamber
walls during chamber pump down.

3.2.4 Computer Monitor System

‘41) was used to monitor the

A computer safety monitor system
system operating conditions and control the shut down procedure in the
experiments described in Chapter 6. A block diagram of this computer
monitor system is shown in Figure 3.3. The primary task of this
computer monitor system is to control experiment time and shut down
sequence. It is also used to ensure the safe operation of the system. As is
shown in Figure 3.3, the system operating pressure, reflected microwave
power and input microwave power readings are used as input signals for

the computer monitor system. The operating status (ON / OFF or OPEN /

54

 

 

 

Swam<m woa~zoz wmacwzoo

 

 

 

‘ Omrating Pressure
‘ Reflected Microwave Power
‘ Input Microwave Power

 

 

 

 

CH4 Gas Flow ON/OFF .
C02 Gas Flow ON/ OFF .
H2 Gas Flow ON/ OFF
Microwave Power ON/ OFF D
Automatic Valve OPEN / CLOSE +

 

 

 

 

Scam-<0) F>HZN§H5UPPUNPI

 

Figure 3.3 Computer Monitor System Block Diagram

 

55

CLOSE) of the flow controllers, microwave power source and the
automatic throttle valve is controlled by the computer monitor system
through a series of relays.

This computer monitor system was designed such that the default
state of the experimental system is in a disabled state where the gas
flows and microwave power are OFF and the automatic valve is CLOSED.
A computer program is needed to enable the experimental system,
monitor the experimental conditions and control the shut-down
sequence. The flow chart of a computer monitor program“ 1’ is shown in
Figure 3.4. As shown, in this program, the operating pressure and
reflected microwave power upper limit and the experiment running time
are first set. The experimental system is then enabled so that the gas
flows and microwave power can be turned on and the automatic throttle
valve can operate in an automatic mode to control the system pressure.
An experimental start-up procedure (see section 6.3.3.2) follows to start
the experiment after which the timer in the computer monitor system is
started. During the experiment, a checking loop is used to compare the
operating pressure, reflected microwave power and timer with the pre-set
values to determine the state of the experiment. If, at any time during the
experiment, the operating pressure and / or the reflected microwave power
go over the pre-set upper limit, the program directs the experiment into
an emergency shut-down sequence. In this sequence. the microwave
power is turned ofi' followed by the shut-ofi' of all gas flows. The
automatic throttle valve is then closed to isolate the vacuum pump from
the process chamber. These three steps follow each other very closely
with the speed of the computer. Otherwise the program directs the
system into a normal shut-down procedure (see section 6.3.3.2) when the

56

 

1 . Set operating pressure and
reflected MW power upper limit
2. Set experiment running time
3. Enable the experimental system
4. Experimental start-up procedure
(see section 6.3.2.2)
5. Start timer

 

 

 

 

l . Check operating pressure

 

 

 

 

 

 

 

 

and reflected MW power
2. Check timer

Timer Operating pressure
or reflected MW

expires
power goes over
the upper limit

I !
Emergency shut-down

Normal shut-down sequence

procedure 1 . Turn off microwave

power

2. Turn ofi‘ all gas flows

3. Close the automatic
throttle valve

(see section 6.3.2.2)

 

 

 

 

 

Figure 3.4 Monitor Program Flow Chart

57

timer expires.
3.3 The First Generation Seven Inch MCPR
3.3. 1 Introduction

The first generation seven inch MCPR (MCPR7- 1) was the first
reactor concept investigated for diamond film growth. It was developed in
the Fall of 1986 and built by J. Asmussen and placed into operation in
the Fall of 1987 at Norton Company in Salt Lake city, Utah. It was based
on a microwave plasma reactor concept invented at Michigan State
University and patented by Asmussen et al.‘42'5°> The technolog is
licensed to Wavemat, Inc. by Michigan State University. Norton Company
has an exclusive sub-license from Wavemat for the application of this
technolog to diamond film deposition. The reactor used in our
investigation is a later version which was commercialized by Wavemat.

Inc.
3.3.2 Reactor Geometry

The cross-sectional views of the MCPR7-1 are shown in Figure 3.5.
As shown, the reactor consists of the cylindrical cavity walls (1) which
form the outer conducting shell of the cavity applicator. The cavity walls
are formed from a 7 inch inside diameter, open ended metallic cylinder. A
water cooled (4 l) sliding short (2), which is electrically connected to the
cavity walls (1) via the finger stock (3). forms the top end of the cavity. It
can be moved back and forth along the longitudinal axis of the

58

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

To Vacuum Pump

 

Quartz Tube
Substrate Holder
Substrate

Top Window

Optical Pyrometer
Power Coupling Probe
Outer Conductor
Base Plate

Water Cooling

LEGEND

$1 Cavity Walls 12
2 Sliding Short 13
3 Finger Stock 14
4 Cavity bottom surface 15
5 Quartz Dome . 16
6 Discharge Zone 17
7 Gas Input Tunnel 18
8 Gas Distribution Ring 40
9 Metal Screen 41
10g Cooling Water Tunnel (42) 0—ring
ll Graphite Base

 

Figure 3.5 Schematic Drawing of MCPR7-1

59

cylindrical cavity walls (1) to change the electrical and physical height of
the cavity applicator. The cavity bottom surface (4) and the water-cooled
(10) base-plate (40) form the lower end of the cavity applicator.
Microwave power is coupled into the cavity applicator through an
adjustable coaxial power input port which is comprised of the power
coupling probe (1 7) and its outer conductor (18).

Reactive gases flow into the discharge zone (6) via the gas input
tunnel (7) inside the base-plate (40). The 9.25 cm i.d. quartz dome (5)
confines the working gas to the lower section of the cavity applicator
where the microwave fields produce a plasma discharge adjacent to the
substrate (14). A metal screen (9), which is attached to the bottom of the
base-plate (40), allows gases to flow into the vacuum pump system but
prevents microwave energy from radiating out of the applicator. The
substrate (14) and substrate holder (13) are placed on top of a quartz
tube (12). The quartz tube (12) stands on a graphite base (11) which in
turn sits on the metal screen (9). A plasma discharge is ignited in the
dome-shaped zone (6) by exciting the cavity in a single discharge loaded
resonant mode. The plasma discharge can be viewed through the top
screened window (1 5), through which the substrate temperature can also
be measured using a optical pyrometer (16).

3.3.3 Reactor Operation

The general experimental operation of this type MCPR has been
described in detail elsewhere <5l'55>. Thus, only a brief description is
given here. Difi'erences fi'om earlier work ‘5 “55’ are operation at high

pressure (30 - 80 Torr) and without a static magnetic field. The ignition,

60

matching, and internal cavity tuning are similar to that reported earlier
for microwave electrothermal thrusters <56) and microwave broad beam
ion sources<51>.

This MCPR can create a microwave discharge when excited in a
single cavity electromagnetic mode. The mode diagram of an ideal 7“
cavity is shown in Figure 3.6557) Each mode was experimentally
evaluated for its potential to deposit diamond films at discharge pressure
of 30 - 80 Torr. A discharge was started by reducing the H2/CH4 gas
pressure to 5 - 10 Torr, and applying microwave power which would then
ignite a discharge that entirely filled the quartz chamber. The discharge
pressure was then increased to an operating condition of 30 - 80 Torr
while length and probe tuning the cavity to a matched condition.

At ignition the discharge completely filled the quartz tube, but as
pressure increased to 30 - 80 Torr, the discharge contracted and
separated fi'om the surrounding quartz walls and assumed a shape
related to the field pattern of the exciting electromagnetic made. That is,
the discharge became “arc like“ and its shape and position varied with
cavity mode excitation. This arc like behavior became particularly evident
at the 40 - 80 Torr pressures.

Experiments demonstrated that one particular mode was superior
to all others for diamond film deposition. This particular mode was
identified as the mm 1 mode by electromagnetic field pattern
measurements. A detailed description of field pattern measurements is
provided in chapter 5. A schematic drawing of the field pattern of the
mom made in a discharge loaded cavity is shown in Figure 3.7. As
shown in Figure 3.5, when excited in this mode, the discharge hovered
over and was in direct contact with the substrate which was placed along

61

538 L. 82 5 no Essa—a «82 o.» 05mm

26:2 Eoeomocv Eu .9.

 

 

 

 

 

vm ON w— up w AV
P L lib b! h P p — p p p m n > p p p p a F
aucuavouu 01.020.71.393.— ucouneou/-1_ a
- r
I to.~
1 N20 mv.N 1
Wed
\\ l 11"
sauce.— a—acuue>ii>u:s:vuum ocean—Lou ‘ t x
n . .
:0
l TQN
we
.1 .ofli — .I
v! r I
s _ .
b b b — h L P _ b D b — D 1- F b 1P P

 

N6

(Kauanbaii iuouosai) ZHD °;

62

TM011 MOde

 

E E

 

 

 

 

 

 

 

Discharge
———-- 8 field line

- - -- if field line

Figure 3.7 Field Pattern of Discharge Loaded TMOI 1 Mode

63

the cavity axis. Under discharge loaded deposition conditions, the cavity
length, Ls, was set at approximately 7.2 cm and the probe position, Lp.
was adjusted to about 1 .8 cm. Deposition experiments were performed
with this electromagnetic mode.

3.3.4 Substrate Holder and Quartz Domes

A variety of substrate (14) materials have been used for diamond
film deposition in this reactor. These materials include silicon wafers.
silicon nitride. carbide drill bits. tungsten wires. etc. The most frequently
used substrate materials are silicon wafers since they are readily
available, common substrate materials in integrated circuits and bench
mark materials for reactor performance determination.

The substrate holders (13) were usually made of graphite,
amorphous carbon or boron-nitride. In most holders. recess areas are
made on the top surface to accommodate the substrate (14) materials to
prevent them from moving during the deposition processes. The typical
substrate holder (13) size had a diameter of 5 cm or smaller.

The cylindrical quartz tubes (12) were cut from a quartz tube with
an outside diameter of about 1". Typical quartz tube (12) length ranges
from 2.5 to 4 cm, depending on the experimental conditions, substrates
and substrate holders.

Besides holding the quartz tube (12), the graphite base (1 1) also
helps to attract the plasma discharge (6) to the substrate (14) so that
reactive species in the plasma discharge (6) can be efl‘ectively utilized for
diamond film growth.

The optimum substrate (14) and substrate holder (13) position was

64

found empirically by numerous experimental runs. The location varied
with substrate size and type. gas pressure and flow, input power. etc.

This reactor is designed to accept several height. 9.25 cm i.d.
quartz domes (5). There is a range for the optimum quartz dome (5)
height and this optimum range depends slightly on substrate size and
location. For quartz dome height below this optimum range, the hot
plasma used for diamond film growth makes contact with the quartz
dome top surface. Film deposition on the quartz dome and in extreme
cases quartz dome overheating and melting may result. For quartz dome
height above this optimum range, a second plasma appears in direct
contact with the quartz dome top surface. weakening the intensity of the
plasma in direct contact with the substrate. Again, film deposition on the
quartz dome and in extreme cases quartz dome overheating and melting
may result. The heights of the quartz domes used in most diamond film
deposition experiments described in this dissertation are 4.35 cm and 6
cm. The O-ring groove (42) of the quartz domes is 2 cm below the cavity
bottom surface (4). Thus, the top of the domes is located 2.35 cm and 4
cm above the cavity bottom surface (4).

3.3.5 Reactor Performance

Good quality diamond films have been deposited using this reactor
on a variety of substrate materials. The actual thin film deposition
performance of this reactor have been described in detail elsewhere<58>.
Thus. only a brief description is given here.

'lypical experimental pressures vary from 30 Torr to 80 Torr using
mixtures of hydrogen (50 sccm - 400 sccm) and methane (0.5 sccm to 3

65

sccm). Under these conditions, absorbed 2.45 GHz power varied from
250 W to 800 W. The substrate area is generally 4 cm by 4 cm or
smaller. The linear growth rate is in the range of 0.4 to 0.8 um/hour.

A typical Raman spectrum of a diamond film (WDF08) deposited in
this reactor under the following experimental conditions is shown in
Figure 3.8.

CH4 flow rate = 2.16 sccm
H2 flow rate = 175 sccm
Pressure = 72.5 Torr

Microwave Power absorbed = 440 W

Though good quality diamond films can be deposited in this
reactor, it has the following limitations. First, the vacuum o-ring seal is
located close to the hot plasma] substrate region. The substrate has an
operating temperature of the order of 700 - 1 100 °C. Hence for any
specific operating pressure, there is a limit on the absorbed power level
and substrate area to prevent the o-ring failure. Second. the power input
comes from the side of the cavity walls and produces an inherent non-
uniform electromagnetic “near” field close to the excitation probe. For a
small area substrate (5 cm in diameter or smaller), the efi'ect is not
significant. But as the substrate area and/ or input power level is
increased. this “near” field effect gets stronger. The discharge is attracted
by the probe’s strong “near” field onto the quartz dome walls, creating a
non-uniform plasma, causing non-uniform film deposition on the

substrate and heating up the quartz walls.

66

79302 5 Banana .5: 3253 a .8 8:825 58mm m.» vac

In. J<UDEHUNE

 

 

 

 

 

08:0“ c.0mmu O.DmN« 00.0hm OD.Omm O0.0NV

— b b b p n h h L b h P p p n by i p P L h:- b p ON c m.“ m
l o . comm

T l

i n. o. some

i l o . ocmm

- 3 03 u 538 >22 .,

a bob mfih u 9:53.:— ..

.. 88m m: u a: -

. 88m 2 .m u wdo -

I l o . 250

I

 

I—ITT

 

 

o . Domes

ALISNBLNI

67
3.4 The Second Generation Seven Inch MCPR

3.4. 1 Introduction

In order to solve the near field problem faced in the MCPR7- l ,
several versions of second generation MCPR were investigated for
diamond film growth.

3.4.2 Reactor Geometry

The second generation MCPR. called the MCPR7-2 in short, is
schematically shown in Figure 3.9. As shown. the difference between this
MCPR and the MCPR7- 1 is that the power probe coupling (19) is now
located at the top end of the cavity, instead of on the side of the cavity.

The sliding short (2) and coupling probe (19) adjustments, La and
[’9' provide the internal cavity impedance tuning mechanism to minimize
the reflected power.

New 7" cylindrical cavity walls (1) were designed and built which
did not have the side feed port that existed in the conventional 7 inch
cavities. This cavity was designed to be longer than the conventional
cavity so that TM012 and TM013 modes can be excited. There are two
windows /ports (22a,22b) on the new cavity. The cooling air inlet (22a) is
for air input to cool the cavity and quartz dome (5). The window/ air
outlet (22b) serves as the exhaust port for the cooling air and also as a
viewing port through which the substrate temperature is measured by an
optical pyrometer.

 

 

 

L.

220

I, I
’I ”l”!
I,
I It, I

 

 

m—OI-

 

 

 

 

 

LEGEND.

Cavity Walls
Sliding Short
Finger Stock
Covuty Bottom Surface
Quartz Dome
Plasma Discharge
Gas Input Tunnel
Gas Distribution Ring
Metal Screen
0 Cooling Water Tunnel
1 Graphite Base
12 Quartz Tube
13 Substrate Holder
14 Substrate
19 Input Coupling Probe
20 Outer Conductor
21 Cooling Water
2203 Cooling Air Inlet
22b Window/Air Outlet

d-‘mmVOIUI-hUN-e

 

 

Figure 3.9 Schematic Drawing of MCPR7-2

69

3.4.3 TMOln Modes Excitation

The experimental operation of this reactor is similar to the MCPR7-
I. It uses the same mode diagram of an ideal 7" cavity shown in Fig. 3.6.
and follows a similar starting procedure as that of MCPR7-l. The
difference between the this reactor and MCPR7-1 is the mode that is
used in diamond film deposition.

TMOI 1 mode can be easily excited in this reactor. The plasma
discharge (6) created is symmetric with respect to the axis of the cavity
walls (I). But the probe near field effect still exists in this reactor with
TMOI 1 mode excitation. Under high input power conditions, the plasma
discharge (6) is attracted to the top surface of the quartz dome (5) since
the coupling probe (19) is located close to the top surface of the quartz
dome with mm 1 mode excitation.

To eliminate the near field efi'ect. this reactor is excited using either
the discharge loaded TM012 or TM013 mode. The field patterns of man
and TM013 modes are shown in Figure 3.10. With TM012 or TMOI3 mode
excitation, the cavity length Is is either doubled or tripled from that of
TMOI 1 mode excitation. Hence the near field of the coupling probe (19) is
moved away from the quartz dome region and now has little efi'ect on the
geometry of the plasma discharge (6) that is created. Under discharge
loaded mom made excitation. the cavity length, Ls. was set at
approximately 14.4 cm and the probe position. LP. was adjusted to about
3.2 cm. Deposition experiments with this reactor (MCPR7-2) were
performed with this electromagnetic mode. Under discharge loaded
TMOI3 mode excitation. the cavity length, Ls. was set at approximately
21.6 cm and the probe position, Lp. was adjusted to about 3.2 cm. TM013

7O

882 232... can «82... 833 «mg do menses 225 2 .m aha—e

one: 3::

 

 

fl.l..---.nl
A U

 

 

 

33:35

  
    

one: 3.:

 

m
L

nnWlul- ii“ no

fi-l----..,.l
A V

as. swan-1......
as. so: will

 

 

 

71

made was used in the later part of the experiments conducted in the
third generation MCPR (MCPR7-3) which is described in section 3.6.

Another advantage of this reactor over the MCPR7- I is its reliable
operation. In the side feed configuration used in MCPR? -1 , since
competing degenerate and near degenerate TM and TE modes can all be
excited. there exists interference from degenerate and near degenerate
dipole modes to the symmetric mode excitation that is desirable for
diamond film deposition. But in the end feed configuration used in this
reactor, only symmetric TM modes are excited. Hence, the mode
excitation zone in the cavity are better separated and the interference
from degenerating TE modes is eliminated. This makes the operation of
this reactor easier and more reliable under all the substrate operating
conditions than MCPR7- l .

3.4.4 End Feed Assembly

Initially. end feed assembly with 7/8" inside diameter probe sleeve
(20) and 0.35" diameter coupling probe (19F59> was tested for diamond
film growth in February, 199 1 . Good quality diamond films were
deposited on silicon wafers. After a few experiments, it was noted that
the finger stock (3) on the sliding short (2) became fragile. It was over
heated since the sliding short (2) to which the 7 / 8" end feed assembly is
attached and the cavity walls (1) do not have water cooling to cool the
finger stock (3) in the high input power conditions required for diamond
film deposition. Also, the 7/8" coaxial cable was too small to handle the
high power (over 1 kW) required for diamond film deposition on larger
surface areas (5 cm in diameter or larger).

72

To solve the over heating and power handling problems. a new end
feed with 1 5/ 8" inside diameter probe sleeve (20) and 5/8" diameter
coupling probe (19) was designed and built. It was designed to be
attached to a water cooled (21) sliding short (2). This new end power feed
assembly has been used to operate at 4 kW over 20 hours in continuous
experiments and over 2000 hours in total running time without any
noticeable damage.

Air cooling is used to cool the cavity walls (1) and quartz dome (5)
under high power input conditions. The cooling air inlet (22a) was
initially designed and built to be 0.5” in diameter and compressed air
was used as the cooling air supply. It was a noisy and unreliable air
cooling system since it sometimes overloaded the air compressor in the
Engineering Research Complex. That air cooling system was replaced by
a new design. In the new design, the cooling air inlet (22a) was enlarged
to 2" in diameter and double screened window was soldered onto the
cavity walls (I) to ensure good electrical continuity of the cavity walls (I).
A Dayton (model No. 4C443A) air blower with free air delivery speed of
100 cfm is used to supply the cooling air through the enlarged cooling air
inlet (22a). This new system is a much quieter and more reliable air
cooling system.

3.4.5 Reactor Performance

This reactor has been used to deposit diamond films over larger
areas than that in MCPR7-1. The largest area covered was 5.7 cm in
diameter. The linear growth rates are in the same range as in MCPR7-l.
namely, 0.4 - 0.8 um/hour.

73

The 7/ 8 inch coaxial power end feed system was used for diamond
film growth on 3 cm x 3 cm square silicon substrates with the following
typical experimental conditions:

CH4 flow rate = 2 seem

H2 flow rate = 300 sccm

absorbed microwave power = 620 Watts
Pressure = 45 Torr

The 1 5/8 inch coaxial power end feed system was used for
diamond film growth on 5.7 cm diameter silicon substrates with the
following typical experimental conditions:

CH4 flow rate = 3 sccm

H2 flow rate = 300 sccm

absorbed microwave power = 1600 Watts
pressure = 40 Torr

A typical Raman spectrum of a diamond film (EDF-2) deposited in
this reactor under the following experimental conditions is shown in
Figure 3. I 1.

CH4 flow rate = 2 sccm
H2 flow rate = 300 sccm
absorbed microwave power = 560 Watts

pressure = 50 Torr

With the end feed. this reactor design solves the non-uniformity
problem caused by the probe “near’ field, but the o-ring over-heating
problem that existed in MCPR7- 1 was still a problem and limited the.

performance of this reactor.

74

«SECS 5 2330309 E—E 93830 a no 203$ :55 u u .m van—ME

A 753 6:63am:
can.” com. com. cow. oom_ cow. cow.

 

 

 

 

 

- t u - u - u - q _ u u i -
1 l 88
n
u .. 89.
w.
.IO.
6
. u
.. .. 88 S.
cm-
a
l I .. 88
3 can u 338 3:
bob on n 9:6qu
:88 8n n a:
1 :88 a u Ea - 82
p - . n p n p b . -

 

 

75
3.5 Microwave Coupling Methods

3.5. 1 Introduction

A series of experiments were performed to find the optimum
configuration for the excitation of microwave discharges for diamond film
deposition over large surface areas. The following excitation variations
were compared: (1) loop and probe coupling, (11) cavity end feed and side
feed. and (III) TMOH, TMmg. TM013 and TEoumodes excitation.

3.5.2 Loop Coupling and Probe Coupling

In an attempt to excite TEOI 1 mode, whose field pattern is shown
in Figure 3.12, a loop coupling system was designed. It is schematically
shown in Figure 3.13. The loop coupling system was realized by first
designing and constructing a series of open loops (arc) (35) from small
brass rods. The probe in the side feed was taken out and replaced by '
these open loops. The open loop was attached at one end to the end of
the center conductor of the power input port and pressed against the
outer conductor (18) wall at the other end. The loop can be moved
horizontally to change the area of the loop and rotated to change the
orientation of the loop inside the cavity. A new outer conductor tube for
the power input port was designed and built. To eliminate microwave
radiation leakage during the linear movement and rotation of the loop,
this outer conductor tube did not have the narrow opening machined on
the original outer conductor tube for the indication of probe length in the
probe coupling.

”

 

76

TEO11 Mode

 

 

 

—-—- 8 field line

— -—-3£ field line

Figure 3. 12 Field Pattern of T'Em 1 Mode

77

 

 

 

 

 

 

 

 

 

 

To Vacuum Pump

 

LEGEND

1 Cavity Walls 10 Cooling Water Tunnel
2 Sliding Short 11 Graphite Base

3 Finger Stock 12 Quartz Tube

4 Cavity bottom surface 13 Substrate Holder

5 Quartz Dome 14 Substrate

6 Plasma Discharge 15 Top Window

7 Gas Input Tunnel 16 Pyrometer Sensor

8 Gas Distribution Ring 18 Outer Conductor

9 Metal Screen 35 Power Coupling Loop

 

Figure 3. 13 Schematic Drawing of a Loop Coupling System

78

The cavity short length Ls. loop area and loop orientation were the
three parameters varied to excite loaded cavity modes and match
microwave power into the loaded cavity. It was found that the TM01 1
mode was the easiest mode to be excited and the associated plasma was
produced. The TEOI 1 mode was not excited using this loop. When the
TM01 1 mode was excited with the loop. the coupling was not emcient.
possibly due to the loose contact between the outer conductor (18) wall
and the open loop (35). This loose contact problem can not be easily
solved since the loop is rotated and also moved linearly in order to match
the input power into the cavity. Arcing takes place at the loose contacts
at high input power conditions. Hence, loop coupling can not be eaSily
used at high input power conditions which are required for diamond film
growth over larger surface areas.

Compared with loop coupling. probe coupling does not have the
loose contact and arcing problems. Thus it became the favorable choice
for the excitation of plasma for diamond film deposition.

3.5.3 ‘ End Coupling and Side Coupling

An end feed can excite symmetric TM modes and keep the near
field away from the plasma/ substrate region by exciting higher order
modes in diamond film growth process. These are advantages over the
side feed where the probe is located on the cavity side wall and a non-
symmetric field is always produced. At high input power conditions, a
reactor using side feed faces the problem caused by near field if it is to be
operated at lower order modes. It also faces the competing excitation
problem caused by degenerating or closely degenerating dipole TE and

79

TM modes if it is to be operated at higher order modes. By contrast, the
end feed excitation makes the operation of the diamond film deposition

reactor easy and reliable by selective excitation of symmetric TM modes.
It is a favorable choice for the excitation of microwave plasma for uniform

diamond film deposition on larger surface areas.

3.5.4 Mode Excitation

The mode patterns of TMon- T'Mom, T'Mm3, and TECH are shown
in Figures 3.7. 3.10. and 3.12. In side feed probe coupling and loop

coupling. TMOI 1 mode is the easiest mode to be excited but the near
electromagnetic fields attract the plasma to the side walls in the case of
high power input. In end feed probe coupling. TED; 1 mode can not be
easily excited and TMOI 1. TM012 and Til/1013 modes are easily excited. But
in the case of TMOI 1 mode excitation, the near fields attract the plasma
to the quartz disk top wall and limit high power input. The Til/[012 or
mom made is the ideal mode to be used for plasma excitation for
diamond film deposition in the end feed probe coupling configuration.
since the near field effect can be reduced by keeping the probe away from
the quartz disk, and high power can be input to produce a large

cylindrically symmetric plasma.

80
3.6 The Third Generation Seven Inch MCPR

3.6. 1 Introduction

In order to be able to deposit diamond film uniformly on large
surface areas, a new 7 inch microwave plasma disk reactor was designed

and built. It was first tested for diamond film growth in May, 1991.

3.6.2 Reactor Configuration

3.6.2. 1 Introduction

This reactor was designed to operate in various configurations.
These configurations include the basic configuration. forced flow
configuration. jet configuration, lower cavity resonance configuration and
down stream configuration. etc. A detailed description of each
configuration is provided in following sections. Preliminary versions of
these configurations have been investigated for diamond film growth.

3.6.2.2 The Basic Configuration

The principle components of the basic configuration reactor are
displayed in the cross-sectional view of Figure 3.14. This reactor uses the
optimized end feed probe coupling. The power feed is the l 5/ 8 inch
coaxial power input assembly (19, 20). The end feed assembly (19,20)
along with the sliding short (2) form top end of the cavity. The lower
section of the cavity consists of the bottom surface (4), the base-plate

81

 

1 Cavity Walls 23 Gas Input Tunnel

2 Sliding Short 24 Distribution Plate

3 Finger Stock 25 Air Cooling Tunnel

4 Cavity Bottom Surface 26 Cooling Water Tunnel
5 Quartz Dome 27 Laser Port

6 Plasma Discharge 28 Quartz Tube

13 Substrate Holder 29 Metal Tube

14 Substrate 30 Metal Choke

19 Power Coupling Probe 31 Metal Plate

20 Probe Sleeve 43 Base Plate

21 Cooling Water 44 Quartz Flange
22a; Cooling Air Inlet 45 O-ring Seal
22b Window/Air Outlet

Figure 3.14 The Basic Experimental Configuration

82

(43). a metal tube (resonance breaker) (29), a metal plate (3 1) and a metal
choke sleeve (30).

The reactive gases, which are supplied through the source gas
input tunnel (23) and annular source gas ring on the gas distribution
plate (24). is confined at the lower section of the cavity by the quartz
dome (5). The quartz dome (5) is enlarged in both diameter and height.
Water cooling is incorporated in the sliding short (2) and the base plate
(43) by using annular water cooling tunnels (2 I , 26). There is also a air
cooling tunnel (25) built in the base plate (43) for the cooling of the
quartz dome walls under high input power operations. Windows (27) are
incorporated in the base plate for laser induced fluorescence diagnosis of
the microwave plasma under diamond film deposition conditions. A top/
cross-sectional view of the base plate (43) is shown in Figure 3. 15.

The substrate (14) to be coated with diamond film lays on top of a
substrate holder (13) which is supported by a quartz tube (28). Different
height quartz tubes (28) were used to change to the position of the
substrate with respect to the plasma discharge (6). If available, a
mechanical moving stage can be used to change the substrate position.
The metal tube (29) was a 3 inch outside diameter stainless steel tube. It
functioned as a resonance breaker. It ensured that the plasma discharge
(6) stayed on top of the substrate by breaking the cavity resonance and
the discharge excitation condition in the smaller cavity underneath the
substrate. By ensuring that the plasma discharge (6) is only created
above the substrate, the resonance breaker (29) improves the emciency
of the microwave input power. It also improves the reliability and
repeatability of the experimental results and the area on a substrate that
can be uniformly coated with diamond film. The metal choke (30) sleeve

83

27-

 

LEGEND.

$5; Quartz Dame

6 Plasma Discharge

23 Gas Input Tunnel

25 Air Cooling Tunnel
26 Cooling Water Tunnel
27 Laser Port

43 Base Plate

Figure 3.15 The Base-plate of MCPR7-3

84

provides a floating end of the cavity and a choke of the microwave
radiation.

This design minimizes the plasma volume by creating a
hemisphere shaped plasma adjacent to the substrate. The cylindrical
symmetry of the tuning mechanism and system configuration ensures
that the plasma generated has an inherent cylindrical symmetry. This
reactor is mounted on a vacuum chamber with chamber walls and a

chamber outlet leading to a vacuum pump.

3.6.2.3 The Forced Flow Configuration

Figure 3.16 is a schematic drawing of a forced flow configuration.
This configuration is different from the basic configuration by the way
the gas flow passage is controlled. In the basic configuration. the gas flow
is not controlled. it flows naturally inside the quartz dome (5). Whereas
in this forced flow configuration. a metal plate (3 l) and a quartz tube (28)
are put together to force the gas to flow through a flow pattern regulator
(32). The flow pattern regulator (32) is a plate with a series of holes.
directing the way that the gas flows through the plasma and the
substrate. This configuration increases not only the efficiency of the
source gas but also the uniformity of the coating by influencing the
shape of the plasma discharge (6) through changing the flow pattern into
the plasma discharge (6).

Figure 3. 17 (a) shows one example of flow regulator. where a
number of smaller substrates are being coated. Variation of input power.
pressure and flow rates along with the flow pattern regulator produces
uniform plasma over the smaller substrates and ensures uniform

85

 

LEE!!!

1 Cavity Walls 22b) Window/Air Outlet
2 Sliding Short 23 Gas Input Tunnel
3 Finger Stock 24 Distribution Plate
4 Cavity Bottom Surface 25 Air Cooling Tunnel
5 Quartz Dome 26 Cooling Water Tunnel
6 Plasma Discharge 27 Laser Port

1 Substrate Holder 28 Quartz Tube

14 Substrate 29 Metal Tube

19 Power Coupling Probe 31 Metal Plate

20 Probe Sleeve 32 Flow Regulator

21 Cooling Water 43 Base Plate

229) 900“"9 N? W“ 80 Optical Pyrometer

Figure 3. 16 The Forced Flow Experimental Configuration

86

 

9:" ' e e

’e’e’e 'e’e’o’e’e
e e 9 >5 0 e e e

50.0.5033;

‘ O ' ...

 

 

 

(b)

 

Figure 3.17 Flow Pattern Regulators

87

coatings on them. Figure 3. 17 (b) shows another example of flow pattern
regulator. where a large piece of substrate is being coated. Variation of
input power. pressure and flow rates along with the flow pattern '
regulator produces a uniform plasma over the substrate so that a

uniform diamond coating can be obtained.
3.6.2.4 The Jet Configuration

Figure 3.18 is a schematic drawing of a jet configuration where the
substrate (14) and substrate holder (13) are located in a region separated
from the plasma discharge (6). The metal plate (3 l) and quartz tube (28)
are put together to force the gas flow through the jet grid (33) which is an
electrically conducting plate with a series of holes. The size of the holes.
the flow rate and the vacuum pump’s pumping speed determine the
pressure difi'erence between the plasma discharge (6) region and
chamber region wherein the substrate (14) is located. In this
configuration. the reactive species are forced to flow at a higher speed
over selected areas so that the deposition rates at the selected areas are

increased.
3.6.2.5 The Lower Cavity Resonance Configuration

Figure 3.19 shows a lower cavity resonance configuration where
the substrate (14) is located at the lower section of the cavity without the
use of metal resonance breaker (29) shown in Figure 3.14. This was the
first configuration used for diamond film growth in the third generation
MCPR.

88

lflflllllflllflllflflllllllllll

 

CD

31
LEGEND
1 Cavity Walls (22b) Window/Air Outlet
2 Sliding Short 23 Gas Input Tunnel
3 Finger Stock 24 Distribution Plate
4 Cawty Bottom Surface 25 Air Cooling Tunnel
5 Quartz Dome 26 Cooling Water Tunnel
6 Plasma Discharge 27 Laser Port
13 Substrate Holder 28 Quartz Tube
14 Substrate 29 Metal Tube
19 Power Coupling Probe 31 Metal Plate
20 Probe Sleeve 33 Jet Grid
21 Cooling Water 43 Base Plate

22a) Cooling Air Inlet

Figure 3. 18 Jet Experimental Configuration

89

 

LEGEND

1 Cavity Walls 22a Cooling Air Inlet
2 Sliding Short 22b Window/Air Outlet
3 Finger Stock 23 Gas Input Tunnel
4 Cavity Bottom Surface 24 Distribution Plate
5 Quartz Dome 25 Air Cooling Tunnel
6 Plasma Discharge 26 Cooling Water Tunnel
1 Substrate Holder 27 Laser Port

14 Substrate 28 Quartz Tube

19 Power Coupling Probe 30 Metal Choke

20 Probe Sleeve 31 Metal Plate

21 Cooling Water 43 Base Plate

Figure 3. 19 Lower Cavity Resonance Configuration

90

3.6.2.6 The Down-stream Configuration

Figure 3.20 shows a down-stream configuration where the
substrate (14) is located outside the cavity and plasma discharge (6)
region.

The substrate can be biased with a voltage during the deposition.
The temperature of the substrate can be adjusted by a heater or cooling
tunnels. Permanent magnets can be placed in the base-plate (43) around
the plasma discharge (6) region to enhance plasma formation at low

deposition pressures.

3.6.3 Reactor Operation

The reactor operation of the third generation 7" MCPR (MCPR7-3)
is similar to that of MCPR7-2. The main difference is that TM013 mode is
used in later part of the experiments conducted in MCPR7-3 where the
cavity length. Ls. was set at approximately 21.6 cm and the probe length.
Lp. was adjusted to about 3.2 cm.

3.6.4 Reactor Performance
3.6.4. 1 Introduction

Preliminary versions of the five operating configurations have been
investigated for diamond film growth. A summary of the reactor

performance in the preliminary versions of is given below. Detailed
description of reactor performance in the force flow configuration is given

91

 

34
LEM 23
1 Cavity Walls '22o Cooling Air Inlet
2 Sliding Short 22!: Window/Air Outlet
3 Fin er Stock 23 Gas Input Tunnel
4 Cavity Bottom Surface 24 Distribution Plate
5 Quartz Demo 25 Air Cooling Tunnel
6 Plasma Discharge 26 Cooling Water Tunnel
13 Substrate Holder 27 Laser Port
14 Substrate 28 Quartz Tube
19 Power Coupling Probe 31 NOW Plate
20 Probe Sleeve 34 Heater
21 Cooling Water 43 Base Plate

Figure 3.20 Down-stream Experimental Configuration

92

in chapter 6.

3.6.4.2 The Basic Configuration

Diamond film was deposited on 4" silicon wafer under the following
experimental conditions:
H2 gas flow = 300 sccm
CH4 gas flow = 3 seem
Pressure = 20 Torr
MW power absorbed = 2 kW
Seeding: scratched with 1 pm diamond powder.

The resultant diamond film deposition rate was slow (~ 3 mg/ hr).
possibly due to low operating pressure and power. Both graphite and
amorphous carbon substrate holders were used in diamond film
deposition processes in this configuration. It was found that higher film
growth rates were obtained with the amorphous carbon substrate
holders. possibly due to easier carbon dissolution from the amorphous
carbon holder into the reactive gas mixture. The amorphous carbon
holders were distorted in the high temperature deposition processes.

3.6.4.3 The Forced Flow Configuration

Using the forced flow configuration. a 12 cm diameter discharge
was created with 200 sccm hydrogen and 2 seem methane gas flow at 40
Torr with 4 kW 2.45 GHz absorbed power. The discharge area produced
is larger than any discharge area reported in the open literature. The
discharge created had a symmetric. hemisphere shape over the substrate

93

surface. Diamond films have been deposited uniformly on 4 inch
diameter silicon wafers. The uniform coated area is larger than any
coated area reported in the open literature. A detailed description of the
reactor performance in this configuration is provided in chapter 6.

3.6.4.4 The Jet Configuration

A preliminary version of this configuration was used for diamond
film deposition. It is schematically shown in Figure 3.21. In this
preliminary version. gas was forced to flow through the center hole on
the graphite grid. The center hole is 0.5 inch in diameter and the
substrate was located approximately 1 cm below the graphite grid.
Without using any cooling mechanism. increased deposition rate (~ 1
pm] hr) over 1.5 cm diameter area had been achieved. Typical
experimental conditions are as follows:

H2 flow rate = 400 sccm

CH4 flow rate = 6 sccm

C02 flow rate = 2 sccm

pressure = 95 Torr

absorbed microwave power = 1.67 kW

The deposition rate was limited because of substrate melting at
high pressure and temperature conditions. It is expected that by adding
cooling tunnels underneath the substrate to control the substrate
temperature. higher pressure and power can be used to further increase
the deposition rates.

94

 

LEEEIEIQ

1 Cavity Walls (22b) Window/Air Outlet
2 Sliding Short 23 Gas Input Tunnel
3 Finger Stock 24 Distribution Plate
4 Cavity Bottom Surface 25 Air Cooling Tunnel
5 Quart: Dome 26 Cooling Water Tunnel
6 Plasma Discharge 27 Loser Port

13 Substrate Holder 28 Quartz Tube

14 Substrate 29 Metal Tube

19 Power Coupling Probe 31 Metal Plate

20 Probe Sleeve 33 Jet Grid

21 Cooling Water (43 Base Plate
22a) Cooling Air Inlet

Figure 3.2 1 Preliminary Jet Experimental Configuration

95
3.6.4.5 The Lower Cavity Resonance Configuration

Diamond film was deposited on a 3 inch silicon wafer under the
following experimental conditions:
H2 gas flow = 200 seem
CH4 gas flow = 2 seem
Pressure = 30 Torr
MW power absorbed = 2 kW
The resultant diamond film was fairly uniform. The deposition rate
was ~ 3.56 mg/hr. It was found that a plasma discharge (6) generated in
this configuration had to compete with a discharge generation process
underneath the substrate holder (13). The competing process made the
starting procedure complicated and unreliable.

3.6.4.6 The Down-stream Configuration

A preliminary experiment with this configuration was conducted in
which a 3 inch substrate heater (34)<6°> was used to control the
substrate temperature.

A schematic drawing of the heater is shown in Figure 3.22.<6°’ As
shown. it is a resistant heater and consists of three layers. a top boron-
nitride (BN) (46). a pyrolytic graphite (47) and a lower boron-nitride (BN)
(48) layer where the pyrolytic graphite layer is the middle layer. The
exposed pyrolytic graphite areas (49) at the two ends are electrical
contacts. This heater was specified to meet the following heating
requirements. namely, it (a) has a 3" heating zone. (b) can achieve

temperatures up to 1 500 °C. (c) can achieve temperature uniformity of :t

96

 

Heater
46 Top BN layer
47 Pyrolytic Graphite
48 Lower BN layer
49 Contacts

Figure 3.22 Schematic Drawing of a 3" Substrate Heater

97

10 °C over the heating zone. (d) operates with 60 Hz. 0 - 200 Volts AC
power. (e) draws a maximum current of 15 A and (i) can be used at
pressures from 1 mTorr to 100 Torr in H2/CH4 gas mixture environment.
The experimental conditions were:
H2 gas flow = 400 seem
CH4 gas flow = 4 seem
Pressure = 15 Torr
MW power absorbed = 1.5 kW
Heater voltage = 100 Volts. 60 Hz
The resultant film growth rate was low and the film uniformity was

poor.

3.7 Comparison of Seven Inch MCPRs

This third generation MCPR offer advantages over the reactors
described in chapter two and the first two generations of MCPRs in that
it creates a large symmetric, disk to hemisphere shaped discharge for
uniform diamond film deposition over large surface areas. The symmetry
of the plasma discharge is ensured by the symmetric. end feed power
input configuration. The power input system. excitation probe and probe
sleeve. have large dimensions so that high microwave power (2 to 5
kilowatts) can be delivered into the cavity. High input power. law near
field efi'ects. a large quartz disk and the design configuration of keeping
the vacuum seal away from the heated areas ensure that a large
diameter plasma can be created. The equator plane of the hemispherical
shaped discharge is created and defined by the cavity bottom surface
with the help of a metallic step near the outside surface of the quartz

98

disk and the lower cavity resonance breaker. The quartz disk ensures
only the plasma adjacent to the substrate is produced and the efficiency
of the absorbed power is maximized. The substrate is located near the
equator of the hemisphere. hence the spacial uniformity of the
distribution of reactive species and the thickness of deposited diamond
film is ensured.

This reactor employs some of the basic features of the conventional
microwave-cavity plasma reactors (MCPR). They are: (l) cylindrical
cavity. (2) internal matching. i.e., sliding short and variable probe; (3)
excitation of T'Mmn modes for deposition of films. However. there are
some important difi'erences between the MCPR7-3 and the conventional
microwave reactors. They are: (I) the probe is mounted on the sliding
short. not on the cavity side walls. thus all the internal cavity
adjustments take place on the sliding short: (2) the probe is located at
the center'of the short and thus only TM modes are easily excited; (3) the
probe is located far away from the discharge region so that the probe
near field does not interfere with the mom made electromagnetic fields
within the discharge region; (4) larger diameter quartz disk can be used
and (5) much higher microwave power can be delivered to the reactor to
create a larger diameter plasma: (6) with the help of the lower cavity
resonance breaker. the cavity bottom surface and the metallic step
defines the equator of the symmetric. disk-like or hemisphere shaped
plasma when the cavity is excited in T'Mmn modes: (7) the substrate
holder is adjustable in that it can be moved up and down independently.
this feature together with the independent sliding short and excitation
probe movement allow the movement of the relative position of the

substrate with respect to the plasma so that the optimum deposition

99

conditions can be reached: (8) the forced reactive gas flow through the
plasma near the substrate improves the efficiency of the reactive gases.
and the uniformity of the deposited film; (9) film deposition down stream
from the plasma; (10) independent and selected area substrate heating or
cooling increases the deposition areas and film uniformity.

3.8 Fourteen and Eighteen Inch MCPRs

3.8. 1 Introduction

The feasibility of diamond film growth over larger surface areas
using larger diameter cavity MCPRs was investigated using 14" and 18"
side feed MCPRs with 2.45 GHz and 915 MHz power. The test results are
described in sections 3.8.2 and 3.8.3.

3.8.2 Fourteen Inch MCPR

The feasibility of diamond film growth using a side power feed 14"
MCPR<61> and 2.45 GHz power was tested using the Chuang 1.2 kW
2.45 GHz power source. The schematic drawing of the experimental set-
ups are shown in Figure 3.23 for substrate supported by a quartz tube
and Figure 3.24 for substrate supported by a metal tube. As shown.
these schematic drawings are similar to that shown in Figure 3.5. The
main difi'erences are that (a) the cavity inside diameter is now 14". (b) the
vacuum seal is placed away from the discharge zone and (c) the quartz
dome is 10" in diameter and 3.5” in height.

It was possible to couple microwave power into the cavity and a

§

Sliding Short

f

7/////////////////////////////////////////////////////////////////////////////////////////////.

W

 

 

’///////////////////////////////////l

‘ 14.0" Microwave Cavity

7
Z
4

Quartz Dome

Obe . ///////////////////A

KW. \

Substrate
& Holder

 

 

. ’//////////////////////////////////////////////////////////////////////////l

7 ’////////////////////////l

//.,

\\\\\\\

    

 

 

 

Quartz TUbC —.»

. . . . O .
z e e e e e’e'e’e e’e’e’e’e’e’e’e’e’e’e’e :e' 'e' e:e:e'e:e'e’ eeeeeeeeeeeeeeeeeeeeeeeeeeeee
eeee 'e‘: e‘e. e. e e‘e‘e'e‘e‘e'e'e‘e’e :' ‘ '

 

 

OOOOOOOOOOOOOOOOOOOOO
......................................

llllllllllllllllllllllllilllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllll lllllllllllllllllllllllllllllllllllllllllllllIllllllllllllllllllllllllllllllllllllllllllll.ll

Perforated
Metal Screen

 

 

 

 

 

 

 

\V/

To Vacuum Pump

Figure 3.23 Substrate Supported by a Quartz Tube in 14" Cavity

Sliding Short

s\\\\\\\\‘

7/////////////////////////////////////////////////////////////////////////////////////////////.

14.0" Microwave Cavity

 

 

' '///////////////////////////////////1

/
f
2

Probe —>
Quartz Dome

 

/
Z
2

 

Plasma

- Substrate

& holder

7/////////////////////////////////////////////////////////////////////////1

/////////////////////////4 '

 

\\\\\\‘

I
g
/
4

s

    

b .

 

IIIIIIII’IIIIIIIII’

\
E::.. p . \\‘\\ «M Support \\

''''''''' ' °°°°°°° \\\\\\\\\\\\\\\\\\\\\\\\ :zzszzzzzzstztzzzzztzz:sztztzzzzezezizaé
lIillllllllllllllllllllll||||||lllllllllllllllllllllIlllllllllllllllllllllllllllllllllllllllll IlllllllllIlllllllllllllllllllllllllllllllllllll||llll'lllllllllllllllllllll lllllllllllllll

 

 

 

%//j

 

 

Perforated
Metal Screen

To Vacuum Pump

Figure 3.24 Substrate Supported by a Metal Tube in 14" Cavity

102

plasma disk can be produced. The plasma disk was approximately-3 inch
in diameter at 20 Torr with 700 Watts power input. The plasma disk was
suspended above the substrate surface.

From the resonant cavity lengths for the TM modes displayed in
Figure 3.25 and Table 3.1. it is expected that mom mode can be excited
in a 14 inch cavity without much interference from the other TM modes.
With an end feed and a properly designed base plate. it is expected that
discharge loaded mom mode can be excited in a 14" MCPR. Diamond
film may be deposited in this reactor with a substrate moving stage and
sumcient microwave power (3 - 15 kW).

3.8.3 Eighteen Inch MCPR

The concept of plasma discharge generation over larger surface
area using 915 MHz power in a larger cavity was proposed by J.
Asmussen.

To test this concept. an 18" MCPR<62> with side feed coupling was
used with a 500 W. 915 MHz power supply. The schematic drawing of
the 18" MCPR is similar to that shown in Figure 3.5. The cavity inside
diameter is 18". The side feed probe sleeve has an inside diameter of 1 5/
8" and the coupling probe has an outside diameter of 5 / 8".

After a series of experiments. it was found that the plasma
discharge loaded TMOI 1 mode can be easily excited. But the available
power level was too low to be used for actual diamond film growth over a
significant area.

The feasibility of diamond film growth using the 18“ MCPR and
2.45 GHz power was tested using the Chuang 2.45 GHz 1.2 kW power

103

5.58 .3 s 882 5:8 £65. 338 35888 and 8&2

“£8 58...... team 338

 

on mm mm em NN ON a m- 3 N— o— m m r
p _ . _ . _ . _. . h . _ . _ . _ u L . _ . _ u h u
o O O
86:; was: 2.25 1
O O O .I
.255 5:5 3:8
o o o 1
$955 was: 355
38.5 38:: :82: r
O O 0
an 55 a 55 c Sac
O O O 1
855 a was: o case
535 9.25 I
O
a CNN}...
$8.5 ..

 

 

 

0.0

Cd

0+

(d‘UIx

ad

0d

0.0—

 

 

104

 

Table 3.1 Resonant Cavity Lengths of TM Modes in 14" Cavity

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

X(n.p) Cavity Mode Cavity Length (cm)
2.405000 (TMOI 1) 6.34694 1
2.405000 (IM012) 12.69388
2.405000 (TM013) 19.04082
3.832000 (TM 1 1 1) 6.746388
3.832000 (TM 1 12) 13.49278
3.832000 (I'M 1 13) 20.239 16
5.136000 (TM21 1) 7.407754
5.136000 (TM212) 14.81551
5. 136000 (TMZ 13) 22.22326
5.520000 (T‘MOZ 1) 7.689620
5.520000 (TMO22) 15.37924
5.520000 (TMO23) 23.06886
6.380000 (FMS 1 1) 8.564938
6.380000 (TM3 12) 17. 12988
6.380000 (M3 13) 25.6948 1
7.016000 (’I‘MlZl) 9.577787
7.016000 (IM122) 19.15557
7.016000 (I'll/1123) 28.73336
7 .588000 (TM41 1) 1 1 .02720
7 .588000 (TM4 12) 22.0544 1
7.588000 (TM413) 33.08161
8.417000 (I'll/I22 I) 15.86860
8.417000 (I‘MZ22) 3 I .7372 1
8.771000 W51 1) 22.24383
8.771000 (TM5 12) 44.48766

 

 

 

 

105

source. It was found that it was difficult to couple the 2.45 GHz power
into the cavity efficiently to create a centered plasma. This could be due
to the large number of degenerate and nearly degenerate modes present.
The instability of the power source may have also contributed to this
inefficient power coupling problem.

CHAPTER FOUR

THE MICROWAVE-CAVITY JET REACTORS (MCJR)

4.1 Introduction

Diamond film deposition with high growth rates is desirable in
many applications. One method to increase the growth rate is to increase
the reactive species concentrations over the growth surface. The
microwave cavity jet reactor (MCJR) is one type of plasma reactor that
may be able to deposit diamond films with high growth rates. In the
microwave cavity jet reactor. the excited species from the microwave
discharge are directed through a nozzle onto a substrate. Since the
microwave discharge can be maintained at high pressure (100 Torr to 1
atmosphere). the number of active species available for diamond film
growth is high. Thus higher growth rates are expected. One advantage
that the MCJRs have over some other types of jet reactors, such as DC
are jet reactors. is that MCJRs do not have electrodes. thus eliminating
the possible contamination from the electrodes. The MCJR was originally
developed at MSU as microwave electrothermal thrusters< 10' 1 1’ .

In this chapter. three microwave cavity jet reactors. the first.
second and third generations of microwave cavity jet reactors. MCJR- 1 .
MCJR-2 and MCJR-3. are described. The MCJR-1 and MCJR-2 have
been developed and investigated for diamond film growth. The MCJR-3

106

107

was designed but has not been built due to lack of funding.

The experiments described in this chapter are preliminary
experiments in investigating the potential of microwave cavity jet reactors
in diamond film deposition. Diamond films have been deposited with a
difi'erent type of microwave jet reactor‘37°38’ which was reviewed in
chapter 2. Linear growth rates in the order of 12 rim/hour have been
reported with that reactor.<37'38>

4.2 Experimental Systems

4.2. 1 Introduction

All the microwave cavity jet reactors that are described in this
dissertation use microwave power as their energr source and vacuum
pumping system is used for reactive gas supply and operating pressure
control. The microwave power supply. gas flow and vacuum pump
system used in the MCJRs are described below in Sections 4.2.2 and
4.2.3.

4.2.2 Microwave Power Delivery System

The microwave circuit used to deliver microwave power to the jet
reactors is similar to that described in Section 3.2.2. The only difi'erence
was the microwave power source that was used. In the experiments
described in this chapter. a Thermex (Model No. 4074) 2.45 GHz. 2.5 kW

power source was used.

L.

108

4.2.3 Flow Control and Vacuum Pump Systems

A schematic drawing of the gas flow control and vacuum pump
system is shown in Figure 4.1. As shown. the reactive gases. high purity
(99.99% or better) H2 and CH4. are supplied by cylinder gas tanks (1)
and (2). The gas flow rates are measured by mass flow meters (3) and
controlled by needle valves (4). The flow meters measure flow rates with
full scale flow ranges of 10 seem (CH4) and 500 seem (H2). The reactive
gases are combined and mixed and as they flow into the quartz tube (5)
which extends into the microwave cavity applicator (6). They flow
through the cavity applicator (6) and the base plate (7) into the process
chamber (8). A mechanical roughing pump (10) pumps the reactive gases
through a manual valve (9) and out into the exhaust (1 1). The flow rate of
purge gas supplied by nitrogen gas cylinders (12) is measured by a tube
flow meter (13) and controlled by needle valves (14). It is added to the
exhaust (1 1) to dilute the flammable reactive gases.

The pressure in the process chamber region is measured by a
Hastings thermal conductivity vacuum gauge (15) with meaningful
pressure measurement range of 5 - 200 mTorr and a MKS type 220B
10.000 Torr full scale baratron pressure gauge (16). The thermal
conductivity vacuum gauge (1 5) was used to monitor chamber base
pressure and the MKS pressure gauge was used to measure the process
pressure. The chamber pressure is controlled by the gas flow rates and
the manual throttle valve.

The instrumentation feed through port (17) on top of the process
chamber is used for heater power supply and thermal couple electrical
signal transmission. With this instrumentation feed through. the

(17) Instrumentation

Feed
Through
Port

 

 

  
   
 
   

109

(18) Vacuum

 

 
    
  

(19) View (9) Manual Valve

Feed
Through
Port

 

*Port '

 

  
 

 

 

 

 

(10) Mechanical
Roughing
Pump

 

 

 

 

 

 

 

V
i
\
s
t
i
g l__l
§ (21)'IShu:epension
§ (16) Baratron
9 Pressure
8 Gauge
3 _
Substrate
(15) Thermal
Conductivity
(7) Base Vacuum Gauge
Plate
(6) Microwave “’ (5) Q‘mz Tube
cavity —%—a

Applicator

L

 

 

 

(4)Needle (3) Flow
Valves Meters

 

 

 

Figure 4.1 Flow Control and Vacuum System for MCJRs

110

temperature of the substrate can be measured and an independent
substrate heater can be used to vary the substrate temperature.The
vacuum feed through port (18) is used to suspend the substrate (20) and
change its position via the suspension tube (21). The downward facing
substrate (20) is mounted on the lower end of the suspension tube (21).
The upper end of the suspension tube (2 1) is a half inch stainless steel
tube which can be sealed by a half inch ultra-torr compressed o-ring
vacuum fitting (18) welded to the center of the process chamber (8) top
surface. When cooling liquid or gas flows inside a hollow suspension tube
(21). the substrate (20) can be cooled. Part of the-process chamber can be
viewed through the view port (19).

4.3 The First Generation MCJR
4.3. 1 Introduction

The first generation MCJR (MCJR- 1) was the first jet reactor
investigated for diamond film growth. It was developed in 1988 and was
built from an existing microwave cavity and an existing stainless steel
chamber.

4.3.2 Reactor Geometry

A schematic drawing of this reactor is shown in Figure 4.2. As
shown. This reactor has a cylindrical side-wall (1) which forms the outer
conducting shell of the cavity applicator. A water-cooled (2 1) sliding short
(2) forms the lower end of the cavity applicator. The sliding short (2) can

111

First Generation MCJR (MCJR—1)

 

 

LEGEND

E1 Cavity Walls 13 Linear Feed Through
2 Sliding Short 14 O-ring Seal

3 Finger Stock 21 Cooling Water

4 Cavrty Top Surface 22 Brass Top Plate
5 Quartz Tube 25 Cooling Water

6 Plasma Discharge 26 Window/Air Outlet
7 Gas Input 27 Cooing Air Inlet
9 Outer Conductor 28 Process Chamber
10 Power Input Probe 29 Ultra-Torr Fitting
11 Substrate 3O Brass Collar

12 Substrate Holder

 

Figure 4.2 Schematic Drawing of MCJR-1

112

be moved along the cylindrical axis of the cavity wall (1) to change the
position of the lower end of the cavity applicator. A water cooled (25)
brass top plate (22) with the cavity top surface (4) form the top end of the
cavity applicator. The brass top plate (22) and the cavity wall (1) are
soldered together. Microwave energy is coupled into the applicator
through the power input probe (10) and its coaxial outer conductor (9).

Reactive gas mixture enters the cavity applicator from the gas
input port (7) where it is contained inside the 1" outside diameter quartz
tube (5). The gas mixture flows upward. it is heated. ionized and
dissociated inside the cavity applicator. The microwave discharge then
flows over and around the substrate (1 1) and substrate holder (12) into
the stainless steel process chamber (28) after which it is pumped out by
a vacuum pump.

The “ultra-torr" vacuum fitting (29) near the bottom of the stainless
steel process chamber (28) serves as a vacuum seal port to the quartz
tube through the compressed o-ring seal (14). The quartz tube inner
space near the “ultra-torr" vacuum fitting (29) functions as the nozzle for
this jet reactor. T‘wo inch diameter openings exist at the center of the
sliding short (2) and the brass top plate (22) so that quartz tubes of less
than 2 inch in outside diameter can extend through the center of the
microwave applicator. Brass collars (30) are used to reduce the size of
the opening to fit smaller quartz tubes and to reduce the microwave
radiation leakage from the cavity applicator. The “ultra-tori" vacuum
fitting (29) on the stainless steel chamber is an 1 inch ultra-torr fitting.
hence quartz tubes (5) of 1 inch in outside diameter and brass collars
(30) of 1 inch in inside diameter were used when experimenting with this

reactor concept.

113

Besides being cooled by the water cooling tubing (21) on the sliding
short (2) and the water cooling tubing (25) on the brass top plate (22).
this reactor is also cooled by cooling air. The cooling air entered the
cavity applicator through the cooling air inlet (27) and exited through the
screened window/ air outlet (26) which also served as a viewing port.

4.3.3 Reactor Operation

The experimental operation of this MCJR has been described in
detail else where “0:115:35”. Thus. only a brief description is given
here. Difi'erences from earlier work (10' 1 15354) are operation at lower
pressure (30 - 80 Torr) and lower flow rates (100 - 400 seem). The
ignition. matching. and internal cavity tuning are similar to that reported
earlier for microwave electrothermal thrusters < l 1’.

The mode diagram of an ideal 7" cavity is shown in Figure 3.6. This
MCJR can create a microwave discharge when excited in a single cavity
electromagnetic mode. A discharge was started by reducing the H2/CH4
gas pressure to 5 - 1 0 Torr, and applying microwave power which would
then ignite a discharge that filled the quartz tube. The discharge
pressure was then increased to an operating condition of 30 - 80 Torr
while adjusting the length. L5. and the probe. 11" tuning the cavity to a
matched condition.

When the sliding short length L.1 is adjusted to about 7.45 cm and
the probe length Lp to 8. 1 mm. the discharge loaded mm 1 mode is
excited. A schematic drawing of the field pattern of the TMOI 1 mode in a
discharge loaded cavity is shown in Figure 3.7. As shown in Figure 4.2.
with the help of the brass collar (30). the substrate (1 1) and substrate

114

holder (12). only the plasma near the top of the cavity applicator is
excited. With this excitation configuration. the plasma is brought close to
the substrate so that the reactive species can emciently flow over the
substrate. Deposition experiments were performed with this excitation
configuration.

Comparing this microwave plasma reactor with the other
microwave cavity plasma reactors (MCPR) described in chapter 3. the
structure of this reactor has interesting features. Referring to Figure 4.2.
the MCJR is mounted upside down compared with MCPR. This idea was
brought up by J. Asmussen in 1986. In this MCJR. the reactive gases
flow upward and are heated and dissociated by absorbing microwave
energy in the cavity applicator. The reactive species generated in a hot
discharge (6) naturally flows upward towards the substrate (1 1). This is
due to the fact that under the same pressure and for the same gas.
higher temperature gas has a lower density and tends to flow naturally
upward in a cooler. higher density gas environment. Therefore. reactive
species are effectively utilized for diamond film growth. Another
interesting feature about this configuration is that the input reactive
gases are drawn toward the hot discharge (6) when the reactive species
generated in the hot discharge (6) flows up and around the substrate (1 1)
toward the pump. Hence. the input reactive gases are emciently utilized
for diamond film growth in this mounting and gas flow configuration.

4.3.4 Reactor Performance

The first generation MCJR has been tested in diamond film growth
uperiments from September. 1988 to February. 1989 in more than 1 7

115

experiments. When the substrates (1 I) were lowered until they were in
contact with the discharge (6). diamond films were grown on the
substrates. The substrates used were typically 1 cm x 1 cm silicon wafers
and silicon nitride.
A typical experiment (DTF04) was conducted under the following

experimental conditions:

CH4 flow rate = 1 seem

H2 flow rate = 120 sccm

pressure = 50 Torr

cavity short length 1.8 = 7 .45 cm

probe length L1) = 8.1 mm

absorbed microwave power = 300 Watts

There are a few limitations in the first generation MCJR. First. it is

not a “true” jet reactor since the “nozzle” diameter is the inside diameter
of the quartz tube (5). which is close to 1 inch. Second. because of the
long “neck” in the ultra-torr fitting (29). the discharge (6) is usually
located far below the open area of the stainless steel chamber and the
substrate (1 1) is usually lowered into the “neck" area for diamond film
deposition. This configuration puts a limit on the size of the substrate to
be smaller than the inner area of the quartz tube. which in the present
reactor is less than 1” in diameter. Third. there are severe limitations on
the input power level and operating pressure in the operation of this
reactor. Since the discharge is in close contact with the quartz walls
which has a melting temperature in the same range as the temperature
of reactive gases in a discharge (6) suitable for diamond film growth. the
deposition pressure was limited to less thanlOO Torr and the power
input was limited to less than 500 Watts so that the quartz tube walls

1 16
would not melt. And finally. the experimental reproducibility in this

reactor was poor because materials were deposited on the quartz tube

walls.
4.4 The Second Generation MCJR
4.4. 1 Introduction

To reduce the nozzle size and scale up the substrate area, the
second generation jet reactor (MCJR-2) was developed in 1989. New top
end plate of the cavity applicator and lower end plate of the stainless
chamber were designed and built. Diamond film deposition experiments

were conducted with this reactor.
4.4.2 Reactor Geometry

A schematic drawing of this reactor is shown in Figure 4.3. As
shown, the new end plate (23) of the cavity applicator is machined from a
brass plate. It has an 1 . 1 inch diameter opening at the center. A water
cooling tunnel (16) is incorporated in the plate which is located close to
the discharge (6) zone so that emcient cooling can be achieved. The 0-
ring spacer (18) functions primarily as a shock resist mechanical support
to the quartz plate (32). A new cavity shell which does not have an end
plate soldered onto it is used as the cavity walls (1). Finger stock (8) is
used to electrically connect the brass end plate (23) to the cavity walls
(1).

The lower end of the stainless steel chamber which was used in

 

117

Second Generation MCJR (MCJR-2)

 

 

 

 

 

 

 

 

 

 

 

 

 

26--5i
LEGEND
1 Cavity Walls 16 Cooling Water
2 Sliding Short 17 Cooling Water
3 Finger Stock 18 O-ring Spacer
4 Cavity Top Surface 19 Jet Nozzle
5 Quartz Tube 21 Cooling Water
5 Plasma Discharge 23 Brass End Plate
7 Gas Input 26 Window/Air Outlet
8 Finger Stock 27 Cooling Air Inlet
9 Outer Conductor 28 Process Chamber
10 Power Input Probe 31 5.5. Efld Plate
11 Substrate 32 Quartz Flan e.
12 Substrate Holder 33 Ultra-Torr Itting
13 Linear Feed Through 34 Outer Quartz Tube
15 O-ring Seal 35 Cooling Gas Inlet

 

Figure 4.3 Schematic Drawing of MCJR-2

118

MCJR-1 was cut ofi‘ and a Varian 8 inch diameter stainless steel Conflat
flange with 6 inch opening near the center was welded onto the stainless
steel chamber. This welded plate serves as a vacuum bolt joint between
the stainless steel chamber (28) and the stainless steel end plate (31).
The stainless steel end plate (3 l) is made from a Varian 8 inch stainless
steel Conflat blank flange. There is a 3.8 inch opening near the center so
that substrates up to 3.8 inch in diameter can be used in the stainless
steel chamber. A water cooling tunnel (l 7) is incorporated in the plate for
the chamber cooling. The O-ring seal (15) on this plate serves as the
vacuum seal between the stainless steel end plate (3 l) and the quartz
plate (32).

The jet assembly is made of an one inch quartz tube (5) welded to
the center of a quartz plate (32). which is 5.5 inches in diameter and
0.25 inch thick. A 1/32 inch diameter hole was drilled at the center of
the quartz plate (32). It serves as the jet nozzle (19). The one inch quartz
tube (5) passes through the cavity applicator while the quartz plate (32)
is sandwiched between the brass end plate (23) and stainless steel end
plate (3 1). Ultra-torr fitting (33) seals the lower end of the quartz tube
and makes the gas flow transition from stainless steel tubing to the
quartz tube (5). On the outside of the 1 inch quartz tube (5). another
quartz tube (34) of 1.5 inch in inside diameter is used to transport the
safety and cooling N 2 gas to the discharge (6) region. This outer quartz
tube (34) is sealed to the inner quartz tube (5) near the lower end of the
sliding short (2) and left open near the discharge region (6). N2 gas enters
this outer quartz tube (34) from the gas inlet (35) and exits into the cavity
applicator near the discharge zone (6). This double tube cooling
mechanism is efficient since it forces the cooling gas to flow close to the

119

hot discharge zone (6).

4.4.3 Reactor Performance

This second generation MCJR has been investigated for diamond
film growth from November. 1989 to January. 1990. More than 17
experiments were performed with this reactor. Diamond-like films have
been grown on silicon substrates. The operation of this reactor is similar
to that of the MCJR-l.
The Raman spectrum of a typical film (JDF- 10) deposited under
the following experimental conditions is shown in Figure 4.4:
CH4 flow rate = 3.5 sccm
H2 flow rate = 504 sccm
Ar flow rate = 252 sccm
pressure = 84 Torr
cavity short length L5 = 7.12 cm
probe length Lp = 3 cm
absorbed microwave power = 350 W

The discharge generated in this reactor was almost in direct
contact with the quartz tube walls. The pressure was limited to 100 Torr
and the power input was limited to 500 Watts so that the quartz tube
walls would not melt. It was not possible to generate a high temperature
discharge to pass through the nozzle (19) to achieve the substrate (1 1)
temperature required for diamond film growth. The reproducibility was
poor because materials were deposited on the quartz tube walls.

120

«£32 5 essence SE a ea 5.58% 58mm 3 «has.

20 13001.”: out

 

      

 

 

 

o.o«m« 0.0mm... o.omw« oo.onm 00.0mm oo.owv
b p p p p p b b b n n p p p p p p P W b W p p p — ho . mom
H B omm "am.
1 50 m n _ .
00 Cum
. .5 a: u 9. n
. t8. vw n a .A
w Eoow NmN n a i
. :88 «om u a: nil .
...I Boom Dd u «20 4 oo CNN
I..L co. 6mm
4
I... oo . ovm
.I... p p P L L p p a p m p L p m m - wl bl m‘ l- »i b p - o . omo.“

 

 

ALISNELNI

121

4.5 The Third Generation MCJR

4.5. 1 Introduction

In order to keep the quartz walls away from the hot plasma
discharge zone. the third generation MCJR (MCJR-3) was designed. With
this design. the quartz wall erosion and reproducibility problems faced by
the previous two generations of MCJR may be solved. Much higher power
input and higher operating pressure experiments may be conducted so
that higher diamond film growth rates may be achieved.

4.5.2 Reactor Geometry

A schematic drawing of this reactor is shown in Figure 4.5. As
shown. new brass end plate (24) of the cavity applicator and new end
plate (31) of the stainless steel chamber are designed.

The brass end plate (24) has a 4.62 inch opening near the center.
Water cooling tunnels (16) are incorporated in this plate. O-ring spacer
(18) is used to support the quartz flange (32). Finger stocks (8) are used
to electrically connect the brass end plate (24) to the cavity walls (1).

The stainless steel end plate (31) is designed to be machined from
a Varian 8 inch Conflat blank flange. It has a 0.5 inch opening near the
center. A series of stainless steel nozzle inserts with difi'erent nozzle sizes
are designed to be fit into the 0.5 inch opening. An enlarged drawing of
the nozzle insert is shown in Figure 4.6. Water cooling (17) is
incorporated in this plate and O-ring seal (15) is used to seal the quartz
flange (32) to the stainless steel end plate (31).

122

Third Generation MCJR (MCJR—3)

 

 

 

 

31 viz-Am
. ‘ 2 32 --6 .1433. 24
I I 1|
4 l l L :: 27
I .. I L—
8 20
| | IO-HCZZT
26 I 2 5
I l ""
2, l l
1 l |
I |
Ill ‘
7
LEGEND
1 Cavity Walls 15 O-ring Seal
2 Sliding Short 16 Cooling Water
3 Finger Stock 17 Cooling Water
4 Cavity Top Surface 18 O-ring Spacer
5 Quartz Tube 19 Jet Nozzle
6 Plasma Discharge 20 Quartz Dame
7 Gas Input 21 Cooling Water
8 Finger Stock 24 Brass End Plate
9 Outer Conductor 26 Window/Air Outlet
10 Power Input Probe 27 Cooling Air Inlet
11 Substrate 28 Process Chamber
12 Substrate Holder 31 5.5. End Plate
13 Linear Feed Through 32 Quartz Flange

 

Figure 4.5 Schematic Drawing of MCJR-3

123

tons «3.2 05 do 358a Susan .3. Emma

32>
32> aoh 3v cozuomlmmoco on

 

 

 

 

 

124
A quartz assembly is used to transport and confine the reactive gas

mixture. The quartz assembly is made of a one inch quartz tube (5)

welded to a hemispherical quartz dome (20) which has an inside
diameter of 4.25 inches and an outside diameter of 4.5 inches. This

hemispherical quartz dome (20) is welded near the equator to a quartz
flange (32) of 6 inch in diameter and 0.25 inch thick. The one inch quartz

tube (5) passes through the sliding short (2) of the cavity applicator and
the quartz dome is housed inside the cavity applicator. The quartz flange
(32) is sandwiched between the brass (24) and stainless steel (31) end

plates.

4- 5.3 Remarks

This third generation MCJ R has been designed but has not been
built. It is expected that the discharge and quartz wall direct contact
problem and the poor reproducibility problem encountered in first and
second generation MCJR will be solved with this jet reactor. Higher
Operating pressure and higher input power which are desirable for
diamond film growth at higher growth rates are expected to be achievable

With this jet reactor.

CHAPTER FIVE

MICROWAVE ELECTRIC FIELDS IN THE MCPR

5 - 1 Introduction

In order to develop microwave reactors that can deposit diamond
films over large surface areas and at high growth rates. it is important to
improve the knowledge of the fundamental plasma/ chemical reactions of
film deposition and to develop an understanding of the electromagnetic

field / plasma interactions.
When an ideal cylindrical cavity of radius a is excited with a

resonant mode (team or T'anq modes) at frequency fo. the cavity
lengths L, are determined by the following equationsfss’

-l/2

(L932: = qna(41r2a2f32u-x:p) (eq. 5.1)
for 'l‘anq modes. and
LTE- 413th -'2 "W ( 52)
( r) npq " qna( a 08" xnp) cc} °

for T3,,“ modes. where n. p and q are non-negative integers. xup are

zeros of the Bessel function. Jn(x). and x'up are zeros of the derivative of
tlle Bessel function. J 'n(xJ.

125

 

 

126

Each of these resonant modes. T'Eupq or Tanq. has well defined
electric and magnetic field patterns which are expressed as the analytical
solutions to the Maxwell's equations with perfectly conducting cylindrical
walls as the boundaries.

When the cavity is excited with a plasma discharge. the electric
and magnetic field patterns of these resonant modes are disturbed. The
determination of electromagnetic field patterns in a general plasma

discharge loaded cylindrical cavity from analytically solving the Maxwell’s
equations has not been achieved yet. This is due to the fact that plasma
discharge can be created at various conditions. such as from sub-mTorr
to over 1 atmosphere pressures. with various geometries and various
electromagnetic properties. The electromagnetic properties of these
plasma discharges as of yet have not been well established. Also. it is still
a challenge to obtain solutions to Maxwell's equations in a cavity loaded
with electromagnetically lossy material. such as the plasma discharge.

This chapter describes the techniques and results of the

measurement of electric field strengths along the cavity inside walls in a
plasma discharge loaded cylindrical cavity (MCPR7-1).<1°'l3'14'7°’ The
ele Ctromagnetic (microwave) / plasma conditions required to grow
mankind thin film are discussed. Specifically. the (relative) spatial
Vanations of the exciting electromagnetic field patterns are measured.
mese spatial electric field measurements and intensities are then related
to other important experimental parameters such as gas mixture. flow
rate . input power. substrate temperature. and discharge pressure. etc.

127
5.2 Experimental Measurement System

The electromagnetic field pattern measurement system consists of
a vertical bar shown in Figure 5. l and two horizontal cylindrical bars
enclosing the cavity as shown in the cross sectional view of FigUre 5.2.
They are soldered onto the outside surface of the cavity side walls. The
top and bottom horizontal cylindrical bars are located at (A) 50 mm and
(B) 20 mm from the cavity bottom surface. respectively. A series of holes.
about 2.2 mm in diameter. are drilled into the vertical and the horizontal
bars and completely through the inside surface of the cavity walls. The
spacing between the holes on the vertical bar is 10 mm and the holes on
the horizontal circles have an angular spacing of 7.5°.

A semirigid micro-coaxial diagnostic probe is inserted into the
holes until the tip of the probe is flush with the inside surface of the
cavity side walls (see insert on Figure 5.2). Power readings. which are

proportional to the square of the rms electrical field strength normal to
title inside cavity walls. are obtained by connecting a power meter to the
Outlet end of the micro-coaxial probe during the experiments.<10>

The relationship between the diagnostic probe power reading PF
and the actual rms electrical field strength E. where E is defined by
E2 = E. was determined by using a 6" diameter empty cylindrical
brags cavity.<56.10.70>

The calibration procedure was performed as follows. The empty
cylindrical brass cavity was excited with the 1mm 1 mode and several
2 r 45 GHz input power levels were used. The corresponding probe power
t‘ea-dings were taken with a power meter attached to the other end of the
probe whose tip was set flush with the inside surface of 6" cavity walls.

128

 

 

 

 

 

 

 

 

 

To Vacuum Pump

 

LEGEND

1 Cavity Walls 11 Graphite Base

2 Sliding Short 12 Quartz Tube

3 Finger Stock 13 Substrate Holder

4 Cavity bottom surface 14 Substrate

5 Quartz Dome 15 Top Window

6 Plasma Discharge 16 Pyrometer Sensor

7 Gas Input Tunnel 17 Input Coupling Probe
8 Gas Distribution Ring 18 Outer Conductor

9 Metal Screen 36 Vertical Bar

10) Cooling Water Tunnel A) 8: (B) Horizontal Rings

 

Figure 5.1 Electric Field Measurement System

129

     

Circumferential
angle

Microcoaxial ' 0. (350.) """" "" "
diagnostic (C) i\ s- (C)

  
   
  

 

Probe

 

 

 

Side window

Figure 5.2 Horizontal Cross-sectional View

 

130

When the probe was inserted into the probe hole at z = 3.7 cm above the
cavity bottom surface and the cavity length d is adjusted to 7.795 cm for
TMOI 1 mode excitation. the following table of diagnostic probe power
reading vs. input power was obtained.

Table 5. 1 Diagnostic Probe Power Reading versus Input Power

 

 

 

 

 

 

 

 

Incident Power Reflected Power Probe Power
P1 (W) Pr (W) Pp (um
0.8526 0.0229 48
0.5858 0.0134 33
0.3608 0.0046 20.5
0.1093 0.001 1 6.3

 

and their ratio Pd/Pp is generated.

and the following table of absorbed power Pd = P1 - Pr. probe power Pp

Table 5.2 Absorbed Power and Probe Power Ratio

 

 

 

 

 

 

 

Absorbed Power Probe Power Power Ratio
Pd (W) Pp oiwi Pd/Pp
0.8297 48 1.73 x 104
0.5724 33 1.73 x 04
0.3560 20.5 1.74 x 104
0.1082 6.3 1.72 x 104

 

 

 

 

 

where we see the power ratio stays constant when the absorbed power is
manged. i.e., Pd/Pp = 1.73 x 104.

In order to establish the relationship between the electric field
s‘tr'itngth E near the cavity inside walls and the probe power reading Pp.
.1116 relationship between the electric field strength E near the cavity

131

inside walls and the absorbed power Pd. is first derived. Thereafter. the
relationship between E and PD can be established from Pd/Pp = 1.73 x
104.

Using the perturbation theory for lossy conducting surfaces and
assuming that the power absorbed by the empty cavity is dissipated on
the cavity walls. the following power balance equation is established.<65>

_s 2
Pa = it§l1il ds (eq. 5.3)
whereas for brass material at f0 = 2.45 GHz. 91 is given by,<65>
9: = 5.01 x 10" f0 = 0.0248 (eq. 5.4)

For TMOI 1 mode. H field has mainly the 11 component. which is

given by.<65>

H4, = C(%)J,(x§r)cos(5;) (eq. 5.5)

where x0, = 2.405. a = 3" = 7.62 cm and C is a constant.
Using equations 5.4 and 5.5. we get. from equation 5.3.

_‘2
Pd = 9i§IH| ds =

2 d a
C2 (%) SR [21:an (x01) Icosz (“3.2) dz + 212nm]? (fa—21” dr]
0 0

d
czxglsiaf (x01) ["X + 211'] (eq. 5.6)

 

132

Since x0, = 2.405. J1(xm) = 0.5191. d/a = 7.795/7.62 = 1.02. Si =
0.0248. we have. from equation 5.6.

Pd = 0.38702 (eq. 5.7)

The electric field E near the cylindrical cavity inside walls has
mainly the component normal to the walls. i.e., E = IEfl. where E, is
given by the following equation,‘65>

excl“ 1‘2
Eli”. = mJl‘xoflsm‘a)
which leads to.
02x3 2:2 n
2 _ l 2 2 _
|Er| If“ - ——m282a2d2J 1 (x01) sin ( (1) (eq. 5.8)

where a) = 21! x 2.45 x 109, z/d = 3.7/7.795 = 0.43. When the other
parameters used to obtain equation 5.7 are also used. we obtain. from

equation 5.8.
[Elflrza = 2.34x 107C2 (eq. 5.9)

From equations 5.7 and 5.9. we obtain the following relationship
between the electric field strength E near the cylindrical cavity inside
Walls. E = IErl . and the absorbed power. Pd.

[E'lzlrm1| = 6.37 x 107Pd (eq. 5.10)

133

Using the result from table 5.2. i.e., Pd/Pp = 1.73 x 104. we obtain
the following relationship between the electric field strength E near the

cavity inside walls and the probe power reading. Pp.

E2 = |Er|2|r_a = 1.1 x 10”?p (eq. 5.11)
where the unit of E is volt/m and the unit of PI) is W.

5.3 Measurement Results
5.3. l Electromagnetic Field Patterns

The axial and circumferential electric fields were measured while
the MPDR was operating under diamond thin film deposition conditions.
An example of a field pattern measurement is displayed in Figure 5.3 for
the case of a discharge that is symmetrically centered over the substrate.
The experimental conditions were: pressure = 65 Torr. H2 flow rate = 300
scam. CH4 flow rate = 1.5 scorn. absorbed microwave power = 590 W and
substrate temperature ~ 975 °C. As shown in Figure 5.3. the electric field
at a constant height is nearly independent of circumferential angle ¢- The
Slight variation in intensity versus 0 can be attributed to the near field of
the exiting probe. Also shown in Figure 5.3. the axial field was measured
and is compared to an empty waveguide T'Mm standing wave (dashed
.line).

These field measurements clearly show that the mode is 4)
independent. and indicate that the plasma loaded TMOI 1 mode was

Electrical Field Strength (x 1 04 volt/m)

134

 

 

 

 

 

 

 

 

 

 

 

 

2.5
2 - -
Top horizontal circle
15 F ++r++ ++++++++ ‘1
' v+4+L++H-'fi" '3 I”. $3.... r+++r+v
.000 000.,
1 imzsssz“¢“3 "0;. .0. . .311
Bottom horizontal circle
0.5 - -
O l 1 _.l_ ...i
0 90 180 270 360
Circumferential Angle (degree)
1.6
1.2 - .
0.8 - -
0.4 - -
Vertical direction
0 i 2 f"; 4 5 6 7
Height Along Cavity Wall (cm)

, Figure 5.3 Measured Electric Field Strength Distribution

135

excited during deposition.

The shape of the electric field patterns shown in Figure 5.3 are
typical of the electric field measured for a 4’ symmetric. axially centered
discharge operating under a wide range of different experimental
conditions. As experimental conditions change. i.e., pressure from 30 -
70 Torr. gas flow and input power. only a slight variation of L5 and Lp
(less than a few mm) are required to maintain a well matched and
centered discharge. However. the discharge can also be maintained in an
off-centered and non-symmetrical position if Is and 1..p are changed from
the optimal discharge centered conditions and / or if the substrate itself is
positioned off the center of the cavity axis. When this occurs. the
electromagnetic field patterns are altered. The vertical field intensities
retain the standing wave field pattern shown in Figure 5.3. however the
circumferential electric field displays a skewed pattern where the electric
field intensities are the highest near the coupling probe (0 ~ 180°) and
decrease as 0 varies form 180° to 0 or 360 degrees. Film growth is
usually non-uniform and non-symmetric under these experimental
conditions.

Figure 5.4 summarizes the measured. plasma loaded TMOI 1 mode
electric and magnetic field patterns and their relationship to the

discharge and substrate. The discharge is formed at the open end of the
Cavity inside the quartz disk and its length can be as long as a quarter
Wavelength ~3.6 cm. The most intense electric field components are
tangential to the discharge boundary. As shown. the discharge is
Completely separated from the quartz walls and is in direct contact with
the substrate and holder. The metal base-plate walls adjacent to the
discharge and the substrate help stabilize the are. In this configuration

136

 

  
   

  

— - - ‘ . ' - - -
. v A'A'A'A'A'A'A'A'A'A'AVA‘ .
' - V‘V‘V‘VAV‘V‘V‘V‘V‘V‘V‘VA '
H - - , v . ' . u-r

 

 

 

 

 

 

 

 

 

 

 

 

 

 

P ”‘1
\d /
\ I
O —— O
O O O O
0 00 00.5

—"8 field line
---—-% field line

Figure 5.4 Measured Field Pattern in MCPR

137

ions and free radicals derived from hydrogen and methane impinge
directly on the substrate heating the wafer by conduction and convection
and provide the necessary species for the appropriate substrate surface
chemistry.

To further illustrate the relationship between electric field and
discharge boundary. a comparison of plasma discharge excited in a
rectangular cavity and that in our cylindrical cavity is shown in Figure
5.5. Since the electric field has different distribution in the two different
cavities. the plasma discharge takes on a difl‘erent shape. What the two
difi'erent configurations have in common is that the intense electric field
adjacent to the discharge are tangential to the discharge boundary. This
relationship can serve us as a guide in designing microwave cavities for

a specific applications.
5.3.2 Power Balance and Discharge Power Density

The microwave power coupled into the plasma loaded applicator is
given by Pd = P, - P,. where P1 is the incident power and Pr is the reflected
power. The power coupled into the applicator Pd divides itself between
the power absorbed in the conducting applicator walls Pb. and the power
delivered to the discharge load Pa. The exact division of the power Pd
between the walls and the discharge load depends on the relative losses
in the discharge versus the losses in the applicator walls.

The experimentally measured electric field intensities shown in
Figure 5.3 are typical order of magnitude electric field strengths inside
the cavity under diamond film thin film deposition conditions. These
measured electric field strengths can be used to estimate the power loss

138

 

cEcE cEccE use causes—5m 88:85 he 8:3...

cc: 2c: all..-
cc: 2c: will

 

 

 

 

bibli

 

 

 

 

 

  

.QQQQQ'

 

 

Q\§§\§\‘\O

 

 

 

 

 

Q.
Q.
\\
Q. \ ‘
\\
.\
Q .\.\\§§§

V n ssssss

V
/
i
l
T—
l
l
|
f-

 

 

\
C‘ \QCQI‘.‘
.\ \Q.§§§§\

- o \Q‘QQQ
O
O

 

 

 

-—-r
__‘l
0 \\§ 0
0 .0. Q
Q \Q. Q
\ sss Q
Q 0“ Q
Q 005 Q
Q \\
\ Q.
QIQG
l

O

O

.ssssssssss -

. sssssssss
ssssssssss
”3333

 

V

 

 

 

 

 

.9ch 50.5530

/.......\ 0‘

3300 5.39.3001

 

139

on the cavity walls and the cavity loaded 9 under diamond film
deposition conditions.

To estimate the power loss on the cavity walls. the measured
electric field strength along the cavity inside walls is first expressed
empirically by the following equation. from Figure 5.3.

NZ
|Er||r=a = 1.45x 10451.1(?) (eq. 5.12)

From equations 5.8 and 5.12. we obtain the following equation to
determine the constant C.

(tread

4
C = 1.45x 10
x01“J1(x01)

 

(eq. 5. 13)

where for the MCPR used in the field pattern measurements. a = 3.5" =
8.89x10'2 m and d = 7.2x10'2 m. Hence. the constant C is given by.

C = 3.22 (eq. 5.14)

From equations 5.6 and 5.14. the power loss on the cavity walls is
given by

Pb = szglflJ? (x01) [$421!] = 3.54 (W) (eq. 5.15)

Therefore, the data shown in Figure 5.3 yields an estimated
absorbed wall power of less than 4 W indicating most of the 590 W of
absorbed power is coupled to the discharge / substrate holder. The
absorbed microwave power efficiency is greater than 99%.

140

The Q factor of the plasma discharge loaded cavity applicator is
estimated by first calculating the energy stored in the cavity applicator.

Assuming that the electromagnetic fields were only slightly altered
by the presence of the discharge and that the electromagnetic fields in
the discharge were not much different from those when the discharge
was not present. the electric field components. E,. E, and Ez can be

expressed by the following equations.<65>

 

_ (3"‘01’t xol “Z
i' - MJI (Tr) sin(—d-) (eq. 5.16)
E¢ = 0 (eq. 5.17)
C 2
E2 = x012J0(x—°-lr)—cos(d—z) (eq. 5.18)
jcoea2

The energ stored in the cavity applicator is given by.<65>
w = 2We = eJIEl W = 418,) dV+eJ|Ez| dV =

2 2 d
=8 e(—) (:21) 2r[(":)Ism2(g z)d zIJ2(fll-r)rdr+
d2 0

2 d
+ (L— :1)Icos(—z)dzIJg(x—91r)rdr =

O

02 d
= 1! zzxglJ ( KM ) 2+(xfi) 2] (eq. 5.19)
2012 e

 

where C is given by equation 5. 14. Hence the energy stored in the cavity
applicator is.

141
w = 2.3x 10'6 (J) (eq. 5.20)

Since the absorbed microwave power by the plasma discharge

loaded cavity applicator is Pd = 590 W. the loaded cavity 9 factor is given

by.<65>

o = — = so (eq. 5.21)

Thus. the absorbed microwave power is very emcient. i.e., Pd ~ Pa.

A visual estimation of the discharge volume is 100 cm3 at the low
pressure of 20 Torr. As pressure increases to 70 Torr. the volume
decreases (for a constant power of 590 W) to 30 cm3. Thus. the discharge
absorbed power density varies from approximately 6 W/cm3 to 20 W/

01113.

5.3.3 Electric Field Strength

Diagnostic probe power readings were taken when chamber
pressure. hydrogen and methane flow rates. and microwave power input
were varied. Each set of measurements were taken for a o symmetric.
well-centered discharge. The results are summarized in Figures 5.6 and
5.7 in which diagnostic probe power is plotted vs. input power. flow rate.
and pressure. The measured diagnostic probe power is proportional to
the square of electric field. i.e., it is proportional to the energy stored in
the loaded cavity fields. The power absorbed. Pd, is equal to the total
power dissipated within the cavity. Thus. the ratio of the probe power to
the absorbed power is proportional to the discharge loaded cavity Q.

Figure 5.6 shows under a constant pressure of 70 Torr and three

142

8 | r I

 

100/0.5 sccm +-
200/1.0 sccm -t— ..

300/1 .5 sccm -a—
Pressure = 70 Torr

>1
UI
I

Coaxial Probe Power (x 10'5 W)
0‘
UI

 

 

 

 

5.5 - ..
5 l l I
400 450 500 550 600

Absorbed Microwave Power (W)

Figure 5.6 Coaxial Probe Power versus Microwave Power and Flow Rate

143

 

  

 

 

 

8 u 1
§ 75 - /
“P
53
5
as 7 '
E
D:
g 6.5 - +— 55 Torr
-i— 60Torr
g 4— 65 Torr
O 6 . *— 70 Torr
H2 = 200 sccm
CH4 = 1 sccm
5.5 ' '
450 500 550 600

Absorbed Microwave Power (W)

Figure 5.7 Coaxial Probe Power versus Microwave Power and Pressure

144

difl’erent flow rates. the diagnostic probe power. and hence the square of
the cavity electric field strength, is approximately proportional to the
power input and only slightly influenced by the flow rates.

Figure 5.7 shows for a constant flow of 200 sccm H2 and 1.0 sccm
CH4. the probe power is approximately proportional to the input power
and is only slightly influenced by the variation in pressure.

These measurements demonstrate that the discharge loaded cavity
9 varies little (~ 60) over the difl‘erent flow rates. pressures and input
powers. The absorbed microwave power emciency is high over the wide
variations of experimental conditions shown in Figures 5.6 and 5.7.
Thus. the power absorbed by the applicator is essentially equal to the
power dissipated in the plasma/ substrate holder.

5.3.4 Substrate Temperature

Substrate temperature plays important role in diamond film
growth process. It afl‘ects diamond film growth rate and morpholog. two
of the leading figures of merit in diamond film growth process. Figures
5.8 and 5.9 show substrate temperature versus absorbed microwave
power, flow rates and pressure.

Figure 5.8 shows under a constant pressure of 60 Torr and three
different flow rates. the substrate temperature increases as absorbed
power is increased and is only slightly influenced by the flow rates in our
reactor.

Figure 5.9 shows for a constant flow of 200 sccm H2 and 1.0 sccm
CH4. the substrate temperature increases when the absorbed power and

pressure are increased.

145

960 j u

+— 100/0.5 sccm
_,_ ZOO/1.0 sccm

 

-9— BOO/1.5 sccm

Pressure = 60 Torr

Substrate Temperature (°C)

 

 

 

920 - . -
900 ' '
450 500 550 600

Absorbed Microwave Power (W)

figure 5.8 Substrate Temperature vs. Microwave Power and Flow Rate

146

 

 

 

 

l l l
1000 - "
6
9..
ES 960 - .
4.;
i3
8.
93
E-'
*3 70 Torr +-
‘8 920 b 65 Torr -i- .
'3 60 Torr -a-
U) 55 Torr -x—
H2 = 200 sccm
CH4 = 1 sccm
880 ' ' '
450 500 550 600 650

Absorbed Microwave Power (W)

Figure 5.9 Substrate Temperature vs. Microwave Power and Pressure

147

5.4 Discussion

These experiments indicate that under the diamond thin film
deposition conditions of 30 to 80 Torr the microwave discharge is in a
transition between an am-bipolar difl’usion controlled discharge and a
freely floating microwave arc.<56'63-66’68> The discharge is excited by the
lossy plasma loaded TMon cylindrical cavity mode where, as shown in
Figure 5.4. the most intense electric field adjacent to the discharge is
tangent to the discharge boundary. This mode couples to the discharge
very emciently and has been used to excite high pressure “arc like"
microwave discharges for other applications.‘56°63"66'68> It also has the
same relationship between electromagnetic field patterns and the
discharge boundaries that was present in the earliest microwave plasma
diamond thin film deposition experimentsfao’ as shown in Figure 5.5.

Since the microwave arc appears to be fundamental to the
observed diamond film deposition, several of its features are briefly
discussed here to provide a better understanding of the discharge. The
microwave are. like lower-frequency and DC arcs. is a thermally in-
homogeneous discharge. It has a hot central core. and thermal gradients
exist between the discharge center and surrounding walls. Microwave
energy is readily coupled into the electron gas in the hot discharge center
because of its reduced gas density. and because neutral gas species are
also readily ionized and excited in the hot discharge region.

Major energy losses form the discharge occur by heat conduction.
convection. and radiation. When large temperature gradients are present.
heat conduction losses become an important loss process. This loss

mechanism includes the contribution to the heat conductivity by

148

molecules. atoms. electrons. and ions. and also chemical reactions such
as the transport of dissociation energy and ionization energy to the arc
fringes by free radicals and ions. Owing to the high-pressure
environment. electron, ion and free radicals recombine quickly outside
the hot central core or on a solid surface in contact with the are core and
thus convert their dissociation. ionization. and excitation energy into
thermal energy. The result is a discharge with radially varying gas
temperature. ionization rate. and volume recombination rate. Gas
temperature. ionization, dissociation. etc.. are highest in the center of the
discharge. while volume recombination and de-excitation of the difl'erent
species increase away from the discharge center as the cooler, denser gas
regions near the walls are approached. The central discharge core gas
temperatures vary with gas type and pressure and can be in excess of
2000 °K.<66‘68> while temperatures extemal to the discharge are
controlled by wall temperatures and the temperatures of the gas flowing
around the discharge. If gas temperatures are 2000 °K or more in
hydrogen gas. then atomic hydrogen can be created by thermal

dissociation as well as free electron impact.<69>

5.5 Summary

The following conclusions were reached from this study, under
diamond film deposition conditions in MCPR7 -1:
(l) The excitation mode was identified to be discharge loaded

TMOI 1 mode.
(2) Maximum electric field intensity in the cavity was ~ 150 V/

149

(2) Discharge loaded cavity quality factor was ~ 60.

(3) The tangential component of the 175 field is the main
discharge excitation field.

(4) Average power density in the discharge was ~ 6 W/cm3 at 20
Torr and 20 W/cm3 at 70 Torr.

(5) Under a constant pressure, the square of the cavity electric
field strength is approximately proportional to the power input and only
slightly influenced by the variation in flow rates.

(6) At a constant gas flow, the square of the cavity electric field
strength is approximately proportional to the power input and only
slightly influenced by the variation in pressure.

(7) Under a constant pressure. the substrate temperature
increases with the power input (~ 0.4 °C/W) and is only slightly
influenced by the variation in flow rates.

(8) At a constant gas flow. the substrate temperature increases
with the power input (~ 0.4 °C/W) under each working pressure. Also. at
the constant gas flow and input power. the substrate temperature
increases (~ 3.5 °C/T‘orr) with increasing pressure.

CHAPTER SIX

DIAMOND FILM DEPOSITION IN THE MCPR

6.1 Introduction

In various practical applications. diamond films with various
thicknesses and morphologies are desired. In order to determine the
growth rates and morphologies of diamond films deposited under various
experimental conditions and hence characterize the performance of the
third generation microwave cavity plasma reactor for diamond film
growth. diamond films are grown on three and four inch silicon wafers
under various experimental conditions. In this chapter. the diamond film
growth rates. growth emciencies and morphologies are presented with
respect to variations in substrate temperature. gas composition and flow
rate. Typical uniformity profile and Raman spectrum are also presented.

The objective of depositing diamond films 'on 3" and 4" silicon
wafers is to achieve the highest possible growth rate with good uniformity
(judged by visual inspection).

The sections in this chapter are arranged as follows. Section 6.2
describes the experimental methodolog' by which the reactor is
characterized. Reactor block diagram and experimental parameters are
used to describe the experimental methodology. Section 6.3 describes the
experimental configuration and the common experimental procedures

150

151

used in the parametric study. The characteristic behaviors of this
experimental configuration are described. Section 6.4 describes the
results of an experimental parametric study of diamond film deposition
on 3" silicon wafers. The goal of this study is to achieve the highest
growth rate possible since achieving good uniformity (judged by visual
inspection) over 3" silicon wafer is possible under a wide range of
experimental conditions in this MCPR. Film growth rates. growth
emciencies and morphologies of the deposited diamond films under
various deposition conditions are described. Typical uniformity profile
and Raman spectrum are also presented. Section 6.5 describes a study
of experimental set-up for diamond film deposition on 4" silicon wafers.
Here film uniformity and growth rate are the judging criteria. Section 6.6
describes a parametric study of diamond film deposition on 4" silicon
wafers. The goal of this study is to achieve the highest growth rate
possible while maintaining good uniformity Gudged by visual inspection).
There exists a compromise between the uniformity and growth rate in
diamond film deposition on 4" silicon wafers in the present reactor. When
reasonable uniformity is required. the main limiting factor on the growth
rate comes from the safe operating temperature limit and the coating
speed of the quartz dome. film growth rates. growth emciencies and
morphologies of diamond films deposited under various experimental
conditions are described. Typical uniformity profile and Raman spectrum

are also presented.

152

6.2 Experimental Methodology
6.2. 1 Introduction

In order to orderly investigate the performance of the MCPR. it is
essential to identify the experimental parameters and establish a method
of evaluating the operational performance of the MCPR. This section
describes the various experimental parameters. methods and “figures of
merit” used in the evaluation.

The deposition experiments conducted in MCPR have many
experimental parameters. The reactor 1/0 block diagram shown in Figure
6. 1 displays the relationship between these experimental parameters.
These experimental parameters can be divided into three groups: (1)
independent input experimental parameters. (2) internal. dependent
experimental parameters. and (3) external. output parameters. Each of
them is summarized in more detail below.

I. Independent Input Experimental Parameters
(i) Substrate size. shape and other physical properties
(ii) Reactor geometry parameters
(a) cavity applicator construction
(b) quartz dome geometry
(c) base-plate design
(d) gas flow configuration
(e) substrate holder geometry and material
(i) substrate location .
(g) end feed vs. side feed
(h) electromagnetic mode excitation and cavity tuning

153

Reactor geometry parameters

Seeding procedures

 

Pressure -—‘H

N ‘—> Growth rate

  
   
   
 

 

  
 

. Substrate temperature ———> MOTPhOIOEY
Microwave
—-
power Discharge volume Uniformity
Power density
Gas . ___> Raman spectra
composition Residence time

 
  

etc. Efficiency

 
 

T0131 ____~
flow rate

 

 

- etc.

 

Start-up and shut-down procedures
Deposition time

etc.

Figure 6.1 Reactor I / 0 Block Diagram

(iii)

etc.)

(w)

154

(i) vacuum system - pumping speed. volume. etc.
Deposition process parameters

(a) substrate seeding procedure

(b) start-up and shut-down procedures

(c) input power variation vs. time (CW vs. pulsed power.

(d) deposition time. etc.

Macroscopic controllable input parameters

(a) operating pressure. p '
(b) absorbed microwave power. Pt

(c) gas composition expressed in % of H2 gas flow. i.e. H2.

CH4/H2. COg/Hz. th.

(d) total input gas flow rate. ft.

11. Internal, Dependent Experimental Parameters

(i)
(ii)
(iii)
(M
(V)

Substrate temperature, T‘s

Discharge volume. Vd. and deposition area. Ad
Discharge power density. w

Gas residence time. tr

Electric field strengths and electromagnetic field

distributions within the reactor

(vi)

Densities of discharge species

(a) densities of charged species

(b) densities of excited species

(c) densities of atomic and molecular species

((1) variation of these densities within the reactor

(especially above the substrate)

(vii)

Energy distributions and temperatures of discharge species

155

(viii) Flow patterns of discharge species within the reactor
(ix) Fluxes of discharge species onto the substrate
Ill. External. Output Parameters
(1) Film linear growth rate (um/ hr) and weight gain. 11 (mg/hr)
(11) Film morphology
(iii) Film uniformity
(iv) Film Raman spectra
(V) Film growth emciency
(a) growth rate vs. power. Pt
(b) growth rate vs. power density. w
(c) growth rate vs. power flux. Pt/Ad
((1) growth rate vs. total flow rate. ft

(e) carbon atom conversion efficiency

6.2.2 Independent. Input Erqrerimental Parameters

It is obvious that the experiment here has many independent input
parameters. Thus in order to develop an understanding of the
operational performance of the reactor. it is best to empirically optimize
and fix some of the independent experimental parameters.

The substrate size and physical properties are usually fixed. Three
and four inch silicon wafers are used in the experiments described in
this chapter. The cavity applicator construction. quartz dome geometry.
base-plate design. gas flow configuration. substrate holder geometry and
material. substrate location. power feed configuration. and cavity
electromagnetic mode excitation are empirically optimized to create the
optimum discharge geometry for uniform substrate heating and

156

minimum quartz dome heating and coating.

Among the deposition process parameters. the seeding and start-
up procedures are empirically optimized to generate reproducible and
high diamond nucleation density. The shut down procedure is chosen
such that the formation of non-diamond carbon on the deposited film
surface is minimized. CW power is used in all the experiments described
in this chapter. The deposition time is fixed for each set of experiments.
More detailed description of these independent parameters is given in
relevant sections.

The four macroscopic controllable input experimental parameters.
i.e. . pressure (p). absorbed microwave power (Pt). gas composition and
total gas flow rate (ft). are varied during the experiments in order to
understand the characteristic behavior of MCPR under diamond film
deposition conditions and to optimize the diamond film deposition
process in MCPR.

6.2.3 Internal, Dependent Experimental Parameters

The internal. dependent experimental parameters vary with
changes in the independent experimental parameters. Substrate
temperature T3 and discharge volume Vd are two important internal
experimental parameters in the diamond film deposition process. The
substrate temperature Ts greatly afi'ects the growth rate. morphology and
Raman spectra of the deposited film. The discharge volume Vd
determines the deposition area and film uniformity.

These internal experimental parameters are determined and / or
defined as follows:

157

substrate temperature T,3 is measured by an optical pyrometer.

discharge volume Vd is estimated by visual estimation.

the discharge power density w is defined by

w = Pt/Vd
where Pt is the absorbed microwave power. and
the gas residence time t, is defined by
tr = qu/ft (sec)
where p (expressed in Torr) is the pressure. Vq (expressed in liters) is the
quartz dome volume above the substrate and ft (expressed in Torr-liter/
sec. 1 Torr-Iiter/sec = 79.05 sccm) is the total flow rate.

These dependent internal experimental parameters vary in a non-
linear fashion when the input experimental parameters vary. In order to
control the deposition process. these parameters must be controlled. It is
also ofien desirable to measure some of these internal experimental
parameters in-situ during the deposition process. If these measurements
are performed versus time during the deposition process. the
understanding of the deposition process can be greatly improved.

The other intemal. dependent experimental parameters will not be
discussed in this dissertation.

6.2.4 External. Output Experimental Parameters

The reactor performance is evaluated in terms of the external.
output experimental parameters. These output parameters can be
measured and / or calculated after the film deposition is completed. Film
growth rate is expressed in terms of both linear growth rate (um/hour)
and weight gain. u (mg/hour). Weight gain is obtained from the

158

deposition time and weighing the wafer before and after the deposition
process using a weight balance. The linear growth rate is calculated fi'om
the weight gain. the area of the wafer and the density of diamond. 3.51
g/cm3.

The film morphology is obtained fi'om observation under the optical
microscope and photographs taken with Scanning Electron Microscope
(SEM). The film uniformity is obtained by both visual estimation and
laser interference reflection techniquefn’ The film quality is
characterized by Raman spectroscopy.

The following “figures of merit” are used to quantify the film growth
emciency of the MCPR:

(1) k1 (mg/kW-hr) = u/Pt: 11 versus Pt. growth rate versus
absorbed microwave power.

(2) k2 (mg-cm3/kW-hr) = u/w: u versus w. growth rate versus
power density.

(3) k3 (mg-cmzlkW-hr) = u/S: u versus 5, growth rate versus
power flux S. S is defined by S = Pt/Ad. where Ad is the area of the
substrate.

(4) gas flow efficiency. k4 (mg/liter) = u/ft: 11 versus ft. growth
rate versus total flow rate.

(5) carbon conversion efiiciency. k5 (%): it is defined as the
percentage of carbon atoms in the input gases that are converted into the
diamond film.

These output parameters allow the MCPR to be compared with
other reactors and allow the comparison of the MCPR as experimental
input conditions change. It is also noted here that the optimization of one
performance criteria. such as uniformity. may not optimize another

159

performance criteria. such as growth rate.

6.3 Experimental Operational Characteristics

6.3. 1 Introduction

In this section. the basic experimental configuration for diamond
film deposition on 3" and 4" silicon wafers is described. The common
experimental techniques. i.e., seeding. start-up and shut-down
procedures. are described in detail. The operational characteristics of
this experimental configuration. i.e. , internal experimental parameters
vs. input experimental parameters. are presented.

6.3.2 Experimental Configuration

The forced flow experimental configuration which was described in
detail in chapter 3 is the experimental configuration used for diamond
film deposition on 3" and 4" silicon wafers. A schematic drawing of this
experimental configuration is displayed in Figure 3. 16.

Figures 6.2 and 6.3 display the schematic drawings of two cavity
shells used in the experiments. The main difi'erence between the two
cavity shells is that there exist four small windows (optical access
windows for emission and laser induced fluorescence spectroscopy
experiments) in the cavity shell (11). The microwave power coupling
emciency of cavity shell (11) is better than that of cavity shell (I). But the
discharge generated with cavity shell (I) is confined to the silicon wafer
better than that with cavity shell (11). The difi‘erences are especially

160

 

Viewing
window

 

 

 

 

   

Air 'tnc

Figure 6.2 Schematic Drawing of Cavity Shell (1)

 

161

 

 

 

Viewing
window

 

 

 

:

Optical window

 

 

 

 

Figure 6.3 Schematic Drawing of Cavity Shell (II)

162

obvious when diamond films are deposited on 4" silicon wafers. For
diamond film deposition on silicon wafers, cavity shell (I) is a more
favorable choice.

The quartz domes made by difierent manufacturers and those
constructed at difierent times by the same manufacturer result in small
difi'erences in geometry. These small difi'erences in geometry do not have
much efi'ect on diamond film deposition on substrate surfaces which are
3.25” in diameter or smaller. But when diamond films are deposited on
4" silicon wafers. these small differences in geometry influence the
optimum deposition condition. cavity tuning and the resulting film
morphology. uniformity. and growth rate. etc.

In the present experimental configuration. the substrate holder
(13) and flow regulator (32) are combined into one holder (13:32). The
metal tube (29) used here has an outside diameter of 3". inside diameter
of 2.875” and a length of 1.5”.

6.3.3 Common Experimental Procedures

In this section. the common experimental procedures. i.e., seeding.
start-up and shut down procedures. which are used in all the
experiments discussed in this chapter. are described in detail.
6.3.3. 1 Seeding Procedures

Substrate seeding has been used prior to deposition experiment to

enhance diamond nucleation density. Three types of seeding procedures
have been used in the experiments described in this dissertation. They

163

include (1) seeding by hand scratching substrate with 0.25 um diamond
paste. (2) seeding by hand scratching substrate with 1 pm diamond
powder and (3) seeding by photo resist seeding method<22'72>.

The first two methods are self-explanatory. They are simple to
perform but the repeatability is poor. The third method is the best in
terms of both repeatability and nucleation density and it is the seeding
method used in all the experiments described in this chapter.

The photo resist seeding method was derived fi'om the standard
photolithography procedure used in integrated circuit fabrication. In this
method. diamond powder and photo resist are first mixed ultrasonically.
The resulting mixture is then used to cover the entire wafer. The wafer is
in turn spined in a spinner which results in a uniform coating on the
wafer. Diamond seeds are embedded in this uniform coating.

A standard recipe‘73> of the photo resist seeding procedure is as
follows:

(a) mix 142 mg Amplex <74> o. 1 pm diamond powder with 16
ml of Shipley<75> type A photoresist thinner.

(b) ultrasonically mix the mixture for 15 minutes.

(c) add 42 ml of Shipley 1470 photo resist.

(d) ultrasonically mix the mixture for 1 5 minutes.

(e) apply the mixture onto the wafer.

(1) spin the wafer with the following spin speed and time: 0 -
4000 rpm in 10 seconds. stay at 4000 rpm for 30 seconds and slow down
to 500 rpm in 5 seconds.

(g) bake the wafer at 135 °C for 30 minutes.

164

6.3.3.2 Start-up and Shut-down Procedures

The computer monitor system described in section 3.2.4 is used in
the experiments described in this chapter. The following start-up and
shut-down procedures are followed.

I. Start-up procedure:

(a) evacuate the system to below 5 mTorr.

(b) turn on H2. CH4 and C02 gas flow.

(c) turn on 1.5 kW microwave power when pressure reaches
10 Torr.

(d) increase microwave power as pressure increases.

(e) start timing when pressure and microwave power reach
set levels.

In the present experimental system. the pressure rise time is about
10 Torr/min with 400 sccm H2 gas flow.

II. Shut-down procedure:
(a) turn ofi' CH4 and C02 gas flow together.
(b) wait 3 or 5 minutes for process self-cleaning.
(c) turn 06 microwave power.
(e) turn off H2 gas flow.

(1) evacuate the experimental system.

6.3.4 Experimental Operational Characteristics

It was shown in section 6.2 that the microwave cavity plasma

reactor can be characterized by a set of input. internal and output
experimental parameters. As displayed in Figure 6.1. the substrate

165

temperature. Ts. and discharge volume. Vd. are internal experimental
parameters which depend on all independent input experimental
parameters. In this section. the characteristic behavior of the internal
experimental parameters is investigated as the input experimental
parameters. i.e.. pressure. microwave power. gas composition and total
gas flow rate. are varied.

The forced flow experimental configuration shown in Figure 3. 16 is
used in all the experiments described in this section. Here. the
supporting quartz tube (28) has an inside diameter of 95 m. an outside
diameter of 100 mm and a length of 50 mm.

Three inch silicon wafers are used as substrates in these
experiments. They have the following physical properties: prime grade. p-
type. Boron doped. <100> orientation. 1 - 10 ohm-cm in resistivity. 76
mm in diameter. and 356 - 406 um in thickness.

To help describing the characteristic behavior of the experimental
reactor. the following nomenclatures are defined:

p <--> Pressure.

Pt <--> Absorbed Microwave Power.
ft <--> Total Gas Flow Rate.

Cg <--> Gas Composition.

Ts <--> Substrate Temperature.

Vd <--> Discharge Volume.

The experimental procedures to obtain the characteristic behavior
of the internal experimental parameters are as follows. A diamond film
was deposited on a 3" silicon wafer under the following experimental
conditions: H2 flow rate = 400 sccm. CH4 flow rate = 6 sccm. C02 flow

rate = 2 seem, pressure = 5 1 Torr. absorbed microwave power = 2.34 kW.

166

substrate temperature ~ 900 °C and deposition time = 5 hours. This
diamond film was estimated to be more than 2.5 pm thick. judging from
previous experiments conducted under similar experimental conditions.
All the measurements were conducted after this diamond film had been
formed. The substrate temperature. T8. was measured by an optical
pyrometer which was mounted on a tripod and fixed at the same location
throughout the entire experiment. The discharge volume. Vd. was
estimated by visual estimation. The neighboring substrate temperature
readings and discharge volume estimations were recorded by varying a
minimum number of input experimental parameters and waiting at least
5 minutes for the experiment to reach steady state.

Figure 6.4 shows the variation of substrate temperature Ts vs.
pressure p and absorbed microwave power Pt. As shown. the substrate
temperature Ts is very sensitive to variations in pressure p and only
varies slowly as the absorbed microwave power Pt is varied.

The discharge volume Vd is sensitive to both the variation in
pressure p and absorbed microwave power Pt. Keeping pressure p
constant. the discharge volume Vd increases with increasing absorbed
microwave power Pt. Keeping the absorbed microwave power Pt constant.
the discharge volume Vd decreases with increasing pressure p.

In Figure 6.4. the lower power limit is defined by the minimum
power needed to generate a discharge volume (thmn that covers a 3"
diameter substrate area. The upper power limit is defined by the
maximum power that can be used to operate the reactor safely without
over-heating the quartz dome. The upper discharge volume limit (Vd)m
is the volume of the quartz dome above the substrate.

Figure 6.5 shows the variation of substrate temperature Ts vs. CH4

Substrate Temperature, Ts (°C)

167

 

sso - '

 

y
950 -

 

 

 

 

 

 

p = 55 Torr
9 2 O _ 3...: M -
(vd)[njn .3" of:
\j’ p = 50 Torr
890 - .
860 I. w O... '-
830 - -
p = 40 Torr H2 = 400 sccm
CH4 = 2 sccm
3" silicon wafer
800 " 1 1 1 1 'l
1.9 2.2 2.5 2.8

Absorbed Microwave Power, Pt (kW)

Figure 6.4 Substrate Temperature versus Pressure and

Absorbed Microwave Power

168

 

 

 

 

 

 

 

 

896 I 1 l l l 1
8’ 894 -
V
Hm
a)“
5 892 -
cu
H
O
E”
0 890 -
I"
8
a:
3:: p = 50 Torr
:3 H2 = 400 sccm
‘0 3" silicon wafen
886 J l I l l l
0 1 2 3 4 5 6
CH4 Flow Rate (sccm)

Figure 6.5 Substrate Temperature versus CH4 Concentration

169

concentration. As shown. the substrate temperature Ts is not sensitive to
variation in CH; concentration.

Figure 6.6 shows the variation of substrate temperature Ts vs. C02
concentration. As shown. the substrate temperature Ts varies slowly with
low 002 concentration and decreases when 002 concentration is further
increased. Low C02 concentrations are used in the diamond film
deposition experiments described in this chapter. Figures 6.5 and 6.6
show that the substrate temperature Ts is not sensitive to variation in
gas composition Cg. It is also found by visual estimation that the
discharge volume Vd is not sensitive to variation in gas composition Cg.

Figure 6.7 shows the variation of substrate temperature T$ vs.
absorbed microwave power Pt and total flow rate ft. As shown. the
substrate temperature Ts increases gradually when the absorbed
microwave power P. is increased and it is not sensitive to variations in
total gas flow rate ft. It is also found by visual estimation that the
discharge volume Vd is not sensitive to variations in total gas flow rate ft.

6.3.5 Summary

The following conclusions can be drawn from the experiments
described in the previous section:

(I) Substrate temperature Ts increases sharply with increasing
pressure (~ 6 OC/'l‘orr) and at constant pressure. Ts increases gradually
(~ 0.02 °C/W) with increasing absorbed microwave power. It is not
sensitive to variations in gas composition and total flow rate.

(2) Discharge volume Vd decreases with increasing pressure and

increases with increasing absorbed microwave power. It is not sensitive

Substrate Temperature, Ts (°C)

170

 

 

 

 

 

 

 

 

 

892 I I I I I
1\
890 - ‘
888 -
886 -
884 -
H = 400 sccm
332 . CH4 = 4 seem
= 50 Torr
t = 2.32 kW
880 _ 3" silicon wafer
J l l l l
0 l 2 3 4 5
C02 Flow Rate (sccm)

Figure 6.6 Substrate Temperature versus C02 Concentration

Substrate Temperature, Ts (°C)

896

894

892

890

886

171

 

 

$88 88sccm
sccm

SCCIII
sccm
SCCID

Hi1? ‘

  

   

l l l l

   
      
  

 

 

p = 50 Torr
CH4/H2 = 1%
3" silicon wafer

 

 

l l

 

2. l 2.2 2.3 2.4

2.5 2.6

Absorbed Microwave Power, Pt (kW)

Figure 6.7 Substrate Temperature versus Total Flow Rate and

Absorbed Microwave Power

 

172

to variations in gas composition and total flow rate.

Using the definitions of discharge power density w and power flux

S.

W = Pt/Vd

S = Pt/Ad
where Ad is the substrate area. the following additional conclusions can
be drawn.

(3) Power density. w. W with increasing pressure and at
constant pressure. it is not sensitive to variation in absorbed microwave
power. That is. under a constant pressure. as the absorbed microwave
power increases. the discharge volume also increases. resulting in small
variation in absorbed power density. Power density is not sensitive to
variation in gas composition and total flow rate.

(4) Power flux S increases with increasing absorbed microwave
power. It is not sensitive to variation in gas composition and total flow
rate.

These results on the characteristic behavior of the MCPR under
diamond film deposition conditions demonstrate that the reactor behaves
in an unique repeatable fashion. A set of experimental curves can be
used to describe the substrate temperature variations as pressure.
absorbed microwave power. gas composition and total flow rate vary.
This set of curves can be used to understand and describe the
experimental behavior of the reactor. The allowable experimental
operating space is defined by these curves. If certain substrate
temperature conditions are desired. these curves can be utilized to
determine the pressure and absorbed microwave power required. These
curves serve as the experimental reactor operating energy “road map”.

173

6.4 Diamond Film Deposition on 3" Silicon Wafers
6.4. I Introduction

Diamond films with various physical properties are deposited on 3"
silicon wafers under various experimental conditions. In order to
optimize the diamond film deposition process. it is necessary to study the
co-relation between the physical properties of the deposited diamond
films and the experimental conditions. Hence a parametric study of
diamond film deposition on 3" silicon wafers was conducted. In this '
study. the following experimental parameters are considered:

I The independent input experimental parameters:
(1) absorbed microwave power. Pt.
(2) pressure. p.
(3) gas (H2. CH4. CO?) composition. and
(4) total flow rate. ft.
II. The internal experimental parameters:
(1) substrate temperature T3. and
(2) gas residence time. t,.
III. The output experimental parameters:
(1) film linear growth rate (um/hour).
(2) film weight gain 11 (mg/hour).
(3) film growth efiiciencies:
(1) growth rate vs. power flux. i.e.. 1.1 vs. S.
(ii) gas flow emciency: k4 (mg/liter) = u/ft.
(iii) carbon conversion emciency. k5 (%).
(4) film morphology.

174

(5) typical film uniformity. and
(6) typical film Raman spectrum.

In this study. diamond films are deposited on 3" silicon wafers
using the forced flow configuration shown in Figure 3.16. The physical
properties of the 3" silicon wafers and quartz tube (28) used in these
experiments are the same as that described in section 6.3.4.

Descriptions of relevant experimental operational energy and
chemistry space are also included whenever possible.

6.4.2 Film Growth Rate and Growth Emciency
6.4.2. 1 Efiects of Substrate Temperature

Substrate temperature is an important internal experimental
parameter in diamond film deposition process. In this study. the
substrate temperatures are achieved by choosing the appropriate
pressures. The absorbed microwave powers are adjusted such that the
discharge volumes are large enough to cover the silicon substrates.

Figure 6.8 displays diamond film linear growth rate and weight
gain versus substrate temperature and microwave power flux with gas
flow rates of 200 sccm H2 and 4 sccm CH4. The pressures and absorbed
microwave powers used to achieve the three substrate temperatures are
shown in Table 6.1 located above Figure 6.8.

Figure 6.9 (a) shows the locations of these substrate temperature
conditions on the characteristic energr “map”. Figure 6.9 (b) displays the
relevant carbon conversion emciency and gas flow emciency vs. substrate

temperature.

Growth Rate (um / hour)

175

Table 6.1 Substrate Temperature Conditions (I)

 

850

 

 

 

 

 

 

 

 

 

 

 

 

Temperature, Ts (°C) 750 950
Pressure. p (Torr) 33 45 57
MW power. Pt(kW) 1.47 2.05 2.51

Power Flux, S (W / cm2)
1 32 45 55
n I I 15.80
Hz = 200 sccm
CH4 = 4 seem
0-8 ' 3" silicon wafer “12.64
0.6 - ' 9.48
0.4 - . 6.32
0.2 - - 3.16
L l I O
750 850 950

Substrate Temperature, Ts (°C)

Figure 6.8 Linear Growth Rate and Weight Gain versus

Substrate Temperature and Power Flux (1)

Weight Gain (mg/ hour)

 

1 I I I I I I I

 

H2 = 200 sccm A g - #4
95 CH4 = 4 seem / ‘

 

3" silicon wafer x—— - ‘n‘ 5 33‘

 

 

. Highest growth rate point

 

Substrate Temperature. Ts (°C)
00
M

.4 11.6 11.8 2 22 2:4 2:6 2:8-
Absorbed Microwave Power. Pt (kW)
(21)

 

 

 

 

A I I I O.

.53 9 95 E

8‘ a

5 no

*5 .5.

a 7 . . 0.74 8‘

‘fl 5

5 as

g s

G 5 ' . 0.53 3

o .9.

C: In
3

:3 o

o 3 l l l 0.32

750 850 950
Substrate Temperature. Ts (°C)
(b)

Figure 6.9 (a) Location of Substrate Temperature Conditions on Energy
Map and (b) Carbon Conversion Emciency and Gas
Flow Emciency versus Substrate Temperature (I)

177

Figure 6. 10 displays diamond film linear growth rate and weight
gain versus substrate temperature and microwave power flux with gas
flow rates of 200 sccm H2. 2 seem CH4 and 2 seem C02. The pressures
and absorbed microwave powers to achieve the five temperatures are
shown in Table 6.2 located above Figure 6. 10.

Figure 6.1 1 (a) shows the locations of these substrate temperature
conditions on the characteristic energy “map”. Figure 6.1 1 (b) displays
the relevant carbon conversion efiiciency and gas flow efiiciency vs.
substrate temperature.

Figure 6.12 shows diamond film linear growth rate and weight gain
versus substrate temperature and microwave power flux with gas flow
rates of 400 sccm H2, 4 seem CH4 and 2 seem C02. The pressures and
absorbed microwave powers used to achieve the three temperatures are
shown in Table 6.3 located above Figure 6. 12.

Figure 6.13 (a) shows the locations of these substrate temperature
conditions on the characteristic energ' “map”. Figure 6.13 (b) displays
the relevant carbon conversion emcieney and gas flow efiicieney vs.
substrate temperature. .

As displayed in Figures 6.8 through 6.13, the linear growth rate,
weight gain, carbon conversion efficiency and gas flow emciency generally
experience peak values when the substrate temperature is varied
between 700 °C and 1000 °C. It is also noted from these figures that the
peak values are gas composition dependent. The locations of these gas
compositions on the Bachmann C-H-O phase diagram are shown in
Figure 6.14. It is also shown in Figures 6.9(a), 6.ll(a) and 6.13(aJ that
the combinations of pressures and absorbed microwave powers to

achieve these peak values fall within the boundaries of the characteristic

178

Table 6.2 Substrate Temperature Conditions (11)

 

 

 

 

 

 

 

Temperature, Ts (°C) Pressure, p fl‘orr) MW power. Pt (kW)
750 33 l .48
800 39 l .76
850 45 2.05
900 51 2.34
950 57 2.64

 

 

 

Growth Rate (um / hour)

Power Flux, S (W / cm2)
33 39 46 52 59

 

0.3 . r . . 4.74
H2 = 200 sccm
CH4 = 2 sccm
C02 = 2 sccm A
3" silicon wafer a
0.25.- . 3.95 ,8
Q)
E
U
0.2 - . 3.16 .3
Ti":
3
0.1.. ' I 2.37

 

 

 

I l I
750 800 850 900 950
Substrate Temperature, Ts (°C)

Figure 6.10 Linear Growth Rate and Weight Gain versus
Substrate Temperature and Power Flux (II)

 

 

l I I I I 1 I I

H2=200 sccm - . AA .,.

950' CH4 = 2 seem W ’ -
C02 = 2 sccm

3" silicon wafer

     

 

\l

M

Q
I

Q. Highest growth rate point

 

Substrate Temperature, T8 (°C)
3
9

 

.4 11.6 133 2 2:2 2:4 2:6 2:8
Absorbed Microwave Power, Pt (kW)

(a)

 

 

 

 

3 3 5 . 037 A
6* 5
§ s.
. a
5 0.32 g5
E a
{’3 e
3 2.5 - - It)
8 0.26 3
O C
a E
B 3

. (D
g 2 75 5 5 0.21

Susbstrate Temperature, Ts (°C)
(b)

Figure 6. l 1 (a) Location of Substrate Temperature Conditions on Energy
Map and (b) Carbon Conversion Emciency and Gas
Flow Emcieney versus Substrate Temperature (11)

180

Table 6.3 Substrate Temperature Conditions (111)

 

 

 

 

 

 

 

 

 

 

 

 

 

Temperature. Ts (°C) 927 964 100 1
Pressure, p (Torr) 50 55 60
MW power, Pt (kW) 2.61 2.67 2.74
Power Flux, S (W /cm2)
0 55 58 59.3 60.9- 8 69
. I l I I °
H2 = 400 sccm
CH4 = 4 sccm
C02 = 2 sccm
.5 0.5 . 3" silicon wafer . 7.90
.8
E
E; 0451- - 7.11
2%
{5 0.4 _ - 6.32
035 I I I 5.53
927 964 1001

Substrate Temperature, Ts (°C)

Figure 6.12 Linear Growth Rate and Weight Gain versus

Substrate Temperature and Power Flux (111)

Weight Gain (mg/ hour)

 

 

 

 

 

a 100$ ' ' '
2.. H2 = 400 sccm
I" 950 - CH4 = 4 sccm
:1. C02 = 2 sccm
3" silicon wafer

E 900-

g
Q 8 50 _ .W .
g . A #A
E5 800- ‘ 1' i V - .
H
in?

m 750 - , . "
.3 0 Highest growth rate point
(D

7 ' '

 

 

 

.4 1‘6 1'3 2 2:2 2:4 21.6 2.8

Absorbed Microwave Power, Pt (kW)
(a)

 

 

 

 

95. 4.5 . . . 0.36 E
3‘ 3
.§ no
g .5.
m 4 . . 0.32 8‘
G 5
O

”‘5‘ E
‘9 m
E 3.5 - - 0.28

o 3
O .9.
a e.
3 . 5
8 3‘ 927 9'64 166—1 0'24

Substrate Temperature, Ts (°C)
(b)

Figure 6.1 3 (a) Location of Substrate Temperature Conditions on Energy
Map and (b) Carbon Conversion Emciency and Gas
Flow Emciency versus Substrate Temperature (III)

1
82

8
an
5
0mm
3
a
O
-3
-0
o
5
G
O
m
G
O
:32.
a
00
mm
0
.«O
a
O
a
83
a.
u
.m
35w
E

  

N
m
s... .
“mm
{We
0
o
"n

a
N
”mCF E U
N
O
3
won “0
IO
M m
0.
W

182

E 8335 893 0-3-0 05 co econ—moaaoo new no 5583 3 .w oar.

  
  

A.
we; “no u £\ «8
I N .
rm. 5 I :\£o .n
5 «hi So
5 Li :6 ”N
“a [£35 3
a acogmanoo moo

183
energy “map”.

In the growth rate and growth efficiency vs. gas composition and
total flow rate experiments that follow, the substrate temperature is
chosen to be fixed at either 900 °C or 950 °C. The CH4 concentration is
varied first in the growth rate vs. gas composition experiments.

6.4.2.2 Effects of CH4 Concentration

Figure 6.15 displays the diamond film linear growth rate. weight
gain and carbon conversion eificiency vs. CH4 gas concentration under
the following experimental conditions: H2 flow rate = 400 sccm, pressure
= 5 1 Torr. absorbed microwave power = 2.34 kW, substrate temperature
~ 900 °C and deposition time = 6 hours.

As shown. the film linear growth rate and weight gain experience
peak values with CH4 to H2 volume ratio of about 6/ 400 = 1.5%. The
carbon conversion efficiency decreases as CH4 concentration is
increased.

Figure 6.16 displays the diamond film linear growth rate, weight
gain and carbon conversion emciency vs. CH4 gas concentration under
the following experimental conditions: H2 flow rate = 400 sccm. pressure
= 57 Torr, absorbed microwave power = 2.5 kW. substrate temperature ~
950 °C and deposition time = 6 hours.

As shown. the film linear growth rate and weight gain experience
peak values with CH4 to H2 volume ratio of about 5/400 = 1.25%. The
carbon conversion efficiency decreases when CH4 concentration is
increased beyond this optimum CH4 concentration.

Figure 6.1 7 displays the locations of the gas compositions used in

Carbon Conversion Efliciency (%)

Growth Rate (um / hour)

 

 

 

 

 

 

  
  
   

 

 

 

7 I I I I I
6 _ \\ ..
5 r- .
4 . .
3 3 4 3 6 7
0.7 . v . . I 1 1.06
Hz = 400 sccm
Ts ~ 900 °C
= 51 Torr A
0.6 , ;= 134 kW . 9.48 B
3 ' silicon wafer ,8
\
no
.8.
0.5 - - 7 90 a
O
H
0 4 6 32 go
. - - . o
3
1 I 1 1 I 474
0'3 3 4 5 6 7
CH4 Flow Rate (sccm)

Figure 6. 15 Linear Growth Rate. Weight Gain and Carbon
Conversion Emcieney versus CH4 Concentration (I)

Carbon Conversion Efficiency (%)

Growth Rate (um / hour)

0.7

0.6

.o
A

.9
m

Figure 6.16 Linear Growth Rate. Weight Gain and Carbon

185

 

 

 

 

#1-
M.
a.

4.5 5.5

 

I I I I I 11.06

 

H2 = 400 sccm
Ts ~ 950 °C
= 57 Torr -
= 2.5 kW

3' silicon wafer

 

l

 

4 4.5 5 5.5 6

CH4 Flow Rate (sccm)

Conversion Emciency versus CH4 Concentration (11)

. 9.48

.c'» .1:
3 8
Weight Gain (mg/ hour)

9
\l
A

S”
y—c
ON

186

a: 58me 893 0-2-0 05 co mcogmoqfioo new Co 66:83 2 .m ennui

“5 f+ov\o x. o
_’_—____L__—_____r___

   

J
/
J
$0
nw.

an; .. 731:6 ”e
e $5.796? #35 no

187

Figures 6.15 and 6.16 on the Bachmann diagram.
The experimental parameters used to obtain the peak growth rate
values are used as the starting experimental parameters in the following

sections.

6.4.2.3 Efi'ects of C02 Concentration

Figure 6.18 displays the diamond film linear growth rate, weight
gain and carbon conversion efiicieney vs. C02 gas concentration under
the following experimental conditions: H2 flow rate = 400 sccm. CH4 flow
rate = 6 sccm, pressure = 51 Torr, microwave power absorbed = 2.34 kW,
substrate temperature ~ 900 °C and deposition time = 6 hours.

As shown, the film linear growth rate. weight gain and carbon
conversion efficiency experience gradual decline as the 002 gas
concentration is increased. Crystalline diamond films are deposited with
these four gas compositions.

Figure 6.19 displays the film linear growth rate, weight gain and
carbon conversion efficiency vs. 002 gas concentration under the
following experimental conditions: H2 flow rate = 200 sccm, CH4 flow rate
= 6 seem. pressure = 51 Torr, microwave power absorbed = 2.46 kW, and
deposition time = 6 hours.

As shown, the film linear growth rate, weight gain and carbon
conversion efiiciency experience sharp increases as C02 flow rate is
increased from 3 to 3.5 sccm, where a dark film is deposited with 3 sccm
C02 gas flow and a crystalline diamond film is deposited with 3.5 sccm
C02 gas flow. The film linear growth rate, weight gain and carbon

conversion efficiency are decreased as C02 gas flow rate is further

Carbon Conversion Efficiency (%)

Growth Rate (11ml hour)

188

 

 

 

 

 

 

   
    

 

 

 

0 l 2 3
0.7 7 I I l 1 1.06
H = 400 sccm
a? = . ism
BS: 1 orr
0-6 " 3" silicon wafer ' 9-48
0.5 - ' 7.90
0.40 1 2 3 6.32

C02 Flow Rate (sccm)

Figure 6.18 Linear Growth Rate, Weight Gain and Carbon

Conversion Emciency versus C02 Concentration (I)

Weight Gain (mg/ hour)

Carbon Conversion Efficiency (%)

Growth Rate (um / hour)

189

 

 

 

 

 

 

 

 

 

4 I I j
3.5 F ' -
3 l- .
2.5 h -
“—73 3.5 4 4'5— '
0.8 . . u . 12.64
0.7 - ‘ 11.06
0.6 - - 9.48
EH: 2060 sccm
0.5 - :45; T3551“ - 7.90
= 2.46 kW
3" silicon wafer
04 . . . 6.32

 

3 3.5 4 4.5
C02 Flow Rate (sccm)

Figure 6. 19 Linear Growth Rate. Weight Gain and Carbon
Conversion Emciency versus C02 Concentration (II)

Weight Gain (mg/ hour)

190

increased from 3.5 sccm to 4.5 sccm. Crystalline diamond films are
deposited in the later three cases.

Figure 6.20 shows the locations of the gas compositions used in
Figures 6.18 and 6.19 on the Bachmann C-I-I-O phase diagram.

From Figures 6.15 and 6.18, we see the optimum gas composition
for high rate diamond film growth at 900 °C is the following: H2 flow rate
= 400 sccm and CH4 flow rate = 6 sccm. This gas composition is used in

the following growth rate versus total flow rate experiments.
6.4.2.4 Efi'ects of Total Flow Rate

Figure 6.2 1 displays the diamond film linear growth rate, weight
gain and carbon conversion emcieney vs. total gas flow rate and gas
residence time under the following experimental conditions: CH4 to H2
volume ratio = 1.5%, pressure = 5 1 Torr, absorbed microwave power a
2.34 kW. substrate temperature ~ 900 °C and deposition time = 6 hours.

Figure 6.22 displays the diamond film linear growth rate. weight
gain and carbon conversion efficiency vs. total gas flow rate and gas
residence time under the following experimental conditions: CH4 to H2
volume ratio = 1.25%, pressure = 57 Torr, absorbed microwave power =
2.5 kW, substrate temperature ~ 950 °C and deposition time = 6 hours.

As shown in Figures 6.21 and 6.22, the film linear growth rate and
weight gain increase with increasing total flow rate when the total flow
rate is low and gas residence time is high. As the total flow rate is further
increased and the gas residence time decreases, the linear growth rate
and weight gain start to saturate. The carbon conversion efficiency

decreases with increasing total flow rate and decreasing gas residence

191

 

=5 apnea 82m 0-3-0 05 8 88.8688 30 do 8:83 one “Ema

en «a $183 5 o

—-_bp—-__-_-_-.—————_—P__~_—_

 

\ again. 7 £\ So

\ an... \fo" <

«Educ- {Noe .
\ an..n£\ .5 u o

192

Residence Time, tr (seconds)

 

 

 

 

 

 

 

 

 

0.8 5.2 2.6 1.3 8:7 635 12.64
—- . .1106 ._
a 0.7 5
O O
J: .l:
\ 0.6L - 9.48 Q,
E. .8.
B 0.5- - 7.90 g
g o
g 0.4- - 6.32 a)
e '8
0 0.3- - 4.74 3

0.2 - - - - - 3.16

50 100 200 300 400
(a)
A 25 I I t l I
SE (1
2‘ 2 -
.2
O
E 15- .
c:
.9
(I) u
a; 101-
5
o 5- -
t:
.8
:3 50 1 00 3
0
H2 Flow Rate (sccm)
(b)

Figure 6.2 1 (a) Linear Growth Rate, Weight Gain and (b) Carbon
Conversion Efiiciency versus Total Flow Rate and Gas Residence Time (I)

193

Residence Time, t, (seconds)

0.75.8 ?9 14:5 9.7 7911.66

 

 

A
V

  

 

 

 

—- '5
g 0.5 - 9.48 E
,g \
\ 50
a 13.
f 0.5 - 790 .5
H

69 53
§ 0.4 [125% - 6.32 g"
6 i:- ~55 03% 3

3silicon wafer
03 I l 1 1 1 474
50 100 200 300 400
(a)

 

20- 4

 

 

 

 

 

030—700 2‘00 300 400

H2 Flow Rate (sccm)
(b)

Carbon Conversion Efficiency (%)

Figure 6.22 (a) Linear Growth Rate, Weight Gain and (b) Carbon
Conversion Emciency versus Total Flow Rate and Gas Residence Time (II)

194

time.
6.4.3 Film Morphologr
6.4.3. 1 Introduction

Scanning Electron Microscopy (SEM) is used to identify the
morphologies of films deposited under various experimental conditions.
The films deposited under the various experimental conditions described
in section 6.4.2 are characterized.

6.4.3.2 Effects of Substrate Temperature

Figure 6.23 displays the SEM photographs of three films deposited
with substrate temperatures: (a) 750 °C. (b) 850 °C and (c) 950 °C. The
pressures and absorbed microwave powers used to achieve these
temperatures are listed in Table 6.1 located above figure 6.8. The other
deposition conditions are: H2 flow rate = 200 sccm. CH4 flow rate = 4
sccm, and deposition time = 4 hours.

As shown. the deposited films evolve from “cauliflower” film at 750
°C, to crystalline film at 850 °C and small grain film at 950 0C.

Figure 6.24 displays the SEM photographs of five films deposited
with substrate temperatures: (a) 750 °C. (b) 800 °C, (c) 850 °C, (d) 900
°C and (e) 950 °C. The pressures and absorbed microwave powers used
to achieve these temperatures are listed in Table 6.2 located above Figure
6. 10. The other deposition conditions are: H2 flow rate = 200 sccm. CH4

flow rate = 2 sccm, C02 flow rate = 2 sccm, and deposition time = 4

(13)

$5.43»

IBM 121681

35:;-

 

Figure 6.23 Effects of Substrate Temperature on Film Morphology (1)
H2 = 200 sccm. CH4 = 4 seem and
= (a) 750. (b) 850 and (c) 950 °C

196

0
6
B
8
1
X

15KU

 

Figure 6.23 (cont'd)

(a)

(b)

 

Figure 6.24 Effects of Substrate Temperature on Film Morphology (II)
H2 = 200 sccm. CH4 = 2 seem, 002 = 2 seem and
T8 = (a) 750. (b) 800. (c) 850. (d) 900 and (e) 950 °C

122101

B
B
8
. 5
1
X

15KU

(d)

XISvBBB

15KU

 

d)

Figure 6.24 (cont’

199

 

Figure 6.24 (cont'd)

(e)

200

hours.

As shown, the films deposited under these conditions are
crystalline films with slightly different average grain sizes. The average
grain size increases gradually as the substrate temperature is increased
from 750 °C to 950 0C.

6.4.3.3 Effects of CH4 Concentration

Film morphology vs. CH4 concentration is shown in Figure 6.25
where the SEM photographs of six films are shown. These films are
deposited under the following experimental conditions: H2 flow rate =
400 sccm, pressure = 5 1 Torr, microwave power absorbed = 2.34 kW.
Substrate temperature ~ 900 °C. deposition time = 6 hours and CH4 flow
rate (a) 3. (b) 4, (c) 5. (d) 6. (e) 6.5 and (fl 7 sccm.

As shown, the average grain size of the films decreases gradually
as the CH4 flow rate is increased fiom 3 to 7 sccm.

Figure 6.26 displays the SEM photographs of five films deposited
under the following experimental conditions: H2 gas flow = 400 scan,
pressure = 57 Torr, absorbed microwave power = 2.5 kW, substrate
temperature ~ 950 °C, deposition time = 6 hours and CH4 flow rate (a) 4,
(b) 4.5, (c) 5, (d) 5.5 and (e) 6 seem.

As shown, the average grain size of the films decreases gradually

also as the CH4 flow rate is increased &0m 4 to 6 sccm.

(b)

 

Figure 6.25 Effects of CH4 Concentration on Film Morphology (I)
H2 = 400 sccm, p = 51 Torr, Pt = 2.34 kW. T$ ~ 900 °C, and
CH4 = (a) 3, (b) 4. (c) 5. (d) 6. (e) 6.5 and (fl 7 sccm

(e)

 

(d)

 

Figure 6.25 (cont'd)

(e)

.a
.a
a

X18

15KU

(fl

X28, 809

15KU

 

nt'd)

Figure 6.25 (co

204

15KU XIB-BBO

(b)

 

Figure 6.26 Effects of CH4 Concentration on Film Morphology (11)
H2 = 400 sccm. p = 57 Torr, Pt = 2.5 kW. Ts ~ 950 °C. and
CH4 = (a) 4, (b) 4.5, (c) 5. (d) 5.5 and (e) 6 seem

(C)

A
15Ku x10.eee

(d)

 

Figure 6.26 (cont'd)

206

B
0
8
0
1
X

15KU

 

Figure 6.26 (cont'd)

207

6.4.3.4 Efi'ects of C02 Concentration

Film morphology vs. C02 concentration is shown in Figure 6.27
where the SEM photographs of four films are shown. These films are
deposited under the following experimental conditions: H2 flow rate =
400 sccm, CH4 flow rate = 6 sccm. pressure = 51 Torr, absorbed
microwave power = 2.34 kW. substrate temperature ~ 900 °C, deposition
time = 6 hours and 002 flow rate (a) O, (b) 1, (c) 2. and (d) 3 sccm.

As shown, the average grain size of the films increases gradually as
the C02 flow rate is increased from 0 to 3 seem.

Figure 6.28 displays the SEM photographs of four films deposited
under the following experimental conditions: H2 flow rate = 200 sccm,
CH4 flow rate = 6 seem. pressure = 51 Torr, absorbed microwave power =
2.46 kW, deposition time = 6 hours and C02 flow rate (a) 3, (b) 3.5. (c) 4.
and (d) 4.5 sccm.

As shown, the average grain size of the films increases gradually

also as the C02 flow rate is increased from 3 to 4.5 sccm.
6.4.3.5 Efi'ects of Total Flow Rate

Film morphology vs. total flow rate is shown in Figure 6.29 where
the SEM photographs of five films are shown. These films are deposited
under the following experimental conditions: CH4 to H2 volume ratio =
1.5%. pressure = 51 Torr, absorbed microwave power = 2.34 kW,
substrate temperature ~ 900 °C, deposition time = 6 hours and H2 flow
rate (a) 50, (b) 100, (c) 200. (d) 300 and (e) 400 sccm.

As shown, the average grain size experiences first a slight increase

208

122701

6
B
8
B
1
X

 

(c) 2 and (d) 3 seem.

51 Torr. Pt = 2.34 kW, Ts ~ 900 0C.

p:

= 6 sccm.
= (a) 0. (b) 1

and C02

400 sccm. CH4

Figure 6.27 Effects of 002 Concentration on Film Morphology (1)

H2

209

(e)

 

(d)

 

Figure 6.27 (cont'd)

(a)

(b)

 

Figure 6.28 Effects of C02 Concentration on Film Morphology (11)
H2 = 200 sccm. CH4 = 6 sccm, p = 51 Torr, Pt = 2.46 kW.
and 002 = (a) 3. (b) 3.5, (c) 4 and (d) 4.5 sccm.

211

(c)

125403

X10,000

(d)

125701

luv:

0
p0
0
0
1
.X

15KU

 

t'd)

Figure 6.28 (con

212

(a)

 

(b)

 

Figure 6.29 Effects of Total Flow Rate on Film Morphology (l)
CH4/H2 = 1.5%. p = 51 Torr. Pt = 2.34 kW, Ts ~ 900 °C, and
H2 = (a) 50. (b) 100. (c) 200. (d) 300 and (e) 400 sccm.

213

(e)

(d)

 

Figure 6.29 (cont'd)

214

6
:1.
a: ‘r

v‘i‘
AL.
“1'4

. ‘s
: 1.

15KU

2..

0
0
0
0
1
X

 

Figure 6.29 (cont'd)

215

when the H2 flow rate is increased from 50 to 100 sccm and then a slight
gradual decrease as the H2 flow rate is further increased toward 400
sccm.

Figure 6.30 displays the SEM photographs of five films deposited
under the following experimental conditions: CH4 to H2 volume ratio =
1.25%, pressure = 57 Torr, absorbed microwave power = 2.5 kW,
substrate temperature ~ 950 °C, deposition time = 6 hours and H2 flow
rate (a) 50, (b) 100, (c) 200, (d) 300 and (e) 400 sccm.

As shown, the average grain size experiences first an increase
when the H2 flow rate is increased from 50 seem to 100 sccm and then a
slight gradual decrease as the H2 flow rate is further increased toward

400 sccm.

6.4.4 Typical Raman Spectrum

figure 6.31 displays the Raman spectrum<76> of a diamond film
deposited on a 3" silicon wafer under the following experimental
conditions: H2 flow rate = 400 sccm. CH4 flow rate = 4 sccm, C02 flow
rate = 2 seem. pressure = 50 Torr, absorbed microwave power = 2.61 kW
and substrate temperature ~ 927 °C. The sharp peak near 1332 cm'1
indicates the excellent quality of the deposited diamond film.

6.4.5 Typical Uniformity Profile
The thickness profiles of a diamond film are displayed in Figures

6.32 and 6.33.<7 1’ They were obtained by laser interference [reflection
technique. The diamond film was deposited on a 3" silicon wafer under

216

15KU X10y000

 

Figure 6.30 Effects of Total Flow Rate on Film Morphology (II)
CH4/H2 = 1.25%, p = 57 Torr. Pt = 2.5 kW, Ts ~ 950 °C, and
H2 = (a) 50. (b) 100. (c) 200. (d) 300 and (e) 400 sccm.

(b)

(C)

(d)

 

Figure 6.30 (cont'd)

218

124101

,e
.__a
.0
a
1
x

15KU

 

Figure 6.30 (cont'd)

219

0.0«ma

I

r I

 

92.82 5 Banana SE ceased ..m a so 358% 58am 86 cams

- P h b — b

o.Ovm«

h

 

 

 

 

20 J<UOEQHUME

0.0wmu

o
b

_ b b b

oo.omm

D

L

oo.oon

 

do Rm z m...
22 as n E
has on u a
.aEQXuNnufluO
fiafiwuyuaflfiu

Boom Dov u «I

00.0mv

3 _

 

 

v.vmo«

0.00Nm

o.oovm

o.oowh

o.oomm

o.ooom«

XiISNBiNI

220

Eur.— ucoEmE ..n a .8 0895 305—08 Ecozoomémouo and van—ma

A 38255.2 v Ratam 8%?
am 2. 2 o o. - om- cm-

 

.00 «mm a we
Bx endure
to? mm H a
EoommanU . . . . .
F... 508$"le ............ ............ ............. ......... :—

Eoom Dov u «m

............ m. ............ m, ............ W ............. W ......... E.

. o a o u o a

v i I I c 0 O

u o u o o o o o
I ......... . ............ . ............ , ............ . .......................... ............ . ......... im—

. c c o u o o

o a o o o o o

o c u o o o o

u u o o o o 0
~ . o o o o a
s . u o o n o
0
c . o a n o .
vl ................................................................................................ 1
u a . o c o o
u u c o o c o
o . s o u u u
o o o

 

 

 

 

 

( SUODFW ) ssauxami ulna

221

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

55 Torr
2.67 kW

P

C02 = 2 seem
Pi

H2 = 400 sccm
CH4 = 4 seem

 

( suomnn ) tuna

   

Wafer Surface ( mm )

Figure 6.33 Three Dimensional Thickness Profile of a 3" Diamond Film

222

the following experimental conditions: H2 flow rate = 400 sccm, CH4 flow
rate = 4 sccm, C02 flow rate = 2 sccm, pressure = 55 Torr, absorbed

microwave power = 2.67 kW and substrate temperature ~ 964 oC.
6.4.6 Summary

From the parametric study of diamond film deposition on 3" silicon
wafers, the results can be summarized as follows:

(1) There exists an optimum substrate temperature range for
high rate diamond film growth with a suitable gas composition. The
growth rate is gas composition dependent.

(2) There exists an optimum CH4 concentration range for high
rate diamond film growth at a suitable substrate temperature.

(3) 3 Addition of C02 to a H2/CH4 discharge tends to dilute some
of the reactive species in the discharge.

(4) The growth rate increases with increasing total flow rate in
the region of 50 -200 sccm and saturates in the region of 300 - 400
sccm, keeping gas composition and substrate temperature constant. This
suggests:

(i) at the low flow rates of 50 - 200 sccm, the deposition rate
is limited by chemically active species hitting the substrate,

(ii) at the higher flow rates of 300 - 400 sccm. some of the
input gas by-passes the deposition process.

(5) Carbon conversion efficiency decreases with increasing total
flow rate and decreasing gas residence time.

(6) Film morphologies depend on substrate temperature. gas

composition and total gas flow rate.

223

(7) Within all the CH4 volume concentrations used, the average
grain size of deposited films decreases as CH4 volume concentration is
increased.

(8) Within all the C02 volume concentrations used, the average
grain size of deposited films increases as CO2 volume concentration is
increased.

(9) Raman spectrum exhibits the excellent characteristics of a
diamond film deposited.

(10) Diamond film with excellent uniformity (better than 2%) has
been deposited.

(l 1) Higher deposition rates may require changing the gas flow
configuration.

(12) The maximum diamond film grth rate obtained on a 3"
silicon wafer is ~ 0.67 pm / hour. The experimental conditions are: H2 flow
rate = 400 sccm, CH4 flow rate = 6 sccm. absorbed microwave power =

2.34 kW, substrate temperature ~ 900 oC and pressure = 5 1 Torr.

6.5 Study of Experimental Set-ups with 4" Silicon Wafers

6.5. 1 Introduction

When diamond films are deposited on 4" silicon wafers in the
present reactor, the experimental set-ups affect the film growth rate,
morphology and uniformity. Using growth rate and uniformity as judging
criteria. a series of experiments are conducted to obtain optimum
experimental set-ups for uniform diamond film deposition on 4" silicon
wafers. The following independent input experimental parameters are

224

considered: (1) cavity shell geometry. (2) quartz dome geometry. (3) wafer
holder material, (4) wafer location. (5) cavity excitation mode and (6)
seeding mixtures.

In this study. diamond films are deposited on 4" silicon wafers
using the forced flow configuration shown in Figure 3. 16. Here the
supporting quartz tube (28) has an inside diameter of l 10 m, an
outside diameter of 1 1 5 mm.

The 4" silicon wafers used in these experiments have the following
physical properties: prime grade. p-type. boron doped. <100> orientation,
24 - 36 ohm-cm in resistivity. 99.5 -100.5 mm in diameter. and 500 -
550 pm in thickness.

6.5.2 Cavity Shell Geometry

The two cavity shells shown in Figures 6.2 and 6.3 are used for
diamond film deposition on 4" silicon wafers. Diamond films were
deposited in the reactors employing the two cavity shells under the
following experimental conditions: H2 flow rate = 400 sccm. CH4 flow rate
= 4 sccm, C02 flow rate = 2 seem. pressure = 47.5 Torr. absorbed
microwave power = 2.58 kW, quartz tube length = 5.09 cm. and
substrate holder material = graphite. The two reactors exhibit the
following difierences as displayed in Table 6.4.

Table 6.4 Comparison of Two Cavity Shells

 

Cavity Shell (1) Cavity Shell (11)
Coupling Efficiency 74.5% 82.8%
Growth Rate (mg/ hour) 10.6 9. l

 

 

 

 

 

 

 

225

Hence. under otherwise the same experimental conditions, the
microwave power coupling efficiency is higher but the diamond film
growth rate is slower in the reactor employing cavity shell (11). It was also
found that when diamond films are deposited on 4" diameter silicon
wafers, the discharge created in the reactor employing cavity shell (II)
tends to float toward the top surface of the quartz dome, which caused
the quartz dome to be coated faster.

(The reactor employing cavity shell (I) has been found to be a

.-L x1. 2 H~"‘"fi
AI

preferable choice for diamond film deposition on 4" silicon wafers with
the present experimental configuration. It is used in all the experiments
that follow.

6.5.3 Quartz Dome Geometry

With the same drawing, quartz domes made by difierent
manufacturers and those constructed at different times by the same
manufacturer present small differences in geometry. When diamond
films are deposited on 4“ silicon wafers. these small differences in
geometry affect the optimum deposition conditions and the resulting film
morphology, uniformity, and growth rate, etc.

TWO quartz domes (Qd#1 and Qd#2) made by the same
manufacture were used for diamond film deposition on 4" silicon wafers
under the following experimental conditions: quartz tube length = 5.2
cm. substrate holder material = graphite, H2 flow rate = 400 sccm. C02
flow rate = 2 sccm, CH4 flow rate = 6 sccm, pressure = 35 Torr, absorbed
microwave power = 2. 1 5 kW and substrate temperature ~ 780 °C. The
deposited films show differences in both growth rate and morpholog as

226

displayed in Table 6.5 and Figure 6.34:

Table 6.5 Comparison of Two Quartz Domes

 

 

 

 

 

 

 

Qd#1 Qd#2
Morphology Figure 6.34 (a) Figure 6.34 (b)
Growth Rate (mg/ hour) 9.42 8.6

 

6.5.4 Substrate Location

Substrate location affects the substrate temperature and the
resulting film properties. such as growth rate. morphology and
uniformity. etc. In the present experimental configuration, the substrate
location is varied by changing the length of quartz tube (28).

I Two experiments were conducted with quartz dome # 1(Qd# l)
and quartz tubes of lengths of 5.09 cm and 5.2 cm. The experimental
conditions are: H2 flow rate = 400 sccm, CH4 flow rate = 6 sccm, C02
flow rate = 2 sccm, pressure = 35 Torr, substrate holder material =
graphite. and absorbed microwave power = 2.12 kW

The effects of substrate location on substrate temperature Ts,
growth rate and uniformity are displayed in Table 6.6, where the
uniformity is judged by visual inspection.

Table 6.6 Efi'ects of Substrate Location (I)

 

Substrate Location Ts (°C) Growth Rate Uniformity

 

5.09 cm 790 10.9 mg/hour good

 

 

 

 

 

5.20 cm 776 9.42 mg/hour fair

 

 

a)

(

227

0
0
0

X10

15KU

 

(b)

m
0
ha.
mun
Mam
=)
mpw
mmm
my.“
66.0
m__9
sea
eC..m
mm“
DEM».
m...
u..%
9%..
«Ch.
S p o
t
mm
355
EWJ
442
W=F
mu

228

As shown. the better choice of quartz tube lengths is 5.09 cm.
judging from both growth rate and uniformity.

II. Three experiments were conducted with quartz dome #2
(Qd#2) and three quartz tubes of lengths of 5.05, 5.09, and 5.2 cm. The
experimental conditions are as follows: H2 flow rate = 400 sccm. CH4
flow rate = 6 sccm. C02 flow rate = 2 sccm. pressure = 35 Torr, and
absorbed microwave power = 2. 16 kW.

The effects of substrate location on substrate temperature Ts,
growth rate and uniformity are displayed in Table 6.7. where the
uniformity was judged by visual inspection.

Table 6.7 Efi'ects of Substrate Location (ll)

 

 

 

 

 

 

Quartz tube length Ts (°C) Growth Rate Uniformity
5.05 cm 791 8.7 mg/hour fair
5.09 cm 784 9.3 mg/hour good
5.20 cm 777 8.6 mg/hour fair

 

 

 

 

 

As shown, the better choice of the quartz tube lengths is again
5.09 cm. judging from both growth rate and uniformity.

5.09 cm long quartz tube is most frequently used in the
experiments where diamond films are deposited on 4" silicon wafers.

6.5.5 Cavity Mode Excitation

The microwave cavity applicator can be excited by TMmn modes for
diamond film deposition, where n is a positive integer. In the present

229

experimental reactor, the excitation of TMm 1 mode is not possible due to
the conflict between cavity short length and excitation probe length
requirements. TM012 and TM013 modes were investigated for their
potentials in diamond film deposition experiments. When the cavity is
excited with T'Mmz mode. the near field efi‘ect from the probe is slightly
present, which causes the plasma to be attracted slightly towards the top
of quartz dome. When the cavity is excited with mom mode, the near
field effect is almost non-existent. Higher order modes. such as TMOM
and mom modes. etc.. were not used since microwave power loss on the
cavity walls is of concern.

The following two experiments were conducted to quantitatively
compare the effects of mo” and Tit/1013 modes. The experimental
conditions were: quartz tube length = 5.09 cm. substrate material =
graphite, H2 flow rate = 400 sccm. CH4 flow rate = 6 sccm. C02 flow rate
= 2 sccm, pressure = 35 Torr, and absorbed microwave power = 2.06 kW.
The effects of cavity mode excitation on substrate temperature, T .
growth rate and uniformity are displayed in Table 6.8. where the
uniformity is judged by visual inspection.

Table 6.8 Effects of Cavity Mode Excitation

 

 

 

 

 

 

 

Excitation Mode Ts (°C) Growth Rate Uniformity
TM012 806 12 mg/hour good
Tit/1013 790 10.8 mg/hour better

 

 

As shown. the diamond film deposited with TMOI3 mode is slightly
more uniform than that with TM012 mode. But the average growth rate of
the diamond film deposited with mom mode is slightly lower than that

230

with mom mode.

Comparing to mom mode excitation. TM013 mode excitation has
the following advantages: (1) there is nearly no near field effect from the
excitation probe. which means that the quartz dome is coated slower
during the diamond film deposition process, and (2) the deposited
diamond film is more uniform. Hence. TMOI3 mode excitation is used in
most experiments described in this chapter.

6.5.6 Seeding Density

In the photo resist seeding method, difi'erent seeding densities
result from different mixtures of diamond powder. photo resist and photo
resist thinner. The effect of seeding density on diamond film growth was
examined by depositing diamond films under otherwise identical
experimental conditions. The two seeding mixtures examined are:

Seeding mixture #1: (a) mix 250 mg 0.1 um diamond powder
with 25 ml of photo resist, (b) ultrasonically mix the mixture for an hour,
(c) apply the mixture onto a silicon wafer, (d) spin the wafer at 4000 rpm
for 30 seconds. (e) bake the wafer at 135 °C for 30 minutes.

Seeding mixture #2: (a) mix 250 mg 0.1 pm diamond powder
with 53 ml of photo resist and 20 ml of photo resist thinner, (b)
ultrasonically mix the mixture for an hour. (c) apply the mixture onto a
silicon wafer, ((1) spin the wafer at 4000 rpm for 40 seconds. (e) bake the
wafer at 135 °C for 30 minutes.

The deposition conditions are: H2 flow rate = 400 sccm, CH4 flow
rate = 4 seem, C02 flow rate = 2 sccm, pressure = 47 .5 Torr. absorbed

microwave power = 2.58 kW, quartz tube length = 5.2 cm. substrate

 

231

holder material = graphite and substrate temperature ~ 860 °C.
The average growth rates are similar with 9.46 mg/ hour for
seeding mixture #1 and 9.8 mg/hour for seeding mixture #2.

6.5.7 Summary

From the study of experimental set-ups for diamond film
deposition on 4" silicon wafers, the results can be summarized as
follows, in diamond film deposition on 4" silicon wafers,

(1) Cavity shell geometry influences diamond film growth. Cavity
shell (I) is a preferable choice.

(2) Variation in quartz dome geometry influences diamond film
growth rate and morphology.

(3) Substrate holder material influences discharge geometry and
substrate heating.

(4) Substrate location influences film growth rate and
uniformity. 5.09 cm long quartz tube is most frequently used substrate
support tube.

(5) Cavity excitation modes influence film growth rate and
uniformity. TMO 13 mode is a preferable choice.

(6) Diamond film growth is not sensitive to slight variation in
seeding density.

232

6.6 Diamond Film Deposition on 4" Silicon Wafers

6.6. 1 Introduction

Diamond films with various physical properties are deposited on 4"
silicon wafers under various experimental conditions. In order to
optimize the diamond film deposition process to achieve the objective of
depositing diamond films with high growth rate and good uniformity
(judged by visual inspection), it is useful to study the co-relation between
the physical properties of the deposited diamond films and the
experimental conditions. Hence a parametric study of diamond film
deposition on 4" silicon wafers was conducted. In this study, the
following experimental parameters are considered:

I The independent input experimental parameters:
(1) absorbed microwave power. Pt.
(2) pressure. p.
(3) gas (H2, CH4, C09) composition. and
(4) total flow rate, ft.
11. The internal experimental parameters:
(1) substrate temperature Ts, and
(2) gas residence time. tr.
III. The output experimental parameters:
(1) film linear growth rate (um/hour),
(2) film weight gain u (mg/hour),
(3) film growth efficiencies:
(i) growth rate vs. power flux, i.e.. 11 vs. S,
(ii) gas flow emciency: k4 (mg/ liter) = u / ft,

233

(iii) carbon conversion emciency. k5 (%).
(4) film morphology.
(5) typical film uniformity, and
(6) typical film Raman spectrum.

In this study, diamond films are deposited on 4" silicon wafers
using the forced flow configuration displayed in Figure 3.16. The
supporting quartz tube (28) has an inside diameter of l 10 m, an
outside diameter of 1 15 mm and a length of 50.9 mm. The physical
properties of the 4" silicon wafers are the same as that described in
section 6.5.1.

Description of relevant experimental chemistry space is also
included whenever possible.

6.6.2 Film Growth Rate and Growth Emciency
6.6.2. 1 Efi'ects of Substrate Temperature

Figure 6.35 displays linear growth rate and weight gain vs.
substrate temperature and power flux under the following experimental
conditions: H2 flow rate = 400 sccm, CH4 flow rate = 6 sccm. and C02
flow rate = 2 sccm. The pressures and absorbed microwave powers used
to achieve these temperatures are listed in Table 6.9 located above Figure
6.35.

As shown, the average linear growth rate and weight gain are
increased as the substrate temperature is increased fiom 8 13 0C to 900
°C. But the film uniformity (from visual inspection) decreases as the
substrate temperature is increased from 813 °C to 900 °C.

Growth Rate (um/ hour)

234

Table 6.9 Substrate Temperature Conditions (4")

 

 

 

 

 

 

Temperature. Ts (°C) Pressure. p (Torr) Power. Pt (kW)
900 45 2.43
836 40 2.28
825 40 2.24
813 . 37.5 2.2

 

 

 

 

Power Flux. S (W / cmz)

 

 

 

O. 6 28.16 2?.1 29.6 31.6 16.22
Hz = 400 sccm
C = 6 seem
0.55 ' cg: = 2 seem ' 14.86
4" silicon wafer
0.5 - " 13.51
0.45 - - 12.16
0.4 - - 10.81
0.35 ' ' 1 I I945

 

813 825 836 900
Substrate Temperature. Ts (°C)

Figure 6.35 Linear Growth Rate and Weight Gain versus
Substrate Temperature and Power Flux

Weight Gain (mg/ hour)

235

Figure 6.36 displays carbon conversion efilciency and gas flow
efiiciency versus substrate temperature and power flux. The carbon
conversion efiiciency and gas flow emciency are increased as substrate
temperature is increased from 813 °C to 900 0C.

In the present reactor. there exists a compromise between the two
objectives. i.e.. high growth rate and good uniformity, in diamond film
deposition on 4" silicon wafers. When a good uniformity is required. the
upper limit on growth rate comes from the upper limit on the substrate
temperature. The substrate temperature is primarily limited by the safety
operation temperature limit and the coating speed of the quartz dome. As
is shown in Figure 6.35. higher growth rate is achieved with higher
substrate temperature. To improve the film uniformity in the higher
substrate temperature case. more absorbed microwave power is required.
The hot discharge is then in close contact with the quartz dome. heating
the quartz dome up to its unsafe operation temperature region and
coating the quartz dome at the same time. A compromise range for the
substrate temperature to serve both obj ectives is between 800 °C to 850
°C.

In the growth rate vs. gas composition and total flow rate
experiments that follow. the substrate temperature is chosen to be fixed
at 845 oC. The CH4 concentration is varied first in the growth rate vs. gas
composition experiments.

6.6.2.2 Effects of CH4 Concentration

Figure 6.37 displays the film linear growth rate. weight gain and
carbon conversion efficiency vs. CH4 gas concentration under the

. ‘—

236

Power Flux, S (W / cm2)

 

 

 

 

 

6 28.'6 29'.1 2'96 31-6 0.63
A H2 = 400 sccm
e\‘: CH4 = 6 seem
8‘ C02 = 2 seem E
c: 5 5 _ 4" silicon wafer .3058 35
Q) ‘ \
'8 1:0
5 *8"
a E‘
O
.0...) 5 _ 40.52 3
is é’
g {:1
3 .2
(1..

o 4. _ _-o.47
g 5 . 33
O 0

1 I 1 . 10.42

813 825 836 900

Substrate Temperature. T$ (°C)

Figure 6.36 Carbon Conversion Emciency and Gas Flow Efficiency
versus Substrate Temperature and Power Flux

Carbon Conversion Efiiciency (%)

Growth Rate (urn / hour)

237

 

 

 

 

 

 

  
 

 

 

4.5 I l I i I
4 .-
3.5 -
3 5'5 6 61.5 5
(a)
045 . 12.15 E
8
Eu
5
0.4 - H2 = 400 scorn . 10.3 5
C02 = 3 seem a
T5 ~ 845 °C "5.0
p = 40 Torr "63
Pt = 2.24 kW 3
4" silicon wafer
0'35 5 5.5 6 6.5 7 9'45
CH4 Flow Rate (sccm)
(b)
Figure 6.37 Linear Growth Rate, Weight Gain and Carbon

Conversion Efficiency versus CH4 concentration

Ir-

238

following experimental conditions: H2 flow rate = 400 sccm. C02 flow rate
= 3 sccm, pressure = 40 Torr. absorbed microwavepower = 2.24 kW,
substrate temperature ~ 845 °C and deposition time = 8 hours.

As shown. the film linear growth rate and weight gain experience
peak values with CH4 to H2 volume ratio of about 6.5/ 400 = 1.625%. But
the film uniformity (judged by visual inspection) in this case is not as
good as that in the case where the CH4 to H2 volume ratio is 6/ 400 =
1.5%. The following experiments of growth rate versus C02 concentration
start from the gas composition of the later case. Also shown in Figure
6.37 is that the carbon conversion efficiency increases initially when CH4
flow rate is increased from 5 to 5.5 sccm and decreases when CH4 flow
rate is further increased. The location of these gas compositions on the

Bachmann C-H-O phase diagram is shown in Figure 6.38.

6.6.2.3 Effects of C02 Concentration

Figure 6.39 displays the film linear growth rate. weight gain and
carbon conversion efficiency versus C02 gas concentration under the
following experimental conditions: H2 flow rate = 400 sccm. CH4 flow rate
= 6 sccm. pressure = 40 Torr. absorbed microwave power = 2.24 kW,
substrate temperature ~ 845 °C and deposition time = 8 hours.

As shown. the film linear growth rate and weight gain experience
peak values with C02 to H2 volume ratio of about 2.5/ 400 = 0.625%. But
the film uniformity (judged by visual inspection) in this case is not as
good as that in the case where the C02 to H2 volume ratio is 3/ 400 =
0.75%. The following experiments of growth rate vs. total flow rate start
from the gas composition of the later case. Also shown in Figure 6.39 is

239

 

CE Edema—Q 3.8.3 0-2-0 05 co mGoEmOQEoO 30 Co .8333 mud 0.59..”

km $+8\o N. o
____L______F________

   
  

J
«.W/
x
rw

an; n £\ N8
«2.78.... ~:\ :8 n o

Carbon Conversion Efficiency (%)

Growth Rate (um / hour)

240

 

 

 

 

 

 

 

 

5 I I I Y
4.5 - -
4 . .
3.5 —- -
3 i a ‘3 a
(a)
8 — 400 sccm
o 45 TH 6%06‘“ 1215
' - s=‘~'4§)A“?orr - °
F: 2.24 kW
4" silicon wafer
04 . .10.8
1 l l I 9.45
0'35 1 2 3 4
C02 Flow Rate (sccm)
(b)

Figure 6.39 Linear Growth Rate, Weight Gain and Carbon
Conversion Efficiency versus 002 concentration

Weight Gain (mg/ hour)

 

241

that the carbon conversion efficiency experiences a peak value at the
peak growth rate point and decreases as the C02 flow rate is further
increased. The location of these gas compositions on the Bachmann C-H-
0 phase diagram is shown in Figure 6.40.

6.6.2.4 Effects of Total Flow Rate

Figure 6.41 displays the film linear growth rate, weight gain and
carbon conversion efficiency versus total gas flow rate and gas residence
time under the following experimental conditions: CH4 to H2 volume ratio
= 1.5%, 002 to H2 volume ration = 0.75%, pressure = 40 Torr, absorbed
microwave power = 2.24 kW, substrate temperature ~ 845 °C and
deposition time = 8 hours.

As shown, the film linear growth rate and weight gain initially
experience an increase as the H2 flow rate is increased from 100 to 300
sccm. They experience a slight decline as the H2 flow rate is further
increased from 300 to 400 sccm. Also shown in Figure 6.41 is that the
carbon conversion efficiency decreases with increasing total gas flow rate

and decreasing gas residence time.

6.6.3 Film Morphology

6.6.3. 1 Introduction

Scanning Electron Microscopy (SEM) is used to identify the

morphologies of films deposited under various experimental conditions.
The films deposited under the various experimental conditions described

242

E gamma ommnm 0-3-0 05 .8 23239800 30 Ho 5583 ové charm

RN 9:83 .5 o
___b__b__L______—__L

  

0
%
$
(w

 

Simeon £\ N8
6 .3.— " £\ :8 u o

243

Residence Time, tr (seconds)

 

 

    
  

 

 

 

 

 

 

 

 

20 10 6.7 5
’5 0.45- .12.15 a
O
4: Q
\ N)
E. a
0
g 0.4 - 8 ~ 845 °C . 10.8 5
8H: 40 Torr H
g 4/H2 = 1.5% go
o coz/Hz = 0.75% o
5 Pt = 2.24 kW 3
”-35 1230 2‘55 166 466—‘9'45
(a)
g 16 I I I I
8‘
i:
2
6:53 12 -
m
r:
.2
s -
E
0
o
8 4 .
.0
a l l l l
0 100 200 300 400
H2 Flow Rate (sccm)
(b)

Figure 6.41 (a) Linear Growth Rate, Weight Gain and (b) Carbon

Conversion Emciency versus Total Flow Rate and
Gas Residence Time

244

in section 6.6.2 are characterized.

6.6.3.2 Effects of Substrate Temperature

Figure 6.42 displays the SEM photogaphs of four films deposited
with substrate temperatures: (a) 900 °C, (b) 836 °C. (c) 825 °C. and (d)
8 13 °C. The pressures and absorbed microwave powers used to achieve
these temperatures are listed in Table 6.9 located above Figure 6.35. The
other deposition conditions are: H2 flow rate = 400 sccm, CH4 flow rate =
6 seem. C02 flow rate = 2 seem, and deposition time = 8 hours.

As shown, the films deposited under these conditions are
crystalline films with different average gain sizes. The film deposited at
900 °C has the largest average gain size and the film deposited at 813
°C has the smallest average gain size.

6.6.3.3 Efi'ects of CH4 Concentration

Film morphology versus CH4 concentration is shown in Figure
6.43 where the SEM photographs of five films are displayed. These films
are deposited under the following experimental conditions: H2 flow rate =
400 sccm, 002 flow rate = 3 seem, pressure = 40 Torr, absorbed
microwave power = 2.24 kW, substrate temperature ~ 845 °C, deposition
time = 8 hours and CH4 flow rate (a) 7, (b) 6.5, (c) 6, (d) 5.5, and (e) 5.

As shown, the film deposited with 7 seem CH4 gas flow has very
small average gain size. The average gain size jumps up when the CH4
flow rate is reduced from 7 to 6.5 sccm. It then decreases gadually as
the CH4 flow rate is further reduced from 6.5 to 5 seem.

(a)

(b)

ISKU X10,000

 

Figure 6.42 Effects of Substrate Temperature on Film Morphology
H2 = 400 sccm. CH4 = 6 seem. C02 = 2 seem. and
T5 = (a) 900. (b) 836, (c) 825 and (d) 813 °C.

15KU

 

246

(c)

X16,800

(d)

Figure 6.42 (cont'd)

(a)

u d l

880

X10:

(b)

 

Figure 6.43 Effects of CH4 Concentration on Film Morphology
H2 = 400 sccm, C02 = 3 seem. p = 40 Torr. Pt = 2.24 kW, Ts = 845 °C.
and CH4 = (a) 7. (b) 6.5. (c) 6. (d) 5.5, (e) 5 seem.

248

 

Figure 6.43 (cont'd)

(c)

(d)

249

 

Figure 6.43 (cont'd)

(e)

250

6.6.3.4 Effects of C02 Concentration

Film morphology versus C02 concentration is shown in Figure
6.44 where the SEM photogaphs of five films are displayed. These films
are deposited under the following experimental conditions: H2 flow rate =
400 sccm. CH4 flow rate = 6 seem, pressure = 40 Torr, absorbed
microwave power absorbed = 2.24 kW. substrate temperature ~ 845 °C,
deposition time = 8 hours and C02 flow rate (a) l, (b) 2, (c) 2.5, (d) 3 and
(e) 4 seem.

As shown, the films deposited with l and 2 seem C02 gas flow
have very small average gain size. The average gain size jumps up when
C02 flow rate is increased to 2.5 sccm. It then experiences a gadual
decline as the C02 flow rate is further increased to 4 seem.

6.6.3.5 Efi'ects of Total Flow Rate

Film morphology versus total flow rate is shown in Figure 6.45
where the SEM photogaphs of four films are displayed. These films are
deposited under the following experimental conditions: CH4 to H2 volume
ratio = 1.5%. C02 to H2 volume ratio = 0.75%. pressure = 40 Torr,
absorbed microwave power = 2.24 kW, substrate temperature ~ 845 °C,
deposition time = 8 hours and H2 flow rate (a) 400, (b) 300, (c) 200, and
(d) 100 sccm.

As shown, the average gain size stays about the same when the
H2 flow rate is decreased from 400 to 300 sccm. It then experiences a
gadual decline as the H2 flow rate is further decreased from 300 to 100

seem.

 

(b)

 

Figure 6.44 Effects of C02 Concentration on Film Morphology
H2 = 400 sccm, CH4 = 6 seem. p = 40 Torr, Pt = 2.24 kW, Ts ~ 845 °C.
and CO2 = (a) 1, (b) 2, (c) 2.5. (d) 3 and (e) 4 seem.

(c)

: A,“

ISKU X10,909

(d)

 

Figure 6.44 (cont‘d)

253

 

Figure 6.44 (cont‘d)

(e)

254

ISKU x16,eae

 

Figure 6.45 Effects of Total Flow Rate on Film Morphology
CH4/H2 = 1.5%, CO2/H2 = 0.75%. p = 40 Torr. Pt = 2.24 kW.
Ts ~ 845 °C. and H2 = (a) 400, (b) 300. (c) 200 and (d) 100 sccm.

(b)

(e)

15KU X10,880

(d)

 

Figure 6.45 (cont’d)

256
6.6.4 Typical Raman Spectrum

Figure 6.46 displays the Raman spectrum<76> of a diamond film
deposited on a 4" silicon wafer under the following experimental
conditions: H2 flow rate = 400 sccm, CH4 flow rate = 6 seem, C02 flow
rate = 3 seem, pressure = 40 Torr, absorbed microwave power = 2.24 kW
and substrate temperature ~ 845 °C. The sharp peak near 1332 cm'1
indicates the excellent quality of the deposited diamond film.

6.6.5 Typical Uniformity Profile

The thickness profiles of a diamond film deposited on a 4" silicon
wafer are displayed in Figures 6.47 and 6.48571) They were obtained by
laser interference / reflection technique. The diamond film was deposited
under the following experimental conditions: H2 flow rate = 400 sccm,
CH4 flow rate = 6 seem, C02 flow rate = 3 seem. pressure = 40 Torr,
absorbed microwave power = 2.24 kW and substrate temperature ~ 845
°C.

6.6.6 Summary

From the parametric study of diamond film deposition on 4" silicon
wafers, the results can be summarized as follows:

(1) There exists a compromising substrate temperature range
for diamond film deposition with high gowth rate and good uniformity.
When a good uniformity is required, the upper limit on film gowth rate
is primarily the safety operation temperature limit and the coating speed

 

257

msmmoz 5 83388 EB 3285 a. a mo 838% :25 98 88E

:0 J<UOIQHUmm
o.o«m« o.ovm« o.omm« oo.omm 00.00h

b b r b _ P b P h - b b h b P b I L b by b b P b L

 

.oo mom 1 m...
- Bx was u so 1
1 t8. 9V u a -
. Boom m u «00 -
I Boom m u 250 l
. Boom Dov u a:

 

 

 

 

 

 

00.0mv

o.vnm«

o.oomm

o . oovm

o . comm

o . 00mm

0.002;

klISNBlNI

258

 

B-E 93:85 ..¢ 8 .«o oaoum mmo—Eo—Fr Bonoommmooo no.0 oar—

ov on

A 382552 0 883m 8.33

 

as. was u on.
to... ov u a

Boom m u «00

Boom m u «:0

Boom oov u a:

.oo mama m...

 

 

 

 

 

I 000000000 O OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO f OOOOOOOOOOOOOOOOOO § OOOOOOOOOOOOOOOOOOOOOOOOOOOO l
o
.
o
u
o o
{I} o
I oooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo l
o o
o
o o
a u
o o
n o
o o
O o
p p - _ -

 

 

nd

w.—

ON

w...

( SUOJDIW ) ssamlom ulna

.........

IIIIIIII
vnmwmn—uq

259

5 I O O

 

 

 

6 sccm

H2 = 400 sccm

CH4

 

 

 

 

 

 

 

( woman ) was

 

Wafer Surface ( mm )

Figure 6.48 Three Dimensional Thickness Profile of a 4" Diamond Film

260

of the quartz dome.

(2) There exists an optimum CH4 concentration range for high
rate diamond film gowth at a suitable substrate temperature.

(3) Addition of 002 to H2 /CH4 discharge tends to dilute some of
the reactive species in the discharge. .

(4) The gowth rate increases with increasing total flow rate in
the region of 100 -200 sccm and saturates in the region of 300 - 400
sccm, keeping gas composition and substrate temperature constant. This
suggests:

(i) at the low flow rates of 100 - 200 sccm, the deposition
rate is limited by chemically active species hitting the substrate.

(ii) at the higher flow rate of 300 - 400 sccm, some of the
input gas by-passes the deposition process.

(5) Carbon conversion efficiency decreases with increasing total
flow rate and decreasing gas residence time.

(6) Film morphologies depend on substrate temperature. gas
composition and total gas flow rate.

(7) Raman spectrum exhibits the excellent characteristics of a
diamond film deposited.

(8) Diamond film with excellent uniformity (better than 2%) has
been deposited.

(9) Diamond film with excellent uniformity and gowth rate is
deposited on a 4" silicon wafer under the following experimental
conditions: H2 flow rate = 400 sccm. CH4 flow rate = 6 sccm, C02 flow
rate = 3 seem. absorbed microwave power = 2.24 kW, substrate
temperature ~ 845 °C, pressure = 40 Torr and gowth rate ~ 0.43 am]

hour.

 

CHAPTER SEVEN

CONCLUSIONS AND RECOMMENDATIONS

7 . 1 Summary of Significant Results

7. 1. 1 Introduction

The research and development described in this dissertation have
lead to the successful creation of a working prototype reactor (MCPR7-3)
of an improved microwave cavity plasma reactor concept and a U. S.
patent application< 17’ . Diamond films with uniformities better than 2%
have been deposited on 3" and 4" silicon wafers with this prototype
reactor. The linear gowth rates obtained are 0.89 m/ hour on 2" silicon
wafers, 0.67 um/hour on 3" silicon wafers and 0.43 um/hour on 4"
silicon wafers. The performance “figures of merit” of diamond film
deposition reactors were developed to compare the MCPRs and other
reactors. Three generations (MCPR7-l , MCPR7-2 and MCPR7-3) of
microwave cavity plasma reactors have been investigated and / or
developed for diamond film deposition over larger substrate areas. Three
generations (MCJR- 1, MCJR-2 and MCJR-3) of microwave cavity jet
reactors have also been developed and / or designed to deposit diamond
films. These reactors have the potential to deposit diamond films with
higher gowth rates. A method to obtain the operational characteristics of

261

262

the MCPRs was developed. The diamond films deposited in the MCPR7-3
prototype reactor were investigated with respect to their gowth rate and
quality as the input experimental parameters were varied. The electric
fields in the microwave cavity plasma reactor (MCPR7- l) were measured
to develop a better understanding of the electromagnetic field / plasma
interactions during diamond film deposition process. These important
accomplishments are summarized in geater detail in the sections that

follow.

7.1.2 MCPR Development

An improved microwave cavity plasma reactor was developed. This
reactor uSes the microwave plasma technology that was developed at
MSU in earlier investigations. That is. the new reactor still employs single
mode excitation and two internal independent tuning variables. i.e..
cavity short and excitation probe length. are used to minimize the
reflected microwave power from the cavity applicator.

Improvements over earlier microwave diamond film reactors are as
follows:

(1) The excitation probe is placed on the sliding short end of the
cavity applicator instead of on the side of the cavity. Near field effects are
eliminated with the use of higher order axial modes. Only TM modes are
excited and the interference from nearby TE modes is eliminated. mom
and TMOI3 modes are used for diamond film deposition. The reactor
operation is more reliable and the reactor is suitable for industrial
applications.

(2) A resonance breaker was placed in the applicator to create a

 

263

stable discharge above the substrate surface.

(3) A larger (5") diameter quartz dome is used and discharges up
to 12.5 cm in diameter are created.

(4) Higher power input (from I .5 kW to 4.5 kW or more) capacity
is achieved with this reactor.

(5) An improved internal gas flow configuration was developed.
Reactive gases are forced to flow close to the substrates and the gas
efficiencies are improved.

This reactor has been used to deposit diamond films in the 20 to
95 Torr pressure range. It can be used to deposit uniform diamond films
on substrate surfaces up to 10 cm in diameter. even without substrate
heating or cooling. Diamond film uniformity better than 2% over 10 cm
diameter silicon wafer has been achieved.

The typical experimental conditions for diamond film deposition on
3" and 4" silicon wafers are as follows:

I. Diamond film deposition on 3" silicon wafers:

pressure = 51 Torr.

H2 flow rate = 400 seem, CH4 flow rate = 6 seem.

absorbed microwave power = 2.34 kW,

substrate temperature ~ 900 °C,

weight gain = 10.6 mg/hour and linear gowth rate = 0.67
um / hour.

II. Diamond film deposition on 4" silicon wafers:

pressure = 40 Torr,
H2 flow rate = 400 sccm, CH4 flow rate = 6 sccm, 002 flow
rate = 3 sccm, .

absorbed microwave power = 2.24 kW,

 

264

substrate temperature ~ 845 °C,
weight gain = l 1.7 mg/hour and linear gowth rate = 0.43

um / hour.

7 . 1 .3 Microwave Plasma Jet Reactor Development

7 . 1 .3.1 Microwave Cavity Jet Reactors

Three generations of microwave cavity jet reactors have been
developed for high rate diamond film deposition.

The first two generation reactors were investigated for diamond
film deposition. One inch diameter quartz tubes were used as the plasma
confinement chamber. The operating pressure and input microwave
power were limited due to quartz tube over-heating. The total gas flow
rate was limited by the pumping speed of the vacuum system.

The third generation reactor was desigied in which the discharge
confinement chamber was enlarged to be a dome. This desigi should
overcome the confinement chamber over -heating problem. It was not

built due to lack of funding.

7.1.3.2 Jet Configuration in MCPR

Preliminary investigation of diamond film deposition using a jet
concept was conducted in the new MCPR7-3. Increased deposition rate (~
1 pm / hour) over a 1.5 cm diameter area was obtained without substrate
cooling. It is expected that higher deposition rates are possible with
proper substrate cooling.

265

7.1.4 Reactor Operational Characteristics

A method to experimentally characterize a microwave plasma
reactor was developed. It was described in sections 6.2 and 6.3 of this
dissertation. This method probably can also be used to experimentally
investigate other plasma reactors. It can be used to locate the required
experimental conditions when diamond films with desired properties.
such as film thickness, morphology, and uniformity, etc. are deposited.

When the discharge occupies only part of the quartz dome volume,
the following conclusions were reached from the study of operational
characteristics of MCPR:

(l) Substrate temperature, Ts. is primarily determined by
operating pressure. Substrate temperature Ts increases sharply with
increasing pressure (~ 6 °C /Torr) and at constant pressure, Ts increases
gadually (~ 0.02 0C /W) with increasing absorbed microwave power. Ts is
not sensitive to variation in gas composition and total flow rate.

(2) For a constant input power, the discharge volume, Vd,
decreases with increasing pressure. At a constant pressure. Vd increases
with increasing absorbed microwave power. Va is not sensitive to
variations in gas composition and total flow rate.

(3) Average discharge power density. w, increases with
increasing pressure and at constant pressure, it is not sensitive to
variation in absorbed microwave power. That is, under a constant
pressure, as the absorbed microwave power increases, the discharge
volMe also increases, resulting in only small variations in absorbed
power density. Power density is not sensitive to variations in gas

composition and total flow rate.

266

(4) Power flux S to the substrate increases with increasing
absorbed microwave power. It is not sensitive to variations in gas

composition and total flow rate.

7.1.5 Diamond Film Deposition on 3" Silicon Wafers

A parametric study of diamond film deposition on 3" silicon wafers
was conducted. The results of this study can be summarized as follows:

(1) There exists an optimum substrate temperature range (~ 800
- 1000 °C) for high rate diamond film gowth with a suitable gas
composition. The gowth rate is gas composition dependent.

(2) There exists an optimum CH4 concentration range for high
rate diamond film gowth at a suitable substrate temperature.

(3) Addition of C02 to H2/CH4 discharge tends to dilute some of
the reactive species in the discharge.

(4) The gowth rate increases with increasing total flow rate in
the region of 50 -200 sccm and saturates in the region of 300 - 400 -
sccm, keeping gas composition and substrate temperature constant. This
suggests:

(i) at the low flow rates of 50 - 200 seem. the deposition rate
is limited by chemically active species hitting the substrate.

(ii) at the higher flow rates of 300 - 400 sccm, some of the
input gas by-passes the deposition process.

(5) Carbon conversion efficiency decreases with increasing total
flow rate and decreasing gas residence time.

(6) Film morphologies depend on substrate temperature, gas

composition and total gas flow rate.

 

267

(7) Within all the CH4 concentrations used, the average gain
size of deposited films decreases as CH4 concentration is increased.

(8) Within all the 002 concentrations used, the average gain-
size of deposited films increases as C02 concentration is increased.

(9) Raman spectrum exhibits the excellent characteristics of a
diamond film deposited.

(10) Diamond film with excellent uniformity (better than 2%) has
been deposited.

(1 1) Higher deposition rates may require changing the gas flow
configuration.

(12) The maximum diamond film gowth rate obtained on a 3"
silicon wafer is ~ 0.67 ml hour. The experimental conditions are: H2 flow
rate = 400 sccm. CH4 flow rate = 6 seem, absorbed microwave power =
2.34 kW, substrate temperature ~ 900 °C and pressure = 5 1 Torr.

7.1.6 Diamond Film Deposition on 4" Silicon Wafers

7.1.6. 1 Study of Experimental Set-ups

A study of experimental set-ups for diamond film deposition on 4”
silicon wafers was conducted, the results of this study can be
summarized as follows:

(1) A slight variation in cavity shell geometry influences
diamond film gowth. Cavity shell (I) is a preferable choice.

(2) A slight variation in quartz dome geometry influences
diamond film gowth rate and morphologr.

(3) Substrate holder material influences discharge geometry and

268

substrate heating. Graphite is a preferable substrate holder material.

(4) Substrate location influences film gowth rate and
uniformity. 5.09 cm long quartz tube is most frequently used substrate
support tube.

(5) Cavity excitation modes influence film gowth rate and
uniformity. TM013 mode is a preferable choice.

(6) Diamond film gowth is not sensitive to slight variations in
seeding density.

7.1.6.2 Diamond Film Deposition on 4" Silicon Wafers

A parametric study of diamond film deposition on 4" silicon wafers
was conducted, the results of this study can be summarized as follows:

(1) There exists a compromising substrate temperature range
for diamond film deposition with high gowth rate and good uniformity.
When a good uniformity is required. the upper limit on film gowth rate
is primarily the safe operating temperature limit and the coating rate on
the quartz dome.

(2) There exists an optimum CH4 concentration range for high
rate diamond film gowth at a suitable substrate temperature.

(3) Addition of C02 to H2 [CH4 discharge tends to dilute some of
the reactive species in the discharge.

(4) The gowth rate increases with increasing total flow rate in
the region of 100 -200 sccm and saturates in the region of 300 - 400
sccm, keeping gas composition and substrate temperature constant. This
suggests:

(i) at the low flow rates of 100 - 200 sccm, the deposition

269

rate is limited by chemically active species hitting the substrate.
(ii) at the higher flow rate of 300 - 400 sccm, some of the
input gas by-passes the deposition process.

(5) Carbon conversion efficiency decreases with increasing total
flow rate and decreasing gas residence time.

(6) Film morphologies depend on substrate temperature, gas
composition and total gas flow rate.

(7 ) Raman spectrum exhibits the excellent characteristics of a
diamond film deposited.

(8) Diamond film with excellent uniformity (better than 2%) has
been deposited.

(9) Diamond film with excellent uniformity and gowth rate is
deposited on a 4" silicon wafer under the following experimental
conditions: H2 flow rate = 400 sccm, CH4 flow rate = 6 seem. C02 flow
rate = 3 seem, absorbed microwave power = 2.24 kW, substrate
temperature ~ 845 °C, pressure = 40 Torr and gowth rate ~ 0.43 ml

hour.

7. 1 .7 Reactor Performance “Figures of Merit”

7 .l.7 .1 Performance “Figures of Merit" of MCPR

Performance “figures of merit“ were developed for comparison
between different goup of reactors and for comparison of the same
reactor under different operating conditions.

I. The MCPRs have the following performance “figures of merit“
as listed in Table 7 . 1.

270

Table 7 . 1 Summary of "Figures of Merit” of

Diamond Film Deposition Experiments in MCPR

 

 

 

 

 

 

 

 

 

 

 

 

 

Gas
Linear Energr Carbon
Exp. gowth Dep. Weight effic. flow conv.
area gain effic.
No. rate (cm2) (m g /hr) (mg/ (m g / effic.
(um/hr) kW-hr) (%)
liter)

MCPR 2.25- 4.5- 0.19- 3.5-
7- l 0.4-0.8 16 4.5 9 0.38 7
MCPR 2.8- l .7 5- 0. 15- 2.9-

7-2 0.4-0.8 20 5.6 3.5 0.3 5.8
MCPR
7-3 0.89 20 6.2 3.3 0.26 2.4
NT#4
WT#27 0.67 45 10.6 4.5 0.44 5.5
WT#29 0.65 45 10.3 4.4 0.57 7 . 1
WT#30 0.62 45 9.75 4.2 0.8 l 10. l
WI‘#3 1 0.5 1 45 8 3.4 1 .3 16.6
WF#75 0.43 78 1 1.7 5.2 0.49 4
WF#84 0.44 78 l 1.9 5.3 0.66 5.5
WF#85 0.42 78 l 1.4 5.1 0.95 7.9
WF#86 0.37 78 9.9 4.4 1.65 13.7

 

 

 

 

 

 

 

 

271

11. Comparing the experiments conducted in the MCPRs. the
following conclusions can be reached:

(1) With the same deposition area, the diamond films deposited
in MCPR7-3 have the highest linear gowth rate and weight gain.

(2) Among the uniform diamond films deposited in MCPR7-3, as
the deposition area increases, the linear gowth rate decreases and the

weight gain increases.

7. 1.7 .2 Comparison with other Reactors

Comparing with other reactor concepts reviewed in chapter 2 and
referring to Table 2.1 , the following conclusions can be reached:

(1) The linear gowth rate in MCPR is lower than that in most
high pressure (higher than 100 Torr) reactors.

(2) The uniform diamond film deposition area (78 cm2) in MCPR
is larger than that in the other reactors reported in the literature.

(3) The weight gain of the diamond films deposited in MCPR is
higher than that in the other low pressure (less than 100 Torr) reactors.
It is lower than that in some high pressure (more than 100 Torr) reactor.

(4) The energr efficiency of MCPR is comparable to most other
diamond film deposition reactors.

(5) The gas flow efficiency of MCPR is higher than most high
pressure diamond film deposition reactors.

(6) The carbon conversion efficiency of MCPR is higher than that
of other diamond film deposition reactors.

(7) The excellent uniformity (better than 2%) of the diamond
films deposited in MCPR is better than that in the other reactors reported

272

in the literature.

7.1.8 _ Measurement of Electric Fields in MCPR

The (relative) spatial variations of the exciting electromagnetic field
patterns in MCPR7- 1 under diamond film deposition conditions were
measured. These spatial electric field measurements and intensities are
related to other important experimental parameters such as gas mixture,
flow rate. input power, substrate temperature, and discharge pressure.
etc.

The following conclusions were reached from this study. under
diamond film deposition condition in MCPR7-1:

(1) The excitation mode was identified to be discharge loaded

mm 1 mode.
(2) Maximum electric field intensity in the cavity was ~ 150 V/

(2) Discharge loaded cavity quality factor was ~ 60.

(3) The tangential component of the E field is the main
discharge excitation field.

(4) Average power density in the discharge was ~ 6 W/cm3 at 20
Torr and 20 W/cm3 at 70 Torr.

(5) Under a constant pressure, the square of the cavity electric
field strength is approximately proportional to the power input and only
slightly influenced by the variation in flow rates.

(6) At a constant gas flow, the square of the cavity electric field
strength is approximately proportional to the power input and only
slightly influenced by the variation in pressure.

273

(7) Under a constant pressure, the substrate temperature
increases with the power input (~ 0.4 °C/W) and is only slightly
influenced by the variation in flow rates.

(8) At a constant gas flow. the substrate temperature increases
with the power input (~ 0.4 °C/W) under each working pressure. Also. at
the constant gas flow and input power, the substrate temperature
increases (~ 3.5 °C/Torr) with increasing pressure.

7.2 Recommendations for Future Research

(1) For diamond film deposition in MCPR7-3, higher gowth
rates may be achieved by using higher pressure (100 - 200 Torr). higher
power (3 - 6 kW) and substrate cooling. A substrate cooling stage has
been built. Different gas flow configurations should also be investigated
for their potential to improve the diamond film gowth rates.

(2) The MCPR concept may be scaled up to larger diameter (14 -
20 inch) cavities by dropping the microwave input power frequency to
915 MHz. Larger area substrates may be covered with increasing input
microwave power. Also, with proper reactor design, the diameter of the
cavity operated with 2.45 GHz power may be enlarged to 8-12 inches.

(3) The improved microwave cavity jet reactor (MCJR-3) should
be built and investigated for its potential in diamond film deposition with
higher gowth rates.

(4) The jet configuration in MCPR7-3 offers an interesting
alternative approach to deposit diamond films with higher gowth rates.
Experiments should be conducted to investigate this potential, using a
substrate cooling system.

274

(5) In order to obtain a better understanding of the
electromagnetic field / plasma interactions, electric field measurements in
MCPR7-3 should be conducted. A new cavity with electric field
measurement probe holes has been built.

(6) In order to improve the knowledge of the fundamental
plasma/ chemical reactions of diamond film deposition and to be able to

 

better control the diamond film deposition process. it is necessary to

"1

identify the discharge species essential for diamond film deposition. It is
helpful to characterize the microwave discharge under diamond film
deposition conditions by emission spectroscopy. actinometry and laser
induced fluorescence (LIF) spectroscopy. The knowledge obtained in this
investigation will be beneficial to achieve single crystal diamond film
gowth which is essential for some industrial applications, such as
diamond electronic and optical devices.

(7) For some industrial applications. it is necessary to deposit
diamond film on substrates at low substrate temperature. The research
on diamond film deposition at lower substrate temperatures and / or low

pressures should be conducted.

 

(8) For different industrial applications, diamond films with
different morphologies are required. Since only a preliminary
investigation of diamond film morphologies versus experimental
conditions was conducted, more detailed investigation of diamond film
morphologies versus experimental conditions should be conducted.

(9) Alternative gas compositions, such as carbon monoxide (CO),
oxygen (02), acetone (CH3COCH3) and acetylene (C2H2), etc.. should be
considered to improve the gas flow efficiency and diamond film gowth

rate.

LIST OF REFERENCES

[ll

[2]

[3]

l4]

l5]

[6]

[7]

[8]

[9]

[10]

LIST OF REFERENCES

S. Tennant, “On the nature of the diamond,“ Phil. Trans. Roy.
Soc. London, 87, 123 (1797).

J.C. Angus and QC. Hayman, Science 241, 913 (1988).

R. Orr, "The impact of thin film diamond on advanced
engineering systems," presented at the global business and
technical outlook for high performance inorganic thin film
coatings, Gorham advanced materials institute, Monterey,
California. October 30 - November 1, 1988.

F. P. Bundy, H. T. Hall, H. M. Strong, and R. H. Wentof, Jr.,
“Man-made diamond,“ Nature, 176, 51 (1955).

W. G. Eversole, U. S. Patents 3030187 and 3030188.

J. C. Angus, H. A. Will, and W. S. Stanko. J. Appl. Phys. 39,
2915 (1968).

B. V. Deryaguin. D. V. Fedoseev, V. M. Lukyanovich. B. V.
Spitsyn, V. A. Ryabov, and A. V. Lavrentyev, J. Cryst. Growth 2,
380 (1968).

B. V. Spitsyn, L. L. Bouilov and B. V. Deryagin, J. Cryst.
Growth 52, 219 (1981).

S. Matsumoto, Y. Sato, M. Kamo, N. Setaka, Jpn. J. Appl. Phys.
21, part 2, 183 (1982).

J. Rogers, “Properties of steady-state. high pressure argon
discharges, " Ph.D. Dissertation, Michigan State University,

275

[ll]

[12)

[13]

[14]

[15]

[16]

[17]

[18]

[19]

276

1982.

S. J. Whitehair, “Experimental development of a microwave
electrothermal thruster,” Ph.D. Dissertation, Michigan State
University (1986) .

M. Dahimene, “Development of a microwave ion and plasma
source immersed in a multicusp electron cyclotron resonant
magnetic field,“ Ph.D. Dissertation, Michigan State University
(1987).

 

; x

J. A. Hopwood, “Macroscopic properties of a multipolar electron
cyclotron resonance microwave cavity plasma source for
anisotropic silicon etching,“ Ph.D. Dissertation, Michigan State
University (1990) .

J. Zhang, B. Huang, D. K. Reinhard and J. Asmussen, J. Vac.
Sci. Technol. A 8. 2124 (1990).

J. Zhang, G. King, T. Grotjohn, and J. Asmussen, “Diagnostic
measurements of a resonant cavity microwave plasma diamond
deposition reactor,” presented at Third International Conference
on the New Diamond Science and Technology, Heidelberg.
Germany, August 31 - September 4, 1992. -

P. Mak, G. King, M. Ulczynski, J. Zhang, T. Grotjohn, and J.
Asmussen, “Experimental diagiosis of low pressure microwave
discharges during diamond thin film deposition,“ presented at
Third International Conference on the New Diamond Science
and Technology, Heidelberg, Germany, August 3 l - September
4, 1992.

 

J. Asmussen and J. Zhang, "Improved apparatus for the coating
of material on a substrate using a microwave or UHF plasma.”
U. S. Patent Application, (May 28, 1992).

S. Matsumoto, Y. Sato, M. Kamo and N. Setaka, Jpn. J. Appl.
Phys. 21. 183 (1982).

S. Matsumoto, Y. Sato, M. Tsutsumi and N. Setaka, J. Mater.

[20)

[21]

[22]

[231

[241

[25]

[26]

[27)

[28]

[29)

[30]

[31]

[32]

277

Sci., 17, 3106 (1982).

Y. H. Lee, P. D. Richard, K. J. Bachmann, and J. T. Glass. Appl.
Phys. Lett. 56, 620 (1990).

K. Baba, Y. Aikawa, and N. Shobata, J. Appl. Phys. 69, 7313
(1991).

A Masood, “Technology of electronic properities of CVD
diamond film microsensors for thermal signals,” Ph.D.
dissertation, Michigan State University (1992) .

K. Suzuki, A. Sawabe, H. Yasuda and T. lnuzuka, Appl. Phys.
Lett. 50, 728 (1987).

K. Suzuki, A. Sawabe, and T. lnuzuka, Jpn. J. Appl. Phys. 29,
153 (1990).

S. Matsumoto. I. Hosoya and T. Chounan, Jpn. J. Appl. Phys.
29, 2082 (1990).

K. V. Ravi, C. A. Koch, H. S. Hu, and A. Joshi, J. Mater. Res., 5,
2356 (1990).

L. M. Hanssen, W. A. Car-rington, J. E. Butler and K. A. Snail,
Mater. Lett. 7. 289 (1988).

S. Matsumoto, J. Mater. Sci. Lett., 4, 600 (1985).

S. Matsumoto. M. Hino and T. Kobayashi, Appl. Phys. Lett.. 51,
737 (1987).

M. Kamo, Y. Sato, S. Matsumoto and N. Setaka, J. Cryst.
Growth 62, 642 (1983).

H. Kawarada, K. S. Mar and A. Hiraki, Jpn. J. Appl. Phys. 26,
L1032 (1987).

J. J. Chang, T. D. Mantei, R. Vuppuladhadium and H. E.
Jackson, Appl. Phys. Lett. 59. 1170 (1991).

 

[33]

[34]

[35]

[36]

[37]

[38]

[39]

[40]

[41]

[42]

[43]

278

T. D. Mantei, Third Diamond Workshop, Detroit, MI, March,
1992.

Y. Liou, A. Inspektor, R. Weimer and R. Messier, Appl. Phys.
Lett. 55. 631 (1989).

P. K. Bachmann, W. Drawl, D. Knight, R. Weimer, and R. F.
Messier, in Extended Abstracts Diamond and Diamond-like
Materials Synthesis, edited by G. J ohnson, A. Badzian, and M.
Geis, Materials Research Society, Pittsburgh, PA, 1988, p. 99. '

P. Bachmann, D. Leers and H. Lydtin, Diamond and Related
Materials. 1. l (1991).

Y. Mitsuda, T. Yoshida and K. Akashi, Rev. Sci. Instrum. 60,
249 (1989).

K. Takeuchi and T. Yoshida, J. Appl. Phys. 71. 2636 (1992).

L. Mahoney, “The desigr and testing of a compact electron
cyclotron resonant microwave-cavity ion source.” MS. Thesis,
Michigan State University (1989).

J. T. Salbert, “Anodic gowth and cathodic removal of silicon
dioxide layers utilizing an electron cyclotron resonant
microwave plasma disk reactor,“ Ph.D. Dissertation, Michigan
State University (1992)

The computer monitor system was designed and built by Z. M.
Rummler with help from M Ennis. The monior progam was
written by Z. M. Rummler at Michigan State University, East
Lansing, Michigan.

J. Asmussen, J. Root, “Ion generating apparatus and method
for the use thereof,“ U. S. Patent No. 4,507,588, March 26,
1985.

J. Asmussen, S. Nakanishi, J. Malinkey and S. Whitehair,
“Microwave electrothmster,’ patent applied for by NASA-Lewis.

[44]

[45]

[46)

[47]

[48]

[49]

[50]

[51]

[52]

[53]

[54]

[55]

279

J. Asmussen, D. K. Reinhard, “Method for treating a surface
with a microwave plasma and improved apparatus,“ U. S.
Patent No. 4,585,688, April 29, 1986.

J. Asmussen, D. K. Reinhard, “Microwave or UHF plasma and
improved apparatus,“ U. S. Patent No. 4,630,566, Dec. 23,
1986.

T. Ropple, D. K. Reinhard and J. Asmussen, “Dual plasma
microwave apparatus and method for treating a surface,“ U. S.
Patent No. 4,691,662. Sept. 8, 1987.

J. Asmussen, D. K. Reinhard and M. Dahimene, “Improved
plasma generating apparatus using magrets and method,“ U. S.
Patent No. 4,727, 293, Feb. 23, 1988.

J. Asmussen, “Method for treating a material using
radiofrequency waves,“ U. S. Patent No. 4,777,336, Oct.
1 1 . 1988.

J. Asmussen, “Improved microwave apparatus,“ U. S. Patent
No. 4,792,772, Dec. 20, 1988: also Canadian Letters Patent No.
1.287.666, Aug. 13. 1991.

J. Asmussen, D. K. Reinhard, “Improved plasma reactor
apparatus and method for treating a substrate," U. S. Patent
No. 4.943.345. July 24, 1990.

J. Root and J. Asmussen, Rev. Sci. Instrum., 56, 154 (1985).

T. Roppel, D. K. Reinhard, and J. Asmussen, J. Vac. Sci.
Technol.. B4, 295 (1986)

J. Asmussen and M. Dahimene, J. Vac. Sci. Technol., B5, 328
(1987).

J. Asmussen, J. Vac. Sci. Technol., A7, 883 (1989).

L. Mahoney, M. Dahimene, and J. Asmussen, Rev. Sci.
Instrum., 59, 448 (1988).

[56]

[57)

(58)

[59]

[60]

[61]

[62]

[63]

[64)

[65]

[66)

[67]

[68]

280

S. Whitehair, J. Asmussen and S. Nakanishi, J. Propul. Power,
3, 136 (1987).

This figure was provided by J. Asmussen, Michigan State
University, East Lansing, Michigan.

B. Huang, “Electrical properties and physical characteristics of
polycrystalline diamond films deposited in a microwave plasma
disk reactor,“ Ph. D. Dissertation, Michigan State University
(1992).

Desigred and built by W. Richardson in 1990 at Michigan State
University, East Lansing, Michigan.

Desigied and built by Union Carbide Coatings Service Co. ,
Cleveland, OH.

Developed by Wavemat Inc. , Plymouth, Michigan.

Designed and built by F. C. Sze at Michigan State University,
East Lansing, Michigan.

J. Asmussen, R. Mallavarpu, J. R. Hamann, and H. C. Park,
“The design of a microwave plasma cavity", Proc. IEEE, 109
(1974).

R. Mallavarpu, M. C. Hawley and J. Asmussen, “Behavior of a
microwave cavity discharge over a wide range of pressures and
flow rates,” IEEE Tran. on Plasma Science, PS-6, 341-354
(December, 1978).

R. F. Harrington, Time-Harmonic Electromagretic Fields
(McGraw-Hill, New York, 1961)

R. L. Kapitza, Sov. Phys. JETP 30. 973 (1970).

L. M. Baltin, V. M. Batemin, I. I. Devyatkim, V. P. Lebedeva, N.
I. Tsemko, High Temp. 9, 1019 (1971).

Y. Arata, S. Miyake, A. Kubayashi, and S. Takuchi, Jpn. J.

[69)

(70]

[71)

[72]

[73)

[74]
(75)

I76)

281

Phys. 40. 1456 (1976).

R. Chapman, J. Filpus, T. Morin, R. Snellenberger, J.
Asmussen, M. Hawley, and R. Kerber, J. Spacecraft 19, 579
(1982).

J. Asmussen, H. H. Lin, B. Maming, and R. Fritz, Rev. Sci.
Instrum. 58. 1497 (1987).

Film uniformity measurements by laser interference / reflection
technique were conducted by T. Theissen at Michigan State
University, East Lansing, Michigan.

A. Masood, M. Aslam, M. A. Tamor and T. J. Potter, J.
Electrochem. Soc. 138, L67 (1991)

A. Masood, private communication, Michigan State University,
East Lansing, Michigan.

Amplex, 16 Britton drive, bloomfield, connecticut 06002
Shipley company, 9 1 Bartlett Street, Marlboro, MA 02 162

The Raman spectra were provided by Scott Martin at Norton
Diamond Film, N orthboro, MA.

 

...a. a»....k...fl.u.ui..:...m
1 $01. “Rhoda“

g“. 332......

53¢- r...
. .1! !¢ a:
21

 

a: , r .
. lvfiflnu.

.
«3.0... . 13:3:
kirk?

1.. .m
(-