TOWARD THE RAPID GROWTH OF HIGH-QUALITY, POLYCRYSTALLINE RIMLESS,
AND LARGE AREA SINGLE CRYSTAL DIAMOND SUBSTRATES
By
Amanda Charris-Hernandez

A DISSERTATION
Submitted to
Michigan State University
In partial fulfillment of the requirements
for the degree of
Electrical Engineering - Doctor of Philosophy
2017

ABSTRACT
TOWARD THE RAPID GROWTH OF HIGH-QUALITY, POLYCRYSTALLINE RIMLESS,
AND LARGE AREA SINGLE CRYSTAL DIAMOND SUBSTRATES
By
Amanda Charris-Hernandez
Silicon has been the common choice as semiconductor material since the 1960s, and
it still constitutes more than 90 percent of the electronic device types available in the
market. The most pressing and visible problem with silicon is its thermal conductivity. The
use of cooling systems is inefficient and serve as a major source of waste. Electronic
applications require an alternative to silicon that enables devices to be smaller, cooler,
faster, powerful, and cleaner. Diamond is the ideal alternative material; it can perform at
higher temperatures than silicon without degrading, it can be efficiently cooled, and it
tolerates high voltages before breaking down. Therefore, the growth of high-quality single
crystal diamond is imperative for further advancing optical and electronic applications of
diamond. Industrial applications require large and defect-free SCDs grown at high growth
rates and at low cost. The primary objective of this investigation is to address the big
challenges facing the diamond growth: (1) the rapid growth, (2) the growth of high quality,
and (3) the large volume growth of single crystal diamond (SCD). In this thesis research
activity, the growth of SCD is performed in a MSU-design microwave plasma cavity
reactor at a constant pressure of 240 Torr and high discharge power densities that are >
500W/cm3. SCD substrates were successfully grown using specially selected and
optimized pocket holder designs. These pocket holder designs create an appropriate
thermal environment to shield the diamond substrate from the intense microwave

discharge. A growth recipe using these pocket holders was developed. This growth recipe
controls the substrate temperature (Ts) and the incident microwave power (Pinc) in a
prescribed function of growth time. This growth recipe allows for a polycrystalline diamond
(PCD) rimless, uniform single SCD growth process. In one continuous process run PCD
rimless and increased diamond surfaces areas of over 2 times were achieved.
Specifically, an extensive exploration of the pocket holder diamond growth process was
experimentally performed using 3.5 mm x 3.5 mm diamond seeds. During these
experiments, the substrate temperature Ts was held constant at 1020°C ± 5°C while the
growth time, input power, and pocket depth were varied. Thick (up to 1.9 mm) and
enlarged surfaces (up to 2 times) were grown in one continuous process run. After laser
cutting and polishing high quality diamond plates with low-stress and low defect densities
were obtained. By varying the pocket aperture, enlarged PCD rimless surfaces (up to 2.3
times) were grown in one process run. It was found that, there is a range of optimal pocket
depths and apertures that can grow SCD both vertically and horizontally. One more
diamond growth step was added to the smooth and large diamond top surfaces of the
first growth step. In these extra growth experiments, the resulting grown SCD substrates
had a top surface area that was enlarged up to 2.7 times. Finally, the lateral SCD growth
versus time was photographically evaluated while the growing diamond substrate still was
completely in the pocket. Results show the lateral growth of SCD as a function of time
and also show the PCD growth on the substrate holder as a function of the time. This
photographic technique is very useful in investigating in-situ CVD diamond growth as a
function of time and can be employed in many other CVD diamond growth experiments.

Copyright by
AMANDA CHARRIS-HERNANDEZ
2016

Diamonds are forever, just as:
Sonia, Jonathan, Daimer, Ainara, and Sugi

v

TABLE OF CONTENTS

LIST OF TABLES ............................................................................................................ x
LIST OF FIGURES ..........................................................................................................xi
CHAPTER 1 .................................................................................................................... 1
INTRODUCTION ............................................................................................................. 1
1.1 Motivation .............................................................................................................. 1
1.2 Research Objectives.............................................................................................. 6
1.3 Thesis outline ...................................................................................................... 11
CHAPTER 2 .................................................................................................................. 14
LITERATURE REVIEW: DIAMOND MATERIAL AND SINGLE CRYSTAL DIAMOND
GROWTH VIA MICROWAVE PLASMA ASSISTED CHEMICAL VAPOR DEPOSITION
...................................................................................................................................... 14
2.1 Introduction .......................................................................................................... 14
2.2 Structure and Material Properties of Diamond ..................................................... 14
2.3 Historical Overview of the CVD Diamond Process .............................................. 18
2.4 Processes of Diamond Growth ............................................................................ 19
2.4.1 Phase Diagram of Carbon ......................................................................... 19
2.4.2 HPHT Diamond Growth ............................................................................. 20
2.4.3 CVD Diamond Growth ............................................................................... 22
2.5 Physical and Chemical Aspects of CVD Diamond ............................................... 22
2.5.1 General Overview of the Surface Reactions .............................................. 22
2.5.2 The Specific Diamond Growth Reactions .................................................. 27
2.5.3 Growth of Defects ...................................................................................... 31
2.5.4 Numerical Modeling of the Harris-Goodwin Model .................................... 33
2.6 MPACVD Growth of Diamond ............................................................................. 35
2.6.1 Pyramidal Hillocks and Non-Epitaxial Crystallites .................................. 38
2.6.2 Step bunching ........................................................................................ 40
2.6.3 Effect of Growth Parameters .................................................................. 42
2.7 Review of Single Crystal Diamond Growth via MPACVD .................................... 53
2.7.1 Carnegie Institute of Washington, USA ..................................................... 54
2.7.2 Diamond Research Center, AIST, and Osaka University, Japan............... 56
2.7.3 LIMHP/CNRS, Paris, France – Recent Results ......................................... 70
2.7.4 Gemological Institute, China University of Geosciences, China ................ 85
2.7.5 Michigan State University, USA ................................................................. 89
CHAPTER 3 ................................................................................................................ 101
DESCRIPTION OF THE EXPERIMENTAL EQUIPMENT AND EXPERIMENTAL
PROCEDURE ............................................................................................................. 101
3.1 Introduction ........................................................................................................ 101
3.2 Experimental Setup ........................................................................................... 101
vi

3.2.1 Microwave Power Supply subsystem ...................................................... 103
3.2.2 Process Gas Delivery Subsystem ........................................................... 104
3.2.3 Vacuum Pumping Control Subsystem ..................................................... 104
3.2.4 Air and Water Cooling Control Subsystem .............................................. 106
3.2.5 Exhaust Line Subsystem ......................................................................... 108
3.2.6 Optical and Photographic Diagnostic Subsystem .................................... 109
3.2.7 The Main Operations Windows ................................................................ 113
3.2.8 Method of System Operation ................................................................... 116
3.3 Experimental Apparatus – MSU Reactor C ....................................................... 117
3.3.1 Overview.................................................................................................. 117
3.3.2 Electromagnetic Focus ............................................................................ 118
3.3.3 Description of the Classic Reactor ........................................................... 120
3.4 Diamond Seeds for the Growth Process............................................................ 124
3.4.1 The Diamond Seed Cleaning Procedure – Wet Chemistry ...................... 125
3.5 The Pocket Holder ............................................................................................. 127
3.5.1 Pocket Holder Preparation....................................................................... 128
3.5.2 Pocket Substrate Holder Configuration ................................................... 128
3.6 System Startup .................................................................................................. 135
3.6.1 Introduction .............................................................................................. 135
3.6.2 The Diamond Seed Loading into the Microwave Cavity Reactor ............. 136
3.6.3 Discharge ignition .................................................................................... 137
3.6.4 Microwave Plasma Assisted CVD Processing ......................................... 138
3.6.5 System Shutdown .................................................................................... 140
3.6.6 Unloading the diamond substrate ............................................................ 141
3.7 Grown Diamond Substrates Measurements ...................................................... 141
3.7.1 Measurement of Substrate Temperature ................................................. 142
3.7.2 Linear and Lateral growth rate (V100 and V111) ......................................... 142
3.8 Main Analytical Instruments ............................................................................... 144
3.8.1 SCD Surface Morphology and Surface Area ........................................... 144
3.8.2 Scanning Electron Microscopy (SEM) ..................................................... 145
3.8.3 Secondary Ion Mass Spectrometry Analysis (SIMS) ............................... 146
3.8.4 Birefringence Imaging .............................................................................. 147
3.8.5 Etch Pits Analysis .................................................................................... 148
CHAPTER 4 ................................................................................................................ 151
MPACVD GROWTH OF SINGLE CRYSTALLINE DIAMOND SUBSTRATES WITH PCD
RIMLESS AND EXPANDING SURFACES ................................................................. 151
4.1 Introduction ........................................................................................................ 151
4.2 Background ....................................................................................................... 152
4.3 Experimental details .......................................................................................... 154
4.4 The growth process recipe ................................................................................ 155
4.5 Experimental results .......................................................................................... 157
4.6 The growth strategy ........................................................................................... 161
4.7 A comparison to earlier results .......................................................................... 163
4.8 Summary remarks ............................................................................................. 164

vii

CHAPTER 5 ................................................................................................................ 166
SINGLE CRYSTAL DIAMOND GROWTH AT CONSTANT SUBSTRATE
TEMPERATURE (Ts) .................................................................................................. 166
5.1 Exploring Constant Substrate Temperature and Constant High Pressure SCD Growth
Using Variable Pocket Holder Depths ......................................................................... 166
5.1.1 Research Highlights........................................................................................ 166
5.1.2. Introduction .................................................................................................... 167
5.1.3 Background .................................................................................................... 168
5.1.4 Experimental Details ....................................................................................... 171
5.2. The Single Crystal Diamond Growth Process ...................................................... 176
5.2.1 The SCD growth recipe: constant temperature deposition ............................. 176
5.2.2 Stopping the SCD growth process .................................................................. 179
5.2.4 Varying the input power and adjusting the reactor for the different pocket holder
depth recipes. .......................................................................................................... 183
5.3. Experimental Results ........................................................................................... 183
5.3.1 SCD Growth versus time ................................................................................ 183
5.3.2 SCD growth versus pocket holder depth ........................................................ 188
5.4 SCD Plate Characterizations ................................................................................. 196
5.4.1 Overview ......................................................................................................... 196
5.4.2 Birefringence imaging ..................................................................................... 197
5.4.3 Etch pits analysis ............................................................................................ 200
5.4.4 Secondary ion mass spectroscopy (SIMS) ..................................................... 204
5.5 SCD growth versus methane concentration .......................................................... 205
5.5.1 Growth Conditions .......................................................................................... 205
5.5.2 Linear and Lateral Growth Rates .................................................................... 206
5.5.3 Surface morphology........................................................................................ 208
5.6 Summary Remarks................................................................................................ 210
CHAPTER 6 ................................................................................................................ 213
SINGLE CRYSTAL DIAMOND GROWTH: (1) USING VARIABLE POCKET HOLDER
APERTURES AND (2) ADDING AN RE-GROWTH EXPERIMENTAL RUN ............... 213
6.1 Introduction ........................................................................................................... 213
6.2 Experimental Details: Variation of the Pocket Aperture ......................................... 215
6.3 Experimental Results: Variation of the Pocket Aperture ........................................ 216
6.3.1 SCD Growth versus time ................................................................................ 216
6.3.2 SCD growth versus growth time, pocket width, and pocket depth: the evolution of
the crystal versus time. ............................................................................................ 225
6.3.3 Summary Remarks: Variation of the Pocket Aperture .................................... 233
6.4 Experimental Details: SCD Re-growth and Expanding Surfaces .......................... 234
6.5 Experimental Results: SCD Re-growth and Expanding Surfaces ......................... 236
6.5.1 SCD Re-growth Criteria .................................................................................. 236
6.5.2 SCD Re-growth – Expanded surfaces ............................................................ 239
6.5.3 Summary Remarks: SCD Re-growth and Expanding Surfaces ...................... 243

viii

CHAPTER 7 ................................................................................................................ 244
PHOTOGRAPHIC EVALUATION OF THE SCD SURFACE GROWTH VERSUS TIME
.................................................................................................................................... 244
7.1 Introduction ........................................................................................................... 244
7.2 Photographic Images Processing and Video Creation .......................................... 245
7.3 Data Acquisition .................................................................................................... 248
7.4 Preliminary Results ............................................................................................... 250
7.5 Summary Remarks................................................................................................ 255
CHAPTER 8 ................................................................................................................ 256
SUMMARY AND FUTURE WORK .............................................................................. 256
8.1 General Summary ................................................................................................. 256
8.2 Accomplishments .................................................................................................. 258
8.3 Publications Resulting from this Dissertation Research ........................................ 260
8.4 Future Work .......................................................................................................... 261
REFERENCES ............................................................................................................ 262

ix

LIST OF TABLES

Table 2. 1 Principles properties of diamond [21]. ......................................................... 17
Table 2. 2 Growth rates ranges in MPACVD diamond growth [49]–[53]. ...................... 38
Table 2. 3 Carnegie Institution of Washington diamond growth parameter reported ..... 56
Table 2. 4 Growth parameters for SCD by Diamond Research Center, Japan ............. 58
Table 2. 5 LIMHP/CNRS, Paris, France growth parameter reported ............................. 71
Table 3. 1 Theoretical TM013 and TM012 modes ........................................................... 120
Table 3. 2 Measured dimensions on HPHT seed ........................................................ 125
Table 5. 1 Growth parameters for SCD growth using the d 2 = 2.3 mm and d3 = 2.6 mm
pocket holder depths.............................................................................. 184
Table 5. 2 Summary of different growth versus depth experiments for eleven samples at
a substrate temperature of 1020 ± 5°C. ................................................. 189
Table 6.1 Growth parameters for SCD growth using the w2 = 6.6 mm and w3 = 7.0 mm
pocket holder apertures at a constant pocket depth of d = 2.6 mm ....... 217
Table 6.2 Growth parameters for SCD growth using w3 = 7.0 mm pocket holder aperture
at pocket depth of d = 2.9 mm ............................................................... 222
Table 7.1 Measured and calculated values of the SCD areas and SCD lateral growth rate
for ACH64, ACH55, and ACH78. An instrument manually collected the
measured values, and the software collected the calculated. ................ 251
Table 7. 2 Calculated values of the PCD inward growth on the pocket wall for experiments:
ACH64, ACH55, and ACH78. ................................................................ 252

x

LIST OF FIGURES

Figure 2. 1 FCC structure of the diamond crystal [21]. ................................................. 15
Figure 2. 2 Pressure-temperature phase diagram for carbon [31]. ................................ 20
Figure 2. 3 Schematic of the physical and chemical processes occurring during the CVD
growth of diamond [34], [38]. ................................................................... 24
Figure 2. 4 A schematic of the reaction process occurring at the diamond surface leading
to the stepwise addition of CH3 species and diamond growth [33]. ......... 26
Figure 2.5 Process map in [CH3]–[H] space showing the operating ranges of the main
CVD diamond growth processes (from [9]). The evolution of the operating
point of the LIMHP reactor as a function of working pressure (from 50 to
300 mbar) and methane concentration. ................................................... 34
Figure 2. 6 Map of surface morphologies as a function of misorientation angle from the
(001) and methane/hydrogen ratio. Three types of morphologies appear.
Atomically flat diamond is preferred, which only grows in the region (III) [43].
................................................................................................................. 36
Figure 2.7 Optical microscope images of the epitaxial CVD diamond films grown at
1000°C (a) Sample overview (b) Magnification 20x [49]. (c) Optical
micrograph of a 100 µm thick diamond film grown on an on-axis (001)
oriented Ib substrate [57]. (d) and (e) CVD diamond film is grown at high
temperature (1000°C and 920°C) exhibiting growth pyramids [56], [58]. . 39
Figure 2.8 Mechanism of two-dimensional nucleation on a wide terrace, beginning of
hillock growth [59]. ................................................................................... 40
Figure 2. 9 Schematic view of step flow growth showing the diffusion of growth species
to steps and subsequent flow of steps to the right [59]. ........................... 41
Figure 2. 10 Optical micrographs of the grown surfaces of CVD diamond with (a), (b) no
intentional N2 addition and (c) 2 ppm of N2 and (d) 6 ppm of N2 [58], [63].
................................................................................................................. 42
Figure 2.11 Diamond growth rates as a function of substrate temperature at 4% methane
concentration and a 3.8 kW microwave power [54]. Here Cme is the methane
concentration. .......................................................................................... 43

xi

Figure 2. 12 The influence of the methane concentration in the diamond growth rate [58]
................................................................................................................. 45
Figure 2.13 Optical microscope images of the diamond films grown at 850°C with
methane concentration of (a) 2% (b) 4% (c) 6% and (d) 8% [49]. ............ 47
Figure 2.14 Evolution of the growth rate of the CVD diamond films as a function of the
injected MW power density in the discharge using 4% of CH4 (square dots)
or 7% of CH4 (triangular dots) [75]. ......................................................... 50
Figure 2.15 Optical microscope images of the surface morphology of epitaxially grown
diamond films deposited on (001) oriented diamond substrates without (a,
b) and with (c, d) nitrogen added to the gas phase [66]. .......................... 52
Figure 2.16 Carnegie Institution’s Geophysical Laboratory diamond is grown at high
growth rates with and without nitrogen [14], [87]. (a) SCD plate, (b) 0.2 – 2.2
carat brilliantly cut and polished nitrogen doped light brown colored and
near colorless SCD substrate, (c) examples of as-grown SCD substrates
after annealing treatment and (d) in image (A) 2.4 carat single-crystal CVD
diamond (B, C, D) 2.3 carat diamond after cut and polished [88]. ........... 55
Figure 2.17 A 4.65 carat, 10 mm thick SCD substrate grown with 24 deposition runs [10]
................................................................................................................. 58
Figure 2.18 Large CVD diamonds grown by multi-steps high growth rate synthesis
process. (a) 10 mm thick, 4.7 carat CVD diamond, (b) 9.6 mm thick, 3.5
carat CVD diamond, (c) 8.7 mm thick, 4.4 carat CVD diamond [11]. ...... 59
Figure 2.19 Schematics of newly designed substrate holder together with corresponding
results of simulation of plasma, pictures of hydrogen plasma and observed
diamond growth rates [102]. .................................................................... 61
Figure 2.20 Schematic illustration of ‘‘enclosed type’’ and ‘‘open type’’ holders. The final
depth of diamond after the growth “d” is indicated [10]. ........................... 62
Figure 2.21 Optical microscope images of grown diamond after 1st growth (left) and
repetition of growth (right) for open and enclosed type holder. Thicknesses
of grown diamonds (t) are indicated in the pictures. A dashed circle indicates
a crack formed by repetition of high rate growth [10]. .............................. 63
Figure 2.22 Lift-off process with ion implantation [108]. ................................................ 64

xii

Figure 2.23 Photograph of freestanding single-crystal CVD diamond plates produced by
a lift- off process using ion implantation [94]. ........................................... 65
Figure 2.24 Schematic illustration of a crystal enlarging process by a combination of liftoff process and side-surface growth [94]. ................................................ 66
Figure 2. 25 (a) Photograph of a half-inch seed plate obtained by cutting and polishing
the diamond in Fig. 5. (b) Photograph of a half-inch nitrogen-doped singlecrystal CVD diamond plate produced under high growth rate (32 µm/h) from
the seed plate by a lift-off process using ion implantation. (c) Polarized light
microscope image of the half-inch single-crystal CVD diamond plate [94].
................................................................................................................. 68
Figure 2. 26 Method steps to produce clones and tiled clones [109]. ............................ 69
Figure 2. 27 An image of typical tiled clones [109]. ....................................................... 70
Figure 2.28 Diamond growth rate as a function of nitrogen addition in the gas phase
(MWPD = 95 W/cm3). Optical and PL images obtained under UV light for an
undoped free-standing CVD diamond crystal and a 2 ppm N2 doped [56].
................................................................................................................. 73
Figure 2.29 Optical images of the samples grown with (a) no intentional N2 addition, (b) 2
ppm of N2 (c) 6 ppm of N2 and (d) 10 ppm of N2 [63]. .............................. 74
Figure 2. 30 Effect of growth temperature. Confocal laser microscopy images of 160 µm
thick CVD films grown at (a) 920 ◦C, 7 µm/h, (b) 870 ◦C, 6 µm/h doped [56].
................................................................................................................. 75
Figure 2.31 Optical image of the free-standing CVD diamond samples obtained after laser
cutting and polishing and grown with (1) no intentional N2 addition, (2) 2
ppm of N2, (3) 4 ppm of N2 in the gas phase [63]..................................... 76
Figure 2.32 Optical images of the diamond films grown at 95 W/cm 3 and 850 °C with (a)
2% CH4, (b) 4% CH4, (c) 6% CH4 and (d) 7.2% CH4 [58]. ........................ 77
Figure 2. 33 From left to right, spatial distributions of microwave power density (MWPD),
electron density (Ne), electron temperature (Te), gas temperature (Tg) and
atomic hydrogen density ([H]) for a pure hydrogen plasma ignited inside the
LIMHP reactor [9]. .................................................................................... 78

xiii

Figure 2. 34 Schematics of the initial diamond substrate before polishing (a) and of
substrates polished into a pyramidal-shape (b) type A, 20° {100}misoriented, (c) type B, 20° {110}-misoriented [114]. ............................... 79
Figure 2. 35. Optical images and 3D representation of the sample grown onto 20° {100}misoriented pyramidal-shape substrates after several growth interruptions.
The total thickness of the CVD layer is (a) 90 μm, (b) 270 μm, (c) 500 μm
[114]. ........................................................................................................ 80
Figure 2.36 Schematic of the complete procedure of self-assembled metallic masks [121]
................................................................................................................. 81
Figure 2. 37 Schematic representation of the experiments (a) Test sample, (b) Reference
sample. After each etching step and the resulting final morphology is shown
in a (c) test sample, (b) reference sample [121]. ...................................... 82
Figure 2. 38 A thick CVD diamond layer grown on an HPHT diamond substrate hollowed
out by a large hole at the (b) beginning, (c) before, and (d) after merging of
the growth fronts and disappearance of the hole [122]. ........................... 83
Figure 2. 39 Laser microscope images of the CVD film grown on an HPHT substrate with
a macro hole after plasma etching to reveal dislocations: (a) full-size image
showing the underlying square substrate. (b), (c) zoomed into the regions
above the substrate and above the hole, respectively, (d) PL UV image of
the freestanding CVD diamond plate. (e) Birefringence image of the
freestanding plate under cross-polarizers. The white dashed square
indicate the position of the edges of the removed substrate [122]. .......... 84
Figure 2. 40 The schematic diagrams of the open type substrate holders and enclosed
type substrate holders with different recess depths. (The yellow squares
represent the seed crystals placed on the substrate holder or in the recess)
[123]. ........................................................................................................ 86
Figure 2. 41 (a) HPHT seed grown by using an open type substrate holder (d=0mm); (b)
HPHT seed grown by using an enclosed type substrate holder (d=2mm)
[123]. ........................................................................................................ 87
Figure 2. 42 (a) The influence of the recess depth on the main surface growth rate. (b)The
influence of the recess depth on the lateral surface and main surface growth
rate ratio. Insets: the cross-section schematic diagrams of the samples
grown with different recess depths [123].................................................. 88
Figure 2. 43 The influence of the recess depth on the FWHM of the diamond layers at the
center part and the periphery [123]. ......................................................... 89
xiv

Figure 2. 44 Comparison of reactor A, B and C (a) absorbed power density versus
pressure and (b) growth rate versus pressure [18], [132]. ....................... 91
Figure 2. 45 A cross-sectional view of SCD synthesis versus time in an open holder (a,
b, c). d) The synthesized diamond after 35.5h of deposition time [20]. .... 92
Figure 2. 46 Cross-sectional views of SCD growth process versus time in a pocket holder
[20]. .......................................................................................................... 93
Figure 2. 47 Optical microscopy images of as-grown SCD substrates with (a) reduced
PCD rim and (b) without a PCD rim [124]. ............................................... 94
Figure 2. 48 Birefringence images of a 221 μm thick CVD SCD plate with exposure times
of (a) 500ms (b) 2000ms and (c) 5000ms [124]. ...................................... 95
Figure 2. 49 Photographs of the discharge above a 3.5 mm x 3.5 mm SCD substrate in a
1.5-inch molybdenum holder as the pressure is increased from 120 to 400
Torr. The absorbed power is kept constant at 2100 W [12]. .................... 96
Figure 2. 50 Discharge power density and volume as a function of the pressure for the
discharges utilizing Pabs= 2100 W and 3% CH4 as shown in Figure 2.49.
The differentiation between the previously investigated moderate pressure
regime and the high pressure regime is indicated by the green dashed line
[12]. .......................................................................................................... 97
Figure 2. 51 Operational field map for Reactor B using a SCD substrate over the entire
pressure regime ....................................................................................... 98
Figure 2.52 Linear growth rate and weight gain as a function of the process pressure and
................................................................................................................. 99
Figure 2.53 Top surface of SCD films grown in the pressure range between 240 and
400Torr [12]. .......................................................................................... 100
Figure 3. 1 Overall setup of MSU microwave plasma cavity Reactor C system [85]. .. 102
Figure 3. 2 Components of standard microwave power supply subsystem ................ 103
Figure 3. 3 Cooling water supply ................................................................................. 108
Figure 3. 4 Optical and Photographic subsystem on microwave cavity reactor C. (1)
Pyrometer, (2) CANON camera. ............................................................ 110

xv

Figure 3. 5 Camera Arduino control map for microwave cavity reactor C [134]. ......... 112
Figure 3. 6 (a) Diagram of the Arduino circuit and (b) photograph of the completed
operational Arduino circuit [135]. ........................................................... 113
Figure 3. 7 The standard view panel operation window .............................................. 114
Figure 3. 8 The schematic view panel operation window ............................................ 115
Figure 3. 9 Electromagnetic fields, TM01n modes (n = 2 left, n = 3 right) [137]. ......... 118
Figure 3. 10 Three typical discharges (in green) that result from exciting the cylindrical
cavity applicator with the TM013 mode [137]. The red lines are the magnetic
field patterns and the blue lines are the electric field patterns. .............. 119
Figure 3. 11 Cross-sectional view of the MCPR [139]. ................................................ 122
Figure 3. 12 Optical microscope image of the side view of the 3.5 mm x 3.5 mm x 1.4 mm
HPHT seed. ........................................................................................... 125
Figure 3. 13 Isometric drawing view of a 64.7 mm substrate holder. .......................... 129
Figure 3. 14 (a) A schematic side view of the microwave cavity plasma Reactor C with a
(b) zoom of the substrate holder setup. ................................................. 130
Figure 3. 15 Side cross sectional view of conventional pocket holder (unit: mm) that was
used in the experiments in Chapter 4. ................................................... 132
Figure 3. 16 Drawings of the pocket holder for SCD growth at variable pocket depths (unit:
inches) that was used in the experiments in Chapter 5. ........................ 133
Figure 3. 17 Side and top view of the pocket holder with a varied width (unit: mm) .... 134
Figure 3. 18 Side view of the pocket holder used for SCD re-growth process (unit: mm)
that are described in Chapter 6.............................................................. 135
Figure 3. 20 Plasma in an ideal contact with the substrate ......................................... 138
Figure 3. 21 Operating field maps and matching, i.e., Pref/Pinc, versus absorbed power for
three constant pressure conditions: 120, 180, and 240 Torr [141]. ........ 140

xvi

Figure 3. 22 Optical microscope image of the side view of the as-grown diamond
substrate. The red arrows show the lateral and linear growth direction. The
dashed red lines identified the dimensions of l1 and l2. .......................... 143
Figure 3. 23 Optical microscope image of the top view of the as-grown with a white circle
drawn around the diamond shape. ........................................................ 145
Figure 3. 24 Schematic of SIMS analysis system [143]............................................... 147
Figure 3. 25 (a) optical image of ACH09 plate and (b) image enhanced ..................... 149
Figure 3. 26 General view of the images processing tool. ........................................... 150
Figure 4. 1 Thick PCD rim growth on SCD substrate using a conventional open holder
growth method. ...................................................................................... 153
Figure 4. 2 Growth process recipe displaying a process cycle variation of T s and Pinc
versus growth time for substrate ACH23 (shown in Figure 4.3). ............ 156
Figure 4. 3 Optical microscope images of (a) ACH23. Close up views of the substrate in
(b) – (d) show rimless growth with larger SCD area. ............................. 158
Figure 4. 4 Optical microscope images of (a) ACH24. Close up views of the substrate in
(b) – (d) show rimless growth with larger SCD area. ............................. 159
Figure 4. 5 Side view images of (a) ACH23 and (b) ACH24 indicate outward growth of the
CVD SCD substrates. ............................................................................ 160
Figure 4. 6 SEM images of (a) ACH23 and (b) ACH24 indicate PCD rim free CVD SCD
substrates. ............................................................................................. 160
Figure 4. 7 Outward growth of the CVD SCD substrates [20]. .................................... 164
Figure 5. 1 Conventional pocket holder design geometry for SCD growth (top side). The
pocket substrate holder placed on the reactor cooling stage (bottom side)
............................................................................................................... 173
Figure 5. 2 Reactor operating field map ...................................................................... 177
Figure 5. 3 Diamond growth recipe versus time for pocket depth of (a) d 2 = 2.3 mm and
(b) d3 = 2.6 mm ...................................................................................... 178

xvii

Figure 5. 4 (a) As-grown uncleaned SCD with the top surface is still entirely within the 2.6
mm deep pocket; (b) cross-sectional schematics when the grown diamond
(in blue) is still within the pocket; (c) optical microscope images of side and
top views of as-grown and cleaned diamond substrate (ACH02). Note that
the bottom yellow region is the original HPHT diamond seed. The green
square represents the the original seed boundary. ................................ 180
Figure 5. 5 The ACH22-b SCD which has grown out the pocket; (b) the associated crosssectional schematic of the grown SCD when it has grown out of pocket; (c)
optical microscope images of side and top view of a grown diamond
substrate (ACH22-b). ............................................................................. 182
Figure 5. 6 Optical microscope top surface images of substrates samples listed in Table
1. For d2 = 2.3 mm (a) 10 hours, (b) 24 hours, and (c) 48 hours. For d3 = 2.6
mm (d) 10 hours, (e) 24 hours, (f) 48 hours and (g) 60 hours. ............... 186
Figure 5. 7 Optical microscope images of the typical samples shown in Table 2 that were
grown in pocket depths of (a) 2.0 mm, (b) 2.3 mm, (c) 2.6 mm and (d) 2.9
mm. ........................................................................................................ 190
Figure 5. 8 Morphology shapes and the growth rate versus pocket holder depth for
experimental data for Table 5.1 and Table 5.2. The square surface is
associated with pocket depths of 2.0 and 2.3 mm. The circular surface is
associated with long runs of the pocket depth of 2.6 mm, and the octagon
shape surface is associated with the shorter runs that used the pocket
depths of 2.6 and 2.9 mm. ..................................................................... 193
Figure 5. 9 The ratio between lateral growth rate, V110, and vertically growth rate, V100,
versus pocket depth. Insets are SEM images of the corners of the SCD
substrates. ............................................................................................. 194
Figure 5. 10 Plot summarizing the normalized lateral CVD area gain versus vertical
thickness gain for different pocket depths for the experimental data listed in
Tables 5.1 and 5.2. ................................................................................ 195
Figure 5. 11 CVD plates after laser cutting and mechanical polishing. The measured
thickness gain for each sample was: 274 μm for ACH22, 255 μm for ACH20
and 600 μm for ACH33. ......................................................................... 197
Figure 5. 12 Examples of two as-grown SCD with their corresponding birefringence
images. (a) ACH20 and (b) ACH33. ...................................................... 199
Figure 5. 13 (a) Etched surface of the polished CVD plate ACH33 and (b) Etched surface
of a unpolished CVD plate ACH20. ........................................................ 201
xviii

Figure 5. 14 SEM images of the pits in three different areas on CVD plate ACH09 (a) one
side at the edge, (b) somewhere at the center (c) somewhere on the center
of the top surface. .................................................................................. 203
Figure 5. 15 Comparison of the etch pit density with recent results. ........................... 204
Figure 5. 16 Nitrogen concentration in (a) SCD CVD plate ACH33 and (b) SCD CVD plate
ACH20. The red line indicates the detection limit of the instrument. ...... 205
Figure 5.17 Growth rates of CVD diamond grown as a function of the methane
concentration under growth conditions of 3.5 kW - 2.4 kW, 1020 °C and 240
Torr. ....................................................................................................... 207
Figure 5.18 Optical microscope images of the top surface of grown SCD substrates for
methane concentrations of (a) 4%, (b) 5%, (c) 6% and (d) 7%. ............. 209
Figure 5. 19 Optical high magnification images of substrate grown at (a) 4 % methane
and (b) 5 % methane. ............................................................................ 210
Figure 6.1 Optical microscope top surface images of substrates listed in Table 6.1, for w2
= 6.6 mm (a) 12 hours, (b) 16 hours, (c) 30 hours, (d) 30 hours, (e) 48 hours,
(f) 50 hours, and (g) 55 hours. Note that ACH55 has smooth edges and a
decreasing top surface........................................................................... 218
Figure 6.2 Optical microscope top surface images of substrates listed in Table 6.1, for w3
= 7.0 mm (a) ACH79 after 12 hours of growth, (b) ACH71 after 30 hours,
and (c) ACH78 after 45 hours of growth. ............................................... 221
Figure 6. 3 Optical microscope top surface and side view of images of substrates listed
in Table 6.2, for w3 = 7.0 mm a constant pocket depth of d= 2.9 mm (a) 45
hours, (b) 40 hours and (c) 50 hours. Note: that the substrate ACH72 have
a dark spot on the left bottom side it is a crack on the top surface that was
produced at the experimental run when the substrate was removed from the
pocket. ................................................................................................... 223
Figure 6.4 The total average lateral growth rate versus pocket aperture at 12 and 45 hours
of growth time ........................................................................................ 224
Figure 6.5 Normalized area gain versus growth time at different pocket configurations
from experimental data provided in Table 6.1, Table 6.2, and selected
results of Chapter 5 grown in pocket depth of 2.6 mm. .......................... 226

xix

Figure 6.6 Plot summarizing the normalized lateral CVD area gain versus growth time for
different pocket apertures and depths for the experimental data listed in
Tables 6.1, Table 6.2, and selected data from Table 5.2. ...................... 227
Figure 6. 7 Photographic images in the frame time of 0, 12, 24, 40, 50 and 55 hours. The
opaque yellow polygon layer represents the HPHT seed inside the pocket.
The orange polygon layer identifies and outlines the grow SCD during the
indicated time frames. ............................................................................ 229
Figure 6. 8 The ratio between lateral growth rate, V110, and vertically growth rate, V100,
versus growth time for three different pocket holder apertures. ............. 230
Figure 6.9 Plot summarizing the normalized lateral CVD area gain versus vertical
thickness gain for different pocket holder geometries for the experimental
selected data from Table 6.1 and Table 6.2 and the results of Chapter 5.
Short run data and long run data from this section are indicated by the grown
and res data points respectively. ........................................................... 232
Figure 6.10 Schematic of the re-growth process. After the 1st growth step process, the
grown substrate is placed in an optimized substrate holder................... 236
Figure 6.11 Optical microscope images of (a) top view of ACH26 after the 1st growth step,
(b) top view of ACH26-2 after the 2nd growth step, and (c) side view of the
ACH26-2. ............................................................................................... 238
Figure 6. 12 Optical microscope side view images of ACH36 after the (a) 1st growth step
and (b) 2nd growth step. ........................................................................ 239
Figure 6. 13 Optical microscope side view of (a) and (b) substrate ACH23 after the 1 st
and 2nd growth step, (a) and (b) substrate ACH27 after the 1st and 2nd growth
step, and (e) top view and (f) side view of substrate ACH34 after the 1st and
2nd growth step....................................................................................... 241
Figure 6. 14 Normalized lateral area gain versus vertical thickness gain .................... 242
Figure 7.1 Experimental setup for photographs and video recording for SCD lateral
growth analysis ...................................................................................... 247
Figure 7.2 Schematic of the data acquisition process. The photographed experiment
showed in the flow diagram is ACH78. A total of 1277 pictures were
captured. ................................................................................................ 249

xx

Figure 7.3 The outward SCD growth and the inward PCD growth versus growth time for
substrate ACH64. .................................................................................. 253
Figure 7.4 The outward SCD growth and the inward PCD growth versus growth time for
(a) substrate ACH64 and (b) substrate ACH78. ..................................... 254

xxi

CHAPTER 1

INTRODUCTION

1.1 Motivation
Over centuries, diamond has been used as a precious gemstone and as an ultra-hard
industrial material. Diamond unites a rich collection of distinctive physical and chemical
properties, which makes it a valuable and strategic material for many engineering
applications. The exceptional compact structure of diamond arises from two basic
concepts: (i) the carbon atoms are small and light, with short range bonds in the diamond
structure [1], and (ii) the carbon atoms form very strong covalent bonds [2]. It has long
been acknowledged that diamond is an excellent material with other outstanding
properties, such as optical, mechanical, thermal, electrochemical, chemical, and
electronic properties [1], [3], [4], [5] that also outclass competing materials. When these
properties are combined, they offer an engineering material with a tremendous potential
for use in a wide range of applications in optics, semiconductors, microelectronics,
biomedical and manufacturing engineering. However, despite its extraordinary potential
for a wide variety of scientific and industrial applications, the high cost and the low volume
of diamond substrate manufacturing limits the diamond material market to just a few highend applications. Currently, the researchers are attempting to overcome the low volume
1

and small physical size of the CVD diamond by developing technologies that grow highquality diamond at low manufacturing cost and larger sizes.
In order to overcome the availability and crystal quality issues, scientists have
developed different technologies for diamond growth in the laboratory. Currently, the two
major approaches that can growth diamonds are the high-pressure high-temperature
(HPHT) process and the microwave plasma assisted chemical vapor deposition
(MPACVD) process.
The HPHT technique to grow diamond from graphite to a single crystal seed that is
suitable for jewelry was first to develop during the last several decades [6], [7]. The
diamond crystals produced with the HPHT technique can have high purities but in the
past, have been limited in size and cost. However, now HPHT crystals are grown up to 3
carats and are available commercially. The high cost is due to the large size of the
manufacturing equipment, namely the large hydraulic presses that are required to grow
these diamonds. To increase the sizes of the HPHT diamonds, the size of the hydraulic
presses become quite large, resulting in an expensive cost. After being polished and
being faceted, HPHT diamonds are rarely larger than 1.5 carats. In the past diamond
crystal grown by HPHT, seed technique has not been suitable for electronic and optical
applications. However, in the future, the availability of large and low-cost single crystal
HPHT diamonds would not only impact the gem fine jewelry market but also would open
up many more research and industry applications.
Recently, the MPACVD process has been identified as a very promising method for
commercially growing high quality diamond [8]. In this process, the microwave power is
coupled into the reactor via a coupling waveguide located at the top of a plasma cavity

2

reactor. A vacuum pump system is attached to the inside of the cavity applicator and
maintains a constant working pressure while the input gases are flow into the bell jar
located inside the cavity applicator. Compared with other diamond growth methods, the
MPACVD process provides controlled and repeatable growth conditions versus time and
the results can be reproduced at a reasonable cost. During the current decade, many
scientists and engineers have developed standard MPACVD single crystal diamond
(SCD) growing processes which overcome some of the constraints of the HPHT process.
The CVD technology can provide diamond substrates with much larger sizes, higher
quality and with high growth rates, i.e. 50 µm – 100 µm [9]–[11] [12] at high process
pressures. This indicates that the MPACVD diamond growing method has the potential
for becoming low-cost and high-quality diamond growth process.
The diamond growth by MPACVD has been investigated over a wide window of
growth parameters, such as pressures, nitrogen content, methane concentration and
reactor configurations. Most notably C.S. Yan et al. [13] in 2002, reported that MPACVD
SCD crystals can be produced at growth rates up to 150 µm/h at pressures of 200 Torr
by adding a small amount of nitrogen into the gas fed (1 – 5 % N2/CH4). A few years later
the followed up their previous results by demonstrating a ten carats of CVD grown SCD
1 cm in thickness at growth rates of up to 100 μm/h at pressures up to 300 Torr with the
addition of nitrogen concentrations (10 % N2/CH4) [14].
Most recently, In 2017 E.V.Bushuev. et al. [15] reported the growth of CVD diamond
films at a constant pressure of 130 Torr and variable methane concentrations of 1%, 4%,
7%, 10%, and 13%. They found that growth rates strongly depended on the methane
concentration and substrate temperature. The low growth rate (0.5 µm/h) was obtained

3

at 1% CH4/H2 at substrate temperature (Ts) = 800 °C. The higher growth rate (82 µm/h)
was achieved at a high methane concentration of 13% at T s = 965 °C ± 15 °C. These
results were consistent with the results measured earlier at Michigan State University by
Lu et al. in 2013 [16].
Silva et al. [9] developed a numerical, theoretical plasma model that was used to
calculate the species concentrations of a typical cylindrical microwave plasma reactor.
This model calculated the species distribution of the different gas species in plasma
discharge for diamond growth as a function the growth parameters, such as pressures,
methane concentration, and substrate temperature. They concluded that at high
temperatures, the desorption of CH3 and C2H2 species becomes more rapid, and
competes with the carbon incorporation reactions. At high pressures, there is a significant
increase in atomic hydrogen density and the CH3 radical density. As the operating
pressure is increased both the hydrogen species densities, and the CH3 radical densities
increased, but the hydrogen density increased versus pressure 10 – 1000 times faster
than the CH3 concentration. Therefore, both the growth rate and the diamond quality
increases [17].
This thesis investigation aims at encouraging the growth of high-quality single crystals:
(1) at a very high growth rates and in an almost nitrogen-free growth process, (2) growth
with an enlarged diamond top surface, and (3) SCD growth with a significantly reduced
polycrystalline diamond (PCD) rim. The diamond growth for this thesis has been
conducted with an MSU-design microwave plasma cavity reactor (MPCR) which is
referred here to as Rector C. This reactor was initially designed and experimentally
operated under the efficient and safe conditions as described by Y. Gu et al. [18].

4

This thesis investigated the growth of diamond versus various experimental variables
under constant pressure condition of 240 Torr. The external input conditions versus time
included (1) variable substrate temperatures, (2) variable input powers, (3) constant
substrate temperatures, and (4) variable methane concentrations. Another variable that
was investigated was the size and the shape of the substrate pocket holder. The pocket
holder was introduced by Asmussen et al. [19] and later investigated by Nad et al. [20].
One goal of this thesis investigation was to explore the influence of the pocket holder
geometry on the SCD growth process. The variable geometry of the pocket included: (1)
several pocket holder depths, (2) several pocket holder apertures, (3) specially designed
pocket holders that were used in large diamond seed growth experiments, and (4)
specially designed pocket holders that were used in diamond re-growth substrates
experiments.
Thus, this thesis investigates a broad range of growth parameters in order to achieve
the rapid growth of single crystal diamond plates and cubes. The quality of the grown
diamond materials was enhanced by eliminating the polycrystalline diamond that typically
surrounds the diamond top surface. The top surfaces of the grown diamond were larger
when compared to the grown diamond in previous MSU experimental thesis results. This
research also includes some characterization techniques that allows the determination of
quality of the grown substrates, such as, birefringence imaging, SIMS analysis, and SEM
images. These characterization techniques are employed to evaluate the grown diamond
material

5

1.2 Research Objectives
In this thesis research, the deposition of high quality, PCD rimless single crystal
diamond SCD via MPACVD is investigated. The overall objective of this investigation is
to develop experimental methods that enable the rapid growth of high quality, large
volume, SCD using the pocket substrate holder configuration.
Ultimately it is desired to achieve this diamond growth in one controlled long process
step. For this investigation, a series of experiments were performed that explore the
growth of SCD over a wide range of input experimental variables. The specific
experiments are described below and are divided into six tasks. Each task is described
below as a separate set of experiments.

i)

Task one

This set of experiments used a pocket holder that was held fixed and then the
substrate temperature and microwave input power were varied as a function of the time.
i.e. the experiments varied the process recipe, Pinc and Ts versus time. This task
demonstrated that the application of a specific process recipe, i.e. the real-time control of
the input microwave power and hence Ts versus time, over the process cycle, enabled
the growth of thicker and larger area SCD while it also eliminates the PCD rim.
Specific objective:
SCD growth is investigated with the following variables:
a) Constant pocket holder depth (d) and width (w) of d = 2.3 mm and w = 6.0 mm,
respectively.

6

b) Microwave input power was varied during the process cycle between of 2.9 kW
– 1.9 kW
c) Constant pressure of 240 Torr
d) Constant methane concentration 5%
e) The substrate temperature was varied as a function of the time over the range
of 985°C – 1040°C
The detailed results of these experiments are covered in Chapter 4 of this thesis.

i)

Task two

This task explored the SCD growth in a pocket holder by varying pocket depths,
methane concentrations, and the process cycle. It was desired to grow thick, high-quality
SCD without a PCD rim. Therefore, this experimental activity identified the experimental
conditions that produced SCD growth without a PCD rim. The experiments were divided
up into two sets: (1) when only the pocket holder depth was varied and (2) when only the
methane concentration was varied. The experimental results identify pocket holder
geometries that yielded PCD rimless growth and associated expanded top SCD surfaces.
Specific objective
This set of the experiments were made under a constant substrate temperature and
a constant pressure condition. The microwave input power was varied versus time in
order to hold the substrate temperature (T s) constant. In particular, these experiments
explored pocket holder growth under the following experimental conditions:
a) Constant substrate temperature (Ts) of 1020°C ± 5°C
b) Constant pressure of 240 Torr

7

c) Constant pocket holder aperture of 6.0 mm for four different pocket holder
depths, i.e. 2.0 mm, 2.3 mm, 2.6 mm and 2.9 mm
d) Variable process times between 12 hours and 60 hours
e) Variable microwave input power between 2.0 kW and 3.5 kW.
f) Variable methane concentration between 4% - 7%
The results of these experiments are covered in Chapter 5 of this thesis.

ii)

Task three

These set of experiments explored the variation of the geometry of the pocket holder
to promote the outward growth of the diamond surfaces. Based on the results of task one
and two, two pocket holder depths were used: 2.6 mm and 2.9 mm and four different
pocket aperture values were varied: from 6.0 mm, 6.6 mm, 7.0 mm, and 7.6 mm
Specific objective:
The objective was to improve the diamond outward growth of the diamond substrates
by enlarging the top surface area. Diamond growth was investigated under the following
experimental conditions:
a) Constant pressure of 240 Torr
b) Constant substrate temperature (Ts) of 1020°C ± 5°C
c) Constant methane concentration (5%)
d) Variable process time. i.e. 15 hours – 55 hours
e) Two different pocket depths, i.e. 2.6 mm and 2.9 mm with the following pocket
aperture widths of 6.0 mm, 6.6 mm, 7.0 mm, and 7.6 mm
f) Variable microwave input power between 2.0 kW and 3.0 kW

8

The results of these experiments are described in Chapter 6 of this thesis.

iii)

Task four

Continue the growth process by adding one additional growth step in order to grow
thicker and larger area SCD plates.
Specific objective:
Based on the enlarged areas obtained from task one, two, and three the primary
objective of this task is to add one growth step and continue expanding the thickness and
the top surface area SCD. Re-growth was performed with the following variables:
a) Specially design pocket holder configuration, i.e. d = 3.6 mm, w = 6.2mm
b) Microwave input power of 3.0 kW – 2.0 kW
c) Constant pressure of 240 Torr
d) Constant methane concentration of 5%
e) Variable process time of 45 hours
f) Constant substrate temperature of 1020°C ± 5°C
The results of these experiments are covered in Chapter 7 of this thesis.

iv)

Task five

This set of experiments explored the SCD growth when the diamond seed area is
increased. The diamond growth process was carried out on large HPHT diamond seeds
of 5.0 mm x 5.0 mm and 7.0 mm x 7.0 mm. Also, a few experiments using 7.0 mm x 7.0
mm microwave plasma CVD plates were performed.
Specific objective:

9

To reproduce the results obtained in previous tasks by using large HPHT seed and
expand the final grown diamond area. The SCD growth was performed with the following
variables:
a) Specially design pocket holder configuration, i.e. d = 2.6 mm, w = 7.7 mm
b) Constant pressure of 240 Torr
c) Variable process time between 12 and 45 hours.
d) Microwave input power of 3.0 kW – 2.0 kW
e) Constant substrate temperature of 1000°C ± 5°C
The results of these experiments are covered in Chapter 8 of this thesis.

v)

Task six

This tasks photographically monitors the SCD growth as a function of the time. This
set of experiments reveals the evolution of the top SCD surfaces versus time.
Photographs were taken during every minute throughout the process cycle in order to
analyze the diamond surface evolution during the growth. The results of these
experiments are covered in Chapter 9.

10

1.3 Thesis outline
This thesis is composed of 10 chapters. The thesis content is summarized in the
paragraphs below.
Chapter 1 presents a short introduction and overview of the research project. It
includes research motivations and research objectives. The major thesis objective is
divided into six tasks where each task is described as a separate group of experiments.
Chapter 2 presents a literature review, including the structure and properties of the
diamond material and a historical overview of the CVD diamond growth process. The
diamond growth processes, i.e. HPHT diamond and CVD diamond growth are discussed.
The physical and the chemical aspects of the CVD diamond were also mentioned here.
Although the mechanism for diamond growth by CVD is a complex process, this chapter
also describes in detail important factors that have to be considered in order to grow
diamond by the CVD process. The growth of possible defects may also occur during the
CVD growth process. Thus, this chapter presents a brief review of defects in the growing
diamond is presented. The effect of the experimental growth variables, such as methane
concentration, substrate temperature and the growth versus time aspects are also
described. Lastly, a review of SCD growth via MPACVD using a pocket holder is
described.
Chapter 3 presents some of the diamond growth experimental procedure and provides
detailed description of the MSU Reactor C configuration that was used in this thesis
investigation. Here an extensive description of the experimental apparatus is shown. The
appearance, properties, and cleaning procedures of the diamond seeds are presented.
Also, a description of the pocket holder configuration that was used in this thesis is
11

presented. The techniques used to characterize the grown diamond samples are also
presented.
Chapter 4 investigates the SCD pocket growth process. This Chapter defined
conventional, benchmark pocket holder configuration, i.e. the depth, d, and the width, w,
of the pocket were held constant at 2.3 mm and 6.0 mm respectively. The process is
carried out with this constant pocket geometry and with both a variable substrate
temperature and a variable microwave input power as a function of the time. Flat and
smooth diamond surfaces were grown with enlarged diamond surface areas due to a high
lateral growth rate. These growth results were compared with earlier results [Nad] and
showed an improvement in the diamond lateral growth area.
Chapter 5 presents the results of an investigation that explored the SCD growth which
was held at a constant substrate temperature of 1020 °C ±5°C and a constant pressure
of 240 Torr, and were conducted with a constant pocket aperture and with variable pocket
depths. The pocket depths were: 2.0 mm, 2.3 mm, 2.6 mm and 2.9 mm. Results revealed
that the deeper pockets do not grow larger area SCD crystals but have a reduced lateral
growth rate and a reduced area gains and often lead to final SCD crystals with smaller
ending surface areas than the initial seed surfaces areas. Shallow pockets have
enhanced lateral growth rates and larger area gains. Also, in this chapter, the importance
of stopping the diamond growth process before it grows out of the pocket is described in
detail. The results from the experimental variation of the methane concentration for a fixed
pocket holder are also presented. The results of the characterization techniques, such as
SIMS analysis and birefringence images on selected CVD plates are presented.

12

In order to promote the outward diamond growth, Chapter 6 presents a set of
experiments in which the pocket holder depth was held constant while the pocket aperture
varied, i.e. the pocket widths were 6.0 mm, 6.6 mm and 7.0 mm. The results showed for
a single experimental run the lateral diamond surface could be up to 2.3 times size than
the initial original HPHT diamond seed area. The results of the SCD re-growth
experiments are shown. One more diamond growth step is added to the smooth and large
diamond top surfaces of the first growth step. In these extra growth experiments, the
diamond top surface was further enlarged. A new pocket holder was designed to allow
the expansion of the top surface area. A new holder configuration is described. The
resulting grown SCD substrates had a top surface area that was 2.5 times original HPHT
diamond seed area.
In Chapter 7 photographs were taken during several typical diamond growth
processes in order to analyze the plasma variation and the lateral SCD growth as a
function of the growth time. The results show top surfaces area growths as a function of
time and also show the PCD growth on the holder as a function of the time. This result
indicated that the PCD growth on the holder should be minimized.
Finally, Chapter 8 presents the conclusions and summary of this thesis research. Also,
different future activities are suggested to continue to expand the diamond lateral growth.

13

CHAPTER 2

LITERATURE REVIEW: DIAMOND MATERIAL AND SINGLE
CRYSTAL DIAMOND GROWTH VIA MICROWAVE PLASMA
ASSISTED CHEMICAL VAPOR DEPOSITION

2.1 Introduction
Diamond is one of the most unique materials which have ever been discovered. It has
a group of excellent physical and chemical properties. Whether it comes from nature or it
is human-made, it has superior material properties when compared with all other
materials. However, currently, the lab-grown diamond materials usually have few defects
that are commonly seen in the natural diamonds. Among the diamond growth methods,
chemical vapor deposition (CVD) of diamond has recently received a great deal of
attention in the material sciences because it has been used in many new applications of
diamond that had previously been either too expensive to implement or too difficult to
make.
2.2 Structure and Material Properties of Diamond
Diamond is the sp3-bonded allotrope of carbon. Figure 2.1 shows a unit cell of a
diamond crystal, which is composed of eight corner atoms, six face-centered atoms and
14

four internal atoms along the cube diagonals. A diamond unit cell contains eight C atoms.
The crystal structure can be viewed as two interpenetrating FCC lattices, offset from one
another along a body diagonal by one-quarter of its length. Each carbon atom bonds with
four other carbon atoms to form a tetrahedral configuration.

Figure 2. 1 FCC structure of the diamond crystal [21].

At room temperature, the lattice constant a0 for diamond is equal to 3.567 Å (0.3567
nm), from which a few other quantities can be derived. The C-C bond length d is equal to
0

a quarter of the cube diagonals, that is: d  3a / 4  1.54 
The corresponding C atomic radius, R, is given by:

R

d
 0.077nm
2

(1)

And its Atomic Packing Factor (APF) can be calculated as:
APF 

volume.of .atoms.in.aunit.cell Vs

total.unit.cell.volume
Vc

(2)

3
The volume of a sphere is 4 / 3R , since there are eight atoms in diamond unit cell,

therefore,
15

4
32
Vs  (8) R 3  R 3
3
3

(3)

And the total unit cell volume for diamond is Vc = a03. From equation (1), we have
R  d /2 

3
3
a0 / 2 
a0
4
8

(4)

Therefore,
3

32   3 a 3 
3  8 0
Vs

  0.34
APF 

3
Vc
a0

(5)

The density can be calculated by


nA
Vc N A

(6)

Where:
n is the atomic number in each diamond unit cell;
A is the atomic weight, for carbon, it is 12.011 g/mol;
NA is the Avogadro number; it is 6.023x1023 atoms/mol
Hence,


(8.atoms / unit.cell )  (12.011g / mol )
 3.515 g / cm 3
3
23
(0.357nm)  (6.023x10 .atoms / mol )

(7)

This value is much higher than that of graphite (approximately 2.3 g/cm3)
The structure of diamond accounts for its outstanding properties. Some of the outstanding
properties of diamond are listed below in Table 2.1.

16

Table 2. 1 Principles properties of diamond [21].
Property
Hardness
Strength, tensile
Friction Coefficient
Sound velocity
Density
Young’s modulus
Thermal expansion coefficient
Thermal conductivity
Debye temperature
Optical transmissivity (from nm to far IR)
Dielectric constant
Electron mobility
Hole mobility
Electron saturation velocity
Work function on (111) surface
Bandgap
Resistivity

Value
1.0x104
>1.2
0.03
1.8x104
3.52
1.22
1.1x10-6
900 – 2320
2000
225
5.7
1800 – 2200
1200 – 1600
2.7x107
Negative
5.45 – 5.50
1013 – 1016

Units
kg/mm2
GPa
Dimensionless
m/s
g/cm3
GPa
K-1
W/mK
K
Dimensionless
Dimensionless
cm2/V-sec
cm2/V-sec
cm/s
eV
eV
Ω-cm

These properties demonstrate the suitability of diamond as a potential candidate to be
an engineered research material for many applications. It is well-known that diamond is
the hardest substance found in nature, and it is well-know that diamond is four times
harder than the next hardest natural mineral, corundum (sapphire and ruby) [22].
Regarding diamond’s physical properties, diamond is the ultimate material in several
ways including: (1) its extreme hardness, (2) it has a value of 10 on the Mohs scale of
mineral hardness, and (3) it is at the top of the hardness scale. Diamond has excellent
properties including a wide-range of optical properties; excellent thermal conductivity; low
coefficient of thermal expansion; the highest melting point; and remarkable
semiconductor properties [23].

17

2.3 Historical Overview of the CVD Diamond Process
The synthesis of diamond has been pursued ever since the French chemist Antoine
Laurent Lavoisier discovered that diamond was a form of crystalline carbon [24]. The first
report describing successful diamond synthesis using HPHT methods was published in
1955 [25], in which diamond was crystallized from solvated metal carbon at high
pressures of pressures of 7 – 10 GPa and temperatures of ~1800°C. Experiments with
low-pressure synthesis had been ongoing in both the United States and the Soviet Union
during the same period. In 1962 the first patent emerged by William Eversole at the Union
Carbide Corporation [26] who reported diamond synthesis under low pressure by a CVD
method. The diamond was grown on diamond substrates with the use of carboncontaining gasses at a pressure of 10 bar and a temperature of around 800 – 1000°C. In
the mid-1970’s Spitzyn a Soviet group [27] showed that diamond crystallization by CVD
should preferably be carried out at reduced pressure. Also, he reported diamond crystals
up to several tens of microns in thickness were also grown on non-diamond substrates.
This is a significant achievement as growth had only previously been possible on diamond
substrates. In 1974 research on CVD diamond growth was initiated in Japan. During
1981 S. Matsumoto at al. [28] reported using a Hot Filament CVD technique in which
atomic hydrogen was incorporated into the growth process to give good quality films on
some substrates with significantly better deposition rates, making lab-grown diamond for
the first time potentially commercially attractive. They also were the first to report on
rapidly growing diamond by microwave plasma assisted CVD. Diamond was grown at
pressures of 10 - 100 Torr and temperatures were around 800 – 1000°C. Widespread
interest developed from there, and at present many other plasmas assisted, and thermally
18

assisted CVD deposition methods have been established. The deposition method that
currently has the best potential for obtaining high-quality diamond substrates is the
microwave plasma assisted CVD (MPACVD) [29].

2.4 Processes of Diamond Growth
2.4.1 Phase Diagram of Carbon
Figure 2.2 shows the phase diagram of carbon, describing the equilibrium phase
structure of carbon at different temperatures and pressures [30]. It can be seen from the
Figure 2.2, under relatively low pressures (the bottom region of the Figure), graphite, the
most common form of carbon, and is stable from room temperature up to its melting
temperature at approximately 4000°C. Whereas diamond is meta-stable at low-pressure,
low-temperature conditions and stable only at very high pressures. Diamond exists in
nature. However, natural diamond is scarce and costly. This has motivated scientists
around the world for more than one century to explore the laboratory growth of diamond
using various techniques with different C sources. At present high-temperature high
pressure (HPHT) and chemical vapor deposition (CVD) are the most two commonly used
techniques for diamond [29].

19

Figure 2. 2 Pressure-temperature phase diagram for carbon [31].

2.4.2 HPHT Diamond Growth
Since both diamond and graphite are pure carbon allotropes, and graphite is richly
available in the earth, the synthesis of diamond using graphite is reasonable. Although
the difference in standard enthalpies for diamond and graphite is only 2.9 kJ.mol-1 [30],
a large activation barrier separates these two carbon phases and prevents the mutual
conversion between them at room temperature under atmospheric pressure conditions.
The phase diagram of carbon, shown in Figure 2.2, indicates the appropriate
directions towards the conversion of graphite to diamond. Although graphite is stable at
room temperature and low pressures, diamond, whose density is much higher than
graphite (3.52 g/cm3 for diamond and 2.25 g/cm3 for graphite), becomes more stable with
20

high pressure. However, the thermal energy at room temperature is not enough to break
the C-C bonds in graphite, and the diffusion of carbon is very difficult. Thus, the high
temperature is also a requirement to convert graphite into diamond. The fact that natural
diamond is produced at high pressure and high temperature in volcanic shafts also
suggests the possibility of diamond growth through heating graphite under extremely high
pressures. The HPHT diamond growth process essentially duplicates the natural diamond
formation process, and the synthesized diamond exhibits similar properties [4]

to

naturally formed diamond.
The first two reports describing successful diamond growth using HPHT methods were
first published by Everlose in 1955 [25], and by the General Electric in the 1960s [31].
Growth pressures as high as 12.5 GPa and associated temperatures as high as 3000 ºC
were reported. Since then, the HPHT technique has been the main method to produce
“industrial diamond” for several decades. During the HPHT process, the transformed
diamond must be cooled under high pressure in order to avoid its transformation back
into graphite. However, these conditions are very harsh and costly to achieve.
Fortunately, it is possible to overcome these limitations via solvent-catalyst reaction [4].
The solvent-catalyst process was invented and developed by General Electric [4], [32].
It established a reaction route with activation energy lower than that of the direct
transformation, and thus a faster transformation under more benign conditions is feasible.
As a result, the solvent-catalyst HPHT diamond growth is readily performed and has been
widely used in industry to produce a diamond. The solvent-catalysts used during the
HPHT process are transition metals and their alloys, including Fe, Co, Cr, Ni, Pt and their
alloys. These metal-solvents can break the bonds between carbon atoms, dissolve

21

carbon and improve the diffusion of carbon to the growing diamond surface. The most
common catalysts currently applied include Fe-Ni and Co-Fe alloys [4].
The main advantages of the HPHT grown diamond over the natural one are the ability
to control the uniformity, size, crystal habit, crystal friability, the higher general quality and
consistency [33]. Research in HPHT diamond growing is continuing in the direction to
lower the production costs and to increase the crystal size.

2.4.3 CVD Diamond Growth
Besides the high processing cost, the diamond synthesized by HPHT is usually in the
form of powders/particles or small bulk materials with the size ranging from nanometers
to a few millimeters. These disadvantages limit its more widely technological applications
[33]. The discovery that it is possible to grow PCD diamond films in the 1980s has further
increased the world's attention improving diamond technologies. These experiments by
Matsumoto et. al. [34], [35] demonstrated that it is possible to produce polycrystalline
diamond thin films by CVD techniques. Their experiments typically used hydrogen (H2)
and methane (CH4) gas mixtures at low pressures (vacuum < 75 Torr). Their CVD grown
diamond displayed mechanical, tribological, and even electronic properties comparable
to those of natural diamond, as shown in Table 2.1 [15-19].

2.5 Physical and Chemical Aspects of CVD Diamond
2.5.1 General Overview of the Surface Reactions
The homoepitaxial growth, by which diamonds grow from diamond crystal seed, has
been intensively studied since the 1960’s. At first sight, the reaction in CVD of diamond
22

seems to be very simple and only involves the decomposition of the carbon-containing
species, e.g. CH4 as follows [4]

CH 4 
 activation 
 C  2H 2

(1)

As a matter of fact, the mechanism of growing diamond by CVD is a complex
integration of many factors and has not been fully understood. Nevertheless, it has been
found that there are two essential conditions which are crucial for the process [36]: (a) the
activation of the carbon containing species and (b) the presence of the atomic H.
The growth of diamond from the chemical vapor phase takes place through a group
of chemical reactions with a competition between the formation of diamond structures
and non-diamond (including graphitic) structures. In the diamond growth using hydrogen
(H2) and methane (CH4) gas mixtures as the reactants, Figure 2.3 demonstrates the
possible physical and chemical processes that are involved in CVD diamond growth [33],
[37].
Firstly, reactive gasses of CH4 and H2 are mixed and then the gas mixture is
introduced into the reactor. Then they pass through a high temperature or energetic
region where the molecules in the gas phases are broken and activated to form reactive
radicals and atoms by, for example a microwave plasma. It is believed that CH 4
decomposes into two primary stable species: the methyl radical (CH3) and acetylene
(C2H2) [38]. Beyond the activation region, these reactive species continue to be
transported and mixed by forced flow, diffusion as well as convection throughout the
reactor and undergo a complex set of chemical reactions until they strike the substrate
surface.

23

Figure 2. 3 Schematic of the physical and chemical processes occurring during the CVD
growth of diamond [34], [38].

On the substrate surface, various processes may occur. The species can either be
absorbed by the surface or be desorbed again back into the gas phase. They can diffuse
along and into the substrate surface or react with the substrate to form various structures
depending on the processing conditions and exposed substrate surface states.
Hydrogen, plays an essential role in the deposition of diamond thin films by CVD it is
created by dissociation of H2 due to the impact of electrons or due to other thermal
dissociation processes. Atomic hydrogen, which promotes reactions, is among key
ingredients in the chemical vapor deposition process [39]. The main functions of atomic
H in the diamond deposition include:

24

1) Keeping diamond lattice stable by terminating the dangling bonds (presence of
high-energy sites) on the surface of diamond films.
2) Creation of dangling bonds for carbon-containing radicals, e.g. CH3, to be attached
to the growing diamond surface by reacting with the H on the diamond surface to
form H2.
3) Creation of CH3 radicals by reacting with CH4. Even though CH3 radicals are
predominantly created through the decomposition of CH4 by thermal dissociation
or impact dissociation, the reaction between H atoms and CH4 will also create CH3
radicals, which can then be attached to suitable surface sites to grow diamond.

25

Figure 2. 4 A schematic of the reaction process occurring at the diamond surface leading
to the stepwise addition of CH3 species and diamond growth [33].

A schematic diagram of growth of diamond involving atomic hydrogen on the (100)
plane is shown in Figure 2.4 [40]. Diamond growth can be considered as a stepwise
addition of carbon atoms to the existing diamond lattice, catalyzed by the presence of
excessive atomic H. At first, the reaction between the atomic H in the gas phase and the

26

H atom bonded to the sp3 carbon dangling bond induces the formation of H2 gas, leaving
behind a reactive surface site. CH3 radicals (gas phase) then collide and react with the
reactive surface site, adding a carbon atom to the lattice. The same process of H atom
abstraction and CH3 addition may happen on the adjacent site. The further H abstraction
on one of the chemisorbed groups creates a radical, which attracts another nearby carbon
group to form the ring structure, locking the two carbon atoms in the diamond lattice.

2.5.2 The Specific Diamond Growth Reactions
Numerous synthesis and decomposition reactions are involved in the diamond growth
process. Goodwin’s [17] scaling model summarizes the essential growth sub-processes
and highlights the essentials of diamond CVD growth. As mentioned earlier, once the
active species form, there are several sub processes taking place on the surface of the
diamond growth: activation of surface sites, adsorption/desorption of active radicals, and
incorporation of adsorbates into the diamond crystal lattice. The following are the four
basic sub processes included in this model:
(1) Surface Activation
A steady-state surface radical site coverage is established in this process. In the
growth atmosphere, there are many active species besides the precursor gases
themselves. The atomic hydrogen and active carbon radicals are part of the active
species. Carbon atom on the surface of the diamond layer has one dangling bond, which
is usually terminated by hydrogen. The surface activation process is to remove this
termination hydrogen and leave the free active carbon atom with an open dangling bond

27

(see Figure 2.4). The following reaction describes this abstraction of surface-terminating
hydrogen by gas-phase atomic hydrogen:
k2

C H  H  C d*  H 2
*
d

(2)

k1

Here CdH represents a generic hydrogenated surface site, and C*d represents the site
with an open dangling bond after the hydrogen is removed (a radical carbon site). k1 is
the reaction rate constant for the flux of atomic hydrogen [H] on the surface. k 2 is the
reaction rate constant for the adsorption of an atomic hydrogen on the surface. This
radical carbon site, C*d reacts with atomic hydrogen on the surface and this reaction is
expressed as:
k2

C  H  Cd H
*
d

(3)

k1

With these two reactions, the rate of creating C*d is expressed as:

r1  k1 [C d H ][ H ]  k  2 [C d H ],

(4)

And the rate of recombining C*d back to CdH is:

r1  k1[Cd* ][ H 2 ]  k2 [Cd* ][ H ]

(5)

(2) Adsorption
In the adsorption process, active hydrocarbon species in the atmosphere attach to the
diamond surface at the radical sites (the ones with open dangling bonds) and combine
with the diamond carbon atoms. This process is described as:
k3

C  C n H m  Cd A
*
d

(6)

28

Where CdA represents the adsorbates formed by the active carbon radical species (CnHm)
and surface diamond carbon atoms (C*d) and k3 is the reaction rate constant for the
adsorption of a carbon containing radical on the surface.
(3) Desorption
Desorption is the reverse process of adsorption. In this process, the surface
adsorbates return to the gas phase either thermally or chemically. In the thermal
desorption process, the carbon radical species are removed from the surface via the
thermal energy, which is expressed as:

k4

Cd A  Cd*  Cn H m

(7)

Where k4 is the reaction rate constant for the thermal desorption of a radical from the
surface. In the chemical desorption process, the carbon species are removed with the
assistance of atomic hydrogen. The atomic hydrogen returns the attached carbon radical
to the gas phase as the carbon precursor species. Chemical desorption is usually called
“etching”. Desorption of atomic hydrogen is explained as:

k5

C d A  qH  C d*  products

(8)

Where k5 is the reaction rate constant for the abstraction of an atomic hydrogen from the
adsorbed carbon containing radical. Desorption is a critical process for the crystalline
diamond growth, where the incorporation of the growth species occurs primarily at surface
steps, which then move across the crystal surface. The desorption sub-process serves
to remove the undesired adsorbates on the substrate surface and facilitates diamond
growth.
29

(4) Incorporation
This is the process by which diamond grows. In this process, the adsorbates C dA are
incorporated into the diamond lattice with the assistance of atomic hydrogen, as indicated
by the reaction:
k5

Cd A  (m  1) H  Cd H  (m  1) H 2

(9)

Where m represents heavier molecules. During this process, atomic hydrogen promotes
the incorporation by assisting the removal of the byproducts of the gas phase. The
incorporated portion of the active carbon radical species stays at the substrate and forms
a new layer of diamond.
The conditions that determine the rate of the four mechanisms listed above are
surface temperature, gas phase atomic hydrogen concentration at the surface, and gas
phase radical carbon concentration at the surface. In the steady-state conditions, this set
of reactions can be combined to give a growth rate G of:

G  k3

ns
nd

 k1

 k1  k 2

 [CH 3 ][ H ]

 k 4  [H ]
k5

(10)

Where ns is the surface site density (2.61 x 10-9 mol/cm2), and nd is the molar density of
diamond (0.2939 mol/cm3). With the simplification of the growth process, Harris and
Goodwin [17] predicted the steady state radical fraction and simplified the growth rate
equation (10) further:

G  1.8 x1011

[CH 3 ][ H ]
5 x10 9  [ H ]

(11)

30

The chemical growth reactions have been investigated from early 1990’s, and many
of these complicated growth processes have been simplified into several chemical
reactions as discussed in the references [17], [41], [42].

2.5.3 Growth of Defects
The challenge in the diamond growth process with higher rates is to do so while
maintaining high-quality diamond growth, i.e. low defect density. There are many types
of point and extended defects in diamond CVD growth process, such as sp2 carbon,
substitutional impurities, vacancies, interstitials, dislocations, twins, etc. [17]. Their
generation depends on the local chemical environment. Out of different possible defects,
Goodwin’s scaling model predicts the incorporation of sp2 carbon.
The underlying assumption of the defect generation model is that defects are
generated when an adsorbate reacts with a nearby adsorbate before it is fully
incorporated into the lattice. Once two of them react to form a sp2 group, they are usually
overgrown as a sp2 defect. The reaction is written as:
k def

Cd A  Cd A  HCg  Cg H

(12)

Therefore, the defect fraction, Xdef in the grown surface will be:

X def 

k def [C d A] 2
(13)

k i [C d A][ H ]

Where kdef[CdA]2 is defined as the rate of defect generation and ki[CdA][H] is the rate of
diamond formation. According to equation (10):

31

[C d A] 

nd G
ki [H 2 ]

(14)

And from equation (12) and (13):

X def 

k def nd G
ki2 [H ]2

(15)

Where ki is a constant at a given temperature. Therefore, the defect fraction in the grown
diamond grown is proportional to the growth rate and inversely proportional to the square
of the amount of atomic hydrogen on the surface at a constant substrate temperature. In
general, if the defect generation reaction depends on [H] and the diamond formation
reaction has more than 1st order dependence on [H], the defect fraction in the grown film
will be:

X def 

kdef nd G
(16)

ki2 [ H ]n

This relationship predicts that the film quality and growth rate are inversely related, and
the increase of atomic hydrogen at the surface improves film quality.

32

2.5.4 Numerical Modeling of the Harris-Goodwin Model
F. Silva et al. [9] developed numerical plasma models to compute the [H] and [CH3]
concentrations in the gas phase as a function of the deposition parameters, i.e. pressure,
microwave power, and methane concentration. They noticed that in their reactor the
electron temperature drops from 15000 to 10000 K, and the gas temperature varies from
2000 to 3600 K as the pressure increases from 37.5 Torr to 225 Torr. This is due to the
increase in electron-heavy species collision frequency which results in an enhanced
energy transfer from the electron gas to the heavy molecule species. They pointed out
that, at low pressure (< few tens of Torr), H is mainly produced by the direct electronimpact dissociation reaction:

e   H 2  e   2H

(17)

However, at higher pressure, H is mainly produced by the thermal dissociation of
hydrogen molecules:

H 2  H  3H

(18)

This result shows that as pressure increases and the gas temperature increases to more
than 1400 K, then there is an associated significant increase in atomic hydrogen density
[H]. The modeling results also show that at low pressures, the CH3 radical production in
the plasma bulk is governed by H-atom concentration through the reaction:

CH 3  H  M  CH 4  M

(19)

Here M is the third body. At high pressure, CH3 production is still due to CH4 dissociation
through collisions with H atoms through the following reaction:

CH 4  H  CH 3  H 2

(20)
33

This reaction mainly takes place very close to the substrate surface. In fact, the
production rate of CH3 also depends on the gas temperature. An optimal CH3 production
is obtained for gas temperatures of 1200 – 2200 K. Also, they calculated for their
MPACVD reactor the operating points regarding H-atom and CH3-radical concentrations
for different pressures and input power discharge conditions and then plotted them on the
process map from Goodwin [17] in [CH3]-[H] space showing the operating ranges of the
primary CVD diamond growth processes. The process map is shown in Figure 2.5.

Figure 2.5 Process map in [CH3]–[H] space showing the operating ranges of the main
CVD diamond growth processes (from [9]). The evolution of the operating point of the
LIMHP reactor as a function of working pressure (from 50 to 300 mbar) and methane
concentration.
34

In Figure 2.5 the flux of the methyl radicals is plotted against that of the atomic
hydrogen. This plot shows a dotted black line, which indicates the constant graphitic
defect fraction contours (equation 16). The solid black lines indicate the constant growth
rates on the (100) oriented diamond surface (equation 11). For a given methane
concentration, the increase in the pressure from 37.5 to 225 Torr leads to three orders of
magnitude increase of the H-atom concentration and one order of magnitude increase of
CH3 radical concentration. The results show an increase of growth rate to several tens of
μm/h. As one would expect, at a constant pressure, the relative defect density with an 8%
methane concentration is higher than that with a 1% methane concentration. However,
for a constant methane concentration, the diamond quality can be improved by increasing
the operating pressure. Consequently, it shows that increasing methane concentration at
high pressure leads to a significant increase in the growth rate without compromising the
film quality. That implies that at higher pressures the growth rate increases and highquality diamond can be grown with higher methane concentrations than at lower pressure.
The high diamond quality, high growth rate synthesis window increases with the highpressure operation

2.6 MPACVD Growth of Diamond
This section reviews the basic principles of nucleation and experimental growth results
on a single crystal diamond surface deposited by CVD process. Before reviewing
previous research on diamond growth first the important aspects of homoepitaxy growth
are discussed.

35

Figure 2. 6 Map of surface morphologies as a function of misorientation angle from the
(001) and methane/hydrogen ratio. Three types of morphologies appear. Atomically flat
diamond is preferred, which only grows in the region (III) [43].

Diamond growth by the microwave plasma assisted CVD (MPACVD) method was
grown first by Kamo et al. in 1988 by growing on a polycrystalline diamond substrate [44].
After this demonstration, the investigations moved forward to the homoepitaxial diamond
growth on natural or HPHT Type Ib substrates. At the early stage of the investigations,
the reported growth rates were typically lower than 1µm/h [45]–[47]. Also, methane
concentrations, usually defined as the ratio of methane to hydrogen flow rates, had to be
36

maintained at 0.025% to suppress the growth of defects. Results revealed visible defects
in the homoepitaxial diamond growth: no-epitaxial crystallites (UC), pyramidal hillocks
(PH), and macro-step bunching [43], [47]. Takeuchi et al. [43] grew homoepitaxial
diamond films on substrates with different misorientation angles from the (001). They
produced a map of diamond surface morphologies as a function of misorientation angle
and methane concentration as shown in Figure 2.6.
However, additional works demonstrated that a different region of parameter space
could be effective in growing diamond at much higher rates. Yan et al. grew diamond at
rates approaching 40 µm/h in the presence of nitrogen [48] with a growth temperature of
1500°C, pressure between 150 – 200 Torr and methane concentrations from 1% to 6%.
Growth rates were increased to 150 µm/h by introducing 1% – 5% N2 relative to methane
flow, and methane flow up to 12% of the hydrogen flow [13].
At a later stage, by lowering the growth temperatures to 700 – 800°C and increasing
the microwave power density, fairly high growth rates were achieved in the absence of
nitrogen [49]–[53]. Also, whereas gas pressures previously used ranged from 20 - 50
Torr [54], for high growth rates the experimental pressures were increased to 100 – 200
Torr. The microwave power densities were around 5 W/cm 3 but for high growth rates
increased to 30 – 100 W/cm3 [54]. Table 2.3 summarizes the growth pressures, the
growth rates, and the power density ranges of these experiments.

37

Table 2. 2 Growth rates ranges in MPACVD diamond growth [49]–[53].

Parameter

Growth rate
< 1µm/h

Power density (W/cm3)
Substrate Temperature (°C)
Pressure (Torr)

5
600 – 900
20 – 50

Growth rate
100 µm/h with N2
20 µm/h without N2
30 – 100
700 – 1200
100 – 200

2.6.1 Pyramidal Hillocks and Non-Epitaxial Crystallites
On the (100) diamond surface, the most frequent defect is the pyramidal hillock, either
with a flat top surface or with a sharp truncated top containing a non-epitaxial crystallite
[55]. The sides of this hillock are aligned with [110]. Under certain growth conditions, such
as substrate temperature and methane concentration, the hillock grows very rapidly
compared to its surroundings [49]. Figure 2.7 shows a homoepitaxial diamond surface
with pyramidal hillocks; the pyramids are usually truncated with non-epitaxial crystallites
on the top. Tallaire et al. [56] reported above 900°C a roughening of the crystal surface
due to the formation of pyramidal hillocks defects, while at reasonably lower temperature
the surface remains smooth.

38

Figure 2.7 Optical microscope images of the epitaxial CVD diamond films grown at
1000°C (a) Sample overview (b) Magnification 20x [49]. (c) Optical micrograph of a 100
µm thick diamond film grown on an on-axis (001) oriented Ib substrate [57]. (d) and (e)
CVD diamond film is grown at high temperature (1000°C and 920°C) exhibiting growth
pyramids [56], [58].

Lee et al. [59]–[61] reported the type of growth strongly depends on the diffusion of
the growth species on the surface, the rate at which growth species are deposited on the
surface from the vapor phase, and the width of the terraces on the growth surface. If the
terrace width is long compared to the diffusion length, then the growth species tend to
deposit on the middle of the terraces, instead of diffusing to the steps. The results in two39

dimensional nucleation, as shown in Figure 2.8. Then, the two-dimensional nuclei provide
shorter steps for further growth along the sides of the hillock. Non-epitaxial crystallites
grow on surface defects by enhanced secondary nucleation [59]. At higher temperature,
these non-epitaxial crystallites grow very fast and provide step edges for further growth.

Figure 2.8 Mechanism of two-dimensional nucleation on a wide terrace, beginning of
hillock growth [59].

2.6.2 Step bunching
If the diffusion length of the growth species is larger than the terrace width, the growth
species diffuse toward the steps [59] because of the lower potential there. In this growth
model, the steps appear to flow across the growth surface. A schematic view of the step
flow growth mechanism is shown in Figure 2.9. At high temperatures and low methane
concentrations a step flow increases on the diamond surfaces. Several reports have
appeared on single crystal diamond growth in step flow mode on the substrate surfaces
[43], [45], [58], [62]–[65].

40

Figure 2. 9 Schematic view of step flow growth showing the diffusion of growth species
to steps and subsequent flow of steps to the right [59].

During growth, impurities or defects may obstruct the transport of diamond growth
species to a particular step [66], so that flow ceases, but other steps will flow to that
position and become pinned. If this process continues, many steps will become bunched
at that position leading to a macroscopically high step. That is referred to as macro-step
bunching. Figure 2.10 shows macro-steps bunching in homoepitaxial diamond growth.

41

Figure 2. 10 Optical micrographs of the grown surfaces of CVD diamond with (a), (b) no
intentional N2 addition and (c) 2 ppm of N2 and (d) 6 ppm of N2 [58], [63].

2.6.3 Effect of Growth Parameters
2.6.3.1 Substrate Temperature

The substrate temperature is known to be a most important parameter in SCD growth.
Diamond has been grown over a large range of temperatures of several hundred
centigrades, focused around 700 – 1300 °C. Numerous investigations have shown
significant effects on the growth rate and crystal morphology of the CVD diamonds at
different substrate temperatures.

42

The diamond growth rate, Rg was modeled with an Arrhenius relationship using an
associated activation energy, Ea and the growth temperature, T [67], [68]. The general
expression is:

Rg  Ae(  Ea / K BT )
The relation showed a substrate temperature between 800 – 1000°C for a maximum
growth rate. Diamond growth in this temperature region is controlled by chemical reaction
kinetics rather than by surface diffusion.

Figure 2.11 Diamond growth rates as a function of substrate temperature at 4% methane
concentration and a 3.8 kW microwave power [54]. Here Cme is the methane
concentration.

43

Above 1000 °C, the growth rate no longer obeys the above relation but saturates. The
growth changes to a diffusion-controlled mode, with a weaker temperature dependence.
Values of Ea for temperatures between 735 – 970°C were reported as 8 ± 3, 18 ± 2 and
12 ± 4 kcal/mol, respectively, for (100) and (110) and (111) diamond surfaces [69]. A.
Gicquel et al. [67] reported in the temperature range 700 - 850°C activation energy in the
range 11 – 16 kcal/mol. Different growth conditions and reactor geometries will change
Ea, so that the use of these parameters may be sensitive to local conditions. Figure 2.11
shows a plot of growth rate versus substrate temperature.
At temperatures above 1200°C, graphite starts to form [54]. At temperatures lower
than 700°C, the diamond growth rate is very low, and the diamond surfaces are rough
and defective with a non-optimal crystal morphologies [58]. Furthermore, the growth of
smooth and flat diamond surfaces also has been reported at 700°C [50], This apparent
contradiction may not exist if growth also depends on other parameters such as methane
and nitrogen concentrations, and reactor geometries. Usually, there is a small
temperature window in which defect-free diamond can be grown.

2.6.3.2 Methane Concentration
In this thesis, investigation methane is used as the main source of carbon to grow
diamond, and the amount of methane present it usually referred to the methane
concentration. Methane concentration is defined as the flow rate ratio of methane gas to
total gas, which is typically composed mainly of hydrogen. Because the methane gas flow
rate is small under typical growth conditions, the methane concentration is often defined
as the flow rate ratio of methane gas to hydrogen gas (CH4/H2).

44

Methane concentration in the gas phase significantly affects the diamond growth rate
and the morphology of diamond. It has been shown the growth rate is found to be linearly
proportional to methane concentration. Figure 2.12 shows a plot of growth rate versus
methane concentration at 850°C without nitrogen from Achard et al. [58]. The results
indicate a slightly linear relationship between methane concentration and growth rate. In
contrast, Bogdan et al. [70] reported a sublinear result in which the surface topography
and the growth rate does not change dramatically by increasing the methane content from
6% to 15% methane concentration. Teraji et al. [54] reported a linear relationship from
1% to 5% methane, then a linear dependence up to 32% methane at 890°C. Whereas;
the growth rate always increases as methane concentration increases, the proportionality
coefficient can be sensitive to other growth parameters.

Figure 2. 12 The influence of the methane concentration in the diamond growth rate [58]
45

Furthermore, some authors reported that the variation of methane concentration
improved the surface morphology on the diamond crystal. Bogdan et al. [70] reported a
smooth diamond surface grown at 15% methane concentration at 700°C. Achard et al.
[58] published diamond surfaces with macro-step bunching when they increased the
methane concentration from 4 to 6%. At 2% methane concentration, they found the
diamond surface to be highly pitted.
Some authors reported the methane concentration should be reduced below 1% to
obtain high-quality diamond films with a flat surface and non-epitaxial crystalline [71], [72].
However, the growth rate under such conditions is very low (<1 µm/h) which is not
desirable for the growth of thick diamond substrates. Tallaire et. al. [49] subsequently
reported when the concentration of CH4 was increased; the surface morphology was
improved.

46

Figure 2.13 Optical microscope images of the diamond films grown at 850°C with
methane concentration of (a) 2% (b) 4% (c) 6% and (d) 8% [49].

Figure 2.12 shows the optical microscope images of the CVD diamond films grown at
850°C with 2%, 4%, 6% and 7% of methane concentration (CH4). With 2% of CH4, the
growth rate was 2 µm/h, and the film surface was very rough, exhibiting a significant
number of pits. It was suggested that this could be a consequence of a high etch rate by
atomic hydrogen during growth enhanced at such high microwave power. This statement
was also reported by Lee et al. [59] in which at 6% and 8% CH4 the diamond morphology
showed micro steps with the surface close to the single domain structure, at 1% and 2%
CH4 concentration. The surface shows hillocks, and random growth morphology occurred
with the double domain structure.
47

2.6.3.3 Process Gas Pressure
The process gas pressure has an influence on the size and energetics of the
microwave plasma. As pressure increases, the size of the plasma ball becomes smaller
since the mean free path of the ions becomes shorter. This also has been experimentally
demonstrated [16], [73]. This influences the substrate temperature, since the power
density may become greater in the vicinity of the substrate. In a microwave plasma CVD
system, the electrons in the plasma are accelerated by the high-frequency electric field,
but the ions cannot follow the microwave field, due to their heavy mass, leading to a low
ion temperature. The ion temperature can be increased by enhancing collisions with hot
electrons, brought about by a higher gas density. When the total gas pressure is low, the
ionization rate in the plasma is also low with only a relatively small number of electrons
available for creating atomic hydrogen and carbon radical species. This results in low
densities of growth radical species and results in a low growth rate that is attributable to
low excitation density of these species.
At higher pressures and power densities, the ionization rate is higher, and the
frequency of collisions between the ions and electrons become larger, which increases
the ion gas temperature. That ultimately leads to an increase in the density of carbon
growth species and atomic hydrogen. Mokuno et al. [11] found the growth rate increasing
almost linearly with pressure in their microwave plasma CVD reactor in the presence of
nitrogen. Li et al. [74] investigated the growth rate and quality of diamond film with
pressure up to 125 Torr. They reported a linear dependence on the growth rate and the
pressure when operating with a fixed methane concentration. Also demonstrated was that

48

the diamond quality increased with increasing pressure. Yan et al. [48] reported a fivefold increase in growth rate when the pressure was changed from 60 to 200 Torr.

2.6.3.4 Microwave Power Density
The microwave plasma density on the substrate surface is enhanced with increasing
the gas pressure because the plasma density increases as pressure increases. The
diamond growth rate is strongly affected by increasing the radical species densities on
the substrate. Consequently, the growth rate increases as pressure increases.
CH3 carbon radicals can be formed at the plasma boundary because CH4 molecules
that are decomposed in the plasma center are possibly cooled at the boundary because
of that region’s low gas temperature. When the microwave input power which creates the
plasma is weak while the pressure is substantially high, recombination reactions among
precursors become considerable. Consequently, the methane concentration dependence
on the growth rate changes from linear to sublinear with the pressure increasing [54]. This
fact implies that the microwave power density must be high to produce a sufficient quantity
of radicals when the pressure increased.
Tallaire et al. [75] observed a three-fold increase, from 3 µm/h to 8.5 µm/h in growth
rate, with a change of microwave power density from 65 to 125 W/cm3. They also used a
pulsed microwave source to increase the peak power density and while maintaining a
constant average power density. They observed an increased growth rate of 40%, with
the conclusion that peak power density, not average power density, is responsible for the
enhanced growth rate. The deposition rate also increased by increases in both the power
density and the methane concentration. They found that the growth rate can be increased

49

six-fold by increasing both methane concentration up to 7% and microwave power density
up to 125 W/cm3. The results are shows in Figure 2.14.

Figure 2.14 Evolution of the growth rate of the CVD diamond films as a function of the
injected MW power density in the discharge using 4% of CH4 (square dots) or 7% of CH4
(triangular dots) [75].

2.6.3.5 Nitrogen
Nitrogen is an important impurity that plays an essential role in the growth of CVD
diamonds. Nitrogen forms several defects centers in diamond and reduces the optical
transparency in the region above 1.7 eV and show an absorption band in the infrared at
wavelengths beyond 9 µm due to the enhancement of defect induced one-photon
absorption [76]. Small amounts of nitrogen into the gas phase also has been shown to
have a large influence on the morphological and growth rate of CVD diamond.
50

The nitrogen can change the diamond surface texture to a pattern of crystallites with
rectangular coplanar (001) facets, thereby reducing the surface roughness and improving
the crystalline quality [77]. Furthermore, it generally induces an increase in the deposition
rate by factor of 2 to 10 [13], [52], [58], [66], [78], [79]. These authors suggested that a
controlled amount of nitrogen in the gas phase enhances the abstraction of hydrogen on
the diamond surface. This increases the diamond phase and influences the growth rate.
However, has been found that significant amounts of nitrogen cause deterioration of the
films or even inhibit diamond growth [80].
It has been shown from many experiments that the influence of nitrogen on the growth
and morphology of diamond grown (100) and (111) faces is completely different. Theije
et. al. [66] have shown that diamond films which grown on (001) surfaces in the presence
of nitrogen have an entirely different morphology from diamond films grown without
nitrogen. Whereas the latter is quite smooth with many squares, shallow growth hillocks
and grown polycrystalline on the edges, the morphology of the diamond films with
nitrogen is determined by the presence of step bunches, see Figure 2.14. The nitrogen in
the films acts as an impurity impinging on a growing surface, thereby destabilizing uniform
step trains by hampering the propagation of individual steps, leading to a step bunching
instability [81]. So far, the mechanism by which nitrogen influences the (001) diamond
growth is not well understood. The mechanism such as surface poisoning or the
generation of additional growth steps has been considered [77]. According to a theory
based on defect-induced stabilization of diamond growth [82], diamond growth is only
thermodynamically possible by the creation of vacancies, since a diamond vacancy can
be created with lower energy than a graphite vacancy. If nitrogen is doped into diamond,

51

the Fermi level at the growth surface is shifted towards the conduction band, which
reduces the energy of vacancy formation in diamond, but not in graphite, so that diamond
growth is favored over graphite.

Figure 2.15 Optical microscope images of the surface morphology of epitaxially grown
diamond films deposited on (001) oriented diamond substrates without (a, b) and with (c,
d) nitrogen added to the gas phase [66].

The morphology of the diamond grown on (111) oriented shows a different
morphology upon nitrogen addition. Without nitrogen addition, these surfaces are a

52

smooth with many shallow growth hillocks and a few micro twins. However, upon nitrogen
addition, a significant increase in twin formation is observed [83].
A lot of nitrogen concentrations were found to deteriorate the diamond film and
decrease the growth rate. That is attributed to the increased amount of carbon/nitrogen
ratio, which would reduce the concentration of growth promoting CH3 radicals in the gas
phase. The carbon/nitrogen ratio may competitively adsorb onto vacant diamond growth
sites and then induce increased twinning and stacking faults [84].
The role of nitrogen in the growth and the diamond surface continues to be of interest
to the diamond researchers either to considerably improve the growth rate, the
morphology or to improve the diamond applications. In this thesis investigation, no
nitrogen is added to the diamond growth process.

2.7 Review of Single Crystal Diamond Growth via MPACVD
The overview of single crystal diamond growth via microwave plasma assisted
chemical vapor deposition is focused on the role of growth parameters on the diamond
growth process. A SCDs growth review is summarized below by groups of researchers
worldwide. Each group has contributed enormously to develop strategies to grow
diamonds. The following is a compilation of recent breakthroughs since 2014. Previous
SCD growth investigations have been described in other MSU Ph.D. research
dissertations, for example from J. Lu and S. Nad [85], [86].

53

2.7.1 Carnegie Institute of Washington, USA
Researchers at the Carnegie Institution’s Geophysical Laboratory have demonstrated
growth up to 10-carat SCD at rapid growth rates (150 micrometers per hour) depending
on the growth conditions, using a microwave plasma assisted chemical vapor deposition
(MPACVD) process. The diamond crystals were grown at high temperatures range of
1100 - 1300°C. The high growth rate was achieved by adding a small amount of nitrogen
gas into the conventional hydrogen and methane gas mixture at high temperatures.
However, the high growth rate is limited by the growth time due to the thick polycrystalline
diamonds which are also deposited on the substrate holder [13]. To avoid the PCD
growth, they performed the growth in multiple runs, in which the run was stopped and
crystal was taken out to remove the polycrystalline diamond that was formed during a
deposition run. After PCD removable the run was resumed and then the growth was
continued again. This careful control of growth was also believed helpful to prevent
cracks. The details of the process were not provided.
An example of a diamond produced by the above technique is shown in Figure 2.16
(a). The large crystal was cut from a 13.5 carat rough diamond block grown at around 50
μm/h in the absence of impurities other than hydrogen. Optical microscopy revealed that
the CVD diamond is clear and relatively free of inclusions and cracks. There are no visible
layers (or growth interfaces) and striations from cross-sectional view within the thickness
above 5 mm. The as-grown, 2.3-carat, brilliant cut diamond was color designated as near
colorless. It also has high optical quality and clarity without visible layers.
The results showed the growth morphology and diamond color strongly depend on
temperature. Additionally, the results showed smooth and dark brown diamond substrates

54

at 1000 – 1100°C brown at 1100 – 1200°C, and tinted yellow at 1200 – 1220°C. Step
flow-type with pyramid-like octahedra tinted yellow formed at 1230 – 1400°C, spherical
black diamond-like carbon was produced below 1000°C, and twinned or polycrystalline
diamond resulted at temperatures over 1300°C.

Figure 2.16 Carnegie Institution’s Geophysical Laboratory diamond is grown at high
growth rates with and without nitrogen [14], [87]. (a) SCD plate, (b) 0.2 – 2.2 carat
brilliantly cut and polished nitrogen doped light brown colored and near colorless SCD
substrate, (c) examples of as-grown SCD substrates after annealing treatment and (d) in
image (A) 2.4 carat single-crystal CVD diamond (B, C, D) 2.3 carat diamond after cut and
polished [88].

The temperature range for smooth, yellow-tint growth is a very narrow 1200 – 1220°C.
Figure 2.16 (a) and Figure 2.16 (c) shows a photograph of an as-grown CVD polished
55

diamond. It was grown with at a growth rate of 58 µm/h at 1200°C. The improved growth
techniques for producing colorless SCD have been summarized in Table 2.3.

Table 2. 3 Carnegie Institution of Washington diamond growth parameter reported
Carnegie Institution of Washington, U.S.A.
Growth
parameter
Reactor type
Substrate seeds
Substrate pretreatment

Description

Reference

ASTeX/Seki reactor
HPHT type Ib (9 mm x 9 mm) and synthesized
SCD with (100) surfaces
cleaned ultrasonically with acetone.

[14]
[13], [14]
[14]

Pressure

5-20% CH4/H2, 0.2-10% O2/CH4. High purity
methane (99.9995%) and 0.2 – 5% N2/CH4
150 – 300 Torr

[14], [88], [89]

Input power

3-5 kW

[14], [88], [89]

Discharge power
density
Substrate
temperature
Growth rate

50 – 100 W/cm3

[88]

Gas chemistry

1000 °C to 1500 °C
40 – 100 μm/h.

[14]

[14], [88], [89]
[90]
[79][90]

2.7.2 Diamond Research Center, AIST, and Osaka University, Japan
The Diamond Research Center of Japan National Institute of Advanced Industrial
Science and Technology (AIST) from Osaka University in Japan has been working on the
homoepitaxial growth of diamond during the past 14 years [10], [11], [91]–[95] [96]–[100].
They have achieved significant progress in the SCD growth process, which is
summarized here in two groups, (1) high rate repetitive SCD growth, and (2) effect of the
Molybdenum holder on the diamond growth process.

56

In order to grow large area CVD SCD substrates with the help of the high growth rate,
the AIST research group implemented a method called, (1) high rate repetitive SCD
growth. In this approach consists of adding a multi-step growth process on the same
substrate to grow very thick SCD substrates. They reported high growth rates which
ranged from 30 to 150 μm/h at pressures up to 220 Torr with nitrogen additions of over
0.1% [10], [11], [91]. In this approach it was necessary to interrupt the growth within a
limited time and then clean the substrate holder to avoid the build-up of the polycrystalline
diamond on the substrate holder. Figure 2.17 and Figure 2.18 shows large diamond
crystals with thicknesses of 8.7 – 10 mm and weight of 3.4 – 4.5 carat were successfully
grown on a 3.5 mm x 3.5 mm x 0.5 mm diamond seed after 24 separate deposition steps
at 68 µm/h for a total of 150 hours. The growth temperature was 1130 °C, and the
pressure was 24 kPa (180 Torr). The flow rate of nitrogen was 0.6 sccm (0.12% N 2/ H2)
[10]. The high rate SCD growth was performed using a standard [11], [91] 2.45 GHz, AX5250 5 kW-MPCVD system produced by Seki Technotron Corp. The specific
experimental conditions are summarized in the following table:

57

Table 2. 4 Growth parameters for SCD by Diamond Research Center, Japan
Diamond Research Center, AIST, and Osaka University, Japan
Growth
Description
Reference
parameter
Reactor type
Seki Technotron Corp
[14]
Substrate seeds
HPHT type Ib (3 mm x 3 mm x 0.5 mm)
[10], [11]
Substrate preHydrogen etching for 30 minutes
[10], [11]
treatment
500 sccm H2 (6N) / 60 sccm CH4 (6N) / 0.6-1.8
Gas chemistry
sccm N2 (4N) (12% CH4/ H2, 0.12-0.36% N2/
[10], [11], [91]
H2).
[10], [11],[14],
Pressure
130 - 220 Torr.
[91]
[10], [11],[14],
Input power
1 – 3.7 kW.
[91] [94]
Discharge power
[10], [11],[14],
50 – 150 W/cm3
density
[91] [94]
Substrate
[10], [11],[14],
1000°C – 1300°C and 1130°C
temperature
[91] [94]
30 – 100 μm/h. (was estimated from the net
Growth rate
[11], [91]
weight gain)

Figure 2.17 A 4.65 carat, 10 mm thick SCD substrate grown with 24 deposition runs [10]
58

The high growth rate was enhanced due to the nitrogen on the gas feed. Even though
the side surface was covered with polycrystalline diamond, the top surface of the
synthesized SCD maintained smooth and flat surface morphology. By investigating the
Raman spectra over a cross section that was cut from one large synthesized SCD crystal,
the thick diamond crystal was showed to be of high quality (FWHM < 1.85 cm -1) [92]. Note
in Figures 2.17 and 2.18; that the grown SCD area is significantly smaller than the original
HPHT seed due to the thick polycrystalline rim around the SCD grown surface. This
indicates that the thicker the grown SCD crystal, the smaller is the area of the top surface.

Figure 2.18 Large CVD diamonds grown by multi-steps high growth rate synthesis
process. (a) 10 mm thick, 4.7 carat CVD diamond, (b) 9.6 mm thick, 3.5 carat CVD
diamond, (c) 8.7 mm thick, 4.4 carat CVD diamond [11].

59

In the research topic (2) numerical and experimental studies of several experimental
and numerical simulation data to show the distribution of gas and electric field intensity
within the conventional reactor configuration [101].
For the research topic (2) effect of the Molybdenum holder on the diamond growth
process, a new molybdenum substrate holder designs were fabricated to increase the
growth rate over time [10], [102]. The designs also were simulated to demonstrate the
effect of different reactor variables, such as substrate temperature, gas flow distribution,
and electric field intensity, over time [103], [104]. The schematics of the newly designed
substrate holders together with corresponding results of the numerical calculation of
plasma, the pictures of hydrogen plasma and the observed diamond growth rates are
shown in Figure 2.19. The numerical calculation showed that the magnitude of the
electrical field around the substrate increases as the radius decreases.

60

Figure 2.19 Schematics of newly designed substrate holder together with corresponding
results of simulation of plasma, pictures of hydrogen plasma and observed diamond
growth rates [102].

As a result, the diamond growth rate of the 8-mm holder was 30 µm/h, and for the
holder with 36 mm the growth rate was 5 µm/h. However, after optimizing the 8-mm
holder’s shape, they obtained a growth rate of 50 µm/h. The configurations of these
substrate holders were previously called an open holder and an enclosed holder [10].
61

Figure 2.20 shows a schematic illustration of two types of the substrate holders. For
enclosed type holder, "d" is defined as the final depth of diamond relative to the top
surface of the holder after the growth. As since the thickness of diamond increases during
the growth, "d" is defined as an average depth of diamond during the growth.

Figure 2.20 Schematic illustration of ‘‘enclosed type’’ and ‘‘open type’’ holders. The final
depth of diamond after the growth “d” is indicated [10].

They observed that the substrate holder configuration plays a significant role in the
quality and the growth rate. The substrates grown in an enclosed holder displayed a
smooth surface as compared to those grown in the open holders as shown in Figure 2.21.
After the first run, the grown diamond substrates by the open type holder indicate the
promotion of edge growth. In contrast, the grown diamond substrates by the enclosed
type holder show macroscopically flat surface morphology, and there is no promotion of
growth at the rim part compared with those by open type holder.

62

Figure 2.21 Optical microscope images of grown diamond after 1st growth (left) and
repetition of growth (right) for open and enclosed type holder. Thicknesses of grown
diamonds (t) are indicated in the pictures. A dashed circle indicates a crack formed by
repetition of high rate growth [10].

After the 2nd growth for open type holder, though relatively high growth rate (87 µm/h)
is obtained, a crack appeared in the center part, and no additional growth cannot be
applied without polishing the surface. In contrast, by using enclosed type holder, the
surface morphology is dramatically improved, and smooth and flat surface morphology
without growth hillocks or no epitaxial crystallites was obtained even after the 5th growth.
See Figure 2.21.
In order to increase the single crystal diamond surface area, one additional approach
was reported from AIST group; it is the production of diamond plates by a lift-off process.

63

As a standard procedure, the fabrication of SCD plates involves a conventional laser
cutting method, in which the grown substrate is separated from the HPHT seed. However,
during the laser cutting method and polishing, several hundreds of microns of diamond
are lost during the cutting process [12]. Thus a lift-off process using ion implantation was
developed to produce large and thick SCD plates and reduce material losses [94], [105]–
[110].

Figure 2.22 Lift-off process with ion implantation [108].

The steps of the lift-off process is shown in Figure 2.22. The lift-off process consists
of a high energy ion (3 MeV) beam being injected into the top surface of the seed
substrate, resulting in the formation of a graphitic layer beneath the top surface (first two
steps are shown on Figure 2.22). The ions are implanted with a shot of 2 x 1016 ions/cm2.
After this, an SCD layer was grown on the top surface of the substrate by using the high
growth rate diamond method described in the section above (third image from left in
Figure 2.22). The next step is etching away the graphitic layer using a non-contact
electrochemical etching method [111]. The grown SCD plate was obtained as a
64

freestanding wafer (right-hand image in Figure 2.22). Due to the thickness of the graphitic
layer and depth of this layer from the top surface are substantially small, a few
micrometers, the loss of material from the seed crystal is always negligible.

Figure 2.23 Photograph of freestanding single-crystal CVD diamond plates produced by
a lift- off process using ion implantation [94].

Figure 2.23 shows a photograph of two plates, which were produced by the lift-off
process using ion implantation. The first plate is a nitrogen-doped diamond (left hand on
the Figure 2.23) and the second plate is an undoped single crystal diamond plate. The
nitrogen-doped CVD diamond grown four-times at approximately at a growth rate of 34
µm/h and was successfully separated from the seed to produce a plate to a thickness of
470 µm. The undoped CVD plate also growth four times at an average growth rate of 7.8
µm/h. Compared with the nitrogen-doped diamond which is brown, the undoped diamond
is colorless. The plates were successfully separated from the HPHT seed by the lift-off
process. However, there is still a size limitation for the CVD diamond plates produced.
65

The size was limited by the size of the original seed. Then they tested a new process to
fabricate larger CVD diamond plates by combining the lift-off process and side surface
growth [94]. The process steps are shown in Figure 2.24.

Figure 2.24 Schematic illustration of a crystal enlarging process by a combination of liftoff process and side-surface growth [94].

An HPHT diamond seed with the size of 10 mm x 10 mm was used to assemble a
single crystal CVD diamond plate with the same cross section and a thickness of 0.4 mm
by the lift-off process (steps 1 and 2, Figure 2.24). The assembled plate was then grown
to a thickness of 3.7 mm with the multi-step high growth rate diamond process with 0.12%
nitrogen addition (step 3, Figure 2.24). The resultant SCD was laser cut and polished to
create side (100) faces. A diamond layer of a thickness of 3 - 4 mm was grown on three
of the faces (step 4, Figure 2.24). This produced a half-inch (12.6 mm x 13.3 mm x 3.7
mm) SCD. This diamond was laser cut to form two pieces of a (001) seed plate.

66

One of the plates was polished and used as a seed substrate for the lift-off process
(step 5, Figure 2.24). Thick diamond of 0.2 mm was grown on the seed by the high-rate
growth of 32 µm/h with nitrogen addition, and the film was successfully separated from
the seed. Figures 2.25 (a) and (b) shows photographs of the polished seed plate and a
separated half-inch single-crystal diamond plate. The results indicated the combination
of both processes is effective for enlarging the area of single-crystal CVD diamond plates.
Also, the enlarged seed plate can be used to produce large undoped single- crystal CVD
diamond plates.
Further, the AIST group introduced the so-called “mosaic wafers” process. This
method was developed to fabricate 1-inch size diamond wafer [108]–[110]. They
introduced an efficient way to build a mosaic wafer with smooth and almost invisible
boundaries. The key point of the technique is the use of same single seed substrates
which are “cloned” from the same “mother” substrate using lift-off process. The final
diamond wafer made of those clones is referred as “tiled clones” [109].
The method to produce clones and tiled clones is shown in Figure 2.26. By using the
lift-off process with the ion implantation method described above, “clones,” i.e. thin SCD
plates, were obtained. Each “clone” SCD plate had similar crystallographic characteristics
over the large cross section area of the seed substrate. For example, in steps (1 – 4) of
Figure 2.26, four clone substrate plates are grown from one seed substrate with a size of
about 10 mm x 10 mm.

67

Figure 2. 25 (a) Photograph of a half-inch seed plate obtained by cutting and polishing
the diamond in Fig. 5. (b) Photograph of a half-inch nitrogen-doped single-crystal CVD
diamond plate produced under high growth rate (32 µm/h) from the seed plate by a lift-off
process using ion implantation. (c) Polarized light microscope image of the half-inch
single-crystal CVD diamond plate [94].

These clones were aligned manually with each other to within less than 500 μm, to
make a tiled clone (step 5 in Figure 2.26). The diamond layers were grown on both sides
of the tiled clone wafer. A free-standing clone of the tiled clone can be produced by using
lift-off process again (step 6 in Figure 2.26).

68

Figure 2. 26 Method steps to produce clones and tiled clones [109].

An image of a typical free-standing clone of a tiled clone is shown in Figure 2.27. The
arrows in Figure 2.27 indicate the locations of the boundaries between the connected
clones. Each connected clone was about 10mm x 10mm, which is the same as the seed
substrate, and gave a free-standing wafer around 20mm x 20mm. The shape of the tiled
clones was almost square. From the Figure 2.27, one can see the boundaries between
the connected clones are almost invisible. In addition, the fabricated wafer has no obvious
cracks. The results show that the use of cloned substrates is very efficient in enlarging
diamond plate area. However, increasing the number of the constituent diamond plates
can cause degradation in quality and cracks on the diamond surface. Thus, in order to
further fabricate high-quality large area SCD plates, enlargement of the SCD substrate
still appears to be very important.

69

Figure 2. 27 An image of typical tiled clones [109].

2.7.3 LIMHP/CNRS, Paris, France – Recent Results
The diamond group of Laboratoire des Sciences des Procédés et des Matériaux,
France has explored the SCD growth process both scientifically and technologically,
regarding size, purity and crystalline quality of the diamond. Their contributions have been
in diamond growth experiments and diamond growth simulation [9], [42], [49], [58], [63],
[65], [112]–[114]. The main research focus of the France group can be classified into four
research topics: (1) effect of growth parameters specifically for the homoepitaxial
deposition of high-quality thick diamond films [49], [58], [63], [115]–[117]. (2) Microwave
engineering plasma modeling; here, the influence of the deposition process parameters
on the production rate of the growth species and CVD diamond growth kinetics, and
different designs of plasma reactors [9], [118] [119]. For large surface area homoepitaxial
CVD diamond growth, (3) the influence of the presence of multiple growth sectors on the
substrate surface and geometric modeling of CVD diamond crystal growth was
70

investigated [112], [113], [120]. Most recently they investigated (4) the reduction of
dislocations by using a self-assembled metallic mask and by the growth of diamond on
an HPHT diamond substrate hollowed out by a large hole [121], [122].
In research topic (1) homoepitaxial thick diamond substrates were grown by a
microwave plasma CVD at high microwave power densities at temperatures ranging from
800 °C to 950 °C and with nitrogen addition [49], [58], [115], [117]. It was shown that the
growth rate could be increased by factors of up to 2.5 by adding small concentrations (2
to 10 ppm) of nitrogen to the gas phase. The specific experimental conditions are
summarized in the following table:

Table 2. 5 LIMHP/CNRS, Paris, France growth parameter reported
LIMHP/CNRS, Paris, France
Growth
parameter
Reactor type
Substrate
seeds
Substrate pretreatment
Gas chemistry
Pressure
Input power
Discharge
power density
Substrate
temperature
Growth rate

Description

Reference

LIMHP reactor

[9], [118] [119].

HPHT type Ib (3 mm x 3 mm x 1.5 mm)

[56]

Wet-chemical cleaning and then a 2-hour H2/O2
plasma etching of the HPHT substrates.
purified hydrogen (9N), 2-7% highly pure
methane (6N) and nitrogen. Nitrogen
concentration was varied between 0 and 200
ppm.
50 - 300 mbar
2 – 4 kW.

films [49], [58],
[63], [115]–[117]

60 – 130 W/cm3

[9], [118] [119]

800°C – 1000°C

[49], [58], [63],
[115]–[117]
[49], [58], [63],
[115]–[117]

5 – 55 μm/h

71

[49], [58], [63],
[115]–[117]
[9], [118] [119]
[9], [118] [119]

They claim that the addition of nitrogen as low as a few ppm in the gas phase resulted
in a strong increase in the growth rate. At a pressure of 220 mbar (165 Torr), 10 ppm of
N2 improves growth rates by more than a factor of 2. This is shown in Figure 2.28. This
low amount of nitrogen enabled the growth of uncoloured “optical-grade” free-standing
CVD diamond film. It was reported that a maximum deposition rate of 55 μm/h was
achieved by introducing 200 ppm of nitrogen at 875°C. Before polishing grown diamond
substrates, surface morphology is shown in Figure 2.29. The images reveal that the
nitrogen incorporation on the growth process drastically deteriorates the morphology of
the diamond films. When the diamond was grown without intentional nitrogen addition
(Figure 2.29 (a)) the surface was very smooth and rough morphologies without no nonepitaxial crystallites, were obtained when the amount of added nitrogen was not greater
than 2 ppm (Figure 2.29 (b), (c) and (d)). Thus, the addition of N2 usually prevents the
formation of twins on (100) surfaces allowing extending the maximum deposition
temperature to a higher value but at the expense of crystal purity. The acceptable
temperature range is relatively narrow and there is only little room for growth rate
improvement using this parameter [56].

72

Figure 2.28 Diamond growth rate as a function of nitrogen addition in the gas phase
(MWPD = 95 W/cm3). Optical and PL images obtained under UV light for an undoped
free-standing CVD diamond crystal and a 2 ppm N2 doped [56].

It was also reported that going to high a temperature under high purity conditions
plagues the surface morphologies for single crystal diamond. Above 900°C, we observe
a roughening of the crystal surface due to the formation of typical pyramidal defects
(Figure 2.30 (a), 920 C, 7 µm/h) while at reasonably lower temperature the surface
remains smooth (Figure 2.30 (b), 870°C, 6 µm/h). At high-temperature twinning is
promoted on existing defects such as screw dislocations emerging at the surface.
Moreover, a favorable ratio of growth rates in (100) and (111) directions (defining the α
parameter) ensures that once formed twins will survive and overgrow [56].

73

Figure 2.29 Optical images of the samples grown with (a) no intentional N2 addition, (b) 2
ppm of N2 (c) 6 ppm of N2 and (d) 10 ppm of N2 [63].

The high-quality free-standing CVD diamond plates are shown in Figure 2.31. From a
macroscopic point of view, all the crystals have very good optical quality. They are all very
clear and relatively free of inclusions. The black corners of the samples (2) in Figure 2.31
indicate that the (111) triangular corners were more defective by nitrogen than the (100)faces. Thus the (100) top surfaces are virtually free of defects regardless of the amount
of incorporated nitrogen.

74

Figure 2. 30 Effect of growth temperature. Confocal laser microscopy images of 160 µm
thick CVD films grown at (a) 920 ◦C, 7 µm/h, (b) 870 ◦C, 6 µm/h doped [56].

Additionally, the methane concentration was varied while the power density of the
substrate temperature was held constant, i.e. 95 W/cm 3 and 850°C respectively. The
result showed a growth rate that increases from 2 to 16 µm/h as the concentration of
methane increases from 2% to 7.2%. This behavior is attributed to the variation of the
concentration of CH3 and atomic hydrogen at the diamond surface [49], [58]. The
morphology of the grown diamonds is shown in Figure 2.32. With 2% of CH4, the film
surface is very rough exhibiting a large number of etch pits. This could be a consequence
of a high etch rate by atomic hydrogen during growth at such high microwave power
density and a low percentage of CH4. The etch-pits progressively disappeared and the
surface improved up to 4% of methane. The films were grown at 6% and 7.2% of methane
shown good quality with no no-epitaxial crystallites, but step bunching could be observed
in some parts of the samples (Figure 2.32 (c) and (d)) [58].

75

Figure 2.31 Optical image of the free-standing CVD diamond samples obtained after laser
cutting and polishing and grown with (1) no intentional N2 addition, (2) 2 ppm of N2, (3) 4
ppm of N2 in the gas phase [63].

In research topic (2) microwave engineering plasma modeling; the influence of the
deposition process parameters on the production rate of the growth species and CVD
diamond growth kinetics, and different designs of plasma reactors have been studied.
Using their reactor geometry, they have developed plasma models to compute the [H]
and [CH3] in the gas phase as a function of the deposition parameters, i.e. pressure,
microwave power and methane concentration [9], [118], [119]. Their investigations show
that the use of high microwave power density (MWPD) plasmas are necessary to promote
atomic hydrogen concentrations that are high enough to ensure the deposition of high
purity diamond films at large growth rates. MWPD, electron density (Ne), electron
temperature (Te), gas temperature (Tg) and the atomic hydrogen [H] density has been
modeled extensively for a pure hydrogen discharge. Figure 2.33 shows the evolution of
the distribution of these quantities for a working pressure that ranges from 50 to 300 mbar.

76

Figure 2.32 Optical images of the diamond films grown at 95 W/cm 3 and 850 °C with (a)
2% CH4, (b) 4% CH4, (c) 6% CH4 and (d) 7.2% CH4 [58].

The microwave power was adjusted at each pressure so as to maintain a constant
plasma volume [9]. The result revealed the importance of working at high pressure in
order to obtain simultaneously high growth rates and high film quality. Also, the
importance of the substrate holder to provide precise control of the surface temperature,
which is a challenging at high pressure.

77

Figure 2. 33 From left to right, spatial distributions of microwave power density (MWPD),
electron density (Ne), electron temperature (Te), gas temperature (Tg) and atomic
hydrogen density ([H]) for a pure hydrogen plasma ignited inside the LIMHP reactor [9].

In research topic (3) they investigated the influence of the presence of multiple growth
sectors on the substrate surface and geometric modeling of CVD diamond crystals [112],
[113], [120]. In 2008 a geometric model of the growth of any cubic lattice material was
introduced by F. Silva et al [120]. The model showed a complete mapping of the final
morphology state of growing diamond, as a function of the growth rates of the crystalline
planes: (100), (111), (110), and (113). This geometric model was subsequently extended
with experimental results [56]. The results reported indicated the dependence of the
78

growth parameters on the diamond crystal faces. The finding showed that (113) faces
usually have the lowest growth rate and rough surfaces. The (111) face have high growth
dislocations whereas (100) and (110) faces have a higher average growth rate and low
densities dislocations.

Figure 2. 34 Schematics of the initial diamond substrate before polishing (a) and of
substrates polished into a pyramidal-shape (b) type A, 20° {100}-misoriented, (c) type B,
20° {110}-misoriented [114].

In an attempt to solve those issues, the growth on pyramidal-shape substrates having
different angles and orientations was carried out. The schematic of the initial diamond
substrate and the substrate into a pyramidal-shape are shown in Figure 2.34. The growth
process was performed at substrate temperature of 850 °C with 5% methane
concentration at a power density of 100 W/cm3. The HPHT substrate has both their lateral
and top faces oriented along the (100) directions. They were then polished into a
pyramidal-shape for which the lateral sides of the pyramid were inclined by an off-angle
of 20°, 30° or 40° along either the (100) directions.
79

Figure 2. 35. Optical images and 3D representation of the sample grown onto 20° {100}misoriented pyramidal-shape substrates after several growth interruptions. The total
thickness of the CVD layer is (a) 90 μm, (b) 270 μm, (c) 500 μm [114].

The grown substrates show the different diamond faces morphologies as well the
lateral growth in (110), (111), and (113) after the first growth run (90 µm) as seen in Figure
2.35 (a). It was also demonstrated the high growth rate of the lateral growth was led by
the (110) face. After 500 µm thickness, a partial overgrowth was observed in the samples.
The minimum thickness required for complete disappearance of these faces reaching up
to 2.2 mm. As was expected in the central part the results showed high etch-pit density.
The other part of the crystal around the central square shows a reduced etch-pit density
[114].
Following up these results, research topic (4) was the reduction of dislocations by
using the self-assembled metallic mask and by the growth of diamond on an HPHT
80

diamond substrate hollowed out by a large hole. First, a new growth strategy was
introduced aiming at preventing threading dislocations from propagating into CVD
diamond layers. It was based on the selective masking of existing defects revealed at the
surface of the substrates by Pt nanoparticles. The procedure was used to selectively
mask substrate defects with metallic nanoparticles in an attempt to decrease dislocation
densities. See Figure 2.36. Step 1 aims at revealing extended defects at the crystal
surface by an adapted etching treatment. Step 2 is the coating of the surface by a thin
CVD platinum film. Step 3 is a thermal treatment so that nanoparticles self-assemble to
the etch-pits on the surface. Step 4 is the PACVD diamond overgrowth to embed Pt
particles. Step 5 is a final etching treatment to reveal and count extended defects. The
full process can be repeated several times to improve its efficiency as illustrated by the
red arrow in Figure 2.36.

Figure 2.36 Schematic of the complete procedure of self-assembled metallic masks [121]
The experiments carried out consist of three successive growth runs with a thickness
of about 80μm. A Test sample was etched and masked with metallic nanoparticles at
81

each step following the procedure described in Figure 3.36. Another reference sample
was etched but not masked coated with Pt. Finally, a short etching run aiming at revealing
dislocations was performed leading to the appearance of etch-pits at the surface. The
schematic representation of the experiments is shown in Figure 2.37 (a) and (b) and the
results in Figure 2.37 (c) and (d). Results revealed that in the reference sample the defect
density increased with every growth layers up to ~ 1 x 107 cm-2. On the other hand, after
the third growth run the test sample revealed a decrease in the defect density to ~ 106
cm-2. The values showed satisfactory results on the reduction of the defects. Although
the defect density is still high, it remains low as compared to HPHT substrates.

Figure 2. 37 Schematic representation of the experiments (a) Test sample, (b) Reference
sample. After each etching step and the resulting final morphology is shown in a (c) test
sample, (b) reference sample [121].

82

Lastly, in the research topic (4) Tallaire et al. [122] investigated the reduction of
dislocations by lateral growth over a macroscopic hole. Standard cubic-shape HPHT Ib
substrates were hollowed out to create a square-shaped through hole (Figure 2.38 (a))
with 2 × 2 mm² dimensions.

Figure 2. 38 A thick CVD diamond layer grown on an HPHT diamond substrate hollowed
out by a large hole at the (b) beginning, (c) before, and (d) after merging of the growth
fronts and disappearance of the hole [122].

The growth lasted for 112 h, resulting in a final thickness of 980 µm (i.e., a normal
growth rate of 8.75 µm/h). After that time, it is noticeable that the central hole completely
disappeared, leaving a top surface free of any un-epitaxial defects. The central part is
smooth with macro steps bunched preferentially along (110) directions at the edges as
seen in Figure 2.39 (a).
83

Figure 2. 39 Laser microscope images of the CVD film grown on an HPHT substrate with
a macro hole after plasma etching to reveal dislocations: full-size image showing the
underlying square substrate. (b), (c) zoomed into the regions above the substrate and
above the hole, respectively, (d) PL UV image of the freestanding CVD diamond plate.
(e) Birefringence image of the freestanding plate under cross-polarizers. The white
dashed square indicate the position of the edges of the removed substrate [122].

The defect density in the center was found to be around 2 × 103 cm−2, while it was at
least an order of magnitude higher in the region above the substrate (Figure 2.39 (b) and
(c)). The PL UV image where the position of the substrate is indicated by dashed white

84

lines shows only light blue luminescence in Figure 2.39 (d). The birefringence image
shows only limited birefringence with a low number of cross-shape dislocations. On the
other hand, the region outside the dotted square exhibits much stronger birefringence,
see Figure 2.39 (e).
In summary, the results indicate a tremendous improvement in the defect density on
the CVD diamond substrates. The hollowed diamond seed seems to be an excellent
approach to grow high-quality CVD diamond at low cost.

2.7.4 Gemological Institute, China University of Geosciences, China
G. Wu et al. [123] recently used a specially designed substrate holder with a circular
recess, in which the seed crystal was placed. By designing substrate holders with different
recess depths and a seed crystal with the different side surface, crystallographic
orientations determine the influence of the recess depths and the crystallographic
orientation of seed sides on the growth quality. They utilized a conventional open holder,
in which the seed is placed in a flat holder and an enclosed holder. They also used an
enclosed holder. This substrate holder was varied for four different recess depths, i.e. 0.5
mm, 1.0 mm, 1.5 mm and 2.0 mm. The schematic of the substrate holder is shown in
Figure 2.40. The experiments were performed on (100)-oriented 3.5 x 3.5mm x ~ 1.4 mm
HPHT type Ib seeds under the following experimental conditions: the pressure of 80 –
100 Torr, 3 – 5% methane concentration, 0.1 – 0.5 sccm of N2 and an input power in the
range of 1.8 – 2.2 kW. All samples were pre-treated in H2 plasma for 1 hours.
Figure 2.41 (a) shows an optical image of the 150-μm-thick diamond layer grown using
the open type substrate holder with the growth conditions of 850 – 900 °C, 82 Torr, and

85

a CH4/H2 ratio of 3% and 0.5 sccm of N2. The sample that was grown used the open type
substrate holder (d = 0 mm) showed a concave surface morphology with overgrowth at
the periphery compared to the flat surface of the center part. In contrast, they also grew
on a HPHT seed placed in enclosed holder. Figure 2.41(b) shows the 110 μm thick
diamond layer grown in the enclosed type substrate holder (d = 2 mm) under the same
conditions as for the experiments of Figure 2.41(a). The result shows an overall smooth
surface and less polycrystalline material at the periphery.

Figure 2. 40 The schematic diagrams of the open type substrate holders and enclosed
type substrate holders with different recess depths. (The yellow squares represent the
seed crystals placed on the substrate holder or in the recess) [123].
The partial results demonstrated that as the recess depth increases, the formation of
polycrystalline diamonds on the periphery is controlled effectively. The growth rate
versus recess depth was evaluated on the grown substrates; the results showed when

86

the recess depth increases, the growth rate of the samples decreases significantly,
especially without the addition of nitrogen, as seen in Figure 2.42 (a).

Figure 2. 41 (a) HPHT seed grown by using an open type substrate holder (d=0mm); (b)
HPHT seed grown by using an enclosed type substrate holder (d=2mm) [123].

The lateral surface and the top surface growth rate ratio were plotted versus recces
depth in Figure 2.42(b). The results showed that the lateral growth of the sample
decreased significantly as the recess increased.
Raman spectroscopy was performed at the periphery and center of the grown
samples as a function of the recess depth to evaluate the quality of the crystals. See
Figure 2.43. The crystalline quality of the center part varies little when d ≤ 1mm, and then
degrades significantly when d= 2 mm. The substrate holder obtains the best crystalline
quality at the center part with the Raman shift FWHM of 3.4 cm −1 with a recess depth of
d= 0.5 or 1mm, in which the top surface of the seed crystal is above or at the same level
as the surface of the substrate holder.
87

Figure 2. 42 (a) The influence of the recess depth on the main surface growth rate.
(b)The influence of the recess depth on the lateral surface and main surface growth rate
ratio. Insets: the cross-section schematic diagrams of the samples grown with different
recess depths [123].

When d= 2mm, the FWHM of the sample at the center part is 4.1 cm−1, while the
crystalline quality at the periphery is better than that at the center part. As the recess
depth increases to when d= 2 mm, the crystalline quality when d= 2mm, already
surpasses that at the center part. At d ≤ 1mm, the top surface of the seed crystal can
contact the plasma to some extent, and the plasma above the center part of the seed
crystal is quite uniform. Hence, a variation of the recess depth has little effect on the
crystalline quality of the center part of the diamond layer.

88

Figure 2. 43 The influence of the recess depth on the FWHM of the diamond layers at the
center part and the periphery [123].

In summary, this investigation showed that as the recess depth increases the
polycrystalline diamond growth was controlled and was reduced adequately. Samples
with a good crystalline quality were grown at low growth rates (3 µm/h). The best-grown
substrates considering the crystalline quality and growth rate were grown in substrate
holder with a recess depth of d= 1 mm.

2.7.5 Michigan State University, USA
The diamond growth research at Michigan State University has provided an extensive
understanding of CVD diamond growth since the late 1990s. The relevant research topics
in the MSU laboratories includes: (1) the development of microwave plasma reactor to
enhance the growth rate [16], [18], [73], (2) substrate holder configurations to improve

89

SCD quality [20], [86], [124], [125], and most recently (3) the devilment of SCD growth
process at high-pressure [12].
In research topic (1) three generations of microwave plasma cavity reactor (MPC)
were developed and called reactor A, B, and C. The first generation of the MPC, was
designed to be operated at pressures in the range of 60 – 100 Torr for polycrystalline
diamond growth [126], [127], [128], [129],[130] . The next generation design modified the
reactor to allow diamond growth at higher pressures, higher power densities and higher
growth rates. A cooling stage was added and PCD films were grown at 145 Torr [126],
[131]. As a relevant output, the polycrystalline films were grown at 7 µm/h with a power
density up to 43 W/cm3 [73]. This new design identified as here Reactor A improved the
growth rate, and increased the operating pressure to 180 Torr. An additional important
modification was the further improvement of the cooling stage and the substrate holder,
i.e. a reduction in the diameter, to enhance the operating pressure and to allow operation
at higher microwave power densities. The new modification MPCR was called reactor B
and was initially redesigned to operate at pressures up to 240 Torr. Initially, polycrystalline
diamond films were grown at 240 Torr with at 475 W/cm3 power densities at an
approximate growth rate of 21 µm/h [73]. Further experiments performed by Lu et al. [16]
demonstrated an increased growth rate in SCD substrates up to 45 µm/h at 240 Torr due
to additions of nitrogen into the gas fed and a higher methane concentration of 7%. The
crystal quality was determined by Raman spectroscopy and indicated low stress with
better quality that HPHT type Ib seeds.
Y. Gu et al. [18], [132] designed the last generation of MSU microwave plasma cavity,
called reactor C for the purpose of operating at higher: pressure, power densities and

90

growth rates. Experimental results showed safe and efficient operational conditions at
pressures from 240 to 300 Torr. A comparison of the SCD growth regime is shown in
Figure 2.44 for all three reactor designs. Reactor C presents significant advantages over
the reactor A and B. Figure 2.44 (a) displays evidence that the power density reaches to
1000 W/cm3 at 300 Torr. It is about twenty times as high as that for reactor A and two
times as high for reactor B. The growth rate also was enhanced. See Figure 2.44 (b). At
7% methane concentration in which reactor B reached to 45 um/h, the growth rate for
reactor C was 60 um/h. This fact confirmed the advantage of reactor C over the previous
designs in terms of high pressure operating conditions. All diamond substrates discussed
in this thesis investigation have been grown with reactor C. The design detail of reactor
C have been presented in detail in Chapter 3.

Figure 2. 44 Comparison of reactor A, B and C (a) absorbed power density versus
pressure and (b) growth rate versus pressure [18], [132].

91

In addition to the development of microwave plasma reactors Asmussen et al. [133]
also introduced a different molybdenum substrate holder designs for growing the SCD.
Initially, the shallow recess holder was used exclusively to constraint the diamond seed
and ensure a constant holding position during the growth process. Nad et al. [20] studied
the variation of the substrate holder geometry on the SCD growth and how the growth of
the PCD rim could be overcome. The first geometry explored was called open holder. In
an open holder, SCD growth experiments were used as a group of controlled benchmark
experiments. In this first geometry, the HPHT diamond seed is placed directly on a flat
molybdenum substrate holder as is shown in Figure 2.45 (a). During the growth process,
the top, edges and the side surfaces of the seed are directly exposed to the active
discharge and to the impressed microwave electric fields that exist in the substrate
discharge boundary interface.
The open holder experiments observations showed that during the growth process, SCD
is grown on the top surface of the seed substrate, Figure 2.45 (a).

Figure 2. 45 A cross-sectional view of SCD synthesis versus time in an open holder (a,
b, c). d) The synthesized diamond after 35.5h of deposition time [20].

92

Simultaneously PCD is also grown on the top edges and the sides of the seed. As
shown in Figure 2.45 (b), an additional PCD layer is also grown on the top of the
molybdenum holder. As the synthesis process proceeds versus time (Figure 2.45 (c)), the
PCD layers increase and a PCD rim completely encloses the SCD top surface edges.
Eventually, a PCD rim grows vertically upward, horizontally outward and radially inward
into the SCD top surface. As a result, decreased useful SCD surfaces were obtained. A
thick PCD rim as shown in Figure 2.45 (d) was formed completely around the top surface.

Figure 2. 46 Cross-sectional views of SCD growth process versus time in a pocket holder
[20].
In order to improve the localized control of the SCD synthesis process, a new holder
design geometry called pocket holder was proposed. In the pocket holder, the diamond
seed is placed within a recess, or a “well” that adjusted into the water-cooled molybdenum
93

substrate holder, as is shown in Figure 2.46 (a). The pocket was designed for the purpose
to protect the SCD seed from the intense microwave plasma and from any microwave
fields that may be concentrated on the hot edges of the substrate in the open holder
configuration. Successful SCD growth was demonstrated, under the pocket growth
conditions the diamond grew with a greatly reduced PCD rim or with a completely
eliminated PCD. Examples are displayed in Figure 2.47 (a) and (b) respectively.

Figure 2. 47 Optical microscopy images of as-grown SCD substrates with (a) reduced
PCD rim and (b) without a PCD rim [124].

The pocket holder effect was proven by the following experiment. One HPHT seed
was cut at an angle to the (100) growth direction. The angled seed allowed that one side
of the growing diamond to grow out of the pocket while the other edges remain into the
pocket. The left edge of the substrate shows a PCD rim in contrast to the right edge. This
difference is due to the fact that the left edge emerges out of the recess of the pocket

94

holder with increasing growth time before the other edges. Thus, that side is subjected to
the plasma and the intense microwave fields and a PCD forms. The effect of the rim on
the synthesized SCD plate is clearly visible in Figure 2.48 which shows a birefringence
images with an exposure time of 500 ms, 2000 ms, and 5000ms. The right edge of the
CVD SCD plate exhibits very low stress as compared to that in the left edge.

Figure 2. 48 Birefringence images of a 221 μm thick CVD SCD plate with exposure times
of (a) 500ms (b) 2000ms and (c) 5000ms [124].

In summary, by utilizing the pocket holder design the successful synthesis of CVD
SCD substrates without the formation of any PCD rim on the CVD grown SCD was
demonstrated. The reduction of the PCD rim in the synthesized SCD substrates leads to
stress-free high-quality CVD SCD plates.

95

Figure 2. 49 Photographs of the discharge above a 3.5 mm x 3.5 mm SCD substrate in a
1.5-inch molybdenum holder as the pressure is increased from 120 to 400 Torr. The
absorbed power is kept constant at 2100 W [12].

The recent results at MSU include exploring the growth of SCD at high operating
pressure up to 400 Torr. The first high-pressure exploration was developed by Muehle
[12] utilizing the highly filtered continuous microwave power supply in a reactor B design
to determine the power density up to 400 Torr. A set of experiments were performed at
constant microwave input power of 2100 W while the pressure increased. Figure 2.49
shows the photographs of the discharges between 120 and 400Torr. The estimated
plasma volumes and the calculated absorbed discharge power densities are plotted
versus pressure in Figure 2.50. The graph shows that the plasma volume is decreasing
for increasing process pressure while the absorbed input power level is held constant.
This results in an increase of the absorbed discharge power density with increasing
pressure. The discharge showed a volume of 7.6 cm3 at 120 Torr and 3.1cm3 to 400 Torr
for a constant absorbed power of 2100 W shown in Figure 2.50.

96

It was observed that the absorbed discharge power density had linear increase from
275.4 to 526.4 W/cm3 for moderate pressures from 120 to 300 Torr. While at higher
pressures up to 400 Torr the absorbed power density increases super linear to 671.3
W/cm3. These results show a significant increase in the absorbed power density at highpressures.
The second exploration, a reactor operational field map at the pressure from 120 Torr
to 400 Torr. The analyzing of the substrate temperature and the pressure as a function of
the absorbed power was presented. The changes in absorbed power densities when
increasing the overall absorbed input power for certain pressures were observed,
specifically for the pressures of 320, 360 and 400 Torr.

Figure 2. 50 Discharge power density and volume as a function of the pressure for the
discharges utilizing Pabs= 2100 W and 3% CH4 as shown in Figure 2.49. The
differentiation between the previously investigated moderate pressure regime and the
high pressure regime is indicated by the green dashed line [12].
97

The substrate temperature and pressure as a function of the absorbed power were
also studied. The results for the entire 120 to 400 Torr pressure regions are plotted in
Figure 2.51. It can be seen that all curves can be described with a linear approximation
with the higher substrate temperature at pressures > 300 Torr with absorbed power >
2500 W.

Figure 2. 51 Operational field map for Reactor B using a SCD substrate over the entire
pressure regime

Additional to the plasma analysis, a set of four growth experiments were performed
lasting 20 hours at 5% methane concentration at the pressure windows between 180 and
400 Torr. The experimental growth rate and weight gain as a function of the process
pressure were measured and are plotted in Figure 2.52. The data showed an almost
linear increase in the growth rate could be seen up to 380 Torr. The growth rate tripled
from 9.3 to 28.1 μm/h when the pressure increased from 180 to 380 Torr. When the

98

pressure is increased further from 380 Torr to 400 Torr, the growth rate flattens out. This
was attributed as the pressure increases the distance between the discharge and the
substrate also increases, that means that the region of high density [CH3] creation
decreased directly above the substrate [12]. The increase of the weight gain as function
of the pressure follows the growth rate behavior.

Figure 2.52 Linear growth rate and weight gain as a function of the process pressure and
comparing the growth rates to previously reported results by Lu et al. [16], [12].

The morphology of the grown SCD films after deposition without additional processing
can be seen in Figure 2.53. It was observed no PCD formation occurred throughout the
entire pressure regime. The experiments used a pocket holder design similar to the
design that already was optimized for rimless growth [20], [125]. The grown substrates

99

have an enlarged top SCD area of ~ 1.4 times that the original HPHT seed area. The top
SCD surfaces seem to not have major defects.
Thus, successful safe and efficient operating high-pressure conditions up to 400 Torr
for MSU reactor B have been found. The growth rate and substrate temperature showed
a linear dependence with the pressure. SCDs with a smooth top surface and with reduced
PCD rim were grown at 400 Torr by using a pocket holder. Overall, results indicated the
advantage of increasing the pressure to enable a rapid growth of SCD substrates.

Figure 2.53 Top surface of SCD films grown in the pressure range between 240 and
400Torr [12].
100

CHAPTER 3

DESCRIPTION OF THE EXPERIMENTAL EQUIPMENT AND
EXPERIMENTAL PROCEDURE

3.1 Introduction
This chapter presents a description of the microwave plasma assisted chemical vapor
deposition (MPACVD) reactor system. This reactor system was used in all the
experiments described in this thesis investigation. The first section begins with an
overview of the typical experimental microwave plasma cavity Reactor C system. The
second section, describes the associated sub-systems that comprise the classical
microwave reactor system. The internal parts of the experimental system and standard
operation are also discussed. The next section, describes in detail the pocket holder
configuration. Finally, a brief discussion of the main analytical instruments is presented.

3.2 Experimental Setup
A typical schematic of the entire microwave cavity plasma reactor C system is shown
in Figure 3.1, it illustrates the basic system design. The schematics below show the
important components of the system and subsystems. The experimental sub-systems
are described as follows: (1) microwave power delivery, (2) the process gas delivery, (3)
101

the vacuum pumping control, (4) the pressure control, (5) the air and water control, (6)
the exhaust line, (7) the optical and photographic diagnostics, (8) the computer control,
and (7) the microwave cavity reactor subsystems.

Figure 3. 1 Overall setup of MSU microwave plasma cavity Reactor C system [85].

102

3.2.1 Microwave Power Supply subsystem
Figure 3.2 (a) shows the key components of a microwave power supply processing
subsystem. The microwave power is generated in the power supply, transferred via
waveguides to a coupling probe and coupled into the material loaded applicator. In the
case of microwave plasma processing the “material load” consists of the cooling stage,
substrate holders, the substrate material, and the processing plasma. The continuously
variable 10 kW adjustable from 10 -100% microwave power is displayed in Figure 3.2 (b)
and is equipped with a microwave power supply of a frequency of 2450 MHz ±10 MHz
and a 15 kW CW S-band magnetron with a typical 68% efficiency. The output of the
magnetron power supply is connected to a circulator that isolates and protects the power
supply from any power that may be reflected from microwave reactor. The reflected power
is led to a water-cooled dummy load which is impedance matched.

Figure 3. 2 Components of standard microwave power supply subsystem
103

The absorbed power Pabs is Pinc - Pref, which is measured as the incident power, P inc,
minus the reflected power, Pref, and is one of the major input variables. Microwave power
measurements are made using dual-directional power couplers inserted in the waveguide
between the circulator and microwave reactor.

3.2.2 Process Gas Delivery Subsystem
The gases used in this thesis for the diamond growth process are: nitrogen, hydrogen,
and methane. The compressed feed gases had the following compositions: 99.9995%
purity (Total impurities < 5 ppm, N2 < 3 ppm) for Hydrogen, 99.999% purity (Total
impurities < 10ppm, N2 < 5 ppm) for Methane. Mass flow controllers (MFC) are devices
used to measure and control the flow of gases. The MKS Type 1179A general purpose
mass flow controller controls the processing gases at a particular range of flow rates. A
signal is sent to the control panel. The total flow rate ranges from tens to hundreds of
sccm (standard cubic centimeters per minute) in this diamond growth process. These
controlled flow rate feed gases are mixed and delivered to the discharge chamber.

3.2.3 Vacuum Pumping Control Subsystem
The pressure control subsystem is used to maintain the desired pressure in the
deposition chamber. During the process, the vacuum pumping system pumps the reacted
gases out of the deposition chamber continuously. In this thesis diamond CVD processes
operated in the pressure range of a 10 Torr to 240 Torr. The pressure was maintained by
using a mechanical roughing pump. A TRIVAC D 16 BCS two-stage rotary vane vacuum
pump was used. The pressure in the deposition chamber was maintained by an automatic

104

throttle valve located between the chamber and the pump. The pressure was measured
by a pressure gauge MKS Type 141A which is a variable capacitance sensor consisting
of rigidly attached capacitive electrodes located on the back or reference side of a metal
diaphragm. The reference side is permanently evacuated and sealed and thus makes the
pressure measurement independent of the gas type or composition. When pressure is
applied to the diaphragm; its deflection produces a change in the distance between the
electrodes and the diaphragm and a resultant capacitance change. The signal is sent to
a pressure controller in the main control panel. The control system automatically adjusts
the throttle valve to achieve the desired pressure in the deposition chamber.

3.2.4 Pressure Control Subsystem
The pressure control subsystem is used to maintain the desired pressure in the
deposition chamber. During the process, the vacuum pumping system will pump the
reacted gases out of the deposition chamber continuously. In this thesis investigation,
most of the diamond growth processes operated at the pressure of 240 Torr. The pressure
was maintained by using a mechanical roughing pump. In this research system, a
TRIVAC D 16 BCS two-stage rotary vane vacuum pump is used. The pressure in the
deposition chamber is maintained by an automatic throttle valve located between the
deposition chamber and the pump. The pressure is measured by a pressure gauge MKS
Type 141A. When pressure is applied to the diaphragm, its deflection produces a change
in the distance between the electrodes and the diaphragm and a resultant capacitance
change. The signal is sent to a pressure controller in the main control panel. The control

105

system automatically adjusts the throttle valve to achieve the desired pressure in the
deposition chamber.
The system leak rate is an important variable in CVD diamond growth. The pressure
changing rate (mTorr/hr) needs to be estimated based on the residual gas, N 2, from the
leaking of the vacuum system. The system leak rate is performed before of the diamond
growth process. After the system is pumped down the throttle valve is closed manually
by pressing close. The pressure before the throttle valve is closed is noted and the
pressure after approximately one hour is noted just before opening the throttle valve.
These values are namely as Pinitial and Pfinal respectively. After an “overnight pumping
down” the base pressure is ≈ 4.0 x 10-6 Torr. Since the leaking gas is air and air is mostly
nitrogen, we can estimate the residual nitrogen from the leaking of the vacuum system
as:

Leak _ rate(mTorr / h) 

Pfinal  Pinitial
time

Here the time is the number of hours which the throttle valve is closed, i.e. ≈ 1 hour.
For all experiments reported in this thesis, the N2 leak was between 0.1 and 2 mTorr/h.

3.2.4 Air and Water Cooling Control Subsystem
The cooling system used is a closed circuit with a minimum water inlet temperature of
19°C. Water recirculation systems are preferred since, apart from saving water, they help
to assure a high standard of coolant purity. The water used for cooling must not contain
any suspended matter which may block the coolant system. It must also be free of
corrosives which might attack the metal lines on the system. To prevent electrolysis, the

106

water must have high electrical resistance. The mineral content and the electrical
conductivity of the cooling water must be periodically checked and improved, if necessary.
The cooling water must also be free from all traces of greasy substances, as even a
thin film of grease may form a critical barrier on the cooler causing extreme anode
temperature. Greasy, oily, or metal particles in the water may be removed by repeated
flushing of the cooling ducts with a rinsing agent. The cleaning process should be
terminated with a flushing with distilled water. The cause for such greasy deposits will
usually be found somewhere in the cooling system, i.e. water pump, hose, etc. Also, the
entire system must be periodically cleaned by flushing and rinsing. The cooling water
supply in the microwave generator is described in Figure 3.3.
For air cooling, there are usually two air blowers used to cool the quartz bell jar through
an inlet in the cavity wall. An air blower is also placed beside the magnetron to cool the
power supply from outside when the growth process is running. An air fan is positioned
around the cavity to cool the external cavity walls.

107

Figure 3. 3 Cooling water supply

3.2.5 Exhaust Line Subsystem
The exhaust subsystem is one of the safety precautions. A nitrogen purge is used as
a dilution of the hydrogen-dominated exhaust gases at the exit of the mechanical roughing
pump. Usually, the flow rate of a nitrogen purge is 10 times of the total flow rate of
flammable output gases. The nitrogen dilution thus reduces the methane and hydrogen
exhaust gas concentration to a nonflammable level at the exit of the mechanical roughing
pump. The safety interlock system measures the pressure of the processing gases to
prevent any over pressurizing when hydrogen is present in the system.

108

3.2.6 Optical and Photographic Diagnostic Subsystem
The optical and photographic subsystem is composed of a pyrometer to measure the
substrate’s temperature, and a digital camera to capture images of the discharge and the
diamond seeds.
The portable optical emission pyrometer (IRCON Ultimax Infrared Thermometer UX
CL1; Figure 3.4 (1)) is used to measure at a fixed wavelength (“two-color”) of 0.96 μm.
The reading temperature range is between 600 - 3000 °C, while the emissivity during the
diamond growth process was 0.1. As shown in Figure 3.4, the pyrometer was positioned
at a high angle in front of the screened window with visibility through the cavity wall. The
pyrometer was set on a tripod at all times to keep the measurement results consistent.

109

Figure 3. 4 Optical and Photographic subsystem on microwave cavity reactor C. (1)
Pyrometer, (2) CANON camera.
The position and height of the tripod were not changed throughout all experiments.
Since the substrate position was below the screened window, and the thermal emission
was focused through the quartz bell-jar at an incident angle of approximately 65° to the
center of the sample. The measurements were usually taken multiple times, and the most

110

repeatable reading was recorded. Once the microwave input power was changed, a
stable substrate temperature reading was obtained after a 1 - 2 minutes lapse. The
position of the pyrometer, the incident angle, and the measured position of the sample
were held constant during the experiments. It was also understood that due to the intense
emission from the discharge, the measurement of the absolute value of the substrate
temperature by a pyrometer could not be accurate. Nevertheless, the measurements
presented here are repeatable and thus serve from run to run as a relative experimental
substrate temperature measurement that can be compared with previous works.
During the diamond growth process, a series of photographs were taken on (1) the
discharge hovering over the substrate and (2) the growing diamond surface as the
microwave input power varied. A CANON EOS 20D 8.2-megapixel semi-professional
digital single-lens reflex camera with 60-mm lens was used to capture the images at a
high angle through another screened window located adjacent to the pyrometer (Figure
3.4 (2)). The window was also just above the top of the bell jar. The camera was located
at a fixed position on a tripod outside of the reactor at an incident angle of approximately
75° to maintain a comparable fixed reference position. For comparison between the
images, all parameters were held constant for all experimental conditions - such as the
focal length was kept at 60 mm, and the camera was perfectly focused on the substrate
rather than the bell jar or cavity spyhole (screened window).

111

Figure 3. 5 Camera Arduino control map for microwave cavity reactor C [134].
To carefully control the plasma discharge and the diamond growth evolution, several
photographs were automatically taken every two minutes by an Arduino remote control
circuit; see Figure 3.5. The Canon EOS 20D camera has a port on the right side for a
remote switch with a type N3 terminal in which an appropriate pinout was connected. The
pinout used was a 2.5 mm (sub-mini) head with 3 cores ((1) ground, (2) focus, and (3)
shutter). One side of the pin was connected to the remote camera port, and the other was
connected to pin 8 on the Arduino circuit. The Arduino connects to 2 different channels
on the relay board to control the focus and the shutter of the camera. The relays are used
to isolate the camera circuit from that of the Arduino. A complete diagram of the circuit is
shown in Figure 3.6. By connecting the camera to an Arduino, all three contact sections
are shorted together, causing the camera shutter to be released. The core (1) is the
112

ground connector. Therefore, the core (1; ground) and (3; shutter) causes the camera to
trigger. The first digital image was captured 5 seconds after the power switch is turned
on, the next images are captured and stored every two minutes.

Figure 3. 6 (a) Diagram of the Arduino circuit and (b) photograph of the completed
operational Arduino circuit [135].

3.2.7 The Main Operations Windows
Two windows provide a system wide collection of tools for operating the system. The
Standard view panel is shown in Figure 3.7 and provides controls and indicators grouped
by function. The schematic view window is shown in Figure 3.8 and provides controls and
113

indicators organized according to the schematic block diagram view. In both views, you
can observe the current status of the machine and can access and adjust control setpoints for subsystems or components, far as the adjustment is allowed by the current
interlock or operation status of the reactor.

Figure 3. 7 The standard view panel operation window

114

Figure 3. 8 The schematic view panel operation window

In general, this window is used to display the state of the tool. Some control functions
(substrate load/unload and clear faults) are available directly from this window. Most other
control functions (such as adjusting the setting of power, vacuum pressure or mass flow)
are available by opening the “Job Panel” for a specific subsystem. As an introduction, the
Operations Window is divided as follows:
1. Upper Left Pane: Interlock Status
2. Lower Left Pane: Sensor Readings
3. Middle Panes: Subsystem settings and status.
4. Right Panes: System Functions
5. Bottom Pane: Automatic Recipe Control & Status
115

The upper left pane in the standard panel window displays the current state of system
interlocks and limits. Generally, the color-coding of these simulated LED indicators is as
follows: (1) Green: status is okay, ready for operation, (2) Yellow: status is not in operation
range, and (3) Red: fault status, out of range during operation.

3.2.8 Method of System Operation
This section provides information on the efficient and safe operation of the tunable
microwave cavity plasma reactor C as well as to gain insight on how process parameters
control the plasma discharge. The system is automated with computer control, and is
substantially equipped with in-situ diagnostics and safety interlocks. Those features
enable the user to control the current system state based on frequently observing the
plasma as well as the overall system data such as pressure, reflected power, substrate,
and system component temperatures during the process. The microwave power delivery
system is designed to efficiently couple the microwave energy to the processing plasma
for deposition and etching applications. The pressure and the microwave power can be
adjusted any time during the process, which influences, for example, the dimension of the
plasma discharge and the seed surface temperature. It is self-evident that there must be
an optimum for a given process; the discharge should not be too large to avoid contact
with the quartz dome.

116

3.3 Experimental Apparatus – MSU Reactor C
3.3.1 Overview
The reactor setup employs a microwave discharge and uses a vacuum chamber with
a rotary pump that runs continuously during the process. The pressure inside the chamber
is kept constant, by a throttle value, while all gases are fed into the chamber with desired
and steady flow rates. Unlike HPHT synthesis, which is typically carried out at pressures
in excess of 5 GPa (~50,000 atmospheres), diamond CVD is usually carried out at a
fraction of one atmosphere mainly for the sake of safety concerning the use of explosive
gas mixtures and because of the limits for plasma generation. The conditions of low
pressure (sub-atmospheric pressures) and low temperature for diamond CVD are actually
in favor of growth of graphite instead of the diamond according to the phase diagram of
carbon (Figure 2.2) [136], due to the fact that graphite is thermodynamically stable while
diamond is kinetically stable or so-called metastable.
Molybdenum is used as the substrate holder for homoepitaxy growth of SCD. The
microwave energy is introduced into the chamber to generate a discharge. The gas
molecules are dissociated to generate carbon atoms for deposition onto the substrate
surface. Microwave plasma assisted CVD (MPACVD), is currently the most widely used
technique for diamond deposition. The diamond substrate is heated by, the bombardment
by energetic plasma species, and the exothermic recombination of radicals, such as
atomic hydrogen, on the diamond growing surface.
The plasma formed typically has a ball or hemispherical shape and sits right on top
the substrate, which increases the efficiency of mass transport of diamond growing
radicals to the diamond surface for deposition.
117

3.3.2 Electromagnetic Focus
The cavity is a hollow metallic cylinder. Since the cylinder wall is electrically
conductive, the amplitude of the electric field parallel along the cylinder wall must be very
small. By applying these boundary conditions on the electromagnetic waves, many
characteristic transverse electric field patterns can occur in the hollow cylinder. Each field
pattern focuses the microwave power at defined locations inside the cavity.
Two transverse magnetic (TM, no magnetic field in the direction of radiation
propagation) modes that are useful for the purpose of establishing a localized plasma
discharge are shown in Figure 3.9.

Figure 3. 9 Electromagnetic fields, TM01n modes (n = 2 left, n = 3 right) [137].

118

The electromagnetic power distribution in the TM01 modes is constant in the annular
direction around the vertical axis of the cavity and has a center maximum in the radial
direction. The microwave radiation is coupled into the cylinder via an antenna (probe)
reaching a quarter wavelength through the top plate of the cylinder and the region of
concentrated power on the very bottom of the cavity is selected for the plasma formation.
The TM013 mode offers a larger distance between the near-field of the probe and the
discharge region. Figure 3.10 shows three discharges that can result from the TM 013
mode excitation.

Figure 3. 10 Three typical discharges (in green) that result from exciting the cylindrical
cavity applicator with the TM013 mode [137]. The red lines are the magnetic field patterns
and the blue lines are the electric field patterns.

119

The actual dimensions of the intensity field depend on the dimensions of the cylinder
(inner diameter and length) and the frequency of the microwave radiation. The cavity
length (short position) Ls and the penetration depth of the probe Lp have been calculated
for two cavity designs and are summarized in Yajun Gu’s PhD thesis [132] and in papers
[18], [138].
Note the theoretical dimensions given in Table3.1 are used as starting points for the
actual cavity to a diamond deposition process. These numbers could have varied by
several mm to optimize the process either way during or before diamond deposition and
are the same from one substrate holder geometry to another substrate holder geometry.
Table 3. 1 Theoretical TM013 and TM012 modes
Frequency in MHz
Cavity radius in cm
Mode
Calculated cavity height Ls in cm
Calculated cavity height Lp in cm

2450
8.90
TM013
TM012
21.6
14.4
8.2
8.2

915
21.75
TM013
TM012
60.2
40.1
3.1
3.1

All the experiments carried out in this thesis research employed Reactor C that used
a hybrid mode excitation as described in detail in reference [18], [132]. A hybrid TM01n
mode plus a TEM001 mode was excited. Then Ls was 16.20 cm

3.3.3 Description of the Classic Reactor
To use the cavity to drive a plasma discharge a microwave transparent quartz dome
is placed at the bottom of the cavity. This allows containing the plasma in a low-pressure
environment inside the quartz dome while the cavity is at atmospheric pressure. The
reactor design described below creates a hemisphere-shaped plasma in close contact
with the substrate and thus allows the coating of large substrate surfaces. For a 2.45 GHz
120

excitation, diamond films can uniformly be deposited over three to four inches diameter
(80 cm2) substrates, and when excited with 915 MHz the substrate deposition width can
be scaled to over eight inches and deposition areas can exceed 320 cm 2. Additionally,
this reactor concept is also operationally versatile. That is, it can be adjusted to deposit
diamond over a wide range of experimental conditions; i.e. it can deposit diamond on a
variety of substrates, at pressures of 0.5-300 Torr, with variable power levels of (1) 500
W – 6 kW at 2.45 GHz [18] and (2) 3 kW – 15 kW at 915 MHz [138] .

121

Figure 3. 11 Cross-sectional view of the MCPR [139].

122

Figure 3.11 shows a cross-sectional view of a typical microwave cavity plasma reactor
configured for operation in the thermally floating substrate mode because it is placed in a
quartz tube (17) located on the holder base plate. As shown, the cavity applicator sidewall
(1) is made of a 17.8 cm inside diameter cylindrical brass tube. This brass tube forms the
outer conducting shell of the cavity applicator and is electrically shorted to a water-cooled
baseplate assembly (2-4) and a water-cooled (21) sliding short (8) via finger stock (9).
Thus, the cylindrical volume bounded by the sliding short, the cavity applicator sidewall
and the base plate forms the cavity applicator electromagnetic excitation region. The
TM013 electromagnetic mode is excited inside the 17.8 cm diameter cavity applicator by
applying 2.45 GHz microwave power and length tuning the applicator to about 21.3 cm.
Once the microwave energy is coupled into the cavity and the reactor pressure is reduced
to approximately 4 Torr a discharge can be ignited inside the quartz dome with 500 –
1000 W of microwave energy.
It has been observed experimentally that the discharge loaded TM 013 mode will
produce an ellipsoid-like discharge (12) which is in good contact with the substrate (7).
Continuous wave (CW) microwave power, typically at 2.45 GHz, is coupled into the
cylindrical cavity applicator through a mechanically tunable coaxial excitation probe (11).
This probe is the center conductor of the coaxial waveguide (10), which is attached to the
center of the sliding short (8). The movement of the sliding short sliding changes the
applicator length, Ls, while the variation of the coaxial probe (10) position adjusts the
depth of penetration of the coaxial excitation probe, L p, inside the cavity.
Positions for discharge ignition are Ls = 21.3 cm and Lp = 3.0 cm for the classic reactor.
Ignition positions for Reactor C, are Ls = 16.2 cm and Lp = 3.0 cm. Then as the reactor

123

system and discharge are brought up to the desired steady-state diamond deposition
process conditions the applicator length, Ls, is iteratively adjusted to reduce the reflected
power and to achieve the desired process results such as deposition uniformity. The
probe depth, Lp, is kept constant for all experiments.

3.4 Diamond Seeds for the Growth Process
Commercially available high-pressure high temperature (HPHT) diamond single
crystal plates, type Ib (100), with typical sizes of 3.5 mm x 3.5 mm x 1.4 mm from
Sumitomo Electric Industries were used as seed substrates for homoepitaxial growth of
CVD diamond single crystals. For most SCD growth experiments presented in this
dissertation, the dimensions of the HPHT diamond seeds that were used were 3.5 mm x
3.5 mm and have a thickness size of ≈ 1.4 mm.
Note that the two biggest surface areas of the HPHT seeds were mechanically
polished, and the other four faces were not. Moreover, the most commonly used HPHT
seeds in this thesis were the 3.5 mm x 3.5 mm seeds listed in Table 3.2. These seeds
have a slight difference between the top and bottom areas dimensions. See Figure 3.12.
Although both surfaces were mechanically polished, the top surface was always used as
deposition surface in the experiments, and the seed was always placed with 3.5 mm x
3.5 mm face up in the pocket holder. HPHT seeds have a golden-yellow color because
the diamond contains 10 – 1. 0 ppm nitrogen in the crystal. They are classified as type Ib
diamond. These diamond seeds typically have defect densities of 104 – 105 cm-2, which
is considered high [140].

124

Table 3. 2 Measured dimensions on HPHT seed

Figure 3. 12 Optical microscope image of the side view of the 3.5 mm x 3.5 mm x 1.4 mm
HPHT seed.

3.4.1 The Diamond Seed Cleaning Procedure – Wet Chemistry
Prior to each deposition experiment the HPHT single crystal diamond seed was first
cleaned by a wet chemical cleaning procedure in order to remove any metallic, graphite
and organic material contamination. The cleaning procedure is described in the following
steps:
1. Nitric acid (40 mL) + Sulfuric acid (40 mL) mixed in a pyrex beaker, placed on the
125

heater (set to 300 °C) for 30 minutes. Then, rinse the sample in DI water.
2. Hydrochloric acid (40 mL) in a pyrex beaker, placed on the heater (set to 300°C)
for 20 minutes. Then, rinse the sample in DI water.
3. Ultrasonic cleaning: Ultrasonic bath cleaning with Acetone (30 mL) in a pyrex
beaker for 15 minutes.
4. Ultrasonic cleaning: Ultrasonic bath cleaning with Methanol (30 mL) in a pyrex
beaker for 15 minutes.
5. Final rinsing and drying: Rinse seed with DI water and blow with nitrogen to remove
water.
Once the cleaning procedure was completed, the substrate seed is placed in the
substrate holder and loaded into the reactor chamber.
After the diamond growth was finished and the grown diamond was taken out of the
reactor, it was cleaned following the same wet-chemical cleaning procedure which was
described in above in the following order:
1. Nitric acid (40 mL) + Sulfuric acid (40 mL) mixed in a pyrex beaker, placed on the
heater at 300 °C
2. The process above was repeated until the grown substrate was clean. This
cleaning varied from 1 – 3 hours depending on the final conditions of the substrate.
3. Ultrasonic cleaning: Ultrasonic bath cleaning with Methanol (30 mL) in a pyrex
beaker for 20 minutes.
4. Blow with nitrogen, and then place the cleaned diamond in a pantry box.
The cleaning process was intended to remove oils, any metallic, graphite and organic
material contamination. The substrate cleaning process should be done in an appropriate

126

laboratory facility. The acidic cleaning steps must be carried out in a fume hood following
relevant operational instructions. Before and after each experiment the quartz bell jar is
cleaned with ethanol to remove any graphite residues on the dome walls. Often the grown
SCD plate was removed from the seed by laser cutting. Then the deposition surface and
the bottom surface of the CVD diamond plate were mechanically polished, and then a
single crystal CVD diamond plate was fabricated for characterization.

3.5 The Pocket Holder
The holder dimensions play an important role in the deposition of large SCD
substrates. Many experiments have used the conventional open holder for CVD diamond
growth. The “open holder geometry” is defined here as a holder geometry where the
diamond seed is placed on a large flat open holder surface. This has been reported in
[20]. These experiments indicated that if one used a flat open holder design, thereby
exposing the diamond directly to the discharge and directly to the microwave
electromagnetic fields, a PCD rim is deposited and forms around the grown SCD material.
Reference [123] recently also has demonstrated a decrease in the diamond crystal quality
when an open holder configuration is used.
In this thesis investigation, different pocket holder designs have been used for the
diamond growth experiments. All pocket holders used in this thesis have been
manufactured from molybdenum obtained from Ed Fagan Inc. and made by the machine
shop in the Department of Physics at MSU.
The variable geometry of the pocket included: (1) different pockets holder depths, (2)
different pocket holder apertures, (3) specially designed pocket holders that were used in
127

the large diamond seed growth experiments, and (4) specially designed pocket holders
that were used in the diamond substrate re-growth experiments

3.5.1 Pocket Holder Preparation
In order to reuse the holder and start the growth process under the same conditions,
the holders had to be cleaned before each experimental run. Before every growth process
the molybdenum, substrate holders are cleaned according to the following suitable
cleaning procedure:
1.

The molybdenum substrate holders were sand blasted in order to ensure and
promote desirable growth process conditions.

2.

Ultrasonic cleaning of all molybdenum pieces with methanol

3.

Rinsing and nitrogen blowing; and

4.

Drying the Moly pieces in an oven at ~ 100 °C for approximately 20 minutes.

Thus, the sand blasting and the other cleaning steps removed all PCD that remained from
any earlier diamond growth run. Hence every new experimental run began under the
same starting conditions with the same clean substrate holder.

3.5.2 Pocket Substrate Holder Configuration
The pocket holder configuration and dimensions are an important component of the
diamond growth system and have a significant role in the diamond growth process. The
substrate holder consists of two separated parts. One is named the substrate holder and
the second is called the pocket holder. Both are shown in Figure 3.14 (b) (2) and (3)
respectively. The substrate holder (2) is a 64.7 mm diameter holder with holes drilled

128

around the circumference. See Figure 3.13. These holes allow the exhaust gases to flow
out of the system via the underlying quartz tube. This quartz tube directs the exhaust
gases to flow out of the system through 4 holes drilled into the cavity bottom conducting
short plate (4). A circular groove, concentric to the cooling stage holds a quartz tube (1)
shown in Figure 3.14 (b). The inner concentric part of the substrate holder has placed
directly above and in contact with the top of the cooling stage, see Figure 3.13 (2). The
substrate holder sits slightly separated at the top of the tube to allow for the vertical
expansion of the quartz tube. For all the diamond growth process reported here, the
substrate holder was held constant with the dimensions as indicated in Figure 3.13.

Figure 3. 13 Isometric drawing view of a 64.7 mm substrate holder.

129

Figure 3. 14 (a) A schematic side view of the microwave cavity plasma Reactor C with a
(b) zoom of the substrate holder setup.

130

The conventional pocket holder geometry has a 29.2 mm diameter top molybdenum
piece (3). This diameter fits in the inner diameter of the concentric part of the substrate
holder of 64.7 mm (2). Both pieces fit on and directly above the cooling stage (Figure
3.14 (b)). The pocket holder has a centered “pocket” which contains a step bottom. The
holder is shown in Figure 3.15 with the important dimensions indicated. The width of the
pocket is 6.0 mm, while the HPHT seed used in this investigation has a width of 3.5 mm
on the bottom side. The bottom step is needed to keep the seed centered inside the
pocket. The dimensions of the bottom step are 4.2 mm x 4.2 mm in the area and have a
height of 1.0 mm. The height of the upper step is 1.3 mm, and the depth of the pocket is
therefore 2.3 mm. The results of the grown diamonds using this conventional pocket are
presented in Chapter 4.
In addition to the conventional pocket holder shown in Figure 3.15, in this thesis
investigation, four different pocket holder geometries were used. The first pocket
geometry strategically varied the pocket depth while the pocket aperture was held
constant. The pocket depths were 2.0 mm, 2.6 mm and 2.9 mm. Here, the only modified
dimension is the height of the upper step, as is shown in Figure 3.16. The thickness of
the pocket holder with a depth of 2.9 mm might be slightly readjusted to allow the diamond
growth to take place within the appropriate substrate temperature window. The results of
the grown diamonds using this conventional pocket are presented in Chapter 5.

131

Figure 3. 15 Side cross sectional view of conventional pocket holder (unit: mm) that was
used in the experiments in Chapter 4.
132

Figure 3. 16 Drawings of the pocket holder for SCD growth at variable pocket depths (unit:
inches) that was used in the experiments in Chapter 5.

The second holder geometry varied the pocket holder aperture or width strategically.
In this new design, the pocket depth was held constant at 2.6 mm while the aperture
length, w, varied from 6.6 mm, to 7.0 mm and 7.6 mm. A drawing of this pocket holder is
shown in Figure 3.17. Also, the thickness of the holder is held constant at 3.2 mm. Note,

133

the bottom pocket was varied from 4.2 mm to 3.8 mm to avoid slight variations on the
seed position. These holders were used in the experiments that are described in Chapter
6.

Figure 3. 17 Side and top view of the pocket holder with a varied width (unit: mm)

The third pocket holder geometry was designed for the SCD re-growth experiments.
Here, a second growth step was added on top of the grown SCD substrates which usually
have a thickness of 1.5 mm and lateral growth area of 26 mm2. Therefore, the new pocket
holder design has a strategical depth of 3.6 mm and a pocket width of 8.2 mm as is shown
in Figure 3.18. The results of the grown diamonds using this conventional pocket are
presented in Chapter 6.
The fourth pocket holder was designed for the SCD growth process on large area
HPHT seeds. Here the depth of the bottom pocket and the width changed based on the
area of the seeds. An example of this pocket is shown in Figure 3.19. This holder is
designed for HPHT seed of 5.0 mm x 5.0 mm x 1.2 mm. In the event of larger seeds, just

134

the bottom pocket is modified and the pocket depth is held constant at 2.6 mm. The results
of the grown diamonds using this conventional pocket are presented in Chapter 7.

Figure 3. 18 Side view of the pocket holder used for SCD re-growth process (unit: mm)
that are described in Chapter 6.

3.6 System Startup
3.6.1 Introduction

Before starting a new experiment, the mechanical and electrical components of the
system are turned on before the process is started. At the start of each experimental run,
the system is pumped down over night, i.e. ~ 12 hours in order to ensure that all the air
inside the chamber is evacuated and there is no leak and residual gas left inside the
chamber. Accordingly, it is suitable to employ the following procedure described below in
Sections 3.6.2 - 3.6.6
135

3.6.2 The Diamond Seed Loading into the Microwave Cavity Reactor
Once the seed was cleaned, the system was prepared for loading the seed into the
cavity reactor and then to initiate the experimental run. Typically, the setup of cavity
Reactor C for loading the diamond seeds is shown in Figure 3.14 (b). The system is
comprised of a stainless-steel cooling stage, a quartz tube (1), three plastic O-rings, and
the molybdenum holders (2) and (3). The loading process is as follows:
1. The quartz bell jar is shown in Figure 3.10 (a) and the quartz tube shown in Figure
3.10 (b) (1) are cleaned with ethanol.
2. The substrate holder (3) with the diamond seed is carefully placed above, and a
good contact with the cooling stage and then the quartz bell jar is set into place
and during pump down the O-rings seal the vacuum.
3. The door of the vacuum chamber (not shown) is tightened, and then the vacuum
pump is turned on.
4. Click the “Rough Down” button on the computer software (open the throttle valve)
to allow the pumping down of the system.
5. The outer shell of the cavity is mechanically moved down slowly until is aligned
with eight screws on the upper side of the cooling base plate. Nuts are tightened.
6. Usually, the system is pumped down overnight (~ 12 hours) before the discharge
ignition.

136

3.6.3 Discharge ignition
After an ~12 hours of pumping down to about 5.0 x 10-6 Torr, the discharge ignition is
as follows:
1. Turn on all the gas inlet valves, i.e. nitrogen, hydrogen, and methane.
2. Turn on the cooling water and the microwave power supply
3. Adjusted the cavity applicator sliding short to an adequate length position. This
length is about 16.25 ± 5 cm.
4. The discharge is ignited by adjusting the pressure to 10 Torr and then by increasing
the instant microwave power and coupling this power into the applicators.
5. After discharge ignition, the cavity applicator may have to be further adjusted by a
slight tuning of the sliding short until the incident microwave power is matched into
the cavity (reduce the reflected power to close to zero). For plasma processing a
good to excellent power, match means that almost all of the microwave incident
power is coupled into the applicator. When comparing incident and reflected power
levels, the reflected power should be less than 15% of the incident power. Note
that it is not necessary to achieve “zero reflected power.”
6. Once the microwave discharge is created the operating pressure is increased to
the desired process conditions (flows, pressure, power). While increasing the
pressure from the ignition pressure of a few Torr to higher processing pressure the
sliding short and input microwave power are readjusted to place and keep the
discharge over and in good contact with the substrate. After the desired pressure
and the desired gas flow are reached, the microwave power is further adjusted to
achieve an excellent temperature uniformity over the substrate.

137

3.6.4 Microwave Plasma Assisted CVD Processing
As outlined above the objective of tuning the microwave power delivery (i.e. short,
reflected power) and the process parameters (i.e. pressure, flows, power) is to achieve a
uniform distribution of the desired temperature across the substrate surface. A uniform
temperature distribution is needed to achieve a uniform diamond growth in terms of
thickness distribution and material quality.

Figure 3. 19 Plasma in an ideal contact with the substrate

Figure 3.20 is a sketch of the plasma discharge region in ideal contact with the
substrate surface. The plasma covers the entire substrate surface area but does not
extend far beyond and thus remains contained and does not touch the quartz dome. If
the discharge were too small, non-uniform deposition would occur.

138

The plasma discharge volume depends on pressure and absorbed power. At a given
pressure increase in power will increase the size of the discharge. At a given power an
increase of the pressure will reduce the size of the discharge. Subsequently, there exists
an optimal region in the pressure-power parameter space in which the plasma discharge
volume will just cover the substrate surface. See (see Figure 3.21).
The absolute temperature is critical for the particular diamond deposition process and
may range from 600 - 1100ºC. Higher temperatures yield higher growth rates. The
temperature depends strongly on the pressure and weakly on the absorbed power (see
Figure 3.21). As a consequence, the pressure of the system and the reactor cooling
systems are adjusted to achieve the desired growth temperature window. Then, the input
power is adjusted to vary the plasma discharge volume over the processing surface area.
The details are shown in Figure 3.21 and also see reference [141], [142] for further details.

139

Figure 3. 20 Operating field maps and matching, i.e., Pref/Pinc, versus absorbed power for
three constant pressure conditions: 120, 180, and 240 Torr [141].

3.6.5 System Shutdown
Once the experimental run has been completed: the software is turned off in the
following order:
1. The microwave input power is decreased to 2000 W
2. The methane is turned off
3. The pressure is turned down twenty Torr by twenty Torr until 100 Torr
4. Then the microwave input power is turned off
5. Hydrogen gas flow is set at zero (0)
6. Throttle valve is opened

140

The system is turned off in the following order:
1. Hydrogen and methane gas lines
2. Microwave power supply
3. Water chiller

3.6.6 Unloading the diamond substrate
Once the system is cooled down, the diamond substrate is unloaded from the reactor.
1. Release the eight nuts in the outer cavity shell and lift off the cavity shell via the
mechanical system.
2. Open the nitrogen gas line (nitrogen flows into the vacuum chamber during the
venting process)
3. Close the throttle valve and waited until the pressure is ≈ 750 Torr, then the
vacuum chamber door can be pulled open
4. Remove the bell jar, take out the substrate holder with the grown diamond still
inside the pocket holder.
5. Place back the bell jar, open the throttle valve then to pump down the reactor
chamber in order to leave the system under vacuum.
Leave the system under vacuum until loading for the next experiment.

3.7 Grown Diamond Substrates Measurements
Before, during and after the growth process or the hydrogen etching experiments,
several different measurements were taken on the substrate. There were four important

141

diamond growth outputs: (1) the substrate temperature measurement, (2) linear growth
rate, (3) the weight diamond gain, and (4) lateral growth rate.

3.7.1 Measurement of Substrate Temperature
Throughout the growth process, the substrate temperature was measured with a 0.96
μm wavelength monochromatic optical pyrometer with an emissivity of 0.1. Two-color
pyrometer temperature measurements were also made under the same experimental
conditions and were approximately 100 K higher than the one-color measurements over
the 800 K–1200 K temperature range of the experiments presented here. As shown in
Figure 3.4 the pyrometer was placed on a tripod which was kept fixed at a constant
position for all temperature measurements to ensure that all measures are under the
same experimental conditions.

3.7.2 Linear and Lateral growth rate (V100 and V111)
This thesis research measures two diamond growth rates, i.e. linear growth rate, V100
and lateral growth rate, V111. These individual notations owe its name to the preferential
directions growth directions on the diamond seed.
The linear growth rate is calculated by the measurement of the thickness of the
substrate before and after the growth process by a linear encoder Model: Solatron
DR600. Five measurements points on the top surface on the diamond seed were taken
to calculate the thickness. One point was in the center, and the other four points are near
each corner. Once the measurements were completed the values were averaged. These
measured values were named the initial linear encoder, Linitial. The same method was

142

used for the grown diamond surfaces; these are named after linear encoder, Lafter. After
these measured were made the following equation calculates the linear growth rate:

V(100 ) ( m / h) 

( Lafter )  ( Linitial )
time

Time refers to the total diamond growth process in hours. The lateral growth rate is
calculated with the dimensions of the outward growth lengths l1 and l2 as is shown in
Figure 3.22. Where L is the total length of the grown substrate, (i.e. L= l1 +l2 + 3.6 mm).
The measurements of these dimensions are recorded using digital microscope
KEYENCE VHX-5000; usually at 30X magnification.

Figure 3. 21 Optical microscope image of the side view of the as-grown diamond
substrate. The red arrows show the lateral and linear growth direction. The dashed red
lines identified the dimensions of l1 and l2.

143

Once the measurements are completed, the following equation calculates the lateral
growth rate:

V(110) ( m / h) 

( L  3.6) 2
time

In the event that outward growth is not uniform on both sides, the lateral growth rate would
be considered as an average of both dimensions, i.e. l1 and l2. Only the linear and lateral
growth rates data are reported in this dissertation since they yielded a more accurate
linear growth rate than weight gain data.

3.8 Main Analytical Instruments
3.8.1 SCD Surface Morphology and Surface Area
A digital microscope KEYENCE VHX-5000 was used to capture the optical images of
the top surfaces. Optical images of the top surface were taken of each experimental run
to analyze the surface morphology and the crystal shapes of the SCD grown substrates.
The top surface area was then calculated from these images by drawing the outline of the
irregular top surface and using the software VHX-500 to determine the area based on the
shape and size of the top surface. The irregular shapes and the jagged edges of the top
surfaces were taken into account by the VAX-500 software integration techniques as is
shown in Figure 3.23. With this instrument, the defects, surface smoothness, and
transparency on the diamond top surface were also observed and analyzed.

144

Figure 3. 22 Optical microscope image of the top view of the as-grown with a white circle
drawn around the diamond shape.

3.8.2 Scanning Electron Microscopy (SEM)
SEM is one of the most widely used surface characterization techniques. It images
the sample surface by scanning it with a high-energy electron beam. Electrons are
thermionically emitted from an electron gun fitted with a filament cathode. Tungsten is
normally used as the filament in thermionic electron guns, due to its high melting point
and low vapor pressure allowing it to be heated for the easy electron emission. When the
emitted electrons pass through a condensing lens, made by a coil of current-carrying wire,
they are collimated into a small beam by forces from a focusing magnetic field. After the
condensing lens, there is another coil which scans the beam of electrons over the sample.
The electron beam is finally focused to a small spot on the sample surface by a final
objective lens. All the SEM components are contained inside a vacuum chamber to
minimize electron-gas interactions. The SEM images reported in this thesis were taken in

145

the MSU Engineering Composite Center by using a Carl Zeiss Variable Pressure SEM
EVO LS25.

3.8.3 Secondary Ion Mass Spectrometry Analysis (SIMS)
Secondary ion mass spectrometry (SIMS) is an analytical technique that detects very
low concentrations of dopants and impurities. It can provide elemental depth profiles over
a depth range from a few angstroms to tens of micrometers. SIMS is a good analytical
method with excellent detection sensitivity for impurities, with ppm or lower detection
sensitivity. Compared to Raman spectra, it can directly detect the impurities
concentrations in the diamond crystal and determine the quality of the diamond. However,
due to the high cost of the measurement, in this thesis investigation, only a limited number
of SCD samples were analyzed by SIMS.
Figure 3.24 shows a schematic diagram of a SIMS analysis system. As shown in the
figure, a focused primary ion beam (usually O+ or Cs+) was generated by an ion gun [143].
The sample was sputtered or etched with the ion beam. Secondary ions formed during
the sputtering process were extracted and analyzed by using a mass spectrometer
(usually a quadrupole or magnetic sector). The secondary ions can range in concentration
from matrix levels down to sub-ppm trace levels.
The SIMS analysis results reported in this thesis investigation was done by Evans
Analytical Group [143]. The SCD surfaces were etched by the ion beam and profiled over
a depth of 12 µm. The detection limits were 2 x 1015 atoms/cm3. The particular element
that was chosen to detect was nitrogen; it is an important impurity that is presented in this
CVD growth experiments that occur in plasma Reactor C.

146

Figure 3. 23 Schematic of SIMS analysis system [143].

3.8.4 Birefringence Imaging
Birefringence is the optical property of a material having a refractive index that
depends on the polarization and propagation direction of light. The birefringence is often
quantified as the maximum difference between refractive indices exhibited by a crystal
material [144].
Birefringence imaging is an effective method for checking the stress level in the
crystal. By examining the birefringence patterns of grown single crystal diamonds, the
spatial distribution of internal crystal stresses could be obtained. Like optical transmission
measurements, the birefringence imaging also required diamond plates with smooth and
flat surfaces which should be parallel with each other.
Here, an optical microscope (Nikon Eclipse ME600) was used to take the
birefringence images. Two quarter-wavelength retardation filters were positioned at 90°
147

offset rotation. The diamond plates were placed between the filters. The microscope was
used in transmission mode. In this setup, an unstressed cubic diamond crystal that was
optically isotropic would appear black in the image. However, stress and strains in the
crystal could make the optical properties anisotropic and cause birefringence, which lead
to an observable light intensity distribution through the crystal.

3.8.5 Etch Pits Analysis
Hydrogen plasma etching of the diamond surface is a simple method to reveal defects
and dislocations. The hydrogen etching process attacks the weak chemical bonding
around defects and dislocations in the diamond surface, leading to the formation of etch
pits. Two main sources of extended defects in CVD crystals can be identified: (i)
dislocations induced by the substrate’s surface state and its polishing, (ii) dislocations
that directly originate from defects in the bulk of the substrate such as stacking faults. The
generated dislocations tend to thread through the CVD film parallel to the growth direction
[145], [146] .
Etch pits were formed on the free-standing CVD plates and HPHT seed oriented (100)
using a hydrogen plasma. Photographs of the plates and seeds were taken at different
magnification, such as 30 x, 100 x and 700 x to observe the etch pits patterns.
In order to calculate the etch pit density, the images were processed by using MATLAB
Image Segmentation. This image processing method involves a new MATLAB code to
improve the contrast of the pictures and thus made an easy way to detect the defects on
the surface. The MATLAB images processing tool creates uncluttered images which
satisfy the conditions to the programming to detect the defects on the surface. The

148

procedure to process the images is broken into three steps: (1) the digital image
processing, (2) code generation (3) identify and correct possible errors in the defects
recognition.
Step (1): A selected image is imported into the software. Then, the image is enhanced
by contrast adjustment, morphological filtering, deblurring, and other image enhancement
tools. An example is shown in Figure 3.25. The images enhanced may also include
removing noise, removing motion blur, reconstruction, and other morphological
operations.

Figure 3. 24 (a) optical image of ACH09 plate and (b) image enhanced

Once the image is enhanced, it is analyzed. Image analysis is the process of
extracting meaningful information from images such as finding shapes, counting objects,
identifying colors, or measuring object properties. The toolbox provides comprehensive
functions for image analysis tasks such as statistical analysis and property measurement.

149

These tools are selected based on the characteristic of the image. An example of the
image processing in a CVD plate, ACH22, is shown in Figure 3.26.

Figure 3. 25 General view of the images processing tool.

Once the image is completed, the step (2) is to generate the MATLAB code. Each
image processed generates a code that should be exported to MATLAB platform. These
codes are added to the complete analysis code that was previously created in MATLAB.
Finally, the program used a set of reference-standard algorithms and visualization
functions for analysis and proper measurement of the etch pits on the surface.
This new etch pits method allows one to have accurate etch pits density
measurements in short periods of time. Each image can be processed in 15 - 20 minutes,
in comparison with the conventional method in which the process had usually taken in 2
- 5 hours.
150

CHAPTER 4

MPACVD GROWTH OF SINGLE CRYSTALLINE DIAMOND
SUBSTRATES

WITH

PCD

RIMLESS

AND

EXPANDING

SURFACES

4.1 Introduction
Single crystal diamond (SCD) growth was performed in optimized pocket substrate
holders at a high pressure of 240 Torr and a high-power density of 500W/cm3. In an effort
to overcome the challenges of growing large area SCD substrates without a
corresponding polycrystalline diamond (PCD) rim, a growth recipe using these pocket
holders was developed. This growth recipe controls the substrate temperature (T s) and
the incident microwave power (Pinc) in a prescribed function of growth time. Through this
process, the feasibility to enlarge the SCD substrate in situ, i.e., during the growth itself
is shown. By allowing the temperature to increase from 980 ͦ C to 1040 ͦ C, then reducing
the temperature, and then allowing it to drift up again, the deposition process alternates
between the fast growth of the different crystal directions (i.e., {110}, {111}, and {100})
and a slow growth to smoothen the top surface. This leads to an increased lateral SCD
151

growth. The slow growth of the crystal faces in turn leads to a smooth and enlarged top
surface. Certain strategies such as the termination of the growth process at the
appropriate time are critical in obtaining flat and smooth SCD surfaces without the
formation of any PCD rim. The SCD substrates grown via this method have been
analyzed by optical and scanning electron microscopies. The lateral SCD surface area
increased between 1.7 and 2 times greater than the initial seed surface area during one
continuous run. The deposited SCDs have high growth rates of ≈ 30 μm/h resulting in
smooth, flat and rimless substrates, hence indicating the improvement in the quality and
morphology of the deposited substrates.

4.2 Background
As the world is becoming more technologically advanced, engineering applications
require novel materials. The unique and excellent properties of diamond make it one such
special material that can be utilized in high-end optical and electronic applications. In the
past, it has been very difficult to grow and manufacture both large area and high-quality
diamond that is needed for many commercial applications. Recently, it was pointed out
that microwave plasma assisted chemical vapor deposition (MPACVD) offers the most
convenient method of growing high-quality SCD substrates [29]. Different MPACVD
techniques like the mosaic wafer and side growth method, have been utilized by several
groups to grow large area SCDs ([94], [109], [147], [148]). However, the deposition of
large area SCD substrates via MPACVD still presents challenges; i.e. during the growth
process, the usable SCD area shrinks due to the unwanted growth of a PCD rim around
the SCD substrate. Figure 4.1 shows the results of such a conventional MPACVD SCD
152

growth. The SCD was grown with CH4/H2 = 5%, and Ts ≈ 1020 ͦ C for 19.5 h. In order to
prevent process runaway during the growth cycle, the pressure and power were reduced
from the initial values of 240 Torr and 1.9–1.6 kW, to their ending values of 130 Torr, and
2.3–1.9 kW. Mokuno et. al [10] observed that when using an enclosed substrate holder
the SCD surface initially slightly increases during the 0.5 mm to 1 mm of growth and then
as the growth process proceeds, it decreases and becomes surrounded by a
polycrystalline rim. In order to address the challenge of rim growth and also to
simultaneously enlarge the grown SCD size, it is imperative to develop growth processes
that minimize the PCD rim formation while increasing the lateral SCD surface area during
growth.

Figure 4. 1 Thick PCD rim growth on SCD substrate using a conventional open holder
growth method.

A substrate holder, identified here as a pocket holder, was designed and built in an
effort to overcome these problems. In a series of exploratory experiments, this holder
153

demonstrated the growth of thick (0.4 – 1.6 mm) and low-stress SCD with almost no rim,
and created a SCD top surface with a larger lateral area (1.23 – 2.5 times) than the
original seed area [20]. These substrates grown in a pocket holder were of type IIa
diamond quality. Recently, G. Wu et.al., investigated the growth dependence versus the
enclosed holder depth at modest pressures (80 – 100 Torr) [123]. Presented here are
experimental details of an SCD process recipe that utilizes a pocket holder and enables
both the growth of the larger area and rimless SCD substrates. The pocket holder process
was optimized under high pressure (240 Torr) and high microwave power density
conditions for large area SCD growth by varying versus time two important growth
process variables: substrate temperature (Ts) and incident power (Pinc). By real-time
controlling the input microwave power and hence T s, this optimized process recipe
enabled the growth of thicker and larger area SCD without creating a PCD rim.

4.3 Experimental details
The experiments were carried out in the microwave cavity plasma reactor (MCPR) C
configuration. The design and operation of this reactor has been discussed in a previous
publication [18] and also is detailed in Chapter 3 of this thesis. The residual nitrogen
concentration caused by any vacuum system leaks or input gas impurities was ~10 ppm.
The reactor was operated in a safe and efficient region as described in ([16], [141]). Such
a reactor design allows the growth of SCDs at high pressures (> 240 Torr) and high-power
densities (> 500 W/cm3) over long growth durations (10 – 72 hours). This pocket holder
design was presented in Chapter 3 as a conventional substrate pocket holder depth, i.e.
d = 2.3 mm and w = 6 mm. For the specific experiments mentioned here, the constant
154

variables include the growth on HPHT type Ib (100) oriented seeds at a pressure of 240
Torr (i.e. 320 mbar), a hydrogen flow rate of 400 sccm, and a methane flow rate of 20
sccm (i.e. 5% CH4/H2). In all experiments, the reactor was always well matched. Thus,
the input power was approximately equal to Pinc. The varying parameters are (a) Pinc: from
3 kW to 1.96 kW and (b) growth times between 40 – 46 hours.

4.4 The growth process recipe
The process recipe independently varies Pinc versus time (see Figure 4.2) to
intelligently control Ts over the process cycle. An example of the variation of P inc and Ts
versus time for such a process recipe is shown in Figure 4.2. This recipe was used to
grow the SCD, ACH23, shown in Figure 4.2 and Figure 4.3 displays the independently
controlled variation of Pinc versus growth time (red curve) and the corresponding Ts
variations (blue curve).

155

Figure 4. 2 Growth process recipe displaying a process cycle variation of Ts and Pinc
versus growth time for substrate ACH23 (shown in Figure 4.3).

When operating at these high pressures and high-power densities, as the SCD grows,
real temperature control is necessary to keep T s within the 1040°C – 980°C SCD growth
window. Normally during a process cycle, Ts drifts upward, i.e. it increases, versus time
as the SCD grow [20]. As a result, the input power must be decreased versus time in
order to keep Ts within the SCD growth window. In Figure 4.2 this decrease is shown over
the entire process cycle as the continuously decreasing Pinc versus time curve.
Superimposed on this required steady power decrease are several short temperature
upward drifts that the input process recipe allows to occur over a fraction of the process
cycle. These are necessary in order to provide desirable deposition conditions on the
substrate during specific portions of the process cycle. The strategy behind these short
time temperature variations is discussed below.
156

4.5 Experimental results
Using the variation in Ts and Pinc versus growth time (shown in Figure 4.2) SCD
substrates were grown in a 2.3 mm deep pocket holder. Figures 6.3 and 6.4 show
examples of rimless and larger area SCD growth. Note the improvement of the top surface
over the conventionally grown substrate shown in Figure 4.1. ACH23 (Figure 4.3) was
grown in one step for 40 hours, with a thickness of 1.2 mm, and a growth rate of 30 µm/hr.
Similarly, ACH24 was grown in one step for 46 hours, with a thickness of 1.34 mm and a
growth rate of ~29.2 µm/hr. High magnification optical microscope images of these two
substrates, in Figures 4.3 and 4.4, show growth without a rim. The top surface of the CVD
SCD substrate is smooth and flat with no major defects on the surface. The dimensions
of the growth surface for these two substrates were 23.28 mm 2 and 27.30 mm2
respectively. Thus, the lateral surface area of the CVD SCD substrate increased by ~1.7
times for ACH23 and ~2 times for ACH24 over the initial SCD seed surface area (i.e.
13.69 mm2). The outward growth of the SCD substrates is clearly seen in Figure 4.5 (a)
and (b).

157

Figure 4. 3 Optical microscope images of (a) ACH23. Close up views of the substrate in
(b) – (d) show rimless growth with larger SCD area.

The substrates were also evaluated under scanning electron microscope (SEM) and
the images are shown in Figures 4.6 (a) and (b). No rim was found around the edges of
the surfaces and the top surface was observed to be flat with no non-epitaxial growth at
the edges and corners. It was also observed that the (110), (111) and the (100) crystal
faces grew simultaneously during this process leading to the enlargement of the CVD
SCD substrate.

158

Figure 4. 4 Optical microscope images of (a) ACH24. Close up views of the substrate in
(b) – (d) show rimless growth with larger SCD area.

159

Figure 4. 5 Side view images of (a) ACH23 and (b) ACH24 indicate outward growth of the
CVD SCD substrates.

Figure 4. 6 SEM images of (a) ACH23 and (b) ACH24 indicate PCD rim free CVD SCD
substrates.

160

4.6 The growth strategy
As has been discussed in [113], [149], the growth rates of (100), (110) and (111)
crystal faces and hence α, and β growth parameters [56], are dependent on Ts. Thus far,
previous publications on diamond growth discuss the growth of SCDs with a constant
temperature over a certain growth time [148], [150]. The objective of this work is to control
the concentration of radical species density over the seed substrate and T s versus time
in order to obtain a high, uniform growth rate without the formation of a PCD rim. To
achieve this objective, Pinc and hence Ts were varied versus growth time.
It has been observed that the growth rates of (111) and (113) faces increase at higher
Ts [149]. Therefore, in this process for the first few hours, Ts is kept high at ~1020°C. At
this high Ts, the entire seed surface is at a uniform temperature which allows the hot and
dissociated gas radicals to deposit uniformly over the entire substrate surface. As a result,
the corners, edges and the center of the substrate start growing uniformly without a rim.
Once the growth has been initiated for the entire surface in the first 10 – 12 hours, it has
been observed in our experiments that the diamond grows uniformly over the entire
substrate. Ts was then lowered versus time to ~980 – 990°C in small steps of ~ 5°C by
decreasing Pinc. As growth proceeds, the substrate thickness increases, and T s also
increases since the substrate surface approaches closer to the intense plasma. During
this process, the temperature is controlled and allowed to drift upward to ~1020°C
following which Ts has reduced again to ~990°C by reducing Pinc. Such a temperature
variation is followed throughout the process cycle. This real-time temperature variation
both promotes the outward growth and produces a smooth, improved morphology of the
SCD crystal.
161

An example of the variation of Ts and Pinc as a function of growth time is shown in
Figure 4.2 for substrate ACH23 (Figure 4.3). Initially when Ts is high (~1030°C), the
growth rates of the different crystal faces are also high allowing the outward SCD growth.
For the next several hours, the temperature was first lowered to ~1000°C (to point “c” in
Figure 6.2) by lowering Pinc (to point “a” in Figure 4.2). A hypothesis is that this lower
temperature growth aids in obtaining a smooth top SCD surface. Although the growth
rates decrease slightly, the entire surface grows more uniformly and improves the surface
morphology. During the ensuing growth period, Pinc is kept constant from point “a” to “b”
in Figure 4.2. However, during that period, as seen in Figure 4.2, Ts drifts upward from
point “c” (~1000°C) to “d” (~1040°C), since the top of the growing crystal surface
approaches the intense plasma as a function of time. This again allows the fast growth of
the crystal faces, especially in the <110> and <111> directions. After this the temperature
is lowered to 996°C (point “e” in Figure 4.2) by reducing Pinc. The temperature variations
between the high (like point “f” in Figure 4.2) and lower (like point “e” in Figure 4.2)
temperature states promote the different growth rated of the different crystal orientations
versus time. These temperatures fluctuations are carried out throughout the growth cycle
(Figure 4.2). An important fact to note is that near the end of the run, T s is reduced from
point “f” to “g” (~985°C). This temperature decreases at the end of the run and then
stopping of the process at the right time is critical in preventing the SCD from penetrating
into the discharge. When Ts and Pinc are not controlled and the substrate grows into the
discharge, a thick PCD rim has been observed the rim is similar to that shown in Figure
4.1. The strategies mentioned here ensure the growth of a flat and even surface with

162

sharp edges and without any rim (see Figure 4.6). Growth processes with similar
temperature variation were followed for the other substrates discussed here.

4.7 A comparison to earlier results
In order to quantify the outward growth of the CVD SCD substrates, the lateral area
gain (i.e. final CVD SCD top area/original seed top area) of the final growth surfaces was
plotted as a function of the vertical thickness gain as shown in Figure 6.7. Except for the
colored data points, this plot is reproduced from Figure 15 in [20]. The black data points
from the original figure are used as comparative reference data. The data points for
substrates ACH23 and ACH24 are represented by the yellow and red data points in Figure
4.7, respectively. The data points from these substrates, located within the dashed oval
in Figure 4.7, correspond to the trend observed in the substrates grown earlier with a
pocket holder using the MCPRs ([18], [73]) operating at different temperature and power
conditions. The blue data point was also grown at the same recipe conditions and the
black data within the dashed oval was grown by Y. Gu [132] using a similar recipe growth
method. However, with the variation of Ts as described in Figure 4.2, it is now possible to
obtain a larger lateral growth of the final CVD SCD substrate surface along with a high
vertical thickness gain.
This growth process allows the deposition of high quality and large, i.e. ~2 times the
initial seed area SCD substrates. By varying Ts and Pinc as a function of the growth time,
a high growth of (110) and (111) crystal faces is enabled such that a high lateral growth
rate may now be achieved. PCD rim growth was also eliminated. Hence, without the PCD
rim, the grown SCDs may now be analyzed more effectively. These rimless substrates
163

also reduce post processing time as it is not necessary to trim the PCD off. With the lack
of rim growth, the stress and hence cracking in the crystals will also be reduced
considerable, thereby improving the quality of the grown substrates. It is feasible to build
on this process and develop additional similar growth recipes for the further enlargement
and growth of larger and thicker SCDs with single to multi – step growth runs.

Figure 4. 7 Outward growth of the CVD SCD substrates [20].

4.8 Summary remarks
The growth recipe described in this research is a special case of the MPACVD growth
of SCDs in a pocket holder. This particular case of MPACVD growth process depends on
the pocket holder dimensions, the behavior of P inc versus time and also on the
corresponding behavior of Ts versus time. However, it was developed empirically. There
164

may be other possible special recipes to grow different PCD rimless SCD shapes. Some
of these are currently under investigation.

165

CHAPTER 5

SINGLE CRYSTAL DIAMOND GROWTH AT CONSTANT
SUBSTRATE TEMPERATURE (Ts)

5.1 Exploring Constant Substrate Temperature and Constant High Pressure SCD
Growth Using Variable Pocket Holder Depths

5.1.1 Research Highlights
•

Polycrystalline diamond (PCD) rimless, single crystal diamond (SCD) growth is
demonstrated in an optimized pocket holder.

•

A constant substrate temperature process control recipe over the entire process
cycle is demonstrated.

•

The influence of the pocket holder depth on the morphology and area of the grown
SCD is revealed.

166

5.1.2. Introduction
SCD substrates were successfully synthesized using an optimized pocket holder
design. The pocket holder design creates an appropriate thermal environment to shield
the diamond substrate from the intense microwave discharge. This substrate
configuration allows for a polycrystalline diamond (PCD) rimless, uniform single crystal
diamond (SCD) growth process. Square shaped pocket holders with a constant width but
with varying depths were used. The SCD growth procedures were carried out via
microwave plasma assisted chemical vapor deposition (MPACVD) in a 2.45 GHz
microwave cavity plasma reactor C [18] at a constant pressure of 240 Torr and high power
density of ~ 500 W/cm3. By continuously adjusting incident microwave power (Pinc), the
substrate temperature (Ts) was held at 1020°C ± 5°C throughout the entire process cycle.
Under these growth conditions, the crystal morphologies exhibited a smooth and flat
surface. The average growth rate of the SCDs varied between 30 – 24 μm/h as the pocket
depth increased from 2.0 to 2.9 mm. The PCD rim was eliminated for all pocket depths
and the shape of the final, as grown, top SCD surface varied from a square to a circular
shape as the pocket holder depth increased. When using the 2.0 mm and 2.3 mm
pockets, the final top SCD surface area increased to almost twice the original HPHT
diamond seed area and the lateral growth rate was slightly larger than the vertical growth
rate. Birefringence imaging of the grown CVD diamond indicated low stress and SIMS
analysis indicated nitrogen concentration levels in the freestanding, CVD diamond plates
were in the range of 50 ppb to 140 ppb.

167

5.1.3 Background
A current challenge facing the SCD growing research community is the development
of plasma assisted CVD process methods that enable the growth of large area and highquality SCD crystals while operating under high growth rate, high pressure diamond
deposition conditions [8]. A major problem with the growth of large area SCD substrates
is that as the diamond grows thicker, the top surface area of the crystal becomes smaller.
As a result, the top SCD growth surface is smaller than the original HPHT diamond seed
surface, and the shrinking top surface is surrounded by a thick PCD rim [11], [20].
Currently, several research groups are trying to overcome these problems by
developing alternative large area growth methods. Several of these methods are (1) the
mosaic growth approach [148], (2) the flipped crystal growth approach [110] and (3) the
flipped side growth approach [108]. While progress has been made in growing SCD
substrates by these methods, each has problems such as limited growth size, growth
process starts and stop interruptions, crystal stress and cracking, and excessively long
growth times that are required to grow the large diamond substrates. The requirement of
rapidly and efficiently growing high quality SCD imposes the additional condition of
growing diamond under high growth rate and high-pressure diamond deposition
conditions.
There are still considerable technical obstacles that must be overcome before robust
and efficient SCD growth processes are available. These processes must enable the
rapid growth of large and high quality SCD substrates. Therefore, fundamental microwave
plasma assisted chemical vapor deposition (MPACVD) SCD growth approaches are
investigated in this chapter. These growth approaches has shown promise to rapidly grow
168

high-quality SCD, without the formation of PCD rim and also with an expanding SCD
growth surface area.
One such fundamental experimental SCD growth approach is to investigate MPACVD
SCD growth using a “pocket holder” [3] [125]. Recent experiments have demonstrated
that by employing a ‘pocket holder” MPACVD process it is possible to grow thick (> 1.4
mm) and high quality SCDs while simultaneously reducing the PCD rim that surrounds
the top growing surface [20], [124]. The specific process consists of the following
experimental procedure: (1) placing a HPHT diamond seed in an optimized pocket holder
geometry, (2) controlling the substrate temperature, T s, versus time over the entire
process cycle, and then (3) ending the process run when the SCD grows out of the holder
configuration. By employing such growth strategies thick (200 µm – 1.4 mm), freestanding SCD plates were produced in one uninterrupted process step. When the process
was extended to a few separate growth steps, thick SCD diamond plates and cubes (~ 5
mm × 5 mm × 4.5 mm) were produced [20]. Thus, pocket holder diamond growth can not
only reduce the growth of PCD rims that are usually observed during MPACVD growth,
but the process can expand the top SCD surface laterally as the SCD thickness increases
[20], [124]. Recently we have shown that the application of a specific process recipe, i.e.
the real – time control of the input microwave power and hence T s versus time, over the
process cycle, enabled the growth of thick and larger area SCD while eliminating the PCD
rim [125]. The experimental details of this process are presented in the next chapter,
chapter 6 of this thesis.
Another recent investigation summarized results from experiments that also employed
pocket holder MPACVD SCD diamond growth versus the holder depth [123]. The

169

experiments were carried out at 80-100 Torr and with 3-5% methane growth environment.
The holder depth was varied from a flat holder to a depth of 2.0 mm to grow several
hundred-micron thick SCD plates. When operating at these low and moderate pressures,
the use of the pocket holder reduced the diamond growth rate but improved the crystalline
quality and the internal stress of the grown SCD. It is important to note that the
experiments that are presented here were executed under considerably different
conditions than those in reference. The diamond growth is performed at the high pressure
of 240 Torr, with high power density discharges of approximately 500 W/cm3, and over a
large range of pockets depths. Hence the diamond growth rate, diamond size, and
diamond crystal quality were expected to be different over previously reported results
[123].
In previous publications [20], [124], [125] have described some specifically optimized
higher pressure and MPACVD growth results. Only a few data points were presented in
[20]. In [125] using a fixed pocket geometry depth of 2.3 mm experiments were described
where the input power and substrate temperature were varied versus time over the
process cycle. In summary, the results presented in [20], [125] explore pocket holder
growth over a limited experimental window.
In this chapter, the objective is to explore diamond growth over a larger and different
experimental input window. The process cycle time is determined by the time it takes to
grow the SCD out of the pocket. Four different pocket holder depths, i.e. four different
substrate holder geometries, are investigated. An input power versus time recipe was
established for each pocket depth in order to achieve the desired constant substrate
temperature over an entire process cycle run. Therefore, a different process recipe was

170

developed for each pocket holder depth. During each entire process cycle run, the
substrate temperature is held approximately constant within 1020°C ± 5°C by adjusting
the input microwave power versus time. Our results show that rimless growth can be
achieved over a range of pocket depths and input power versus time process cycles.
However deeper pockets yield smaller final top surfaces and in fact can result in final
surfaces that are smaller the original seed surface. These experiments provide a deeper
understanding of the pocket holder growth versus the pocket depth/input power variable,
as well as provide insight on how to further improve SCD growth in a particular pocket
holder geometry.

5.1.4 Experimental Details
The growth procedure was carried out via MPACVD in a 2.45 GHz microwave cavity
plasma, reactor C [18] at 240 Torr and by using a high power density discharge of
approximately 500 W/cm3. The residual nitrogen concentration caused by any vacuum
system leaks or input gas impurities was held below 10 ppm. In the experimental runs
presented here the nitrogen concentration was less than 10 ppm. However, the nitrogen
concentration could vary from 5 ppm to 9 ppm from run to run. SCDs were grown on
HPHT type 1b, (100)-oriented (on all sides) single crystal diamond seeds for 10 – 74
hours with H2 flow rates of 400 sccm and a methane concentration of 5%. A feed gas
mixture of hydrogen and methane gases with purity levels of 99.9995% and 99.999%
respectively was used for all experiments.
As is discussed in [20], the holder dimensions play an important role in the
deposition of large SCD substrates. Many experiments have used the conventional open
171

holder for CVD diamond growth. The “open holder geometry” is defined here as a holder
geometry where the diamond seed is placed on a large flat open holder surface. This has
been reported in [10], [11], [20], [151]. These experiments indicated that if one used a flat
open holder design, thereby exposing the diamond directly to the discharge and directly
to the electromagnetic field, a PCD rim is deposited and forms around the grown SCD
material. Reference [150] recently also has demonstrated a significant decrease in the
total top surface area of the SCD when using the open holder configuration.
The SCD was placed in a pocket holder as discussed in [20] and a series of pocket
holder experiments were performed. The generic pocket holder shown in Figure 5.1 (a),
was used in all the experiments. Two important variables when growing a large, smooth
and flat top diamond surface are the pocket holder depth (dn) and the pocket width (w) as
is shown in Figure 5.1 (a). The HPHT diamond seeds used in all experiments were placed
in four different pocket holder depths. The pocket width, w was held constant at 6.0 mm
for all pockets. The pocket holder was redesigned to accommodate the different pocket
depths. The redesign slightly adjusted the holder thickness, c, to allow the diamond
growth to take place within the appropriate substrate temperature window. During each
process cycle, a pocket holder was placed in the reactor as shown in Figure 5.1 (b).

172

Figure 5. 1 Conventional pocket holder design geometry for SCD growth (top side). The
pocket substrate holder placed on the reactor cooling stage (bottom side).

Before loading the diamond seeds into the reactor, they were carefully cleaned with
two freshly prepared chemicals etch solutions. The first solution was made from a nitric
and sulfuric acid mixture in a [1:1] ratio heated at 350°C. The second solution was made
up of hydrochloric acid heated at 350°C. The HPHT diamond seed was then cleaned by
mild agitation in a beaker containing acetone and then methanol. Throughout the growth
process, the substrate temperature was measured with a 0.96 μm wavelength
monochromatic optical pyrometer with an emissivity of 0.1. Two-color pyrometer
173

temperature measurements were also made under the same experimental conditions and
were approximately 100 K higher than the one-color measurements over the 800 K –
1200 K temperature range of the experiments presented here. The pyrometer was placed
on a tripod which was kept fixed at a constant position for all temperature measurements.
This ensured that all temperature measurements were made under the same
experimental conditions. By using the emissivity of 0.1, the measurements presented
here can be directly compared with measurements from previous publications [16], [18],
[20]. All the diamond growing experiments made use of a 2.45 GHz, 10 kW
Cober/Muegge microwave magnetron power supply that was operating at the low end of
its recommended power operating ranges, i.e. ~2 kW, and additionally was not well
filtered. Thus, the microwave power supplied was not constant but was both frequency
and amplitude modulated at 120 Hz. The resulting discharges at times even visually
flickered. Photographs of the discharge revealed that the visual discharge oscillated with
respect to time at 120 Hz during the complete process cycle, and sometimes depending
on the power and pressure input conditions would turn off during part of the 120 Hz cycle
[152]. Thus, all the experimental results described here are considered to be excited by
an unfiltered microwave power supply, i.e. are not excited by CW microwave power. While
it is well known that pulsed microwave power in the range of 100 – 250 Hz can influence
MWPACVD diamond growth rates [58], [75], [153]–[155], it is difficult to say how varying
the discharge that is used here might change the growth rate over the growth rates from
a purely CW excited discharge. However, the variability of the power supply is pointed
out here, because it might result in growth rates that are considerably different from the
growth rates from a purely CW excited discharge.

174

The CVD diamond growth procedure involved two steps. The first was a three-hour
H2 plasma etch step. The following process parameters were used for the plasma etching:
the H2 flow rate was 400 sccm and the substrate temperature was held at 850°C – 980°C.
The second step was the diamond growth step. Specifically, SCD growth was
experimentally investigated over the following subset of process variables: (1) pocket
holders with a width of w = 6.0 mm and four different depths of d1 =2.0 mm, d2 = 2.3 mm,
d3 = 2.6 mm, and d4 = 2.9 mm (see Figure 5(a)), (2) input powers being adjusted versus
time between 3.5 kW and 1.8 kW, and (3) growth times being varied between 10 – 74
hours. Thus, the pocket depth dimensions were varied from 1/3 to 1/2 of the width of the
top of the pocket opening.
The following experimental variables were held fixed: (1) HPHT seed size, 3.7 mm x
3.7 mm x 1.4 mm, (2) total flow rate = 420 sccm, and (3) methane concentration = 5%.
The substrate temperature was monitored after every thirty minutes during all
experiments and was held approximately constant at 1020 °C ± 5°C during each
experiment. During the process run the discharge pressure was 240 Torr, and the input
microwave power recipe varied versus time over the process cycle run. Input power was
varied from run to run between the range of 3.5 kW to 1.8 kW due to the varying pocket
depths and the thicknesses, c, of the different substrate holders.
The thickness gain of the grown SCD substrate was measured by a Solartron DR600
linear encoder measurement with an accuracy of 0.01 mm. The average diamond
thickness was determined by measuring at the four corners and at the center of each
grown substrate and then these five measurements were averaged. A digital microscope
KEYENCE VHX-5000 was used to capture the optical images of the top surfaces. Optical

175

images of the top surface were taken of each experimental run. The top surface area was
then calculated from these images by drawing the outline of the irregular top surface and
software VHX-500 determined the area based on the shape and the size of the top
surface. The irregular shapes and the jagged edges of the top surfaces are taken into
account by the VAX-500 software integration techniques. Variable pressure scanning
electron microscopy (VP - SEM), a Zeiss EVO LS 25 was performed to analyze the
surface morphology and the crystal shapes of the SCD grown substrates. The quality of
the grown SCD substrates was also investigated by using birefringence imaging and
SIMS analysis techniques.

5.2. The Single Crystal Diamond Growth Process
5.2.1 The SCD growth recipe: constant temperature deposition
The growth recipe required varying incident power (Pinc) as a function of the growth
time to hold Ts constant throughout an entire growth process run. Thus, Pinc was adjusted
during an individual CVD growth process run to compensate for the T s variations. For a
better understanding of the process cycle, an example of the variation of P inc and Ts
versus time as is shown in Figure 5.2, for a typical generic reactor. The first step is to
place the HPHT diamond seed in the pocket holder. Next, the gases flow into the quartz
bell jar chamber; the discharge is ignited and the initial operating condition is established
at Ts ≈ 1020°C, p = 240 Torr at time t = 0.
If Pinc is held constant as the deposition time increases, then as is shown in Figure 5.2
Ts increases versus time from operating point A to point B. This occurs because, as
diamond deposition proceeds, the CVD SCD volume increases, and the SCD grows
176

thicker up into the discharge. Other reactor characteristics also change over time, such
as PCD deposition on the substrate holder, film deposition on the reactor walls, etc. These
changes result in a varying reactor operating field map curve as a function of the growth
time.

Figure 5. 2 Reactor operating field map

In order to achieve a constant Ts during the growth process, Pinc must be continuously
decreased throughout the process cycle. Examples of such a variation of the Pinc as a
function of the growth time is shown in Figure 5.3 (a) and (b). By continuously adjusting
the Pinc (red curve in Figure 5.3 (a) and the blue curve in Figure 5.3 (b)) Ts is held
approximately at ~1020°C throughout the entire process (black points in Figure 5.3(a) and
(b)). Each pocket holder depth had different reactor deposition conditions. Therefore, an
individual power versus time growth recipe had to be developed for each pocket holder
depth. As a result, both the input power variation and the deposition cycle times are
177

different for each holder depth. However, the power variation versus time recipes have
similar power versus graphical time shapes; i.e. the power variation versus time initially
starts at a high initial value and then gradually decreases versus time until the process
cycle is ended.
A typical experimental variation of power versus time recipe is as follows: the initial
starting input power is 3000 W and with a corresponding measured discharge microwave
power density of 494 W/cm3. As the process cycle develops versus time the input power
is continuously reduced and the process ending input power is 2600 W with a
corresponding microwave power density of 428 W/cm3. Thus, over the process cycle,
both the input power and the discharge power density are reduced by a similar 13%.

Figure 5. 3 Diamond growth recipe versus time for pocket depth of (a) d2 = 2.3 mm and
(b) d3 = 2.6 mm.

178

5.2.2 Stopping the SCD growth process
An important step during the growth cycle is ending the process at the right time. The
growth process cycle is ended when the diamond starts to grow out of the pocket. This
prevents the growth of the PCD rim. An example of the top surface of the diamond crystal
when growth is stopped before the top surface protrudes out the pocket is shown in Figure
5.4 (a). The smooth part in the center of the photo is the top surface of the grown diamond,
and the rough surface surrounding the SCD, is the PCD which grows on the holder. In
Figure 5.4 (b) the yellow rectangle represents the original HPHT cross-sectional area,
and the blue shape represents the grown CVD diamond cross-sectional area. Figure 5.4
(c) displays optical microscope images from a single process cycle run of 60 hours
(ACH02), with a thickness of 1.60 mm and growth rate of 26.5 μm/h. The diamond was
grown using a pocket holder with a depth of d3 = 2.6 mm and was stopped before the top
surface grew out of the pocket. The top growth surface shows a circular surface
morphology with a flat and smooth surface. This is displayed in Figure 5.4 (b). There is
no evidence of PCD on the top surface or on the edges of the grown substrate.

179

Figure 5. 4 (a) As-grown uncleaned SCD with the top surface is still entirely within the 2.6
mm deep pocket; (b) cross-sectional schematics when the grown diamond (in blue) is still
within the pocket; (c) optical microscope images of side and top views of as-grown and

180

Figure 5.4 (cont’d)
cleaned diamond substrate (ACH02). Note that the bottom yellow region is the original
HPHT diamond seed. The green square represents the original seed boundary.

In contrast, Figures 5.5 (a) and (c) displays optical microscope images for a process
cycle of 48 hours (ACH22-b), with a thickness of 1.51 mm and growth rate of 31.5 μm/h.
The diamond substrate was grown using a pocket holder with the depth of d2 = 2.3 mm.
Here the process was not stopped before the SCD protruded into the discharge. When
that happens, the SCD top surface grows into the discharge, and a PCD rim is formed on
the edges as shown in Figure 5.5 (b) and (c). Note in Figures 5.5 (a) the smooth part in
the center of the photo is the top surface of the grown diamond with a PCD rim. The rough
surface, surrounding the growing SCD, is the PCD which grew on the holder and into the
top surface of the SCD. In Figure 5.5 (b) the dark blue area represents the PCD diamond
grown on the SCD surface, the light blue represents the grown SCD material, and the
yellow rectangle represents the HPHT diamond seed. These examples demonstrate the
importance of stopping the experiments at the appropriate growth time.

181

Figure 5. 5 The ACH22-b SCD which has grown out the pocket; (b) the associated crosssectional schematic of the grown SCD when it has grown out of pocket; (c) optical
microscope images of side and top view of a grown diamond substrate (ACH22-b).

182

5.2.4 Varying the input power and adjusting the reactor for the different pocket
holder depth recipes.
The pocket depths, dn, and the associated thicknesses, c, of the molybdenum holder
design shown in Figure 5.1 (a) were varied from one pocket depth experiment to another.
Additionally, in order to achieve a constant temperature growth recipe from process cycle
to process cycle both of the input power and the holder thickness, c, were increased as
dn was increased. Thus for each particular dn run the reactor had to be: (1) slightly
readjusted to operate in the safe and efficient operating regime [141] and (2) slightly
adjusted over a different input power range as dn was varied. When using deeper pockets
and thicker holders, operation at the higher input power ranges was required. All the
different growth recipes for the different pocket depths looked similar versus time to the
recipe shown in Figure 5.2. The major difference between the recipes was that the deeper
pockets required higher input powers over the entire process cycle. For example, the
beginning input powers for the 2.0 mm and 2.9 mm pocket depths were 2.3 kW, and 3.5
kW respectively. As the process cycle evolved versus time for all of the pockets, the input
power decreased with time and a similar way, but the deeper pockets had higher power
levels over the entire process cycle recipe.
5.3. Experimental Results
5.3.1 SCD Growth versus time
Two groups of exploratory experiments were performed to improve the understanding
of the SCD growth process and to identify the right time to stop the process. These
exploratory experiments were conducted using two typical diamond pocket depths: d2 =
183

2.3 mm and d3 = 2.6 mm. The process cycle times depended on the size of the pocket
depths. For d2, the cycles times were 10, 24, and 48 hours; while for d3 the cycle times
were: 10, 24, 48, and 60 hours. Table 1 summarizes the different growth parameters and
output variables for the process run identified as the three ACH22 cycle runs and another
four ACH21/ACH02 process cycle runs. All seven experiments represented individual
experimental runs that were started from a new seed. T s was kept constant at 1020°C
throughout each individual process cycle run. A Ts measurement was recorded and the
power input was adjusted every thirty minutes. The growth rates that are summarized in
Table 5.1 are the thickness gain divided by the growth time. Thus, this growth rate is an
average growth rate over the entire process cycle time. The instantaneous growth rate
may vary from this average rate during the process cycle.

Table 5. 1 Growth parameters for SCD growth using the d2 = 2.3 mm and d3 = 2.6 mm
pocket holder depths.
Pocket
depth
d2
=2.3mm

Sample

Process Thickness
cycle
gain (mm)
(h)
ACH22
10
0.274
ACH22-a
24
0.717
ACH22-b
48
1.520

d3=2.6mm

ACH21
10
0.289
ACH21-a
24
0.705
ACH21-b
48
1.464
ACH02
60
1.600
*Useful SCD surface: **Includes the PCD rim

184

Average
growth rate
(μm/h)
27.3
30.0
31.5
28.9
29.4
30.5
26.5

Area Normalized
(mm2)
area
16.7
20.95
21.7
16.8
16.6
18.6
13.2

1.22
1.53
(2.36)**;
(0.87)*
1.22
1.21
1.35
0.97

Figures 5.6 (a) – (g) display images of the top surface areas of the substrates that are
listed in Table 1. All the grown substrates display a smooth and flat top surface without
any defects or PCD except for the sample ACH22-b, which is smooth and flat in the
center, but, as is shown in Figure 5.6 (c), the top surface is surrounded by an outer PCD
rim. The as grown surfaces are very flat. The corners do not differ more than ± 2 µm when
compared with the center thickness. As discussed previously in section 3.2, sample
ACH22-b grew out of the pocket and into the plasma for approximately the last 10 hours
of its process cycle leading to the formation of PCD.
The data in Table 5.1 and Figures 5.6 (a) and (d) demonstrate that after the 10 hours
of the deposition process, even though two different pocket depths were used; the
resulting CVD grown diamonds surfaces were very similar at this early time in the growth
cycle. The top surface areas of both samples were the same, 16.7 mm 2 with d2 and 16.8
mm2 with d3.
After 24 hours, the top surface area and the average growth rate decreased slightly
as the pocket depth increased. From Table 5.1 it can be seen that the deeper pocket
appears to have a somewhat reduced average thickness growth and also restricts the
outward growth (see for example – ACH21-a versus ACH22-a). Subsequently, samples
ACH21-b and ACH22-b which were grown for 48 hours, displayed dramatically different
output growth results and morphologies. These differences were not surprising since
ACH22-b, which was grown in a 2.3 mm depth pocket, was protruding from the pocket at
the end of its cycle and ACH21-b was still inside the pocket when the process was
stopped. At the end of its cycle, ACH22-b had a PCD rim as is shown in Figures 5.5 and
5.6 (c).

185

Figure 5. 6 Optical microscope top surface images of substrates samples listed in Table
1. For d2 = 2.3 mm (a) 10 hours, (b) 24 hours, and (c) 48 hours. For d3 = 2.6 mm (d) 10
hours, (e) 24 hours, (f) 48 hours and (g) 60 hours.

186

The top surface of ACH22-b protruded out of the pocket during the last approximately
10 hours of the process cycle. Once the top surface grew out of the pocket, the top surface
was directly exposed to the intense plasma discharge, and the electromagnetic fields are
concentrated on the edges of the diamond, and the deposition process accelerates and
appears similar to that which shown in Figure 5.5. Consequently, the substrate
temperature is non-uniform, and this results in the growth of a PCD rim. This PCD rim is
responsible for the increased top surface area of 21.7 mm2 for ACH22-b as indicated in
Table 1. The thick PCD rim also had grown inward and upwards on the SCD surface.
Hence, the useful SCD surface area decreased. In this case, the top useful single crystal
surface area decreased to 0.87 times of the initial original HPHT diamond seed area (see
Figures 5.5 (b) and 5.5 (c)).
At the end of the process cycle, ACH22-b measured a thickness gain of 1.520 mm
and an average thickness growth rate of 31.5μm/h. This run can be compared to the
similar ACH21-b run displayed in Figure 5.6 (f). The difference between these two runs
is that ACH21-b run was ended before it protruded from the 2.6 mm depth pocket and as
a result, it displays no PCD rim. Sample ACH21-b has a thickness gain of 1.464 mm and
an average growth rate of 30.5 μm/h. Its top surface remained below the final holder
surface. The ACH02 experiment lasted 60 hours. When the run was completed, the top
surface of the SCD was just starting to protrude out of the pocket. Figure 5.6 (g) shows
that the top surface is smooth, flat, and without any visual defects and is not surrounded
by a PCD rim. Its shape is circular and the normalized surface area is 0.97, which is just
slightly smaller than the original square HPHT diamond seed area.

187

5.3.2 SCD growth versus pocket holder depth
The results from eleven different growth versus depth experiments are summarized in
Table 2 and in Figures 5.7 – 5.9. The experiments were performed using constant Ts
deposition recipes, i.e. 1020 ± 5°C, similar to the one shown in Figure 5.3. The growth
time was varied from 40 hours to 74 hours for each run. The length of the process cycle
was experimentally determined, by visually observing the location of the top surface and
then stopping the process before the SCD protruded into the discharge as discussed in
section 5.2.2. The experiments that were not stopped on time developed a PCD rim,
similar to ACH22-b. Those experiments were not included in the table because they were
grown out of the desirable growth recipe conditions. The average growth rates listed in
Tables 5.1 and 5.2 are plotted versus pocket depth in Figure 5.8. As indicated in Figure
5.8 the growth rate is a maximum for the shallow pockets and decreases as the pocket
depth increases.
Figures 5.7 (a) – (d). Although Figure 5.7 just shows four of the eleven samples, all
the other samples listed in Table 5.2 displayed a smooth top surface without a PCD rim
around the edges of the top surfaces. It was also observed that the grown diamond in
pocket holder depths of 2.0 mm and 2.3 mm exhibited jagged edges and a square top
surface. These samples grew outward and upward as a function of the time at an average
growth rate of approximately 30 μm/h. As a result, expanded areas as large as 26.5 mm2
were obtained. For these pockets, the total area increased to almost twice, i.e. 1.6 – 1.94
times, of the original HPHT diamond seed area.

188

Table 5. 2 Summary of different growth versus depth experiments for eleven samples at
a substrate temperature of 1020 ± 5°C.

Pocket
depth
d1= 2.0
mm
d2= 2.3
mm
d3= 2.6
mm

Sample Pocket Process Thickness
Average
Normalized
depth
cycle
gain (mm) growth rate
area
(mm)
(h)
(μm/h)
ACH31
2.0
40
1.20
30.0
1.94
ACH44
2.0
40
1.17
29.5
1.61
ACH26
2.3
42
1.22
29.1
1.78

ACH09
2.6
50
1.40
28.0
1.40
ACH212.6
48
1.46
29.8
1.32
b
ACH27
2.6
50
1.43
28.6
1.33
ACH08
2.6
55
1.49
27.1
1.35
ACH02
2.6
60
1.60
26.5
0.96
ACH11
2.6
74
1.93
26.1
0.57
d4= 2.9
ACH35
2.9
45
1.06
23.5
1.10
mm
ACH47
2.9
50
1.38
27.3
0.85
Optical microscope images of selected samples from Table 5.2 are displayed in

189

Figure 5. 7 Optical microscope images of the typical samples shown in Table 2 that were
grown in pocket depths of (a) 2.0 mm, (b) 2.3 mm, (c) 2.6 mm and (d) 2.9 mm.

The samples grown in pocket depths of 2.6 mm and 2.9 mm exhibited smooth edges
rather than the jagged edges that were observed in the shallower pockets. The measured
average growth rates were 23 – 30 μm/h (see Figure 5.8). However, the final resulting
shapes were octagonal and circular and the final surface area gain varied from 1.5 times
to less than the original seed area; i.e. the deeper the pocket, the smaller the final surface

190

area. One sample, ACH11, exhibited a decreased area gain of 0.57 from the original seed
area after 74 hours of growth. It was found that longer process cycles and deeper pocket
depths had smaller area gains.
The shape of the top surface varied as the crystal thickness increased. The final
crystal shape was a function of growth time and pocket depth. This can be more clearly
understood by referring to the data presented in Table 5.1 and Figure 5.6. As is shown in
Figures 5.6 (d) and (e) after 10 and 24 hours respectively of the process cycle, the
diamond was grown uniformly on the entire surface, and the top area was almost square
shaped. However, after continuing the growth versus time beyond 24 hours the surface
area decreased and changed from square shaped to the octagon shape and even for a
longer run, ACH02, into a circular shape. This is indicated in the data presented Figures
5.6 (d) – (g) and in Table 5.1, and pictorially by the red line shapes surrounding the images
of the SCD top surfaces in Figure 5.8.
When using the 2.3 mm pocket the growth rate increases with time. For example, see
the data points in Table 1 for 2.3 mm pocket depth. The lower average growth rates are
from, the shorter growth times, and the average growth rate gradually increases as the
time increased. This suggests that the instantaneous growth rates slightly increase as the
diamond grows thicker and the top surface approaches the pocket opening.
The data for the 2.6 mm pocket still depends on time but in a different way than 2.3
mm pocket depth. When the growth times vary between 48 to 55 hours the average
growth rates vary over the narrow range of 27 – 30 µm/hr and the final top surface shapes
are an octagon. However, if the growth time is much larger, i.e. 60 – 74 hours the growth
seems to decrease with time and the final shape changes to a circle.

191

This is illustrated by the two data points: (1) ACH02 grown for 60 hours and (2) ACH11
which was grown for 74 hours. In these longer runs, the average growth rate decreases
and the final top surface changes from an octagon to a circle and the surface area are
smaller than the original seed. During these two longer runs the diamond grows out of the
original molybdenum pocket but is still in an “effective extended new pocket” that is
created by the additional deposited PCD on the holder. This extended pocket
phenomenon as already mentioned in reference [20]. As the growth proceeds, the
average deposition rate decreases and the shape of the top surface becomes a circle,
but there is no PCD rim formed.

For a better understanding of how the pocket depth influences growth, the results in
Table 5.2 were summarized in Figure 5.9. It is a plot of the lateral growth rate (V110) and
the vertical growth rate (V100). It displays the growth rate ratio between the outward
surface, represented by V110 and upward surface, represented by V100. The ratio between
both rates was calculated with the help of scanning electron microscope (SEM) images.

192

Figure 5. 8 Morphology shapes and the growth rate versus pocket holder depth for
experimental data for Table 5.1 and Table 5.2. The square surface is associated with
pocket depths of 2.0 and 2.3 mm. The circular surface is associated with long runs of the
pocket depth of 2.6 mm, and the octagon shape surface is associated with the shorter
runs that used the pocket depths of 2.6 and 2.9 mm.

Images of the corner of each substrate reveal the crystal shapes and allow the
monitoring of the diamond outward growth. The SEM images show that substrates that
were grown in pockets depths of d1 and d2 lead to a fast and a uniform growth of the entire
surface with the resulting final square area. For example, a fast-outward growth was
observed for sample ACH31 that grew in the shallow pocket (d1 = 2 mm). In contrast,
193

sample ACH35 which grew in a deeper pocket (d4 = 2.9 mm) showed a very slow outward
growth as compared to the growth in the vertical direction.

Figure 5. 9 The ratio between lateral growth rate, V110, and vertically growth rate, V100,
versus pocket depth. Insets are SEM images of the corners of the SCD substrates.

Figure 5.10 displays a plot of the normalized lateral area gain versus the vertical
thickness gain for the data listed in Table 5.2. The vertical axes, (normalized lateral area
gain) is the final CVD lateral area gain, i.e. the final CVD lateral area divided by the initial
HPHT seed area (3.7 x 3.7 mm2). The dashed horizontal line represents the original
normalized HPHT seed surface area. For the case considered in this investigation, the
194

seed surface area is 13.7 mm2. In Figure 5.10, this area has been normalized with respect
to the seed area of 13.7 mm2.

Figure 5. 10 Plot summarizing the normalized lateral CVD area gain versus vertical
thickness gain for different pocket depths for the experimental data listed in Tables 5.1
and 5.2.

The purple line in Figure 5.10 is the “best fit” curve that was drawn in reference [20]
between experimental data points. It indicates the behavior of the lateral area gain versus
thickness gain data points that were presented in [20]. It is reproduced in Figure 5.10 as

195

a reference curve for comparison with the experimental data that is presented in Tables
5.1 and 5.2.
In Figure 5.10 the experimental data points from Table 5.2 are grouped by pocket
depth. It can be readily observed that the shallow pockets have a larger area gain than
the deeper pockets. d1 and d2 pockets yield the largest normalized area gain. These two
pocket holders can grow SCD with area gains of 1.6 – 1.95 before their top surfaces
protrude out of the pocket. The deeper pockets, i.e. d3 and d4 have top surfaces that are
smaller than the purple benchmark curve. The deeper pockets and the longer runs end
up with the smallest process ending surfaces. The long duration runs resulted in a smaller
or no lateral area gain. See for example ACH11 and ACH47. Their top surfaces are
smaller than the original seed. The experimental data for the 2.6 mm pocket clearly
indicates that the thicker crystals produce a smaller top surface. The data indicate that
under the initial constraints of this investigation there is an optimum pocket depth for an
area gain. That occurs for the shallow pockets; i.e. for d = 2.0 mm – 2.3 mm.

5.4 SCD Plate Characterizations
5.4.1 Overview
In order to characterize the grown substrates, freestanding, SCD plates were obtained
by laser cutting perpendicular to the growth direction to remove the HPHT seeds from the
CVD grown substrates. Then the plates were mechanically polished on both sides to
reduce roughness. Examples of these freestanding plates are shown in Figure 5.11. Note
that the edges of the plate ACH33 were also laser trimmed and polished. CVD diamond
grown in the reactor C has been characterized by Raman spectroscopy and IR-UV
196

transmission measurements, and the results have been reported elsewhere [16], [18].
The Raman FWHM were as low as ~ 1.75 – 1.8 cm-1 and the transmission measurements
showed that the SCD grown with reactor C was of similar quality to type IIa diamond.
Summarized below are typical birefringence and SIMS measurements performed on
diamond plates obtained from the ACH20 and ACH33 experiments.

Figure 5. 11 CVD plates after laser cutting and mechanical polishing. The measured
thickness gain for each sample was: 274 μm for ACH22, 255 μm for ACH20 and 600 μm
for ACH33.

5.4.2 Birefringence imaging
Birefringence imaging was performed to obtain information about the distributions of
internal stress and dislocations throughout the crystal lattice. The images of the CVD

197

plates were taken with a Nikon ME600 optical microscope and with an external polarizer
and analyzer. The birefringence images were taken at 5000 ms exposure time. Figures
5.12 (a) and (b) show two examples of the as – grown SCDs along with the birefringence
image of their associated, freestanding diamond plate.
The first SCD CVD sample, ACH20, is shown in Figure 5.12 (a) along with the
birefringence image of its freestanding plate. ACH20 was grown for 10 hours in a pocket
holder d2 = 2.3 mm. The measured thickness gain was 255 μm with a growth rate of 28
μm/h. The birefringence image is dark indicating low stress in the crystal structure. This
can be observed especially in the center of the plate. However, the three edges show
some bright spots indicating some stress. This stress is probably related to the jagged
edges that were observed (See Figures 5.6 (a) and 6 (b)) in the CVD growth results of
the d3 pocket.
The second SCD CVD sample ACH33 grown for 24 hours with a measured thickness
gain of 600 μm thick with a growth rate of 27 μm/h in a holder depth of d3. See Figure
5.12 (b). Similar to the ACH20 birefringence, the birefringence ACH33 image displays a
dark low-stress interior with the exception of two butterfly-shaped patterns. The edges
show three bright spots indicating some small points of stress at the edge of the sample.
The two butterfly shaped patterns seen in the birefringence image of ACH33 are typical
of threading dislocations which are defects in the lattice structure and are observed along
the (100) growth direction ([156], [157]). The two butterfly patterns also coincide with the
2 globular shaped defects on the surface of the as-grown ACH33 substrate (Figure 5.12
(b)). These are most likely due to bundles of dislocations originating from the underlying
seed or originate from the surface of the seed. This is not a typical result. The diamond

198

plates that are more commonly grown appear similar to that shown in Figure 5.12 (a).
However, Figure 5.12 (b) is displayed here to illustrate what can occur when the original
seed or the initial seed deposition surface had some imperfections.

Figure 5. 12 Examples of two as-grown SCD with their corresponding birefringence
images. (a) ACH20 and (b) ACH33.
199

5.4.3 Etch pits analysis
Etch pits analysis was performed on the surface of a free-standing CVD plates,
ACH09, ACH20, and ACH33 using hydrogen plasma etching. The substrates were etched
at 240 Torr for 1 hour with a pure hydrogen discharge. The substrate temperature was
measured at ≈ 1000°C at a constant microwave power of 2.5 kW. The average etch rate
was between 2 - 5 µm/h. The etch pits densities were calculated with MATLAB images
processing presented in Chapter 3, section 3.8.5.
Figure 5.13 (a) shows an optical microscope image of the polished and trimmed
ACH33 plate, and an enlargement selected area on the surface (marked in a black circle).
A large number of etch pits are concentrated around the polishing grooves on the surface.
The recorded etch pit density of such plate was ~ 1.3x103 cm-2, and the etch rate was ≈
2.1 µm/h. As shown in the images, the pits were either not present around the middle of
the plate or they were minimal or were present at low density. The birefringence image
for this plate exhibited low stress, especially in the center. Even though these results were
anticipated, it was also expected to observe pits around the two butterfly patterns
revealed for birefringence image (Figure 5.12 (a)). However, etch pits were not present.
The result suggests the need of a more aggressive etching were the temperature and the
process time are increased. Figure 5.13 (b) displayed optical microscope images of the
CVD plate, ACH20, on which the top surface remained unpolished. After 1 hour of plasma
hydrogen exposure, ≈ 1.8 µm on the uneven grown surface was etched. Random
concentrations of pits were observed along the longitudinal direction, i.e. (010) direction,
as indicated in the enlarged image of the selected area in Figure 5.1 3 (b). These pits

200

distribution could be attributed to dislocations that arise from the HPHT seed. The
average etch pit density was calculated to be ≈ 3.8x104 cm-2.

Figure 5. 13 (a) Etched surface of the polished CVD plate ACH33 and (b) Etched surface
of a unpolished CVD plate ACH20.

201

The shape and dimensions of the pits were analyzed and calculated with SEM images.
Figure 5.14 shows SEM images of the etched diamond area on the unpolished CVD plate
ACH09. The etch pit density was calculated to be ≈ 1x10 4 cm-2, and the etch rate was ≈
5.3 µm. Cross-sectional profiles along the lines crossing the center of the pits are shown
in Figure 5.14 (a) and (b). These images show hollow pits symmetrically aligned with the
(110) growth direction. The depth of the pits was measured and varied between 1.8 µm
– 2.2 µm. The width was measured between 2.5 µm – 9.5 µm, as shown in Figure 5.14
(c).

202

Figure 5. 14 SEM images of the pits in three different areas on CVD plate ACH09 (a) one
side at the edge, (b) somewhere at the center (c) somewhere on the center of the top
surface.

The results of the etch pits densities were compared with some recent results reported
from 2016. See Figure 5.16.
All the plates etch pits densities calculated have significant low values in comparison
with the HPHT seed, even in comparison with previously reported results [142], [124].
The polished plate (ACH33) showed the lowest defect density on the surface. However,
an aggressive etch is recommended. These results still seem high as compared with the
most recent results reported by Tallaire et al. [122] in which the etch pits density was ≈

203

2x103 cm-2. Nevertheless, their investigation used a different growth approach. See
Chapter 2, section 2.6.3. Despite the tremendous improvement size and quality of the
CVD diamond substrates, the objective of low dislocation density material is still
unachieved.

Figure 5. 15 Comparison of the etch pit density with recent results.

5.4.4 Secondary ion mass spectroscopy (SIMS)
SIMS measurements were performed in order to verify the nitrogen concentration in
the grown CVD plates. SIMS was carried out by Evans Analytical Group (EAG). Figure
5.13 shows the depth profile of nitrogen concentration in SCD CVD plates ACH33 and
204

ACH20. The red line shown in the two figures indicates the detection limit (2x1015
atoms/cm3) of the instrument provided by EAG. As shown in the Figure 5.13 the nitrogen
concentration is constant until at least 13 µm depth into the bulk of the CVD sample. The
nitrogen concentrations in the CVD plate ACH33 is approximately 50 ppb, and for ACH20
it was 140 ppb.

Figure 5. 16 Nitrogen concentration in (a) SCD CVD plate ACH33 and (b) SCD CVD plate
ACH20. The red line indicates the detection limit of the instrument.

5.5 SCD growth versus methane concentration
5.5.1 Growth Conditions
Methane concentration is defined as the flow rate ratio of methane gas to total gas,
which is typically composed of hydrogen. Because the methane gas flow rate is negligibly
small under typical growth conditions, the methane concentration is often defined as the

205

flow rate ratio of methane gas to hydrogen gas (CH4 /H2). Here, effects of the methane
concentration on the diamond growth process are discussed.
The effect of the CH4 concentrations on the growth surface morphologies and growth
rate was studied at a constant pressure of 240 Torr and a constant substrate temperature
of 1020°C ± 5 °C. The substrate temperature was controlled by the adjusted the
microwave input power. The input power was relatively high between 2.8 kW – 3.3 kW
for this set of experiments. Four experiments were performed in a pocket holder with d =
2.6 mm and w = 6.0 mm at 4%, 5%, 6% and 7% methane concentration. The experiments
were conducted at 48 hours, however, at 7% CH4 the experiment had to be stopped due
to the soot formation on the quartz dome walls.

5.5.2 Linear and Lateral Growth Rates
Figure 5.17 shows the growth rate of grown diamond substrates as a function of the
methane concentration. The plot on the insert on the bottom side of the figure is the
variation on the microwave power as a function of the methane concentration. The higher
input power was 3.5 kW with a total microwave power density (MWPD) calculated of 580
W/cm3 at 4% methane concentration. The lowest input power was 2.4 kW with a total
MWPD calculated of 404 W/cm3 at 7% methane concentration.
In the range from 4% to 6%, the linear and lateral growth rate showed a strong linear
dependence with methane. When methane increases from 4 to 6, the linear growth rate
also increased from 23.7 µm/h to 29.0 µm/h. The high MWPD with a moderate methane
concentration (4% - 6%) create an appropriate diamond growth condition which increases
the growth rates while the quality of the surfaces is improved. This is attributed that the
206

density number of atomic hydrogen increased with increasing the plasma density, the
hydrogen etching effect on the diamond growth rates is important under moderates
methane concentrations at high MWPD conditions. Similar results have been reported by
[49], [54], [158], [12], [16].

Figure 5.17 Growth rates of CVD diamond grown as a function of the methane
concentration under growth conditions of 3.5 kW - 2.4 kW, 1020 °C and 240 Torr.

An important observation is the lateral growth rate at 4% methane, it was calculated
as zero. There was no lateral growth on the grown CVD diamond substrate. It might be
due to the pocket depth; Figure 5.9 showed that the lateral growth rate for substrates

207

grown in the same pocket depth (d = 2.6 mm) are low in comparison with substrates
grown by using shallow pockets. This fact suggests as the pocket depth increases, the
lateral growth rate of the substrates decreases significantly and is decelerated at low
methane concentrations.
Although some authors have extensively reported an increased growth rate at high
methane concentrations [12], [16], [49], in this study, the linear growth rate at 7% methane
decreased. The growth rate of 7% was measured at 25 µm/h while the lateral growth rate
remains almost the same, i.e. 26 µm/h. This is a decrease in the growth rate of 13.8%
over the growth rate of the grown CVD diamond at 6%. As a suggested by Teraji et al.
[54], this could be a consequence of the incorporation of graphite in the diamond crystal.
Diamond growth is difficult to continue if the diamond growth surface is covered
considerably with non-diamond phases.
It was expected a linear trend in the growth rate for as a function of the methane
concentration at a constant temperature. However, a decreased of almost 14% was
observed in the substrate which grew at higher methane concentration. This experiment
was performed at low MWPD, calculated at 404 W/cm 3. This result indicated a decrease
of 30%. This fact supports that the growth rate mainly depends on substrate temperature
which is mainly controlled by microwave power. It is required to increase both parameters,
i.e. MPWD and Ts, in order to have high lateral and linear growth rates.

5.5.3 Surface morphology
Figure 5.18 shows the optical microscope images of the top surface of the CVD
diamond substrates grown at 1020 °C with 4%, 5%, 6% and 7% of CH4. The substrate
208

grown at 4% exhibited a smooth surface with small step-bunching, as is shown at high
magnification image in Figure 5.19 (a). However, this step-bunching does not represent
significant changes in the top surface morphology. This morphology could be a
consequence of a high etch rate by atomic hydrogen during growth enhanced at high
microwave power (3.5 kW). Similar results have been reported by [49], [59].

Figure 5.18 Optical microscope images of the top surface of grown SCD substrates for
methane concentrations of (a) 4%, (b) 5%, (c) 6% and (d) 7%.

209

With 5% and 6% of CH4, the diamond top surfaces displayed very smooth
morphologies without any defects and with no evidence of PCD as is shown in Figure
5.18 (b) and (c). High magnification image of substrate grown at 5% methane is displayed
in Figure 5.19 (b), a perfectly flat surface is revealed.

Figure 5. 19 Optical high magnification images of substrate grown at (a) 4 % methane
and (b) 5 % methane.

With 7% of methane, step-bunching were observed macroscopically in some parts of
the top surface, mainly in the center, Figure 5.15 (d). Although the top surface is not
perfectly flat, there are no major defects on the surface.

5.6 Summary Remarks
Thick (1.2 – 1.9 mm) high quality and low-stress diamond crystals were grown in a
pocket holder in one continuous process cycle run. It was shown, through diamond growth
versus time studies, that MPACVD SCD grows without a PCD rim if during the growth
cycle the top surface of the SCD remains inside and below the pocket opening. If the SCD
210

protrudes out of the pocket a PCD rim forms. At high pressures of 240 Torr, PCD rimless
SCD can be routinely grown inside a pocket holder, and the pocket holder is an excellent,
hot, thermal environment to grow SCD and to investigate the fundamentals of SCD
growth.
The final (process ending) shape and area of the top SCD surface varies with pocket
depth and process time. Shallow pockets (d = 2.0 mm and 2.3 mm) result in square SCD
surfaces, and deep pockets (d = 2.6 mm and 2.9 mm) result in the octagon and circular
shapes. Shallow pockets have enhanced lateral growth rates and larger area gains.
Deeper pockets do not grow larger area SCD crystals but have reduced lateral growth
rates and reduced area gains and often result in final SCD crystals with smaller ending
surface areas than the areas of the ordinal seeds.
Given the initial constraints of this study, i.e. given the dimensions of our pocket
opening, the pocket depths, deposition time and input powers, there is a range of optimal
pocket depths that can grow SCD both vertically and horizontally. Those pocket depths
appear to be 2.0 – 2.6 mm (Figure 9). In these pockets, the growth rate was ~ 30 – 27
µm/ h. For the shallow pockets, the final top SCD surface area increased to almost twice
the original HPHT diamond seed area and then the lateral growth rate was close to the
vertical growth rate. These experiments have demonstrated that moving to higher
pressure, higher power density and low N2 growth conditions result in high growth rates
with improved quality and almost stress-free SCD. Thick, PCD rimless, high quality, and
low-stress SCD were grown under these conditions.
The etching experiments on the grown SCD plates have displayed low defect densities.
The analysis of such plates with etch pit density measurements has shown a reduction in

211

etch pit density in CVD SCD plates to ≈ 104 etch pits/cm2. Further work needs to be done
to optimize the growth process to produce even lower defect SCD plates.
Smooth SCD surfaces were grown at 4%, 5%, and 6% methane and the lateral growth
was significantly increased with methane. A higher methane concentration of 7% resulted
in a change of surface morphology and a decrease on the linear growth rate. Further work
is needed investigate the variation of methane concentration as a function of the growth
rate, MWPD and substrate temperature.

212

CHAPTER 6

SINGLE CRYSTAL DIAMOND GROWTH: (1) USING VARIABLE
POCKET HOLDER APERTURES AND (2) ADDING AN REGROWTH EXPERIMENTAL RUN

6.1 Introduction
Chapters 4 and 5 investigated SCD pocket growth where the pocket was fixed with a
constant reference geometry, and then the pocket depth was varied from that fixed
reference geometry. In this chapter, SCD pocket growth is experimentally investigated
to understand how this SCD growth technique can be extended to further enlarge both
the grown crystal thickness and the top growing surface area. Two groups of experiments
are performed in this chapter: in group (1) the pocket aperture is increased and varied to
explore SCD growth as a function of pocket aperture size, and in group (2) SCD growth
is investigated where using newly designed holders another experimental run is added to
the diamond growth of the first run.
It was found in Chapter 5 that shallow pockets grew diamond substrates with jagged
edges and the largest top surface areas. On the other hand, deeper pocket grew thick
213

diamond substrates with smooth edges, flat surfaces, and reduced top surface areas. In
the research activities of Chapter 5 only the pocket depth was varied in order to
investigate diamond pocket holder growth while still producing a flat top surface with
smooth edges.
Based on the results of Chapter 5, it was decided to investigate the SCD growth where
the pocket depth was held constant while the pocket aperture was varied. In this chapter,
the holder depth was held constant at d = 2.6 mm or 2.9 mm for three different pocket
aperture widths of: w1 = 6.0 mm, w2 = 6.6 mm, and w3 = 7.0 mm. Also, a set of
complementary experiments were performed with constant pocket depths of d = 2.9 mm
again using three different apertures, i.e. w1, w2, w3. The substrate holder was redesigned
in order to accommodate the different aperture widths. The redesign had constant
dimensions, such as holder thickness (c = 3.2 mm) and inner pocket width (1 mm). The
pocket aperture width was the only parameter adjusted. See Figure 3.16 for the details of
the substrate holder.
In this chapter, the experimental data revealed that after increasing the pocket holder
aperture, the lateral growth was enhanced substantially. Significant improvements in the
lateral growth rate were observed at the early stage of each experiment. However, the
thick PCD which grew on the substrate holder and in the pocket walls influenced the
lateral growth. It was difficult to prevent the growth of PCD on the substrate holder. In
order to avoid or minimize the PCD growth on the substrate holder and there by interrupt
the lateral SCD growth on the growing SCD substrate, a set of re-growth experiments
were performed. Here, one more growth step is added to the smooth and large diamond
top surface of the first growth step. A few selected, grown SCD substrates were used for

214

re-growth experiments, and a new pocket holder was designed to allow the expansion of
the top surface area. The pocket holder depth was increased from d= 2.6 mm to d= 4.0
mm and the aperture width from w= 7.0 mm to w= 8.2 mm. At the end of the first growth
step, the grown diamonds were taken out of the substrate holder, were acid cleaned and
then were placed in the new pocket holder design for the re-growth experiments. In these
additional growth experiments, the diamond top surface was further enlarged.

6.2 Experimental Details: Variation of the Pocket Aperture
The set of experiments were carried out in the MPACVD reactor C configuration at
240 Torr and microwave power density of approximately 500 W/cm3. SCDs were grown
on 3.5 mm x 3.5 mm x ≈ 1.4 mm HPHT type Ib diamond seeds for 12 – 60 hours with H2
flow rates of 400 sccm and a methane concentration of 5%. During each experiment, the
substrate temperature was monitored every thirty (30) minutes and was held
approximately constant at 1020°C ± 5°C during each experiment. Figure 5.3 provides
examples of a typical growth process recipe. In order to prevent the growth a PCD rim
the growth process cycle was ended when the diamond started to emerge from and grow
outside of the pocket.
The pocket holder was redesigned from the holders used in Chapter 5 to
accommodate the different size pocket apertures. The redesign slightly adjusted the
holder thickness, c, for the pocket depths of d = 2.9 mm to allow the diamond growth to
take place within the appropriate substrate temperature window. During each process
cycle, a substrate pocket holder was placed in the reactor as shown in Figure 5.1 (b).

215

6.3 Experimental Results: Variation of the Pocket Aperture
6.3.1 SCD Growth versus time
Two groups of experiments that investigated the upward and outward growth of the
SCD at a constant pocket holder depth of d = 2.6 mm were performed. These experiments
used two different pocket apertures: w2 = 6.6 mm and w3 = 7.0 mm. The cycle varied from
12 hours to 55 hours, depending on the aperture. For w2= 6.6 mm the process cycle runs
times were 12, 16, 24, 30, 48, 50 and 55 hours. The process cycle runs times for w3= 7.0
mm were 12, 30 and 45 hours. Table 6.1 summarizes the different growth parameters
and output variables for each aperture size. For all the experiments presented, Ts was
kept constant at 1020°C throughout each individual process cycle run by adjusting the
microwave input power versus time. In all of these set of experiments, the growth process
was stopped before the top surface grew out of the pocket.
The linear growth rates that are summarized in Table 6.1 are the thickness gain
divided by the growth time. Thus, this growth rate is an average growth rate over each
entire process cycle for each experimental run. The lateral growth rates are calculated
with the dimensions of the outward growth length divided by the run cycle growth time.
See section 3.7.2 for details on the lateral growth calculations.
Optical microscope images from Table 6.1 are shown in Figure 6.1 (a) – (f) and Figure
6.2 (a) – (d). Figure 6.1 shows images of top and side views of the substrates that were
grown using aperture w2 = 6.6 mm. Figure 6.2 shows images of top and side views of the
substrates that were grown using aperture w3 = 7.0 mm. All the samples listed in Table
6.1 exhibited smooth and flat top surfaces without a PCD rim around the edges of the top
surfaces.
216

Table 6.1 Growth parameters for SCD growth using the w2 = 6.6 mm and w3 = 7.0 mm
pocket holder apertures at a constant pocket depth of d = 2.6 mm

Linear
Pocket
Process
Thickness growth
width Sample
cycle
gain (mm)
rate
(w)
(h)
(μm/h)
ACH54
12
0.335
27.1

6.6
mm

7.0
mm

Lateral
growth
rate
(μm/h)
48.0

Area Normalized
(mm2)
area
22.6

1.65

ACH56

16

0.450

28.0

38.3

23.2

1.70

ACH53

24

0.676

28.2

27.0

24.0

1.75

ACH52

30

0.828

27.6

22.1

24.3

1.78

ACH57

48

1.300

27.0

14.2

24.6

1.82

ACH58

50

1.440

28.9

14.0

24.6

1.80

ACH55

55

1.530

27.8

10.0

22.0

1.60

ACH79

12

0.348

29.0

52.2

23.6

1.73

ACH71

30

0.855

28.5

27.4

27.5

2.01

ACH78

45

1.310

28.8

20.0

28.7

2.10

Additionally, all the grown SCD have square crystal top surfaces. However, many of
the surfaces have jagged edges while a few surfaces have smooth edges. For example,
samples ACH54, ACH56, ACH53, ACH52, ACH57, ACH58, ACH55, ACH79, ACH71, and
ACH78 all have jagged edges. Sample ACH58 exhibits partially smooth and partially
jagged edges while ACH55 exhibits only smooth edges. When using pocket holder
apertures of w2= 6.6 mm and w3= 7.0 mm, the linear upward growth rate remained
unchanged between 28 – 29 μm/h while the top surface area slowly increased as a
function of the time up to 45 hours and then slightly decreased

217

Figure 6.1 Optical microscope top surface images of substrates listed in Table 6.1, for w2
= 6.6 mm (a) 12 hours, (b) 16 hours, (c) 30 hours, (d) 30 hours, (e) 48 hours, (f) 50 hours,
and (g) 55 hours. Note that ACH55 has smooth edges and a decreasing top surface.

218

It also was observed that the substrates were grown within the range from 12 to 55
hours using aperture w2 = 6.6 mm grew upward as a function of the time at approximately
the same average growth rate, i.e. an average linear growth rate of 28 – 29 µm/h.
However, the lateral growth rates varied as a function of the time from 48 µm/h at 12
hours to 10 µm/h at 55 hours. This experimental data clearly indicates the lateral growth
rate is faster at the early stages of the growth process. After 48 hours of the growth
process, the lateral growth slowed considerably, while the vertical linear growth continued
at ~ 27 – 28 µm/h. The grown substrates exhibited almost the same normalized top
surface normalized area of approximately 1.70 – 1.80 times.
The fast-lateral growth rates during the first stages of growth that are observed from
the data in Table 6.1 were first noted in work described in Chapter 4 sections 4.4 and 4.6.
In the experiments described in Chapter 4, the input power was increased during the early
stages of the process cycle in order to enhance the early lateral growth rates. The results
presented here in Tables 6.1 and 6.2 strengthen the observation of early fast lateral
growth rates if the coupled variable, input power/ substrate temperatures, is maintained
at the appropriate levels
Smooth edges were observed in substrates grown with longer process cycles, i.e. 50
and 55 hours. The maximum total outward lateral growth was calculated for those
substrates grown at 48 hours and 50 hours with an average normalized top surface area
of 1.81 times more than the initial HPHT diamond seed area. However, a decrease of
11% in the area gain was observed in the ACH55 substrate which grew with the longest
process cycle time of 55 hours. This is attributed to the thick PCD layer that grew over
time in the top of the substrate holder and around the pocket walls suppressing the SCD

219

outward growth. Thus, process cycles longer than 50 hours resulted in smaller final top
surface areas. Similar results were found in Chapter 5 by using the same pocket depth
(d = 2.6 mm) and a smaller aperture width of w1= 6.0 mm.
Figures 6.2 (a), (b) and (c) display smooth and flat surfaces for the three substrates grown
using a pocket holder with w3 = 7.0 mm and a pocket holder depth of 2.6 mm. These
substrates were grown with growth process cycles of 12, 30 and 45 hours. The measured
average total linear growth rate was 29 μm/h. As in the previous case, the faster lateral
growth rates were for the substrates grown for the shortest time of 12 hours. It was
measured at 52.2 μm/h. The total average lateral growth rate of 20 μm/h was calculated
for the substrate grown for 45 hours, which was the longest experimental run. As a result,
expanded areas as large as 28.7 mm2 were obtained in a single run. For this pocket holder
configuration, i.e. d = 2.6 mm and w3 = 7.0 mm, the total area increased up to 2.1 times
of the initial HPHT diamond seed area. When using pocket holder apertures of w 2= 6.6
mm and w3= 7.0 mm, the linear growth rate remained unchanged between 28 – 29 μm/h
while the top surface area increased as a function of the time up to 50 hours.
For a better understanding of how the pocket aperture size influences the lateral
growth of the SCD, a separate additional set of experiments were performed using a fixed
pocket holder configuration of d = 2.9 mm and w3 = 7.0 mm. The results from the fourdifferent growth process are summarized in Table 6.2 and Figure 6.3 (a) – (c).

220

Figure 6.2 Optical microscope top surface images of substrates listed in Table 6.1, for w3
= 7.0 mm (a) ACH79 after 12 hours of growth, (b) ACH71 after 30 hours, and (c) ACH78
after 45 hours of growth.

221

Table 6.2 Growth parameters for SCD growth using w3 = 7.0 mm pocket holder aperture
at pocket depth of d = 2.9 mm
Linear Linear
Pocket
Process
Thickness growth growth Area Normalized
width Sample
cycle
gain (mm)
rate
rate
(mm2)
area
(w)
(h)
(μm/h) (μm/h)
ACH73
12
0.309
25.8
52.4
23.6
1.73
7.0

ACH67

40

1.00

25.1

24.1

30.6

2.24

mm

ACH70

45

1.12

25.0

20.9

30.1

2.20

ACH72

50

1.23

24.6

20.1

31.5

2.30

The experiments were performed using constant T s deposition recipes, i.e. 1020 ±
5°C. The growth time was varied at 12, 40, 45 and 50 hours for single individual runs.
The grown substrates exhibited smooth and flat surfaces with jagged edges and square
top surfaces. These substrates grew upward as a function of the time at an average linear
growth rate of ≈ 25 µm/h. As a result, expanded areas were obtained as large as 31.5
mm2, and the total area gain increased to 2.3 times of the initial HPHT diamond seed
area. This increase is the highest obtained for CVD growth on a standard (3.5 mm x 3.5
mm) HPHT diamond seed in a single run.
The results from Chapter 5 showed grown substrates that displayed different crystal
shapes, i.e. square, octagon, circle, depending on the pocket depth. Contrary to those
results, all the substrates listed in Table 6.1 and Table 6.2 that were grown with apertures
of 6.0 mm and 7.0 mm displayed the same final square crystal shape.

222

Figure 6. 3 Optical microscope top surface and side view of images of substrates listed
in Table 6.2, for w3 = 7.0 mm a constant pocket depth of d= 2.9 mm (a) 45 hours, (b) 40
hours and (c) 50 hours. Note: that the substrate ACH72 have a dark spot on the left
bottom side it is a crack on the top surface that was produced at the experimental run
when the substrate was removed from the pocket.

It was observed for the data from Table 6.1 and 6.2 that the total average lateral
growth rate decreases as a function of the growth time. Figure 6.4 summarizes the total
lateral growth rate for all the pocket apertures (including data from Chapter 5) at 12 and
45 hours of growth time. At 45 hours, the lower growth rates are from, the narrow pocket
aperture, i.e., 6.0 mm and the lateral growth rate gradually increases as the aperture size
is increased. Using the same pocket aperture, i.e. 6.0 mm, the total lateral growth rate
decreases from 24.9 µm/h to 7.2 µm/h as the run is extended from 12 hours to 45 hours.

223

A higher initial lateral growth rate of 52.2 µm/h was measured for the wider pocket
aperture, i.e. 7.0 mm at 12 hours of growth time. For all shorter growth times, these results
indicated that the lateral growth rate is significantly larger, i.e. it more than doubles, as
the pocket aperture is increased.

Figure 6.4 The total average lateral growth rate versus pocket aperture at 12 and 45 hours
of growth time

Results from Chapter 5 and these results indicated that the linear vertical growth rate
becomes slower with deeper pocket holders. Also, the top surface areas enlarged as the
pocket aperture size increased. Based on these experimental results it is speculated that

224

the diamond could have grown even larger with enlarged surface areas and with slight
thickness gains by using deeper pockets (3.0 mm – 3.3 mm) and wider apertures (7.3
mm – 7.6 mm).

6.3.2 SCD growth versus growth time, pocket width, and pocket depth: the
evolution of the crystal versus time.
To better understand how the pocket aperture plays a significant role in the lateral
growth of CVD diamond substrates, the results of the normalized area gain versus growth
time are plotted in Figure 6.5. All the substrates listed in Table 6.1 and Table 6.2, as well
as selected results from Chapter 5 for a pocket holder with d = 2.6 mm and w = 6.0 mm,
are summarized in Figure 6.5.
The experimental data points are grouped by the pocket width at a constant pocket
depth of 2.6 mm. Two pairs of eight substrates, which were grown in separate pockets
with w1 = 6.0 mm and w2= 6.6 mm, are were plotted and represented by the red and the
blue curves, respectively. Three substrates which were grown in a pocket with of w3 = 7.0
mm, and d = 2.6 mm are also plotted and represented by the orange curve. Also included
are a set of four experiments grouped by fixed pocket geometry of d= 2.9 mm and w= 7.0
mm (the black curve in the figure). All the curves begin at t=0 and grow from the original
HPHT seed surface area of 13.7 mm2: Note that at t = 0, i.e. at the start of the experiment,
the normalized area is one.

225

Figure 6.5 Normalized area gain versus growth time at different pocket configurations
from experimental data provided in Table 6.1, Table 6.2, and selected results of Chapter
5 grown in pocket depth of 2.6 mm.

It can be readily observed that the pockets with wider apertures reached larger area gains
than the narrower pockets apertures. The pockets with an aperture width of w 3= 7.0 mm
yield the largest normalized area gain. This particular pocket holder geometry (at d = 2.6
mm) can grow SCD with area gains up to 2.1 times before their top surfaces start to
decrease.

226

Figure 6.6 Plot summarizing the normalized lateral CVD area gain versus growth time for
different pocket apertures and depths for the experimental data listed in Tables 6.1, Table
6.2, and selected data from Table 5.2.

The deeper and wider pocket geometry, i.e. d = 2.9 mm and w3= 7.0 mm, can grow
SCD with a top surface area gain up to 2.3 times in a single run. The narrow apertures
and the longer runs resulted in smaller final SCD surface area gains. Opposite to this
trend, the wider apertures and the longer runs resulted in a larger top surface area gains.
The experimental data displayed in Figure 6.6 is broken into three growth time regions.
Growth region I describe the initial growth over the first 12 -15 hours where lateral growth
rates are high, i.e. > 25 µm/h and exceed the vertical growth rate. Rapid lateral growth
227

occurs in this time region for all four pocket holder geometries and the fastest lateral
growth rate, i.e. 52 µm/h was measured for the wider pocket apertures. Growth region II
is called the “crystal thickness growth” region. Here, the lateral growth is significantly
slowed over time by the presence of the pocket walls, while the linear growth continues
at a rate of 27 – 29 µm/h. Also, during this time a thick PCD layer starts to grow over the
substrate holder. Growth region III is identified as the PCD overgrowth region. This growth
region is dominated by the growth of a thick PCD layer on the top substrate holder. In
region II the PCD layer on the substrate holder thickens with time and after 45 hours; i.e.
at the beginning of region III growth, the PCD starts to grow into and fill the aperture
opening. In this region, the effective aperture width starts to decrease and thus the top
growing surface of the SCD decreases. At the end of a long run that extends into region
III, the total top surface area can be less than the initial HPHT seed area. The formation
of PCD on the substrate holder thus limits the growth of an ever expanding SCD top
surface. This indicates that the growth of PCD on the holder should be avoided.
An example of the growth behavior in these three regions is presented in Figure 6.7.
Photographic (see Chapter 7 for information on the photographic technique) images were
captured in each growth region to evaluate the SCD lateral growth. In Figure 6.7 the
opaque yellow polygon layer represents the original HPHT seed inside the pocket. Six
polygons were copied and carefully overlaid on each growth stage of the diamond. The
orange polygon layer outlines the growing SCD during the indicated time frames of 12,
24, 40, 50 and 55 hours. As can be seen from the photographs, the lateral growth occurs
at the early stage of growth, i.e. growth region I. In growth region II, after 40 hours
onwards the lateral growth is slowed, and the linear vertical growth remains steady at ~

228

28 µ/hr. Finally, in growth region III, the overgrowth of PCD on the substrate holder and
into the walls of the pocket is observed which either interrupted the lateral growth or
decreased it. This is clearly confirmed from the analysis of the photos that is presented in
Chapter 7.

Figure 6. 7 Photographic images in the frame time of 0, 12, 24, 40, 50 and 55 hours. The
opaque yellow polygon layer represents the HPHT seed inside the pocket. The orange
polygon layer identifies and outlines the grow SCD during the indicated time frames.

Figure 6.8 displays a plot of the lateral growth rate (V110) and the vertical growth rate
(V100). The experimental data in the curves in this figure is calculated from the data
229

displayed in Fig. 6.7. It displays the growth rate ratio between the outward, lateral surface
growth, represented by V110 and upward surface growth, represented by V110 and upward
surface, represented by V100. The ratio between both rates was calculated with the data
in Figure 6.7 and from Table 6.1 and 6.2. The substrates that were grown in the wider
pocket apertures, i.e. 7.0 mm lead to an initial fast and uniform growth of over the entire
surface resulting rapidly expanded the surface area at the beginning of the process cycle.
For example, the faster growth was observed for a substrate that grew in the wider pocket
(V110/V100 = 1.7).

Figure 6. 8 The ratio between lateral growth rate, V110, and vertically growth rate, V100,
versus growth time for three different pocket holder apertures.

230

In contrast, the substrates that grew in the narrow pocket (w = 6.0 mm) showed a
slower initial growth rate. After the initial fast lateral growth period during region I the
lateral growth rate decreased as a function of the time. This behavior was true for all
apertures. The narrow apertures and the longer runs resulted in smaller final SCD surface
areas. Opposite to this trend, the wider apertures and the longer runs resulted in a larger
top surface area. This is caused by the PCD growing on the substrate holder and into the
pocket walls as a function of the time in region II and III as a function of the time. Hence,
the longer runs make the effective pocket aperture decrease over time resulting in smaller
final SCD top surfaces areas. The wider apertures keep the aperture wide enough to
allow for more outward growth and eventually the further enlargement of the final surface
area before the PCD growth modifies the process.
In region III the diamond surface area could either continue to increase or start to
decrease versus time depending on the aperture. The data indicate that after 45 hours
(growth region III) the diamond surface area is determined by the aperture size. The
lateral growth shows a trend that implies the continued to lateral growth until the top
surface area fills the wider apertures. The black arrows in Figure 6.6 are used to represent
ether an increasing (up) or a decreasing (down) top surface area growth trend.

231

Figure 6.9 Plot summarizing the normalized lateral CVD area gain versus vertical
thickness gain for different pocket holder geometries for the experimental selected data
from Table 6.1 and Table 6.2 and the results of Chapter 5. Short run data and long run
data from this section are indicated by the grown and res data points respectively.

In order to compare, the recent experimental data with the results found in Chapter 5,
Figure 6.9 shows a reference plot of the normalized lateral area gain versus the vertical
thickness gain. This figure was used in Chapter 5 (Figure 5.10). In Figure 6.9 all the
experimental data points from Chapter 5 are grouped by pocket depth and are
232

represented by all black polygons. Selected data points from Table 6.1 and Table 6.2 are
represented either: (1) by the red polygons which represent the long run data, or (2) by
the green polygons which represent the short run data from Tables 6.1 and 6.2.
For all the experiments presented in this chapter, the longer runs that extend into
region III ends up with the smallest surfaces areas. See for example ACH55. The shorter
runs up to 45 hours end up with enlarged surfaces areas. This is caused by the PCD
growth on the substrate holder. See for example ACH67 and ACH78. The experimental
data for the w= 7.0 mm holder aperture clearly indicates a significant improvement in the
enlargement of the final diamond top surface.
A comparison of the short run and long data points indicated the growth in the three
growth regions: (1) a quick lateral growth (see ACH54 and ACH79) over the first 0.3 mm
of thickness, (2) a slower lateral growth over the next 0.7 mm of thickness growth and (3)
a decrease in top surface area over the next ~ 0.5 mm of thickness growth. This already
has been indicated in Figure. 6.6. The short run data points indicate that the wider
apertures allow the rapid initial growth of enlarged surface areas.

6.3.3 Summary Remarks: Variation of the Pocket Aperture
Enlarged diamond top surfaces were grown in variable pocket holder apertures as
one continuous growth process. It was discovered that the long growth experiments could
be divided into three growth regions: (1) region I where the lateral growth was high, (2)
region II with (a) a steady thickness (vertical) growth and (b) a diminished lateral growth
that is limited by the proximity of the pocket walls and (3) region III where the growth is
effected and limited by the growth of PCD on the substrate holder and into the pocket
233

aperture itself. Additionally, it was found, the linear and lateral growth rate decrease as
the pocket depth increased from 2.6 mm to 2.9 mm. Those results are similar all holder
apertures, i.e. 6.0 mm, 6.6 mm and 7.0 mm. More specifically wider pocket apertures
have a significantly enhanced lateral growth rates and larger areas gain. Narrow pocket
apertures have an increased top surface area up to 45 hours of the growth process. After
45 hours, a trend to decrease the lateral growth was observed in all the substrates grown
in a 2.6 mm pocket depth and apertures of 6.0 mm and 6.6 mm. See example blue and
red curves in Figure 6.5. The larger surface areas were grown on the pocket holder with
d = 2.9 mm and w = 7.0 mm. The normalized lateral area gain measured was 2.3 times
that the initial HPHT diamond seed.

6.4 Experimental Details: SCD Re-growth and Expanding Surfaces
The set of experiments were carried out in the MCPR reactor C configuration at 240
Torr and microwave power density of 700 W/cm 3. SCDs were grown on smooth and flat
CVD SCD top surfaces with areas of approximately of 5 mm x 5 mm x ≈ 1.4 mm for 40 –
50 hours with H2 flow rates of 400 sccm and a methane concentration of 5%. The
substrate temperature was carefully monitored and was held approximately constant at
1020°C ± 5°C during each experiment. The growth process cycles were terminated when
the diamond started to grow outside of the pocket. The experiments presented here are
exploratory in nature and do not represent the results from a complete and optimized
investigation.
A new pocket holder was designed to accommodate the larger grown substrates. The
substrate holder dimensions were selected based on the earlier results from Chapter 6,
234

i.e. the larger aperture pocket holder designs. The design adjusted the holder thickness,
c, the pocket depth, and the pocket aperture (c = 4.5 mm, d = 4.0 mm and w = 8.2 mm).
This design is displayed in Figure 3.18. The dimensions of the pocket holder allow that
the diamond growth takes place within the appropriate substrate temperature window.
During each process cycle, a pocket holder was placed in the reactor as shown in Figure
5.1 (b).
Long growth process cycles are imperative to grow thick and large volume SCDs.
However, as has been discussed in Chapter 5 and section 6.3.2, the high growth rate of
PCD on the substrate holder disturbs the SCD growth. Therefore, the re-growth is an
alternative growth method to one continuous SCD growth process. The re-growth process
involves adding an additional growth step on the top of one already grown substrate with
smooth and enlarge top surface areas. Before the second growth step, a new pocket
holder was designed and then was used in the second growth step. After the 1st growth
run was completed, the substrates were cleaned and analyzed. The cleaning process
was described in section 3.4.1.

235

Figure 6.10 Schematic of the re-growth process. After the 1st growth step process, the
grown substrate is placed in an optimized substrate holder.

Figure 6.10 shows a flow diagram of this process. This figure describes the regrowth process using an optimized pocket holder design.

6.5 Experimental Results: SCD Re-growth and Expanding Surfaces
6.5.1 SCD Re-growth Criteria
In order to continue expanded areas without producing defect on the new surfaces,
the 2nd growth entirely depends on the 1st grown surface. The first diamond growth step
must satisfy certain growths criteria. The following criteria were listed in order to continue
expanding the top surface at the same time that the smoothness and uniformity of the
diamond surfaces are maintained. The first growth step should have:

236

1. Smooth and flat surface, to avoid the propagation of the defects in the second
growth step
2. With no evidence of PCD and with no jagged edges
3. Evenly grown top surfaces, i.e. ~ 4.5 mm x 4.5 mm
4. Due to the pocket design, the thickness of the substrate should be more than 1.0
mm.
An example of two substrates that do not meet these growth criteria are shown in
Figure 6.11 and Figure 6.12. Figure 6.12 displays optical microscope images of the grown
substrate ACH26 after the 1st and 2nd growth steps. ACH26 was first grown in one step
for 46 hours, with a thickness of 1.22 mm and a growth rate of ~29.1 µm/hr. The top
surface of the SCD substrate was uneven with no major defects on the surface and had
jagged edges. The dimensions of the first growth surface for this substrate was 24.08
mm2. Thus, the lateral surface area of the CVD SCD substrate increased by ~ 1.8 times
over the initial SCD seed surface area (i.e. 13.69 mm2).

237

Figure 6.11 Optical microscope images of (a) top view of ACH26 after the 1st growth step,
(b) top view of ACH26-2 after the 2nd growth step, and (c) side view of the ACH26-2.

During the second growth step, the surface was expanded up to 34.8 mm2. This is ~
2.6 times over the initial HPHT seed and 1.4 times over the 1st growth step. It is an
increase of growth area of 40% over the 1st growth step. It was observed significant
defects on the top surface, and the edges maintained their jagged appearance. It is
considered that the uneven surface was propagated up to cause a crack on the top
surface, which is identified as the black band on the top surface. See Figure 6.11 (b). The
2nd growth was performed for 45 hours at a growth rate of 25.4 µm/h.
Another example is shown in Figure 6.12. ACH36 was grown in one step for 43 hours,
with a thickness of 1.0 mm and at a growth rate of ~23 µm/hr. The top surface of the CVD
SCD substrate is a smooth and flat surface, with no evidence of PCD and with smooth
edges. The dimensions of the growth surface of this substrate was 15.2 mm2. Thus, the
lateral surface area of the CVD SCD substrate increased by ~ 1.1 times over the initial
SCD seed surface area (i.e. 13.69 mm2).

238

After 30 hours of the 2nd growth step, the substrate exhibited almost the same
morphology of the 1st growth. The substrate ACH36-2 displayed smooth and flat surface.
The top surface area increased up to 19.5 mm 2. It is ~ 1.28 times over the initial HPHT
seed and just 1.42 times over the 1st growth step.

Figure 6. 12 Optical microscope side view images of ACH36 after the (a) 1st growth step
and (b) 2nd growth step.

These examples demonstrate the importance of conducting the re-growth process on
expanded, smooth and flat surfaces to enable the continuation of growing expanding
surfaces without major defects.

6.5.2 SCD Re-growth – Expanded surfaces
Three different re-growth experiments were performed to continue expanding the top
diamond surfaces. These experiments used a constant temperature of 1020°C ±5°C over
239

the growth process and used a 4.0 mm deep pocket holder. Figure 6.13 (a) - (d) shows
examples of the PCD rimless and enlarged SCD areas of the first runs, ACH23, ACH27
and ACH73. Note the improvement on the top surface over the substrate ACH26 showed
in Figure 6.11. ACH23 was grown in one step for 40 hours, with a thickness of 1.20 mm,
and a growth rate of 30 µm/hr. Similarly, ACH27 was grown in one step for 45 hours, with
a thickness of 1.16 mm and a growth rate of ~ 26.6 µm/hr. Both SCD CVD substrates
were selected for regrowth.
The top surfaces of the both CVD SCD substrates are smooth and flat with no major
defects on the surface. The dimensions of the growth surface for these two substrates
were 23.28 mm2 and 22.1 mm2, respectively. Thus, the lateral surface area of the CVD
SCD substrate increased by ~ 1.7 times for ACH23 and ~ 1.6 times for ACH27 over the
initial SCD seed surface area. The outward growth of the SCD substrates is clearly seen
in Figure 6.13 (a) and (c).
The substrates were analyzed and then placed in the optimized pocket holder to
continue growth. As a result, as shown in Figure 6.12 (b) and (d), expanded top surfaces
have been successfully grown. ACH23-2 was re-grown for 50 hours at a growth rate of
28.7 μm/h and now had a total thickness of 2.64 mm. The red dotted lines in the
substrates represent the boundary between the first and second growth steps. After a
total time of 90 hours, the dimension of the total top surface was 34.22 mm2. The lateral
surface area of the CVD SCD substrate increased by 2.5 times. ACH27-2 was re-grown
for 45 hours, produced a total thickness of 2.32 mm, and had a growth rate of 25.2 μm/h.
The total growth time was 90 hours, the dimension of the total top surface was 27.56 mm2
and the total top surface increased by 2.1 times.

240

Figure 6. 13 Optical microscope side view of (a) and (b) substrate ACH23 after the 1st
and 2nd growth step, (a) and (b) substrate ACH27 after the 1st and 2nd growth step, and
(e) top view and (f) side view of substrate ACH34 after the 1st and 2nd growth step.

241

Finally, ACH34 was grown in one step for 20 hours, with a thickness of 552 µm, and
a growth rate of 27.6 µm/hr. The second growth step was conducted for 40 hours at a
growth rate of 27.5 µm/hr with a total thickness gain of 1.65 mm. The resulting CVD SCD
substrate increased by 2.2 times.
The normalized lateral area gain versus the vertical thickness gain for the re-growth
exploratory details plotted in Fig. 6.14. All the experiments that were regrown on uneven
and non-expanded surfaces developed defects and smaller surfaces, similar to ACH26
and ACH36. Those experiments were not included in the plot of Figure 6.14 because they
were re-grown out of the desired growth criteria.

Figure 6. 14 Normalized lateral area gain versus vertical thickness gain

242

In Figure 6.14, three regrown data points were plotted and are compared with the
growth on typical substrates described in Chapter 5. The results of these experiments
show the improvement of the thickness and the top surface area. The linear thickness
increased from approximately 1.5 mm to 2.6 mm, and the normalized area gain increased
from approximately 1.9 times to 2.5 times. The two dotted horizontal lines in the plot
represent the normalized original HPHT seed area, and the maximum area gain may be
achieved in one growth step with the pockets used in the experiments that are described
in Chapter 5. The pink line in Figure 6.14 is the best fit curve that was drawn in reference
[20] between experimental data for previous results reported.

6.5.3 Summary Remarks: SCD Re-growth and Expanding Surfaces
After a designing, a new substrate pocket holder, thick CVD SCD substrate up to 2.6
mm were grown when one additional growth step was added. The total top surfaces were
re-grown and had final surface areas of up to 2.5 times over the initial HPHT seed. This
new pocket holder geometry can re-grow SCD both vertically and horizontally. For this
pocket geometry, the final top surface is expanded while the smoothness remains the
same. These exploratory experiments have demonstrated that by adding a second growth
step and by growing at constant substrate temperature additional CVD SCD can be added
to the SCD grown in the first step. The newly grown SCD had high lateral growth rates
and further enlarged the final SCD surface areas.

243

CHAPTER 7

PHOTOGRAPHIC EVALUATION OF THE SCD SURFACE
GROWTH VERSUS TIME

7.1 Introduction
The broad investigation carried out in previous chapters revealed that the growth of
different SCD crystal shapes closely depends on growth time, substrate temperature, and
pocket geometries. It was also found that the growth of PCD on the substrate holder plays
a decisive role in the growth of large volume SCDs. Presently, the maximum SCD area
gain, achieved for a single growth run is 2.3 times and when a second growth run was
added the top surface could be expanded 2.6 times. Before research activities move
forward in the exploration of more pocket holder geometries, it is important to understand
and assess the growth evolution of the SCDs versus time in order to continue to enlarge
the crystal area even further than 2.6 times.
This chapter presents a new experimental technique that attempts to measure the insitu evolution of the lateral SCD growth versus time by using photographs and video
recording. Photographs are helpful to observe the lateral growth at a specific growth time.
244

The videos provide further information about the plasma behavior and also provide an
overview of actually how the SCD top surface evolves and grows versus time.
The data presented here was collected from a few experimental runs already shown
in previous chapters. Thus, here is an exploratory study based on high-resolution
photographs. The experimental data presented here does not include all the data that
was measured, but due to the large number of photos that were measured many of the
photos could be omitted without affecting the accuracy of the final results. The
photographic plasma diagnostic techniques presented here have proved very useful in
recording the plasma and the diamond growth behaviors versus time, and are expected
to be very usefully employed in future experimental microwave plasma studies at MSU.

7.2 Photographic Images Processing and Video Creation
High-resolution pictures of the growth process were taken using a CANON EOS 60D
8.2-megapixel semi-professional digital single-lens reflex camera with 60-mm lens. The
photographs were taken at a high angle through a screened window located adjacent to
the pyrometer. The window is at the top of the bell jar. The camera was located at a fixed
position on a tripod outside of the reactor at an incident angle of approximately 75° to
maintain a comparable fixed reference position. For comparison between the images, all
parameters were held constant for all experimental conditions. For example, the focal
length was kept at 60 mm, 1/5 focal aperture, ISO 100 and the camera was perfectly
focused on the substrate rather than the bell jar or cavity spy hole (screened window).
The experimental setup is displayed in Figure 7.1. The Canon EOS 60D camera has a
port on the right side of a remote switch with a type N3 terminal in which an appropriate
245

pinout was connected. The timer was set to take a picture every 2 minutes. See section
3.2.6 for more details about the camera settings. The video of each set of experiments
was created based on a sequence of the photographs captured during each growth
process.
All photographs were taken at the following operational condition: the pressure of 240
Torr, methane concentration of 5% and total gas flow of 420 sccm. The camera
automatically takes pictures after the plasma ignition.

246

Figure 7.1 Experimental setup for photographs and video recording for SCD lateral
growth analysis

The growth process continues operating in prescribed growth recipe, i.e. constant
substrate temperature, while the camera captures the SCD lateral growth over time. At
the end of the process, all images are processed and analyzed.

247

7.3 Data Acquisition
The process of data acquisition presented here is the result of software created
specifically for the analysis of the growth of diamond inside a pocket holder based on
image analysis. The camera setup and software are still under exploratory analysis and
development.
After processing the videos from 15 different experiments, the horizontal growth of the
SCD over time was clearly observed. Also, the initial fast growth of the PCD on the
substrate holder and into the pocket walls was captured after approximately 12 hours.
Thus, it was decided to select several images from different growth times to analyze the
SCD and PCD growth carefully. During the lengthy growth process experiments, the
number of images to select and analyze became larger. Typically, increments in sets of
20 photos were chosen until reaching a total of approximately 80 photos per experiment.
Hence, the software displayed the first photo where the area of the SCD and PCD are
selected. The process is repeated until the latest photo of the set. The complete process
is schematically represented in Figure 7.2.

248

Figure 7.2 Schematic of the data acquisition process. The photographed experiment
showed in the flow diagram is ACH78. A total of 1277 pictures were captured.

After the process is completed, the area of the SCD and PCD as a function of the time
is provided by the software. Since this is an image-processing software, all the
measurements are in pixels versus time. The equation to convert the area from pixels to
mm2 can be written as follow:

 Ai

  [mm2 pixels] 
 An

[mm 2 ]
 A f [ pixels]  ASCD [mm 2 ]
pixels

249

(21)

(22)

where Ai is the initial area measured on the seed at t = 0, i.e. ~ 3.6 mm x 3.6 mm for
one HPHT seed, An is the area at n time - measured in pixel - calculated by the software,
and Af is the final area. The result of this ratio is multiplied by the final area calculated by
the software. As a result, the area is converted to mm2 for subsequent comparison with
the area measured manually with calibrated instruments. This latter data was presented
in chapter 5 and 6. Also, the lateral growth rate can be calculated using these values. A
similar approach is used to calculate the area of the PCD that grew into the pocket. The
area is calculated using the difference between the area of the pocket aperture at t = 0
and the area of the pocket after the growth process is completed.

7.4 Preliminary Results
Here are presented the results of three experiments that grew in different pocket
apertures, such as ACH64, ACH55, and ACH78 that were grown in pocket apertures of
w1 = 6.0 mm, w2 = 6.6 mm and w3 = 7.0 mm, respectively. The photographed experiment
ACH64 was grown at constant substrate temperature for 48 hours at a growth rate of 26
um/h. The measured total thickness was 1.21 mm, and the total top surface area was
17.65 mm2. Thus, the substrate top surface area grew 1.3 times. The total number of
photos captured for this experiment was 1030. The experimental data for substrates
ACH55 and ACH78 were given in Table 6.1. A total amount of 430 and 1,277 photos were
collected for the experiments ACH55 and ACH78 respectively. All the images were
uploaded and processed by the software. Increments in sets of 15 photos were selected
to calculate the lateral growth area on the SCD and the PCD on substrate ACH64.
Similarly, increments in sets of 5 photos were chosen on substrate ACH55 and
250

increments in sets of 20 photos were selected on substrate ACH78. As a result, the initial
area, the final area and the lateral growth rate calculated by the software for the SCD in
summarized in Table 7.1. Also, the table displays the measured values for each grown
substrate, i.e. measured by an instrument. Table 7.2 summarized the initial area, the final
area and the inward PCD growth calculated by the software.

Table 7.1 Measured and calculated values of the SCD areas and SCD lateral growth
rate for ACH64, ACH55, and ACH78. An instrument manually collected the measured
values, and the software collected the calculated.

Sample

ACH64

ACH55

ACH78

V (110)

Method

Ai (mm2)

Af (mm2)

Measured

13.0*

17.6

6.3

Calculated

12.9

14.3

1.9

Measured

13.0*

22.0

10.0

Calculated

12.8

17.8

5.6

Measured

13.0*

28.7

20.0

Calculated

12.2

25.4

16.0

* This is the HPHT seed area.

251

(µm/h)

Table 7. 2 Calculated values of the PCD inward growth on the pocket wall for
experiments: ACH64, ACH55, and ACH78.

Sample

ACH64

ACH55

ACH78

Inward growth rate

Method

Ai (mm2)

Af (mm2)

Measured

36.0*

--

--

Calculated

36.8

20.43

15.4

Measured

43.6*

--

--

Calculated

43.5

22.5

16.8

Measured

49.0*

--

--

Calculated

48.2

33.1

13.8

(µm/h)

*These are the pocket aperture areas.

Note: that the pocket aperture width represents the initial PCD area and the pocket
aperture size after completed the experiment represents the final PCD area. Thus, the
PCD inward growth rates were calculated using the length of the PCD that grew into the
pocket. It is the difference between the pocket width before and after the growth process.
The software calculated the final PCD areas for all these exploratory analyses. These
values were not measured for any of the experiments. Hence, the PCD inward growth
rates presented here are based on the experimental data that was measured
photographically by the images and not by the final experimental measurements made
after a run was completed.
The values provided from the photo software calculations differs from the experimental
data that was measured presented in Chapters 5 and 6. The SCD final area had a
calculated percent error of 23% for the substrate ACH64 and ACH55 and 13% for ACH78.
The lateral SCD growth rate had a calculated percent error within 25 - 44% between the
252

experimental data and the image-based data. Despite the differences in values, the
growth regions identified in Chapter 6 can also be observed using this approach.

Figure 7.3 The outward SCD growth and the inward PCD growth versus growth time for
substrate ACH64.

The outward SCD growth and the inward PCD growth versus time for substrate
ACH64 are plotted in Figure 7.3. The images on the insert on the bottom side of the figure
are the photographs at the beginning and just before finishing the experiment. The basedimages data showed the fast SCD lateral growth on the early stage of the growth process
during approximately the first 15 hours. Then, the lateral growth of the SCD is slowed,
and the PCD starts the rapid growth into the pocket walls, and as results, the pocket
253

aperture decreases. See the red curve in Figure 7.3. Finally, the latest growth stage is
led by the faster PCD growth into the pocket and the reduction in the final SCD top surface
area. The image-based analysis for ACH64 confirmed the results found in Chapter 5 and
6 where the three growth regions were identified.

Figure 7.4 The outward SCD growth and the inward PCD growth versus growth time for
(a) substrate ACH64 and (b) substrate ACH78.

Figure 7.4 (a) and (b) shows the outward SCD growth and the inward PCD growth
versus time for substrate ACH64 and ACH78 respectively. Both figures followed the same
trend observed for ACH64. Note that the gap between the SCD area and the PCD area
decreases, due to the decrease of the pocket aperture. Also, this can be seen in the insert
images on the bottom side of each figure.

254

7.5 Summary Remarks
Photographs were automatically taken every two minutes of the entire growth process
in order to observe the SCD growth over time. The photos show the evolution of the lateral
SCD growth as a function of the time. Also, they show the overgrowth of the PCD on the
substrate holder and into the walls of the pocket.
It was confirmed by the images processing method that after ~ 45 hours the SCD area
decreased due to the growth of PCD into the pocket. The PCD inward growth rate into
the pocket was calculated as high as ~ 17 um/h.
The SCD lateral growth calculations based on the images showed low values in
comparison with the values measured manually and reported in chapter 5and 6. The
calculated error was up to 44% for the substrates grown with a longer run. These
inaccurate results are in part attributed to the angled camera position. Also, due to the
distortion of the image that might have occurred in looking through the quartz dome.
This image-based approach has been not optimized for accurate values. Further
research on the optimization the method and the relocation the camera will be needed.
However, this approach is expected to be very usefully employed in future experimental
microwave plasma studies at MSU.

255

CHAPTER 8

SUMMARY AND FUTURE WORK

8.1 General Summary
This thesis explored MPACVD the MPACVD growth of SCD using the “pocket holder”
geometry shown in Figure 3.15 and Figure 5.1. One important objective was to learn how
the pocket holder geometry influences the diamond deposition process. With this
objective in mind an experimental investigation was carried out which primarily varied the
pocket dimensions while all other input variables were held fixed. However, the
experimental plan deviated from this plan by making a few necessary exceptions and
allowing the following input variables also to vary: (1) in order to maintain a constant
substrate temperature, the input power had to be varied over time within each process
cycle run. (2) In one set of experiments, the substrate temperature was purposely varied
versus time in in order to achieve an improved final grown SCD. (3) The experimental run
time was allowed to vary from 10 hours to in a few cases to over 90 hours, and (4) in one
set of experiments the methane concentration was varied. The experiments varied the
pocket dimensions, like the depth and aperture width, one small step per experimental
run at a time. Using the MCPR reactor C (see Figure 3.14) single crystal diamond growth
256

was experimentally investigated over a large input variable space. Individual experimental
run times were varied between ~10 hours to 60 hours.
Especially, the following experimental variables were varied one variable at a time per
run: (1) the pocket depth, di, between 2.0 mm – 4.0 mm, (2) the pocket aperture width,
wi, between 6.0 mm and 7.0 mm and (3) the process cycle run time from about ~10 hours
to about 60 hours for the single run experiments and it was varied up to 90 hours for the
re-growth experiments, and in (4) one set experiments the methane concentration was
varied between 4% and 7%.
The following input variables were held constant: (1) excitation frequency was fixed at
2.45 GHz; (2) flow rate was held at ~ 420 sccm; (3) for most of the experiments the
methane concentration was fixed at 5%, and then one set experiment it was varied from
4% to 7%; (4) the operating pressure held constant at 240 Torr; and as indicated in the
paragraph above (5) the deposition time was varied between 10 hours and 60 hours. The
coupled input power/substrate temperature input variable was varied by adjusting the
substrate thermal cooling rate and the input power where the reactor operated between
the substrate temperature SCD deposition window of 950 – 1200 °C. Typically the
experiments required between 2.0 kW – 3.5 kW of input power to achieve the desired
substrate temperature.

257

8.2 Accomplishments
High quality, stress-free SCDs and PCD rim free SCD were grown with both an
expanding surface top surface and an increasing thickness [125], [159]. Linear growth
rates were between 24 – 30 µm/h and surface area gains of 2.1 – 2.3 times for a single
run and 2.5 times for an additional growth step. These experiments have demonstrated
that moving to high pressure, higher power density and low N2 growth conditions result in
high growth rates with improved quality and almost stress-free SCD. Thick, PCD rimless,
high quality, and low-stress SCD were grown. The etching experiments on the grown
SCD plates have displayed low defect densities. The analysis of such plates with etch pit
density measurements has shown a reduction in etch pit density in CVD SCD plates to ≈
104 etch pits/cm2 [160].
It was discovered that PCD grown rim could be eliminated by using the proper pocket
size, and by stopping the run before the SCD diamond emerged from the pocket. The
process has to be stopped before the growing SCD starts to emerge from the pocket. It
was found that a PCD rim could reappear when the methane concentrations were varied
above > 6% and also simultaneously shoot formation in the reactor also occurred under
these conditions.
It was found that SCD growth in a pocket could grow PCD rimless SCD with expanding
surface areas. The growth of thicker and larger area SCD without creating a PCD rim was
facilitated by real time controlling the process, i.e. varying power and substrate
temperature, versus time. Large area and PCD rimless growth could be optimized by
applying the appropriate input power/substrate temperature process recipe.

258

It was discovered that the evolution of the diamond crystal growth in the pocket growth
could be divided into three growth regions: (1) region I where the lateral growth was high,
(2) region II with (a) a steady thickness (vertical) growth and (b) a diminished lateral
growth that is limited by the proximity of the pocket walls and (3) region III where the
growth is affected and limited by the growth of PCD on the substrate holder and into the
pocket aperture itself.
In the initial growth region I, the lateral SCD growth can be high (~ 52 µm/h) if the
appropriate power/temperature recipe is applied. Typically, this requires that the Ts >
1000°C. Then the initial SCD growth is over the entire seed surface, and also the SCD
grows both vertically and horizontally. This leads a SCD growth with an initially expanding
surface. In this growth region, the growing SCD surface area expands horizontally at high
growth rates in the pocket and approaches and is limited by the pocket walls.
In growth region II is mainly vertical. The lateral growth continues although at a much
slower rate. The lateral growth is limited by the presence of the pocket walls. Vertical
growth continues throughout this region at about the same rate of typically, 27 – 29 µm/h.
In growth region III the SCD growth is affected and limited by the growth PCD on the
top surface of the substrate holder, and eventually into the aperture of the pocket and
there by reduces the effective deposition aperture area. The top surface area contracts
and become smaller; i.e. the lateral growth becomes negative. This is undesirable, and
the PCD growth on the substrate holder should be avoided.
As the pocket aperture is increased, the upward growth rate is almost steady at 27 29 µm/h. The growth rate slightly increases as the aperture width increases from 6.0 mm
to 7.0 mm. Upward growth rate only slightly increases as time increases from 12 – 50

259

hours, and decreases at 55 hours. In all the experiments, the initial (10 – 15 hour) lateral
growth rate is very high, for example, 48 – 52 µm/h, in the early phases of growth in region
I. In region II the lateral growth rate significantly slows as the lateral growth is inhibited by
the presence of the walls of the pocket.
Finally, this thesis activity introduces a new photographic method to evaluate the top
surfaces versus growth time. The photographic plasma diagnostic techniques have
proved very useful in recording the plasma and the diamond growth behaviors versus
time. The results of the growth evaluation present significant differences with the
measured manually results. The accuracy of these first photo measurement results
should be improved in future diamond growth research activities at MSU. More work
needs to be done in order to employ this method in future microwave plasma studies.

8.3 Publications Resulting from this Dissertation Research
The detailed experimental SCD growth in a fixed pocket holder geometry at variable
substrate temperature and variable microwave input power as a function of the time
described in this dissertation has been published [125].
The detailed experimental SCD growth in a variable pocket holder depth at a constant
substrate temperature versus time that is described in this dissertation has been
published [159].

260

8.4 Future Work
Based on the current results presented in this dissertation, there are a number of
recommendations that should be investigated further. Listed below are examples of future
work:
1. Explore the SCD growth on larger initial starting seeds. For example, the future
experiments should start with 7.0 mm x 7.0 mm seeds and bigger starting pockets.
2. Explore SCD growth with a series of three to ten short 10 – 15-hour experiments

that use aperture widths and depths that favor the large initial lateral growth rates
and growing relatively thin CVD SCD but with large area top surfaces. Each
experimental run would deposit about a 0.3 – 0.5 mm thick CVD substrate with a
top surface that is expanding at the rate of a surface gain of 1.5 – 1.7 per run.
3. Explore the growth process at even deeper and wider pocket, such as d = 3.3 mm

and w= 8.0 mm. This would further increase the lateral growth rate in shorter runs.
4. Re-growth in surfaces with expanded areas more than 2.3 times. It could allow
further increased the top surface.
5. Investigate the growth process in a pocket holder at higher pressures, such as 450
– 500 Torr. The discharge power density and species densities would increase
further, which might increase the final diamond size, increase the diamond growth
rate and could improve the diamond quality even further.
6. Explore the photographic evaluation method. This would be useful to evaluate insitu the lateral diamond growth. This suggests re-designing the cavity shell and
installing a camera closely and perpendicular to the top of the seed.

261

REFERENCES

262

REFERENCES

[1]

J. J. Gracio, Q. H. Fan, and J. C. Madaleno, “Diamond growth by chemical vapour
deposition,” J. Phys. D Appl. Phys. J. Phys. D Appl. Phys, vol. 43, no. 43, pp.
374017–374017, 2010.

[2]

E. W. John Wilks, Properties and Applications of Diamond, 1st ed. 1991.

[3]

R. S. Balmer et al., “Chemical vapour deposition synthetic diamond: materials,
technology and applications.,” J. Phys. Condens. Matter, vol. 21, no. 36, p. 364221,
2009.

[4]

H. O. Pierson, “HANDBOOK OF CARBON , GRAPHITE , DIAMOND AND
FULLERENES by.”

[5]

M. Gabrysch, “Electronic Properties of Diamond.”

[6]

A. S. Barnard, The Diamond Formula. Diamond Synthesis: a gemmological
perspective. Oxford: Butterworth-Heinemann, 2000.

[7]

J. E. Field, “The Properties of Natural and Synthetic Diamond,” Cryst. Res.
Technol., vol. 28, no. 5, p. 602, 1993.

[8]

M. SSchreck, J. Asmussen, S. Shikata, J.-C. Arnault, and N. Fujimori, “Large-area
high-quality single crystal diamond,” MRS Bull., vol. 39, no. June, pp. 504–510,
2014.

[9]

F. Silva, K. Hassouni, X. Bonnin, and A. Gicquel, “Microwave engineering of
plasma-assisted CVD reactors for diamond deposition.,” J. Phys. Condens. Matter,
vol. 21, no. 36, p. 364202, 2009.

[10] Y. Mokuno, A. Chayahara, Y. Soda, Y. Horino, and N. Fujimori, “Synthesizing
single-crystal diamond by repetition of high rate homoepitaxial growth by
microwave plasma CVD,” Diam. Relat. Mater., vol. 14, no. 11–12, pp. 1743–1746,
2005.
[11] Y. Mokuno, A. Chayahara, Y. Soda, H. Yamada, Y. Horino, and N. Fujimori, “High
rate homoepitaxial growth of diamond by microwave plasma CVD with nitrogen
addition,” Diam. Relat. Mater., vol. 15, no. 4–8, pp. 455–459, 2006.
[12] M. Muehle, “PROCESS DEVELOPMENT TO SOLVE CRITICAL CHALLENGES
MOVING TOWARDS TWO INCH SINGLE CRYSTALLINE DIAMOND WAFERS,”
Michigan State University, 2017.
[13] C. S. Yan, Y. K. Vohra, H. Mao, and R. J. Hemley, “Very high growth rate chemical
263

vapor deposition of single-crystal diamond,” Proc. Natl. Acad. Sci., vol. 99, no. 20,
pp. 12523–12525, 2002.
[14] Q. Liang et al., “Recent advances in high-growth rate single-crystal CVD diamond,”
Diam. Relat. Mater., vol. 18, no. 5–8, pp. 698–703, 2009.
[15] E. V. Bushuev et al., “Express in situ measurement of epitaxial CVD diamond film
growth kinetics,” Diam. Relat. Mater., vol. 72, pp. 61–70, 2017.
[16] J. Lu, Y. Gu, T. A. Grotjohn, T. Schuelke, and J. Asmussen, “Experimentally
defining the safe and efficient, high pressure microwave plasma assisted CVD
operating regime for single crystal diamond synthesis,” Diam. Relat. Mater., vol. 37,
pp. 17–28, 2013.
[17] D. G. Goodwin, “Scaling laws for diamond chemical-vapor deposition. I. Diamond
surface chemistry,” J. Appl. Phys., vol. 74, no. 11, pp. 6888–6894, 1993.
[18] Y. Gu, J. Lu, T. Grotjohn, T. Schuelke, and J. Asmussen, “Microwave plasma
reactor design for high pressure and high power density diamond synthesis,” Diam.
Relat. Mater., vol. 24, pp. 210–214, 2012.
[19] “20161108 Issued Patent TEC2007-0078-01.pdf.” .
[20] S. Nad, Y. Gu, and J. Asmussen, “Growth strategies for large and high quality single
crystal diamond substrates,” Diam. Relat. Mater., vol. 60, pp. 26–34, 2015.
[21] K. E. S. and J. P. Dismukes, Synthesis diamond: emerging CVD science and
technology, 1st ed. 1997.
[22] A. G. Inc, “No Title.” [Online]. Available: http://www.galleries.com/.
[23] Jeremy Harvey, “No Title,” University of Bristol, Physical and Theoretical Chemistry.
[Online]. Available: http://www.chm.bris.ac.uk/pt/harvey/gcse/covalent.html.
[24] J. E. Field, The Properties of Diamond, 1st editio. Academic Press Inc., 1979.
[25] F. P. BUNDY, H. T. HALL, H. M. STRONG, and R. H. WENTORF, “Man-Made
Diamonds,” Nature, vol. 176, no. 4471, pp. 51–55, 1955.
[26] W. G. Eversole, “No Title,” United States Patent No. 303187 and 303188 (1962)No.
303187 and 303188, 1962.
[27] B. V. Spitsyn, L. L. Bouilov, and B. V. Derjaguin, “Vapor growth of diamond on
diamond and other surfaces,” J. Cryst. Growth, vol. 52, no. PART 1, pp. 219–226,
1981.

264

[28] S. Matsumoto, Y. Sato, M. Kamo, and N. Setaka, “Vapor Deposition of Diamond
Particles from Methane,” Jpn. J. Appl. Phys., vol. 21, no. Part 2, No. 4, pp. L183–
L185, 1982.
[29] S. S. Matthias Schreck , Jes Asmussen and and N. F. Jean-Charles Arnault,
“Large-area high-quality single crystal diamond,” MRS Bull. supl. CVD Diamond-Research, Appl. Challenges, vol. 39.6, p. 504–510., 2014.
[30] D. D. Hof, J. A. Dinsmore, S. Barber, R. Suhr, and T. R. Scofield, “Advocacy : The
T . R . A . I . N . E . R . Model,” vol. 2, no. 1, pp. 15–28, 2009.
[31] C. Station, “General electric company schenectady, new york,” Power, vol. 3, no.
4, pp. 2030–2036, 1987.
[32] H. M. BUNDY, F. P. Strong, “Behavior of Metals at High Temperatures and
Pressures,” Solid State Physics-Advances Res. Appl., vol. 13, pp. 81–146, 1962.
[33] B. W. Paul M, “Diamond thin films: a 21st-century material,” Phil. Trans. R. Soc.
Lond. A, vol. 358, no. Tennant 1797, pp. 473–495, 2000.
[34] S. Matsumoto, Y. Sato, M. Tsutsumi, and N. Setaka, “Growth of diamond particles
from methane-hydrogen gas,” J. Mater. Sci., vol. 17, no. 11, pp. 3106–3112, 1982.
[35] J. E. B. F. G. Celii, “Diamond Chemical Vapor-Deposition,” no. 42, pp. 643–684,
1991.
[36] M. Kamo, “Synthesis of diamond from gas phase and its properties,” Synthesis
(Stuttg)., vol. 7, no. 2, pp. 7–12, 1990.
[37] T. F. Diamond, C. Vapour, and D. Methods, “Thin Film Diamond by Chemical
Vapour Deposition Methods,” pp. 21–30, 1992.
[38] K. E. Spear, “Diamond -Ceramic Coating of the Future,” Perspective, vol. 91, no.
72, pp. 171–91, 1989.
[39] M. Frenklach, “The role of hydrogen in vapor deposition of diamond,” J. Appl. Phys.,
vol. 65, no. 12, pp. 5142–5149, 1989.
[40] M. N. R. Ashfold, P. W. May, C. A. Rego, and N. M. Everitt, “Thin film diamond by
chemical vapour deposition methods,” Chem. Soc. Rev., vol. 23, no. 1, p. 21, 1994.
[41] S. J. Harris, “Mechanism for diamond growth from methyl radicals,” Appl. Phys.
Lett., vol. 56, no. 23, pp. 2298–2300, 1990.
[42] S. Farhat, C. Findeling, F. Silva, K. Hassouni, and A. Gicquel, “Role of the plasma
composition at the surface on diamond growth,” J. Phys. IV Fr., vol. 8, p. Pr7-391Pr7-399, 1998.

265

[43] D. Takeuchi, H. Watanabe, S. Yamanaka, H. Okushi, and K. Kajimura,
“Homoepitaxial diamond films grown by step-flow mode in various misorientation
angles of diamond substrates,” Diam. Relat. Mater., vol. 9, no. 3, pp. 231–235,
2000.
[44] M. Kamo, H. Yurimoto, and Y. Sato, “Epitaxial growth of diamond on diamond
substrate by plasma assisted CVD,” Appl. Surf. Sci., vol. 33–34, no. C, pp. 553–
560, 1988.
[45] K. Hayashi, S. Yamanaka, H. Okushi, and K. Kajimura, “Stepped growth and
etching of (001) diamond,” Diam. Relat. Mater., vol. 5, no. 9, pp. 1002–1005, 1996.
[46] D. Takeuchi et al., “High quality homoepitaxial diamond thin film synthesis with high
growth rate by a two-step growth method,” Diam. Relat. Mater., vol. 8, no. 6, pp.
1046–1049, 1999.
[47] H. Watanabe, D. Takeuchi, S. Yamanaka, H. Okushi, K. Kajimura, and T.
Sekiguchi, “Homoepitaxial diamond film with an atomically flat surface over a large
area,” Diam. Relat. Mater., vol. 8, no. 7, pp. 1272–1276, 1999.
[48] C.-S. Yan and Y. K. Vohra, “Multiple twinning and nitrogen defect center in chemical
vapor deposited homoepitaxial diamond,” Diam. Relat. Mater., vol. 8, no. 11, pp.
2022–2031, 1999.
[49] A. Tallaire, J. Achard, F. Silva, R. S. Sussmann, and A. Gicquel, “Homoepitaxial
deposition of high-quality thick diamond films: Effect of growth parameters,” in
Diamond and Related Materials, 2005, vol. 14, no. 3–7, pp. 249–254.
[50] G. Bogdan et al., “Growth and characterization of near-atomically flat, thick
homoepitaxial CVD diamond films,” Phys. Status Solidi Appl. Mater. Sci., vol. 202,
no. 11, pp. 2066–2072, 2005.
[51] T. Teraji and T. Ito, “Homoepitaxial diamond growth by high-power microwaveplasma chemical vapor deposition,” J. Cryst. Growth, vol. 271, no. 3–4, pp. 409–
419, 2004.
[52] O. A. Williams and R. B. Jackman, “High growth rate MWPECVD of single crystal
diamond,” Diam. Relat. Mater., vol. 13, no. 4–8, pp. 557–560, 2004.
[53] O. A. Williams and R. B. Jackman, “Homoepitaxial growth for surface conductive
device applications,” Diam. Relat. Mater., vol. 13, no. 2, pp. 325–328, 2004.
[54] T. Teraji, “Chemical vapor deposition of homoepitaxial diamond films,” Phys. Status
Solidi, vol. 203, no. 13, pp. 3324–3357, 2006.
[55] C. Wang, M. Irie, and T. Ito, “Growth and characterization of hillock-free high quality
266

homoepitaxial diamond films,” Diam. Relat. Mater., vol. 9, no. 9, pp. 1650–1654,
2000.
[56] A. Tallaire, J. Achard, F. Silva, O. Brinza, and A. Gicquel, “Growth of large size
diamond single crystals by plasma assisted chemical vapour deposition: Recent
achievements and remaining challenges,” Comptes Rendus Physique, vol. 14, no.
2–3. pp. 169–184, 2013.
[57] T. Bauer, M. Schreck, H. Sternschulte, and B. Stritzker, “High growth rate
homoepitaxial diamond deposition on off-axis substrates,” Diam. Relat. Mater., vol.
14, no. 3–7, pp. 266–271, 2005.
[58] J. Achard, A. Tallaire, R. Sussmann, F. Silva, and A. Gicquel, “The control of growth
parameters in the synthesis of high-quality single crystalline diamond by CVD,” J.
Cryst. Growth, vol. 284, no. 3–4, pp. 396–405, 2005.
[59] N. Lee and A. Badzian, “A study on surface morphologies of (001) homoepitaxial
diamond films,” Diam. Relat. Mater., vol. 6, no. 1, pp. 130–145, 1997.
[60] N. Lee and A. Badzian, “Effect of misorientation angles on the surface
morphologies of (001) homoepitaxial diamond thin films,” Appl. Phys. Lett., vol. 67,
no. 1, p. 2011, 1995.
[61] N. Lee and A. Badzian, “Effect of methane concentrations on surface morphologies
and surface structures of (001) homoepitaxial diamond thin films,” Appl. Phys. Lett.,
vol. 67, no. 1, p. 2011, 1995.
[62] K. Hayashi, S. Yamanaka, H. Watanabe, T. Sekiguchi, H. Okushi, and K. Kajimura,
“Diamond films epitaxially grown by step-flow mode,” J. Cryst. Growth, vol. 183, pp.
338–346, 1998.
[63] A. Tallaire et al., “Characterisation of high-quality thick single-crystal diamond
grown by CVD with a low nitrogen addition,” Diam. Relat. Mater., vol. 15, no. 10,
pp. 1700–1707, 2006.
[64] P. Martineau, M. Gaukroger, R. Khan, and D. Evans, “Effect of steps on dislocations
in CVD diamond grown on {001} substrates,” Phys. Status Solidi Curr. Top. Solid
State Phys., vol. 6, no. 8, pp. 1953–1957, 2009.
[65] A. Tallaire et al., “Dislocations and impurities introduced from etch-pits at the
epitaxial growth resumption of diamond,” Diam. Relat. Mater., vol. 20, no. 7, pp.
875–881, 2011.
[66] F. K. De Theije, J. J. Schermer, and W. J. P. Van Enckevort, “Effects of nitrogen
impurities on the CVD growth of diamond: Step bunching in theory and experiment,”
Diam. Relat. Mater., vol. 9, no. 8, pp. 1439–1449, 2000.
267

[67] A. Gicquel, E. Anger, M. F. Ravet, D. Fabre, G. Scatena, and Z. Z. Wang, “Diamond
deposition in a bell-jar reactor: influence of the plasma and substrate parameters
on the microstructure and growth rate,” Diam. Relat. Mater., vol. 2, no. 2–4, pp.
417–424, 1993.
[68] J. Asmussen and D. K. Reinhard, Diamond Films Handbook. Marcel Dekker, 2002.
[69] C. J. Chu, R. H. Hauge, J. L. Margrave, and M. P. D’Evelyn, “Growth kinetics of
(100), (110), and (111) homoepitaxial diamond films,” Appl. Phys. Lett., vol. 61, no.
12, pp. 1393–1395, 1992.
[70] G. Bogdan, M. Nesládek, J. D’Haen, K. Haenen, and M. D’Olieslaeger,
“Freestanding (100) homoepitaxial CVD diamond,” Diam. Relat. Mater., vol. 15, no.
4–8, pp. 508–512, 2006.
[71] D. Takeuchi, H. Watanabe, S. Yamanaka, H. Okushi, and K. Kajimura, “Defects in
device grade homoepitaxial diamond thin films grown with ultra-low CH4/H-2
conditions by microwave-plasma chemical vapor deposition,” Phys. Status Solidi aApplications Mater. Sci., vol. 174, pp. 101–115, 1999.
[72] T. Teraji, S. Mitani, and T. Ito, “High rate growth and luminescence properties of
high-quality homoepitaxial diamond (100) films,” Phys. Status Solidi Appl. Res., vol.
198, no. 2, pp. 395–406, 2003.
[73] K. W. Hemawan, T. A. Grotjohn, D. K. Reinhard, and J. Asmussen, “Improved
microwave plasma cavity reactor for diamond synthesis at high-pressure and high
power density,” Diam. Relat. Mater., vol. 19, no. 12, pp. 1446–1452, 2010.
[74] X. Li, J. Perkins, R. Collazo, R. J. Nemanich, and Z. Sitar, “Investigation of the effect
of the total pressure and methane concentration on the growth rate and quality of
diamond thin films grown by MPCVD,” Diam. Relat. Mater., vol. 15, no. 11–12
SPEC. ISS., pp. 1784–1788, 2006.
[75] A. Tallaire, J. Achard, F. Silva, and A. Gicquel, “Effect of increasing the microwave
density in both continuous and pulsed wave mode on the growth of monocrystalline
diamond films,” Phys. Status Solidi Appl. Mater. Sci., vol. 202, no. 11, pp. 2059–
2065, 2005.
[76] L. S. Pan, Diamond: Electronic Properties and Applications. KLUWER ACADEMIC
PUBLISHERS, 1995.
[77] R. Locher, C. Wild, N. Herres, D. Behr, and P. Koidl, “Nitrogen stabilized ???100???
texture in chemical vapor deposited diamond films,” Appl. Phys. Lett., vol. 65, no.
1, pp. 34–36, 1994.
[78] W. Müller-Sebert, E. Wörner, F. Fuchs, C. Wild, and P. Koidl, “Nitrogen induced
268

increase of growth rate in chemical vapor deposition of diamond,” Appl. Phys. Lett.,
vol. 759, p. 759, 1995.
[79] Q. Liang et al., “Enhanced growth of high quality single crystal diamond by
microwave plasma assisted chemical vapor deposition at high gas pressures,”
Appl. Phys. Lett., vol. 94, no. 2, p. 24103, 2009.
[80] S. A. Catledge and Y. K. Vohra, “Effect of nitrogen addition on the microstructure
and mechanical properties of diamond films grown using high-methane
concentrations,” J. Appl. Phys., vol. 86, no. 1, pp. 698–700, 1999.
[81] I. Sunagawa, “Growth and Morphology of Crystals,” Forma, vol. 14, no. 358, pp.
147–166, 1999.
[82] Y. Bar-Yam and T. D. Moustakas, “Defect-induced stabilization of diamond films,”
Nature, vol. 342, pp. 786–787, 1989.
[83] G. Z. Cao, J. J. Schermer, W. J. P. van Enckevort, W. A. L. M. Elst, and L. J. Giling,
“Growth of {100} textured diamond films by the addition of nitrogen,” J. Appl. Phys.,
vol. 79, no. 3, pp. 1357–1364, 1996.
[84] D. Raabe and T. Liu, “Influence of nitrogen doping on growth rate and texture
evolution of chemical vapor deposition diamond films,” Appl. Phys. Lett., vol. 94,
no. 2, pp. 2–5, 2009.
[85] J. Lu, “SINGLE CRYSTAL MICROWAVE PLASMA ASSISTED CHEMICAL
VAPOR DIAMOND SYNTHESIS AT HIGH PRESSURES AND HIGH POWER
DENSITIES,” 2013.
[86] S. Nad, “GROWTH AND CHARACTERIZATION OF LARGE , HIGH QUALITY
SINGLE CRYSTAL DIAMOND SUBSTRATES VIA MICROWAVE PLASMA
ASSISTED CHEMICAL VAPOR DEPOSITION By Shreya Nad A DISSERTATION
Submitted to Michigan State University in partial fulfillment of the requirements fo,”
2016.
[87] M. A. Elliott, P. W. May, J. Petherbridge, S. M. Leeds, M. N. R. Ashfold, and W. N.
Wang, “Optical emission spectroscopic studies of microwave enhanced diamond
CVD using CH 4 /CO 2 plasmas,” Diam. Relat. Mater., vol. 9, pp. 311–316, 2000.
[88] Y. Meng et al., “High optical quality multicarat single crystal diamond produced by
chemical vapor deposition,” Phys. Status Solidi a-Applications Mater. Sci., vol. 209,
no. 1, pp. 101–104, 2012.
[89] Q. Liang et al., “Enhancing the mechanical properties of single-crystal CVD
diamond,” J. Phys. Condens. Matter, vol. 21, no. 36, p. 364215, 2009.

269

[90] Y. Meng et al., “Enhanced optical properties of chemical vapor deposited single
crystal diamond by low-pressure/high-temperature annealing.,” Proc. Natl. Acad.
Sci. U. S. A., vol. 105, no. 46, pp. 17620–17625, 2008.
[91] A. Chayahara et al., “The effect of nitrogen addition during high-rate homoepitaxial
growth of diamond by microwave plasma CVD,” Diam. Relat. Mater., vol. 13, no.
11–12, pp. 1954–1958, 2004.
[92] Z. B. Feng, A. Chayahara, Y. Mokuno, H. Yamada, and S. Shikata, “Raman spectra
of a cross section of a large single crystal diamond synthesized by using microwave
plasma CVD,” Diam. Relat. Mater., vol. 19, no. 2–3, pp. 171–173, 2010.
[93] H. Yamada, A. Chayahara, Y. Mokuno, and S. ichi Shikata, “Numerical and
experimental studies of high growth-rate over area with 1-inch in diameter under
moderate input-power by using MWPCVD,” Diam. Relat. Mater., vol. 17, no. 7–10,
pp. 1062–1066, 2008.
[94] Y. Mokuno, a. Chayahara, H. Yamada, and N. Tsubouchi, “Improving purity and
size of single-crystal diamond plates produced by high-rate CVD growth and lift-off
process using ion implantation,” Diam. Relat. Mater., vol. 18, no. 10, pp. 1258–
1261, 2009.
[95] H. Yamada, A. Chayahara, and Y. Mokuno, “Short-pulse excitation of microwave
plasma for efficient diamond growth,” Appl. Phys. Lett., vol. 109, no. 9, p. 92102,
2016.
[96] H. Yamada, A. Chayahara, Y. Mokuno, Y. Soda, Y. Horino, and N. Fujimori,
“Modeling and numerical analyses of microwave plasmas for optimizations of a
reactor design and its operating conditions,” Diam. Relat. Mater., vol. 14, no. 11–
12, pp. 1776–1779, 2005.
[97] H. Yamada, A. Chayahara, Y. Mokuno, Y. Horino, and S. Shikata, “Numerical
analyses of a microwave plasma chemical vapor deposition reactor for thick
diamond syntheses,” Diam. Relat. Mater., vol. 15, no. 9, pp. 1389–1394, 2006.
[98] H. Yamada, A. Chayahara, and Y. Mokuno, “Simplified description of microwave
plasma discharge for chemical vapor deposition of diamond,” J. Appl. Phys., vol.
101, no. 6, 2007.
[99] H. Yamada, A. Chayahara, Y. Mokuno, and S. ichi Shikata, “Simulation with an
improved plasma model utilized to design a new structure of microwave plasma
discharge for chemical vapor deposition of diamond crystals,” Diam. Relat. Mater.,
vol. 17, no. 4–5, pp. 494–497, 2008.
[100] H. Yamada, A. Chayahara, Y. Mokuno, and S.-I. Shikata, “Microwave plasma
generated in a narrow gap to achieve high power efficiency during diamond
270

growth,” Diam. Relat. Mater., vol. 18, no. 2–3, pp. 117–120, 2009.
[101] H. Yamada, “Numerical Simulations to Study Growth of Single-Crystal Diamond by
Using Microwave Plasma Chemical Vapor Deposition with Reactive ( H , C , N )
Species,” vol. 51, no. 100, pp. 1–7, 2012.
[102] Y. Horino, A. Chayahara, and Y. Mokuno, “High-Rate Growth of Large Diamonds
by Microwave Plasma Chemical Vapor Deposition with Newly Designed Substrate
Holders,” New Diam. Front. Carbon Technol., vol. 16, pp. 63–69, 2006.
[103] H. Yamada, A. Chayahara, Y. Mokuno, Y. Horino, and S. Shikata, “Simulation of
temperature and gas flow distributions in region close to a diamond substrate with
finite thickness,” Diam. Relat. Mater., vol. 15, no. 10, pp. 1738–1742, 2006.
[104] H. Yamada, A. Chayahara, Y. Mokuno, Y. Horino, and S. Shikata, “Simulation of
microwave plasmas concentrated on the top surface of a diamond substrate with
finite thickness,” Diam. Relat. Mater., vol. 15, no. 9, pp. 1383–1388, 2006.
[105] Y. Mokuno, A. Chayahara, and H. Yamada, “Synthesis of large single crystal
diamond plates by high rate homoepitaxial growth using microwave plasma CVD
and lift-off process,” Diam. Relat. Mater., vol. 17, no. 4–5, pp. 415–418, 2008.
[106] N. Tsubouchi, Y. Mokuno, H. Yamaguchi, N. Tatsumi, A. Chayahara, and S.
Shikata, “Characterization of crystallinity of a large self-standing homoepitaxial
diamond film,” Diam. Relat. Mater., vol. 18, no. 2–3, pp. 216–219, 2009.
[107] N. Tsubouchi, Y. Mokuno, A. Chayahara, and S. Shikata, “Crystallinity of
freestanding large undoped single crystal diamond plates produced using pre-ionimplanted substrates and lift-off processes,” Diam. Relat. Mater., vol. 19, no. 10,
pp. 1259–1262, 2010.
[108] H. Yamada, A. Chayahara, Y. Mokuno, N. Tsubouchi, S. I. Shikata, and N. Fujimori,
“Developments of elemental technologies to produce inch-size single-crystal
diamond wafers,” Diam. Relat. Mater., vol. 20, no. 4, pp. 616–619, 2011.
[109] H. Yamada, A. Chayahara, H. Umezawa, N. Tsubouchi, Y. Mokuno, and S. Shikata,
“Fabrication and fundamental characterizations of tiled clones of single-crystal
diamond with 1-inch size,” Diam. Relat. Mater., vol. 24, pp. 29–33, 2012.
[110] H. Yamada, A. Chayahara, Y. Mokuno, H. Umezawa, S. I. Shikata, and N. Fujimori,
“Fabrication of 1 inch mosaic crystal diamond wafers,” Appl. Phys. Express, vol. 3,
no. 5, pp. 1–4, 2010.
[111] M. Marchywka, P. E. Pehrsson, D. J. Vestyck, and D. Moses, “Low energy ion
implantation and electrochemical separation of diamond films,” Appl. Phys. Lett.,
vol. 63, no. 25, pp. 3521–3523, 1993.
271

[112] F. Silva et al., “Single crystal CVD diamond growth strategy by the use of a 3D
geometrical model: Growth on (113) oriented substrates,” Diam. Relat. Mater., vol.
17, no. 7–10, pp. 1067–1075, 2008.
[113] F. Silva et al., “High quality, large surface area, homoepitaxial MPACVD diamond
growth,” Diam. Relat. Mater., vol. 18, no. 5–8, pp. 683–697, 2009.
[114] A. Tallaire et al., “Growth strategy for controlling dislocation densities and crystal
morphologies of single crystal diamond by using pyramidal-shape substrates,”
Diam. Relat. Mater., vol. 33, pp. 71–77, 2013.
[115] J. Achard, F. Silva, O. Brinza, A. Tallaire, and A. Gicquel, “Coupled effect of
nitrogen addition and surface temperature on the morphology and the kinetics of
thick CVD diamond single crystals,” Diam. Relat. Mater., vol. 16, no. 4–7 SPEC.
ISS., pp. 685–689, 2007.
[116] A. Secroun, A. Tallaire, J. Achard, G. Civrac, H. Schneider, and A. Gicquel,
“Photoconductive properties of lightly N-doped single crystal CVD diamond films,”
Diam. Relat. Mater., vol. 16, no. 4–7 SPEC. ISS., pp. 953–957, 2007.
[117] P. Djemia, A. Tallaire, J. Achard, F. Silva, and A. Gicquel, “Elastic properties of
single crystal diamond made by CVD,” Diam. Relat. Mater., vol. 16, no. 4–7 SPEC.
ISS., pp. 962–965, 2007.
[118] F. Silva, X. Bonnin, J. Scharpf, and A. Pasquarelli, “Microwave analysis of PACVD
diamond deposition reactor based on electromagnetic modelling,” Diam. Relat.
Mater., vol. 19, no. 5–6, pp. 397–403, 2010.
[119] A. Gicquel et al., “Quantitative analysis of diamond deposition reactor efficiency,”
Chem. Phys., vol. 398, no. 1, pp. 239–247, 2012.
[120] F. Silva, X. Bonnin, J. Achard, O. Brinza, A. Michau, and A. Gicquel, “Geometric
modeling of homoepitaxial CVD diamond growth: I. The {100}{111}{110}{113}
system,” J. Cryst. Growth, vol. 310, pp. 187–203, 2008.
[121] M. Naamoun et al., “Reduction of dislocation densities in single crystal CVD
diamond by using self-assembled metallic masks,” Diam. Relat. Mater., vol. 58, pp.
62–68, 2015.
[122] A. Tallaire, O. Brinza, V. Mille, L. William, and J. Achard, “Reduction of Dislocations
in Single Crystal Diamond by Lateral Growth over a Macroscopic Hole,” Adv.
Mater., vol. 29, no. 16, pp. 1–5, 2017.
[123] G. Wu, M.-H. Chen, and J. Liao, “The influence of recess depth and crystallographic
orientation of seed sides on homoepitaxial growth of CVD single crystal diamonds,”
Diam. Relat. Mater., vol. 65, pp. 144–151, 2016.
272

[124] S. Nad and J. Asmussen, “Analyses of single crystal diamond substrates grown in
a pocket substrate holder via MPACVD,” Diam. Relat. Mater., vol. 66, pp. 36–46,
2016.
[125] S. Nad, A. Charris, and J. Asmussen, “MPACVD growth of single crystalline
diamond substrates with PCD rimless and expanding surfaces,” Appl. Phys. Lett.,
vol. 109, no. 162103, pp. 0–4, 2016.
[126] K. Kuo and J. Asmussen, “An experimental study of high pressure synthesis of
diamond films using a microwave cavity plasma reactor,” Diam. Relat. Mater., vol.
6, pp. 1097–1105, 1997.
[127] S. S. Zuo, M. K. Yaran, T. A. Grotjohn, D. K. Reinhard, and J. Asmussen,
“Investigation of diamond deposition uniformity and quality for freestanding film and
substrate applications,” Diam. Relat. Mater., vol. 17, no. 3, pp. 300–305, 2008.
[128] J. Zhang, “An investigation of electromagnetic field patterns during microwave
plasma diamond thin film deposition,” J. Vac. Sci. Technol. A Vacuum, Surfaces,
Film., vol. 8, no. 1990, p. 2124, 1990.
[129] J. Zhang, “Experimental development of microwave cavity plasma reactors for large
area and high rate diamond film deposition,” Michigan State University, 1993.
[130] D. Zhang and T. Cui, “Department of Mechanical Engineering , University of
Minnesota , Minneapolis , Minnesota , USA,” vol. 1996, no. l, pp. 362–365, 2011.
[131] K. Kuo, “Microwave assisted plasma CVD of diamond films using thermal-like
plasma discharge,” Michigan State University, 1997.
[132] Y. Gu, “THE NEW GENERATION MICROWAVE PLASMA ASSISTED CVD
REACTOR FOR DIAMOND SYNTHESIS,” 2011.
[133] J. Asmussen et al., “Process and Apparatus for diamond Synthesis,” US 9487858
B2, 2009.
[134] Canon, “CANON EOS 20D Digital Manual,” 2014. [Online].
http://gdlp01.c-wss.com/gds/9/0900000259/01/EOS20DIM-EN.pdf.

Available:

[135] R. Diaz and A. Charris, “DS4 Box: Complement for DS4 (Diamond System 4).
Michigan
State
University,”
2017.
[Online].
Available:
https://github.com/RDiazPR/DS4.
[136] Http://www.bris.ac.uk/Depts/Chemistry/MOTM/diamond/diamond.htm, “University
of Bristol, School of Chemistry.” .
[137] J. Asmussen, “Microwave Plasma Source Technologies: A Fifty Year Evolution
273

from Unwanted Discharges to Free Radical Sources, to Low Pressure and
Temperature Plasma Processing, to Gem Quality Diamond Synthesis,” in AVS
62nd International Symposium, 2015.
[138] J. Asmussen, T. a. Grotjohn, and T. Schuelke, Advances in Plasma Synthesis of
UNCD Films, Second Edi. Elsevier Inc., 2012.
[139] W.-S. Huang, “Microwave Plasm a A ssisted Chemical Vapor Deposition of U ltrananociystalline Diamond Films VOLUME I,” Michigan State University, 2004.
[140] “HPHT
Sumitomo.”
[Online].
http://www.sumicarbide.com/materialsgroup/crystalu.htm.

Available:

[141] S. Nad, Y. Gu, and J. Asmussen, “Determining the microwave coupling and
operational efficiencies of a microwave plasma assisted chemical vapor deposition
reactor under high pressure diamond synthesis operating conditions,” Rev. Sci.
Instrum., vol. 86, no. 7, 2015.
[142] J. Lu, Y. Gu, and J. Asmussen, “A Description of The Experimental Microwave
Discharge Behavior Versus Pressure , Power and Reactor Geometry for MPACVD
Diamond Synthesis Reactors,” vol. 43, no. 15, p. 153001, 2012.
[143] L. EAG, “SECONDARY ION MASS SPECTROMETRY (SIMS).” [Online]. Available:
http://www.eaglabs.com/mc/secondary-ion-mass-spectrometry.html.
[144] R.
C.
Microscopy,
“Optical
Birefringence.”
[Online].
http://www.olympusmicro.com/primer/lightandcolor/birefringence.html.

Available:

[145] O. A. Ivanov, A. B. Muchnikov, V. V Chernov, S. A. Bogdanov, A. L. Vikharev, and
J. E. Butler, “Experimental study of hydrogen plasma etching of (1 0 0) single crystal
diamond in a MPACVD reactor,” Mater. Lett., vol. 151, pp. 115–118, 2015.
[146] N. Tsubouchi, Y. Mokuno, and S. Shikata, “Characterizations of etch pits formed
on single crystal diamond surface using oxygen/hydrogen plasma surface
treatment,” Diam. Relat. Mater., vol. 63, pp. 43–46, 2016.
[147] Y. Mokuno, A. Chayahara, H. Yamada, and N. Tsubouchi, “Large Single Crystal
Diamond Plates Produced by Microwave Plasma CVD,” Mater. Sci. Forum, vol.
615, no. 617, pp. 991–994, 2009.
[148] H. Yamada, A. Chayahara, Y. Mokuno, Y. Kato, and S. Shikata, “A 2-in. mosaic
wafer made of a single-crystal diamond,” Appl. Phys. Lett., vol. 104, no. 10, pp. 10–
14, 2014.
[149] O. Brinza et al., “Dependence of CVD diamond growth rate on substrate orientation
as a function of process parameters in the high microwave power density regime,”
274

Phys. Status Solidi Appl. Mater. Sci., vol. 205, no. 9, pp. 2114–2120, 2008.
[150] C. J. Widmann, W. Müller-Sebert, N. Lang, and C. E. Nebel, “Homoepitaxial Growth
of Single Crystalline CVD-diamond,” Diam. Relat. Mater., vol. 64, pp. 1–7, 2016.
[151] A. B. Muchnikov, A. L. Vikharev, A. M. Gorbachev, D. B. Radishev, V. D. Blank,
and S. A. Terentiev, “Homoepitaxial single crystal diamond growth at different gas
pressures and MPACVD reactor configurations,” Diam. Relat. Mater., vol. 19, no.
5–6, pp. 432–436, 2010.
[152] J. Lu, “Single Crystal Microwave Plasma Assisted Chemical Vapor Diamond
Synthesis at High Pressures and High Power Densities,” 2013.
[153] A. B. Muchnikov, A. L. Vikharev, A. M. Gorbachev, and D. B. Radishev,
“Comparative study of homoepitaxial single crystal diamond growth at continuous
and pulsed mode of MPACVD reactor operation,” Diam. Relat. Mater., vol. 20, no.
8, pp. 1225–1228, 2011.
[154] A. L. Vikharev et al., “Comparison of pulsed and CW regimes of MPACVD reactor
operation,” Diam. Relat. Mater., vol. 12, no. 3–7, pp. 272–276, 2003.
[155] A. Tallaire et al., “Multiple growth and characterization of thick diamond single
crystals using chemical vapour deposition working in pulsed mode,” J. Cryst.
Growth, vol. 291, no. 2, pp. 533–539, 2006.
[156] Y. Kato, H. Umezawa, H. Yamaguchi, and S. Shikata, “X-ray topography used to
observe dislocations in epitaxially grown diamond film,” Jpn. J. Appl. Phys., vol. 51,
no. 9, 2012.
[157] H. Pinto and R. Jones, “Theory of the birefringence due to dislocations in single
crystal CVD diamond,” J. Phys. Condens. Matter, vol. 21, p. 364220, 2009.
[158] T. Teraji, M. Hamada, H. Wada, M. Yamamoto, and T. Ito, “High rate growth and
electrical/optical properties of high-quality homoepitaxial diamond (100) films,”
Diam. Relat. Mater., vol. 14, no. 11–12, pp. 255–260, 2005.
[159] A. Charris, S. Nad, and J. Asmussen, “Exploring constant substrate temperature
and constant high pressure SCD growth using variable pocket holder depths,”
Diam. Relat. Mater., vol. 76, no. December 2016, pp. 58–67, 2017.
[160] T. Grotjohn, A. Hardy, M. Becker, A. Charris, J. Asmussen, and T. Schuelke,
“Diamond Based Power Electronics: Material and Devices - MRS Europe,” 2016.

275