-
U

.1.

..
1. h.
. :uJaw a
t 0.. I
.
. L ..
0u.u.

. .c
an . . . .
”.viummw.

nu,a. .
INN;

h

. ..H....bdrww.4 . .

L ,‘nflyv. .

- . as.

,
..t\ anM I
wwwmmwmmmwl uggnwuu

IV an n v
... “Wmmwfi a? o

5|
‘

.. Wm
. n . f-AAVWVMIV
.Eywm u.

.7 lawman”.
o Ju.fvfln.:v0w
0R ura‘udlww 3H.“

. .v ... 03.4.7.3.
g. WW}. mfg... , .

1w“ 1.

o
9.
Mr vb
.

s r da1.|‘.00

. nu-
$v0‘.k.o.
‘0 bl.
.

’3‘;

.L'Ii

‘1- .%.«-L, . I
unnlnvdunw:”HWkn|H
3.0%.. -

 

xx! 11 .‘v
1 . 3..
1.5%”...

r!

I“

s u I.“ III. Vb
n yu.anMuv. “Wm
. P...

‘Efiiizin .

Quit. u
‘ y ‘5. .

 

 

 

. 4
VA‘
ha .v
‘ l
.701.
I.f.
-’~

.
, .sIthalonrnv-
. b a A l

 

”I a
Hw. . .
. lanai}-
. ‘. WJ(MMm-Juihunnrn

E

vv.n

nouummmMflMrhAduv - .

.LM oli‘luu.v.£uhnliv
.

 

THESIS

                                                             

           

WI IIIIIIIIIIIIIIIIIIIII

31293 01716 3548

This is to certify that the

dissertation entitled

FABRICATION AND CHARACTERIZATION OF
DIAMOND FIELD EHITI'ERS FOR
FIELD EHISSION DISPLAYS

presented by

Dongsung Bong

has been accepted towards fulfillment
of the requirements for

PhD degree in Electrical Engineering

 

 

fizz/«7%»

 

Major professor

Date €Z/7/97

MSU is an Affirmative Action/Equal Opportunity Institution 0- 12771

 

 

LIBRARY

Michigan State
University

 

 

 

PLACE IN RETURN BOX
to remove this checkout from your record.
TO AVOID FINE-S return on or before date due.

I MTE DUE DATE DUE DATE DUE

m.
I’DECIO i “2001

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1M mu

FABRICATION AND CHARACTERIZATION OF
DIAMOND FIELD EMITTERS FOR
FIELD EMISSION DISPLAYS
By

Dongsung Hong

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Electrical Engineering

1 997

ABSTRACT

FABRICATION AND CHARACTERIZATION OF
DIAMOND FIELD EMITTERS FOR
FIELD EMISSION DISPLAYS

By

Dongsung Hong

New materials, with low work function and high chemical immunity, are needed to
produce low-cost and stable field emission display (FED). The unique intrinsic properties
of diamond make it an excellent material for field emitter.

In the present work, using an IC-compatible fabrication process, a field emission
testchip was designed, fabricated, and tested. The chip contains a number of test devices
including a 1x4 pixel triode FED which was demonstrated for the first time. The testing of
the chip revealed that it is important to enhance field emission current density for further
development of FED.

Using different film growth conditions and post-deposition carbon implantation, the

effect of defect on the field emission current density was systematically studied. The CH4/
H2 ratio, grain size, resistivity, and implantation dose in the ranges of O.5-2%, 0.3-1 .5 pm,
1.7-189 (2cm, and 5x105-5x106 cm'z, respectively, were used to vary the defect density in

the film.

Based on the field emission data collected from a variety of samples, it was found that

emission from diamond is enhanced when (i) sp3/sp2 is low, (ii) peak at 1332 cm" is
wider, (iii) grain size and the roughness of film are small, (iv) film is highly doped, and (v)

ion itnplantation dose is high while energy is low. The results seem to suggest that field

emission from polycrystalline diamond is affected by (i) defects and (ii) field

enhancement at the grain tips. As all samples were treated in hydrogen at 900 °C, the

electron affinity may be same for all samples.

I II'ouIt‘
iii. IIS gu
WI Ike
and D; KI
In: of III:

Man 1
mi j» f}.
3.331111 I a
gimme :
:‘ifiilfil In

I331] II
III Mam
IIIIIEAI !

ImfliI}

235336. u

ACKNOWLEDGEMENTS

I would like to express my sincere gratitude to my advisor, Dr. Dean M. Aslam. With-
out his guidance and financial support, this work wouldn’t have been possible. I also
would like to thank my committee members, Dr. Reinhard, Dr. Grotjohn, Dr. Grummon,
and Dr. Klomparens. Their valuable comments on my research resulted in great enhance-
ment of the quality of this work.

Many thanks go to Dr. O’Kelly, Dr. F remon, and friends at the Graduate School. Two
and half years of experiences at the Graduate School enriched my personal and social
growth. I also would like to thank all my, labmates. Camaraderie between us helped us to
overcome some difficulties I encountered throughout this work and to teach how to work
together in harmony.

I can not forget all the support from Korean EE students and their family when I need
help. Many friends whom I have known for years and whom I became to know here at
Michigan State helped me in many occasions.

Finally but not least, I would like to thank my parents and family in Korea. Their

patience, unconditional support, and love gave me the strength to finish this work.

TABLE OF CONTENTS

LIST OF TABLES viii
LIST OF FIGURES ix
LIST OF ABBREVIATIONS xiv
1 RESEARCH MOTIVATION AND GOALS ................ 1
1.1 Introduction .............................................. 1
1.2 Objective of This Work ..................................... 2
1.3 Dissertation Organization .................................... 3
2 BACKGROUND ...................................... 6
2.1 Introduction .............................................. 6
2.2 Derivation of Fowler-Nordheim Equation ....................... 6
2.3 Fowler-Nordheim Plot ...................................... 21
2.4 Diamond Field Emitters ..................................... 23
2.4.1 Models ............................................ 30
2.5 CVD Diamond Infrastructure ................................ 32
2.5.1 Nucleation ......................................... 32
2.5.2 Patterning .......................................... 35
2.5.2.1 Photographic Methods .......................... 38
2.5.2.2 Direct-write Methods ........................... 38
2.5.3 ' Doping ............................................ 38
2.5.4 Metallization ....................................... 40
3 DEPOSITION AND CHARACTERIZATION SYSTEM ..... 42
3.1 Introduction .............................................. 42
3.2 Deposition Systems ........................................ 42
3.2.1 HFCVD ........................................... 42

3.2.2 RFCVD ........................................... 50

3.3 Film Characterization ....................................... 50

3.4 Field Emission Characterization System ........................ 53

3.4.1 Anodes Designs ..................................... 58

3.4.1.1 W Probe Anode ................................ 59

3.4.1.2 Brass Column Anode ........................... 59

3.4.1.3 Lateral Anode ................................. 59

3.4.1.4 Glass Anode .................................. 60

3.4.1.5 Phosphor Anode ............................... 60

3.4.1.6 Built-In Anode ................................ 60

3.5 Summary ................................................ 64

4 IC COMPATIBLE FIELD EMISSION TECHNOLOGY ...... 65

4. 1 Introduction .............................................. 65

4.2 New and Enhanced Patterning Technique ....................... 65

4.3 Preliminary Samples ........................................ 67

4.4 Discrete Samples with Built-in Anode .......................... 70

4.4.1 Vertical Structure .................................... 70

4.4.2 Lateral Device ....................................... 73

4.5 Testchip I ................................................ 76

4.6 Testchip II ................................................ 78

4.7 Testchip III ............................................... 80

4.8 Summary ................................................. 86

5 FIELD EMISSION RESULTS ................................ 89

5. 1 Introduction .............................................. 89

5.2 Preliminary Samples ........................................ 89

5.2.1 Setup .............................................. 89

5.2.2 Results ............................. i ................ 91

5.3 Discrete Samples with Built-in Anode .......................... 97
5.4 Testchip I ................................................ 103

vi

55 Te
56 Te
5.7 Di
5.8 SK

6 APPLIt

6.2 E

6.3 l

6.4

7 COM
7.2

«Wm
M’PENI

BIBLIOI

5.5 Testchip II ................................................ 103

5.6 Testchip III ............................................... 107

5.7 Discussions ............................................... 118

5.7.1 Effect of Diamond Deposition Condition .................. 119

5.7.2 Effect of Implantation ................................ 129

5.8 Summary ................................................. 141

6 APPLICATIONS ............................................. 143
6.1 Introduction .............................................. 143

6.2 Diamond Field Emitter Display ............................... 143

6.2.1 Diamond Field Emitter Display with Bridge Shaped Grid ..... 144

6.2.1.1 Fabrication ................................... 144

6.2.1.2 Testing ...................................... 147

6.2.2 Diamond Field Emitter Display with Self-Aligned Grid ...... 150

6.2.2.1 Fabrication ................................... 154

6.2.2.2 Testing ...................................... 157

6.3 Field Emitter Pressure Sensor ................................. 165

6.3.1 Simulation .......................................... 166

6.3.2 Fabrication ......................................... 166

6.3.3 Testing ............................................ 172

6.4 Summary ................................................. 175

7 CONCLUSIONS AND FUTURE RESEARCH ................ 176
7.1 Summary of Conclusions .................................... 176

7.2 Future Research ........................................... 177
APPENDIX. A 178
APPENDIX. B 185
BIBLIOGRAPHY 189

vii

llSl OF

3.2 Com
3.1 list
3.3 list
3.3 List
4.1 Suit
3.1 Sun

Imp

LIST OF TABLES

2. 1 Summary of various diamond field emitters ......................... 26
2. 2 Comparison of two types of diamond powder loaded fluids ............. 36
3. 1 List of pressure readouts. ....................................... I 56
3. 2 List of power supplies. ......................................... 56
3. 3 List of measurement instruments .................................. 57
4. 1 Summary of mask steps for each structure in test chip 111 ............... 82
5. 1 Summary of sample preparation and characterization .................. 120
5. 2 Implantation energy and dose for four samples ...................... 136

viii

LIST OF FIGURES

l.

2.

1

1

An overview of the development of diamond field emitter .............. 4
Surface potential barrier of a metal-vacuum system ................... 7
t(y) and v(y) as a function of y ................................... 22

Band diagram including the hypothetical bands in the energy gap [53]. . . . 31

Emission models for a Si/diamond system: tunneling of electrons through the
bent band gap and escaping into a vacuum from the surface under NEA
conditions [3 7] ................................................ 31

Simplified model of electron emission through dielectric layer, where
electrons tunnel through metal/dielectric interface [40] ................ 33

For substitutional nitrogen. Donors are ~l .7eV below Be. In equilibrium a
depletion region forms at the metal-diamond interface. The circles represents

the energy position of the donors [42] .............................. 33
Energy levels for a metal-diamond interface with the metal biased to -6V

with respect to the diamond [42] .................................. 34
Schematic diagram of a diamond cathode in a diode structure [42] ....... 34
Schematic of HFCVD constructed. ............................... 44

(a) AN SYS simulation of temperature profile due to sagging filament. (b)
Temperture of the substrate as a function of the distance from the center. . 47

Typical morphology by SEM (a) and Raman spectrum (b) of the film grown

by HFCVD ................................................... 51
Schematic of automated characterization system built. ................ 54
Six different anodes used in this study. ............................ 61

Process sequence for the patterning of (a) DPLFl and (b) DPLF2 using
photographic method ........................................... 66

Patterning results of (a) DPLFl and (b) DPLF 2 samples by photolithography
........................................................... 68

Comparison of (a) conventional patterning methode and (b) double layer

method ...................................................... 69

4.4 Cm
dot
4.5 (a)
0ft
1,6 Dis
4.? (at
4.8 Te:
pat
0V
4.9 le
sat
co:
4.10 V;
lll T(
4.12 It
4'13 S]
5.1 T!
52 1.
a
ll
5.3 I-
it
5.4 I
l
l
5.5 ,

4.4
4.5

4.6
4.7

4.8
4.9
4.10
4.11

4.12

4.13

5.1

5.2

5.3
5.4
5.5

5.6

Cross sectional views of preliminary samples with continuous film and (b)
dot patterned film. ............................................ 71

(a) Top view, (b) cross-section view, and (c) magnified view of section A-A’

of vertical type discrete sample with built-in anode. .................. 72
Discrete samples with built-in anode grown on quartz substrate. ........ 74
(a) Cross sectional view and (b) top view of lateral device .............. 75

Test chip I (a) Diamond pattern (b) Sacrificial photoresist pattern (c) Metal
pattern I ((1) Metal pattern II which covers most of diamond with metal (e)
Overlapped view (t) completed wafer .............................. 77

Test chip II (a) Diamond pattern (b) wafer picture corresponding to (a), (c)
sacrificial phtoresist pattern ((1) Metal pattern (e) Overlapped view (f)
completed wafer ............................................... 79

Variety of structures on test chip 111. .............................. 81

Top and cross sectional views of type A device in different process steps

........................................................... 84
Top and cross sectional views of type B device in different process steps

........................................................... 85
SEM and cross-sectional view of lateral process in each process step ..... 87

Test configuration of the samples; continuous film (a) and patterned film (b)
........................................................... 9O

I-V curves of diamond field emitter for blunt and sharp anodes; emitter to

anode separation is 50 um and the measurement pressure is 10'7 Torr. The
inset shows the corresponding F-N plot. ........................... 92

I-V curves for different anode to cathode distances for the sharp anode at 10'7
torr . ....................................................... 93

I-V curves for two different pressures; the emitter to anode separation is 200

um and the sharp anode is used. ................................. 95
The field emission pattern measured (a) for continuous film in position 2 of

Figure 5.1(a) and (b) for patterned film ............................. 96
Measurement setup for (a) vertical device and (b) lateral device. ........ 98

5.17

3211

3.11

)1:
mi

.'. . (at

8C

5.7

5.8

5.9

5.10

5.11

5.12

5.13

5.14

5.15

5.16

5.17

5.18

5.19

5.20

5.21

5.22

5.23

I-V and F-N data for four different vertical devices. .................. 99
I-V and F-N data for two different lateral devices ..................... 100

I-V and F-N data for two different vertical devices seed with low and high
nucleation density. ............................................ 101

I-V and F -N data for two different lateral devices seed with low and high
nucleation density. ............................................ 102

Measurement setup for cell 224 of tetschip 1. Eight marked cells were
measured. ................................................... 104

(a) I-V and F -N data of cells marked circle in Figure 5.11. (b) I-V and F-N of
a cell from square marked and one cell from circle marked in Figure 5.11

........................................................... 105
I-V and F-N data for samples with metal pattern I and metal pattern II in test
chip I. ...................................................... 106
Measurement setup for test chip II (a) vertical device (b) lateral device

........................................................... 108
l-V and F -N data for vertical devices with different emitter area ......... 109

I-V and F -N data for lateral devices with different emitter to anode distance

........................................................... 110
I-V and F-N data for lateral devices with different emitter shape with samle
distance. .................................................... 111
Measurement setup for (a) type A and (b) type B device. .............. 113

I-V curve and F-N plot of diamond field emitter for type A and B device
........................................................... 114

I-V curves of type A (inset) and type B devices for continuous and array type
emitters with the same total area .................................. 115

Measurement setup for lateral device from Al on thop of diamond film (a)
and Al on Si backside. ......................................... 116

I-V and F-N data for lateral device with different emitter contact. ....... 1 l7

I-V curves, Raman spectra, and SEM pictures of samples for samples with
different CH4 concentrations ..................................... 121

xi

6.1

5.4

F0

5.24

5.25

5.26

5.27

5.28

5.29

5.30

5.31

5.32

5.33

5.34

6.1

6.2

6.3

6.4

6.5

6.6

I-V curves, Raman spectra, and SEM pictures of samples for samples with

doped and undoped film. ....................................... 123
I-V curves, Raman spectra, and SEM pictures of samples for samples doped
with two different doping methods. ............................... 124
I-V curves, Raman spectra, and SEM pictures of samples for samples with
doped only layer and doped layer over undoped layer. ................ 125
I-V curves, Raman spectra, and SEM pictures of samples for samples with
different grain size ............................................. 127
Four different structures for ANSYS simulation: (a) Type I, (b) Type II, (c)
Type III, and ((1) Type IV. ...................................... 130
Simulated electric fields at the tip for different types for anode to emitter
distances of l and 10 um ........................................ 134
Measured emission current as a function of field for anode to emitter

distances of 1.6 and 50 um. ..................................... 135
SEM image and Raman spectra of implanted and unimplanted area. ..... 137
I-V curves for four samples measured at Michigan State. .............. 138
F-N curves for samples (a) HI-S and HI-9 measured at Varian ........... 138

Residual gases for HI-2 after 350 0C 16 hr bake out: two graphs are used for
clarity ....................................................... 140

FED cell fabrication process; cross sectional views and SEM micrographs.
........................................................... 145

A magnified view of a display cell showing vacuum gap between gate and
emitter. ..................................................... 148

Experimental setup for I-V measurement and (b) I-V curve and F-N plot of a
display cell when tested in a diode configuration. .................... 149

Setup for emission image from a triode display cell and its corresponding
image. ...................................................... 151

Setup for measuring emission image from four triode display cell array and
its corresponding image ........................................ 152

A failed display cell due to a cluster of diamond particles. ............. 153

xii

6.1!}

Ell

6.13

.1. l

. Des

dep

gal:
Sat
Me

gm

Elli
‘-

6.7

6.8

6.9

6.10

6.11
6.12

6.13

6.14

6.15

6.16

6.17

A.1

Design view, cross sectional view, and wafer photography (a) after diamond
deposition, (b) afier emitter contact etch followed by Cr evaporation, (c) after
gate definition. ((1) Design view and finished wafer. Two different gate

structure are also shown ......................................... 155
Samples with (a) cracks and (b) no crack. .......................... 161
Measurement setup for self-aligned diamond FED in triode mode. Emitter is
grounded at the pad shown in Figure 6. 7(b) ......................... 162
Anode current Ia and gate current Ig as a function of anode voltage Va and
gate voltage Vg ................................................ 163
Emission pattern of the FED chip in triode configuration ............... 164
AN SYS simulation of diaphragm. ................................ 167
Displacement value of diaphragm along A-A’ of Figure 6.12. Unit is in um
........................................................... 168
Approximation of pressure sensor to calculate total current ............. 169
Cross-sectional views of a pressure sensor at each process step and the
corresponding pictures of top view. ............................... 170

(a) Measurement setup and (b) I-V and F-N plot of a diamond pressure sensor
when measured at a pressure of 10'6 Torr and 10'3 Torr ................ 173

Measurement setup for a pressure sensor in atmosphere ................ 174

Schematic of RFCVD system. ................................... 184

xiii

lIST 0F 1

i-N: Fowler-
ll: integrate.
lie: \loltbd.
FED: letd Et
l'CD; liquid
Sit Silicon

CVD: Chem:
ill: Negati
RGA: Residt
3061313111 C
HFCVD; M}
hiCl'D; H6
Pitt]; R3
Station
W Atom
SEM; Scam
3Pll: Dian

I‘m. Catho

LIST OF ABBREVIATIONS

F-N: Fowler-Nordheim

IC: Integrated Circuit

Mo: Molybdenum

FED: Feid Emission Display

LCD: Liquid Crystal Display

Si: Silicon

CVD: Chemical Vapor Deposition

NEA: Negative Electron Affinity

RGA: Residual Gas Analyzer

SOG: Spin On Glass

MPCVD: Microwave Plasma Chemical Vapor Deposition
HF CVD: Hot Filament Chemical Vapor Deposition
RFCVD: Radio Frequency Chemical Vapor Deposition
SiC: Silicon Carbide

AF M: Atomic Force Microscopy

SEM: Scanning Electron Microscopy

DPLF: Diamond Powder Loaded Fluid

CRT: Cathode Ray Tube

SiOz: Silicon Dioxide

Cr: Chrome

ITO: Indium Tin Oxide

Zn: Zinc

xiv

130'. Z11
I-l’: Cur

ll: lung

ZnO: Zinc Oxide
I-V: Current versus Voltage

W: Tungsten

XV

1.1

. I
' 'Ytr‘g
‘il 4 .

- .'. .
ui't‘9g-

““ L. [Jr

CHAPTERl

FmfiEARCHhMTHWATHHJAND
GOALS

1.1 Introduction

The theory for field emission was first proposed by Fowler and Nordheim (F—N) in
1928 [1], but its realization in the field of vacuum microelectronics could be possible
through the development of modern integrated circuit (IC) technologies. Despite its
advantage of vacuum transport of electrons over electron transport in solid, solid state
devices have substituted major part of vacuum devices since technological development in
early fifties mainly due to compactness. Ironically vacuum microelectronics area is
revitalized thanks to the technological development of 1C fabrication that helps make it
possible to produce small vacuum devices.

Fabrication of microtip made of molybdenum (Mo) by Spindt et a1. [2] in the mid
seventies opened a new era for field emission devices. A variety of materials and structures
were attempted by a number of research groups thereafter to improve the performance of
devices with respect to their application needs [3, 4, 5]. Application possibilities of field
emitter devices range from ultra fast switches, microwave amplifiers and generators, flat
panel display devices, intense electron/ion sources, multiple electron sources, new

electron beam lithography tools to miniature electron excitation devices [3, 4]. Mainly due

been:

t I
Dist-

;:l

-._,t_

R.

r‘

to its colossal market prospect, applying field emission devices to low cost and high
performance flat panel display of modern information society has been major thrust of

field emission research.

Low power consumption; wide viewing angle; and bright, fast, and high contrast
image qualities are the advantages of field emission displays (FED) over the conventional
liquid crystal displays (LCD). The first prototype FED, based on metal microtip emitters
first fabricated by Spindt et al., was reported in 1991 [6]. Si based prototypes have also
been reported. However, in a silicon (Si) or metal FED, high work function (which means
high supply voltage) and phosphor contamination of the tip are some of the problems.
High work function (typically 4 eV) makes the field enhancement necessary which is
achieved by using microtip emitters. The microtip fabrication increases the processing
cost. New materials, with low work function and high chemical immunity, are needed to
produce low-cost and stable FED.

Recent advances in diamond film technology have led to inexpensive polycrystalline
chemical-vapor deposited (CVD) diamond films on non-diamond substrates [7,8]. Due to
its negative electron affinity (NBA) [9], immunity to chemical attack, hardness and very
high thermal conductivity (highest at 300 K), diamond is an excellent material for field
emitters, especially for FED. Field emission was demonstrated from diamond or diamond-
like materials [10, 11, 12, 13, 14, 15]. Although a diode-based diamond-like FED has been
demonstrated [12], many issues, related mainly to the diamond quality and process tech-

nology, must be resolved before any reliable diamond FEDS can be realized.

1.2 Objective of This Work

Early works on diamond field emitters have successfully demonstrated the excellent

emissmr
:zissittr

irzsstttr

retort:
4'.

mill-101'.
{meat

its i1

emission properties of diamond field emitters. Being in an early stage of diamond field
emission research, these studies are mainly focused on initial measurement of field
emission from various diamonds itself leaving some unanswered questions mainly
associated with the mechanism of field emission from diamond. To be able to answer

these questions, it is necessary to study correlation between emission characteristic and

diamond properties such as sp3/sp2 ratio, doping concentration and grain size. To apply
diamond’s unique properties to real world application, it is important to develop
fabrication technology for diamond field emitters, compatible to standard IC process.

Thus, in order to achieve these two main objectives. one has to address following issues.

(1) The development of a fabrication technique for diamond field emitter compatible

with standard IC fabrication processes using testchip approach.

(2) Design and systematic execution of a series of experiments to collect emission
data from a number of different film growing conditions and post growth

treatment S .

(3) Demonstration of application of the developed fabrication technology into

diamond FED in triode mode.

Figure 1.1 illustrates a number of essential tasks to be done in order to realize
diamond FED in triode configuration. Work covered in this research is limited to those

tasks with solid frame.

1.3 Dissertation Organization

Research motivation and goals are defined in Chapter 1. In Chapter 2, a complete
derivation of the F-N theory of field emission, diamond properties, the current

technologies of diamond field emitters, and suggested models are described. Brief review

«55:.» Eva 1:255. no «58:205.. 2: me 33:25 =< A.“ 953%

 

 

=o_§:m_mE_
.3 omens—o

 

 

 

 

EoanBSn

 

cosmozamao.

 

 

 

 

 

=oU Om.»

 

 

 

 

 

 

 

 

 

ofim 5me

28:53:
:os_momov-.mom

 

 

 

 

 

 

 

 

 

 

526:8

 

 

cosmutoaofimcu

 

 

 

emu 25qu

 

 

 

 

 

wcfioc

 

 

538w Lo 69cm

 

 

 

 

 

 

2:2 mmm\mmm

 

 

 

 

connotes.“ Ego and.

 

 

:ocaotnfl x38

 

 

Ea amuse E238...

 

 

0.25023

 

EoEmEoBQ
$22508.

 

 

 

 

 

 

2698 5.5.5

 

 

 

 

:f CVD .
.‘CftSTJCll
faint-.160
etglm Lt.
{ii 11mm

treasure s

of CVD diamond infrastructure is given as well. Chapter 3 deals with the design and
construction of two diamond CVD systems and characterization system including anode
design. Chapter 4 focuses on the development of IC compatible diamond field emitter
fabrication technology. Characterization of various samples and qualitative effort to
explain the measurement results are discussed in Chapter 5. In Chapter 6, two applications
of diamond field emitter are demonstrated, namely, diamond FEDS in triode mode and

pressure sensor. Summary and conclusions are given in Chapter 7.

CHAPTER 2

BACKGROUND

2.1 Introduction

In this chapter, derivation of F-N equation will be presented in detail. Later, diamond
field emitters reported so far by others and some explanations of emission process from
diamond will be summarized. Nucleation, patterning, doping, and metallization

technologies are reviewed briefly.

2.2 Derivation of Fowler-Nordheim Equation

Let’s consider the surface potential energy barrier shown in Figure 2.1 denoted by the

solid curve. This curve includes the image potential component e2/4z and applied field
effects component er with respect to how they affect a step barrier when combined. The

potential energy barrier can be approximated by

2
V(z) = EF+¢—:—Z—er (2.1)

= 0 for z>0

for z<0 g
The peak of the barrier described by Equation 2.1 occurs at zmax, where 2max is defined

by

let

Vacuum
level

 

 

Metal Vacuum

Image potential effect only

Applied field effect only

 

Figure 2.1. Surface potential barrier of a metal-vacuum system.

41/- 3
dz 4z 2

 

which gives

- 8 1/2
zmax "' 1?:

At z=zmax, Equation 2.1 becomes:

e2 6 1/2
Vmax = V(zmax) = EF+¢——__—172_8F(4F)

8
43;.)

3/2 1/2 3/2 1/2
v -E +4) 8 e F
m..- F “—2—“?—

3 1/2
vmax: EF+¢-(e F)

(2.2)

(2.3)

(2.4)

(2.5)

(2.6)

(2.7)

For a theory of electron emission from metal surfaces, it is necessary to consider (i) the

number of electrons with energy between W and W + dW normal to the surface

impinging on the surface barrier, N(W), and (ii) the tunneling probability of electron

through the barrier, D(W). Then P(W)=N(W)D(W) will be the number of electrons per

area per second per energy range dW, that penetrate the barrier. Hence, the emitted

current density, i.e. the number of electrons emitted per unit area per unit time, in the

energy range W and W+dW is given by:

J = eN(W)D(W)dW

(2.8)

li the

she: ‘

I " V
Lilli l
H

l: -

“it?! 516;,

If the electron energy range extends from —oo to co , Equation 2.8 becomes:

J = e r N(W)D(W)dW (2.9)

Let’s start to evaluate N(W) first. Assuming the electron follows Fermi-Dirac statistics,

N(W) is given by [16]
N(W) = 4—Z3t—nfif(E)dE (2.10)

where W is the energy of tunneling electron and f(E) is Fermi-Dirac distribution function.

Using the Fermi-Dirac distribution function

 

41tm l
N(W) — 7]; (E—EF) dE (2.11)
1 + exp[———]
kBT
Let
H = (212)
kBT '
then
dE
__ = d .
kBT x (213)
dE = 1(3de (2.14)

Limits of integration change according to Equation 2.12:

 

oo——)oo (2.15)

W W—EF
—) 2.16
kBT ( )

Then Equation 2.11 becomes

which. '

ECOEC

10

_ 41tm kl?T
N(W) — fifi'EFITe—mdx (2.17)
kBT

which, using standard form from integration table [17]

 

 

 

 

 

 

 

 

 

 

 

I dx x = ln 1 _ (2.18)
1+ 8 1+ e x
becomes
N(W) — 3 n _x (2-19)
h 1+e W-E,
kBT
41kaBT 1
N(W) = 3 n( _°°)- Il'l fl—W—E— (2.20)
h l+e - k TF
1+e B
w—E,r fl
4nkaT “1;?“
N(W) = 3 ln(l)—ln(1)+ln 1+e J (2.21)
h
41tkaT W—EF
N(W): 3 ln[l + exp(— )] (2.22)
h kBT

Each of these electrons has a probability D(W) to be transmitted through the surface
potential barrier. Transmission coefficient for the potential barrier shown in Figure 2.1 is

given by [16],

 

32an
h2

 

D( W) = exp[—J3

Z1

[V(z) - Wldz] (2.23)

11

when W < Vmax , z, and 22 are two real roots of V(z)-W=0 as shown in Figure 2.1. Using

Equation 2.1 which expresses the potential energy in terms of the applied field and image

force effect, the above equation can be written as:

 

 

h2
Z' 1 R4
1 l

:2 32an 2
D(W) -_- exp[—J [EF+¢—W—:—Z—er]dz]

Let’s find two real roots of the V(z)-W=0

2
V(z)—W=EF+¢—w—:—Z—e1~"z = 0

7

h

(EF+(l)—W)z—eZ—er2 = 0

 

2 (EF+¢—W) e
Z eF z 4F

 

 

 

(Ewe—W)+ (EF+¢—W)2_ie_
eF _ eze 4F

 

2

 

 

 

 

(EF+¢—W)+J(EF+¢-W)2_(EF+¢-W)2 e317

 

 

 

 

 

2
, _. eF e2 F2 e2 F2 (EF+¢‘W)
’ 2
(EF+¢—W) (EF+¢-W>J e3F
i F 1— 2
7_ 8F e (EF+¢—W)
“ ’ 2

 

 

 

(EF+¢‘W) e3F
z = 2eF 1i 1— 2
(Ep+¢-W)

(2.24)

(2.25)

(2.26)

(2.27)

(2.28)

(2.29)

(2.30)

(2.31)

std the

ill]

[Sing E.

warm

12

E —W 3
22 = ( F+¢ ) 1i 1_ e F (2.32)
21 28F (EF+¢—W)2

The above defines the limits of integration for the transmission coefficient expression.

 

 

 

These limits correspond to the distance the electron must travel in tunneling through the

potential barrier. The shorter the distance, the higher the probability of tunneling.

Let

ale-”F

y = ¢—————+ EF— W (2.33)
and the integration variable:
2eF
: ml (2.34)
Then
t6 + E F — W
= —_—2eF u (2.35)
+ E — W
dz = q-J—zfi—du (2.36)

Using Equation 2.32 and 2.33

 

E W

32: (F1: )11 +~/1-y 2] (2.37)
E W

Z1= (F-Hb )[1-YIJ1-21 (2.38)

2eF

 

Substituting Equation 2.37 and 2.38 into Equation 2.34

u2 = (1+Jl—y2) (2.39)

lhfithiE

'hnlibeco

hngmesut

31121131111 1161

13

u1 = (l-A/l—yz) (2.40)

Then R1 in Equation 2.24 becomes:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2 2
321nm e ¢+EF—W
R = E —W— — —— .4
‘ h2 F”) 4(pug—W 1 2eF “1 ‘2 1’
ll 1 2eF “1
2
32am 83F ¢+EF‘W
R = E —W— — .4
‘ [1:2 if”) 2(¢+EF—W)u 2 “i ‘2 2’
+E —We3F —
12(¢+EF-W)u—(¢ F ) 2 —(¢+EF-W)u2
R _ 16an (¢+EF—W)

1 _ 2 (2.43)

t " “

2 2
R _ 16an 2(¢+EF—W)u-(¢+EF—W)y -(¢+EF-W)u

1- 2 (2.44)

h u

2 2
4 2 — —
R] = 7% m(¢+EF—W)[i—i—u] (2.45)
Then 1] becomes:
2 (¢+EF-W)

II - [ZR] 2eF du (2.46)
_ 2“«/’"(¢+EF_W)3U+ 1_)'2)~/—u2+2u—y2d 247
I _ heF I u u (' )

(l-Jl—yz)

L l

 

I2
Using the substitution n2 = u in Equation 2.47 brings the integral 12 to standard form

which can be evaluated by an elliptical integral:

l1 ..
1‘1.

in

J(l +x/l —)'2)

12:

12:

14

2min = du

u2—)l+«/l-y2
n2—)JI+Jl—y2

n]——> l--:(/l—y2

 

 

_ 4 2 2_ 2
l J“*? 1mm
,———— Tl
l—x/l-y2

J(l +J1-y3)

2 j «[414 + 2112 - .vzdn
l I

Jl—fi—yz R2

 

Add and substract x/l — yz, n l — yz, and 1 to R2 to get:

R2 = —ll4+112+112+«/1-.V2-«/1-y2+ll«/1- 2—n,/1—yz-l+ l—y2

which after factorization becomes:

R2 = (1+~I1- 2-Tl‘°')(112-(1-'J1-y2))

Equation 2.54 now becomes:

a: 41+ l—-y2

R2 = (aZ-n2)(n2—b2)

(2.48)

(2.49)

(2.50)

(2.51)

(2.52)

(2.53)

(2.54)

(2.55)

(2.56)

(2.57)

(2.58)

15

 

I2 = 2J:«/(02-112)(112-b2)0"1 (2.59)

Equation 2.47 now becomes:

 

 

4 EF —W
II = “Jm(¢h:p )3 }J(a2— n2)(n2—b2)dn (2.60)

 

Using elliptic integral table [17]

 

1t/2

 

 

 

 

4n./m(¢+Ep-— W) 2a (212.1132)“
11 = heF 3l:——— J1 x/l -k7-sin20d9— b2 {11225111291
I = SKA/m(¢+EF—W)3a[(a2+b2)E(k)—b2K(k)] (261)
1 3heF 2 '

l I
V1

where ‘K(k)’ and ‘E(k)’ are elliptic integrals of the first and second kind respectively:

 

 

n/Z d9
K(k) = (2.62)
‘(E 4/1 —kzsin20
n/2
E(k) = J1 l—kzsin20d0 (2.63)
0
And,
2_ 2
k2 = a 2b (2.64)
a

 

Vl = ./1+./1-y2[l+“/; 1’2:llfi’y2(E(/c)—(1—./1—y2)K(k))] (2.65)

Vl = J1+,/1—y2[E(k)—(l—,/1—y2)K(k)] (2.66)

16

where:

v(y) = 2~1/2,/1+ ./1 —y2[E(k) — (1 — ./1—y2)K(k)] (2.67)

Then 11 becomes:

 

8nJ2m(¢ + E,r — W)3
l _ 3heF

 

v(y) (2.68)

The transmission coefficient can then be expressed as:

 

 

(2.69)

3heF

8 2 E _w3
D(W)=exp[ n m(¢+ F ) 0)]

Equation 2.9 can be written as

 

 

 

oo W—E
4nmk3r — F) 81: 2m(¢+E -W)3 ‘
J = eJTln[l +e kBT jexp[ f 3heFF v(y) dW (2.70)

I 1 G(W) I r
1

N (W) D(W)

 

 

 

 

Let’s evaluate the above expression at (or near) absolute zero temperature.

First consider the exponent, G(W) in the transmission coefficient portion of Equation 2.70:

 

8nj2m(¢ + EF — W)3 .

G(W) = - 3h”. V(.)) (2.71)

 

with

x/e3F

= m (2.72)
F _

y

Since electrons have energies in the neighborhood of the Fermi level Ep, using Taylor’s

expansion theorem:

l7

G(W) = G(EF)+G'(EF)(W-EF)+ ”(21"F__)'(W EF)2+

Only first two terms will be considered here.

G(W) = G(EF) + G'(EF)(W—EF)

Using Equation 2.7], G(W) at W=EF is given by

0(5)) = ————8“Wv(@)

3heF 6

G’(W) in the second term is given by:

 

 

, _ d [ 8nj2m(¢+EF—W)3 W
G (W) ‘ 21w] 3heF V(¢+EF-

Using

(uv)' = u’v + uv'

u’v is given by

—-81'l:
3heF

 

and uv’ is given by

 

 

3h 4F[./2m(¢+EF- W)3— —v(y)]

Equation 2.76 becomes:

w)]

Eamon + EF — W)3)-1/22m3(¢ + EF — W)2v(y)]

 

 

 

._ —3 E -W2m
G’(W) = 8K[ (¢+ F )

3"” J2m(¢ + EF — W)3

Using

v(y)+JZm(¢+EF— W) 3—v(y):|

(2.73)

(2.74)

(2.75)

(2.76)

(2.77)

(2.78)

(2.79)

(2.80)

0] ll

lift-'1)

~16 1’13

18

dv _ dvdy

 

 

 

 

 

271i - m ‘38"
LL = dv x/e3F (2 s?)
“W d-V(¢+EF— W)2
_ —3 +5 —W 2m /3 ,
G'(W)= 811: |: ((1) F ) v(y)+,\/2im((l)+EF—W)3 e F 2d) (2.87))
3heF fim(¢+EF_ W)3 (¢+EF_ W) (1).

 

 

 

 

 

 

 

_ _——8nJ2m(¢+EF W)_ _ 3v.v( )+ {—6316 dv (224)
3heF (0+EF—W)d_v '
4nj2m(¢+EF—W) 2de
._ heF [I (y) —— 3.213;] (2.85)
Let t(y) = v(y)—Zy-g—VU) then
3 dy
41c /2m(¢+E -_W)
G’(W) = heF F" t(y) (2.86)
G'(EF)= Wfly) (2.87)
Equation 2.74 becomes
(2.88)

G(W) = -8“‘2m¢3"( 3F F1 4“ ——2mem¢t(y ><W— EF)

3heF (1)

As the temperature approaches absolute zero,

l9

 

_w E,r
kBTln[l+e KT J = O whenW>EF
(2.89)
=EF—W whenW <EF
Thus, from Equation 2.70, it follows that:
J = 0 when W > EF
e41tm E; (2.90)
J =

 

h [(EF— W)exp[G(W)]dW whenW<EF

—oo

As at T=O K the highest energy is EF, the limit of integral in Equaion 2.90 changed to —oo

 

 

 

 

t0 EF.
Using Equation 2.74:
Er
J- — etc—:11! exp[G(EF)+ G’ (EF)(W— EF)](EF— —W)dW (2.91)
e41tm
J = ——exp[G(EF)—G ’(EF)EF]JLI: exp[G’ (EF)W](EF— W)dW (2.92)
L I13“ I
Cl 13
I3 = fr EFexp[G'(EF)W]dW-J-EF Wexp[G’(EF)W]dW (2.93)
l l l I
I4 15
First integral 14 becomes
EF 5”
I- . 4
4— G —(——,EF)exp[G’ (EF)W] (29)

 

For second integral 15, let’s use integration by part

20

 

[u'v = uv—Iuv' (2.95)
CXPIG’(EF)W]
Let u' = ex [G'(E )W] , = , ,v=W,v’=l
" F G (12,.)

Then second integral becomes

exp[G’(EF)W]W EF

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

F l
= , +J.E , ex G’ E W dW (2.96)
exp[G’(E )W]W EF 1 EF
- - , F + 2 exp[G'(Ep)W1 (2.97)
G (5;) _ G’ (EF)
Let’s combine the first integral and second integral
E; , E E
EF ' exPlG (Ep)WlW F 1 F
I = , ex [6’ E )W) — , ——ex [6’ E W] (2.98)
E E‘ 2,.
= G, g exp[G’(EF)W] — 2 exp[G’(EF)W][WG’(EF)— 1] (2.99)
( r) -0. G’ (EF)
EF I I
———’exp[G (EF)EF] - exp[G (EF)EF][EFG (EF) — 1] (2.100)
=0 (Ep) 0’2 F
1 , , ,
= ,2 exp[G (EfiEM (EH74 (EF)+1] (2.101)
G (EF)
ex G’ E E
I3 = pl 2( F) F] (2.102)
0' (EF)
Then I becomes:
G’(EF)EF ,
J = e41tme (eG(EF)—G (EHEF) (2.103)

 

h3 G’2(EF)

21

e41tm

= 3 ’2 exp[G(EF)] (2.104)
h G

(5;)

Substituting G(EF) from Equation 2.75 and G’(EF) from Equation 2.87

 

3 2 ('— 3/2
= _eL_ex [ 8“ 2"“1’ v(y)] (2.105)
81th¢t2(y) 3heF
where
2 d
t(y) = V(y)—§y-d—Vv'(y) (2.106)
and

v(y) = 2-“241 +4/1—y2[E(k)—(1—~/1—y2)K(k)] (2.107)

Substituting all numerical parameter of
m=9.11x10'31 kg

k3: 1.38x10'23 J/K

h=6.63x10'34 Js

e=1.6x19"9c

‘6 2 3/2
J = 1.54X10 F exp[—6.83X107q_)ir_v(y)] (2.108)

WU)

 

2.3 Fowler-Nordheim plot

t2(y) vs. y and v(y) vs. y curves are shown in Figure 2.2. At (or very near) absolute

zero temperature, y value is close to 0 since electron enegry is close to Ep. In this case,

we can approximte t2(y)=v(y)=l for ease of discussion. Let d = the emitter-to-gate

1.4

1.2

0.8

0.6

0.4

0.2

22

 

l

 

I

T

 

t(y)

V(y)

 

0.2

Figure 2.2. t(y) and v(y) as a fuction of y.

 

EL. 5

.t I
L“;

0?

N011

1
Cite:

‘&

with.

“4x. a;
‘-

23

distance such that F=V/d. Now let A: the emitting area in cmz. These two assignments
allow us to put the F-N equation in Equation 2.113 in terms of the physically tractable

quantities of current, I and voltage V. 80

2 3/2

 

 

 

 

J = 1.54 X 10-6V—2exp[-6.83 x 10m d] (2.109)
V
(pd
01'
3/2
L2 = 1.54 x10_6—A—2-exp[—6.83 x1074) ‘1] (2.110)
V (M V
Now take the log of both sides
3/2
logiz- = 1og(1.54x 10’6A§)—6.83x 1071’ dloge (2.111)
V (pd V
or
3/2
iogi2 = log(1.54 x10—6—A-3)— 2.97 x 1074’ d (“‘2’
V (M V

_ Clearly, a plot of the Iogi2 vswl/ will have straight line. This plot is called F-N plot and
V

often used to confirm field emission.

2.4 Diamond Field Emitters

The chemical, thermal, and electrical properties of the emitter material play an
important role in the design of on-chip vacuum tubes. To achieve high current density at
low applied voltages, two approaches are feasible. One approach is to sharpen the

emitters to enhance the electric field, resulting in a greater field effect. Another approach

C0,":

lq~

11.11.?
{:3 [hi

3171110

MM“
. 1'
1'

24

is to use emitter materials with low work function. Thus less energy is required to emit
electrons into vacuum. In other words, the emitters (if semiconductors are used) should
have a low (or even negative) electron affinity so that lower electric fields are required
for emission.

Low electron affinity is just one of many desired emitter properties. In addition to a
low electron affinity, the ideal emitter would have a high carrier concentration and high
mobility, and thus a high electrical conductivity (from O=nqu). It would also have a high
electric field breakdown strength so as to withstand the locally high fields. The thermal
conductivity should be high so the heat generated by the high current densities could be
quickly dissipated. Finally, the emitter should be chemically inert since impurities in the
emission surface can increase the material work function. Early metal emitters suffered
from impurity contamination [18] and structural fragility, although their high electrical

conductivity made them promising emitter candidates. Si microtips have been shown to

withstand high current densities (J > 1000 A/cm'z), but work function raising impurities
continue to be a problem, along with current instability and the need for an ultra high
vacuum environment [19]. Diamond, however, has the possibility of overcoming many
of these obstacles and greatly improving device performance, particularly in harsh

environments.

Diamond’s high electric field breakdown coefficient of 107 ch'l is 30 times that of
Si and 1.5 times that of GaAs [20], meaning locally high electric fields could be

produced and the device would continue to function properly. The electron saturation

velocity in diamond is nearly twice that of silicon at 2.7)(107 cms'l[20], which would

facilitate high speed device applications. Diamond is nearly impervious to chemical

25

attack and therefore work function raising contamination encountered in metals and Si
would be of a lesser concern. The resistance of diamond to chemical attack means that
the device may be operated in a lower vacuum environment, thus relaxing fabrication
demands. The thermal conductivity of diamond is unmatched at 20W/cm-K (for type 11b)
[19] and is nearly 5 times that of copper, so it should be able to quickly dissipate the
power induced by the high current densities. In the (l l 1) direction, the electron affinity
of diamond is very small and perhaps even negative [9], so emission could be realized at
lower applied fields. For type IIb diamond the electrical resistivity is on the order of 10-
1000 ohm-cm [21], so conductivity of type IIb diamond is resonably high.
I Recently, the research activity in field emission from diamond has seen an
exponential growth [10, ll, 12, 13, 14, 15] which is due mainly to its NBA [9] and
immunity to chemical attack. Low anode voltages, simplified fabrication process, less
stringent vacuum requirements, and high stability are some of the consequences of
diamond’s unique properties. The simplified fabrication process is expected to result in
lower cost.

The field emission has been demonstrated from a diamond junction device [10],
homoepitaxial diamond [11], amorphous diamond [12], polycrystalline diamond [13,15]
and diamond coated Si tip [14]. The emission was observed for electric fields in the

range of 0.03 - 0.5 MV/cm which is two orders of magnitude lower than that for typical

Si or metal emitters. Current densities of up to 10 Acm'2 have been reported. Prototype
FED based on diamond and carbon have been reported [12, 22] as well. Carbon based

field emitters reported are summarized in Table 2.1.

.01....1-OCIII-I Iv-ICF- I‘ll-IllIl-u-I IIIIVI.III.I .III

Quill“! IIIIIII aAI

26

 

988::

 

 

an 95mm 28.21 328811

:2 99.3 >on .. .1 <1 m: 1c a 838 0.5

an gums. a1\>o~ ,. m <1 E .. 1 SE 0.5 818 z

.85 a:

as 05.1: a 338 2255

NEUZanmnseH

E1S§usem 08.: a.11 as 1:

as 96% >81n§s> a <8 8 use; a 838 25.55

a:

:3 96 a 338 2835

«Emma 8a Esau? SE 5530 £52m

$3 2858 12211 51S 5.2m <1 8.1 25:5 BREE

a

Rs gum: E1S-um <1$u1 8 :21 .5 SE So

.31 gums >879 >8mns> <1wns1 1: a 8180 3a

concmo

:2 > omm _ .(> <1 cm 1221 8%.: 250
00:95 0 O OE an: acne m 20$ .5 H hwmmcov 0.5 OE m Co a

a. a n 5 .H. O > Rug—Boa 3:991 8250 10 H 32:6 H H ”M E. H 2

 

33:50 23a 9855:. 33...; .3 FEES—6 flu ~33.

27

 

a:
25:86 825855

 

 

51 96mm E13 2 (m <2 S .. H
RE 28.558505 En\>_zonm_ Em 25:85
:E 59E: E\>2oo_um <1 02 (H 5E 25:85 5>U
SE 5>U 02 288 25:85
32 5>UnE 5E 25:86 5>U
a:
:5 55.1: <1 :8 L a 888 2285
5:
K2 5R5: <3 25 G 288 25:85
25552 a: G 388
$2 88 222180 5825 >oom _ z> <3; 5580 50:55:31
:2 5512 51288 15252". SE 8886 55
$2 5280 5.82
25558 EOE
5528.: 5 988
an 96mm > So u> 1521 8w": 0883 2285

0258 o o 2: :2 .53 m 20: 5 p. .3285 2 2: m :o 8

a. m o 5 .H. 5 > 82855 25%? 5250 52.12.50 5 a Q 8. a 2

 

815.80 .3 as“...

28

 

a:

 

 

 

83 mm 5 558250580
5:55 25:

RS 525 :28 25520
SE 252:8

51 > o? z > NEQZEE: . 888 82812
a .5 mEE

SE Q>0n=2 > com: > <1 mm A 25:85 3:553
Hm : Q>0m2 E\>2omvm <1 m T. H 5E 25:85 Q>0
a:

$2 Q>0m2 _m 5588 25:85
552%

E1 555 <1 33 a: H a 838 2835
a:

GE D>0mm E 588 25:85
558585 :55 52

$3 682552 58556.“— >om_ 2> <1 2:: _ 28 _m .5 mEE 0:N
25

-55 5580 52:

:1; > oooz> 25:4 8:555 SE .55.
cop—255% 55:55 525qu m 25: 5 _. $7.5: 255.5» a. 82582

> 825:5 252?. 5250 5 m 5250

 

82.5.18 .3 2.3

> -uq-:flihc‘ .il‘m-fl‘ <

nyd-u-u-unhfisU .-.N Qua-3.?

‘ 1..-!“ 1

29

 

 

 

5:
$2 >x mHu> <5 :5 t H 5:5 52:80 852
:2 «quoum <1 n.2,: a 8 05

8:25 o 55 :2 :5: m 22H 5 H 13:55: 2: o m :5 8

h— m 55 .H. on > 8555: :5:an71 5250 5 H 5250 : P: d 8. . H2

 

255250 .H.N 53:9

2.4.1

30

2.4.1 Models

Presently, the physics of electron emission from diamond is not well understood. First
theoritical study by Huang et al. [53] found that emission from surface states is shown to
be capable of producing the current density with field magnitudes comparable to
experiments. The question of how these electrons are transported to the surface states
needs to be answered. One possible mechanism they propose is that the elecron transport
can take place through the defect states.

If defect concentration is significant, the electron states in these defects could form a
band or bands. In their model, two subbands in the intrinsic band gap are postulated,
which may be generated, for example, by defects or impurities. Band structure they
postulated is shown in Figure 2.3. It is suggested that the defect bands can transport
electrons to the unoccupied surface band located 1 eV below the conduction band, which,
if occupied under applied field, can emit electrons to vacuum.

Givargizov et al. [37] proposed a model for electron emission from Si tip coated with
thin diamond. They assumed that diamond is perfect crystal with low doping
concentration. The proposed emission process is illustrated in Figure 2.4 for a Si/diamond

interface, assuming that properties of a heavily doped Si/diamond heterojection are similar

to metal/diamond Schottky junction. At n<lO15 cm'3, the bend bending effect at the Si/
diamond interface is small and may be neglected. Electrons tunnel through the bent band
gap of the diamond, and drift to the surface and escape for vacuum. NEA of diamond
surface was assumed in a model.

Choi et al. [40] proposed a model similar to Givargizov’s model for electron emission
from diamond coated Mo emitter. They consider an undoped or slightly doped diamond
layer on M0. The band bending effect at the metal/diamond interface is so small that it can
be neglected. The energy barrier height at the metal/diamond (intrinsic) interface is very

nearly Eg/Z. They assume that diamond has a NEA surface, as a result, the emission would

31

 

 

 

 

 

Figure 2.3. Band diagram including the hypothetical bands in the

energy gap [53].

 

e——w
”$1

 

Dianond

 

Uacuun

Figure 2.4. Emission models for a Si/diamond system: tunneling of
electrons through the bent band gap and escaping into a
vacuum from the surface under NEA conditions [37].

5:;er t
conducti
sacuum

Gei~
.5rc5 I
Mon.
cEtctror
daring
itch 1
{ms :
h: s_\'
comp;

ms! 1

If

it

32

depend only on tunneling of electrons through the Mo/diamond Schottky barrier into the
conduction band of diamond. The resultant energy band diagram for metal/diamond!
vacuum is shown in Figure 2.5.

Geis et al. [42] and Lerner et al. [54] explained the emission performance of their
device by the stable NEA of diamond, which allows for injection of electrons from
diamond into vacuum. They argue that emitter operation is limited by the injection of
electrons into the diamond at the back metal-diamond interface, which depends upon the
doping of the diamond and the roughness of that interface. They consider the case in
which the diamond is doped with an electron donor. Substitutional nitrogen, for example,
forms a deep trap 1.7 eV below the conduction band. Figure 2.6. shows the energy levels
for synthetic-type Ib diamond containing substitutional nitrogen, which forms a
comparatively shallow trap. Since the donor energy is well above the Fermi energy of
most metlas, a depletion region will from in equibrium. When diamond is positively
biased, electrons can tunnel into the conduction band as illustrated in Figure 2.7. Once

electrons are populated in the conduction band of the diamond, due to NBA of diamond

electrons can escape into the vacuum with little to no electric filed, 0-1 Vum". This

concept is shown in Figure 2.8.
2.5 CVD Diamond Infrastructure

2.5.1 Nucleation

To grow diamond films effectively on non-diamond substrates, the substrate is usually
treated to enhance nucleation density. Abrasive polishing of the substrate with various
grain sizes of diamond [55, 56], c-BN, or SiC [57] powder is one of the commonly
utilized methods. Ultrasonic agitation of the substrate with diamond powder suspension

[58, 59, 60, 61] is another most commonly used pre-treatment technique. Bias enhanced

33

 

 

 

 

Figure 2.5. Simplified model of electron emission through
dielectric layer, where electrons tunnel through
metal/dielectric interface [40].

 

 

......
................

.__.~. a, ----- Q o- -o- -o-- -@-- -o- - o- -- F-c

  

" v .- -.lllll ....
m..." ...... . ................
an. O oooooooooooooo

DIAMOND

 

Ev

Figure 2.6. For substitutional nitrogen. Donors are ~l.7eV
below EC. In equilibrium a depletion region forms
at the metal-diamond interface. The circles repre-
sent the energy position of the donors [42].

34

DIAMOND

 

 

 

Ev

Figure 2.7. Energy levels for a metal-diamond interface with the
metal biased to -6V with respect to the diamond
[42].

    

" / .33.“:
METAL- DIAMOND- VACUUM-
DIAMOND VACUUM METAL
INTERFACE INTERFACE INTERFACE
4 0V
V 0.7 0V
1 4.5 eV
I

 

 

Figure 2.8. Schematic diagram of a diamond cathode in a diode
structure [42].

35

nucleation (BEN) has also been reported for effectively creating nucleation sites on silicon

substrates [62] and preparing highly oriented diamond films [63, 64]. Although these

methods successfully enhanced the nucleation density up to 10cm'2 [55, 59, 65] most of
these techniques cause surface damage and have a poor uniformaity [57].

Nucleation method used in this experiment is an IC-compatible nucleation technique
which does not cause surface damage [66]. Diamond Powder Loaded Fluids (DPLF) with
different carrier fluids, mean powder sizes and densities are applieded on substrates by
various means to enhance the nucleation density. The idea of this method is to spread
diamond crystals, suspended in carrier fluids, on the substrate surface. During the diamond
deposition process, the carrier fluids are evaporated at initial stages leaving behind the
diamond particles which act as seeds for diamond growth. DPLFs are applied to substrates
through direct writing, spinning, spraying or brush-painting. The technical details of two
different DPLFs are summarized in Table 2.2. Further detail can be found in Ref. [67].

The DPLF] are commercially available diamond powder suspensions, 1/40 SQG,
from Du Pont Chemicals-Repauno Plant, NJ. The DPLF2 is prepared by suspending
diamond powders into photoresist [68, 66]. In contrast to other nucleation procedures,
DPLF method does not cause any damage to the substrate surface. This method is highly
compatible with standard integrated circuit (IC) processes and amenable to patterning,

doping, and coating of 3-D samples.

2.5.2 Patterning

In order to fulfil a variety of applications and desirable structures, growth of patterned
diamond thin films is desired. Patterning techniques are, in general, distinguished into two
categories, selective deposition and selective etching.

The selective deposition is achieved by pre-deposition nucleation on the desired area,

or by masking the undesirable area during the diamond deposition. Hirabayashi et al. used

36

Table 2.2. Comparison of two types of diamond powder loaded fluids.

 

 

 

DPLFl DPLF2
Carrier Fluid Water Photoresist
Mean powder size 0.038 pm 0.101 mm
Density 4O Carats/liter 12 Carats/liter
Nucleation density ~10” cm'2 _108 cm'2

Application methods

Patterning

Spary, brushing, or direct
writing

Photolithography, spray
with shadow mask, or
direct writing

Spray, spin, or direct
writing

Photolithography, spray
with shadow mask, or
direct writing

 

37

Ar sputtering in undesired growth regions to suppress nucleation after substrates were

pretreated by ultrasonic method [61]. Ma et al. reported selective deposition onto Si02

stripes patterned on a Si substrate [58]. The Ar+ ion beam was used to suppress the

nucleation on both Si and Si02 except in the shadows cast by the downwind edge of the
Si02 stripes. The selective nucleation was successfully achieved by masking substrates

during the ultrasonic treatment by Masood et al. [66]. Leksono and coworkers reported
successfully the patterning of diamond films by lift off process on Si wafer. Thin ZnO film

and amorphous silicon were used as sacrificial layers for lift off processes [69]. Si02 or
Si3N4 as a masking layer during the diamond deposition was successfully reported by

Masood et al. [66] and Roppel and coworkers [70].
The patterning of continuous diamond films by selective etching method was reported

by Masood et al. [71,66]. Using a rapid thermal processor (RTP), diamond films were
successfully etched in oxygen environment at 700 0C. The SiOZ or Si3N4 were used as

masks to block the undesirable area during etching process. ECR etch of diamond using

02 plasma was performed as well [72].

However, most of these techniques, which involve selective etching or selective
deposition, either cause damages on substrate surfaces or have poor reproducibility.
Recently, a novel patterning process, photolithographic patterning techniques, has been
developed at Michigan State University to selectively deposit high quality polycrystalline
diamond films through standard photolithographic processes [66]. This technique not only
achieves excellent selectivity without surface damages, but is also compatible with the
existing integrated circuit fabrication technology. This technique is adopted in the present

research to pattern CVD diamond films.

min i
he ran
1m 2
innit}
was f0
Th5 ra

mge

r—J

(I.

38

2.5.2.1 Photolithographic Methods

The mostly used patterning technique in this research was to prepare samples coated
with DPLF2. In this technique, DPLF2 is spin-coated on substrates, Si or oxidized Si
wafer in most cases, and is patterned by standard photolithography. The spin rate was in
the range of 2000 - 4000 rpm. An enhanced diamond nucleation density was observed at
lower spin rate, however, the seeding uniformity seems inferior. When higher nucleation

density is required, DPLF2 was coated multiple times. An optimal spin speed of 3000 rpm

was found to possess a nucleation density on the order of 108 cm'2 with good uniformity.

The ratio of nucleation density in the undesired area and that in the desired area is in the

range of 7.57 x 10‘3 - 3.38 x 10‘4[67].

2.5.2.2 Direct-write Patterning Method

In direct-write patterning technique, DPLF] are used as seeding materials and are
applied to substrate surfaces via a fine spray nozzle. The minimum pattern size can be
varied by selecting the opening sizes of nozzles. In initial experiments, a capillary of wire
bonder, with an opening of ~40 um (1.5 mils) was fixed in front of syringe needle and was

used as a spray nozzle. The width of lines manually created on Si or oxidized Si wafers are

in the range of 125 - 600 pm. A substrate temperature of 65 - 75 0C is found optimal to dry
the seeding fluid, DPLF], before it spreads out. In case of DPLF2, spary gun with
compressed air was also used with shadow mask. G. Yang has designed and built

computerized direct writing system [67].

2.5.3 Doping

As a wide band gap semiconductor material, Eg=5.5 eV, CVD diamond films

deposited without intentional doping are usually good insulators. As a result use of

39

diamond as an electronic material requires a control over the ability to dope it p- and n-
type. In general, doping of semiconductors is achieved by diffusion, ion implantation, or
in situ doping. Because of the low diffusibility of most elements in diamond, diffusion is
impractical. Although ion implantation can achieve active doping, the damages of crystal
structure were introduced [73]. Therefore, in situ doping was commonly utilized in
preparing semiconducting diamond films. Boron (B), aluminum (Al), phosphor (P),

lithium (Li), and nitrogen (N) are potential dopants of diamond. Although the NH4H2PO4
and (CH3O)3P were reported to produce n-type semiconductive CVD diamond films, the

conductivity is too low for electronics application [74, 75]. Other attempts to obtain n-tupe
doping were unsuccessful until now [76, 77]. P-type diamond films were successfully
prepared by using born as an in-situ dopant.

A number of in situ doping techniques involving gaseous, liquid and solid sources to
incorporate boron into CVD diamond films have been successfully used to produce the p-
type semiconducting diamond films. Solid sources (B203 [78, 79], boron powder [68])
have the advantages of chemical stability and simplicity. However, non-uniformity of
doping are observed and it easily contaminates deposition chamber. Gaseous source

(BzH6 [80]) has a better controllability, but is poisonous and explosive. Liquid source
(B(CH3O)3 [78, 81]) is also used to prepare p-type diamond and is reported to be
nonpoisonous, stable, and controllable. Due to the safety concern, only solid sources are
used in this research.

High purity (5N8) amorphous boron powders and B203 boron wafers purchased from

Techneglas Co., Perrysburg, OH are used to fabricate semiconductive diamond films via in
situ doping technique in this work. Specially designed holders, as shown in Figure 3.4
later, made of 1 mm thick Mo plate were used to introduce boron powder into process
chamber. The doping sources were placed on substrate holder during the deposition.

Although relatively uniform doping profile was confirmed by secondary ion mass

speCUD'
wider
burs

n‘nn

nnnn
:kim
nenn

Th
fiAu
mm

s.

‘é-Lh E

40

spectroscopy (SIMS) measurement [71], depending on the relative position of born

powder holder, non-uniformity is found. A layer of undoped diamond under the p-type
layer was found to enhance the quality of p-type diamond films, measured in terms of sp3/

sp2 ratio by Raman spectroscopy, even for very high d0ping levels [71, 82].

2.5.4 Metallization

Metallic contacts are essential for achieving electronic devices. It is desirable that the
resistance of devices be determined mainly by the resistivity of the bulk material and that
the influence of contact resistance be minimized. In addition, the good adhesion is
essential for developing the contact structure.

The contact resistance is found to decrease with an increase in the doping level [83].
Ti/Au, Ta/Au, Ti/Pt, and Ti/Mo/Au were reported by several workers to provide good
contacts with diamond films [68, 84, 85, 86].

Since resistance to high temperature is not required for contacts in diamond field
emitter research and single target metal evaporation facility was not readily available, A]

were tested on the samples which diamond films were synthesized on SiOz substrates. A1
with a thickness in the range of 0.6 - 1 pm were thermally evaporated under a pressure less

than 10'6 torr. Patterns were defined either by shadow masks or by photolithographic
processes and wet etching.

After the metallization, samples with contact metals of A] were thermally annealed at
500 0C for 20 minutes, in N2 by tubular furnace constructed by G. Yang or rapid thermal
processor (RTP). Al contact layers show good adhesion on diamond and Si02 (if any).

The ohmic contacts of A1 with polycrystalline diamond films are confirmed by I-V
measurements at room temperature. A Keithley 224 programmable current source along

with a Fluke 8840A multimeter are used for I-V characterization. The voltage reading was

41

taken right after current was applied to minimize the self-heating effect especially at

current > 10’4 A.

CHAPTER 3

DEPOSITION AND
CHARACTERIZATION SYSTEMS

3.1 Introduction

Prior to the study of field emission (FE) from diamond film, it is necessary to build a
film deposition and an automated FE characterization system beforehand. Design and
construction of the deposition systems and the characterization system are described in
detail. In addition, film characterization techniques are described. Anode designs used in

this study are also discussed.

3.2 Deposition Systems

The ability to grow Optimum films for FE is essential for the realization of diamond
field emitter application. With the exception of some samples, all diamond films used in
the present research were prepared by a hot filament chemical vapor deposition system
(HFCVD) [87]. In addition to the description of HFCVD system, the modification of

RFCVD system are described in this section.

3.2.1 Hot Filament Chemical Vapor Deposition System

The HFCVD reactor used in this work was designed by S. Sahli and A. Masood after a

42

43

similar system used at Ford Motor Company’s Research Lab. It was constructed at
Michigan State university by S. Sahli [88] and D. Hong. Original system at Ford can grow
only up to two-inch samples. In the design of new system, consideration was given to
grow uniform diamond film up to four-inch wafers. Some modifications and optimizations
were incorporated during the course of this work to improve the reproducibility,
deposition uniformity, growth rate, and film quality. A simplified schematic diagram of
this system is shown in Figure 3.1. Based on the operational function, a HFCVD system

can be categorized into the following five major subsystems.
(1) Chamber
(2) Filament
(3) Sample stage
(4) Vacuum system
(5) Pressure and flow controls

0 Chamber

The HFCVD reactor consists of an 18-inch diameter stainless steel vacuum chamber
with a 10 inch diameter access door. 1/4 inch stainless tubing is wrapped around the top
portion of the chamber for cooling purpose. When it is in operation, additional cooling by

blower is necessary to prevent the chamber from overheating.

0 Filament

In original design, ten 5-inch long and 0.005-inch thick Ta wires in a parallel

configuration were horizontally mounted on a 4.5 x 5 inch2 frame via 20 Mo hooks. The

1 mm
DIAMETER
HOLES

coco
ooooo
I F

BORON POWDER HOLDER

1

 

 

 

 

 

 

 

H2: :03: ‘ THERMOCOUPLE
CH4 CONT. / SUBSTRATE
|

 

  
     
 

/ PYROMETER

 

 

 

 

 

TEMP.
CONTROL

 

 

 

 

N2
lVACUUM

 

 

CONTROL

PRESSURE *
PUMP

 

 

 

Figure 3.1. Schematic of HFCVD system constructed.

45

supporting frame consists of two Mo bars which serve as electrodes. The 4.5 x 5 inch2
filament array was utilized to ensure a uniform deposition over a 4 inch wafer. This
configuration caused an unacceptable amount of sagging due to lack of tension applied to
thin wires. Same configuration with 0.02-inch thick wires was attempted. This new
scheme reduced the amount of sagging within tolerable range. However, the total current
exceeded the specification of 50 A power supply. The supporting frame had to be modified
to the combination of two BN bars and two Mo bars to accommodate filaments in a series

configuration such that total current can be reduced down to less than 50 A. The filament

typically draws l8 - 22 A current at nominal temperature in the range of 2200 - 2400 oC.
The filament temperature is monitored by an optical pyrometer (Williamson 8125PS-G-C)
through a 6 inch glass view port. Monitored signal is fed into Yokogawa UT35 controller
to maintain filament temperature at specified value. Enhancement efforts are being done to
change filament assembly in a vertical fashion that would eliminate sagging of the

filament.

0 Sample stage

Original sample stage used was 2” diameter heater assembly powered by Research
Inc.’s type 663F power supply. Temperature was measured with type K thermocouples at

both the t0p and back side of substrate during deposition process. Sensed signal was

feedbacked to Partlow MK2000 controller. It was found that 4.5 x 4.5 inch2 filament array
generates enough heat to maintain substrate at normal deposition temperature without
heating the substrate. In order to grow diamond film on four-inch wafer, sample stage was
changed to four-inch wafer, graphite plate, or graphite column. The supporting frame was
specially designed with a vertical movement mechanism to adjust the separation of
filament array and substrate in the range of 0.5 - 1 cm. Due to the sagging of the filament,

temperature at the surface of the substrate was expected to be non—uniform. A temperature

46

profile of four-inch substrate was characterized by measuring temperatures at several

points of substrate using thermocouples during a test deposition. A ~100 °C difference
was found between the temperatures measured at the center and the edge of the top side of
substrate. The substrate temperature then was monitored at the edge in the actual
deposition to minimize the blocking area by thermocouples. ANSYS simulation of

sagging filament also confirmed that difference between simulated temperatures at the

center and the edge of the top side of substrate is ~120 °C. This value almost coincides
with the measured value. Figure 3.2 shows simulated structure and temperature plot
respectively. During the test run, filament temperature was controlled. It later turned out to
be inefficient due to the following reasons. First, due to the sagging of the filament as a
function of time, focusing aperture of pyrometer didn’t measure same area of filament.
Secondly, small vibration of the system could change temperature reading drastically
which would enormously fluctuate filament current. It was found, after numerous test run,
that controlling substrate temperature is better method to achieve more uniform diamond
film. Later, machine was reconfigured, so that thermocouple signal of substrate

temperature is fed into filament controller.

0 Vacuum system

A Varian SD-200 rotary pump is used for evacuation. This configuration is sufficient

enough to evacuate chamber down to 10 mTorr within 30 minutes. In addition, N2 is used
to purge and backfill the chamber.
0 Pressure and flow controls

The reactant gas mixture consists of ultra-pure grade (99.995%) methane (CH4) and

hydrogen (H2). The flow rates of these gases are independently controlled by MKS type

47

 

 

 

 

 

 

 

Temperature
C:
:23
-
Filament =
-
-
v -
' " if “hotter
Substrate
(a)
880 I . r .
860 ~ .
g 840- -
9
:3 __ -4
E 820
(D
E
(”800- ~
.—
780 ~ 5
760 - 5
-l.5 -0.5 0.5 1.5

Distance from center (inch)
(b)

Figure 3.2. (a) ANSYS simulation of temperature profile due to sagging fila-
ment. (b) Temperature of the substrate as a function of the dis-
tance from the center.

48

1159B mass flow controllers and read by MKS Model 247C 4 channel readout. The
operation pressure is monitored through a MKS type 122A baratron pressure gauge and is
controlled by a MKS type 250 pressure controller via a type 248A upstream valve. The
base pressure, measured by a MDC thermocouple vacuum gauge, is in the range of l - 10

mTorr.

° Deposition parameters

Since CVD growth of diamond is essentially in metastable phase, the selection and
consistency of processing parameters are very important in achieving reproducible high
quality diamond films [8]. The ranges and typical values of deposition conditions used

throughout this research are summarized as follows. Values in the parenthesis are typical

values used:
(1) Gas composition: CH4 : H2 = 0.5% - 2% (1%)
(2) Gas flow rate: CH4 : l - 5 sccm (4 sccm)
H2 : 200 — 500 sccm (400 sccm)
(3) Filament temperature: 2200 - 2400 °C (2300 0C)
(4) Substrate temperature: 850 - 950 °C (900 °C)
(5) Operation pressure: 40 - 50 Torr (50 Torr)
(6) Substrate to filament distance: 0.5 - 1.5 cm (1 cm)
(7) Doping source: Boron powder or 8203 boron wafer

0 Operation procedure

49

The operation sequence of HFCVD is reported to have a great influence on the

deposited film quality [71]. The operation steps used through out this research are

described as follows.

Start up procedures:

(1) Samples to be deposited and doping sources (if any) are loaded on sample holder.

(2)

(3)

(4)

(5)

The distance of filaments and substrate is adjusted to be ~l cm. (The distance

varies with the sag of filaments.)

Process chamber is evacuated down to a base pressure of less than 10 mTorr.

Hydrogen is then introduced into the process chamber.

When the chamber pressure reaches > 20 Torr, the filament temperature is

brought up slowly to the nominal temperature of 2300 °C. The slow adjustment
of filament current is desired to avoid any damage and to ensure long life time of

the filament.

CH4 is switched on only after the substrate temperature reaches desired

temperature of 900 0C.

Shut down procedures:

(1)

The CH4 is turned off to terminate the deposition processes, while the H2 remains

on for another 5 to 10 minutes. During this period, the hydrogen is believed to be
the preferential etchant to remove a thin conducting carbonaceous layer on the

surface of CVD diamond films [71]. This step makes the surface H2 passivated.

No conduction is found in the case of undoped films after this treatment.

50

(2) The applied filament current is decreased slowly down to zero. The chamber is
then evacuated to vent out process gases. Samples are unloaded after the system

is cooled down and backfilled with N2.

Typical morphology and Raman spectrum of the film grown by HFCVD is shown in
Figure 3.3.

3.2.2 Radio Frequency Chemical Vapor Deposition System

A radio frequency chemical vapor deposition (RFCVD) reactor is currently under
construction in collaboration with Dr. N. Abu-Ageel (Appendix A). RFCVD systems,
used for diamond growth in the late eighties and early nineties, were found to lead to
lower quality films [89, 90, 91, 92]. As such films are more appropriate for FE, an existing

Plama-Therm system is being modified.

3.3 Film Characterization

The quality of diamond films was monitored by secondary electron microscopy
(SEM), Raman spectroscopy, and atomic force microscopy (AFM). The SEM and AFM

provided a direct vision of film surface morphology. The Raman spectroscopy provides a
direct evidence of existence of diamond through its characteristic peak at 1332 cm]. Peak

values at 1332 cm'1 and 1580 cm‘1 were compared to estimate sp3/sp2 ratio. AFM was

used to inspect small grain films.

- Raman spectrosc0py

In Raman effect, discovered by physicist Sir Chandrasekhara Venkata Raman of India,
light is scattered by an atom or a molecule that changes its state. There is a corresponding

discrete decrease in the frequency of scattered radiation. Raman spectroscopy is widely

51

ISKU >416,8€1€l

 

 

€4.5-

Relative intens
A

OD

 

 

 

 

2 E r l l r r
T1 00 1200 1 300 1 400 1 500 1 600 1700
Wave number shift (1/cm)
(b)
Figure 3.3. Typical morphology by SEM (a) and corresponding Raman
spectrum (b) of the film grown by HFCVD.

so:
of t

5
{CL

31
Tl
I“

52

used in the analysis of materials and the identification of trace elements. In diamond
related research, Raman spectroscopy is a powerful tool to identify the presence of sp2 and

sp3 bonding in diamond films [93]. The wave number shift of I332 corresponds to sp3
diamond peak.

Raman system used in this research was built by J. Mossbrucker using monochrometer
donated by Ford. It consists of laser, optics, monochrometer, photomultiplier tube,

counter, and computer. The 514.5 nm argon laser line (green) is used. It has a laser power

of 600mW and a spectral resolution of 9 cm". The monochrometer is Czemy-Tumer

scanning type where both gratings are mounted on the same axis.

0 Scanning electron microscopy (SEM)

SEM consists of an electron gun and electron detectors in a vacuum chamber. Images
are constructed by collecting the secondary electrons emitted from samples due to the
incident electron beam. The image then is displayed on a cathode ray tube (CRT) for
direct vision [94]. Because of the large magnification range of 10 X to 300,000 X [95] and
great depth of field, SEM is a very powerful tool for studying the morphology and visual
analysis of diamond films. It is widely utilized to inspect the surface morphology, crystal
orientation, grain sizes, and film thickness. It is also used to monitor the nucleation
densities and patterning. The need for a conducting specimen somewhat limits its utility
for undoped films on insulating substrates. Three different SEMs were used for this
research.

Jeol SEM with W filament at Composite Material and Structure Center was used
often. It is simple to operate and it can accommodate large size samples. Jeol SEM at

Electron Optics Center is equipped with LaB6 gun. It makes saturation of filament easy.

Another advantage is that it is fully computer controlled. As a result, it is easy to save

images into data file. Usage of this machine was hindered by the size limitation of the

33
Cr

'53

Cfrf

1r .
I“ L

53

sample that can be loaded. Environmental SEM at Composite Material and Structure
Center was also used for some samples. This special SEM allows non-conducting
samples. It was beneficial to load samples without conductive layer coating. It was
especially useful when testchip is fabricated. After each step of wafer processing, SEM

picture could be taken without the need for the conductive layer coating.

0 Atomic force microscopy (AF M)

Atomic Force Microscope (AFM), also called as Scanning Force Microscope (SFM),
can measure the force between a sample surface and a very sharp probe tip mounted on a
cantilever beam having a spring constant of about 0.1-1.0 N/m, which is more than an
order of magnitude lower than the typical spring constant between two atoms. Raster
scanning motion is controlled by piezoelectric tubes. If the force is determined as a
function of samples’s position, then the surface topography can be obtained. Detection is
most often made optically by interferometry or beam deflection. In AFM measurements,
the tip is held in contact with the sample. Spatial resolution is a few nanometers for scans
up to 130 run, but can be at the atomic scale for smaller ranges. Both conducting and

insulating materials can be analyzed without sample preparation.

3.4 Field Emission Characterization System

As field emission characterization systems are commercially not available,
characterization system is designed and built. It consists of vacuum chamber, pumps,
pressure gauges and readouts, power supply, measurement instruments, and computer

system. Schematic of the system is shown in Figure 3.4.

- Chamber

8.5-inch diameter vacuum chamber was machined at the Physics shop. It has 8 flange

 

 

Pressure

54

Vacuum Chamber

I”

 

Readout

 

 

 

 

 

 

_1

Pressure

 

Sample :r—J

 

Anodir

 

 

 

 

Gauge

 

 

 

Gate Valve __0

 

 

 

 

 

 

 

Cooling
Unit

 

 

 

Turbo Pump

 

 

 

Water line

 

 

 

 

Rotary Pump

 

 

1

Exhaust

 

 

l

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I-limitin g
Resistor V meter
J
J.
I meter
Power
Supply
Daisy-chained
IEEE488

 

 

 

 

 

 

 

Figure 3.4. Schematic of automated characterization system built.

C01

5i;

TL-
14?.-

55

connections and can be accessed from the top. The following ports were attached to the

eight flanges.
(1) Gate valve
(2) Thermocouple pressure sensor
(3) Barotron pressure gauge
(4) Electrical feed through
(5) Mechanical feed through
(6) Ion gauge
(7) N 2 purge
(8) View port

0 Vacuum system

The chamber, equipped with two stage pumping system, is connected to the turbo
pump (Ley bold-Heraeus Turbo Tronik NT150/360) through gate valve. Turbo pump is
water cooled by closed loop Neslab CFT-33 refrigerated recirculator. Varian SD-200

mechanical pump is connected to the turbo pump in series. With this configuration

vacuum chamber can be evacuated down to 10'6 Torr range.

0 Pressure readout

Four types of vacuum readouts were used to check different ranges of vacuum levels.

They are summarized in the Table 3.1.

56

Table 3.1 List of pressure readouts.

 

 

 

 

 

 

 

Readout type Range
MKS barotron 122AA-01000AB with PDR-C- 2 - 1000 Torr
2C MKD power supply readout
MKS barotron 122AA-00002AB with PDR-C- 1-2000 mTorr
2C MKD power supply readout
MDC thermocouple vacuum gauge 1 - 1000 mTorr
Varian 0571-K2471-303 ionization gauge 10-8 _ 10 -4 Torr
Table 3.2 List of power supplies.
Model Rating $333.33.,
Hewlett Packard Harrison 61 10A DC power 3000V, 5mA No
supply
Keithley model 248 high voltage supply SOOOV, 5mA Yes
Keithley 230 programmable voltage source 100V, lOOmA Yes
HP model 712B power supply 500V, 200mA No
Universal Electronics regulated power sup- 500V, 200mA No
ply model 520A
Keithley 220 programmable current source lOOmA, 100V Yes

Keithley 224 programmable current source lOOmA, 100V Yes

 

57

Table 3.3. List of measurement instruments.

 

 

 

Model Computer controllability

Fluke 8506 Thermal RMS digital No
multimeter

Keithley 181 nanovoltmeter Yes
Keithley 169 multimeter No
Fluke 77 multimeter No
HP 3435 digital multimeter No
HP 34757 multimeter Yes
Keithley 595 quasistatic CV meter Yes

Fluke 8840A multimeter No

 

5.“
TL!‘

CIT

I75,

58

0 Power supply units

Various power supply units were used. Two different types of power supplies used
were DC volt and DC current. All the power supplies used were listed in the following

table 3.2.

0 Measurement readout units

Measurement readout was performed by either measuring voltage drop over the

resistor or measuring current. Instruments used are listed in Table 3.3.

0 Computer system

A computer with IEEE 488 interface card was employed to automate the
characterization system. Computer, power supply, and measurement instruments were
daisy chained as shown in Figure 3.5. Control program was developed by using BASIC
program with built-in library functions furnished with IEEE 488 card vendor, Hewlett
Packard. Source code is shown at appendix A. This control program allows automatic
measurement.

An effort to build new characterization system is led by C. Koellner. This system is
similar to one already discussed. Major addition to this new system is fully computer
controlled XYZ stage with a resolution of 1 pm and microscope with camera for field

emission pattern acquisition. This will enable us to map the wafer in a great detail.

3.4.1 Anode Designs

Six different types of anodes were used to characterize samples. Each has its own pros
and cons. The best anode type for a particular measurement situation was selected. They

are shown in Figure 3.5.

15' q
Last

.‘ TI
p13

59

3.4.1.1 W Probe Anode

Figure 3.5 (a) shows tungsten probe anode used in the characterization of discrete
samples, W probe was used often due to its simple setup. Sample is fastened on substrate
holder and probe is placed on top of the sample. Advantage of W probe anode is that there
is no insulating material involved in but the separation between anode and emitter is not

precisely controlled. Two SEM pictures of two W probes used are shown in Figure 3.5 (b).

3.4.1.2 Brass Column Anode

In this type of anode, insulating material is placed between sample and the mirror
grade polished brass column. In some cases, Si wafer coated with Al was substituted by
brass column. This configuration is shown in Figure 3.5 (c). Its advantage was that we
know the separation precisely by knowing the thickness of the insulating material, which
is quartz sheet. Usual thickness of the Q2 plate used is in the range of ~100 um. In some
cases, 100 pm thick plate is etched with a solution of HF mixed with water to get thinner

plate such as 50 um.

3.4.1.3 Lateral Anode

In this case, MKS mechanical feed through was used. Its maximum travel range is 1
inch and minimal resolution is 25.4 um. Sample is mounted on the sample holder with
clamp and positioned in a vertical fashion. Detailed setup is shown in Figure 3.5 (d). At
first, anode and emitter contact was made. Contact were detected by measuring resistance
between anode and cathode. Upon finding contact between anode and cathode, 1 notch of
mechanical feedthorugh was moved back to make a separation of 25.4 um and another 1

notch of mechanical movement was added to get 50.8 um separation.

(II .

A

I'll

'AJ

60

3.4.1.4 Glass Anode

Figure 3.5 (e) shows that final version of glass anode made in collaboration with Dr. S.
Kwon. Samples are placed on top of the glass anode upside down. Two mask fabrication
steps were required to make the glass anode. First mask defines the grooves in glass.
Depth and width of the groove can be controlled precisely by the etching time. Sodalime
glass were used. Next, Al is thermally evaporated on the glass and second mask is used to
remove Al outside groove area. With small gap between anode and emitter it was possible
to demonstrate low voltage emission. This glass anode was very useful to calculate current

density.

3.4.1.5 Phosphor Anode

This type of anode was provided by Zenith Corporation. Transparent conducting layer,
Indium-Tin-Oxide (ITO), is coated on 1 x 2 inch sodalime glass. On top of ITO, phosphor
material (ZnO:Zn), which transfer electron energy to light emission, was coated. Samples
had to be separated by quartz insulator. This phosphor anode is very useful to record
emission spot and to check whether or not measured current is emission current quickly. It
is expected that we only see visible spots when forward biased, if it were real emission
current. If visible spots were found at both same forward and reverse bias, it would be

discharge current due to vacuum breakdown. Phosphor anode is shown in Figure 3.5 (f).

3.4.1.6 Built-In Anode

Some discrete samples and field emitters in testchips includes built-in anode, so that
no other anode system is necessary. This will be discussed in Chapter 4 in greater detail.

One example of built-in anodes, a bridge shaped Al anode over emitter, is shown in Figure

3.5 (g).

61

W probe

 

 

  
 

quartz spacer Diamond

 

 

Figure 3.5. Six different anodes used in this study.

62

 

 

 

(d)

 

   
     
   

Sample in
upside down
position

 

(e)

 

Figure 3.5. Continued.

63

Glass coated with ITO

   

ZnO:Zn phosphor

 

 

 

  
 

Emitter elec o o .

 

 

(g)

 

 

 

Figure 3.5. Continued.

3.5 Summary

In this chapter, the design and construction of HFCVD system is discussed in detail.
Modification of RFCVD for low temperature diamond deposition is described as well.
Low temperature deposition is especially important to meet low cost, large area diamond
deposition, which is crucial to commercialization of diamond FED in the near term. In
section 3.3, an overview of commonly used film characterization tools, SEM, Raman
spectroscopy, and AFM is given. In section 3.4, the design and construction of an
automated characterization system is explained. Finally, six different types of anodes used

in the measurement are introduced.

CHAPTER 4

IC COMPATIBLE FIELD EMITTER
TECHNOLOGY

4.1 Introduction

In this chapter, new patterning techniques for DPLF] and enhanced technique for
DPLF2 are beriefly mentioned. Chronological efforts to achieve IC compatible field
emitter technology are discussed. Starting from preliminary and simple discrete samples
to establish key technologies, three unique testchips were designed and fabricated.

Detailed explanations of these three testchips are given.

4.2 New and Ehanced Patterning Technique

This section deals with the modification of photolithographic patterning technique to
incorporate patterning of samples coated with DPLF]. In addition, problems, difficulties
and proposed solution are also included in this section.

In the case of DPLF1,both a combination of photolithographic methods and direct
writing methods are used for patterning. For a photolithographic method for DPLFl, as
shown in Figure 4.1, a sacrificial layer of photoresist is uniformly spin-coated on the
substrate with a high spin rate of 4500 rpm. This photoresist layer is for preventing

diamond powder from touching the substrate directly. The seeding fluid then is manually

65

66

DPLF1

 

 

 

 

OOOOOOOOOO

 

Ilf’ii‘iell

(b)

 

(a)

Figure 4.1. Process sequence for the patterning of (a) DPLF1 and (b)
DPLF2 using photolithographic method.

67

brushed over. The second layer photoresist is spin-coated at a speed of 4000 - 4500 rpm to
prevent the seeding material removed during the developing process. Due to the poor
transparence of DPLF1 , a fully exposure of the first photoresist layer prior to the coating
of seeding fluid is helpful to improve the selectivity. However, it may cause the problem of
over developing resulting a poorly defined pattern edge. The schematic of DPLFs
patterning procedure by standard photolithography is shown in Figure 4.1. SEM pictures
of diamond structures prepared by both patterning procedures are shown in Figure 4.2 for
comparison. The wavy edge observed on DPLF1 pattern is mainly caused by the over
developing of the first layer photoresist and the non-uniformity of DPLF1 seeding fluid
coated by hand brushing. A uniform and well controlled application technique is essential
to improve the pattern quality of DPLF1 seeded samples.

As shown in Figure 4.1 (b), there are a number of diamond stray particles in unwanted
area when DPLF2 was applied. These are due to remaining diamond particles, which are
not washed away completely during the development process of the diamond-loaded
photoresist. They can affect the yield if they form a continuous film and if their size
becomes too large. A double layer process is developed to prevent diamond particles from
reattaching to the surface in undesired areas. Comparison between conventional method
and double layer method is shown in Figure 4.3. Essence of double layer process is that
pure photoresist is coated underneath DPLF2 to prevent diamond particle from touching

the substrate directly.

4.3 Preliminary Samples

Early diamond field emitter samples were B-doped polycrystalline films on p-type Si
wafers. Nucleation is achieved by spin coating of DPLF2 as described in section 2.5.1.
One group of samples are not patterned in order to grow continuous film on the substrate.

Another group of samples were patterned using spray method with shadow mask

68

 

Figure 4.2. Patterning results of (a) DPLF1 and (b) DPLF2 samples by
photolithography.

69

NEW
CURRENT DOUBLE-LAYER
METHOD METHOD

 

 

.. mmm 11111111

 

 

(b)

Figure 4.3. Comparison of (a) conventional patterning method and (b)
double layer method.

70

described in section 2.5.2 to create a collection of diamond dots on the substrate. Cross

sectional view of the samples are shown in Figure 4.4. The samples are subsequently
placed in the HFCVD reactor. The filament temperature in HFCVD was 2300 °C which

maintained the substrate temperature, sensed by a thermocouple, at ~900 0C. The

deposition atmosphere consisted Of 1% CH4 in H2 at 50 Torr. The separation between

filament and substrate was approximately 5 mm. Using a diamond deposition rate of
approximately 0.25 pm per hour, a film thickness of 2 um was achieved. In—situ doping
Of boron was accomplished by placing a small container with pure boron powder on the

substrate holder as describe in section 2.5.3. Such an arrangement can provide resistivity
values in the range of 20-100 52cm [96]. Al was thermally evaporated on the backside of Si

wafer to provide an ohmic contact. The samples were annealed at 400°C in nitrogen for 30
minutes. With the exception of diamond deposition, all the sample fabrication steps were

completed in a class 100 area in the cleanroom.

4.4 Discrete Sample with Built-In anode

Two types of simple samples with built-in anode were fabricated, namely vertical
structure and lateral structure. Although it is required to pattern diamond film and Al
anode, very crude first sample with built-in anode was made even without any mask or
with simple transparency mask, metal shadow mask, or crude shadow mask made by Al

foil.

4.4.1 Vertical Structure

Figure 4.5 (a) shows top view and (b) shows magnified cross sectional view along A-
A’ of the first diamond field emitter structure with built-in anode fabricated on a piece of

oxidized Si wafer. Figure 4.5 0)) shows step-by-step coross-sectional view. A 3-micron

71

 

 

Al
(a)

Diamond

 

 

 

 

Top view

 

 

 

Al

 

(b)

Figure 4.4. Cross sectional views of preliminary samples with (a) continuous film
and (b) dot patterned film.

72

Diamond

_______ s102

 

Si

 

 

 

Photoresist

_1:]_

 

 

 

 

Al

 

 

 

 

 

 

 

 

iflfium 988383

 

Figure 4.5. (a) Top view, (b) cross-section view, and (c) magnified view of section A-A’
of vertical type discrete sample with built-in anode.

73

thick layer Of SiOZ, deposited at 400 0C on p-type (100) Si, is annealed in N2 at 1000 0C

for 30 min to improve its insulating properties. Patterning and nucleation of diamond was
achieved by manual direct writing of DPLF1 or DPLF2 using a syringe method described
in 2.5.2.2. Structure shown in Figure 4.5 used DPLF2. B-doped and strip-shaped patterned
diamond film is grown on top of substrate. Photoresist, patterned by syringe again, was

applied on top of the diamond film. Width of this sacrificial photoresist is wider than the

width of diamond line. Photoresist is baked at 120 °C for 30 minutes to harden it before
putting thin Al layer on it. Using shadow mask, Al layer is evaporated on top of
photoresist layer. Width of Al line is narrower than underlying photoresist layer. Thus
photoresist underneath Al layer can be removed. To remove photoresist only without
attacking Al, either acetone or Shiply photoresist remover was used. After photoresist is
removed, bridge shaped A] anode is constructed over diamond film with vacuum equal to
the thickness of sacrificial photoresist layer. A] is also selectively evaporated at the ends of
each diamond lines to make ohmic emitter contacts. Same structures were also fabricated
on top of quartz (QZ) plate as shown in Figure 4.6. Adhesion of diamond film to Q2 plate
was as good as to oxidized Si wafer. Two vertical devices are on QZ plate, one in the far
left and another one in the far right. Color Of diamond lines in these two vertical devices
are different. One in the left used DPLF1 and one in the right used DPLF2. This simple
device enables us to characterize its current versus voltage without any external anode

attached. It permits great accuracy of anode to emitter distance and emitting area.

4.4.2 Lateral Device

Figure 4.7 (a) shows cross sectional view along with A-A’ and (b) shows top view of
the first lateral diamond field emitter structure with built—in anode fabricated on a piece Of
oxidized Si wafer. B-doped and patterned diamond film is grown on top of oxidized Si

wafer. Sample is then covered with thermally evaporated Al. To create the distance

74

A vertical device with A vertical device with
high n leation density film low nucleation density film

        

Two lateral devices with

Two lateral devices With low nucleation density film

high nucleation density film

Figure 4.6. Discrete samples with built-in anode grown on quartz substrate.

75

Diamond

SiO2

 

 

Si

 

 

AI

K1
(9
3‘)
v4
('3‘.
Si

1 QBt-km

 

(b)
Figure 4.7. (a) Cross sectional view and (b) top view of lateral device.

76

between anode and emitter, photolithography and subsequent etching was used to pattern
A]. It was impossible to use shadow mask due to the difficulty of aligning of small
dimension. Transparency mask, which has a gap between dark pattern of ~125 um, was
used. As shown in Figure 4.7 (b), minimum gap created is in the range of 20 um. Emitter

contact is Al on part of diamond film and anode contact is Al on SiOZ. Four lateral

structures made on QZ are shown in Figure 4.6. Again, two in the left side are nucleated
and patterned with DPLF1 and two in the right side were by DPLF2. Compared to vertical
structure described in the previous section, immediate advantage of lateral device is its
high yield. All the lateral devices fabricated exhibited successful Operation while some of
vertical structures suffered collapsed bridge partly due to surface tension during sample
drying process and partly due to wafer handling error. Another advantage was that it was
easy to control the gap between anode and cathode by shifting alignment. In case of
vertical device, the gap between anode and cathode was controlled by the thickness of

photoresist.

4.5 Testchip I

Encouraged by the work in section 4.4, a l x 2 inch2 testchip was designed and
fabricated. Two chips were placed on 4 inch wafer to check uniformity of diamond film
quality over 4 inch wafer from newly built HFCVD system. Each chip consists of eight
identical vertical devices. Figure 4.8 (f) shows fabricated 4 inch wafer. Each emitter is
fundamentally same as one shown in Figure 4.5 except the fact that size is well controlled
in the testchip. Figure 4.8 (a) - ((1) show the top view of four masks used in the fabrication.

Layout of masks was done using Framemaker and pattern size at during layout was ten
times larger than size appeared on the wafer. Each layer is printed in the high resolution
black and white printer. It was reduced two times due to the limitation of reduction lens Of

the equipment, and printed on transparency at the Instructional Media Center. These

77

4.
+
II

II
+
+

(a) (b)

llll, +nllfi.

 

 

 

 

u.
I
I

 

 

 

 

‘llll“ *iil

(c) (d)

 

 

l—ll—I

 

 

 

 

 

 

 

 

 

(6)

Figure 4.8. Test chip I (a) Diamond pattern (b) Sacrificial photoresist pat-
tern (c) Metatl pattern I (d) Metal pattern II which covers most
of diamond with metal (e) Overlapped view (I) completed wafer.

78

transparency masks were cut and pasted on 1.5 mm thick 4 x 4 inch glass. Minimum
feature size possible with this transparency mask was in the range of 50 um. Although the
contrast of the pattern in the trancyparency masks was not as high as Cr mask, use of
tranceparency masks in the design rule of > 50 pm is acceptable and used successfully
before [71].

Three mask steps were involved to fabricate this sample. First mask (Figure 4.8a)
defines diamond line. Second mask (Figure 4.8b) patterns sacrificial photoresist layer.
Third mask (Figure 4.8c and d) cuts Al layer. Two different types Of third mask were used.
As shown in Figure 4.8 (c), metal pattern I has Ohmic emitter contact only at the two ends
of diamond strip. Metal pattern 11 shown in (d) has A] pattern covering almost all the
diamond line except emitter area reducing resistance. It is expected that electrons move
either bulk or surface of the diamond film from the metal ohmic contact to the emission
area when metal pattern I is used. However, in case Of metal pattern 11, due to high
conductivity of metal on top Of relatively low conducting diamond, it is expected that
electrons prefer to move through metal until they reach emitting area. Different emitting
behavior due to two different conduction mechanism is postulated. Nucleation and
patterning of diamond line was done by photographic method using DPLF2. Figure 4.8 (e)

shows overlapped schematic of the original design. Completed wafer is shown in Figure

4.8 (f).

4.6 Testchip 11

After confirming uniformity of HFCVD system, 1 x 1 inch2 testchip with three-mask
step was designed and fabricated to include lateral device and have emitters in different
geometric values such as size of emitter area and anode to emitter distance. Figure 4.9
shows top view of designed pattern and corresponding pictures taken under the Optical

microscope.

(c)

 

'l-

 

(b)

-.|'i‘.'.—'—". -."_—.T1'l.-

TEE-

(d)

I
I
0

:ln‘\' ""1 4 7;.
llll ‘ I Ht“
9

(1)

Figure 4.9. Test chip II (a) Diamond pattern (b), wafer picture correspond-
ing to (a), (c) sacrificial photoresist pattern ((1) Metal pattern (e)
Overlapped view (1') completed wafer.

80

Identical to masks in testchip I, first mask (Figure 4.9a & b) defines diamond pattern,
second mask (Figure 4.9c) sacrificial layer, and third mask (Figure 4.9d) metal layer.
Again transparency masks were used for testchip II. Upper left comer has 16 vertical
emitters. Widths of diamond lines are 50, 100, 150, and 200 um, respectively. Widths of
Al lines are same as diamond lines from top to bottom. As a result, nine different
combinations of emitter area are created. This matrix combination was intentionally
designed to compare current density from different diamond emitter area. In a vertical
device, separation between anode to emitter, d, is same for all 16 structures. (1 is same as
thickness of sacrificial photoresist layer.

Other three quarters Of the testchip were devoted to the lateral devices. Every lateral
emitter has different separation between the anode and emitter structures varying from 10
to 50 um. Emitters at upper right comer have wider diamond line with small saw teeth like
pattern. Emitter area for lateral type emitter is diamond thickness multiplied by width of
diamond line. Emitters at lower right comer have pointed shape. Two saw shaped emitters,

one in lower left and another in lower right comer, have largest emission area.

4.7 Testchip III

Upon testing all the necessary components of fabrication technology throughout
previous sections, a diamond field emitter testchip ID was designed and fabricated. A
completed 4 inch wafer is shown in Figure 4.10 (a). Composite chip consists of four
different areas, MEMS area, piezoresistive sensor area, thermistor and heater area, and
field emitter area. Field emitter area includes two types of vertical emitters, lateral devices,
diode type FED, triode type FED, and pressure sensors. FEDS and pressure sensors will be
discussed in Chapter 6 in greater detail. Figure 4.10 (b) shows magnified views of all kinds
of devices included in the testtestchipchip IH. Total number of masks was six. Detailed

information is summarized in Table 4.1. Not all layers were was necessary to build each

81

    
 

  

lulu-l-

 

ypc A verticu - .
. \4 .
emitter -

Triddc FED cell

.
W‘éaém’q

Figure 4.10. Variety of structures on test chip III.

82

Table 4.1. Summary of mask steps for each structure in test chip III

 

Mask Pattern Type A Type B Lateral Diode Triode Pressure
vertical vertical

 

 

number defined . . device FED FED sensor
devrce devrce
1 Oxide etch x x x x
2 Diamond x x x x x x
pattern
3 Cr electrode x x
4 Sacrificial x x x x
layer
5 lst Al layer x x x x x

6 2nd Al layer x

 

83

structure. For example, mask number 3, Cr electrode patterning and mask number 6,
second Al patterning were not necessary for type A vertical structure.

Composite chip was designed by layour editor L-Edit and saved in GDSII format by
each member of the group [97]. D. Hong finally combined all the data. This data was later
merged together and converted to MEBES format to generate chrome masks using
electron beam exposure system. Four inch chrome mask set with minimum feature size of
4 mm was generated at Samsung (S. Korea).

Although some techniques might have been explained briefly in the previous sections
fabrication steps will be explained in greater detail again in this section because this
section is one of the main contribution of this research. Two different types of vertical field
emitter structures were fabricated. For the type A, targeted for pressure sensor
applications, p-type diamond film was grown directly on p—type Si. For the type B,
targeted for display applications, the diamond film was grown on oxidized Si and
chromium was used as the contact to the diamond film.

The fabrication process for the types A and B is depicted in Figure 4.11 and Figure

4.12 respectively. A 3-micron thick layer of SiOz, deposited at 400 °C on p-type (100) Si,

is annealed in N2 at 1000 0C for 30 min to improve its insulating properties. For the type A

structure, the oxide is patterned using buffered oxide etch and is used as a mask for Si
etching (Figure 4.11a). Before the diamond growth, the seeding and patterning is accom-
plished by photographic method using DPLF2 described earlier. The mixture is spin-

coated and patterned using a standard lithographic process. As the diamond is typically

deposited at 900 °C, the photoresist evaporates leaving behind the diamond particles
which act as seeds for diamond growth by the CVD process [66]. Using HFCVD reactor,

p-type polycrystalline diamond is grown on Si for type A devices (Figure 4.11b) and on

 

 

 

Oxide
l Si |
(a)
Diamond
(b)
Sacrificial layer
(C)
Al

((1)

84

 

 

 

 

Figure 4.11. Top and cross sectional views of type A device in different process steps.

85

(a)

Cr

0))

(C)

 

(d)

Figure 4.12. Top and cross sectional views of type B device in different process steps.

86

SiOz for type B devices (Figure 4.12a).

For the type B devices, approximately 2,000 A thick Cr is thermally evaporated on
diamond and patterned. Using wet etching, the thickness of the Cr layer on diamond is
reduced. As the diamond surface is rough (Figure 4.12b), part of its surface is exposed
after etching. Since some diamond emitters consist of a continuous film while Others
consist of patterned dots, we show SEM micrographs of a dot patterned film in Figure
4.12(a) and a continuous film in Figure 4.12 (b). Photoresist, serving as a sacrificial layer,
is spin coated at a speed Of ~1000 rpm and is patterned (Figure 4.11c & Figure 4.12c). Al,
thermally evaporated on top of the photoresist layer, is patterned and part of the sacrificial
photoresist is now exposed to remove it. After removing the sacrificial layer, the
separation between A1 and diamond, computed by taking into account the thicknesses of
sacrificial photoresist, oxide, etched Si and diamond film, is in the range of 9~10 um for
type A devices and 2~3 um for type B devices. Al is also evaporated on the backside Of the
wafer to provide an ohmic contact for type A devices (Figure 4.11d). For type B devices
(Figure 4.12d), Cr is the cathode contact.

The fabrication process for the lateral device is depicted in Figure 4.13. Oxide and Si
in the lateral device area is also etched at the same time, Cross sectional view and SEM
picture of lateral device is shown in Figure 4.13. in the previous section, Ohmic contact to
emitter in the lateral device is provided by Al only on top of diamond film. In testchip HI,
ohmic contact to emitter can be achieved either from Al layer on top of diamond or A1 at
the backside of wafer. This variation enables us to study the effect of metal to diamond

interface on the emission characteristics.

4.8 Summary

In this chapter, the development of diamond field emitter devices using testchip

method is described in chronological order. By integrating all the technologies available

87

   

Diamond

Gap

 

Diamond

Figure 4.13. SEM and cross-sectional
view of lateral device in
each process step.

88

and newly developed, IC compatible diamond field emitter fabrication technology has

been achieved.

CHAPTER 5

FIELD EMISSION RESULTS

5.1 Introduction

In this chapter, measurement data on a variety of samples described in Chapter 4 will
be presented. Up to section 5.6, factual data will be presented without trying to answer the
questions raised. In section 5.7, some questions are raised and additional experiments,
which may help answer the questions, are executed. Combining all the data, preferred
conditions for field emission from diamond film are summarized. An interpretation of

these observations is discussed qualitatively.

5.2 Preliminary Samples

The main aim of the first measurement experiment was to qualify the functionality of

characterization system and characterization method itself.

5.2.1 Setup

Figure 5.] depicts a setup for sample testing for a preliminary sample described in
section 4.3. Setup for continuous film sample is shown in Figure 5.1 (a) and setup for dot
patterned film sample is shown in Figure 5.1 (b), respectively. For continuous film

sample, the voltage drop across the resistor R, measured by a nanovoltmeter, is used to

89

90

 

 

Glass
iTo

 

 

 

 

ZnO:Zn

 

Si L Diamond

llllllllllllllllllllllllIIIIIIIIIIIIIIIII
Al

 

 

 

 

 

 

 

 

 

 

 

 

 

 

(b)

Figure 5.1. Test configuration of the samples; continuous film (a) and
patterned film (b).

91

compute the emission current at switch position 1 using sharp and blunt tungsten (W)
anodes described in section 4.4. To protect the measurement instrument, two resistors
were connected in series. One has at least an order of magnitude higher value than another
one. Voltage was read at the smaller resistor. In case there is are or sudden vacuum
breakdown, which makes all the applied potential across the resistors, at least only one
tenth of applied potential is dropped at the measuring resistor. At position 2, phosphor
anode is used to record emission pattern. This phosphor anode is very useful to record
emission spot and to check quickly whether or not measured current is emission current. It
is expected that we only see visible spots in forward bias, if it were real emission current.
If visible spots were found for the same forward and reverse bias, it would be discharge
current due to vacuum breakdown. For dot patterned film, emission pattern was inspected

only by using phosphor anode.

5.2.2 Results

The field emission current was studied as a function of anode geometry, anode-to-
cathode separation, and pressure. Figure 5.2 shows the emission current for sharp and
blunt anodes shown in Figure 3.5 (b); the corresponding F-N plot is shown in the inset.
The difference in the current for the two anode types is more pronounced in upper anode
voltage range where the emission current shows a F-N behavior. The deviation from F-N
behavior in the lower voltage range may indicate the presence of non-emission currents. It
may also be due to the fact that emission area is larger when blunt anode was used. The
reverse current was an order of magnitude lower than the values shown in Figure 5.2.

Figure 5.3 shows the emission currents for anode to cathode separations of 50 and 100
um for the sharp anode. The F—N curve becomes steeper with increasing separation
indicating a smaller geometric factor. Part of the emission current may be from tiny grain

tips on the polycrystalline diamond surface. The distance dependence of F—N field

CURRENT(uA)

92

d=50 um, p=10'7 torr

 

 

 

 

 

 

 

 

2
10 ' T ' 1
xSHARP ANODE ‘
1 OBLUNTANODE o o o
O
10 5 ° 1
O .
O o " x 1
O 0 x ‘4 . E
10 o o 1" x,
5 ° " E165 '
_1 ~ x0 3 -6 I 0..
10 r 0 “'2 ‘0
a} 1o"fl 4 ~
_2’ ° 2W(W)4 x103
10 A ‘ 4
200 300 400 500 600

ANODE POTENTIAL (VOLTS)

Figure 5.2. I-V curves of diamond field emitter for blunt and sharp anodes;
emitter to anode separation is 50 um and the measurement pres-
sure is 10'7 torr. The inset shows the corresponding F-N plot.

PA)

CURRENT(

93

SHARP ANODE, p=10'7 torr

 

 

 

 

 

 

 

 

1
,, i
d=50um x
. / x
o a? "
1 0 _ 0°00 d=100um 4
00° " .
000° X 10-4 1
1 1 0° A 10-5 1
— o N
1 O 7 0° 8 g 1.... \.... 4
ۤ1dq ._ .
.. _ 4 '_ -
. 1° 0 2 4 _3 6 .
_2 1N(1/V) X10
1 0 ‘ ‘ ‘
200 400 600 800 1 000
ANODE POTENTIAL (VOLTS)

Figure 5.3. I-V curves for different anode to cathode distances for
the sharp anode at 10'7 torr .

94

emission current affecting the geometric factor has been Observed for microtip emitters

[98, 99]. The emission was reported for anode-tO-cathode separation of 60 um for

tungsten-coated silicon pyramid [99] and up to 500 um for diamond field emitter [13].
The pressure dependence of field emission current for hydrogen ambient is shown in

Figure 5.4. The test chamber was evacuated and flushed with H2 many times and the

pressure was kept at 10’7 or 10'2 Torr during the measurement. Field emission behavior is

continued even though absolute value of current density is decreased as the pressure inside

vacuum chamber is raised up to 10'2 Torr range. The field emission at a pressure Of 10'2

Torr in H2 environment may be due to diamond’s chemical immunity to residual gases

which are known to be adsorbed to the surface of other materials and raise their work

functions. This may also be related to the enhancement of field emission found for H2 and
hydroxide (OH) ambient [100]. Kim et al. [33] have also measured constant emission
current for pressure range of 10'6 ~ 10‘4 Torr from diamond wedge fabricated by mold

technique. Emission from Si tip dropped very sharply at pressure higher than 2x10‘4 Torr
[101].

Using a phosphor anode, the field emission from diamond was confirmed by light
emission observed in position 2 (Figure 5.1) of continuous sample and dot patterned
sample. The anode—tO-emitter separation was kept at 100 um using thick quartz sheet. As
shown in Figure 5.5, the bright spots, caused by electrons impinging upon the phosphor-
coated anode, indicate emission only from certain points from the diamond film. In case of
dot patterned film, the light emission pattern does not correspond to that on the film. An
increase in anode voltage caused an increase in the number of emission sites and the spots
became brighter. For negative anode voltages, no light emission was observed. This
proves that measured current is not discharge current. Bright spots can be seen at emission
current as low as 50 nA. Although emission area for dot patterned film was smaller than

that of continuous film, number of emission sites was higher and larger current was

95

SHARP ANODE, d=200 um

 

 

 

 

 

 

 

 

0
1o . 1 . . .
I 10 v 1
. o
"‘7‘ 10_71 1 O
i o
-1 V1043
10 a? 0 .
I- -9 O
C 10 6 8 1o 12 o i
i 1N(1N) x104 .
'1’ . o x .
E -2
I1 0 r O x ..
g x
0 o
X i
o
4 o x o p=10‘7 torr
-3 x p=10'2 torr
1o . .. . .
800 1 000 1 200 1400 1600
ANODE POTENTIAL (VOLTS)

Figure 5.4. I-V curves for two different pressures; the emitter to

anode separation is 200 um and the sharp anode is used.

96

CONTINUOUS
FILM

 

(a)

PATTERN ED
FILM

 

Figure 5.5. The field emission pattern measured (a) for continuous film in posi-
tion 2 of Figure 5.1(a) and (b) for patterned film.

97

detected.
It is not clear why the emission is non-uniform. Discussion about non-uniform

emission from diamond film will be deferred until section 5.7.

5.3 Discrete Samples with Built-in Anode

Figure 5.6 depicts a setup for sample testing for discrete samples with built-in anodes
described in section 3.4.1.6. Setup for vertical device is shown in Figure 5.6 (a) and setup
for lateral device is shown in Figure 5.6 (b), respectively. Figure 5.7 shows I-V data of
vertical devices on oxidized wafer. Since sacrificial photoresist was not spin coated, gap
between anode and emitter is in the range of 100 um as shown in Figure 4.5 (c). Due to
manual diamond patterning, different photoresist thickness, and non-uniform A] width,
there are some variation among data. Different tum-on voltages reflect different anode to
emitter distances and variation on current at the same voltage may be from the variation of
emitter size. Figure 5.8 shows I-V data for lateral devices on oxidized Si wafer. For lateral
devices, tum-on voltage is much lower than vertical devices because anode to emitter
distance is in the range of ~25 um. Also, data shown in Figure 5.8 is close to each other
than data shown in Figure 5.7. It may be due to the fact that anode to emitter distance was
well controlled for lateral device by using precise aligner. Figure 5.9 shows I—V data for

vertical devices on Q2 sample (Figure 4.6). Nucleation densities for two different devices

were ~10ll cm'2 and ~108 cm‘2

, respectively. Diamond film seeded with high nucleation
density exhibits slightly higher current and low tum-on voltage. As shown in Figure 5.10,
for lateral devices on QZ substrate higher emission current is again found in case of high
nucleation density film. These measurement data is promising to future application of

built-in structures. The main purpose Of this study was the successful implementation of

built-in anode and technological development for field emission testchip.

98

 

(b)

Figure 5.6. Measurement setup for (a) vertical device and (b) lateral device.

99

 

 

 

 

 

 

 

 

 

“1 l-V
1AX10 , . . .
_,o F-N 4‘
10 - -
1.2 T _
1o“‘~
1 h -i
1.
E10 ‘2- 4
E - d»
< 0.8 T ‘13 3
T: 10 » 4 +
5
g 0.6 - .
0 1° 65 7 7 s a 875 S 9 5 10
1N x10" +
0.4 T + x 4
at +
O I
0.2 T x
a + O 5 o O
o , *9 r—TTEEE‘gz‘T 9 O . l
1000 1100 1200 1300 1400 1500

Potential (V)

Figure 5.7. I-V and F-N data for four different vertical devices.

Current (Amp)

1.5

100

x 10.5 H,

 

 

I

 

 

 

 

 

I

0 - 72:.- '. 1 1
240 260 280 300 320 340 360

Potential (V)

Figure 5.8. I-V and F-N data for two different lateral devices.

 

380

Current (Amp)

101

 

1.5 l I l I

 

    
 
 
  
   
  

High nucleat' n
densigy x

 

 

 

-‘l4

 

P
01

Low nucleation
density 0

  
 

 

 

 
 

 

 

v

 

'1200 1300 1400 1500
Potential (V)

1000 1100-

Figure 5.9. I-V and EN data for two different vertical devices seeded
with low and high nucleation density.

102

 

 

 

 

 

 

 

 

 

 

5 l-V
x 1
8 o I I I I I I
10‘9 ,
I
6 "' 10 10E. “
2 —1r . .
g g“) :1 High nucleat on
S, i densrty x
2:"; 4 T 10 ‘2 x -
5 i
0 10-13 4
2.6 2 8 3 3.2 3 4 3 6 3 8 4
IN x 10 3 x O
2 - Low nuéI-tiori) ,
" densit
x x x O
’ 0
° 0
- - e ‘3 ' O
0 n A l ._._—- ' x " l I l
240 260 280 300 320 340 360 380

Potential (V)

Figure 5.10. I-V and F-N data for two different lateral devices seeded

with low and high nucleation density.

103

5.4 Testchip I

Figure 5.1 1 depicts a setup for sample testing for testchip 1. Cell notation in the form
of "XYZ" are used to identify a cell in a wafer. The "X" specifies the chip number of the
wafer. The "Y" denotes row number of each chip. The "Z" denotes column number of each
chip. For example, the 224 denotes the chip NO. 2, row No. 2, column NO. 4 which is
connected to the test circuit in Figure 5.11. In Figure 5.12 (a) I-V data is shown for cells
marked circle in Figure 5.11, which are away from the center of the wafer. Although there
is a variation in current level, it can be said that diamond quality, grown by newly built
HFCVD, is similar and uniform over 4 inch wafer. When compared I-V data taken from
cells marked in circle to data taken for cells marked in square, cells close to the center
exhibit a little bit larger current at the same voltage. As marked on the center of the wafer,
two boron powder holder were located at the center of the wafer. Resistivity of cells close
to the center is smaller than cells away from the center due to non-uniform doping.
Reduced resistivity might have played a role in higher current. This will be investigated
further in section 5.7. Figure 5.13 shows data for samples made by using metal pattern I
and metal pattern H. The wafer where diamond film covered with metal on the top of the
diamond line (metal pattern II shown in Figure 4.8d) shows higher current than the wafer
where diamond pattern with metal only on the pad (metal pattern I shown in Figure 4.8c).
For the samples using metal pattern I, electrons have to move through diamond to reach to
emission site from the pad. But in case of sample by using metal pattern 11, electrons are

likely to move through metal just before the emission site. Sample using metal pattern II

shows higher current at lower voltage.

5-5 Testchip II

To include lateral devices, testchip II is designed and fabricated. Figure 5.14 shows

meaSUI-ement setups for vertical devices and lateral devices. In Figure 5.15, I-V curve for

Column 1

Column 4

Row 1

Chip 1
Row 2

Chip 2

NANO-

VOLT
METE Fl

 

 

Location of B—powder
holder

Figure 5.11. Measurement setup for cell 224 of testchip I. Eight marked cells
were measured.

3.5 r I I I T I I I

 

 

I“
01
I

 

 

Current (Amp)

 

I

0.5

 

 

 

 

O ' - -
28 30 32 34 36 38 40 42 44 46
Potential (V)

x 10"4 1",

 

 

0.8 T

.0
a:
I

 

 

 

Current (Amp)

.0
A

0.2 T

 

 

 

 

28 30 32 34 36 318 40 42 44 46
Potential (V)

Figure 5.12 (a) l-V and F-N data of cells marked circle in Figure 5.11. (b)

I-V and EN of a cell from square marked and one cell from

circle marked in Figure 5.11.

Current (Amp)

106

 

 

   
 
 
  
   

 

 

 

 

 

 

-5 _
x 10 1 V
8 I I I I I I T f
_7 F-N
10 E
X
6 __ 10.8 1 _‘
X
210..» Metal
2 1
— pattern II
4 I- 1040 14 —4
1
O
10*” . A i + 4 A
0.022 0.024 0.026 0.028 No.03 0.032 0.034 0.036 O
1
2 _ -
Metal
pattern I
A A l l

 

 

 

o I
28 30 32 34 36 38 40 42 44 46

Potential (V)

Figure 5.13. I-V and F-N data for samples with metal pattern I and
metal pattern 11 in test chip I.

107

vertical devices with different emission area are shown. Since sacrificial photoresist layer
was spin-coated for the testchip II, vacuum gap is believed to be same for all sixteen
devices. It is found that measured currents don’t grow in proportion to the emitting area.
Ernission from diamond seems to be non-uniform. Figure 5.16 shows, I-V for lateral
device with same emission area but different anode to emitter distance. At same field, it

seems that emitter with large distance to anode draws slightly larger current than emitter

with smaller distance.

Figure 5.17 shows lateral device with same anode to emitter distance but different

emitter shape.

5-6 Testchip 111

Using the setups shown in Figure 5.18 (a) for type A and in Fig. 5.18 (b) for type B,

the current-voltage (I—V) measurements are taken by placing the samples inside a vacuum

Chamber with a pressure of 10'6 Torr. The current density I measured as function of field

strength F for the diode structure is shown in Figure 5.19 for type A and B devices. The
current density measured at 0.2MV/cm is approximately 0.5A/cm2 for type A device and

O. l A/cm2 for type B device. A number of factors may have contributed to the difference in
the current densities for devices A and B. First, the diamond film is partly covered with Cr
for type B devices, reducing the effective emission area. Secondly, the emitter contact
resistance might have affected the emission at high emission current values, as the emitter
with smaller contact resistance was found to exhibit higher current density [6]. It may be
noted that for the type A device, the metal semiconductor contact is Ohmic. Lastly, the
different anode-to-emitter spacing for type A and B may affect the field at the emitting
surface. The diamond surface consists Of tiny tips of diamond due to the surface
roughness. It was found [102] that, for the same field, a larger anode—tO-emitter spacing

results in a higher current. It is argued that as the anode is moved away from the emitter,

108

 

 

 

(b)

Figure 5.14. Measurement setup fo testchip II (a) vertical device (b) lateral
device.

Current (Amp)

109

Actual area=250015000210000220000 um2
Area ratio=1:2:4:8
Current ratio=1:l.12: 1.36:3.5

x 10’5 1"V

 

 

2.5

I
l

      

Increasing
area

 

 

 

I

0.5

 

 

 

 

46

Potential (V)

Figure 5.15. I-V and EN data for vertical devices with different emitter
area.

110

 

3 .-
d=lO 11m d=20 um d=30 um d=40 um d=53 lmi

D

2.5 T .

 

Current (Amp)

 

 

00 200 300 - 400 I 500 -. 600 700 800
Potential (V)

F—N

 

 

INN

d —l A
O O O
.1. l. .1.
N d O

:2.

II

x N

o

T:

B

d=SO |.tm

 

 

 

0 0.002 0.004 N 0.006 0.008 0.01
1

Figure 5.16. I-V and EN data for lateral devices with different emitter to
anode distances.

lll

 

 

3.5 T

 

I“
U"
l

 

 

Current (Amp)
N

.3
U"
r

 

0.5 T

 

 

0 “‘

 

 

 

28 30 32 34 36 38 40 42 44
Potential (V)

Figure 5.17. I-V and F-N data for lateral devices with different emitter
shape with same distance.

46

112

the area of relatively uniform electric field under the probe expands, and more emitters can
contribute to the measured current. Our simulation studies and measurements, which will
be discussed later in detail, support this argument.

The testchip contains a number of type A and B device structures a consisting of
continuous or patterned diamond film as the emitter. As shown in Figure 5.20, for both
I: ypes of emitters, the current density for an array of dots is higher than that of a continuous

{i ]m for equal areas of the emitters. As the continuous and patterned films are prepared

under similar deposition conditions, their doping level, surface morphology, sp3/sp2 ratio
and surface termination are expected to be similar. Using a phosphor-coated glass plate as
an anode, it was found that the emission from doped diamond films is non-uniform [15].
In another study [103], the emission from isolated diamond particles showed a higher
current density than that from a continuous film. Although our results seem to suggest a
larger number of emission sites for the patterned emitter, it is not clear whether the
enhanced emission takes place at edges.
In Figure 5.21 measurement setup for lateral device is shown. Two different emitter
contacts were possible. One from the top and another from the backside of the wafer. In
F i gu re 5.22, I-V data for two different emitter contacts were depicted. Higher current was
detected when emitter contact was established from the backside of the wafer. This trend
is similar to the vertical device where type A device which has emitter contact at the

baCkside of the wafer showed higher current density than type B device.

O

(a) Type A

 

 

 

 

 

 

d: 9~10um

 

@‘L‘LWV‘ (b) Type B

 

 

 

 

 

d: 2~3 um

 

 

 

 

 

 

 

 

Figure 5.18. Measurement setup for (a) type A and (b) type B device.

114

 

 

 

 

0.6 . . .
0.5- X -
5.; 0.4- x TypeA O 5
O x
2 o
V __ J
g 0.3 x O
(I) O
8 x o
'0 0.2- O i
’5 x o
2
5 X 0
0 0.1” X o 0 Type B J
x x o O
O—ex-akxcx 3 x0 9 O O i i
0 1 2 3 4
Field (x105 V/cm)

Figure 5.19. I-V curve and F -N plot of diamond field emitter for type A and B
device.

Current density (A/cmz)

115

 

 

 

 

 

 

0.7 r . .

a?

2 g i O

0.6 E

g 1.5) O .-

1 3

0-5 T5 1 ::: 0 .-

ac) III

U 05 1 O
0.4T *5 '

e . I O x

L G A---nnflll' A .
03- a 0 0.5 1 1.5 2.0 2.5 O x
' X

Field (x105 V/cm) o x
O X
0.2" o x
x I
O X
0.1” O x
O X X
0 x
O X
Q X

O—e—e—e—lfi—Q 1 L 1

 

 

O
.3
N

0.)

Field (x105 V/cm)

Figure 5.20. I-V curves of type A (inset) and type B devices for

continuous and array type emitters with the same total
area.

116

v
(a)

A] anode —__L_
:7 4:1. A]

   
  
 

 

 

Diamond

 

(b) _

Figure 5.21. Measurement setup for lateral device from Alon top of diamond
film (a) and Al on Si backside.

O

O.
«DEG; ~C0t30

AU.

0.2

117

 

 

 

 

 

 

 

 

 

1 2 x 10“ "V
.8 F—N
10 - -
1 I!
10‘9 {
10"°{ Bottom contact
0.8 T E x -
g 10‘11 r
s i
E 0 6 T 10 ‘2. i
a) " x
’5 , . .
0 1° 3 3.5 4 4.5 s
0.4 _ 1N x 10“1 x _
x O
0.2 " a (D
x x o ’ op contact
X . ’ o O O
|-. ; - .— T": .:__.-_- " ' " l l
200 220 240 260 280 300
Potential (V)

Figure 5.22. I-V and F-N data for lateral device two different emitter
contact.

To

533mb

118

5.7 Discussions

The results in the preceding sections raise the following questions:

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

Why electron emission is possible at 10'2 Torr? (section 5.2)

Why diamond seeded with high nucleation density shows higher current than

diamond film seed with low nucleation density? (section 5.3)
Why low resistivity film exhibits higher emission current? (section 5.4)

Do different contacts to diamond affect emission current from diamond? (section

5.4 and 5.6)
Why emission from diamond is non uniform? (section 5.4 and 5.6)

Why emission current seems to be higher even at same field when anode to

emitter distance is larger? (section 5.5 and 5.6)?

Why dot patterned film shows more current than continuous film? (section 5.2

and 5.6)
How does diamond film quality relate to the emission characteristics?

What is the best emitter structure for achieving high current densities needed for

FED?

To address some of the questions two additional experiments were conducted as

described in the following sections.

ll

b)

run;

of]

of aj

hem

mod

119

5.7 .1 Effect of diamond deposition condition

To study the role of film deposition parameters on the field emission current density,
which may address the questions (2), (3), (8) and (9), three different sets of diamond films

were prepared on p-type Si. Using different CH4 concentrations in hydrogen, the samples

with different sp3/sp2 ratios were produced in the first set. The sp3/sp2 ratio is calculated

by subtracting the base line value from Raman spectrum and comparing count numbers at

1332 cm'1 and 1580 cm". In the second set, the boron doping concentration of films was
varied using solid or powder sources for in-situ doping of diamond [104]. The resistivity
was measured by the four-point probe method. Depending upon the relative position from
either solid source or boron powder crucible, the resistivity varies. For the third set,
diamond films with grain sizes in the range of ~03 to ~15 um and film thicknesses in the

range of ~0.5 pm to ~25 um were prepared using initial nucleation densities in the ranges

of 1011 and 108 cm‘2 [105], respectively. Table 5.1 summarizes the deposition conditions
of all the films used in this study.

For the field emission measurement, a 50 pm thick quartz plate was used as a spacer
between the anode and the diamond emitter. A polished brass column was used as an
anode [15].

Figure 5.23 shows I-V curves, Raman spectra and SEM pictures for samples with

CHMHZ ratios of 0.5, 1 and 2%. With increasing CH4 concentration, a deterioration of sp3/

sp2 ratio and thus the diamond quality is indicated by the decreasing height of the diamond

peak in the Raman spectra. The diamond with low sp3/sp2 ratio exhibits low emission
fields and high current densities. It is particularly interesting to note from the SEM

micrographs that with an increasing CH4 concentration the number of small grains on the

diamond surface increases, resulting in a higher density of grain boundaries. A similar

trend is also observed in films grown in microwave CVD [106].

120

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

N3... 8.. 2.. .033 m N + o mNz § 9,
aNd- NE: T .338 m N + o 3... § we
3.33
and- 85.. 2 T non N + o 3.. 9E 2
Bum?
93.. 3.. N? mon w + o 2.. § 2..
an? NN.o.. $7 o + v we. N; 2
8.? to: o + w 3.. § 3.
NNdz 2.? 2.. Sega m N + o m. T sN 2
:3- MN? NN- 332. m N + e 3.. NS Na
93.. one- mm- .838 m N + e a? $3 z
N-Eo<E_ N-Eo<E _o.o come—o
AEQGV 02:8 +nwoMMn== 05E NEvmu o ammm
a 3.802% £33.63. wimofl A25”: wafimm _ m
20¢ :EmflEm 5390

 

deaafiuoauauaau can 555333 0.95% we tan—55m .~.m 29¢.

 

A<1v «£05530

4

CED fanatic.» 33:6

3

at.

12]

 

Current (uA)

 

Boo

o

2°/o[ ,0
1% a .
.° 0.5%

x
O x
00°00 x"‘

1000 1500 2000 2500

 

 

 

 

 

Potential (Volt)
C
D 7000
3' .
E
-"-'-' 60” 0.5%
'9
s 5000>
3‘ 1%
.6 ml
dc, 2%
E 30qu 1200 1500 1400 1500 1600

Raman Shift (cm'1)

1700

 

2%

Figure 5.23. I-V curves, Raman spectra, and SEM pictures of samples for sam-

ples with different CH4 concentrations.

med
then
Th5

cond

Y‘ 2.

nudi
d0p€l
of gr
dunn
ofno
endm
conUr
presur

man“)

122

A higher CH4 concentration results in lower resistivity as measured by 4-point probe
method. As the density of grain boundaries is enhanced through high CH4 concentration,

their contribution to the conduction may be responsible for the observed low resistivity.
This may suggest that the presence of graphitic phase may be related the enhanced

conduction.

As the high doping levels affect sp3/sp2 ratio, grain boundaries and grain sizes [96], we

studied the effect of doping on the field emission current. As seen in Figure 5.24, the

doped film shows enhanced emitter current. Again, the low sp3/sp2 ratio and high density
of grain boundaries seems to result in a low field and high current density. Although
during the preparation of the undoped sample no dopant source was used, the early stage
of non-continuous film shows some conductivity with resistivity larger than 1 Kflcm as
evident from the 4-point probe measurement. When it is grown longer and becomes
continuous, this conducting behavior disappears probably indicating that B source,
presumably from p-type Si wafer, is now completely covered by diamond film and
diamond film is becoming non-conducting. During the measurements on some undoped
films, arc was observed before the start of the emission, the origin of which is not well
understood.

To produce samples with large differences in doping levels, boron powder or 8203

wafer was used as a dopant source. As shown in Figure 5.25, although CH4 concentration

is the same, sp3/sp2 ratio tends to decrease with increasing B concentration. Highly doped

film shows low field and high current density emission. In earlier experiments [71], it was

found that a diamond film deposited directly on Si results in a lower sp3/sp2 ratio than the
one deposited on top of an undoped diamond film. As shown in Figure 5.26, the emission
current from a doped film deposited directly on Si was higher than that of a film deposited

on top of undoped diamond. In all the above results, the relationship between the low

-- Ax‘Iv 0:105:40

5‘

«223 \Cfihtnh

2 11 (I

(v \nuuWCQHC.

Figl

n:

123

 

 

 

 

20’

A

$15

I: Doped .

5 ,0) Undoped

5 0°

'5 °°

o 5* i
1%00 1500 20.03 2500 3000 3500

Potential (Volt)

 

 

 

 

A AMA

: WVV * ,7

C

D 3500

E 3000 Undoped
b

:5 2500»

h

S. 2000 Doped
5'

'17) 1500

C

o 1AM A L A L 1
E “1100 1200 1300 1400 1500 1600 1700

Raman Shift (cm'1)

 

Figure 5.24. I-V curves, Raman spectra, and SEM pictures of samples for sam-

ples with doped and undoped film.

A<1v «£05530

. w a m a w m m“..-
«2:: >5m5535<w 53:05:.

us

figure

Current (11A)

Intensity (Arbitrary Unit)

at

124

 

 

 

‘J
0 X
20 o X 1
0 x
o X
151 B powder . 1
0° "‘
10* 0° 8203 wafer *
00 X
0° ‘
5» .
n xxx‘x )(
5’00 1000 1500 2000 2500 3000

 

. Potential (Volt)

 

4000'

3500

3000

2500

2000

1500

l
8203 wafer «

. .

 

 

 

1%

00 1200 1300 1400 1500 1600 1700
Raman Shift (cm'1)

 

 

B powder

Figure 5.25. I-V curves, Raman spectra, and SEM pictures of samples for sam-
ples doped with two different doping methods.

A<1v «5505530

«:55: >505tn~5<v >3020~C~

Current (11A)

Intensity (Arbitrary Unit)

 

 

300 1 000 1 500 2000 2500

 

x
1111

Potential (Volt)

 

  
   
  

Undoped+
Doped

Dop
onl

 

 

 

“"i‘ioo 1200 1300 1400 1500 1600 1700

Raman Shift (cm'1)

sample‘ifi \ __

 

Undoped+
Doped

 

Figure 5.26. I-V curves, Raman spectra, and SEM pictures of samples for sam-
ples with doped only layer and doped layer over undoped layer.

cm

1.01

no

filrr
(Tal

repr

126

emission field and (1) low sp3/sp2 ratio and (ii) high grain density is consistently obvious.
Another interesting feature of the results is the presence of both small and large grains.
Lower quality diamond films were found to results in enhanced emission in earlier studies
[102,103].

It would be interesting now to compare the emission behavior of small and large grain

films. As shown in Figure 5.27, the small grain film results in the lowest emission field

(Table 1). In an earlier study, current densities in the range of 10 Acm'z, the highest
reported for diamond, were reported for nanocrystalline diamond [107].

The data in the present study provides an experimental evidence for the enhancement

of current density and reduction in the emission field for diamond films with low sp3/sp2
ratios, high doping densities, and large densities of grain boundaries (i.e. high densities of
small grains). Field enhancement at the grain tips, surface termination, defects, number of
emission sites and grain orientation may have played a role in the field emission. These
considerations lead to a number of questions: Does the fine grain film result in a higher
density of emission sites than a film with a combination of small and large grains? Are
these emission sites related to field enhancement at the tips? Is the presence of defects in
diamond necessary for field emission from p-type or undoped diamond? Is the field
emission from single crystal diamond with a smooth surface [108] related to defect
densities?

Electron emission from flat surface of boron-doped natural diamond has been reported
[108]. In this case, defects, orientation and surface passivation of the diamond may be
among the critical factors. However, in the case of polycrystalline diamond, additional
factors such as grain boundaries, grain size differences, high densities of surface defects
and the field enhancement at grain tips, should be taken into account. The lower threshold

fields and higher current densities for polycrystalline than those of single crystal diamond

have been attributed to lower sp3/sp2 ratios and the related defects found in polycrystalline

a<1v «£05530

5 mm :4. m ”w... M ii

2

35:3 >505535<C 5.0.2055

127

 

 

 

 

 

Zum
30 sample #8
25' ° ,-. ~ :Mfik
2 .. ,. . .
a 15 _
c .. .
g 10 °
11‘
a . . ‘
c) 500.... 1000 1500 2000
Potential (Volt)
5000’
4500
4000

   

 

 

3500t

.1 )1 ~0E3um
3000r .
2“‘lm‘ioo 1200 1:500 1430 1500 Taco 17100
Raman Shift (cm'1)

Intensity (Arbitrary Unit)

 

Figure 5.27. I-V curves, Raman spectra, and SEM pictures of samples for sam-
ples with different grain size.

0m
dmz

fins

111 a
Sepzu
Type
mnm
Mme
entire,
fight.
p-T‘VlOU
53118810
1713.)? in C1

:1

“Wfimc

128

films [102,103]. The role of field enhancement at the grain tips may vary for small and
large grain films, and for films containing both the small and large grains. The density of
emission sites in a particular film may depend on field enhancement, grain size, surface
termination, and defect densities.

When both the small and large grains are present on the diamond surface, the height of
large grains can lead to a larger field at their tips and, therefore, the emission may be
initiated at these sites first when a field is applied. The small grains present between the
large grains will have a lower field at their tips and the field emission will primarily take
place from the highest tips. However, when the number of small grains between the large
grains increases, the density of emission sites in the small grains area increases. Thus, for
a film consisting of small grains only, a larger density of the emission sites may be
observed due to similar fields at the emitting tips. This may be related to the measurement
data found for discrete sample in section 5.3 where film seeded with high nucleation
density exhibited larger current than low nucleation density seeded film.

As shown in Figure 5.28, films with small and large grains, categorized into types I, II,
III and IV, were simulated to study the field enhancement using anode to emitter
separations of l or 10 um. Type I represents small grain film. Grain size is set to 0.5 pm.
Type 11 represents large grain film. Grain size is set to 2 um. Type III denotes 50-50
mixture of small grain and large grain film. Type IV represents mostly small grain film
with occasional large grain film. Figure 5.29 shows electric fields at the tips for the four
emitter types. Data shows that field enhancement is larger when separation between the
neighboring tips and anode to emitter distance is larger. This result is consistent with
previous simulation studies of Si microtip emitters [109,110]. This may suggest that the
emission is enhanced when anode to emitter distance is larger at a given field strength and
may indicate that emission is predominantly coming from grain tips. To verify this effect

experimentally, the same film was measured for anode to emitter separations of 1.6 pm

:::!

i-JI
; '-

[ I l
the
hyd
mor
Con

to lo

5070;

129

and 50 pm. As shown in Figure 5.30, the larger separation results in higher current and
lower threshold field. This may explain question (6) raised in the beginning of section 5.7.
This has also been noticed at the characterization of lateral device of testchip II (section
5.5, Figure 5.16).

The adsorption of H on the surface of diamond is known to lower its electron affinity
[111]. As all our samples were treated in hydrogen plasma in the HFCVD system just after
the diamond film deposition, it is believed that the film surface is passivated with
hydrogen for all samples. It is conceivable that the surface termination with hydrogen is
more pronounced for small grain films due to a larger density of grain boundaries.
Consequently, the enhanced field emission behavior of small grain films may also be due

to low electron affinity.

5.7 .2 Effect of Implantation

Defects can be introduced in diamond through the variation of deposition conditions
[112, 102] or by ion implantation [103, 35]. Zhu et al. [103] studied field emission from
ion-implanted diamond. In their experiment, boron, carbon and sodium ions were
implanted. But only one energy level was selected for each ion source. Although we only
select carbon ion for implantation, we selected two different doses and two different
energy levels in an effort to study the field emission dependence on the depth of damaged

layer as well as the density of damage sites.

Four 1 x 2 cm2 samples consisting of B-doped polycrystalline diamond films grown
on p-type Si by HFCVD system were prepared under similar deposition conditions. Half
of each sample was covered with metal to block implantation so that direct comparison of

emission current from both implanted and unimplanted areas can be made. The
implantation doses and energies were 5x10l6 cm'2 and 5x1015 cm'z, and 20keV and 50

keV, respectively. Samples were implanted with either C2+ or C4+ depending upon their

130

— 09$. :5 Ecru—3.5% m>m2< 58 85:50:53 «53:5 Born and «Sufi

 

Sim
BEEm

 

$—

0)

DO

I:

O

|]IIIIIIII =
“ U]

‘5'?
a

Eocd

 

 

131

l—
0
00
C
O
I-a
H
U)

DIIIIIIII

25
«‘5’?
m3

= 2:5. 3v 0:55.80 .wN.m 0.535

 

 

.oEEm

 

ono=<

 

132

E1N

 

L1
0
no
1:
o
b
V)

UIIIIIIII

E 25 a $555.59 dam 2:»:

 

SEEM

 

ovoc<

 

133

Eim

 

._

0)
DD
8
[JIIIIIIII a

58.33
Beam

2 25 A3 3.2.5.50 .wNm 20E

 

BEEm

 

one—5.

 

2:035: 0.0.5-35

134

 

 

 

 

 

 

 

 

20 . . a o
O
O 10 um
[5, 15 - 1 um
S
E x "‘
'2 'u I
.9 .. 3;. 1 . .
“a la. 10 um
'F 1 O ' o i: '
.. ”5 Wm
Tu
E
o . .
Z I 11 III IV
Emitter Type
5 1 1 1 1
I II III IV
Emitter Type

Figure 5.29. Simulated electric fields at the tip for different types for
anode to emitter distances of 1 and 10 um.

’<-\ *ccxx- uIL

135

Current vs. Field

 

 

 

 

35
30 - *
25 -
3% 20- ° ..
E, 50 um
55 15 ~ 0 ..
10 b O a" 1.6 um
5 b O 0 X x
0L 4 115 PR 3: " 3" I
8 9 10
Field (V/um)

Figure 5.30. Measured emission current as a function of field for anode to
emitter distances of 1.6 and 50 um.

thc‘
{ha

€111.

“as
711.11 l
mfrg

Since

mifisur

G31" blm‘

136

doses. Implantation was done at National Superconducting Cyclotron Laboratory by Dr. T.
Grimm [113]. Detailed sample preparation conditions are summarized in Table 5.2.

Table 5.2. Implantation enegry and dose for four samples.

 

 

 

Sample ID Energy (keV) Dose (cm'z)
111-2 20 5x1015
HI-3 50 5:11015
HI-S 20 5:110l6
III-9 50 5:11016

 

SEM and Raman spectroscopy were done both on implanted and unimplanted areas of
the samples before I-V characterization. As shown in Figure 5.31, SEM pictures reveal
that implanted areas show darker image than unimplanted areas indicating secondary

emission is less from implanted areas. Raman spectra show that implanted areas have

shorter peak at 1332 cm'1 and wider full width half maximum indicating that quality of
diamond is deteriorated due to implantation. Electrical characterization was performed at
two different locations, Michigan State and Varian Associates. First, I-V characterization
of the samples are performed at Michigan State. Brass anode described in section 3.4.1.2
was used for measurement set-up. Figure 5.32 shows I-V curves. I-V data clearly show
that threshold field is lower and current density is higher for higher doses and lower
energies. The data support the fact that shallow defects can enhance the field emission
since emission increases as the implant dose is increased and made shallower. It was also
observed that, for all samples, larger anode to emitter separations led to higher currents.
This may be due to the field enhancement found for larger anode to emitter separation
[112, 35]. Parts of the samples were cut and shipped to Varian Associates for the

measurements. Dr. S. Bandy led the measurement work in collaboration with Mr. L.

Garbini. The area of the sample was about 25 mm2 and the rectangular samples were

~«~£D >505=QL <» 33.0203: PM.»

137

Cross
section

 

Top

view

15KU X2,068

 

 

 

 

 

 

 

 

 

E‘ 3500 f E 5500 -
I: C

D D 5000

E 3000 E

E E

9:: .1;- 4500

‘3 2500 e

S, $ 4000»

> >1

’5 2000 ’E' 35001

‘2 15 ‘2 30

— (1800 1200 1400 1600 1800 — c11300 1200 1400 1600 1800

- -1
Raman Shift (cm-1) Raman Shlfl (cm )

Figure 5.31. SEM image and Raman spectra of implanted and unimplanted area.

\~.::
>
Am

Fl

1(llA)

138

 

 

 

 

102 + 5
.,....’
r 000°°°° ‘
00000")0000 I
300000000 1““.-
0 000000 1“.
r 0000 .111". . m. C. 1
IO 0 000° . In‘.u:::.000:::::::+:’m 4“:§+:
O O x 1"...'. .. ++++++’*T§+
i 0 ' :0::+++o**“** }
l ‘1‘ ++
2 x 4’ +::+ o HI‘S
10 . x HI-9 (
,. * HI-2
r .. + HI-3 1
10-4 1
0 1000 2000 3000
v (V)

Figure 5.32. I-V curves for four samples measured at Michigan State.

ln(1/V2)

ln(1N2)

-28

 

1.16 - 1.2

1000/V

1.24

 

lOOON
(b)
Figure 5.33. F-N curves for sample (a) III-5 and (b) III-9 measured at Varian.

31‘

dr:
sio

531'.

pic

‘11 1|

and

139

attached to a stainless steel conflat blank using indium metal. The anode was a
molybdenum screen of 76.2 um diameter wire by 1016 um mesh attached to a micrometer
drive for positioning above the sample surface. The I-V measurements were made with a
curve tracer. An Inficon QX-2000 residual gas analyzer (RGA) was used for gas analysis
and the system was pumped with an ion pump plus sublimator after being rough pumped
with a turbo backed with a turbo plus diaphragm backing pump.

All samples had the same characteristics when first operated in normal mode, i.e.
drawing current to the anode positioned about 50.8 urn above the emitter surface. Emis-
sion abruptly started in the 800-1000V region after arcing was first experienced at these
same voltages. While arcing, the current flow had the appearance of a gas sustained dis-
charge and no data for F-N relationship could be taken because of current instability. Sam-
ples HI-S and HI-9 were operated drawing current from “anode electrode” to sample with
a continuous power supply at up to a few mA current after testing in the standard mode.
After this treatment stable current could be drawn from the sample up to about a few 11A
and had the characteristics of field emission per F-N curve. Figure 5.33 shows F—N plots

for HI-S and HI-9. Emission currents of up to 100 “A, which is equivalent to current den-

sity of ~O.lAcm'2, were measured on HI-9 before onset of gas discharge destroyed the
emitting site while in the process of trying to achieve higher currents levels. Severe local
disruption of the diamond surface on both samples was seen by SEM after conclusion of
the testing. This may indicate that emission from the surface was non-uniform and small
number of sites contribute to majority of current.

After collecting data on HI-S, the Inficon QX-2000 was attached to the station for
gas analysis. Analysis of sample HI-9 initially indicated nitrogen, CO, water, hydrogen,

methane, argon and carbon dioxide as the components when drawing current to the anode.

140

4x 10'9

 

 

3x109
5
l" -9
,; 2x10
‘5‘:
0
‘3‘ 1x109
0 2 4 6 8
Time(min)
4x10'9'
3.110-9
S
[—
2‘ 2x109 CO
:1
0.. 1x10'9
C&N
0 2 4 6 8

Time (min)

Figure 5.34. Residual gases for HI-2 after 450 °C 16 hr bake out:
two graphs are used for clarity.

e

fro

Sta;

Whe

disc}

:«n
00

F11
results
Sample.
implame
20W :11

Based

141

Sample HI-2 was baked at 450 0C, with the anode in contact with the sample to

prevent movement above the melting point of the indium solder. The RGA charts shown in

Figure 5.34 were taken after a 450 °C bake out with a three day stand. The major gas liber-
ated from the sample was CO. All pressure increases on these charts were related to onset
of current flow. Magnitude of pressure was related to voltage and anode to sample spacing
in all cases. Higher voltage and greater spacing produced higher gas pressure. Sample HI-
3 was given a hydrogen plasma treatment using the curve tracer as a power supply. No
improvement in the emission instability was seen after this treatment. CO was again found
as the major component curing emission from the diamond surface. Measurement at
Varian seems to support vacuum arc discharges preceding high electron field emission
from carbon films suggested by Groning et al. [114]. In our measurement at Michigan
State, it has also been noticed that auto trip-off of power supply took place occasionally
when load impedance changes abruptly. This may be related to the effect of vacuum arc

discharges preceding electron field emission from carbon films.

5.8 Summary

Films with different growth conditions were prepared and careful characterization
resulted in valuable data in a systematic way. Characterization was performed for the

samples, with CH4 concentration varying from 0.5% to 2%, doped with B powder and
B203 wafer, seeded with nucleation density in the range of ~108 cm'2 to 1011 cm'z, and

implanted with carbon at the dose of 5x1016 cm'2 and 5x1015 cm'2 and dose energy of
20keV and 50 keV.

Based on a large amount of field emission data collected from a variety of samples, it

en
er
611

me

QUE:

not a
due 1
[710g]-
imp]
edge e
r31.10 0
We A

3‘15 Edi

142

was found that emission from diamond is enhanced when (i) sp3/sp2 is low, (ii) peak at

1332 cm'1 is wider, (iii) grain in the film is small, (iv) film is highly doped, (v) film is
damaged by ion implantation, especially density of damage is high and location of
damage is shallow, (vi) grain boundary is present, and (vii) roughness film is small. These
seems to suggest field emission from polycrystalline diamond is affected by (i) NEA, (ii)
defects, and (iii) field enhancement effect at grain tips. These results seem to support a
model proposed by Huang et al. [53] and Zhu et al. [102], which explains that electrons
are transportation of electrons to the surface state through the defect states and subsequent
emission from surface state. Other models by Givargizov et al. [37], Choi et al. [40], Geis
et al. [42], and Lerner et al. [54], which are also mentioned in Chapter 2, may be valid to
explain emission from very thin diamond film, either n-type or undoped, coated on Si or
metal but may not be suitable to explain emission from thick diamond film in our case. In
addition to defects, field enhancement factor in grain tips are found to be one of important
factors for emission from polycrystalline diamond film.

Some unanswered questions, which were raised in the beginning of section 5.7, are
questions number (1), (4), (5), and (7). Diamond’s ability to emit electrons in relatively
high pressure may be due to its chemical inertness. Although it was noticed in Figure 5.13
that different contact methods to diamond do change emission current, question (4) was
not addressed in this study. Non-uniform emission from diamond film (question 5) may be
due to non-uniform film morphology. New characterization system being built with a
programmable stage of resolution 1 pm may help answer this question by mapping whole
sample with in-situ Raman or other film characterization system. For question number (7),
edge effect may have played a role. For example, as shown in Figure 5.20, although area
ratio of continuous film vs. dot patterned film for type A and B is 1:1, edge ratio is 2:1 for
type A and 3:1 for type A. Current ratio seems to agree with this ratio but it is not clear

this edge effect is sole reason for enhanced emission.

6.1

an
ch.
Fa

if]

CHAPTER 6

APPLICATIONS

6.1 Introduction

Although application area for field emitters are numerous, driving force for field
emission research has been a realization of FED. In this chapter, fabrication and
characterization efforts of two types of diamond FEDS in triode mode are mentioned.
Fabrication technology of pressure sensor is also presented. Realization of triode FED was

reported for the first time [1 15].

6.2 Diamond Field Emitter Display

The growing need for developing low cost and high performance flat panel displays
has spurred a strong research effort in FED technology. Prototype 6” diagonal color FED
[116] was demonstrated using Spindt type [2] microtip metal field emitters. For this type
of FED, it is necessary to fabricate micron sized micro-tips and grids over large area
panels, which results in higher cost. In addition, even though the lifetime of metal micro-
tips has been demonstrated, use of sulfide based color phosphors is a potential problem
due to contamination of metal tip from loose phosphor powder particles.

A prototype black and white 1” diagonal diamond FED based on diode configuration

has been demonstrated [12] using amorphous diamond films. The major drawback of

143

"2

di

16:

016111):
15 x 15

{Idols

divémOnd

“We;

144

diode structure is relatively high switching voltage. To address this problem, triode type

diamond FED technology has been developed at Michigan State.

6.2.1 Diamond Field Emitter Display with Bridge Shaped Grid

In this section, first diamond field emission triode display cells using p—type
polycrystalline diamond field emitter [15] is described. Using a 4-mask lithographic
process, display cells are fabricated on oxidized Si wafer. Initial measurements of these
display cells in diode configuration show field emission at a gate field of 0.1-0.2 MV/cm
resulting in gate switching voltage as low as 50 volts. When tested with phosphor anode in

triode configuration, successful operation of display cells was confirmed.

6.2.1.1 Fabrication

As briefly described in Chapter 4, triode display cells are fabricated in test chip III
along with other structures. An overview of the test chip containing a six by six array
display has already been shown in Figure 4.7. Fabrication process of a diamond field
emitter triode display cell is almost identical to that of type B device described in section
4.7. The only difference is that grid structure has many holes in it to pass emitted electrons
through them. Although only unique points in FED will be discussed here, whole process
is depicted in Figure 6.1.

The emitter (p-type diamond) area shown in Figure 6.1(a) forms one pixel and consists
of either a continuous film or an array of diamond dots. The array size varies from 2 x 2 to
15 x 15. The SEM micrographs shown in Figure 6.1 are for pixels with different number
of dots. For example, pixels consisting of array of 8 x 8 and 5 x 5 dots are shown in Figure
6.1(a) and 6.1(b) while the one for Figure 6.1(c) shows a pixel consisting of a continuous

diamond film.

The patterned (Figure 6.1e) Al layer has a 15 by 15 array of 4x4 um2 holes to permit

145

 

 

Diamond
1 1 SiO
1 Si 1 2
(a)

 

 

 

 

 

._.__——- ,
SONm 642103

 

 

 

Figure 6.1. FED cell fabrication process; cross sectional views and SEM
micrographs.

146

Photoresist

(d)

 

 

 

 

 

O ”Jay-II. E‘IGEx‘t

' .‘IIV

 

Figure 6.2

Figure 6.]. continued.

ale:

PTO

c011
diar
Cr 1
amt
diam
gate

thicl

6.2..

147

electron supply to the anode. The sacrificial photoresist layer is removed (Figure 6.1f) to
produce gap between gate and emitter.

After removing the sacrificial layer, the separation between A1 and diamond,
computed by taking into account the thicknesses of the sacrificial photoresist layer and the

diamond film, is in the range of 2~3 um (Figure 6.1f). A1 is also evaporated on the part of
Cr pattern to serve as a cathode contact. Completed samples are annealed at 400 °C in N 2

ambient for 30 min using a rapid thermal processor. SEM picture of a completed cell of
diamond field emitter is shown in Figure 6.1(t) and a magnified view of the gap between
gate and cathode is shown in Figure 6.2. Note that even though sacrificial layer was quite

thick, step coverage by Al is still acceptable.

6.2.1.2 Testing

- I-V Data

Fabricated samples are first characterized by current versus voltage (I-V)

measurement in a diode configuration to confirm the field emission from diamond FED

cells. Figure 6.3 shows the measurement setup (a) and the I-V data (b) measured at 10'6

Torr. The emission starts at approximately 20 V leading to an emission field in the range

of 0.07 - 0.1 Mch'l. As evident from the inset of Figure 6.3 (b), the IN2 versus 1N plot

is a straight line showing a typical F-N field emission behavior. The current density
measured at 0.2 Mch'1 is approximately 0.1A/cm2 as compared to reported values in the

range of O.1~10 Acm'z. As compared to the switching voltage of several hundred volts
used in the prototype diamond diode FED [12], a gate voltage in the range of 50 - 100
volts was enough to control the field emission in our triode FED structure described in the

next section.

148

Al gate

Emitter

15KU XIB’BBB

 

Figure 6.2. A magnified view of a display cell showing vacuum
gap between gate and emitter.

149

—b <

 

 

 

 

d= ~3 um

 

 

 

 

 

(a)

 

 

 

 

 

 

 

 

25 -2
1o -
14.00
20rA10-3_ 00 o 1 o
N o O
3 15:10—4- o
.0-0 N
E) g 5 (l O
5 10 1° 0 0.05 0.1 0.15 0.2
4% 1N(1N) °
(5 o
51 o 1
O O
0 A .. o 9 ° . .
0 20 40 60 80

Gate Potential (Volt)
(b)

Figure 6.3. (a) Experimental setup for I-V measurement and (b) I-V
curve and F-N plot of a display cell when tested in a diode
configuration.

as]
am

ins

ex;

pat

[ha

6.4

firs

the

1 50
- Triode Cell Testing

Phosphor anode described in Chapter 3 was used as an anode to test the triode FED
cells. A quartz spacer with a thickness of 0.1 or 1 mm was used between the phosphor
anode and the diamond emitter structure. Figure 6.4 shows such an experimental setup to
inspect light emission from one pixel. The grid and anode potentials are in the ranges of
100 and 500 volts, respectively. The grid or anode currents were not monitored in this

experiment which focussed only on light emission pattern depicted in Figure 6.4. The

pattern was recorded in a vacuum chamber with a pressure of 10'6 Torr.

Figure 6.5 shows experimental setup for light emission from a triode FED with an
array of l x 4 pixels and its emission pattern recorded by a camera. It may be pointed out
that the actual number of holes in the Al grid for every pixel is 225. However, in Figures
6.4 and 6.5, only 4 holes were shown for clarity.

Although the triode diamond FED cells have been demonstrated successfully for the
first time, some problems were encountered during the fabrication process. As shown in
the SEM micrography of Figure 6.1(b), (d), and (f), there are a number of stray diamond
particles in areas where they should not appear. They result from the diamond particles
which could not be washed away completely during the development process of
photoresist mixed with diamond particles. They can affect the yield if they form a
continuous film and if their size becomes too large. Figure 6.6 shows a failed cell due to
clustered diamond particles. Due to these problems, only a fraction of the FED devices
were found to function properly. For later samples shown in Figure 6.1(a) and (c), double
layer patterning technique described in Chapter 4 was applied. It shows reduced number

of stray particles.

6.2.2 Diamond Field Emitter Display with Self-Aligned Gate

Although the first diamond FED in triode mode is demonstrated in the previous

151

 

Emission pattern on phosphor

/ ITO on glass

ZnO:Zn phosphor

   
 

 

 

 

 

 

 

Figure 6.4. Setup for emission image from a triode display cell
and its corresponding image.

152

Emission pattern
on phosphor

 

ITO on glass

ZnO:Zn phosphor

 

 

Figure 6.5. Setup for measuring emission image from four triode
display cell array and its corresponding image

153

 

7; " is}: A)!

Figure 6.6. A failed display cell due to a cluster of diamond particles.

sec
dct

mo

nnr

witl
betv
aver
is at

S}‘Sit‘

on-g
Fflni
l0éis

Oxide

154

section, it had low yield and high gate current. It also requires um range lithography to
define small gate halls and accurate alignment. In this section, diamond FED in triode

mode with more relaxed and self-aligned gate structure is developed and demonstrated.

6.2.2.1 Fabrication

Detailed fabrication process is shown in Figure 6.7. Four transparency masks with
minimum feature size of ~50 were used. First layer defines diamond pattern, second layer
emitter contact, third layer gate hole, and fourth layer metal. Oxidized four inch wafer was
used as a substrate. One wafer consists of two display chips.

One chip consists of twelve 2.5mm wide and 35mm long diamond resistors, doped
with B powder and grown on 4" oxidized Si wafer using HFCVD system. Separation
between each resistor is O.5mm. Resistance of diamond cathode is 20 ~ 32 k9. Assuming
average film thickness of 2 um, resistivity is in the range of 2.86 ~ 4.57 Qcm. This value
is approximately in accordance with the resistivity value measured by four point probe
system. Patterned diamond film is shown in Figure 6.7 (a).

Due to the limitation of low temperature oxidation facility and the simplicity of spin-
on-glass (SOG) process, SOG was selected for an insulating layer. First SOG used was
Filmtronics’ SOG (Model No. GF51 1F). GF511F is an alcohol solution which is applied
to a semiconductor surface to yield a pure Si02 film similar in characteristics to a pyrolitic
oxide. SOG yields a pyrolytic oxide essentially free of sodium ions or other undesirable
elements deleterious to semiconductor devices. Films may be formed on a wide variety of
flat surfaces by spinning, spraying or dipping. For precise control of flatness and greatest
uniformity from wafer, spinning is the recommended procedure for applying SOG film.

After spinning, the film remaining on the surface of the wafer must be hardened or
densified by heat treatment. The heat treatment is carried out at temperatures from 200 °C

to temperatures in excess of 800 or 900 °C in air, oxygen or nitrogen. The extent of the

155

A A, Design view

Diamond

Cross sectional view

- across A—A’

Oxidized Si

 

 

Wafer photography

 

Figure 6.7 (a). Design vew, cross sectional view, and wafer photography
after diamond deposition.

Cl
l6
“l
fih

At

obt
aHt
cha

SK]

50m
EXpa
SOG
Care

Show!

SOG.
Ihicks

Will {C

156

densification of the film is indicated by the etch rate in dilute HF solutions. A film
densified by heat treatment at 200 0C for 15 minutes will exhibit an etch rate 1000 times

faster than a thermal oxide. A similar heat treatment at 900 °C will yield a film whose etch
rate is two to three times faster than a thermal oxide. In addition to the reduction in rate of

etching by HF, heat treatment also affects the hardness of the film, as evidenced by its

resistance to scratching with a steel stylus. Heat treatment at 800 °C will yield a film

which will exhibit scratch resistance comparable to a thermal oxide. The thickness of Si02
film formed from Silicafilm is determined by the spin speed when the material is spun on.

At 3000 rpm, a film thickness of 2000 A is obtained after heat treatment at 200 °C. The
film thickness decreases as spin speed is increased. FIlms up to 8000 A thick may be
obtained by successive applications of material applied while the wafer is spinning and
allowing 5 or 10 seconds between successive application. The dielectric and optical
characteristics of films formed from SOG are quite similar to the characteristics of pure

Si02 formed by the oxidation of silicon at high temperature.
In this experiment, SOG was spin-coated three times at 3000 rpm to get at least 8000

A. Sample is baked in convection oven at 200 °C. When sample is cooled down rapidly,
some cracks in SOG film are found due to rough nature of diamond film and thermal

expansion coefficient mismatch between diamond and SOG film. To assure crack free

SOG film, slow ramping down heat cycle of less than 3 °C/min is required. Although great
care has been taken, crack problem hindered further process development. Cracks are
shown in Figure 6.8 (a).

Emulsitone’s silicafilm 10,000 was used to get crack free film instead of Filmtronics’
SOG. Silicafilm 10,000 is similar to Filmtronics’ SOG GF511F with the exception that a
thickener has been added to the solution. When applied by spinning a thick glassy layer

will form, which must be heated to remove the thickener. After spinning at 3000 rpm the

fil
tr:
b)
Ft
he
fill
ho

6.5

the

mc

Tw
dia
SUI

[WC

fro:

ho}:

ShOl

157

film should be heated at 200 °C for 15 minutes in air to densify the film. The heat

treatment is followed by heating at 450-500 °C for one hour in air to remove the thickener

by oxidation. Silicafilm 10,000 etches more rapidly than more conventional Si02 films.

For example, a typical etch rate in 5% HF solution at 25 °C is 150 A/second for a film
heated at 500 0C. Silica 10000 was spin-coated three times at 3000 rpm to get 3 ttm thick

film. Sample is baked in convection oven at 200 °C and baked in furnace at 500 °C for one
hour. Crack problem has dramatically reduced. A sample without crack is shown in Figure
6.8 (b). Occasional crack was still observed if heat cycle is not properly controlled.

At the two ends of each resistor area SOG is etched to make emitter contacts. Cr is
thermally evaporated on top of diamond where SOG is etched as well as on top of SOG in
most area. This step is shown in Figure 6.7 (b). Cr is patterned to create gate holes.
Through open Cr area, SOG is etched and it is intentionally overetched to create undercut.
Two different gate structures were used to create either one by one 200 um wide hole in
diameter or two by two 100 um wide hole to check emission properties from different
structures. Half of gate mask has one by one 200 um wide gate holes and another half has
two by two 100 um wide holes as shown in Figure 6.7 (c). Since gate hole size is very
large compared to SOG thickness, it is not expected to have emission current contribution
from the center of the gate hole. Majority of current is expected to initiated near the gate
hole edges.

Cr is etch once again to separate each six row and twelve columns. Finished sample is

shown in Figure 6.7 ((1).

6.2.2.2 Testing

Figure 6.9 shows measurement setup for the sample. Gate voltage was varied from 40

to 80 Volts. Anode voltage was varied from 100 to 500 Volts. While emission pattern is

158

 

Cr

Cross section across
A-A’

 

Cr

 

Cross section across
B -B ’

 

 

Figure 6.7(b). Design view, cross sectional view, and wafer photography after
emitter contact etch followed by Cr evaporation.

159

 

 

Figure 6.7(c). Design view, optical microscopy picture and cross sec-
tional view of the sample after gate definition.

160

 

 

 

 

 

Figure 6.7(d). Design view and finished wafer. Two different
gate structures are also shown.

161

 

Figure 6.8. Samples with (a) cracks and (b) no crack.

162

Phosphor Anode

 

 

Figure 6.9. Measurement setup for self-aligned diamond FED in triode
mode. Emitter is grounded at the pad shown in Figure 6.7(b).

la (Amp)

Ig (Amp)

163

 

 

 

 

 

 

 

 

 

 

—4
1 0 ; r r
-5 I I ‘
1 O y 8 g g
i
. O Vg=40V 1
10‘6 x vg=6ov .3
* Vg=80V 1
1o-7 1 1 e
1 00 200 300 400 500
Va (Volts)
—4
1 0 ~ 1
o vg=4ov ‘
10.5 X Vg=60V
# * Vg=80V
-6? .
1 O f I 3.: x
-7,
1 0 * ‘ ‘
1 00 200 300 400 500
V0 (Volts)

Figure 6.10. Anode current Ia and gate current Ig as a function of
anode voltage Va and gate voltage Vg

 

 

 

.‘ Ill-J‘-

Common CI‘IHIICI' contact

 

Cell 7,1 Common gate contact

Cell 7.3 Anode contact

Figure 6.11. Emission pattern of the FED chip in triode configuration

165

recorded in the camera, gate current and anode current were monitored. Figure 6.10 (a)
shows the anode current characteristics for gate voltage of 40, 60, and 80 Volts and Figure
6.10 (b) shows the corresponding gate current characteristics. These are typical I-V

characteristics for triode structure [117]. Ia/Ig ratio of ~10 is found. This value is in the

better end of reported value of 0.1 - 10 [42]. Figure 6.11 shows emitted pattern. Although
it was successful to have two pixels working, not all the pixel emitted. Further process and
design improvement is needed to address these problems. However, the purpose of this
experiment was to demonstrate the possibility of new self-aligned process.

In Chapter 5, it was revealed that films grown with high nucleation density resulted in
higher current density and lower tum-on voltage. Another advantage to use films with high
nucleation density in this experiment is smoothness of high nucleation density film. It
would reduce crack problem encountered in this experiment with low nucleation density

film.

6.3 Field Emitter Pressure Sensor

Commercially available piezoresistive and capacitive pressure sensors are subject
to limitations of temperature and sensitivity, respectively. Due to their temperature inde-
pendence and radiation hardness, the use of field emitter devices for pressure or displace-
ment sensors has been proposed [4,118]. Although simulation of field emitter pressure
sensors shows high sensitivity [118], nobody has produced pressure sensor based on field
emission yet. Using the fabrication technology developed, we demonstrate the design and
fabrication of a pressure sensor for the first time [119]. Main technique used is self-vac-
uum-sealing process [120]. The field emission behavior was confirmed in vacuum cham-

ber.

 

166

6.3.1 Simulation

To study the displacement of diaphragm depending upon geometries and materials of
devices, ANSYS finite element analysis was performed. Actual geometric and material
data are fed into the simulator. Input variables to the simulation are geometric parameters
of devices such as thickness, width, and length of diaphragm. Also material properties
such as Young’s modulus, Poisson’s ratio and shear modulus are used as input parameters.
Applied pressure to the diaphragm is used as a boundary condition to the problem.
Constructed model was then meshed, solved and resulted in displacement. Figure 6.12
depicts a sample contour plot of displacement while Figure 6.13 shows displacement plot.
Having known displacement value, assuming uniform field emission from whole cathode
area, total current from a device is calculated from F-N equation as follows. We can
calculate the current for each element by approximating diaphragm as a collection of flat
squares of same size with different distance to the emitter. Summation of all the squares
with different distance (1 resulted in total current. This concept is shown in Figure 6.14. It
was found that 4 pm thick A1 diaphragm, which was diaphragm thickness of fabricated

sample, can yield curve an order of magnitude change in anode current per applied

pressure of 40 dyne/cmz.

6.3.2 Fabrication

Pressure sensors were also fabricated in test chip III. The complete fabrication process
for diamond field emitter pressure sensor is shown in Figure 6.15. Fabrication process is
identical to type A devices described in Chapter 4 up to first A1 deposition. Explanation
for the fabrication process is omitted up to this point. Minor modification to type A device
is incorporated to achieve self vacuum sealing. Al is patterned to form holes and part of
sacrificial photoresist is now exposed to be removed (Figure 6.15c). The sacrificial layer

underneath A1 is removed using either acetone or remover. Wafer is then cleaned with de-

167

yam

manages Mo oBmoQ

635%? «a 522:5“ @672 .2.» 2:5

 

168

 

(,1oxx—t)

  
    
 

22.426

        

1
33.941

ha.97ll 25'456

12.728 21.213

DIST

0.329

8,485 1 38.184

4.243

29.698

Figure 6.13. Displacement value of diaphram along A-A’ of Figure 6.12.
Unit is in um.

169

 

Bent diaphragm apporximation

 

Emitter surface

Figure 6.14. Approximation of pressure sensor to calculate total current.

(a)

(b)

(C)

 

:L_—‘: Oxide

 

170

Si

 

 

Diamond

  

 

Sacrificial layer

 

 

Aluminum

 

 

 

Figure 6.15. Cross-sectional views of a pressure sensor at each pro-
cess step and the corresponding pictures of top view.

(d)

(e)

 

 

171

 

 

 

Figure 6.15. Continued.

 

 

172

ionized (DI) water numerous times to ensure cleaning of hollow inside. Al is evaporated
once again at grazing incidence angle to seal the sensor and patterned appropriately. SEM
pictures in Figure 6.15 ((1) shows holes just before its complete closure. Sealing
evaporation has to be done three to four times in Al evaporator. In general, with maximum
Al clip loading, one time Al evaporation resulted in 1 pm thick film. It was necessary to
repeat this procedure to completely seal the device. Al is then etch to separate each
devices. Completed structure is shown in Figure 6.15 (e). The fabricated device is
intended to be self-vacuum sealed and can be operational at the atmosphere if successful
vacuum sealing is achieved. The anode is processed as a diaphragm that can be displaced
by applying pressure, which will eventually change the field emission current. A] is also

evaporated on the backside of wafer to serve as a cathode electrode (Figure 6.15e).
Completed samples are annealed at 400 0C in N2 ambient for 30 min. The current
voltage (I-V) data is taken by placing the wafer inside a vacuum chamber with a pressure

of 10'6 Torr first to make sure emitter itself is functioning its successful field emission

behavior.

6.3.3 Testing

The I-V characteristic of one pressure sensor as a field emitter itself is shown in

Figure 6.l6(b), the corresponding F-N plot is shown in the inset. This data is taken for a

pressure sensor with 40 x 40 um2 diamond emitter inside high vacuum chamber. The
exponential I-V curve and the straight—line F-N plot indicate that measured current is

indeed field emission current. The current density measured at 0.2MV/cm is approxi-

mately 0.5A/cm2 as compared to reported values in the range of 0.1~10A/cm2. Field emis-

sion behavior is continued even though absolute value of current density is decreased as

173

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

(a)
1
mo:
10 i 10'6 Torr ° ° 0 1
_13 o ° 10'3 Torr
10 g 0 x x x x i
2 —2E o x x X x 1
310 E o x x x F-N Plot 1
41-! i x -2
-3 x 10 .
g) 10 r o x 1
S r 10 ‘75”
0 104? x N 1
q) x SIG-6' IT“
'0 —5'
8 10 r 1‘o_8 1
< -6: -10 - .
10 gr x 1° 0 0.02 0.04 0.06 1
10—73 . . 11N
0 50 100 150 200
Anode Potential (Volt)

(1))

Figure 6.16. (a) Measurement setup and (b) I-V curve and F -N plot of a dia-

mond pressure sensor when measured at a pressure of 10‘6 Torr
and 10'3 Torr.

174

 

N2 BLOW

 

 

 

 

 

 

Figure 6.17. Measurement setup for a pressure sensor at atmosphere.

175

the pressure inside vacuum chamber is raised up to 10'3 Torr range. This is due to effective
work function change of emitter surface. As we have reported earlier, diamond field emit-
ter operates at relatively higher pressure than field emitter devices made of materials other
than diamond.

True pressure sensor operation is not achieved at this time. Applied external force,

preferably pressurized air or N2 blow to anode membrane could bend anode membrane

since anode is very thin diaphragm. This bending decreases the distance between anode

and cathode and subsequently increases the electric field affecting the field emission

current. Simulation study showed that ~40 dyne/cm2 force can change the emission
current by one order of magnitude. In our case, when tested in atmospheric pressure with
setup shown in Figure 6.17, pressure sensor didn’t show measurable emission current.
When this non working sensor was put inside the vacuum chamber back, reproducible
emission current was again measured. This implies that vacuum sealing is not properly
performed at this moment. One of the possible reasons for inadequate vacuum sealing may
be due to porous Al anode diaphragm. Depending on Al evaporation condition, porous Al

films have been witnessed [121].

6.4 Summary

In this chapter, the design, fabrication and characterization of two kinds of triode type
diamond FEDS and the pressure sensor were described. The main purpose of these works
was to demonstrate the possibility of these applications. Light emission pattern is recorded

for a 1 x 4 pixel triode FED for the first time.

CHAPTER 7

CONCLUSIONS AND FUTURE
RESEARCH

7.1 Summary and Conclusions

0 Development of IC compatible fabrication process for diamond FED in triode

Using a testchip, IC compatible fabrication process for diamond FED in triode mode
was developed for the first time. As a result, a unique testchip was designed and
fabricated. Operation of diamond FED in triode mode was reported first time. Key
technology developed was constructing a built-in anode to the structure. Self-aligned

process technology was also successful.

0 Design and execution of experiments to collect emission data from wide variety

of diamond samples

Films with different growth conditions were prepared and careful characterization
resulted in valuable data in a systematic way. Characterization was performed for the

samples, with CH4 concentration varying from 0.5% to 2%, doped with B powder and

B203 wafer, seeded with low and high nucleation density, and implanted with carbon at

the dose of 5x1016 cm’2 and 5x1015 cm’2 and dose energy of 20keV and 50 keV.

176

177

0 Identifying best diamond for field emission and explanation of results

Based on a large amount of field emission data collected from a variety of samples, it
was found that emission from diamond is enhanced when (i) sp3/sp2 is low, (ii) peak at

1332 cm'1 is wider, (iii) grain size in the film is small and roughness of film is small, (iv)
film is highly doped, (v) film is damaged by ion implantation, especially density of
damage is high and location of damage is shallow, and (vi) grain boundary is present.
These seems to suggest field emission from polycrystalline diamond is affected by (i)

defects and (ii) field enhancement effect at grain tips in addition to NBA of diamond.

7 .2 Future Research

Although in the present research, realization of triode mode diamond FEDs were first
accomplished, there is a need in the following areas for introduction of the diamond FED

in the near future.

(i) The better understanding and prediction of electron emission mechanisms of
semiconducting CVD diamonds is important to optimize the performance of

diamond FED.

(ii) It is important to establish a technology to grow diamond on a large glass
substrate to compete with LCD market. Low temperature diamond deposition

system with uniformity over large area is essential.

(iii) Efforts should be made to increase emission site density to achieve high

current density.

178

APPEN DEX A.

A simplified schematic diagram of this system is shown in Figure A.1. Based upon the
Operational function, a RFCVD reactor can be categorized into the following seven major

subsystems.
(1) Chamber
(2) RF power system
(3) Pressure and flow controls
(4) Vacuum system
(5) Temperature readout and control

(6) Process sequencers

° Chamber

The main chamber and fixture assembly are constructed of aluminum which
provides RF shielding and grounding.The substrate electrode contains a coiled resistance
heater that provides uniform high temperature control. The aluminum fixtures, for the
most part, are not attacked by reactive gasses typically used in deposition applications and
are easily cleaned in a hot reactive gas plasma. Fixture surfaces are smooth to minimize
plasma arcing and to facilitate cleaning. The lift assembly is designed to raise the top plate
and electrode, allowing complete access to the substrate surface. The air design allows

smooth movement and high weight capacities.

0 RF power system

179

The RF generator is 13.56 MHz crystal controlled with 3 KW output. Automatic
power control is used in conjunction with the RF generator and allows precise regulation
of output power. Multiple preset power levels provide control. Automatic tuning allows
for constant impedance matching which enables a predetermined Watt density to be
maintained within the plasma. The results are exact process control and a virtual "hands-
off" operation. The electrode voltage potentials have a direct relationship to process

conditions within the reactor.

0 Pressure and flow controls

The mass flow control allows precise metering of process gasses. The system
controller is activated by manual or automated means and attains preset conditions
through closed circuit monitoring. Total, individual, or ratio flow values are digitally
displayed. Rotometers are provided for use during the process purging cycle where precise
flow monitoring is not required. Solenoids activate gas flow and are manually or
automatically activated. Chamber pressure is automatically adjusted by a valve, driven in a
closed loop, capable of maintaining a preset level regardless of varying conditions. The

valve controller can be operated in either the manual or automatic modes.

0 Vacuum system

The gas control system accepts multiple gasses and is constructed of stainless steel.
This component is an integral part of the vacuum system and is subsequently vacuum
tight. Flushing of the process channels with an inert gas is easily accomplished during the
cleaning cycle of the reactor. The vacuum system allows pump down of the reaction
chamber to a low pressure, thus eliminating background gasses that may contaminate the
process. Pressure and flow requirements are met through the use of an automated

throttling type process valve. A pressure controller drives the throttle valve

180
0 Temperature readout and control

The temperature controller and readout displays and regulates the temperature of the
substrate. This controller monitors temperature through two independent control loops
using separate sensors. An over-temperature condition is prevented by the use of a sensor
located on the substrate. The temperature of this substrate is precisely controlled by a
resistance heater which is capable of rapid cycling for heating, independent of plasma

power.

0 Process sequencer

The manual processing sequencer is a central switching network that allows manual
sequencing from a single panel. The optional automatic process sequencer controls the
total deposition process and can be programmed for "one button" operation. The
sequencer is microprocessor-based, and is capable of storing four programs, a total of 204
steps. Each step can make a decision based on 8 inputs and provide up to 16 outputs in a
standard configuration. Also included is a count-down timer that is programmable from .1

second to 99.99 hours.

0 Operation procedure

The operation sequence of RFCVD system is carefully laid out to operate machine in
different condition than original design without damaging the system.

Starting the system:

(1) Turn on the cooling water.

(2) Turn on the power switches: located on the system front panel.

(3) Turn on the nitrogen flow: use the lever in the backroom.

(4)

(5)

(6)

(7)

(8)

(9)

(10)

(11)

(12)

(13)

181

Turn on the methane flow switches: located in the backroom. Ar flow switches

are always on.

Vent the system: turn up the vent switch. Wait for 3 minutes.

Open the system: Turn the hoist switch up. Push the two white knobs in to open

the system.

Clean the system. Load the sample.

Close the system: Turn the hoist switch down. Push the two white knobs in to

close the system.

Turn the vent switch down. Turn the roughing switch up to evacuate the system.

Wait until the pressure is near 40 mTorr. Turn the heat exchange on.

Set the total pressure in the chamber to the desired value. For example: 1.0 at 10,
1 and 0.1 Volts means 1, 0.1 and .01 Torr, respectively. Rotate the control switch

from "open" mode to "automatic" mode.

Turn up the process switch and the Ar flow switch on the front panel.

Increase the Ar flow using the Ar flow meter (meter #3). Wait until the pressure

stabilizes at the desired value.

(14) Turn up the methane flow switch on the front panel.

(15)

(16)

Increase the methane flow using the methane flow meter (meter #2).

Turn on the RF Generator. Wait 1 minute until the "RF OFF" switch turns on.

 

182

(17) Push "RF ON" switch. Increase the power using the power knob. Watch the

incident and reflected powers. Use the automatic mode of Tuning and Load to

reduce the reflected power.

(18) Watch the substrate temperature and the upper stage temperature during

deposition.

Shut down procedures:

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

Turn RF plasma off. Decrease the power knob back to zero. Let the cooling fan

run for 3 minutes. Turn off RF generator power.

Turn Ar and methane flow switches down. Reduce the flow meters reading back

10 Zero.

Rotate the control switch from "automatic" mode to "open" mode. Wait until

system cools down. Turn cooling water off.

Turn the roughing valve down (i.e. No pumping down).

Vent the system: Turn the vent valve up. Wait for 3 minutes.

Open the system: Turn the hoist switch up. Push the two white knobs in to open

the system.

Take your sample(s) out.

Close the system: Turn the hoist switch down. Push the two white knobs in to

close the system.

Turn the vent switch down. Turn the roughing switch up to evacuate the system.

we
I

183

(10) Wait until the pressure is near 40 mTorr.

(l 1) Turn the roughing valve down.

(12) Turn off the power switches: located on the system front panel.

(13) Turn off the methane flow switches: located in the backroom.

 

4
00

.538: 96...: a 88:23 .3. 2:2...

 

3.5.80 >65
30 330E

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

+

 

 

 

.9980
530m
£3893.

1 I .T 83350
830m "B

 

 

 

 

 

 

 

hoocmswom n
l l l 1 $805
1 3282
_ a
— a
I .8250 .1
8:30am _
3:50 _
v>—d> _ QE
a
238mb _ 8.593 op.
35:00
8388505
_
_
7 — _
.8280
13880.“th

 

x5302
wEnBaE
$888.3

 

 

 

Howcmcoxm
“mom

 

 

 

 

 

 

I|._

 

 

 

185

APPENDIX B.

5 'Copyright Hewlett-Packard 1984, 1985

10 '

15 ' Set up program for MS-DOS HP-IB I/O Library

20 ' For use independent of the PC instrument bus system

25 '

30 DEF SEG

35 CLEAR ,&HFE00

40 I=&HFE00

45 '

50 ' PCIB.DIR$ represents the directory where the library files
55 ' are located

60 ' PCIB is an environment variable which should be set from MS-DOS
65 ' i.e. A:> SET PC1B=A:\LIB

70 '

75 ' If there is insufficient environment space a direct assignment
80 ' can be made here, i.e

85 ' PCIB.DIR$ = "A:\LIB"

90 ' Using the environment variable is the preferred method
95 '

100 PCIB.DIR$ = ENVIRON$("PCIB")

105 I$ = PCIB.DIR$ + "\PCIBILC.BLD"

110 BLOAD I$,&HFE00

115 CALL I(PCIB.DIR$, 1%, 1%)

120 PCIB.SEG = 1%

125 IF J%=0 THEN GOTO 160

130 PRINT "Unable to load.";

135 PRINT " (Error #";J%;")"

140 STOP

145 ’

150 ' Define entry points for setup routines
155 '

160 DEF SEG=PCIB.SEG

165 0.8 =5

170 CS = 10

175 IV =15

180 LC =20

185 LP =25

190 LD.FILE = 30
195 GET.MEM = 35

200 LS =40
205 PANELS =45
210 '

215 ' Establish error variables and ON ERROR branching
220 '

186

225 DEFERR = 50

230 PCIB.BRR$ = STRING$(64,32)

235 PCIB.NAME$ = STRING$(16,32)

240 CALL DEF.ERR(PCIB.ERR,PCIB.ERR$,PCIB.NAME$,PC[B.GLBERR)
245 PCIB.BASERR = 255

250 ON ERROR GOTO 410

255 ’

260 J=-1

265 I$=PCIB.DIR$+"\HPIB.SYN"

270 CALL O.S(I$)

275 IF PCIB.BRR<>0 THEN ERROR PCIB.BASERR

280 '

285 ' Determine entry points for HP-IB Library routines

290 '

295 1:0

300 CALL I.V(I,IOABORT,IOCLEAR,IOCONTROL,IOENTER)
305 IF PCIB.BRR<>0 THEN ERROR PCIB.BASERR

310 CALL I.V(I,IOENTERA,IOENTERS,IOEOI,IOEOL)

315 IF PCIB.BRR<>O THEN ERROR PCIB.BASERR

320 CALL I.V(LIOGE'I'I‘ERM,IOLLOCKOUT,IOLOCAL,IOMATCH)
325 IF PC1B.ERR<>0 THEN ERROR PCIB.BASERR

330 CALL I.V(I,IOOUTPUT,IOOUTPUTA,IOOUTPUTS,IOPPOLL)
335 IF PCIB.BRROO THEN ERROR PCIB.BASERR

340 CALL I.V(I,IOPPOLLC,IOPPOLLU,IOREMOTE,IORESET)
345 IF PCIB.BRR<>0 THEN ERROR PCIB.BASERR

350 CALL I.V(I,IOSEND,IOSPOLL,IOSTATUS,IOTIMEOUT)
355 [F PCIB.BRR<>O THEN ERROR PCIB.BASERR

360 CALL I.V(I,IOTRIGGER,IODMA,J,J)

365 IF PCIB.BRROO THEN ERROR PCIB.BASERR

370 CALL CS

375 I$=PCIB.DIR$+"\HPIB.PLD"

380 CALL L.P(I$)

385 IF PCIB.BRR<>0 THEN ERROR PCIB.BASERR

390 GOTO 475

395 '

400 ' Error handling routine

405 '

410 IF ERR=PCIB.BASERR THEN GOTO 425

415 PRINT "BASIC error #";ERR;" occurred in line ";ERL

420 STOP

425 TMPERR = PCIB.BRR

430 IF TMPERR = 0 THEN TMPERR = PCIB.GLBERR

435 PRINT "PC Instrument error #";TMPERR;" detected at line ";ERL
440 PRINT "Error: ";PCIB.ERR$

445 STOP

450 '

187

455 ' COMMON declarations are needed if your program is going to chain

460 ' to other programs. When chaining, be sure to call DEFERR as

465 ' well upon entering the chained-to program

470 '

475 COMMON PCIB.DIR$,PCIB.SEG

480 COMMON LD.FILE,GET.MEM,PANELS,DEF.ERR

485 COMMON PCIB.BASERR,PCIB.ERR,PCIB.ERR$,PCIB.NAME$,PCIB.GLBERR
490 COMMON
IOABORT,IOCLEAR,IOCONTROL,IOENTER,IOENTERA,IOENTERS,IOEOI,IOEO
L,IOGETTERM,IOLLOCKOUT,IOLOCAL,IOMATCH,IOOUTPUT,IOOUTPUTA,IO
OUTPUTS,IOPPOLL,IOPPOLLC,IOPPOLLU,IOREMOTE,IORESET,IOSEND,IOSPO
LL,IOSTATUS,IOTIMEOUT,IOTRIGGER,IODMA

495 '

500 FALSE = 0

505 TRUE = NOT FALSE

510 NOERR = 0

515 EUNKNOWN = 100001!

520 ESEL = 100002!

525 ERANGE = 100003!

530 ETIME = 100004!

535 ECTRL = 100005!

540 EPASS = 100006!

545 ENUM = 100007!

550 EADDR = 100008!

555 COMMON FALSE, TRUE, NOERR, EUNKNOWN, ESEL, ERANGE, ETIME,
ECTRL, EPASS, ENUM, EADDR

560 '

565 ' End Program Set-up

570 ' User program can begin anywhere past this point

575 '************************* DECLARATIONS ************************
578 ' open "hongdat" for output as #1

590 K20=712 : MPSSO=709 : DVM=705

592 CSM=719 : HPDVM=722

595 KEY OFF

1000 "'**************************** IV ******************************
1005 CLS

1010 PRINT

1015 PRINT ""

1020 PRINT " INPUT YOUR CHOICE OF THE FOLLOWING:"

1025 PRINT " 1. I FOR IV MEASUREMENTS "

1030 PRINT " 2. R TO RETURN TO MAIN MENU "

1035 PRINT ""

1040 INPUT " COMMAND "; AS

1045 IF A$ = "1" OR A$ = "i" THEN GOTO 1059

1050 IF A$ = "R" OR A$ = "r" THEN GOTO 605

1055 GOTO 1040

 

188

1059 PRINT " TURN ON HP3457A, set the voltage"

1060 input "Name of file to save this measurement data] "; MS

1061 input "Name of file to save this measurement data2"; n2$

1065 open n1$ for output as #1 : open n2$ for output as #2

1066 input "input Voltage ";inv

1067 write #1, inv : for r=1 to 10000 : next r

1070 'for s=1 to 10

1075 C$="RESET" : DEV=HPDVM : GOSUB 3600 'initialize voltmeter
1100 'FOR T=1 TO 2000 : NEXT T ' WAIT TILL CURRENT IS SET
1120 C$="DVC" : DEV=HPDVM : GOSUB 3600 'READ VOLTS
1125 V=VAL(info$) : PRINT inv;V

l 130 write #1, V : write #2, inv, v

1135'nexts

1145 INPUT " set another voltage (Y/N)";TMP$

1150 IF TMP$="Y" OR TMP$="y" THEN GOTO 1066

1 155 close #1: close #2

1160 GOTO 1010

3595 '"****************** I/O **************

3600 LENGTH=LEN(C$)

3605 CALL IOOUTPUTS(DEV,C$,LENGTH)

3610 MAX.LENGTH = 30

3615 ACTUALLENGTH = 0

3620 INFO$ = SPACE$(MAX.LENGTH)

3625 CALL IOENTERS(DEV,INFO$,MAX.LENGTH,ACTUAL.LENGTH)
3630 INFO=VAL(INFO$)

3635 RETURN

7600 END

BIBLIOGRAPHY

10.

11.

189

BIBLIOGRAPHY

R.H. Fowler and L. Nordheim, “Electron Emission in Intense Electric Fields”, Proc.
Roy. Soc. London, ser. A, vol. 119, p. 173, 1928.

C. A. Spindt, I. Brodie, L. Humphrey, and E. R. Westerberg, “Physical properties of
thin film field emission cathodes with molybdenum cones”, J. Appl. Phys. 47, p.
5248, 1976.

T. Utsumi, “Vacuum microelectronics: what's new and exciting”, IEEE trans.
Electron Devices, ED-38, p. 2276-83, 1991.

H. H. Busta, “Review Vacuum microelectronics-1992”, J. Micromech. Microeng., 2,
p.43,1992.

I. Brodie and P. R. Schwoebel, “Vacuum Microelectronic Devices”, Proc. of IEEE,
82,p.1006,l994.

A. Ghis, R. Meyer, P. Rambaud, F. Levy, and T. Leroux, “Sealed vacuum devices:
fluorescent microtip displays”, IEEE Trans. Electron Devices, ED-38, p. 2320-2,
1991.

S. Matsumoto, Y. Sato, M. Kamo and N. Setaka, "Vapor Deposition of Diamond
Particles from Methane," Jpn. J. Appl. Phys., vol. 21, pp. L183-L185, 1982.

J. C. Angus and CC. Hayman, "Low-temperature, Metastable Growth of Diamond
and Diamond-like Phases," Science, vol. 241, pp. 913-922, 1988.

E]. Himpsel, J.A. Knapp, J.A. Van Vechten, and DE. Eastman, “Quantum
Photoyield of Diamond(111)- A Stable Negative-Affinity Emitter”, Phys. Rev. B,
vol. 20, No. 2, p. 624, 1979.

M. Geis, N.N. Efremow, J.D. Woodhouse, M.D. Aleese, M. Marchywka, D.G.
Socker, and LP. Hochedez, “Diamond Cold Cathode”, IEEE Elec. Dev. Lett., vol.
12, No.8, p. 456, 1991.

M. Geis, J.C. Twichell, C.O. Bozler, D.D. Rathman, N.N. Efremow, K.E. Krohn,
M.A. Hollis, R. Uttaro, and TM. Lyszczarz, Proc. of 6b Int. Vacuum

12.

13.

14.

15.

l6.

17.

18.

19.

20.

21.

22.

190

Microelectronics Conf., p. 160, 1993.

C. Xie, C. N. Potter, R. L. Fink, C. Hilbert, A. Krishnan, D. Eichman, N. Kumar, H.
K. Schimidt, M. H. Clark, A. Ross, B. Lin, L. Fredin, B. Baker, D. Patterson, and W.
Brookover, “Use of diamond thin films for low cost field emission displays”, Proc. of
7th Int. Vacuum Microelectronics Conf., p. 229, 1994.

NS. Xu, R.V. Latham, and Y. Tzeng, “Cold Electron Emission Characteristics of
Broad-Area Diamond Coated Mo Electrodes”, Proc. of 2nd Int. Conf. on the
Applications of Diamond Films and Related Materials, p. 779, 1993.

J. Liu, V. V. Zhimov, A. F. Myers, G. J. Wojak, W. B. CHoi, J. J. Hren, S. D. Wolter,
M. T. McClure, B. R. Stoner, and J. T. Glass, “Field emission characteristics of
diamond coated silicon field emitter”, J. Vac. Sci. Technol. B 13(2), p. 422, 1995.

D. Hong, and M. Aslam, "Field Emission from P-type Polycrystalline Diamond
Films," J. Vac. Sci. Technol. B, vol. 13, no. 2, p. 427, 1995.

A. Modinos, “Field, thermionic, and secondary electron emission spectroscopy”,
Plenum Press, New York, 1984.

H. B. Dwights, “Tables of Integrals and other Mathematical Data”, The Macmillan
Company, New York, 1934.

C. A. Spindt, C. E. Holland, A. Rosengreen, and I. Brodie, “Field-emitter arrays for
vacuum microelectronics”, IEEE Trans. Electron Devices, ED-38, p. 2355-63, 1991.

G. G. P. van Gorkan and A. M. E. Hoeberechts, “Performance of Silicon Cold
Cathodes”, J. Vac. Sci. Tech. B, vol. 4, p. 108-111, 1986.

A. T. Collins, “The electrical and optical pr0perties of diamond; do they favor device
applications?”, Diamond, SiC, and Related Wide Bandgap Semiconductors, vol.
162, J. T. Glass, R. Messier, N. Fujimori Eds., Pittsburg, PA; Material’s Research
Society, p. 3-14, 1990.

P. K. Bachmann and R. Messier, “Emerging technology of diamond thin films”,
C&EN Special Report, May 15, 1989.

A. Y. Tcherepanov, A. G. Chakhovskoi, and V. B. Sharov, “Flat panel display
prototype using low-voltage carbon field emitters”, J. Vac. Sci. Technol. B 13(2), p.
482. 1995.

 

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

191

Y. shoda, D. M. Tanenbaum, S. W. Turner, and H. G. Craighead, “Fabrication of
arrayed glassy carbon field emitters”, J. Vac. Sci. Technol. B 15(2), p.343, 1997.

S. Lee, B. K. Ju, Y. H. Lee, D. Jeon, M. H. Oh, “Fabrication and field emission study
of gated diamondlike-carbon-coated silicon tips”, J. Vac. Sci. Technol. B 15(2), p.
425, 1997.

K. C. Park, J. H. Moon, S. J. CHung, M. H. Oh, W. I. Milne, and J. jang,
“Relationship between field emission characteristics and hydrogen content in
diamondlike carbon deposited by the layer-by-layer technique using plasma
enhanced chemical vapor deposition”, J. Vac. Sci. Technol. B 15(2), p. 428, 1997.

K. C. Park, J. H. Moon, S. J. Chung, J. H. Jung, b. K. Ju, M. H. Oh, W. I Milne, M.
K. Han, and J. Jang, “Field emission properties of ta-C films with nitrogen doping”,
J. Vac. Sci. Technol. B 15(2), p. 431, 1997.

E. I. Givargizov, V. V. Zhimov, N. N. Chubun, and A. B. Voronin, “Diamond cold
cathodes for electron guns”, J. Vac. Sci. Technol. B 15(2), p. 442, 1997.

V. V. Zhimov, E. I. Givargizov, A. V. Kandidov, B. V. Seleznev, and A. N. Alimova,
“Emission characterization of diamond-coated Si field emission arrays”, J. Vac. Sci.
Technol. B 15(2). [3446, 1997.

E. I. Givargizov, V. V. Zhimov, N. N. Chubun, and A. N. Stepanova, “Fabrication of
field emission diaplay prototype based on Si field emission arrays with diamond
coating”, J. Vac. Sci. Technol. B 15(2), p. 450, 1997.

K. C. Park, J. H. Moon, S. J. Chung, and J. Jang, “Field emission of nitrogen doped
diamondlike carbon films deposited by plasma enhanced chemical vapor
deposition”, J. Vac. Sci. Technol. B 15(2), p. 454, 1997.

S. Lee, S. Lee, S. Lee, D. Jeon, and K. R. Lee, “Self-aligned silicon tips coated with
diamondlike carbon”, J. Vac. Sci. Technol. B 15(2), p. 457, 1997.

W. P. Kang, J. L. Davidson, M. A. George, I. Milosavljevic, Q. Li, J. F. Xu, and D. V.
Kems, “Physical characterization of diamond pyramidal microtip emitters”, J. Vac.
Sci. Technol. B 15(2), p. 460, 1997.

S. Kim, B. K. Ju, Y. H. Lee, B. S. Park, Y. J. Baik, S. Lim, and M. H. Oh, “Emission
Characterixtic of Diamond-Tip FIeld Emitter Arrays Fabricated by Transfer Mold
Technique”, J. Vac. Sci. Technol. B 15(2), p. 499, 1997.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

192

Y. A. Grigoriev, A. I. Petrosyan, V. V. Penzyakov, V. G. Pimenov, V. I. Rogovin, V. I.
Shesterkin, V. P. Kudryashov, and V. C. Semyonov, “Experimental study of matrix
carbon field-emission cathodes and computer aided design of electron guns for
microwave power devices, exploring these cathodes”, J. Vac. Sci. Technol. B 15(2),
p.503,1997.

W. Zhu, G. P. Kochanski, S. Jin, and L. Seibles, “Electron field emission from
chemical vapor deposited diamond”, J. Vac. Sci. Technol. B 14(3), p. 2011, 1996.

A. F. Myers, S. M. Camphausen, J. J. Cuomo, J. J. Hren, J. Liu, and J. Bruley,
“Characterization of amorphous carbon coated silicon field emitters”, J. Vac. Sci.
Technol. B 14(3). p. 2024, 1996.

E. I. Givargizov, V. V. Zhimov, A. V. Kuznetsov, and P. S. Plekhanov, “Cold
emission from the single-crystalline microparticle of diamond on a Si tip”, J. Vac.
Sci. Technol. B 14(3), p. 2030, 1996

V. V. Zhimov, A. B. Voronin, E. I. Givargizov, and A. L. Meshcheryakova,
“Emission stability and high current performance of diamond-coated Si emitters”, J.
Vac. Sci. Technol. B 14(3), p. 2034, 1996.

A. Lamouri, Y. Wang, G. T. Mearini, I. L. Krainsky, J. A. Dayton, Jr., and W.
Mueller, “Electron emission observations from as-grown and vacuum-coated
chmical vapor deposited diamond”, J. Vac. Sci. Technol. B 14(3), p. 2046, 1996.

W. B. Choi, J. Liu, M. T. McClure, A. F. Myers, V. V. Zhimov, J. J. Cuomo, and J. J.
Hren, “Field emission from diamond coated molybdenum field emitters”, J. Vac. Sci.
Technol. B 14(3), p. 2050, 1996.

N. Pupeter, A. Gohl, T. Habermann, E. Mahner, G. Muller, H. Piel, Ph. Niedermann,
and W. Hanni, “Field emission measurements with um resolution on chemical-
vapor-deposited ploycrystalline diamond films”, J. Vac. Sci. Technol. B 14(3), p.
2056, 1996.

M. W. Geis, J. C. Twichell, and T. M. Lyszczarz, “Diamond emitters fabrication and
theory”, J. Vac. Sci. Technol. B 14(3), p. 2060, 1996.

W. P. Kang, J. L Davidson, M. Howell, B. Bhuva, D. L. Kinser, D. V. Kems, Q. Li,
and J. F. Xu, “Micropattemed plycrystalline diamond field emitter vacuum diode

arrays”, J. Vac. Sci. Technol. B 14(3), p. 2068, 1996.

L. A. Chemozatonskii, Z. Ya. Kosakovskaya, Yu. V. Gulyaev, N. I. Sinitsyn, G. V.

45.

46.

47.

48.

49.

50.

51.

52.

53.

54.

55.

193

Torgashov, and Yu. F. Zakharchenko, “Influence of external factors on electron field
emission from thin-film nanofilaent carbon structures”, J. Vac. Sci. Technol. B 14(3),
p.2080,l996.

T. Xie, W. A. Mackie, and PR. Davis, “Field emission from ZrC films on SSi and
Mo single emitters and emitter arrays”, J. Vac. Sci. Technol. B 14(3), p. 2090, 1996.

E. I. Givargizov, “Silicon tips with diamond particles on them: New field emitters?”,
J. Vac. Sci. Technol. B 13(2), p. 414, 1995.

V. V. Zhimov, E. I. Givargizov, and P. S. Plekhanov, “Field emission from silicon
spikes with diamond coatings”, J. Vac. Sci. Technol. B 13(2), p. 418, 1995.

T. Asano, Y. Oobuchi, and S. Katsumata, “Field emission from ion-milled diamond
films on Si”, J. Vac. Sci. Technol. B 13(2), p.431, 1995.

Yu. V. Gulyaev, L. A. Chemozatonskii, Z. Ja. Kosakovskaja, N. I. Sinitsyn, G. V.
Torgashov, and Yu. F. Zakharchenko, “Field emitter arrays on nanotube carbon
structure films”, J. Vac. Sci. Technol. B 13(2), p. 435, 1995.

V. V. Zhimov and E. I. Givargizov, “Chemical vapor deposition and plasma-
enhanced chemical vapor deposition carbonization of silicon microtips”, J. Vac. Sci.
Technol. B 12(2). P633, 1994.

J. Liu, U. T. Son, A. N. Stepanova, K. N. Christensen, G. J. Wojak, E. I. Givagizov,
K. J. Bachmann, and J. J. Hren, “Modification of Si field emitter surfaces by
chemical conversion to SiC”, J. Vac. Sci. Technol. B 12(2), p. 717, 1994.

W. A. Mackie, R. L. Hartman, M. A. Anderson, and P. R. Davis, “Transition metal
carbides for use as field emission cathodes”, J. Vac. Sci. Technol. B 12(2), p. 722,
1994.

Z. H. Haung, P. H. Cuttler, N. M. Miskovsky, and T. E. Sullivan, “Calculation of
electron field emission from diamond surfaces”, J. Vac. Sci. Tech. B 13(2), p. 526,
1995.

P. Lerner, P. Cutler, and N. M. Miskovsky, “Theoretical analysis of field emission
from a metal diamond cold cathode emitter”, J. Vac. Sci. Technol. B 15(2), p. 337,
1997

R. Erz, W. Dotter, K. Jung and H. Ehrhardt, "Preparation of Smooth and
Nanocrystalline Diamond Films," Diamond and Related Materials, vol. 2, pp. 449-

 

56.

57.

58.

59.

60.

61.

62.

63.

65.

66.

194

453, 1993.

Y. Hayashi, W. Drawl, R. W. Collins, and R. Messier, "In-process Ellipsometric
Monitoring of Diamond Film Growth by Microwave Plasma Enhanced Chemical
Vapor Deposition," Appl. Phys. Lett., vol. 60, no. 23, pp. 2868-2870, 1992.

E. J. Bienk and S. S. Eskildsen, "The Effect of Surface Preparation on the Nucleation
of Diamond on Silicon," Diamond and Related Materials, vol. 2, pp. 432-437, 1993.

J. S. Ma, H. Kawarada, T. Yonehara, J. I. Suzuki, J. Wei, Y. Yokota and A. Hiraki,
"Selective Nucleation and Growth of Diamond Particles by Plasma-assisted
Chemical Vapor Deposition," Appl. Phys. Lett., vol. 55, no. 11, pp. 1071-1073, 1989.

M. P. Everson and M. A. Tamor, "Studies of Nucleation and Growth Morphology of
Boron-doped Diamond Microcrystals by Scanning Tunnelling Microscopy," J. Vac.
Sci. Technol. B, vol. 9, no. 3, pp. 1570-1575, 1991.

S. Iijima, Y. Aikawa, and K. Baba, "Early Formation of Chemical Vapor Deposition
Diamond Films," Appl. Phys. Lett., vol. 57, no. 25, pp. 2646-2648, 1990.

K. Hirabayashi, Y. Taniguchi, O. Takamatsu, T. Ikeda, K. Ikoma, and N. Iwasaki-
Kurihara, "Selective Deposition of Diamond Crystals by Chemical Vapor Deposition
Using a Tungsten-filament Method," Appl. Phys. Lett., vol. 53, no. 19, pp. 1815-
1817, 1988.

B. W. Sheldon, R. Csencsits, J. Rankin, and R. E. Boekenhauer, "Bias-enhanced
Nucleation of Diamond During Microwave-assisted Chemical Vapor Deposition," J.
Appl. Phys, vol. 75, no. 10, pp. 5001-5008, 1994.

BR. Stoner, G. H. Ma, S. D. Wolter, and J. T. Glass, "Characterization of Bias-
enhanced Nucleation of Diamond on Silicon by in vacuo Surface Analysis and
Transmission Electron Microscopy," Phys. Rev. B, vol. 45, pp. 11067-11084, 1992.

X. Jiang, C. P. Klages, R. Zachai, M. Hartweg, and H. J. Fusser, "Epitaial Diamond
Thin Films on (001) Silicon Substrates," Appl. Phys. Lett., vol. 62, pp. 3438-3440,
1993.

M. P. Everson and M. A. Tamor, "Investigation of Growth Rates and Morphology for
Diamond Growth by Chemical Vapor Deposition," J. Mater. Res., vol. 7, no. 6, pp.
1438-1443, 1992.

A. Masood, M. Aslam, M. A. Tamor and T. J. Potter, "Techniques for Patterning

 

67.

68.

69.

70.

71.

72.

73.

74.

75.

76.

77.

195

CVD Diamond Films on Non-diamond Substrates," J. Electrochem. Soc., vol. 138,
no. 11, Pp. L67-L68, 1991.

G. S. Yang, “Technology of CVD Diamond Thermistor and THin Film Heater”, Ph.
D. dissertation, Michigan State Univerisity, 1996.

M. Aslam, G. S. Yang, and A. Masood, "Boron-doped Vapor-deposited Diamond
Temperature Microsensors," Sensors and Actuators: A., vol. 45, no. 2, pp. 131-137,
1994.

M. W. Leksono and H. R. Shanks, "Patterning of Microwave Plasma Deposited
Diamond Films," Mat. Res. Soc. Symp. Proc., vol. 189, pp. 87-90, 1991.

T. Roppel, R. Ramesham, C. Ellis, and S. Y. Lee, "Thin Film Diamond
Microstructures," Thin Solid Films, vol. 212, pp. 56-62, 1992.

A. Masood, "Technology and Electronic Properties of Diamond Film Microsensors
for Thermal Signals," Ph. D. dissertation, Michigan State University, 1991.

R. N. Chakraborty, “Post-depositional processing of diamond film using electron
cyclotron resonance plasmas”, Ph.D. Dissertation, Michigan State University, 1995.

P. R. de la Houssaye, C. M. Penchina, C. A. Hewett, J. R. Zeidler, and R. G. Wilson,
"Hall Mobility and Carrier Concentration versus Temperature for Type IIa Natural
Insulating Diamond Doped with Boron by Ion Implantation," J. Appl. Phys., vol. 71,
no.7, pp. 3220-3224, 1992.

K. Bekku, Y. Mori, N. Eimori, H. Makita, A. Hatta, T. Ito, and A. Hiraki, "Growth
and Characterization of CVD Diamond Films Doped with Phosphorous," Advances
in New Diamond Science and Technology, Ed. S. Saito, N. Fujimori, O. Fukunaga,
M. Kamo, K. Kobashi and M. Yoshikawa, MYU, Tokyo, pp. 701-704, 1994.

L. J. Giling, Y. S. Wang, G. Z. Cao, G. J. Bauhuis, and P. F. A. Alkemade, "Diamond
Films Co-doped with Phosphorus and Nitrogen," Advances in New Diamond Science
and Technology, Ed. S. Saito, N. Fujimori, O. Fukunaga, M. Kamo, K. Kobashi and
M. Yoshikawa, MYU, Tokyo, pp. 359-362, 1994.

J. Bemholc, S. A. Kajihara, and A. Antonelli, "Theory of Doping of Diamond," Mat.
Res. Soc. Symp. Proc., vol. 242, pp. 323-333, 1992.

G. Z. Cao, F. A. J. M. Driessen, G. J. Bauhuis, L. J. Giling, and P. F. A. Alkemade,
"Homoepitaxial Diamond Films Codoped with Phosphorus and Nitrogen by

 

78.

79.

80.

81.

82.

83.

84.

85.

86.

196

Chemical Vapor Deposition," J. Appl. Phys., vol. 78, no. 5, pp. 3125-3131, 1995.

J. G. Ran, C. Q. Zheng, J. Ren, and S. M. Hong, "Properties and Texture of B-doped
Diamond Films as Thermal Sensor," Diamond and Related Materials, vol. 2, pp.
793-796, 1993.

G. Kawaguchi, J. Nakanishi, A. Otsuki, T. Oku, and M. Murakami, "Dependence of
Contact Resistance on Metal Electronegativity for B-doped Diamond Films," J.
Appl. Phys., vol. 75, no. 10, pp. 5165-5170, 1994.

B. R. Stoner, C. T. Kao, D. M. Malta, and R. C. Glass, "Hall Effect Measurements on
Boron-doped, Highly Oriented Diamond Films Grown on Silicon via Microwave
Plasma Chemical Vapor Deposition," Appl. Phys. Lett., vol. 62, pp. 2347-2349,
1993.

C. F. Chen and S. H. Chen, "Effects of Annealing on the Electrical Properties of
Boron—doped Diamond Films," Advances in New Diamond Science and Technology,
Ed. S. Saito, N. Fujimori, O. Fukunaga, M. Kamo, K. Kobashi and M. Yoshikawa,
MYU, Tokyo, pp. 709-712, 1994.

M. A. Plano, S. Zhao, C. F. Gardinier, M. I. Landstrass, D. R. Kania, H. Kagan, K.
K. Gan, R. Kass, L. S. Pan, S. Han, S. Schnetzer, and R. Stone, "Thickness

Dependence of the Electrical Characteristics of Chemical Vapor Deposited Diamond
Films," Appl. Phys. Lett., vol. 64, no. 2, pp. 193-195, 1994.

C. A. Hewett, M. J. Taylor, J. R. Zeidler, and M. W." Geis, "Spectfic Contact
Resistance Measurements of Ohmic Contacts to Semiconducting Diamond," J. Appl.
Phys., vol. 77, no. 2, pp. 755-760, 1995.

J. Nakanishi, A. Otsuki, T. Oku, O. Ishiwata, and M. Murakami, "Formation of
Ohmic Contacts to p-type Diamond Using Carbide Forming Metals," J. Appl. Phys.,
vol. 76, no. 4, pp. 2293-2298, 1994.

Y. Nishibayashi, N. Toda, H. Shiomi, and S. Shikata, "Thermally Stable Ohmic
Contact to Boron Doped Diamond Films," Advances in New Diamond Science and
Technology, Ed. S. Saito, N. Fujimori, O. Fukunaga, M. Kamo, K. Kobashi, and M.
Yoshikawa, MYU, Tokyo, pp. 717-720, 1994.

B. A. Fox, D. M. Malta, H. A. Wynands, S. R. Sahaida and J. A. von Windheim,
"High Temperature Electrical Measurements and Properties of Diamond,"

Proceeding of 2nd International High Temperature Electronics Conference, vol. 2,
pp. VI3-8, Charlotte, NC, 1994.

 

87.

88.

89.

90.

91.

92.

93.

94.

95.

96.

97.

98.

197

S. Matsumoto, Y. Sato, M. Tsutsumi and N. Setaka, "Growth of Diamond Particles
from Methane-hydrogen Gas," J. Materials Science, vol. 17, pp. 3106-3122, 1982.

S. Sahli, “Electronic Characterization and Fabrication of CVD diamond
Piezoresistive Pressure Sensors”, Ph.D. Dissertation, Michigan State University,
1997.

I. Watanabe, T. Matsushita, and K. Sasahara, “Low-Temperture Synthesis of

Diamond Films in Thermoassisted RF Plasma Chemical Vapor Deposition”, Jpn. J.
Appl. Phys. 31, 1428, 1992

Y. Shimada, N. Mutsukura, and Y. Machi, “Synthesis of diamond using Fe Catalysts
by RF plasma chemical vapor deposition method”, Jpn. J. Appl. Phys. 31 (Part 1,
6B), p. 1958, 1992

G. Amaratunga, A. Putnis, K. Clay, and W. Milne, “Crystalline diamond growth in
thin films deposited from a CH4/Ar rf plasma”, Appl. Phys. Lett. 55 (7), p. 634,
1989

R. A. Rudder, G. C Hudson, J. B. Posthill, R. E. Thomas, R. C. Hendry, D. P. Malta,
R. J. Markunas, T. P. Humphreys, and R. J. Nemanich, “Chemical vapor

depositionof diamond films from water vapor rf-plasma discharges”, Appl. Phys.
Lett. 60 (3). P. 329, 1992

Y. Namba, E. Heidrarpour, and M. Nakayama, "Size Effects Appearing in the Raman
Spectra of Polycrystalline Diamond," J. Appl. Phys., vol. 72, no. 5, pp. 1748-1751,
1992.

S. L. Flegler, J. W. Heckman, and K. L. Klomparens, Scanning and Transmission
Electron Microscopy, W. H. Freeman and Company, New York, 1993.

J. B. Bindell, "Scanning Electron Microscopy," Encyclopedia of Materials
Characterization, Ed. C. R. Brundle, C. A. Evans, Jr., and S. Wilson, pp. 70-84,
Butterworth-Heinemann, Boston, 1992.

A. Masood, M. Aslam, M. A. Tamor, and T. J. Potter, "Synthesis and Electrical
Characterization of Boron-doped Thin Diamond Films," Appl. Phys. Lett., vol. 61,
pp. 1832-1834, 1992.

D. Schulz, S. Shahzaman, S. Sahli, G. Yang and D. Hong.

D. Hong, M. Aslam, M. Feldmann, and M. Olinger, “Simulations of fabricated field

99.

100.

101.

102.

103.

104.

105.

106.

107.

108.

109.

198

emitter structures”, J. Vac. Sci. Tech. B 12(2), p. 764, 1994.

H. H. Busta, R. R. Shadduck, and and W. J. Orvis, IEEE trans. on EDs, Vol. 36, No.
11, 2679 (1989).

M. N. Yoder, Proc. of SPIE, vol. 2151, Diamond-Film Semiconductors, 72 ( 1994).

M. Aslam, P. Klimecky, G. P. Myers, H. H. Busta, B. J. Zimmerman, B. E. Artz,

.L.W. Cathey and R. E. Elder, “Triode characteristics and vacuum considerations of

evaporated silicon microdevices”, J. Vac. Sci. Tech. B 11(2), p. 422, 1993.

W. Zhu, G. P. Kochanski, S. Jin, and L. Seibles, “Defect-Enhanced Electron Field-
Emission from Chemical-Vapor-Deposited Diamond”, J. Appl. Phys. 78 (4), 2707
(1995).

W. Zhu, G. P. Kochanski, S. Jin, L. Seibles, D. C. Jacobson, M. McCormack, and A.
E. White, “Electron Field-Emission from Ion-Implanted Diamond”, Appl. Phys.
Lett. 67(8), 1157 (1995).

I. Taber, M. Aslam, M. A. Tamor, T. J. Potter, and R. C. Elder, "Piezoresistive
Microsensors Using P-type CVD Diamond Films," Sensors and Actuators: A, vol.
45, pp. 35-43, 1994.

G. S. Yang and M. Aslam, "Ultrahigh Nucleation Density for Growth of
Smooth Diamond Films," Appl. Phys. Lett., vol. 66, no. 3, pp. 31 1-313, 1995.

S. J. Kwon, Y.. H. Shin, D. M. Aslam, Y. B. Li, and J. D. Lee, “Low vonage emission
Characteristics of the undoped Polycrystalline-diamond field Emitter by MPCVD”,
accepted for the presentation at the 10th Int. Vacuum Microelectronics Conf.,
Kyungju, Korea, Aug. 1997.

C. Xie, N. Kumar, C. C. Collins, T. J. Lee, H. Schmidt, and S. Wagal, “Electron
Field Emission from Amorphic Diamond Thin Films”, Tech. Dig. of the 6th Int.
Vacuum Microelectronics Conference, 162 (1993).

P. Baumann, S. P. Bozeman, B. L. Ward, and R. J. Nemanich, “Characterization of
Metal-Diamond Interfaces: Electron Affinity and Schottky Barrier Height”,
presented at Diamond ’96 Conference, Troy, France, Sep. 1996.

H. C. Lee and R. S. Huang, “Simulation and Design of Field Emitter Array”, IEEE
Electron Dev. Lett., vol. 11, 579 (1990).

110.

111.

112.

113.

114.

115.

116.

117.

118.

119.

120.

121.

199

D. Hong and D. M. Aslam, “Simulation study of microtip field emitter arrays in
triode configuration”, accepted for the presentation at the 10th Int. Vacuum
Microelectronics Conf., Kyungju, Korea, Aug. 1997.

M. Geis, J. C. Twichell, J. Macaulay, K. Okano, “Electron Field-Emission from
Diamond and Other Carbon Materials after H2, 02 and Cs Treatment”, Appl. Phys.

Lett. 67 (9), 1328 (1995).

D. Hong and D. M. Aslam, “Technology and Characterization of Diamond Field
Emitter Structures”, accepted to IEEE Elec. Dev., 1997.

D. Hong, D. M. Aslam, T. Grimm, L. Garbini, and S. Bandy, “Field Emission from
Carbon Implanted Polycrystalline Diamond Film”, accepted for the presentation at
the 10th Int. Vacuum Microelectronics Conf., Kyungju, Korea, Aug. 1997.

O. Groning, O. M. Kuttel, E. Schaller, P. Groning, and L. Schlapbach, “Vacuum arc
discharges preceding high electron field emission from carbon films”, Appl. Phys.
Lett. 69 (4), P. 476, 1996.

D. Hong and M. Aslam, “Diamond Field Emitter Triode Display Cells”, Proceeding
of 8th Int. Vacuum Microelectronics Conf., p. , 1995.

Pixel International, 7th Int. Vacuum Microelectronics Conf., Grenoble, France, July
4-7, 1994.

H. H. Busta, B. J. Zimmerman, M.C. Tringides, C. A. Spindt, IEEE Trans. Electron
Devices, ED-38, p. 2558, 1991.

H. C. Lee and R. S. Huang, “A theoretical study on field emission array for
microsensors”, IEEE Trans. Electron Devices, ED-39, p. 313-24, 1992.

D. Hong and M. Aslam, “Diamond Field Emitter Pressure Sensor”, Proceeding of
8th Int. Vacuum. Microelectronics Conf., p., 1995.

S. Zum, Q. Mei, C. Ye, T. Tamagawa, and D. L. Polla, “Sealed vacuum electronic
devices by surface micromachining”, Proc. IEDM 91 (Washington DC.) (New York:
IEEE), P. 205-8, 1991.

G. P. Myers, M. Aslam, P. Klimecky, L.W. Cathey, R. E. Elder, and B. E. Artz, J.
Vac. Sci. Tech. B 11(3), 642 (1993).

"I111111111111111111111111“