33$ .. .d
1 1 .i \ t t.

1.... .531...
. gal-X!- n

1135123.

 
 

 

{1; .3
. 1.1. .

 

:.
Lia“.

 

.: .3 L.
1.2. pair’s; :4.

. .
..,.. 5.:
. run In

 

£52... n.
I! . xi 3.

I.

a: g... .
mean??? u .‘ .Z

L

31.

,3
. .3.
2. as...

.

.

magma

fiflg.

2m ‘
m; ..
. vmmflua-

‘
»

 

 

w.

, a3. wag...

 

51:53:;

far».

9

I...) .5“.

ta}.
1-1.» 1

.2 . 13...... l.

Ix 5:11.
: .3 : s:

(5111...:

saliva-Lari}.
at
.2.

aui~ :
-\ i.\ :v

 

 

 

If. 53!.

.9.

 

1.43:.

“1.2!: .r.
”fin. 6......

4

st

.

mdfix

i.
.

 

 

....
Inc ”I

if?“ ‘. , ..
.. 3,. Man w“?
.3.
fig”.

3

.u
l...-
I

 

.. “am?
.. y a.
up? mm Wlfi

It!!!
a

n in
s I:
s u

iii

0.0%“...

.1

.4}:
fix:

, «can;

t .11.

.3...“
2:3

9..
he. «

.Ifi

53;.

1.5111. 3‘

 

I...

3 A»: :

1..

e
. {.15.
r a...

t:

.
5.7. 31.3.. 1.311%.

.
.
3 .

 

 

i!

a
«I... PI‘! 3

.2. E. i A
I I,

an;

.: v3. 2&2“.

.
1.

.3. 5 3.
ishflrvhics

V!

.4
‘I17.

5::- .3.

i._)...i..uo

\iltfjhpt

E:
l i’\‘.

av...

, 3m...)

7w: 5.0

:73 l.

foray»?
5"?“«4 .w h. 340%..

711
3|

.
as. 3
A 3be 1.14.
w .Is....W.ru.J
.

,.

., .333”?

.5..é....2 .

1

1.0.? .3473:
./ 2: {ul
.1135}: .
‘13..

it...

flan-{01' {.1
Btu. 9.1

Wait-.1:

v :10! .4. .ill

29:. 1 33.31:.
#.s 572:.

«A 5:: tv

 

.1!!! >

      

2 LIBRARY
1 o 04 Michigan State
I University

This is to certify that the
dissertation entitled

FABRICATION AND SITE SPECIFIC GROWTH OF
NANOWIRES

presented by

CHUN-I WU

has been accepted towards fulfillment
of the requirements for the

PhD. degree in Electrical @ineerinL

 

 

 

R J ———-a
/ avstlcmw

Major Profe'ssot‘s Signature

4/14/04?

Date

MSU is an Affirmative Action/Equal Opportunity Institution

 

—.-.-.--ct

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

 

 

DATE DUE DATE DUE DATE DUE
. r o 2008
JUN 2‘ 2007 M3113 n R
ucUelTileUVV

 

 

 

 

 

 

 

 

 

 

 

 

 

2/05 p:/ClRC/DateDue.indd-p.1

FABRICATION AND SITE SPECIFIC GROWTH OF NANOWIRES
By

Chun-I Wu

A DISSERTATION

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

DOCTOR OF PHILOSOPHY
Department of Electrical and Computer Engineering

2006

ABSTRACT
FABRICATION AND SITE SPECIFIC GROWTH OF
NANOWIRES
By

Chun-I Wu

Nanotechnology has the promise to affect many areas of study including
structural, materials, biological, and electrical sensors and circuits. In this
presentation, fabrication routes to growing whiskers and nanowires of various
materials will be presented. The materials investigated include gallium arsenide
(GaAs), indium antimonide (InSb), a clathrate material SI'sGaléGC30, and germanium
(Ge). The experiments were designed to utilize the vapor-liquid-solid (VLS) growth
mechanism, where a metal catalyst is used to mitigate the grth of the nanowires.
The results for germanium, however, show a strong enhancement of growth upon the
exposure of the samples to air while at the growth temperature. The role of oxide
assisted (0A) growth in combination with the VLS mechanism will be discussed.

Nanowires of GaAs, InSb, and SrgGamGew showed compositions that were
deficient in the more volatile elements. Volume confinement studies will also be

presented for the clathrate material.

Growth of germanium dioxide (GeOz) whiskers and nanowires was found to be
highly selective to the locations of the metal catalyst (gold). Examples of site specific
growth of the GeOz wires will be presented to Show nanobridges on silicon substrates,
and nanowire bundles inside tapered micropipette structures. The grth technique is
relatively simple, and applicable for scaleup procedures. Images of the whiskers and

nanowires were taken with SEM and TEM.

Copyright by
Chun-I Wu
2006

To my parents and my sisters.

ACKNOWLEDGEMENTS

I would like to express my sincere gratitude to my advisor, Dr. Tim Hogan. This
work won’t have been possible without his guidance, constant encouragement and
continuing support. Besides academics, his goodness, dedication, honesty and equal
respect for everyone have deeply moved me. I would also like to thank my committee
members, Dr. Ayres, Dr. Grotjohn, and Dr. Fan for their valuable comments on the
project.

Finally I would like to express my gratitude to my parents and family members

for their patience, unconditional support and confidence in my abilities.

vi

TABLE OF CONTENTS

LIST OF TABLES ............................................................................. ix

LIST OF FIGURES ............................................................................. x

Introduction ...................................................................................... 1
Chapter 1 Theory

1.1 Different growth method ................................................................... 6

1.1.1 Chemical vapor deposition ......................................................... 6

1.1.2 Physical Vapor Deposition .......................................................... 9

1.2 Pulsed Laser Deposition ................................................................... 25

1.2.1 History and Fundamental of pulsed laser deposition .......................... 25

1.2.2 Pulsed Laser Deposition: Mechanisms of Ablation ............................ 27

1.2.2.1 Collisional ........................................................................ 31

1.2.2.2 Thermal Evaporation .......................................................... 32

1.2.2.3 Electronic sputtering ........................................................... 33

1.2.2.4 Exfoilation ....................................................................... 34

1.2.2.5 Hydrodynamic Sputtering .................................................... 35

1.2.3 Particulates Generate By Pulsed Laser Deposition ........................... 38

1.3 Thermodynamics and Chemical potential .............................................. 43

1.3.1 Binary solution ...................................................................... 46

1.3.1.1 The Gibbs Free Energy of Binary Solutions ................................. 47

1.3.2 Chemical potential ................................................................. 48

1.4 Growth mechanism ....................................................................... 50

1.4.1 Vapor-Liquid-Solid (V LS) grth mechanism ............................... 50

1.4.2 Solid-liquid-solid ................................................................... 57

1.4.3 Oxide-Assisted Growth ........................................................... 58

1.5 Vacuum Sciences .......................................................................... 60

1.5.1 Rate at Which Molecules Strikes a Surface ........... . ......................... 62

1.5.2 Ideal Gas Law Pressure ............................................................. 65

1.6 Gas Flow and Conductance ............................................................... 71

Chapter 2 Growth of III-V Semiconductor Nanowires (GaAs)

2.1 Introduction ................................................................................. 73
2.2 Experiments ................................................................................. 74
2.3 Discussion .................................................................................... 99
2.4 Conclusion ................................................................................. 102

Chapter 3 Growing nanowires from SrgGa16Ge3o
3.1 Introduction .............................................................................. 103
3.2 Experiments ............................................................................. 104

vii

3.3 Discussion ................................................................................. 123

Chapter 4 Growth nanowires from germanium

4.1 Introduction ........................ 125
4.2 Experiments ............................................................................... 127
4.3 Theory ...................................................................................... 155
4.4 Discussion ................................................................................. 157
4.5 Conclusion ................................................................................ 157
Chapter 5 Gold Pattern experiment
5.1 Introduction ............................................................................... 159
5.2 Experiment ................................................................................ 160
5.3 Gold Deposited at Room Temperature ................................................ 160
5.4 Gold Deposited at High Temperature ................................................... 165
5.5 Conclusion ................................................................................ 170
Chapter 6 Nanobridge
6.1 Introduction ................................................................................. 172
6.2 Experiments ................................................................................ 174
6.3 Conclusion .................................................................................. 189
Chapter 7 Conclusions ..................................................................... 190
Chapter 8 Scanning Electron Microscope .......................................... 193
Chapter 9 System Set Up .................................................................. 207
Appendix A ...................................................................................... 213
Appendix B ...................................................................................... 217
Appendix C .............................................................. l ........................ 218
Appendix D ...................................................................................... 219
References ........................................................................................ 223

viii

LIST OF TABLES

Chapter 1

Table 1. Qualitative comparison of deposition rates of various growth techniques [31] 8

Table 2. Experimental values for the binding energy, Edes, and the activation energy for
diffusion, E s, for some typical systems .............................................................. 13

Table 3. Temperature of metling, Boiling, and a Vaporization Rate of l nm/pulse for
Selected Solids. [31] .......................................................................................... 33

Table 4. Quantities needed for evaluating thermal stress by means of Equation 5 [31]... 34

Table 5. The different shape and size between the three phases ....................................... 39
Appendix

Table 6. Properties of quartz tube ................................................................................... 213
Table 7. Specifications for Lamda Physik 210i KrF Excimer Laser .............................. 214
Table 8. Beam box specifications .................................................... l ............................... 214
Table 9. Specifications for the optical window .............................................................. 215
Table 10. Specification for the plano-convex lens .......................................................... 215
Table 11. Specification for the high power UV laser mirror .......................................... 216
Table 12. Specifications for the aluminum optics stand ................................................. 216
Table 13. Calculated Concentrations and Associated Etch Rates ................................... 220

ix

LIST OF FIGURES

Chapter 1

Figure 1. Chemical vapor deposition process ..................................................................... 7
Figure 2. Schemiatic of film surface with various edges, pits, and clusters identified ..... 10

Figure 3. Free-energy change (A G) as a function of cluster (r‘ > r) or stable nucleus
(r > r.) size. r. is critical nucleus size, and A G. is critical free-energy barrier
for nucleation. .................................................................................................... 15

Figure 4. Total free energy (surface plus volume free energies) of cluster formation vs.
size (r) for different supersaturations (S). .......................................................... 17

Figure 5. Schematic representation of three film growth modes. 6 is the overlayer
coverage in monolayers (ML). ........................................................................... 18

Figure 6. Schematic of basic atomistic processes on substrate surface during vapor
deposition. .......................................................................................................... 19

Figure 7. Schematic representation of MBE deposition of a thin film. The film thickness
and composition is defined by controlling the individual fluxes. ...................... 21

Figure 8. A schematic representation of the sputtering process. ...................................... 23

Figure 9. Formation of cone structures during laser ablation, indicating a collisional
interaction of particles with the target surface. .................................................. 31

Figure 10. Droplet formation in the gold target. Sputtering of polycrystalline Au in air
after 4320 pulses from an ArF excimer laser (193nm) of 12 ns and 2.5J/cm2.

(From Kelly et a1. 1985) .................................................................................. 35
Figure 11. Plasma gas-dynamic expansion [31]. .............................................................. 36
Figure 12. Laser produced plasma plume from a metal target [32]. ................................. 37

Figure 13. Variation of Gibbs free energy with the arrangement of atoms. Schematically,
the configuration GZ has the lowest free energy which results in a system that
is at stable equilibrium. The configuration G1 is a metastable equilibrium. .45

Figure 14. Variation of G1 (the free energy before mixing) with alloy composition. ...... 47

Figure 15. Phase diagram for the gold-silicon system. ..................................................... 52

Figure 16. Gold droplets form an alloy with the silicon. The alloy has a eutectic point

near 360°C, above which it is molten. ............................................................ 55

Figure 17. All directions of molecular velocities are consistently probable. .................... 63

Figure 18. A vector always represents the direction and the magnitude of its velocity. .. 63

Chapter 2
Figure 19. Nanowires grown from a piece of GaAs. The scale bar is 500 nm. ............... 75
Figure 20. The nanowires grew from the GaAs chunk. The scale bar is 511m ................. 76
Figure 21. These nanowires were a few microns in length. The scale bar is 2 pm. ......... 77
Figure 22. The nanowires grew from the edge. The scale bar is 2pm. ............................ 78
Figure 23. Histogram showing the diameter of nanowires. Total sample size: 42.
Average Diameter: 65 nm .............................................................................. 79
Figure 24. Histogram showing the height distribution of nanowires. Total sample size :
42. Average length of the nanowire is 2.7 um. .............................................. 80
Figure 25. The EDS analysis of nanowire ....................................................................... 81
Figure 26. Under these fabrication conditions, a very high density of nanowires formed.
The scale bar is 20 um. ................................................................................... 83
Figure 27. The diameters of these nanowires were uniform. The scale bar is 5 pm. ...... 84
Figure 28. This zoom-out image illustrates that nanowires connected with each other in
order to form a wool-like structure. Nonetheless, there was no gallium
arsenide chunk lefi inside this structure. ......................................................... 84
Figure 29. The nanowires formed a thick layer on top of the GaAs chunk. .................... 85
Figure 30. The SEM image shows nanowires. In addition, the same image illustrates that

Figure 31.

Figure 32.

the gallium arsenide surface was underneath the layer of nanowires and there
were some nanowires left on the surface. The length of nanowires was tens of

microns. The scale bar is 5 pm. ...................................................................... 86
EDS analysis on the grown gallium oxide nanowires. ................................... 87
Histogram showing the diameter distribution of the nanowires. The average

diameter 18 81 nm. ........................................................................................... 88

xi

Figure 33. The SEM image shows the nanowires grew on the piece of gallium arsenide.
These nanowires were a few microns in length. The scale bar is 10 um. ...... 89
Figure 34. The SEM image shows the nanowires grew on top of gallium arsenide chunk.
......................................................................................................................... 90
Figure 35. The SEM image show the nanowires grew on top of gallium arsenide chunk.
The scale bar is 2 pm. ..................................................................................... 91
Figure 36. Histogram showing the diameter of nanowires. Total sample size: 40. ......... 92
Figure 37. EDS analysis on the grown gallium oxide nanowires .................................... 93
Figure 38. The SEM image of a piece of GaAs. No gold was deposited on this sample
before growth. ................................................................................................. 94
Figure 39. TEM image shows nanowire grown from GaAs ............................................ 95
Figure 40. TEM image shows nanowire grown from GaAs ............................................ 96
Figure 41. TEM image shows nanowire grown from GaAs ............................................ 97
Figure 42. EDS analysis of nanowires ............................................................................. 98
Chapter 3
Figure 43. The SrgGa15Ge30 structure. In the framework, the black atoms occupy the 6c

Figure 44.

Figure 45.

Figure 46.

Figure 47.

Figure 48.

Figure 49.

sites, the light gray atoms occupy the 161' sites, and the dark gray atoms
occupy the 24k sites. In the cavities, the gray atoms are at the Sr(1) 2a sites

and the black atoms are at the Sr(2) 24k sites [51]. ....................................... 103
The SEM image shows that nanowires were scattered on the surface of
SrgGamGe3o powders. The scale bar is 5 pm. ............................................. 105
The SEM image shows that the dark region is a quartz .plate. The scale bar is
2 um. .............................. 106
The SEM image shows that nanowires were mostly scattered on the top
surface of the SrgGa16Ge30 powder. The scale bar is 5 11m. ........................ 106
Histogram showing the diameter of nanowires. Total sample size: 22.
Average diameter is 52 nm. .......................................................................... 107
EDS analysis on the nanowires grew on SrgGaleGe3o ................................. 108
The SEM image shows the nanowires around the SrgGa16Ge3o powder. 110

xii

Figure 50. The nanowires scattered around the SrgGa16Ge3o powders. ........................ 110
Figure 51. The SEM image shows the distinctive structure of the nanowires. .............. 111
Figure 52. The average diameter was 40 nm and the average length was 1.5 um. ....... 112
Figure 53. The average diameter of nanowires was about 100 nm. ............................... l 12
Figure 54. Histogram showing the diameter of nanowires. Total sample size: 36. The
average diameter is 62.22 nm. ...................................................................... 113
Figure 55. EDS analysis on the grown nanowires ......................................................... 114
Figure 56. The length of nanowire was over 4 pm . Additionally, some nanowires were
observed to form as a ball shape in the tip. ................................................... 116
Figure 57. The diameter of nanowires ranged from 40 nm to 200 nm. The scale bar is
500 nm. ......................................................................................................... 116
Figure 58. Nanowires on top of the quartz plate. The scale bar is 2 pm. ..................... 117
Figure 59. The dark region is the quartz plate. The scale bar is 2 11m. ......................... 118
Figure 60. The scale bar is 10 um. ................................................................................. 118
Figure 61. Histogram showing the diameter of nanowires. Total sample size: 36.
Average diameter 43 nm. .............................................................................. 119
Figure 62. The diameter was about 60 nm and the length was about 1.5 pm ................ 120
Figure 63. The diameter was about 40 nm. .................................................................... 120
Figure 64. The density of these nanowires was unifome distributed over the quartz
substrate. ....................................................................................................... 121
Figure 65. The length was about 2 pm and the diameter was about 120 nm ................. 122
Figure 66. The scale bar is 20m. .................................................................................. 122
Figure 67. Histogram showing the diameter of nanowires. Total sample size: 42. Average
diameter is 55.24 nm ..................................................................................... 123
Chapter 4
Figure 68. The experimental setup, for investigating the growth of GeOz nanowires. . 128

Figure 69. The SEM image shows the edge of the gold coin. The scale bar is 200 pm. 129

xiii

Figure 70.

Figure 71.
Figure 72.
Figure 73.
Figure 74.
Figure 75.

Figure 76.

Figure 77.

Figure 78.
Figure 79.
Figure 80.

Figure 81.

Figure 82.

Figure 83.

Figure 84.
Figure 85.
Figure 86.
Figure 87.
Figure 88.

Figure 89.

The SEM image of nanowires on the surface of gold coin. The scale bar is 10

um. ................................................................................................................ 130
The SEM image shows the whiskers at the edge. The scale bar is 2 pm. 131
The whiskers with kinked shape. The scale bar is 2 11m. ............................ 131
Away from the observed area, no nanowires were observed. ...................... 132
EDS analysis on the nanowires. .................................................................. 133
The silicon wafer was covered with dense germanium oxide nanowires ..... 135
The growth of nanowires on the edge of a silicon wafer. The scale bar is
10 um. ........................................................................................................... 136
The gold balls were appeared at the tip of nanowires. The scale bar is 5 1111.37
The image of a nanowire at a higher magnification (1 pm scalebar). .......... 138
EDS analysis on the nanowires. ................................................................... 139
The silicon wafer only partially covered by nanowires (1mm scalebar) ...... 141
A closeup image of a scratched region showing that the grth of nanowires
only occured in the gold coated areas. The scalebar is 10 um ..................... 141
The dense of nanowires were observed on the surface of silicon wafer. The
scale bar is 5 um. .......................................................................................... 142
Histogram showing the diameter of nanowires. Total sample size: 20. The
average diameter is 285 nm. ......................................................................... 143
EDS analysis on the grown nanowires. ........................................................ 144
The image of the sample set. ........................................................................ 146
No growth happened in the “S” (1mm scale bar). ........................................ 147
Dark region is no gold zone. The scale bar is 20 um. ................................. 148
Nanowires extending into no gold zone. The scale bar is 5 pm. ................. 148
The nanowires had a bright dot at tip (1 pm scalebar). ................................. 149

xiv

Figure 90. Dense nanowires grown onto the surface (5 pm scalebar). .......................... 150
Figure 91. Higher magnification SEM image of nanowires (2 pm scalebar) ................ 150
Figure 92. SEM EDS analysis on the nanowires. .......................................................... 151
Figure 93. TEM image shows the gold ball at the tip of the nanowires (0.2 pm scalebar)..
152
Figure 94. TEM image shows the nanowires have diameter about 40 nm (0.2 pm
scalebar) ........................................................................................................ 153
Figure 95. TEM image shows the nanowires have diameter about 65 nm (20 nm
scalebar) ........................................................................................................ 153
Figure 96. TEM image shows the gold ball at tip of nanowires (20 nm scalebar). ....... 154
Figure 97. Selected area electron diffraction (SAED) pattern of nanowires. ................ 154
Chapter 5
Figure 99. The darker location is Si and the lighter locations are SiOz. The scale bar is 20
um. ................................................................................................................ 161
Figure 100. The area identified in this image is near the edge of the half-circle pattern.
The scale bar is 2 pm. ................................................................................ 161
Figure 101. The image is near the edge of the half-circle pattern. The darker location is
Si and the lighter locations are SD; The scale bar is 2 11m. .................... 162
Figure 102. The image is inside the pattern. The background is Si. and the white dots are
Au. The scale bar is 1 um. ......................................................................... 162
Figure 103. The dark part is Si and the light ball 1S Au. This area is inside the pattern.
The scale bar 18 10 um. .............................................................................. 162
Figure 104. This is a half-circle pattern. The white chunk is silicon. The scale bar is 20
um. ............................................................................................................. 162
Figure 105. The average diameter of the Au balls was 100 nm. The scale bar is 1 pm. 163
Figure 106. This is a half-circle pattern. The scale bar is 20 um ................................... 163
Figure 107. The darker location is Si and the lighter locations are SiOz ....................... 163
Figure 108. 10.5 nm of sputter deposited Au followed by an annealing at 700° C for 5

mins. The scale bar is 10 um. .................................................................... 163

XV

Figure 109.

Figure 110.

Figure 111.

Figure l 12.

Figure 113.

Figure 114.
Figure 115.
Figure 116.
Figure 117.

Figure 118.

Figure 1 19.

Figure 120.
Figure 121.

Figure 122.

Figure 123.
Figure 124.

Figure 125.

The image is near the edge of the half-circle pattern. The scale bar rs 2

um....... .. ................................................................................................. .164
The image is near the edge of the half-circle pattern. The scale bar is 2
11m......... .................................................................................................. 164

The image is near the edge of the pattern. The dark location is Si and white

dots are Au. The scale bar is 2 pm. ............................................................ 164
The dark location is Si. The scale bar is 5 pm ............................................ 164
The Au ball was observed to be in the gap as a bridge. The scale bar is 10
um. ............................................................................................................. 165
The area in this image is in the edge of pattern. The scale bar is 10 pm... 165
The SEM image shows the square pattern. The scale bar is 50 um. ......... 166
The SEM image shows the edge of pattern. The scale bar is 2 um ............. 166
The Au ball was found outside. .................................................................. 166
The SEM image shows gold ball scattered outside the pattern. The scale bar
is 50 um ...................................................................................................... 166
This gap is not big enough to have a big Au ball produced in the gap. The
scale bar is 2 pm. ....................................................................................... 167
The Au balls are inside the pattern. The scale bar is 10 um ....................... 167
The image is near the edge of the pattern. The scale bar is 20 um. ........... 167

The temperature was raised to 700° C and Au was deposited by pulsed laser
deposition (PLD) for 1 minute. The substrate was then maintained at 700° C
for 5 minutes before reducing power to the tube furnace. The shape of the
Au ball looks like a bridge. The scale bar is 2 pm ..................................... 168

The image is taken near the edge of the pattern. The scale bar is 5 pm ..... 168
The image is taken near the edge of the pattern. The scale bar is 2 um ..... 168
The temperature was raised to 700° C and Au was deposited by pulsed laser

deposition (PLD) for 1 minute and reducing power immediately. The scale
bar is 20 um. .............................................................................................. 169

xvi

Figure 126.

The temperature was raised to 700° C and Au was deposited by pulsed laser
deposition (PLD) for 1 minute and reducing power immediately. ............ 169

Figure 127. The mask designed for patterning the Si02 layer using a positive photoresist

(Shipley 1813). ........................................................................................... 170
Chapter 6
Figure 128. Schematic diagrams of (a) metal catalyst and (b) lateral nanowire growth.
..................................................................................... 174
Figure 129. The nanowires grew across the trench. The scale bar is 50 um. ............... 175

Figure 130.
Figure 131.

Figure 132.

Figure 133.
Figure 134.
Figure 135.
Figure 136.
Figure 137.
Figure 138.
Figure 139.
Figure 140.
Figure 141.
Figure 142.
Figure 143.
Figure 144.

Figure 145.

Nanowires grew across the 70 um wide trench. The scale bar is 50 um .. 176

Dense nanowires grew across the trench. The scale bar is 50 um. ............ 177
The whiskers with large diameters grew on the side wall. The scale bar is 50
um. ............................................................................................................. 178
The scale bar is 50 um. ............................................................................... 178
The scale bar is 50 um. ............................................................................... 178
The nanowires grew in the region with different widths. ........................... 179
Nanowires inside the trench. The scale bar is 50 um. ................................ 180
Nanowires inside the trench. The scale bar is 50 um. .' ............................... 180
Nanowires grew along the comer. .............................................................. 180
The growth of nanowires in the trench. ...................................................... 182
The SEM image shows nanowires grew across the trench ......................... 183
The nanowires grew across the 80 wide trench. ......................................... 184
The scale bar is 20 um. ............................................................................... 184
The scale bar is 20 um. ............................................................................... 184
The growth of nanowires in the trench. ...................................................... 185
The scale bar is 20 um. ............................................................................... 185

xvii

Figure 146. In the SEM image of whiskers grew across the trench. The scale bar is

20 um. ........................................................................................................ 186
Figure 147. The scale bar is 10 um. ............................................................................... 187
Figure 148. The scale bar is 10 um. ............................................................................... 187
Figure 149. The scale bar is 20 um. ............................................................................... 187
Figure 150. The scale bar is 20 um. ............................................................................... 187
Figure 151. The scale bar is 20 um. ............................................................................... 187
Figure 152. The scale bar is 20 um. ............................................................................... 188
Chapter 8
Figure 153 Schematic of an SEM ................................................................................. 194
Figure 154. Interaction-volume variation with accelerating voltage. ............................ 200
Figure 155. Secondary electron escape, edge effects, and image production ................ 201
Figure 156. Schematic of an EDS detector attached to an SEM. .................................. 205
Figure 157. Quartz tube furnace system ........................................................................ 208
Figure 158. Autosketch of optical-side drawing ............................................................ 209
Figure 159. Autosketch of endcap ................................................................................. 210
Figure 160. Autosketch of endcap ................................................................................. 211
Figure 161. Autosketch of endcaps assembled .............................................................. 212

xviii

Introduction

Following Richard Feynman’s best known comment “There is plenty of room at
the bottom [1],” emerged a series of dominant interest in exploring subjects concerning
reduced dimensional structures, such as microelectromechanical system (MEMS),
quantum wells, nanowires, nanotubes, and nanodots [2 - 4 ]. Among these topics,
semiconductor nanowires are deemed to be an innovation by means of the application of
the Moore’s law and the capabilities of manufacturing ever smaller metal oxide
semiconductor field effect transistors (MOSFETs). With the ongoing evolution of
reducing the length of the channel, it appears that corresponding technological advances
should have been developed to support usable threshold voltages. In other words, the
desired advances should be able to sustain the increased coverage of the channel surface
by the gate through structures such as the Double Gate FET, or F inF ET [5]. In particular,
the fabrication of nanowires by self assembly deserves more attention with the
improvement of semiconductor nanowires fabrication techniques [6 , 4]. In general,
applications of the nanowires structures of interest encompass field effect devices,

mechanically resonant structures, and sensors.

Nanotechnology

Nanotechnology deals with various structures of matter having dimensions of the
order of a billionth of a meter. A recent definition of nanotechnology is
“Nanotechnolgoy is the understanding and control of matter at dimensions of roughly 1

to 100 nanometers [7].” This dimension range is especially drawing attention where

interdisciplinary fields of research are emerging [8,9,10]. The nanoscale is unique
because it is the size scale where the commonplace properties of materials, including
conductivity, hardness, or melting point, and the more exotic properties of the atomic and
molecular world, such as wave-particle duality and quantum effects converge [11]. On
the nanoscale, the most fundamental properties of materials and machines are closely
related to their size, the condition that is not observed if on other scales. If the size
reduces to or smaller than de broglie wavelength, the property changes. For example,
after calculation we can get the de broglie wavelength of germanium is 17.98 nm. That
means once the size of germanium smaller than 17.98 nm, the material properties will

different from other scale.

One-dimensional nanostructure

The reason why semiconductor nanowires and nanotubes, one-dimensional
structures with diameters not exceeding a few hundred nanometers, can exhibit a
breakthrough in terms of electronic and optical properties is related to their unique
structural one-dimensionality and the effects of quantum confinement of carriers in two
dimensions [12]. Sitting on the advantage of the selection of various compositions and
band structures, the one-dimensional semiconductor nanostructures are deemed to be the
essential components for constructing likely diverse applications of nanoscale devices.
Considering this feature of one-dimensional nanostructures, current research has mainly
paid attention to a logical synthetic control of one-dimensional nanoscale building blocks,
the new characterization of properties, and the fabrication of devices would be taken into

account by the use of nanowire building blocks, and by integrating nanowire elements

into architectures that are complicated in functions [13]. Corresponding with the topic of

this dissertation, nanowires become the focus in the following discussion.

Application

Having specific and distinguished characteristics, such as the small sizes and high
surface-to-volume ratios, one-dimensional nanostructures have commanded great
attention by their useful and newsworthy properties. The control of the surfaces and
internal interfaces also equip one-dimensional nanostructures with unique stability effects.
Certain fundamental limitations pose problems for the realization of this effort. Size-
dependent thermal conductivity in nanostructures is one of them. Under the influence of
increased boundary scattering and lessened phonon group velocities resulting fi'om the
phonon confinement, the phonon transport is expected to be blocked in thin 1D
nanostructures. Nonetheless, if employing an alternative perspective, poor heat transport

is useful for thermoelectric materials, which are characterized by a figure of merit :ZT =

aZT/[pOcp + 1(6)], where a, T, p, xp and K6 represent the Seebeck coefficient, absolute

temperature, electronic resistivity, lattice thermal conductivity and electronic thermal
conductivity, respectively. It is found that as the proceeding of the phonon transport
worsens, the value of the figure would improve. In fact, the Dresselhaus research group
has predicted that ZT can be increased above bulk values in thin nanowires as long as the
diameters, compositions and carrier concentrations are carefully manipulated [14, 15].
This contention has been confirmed by different levels of experiments.

Various types of nanowires have been fabricated using different methods, including

evaporation [16], laser ablation [17], chemical vapor deposition (CVD), and electron

beam lithography [18]. Moreover, nanowires that have been the target of studies range
from silicon, germanium, III-V material, to II-VI material.

In the last decade, silicon nanowires have attracted plenty of research interests
owing to the fact that silicon has certain attributes useful in applications of nanodevices.
A couple of methods have been proved to be viable growth mechanisms. Metal catalyst-
assisted and the oxide assisted (0A) methods are evaluated as the most accepted ones
among the many methods discussed. Metal catalyst-assisted growth mechanism
functions in connection with the widely known vapor-liquid-solid (V LS) mechanism [6].
Technically, the silicon vapor source in metal catalyst-assisted mechanism is obtained
through silane which is created by using CVD method or by laser ablating the metal
catalyst in physical vapor deposition (PVD). Without using the metal catalyst, the oxide-
assisted method is making use of the metal oxide powder to grow nanowires.

Currently, there is more and more research focused on the silicon material as well
as on III-V and II-VI material. III-V and II-VI semiconductor nanowires are grown by
means of the laser-assisted catalystic growth (LCG) method [19]. In addition, the
nanowire products are collected at the down-stream cold end of the furnace.

One challenge that needs to be addressed is how to grow nanowires in a way that is
appropriate to the device fabrication from a manufacturing perspective. The
exceptionally high surface-to-volume ratio of thin one-dimensional nanostructures has
enabled them to function well with inherently high sensitivity and short response time.
A common sensing device can be produced by configuring high performance, field-effect
nanowire transistors, in which the specific function of sensing is achieved by coupling a

recognition group and the surface of the nanowire. The device incorporates a Si

nanowire with distinct p- or n-type doping, source drain electrodes that are insulated from
the environment, and a microfluid device for delivery of solutions. Since the Si nanowire
is suspended in ethanol and flow aligned on substrates, and the contact leads are
characteristic of emerging with electron —beam lithography, for manufacturing purposes,
it would be beneficial to grow the device directly from substrates.

In this study, we investigated the growth of different nanowires and the control of

the growth of nanowires. The procedures encompass exploring various materials,
including GaAs, InSb, SrgGa16Ge3o and Ge, and different methods and approaches in

order to achieve the above research goal. In particular, the deposition scheme was
accomplished by the use of the pulsed laser deposition (PLD) technique, and KrF excimer
laser and the nanowire growth were utilized based on the vapor-liquid-solid (V LS)
mechanism using gold as the mediating metal catalyst. The role of oxygen in the metal-
catalyzed nanowire growth process is also studied in order to gain a deeper understanding

of the two nanowire growth methods.

CHAPTER 1 THEORY

1.1 Different grow methods

The vapor-liquid-solid growth mechanism was reported by Wagner and Ellis in
1964 [20]. In their report, a chemical vapor deposition technique was used to fabricate
high aspect ratio silicon wires of with diameters of 200 um. Many reports have followed
this initial work [21,22 ,23 ], with more recent efforts focused on diameter in the
submicron and nanoscale with reported diameters down to 3 nm [24,25]. Both chemical
vapor deposition (CVD), and physical vapor deposition (PVD) techniques have been
utilized in the fabrication of nanowires through the vapor-liquid-solid technique. More
recently, an oxide assisted (0A) technique has been proposed as another mechanism for
the growth of nanowires [26]. In this section, I review the various deposition techniques,

and identify the similarities and difference between them.

1.1.1 Chemical Vapor Deposition (CVD)

Chemical vapor deposition (CVD) is a process which is widely utilized for
depositing thin films of various materials. In addition, there are a wide range of
applications of CVD, such as the fabrication of microelectronic devices and the

deposition of protective coatings.

\ \

 

 

Source Gas Gaseous Products

 

 

 

 

 

 

 

 

Substrate

 

Heat Flux

 

 

 

Figure 1. Chemical vapor deposition process

Usually, in a CVD process, the substrate is exposed to one or more volatile
precursors, whose reaction and decomposition on the substrate surface lead to the
formation of the desired deposit.

As the deposited surface takes shape, the gas mixture is heated as it approaches the
surface which is placed upon a heated substrate. Reactant gases, which are often diluted
in a carrier gas, come into the reaction chamber at room temperature. Depending on the
process and operating conditions, the reactant gases may go through consistent chemical

reactions in the vapor phase before hitting the surface.

 

 

 

 

 

 

 

Method Energy Multi- Vacuum Typical >10
Range(eV) Element Requirement Deposition mTorr
Rate Rate (A/s) Reactive
Control Gas
CVD 0.1 Moderate HV 20 Yes
MOCVD 0. 1 Moderate UHV 10 Yes
MBE 0. 1 Moderate UHV 3 No
PECVD 0.1-500 Moderate HV 20 Yes
PLD 1-1000 Easy HV 10 Yes
TE 0. 1 Difficult HV 20 No

 

 

 

 

 

 

 

 

Table 1. Qualitative comparison of deposition rates of various growth techniques [31]

Close to the surface thermal, the momentum, and the layers of the chemical
concentration boundary change with the status of the gas stream. They form as the gas
streams heats, and then slow down due to the viscous drag, and the chemical composition
alters accordingly. On the other hand, heterogeneous reactions of the source gases or
reactive intermediate species take place at the deposition surface which produces the
deposited material. By this way, byproducts resulting from the gaseous reaction are
carried out of the reaction chamber.

Plasma-enhanced CVD (PECVD) is the CVD process that makes use of plasma to
improve chemical reaction rates of the precursors. What’s important is that PECVD

process enables the deposition to occur at lower temperatures, which plays an important

role in the semiconductors producing.

1.1.2 Physical Vapor Deposition (PVD)

Physical vapor deposition involves atoms and/or molecules of a material that are
deposited from the gas phase (or vapor) onto a substrate without a chemical reaction step
taking place. Various mechanisms can be used to get the material into the gas phase
including, thermal evaporation, sputtering, pulsed laser deposition, cathodic arc, and
molecular beam epitaxy. During the PVD process, material fiom a liquid or solid source
is transported through the gas phase to a substrate. In placing the atoms or molecules
from the source into the vapor phase, the particles are raised in energy. As the particles
arrive, their energy is taken away by a cooler substrate surface facing the source, which
allows the particles to form a solid thin layer. The atoms or molecules in the vapor phase
will condense on the substrate after some surface diffusion of the adatoms.

Accommodation of the adatoms occurs as the Gibbs free energy, AG, is sufficiently

reduced during condensation. Since AG is further reduced for adatoms that form clusters,
or accommodate to step edges, there is a preferential growth of films along such defects,

and/or in island formations. These mechanisms for film grth are summarized in blow.

Hole fdge
Atom on Step

surface _
r 1):?"- . ..

    

Crevice

 

 

Screw
Dislocation Corner

 

 

Figure 2. Schemiatic of film surface with various edges, pits, and clusters identified.

Assuming the incident kinetic energy is not too high; an atom striking a surface
will lose its velocity component normal to the surface in a short time. The incident vapor
atom is then physically adsorbed and called an adatom. It may or may not be completely
thermally equilibrated with the surface. If not equilibrated, it can move over the surface
by jumping fiom one potential well to another as driven by thermal activation from the
surface and/or by its own kinetic energy parallel to the surface. The adatom has a finite
stay or residence time on the surface during which it may interact with other adatoms to
form a stable cluster and be chemically adsorbed (incorporated into the surface) with the
release of the heat of condensation. If not adsorbed, the adatom reevaporates (desorbs)
into the vapor phase. Therefore, condensation is the net result of an equilibrium between

the adsorption and desorption processes.

10

Sticking Coefficient

Sticking coefficient is the probability that an impinging atom will be incorporated
into the substrate (surface) is called the “condensation” or “sticking” coefficient. This
probability is given as the ratio of the number of atoms incorporated into the film to the

number of atoms impinging on the surface.

_ Nadh

— Equation 1
S Ntot

where Nadh = number of atoms adhering to the surface and Ntot = total number of atoms

arriving at the substrate.
Accommodation Coefficient

The accommodation coefficient, “r is a measure of the degree of thermal equilibration

with the surface and is given by:

.— E-—E .
aT = T’ T, =—‘—L Equatron 2
Iii—Ts Ei_Es

 

where temperatures, T’s, and kinetic energies, E’s, correspond to the equivalent rms
values for the incident (1'), reevaporated (r) atoms, and the substrate (s) respectively.
Several theoretical investigations considered the problem of a head-on collision of
an atom with a one-dimensional lattice consisting of spring-connected masses. For
nearly equal masses of the impinging atom and the substrate lattice atom, a unity sticking

coefficient should be obtained for incident kinetic energies up to 25 times the desorption

energy, Edes- Since, for metals, Edes 2 1-4 eV, vapor atoms of equivalent beam

temperatures of the order of a million degrees should be physically adsorbed. Of course,

a 3D lattice of spring connected masses would be stiffer, decreasing the expected sticking

11

coefficient. For light impinging atoms relative to the substrate atoms, and/or for high

incident kinetic energy, the sticking coefficient can be much less than unity.

As mentioned, the adatom has a finite stay or residence time, Is, on the surface given by

1 Edes
: _e kBT
V

rs Equation 3

where Edes = energy of desorption, v = surface vibrational frequency (~10l2 Hz), k3 is

Boltzmann’s constant (8.618x10"5 eV/K). It has been theoretically predicted that almost

all of the initial energy Ei is lost in the first three lattice vibrations and that the time to

. . . . 2 .
thermally equrlrbrate wrth the substrate, re, rs about Te S — or about a prcosecond.
v

At high binding energies, rs is large (re is small - v large) and thermal equilibrium
occurs rapidly. Here an adatom can be considered as localized and it will diffuse by
discrete jumps.

If Edes ~ kBT, they do not equilibrate rapidly but remain “hot” (usually having as <

l), and the adatoms can be viewed as a 2D gas (where the kinetic theory of gasses can be

applied).
During IS, the adatoms will diffuse an average distance, 5(- , of

Es Edes ‘Es

27:,l2Dsrs =am/2vrse ZkBT =J2a0e ZkBT Equation4

where a0 is the jump distance between adsorption sites on the surface (on the order of

 

atomic dimensions), E S is the activation energy for a surface diffusion jump, and D5 is

the surface diffusion coefficient:

12

Es

 

2 _kBT

Ds 1‘ 001/8

Equation 5

Several values of desorption and activation energies are given in the following table.

These values are very sensitive to the cleanliness of the films as shown for Cd on clean

and contaminated Ag substrates.

Table 2. Experimental values for the binding energy, Edes, and the activation energy for
diffusion, E,, for some typical systems.

 

 

 

 

 

 

Condensate Substrate EdeSa eV Es: eV
Ag NaCl 0.2
Ag NaCl 0.15 (evaporation)
0.10 (sputtering)
Al NaCl 0.6
Mica 0.9
Ba W 3.8 0.65
Cd Ag (fresh film) 1.6
Ag, glass 0.24
Cu Glass 0.14
Cs W 2.8 0.61
Hg Ag 0.11
Pt NaCl 0.18
W W 3.8 0.65

 

 

13

Experimentally the value of Edes has typically shown a value of about 4 times the

diffusion activation energy E S.

Homogenous Nucleation
For the homogenous nucleation of a spherical solid phase of material, the change in

free energy can be described as

4 .
AG = 370:3AGV + 4m2 7 Equation 6

where r is the radius of the sphere, y is the surface free energy per unit area, and A0,, is

the change in free energy per unit volume as given by

P
...k_311n_"_

A0,:
(2 P,

Equation 7

Here 0 equals the atomic volume of the solid, Pv is the pressure of the supersaturated
vapor, and PS is the vapor pressure at the surface of the solid. With the supersaturation,

S, defined as (Pv — Ps)/PS we could write

-525

A0,, = ln(1+ S) Equation 8

Nucleation will not occur unless there is supersaturation (S > 0) and A0,, is negative.

An energy barrier to the formation of stable nuclei is evident when the derivative of

(Equation 6) with respect to r is set equal to zero as shown in Figure 3.

14

Free
Energy
Change
AG

 

 

 

Figure 3. Free-energy change (AG) as a function of cluster (r. > r) or stable nucleus
(r > r ) size. r is critical nucleus size, and AG is critical free-energy barrier
for nucleation.

The critical radius is found to be

r* = — 27 Equation 9
A0,,

 

or using this in (Equation 6) gives the critical free energy banier for homogenous

nucleation

= —— Equation 10

The nucleation rate for this sphere is

15

IV = N‘A‘w Equation 11

where

1.:
AG

:11
N = nse kBT

Equation 12
with n S = density of all possible nucleation sites, and M = equilibrium concentration of
stable nuclei, (0 is the rate at which atoms impinge on the the nuclei of critical area,

Al= 41tr2.

P —P N
w=aS( v S) A Equation13
ZWRT

 

or upon substitution

*
AG

- — —P
N=nse kBT 47‘72 as(Pv s)NA

Equation 14
27zMRT

This shows N is strongly dependent on AG‘ which is strongly dependent on the
supersaturation, S as was shown in (Equation 8). The effect of supersaturation

on AG‘ is shown in Figure 4.

l6

Surface
Energy

AGO

 

 

 

Volume
Energy

Figure 4. Total free energy (surface plus volume free energies) of cluster formation vs.
size (r) for different supersaturations (S).

Capillarity Theory

The capillarity theory extends the homogenous theory to the nucleation of a film on
a planar substrate. The three different growth modes of a film on a substrate are
Volmer-Weber (or island), Frank-van der Merwe (or layer), and Stranski-Krastanov (or

layer plus island) are summarized in Figure 5.

17

Coverage

 

 

 

 

 

 

 

 

 

 

 

 

0<1ML 1<9<2 0>2ML

Mechanism

3-D Island

GmW‘h _a_=._ n m n m
2-D Layer-by- -——1
layer Growth __, ,—

Stranski-

Krastanov r1 J1

Growth -—1 r—- ——m

 

 

 

 

 

 

 

 

 

Figure 5. Schematic representation of three film growth modes. 9 is the overlayer
coverage in monolayers (ML).

Here three interface tensions must be considered as shown in Figure 6, to replace the

interface energy, 7, in (Equation 2).

18

Deposition

 

 

 

Desorption
Vapor
0
Film nucleus
v 17 Wt“
9 O
K O m
st 0
Substrate
'st

Figure 6. Schematic of basic atomistic processes on substrate surface during vapor

deposition.

The substrate-to-vapor, st, film-to-vapor, yvf, and film-to-substrate, 713 are related by
st = yfi + y” c036 Equation 15
And (Equation 2) becomes

AG = a3r3AGv + alrzyfl + azrzyfs — azrzysv Equation 16

G = 0 which gives
dr

 

. . . ’9 .
The crrtrcal cluster srze, r , is found from

 

r: :_2|,aly1gf “12(7):: ’73vll

2 Equation 17
3a 3 AGV

and

a: = 40171f +02(7fs —7sv)]3

AG
27a§AGE

 

Equation 18

19

In general, PVD is used to deposit films with thickness ranging from a few
nanometers to a few micrometers. In addition, PVD can be used to create multilayer
structures by supplying different source materials during the process. The substrate size
and material can be varied.

PVD processes can be used to deposit films of elements and alloys as well as
compounds using reactive deposition processes. In reactive deposition processes,
compounds are formed by the reaction of depositing material with the ambient gas
environment. Quasi-reactive deposition is the deposition of films of a compound
material from a compound source. During the transport and condensation process,
having a partial pressure of reactive gas in the deposition environment can compensate
the loss of more volatile species or less reactive species. A couple of examples

representative of the physical deposition are given below:

Molecular beam epitaxy

Molecular beam epitaxy (MBE) is a beam of molecules or atoms which emanate
from an evaporation or effusion source. The beam of particles travels across an
evacuated chamber with few if any collisions, and deposits onto a substrate. Throughout

the process, atoms are deposited one layer at a time, under ultra-high vacuum (UHV

-10

10 — 10'11 Torr), which leads to the formation of a perfect crystal. At this pressure, the

mean-free path of gases is several orders of magnitude larger than then distance between

the source and substrate. At very low pressures, the mean-free path can be approximated

by [27]:

20

_ 5x10—3
P

L

 

Equation 19

In solid-source MBE, pure elements such as indium and antimony are heated in
separate effusion cells furnaces until they each slowly begin to evaporate. The
evaporated elements then condense on a heated substrate, where they react with each

other and create indium antimony.

Atomically clean substrate at temperature T

 

Ultra-high vacuum

Antimony flux Indium flux

Figure 7. Schematic representation of MBE deposition of a thin film. The film
thickness and composition is defined by controlling the individual fluxes.

A computer controls shutters in front of each effusion cell, allowing precise control
of the thickness of each layer, down to a single layer of atoms. Intricate structures of
layers of different materials may be fabricated this way. Such control has allowed the
development of structures where the electrons can be spacially confined, giving quantum
wells or even quantum dots. Such layers are now a critical part of many modern
semiconductor devices, including semiconductor lasers and light emitting diodes.

The ultra-high vacuum environment within the growth chamber is maintained by a
system of cryopumps, and cryopanels, chilled using liquid nitrogen to a temperature of 77

kelvins (-l96 degrees Celsius). The wafers on which the crystals are grown are mounted

21

on a rotating platter which can be heated to several hundred degrees Celsius during
operation.

The main advantages of MBE are the low growth rate, which is not conditioned by
wafer temperatures and facilitates a precise control of film thickness, and high purity
where only elements that are components in the deposited film are in the vapor phase
during deposition. In this way, chemical reaction byproducts can not be incorporated as

impurities as can be the case with CVD methods.

Thermal evaporation is one of the oldest techniques of depositing thin films.
Evaporation uses a heater to melt the material and raises its vapor pressure to a useful
range. The vapor pressure is a pressure at which the vapor of the targeted substance is in
equilibrium with its liquid or solid forms. Employing the thermodynamic principle, the

number of molecules departing a unit area of evaporant per second, or flux F, is given by:
F = N ¢ exp— (£9) Equation 20
kT

where N ¢ is a slowly variable fiinction of the temperature (T) and ¢e is the activation

energy (in eV) needed to evaporate cone molecule of the material. The activation energy
for evaporation is associated with the enthalpy of the formation of the evaporant, H, as

H=¢e xexN. Equation 21

This is executed in a high vacuum system, which could allow the vapor to approach
the substrate without reacting with or spreading out against other gas-phase atoms in the
chamber, and could reduce the consolidation of unwanted substances from the residual

gas in the vacuum chamber [28]. To avoid contamination, only those materials with a

22

much higher vapor pressure than the heating element can be deposited. Thermal
evaporation of compounds and alloys is not commonly done, since the individual

elements typically have different evaporation rates.

Sputtering relies on ions to strike materials from a target, in which a small number of
atoms are obtained at a time. The physical sputtering process entails the physical
vaporization of atoms from a surface via momentum transfer realized by bombarding

energetic atomic sized particles.

Incident ion _
Reflected ron or atom

O O

O Souttered atom

0

Surface Secondarv electron

 

O Implanted ion

Figure 8. A schematic representation of the sputtering process.

Commonly, the energetic particles are ions of a gaseous material being accelerated in an
electric field [29]. For example, during the process of sputtering, the target, such as a

disc of the material to be deposited, at a high negative potential, is bombarded repeatedly

23

with positive argon ions. The target material is sputtered away principally in the form of
neutral atoms through momentum transfer and then ejected surface atoms are deposited
onto the substrate placed on the anode. The feature of allowing a great deal of materials
available to use makes sputtering a preferred option over evaporation in many
applications [28]. As shown in Equation 4, the amount of materials, W, sputtered from
the cathode is inversely related to the gas pressure P, and the anode-cathode distance, d:

W_kW

— — Equation 22
Pd

where V is the working voltage, i is the discharge current and k is a proportionality
constant. The target can be maintained at a low temperature by water cooling, and
compounds and alloys can be deposited with good transfer of stoichiometry from the

target to the substrate.

Pulsed laser deposition (PLD) has emerged as one of the simplest and most versatile
methods for the deposition of thin films of a wide variety of materials. The stochiometric
removal of constituent species from the target during the ablation process is one of the
main advantages in PLD compared with other deposition techniques. Different thin film
materials have been deposited by the use of PLD, including metals, semiconductors,
nitrides, dielectrics, oxide, organic compounds/polymers, and ternary compounds. It has
been described as an optical sputtering technique, and with the power source external to

the deposition chamber, multiple chambers can use the same laser [30].

24

1.2 Pulsed Laser Deposition (PLD)

1.2.1 History and Fundamental of pulsed laser deposition
Introduction

Pulsed laser deposition (PLD) is a technique commonly used to deposit thin films
[31]. This technique involves the evaporation of a solid target by short, high-energy laser
pulses. The laser beam vaporizes the surface of the target, and the vapor condenses on a
substrate placed near the target.

The PLD technique has the following advantageous characteristics: 1. Simple: a
laser beam vaporizes a target surface, producing a film with the same composition as the
target. 2. Versatile: many materials can be deposited in a wide variety of gases over a vast
range of gas pressures. 3. Cost-effective: one laser can serve many vacuum systems. 4.
Fast: high quality samples can be grown in 10 or 15 minutes. 5. Scalable: complex oxides

move toward the volume production [31].

Historical Development of Pulsed Laser Deposition

1916 - Albert Einstein proposed a hypothesis regarding the process of the stimulated
emission.

1960 - Theodore H. Maiman built the first optical maser by means of a rod of ruby as the
lasing medium.

1962 — Breech and Cross utilized the ruby laser to vaporize and excite atoms from a solid
surface.

1963 - Askar’yan made contributions to the development of the laser beam to liquid

surface interactions.

25

1963 — Meyerand and Haught investigated the laser beam to gaseous material
interactions.

1965 - Smith and Turner conducted studies on semiconductors, dielectrics,
chalcogenides, and organometallic materials.

1970s -Q -switches were used to produce short optical pulses under the condition of

peak power.

2(1)
680"

>108W/cm2. E:

 

E = electric field of electromagnetic wave (V/cm)

(I) = power density in (W/ cmz)
£0= dielectric constant in vacuum (8.85 x 10'12 F/m)

n = refractive index

c = velocity of light (3.00 x 108 m/s)

1975 — A complex stoichiometry took form with the low temp. superconductor Re 322

(Desserre and Eloy)

1979 — (Gapanov et al., 1979) semiconductor superlattices. Also demonstrated reactive
deposition in 1979.

1980s - PLD as a technique for film growth earned recognition and drew a great deal of
interest. Primarily, PLD was used to formulate crystalline thin films with
epitaxy quality. In addition, PLD was present in the growth of the ceramic
oxide, nitride films, metallic multilayers, and various superlattices.

1990s — The swift development of the laser technology rendered PLD more desirable.

Newly developed lasers were characteristic of higher repetition rate than the

26

early ruby lasers, which made the growth of thin films possible. Then, reliable
electronic Q-switches laser was adopted to generate optical pulses of extremely
short length. It is this factor that created the opportunities for PLD to be used to
attain the congruent evaporation of the target and to deposit stoichiometric thin
films. In general, this significant breakthrough resulted in the development of
the laser with a high efficient harmonic generator and excimer lasers delivering
UV radiation. The non-thermal laser ablation of the target material was
influenced and became exceedingly efficient thereafter. Additionally, using
PLD to synthesis buckminster fullerness and nanopowers was another example
of its application.

20008 — Some production-related issues began to gain interests, such as reproducibility,
the large-area scale-up, and the multiple-level. In other words, the attention

awarded to the thin film fabrication reached its peak time.

1.2.2 Pulsed Laser Deposition: Mechanisms of Ablation
Introduction

Although the Pulse Laser Deposition (PLD) method is conceptually simple, the
underlying ablation mechanisms are related to complex physical principles, such as
collisional, thermal evaporation, electronic sputtering, exfoliation and hydrodynamics
sputtering. In a typical PLD process, a target is placed in a vacuum chamber, and a pulsed
laser beam converts the target material from solid into a vapor, often with a plasma or
plume formed above the target, the evaporated species have high translational energy,

and expand away from the target toward the substrate where condensation occurs. The

27

main components of the above physical process include a laser, optics, and a vacuum
system. In general, the thin film growth process using PLD can be divided into the
following four phases:
1. The connection of a high flux of laser radiation with the target and surface and
the laser-plume interaction.
2. Dynamic ranges of ablation materials.
3. Deposition of the ablation materials with the substrate.

4. Nucleation and growth of a thin film on the substrate surface.

Each phase in PLD is important in fabricating crystalline thin films with epitaxial
quality and those having stoichiometric, uniform and small surface roughness
composition. In the first stage, the pulsed laser beam is focused onto the surface of the
target. The high flux densities and short pulse duration are conducive to the rapid heating
of all elements in the target to their evaporation temperature. Materials are then removed
from the target surface and are ablated out with stoichiometry as in the target. The
fluences of the laser shining on the target determine the immediate ablation rate.

Afler the laser-plume interaction, during the second stage of the thin film growth,
the emitted materials are inclined to move towards the substrate under the influence of
the gas-dynarnic expansion of plasma created by irradiating a solid with laser pulses, and

to demonstrate the forward peaking phenomenon. The spatial thickness is a parameter

that is varied as a function of cos" q , and the spot size of the laser and the plasma

temperature exert major effects on the uniformity of the deposited film. In addition, the

28

distance between the target and the substrate is a deciding factor that affects the angular
spread of the ablated materials.

The third stage of deposition of the ablation materials with the substrate is critical
in determining the quality of the produced thin film. The reason is that the ejected high-
energy species have effect on the substrate surface and might incur a variety of damage to
the substrate. To be more specific, these energetic species can sputter some of the
surface atoms and a collision region will be shaped between the incident flow and the
sputtered atoms. Then, the film could be produced after a therrnalized region is formed.
In fact, the collision region is utilized as a source for the condensation of particles. If the
condensation rate is higher than the rate of particles resulted from the sputtering, the
thermal equilibrium condition can be achieved fast and the film will grow on the
substrate surface under the condition of the obstruction of the direct flow of the ablation
particles and the thermal equilibrium reached.

In the final stage of the nucleation-and-growth of crystalline films depends the
condensing material, temperature and the properties of substrate. The nucleation process
depends on the interfacial energies between the three phases present, the vapor atom,
condensing layer, and the substrate. It is described that the minimum-energy shape of a
nucleus is like a cap. The deposition rate and the substrate temperature are such factors
that might affect the critical size of the nucleus. In the case of the large nuclei, which are
a characteristic of small supersaturation, they create insulated patches of the film on the
substrate which are produced thereafter and coalesce together. While the supersaturation
increases, the critical nucleus is reduced until its height extends to the atomic diameter

and its shape is equal to that of a two-dimensional layer. If the scale of the

29

supersaturation is large, the layer-by—layer nucleation will occur for partially wetted
foreign substrates.

In order to grow the crystalline film, the surface mobility of the adatom, or the
vapor atoms, is required. Usually, the adatom will diffuse through different atomic
distances before holding to a stable position within the newly formed film. The adatom's
surface diffusion ability is determined by the surface temperature of the substrate. In
general, the high temperature is conducive to the rapid and defect-free crystal growth,
while at low temperature or large supersaturation, the energetic particle effect might deter
the crystal growth and result in disordered or even amorphous structures.

In the PLD process, during which the short laser pulsed duration (~10 ns) affecting
the small temporal spread (<=10 ms) of the ablated materials, the deposition rate can be
quite high (~10 urn/s). As a result, a layer of thin film can be produced with the aid of
the preferred layer-by—layer nucleation. In addition, since the deposited particles can be
highly energetic (typically 1-10eV in energy), surface diffusion of the adatoms is
enhanced, and surface temperatures can be higher than the bulk substrate. These
mechanisms can help to lower the substrate temperature needed for crystalline film
growth.

As mentioned before, mass removal takes place from the target surface where the
laser ablates. The whole process includes various complex mechanisms. At sufficiently
high flux densities and short pulse duration, elements in the target get more energy by
rapidly heated up. When the target surface material has gained enough energy, particles

are dissociated from the target surface. Five mechanisms for the emission of target

30

material have been identified as: Collisional, Thermal Evaporation, Electronic Sputtering,

Exfoliation, and Hydrodynamic sputtering.

1.2.2.1 Collisional

Collisional mechanisms are present for ion and electron bombardment of a target,

but not for photon bombardment. Nonetheless, indirect collisional effects do exist with

photons. It has been shown that as plasma takes shape during the laser-surface

interaction. The plasma formed above the target surface by photon bombardment will

include creating electrons and ions, which collisionally interacts with the target.

 

Higher magnification views of cone
structures.

ION BEAM

Sketches relevant to the evolation of an
asperity on an ion-bombarded surface.

Figure 9. Formation of cone structures during laser ablation, indicating a collisional
interaction of particles with the target surface [31].

2512 + 4M1MzEr

A
E: 2 2
mzc (M1 +M2) M2

 

2
E
=—lx2.l47x10_9 +

4M1M2E1

2 in eV Equation 23
(M 1 + M2)

31

E refers to the maximum energy transfer, E1 represents the binary collision between a
particle with energy, and m1 is rest mass, m2 , a target atom of mass. M1 and M 2 are

the atomic weights of the two particles.

1.2.2.2 Thermal Evaporation

Thermal evaporation may require temperatures well above the melting or boiling
points. Thermal evaporation commonly indicates a process during which the laser energy
is converted to the lattice vibration energy proceeding to the bonds breaking. Some

atomic-sized materials from the target surface will be released thereafter.

Kelly and Rothenberg (1985) used the following equation describing thermal
sputtering which is based on the application of the vaporizing flux from a condensed
phase:

Vaporizing flux = condensing flux

=p(2mnkBT)_1/2

 

k TV } x (271mk B T) ‘1 / 2 atoms / mzs Equation 24
B

= P0 {exp
where p is the pressure under the condition of the existence of the equilibrium vapor, p0

is a constant, and AH v is the heat of the vaporization while the condensation efficiency

is assumed to be unity. The total loss from the target is demonstrated as

_ — AH _ .
Depth/pulse = @- (271ka) 1/ 2 x I00 exp(——‘i)T 1/ 2(it Equatron 25
nc 0 kBT

where nc is the number density of the condensed phase.

32

1/2
A A
Depth/pulsez (pm T r/Ml/ZAHV) x1.53 x 106m/ pulse Equation 26

where T is the maximum surface temperature, M is the molecular weight of the
vaporizing species, and AH v is in electron volts.

Table 3. Temperature of metling, Boiling, and a Vaporization Rate of 1 nm/pulse for
Selected Solids. [31]

 

 

 

 

 

Solid Melting Point (K) Boiling Point at latrn (K) Temperature for a
Vaporization Rate of
lnm/pulse (K)

Au 1336 3 1 33 4600

Si 1685 1685 4600

MgO 3098 3530 4200

Si02 1995 3200 3800

 

 

 

 

 

 

In fact, the vaporizing flux equation is a standard thermal evaporation formula and the

condensation efficiency is one, or unity.

1.2.2.3 Electronic Sputtering

Electronic sputtering involves the electronic excitation which leads directly to the
bond breaking. The electronic excitation begins with the absorption of the laser energy,
which results in the excitation of the outer shell electrons. The atoms and ions are
liberated if the excitation energy is greater than the binding energy. For example, a laser

of wavelength 1 = 248nm, hv = 5.01 eV, so for materials with the bandgaps less than 5

33

eV, each photon can excite an electron across the bandgap. One characteristic of this

mechanism is the existence of a threshold energy density for the particle emission.

1.2.2.4 Exfoliation

Exfoliation of materials can occur when a target is repeatedly thermally shocked
yielding flakes of materials that break from the surface. It typically occurs when the
system has a high linear thermal expansion (AL/L), where L is the thickness that is heated
and AL is the change in thickness, a high Young’s modules (15), a high melting point
(TM), and the laser-induced temperature excursions approach but do not exceed T M. The
thermal shocks occur repeatedly and, since they are not relieved by melting under the
above conditions, they eventually lead to cracking.

According to the argument of Kelly, the thermal shock can be measured in terms of

thermal stress which can be expressed as

stress = PAL— Equation 27

Table 4. Quantities needed for evaluating thermal stress by means of Equation 5 [31].

 

 

 

 

 

Solid Linear Thermal Expansion Young’s Modulus, E EAL/L0
of the Solid of solid (1010xPa) (joloxpa)

Au 0.018 8.0 0.12

Si 0.0054 1 1.3 0.06

MgO 0.045 26 1.2

8102 0.00077 7.2 0.006

 

 

 

 

 

 

34

Evidenced from the table, exfoliational sputtering will not occur with low melting
metals. In addition, due to the influence of the low thermal stress, gold and silicon, in

particular, are not expected to experience exfoliational sputtering.

1.2.2.5 Hydrodynamic Sputtering
Hydrodynamic sputtering comprises the formation of droplets on the surface of the
target material caused by surface melting, and the eventual expulsion of droplets as a

consequence of the transient melting.

 

Figure 10. Droplet formation in the gold target. Sputtering of polycrystalline Au in air
after 4320 pulses from an ArF excimer laser (193nm) of 12 ns and 2.51/cm2.
(From Kelly et al. 1985)

Plasma gas-dynamic expansion
Semiquantitative models have provided expressive power for the expansion
dynamics of a laser-generated plasma plume. Gas dynamics are deemed to be a factor

influencing spatial and velocity distributions and the hydrodynamic equations can be

35

used to model the gas cloud moving away from the target. The model is focused on
analyzing the density of the gas that is time-and space-dependent and the angular velocity
distribution of the ejected particles. In the simulation process, the particle motion is
developed in the direction x-axis perpendicular to the two lateral directions of the target
plane, y and z. The process can be divided into following steps: interaction of the laser
with the target resulting in the formation of the plasma, the isothermal expansion, and the
adiabatic expansion of the gas cloud.

According to the above illustration, the first two regimes take place during the laser
pulse, and the latter one is especially applied after the laser pulse is over. In the
isothermal expansion, since the description of the relation between mass, momentum and
plasma energy is required, the gas-dynamic equations of mass and momentum
conservation are preferred. The velocities are determined by the pressure gradient, which
influences the development of the expansion into an anisotropic form. When the laser
pulse ends, the regime changes to the unsteady adiabatic expansion (UAE), in which no

injection of particles is considered into factors controlling the expansion.

Surface
of target
. Unsteady adiabatic
Free fllght i expansion 5 Knudsen Z
’l‘ "‘ layer "

I
I
m W ”—— .—u——- .—. * .

+ ++ +++4

Flow velocity Gas-kinetic velocity

 

Figure 11. Plasma gas-dynamic expansion [31].

36

The dimensions of the gas cloud govern the expansion of the plasma in three
directions once it is formed. For example, when the influence from the target is large, the
highest velocities are in the direction of the smallest dimensions and vice versa, which
helps produce an ellipsoidal plume that is stretched out from the surface. In addition,
transverse expansions are created, which are determined by two lateral dimensions of the
gas cloud when the plume is close to the target. In this situation, the forward expansion
is the preferred direction. For example, an elliptical plume is produced by a symmetrical,
or circular, spot. Nonetheless, in the situation of the non-symmetrical spot, the expansion
velocity is larger for the direction where the spot is smaller initially. At that time, due to
the influence of the laser spot anisotropy, the pseudorotation, or flip-over, of the plume

shape will take place.

Region—I

. Region-III

Target
Material

Refill-31%”

 

Laser produced plasma plume from a metal target

Figure 12. Laser produced plasma plume from a metal target [32].

37

In terms of physical composition, the laser induced plasma can be divided into
three regions as shown in the above image.

Region-I is the essential part, in which the emission of the plasma occurs close to
the sample surface. In the sample surface, the temperature is set at maximum and nearly
all the species are in inonized state.

Region-II is the middle region, in which both neutral and a certain number of
molecular species appear with the exception of ionized species.

Region-HI is the extended region of the plasma. In this region, the plasma
temperature is lower compared with other two regions, and the molecular species with

larger density are present.

1.2.3 Particulates generates by pulsed laser deposition

In general, the quality of the deposited film is mainly determined by the
crystallinity of the lattice and the surface smoothness. The generation of particulates by
pulsed ablation is the major factor influencing the smoothness of the resulting thin film.
To be more specific, the nature of particulates covering the surface of thin films,
including the generation rate, the chemistry, the energy, the size, and the microstructure,
is determined by the process condition and the type of the material [33]. In particular, in
the case of high-performance optical and electronic applications, due to the rigid
constraints on surface smoothness, the demand for the particulate density and its size is
extremely high.

In the following paragraphs, details about particulates are discussed in five sections.

The first is related to their participation characterization.

38

Table 5. The different shape and size between the three phases.

 

 

 

Solid Liquid Vapor
Shape Irregular Spherical Spherical or
polyhedral
Size Sub-micron Micron nanometer

 

 

 

 

 

 

Based on the status of the original matter, there are three categories defining
particulates produced by pulsed laser ablation, especially at the moment when they are
expelled from the target. The original matter could be in solid, liquid, or vapor state.
The origin of the particulate generation is multifaceted as shown in the Table 5.

The second section is pertinent to the chemistry of particulates. Since it was found
that films produced by PLD process have chemistry quite similar to that of the target, a
mechanism meant to create an environment inducing the exhaustion of lower melting
temperature element in the particulates is viable. According to this mechanism, the
composition of the ejected liquid droplet positioned above the surface is hypothesized to
be fairly similar to that of the target. In addition, the depletion of the low melting
temperature element takes place principally as it is moving close towards the substrate.
For example, in the case of a particulate of radius r, a depletion layer of thickness is
assumed to be 1, hence, the normalized resultant content of the depleted element can be
obtained as

RESULTANT%
INITIAL%

 

= (5:53 Equation 28
r

In the third section of the characteristic of microstructures, the focus is on the

cooling rate that is found to be an important factor deciding the microstructure of

39

particulates. The reason is that particulates become solid as a result of molten droplets
ejected fiom the target. Currently, the cooling rate is measured by two different methods.
When the ablation experiment is operated in vacuum, cooling is mainly noticed as radiant
rather than convective. Put it in simples, the cooling rate can be expressed as

6

pp

T = ( )£0(T4 — T04) Equation 29

 

where D is the particulate diameter, p is the material density, Cp is the heat capacity of
the material, a is the radiant emissivity, cr is the Stefan-Boltzmann constant, T is the
temperature of the particulate, and T o is the surrounding temperature. From this formula,
it is known that as the cooling rate is high, the solidified particulates tend to have finer
microstructures. Moreover, this formula is also applicable to the conventional powder
production process, such as atomization and splat quenching.

The fourth part of the discussion is focused on the effects of processing parameters,
two of which are the laser fluence and the laser wavelength. In the situation of using a
specific material and setting the laser wavelength within a certain range, the laser fluence
on the target was found to be able to exert most significant effect on the particulate size
and density. Fluence, which is to specify the energy delivered in a given time, provides a
powerful foundation deciding how much the tissue will be heated by the laser
pulse. Higher fluence means more heating is required, which has more effect on the
tissue. Among the various factors related to the setting of treatment levels, the laser
fluence is deemed as the most influential one. The laser fluence is measured in joules per
square centimeter, rather than the power or pulse energy per unit area as in the case of
usual definition. The fluence is derived from dividing the pulse energy by the area of the

spot, whose size has determining power on the fluence. In other words, varying the laser

40

power or the laser spot size can result in changes in the laser fluence. Usually, if the laser
power is set constant, the particulate number density is higher with tighter focus.
Particularly, a threshold laser fluence exists, below which the particulates are negligible.

In addition, the impact of the laser wavelength A on the effectiveness of the
absorption of the laser power into the target has been observed. In certain cases of
metals, the absorption coefficient or decreases with the declining A, which leads to a
larger laser penetration depth in metals in the Ultraviolet (UV) range compared with that
in the infrared (IR) range. Aside from these observations, for other materials, by taking
different absorption mechanisms into account, such as lattice vibration, free carrier
absorption, impurity centers, or bandgap transition, the definition of the relationship
between the absorption coefficient and wavelength is far more complex. For example, in
the case of oxide superconductors, unlike the above conclusion obtained from metals, the
laser penetration depth appears to be larger in the near IF than in the UV range. On the
whole, it is estimated that the principle impact that the laser wavelength has on the
generation of particulates is ascribed to the difference in the absorption coefficient when
various laser wavelengths are used.

In the final section, the target-to-substrate distance is discussed, whose effect is
reflected in the angular spread of the ejected flux if the film is deposited in vacuum. At a
larger target-to-substrate distance, or with the influence from the ambient pressure,
particulates might occur; especially diverse types of particulates may emerge depending
on the position of the substrate. This situation usually arises during the PLD process

conducted in a poor vacuum. Normally, particulate trajectories are in various directions

41

if a defocused laser beam is used, compared with the condition in which an extremely
focused beam is used to help generate a collimated jet of particulates.

In fact, the effects of the target-to-substrate distance and the ambient pressure could
be interrelated. The improved collisions between the laser-produced plumes will affect
the increase of the background gas pressure. This argument is supported by the study of

Dyer et a1 (1990). According to their investigation, E / R, is used as a scaling parameter
for measuring the plume range, where E is the laser-pulse energy and R, is the

background gas pressure. Furthermore, the length of the plume can be explained by

L oc (E / p0)“ 37 , where 7 represents the ratio of specific heats of the elements in the

plume. If the target-to-substrate distance is greatly smaller than L, there are no
discernible changes in the particulate size and the density. Nonetheless, when the target-
to-substrate distance increases, the quantity of smaller particulates will decline compared
with that of entire particulates. Then, a few larger particulates will emerge, which
implies a merge takes place during flight. Moreover, as soon as this distance is large
enough, or the substrate is located far beyond L, the adhesion to the substrate is reduced
to a certain degree. The substrate is composed of the ejected matter, including the

particulates and atomic species.

42

1.3 Thermodynamic and chemical potential

In terms of being predictive of whether an alloy is in equilibrium, thermodynamics
is deemed to be a useful tool, especially during phase transformations in which changes
toward equilibrium are usually the main concerns. In fact, this is also the main
contribution that thermodynamics has made to physical metallurgy [34]. In normal
situation under which the temperature and pressure is constant, the relative stability of a
system during phase transformations is determined by its Gibbs free energy (G). The
Gibbs free energy of a system is defined by the equation

G = H — TS Equation 30
where H is the enthalpy, T the absolute temperature, and S the entropy of the system.
Enthalpy is a measure of the heat content of the system and is induced by the following
equation

H = E + PV Equation 31
where E is the internal energy of the system, P the pressure, and V the volume. The
internal energy occurs from the total kinetic and potential energies of the atoms within
the system. There are two sources of the kinetic energy, one of which is from the atomic
vibration in solid or liquid status. The other one is from translational and rotational
energies for the atoms and molecules within a liquid or gas. In the latter condition, the
interactions between the atoms within the system give rise to the desired energy.
Moreover, changes in the internal energy and the volume of the system will determine the
status of the absorbed or evolved heat when a transformation or reaction takes place.
Then, this situation can be applied to the two factors, pressure and volume, made of the

term PV, and result in the condition that, at constant pressure, the heat absorbed or

43

evolved is produced corresponding with the change in H. Usually, PV is comparatively
small with E when operating under condensed phases, including solids and liquids, which
lead to the possible difference between H and E. The remaining factor that is related to
the term G is S, the entropy of the system, which is to measure the randomness of the
system.

A definition of a system to be in equilibrium means it is in the most stable state,.
Applying the laws of classical thermodynamics, a closed system, such as through the
fixed mass and composition, if set at constant temperature and pressure, will be in
equilibrium with the aid of the lowest possible value of the Gibb free energy, or AG =0.
According to the definition of G, if the state is described to be at the highest stability, at
the minimum energy caused by a low enthalpy and high entropy. At low temperatures,
since solid phases have the strongest atomic binding, which is contributive to the lowest
internal energy, they are preferably stable. On the other hand, at high temperatures,
phases are influenced by the negative product of the absolute temperature, T, and the
entropy of the system, S, and therefore are demonstrated with most stable state
characteristic of more freedom of atom movement, liquids and gases.

The following chart illustrates the lowest possible fiee energy defining the

equilibrium condition.

44

Gibbs Free
Energy (G)

 

 

 

G1 G2

Figure 13. Variation of Gibbs free energy with the arrangement of atoms.
Schematically, the configuration G2 has the lowest free energy which
results in a system that is at stable equilibrium. The configuration G1
is a metastable equilibrium.

If placing these configurations onto the Figure 13, in which different atomic
configurations can be represented by points along the abscissa, the configuration G2
would be seen to be at the stable equilibrium state. At this point, minor changes in the
arrangement of atoms to a first approximation create merely no changes in G, which
could be representative of the example of AG =0. Nonetheless, if desiring to create the
value AG =0, other arrangements, such as G1, could potentially make contributions by
which placing at a local minimum in free energy and therefore also satisfying AG =0.
This condition, however, does not have the lowest possible value of G. The condition at
G1 is called a metastable equilibrium state. In reality, during the states for which AG ¢ 0,
called intermediary states, the unstable status could be only realized lasting a short period
of time. On the other hand, if the effect of thermal fluctuations on the arrangement of

atoms is sufficiently high, atoms can rearrange into one of the free energy minima. A

45

system can transform into a new stable (or metastable) equilibrium state with a change of
temperature or pressure.

Since any transformation leading to a decline in Gibbs free energy is possible, a
definite criterion for any phase transformation is

AG = GZ-GI < 0 Equation 32

where GI and G2 represent free energies of the early and final states, respectively. Being
able to skip a series of intermediate metastable states, the transformation could arrive at
the stable equilibrium state directly.

By definition, the intensive properties of a system refer to those that are
independent of the size of the system, such as T and P, while the extensive properties are
those that are changing in proportion to the quantity of materials in the system, such as V,

E,H,SandG.

1.3.1 Binary solutions

Similar with single component systems, in which all phases have the same
composition and the equilibrium simply encompasses pressure and temperature as
variables, the composition is also variable in the case of alloys. Nonetheless, in order to
understand phase changes in alloys, an appraisal of how the Gibbs free energy of a given
phase is required, which is determined by the composition, the temperature, and the

pressure.

46

1.3.1.1 The Gibbs Free Energy of Binary Solutions
In order to obtain the Gibbs free energy of binary solutions in alloys, a calculation

of individual atoms in the system is necessary, which is derived from the free energies of
pure atoms. For example, during the mixing of atom A and B, the two materials are
mixed to create a solid material, or the system, in a flexible way, which is to set one mole
mixed with X mole of A and Y mole of B. The following formula can demonstrate the
relationship between X and Y

X + Y = 1 Equation 33
where X represents the mole fiaction of A in the alloy and Y is the mole fraction of B.

Thus, the total free energy of the system before the process of mixing is shown as

Gtotal :2 XGA + YG B Equation 34

where G A is the free energy of material A and G B is the free energy of material B

 

 

 

>0
on

Figure 14. Variation of G1 (the free energy before mixing) with alloy composition.

47

Due to the difference in Gibbs free energy, during the mixing process, the total free
energy of the system will not remain stable, such that after the mixture is formed, the

total free energy becomes Gmtal'
610101.: Gtotal + AG Equation 35

where AG is the difference in Gibbs free energy before and after the mixing.
Combining the condition of Gmm, = H — TS , Gmml': H '—T S', and AH = H '—H and
AS = S'—S , we could get the following conclusion
AG = AH — TAS Equation 36
AH represents the change in heat during the process of mixing as well as the change in

internal energy (E) which is caused by the mixing. On the other hand, AS refers to the

difference in entropy before and after the mixing.

1.3.2 Chemical Potential

By definition, the chemical potential of a fixed entropy and a volume
thermodynamic system refers to the possibility of energy changes in the system with the
introduction of an additional particle. If a system has various species of particles, a
variety of different chemical potentials would take place, which is associated with each
species of particles respectively. In the situation, the change in energy is explained by
the growing number of the particles of that specific species which is increased by one. In
terms of the investigation of systems composed of reacting particles, the chemical
potential is especially critical. For example, in the case of two species 1 and 2, particles

belonged to species 1 can transform into those of species 2 and vice versa. A

48

supersaturated design mixed with water liquid (species 1) and water vapor (species 2) can
exemplify this kind of situation.

According to the second law of thermodynamic, which states that heat plays an
important role in energy conversions in that all forms of energy can be converted into
heat, nonetheless, the reverse process cannot be realized by way of transforming the
converted heat back to its original form. In other words, the conversion process
involving heat cannot be initiated without loss. Based on this fact, any increase in one
chemical potential would lead to a net release of energy of the system in the form of heat,
during which the species of increased potential transform into the other species.
Similarly, if the reverse transformation occurs, a net gain of energy would emerge in the
form of heat. If the system is in equilibrium, the chemical potentials of the two species
should have been equal. In the case of alloys, if atoms are added or removed,
understanding how the free energy of a given phase will change is meaningful. For

instance, if dn A moles of species A, is added or removed to a large amount of a phase at
constant temperature and pressure, the size of the system will increase by dn A . Thus, the
total free energy of the system will increase by AG’.

More importantly, if dn A is small enough, AG’ will be produced in proportion to

the amount of A that is added. Following this rationale, we can get the equation shown

as
A02 ,u Adn A Equation 37
where ,u A is called the partial molar free energy of A, or the chemical potential of A in

the phase, with the proportionality set as constant. Since ,u A is determined by the

49

composition of the phase, dn A must be small enough not to influence the change in the

composition.

6G'

2 _ E nation 38
#A (anA)T,P,nB C1

The symbol G’ represents the Gibbs free energy, which could simply refer to the whole
system. Usually, since the symbol G is used to signify the molar free energy, it refers to

the fact that it is independent of the size of the system.

1.4 Growth Mechanism

1.4.1 Vapor-Liquid-Solid (VLS) Growth Mechanism
Introduction

Normally, the nanowire synthesis involves a vapor phase, such as their growth by
means of chemical vapor deposition, physical vapor deposition or pulsed laser deposition.
Among the various nanowire growth methods, the vapor-liquid-solid (V LS) mechanism,
which was proposed by Wagner et a1. as for the growth of single-crystal silicon whiskers,
is particular popular [35]. Moreover, Givargizov has developed the growth model and
delineated it within the framework of kinetics [3 6]. The VLS reaction was adopted by
Westwater to grow silicon nanowires by using silane as the Si source gas and Au as the
mediating catalyst [37].

In the VLS growth mechanism, a liquid metal region is formed by means of raising
the temperature of the metal and semiconductor above the eutectic temperature. The

liquid metal region also serves as an energetically preferred location supporting the

50

absorption of gas-phase reactants, which is related to the high sticking coefficient of the
molten alloy. At least two basic mechanisms of the eutectic alloy formation can be
emphasized currently [38]. The first is the formation of the eutectic alloy. Before the
formation of the eutectic alloy, the intermediate layer of supersaturated solid solution is
formed which is followed by the growing defects concentration into these layers and the
increase of the free energy of the system. Actually, the increase of the free energy is the
reason why the eutectic systems are easily melting. Different from the first mechanism,
there is no formation of solid solution in the second mechanism. Based on the
application of the adhesion theory and the mechanism of contact melting, the developing
surface force between the atoms in the two surfaces takes place in micro-regions of
physical contact between the conjugated crystal phases, and the transition epitaxial zone
can be formed without the formation of solid solution. Actually, the formation of these
transition epitaxial zones would result in the subsequent reduction of differences in
lattices parameters between solid phases at the interface region under the condition of the
stress present at the interface. In terms of the metal addition to the target, Wagner and
Ellis have made great contributions. According to the study of Wagner and Ellis, since
epitaxial structures are defined in location and diameter by the metal catalyst, the addition
of such metals would catalyze their growth. The increasing gathering of the
semiconductor encourages the achievement of the supersaturation condition in the metal-
semiconductor alloy, which prompts the semiconductor phase to detach from the alloy
and to form the next epitaxial layer at the base of the liquid. Then the cluster

supersaturates and grows into one-dimensional wire of the material, under which the

51

alloy cluster is produced. Experimental parameters, such as temperature and pressure,

and the nature of the metal catalyst determine the formation of the produced wire.

 

1eoo_................---.
1400'
1200}

1 000

  
 

800 I

Temperature (°C)

 

 

 

 

(AuSi)l + SiS 1
600 ' .V 1
400 ' -‘
. | . .
s— Au Sr—t
200 A A A l A A A l A A A l A A J l A A A
0 20 40 60 80 100
Atomic % Si

Figure 15. Phase diagram for the gold-silicon system.

This phase diagram plots the atomic percentage of Au and Silicon along the X-axis
and temperature along the Y-axis. In the Region 1, since the temperature in the Au-Si
system is set below the eutectic temperature (~3 60° C), which is the minimum melting
temperature of any possible alloy, no molten ally is formed. The reason is that eutectic
implies that an alloy has a sharp melting point. Eutectic, or eutectic mixture, is a mixture
of two or more elements which has a lower melting point than any of its components.
Moreover, one characteristic of eutectic corresponds with the possibility that specific
composition has the lowest melting point of an alloy. For example, if an AuSi alloy has

19.5 atomic % of Si and 80.5 % of Au, it will melt at T=360° C. On the other hand, in

solid phases composed of solid forms of pure Au and Si, solid up to 1063° C and 1412° C,

52

respectively, Au and Si will coexist as solids in every atomic percentage of these two
elements at the interface between them.

In the Region II, the solid form of gold and silicon starts to form liquid alloy if the
following conditions are met: the temperature is higher than eutectic temperature
(~360°C) and the weight percent concentration of silicon is raised to a certain level,
usually 82%. When these conditions are present, a molten alloy would be formed.
Region H is a biphasic region; in which the solid gold and Au-Si molten alloy coexist. If
more volume of the silicon is added into the system, the molten alloy will absorb more
solid gold, which boosts the size of the Au-Si alloy droplet. Due to the characteristic of
high sticking coefficient of the molten alloy, in which a large proportion of added atoms
are adhering to its substrate, a liquid metal region is formed and provides a place
attracting the absorbed gas-phase reactants.

In the Region 111, due to the further absorption of atoms from the vapor phase, the
silicon percentage will increase with time at certain temperature, which results in the
increase of the gold and silicon droplet. The alloy remains liquid within this region while
no solid phase exists. Nonetheless, if there are more and more silicon atoms added into
the system, the whole system is pushed to fall into the Region IV, in which the solid
silicon begins to emerge in the system with other variables, including temperature and
pressure, remaining constant.

In the Region IV, the gold and silicon alloy remains in liquid state. With the
atomic percentage of silicon increasing, the silicon crystal starts to grow. As the silicon
concentration increases over 50 atomic percent at 1000 ° C, the silicon will begin drop

out of the molten alloy as a solid, lowering the concentration of silicon in the melt, and

53

moving to the left on the phase diagram. Through this mechanism, the diameter of the
resulting nanowire depends on the size of the Au-Si droplet. As more silicon added into
the system the alloy again supersaturates and more silicon solidifies from the melt in a
repetitive manner to continue the nanowire growth. The size of the Au-Si liquid alloy is
also determined by the temperature of the substrate. The higher the substrate temperature,
the larger the Au-Si liquid alloy becomes, which influences the formation of bigger
diameter nanowires.

As the temperature of the system is raised above the eutectic point (~ 360° C in
Figure 15), a molten alloy is formed. For proper growth of the nanowires it is desired to
have the impinging silicon adatoms to preferentially stick at the molten alloy locations.
Since the sticking coefficient of a liquid is very high, conditions for reducing the sticking
coefficient at all other locations (silicon on silicon) are sought. As described previously,
this condition can be met by raising the temperature of the substrate, and thus enhancing
the surface diffusion of the adatoms. As the alloy supersaturates with silicon, the solid
silicon drops out of the molten alloy to form the next layer of the nanowire, and the

molten alloy is pushed to the top of the wire.

54

Au particles

A7 \

 

 

 

 

Figure 16. Gold droplets form an alloy with the silicon. The alloy has a eutectic point
near 360°C, above which it is molten.

By using VLS technique, controlling the metal droplet location the location of nanowires
can also be controlled. We initially investigated different patterned shapes to see if it was
possible to control the location of the metal droplets to specific locations within the

pattern.

Kinetics

In kinetic studies the crystal growth rate is usually measured as a function of the
driving force, such as supersaturation. For the growth of whisker, the discovery of the
dependent relationship between the axial growth rate and the whisker diameter is
especially enlightening. A method has been proposed to the sort out the determining

factors of the supersaturation and those of the kinetic coefficients of crystallization, under

55

the condition that the specific energy of the prismatic whisker surfaces is already
specified. This method is based on the application of the Gibbs-Thomson equation and
certain kinetic laws of growth.

Since the vapor pressure and solubility of Si increase with the diminishing diameter
of the whisker, the growth rate decreases under the influence of the decreasing
supersaturation. According to Gibbs-Thomson’s contention, the decrease of the

supersaturation as a function of the whisker diameter can be given as

An = A,u0 — 4Qa/ d Equation 39
where A,u is the valuable difference between the chemical potentials of Si in the vapor
phase and those in the whisker, Ago is the similar difference at a plane boundary

((1 —> 00), a is the particular free energy of the whisker surface, and Q is the atomic
volume of Si.

Since the dependence of the growth rate V on the supersaturation A,u / kT , where k
and T have the usual implications, is unidentified in advance, experimental data must
provide related information effectively. For the singular faces, the dependence is not
linear and, however, in many cases is of nth power

V ~ (Ap/ kT)" Equation 40
This proportionality can be expressed as an accurate equation:

V = b(A,u/ kT)" Equation 41
where b is the coefficient independent of the supersaturation. Based on this equation, it is

concluded that this independence can be used as a criterion determining n from

experiments. Furthermore, substituting A,u , we can obtain

56

Vl/n = Allo bl/n _ 4Qab1/n _}_

Equation 42
kT kT d

Vl/n

which is equivalent to a linear dependence of opposed to l/d.

1.4.2 Solid-liquid—solid growth mechanism

Solid-liquid-solid is corresponding to the mechanism making use of the sequence
of vapor —liquid —solid. In the case of Ni thin film on Si (111), due to the relatively high
solubility of Si in SizNi eutectic alloy, more Si atoms will diffuse through the solid
(substrate)-liquid interface into the liquid phase, such as Ni-Si droplets. A second liquid-
solid (nanowire) interface will take shape when the liquid phase becomes supersaturated
owing to thermal or compositional fluctuations, which leads to the growth of SiNWs.
Since this growth process involves the phase changes of solid-liquid-solid, it is named as
the SL8 growth, which is actually equivalent to the VLS mechanism.

From a kinetic perspective, due to the fluctuation in Si—Ni liquid droplets, the
concentration gradient and the supersaturation are driving forces for this uninterrupted
diffusion of Si atoms from the substrate through the substrate-liquid interface to the
liquid droplets, and then through the liquid droplets-nanowires to load themselves onto
nanowires. On the other hand, the carrier gas Ar2/Hz will run into the surface of the
semisphere-shaped Si-Ni liquid droplets, and will swap energy and momentum with
atoms at the surface of the Si-Ni nanoparticles, stimulating the occurrence of overcooling
at the surface of the droplets. This overcooling is vital to commence a favored

unidirectional growth of a-SiNWs.

57

1.4.3 Oxide assisted growth

Introduction

Supported by the result of the experiment which showed that the metal catalyst was
not observed in Si nanowires even when metals were mixed in the target, a different
model has been proposed. Moreover, it was discovered that metal was not a necessary
component of the Si nanowire synthesis through laser ablation. Instead, Si02 was the
special and effective catalyst, which could fundamentally improve the growth of Si
nanowires [39].

In this oxide-assisted growth (OAG) method, oxides, rather than metals, play an
important role in inducing the nucleation and the growth of nanowires [40]. Because of
the metal-free property, OAG not only supports large-quantity production of nanowires
with controllability, but also yields nanowires characteristic of metal-free high purity.

The understanding of the OAG mechanism still has not developed into mature
stage. Nonetheless, based on the success of the growth of nanowires from different

materials, this growth mechanism deserved better attention.

Methods

S. T. Lee etc. [40] contended that the grth mechanism is assisted by silicon oxide.

Specially, the vapor phase of SixO (x>1) generated through thermal evaporation is the

main determinant. It is assumed that the nucleation of nanoparticles generally takes place

58

at the substrate by means of different decompositions of silicon oxide at the relatively
low temperature of 930°C as shown below.
Si x0 —> Si x_1 + Si0(x > 1) Equation 43
and
2Si02 —9 Si + Si02 Equation 44
These decompositions lead to the precipitation of silicon nanoparticles, such as the
nuclei of Si nanowires, which are covered with shells of silicon oxide as observed.

Essentially, a large amount of Si nanowires have been synthesized through thermal
evaporation by mixing silicon and Si02 powder. In order to form Si nanoparticles, or

nuclei of nanowires, Si oxide vapor generated from the powder mixture first condensed
on the substrate and then decomposed. A Si nanowire nucleus was composed of a
polycrystalline Si core with a high density of defects and a silicon oxide shell. The
grth mechanism was proposed to be properly associated with the imperfect structure of

Si crystal cores and SiO.

During evaporation, since SixO vapor was recurrently generated, the nucleation

could occur with different crystalline directions, either on the side surfaces or the tips of
nanowires. If the nucleation takes place on the side surfaces, the forking of the
nanowires is often observed. On the other hand, if it takes place on the tips of nanowires,
a re-nucleation will occur. Especially, since the nuclei formed on the tips in an adverse
growth direction could not grow fast, a re-nucleation shall inevitably occur again, which

will lead to the formation of nanoparticle chains.

59

Nanowires
Nanowires of different materials utilizing this method have been grown primarily

either by thermal evaporation or laser ablation. For example, Si nanowires have been
synthesized through thermal evaporation by mixing silicon and Si02 powder [39,40]. In
a similar fashion, the synthesis of MgO nanorods has been achieved by mixing Mg and
CuO powder [41], and ZnO nanowires are the production by combining ZnO and SnOz,

[42].

1.5 Vacuum Sciences

The kinetic theory of Gasses

The kinetic theory of gases was initially developed by Maxwell and was then
investigated extensively by many following investigators. The main contribution of this
theory lies in its useful explanations for both the bulk properties of a gas and the transport
properties. The former includes elements such as the pressure, molecular impingement
rate, and the distribution of molecular velocities, while the latter involves viscosity,
thermal conductivity, and diffusion. This division is based on the employment of
dynamics of the motion of individual molecules.

Essentially, the basis of the kinetic theory of gases is in close relation to two
primary assumptions. The first premise hypothesizes that molecules of a gas are in
constant motion, and this motion is significantly connected with macroscopic properties
known as the temperature and pressure, distinguishing the state of the gas within a

specific small region. The second hypothesis is that gas is composed of small particles.

60

In the gaseous state, these particles are monatomic or polyatomic molecules at moderate
temperatures while at higher temperatures, they may be fully unconnected and
transformed into atoms or even into positive ions and electrons in order to give shape to
plasma.

It is assumed that the core of a mass of the molecule is moving toward the same
direction with a constant velocity between collisions with other molecules. The forces
between molecules are barely noticed only when the distance between the molecular
centers and molecules is reduced to the range known as the mean molecular diameter.
Velocity is viewed as a vector which must be measured in relation to some rigid
reference frame. The magnitude of the velocity vector is defined as the speed of the
molecule. Since kinetic energy is proportionally related to the square of the molecular
velocity, it is a scalar quantity.

By definition, temperature is quantity proportional to the average kinetic energy of
the translation of the particles in a reference frame, which is moving with any fluid flow
within a specific small region. In the case of monatomic molecules, the effect of
increased temperature is mainly discerned through amplified translational energy of the
molecules. On the other hand, as for diatomic and polyatomic molecules, increased
rotational energy of the molecule approximately one or more axes and increased
vibrational energy of the component atoms regarding mean positions of equilibrium are
both originated from increased temperatures by means of intermolecular collisions. The

assumptions involved in this model are as follows:

1 Any fixed volume of gas incorporates a great deal of molecules.

61

2 The influence of molecules only emerges when molecules collide. Between
collisions with each other or with the container walls, they travel in the straight
direction.

3 All directions of molecular velocities are of identical possibility.

4 Collisions of one molecule with another or with surfaces that are positioned open to
the gas are entirely flexible. All surfaces are regarded to be smooth.

5 Although there are any possible magnitudes in relation to the speed of a given
molecule, the number of molecules with speeds in a specific range is constant with

time.

1.5.1 Rate at Which Molecules Strike A Surface.

In order to study the role of pressure in the kinetic theory, the discussion about the
molecular impingement rate is critical. The impingement rate refers to the number of gas
molecules that hit a specified area in the system per unit time. The molecular

impingement is adopted to formulate pressure in the kinetic theory. It is usually stated in

units of molecules / sec— cm 2 .
As described in the Assumption 3, in Figure 17 all directions of molecular
velocities are consistently probable, which considers the same probability of all directions

of molecular velocities.

62

‘
7

Figure 17. All directions of molecular velocities are consistently probable.

A vector always represents the direction and the magnitude of its velocity. We can

relocate all vectors to a common origin and to a sphere of radius built close to this origin,

Q

shown as follows:

Figure 18. A vector always represents the direction and the magnitude of its velocity.

63

In this structural expression, the velocity vectors cross the surface of the sphere at
many points, equivalent to the number of molecules intersected in the system. The
average number of intersections per unit area is expressed as:

A,— = 1— Equation 45
S 4m2

Hence, the number of intersections on any part of the surface of the sphere, (15, is

dN = {—N2—}1S Equation 46
4m

which is independent of where (15 is located on the sphere.
Expanding this formulation further and defining a specific surface element, dS, in
terms of the polar coordinates 9 , ¢ , an equation can be developed as follows:
d5 = r2 sin wadas Equation 47
representing the number of molecules whose velocity vectors intersect this constituent of

area. This indicates the number of molecules whose velocities are directed within a polar
angle from 6 to 6 + (16’ and azimuthal angle from ¢ to ¢ + d¢ , which is called d 2N“,

is where
2 N . .
d N 9 ¢ = [-—)(srn wow) Equation 48
’ 47:

The result mentioned above can be used to identify the molecular impingement rate,
the rate at which molecules from the gas run into a unit area of the surface that are
exposed to the gas, either the container surface or any area made within the bulk of the

gas. Consequently, this could be realized by assuming a random element, (IA, of any

64

surface exposed to the gas, and making a line normal to the plane of dA, and a reference

plane posed at a 90 degree angle against the normal.

3
N945”. = d n9,¢,vdv

 

 

vdn Equation 49
= dAdt( v )sin (9cos 6d6d¢
47:
N0¢v l
I = ’ ’ = — dn sin6coséld3d E uation 50
9”” dAdt in} " ¢ q

By integrating the equation above with respect to 9 from zero to rt/2 and respect to (p from

zero to Zn, yielding

 

1,, = (-—1—)vng (27:)(1) Equation 51
47: 2
So the total impingement rate for all molecules, irrespective of their speeds, is
1 =1/4[vdn, Equation 52
We can rewrite the equation in terms of to as
19 ,5 v = (ijvdnv cos (5th Equation 53
’ ’ 47:
The total impingement rate for all velocities per unit solid angle is thus
I a, = (——l—) cos (9! vdnv Equation 54
47:
Ia, = l 715008 6 Equation 55
4 7:

This is known as the Cosine Law or Cosine Rule.

1.5.2 Ideal Gas Law Pressure
Understanding of the impingement rate may be beneficial to developing an
expression for the pressure influencing the container walls originated from the collisions

of the gas molecules with the walls. The pressure refers to either the force exerted per

65

unit area by the gas molecules on the walls of the container or the rate at which the
momentum is transmitted to the unit area of wall surface.
p=nK T Equation 56
According to the kinetic theory, due to the impact of molecules on the confined
walls, a gas yields a pressure on these walls. Since during the process of the gas exerting
pressure on the inactive solid wall of the enclosure, no energy is lost from the gas, it
appears that each molecule is projected back from the wall with the same speed as that
with which it strikes, but in the reverse direction regarding the normal. In other words,
these resulting impacts are absolutely elastic.
It is hypothesized that a molecule of a mass m moves toward a smooth wall surface
belonging to the x, y plane with the velocity component, vz, vertical against the wall.
Since the rebounding rate of a molecule is consistent, the change of momentum per

impact is 2mvz. Then there are v molecules striking the unit area, which results in
2mvzv. Nonetheless, the pressure, P, on a wall is known as the rate at which the
momentum is lent to a unit area of surface. Therefore,

ZmZ vzv = P Equation 57
where the summation is total values of vzv and all molecules are presumed to have the

same mass m.

The final step is to calculate v. Among all of the molecules within a volume AV

stretching out from a small area, AxAy, of the wall by a distance Ivzdtl, where dt is a

short interval of time. In equilibrium, only a half of the molecules will be moving with

velocity components 172 toward the wall. Let 71, represent the number of molecules per

66

unit volume in the gas within AV that have a velocity component of either v, or —v,.
Then the flux rate against the wall will be v = n sz /2 , and 2m ZVZV = P becomes
P = mznzvg‘ Equation 58

The sum of the speed of a molecule with velocity components is the square root of the

quantity
v 2 VI + v

y + v, Equation 59

Defining the averages

<v£ > = (Z nxvg)/n, Equation 60
(vi) = (Znyvifln,

2 2
<1», > = (anvz )/n,
where n is the total number of molecules per unit volume without considering velocity

components. The following symmetry could be derived accordingly (v2 > = (v2) = <v2>

x y z

and
2 _ 2 _ 2 -
<v >— 3<vz> — 3(znzvz)/n Equation 61
Then from P = m anvz2 , the basic equation for the pressure is obtained as
l 2 .
P --= 3 nm<v >. Equation 62

Since

mn = p Equation 63

67

which indicates that, at constant temperature, their pressure is altered in correspondence
with the density, or inversely with the volume. This is known as Boyle ’5 law.

It is known that if two different gases, originally at the same temperature, are
mixed, no change in temperature will occur. This result is proved to be true independent
of the control of volumes. As a result, the average kinetic energy of molecules must be
the same for all gases at any given temperature, and similarly, the rate of increase with
temperature must be identical for all gases. The temperature could be defined in terms of

the average kinetic energy per molecule, which is shown as the following relation
2 _ 2 _ 2 _ l 2 __ l .
(manvx /2n) —(mZ nyvy /2n) — (manvz /2n) — Em v -- 2kT Equation 64

For each of the three degrees of freedom of translational motion, where T is the absolute
temperature, and k is a universal constant, or the Boltzmann’s constant. The total mean

translational energy is then
I 2 2 3 .
Em<v > = —-mv, = 2kT Equation 65

where v, is identified as the root-mean —square velocity. The total kinetic energy of

molecules in a volume V will be
I 2 3 .
-2- mn<v >V = -2—PV Equatron 66

Following Boyles’s law, it could be expressed in the form
P = nkT Equation 67
which is known as the Charles’ law.
In order to evaluate the behavior of gas in phase reactions, gaseous beams, heat

flows, and the surface bombardment, it is imperative to develop a model of gas

68

molecules. The kinetic theory of gases views gaseous molecules as hard spheres. The
probability distribution of velocities is based on the application of the Maxwell speed
distribution. For a simple monotonic gas, the probability of a molecule possessing a

certain speed is given by
P(v) = 47z[—’—n—-]3/ 2 vze_mv2 /2kT Equation 68
27dcT

where m is the mass of the molecule, k is Boltzmann’s constant, v is the magnitude of the
velocity, and T is the temperature in Kelvin. The speed, or the average magnitude of the

velocity, is given by

M E = I: vP(v)dv = 1/8_k_7; Equation 69
7071

The standard component of the velocity in any direction is given by
vx = 17 = 17, = — Equation 70

and the root mean square velocity is given by

vrms = 4%]: Equation 71

Since the direction of the thermal velocity is random, if there are no applied forces
on the outside, the average velocity is zero. If a small pressure gradient is forced upon
the gas, a net macroscopic flow will emerge, from high to low pressure. On the
microscopic level, this excited flow velocity vector is imparted upon the much larger
thermal velocity. Although individual atoms may be moving against the pressure
gradient at any second, normally the flow will travel from high to low pressure.

Gas phase collision is one of the crucial mechanisms specified for gases changing

velocity at atmospheric pressure. It is hypothesized that a molecule of diameter d is

69

moving arbitrarily in the gas. If another molecule of the same type is within a distance d

of the estimated path of the first atom, a collision takes place. Then, the molecule has a

collision cross section of 72:12 , which, by the simple hard sphere approximation, is four
times the definite cross section of a single molecule. The probability that a collision will

be initiated over a distance L is given by
P = L711 2 n Equation 72
where n is the number of gas molecules per unit volume. Setting P z 1, the average

distance between collisions can be estimated as

l

 

,1 z 2 Equation 73
7rd n
These collisions are usually called the mean free path 71.
A more accurate rigor statistical treatment gives [43]:
2 ~ ——1—— Equation 74
find 2n

Since n is not uniformly defined, it must be calculated from a macroscopic equation of
state. Thus, the optimal gas law is frequently used

71 = fl = 3 Equation 75
V kT

where P is the pressure of the chamber. Combining equations,

kT

A =—
fizzdzp

Equation 76

7O

1.6 Gas Flow and Conductance

The equations are used to calculate pumping speeds and gas flows. In measuring
the flow of a simple liquid, such as water, a common unit might be the volumetric flow
rate, or gallons per hour. Although volumetric flow rate is usually used in recounting gas
flows, especially in delineating pumping speeds, due to the fact that gases are much more
compressible than liquids, the use of the gas flow in this approach incurs problems. In
order to avoid this problem, the throughput is used to express the amount of gas flowing
through a system. The mass of a gas in some volume V is

G = pV Equation 77
where p is the mass density (m x n ). Then the mass flow rate is

_dG

— — Equation 78
dt

qm

The throughput of a gas Q, which has units of pressure-volume/time, is given by

Q = q,,, f— Equation 79
,0

Generally, gas flows are measured in terms of a standard volume, which refers to
the condition that the volume of an equivalent amount of gas would occupy at 0°C and 1
atm of pressure. While one mole of gas occupies 22.4 L at standard conditions, one

standard liter is 1/22.4 moles. Pumps are usually defined in terms of pumping speed, S p

dV
Sp =—Q-=——p Equation 80
Pp dt

where PI, is the inlet pump pressure. For instance, a pump rated for 1000 l/ min (slm),

will pump 1000 slm at 1 atm inlet pressure. If the inlet pressure is 0.1 atm and the

71

pumping speed at this pressure stays at 1000 1/min, it can be concluded the maximum gas
flow that this same pump could allow is 100 slm. Furthermore, the pumping speed
normally is not a constant. Instead, it is closely related to the gas being pumped and the

inlet pressure.

72

Chapter 2 Growth of III-V Semiconductor Nanowires (GaAs)

Studies of one-dimensional (1D) nanoscale materials have attracted substantial
research interest during the past decade concerning their specific physical properties and
potential applications. III-V group semiconductors are of particular interest for their
electronic and optoelectronic properties. Generally, efforts have focused on the synthesis
and the characterization of III-V semiconductor colloids and nanocrystals [44], while
fabrication techniques have encompassed the evaporation, the laser-assisted catalytic
growth (V LS, at above 680° C) [45,46], and the oxide-assisted fabrication (~750° C)

[47].

2.1 Introduction

A series of experiments were designed to investigate the growth of gallium
arsenide nanowires. In the course of the experiment, however, it was found that the
nanowires formed were deficient in arsenic. Furthermore, oxygen was readily
incorporated into the nanowires. A set of experiments was then investigated for the
growth of gallium oxide nanowires from the GaAs source material. It is well known that
the deposition of gold onto III-V substrates can catalyze a series of chemical reaction at
higher temperatures. Since the Pauling electronegativity of Au is higher than that of the
elements of Group V, Au can act as an oxidizing agent to displace the less
electronegative elements of Group V by combining the highly electropositive Group III
metals. In these experiments, GaAs pieces were first rinsed with acetone and methanol.
Then gold was deposited either on the substrate at room temperature via sputter

deposition, followed by an annealing step, or onto the substrate at a high temperature of

73

deposition via laser ablation. Various temperatures were investigated with results shown

as follows.

2.2 Experiments
Experiment #1 :

A gold layer was sputter deposited at room temperature to a thickness of 20 nm,
and the sample was subsequently annealed to 970° C. The GaAs target was placed
approximately 4 cm from the substrate and ablated by pulsed laser deposition (PLD) for 1
minute and 30 seconds. The pulse was 10 Hz at 500 m].

The experiment was conducted within a quartz tube (outer diameter, 90 mm;
length, 80 cm) heated by a horizontal tube furnace. The sample was positioned at the
center of the quartz boat, which was placed at the center of a quartz tube which was
evacuated to about 30 mTorr. A background gas of argon was introduced into the
chamber at a flow rate of 200 standard cubic centimeters per minute (sccm), which was
maintained throughout the whole reaction period.

All the samples were examined and photographed using the JEOL Scanning
Electron Microscope (SEM) and EDS (Energy Dispersive X—Ray Spectroscopy) was

concurrently carried out to determine the elements present in the sample.

74

 

Figure 19. Nanowires grown from a piece of GaAs. The scale bar is 500 nm.

In the SEM image of Figure 19, the upper left white area is the GaAs chunk. This
GaAs sample was placed in front of a separate gallium arsenide target which was ablated
by a KrF excimer laser for 1 minute and 30 seconds, at 10 Hz and 500 ml. The
nanowires shown in the SEM image have diameters ranging from 30—90 nm and their
lengths are a few microns. Most of the nanowires appear to be straight forming to a sharp
tip. A characteristic of the vapor-liquid-solid growth mechanism is the formation of an
alloy ball at the tip of the nanowires. The absence of this characteristic ball suggests that
either a different growth mechanism is taking place such as the oxide assisted growth, or

the gold ball is being incorporated into the nanowire during system cool down.

75

 

Figure 20. The nanowires grew from the GaAs chunk. The scale bar is 511m.

76

 

Figure 21. These nanowires were a few microns in length. The scale bar is 2 pm.

Figure 21 is the SEM image of gallium oxide nanowires grew from the gallium
arsenide chunk. These nanowires were a few microns in length and were primarily found
on the edge of the sample. The upper and right white areas are the gallium arsenide

chunk. These nanowires were 2-3 microns in length.

77

 

Figure 22. The nanowires grew from the edge. The scale bar is 211m.

In Figure 22, the SEM image shows gallium oxide nanowires grew from the GaAs chunk.
These nanowires were a few microns in length and their growth mainly took place along
the edges. The SEM image also shows the presence of some triangular shape structures

scattered throughout the image.

78

 

 

(rm)

 

 

 

Figure 23. Histogram showing the diameter of nanowires. Total sample size: 42.
Average Diameter: 65 nm

79

 

Lenghofnarowires

10th

 

O

23 24 25 26 27 28 29 3 31
(alabaster)

 

 

 

Figure 24. Histogram showing the height distribution of nanowires. Total sample size :
42. Average length of the nanowire is 2.7 um.

80

 

 

 

 

Figure 25. The EDS analysis of nanowire

81

8.2: See .28
-.. ... ... ... ... 83 80.0 80.0 82 08.9 32
as .1 AS 3.0.: m3 .1 m3 :2 83 83 E: 83 x2
... ... ... ..- ... 83 80.0 08.0 83 82 :0
3o .1 3.8 3.3.. 3.3.. 8.: 33 $3 82 83 98.9 x3
:3. one 8.3+ 8.0.: m3 2% 83 83 and 83 x2
.83.. at: 8.3+ 3.3.. 8.2 so.” 83 23 83 03° v5
85 5 £0.22, 55 2:
.a :52 as 5oz .552 $2925 «.5925 “ZN u < N came. 225m.

33 23% 95:58
o_. m m x. o m v m m r

w< mo ..n . “ ..a 8m

WW .n a coo?

.m .H w oomr

«my w ooou

_ w comm

_ w coon

no

Figure 25 shows the EDS analysis taken on the nanowires shown in Figure 21. The
EDS analysis shows the components of nanowires were gallium and oxygen. The
histograms in Figure 23 and Figure 24 show the diameter and length distributions of the
nanowires. The average dimensions of the grown nanowires were about 65 nm in
diameter and 2.7 pm in length. These nanowires were not distributed evenly and
consistently on the surface, but tended to grow in clustered areas.

A similar experiment was conducted with the same temperature and pressure
conditions. The only difference was the substrate was placed behind the GaAs target.
The results were similar to those reported above showing that the location of the GaAs

target did not affect the nanowires growth.

Experiment #2:

Experiment #1 described the first evidence of gallium oxide nanowires growth on
the GaAs substrate. In Experiment #2, nanowires dependence on annealing time is
reported. A gold layer was sputter deposited at room temperature to a thickness of
20 nm, and the sample was subsequently annealed at 970° C for 40 rrrinutes. Two

samples of different sizes were placed inside the chamber.

82

 

Figure 26. Under these fabrication conditions, a very high density of nanowires
formed. The scale bar is 20 um.
The SEM images in Figure 26 and Figure 27 show nanowires of several microns in
length with an average diameter of approximately 80 nm. The size of this GaAs substrate
was approximately 0.2 mm by 0.2 mm. The diameters of these nanowires were more

uniform than for the shorter growth times of the previous section.

83

 

Figure 27. The diameters of these nanowires were uniform. The scale bar is 5 run.

 

Figure 28. This zoom-out image illustrates that nanowires connected with each other

like structure. Nonetheless, there was no gallium

in order to form a wool

arsenide chunk left inside this structure.

84

The SEM images in Figure 29 and 30 were taken from a larger substrate of gallium
arsenide (approximately 0.8 mm by 0.8 mm). The nanowires formed a thick layer on top
of the piece of GaAs. From the SEM image in Figure 29, the GaAs chunk could be
observed to be placed on the left side while a thick layer of nanowires were on the top.

These nanowires were a few microns in length.

 

Figure 29. The nanowires formed a thick layer on top of the GaAs chunk.

85

 

Figure 30. The SEM image shows nanowires. In addition, the same image illustrates
that the gallium arsenide surface was underneath the layer of nanowires

and there were some nanowires left on the surface. The length of
nanowires was tens of microns. The scale bar is 5 11m.

The EDS analysis in Figure 31 shows the nanowires consisted primarily of gallium

and oxygen.

The histogram in Figure 32 shows the diameter distribution of the grown gallium

oxide wires. The average dimension of the grown nanowires was about 81 nm in

diameter.

86

 

 

 

 

or m A o m w m N F
maxim . 3....3. . a a ,1 11.1--.. 1.1... a .._ mania.
8 a a. . com
8 Wm .- 83
A T 82.
T ooom
mo
88. 8.8_ A .38
... ... ... .. 1 88.8 88.8 88 88.8 88.8 A :4
8.8.: 83 8.3.. 8.8.: 8.4 83 83 83 :2 88 _ v.8
... ... r .. .. 88.8 83 88.8 88.8 88.8 A 88
8.8.: 8.8 8:. 8:. 3.8 83 88 88 83 88.8 A 80
8:. 8.8 8.8.: 8.3.. 8.8 83 88.8 8: 88 N88 _ ca
8.2.. 8.8 8.3.. 8.3.. 83 :3 83 8; :8 8.8. v5
atm :m oi; Bum _ 0::
8 .88 a. 82 .882 a. .88; as .88; as u < N 085. 8888

Figure 31. EDS analysis on the grown gallium oxide nanowires.

The density of nanowires was found to be significantly higher in Experiment #2

compared to the results of Experiment #1. These nanowires were distributed evenly and

consistently on the surface of the piece of GaAs.

87

In Experiment #1, the sample stayed in 970° C for 90 seconds while the laser
ablated the GaAs target. The results showed these nanowires were not distributed evenly
on the surface. In Experiment #2, the GaAs was annealed at 970° C for 40 minutes, and

showed a significant increase in the density and distribution of nanowires.

 

Total sample: 40

Count

 

50 60 70 80 90 100 1 10 120 130

Dlametor (nm)

 

 

 

Figure 32. Histogram showing the diameter distribution of the nanowires. The average
diameter is 81 run.

Experiment #3:
To study the affect of the annealing temperature, the same experimental conditions

were used except the annealing temperature. A gold layer was sputter deposited at room

temperature to a thickness of 20 nm, and the sample was subsequently annealed at 800° C

for 40 minutes.

88

 

V. [10711121.

Figure 33. The SEM image shows the nanowires grew on the piece of gallium
arsenide. These nanowires were a few microns in length. The scale bar is

10 um.

Figure 34 shows a broader view indicating the uniform coverage of the substrate by
nanowires. The shape in the middle of the image was formed after picking up the sample
with tweezers. It clearly indicates that the piece of gallium arsenide was covered with a
thick layer of nanowires with uniform length. Nonetheless, not all of the germanium

arsenic turned into nanowires.

89

 

Figure 34. The SEM image shows the nanowires grew on top of gallium arsenide
chunk

90

 

Figure 35. The SEM image show the nanowires grew on top of gallium arsenide
chunk. The scale bar is 2 11m.

The diameter of nanowires was quite uniform as indicated in the histogram shown
in Figure 18. The average dimension of the grown nanowires was about 78.5 nm in

diameter.

91

 

Count

60

 

70

Total samle: 40

 

80

(nm)

100

 

 

Figure 36. Histogram showing the diameter of nanowires. Total sample size: 40.

Comparing Experiment #3 with Experiment #2, we do not find significant

differences in the results; however below 700° C annealing, no growth was observed.

The EDS analysis in Figure 37 shows gallium and oxygen as the predominant

components of the nanowires. Arsenic also appeared in the analysis; however, its amount

was extremely small compared with that of gallium and oxygen.

92

 

 

 

 

8.8_ 8.88 88
... .. ... ... ... 88.8 88 88 88.8 88 8
8.8.. 8 8.8.: 8.8.. 3.8 83 83 83 8: 88 x8
... ... ... ... .. 88.8 88 88 88 8.8 do
8.3.. :8 8.1.. 8:. 8.: 83 88 88.8 83 88 v80
8:. 8.8 8.3.. 8.8.: 8.8 88 88.8 8:. :8 88 v5
55 :m 88.25 Loam oz:
8 88 a. .88 .88 .8888 8888 8N a. < N 88.x 888
or m m N. o m e m N 8
oo - 8m
«0
.- coo?
83
.3

Figure 37. EDS analysis on the grown gallium oxide nanowires

93

 

Figure 38. The SEM image of a piece of GaAs. No gold was deposited on this sample
before growth.

Figure 38 is a SEM image of GaAs which did not have a layer of gold deposited on
it. This piece of GaAs was placed next to a separate sample that did have a layer of
deposited gold. There were still some nanowires on the surface; however, the density of

nanowires is significantly smaller, and the nanowire clusters are loosely bound to the

surface of the substrate.

94

TEM image analysis

 

Figure 39. TEM image shows nanowire grown from GaAs

95

 

Figure 40. TEM image shows nanowire grown from GaAs

96

 

Figure 41. TEM image shows nanowire grown from GaAs

97

'3 past r' rim 1

 

 

 

I . vjfi—r vvvvvvv I'vv—V vvvvvv ‘ v 1'. I vvvvvvv ‘v—r

 

 

 

0 2 4 8 8 1 0
uII Scale 8255 cts Cursor: 0.000 kev
Element mht% Atomic%
O K 31.07 65.46
Cu K 26.00 13.79
Ga K 42.93 20.75

Totals 100.00

Figure 42. EDS analysis of nanowires

The TEM images show a nanowire diameter of approximately 70 nm. The atomic planes
shown in Figure 41 have a spacing of approximately 3.46 A. In Figure 42, the EDS

analysis shows gallium and oxygen in a ratio of approximately 1:3 respectively.

98

2.3 Discussion

Current studies of Au/III-V interfaces have pointed out that the “interdiffusion”
between the metal and semiconductor occurred after samples were heated to relatively
low temperatures, such as 573 K and other degrees of same effect. Meanwhile, the fact
that this diffusion is essentially the result of a chemical reaction has been identified. The
general pattern derived from these studies is that heating a Au film/III-V substrate system
will lead to the decomposition of the semiconductor near the interface and the release of
the volatile elemental Group V via the grain boundaries of the metal film [48]. The study
of the Au-Ga-As system indicated that the bulk Au and GaAs are not thermodynamically
stable with respect to one another [49] since Pauling electronegativity of Au is higher
than that of the element of Group V. Thus, Au could act as an oxidizing agent and
displace the less electronegative Group V elements by combining the highly
electropositive Group III metals.

The investigation concerning chemical reactions occurring at the interface between
a metal film and a III-V compound semiconductor has attracted a great deal of attention.
Nonetheless, the measure of the reaction spontaneity is the change in the Gibbs free
energy of the system, (AGR), which takes into account the entropy change of the reaction
(ASR) of a system as well as AHR. In our discussion, since the assumption was that
metal/semiconductor reactions only involve solid-phase products, the contribution of the
entropy was considered insignificant. The phenomenon of group V elements escaping
from III-V compound semiconductors in gas-phase has been observed during reactions of
thin metal films with substrates [50]. Hence, the entropy change is required for the gas-

phase elements formation. This change is positive, especially for a vaporization process.

99

The system with a positive entropy reaction will ultimately become thermodynamically
spontaneous at particularly high temperatures; therefore, either the entropy or the reaction
is considered to be an important factor. The determination of AG; as a function of
temperature refers to the temperatures at which a reaction may occur spontaneously.
Expanding on the thermodynamic principles proposed by Gibbs in 1875, Gibbs’
phase rule identifies the possible number of degrees of fi'eedom in a system at
equilibrium, in terms of the number of separate phases and the number of chemical

constituents in a system.
F = C-P+2 Equation 81

where F is the number of degrees of freedom, C the number of chemical constituents, and
P is the number of phases. In the case of the system of Au-III-V with three components,
the maximum number of phases that may coexist is P=3. With the Au film deposited
onto the III-V substrate, the quantity of the II-V material is larger than that of the Au thin

film in this system. The reaction:
Au+III-V—)Au-III+V Equation 82

The Au thin film reacts with the III-V substrate in order to form the intermetallic
compound Au-III and elemental (element) V. In particular, the system will not reach the
equilibrium until all of the Au is consumed. This system is an open system, which
indicates that, during a heating cycle, the vapor pressure of the Group V elements
becomes appreciable. There are four phases present at equilibrium (III-V), Au-III, and

the elements of Group V in solid and gas-phases. As the gaseous species are

100

continuously being pumped out, the vaporization process will continue until all the solid
elemental group has escaped into the gas phase.

The reactions that have been observed between deposited Au films and GaAs
substrates were mainly driven by the loss of the gas-phase As in an open system during
the experiment. After losing a certain amount of As, the composition of solid phases that
were initially on the Au-GaAs pseudobinary shifted to the Ga—rich side, where the
equilibrium phases are Au and GaAs phases. With the loss of As in an open system, all
the Au may be consumed. Thus, a series of reactions liberating gas-phase As species
become feasible accordingly, in which the succeeding reaction occurs once the gold

containing reactants had been consumed:

7Au + 2GaAs —-) Au7Ga2 + 2A8
Au7Ga2 + GaAs —) AU7Ga3 + As
Au7Ga3 + 4GaAs —> 7AuGa + 4As

AuGa + GaAs —) AuGa; + As

Based on the thermodynamic equations mentioned in the previous section, an equation

can be deduced as follows:
AGR=AHR-TASR Equation 83

A reaction will occur spontaneously when AGR<0. If both AHR and ASR are positive,
AGR will be positive when T is small.

TASR=AHR- AGR

T = (AHR- AGR)/ ASR Equation 84

101

We can find the point AGR = 0 when the temperature is AHR/ ASR, This is the point when

the reaction will first become thermodynamically feasible.

2.4 Conculsion

In this chapter, various growth conditions have been employed to grow nanowires
from GaAs in the fumace system. Nanowires with different densities were found in
different growing conditions. From the SEM image in Figure 38, the significant growth
of nanowires only occurred when an additional gold was deposited on the gallium
arsenide chunk. The EDS analysis shows that nanowires were gallium oxide nanowires
with the gallium to oxygen ratio ranging from 1:1 to 1:3. To our knowledge, this is the
first observation of gallium oxide nanowires being synthesized through the simple
physical evaporation of the gold deposited on a GaAs chunk. In all cases, a gold ball was
found absent at the tips of the nanowires, however, when no gold was deposited on the

GaAs, no nanowires grew.

102

Chapter 3 Growing Nanowires from SrgGa16Ge3o

3.1 Introduction

Through independent research in the laboratory, an opportunity to investigate a
new material consisting of SrgGa16Ge3o was capitalized on. This material is a type-I
clathrate and displays glass-like thermal conductivities along with good charge carrier
mobility, which makes it an interesting material for therrnoelectrics research. We
produced nanowires from SrgGawGem simply by mixing it with Au nanoparticles and
annealing. To our knowledge, this material has not previously been grown in the form of
nanowires. Hence, the successful growth of clathrate nanowires would establish the
foundation for further studies of this thermoelectric material. Unfortunately, EDS
analysis during the TEM imaging shows the nanowires primarily consist of gallium, with

no trace of Sr and Ge detected.

 

Figure 43. The SI‘gGamGem structure. In the framework, the black atoms occupy the 6c
sites, the light gray atoms occupy the 16i sites, and the dark gray atoms
occupy the 24k sites. In the cavities, the gray atoms are at the Sr(l) 2a sites
and the black atoms are at the Sr(2) 24k sites [51].

103

3.2 Experiments
Experiment #1:

SrgGa16Ge3o powder was mixed with gold nanoparticles and annealed at 810° C for
120 minutes. The SrgGa16Ge3o pile was mixed with colloidal Au of 20nm nominal size
and with polylisine-L to avoid colloidal clustering. The sample was placed on top of a
quartz plate, and subsequently annealed to 810° C for 120 rrrinutes. The experiment was
conducted within a quartz tube (outer diameter, 90 mm; length, 80 cm) heated by a
horizontal tube furnace. The sample was put on a quartz boat and placed at the center of
the quartz tube that was evacuated to about 30 mTorr. A background gas of argon was
used at a flow rate of 200 sccm and this flow was maintained throughout the whole
reaction. The pressure inside the tube increased from 30 mTorr to 150 mTorr after the
argon gas entered into the tube.

The SEM image in Figure 44 shows that the nanowires were scattered on the
surface of the ngGa|6G630 powder. The dark area on the left side is the quartz plate.
Nanowires with a diameter of 50 nm grew on the SrgGa16Ge30 powder surface and mixed
with some whiskers of bigger size. The nanowires were found to grow in select regions
and were not notably straight. Some nanowires were observed on the quartz plate close
to the SI’gGamGem powder.

In the SEM image in Figure 45, the dark region is the quartz plate. The bright area
in the bottom left of the image is the SrgGa16Ge3o powder. We found that most of
nanowires distributed on the bottom half of the image were close to the SrgGa16Ge3o
powder region. Nanowires close to the srsGaléGC30 powder particles were found in

higher densities with lengths of approximately 5 pm and average diameters of 30 nm,

104

 

farther away from the powder particles, the density and length of the nanowires gradually

decrease to negligible quantities at distances of 11 um from the powder particles.

 

Figure 44. The SEM image shows that nanowires were scattered on the surface of
SrgGamGem powders. The scale bar is 5 pm.

The SEM image in Figure 46 shows that nanowires were mostly scattered on the top

surface of the SrgGalsGem powder. A mixture of larger diameter and smaller diameter

nanowires can be seen by comparing the center of the image, to the nanowires at the

perimeter of the powder particle. The dark area on the right side is the quartz plate. The

average length of nanowires was 4 pm and the average diameter was 70 nm. The scale

bar is 5 11m.

105

 

Figure 45. The SEM image shows that the dark region is a quartz plate. The scale bar
is 2 11m.

 

Figure 46. The SEM image shows that nanowires were motly scattered on the top
surface of the STgG316Ge30 powder. The scale bar is 5 pm.

106

 

Sarplesize: 22

I 561651

 

 

 

 

Figure 47. Histogram showing the diameter of nanowires. Total sample size: 22.
Average diameter is 52 nm.

The histogram in Figure 47 shows the diameter distributions of the nanowires. The

average dimension was about 52 nm in diameter.

107

0..

m0

 

 

 

8.8. 8.88 .28
..- ..- :. ... 88 88.8 88 88 88.8 2 2
:8 .1 8.8 8.8.: 88-: N8 33 888 88.8 83 88.8 38
-.. ... ... ... i 88 888 88 88 888 :5
-.. ... ... ... .. 88 88.8 88 88 888 .5
8.8-: 3.: 8.2+ 5.: 3.8 83 83 :3 83 88 Em
... ... ... t. .. 88 88 83. 88 8.8 now
8.8.: NS 3.8 .1 x8 .1 8.3 83 .88 83 83 88 x8
... ... ..- ... ... 88 88.8 88 88 88 one
88.: 8.8 8.3.. 8.8.: 8.: 83 38 83 83 :8 38
8.7: 8.8 8.8 .1 88.: 8.8 38 83. 83 88 98.8 go
8.8 .1 8.8 8.1.. 8.7: :8 88 83 N83 88 88 v6
..otm ....w o\o.a>> Aotm 0:3
8 2oz 8 .82 .582 8 8%; 8 ~88; u<N u < N some 828%
m w n o m e m N _.
U." ..=< mo _
Ma r ooom
.2. 88
um 88.

Figure 48. EDS analysis on the nanowires grew on Sf8G316G630

Because the nanowires grew on the surface of the SrgGa16Ge3o powder, the EDS

still shows the composition of SrgGamGem.

108

Experiment #2:
A similar experiment was conducted in the same condition. The only difference
was the sample was not mixed with gold nanoparticles. No nanowires were observed in

the experiment.

Experiment #3:
A similar experiment was conducted in the same condition. The only difference
was the sample was annealed for shorter period of time. The results were similar to

Experiment #1. The nanowires were scattered on the surface of the SrgGa16Ge3o powder.

Experiment #4:

SrgGamGem powder was mixed with gold nanoparticles were annealed at 825° C
for 120 minutes. We slightly increased the annealing temperature to 825° C and used the
same annealing time, 120 minutes. We used silicon wafer instead of quartz plate.

In Figure 49 and 50, the dark area is the silicon wafer. We found that the density of
nanowires was higher close to the powder. Most nanowires were found within a 5 pm
region around isolated powder particles, and higher densities of nanowires were found
between clusters of powder particles as shown in Figure 8.

In the SEM image in Figure 49 shows the nanowires on the quartz plate were
scattered around the SrgGamGem powders. In addition, there were some nanowires

growing on the surface of the powder.

109

 

Figure 50. The nanowires scattered around the SrgGamGew powders.

110

The SEM image in Figure 50 shows the nanowires on the quartz plate were
scattered around the Sf3G816G630 powder and growing on the surface of the powder. This
suggests. that a vapor transport of one or more component from the material is occurring
during growth of the nanowires.

In Figure 51, the nanowires show a uniform diameter along the length of the
nanowires ranging from 40 nm to 100 nm and the average length was 1.5 pm. These
ball-shape tips were observed in the majority of wires characteristic of the vapor-liquid-
solid growth mechanism. Most of the wires were remarkably straight without being

kinked.

 

Figure 51. The SEM image shows the distinctive structure of the nanowires.

111

 

Figure 53. The average diameter of nanowires was about 100 nm.

112

 

Totasarplesizezx

ISaiest

Coun
O—wanbUICDNmO

 

 

 

 

Figure 54. Histogram showing the diameter of nanowires. Total sample size: 36. The
average diameter is 62.22 nm.

The SEM image in Figure 53 shows nanowires that grew on the surface of the
SrsGalgGego powder. These nanowires had gold balls formed at the tips and the average
diameter of the nanowires was about 100 nm. The histogram in Figure 54 shows the
diameter distributions of the grown nanowires. The average dimension was about

62.2 nm in diameter.

113

or

It‘d" ' CII.
1 ‘
.

m

lull

 

[x

 

 

 

hm . a
we we a To 82
w ooom

mo

8 coon
r 08¢
88: 888 .98
.. ... ... ... ... 88 88 88 88 888 .25
... ... ... ... ... 888 88 88 88 88.8 3
8.8.: 8.3. 3.: 3-: 8.: 83 83 83 83 88 Em
... ... ... ... ... 88 88.8 88 88 88 48
88.: :8: 8.8.: 88.: 8.8 83 888 88 R3 888 v8
... ... ... ... .. 88 88 88 88 88.8 .26
8.: «8 8.8.: 88.: 8.: 83 88 88 :3 88 x8
83.: 8.8 88.: 88.: 8.8 88 88 :8 88 88 v5
88.: 8.8 88.: 88.: :3 8: 83 88 88 88 xo
85 5 0:8 85 8:
as 82 8 82 :82 588; 888; “E u. < N 885. 888

Figure 55. EDS analysis on the grown nanowires

114

Because the nanowires grew on the surface of the SrgGamGem powder, the EDS
was taken on a region of nanowires with a powder particle as a substrate, thus the

analysis still shows the composition of SrgGaléGem.

Experiment #5:

SrgGa16Ge3o powder was mixed with gold nanoparticles and annealed at
temperature at 740° C for 40 minutes with quartz plate on top of it. From the previous
experiments, we found out the density of nanowires was higher in the region closer to the
SrgGawGem powder. In the following experiment, we used this phenomenon to increase
the amount of nanowires.

The SrgGamGem pile was mixed with colloidal Au of 20 nm nominal size and with
polylisene-L to avoid colloidal clustering, and the sample was subsequently annealed at
740° C for 40 minutes. We put a piece of quartz plate on top of the pile as a local vapor

confinement technology.

115

 

Figure 56. The length of nanowire was over 4 um. Additionally, some nanowires
were observed to form as a ball shape in the tip. (2pm scalebar).

 

Figure 57. The diameter of nanowires ranged from 40 to 200nm. (500nm scalebar).

116

The SEM images show the growth of nanowires from the SI‘3G315GC30 powder
(Figure 57 and 58). The nanowires had gold balls formed at the tips, which suggests the
dominance of the VLS grth mechanism.

In the SEM image in Figure 58 and 59 the nanowires were observed on top of the
quartz plate. Most wires appear straight and the diameters were relatively uniform.

Again gold balls can be seen at the tips.

 

Figure 58. Nanowires on top of the quartz plate. The scale bar is 2 pm.

117

 

Figure 60. The scale bar is 10 pm.

118

The SEM images show nanowires grown from the SrgGaléGe3o powder (Figure 58

through Figure 60). The nanowires were very straight with gold balls formed at the tips.

 

Totd mrple size: 36

 

 

 

 

 

I Series1

 

 

 

 

Count

c3388

 

 

 

 

 

 

 

Figure 61. Histogram showing the diameter of nanowires. Total sample size: 36.
Average diameter 43 nm.

The histogram in Figure 61 shows the diameter distributions of the grown

nanowires. The average dimension was about 43nm in diameter.

Experiment #6:

We increased the annealing temperature to 950°C and used the same annealing
time, 40 minutes. A higher density of nanowires was found for this sample compared to
previous experiments as shown in Figure 62. The average diameter was about 60 nm and
the length was about 1.5 um. In the SEM image in Figure 63, was taken in the quartz
area far removed from the stGaléGC30 powder. The diameter was about 40 nm and the
length was about 1.5 pm. The scale bar is 10 pm. The density of these nanowires was

uniformly distributed over the quartz substrate as shown in Figure 64.

119

 

Figure 63. The diameter was about 40 nm.

120

 

Figure 64. The density of these nanowires was uniformly distributed over the quartz
substrate.

In the SEM image in Figure 65, the high density of nanowires was observed to
emerge in the quartz plate outside the SrgGa15Ge3o region. The length was about 2 pm
and the diameter was about 120 run. The scale bar is 2 pm. In the SEM image in Figure
66, the zoom out image shows nanowires were grown between the powder particles.
There is a difference in the diameter of the nanowires. Because some images are closer to

powder particles, and others are farther away.

121

 

Figure 66. The scale bar is 20pm.

122

 

 

 

I 891831

 

 

 

 

 

 

 

 

 

 

 

 

Figure 67. Histogram showing the diameter of nanowires. Total sample size: 42.
Average diameter is 55.24 nm

The histogram shows the diameter distributions of the grown nanowires. The

average dimension was about 55.2 nm in diameter.

3.3 Discussion

In experiments of growing nanowires from SrgGa16Ge3o, we have produced
nanowires from ngGal6GC3o by simply mixing it with Au nanoparticles and using a local
vapor confinement technique. Analysis of the nanowires shows we were not successful
in growing SrgGa16Ge3o nanowires, instead Ga203 nanowires were formed. A couple of
factors are observed to have affected the results, and gold is one of them. It was shown
that there was no grth of nanowires taking place without having gold nanoparticles
mixed with the sample. From the images shown, the gold balls clearly appeared at the
tips, which was an indication that the grth mechanism was the well-known vapor-
liquid-solid mechanism. From the TEM analysis, the nanowires were found to be

gallium oxide nanowires. Contrary to our expectation, strontium and germanium were

123

not present in the nanowires. In addition to gold, the second factor influencing the
growmg of nanowires from SrgGamGem is the vapor pressure. By confining the source
SrgGamGem material between quartz plates and adding gold nanoparticles, a significant
increase in the density of grown nanowires was observed. In addition, the diameter of
wires became uniform and nanowire grth could be found over broad regions of the
quartz substrate. Without the additional confinement on the source material, nanowires

were only found in close proximity to the source material.

124

Chapter 4 Growth of Nanowires from Germanium

One-dimensional semiconducting oxides nanostructures which have fascinating
stability and optical properties have attracted much attention because of their potential
applications in different fields such as optoelectronics and sensing. Germanium dioxide,
as a promising material, has been extensively studied on its optical and electronic
properties, and the grth of germanium dioxide nanowires has been studied by different
methods. Thermal annealing of germanium under sub-atmosphere condition leads to the
growth of a dense distribution of wires with diameters ranging from tens to hundreds of
nanometers and lengths of up to tens of microns on the substrate surface. This study
simply uses oxygen, which exists in the air, to react with germanium at certain
temperatures to form nanowires. The products are characterized by scanning electron
microscopy (SEM), transmission electron microscopy (TEM) as well as selected area
electron diffraction (SAED) pattern. The observations show the nanowires are crystalline

GeOz nanowires.

4.1 Introduction

Therefore, in order to meet the demands of further applications, it is imperative to
make use of the effect by combining nanowires and related optical properties and specific
sizes. Germanium dioxide is a blue luminescence material with a high refractive
property, and low optical loss [52]. Such properties have found great interest in the

fabrication of low loss optic fibers composed of Si02 and GeOz. Considering these

125

 

superior properties and their impact on the industry, germanium dioxide deserves further

investigation

Germanium Dioxide (GeOz)
This is the most common solid compound of germanium. Germanium dioxide
takes shape in one amorphous and two crystalline modifications:

1) A glassy amorphous form, refractive index 1.607, density 3.64 g/(cm3) at 20° C.

2) A tetragonal rutile form, refractive index 2.05, density 6.24g /(cm3) at 20° C.

3) White hexagonal quartz modification, refractive index 1.735, density 4.7g /(cm3) at
18° C.

The tetragonal form is practically insoluble in water, while the hexagonal and the
amorphous modifications have low solubilities; 0.45 and 0.52 % respectively, at 25° C.
Aqueous solutions are acidic due to the formation of metagermanic acid, H2Ge03. The
hexagonal modification converts to a tetragonal crystal system when heated at 350° C in
water under pressure. Both crystalline forms convert to a glass-like amorphous GeOz
when heated at 1100° C [53]. It is a stable, unreactive white powder which melts to a
glass and can be used as a component of glass which is more refiactive than the matching
silicate glasses. It is readily obtained by the hydrolysis of GeCl4, generating a
microcrystalline form of the hexagonal modification of GeOz, bearing the structure
similar to that of low quartz. This is often called the “soluble” form that dissolves in cold
water to the extent of 4g/l. In a similar fashion, the same crystalline modification is
obtained by keeping GeOz glass at 1080° C until it devitrifies. The only other

modification of GeOz is the “insoluble” or tetragonal form, which is obtained by the

126

hydrothermal conversion of the “soluble” form through heating with water at 355° C for
100 hours. The “insoluble” form is stable up to 1033° C, where it inverts slowly to the
“soluble” form, which will melt at 1116° C [54].

To produce GeOz nanowires, various mechanisms have been utilized, including
physical evaporation [55], thermal treatment [56], carbon nanotube confined reaction of

Ge [57], laser ablation [58], heating in oxygen environment [59], thermal oxidation [60],

and carbotherrnal reduction reaction [61].

4.2 Experiments
Experiment #1:

A layer of gold was deposited on a germanium wafer at 850° C. The germanium
chunk was put on top of the silicon wafer and subsequently annealed to 850° C. After
reaching 850°C, a KrF laser was used (500 ml, 10 Hz for 90 seconds) to ablate a gold
target. Room air was allowed to flow into the hot chamber for 10 minutes to increase the
pressure to 350 Torr. Then the vacuum pump was turned on to evacuate the chamber,
and the fumace was turned off until to cool down.

The experiment was conducted within a quartz tube (outer diameter, 90mm; length,
80 cm) heated by a horizontal tube fiimace. The sample was put at the center of a quartz
boat that was placed at the center of the quartz tube. The tube was evacuated to about

30 mTorr. Figure 68 shows the experimental setup.

127

 

Gold coin

 

Laser

\

Germanium ......f _. ‘ 1 f
Silicon J

Observed area

 

 

 

 

 

 

 

Figure 68. The experimental setup, for investigating the growth of GeOz nanowires.

Figure 68 shows the setup and the observed are, of which the SEM images were
taken. The SEM image in Figure 69 was taken in the observed area as shown in Figure
68. According to the setup system in Figure 68, the only opening for gas flow in this
system is the gap between the gold coin and the tube. Afier the experiment, nanowires
were found to emerge along the gap on side of the gold coin. The SEM image in Figure
69 shows the edge of the gold coin. The nanowires did not find on the germanium or

silicon wafer. The nanowires were found on the surface and the edge of the gold coin.

128

 

Figure 69. The SEM image shows the edge of the gold coin. The scale bar is 200 pm.

129

 

10 pm

Figure 70. The SEM image of nanowires on the surface of gold coin. The scale bar is
10 pm.

The SEM image in Figure 70 shows that the nanowires were over 30 pm in length.

The diameters of nanowires were between 100 to 1000 nm, and those with smaller

diameters were found to be much straighter than the larger diameter wires. The larger

diameter nanowires occasionally changed direction during their growth creating the
kinked shape.

The SEM image in Figure 71 shows the wires at the edge of the gold coin. The

diameters of the wires shown in Figure 71 ranged from 100 to 1000 nm and the larger

wires were kinked. The smaller nanowires are more clearly seen in this image as straight

and short in length.

130

 

Figure 72. The whiskers with kinked shape. The scale bar is 2 um.

131

Figure 72 shows the SEM image of the germanium oxide nanowires. The kinked
nanowires had an average diameter of 1 um. Between those large wires, there were some

nanowires with an average diameter of 100 nm.

 

Figure 73. Away from the observed area, no nanowires were observed.

Figure 73 shows the SEM image taken at the edge of the gold coin that was far
from the chamber exhaust opening. There were no nanowires observed in this area. From
this observation it was hypothesized that germanium was volatilizing into the gas stream
such that the edge of the gold coin that was exposed to the highest concentration of

germanium (or germanium oxide) gas molecules was the location of nanowires growth.

132

 

 

 

 

 

 

 

 

8 8 ... . . . . m ... m m w .

...... H. 4;-..

. -82

8 h -83

-88

18.8

3 -88
8.8. 8.8. 82

.. .. .. .. .. 88 88 88 88 88 88
.8... 8.8 8:. 8:. 8.: 8: 8: 88 8: 88 .8
2:. 88 .8... .8... 8.8 8. 88 88 e8 e8 8
8:. 8 8.3. .8... .8 82 83 88 :8 88 8
85 5 .85) 85 8:
.. 8.. .. 88 .82 ... .8... .. .888 .8 a < N .88. .888

Figure 74. EDS analysis on the nanowires.

133

In Figure 74 the EDS analysis on the nanowires showed the composition of
nanowires were germanium and oxygen. The germanium to oxygen ratio was about 0.68
for an average composition of Geo}, with x = 0.53.

The germanium oxide nanowires and whiskers on the edge of the gold coin might
indicate that the gold did play a role as metal catalyst. The following experiments were

designed to test this assumption.

Experiment #2:

A layer of gold was deposited on top of germanium wafer at room temperature.

The piece of germanium was placed on top of an uncoated silicon wafer and another
piece of silicon wafer with gold deposited (~10nm thick) was put on top of it,
subsequently annealed to 850° C. It was then lefi at 850° C for 40 minutes with Ar/H
(95/5) flow at 200 sccm to reach 10 Torr. Then the Ar/H flow was shut off and room air
was allowed to flow into hot chamber for 1 minute until a pressure 350 Torr was
obtained. The vacuum pump was then turned on, and the fumace was turned to initiate
the cool down.

The experiment was conducted within the same quartz tube heated by a horizontal
tube furnace. The sample was placed at the center of quartz boat. The quartz boat was
placed at the center of a quartz tube, which was evacuated to about 30 mTorr. A
background gas of argon flow of 200 sccrn traveled along the quartz tube, which was
maintained in state throughout the entire reaction time. The pressure inside the tube

increased from 30 mTorr to 150 mTorr afier the argon gas entered the tube.

134

The SEM image in Figure 75 shows the silicon wafer with a layer of gold was
covered with dense germanium oxide nanowires on its surface, which helps us to prove
the assumption, which germanium oxide nanowires and whiskers on the edge of the gold
coin might indicate that the gold did play a role as metal catalyst. Most of the nanowires

were in very straight shape.

 

Figure 75. The silicon wafer was covered with dense germanium oxide nanowires.

135

\

.\ . \

\

“X‘\ A‘ Q \

Figure 76. The growth of nanowires on the edge of a silicon wafer. The scale bar is
10 um.

 

The SEM image in Figure 76 shows the growth of nanowires on the edge of the
silicon wafer. The nanowires tapered off and the accompanied bright dots were at the tip.
The gold dots at the tips clearly confirmed the VLS growth mechanism.

On the silicon surface that was coated with gold, a high density of nanowires were
found with diameters that tapered to points at the tips which terminate at a gold dot as
shown in Figure 77. Nanowires had diameters of 90 nm at the tip; however, in some

sections of the wires, the diameter was 2 pm.

136

 

Figure 77. The gold balls were appeared at the tip of nanowires. The scale bar is
5 pm.

The image of higher magnification is shown in Figure 78 where the growth of
germanium oxide nanowires is observed. The diameter of nanowires was about 100 nm,

and the diameter of the accompanied bright dots was 200 nm.

137

 

Figure 78. The image of a nanowire at a higher magnification (1 pm scalebar).

138

 

8§ 8.2: 52
... ... ... ... 82 83 83 83 80.0 ..8
$0... 8.9. 2.1.. 8.1+ 25 E: 83 82 $3 $2 v30
3... Rd... 5.3.. $5.? 8.2 no: 83 8E 32 ”mod xo
”mm .1 8“ $3.. 3.3. E 83 83 03.2 83 82 v8

 

as 5 ...; as .5
.. so... ... S. .52 .15.; ... E... E a < N 2...... sea

 

58 35% $5555

.—
_.

 

183
r88
.. 8mm

 

 

8

63 mm "use 29?

Figure 79. EDS analysis on the nanowires.

In the Figure 79 EDS analysis on the nanowires shows the composition of the

nanowires were germanium and oxygen. The germanium to oxygen ratio was about 0.88.

139

The result shows the silicon wafer with a layer of gold was covered with dense
germanium oxide nanowires on its surface. We wanted to verify that the growth

mechanism was metal catalyst. So the following experiments were conducted.

Experiment # 3:

The experiment was designed to study the effects of gold on the growth of GeOz
nanowires. A 10 nm thick layer of gold was deposited onto a silicon wafer. A small area
was then scratched to remove some of the gold. The scratched sample was then placed
on top of a piece of germanium which was on top of a silicon wafer. The Si/Ge/Si(with
scratched Au film) was heated in a quartz tube to 850° C. It was then left at 850° C for
40 minutes with Ar/H (95/5) flow at 200 seem to reach 10 Torr. Then the Ar/H flow was
shut off and and room air was allowed to flow into the hot chamber for 1 minute until a
pressure of 350 Torr was reached. The vacuum pump was then turned on, and the
fiimace was turned off.

The resulting images are shown in Figure 80 through Figure 82 and show that the
gold was essential for the nanowire growth. The SEM image in Figure 80 shows the
silicon wafer having 10 nm thick gold coated. Most areas on the surface were covered
with a white wool-like layer, which was essentially made of germanium oxide nanowires.
Before the experiment, tweezers were carefully used to scratch a few lines on the gold-
coated layer. After the experiment, no growth of nanowires was observed along these

lines.

140

. 1mm A "l

 

. A
Figure 80. The silicon wafer only partially covered by nanowires (1mm scalebar).

‘51". . . . .
Figure 81. A closeup image of a scratched region showing that the growth of
nanowires only occured in the gold coated areas. The scalebar IS 10
pm

 

141

The SEM image in Figure 81 shows a dark region in the center where the gold
layer was mechanically removed before heating the substrate. The wires in the growing
region had long length that could be extended into the dark region. Some nanowires were
over 40 pm in length with diameters of less than 100 nm. Bright dots could also be found
at their tips. Nanowires with larger diameters of ZOO-300 nm usually exhibited more

kinking and bright dots at the tips could not be found.

.1

 
  
  
  
  
    
 

:59
"fl.
22>

i.

is

  

U
a.

.r 1."
~

9g a.
h“:
is.

p .1 .
32'.»-
.':-F. ‘

‘V ‘
y.
'5
’
, C
I.“

.‘v
1‘ a

...... .-_~.--~-'.;£

--

Figure 82. The dense of nanowires were observed on the surface of silicon wafer. The
scale bar is 5 pm.
The SEM image in Figure 82 shows that the nanowires grew in the gold-rich area.
The diameters of nanowires ranged from 100 nm to 500 nm. The bright dots were found
at the tips of nanowires with smaller diameters. The kinked shape was usually present in

nanowires with big diameters.

142

The histogram in Figure 83 shows the diameter distributions of the grown wires.
The average dimensions of the grown wires were about 285 nm in diameter. In Figure 84

EDS analysis on the nanowires showed the average composition to be GeOz.Jr with

x = 0.8.

 

Total sample size: 20

Count

I Series1

 

 

(nM)

 

 

Figure 83. Histogram showing the diameter of nanowires. Total sample size: 20. The
average diameter is 285 nm.

143

 

8.02 8.2: EE
... ... 89° 805 806 Bod Sad .30
8.7? 8.9. Edi: 5.2+ 3.2. $3 83 ago 23 Bad x3

3.2+ 33 2.7: 2.7? 8.3 83 89.0 83 mood mood xo

 

55 E 35> 55 0::
.N. .52 .N. go... .552 :90; $29.; an a < N one; Sam

 

.98 as... .2258

:ormwhomvmmr

 

. — .
.13.-..“1‘931hnf6.-uflhiuviniflirublifiéuuflvfiafiluiiifiig3.3....... it. i. it..- .1343- .rlw - . ..l a , . . . Linutulllqr
.. .- . . . ..~_
.\
_.
.. . . u
. . II
. ,. .
.
. ... .
.. V H
. . .. CON

~

90 .. 00?

Figure 84. EDS analysis on the grown nanowires.

144

Experiment #4:

A piece of germanium was put on top of two silicon wafers, one is gold coated and
the other is uncoated. Then a second piece of silicon with a gold coating (~10 nm thick)
was placed on top of germanium, and the stack was subsequently annealed to 850° C.
Before annealing, the gold coated substrate was scratched in the form of the character
“S”. Then the system was left at 850° C for 40 minutes with Ar/H (95/5) flowing at 200
seem to reach 10 Torr. The Ar/H flow was then shut off and room air was allowed to
flow into the hot chamber for 1 minute to reach a pressure of 350 Torr. Then the pump

was turned on and the furnace turned off to allow the system to cool.

145

 

 

Figure 85. The image of the sample set.

Figures 85 through 87 show successively higher magnification of the top sample
and demonstrate the control over grth location for the GeOz nanowires. The “S”

shaped scratch was formed using a diamond pen.

146

 

Figure 86. No growth happened in the “S” (1mm scale bar).

Figures 87 and 88 show different portions of the “S” and although nanowires
extend over the scratched region, termination of the growth by removal of the gold is
rather abrupt. Of the various nanowire materials studied in this thesis, GeOz has shown
the greatest control over location when using gold as the metal catalyst. The nanowires
have lengths over 20 pm, showing some tapering along the length of the wires, and a
bright ball at the tip of each wire as in Figure 89. The diameter of the wire is

approximately 100 run near the tip, with a 150 nm diameter gold ball.

147

:‘T‘ ‘r '5' "f. .t‘
1.1"} I.“ “3.13: 'y .5 =

”“ ‘.¥ a" V

 

T5 9.3% ..
T‘Xfi'.
} .

i

.\. .
w

: .11":

;.-.
f ' .
.
i-
,

r .

 

Figure 88. Nanowires extending into no gold zone. The scale bar is 5 pm.

148

 

Figure 89. The nanowires had a bright dot at tip (1 pm scalebar).

Far from the scratched region, the nanowires grew with high density as shown in
Figure 90 and Figure 91. The diameters of nanowires ranged fiom 100 nm to 400 nm.
EDS analysis of the nanowires in Figure 92 shows an average composition of Ge02_,,
with x= 0.6 suggesting that there is either a mixture of G60 and GeOz nanowires, or
some unoxidized germanium is being deposited. Further control over the oxygen partial
pressure would be needed to investigate the compositional control of the nanowires,
however it was found that nanowires did not grow when the pressure was maintained at
30 mTorr, and no room air was allowed into the chamber. It was thus concluded that
both oxygen, and the metal catalyst (gold in this study) were required (in addition to the

proper growth temperature) for successful nanowires growth.

149

 

“'5’ i" "72”” ..c
M» ‘55.:

u
‘
v

o

' a

awe * .
a“ /’ix/’%fi’i4?; [(4%«7»
, ’ ' 9 ‘ ’L

  

    
  
  

 

r. . a»...
w" gas-:94! , ' 5 « .
Mamas. .2" . .

 

Figure 91. Higher magnification SEM image of nanowires (2 pm scalebar)..

150

 

8.09 8.02.

+8-: 8.: N2... N2-.. 82
8:. as... 02:. 02:. $8
85 :m $.02, 85

0N. .82 .No so... .532 .N. 2925 .N. 2%;

oood oood 80.0 08.0 oood
3o; coo; momd mace ommd
«a: mmmd on; momd v3.0

“ZN ... <

. E8 2.3.... 95:55

.... 0r 0 w k. w

m v m N r

.80»
._ 3
x 00
x o

2:

 

N 255. EoEom

 

 

 

00

$0

I. n. I .~ I ..l..-:.t.¥ It. ..FI!.I:.\I.|C n.-. ».luurq.l.lur'.rt...

Ito-ll

4.1.3.5...“ . "

6.0

..MI.

... com

. 00v
. 000
- 000
.000 r

 

Figure 92. SEM EDS analysis on the nanowires.

151

TEM analysis

TEM samples were prepared by sliding a TEM copper gid on the surface of
nanowire covered substrate. Many nanowires would break off and remain on the grid,
and no post processing was needed. In some cases, the surface of a nanowire covered
substrate was scraped with a razor blade, over a container of methanol. The methanol
was then drawn into a pipette, and a small drop placed on a TEM gid. After evaporation,

the TEM analysis was taken as shown in Figures 93 through 96.

 

Figure 93. TEM image shows the gold ball at the tip of the nanowires (0.2 pm
scalebar).

152

 

Figure 94. TEM image shows the nanowires have diameter about 40 nm (0.2 pm
scalebar).

 

Figure 95. TEM image shows the nanowires have diameter about 65 nm (20 nm
scalebar).

153

 

Figure 96. TEM image shows the gold ball at tip of nanowires (20 nm scalebar).

Selective area electron diffraction (SAED) shows ordering in the nanowires, but was

inconclusive for determining the orientation and lattice constant as shown below.

 

Figure 97. Selected area electron diffraction (SAED) pattern of nanowires.

154

4.3 Theory

Germanium forms a wide range of binary compounds and eutectics with many
metals and metalloid elements. In short, the metal is stable in air and water at ambient

temperatures. Nonetheless, it reacts with oxygen at elevated temperatures engendering
divalent and tetravalent oxides, GeO and GeOz. The amount of GeOz will gow with the

increasing time of exposure to oxygen [62].
Germanium forms inorganic ionic and covalent compounds in both its +2 and +4

oxidation states [63]. Its +2 oxidation state is similar to that of a singlet state carbine in
which Ge is sp2 hybridized with a lone pair of electrons and an empty p orbital. This

electronic configuration results in diverse structural types, which are determined mostly
by steric factors ranging from V-shaped monomeric molecules to polymers. The stability
of the +2 state is geater than Si and less than Sn. In its +4 oxidation state, Ge

compounds are four-, five-, and six-coordinate. Some developments have been focused
on the isolation of stable monomeric GeIV compounds with an N or S double bond to
germanium [63]. Germanium oxidation state +4 is more stable than the +2 oxidation
state [53]. Therefore, germanium will oxidize easily in air to GeOz. Germanium dioxide

has a high refractive index and infrared transmission, and is utilized in industrial glasses

such as fiber optics.

We suggest that the process for the formation of the Ge02 nanowires involves four

steps. First, germanium is heated in the air and yields germanium (IV) dioxide, Ge02

[64] as a thin layer on top of the germanium, Ge.

Gem + 020’) ———9 Ge02(s) Equation 85

155

 

Second, the surface germanium dioxide continues to be heated with germanium
while another chemical reaction takes place. The germanium dioxide, GeOz, reduces to

germanium monoxide, GeO. The metal also reduces the tetravalent oxide to the divalent

oxide once heated at elevated temperatures

Gem + GeOz(s) $531+ 20.20“) Equation 86
GeOz by itself is stable even under the condition of higher temperatures, but, when it
comes to contact with Ge, GeOz reacts with Ge at > ~710°C, converts to GeO, and then
sublirnes.

Third, it is known that GeO can be removed from Ge surfaces through thermal

annealing in a vacuum. Then GeO starts to vaporize above ~710°C.

Ge0(s) —fl9—> Geo“) Equation 87
GeOU) iii) GeOM Equation 88

Fourth, GeO oxidation in air to GeOz is rapid above 550°C [63]

264200,) + 020’) ——> 2Ge02(v) Equation 89
The gold liquid droplet is a preferred location for deposition from the vapor, which

causes the liquid to become supersaturated with G602“) nanowires gowing from the

gold seed on top of the silicon wafer via the vapor-liquid-solid mechanism [65]. The
presence of a liquid layer in contact with the gowing crystal has important consequences.
The surface of the liquid has a large accommodation coefficient and is therefore a
preferred site for deposition. Hence, the liquid becomes supersaturated with materials

supplied from the vapor, and the crystal growth occurs through precipitation at the solid-

156

 

liquid interface. For a condensation reaction, the gowth rate J (atoms/ cm2 sec) of an

extensive crystal surface can be estimated from gas kinetics as:

l
J = aapo (271ka) 2 Equation 90

where a is the accommodation coefficient, 0': (p— p0)/p0 is the supersaturation of

the vapor in which p is the vapor pressure, and p0 is the equilibrium vapor pressure of

the solid at temperature T. The faction of approaching atoms which become

accommodated on the gowing surface is a [66].

4.4 Discussion

After germanium (melting point: 938.3 °C) was exposed to air for 1 minute at
850°C, large quantities of white wool-like wires were found on top of the patterned
silicon wafer on which a 10 nm thick layer of gold had been deposited, then scratched.
The SEM images show that GeOz nanowires grew only where the gold was deposited.
The SEM images also show the diameters of nanowires varied from 50 nm to several
hundred nanometers, and they were tens of microns in length. TEM analysis showed

more definitively the gold tips at the ends of the nanowires, and indicated ordering of the

atoms in the nanowires.

4.5 Conclusion

In summary, the thermal treatment of the germanium at 850° C under air flow led to a
series of reactions forming nano- and microwires with diameters ranging from tens to

hundreds of nanometers, and whose lengths were up to tens of microns on the substrate

157

 

surface. In contrast to other methods, the manipulation of controlling lower temperature
and shorter reaction time appears to have evoked these chemical reactions. Moreover,
the longer exposure to air resulted in the formation of nanowires with larger diameters.
From these findings, it could be concluded that since nanowires only gow on the gold
surface, the VLS mechanism plays an important role in the gowth of nanowires. The
presence of gold at the tip of the wire, as verified by EDS analysis, however the presence

of oxygen during gowth is also essential for gowth of the nanowires.

158

 

Chapter 5 Gold pattern experiment

5.1 Introduction

In the vapor-liquid-solid (VLS) technique, a metal is chosen based on the phase
diagam between that metal and the nanowire materials of interest. As an example, the
phase diagam for silicon and gold is shown in the previous section. If the system is
raised in temperature above the eutectic point (~ 360° C), a molten alloy is formed. Since
the sticking coefficient of a liquid is very high, a preferential deposition of silicon at this
location can occur at higher deposition temperatures where the sticking coefficient of
silicon on silicon can be reduced. As the alloy is supersaturated with silicon, the solid

silicon drops out to form the next layer of nanowires, and the molten alloy is pushed to

\ -"_ _-., --.
/ ~\
\
I
v ’0
S.
' n
I h
I r,
.,
/ \ ~ ,
/ . , O ‘
I‘ 't
k a
‘ i ‘ E
a t
3‘ + .“
~. 3
t i‘ . ’
l ,
.--" c-N.,_ ~ . ,
__, x . , I
/ ‘ ’ .
.' \ .
r ‘4.
\ _ ' ' ‘1 u
. .i .

the top of the wire.

     

 

Figure 98. Gold droplets form an alloy with silicon.
The alloy has a eutectic point near 360°C,
above which it is molten.

In the VLS technique, the location of the nanowires can be controlled by

controlling the location of the metal droplet. Reported here are the results of

159

investigating various patterns etched through silicon dioxide to determine if a preferential

location for the metal droplets could be found within the pattern.

5.2 Experiment

A series of experiments were devised to investigate the effects of the control over
the location and size of gold particles on silicon (111) substrates. It is well known that
the deposition of gold onto substrates containing defects can result in a “decoration” of
those defects by gold particles. This is caused by the reduction in free energy when the

droplets as described in Figure 3 are located at a step edge on the substrate. At such

locations a larger portion of the total droplet surface is governed by 'st- Therefore, it is

expected that the gold particles will preferentially reside at these step edges or at the
corners of step edges where the free energy is even further reduced.

In these experiments, the silicon substrates were first oxidized. The oxide layers
were then patterned through standard lithography processes, as described in the “Buffered
Etches” section. Five different shapes were patterned through the oxide layer. Gold was
then either deposited on the substrate at room temperature, followed by an annealing step,
or it was deposited onto the substrate at a high temperature of deposition. Various

temperatures were investigated with results shown below.

5.3 Gold Deposited at Room Temperature

An oxide layer was first gown to a thickness of 100 nanometers, and then
patterned. A gold layer was then sputter deposited at room temperature to a thickness of

10.5 nanometers, and the wafer was subsequently annealed to the temperature indicated.

160

 

Figure 99 through Figure 104 show the resulting SEM images. Due to the surface
tensions, Au will ball up to form surface droplets as shown in the figures. The diameter
of the Au ball in this experiment varied from 0.1 um to 0.5 um. Based on the phase
diagam, the balls were an alloy of gold and silicon when annealed to these temperatures.
Figures 99, 103, and 104 show the gold preferentially coalesces at and near the holes

through the oxide layer.

   

Figure 99. The darker location is Si and Figure 100. The area identified in this
the lighter locations are Si02. image is near the edge of the half-circle
The scale bar is 20 um. pattern. The scale bar is 2 pm.
Figure 99 through Figure 104 show the silicon wafer with 10.5 nm of Au followed
by a 550° C annealing for 5 minutes. The darker regions in these images correspond to

the silicon surfaces, and the lighter regions to the Si02 films. The density of Au balls is

seen to be higher on Si than the Si02 locations.

161

4
l -

Figure 101. The image is near the edge of Figure 102. The image is inside the

the half-circle pattern. The pattern. The backgound is
darker location is Si and the Si. and the white dots are Au.
lighter locations are Si02. The The scale bar is 2 pm.

scale bar is 4 pm.

.-

Figure 103. The dark part is Si and the Figure 104. This is a half-circle pattern.
light ball is Au. This area is The white chunk is silicon. The scale bar
inside the pattern. The scale bar is 20 pm.
is 10 pm.

 

Figure 105 through Figure 107 are SEM images of the silicon wafer with a deposited
layer of Au at room temperature. The thickness of the Au layer was 10.5 nanometers,
and the silicon wafer was annealed at 620° C for 5 minutes. At this annealing

temperature, larger Au balls (about 500 nm diameter) were observed at silicon locations

162

on the substrates, which suggests a larger surface diffusion occurring on the silicon
dioxide surface. The average diameter of the Au particles was estimated from these

images to be 100 nm.

.-

Figure 105. The average diameter of the Figure 106. This is a half-circle pattern.
Au balls was 100 nm. The The scale bar is 20 um.
scale bar is 2 pm.

.-

Figure 107. The darker location is Si and Figure 108. 10.5 nm of sputter deposited
the lighter locations are Si02. Au followed by an
annealing at 7000 C for 5
mins. The scale bar is 10
pm.

163

 

.-

Figure 109. The image is near the edge of Figure 110. The image is near the edge
the half-circle pattern. The of the half-circle pattern.
scale bar is 2 um. The scale bar is 2 pm.

Figure 108 through Figure 112 are SEM images of the silicon wafer with a 10.5 nm
thick sputter deposited layer of Au at room temperature. The silicon wafer was then
annealed at 700° C for 5 minutes. Some large Au balls of about 500 nm diameter were
seen at silicon locations. This suggests a higher sticking coefficient in the silicon regions
and a high surface diffusion of Au. The average diameter of the Au particles v 2

than 100 nanometers.

.-

Figure 111. The image is near the edge of Figure 112. The dark location is Si.
the pattern. The dark location is The scale bar is 5 pm.
Si and white dots are Au. The
scale bar is 2 pm.

 

164

 

From Figure 99 through Figure 112, three different annealing temperatures were
investigated in order to determine the morphological influence of temperature on gold
balls. Larger Au balls were observed in the silicon regions of the substrates for higher
annealing temperatures; however, in all cases, the gold was evenly distributed throughout
the silicon regions and did not preferentially collect at the Si-SiOz edges. This could be
ascribed to the relatively thin oxide layers used, and the annealing temperatures

investigated.

5.4 Gold Deposited at High Temperatures

From Figure 113 through Figure 122, the Si02 layer on the silicon wafer was first
patterned and subsequently placed inside a tube furnace system. The temperature was
then raised to 700° C, and Au was deposited by pulsed laser deposition (PLD) for 1
minute. The substrate was then maintained at 700° C for 5 minutes before reducing power

to the tube furnace.

    

10 um

Zum

 

Figure 113. The Au ball was observed to be Figure 114. The area in this image is in
in the gap as a bridge. The the edge of pattern. The
scale bar is 10 um. scale bar is 2 pm.

165

Similarly, it was found that gold did not preferentially locate at the silicon — silicon
dioxide edges; rather, it was distributed evenly on the surface. It appears that silicon
dioxide layers did show pits and holes through the oxide at locations where it was not
patterned. Gold preferentially deposited at silicon locations, and not on the oxide layer.
We are, therefore, in the process of investigating thicker oxide films both to increase the
step height between Si and Si02 and to avoid pinholes through the oxide films and the

degradation of the oxide films at higher temperatures.

   

50 um zum

Figure 115. The SEM image shows the Figure 116. The SEM image shows the
square pattern. The scale bar edge of pattern. The scale
is 50 um. bar is 2 pm.

211m

   

Figure 117. The Au ball was found Figure 118. The SEM image shows gold
outside the pattern. The scale ball scattered outside the
bar is 2 pm. pattern. The scale bar is 75
pm.

166

 

Figure 119. This gap is not big enough to have a big Au ball produced in the gap. The
scale bar is 2 pm.

   

Figure 120. The Au balls are inside the Figure 121. The image is near the edge of
pattern. The scale bar is 10 the pattern. The scale bar is 20
um. pm.

167

 

 

Figure 122. The temperature was raised to 700° C and Au was deposited by pulsed
laser deposition (PLD) for 1 minute. The substrate was then maintained at
700° C for 5 minutes before reducing power to the tube fumace. The shape
of the Au ball looks like a bridge. The scale bar is 2 pm.

 

5 pm 2pm

Figure 123. The image is taken near the Figure 124. The image is taken near the
edge of the pattern. The scale edge of the pattern. The scale
bar is 5 pm. bar is 2 pm.

168

 

Figure 125. The temperature was raised Figure 126. The temperature was raised

 

to 700° C and Au was to 700° C and Au was
deposited by pulsed laser deposited by pulsed laser
deposition (PLD) for 1 deposition (PLD) for 1
minute and reducing power minute and reducing power
immediately. The scale bar immediately.

is 40 um.

Figure 125 and Figure 126 were obtained using laser ablation of gold for 1 minute. After
this deposition, the temperature was rapidly decreased. These figures show that the
number of large Au balls has decreased relative to the experiments where the substrate

was maintained at 700° C for 5 minutes after the deposition.

169

Mask design

 

A
1:1 D
Q /
A A
A [:1 Q
A
D Q
[:1 Q C] Q
Q / Q / Q If Q 7

Figure 127. The mask designed for patterning the Si02 layer using a positive photoresist
(Shipley 1813).

We have 110 sets in this mask design (Figure 127). The resolution of this mask is 5080

dpi. The mask was first drawn in AutoSketch, and then we converted it to a pdf file with

5080 dpi. The file was sent to a local company, Info Tech, for print.

5.5 Conclusion
Initial investigations aiming at controlling the location and size of the gold metal
catalyst particles on the surface of silicon have shown a preferred deposition of Au onto

silicon locations, particularly for the depositions of Au on the substrates at higher

170

temperatures. The expected preferential deposition of Au at the edges between silicon
and silicon dioxide was not observed under the conditions studied. Further studies are
under way to investigate thicker oxide films and a broader range of deposition

temperatures.

 

171

Chapter 6 Germanium Oxide Nanobridge

This chapter aims at investigating the lateral growth of high density metal-
catalyzed germanium oxide nanowires and the bridging of nanowires between two
vertical silicon sidewalls, which can be developed as electrodes used for an electronic
device. A variety of methods to produce metal-catalyzed nanowires were used. Afier the
catalytic metal was placed on the vertical silicon surfaces, a thermal annealing process
was initiated to gow nanowires and to eventually form ‘nanobridges’. Nanobridges have
certain unique properties, such as a high surface-to-volume ratio and the capability of

connecting two electrodes, which allow them to be effectively used as nanosensors.

6.1 Introduction

Possessing a high surface-to-volume ratio renders free-standing nanostructures very
sensitive to the charged species absorbed on their surfaces. More specifically, this
property makes them especially attractive for sensor applications. The lateral gowth of
GaAs nanowires by metalorganic vapour phase epitaxy (MOVPE) on a GaAs substrate
has been reported [67,68]. The lateral gowth of silicon nanowires between two silicon
surfaces through chemical vapour deposition (CVD) also has been reported [69].
Moreover, using one-dimensional nanostructures to synthesize semiconducting oxides
that have the feature of stability, and optical properties of interest has attracted much
attention. This has led to the investigations of their potential applications in different
fields, such as optoelectronics, being used as detecting tips for atomic force microscopy

(AFM), and sensing. Germanium dioxide, as a promising material, has been extensively

172

‘-

 

studied regarding its optical and electronic properties, and the gowth of germanium
dioxide nanowires has been studied through different methods. Thermal annealing of
germanium under the sub-atmosphere condition leads to the growth of densely distributed
wires with diameters ranging from tens of nanometers to hundreds of nanometers, and
their lengths are up to tens of microns on the substrate surface. In the previous chapter,
the gowth of germanium oxide nanowires has been discussed in detail. In this chapter,
the focus is on the lateral gowth of germanium dioxide nanowires. Using semiconductor
nanowires, researchers have succeeded in demonstrating nanowire sensors for biological
species [70]. In the current research, the underlying mechanism for nanowire sensors is a
field effect [71]. Researchers have used a fluid flow method to assembled nanowires
into parallel arrays and clusters, and electrical contacts were defined by e-beam
lithogaphy. The approach of connecting electrodes and nanowires one at a time cannot
be used for manufacturing devices, although it is useful for understanding the
characteristics of nanowires. Hence, the mass self-assembling technique is necessary to
enable germanium oxide nanowires to bridge between electrodes.

In this chapter, the lateral gowth of metal-catalyzed germanium oxide nanowires is
reported, from one vertical Si plane to the other vertical plane. By using a large number
of such ‘nanobridges’ in parallel, the desired high-surface area can be obtained in a small
volume. Nanowires should gow laterally from one sidewall toward the opposing
sidewall of the trench. The surface-to-volume ratio is key to the performance of
nanosensors. Therefore, the distance between the two vertical surfaces plays an
important role. In the example of this chapter, the distance between the two vertical

surfaces ranges from 50 to 70 pm.

173

t i H non-f“

fit :.

 

 

 

 

 

 

(b)

Figure 128. Schematic diagams of (a) metal catalyst and (b) lateral nanowire gowth.

6.2 Experiment
Experiment #1:

A piece of germanium was placed on top of a silicon wafer. Another piece of
silicon wafer with gold deposited (~10 nm thick) was put on top of the germanium and
the stack was annealed to 850° C. It was then left at 850° C for 40 minutes with an Ar/H
(95/5) flow of 200 seem to reach 10 Torr. Then the Ar/l-l flow was stopped and room air
was allowed to flow into the hot chamber for 1 minute to reach pressure 350 Torr. Then
the pump was turned on, and the furnace was turned off in order to cool.

The experiment was conducted within a quartz tube (outer diameter, 90mm; length,
80 cm) heated by a horizontal tube furnace. The sample was put at the center of a quartz

boat placed in the middle of the quartz tube, which was pumped down to about 30 mTorr.

174

Result

Photoresist was spun on the (110) silicon wafer and patterned, and a KOH etch was
used to form trenches of various widths. Before removing the photoresist, a gold film
was deposited to coat the inside of the trenches. The photoresist was then removed along
with any gold layer on top of the photoresist and germanium oxide nanowires gown as
described above. The SEM image in Figure 129 shows that the germanium oxide

nanobridge gew across the 70 um trench.

 

Figure 129. The nanowires gew across the trench. The scale bar is 50
pm.
The gold was deposited on both sidewalls of trenches and the germanium oxide
nanowires gew predominantly only where the gold was deposited. From the image, the
nanowires gew in preferred directions that were almost perpendicular to each other, and

most nanowires were very straight. There were also some whiskers with larger diameters

175

and shorter lengths gowing on the sidewalls. The diameter of the whiskers was about

2pm.

A

‘V

. ET'W' ”"13"?" “ha . t ~
aaauaeawa $88858

fi‘
- _

5'0 i311]

 

Figure 130. Nanowires gew across the 70 pm wide trench. The scale bar is 50 um

Figure 129 and Figure 130 show the SEM images of germanium oxide nanowires
gowing across the trench 70 pm in width to form nanobridges. The gold was deposited
on the sidewall of the silicon wafer trench only. Therefore, the gowth of nanowires
mainly took place inside the trench where gold was deposited, and little to no gowth can

be found outside of the trenches.

176

 

 

Figure 131. Dense nanowires gew across the trench. The scale bar is 50 pm.

Figure 131 shows the SEM image of a high density of germanium oxide nanowires
gowing across the silicon wafer trench. The shorter nanowires, which did not gow
across the trench, show bright dots at the tip. Most nanowires were straight and long
enough to bridge the 70 um trench. Several larger diameter whiskers were found along
the sidewall and exhibit more kinking along the length of the wire. These larger whiskers
can be more clearly seen in Figure 132 where it is also easily observed that many wires

are also at an angle up and out of the trench.

177

   

Figure 132. The whiskers with large diameters gew on the side wall. The scale bar is
50 um.

   

Figure 133. The scale bar is 50 um. Figure 134. The scale bar is 50 pm.

178

Figure 133 and Figure 134 show the SEM images of germanium oxide nanowires
gowing in a wider trench, which was 85 pm in width. The density of nanowires inside

wider trench had lower density than that in narrower trench.

‘ Mfiwh’de-euastt: ..
“t A': .

.-

a

 

...- 'n A. ,> . .141“. -- '

..

. {:14 .ggfiéfigfiwdImp 101,; pm

 

Figure 135. The nanowires gew in the region with different widths.

The SEM image in Figure 135 shows that the nanowires gew inside trenches of
different widths. In this region, nanowires gowing in the narrower trench tended to
exhibit higher density and longer length. This phenomenon was also observed at the
comer of the center island where the nanowires had higher density and longer length in
the narrow trench than in the wider trench. The widths of these trenches were 80 and

115 pm, respectively.

179

 

 

Figure 136. Nanowires inside the Figure 137. Nanowires inside the trench.
trench. The scale bar is 50 The scale bar is 50 um.
um.

-—5'0 [rm—I

 

Figure 138. Nanowires gew along the corner.

Figure 136 and Figure 137 show the SEM images of nanowires gowing inside the
trench. The SEM image in Figure 136 shows that the whiskers with the larger diameters

gew in the upper portion of the trench (near the surface of the substrate). The SEM

180

 

image in Figure 137 illustrates the kinked whiskers with larger diameters emerged in the
upper part of the trench while in the bottom part, nanowires with smaller diameters were
observed. Figure 138 shows the SEM image of the germanium oxide nanobridge

intersecting with the trench.

Experiment #2: p,—
Gold nanoparticles were applied on the silicon wafer before the experiment. The s

germanium was placed on top of a silicon wafer and another piece of silicon wafer with

 

gold nanoparticle (20 nm) was placed on top of the germanium, and the stack of samples
was subsequently annealed to 850° C. Then the system left at 850° C for 40 minutes with
Ar/H (95/5) flow at 200 seem to reach 10 Torr. The Ar/H flow was then stopped and
room air was allowed to flow into the hot chamber for 1 minute to reach pressure of

350 Torr. The vacuum pump was then turned on, and the furnace turned off to cool the

chamber.

181

Result

 

Figure 139. The gowth of nanowires in the trench.

The SEM image in Figure 139 shows the nanowires gew on the sidewall of the
trench. The scale bar is 20 pm. The gold nanoparticles (20 nm) were applied to the top

of the silicon surface before the experiment.

182

 

 

 

 

Figure 140. The SEM image shows nanowires gew across the trench.

Figure 140 is the SEM image of germanium oxide nanowires gowing inside the
trench. The scale bar is 20 pm. A few germanium oxide nanowires made it across the
trench. Afier the sample was exposed to the air for 1 minutes, the nanowires gew.
Compared with the 10 nm gold deposited on the surface, the density of germanium oxide

nanowires was lower in certain areas of the silicon wafer.

183

 

 

 

Figure 141. The nanowires gew across the 80 wide trench.

The SEM image in Figure 141 shows the gowth of nanowires from the center
island (lower right of the image) toward the sidewall of the trench. The nanowires

tapered off to a point.

 

-‘

Figure 142. The scale bar is 20 um. Figure 143. The scale bar is 20 pm.

184

 

 

From Figure 142 through Figure 145, it appears that only a few nanowires crossed

the trench. Most nanowires were shorter than 20 pm in length.

   

Figure 144. The gowth of nanowires in Figure 145. The scale bar is 20 um.
the trench.
Discussion
The nanowires gown from the gold nanoparticles (20 nm) had lower density
compared with those gown from the gold layer with 10 mm thickness. Diameters of the
nanowires appears much more uniform by using the gold nanoparticles, and the larger
whiskers near the surface of the trench sidewalls were not observed when the gold
nanoparticles were used. A high density of nanowires can be seen at the surfaces of the
trench sidewalls, with a few nanowires exhibiting much higher gowth rates and

extending across the trench.

Experiment #3:
A layer of gold was deposited on the silicon wafer at high temperature. The piece
of germanium was put on top of the silicon wafer and subsequently annealed to 850° C.

After reaching 850° C, room air was allowed to flow into hot chamber to reach pressure

185

 

400 Torr. Then the KrF laser was used at 500 ml, 5 Hz for 90 seconds to ablate the gold
target. The system was then lefi at 850° C for 8 minutes with Ar/H (95/5) flow at

200 sccm. Then the pump was turned on, and the furnace was turned off to cool.

Results
The SEM image in Figure 146 shows that the whiskers grew on the sidewall of the
trench. Many whiskers show 90° bends along the length of the wire, suggesting a

tetragonal crystal structure. The dimension of these whiskers was about 4 pm in width.
F g “m

r l w.

1-

“451 i

 

Figure 146. In the SEM image of whiskers gew across the trench. The scale bar is
20 pm.

186

 

 

Figure 147. The scale bar is 10 um. Figure 148. The scale bar is 10 um.

Figure 147 and Figure 148 are the SEM images of germanium oxide whiskers

gowing on top of the silicon wafer. Half of the whiskers in the images tapered off along

 

the crystal structure to a point while the remainder abruptly terminate in their gowth.

 

Figure 151. The scale bar is 20 pm.

187

 

Figure 152. The scale bar is 20 um.

Figure 149 through Figure 152 show the SEM images of germanium oxide
whiskers gowing on top of the silicon surface. Many of the whiskers in these images

show sharp 90° corners on one or more of their edges in cross-section.

Discussion

The germanium oxide whiskers gown fi'om the laser ablated gold target had larger
diameters than pervious results and appeared with many 90° edges. The size of these
wires make such observations easier to identify. Also seen, are approximately half of the
wires tapering to a point, with the others abruptly terminating growth. For the later wires,

the geometry of the cross sections can be clearly seen.

188

 

6.3 Conclusion

In another experiment of gowing germanium oxide nanowires on top of the pre-
pattem silicon wafer, the control of the gowth of nanowires was recogiized to be
important. By placing gold on the sidewalls of the trench, the lateral gowth of
germanium oxide nanowires produced the first germanium oxide nanobridge, which is
deemed as a sigiificant breakthrough in the nanowire research. The nanobridge can be
formed with different lengths and diameters by using different methods to place gold on

the sidewall and controlling the time exposed to the air.

189

 

‘17.“

Chapter 7 Conclusion

In this study, the tube fumace system allowed the experiments to take place in a
wide range of operating environments: the pressure regions could vary from 10'3 to
500 Torr, and the temperature range could reach as high as 1000° C. This wide operating
range allowed us to investigate a variety of materials, such as gallium arsenide, indium
antimonide, SrgGa16Ge3o, and germanium, and thus to speculate, for future studies on the
conditions under which nanobridges made from these materials can be produced.

In the gallium arsenide nanowires gowth experiment, the nanowires only emerged
when an additional gold layer was deposited on top of the piece of gallium arsenide. The
EDS analysis showed that the nanowires were actually gallium oxide nanowires. To our
knowledge, this is the first observation of gallium oxide nanowires that have been
synthesized through the simple thermal annealing of gold deposited on GaAs. The
absence of gold balls at the tips of the nanowires suggests the gowth mechanism may not
be VLS-based. Instead, the solid-liquid-solid mechanism might be more appropriate to
explain the gowth mechanism. Since the Pauling electronegativity of Au is higher than
that of the elements of Group V, arsenic is absent and Au acts as an oxidizing agent and
displaces the less electronegative elements of Group V by combining the highly
electropositive Group III metals. In addition, gold plays a major role in determining the
quantity of gallium oxide nanowires. The gallium arsenide without the gold layer
showed no trace of nanowire gowth. The diameters of gallium oxide nanowires were
affected by varying the gowth temperature. The experiment at lower temperatures

produced nanowires with larger diameters. From these results, two factors are identified

190

 

to affect the gowth of gallium oxide nanowires: the thickness of gold and the
temperature.

In the SrgGamGe3o nanowire gowth experiment, we produced the nanowires from
SrgGamGem by simply mixing it with Au nanoparticles and using a local vapor
confinement technique. No gowth of nanowires took place without the presence of gold
nanoparticles. In the images shown in chapter 5, the gold balls appeared at the tips,
which was an indication that the gowth mechanism was the well-known vapor-liquid-
solid mechanism. From the TEM analysis, the nanowires were found to be gallium oxide
nanowires, and, strontium and germanium were not present in the final component of
these nanowires. In addition to gold, the second factor influencing the gowth of
nanowires from SrgGa16Ge3o was vapor pressure. After a piece of quartz plate was added
on top of the sample, the quantity of nanowires dramatically increased. Nanowires thus
gew after a piece of quartz plate was added. Most importantly, these newly formed
nanowires had uniform diameters.

In the experiment of gowing nanowires from germanium, the germanium
nanowires were gown using the gold mediated VLS gowth mechanism. After a series
of chemical reactions, germanium oxide nanowires were produced below its melting
point. There are two critical factors that affected the gowth of nanowires. First, gold
was an important factor. Our experiment showed the absence of gold resulted in no
gowth of nanowires; however, the presence of gold at the tip of wires, as verified by the
EDS analysis, suggested the VLS gowth mechanism. Second, oxygen was also critical

in the nanowires gowth process: No gowth will occur if oxygen is not involved in the

191

v
1‘.

 

system. Additionally, the size of nanowires was determined by the time they were
exposed to air at high temperature.

In another experiment of gowing germanium oxide nanowires on top of the etched
silicon wafer, control of the gowth of nanowires was important. By placing gold on the
sidewall of the trench, the lateral gowth of germanium oxide nanowires produced the
first germanium oxide nanobridge: a breakthrough in our nanowire research. By using
different methods to place gold on the sidewalls and controlling the time exposed to the
air nanobridge can be formed with different lengths and diameters.

In the nanoscale world, the study of materials involves the selection of appropriate
mechanisms, as well as the considerations of the interference from different variables.
Even a small change in the experiment condition will influence the result of the gowth to

some degee.

192

 

Chapter 8 Scanning Electron Microscope

Initially commercialized in 1965, the scanning electron microscope (SEM) has
been identified as an exceedingly valuable instrument in terms of scientific contributions.
In general, the popularity and advantages of SEM can be demonstrated in various ways.
For instance, SEM has a large depth of field, in which the sample can be precisely
focused at one time. Also, its magnification is up to four hundred times geater than that
of a light microscope.

Currently the resulting image quality derived from the SEM can be much geater
due to its depth of field. The resolution of SEM is comparatively higher than that of a
light microscope, the feature that enables the sample to be investigated at a higher
magnification than what is possible while using a light microscope. Specifically, this

instrument has a wide range of magnification, normally between about 10x and 100,000x.

Theory of SEM Operation

A basic diagam of an SEM is shown in Figure 1. The electron gun, which is
located at the upper part of the column, produces an electron beam. The top portion is the
electron source, a hair-thin and V-shaped filament of fine tungsten wire. It is being
heated by passing a high voltage and low amperage current. Electrons boiling off it are
divided along the column as a tightly coherent beam, aimed directly at the specimen. The
beam accrues through the anode and is condensed by the condenser lens, followed by the

step of being focused as a very finely-tuned point on the specimen via the objective lens.

193

l

 

The cohesiveness and direction of the beam is controlled by electromagnetic lenses.

 

  
  
  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Electron
gun
Condenser
ens
Scan Coils
Scan
Generator Objective
lens
Computer E] n\ Final
\ aperture
C th d -
tulieo e ray \\ Electron
beam
Vi
Preamplifier r—— Sample
Photomultiplier Tube/ \ Scintillator
Light pipe
Faraday ca e

 

 

 

Figure 1. Schematic of an SEM [72]

The electron beam is deflected back and forth by a magnetic field which is produced by
coils. The coils are located within the objective lens, and are energized by the scan

generator. This varying voltage is produced by the scan generator.

194

 

When a beam of electrons strike on an object, it affects other electrons to be loosely
attached to the object itself. These secondary electrons, primarily collected by the
detector near the specimen, scatter in directions determined by the angle of incidence of
the beam, and by the surface topogaphy of the specimen. The signals are converted to a
voltage, and then amplified. The amplified voltage is then applied to the gid of the
cathode-ray tube (CRT) and modulates the intensity of the brightness of the spot on the
CRT. This modulation is meant to demonstrate that the brighter spot on the CRT is
produced by a larger voltage in the detector. This larger voltage is created by a larger
number of secondary electrons positioned comparatively higher on the sample surface.
Then, the SEM image is composed of thousands of spots characteristic of varying

intensity on the face of a CRT that corresponds to the topogaphy of the sample [72].

195

Everhart-Thornley Detector

The Everhart-Thomley detector is a major component of the SEM. In short, the
innovative Everhart-Thomley detector has made the formation of images using the
secondary electron signal possible. Moreover, the corresponding result is higher in terms
of potential resolution when using this signal.

The frontal facet of the detector is mainly made of a structure called Faraday cage or
collector screen. By definition, since the Faraday cage either takes the form of a wire-
mesh or a metal ring, it is usually maintained at a positive potential on the order of a few
hundred volts, which is intended to efficiently collect nearly all of the secondary
electrons emitted from the sample. The scintillator placed at the end of the Faraday cage
has a thin conductor coating so as to be maintained at a higher positive potential of
several kilovolts. If the secondary electrons are inside the Faraday cage, they will speed
up by the force of highly positive electrical charge applied to a scintillator acting as a
collector. When the secondary electrons hit the scintillator, light is generated thereafter
and is directed towards the photomultiplier through the light pipe. Then the
photomultiplier outputs a signal pr0portional to the number of secondary electrons
collected.

Photons which enter the photomultiplier tube collide with the first electrode,
inducing it to emit electrons. Then inside the tube, the electrons leap back and forth

among a series of charged plates, and the number of electrons is increased in a cascading
way, which eventually creates a gain of 106 electrons with each bounce. Subsequently,

the output voltage of the photomultiplier is amplified to a larger degree by the

preamplifier, whose output voltage finally modulates the intensity of the spot on the CRT.

196

 

There is one factor influencing the output signal from the electrons: the position of the
scintillator. In normal situation, it is located line-of-sight with the sample. Due to this
variable, the backscattered electrons that regularly are too active to be deflected to a large
degree by a 200 V potential will also output a sigral in the detector, even under the
condition that a negative potential is applied to the Faraday cage. In this situation, we
term it as a pure backscattered electron detector. By adjusting the cage potential, the
tweaking of the topogaphic contrast given by the detector is made possible.

Typically, the Everhart-Thornley detector can detect most of the secondary electrons
that depart from the surface of the specimen since these electrons are low in energy, and
can be deflected simply through angles of extremely large size. Moreover, the detector
can identify the backscattered electrons, which are openly moving toward the scintillator.
Due to the high energy carried, however, the backscattered electrons are not easily
deflected.

The Faraday cage is a required tool for shielding the electron beam from the 12,000-
V positive electrical charge on the scintillator. The absence of the cage will cause the
charge to deflect the electron beam. Nonetheless, since the action of the 12,000-V force
field is limited by the Faraday cage that is external to this field, the beam of electrons is
impacted merely by the 300-V force field of the Faraday cage itself [72].

The main function of scintillator is to convert the secondary and the backscattered

electrons into a burst of light.

197

 

The Display Output

The display output produces the image in the form of a television-type picture on a
CRT and the images from the SEM are produced on the CRT. The CRT beam is not
static. Rather, it swings back and forth across the face of the tube and its swing is
synchronized with that of the electron beam in the column. A SEM usually has two types
CRTs, one of which is for recording while the other is for viewing. The viewing CRTs
exhibit a delay of approximately several seconds after the scan moves across the screen.
As the image persists for a while, it is easy to be observed. However, in usual cases, the
viewing CRT is exceptionally coarse-gained and with poor resolution. With this
consideration, the work of photogaphy is often achieved via a recording CRT that has a
superior gain as well as the higher resolution. A camera is mounted in front of the CRT,

and the image is recorded line by line. [72].

198

 

Specimen-Beam interactions

Specimen interaction is what makes Electron Microscopy possible. When a beam of
electrons hits on an object, it knocks other atoms loose from the object itself. The
interactions generate diverse secondary products, such as electrons of different energy,
and x-rays. The main purpose of these secondary products is to produce the images of
the sample as well as to collect additional data from the sample. That information can use
to form the secondary electrons image, backscattered electrons images and EDS analysis.

When the electron beam arrives to the surface of the sample, some of the electrons
keep moving into the sample. The electrons which move inside the sample might be
interacting with the sample atoms. The interaction between electron and sample atoms
would give the sample atoms energy to scatter. At the same time, the electrons energy
would lose because the energy is used for scattering the sample atoms. And that also
changes the direction of the electron movement. Because of many interactions, the atoms
might gradually lose energy. The interaction is a scattering process where no sharply
defined limits exist regarding the boundary of the scattering. A defined limit may be
placed upon the incident electron energy left as well as on the region of the interaction
specified. Thus the illustration would focus on the area where the interactions occur with
most possibility. Typically, the interaction is identified as pear-shaped. However, from
the samples observed, the volume, both in terms of the measurement of depth and width,
of the interaction change directly with the accelerating voltage, as shown Figure 2, and

inversely with the average atomic number of the sample.

199

 

 

 

Electron Electron Electron

 

beam beam beam
Sample
surface
LOW Medium .
accelerating accelerating High .
voltage voltage accelerating
voltage

Figure 2. Interaction-volume variation with accelerating voltage.

The interactions between incident electrons from the electron beam and the sample
atoms are described to be both elastic and inelastic. Elastic interactions take place
between incident electrons and the nucleus of atoms of the sample that are distinguished
by a large-angle deflection of the incident electron. Meanwhile, the minor energy loss by
the incident electron is also observed. On the other hand, inelastic interactions occur
between the incident electrons and the orbital shell electrons of the atoms of the sample,

which are characterized by a small-angle deflection of the incident electron [72].

200

Secondary Electrons Produce the Standard SEM images
In scanning electron microscopy, an inelastic interaction is responsible for
producing secondary electrons, which are mostly created by interactions between incident

electrons and weakly bound conduction-band electrons in the atoms of the sample.

    
 
  
 

Electron

Many beam

secondary

electrons

escape
Few
secondary
electrons
escape

43

Figure 3. Secondary electrons escape, edge effects, and image production [72].

Because the average energy of secondary electrons is about 3 eV to 5 eV, which is
evaluated as low energy, secondary electrons are thus attracted easily by the positive high
voltage on the Faraday cage. This attraction can detect all secondary electron in all
directions. From the standard image in the SEM, it could be observed that secondary
electrons are the main components. The creation of Secondary electrons occurs in the
entire region of the specimen-beam interaction. Nonetheless, due to their
characteristically low energy, they are strongly absorbed by the sample while only those
produced near the surface of the specimen can break out with the aid of the decreased

path to the surface. There are, therefore, two factors influencing the contribution of

201

secondary electrons in the SEM image: the absorption and escape of secondary electrons.
These factors can dictate whether secondary electrons are capable of producing a
predominantly topogaphical image in the SEM. It is found that small projections on the
sample surface occupy the areas with a shorter path length of the escape of secondary
electrons than do the flat areas as shown in Figure 3. These protruding areas emerge
brightly on the image. Since the escape of secondary electrons takes place with a small
volume among the total amount of the specimen-beam interaction, the image of the
secondary electron is able to provide the image of highest resolution [72].
X-ray Analysis

During the specimen-beam interactions, a beam of electrons impinges on an object;
it knocks other atoms loose from the object itself. The interactions generate diverse
secondary products, such as electrons of different energy, and x-rays. That information
can use to form the secondary electrons image, backscattered electrons images and EDS
analysis. In terms of the origin, x-rays are produced from the whole area of the specimen-
bearn interaction, found as secondary electrons; back scattered electrons, and Auger
electrons. Nonetheless, x-rays are photons, rather than electrons. Furthermore, they have
a much lower probability of losing their energy due to inelastic scattering. X-rays are
produced in the electron microscopes, such as SEM and TEM. Accessory attachments
enable the detection and analysis of x-rays, which can be used to determine the presence,
amount, and distribution of the elements in the sample. What renders x-ray analysis
beneficial is that there is no need to destroy the sample. Moreover, it is complete on a
spatial basis, which means that there is a comparison of composition available between

one part and an adjacent part within a sample.

202

 

Unlike many analytical techniques, x-ray analysis is not as sensitive comparable to
neutron activation, atomic absorption, and inductively coupled plasma spectroscopy,
which have limits of detectability measured in parts per million or better. Typically, the
value of concentration of a specific element that can be detected for an entire sample in
an SEM is 0.08% by weight, while the counterpart for a thin section in the TEM is
0.004% [72].

It can be defined that x-rays originate from inelastic scattering as a vacancy
existing in an inner orbital shell packed with an electron from a shell of higher energy. It
is this difference of energy between the shells that is emitted in the form of an x-ray. The
element of the energy and the wavelength of the x-ray created is the main attribute. Its
energy and wavelength are related by the equation lt=1.2398/E, where it is the wavelength
in nanometers, and E is the energy in keV. In fact, the actual energy of x-rays might
change, ranging from 2 to 3 eV based on the Heisenberg Uncertainty Principle.

Each x-ray produced is given a name based on the name of the shell (K, L, M, N)
where the vacancy was created as well as on the number of orbital shell jumps made by
the electron that filled the vacancy. Essentially, an element often produces more than one
type of x-ray because multiple beam electrons strike the sample with each producing a
different type of interaction. Each type of x-ray is called a line. If there are sufficient x-
rays of a given line, an x-ray line peak in the spectrum analyzed is created. Each x-ray
line has a representative energy and wavelength, and those x-rays from a given shell are
termed as a family, such as the K family line and the M family line.

In order to remove an electron from a given shell, the beam of electrons must have

certain energy, called the critical excitation energy, especially under the condition that the

203

 

energy is not sufficient, resulting in no production of x-ray lines from that shell (or
subshell). The energy of the beam is decided by the accelerating voltage of the electron
microscope.

While conducting analyses of these specific elements, critical excitation energy
differences between elements will lead to differences in the analytical spatial resolution.
In thick samples the beam of electrons eventually will lose all of their energy, however in
thin sections only a small percentage of the energy is lost. The differences between the
analytical spatial resolutions of the various elements are more discernible in whole
samples than they are in thin sections. Thus, the main differences in the analytical spatial
resolution for different elements are primarily ascribed to the geater variation in energy
of the beam electrons in thick samples.

Additionally, the production of characteristic x-rays is accompanied by the creation
of noncharacterisitc x-rays. These x-rays are collectively called backgound x-rays,
continuum x-rays, or bremsstrahlung x-rays, which are produced following the
deceleration of the beam electron when it enters the coulombic field between the nucleus
and the tightly bound electrons. The energy of the background x-rays varies from zero to
the energy of the electron beam. Usually, in a normal spectrum, characteristic x-ray

peaks are placed on top of the background x-rays [72].

204

 

Measuring the energy of x-rays

The elements in a sample can be determined by measuring the energy of the x-rays
produced. The measurement of energy is called energy-dispersive spectroscopy (EDS).
The EDS detector attaches to the column of an electron microscope. A number of
components, including collimator, window, detector crystal, field-effect transistor, and a

liquid nitrogen dewar, comprise the detector as shown in Figure 4 [72].

 

 

 

 

Stray l 1‘
X ray \\ l

Electron
Beam

X ray
W ......
3/

\ \ Collimator
\ Crystal

Flela-CIICCI transistor

 

 

 

 

 

 

 

 

 

Figure 4. Schematic of an EDS detector attached to an SEM [72].

205

 

 

 

Peak Identification
Peaks will be automatically specified during the spectrum accumulation.

Nonetheless, to identify a more accurate peak, a number of things need to be consider,

1. Confirm that only these seven elements are excluded: H, He, Li, Fm, Md,
No, Lr.

2. Ensure that the peak intensities match those of the displayed lines, such as
Ka and KB pairs.

3. Select a small peak. In order to observe all the possible fits for the
designated peak, it is necessary to check from Be to Z+. If any of the
displayed KLM lines approach the peak, then it is imperative to investigate

the match and decide if it is a valid match.

Discussion
To avoid misidentifying by software in the use of EDS identification, the steps as
mentioned earlier need to be followed to obtain the accurate data. Moreover, since the
default line is not always the best, another necessary step is to determine if the best KLM
line was used in the quantitative analysis calculation. Thus, it is necessary to select the
line with highest counts that will be beneficial to acquire the quantitative analysis with

more accuracy.

206

 

Chapter 9 System set up

The photogaph of the furnace system are shown in Figure 157. The outer
diameter of the ceramic fumace is 10 inches while the inner diameter is 4 inches. The 48-
inch long quartz tube housing the substrate and the target has a diameter of 85mm. The
ends of the quartz tube are covered with the air-cooled end caps made of stainless steal.
The vacuum was provided by the Tri-Scroll roughing pump (dry pump). A ConvecTorr
vacuum gauge from Varian was used to measure the pressure inside the tube. A K-type
thermocouple was placed between the quartz tube and the furnace, and was connected to
a PID fuzzy logic-based controller. The sample and target were placed on top of a quartz
boat that was placed inside the tube near the center of the furnace. The laser beam then
passed into the furnace through the fused silica optical window, near the Ar gas inlet.

The following figures show the Autosketch drawings and the photo gaphs for the
stainless steel end caps that were designed and manufactured for the furnace system. The
figures below show the drawings and the photogaphs of the optical-side endcap, which
was desigied to accommodate the transmission of laser into the furnace as well as the gas

inlet. The second endcaps shows the gas outlet for a vacuum port to the chamber.

207

 

 

Figure 157. Quartz tube furnace system

208

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

T ..SNN l! :BNO l
11 IT .83 gang...
tl/
.3. , MW/
; a 3+
. /
/./
.. // 4
N I, // O
.86 / ...
..mmNé .950 bum.” nomad I L
N ..
..N
\ \.. x
t x .. x
f .N. \\ O
8.5.. _ .
l L.— IMWVO \W..\\
E _ ....
.t. _ k
..mmmd
I'lnL
B .23.: com
.92. .11 1|: 8.0... 2:33

Figure 158. Autosketch of optical-side drawing

209

 

Illulbflwllll

 

 

..mNN.v .95.»

 

 

“—

..mnmd

 

 

 

 

I'll

v.1[l

 

 

 

 

 

..mwmd

 

 

..nrnd LI!

 

 

 

med a It
. 8
‘I .meo yfiaiwl/
.1. r
F :
.83 d A w \ V ‘
, ....
o /..,/ \\ 0
xx // \ x
. x. x \\
.x// ./ \\\\\

 

 

 

./. \.
I’ll/III: .I:III!I.IX\1\..\ \HW\
.wa ..-...-..\u...
n O 0 {fax 0
r _ x
on 557v a co m

Figure 159. Autosketch of endcap

210

 

 

 

 

 

..omm.m

 

 

 

 

 

 

 

 

 

 

 

 

..omN.m 1‘

Sagaosv
sfipsoue: lJ

 

 

 

 

fio o

 

B .93.: 8m \
8894

 

Figure 160. Autosketch of endcap

211

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

:
l ' . i
C U
Q J
-p.
[ i ' I
L] L. —I._l

 

 

 

 

 

 

Figure 161. Autosketch of endcaps assembled

212

1/4-20 thru holes

APPENDIX A

Systems/Components Specifications

This section describes the components that have been used during the research. It is
separated into several categories according to the list shown below. Components that
relates to vacuum are treated entirely under Vacuum Technology in Section E.

1. Furnace system

2. KrF Excimer Laser

3. Optics

1. Furnace system
The quartz tube was used to house the substrate/target in the filmace system. The
inside diameter of the tube was 85mm. Some of the key properties of the quartz tube are
good thermal shock and stress resistance, low thermal conductivity, good strength, wear
resistant, and most importantly, usable at high temperature. Some other properties are

listed below.

Table 6. Properties of quartz tube

 

 

 

 

 

Density 2.2 g/cm3
Thermal Conductivity 1.4 W/mK
Coefficient of Thermal Expansion 5. 5 x 10-7/C
Maximum Use Temperature 1683 °C
Manufacturer Technical Glass Products, Inc

 

 

 

 

213

 

2. KrF Excimer Laser
The Lamda Physik LPX 210i KrF Excimer Laser is used to ablated the targets

inside the furnace. The specifications for this laser are listed below.

Table 7. Specifications for Lamda Physik 210i KrF Excimer Laser

 

 

 

 

 

Wavelength (nm) 248
Maximum Repetition Rate (Hz) 100
Pulse Duration (ns) 25
Beam Dimension (mm2) 5'12 x 23
Average Power(W) 65

 

 

 

 

3. Optics
The optics used in the experiment comprises of a UV-graded optical materials
which includes plano-convex lens, laser mirror, fused silica optics window and Plexiglas
beam box.
a. Plexiglas Beam Box
The purpose of the UV—graded beam box is to protect the researchers during the

laser ablation. The specifications are listed below.

 

Material UV blocking plexiglas

 

Dimension (inches) 48 x 7.25 x 14.25

 

 

 

 

Table 8. Beam box specifications

214

 

b. Optics Window
Deep UV optics window was used to couple the laser into the furnace system. The

specifications are listed below.

 

 

 

 

 

Material Deep UV grade fused silica
Surface Quality 40-20
Wavelength (nm) 0.2 - 2 pm
Nominal Flange Diameter 3.7 inches and 2.75 inches
Manufacturer Insulator-Seal

 

 

 

 

Table 9. Specifications for the optical window

c. Plano-Convex Lens
The round UV—graded fused silica plano-convex lens is used to direct and focus the

laser beam onto the target. The specifications are listed below.

Table 10. Specification for the plano-convex lens

 

 

 

 

 

Material UV—grade fused-silica
Wavelength (nm) 248
Diameter (inches) 2

Surface Quality 60-40
Manufacturer Oriel Optics

 

 

 

 

215

 

d. Laser Mirror

The laser mirror was used to redirect the laser beam in the experiment conducted in

the fumace system. The specifications are listed below.

 

Material

UV grade fused silica

 

Dimension (diameter x thickness)

2 x 0.375 inches

 

Surface finish

20-10 laser grade polish

 

Reflectance (248 nm)

> 97% (45°AOI)

 

Manufacturer

 

 

Acton Research Corp

 

Table l 1. Specification for the high power UV laser mirror

e. Optical Stand

The two aluminum optics stand was used to support the lens and the mirror. A

single piece of the optics stand is shown in Figure. The specifications are listed below.

 

Material

Aluminum

 

Dimension (inches)

 

 

24 x 2.65

 

Table 12. Specifications for the aluminum optics stand

216

 

 

 

APPENDIX B

Vacuum Technology

1. Pressure Units

The System International (SI) pressure unit is Pascal (Pa). It corresponds to 1

2 . . .
N/m . A umt of pressure un1t commonly used 15 torr, where pressure of one atmosphere

at the sea level (0°C) corresponded to 760m in height of mercury column. The
following relationship are used for pressure conversion from one unit to another.

1 torr = 133.3Pa

1 atm = 1.013 x 105 Pa = 760 torr

2. Vacuum pump
The Dry scroll roughing pumps are used to provide the vacuum in the system.
3. Pressure Measurement

a. Vacuum gauge controller.

The senTorrTM vacuum gauge controller was used to monitor the pressure level
inside the system.

b. Thermocouple Gauge

Thermocouple gauge are rated to measure the pressure in the range of atmosphere

down to 10.3 torr. In this study, ConvecTorrTM thermocouple gauge from Varian

is used as shown.

217

 

APPENDIX C

Germanium Dioxide

Its erythropoietic effect, the possible bacteriocidal or fungicidal effect of

organogermanium compounds, prompts the stimulation of the formation of red blood

cells (srythrocytes) by injections of GeOz, this was first reported in 1922, raising a series

of interest in the treatment of anemia. The effect appears to vary with individuals and
with the kind of anemia being treated. Moreover, the curing results are merely
temporary. In all cases, the effect is less evident than with repeated small doses of

arsenic trioxide. Nonetheless, the consequent danger of poisoning also is comparatively

less. Based on the hypothesis that the low solubility of Ge02 in water might be blamed

for the marginal and inconsistent results, higher levels of germanium in the form of the

highly soluble dimethylgermanium oxide were administered to hamsters. While these

demonstrated the non-toxic behavior of (CH3)zGeO, the effect on erythrocyte levels was

inconclusive.

218

 

APPENDIX D

Buffer Etches

To fabricate the windows through the oxide layer, a standard lithography technique was
used. A positive photoresist (Shipley 1813) was used, followed by a buffered
hydrofluoric acid etch of the Si02 film to open windows through the oxide layer to the

silicon substrate.

It is important to note that Si02 is an amorphous material, which etches equally

well in all directions, exhibiting isotropic etching. In other words, when an oxide etch
depth of lnm is required, a lateral etch of lnm will also occur. Furthermore, the

thickness of any subsequently deposited layer is directly impacted by the wall shape and
slope angle of the etched Si02. The controllable parameters in wet etching are time,
temperature, solution concentration, and solution recirculation.

The etching mechanism of Si02 has been described by Prigongine, et al. as follows:

 

Protons are adsorbed to oxygen on the surface featuring the
I. strongest basicity, which functions as the proton accepter.
11. Oxygen which adsorbs the proton needs the valence electron.
Oxygen obtains the valence electron from neighboring silicon
111- which has many electrons.
As silicon gives the valence electron to oxygen, the electron
IV. density around silicon decreases. Consequently the silicon-oxygen
bond gets weaker, and the bond is eventually broken.

 

 

 

219

 

 

V. When the silicon-oxygen bond is broken, silicon becomes positive.

VI. As a result, HFz' is coordinated to facilitate etching.

 

 

 

Si02 (s) + 4HF(aq) —+ SiF4 (g) + 2H20

Hydrogen fluoride (HF) is rarely used at full strength because of the consequent

high etch rate, which is difficult to control. In our experiment, we used 1:8 HF to

ammonium fluoride (NH4F).

In a study of various dilute HF solutions, in which NH4F was used as a buffering

agent, Judge used the conditions and constants described by Mesmer and Baes to

determine the effect of the solution composition on the etch rate of undoped Si02. The

results of this study are presented in Table 13.

Table 13. Calculated Concentrations and Associated Etch Rates.

 

 

 

 

Calculated Concentrations Si02 Etch
Rate
Mx103°T=25°C ° .
’ A/ mm
[HF] [HFZ’ ] [F’] [Hi ]
13 87 677 0.0255 44.34

 

220

 

 

41 159 409 0.129 87.12

 

 

 

 

81 189 242 0.435 1 12.26
109 191 182 0.778 124.08
175 173 102 2.24 133.92
266 128 49.4 7.01 118.26

 

Using thF to dilute the etchant is common in the semiconductor industry and is
frequently termed Buffered Oxide Etch (BOE). Parisi and his coworkers found that the
activation energy for the dissolution of Si02 in BOE solutions was 9.9 kcal/mole.
Furthermore, based on their empirical results, they determined that the depth etch rate

(7) could be approximated by the following rate equation:

7 = 4.5 x109[HF]exp(—4980/T)

where 7 is in angstroms per minute, [HF] is the molar concentration of HF, and T is the

absolute temperature in degrees Kelvin. This equation is valid for etchant solutions with
molar concentrations of HF between 0.9 and 3.8 moles per liter and within the
temperature range of 25 to 55° C.

The taper of the resulting slope is a function of the depth etch rate and the lateral

etch rate (7,, ). The angle 9 that the taper forms with the wafer surface is given by
sin 6 = 7 / ye

A variety of factors have a direct influence on the lateral etch rate. These factors

include the type of the photoresist used and the condition of the exposed

221

 

oxide/photoresist interface prior to the etching. In addition, if a wetting agent is used, a
correlation between the lateral etch rate and the HF concentration or temperature cannot
be precisely defined, given the variety of influences on this etch rate. However, it has
been discovered that the edge profiles of the taper follow the trend of increasing with

higher molarity.

222

 

References

1 Published in the February 1960 issue of Caltech's Mineerinflnd Science, and
available at l_|ttp:l/www.zyvex.com/nanotecyfeynmanmtml (09/06/2005).

2 An excellent summary of many research efforts can be found at:

http://www.nano.gov/htmI/research/rnews.html (09/06/2005).

3 B. Messer, J. H. Song, and P. Yang, “Microchannel Networks for Nanowire
Patterning,” J. Am. Chem. Soc., vol 122, pp 10232-10233, July 2000.

4 C. M. Lieber, “One-Dimensional Nanosturctures: Chemistry, Physics &
Applications”, Solid State Communications, vol 107, pp 607-616, 1998.

5 Y. Liu, K. Ishii, T. Tsutsumi, M. Masahara, T. Sekigawa, K. Sakamoto, H.
Takashima, H. Yamauchi, and E. Suzuki, “Systematic Electrical Characteristics of
Ideal Rectangular Cross Section Si-Fin Channel Double-Gate MOSFETs Fabricated
by a Wet Process,” IEEE Transactions on Nanotechnology, vol. 2, no. 4, pp. 198-204,
2003.

6 Givargizov, E. I., “Grth of whiskers by the vapor-liquid-solid mechanism,” In E.
Kaldis, Ed., Current Topics in Materials Science, vol. 1, pp. 79-145, 1978.

7 A recent definition of nanotechnology is “Nanotechnology is the understanding and
control of matter at dimensions of roughly 1 to 100 nanometers, where unique
phenomena enable novel applications. Encompassing nanoscale science, engineering
and technology, nanotechnology involves imaging, measuring, modeling, and
manipulating matter at this length scale.At the nanoscale, the physical, chemical, and
biological properties of materials differ in fundamental and valuable ways from the
properties of individual atoms and molecules or bulk matter. Nanotechnology R&D is
directed toward understanding and creating improved materials, devices, and systems
that exploit these new properties.” This definition can be found at :

http://www.nano.gov/html/facts/whatlsNano.html (10/03/05).

8 M. H. Dvoret, D. Esteve, and C. Urbina,” Single-electron transfer in metallic
nanostructures ” Nature, vol. 360, pp. 547, 1992.

223

 

9 A. P. Alivisatos,” Semiconductor Clusters, Nanocrystals, and Quantum Dots
” Science, vol. 271, pp 933, 1996.

10 B. Gilbert, F. Huang, H. Zhang, G. A. Waychunas, and J. F. Banfield, “Nanoparticles:
Strained and Stiff.” Science 305: 651-654, 2004.

11 M. A. Ratner and D. Ratner, “Nanotechnology,” Prentice Hall PTR, Nov. 8 2002.

12 X. Duan, Y. Huang, R. Agarwal and C. M. Lieber, "Single-Nanowire Electrically
Driven Lasers," Nature 421, 241-245 (2003).

13 M. Law, J. Goldberger, and P. Yang, “Semiconductor nanowires and nanotubes,”
Annu. Rev. Mater. Res. Vol. 34, pp. 83-122, 2004.

14 LB. Hicks and MS. Dresselhaus, Phys. Rev. B 47, p. 12727, 1993.
15 Y.-M. Lin and MS. Dresselhaus, Phys. Rev. B 68, 075304, 2003.

16 H. J. Dai, E. W. Wong, Y. Z. Lu, S. S. Fan, and C. M. Lieber,” Synthesis and
Characterization of Carbide Nanorods” Nature, Vol. 375, pp769, 1995.

17 J. Hu, T. W. Odom, and C. M. Lieber, “Chemistry and Physics in One Dimension:
Synthesis and Properties of Nanowires and Nanotubes” Acc. Chem. Res. Vol. 32 pp
435, 1999.

18 H. Namatsu, Y. Watanabe, K. Yamzaki, T. Yamaguchi, M. Nagase, Y. Ono, A.
Fujiwara, and S. Horiguchi, “Fabrication of Si single-electron transistors with precise

dimension by electron-beam nanolithography,” J. Vac. Sci. Technol. B, Vol. 21, (1),
Jan/Feb 2003.

19 X. Duan, and C. M. Lieber, ”General synthesis of compound semiconductor
nanowires,” Advanced materials, Vol. 12, No. 4, pp298-302, 2000.

20 R. S. Wagner, W. C. Ellis, “Vapor-Liquid-Solid Mechanism of Single Crystal
Growth,” Applied Physics Letters, vol. 4, no. 5, pp. 89-90, 1964.

224

 

21

22

23

24

25

26

27

28

29

30

31

J. Westwater, D. P. Gosain, S. Usui, “Si Nanowires Grown via the Vapour-Liquid-
Solid Reaction,” Physica Status Solidi. A, vol. 165, pp. 37-42, 1998.

E. I. Givargizov, A. N. Stepanova, L. N. Obolenskaya, E. S. Mashkova, V. A.
Molchanov, M. E. Givargizov, I. W. Rangelow, “Whisker Probes,” Ultramicroscopy,
vol. 82, pp. 57-61, 2000.

I. W. Rangelow, F. Shi, P. Hudek, P. Grabiec, B. Volland, E. I. Givargizov, A. N.
Stepanova, L. N. Obolenskaya, E. S. Mashkova, and V. A. Molchanov,
“Micromachined ultrasharp silicon and diamond-coated silicon tip as a stable field-
emission electron source and a scanning probe microscopy sensor with atomic

sharpness,” Journal of Vacuum Science and Technology B, vol. 16, no. 6, pp. 3185-
3191, 1998.

J. Hu, T. W. Odom, C. M. Lieber, “Chemistry and Physics in One Dimension:
Synthesis and Properties of Nanowires and Nanotubes,” Accounts of Chemical
Research, vol. 32, no. 5, pp. 435-445, 1999.

E. I. Givargizov, “Growth of Whiskers by the Vapor-Liquid-Solid Mechanism,”
Chapter 3 in Current Topics in Materials Science, Volume 1, ed. E. Kaldis, north-
Holland Publishing Company, 1978.

R.-Q. Zhang, Y. Lifshitz, S. T. Lee,“Oxide-Assisted Growth of Semiconducting
Nanowires,” Advanced Materials, vol. 15, no. 7-8, pp. 635-640, 2003.

Fundamentals of Semiconductor Processing Technology, Badih El-Kareh, pp. 142
Fundamentals of Microfabrication, M. J. Madou, CRC Press, Mar. 13, 2002.

Handbook of Physical Vapor Deposition (P VD) Processing, D. M. Mattox, William
Andrew Inc., Mar. 1. 1998.

Handbook of Nanophase and Nanostructured Materials, Z. L. Wang, Y. Liu, and Z.
Zhang, Springer, pp204, 2002.

Pulsed Laser Deposition of Thin Films, ed. D. B. Chrisey and G. K. Hubler, New
York, John Wiley & Sons, Inc, 1994.

225

32 http://www.photonics.cusat.edu/Research Laser%20[nduccd%20Plasma.html
(as 10/03/05)

33 Pulsed Laser deposition of thin frilms, D. B. Chrisey. Chap. 6.

34 Phase transformations in Metals and Alloys, D.A. Porter.

35 R. S. Wagner, W. C. Ellis, “Vapor-Liquid-Solid Mechanism of Single Crystal
Growth,” Applied Physics Letters, vol. 4, no. 5, pp. 89-90, 1964.

36 E. I. Givargizov, “Fundamental Aspects of VLS Growth,” Journal of Crystal Growth,
vol. 31 pp20-30, 1975.

37 J. Westwater, D. P. Gosain, S. Tomiya, S. Usui, and H. Ruda, “Growth of Silicon
nanowires via gold/silane vapor-liquid-solid reaction,” Journal of vacuum science &
Technology B, vol. 15, no. 3, pp. 554-557, 1997.

38 B. Bokhonov, and M. Korchagin, “In situ investigation of stage of the formation of

eutectic alloys in Si-Au and Si-Al systems,” Journal of Alloys and compounds, vol.
312, pp 238, 2000.

39 N. Wang, Y. H. Tang, Y. F. Zhang, C. S. Lee, I. Bello and S. T. Lee, “Si nanowires
grown from silicon oxide,” Chemical Physics Letters, Vol. 299, Issue 2, pp 237-242,
Jan. 1999.

40 Rui-Qin Zhang, Yeshayahu Lifshitz, Shuit-Tong Lee, “Oxide-Assisted Growth of
semiconducting Nanowires,” Advanced materials, 15, no.7-8, April 17, 2003.

41 N.-G Ma, C.-J. Deng, P. Yu, M. Aravind, and D. H. L. Ng, “Formation of MgO
Nanorods in the reaction zone of a Mg-CuO powder mixture by in-situ reaction,”
Philosophical Magazine, Vol. 84, No. 1, pp 69-80, Jan. 2004.

42 P. Gao and Z. L. Wang, “Self-Assembled Nanowire-Nanoribbon Junction Arrays
ZnO,” The journal of physical chemistry B, Vol. 106, Num. 49, pp12653-12658, Dec.
2002.

226

 

43

44

45

46

47

48

49

50

51

52

53

The science and engineering of microelectronic fabrication, S. A. Campbell, Oxford
university press, 2001.

X. Duan and C. M. Lieber, “Synthesis and optical properties of gallium arsenide
nanowires,” Applied Physics Letters, Vol. 76, pp. 1116-1118, 2000.

J. Hu, T. W. Odom, and C. M. Lieber, “Chemistry and Physics in One Dimension:
Synthesis and Properties of Nanowires and Nanotubes,” Acc. Chem. Res., Vol. 32,
pp.435-445, 1999.

X. Duan and C. M. Lieber, “Laser-Assisted Catalytic Growth of Single Crystal GaN
Nanowires,” J. Am. Chem. Soc., vol. 122, pp. 188-189, 2000.

W. S. Shi, Y.F. Zheng, N. Wang, C.S. Lee, and S. T. Lee, “Synthesis and
mircrostructure of gallium phosphide nanowires,” J. Vac. Sci. T echnol. B, Vol. 19,
pp.1115-1118, 2001.

C. T. Tsai, and R. S. Williams, “Solid phase equilibria in the Au-Ga-As, Au-Ga-Sb,
Au-In-As, and Au-In-Sb ternaries,” J. Mater. Res., Vol. 1, Num. 2, Mar/Apr 1986

M. B. Panish, J. Electrochem. Soc. Vol. 114, pp517, 1967.

J. R. Arthur, “Surface stoichiometry and structure of GaAs,” Surface Science, Vol. 43
num. 2, pp449-461, 1994.

Y. Zhang, P. L. Lee, G. S. Nolas, and A. P. Wilkinson, “Gallium Distribution in the
clathrates SP8G316G630 and Sr4Eu4Ga16Ge3o”, Applied Physics Letters, Vol. 80,
pp2931-2933, 22 April, 2002.

K. Saito and A. J. Ikushima, “Reduction of light-scattering loss in silica glass by the
structural relaxation of “frozen-in” density fluctuations”, Appl. Phys. Lett., vol. 70
(26), pp.3504-3506, June 1997.

P. Patnaik, Handbook of inorganic chemicals, New York : McGraw-Hill, 02003

227

 

54

55

56

57

58

59

60

61

J. C. Bailar, Jr., H. J. Emeléus, Sir Ronald Nyholrn and A. F. Trotman-Dickenson,
Comprehensive inorganic chemistry, Pergamon Press; distributed by Compendium
Publishers [Elmsford, N.Y., 1973]

Z.G. Bai, D.P. Yu, H.Z. Zhang, Y. Ding, Y.P. Wang, X.Z. Gai, Q.L. Hang, G.C.
Xiong, and SO. F eng, "Nana-scale GeO2 wires synthesized by physical evaporation",
Chemical physics letters, vol. 317, pp504-509, Feb. 2000.

P. Hidalgo, B. Mendez and J. Piqueras, " GeO2 nanowires and nanoneedles grown by

thermal deposition without a catalyst", Nanotechnology, vol. 16, pp2521-2524, Sep.
2005.

Y. Zhang, J. Zhu, Q. Zhang, Y. Yan, N. Wang, and X. Zhang, “Synthesis of GeO2

nanorods by carbon nanotubes template”, chemical physics letters, vol. 317, pp. 504-
509, Feb. 2000.

Y. H. Tang, Y. F. Zhang, N. Wang, 1. Bello, C. S. Lee, and S. T. Lee, “Germanium
dioxide whiskers synthesized by laser ablation”, Applied physics letters, vol. 74 (25),
pp. 3824-3826, Feb. 1999.

H. Y. Dang, J. Wang, and S. S. Fan, "The synthesis of metal oxide nanowires by
directly heating metal samples in appropriate oxygen atmospheres", Nanotechnology,
vol. 14, pp738-741, May 2003.

J .-Q. Hu, Q. Li, X.-M. Meng, C.-S. Lee, and S.-T. Lee, "Synthesis and
nanostructuring of patterned wires of GeO2 by thermal oxidation", Advanced
materials, vol. 14, no. 19, pp1396-l399, Oct. 2002.

X. C. Wu, W. H. Song, B. Zhao, Y. P. Sun, and J. J. Du, "Preparation and
photoluminescence properties of crystalline GeO2 nanowires", Chemical physics
letters, vol. 349, pp210-214, Oct. 2001.

62 K. Prabhakaran, and T. Ogino, “Oxidation of Ge(100) and Ge(111) surfaces: an UPS

63

and XPS study,” Surface Science, vol. 325, pp.263-271, 1995.

R. B. King, Encyclopedia of inorganic chemistry, Chichester ; New York : Wiley,
c1994.

228

 

64

65

66

67

68

69

7O

71

72

GA. Jacobson, Encyclopedia of chemical reactions, New York, Reinhold Pub. Corp.,
1946-59.

R.S. Wager, and WC. Ellis, “Vapor-liquid-solid mechanism of single crystal
growth,” Applies physics letters, vol. 4, pp89-90, March 1964.

A. P. Levitt, Whisker technology, New York, Wiley-Interscience 1970.

K. Haraguchi, K. Hiruma, T. Katsuyama, K. Tominaga, M. Shirai, and T. Shimada,
“Self-orgainized fabrication of planar GaAs nanowhisker arrays”, Appl. Phys. Lett. ,
Vol. 69, pp 386-387, 1996.

K. Haraguchi, K. Hiruma, K. Hosomi, M. Shirai, and T. Katsuyama, “Growth

mechanism of planar-type GaAs nanowhiskers”, J. Vac. Sci. T echnol. B, Vol. 15, pp
1685-1687, 1997.

M. S. Islam, S. Sharma, T. I. Kamins, and R. S. Williams, “Ultrahigh-density silicon
nanobridges formed between two vertical silicon surfaces”, Nanotechnology, Vol. 15,
pp L5-L8, 2004.

W. U. Wang, C. Chen, K.-H. Lin*, Y. Fang“, and C. M. Lieber, “Label-free detection
of small-molecule—protein interactions by using nanowire nanosensors”, Perceedings
of the National Academy of Sciences of the united States of America, Vol. 102,
pp3208-3212, 2005.

Y. Cui, and C. M. Leber, “Functional Nanoscale Electronic Devices Assembled
Using Silicon Nanowire Building Blocks”, Science, Vol. 291, pp 851-853, 2001.

Scanning and Transmission Electron Microscopy, S. L. Flegler, J. W. Heckman, Jr,
and K. L. Klomparens, New York, Oxford University Press 1993.

229

 

    
 

ITY BRA IE

11

52

MICMIG TE

1111111111111111111