Abstract
SOWERS, ANDREW THOMAS. Characterization of Field Emission Properties of
Nitrogen-Doped Diamond. (Under the direction of Robert J. Nemanich.)
This study explores the field emission properties of nitrogen-doped diamond
films. For this investigation, over 70 nitrogen-doped diamond films were deposited on
silicon and molybdenum using microwave plasma CVD with varying process
parameters. It was observed that the addition of low amounts of nitrogen initially
enhances the growth rate by a factor of 4. For [N]/[C] ratios greater than 0.3 the
growth rate decreases with increasing nitrogen addition. Raman spectra of these
diamond films indicate that the addition of nitrogen during growth reduces film
quality and increases the presence of non-diamond bonding in the films. Under
certain conditions, films can be grown which exhibit photoluminescence bands at
1.945 eV and 2.154 eV that are attributed to single substitutional nitrogen. Field
emission characteristics were measured in ultrahigh vacuum with a variable distance
anode technique. For samples grown with gas phase [N]/[C] ratios less than 10, stable
field emission was preceded by arcing events which produce damage to the film and
substrate. Samples grown at higher [N]/[C] content could be measured prior to an
arcing event.
Contrary to other reports on nitrogen-doped diamond, these
measurements indicate relatively high threshold fields (>100 V/µm) for electron
emission. A field emission model, based upon the compensation of the nitrogen
donors by in-gap defects, is proposed for these nitrogen-doped diamond films.
CHARACTERIZATION OF FIELD EMISSION PROPERTIES OF
NITROGEN-DOPED DIAMOND
by
ANDREW THOMAS SOWERS
A dissertation submitted to the Graduate Faculty of
North Carolina State University
in partial fulfillment of the
requirements for the Degree of
Doctor of Philosophy
DEPARTMENT OF PHYSICS
Raleigh
1999
APPROVED BY:
David E. Aspnes
Griff L. Bilbro
David L. Dreifus
Linda S. Plano
Robert J. Nemanich
Chair of Advisory Committee
ii
I dedicate this dissertation to my family, and wish to thank them for their
many years of love and support throughout my education and research.
ii
iii
Biography
Andrew Thomas Sowers was born on September 11, 1970 in San Diego,
California. He was adopted by his parents, Ersul and Virginia Sowers, when he was 6
weeks old. His father was a commissioned officer in the US Navy. When Andy was 2
years old, his family moved to Virginia Beach, Virginia because his father was
transferred from Alameda to Norfolk. Eight years later his family relocated to Eden,
North Carolina when his father retired from the Navy. He showed an interest in Math
and Science at an early age, which was encouraged by his parents. Andy attended
J.M. Morehead High School, and graduated in June 1988. He decided that North
Carolina State University was where he wanted to begin his college career majoring in
Physics. He received his B.S. in Physics in 1992. Soon thereafter, he began working
on his Ph.D. in Physics with a minor in Electrical Engineering, at North Carolina State
under the direction of Dr. Robert J. Nemanich. His initial work focused on the
characterization of diamond films using Raman scattering spectroscopy. In 1993, the
Surface Science Laboratory acquired an ASTeX diamond deposition system from BP
Oil in Cleveland, Ohio. Since that time, Andy’s primary research topic has been the
growth and characterization of the field emission properties of nitrogen-doped
diamond films. He completed his doctoral degree on January 15, 1999.
iii
iv
Acknowledgements
This dissertation was by no means an individual effort. When you’ve been a graduate
student for almost seven years, you encounter a multitude of people that provide
invaluable insight, assistance, and support. I would like to take this opportunity thank
the following people for their contributions to my education and research:
Dr. Robert Nemanich, for providing me the opportunity to work in the Surface Science
Laboratory with a great collection of people. I also appreciate his scientific insight,
valuable guidance, and support throughout my graduate studies at NCSU.
I would also like to thank the other members of my advisory committee, Dr. David
Aspnes, Dr. Griff Bilbro, Dr. David Dreifus, and Dr. Linda Plano for their guidance
and review of this dissertation. I am especially grateful for the commitment of Drs.
Linda Plano and David Dreifus who agreed to remain on my committee despite being
relocated from North Carolina.
Cecilia Upchurch, who is without a doubt, the world’s greatest secretary. Thank you
for helping me with a multitude of administrative details, but most of all, for being
such a great friend.
I must take this opportunity to thank all the present and former members of Surface
Science Laboratory for providing assistance ranging from bolting on a UHV flange to
solving difficult research issues.
I would also like to thank several former SSL members who were invaluable in the
initial stages of my research: Dr. David Aldrich, Dr. John Barnak, Dr. Mark
Benjamin, Dr. Sean King, Dr. Jay Montgomery, Dr. Tom Schneider, and Dr. Jaap van
der Weide. The following post-docs were also an exceptional source for information
on diamond: Dr. Steve Bozeman, Dr. Eliane Maillard-Schaller, and Dr. Peichun Yang
Many thanks to: Erica Robertson and Shawn Wagoner for help with the diamond
chamber; Brandon Ward for support with Field Emission; Dr. Leah Bergman for her
assistance with interpreting Raman and PL; Hoon Ham for his miscellaneous technical
and computer advice; and Peter Goeller for SEM analysis. I am grateful to the
following people for several recent lunch conversations: Dimitri Alexson, Jim
Christman, Steve English, Jeff Hartman, Ed Hurt, Christian Koitsch, Mike O’Brien,
Kieran Tracy, and Morgan Ware.
iv
v
PAMS Machine Shop (Tim Harvell, Frank Milkowski, Hai Bui, Mike Gregory, and
Ronna Klionsky), for helping me design and build much of the equipment used in this
research. I am especially grateful for all the rush jobs, and for your introducing me to
the “Old Fashioned” country ham biscuits.
PAMS Research (Pam Baker, Dee Gill, Nadine Ward), for providing me with
countless purchase orders, handling rush jobs, and for managing to pay all the bills,
despite my habit of losing or throwing away packing slips.
NCSU Microelectronic Fabrication Laboratory (Joan O’Sullivan, Elizabeth Condon,
Henry Taylor, and Dr. Dick Kuehn), for providing access to the microelectronics lab,
supplying assistance with equipment, and giving me valuable cleanroom experience.
The entire staff of Kobe Steel USA, Inc., for being an inexhaustible source of
information about the growth and properties of diamond.
Last, but certainly not least, I cannot sufficiently express my appreciation to my wife,
Darci Sowers, for her love, support, and understanding. I always anticipated the
writing of my dissertation would take about 3 months, and without her
encouragement, I would not have be able to finish it in less than half that time. Also, I
appreciate her time spent proofreading sections of this manuscript, the sacrifices that
she made, and for her ability to make me laugh during particularly stressful times
while performing research and writing this dissertation.
v
vi
Table of Contents
LIST OF TABLES ........................................................................................................ix
LIST OF FIGURES .......................................................................................................x
1. Introduction and Scope of Dissertation ...................................................................1
1.1. Motivation and Introduction .............................................................................1
1.2. History of Diamond ..........................................................................................3
1.3. Scientific Research on Diamond ......................................................................7
1.4. Properties of Diamond ....................................................................................10
1.5. Diamond-Based Cold Cathodes ......................................................................11
1.6. Scope of Dissertation and Outline ..................................................................14
References ..............................................................................................................16
2. Experimental Details ..............................................................................................25
2.1. Growth of Diamond by Microwave Plasma CVD ..........................................25
2.1.1. Theory of Diamond CVD ....................................................................26
2.1.2. The Microwave Plasma CVD Diamond Chamber ..............................28
2.1.3. Laser Reflectance Interferometry ........................................................32
2.2. Characterization of Diamond Films ...............................................................35
2.2.1. Raman Scattering and Photoluminescence Spectroscopy ...................36
2.2.2. Microscopy ..........................................................................................40
2.2.3. Field Emission Characterization ..........................................................41
References ..............................................................................................................47
vi
vii
3. Growth and Properties of Nitrogen-Doped Diamond Films ..................................65
3.1. Introduction ....................................................................................................65
3.2. Influence of Nitrogen upon the Growth of Diamond Films ...........................67
3.3. Characterization of Nitrogen-Doped Diamond Films ....................................69
3.4. Growth of Nitrogen Doped Diamond: Chemistry & Growth Mechanisms ...72
3.5. Other Research ...............................................................................................73
3.6. Conclusions ....................................................................................................74
References ..............................................................................................................75
4. Introduction and Review of Field Emission from Diamond ..................................82
4.1. Introduction and Historical Background ........................................................82
4.2. Field Emission Research on Metal Cathode Arrays .......................................84
4.3. Field Emission from Semiconducting Diamond ............................................86
4.4. Field Emission Measurement Techniques ......................................................87
4.5. Field Emission Results from Diamond ...........................................................88
4.6. Conclusions ....................................................................................................91
References ..............................................................................................................93
5. Field Emission Properties Of Nitrogen-Doped Diamond Films ..........................102
5.1. Introduction ..................................................................................................103
5.2. Experimental .................................................................................................106
5.3. Results ..........................................................................................................112
5.4. Discussion .....................................................................................................120
5.5. Conclusions ..................................................................................................126
References ............................................................................................................128
vii
viii
6. Summary and Future Research ............................................................................145
6.1. Summary of Results ......................................................................................145
6.2. Future Research ............................................................................................147
References ............................................................................................................149
Appendices .................................................................................................................150
Appendix A. Characterization of Highly Oriented Diamond Films ..................151
A.1. Introduction ............................................................................................151
A.2. Theoretical Background .........................................................................154
A.2.1. (100) Surfaces ..............................................................................155
A.2.2. (110) Surfaces ..............................................................................157
A.2.3. (111) Surfaces ..............................................................................159
A.3. Experimental ...........................................................................................162
A.4. Results and Discussion ...........................................................................163
A.5. Conclusions ............................................................................................169
References ......................................................................................................171
Appendix B. Thin Films Of Aluminum Nitride And Aluminum
Gallium Nitride For Cold Cathode Applications ...........................................183
References ......................................................................................................192
viii
ix
List of Tables
Table 1.1. Properties (at room temperature) and applications of diamond. ................18
Table 3.1. The classification scheme for single crystal diamond. Typical
materials properties are also shown. ..........................................................78
ix
x
List of Figures
Figure 1.1. (a) Schematic of the most commonly observed morphologies for
natural diamond single crystals. A photograph of several typical rough
diamonds is shown in (b). ..................................................................................19
Figure 1.2. (a) The Cullinan diamond (3,025 carats) as it was discovered in
the Premier mine in South Africa in 1905. The nine major gemstones,
which were cut from the original Cullinan, are shown in (b). ...........................20
Figure 1.3. Ball and stick model of the diamond crystal structure illustrating
the tetrahedral bonding of the carbon atoms. .....................................................21
Figure 1.4. The quantum efficiency for photoemission for a hydrogen
terminated (111) diamond surface as a function of incident photon energy.
Himpsel et al. suggested that the extraordinarily high quantum yield
indicated a negative electron affinity. ................................................................22
Figure 1.5. Band diagrams for semiconductor surfaces, which exhibit a
positive electron affinity (χ>0) and a negative electron affinity (χ<0). .............23
Figure 1.6. Simplified energy band diagram for an ideal negative electron
affinity cold cathode emitter. .............................................................................24
Figure 2.1. Schematic of the different processes which are suspected to lead
to the deposition of diamond by chemical vapor deposition. .............................50
Figure 2.2. Schematic of the microwave plasma CVD chamber used to
deposit diamond thin films in this study. ...........................................................51
Figure 2.3. Drawing of the ASTeX microwave plasma CVD chamber with
the inductively heated sample stage. ..................................................................52
Figure 2.4. Drawing of the ASTeX microwave plasma CVD chamber with
the modified sample stage. This configuration is compatible with the
molybdenum sample holders. .............................................................................53
Figure 2.5. A model for the reflectivity from the growing diamond surface.
In this case, the incident laser beam is normal to the diamond surface. ............54
Figure 2.6. The reflectance for the vacuum/diamond/silicon model as a
function of diamond film thickness. ...................................................................55
Figure 2.7. The reflectivity of the growing diamond surface taken during a
diamond film deposition. ....................................................................................56
x
xi
Figure 2.8. A schematic of the integrated UHV surface characterization and
processing system. ..............................................................................................57
Figure 2.9. Schematic of possible photon-lattice scattering events. The
scattering due to optic vibrations or molecular vibrations/rotations is
termed Raman scattering. ...................................................................................58
Figure 2.10. Raman scattering spectra obtained from single crystal diamond,
graphite, and a typical CVD diamond film. .......................................................59
Figure 2.11. Photoluminescence spectra collected at 300 K and 10 K of a
nitrogen-doped diamond film. ............................................................................60
Figure 2.12. Photograph of the optical characterization system used to collect
Raman and Photoluminescence spectra. ............................................................61
Figure 2.13. Diagram of the hydrogen plasma cleaning chamber. ............................62
Figure 2.14. Schematic of the distance-variable field emission characterization
system. ................................................................................................................63
Figure 2.15. Illustration of the method used to determine the average threshold
field required for field emission. ........................................................................64
Figure 3.1. Raman scattering spectra of nitrogen-doped diamond films with
varying amounts of nitrogen addition. ...............................................................79
Figure 3.2. SEM micrographs of two CVD diamond films prepared with
1.5% methane. (a) without nitrogen admixture, (b) with 60 ppm nitrogen
admixed. .............................................................................................................80
Figure 3.3. SEM micrographs of two diamond films grown by microwave
plasma CVD from an H2/C6H14 plasma. (a) with 100 ppm nitrogen
admixed, (b) with 1 vol.% nitrogen admixed. ...................................................81
Figure 4.1. Schematic of the potential energy diagram for the metal/vacuum
interface with a large applied external field. In addition, FowlerNordheim type tunneling through this potential barrier is illustrated with a
simplified electron wavefunction. ......................................................................97
Figure 4.2.
Schematic of a Spindt cathode. .............................................................98
Figure 4.3. Cross sectional view of a field emission characterization
apparatus with “parallel plate” geometry. ..........................................................99
xi
xii
Figure 4.4. Schematic diagram of the experimental setup for a position
variable anode testing system. ..........................................................................100
Figure 4.5. Schematic diagram for a field emission energy distribution
apparatus. ..........................................................................................................101
Figure 5.1. Energy band diagram illustrating the field emission mechanism
proposed by Geis et al. for nitrogen doped diamond. ......................................130
Figure 5.2. Growth rates from several nitrogen-doped diamond films with
[N]/[C] ratios from 0 to 48. ..............................................................................131
Figure 5.3. Raman scattering spectra for several nitrogen-doped films
deposited on silicon at ~900ºC with [N]/[C] ratios from 0 to 48. ....................132
Figure 5.4. The FWHM of the diamond Raman line of the spectra in Figure 3
as a function of [N]/[C] in the process gas. ......................................................133
Figure 5.5. Photoluminescence spectra recorded at room temperature for
nitrogen-doped diamond films grown on silicon with gas phase [N]/[C]
from 0.2 to 1.0. .................................................................................................134
Figure 5.6. Low temperature PL spectra recorded at 10K for a nitrogen-doped
film grown on molybdenum. ............................................................................135
Figure 5.7. Scanning electron micrographs of the surfaces of several
nitrogen-doped films. .......................................................................................136
Figure 5.8. Threshold voltage as a function of distance for a nitrogen-doped
film in which arcing occurred during computer controlled auto-approach. .....137
Figure 5.9. Current-voltage curve during a field emission measurement in
which an arcing event occurred. .......................................................................138
Figure 5.10. Field emission measurements from a nitrogen-doped films grown
on molybdenum with [N]/[C] =10 in which an arc occurred during the
measurements. ..................................................................................................139
Figure 5.11. Scanning electron micrograph of the arc-damaged region from the
nitrogen-doped film studied in Figure 5.10. .....................................................140
Figure 5.12. Raman scattering spectra for an undoped carbon film with high
sp2 content. .......................................................................................................141
Figure 5.13. Field emission measurements of a carbon film with high sp2
content used to characterize the field emission testing system. .......................142
Figure 5.14. PEEM images of a hydrogen terminated nitrogen-doped diamond
film after 800ºC and 1000ºC anneals. ..............................................................143
xii
xiii
Figure 5.15. A schematic energy band diagram illustrating the defect enhanced
field emission model from compensated nitrogen-doped diamond. ................144
Figure A.1. Theoretical anisotropy for the Raman active mode of the
diamond crystal structure for the (100), (110), and (111) crystal
orientations. ......................................................................................................173
Figure A.2. Schematic of the Raman scattering spectroscopy system used in
this investigation. .............................................................................................174
Figure A.3. The angular dependence of the Raman intensity for a (100)
surface of silicon. .............................................................................................175
Figure A.4. The angular dependence of the Raman intensity for a (110)
surface of silicon. .............................................................................................176
Figure A.5. The measured angular dependence of the Raman intensity for a
(111) surface of silicon. ....................................................................................177
Figure A.6. (a) The polarization sensitive Raman scattering measurements
for a 30 µm thick (100) HOD film. (b) An SEM micrograph taken of the
surface indicates (100) texture of the film. ......................................................178
Figure A.7. (a) The polarization sensitive Raman scattering measurements
for a (100) textured, but not oriented film. (b) An SEM micrograph taken
of the surface indicates (100) texture of the film. ............................................179
Figure A.8.
SEM micrographs of HOD films with varying film thickness. .........180
Figure A.9. Polarization sensitive Raman scattering measurements for the
HOD films shown in Figure A.8. .....................................................................181
Figure A.10. The quality factor, R, as a function of HOD film thickness. .............182
Figure B.1.
Schematic of the cold cathode structures. ..........................................193
Figure B.2. (a) Current-voltage characteristics for four cathode devices. (b)
Grid and collector currents obtained from an AlGaN cathode with
Vg=20V. ...........................................................................................................194
xiii
Chapter 1
Introduction and Scope of Dissertation
6.1
Motivation and Introduction
Electron beams are essential components for many electronic applications
ranging from commonly found television sets to microwave amplifiers and advanced
analytical tools such as electron microscopes. In many applications, electrons are
emitted from a conducting material by heating to the point where the electrons have
sufficient energy to overcome the work function barrier. These emitters are termed
thermionic electron sources. Recently, electron sources utilizing field emission have
been developed. For these sources, electrons are emitted by applying a sufficiently
high electric field such that the electrons tunnel through the work function barrier.
While thermionic electron sources are suitable for many applications, the use of field
emission sources may lead to improved performance in many existing applications
and, more importantly, to new technologies.
Recently, much research and
development has been devoted to the development of vacuum microelectronics for
new applications such as field emission-based flat panel displays and high power
microwave amplifiers.
2
Field emission from metals has been investigated for many years, and it has
been established that low work function metals emit electrons more readily than
metals with a higher work function. Recently, field emission flat panel displays have
been fabricated by several corporations (i.e. Candescent, Micron, Motorola, etc.).
These prototype displays typically utilize pointed metallic structures in order to
concentrate the electric field at the emitter. However, the high fields around the
emitter can also lead to the erosion of the tips (and subsequent degradation of
performance) by the back sputtering of positive ions.
A new approach being considered for field emission sources is to employ
semiconducting materials such as diamond, which exhibit a negative electron affinity
(NEA). For NEA diamond surfaces, electrons in the conduction band can be directly
emitted into vacuum without overcoming an energy barrier. Consequently, it has been
anticipated that high emission currents at modest applied electric fields could be
achieved with diamond surfaces. Although there have been several reports indicating
field emission from diamond at low applied fields, the mechanisms governing the field
emission process from diamond are not completely understood.
The research
described in this dissertation was motivated by reports that indicate that electrons can
be injected into the diamond conduction band for nitrogen-doped diamond and
consequently emitted into vacuum at a very low applied field.
2
3
6.2
History of Diamond
It is believed that diamonds were first discovered by the Dravidians in ancient
India seven to eight centuries before the birth of Christ [1.1-2]. Unfortunately, these
were prehistoric times and not much is known about these extraordinary people. It is
speculated that the earliest diamonds were found in alluvial (riverbed) deposits along
the banks of either the Kistna River or the Godaveri River in southern India [1.1].
Consequently, the nearby city of Golconda became the center of diamond trade. It is
interesting to note that India continued as the sole source of diamonds until the early
eighteenth century when dwindling supplies encouraged the discovery of diamonds
elsewhere.
In these early times, the smallest units of weight adopted in commerce were,
generally, specified seeds. The ancient pearl dealers in the Orient made the striking
discovery that the dried seeds from the pods of the carob or locust-pod tree (ceratonia
siliqua) were remarkably uniform in size and weight. These traders observed that
these carob seeds always weighed within one or two percent of each other no matter
how old the tree, or where it grew, or from which part of the pod the seeds were
extracted [1.1-3]. As a result, the carob seed became the standard unit of weight for
measuring pearls. The carob tree is commonly found throughout the Mediterranean
and the Far East. Consequently, as multi-national trade between the Far East and
Europe developed, carob seeds were also adopted as the standard unit of weight for
diamond and other gemstones. This unit of weight became known as the carat, after
3
4
the Greek name for the carob tree, the keration. Interestingly, the carat is still used
today to express the weight diamond and other gemstones. In 1907, the International
Committee on Weights and Measures in Paris adopted what is now known as the
metric carat, which is defined to be exactly 200 milligrams.
That diamond was an extremely hard substance was readily apparent to these
early discoverers. Indeed, there is evidence that diamond was initially used as a tool
for carving (and later as a weapon!) rather than coveted as a precious gemstone [1.1].
With time, the extreme properties of diamond became attributed to magical and
supernatural powers [1.1-3]. In fact, the Indian word for diamond is vajra, which is
also the same word used to denote the thunderbolt, another form of natural power
[1.1]. As a result, the extreme value of diamond shifted from its use as a tool to a
supernatural icon. It is not surprising then that countless myths, legends, and tales
regarding the origin and magical properties of diamond were propagated to western
civilizations through development of trade routes between Europe and the Far East.
Even as late as the 16th century, the magical powers of diamonds were still being
exploited.
It was thought that diamond powder could cure stomach disorders.
In 1532, Pope Clement VII died when doctors failed to cure his aliments with a
prescription of fourteen spoonfuls of powdered gems (including diamond) [1.2].
Interestingly, the mysticism surrounding diamond continues to some extent even to the
present day.
For example, the famous Hope diamond has been associated with
horribly bad luck for its owner [1.3].
4
5
The cutting and polishing of diamonds into the complex facetted gemstones
with which we are now familiar is a relatively recent practice. For centuries, rough
diamonds were kept only as talismans as a symbol of strength, and often not worn at
all. During this time, stones having natural facets in the shape of an octahedron were
the most prized. A photograph of several rough diamonds exhibiting octahedral and
irregular faceting is shown in Figure 1.1. Although these rough octahedral diamonds
were occasionally set in rings, a Hungarian queen’s crown set with uncut diamonds,
dating from approximately 1074 BC, is perhaps the earliest example of diamond
jewelry [1.3]. The art of polishing and faceting diamond gemstones has often been
accredited to Louis de Berquem of Bruges in 1476 [1.3]. However, there is significant
evidence that diamond polishing techniques were known to the Indians up to 150 years
earlier [1.3]. In 1477, the court counsel of the Archduke Maximilian of Austria
ordered him to prepare for his marriage to Mary of Burgundy with two rings, one set
with a diamond as an engagement ring, the other a gold band, the wedding ring [1.1].
Thus began the tradition of giving diamond engagement rings as a token of enduring
love. This tradition has been partially responsible for the prosperity of the diamond
industry to the present day.
In the early 1720’s Portuguese prospectors looking for gold discovered
diamond in the Minas Geraes province of Brazil. Like those found in India, these
diamonds were collected from alluvial deposits along riverbeds.
The Brazilian
diamond industry hit its peak in the mid nineteenth century producing over a quarter
of a million carats of diamond a year [1.1]. However, the Brazilian monopoly on the
5
6
diamond world industry was shattered upon the discovery of the vast diamond fields in
South Africa in 1867 [1.1-3]. In contrast to the alluvial diamond fields in India and
Brazil, diamonds were found in volcanic “pipes” through which molten rock now
cooled and hardened-rose up from deep within the Earth. This rock, referred to as
Kimberlite, was found incredibly rich in diamond.
From the frenzied rush of
prospectors to the area emerged the present-day industry giant, De Beers Consolidated
Mines Limited [1.1,3]. The world’s largest diamond, the Cullinan diamond, was
found in these South African deposits in 1905. Weighing 3,025 carats, the Cullinan
diamond was about the size of a man’s fist. A photograph of the uncut Cullinan
diamond is shown in Figure 1.2 along with the nine gemstones that were fashioned
from it. Today, South Africa continues to supply a large fraction of the world’s
diamond supply. In addition, diamonds are also mined in a number of countries
including Australia, Botswana, Russia, Canada, and the United States to name a few.
Today, it is estimated that less than 20% of the diamond mined and collected
are suitable for use as gemstones. Most natural stones are delegated for use in a
myriad of industrial applications [1.1-3]. Diamond-studded rotary drill bits and saws
are used to drill oil wells, bore holes through solid rock, or cut asphalt, stone, and
concrete. A majority of the low-grade diamond is crushed into powder for use in
grinding and polishing applications. Very thin wire is manufactured by pulling thick
wire through a series of graduated diamonds with small holes drilled through them. In
addition to these abrasive applications, diamonds have also been developed for x-ray
windows, radiation detectors, heat spreaders, and surgical knives just to name
6
7
a few [1.2]. With the development of techniques for the synthesis of bulk diamond
and diamond thin films, the number of potential applications for the unique materials
properties of diamond have skyrocketed. With these capabilities, it is now possible
that diamond can be engineered for specific applications. Some of these applications
will be discussed below.
6.3
Scientific Research on Diamond
Perhaps the first recorded “scientific” study of the properties of diamond was
performed by the Roman philosopher Gaius Plinius Secundus (23-79 AD), commonly
referred to as Pliny the Elder [1.1-2]. In his vast encyclopedia, Natural History, Pliny
recorded several observations of a material he called adamas, the Greek word for
unconquerable.
Interestingly, from the word adamas we have derived the word
diamond. Pliny divided his adamas stones in to six varieties, one of which we believe
were diamonds.
The others we speculate were not diamond, but rather other
crystalline minerals such as quartz, etc. For the variety that we now call diamond,
Pliny writes that they can be ‘tested upon an anvil, and they are so recalcitrant to
blows that an iron hammer head may be split in two and even the anvil itself be
unseated.’ That this observation was false was certainly known to the ancient Indians
who fashioned diamond splinters into tools and weapons. Nonetheless, the confusion
between the hardness of diamond and inherent “strength” of the crystal exists even
today. Later Pliny writes that the strength of diamond can be mitigated by immersing
7
8
the stones in warm goat’s blood. Pliny also wrote that diamond could neutralize
poison if eaten, and would cure insanity and drive away worry if worn. Indeed,
accounts such as these were largely responsible for perpetuating the mysticism of
diamond for centuries.
However, as scientific methods and understanding improved, the myths and
fantasy surrounding diamond gradually began to disappear. In 1694, G. Averani and
C. A. Targioni conducted experiments on heating diamonds [1.2,4]. Using a wide
aperture lens, they focused intense sunlight upon diamonds and observed that they
vaporized. They were uncertain as to what had happened and did not study the
phenomenon further.
In 1772, Antoine-Laurent Lavoisier carried out a series of
experiments to study the role of oxygen in combustion [1.2,5]. In this study, he
expanded upon the early work of Averani and Targioni by focusing sunlight upon
diamonds placed inside a closed bell jar. In this way, he could examine the products
of the reaction.
He observed that the gas resulting from the destruction of the
diamonds had the same properties of the products formed during the burning of
charcoal. Consequently, Lavoisier speculated that diamond and charcoal were closely
related. These results were expanded by the similar studies of Smithson Tennant,
William Allen and William Hasledine Pepys, and Humphry Davy [1.2,6-8]. By the
early 19th century, it was agreed that diamond was a crystalline form of carbon.
Although it had been determined that diamond was a crystalline form of
carbon, there was much dispute over the crystalline structure of diamond. In nature
diamonds were found which exhibited octahedral, tetrahedral, or more commonly no
8
9
symmetry at all. As a result, debates over the crystalline structure raged throughout
the 19th century. In 1912, W.H. Bragg discovered that the arrangement of atoms in a
crystal could be found by diffracting x-rays from them. This technique was quickly
applied to the problem of the structure of diamond, and it was decisively determined
that diamond exhibited a cubic crystalline structure [1.2,9].
In 1934, Robertson, Fox, and Martin reported that natural diamonds could be
classified into two types according to their optical properties [1.10]. They speculated
that differences in absorption and transmission were produced by differences in
crystalline perfection. Today, we know that the concentration of nitrogen-related
defects in diamond were responsible for the different optical characteristics observed
by Robertson and co-workers. However, their classification has formed the basis for
the scientific classification of single crystal diamonds, which is still used today.
The atmosphere was right for the rapid discovery of techniques for the
synthesis of diamond, and research has continued to the present time. The discovery
of diamonds in volcanic structures in South Africa provided valuable clues for the
possible synthesis of synthetic diamond. After many years of theoretical studies and
unsuccessful attempts by many groups, General Electric announced in 1955 that
diamonds
had
been
synthesized
by
a
high
pressure,
high
temperature
process [1.2-3,11]. Interestingly, successful efforts to synthesize diamond by low
pressure vapor deposition predated this announcement [1.2,12-14].
However,
chemical vapor deposition of diamond did not receive significant attention until the
quality and growth rates were increased by growth techniques introduced in the 1980s.
9
10
6.4
Properties of Diamond
The diamond crystal structure consists of a face centered cubic lattice. The
basis consists of two atoms offset by ¼ of the body diagonal of the cube. A schematic
of the diamond crystal structure is illustrated in Figure 1.3. Each carbon atom is
covalently bonded (sp3) to its four nearest neighbors, which are in a tetrahedral
arrangement around it.
At room temperature, the lattice constant of diamond is
3.567 Å. As a result, diamond is the most atomically dense material known with an
atomic density of 1.76×1023 cm-3.
The strong sp3 bonding character in diamond leads to a number of the extreme
properties of diamond. The most common, of course, is the unparalleled hardness of
diamond. Natural diamonds exhibit a Knoop hardness up to 110 GPa. It is this
property which has made diamond such an important material as an abrasive. In
addition to its extreme hardness, the strong bonding in diamond leads to an extremely
rigid lattice. This rigidity along with the low atomic mass of carbon leads to high
phonon velocities. Consequently, diamond can exhibit a thermal conductivity up to
2500 W/m⋅K, the highest for any material measured to date. This high thermal
conductivity makes diamond an obvious choice for thermal management applications
where diamond can be used as a heat spreader. More recently, the high phonon
velocity of diamond has also been employed in diamond-based surface acoustic wave
(SAW) devices [1.12-13,15].
Finally, the strong bonding in diamond leads to a
material that is chemically very stable. In fact, there is no chemical known that can
10
11
etch or oxidize diamond at room temperature. Therefore, diamond may be an ideal
anti-corrosion and/or anti-wear coating for harsh chemical environments.
In addition to the above properties, diamond is also considered as a wide
bandgap semiconductor with an indirect bandgap of 5.47 eV. Consequently, high
quality diamond is observed to be transparent from the UV (225 nm) to the long
infrared (>25 µm), with the exception of an infrared absorption band from 2 to 7µm.
Consequently, diamond has been developed for optical windows.
In addition to
optical applications, many reports have claimed that diamond may be an exceptional
material for high power, high temperature electronic devices. Diamond can support
fields up to ~107 V/cm before electrical breakdown occurs. In addition, the low
atomic mass of carbon allows very high carrier saturation velocities (~107 cm/s).
These properties combined with electron and hole mobilities of 2200 cm2/V⋅s and
2000 cm2/V⋅s, respectively, indicate the potential for diamond microelectronic
applications. Although p-type diamond can be synthesized using boron as a dopant,
there has been difficulty in obtaining a reliable n-type dopant for diamond. This fact
has severely hindered the development of diamond-based electronic devices. Some of
the properties of diamond along with potential applications can be found in Table 1.1.
6.5
Diamond-Based Cold Cathodes
The potential of diamond as a cold cathode emitter was first recognized by
Himpsel and co-workers in 1979 [1.16]. In these experiments, the quantum efficiency
11
12
for photoemission was measured as a function of incident photon energy. It was
observed that electron emission from hydrogen terminated (111) surfaces occurs for
incident photon energies greater than the diamond bandgap (hν=5.47 eV). In addition,
an extraordinarily high quantum efficiency (~70%) was observed for higher incident
photon energies. These results are illustrated in Figure 1.4. Himpsel and co-workers
suggested that these results indicated that the (111) hydrogen terminated diamond
surface exhibited a negative electron affinity (NEA).
The electron affinity is defined as the energy required to remove an electron
from the conduction band minimum to a distance far from the semiconductor. For
most semiconductors, the vacuum level lies above the conduction band minimum. As
a result, a positive electron affinity surface is observed for most semiconductors. The
presence of a positive affinity creates an energy barrier, which prevents low energy
electrons in the conduction band from escaping into vacuum. For some wide bandgap
semiconductor surfaces, the vacuum level lies below the conduction band minimum
and a NEA surface is observed. In contrast to positive affinity surfaces, no energy
barrier exists at the surface. Consequently, conduction band electrons with kinetic
energies as low as the conduction band minimum have a high probability of escaping
into vacuum. Simplified energy band diagrams for positive and negative affinity
surfaces are illustrated in Figure 1.5.
The presence of a NEA for the hydrogen terminated diamond (111) surface has
been verified by Pate using photoemission spectroscopy [1.17]. Today, it is generally
accepted that the hydrogen terminated (111) surface exhibits a NEA [1.18-19]. More
12
13
recently, the (100), (110), and (311) hydrogen terminated diamond surfaces have also
been shown to exhibit a NEA [1.20-23]. These results demonstrate that, unlike metal
surfaces, these NEA diamond surfaces will not limit the electron emission. As a
result, much research has been directed to the development of diamond-based cold
cathodes.
While the ideal cathode material would exhibit a NEA as noted above, other
properties are equally if not more important. In the most basic sense, field emission
from a semiconductor involves the supply of electrons to the material, transport
through the bulk, and finally emission at the surface. To achieve these properties,
n-type semiconducting characteristics are desired. Highly doped n-type material will
allow low resistance contacts and provide electrons for transport through the material.
An energy band diagram of an ideal NEA cold cathode emitter is shown in Figure 1.6.
To date, efforts have been limited by the lack of a reliable shallow n-type dopant for
diamond.
Consequently, it is evident that the most critical problem remaining to be
solved is supplying conduction electrons to the diamond NEA surfaces. Recently,
Geis et al. have reported that single substitutional nitrogen doping in diamond may
prove helpful for electron emission [1.24-25]. In this study, it was suggested that a
Schottky barrier forms when nickel is used as a backside contact. For high nitrogen
concentrations, a narrow depletion layer forms and electrons can be injected through
this barrier into the conduction band and emitted at the NEA surface. More recently,
Okano et al. have reported threshold fields less than 1 V/µm for nitrogen-doped
13
14
diamond films grown by hot-filament CVD using urea as a nitrogen doping source
[1.26]. They suggest that extraordinary field emission characteristics were attributed
to the high concentration of nitrogen incorporated into the films.
Despite these
studies, the mechanisms governing field emission from nitrogen doped diamond are
still undetermined.
6.6
Scope of Dissertation and Outline
This dissertation represents the culmination of a study of the field emission
properties of nitrogen doped films. In this study, nitrogen doped diamond films were
synthesized by the addition of nitrogen gas to the process gas during microwave
plasma chemical vapor deposition.
Following growth, the properties of these
nitrogen-doped diamond films were characterized by a number of techniques
including Raman and photoluminescence (PL) spectroscopy and optical and scanning
electron microscopy (SEM). The field emission characteristics were examined using a
variable distance anode technique as well as photoemission electron microscopy
(PEEM).
Chapter 2 of this dissertation covers the experimental apparatus and
techniques used to synthesize and characterize the nitrogen-doped diamond films.
Specifically, the variable distance anode technique developed for the characterization
of the field emission properties of diamond films is discussed in Chapter 2.
A
literature review of the growth and properties of nitrogen-doped diamond is presented
in Chapter 3. Chapter 4 highlights relevant research focused upon field emission from
14
15
diamond. The results from the study of the field emission properties of nitrogen doped
diamond are presented in Chapter 5 along with a possible field emission mechanism.
Finally, conclusions and possible areas of future research are presented in Chapter 6.
15
16
References
1.1
J.Y. Dickinson, The Book of Diamonds, (Crown Publishers Inc., New York,
1965).
1.2
G. Davies, Diamond, (Adam Hilger Ltd., Bristol, 1984).
1.3
S. Tolansky, The History and Use of Diamond, (Shenval Press, London, 1962).
1.4
G. Averani and C.A. Targioni, G. Litterati Ital. 8 p.221 (1711).
1.5
A.L. Lavoisier, Memoire Academie des Sciences, p.564 (1772).
1.6
S. Tennant, Phil. Trans. R. Soc. 87 p.97 (1797).
1.7
W. Allen and W.H. Pepys, Phil. Trans. R. Soc. 97 p.267 (1807).
1.8
H. Davy, Phil. Trans. R. Soc. 27 p.39 (1809).
1.9
W.H. Bragg and W.L.Bragg, Proc. R. Soc. A 89 p.277 (1912).
1.10 R. Robertson, J.J. Fox, A.E. Martin, Phil. Trans. R. Soc. A 232 p.463 (1934).
1.11 General Electric Co., press release, February 15, 1955.
1.12 Diamond: Electronic Properties and Applications, eds. L.S. Pan and D.R.
Kania (Kluwer Academic Publishers, Boston, 1995).
1.13 Handbook of Industrial Diamonds and Diamond Films, eds. M.A. Prelas, G.P.
Popovici, and L.K. Bigelow (Marcel Dekker, New York, NY, 1998).
1.14 The Physics of Diamond, eds. A. Paoletti and A. Tucciarone (IOS Press,
Amsterdam, 1997).
1.15 J.T. Glass, B.A. Fox, D.L. Dreifus, and B.R. Stoner, MRS Bulletin 23, p.49
(1998).
1.16 F.J. Himpsel, J.A. Knapp, J.A. van Vechten, and D.E. Eastman, Phys. Rev. B 20
p.624 (1979).
1.17 B.B. Pate, Surf. Sci. 165 p.83 (1986).
16
17
1.18 J. van der Weide and R.J. Nemanich, Appl. Phys. Lett. 62 p.1878 (1993).
1.19 R.J. Nemanich, L. Bergman, K.F. Turner, J. van der Weide, and T.P.
Humphreys, Trieste Semiconductor Symposium on Wide Bandgap
Semiconductors, Physica B 185 p.528 (1993).
1.20 J. van der Weide, Z. Zhang, P.K. Baumann, M.G. Wensell, J. Bernholc, and R.J.
Nemanich, Phys. Rev. B 50 p.5803 (1994).
1.21 J. van der Weide and R.J. Nemanich, J. Vac. Sci. Technol. B 12, 2475 (1994).
1.22 P.K. Baumann and R.J. Nemanich, Diamond Relat. Mater. 4 p.802 (1995).
1.23 P.K. Baumann and R.J. Nemanich, Surf. Sci. 409 p.320 (1998).
1.24 M.W. Geis, J.C. Twichell, N.N. Efremow, K.E. Krohn, and T.M. Lyszczarz,
Appl. Phys. Lett. 68 p.2294 (1996).
1.25 M.W. Geis, J.C. Twichell, and T.M. Lyszczarz, J. Vac. Sci Technol. B 14 p.2060
(1996).
1.26 K. Okano, S. Koizumi, S.R.P. Silva, G.A.J. Amaratunga, Nature 381 p.140
(1996).
17
18
Value at Room
Temperature
Applications
Lattice Constant
3.567 Å
Density
3.515 g/cm3
Knoop Hardness
110 GPa
Polishing and Grinding
Young's Modulus
1050 GPa
Wear-Resistant Coatings
Coefficient of Friction
0.05 - 0.15
Longitudinal Phonon
Velocity
1.8×106 cm/s
Speaker Diaphragms
Thermal Conductivity
2500 W/m·K
Thermal Management
Linear Expansion
Coefficient
1×10-6 K-1
Breakdown Field
1×107 V/cm
Carrier Saturated Velocity
2.7×107 cm/s (electrons)
1.0×107 cm/s (holes)
2200 cm2/V·s (electrons)
2000 cm2/V·s (holes)
Carrier Mobility
Cutting Tools
Heat Sinks, Spreaders, etc.
"Active" and "Passive"
Electronic Devices
For high power,
high frequency, and
high temperature
applications
Intrinsic Resistivity
>1016 Ω·cm
Johnson's Figure of Merit
4.30×1013 V/s
Keyes' Figure of Merit
7.68×109 W/s·K
Bandgap
5.47 eV
Absorption Edge
225 nm
Optical Windows &
Coatings
Dielectric Constant
5.7
Radiation Detectors
Index of Refraction
2.41 for λ=632.8 nm
Electron Affinity
Negative, χ≈-1 eV
Transistors:
Bipolars, FETs, etc.
Gemstones
Vacuum Microelectronic
Devices
Table 1.1. Properties (at room temperature) and applications of diamond.
From [1.12-14].
18
19
(a)
(b)
Figure 1.1. (a) Schematic of the most commonly observed morphologies
for natural diamond single crystals. Note that although natural diamond
crystals stones exhibit octahedral or tetrahedral symmetry, most stones
mined today are irregular with no apparent symmetry. A photograph of
several typical rough diamonds (uncut, unpolished) is shown in (b).
From [1.2].
19
20
(a)
1
2
3
4
(inches)
(b)
Cullinan I
530.2 carats
Cullinan II
317.4 carats
Cullinan III
99.4 carats
Cullinan IV
63.6 carats
Cullinan V Cullinan VI Cullinan VII Cullinan VIII Cullinan IX
18.8 carats 11.5 carats
8.8 carats
6.8 carats
4.4 carats
Figure 1.2. (a) The Cullinan diamond (3,025 carats) as it was discovered in
the Premier mine in South Africa in 1905. The nine major gemstones,
which were cut from the original Cullinan, are shown in (b). These gems
are now part of the British Crown jewels. From [1.1].
20
21
Figure 1.3. Ball and stick model of the diamond crystal structure
illustrating the tetrahedral bonding of the carbon atoms. The diamond
crystal structure consists of a face-centered cubic lattice with a two atom
basis. The basis consists of two carbon atoms offset by 1/4 of the body
diagonal. The lattice constant, a, for diamond is 3.567Å, making diamond
the most atomically dense material known.
21
22
Figure 1.4. The quantum efficiency for photoemission for a hydrogen
terminated (111) diamond surface as a function of incident photon energy.
This figure illustrates that for incident photons of energy of ~16 eV a
quantum efficiency of ~0.7 was observed. Himpsel et al. suggested that the
extraordinarily high quantum yield indicated a negative electron affinity.
From [1.16].
22
23
Evac
χ>0
EC
EV
Most Semiconductors
EC
χ<0
Evac
EV
NEA Semiconductor
Figure 1.5. Band diagrams for semiconductor surfaces, which exhibit a
positive electron affinity (χ>0) and a negative electron affinity (χ<0).
23
24
EC
χ<0 (NEA)
EF
Evac
EV
Metal
n-type Diamond
Vacuum
Figure 1.6. Simplified energy band diagram for an ideal negative electron
affinity cold cathode emitter.
24
25
Chapter 2
Experimental Details
2.1
Growth of Diamond by Microwave Plasma CVD
The growth of diamond by chemical vapor deposition (CVD) has been
investigated since the 1950’s [2.1-4]. However, it was not until the mid-1980’s that
CVD processes were developed that could deposit high quality diamond films at
reasonable deposition rates [2.5-9]. The growth of diamond employing a microwave
discharge is one of these techniques [2.7]. This process termed microwave plasma
CVD (MWCVD) is the method used to synthesize diamond films in this study.
With the development of these deposition techniques, the theory of diamond
CVD has been extensively investigated [2.10-14].
In the following section, a
simplistic overview of the theory of CVD growth of diamond is presented.
In
particular, this discussion is limited to the growth of diamond by the MWCVD
technique using hydrogen and methane and the growth precursors. Furthermore, this
discussion focuses upon the growth of diamond on diamond surfaces. More detailed
accounts of diamond growth and nucleation of diamond on non-diamond surfaces
25
26
have been published elsewhere [2.13-14]. Finally, in Section 2.1.2, a description of
the microwave plasma CVD system used in this study is presented.
2.1.1 Theory of Diamond CVD
In typical CVD processes where the pressures are low, gas phase reactions are
often negligible or are limited to the decomposition of a few precursors (e.g. use of
organometalic compounds, etc.).
However, in diamond CVD, there is general
agreement that free radicals play a crucial role in the deposition of diamond through
gas phase and surface reactions. In particular, as will be discussed below, atomic
hydrogen has been identified as an essential component in the deposition of diamond
[2.15-17]. In addition, atomic hydrogen is also responsible for the preferential etching
of sp2 bonded carbon, which is also deposited during diamond growth.
In microwave plasma CVD reactors, atomic hydrogen is produced within the
plasma discharge. Free electrons in the plasma are accelerated by the electromagnetic
fields inside the microwave cavity. Through this coupling mechanism between the
electrons and the applied external radiation, electrons can reach energies of several
electron volts (1 to 4 eV on average).
Consequently, atomic hydrogen can be
produced through the following reaction:
H 2 + e− → H + H + e− .
(1)
The steady state concentration of atomic hydrogen present in the CVD reactor is
determined by the balance of atomic hydrogen production and recombination. For
26
27
typical conditions used for the deposition of diamond, recombination of atomic
hydrogen is a slow process. Therefore, atomic hydrogen alone is likely to diffuse to
the substrate or chamber walls before recombining with other atomic hydrogen
species. So far, we have not mentioned the influence of small concentrations of
hydrocarbon species (such as methane) in the plasma. The presence of methane in the
plasma provides other possible recombination pathways such as:
H + CH 4 → CH 3 + H 2
(2)
H + CH 3 → CH 4 .
(3)
or similarly
In addition to providing recombination mechanisms for atomic hydrogen, the presence
of hydrocarbon species in a hydrogen plasma presents a complex chemical
environment. Hydrocarbon species react with each other and with hydrogen to form
various other species in the plasma. The equilibrium concentration of hydrocarbon
species is determined by the multitude of possible reaction pathways. There has been
much debate and research over the particular growth species responsible for diamond
growth. One of the most accepted models is that methyl radicals are the predominant
growth precursor for the formation of diamond [2.18-21].
One of the most important aspects of diamond surface chemistry is the
interaction of atomic hydrogen with the growing diamond surface. Typical CVD
growth temperatures exceed 800°C. Under ultra-high vacuum conditions, hydrogen
desorbs from diamond (111) and (110) surfaces at temperatures as low as 800°C and
27
28
stable surface reconstructions are observed [2.22-23].
However, during diamond
CVD, the constant flux of atomic hydrogen from the plasma is thought to stabilize the
growing surface by preventing reconstruction.
In contrast to hydrogenating the
diamond surface, atomic hydrogen can also abstract hydrogen from the diamond
surface through the following reaction:
C d H + H → C *d + H 2
(4)
where CdH represents a hydrogen terminated surface site and Cd* is an equivalent site
without hydrogen termination. This abstraction process produces a chemically active
site on the diamond surface, which can react with another hydrogen atom or growth
species. Upon reaction with an active site, a hydrocarbon can either desorb from the
surface, or become incorporated into the diamond surface. Incorporation (and thus
diamond film growth) occurs when the growth species loses all its hydrogen atoms
through hydrogen abstraction. A schematic of the theory of CVD diamond growth is
illustrated in Figure 2.1.
2.1.2 The Microwave Plasma CVD Diamond Chamber
In this study, a modified high-pressure microwave source (HPMS)
manufactured by Applied Science and Technology (ASTeX) was used to deposit
diamond films. This chamber was modified from its initial design in order to become
part of the integrated UHV surface characterization and processing facilities located in
28
29
the Surface Science Laboratory. A schematic of this diamond deposition system is
illustrated in Figure 2.2.
Microwave radiation is supplied to the CVD reactor by an ASTeX S1500i
microwave power generator. This system consists of a high-performance 2.45 GHz
(λ=12.2 cm) magnetron power head capable of producing up to 1500 W of output
power.
This microwave radiation is delivered to the diamond chamber using
rectangular aluminum (WR284) waveguides. To protect the magnetron from reflected
power, a three-port circulator equipped with a water-cooled dummy load is included
on the system. This circulator has the property that any power reflected from the
diamond plasma chamber is diverted to the dummy load. The water-cooled dummy
load is capable of absorbing 500 W of microwave power on a continuous basis or up
to 1500 W intermittently. To minimize the reflected power from the plasma chamber,
a three-stub tuner is included on the waveguide assembly. This device matches the
impedance of the plasma system using impedance transforming tuning slugs that can
be independently positioned.
These waveguide components deliver the microwave radiation to a symmetric
plasma coupler (SPC), which connects to the quartz window of the double walled
water-cooled stainless steel vacuum process chamber. The SPC (patented by ASTeX)
employs an antenna to convert from rectangular to cylindrical symmetry.
The
combination of the SPC, cylindrical chamber, and the sample stage form a microwave
cavity. The sample stage forms the bottom of the waveguide and provides additional
tuning of the microwave cavity. When properly tuned, a TM01 mode is supported by
29
30
the cavity and a plasma ball forms above the sample surface in the regions of high
electric field. Two sample stage configurations can be used depending upon whether
in vacuo sample transfer to the integrated surface characterization and processing
system is desired.
The original HPMS sample stage can accommodate substrates up to 4” in
diameter. A graphite susceptor inductively heated by a RF coil, provides uniform
heating up to ~900°C. Although this heater incorporates a thermocouple, a one color
optical pyrometer is used to accurately determine surface temperature during growth.
Since 1” substrates were usually used, a molybdenum plate was used to cap the
graphite susceptor and position the substrate within the reactor. This plate prevents
the unwanted etching of the graphite susceptor and subsequent increase of
hydrocarbons in the plasma. In this configuration, the substrates are manually placed
upon the sample stage. As a result, this stage is incompatible with the sample holders
used in the integrated surface characterization and processing system. However, an
advantage to this configuration is that a bias can be applied to the sample stage for
biased enhanced nucleation (BEN). A schematic of the graphite susceptor sample
stage is shown in Figure 2.3.
For some experiments, it is desirable to deposit a diamond film and transfer the
sample in vacuo without exposing the surface to ambient conditions. For this reason,
a new sample stage was developed which was compatible with the molybdenum
sample holders used in the integrated surface characterization and processing system.
Unfortunately, any modification to the microwave cavity usually results in the loss of
30
31
plasma stability. Figure 2.4 illustrates the modified sample stage, which is compatible
with the molybdenum sample holders. The original graphite susceptor is coaxially
bored so that the sample holder can be recessed even with the top of the sample stage.
The sample holder can be moved up and down by a linear vacuum feedthrough. This
movement allows the sample holder to be positioned at various distances from the
plasma. Furthermore, this movement is necessary to raise the sample holder for
transfer out of the reactor.
In addition to some loss of plasma stability, this
configuration has other disadvantages. This configuration does not employ a substrate
heater. As a result, sample temperature is determined by direct interaction with the
plasma discharge and must be measured using an optical pyrometer. This limits the
parameter space available for diamond deposition. Additionally, biased enhanced
nucleation has not been achieved with this sample stage.
As mentioned previously, the CVD reactor is a water-cooled stainless steel
chamber. In addition, conventional UHV vacuum components, such as conflat flanges
and vacuum feedthroughs, are employed. Vacuum is established in the reactor using a
Leybold D16B rotary vane mechanical pump, and the base pressure of the system
is ~10-4 Torr. Convectron (Granville-Phillips) and baratron (MKS) gauges are used to
monitor the pressure in the system. During growth, the process pressure is maintained
using a feedback loop between the baratron gauge and a MKS throttle valve. Process
gases (e.g. hydrogen, methane, etc.) are admitted at the top of the chamber near the
quartz window. Precise control of the gas flow is achieved using MKS mass flow
controllers. The process gases are mixed in a gas manifold prior to entering the
31
32
chamber. Nitrogen doped diamond films were synthesized by the addition of nitrogen
to the process gas mixture.
For small nitrogen addition, this was achieved by
admitting a mixture of nitrogen diluted in hydrogen. For larger nitrogen process
concentrations, pure nitrogen was admitted directly to the gas manifold. More details
regarding the nucleation and growth of nitrogen doped films are discussed in
Chapter 5.
In addition to the vacuum deposition chamber, an intermediate chamber is used
as a pressure lock between the mechanically pumped deposition chamber and the
UHV transfer line. This intermediate chamber is pumped by a turbomolecular pump
and has a base pressure of 10-9 Torr. After growth, the samples are held in this
intermediate chamber until the pressure is acceptable for transfer (~10-8 Torr).
2.1.3 Laser Reflectance Interferometry
Laser reflectance interferometry (LRI) was an additional capability added to
the ASTeX HPMS diamond CVD reactor. This technique provides in situ growth rate
information by monitoring the reflectance of the growing diamond film for thin
diamond films grown on polished non-diamond substrates [2.25-26].
The LRI
apparatus used to monitor the growth in this study consisted of a HeNe laser
(λ=632.8 nm) directed at normal incidence to the substrate. A silicon photodiode was
used to monitor the reflected beam.
32
33
The reflectance of the diamond surface can be modeled by considering the
reflectivity of a three-layer dielectric stack (i.e. vacuum, diamond, and silicon) at
normal incidence, as shown in Figure 2.5. The electromagnetic boundary conditions
require that the tangential component of both E and H are continuous at both
interfaces. Assuming a plane wave solution to Maxwell’s equations of the form:
E(x, t ) = E 0 ei (k ⋅x −ωt )
(5)
where E and H are related by:
H=
c
k ×E.
ω
(6)
The continuity of the tangential component of E at z=0 requires that:
E i − E r = E t 1 − Eb .
(7)
Similarly, for the tangential component of H at z=0, we have:
n1 Ei + n1 Er = n2 Et1 + n2 Eb .
(8)
Analogously, for the interface at z=d these boundary conditions require the following
for the tangential components of E and H, respectively:
Et1e ik2d − Eb e − ik2d = Et 2 e ik3d
(9)
n2 Et1e ik2d + n2 Eb e − ik2d = n3 Et 2 e ik3d .
(10)
The relative amplitude of the reflected wave can be explicitly determined from the
system of four equations above. Thus, we have:
E r r12 + r23 e 2ik2 d
=
Ei 1 + r12 r23 e 2ik2 d
(11)
33
34
with
r12 =
n −n
n2 − n1
and r23 = 3 2 .
n2 + n1
n3 + n2
(12)
The reflectivity of the three-layer structure is then given by:
E
R= r
Ei
2
(13)
The reflectivity at normal incidence of the vacuum/diamond/silicon stack is plotted as
a function of diamond film thickness in Figure 2.6. It can be shown that the period
between successive maxima or “fringe” corresponds to a thickness change of λ/2n2.
For illumination with a HeNe laser with (λ=632.8 nm), each period corresponds to a
growth of 0.13 µm for diamond. The value is valid assuming the index of refraction
of the growing film is the same as that reported for bulk diamond. As a result, LRI
allows the determination of the total thickness deposited. By calculating the amount
of time required for each interference fringe, the growth rate can be calculated.
In practice, the LRI behavior deviates slightly from the above treatment. In
Figure 2.7, the reflected intensity taken during a typical growth run is shown. In
particular, sharp cusps are observed at the minima of each cycle. The exact reason for
this behavior is not known, although it is speculated that scattering from the diamond
grains or from the rough silicon surface may play a role. In addition to the cusps, it is
observed that the overall reflectivity decreases for long deposition runs.
As the
diamond film deposition progresses, the diamond grains in the film grow larger. This
has the effect of increasing the surface roughness, which increases light scattering
34
35
from the surface and reduces the reflectivity. Despite these deviations from the
electromagnetic theory presented above, the periodicity between interference fringes is
observed to remain the same. As a result, film thickness and growth rate information
can still be accurately obtained.
2.2
Characterization of Diamond Films
After diamond film deposition, the samples were typically removed from the
chamber and characterized by several techniques. To evaluate diamond film quality,
Raman scattering spectroscopy was used. The presence of optically active defect
centers was verified by photoluminescence measurements. In addition, optical and
secondary electron microscopy was used to characterize the film morphology and to
identify surface features.
Following this analysis, the samples were then mounted on molybdenum
sample holders and introduced into the integrated UHV surface characterization and
processing system for field emission analysis. A schematic of this system is shown in
Figure 2.8. In order to remove contamination from the diamond film surface, a brief
radio frequency (RF) hydrogen plasma exposure was usually employed prior to field
emission characterization.
After field emission testing, the diamond films were
occasionally examined using x-ray photoemission spectroscopy (XPS), Auger electron
spectroscopy (AES), and ultraviolet photoemission spectroscopy (UPS). However,
these techniques were of limited use due to the insulating nature of the films examined
35
36
in this study. As a result, a detailed account of the experimental apparatus for these
techniques is not provided here.
For excellent reviews on these surface science
characterization techniques consult Ref #2.27.
2.2.1 Raman Scattering and Photoluminescence Spectroscopy
In elementary terms, the Raman effect in crystals can be described as the
interaction of light with the crystal vibrational modes. More precisely, the Raman
effect is the inelastic scattering of photons by lattice vibrations (or phonons). Incident
light of energy Eo can interact with the crystal to create or annihilate one or more
phonons [2.28-29]. In the case where a lattice vibration is created by this scattering
event, the incident photon must contribute energy to sustain the phonon.
Consequently, the scattered photon energy is reduced by Evib or the energy of the
phonon created by the interaction. The process where a phonon is created is known as
a Stokes interaction. On the other hand, if a phonon is destroyed in the Raman event,
the energy of the scattered photon increases by the energy of the phonon, Evib. These
interactions are termed anti-Stokes processes. The Stokes interaction is the most
probable scattering event, since the anti-Stokes interaction depends upon the
availability of the crystal vibrations. As a result, Raman analysis of the diamond films
grown in this study has focused upon the Stokes component. Possible Raman
interactions between photons and phonons along with the resulting spectra are shown
in Figure 2.9.
36
37
It should be mentioned that only lattice vibrations having certain types of
symmetry give rise to Raman scattering. These phonons are said to be Raman active.
Furthermore, this crystal symmetry also produces selection rules whereby Raman
active modes may be forbidden for some crystallographic orientations [2.28-29]. In
general for polycrystalline diamond films, these selection rules are relaxed due to the
random orientation of the grains.
However, these selection rules were used to
characterize the crystallinity of a new class of diamond films, which consist of highly
oriented (100) grains. More details of this Raman analysis of highly oriented diamond
(HOD) films can be found in Appendix A.
As discussed above, Raman scattering provides information about the
vibrational properties of the crystal. Single crystal diamond has a triply degenerate
Raman active mode, which appears in the Raman spectra at 1332.5 cm-1 [2.30]. Other
forms of carbon such as graphite have different Raman active modes owing to their
different crystal structure and symmetry. Further analysis of the peak corresponding
to the Raman active mode can yield additional information about the diamond film.
The width of the diamond characteristic peak is inversely related to the phonon
lifetime. The phonon lifetime is reduced by a number of effects such as reduced grain
size, defects, etc. [2.31-32]. Consequently, the full width at half maximum (FWHM)
of the diamond Raman line can be used to quantify the crystalline quality of a
diamond film. The presence of strain in the diamond film can also be observed using
Raman scattering spectroscopy [2.31-32]. Shifts in the diamond Raman line to higher
frequencies indicate compressive strain in the film, while shifting to lower frequencies
37
38
is indicative of tensile strain. As a result, Raman scattering spectroscopy has become
an established technique for evaluating the bonding in diamond CVD films. The
Raman scattering spectra from single crystal diamond, graphite, and a typical CVD
film are shown in Figure 2.10.
In addition to the Raman effect, photons also interact with the electronic states
in a solid. An incident photon can promote an electron into an excited state through
the process of absorption.
If the subsequent transition back to the ground state
involves the emission of a photon, this light is referred to as photoluminescence (PL).
The transition back to the ground state is referred to as the “zero-phonon transition” if
there is no phonon interaction involved. The corresponding feature in the PL spectra
is known as the zero-phonon line (ZPL). However, for many optical centers, there is a
significant interaction between electronic and vibrational transitions. As a result, an
electron in an excited state may recombine through the simultaneous process of giving
energy to the lattice and the emission of a photon. The energy of the emitted photon
in this case is less than the energy of a photon emitted in a zero-phonon transition.
Consequently, side bands to the ZPL are formed in the PL spectra. This electronlattice interaction is strongly temperature dependent. At higher temperatures, there are
more phonons present in the crystal and thus a higher probability of electron-lattice
transitions. This results in an increase in the sideband intensity at the expense of
intensity of the ZPL. Conversely, at low temperatures, vibrational participation is
reduced producing a sharper more intense ZPL in the PL spectra. For this reason,
many PL experiments are conducted at low temperatures for precise identification of
38
39
optical centers.
This temperature effect upon the photoluminescence spectra is
illustrated in Figure 2.11. It should also be mentioned that in addition to lattice
interactions, there are other processes, which influence photoluminescence.
In
particular, a high concentration of defects reduces luminescence by providing nonradiative recombination paths.
Many
optical
centers
have
been
photoluminescence spectroscopy [2.14].
identified
in
diamond
using
In general, these optical centers are
associated with particular defect structures, which provide electronic states within the
bandgap.
Each optical center has its own electronic levels and thus can be
distinguished from other centers by its luminescence “finger-print.” Although PL
spectroscopy allows for the precise determination of the energy between the excited
and ground state of an optical center, this technique cannot, in general, determine the
location of the defect level within the bandgap.
In this study, photoluminescence spectroscopy was used to identify the
presence of single substitutional nitrogen in diamond CVD films.
substitutional
nitrogen
cannot
be
directly
observed
in
Single
photoluminescence
measurements. However, the combination of substitutional nitrogen with one or more
vacancies gives rise to optical centers that can be observed in PL spectra at 1.945 eV
and 2.154 eV [2.33-37]. The optical center with a ZPL at 1.945 eV has been identified
with single substitutional nitrogen and a vacancy, in nearest neighbor positions,
arranged in a trigonal symmetry [2.33-37]. The second center with a ZPL at 2.154 eV
has been associated with a substitutional nitrogen atom with an even (at least two)
39
40
number of vacancies [2.35-36]. These nitrogen related optical centers have been
observed in single crystal diamond where nitrogen is primarily incorporated in isolated
substitutional sites. This type of single crystal diamond is classified as Type Ib
diamond (see Table 3.1).
The experimental setup used for Raman and photoluminescence spectroscopy
is shown in Figure 2.12. The system consists of an ISA U1000 scanning double
monochromator equipped with a photomultiplier tube (PMT), which was specifically
designed for high resolution Raman scattering applications. The focal length of the
monochromator is 1 meter, and the spectral resolution is 0.15 cm-1. An argon ion laser
was used as the excitation source. Typically, the 514.5 nm line was used for both
Raman and PL analysis.
This optical system allows spectra to be acquired in two modes depending
upon the desired spot size. In the “micro” configuration, the laser beam is focused to a
spot size of ~5 µm in diameter. This is facilitated using an Olympus BH2 microscope.
In the “macro” configuration, a spot size of approximately 2 mm × 100 µm is achieved
using a cylindrical lens. The macro configuration has the additional capability that
samples can be mounted to a Janis CCS-350 closed loop refrigeration system. Using
this cold stage, spectra can be obtained at temperatures from ~10 K to 300 K.
40
41
2.2.2 Microscopy
Following optical characterization described above, the nitrogen-doped films
were also examined using conventional optical microscopy as well as scanning
electron microscopy (SEM).
Using an Olympus BX60 optical microscope, the
diamond surfaces were examined for scratches or other large defect structures which
could influence the field emission characteristics by enhancing the field at sharp defect
sites. The advantages of optical microscopy are that it is an inexpensive and quick
method to characterize the surface. However, magnification and depth of focus are
limited by the wavelength of the light. In order overcome these limitations SEM was
used to characterize the morphology and identify small defects on the diamond
surface. For this characterization, a JEOL 6400 SEM was used. Following field
emission measurements, these microscopy techniques were again employed for the
identification of any changes to the diamond surface caused by micro-arcs, etc.
2.2.3 Field Emission Characterization
In order to remove contamination from ambient exposure and to terminate the
diamond surface with hydrogen, the nitrogen-doped diamond films were typically
exposed to a remote hydrogen plasma before field emission testing. A schematic of
this hydrogen plasma system is illustrated in Figure 2.13. The plasma system is
evacuated using a turbomolecular pump and has a base pressure of ~1×10-8 Torr.
During plasma exposure, hydrogen is introduced into the chamber through a 25 mm
41
42
diameter quartz tube. For these studies, the hydrogen flow rate was held at 79 sccm
using a mass flow controller. This flow rate results in a pressure of ~25 mTorr in the
plasma chamber. A 12-turn copper coil around the quartz tube is used to inductively
couple RF power to the plasma. Using an impedance matching network, 50 W of RF
power was used in these plasma exposures. The sample surface faces the plasma and
is located ~40 cm from the end of the quartz tube. During plasma exposure, the
sample surface is heated to ~500°C by the use of a tungsten filament located in close
proximity to the back of the sample. Following hydrogen plasma treatment, the
samples were transferred to the field emission chamber via the UHV transfer line.
A schematic of the apparatus used to characterize the field emission properties
of diamond films is illustrated in Figure 2.14. The sample holder (and sample) is
mounted to the field emission testing stage by the use of the bayonet pins on the
holder. In order to minimize vibration, the holder is held firmly in place by the use of
spring clips, which push firmly against the sample surface. Furthermore, the entire
testing stage is manufactured from a single piece of stainless steel. As a result the
effects of vibrations are minimized since the system and sample move in concert
rather than separately. The chamber is pumped by a turbomolecular pump and an ion
pump.
With this pumping configuration, a base pressure of ~2×10-9 Torr was
achieved.
The anode is attached to a linear translation stage, which is controlled by a
UHV stepper motor.
Using reduction gearing, one step of the stepper motor
corresponds to an average translation of the anode stage by 55 nm. Consequently, the
42
43
field emission properties of the diamond film can be obtained at multiple anode-tocathode distances. A disadvantage to this translation stage is that the absolute distance
between the sample surface and the anode is not known with the required precision.
However, the relative distance between measurements is known to a high degree of
accuracy. As a result, an analysis technique was developed, which is based upon the
relative distance between measurements. This technique is described in more detail
below.
A cylinder of molybdenum was chosen for the anode for this apparatus. To
eliminate unwanted effects from the edges of the cylinder, the end of the anode was
rounded to a high radius of curvature. Prior to installation into the field emission
testing chamber the end of the anode was polished with diamond grit (down to 1 µm).
Following this polishing step, the anode was ultrasonically cleaned to ensure the
removal of residual polishing material.
Anodes with diameters ranging from
1 to 3 mm could be used in the system depending upon the desired area of
characterization.
The required electric field for field emission is produced by applying a voltage
between the sample and the anode, which is located in close proximity to the sample
surface. Applied voltage ranging from 0 to 1100 V was provided by a computer
controlled Keithley 237 source measure unit (SMU). The SMU has the ability to
simultaneously source a voltage and measure the current in the circuit. An additional
feature of the SMU is the capability of limiting the current below what is known as the
compliance value. This circuit prevents damage to the SMU as well as the rest of the
43
44
system in the event of a short circuit, which may occur during high voltage
breakdown.
In these measurements, a value of 1×10-8 A was chosen as the
compliance value. In addition, triaxial cables between the SMU and field emission
chamber were used to reduce the electrical noise in the system. With this cabling, the
electrical noise was limited to ±1×10-11 A.
Although the anodes used in this study were rounded, it is assumed that the
electric field between the anode and the sample surface can be modeled by the parallel
plate geometry. This assumption is valid since the radius of curvature is very large in
comparison to the distance, d, between the anode and the sample (cathode). In the
parallel plate geometry the electric field is given by:
E=
V
d
(14)
where V is the applied voltage between the anode and the cathode. As a result, if a
particular electric field, Ethreshold, is required for field emission from a cathode surface,
then the required applied voltage is given by:
Vthreshold = E threshold d .
(15)
From this expression, it is evident that the voltage required for emission is linearly
dependent upon d, the distance between the anode and the sample (cathode). If the
anode to sample distance is increased, there is a corresponding increase in the voltage
necessary to induce field emission. On the other hand, if this distance is decreased,
there is an analogous reduction in the voltage required for field emission. This linear
44
45
relationship allows the determination of the required threshold field without knowing
the absolute anode to cathode separation.
For any given field emission measurement, a family of current-voltage (I-V)
curves is recorded with each curve corresponding to a different anode-to-sample
spacing. The field emission measurements begin with the anode positioned at an
unknown distance away from the sample surface. The stepper motor count is recorded
and an I-V sweep is initiated. This I-V sweep consists of gradually ramping up the
voltage in 5 volt increments until the measured current reaches compliance.
At
compliance, the sweep is reversed and the applied voltage is reduced to zero.
Following this I-V measurement, the anode is moved closer to the sample by a known
distance, and the cycle is repeated until at least 5-10 curves are collected. As expected
for parallel plate geometry, the I-V curves shift to lower voltage values with
decreasing anode to cathode distance.
Due to the exponential nature of the field emission I-V curves, it has become
practice to define the “turn-on” voltage or threshold voltage in terms of a specific
current value. In this study, the voltage that results in a current value of 0.5 nA was
chosen to represent the threshold voltage for electron emission.
Each threshold
voltage is then plotted versus distance relative to the first I-V curve and, as expected,
the resulting graph was linear. Upon fitting the data to a straight line, the slope
represents the average field for the threshold current emission. This method for
determining the average threshold field does not rely upon the absolute anode to
sample spacing, but rather an accurate measurement of the change in distance of the
45
46
anode with respect to the sample. In addition, this technique has the advantage that
the anode is never in contact with the sample.
This technique used for the
determination of the average threshold field is illustrated in Figure 2.15.
In general, the electric field between the anode and the sample surface can only
be precisely modeled by the parallel-plate geometry discussed above for a metal
sample. For metals, there is no electric field penetration into the cathode (sample)
surface and we obtain the familiar relation:
E=
V
d
(16)
However, for semiconductors there will be, in general, some field penetration into the
sample which will influence the value of the electric field between the sample and
cathode. The extent of the field penetration for semiconductors depends not only upon
the dielectric constant of the material, but also upon the dopant concentration. For
heavily doped semiconductors, the field penetration is reduced due to the presence of
large concentration of free carriers. For the nitrogen-doped diamond films grown in
this study, it is not completely evident how much field penetration will occur.
Typically, the thickness of the nitrogen-doped diamond film is typically much smaller
than the anode to cathode distance. Consequently, it is assumed that field penetration
minimally influences the field emission characterization of these films. The role of
field penetration and its influence upon the variable distance anode technique
employed in this investigation is a topic of ongoing research.
46
47
References
2.1
H. Meincke, Schweiz. Arch. Angew. Wiss. Tech. 23 p.85 (1957).
2.2
W.G. Eversole. “Synthesis of Diamond,” US Patent No. 3,030,187 (1962).
2.3
J.C. Angus, H.A. Will, W.S. Stanko, J. Appl. Phys. 39 p.2915 (1968).
2.4
B.V. Deryagin and D.V. Fedoseev, Russ. Chem. Rev. 39 p.783 (1970).
2.5
R. Mania, L. Stobierski, and R. Pampuch, Cryst. Res. Technol. 16 p.785 (1981).
2.6
S. Matsumoto, S.Y. Sato, M. Kamo, and N. Setaka, Jpn. J. Appl. Phys. 21
p.L183 (1982).
2.7
M. Kamo, Y. Sato, S. Matsumoto, and N. Setaka, J. of Crystal Growth 62 p.642
(1983).
2.8
K. Kurihara, K. Sasaki, M. Kawarada, and N, Koshino, Appl. Phys. Lett. 52
p.437 (1988).
2.9
Y. Hirose, S. Amanuma, N. Okada, and K. Komaki, First International
Symposium on Diamond and Diamond-Like Films, eds. A.P.J. Dismukes
Meyerson, Moustakas, K. Spear, K. Ravi, and M. Yoder, (The Electrochemical
Society, Pennington, NJ, 1989) p.80.
2.10 W.A. Yarbrough and R. Messier, Science 247 p.688 (1990).
2.11 F.G. Celii and J.E. Butler, Ann. Rev. Phys. Chem. 42 p.643 (1991).
2.12 J.C. Angus, Thin Solid Films 216 p.126 (1992).
2.13 The Physics of Diamond, eds. A. Paoletti and A. Tucciarone (IOS Press,
Amsterdam, 1997).
2.14 Handbook of Industrial Diamonds and Diamond Films, eds. M.A. Prelas, G.P.
Popovici, and L.K. Bigelow (Marcel Dekker, New York, NY, 1998).
2.15 M. Frenklach, J. Appl. Phys. 65 p.5142 (1989).
2.16 W.L. Hsu, J. Vac. Sci. 6 p.1803 (1988).
47
48
2.17 W. Banholzer, Surf. Coat. Tech. 53 p.1 (1992).
2.18 S.J. Harris, A.M. Weiner, and T.A. Perry, Appl. Phys. Lett. 53 p.1605 (1988).
2.19 C.J. Chu, M.P. D’Evelyn, R.H. Hauge, and J.L. Margrave, J. Mater. Res. 5
p.2405 (1990).
2.20 C.J. Chu, M.P. D’Evelyn, R.H. Hauge, and J.L. Margrave, J. Appl. Phys. 70
p.1695 (1991).
2.21 M.P. D’Evelyn, C.J. Chu, R.H. Hauge, and J.L. Margrave, J. Appl. Phys. 71
p.1528 (1991).
2.22 J. van der Weide, Z.Zhang, P.K. Baumann, M.G. Wensell, J. Bernholc, and R.J.
Nemanich, Phys. Rev. B 50 p.5803 (1994).
2.23 P.K. Baumann and R.J. Nemanich, Surf. Sci. 408 p.320 (1998).
2.24 J.E. Butler and R.L. Woodin, Philos. Trans. R. Soc. London 342 p.209 (1993).
2.25 B.R. Stoner, B.E. Williams, S.D. Wolter, K. Nishimura, and J.T. Glass, J.
Mater. Res. 7 p.257 (1992).
2.26 C.H. Wu, W.H. Weber, T.J. Potter, and M.A. Tamor, J. Appl. Phys. 73, p.2977
(1993).
2.27 D.P. Woodruff and T.A. Delchar, Modern Techniques of Surface Science
(Cambridge University Press, Cambridge, 1994).
2.28 R. Loudon, Adv. Phys. 13 p.423 (1964).
2.29 W. Hayes and R. Loudon, Scattering of Light by Crystals (Wiley, New York,
1978).
2.30 S.A. Solin and A.K. Ramdas, Phys. Rev. B 1 p.1687 (1970).
2.31 R.J. Nemanich, J.T. Glass, G. Lucovsky, and R.E. Schroder, J. Vac. Sci.
Technol. A 6 p.1783 (1988).
2.32 L. Bergman, B.R. Stoner, K.F. Turner, J.T. Glass, and R.J. Nemanich, J. Appl.
Phys. 73, p.3951 (1993).
2.33 G. Davies, Chemistry and Physics of Carbon 13 p.1 (1977).
2.34 G. Davies and M.F. Hamer, Proc. R. Soc. Lond. A 348, p.285 (1976).
48
49
2.35 G. Davies, J. Phys. C: Solid St. Phys. 12 p.2551 (1979).
2.36 A.T. Collins and S.C. Lawson, J. Phys. C: Condens. Matter 1 p.6929 (1989).
2.37 L. Bergmann, M.T. McClure, J.T. Glass, and R.J. Nemanich, J. Appl. Phys. 76
p.3020 (1994).
49
50
Figure 2.1. Schematic of the different processes which are suspected to
lead to the deposition of diamond by chemical vapor deposition.
From [2.24].
50
51
3 Stub
Tuner
Microwave
Power Head
Isolation
Gate Valves
Symmetric
Plasma Coupler
Intermediate
Chamber
UHV Transfer
Line
HPMS
CVD Reactor
Turbo Pump
Gas Manifold
Power Supplies &
Control Electronics
Induction Heater
Matching Network
Mechanical
Pumps
Figure 2.2. Schematic of the microwave plasma CVD chamber used to
deposit diamond thin films in this study.
51
52
Microwaves In
Quartz
Microwave
Window
Double-Walled
Stainless Steel
Chamber
Gas Inlet
Cooling
Water In
Graphite
Susceptor
Plasma
Ball
Thermocouple
RF Heater
Coil
Pumping
Port
Cooling
Water Out
Alumina
Heater
Support
Stainless Steel
Heater Shell
Figure 2.3. Drawing of the ASTeX microwave plasma CVD chamber with
the inductively heated sample stage. This sample heating stage is capable
of heating the samples up to ~900°C. Higher surface temperatures are
possible as a result of plasma heating.
52
53
Microwaves In
Quartz
Microwave
Window
Gas Inlet
Double-Walled
Stainless Steel
Chamber
Cooling
Water In
Sample
Holder
Plasma
Ball
Modified
Graphite
Susceptor
Pumping
Port
Alumina
Heater
Support
Stainless Steel
Heater Shell
Cooling
Water Out
Sample Holder Stage is
Connected to a Linear Motion
Vacuum Feedthrough
Figure 2.4. Drawing of the ASTeX microwave plasma CVD chamber with
the modified sample stage. This configuration is compatible with the
molybdenum sample holders. The substrate temperature is regulated by the
microwave power and the distance between the sample and plasma ball.
53
54
Ei
Hr
Et1
Er
Hi
Hb
Et2
Eb
Ht1
Ht2
d
Vacuum
n1=1.0
Diamond
n2=2.4
z=0
Silicon
n3=3.882
+0.019i
z=d
Figure 2.5. A model for the reflectivity from the growing diamond surface.
In this case, the incident laser beam is normal to the diamond surface.
54
55
0.4
Reflectance
0.3
0.2
0.1
0
0
0.2
0.4
0.6
0.8
1
1.2
Diamond Film Thickness (µm)
Figure 2.6. The reflectance for the vacuum/diamond/silicon model as a
function of diamond film thickness.
55
Reflectivity (Arbitrary Units)
56
0
100
200
300
400
500
Growth Time (Minutes)
Figure 2.7. The reflectivity of the growing diamond surface taken during a
diamond film deposition. Note the sharpened cusps at the “bottom” of each
interference fringe.
56
57
Field Emission
Characterization
XPS
AlN/GaN
Gas Source MBE
E-beam Metallization
LEED/AES
Solid Souce
MBE
Angle-Resolved
UPS
ECR Nitrogen
Plasma
Load Lock
Microwave Plasma
Diamond CVD
Wafer Bonding
Hydrogen Plasma
Oxygen Plasma
Figure 2.8. A schematic of the integrated UHV surface characterization and
processing system. Samples (up to 1 inch in diameter) are mounted to
molybdenum sample holders and introduced into the vacuum system
through the load lock.
57
58
Eo
Eo– Evib
Eo
Evib
Evib
Stokes
Interaction
Eo– Evib
Eo+ Evib
Anti-Stokes
Interaction
Eo
Eo+ Evib
Figure 2.9. Schematic of possible photon-lattice scattering events. The
scattering due to optic vibrations or molecular vibrations/rotations is termed
Raman scattering. (The scattering of light by acoustic modes is called
Brillouin scattering.) The spectra of the Raman scattered photons for
Stokes and anti-Stokes interactions are illustrated below each schematic.
58
1332
59
1580
Intensity (Arbitrary Units)
Single Crystal
Diamond
1480
1332
1350
Graphite
1136
Typical C VD
Diamond Film
1000
1200
1400
1600
1800
-1
Wavenumber Shift (cm )
Figure 2.10. Raman scattering spectra obtained from single crystal
diamond, graphite, and a typical CVD diamond film. These spectra were
collected using the 514.5 nm line of an argon ion laser.
59
2.154 eV
(ZPL)
1.945 eV
(ZPL)
Intensity (Arbitrary Units)
60
T=10K
T=300K
1.5
1.7
1.9
2.1
2.3
Energy (eV)
Figure 2.11. Photoluminescence spectra collected at 300 K and 10 K of a
nitrogen-doped diamond film. At 300 K there is significant electron-lattice
interaction, which reduces the ZPL intensity and produces broad sidebands.
Conversely, at lower temperatures the electron-lattice interaction is reduced.
This produces sharper, more intense ZPL features as shown above.
60
61
Cooling
Stage
PMT
Monochromator
Macro
Stage
Microscope
Argon Ion
Laser
Computer
Figure 2.12. Photograph of the optical characterization system used to
collect Raman and Photoluminescence spectra.
61
62
Gas Inlet
Gate Valve
Quartz
Tube
RF Coil
Gate Valve
UHV Transfer
Line
To
Oxygen
Plasma
Chamber
~40 cm
Tungsten
Heater &
Thermocouple
Sample
(facing up)
To Turbopump
Figure 2.13. Diagram of the hydrogen plasma cleaning chamber. Purified
hydrogen gas is introduced into the chamber through the quartz tube. A
plasma is induced by the RF coil. The sample, located ~40 cm away from
the quartz tube, faces the discharge.
62
63
Sample on
Sample Holder
Anode
Electrical
Connection
to Anode
Teflon
Insulator
UHV Stepper
Motor
Worm
Anode
Translation
Stage
1/4-80 Lead
Screw
Worm Gear
Figure 2.14.
Schematic of the variable distance field emission
characterization system. The distance between the sample (fixed) and the
anode (moveable) is adjustable using a UHV stepper motor. A field is
produced by applying a voltage between the anode and the sample. For
clarity, the frame, which supports the sample holder and the anode
translation stage, is not shown in the figure. The entire apparatus is located
inside a stainless steel vacuum chamber with a base pressure
of ~2×10-9 Torr.
63
Measured Current (A)
64
(a)
10
-9
10
-10
10
-11
Decreasing
Anode to
Cathode
Distance
0
200
400
600
800
Voltage for 0.5nA (V)
Applied Voltage (V)
800
(b)
600
400
Slope is the
average
threshold field
200
0
14
12
10
8
6
4
2
0
-2
Relative Distance ( µ m)
Figure 2.15. Illustration of the method used to determine the average
threshold field required for field emission. First, a family of I-V curves is
collected, each at a smaller anode-cathode separation. These curves are
shown in (a). Next, for each I-V curve the voltage required for the
threshold current value (0.5 nA, marked in (a)) is plotted versus relative
distance as shown in (b). The relative distance is the distance the anode has
been moved from the first I-V measurement. The slope of the line shown
in (b) is represents the average threshold field for this measurement.
64
65
Chapter 3
Growth and Properties of Nitrogen-Doped Diamond Films
6.7
Introduction
Nitrogen has long been recognized as one of the most common impurities in
natural and synthetic high pressure, high temperature (HPHT) diamond. The presence
of nitrogen in single crystal diamond has been widely investigated and has been shown
to influence mechanical, thermal, optical, and electrical properties [3.1-3]. In fact,
single crystal diamonds are now scientifically classified according to nitrogen content
and defect complexes. This classification scheme for diamond is summarized in
Table 3.1.
The development of low-pressure CVD techniques for the synthesis of
diamond thin films in the mid-1980’s has presented the possibility of controlled
incorporation of dopants, such as nitrogen, during growth.
Consequently, many
studies have focused upon the effect of nitrogen on the growth of diamond and the
properties of nitrogen-doped diamond films.
In 1990, Yokota and co-workers were one of the first groups to report the
growth and properties of nitrogen-doped diamond films [3.4]. This study focused
upon cathodoluminescence (CL) measurements of nitrogen-diamond films prepared
65
66
by microwave plasma CVD with the addition of nitrogen gas to the CO/H2 plasma
growth environment. Their CL spectra indicated the presence of a luminescent center
at 2.15 eV, which has been attributed to a nitrogen-vacancy center in single crystal
diamond. CL images of isolated diamond nuclei using with the 2.15eV line indicated
that (111) surfaces were luminescent, while the (100) surfaces were not. As a result,
they speculated that nitrogen is preferentially incorporated into the (111) surfaces.
Since the report by Yokota et al., the growth and characterization of nitrogendoped diamond has been extensively studied by many groups.
Nitrogen-doped
diamond films have been synthesized by several different deposition techniques such
as microwave plasma CVD, hot filament CVD, oxyacetylene flame deposition, and so
on. Furthermore, the process conditions employed to grow these films (e.g. chamber
pressure, substrate temperature, methane concentration, source of nitrogen etc.) are
widely varied among the published reports. For this reason, it is difficult to assess the
influence of nitrogen upon the growth and properties of diamond in terms of absolute
concentrations in the process gas mixture. Fortunately, the addition of nitrogen to the
growth environment seems to have similar effects for the different deposition
techniques despite that fact that absolute nitrogen concentration varies significantly.
What follows below is an overview of the present understanding of the growth and
properties of nitrogen-doped diamond thin films.
66
67
6.8
Influence of Nitrogen upon the Growth of Diamond Films
By measuring Raman scattering spectra, Bachmann and co-workers concluded
that the addition of 0.15% nitrogen to a microwave plasma with 0.75% methane in
hydrogen enhances diamond film quality [3.5]. Other groups have also reported an
increase in the diamond line intensity as well as a reduction of non-diamond
contributions in the Raman spectra with the addition of small amounts of nitrogen
[3.6-12].
Raman spectra illustrating the improvement in diamond quality with
nitrogen addition are shown in Figure 3.1.
Locher et al. also studied the effect of low process concentrations of nitrogen
upon the growth of diamond films by microwave plasma CVD [3.6]. In addition to
Raman scattering analysis, the morphology of films was imaged using scanning
electron microscopy (SEM). Without nitrogen addition, the films exhibited a finegrained structure without facets. However, when 60 ppm of nitrogen was admitted
during growth, a highly textured film with coplanar (100) facets was obtained. In
Figure 3.2, SEM images depicting the influence of nitrogen upon diamond surface
morphology are shown. The formation of highly textured (100) diamond films with
nitrogen addition has been well documented in the literature by several other groups
using various deposition techniques and process conditions [3.7,13-17]. X-ray texture
analysis has also been performed on nitrogen-doped textured films [3.6-7,18,19]. This
analysis indicates that a transition occurs from (110) to (100) texture as the nitrogen
concentration in the growth environment increases.
67
68
Rutherford backscattering (RBS) ion channeling experiments have been
performed on homoepitaxial nitrogen doped diamond films prepared by microwave
plasma CVD [3.20]. For these films, it was determined that dislocations were the
primary defect responsible for dechanneling.
Furthermore, the addition of small
concentrations of nitrogen during growth resulted in films with lower dislocations.
This is in agreement with the Raman scattering and SEM results discussed above.
However, if nitrogen is added during growth in larger concentrations, the
diamond quality has been shown to deteriorate. For high nitrogen concentrations, the
diamond Raman line decreases and broadens [3.5-12]. In addition, peaks due to nondiamond bonding become more prominent with increasing nitrogen addition to the
process gas. At very high nitrogen concentrations (>2% nitrogen), Bergman et al.
observed a Raman signature which may be due to nitrogen and carbon bonding in the
films [3.21]. The decrease in film quality at higher nitrogen concentrations has also
been observed with SEM [3.7-8,22-23].
In this regime, the surface morphology
deteriorates resulting in no obvious crystal faceting. SEM images of nitrogen-doped
films grown with high nitrogen concentrations are shown in Figure 3.3.
In addition to influencing the diamond film quality, nitrogen has also been
shown to affect the growth rate of diamond. Müller-Sebert and co-workers studied the
growth rate of diamond films prepared by microwave plasma CVD [3.24]. In this
report, the methane fraction was held at 3% and nitrogen was added from 5 to 25 ppm.
It was observed that the growth rate of diamond could be increased by a factor of five
with the addition of 25 ppm of nitrogen. This growth rate enhancement has also been
68
69
reported by others [3.7-8,18,20]. In the high nitrogen regime where the diamond
quality diminishes, the growth rate has also been shown to decrease [3.8-9,12,18].
The above discussion has centered upon the growth of nitrogen-doped
diamond by microwave plasma CVD.
However, many groups have synthesized
nitrogen-doped films with other deposition techniques such as hot filament CVD
[3.12,14,25-28] and oxyacetylene flame deposition [3.14,29-30]. Nitrogen is very
difficult to dissociate, with a bond strength of 945 kJ/mol. In plasma deposition
systems, dissociation is possible through collisions with hot electrons. However, in
hot-filament and the flame deposition techniques, dissociation occurs through
primarily thermodynamic means.
Consequently, gases other than nitrogen (e.g.
ammonia) are often used for the deposition of nitrogen-doped diamond. Despite
differences in nitrogen doping sources, analogous observations to that described above
have been reported for these techniques.
6.9
Characterization of Nitrogen-Doped Diamond Films
There has been considerable interest in characterizing the properties of the
nitrogen-doped films and comparing these properties with those of natural and
synthetic single crystal diamond.
As discussed earlier, cathodoluminescence
measurements by Yokota et al. indicated the presence of the bands due to the nitrogenvacancy system commonly observed in Type Ib diamond where nitrogen is
incorporated substitutionally [3.4]. Measurements by other groups have also verified
69
70
the presence of these nitrogen-vacancy centers in cathodoluminescence spectra
[3.14,31]. The influence of nitrogen doping upon the photoluminescence spectra has
also been extensively investigated [3.6,13,21]. These reports have also identified the
nitrogen-vacancy centers associated with single substitutional nitrogen.
Despite the identification of these nitrogen-vacancy centers, a quantitative
measurement of the concentration of substitutional nitrogen is not possible with these
optical methods. For this reason, there has been a significant research using analytical
techniques such as secondary ion mass spectrometry (SIMS) [3.22-23,25,32], electron
paramagnetic resonance (EPR) [3.15,32-33], nuclear reaction analysis (NRA)
[3.14,34], and electron recoil detection (ERD) [3.16,33] to determine the nitrogen
doping concentration.
Using SIMS, Mort et al. have reported that nitrogen concentrations up
to 4×1019 cm-3 have been measured for diamond films grown by hot filament CVD
[3.25]. In addition, these results indicate substantial nitrogen uniformity throughout
the film. By doping with isotopical 15N2 gas, Samlenski and co-workers used NRA to
determine the concentration of incorporated
15
N in (100) and (111) homoepitaxial
films grown by microwave plasma CVD [3.14,34]. These reports, indicated nitrogen
concentrations up to 2.3×1018 cm-3.
In addition, their measurements revealed a
preferred incorporation of nitrogen into (111) growth surfaces by a factor of 3-4.
These studies quantitatively verified the earlier observations by Yokota et al. [3.4].
Unfortunately, SIMS and NRA evaluate the total concentration of species
present in the film and cannot distinguish between substitutional and aggregated
70
71
nitrogen.
For this reason, EPR has been used to determine the amount of
substitutional nitrogen in diamond thin films. EPR detects the presence of impurities
and defects with an unpaired electron. For this reason, EPR has become a valuable
technique in determining the concentration of substitutional nitrogen. Hoinkis et al.
used SIMS and EPS in a complementary fashion to determine that nitrogen is
predominantly incorporated in substitutional form [3.32].
The electrical characteristics of nitrogen-doped diamond films have also been
studied [3.15,25,33,35-36]. Mort and co-workers studied the effect of nitrogen doping
upon the temperature dependence of the electrical conductivity [3.25]. A decrease of
several orders of magnitude was observed in the conductivity for temperatures at or
above room temperature.
This decrease was attributed to a compensation effect
between nitrogen donors and the background “p-type” defects. Similar measurements
by Jin and Moustakas obtained an activation energy of ~1.5 eV for nitrogen-doped
diamond films [3.35]. This is in good agreement with the value of 1.7 eV reported for
single substitutional nitrogen in single crystal Type Ib diamond [3.1-3]. By studying
the photoconductivity of nitrogen doped films, Graeff et al. also observed a donor
level ~1.5 eV below the conduction band minimum [3.33]. In addition, their results
indicated that these nitrogen donor states were compensated by defects in the film.
The thermal properties of nitrogen doped diamond CVD films have also been
studied [3.24,37]. In a study by Bachmann and co-workers, diamond films were
synthesized by various growth techniques and chemistries.
They found that the
thermal conductivity sharply decreases for diamond films grown with nitrogen [3.37].
71
72
They observed that the addition of nitrogen to the growth environment increases the
concentration of sp2 carbon in the deposited films. It was suggested that sp2 carbon is
responsible for the reduction of the phonon mean free path by scattering the phonons
in the crystal. The reduction of the phonon mean free path is then observed in a
decrease in the thermal conductivity.
6.10 Growth of Nitrogen Doped Diamond: Chemistry & Growth Mechanisms
Recently, emphasis has been placed not necessarily upon the growth and
characterization of nitrogen-doped diamond, but on the actual growth mechanisms.
Many groups have monitored the chemistry during growth using optical emission
spectroscopy (OES) [3.6,9,17-19,24] or molecular beam mass spectrometry (MBMS)
[3.28,38]. OES has the disadvantage that only optical transitions from excited species
can be observed.
Similarly, with MBMS only stable species can be monitored.
Additionally, the instrumentation may also effect the observed chemistry.
Nevertheless, these techniques indicate that nitrogen addition results in the formation
of HCN or CN radicals. The concentration of HCN and CN increase with increasing
nitrogen addition. It has been observed that for low nitrogen content, the other growth
species remain unchanged [3.17].
However, for large amounts of nitrogen the
concentration of CH radicals decreases significantly with increasing nitrogen
content [3.9].
72
73
These studies suggest that at low nitrogen process conditions, HCN or CN may
abstract hydrogen from the diamond surface and therefore alter the growth process.
Since these species have low desorption rates, the growth rate increases due to an
increased incorporation of these species. However, at high nitrogen concentrations,
the growth rate of diamond is suppressed due to the reduction of CH precursors
through the formation of HCN or CN.
6.11 Other Research
Other groups have intentionally introduced nitrogen into the growth
environment for reasons other than the growth and characterization of nitrogen-doped
diamond films. In an early study, Badzian and co-workers introduced large amounts
of nitrogen (>20%) during growth in an attempt to synthesize C3N4 [3.39].
In
addition, they explored the possibility of depositing films without hydrogen. Despite
their efforts, only highly defective carbon deposits were obtained. Since this report,
many groups have also been unsuccessful in their attempts to synthesize C3N4 by the
addition of nitrogen containing species to diamond growth environment [3.27,38].
In an attempt to grow phosphorus-doped n-type diamond, Cao and co-workers
discovered that by codoping with nitrogen and phosphorus, the phosphorus doping
efficiency could greatly increased [3.40-42]. Despite incorporation of nitrogen and
phosphorus at concentrations of 6×1019 cm-3 and 3×1019 cm-3, respectively, the films
were too insulating for Hall measurements.
73
74
6.12 Conclusions
The growth and properties of nitrogen-doped diamond films are well
documented in the literature. It has been observed that the addition of small amounts
of nitrogen to the growth environment increases crystalline quality, growth rate, and
promotes (100)-textured films.
Conversely, at higher nitrogen concentrations the
diamond film quality decreases, along with a commensurate decrease in the growth
rate and morphology. Analysis of high quality nitrogen-doped films indicates that
nitrogen occupies primarily substitutional sites with concentrations on the order of
1018 to 1019 cm-3. Optical, electrical, and thermal characterization reveal that nitrogendoped diamond films are very similar to the properties single crystal Type Ib diamond.
However, electrical compensation of nitrogen donors was observed for some nitrogendoped films.
74
75
References
3.1
The Properties of Diamond, ed. J.E. Field (Academic Press, London, 1979).
3.2
The Physics of Diamond, eds. A. Paoletti and A. Tucciarone (IOS Press,
Amsterdam, 1997).
3.3
Handbook of Industrial Diamonds and Diamond Films, eds. M.A. Prelas, G.P.
Popovici, and L.K. Bigelow (Marcel Dekker, New York, NY, 1998).
3.4
Y. Yokota, H. Kawaranda, and A. Hiraki, MRS Symp. Proc. 162 p.231 (1990).
3.5
P.K. Bachmann and D.U. Weichert, Diamond Relat. Mater. 1, p.422 (1992).
3.6
R. Locher, C. Wild, N. Herres, D. Behr, and P.Koidl, Appl. Phys. Lett. 65 p.34
(1994).
3.7
S. Jin and T.D. Moustakus, Appl. Phys. Lett. 65 p.403 (1994).
3.8
S. Bohr, R. Hauber, and B. Lux, Appl. Phys. Lett. 68 p.1075 (1996).
3.9
H. Chatei, J. Bougdira, M. Rémy, P. Alnot, C. Bruch, and J. Krüger, Diamond
Relat. Mater. 6, p.107 (1997).
3.10 H. Chatei, J. Bougdira, M. Rémy, P. Alnot, C. Bruch, and J. Krüger, Diamond
Relat. Mater. 6, p.505 (1997).
3.11 T. Vandevelde, M. Nesladek, K. Meykens, C. Quaeyhaegens, L.M. Stals, I.
Gouzman, and A. Hoffman, Diamond Relat. Mater. 7, p.152 (1998).
3.12 A. Afzal, C.A. Rego, W. Ahmed, and R.I. Cherry, Diamond Relat. Mater. 7,
p.1033 (1998).
3.13 K. Kania and P. Oelhafen, Diamond Relat. Mater. 4, p.425 (1995).
3.14 G.Z Cao, J.J. Schermer, W.J.P. van Enckevort, W.A.L.M. Elst, and L.J. Giling,
J. Appl. Phys. 79 p.1357 (1996).
3.15 M. Fanciulli, S. Jin, and T.D. Moustakas, Physica B 229, p.27 (1996).
75
76
3.16 A. Bergmaier, G. Dollinger, T. Faestermann, C.M. Frey, F. Ferguson, H.
Guttler, G. Schultz, and H. Willerscheid, Diamond Relat. Mater. 5, p.995
(1996).
3.17 T. Vandevelde, M. Nesladek, C. Quaeyhaegens, and L. Stals, Thin Solid Films
290-291 p.143 (1996).
3.18 T. Vandevelde, M. Nesladek, C. Quaeyhaegens, and L. Stals, Thin Solid Films
308-309 p.154 (1997).
3.19 T. Vandevelde, M. Nesladek, K. Meykens, C. Quaeyhaegens, L.M. Stals, I.
Gouzman, and A. Hoffman, Diamond Relat. Mater. 7, p.152 (1998).
3.20 R. Samelski, J. Schmälzlin, R. Brenn, C. Wild, W. Müller-Sebert, and P. Koidl,
Diamond Relat. Mater. 4, p.503 (1995).
3.21 L. Bergman, M.T. McClure, J.T. Glass, and R.J. Nemanich, J. Appl. Phys. 76
p.3020 (1994).
3.22 M. Nesládek, K. Meykens, L.M. Stals, C. Quaeyhaegens, M. D’Olieslaeger,
T.D. Wu, M.Vaněčk, and J. Rosa, Diamond Relat. Mater. 5, p.1006 (1996).
3.23 M. Nesládek, M.Vaněčk, and L.M. Stals, Phys. Stat. Sol. 154, p.283 (1996).
3.24 W. Müller-Sebert, E. Wörner, F. Fuchs, C. Wild, and P. Koidl, Appl. Phys. Lett.
68 p.759 (1996).
3.25 J. Mort, M.A. Machonkin, and K. Okumura, Appl. Phys. Lett. 59 p.3148 (1991).
3.26 P.W. May, B.R. Burridge, C.A. Rego, R.S. Tsang, M.N.R. Ashfold, K.N.
Rosser, R.E. Tanner, D. Cherns, and R. Vincent, Diamond Relat. Mater. 5 p.354
(1996).
3.27 H. Spicka, M. Griesser, H. Hutter, M. Grasserbauer, S. Bohr, R. Haubner, B.
Lux, Diamond Relat. Mater. 5, p.383 (1996).
3.28 R.S. Tsang, C.A. Rego, P.W. May, M.N.R. Ashfold, and K.N. Rosser, Diamond
Relat. Mater. 6, p.247 (1997).
3.29 R.J.H. Klein-Douwel, J.J. Schermer, J.J ter Meulen, Diamond Relat. Mater. 7,
p.1118 (1998).
3.30 C.A. Wolden, C.E. Draper, Z. Sitar, and J.T. Prater, Diamond Relat. Mater. 7,
p.1178 (1998).
76
77
3.31 S. Sonada, J.H. Won, H. Yagi, A. Hatta, T. Ito, and A. Hiraki, Appl. Phys. Lett.
70 p.2574 (1997).
3.32 M. Hoinkis, E.R. Weber, M.I. Landstrass, M.A. Plano, S. Han, D.R. Kania,
Appl. Phys. Lett. 59 p.1870 (1991).
3.33 C.F.O Graeff, E. Rohrer, C.E. Nebel, M. Stutzmann, H. Güttler, and R. Zachai,
Appl. Phys. Lett. 69 p3215 (1996).
3.34 R. Samelski, C. Haug, R. Brenn, C. Wild, R. Locher, and P. Koidl, Diamond
Relat. Mater. 5, p.947 (1996).
3.35 S. Jin and T.D. Moustakus, Appl. Phys. Lett. 63 p.2354 (1993).
3.36 B.B. Li, M.C. Tosin, A.C. Peterlevitz, and V. Baranauskas, Appl. Phys. Lett. 73
p.812 (1998).
3.37 P.K. Bachmann, H.J. Hagemann, H. Lade, D. Leers, D.U. Weichert, H. Wilson,
D. Fournier, and K. Plamann, Diamond Relat. Mater. 4, p.820 (1995).
3.38 P.W. May, B.R. Burridge, C.A. Rego, R.S. Tsang, M.N.R. Ashfold, K.N.
Rosser, R.E. Tanner, D. Cherns, and R. Vincent, Diamond Relat. Mater. 5,
p.354 (1996).
3.39 Badzian, T. Badzian, and S.T. Lee, Appl. Phys. Lett. 62 p.3432 (1993).
3.40 G.Z Cao, W.J.P. van Enckevort, L.J. Giling, and R.C.M. de Kruif, Appl. Phys.
Lett. 66 p.688 (1995).
3.41 G.Z Cao, L.J. Giling, and P.F.A. Alkemade, Diamond Relat. Mater. 4, p.775
(1995).
3.42 G.Z Cao, F.A.J.M. Driessen, G.J. Bauhuis, L.J. Giling, and P.F.A. Alkemade, J.
Appl. Phys. 78 p.3125 (1995).
77
78
Type Ia
Type Ib
Type IIa
Type IIb
Nitrogen
Nitrogen
Nitrogen
Boron
(In Aggregated
Form)
(Isolated
Substitutional)
(Little or no
nitrogen)
(Isolated
Substitutional)
Impurity
Concentration:
1019-1021 cm-3
1018-1020 cm-3
≤ 1013 cm-3
1014-1020 cm-3
Typical Color:
White to
Yellow
Yellow
White
Blue/Gray
>1016 Ω⋅cm
ED=~4.0 eV
>1013 Ω⋅cm
ED=~1.7 eV
>1016 Ω⋅cm
10-1-104 Ω⋅cm
EA=0.375 eV
Thermal
Conductivity
(at 300 K):
500-1800
W/m⋅K
~900
W/m⋅K
2000-2500
W/m⋅K
1840-2020
W/m⋅K
Natural
Abundance:
~98%
≤ 1%
Very Rare
Extremely
Rare
Main Impurity:
Electrical
Resistivity:
Table 3.1. The classification scheme for single crystal diamond. Typical
materials properties are also shown. Data taken from [3.1-3].
78
79
Figure 3.1. Raman scattering spectra of nitrogen-doped diamond films
deposited with the indicated [N]/[C] ratios in the gas phase. A He-Ne laser
(λ=632.8 nm) was used as the excitation source. From [3.7].
79
80
(a)
(b)
Figure 3.2. SEM micrographs of two CVD diamond films prepared with
1.5% methane. (a) without nitrogen admixture, (b) with 60 ppm nitrogen
admixed. From [3.6].
80
81
(a)
(b)
Figure 3.3. SEM micrographs of two diamond films grown by microwave
plasma CVD from a H2/C6H14 plasma. (a) with 100 ppm nitrogen admixed,
(b) with 1 vol.% nitrogen admixed. From [3.22].
81
82
Chapter 4
Introduction and Review of Field Emission from Diamond
4.1
Introduction and Historical Background
In 1897, Wood first reported the observation of cold field emission [4.1].
Subsequently, other groups reported the properties of vacuum gaps under large applied
voltages [4.2-4]. Prior to vacuum breakdown, these studies indicated that a small but
detectable current flows between the electrodes. It was also observed that these
“prebreakdown” currents increased rapidly with increasing gap voltage until vacuum
breakdown occurred. Despite these and other reports, the physics describing these
prebreakdown or field emission currents remained undetermined.
In 1928, Fowler and Nordheim developed a quantum mechanical theory of
field emission from metallic surfaces [4.5-7]. In this treatment, the field emitted
current was determined by calculating the transmission coefficient (or tunneling
probability) through the modified surface potential barrier. From simple electrostatic
theory, an electron outside a conductor is attracted to the metal surface by its “image”
charge. The magnitude of this attractive force is given by:
F=
1  e2

4πε o  4z 2



(1)
82
83
where εo is the permittivity of free space, e is the charge on an electron, and z is the
distance from the metal surface. The resultant potential combining both the image
potential and the applied field is shown in Figure 4.1. In the Fowler-Nordheim theory,
the total electron emission current density, j(E), is given by:
∞
j (E ) = e ∫ N (ε )D (E , ε )dε
0
(2)
where E is the applied field, ε is the electron energy, N(ε) represents the number of
electrons impinging upon the surface barrier, and D(E,ε) is the transmission
coefficient. At low temperatures (T ≤ 300 K) where only a few electron states above
the Fermi level are occupied, the above expression can be evaluated using the WenzelKramers-Brillouin (WKB) approximation to give the Fowler-Nordheim equation:
 8π 2m e Φ 3 2 ν(y ) 
e 3E 2

j (E ) =
exp


heE
8πhΦ t 2 ( y )
3


(3)
where h is Planck’s constant, Φ is the work function of the metal, me is the mass of an
electron, and t(y) and ν(y) are the Nordheim elliptical functions. Often t(y) and ν(y)
are taken as unity so the Fowler-Nordheim equation can be written in the following
form:
 − C 2Φ 3 2 
C 1E 2

j (E ) =
exp
Φ
E


(4)
where C1 and C2 are constants. The success of the Fowler-Nordheim field emission
model from metal surfaces was one of the early experimental verification of quantum
mechanical tunneling.
83
84
For electrodes in a parallel plate configuration, the electric field can be
written as:
E=
βV
d
(5)
where V is the applied voltage, d is the distance between the anode and cathode, and β
is the geometric field enhancement factor of the cathode surface. For example, for a
pointed needle of length l and radius r, β can be modeled as [4.7-8]:
β=
2
.
r ln (2l / r )
(6)
Whereas for flat surfaces, we have β equal to 1. With this model the Fowler
Nordheim equation can be expressed as a function of applied voltage:
2
 − C2Φ 3 2d 
C  βV 

 .
(7)
j( V ) = 1 
exp

Φ d 
β
V


In this form, it is evident that for a given voltage the field emission current will be
maximized for low work function materials and/or by employing pointed emitters with
large field enhancement factors.
4.2
Field Emission Research on Metal Cathode Arrays
In order to use the Fowler-Nordheim equation for metals, the majority of the
early field emission research was devoted towards the transition and refractory
metals [4.9]. However, the large fields (>1000 V/µm) required for emission from
planar metallic surfaces has limited their applicability. Consequently, there has been
much emphasis placed on the development of sharp, pointed field emitters in array
84
85
structures. A major development in the field emission arena was the introduction of
the Spindt cathode in 1968 [4.10].
The Spindt cathode is fabricated from the
deposition of molybdenum through an ever-decreasing aperture. A schematic of this
vacuum microelectronic device is shown in Figure 4.2.
The field emission
characteristics of the Spindt cathode have been refined since its introduction [4.11,12].
Spindt has reported emission of 10 µA for a single molybdenum emitter [4.13].
Similar Spindt-type cathodes have also been fabricated from heavily n-doped silicon
[4.14,15]. Although a semiconductor, degenerately n-doped silicon has a conductivity
similar to metals, and the Fowler-Nordheim equation can still be employed. Emission
currents of up to 50 µA per tip for an applied voltage of 150 V have been reported for
these silicon-based emitters [4.16]. Despite these developments, Spindt-type cathodes
are susceptible to degradation from back-sputtered positive ions and the adsorption of
species on the emitting surfaces [4.11,17].
In order to alleviate the problem of cathode sputtering by self-generated ions,
the reduction of the required operating voltage of field emission devices has been
identified as a primary goal in ongoing research. One approach to this problem has
been the use of low work function materials (e.g. cesium, potassium, and sodium) as
coatings for field emission devices [4.18].
However, widespread use of these
materials as coatings has been hampered by their materials properties. For example,
these materials are highly reactive, and extreme measures are needed to prevent the
surface from forming a high work function oxide.
85
86
4.3
Field Emission from Semiconducting Diamond
With the discovery of the diamond negative electron affinity surfaces and the
development of diamond deposition techniques, it has been suggested that diamond
would be an ideal material to coat Spindt-type field emitters or to replace them
altogether [4.19,20].
It must be emphasized at this point that diamond is a
semiconductor and the field emission characteristics may not be governed by the
Fowler-Nordheim equation for metals. Detailed derivations and discussion of the
equations describing emission from semiconductors has been published in many books
and articles [4.21-25].
Although the functional form governing emission from
semiconductors is similar to the Fowler-Nordheim equation, modifications are made to
take into account field penetration into the semiconductor, the presence of surface
states, and whether the emission originates from the valence band, conduction band, or
impurity states.
In contrast to metals, the field emission process for semiconductors combines
four effects: (1) electron supply to the semiconductor, (2) transport through the
semiconductor, (3) emission into vacuum, and (4) transport through vacuum to the
anode. The initial studies by Himpsel et al. [4.19] and Pate [4.20] demonstrate that,
unlike metal surfaces, the emitting surface will not limit the emission for hydrogenterminated (111) diamond surfaces.
More recently, the (100), (110), and (311)
diamond surfaces have also been observed to exhibit a NEA [4.26-29]. This presented
the possibility that diamond based cathodes could be fabricated which exhibit high
86
87
emission currents at extremely low applied fields. In addition, flat diamond surfaces
could replace the Spindt-type cathodes such that cathode degradation due to sputtering
could be minimized or eliminated.
These results were largely responsible for
stimulating the development of diamond based field emitters.
4.4
Field Emission Measurement Techniques
The characterization methods used to study the field emission properties of
diamond are discussed below. One of the simplest and most common techniques used
to study the field emission properties of diamond is the “parallel plate method”
[4.30-33]. This characterization technique is illustrated in Figure 4.3. With this
procedure, the cathode surface is separated from a large flat anode by a thin
(typically 50-100 µm) insulating spacer. Common spacer materials include Kapton
foil and glass fibers. This method has the advantage that the anode can be replaced
with phosphor or indium tin oxide (ITO) screen to image the emission sites. However,
this technique is very sensitive to misalignment between the anode and cathode. In
addition, since the entire cathode surface is explored, the current voltage
characteristics may be dominated by strong emission areas not representative of the
whole diamond surface. Finally, obtaining the required high vacuum may be more
difficult for this arrangement. Another characterization procedure that has gained in
popularity is the “variable distance anode technique” [4.34-37]. A schematic of this
technique is shown in Figure 4.4. With this method, current versus voltage curves can
87
88
be obtained at various distances from the cathode surface. Anode-to-cathode distances
from a few microns to a hundred or more are possible. To avoid field enhancement at
the edges of the anode, the corners are usually rounded to a high radius of curvature,
or alternatively a spherical anode is used.
This technique has the advantage of
determining the field emission properties as a function of distance. A disadvantage is
that this apparatus is more complex than its “parallel plate” counterpart. In addition,
the anode-to-cathode distance is often not known with absolute precision. As a result,
this distance is often determined by an additional technique or analysis that might
introduce additional error into the field emission characterization. A third technique
that has been used to characterize the field emission from diamond is field emission
energy distribution (FEED) measurements [4.38-41]. This method is probably the
most complex, yet potentially the most powerful of the techniques discussed thus far.
In FEED, field emission occurs as a result of a potential applied between the sample
and a grid or sharpened tip. Most of the electrons are collected by the anode, however
a small fraction escape collection and are energy analyzed. By reference to the Fermi
level, this energy analysis can yield information regarding the origin of the emitted
electrons and possible field emission measurements. FEED is illustrated in Figure 4.5.
4.5
Field Emission Results from Diamond
In 1991, C. Wang et al. was the first to report field emission from diamond
surfaces [4.42]. Since then, literally hundreds of studies on field emission have been
88
89
published internationally. It is beyond the scope of this review to cover all aspects of
field emission from diamond. However, there are several significant reports, which
have contributed to the present understanding of the field emission properties of
diamond and diamond films.
Studies by Zhu et al. have indicated a correlation between the defect density
present in undoped and p-type diamond films and the required field for emission
[4.34,43-44]. In these reports, it was found that the presence of high defect densities
led to a reduction of the threshold field. It was proposed that the defects in the film
created defect subbands within the bandgap of the diamond. It is speculated that these
defect bands may allow for electron transport to the emitting surface via an electron
hopping mechanism. In theoretical investigations, Huang et al. have also suggested
that states in the bandgap must be present and must participate in the emission process
to account for the low field emission from diamond [4.45-47]. Many other groups
have seen this defect enhanced effect [4.33,48-49]. In fact, much research has now
been devoted toward studying the field emission properties of diamond-like-carbon
and other carbon-based materials with high concentrations of defects.
Other groups have suggested that the presence of graphite inclusions may also
explain the low field emission behavior of diamond films [4.38,42,50-51]. When an
external electric field is applied, the field penetrates into the diamond and interacts
with small non-diamond carbon particles forming conducting channels through the
diamond. These conducting channels will bias the inclusions to nearly the substrate
89
90
potential. This effect results in an extremely high field at these particles, and as a
result there is a high probability that emission will occur.
Bandis and Pate have studied the field emission mechanism for single crystal
p-type diamond using simultaneous field emission and photoemission experiments
[4.39].
From these experiments, it was found that the field-emitted electrons
originated from the valence band maximum. The temperature dependence of the field
emission from p-type diamond films was studied by Glesener and Morrish [4.52].
These results showed no dependence of the field emission properties with temperature.
They indicate this behavior is consistent with emission from the valence band.
Gröning et al. have reported that stable field emission from diamond and
amorphous carbon films is preceded by an arcing or “activation” event [4.35]. This
arcing event leads to the formation of craters, molten areas, and surface debris.
Following the arcing event, there was a reduction of the threshold field required for
emission, as well as an increase in emitted current.
It is suggested that the
improvement in the field emission characteristics is due to the formation of sharp field
enhancing protrusions. Arcing and corresponding improvements in the field emission
characteristics have also been reported by many other research groups [4.32,53].
It should be emphasized at this point that all the mechanisms proposed for field
emission thus far do not involve emission from the conduction band. Although the
negative electron affinity surface of diamond may play a role in enhancing the field
emission properties, these mechanisms indicate that it alone is not responsible for the
observed low field emission from diamond. As a result, considerable research has
90
91
been devoted towards injecting electrons into the diamond conduction band.
In
principle, the use of n-type diamond would solve this problem. Then electrons could
be supplied to the diamond conduction band via a low resistance ohmic contact. The
electrons would then be transported through the bulk and emitted at the negative
electron affinity surface. To date, however, n-type diamond material is still not
routinely available. Although substitutional nitrogen is a relatively deep donor located
1.7 eV below the conduction band, experimental and theoretical studies have indicated
that high concentrations of nitrogen enhance the field emission characteristics
[4.36,54-57]. It has been proposed that a Schottky contact with a narrow depletion
region is formed between the diamond/metal interface. Upon large applied negative
bias, electrons can tunnel into the conduction band through this barrier and are emitted
at the negative electron affinity surface. Despite these reports, the actual role of
nitrogen in the field emission process still needs to be explored. As a result, this
dissertation is directed towards understanding the effects of nitrogen doping on the
field emission process. The field emission characteristics of nitrogen doped diamond
films are presented in the next chapter.
4.6
Conclusions
Many interesting experiments have been conducted to investigate the field
emission characteristics of diamond. The results, indicate that low field electron
emission from diamond based materials is possible. However, the actual mechanisms
91
92
governing the field emission process remains largely undetermined. Field emission
models involving defect bands in the bandgap as well as graphitic inclusions have
been proposed to enhance the field emission properties of diamond. Simultaneous
photo and field emission measurements from p-type diamond have indicated that the
field-emitted electrons originate from the valence band. None of these proposed
emission mechanisms involve emission from the conduction band. As a result, the
negative electron affinity of diamond plays a minor role.
Emission from the
conduction band has been proposed for nitrogen-doped diamond. However, this has
not been firmly established.
Because of its negative electron affinity property, much attention has been
devoted toward developing diamond as an efficient field emitter. However, other
wide bandgap materials, such as AlN, AlGaN, and BN have also been shown to
exhibit a negative electron affinity [4.58-59]. AlN and AlGaN have the advantage that
they can be grown heteroepitaxially on SiC substrates. However, like diamond, n-type
AlN is not yet available. Field emission studies on these wide bandgap materials are a
focus of ongoing research. In Appendix B, the fabrication and operation of cold
cathode structures utilizing AlN and AlGaN as the field emitting layers are discussed.
92
93
References
3.43 R.W. Wood, Phys. Rev. Ser. I 5 p.1 (1897).
3.44 R.F. Earhart, Phil. Mag. 1 p.147 (1901).
3.45 G.M. Hobbs, Phil. Mag. 10 p.617 (1905).
3.46 R.A. Millikan and R.A. Sawyer, Phys. Rev. 12 p.167 (1918).
3.47 R.H. Fowler and L. Nordheim, Proc. Roy. Soc. A119 p.173 (1928).
3.48 L. Nordheim, Proc. Roy. Soc. A121 p.626 (1928).
3.49 R. Gomer, Field Emission and Field Ionization (Harvard University Press,
Cambridge, Massachusetts, 1961), reissued as an American Vacuum Society
Classic (American Institute of Physics, New York, NY, 1993).
3.50 High Voltage Vacuum Insulation, ed. R.V. Latham, (Academic Press, San
Diego, 1995).
3.51 A. Modinos, Surf. Sci. 70 p.52 (1978).
3.52 C.A. Spindt, J. Appl. Phys. 39 p.3504 (1968).
3.53 C.A. Spindt, I. Brodie, L. Humphrey, and E.R. Westerberg, J. Appl. Phys. 47
p.5248 (1976).
3.54 P.R. Schwoebel, and I. Brodie, J. Vac. Sci Technol. B 13 p.1391 (1995).
3.55 C.A. Spindt, IEEE Trans. Electron Dev. 38 p.2355 (1990).
3.56 R.N. Thomas, R.A. Wickstrom, D.K. Schroder, and H.C. Nathanson, Solid State
Electron. 17 155 (1974).
3.57 G.G.P. van Gorkom, and A.M.E. Hoeberechts, J. Vac. Sci Technol. B 4 p.108
(1986).
3.58 I. Brodie and C.A. Spindt, Adv. Electron. Electron Phys. 83 p.1 (1992).
3.59 R. Smith, J. Phys. D 17 p.1045 (1984).
3.60 S. Iannazzo, Solid State Electron. 36 p.301 (1993).
93
94
3.61 F.J. Himpsel, J.A. Knapp, J.A. van Vechten, and D.E. Eastman, Phys. Rev. B 20
p.624 (1979).
3.62 B.B. Pate, Surf. Sci. 165 p.83 (1986).
3.63 R. Gomer, Field Emission and Field Ionization (Harvard University Press,
Cambridge, Massachusetts, 1961), reissued as an American Vacuum Society
Classic (American Institute of Physics, New York, NY, 1993).
3.64 R. Stratton, Phys. Rev. 125 p.67 (1962).
3.65 R. Stratton, Phys. Rev. A 135 p. 794 (1964).
3.66 A. Modinos, Field, Thermionic, and Secondary Electron Emission Spectroscopy
(Plenum, New York, 1984).
3.67 K.L. Jensen, J. Vac. Sci Technol. B 13 p.516 (1995).
3.68 J. van der Weide, Z. Zhang, P.K. Baumann, M.G. Wensell, J. Bernholc, and R.J.
Nemanich, Phys. Rev. B 50 p.5803 (1994).
3.69 J. van der Weide and R.J. Nemanich, J. Vac. Sci. Technol. B 12, 2475 (1994).
3.70 P.K. Baumann and R.J. Nemanich, Diamond Relat. Mater. 4 p.802 (1995).
3.71 P.K. Baumann and R.J. Nemanich, Surf. Sci. 409 p.320 (1998).
3.72 K. Okano, S. Koizumi, S.R.P. Silva, G.A.J. Amaratunga, Nature 381 p.140
(1996).
3.73 J.S. Lee, K.S. Liu, and I.N. Lin, J. Appl. Phys. 82 p.3310 (1997).
3.74 N.A. Fox, S. Mary, T.J. Davis, W.N. Wang, P.W. May, A.Bewick, J.W. Steeds,
and J.E. Butler, Diamond Rel. Mater. 6 p.1135 (1997).
3.75 K.H. Park, S. Lee, K.H. Song, J.I. Park, K.J. Park, S.Y. Han, S.J. Na, N.Y. Lee,
and K.H. Koh, J. Vac. Sci. Technol. B 16 p.724 (1998).
3.76 W. Zhu, G.P. Kochanski, S. Jin, and L. Seibles, J. Appl. Phys. 78 p.2707 (1995).
3.77 O. Gröning, O.M. Küttel, E. Schaller, P. Gröning, and L. Schlapbach, Appl.
Phys. Lett. 69 p.476 (1996).
3.78 M.W. Geis, J.C. Twichell, N.N. Efremow, K.E. Krohn, and T.M. Lyszczarz,
Appl. Phys. Lett. 68 p.2294 (1996).
94
95
3.79 F. Lacher, C. Wild, D. Behr, P. Koidl, Diamond Rel. Mater. 6 p.1111 (1997).
3.80 N.S. Xu, R.V. Latham, Y. Tzeng, Electron. Lett. 29 p.1596 (1993).
3.81 C. Bandis and B.B. Pate, Appl. Phys. Lett. 69 p.366 (1996).
3.82 R. Schlesser, M.T. McClure, W.B. Choi, J.J. Hren and Z. Sitar, Appl. Phys. Lett.
70 p.1596 (1997).
3.83 R. Schlesser, M.T. McClure, B.L. McCarson and Z. Sitar, J. Appl. Phys. 82
p.5763 (1997).
3.84 C. Wang, A. Garcia, D.C. Ingram, M. Lake, M.E. Kordesch, Electron. Lett. 27
p.1459 (1991).
3.85 W. Zhu, G.P. Kochanski, S. Jin, L. Seibles, D.C. Jacobson, M. McCormack, and
A.E. White, Appl. Phys. Lett. 67 p.1157 (1995).
3.86 W. Zhu, G.P. Kochanski, S. Jin, and L. Seibles, J. Vac. Sci Technol. B 14 p.2011
(1996).
3.87 Z.H. Huang, P.H. Cutler, N.M Miskorvsky, and T.E. Sullivan, Proceedings of
the 7th International Vacuum Microelectronics Conference, Grenoble, France,
p.92 (1994).
3.88 Z.H. Huang, P.H. Cutler, N.M Miskorvsky, and T.E. Sullivan, Appl. Phys. Lett.
65 p.2562 (1994).
3.89 Z.H. Huang, P.H. Cutler, N.M Miskorvsky, and T.E. Sullivan, J. Vac. Sci
Technol. B. 13 p.526 (1995).
3.90 N.A Fox, W.N. Wang, T.J. Davis, J.W. Steeds and, P.W. May, Appl. Phys. Lett.
71 p.2337 (1997).
3.91 A. Wisitsora-at, W.P. Kang, J.L Davidson, and D.V. Kerns, Appl Phys Lett. 71
p.3394 (1997).
3.92 N.S. Xu and R.V. Latham, J. Phys. D 19 p.477 (1986).
3.93 N.S. Xu, Y. Tzeng, and R.V. Latham, J Phys. D 26, p.1776 (1993).
3.94 J.W. Glesener and A.A. Morrish, Thin Solid Films 290/291 p.153 (1996).
3.95 P.W. May, S. Höhn, W.N Wang, and N.A. Fox, Appl. Phys. Lett. 72 p.2182
(1998).
95
96
3.96 M.W. Geis, J.C. Twichell, and T.M. Lyszczarz, J. Vac. Sci Technol. B 14 p.2060
(1996).
3.97 K. Okano, S. Koizumi, S.R.P. Silva, G.A.J. Amaratunga, Nature 381 p.140
(1996).
3.98 P. Lerner, P.H. Cutler, and N.M Miskorvsky, J. Vac. Sci Technol. B 15 p.337
(1997).
3.99 P. Lerner, N.M Miskorvsky, and P.H. Cutler, J. Vac. Sci Technol. B 16 p.900
(1998).
3.100 M.C. Benjamin, C. Wang, R.F. Davis, and R.J. Nemanich, Appl. Phys. Lett. 64
p.3288 (1994).
3.101 M.J. Powers, M.C. Benjamin, L.M. Porter, R.J. Nemanich, R.F. Davis, J.J
Cuomo, G.L. Doll, and S.J. Harris, Appl. Phys. Lett. 67 p.3912 (1995).
96
97
Applied Field
Image Potential
Φ
Net Potential
e-
Metal
Vacuum
Figure 4.1.
Schematic of the potential energy diagram for the
metal/vacuum interface with a large applied external field. In addition,
Fowler-Nordheim type tunneling through this potential barrier is illustrated
with a simplified electron wavefunction.
97
98
Metal Tips
Metal Gate
SiO2 Spacer
~1 µm
~1 µm
Silicon Substrate
(Note: Not to Scale)
Figure 4.2. Schematic of a Spindt cathode. Typically, the cone-shaped
emitter is fabricated from molybdenum or heavily doped n-type silicon.
Although emission currents as high as 50 µA per emitter have been
observed, these structures are susceptible to degradation from backsputtering of positive ions.
98
99
SMU
Insulating Spacer
(50-100 µm)
A
Anode or ITO Screen
Sample
(Not to Scale)
Figure 4.3. Cross sectional view of a field emission characterization
apparatus with “parallel plate” geometry. Although this test device is
placed in high vacuum, obtaining the required vacuum between the sample
and the anode or ITO screen may be difficult.
99
100
Rounded Anode
(~1-3 mm dia.)
Variable Distance Anode
(step size = ~0.05 to 3 µm)
SMU
-
e
Sample
A
UHV Chamber
Figure 4.4. Schematic diagram of the experimental setup for a variable
distance anode testing system.
100
101
Hemispherical Energy
Analyzer
ee-
Sample
Sample Holder
Figure 4.5. Schematic diagram for a field emission energy distribution
apparatus. Note that in this diagram, a sharpened tip is used to extract
electrons from the sample. However, in some systems a metal grid is
employed to provide the necessary field for emission.
101
102
Chapter 5
Field Emission Properties Of Nitrogen-Doped Diamond Films
A.T. Sowers, B.L. Ward, S.L English and R.J. Nemanich
(Submitted to the Journal of Applied Physics)
Abstract
This study explores the field emission properties of nitrogen-doped diamond grown by
microwave plasma CVD. Over 70 nitrogen-doped diamond samples were grown on
silicon and molybdenum under varying process conditions. Under certain conditions,
films can be grown which exhibit photoluminescence bands at 1.945 eV and 2.154 eV
that are attributed to single substitutional nitrogen. Field emission characteristics were
measured in ultrahigh vacuum with a variable distance anode technique. For samples
grown with gas phase [N]/[C] ratios less than 10, damage from micro-arcs occurred
during the field emission measurements. Samples grown at higher [N]/[C] content
could be measured prior to an arcing event. Contrary to other reports on nitrogendoped diamond, these measurements indicate relatively high threshold fields
(>100 V/µm) for electron emission. We suggest that the nitrogen in these films is
102
103
compensated by defects. A defect-enhanced electron emission model from these films
is discussed.
5.1
Introduction
Diamond has been considered as a material of choice for next-generation cold
cathode materials. The strong sp3 bonding character in diamond leads to a material
with unique mechanical, chemical, and electrical properties that are particularly
suitable for the harsh environments found in cold cathode applications. In addition,
properly prepared diamond surfaces have been shown to exhibit a negative electron
affinity (NEA) [5.1-5].
A NEA occurs when the vacuum level lies below the
conduction band minimum at the semiconductor/vacuum interface. The presence of a
NEA for a semiconductor means that electrons in the conduction band can be freely
emitted into vacuum without a barrier.
This principle has motivated extensive
research in the development of diamond based cold cathodes for vacuum
microelectronics including flat panel displays and high power microwave amplifiers.
While the ideal cathode material would exhibit a NEA as noted above, other
properties are equally if not more important. In the most basic sense, field emission
from a semiconductor involves the supply of electrons to the material, transport
through the bulk, and finally emission at the surface. To achieve these properties, ntype semiconducting characteristics are desired. Highly doped n-type material will
allow low resistance contacts and provide electrons for transport through the material.
103
104
To date, efforts have been limited by the lack of a reliable shallow n-type dopant for
diamond.
Nitrogen has a high solubility in diamond and is found in both natural and
synthetic diamonds. In synthetic high pressure, high temperature (HPHT) type-Ib
single crystal diamond, nitrogen is present in primarily single substitutional form with
a relatively deep donor level located ~1.7 eV below the conduction band minimum
[5.6,7]. In 1996 Geis et al. reported an enhancement of field emission properties of
single crystal nitrogen-doped diamond [5.8]. In that study, nickel was deposited as the
back contact on nitrogen-doped crystals. It was reported that this metallization created
a Schottky contact forming a narrow depletion layer. In this model the applied voltage
drops primarily across the depletion layer at the nickel/diamond interface. Electrons
tunnel into the diamond conduction band through the narrow depletion region and are
emitted into vacuum at the NEA surface. A simplified band diagram illustrating this
emission mechanism is shown in Figure 5.1. More recently, Geis et al. have reported
another electron emission mechanism from type-Ib single crystal diamond [5.9]. This
model is a surface-emission model in which electrons tunnel from the metal back
contact into surface states at the interface of nitrogen-doped diamond and vacuum.
Okano et al. have reported threshold fields less than 1 V/µm for nitrogendoped diamond films grown by hot-filament CVD using urea as a nitrogen doping
source [5.10]. Using Rutherford backscattering, the nitrogen content in the deposited
films was found to be ~1020 cm-3. However, no evidence of single substitutional
nitrogen doping was presented. It is possible that the nitrogen incorporated into these
104
105
films is segregated at the grain boundaries or in nitrogen aggregates commonly
observed in natural diamond crystals. These urea-doped diamond samples have also
been studied by Matsuda and co-workers [5.11].
Using X-ray Photoelectron
Spectroscopy (XPS), significant amounts of oxygen and tungsten incorporation were
observed (>20 atomic % and ~3 atomic %, respectively). It is not evident what role
these additional impurities may play in the field emission process. In these studies the
threshold fields required for electron emission were higher (from 8.0 to 4.5 V/µm)
than reported by Okano et al. and arcing was observed during some measurements.
Despite these published reports, the mechanisms governing field emission from
nitrogen-doped diamond have not yet been determined.
This study explores the field emission properties of nitrogen-doped diamond
grown by microwave plasma CVD. The reports from Geis et al. discussed above
indicate that electrons may be injected into the diamond conduction band through the
substrate/diamond interface. As a result, the role of this interface was studied by
depositing nitrogen-doped diamond on both silicon and molybdenum substrates using
various nucleation techniques.
Evidence of single substitutional nitrogen
incorporation was confirmed by photoluminescence measurements.
The electron
emission properties of these diamond films were examined with field emission
characterization, and photoemission electron microscopy (PEEM). In contrast to the
reports described above, the results of this study indicate that nitrogen doping
produces diamond films with average threshold fields often exceeding 100 V/µm. In
addition, arcing is frequently observed during field emission characterization causing
105
106
extensive physical damage to the sample. A mechanism describing field emission
from these nitrogen-doped diamond films is also presented.
5.2
Experimental
Nitrogen-doped diamond films were deposited in a commercially available
ASTeX HPMS stainless steel microwave (2.45 GHz) plasma CVD deposition
chamber. In situ growth rate and film thickness information was monitored using
laser reflectance interferometry (LRI). For this technique, a HeNe laser (λ=632.8 nm)
was directed onto the substrate at normal incidence. During growth, interference
between reflections from the vacuum/diamond and diamond/substrate interfaces
causes the reflected light to oscillate as a function of time. By recording the reflected
intensity with a silicon photodiode, the thickness of the film can be measured by
counting the number of interference “fringes.” Details of using the LRI technique to
monitor diamond growth may be found elsewhere [5.12,13].
The conventional gas mixtures of hydrogen and methane were used as the
growth precursors. Two sources of nitrogen were used depending on the desired
nitrogen concentration in the process gas. For low nitrogen process concentrations a
mixture of nitrogen (2.11%) diluted in hydrogen was used.
For high nitrogen
concentrations zero-grade nitrogen (99.998% minimum purity) was directly admitted
to the process gas. With these two sources, nitrogen could be added as an impurity to
the process gas with gas phase atomic nitrogen to carbon ratios ([N]/[C]) from 0 to 80.
106
107
Polycrystalline diamond films containing nitrogen were deposited on 25mm
diameter n-type (1 Ω·cm) (100) silicon or mirror polished molybdenum substrates.
Either diamond grit polishing or biased enhanced nucleation (BEN) was employed to
enhance nucleation. The diamond scratched substrates were hand polished for 10
minutes using 1-2 µm diamond powder applied to a nylon polishing cloth. Before
loading into the chamber, the scratched substrates were cleaned ultrasonically in
acetone and methanol to remove residual diamond powder. For depositions using
BEN, the substrates were loaded into the diamond growth chamber without any
pretreatment. Before BEN, the untreated substrates were exposed to a hydrogen
plasma in the diamond growth chamber for 10 minutes. Nucleation densities of 109
cm-2 or greater have been achieved with these nucleation techniques. This is critical
for the formation of continuous films used in this study.
For both the scratched and BEN samples, diamond nucleation was achieved at
~760ºC surface temperature, 600 W microwave power, 20 Torr chamber pressure, and
at a flow rate of 400 sccm using process gases consisting of 2 vol.% methane in
hydrogen. The samples employing BEN were biased –200 V during the nucleation
phase of growth. Nucleation time was determined by monitoring the reflected LRI
beam for an initial drop in reflectivity indicating sufficient nucleation [5.12]. For both
techniques, the nucleation time was ~21 minutes for most samples. Following the
nucleation step, the substrate temperature, microwave power, chamber pressure, and
process gases were changed to the growth conditions.
107
108
Nitrogen-doped diamond films were grown at substrate temperatures from
800-900ºC, 1300 W microwave power, and 50 Torr chamber pressure. The growth
process gases consisted of 0.5 vol.% methane and 0-12 vol.% nitrogen in hydrogen at
a total flow rate of 500 sccm. All samples were grown ~1 µm thick by monitoring the
LRI oscillations.
After growth, room temperature micro-Raman and photoluminescence spectra
were recorded with an ISA U-1000 scanning double monochromator using the 514.5
nm line of an argon ion laser as the excitation source. The laser beam was focused on
the samples to a spot size of ~3 µm diameter using an Olympus BH-2 microscope.
Low temperature PL measurements were also measured using the same
monochromator in a macro configuration. For these measurements the sample was
mounted to a Janis CCS-350 closed cycle helium refrigerator system. In the macro
configuration the laser was focused to a rectangular spot size approximately
100 µm x 2 mm.
The samples were examined using an Olympus BX60 microscope with
magnifications up to 500x to identify large surface defects and/or damage both before
and after field emission measurements. To evaluate the diamond film morphology and
to distinguish smaller damage resulting from field emission measurements, the
diamond thin films were imaged with a JEOL 6400 field emission scanning electron
microscope (SEM).
After optical characterization, the nitrogen-doped diamond samples were
mounted on molybdenum sample holders and transferred into the loadlock of a
108
109
multichambered ultrahigh vacuum (UHV) system. This UHV system is equipped with
a ~14 m long linear transfer mechanism which interconnects 10 different analysis and
surface processing chambers. The chambers used in this study include hydrogen
plasma processing, XPS, Auger Electron Spectroscopy (AES), and variable distance
anode field emission.
The UHV system has been described in more detail
elsewhere [5.14].
Before field emission measurements were obtained, the diamond samples were
typically exposed to a remote hydrogen plasma to remove adsorbed species and
hydrogen terminate the surface. The hydrogen plasma clean consisted of a 1 minute
exposure to a 50 W RF plasma discharge at ~25 mTorr. The sample temperature was
held at 500ºC during the plasma treatment. This atomic hydrogen exposure has been
shown to induce a NEA on several different diamond surface orientations [5.3]. The
nitrogen-doped diamond films were also examined with XPS and AES to identify the
chemical composition at the sample surface. XPS analysis was performed using
Al Kα (hν=1486.6 eV) radiation and a VG CLAM II electron analyzer. AES spectra
were obtained using a beam voltage of 3 keV and an emission current of ~1 mA using
a Perkin-Elmer cylindrical mirror analyzer CMA.
The field emission measurements employed a variable distance anode, which
was stepped toward the surface and current-voltage (I-V) measurements were obtained
at various anode-to-sample distances. The measurements were conducted in a UHV
environment with pressures typically <5x10-9 Torr.
A cylinder of molybdenum
(3 mm or 1 mm in diameter) was chosen as the anode for these measurements. The
109
110
end of the cylinder was either polished flat or polished to a very high radius of
curvature (typically >5 mm) to minimize edge effects. The anode is mounted on a
stage that is coupled to a UHV stepper motor. The stepper motor controls the distance
between the anode and the sample such that one step of the motor corresponds yields a
translation of the anode by 55 nm.
The I-V measurements are acquired with a
computer controlled Keithley 237 Source Measure Unit (SMU). The SMU has the
ability to simultaneously source a voltage and measure a current. A current limiting
circuit is also included within the SMU so that no voltage is applied that causes the
current to exceed 1x10–8 A.
Although the anodes used in this study were rounded, it is assumed that the
electric field between the anode and the sample surface can be modeled by the parallel
plate geometry. This assumption is valid since the radius of curvature is very large in
comparison to the distance, d, between the anode and the sample (cathode). In the
parallel plate geometry the electric field is given by:
E=
V
d
(1)
where V is the applied voltage between the anode and the cathode. As a result, if a
particular electric field, Ethreshold, is required for field emission from a cathode surface,
then the required applied voltage is given by:
Vthreshold = E threshold d .
(2)
From this expression, it is evident that the voltage required for emission is linearly
dependent upon d, the distance between the anode and the sample (cathode). If the
110
111
anode to sample distance is increased, there is a corresponding increase in the voltage
necessary to induce field emission. On the other hand, if this distance is decreased,
there is an analogous reduction in the voltage required for field emission. This linear
relationship allows the determination of the required threshold field without knowing
the absolute anode to cathode separation.
For any given field emission measurement, a family of I-V curves is recorded
with each curve corresponding to a different anode-to-sample spacing. Initially, the
anode is positioned at some unknown distance above the sample. The stepper motor
count is recorded and an I-V curve is collected. Next, the anode is moved closer to the
sample by a known distance and the cycle is repeated until at least 5-10 curves are
collected. As expected, the I-V curves shift to lower voltage values with decreasing
anode to cathode distance.
Due to the exponential nature of the field emission I-V curves, the “turn-on”
voltage or threshold voltage must be defined in terms of a specific current value. In
this study, the voltage that results in a current value of 0.5 nA was chosen to represent
the threshold voltage for electron emission. In order to prevent micro-arcs for these
nitrogen-doped diamond films, it was necessary to limit the current to 0.5 nA and less.
Each threshold voltage is then plotted versus distance relative to the first I-V curve,
and as expected, the resulting graph was linear. Upon fitting the data to a straight line,
the slope represents the average field for the threshold current emission. This method
for determining the average field does not rely upon the absolute anode to sample
spacing, but rather an accurate measurement of the change in distance of the anode
111
112
with respect to the sample. In addition, this technique has the advantage that the
anode is never in contact with the sample.
The electron emission characteristics of these samples were imaged using
PEEM in a high-resolution system from Elmitec.
In PEEM measurements,
photoelectrons are excited by ultraviolet radiation. These photoelectrons are
accelerated into an electromagnetic immersion lens, which has a lateral resolution of
10 nm. In this lens, 20 kV is applied over a distance of 2 mm resulting in an average
field of 10 V/µm at the sample surface. Once the photoemitted electrons accelerate
through this objective lens, they are magnified and focused by a series of lenses onto a
microchannel plate and phosphor screen. A charge coupled device (CCD) camera
records the image. The spontaneous emission from a free electron laser system was
used as the ultraviolet light source. The light was obtained from the OK-4 UV free
electron laser (FEL) at the Duke University Free Electron Laser Laboratory. The
PEEM technique and applications have been described in greater detail
elsewhere [5.15-16].
5.3
Results
The goal of this work is to investigate the role of nitrogen doping on the field
emission properties of diamond. Previous field emission results from intrinsic and
p-type diamond films have indicated lower threshold fields for samples with higher
defect densities as measured by Raman scattering spectroscopy [5.17]. As a result,
our efforts were to produce high quality nitrogen-doped films, and not to deliberately
112
113
deposit highly defective films. For this reason, diamond films were grown with
0.5 vol.% methane. Undoped diamond films grown under these conditions are known
to result in films of high quality. Despite this objective, as will be seen below, the
addition of nitrogen to the process gas changes the growth chemistry and results in
lower film quality.
In this study over 70 nitrogen-doped films were grown with gas phase [N]/[C]
ratios from 0 to 48. It was observed that the addition of low amounts of nitrogen
initially enhances the growth rate by a factor of ~4.5. However, for [N]/[C] ratios
greater than 0.3 the growth rate decreases with increasing nitrogen addition.
Ultimately for [N]/[C] ratios greater than 48, no deposition is observed, and the silicon
substrates are visibly roughened. The growth rates for the nitrogen-doped diamond
films grown in this study are shown in Figure 5.2.
The Raman spectra for nitrogen-doped diamond films grown on silicon at
~900ºC with various gas phase [N]/[C] ratios are shown in Figure 5.3. For reference,
the spectra for a diamond film grown without nitrogen is also included in the figure.
As stated earlier, the addition of low quantities of nitrogen leads to a decrease in
diamond film quality. In addition to the diamond Raman line at 1332 cm-1, peaks
associated with graphite at ~1350 cm-1 and ~1580 cm-1 are present in the spectra and
become more prominent with increasing nitrogen content in the process gas. Other
peaks from microcrystalline diamond and sp2 bonding in diamond are also visible at
1140 cm-1 and ~1500 cm-1, respectively. At high nitrogen concentrations, peaks (at
1190 cm-1 and 1550 cm-1) possibly attributed to N-C complexes are evident in the
113
114
spectra, however more work needs to be performed to determine their identity [5.18].
In Figure 5.4, the FWHM of the diamond Raman line for several diamond films is
plotted as a function of [N]/[C]. We can conclude from Figures 5.3 and 5.4 that as the
nitrogen content in the process gas increases, the amount of non-diamond carbon in
the deposited film increases and the defect density with in the diamond crystallites
also increases.
For gas phase [N]/[C] ratios from 0.1-1.0 and growth temperatures of ~900ºC,
nitrogen-doped diamond films can be deposited which exhibit PL bands attributed to
nitrogen+vacancy optical centers.
These bands are a characteristic of single
substitutional nitrogen doping in diamond seen in type Ib HTHP synthetic diamond.
The room temperature PL spectra for several nitrogen-doped diamond samples
exhibiting these luminescence features are shown in Figure 5.5. The zero-phonon
lines (ZPL) for these nitrogen-related bands are found at 1.945 eV and 2.154 eV. The
intensity of these nitrogen related centers decreases with increasing nitrogen process
content. A possible explanation for this trend is the quenching of the luminescence in
the presence of a high concentration of defects. This is consistent with the Raman
spectra shown in Figure 5.3.
It should also be noted that nitrogen addition to the process gas also seems to
enhance the 1.680 eV band that has been attributed to silicon incorporation in diamond
films. One possible explanation is that nitrogen containing species in the plasma
increases the silicon etch rate during growth, causing increased silicon incorporation.
A possible unwanted effect of silicon incorporation could be the compensation of the
114
115
single substitutional nitrogen donors. For this reason, molybdenum was also chosen
as a substrate material.
Nitrogen-doped films grown on molybdenum substrates
exhibit no silicon related luminescence.
At higher temperatures (i.e. room temperature), there is significant electronlattice interaction. The effect of this interaction is to reduce the ZPL intensity of the
nitrogen related luminescence while increasing the vibronic sideband intensity.
Conversely at low temperatures, there are less phonons available in the material to
interact with the optical center transitions.
For this reason, photoluminescence
measurements were also recorded at low temperatures for the nitrogen-doped films in
this study. PL spectra recorded at room temperature and 10 K for a nitrogen-doped
diamond film grown on molybdenum are shown in Figure 5.6. The effect of this
electron-lattice interaction is easily distinguished. The presence of these PL bands for
these films indicates that nitrogen is being incorporated into the diamond providing a
donor level located ~1.7 eV below the conduction band minimum.
Figure 5.7 shows the surface morphologies of several diamond films examined
in this study. Without nitrogen addition, the diamond film morphology typifies high
quality diamond growth with well faceted grains. However as seen in the figure, the
addition of nitrogen degrades this faceting and ultimately produces extremely finegrained diamond films at the highest nitrogen process gas concentrations.
After H-plasma cleaning, nitrogen-doped samples were analyzed with XPS and
AES to identify the chemical composition of the films. Despite evidence of nitrogen
incorporation in the PL spectra, only carbon was observed in XPS and AES. For
115
116
reference, the minimum detection limits of XPS and AES are ~0.1 atomic % and
~1.0%, respectively.
It should be stressed at this point that neither oxygen nor
metallic impurities were incorporated during growth.
Initially, the distance variable anode technique was used to characterize the
field emission properties of the nitrogen-doped films using procedures developed for
other wide bandgap materials. Several experiments were performed to evaluate the
effect of nitrogen doping upon the threshold field required for electron field emission.
However, these earliest measurements produced widely varying and often
irreproducible results. In fact, threshold fields for emission often varied between
9 V/µm and 80 V/µm for different regions on the same sample. Three field emission
measurements on a nitrogen-doped film illustrating this instability are shown in
Figure 5.8.
It was only after the diamond surfaces were examined after field emission that
the reason for this unstable behavior was evident. Micrographs of these surfaces
revealed arc-damaged sites similar to features reported by Gröning et al. [5.19]. It was
established that the damage occurred during a computer controlled auto-approach
sequence. Consequently for later measurements, a manual approach technique was
used to place the anode appropriately above the sample surface.
Despite the refinements in the field emission measurements, micro-arcing was
still observed for all films with [N]/[C] ratios less than 10. It should be pointed out
that this behavior was observed regardless of the nucleation technique or substrate
material. For these films, micro-arcing occurred during the first I-V measurement just
116
117
after the current rose above the baseline noise level (±2 × 10-11 A) in the test system.
Micro-arcs are detected by monitoring the I-V curves for large discontinuous jumps in
the measured current. Figure 5.9 shows an I-V curve illustrating an arcing event. If
field emission measurements are continued after micro-arcing, average threshold
fields for electron emission again range from 9-80 V/µm depending upon the
magnitude of the damage to the film and substrate. It is evident that measurements
from arc-damaged surfaces are not indicative of the nitrogen-doped diamond film
properties, but rather from the damaged material and sharp protrusions from the
surface produced by the micro-arc.
For nitrogen-doped films with [N]/[C] ≥ 10, field emission could be observed
without being preceded by an arcing event. Usually for these samples, approximately
five I-V curves could be collected before an arcing event occurred. Analysis of data
collected before arcing, indicate that the average threshold fields required for emission
are 100-300 V/µm. After arcing, the threshold fields are again reduced to 9-80 V/µm.
Figure 5.10 shows the field emission data collected from a sample grown with
[N]/[C]=10. During this experiment an arc occurred at the end of the third I-V
measurement. There is a distinct difference in the I-V curves collected before and
after the arcing event. Both the shape and signal to noise ratio of the I-V curves
change dramatically after the arc indicating a change in the emission mechanism. As
shown in Figure 5.10c, analysis of the I-V data indicate that the threshold field
required for emission before and after the arcing event is 250 V/µm and 73 V/µm,
respectively. Figure 5.11 is an SEM micrograph of the arc-damaged region from these
117
118
measurements. EDS analysis of this region indicate that the irregular material on the
surface is molybdenum from the anode.
To ensure that the arcing behavior observed for the nitrogen-doped films was
not a characteristic of the field emission system, several undoped carbon films with
high sp2 content were examined.
These films were prepared in the same CVD
chamber used for the deposition of the nitrogen-doped films grown for this study. The
process conditions for these films were as follows: 10% methane in hydrogen, 900ºC
substrate temperate, 900 W microwave power, and 20 Torr chamber pressure. The
Raman scattering spectra for one of these films is shown in Figure 5.12.
These samples were loaded directly into the field emission testing system
without any surface pretreatment. Unlike the nitrogen-doped films, these samples
exhibited field emission at exceptionally low fields without arcing.
Figure 5.13
illustrates the field emission data taken from one of the measurements. The analysis
of the I-V curves taken at several anode to cathode distances indicates that the
threshold field required for 0.5 nA is ~4 V/µm. This value has been reproduced many
times across the sample surface, as well as among several identically prepared
samples. The emission mechanism for these high sp2 containing carbon films is
unclear and is a topic of ongoing research. These field emission results indicate that
the arcing behavior, which was observed for nitrogen-doped diamond, is a property of
the films, not the field emission apparatus. Possible mechanisms, which produce the
micro-arcing behavior, will be discussed in the next section.
118
119
Ultraviolet photoemission spectroscopy (UPS) is a method of choice for
identifying the presence of an NEA for p-type diamond surfaces. However, the
nitrogen-doped diamond films in this study are highly insulating, which causes
charging during UPS analysis. UPS measurements are highly sensitive to charging,
which makes this technique unsuitable for the characterization of these nitrogen-doped
diamond films.
Photothreshold measurements can also be used to identify an NEA surface.
The presence of an NEA for diamond surfaces has been associated with hydrogen
termination [5.20]. This hydrogen termination has been shown to be very stable even
after atmospheric exposure. However, hydrogen desorption occurs when diamond
films are annealed in excess ~950ºC [5.21]. These hydrogen free surfaces have been
shown to exhibit positive electron affinities (PEA).
Thus, observation of the
photoyield before and after annealing can indicate the presence of a NEA.
Photothreshold measurements of nitrogen-doped diamond films were
performed using PEEM along with the tunable spontaneous emission from the Duke
University OK-4 ultraviolet free electron laser. In these experiments, nitrogen-doped
films were exposed to a remote hydrogen plasma for one minute before being
transferred ex situ to the PEEM/FEL facility. PEEM images of a nitrogen-doped
diamond film taken with 5.4 eV and 6.0 eV ultraviolet radiation after annealing to
800ºC and 1000ºC are shown in Figure 5.14. For 5.4 eV (hν ≈ Eg) excitation, there is
a distinct decrease in the photoyield after annealing to 1000ºC. These results suggest
that the original hydrogen terminated surfaces exhibited an NEA. Conversely, for
119
120
6.0 eV radiation no apparent change in the photoyield is observed with annealing.
This is consistent since for 6.0 eV excitation hν > Eg + χ for either a positive or
negative affinity surface.
5.4
Discussion
Using optical emission spectroscopy (OES), several groups have reported the
effect of nitrogen addition to H2-CH4 plasmas during diamond growth [5.22-24]. In
these studies, nitrogen addition to the plasma produces two additional bands in the
emission spectra due to CN radicals and N2. This is direct evidence that nitrogen can
be effectively dissociated by the microwave plasma environment.
It has been
proposed that species such as CN or HCN can abstract hydrogen from the diamond
surface and therefore alter the growth process. Cao et al. have suggested that the
growth rate increase at low nitrogen concentrations is due to the increased
incorporation of these species due to lower desorption rates [5.25]. At higher nitrogen
concentrations, CH species, which have been linked to diamond growth, are reduced
in the plasma through the following possible reactions [5.23]:
CH + N → H + CN
(3)
CH + N2 → HCN + N.
(4)
or
The reduction of CH species leads to decreased crystalline quality along with slower
growth rates. Ultimately at the highest nitrogen concentrations, no deposition occurs
due to the efficient etching of carbon by the CN species. These reports are consistent
120
121
with the observations of growth rate and crystalline quality of the nitrogen-doped
diamond films grown in this study.
Despite the increase of structural defects and non-diamond phases observed in
the Raman spectra upon the addition of nitrogen, photoluminescence measurements
indicate single substitutional nitrogen incorporation. This means that these films
possess the same nitrogen donor level observed in type Ib HPHT synthetic diamond
single crystals. For thin diamond films, the concentration of substitutional nitrogen is
difficult to quantify by analytical methods due to the lack of sensitivity for nitrogen,
and the tendency for nitrogen to form aggregate defect centers.
The nitrogen
concentration of homoepitaxial nitrogen-doped diamond films was investigated by
Samlenski et al. using nuclear reaction analysis [5.26]. In that study, the nitrogen
incorporation for the (100) and (111) growth sectors was ~6.8 × 1017 cm-3 and
~2.3 × 1018 cm-3, respectively.
As a result, the single substitutional nitrogen
concentration of the films grown in this study is estimated between 1017 cm-3 and
1018 cm-3. In comparison, the single substitutional nitrogen concentration found in
type Ib HPHT single crystal is ~1019 cm-3.
As stated earlier, field emission from a semiconductor involves (i) the supply
of electrons to the semiconductor, (ii) transport of electrons to the surface, and
finally (iii) the emission from the surface. In general, field emission measurements
will reflect aspects of each of these processes. However, photoemission studies allow
the characterization of emitting surface without being greatly influenced by electron
supply or transport. In photoemission measurements, the electrons are photoexcited
121
122
into the conduction band levels near the surface. By observation of the electrons
emitted from the surface, the electron emission properties of the surface can be
characterized.
Hydrogen desorption/photothreshold experiments taken with the PEEM/FEL
indicate that the nitrogen-doped diamond grown for this study exhibits an NEA when
hydrogen terminated. In addition, the PEEM images show emission from all surfaces
with stronger emission from the edges near the surface. The observed enhanced
emission from the edges is likely due to field enhancement from the applied field in
the microscope.
This type of photoemission was observed uniformly across the
sample. In contrast, field emission from carbon surfaces has been shown to originate
from “hot-spots” randomly distributed across the film.
In contrast to the nitrogen-doped diamond films prepared by Okano et al., the
film examined in this study do not contain oxygen and tungsten. In some films PL
features attributed to silicon incorporation in diamond have been observed.
For
comparison, nitrogen-doped films were grown on mirror polished molybdenum
substrates. For these films, no silicon related luminescence was observed. However,
the field emission properties for the samples grown on either molybdenum or silicon
were the same. This suggests that the silicon incorporation observed for some of the
films does not appreciably influence the emission mechanism.
Although Geis et al. have reported that electrons can be injected into the
diamond conduction band through a Schottky barrier, no evidence of this type of
emission was observed for the nitrogen-doped diamond films examined in this study.
122
123
In the Geis model the applied voltage between the anode and cathode drops primarily
across the Schottky barrier. Consequently, a very small field is produced between the
diamond surface and the anode. Thus, the field emission characteristics are weakly
dependent upon the anode to cathode distance.
In contrast, all field emission
measurements in this study exhibited strong linear dependence upon the anode to
cathode distance. This dependence suggests that the voltage is dropped across the
anode to cathode vacuum gap, rather than the backside contact.
Zhu et al. studied the field emission properties of undoped and p-type diamond
films with varying diamond quality [5.17]. In that study, there was no clear trend
between the content of non-diamond phases in the Raman spectra and the field
emission properties. As a result, it was proposed that the presence of graphitic defects
alone did not account for the emission characteristics of diamond. On the other hand,
a strong correlation between the concentration of structural defects (quantified by the
FWHM of the diamond Raman lineshape) and the threshold field required for field
emission was observed. Specifically, when the FWHM of the diamond peak was
greater than 7 to 11 cm-1 the threshold fields were typically less than 50 V/µm. It was
suggested that the diamond defects create additional energy bands within the bandgap
of diamond and thus contribute electrons for emission at low fields. In contrast to that
work, the field emission characteristics of the nitrogen-doped films in these
experiments do not correlate with the FWHM of the diamond Raman peak.
By studying the interaction of the nitrogen donors and the defects present in
the films, the discrepancies between our field emission results and the reports of Zhu
123
124
and Geis may be explained.
In 1991, J. Mort et al. measured the electrical
conductivity of undoped and nitrogen-doped films deposited by hot filament CVD
[5.27]. In that study, electrical measurements of undoped diamond films indicated the
conductivity was due to transport through in-gap defect states. Furthermore, it was
established that these states were identified as acceptors. For nitrogen-doped films, it
was observed that the electrical conductivity was less than that for undoped diamond
films by a factor of 103 at room temperature and 106 at 400 K. This effect was
attributed to the compensation of the defect acceptor states by nitrogen donors.
The effects of compensation have implications upon the Geis emission model.
If the nitrogen-doped diamond is compensated, then the formation of a Schottky
depletion layer will be inhibited and electron injection into the conduction band will
not occur. This is consistent with the field emission results presented in this study. In
addition, we observed no dependence of the field emission properties upon the
nucleation method or substrate. This suggests that the field emission mechanism is
limited more by the bulk diamond film than the supply of electrons to the film.
We suggest that the defect enhanced field emission model proposed by
Zhu et al. can be extended to include compensation effects observed in this study. At
low nitrogen concentrations, the in-gap defect states are compensated by the nitrogen
donors. Therefore, the density of defect states available for “hopping” conduction is
reduced thus greatly increasing the threshold field required for electron emission. As
the nitrogen concentration in the process gas is increased, more defects in the film are
created. It is assumed that despite the increase in the nitrogen gas concentration, the
124
125
amount of nitrogen incorporated in the diamond film remains relatively constant.
Eventually at the highest nitrogen concentrations, there are enough in-gap defect states
present in the material to overcome the effects of nitrogen compensation. As a result,
defect enhanced electron field emission begins to occur. This emission mechanism is
illustrated in Figure 5.15.
As mentioned previously, field emission is often preceded by micro-arcing for
nitrogen-doped diamond. This behavior is similar to that reported by Gröning et al.
[5.19]. In this study, it was reported that the localized pressure between anode and
cathode increases due to electron stimulated desorption from electrons striking the
anode. Ultimately, a discharge occurs creating craters on the diamond surface. This
model suggests that arcing would be a property of the anode material not the cathode.
However, our experiments have indicated that field emission can be obtained from
carbon films with high sp2 content without arcing. In addition with pressures less than
10-8 Torr, one would expect that eventually all adsorbed species on the anode would
be removed and the arcing behavior would cease. Our experiments have not indicated
a decrease in arcing with time.
As a result of these observations, we suggest that the increase in local pressure
is due to properties of the cathode, not the anode. Field emission studies from
diamond have indicated that electrons are emitted from random emission sites across
the sample. This “spotty” emission coupled with the highly insulating nature of the
nitrogen-doped films can produce large resistive heating on a microscopic scale. We
suggest that this heating causes trapped gases (e.g. hydrogen, nitrogen etc.) in the
125
126
diamond film to be released causing vacuum breakdown within the gap. In this study,
samples with the highest nitrogen process gas concentrations exhibit higher
conductivity and are less susceptible to arcing.
5.5
Conclusions
The objective of this work was to grow high quality nitrogen-doped diamond
films and to determine the role of nitrogen in these films. However, the diamond film
quality is diminished by even small nitrogen concentrations in the process gas. The
field emission properties of films with [N]/[C] gas phase concentrations up to 48 have
been measured. Despite evidence of single substitutional nitrogen doping, detectable
field emission in most cases is preceded by an arcing event that causes damage to the
film and substrate and drastically changes the emission properties. Analysis of I-V
data from the films that exhibit electron emission prior to arcing indicate that
extremely high fields (100-300 V/µm) are required for field emission. It is likely that
the required threshold fields for the samples that exhibit arcing before emission exceed
300 V/µm.
We suggest that the nitrogen donors incorporated into these diamond films are
compensated by in-gap defect states.
By considering compensation effects, the
apparent contradiction between the defect enhanced field emission model proposed by
Zhu et al. and the field emission characteristics of the nitrogen-doped diamond films
in this study can be explained. This compensation reduces the density of in-gap states
and increases the threshold field required for electron emission. It is evident that these
126
127
compensation effects need to be eliminated in order to investigate the properties of
nitrogen in CVD diamond films.
Acknowledgements
We gratefully acknowledge the Duke University Free Electron Laboratory for access
to the OK-4 UV FEL. This project was supported through the Office of Naval
Research.
127
128
References
5.1
F.J. Himpsel, J.A. Knapp, J.A. van Vechten, and D.E. Eastman, Phys. Rev. B 20,
p.624 (1979).
5.2
B.B. Pate, Surf. Sci. 120, p.83 (1986).
5.3
J. van der Weide, Z. Zhang, P.K. Baumann, M.G. Wensell, J. Bernholc, and R.J.
Nemanich, Phys. Rev. B 50, p.5803 (1994).
5.4
J. van der Weide, and R.J. Nemanich, J. Vac. Sci. Technol. B 12, p.2475 (1994).
5.5
P.K. Baumann and R.J. Nemanich, J. Appl. Phys. 83, p.2072 (1998).
5.6
R.G. Farrer, Solid State Commun. 7, p.685 (1969).
5.7
R.G. Farrer and L.A. Vermeulen, J. Phys. C: Solid St. Phys. 5, p.2762 (1972).
5.8
M.W. Geis, J.C. Twichell, J. Macaulay, and K. Okano, Appl. Phys. Lett. 67, p
1328 (1995).
5.9
M.W. Geis, N.N. Efremow, K.E. Krohn, J.C. Twichell, T.M. Lyszczarz, R.
Kalish, J.A. Greer, and M.D. Tabat, Linc. Lab J. 10, p.3 (1997).
5.10 K. Okano, S. Koizumi, S.R.P. Silva, and G.A.J. Amaratunga, Nature 381, p.140
(1996).
5.11 R. Matsuda, K. Okano, and B.B. Pate, Mat. Res. Symp. Proc. 509, p.59 (1998).
5.12 B.R. Stoner, B.E. Williams, S.D. Wolter, K. Nishimura, and J.T. Glass, J.
Mater. Res. 7 p.257 (1992).
5.13 C.H. Wu, W.H. Weber, T.J. Potter, and M.A. Tamor, J. Appl. Phys. 73, p.2977
(1993).
5.14 J. van der Weide, R.J. Nemanich, Phys. Rev. B 49, p.13629 (1994).
5.15 E. Bauer, Inst. Phys. Conf. Ser. 119, p.1 (1991).
5.16 L.H. Veneklasen, Rev. Sci. Instrum. 63, p.5513 (1992).
5.17 W. Zhu, G.P. Kochanski, S. Jin, L. Seibles, D.C. Jacobson, M. McCormack, and
A.E. White, Appl. Phys. Lett. 67, p.1157 (1995).
128
129
5.18 L. Bergman, M.T. McClure, J.T. Glass, and R.J. Nemanich, J. Appl. Phys. 76,
p.3020 (1994).
5.19 O. Gröning, O.M. Küttel, E. Schaller, P. Gröning, and L. Schlapbach, Appl.
Phys. Lett. 69, p.476 (1996).
5.20 J. van der Weide, R.J. Nemanich, Appl. Phys. Lett. 62, p.1878 (1993).
5.21 P.K. Baumann and R.J. Nemanich, Surf. Sci. 409 p.320 (1998).
5.22 R. Locher, C. Wild, N. Herres, D. Behr, and P. Koidl, Appl. Phys. Lett. 65, p.34
(1994).
5.23 T. Vandevelde, M. Nesladek, C. Quaeyhaegens, and L. Stals, Thin Solid Films
290-291, p.143 (1996).
5.24 H. Chatei, J. Bougdira, M. Rémy, P. Alnot, C. Bruch, and J. Krüger, Diam.
Relat. Mater. 6, p.107 (1997).
5.25 G. Cao, J. Schermer, W. van Enckevort, W. Elst, and L. Giling, J. Appl. Phys.
79, p.1357 (1996).
5.26 R. Samlenski, C. Haug, R. Brenn, C. Wild, R. Locher, and P. Koidl, Appl. Phys.
Lett. 67, p.2798 (1995).
5.27 J. Mort, A. Machonkin, and K. Okumura, Appl. Phys. Lett. 59, p.3148 (1991).
129
130
+
+
E
+
+
+
fm
e
-
E
+
+
c
e
+
E
E
fs
E
Type Ib Diamond
vac
d
E
Nickel
-
v
Vacuum
Figure 5.1. Energy band diagram illustrating the field emission mechanism
proposed by Geis et al. In this model, a depletion layer is formed when an
electric field is applied across the diamond. Electrons tunnel from the metal
contact into the conduction band of the diamond, and are emitted at the
NEA surface. This figure is shown for a large applied bias.
130
131
Growth Rate (µm/h)
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0
10
20
30
40
50
[N]/[C] ratio
Figure 5.2. Growth rates from several nitrogen-doped diamond films with
[N]/[C] ratios from 0 to 48. Growth rate information was obtained by
monitoring the LRI oscillations during growth.
131
1190
1550
1332
Intensity (Arbitrary Units)
132
1475
[N]/[C]=10
1140
1350
[N]/[C]=48
[N]/[C]=1.0
[N]/[C]=0.1
× 1/3
[N]/[C]=0
1000
1200
1400
1600
1800
-1
Wavenumber Shift (cm )
Figure 5.3. Raman scattering spectra for several nitrogen-doped films
deposited on silicon at ~900ºC with [N]/[C] ratios from 0 to 48. The
addition of nitrogen to the process gas increases non-diamond bonding in
the films.
132
133
-1
Diamond FWHM (cm )
30
25
20
15
10
5
0
0
10
20
30
40
50
[N]/[C] ratio
Figure 5.4. The FWHM of the diamond Raman line of the spectra in
Figure 5.3 as a function of [N]/[C] in the process gas.
133
Intensity (Arbitrary Units)
134
[N]/[C]=1.0
[N]/[C]=0.8
[N]/[C]=0.6
[N]/[C]=0.4
2.154 eV
1.945 eV
1.5
1.7
1.9
2.1
[N]/[C]=0.2
2.3
2.5
Energy (eV)
Figure 5.5. Photoluminescence spectra recorded at room temperature for
nitrogen-doped diamond films grown on silicon with gas phase [N]/[C]
from 0.2 to 1.0. The presence of the luminescence bands at 1.945 and
2.154eV indicate single substitutional nitrogen incorporation.
134
2.154 eV
1.945 eV
Intensity (Arbitrary Units)
135
1.5
1.7
1.9
2.1
2.3
2. 5
Energy (eV)
Figure 5.6.
Low temperature PL spectra recorded at 10K for a
nitrogen-doped film grown on molybdenum. Low temperatures enhance
the nitrogen related luminescence because electron lattice interactions are
reduced.
135
136
(a) [N]/[C]=0
(b) [N]/[C]=0.1
(c) [N]/[C]=1
(d) [N]/[C]=10
(e) [N]/[C]=48
Figure 5.7. Scanning electron micrographs of the surfaces of several
nitrogen-doped films examined in this study. For reference, an undoped
diamond film is also shown.
136
Voltage for 0.5nA (V)
137
1000
61 V/µm
800
600
82 V/µm
400
36 V/µm
200
0
10
8
6
4
2
0
Relative Distance (µm)
Figure 5.8. Threshold voltage as a function of distance for a nitrogen-doped
film in which arcing occurred during computer controlled auto-approach.
The slope of each line represents the threshold field required for 0.5nA of
emission current.
137
Current (A)
138
10
-8
10
-9
10
-10
10
-11
Arc Occurred
0
200
400
600
800
1000
Applied Voltage (V)
Figure 5.9. Current-voltage curve during a field emission measurement in
which an arcing event occurred. Note the discontinuous jump in the current
at the end of the curve.
138
Current (A)
139
10
-9
10
-10
10
-11
(a)
Current (A)
0
10
-9
10
-10
10
-11
Arc Occured
200
400
600
800
1000
400
600
800
1000
Applied Voltage (V)
(b)
0
200
Applied Voltage (V)
Voltage for
0.5nA (V)
900
800
(c)
700
After Arc:
73 V/µm
600
500
Before Arc:
250 V/µm
400
300 2.5
2
1.5
1
0.5
0
Relative Distance (µm)
Figure 5.10. Field emission measurements from a nitrogen-doped films
grown on molybdenum with [N]/[C] =10 in which an arc occurred during
the measurements. Current-voltage measurements are taken at several
anode-to-sample distances and are shown in (a) and (b). I-V measurements
recorded before the arcing event are shown in (a). During the third I-V
measurement an arc occurred. The I-V measurements taken after the arcing
event are shown in (b). It is evident that the shape and the signal to noise
ratio changes after the arc. The threshold field analysis for this data set is
shown in (c). The threshold fields required to obtain 0.5nA of current
before and after the arc are 250 and 73 V/µm, respectively.
139
140
Figure 5.11. Scanning electron micrograph of the arc-damaged region from
the nitrogen-doped film studied in Figure 5.10.
140
1150
1580
Intensity (Arbitrary Units)
1341
141
1000
1200
1400
1600
1800
-1
Wavenumber Shift (cm )
Figure 5.12. Raman scattering spectra for an undoped carbon film with
high sp2 content.
141
142
Current (A)
10
-9
10
-10
10
-11
(a)
450
500
550
600
650
Applied Voltage (V)
Voltage for 0.5nA (V)
700
(b)
650
600
550
4.5 V/µm
500
450
400
350
40
30
20
10
0
Relative Distance (µm)
Figure 5.13. Field emission measurements of a carbon film with high sp2
content used to characterize the field emission testing system.
Current-voltage measurements are taken at several anode-to-sample
distances and are shown in (a). The threshold field versus relative distance
for this data set is shown in (b). The threshold fields required to obtain
0.5nA of current are ~4 V/µm.
142
143
(a) after 800°C anneal, hν=6.0 eV
(c) after 1000°C anneal, hν=6.0 eV
1 µm
(b) after 800°C anneal, hν=5.4 eV
1 µm
(d) after 1000°C anneal, hν=5.4 eV
1 µm
1 µm
Figure 5.14. PEEM images of a hydrogen terminated nitrogen-doped
diamond film after 800ºC and 1000ºC anneals. Top and bottom images
were obtained using 5.4 eV and 6.0 eV excitation, respectively. The
decrease in photoyield for 5.4 eV radiation after the 1000ºC anneal indicates
that there is a transition between an NEA and PEA surface.
143
144
E
e
-
+
+
f
+
+
E
+
+
c
+
+
+
e
E
E
Silicon
(n-type)
N-doped CV D
Diamond
-
vac
v
Vacuum
Figure 5.15. A schematic energy band diagram illustrating the defect
enhanced field emission model from compensated nitrogen-doped diamond.
144
145
Chapter 6
Summary and Future Research
6.13 Summary of Results
The motivation for investigating the role of nitrogen in the field emission
process was provided by the reports of Geis et al. and Okano et al. [6.1-2]. These
studies indicated that substitutional nitrogen doping lowered the threshold field
required for field emission for single crystal and polycrystalline diamond. Perhaps the
most exciting results of these investigations were the emission models that indicated
that electrons were injected into and subsequently emitted from the conduction band.
However, other research groups have not yet confirmed similar results for nitrogendoped diamond. This investigation was conducted in an attempt to provide a better
understanding of the field emission process for nitrogen-doped diamond.
In the present study, nitrogen-doped diamond films were synthesized by
microwave plasma chemical vapor deposition through the controlled addition of
nitrogen to the process gas. It was observed that the addition of small quantities of
nitrogen during growth produced films, which exhibited photoluminescence bands
attributed to single substitutional nitrogen.
Furthermore, the addition of low
concentrations of nitrogen initially enhances the growth rate by a factor of 4.
145
146
However, for high nitrogen process concentrations, Raman scattering spectra and
scanning electron micrographs indicate a decline in film quality. These observations
are consistent with reports from other groups on the influence of nitrogen doping upon
the growth and properties of diamond films (as cited in Chapter 3).
In order to characterize the field emission properties of these nitrogen-doped
diamond films, a variable distance anode technique was developed.
With this
technique, field emission properties of a sample are investigated using several anode
to cathode distances. In contrast to the reports of Geis et al. and Okano et al.,
relatively high fields (100-300 V/µm) were required for field emission for the films
examined in this investigation. In addition, in most cases detectable field emission is
preceded by an arcing event that causes damage to the film and substrate and
drastically changes the emission properties. This arcing behavior has been observed
for films deposited on either silicon or molybdenum substrates. Furthermore, no
correlation between the field emission characteristics and film thickness was observed.
A compensation/defect field emission model has been proposed for these nitrogendoped diamond films. This model is an extension of the defect-enhanced model
proposed by Zhu et al. [6.3]. In the defect-enhanced model, it is speculated that the
presence of high concentrations of defects in the diamond film creates additional
bands within the diamond bandgap. These defect bands enhance the field emission
properties through hopping conductivity mechanisms.
However, for the
compensation/defect model for nitrogen doped films, it is suggested that these defect
146
147
bands are compensated by the presence of the nitrogen donors. This compensation
effect, in turn, results in higher threshold fields required for field emission.
6.14 Future Research
The present research has suggested that nitrogen-doped diamond grown by
microwave plasma chemical vapor deposition may not be a suitable material for
vacuum microelectronics. However, the use of other nitrogen dopant sources possibly
in conjunction with new deposition techniques may allow the synthesis of nitrogendoped diamond comparable to Type Ib single crystal diamond. Consequently, these
films may exhibit field emission characteristics similar to those reported by Geis et al.
Recently, advances in obtaining n-type diamond have been made using
phosphorus as an n-type dopant [6.4-5]. These reports have indicated that phosphorus
provides a donor level with an activation energy from 0.12 to 0.43 eV. To date,
phosphorus doped diamond has received the most attention for electronic device
applications. However, the development of n-type phosphorus-doped diamond may
facilitate electron emission from the conduction band of diamond using ohmic
contacts.
In order to fully utilize the NEA property of diamond surfaces, emission from
the conduction band will be required.
However, diamond-based films with
technologically feasible field emission characteristics are rapidly being developed.
Often these films contain high concentrations of sp2 bonded carbon and may not
147
148
exhibit a NEA. Although many field emission mechanisms have been proposed, more
research is needed to determine the mechanisms that govern field emission from
diamond-based films. Finally, Okano et al. have suggested that high concentrations of
nitrogen were responsible for the low threshold fields required for field emission [6.2].
Recently, XPS analysis of these films by Matsuda and co-workers has indicated a
significant concentration of tungsten in the diamond films [6.6]. It is uncertain what
influence metallic impurities such as tungsten may have upon the field emission
process.
In fact, it may be suggested that the low threshold fields observed by
Okano et al. can be attributed to the presence of these metallic impurities. Preparation
and characterization of the field emission properties of metal/diamond composites is
suggested as a topic of future research.
148
149
References
6.1
M.W. Geis, J.C. Twichell, J. Macaulay, and K. Okano, Appl. Phys. Lett. 67, p
1328 (1995).
6.2
K. Okano, S. Koizumi, S.R.P. Silva, and G.A.J. Amaratunga, Nature 381, p.140
(1996).
6.3
W. Zhu, G.P. Kochanski, S. Jin, L. Seibles, D.C. Jacobson, M. McCormack, and
A.E. White, Appl. Phys. Lett. 67, p.1157 (1995).
6.4
T. Nishimori, K. Nakano, H. Sakamoto, Y. Takakuwa, and S. Kono, Appl. Phys.
Lett. 71 p.945 (1997).
6.5
S. Koizumi, M. Kamo, Y. Sato, H. Ozaki, and T. Inuzuka, Appl. Phys. Lett. 71
p.1065 (1997).
6.6
R. Matsuda, K. Okano, and B.B. Pate, Mat. Res. Symp. Proc. 509, p.59 (1998).
149
150
Appendices
150
151
Appendix A
Characterization of Highly Oriented Diamond Films
4.7
Introduction
Unless special techniques are employed to enhance nucleation, chemical vapor
deposition (CVD) of diamond on non-diamond substrates generally produces sparse
nuclei with densities less than 105 cm-2 [A.1-3]. The predominant reason for this low
nucleation density is the high surface energy of diamond relative to the substrate
material.
This condition makes it difficult to satisfy the free energy conditions
required for monolayer by monolayer growth.
Consequently, diamond tends to
nucleate and grow three-dimensionally in a Volmer-Weber growth mode [A.2].
This growth behavior leads to the synthesis of continuous polycrystalline
diamond films by growing out three-dimensional nuclei until they coalesce.
Therefore, a high nucleation density is necessary to produce smooth and continuous
thin films. Numerous techniques to enhance the nucleation density of diamond on
various substrates have been developed [A.1-2]. Abrading the substrate surface with
diamond powder prior to growth is probably the simplest and most commonly used
method to enhance diamond nucleation.
These abrasive treatments vary from
ultrasonic agitation in diamond powder slurries to the application of diamond pastes
151
152
with cotton swabs. It is generally accepted that these abrasive treatments promote
diamond nucleation through the formation of high energy defects, which serve as
nucleation sites. In addition, residual diamond particulates may also serve as seed
crystals. Although nucleation densities from 106 to 1010 cm-2 have been obtained,
these abrasive techniques may not be optimal for thermal, electrical, and optical
applications due to contamination and damage to the substrate surface [A.1-2].
Consequently, there has been much research devoted towards developing new
nucleation techniques and potential substrates for heteroepitaxial diamond
growth [A.2].
In 1990, Yugo and co-workers reported a significant increase in the nucleation
density for pristine silicon substrates by negatively biasing the substrate during
microwave plasma CVD [A.4]. This nucleation technique has been termed biased
enhanced nucleation (BEN).
Since this early report, BEN has been extensively
studied and refined to allow “heteroepitaxial” deposition on β-SiC and the synthesis of
highly oriented diamond (HOD) films on (100) silicon substrates [A.5-7]. In this
chapter, the term heteroepitaxy will be used to designate a registry between the
substrate lattice and the diamond film and does not imply that a complete
heteroepitaxial film has been deposited. The synthesis of HOD films on (100) silicon
is a multi-step growth process [A.7-8]. The silicon substrate is initially carburized to
form a SiC interlayer. Following the carburization step, BEN is employed to obtain a
high percentage (up to ~50%) of oriented (100) diamond nuclei. Next, the growth
conditions are chosen which result in the preferential growth of the (100) grains.
152
153
Consequently, a continuous film of slightly misoriented (100) grains is obtained. It
has been suggested that this misorientation is due to the large lattice mismatch
between silicon and diamond (34%) [A.9]. Verwoerd considered the possibility that
heteroepitaxy arises from matching three diamond unit cells to two silicon unit cells
[A.10-11]. Furthermore, Tucker and co-workers calculated that a 9.5° tilt of the
diamond lattice with respect to the silicon lattice is necessary to provide a perfect 3:2
registry of the diamond and silicon planes [A.9]. This calculation is consistent with
high-resolution transmission electron micrographs, which indicate that diamond grains
exhibit an approximate tilt of 9° with respect to the silicon [A.9].
Raman scattering spectroscopy has become a valuable tool for the
characterization of diamond films [A.1,12]. Raman spectroscopy is a quick, nondestructive technique, which can indicate film quality, stress in the film, and the
presence of non-diamond bonding in the film. In the diamond research community, an
often-overlooked property of Raman scattering is the presence of polarization
selection rules. The predominant reason for this is that most diamond films are
polycrystalline in which these polarization selection rules are relaxed due to the
random orientation of the grains. However, for single crystal or oriented films these
selection rules can be used to determine the crystallographic properties of the diamond
films.
In this study, the crystallographic properties of (100) HOD films grown on
silicon were examined using polarization sensitive Raman scattering spectroscopy. Xray diffraction techniques are commonly used to evaluate the crystallinity of diamond
153
154
films [A.13-16]. The purpose of this study was to indicate that typical Raman systems
could be modified at little expense to provide a complimentary technique to these Xray measurements. The polarization selection rules for the Raman active mode of the
diamond lattice were calculated for several low index planes.
Following these
theoretical derivations, these selection rules were verified using single crystal diamond
and silicon samples. Finally, the polarization dependence of several textured and
highly oriented diamond films were examined with this technique. These polarization
dependent Raman measurements indicate that the HOD film is oriented with the
silicon substrate.
Furthermore, these measurements suggest that the diamond
crystallinity improves with increasing film thickness.
These results are in good
agreement with X-ray analysis of similar films [A.13-16].
4.8
Theoretical Background
Each first-order Raman scattering event involves the annihilation of an
incident photon of frequency ωi, the creation of a scattered photon of frequency ωs,
and the creation or annihilation of a phonon of frequency ω. Typically, we choose to
concentrate on Stokes scattering events where phonons are created by the Raman
interactions such that ωi=ωs+ω. The problem of calculating the Raman scattering
efficiency or intensity has been investigated by Loudon [A.17-19]. In this quantum
154
155
mechanical treatment, the Raman scattering intensity was given by the following
expression:
I = A ∑ (e i R j e s ) .
2
(1)
j
The prefactor, A, is a constant of proportionality related to the Raman cross section,
and ei and es are the unit polarization vectors of the incident and scattered photons,
respectively. The Raman scattering tensors, Rj, are related to the symmetry of the
Raman active vibrational modes of the crystal. Using group theory, these tensors have
been tabulated for the different crystal symmetry classes [A.18-19].
In general, the intensity of the Raman scattered radiation depends upon the
directions of the polarization vectors of the incident light and scattered light relative to
the crystallographic axes. Hence, crystallographic information of the sample can be
obtained by analyzing the anisotropy of the Raman scattered radiation. In this study,
this anisotropy was measured by rotating the sample about the optic axis while
keeping the polarization of the incident and scattered light fixed. A disadvantage of
this technique is that the sample must be precisely mounted in the center of the
rotation stage in order to avoid sample translation that can occur during rotation.
A.2.1 (100) Surfaces
Diamond belongs to the space group Oh7 with two atoms per unit cell. The
Raman active mode has F2g symmetry, which is triply degenerate at the zone-center.
155
156
The Raman tensors referred to the cubic axes for the F2g mode are represented by
[A.18-20]:
(100 )
(100 )
(100 )
0 0 0


R x = 0 0 d
0 d 0


0 0 d


R y = 0 0 0
d 0 0


(2)
 0 d 0


R z = d 0 0
 0 0 0


The (100) superscript is used as a label to indicate the (100) surface diamond. To
obtain the appropriate Raman scattering tensors as a function of rotation angle about
the optic axis (x-axis), we must employ the similarity transformation
R ′ = TRT −1
(3)
where T is the transformation matrix which relates the crystal coordinate system to the
laboratory coordinate system. In this case, the transformation matrix required to rotate
through an angle θ about the x-axis is given by:
0
0 
1


T =  0 cos θ sin θ  .
 0 − sin θ cos θ 


(4)
156
157
Applying this transformation to (100)Rx, (100)Ry, and (100)Rz yields:
0
0
0



(100 )
R x (θ) =  0 d sin 2θ d cos 2θ 
 0 d cos 2θ − d sin 2θ 


(100 )
(100 )
d sin θ d cos θ 
 0


R y (θ) =  d sin θ
0
0 
 d cos θ
0
0 

(5)
d cos θ − d sin θ 
 0


R z (θ) =  d cos θ
0
0 .
 − d sin θ
0
0 

In this study, we have chosen to fix the polarization of the incident light
horizontally. Thus, the incident unit polarization vector ei can be expressed as:
e i = (0 1 0) .
(6)
Similarily, if we measure the component of the Raman scattered radiation that is
polarized in the horizontal direction (by using a linear polarizer), we have:
e s = (0 1 0) .
(7)
Using equations (1), (5), (6), and (7) we can calculate the anisotropy of the Raman
scattered radiation as a function of sample rotation angle:
(100 )
I (θ) ∝ d 2 sin 2 2θ .
(8)
A.2.2 (110) Surfaces
The anisotropy of the Raman scattered radiation can also be calculated for
other single crystal orientations. Once the Raman tensors are transformed into the
157
158
proper reference frame, the calculation proceeds analogously as that described above
for the (100) surfaces. To obtain the Raman tensors for the (110) surface, we must
again employ the similarity transformation. That is,
(110 )
R j =T (100 ) R jT −1 ,
(9)
where the transformation matrix T transforms the axes from the (100) to the (110)
orientation. This transformation can be achieved by a 45° rotation about the z-axis.
This transformation is expressed in matrix form by:
 22

T =  − 22

 0
0

0

1
2
2
2
2
0
(10)
When applied to the (100)Rj tensors, this similarity transformation gives
 0

(110 )
Rx =  0
 2
 2 d
2
2
d

d
0 
2
2
2
2
0
0
d
2
 0

0
2 d 

(110 )
− 22 d 
Ry =  0
0
 2
2
0 
 2 d − 2 d
(110 )
(11)
d 0 0


R z =  0 − d 0
 0 0 0


158
159
Now the derivation proceeds exactly as for the (100) surface. Applying the
similarity transformation again to obtain the Raman tensors as a function of the angle
of rotation about the optic axis (x-axis) gives:


(110 )
R x (θ) = 


0
2
2 d sin θ
2
2 d cos θ
d sin θ
2d sin θ cos θ
2
2
( 2
)
2 d cos θ − sin θ

d cos θ

2
2
2
(
)
θ
−
θ
d
cos
si
n

2
− 2d sin θ cos θ 


(110 )
R y (θ) = 


0
2
2 d sin θ
2
2 d cos θ
d sin θ
− 2d sin θ cos θ
2
2
2
2 d (sin θ − cos θ )

d cos θ

2
2
2
(
)
θ
−
θ
d
si
n
co
s

2
2d sin θ cos θ 
(110 )
2
2
2
2
2
2
2
2
(12)
0
0
d



2
R z (θ) =  0 − d cos θ d sin θ cos θ  .
 0 d sin θ cos θ − d sin 2 θ 


Then using equations (1), (6), (7), and (12) we have:
(110 )
I (θ) ∝ 4d 2 sin 2 θ cos 2 θ + d 2 cos 4 θ .
(13)
A.2.3 (111) Surfaces
As a final example, the anisotropy of the Raman scattered radiation for the
(111) diamond surface is calculated below. To obtain the Raman tensors for the (111)
surface, we must again employ the similarity transformation. For this case, we must
utilize the transformation matrix T which transforms the axes from the (110) to the
(111) orientation. This transformation can be achieved by a rotation of θ = cos −1
( )
6
3
159
160
or ~35.3° about the y-axis.
This transformation is given in matrix form by the
following:
 36

T= 0
 3
 3
0 − 33 

1
0 
6 
0
3 
(14)
When applied to the (110)Rj tensors, this similarity transformation gives:
 − 23 d − 66 d

(111)
R x =  − 66 d
0
 2
3
3 d
 6 d
d

d

2
3d 
2
6
3
3
6
2
 − 23 d

6 d
6 d 

(111)
R y =  66 d
0
− 33 d 
 2

3
2
3d 
 6 d − 3 d
 23 d
0

(111)
Rz =  0 − d
 2
 3 d 0
(15)
d

0 

2
3d 
2
3
160
161
Next, we apply the similarity transformation again to obtain these Raman
tensors as a function of the angle of rotation:
(111)
R x (θ) =


 d6 (
d(
6
(111)
− 23d
2 sin θ −
2 sin θ −
d
6
( 2 sin θ − 6 cos θ)
(sin θ + 3 sin θ cos θ)
)
( 3 cos θ + 2 sin θ cos θ −
6 cos θ )
2d
3
6 cos θ
2
2
d
3
d
6
( 2 sin θ −
6 cos θ
)
( 3 cos θ + 2 sin θ cos θ − 3 sin
(cos θ + 3 sin θ cos θ)
3 sin θ )
2
d
3
2
2
2
2d
3


θ )


R y (θ) =


 d6 (
d(
6
− 23d
6 cos θ +
2 cos θ −
d
6
( 6 cos θ + 2 sin θ)
(sin θ + 3 sin θ cos θ)
)
( 3 sin θ + 2 sin θ cos θ −
6 sin θ )
2d
3
2 sin θ
d
3
2
2
 23d

(111)
R z (θ) =  3 3 2 d sin θ
 3 2 d cos θ
 3
(
d
6
( 2 cos θ −
6 sin θ
)
( 3 sin θ + 2 sin θ cos θ − 3 cos
(cos θ + 3 sin θ cos θ)
3 cos θ )
2
d
3
2
2d
3
3 2
3 d sin θ
2
2
d
3 sin θ − 3 cos θ
)
4
3 d sin θ cos θ
2


4
d
sin
θ
cos
θ

3

2
2
d
3 (cos θ − 3 sin θ )
2


θ )


3 2
3 d cos θ
(16)
Finally, using equations (1), (6), (7), and (16) we have:
(111)
I(θ) ∝ d 2 .
(17)
From these derivations for the (100), (110), and (111) surfaces of diamond, it
is evident that by measuring the anisotropy of the Raman scattered light, the
crystallographic orientation of the sample can be identified.
In Figure A.1, the
intensity of the Raman scattered light is plotted as a function of rotation angle for the
(100), (110), and (111) diamond surfaces.
These graphs are plotted in polar
coordinates. The distance from the origin represents the intensity of the Raman
scattered light and the angle is the angle of rotation about the optic axis. It should also
161
162
be noted that these derivations are not limited to single crystal diamond samples. In
fact, these calculations are valid for any crystal with the diamond crystal structure (e.g.
silicon, germanium, etc.).
4.9
Experimental
A schematic of the scanning Raman spectroscopy system employed in this
investigation is illustrated in Figure A.2.
The system is a conventional Raman
scattering system, which was modified for these polarization sensitive measurements.
An adjustable linear polarizer was placed in the beam path immediately adjacent to the
collection lens as shown in Figure A.2.
This polarizer was oriented with the
transmission axis oriented horizontally. Samples under investigation were mounted on
a rotary stage that could be manually rotated 360° about the optic axis.
The system consists of an ISA U1000 scanning double monochromator
equipped with a photomultiplier tube (PMT), which was specifically designed for high
resolution Raman scattering applications. The focal length of the monochromator is 1
meter, and the spectral resolution is 0.15 cm-1. For this study, the 514.5 nm line of an
argon ion laser was used as the excitation source. In this configuration, the laser beam
is focused using a cylindrical lens to a spot size of approximately 2 mm × 100 µm.
For any given polarization sensitive measurement, a family of Raman spectra
is recorded with each spectra corresponding to a different rotation angle of the sample.
These measurements begin with the sample positioned arbitrarily with respect to the
162
163
incident and observation polarization directions. The angle of rotation on the sample
stage is recorded and a Raman spectrum of the sample is initiated. Following this
measurement, the sample is rotated by 5 to 20°, and a corresponding Raman spectrum
is recorded.
This sequence is repeated until the sample stage has been rotated
through 360°. Following data acquisition, the anisotropy of the Raman scattered light
can be examined by plotting the intensity of the Raman line as a function of the angle
of rotation. For each spectrum, the intensity of the Raman line is determined by
subtracting the background signal from the peak height of the Raman line.
In order to establish standards and verify the theoretical calculations shown
above, this characterization technique was initially used to examine single crystal
diamond and silicon samples.
Subsequently, several diamond films grown by
microwave plasma CVD were studied. These films exhibited varying degrees of
crystalline perfection ranging from randomly oriented polycrystalline films to HOD
films.
The synthesis of these diamond films has been discussed in more detail
elsewhere [A.5,7-8].
4.10 Results and Discussion
The first objective of this investigation was the verification of the theoretical
calculations of the anisotropy of the Raman scattered radiation for several low index
planes of the diamond crystal structure. Since silicon substrates of high quality are
163
164
readily available in many crystallographic orientations, these polarization sensitive
measurements were initially performed on (100), (110), and (111) silicon samples.
The angular dependence of the Raman intensity for a (100) silicon sample is
shown in Figure A.3. As can be observed in the figure, the measured intensity of the
Raman line agrees well with the theoretical dependence calculated above. The minor
deviations from the predicted behavior can possibly be attributed to a slight
misalignment of the sample during the measurements. Analogous measurements of
(110) and (111) silicon samples are presented in Figure A.4 and Figure A.5,
respectively.
Although the calculations above indicate that the (111) surface is
isotropic with respect to rotation about the optic axis, the measured dependence of the
Raman line intensity exhibits a slight three-fold symmetry.
The reason for this
deviation is not completely evident, but may be due to the non-normal incidence of
laser. More research is necessary to explain this behavior. This polarization sensitive
technique has also been used to examine (100) and (111) natural diamond specimens.
These measurements (not shown) precisely display the same characteristics as the
silicon surfaces discussed above.
The second objective of this study was to utilize the polarization sensitive
measurements as a crystallographic probe of highly oriented, (100) textured diamond
films grown on (100) silicon substrates. Diamond is transparent to the 514.5 nm laser
used as the excitation source. This property enables the ability to observe Raman
scattering not only from the diamond film, but also from the underlying silicon
164
165
substrate. Consequently, the orientation of the diamond film with respect to the
substrate can be evaluated.
In Figure A.6a, the angular dependence of the Raman intensity for both the
diamond film and silicon substrate (both have been normalized) is shown for a 30 µm
thick (100) HOD film. A SEM micrograph indicating the surface morphology of this
(100) HOD film is shown in Figure A.6b. Although the measured angular dependence
of the diamond Raman line for the HOD film exhibits the expected symmetry for a
(100) surface, significant deviation from the theoretical response for a (100) surface is
evident.
However, closer analysis reveals that the measured intensity can be
accurately modeled by the following expression:
I(θ) = A sin 2 2θ + B ,
(18)
where the theoretical response for a (100) surface has been modified by the addition of
component which is invariant to rotation.
There are several possible effects that may contribute to this invariant
component. It has been shown that the BEN technique can used to obtain up to ~50%
of (100) oriented diamond grains [A.5,7-8]. To obtain a highly oriented diamond film,
growth conditions are chosen which preferentially favor the growth of these (100)
grains. The observed invariant component in the polarization sensitive measurements
could be produced by the polycrystalline component of the film that is not (100)
oriented. Alternatively, a random distribution of slightly misoriented (100) grains
could also contribute to the observed constant component. In addition, scattering of
light by either grain boundaries or surface roughness may also be partially responsible
165
166
for this invariant component. At the present time, it is not possible to differentiate
between these effects.
Consequently, all are combined into the single invariant
term, B. Interestingly, the polarization sensitive measurements of the silicon substrate
also exhibit an invariant angular component. This result was somewhat unexpected
since the angular dependence of the Raman intensity for single crystal silicon has been
shown above to agree well with the theoretically predicted models. This will be
discussed in more detail below.
These polarization sensitive measurements confirm SEM and X-ray diffraction
observations that indicate that HOD films are predominantly (100) oriented
[A.5,7-8,13-16].
Furthermore, comparison between the angular response of the
diamond film and silicon substrate indicates that the diamond film is aligned with
respect to the silicon (100) planes. In order to further quantify these measurements,
we introduce the following quality factor, R:
R=
A
.
A+B
(19)
As the film quality approaches that of an ideal single crystal, this quality factor
approaches unity. On the other hand for polycrystalline films, R approaches zero. The
quality factor, R, for the 30 µm film shown in Figure A.6 was calculated to be 0.7.
In contrast, the angular dependence of the Raman intensity for a highly (100)
textured (but not oriented) film was also measured. As expected for a randomly
oriented polycrystalline film, the Raman scattered intensity was completely
independent of the angle of rotation. Consequently, a quality factor of zero was
166
167
obtained for this film. The angular dependent measurements for this (100) textured
diamond film, along with a SEM micrograph of the surface, are shown in Figure A.7.
The third and final objective of this investigation was the application of the
polarization sensitive Raman scattering technique to characterize the texture and
crystallographic properties of HOD films as a function of film thickness. For this
portion of the study, a thickness series consisting of 1, 4, 15, and 20 µm thick HOD
films were grown under identical process conditions.
SEM micrographs of this thickness series are shown in Figure A.8. These
micrographs illustrate the texture evolution during the growth of a (100) HOD film.
At 1 µm, approximately 50% of the grains on the surface appear to be oriented with
respect to the silicon substrate. After 4 µm of diamond film growth, the percentage of
oriented grains has increased to 85% and greater than 95% by 15 µm. At 20 µm, the
oriented grains have started to coalesce into larger (100) faces, forming low angle
grain boundaries.
The angular dependence of the diamond Raman line for the thickness series
described above is shown in Figure A.9. At 1 µm in film thickness, there is evidence
that the film is partially oriented with the silicon substrate. This observation is in good
agreement with the SEM micrograph, which indicates that ~50% of the grains are
oriented for this film thickness. The quality factor, R, for this film was calculated to
be 0.27. These results suggest that polarization sensitive measurements can be used to
detect partial orientation for diamond grains less than 0.5 µm in diameter. As the film
thickness increases, the crystallographic quality of the film also improves. The quality
167
168
factor, R, increases from 0.39 for the 4 µm film to 0.55 for the 20 µm diamond film.
This trend is illustrated in Figure A.10, where the quality factors for all HOD films
examined in this study are plotted versus film thickness. These measurements indicate
that this Raman scattering technique can be a valuable tool for the measurement of the
orientation and texture of HOD films.
We now return to the issue regarding the angular response of the silicon
substrate for these (100) HOD films. From Figure A.9, it can be observed that for the
exception of the 1 µm thick film, the angular response for the silicon Raman line
exhibits an invariant component. It is well known that CVD diamond films exhibit a
columnar grain structure due to the three dimensional growth habit of diamond.
Scattering from the random distribution of these columnar grains is a possible
explanation for the deviation of the angular response of the silicon Raman line from
the predicted theory. This is consistent with the observations in Figure A.9. For the 1
µm thick film, the columnar texture is not significantly present since the grains have
not begun to coalesce significantly. Consequently, little scattering occurs and the
angular dependence of the silicon Raman line agrees well with theory. However, for
the thicker HOD films, the grains begin to coalesce and thus there is a higher
probability that scattering from the grain boundaries may occur as the Raman scattered
light passes through the diamond film.
From the discussion above, it is suggested that the constant component for the
silicon angular response is due primarily to scattering since the silicon substrate is of
high crystalline quality. Hence, it is expected that scattering from grain boundaries
168
169
produces an equivalent effect upon the response of the diamond Raman intensity.
Thus, the effect of scattering by the diamond grains can be quantified by calculating
quality factors for the silicon response using the same procedure employed for the
diamond response. Consequently, a normalized quality factor can be computed by
taking the ratio of the diamond to silicon quality factors. That is,
Rnorm =
R diamond
.
R silicon
(20)
As a result, the influence of scattering and the effects of misorientation in the diamond
film can be separated. The normalized quality factors for the HOD films examined in
this study are shown in Figure A.10. Since Raman scattering occurs throughout the
film, it may not be subjected to the same scattering effects from the silicon substrate.
Consequently, this normalized quality factor may overestimate the quality of the
diamond film.
4.11 Conclusions
In this investigation, a technique using polarization selection rules for Raman
scattering was developed to characterize the crystallographic properties of highly
oriented (100) diamond films. This technique, termed polarization sensitive Raman
scattering spectroscopy, involves measuring the anisotropy of the Raman scattered
radiation from a crystalline solid. Prior to the characterization of the HOD films, the
Raman selection rules for (100), (110), and (111) surfaces of the diamond crystal
169
170
structure were calculated. These selection rules were subsequently verified by using
this technique to characterize several single crystal diamond and silicon samples.
Polarization sensitive measurements of HOD films indicate that the films
exhibit (100) texture and are predominantly oriented with the silicon substrate. In
addition, this polarization sensitive Raman scattering technique has identified the
presence of (100) texture for HOD films as thin as 1 µm. Furthermore, this technique
has indicated that the crystallographic quality and orientation improves dramatically
under deposition conditions that promote (100) texture. These observations are in
good agreement with prior SEM and X-ray diffraction studies of HOD films.
Furthermore, polarization sensitive Raman scattering spectroscopy has been shown to
be sensitive to changes in the quality of orientation thus making it a complimentary
technique to SEM and X-ray diffraction measurements.
170
171
References
A.1 Diamond: Electronic Properties and Applications, eds. L.S. Pan and D.R.
Kania (Kluwer Academic Publishers, Boston, 1995).
A.2 Handbook of Industrial Diamonds and Diamond Films, eds. M.A. Prelas, G.P.
Popovici, and L.K. Bigelow (Marcel Dekker, New York, NY, 1998).
A.3 G. Popovici and M.A. Prelas, Phys. Stat. Sol. A 132 p.233 (1992).
A.4 S. Yugo, T. Kimura, and T. Muto, Vacuum 41 p.1364 (1990).
A.5 B.R. Stoner and J.T. Glass, Appl. Phys. Lett. 60 p.698 (1992).
A.6 X. Jiang and C.P. Klages, Diamond Relat. Mater. 2 p.1112 (1993).
A.7 S.D. Wolter, B.R. Stoner, J.T. Glass, P.J. Ellis, D.S. Buhaenko, C.E. Jenkins,
and P. Southworth, Appl. Phys. Lett. 62 p.1215 (1993).
A.8 B.R. Stoner, S. Sahaida, J.P. Bade, P. Southworth, and P. Ellis, J. Mater. Res. 8
(1993).
A.9 D.A. Tucker, D.K. Seo, M.H. Wangbo, F.R. Sivazlian, B.R. Stoner, S.P.
Bozeman, A.T. Sowers, R.J. Nemanich, and J.T. Glass, Surf. Sci. 334 p.179
(1995).
A.10 W.S. Verwoerd, Surf. Sci. 304 p.24 (1994).
A.11 W.S Verwoerd, Diamond Relat. Mater. 3 p.457 (1994).
A.12 P.K. Bachmann and D.U. Weichert, Diamond Relat. Mater. 1, p.422 (1992).
A.13 X. Jiang, C.P. Klages, M. Rosler, R. Zachai, M. Hartweg, and H.J. Fusser, Appl.
Phys. A 57 p.483 (1993).
A.14 B.R. Stoner, C.T. Kao, D.M. Malta, and R.C. Glass, Appl. Phys. Lett. 62 p.2347
(1993).
A.15 B.A. Fox, B.R. Stoner, D.M. Malta, P.J. Ellis, R.C. Glass, F.R. Sivazlian,
Diamond Relat. Mater. 3, p.382 (1994).
A.16 R. Kohl, C. Wild, N. Herres, P. Koidl, B.R. Stoner, and J.T. Glass, Appl. Phys.
Lett. 63 p.1792 (1993).
171
172
A.17 R. Loudon, Proc. Roy. Soc. A 275 p.218 (1963)
A.18 R. Loudon, Adv. Phys. 13 p.423 (1964).
A.19 W. Hayes and R. Loudon, Scattering of Light by Crystals (Wiley, New York,
1978).
A.20 S.A. Solin and A.K. Ramdas, Phys. Rev. B 1 p.1687 (1970).
172
173
(100)
(110)
(111)
(100 )
I (θ) ∝ d 2 sin 2 2θ
(110 )
(111)
I(θ) ∝ 4d 2 sin 2 θ cos2 θ + d 2 cos4 θ
I (θ) ∝ d 2
Figure A.1. Theoretical anisotropy for the Raman active mode of the
diamond crystal structure for the (100), (110), and (111) crystal
orientations. For these polar figures, the distance from the origin represents
the intensity for the Raman line, while the angle corresponds to the angle of
rotation of the sample.
173
174
Argon
Ion
Laser
Scanning
Double
Monochromator
Linear
Polarizer
Collection
Lens
PMT
Reflected
Light
Sample
Rotary
Stage
Figure A.2. Schematic of the Raman scattering spectroscopy system used
in this investigation. In this arrangement 514.5 nm light from an argon ion
laser (polarized horizontally) illuminates the sample under investigation.
The angle of incidence of the laser is ~38.7° with respect to the sample
surface normal. A linear polarizer with its transmission axis oriented
horizontally is mounted behind to the collection lens. The anisotropy of the
Raman scattered light is measured by rotating the sample about the optic
axis.
174
175
Figure A.3. The angular dependence of the Raman intensity for a (100)
surface of silicon. The open circles in the figure represent the experimental
data. The dashed line corresponds to the expected theoretical anisotropy for
the (100) surface.
175
176
Figure A.4. The angular dependence of the Raman intensity for a (110)
surface of silicon. The open circles in the figure represent the experimental
data. The dashed line corresponds to the expected theoretical anisotropy for
the (110) surface.
176
177
Figure A.5. The measured angular dependence of the Raman intensity for a
(111) surface of silicon. The open circles in the figure represent the
experimental data. Although the theoretical calculations indicate that the
(111) surface is isotropic with respect to rotation about the axis, the polar
Raman intensity from the graph above illustrates a three-fold symmetry.
177
178
(a)
(b)
Figure A.6. (a) The polarization sensitive Raman scattering measurements
for a 30 µm thick (100) HOD film. The filled and open circles in the figure
correspond to the response from the diamond film and the silicon substrate,
respectively. These measurements indicate that the film is predominantly
(100) textured and oriented with respect to the crystallographic axes of the
silicon substrate. (b) An SEM micrograph taken of the surface indicates
(100) texture of the film.
178
179
(a)
(b)
Figure A.7. (a) The polarization sensitive Raman scattering measurements
for a (100) textured, but not oriented film. The filled circles in the figure
correspond to the response from the diamond film. Although the film is
(100) textured, the random distribution of grains relaxes the polarization
selection rules. This results in the isotropic angular response illustrated
above. (b) An SEM micrograph taken of the surface indicates (100) texture
of the film.
179
180
1 µm
4 µm
15 µm
20 µm
Figure A.8. SEM micrographs of HOD films for varying film thickness.
These films were grown on (100) silicon under the same process conditions.
180
181
1 µm
4 µm
15 µm
20 µm
Figure A.9. Polarization sensitive Raman scattering measurements for the
HOD films shown in Figure A.8. The filled and open circles in the figure
correspond to the response from the diamond film and the silicon substrate,
respectively.
181
182
Quality Factor, R
1
0.8
0.6
0.4
0.2
0
0
5
10
15
20
25
30
35
Film Thickness (µm)
Figure A.10. The quality factor, R, as a function of HOD film thickness.
The quality factor is computed by analyzing the polarization sensitive
Raman scattering data. This graph illustrates the improvement in the
crystallographic properties as the film thickness increases. The solid
diamonds correspond to the quality factor for the diamond films, while the
solid squares refer to the normalized quality factor.
182
183
Appendix B
Thin Films Of Aluminum Nitride And Aluminum Gallium Nitride For Cold
Cathode Applications
A. T. Sowers, J. A. Christman, M. D. Bremser, B. L. Ward, R. F. Davis,
and R. J. Nemanich
(Published in Applied Physics Letters Volume 71, Number 16, October 20, 1997.)
Abstract
Cold cathode structures have been fabricated using AlN and graded AlGaN structures
(deposited on n-type 6H-SiC) as the thin film emitting layer. The cathodes consist of
an aluminum grid layer separated from the nitride layer by a SiO2 layer and etched to
form arrays of either 1, 3, or 5 µm holes through which the emitting nitride surface is
exposed. After fabrication, a hydrogen plasma exposure was employed to activate the
cathodes. Cathode devices with 5µm holes displayed emission for up to 30 min.
before failing. Maximum emission currents ranged from 10-100nA and required grid
voltages ranging from 20-110 V. The grid currents were typically 1 to 104 times the
collector currents.
183
184
The electrical properties of wide bandgap semiconductors in combination with
the high temperature chemical stability make these materials candidates for use in high
power and high frequency devices. Moreover, wide bandgap semiconductors such as
diamond [B.1], AlN [B.2], and AlxGa1-xN for x ≥ .75 [B.3] show promise for use as
cold cathode materials since these materials have been observed to exhibit a negative
electron affinity (NEA).
The presence of a negative electron affinity for wide bandgap semiconductors
means that electrons excited into the conduction band can be freely emitted into the
vacuum.
Prior studies that detected the presence of a NEA for these materials
employed UV photo excitation. Thus, carriers excited into the conduction band near
the surface could escape and be detected. In contrast, for field emission from a NEA
material, if electrons can be supplied to the conduction band then they would be freely
emitted into the vacuum. An ideal wide bandgap semiconductor would then exhibit a
NEA and also sufficient n-type doping to supply electrons into the conduction band
and to form low resistance contacts. To date, it has proved difficult to produce an
n-type wide bandgap NEA semiconductor. While n-type GaN is routinely obtained by
Si doping, n-type characteristics of AlN have not been confirmed.
Much wide bandgap field emission research has been dedicated to depositing
diamond on various field emitting tips [B.4,5] or depositing selectively grown GaN
pyramids on n-type 6H-SiC [B.6,7]. Indeed, some research has been dedicated to
using diamond films for field emission device applications [B.8,9].
Previous
184
185
experiments from our laboratories show that field emission from nitride materials
exhibits Fowler-Nordheim behavior similar to that from diamond [B.10].
Our
approach then is to use a planar nitride surface as the emitter rather than a structure
that deliberately exploits field enhancement at a sharp projection. The advantage of
this approach is that sputtering and other tip degradation processes are avoided. This
approach is similar to recently fabricated diamond cathodes [B.11].
The basic approach employed in this study is presented in the nitride cold
cathode design shown in Figure B.1. An aluminum grid is separated from the nitride
emitting layer by a SiO2 layer. An array of square emission holes is etched through
the aluminum and SiO2 to the nitride layer. At the bottom of the emission hole, the
electron emission occurs at the vacuum-nitride interface induced by the grid
voltage (Vg) between the aluminum pad and the backside contact.
Two different nitride emitting layer structures are employed. In one structure a
thin AlN layer is deposited on a n-type 6H-SiC substrate. In this case, electrons from
the SiC would be extracted through the AlN layer. AlN has a direct bandgap of
6.2 eV while 6H-SiC has an indirect bandgap of 3.0 eV.
The heterojunction
conduction band offset between the AlN and SiC is expected to be between 1.8 to
2.4 eV with the conduction band of the AlN above that of the SiC [B.2]. The second
approach involves an n-type GaN base layer with an AlxGa1-xN graded layer. This
graded layer varies in composition from x=0.05 at the interface with the GaN base
layer to x=0.90 at the emitting surface. Electrons are supplied to the n-type GaN and
185
186
then emitted through the NEA Al rich AlGaN layer. If the grading is completely
smooth then there would be no sharp barriers for the conduction electrons.
Prior to patterning, the nitride emitting layer is deposited on a 6H-SiC substrate
using the metalorganic chemical vapor deposition system described elsewhere
[B.12,13]. The first structure employed a 1000 Å AlN layer deposited directly on the
SiC substrate. The second structure uses a 1000 Å AlN buffer layer deposited on the
SiC substrate. The n+ GaN base layer is grown on this buffer layer and is 1µm thick
doped to n=1 x 1019 cm-3 with silicon. The graded AlxGa1-xN layer is 0.5 µm thick and
is graded from x=0.05 to x=0.90. The surfaces of both nitride emitting layers are
expected to exhibit a NEA [B.2]. As shown by atomic force microscopy, the nitride
films are smooth with root mean square roughness of ~20 Å on a 1 × 1 µm scan.
However, cracking of the top surface was observed by scanning electron microscopy
(SEM) for the graded AlGaN emitting layer. The cracks were ~0.1 µm wide with an
average domain size of 5 µm2 between cracks.
Fabrication of the cold cathodes is accomplished with a two mask process. A
SiO2 layer (~1 µm thick) is deposited on the nitride layer at 400 °C by low pressure
chemical vapor deposition using diethylsilane and oxygen precursors. The grid layer
is formed with 200-300 nm of thermally evaporated aluminum. The first mask step
creates the emission holes which are either 1, 3, or 5 µm squares. Square holes were
used for convenience of the lithography system. The aluminum is patterned with a
standard aluminum etch. Reactive ion etching (with SF6 and O2) is used to etch the
oxide in order to obtain high aspect ratio features. Since the RIE environment may
186
187
damage the emitting surface, a wet oxide etch is used to etch the last ~0.1 µm of oxide
to expose the nitride layer. The second mask step defines the 1 x 2 mm metal pads
that form the grids. The last processing step is thermal evaporation of 200-300 nm of
aluminum onto the backside of the SiC to form an electrical contact.
The electrical testing system has two electrical probes. One probe is used to
make contact to the grid and the other probe is used to collect the emission current
~1 mm above the holes. In this configuration only the total emission current for a
single device is obtained. The grid voltage can be varied from 0-110 V and the
collector voltage from 0-1100 V using two Keithley source measure units (SMU).
The SMU can simultaneously source a voltage and measure the current through the
circuit. The maximum current is limited by the compliance value which was 1 µA for
the collector probe and 10 mA or 100 mA for the grid probe. The pumping system is
oil free and testing is performed at pressures < 5 x 10-7 Torr.
The nitride cathodes were electrically tested directly after processing and no
collector or grid currents above the noise level were measured for the several devices
tested. It was determined that a post processing clean was necessary to activate
emission from the nitride cathodes. Since the aluminum grids are exposed, we are
limited to cleans which will not etch away the grids or damage the nitride layer. The
samples were cleaned ultrasonically in methanol for 10 min. and then subjected to a
remote hydrogen plasma clean at 25 mTorr and 450 °C for 10 min. A hydrogen
plasma exposure with these parameters has been shown to remove hydrocarbons from
AlN and GaN surfaces [B.14].
Also, a hydrogen plasma will remove residual
187
188
photoresist. After plasma exposure, the sample is immediately transported in air to the
electrical testing system.
Both AlN and AlGaN cathode structures exhibited collector currents well
above the current background level.
Only cathodes with 5µm emission holes
displayed emission, and the cathode lifetimes varied from several up to 30 minutes.
For all measurements shown in this paper, the grid and collector currents were
measured at a constant grid voltage while the collector voltage was varied. These
measurements were repeated at different grid voltages for each device. This procedure
was employed to verify that it was indeed the grid voltage that induced the emission.
No significant direct dependence on collector voltage was observed, but the emission
signal varied significantly at different times. At a constant grid voltage, the varying
collector current is attributed to cathode instability rather than to the changing
collector voltage.
Identical structures but without emission holes were also fabricated to test the
SiO2 properties. These test structures are on the same wafer as the cathodes, and
therefore undergo the exact same processing and plasma treatments.
The oxide
breakdown voltage for these structures was found to be >800 V which is much higher
than the grid voltages of 0-110 V used during normal cathode testing. No collector
currents above the noise level have been measured for the structures without emission
holes.
Figure B.2a shows the average current-voltage data for four cathodes that
operated for ~30 min. As expected, the collector current (Ic) increases with increasing
188
189
grid voltage. The collector current and grid current (Ig) were essentially independent
of the collector voltage and depended mainly upon the grid voltage as shown in Figure
B.2b. This AlGaN cathode had the largest collector current of the graded film devices
measured (Ic ≈ 10 nA). Grid currents for the AlGaN devices were 103 to 104 times the
collector current. The electrical measurement was improved and the noise baseline
was reduced for the AlN devices. Similar data was obtained from the AlN cathodes
with the ratio of the grid current to the collector current ranging from 1 - 100.
The cathodes that functioned followed similar patterns during testing. Initially,
the grid current would be high, either 10 or 100 mA, depending on the SMU
compliance value, and no emission was detected at the collector. Then the grid
current would drop and a collector current could be measured. SEM of the cathodes
after testing revealed evidence of melting of the aluminum grid layer around the
emitting holes and at the point of contact between the aluminum pad and the grid
probe. The emission holes would be enlarged and rounded, and sometimes more
severely damaged.
This effect is attributed to either current flowing along the
sidewalls of the emission holes or electrons emitted from the nitride surface and then
collected on the aluminum grid.
We suspect that after the H-plasma clean, a
conducting residue coats the sides of the emission holes. This creates a short between
the grid and the nitride layer. During initial device testing, a high current flows
through the residue which eventually decomposes the residue, and the current may
also melt areas of the grid around the holes. After this decomposition occurs, an
electric field can build up at the nitride surface and electron emission is detected. We
189
190
also suspect that the devices with 1 and 3 µm holes do not function because of residue
that remains after processing.
It is difficult to make a quantitative comparison between the two nitride
emitting layers. This is primarily due to the limited lifetime of the devices. As a
result, data was not obtained from all devices under identical voltage conditions. In
general, the AlGaN devices required lower grid voltages than the AlN devices to begin
operating. However, once emission was achieved, similar collector currents were
obtained for both emitting layers. In addition, the high grid currents obtained preclude
a Fowler-Nordheim analysis of the collected current.
We suggest several changes for improved operation of the nitride cathode
devices. The AlN buffer layer required for high quality growth may be a significant
barrier for the emission particularly from the graded films.
A top contact type
structure in which the electron supply contact is made to the n-type GaN layer would
circumvent this problem. It is also apparent that improved fabrication processes are
necessary. In addition further studies are required on the effect of various surface
treatments on the electron emission from AlN or AlGaN layers.
The authors express their appreciation to the NCSU Microelectronic
Fabrication Laboratory staff for help with the lithography and processing necessary to
fabricate the cathodes.
We acknowledge Bill Partlow of Northrup-Grumman for
additional field emission measurements and for helpful discussions. We also thank
Chris Hatfield and Griff Bilbro for helpful discussions and device simulations. This
190
191
research is supported by the Office of Naval Research and the Northrup-Grumman
Science and Technology Center.
191
192
References
B.1 J. van der Weide, R. J. Nemanich, Phys. Rev. B 49 p.13629 (1994).
B.2 M. C. Benjamin, C. Wang, R. F. Davis, R. J. Nemanich, Appl. Phys. Lett. 64
p.3288 (1994).
B.3 M. C. Benjamin, M. D. Bremser, J. T. W. Weeks, S. W. King, R. F. Davis, R. J.
Nemanich, Appl. Surf. Sci. 104/105 p.455 (1996).
B.4 E. I. Givargizov, J. Vac. Sci. Technol. B 13 p.414 (1995).
B.5 J. Liu, V. V. Zhirnov, G. J. Wojak, A. F. Myers, W. B. Choi, J. J. Hren, S. D.
Wolter, M. T. McClure, B. R. Stoner, J. T. Glass, Appl. Phys. Lett. 65 p.2842
(1994).
B.6 O. H. Nam, M. D. Bremser, B. L. Ward, R. J. Nemanich, R. F. Davis, Mater.
Res. Soc. Proc. 449 p.107 (1997).
B.7 R. D. Underwood, D. Kapolnek, B. P. Keller, S. Keller, S. P. DenBaars, U. K.
Mishra, Topical Workshop on Nitrides, Nagoya, Japan, September (1995).
B.8 N. S. Xu, Y. Tzeng, R. V. Latham, J. Phys. D. 27 p.1988 (1994).
B.9 M. W. Geis, J. C. Twichell, Appl. Phys. Lett. 67 p.1 (1995).
B.10 R. J. Nemanich, M. C. Benjamin, S. W. King, M. D. Bremser, R. F. Davis, B.
Chen, Z. Zhang, and J. Bernholc, Mater. Res. Soc. Proc. 395 p.375 (1996).
B.11 M. W. Geis, J. C. Twichell, N. N. Efremow, K. E. Krohn, C. Marchi, T. M.
Lyszczarz, Proceedings of the 8th International Vacuum Microelectronics
Conference, (unpublished,1995 ). p.277.
B.12 M. D. Bremser, W. G. Perry, N. V. Edwards, T. Zheleva, N. Parikh, D. E.
Aspnes, R. F. Davis, Mater. Res. Soc. Proc. 395 p.195 (1996).
B.13 M. D. Bremser, W. G. Perry, T. Zheleva, N. V. Edwards, O. H. Nam, N. Parikh,
D. E. Aspnes, R. F. Davis, MRS Internet J. Nitride Semicond. Res. 1, p.8 (1996).
B.14 S. W. King, L. L. Smith, J. P. Barnak, J. Ku, J. A. Christman, M. C. Benjamin,
M. D. Bremser, R. J. Nemanich, and R. F. Davis, Mater. Res. Soc. Proc. 395
p.739 (1996).
192
193
Al Grid
Emission
Holes
SiO2
Collector
Electron Paths
Vc
Al Grid
SiO2
Nitride Layer
Vg
n-SiC
Al Contact
AlxGa1-x N, x=0.05 to 0.90
n+ GaN
AlN
n-SiC
Figure B.1. Schematic of the cold cathode structures. The top view (left)
shows a 5 x 5 array of emission holes. The Al grid is 1 x 2 mm, the
emission holes are either 1, 3, or 5 µm, and the hole spacing varies from
5 to 75 µm. The nitride layer is exposed through the holes. The schematic
cross section (bottom) is across the emission holes. The inset shows the
nitride layers used in the graded AlGaN devices.
193
194
a)
Average Collector Current (A)
10-7
10-8
10-9
10-10
10-11
10-12
AlN Cathodes
AlGaN Cathodes
0
10
20
30 40 50 60 70 80 90 100 110
Grid Voltage (V)
b)
Current (A)
10-4
10-5
10-6
10
Grid Current (A)
Collector Current (A)
-7
10-8
10-9
200
300
400
500
Collector Voltage (V)
600
Figure B.2. (a) Current-voltage characteristics for four cathode devices.
The open and closed triangular markers represent data obtained from
devices with the thin AlN emitting layer and the graded AlGaN emitting
layer, respectively. The up and down triangles are associated with different
devices. The electrical measurements were improved for the AlN devices
which lowered the noise baseline. (b) Grid and collector currents obtained
from an AlGaN cathode with Vg = 20 V.
194
195
195