Livro - Field Emission in Vacuum Microelectronics Por Fursey

Field Emission in Vacuum
Microelectronics
MICRODEVICES
Physics and Fabrication Technologies
Series Editors: Ivor Brodie, Consultant

Paul Schwoebel, SRI International and University of New Mexico
Recent volume's in this series:

COMPOUND AND JOSEPHSON HIGH-SPEED DEVICES
Edited by Takahiko Misugi and Akihiro
ELECTRON AND ION OPTICS
Miklos Svilagyi
ELECTRON BEAM TESTING TECHNOLOGY
Edited by John T.L. Thong
FLELD EMISSION IN VACUUM MICROELECTRONICS
George Fursey
ORIENTED CRYSTALLIZATION ON AMORPHOUS SUBSTRATES
E.I. Givargizov
PHYSICS OF HIGH-SPEED TRANSISTORS
Edited by Juras Pozela
THE PHYSICS OF MICRO/NANO-FABRICATION
Ivor Brodie and Julius J. Murray
PHYSICS OF SUBMICRON DEVICES
Ivor Brodie and Julius J. Murray
THE PHYSICS OF SUBMICRON LITHOGRAPHY
Kamil A. Valiev
RAPID THERMAL PROCESSING OF SEMICONDUCTORS
Victor E. Borisenko and Peter J. Hesketh
SEMICONDUCTOR ALLOYS: PHYSICS AND MATERL\L ENGINEERING
An-Ben Chen and Arden Sher
SEMICONDUCTOR DEVICE PHYSICS AND SIMULATION
J.S. Yuan and Peter Rossi
SEMICONDUCTOR LITHOGRAPHY
Wayne M. Moreau
SEMICONDUCTOR MATERIALS
An Introduction to Basic Principles
B.G. Yacobi
SEMICONDUCTOR PHYSICAL ELECTRONICS
Sheng S. Li
A Continuation Order Plan is available for this series. A continuation order will bring delivery of each new volume
immediately upon publication. Volumes are billed only upon actual shipment. For further information please
contact the publisher.
Field Emission in Vacuum
Microelectronics
George Fursey
Vice-Chairman, Russian Academy of Natural Sciences
Moscow, Russia
Edited by
Ivor Brodie
Consultant
Palo Alto, California
Paul Shwoebel
Senior Research Engineer
SRI International
Menlo Park, California
and
Research Professor of Physics
University of New Mexico
Albuquerque, New Mexico
Kluwer Academic / Plenum Publishers

New York, Boston, Dordrecht, London, Moscow
Library of Congress Cataloging-in-Publication Data
Fursey, George.
Field emission in vacuum microelectronics / George Fursey.
p. cm. — (Microdevices)
Includes bibliographical references and index.
ISBN 0-306-47450-6
1. Vacuum microelectronics. 2. Electromagnetic fields.
I. Title. II. Series.
TK7874.9.F87 2003
621.381—dc21 2002034120
ISBN: 0-306-47450-6
© 2005 Kluwer Academic / Plenum Publishers, New York

233 Spring Street, New York, New York 10013
http://www.wkap.nl/
10 9 8 7 6 5 4 3 2 1
A C.I.P. record for this book is available from the Library of Congress
All rights reserved
No part of this book may be reproduced, stored in a retrieval system, or transmitted in any form or by
any means, electronic, mechanical, photocopying, microfilming, recording, or otherwise, without written
permission from the Publisher, with the exception of any material supplied specifically for the purpose of
being entered and executed on a computer system, for exclusive use by the purchaser of the work
Printed in the United States of America
PREFACE
The field electron emission (FEE) is a unique quantum-mechanical effect of electrons tun-
neling from a condensed matter (solid or liquid) into vacuum. The efficiency of this emission
process is tens of millions of times higher than in other known emission processes. The
extremely high current density in FEE and the fact that no energy is consumed by the emis-
sion process proper afford exceptionally wide possibilities for practical application of this
effect.
Currently, the FEE is being infused with new life due to emergence of vacuum microelec-
tronics, a principally new branch of the micro- and nanoelectronics.
Point field emission cathodes (FECs) find novel applications in the fine-resolution electron
microscopy. Auger spectroscopy, atomic-resolution electron holography, and other areas of
superfine diagnostics of atomic surfaces.
The FEE is also a process capable of initiating and sustaining the generation of high-
power and superhigh-power electron beams (thousands and millions of amperes) through
a phenomenon called the explosion electron emission, which is the basis of the modem
high-current electronics.'*^^"^^^
This monograph is devoted to the phenomenon of FEE per se, its governing relationships,
attainable parameters of the field emission process, and its application to problems of
vacuum microelectronics.
The book is intended for researches in various branches of physics, postgraduate and
undergraduate students, specialists in physical electronics (emission electronics, micro-
and nanoelectronics), engineers, and technologists working in the field of vacuum micro-
electronics, specialists interested in electron microscopy and electron-probe systems of high
resolution.
The research in FEE has a long history and a number of excellent books and reviews had
been written on this subject. Among them there is a review by W. P. Dyke and W. W. Dolan
(1956),^'^ monographs by M. I. Elinson and G. F. VasiFev (1958),^^ a book by R. Gomer
(1961),^^ a review by R. Fischer and H. Neumann (1966),^"*^ monographs by A. G. J. Van
Oostrum (1966)^^ and by L. W. Swanson and A. E. Bell (1973),^^ a volume of review arti-
cles on Non-Incandescent Cathodes edited by Elinson,^^' ^^' ^^' ^^^ a book by A. Modinos
(1984),^^ a topical review on vacuum microelectronics by I. Brodie and C. A. Spindt
(1992),^^ a review by I. Brodie and P R. Schwoebel (1994),^^^ a review by G. N. Fursey
(1996),^^^ and others.
vi PREFACE
Lately, new data on the FEE process have been uncovered and further areas of research took
shape. Besides, stimulated by the needs of the vacuum microelectronics, a strong interest
has developed in some features of the field emission process under specific conditions: in
microwave fields, at extreme current densities, and when FECs operate at very short pulse
durations. New aspects emerged in the problem of stability and formation of the surface of
FECs. Very little attention is given to studies of FEE from nanoscale objects (emitters of
just this size are currently used in the vacuum microelectronics). Renewed interest is seen
toward FEE from semiconductors. In the earlier reviews, this information is either lacking
or insufficient. This monograph has the purpose to fill this gap.
Very important information on the properties of field electron emitters is provided by field
emission microscopy and, accordingly, the data accumulated in field emission imaging
experiments. The field emission patterns represent an immediate source of data on varia-
tion of the submicron geometry of an emitting surface, its work function and many other
characteristics. Of great importance is the unique possibility, afforded by the field emis-
sion microscopy, of studying in situ the processes on the surface taking place in superhigh
electric fields of ~10^ V/cm. Much effort has been invested in order to present in this book
a wide range of images for different materials from metals, semiconductors, and carbon
nanoclasters. Such a range of original field emission images of different objects in a wide
range of electric fields is published for the first time.
Special emphasis is made on the limiting properties of thefieldemission process (Chapter 3).
Distinctive features of the FEE in microwavefieldsare considered in Chapter 4. In Chapter 5,
detailed consideration is given to the FEE from semiconductors, and in Chapter 6, data on
FEE statistics are presented. Great attention is paid to analysis of new ideas in the theory of
FEE (Chapters 1 and 3) and to fundamental, crucial experiments providing deeper insight
into the mechanism of the effect under extremely high electric field (Chapter 3).
In Chapter 8, the effect of FEE is treated as a possible basis for designing various devices
of the vacuum microelectronics. Here we attempted to clarify with all possible detail the
problems of a new class of low-voltage displays (Chapter 8) and the use of point FECs in
electron probe systems (Chapter 7). Special attention is given to the possibility of application
in the vacuum microelectronics of self-organizing structures based on carbon clusters, which
have a very low emission threshold, such as diamond-like films, fuUerenes, and nanotubes.
Initially, the book was supposed to cover also studies of thefieldemission from the supercon-
ducting state and polarization effects in emission from surfaces coated with ferromagnetic
films. However, because of insufficient amount of data on these problems, such a review
was considered premature.
The absence of data on the energy distribution of field emission electrons in this book can
be puzzling. This subject has been excellently dealt with in an original work by R. D. Young
(1959),^^ in a paper by R. D. Young and E. W. Muller (1959),^^ in reviews by L. W. Swanson
and A. E. Bell (1973)^^ and A. G. J. Van Oostrum (1966),^^ and in a book by A. Modinos
(1984).'^^ I could not convince myself to just repeat their account of the problem and refer
the reader to the above publications.
This book would not have been written without the help from my colleagues, associates,
and friends, whose assistance in preparing this monograph cannot be overestimated. First
of all, I would like to express my deepest gratitude to the outstanding specialists in the field
of vacuum electronics, Henry Gray and Charles Spindt, for many discussions, hints, and
PREFACE vii
the access to detailed information concerning their published works. My great appreciation
to professor Francis Charbonnier for his interest in the book and the overall support.
Many thanks are due to Prof. Aivor Brodie and Dr. Paul Schwoebel for taking part in the
work on this book, in particular, the valuable additions they made to the book, editing work
and very useful comments on the manner of presentation of the material.
I am very thankful to Prof. E. I. Givargizov and Prof. A. J. Melmed for making available to
me excellent pictures of field ion images of some semiconductors, and Prof. V. N. Shredink
for the original photographs of field emission pictures of nanoscale emitters.
I am also grateful to Prof. L. M. Baskin, Dr. D. V. Glazanov for the fruitful discussion
and, A. D. Andreev, A. N. Saveliev, V. M. Oichenko, and D. V. Novikov for the technical
assistance in preparing the book for publication and to B. N. Kalinin for translating it into
English. My sincere thanks go to S. I. Martynov for help with artistic design of the book
cover layout.
And, finally, I express my most heartfelt thanks to my wife Ludmila Fursey for her patience
and forbearance, for the work she assumed of listening to and correcting many fragments
of this writing.
HISTORICAL OVERVIEW
The phenomenon of the electron field emission was discovered by R. W. Wood in 1897.^
Initial theoretical insight into this process was provided by W. Schottky (1923),^ who
assumed that the electrons are emitted over a potential barrier at the surface lowered by
the applied electric field. It was subsequently shown experimentally that fields capable of
initiating emission are in fact 10-50 times lower than this value.
R. A. Millikan and C. F. Eyring^ and B. S. Gossling"* discovered (1926) that emission
currents are not affected by temperatures up to 1500 K. Soon afterwards R. A. Millikan,
and C. C. Lauritsen^ found that the emission current varied exponentially with the applied
electric potential. In 1928, R. H. Fowler and L. W. Nordheim developed a theory of field
emission based on quantum-mechanical tunneling of electrons through the surface poten-
tial barrier.^' ^ This theory accurately described the dependence of the emission current on
the electric field and the work function. It also followed from this theory that no external
excitation is required for the initiation of this process (as distinct from thermal and photo
emission). Clear evidence to this was obtained by J. E. Henderson et al. in their studies of
the energy distribution of electrons^"^^ and measurements of the calorimetric effect.^^' ^^
A more rigorous study of the energy distribution was later undertaken by E. W. Miiller,^^' ^"^
R. D. Young,^^ A. G. J. Van Oostrum^'^ and L. W. Swanson.^^' ^^ An important develop-
ment in the study of field emission was the invention of the field emission microscope by
E. W. Miiller in 1936.^^' ^^ It was, in fact, at about this time that the systematic accumulation
of data began on the surface properties of field emitters. With the field emission microscope,
it was possible to investigate many phenomena, such as factors causing instabilities of the
field emission process and the way in which the emitter tip is affected by the applied electric
field, temperature, adsorption of foreign atoms, and electron and ion bombardment. The
high magnification (10^-10^ times) and resolution (10-30 A) of the field emission micro-
scope, in combination with the possibility of actively influencing the object under study
in situ, made it an indispensable tool for studying adsorption, desorption, epitaxy, surface
diffusion, phase transitions, etc. (see reviews).^"*' ^9,22-29
One of the most important results of the quantum-mechanical theory was the prediction of
extremely high field emission current densities, far in excess of those possible with thermal
electron emission. In 1940, R. Haefer^^ made use of the transmission electron microscope
in a quantitative study of emitter's shape and emitting area and experimentally proved the
X fflSTORICAL OVERVIEW
feasibility of achieving high current densities, by obtaining stationary current densities of

j - 10^ A/cm^.
The beginning of basic research in the high current range and the first achievements in the
practical use offieldemission are associated with research by W. P. Dyke and his group.^^"^^
Using high-vacuum equipment, pulsed techniques, and state-of-the-art technologies, these
researchers were able to study electron emission in higher fields and current densities
than those explored previously. Current densities of 10^-10^ A/cm^ were achieved under
pulsed and steady-state conditions, respectively.^ ^' ^^ W. P. Dyke and J. K. Trolan^ ^ observed
appreciable deviations from the Fowler-Nordheim theory in the high-field region. These
deviations manifested themselves as a subexponential dependence of the field emission
current on the applied potential. They ascribed this observation to space charge induced
lowering of the electric field strength near the emitter surface. J. P. Barbour et al.^"^ provided
a more detailed theoretical and experimental validation of this interpretation. T. J. Lewis^^
suggested that the observed departure from the theory could be related to the fact that the
shape of potential barrier differs from that assumed in the image force theory. This difference
becomes greater in the high-field region where barrier dimensions are of the same order as
the interatomic separations, an idea that has been elaborated upon further.^^
A number of important results obtained by the Dyke's group relate to the causes of insta-
bilities and degradation of field emitters.^ ^"^^ It appeared that the main cause of the emitter
degradation was Joule heating of the tip by the emission current.^^' ^^ In the case of clean
and smooth emitter surfaces, the instability develops throughout the emitting tip. This heat-
ing can also occur locally when isolated micro-nonuniformities develop on the emitter as a
result of, for example, ion bombardment of the cathode. By eliminating factors leading to
the formation of micro-nonuniformities, stable operation of the field emission cathode in
the continuous mode for more than 7500 hr was demonstrated.^^"^^ In these experiments,
current densities were close to the maximum achievable values of ~10^ A/cm^).
M. I. Elinson and co-workers"*^^"^ found that the maximum current values were dependent
on emitter geometry and showed that by increasing the tip cone angle, the current density
could be increased by about an order of magnitude without emitter tip damage. Researchers
in this group outlined a program to search for suitable materials for the field emission
cathodes and studied materials based on metal-like and semiconductor compounds, such
as LaB6'*^~^ and ZrC.^^' ^ A number of investigations in the high current density region
have also been performed by G. N. Shuppe's group."^^' "^^
Further progress in research on field emission at extremely high current densities has been
conducted by G. N. Fursey et al. Improvement in the experimental techniques employed
made it possible to increase sensitivity of the pulsed measurements by 5-7 orders of
magnitude."^^' ^^ This allowed for thefirstmeasurements to be made of the maximum current
densities from localized emission areas on the emitter tip crystal. The pulsed measurements
were extended to the range of 10~^-10~^ sec.^^"^"* Experiments under quasi-steady-state
conditions were performed in the range of 10~^-10 sec. In direct current experiments, ther-
mal effects due to field emission at high current densities were demonstrated.^^' ^^ A new
type of instability caused by the spontaneous change of the cathode surface microgeometry
near the thermal destruction threshold^ ^'^^ was discovered. In studies by G. A. Mesyats
and G. N. Fursey, current densities of 10^ A/cm^ were observed with nanosecond range
pulse lengths.^^'^^ Current densities up to 5 x 10^ A/cm^ were demonstrated for field
emission localized to nanometer-scale emitting areas.^^ In experiments by V. N. Shrednik
HISTORICAL OVERVIEW xi
et al., current densities up to (10^-10^^) A/cm^ are recorded from nanometer-sized tips
under steady-state conditions.^^ Recently, G. N. Fursey and D. V. Galazanov, using tips
with an apex radius of '^lO A, were able to reach current densities of 10^^-10^^ A/cxx^P
These current densities are close to the theoretical supply limit of a metal's conduction band
when the electron tunneling probability is unity.
Experimental and theoretical studies have been conducted in order to determine a method
of increasing the stability of the field emission current and preventing ion bombardment of
the cathode.^^' ^^' ^^ In fairly recent studies by our group,^^' ^^ it was shown that applied
microwave fields could reduce the intensity of cathode ion bombardment by several orders
of magnitude due to a repulsive potential for ions near the surface.
The practical application of field emission began with the founding of the Field Emission
Corporation by W. P. Dyke in 1959. This company produced unique commercial instruments
such as pulsed X-ray apparatus for recording high-speed processes^^' ^^ and compact X-
ray sources for medicine.^^ Subsequently, it was discovered^^ that the phenomenon taking
place in these instruments was not only conventionalfieldemission, but a related phenomena
referred to as explosive emission.^^"^^
In the 1960s, A. V. Crewe and co-workers^^"^"^ demonstrated the possibility of using
field emitters as an electron source for atomic-scale resolution electron microscopy. The
field emitter became a promising electron source for Auger spectroscopy,^^'^^ X-ray
microscopes,^^ and semiconductor lithography.^^'^^ Quite recently, reports have been
published on the atomic-scale electron holography using field emitters.^^^^
Of particular interest is the use of field emitters in microwave devices.^^' ^^' ^^ A series of
unique microwave devices has been proposed based on field emission.^^' ^^-^^^' ^^^
K. R. Shoulders^^^ suggested using arrays of field emission cathodes in various microelec-
tronic components and devices. Research along this avenue was initiated in the United States
by C. A. Spindt.^^' ^^^ Interest in this area has grown rapidly over the last 10 years in the
USA,^^"^ France,^^^ Japan,"^^ and Russia. ^^^ The most striking example of the application
of field emission cathodes to an area of technological interest is in flat-panel displays^^^"^^^
that are remarkable in their potential for high brightness, resolution, and low price when
compared to existing displays.
In recent years, new ideas and experimental techniques and approaches have been devel-
oped to address observations that could not be reconciled with previous theoretical models.
Examples are the energy spectra of the field emitted electrons, which show broadening
of the energy spectrum at high current densities,^^^ the presence of high- and low-energy
tails,^^^' ^^^ and additional peaks thought to be due to the bulk band structure and surface
states (see p. 223).^^ In some studies,^^^ additional peaks have been observed using very
small, atomically sharp emitters.
As a consequence of the potential impact of vacuum electronics in the field of nano-
electronics, there is a need for insight into the physics of field emission when emitter
dimensions are equal to or smaller than the width of the surface potential barrier.^^"^"^^^
Again, the inadequacy of the one-particle approximation of the field emission is apparent
in this case.^^^' ^^^' ^^^ Several related studies have been conducted on a variety of top-
ics. F. I. Itskovich^^^' ^^^ indicated the possible effect of the Fermi surface structure on
field emission. Also,^^' 1^2-135 adiabaticity has been considered with electron tunneling in
high-frequency alternating fields. For field emission from metallic electron emitters coated
with ferromagnetic films, polarized electrons have been detected (19: pp. 256-258, 29:
xii HISTORICAL OVERVIEW
pp. 155-164).^^^^'^'*' ^"^^ A number of researchers have also studied statistical processes
in field emission^^^ and found that in some cases many-particle tunneling occurred with
oxidized surfaces and from high-temperature super conducting materials. ^"^^
Lastly considerable advances have been made in research on field emission from semicon-
ductors (Chapter 8, pp. 149, ISO-lSl).-^^ Recently, interest in field emission from diamond
and diamond-like films and fuUerenes has grown dramatically.^^^"^^^' '*^^' ^^' ^^^
CONTENTS
1 Field Electron Emission from Metals 1

1.1 Fowler-Nordheim theory 1
1.2 Thermal-field emission 4
1.3 Extending the theory of field electron emission from metals 6
1.3.1 Deviations from the Fowler-Nordheim Theory and Peculiarities of
Field Electron Emission from Small-Scale Objects 6
1.3.2 The Effect of Fermi Surface Structure 12
1.3.3 Many-Particle Effects 13
1.4 Resume 17
2 Characteristics of Field Emission at Very High Current Densities 19

2.1 Deviations from the Fowler-Nordheim theory in very high electric fields .... 19
2.2 Space charge effects in field emission 19
2.3 Space charge with relativistic electrons 22
2.4 The shape of the potential barrier in high electric fields 26
2.5 Resume 30
3 Maximum Field Emission Current Densities 31

3.1 The theoretical limit of field emission current 31
3.2 Effects preceding field emitter explosion 32
3.2.1 Dyke's Experiments 33
3.2.2 Pre-Breakdown Phenomena 34
3.2.3 Time-Dependent Observations 35
3.2.4 Quantifying Local Effects 36
3.3 Heating as the cause of field emission cathode instabilities 36
3.3.1 Experimental Demonstration of Field Emitter Heating 37
3.3.2 Analysis of Thermal Processes 38
3.3.3 Three-Dimensional Analysis of Emitter Heating 44
3.4 Build-up of the field emitter surface at high current densities: thermal field
surface self-diffusion 47
3.5 The highest field emission current densities achieved experimentally 49
3.5.1 Experimental Current Density Values 50
xiii
V CONTENTS
3.5.2 The Limiting Current Densities Attained Cooling 50

3.5.3 Current Densities from Nanometer-Scale Field Emitters 54
3.6 Resume 56
Field Emission in Microwave Field 57

4.1 Introduction 57
4.2 The adiabatic condition—tunneling time 58
4.3 Experimental verification of Fowler-Nordheim theory 60
4.4 Maximum field emission current densities 62
4.5 Ion bombardment of the cathode 63
4.6 Electron energy spectra: Transit time 64
4.7 Field emission from liquid surfaces 67
4.8 Resume 69
Field Emission from Semiconductors 71

5.1 Introduction 71
5.2 Emitter surface cleaning 72
5.3 Current-voltage characteristics 75
5.4 On preserving the initial surface properties of a field emitter 77
5.5 Voltage drop across the sample and the field distribution in the emitting area 78
5.6 Theory of the field electron emission from semiconductors 81
5.6.1 Stratton's Theory 82
5.6.2 Modeling a Semiconductor Emitter: Statement of the Problem 83
5.6.3 Zero Current Approximation 87
5.6.4 The "Nonzero Current" Approximation with p-Type Semiconductors 88
5.6.5 The Effect of High Internal Fields 90
5.6.6 Field Emission from n-Type Semiconductors 90
5.6.7 Results of Numerical Calculations 91
5.7 Transition processes in field emission from semiconductors 97
5.7.1 Time Variation of the Field Electron Emission 98
5.8 The stable semiconductor field emission cathode 99
5.9 Adsorption on semiconductor surfaces 101
5.9.1 Oxygen Adsorption on Ge 101
5.9.2 Electroadsorption 101
5.10 Resume 103
Statistics of Field Electron Emission 105

6.1 Formulation of the problem 105
6.2 Method of investigation 106
6.3 Field emission statistics from metals 108
6.4 Field emission statistics at cryogenic temperatures 109
6.5 Multielectron field emission from high temperature superconducting
ceramics Ill
6.6 Investigations of field emission statistics for highly transparent barriers 113
6.7 Resume 114
CONTENTS XV
7 The Use of Field Emission Cathodes in Electron Optical Systems:

Emission Localization to Small SoUd Angles 115
7.1 Introduction 115
7.2 The optimum crystallographic orientation of the field emission cathode 116
7.3 Field emission localization by thermal-field surface self-diffusion 117
7.3.1 Localization by Build-Up Processes 117
7.3.2 Localization by Micro-Protrusion Growth 118
7.3.3 Large Micro-Protrusions 119
7.3.4 Fine Micro-Protrusions 122
7.4 Field emission localization by a local work function decrease 124
7.4.1 Basic Principles of Localization by Decreasing the Work Function... 124
7.4.2 ZronW 125
7.4.3 The Influence of the Vacuum Conditions 125
7.4.4 The Dispenser Cathode Approach 128
7.4.5 Stabilization by Oxygen Treatment 128
7.4.6 Shaping of the ZrAV Emission Sites in an Applied Electric Field 129
7.5 Field emission from atomically sharp protuberances 131
7.6 Applications of field emission cathodes in electron optical devices 131
7.7 Resume 136
8 Advances in Applications 137

8.1 Introduction 137
8.2 Short historical review and main development stages 138
8.3 Field electron microscopy 140
8.4 Field emission displays 144
8.4.1 Requirements to the Tips and Lifetime of Matrix Field Emission
Cathodes 148
8.4.2 Comparative Characteristics of the Display Types: Advantages of
Field Emission Displays 150
8.5 Other applications of field emission 152
8.5.1 Miniature Field Emission Triodes and Amplifiers Based on Field
Emission Arrays (FEAs) 153
8.5.2 Microwave Devices for Millimeter Wave Amplification 155
8.5.3 Nanolithography 158
8.5.4 Vacuum Magnetic Sensors 158
8.5.5 External Pressure Sensors 159
8.5.6 Mass-Spectrometers with Field Emission Cathodes 159
8.5.7 Use of Field Emission in Gas Lasers 160
8.6 Arrays made of carbon nanoclusters 161
8.7 Resume 169
References 171
List of Main Notation 193
Index 197
1
FIELD ELECTRON EMISSION FROM METALS
1.1. FOWLER-NORDHEIM THEORY
The field emission (FE) process is a unique type of electron emission as it is due
exclusively to quantum-mechanical effects—tunneling of electrons into vacuum. This phe-
nomenon occurs in high electric fields. This phenomenon occurs in high electric fields
10^-10^ V/cm. In order to produce such high fields using reasonable potentials the emitter
is usually formed into a tip with the apex radius of curvature ranging from tens of angstroms
to several microns. The high electric field narrows the potential barrier at the metal-vacuum
interface sufficiently for the electrons to have a significant probability of tunneling from the
solid into the vacuum.
The quantitative description of the process, as given by the Fowler-Nordheim (FN)
theory,^' ^' ^^^ provides an adequate picture. Quantifying the FE process involves calculating
the FE current density as a function of the electric field. To do this the probability for the
electron to tunnel through the potential barrier must be determined, that is, the barrier
transparency, D, and the flow, N, of electrons incident on the barrier from within the metal
must be calculated and then integrated over the electron energy on which both D and A^ are
dependent. Usually this is the portion of the electron energy whose velocity component is
normal to the barrier.
FN theory is based on the following main assumptions:
1. The metal is assumed to obey the Sonmierfeld free electron model with Fermi-
Dirac statistics.
2. The metal surface is taken to be planar, that is, the one-dimensional problem is
considered. This assumption is accurate because in most cases the thickness of
the potential barrier in fields of 10^-10^ V/cm is several orders of magnitude less
than the emitter radius. Thus, the external field can be taken to be uniform along
the surface.
3. The potential Ui(x) within the metal is considered constant iUi(x) = const =
—Uo). Outside the metal the potential barrier is regarded as entirely due to the
image forces Ux = —e^/4x with the externally applied electric field having no
effect on the electron states inside the metal.
4. The calculation is performed for the temperature 7 = 0 K.
FIELD EMISSION IN VACUUM MICROELECTRONICS
Vacuum level
Fermi level
Band edge
5 10
Metal Vacuum
Figure 1.1. The potential energy of electron Uix) (in eV) as function of the distance x (in A) from the metal
surface. —e^/4x is the image force potential; —eFx is external apphed potential; U{x) is total potential; Up is
total potential well depth in the metal; cp is work function; F is electric field strength.
Under these assumptions, the current density is given by the equation

/»00
j =e n (E,)D(E,,F)dE,, (1.1)
Jo
where e is the electron charge, n(Ex) is the number of electrons per second having ener-
gies between Ex and Ex 4- dEx, incident on 1 cm^ of the barrier surface from within the
metal; Ex = pi/2m is the part of the electron kinetic energy carried by the momentum
component px normal to the surface, m is free electron rest mass, and F is applied electric
field.
The barrier transparency is calculated using the semiclassical method of Wentzel-
Kramers-Brillouin (WKB) approximation.^^^' ^^^ With an applied electric field F the
following potential function
^2
U(x) = eFx (1.2)
Ax
describes the barrier (Fig. 1.1).
The calculations show that for such a potential barrier the transparency is given by
87r(2m)i/2l \Ex\^^^
D(Ex,F) =:exp - -^(y), (1.3)
She
where T^(y) is the Nordheim function.
i^(y) = l-^l\\ + (1 - >;2)l/2]l/2 . [^(^) - {1 - (1 - /)l/2}]/^(^), (1.4)

having for an argument

(^3^)1/2
y= (1.5)
and
rn/l nn/l
da
E(k) K(k) = / {\-k^smâ)^f'^doi (1.6)
Jo
are complete elliptic integrals of the first and second kinds, with
^,^ 2(1-^^/2
l + (l_^2)l/2-
Using (1.3), the FE current density at T = 0 follows the classic FN formula

.3 T72 r .rrfi/2
J = -J-— exp -6.83 • \0^^^{y) , (1.7)
%7th
where cp is the work function. Substituting the values of constants and expressing cp in eV,
F in V/cm, and j in A/cm^ we have,
j = 1.54 . 1 0 - ^ - 2 - — -6.83 . lO^îîy) (1.8)

t^(y)(p I F
where y = 3.79 • IQ-^^. ^/F/(p, t(y) = i}(y) - (2y/3)(di^(y)/dy).

1^ (y) and t (y) had been tabulated^^^ and can be found in a number of works.^^' ^^^ t (y)
in the preexponential factor is close to unity and varies weakly with the argument. In many
cases it is justifiably set to unity. The Nordheim function t^(y) varies significantly with y
and, correspondingly, so does F (Fig. 1.2).
Formulas (1.7) and (1.8) give an excellent description of the experimentally observed
exponential dependence of the emission current on field strength F and work function (p. In
the so-called FN coordinates the functional dependence In 7 = / (1 / F) or, correspondingly,
(a)
1.0
^ 0.5
0.5 1.0
y y
Figure 1.2. The behavior of i^,t, and 5 as a function of >'.

4 FIELD EMISSION IN VACUUM MICROELECTRONICS
\nl = /(!/(/)isastraightlineoverawiderangeofemissioncurrentvalues. Here, / = j-S

is the emission current, with S as the emitting area and F = PU, where ^ is a geometric
quotient determined by the geometry of the vacuum gap.
The Nordheim function i^(y), as noted earUer, is strongly dependent on y (Fig. 1.2),
however it does not significantly affect the linear behavior of the current-voltage character-
istic because i^{y) is very close to a parabolic curve of the form ?> = 1 — by^ (b = const).
In this approximation the additional dependence on F through i^(y) is transferred to the
preexponential factor and has no effect on d(\nj)/d(l/F). The correction to the slope of
the so-called FN characteristic \n(j/F^) = f{\/F) is given by the expression
5(y) = ^ ( y ) - f ^ , (1.9)
2 ay
which, as with t{y), is very close to unity (Fig. 1.2). Then,

^Mp._2.98.10V^5(y) (1.10)
d{\/F)
with (f in eV, F in V/cm, j in A/cm^.
1.2. THERMAL-FIELD EMISSION
At nonzero temperatures the energy spectrum of electrons in a metal will contain

electrons at energies above the Fermi level. These electrons begin to contribute appreciably
into the emission current. Emission of this sort is referred to as thermal-field (T-F) emission.
In this case the limit for the Fermi functions f(Ex, 0) = 1 in formula (1.1) can no longer
be used and expression (1.1) should be integrated using the general Fermi function
f(Ex,T) = ^- (1.11)
Qxp((E,-E^)/kT)-hl
where £ F is the Fermi energy and k the Boltzmann constant.
Calculations of the T-F emission current density have been made by a number of
authors.^^^"^^^ A fairly simple analytic expression can be obtained only for rather low
temperatures,^^^ at T < 1000 K. At higher temperatures the analytic solution is possible
only over limited temperature intervals and the expressions are cumbersome.
The ratio of j (F, T) to that at zero temperature, j (F, 0) may be expressed as^^
7(F, 0) sinTtco'
where
.= '''^'^'^'^1^9.22.lO^^L (1.13)
he F F
Expression (1.12) holds as long as a; < 0.7,^^^ such that j(F, T)/j(F, 0) < 5.
Numerical calculations for different temperatures, work function values, and field
strengths have been made by Dolan and Dyke^^^ using the WKB method.^^^' ^^^ More
rigorous numerical calculations^^^ used a more accurate expression for the barrier
transparency, ^^^ which is applicable throughout the energy range of interest, both above
and below the potential barrier:
D(Ejc,F)=\l-\-Qxp(-2ih~^ j p(x)dxj\ (1.14)
where
1/2
/7(x) = J2mr£^ + ^ + ^ F
and jci and X2 are zeros of the radical defined such that jci < JC2. The results of numerical
calculations^^^ are shown in Fig. 1.3. These curves allow for an estimation of T-F emission
current at various temperatures, work functions, and field strengths. The effect of the
temperature is most pronounced for lowfieldsand low work function values. For example, at
1000 K with a work function value cp = 4.5 eV, j (F, T)/j (F, 0) ^ 5 at F == 2 • lO'^ V/cm
where as at F = 3 • lO'^ V/cm we have j(F, T)/j(F, 0) ^ 1.7. For higher field strengths
the temperature contribution is even less.
10 h
0.215
0.215
0.129
0.215
.5;/cr=0.129
kT=0.2^5
^;/c^=0.215
=-2;kT=0
0.129
/c7=0.215
(p=-4.5;/cr=0.129
1 2 3 4 5 6 7 8
10S/F,cmA/
Figure 1.3. Thermionic-field emission. Numerical calculation of the FE in FN coordinates for metals with different
work function values (<^ = 2 , . . . , 4.5 eV) in the temperature interval 0-3000 K.
At higher temperatures, co > 0.7, the emission process moves into the regime of
Schottky emission (thermal emission through a barrier lowered by an electric field) and
at still higher temperatures to the regime of pure thermionic emission.
For a fixed work function value, the regions over which particular types of emission
dominate will be determined by the temperature and electric field strength.
1.3. EXTENDING THE THEORY OF FIELD ELECTRON EMISSION

FROM METALS
The model of the bulk metal and its surface in FN theory is very simplified. The free
electron model, assumed atomic scale surface smoothness, and the one-particle approxima-
tion. There is also an inconsistency with the quasi-classical approach of combining the flow
of essentially classic particles with the quantum-mechanical phenomenon of tunneling. In
addition, FN theory considers the one-dimensional problem with a potential profile that
only accounts for image forces. Attempts to improve upon field electron emission theory
have been the result of the discovery of new facts; the departure from FN behavior at high
fields and high current densities,^^' î' ^^' ^^' ^^ the presence of high- and low-energy tails in
the emitted electron energy spectrum, ^^^' ^^^ the detection of fine structure in the FE spectra
from certain crystallographic orientations^^ (p. 223), anomalous broadening of the energy
spectrum in very high electricfields,^^^'^^^' ^^^ and the emergence of additional peaks in
the spectra for atomically sharp microtips.^^^
With the development of atomically sharp emitters used in tunneling spectroscopy ^^^
and electron holographys^' s^' ^^ there came a need to understand the degree of localization
of the tunneling process. The physics of the field electron emission process from ultra-
small-size emitters is now exciting a great deal of interest^^^"^^^' ^^^' ^^^ due to the rapid
development of the vacuum nanoelectronics and new means of diagnosing surfaces with
atomic resolution. However, at the moment we lack a theory that permits accurate calculation
of the cathode's operating characteristics. What are available are attractive ideas, tentative
estimates, and suggested methods of solution. The purpose of the next section is to review
these.
1.3.1. Deviations from the Fowler-Nordheim Theory and Peculiarities of

Field Electron Emission from Small-Scale Objects
As shown in a number of works, ^^' ^^' ^^^' ^^^ the one-dimensional approximation gives
a fair description of the electron FE process for atomically smooth emitters having a radius
greater than ^ OA |xm. In this case the width of the potential barrier is significantly less
than the cathode's radius of curvature. Atomic-scale surface roughness and the variation
of the work function between different emitter faces do not result in a significant deviation
from the results obtained with the one-dimensional approximation.^^' ^^^
Vacuum microelectronics employs cathodes having radii in the nanometer range (emit-
ters 10-200 A in radius). Even sharper tips (atomic sharpness) are used as point sources of
electrons in ultra-high-resolution electron spectroscopy, electron holography, and tunneUng
spectroscopy. With these field emitters the radius of curvature is close to or less than the
FIELD ELECTRON EMISSION FROM METALS 7
barrier width, so the assumptions of an one-dimensional barrier and field uniformity over
the apex of the tip are no longer justified. Strictly speaking, for these nanometer emitters
a fundamental revision of the theory is needed. In particular, it is necessary to solve the
three-dimensional Schrodinger using an asymmetric potential barrier, and calculate the
behavior of the potential near the surface accounting for its variation with radius r^ and
polar angle 0 As solving such a problem involves formidable difficulties at present only
rough calculations are being made.
One approach^^^^^^ has been to calculate the potential barrier transparency for the sim-
ple case of a rectangular axially synmietrical barrier formed inside the scanning tunneling
microscope (STM) gap. In this case, the transparency coefficient is given by,
m^^= êxp(-2poÔ ^^^^^^

8Jiy det(Qi - ^2 + 2dQ2^\)
where d\ is the length of the most probable tunneling path, po = yJ2{E — U), X is a
constant, and Q.j is a matrix representing the curvature of surface apex of the STM probe
at the point of its intersection with the most probable tunneling path.
Transparency calculations for a spherically symmetrical FE from an STM tip of radius
^^124-126 yi^l^
D{E) = ^ ^ exp(-2/7oî), (1.16)

4d
where
-1/2
^-=[G+i+^)G+i+^)]
and Rj are the principal radii of curvature of the surface under investigation at their point
of intersection with the most probable tunneling path.
Assuming that the emission originates predominantly at a single atom and using the
Wigner model of the localized state^^^ it was found that
jriFz
DiE) = ?—r— —-, (1.17)
( £ - £ o ) + 5(ri + r2)2'
where £0 is the energy of the localized state and Fi and Tz the localized state decay widths
in the STM tip and the sample, respectively.
Also, in the simpler case of a rectangular barrier profile and with FE from a single atom
^(^> = J77^ exp(-2po4), where ix^l. (1.18)

4(6[ + K)
Tentative estimates show that in the case of the STM gap, accounting for the three-
dimensional character of the barrier makes only for a slight difference to the preexponential
factor when compared to that of the one-dimensional approximation. With experimental
techniques this difference cannot be detected.
Multidimensional tunneling of electrons with application to STM and hence, FE was
also investigated by Lucas et al. and Huang et al.^^^^^^^ In Lucas et al.^^^ the scattering
theoretic technique of localized Green functions was applied to the calculation of emission
current distribution at the planar STM electrode. The obtained results were applied to the
explanation of the lateral resolution of STM. The work^^^ has been devoted to the general
formalism of the WKB approach to the problem of particle tunneling in multidimensional,
in particular, nonseparable, potentials.
Dyke,^^ Elinson,^^ and later in the case of nanoscale emitters Rodnevich^^^ noted that,
a more rigorous calculation of thefieldnear the tip will require taking into account the field
inhomogeneity over the surface. For conventional-size tips Dyke et al.^^^ made detailed
calculations of F(^) and compared these with experimental data for emitters with re >
0.1 |xm. The calculation results were found to be in good agreement with the experiment.
Rodnevich^^^ and Cutler et al.^^^ considered the emission from a sharp micro-
roughness on a planar macro-cathode surface. It was shown that in this case the expression
for the electric potential and, consequently, the electron potential energy in the Shrodinger
equation must be taken in a more complicated form as compared with the usual "plane"
description. ^^^ The usual expression for the potential energy is
^1 = - — - eF^x + £ F + ^ (1.19)
In the case of a planar cathode with a micro-roughness it is correct only where FQ =

const. Taking into consideration that F = /SFQ, where )S is a geometric factor and FQ the
electric field strength at the cathode surface, (1.19) can be written as
U2 = -—- peFox + EF-\-(P (1.20)

4x
Rodnevich^^^ drew attention to the fact that representing the potential barrier in the form
of Ui or U2 raises doubts because the potential distribution near a micro-roughness is
essentially nonlinear and should be written as (Rodnevich^^^)
U3 = -—- <i>(x) + FF + ^ (1.21)

4x
where 4>(jc) is a nonlinear function generally depending on the shape of the micro-
roughness. A suitable form of the function <^(x) should be used with each particular
microtip shape. As shown by Porotnikov and Rodnevich,^^^ the barrier structure, the FE
current density, and the energy distribution markedly depend upon a dimensionless param-
eter M = (p/Foa, where a is the micro-roughness height. For M <^ 1 the barrier structure
U3 is close to U2 and forM^ 1 the barrier practically coincides with U\ (JC) although
the surface field strength is given by ^FQ instead of FQ as for the barrier described by
Ui(x). Analysis of the potential distributions near a sharp micro-roughness for different tip
shapes shows that over a certain range of M values, 3 . lO""^ < Af < 30, current-voltage
characteristics are nonlinear and appear more intricate than predicted by the FN theory.
Similar ideas, but in a more consistent and rigorous manner, have been discussed by
Culter et al. and Jun He et al.^^^' ^^^ In these studies a numerical calculation has also been
made for the model situation that corresponds to a small-gap Spindt cathode design. In
Cutler et al.^^^ the influence of emitter nonplanarity upon the barrier shape was theoret-
ically investigated and current-voltage characteristics were calculated. It was found that
taking into account of the field strength variation along the direction away from the emitter
(potential function nonlinearity), and, consequently the corrected barrier shape leads to the
changes of current-voltage characteristic nonlinearity, that is, to deviation from FN theory
predictions. Cutler et al.^^^ described this effect as "dramatic increase in the current density
relative to the FN result for planar model of the tip with the same voltage applied in each
case"^^^ (p. 388). We note, that this increase is connected with the ")6-factor"—field strength
increase on the emitting surface of the tip to be compared with the abovementioned "plane"
case. We suppose that more correct comparison of the results must be done for identical val-
ues of the field strength at the emitting surface.^^^ For identical F values at the surface, the
shape of the potential barrier will be changing as the distance from the cathode is increased,
as shown in Fig. 1.4. It is evident from the figure that decreasing the field emitter radius
causes a marked increase of the barrier width and, consequently, the drop in the barrier
transparency and the tunnel current. So, the "^-factor" effect will be correctly included in
the interpretation of results in both cases. The effect of this factor can also be seen from
Cutler et al.^^^ The increase of current density 7, plotted as a function of voltage can be
compared with the decrease of J when plotted as a function of F = ^V. Jun He et al.^^^ is
a theoretical investigation of the interaction of electron and image charges for emitters of
various geometries. The obtained results are compared with the planar case, again for the
same anode-cathode voltages.
Calculations^^^ show that image potential for the small-scale field emitters (10-100 A)
practically is the same as for the planar cathode.
To illustrate specific features of the field electron emission process in small-size emit-
ters, Fursey and Glazanov^^^ performed numerical calculations of J-F relationships in
usual FN coordinates and the total energy distribution of emitted electrons. Calculation
results for the spherical cathode model were compared with the planar model.
10
'^^^ Fermy level

> 5
>2'
vS^
\ V \ ^ ^ Re = 40A
g> \ N. ^ ^ ^ ^ ^
(D \ >w ^**'"**""*-*^
C \ X.
0) 0 ^-^
]1
\ \
\â=8oA
^«v
c
©
o -5 \
Q.
c \
o \
B \
0)
\
UJ
-10 Meta1 Vacuum \
-15 -
-20 0 20 40 60
Distance from metal surface (A)
Figure 1.4. Dependence of the barrier shape upon emitter radius RQ for tungsten.219
For description of the emission from sharp microtips with the apex curvature radius
RQ a spherically symmetric model was chosen. In this model the potential variation with
distance x from the emitting surface is given by
U(x) = - ^ - eFox ( ^ ^ ] +E^-\-(p (1.22)

Ax \x-\-ReJ
where FQ is the field strength at the emitting surface.

The results are presented in Fig. 1.5 for RQ = 80 and 40 A, respectively. The deviation
from the FN plot (lower current density J for the same values of FQ) is more significant for
small RQ values. The reason is that the barrier width is larger, and therefore D{Ex) is less
for sharp emitters due to the widening barrier shape (Fig. 1.4).
The barrier can be considered plane if
where d is the barrier width for electrons near the Fermi level. For a plane barrier d ^ (p/F.
Noticeable FE is found for fields 0.3-0.9 V/A, which corresponds to the range of d values
(cp = 4.5 eV for tungsten) from 15 A (F = 0.3 V/A) to 5.0 A (F = 0.9 V/A). So, the
22-1
204
184 Plain model
•^ 16 Emitter radius Hg=80 A

>
u:
CVJ
14
^
^"
12
Ô)
Emitter radius RQ=40A
10
64
0.5 1 1.5 2 2.5 3

1/F(V/A)-^
Figure 1.5. DeviationfromFN plot for a small-scale objects (numerical calculation for tungsten) 219
^ ^ 1 I r-
1.5 2 2.5 3 3.5 4 4.5 5 5.5 6
Electron energy (eV)
Figure 1.6. Electron energy distribution from tungsten emitters for different radii.^^^
condition RQ » d, with the emitter radiusfixed,is violated in weakfieldsmore significantly

than shown in Fig. 1.5.
Energy distributions of emitted electron in high fields are presented in Fig. 1.6. One
can see that the area under the curve is greater for the plane case. Also, the width of energy
distribution curve is less for sharp emitters (RQ = 20, 40 A), because the barrier width
increase is more significant (compared with planar case) for low-energy electrons.
Theoretical and experimental investigations of the field electron emission from nano-
emitters are important for technology of point electron sources, particularly those used
in vacuum microelectronics devices, and with regard to the fundamental aspects of the
emission phenomena involving different areas of physics, such as quantum mechanics,
solid state physics, surface physics, etc.
We note also that in the case of nano-emitters traditional theoretical models such as
"solid state," "free-electrons," or "band-structure," etc. need to be analyzed more extensively
and more thoroughly. It is important to bear in mind that, as shown experimentally, in

specially formed nanotips the emitting area includes just a few atoms. (In Fursey et al."^^^
the curvature radii of micro-emitters were found to be in the range 10-80 A.) Also, ideas
relating to emission from a single atom at the emitter apex are discussed.^^^
1.3.2. The Effect of Fermi Surface Structure
The FN theory is based on the Sommerfeld's free electron model. This means that
the FN equation has been derived using the Fermi-Dirac energy distribution function. The
actual situation might prove to be much more complex.
The modem electronic theory of metals is based on the idea that electrons in metals
behave as quasi-particles displaying a complex energy dispersion law.^^^ One of the first
serious attempts to consider the problem of the electron FE with an arbitrary dispersion
law is to calculate the transparency coefficient^^^' ^^^ using the Bloch function formalism.
In this case, the Fermi surface is traversed by an axis P^ that is perpendicular to the cath-
ode surface, then the FN formula (1.7) based on the free electron theory holds (to within
the preexponential factor). In those cases where P^ axis does not traverse the Fermi surface
(so-called open Fermi surfaces), then due to the conservation of tangential quasi-momentum
for the electron exiting the metal, the work function, (p, in the exponential term in Eq. (1.7)
must be replaced with some effective quantity </?*>(/?. It has also been shown that the
energy spectrum is different than that obtained using the free electron model.^^^
Theoretically, it was found that the temperature dependence of the electron FE differs
in principle from what is predicted by the free electron model, namely, the emission current
in this case decreases instead of increasing as in the free electron model.
Despite the significance of these results,^^^' ^^^ the effects have not yet been observed
experimentally. This seems to be due to the fact that typically the most strongly emitting
areas of the cathode tip crystal (high index planes) are those associated with Fermi surfaces
perpendicular to P^.^^^ Thus the anomalies should be looked for in the low-index crystallo-
graphic areas.^^^' ^^^ Such features for {100} and {110} faces have been revealed primarily
in the energy spectrum of the tunneling electrons ^^ as will be shown below.
Note that the shape of the constant energy Fermi surfaces might affect the FE, not so
much through the exponential term in the expression for emission current, as through the
preexponential factor, which is significantly different than that given by the free electron
model. Research performed in our laboratory^^^' ^^^ showed that the experimentally deter-
mined constant in the preexponential factor coincides, to within order of magnitude, with
the value calculated from the FN theory only in the case of tungsten. For other materi-
als, such as molybdenum, niobium, and tantalum, experimental values might differ from
theoretical values by several orders of magnitude.
Unfortunately, due to the mathematical difficulties involved, the formalism developed
by Itskovich^^^' ^^^ did not yield the value of the preexponential factor. As he noted, the
calculations could be performed only in the case of nearly free electrons. Attempts to
calculate the preexponential factor with an arbitrary dispersion relationship were undertaken
by Gadzuk^^^ and Politzer and Cutler. ^^^' ^^^ They showed that for tunneling from the c-band
the preexponential factor might reduce the emission by a factor of 100-1000.
Gadzuk^^^ also calculated D for a Bloch metal using tight binding wavefunctions and
transfer Hamiltonian methods. Gadzuk^^^ further postulated a tight binding <i-band that
is described by a linear dispersion of the form E = b(k), where b is a. parameter. The

Unear E{k) relationship does not greatly alter the functional form of the energy distribution
expression from that of a quadratic E(k) relationship; however, the transmission coefficient
of J-band tunneling appears to be reduced relative to ^-band tunneling by a preexponential
factor of 10-2-10-^
1,3.3. Many-Particle Effects
There are many phenomena, which cannot be described in terms of the one-electron
approximation. These problems include the description of the variation of the potential
barrier within a small distance from the surface (less than two or three angstroms) and the
analysis of correlation effects involving electron-electron and electron-phonon interactions
in the near-surface region.
FN theory as well as subsequent theories, including the general theory of electron emis-
sion from metals,^^^ is essentially an one-electron theory. With the progress in quantum-field
methods with statistical physics it became possible to develop a multielectron theory of
field electron emission. One of the first attempts to apply this approach was undertaken by
Gadzuk.^^^ An analysis of the emergence of these ideas was made by Baskin.^^^ Gadzuk
used a technique employed previously in solving the problem of Josephson tunneling ^^^
in which the initial state of the anode-cathode system prior to emission is described by
two independent Hamiltonians: the cathode Hamiltonian //L (assumed to be located on the
left-hand side) and the anode Hamiltonian H^ (located on the right-hand side).
The external field applied to the cathode inducing electron emission is equivalent to
the introduction of a tunneling Hamiltonian Hj, which mixes states on the right and left
thereby causing an effective flow of particles from one side to the other. Ultimately, the
problem is reduced to the use of a complete Hamiltonian
H = HL-\-HK-\-HT (1.23)
where HL = J2k,cr ^kC^Ck and HR = ^^Â^d^dq are unperturbed Hamiltonians,

and C"^, C, J"^, J are the electron generation and annihilation operators on the left (right)
obeying the Fermi anticommutation rules. In the absence of electron pairing the C and d
operators commutate; in the general case these are quasi-particle operators.
The tunneling Hamiltonian can be written, ^^^
WT - X ] ^T^k,Cld,a + n,dâCka). (1-24)

k,q,a
Here, the matrix element Tkq was expressed in the form:^^^
(1.25)
The exponent in the Eq. (1.25) can be approximated by^

rX2
{<
• - E F
exp I - 2 / k±(x)dx : exp -c-\- (1.26)
^1
where as usual
3heFs
heFs
d =
£'F is the Fermi energy; p± the density of states with the momentum normal to the surface
barrier.
The current can be written as
j=e{NL). (1-27)
where the number of particles operator
NL = Y^C^Ck (1.28)
k
obeys the equation

iNL = [NL.H] = [NL,HT] (1.29)
The last equation is based on the property that the A'^ operator conunutates both with
HL and H^. Using Eqs. (1.29) and (1.24), the current can be written,
; = -ei([NL. Hj]) = -2e]mYTkq{C-^dq). (1.30)
Expression {C^dq) in an abridged form will be
{C^q) = {S-Ht)Cj^(t)dq(t)S(t)), (1.31)
where Ck(t) = exp(iHLt)Ck expi-iH^t) and
5(0 = exp - / j txp{r)t')Hi{t') dt' .
Assuming that the spectral function of the electron gas in the metal is '^5(£'^ ~ ^q)^
then calculation of the matrix elements using Green's function method gives^^^ the usual
expression for the FN function
j = {27Tmc(f/h^)e-'' (1.32)
Calculations show generally that the magnitude of the emission current depends on
the properties of the anode metal due to the assumption that the anode states are the final
states of electrons after tunneling. This is true only in those cases where the anode-cathode
separation is less than the electron correlation length. Experimentally, as a rule, the opposite
is true. As the final state, the electron state in a uniform field should be chosen.
It follows from Eq. (1.32) that the validity criterion for the FN equation is the closeness
of the spectral function of electrons in the metal to the delta-function. Because of the fact
that this is always the case for the electrons near the Fermi surface ^^^ the one-electron
approximation will be valid when most of the emitted electrons are in states close to the
Fermi level.
Gadzuk^^^ formulated a more rigorous statement of the problem, which, in principle,
can account for many-particle effects. With this formulation the validity of the one-electron
approximation appears to be more justified. The Hamiltonian HL used by Gadzuk does not
take into account electron correlation.
Lea and Gomer^^^ and, independently, Gadzuk, and Plununer^^^ noticed that high- and
low-energy tails are observed in the measured total-energy distributions of electrons for a
number of faces {111}, {120}, and {112} (Fig. 1.7). Lee and Gomer suggested that the
high-energy tails in the electron energy distributions are manifestations of electron-electron
scattering in the emitter crystal, making possible the simultaneous emission of electrons
in pairs. The emergence of such a pair can occur in the following way. Emission of an
electron that resides below the Fermi level at energy Ef: — A generates a "hot" hole in the
emitter near-surface region (Fig. 1.8). A conduction electron scattered by this hole acquires
an energy Ep + A and so is raised above the Fermi level where the tunneling probability
P2 is greater than the tunneling probability Pi of the first electron. Because of the very
short electron-electron tunneling time (~10~^^ sec) both electrons find their way into the
vacuum simultaneously as a pair. The probability of the simultaneous tunneling of the two
electrons is proportional to a product of the tunneling probabilities for each electron and
may be represented in the form^^^
•-c((^-A)^/2
p = Pi. P2 ^ exp e^p r ^ ( ^ ^ ^ ) 1 ^ exp(-2c^^/VF), (1.33)
5
F(V/A)
4 0.370
^ ^^ 0.355
l\/0.326
53 -
/ •
//
2
4
>
VV
--1> \NX.
1
1 1 1 1 1
-3 - 2 - 1 0 1
e = E-Ep(eV)
Figure 1.7. Electron energy distributions (normalized) of field-emitted electron from tungsten (111) at 78 K, 122
Figure 1.8. Schematic diagram of FE electron pair tunneling.
1 I I 1
Q
• (120) / J
y
• (112) J 1
A (111) /• I
-10
? \
o -11 #•
Q.
8
^^
•
1H
3)
-12
-13
-14
-{ ' T
? /
/
• / , , , "
-10 -9 -8 -7
lg'/o,a/'(A)
Figure 1.9. Experimental observation of quadratic dependence of the high-energy electron current (so-called pair
electrons) versus the total current in the probe. ^^^
where ^ and F denote, as usual, the work function and the appUed field, respectively, and
c — 0.683 in eV, A units.
As the total current, by Eq. (1.7), is proportional to Qxi^{-lc(p^/'^/F), the current due
to paired electrons should be proportional to the square of the total current
^couple ^ total (1.34)
It has been shown experimentally that /couple is indeed proportional to the square of the
total current (Fig. 1.9).^^^' ^^^' ^^^ It can be seen from Fig. 1.9 that at high current densities
the fraction of current due to "paired" electrons can be several percent. The idea of pair
tunneling has been proposed to explain an anomalously high work function value found for
the {011} face of tungsten.^^^' ^^^
As noted in the Introduction to this section, a many-particle problem, though of entirely
different character than that described above, must be solved to describe the variation in
the potential barrier over atomic distances from the surface. An analysis of the barrier
configuration near the surface accounting for this variation was attempted by Bardeen^^^
and will be considered further in the next chapter.
1.4. RESUME
The Fowler-Nordheim theory gives a fairly adequate account of the basic features of
field electron emission and its main deductions have a solid experimental corroboration.
1. The tunneling nature of the emission has been unambiguously proved in direct
experimental determinations of the energy distributions of FE electrons by the
retarding potential method.
2. The FN theory explains the exponential dependence of j pm F, or, to put it another
way, the linear form of the current-voltage characteristics of the field electron
emission (ln(0 = f(l/u)). The exponential variation of 7 with cp^^'^ has been
confirmed experimentally.
3. It should be noted that accurate comparison of the theory and experiment is impeded
by difficulties encountered in measurements of the electric field strength F. The
error in measurements of F under standard experimental conditions is usually not
less than 15 percent, posing problems in the use of the theory for accurate absolute
measurements.
4. The theory of T-F emission also has a reliable support in experiment. How-
ever, the FE emission theory does not provide simple analytical relationships.
Straightforward formulas are available only for the region of relatively low temper-
atures (formulas 1.12 and 1.13). Formulas for high temperatures are cumbersome;
therefore, for specific cases numerical calculations are more preferable. Such
calculations are illustrated in Fig. 1.3.
5. As to the refinements of the theory (Section 1.3), it should be noted that all of them
are of the fundamental nature but as yet could not be reliably confirmed because
of insufficient accuracy of measurements.
6. The strongest effect is the deviation from the traditional FN theory of the current-
voltage characteristics in the case of nanometer-scale field emitters. Preliminary
studies (Subsection 1.3.1) provided consistent explanation of this effect assuming
nonlinear variation of the potential near the emitter surface. This case is of partic-
ular importance because of the widening use of the nanometer-scale field emitters
in various vacuum microelectronics devices. Further thorough studies are required
in order to get a clearer insight into the field electron emission of such small-scale
emission centers.
This page intentionally blank
CHARACTERISTICS OF FIELD EMISSION AT
VERY HIGH CURRENT DENSITIES
2.1. DEVIATIONS FROM THE FOWLER-NORDHEIM THEORY IN

VERY HIGH ELECTRIC FIELDS
Basic research into Field Emission (FE) at high current densities was initiated by
Dyke and his colleagues in the mid-1950s.^^"^"^ Modem vacuum techniques, pulsed voltage
methods, and other advanced technologies allowed their use of higher electric fields and
higher emission current densities than were previously possible. Pulsed current densities of
10^ A/cm^ with tungsten (W) cathodes and, later, up to 10^ A/cm^ in the dc mode^^ were
achieved in electric fields 5-8 x 10^ V/cm.
Deviations from Fowler-Nordheim (FN) theory (Fig. 2.1) were first detected on W
emitters in highfields.^^'*^'^^'^^'^'^'^^^ Such deviations were also observed for Mo, Ta,
Re, W2C by Fursey with co-workers^^'^' ^^^ and by Elinson and Kudintseva"^^ for ZrC
andLaB6.
Dyke and Trolan^^ interpreted the deviations from linearity in the FN current-voltage
characteristics as due to lowering of the electric field near the cathode due to space charge
from the emitted electrons. Lewis^^ suggested a connection between the deviation in the
current-voltage characteristics and the dissimilarity between the actual shape of the potential
barrier at the metal-vacuum boundary and the approximate shape used in the image force
potential theory. This dissimilarity is pronounced in strong electric fields where the barrier
thickness approaches interatomic distances. We will discuss these ideas in more detail in
the following sections.
2.2. SPACE CHARGE EFFECTS IN FIELD EMISSION
The effect of space charge on FE was first discussed briefly by Stem, Gosling, and
Fowler. 1^^
It follows from the Poisson equation for a potential distribution between infinite planar
electrodes that
- ^ = -k-j-U-'f\ (2.1)
19
E\
•. C
\
\
\
- -12 \
c
\
AA
-16 A
A
A
A
I A,
-20
1.0 2.0 3.0
lo^/u.v-^
Figure 2.1. DeviationsfromFN theory at high electricfields:AC—Region of FN current-voltage characteristics;
CE—region of anomalous behavior.^^
where U is the potential, x is the coordinate; k = 2jr(2me/^)^/^, and j is the current

density. The solution of Eq. (2.1) with the boundary conditions
Û=o = 0 U\,=d = U^ dU/dx\,ô =0 (2.2)
(typical for thermionic emission) leads to a well-known formula of Boguslavskii-Langmuire
3/2
J = 9kd^ •â (2.3)
where d is the inter-electrode separation.

In the case of FE the electric field strength, F, at the emitter surface is nonzero. Hence,
the boundary conditions (2.2) take the form
Û=o = 0 U\,=d = U, dU/dx\,=o = F. (2.4)
The relationship between the electric field, current density, and the potential was
found,^^^ by integrating Eq. (2.1) using boundary conditions (2.4), as
1/2"
U^ 16 1/9 U.
F = 1 - -n(m,/2e)'^^j • - ^ (2.5)
Since F = U^^/d, if space charge is ignored, the criterion for neglecting the space
charge effect is.
1/2
16 w. t/^.
«1. (2.6)
This problem can be analyzed in more detail.^^ The validity of choosing the one-
dimensional approximation to represent the actual geometry of the point-cathode diode
CHARACTERISTICS OF FIELD EMISSION AT VERY HIGH CURRENT DENSITIES 21
2 4
Figure 2.2. Theoretical dependence of the current density on the applied voltage 34
was justified by the fact that the space charge affecting the electric field near the emitting
surface could be considered as concentrated within a very small distance from the cathode.
By solving Eq. (2.5) in combination with the FN equation,
7=AF2exp(-5/F), (2.7)
where A — 1.54 • 10~^/(^; B = 6.83 • 10^(p^^^T^(y), and eliminate the current density j
from Eqs. (2.5) and (2.7), we have an equation linking the field strength F to the potential U
4kAui^^Qxp{-B/F) - 3U^ = 9kÂ^F^d^ e\p{-2B/F) - 3Fd. (2.8)

Equations (2.8) and (2.7) can be used to calculate the current density for fixed values
of F as a function of the applied voltage. The usual relationship between the field and the
potential at zero current, F = f/a/^. follows from Eq. (2.8) at small F. At much larger F,
the exponential terms become dominant, and Eq. (2.8) is reduced to Eq. (2.3).
The parameter d was chosen in Barbour et al.^"* to be equal to 1/13, where fi is
a geometrical factor relating the field strength to the potential at current densities, where
the effect of space charge can be neglected.
Figure 2.2 is a plot of the current density versus voltage. The curve ACE was cal-
culated using Eq. (2.7) neglecting the space charge; the curve ACD was calculated using
Eqs. (2.7) and (2.8) and allowing for the space charge; the BD curve shows the Boguslavskii-
Langmuire law (2.3). The initial part of the curve coincides with the FN straight line at small
values of U. As U increases, the curve becomes sublinear and asymptotically approaches
the Boguslavskii-Langmuire curve.
The effect of space charge on FE for a spherical diode has also been investigated. ^^^"^^^
The Poisson equation in the case of a spherically symmetric electron flow has the form
d_ ydU
'd? dr \]- (U{r) + UQ)Î^'

(2.9)
where Uo is the initial electron energy, assuming, for simplicity, that at the cathode surface
(7 = 0. The exact solution of the equation with boundary conditions like Eq. (2.4) was found
by Poplavskii,-^^^ but his solution is cumbersome and cannot be presented in elementary

functions.
Isenberg^^^ estimated the additional difference of potentials, AUm, required, account-
ing for space charge, to make the field strength near the cathode equal to the value found in
the absence of space charge for the same total current:
AUrr, = ^ = J - ^
K'"^-)'
In - ^ - 1 , (2.10)
where / = j^^^l is the total current in a spherical diode, r\ is the radius of the cathodes,
ra is the radius of the anodes, and r^^ ^ ry^. Furthermore, the formula is valid only if
lAU/Uîl <C 1, and thus can be used for estimating the effect of space charge only when
deviations from the FN are small.
The AUm values determined from Eq. (2.10) have been compared with exact calculated
values from the Poplavskii's formulae.^^^ The maximum error in AUm calculated from
Eq. (2.10) is not greater than a few percent when the emission current is reduced by a factor
of two due to space charge.
Kompaneets^^^' ^^^ analyzed the asymptotics of the exact solution of Eq. (2.10) found
by Poplavskii^^^ and was able to simplify the system of equations and to obtain an estimate
of AUm. having the same order of accuracy as those given by Eq. (2.10).
2.3. SPACE CHARGE WITH RELATIVISTIC ELECTRONS
The preceding discussion of space charge is relevant when the velocities of the electrons
are small compared to the velocity of light (accelerating voltages of tens of kV). Relativistic
effects influence the electron mass and, therefore, the electron velocity can be a factor in the
emission process for electron energies in the range of hundreds of kV.^^^' ^^^ It is obvious that
retardation of the electron motion due to the relativistic effects increases the effect of space
charge. It has been shown by calculation^^^ in that the voltage required to maintain a given
current density in the planar diode model is increased several times when relativistic effects
are accounted for. In the majority of real cases, however, the field cathode is manufactured
as a tip and, hence, the spatial distribution of the beam current density is similar to the
distribution in a spherical diode.
Accounting for the nonuniformity of the potential distribution in a real diode by adjust-
ing the inter-electrode distance J in a planar diode, based on the assumption that most
of the screening charge is concentrated within a distance approximately equal to the tip
radius, is no longer a good approximation. Qualitatively, the reason for this lies in the fact
that the total screening charge in the inter-electrode gap, d, at high currents will still be
quite large despite the fact that the space charge density decreases rapidly away from the
cathode surface. The error involved is especially significant when relativistic effects cause
the density of the screening charge pe to decrease as 1/r^, that is, much slower than in
the nonrelativistic approximation (po ^ r~^[U{r)]~^^^, where U{r) is the electric field
potential).
A more rigorous approach is to solve a system of equations involving the Poisson
equation and energy conservation and the continuity equation for the current density in
a spherical diode model^^-^
V^(7=47rpe, (2.11)
m^c^ -\-eU = meC^Cl - v^/c^)'^^^, (2.12)
V . ; = 0, (2.13)
with boundary conditions
U\r=r,=0 dU/dr\r=r,=F. (2.14)
The emission current density j = j\r=r\, is determined by the FN equation (2.7). The
approximate solution of Eqs. (2.11-2.14) has the form
+ ln[y;K2T1g-(l + , ) ] - j ; ; ^ , , „ [ . (2.15)
where AUra, as in Eq. (2.10), is the correction accounting for space charge when relativistic
effects are important, and r] = eFr\^/mQC^. Equation (2.15), like Eq. (2.10), was obtained
under an assumption that ra > r^ and its use is justified only if
|A^m/ôl « 1, (2.16)
where Lô is the potential in a spherical diode in the absence of space charge and is related
to the electric field strength F by
Uoir,) = Fr^[l - (ry,/r,)l (2.17)
In accordance with Eqs. (2.7) and (2.15-2.17), a fixed current density JQ and fixed
anode voltage U^ correspond to a fixed F value. Using these relationships, one can plot
the nonlinear, in lg[7(l/f/a)] coordinates, current-voltage characteristics of a spherical
diode. As before, the approximate character of the solution restricts the validity of these
relationships to small deviations from the FN law.
One can derive from Eqs. (2.15-2.17) an explicit formula for the current density at
which the deviation occurs by assuming, as the deviation criterion, twice the current value
calculated from the FN formula disregarding the space charge and relativistic effects at the
same applied voltage.
. [\n(2FyH2e/m,)'/^) Inrk ,
In jdev = In [ 47t(2-\-B(py^/F) J 2
iiîl.
- In I n — ^ - 1 . (2.18)
n. J
The dependence of the deviation-point current on the emitter radius for two work
function values is shown in Fig. 2.3.
The current-voltage characteristics begin to deviate from the straight line at a current
value that increases with emitter radius and decreases with increasing work function. The
r^, cm
Figure 2.3. Dependence of the current density on the apphed vohage when the influence of electron space charge
is detected:^^^ 1: <^ = 4.5 eV; 2: (p = 2.0 eV.
calculation predicts that the influence of relativistic effects must be taken into account for
emitter radii greater than 10~^ cm. At high FE currents in the spherical diode,
/ > 1.7 10^ A, (2.19)
the system of Eqs. (2.11-2.13) is simplified and an approximate expression can be written
forL^:
U = (//c)[ln(ra/rk) - 1]. (2.20)
It follows from Eq. (2.20) that the value of the current is not dependent on the work
function of the cathode material. FE current is limited only by space charge. The same
analysis for a cylindrical electrode system leads to
U = (47r/c)rark;e. (2.21)
Note that the cylindrical model applies in the practically important case of a sharp
edge-plane system (blade cathodes). If the conditions (2.16) and (2.19) are not satisfied,
the approximate analysis is irrelevant and one must solve Eqs. (2.11-2.13) numerically. In
Fig. 2.4, the current-voltage characteristics of spherical diodes with W cathodes are shown.
The plots were made from the results of numerical calculations.^^^^ The characteristics
include a rather extended portion where the increase in current with voltage is slower than
in the linear region. (The increased slope of the characteristic at high voltages is related to
scaling effects.)
This sublinearity is pronounced for HC ^ 10~^ cm and at current densities j >
lO'^ A/cm^. The decrease of current due to the space charge in this case is so sig-
nificant that it becomes very difficult to achieve current densities of 10^-10^ A/cm^,
which a FE cathode can easily produce when space charge is not important. Such
current densities have been observed experimentally in pulsed regimes on sharp tips
(r\ = 10~^-10~^ cm).^^'"*^'^^' ^^^' ^^•^' 174,204 p^j. example, to achieve a current density
of 7e = 10^ A/cm^ at rk = 10~^ cm, the voltage must be increased by roughly an order of
magnitude over the value obtained when space charge can be neglected.
Distributions of the potential and thefieldin a spherical diode have also been calculated
by numerical methods,^^^ see Figs. 2.5 and 2.6. It should be noted that at relativistic electron
energies and high current densities, the potential in the diode rises rather monotonically
CHARACTERISTICS OF FIELD EMISSION AT VERY fflGH CURRENT DENSITIES 25
0.5 1.0 1.5 2.0

W^/bU,cmN
Figure 2.4. Current-voltage characteristics of the spherical diode with a W cathode accounting for space charge
and relativistic effects {(p = 4.5 eV, fi = F/U 2X j = 0). 1: FN line ignoring the effects of space charge;
2: rj, = 10~^ cm, 3: r^ = lO--"^ cm, 4: rj, = IQ-^ cm.Ô^
30
>
? 20
CVJ
I
CVJ
II II
CM CVJ
3 10
1 1 1 1 1
0.2 0.4 0.6 0.8 0.01 0.03 0.05 0.07
r, cm r, cm
Figure 2.5. The potential distribution in the spherical diode for different emission current densities: 1: ; =
10^ A/cm2; 2: j = lO'^ A/cm^; 3: ; = 10^ A/cm^ (rj, = 10"^ cm).203
despite the large difference between the cathode and anode radii. The field strength under
these conditions is at a maximum some distance from the cathode surface instead of at the
cathode surface (Fig. 2.6).
In conclusion, we note that the spherical diode only approximates the point diode
potential symmetry. The principal dissimilarity is that in this spherical diode model, the
intrinsic magnetic field due to the emission current is not taken into account. Recall that
the ratio of magnetic component of the Lorentz force to the force of electrostatic repulsion
300
E
o 200
II
100
0)
r, cm
Figure 2.6. The electric field distribution in the spherical diode for different emission current densities: 1: ; =
10^ A/cm^; 2: j = 10^ A/cm^; 3: ; = 10^ A/cm^ {r^ = lO'^ cm).^^-'^
in an electron beam is of the order of v^/c^. If the electron energy in is the MeV range,
the electron velocity is then comparable to the velocity of light, and the resulting magnetic
field can significantly affect the emitter's current-voltage characteristics.
2.4. THE SHAPE OF THE POTENTIAL BARRIER IN

HIGH ELECTRIC FIELDS
FN theory and its corollaries have been well validated with W emitters in fields of
~ 4 • 10^ V/cm. The assumption of a smooth surface and the use of the simpHfied model of
the mirror image forces potential (Fig. 1.1) are quite justified at these field strengths.
In higher electric fields, the dimensions of the potential barrier at the solid-vacuum
interface become on the same order as interatomic distances and the radii of close-range
interactions. The width of the image force barrier, AJC = |jci — JC2|, where jci and X2 are the
classical turning points (Fig. 2.7a), can be easily calculated if the kinetic energy of electrons
in the direction of emission is zero at the turning points. The electron energy at the Fermi
level is Ex = —(p. If the potential function is the image potential and that of an external
field F, then
Ex = —e /4x — eFx (2.22)
and
/:^x = {(p^ -e^F)^f'^leF. (2.23)
The distance JCm from the interface of the potential maximum and the barrier height
//m are found from the condition of the potential extremum
x^ = l/2(^/F)i/2, (2.24)
Hra = (p-(e'F)1/2 (2.25)
In Fig. 2.7, the plots of AJC(F), Xm, and Hm(F) are shown.
(a) ,
1
-^m
V2
X
^ X
Dx
(b) 14 (c) 4.0
3.2
>
0) 2.4
1.6
0.8
3 5 7 9 1 3 5 7
F10^V/cm FlO^V/cm
Figure 2.7. The dependence of the barrier width on image forces (a) the position of the maximum in the barrier,
(b) the barrier width and the position of its maximum as a function of appHed field, (c) the maximum height of
the barrier as a function of applied field.
It can be seen that in fields exceeding 5 10^ V/cm, the barrier width and jCm values
are of the same order as the lattice constant for W (a = 2.37 A). It is obvious that the
previous assumptions regarding the metal surface in such strong fields are unreliable, and
it is necessary to know the exact nature of forces acting on the electron near the surface.
In this case, the surface can no longer be treated as structureless and perfectly smooth.
The interaction with surface atoms affects the electron motion and, strictly speaking, to
describe the potential one needs to solve a multidimensional, quantum-mechanical many-
body problem (see also Chapter 1). This problem has not been solved in the general case.
Attempts to solve the problem in the three-dimensional case, even in the one-electron
approximation, have failed due to the mathematical difficulties caused by the asymmetry
of the forces at the metal-vacuum boundary.
The most comprehensive quantum-mechanical theory describing the potential barrier
for electrons near the metal boundary is by Bardeen.^^^' ^^^ The problem was solved by a
self-consistent field method in the framework of Hartree-Fock approximation. The impor-
tant result of Bardeen's research^^^ was the demonstration of the fact that the barrier due to
the image potential is a good asymptotic approximation to the potential at sufficiently large
distances from the surface.
Juritchke^^^ found that, in quantum-mechanical terms, one can interpret the classic
polarization of a metal in terms of an electron and a "hole" coupled by the exchange
interaction and moving in opposite directions.
Bardeen's potential is shown in Fig. 2.8a. The potential function outside the metal is a
smooth curve coinciding with the image force law at large values of x and asymptotically
x=0
Figure 2.8. Different models of the potential barrier at metal-vacuum interface: (a) no external electric field;
(b) with an external electric field F. 1: image potential barrier; 2: Bardeen's barrier; 3: Cutler-Gibbons barrier;
4: Seits-Vasiliev-van Gostrum barrier.
approaching the energy of the lowest state £3 in the metal. The potential curve has a
shallow minimum in the near surface region and decays quickly from the surface. Loucks
and Cutler^^^ have accounted for Coulomb correlation by using Paince-Bohm's formalism
and more carefully calculating the exchange portion of the potential. The correction led to
only a small increase in the potential barrier height.
It is easily seen from Fig. 2.8 that with Bardeen's potential, the barrier under the applied
electric field is wider and higher than that given by the image potential barrier alone. It is also
obvious that the dissimilarity should be greater in the high field region where the influence
of the near-surface potential is important. Lewis^^' ^^^ was the first who called attention
to the fact. His qualitative analysis has shown that this effect should lead to progressive
damping of the emission current with increasing field strength. A more rigorous analysis
of the influence of the barrier shape on the current-voltage characteristics in strong electric
fields was made by Cutler and Nagi.^^ Since Bardeen's potential, and that of Juretchke and
Loucks,-^^^' ^^^ have a complicated analytical form and do not readily lend themselves to
exact calculation of the transmission coefficient from the Schrodinger equation. Cutler and
Nagi-^^ used a simplified expression for the potential that included a quantum-mechanical
correction obtained by Cutler and Gibbons.'^^^ The function has the form.
^2 ^2
U(x) = -— -\-h—^ = E{x)-\-hx~^\E{x)\ atx > JCS > 0
4JC AX^ (2.26)
U(x) = —Ea at jc < jCs
where /i is a parameter depending on the metal and has been calculated to be on the order of
0.069 A for £a = 10 eV,^^' ^^^ and 0.9 A for E^ = 9.3 eV.î^ Although this approximation
is rough and does not consider details of the barrier structure, it nevertheless reflects the
basic features of Bardeen's potential: transition to the image potential at large distances
from the metal, the occurrence of the potential minimum near the surface, and the reduction
of the force acting on the electron in comparison with the image force alone. FN curves
calculated using this modified potential are shown in Fig. 2.9b.
CHARACTERISTICS OF FIELD EMISSION AT VERY fflGH CURRENT DENSITIES 29
(a)
-3
F=5-10^V/cm ::5
>
5-
-7 F=10»V/cm
10 2 4
x,A lO^/F.cmA/
Figure 2.9. (a) Plot of the potential energy near a metal surface as a function of distance at different fields;
(b) illustration showing the deviation from linearity of the current-voltage characteristics due to changes in
barrier shape by the image force: 1: image potential; 2: simphfied potential of Cutler-Gibbons; 3: potential of
Seits-Vasiliev-van Gostrum.
Earlier, Vasiliev^^^ used the potential function proposed by Seits^^^ to analyze the FE
current in strong fields (later this potential was used by van Oostrom^^ in his calculations).
The image potential has a discontinuity as jc -> 0. This discontinuity can be eliminated by
writing the potential function in the form
U{x) = -e^/{Axê^/E^). (2.27)
Vasiliev also used a more general expression derived by Miller and Good^^^ for the
transmission calculations which is valid throughout the range of the electron energies, both
below and above the top of the barrier:
D{E,F) 1+exp-J- / \2m^{E^ ^ eFx + U{x))]^''^ dx (2.28)
The current-voltage characteristic calculated with the Seits-Vasiliev-van Oostrom potential

is also shown in Fig. 2.9b.
As shown in Fig 2.9, the use of the corrected form of the potential near the surface
leads to the deviation of the current-voltage characteristic at lower currents. It is important
to note that the values of the current density and field, at which deviation from FN theory
occurs, are about the same as for which space charge affects the emission characteristics FE.
2.5. RESUME
Summarizing this chapter, its main results are as follows.

1. In the region of high fields and current densities, a noticeable deviation is observed
from the FN law, which is apparent as a slowing down of the current rise with
increasing voltage (Figs. 2.1-2.3). This effect is observed not only for tungsten
emitters but also for Mo, Ta, Re, W2C and other materials, where high current
densities (> 10^ A/cm^) can be reached. This deviation is stronger at high current
densities and in materials with lower work function.
2. Theoretical analyses and experiments indicate that the slowing down of the FE
current rise with increasing electric field strength is caused by two factors: build-up
of the space charge of emitted electrons and increasing contribution from the
potential in the near-surface region because of quantum-mechanical effects. Under
conditions usual in a FE experiment and at r^ = 10~^ cm, voltages up to 10-20 kV
and work function values ofcp^ 4-4.5 eV are both effective. At greater KQ and
small (p, as well as at very high current densities, the space charge becomes a
dominant factor.
3. In blunt emitters (r^ > 10~^ cm) at high anode voltages (Ma > 300 kV), rela-
tivistic effects become clearly apparent in the field electron emission. According
to calculations, the space charge of relativistic electrons can change the anode
voltage necessary for obtaining high current densities (10^-10^ A/cm^) by more
than an order of magnitude. Under ultra-relativistic conditions (at Ma > 10 MV
for tungsten emitters), the emission current becomes practically independent of
the work function and rises Unearly with voltage (see Eq. (2.20)).
MAXIMUM FIELD EMISSION CURRENT DENSITIES
3.1. THE THEORETICAL LIMIT OF FIELD EMISSION CURRENT
One of the most remarkable results of field emission (FE) quantum theory is the pre-
diction of very high emission current densities. Very high current densities are possible due
to two factors: (1) No energy is required for maintaining the emission process if electrons
exit the solid by a tunneling mechanism, that is, the emitter does not need to be heated,
irradiated, or otherwise excited by some external energy source; and (2) there is a very large
reservoir of electrons near the Fermi level of a metal.
As the density of conduction band electrons in metals is of the order of 10^^-10^^ cm~^,
it can be assumed that the electrons in the metal flowing toward the metal/vacuum interface
can sustain a current density of about 10^^ K/cvc?-?^^ In the limiting case, of a potential
barrier transparency of 1 (i.e., the barrier is fully eliminated by the external electric field),
the maximum electron current density passing through a metal-vacuum interface can be
expressed in the free electron model as,
j = e]jjfip)n.—--^, (3.1)
(p,n)>0
where, e and m are the electron charge and mass, respectively, p is the electron momentum,
n is the unit vector normal to the emitting surface, and / is the Fermi function. At T =
0 K, f(p) equals unity at the Fermi surface, and f(p) outside it, and Eq. (3.1) takes a more
straightforward form:
(p,n)>0
For a spherical Fermi surface, Eq. (3.2) in spherical coordinates has the form:
j= - d(t) do' sin6> • cos6> / p^-^, (3.3)

m Jo Jo Jo ^
where pp is the momentum of an electron at the Fermi surface. Integration of Eq. (3.3)
gives a simple relation for the limiting emission current due to conduction band electrons
31
in the metal
j= -f-^ = 4.3 . 10Ê^ [A/cm^], (3.4)
where £ F is the Fermi energy (in eV) measured from the bottom of the conduction band.
The above treatment can be extended to cover the case of eUiptic energy bands where
the dispersion law has the form:
E, = jl + ^ + f . ,3.5)
2m X 2m y 2m z
Assuming that x is normal to the emitting surface, and introducing the new variables
Px/^x = C^. Py/^y = ^^. PJ/f^z = y}. we have
I7I = ^ / / / "^^'^^ d^-di^-dX' Jtr. (3.6)
where 7tr is the Jacobian of the transition from coordinates px, Py, Pzô^.i^, x- Integration
of Eq. (3.6) gives a formula for the limiting emission current in the case of elliptic Fermi
surfaces
7max = 4.3 • 10^m:(m;m*)i/2£2^ (3.7)
where m* = mx/m, m* = my/m, and m* = m^/m are the ratios of the effective mass
components to the free electron mass.
Note that the limiting emission current density 7niax for the elliptic Fermi surface can
be lower than in the case of the spherical Fermi surface for particular ratios of effective
msisses mx, my, mz.
The FE current density is usually lower than that predicted by Eqs. (3.4) and (3.7)
due to the initiation of a vacuum arc because of the onset of FE instability at high current
densities. When an arc is induced, the cathode begins to melt and ceases to be a controllable
FE cathode.
3.2. EFFECTS PRECEDING FIELD EMITTER EXPLOSION
A very important case of vacuum breakdown initiation is observed with so-called

"conditioned electrodes" in high vacuum. This mechanism involves FE from areas on
the electrode surface where the field strength is enhanced due to micro-protrusions such
as microtips. High current causes the electrical explosion of these micro-protrusions and
the formation of a dense plasma near the cathode surface facilitating breakdown and arc
initiation. This phenomenon is of a very general character.
This mechanism not only facilitates vacuum breakdown but also plays an important
role in the subsequently ignited vacuum arc. It is possible to develop a model of the
pre-breakdown situation at an individual microtip and to study quantitatively emission
phenomena and processes at its surface in the time period just preceding the breakdown.
These investigations allow one to experimentally study the initiation phase of break-
down and provide a key to understanding the mechanism of vacuum breakdown. This
MAXIMUM FIELD EMISSION CURRENT DENSITIES 33
approach was suggested and brilliantly conducted in the well-known studies of Dyke and
his collaborators.^^"^^ Such studies were continued"^^''^^'^^' ^^'^' ^^^ and further advanced
with improved experimental techniques including increased experimental sensitivity (5-6
orders of magnitude), larger pre-breakdown time ranges, and special conditions that enabled
the observation of individual stages of the process. Mention should be made of the use of
the nanosecond time scale techniques for the excitation and registration of pre-breakdown
processes,^^' ^^ the use of long current pulses, 10~^-10~^ sec, to simulate quasi-stationary
conditions,^^ and the extensive utilization of the FE microscopy and electron transmission
and high resolution shadow microscopy.
Increased sensitivity enabled studies of emission processes localized to surface areas
less than 10"^ cm in size^^' ^^' ^'^^' ^'^^' 2^^' ^^^ at temperatures down to 4.2 KP"^
3.2.1. Dyke's Experiments
In a set of elegant experiments by Dyke and co-workers,-^^"^^ it was shown that

the destruction of a field cathode is caused by large emission currents. In the case of
atomically clean and smooth emitters, instabilities can develop due to the formation of
micro-protrusions at the surface as a result of ion bombardment or other reasons.^^' ^^' ^^
Elimination of the conditions leading to formation of micro-protrusions allows one to
achieve current densities of 10^ A/cm^ dc^^'-^^ or lOÂ/cm^ for a pulse duration of
lO-Sec.^^'^2
Using pulsed excitation and registration in their investigations on tungsten (W) emitters.
Dyke and co-workers were able to identify and control a number of effects during the
transition to vacuum breakdown precedingfieldemitter destruction.^^ One of these effects is
the occurrence of a brightringin the emission images ("ring" effect)—^Fig. 3.1a; the other is a
spontaneous increase in the emission current during a voltage pulse ("tilt" effect)—Fig. 3.1b.
Figure 3.1. Observation of pre-breakdown phenomena [32]: (a) "ring" effect; (b) "tilt" effect—increase in current
with time with voltage held constant, (top current oscillogram, / = 10"^ A), (bottom voltage oscillogram,
U =9.12 kV), pulse duration 0.5 îsec.
Explosion
f(2^sec)-
Satu ration
f(4)isec)-
Figure 3.2. Current saturation effect in the pre-breakdown stage.^^- ^^ [ 1-6]: current versus time traces for different
values of pulse voltage Un > Um for n > m.
Unless the emitter is destroyed, the pre-breakdown effects are fully reversible. The Dyke
experiments were verified and advanced in investigations of Elinson,^^' ^ Shuppe,"^^' "^^ and
in early works of Fursey's team.^^"^^' ^^^' ^^^
Both these effects have been interpreted as being caused by emitter heating due to the
electron current, that is. Joule, or resistive, heating. Results of preliminary calculations
based on joule heating^^ were consistent with the model. It was suggested that heating of
the field cathode tip leads to an increase of the emission current because cold FE gives
way to thermal FE. Higher emission current raises the cathode temperature still further and
so on until the field cathode is destroyed. At the final stage of the process, evaporation
and ionization of the cathode surface atoms increase the electron concentration, further
accelerating the process.
3.2.2. Pre-Breakdown Phenomena
New features in pre-breakdown phenomena were observed by utilizing more precise

techniques of generating and regulating the pulsed fields and considerably increasing the
sensitivity to pre-breakdown currents (6-7 orders of magnitude increase in sensitivity)."^^' ^^
First, pre-breakdown effects were identified more clearly (Fig. 3.2). It was found that no
avalanche-like current increase occurred if high current densities were approached at a slow
rate and, on the contrary, a saturation region was observed (Fig. 3.2). This fact led to the
suggestion that explosive breakdown occurs not as a result of a gradual evaporation of the
emitter material, but as an electrical explosion of the entire apex of the field emitter.'^^' ^^' ^^
The second notable feature was that the "ring" observed by Dyke was not a single ring,
but in fact, a number of concentric rings of oscillating intensity."*^' ^^' ^^"^ In a number of
cases, these rings overlay the central part of the emission image (Fig. 3.3). The mechanism
^^> (b)
(c)
Figure 3.3. Multiple rings in the FE pattern: (a) W;^ ^ (b) and (c) Ta. ^^"^ Photograph in (c) is taken with a spherical
microscope (side view).
of the ring formation is not completely understood at present. Their outward appearance
and multiplicity enable us to presume that these rings are due to an electro-optic effect.
The pre-breakdown effects (the "ring" and the "tilt") are observed not only on tungsten
but represent general manifestations of the pre-breakdown state for a number of refrac-
tory materials. They have been observed with tantalum (Fig. 3.3b,c),^^^ molybdenum,^^^
rhenium,^^^ tungsten carbide,^^"^ niobium,^^' ^'^ and systems involving absorbates.^^' ^^
3.2.3. Time-Dependent Observations
To understand pre-breakdown phenomena, we need to quantify the time dependence

of the effects that are observed. Time dependent processes were investigated by carefully
approaching the critical pre-breakdown state at different field pulse lengths.'^^"^^' ^^' ^^
Experiments have been conducted with pulse lengths from 10 sec in duration^^ down
through the millisecond range^^'^^ to the 10~^-10~^ sec regime.^^'^'^'^^"^ Current versus
time traces in the pre-breakdown phase have similar general features over a wide range of
pulse time intervals. However, the mechanism of the phenomenon appears to be different
for different pulse durations.
In the nano- and microsecond regions, the current variation is due to a transition
from FE to thermal FE by heating of the entire emitter tip. In the millisecond regime^^' ^^
(a) Voltage "LJ" ^^^
Figure 3.4. Probe experiments with registration of high FE current densities.'*^'^^ (a) Tube with a rotating field
emitter on a gimbal. 1: windows for photoregistration. 2: luminescent screen. 3: probe hole. 4: collector of emitted
electrons. 5: collector appendix. 6: FE cathode. 7: conducting film screening the collector from high-frequency
interference, (b) Current and voltage oscillograms: 1,2,3: current oscillograms for different pulsed voltage values.
and under quasi steady-state conditions,"^^ modification of the emitter geometry under the
influence of the field and thermally activated surface diffusion play a dominant role.
3.2.4. Quantifying Local Effects
High measurement sensitivity has enabled investigation"^^' ^^ of the FE currents in

the pre-breakdown phase from separate areas of the emitting surface using a probe-hole
(Fig. 3.4). Figure 3.5 shows current oscillograms taken in the ring and central parts of the
FE image. Note that the current in the ring appears with a considerable delay.
Current in the central portion of the image (Fig. 3.5) varies very much like the total
current from the tip. The "inertial" behavior of the current in the ring shows that it may be
influenced by emitter tip Joule heating.
3.3. HEATING AS THE CAUSE OF FIELD EMISSION

CATHODE INSTABILITIES
At present the hypothesis that the FE current is limited by thermal destruction of

the emitter has been demonstrated experimentally and confirmed by rigorous theoretical
calculations.
(a) Probe (b)
(c)
Figure 3.5. Emission current from the ring area and central area of the emitting surface.'*^ "^^ (a) Schematic
emission pattern for W and the probe-hole. (1,2: different probe positions); (b) W emission pattern at maximum
current density and the view of probe hole; (c) emission current versus time for the ring and central area with a
fixed pulsed voltage.
3.3.1. Experimental Demonstration of Field Emitter Heating
Studies of the ring emission component of the total current have shown that it has a
considerable time delay (Fig. 3.5). This delay may be considered as a direct proof of the
emitter tip heating. If this is true, intentional heating of the tip by an external source will
result in the shift of ring current to the beginning of the pulse.^^ Experimental results shown
in Fig. 3.6a demonstrate emitter heating in the pre-breakdown phase.
Another means of ascertaining the role of emitter heating is by observing the behavior
of adsorbed layers,^^ as surface migration of this coating can serve as a reliable indication
of the emitter heating. In one experiment, a layer of Barium (Ba) is evaporated onto the
side of the tip surface. As seen from Fig. 3.6b, up to the pre-breakdown phase, the emission
image remains unchanged. Self-heating of the tip causes migration of Ba and the formation
of a uniformly distributed layer (Fig. 3.6c).
The migration process can also be correlated with changes in the emission current and
their occurrence during the voltage pulse see Fig. 3.6d. During the time interval (ô — î),
the emitter heats up to a temperature sufficient to initiate surface migration; during interval
(ti — t3), the migrating layer passes the probe-hole and the collector current rises rapidly;
during the time interval (^3 — ^4), the edge of the Ba deposit spreads over the entire portion
of the surface covered by the probe-hole and the subsequent variation in collector current is
(a) 300 K
(b)
(c)
(d) '3 U
\—^
f(4nsec)-
U
Figure 3.6. Experimental evidence of field emitter heating at maximum current densities, (a) Emission current
from the ring at different temperatures Ti = 300 K, T2 = 1500 K; (b) and (c) migration of Ba across W surface
in the pre-explosion phase (b: initial image, c: final image);^^' ^^^ (d) current through the probe-hole versus time
at fixed pulsed voltage.
small. By measuring the time it takes the Ba layer to pass the probing orifice and estimating
the magnification of the FE microscope, the migration rate and temperature (from the
Arrhenius equation) can be determined. According to these estimates, Ba migration is
initiated at the emitter temperature of ^2000-2500 K.
3.3.2. Analysis of Thermal Processes
The experiments'*^'^^ conclusively demonstrate that with clean electrodes, pre-

breakdown effects are due to Joule heating of the field emitter. Because of thermal
instabilities in the field cathode tip, a transition from FE to explosive emission and vacuum
breakdown occurs7^ The emitter tip is heated as a result of two processes: Joule heating
and heating due to the Nottingham effect.
3.3,2.1. Nottingham Effect
For many years it was believed that the major cause of emitter destruction was Joule
heating. However, experiments by Swanson et al.^^^ as well as calculations by Levine^^^
indicate that the dominant contribution to the thermal balance during FE is due to a purely
quantum-mechanical energy exchange process, referred to as the Nottingham effect.^^^
According to this energy exchange mechanism, most of the electrons tunnel from energy
levels below the Fermi level, that is, their energy is lower than that of the conduction band
electrons that replace them (Fig. 3.7). As the incoming electrons are more energetic than
the tunneling electrons, energy evolves near the surface (within roughly one electron mean
free path). At high current densities, the energy associated with the Nottingham effect could
exceed the energy due to Joule dissipation.
Together with Joule heating, Nottingham heating causes a rapid temperature rise in the
emitter tip. It is noteworthy that after reaching a temperature referred to as the inversion
temperature, the Nottingham effect first vanishes when the spectrum of emitted electrons
becomes symmetric about the Fermi level (Fig. 3.7). Then, as the temperature increases
beyond the inversion temperature, the Nottingham effect begins to cool the cathode because
most of the emitted electrons come from energy levels above the Fermi level and, thus,
extract energy from the cathode.
Joule heating and the Nottingham effect together can heat the cathode faster than each
of them alone. As we will see, in this case, an interesting situation can arise because at
some point the surface temperature drops considerably below the temperature inside the
cathode.
Vacuum
level
Fermi
level
Bottom of
conduction band
Figure 3.7. Illustration of Nottingham effect.

3.3.2.2. Theory of Nottingham Effect
The energy exchange in the Nottingham effect H^ is represented by the energy flow
conveyed to the metal of taken from it by emitted electrons. This energy flow is equal to
the number of electrons emitted in a unit time, that is, to the emission current / multiplied
by the difference in average energies s between the field emitted electrons £em and the
conduction band electrons Ecd'-
/ / N = / • £. (3.8)
where 6 = £em - ^cd-
In the Nottingham model,^^"* it is assumed that the average energy of conduction band
electrons (£cd) is equal to the Fermi energy (Ecd = Ef:), and accordingly,
S = E ^ r a - ^F. (3.9)
Note that this assumption is true only for metals with spherical isoenergetic Fermi
surfaces. For the complex band structure of transition metals, this condition is not an
accurate assumption (for details see Seits^^^ and Mott and Jones^^^). However, as it is
very difficult to perform analytic calculations without this simplifying assumption, in the
discussion below, it will be assumed that Ecd = ^F.
The average exchange energy per one emitted electron is
/ 6- P(s)de
e= ' ^ , (3.10)
/ P(e)de
where e = E — E^ and P{e) ds is the total energy distribution of emitted electrons relative
to the Fermi level.
Here,230
^ ^ 70 • / ( g ) ' exp(g/^)
['-r^^K-^)'*'] <'"'
with 70 being the emission current density as defined by the Fowler-Nordheim (FN)
Eq. (1.8); d, a factor depending on cp and F and numerically (in eV) equal to
9.76 • 10"^ • F
d = -p^—— [eV]. (3.12)
Era is the maximum energy at a normal to the direction of emission and 4>p the polar angle
in the y, z plane.
The band structure is accounted for by the integral in Eq. (3.11) and for E^/d ^ I
can be neglected. Under normal conditions of FE, d ^ 0.15-0.25 eV, whereas Em for
most degenerate metals with high Fermi energies is a few eV, so in many cases the term
involving the integral in Eq. (3.11) is negligibly small. An exception is transition metals
with partially filled narrow d-bands. In some crystallographic directions, these metals have
Fermi surfaces extending over small intervals thus making the integral in Eq. (3.11) an
appreciable fraction of P{£)?^^ Because exact values of E^ are not known, in approxi-
mate treatments, the contribution of the structural term is usually neglected. In this case,
substituting the Fermi-Dirac distribution into Eq. (3.11) we have.
_ 70 r exp(g/J) 1
P(e) (3.13)
~ d [l-\-Qxp(6/pd)\'
where
(3.14)
d F
is a dimensionless parameter (<^, eV; T, K; F, V/cm). Substituting Eq. (3.13) into
Eq. (3.10) and integrating gives an expression for e first derived by Levine^^^
e = —Ttpd • ctginp). (3.15)
Analysis of Eq. (3.13) shows that the energy distribution becomes symmetrical for
p = \. Distribution functions for this case as well as for the case p{e) = Q is given in
Fig. 3.7.
The condition p{e) = \ allows one to immediately obtain an expression for the
inversion temperature:
y*
' ' • ' ' • ' ' - ' F.
- (3.16)
~ 2K ~
Therefore, assuming that Ecd is close to the Fermi energy Ep, at 7*, energy exchange due
to the Nottingham effect is zero (^ = 0). A plot of T*{F) is shown in Fig. 3.8.
3500
3000
2500 (p = 2.&ey /
/ ^ (p = 4.5ey
^ 2000
V / y
1500
1000 / /
500
1 y 1 J 1 1
6 8 10 12
F, x10^V/cm
Figure 3.8. Variation of inversion temperature T* with the electric field calculated using Eq. (3.16).
The steady-state temperature distribution along an emitter tip taking into account the
Nottingham effect can be solved in the traditional way as it can be assumed that the energy
exchange is confined to the near-surface region equivalent to the electron mean free path in
a metal (10~^-10~^ cm). Then for standard emitters where r^ > 10~^ cm, the contribution
of the Nottingham effect can be accounted for by boundary conditions :^^^
= -H^(Ts) = -71 pd • ctginp) • - , (3.17)

"37 r=re e
where 7s is a temperature of the emitter apex where r = r^.

For typical emitter tip cone angles (a < 30°),
T(r)\r- 7b, (3.18)
where To is the temperature of the tip far from the apex. Neglecting radiation losses, the
equation for thermal balance under steady-state conditions is
r + ?r + 1 : ^ = 0 , (3.19)
r r^ XT
where p is the resistivity constant, XT is the thermal conductivity coefficient, and r is the
coordinate in the radially symmetric one-dimensional problem^^' ^^^ (Fig. 3.9). Integrating
Eq. (3.19) with the boundary condition given in Eq. (3.17) gives
H^{Ts)-j'rl j^rlp ro
(3.20)
iO H
XT'^'^
r
^(•-S)
Figure 3.9. On the one-dimensional model of cathode tip heating by the ¥E current.
Then the temperature difference between the apex of the emitter tip TQ and its base TQ can
be represented by the formula
Ts To = 1 (3.21)
2XT
where p is assumed independent of temperature and equar^^ to p ^ p(0.8 • Ts); ro =

re/ sin(^/2), re is the radius of the emitter tip's apex. In Eq. (3.21), the first term is due to
the Nottingham effect and the second term is due to Joule heating. The relative contributions
of these terms vary with temperature. At low temperatures, tip heating is mainly Nottingham
heating (Fig. 3.10). At higher temperatures, approaching 7*, the Nottingham effect is close
to the inversion point, and the dominant contribution will be Joule heating. This result is
very important because it implies that for metals having melting points below the inversion
temperature, emitter destruction will be the result of Nottingham heating. All metals with
ordinary work function values (4.5-5 e V) that, therefore, require correspondingly high fields
(F > 5 • 10^ V/cm) for high current density electron emission fall into this category, with
the exception of the most refractory metals such as W, Mo, Ta, and Re.
Experimental verification of the Nottingham effect was made by Drechsler^^^ and by
Swanson et al. and Charbonnier et al.^^^' ^"^^ In general, experiment agrees with the the-
ory. It is observed that the Nottingham effect plays a dominant role in emitter heating at
intermediate temperatures. However, Drechsler^^^ and Swanson et al.^^^ observed consid-
erable differences between the experimentally determined inversion temperature and the
theoretical value. This was observed with clean tungsten {(p = 4.5 eV) and tungsten fol-
lowing the adsorption of electropositive atoms (cp = 2.8 eV). The experimental values
were significantly lower than the calculated value—see Fig. 3.11. They^^^ suggest that
the discrepancy is due to a difference between the conduction electron energy £cd and
£F- A difference of just a few hundredths of an eV between Ecd and Fp^^^ can account
for the observed lowering of T*. Details of the mechanism of this process are not yet
clear. Note, however, that with refractory metals, lower 7* values lead to increased emitter
thermal stability due to the earlier inversion of the Nottingham effect and subsequent emitter
cooling with increasing emission current. The effect is more pronounced at lower cp.
3000
^ 2000
1000
2 3
r, x10~^cm
Figure 3.10. Temperature profile across emitter bulk in steady-state regime calculated for a one-dimensional
model.^'^'^ F — l.S • 10^ V/cm, r^ = 10~^ cm. 1: Joule heating and Nottingham effect; 2: Nottingham effect;
3: Joule heating.
2500
2000
^
V 1500
F. M. Charbonnier et al.^^^
1000 <— L. W. Swanson et al.^^^
/
700 ^•* ê— M. Drechler et a\P^
500
3 4 5 6 7 8 9 10
F,xÔ^\//cm
Figure 3.11. Comparison of inversion temperatures calculated using different potential barrier models.^-^-^
1: Image potential; 2: Cutler-Gibbons potential; 3: Zeitz-Vasilyev-Ostrum potential.
3.3.3. Three-Dimensional Analysis of Emitter Heating
The first theoretical analysis of the emitter tip self-heating was carried out by Dyke
et al.^^ who calculated the value of the steady-state maximum current density j ^ for a
one-dimensional model of the emitter tip. Although this theoretical model was not com-
pletely accurate (the Nottingham effect and the dependence of the conductivity p on
temperature were neglected), with arbitrary choice of the average value of r, satisfactory
agreement with experiment was achieved.
Although the dependence of thermo-physical properties of the cathode material upon
temperature were subsequently considered^^^ while solving a one-dimensional time-
dependent problem of the tip heating, the Nottingham effect, which is an important source
of heating, was not accounted for. The time-dependent problem of emitter heating in the
one-dimensional case (with the Nottingham effect included) was first studied by Litvinov
et al.^^^ and later by Nevrovsky and Rachovsky.^^^^
Mitterauer^^^ numerically solved the problem for a tip having the form of a prolate
ellipsoid of rotation. In this case, thermal conductivity was accounted for only along the
principle axis and the temperature and current density distribution were considered homo-
geneous for each cross-section perpendicular to the tip axis. This approach is, in fact, also
a one-dimensional approach.
In one-dimensional models, the most important near-apex region of the cathode is not
considered. Also, the actual distribution of the heat sources in the volume and over the
surface, and the influence of the temperature on these sources are ignored in this model.
Proper treatment requires the framework of a three-dimensional model.
The three-dimensional problem was solved by Glasanov, Baskin, and Fursey.^^^ The
calculations were conducted for a tip of the form suggested by Dyke,^^' ^^^ which describes
well the typical geometry of real FE cathodes and enables one to calculate the field strengths.
These calculations took into account^^^' ^^^ Joule heating, Nottingham and Thomson^^^
effects, and thermal radiation. The temperature dependence of the resistivity pij), heat
capacity c{T), and surface emissivity b{T) were tabulated.^^^
The system of equations describing the kinetics of emitter heating by emission current
includes the thermal conductivity equation
?\T
8 . c ( r ) - - = V [X{T)S/T] + Gir, t), (3.22)
ot
and the continuity equation for the current density in the bulk of the emitter
div7(?,r)=0. (3.23)
Here, T = T(r, t) is the emitter temperature; 6 is the density, k(T) is thermal conductivity,
G(r, t) is the volume density of heat evolution rate determined by the Joule and Tompson
effects, where
G{r, t) = p(T) • f(7, t) + g{T) . (J(r, t), V r ( ? , t)). (3.24)
and g{T) is a temperature-dependent Tompson coefficient.^"^^ j (r, t) is the emission current

density
/(?, 0 = ^ • [VO(?, t) + a(T)VT{r, t)], (3.25)
p{T) ••
where 4>(r, t) is the electric potential, a(T) is a function related to g{T)?^^
dot
g(T) = - r — . (3.26)
clT
Using Eqs. (3.23) and (3.25) one has an equation for the electric field potential 4>(r, t)
inside the emitter
_L_v*(?.„] = -v[-;^.,r,vr]. (3^27)

Boundary conditions for these equations include Nottingham effect and thermal
radiation from emitter surface
(3.28)
an c e
where a is the Stephan-Boltzmann constant; n is an external normal to the emitter

boundary S; 7(Fs, Ts) is the emission current density depending on the emission surface
temperature 7$ and the electric field strength Fs at the surface. siFs, 7s) is an average value
of the energy evolved at the surface per emitted electron. This problem has been solved
numerically using approximate formulae^^' ^^^ for j{Fs, Ts) and 6(Fs, 7s).
In Fig. 3.12, sets of isotherms showing the dependence of temperature on time are
plotted. The dependence of apex surface temperature Ts and the maximum tip temperature
Tin on the time are shown in Figs. 3.13 and 3.14. Calculations showing the kinetics of
heating a tip with axial symmetry with a form approximated by Dyke et al.'^^^ have also
been performed.
The most important result of these calculations is the formation of the overheated core
inside the emitter apex. The temperature at the surface may be well below the melting point
while in the internal region (the kernel), it may be equal to several tens of thousands degrees
kelvin (Figs. 3.12 and 3.13). This leads to enormous temperature gradients, ~10^ K/cm
and the generation of large thermoelastic stresses. The value of tangential stresses at the tip
surface can exceed 2 10^ Pa^^^ that can destroy the tip before the melting point is reached.
yo=0.2-10«A/cm2
f = 4.8nsec f = 7.4nsec
Overheated "corn"
4800.
4400^ 3600 3200
2800
4000 2400
2000
1600
Figure 3.12. Temperature profiles in the emitter tip 238
Figure 3.13. Temperature as a function of time.^-^^ 1: tip surface temperature; 2: maximum temperature in the
emitter bulk.
6000
y = 2-10Â/cm2
5000
^ 4000 -/
3000 4 \
2000 \ 3
1000 - 1
1 1 1
Figure 3.14. Temperature distribution along the emitter axis of rotational symmetry. 1: 1.6 nsec after application
of electric field; 2: 3.5 nsec; 3: 4.0 nsec; 4: 7.4 nsec.
If the energy input rate is great enough, relative to the time required to destroy the tip by
tangential stress, injection of a stream-liquid mixture into the vacuum gap can occur.
The kinetics of the heating process may be viewed in the following way. During the
initial stage, the main contribution to heating of the emitter is due to the Nottingham effect,
since the temperature of any point of the surface is lower than the inversion temperature.
Eventually, the temperature of the emitting surface near the apex of the tip becomes greater
than the inversion temperature, which induces Nottingham cooling and leads to a displace-
ment of the temperature maximum from the surface into the bulk. Simultaneously, the
rate of rise of the surface temperature decreases. However, complete stabilization of the
temperature at the apex is not achieved.
3.4. BUILD-UP OF THE FIELD EMITTER SURFACE AT HIGH CURRENT

DENSITIES: THERMAL FIELD SURFACE SELF-DIFFUSION
The efficiency of FE cathode heating increases with the appearance of micro-

protrusions due to heat localization. This localization causes a further evolution of the
micro-relief, a resulting local increase in field strength and thus FE current density and,
finally, explosion of one or several of the resulting "micro-tips", destruction of the FE
cathode, and the development of vacuum breakdown.
It has been shown^^' ^^ that at high current densities and with sufficient pulse durations,
rearrangement of the emitter surface takes place, whereby micro-protrusions and sharp
ridges are formed.^"^^ The activation energy for surface diffusion is due to the heating of
cathode by the electron current. If the electric field, and correspondingly the emission
current, pulse duration is hundreds of microseconds to a few milliseconds or longer, then
the rearrangement will always begin before the tip body heats to the critical temperature and
the cathode explodes. The effects are cumulative as each subsequent high current density
pulse creates protrusions, further enhancing the rearrangement effect.
(c)
• 4
3
2
(7^ 1
Figure 3.15. Surface migration withfieldemitter tips heated by the emission current.^ ^' ^^ (a) and (b) FE patterns
in the initial and final stages, respectively; (c) increase in the FE current during square wave voltage pulses of
various amplitudes; (d) shadow-micrograph of the tip after current pulsing.
In earlier studies, this rearrangement process (diffusion) could not be detected directly
because under such conditions, space charge is important and thus the image contrast is low.
To study the build-up effects, after applying the maximum current density, "frozen" traces
of the build-up were observed at lower current densities (^10^ A/cm)—Fig. 3.15. In the
experiments^^' ^^ FM Miiller microscope was used to examine W, Mo, Nb, and Ta cathodes
with tip radii of ro ^ (1-5) • 10"^ cm. It was found that the degree of emission-induced
migration depends on the duration of the current pulse. Migration can be enhanced by
repeated application of a voltage pulse for fixed durations, which causes "accumulation" of
surface micro-roughness. As the surface micro-relief is changing, an appreciable increase
in the emission current at fixed applied voltage occurs in the low current region. The rate of
the current enhancement depends on the maximum current density and the initial emitter
^ r=850K
25
20 1 1
AJ=3.2-10Â/cm^ 1 1 7-_300K
•J=3.05-10Â/cm4 1 •
15
10 1 7J 1
1 1 1 1 t 1 1
2 3 4 5 6 7
n
Figure 3.16. Variation of the FE current as a function of the number of appHed voltage pulses for different values
of; and T}^^ Tpuise = 100|jisec.
temperature (Fig. 3.16). The increase in current as a result of repeated application of voltage
pulses of the same amplitude continues until breakdown sets in.
Emission-induced migration is a key factor determining FE instability at maximum
current densities with atomically clean cathode surfaces over a wide range of times (t ^
10~^-10~^ sec). The maximum current density is limited not only by thermo-physical
parameters of the emitter bulk but to a greater extent by surface self-diffusion, a process
controlled by the migration activation energy.^^' ^^' ^^^
3.5. THE HIGHEST FIELD EMISSION CURRENT DENSITIES

ACHIEVED EXPERIMENTALLY
As shown above, the FE current density attainable in practice is limited by heating of the
emitter tip. A number of experimental attempts to circumnavigate the limitation caused by
the field emitter heating have been conducted. These attempts were based on the following
ideas.
1. As heating is essentially an inertial process, the maximum current density can
be increased by applying electric field pulses of very short duration. In this way,
current densities up to 10^ A/cm^ were achieved using pulse durations of a few
nanoseconds.^^' ^^' ^^^
2. Reduce Joule heating by using superconduction emitter tips. It was found that
the superconductivity was destroyed by the high current densities. By cooling to
temperatures of 2-4.2 K, jmax could be increased two- or threefold.^^^
3. Current densities of 10^-10^^ A/cm^ were obtained in experiments with small

emission spots^^'^^ and tips 10-30 A in size.^^'^^^ Because of small size of
the field emitter, the electrons traverse it without scattering, so that no energy
is dissipated by Joule heating. This will be discussed further in the next section.
3.5.1. Experimental Current Density Values
The limiting current densities obtained by various authors over the past years are
presented in Table 3.1.
Table 3.1. Limiting current densities

Material r, sec y'max, A/cm^ Authors
w do mode 10/ Martin et al.^^ and Dyke et al. (1960)^^

w IQl-lQ-^ 2 10^ Fursey and Kartsev (1970)^^
Mo 10-3 8 10^ Fursey et al. (1984),^^ Krotevich et al. (1985),^^ and
Nb 10-3 3- 10^ Krotevich (1985)^^1
LaB6 3•10-^ IO'7-IO^ Elinson and Vasiliev (1954)^^^ and Elinson and Kudnitseva
ZrC 3•10-^ IO'7-IO^ (1964)252
Ta 4.10-^ 5 10'7 Fursey and Tolkacheva (1963)^^^
Re 4•10-^ (3-5) • 10'7 Fursey (1964)^^^
W 10-7 3 10^ Mesyats and Fursey groups53,54
W 10-^ (5-6) • 10^
W 10-9 10^
3.5.2. The Limiting Current Densities Attained Cooling
It is known that the maximum attainable FE current density is limited by Joule heating
of the lattice and the Nottingham effect. Whereas the Joule heat is released throughout the
bulk of the cathode, the heat produced by the Nottingham effect is localized to the near-
surface region whose thickness is of the order of electron-phonon free path ^X&p. Above
the Debye temperature, this distance is of the order of 100 A, that is, the Nottingham effect
is a surface phenomenon. Lowering the emitter temperature causes a significant reduction
in Joule heating and a considerable increase in Aep, which shifts Nottingham heating to a
bulk effect. Therefore, the maximum FE current density can be increased by decreasing the
emitter temperature.
By the Nottingham mechanism, after an electron is emitted, the remaining conduction
band hole will scatter electrons into states above the Fermi level.^^^ These electrons will
theimalize in time tee, which is a characteristic time of the electron-electron interaction.
After time tep, which is a characteristic time of the electron-phonon interaction, the energy
will be transferred to the lattice. Note that the electron thermalization can only occur if
Tee < Tep. At low temperatures, the thickness of the region in which energy transfer to the
lattice takes place (Aep ~ vpêp* ^ being the velocity of electrons on the Fermi surface),
can be as large as a fraction of a millimeter in pure metals.
The complete theoretical analysis of such a complex situation is possible only in terms
of kinetic theory. However, assuming tee < êp, a metal can be considered as a combination
of two subsystems, one of electrons and the other of lattice atoms.^^"^
The energy balance equations for these subsystems have the form:
371
cc'^ = V(/e • T,) - A(Te. Ti) + 7fe • f(r), (3.29)
01
Ci-~= V(xi • Ti) - A(T„ 7i). (3.30)

ot
Here, Te, Ce, and Xe are the electron temperature, electron heat capacity, and thermal con-
ductivity, respectively; 7i,Ci, and Xi are respective quantities for the lattice subsystem;
A(re, 7i) is a function describing nonlocalized transfer of energy from the electron subsys-
tem to the lattice, yVe is FE current density. The last term in Eq. (3.29) takes account of the
Nottingham energy released in the electron subsystem over the length Aee •
Despite the significant approximations in such an approach with an appropriate choice
of A{TQ,T{) and / ( r ) , it can adequately reproduce the main features of the phenomenon:
the spread of the electron-electron interaction over length Age and the electron-lattice
interaction over length Aep. Functions A{TQ, Ty) and f{r) were taken in the form:
A(re, Ti)=aju [re(3c), TAX)] • exp { " ^ y ^ } ' ^^^^ (3-31)
/(r) = ^.expj-^}, (3.32)
where U{Tt, TO is the usual expression for the rate of local energy transfer to the lattice;^'*^
a, and p are constants derived from the condition that for A,ep -^ 0, Eqs. (3.31) and (3.32)
transform to local relationships;^^^' ^^^ I is a distance from point r to the emitting surface.
Boundary conditions for the system of Eqs. (3.29-3.32) are the temperature stability inside
the emitter and the absence of temperature gradients along the normal n to the emitting
surface S:
Ti{r -> oo) = Teir -> oo) = 7b, (3.33)

dTi _ are
= 0, j\s=J(Tc,F) (3.34)
'dn s ^« s
with initial conditions
T.(t = 0) = Te(t = 0) = To. (3.35)
where 7b is the initial temperature, J(TQ,F) the known function^^"^ (see Eqs. (1.11)
and (1.12), Chapter 1) of the electric field strength F and TQ on the emitting surface.
If spreading of only the electron-electron interaction is taken into account and A is
taken in the form A(re, TJ) = V(TQ - Ti), then neglecting the electronic heat capacity, we
can obtain for a one-dimensional cathode:
B exp{-v7$}
Ti = -exp{-§} (3.36)
\-y
where y = A^gV/xe; ? = X/XQ^; B = Ktj^/Xt\ ^ = êm - ^F is the average Nottingham

energy per electron released in the electronic system during tunneling.
If Aee » (Xe/v^)^^^ = <5, then TQ-T\ = {£ ' 7/vA,ee)exp(-jc/A,ee). Conversely, if
>.ee « 5, Te - Ti = (£ -j/v^) e x p ( - x / 5 ) .
As can be seen from these relationships, for Xee <^ 5, the characteristic size of the
heated region in the electron gas is ~ ^?^^ For Aee ^ 5, the heated region is much larger
and the electron gas temperature much lower than in the former case. As a consequence,
the lattice heating rate is much lower:
£'J
Ti - To (3.37)
Aee + ^
So, accounting only for heat spreading over the length Xee results in a much lower lattice
heating rate.
The system of Eqs. (3.29-3.34) has been integrated numerically for a cone-shaped
emitter. The calculation was carried out for TQ = 4.2 and 300 K. The magnitude of the
electric field F was chosen such that a steep current rise occurred during a time interval of
10 nsec.
Figure 3.17 shows the calculated time dependence of the emission current density. As
seen in the figure, the maximum current density at To == 4.2 K is a factor of 2.8 greater than
at To = 300 K. With increasing Aep, this ratio increases somewhat but Xep cannot be much
in excess of ro/ao (ro is the emitter radius and ao is the emitter cone angle).
An experimental study was carried out on tungsten single crystals cooled to liquid
helium temperature (4.2 K).^^^ The sample was placed in an evacuated tube and was
Figure 3.17. Calculated variation of the maximum current density jemiO' ^'- TQ = 4.2K, 2: TQ = 300K,
ro = 0.2^-m, ao = 20°.
Voltage
Figure 3.18. Schematic of the experimental apparatus. 1: cathode tip; 2: collector; 3: screen; 4: grid; 5: insulator;
6: matched coaxial screen; 7: Dewar vessel with liquid helium.
immersed in liquid helium (Fig. 3.18). The residual gas pressure in the working space
of the device after immersion in liquid helium was 10~^^-10~^^ Torr.
A modified Muller microscope permitted observation of the emission images and inves-
tigation of the FE characteristics in a nanosecond range. Most experiments were carried
out with pulse durations of 10 nsec or in the dc mode. The tip surface was cleaned by
heating to high temperature in the cryostat. The emitting area was determined using known
methods^^ and the emitter radius was derived from current-voltage characteristics or from
electron-optical images of the tip profile.
With samples cooled to 4.2 K, the maximum current density that could be extracted
without damage to the tip was two to three times that could be extracted at room temperature.
Measured current and current densities for various samples at 300 K and 4.2 K are shown
in Table 3.2. The experimental results are in fair agreement with the theoretical model
presented in the beginning of this section.
Table 3.2. Measured maximum current and current densities for various
samples
To = 300 K To = 4.2 K
re, |xm /m,A 7m,Acm"2 r e , |JLm /m,A ;m, A • cm ^
0.28 1.15 3.7 • 10^ 0.14 0.79 10.1 10^

0.29 1.07 3.2 • 10^ 0.15 0.73 8.1 10^
0.37 1.16 2.1 . 10^ 0.18 1.17 9.0 10^
0.38 1.53 2.6 • 10^ 0.25 2.0 8.0 10^
1.0 7.0 1.75 • 10^ 0.29 3.39 10.1 10^
3.5.3. Current Densities from Nanometer-Scale Field Emitters
The highest current densities we obtained when emission was localized to very small
(nanometer scale) areas of the emitter surface. Two types of experiments have been con-
ducted. In the first type,^^' ^^ zirconium (Zr) was deposited onto the surface of an ordinary
tungsten emitter of size 10~^ cm (Fig. 3.19). In the second type,^^' '^^^' ^^ nanometer-scale
micro-protrusions having characteristic dimensions of 10-20 A were formed on the emitter
as a result of surface migration of W atoms in a strong electric field. In both cases, the
local temperature rise, for a given emitter current density, was greatly reduced due to the
cooling action of the remainder of the emitter tip. Note that with nanometer-scale emitting
areas, there is practically no preferential energy deposition in the emitting region because
its size, ~10~^ cm, is substantially smaller than the electron-photon free pat. This means
that the heat produced by the current flowing through the emitter tip is distributed over a
much larger volume than that associated with the emission site alone.
The first experiments with selective adsorption of Zr on W were done by Fursey and
Shakirova.^^ Emission patterns of Zr on W (001) are shown in Fig. 3.19c. The emission
sites in Fig. 3.19c have radii of ~40-50 nm whereas the overall tip radius is 300-400 nm.
Ultra-small emitting spots, with radii of 10-15 nm (Fig. 3.19d) were formed by deposit-
ing a Zr-ZrOAV coating in an electric field (^^5 • 10^ Y/cm)}'^' •^^^ The smallest locaUzed
emission sites achieved had radii of ~ 3 nm. The dimension of the site was determined by
comparing its image with that of the entire tip for which the radius had been determined
independently.
The maximum current density extracted from the type of site shown in Fig. 3.19c in
the pulsed mode (2-5 jxsec duration pulses) was ~5 • 10^ A/cm.^^^ For the sites shown in
Fig. 3.19d, dc current densities up to 10^ A/cm^ were obtained.^"^
The first experiments with micro-protrusions were carried out by Shrednik's group.^^
In his study, dc mode current densities up to 10^ A (in some experiments up to 10^^ AJcrx?)
could be obtained. Recent experiments^^^ have shown that if very small isolated micro-
protrusions are created on the surface, current densities greater than 10^^ A/cm^ and up to
10^^ A/cm^ could be achieved. Illustration of current measured at the nanotip by the probe
is given in the Fig. 3.20.
Theoretical analysis^^^ of emitter heating when the characteristic electron-phonon free
path Aep is much greater than the size of the emitter has shown that for 2rem/êp < 0.6,
(b)
(d)
Figure 3.19. Nanoscale emission area on the field emitter surface.^^ (a) Initial FE image of W (Oil); (b) Image
of a single microprotrusion on W (011) generated by the build-up process; (c) localization of FE due to selective
Zr adsorption on the (001) face of tungsten; (d) image of an ultra-small emission site on the (001) face of tungsten.
Figure 3.20. Probe experiments for W {011} (a) initial FE image; (b) image of nanotips on the surface after
build-up.
current densities j > 2 x 10^^ A/cm^ can be obtained in dc operation without destruction
of field emitter.
Summary of the data for extremely high current densities is shown in Table 3.3.
Table 3.3. Summary of data for extremely high current densities

Record current
Conditions Material densities, A/cm^ Authors
Nanosecond duration W 10^ Fursey et al. (1969),^^

(Vise = 5 nsec, T = 300 K) Kartsev et al. (1970)^^
Deep cooling W 109 Aksyonon et al (1979)^24
(>lse = 10 nsec, T = 4.2 K)
Small emission spot W + Zr 5 109 Fursey and Shakirova
(spot diameter d ^ 100 A, (1966)^^
Vise = ^ i^sec, T = 300 K)
Super small emission spot W-f Zr 10^ Fursey etal.(1995)2'^
(d = 20-30 A, dc mode, T = 300 K)
Nanotips (dc mode, T = 300 K) W lo^-ioio Pavlov et al. (1975)^^
Super small nanotips w lOlO-lO^l Fursey et al. (1998)^22
(emitter radius rem = 10-20 A,
Vise = 100 msec, T = 300 K)
3.6. RESUME
FE is a unique electron emission mechanism, in which no energy is required for

emission:
1. Theoretically the entire flow of electrons incident on the solid/vacuum interface
from inside the solid to be emitted. The theoretical limit is 2-3 x 10^^ A/cm^.
2. In practice, the maximum current density is limited by Joule and Nottingham
processes in the near-surface region.
3. For current pulses having a duration of 10~^-10~^ sec, thermal destruction of
the emitter can be accelerated by thermally activated, field-assisted surface self-
diffusion, and formation of micro-protrusions on its surface.
4. It has been experimentally shown that under special conditions, FE current
densities can approach 10^^ A/cm^, close to the theoretical limit.
FIELD EMISSION IN MICROWAVE FIELD
4.1. INTRODUCTION
Field emission (FE) tips are promising in various microwave cathode applications.
From a fundamental point of view, electron emission in microwave fields has a number of
unusual characteristics.
The main features of FE in a microwave field are:
1. Electron tunneling occurs during very short time intervals ensuring adiabatic
conditions up to frequencies of about lO^'* Hz.
2. The dependence of electron emission on the electric field is nonlinear. Conse-
quently, there is an opportunity to perform nonlinear operations (such as frequency
transformation and multiplication, and microwave field power conversion), and
directly produce electron bunches during the emission process.
3. Ion motion in microwave field reduces ion bombardment of emitter surface.
4. The electron energy distribution shows unique features.
An advantageous property of FE is that it can be switched on and off virtually instan-
taneously, due to the fact that no heating of the emitter is required for emission as it is with
thermionic electron emission. Also, the cathode as an element of microelectronic system is
radiation hard and its performance is temperature insensitive. Field cathodes can be used
in high-power, high-current microwave field devices, and electron optical systems: mag-
netrons, energy conversion devices, high-voltage and high-resolution electron microscopy,
and so on.
In this chapter, we discuss the physical mechanisms of the FE process and the resulting
electron beam behavior in a microwave field. One of the few previous reviews of these
topics is by Charbonnier et al.^^ There are also papers by Russian authors,-^^"^' ^^^ which
are not generally known to foreign researchers. Practical applications of FE in a microwave
field have been briefly reviewed by Brodie and Spindt^^ and Busta,^^ and discussed at a
number of conferences by Gray^^^ and Smith and Gray^^^ and, in more detail, by Parker^^^
and Gulyaev et al.^^^
57
4.2. THE ADIABATIC CONDITION—TUNNELING TIME
In the presence of an ac field, the potential barrier between the solid surface and vacuum
may change so rapidly that tunneling occurs in a time interval that is long relative to that
during which the potential barrier at the surface changes. In this case, the FE process is
no longer adiabatic and Fowler-Nordheim (FN) theory (which is based on the stationary
Schrodinger equation) is no longer valid.
Adiabatic conditions as they relate to FE processes were first analyzed by Bunkin and
Fedorov.^^^ They showed that for frequencies up to roughly 10^^ Hz, the current density is
determined by the conventional FN equation with the intensity of electric field replaced by
its instantaneous value.
A "physical" formulation of adiabatic conditions can be done in terms of tunneling time,
Ttun- The tunneling time is assumed to be the mean time of interaction between a tunneling
electron and the potential barrier. This time parameter determines specific properties of the
tunneling process.^^^' ^^^' 260-262 p^j. instance, when a tunneling electron passes through a
time-dependent barrier, the characteristics of the tunneling process can vary dramatically
if the conditions are not adiabatic.^^^' ^^^ The role of the tunneling time rmn is important in
tunneling with dissipation, wherein the electron "feels" the under-barrier "friction,"-^^^' ^^^
and also in image potential calculations.^^^^ It can be assumed that the electron motion
becomes adiabatic, that is, the particles tunnel through a static potential V(t), if the rate of
change of the applied field is small relative to the tunneling time: (i>|rtunl ^ 1. where co is
the frequency of the field oscillations.
The problem of tunneling time was first formulated by Wigner.^^^ To determine the
time, he described the particle as a wave packet. Scattering of such a packet by a spherical
potential with a finite radius was considered, and the tunneling time was defined as the
time interval between when the centers of mass of the incident and scattered wave packets
crossed the surface of a sphere of radius R:
2R 2d
Tmgncv = — -\-- — [S(k)l (4.1)
V V dk
where v and k are the velocity of the incident particle and the wave vector respectively, and
8(k) is the phase shift between the incident and reflected waves.
This approach was used by Hartman^^^ to describe tunneling through a rectangular
barrier. Here the tunneling time, rnartman, was calculated as the time interval between when
the maximum in the wave packet passes through the left side of the barrier and the formation
of the maximum in the transmitted wave packet at the right side of the boundary:
o
rHartman = h--[8i{k)l (4.2)
dE
where î (k) is the phase difference between the wave function ^ ( 0 near the barrier bound-
aries for a value of the particle energy E. It was shown later^^^' ^^^' ^^^ that such an approach
was not quite correct, because the wave packet was deformed in the barrier region. This
imposes conditions on the shape of the packet and on the barrier thickness, which can-
not be easily combined with one another. Hartman has shown^^"* that in some cases, the
calculations yield a negative tunneling time itun, which has no physical sense.
FIELD EMISSION IN MICROWAVE FIELD 59
An approach to calculating the tunneling time was suggested by Baz.^^^' ^^^ It is based
on a gedanken experiment related to the precession of tunneling electron spin in a weak
magnetic field. Calculating the angle of spin rotation, 0, for the transmitted wave by solv-
ing the stationary Schrodinger equation on the assumption that the spin orientation of the
incident wave is known, one gets the tunneling time, TBaz, as:
rBaz = — , (4.3)
COL
where coi^ is the Larmor precession frequency.

Baz's method was applied to the case of tunneling through a one-dimensional potential
barrier.^^^ The spin rotation angles for the transmitted wave in a plane perpendicular to the
field were determined. Buttiker showed^^^ that in the quantum case there are two angles
of spin rotation in a magnetic field: the first angle (p± in the plane perpendicular to the
magnetic field, and the second one (p\\ in the plane parallel to the field. It so happens that it
is the time
^Buttikerz = , (4.4)
(OL
that is equal to the tunneling time defined for a rectangular barrier in the quasi-
classical case.^^^
The most complete and accurate treatment of the tunneling time problem was by
Sokolovski and Baskin.^^^ They found that the time the electron interacts with the
potential barrier can be calculated by Feinman's procedure of integrating along classical
trajectories.^^^ If the wave function ^(r) corresponds to a state with a determined energy,
the mean traversal time can be represented as
TQ(E) = {t2-ti) f |vI/(r)|Vr, (4.5)
where ti and t2 are the times at which the particle is detected in its initial and final points,
ri and r2 respectively, in a finite region Q.
It was shown that the tunneling time r^ is a complex number, and, in this sense, is not
observable. Nevertheless, it is intrinsically connected with the observable angles of spin
rotation due to Larmor precession and with the adiabatic parameters of the time-dependent
problem.^^^ The relation between r^ and the time derived by Baz's method^^^' ^^^ can be
showni^^'2'73,274jQ^g.
(p± = ft)LRe[r^(ri, r2, r)] (4.6a)
and
(p\\ = a;Llm[r^(ri, r2, r)], (4.6b)
where for the whole angle of rotation,
(P = ((PI-\- (^^)^/2 = coL\TQ{rur2, r)|. (4.6c)
A calculation of the tunneling time rtun through a one-dimensional barrier at a metal-

vacuum boundary determined byU(x) = Uo—eFx—e^/(l6nXQX) shows^^"* that for a wide
potential having a low transparency, Im[rtun(£^)] ^ Re[rtun(£^)]. Thus, it is sufficient to
find the imaginary part of rtun(£^) to estimate the tunneling time. For Uo = lOeV,E = 5 eV,
P changing from 0 to 0.25 {F changes from 0 to 1.6 V/A), the term Im[rtun(£^)] can be
approximated by the expression:
Im[rtun(£)] ^ 1.5[mi/2(ô - E)^'^\ /{eF) . (4.7)
Re[rtun(£^)] is obtained from the expression:
Re[rtun(^)] = 0/2)h/(Uo - E). (4.8)
¥OTF ^ \ V/A: Im[rtun] ^ 10"^^ sec and Re[rtun] ^ 2 x 10~^^ sec.

To conclude the discussion on tunneling time, we note that at present all the tunneling
time estimates for FE yield values ~10~^^ sec. The tunneling time Ttun is a fundamental
parameter defining the applicability of FN theory to the case of microwave field excitation.
Quantitatively, the applicability criterion is expressed through the adiabatic condition:
y=a;|rtunl«l. (4.9)
For practical devices, the parameter y actually determines the upper frequency limit of
~ 10^"^ Hz. This is the fundamental frequency limit for devices based on the direct interaction
of an electromagnetic wave with a field emitter surface, for example, in resonators, etc. In
practice, the actual frequency limit for devices is determined by such parameters as the
interelectrode capacitance, inductance, and resistance.
4.3. EXPERIMENTAL VERIFICATION OF FOWLER-NORDHEIM THEORY
The first experiments demonstrating the validity of the main principles of static FE
theory in the microwave fields were conducted by Dyke's group.^^ Current-voltage char-
acteristics, retarding potential curves, and some conclusions from the FN theory associated
with nonlinearity of the emission process were studied.^^ The incentive for this investiga-
tion was the application of FE cathodes to microwave power amplification and frequency
conversion. Frequencies up to 10 GHz were studied.
It was shown^^' ^^^ that the field electron image of a tungsten (W) emitter is similar
to that obtained in a static field. In more quantitative experiments, the field emitter was
located directly in the gap of an X-band cavity. Assuming that the average emission current
is then measured as a function of the microwave field input power Pin, the dependence of
the logarithm of average current (/) on the input power Pin should be equivalent to the
linear dependence ln(//y^) = f{\/U) predicted by the FN theory. This dependence has
been demonstrated^^ over a range of the emission current varying by at least two orders of
magnitude.
More accurate methods for the study of FE in microwave fields have since been.^^' ^^^
Such studies have combined FE microscopy in microwave fields, measurements of the
dependence of the emission current density onfieldstrength, and the investigation of the time
dependence and of the retarding potential curves over time intervals of 10~^-1 sec. Also,
dc measurements on the specimens could be conducted. The range of emission currents
investigated were increased by two to three orders of magnitude over those previously
utilized^^' ^^^ and the vacuum in the experimental chamber could be varied between 10~^
and 10"^^ Torr. Experiments^^ were performed on W and W2C emitters with ^0.1 |xm
radii of curvature and on liquid metals either contained in narrow capillaries or as thin
layers at a solid needle substrate.^'^^' ^^^
The experiments were conducted using the arrangement shown schematically in
Fig. 4.1. The FE cathode was located in the gap of a microwave prismatic resonator (//103
mode, natural frequency 4.7 GHz). The oscillation wavelength was 6.4 cm, the input power,
Pin, was less than 1.5 x 10^ W. The direct current components of the cathode current /cath
and the current of a beam collector /cou placed in the beam drift space outside the resonator
were measured.
The results of static and microwave field investigations were correlated with one
another. The dependence ln(/) = /(P~^/^) and the FE image obtained in a microwave field
are shown in Fig. 4.2a. One can see that the plot is linear over four orders of magnitude.
An equivalent voltage in the resonator was calculated^^' ^^^ exactly for the well-defined
resonator parameters. The transition from Pin to U permitted us to obtain a generalized
FN dependence that is valid for both static and microwave fields (Fig. 4.2b). The small
divergence from the linearity in the slopes is related to errors in the definition of equivalent
voltage U.
The current in these experiments was varied between 10~^^ and 10"^ A in the static
measurements and between 10~^ and 10~^ A in the pulsed, quasistatic regime. In the
microwave fields, an average current for a pulse was measured. This average current varied
in the range 10~^-10~^ A. The peak currents /peak were approximately ten times greater than
the average values. Current densities were determined from the expression / = / / ( o r ^ ) ,
where a is a coefficient linearly dependent on the electricfield.^^^Forfieldsbetween 3x10^
and 10^ V/cm^, the value of a is 1-2.5 for a surface work function of 4.5 eV. We can see that
uj\^
Figure 4.1. Microwave apparatus:^^^ 1: Modulator, 2: excitation device, 3: attenuator, 4: power amplifier, 5: ferrite
filter, 6: directional coupler, 7: commutator, 8: power measuring device, 9: cavity, 10: diaphragm, 11: screen,
collector, 12: grids, 13: vacuum chamber, 14: optical window, 15: HE cathode.
(a) (b)
Microwave field
-35
5
^ -4
c
-40
-45
1.0 1.3
Figure 4.2. (a) Experimental emission characteristics for a W emitter in a microwave field, (b) The generalized
FN dependence for static and microwave fields.^^' ^^^
these experimental results indicate that the FN law is applicable to FE from dc to frequencies
up to at least 10 GHz.
4.4. MAXIMUM FIELD EMISSION CURRENT DENSITIES
The maximum current densities, jâx^ which may be practically achieved with a
field emitter, are limited by emission current heating induced thermal destruction of
field cathode.^^'^^ Recall that the maximum current obtained with W field emitters is
^10^ A/cm^ for a static field, ^ 5 x 10^-10^ A/cm^ for microsecond pulses, and up to
10^ AJcrr? in the nanosecond duration pulse range^^' ^^ (see Chapter 3).
Maximum allowable current densities in a microwave field have been
measured.-^^^'^^^'-^^^ With peak electric field values of 10^ V/cm, maximum values of
the average current density for pulse durations of about 10-100 |x sec were 5 x 10^ to
10^ A/cm^, and maximum values for a single pulse with a duration of ~10~^^ sec was
10^ A/cm^.
Such values of current densities are greater than those obtained under more lenient
conditions in static pulse fields. It should be noted that the character of energy dissipation
in the microwave field may have some peculiarities. The point is that the extraction of
current in a microwave field takes up much shorter time (a time half a period of field
oscillation) due to a nonlinear dependence of emission current on the field strength. For a
frequency of 4.7 GRzP^' ^^^ this time may be as small as 10~^^ sec, which is shorter than
the electron-phonon interaction time. Therefore, there may be a much smaller evolution
of heat by the emitter than in static field. The difference in allowable current density does
not appear to be much different: microsecond pulses ~10^ A/cm^ in the static pulsed case.
and approximately the same current density in the microwave field with 10 |xsec duration
pulses. Also, at 4.7GHz, the one time is 10~^^ sec for / ~ 10^ A/cm^. This / is similar
to that obtained for nanosecond duration pulses in the "static" case.
4.5. ION BOMBARDMENT OF THE CATHODE
The main cause of the instability of field cathodes is build-up processes on the surface
in a strong electric field due to surface migration.^^ Virtually any mechanism activating the
surface migration processes (ion bombardment, self-heating of the emitter by the emission
current, etc.) may give rise to the appearance of microtips on cathode surface. The increased
field at this microtip leads to increased emission current, cathode heating, and the eventual
development of a vacuum arc.
Under practical vacuum conditions, surface migration is activated by ion bombardment
of the cathode by ions formed from the residual gas atmosphere due to electron impact
ionization. Early experiments with FE cathodes operated in microwave fields showed that
emission stability was substantially higher than that in staticfields.^^^-^^Êvidently, the
reason was a radical change in the character of ion bombardment in the microwave field.^^
The field distribution near the apex of the cathode tip is characterized by high gradients:
about 10^1-10^2 v/cm^. While moving in an ac field with such a gradient, the ions "feel"
a force directed toward the region of the lower field. This force causes the ions to drift into
regions of decreasing field. As the field distribution near the apex of the emitter tip is nearly
spherically synmietric, the equation of ion motion in an ac field of frequency CD is
: 7 ^ - - ^ - f sinM), (4.10)
where FQ is the field at the cathode surface, r^ the field emitter tip radius, r the radial
spherical coordinate, and e^ and m{ are the ion's charge and mass, respectively. Using
Kapitsa's method,^^^ it can be shown^^ that an ion of an energy £", less than £'max, never
reaches the cathode surface where
_ (2-10/3(^./ro^2)2/3(^.^)l/3 ^ ^ ^.
^max — 1 2/ 17x2// 2\ \^-^^)
and a){ = (leiFo/niiro)^^^. The dependence of £max on co is shown in Fig. 4.3a. The ion
trajectories calculated by numerical solution of Eq. (4.10) for ions O"^ having an initial
energy of 10 eV moving toward the emitter at different angles are shown in Fig. 4.3b.
The effect of eliminating ion bombardment in microwave fields has been observed
experimentally^^ by comparing emission stability in microwave and static fields at a residual
gas pressure of 10~^ Torr.
The experiment was aimed at comparing, under adequate conditions, the video pulse
regime, having ion bombardment inevitably present, with the radio pulse regime, where
the ejection of the ions from field nonuniformity occurs. To this end, the change of FE
current with time was investigated in the both cases. The length of video and radio pulses
was varied within a range of 1 0 - M o - i sec, and the pulse repetition, between single-shot
(a)
200
>
100 h
Figure 4.3. (a) Dependence of threshold value of the ion energy £max upon the microwave field frequency a>. (If
E < En , ions do not bombard the cathode surface); (b) trajectories of O"*" ions; ro = 0.5 ixm, &> = 3 x 10^" s
Fo = 6 X 10'7 V/cm.
pulses and a frequency of 10"^ Hz. The radio pulse stuffing frequency was 4.7 GHz. Vacuum
was varied between 10~^ and lO'^Torr. Test measurements to distinctly accentuate the
"ejection effect" were carried out in a vacuum 10~^ Torr.
A significant difference in the emission current behavior for the microwave and static
field was detected (see Fig. 4.4). In the microwave field, one can see that after an initial
decrease in the emission current, due to adsorption processes, the current remains stable
for more than 100 hr (Fig. 4.4, curve 2). In addition, the emission image did not change.
For video pulses, an abrupt current rise was observed after a few minutes of operation at
10"^ Torr (Fig. 4.4, curve 1), and mobile emission centers were observed in the emission
image (Fig. 4.4b). Irreversible changes in the emitter microgeometry and the emission
current lead to explosion of the tip and vacuum breakdown. Besides facilitating devices
applications, the absence of ion bombardment in microwave fields offers new opportunities
for studying adsorption processes with a time independent emitter tip geometry.
4.6. ELECTRON ENERGY SPECTRA: TRANSIT TIME
Electron spectroscopy in an ac field reveals some unusual features in the energy dis-
tributions that are a result of the time varying field. It is very important to quantify the
energy distribution of the electron beam in order to analyze the efficiency of a series of
devices for which the FE cathode is located directly in the resonator. This problem was first
studied by Dyke's group.^^ Retarding potential curves for the electron beam at the output
of the resonator indicate that when a high static field is employed in combination with a
microwave field transit time effects are negligible.^^
Q
10
Figure 4.4. Field emission current stability for WC tip: 1: video pulse regime; 2: radio pulse regime. (The pulse
current, /Q, is 200JJLA for both regimes, pulse length, T, is 100 p.s, pulse repetition frequency is 100Hz, radio
pulse stuffing frequency is 4.7 GHz, and the residual vacuum pressure is 10~^ Torr.)
Transit effects have been investigated theoretically and experimentally.^^^' ^^^ Com-
plete calculations of the energy distributions for the ac fields are difficult, even using
numerical methods. This difficulty is due to the large difference between the tip radius
ro = 0.1-1 |xm and the vacuum gap length h = 1-4 mm. Within a paraxial approximation,
the problem can be solved. The calculated energy distributions of electrons at the resonator
output are shown in Fig. 4.5a. Calculations show that the qualitative character of the curves
is determined by the parameter ai = coH/vi. Here, i;i is the maximum electron velocity in
the near-emitter region (where the field gradient is high), and H is the length of the region
with a homogeneous electric field distribution, FQ. The order of magnitude of parameter
Qfi is equal to the transit phase of the emitted electron. The influence of the parameter
^2 = vi/iôsc on energy distribution is not very significant. Here, VQSC is the velocity of
oscillated electrons in the homogeneous resonator field.
Using special techniques,^^^ we have investigated the energy distribution of electrons
in microwave fields. In our studies, the gap h was 1-4 mm, the frequency was 4720 MHz,
and the emitter radius was 0.1 |xm. Retarding potential curves are shown in Fig. 4.5b. Uo
is the potential at which the collector current becomes equal to zero, and /coU, max is the
collector current at a retarding potential f/ret = 0. The homogeneous field strength FQ in
the resonator determined by the input power was less than 2x10"^ V/cm. Curve 1 shows the
measurement made in the static regime with an experimental error of less than 5 percent.
The poor resolution is still sufficient for analysis of strong transit effects. The effects of a
finite transit time on the energy distribution functions are shown in curves 2-4 in Fig. 4.5b.
Here, ai varies from 7c/2 (curve 1) to n (curve 4). One can see from curve 2 that at
o?i = 7r/2, only one group of electrons exists with the energy in the range of 0.56:max»
where sâx is the maximum electron energy at the resonator output. The half width of
energy distribution is equal to 0.2emax in this case (see curves 2 and 3 in Fig. 4.5a). When
0.2 0.4 0.6 0.8 1.0
Figure 4.5. (a) Electron energy distribution function at the resonator output (tungsten cathode; FQ = 6 x
lO'^ V/cm).^^"^ 1: ai = 0, ^2 = 50; 2: ai = n/2, a^ = 3; 3: OL\ = n/l, ai = 0.3; 4: ai = TT, ^2 = 3;
5:ai = 7z,a2 = I', (b) retarding potential curves. 1: staticfields;2: microwavefields:h = 1.2 mm, UQ = 3.6kV;
3: h = 1.5 mm, UQ = 4.0 kV; 4: h = 2.5 mm, t/o = 6.8 kV; h is the transit space of cavity; UQ is the cut-off
potential).
Qfi = TT (curve 4 in Fig. 4.5b), at least two groups of electrons exist: One with high energies
(0.5£max-^max). and another one with low energies (O-O.Semax)- From curves 4 and 5 in
Fig. 4.5, we can see that the primary cause of the appearance of low-energy electrons is
transit effects. Thus, transit time effects have an important effect upon the electron energy
distribution at the resonator output with FE cathodes even for rather short (~1 mm) gaps
at several centimeter wavelengths. When ai = TT, a substantial fraction of the electrons in
the beam have almost zero energy. Clearly, electron "bunching," that is, the formation of
distinct groups of electrons is possible, as can be seen by the deep minimum in the retarding
potential energy curves.
4.7. FIELD EMISSION FROM LIQUID SURFACES
FE from metals typically occurs in surface fields exceeding 10^ V/cm. In practice,
working at reasonable applied voltages requires forming the cathode into a small tip with a
radius of curvature on the order of 10~'^-10"^ cm. These tips may be made by the traditional
methods of electrochemical etching, etching in an oxygen atmosphere, and formation in a
high electric field.^^^ Interestingly, liquid metals allow one to generate very high electric
fields at their surface with modest voltages due to an electrohydrodynamic instability that
arises.^^^' ^^^^ For instance, typical dimensions of the submicron inhomogeneities produced
with liquid-metal ion sources (LMIS) are on the scale of about 10~^ cm, which produces
fields sufficient for field ion emission. It has been shown^^^ that field electron emission
from liquid metals may also be obtained by the same LMIS techniques if one reverses the
emitter polarity.
It is also possible to form micro-protrusions of the surface of liquid conductors by
using microwave fields to excite microcapillary waves on the surface.^^^ Such an instability
leads to micro-protrusions having characteristic dimensions of the order of 10~^-10~^ cm,
providing field enhancement sufficient to initiate electron emission at rather small average
values of the microwave field in the resonator.
Such experiments have been conducted in a microwave field prismatic resonator (i/103
mode, natural frequency of 4.7 GHz, see Section 4.3). The oscillation wavelength was
6.4 cm and the input power Pin was less than 1.5 x 10^ W. Capillary emitters filled with
Ka-Na alloy or with fusible indium alloys were used. Also, a thin (roughly 1 |xm) layer
of Bi-Sn-Pb alloy deposited at a solid copper needle was studied. Materials having low
melting points (less than 150°C) were chosen to exclude effects due to thermionic emission.
The emission characteristics for various cathodes are shown in Fig. 4.6. The field
of resonator F is proportional to P^^ and that measurements were taken using single
pulses. The pulse durations were chosen to be less than 10~^ sec in order to prevent a
transition of FE to the explosive emission.^^ However, it should be noted that, as distinct
from emitters having a high melting point (e.g., with W), the operational characteristics of
liquid-metal cathodes are restored after a breakdown event. Lastly, it was found that the
emission characteristics of liquid-metal cathodes in microwave fields are reproducible over
a wide range of pressures, up to 10"^ Torr.
Note that the spread in experimental points is less for the microwavefields^^^(Fig. 4.6)
than for the staticfields.^^^This is due to the reproducible response of the liquid surface to
pulses of microwave field power.
It was shown^^^ that there is a finite time lag, ra, for the development of the FE process
at the surface of liquid metal cathodes, ra is related to characteristics of the system that
govern the formation of the electrohydrodynamic instability; field strength, surface tension
forces, etc.
The excitation of FE from liquid-metal cathodes by static and microwavefieldshas been
compared.^^^ It was found that in static fields, FE is not initiated even at fields more than
1.2 1.4 1.6

Prn^/2^.103W-1/2
Figure 4.6. FE characteristics for liquid-metal surface in microwave field. 1: K-Na alloy: ip = 5 |JLS, /max =
90 |xA (^capill = 90 jjim); 2: Bi-Sn-Pb alloy: Tp = 12 |xs, /max = 180 \iA; ip is the pulse duration.
Figure 4.7. Frozen signs of the microprotnisions created on the surface of a liquid-metal cathode in a microwave
field.
an order of magnitude larger than those at which the FE is observed with microwave exci-
tation. Therefore, the threshold value of the microwave field required for the development
of instabilities of a conducting liquid surface are smaller than the critical Tonks-Frenkel
field.'^^^' ^^^ Detailed calculations of the threshold values for microwave field excitation
corroborate this conclusion.^^^
Direct experimental evidence of the occurrence of microcapillary waves and of the
formation of fine micro-protrusions at a liquid-metal surface in microwave field has
also been obtained.^^^ Visualization was achieved by freezing the micro-protrusions on
the surface of a thin (1-2 |xm) alloy layer deposited on a Cu needle (Fig. 4.7). A
detailed discussion of initiating the FE processes from liquid-metal surface is given in
Fursey et al.^^^
4.8. RESUME
We summarize the main points concerning the fundamental physical aspects of

operating the FE cathodes at microwave frequencies.
1. FN theory accurately describes the FE process up to frequencies of ^10^"^ Hz.
The condition for validly of the adiabatic approximation to FE is (w|rtunl <^ 1-
Theoretical estimates of the tunneling time give rtun ^ 10"^^ sec.
2. FN theory has been experimentally validated at microwave frequencies over 4-5
orders of magnitude in FE current. Other consequences of the theory, including
the shape of the electron energy distributions have also been verified.
3. Maximum measured FE current densities are on the order of lOÂ/cm^ at
microwave frequencies.
4. Ion bombardment of the cathode surface is virtually eliminated at microwave fre-
quencies by the formation of an effectively repulsive potential for the ions near the
cathode surface. The result is a significant increase in the stability and operational
lifetime of the cathode when compared to dc operation at pressures of the order of
10"^ Torr.
5. The non-homogeneity of the electric field has significant effects on the electron
beam energy distribution in macroscopic microwave devices. Transit time effects
begin to modify the electron energy distribution at frequencies greater than 10^ Hz
with millimeter-range gaps. In vacuum microelectronics devices, such effects may
be neglected at up to frequencies of 100 GHz as the gaps are typically of the order
of 1-10 |xm.
6. Stable FE from the surface of liquid metals has been demonstrated. Microcapillary
waves are formed at the liquid-metal surface, their crests forming an array of
field emitters. The dimensions of the surface instabilities can be of the order of
nanometers.
7. The unique properties of FE offer many intriguing opportunities for applications in:
(a) "macroscopic," high-current devices: beam-bunching injectors, energy convert-
ers, high-power traveling wave tubes, klystrons, magnetrons, etc.; (b) high-energy
electron optical devices; in particular, very-high resolution electron microscopes
having beam energies of the order of several MeV; (c) vacuum microelectron-
ics devices (frequency multipliers, amplifiers, traveling wave tubes, miniature
electron sources for electron optical systems used for microscopy, holography,
lithography, etc.).
FIELD EMISSION FROM SEMICONDUCTORS
5.1. INTRODUCTION
The interest in field emission (FE) from semiconductors is stimulated by the following
unique features of this process:
1. FE provides a source of valuable information on the physics of solid state surfaces.
FE and ion emission microscopy, which enable visualization of solid state surfaces with
high resolution and magnification yield data on a number of important characteristics of the
surface such as the work function, ionization potential of surface atoms, electron affinity,
surface tension, evaporation energy of the surface atoms, and activation energies for surface
transport processes. Knowledge gained by these techniques in studies of surface phenomena
on metals are widely known.^^' ^^' ^^' •^^^ ^^' ^^^ Initially the application of these techniques
to semiconductors presented many problems due to the difficulty of cleaning the surface
of a semiconductor emitter and obtaining a symmetric emission image. In this chapter the
issue of semiconductor tip surface cleaning is given special consideration.
2. Distinct from metals, a semiconductor offers numerous ways of varying the charac-
teristics of the emission process by control of carrier concentration in the emitter bulk, thus
making unique electron devices possible. In recent years a number of ideas have been gen-
erated which might lead to new systems in vacuum electronics. It appears possible to have
a limited carrier concentration in the near-surface region of a semiconductor and control
the FE process by generating or injecting carriers into this region.
3. Application of semiconductors in vacuum microelectronics is aided by the fact that
for some semiconductor materials, such as Si, the basic technology of fabricating complex
structures has been developed.
4. Local deviations from electrical neutrality in the emitter bulk could create an electric
field near the surface thereby reducing the threshold externally applied field for electron
emission.^^'^ ^^
5. Finally, a semiconductor microcrystal/field-emitter can produce an extremely high
electric field (10^ V/cm) near its surface, which cannot be produced by any other means.
Thus it may be possible to study surface processes and transport: adsorption, desorption,
catalysis, surface diffusion in strong electric fields.
71
FE from semiconductors, relative to metals, is a much more complicated process due

to the low carrier concentration in the emitter bulk. The low carrier concentrations allow
for field penetration into the semiconductor, causing band bending and nonlinearity of
the current-voltage characteristics in Fowler-Nordheim (FN) coordinates. Under certain
conditions these same features can make the FE current very thermal- and photosensitive.
The capture of free carriers by traps can cause a deviation from the electrical quasi-
neutrality in the emitter bulk and lead to modifications of thefielddistribution at the surface.
A semiconductor emitter is very difficult to prepare because surface cleaning, sharpening,
and heating can lead to irreversible changes in its initial properties. These irreversible
changes can be difficult to identify and lead to experimental results that are difficult to
reproduce. For many years this problem prevented systematic studies of the field elec-
tron emission in semiconductors and thereby the development of an adequate model and
quantitative theory of the phenomenon.
Reproducible results can be obtained only if one begins with a clean surface using
condition and treatment procedures that do not change the sample's initial properties. We
consider this matter very important and therefore assigned to it two subsections, 5.2 and
5.4. In this chapter we will concentrate on the most reliable results obtained on emitters
with atomically clean surfaces and reproducible initial properties.
In earlier reviews some experimental observations were presented. As regards theory,
a number of rather abstract mechanisms were suggested, which, as was shown afterwards,
either had no effect on the process at all or represented factors of second or third order.
Data on FE from semiconductors wasfirstreviewed in Elinson and VasiFev^^ and the
most substantial review was given in Fischer and Neumann. ^"^^ Comprehensive efforts to
elicit general properties of the phenomenon and give a consistent account of the suggested
mechanisms were undertaken in Fursey and Uvov^^^ and Baskin et al.,^^^ as well as in a
fairly recent review by Modinos,^^ a paper by Fursey and Baskin,-^^^ and in recent works
by Fursey. ^^^
5.2. EMITTER SURFACE CLEANING
In FE experiments, especially with semiconductors, it is most important in sam-

ple preparation not to change its initial properties during its cleaning and obtaining FE
patterns.
The most effective technique of cleaning a semiconductor tip of surface impurities is
desorption in a strong electric field, either in vacuum or in a hydrogen atmosphere. The
latter has been used in cleaning Ge,^^^ Si,^^^ GaAs,^^^'^"^^ and CdS.^^'^ Field-stimulated
reactions between the impurity atoms and hydrogen on the sample surface greatly facilitate
the desorption process. However, as hydrogen diffusion into the sample bulk can change
the characteristics of the emitter, vacuum desorption cleaning is preferable. Until recently,
cleaning of the emitter surface by field desorption was an extremely laborious process. The
main problem was the "explosion" of the emitter tip often in the final stage of cleaning.
Emitter destruction was initially associated with fatigue phenomena.^^^'^^^ Later, it has
been shown^^^"^^^ that emitter instability is the result of the fact that the electric field
needed to remove well bound surface impurity atoms is typically much higher than the field
required to evaporate the tip material. Thus impurity removal can often result in an abrupt,
explosive-like evaporation of the semiconductor material and destruction of the tip.
FIELD EMISSION FROM SEMICONDUCTORS 73
Stability of the semiconductor tip during field-desorption cleaning is sensitive to the

temperature and the duration of sample heating during sample pretreatment and vacuum
chamber outgassing.^^^ For example, it has been found that the field-desorption cleaning of
Ge in vacuum is feasible only if the sample temperature does not exceed 673 K.^^^ Lowering
of the pretreatment temperature to 473 K in most cases allowed for contamination removal
without destruction of the semiconductor tip.
The reasons for the change in the properties of Ge samples during heating have been
analyzed.^^^^^^ The formation of an adherent coating occurs as a result of conversion of
various germanium oxides to Ge02. Heating of the tip to temperatures above 673 K, even
for a short time stabilized the oxide layer to the point that it is impossible to remove by field
desorption without destroying the tip. An FE pattern of Ge with an adherent oxide layer at
the periphery of the emitting region is shown in Fig. 5.1.
Lowering the temperature to 473 K during pre-treatment facilitated surface cleaning
in the case of Si as well. At this temperature the surface of the tip can be cleaned by
field-desorption alone.
Figure 5.2a shows a FE pattern of a Ge tip oriented in the (Oil) direction and cleaned
by field desorption in ultra-high vacuum.^^^ During preliminary treatment the temperature
did not exceed 473 K. The FE pattern is low-contrast and details of the crystal structure
are poorly resolved.^^^' ^^^' ^^^ After a brief heating to 423-673 K for 5-10 sec a distinct
symmetry emerged as shown in Fig. 5.2b. An interesting feature of the freshly cleaned
Ge surface is the bright spots in areas of the cube faces (Fig. 5.2a). This local emission
enhancement was observed in the past^^"* and attributed to the migration of atoms towards
the apex of the tip where the field strength is highest. However, it has been shown^^^' ^97-300
that (100) faces in other locations as well (even those at the edge of the emitting region)
have enhanced emission, Fig. 5.2a. Investigations of FE from different areas of the emitting
surface of Ge ^^^ have been used to determine the difference in work function values between
cube faces and the central part of the emitting surface which was found to be 0.3 eV. Such
a large difference in work function contrasts with previous results obtained using other
methods. This discrepancy is likely due to the fact that Ge surfaces cleaned using field
differ significantly from those cleaned via ion bombardment.^^
Figure 5.1. FE micrograph of a Ge tip with an adherent oxide at the periphery of the emitting region.-^^^
(b)|
Figure 5.2. FE pattern of Ge"^^' after (a) field desorption; (b) thermal anneal.
Figure 5.3. FE patterns of Si:^^^ (a) annealed at 1610 K; (b) annealed at 1650 K.
The first FE patterns of silicon were obtained by D'Asaro^^^ after heating the tip to high
temperatures. Later, better quality FE images were obtained by Allen^^"^ and Perry.^^^ They
both used extensive heating and field desorption of the tip sample. Field-emission patterns
of Si cleaned by field desorption in vacuum are shown in Fursey and Ivanov^^^ and Fursey
and Egarov.^^"* Excellent FE patterns of Si cleaned by field desoption in hydrogen were
published by Busch and Fischer.^^^ These authors have also obtained superb FE patterns of
Si after tip annealing (Fig. 5.3). FE patterns have also been obtained for some compound
semiconductors: GaAs,^^^' ^45,346 JÂS,^^^ InSb,^^^ and CdS.^^'^
Field ion imaging of semiconductors is very difficult because image fields are often
higher than the evaporation field of the emitter material. Nevertheless, such patterns have
been obtained. Melmed and Stem^^^ and Melmed and Givargizov^^^ have taken high
quality field ion patterns for Ge, Si, GaAs, and InAs (Fig. 5.4) using neon and hydro-
gen as the image gases and employing a micro-channel plate intensifier. These field ion
Figure 5.4. Field ion micrographs at ~30K (courtesy A. J. Melmed and E. I. Givargizov): (a) n-type Ge (111)
imaged with neon (~62 nm diameter image); (b) Si, 111 imaged with neon (~56 nm diameter image); (c) B-GaAs
(111) imaged with hydrogen (~56 nm diameter image); (d) InAs (111) imaged with hydrogen (~44 nm diameter
image).
micrographs were obtained for semiconductor whiskers grown by the vapor-liquid-solid

(VLS) method.^^^ The individual whiskers were removed from substrates and epoxyed to
tungsten wire supports. Electropolishing was used to sharpen the whiskers for field ion
309
microscopy. "^
5.3. CURRENT-VOLTAGE CHARACTERISTICS
The relationship between the FE current and electric field (or voltage) has been the
subject of a number of investigations.'^^' ^^^ It was found, in contrast to metals, that in
semiconductors with low carrier concentrations in the conduction band (high-resistivity p-
and n-types) the current-voltage dependence in FN coordinates is nonlinear. Some evidence
of nonlinear current-voltage characteristics appear in early studies by Apker and Taft who
investigated CdS.^^^
The first comprehensive study of the nonlinear current-voltage characteristics was
carried out by Sokol'skaya and Shcherbakov.^^^'^^^ These experiments showed that the
, \ p-Ge296 -6f
-^h" V" r=77K

- V.Jight
I II ^
5> - 8
\ \ 1
-9
-10 1 1 1 1 1
-6
P-Ge3i5
V \^^^ ^^^^ L = 0
-7
\^s!?\
< \ ^"^^^Î
y 253 K
^-8
%195K
-9
1 1 1
-8F GaAs3i6 -6
V^^ 7'=295K
-9 r=300K -7
L=0 •^
1
5 ,038^
-9 ~
XPJDS^k
-11
-10
L=0 ^
, 1 — 1 1 1
1.2 1.6 2.0 2.4 2.8

loVuy-^
Figure 5.5. Nonlinear current-voltage characteristics of various semiconductors; LQ = maximum light
intensity. ^^^'^^2
current-voltage curve can be divided into three regions (Fig. 5.5): region I—linear current-
voltage dependence; region II—slow current variation ("saturation" region)—it is in this
region of the current-voltage characteristic that a strong photo- and thermal sensitivity of
the FE current has been observed; region III—^rapid rise of the current with voltage (in some
cases more rapid than in region I).
Later studies showed that the current-voltage characteristics of this type are typical of
a wide variety of semiconductors, including Sb2S3,^^^ high-resistivity Si,^^^ p-type Si,-^^^
low-resistivity Ge,296 p.type GaAs,^!^'^!^ In2S3,^^^'^^^ CdS,î2,3i3 ^^d CdP2.^^^ These

observations have been summarized^^^ (see Fig. 5.5).
5.4. ON PRESERVING THE INITIAL SURFACE PROPERTIES OF

A FIELD EMITTER
It was noted that the shape of current-voltage characteristics of a semiconductor emitter

can be significantly changed during sample preparation as well as during the course of the
experiment itself. It was found that noticeable changes to the initial, intrinsic properties
of emitter material take place during its heating. This is associated with the contaminants
remaining on the side of the tip surface after electrolytic etching. At elevated temperatures
in vacuum these contaminants diffuse into the tip bulk and modify the doping. In some
cases highly conducting channels form on the sides, electrically shunting the sample's bulk.
Changes of the initial properties offieldemitters were noted in early studies of Si.^^^' ^^^
It was assumed in these studies that during cleaning of the silicon tip by annealing at 150G-
1600 K, rapid diffusion of acceptor impurities into the sample occured, which led to the
formation of a highly degenerate surface p-layer, irrespective of whether the material was
n or p type.
It has also been shown that the FE properties of Si, Ge, and some other semiconductors
change markedly as a result of treatment at much lower temperatures.^^^"^^^' ^^^ Thermal
treatment of Ge and Si samples at 700 K linearized the current-voltage characteristics and
eliminated any photo- or thermal sensitivity. Perceptible changes in the I-V characteristics
are observed even after heating to only 570 K. Heating to this temperature weakens the
current saturation in region II and shifts the appearence of the region to higher current
levels (see Fig. 5.6).
(a) (b)
6h
< 7 ^n
1 2 2 3
10^/L/,V-^
Figure 5.6. Influence of thermal treatment on the current-voltage characteristics:^^ ^ (a) change in initial emission
properties of a p-type Si cathode (14 Q • cm) after heating. T°C: (!) 150; (2) 250; (3) 300; (4) 350; (b) CVC for
a p-type Si cathode (14 ^ • cm) after different thermal treatments. (Previously, the emitter was electrochemically
etched creating stable complexes with metallic ions.) T°C: (1) 250; (2) 300; (3) 350; (4) 400; (5) 500.
The initial properties of a semiconductor emitter are not affected unless it is heated
to above 420-500 K. In this case the current-voltage characteristics do not change, the
photo- and thermal sensitivity remains the same, and the experimental results are completely
reproducible.
It has been found experimentally^^ ^ that irreversible changes of the properties of semi-
conductor field cathodes heated to temperatures in excess of 400 K are principally caused by
metal ions depositing at the emitter surface during electrolytic tip sharpening. These are the
metals which modify the sample surface conductivity upon heating. To reduce the concentra-
tion of metal impurities,^^^ tip treatments in solutions of organic agents that form complexes
with copper and similar metal ions in base solution^^^ were effective. These complexes can
subsequently be easily removed by rinsing with water. The experimental results are shown
in Fig. 5.6b. These experiments demonstrate that the initial properties of semiconductor
emitters subjected to thermal treatment are not changed so strongly if preventive measures
are taken.
5.5. VOLTAGE DROP ACROSS THE SAMPLE AND THE FIELD

DISTRIBUTION IN THE EMITTING AREA
As yet no theories satisfactorily describe field electron emission from semiconduc-

tors. Shapes of the current-voltage characteristics predicted by the Morgulis-Stratton
theory^^^"^^^ do not agree with experimental curves.
In a number of mechanisms that reproduce the essential features of the observed
current-voltage characteristics, that the prevailing four points were suggested are as follows:
1. The nonlinearity of the current-voltage characteristics is due to the voltage drop
AL^c across the emitter. This idea, first suggested by Apker and Taft,-^^^ was later used to
explain the saturation observed in region II of the current-voltage characteristics.^^^' ^^^
However, direct measurements by the retarding potential method, demonstrated that this
interpretation is not correct. Actually the shape of the current-voltage curves in the majority
of p-type semiconductors^^^' ^^^ as well as in high-resistivity n-type semiconductors^^^ is
only slightly affected by A (7c- The extent to which Af/c affects the shape of the current-
voltage curves for Ge and Si^^^' ^^^ can be seen in Fig. 5.7.
2. In a number of experiments the emission images change size when going from
region I to region II of the current-voltage characteristic.^^^' ^^^' ^^"^ Throughout region II
the emission pattern shrinks in size as the voltage is increased. In region III the size of
emission pattern typically stays constant and sometimes increases and becomes more than
in region II. The effect of image shrinkage is dependent upon the electron concentration in
the conduction band and the emitter shape. A similar experimental study of the changes of
emission image size with the applied voltage have been carried out on Ge and Si.^^^' ^^^
Changes in the size of the emission image reflects changes in the structure of the
emitter space-charge region as it spreads. The field around the tip apex becomes more
uniform and the equipotential surfaces less curved leading to lower magnification values
in the FE microscope. Emission image shrinkage is illustrated schematically in Fig. 5.8.
Variations of the size of emission image with applied voltage U and the voltage drop across
the emitter are shown in Fig. 5.7.
(b)
-A 3\\4
-5 \\
'^
< -6 ^
^ \
"• -7
-8
\
-9 1 1 1 1 1 1 -
2.0 3.0 4.0 2 3 4 5 6 7
10^/(7, V"^ io'^/u,v-''
(c) Aa.v
0 50 100 150
> 0.95 h
0.90
Figures.?. Illustration of a weak influence AU and A^ on the nonlinearities of the current-voltage curves:^^^' ^^^
(a) Current-voltage characteristics of Ge at 253 K (no illumination): (1) typical experimental curve; (2) with
correction to the voltage drop; (3) with correction to the variation of ;8( V); (4) FN plot. Current regions I, II, and
III visible in curve 1; (b) analogous to (a) for p-type Si (3 ^ • cm); (c) (1) relative change in the size of the emission
pattern with applied voltage; (2) relative change in size of the emission pattern with tip voltage (Ge at 253 K not
illuminated); (d) analogous to (c) for p-type Si (3 ^ • cm).
Emission image shrinkage means that in region II of the current-voltage characteristic

the field geometric factor fi = F/U is decreasing. If this fact is not taken into account in
calculations of the current-voltage characteristics, the values of the field strength F near the
emitter surface will be overestimated. From the measured relative sizes of emission image,
R/Ro, the distortion of the field profile near the surface can be estimated.
Assuming
(5.1)
Ro ^ Po
and taking account of the fact that R/RQ, as follows from experiments (see Fig. 5.8), is
decreasing linearly with U in region II of the current-voltage characteristic, we have
^ = l-k{U-Uo). (5.2)
(b)
\ "% ' ^ .!
(c) (d)
<p2=const
(^1= const
Figure 5.8. Schematic of the effect of a shrinking emission image: (a) initial emission image of the semiconductor,
RQ = initial image radius; (b) shrunken emission image, R—image radius; (c) At/ is the voltage drop across the
tip over the space-charge region depleted of carriers; (d) the potential and field distribution near a semiconductor
tip with field penetration.
Here ySo is a ^/-independent geometrical factor in region I of the current-voltage character-

istic; C/o is the voltage at which the emission picture starts to shrink; k is the slope of the
function R/RQ = f{U)\ k ^ IQ-^-lO"^ V ^
Assuming that in the range of anode voltages exceeding UQ the dependence of the field
emission current density j on field strength F follows FN relationship, the expression for
the FE current / as a function of U that accounts for variations in fi can be written as
Bip^f^îy)
I = A(U)cxp (5.3)
PoU[l - k(U - Uo)]
where cp is the effective work function, 5 is a constant, i^(y) is the Nordheim function of
argument y =z[(x- I) x e^P(U)]^^^/(x + l)^/^(^ and ae is a permittivity.
The FE current curves for varying field-voltage proportionality factor are plotted in
Fig. 5.7a. As seen in the figure (curve 3), P(U) causes only an insignificant deviation from
the FN linear relationship. Thus, the fact that p = P{U) is evident in an emission image as
the "shrinkage" effect but does not explain the observed kink and the saturation region in the
experimental current-voltage curves.
3. Elinson et al.^^^ explained the slow rise of current at voltages above a certain value
in an assumption that the average electron energy in this voltage range exceeds the electron
affinity due to heating of electrons by the internal field. In this case the potential barrier
transparency is at its highest and the emission is no longer impeded. However, emission
current saturation in samples with large electron affinity, ^ 3 - 4 eV, and a relatively small
energy gap (Si, Ge, and others) cannot be explained by heating of electron gas or emission
over the barrier because the energy of electrons is limited by the energy gap Sg.
4. An analysis of the experimental results for p-type semiconductors with cleaned emit-
ting surfaces indicates that the most probable cause for saturation of the current-voltage
curves is limited by the carrier generation rate in the bulk of the sample. The current sat-
uration is the result of a number of interdependent processes, which reduce the electron
concentration in the near-surface region of the semiconductor, increase field penetration
into the sample, change the field geometry near the emitter, and increase the voltage drop
across the crystal bulk. Under such conditions the emission current is limited by the supply
of carriers to the surface and is somewhat dependent on the barrier transparency and on
the applied potential up to the abrupt rise of the carrier concentration in region III. The
same processes may be responsible for nonlinearities of the current-voltage characteris-
tics in high-resistivity n-type samples. An attempt at a theoretical consideration of such
a mechanism for p-type semiconductors has been undertaken.^^^ This approach has been
elaborated and extended to include the case of n-type semiconductors.^^^
5.6. THEORY OF THE FIELD ELECTRON EMISSION FROM

SEMICONDUCTORS
Despite the considerable accumulation and interpretation of experimental data on field

emission from semiconductors, and establishment of some general relationships, a quaUta-
tive theory of the process has not yet been developed. Different interpretations have been
put forward about individual facts or effects, such as the appearance of various slopes in
current-voltage characteristics, the effect on emission of the heating of the electron gas, the
role of possible quantization of electron states in the emitter near surface region, and so on.
Yet data on FE kinetics have not been treated theoretically. However, a simple review of past
and present theoretical work would warrant the conclusion that FE from semiconductors is
qualitatively understood.
As shown above, the nonlinear current-voltage characteristics of FE are inherent in an
extensive class of semiconductors. As a rule, these are p- or n-type semiconductors with a
low concentration of electrons in the conduction band.^"^^' ^^^ In the saturation region, the
emission current shows photo- and thermal sensitivity, the size of emission image changes,
and the voltage drop across the crystal increases.
The principal aim of the theoretical research on this problem is to reveal the mechanism
leading to the saturation region and the region of subsequent rapid current growth, to
ascertain the role of bulk defects and surface states, and to identify the energy states from
which the electrons tunnel into vacuum. Electron emission from the conduction and valence
bands of n-type semiconductors has been described by Stratton^^^ without accounting for
current saturation. Arthur^^^ directed his attention to the formation of a degenerate inverse
layer in the near-surface region of p-type semiconductor emitters and suggested explaining
saturation of the FE current by analogy with the saturation of the reverse current in a p-n
junction.
A very detailed model was analyzed theoretically by Yatsenko.^^^ Here the system
of anode-vacuum gap/p-type field emitter was considered as a kind of metal-dielectric-
semiconductor (MDS) structure, whose leakage current is essentially the emission current.
Numerically calculated saturation currents of these structures are a few orders of magnitude
lower than those observed experimentally. Later,^^^ an attempt was made to explain this
difference by the high generation rate on the lateral surface of the emitter. The current values
obtained agree with the experimental values to within an order of magnitude.
A theory which takes into account the mutual influence of the space charge region and
the current flowing through it has been developed.^^^ A small deviation of the electron gas
from thermodynamic equilibrium was assumed. This is equivalent to assuming a short effec-
tive lifetime of electron-hole pairs. The calculated values of the emission saturation current
agree with the experimental values to within an order of magnitude. The possible mech-
anisms which reduce the effective lifetime are a high surface generation-recombination
velocity and an increased trapping cross-section in the strong electric field.
The influence of screening of the applied field by the charge of surface states was
studied by Stratton.^^^ Fischer^"^^ has calculated probabilities of ionization of the surface
states by strong electric field and the values of emission current.
The influence of surface states on FE from p-type semiconductors has been analyzed,^^^
and the considerable effect of screening by the charge of surface states in the presence
of a stationary emission current was explained assuming that the surface states exchange
electrons with the energy bands.
5.6.1. Stratton's Theory
In the range of relatively low currents the Morgulis-Stratton theory^^^' ^^"^ gives quite
an adequate description of the experimental results: a linear increase in the logarithm of
the current with inverse value of applied voltage, constant size of the emission image, and
lack of photosensitivity. This theory assumes that due to partial penetration of the electric
field into the near-surface region of emitter the free electron concentration in this region
rises considerably and the electron gas is degenerate. Degeneration is characterized with a
parameter i^ which is the separation of the Fermi level from the bottom of the conduction
band. One of the basic assumptions of the theory is that only an insignificant fraction of
the electrons arriving at the potential barrier tunnel through it into vacuum. Therefore,
the emission process can be treated in a "zero current" approximation and the Fermi level
considered to be the same throughout the sample.
Calculation of the emission current density as a function of field strength is carried out
using the FN method.^ The current-voltage characteristics in Ig j versus F~^ coordinates,
in all cases, considered of emission from the valence or conduction bands for an arbitrary
degree of the electron gas degeneration in the near-surface region yield straight lines. Their
slopes depend on the electron affinity x»the forbidden band width Eg (for emission from
the valence band), £a the acceptor level, and the degeneration parameter i^, which equals
??s near the surface (Fig. 5.9).
Stratton has shown that in the presence of considerable concentrations of positively
charged surface centers the current-voltage characteristics may display characteristic kinks.
Their origin is explained as follows: at low applied field strengths the surface charge field
bends the bands of allowed energies upwards so that initially emission comes from the
Figure 5.9. Energy diagram of the near-surface region of a semiconductor in the "zero-field" approximation, 151
(T^—degeneration parameter; i>s—degeneration parameter near the surface; x—electron affinity; Uix)—energy
at the bottom of the conduction band; ^—level of electrochemical potential.)
valence band. When the applied field becomes high enough to compensate for the sur-
face charge field the band bending changes sign, the concentration of the conduction
band electrons rises, and emission from the conduction band becomes dominant. In the
current-voltage characteristics the transition from valence band emission to conduction
band emission will be seen as a dramatic increase in current. However, curves of this shape
have not been observed in experiments.
A key feature of the measured nonlinear current-voltage characteristics with highly
resistive n-type semiconductors as well as /?-type semiconductors (Fig. 5.5) is the presence
of a saturation region at high fields which is at odds with Stratton's predictions. Thus it
appears that the nonlinearities in the experimental current-voltage characteristics are not
related to the influence of the positive surface charge.
A more rigorous analysis,^^^' ^^^ accounting for the effect of the semiconductor's energy
band structure on the electron emission process and corrections for the effective mass as well
as the surface states, yielded no new predictions as to the general shape of current-voltage
characteristics.
Although Stratton's theory cannot be applied to field electron emission from semicon-
ductors with low equilibrium concentrations of conduction band electrons in high fields,
qualitatively it is in fair agreement with experiment for most semiconductors in low electric
fields (region I of the current-voltage characteristic [Fig. 5.5]).
5.6.2. Modeling a Semiconductor Emitter: Statement of the Problem
The latest concepts ion modeling the semiconducting field emitter, that are supported
by experimental observations have been developed.^^^' ^^^' ^^^
Consider an extrinsic semiconductor with an arbitrary concentration of acceptor and

donor impurities. The contact of the semiconductor to the substrate is assumed to be located
at such a distance that its influence on the processes in the near-surface regions of the tip
may be neglected. It is further assumed that the carrier concentration at any point in the
semiconductor is determined by the quasi-equilibrium level of the electrochemical potential.
The penetration of the field into the semiconductor and the resulting change of the carrier
concentration in the near-surface region is allowed for in the band bending terms and the
problem is solved in a one-dimensional approximation.
Figure 5.10. Energy diagram of a semiconductor emitter in a high electricfield:^^^ (a) p-type semiconductors;
(b) n-type semiconductors.
The energy diagram of the near-surface region of a semiconductor in an electric field is

shown in Fig. 5.10. Here T^(X) is a function characterizing the position of the bottom of the
conduction band relative to the Fermi level at any point in the crystal, i>s the value of this
function at the surface, iôo its value in the sample bulk, U(x) and ^(x) are the energies of
the bottom of the conduction band and of the electrochemical potential level, respectively,
which are referred to an electron at rest at infinity, x is the electron affinity, e^ and s^ are
the energies of the acceptor and donor levels, referred to the bottom of the conduction band.
For the emission current density from the conduction band the following expression,
obtained by Stratton,^^"^ may be used (for a more general case see Christov^^^):
. _ e^F^
Jem —
STvhcp
2-^^1/2^3 exp ( - ^ ) cp^/^^s] ; ^s > 0 (5.4)
_ 47rme(kT)^
Jem ^ exp [ - 5 ; | x ^ / ^ e ( > ^ ) l exp (J^^ ; z^s < 0
where
y=27t{2m)^^^/eh; <p ^ X - ^s\
@{Y) is the Nordheim function of the argument
^^-1\^/2(^3FS)1/2
U + iy (p
in which, though, the degeneration parameter i^s is not merely a function of the field
penetration into semiconductor under condition of zero (negligible) current, as in Stratton's
theory, but calculated by simultaneously solving Poisson's equation
—^ = (n -p-^N- -N^) (5.5)

dx^ X
and the equation for the density of current j flowing through the sample and equal to the
emission current density 7em*
d^
J = -(l^nn + fJ^pP)— (5.6)
Here n, /?, N'^, and A^~ are the concentrations of electrons, holes, and charged impurity
centers, respectively, /x„, /x^ are the electron and hole mobilities, Fs the applied field
strength, and ae the high-frequency dielectric constant.
Assuming a quadratic dependence of the electrons and hole energies on the quasi-
momentum, we can write
""^^'/^j^)' P= «^î/2(-^^) (5.7)

where 8g is the energy gap, a = {rrip/mn)^^^, ^ — An{lmnkT/hî^^^, rrin and nip are the
effective masses of electrons and holes, î/2(?]) the Fermi integral of the order \ defined
in the usual way.
We assume that in formulae (5.7) T is the lattice temperature. The concentrations of
the ionized donor and acceptor centers will be, respectively,
-1
(5.8)
N = A'.[l+2exp(£i^) (5.9)
The simultaneous Eqs. (5.5) and (5.6) are solved taking into account the boundary
conditions
dU 1 eFs
(5.10)
dx x-»-0
dU di; J (5.11)
~d^ dx x->-oo M««oo + M/?Poo
while the band structure of the semiconductor and the doping level (i.e., Woo, "oo^ A^d. and
Aâ) are assumed known. Expression (5.11) may serve as a boundary condition because in the
simultaneous solution of Eqs. (5.5) and (5.6) j is assumed to be an independent parameter
equal to the emission current j = jem(Fs, T^S)- The calculations are much simplified if
a dimensionless energy parameter y(x) = G{x)/kT is introduced, whereupon the space
charge density electric p is expressed as
p(y) = e{n-p-N:^ + N;) (5.12)
and the "effective" total density of mobile charges
b(y)=n + ^p. (5.13)
It is possible to transform the simultaneous Eqs. (5.4)-(5.6), using y as an independent

variable instead of x. By boundary condition (5.11) and denoting E{y) the field strength
inside the sample, we obtain
STtekT
eÊ^{y) -[H(y) + G{y)l (5.14)
f^n^îyoo) ae
where
Hiy) = rP(y)dy, (5.15)

hoc
y dy
G{y) = j f (5.16)
p{y) .
voc eiinb{y)E{y) - j
Here joo = y(^)|jc^-oo and y^ = y{x)\x-^-o = ^sl^T. If jc = 0 is chosen as the upper

integration Umit, then using condition (5.10) expression (5.14) may be rewritten in the form
e-Fl / - « - ' ^ ^ . [ H ( , . ) + G(..)]. (5.17)

ae2 ^2jy2(^y^) ae
Equations (5.14) and (5.17) are very convenient for a qualitative discussion of the depen-
dence of emission current density on the applied field strength, crystal temperature, and
doping level for both p- and n-type semiconductors.
Together with Eq. (5.4) they allow for a numerical evaluation of the current-voltage
characteristics. To determine the spatial distribution of the charge density, field strength,
and the position of the quasi-Fermi level in a semiconductor, it is necessary to know the
dependence of the space coordinate JC on y. This dependence follows from Eq. (5.6) written
in integral form
. = kT r '^"^^'^ dy (5.18)
Jy efMnHy)E(y) - j
similar to integral (13) in Stratton.^^'*
5.6.3. Zero Current Approximation
For low electric fields, where the electron emission currents are still small, the influ-
ence of the current flowing through the sample may be neglected. Assuming in (5.14) and
(5.17) that the density of the current flowing through the sample is zero (the "zero current
approximation"), we obtain
eÊîy) = —kTH(y) (5.19)

ae
and
f^ = '-^kTHiy.) (5.20)
ae"^ ae
It should be remembered, of course, that the emission current density 7em is, in fact
not zero.
The integral H(y) defined by (5.15), is taken in the explicit form
[2 2
H(y) = eA j -^s/iiy) + « - ^ 3 / 2 ( - y - ^g)
+ Y ^"^ I ^ + ^ exp(>^ - co^)] + ^ 1^ 11 + ^ exp(a;d - y ) l } (5.21)
where coq = Sq/kT. This result is similar to that obtained earlier by Seiwatz and Green.^^^
It can be shown that in not very heavily doped semiconductors (N^cog/A <^ 1 and
N(^/A <^ 1 the first term in (5.21) becomes dominant once y = i^/kT > —2.
Using this fact, we obtain from (5.20) and (5.21)
F^ = S7TxAkTl3='3/2(ys) (5.22)
The condition i^s = 0 means that the bottom of the conduction band is at the Fermi
level and the electron gas near the semiconductor surface is practically degenerate (Strictly
speaking, the degeneration sets in already at i>s ^ —IkT).
Substituting into (5.22) the value of §53/2(0) = 0.7685, and with T expressed in K,
we obtain a degeneration criterion for the electron gas near the semiconductor surface
F > Fî* s 1.44 X 10^^1/2 j ^ ^ ^ 3 / 2 / _ ^ y ^ ' v / ^ ^ (5 23)
For the case of Ge and Si (aece = 16, aesi = 12) at T = 300 K and (m„/m)Ge =
0.55, (m„/m)si = 1.08, we have F^^^ = 3.7 x 10^ V/cm; F^^ = 5.7 x 10^ V/cm.
The calculations above indicate that the fields at which degeneration of the electron
gas at the emitter surface sets in are considerably lower than the fields necessary for an
appreciable FE, that is, in the stationary case the emission always comes from the degenerate
conduction band.
From Eq. (5.23) it is seen that the degeneration criterion does not depend on the type
of conduction and the doping level because with approaching degeneration the concentra-
tion of electrons in the near-surface region becomes, in fact, independent of the electron
concentration in the bulk.
At fields corresponding to y > 6, it is possible to make use of the asymptotic behavior
of the Fermi integral ^3/2 (y) ^ fy^^^; this approximation is accurate to within 15 percent.
In this case expression (5.22) is simplified and we get a relationship equivalent to the one
derived by Stratton^^"^
15/.3 ^^/^
^s = Fs^/^ (5.24)
Ll287r2ae(2m„)3/2J
The current-voltage characteristics plotted using the relationships (5.4) and (5.22) are
straight lines. Distributions of the field strength and the electron concentration in the sample
bulk under zero current may be obtained from expressions (5.18) and (5.7).
Numerical calculations for p-Ge at Aâ = 10^^ cm~^ and T = 300 K show that the field
strength and the electron concentration change drastically at a depth of ^10~^ cm, which is
approximately an order of magnitude smaller than the emitter radius. This means that in the
region of small FE currents of the emitter surface may be considered nearly equipotential
and the geometrical factor p = F^/U constant and independent of the applied field, what
is in agreement with experiment (cf. Section 5.5).
5.6.4. The "Nonzero Current" Approximation with p-Type Semiconductors
Both n- and p-type semiconductors in a high electricfieldhave different structures in the

space charge region. The carrier concentration in n-type semiconductors smoothly decreases
away from the crystal surface (Fig. 5.10b), while in /?-type semiconductors a depletion
region is formed directly after the near-surface inversion layer (Fig. 5.10a), in which the
free carrier concentration and, accordingly, the conductivity are much lower than in the bulk.
In p-type semiconductors the current through the sample and thus the emission current are
essentially space-charge limited.
In the region of minimum conductivity the level of the electrochemical potential is
approximately in the middle of the forbidden gap and y = y^ = —o)g/2 -f In ^ , where
a = {lip/în)(i^p/i^n)^^^- According to (5.13) the effective free carrier concentration is

bm where
bm = b{ym) = Avêxp [ ~ ) (5.25)
The position of the Fermi level and b{y) are related by formula j ~ b(y)d^/dx. Obvi-
ously, for a given current value d^/dx is maximum at >^ = y^ and, generally, remains
so throughout the depletion region. At the surface (y = ys) and in the bulk of the sample
(y = yoo)b(y) is large and the electrochemical potential changes slowly. Schematically, the
position of the quasi-Fermi level in the energy diagram for a p-type sample through which
current is flowing is presented in Fig. 5.10a. As d^/dx approaches dU/dx, the integral
of G{y) in Eq. (5.14), depending on j rises sharply and the zero current approximation
ceases to be valid. By means of estimates it can be shown that with increasing current in
Eqs. (5.14) and (5.17), first of all the term containing G(ys) should be considered, because
the contribution of the term proportional to 7^, is approximately two orders less for the
fields which satisfy the condition e^F^/x ^ SneKTGiys). It follows from expression
(5.16) that G(ys) is large if the denominator of the integral expression is small, that is, in
the region where b(y) = bm- The denominator of the integral expression for G{y) is always
positive as it represents the diffusion current. At any arbitrary point of the semiconductor
emitter the limiting value of the current density is given by
j < efiMy)E(y) (5.26)
with a minimum value 2ity = y^.

Noting also that E(y) < Fs/ae, we can write
j < efinbiym)- = eiXnAâe-^^l^^ = /^^ (5.27)

ae ae
The value of the emission current is limited by the supply of carriers from the depletion
region. In the current-voltage curves this limitation is seen as current "saturation", that is,
slower current increase with applied field (Fig. 5.5). The values of the limiting current densi-
ties calculated from (5.27) are for Ge (Aâ = 10^^ c m - ^ T = 3 0 0 K ) / ^ " ' ^ 2 x 1 0 " ^ A/cm^,
and for Si (Aâ = 10^^ c m - ^ T = 300 K ) / ^ " ' ^ 1 A/cm^ which is considerably less than
the critical current densities leading to the emitter destruction. The limiting current density
depends slightly on the semiconductor doping level.
From formula (5.27) one can derive an important relationship between the limiting
current density and temperature. With cog = Sg/kT and noting that the strongest dependence
on T is contained in the exponent, and neglecting the weak dependence of the coefficient
A on temperature, we write
dilnf"^) Eg
(5.28)
d(l/kT)
The result shows the important role of the carrier generation in the depletion region and
of the high temperature sensitivity of the semiconductor field emitter at 7em = /^"^- It is
quite clear that the conductivity of the depletion region may be increased by the generation
of carriers by light (see Fig. 5.5).
5.6.5. The Effect of High Internal Fields
In our calculations the process of carrier generation by strong electric field, which
results in the avalanche current rise prior to breakdown is not considered (region III of the
current-voltage characteristic, Fig. 5.5). The internal field affects the emission current in
the saturation region mainly by changing the carrier mobility. This dependence may be
approximated by the formula IJL(E) = jioiEo/E)^, where the exponent 8 is typically in the
range 0 < 8 < I (depending on the particular mechanism that limits the carrier mobility),
jjio is a constant equal to the carrier mobility in field EQ, and EQ is the field strength value
above which this mechanism becomes dominant. With most semiconductors in the fields
of 10^-10^ V/cm, the major limiting mechanism is generation of optical phonons by the
electrons, in which case 8 = 1.^^^ For weaker fields the dominant mechanism may be
scattering of carriers by acoustic phonons, with 8 = ^.
As shown by the calculations, in the region of transition to saturation the internal fields
in p-type crystals may exceed EQ ^ 10^-10"^ V/cm, so that in estimations of the limiting
emission current density by Eq. (5.27) it may be assumed that 8 = 1. Then
f^ = eAVînoEo] exp ( - ^ ) (5.29)
In this case j ^ ^ does not depend on the field strength.

It can be demonstrated that /^"^ is very close to the saturation current density (the latter
can be determined in a straightforward manner only under complete saturation), and the
transition region from the linear current-voltage relationship to saturation is very narrow.
If the carrier mobility in this case is proportional to E~^, then complete saturation of the
current should be expected. On samples of p-Ge an almost complete saturation has been
observed.^^^ Full saturation has been observed on samples ofp-Si^^^ (Fig. 5.5). In the case
of slowly varying /iniE), the transition to saturation occurs in higher fields and the current
density in this region increases slightly with the applied field.
The results of numerical calculations of the current-voltage characteristics accounting
for the function /x„(£) are discussed in Section 5.6.7.
5.6.6. Field Emission from /f-iype Semiconductors
In n-type samples the diffusion current density in any arbitrary cross section is large
in comparison with the total current density j . Therefore the function H(y) is small every-
where, and the term G(y) dominates in Eq. (5.17). The condition of the applicability of the
zero current approximation for an n-type semiconductor can be written in the form
This condition means that the density of current flowing through the sample should be
much smaller than the limiting current density
/im ^ 'Il^^noo (5.30)

ae
It is evident that at j ^ /^"^ the FN-type j{F) function reduces to Ohm's law if fiiE) is
ignored.
It follows from (5.30) that the saturation current density is proportional to the major-
ity carrier concentration Woo in the sample bulk which in turn depends on the donor
concentration N^ through the relationships
A^d at r > 7b,
where To = 750(m/m„)(A^d/10^^)^/^
According to (5.31), at 7 > TQ the saturation current density /^"^ does not depend on
the temperature and at T < 7b
^(}I1J^^_M. (5.32)
d(l/kT) 2
As l^dl <^ Sg, the temperature dependence of the saturation current density in /i-type semi-
conductors is much weaker than in p-type semiconductors (cf. formula (5.28)). The results
of numerical calculations of current-voltage characteristics for n-type emitters follow in
Section 5.6.7.
5.6.7. Results of Numerical Calculations
The simultaneous Eqs. (5.5) and (5.6) may be reduced to a first-order nonlinear
differential equation of the form
^ = 1^ ^Î^ (5.33)
dy ae E{y) - j/iinb{y)
with the boundary condition
b{yoo)
where joo is determined from the relation p(yoo) = 0.
Equation (5.33) was solved numerically. In the calculation, the current density yem — j
was assumed to be a parameter. With the prescribed emission current density for fixed values
of the parameter j , E{yoo) was calculated and, by Eq. (5.33), E{y) derived (see Fig. 5.11).
From the relationships for the emission current density in Eq. (5.4), E^ may be determined
as a function of jem and y^\ F^ = 7^s(7em» Js)- The values of y^ and F^ corresponding
to the specified emission current density are determined graphically from the equation
FsUcm, ys)/^ = E(ys). The current-voltage characteristics calculated by this method are
given in Fig. 5.12, curves 1 and 2.
As seen in Fig. 5.12a, the transition to saturation in the second case begins at higher
voltages and the corresponding saturation current density is increased by roughly one order
of magnitude. The importance of the dependence of carrier mobility />i„ on the field strength
is clear. In the second case, the saturation region is much flatter, which is in agreement with
the results of the quantitative analysis.
Note that the calculated values of the saturation current density /^"^ are considerably
smaller than the measured values. This discrepancy is likely due to the fact that the emitter
^10^ E, V/cm
-^103
-18 -16 -14 -12
Figure 5.11. Relationship E{y) (p-Ge, Aâ = 10^^ cm~-^, T - 300 K) for several values of the parameter y}^^
(1) 7 = 0; (2) i = 100 A/cm^; (3) j = 150 A/cm^; (4) ; = 200 A/cm^.
model is one-dimensional. Preliminary calculations show that /^"^ for conical emitters is
one or two orders larger. ^^^
Calculation results of current-voltage characteristics for n-type samples are shown in
Fig. 5.12b. Here a strong dependence of the saturation current on the impurity concentration
is observed.
The structure of the space charge region at nonzero current through the sample may be
characterized by the distribution of free carriers in the space charge p{x) and the parameter
y{x) across the sample. Calculations show that the initial decrease and subsequent growth
of the concentration of free charge carriers b{x) are rather sharp. Therefore, the character-
istic dimension of the depletion region / in a /?-type sample may be defined as a distance
between the points in which the free carrier concentration is double the minimum concen-
tration. Figure 5.13 clearly shows broadening of the depletion region with increase of the
field Fs (p-Ge, Aâ = 10^^ cm~^, 7 = 300 K). When the current density saturates, the
broadening of the depletion region continues solely because of more field penetration into
the semiconductor. The result is in good qualitative agreement with the measurements of
the characteristic dimension of the depletion region by light probe method.^^^ The electron
concentration n^ near the semiconductor surface first slowly increases as the applied field is
increased, then over rather wide range of Fs it remains essentially constant and eventually
drops sharply as Fs approaches the value corresponding to the transition to saturation when
the near-surface region is depleted (Fig. 5.14).
Analysis shows that the main features of the FE current-voltage characteristics are
related to the low concentration of free electrons and the existence of capture centers,
including deep traps, inside the semiconductor.^^^' ^^^' ^^^' ^^^
FE from /?-type semiconductors is difficult to analyze. The first difficulty involves
accounting for bipolar conductivity and, hence, generation and recombination processes
in the emitter. The simultaneous solution of the Poisson equation, the generation-
recombination equations, and the equation for the current density in the sample is necessary.
2.2 2.4 2.6 2.8 3.0 3.2 3.4

F;\lO-®xV-''cm
Figure 5.12. (a) Theoretical current-voltage characteristics calculated for p-Ge with the assumption of various
relationships between mobility and strength of the internal field in a crystal of p-Ge, Aâ — 10^^ cm~'^, T =
Mn(0); F < 10^ V/cm
10^ V/cm < F < lO'* V/cm

(1) ^^niF) =
1/2
I1 . /io'\ F > 10^ V/cm
Mn(0); F < 10^ V/cm

(2) fln(F) =
Mn(0) F > 10^ V/cm
(^)"-
(b)
2 3 4
F;\lO-8xV-^cm
Figure 5.12. (b) Theoretical current-voltage characteristics calculated for sample of n-Si with (l) N^
2 10*2cm-2,(2)8.510^2cm-^
10"
10-'
£
o
10"^
10-
3.0 3.2 3.4 3.6 3.8 4.0
Fs, lO^xVcm-"'
Figure 5.13. Change of the extent of the depleted region, /, versus external field Fs (p-Ge, Aâ = 10^^ cm~^,
2 3 4
Fg, lO^xVcm-''
Figure 5.14. Concentration of conduction electrons in the near-surface region, ^s, versus field strength Fs (p-Ge,
iVa = 10^^ cm-3, T = 300 K)}^^
It provides an opportunity to take into account the mutual influence of processes of field
penetration in a semiconductor and a finite conductivity current flow through a specimen. If
j < e(finf^p)^^^niFs/ ae (5.35)
an approximation of a zero conductivity current is reasonable. In this equation />t„ and fjip is
the mobility of electrons, holes, ae the dielectric penetrance of the semiconductor, Fs the field
strength on the field emission border, n/ the electron concentration in a pure semiconductor,
L the length of the tip. This applies to region I current-voltage characteristic.
If Fs increases, the current density can reach values, for which the near-surface layer
is exhausted, and the process of emission begins to depend upon inflow of electrons from
the bulk of the emitter. Our analysis^^^' ^^^ shows that if the FE current is comparable with
the maximum attainable electron inflow rate from the depth of a specimen, it is possible
to separate three areas in a crystal: near-surface n-region, following it /-region and deep
/7-region, where the concentration of holes is approximately equal to the concentration of
ionized acceptors. Of course, such a division is reasonable only if the thickness of these
areas is much greater than the thickness of the transition region between them. It is possible
to show, that this requirement is equivalent to satisfying the inequalities
NI'^ < Fs/(47TxkT)^^^ < Nl'^/ui. (5.36)
So, a similar division is impossible for semiconductors with conductivity, close to the intrin-
sic semiconductors, as well as for semiconductors with high concentration of impurities.
Calculations show, that the generation-recombination processes in the /-area give the
main contribution to the value of the saturation current. The decisions obtained made it
possible to describe two limit cases of system behavior:
1. negligible recombination in the /-region;
2. high speed recombination, when a small deviation from the state of thermo-
dynamical equilibrium of the electron-hole plasma under balance takes place.
When recombination rates in the /-region can be ignored, recombination does not
influence upon the value of the saturation current. The saturation current is determined by
the generation rate of carriers in the /-area. In the case of large recombination rates, the
current is limited not by generation of carriers in the /-area, but by the final conductivity of
this area. Calculations^^^' ^^^ have shown that, at room temperature, the saturation current
density for a p-Si emitter with a cone angle of ^10° cannot exceed 10^ A/cm^. This value
is close to experimentally observed values.
For FE from ^-semiconductors, the saturation current is determined not only by the
generation and recombination rates in the /-area, but also by the volume of this area.^^^
The presence of traps, on which significant charge can reside, leads to a change in this
volume and, thus, to a change in the saturation current. It has been shown that accounting
for electron charge on traps leads to thermal sensitivity of the saturation current. For a
cylindrical emitter,
d{\x\js) £g Ei
(5.37)
d(\/kT) 4
and for a conical emitter.
djlnj.) ^_3 1 3 33^
d{l/kT) 4 ^ 8 ^
Here Eg is the width of a forbidden band, Et the activation energy of traps.
FE of n-type semiconductors has a specific character. For low fields, the emission
current is determined by the barrier transparency D. However, when the field strength
Fs and, accordingly, D increases, the FE current is found to be restricted by the inflow
of electrons from the emitter bulk. So as for p-type semiconductors, the external field
penetrates into the bulk of an emitter up to contact, causing injection of electrons into a
conductivity band. Therefore, "screening" of an external field takes place on the injected
charge. If
Fs/47txe » A^d^, (5.39)
where L is the length of emitter tip, the role of injection of electrons from contact is a
main factor and the emission current is restricted by the space charge of the carriers in the
semiconductor. In this case
j - FIIL. (5.40)
If
Fs/47r«^-A^d^, (5.41)
the injection is small and the emission current is limited by the finite conductivity of the
emitter's material and the FE current density is
7-^/x„FsiVd/ae (5.42)
The investigation of the influence of deep trap states with concentration M and
activation energy ^t upon FE has shown^"^^' ^^^' ^^^ that traps influence on FE only when
Nt>N^, Nt> NcC\p(-6t/kT), (5.43)
where A^c is the density of states at the bottom of a conduction band. Therefore only
deep traps with high concentrations influence the FE processes. This would be expected in
broad-band n-type semiconductors (e.g., CdS).
Calculations show that, generally, in semiconductors of both n- and /7-type there are
five regions of current-voltage characteristics:
1. FN region;
2. First saturation region, j ~ /x„Fg cxp(—St/kT);
3. Second saturation, j ^ fjin(F^/L) cxp(-£t/kT);
4. A region in which the current increases rapidly;
5. Third saturation, j ~ /x^F^^/L.
The region of rapid current increase (4) is not associated with avalanche multiplication of
the carrier, but with the limited filling of traps.
The existence of these regions is shown by experimental data for single crystal CdS.
Figure 5.15 shows the current-voltage characteristics plotted in Ig 4 = /(Ig U^) coordi-
nates. These coordinates are chosen to help identify the power law regions of the curves
h ^ U^' The FN region corresponds to extremely low emission current values and was
not observed on this sample because emission current was very small and the sensitiv-
ity of the apparatus was not enough for current registration. What is seen is that at low
voltages the current-voltage characteristic have regions of linear (n = 1) and quadratic
(n = 2) dependence of emission current on the applied voltage. Saturation of the emission
current following the region in which the current rises rapidly with voltage shows that no
avalanche breakdown takes place in the emitter. When the anode voltage is decreased, the
near-surface electric field supporting the emission is maintained by charged traps. There-
fore, in semiconductors with a high (>10^^ cm~^) concentration of deep trap states, the
current-voltage characteristics show appreciable hysteresis, that is, the curves 4 = /(t/a)
taken with increasing and decreasing anode voltage do not coincide. After complete removal
of the anode voltage the deep traps remain charged for long time periods. When the voltage
is reapplied, the emission current returns to the value it had before saturation (a memory
-10
<
1-^
-12 n=1
-14
2.7 2.9 3.3 3.3 3.5 3.7
ig(â.v)
Figure 5.15. Current-voltage characteristics of a CdS single crystal tip. Regions of linear (« = 1) and quadratic
(n = 2) dependence of the emission current IQ on voltage f/a are shown.^^^
effect). Possibly, this phenomenon can be utilized in designing power-dependent storage

devices.
Recently it was reported^^^"^^^ that nonlinear current-voltage characteristics similar
to those in Fig. 5.5 have been observed for diamond-like films. The hysteresis or divergence
of the characteristics measured with increasing and decreasing voltage was observed as
well. This indicates that the processes taking place during field electron emission from
diamond-like films may have much in common with space-charge region effects in wide
band gap semiconductors.
5.7. TRANSITION PROCESSES IN FIELD EMISSION FROM

SEMICONDUCTORS
The experiments discussed above were conducted in the slow registration mode. That is,
recording of the emission characteristics usually took at least a few minutes. Therefore the
observations were of steady-state FE processes. Although electron tunneling into vacuum
is essentially instantaneous, the establishment of a stationary distribution of charge carriers
and field in the sample is characterized by finite relaxation times. Because the electron
concentration in the conduction band is not large, when the electron supply is depleted by
extracting large current densities, a bulk depletion region is formed and it is the relaxation
processes in this region that determine the character of the field electron emission. Some
data on the inertial characteristics of FE process in semiconductors are given by Arthur^^^
and Sokolskaya and Klimin.^^ Arthur observed the occurrence of hysteresis during the
measurement of transient current-voltage characteristics in p-type Ge. Sokolskaya and
Klimin noted a lag in the variation of field electron emission current from CdS during
pulsed illumination.
Stetsenko^^^ studied photo-field emission from p-Si ( 7 x 1 0 ^ ^ - cm) in a pulsed mode.
It was shown that the current pulses lagged the voltage pulse front by a time interval id which
varied from 10 to 100 |xsec depending on temperature, intensity of illumination, and the
voltage pulse amplitude; decreasing as these parameters were increased, ra was also found
to be very sensitive to the value of the fixed bias voltage. In addition, the emission current
monotonically decreased with time after reaching some peak value. Stetsenko suggested
that this behavior was related to the presence of positively charged surface states having a
concentration of ~10^^ cm~^. The kinetics of field electron emission in high-ohmic Ge and
Si under emitter illumination with rectangular light pulses of a few millisecond duration
has also been studied^^^ and a considerable delay of the photoresponse to illumination was
detected. Characteristic current rise times varied from 10~^ to 10~^ sec depended on the
density of traps in a sample. In this study, a light probe with a spot size of less than 0.2 mm
was utilized. Local illumination of isolated tip areas permitted the size of the depleted
region to be determined. It was shown that the extent of photosensitive area is many orders'
magnitude larger than the emitter radius (about 1 mm in the silicon samples studied).
5.7.1. Time Variation of the Field Electron Emission
Establishment of a stationary concentration in the emitter bulk following application

of a pulsed anode voltage is evident in the current kinetics.^^^' ^^^' ^^^' ^^^ On atomically
clean Ge and Si surfaces the current in region I of current-voltage characteristics lag by a
time interval ra that decreases with increasing anode voltage U^, (Fig. 5.16). The kinetics in
region II are of the most interest: The current 4 first reaches a maximum and then drops to
some steady state value. As Lâ is increased, the maximum in 4 is reached sooner and the
ratio, G, of the maximum current to its steady-state value increases. By the end of region II
the value of G ^ 10. The time variation of 4 reflects the process of forming the steady-state
structure of the space charge in the semiconductor emitter.
In region I the current shows no variation with time and for a square voltage pulse
a square current pulse is obtained. This behavior is to be expected because emission in
this region only depends on the transparency of the potential barrier, and not on any bulk
properties. The carrier generation rate for a given value of emission current is sufficient
to maintain the degenerate state of the electron gas near the emitter surface. As seen in
oscillograms 2-5 of Fig. 5.16a and oscillograms 5-7 of Fig. 5.16b the current drops with
time. This drop appears to be associated with the depletion of electrons in the near-surface
layer of the emitter due to a number of interrelated factors: (1) reduction in the number
of electrons incident on the potential barrier; (2) reduction of the effective field strength
F near the tip apex by redistribution of voltage between the tip and the vacuum gap; and
(3) a decrease in F due to a change of the space charge configuration. The characteristic
time is determined by the time taken to establish a new diffusion-drift equilibrium under
conditions of current flow through the sample.
In region III of the current-voltage characteristic a current rise is observed at
constant U. This current rise is seen in the beginning of region III (oscillograms 6, 7
of Fig. 5.16a and oscillogram 8 of Fig 5.16b), and is probably evidence of carrier multi-
plication in the high internal field of the crystal. Rather sluggish response of the emission
current to the change in carrier concentration due to field-generated carriers can also be
explained by the slow rearrangement of the space charge in the sample, in which the pres-
ence of traps can play an important role. We note that the current increase at the end of
region III causes heating and, in most cases, destruction of the emitter.
The results of studies of the kinetics and residual relaxation effects indicate the impor-
tant role that transient processes in the sample bulk play in establishing steady-state
(a) /
10-^
10-8
4 5
(b) 5
a^ - ^
r \
6
A_ ^
-s.
^
-• t
Figure 5.16. Oscilloscope traces showing the kinetics of the field electron emission: (a) p-type Ge (voltage
pulse duration ipuisg = 5 msec);-^^^ (b) p-type Si; traces 1 and 2 correspond to region I of the I-V characteristics;
traces 3-5 to the transition between regions I and II; trace 6 to region II; traces 7 and 8 to the transition between
regions II and III (rpuiê = 100 msec).-'^^^' ^^^
emission. The absence of residual effects in region I and their presence in region II indicates
that the processes are related to the high electric field in the sample.
The kinetics of FE from clean semiconductor surfaces has been investigated with «-type
silicon.^^"^' ^^^ Si samples doped to a wide range of concentrations and having resistivity
values of 10, 50, 300, 800, and 2000 Q, • cm~^ were studied. It was shown that the variation
with time of the FE current in n-type semiconductors (Si) is qualitatively similar to that
observed in p-typc samples.
5.8. THE STABLE SEMICONDUCTOR FIELD EMISSION CATHODE
As we have seen, experiments and theoretical studies show that the appearance of the
region of slow current variation is related to the limited supply of carriers from the semicon-
ductor bulk. It follows that beyond a certain emission current, the current is unaffected by
a further increase in barrier transparency. Direct evidence of this can be inferred from the
insensitivity of the electron emission to surface contamination, in particular, the residual
gas pressure.^^^
Studies of this insensitivity were^^^ carried out on single crystal p-type silicon with
resistivities of 14 and 200 Q • cm and on 2000 Q • cm n-type silicon. The vacuum was varied
between 5 • 10~^ and 5 • 10~^ Torr with air by a leak value. The experimental procedure
involved: (1) evacuation to ^ 5 • 10~^ Torr; (2) cleaning of the sample surface by field
desorption; (3) collection of I-V data in UHV; and (4) collection of I-V data in various
residual pressures of air.
The key results are shown in Fig. 5.17. It can be seen that the emission current is nearly
independent of the residual gas pressure in the current range corresponding to region II of
current-voltage characteristic (Fig. 5.17b, curve 1) and independent of time (Fig. 5.17b,
curve 2). Under these same vacuum conditions large current fluctuations occurred in
region I of current-voltage characteristic. Note that the extent of the saturation region
of current-voltage characteristic is highly dependent on the doping level of material
(cf. curves 1 and 2 in Fig. 5.17a). For p-Si with p — 2000 Q. - cm, the extent of this
region was >23 kV which is more than 4 times the range of voltage variation correspond-
ing to the linear part of current-voltage characteristic. This result agrees with the theoretical
premise that the extent of the depleted region^^^' ^^^ and, correspondingly, the internal field
(a)
0.5 1.0 1.5 2.0

(104/L/),V-^
(b) 10-s 10"^ 10"^ 10-5

p, Torr
f, hr
Figure 5.17. FE current as a function of residual gas pressurer^^^ (a) experimental I-V characteristics of p-Si:
1—small cone angle (2000 Q, cm); 2—large cone angle (3 Q. cm), (b) FE current as a function of residual gas
pressure (1) and time (2) (at /? = 10~^ Torr). Current values correspond to points a (curve 1) and b (curve 2) of
the current-voltage characteristic in part (a).
gradient and "breakdown strength" of a sample, is inversely proportional to the doping

level.
These results support the proposition that under certain conditions (region II of the
current-voltage characteristic) the electron emission is independent of the barrier trans-
parency and that bulk properties play the dominant role in the emission process. These
results also indicate that under such conditions the FE current is not sensitive to the surface
states introduced by adsorption of residual gases.
5.9. ADSORPTION ON SEMICONDUCTOR SURFACES
The ability to visualize adsorption and diffusion processes on different crystal sur-
faces with a 20-30 A resolution makes the FE microscopy a unique instrument studying
such phenomena. The delay in using field electron microscopy to investigate adsorp-
tion on semiconductors was due to considerable difficulties involved in producing a
clean semiconductor emitter surface. Advances in the semiconductor surface cleaning
process^^^' ^^^' '^^^' ^^^' ^^^' ^^^' ^^^ have allowed for some studies of adsorption.
The adsorption of Ba,^^^ Au and Ti^^''' ^^^ and SiO^^^ on Ge has been investigated.
The adsorption of Ga and As on GaAs has also been studied.^^^
Bakhtizin and Stepanov^^^ carried out first experiments on SiO adsorption on Ge.
Arthur^^^ studied adsorption of Ga and As on GaAs. A number of studies on oxygen
adsorption on Ge was performed.
5.9.1. Oxygen Adsorption on Ge
The first studies of oxygen adsorption on atomically clean Ge were carried out
by Emst,^^' ^^^ Ivanov and Zabotin,^^"^ Mileshkina and Bakhtizin.^^^ The most signifi-
cant observation was the highly anisotropic adsorption of oxygen on different Ge crystal
planes.^^ preferential adsorption was observed on the (100) plane and in the vicinity of
(111) plane on the (123), (144), (112), and (113) planes. These experiments were later
confirmed by Ivanov, Rozova, and Fursey.^"^^ The average sticking coefficient s was found
to be 0.1 from measurement of the time intervals required for the formation of a manolayer
at a given oxygen pressure.^^' ^^^
Using a more sophisticated oxygen source^^^ we reproduced these results of O2 adsorp-
tion on p-type Ge.^"^^ Emission images on Ge surface are shown in Fig. 5.18. It was found
that O2 adsorption affects the nonlinear portion of current-voltage characteristics by lower-
ing the saturation current and appreciably reducing the extent of saturation region. The field
desorption of oxygen resulted in complete recovery of the initial shape of the current-voltage
characteristics.
5.9.2. Electroadsorption
Electroadsorption suddenly accelerate O2 adsorption on Ge at a particular electric

field value Fg.^^^' ^"^^ It was discovered that electroadsorption effects appear at field values
near those required for Ge evaporation. The threshold field F^ can be calculated from the
Figure 5.18. Adsorption of O2 on Ge:^^'^ (a) FE pattern of Ge (111) with a clean surface after the field evaporation
of contaminations. (VD = 24.7 kV, VpE = 5.8 kV); (b) O2 adsorption on the Ge (P02 = ^^~^ ôrr, fexp =
20 min, (/pE = 6.7 kV); (c) thick layer O2 on the Ge. (P02 = ^ x 10"^ Torr, fexp = 100 min, f/pE = 7.4 kV).
equation K = U^/UQ = FS/FQ, where UQ is the voltage required for Ge evaporation, Us is

the voltage at which the acceleration of desorption takes place, and Fe is the known value
of Ge evaporation field. The effects of electroadsorption are illustrated in Fig. 5.19.
The character of the O2 coating on Ge varies under high field conditions (compare
Figs 5.18 and 5.20).^^^ It is possible to produce dense, well bound layers having thicknesses
of a few monatomic layers. The highest O2 adsorption rates are observed in the vicinity
of (001), (111), and (113) faces as evidenced by the dark spots in the emission pattern
(Fig. 5.20).
(a) (b)
6 12 18
t, min
Figure 5.19. Influence of a high electric field on O2 adsorption on Ge: (a) dependence of the "effective work
function" on the time of oxygen exposure: (1) without field; (2) with pulsed electric field, f/ads = 9 kV, f/ads =
0.46 • f/desorpofGe-^^"*^ (b) Current versus time during O2 adsorption on Ge (111): (1) K = 0, (2) K = 0.4;
(3) K = 0.45; (4) K = 0.6; (5) K = 0.7; (6) K = 0.8. {K = f/ads/^desorpof Ge)-^'*^
Figure 5.20. Adsorption of the O2 in a high electric field with Ge (111):"^^^ (a) clean Ge surface obtained by field
desorption {UD = 13.25 kV, f/pE == 3 kV, /pE = 2.5 x 10~^ A); (b) surface after exposure to O2 for 20 min at
U = 6kY (PQ^ = 10-6 Torr, f/pE = 3.63 kV, /pE = 2.5 x 10"^ A).
5.10. RESUME
This chapter outlines the most important studies of FE from semiconductors. The main
conclusions are:
1. In recent years, the technology of preparing atomically clean semiconductor

surface has been developed and FE and field ion microscopy of semiconductors
has advanced. Perfect emission images of a number of semiconductors have been

obtained.
2. The conditions have been established under which FE experiments must be
conducted in order to preserve the initial properties of the semiconductor sample.
3. Key features of FE from semiconductors have been observed experimentally and
described theoretically.
4. Research methods to study adsorption on semiconductor surfaces have been
developed and very interesting results have been obtained.
STATISTICS OF FIELD ELECTRON EMISSION
6.1. FORMULATION OF THE PROBLEM
Several fundamental physical processes involved in field electron emission can be

elucidated through measurements of the statistical distribution of "elementary emission
events," that is, the statistical distribution of the emission of either a single electron or
group of correlated electrons. In the general case, with some probability P/, one may
observe some number Ni of electrons for a given emission event. The function P(N), is
the distribution of A^ over different events.
Investigations of secondary and photoelectron emission have shown that groups con-
taining different numbers of correlated electrons have been observed in the emission current.
The information obtained in these experiments enables one to obtain not only the fundamen-
tal parameters of the process (the secondary emission coefficient, quantum exit, fluctuations)
but also aids in elucidating the physical mechanisms underlying the emission phenomenon.
The unique feature of field emission (FE) is that electrons tunnel into vacuum with-
out external excitation. Investigation of the emission statistics associated with a non-
excited electron gas not only allows us to quantitatively evaluate the applicability of the
single electron approximation to the FE process, but also aids in elucidating fundamental
aspects of solid-state electronics; in particular, those involving electron-phonon inter-
actions and correlation effects that yield information about electron-electron interactions
(see Chapter 1, Section 1.3.3).
In a number of papers, the possibility of groups of electrons tunneling into vacuum,
correlated in space and time, have been discussed. For instance, Young^^^ suggested that the
anomalously large value of the work function of the tungsten (W) {110} plane may be due
to tunnelling electron pairs. The possibility of pair tunnelling was explained either by (1)
fluctuations in the transparency of the potential barrier, (2) lattice thermal oscillations,^^^
or (3) the adsorption of residual gas molecules.^^^ Lee and Gomer^^^ have shown that pair
tunneling would explain the "high energy tail" observed in field electron energy distribu-
tions (see Chapter 1, Fig. 1.7). Lastly, the statistics of FE from semiconductors contains
information about the process of the creation of current carriers when field penetration is
involved and the search for electron pairs in FE from superconducting materials is of special
interest.
105
6.2. METHOD OF INVESTIGATION
The methodology of measuring emission spectra statistics is shown in Fig. 6.1. The
FE tip serves as the cathode. The emitted electrons, after passing through a hole in the
first anode, are accelerated by an auxiliary electrode to increase their energy in order to
facilitate detection. Upon arriving at the detector a signal is produced with an amplitude
proportional to the number of electrons. The pulse amplitude analyzer counts the number
of events associated with the arrival of groups of electrons containing one, two, three, or
more electrons.
The first direct measurement of statistical FE events was reported by Herman.^^^
The emitted electrons were accelerated to 15keV whereupon they entered a gas-filled
proportional counter. The absence of electron groups was observed with an accuracy of
1 percent. Later, observing field electron emission from spontaneous microprotrusions on
metal surface, groups of two, three, and four electrons, were detected^^^ with the help of
a proportional semiconductor detector. The pairs were several percent of the total number
of electrons. Note that in these previous investigations,^^^' ^^^ the emitter surface was not
cleaned, and the measurements were carried out only under "technical" vacuum conditions
(^10-6 Torr).
/ V \
/ . 1 ^ Anode . \
I 2 Anode |
Detector
Amplifier Analyzer
^ p(.)=.^('')
lS{n)
N
2E
ill 1^^
Figure 6.1. Schematic of the experimental apparatus used to study statistical processes in FE.
STATISTICS OF FIELD ELECTRON EMISSION 107
Measurements of FE statistics from atomically clean surface of polycrystalline tung-

sten with a reproducible current-voltage dependence in ultra-high vacuum have been
conducted.^^^' ^^^ This research has shown that FE from W is to a considerable extent a sin-
gle electron emission process. However, systematic investigations of FE statistics became
possible only after improvements in the experimental device^^^ wherein surface cleanliness
and perfection were controlled by emission patterns. FE from different crystallographic
planes was studied using a rotating cathode assembly and special measures were taken to
suppress emission from intermediate electrodes.
Figure 6.2 is a schematic of this improved experimental arrangement.^^"^ The field
emitter (1) is fixed in the rotating system. Motion of the cathode is accomplished with
bellows (2) which enables one to observe any region of the FE image by the probe hole in
the screen (3). The electrons are registered by the semiconductor detector (4). The cathode
may be heated by passing current through the support loop and cooled by liquid nitrogen.
Special attention was paid to make the apparatus highly sensitive to detecting groups of
correlated electrons.. The counting rate was 500 electrons/sec. This value was considered
as the optimal rate because it enabled us to establish the appearance of electron pairs when
they comprised more than 0.1 percent of the total emission current.
Figure 6.2. Device for the investigation of FE statistics.

6.3. FIELD EMISSION STATISTICS FROM METALS
The first complete investigation of the statistics of FE from W is described by Fursey

et al.^^^ Here emission processes from various facets; (001), (111), (110), and (112)
(Fig. 6.3), were investigated for which anomalies in energy distributions of field electrons
have been observed.^^^ In addition, emission processes were studied from all other regions
of the emission pattern, in particular, the (120), (130), (113), (115), (116), and (117) facets.
Pair tunneling was not observed from any crystallographic direction. Some of the spectra
obtained are shown in Fig. 6.4 and analysis of some of the spectra in Table 6.1.
The role of thermal fluctuations in the FE statistics of W has also been studied^^^ in the
temperature range from 77 to 1000 K. For the temperatures investigated, FE had a single
electron character. The influence of the adsorption of residual gases on the FE statistics
was also investigated. In the first of two experiments, the investigations were performed
with different degrees of residual gas coverage on the W surface at a background pressure
of 5 • 10~^^ Torr. In the second experiment the statistics were measured at a pressure of
10"^ Torr, as in Gazier.^^^ Again, many-particle effects were not observed (Figs. 6.4e,
6.4f). These results enable Fursey et al.^^^ to infer that the multipeak spectra observed in
Gazier-^^^ are connected with parasitic secondary emission from intermediate electrodes. In
measurements carried out with an intermediate accelerating electrode, a secondary emis-
sion multipeak spectrum (Fig. 6.4g) was obtained. Later, a detailed investigation of the
FE statistics for tantalum (Ta), molybdenum (Mo), and niobium (Nb)^^^ was conducted.
Figure 6.3. re pattern of W: a, b, c, d - locations of the probing holes for the emission patterns of W corresponding
to crystallographic directions (010), (122), (Oil), and (112) respectively.
(a) (b) (c) (d)

A
3 - l\ (001) 3 (111) 3 1 (110) 3- l\ (112)

•
-f
•
2 2
•
' '
1 - 1 1
J, I ,
-
10
\
30
J\
10
1 1
30
d I ,
10 30
4\ :
•
10 30
10
a^ 30
Figure 6.4. FE statistics spectra for W: (a, b, c, d) for an atomically clean surface; (e) with residual gases adsorbed
in a vacuum of 5 X IQ-Ô j o ^ . (f)
in a vacuum of 10 ^ Torr; (g) with additional acceleration. Vertical scale:
pulses (A^ X 10~^); horizontal scale: energy {ktV)P^
Table 6.1. Relative height of the second peak in

spectra from metals
IS(2)/S(\)] X 10^
Facet (hkl) W Mo Ta Nb
111 8 8 8.5 9
112 9 8.5 9 9
110 8.5 8 9 8
120 9 8 8 9
116 9 9 9 8
001 9 9 8 8.5
130 8.5 — — —
332 9 — — —
The results presented in Table 6.1 show, that in the temperature range of 77-1000 K, FE
from the metals is single particle to within an accuracy of 0.1 percent.
6.4. FIELD EMISSION STATISTICS AT CRYOGENIC TEMPERATURES
The experimental apparatus for the investigation of FE statistics at liquid helium (He)
temperature (4.2 K) is described in Fursey et al.^^^ A schematic of the experimental arrange-
ment used to investigate FE statistics at liquid He temperature is shown in Fig. 6.5. The
Figure 6.5. Schematic of the device for investigating FE statistics at hquid He temperature 376
glass vacuum tube (1) is placed into a cryostat (2). Transfer of the cathode (3) is carried
out by bellows (4) with the help of a mechanical manipulator (5,6). The FE image on the
screen (7) is observed by a fiber optic assembly (8). The semiconductor detector (9) and
the amplifier (10) are maintained at their optimum operational temperatures by heaters. In
this device it is also possible to investigate the electron emission characteristics of super-
conductors in high magnetic fields. The superconducting electromagnet (11) can create a
magnetic field magnitude of 1 T in the region of the emitter.
A comparative investigation of W and Nb at 4.2 K has been reported.^^^' ^^^ It is known
that W remains in the normal (nonsuperconducting) state at 4.2 K. Spectra of the FE statis-
tics have been obtained by probing individual facets (low index crystal planes) such as the
(110), (112), and (120) planes. For all the W facets investigated, FE had a single-electron
character within the accuracy mentioned above (P(2) < 0.1 percent). We also investi-
gated superconduction single crystals of Nb. The measurements of the FE statistics were
conducted on different crystallographic facets of an atomically clean Nb surface. For all
values of current density (10~^-10~'^ A/cm^) and field strength investigated, FE from Nb
in the superconducting state had a single-electron character (Fig. 6.6). The presence of a
magnetic field did not induce any changes in the character of the electron emission from
superconductiong Nb, that is, the spectrum of the FE statistics remained a single-electron
for all values of magnetic field up to 1T.
Figure 6.6. FE spectrum for Nb at 4.2 K. 378
The absence of electron pairs in these experiments may be related to the dispersion of
the electrons and their low collection efficiency by the probe hole in front of the detector.
The space between electron groups on the emitter surface is limited by their correlation
length, which for Nb is equal to 2000 A. Such an area corresponds to a part of the screen of
the FE projector exceeding the dimensions of the usual probe window. For this reason, it is
necessary to study superconductors with a comparatively small correlation length; Nb-Ti
(300A),Nb-Sn(70A),and ceramics (20 A). ^'^'7
FE statistics from a Nb-Ti alloy were measured in the temperature range of 4.2-
300 K.^"^^ This superconductor has a critical temperature of 18 K and a correlation length of
300 A. Emitter tips were prepared by electrochemical etching. Prior to experiments we first
measured the current-voltage characteristics of the Nb-Ti emitters and then observed the
FE pattern. The current-voltage dependence exhibited linear Fowler-Nordheim (FN) char-
acteristics. The emission pattern was symmetric after thermal cleaning of the crystal emitter.
FE statistics of the Nb-Ti alloy were investigated for current densities of 10~^-10~^ A/cm^
from various crystallographic directions. Measurements at He temperature were carried out
in a cryostat where a fiber-optic bundle allowed observation of the FE pattern to check the
degree of cleanliness of the emitter surface. Measurements in the cryostat were carried out
at a vacuum of 10~^^ Torr. Analysis of the spectra of the FE statistics obtained for Nb-Ti
revealed that in the temperature region 4.2-300 K, FE has a single-electron character to an
accuracy of 99.9 percent (Fig. 6.7).
6.5. MULTIELECTRON FIELD EMISSION FROM HIGH TEMPERATURE

SUPERCONDUCTING CERAMICS
For high temperature superconductors FE statistics were reported first in Fursey et al. ^"^^
In this paper many-particle tunneling was observed from yttrium (Y) ceramics (Fig. 6.8).
The maximum fraction of paired events was 10 percent. The fraction of three electron and
four electron groups was roughly 1 and 0.2 percent, respectively.
The sample was sharpened mechanically by diamond polishing. Emission was observed
to occur from separate microtips (protuberances) on the surface of the sample yielding
0)
V)
Q.
Q.
I
X
40
H20
_L_J^
10 20
E, keV
Figure 6.7. FE spectrum for Nb-Ti at 4.2 K.
Hioo
3
Q.
50
A
12.5 25 37.5 50
E, keV
Figure 6.8. ra spectrum for Y ceramics. ^^^
an emission pattern with an irregular structure in the form of separate bright emission
regions. The superconducting properties of the ceramics were checked before and after the
experiments by observing the Weissner effect. The vacuum in the device at 300 K was 10~^
Torr, and at 4.2 K, better than 10"^^ Torr.
Up to the present, current-voltage characteristics, the emission images, and the FE
statistics for Y high temperature superconductor single crystals and for Y and Yi high tem-
perature superconductor ceramics have been investigated.^''^ It has been established that
multielectron FE takes place only from separate, weakly emitting regions of Y-based ceram-
ics. The current-voltage dependence of such samples (Fig. 6.9) has some unusual features
also found with semiconducting field emitters: Both have nonlinear FN characteristics that
are temperature sensitive.
o
I^H o o
o
-ioh ''^ °
o
I I I I I
0.14 0.18 0.22 0.26 0.30
MU, keV-^
Figure 6.9. Current-voltage dependence of Y ceramics.
It is important to note that multielectron FE was observed from 4.2 to 300 K, the critical
temperature being equal to 90 K. It is clear that the effect observed is not related directly
to the superconducting state. This observation cannot be explained at present and requires
additional investigation.
6.6. INVESTIGATIONS OF FIELD EMISSION STATISTICS FOR fflGHLY

TRANSPARENT BARRIERS
The probability of multiparticle tunneling essentially depends on the transparency of

the potential barrier.^^^ A considerable part (several percent) of the tunneling current from
W into vacuum involves electron pairs at FE current densities from 10^ to 10"^ A/crn^}^^
This range of current density was necessary to maintain the number of random double
superpositions at a level of 0.1 percent as the resolution of the amplitude analyzer was
~10-^ sec.
The only practical way to investigate FE statistics at current densities in the range of
10^-10"^ A/cm^ is to use very short detection time intervals (less than 10~^^sec). Such
studies have been conducted.^^^ The idea is based upon quickly deflecting a narrowly
focused electron beam with the leading edge of a nanosecond duration voltage pulse. The
time it takes for the beam to pass through a probe hole placed in front of the detector
is the detection time. The minimum time between the pulses should be larger than delay
time associated with the total detection system (detector and associated electronics). For
conventional devices this time is of the order of 10"^ sec. The experimental apparatus
(Fig. 6.10) includes a FE projector with a probe window, an electron focusing and deflection
system, and a semiconductor detector.
The emission statistics for W was investigated up to current densities of 10 A/cm^.
The spectra obtained (Fig. 6.11) show that the second peak corresponds only to occasional
superpositions; FE has a single particle character. It should be mentioned that the full
capacity of this device was not completely utilized; it is possible to make the time gate
smaller than 10~^^ sec and thereby increase the FE current density to 10"^ A/cm^.
Cathode
1^1 ^^ 'H^'
^=^
Screen Anode
Adjusting plates - ^Dl
Focus system —
Deflecting yoke
Deflecting plates -OS

Collector
Si-detector
Figure 6.10. Schematic of the experimental apparatus for the investigation of FE statistics at high current densities.
30
Q.
50
_L
20 50
E, keV
Figure 6.11. I ^ statistics spectrum for W at a current density of 10 A/cm^.
6.7. RESUME
1. It is to be noted that a correct method to determine the statistical distribution of

"elementary emission events" has been developed by now.
2. Studies of the rate of elementary events and statistics of the field electron emission
in direct experiments confirm the adequacy of the model of free electrons for the
field electron emission from metals with an accuracy of not worse than 0.1 percent.
In some materials, in particular, superconducting Y ceramics, multiple-particle tunneling
has been observed. Explanation of this effect has not yet been given.
7
THE USE OF FIELD EMISSION CATHODES IN
ELECTRON OPTICAL SYSTEMS: EMISSION
LOCALIZATION TO SMALL SOLID ANGLES
7.1. INTRODUCTION
In this chapter several approaches to the fabrication of high-brightness field emission

(FE) electron sources are discussed. Many electron optical systems require that the electron
beam be formed into a very small spot, of 10 nm or less. The current density per unit solid
angle, or brightness p, of the electron source, is limited by the Langmuir equation to a
value ^max-
Bra^ = jeV/(7rkT). (7.1)
where j is the current density at the cathode, T is the cathode temperature, k is the
Boltzmann's constant, and e is the electron charge. To exceed this Langmuir limit, the
size of the cathode must be reduced, and the current density increased. Field electron
emission meets these requirements. Unique features of FE electron sources are:
1. very high-energy efficiency;
2. available steady-state FE current densities approaching 10^-10^ A/cm^;
3. a very small ('--0.4 eV) electron energy spread;
4. a high beam coherence due to the narrow energy spread and the use of small (atomic
dimension) emission areas.
Conventional tungsten (W) FE emitters have been utilized in a new generation of
high-resolution electron probe devices, such as: scanning electron microscopes having
resolutions approaching several Angstroms; transmission electron microscopes with reso-
lutions approaching one Angstrom over a wide range of electron beam energies; electron
lithography machines; Auger spectrometers; X-ray micro-analyzers; etc.
FE can be further exploited in such devices if the emission process can be localized
to a small area on the cathode surface. Several procedures leading to localization of the
emission process are reviewed in this chapter:
1. Thermally activated, field-assisted, surface self-diffusion, or "thermal-field
build-up" of the emitter surface^ ^"^'*' ^^^' ^^^ can result in localization of the elec-
tron emission by the formation of sharp ridges at the edges of crystal facets in
115
the vicinity of the emitter tip's apex. These ridges serve as regions of local field
enhancement, and thus preferentially emit electrons. Thermal-field build-up using
pulsed fields has recently been shown to produce ridges of atomic sharpness.^"*^
2. Adsorption on the emitter tip surface can result in a decrease in the surface work
function of certain crystal planes, thereby locaUzing the FE to those planes.^^' ^^^ It
has been shown that emission regions having smaller dimensions can be obtained
if high electric fields are applied while heating the adsorbed layer.^^^ The example
of ZrO on W will be discussed.
3. Formation of very sharp points having a single atom at the apex of a tip effec-
tively localizes the emission to that atom by field enhancement. This method of FE
localization offers many opportunities for electron holography.^^^' ^^^ The possi-
bility of developing high-resolution electron transmission microscopy and electron
holography without utilizing focusing electron beam lenses is also discussed.
7.2. THE OPTIMUM CRYSTALLOGRAPHIC ORIENTATION

OF THE FIELD EMISSION CATHODE
A conventional FE cathode whose surface has been cleaned, smoothed, and rounded by
high temperature thermal flashing serves as a virtual point electron source having dimensions
of the order of 10~^ cm. The cathode's surface is comprised of different crystallographic
planes having different work functions and local radii of curvature. By the Fowler-Nordheim
(FN) equation, one sees that these distinctions in geometry and work functions result in a
variation in the emitted electron current density over apex of the cathode tip. If the cathode
has been formed from extruded polycrystalline wire of a material belonging to the body-
centered-cubic (bcc) class, the apex of the tip is typically a (110) oriented crystal grain.
Unfortunately, the (110) direction has a relatively high work function in addition to being
a relatively large and flat crystal plane on the tip's surface, resulting in the emission of a
small electron current density along the tip axis (see Fig. 7.1b). It is preferred in electron
optical applications that the most intense electron emission be directed along the optical
axis of the device. To increase the current density emitted on axis using bcc materials such
as W, the emitter tip cathode is typically fabricated from zone-refined, single crystal wire.
With tungsten, orientations of choice have been (111), (100), and (310) (see Fig. 7.1a,c).
Figure 7.1. Distribution of field electron emission over the cathode surface for different crystallographic
orientationsof a Wtip: (a) (111); (b) (Oil); (c) (001).
FE CATHODES IN ELECTRON OPTICAL SYSTEMS 117
Table 7.1. Comparison of FE electron sources with conventional ones

Parameters of electron source FE (cold cathode) LaB6 (thermoc athode) Tungsten filament
Brightness, A • cm~^ • sr~^ lO^-lQlO 107 10^

Dimension, A <102 10^ >10^
Energy spread, eV 0.22 2.0 2.0
Service life, hr >4000 1000 40
Vacuum conditions, Torr 10-9-10-10 10-7 10-^
Devices utilizing such FE cathodes are already designed, and outstanding results have
been obtained, for example, in electron microscopes developed by Hitachi and Amray. The
advantages of these electron sources, over conventional thermionic cathodes, have been
demonstrated quite unequivocally with several parameters: the electron source brightness
and coherency, the electron beam energy distribution, noise characteristics, and service life.
(Table 7.1).
7.3. FIELD EMISSION LOCALIZATION BY

THERMAL-FIELD SURFACE SELF-DIFFUSION
7.3.1. Localization by Build-Up Processes
Emission into angles approaching 100° at a current density of ^^10^ A/cm^ is typical
for thermally annealed FE cathodes. Such large beam divergences are undesirable in an
electron source since they are associated with low source brightnesses. In addition, a number
of technological difficulties arise, for example, since the majority of the beam current is
intercepted by apertures electron stimulated desorption impedes the maintenance of high-
vacuum conditions in the electron gun. These problems are somewhat alleviated by selection
of the proper emitter tip orientation; however, much more significant improvements in the
source brightness can be achieved through emission localization techniques, such as those
involving build-up phenomena.
It has been shown^^^"^^^ that micro-ridges and crests can be formed in the apex region
of a W emitter by heating the tip to temperatures of 1800-2200 K in the presence of an
applied electric field F2 at the apex of the emitter, where
F2 > (167ry/r)i/2, (7.2)
where y is the surface tension of the emitter tip material, and r is the radius of the tip
endcap. The micro-ridges are formed in a few tens of seconds under these conditions. It
was later shown^^^ that these ridges also form over a time period of several tens of hours at
similar temperatures if the applied field is of the order of Fi = (Sny/r)^^'^.
The effects of such thermal-field processes on the electron emission distribution of
the emitter tip cathodes is shown in Fig. 7.2. It is clear that emission can be enhanced in the
(100), (111), and (013) directions. Note in this figure that the low index planes such as the
(110) and (211) planes grow in size during this process, to keep the surface free energy a
Figure 7.2. Localization of RE to several crystallographic planes by thermal-field build-up.
minimum. The edges formed by these planes are the most highly emitting, due to the large
local field enhancement factor ^ = F /U, where F is the magnitude of the applied electric
field and U the applied voltage. Changes in the electron emission spatial distribution that
occur during such build-up process are shown for the case of (100) buildup in Fig. 7.3.
7.3.2. Localization by Micro-protrusion Growth
The FE current can be even more highly confined if a sharp micro-protrusion is formed
at the apex of the emitter tip. The formation of such micro-protrusions in a high electric
field was observed in early studies by Becker,^^^ Sokolskaya,^^^ Drechsler,^^^ and others.
To date, the mechanism of protrusion formation, as well as their shape and crystallographic
structure, have been only partly established.^^^"^^^' ^^^' ^^^
(a) • ^ ^ • M i i ^ H (b)|
Figure 7.3. FE localization on a built-up W (100) tip: (a) the initial surface; (b) after some build-up; (c) after
further build-up.
At present, three general kinds of such micro-protrusions have been identified: Type I:
comparatively rough structures having dimensions comparable to those of basic grains (for
a typical tip with a radius of 10~^ cm, the protrusion's dimensions are of the order of 3-5 x
10~^ cm); Type 11: fine micro-protrusions having nanometer dimensions (10~^-10~^ cm);
Type III: atomically sharp ridges and crests. The formation of atomically sharp structures
(Type III) was recently observed when very high electric field pulses were applied to (and
the resulting large emitted current densities were extracted from) emitter tips.^"^^
Several experimental difficulties are involved in the investigation of micro-protrusion
growth: (1) All three type of protrusions form simultaneously making it difficult to determine
the type of protrusion which corresponds to a particular emission site by conventional FE
microscopy; (2) High spatial resolution is required for studying the second and third types
due to their proximity to one another. The combination of two microscopy based approaches
must be applied if a better understanding of these growth phenomena is to be achieved:
(a) FE and the high-resolution transmission/scanning electron microscopy^^^' ^^^' ^^^; (b) FE
and field ion microscopy.^^^' ^^^
7.3.3. Large Micro-protrusions
It has been shown that large micro-protrusions are characterized by a slightly truncated,
pyramidal shape. Some typical shapes of this type of outgrowth on a W emitter are shown in
Fig. 7.4. Figure 7.4c shows transmission electron microscope images of the emitter shape,^^^
and Fig. 7.4a are schematics of these images.^^^ Figure 7.4b shows the corresponding field
electron emission images.^^^' ^^^ The FE micrographs were taken after brief heating cycles to
slightly smooth the surface. The fact that the outgrowths are a single crystalline and oriented
normal to the corresponding close-packed crystal facets was confirmed by subsequent field
ion imaging and field evaporation of the protuberances (see Fig. 7.5).^^^' ^^^ It has been
proposed that these outgrowths are the result of "auto-epitaxial" growth of the close-packed
facets, where the source material is comprised of atoms migrating from the sides, i.e. shank,
of the emitter tip.388-390,400,401
The process of micro-protrusion formation is typically performed with the tip held
at a positive potential relative to its counter-electrode (negative field). In this case micro-
protrusion growth occurs simultaneously with field evaporation of the emitter tip, resulting
in an overall shortening of the emitter tip.^^^ This fact was misinterpreted is several
(a) (b) (c)
Figure 7.4. Typical shapes of micro-protrusions on a W tip: (a) schematic representation; (b) FE micrographs;
(c) transmission electron microscope shadow graphs.
(a) ^^^^^^^^m ( b ) ^ ^ ^ ^ ^ ^ ^ ^ H (c)|
Figure 7.5. Micro-protrusions observed in the field ion microscope, 389
Figure 7.6. Field outgrowths at a W surface: (a) positive (negative potential on tip) field growth, observed in a
transmission electron microscope; (b) negative field growth, observed in a scanning electron microscope.
instances and used as an argument against the formation of micro-protrusions by a growth

process.^^^'*^^ These arguments were countered by studies of micro-protrusion forma-
tion in a reversed field, where the emitter tip is held at a negative potential relative to its
counter-electrode (positive field), thereby eliminating the effects of field evaporation.^^"*"^^^
Direct evidence of micro-protrusion outgrowth in high electric fields was obtained by
conventional electron microscopy.^^"^' ^^^ A transmission electron micrograph of a built-up
tip, at a resolution of ^1 nm, is shown in Fig. 7.6a. The pattern of the surface formed
by build-up in a negative field (Fig. 7.6b) is similar to that obtained with a positive field,
although the surface structure formed during build-up in a negative field shows effects of
partial field evaporation of the pyramidal protrusion. The tip profile obtained during build-
up in a positive field is similar to the profiles observed earlier with build-up in a negative
field only at lower imaging resolution.^^^' ^^^
The similarity in the shape of the tips as the result of build-up in both positive and
negative fields is consistent with the formation of micro-protrusions by a growth process.
The suggestion that micro-protrusions are formed due to the removal of surrounding emitter
tip material by field or thermal evaporation^^ ^"^^^ is not consistent with the experimental
results.^^^
7.3.4. Fine Micro-protrusions
Localized electron emission due to build-up processes was observed in early work
with the FE microscope.^^^' ^^^' "^^^ It was immediately realized that this localized emission
was a result of the presence of micro-protrusions on the emitter tip surface. Initially it was
thought that the protuberances were merely random clusters of atoms on the tip surface.^^^
Later it was realized that they were actually crystalline.^^^' ^^^ Typically these protrusions
had a ridge structure with dimensions on the order of 50 A with the radius of curvature of
the ridge's edge of the order of a few angstroms.
It is presumed that the growth mechanism of these small micro-protuberances is similar
to the larger outgrowth discussed previously. These protuberances typically take the shape of
short, pointed hillocks, or comparatively long needles. Both types form most readily: (a) in
the neighborhood of the (110), (111), and (211) facets where the surface is comparatively
rough, where atom migration is hindered, and the conditions for crystal nucleation are
favorable,^^^' ^^ and (b) in the regions of greatest field strength, that is, in regions where
facets meet (see Fig. 7.4).^^^
Comparatively high temperatures (T = 1700-2000 K)^^^' ^^^ and high electric fields,
FB,^^^ where
F3=2(167r>//r)i/2 (7 3)
are the most favorable for the formation of these micro-protrusions. A comparison between
experimental results^^^' ^^^ and Eq. (7.3) is shown in Fig. 7.7. Additional studies of the
conditions leading to the formation of these protuberances were made in work."*^^ The
process of protuberance formation has been shown to be the same for a wide variety of
materials (W, Mo, Ta, Nb, Ir, Re).^^^
10^ Fevapor(1990K)
8
E
6 ^^^^^ ^^^^^ ^ ^ ^ i
5 ^'V^.^^V^^ **V^ *
4
3
^^V*^^ * ^"^H)^
2 +—1
0—2
*—3
^\^]^^\^^ ^3
10^ - ^^''^/^"^Fa
1 1 1 • 1 • • II • •• ~ ^ 1
10"® 2 3 4 5 6 8 10-5 2 3 4 5 6 8 10"^
r, cm
Figure 7.7. Dependence of Fi, F2, and F3 on the tip radius r and experimental values of the fields: (1) FE
measurements;^^^' ^^^ (2) electron microscopy investigations;^^^ (3) Drechsler's results.^^^
The presence of small protuberances allows for the confinement of FE to very small
regions on the emitter tip surface. Several desired characteristics of an emission localization
process allow for: (1) the ability to periodically heat the emitter tip for cleaning; (2) the
ability to reproducibly localize the emission to a given region of the tip surface; (3) high
current stability during operation; and (4) the generation of one emission site (for beam
applications).
It appears that the work of Crewe and coworkers^ ^"^'^ is among the first successful
attempts to use built-up emitter tips in a technological application. Such tips, following
build-up in the (310) direction on W, were used in a scanning electron microscope to
obtain a record resolution at that time (^^5 A at 25 kV). Single uranium and thorium atoms,
suspended in a specially fabricated polymer compound, were resolved. The maximum FE
current was 500 ix A, with the current stability being better than 4 percent over time periods
of-90min.'7i'73
Swanson and Grouser^^^ localized the emission by (100> build-up on W, and thereby
reduced the beam divergence angle to 18°. Pautov and Sokolskaya"^^^ obtained a beam
divergence angle of 12° on tantalum (Ta) in the (111 > direction and noted that similar results
would be expected with niobium (Nb). In this case of (111) build-up, a (111) oriented tip
was used, resulting in confinement of the emission on the tip axis. Single micro-ridges
emerged as a result of the enlargement of the {110} and {112} planes.
Further studies of build-up on (111) oriented tips have been conducted.^^^ By varying
both the field and temperature, it was possible to form a single micro-ridge and remove the
satellites (see Fig. 7.8). The authors also proposed a method of fine micro-ridge formation
in the (Oil) direction by the adsorption of silicon (Si) atoms on a W surface. The possibility
of the formation of a ridge at the (Oil) face is increased significantly in this case and the
ridge has a crystalline structure; FE and field ion micrographs are shown in Fig. 7.9. In these
micrographs the crystallinity of the structure is evident with the lattice constant appearing
to be close to that of a pure W lattice. Experiments with an atom probe have shown that the
micro-ridge is composed of a mixture of Si and W atoms.^^^
Finally, it should be noted that the micro-protrusion formation process is very sensitive
to the presence of adsorbates. The main drawbacks in the use of micro-protrusions for
emission localization in the build-up process are: (a) difficulties in the controlled generation
of micro-protrusions of a pre-set shape in afixedplace at the cathode surface; (b) poor beam
current stability over extended periods of cathode operation, and (c) the difficulty in restoring
the emitter tip emission characteristics following degradation.
Figure 7.8. Field electron images of a W (111) tip with (111) during thermal-field treatment, (a) the initial W tip;
(b) the built-up W tip; (c) localization of the FE by the formation of a single emission site.
Figure 7.9. FE (a) and field ion micrographs (b, c, d) of micro-ridges formed at W (011) crystal planes/
7.4. FIELD EMISSION LOCALIZATION BY A

LOCAL WORK FUNCTION DECREASE
7.4.1. Basic Principles of Localization by Decreasing the Work Function
The selective adsorption of electropositive atoms on particular crystal planes, thereby

resulting in the reduction of the work function of these planes, is one of the most elegant ways
to localize FE on the cathode surface. If the crystal plane upon which selective adsorption
occurs is at the apex of the emitter tip cathode a reasonably paraxial electron beam can be
formed from a conventional thermally annealed emitter tip.
The adsorption of zirconium (Zr) on W is the most illustrative example of selective
work function reduction by adsorption. The absorbed Zr forms small islands in a vicinity
of the (100) facet of W. The first mention of the possibility of FE localization due to the
deposition of Zr atoms on W appeared in a review by Dyke.^^ The emission characteristics
from the adsorbed Zr islands have been investigated under various vacuum conditions and
at maximum current densities.^^'^^^'^^^^^^ Our research group suggested the use of W

tips oriented in the (100) direction.^^^''*^^''^^'^ Comprehensive investigations of the ZrAV
system were also carried out by Swanson and Martin and Tuggle et al."^^^' ^^^
A number of other systems, similar to the ZrAV system have been investigated. Hf on
W and Mo was investigated by Shrednik^^^ who demonstrated that emission localization
similar to that observed with Zr on W occurs. A series of adsorbates (Al, Cr, Ge, Mb, Ti, Si,
Zr, Hf) on W and Mo have also been suggested."*^^ However, the most thoroughly studied
system of Zr on W offers the greatest potential. It is therefore considered in greater detail
below.
7.4.2. ZronW
Thefirstthorough investigation of the topology of Zr layers adsorbed on W was reported

by Shrednik.^^^ Through the use of field ion microscopy, he was able to show that following
the deposition of monolayer coverages of Zr, the bright emission regions in the vicinity of
W (100) planes were due to a local decrease in the work function and not due to local field
enhancement. The minimum work function measured for the W/Zr system"^^^ was 2.62 eV.
It was also shown^^^ that with thicker Zr coverages, the contrast in the emission micrographs
is not only due to the lower work function, but also due to field enhancement by roughness
at an atomic level. It was suggested that the formation of close-packed Zr layers may be
due to the small lattice mismatch between Zr and W. These close packed layers were also
reported by other researchers.^^' ^^^
Actually, two types of structures can be assigned to the Zr overlayers: (1) close-packed
layers yielding fairly homogeneous emission throughout the adsorption region (Fig. 7.10a);
and (2) more loosely packed layers that form a series of concentric rings with a dark region
in the center (Fig. 7.10b). This ring structure (Fig. 7.10c) is clearly visible in the field
ion microscope (see Fig. 7.10d). With the close-packed layers rapid surface migration is
visible at 1400-1500 K. The concentric ring shape of the Zr layers appears to be more
stable, possibly due to the fact that in this configuration, the Zr atoms reside on the terraces
surrounding the {100} facet, which may hinder surface migration. The presence of the low
emission region located in the center of the emission distribution is not convenient for
electron optical device applications.
Illustration of emission localization to a small solid angle using ZrO on W (100) is
presented in Fig. 7.11.
7.4.3. The Influence of the Vacuum Conditions
Several important considerations involved in the use of adsorbates for selective work
function reduction and localization of the emission are: (1) The lifetime of the adsorbate's
properties that lead to emission localization, (2) Reactivation of the localization mechanism,
and (3) Replacement of the adsorbate. A number of experiments have been conducted in
which these issues have been investigated.^^ At vacuums of 10~^ Torr, it was found that the
emission current was stable for ^ 1 hr with the Zr overcoating, whereas for clean W under
the same conditions, decay in the emission current occurred in 10-15 min.
Cathode stability, as well as the reproducibility of the current-voltage characteristics,
over time periods of 136 hr were also investigated. Curve I, in Fig. 7.12 is data accumulated
Figure 7.10. Field electron and ion micrographs of Zr layers on W: (a) close-packed layers; (b) concentric ring
layersr^ (c) concentric ring layers, '-^ (d) hydrogen field ion pattern,. 4 1 3
Phosphor
anode
Figure 7.11. Emission localization to a small solid angle using ZrO on W (100).
-6
-7
^ - % \
^-^ \ ^v XV
^-8 V
]B)
V N\
-9 \
III
\ \ 1
Figure 7.12. Restoration of emission characteristics after operating 2^ coated W in a vacuum of 10~^ Torr, ^
(I) immediately after Zr deposition; (II) after 136 hr of operation (P = 10~^ Torr); (III) after annealing at
T = 1300 K.
immediately following deposition of the Zr overlayer, and Curve II subsequent to 136hr

of cathode operation. It is apparent that the emission characteristics have changed signifi-
cantly during this time period, presumably due to adsorption processes. Note however, that
the original current-voltage characteristic can nearly be restored following heating of the
cathode to 1200-1300 K for 30 sec. as shown in Curve III. This process can be repeatably
reproduced over several degradation/reactivation cycles. Similarly the emission character-
istics following deposition of additional Zr are repeatable if heating times and temperatures
are accurately reproduced.
Several experiments have been conducted in order to examine the possibility of employ-
ing the ZrAV system under less stringent vacuum conditions: (a) the effects of admitting
laboratory air; (b) restoration of current-voltage characteristics in vacuums of 5 x
10~^-10~^ Torr, and (c) the effects of venting to laboratory air to 10~^ Torr followed
by reevacuation to 5 x 10~^ Torr.
The results of these experiments lead to the following conclusions.
(1) The emission characteristics can only be restored by the redeposition of Zr

following venting to atmospheric pressure of laboratory air.
(2) Gradual venting to laboratory air up to pressures of 5 x 10~^-10~^ Torr results
in degradation of the emission confinement almost to the point of the predepo-
sition characteristics. The emission localization is partly restored following heat
treatment to 1200-1300 K, presumably by contaminant desorption.
(3) An abrupt pressure increase to 10~^ Torr substantially reduces the emission, and
localization totally disappears. Localization cannot be restored by heating.
7.4.4. The Dispenser Cathode Approach
A major practical problem associated with the localization of FE by the selective adsorp-
tion of work function reducing materials is the sensitivity of the resulting surface to cathode
surface heating, ion bombardment, adsorption of residual gases, etc. As noted above, the
localizing effect of a monolayer, or nearly a monolayer, coverage is destroyed in poor
vacuum conditions, including of course exposure to the atmosphere. These characteristics
show that regeneration of the cathode is necessary at periodic intervals, a problematic issue
to their use in practical electron optical devices. If restoration of the cathode characteristics
is possible by straight forward techniques, one can consider this type of source for use in
commercial devices.
It has been observed'^^^, 415,416 ^j^^^ ^r may be diffused into the cathode by heating to
form a solid solution. In this case, degraded emission characteristics can be restored simply
by heating the cathode.
Special placement of an external source of Zr near the apex of the tip (see Fig. 7.13)
facilitates the impregnation of the tip's bulk and the transport of Zr atoms toward the emitter
tip by surface migration. Experiments with an electron source of the type shown in Fig. 7.13
were performed.^^^''^^^'"^^^ The FE cathodes could be stored in a closed container under
normal pressure for several months without degradation. Following subsequent installation
into the vacuum system, and heat treatment at 1300-1500 K, the original localized emission
current-voltage characteristics, the shape of the emission site, and the current density
distribution within the emission site were restored. Such results make the ZrAV system
promising for use in commercial electron devices.
7.4.5. StabiUzation by Oxygen Treatment
Swanson^^^' ^^^' ^^^ was thefirstto notice that localization is facilitated if Zr is deposited
on a W surface in the presence of oxygen. In fact, the lowest work functions, highest current
stability, and highest thermal stability (to ^^1800 K) are obtained by the adsorption of ZrO
molecules."^^^' "^^5' "^^^ The thermal stability is very important if the emitters are to be operated
in the Schottky emission regime.^^^' "^^8-410,420
Several other characteristics of the emission properties of the coadsorbed Zr/0 system
are beneficial: (1) The electron current density throughout the emission region is very
uniform, that is, it is possible to obtain an emission site without a dark area in the center;^^^
(2) The ZrOAV layers decrease the level of low-frequency noises in comparison to that
observed with ZrAV layers, a result that is particularly important for the thermal-field
emission regime; and (3) The stability of the average current is higher for longer time
periods. The decrease in low frequency noise is apparently the result of the formation of
a strongly bonded chemisorbed compound that decreases the mobility of the layer.
It is possible to confine the adsorption of Zr (ZrO) to one small {100} plane on the
emitter tip if the emitter tip build-up process is performed in a background pressure of oxygen
prior to deposition. The terraces surrounding the {100} plane are decreased in size and the
subsequent adsorption of Zr (ZrO) does not result in a decreased work function in these
'W<001>
Figure 7.13. Design of afieldcathode with the Zr source mounted at the base of the W tip.^^^'^^^
regions, presumably as a result of an average change in distance between surface atoms.^^^

The procedure for the formation of ZrOAV point cathodes has been described."^^^' ^^^ It has
also been determined that reproducible results can be obtained when oxygen impregnates
the whole volume of a tip near its apex. This requires an oxygen exposure of tens of hours
at a pressure of ^10~^ Torr.
Investigations'*^^' ^^^' ^^^ confirm these data and stable layers were formed in one to two
minutes with residual gas pressures of 10~^-10~^ Torr."*^^ The best operational parameters
for the cathode are obtained with this treatment: a brightness of 10^ A • str~^ • cm~^, a
current drift of 1 percent, and a noise level of 1 percent. The probe currents were <1 |xA.
The maximum probe current obtained in our experiments was 10 |xA at the total current of
several mA."*^^ The current drift at these currents was 2-3 percent with a noise level of 3
percent at 1800 K.
7.4.6. Shaping of the ZrfW Emission Sites in an Applied Electric Field
It has been shown that the direction of Zr migration on W depends on the sign of electric
field.^^' ^^^ With the emitter tip serving as the cathode Zr migrates from the apex toward
the emitter tip shank. With the field direction reversed, the Zr is drawn toward the tip's
apex. By using this reversed emitter tip polarity and heating the emitter tip it is possible
(c)
Figure 7.14. (a) Initial micrograph of a built-up W (100) tip; (b) FE localization at the surface of a built-up ZrAV
(100) tip; (c) scheme of microtip formation at W (100).
to decrease the beam divergence from 15-20° to 5-10° (see Fig. 7.14). The field strength
employed was chosen to be equal in magnitude to that required to draw an electron current
density of ~ 10^ A/cm^ in the FE mode of operation. The temperature at which Zr migration
is observed for the initial deposit is 1300-1500 K. The work function determined from the
relative slopes of FN plots for these layers (Fig. 7.14b) was 2.9 eV.
Recent experiments'*^^'^^^ confirm the possibility of obtaining such narrowly colH-
mated FE beams. A thorough analysis of the FE images indicates that tip build-up is
the process leading to such localization. A region of enhanced field strength emerges at
the apex with the orientation of the tip in the (001) direction, and a micro-ridge appears
during more advanced stages of build-up (see Fig. 7.14c). The area of the {100} facets
decreases, thereby leading to further emission confinement (Fig. 7.14b). The reduced size
of the terraces surrounding the (100) plane prevent the appearance of the emission zones of
concentric rings.
Direct confirmation of the important role played by the build-up in this localization
process is the emergence of a rebuilt W surface coming out from beneath the Zr layer
(Fig. 7.14c). The size of the emission site is of the order of ten's of angstroms. Thus, the
two emission confinement techniques can be combined for "double" localization: forming
a build-up surface and generating an ultrasmall emission site at (100) plane.
7.5. FIELD EMISSION FROM ATOMICALLY SHARP PROTUBERANCES
Electron holography, one of most promising methods for investigating nanostructures

and atomically sized objects, was proposed by Gabor in 1948."^^^ The key to practical
implementation of this idea is the development of an electron source having sufficient
coherence and brightness. FE cathodes offer new opportunities for the development of
electron holography due to their very high brightness (two orders of magnitude higher than
thermionic cathodes (see Table 7.1)^^^ and high coherence. Excellent interference patterns
using an FE cathode were obtained by Zeitler and Ohtsuki^^"^ on a gold film having 10~^ cm
diameter holes. Recently, FE microtips (Spindt-type Mo tips with a gate diameter aperture of
'^l mm)"*^^ were used for the observation of interference fringes with low-energy (140eV)
electrons.
pjj^j.80,82,83 suggested and investigated ultrasharp point emitters having a single atom
at the apex. The emitters were fabricated by field ion techniques, where in the course of
controlled field evaporation of a (111) oriented W tip (Fig. 7.15a). The W trimer is formed
on the (111) plane (see Fig. 7.15b). W atoms are deposited from the vapor phase until one
fills the threefold hollow site formed by the trimer (Fig. 7.15c). The procedure results in an
angular divergence in field ion mode of operation of less than 0.5°.
The formation of atomically sharp outgrowths was also demonstrated by Bihn and
Garcia by a build-up process called the "field-surface melting" technique.^^' ^^^ This process
involves the application of an electric field to a W {100) tip coated with gold. The tip's radius
of curvature was 70-100 nm. The emitter tip is heated by drawing very high current densities
in the pulsed mode of operation. If the surface is rough due to the presence of adatoms,
vacancies, kinks, etc., the action of the electric field and the surface temperature leads to
so-called "field-surface melting" and the formation of nanoprotrusions, that can terminate
in one atom. By rapid cooling of the nanoprotrusion in the applied field, one obtains highly
localized, and thereby coherent, electron emission. Measurements have shown that such a
source, formed by a protrusion of 20-30 nm in height and terminated by one atom provides
emission angles for ions and electrons of -^2° and 4° respectively. The electron beam
provides currents of <1 |xA and an ion beam currents of ^lO"^'* A.
Atomically sharp emission sites can be also formed by this build-up process without
overcoating the emitter with a low melting point material^"^^ as has been demonstrated by
the formation of atomically sharp protuberances and ridges on W, Mo, Ta, Nb surfaces.
7.6. APPLICATIONS OF FIELD EMISSION CATHODES IN

ELECTRON OPTICAL DEVICES
In conclusion we summarize some important parameters of FE cathodes obtained by

different investigators. Maximum brightness measured using built-up tips are of the order
of 10^^ A • sr~^ • cm~^ and the smallest energy spreads are ^0.2 eV using the cathodes
at low temperatures. Substantial interest exists in using ZrO coated W (100) cathodes in
132
Figure 7.15. Preparation of a single-atom tip of W [lll]-oriented single-crystal wires: (a) common field ion
image showing the (111) apex plane with a trimer as well as vicinal planes; point source tips with a significantly
high electric field confined to just (b) three atoms and (c) one atom.221
(U^ o o o o o
o^ o o o o o
*> ^ o
^
o o o o
<N - H T t T t
/—S 00 A A A A
—
'*
s
w fc
H
s-^
c?
H
tya 6
3 ON 0\ ON ON
3 I I I I I I I
o •2 0 0 0 0 0 0 0
(75
O
;5 '-S
U c0
o 0
s
c >
o 'S ^ o o
i3
Q Vi (D
tX)
5
<7 ^ <^l oo oo Î
<7 7
s ^
13 cd J ^ J ^ » 0 cs, erj « 0
d
1J
c '0
< >
*s
Ui
'a
(/i
c C/D
c
cd
h ^ 1
-o X5
S
c X 0
c^
^-^ s<
s
w
w
X
o
(/5 c o X X
o
a^
o
t3
CJ 88
cy5
O
^ o
o o o
VH '5
OD
o cj
I X 88
'E
G
s X
o
cs
X X
O
h
o 6 E
^
13
W)
g »< >S- ^>
"^ ^ ^ ^ -^
*- ^ *^ "S
C/)
"3
(^
o (A CO - ^ ^ w 00
C/5 (D
_o C^ > d
C/5
•c _
o O c« cj
0^ a. •
o c« CO a. ^ Q
t>
C3 VO (N fN <N <S
^
JZ
U D • 3 0 0
a^
r^ 0 •g iS 5
i>
0
<
^
0
i^
0
.5 £ £
zH 's0 l^S o
o (N CO u
r^
^-1 00 ON »0 S8
W 1/5 1/3 CO 00 00
Ssss sss
I U W W w w w
1-^ CT) C/3 0 0 H cr> on
(a) Field Emitter
(d)
Figure 7.16. Scanning electron microscopy with a FE cathode: (a) tip used in the Hitachi S-900 scanning electron
microscope; (b) scanning electron microscopy of the tip apex; (c) ball model of the tip apex; (d) high resolution
image of Au-Pd particles in a carbon substrate.
Figure 7.17. (a) Holograms taken with 80 eV electrons of carbon fibers with the electron point source 2400 nm
away from the carbon sample; (b) closeup of the central region with the source just 550 nm away from the
sample-emitting electrons at 65 eV (bottom); (c and d) results of the numerical reconstruction of the above
holograms.-^^^
the thermal-field emission regime. Maximum current densities for localized emission in
this case were roughly 10^ A/cm^ with minimum electron beam devergencies of ~5°. The
cathode service life in electron optical devices is several thousand hours. Maximum total
current in the beam probe is 10 |xA.
The first attempts at employing FE cathodes in electron microscopy were made by
Crewe et al.^^"^"^ who have succeeded in showing the potential advantages in the use of
such cathodes. They attained a 5 A resolution with a scanning electron microscope at 25 kV.
At present, the most significant applications of FE cathodes in electron microscopes have
been made by Hitachi (see in 423) and Amray firms. The basic operational parameters of
their microscopes are listed in Table 7.2. Figure 7.16 shows several important features and
results of applying FE to scanning electron microscope cathodes.
There have been only a few attempts at applying electron holography using FE.^^' ^^^
Interferograms obtained by Fink et al.^^^ on thin Au films are shown in Fig. 7.17. Electron
holography was demonstrated using a well-coUimated, high-energy electron beam produced
LEEPS setup
Point source Sample Detector
Figure 7.18. Setup for the holography experiments with low-energy electrons.^^^
by a high-resolution electron microscope.-^^-^'^^^ Fink proposed an electron microscopy

method to investigate objects only several micrometers from the electron source.82^ A mag-
nification of more than 10^ is expected. One remarkable feature of this form of electron
microscopy is that no lenses are required, thereby eliminating aberrations (Fig. 7.18).
7.7. RESUME
1. FE cathode is the most pointed, most bright, and most monochromatic of electron
emitter. Usual dimensions of a field electron emitter lie within tenths of a microm-
eter. With special methods, one can reduce the linear dimension of an emitting
spot down to 10^-10"^ cm, or even make the emitter atomically sharp in some
particular cases.
2. Tip FE cathodes are employed in the industrially manufactured transmission
and scanning electron microscopes of superhigh resolution, in Auger-electron
spectrometers, and in the electron holography.
3. A new approach to the development of electron optical systems is being developed
based on the vacuum microelectronics techniques, which permits forming of the
cathode and of electron optics elements in a single block.
8
ADVANCES IN APPLICATIONS
8.1. INTRODUCTION
Vacuum microelectronics (VME) is a new field in micro- and nanoelectronics that

has been developed during the recent few decades. VME is based on the employment of
electrons in vacuum at the dimensions of active elements in tenths and hundredths of a
micrometer. Practically, a field electron emitter (FEE) is used as an active element in VME
systems. By employing microscopic anode-to-cathode gaps, one can build devices with a
field emission (FE) cathode controlled by a voltage in tens of volts.
VME comprises of devices and active system components of micrometer size com-
mensurable with nanometer-scale dimensions. VME presents a sort of an alternative to the
solid-state electronics, employing the ballistic motion of electrons in vacuum instead of the
transport of current carriers (electrons and holes) in semiconductors.
Small dimensions of a FEE provide an opportunity of bringing the size of active ele-
ments down to a few tens of atomic dimensions (i.e., to 10~^-10~^ cm). Some up-to-date
experimental data indicate the existence of elective elements of a size on the atomic scale
(1-3 nm) on such special materials as, for example, carbon nanotubes.
The small dimensions of FEEs also permit achieving of a density of active elements of
up to 10^-10^^ cm~^, and even of up to 10^^ cm~^, if one employs such "self-organizing"
systems as fullerenes, nanotubes, etc.
The advantages of VME as compared to the solid-state electronics are as follows:
1. no dissipation of energy at electron transport in the medium, that is, in vacuum.

That gives the VME devices advantages for certain classes of devices such as those
generating high power at high frequencies;
2. the VME devices are acting with a high speed due to the virtual lack of inertia in
the FE process and a very small electron transport time in the vacuum gap, which
permits designing of fast-acting high-frequency devices;
3. a marked nonlinearity of the FE current-voltage characteristics, which permits
designing of frequency converters and multipliers;
4. no energy expenses on the act of electron emission to quantum origin of the
tunneling process;
5. a high radiation tolerance and heat resistance of the emitters.
137
Main VME applications

1. the most outstanding and widespread application is the production of flat low-
voltage displays of high brightness and high resolution based on field emission
arrays (FEAs);
2. employment of nano-FE cathodes in electron optic systems of superhigh resolution
(in scanning and transmission electron microscopes, electron lithography systems,
Auger electron spectrometers, and in tunnel microscopy);
3. production of active elements for integrated circuits (diodes, transistors, etc.);
4. employment in microwave devices;
5. devising of different types of pressure gauges, magnetic field gauges, etc.;
6. devising of a new type of ion source;
7. Realization of new very efficient production technologies for microelectronics
(lithography, etc.).
8.2. SHORT fflSTOMCAL REVIEW AND MAIN DEVELOPMENT STAGES
A fairly comprehensive history of the subject has been presented in the last few years
in reviews by Brodie and C. A. Spindt,^^ I. Brodie and P. R. Schwoebel,^^^ H. H. Busta,^^
R. Baptist,^^^ and H. F. Gray.i^'429
VME had began from the idea of a first vacuum transistor resistant to high levels of
radiation and high ambient temperatures and having, as does a semiconductor transistor,
a nonlinear current-voltage characteristic. In a vacuum microtransistor, in contrast to a
solid-state one, transport of the charge carriers takes place in vacuum and by a ballistic
method, in structures of micron and submicron dimensions.
The idea of a vacuum transistor had first been suggested by Shoulders in 1961 ^^^ in his
designs of vertical and horizontal microtriodes. Shoulders proposed a technology of active
emitter elements 0.1 |xm in size, applied FE microscopy for visualizing the emitter surface
microstructure and was one of the first to formulate a concept of microminiaturization of
vacuum electronics devices as a way to achieve such new features as high accuracy of
parameter reproduction, stability, low noise level, and others.
The next important phase in the growth of VME was the development of a new class
of cathodes, namely, the FEAs. An important achievement in the technology of FEAs was
integration of the electron collector (anode) with the FE matrix and the "honeycomb" control
electrode (gate) separated by a gap of a few microns, which made possible the addressing
of individual cathode elements and thus design amplifiers, frequency multipliers, current
converters, and elements of display panels. Because of the small separation between the
cathode and the gate these FEAs can be operated at low voltages (hundreds of volts). The first
report of a FEA with a multi-tip system with developed working area was made by Spindt
(1968,^^"^). The FEA proposed by Spindt (Spindt cathode) consisted of a large number of
individual molybdenum emitter tips arranged with a density of more than 10^ cm^ at micron
distances from one another. One more direction of the development of FEAs was the work
on increasing the operating area of cathodes and density of the elements in order to reduce
spatial fluctuations of the emission current.
Work on fabrication of FEAs of an area greater than 140 cm^ was done by Spindt and
Brodie with co-workers.^^"^^^
ADVANCES IN APPLICATIONS 139
Due to modem technological achievements in preparing thin films of submicron

dimensions and advances in electron beam lithography, emitter packing densities of up
to 10^-10^ cm^ were reached and matrix currents averaged over cathode surface raised to
over 1000 A/cm^.'^'*^ In 1989, Makhov^"^^ had demonstrated that FEAs of silicon pyramids
coated with a thin (30 A) film of dielectric (silicon dioxide) can be operated at voltages as
low as 10 V due to lower Schottky barrier at the semiconductor/dielectric boundary and low
electron affinity of the surface of the dielectric.
In 1970, Crost et al."*^^ proposed a technology of flat display panels based on
Spindt's FEAs.
VME received further impetus when FEAs were developed based on tip, blade, conical,
and pyramidal structures of heavily doped silicon. An advantage of amorphous silicon as
a material for FEAs is the highly developed technology of forming microstructures on the
surface of semiconductors relying on the vast experience of mass production of silicon
integrated circuits. The traditional methods of fabricating microcomponents of a silicon
FE cathode are anisotropic selective chemical etching, ion beam lithography, reactive ion
etching, and others. Success in obtaining stable and uniform emission from wide band
semiconductors made them competitive with metals owing to their superior chemical surface
stability and low operating voltages.
In 1986, Gray et al. built on the basis of silicon FEAs, the first efficient vacuum tran-
sistor or a miniature FE triode,^^ which in essence followed the concept of Shoulders.
This vacuum microtransistor has a number of advantages over solid-state semiconduc-
tor transistors: transport of electrons is not affected by properties of the semiconductor
making unnecessary the control of crystallinity and impurity content; due to its micrometer-
scale dimensions, the packing density of integrated circuit components can be more than
10^ per cm^; and because of small cathode-anode separation (a few microns), the electron
transit time is short making possible the generation of high-frequency currents. Fabrication
of the vacuum microtransistor, stimulated the development of vacuum integrated circuit
technology.
In the years to follow, field sources of electrons based on stable FEAs have been
intensively studied as candidates for application in flat display panels of high-brightness,
nanometer-scale electron beam lithography, electro-optical devices of high electronic
brightness and small spot size, microwave amplifiers and generators for frequencies from
100 to 100 GHz, and high-speed digital switching devices of the subpicosecond range.
Work on microwave vacuum electronic devices with high-current FE cathodes was
begun by Brodie and Spindt (1979,^^ 1976^^; Lally et al. 1986^^^). Spindt initiated work
on frequency generators for a range of hundreds of GHz based on FEAs of submicron
dimensions with periodic structure (Spindt, 1992"^^).
First color display of high resolution (3(X) pixels per in.) with individually addressed
FEAs has been demonstrated in 1987 (Holland et al."^^). Work on efficient conmiercial low-
voltage FE displays has been started at PixTech, Motorola, Micron Display Technologies,
FED Corporation, SIDT, Honeywell, Silicon Video, Raytheon, Futaba, and Samsung.
FE from ultrasmall sites at the tip cathode surface was utilized in designing vari-
ous surface diagnostics systems of superhigh resolution, in electron microscopy,'^^^'^^'^
and in atomic resolution electron holography.^^' "^^^ In these applications, advantage was
taken of high brightness and spatial coherence of the electron beam emitted by a point FE
cathode.
Discovery of the scanning tunneling microscopy^^^ stimulated the development of large

number of techniques for visualizing the surface topography with atomic resolution.
Microminiaturization of the electron optics made possible by the use of multielectrode
systems of cylindrical microelectrodes as electrostatic lens coaxial with individual tips in
FEAs paved the way for multiple-beam electron lithography of nanometer range."*^^ This
technique is remarkable for its high efficiency and output in producing matrices of silicon
integrated circuits. Efficient fabrication of individual nanostructures is achieved with a
probing "resist-free" lithography in a scanning electron microscope, where at a potential
difference between the probe and the substrate of several volts a certain composition of
the ambient modification of the substrate surface by the probe is produced as a result of
locaUzed chemical reaction (e.g., of oxidation).^^'^
In the works of Baptist, a possibility, in principle, of producing high-power ion beams
and building mass-spectrometers with the use of FEAs has been demonstrated. ^^^
In 1995, Denis Herve built a compact blue semiconductor laser excited by an electron
source based on an PEA (see Gray^^"*).
In the last decade the technologies of producing efficient FEAs has been actively devel-
oped. Using the method of optical interference lithography, FEAs of large operating area
have been produced having a cell area of 0.1-0.2 ixm.^^"^
Fabrication of large-area multi-tip FEAs with a control electrode having small grid
cells became possible with the use of chemical surface polishing"^^^ eliminating the need
for lithographic methods.
Molecular beam and gas phase epitaxy provided means of dorming nanometer tips
of silicon,"^"^ as well as heterostructures, for example, Si-Ge^^"^ and nanocrystals of
diamond"^^^ on the surface of silicon tips.
A considerable contribution to the development of FEA technology was made by
research groups at LETI and SRI.^^^
A new field of VME originated from the work on FEAs made of materials with low
electron affinity or highly nonuniform nanostructure, from which emission currents can be
obtained at relatively low electric fields (of the order of 10^ V/cm). Of the materials known to
possess such properties and likely to have practical applications the most intensively studied
are films and coats based on nanoclusters, namely diamond-like structures, fullerenes,
tubelenes (nanotubes), and their derivatives."^^' "^^^
8.3. FIELD ELECTRON MICROSCOPY
An important development in the FE investigations had been the invention by MuUer

in 1936 of the FE microscope—the projector.^^ Practically from this moment a systematic
accumulation of the data on surface properties of thefieldemitters started. This device made
it possible to identify many factors causing instability of the FE process, to study the types of
changes induced in the shape of afieldemitter by the electric field, temperature adsorption of
foreign atoms, electron, and ion bombardment. High magnification (10^-10^) and resolving
power (10-30 A) of the field projector together with the possibility of active action on the
object of study (in situ) made it an indispensable tool in the studies of adsorption, desorption,
epitaxy, surface diffusion, phase transitions, and so on.-^^' •^^' '^^' ^^
Fluorescent screen
(a)
Field emitter
(c)|
Figure 8.1. The principle of FE microscopy, (a) Setup; (b) field emission picture of w tip; (c) field ion picture of
w tip.
The FE microscopy is based on the fact that if a fluorescent screen—anode is placed

at a macroscopic distance (a few centimeters) across the path of electrons emitted from a
sharp tip (a few tenths of a micron), then the electron rays will produce on this screen a
projection of the tip apex with a very high magnification (Fig. 8.1a).
Because the electrons emitted from the tip surface scatter almost radially, the magni-
fication of such a microscope-projector is equal to the ratio of the distance between the
screen and the tip to the radius of the tip apex. Actually, the trajectories of electrons are not
quite radial, being affected by the base of the tip and the electrodes holding it; so the elec-
trons move along slightly sloping parabolas and the image on the screen appears somewhat
compressed. Taking this into consideration, the magnification M can be represented by a
simple formula
M= i.-, (8.1)
y r
where y is the compression coefficient (1.5 < y < 2); /? is the anode-cathode distance,
and r is the emitter apex radius. Because the tip apex dimensions are of the order of a
few tenths of hundredths of a micron and the distance R can be made (3-10) cm, the
magnification of such an instrument can be very great, of power up to 10^-10^. Muller^^
had applied this principle of field electron microscopy for the first time in a study of the
emission current density at the surface of emitter tip. He noted that the image obtained is
not uniform but consists of synmietrically arranged bright and dark spots. He understood
that the separate spots represented areas of the emitting surface, where the work functions
are different (see Fig. 8.1). He also realized that because the FE current density depends
exponentially on the work function (j ^ exp(—^^/^/F)), formula (1.8), the bright spots on
the screen corresponded to regions with small work function and the dark spots are areas
where the work function is large. He also found out what the tip apex is like.
The following picture can be imagined: put on a cubic crystal (body-centered, face-
centered, or hexagonal) is a hemispherical "hat," which cuts in the surface a smooth
hemispherical segment composed of sections of different crystallographic planes. It is
known from crystallography that different crystallographic planes have different arrange-
ments and packing densities of atoms. It is also known that the packing density of atoms
determines the work function. Closely packed planes have large work functions; a "loose,"
less closely packed ones have smaller work functions. So, at the surface of the fluorescent
screen due to different densities of the emission beams the component crystallographic
planes are displayed with the high magnification, 10^-10^, mentioned above.
The FE image of a tungsten (W) crystal tip is shown in the right-hand side of Fig. 8.1b.
Instead of a usual single grain it features two single-crystal grains. Comparison of the
emission image with X-ray diffraction data, as well as, comparison of the image parameters
with calculated crystallographic projections provided reliable identification of the planes
seen on the screen. In the emission image in Fig. 8.1b a closely packed {011} facet (in
crystallographic Miller indices*), facets of the {112} type, cubic {100} facet, and so forth.
Also, MuUer was the first to develop the atomic-scale microscopy using his discovery
of a field ionization effect, the so-called field ion microscope.^^^'^^^ An ionic image of
a W tip oriented in the same crystallographic direction is also shown in Fig. 8.1c. In this
image individual atoms can be seen as separate points. It was a great invention, the first in
human history, direct observation of the atomic structure of matter. Regrettably, because of
the limited size of this article, it is not possible to describe here the principle of field ion
microscopy and the outstanding achievements made using this remarkable instrument. This
is a subject to be considered separately.
The FE microscope possesses not only large magnification but also high resolution.
The resolution is understood as the possibility of separate observation of two close points
of an object. In the field electron microscopy the resolution is determined by the wave
nature of an electron, that is, in accordance with the de Broglie relation its wavelength
is A. = h/p, where p = mi; is the electron momentum. If an accelerating voltage U is

applied across the cathode-anode gap of a microscope, the electron wavelength, as can be
easily shown, will be A = 12.3/vÂ(l A = 10~^cm), if U is measured in volts. It is
found that because of diffraction the spreading at accelerating voltages of several kilovolts
usual in emission microscopy is about 8 A. Besides, the spreading occurs also because of
the tangential velocity component of the tunneling electron exiting the surface. The total
spreading in usual conditions of the FE microscopy is 20-30 A. According to estimates in
the resolution 5 of a FE microscope can be calculated by an approximate formula
5 ^ 2,62 • y{rlM{Y)(p^f^)^''^{k) (8.2)
where y is the compression coefficient; r is the emitter radius in E; A: is a coefficient equal
to approximately 5; ^{Y) is the Nordheim function; and cp is the work function in eV.
The emission microscope proved to be a remarkable instrument and made possible
studies of a number of important properties of surface in strong electric fields as well as
various fine effects on the surface.
First of all, with the emission microscope it has been revealed, what is going on at
the surface of the emitter proper, under different conditions: causes of instability in the
operation of field cathode established, its emission characteristics specified, local current
densities from different areas of the emitting surface measured, and others. With the FE
microscope, it is possible to identify, the crystalline structure of materials by comparing the
parameters of an emission pattern with calculated crystallographic projections.
With single crystal tips some phenomena can be studied in strong electric fiields,
10^-10^ V/cm, unattainable with macroscopic objects.
MuUer^^^ has discovered reconstruction of the tip surface occurring due to migration of
the surface atoms in emitters heated in a strong electric field. The fact that the reconstruction
is related to changes in the shape of crystal (its cut) has been demonstrated by Benjamin
and Jenkins"*^^ in brilliant works of Sokolskaya.^^^
It has been shown that the built-up phenomenon plays an exceptionally important role
in initiation of instability of the emission process due to formation on the emitting surface
of sharp "ridges" and small nanometer-size protuberances. Their formation causes local
enhancement of the electric field and a drastic increase of the emission current. An example
of the reconstruction with formation of nanometer protuberances on the tip apex is shown
inFig.8.2(see388,389)
As mentioned above, the resolution of a field microscope is 20-30 A and it is difficult
in principle to further improve it. Yet, with very small protuberances on the surface, the
resolution can be still higher."^^^ In the studies^^^'^"^^ ridges having a "sharpness" of just
2 or 3 atoms have been observed (Fig. 8.3).
The FE microscope allows one to study extremely slow displacements of the surface,
study in detail the surface migration, two-dimensional and three-dimensional evaporation,
and various chemical reactions, including reactions in a strong electric field.
Employing FE microscope, major quantitative characteristics of crystal surfaces can
be studied: work functions, activation energies for migration, desorption, and evaporation
of atoms, the size of two- and three-dimensional formations on the surface, and bonding
energies in two-dimensional crystals. The FE microscopy can be applied in studies of
high-temperature phase transitions and orientational relationships for transitions between
different crystalline modifications.
Figure 8.2. Nano-protrusions on the surface of FE cathodes.
With a microcrystal in the form of a tip, a unique method of obtaining atomically clean
surfaces can be implemented using desorption and evaporation of atoms by electric field^^^
under direct control of the process in the FE microscope. With this method atomically clean
surfaces of a number of semiconductor crystals have been obtained, such as Ge, Si, GaAs,
and others (see Chapter 5).
The above list does not exhaust possible applications of the field electron emission
microscopy. This technique is currently being actively developed and discoveries of new
effects and phenomena can be expected.
8.4. FIELD EMISSION DISPLAYS
One of the most promising applications of FEE at the present time is flat FE dis-
play panels of high-image quality, components for computing and information processing
systems featuring supershort response times (of the order of microseconds) and high radia-
tion and thermal resistance. By replacing thermal cathodes with FE cathodes consisting
of multi-emitter arrays considerably higher currents can be attained and the display
(a)
(b)
Figure 8.3. "Atomically sharp" edges formed by a strong electric field: (a) on the W surface; (b) on the Mo
surface.
Glass faceplate
(tin oxide coated)
Phosphor-coated anode
Vacuum
Dielectric
insulator
Base electrode
(emitter tip cathodes)
Rear vacuum
envelope plate f/^
Figure 8.4. The flat-panel display based on microfabricated field-emitter arrays.^^- "^^^ (a) Schematic diagram
of the display, (b) Photograph of the image of a feline on the black-and-white screen of the LETI flat-panel
display.
characteristics significantly improved, resulting in greater brightness and spatial resolu-

tion (density of pixels), larger viewing angle, and higher speed, as well as lower power
consumption.
Field emission displays (FEDs) are devices comprising an array cathode, an extracting
electrode, and a luminescent screen, where the image is formed (see Fig. 8.4). A con-
spicuous achievement of the past years, mainly connected with works by Spindt,^^"^ is the
fabrication of an ordered array of metal or semiconductor tip cathodes integrated with a
honeycomb anode (gates) and arranged in such a way that facing each tip is a correspond-
ing anode gate (see Fig. 8.5). The small distance between the emitter and the gate (a few
(a) Metal Si02 Dielectric

tips
0.75 |im
1 |Lim-
Figure 8.5. Spindt-type field-emitter cathode array.^^'*^^ (a) Schematic diagram, (b) Scanning electron micro-
scope micrograph of an array, (c) Scanning electron microscope micrograph of a single emitter in the array shown
above.
micrometers), as well as a small (tens of nanometers) tip apex radius solve several important
problems:
1. Fabrication of a low-voltage multi-emitter cathode. FE cathodes with a honey-

comb gate controlled with a voltage of about a hundred volts are now in operation,
with some studies indicating the possibility of reducing the voltage down to
ten volts."^^
2. Drastic reduction of the intensity of ion bombardment of the FE cathode due to
the novel principle of operation of the device, which increases the display lifetime,
makes the emitter operation more stable and eases the requirements to operational
vacuum; a vacuum of the order of 10~^ Torr was shown to be sufficient.
Due to smallness of the gap between the FE cathode and the controlling electrode (gate)
the signal frequency can be high, covering the range of TV signals, while the high speed of
FEDs compared to an ordinary cathode-ray tube ensures high instantaneous brightness of
every pixel and, consequently, high average luminance over the screen area.
The field-emitter-array cathode operates cold, needing no warm-up time; the modula-
tion voltages at the gate are in the same range as in conventional cathode-ray tubes (CRT)
guns; and the gate-to-cathode capacitance does not impose a limitation at television mod-
ulation frequencies (3-30 MHz). The major advantage, however, is that the much higher
currents can be obtained from a similar-sized cathode compared to conventional CRT's.
Higher currents are useful for the high-brightness, large-pixel-count displays required for
high-definition television (HDTV).
For the same frame rate, the dwell time of the beam on a pixel is smaller for HDTV
than for a conventional TV display, thus requiring a higher instantaneous brightness to
maintain a given average brightness for the whole frame. This brightness increase must
be attained by using higher beam currents, since X-ray generation makes it impractical to
increase beam voltages much above their present maximum levels (30 kV).
One of the most promising FED types are the devices with active matrix addressing,
in which every picture element (pixel) has a built-in circuitry facilitating addressing and
recording of information, so that the information capacity of a display can be increased for
the same or even better image perception characteristics. The idea of a planar addressable
matrix as a replacement for ordinary CRTs was first formulated by Crost."*^^
The basic component of such a display is a field electron multi-emitter cathode (array)
with gates (Fig. 8.5). The matrix of the multi-emitter cathode is usually fabricated on a
dielectric substrate such as glass. The entire cathode surface is divided into individual
elements—pixels. Figure 8.6 shows the schematic of an individual pixel forming a three-
color image; each pixel consisting of three parts, one for each color.
8.4.1. Requirements to the Tips and Lifetime of Matrix Field Emission Cathodes
A FEE usually has the form of an elongated tip, cone, pyramid, or blade, or it may
represent random nonuniformities shaped as sharp angles, ridges, and micro-protrusions,
which have close emissive capacities and emit onto the luminescent screen a fairly uniform
flow of electrons.
Faceplate
Horizontal
address
line
(Cathodes)
Field emission
cones
Dielectric
Gate substrate
Oxide
pillar electrodes
87 428
Figure 8.6. The principle of a flat-panel display's color pixel based on field-emitter arrays. '' °
Tip cathodes used in FEDs should keep their shape, withstand the pressure differential
on the casing walls, have conductivity of 10^-10^^ Ohm and be resistant to oxidation by
residual oxygen.
The size of an individual pixel in latest FEDs is usually 0.3 mm. At the present time
efforts are undertaken to reduce it down to 0.1 mm, which will increase the display res-
olution. Each pixel comprises more than 1500 tips distributed over the matrix surface at
a density of 10^-10^ cm~^, making the fluorescence practically uniform within the pixel.
The current density from one pixel in this case exceeds 50 ixA.'*^^
Above the tips, strips of the control electrode (three per each pixel, with gates) are
situated. A pixel is activated by applying a potential corresponding to the threshold value
for FE to the multi-emitter matrix and the required bias potential to the control electrode,
which sets the signal magnitude. The signal can be modulated by varying the pixel turn-on
time. A pixel can also be controlled by individual lines or rows. The electrons emitted
from every pixel are accelerated in the vacuum gap between anode and cathode (screen)
and impinge upon a luminescent screen. If the separation between the anode (screen) and
cathode is much larger than the size of pixel, the images of pixels can overlap reducing
the display resolution. To prevent this, in some designs a system of conducting parallel
electrodes comprising the bases of every separate pixel is introduced in the vacuum gap.
It should be noted that the state-of-the-art lithographic technology of FEAs does not
allow fabrication of FE cathode of large operating surface area. It is reported that the work
is underway on display panels of an operating area up to 15 cm^."^^^ According to reports
by LELI Company, panels of an area up to 1 m^ are being developed.
A relatively inexpensive alternative technology of multi-emitter matrices for dis-
play panels uses dielectric matrices with randomly distributed ultra-fine particles of
semiconductors or conductors, which represent point emission source.^^^
FEDs can be divided into several types differing in the operating voltage and the size
and form of emitters. FEDs of low-voltage type are operating at voltages in the range
100-800 V and high-voltage types at 1000 V and higher.
High-voltage FEDs have longer service life than low-voltage displays and give better
color rendition. If the operating voltage has to be increased to raise the phosphor emission
efficiency, the anode-cathode separation should be increased to prevent an occurrence of
discharge. Intrinsic divergence of the electron beam produced in field electron emission
increases the spot size on the screen necessitating focusing of the electron beam.
Therefore, in high-voltage FEDs, which have higher brightness than low-voltage ones,
the precise color separation poses problems related to switching between separate phosphor
strips within a pixel.
An advantage of the low-voltage FEDs is mainly their lower cost compared with high-
voltage FEDs.
Motorola has developed the best high-voltage FEDs and the Futaba company the
best low-voltage FEDs, and microdisplays 3 cm in size have been developed by the MDT
Corporation."^^^
LETI (France) has fabricated the first monochrome display.^^^"^^^ One of the models
announced by LETI"^^^ is a 6-in. black and white display, in which the screen luminance of
200 Cd/cm^ has been achieved. The accelerating voltage between the cathode and anode is
250 V and the voltage across the gap between the gates and the multiple-tip FE cathode is
80 V. The matrix of the FE cathode contained 256 x 256 pixels, each of an area 0.12 mm.
Power consumption is less than 1 W.
Companies such as LETI, Motorola, and Futaba successfully introduce this technol-
ogy into production of commercial full-color displays. At present the research by these
companies aims at increasing the area of the displays (up to 1 m^) and the efficiency of
phosphors.^^^
A low-voltage FED developed by Futaba (anode voltage of 800 V) has a 7-in. panel
with the attainable screen luminance of up to 300 Cd/cm^."^^^
Pix Tech Company has already produced a pilot batch of monochrome panels with a
screen diagonal of 7 and 12.1 in. Candenset Technologies Corporation and Sony plan for
2001 production of a full-color FED with a screen diagonal of 14 in.
At the IVMC (1999, 2000) conference it was reported that a number of firms began
production of the FE displays. In particular. Motorola will invest about a billion dollars in
production of FEDs.
8.4.2. Comparative Characteristics of the Display Types:

Advantages of Field Emission Displays
For displaying and processing graphic information color and monochrome CRT dis-
plays and liquid-crystal displays (LCD) are now predominantly used. In CRT displays,
electron beams in vacuum pass through a shadow mask producing visible pixels at the
spots, where under the electron impact a layer of phosphor starts to emit light of different
colors. The most promising area of the CRT technology is in HDTV television, which pro-
vides a resolution of more then 150 lines/mm and the pixel size less than 0.1 mm and makes
possible fabrication of large-format monitors with a screen diagonal of more than 20 in.
Standard CRT designs meeting requirements with respect to information content, resolu-
tion, and cost in some cases cannot be used because they are bulky (the monitor depth and
weight grow fast with the screen size), breakable, consume much power, and also degrade
the equipment characteristics. For example, CRTs are impossible to build into portable
computers, they have poor efficiency when used in technological and cosmic installations.
Close competitors of CRTs are LCD. In an LCD, a thin layer of sheet material transmits
or blocks, where appropriate, the backlight (passing color filters). The sheet material is
patterned into small cells, each cell corresponding to a pixel. The advantages of LCDs are
their small weight and depth, and extremely low power consumption. But these displays
are costly, rather slow, have poor contrast, require backlight, and noted for considerable
dependence of their performance on the ambient temperature. Particularly interesting is the
work on development of liquid-crystal full-color displays carried out by some Japanese and
American companies such as Sony, Pix Tech, Nokia, Motorola, and Tektronix.
The main problem with producing large-area LCD displays is making defect-free matri-
ces. The current state of the LC-indicators (production of matrices, LC-materials, packages,
activating integrated microcircuits) reached such a high level that the wide use of active
matrix addressing became possible. Because of successful development of indicators with
active matrix addressing for miniature pocket television sets, fabrication of large screens
for computers, automated work sites, and domestic electronic information systems became
feasible.
As for production of large displays, the most suitable for them appears to be the
technology of plasma panel displays, in which the phosphor is activated by UV radiation
from the plasma of electric discharge.
In the last five years, as noted above, flat-panel displays based on FEAs (multi-emitter
matrices) have found increasingly wider application in electronic devices.
The FED's principle of operation makes them considerably more compact than CRTs.
So, for the same screen area of, say, 12.5 cm, the FED has a thickness of 1.5 cm and a
weight of 285 g compared with CRT's 27 cm and 1450 g, respectively.
Advantages of the matrix addressing scheme for FED control compared with that of
CRT are abandonment of the cathode heating, low-efficiency deflection system, and the
shadow mask for confining the electron beam to a specified point on the phosphor. FEDs are
also practically instantaneous, afford large viewing angles, have ultra-short frame switching
time, low power consumption, high brightness and uniformity of screen luminance required
for displaying and processing of images, stable operation in a wide temperature range, and
compactness.
Image luminance and contrast provided by FEDs are practically independent of the
viewing angle within up to 160°. Highly efficient FED designs ( 3 ^ 0 ImAV) have a bright-
ness of up to 3500 Im that can be changed in increments as small as 0.05 Im. High speed
corresponding to a response time of 20 ns ensures good dynamic definition of the image.
Use of phosphorus as a fluorescent substance provides high quality of color separation,
wide spectrum of color shades and 256 hues of gray (8-bit gray scale). In particular, FEDs
can reproduce scores of the human body color hues, making them suitable for applications
in television for medicine. FEDs demonstrate stable operating characteristics in a wide
temperature range (from —40 to -|-85°C). Because of small distance between the FE tips
and the anode (0.2 mm) the size of the spot on the screen (or the pixel) is only about 0.33 mm.
Table 8.1. Comparative data on parameters of various types of

displays for 1999
Parameter CRT LCD FED
Brightness, Cd/m^ >500^ >50 >300

Contrast >100: 1 >12: 1 100: 1
Viewing angle >100^^ >80° >160°
Pixel size, mm 0.3^ 0.3 0.32
Image stability Fair Excellent Fair
Response time 50pLS > 200 ms 20 ns
Power consumption. W >200 1^ 1^
Maximum screen size (diagonal), in. 38 30 12.1
Notes:
^ Can be operated under direct sun illumination.
^ The value corresponds to HDCRT flat-panel display.
^ For EG Electronics 78FT II 795FT CRT-monitors, the pixel size is 0.24 mm;,
for HDCRT, it is less than 0.15 mm.
^ LCDs with backlight.
Drawbacks of FEDs include possible considerable fluctuations of the emission current

(30 percent on the average) even with absolutely stable applied voltage. Occasional current
surges can bring about destruction of the electrodes. Such occurrences can be prevented by
perfecting the FED design.
The potential of FEDs make them especially attractive for creating system for repro-
duction and processing of optical information in medical apparatus, measuring devices,
video system for the military, and so on.
Table 8.1 gives comparative data on parameters of various types of displays for 1999.
FED characteristics have been taken from advertising information of the Fix Tech Company
posted in the Internet.
8.5. OTHER APPLICATIONS OF FIELD EMISSION
Recent years have seen an explosive growth of VME as scores of its applications
emerged in totally unexpected areas and continue to expand. Traditional application area of
the FEAs is radio industry devices, such as vacuum integrated circuits for fast processing
of signals; miniature microwave tubes for generation of ultrahigh (higher than 1 GHz), and
high (below 0.5 GHz) frequencies; flat display panels for high-resolution television; and
high capacity fast memory cells for operation in a parallel mode at rates of 10 Gbit/s.^^^
VME devices find application in neighboring areas of high-current electronics, lithography,
surface diagnostics, electronic devices resistant to high temperatures, and radiation; and in
generation of high-power electron beams.
Mention should be made of the use of FE cathodes as sources of relativistic electrons
for cosmic research, as well as, in particle accelerators and nuclear reactors; in multiple-
beam electron sources for lithography; high-intensity electron sources in devices for surface
states diagnostics; and in free-electron lasers."^^^
Above (Section 8.4) FEDs have been considered and the use of field emitters as high-
intensity point sources of electrons (see Chapter 7).
Below we briefly present the major ideas, designs and techniques relating to latest
achievements in VME and some other applications.
8.5.1. Miniature Field Emission Triodes and Amplifiers Based on

Field Emission Arrays (FEAs)
First designs of the current-controlled electronic devices with needle-like FE cathodes

have been disclosed by Dyke.^^^' ^^ However, these devices did not find wide application
because needle cathodes could not operate for a long time under forevacuum conditions
and their internal resistance was considerable.
The possibility of producing miniature vertical and horizontal microtriodes (vacuum
transistors), as mentioned above (Section 8.1) was first predicted by Shoulders in 1961.^^^
First vacuum FE triode was reported by Gray in 1986.^^ Originally, it was 2i planar
(lateral) vacuum FE triode (transistor), in which a multi-emitter tip cathode was structurally
combined with a control electrode in the form of metal teeth. In the planar microtriode
electrons move straight upward and their flow converges at the anode-sink at the same
surface (Fig. 8.7).
In three-dimensional microtriodes^^^ of size less than 3 |xm, the most efficient collec-
tion of electrons at the anode is achieved on their trajectory upward from the tip (Fig. 8.8).
•înput
•ôutput
Input Output
Figure 8.7. A schematic of a planar FE microtriode:^^ 1: silicon substrate; 2: insulating layer; 3: silicon tip
emitter (source); 4: metal gate electrode; 5: metal anode electrode (sink); 6: trajectory of electrons.
Metal anode
Metal base electrode
Figure 8.8. A schematic of a vertical FE triode.^^^
The technology of planar microtriodes is less complicated compared with that of three-
dimensional devices; however, planar devices have a principle limitation on the use of metal
tip or blade emitters in contact with a layer of dielectric because the electrons can tunnel
from the Fermi level of the metal into the conduction band of dielectric at lower fields than
necessary for tunneling into vacuum.*
In a microtriode, by varying the potential of the control grid electrode (Fig. 8.9), the
anode current can be modulated.
Vacuum microtriodes have a number of advantages over solid-state devices:
1. higher power (more than 10 W);
2. resistance to ionizing radiation up to the level capable of destroying the anode
material;
3. stability in a wide temperature range (from liquid helium temperature to tempera-
tures in excess of 1000 K).
In order to increase the upper operating frequency of microtriodes from 1.3 GHz to the
centimeter range (30 GHz) it is necessary to reduce the interelectrode capacitances as well
as to increase the packing density and the transconductance of the static characteristics of
the triodes.
Steeper slopes of the I-V characteristics are achieved by increasing the density of tips,
use of blade emitters, and materials with low work function values.
The interelectrode capacitances can be lowered by reducing the cell size of the grid
electrode down to 1 |xm or smaller, reducing the permittivity of the layer of dielectric
between the gate and the base, reducing the distance between the gate and the base, and
using ultrasharp FE cathodes.
Modulation of the grid voltage in the FE cathode by an alternating signal produces
an electronic beam of well-modulated density, making it possible to excite in the electro-
dynamic system electromagnetic oscillations of a frequency equivalent to an input signal.
On this basis have been proposed and fabricated various modifications of amplifiers with
* We do not consider here the technologies of microtriodes based on achievements of the solid-state electronics,
which have been discussed in detail elsewhere.-^^^' ^ ^ ' ^ ^ ' ^^^
0.5-1 îm
0.3 |im ~1
0.7 |am
T
Figure 8.9. Structure of a three-dimensional silicon re microtriode:^^^ 1: anode; 2: grid; 3: intercomponent

insulation; 4: insulation layer of silicon oxide Si02; 5: silicon substrate; 6: silicon tip emitter.
modulation of the emission current by the signal.^^"^ In the most thoroughly studied devices
of this type, femitron, clistrode, the amplification coefficient is 10-12 dB with an efficiency
of up to 70 percent.
Wide amplification band, high efficiency, and compactness are the features of the
amplifiers with FE cathodes, which led to their wide use in cosmos-based, intersatellite and
mobile conmiunication systems, cosmic radars, local information networks, generators,
detectors for radioastronomy, and so on.
8.5.2. Microwave Devices for Millimeter Wave Amplification
Such, properties of the FE cathodes, as no need of heating, lack of inertia (the time
of flight of electrons being less than 10^^ s for a cathode-anode distance of 0.5 |xm), high
average current density (more than 10^ Aleve?- from an array surface^^), and nonlinearity
of the current-voltage characteristics make them especially suitable for use in microwave
devices. Narrow energy distribution curve of the FE electrons (0.22 eV at room temperature
and 0.14eV at liquid nitrogen temperature, see^^^) lowers the intrinsic noise of microwave
devices. Due to sharpness of the electron beams produced by arrays of emitter structures
their dimensions can be reduced, which also reduces the potential at the collector and
raises device efficiency. Highly intense and narrow electron beams obtainable from the
FEAs make possible size and weight reduction of microwave tubes. The introduction of
electron collector serves to increase efficiency by returning the emitted electrons back into
the device.
Microwave devices with FE cathodes can have a very wide range of operating frequen-
cies (from a few GHz to thousands of GHz), linear frequency characteristic, insignificant
variations of the output power and other parameters in the operating range, fast frequency
tuning, and low intrinsic noise levels.
Microwave devices are superminiature triodes of a size much less than the
wavelength.^^^'^^^The main element of the most electronic microwave devices for frequen-
cies up to 1000 GHz is a high-current thin-film FEA (Fig. 8.10) of submicron dimensions.
Such a cathode possesses an optimum combination of the basic device parameters: current
density, total current, pulse duration and repetition rate, and the stability of operation.
The transmission line for a low-power RF signal is formed by a FEA and a grid with
a pitch equal to the spacing of the FEA. The input RF signal modulates the electron beams
produced by every row of the matrix FEAs. The microwave device is designed (Fig. 8.10a)
as an FEA (3) with tip cathodes (4) and two microstrip lines having a common metal strip
with holes (1) and cylindrical microchannels (2), which plays the role of a grid and to which
a potential is applied causing the FE.^^^ To the upper metal strip (5) a potential is applied
capable of ensuring the passage of the electron beam through the microchannels. The entire
(b)
Output
UI
Top
view
oooo
input ^ oooo
strip line oooo
^
ô o o
k
/ \
Field >/ Match n
emission array / stub
(c) (d)
i l l i 25 mkm
i 5 mkm Input -^
=fe
' , ^'
12mkm
Figure 8.10. A schematic of microwave amplifier with the Spindt cathode:^^^' ^^ (a) construction of the device;
(b) resonant amplifier; (c) distributed amplifier; (d) schematic of tetrode amplifier.
system is evacuated. The height of the strip lines is from a few to tens of microns. In this
way, a system of microtriodes is obtained, which have grid and anode circuits coupled by
traveling electromagnetic waves propagating along the strip lines.
Two patterns of strip lines have been suggested: perpendicular (Fig. 8.10b) and parallel
(Fig. 8.10c, distributed amplifier).
For the resonant amplifier the following parameters have been obtained: frequency
—60 GHz; width of the operating frequency band 1 percent; power more than 20 W; gain
factor 10-20 dB; and efficiency >50 percent.^^^' ^^^
In a distributed amplifier the frequency band can be much wider but because of higher
wave resistance the efficiency is lower.^^^
By adjusting the area of FEAs, required cathode currents can be obtained from vacuum
electronic microwave devices. The cathode is usually made as a matrix of conical tips on
a silicon substrate. The cathode material is usually W or Mo;^^ besides, carbides of Zr or
Tl, as well as, heavily doped silicon can be used.^^^ Advances in modem microelectronics
technology made possible fabrication of FE cathodes with practically inertialess control of
the emission current at frequencies up to 1000 GHz.
Because of high nonlinearity of their I-V characteristics, with FE cathodes more effec-
tive modulation of the emission is possible than with hot cathodes, which is made use of
in diodes, clystrodes, and other devices.^^^ Due to this circumstance, in a diode with FE
apart from high FE current density an efficient generation at microwave frequencies can be
reached.
In a clystrode the integrated electron beam excites a microwave signal and is accelerated
by the voltage applied between the cathode and the grid, and then the kinetic energy of the
beam is transformed into the energy of harmonic microwave oscillations, when it passes
the resonator.
Structures prepared by methods of the VME can be used for creating analogs of many
traditional types of electrical vacuum microwave transit-time devices, for example, traveling
wave tubes (TWT), backward-wave tubes (BWT), clystrons, distributed amplifiers and so
on."^^^ These devices are used as heterodynes in radios, driving oscillators in amplifying
circuits, or self-contained generators in some types of radio equipment.
In a TWT a low-power RF signal is fed to the input and passes through a transmission
line formed by the FEA and the cathode grid. The cathode, isolating and anode grids have
a pitch equal to that of the FEA and are made of wires of different diameters; this and
corresponding potentials on the grids enables electron beams to be formed from isolated
tip cathodes. Dimensions of the TWT as well as the electrode potentials are matched to the
input signal frequency in such a way that the input signal passing through the transmission
line formed by the anode and anode grid is amplified by the modulated electron beam from
every row of the tips.
The principle of operation of the BWT is based on interaction of the electron beam
with a backward harmonic of the RF field wave. Prototypes of BWTs generate wide-band
oscillations in a range of 137-311.5 GHz as the control voltage is swept through a range of
0.5-5 kV with an average peak signal power of 100 |xW^^^
Wide-band amplifiers with distributed amplification on thin-film FE cathodes (Mo tip
cathodes for a supply voltage of 100 V) with modified control electrode and heat-resistant
anode in the form of transmission line can be used for fabrication of triode and tetrode
amplifiers with distributed interaction for a output power of more than 10 W in a frequency
range of 0-20 GHz. Such amplifiers are more compact than TWTs. Thin-film FE cathodes
for a current density of up 100 A/cm^ at accelerating voltages not exceeding 3 kV and
operating currents up to 20 mA represent components of a new class of ultrasmall TWTs
of the submillimeter range.
8.5.3. Nanolithography
High brightness (lOÂ/cm^ • sr) of the electron beam as well as the high density of
energy released on the target make the FE cathodes suitable for precision fabrication of the
elements of submicron structures by electron lithography, microengraving, and so forth.
The divergence, angle of the electron beam can be reduced using apertures cutting out the
most intensely emitting spot on the tip surface or using adsorption of foreign atoms onto
the FE cathode surface (e.g., Zr on W), which reduce the work function for electrons. Tips
with narrow solid angles can be formed using thermal-field treatment of the cathode (see
Chapter 7).
By now a large number of works have been published, in which a possibility of pro-
ducing nanostructures without using electron beam lithography is demonstrated. The most
thoroughly developed is the area of "resist-free" lithography in which under a potential
difference of a few volts between the probe and the substrate and a specific ambient near the
probe incidence the surface is modified because of a local chemical reaction (e.g., oxida-
tion). This method was successfully applied for forming nanostructures on a silicon surface
passivated with hydrogen,^^^ as well as on thin films of some metals, such as titan.^^^
Of special interest for FEAs fabrication is the technique of multiple-beam resist-free
electron lithography capable of producing tip emitters of smaller size and greater density. As
a source of electrons in the multiple-beam lithography matrix-controlled field emitter arrays
are used.^^^ The smallest hole in a film of electronic resist obtainable by this technique is
0.5 |xm in diameter. The hole size is limited by aberration effects in simple electronic lens
of the multiple-beam lithography units and electron scattering processes in the substrate.
8.5.4. Vacuum Magnetic Sensors
Magnetic field sensors (e.g., Vacuum Magnetic Sensors (VME)) are devices based
on the use of ultrasmall tip field emitters displaying high sensitivity, low noise level, and
considerable thermal stability. VMS detects the deviation of an electron beam incident
on the anode caused by the Lorentz force thus giving a measure of the magnetic field
induction. The first demonstration of VMSs and their potential for practical application was
done by Sugiyama.^^^ A VMS designs described in^^^ comprised a conical silicon emitter
tip positioned centrally in an orifice of a pulling gate electrode, four lens for focusing
the electron beam, and a two-section anode. The tip was made by reactive ion etch and
subsequent thermal oxidation of a silicon substrate with deposited layer of Si021 |xm in
diameter as a masking coat. The multilayer emitter structure was formed by subsequent
deposition on the substrate of a 0.5 |xm thick layer of dielectric (Si02) and a 0.2 |xm thick
film of niobium. The unit dimensions were 2 x 6 cm for an emitter-anode distance of 2 mm.
The emission threshold voltage applied to the gate was 60 V; at 80 V on the gate and 200
V applied to anode the emission current was 0.6 |xA. The tip was positioned exactly at the
center of the gap between two anode sections, ensuring, in the absence of magnetic field,
8=0 BÔ
/i-/2 = 0 /i-/2^0
^ 0
Figure 8.11. Principle of the magnetic sensor. 105
equal current flows in circuits closing the anode sections. Application of the magnetic field
in the plane of the multilayer emitter created a difference of currents in the circuits of the
two-section anode proportional to the field induction (Fig. 8.11). The sensor was able to
register magnetic field inductions of less than 10 mTl.
8.5.5. External pressure sensors
The main element of the sensor is the FE cathode positioned at a distance of about
1 ^Jlm from an elastic membrane, the collector. The pressure sensors may have diode or
triode configurations.^^^^^^ In both cases the external pressure deflects (bends) the collector
electrode causing variations of the current in the circuit corresponding to the magnitude of
pressure (Fig. 8.12).
8.5.6. Mass-spectrometers with field emission cathodes
Mass-spectrometers with electronic ionization are widely used as sensitive and fast
versatile detectors of molecular beams. Cryopumping is used to bring the residual gas
pressure down to 10~^^ Pa.
Use of hot cathodes in mass-spectrometers introduces an interfering effect of additional
fragmentation of the molecules in the sample analyzed as a result of heating and illumination
by the cathode. FE considerably simplifies the interpretation of mass-spectra because no
heating or illumination is produced by FE cathodes. In the FE ionizer molecules are ionized
as a result of losing or acquring electrons due to the applied field. Molecules ionized in an
area of high field (10^-10^ MV/cm) experience much less fragmentation than in ionization
by electronic impact. In Ortoff and Swanson,^^"* a durable water-etch resistant iridium FE
cathode was used as ionizer. Of special significance for the development of FE cathodes
for use in ionizers are graphite fibers 5-8 |xm in diameter or graphite filaments 1 mm in
diameter. Such FE cathodes are non-sensitive to ion bombardment and spark discharges
reducing the service life of metal FE cathodes.^^^
Collector
Vacuum seal
Figure 8.12. A schematic diagram of a triode-type pressure sensor.529
FEAs considerably increase the ionization volume and thus the sensitivity of the
method.
For mass-spectrometry of nonvolatile compounds a field desorption ionizer has been
proposed.^^^ In field desorption a large electric field gradient is used for inducing electron
tunneling from the sample surface, on which the substance studied is adsorbed. When a
sample molecule looses an electron, it leaves the positively charged surface going over to
the gas phase. The ions thus created enter the mass-spectrometer, are separated according
to their mass-to-charge ratio, and detected. Field desorption has a number of advantages
over other ways of producing ions, especially of polar and nonvolatile substances, namely
the molecules are not vaporized by heating but are available in the adsorbed form right
in the ionization zone. The field desorption ionization is less rough than electron impact
ionization.
8.5.7. Use of Field Emission in Gas Lasers
Electron beams of energy of hundreds of keV and current density 10^-10^ AJcvc? are
used for electron excitation of gas lasers. The cross-section shape and dimensions of the
electron beam should coincide with the working gas excitation chamber surface, on whose
area the laser power is dependent in direct proportion. The electron beam has rectangular
cross-section with uniform current density. The working gas excitation efficiency depends
on uniformity of the input beam with constant current density over the excitation area of
the gas laser. Usually, a blade FEC and a planar thin anode transparent for emitted electrons
are used.
8.6. ARRAYS MADE OF CARBON NANOCLUSTERS

One of the major problems of the modem VME is fabrication of high-efficiency, low-
voltage FEEs. Such field emitters should meet the following basic requirements: large
working area, efficient emission at low voltages (from a few to tens of volts) and fields
(1-10 V/|xm), high density of emission centers, and a controllable and reproducible fab-
rication procedure. The traditional method of fabricating field emitters is based on the
use of multi-needle FE cathodes and precision technological processes based on electron
lithography techniques ac.
As a cathode material, metals and semiconductors are usually used which, unfortu-
nately, have a rather high work function (4-5 eV). An alternative development path in VME
is a relatively inexpensive technology of planar FE cathodes based on materials with low
electron affinity or having high surface microroughness to ensure high local amplification
factor ^ for the electric field. These materials include films and coats of various allotropic
forms of carbon such as nanoclusters.^^' ^^^' ^^^
From analysis of the data accumulated so far, two main approaches to the development
of highly efficient field emitters based on carbon nanoclusters can beoutlined.^^^'^^^The
first approach relies on planar FE cathodes with coats of diamond-like films or thin carbon
films having high density of ordered emission centers. ^^^ In the other approach use is made
of a FE multi-needle cathode with a coat of carbon nanoclusters of a thickness from hundreds
of nanometers to tens of micrometers. The coat lowers the electronic work function and
the emission threshold field strength, and also reduces fluctuations of the emission current,
thus improving the emission characteristics.^^^' ^^
Carbon nanoclusters are self-organizing structures and open a principally new way of
fabricating FE cathodes with unique characteristics, such as very low average magnitude
of the threshold emission fields. In addition, carbon nanoclusters ensure high emission
stability, uniformity of the emission current density over the cathode surface and so on.
Considerable theoretical and practical interest in the ways of obtaining efficient and stable
field electron emission from carbon nanoclusters is reflected in the fact that in the past
two years a considerable number of publications on this subject appeared. However, the
advances in this field have not yet been presented in a systematic manner. A review of the
most recent achievements was presented in Sinitsyn et al."^^
Carbon materials of nanocluster structure having unique properties for use as FEEs can
be divided into the following two basic groups: (1) diamond-like films and (2) coats based
on fuUerenes and carbon nanotubes.
Major preparation methods for these materials are electric arc vaporization, magnetron
and laser sputtering of graphite, as well as ion-plasma deposition with the use of plasma
injected with a gas containing hydrocarbons. Electric arc vaporization of graphite should
be recognized as one of the most promising methods of producing carbon nanoclusters

because of its abiUty to yield various allotropic forms of carbon from diamond to graphite,
including intermediate forms: nanotubes and atomic clusters. It can also yield a variety of
surface morphologies from nanoscale nonuniformities (50-200 nm) to carbon fibers, and
coats of large area can be obtained.
Investigations of the emission properties of various carbon nanoclusters^^^' '^^^' '^^^' "^^^
confirmed that FE from these materials takes place in electric fields more than two orders
of magnitude lower than in metals and semiconductors. The current-voltage character-
istics investigated in a wide range of currents (4-5 orders of magnitude) give a positive
evidence that the mechanism of electron emission from these materials is linked to FE
(Fig. 8.13).
No clear interpretation exists at the moment of the extremely low threshold field for
the electron emission from carbon nanoclusters. A number of suggestions have been made
to explain this phenomenon.
(a) A
-13 h
-14
-15
-16
-17
0.6 0.7 0. 8 0.9 1.0 1.1
10^
Figure 8.13. Current-voltage characteristics for FE from fullerene structure and nanotubes:^^^' "^^^ (a) Current-
voltage characteristics; (b) f^ image of the cathode made from cathode deposit.
However, studies of the surface geometry of diamond-like films andfilmsof amorphous

carbon with the use of atomic force and tunneling microscopes ^62,468,516 ^-^ ^^^ reveal
any microstructures with a field enhancement factor ^ exceeding 5-10, so that the exper-
imentally observed lowering of the threshold field by more than two orders of magnitude
could not be accounted for.
2. To explain the FE mechanism in diamond-like films, a fundamental assumption has
been made, now shared by many researchers, that the electron affinity for the emitter surface
could be negative."^^^
However, this assumption raises some doubts because lowering of the threshold field
for initiating the emission has been observed not only for diamond-like carbon but for a
number of other carbon materials, such as films and coats of nanotubes and fullerenes as
well as for some wide bandgap semiconductors."*^^' ^^^' ^^'^' "^^^ Besides, this interpretation
is unable to explain the lack of enhanced emissive ability in macro-crystals of diamond and
lowering of the emission threshold with transition from microcrystalline to nanocrystalline
structure in diamond films.^^^
3. In studies,^^^' ^^^' ^^^ a new approach has been proposed to explain the low values
of electric fields, at which the FE from carbon nanoclusters is observed. This approach
is based on a hypothesis that in the near-surface region (0.2-1 nm), a negative charge is
accumulated in deep trap states within the bandgap and on defects as well as induced by
polarization of adsorbed dipole-type molecules of impurities. This charge causes drastic
increase of the electric field close to the surface of emitting centers. Calculations carried out
for diamond-likefilms"*^^have shown that at a concentration of the charges trapped by these
centers of about 10^^ cm~^, the local electric field strength near the surface of nanoclusters
can exceed 10^ V/cm even without an externally applied voltage.
In the framework of this model, the experimentally observed relationships receive
proper explanation. Becomes understandable the dependence of emission characteristics
on film thickness, the effect of hysteresis in current-voltage characteristics of FEEs"*^^' ^^^
(with increasing current charging of the deep trap states takes place and the field strength in
localized sites is changing). Ultimately, small values of the threshold voltage for emission
become explainable: the external voltage is required not so much for creating a field at the
surface as for transporting the electrons in the direction normal to the film.
4. Gulyaev and Sinitsyn"*"*^ suggested that the key role is played by structural defects
causing local lowering of the electronic work function.
Topologically, the defects can be produced by missing carbon atoms in the struc-
ture's network, introduced atoms of other elements, mutual interaction of the nanostructure
components, or specific geometry of the nanoclusters formed.
Quantum-chemical calculations of model structures based on isolated or interacting
fullerenes and tubelenes have shown that they have lower ionization potential and hence
lower electron work function than graphite.'*'*^
In Zakharchenko et al.,^^^ an analysis of the electronic spectra of single- and multip
wall carbon nanotubes with open and capped ends was carried out. The inside of the tubes
was doped with atoms of the elements of groups I-IV. It has been shown that the cylindrical
part of the tubes has metallic conductivity with the Fermi level at about 5.5 eV; doping of
the nanotubes with metal atoms (tin, lead) causes widening of the conduction band and
lowering of the barrier at the interface with vacuum down to 3 eV. Near the dome cap at
the top of a nanotube a forbidden band exists, which is characteristic of fullerene clusters.
A conclusion has been made that some walls in open-ended nanotubes are polarized by an
external electric field and the mobility has been estimated of the electrons tunneling through
the potential barrier at the surface of the walls arising due to a positive charge induced at
their surface. It was shown that the positive charge induced at the end of a nanotube can
significantly reduce the width of the potential barrier in the cylindrical part of the nanotubes
and thus considerably enhance the FE current. The induced charge can be significantly
increased due to ionization of atoms of different species, for example, oxygen, chemically
bonded to a carbon atom.
Measurements of the energy distribution function of the electrons emitted from carbon
nanotubes^^^ have shown that the electrons are emitted not from the wide quasi-continuous
band (as in metals) but from certain narrow energy levels having a half width of 0.3 eV
corresponding to states localized at the ends of nanotubes, possibly, related to changes in
the overlap of electronic jr-orbitals in the area of the spherical caps of nanotubes.
In Chen et al.^^^ a dependence of the emission current on orientation of nanotubes in an
externalfieldhas been established. It is shown thatfilmcoats of nanotubes aligned parallel to
the substrate have lower threshold field strength compared with coats of nanotubes oriented
at a normal. This result is an evidence that the dominant emission from the walls is due to
stacking faults of individual graphite layers, which can be observed in a scanning tunneling
microscope.
Experiments with doping of films of carbon nanoclusters with atoms of tin, lead, and
others^^^ showed that doping of a carbon nanotube with, for example, tin atoms may cause
severe deformation of its structure and, as a result, lowering of the ionization potential of
the nanotube by 2 eV and corresponding reduction of the work function for electrons.
In studies by Rakhimov and co-workers,^^^ FE images of diamond-like films were
obtained and dependence of the emission current on the electric field strength investigated.
It has been found that the emission current originated at isolated areas on the sample surface.
High density of emission centers (10^ cm""^) and high surface uniformity of the emission
current were observed. It was demonstrated that the centers of efficient emission consisted
of a large number of spots (of a size 50-100 nm) located in the intercrystalline boundary
areas. An emission mechanism has been proposed based onfieldenhancement at conducting
intercrystalline "ribs" and supported by data of scanning electron microscopy indicating
that the current from intercrystalline areas (of higher conductivity) is much larger than
from apexes of nanocrystalline grains. It is interesting that in the limiting case of a mainly
amorphous diamond-like film, high emissive properties were not observed. Rakhimov's
data for the diamond-like films also confirm the hypothesis of the role of defect sites,
exemplified in this case by the boundaries of nanocrystalline grains.
One of the most important tasks of our investigations was a quantitative study of
the relationship between emission characteristics and kinetics of the FE process, on the
one hand, and the topology of emitting surfaces based on fuUerenes and other carbon
nanoclusters, on the other.^^^' ^^^^ ^^^^ ^^2' ^^^
High-resolution atomic force microscopy was used in Baskin et al., (1999), Dyuzhev
et al. (1998, 1999)^^^' ^^^' ^^^ to determine dimensions of carbon clusters and the curvature
radii of micro-irregularities at the surface of diamond-like films and fullerene coats.
Shown in Fig. 8.14 is a computer reconstruction of the surface element of a diamond-
like film obtained by means of scanning atomic force microscopy. Characteristic heights
Height of tips: 3 5 - 5 5 nm
Surface Physics and Electronics
Research Centre (SPERC) of
University of Telecommunication
Figure 8.14. Atomic force microscopy image of diamond-like film."^^-^' ^^^
of the pyramidal nanocrystals are 30-50 nm and the apex radii of the nanoclusters in some
cases are as small as 5 nm. The density of micro-protrusions of small dimension in some
spots at the surface amounts to 10^ cm~^.
Presented in Fig. 8.15 is a micrograph of adsorbed close-packed spatially ordered
aggregates of Ceo molecules of a size about 20 A obtained by vacuum evaporation onto
a Mo substrate."*^^
The obtained FE images of diamond-like and fuUerene films reveal the presence of
separate small-area centers with a high current density, up to 10^ A/cm^.
Most specimens had current-voltage characteristics obeying the Fowler-Nordheim
(FN) law.
Carbon nanotubes is a potentially promising material for vacuum nanoelectronics. Such
features of the nanotubes as their small dimensions (with curvature radii from 1 to 10 nm),
electrical conductivity (100 Q~^ cm~^), mechanical strength, and chemical stability make
nanotubes a candidate material for future microelectronic elements."^^^
I
100 A I
Figure 8.15. Transmission electron microscopy image of surface with fullerene structures. ^-' ^"'
Gulyaev et SLI.^"^^ have developed a technique for producing films and coats consisting
mainly of single- and multiple-wall carbon nanotubes. The FE characteristics of carbon
nanotubes are close to those of diamond-like films. According to Sinitsyn et al.,^"^^ electronic
work function for films of single-wall nanotubes found from the slope of current-voltage
characteristic in the FN coordinates is less than 0.5 eV. The FE current density averaged
over the cathode surface reached 3 A/cm^.
A special variety of the carbon materials is the so-called "cathode deposit" forming
during arc discharge vaporization of graphite directly on the cathode."^^^' ^^2-534 j^. j^jgh^
prove a very suitable material for the fabrication of FEAs.
Detailed electron microscopy studies including scanning electron microscopy, trans-
mission electron microscopy, and high-resolution tunneling microscopy have shown that
the cathode deposit contains closely packed nanotubes, which in the growth process aligned
along a preferred direction related to the direction of electric field (Fig. 8.16).
Figure 8.16. Scanning electron microscopy nanotubes micrograph.'*^^' ^^^' ^^^
Distribution of the carbon nanotubes in the cathode deposit has been studied and a
method developed for fabricating FE cathode arrays from the deposit material.
Direct observations of the emitting surface of the deposit in a FE microscope have
revealed high density and uniformity of emission centers and low operating voltages. This
opens the prospect of developing a technology of fabricating large-area FE cathodes taking
advantage of self-organization of carbon clusters directly in the growth process. In this case,
there is no need, in a special high-precision technology, for creating a regular pattern of
nanotubes on a substrate of different material. FE images of a multi-emitter cathode made
of the deposit are shown in Fig. 8.17.^^^
The studies of carbon nanoclusters are now in their initial stage, with the number
of publication growing exponentially with time. The fact that a number of new devices
for VME based on carbon materials have already been demonstrated can be considered a
considerable technological achievement.
One of these entirely new types of devices are ecologically safe flat display panels with
high brightness and resolution.'^'*^
FE cathodes based on diamond-like films and carbon nanotubes made possible fabri-
cation of flat monolithic display panels characterized with high operating stability and low
emission current fluctuations (not exceeding 10 percent).
Figure 8.17. FE image for nanotubes.^"^^' ^^^' ^^^
A display with an emitter based on diamond-like carbon film described in Kim et al.^^^
possesses a resolution of 60 lines per in. at an average emission current of 500 mA and
operating voltage of 800 V. The diagonal of this monochrome panel is 2 in. and the screen
brightness is lOOcd/m^.
A monochrome display based on carbon nanotubes reported in Choi et al."^^^ has an
effective screen area of 130 cm^ with resolution of 30 lines per in.; its brightness can reach
450 cd/m^ at an operating voltage of 800 V and average current of 1.5 mA.
Fabrication with the use of nanotubes of a full-color display with a diagonal of 9 in.
was reported in Choi et al.^^"* The FE cathode is made by rubbing in the prepared matrix of a
paste based on ground nanotubes. The pixel size is 200 x 20 |xm. At an electric field strength
of 3 V/|xm, the image brightness is 800,200, and 150 cd/m^ for green, red, and blue lights,
respectively. The high brightness is achieved at a grid-cathode voltage of 100-300 V and
anode voltage of 500-1000 V. The authors mention good reproducibility of the fabrication
process.
8.7. RESUME
1. VME as a principally new field of electronics has taken shape.

2. Outstanding results have been achieved in developing FEAs for planar low-voltage
FEDs, amplifiers, microwave devices, and so on.
3. It has been shown that there are no limitations for TV applications of multi-emitter
FEAs at frequencies up to 330 MHz.
4. No principle limitations exist for the operation of high-frequency devices up to
tens of GHz, which will have been fabricated in the near future.
5. Point FE cathodes have already found application in optoelectronic devices, in
microscopy, spectroscopy, and holography.
6. VME have opened a possibility of integrating in one unit a point electron source
and controlling electrodes for electron microscopy, mass-spectroscopy. Auger
spectroscopy, and other applications (Baptist, Gray).
In 1996, Henry Gray had excellently portrayed the future of VME. His prophecies are
coming true almost word for word now. It seems to me appropriate, as a homage to his
memory, to quote his own words by way of a resume.
"Leading industrial manufacturers all over the world have already started commercial
production of some types of vacuum microelectronic devices. Investments into this field are
expected to amount to 10 billion dollars during the coming five to ten years. For instance,
the Motorola company alone invested about 900 million dollars into the production of field
emission displays in 1999."
REFERENCES
1. R. W. Wood, A new form of cathode discharge and the production of X-rays, together with some notes of
diffraction, Phys. Rev. 5(1), 1-10 (1897).
2. W. Schottky, Uberkalte und warme elektronenentladungen, Z Physik 14, 63-106 (1923).
3. R. A. Millikan and C. F. Eyring, Laws governing the pulling of electrons out of metals by intense electrical
fields, Phys. Rev. 27(1), 51-67 (1926).
4. B. S. Gossling, The emission of electrons under the influence of intense electric fields, Phil. Mag. 1(3),
609-635 (1926).
5. R. A. Millikan and C. C. Lauritsen, Dependence of electron emission from metals upon field strengths and
temperatures, Phys. Rev. 33(4), 598-604 (1929).
6. R. H. Fowler and L. W. Nordheim, Electron emission in intense electric fields, Proc. Royal Soc. (London)
A119(A781), 173-181 (1928).
7. L. W. Nordheim, The effect of the image force on the emission and reflection of electrons by metals, Proc.
Royal Soc. (London) A121(A788), 626-639 (1928).
8. J. E. Henderson and R. I. Badgley, The work required to remove a field electron, Phys. Rev. 38(3), 590
(1931).
9. J. E. Henderson, R. H. Dahlstorm, and F. R. Abott, The energy distribution of electrons in field emission,
Phys. Rev. 41(3), 261 (1932).
10. J. E. Henderson and R. H. Dahlstorm, The energy distribution in field emission, Phys. Rev. 55(5), 473-481
(1939).
11. G. M. Fleming and J. E. Henderson, A search for temperature changes accompanying field emission at high
temperatures, Phys. Rev 54(3), 241 (1938).
12. G. M. Fleming and J. E. Henderson, The energy losses attending field current and thermionic emission of
electrons from metals, Phys. Rev. 58(10), 887-894 (1940).
13. E. W. Miiller, Zur Geschwindigkeitsverteiling der Elektronen bei der Feldemission, Z Physik 120(5/6),
261-269 (1943).
14. E. W. Miiller, Feldemission, Ergebn. d. exakt. Naturwiss 27, 290-360 (1953).
15. R. D. Young, Theoretical total-energy distribution of field emitted electrons, Phys. Rev. 113(1), 110-114
(1959).
16. R. D. Young and E. W. MuUer, Experimental measurement of the total energy distribution of field emitted
electrons, Phys. Rev 113(1), 115-120 (1959).
17. A. G. J. van Oostrum, Validity of the Fowler-Nordheim model for field electron emission. Philips Res. Rep.
Suppl. 11, 1-102 (1966).
18. L. W. Swanson and L. C. Grouser, Total energy distribution of field-emitted electrons and single plane work
functions for tungsten, Phys. Rev 163(3), 622-641 (1967).
19. L. W. Swanson and A. E. Bell, Recent advances in field electron microscopy of metals. Adv. Elect. Electron
Phys. 32, 193-309 (1973).
20. E. W. Miiller, Versuche zur Theorie der Elektronenemission unter der Einwirkung hoher Feldstarken,
Z Techn. Physik 31(22/23), 838-842 (1936).
171
172 REFERENCES
21. E. W. Mtiller, Elektronenmikroskopische Beobachtungen von Feldkathoden, Z Techn. Physik 106(9/10),

541-550 (1937).
22. G. Ehrlich, An atomic view of adsorption, Brit. J. Appl. Phys. 15(4), 349-364 (1964).
23. R. Gomer, Field Emission and Field Ionization. (Harvard University Press, Cambridge, Massachusetts,
1961), 195 p. 2 ed (AIP, New York, 1993).
24. I. L. Sokol'skaya, Field-emission microscopy application for surface diffusion and self-diffusion
investigations, in: Surface Diffusion and Flowing, edited by Ya. E. Geguzin (Nauka, Moscow, 1969),
pp.108-148.
25. M. I. Elinson and G. F. Vasil'ev, Field Emission. (Fizmatgiz, Moscow, 1958), p. 272.
26. G. N. Shuppe and N. S. Zakirov, Electron Emission from Metal Crystals. (Izdat. SAGU, Tashkent, 1959).
27. W. P. Dyke and W. W. Dolan, Field emission, in: Advances in Electronics and Electron Physics, Vol. 8,
edited by L. Marton (Academic Press, New York, 1956), pp. 89-185.
28. G. N. Fursey, V. N. Shrednik, and G. A. Mesyats, High voltage cold cathodes, in: Cold Cathodes, edited by
M. I. Elinson (Sovetskoe Radio, Moscow, 1974), pp. 165-304.
29. A. Modinos, Field, Thermionic, and Secondary Electron Emission Spectroscopy (Plenum, New York, 1984),
375 p.
30. R. Haefer, Experimentelle Untersuchungen zur Priifung der wellenmechanischen Theorie der Feldelec-
tronenemission, Z. Phys. 116(9/10), 604-609 (1940).
31. W. P. Dyke and J. K. Trolan, Field emission: large current densites, space charge and vacuum arc, Phys. Rev.
89(4), 799-808 (1953).
32. W. P. Dyke, J. K. Trolan, E. E. Martin, and J. P. Barbour, The field emission initiated vacuum arc. I.
Experiments on arc initiation, Phys. Rev 91(5), 1043-1054 (1953).
33. W. W. Dolan, W. P. Dyke, and J. K. Trolan, The field emission initiated vacuum arc. II. The resistively heated
emitter, Phys. Rev. 91(5), 1054-1057 (1953).
34. J. P. Barbour, W. W. Dolan, J. K. Trolan, E. E. Martin, and W. P. Dyke, Space-charge effects in field emission,
Phys. Rev. 92(1), 45-51 (1953).
35. E. E. Martin, J. K. Trolan, and W. P. Dyke, Stable, high density field emission cold cathode, J. Appl. Phys.
31(5), 782-789 (1960).
36. W. P Dyke, F. M. Charbonnier, R. W. Strayer, R. L. Floyd, J. P Barbour, and J. K. Trolan, Electrical stability
and life of the heated field emission cathode, J. Appl. Phys. 31(5), 790-805 (1960).
37. T. J. Lewis, Theoretical interpretation of field emission experiments, Phys. Rev. 101(6), 1694-1698 (1956).
38. P. H. Cutler and D. Nagy. The use of a new potential barrier model in the Fowler-Nordheim theory of field
emission. Surf. Sci. 3(1), 71-94 (1965).
39. W. P Dyke, Advances in field emission, Sci. Amer 210(1), 108-118 (1964).
40. M. I. Elinson, V. A. Gor'kov, and G. F. Vasil'ev, Field emission from rhenium, Radiotekhnica i Elektronika
3(3), 307-312 (1958).
41. M. I. Elinson, Influence of gas adsorption at the surface of emitter uponfield-electronemission, Radiotekhnica
i Elektronika 3(3), 438-+39 (1958).
42. M. I. Elinson and G. F. Vasil'ev, Investigation of thefieldemission from lanthanum hexaboride, Radiotekhnika
i Electronika 3(7), 945-953 (1958).
43. M. I. Elinson and G. A. Kudinzewa, Field emission cathodes of high-melting metal compounds, Radio Eng.
Electron. Phys. 1, 1417-1423 (1962).
44. M. I. Elinson and G. A. Kudintseva, Field-electron cathodes of metallic high-melting compounds,
Radiotekhnika i Electronika 1 {9), 1511-1518 (1962).
45. I. I. Gofman, O. D. Protopopov, and G. N. Shuppe, Pulse technique in the study of field emission from
tungsten emitter, Izvestiya Akad. Nauk UzSSR, Fizika- Matematika 6, 72-77 (1959).
46. 1.1. Gofman and G. N. Shuppe, Pulse technique in the study of field emission from tungsten single crystels,
Radiotekhnika i Elekronika 4, 1215-1216 (1959).
47. G. N. Fursey, Field emission from tungsten monocrystal preceding the vacuum arc development,
Radiotekhnika i Elekronika 6(2), 298-302 (1961).
48. I. L. Sokol'skaya and G. N. Fursey, Investigation of the phenomena preceding the tungsten emitter destruction
by pulses of high-density field-electron emission current. Radiotekhnika i Elekronika 7(9), 1474-1483
(1962).
49. G. N. Fursey and G. K. Kartsev, Stabihty of field emission and migration processes preceding development
of a vacuum arc, Zh. Tech. Fiz. 40(2), 310-319 (1970).
REFERENCES 173
50. I. L. Sokol'skaya and G. N. Fursey, The influence of different coatings upon the character of phenomena
preceding the distinction of tungsten emitter by pulses of high density field-electron emission current,
Radiotekhnika i Elekronika 7(9), 1484-1494 (1962).
51. G. N. Fursey, V. E. Ptitsyn, and D. N. Krotevich, Spontaneous migration of the surface atoms at maximum
current densities of the field-electron emission initiating vacuum breakdown, Proc. 11th ISDETV, Berlin,
GDR. (September 24-28, 1984), 1, 69-71.
52. D. N. Krotevich, V. E. Ptitsyn, and G. N. Fursey, Spontaneous build-up of a field-emission cathode at
maximum current densities, Zh. Tech. Fiz- 55(3), 625-627 (1985).
53. G. N. Fursey, G. K. Kartsev, G. A. Mesyats, D. I. Proskurovskii, and V. P. Rotshtein, Field emission initiated
vacuum arc at extremely high field at high current densities. Proc. 9th Int. Conf. Phenomena Ionized Gases,
Bucharest, (1969), p. 88.
54. G. K. Kartsev, G. A. Mesyats, D. I. Proskurovskii, V. P. Rotshtein, and G. N. Fursey, Investigation of the time
characteristics of the transition of field emission to vacuum arc, Dokl. Akad. Nauk USSR 192(2), 309-312
(1970).
55. G. N. Fursey and S. A. Shakirova, Localization of field emission in small solid angles, Zh. Tech. Fiz. 36(6),
1125-1131(1966).
56. V. G. Pavlov, A. A. Rabinovich, and V. N. Shrednik, High local current densities of field emission on steady
state process, Zh. Tech. Fiz. 45(10), 2126-2134 (1975).
57. G. N. Fursey, L. M. Baskin, D. V. Glazanov, A. O. Yevgen'ev, A. V. Kotcheryzhenkov, and S. A. Polezaev,
The specific features of field emission from submicron cathode surface areas at high current densities. 7th
Int. Conf. Vacuum Microelectronics, Portland, Oregon (July 30-August 3,1995). Technical Digest (Electron
Device Society), pp. 504-508, J. Vac. Sci. Technol. B 16(1), 232-238 (1998).
58. M. I. Ehnson, V. A. Gor'kov, and G. F Vasil'ev, Study of a method to reduce bombardment of field-electron
cathode by ions of residual gases, Radiotekhnika i Electronika 2(2), 204-218 (1957).
59. V. A. Gor'kov, M. I. Elinson, and V. B. Sandomirskii, On the role of space charge in initiating of high-density
field-electron emission current, Radiotekhnika i Electronika 7(9), 1495-1500 (1962).
60. L. M. Baskin, L. L. Anan'ev, D. A. Borisov, A. A. Kantonistov, and G. N. Fursey, Ion bombardment
suppression effect of a field emission cathode, Radiotekhnika i Electronika 28(12), 2462-2464 (1983).
61. G. N. Fursey, Field emission in a microwave field, J. Vac. Sci. Technol. B. 13(2), 558-565 (1995).
62. F. M. Charbonnier, J. P. Barbour, J. L. Brewster, W. P. Dyke, and F. J. Grundhauser, Intense nanosecond
electron beams, Trans. IEEENS'U{3\ 789-793 (1967).
63. W. P. Dyke and F. M. Charbonnier, Advanced electron tube techniques. Proceeding of the VI National
Conference (1963).
64. M. I. Ehnson, V. A. Gor'kov, A. A. Yasnopol'skaya, G. A. Kudintseva, Study of pulse field emission at high
current densities, Radiotekhnika i Elektronika 5(8), 1318-1326 (1960).
65. G. N. Fursey, Power electron sources using field emission, in: Cold Cathodes, edited by M. I. Elinson
(Sovetskoe Radio, Moscow, 1974), pp. 241-269.
66. G. N. Fursey and G. A. Mesyats, Explosive field emission at the initial stages of vacuum discharges, in: Cold
Cathodes, edited by M. I. Elinson (Sovetskoe Radio, Moscow, 1974), pp. 269-303.
67. G. N. Fursey and P. N. Vorontsov-Vel'yamnov, Qualitative model of vacuum arc initiation, Zh. Tech. Fiz.
37(10), 1870-1888 (1967).
68. S. P. Bugaev, A. M. Iskol'ski, G. A. Mesyats, and D. I. Proskurovskii, Electron-optical observation of
initiation and development of pulse breakdown of short vacuum gap, Zh. Tech. Fiz. 37(12), 2206-2208
(1967).
69. E. A. Litvinov, G. A. Mesyats, and D. I. Proskurovskii, Field emission and explosive electron emission
processes in vacuum discharges, Uspekhi Fiz. Nauk 139(2), 265-302 (1983).
70. G. N. Fursey, Field emission and vacuum breakdown, IEEE Trans. Electron Insulation EI-20, 659-670
(1985).
71. A. V. Crewe, D. N. Eggenberger, J. Wall, and L. M. Welter, Electron gun using a field emission source.
Rev. Sci. Instrum. 39(4), 576-583 (1968).
72. A. V. Crewe, J. Wall, and L. M. Welter, High-resolution scanning transmission electron microscope, J. Appl.
Phys. 39(13), 5861-5868 (1968).
73. A. V. Crewe and J. Wall, High-resolution capabilities in scanning microscopy. Proc. 7th Int. Congr Electron
Microsc, Grenoble, France, 1970. Paris, Franscaise de Microscope Electroniques, 1970, pp. 1-2.
74. A. V. Crewe, J. Wall, and J. Langmore, Visibility of single atoms. Science 168(3937), 1338-1340 (1970).
174 REFERENCES
75. P. Morin and R Simondet, An electrostatic-field emission microscope for spectroscopic Auger-electron
analysis, J. Physique Colloq. 45, 307-308 (1984).
76. B. Reihl and J. K. Gimzewski, Field-emission scanning Auger microscope (FESAM), Surf. Sci. 189, 36-43
(1987).
77. I. P. Zhizhin, G. N. Fursey, S. A. Shakirova, and A. V. Shishatskiy, For a problem on the getting of electron tips
by field emission and magnetic optics, 24th AU-Union Conference on Field emission electronics. Tashkent
(1970), section 4, p. 34. Izvestiya Akad. Nauk USSR 35(2), 302-306 (1971) (in Russian).
78. N. Samoto, N. Tamura, R. Shimizu, S. Namba, H. Hashimoto, and K. Gamo, A stable high-brightness
electron-gun with Zr/W-tip for nanometer lithography. 1. Emission properties in Schottky-emission and
thermal field-emission regions, Jpn J. Appl Phys. Part 1-Regular Papers and Short Notes 24(6), 166-11 \
(1985).
79. R. W. Devenish, D. J. Eaglesham, C. J. Humphreys, and D. M. Maher, Nanolithography using field-emission
and conventional thermionic electron sources, Ultramicroscopy 28, 324-329 (1989).
80. H. W. Fink, Point source for ions and electrons, Physica Scripta 38, 260-263 (1988).
81. A. Tonomura, Applications of electron holography using a field-emission electron-microscope, J. Electron
Microsc. 33, 101-115 (1984).
82. H. W. Fink, I. Stoecker, and H. Schmid, Holography with low-energy electrons, Phys. Rev. Lett. 65(10),
1204-1206 (1990).
83. H.-W. Fink, H. Schmid, H. J. Kreuzer, and A. Wierzbicki, Atomic resolution in lensless low-energy electron
holography, Phys. Rev. Lett. 67(12), 1543-1546 (1991).
84. J. J. Saenz, N. Garcia, H. De Raerdt, and V. T. Binh, Electron emission from small microtips. Collogue de
Physique, Collogue C5 50(11), C8-73^8-78 (1989).
85. T. Kawasaki, M. Tomita, A. Tonomura, T. Miyada, T. Matsuda, and J. Endo, Development of 350 kv
holography electron-microscope. 2. Magnetic-field superimposed field-emission electron-gun, /. Electron
Microsc. 38, 295 (1989).
86. R. Morin, Point source physics: application to electron microscopy and holography, Microsc. Microanal.
Microstruct. 5, 1-10 (1994).
87. I. Brodie and C. A. Spindt, Vacuum microelectronics, in: Advances in Electronics and Electron Physics,
Vol. 83 (Academic Press, New York, 1992), pp. 1-106.
88. H. H. Busta, Vacuum microelectronics—1992. J. Micromech. Microeng. 2(2), 43-74 (1992).
89. F. M. Charbonnier, J. P. Barbour, L. F. Garrett, and W. P. Dyke, Basic and applied studies of field emission
at microwave frequencies, Proc. IEEE 51(7), 991-1004 (1963).
90. M. I. Elinson and A. V. Gor'kov, The device for generation of short-term electronic packages. Inventor's
Certificate 0137545. Bulletin of Inventions (1961), No 8.
91.1. Brodie, Application of field emission cathodes to microwave power tubes. Tri-Service Microwave Power
Tube Conf, Monterey, California, USA, (1976).
92. I. Brodie and C. A. Spindt, The apphcation of thin film field emission cathodes to electronic tubes, Appl.
Surf. Sci. 2(2), 149-163 (1979).
93. S. T. Smith and H. F. Gray, Distributed amplifier possibilities using field emitter arrays, 1st Int. Conf. Vacuum
Microelectronics (1988), Williamsburg, Virginia, USA.
94. D. Palmer, D. Temple, C. Ball, H. F. Gray, and G. McGuire, Low capacitance high-transconductance, silicon
raAs for microwave amplifier applications, 7th Int. Conf Vacuum Microelectronics (July 1994). Technical
Digest (Societe Francaise du Vide, Paris), p. 297.
95. M. Garven, M. A. Kodis, J. L. Shaw, K. T. Nguyen, and H. F. Gray, FEA emission characterisation and
beam transport for linear microwave power amplifiers, 7th Int. Conf. Vacuum Microelectronics, Portland,
Oregon (July 30-August 3, 1995). Technical Digest (Electron Device Society), pp. 231-234.
96. Yu. V. Gulyaev, A. Gritsenko, V. Efimov, Y Zakharchenko, and N. Sinitsyn, Functional possibility of vacuum
circuits based on distributed microwave systems, 3rd Int. Conf. Vacuum Microelectronics (1990), Monterey,
California, USA.
97. Yu. F. Zakharchenko, N. I. Sinitsyn, and Yu. V. Gulyaev, Distributed generator with extented interaction based
on field emitter arrays, 9th Int. Conf Vacuum Microelectronics, St. Petersburg, Russia (July 7-12, 1996).
Technical Digest (Nevsky Courier, St. Petersburg), pp. 610-613. / Vac. Sci. Technol. B 15(2), 533-534
(1997).
98. P. N. Laily, Y. Goren, and E. A. Nettesheim, A broadband distributed amplifier based on thin-film field-
emission cathodes, 1st Int. Conf. Vacuum Micrielectronics (1988), Williamsburg, Virginia, USA.
REFERENCES 175
99. W. A. Anderson, Frequency limits of electronic tubes with field emission cathodes. Vacuum Micro-
electronics 89 (Inst. Phys. Conf. Sen 99) edited by R. Turner (Institute of Physics, Bristol, 1989),
pp. 217-221.
100. H. G. Kosmahl, A wide-bandwidth high-gain small-size distributed ampHfier with field-emission triodes
(FETRODEs) for the 10 to 300 GHz frequency range, IEEE Trans. Electron Devices ED-36(11), 2728-2737
(1989).
101. V. I. Makhov, Low voltage pulsed magnetron with heating free excitation, 9th Int. Conf. Vacuum Micro-
electronics, St. Petersburg, Russia (July 7-12, 1996). Technical Digest (Nevsky Courier, St. Petersburg),
pp. 449-452.
102. K. R. Shoulders, Microelectronics using electron beam activated machining techniques, in: Advances
in Computing, edited by F. L. Alt (Academic, New York, 1961), pp. 135-293.
103. C. A. Spindt, C. E. Holland, I. Brodie, J. B. Mooney, and E. R. Westerberg, Field emitter arrays applied to
vacuum fluorescent display, IEEE Trans. Electron Devices ED-36(1), 225-228 (1990).
104. H. G. Gray, Vacuum microelectronics 1996: Where we are and where we going, 9th Int. Conf. Vacuum Micro-
electronics, St. Petersburg, Russia (July 7-12, 1996). Technical Digest (Nevsky Courier, St. Petersburg),
pp. 1-3.
105. R. Baptist, Trends and developments of vacuum microelectronics in Europe. 9th Int. Vacuum Microelectronics
Conf (July, 7-12 1966), St. Petersburg, Russia. Le Vide: Science Tecnique et Applications 52(282) 499-516
(1996).
106. R. E. Burgess, H. Kroemer, and J. M. Houston, Corrected values of Fowler-Nordheim field emission function
ny) and S(y), Phys. Rev, 90, 515 (1953).
107. V. I. Makhov, Field emission cathode technology and its apphcation, 4th Int. Vacuum Microelectronics Conf.
Technical Digest (Nagahama, Japan, 1991), pp. 40-43.
108. L. D. Karpov, A. P. Genelev, V. A. Drach, V. S. Zasenkov, Y. V. Mirgorodski, and A. N. Tikhonski, Patterning
of vertical thin film emitters in field emission arrays and their emission characteristics, 9th Int. Conf. Vacuum
Microelectronics, St. Petersburg, Russia (July 7-12,1996). Technical Digest (Nevsky Cur'er, St. Petersburg),
pp.501-504.
109. L. D. Karpov, A. P. Genelev, Y V. Mirgorodski, and A. N. Tikhonski, Patterning and electrical testing of
field emission arrays with novel emitter geometries, 9th Int. Conf. Vacuum Microelectronics, St. Petersburg,
Russia (July 7-12, 1996). Technical Digest (Nevsky Courier, St. Petersburg), pp. 542-546.
110. A. K. Gunguly, P. M. Philips, and H. F. Gray, Linear theory of a field-emitter-array distributed ampUfier,
J. Appl. Phys. 67(11), 7098-7110 (1990).
111. C. A. Spindt, C. E. Holland, P. R. Schwoebel, and L Brodie, Maximizing field-emitter-array transconductance
for microwave applications, 8th Int. Conf. Vacuum Microelectronics, Portland, Oregon (July 30-August 3,
1995). Technical Digest (Electron Device Society), p. 137.
112. B. Gorfinkel and J. M. Kim, Co-development offieldemission displays at scientific research institute "Volga"
(Russia) and Samsung Advanced Institute of Technology (Korea), 9th Int. Conf Vacuum Microelectronics,
St. Petersburg, Russia (July 7-12, 1996). Technical Digest (Nevsky Courier, St. Petersburg), pp. 552-556.
J. Vac. Sci. Technol. B, 15(2), 524-527 (1997).
113. C.-C. Wang, J.-C. Wu, C.-M. Huang, Data fine driver design for a 10^^480 x 640 x 3 color FED, 9th Int.
Conf Vacuum Microelectronics, St. Petersburg, Russia (July 7-12,1996). Technical Digest (Nevsky Courier,
St. Petersburg), pp. 557-561.
114. P. Vaudine and R. Meyer, 'Microtips' fluorescent display, Proc. lEDM 91 (Washington DC) (IEEE, New York,
1991) pp. 197-200.
115. C. Curtin, The field emission display: a new flat panel technology. Invited Paper, Conf Record of 1991 Int.
Display Res. Conf (IDRC 91, San Diego, C4) pp. 12-15.
116. A. Ghis, R. Meyer, P. Rambaud, F. Levy, and T. Leroux, Sealed vacuum devices: fluorescent microtip
displays, IEEE Trans. Electron Devices ED-38(10), 2320-2322 (1991).
117. A. Ting, C. M. Tang, T. Swyden, D. McCatrhy, and M. Peckerar, Field-effect controlled vacuum field-
emission cathodes. 4th Int. Vacuum Microelectronics Conf Technical Digest (Nagahama, Japan, 1991),
pp. 200-201.
118. L. Zhang, A. Q. Gui, and W N. Carr, Lateral vacuum microelectronic logic gate design, J. Micromech.
Microeng. 1(2), 126-134 (1991).
119. 1.1. Gofman, Study of electrostatic emission of electrons from tungsten in wide range of current densities,
Fizika Tv Tela 4(8), 2005-2015 (1962).
176 REFERENCES
120. A. E. Bell and L. W. Swanson, Total energy distributions of field-emitted electrons at high current density,
Phys. Rev. B 19, 3353-3364 (1979).
121. G. Lea and R. Gomer, Evidence of electron-electron scattering from field emission, Phys. Rev. Lett. 25,
804^806 (1970).
122. J. W. Gadzuk and E. W. Plummer, Hot-hole-electron cascades in field emission from metals, Phys. Rev. Lett.
26,92-95(1971).
123. V. T. Binh, S. T. Purcell, N. Garcia, and J. Doglioni, Field-emission electron spectroscopy of single-atom
tips, Phys. Rev. Lett. 69(17), 2527-2530 (1992).
124. M. Yu. Sumetskii and G. V. Dubrovskii, Tunneling through axially-symmetrical barrier, Yademaya Fizika
29, 1406-1413 (1979).
125. M. Yu. Sumetskii, Tunnehng probability through multidimensional potential barriers, TMF 45,64-75 (1980).
126. M. Yu. Sumetskii, About dimensional effects in the tunnehng process, Zh. Tekhn. Fiz. 54(11), 2227-2232
(1981).
127. B. B. Rodnevich, Energy spectrum offieldemission, 9th Int. Conf. Vacuum Microelectronics, St. Petersburg,
Russia (July 7-12, 1996). Technical Digest (Nevsky Courier, St. Petersburg), pp. 77-80.
128. J. W. Gadzuk, Many-body tunnehng-theory approach to field emission of electrons from solids. Surf. Sci.
15,466-482 (1969).
129. J. W. Gadzuk, Band-structure effects in the field-induced tunneling of electrons from metals, Phys. Rev.
182(2), 416-426 (1969).
130. F. I. Itskovich, On the theory of field-electron emission from metals, Zh. Ekspen i Teor Fiz- 50(5), 1425-1437
(1966).
131. F. I. Itskovich, On the theory of field-electron emission from metals, Zh. Ekspen i Teor. Fiz. 52(6), 1720-1735
(1967).
132. F. V. Bunkin and M. V. Fedorov, Cold emission of electrons from surface of metals in strong field of radiation,
Zh. Eksper i Tekh. Fiz. 48, 1341-1345 (1965).
133. L. V. Keldysh, Ionization in the field of a strong electromagnetic wave, J. Exp. Theo. Phys. 47, 1945-1957
(1964).
134. M. Buttiker and R. Landauer, Traversal time for tunnehng, Phys. Rev. Lett. 49, 1739-1742 (1982).
135. D. G. Sokolovski and L. M. Baskin, Traversal time in quantum scattering, Phys. Rev. A 36, 4604-4611
(1987).
136. N. MuUer, W. Eckstein, W. Heiland, and W. Zinn, Electron spin polarization in field emission from EuS-
coated tungsten tips, Phys. Rev. Lett. 29(25), 1651-1654 (1972).
137. N. MuUer, Electron spin polarization in field emission from nickel: effects of surface adsorption, Phys. Lett.
Sen A 54,415^16 (1915).
138. M. Landolt and M. Compagna, Spin polarization of field-emitted electrons and magnetism at the (100)
surface of Ni, Phys. Rev. Lett. 38, 663-666 (1977).
139. M. Landolt, M. Compagna, J. N. Chazalviel, Y. Yafet, and B. Wilkens, New tunnehng experiments from
strong ferromagnets and predictions of the band theory: Resolution of a controversy, J. Vac. Sci. Technol.
14,468^70 (1977).
140. M. Landolt and M. Compagna, Demagnetization of the Ni(lOO) surface by hydrogen adsorption, Phys. Rev.
Lett. 39(9), 568-570 (1977).
141. G. Chrobok, M. Hofmann, G. Regenfus, and R. Sizmann, Spin polarization on field-emitted electrons from
Fe, Co, Ni and rare-earth metals, Phys. Rev. Sen B 15(1), 429^^0 (1977).
142. G. Baum, E. Kisker, A. H. Mahan, W. Raith, and B. Reihl, Field emission of monoenergetic spin-polarized
electrons, Ap/?/. Phys. 14, 149-153 (1977).
143. M. Landolt and M. Compagna, Electron spin polarization in field emission, Surf. Sci. 70, 197-210 (1978).
144. M. Landolt and Y Yafet, Spin polarization of electrons field emitted from single-crystal iron surfaces, Phys.
Rev. Lett. 40, 1401-1403 (1978).
145. A. I. Baz', Ya. B. Zeldovich, and A. M. Perelomov, Scattering, reactions and decay in nonrelativistic quantum
mechanics. (Wiley, New York, 1969), p. 544.
146. E. Kisker, G. Baum, A. H. Mahan, W. Raith, and B. Reihl, Electron field emission from ferromagnetic
europium sulfide on tungsten, Phys. Rev. Sen B 18(5), 2256-2275 (1978).
147. G. N. Fursey, A. V. Kocheryzhenkov, and V. I. Maslov, The quantity of elementary acts and statistics of the
field emission. Surf. Sci. 246(1-3), 365-372 (1991).
REFERENCES 177
148. G. N. Fursey, A. V. Kocheryzhenkov, V. I. Maslov, and A. P. Smimov, Multi-particle tunneling during field
emission from YBa(2)Cu(3)0(7-d), Pis'ma v Zhumal Tekhnicheskoi Fiziki 14(20), 1853-1856 (1988).
149. R. Fischer and H. Neumann, Feldemission aus Halbleitem, Fortschr Phys. 14, 603-692 (1966).
150. G. N. Fursey and O. I. L'vov, News in the investigation of field-electron emission from semiconductors
(review), in: R. Fischer and Kh. Noyman, Field-electron emission from semiconductors, translated by
G. N. Fursey and B. I. Saprykin, edited by I. L. Sokol'skaya (Nauka, Moscow 1971), pp. 137-215.
151. L. M. Baskin, O. I. L'vov, and G. N. Fursey, General features of field emission from semiconductors,
Phys. Stat. Sol. (b) 47,49-62 (1971).
152. G. N. Fursey, Early field emission studies of semiconductors. Appl. Surf. Sci. 94/96, 44-59 (1996).
153. A. T. Rakhimov, B. V. Seleznev, N. V. Suetin, A. V. Kandidov, V. A. Tugarev, and I. A. Leont'ev, Examination
of electron field emission efficiency and homogeneity from CVD diamond films, Proc. Appl. Diamond Conf
1995, Applications of Diamond Films and Related Materials: 3rd Int. Conf, National Institute of Standards
and Technology, Gaithersburg, MD (August 21-24, 1995). Edited by Albert Feldman, National Institute of
Standards and Technology Gaithersburg, MD, USA, pp. 11-14.
154. Yu. V. Gulyaev, L. A. Chemozatonskii, Z. Ya. Kosakovskaya, A. L. Musatov, N. I. Sinitsin, and G. V.
Torgashov, Carbon nanotubes structures a new material of vacuum microelectronics, 9th Int. Conf Vacuum
Microelectronics, St. Petersburg, Russia (My 1-12,1996). Technical Digest (Nevsky Cur'er, St. Petersburg),
pp. 5-9.
155. Yu. V. Gulyaev, N. I. Sinitsyn, G. V. Torgashov, Sh. T. Mevlyut, A. I. Zhbanov, Yu. V. Zakharchenko, Z.
Ya. Kosakovskaya, L. A. Chemozatonskii, O. E. Glukhova, and I. G. Torgashov, Work function estimate for
electrons emittedfromnanotube carbon clusterfilms.9th Int. Conf Vacuum Microelectronics, St. Petersburg,
Russia (July 7-12,1996). Technical Digest (Nevsky Cur'er, St. Petersburg), pp. 206-210. /. Vac. Sci. Technol.
B 15(2), 422-424 (1997).
156. Yu. V. Gulyaev, N. I. Sinitsyn, G. V. Torgashov, Yu. A. Grigoriev, V. I. Shesterkin, A. G. Veselov, Yu. V.
Shvetsov, and V. C. Semyonov, Emission of low-voltage multi-tip carbon matrices coated by carbon clusters,
9th Int. Conf Vacuum Microelectronics, St. Petersburg, Russia (July 7-12,1996). Technical Digest (Nevsky
Cur'er, St. Petersburg), pp. 519-521.
157. F. Y. Chuang, C. Y. Sun, H. F. Cheng, W. C. Wang, C. M. Huang, and I. N. Lin, Electron emission
characteristics of pulsed laser deposited diamond-like films, 9th Int. Conf. Vacuum Microelectronics, St.
Petersburg, Russia (July 7-12, 1996). Technical Digest (Nevsky Cur'er, St. Petersburg), pp. 334-338.
158. H. Kim, Y. W Choi, J. O. Choi, H. S. Jeong, and S. Ahn, I-V characteristics of volcano type Si FEAs with
diamond-like carbon coating, 9th Int. Conf. Vacuum Microelectronics, St. Petersburg, Russia (July 7-12,
1996). Technical Digest (Nevsky Courier, St. Petersburg), pp. 344-348.
159. L. W. Nordheim, Die Theorie der Elektronenemission der Metalle, Z. Physik 30(7), 177-196 (1929).
160. N. Froman and P. O. Froman, JWKB Approximation. Contribution to the Theory. (New-Holland Publishing
Company, Amsterdam, 1965).
161. D. Bohm, Quantum Theory. (Prentice-Hall, New York, 1952).
162. W. W. Dolan, Current density tables for field emission theory, Phys. Rev. 91(3), 510-511 (1953).
163. E. Guth and C. J. MuUin, Electron emission of metals in electric fields, Phys. Rev. 61(5-6), 339-348 (1942).
164. E. L. Murphy and R. H. Good, Thermionic emission, field emission and the transition region, Phys. Rev.
102(6), 1464-1473 (1956).
165. S. G. Christov, General theory of electron emission from metals, Phys. Stat. Sol. 17(1), 11-26 (1966).
166. L S. Andreev, Investigation of electron emission from metal in the range of transition from cold emission to
thermionic field emission, Zh. Tekh. Fiz. 22(9), 1428-1441 (1952).
167. W. W. Dolan and W. P. Dyke, Temperature and field emission of electrons from metals, Phys. Rev. 95(2),
327-332 (1954).
168. M. I. Elinson, F. F. Dobryakova, V. F. Krapivin, and Z. A. Malina, Yasnopol'skaya, On the theory of field-
and thermionic-field emission of metals and semiconductors, Radiotekhnika i Electronika 6(8), 1342-1353
(1961).
169. L. D. Landau and E. M. Lifschitz, Quantum Mechanics. (Nauka, Moscow, 1982).
170. R. Z. Bakhtizin and Yu. M. Yumaguzin, Study of the high-energy tail of the tungsten field-electron
distribution, Poverkhnost': Fiz., Khim., Mekhan. 7, 51 (1987).
171. R. Z. Bakhtizin and Yu. M. Yumaguzin, High energy tails in the total energy-distribution of field emitted
electrons from tungsten, Phys. Stat. Sol. B 147, 241 (1988).
178 REFERENCES
172. G. Binning, H. Rohrer, Ch. Gerber, and E. Weibel, Surface studies by scanning tunneling microscopy, Phys.
Rev. Lett. 49(1), 57-61 (1982).
173. G. N. Fursey and D. V. Glazanov, Field emission properties of ultrasmall Zr spots on W, J. Vac. Sci. Technol.
B 13(3), 1044-1049 (1995).
174. G. N. Fursey, L. M. Baskin, D. V. Glazanov, A. O. Yevgen'ev, A. V. Kotcheryzhenkov, and S. A. Polezhaev,
The specific features of field emission from submicron cathode surface areas at high current densities, /
Vac. Sci. Technol. B, 16(1), 232-237 (1998).
175. P. H. Cutler, Jun He, and N. M. Miskovsky, Theory of electron emission in high fields from atomically sharp
emitters: validity of the Fowler-Nordheim equation, J. Vac. Sci. Technol. B 11(2), 387-391 (1993).
176. Jun He, P. H. Cutler, N. M. Miskovsky, T. E. Feuchtwang, T. E. Sullivan, and M. Chung, Derivation of the
image interaction for non-planar pointed emitter geometries:application tofiledemission I-V characteristics.
Surf. Sci. 246, 348-364 (1991).
177. A. A. Porotnikov and B. B. Rodnevich, On the structure of the potential barrier at a metal boundary, Appl.
Mech. Tech. Phys. 5, 80-86 (1978).
178. V. N. Shrednik, Field emission microscopy from metal-film covers. Ph. D. Dissertation, loffe Institute of
Physics and Technology, USSR Academy of Sciences, Leningrad (1965).
179. V. N. Shrednik, Borders of applicability of the Fowler-Nordheim theory and corrections to the theory, in:
Cold Cathodes, edited by M. I. Elinson (Sovetskoe Radio, Moscow, 1974), pp. 169-177.
180. I. M. Lifshits, M. Ya. Azbel', and M. N. Kaganov, Electronic Theory of Metals. (Nauka, Moscow,
1971).
181. D. N. Krotevich, Study of surface microgeometry and structure in the pre-explosion phase of electron
emission. Ph. D. Dissertation, Prof. M. A. Bonch-Bruevich Electrotechnical Institute of Teleconmiunication,
Leningrad (1985).
182. G. N. Fursey, L. M. Baskin, and D. V. Glazanov, Study of general emissive parameters of sub-nanometer
field emission emitter. 42nd Int. Field Emission Symp., Madison, Wisconsin, USA, (August 7-11, 1995).
183. B. A. Politzer and P. H. Cutler, Model calculation of field emission from 3d bands in nickel. Surf. Sci. 22(2),
277-289 (1970).
184. B. A. Politzer and P. H. Cutler, Theory of electron tunneling from d bands in solid. Application to field
emission. Mater Res. Bull. 5(8), 703-720 (1970).
185. A. M. Brodskiy and Yu. A. Gurevich, Theory of Field Electron Emission from Metals. (Nauka, Moscow,
1973).
186. L. M. Baskin. To the theory of field emission from semiconductors. Ph. D. Dissertation, Institute of automatic
control systems and radioelectronic, Tomsk (1975).
187. I. O. Kulik and I. K. Yanson. Jozephson Effect in Superconducting Tunnel Structures. (Nauka, Moscow,
1970), p. 144.
188. F. R. Schrieffer, Theory of Superconductivity. (Benjamin, New York, 1964).
189. A. A. Abrikosov, L. P. Gor'kov, and I. E. Dzyaloshinski, Methods of Quantum Field Theory in Statistical
Physics. (Fizmatgiz, Moscow, 1962).
190. J. W. Gadzuk and E. W. Plummer. Field emission energy distribution (FEED), Rev. Modem Phys. 45(3),
487-548 (1973).
191. R. Young, Search for field emission of "paired" or correlated electrons. Proc. 14th Field Emission Symp.,
NBS, Maryland and Georgetown University (1967), pp. 7-10.
192. T. Engel and R. Gomer, Search for field-emission of electron pairs. Proc. 14th Field Emission Symp., NBS,
Maryland and Georgetown University (1967), p. 6.
193. J. Bardeen, The image and van der Waals forces at a metaUic surface, Phys. Rev. 58(8), 727-735 (1940).
194. G. N. Fursey and I. D. Tolkacheva, High field emission current and the effects preceding vacuum breakdown
for Ta and Mo emitters, Radiotechnika I Electronika 8(7), 1210-1221 (1963).
195. G. N. Fursey, Impulse field emission from rhenium, Zh. Tekh. Fiz^ 34(7), 1312-1316 (1964).
196. T. E. Stem, B. S. Goshng, and R. H. Fowler, Further studies in the emission of electrons from cold metals.
Royal Soc. Proc. A124, 699-722 (1929).
197. N. B. Isenberg, Role of a volume charge in spherical electron projectors. Zh. Tekh. Fiz. 24(11), 2079-2082
(1954).
198. N. B. Isenberg, Influence of a volume charge on a form of Inl = f{I/ V) characteristic of field cathodes,
Radiotekhnika I Electronika 9(12), 2147-2155 (1964).
REFERENCES 179
199. A. S. Kompaneets, Influence of a volume charge of field emission, Radiotekhnika I Electronika 5(8),
1315-1317(1960).
200. A. S. Kompaneets, Influence of a volume charge on field emission, Doklady AN USSR 128(6), 1160-1162
(1959)
201. R. N. Poplavskii, Potential distribution in a spherical capacitor in case of saturation current, Zh. Tekh. Fiz.
20(2), 149-159. (1950).
202. V. A. Godyak, L. V. Dubovoi, and G. R. Zabolotskaya, Calculation of field-emission current of relativistic
electrons confined by a space charge. Zh. Eksp. Tekh. Fiz. 57(11), 1795-1798 (1969).
203. L. M. Baskin, V. A. Godyak, O. I. Lvov, G. N. Fursey, and L. N. Shirochin, Influence of space charge of
relativity electrons on field emission. Zh. Tekh. Fiz. 57(11), 1795-1798 (1972).
204. G. N. Fursey, A. A. Antonov, and B. F. Gulin, Study of tungstenfieldemission in nanosecond range of pulse
durations, Vestnik Leningrad. University, 10, 71-74 (1971).
205. J. Bardeen, Theory of the work function, Phys. Rev. 49(9), 653-663 (1936).
206. H. J. Juretchke, Exchange potential in the surface region of a free electron metal, Phys. Rev. 93(5), 1140-1144
(1953).
207. T. L. Loucks, P. H. Cutler, The effect of correlation of the surface potential of a free electron metal, J. Phys.
Chem. Solids IS, 105-113 (1964).
208. T. J. Lewis, Some factors influencing field emission and the Fowler-Northeim law, Proc. Phys. Soc. B68,
Part n, No. 43 IB, 938-943 (1955).
209. P. H. Cutler and J. J. Gibbons, Model for the surface potential barrier and the periodic deviation on the
Schottky effect, Phys. Rev 111(2), 394-402 (1958).
210. G. G. Bertold, A. Kuppermann, and T. E. Phipps, Application of numerical methods to the theory of the
periodic deviations in the Schottky effect, Phys. Rev. 128(2), 524-531 (1962).
211. G. F. Vasil'ev, Influence of potential barrier shape at an emitter-vacuum boundary and electric field dis-
tribution along emitter surface upon the current-voltage characteristics of field emission, Radiotekhnika i
Elektronika 5(U), 1857-1861 (1960).
212. F Seitz, Modem Theory of Solids. (McGraw-Hill, New York, 1940), p. 163.
213. S. C. Miller and R. H. Good, A WKB-type approximation to the Schrodinger equation, Phys. Rev. 91(1),
174-179 (1953).
214. C. A. Spindt, A thin film field emission cathode, J. Appl. Phys. 39, 3504-3505 (1968)
215. M. Yu. Sumetskii, Transparency of non-synmietric tunnel micro-contact and semi-classical theory of
scanning tunnehng microscope. Zh. Exper i Teor Fiz. 94(3), 7-22, (1988).
216. A. A. Lucas, H. Morawitz, G. R. Henry, J. P. Vigneron, Ph. Lambin, P. H. Cutler, and T. E. Feuchtwang,
Tunneling through localized barriers with application to scanning tunneling microscopy: New scattering
theoretic approach and results, / Vac. Sci. Technol. A 6(2), 296-299 (1988).
217. Z. H. Huang, T. E. Feuchtwang, P. H. Cutler, and E. Kazes, Wentzel-Kramers-Brillouin method in
multidimensional tunnehng, Phys. Rev. A41(l), 32-41 (1990).
218. W. P. Dyke, J. K. Trolan, W. W. Dolan, and G. Barnes, The field emitter: fabrication, electron microscopy,
and electric field calculations. J. Appl. Phys. 24, 570-576 (1953).
219. G. N. Fursey and D. V. Glazanov, Deviation from the Fowler-Nordheim theory and peculiarities of field
electron emission from small-scale objects (Review). J. Vac. Sci. Technol. B 16(2), 910-915 (1998).
220. P. R. Schwoebel and I. Brodie, Surface-science aspects of vacuum microelectronics, J. Vac. Sci. Technol.
(B) 13(4), 1391-1410 (1995).
221. H.-W. Fink, H. Schmid, and H. J. Kreuzer, In-fine holography, using low-energy electrons and photons:
applications for manipulation on a nanometer scale, J. Vac. Sci. Technol. (B) 13(6), 2428-2431 (1995).
222. G. N. Fursey, L. M. Baskin, D. V. Glazanov, A. O. Yevgen'yev, A. V. Kotcheryzhenkov, and S. A.
Polezhaev, Specific features offieldemissionfi-omsubmicron cathode surface areas at high current densities,
J. Vac. Sci. Technol. (B) 16(1), 232-237 (1998).
223 . G . N . Fursey, Study of electron emittion in extremely high electricfieldand in the condition of transition to the
vacuum arc. Ph. D. dissertation. Institute Of Semiconductors (Russian Academy Of Science), Novosibirsk
(1973), pp. 1-393.
224. M. S. Aksyonov, L. M. Baskin, V. M. Zhukov, N. F Fedorov, and G. N. Fursey, Increase of the region of
localization of the Nottingham effect duringfield-electronemission under low temperature conditions, Izv.
Akad. Nauk SSSR: Ser. Fiz 43(3), 543-546 (1979).
180 REFERENCES
225. L. W. Swanson, L. C. Grouser, and F. V. Charbonnier, Energy exchanges attending field electron emission,
Phys. Rev. 151, 327-340 (1966).
226. P. H. Levine, Thermoelectric phenomena associated with electron-field emission, J. Appl. Phys. 33,582-587
(1962).
227. W. B. Nottingham, Remarks on energy losses attending thermoionic emission of electrons from metals,
Phys. Rev. 58, 906-907 (1940).
228. S. I. Anisimov, B. L. Kapehovich, and T. L. Perel'man, Emission of electrons from the surface of metals
induced by ultrashort laser pulses, Zh. Eksper. i Teor. Fiz. 66(2), 776-781 (1974).
229. N. F. Mott and E. H. Jones, The Theory of the Properties of Metals and Alloys, (Oxford university press,
London, 1936), pp. 305-314.
230. R. S. Stratton, Energy distribution of field emitted electron, Phys. Rev. 135, A, 794-805 (1964).
231. J. Houston, Private communication, See also [225].
232. M. Drechsler and G. Jaehing, Heating and cooling of thin wires by field emission, Z. Naturforsch, 18a,
1367-1369 (1963).
233. I. Engle and P. H. Cutler, The effect of different surface barrier models on the Nottingham energy exchange
process. Surf. Sci. 8, 288 (1967).
234. W. B. Nottingham, Remarks on energy losses attending thermionic emission of electrons from metals, Phys.
Rev. 59, 907 (1941).
235. E. A. Litvinov, G. A. Mesiats, and A. F. Shubin, Calculation of thermo-field emission preceding the explosion
of micro-emitters caused by electron field-emission pulses, Izv. Vysshikh Uchebnikh Zavedenii SSSR: Sen
Fiz. 4, 147-151 (1970).
236. V. A. Nevrovskii and V. I. Rakhovskii, Thermal instability of microprotrusion on a cathode during vacuum
breakdown, Zh. Tekh. Fiz. 50(10), 2127-2135 (1980).
237. J. Mitterauer, P. Till, and M. Haider, The initiation of cathode induced vacuum breakdown by dynamic field
emissionfrommicroprotrusions. Proc. 7th Int. Symp. Discharges Electrical Insulation Vacuum, 1976, Proc.
ISDEIV, Novosibirsk (1976), pp. 83-87.
238. D. V. Glazanov, L. M. Baskin, and G. M. Fursey, Kinetic of the pulsed heating of field-emission cathode
point with real geometry by a high-density emision current, Zh. Tekh. Fiz. 59(5), 60-68 (1989).
239. V. A. Gor'kov, M. I. Elinson, and G. D. Yakovleva, Theoretical and experimental investigations of pre-arc
phenomena during field-electron emission, Radiotekhnika i Elektronika 7(9), 1501-1510 (1962).
240. F M. Charbonnier, R. W. Strayer, L. W. Swanson, and E. E. Martin, Nottingham effect in field and T-F
emission: heating and cooling domains, and inversion temperature, Phys. Rev. Lett. 13, 397-401 (1964).
241. L. D. Landau and E. M. Lifshitz, Electrodynamics ofContinua. (Nauka, Moscow, 1982).
242. Ibid., p. 620.
243. Tables of Physical Constants, edited by L K. Kikoin, Moscow, Atomizdat, 1976.
244. G. N. Fursey, D. V. Glazanov, and S. A. Polezhaev, Field emission from nanometer protrubances at high
current density, IEEE Trans. Dielect. Electric. Insul. 2(2), 281-287 (1995).
245. D. V. Glazanov, L. M. Baskin, and G. N. Fursey, Numerical modeling of the development of thermal
instability of the filed emission initiating vacuum breakdown. XI Int. Symp. Discharges Electrical Insulation
Vacuum, Berlin, German Democratic Republic, September 24-28, 1984, Pro. ISDEIV, (1984), pp. 65-68.
246. L. M. Baskin, D. V. Glazanov, and G. N. Fursey, Influence of thermoelastic stresses on the destruction
processes of field emission cathode points and the transition to explosive emission, Zh. Tekh. Fiz. 59(5),
130-133 (1989).
247. D. N. Krotevich, V. E. Ptitsyn, and G. N. Fursey, Observation of the fine structure of rearranged microcrystal
surface by pulsed filed-emission microscopy, Fiz. Tverd. Tela 28(12), 3722-3724 (1986).
248. M. I. Kaganov, I. M. Lifshitch, and L. V. Tantarov, Relaxation between electrons and the crystalline lattice,
Zh. Eksper i Teor Fiz. 31(2), 232-237 (1956).
249. G. N. Fursey, L. M. Baskin, D. V. Glazanov, A. O. Evgen'ev, A. V. Kotcheryzhenkov, and S. A. Polezhaev,
The specific features of field emission from submicron cathode surface areas at high current densities,
J. Vac. Sci. Technol. B 16(1) 232-238 (1998).
250. G. N. Fursey, B. F. Gulin, and A. A. Antonov, Investigation of the field-electron emission from tungsten in the
nanosecond range of pulse durations, Vestnik Leningradskogho Universiteta: Seria 4. Physics & Chemistry
2, 71-74 (1971).
25 L M. I. Elinson and G. F. Vasil'ev, Investigation of the filed emission from lanthanum hexaboride. Radio Eng.
Electron. 3(7), 123-124 (1954).
REFERENCES 181
252. M. I. Elinson and G. A. Kudnitseva, Field-electron cathode of metallic high-melting compounds,

Radioteknika iElectronika 7(9), 1511-1518 (1962).
253. G. N. Fursey and S. A. Shakirova, Field emission localization within small solid angles, Proc. Electronics
Technique Conf. Electron-beam and photo-electron Devices 1, 99-102 (1969).
254. L. M. Baskin, R. Z. Bachtizin, V. G. Valeev, and G. N. Fursey, Advances in the Field Emission Investigations,
Supplement to the Russian Translation of A. Modinos, Field, Thermonic, and Secondary Electron Emission
Spectroscopy, edited by G. N. Fursey (Nauka, Moscow, USSR, 1985), pp. 232-302 (in Russian).
255. G. N. Fursey, L. L. Anan'ev, and D. A. Borisov, Field emission in microwave field. Proc. FV All-Union
Seminar on Microwave Relativistic Electronics, Moscow, USSR (1985), pp. 1-20 (in Russian).
256. H. F. Gray, The electronics of the 21st century, nanoelectronics. Vacuum Microelectronics 8, Electro/89,1-6
(1989).
257. S. T. Smith and H. F. Gray, Distributed amplifier possibilities using field-emitter arrays, Proc. 1st Int. Conf.
Vacuum Microelectronics, Williamsburg, Virginia, USA (1988), pp. 7-9.
258. R. L. Parker, RF vacuum microelectronics, 7th Int. Conf. Vacuum Microelectronics, Grenoble, France
(1994).
259. Yu. V. Gulyaev, I. S. Nefedov, N. I. Sinitsyn, G. T. Torgashov, Yu. F. Zakharchenko, and A. I. Zhbanov,
Distributed microwave amplifier on field emitter arrays with a nonhomogeneous energy collector. Revue
"le vide, les Couches Minces"-Supplemeny au No. 271, 84-87 (1994).
260. V. Ambageokar, U. Eckem, and G. Schoen, Quantum dynamics of tunneling between superconductors, Phys.
Rev. Lett. 48(25), 1745-1748 (1982).
261. R F. Voss and R. A. Webb, Macroscopic quantum tunneling in 1 \im Nb Josephson junctions, Phys. Rev.
Lett. 47(4), 265-268 (1981).
262. A. M. Gabovich, V. M. Rosenbaum, and A. L. Loitenko, Dynamic image forces in three-layer systems and
field emission. Surf. Sci. 186(3), 523-549 (1987).
263. E. P. Wigner, Lower limit for the energy derivative of the scattering phase shift, Phys. Rev. 98(1), 145-147
(1955).
264. T. E. Hartman, Tunneling of a wave packet, Appl. Phys. 33(12), 3427-3433 (1962).
265. A. Goldberg, H. M. Schey, and J. L. Schwartz, Computer-generated motion pictures of one-dimensional
quantum-mechanical transmission and reflection phenomena, Amer J. Phys. 35(3), 177-193 (1967).
266. P. Schnupp, The tunneling time of an electron and the image force. Thin Solid Films 2(3), 177-183 (1968).
267. A. I. Baz, Life-time of intermediate states, J. Nucl. Phys. (USSR) 4(2), 252-260 (1966) (in Russian).
268. A. L Baz, A quantum mechanical calculation of collision time, J. Nucl. Phys. (USSR) 5(1), 229-235 (1967)
(in Russian).
269. V. F Rybachenko, Time of particle penetrating through barrier, / Nucl. Phys. (USSR) 5(4), 895-901 (1967)
(in Russian).
270. M. Buttiker, Larmor precision and the traversal time for tunneling, Phys. Rev. B 27(10), 6178-6188 (1983).
271. R. Feinman and A. Gibbs, Quantum Mechanics and Path Integrals (McGraw-Hill, New York, 1969).
272. D. G. Sokolovski, Quantum time of flight and adiabatic hmit in scattering problems. Report of the Ministry
High Schools: Physics. (USSR) 3, 52-58 (1988) (in Russian).
273. L. M. Baskin and D. G. Sokolovski, Traversal time in the process of tunnehng. Report of the Ministry of
High Schools: Physics. (USSR) 3, 26-29 (1987) (in Russian).
274. D. G. Sokolovski and L. M. Baskin, Time of flight at tunneling, J. Tech. Phys. (USSR) 55(9), 1838-1840
(1985) (in Russian).
275. Linfield College, McMinnville, Ore., Engineering Report (1953).
276. L. L. Anan'ev, M. M. Bogatskii, D. A. Borisov, A. A. Kantonistov, and G. N. Fursey, Method of field
emission investigation in microwave electric field, Instrum. Tech. Exp. (USSR) 5, 165-168 (1983) (in
Russian) [Instrum. Exp. Tech. (USA) 26, 1181 (1983)].
277. L. M. Baskin, A. A. Kantonistov, G. N. Fursey, and L. A. Shirochin, Pecuharities in explosive emission of
liquid metals in microwave field. Rep. USSR Academy of Sciences (USSR) 296(6), (1987) 1352 (in Russian).
278. A. A. Kantonistov, L N. Radchenko, G. N. Fursey, and L. A. Shirochin, Field emission of liquid metals in
alternating fields. Collogue de Physique, CoUoque C8, supplement au Tome 50,11, C8-203-207 (1989).
279. D. V. Glazanov. Thesis ofkandidate dissert., Tomsk, 1985 (in Russian).
280. L. L. Anan'ev, D. A. Borisov, A. A. Kantonistov, and G. N. Fursey, Study of field emission at high current
densities in microwave field, Proc. IVAll-Union Symp. High-Current Electronics, Tomsk, pp. 26-29 (1982)
(in Russian).
182 REFERENCES
281. M. I. Elinson and V. A. Gorkov, Some peculiarities of field emission cathode operation in microwave field,
Radiotech. Electron. (USSR) 6(2), 336-339 (1961) (in Russian).
282. R. M. Voronkov, V. F. Gass, V. A. Danilichev, and I. A. Smimov, The testing of injector with field emission
cathode in accelerator, Pribory i Tekhnica Experimenta (USSR) 4, 18-20 (1974) (in Russian).
283. L. D. Landau and E. M. Lifschitz, Mechanics (Nauka, Moscow, 1965).
284. L. M. Baskin, E. A. Sazonov, L. M. Chemych, and G. N. Fursey, Influence of transient effects on parameters
of microwave diode with non-heated cathode, Radiotech. Electron. (USSR) 33(10), 2208-2210 (1988)
(in Russian).
285. E. V. MuUer and T. T. Tsong, Field-ion microscopy (American Elsevier Publishing Company, Inc., New York,
1969).
286. L. Tonks, A theory of liquid surface rupture by uniform electric field, Phys. Rev. 48, 562-568 (1935),
287. Ya. Frenkel, On a theory of liquid surface instability in electric field in vacuum, Zh. Tekh. Fiz. 6, 347-351
(1935).
288. K. Hata, R. Ohya, S. Hishigaki, H. Tamura, and T. Noda, Stable field emission of electrons from liquid
metal, Jpn J. Appl. Phys. 26(6), L896-L898 (1987).
289. L. M. Baskin, G. N. Fursey, A. A. Kantonistov, B. N. Movchan, I. N. Radchenko, and L. A. Shirochin,
Microstructure formation at the Uquid surface in microwave fields, Proc. 14th Int. Symp. Discharges
Electrical Insulations Vacuum, Santa-Fe, CA, (1990), p. 112.
290. N. Fursey, L. A. Shirochin, and L. M. Baskin, Field emission processes from liquid-metal surface, J. Vac.
Sci. Technol. B 15(2), 410-^21 (1997).
291. J. R. Arthur, Surface structure and surface migration by field emission microscopy, J. Phys. Chem. Solids
25, 583-591 (1964).
292. G. Busch and T. Fischer, Feldemission aus Sihzium, Phys. Kondens. Materie 1, 367-393 (1963).
293. J. R. Arthur, Gallium arsenide surface structure and reaction kinetics:fieldemission microscoipy and surface
migration, J. Appl. Phys. 37(6), 3057-3064 (1966).
294. F. G. Allen, Field emission from Silicon and Germanium; Filed decorbtion and surface migration, Chem.
Solids 19, 87-99 (1961).
295. V. G. Ivanov, On preparing germanium emitters and obtaining a field emission picture of pure germanium,
Radi. ElectK 10(3), 576-578 (1965).
296. J. R. Arthur, Photosensitive field emission from p-type Germanium, J. Appl. Phys. 36, 3221-3227
(1965).
297. V. G. Ivanov, G. N. Fursey, and I. L. Sokol'skaya, Study of field emission from germanium, Fiz. Tverd. Tela
9(4), 1144-1148 (1967) (in Russian).
298. L. Ernst, tJber die Oberflachenreinigung von Germaniumspitzen durch Felddesorption im Feldemissions-
mikroskop, Phys. Status Solidi 7, K61-K64 (1964).
299. L. Ernst, Study of the field desorption of oxidized germanium tips in a field emission microscope, Phys.
Status Solidi 14, K107-K109 (1966).
300. L. Ernst, Feldemissions-Mikroskopie von Germanium, Phys. Status Solidi 19, 89 (1967).
301. L. A. D'Assaro, Field emission from silicon, J. Appl. Phys. 29, 33-34 (1958).
302. R. L. Perry, Experimental determination of the current density electric field relationship of silicon field
emitters, J. Appl. Phys. 33, 1875-1883 (1962).
303. G. N. Fursey and V. G. Ivanov, Field-electron emission from p-type Si cleaned by field desorption,
Fiz. Tverd Tela 9(6), 1812-1818 (1967) (in Russian).
304. G. N. Fursey, V. G. Egorov, S. P. Manokhin, and M. R. El'Nimr, Relaxation effects during field emission
from silicon, Fiz. Tverd Tela 15(5), 1360-1363 (1973) (in Russian).
305. S. G. Truxillo, J. C. Blair, N. G. Einspruch, and R. Stratton. High-field electron emission from indium
arsenide, / Chem. Phys. 44, 1724 (1966).
306. W. R. Savage, High-field electron emission from indium antimonide. Solid State Commun. 1, 144-147
(1963).
307. J. Marien and J. Loosveld, Field emission studies on clean symmetrical surfaces of cadmium sulfide, Phys.
Status Solidi (a) 48, 213 (1971).
308. A. J. Melmed and R. J. Stem, Field-ion microscopy of silicon, Surf. Sci. 49, 645-656 (1975).
309. A. J. Melmed and E. I. Givargizov, unpublished research, NBS, 1979.
310. E. I. Givargizov, Krist. Tech. 10 (1975) 473 (in Russian); in: Current Topics in Materials Science, edited by
E. Kaldis (North-Holland, Amsterdam, 1978), p. 79.
REFERENCES 183
311. U. Apker and E. A. Taft, Field emission from photoconductors, Phys. Rev. 88, 1037-1038 (1952).
312. I. L. Sokol'skaya and G. P. Shcherbakov, Study of strong field on field-electron emitters of CdS crystal, Fiz.
Tverd. Tela31(ll 167-175 (1961).
313. G. P. Shcherbakov and I. L. Sokol'skaya, Experimental investigation of energy distribution of filed-emitted
electrons from CdS single crystals, Fiz. Tverd. Tela 4(12), 3526-3536 (1962).
314. O. V. Golovanova and A. I. Klimin, Sb2S3 field-electron emission, Radi. Electr. 10, 491^97 (1965)
(in Russian).
315. P. G. Borzjak, A. F. Yatsenko, and L. S. Miroshnichenko, Photo-field-emission from high-resistance silicon
and germanium, Phys. Status Solidi 14, 403^11 (1966).
316. A. F. Yatsenko, Photo-field emission from high-resistance p-type semiconductors, Elektr. Tekhnika, Ser, 4:
Elektr.-luchevye i Fotoelektr. Pribory (in Russian) 2, 169-175 (1968).
317. Z. P. But, A. E. Kravzov, and A. F. Yatsenko, Field electron emission from semiinsulator galium arsenide
p-type, XIII Conf. Emission Electronics, Moscow (1968), p. 42.
318. H. Neumann, Field emission from In2S3, Z. Naturforsch. 22a(7), 1012-1019 (1967).
319. H. Neumann, Field emission from silver-doped cadmium antimonide, Z. Naturforsch. 23a(8), 1240
(1968).
320. L. M. Baskin, O. I. L'vov, and G. N. Fursey, General features of field emission from semiconductors, Phys.
Status Solidi (b) 47,49-62 (1971).
321. N. V. Egorov, G. N. Fursey, and M. K. El'Nimr, Preservation of semiconductor field emission cathode
properties, Zh. Tech. Fiz. 44(5), 1117-1118 (1974) (in Russian).
322. R. R. Stead and O. W. Richardson, USA Patent No. 2.871.111.
323. N. D. Morguhs, On Shottky effect in compound semiconductor cathodes, Zh. Eksp. i Teon Fiz. 16,959-954
324. R. Stratton, Fieldemission von Halbleitem, Proc. Phys. Soc. B68, 746-756 (1955).
325. R. Stratton, Theory of field emission from semiconductors, Phys. Rev. 1251, 67-82 (1962).
326. M. S. Kagan, N. M. Lifshits, A. L. Musatov, and A. A. Sharonov, Field-electron emission from high-resistivity
germanium, Fiz. Tverd. Tela. 6, 723-727 (1964) (in Russian).
327. I. L. Sokol'skaya and G. P. Scherbakov, Nonhnearity of the current-voltage characteristics of field emitter
from CdS single crystals, Fiz. Tverd. Tela 4(1), 44-51 (1962) (in Russian).
328. G. N. Fursey and N. V. Egorov, Field emission from p-type Si, Phys. Status Solidi 32, 23-29 (1969).
329. G. N. Fursey and R. S. Bakhtizin, Nonlinear current-voltage characteristics of p-type Ge, Fiz. Tverd. Tela
11, 3672-3674 (1969) (in Russian).
330. M. I. Ehnson, A. G. Zdan, V. F. Krapivin, Zh. B. Lipkovskii, V. N. Lutskii, and V. B. Sandomirskii, The
theory of "non contact" emission of hot electrons from semiconductors, Radiotekhnika i Elektronika 10,
1288 (1965).
331. N. D. Morguhs, On the problem of filed-electron emission of complex semiconductor cathodes, Zh. Tech.
Fiz. 17(9), 983-986 (1947).
332. S. G. Christov, Phys. United theory of thermionic and field emission from semiconductors, Phys. Status
Solidi 21,159-113 (1967).
333. P. G. Borzyak, A. F. Yatsenko, and L. S. Miroshnichenko, Photo-field—emission from high-resistance siUcon
and germanium, Phys. Status Solidi 14,403^11 (1966).
334. G. N. Fursey, I. L. Sokolskaya, and V. G. Ivanov, Field emission from p-type germanium, Phys. Status Solidi
22,39-46(1967).
335. G. N. Fursey, M. I. Kaplan, and O. I. Lvov, On the theory of field-electron emission from p-type
semiconductors, Vestnik Leningrad Univ. 16, 167-170 (1968).
336. R. Seiwatz and M. Green, Space-charge calculations for semiconductors, J. Appl. Phys. 29, 1034-1340
(1958).
337. E. M. Conwell, High field transport in semiconductors (Academic Press, NY 1967), W. Shockley, Bell Syst.
Tech. J. 30,990 {195\).
338. A. F. Yatsenko, On a model of photo-field-emission from p-type semiconductors, Phys. Status Solidi (a)
1(2), 333-348 (1970).
339. Z. P. But, L. S. Miroschenko, and A. F. Yatsenko, Temperature dependence of field emission from p-type
sihcon with various bulk doping, Uknfiz. Zh. 17, 949-955 (1972).
340. R. Fischer, Field emission from Tanmi surface states in the one-dimensional model, Phys. Status Solidi 13,
K5-K8 (1966).
184 REFERENCES
341. L. M. Baskin, O. L. Lvov, and G. N. Fursey, On the theoi-y of filed emission from p-type semiconductors,
Phys. Status Solidi (a) 42, 757-767 (1977).
342. G. N. Fursey and L. M. Baskin, Features of field emission from semiconductors, Mikroelectronika 26(2),
117-122 (1997) (in Russian).
343. K. Yokoo, Functional field emission for highfrequencywave application, Proc. 12th Int. Microelectronics
Conf., Darmstadt, Germany, (July 6-9, 1999), 206-207.
344. K. Yokoo, New approach for high frequency electronics based on field emission array, Proc. 2nd Int.
Workshop Vacuum Microelectronics^ Wroclaw, Poland, (July 11-13, 1999), pp. 38-40.
345. V. G. Ivanov, V. S. Danilova, and O. L. Volosova, Oxygen adsorption on atomically clean surface of gallium
arsenide, Vestnik Novgorod University. Ser. Estestv. i Tekhn. Nauki No. 3, 8-12 (1996).
346. V. G. Ivanov, Phenomenon of intrinsic breakdown in the field emission p-type semiconductors, IX Int.
Vacuum Microelectronics Conf., St. Petersburg, Russia, pp. 148 (1996).
347. V. G. Ivanov, T. G. Rozova, G. N. Fursey, and T. P. Smimova, Field emissionfromgermanium during oxygen
adsorption, Fiz. Tverd. Tela 16,495-501 (1974).
348. T. G. Rozova, V. G. Ivanov, and G. N. Fursey, Influence of a strong electric field on the adsorption of oxygen
on germanium, Fiz. Tverd. Tela 17(1), 64-66 (1975).
349. V. G. Ivanov, T. P. Smimova, and G. N. Fursey, Threshold effect of oxygen electric adsorption on the
individual faces of single crystal germanium in a field-emission microscope, Poverkhnost, 12, 128-130
350. G. N. Fursey and N. V. Egorov, Stable semiconductorfield-emissioncathode, Zh. Tech. Fiz. 42(5), 1090-1092
(1972).
351. L. M. Baskin, N. M. Egorov, V. E. Pritsin, and G. N. Fursey, Effect of deep capture centers on the emissivity
of wide-gap semiconductorfield-emissioncathodes. Pis'ma Zh. Tech. Fiz. 5, 1345-1348 (1979).
352. L. M. Baskin and G. N. Fursey, Decisive role of the deep states in a vacuum breakdown initiated in presence
of dielectrical inseptions, Proc. 13th ISDEIV, Paris, Part 1, pp. 31-33 (1988).
353. L. Ernst, Adsorption of oxygen on atomically clean germanium surfaces examined in the field emission
microscope. Surf. Sci. 6,487-491 (1967).
354. V. G. Ivanov and Zabotin, A study of adsorption of oxygen on atomically clean germanium surfaces in field
emission microscope, XIV All-Union Conf. Field Emission Electronics. Tashkent, pp. 70-71 (1970).
355. R. Z. Bakhtizin and V. I. Stepanov, Influence of the adsorption of SiO molecules onfield-electronemission
from p-type Ge. Fiz. Tverd Tela 14(1), 294-296 (1972).
356. I. L. Sokol'skaya, V. G. Ivanov, and G. N. Fursey, Study of barium adsorption on germanium by field-emission
microscopy, Phys. Status Solidi 21, 789-795 (1967).
357. R. Z. Bakhtizin and I. L. Sokol'skaya, Titanium adsorption on the p-type Ge single crystals, Fiz. Tverd. Tela,
12(10), 2815-2820 (1970).
358. I. L. Sokol'skaya, N. V. Mileshkina, and R. Z. Bakhtizin, Gold adsorption on clean germaniumfieldemitters,
Phys. Status Solidi (a) 14(2), 417-^22 (1972).
359. N. V. Mileshkina and R. Z. Bakhtizin, Oxygen adsorbtion on the atomically clean surface of germanium
field emitter. Surf. Sci. 29, 644-652 (1972).
360. V. G. Ivanov, Experimental study of field emission from semiconductors. Doctor's dissertation. St.
Petersburg, State University, pp. 1-316 (1999).
361. J. T. Trujillo, A. G. Chakovskoi, and C. E. Hunt, Effect of vacuum conditions on low frequency noise in
silicon field emission devices. Proc. IXth Int. Vacuum Microelectronics Conference, St. Petersburg, pp.
133-137 (1996).
362. N. V. Egorov, G. N. Fursey, A., and V. Kotcheryzhenkov. Kinetics effects infieldemissionfromhigh-resistive
n-type Si. Fiz. Tverd Tela 15(3), 892-894 (1973).
363. P. G. Borzjak, V. F. Bibik, A. E. Kravzov, B. V. Stetsenko, A. F. Yatsenko, Kinetics offieldelectron emission
from hieghomic semiconductors. Report to the XIII All-Union Conference on Field Emission Electronics,
Moscow, p. 41,(1968).
364. I. L. Sokol'skaya and A. I. Klimin, Field-electron emission from cadmium sulphide and selenide. I Influ-
ence from an electric field and temperature, Vestnik Leningradskogo Universiteta: Serya Fiz. 4(1), 34-41
(1961).
365. B. V. Stetsenko, A. F. Yatsenko, and L. S. Miroshnichenko, Effect of surface states on transient processes in
sihcon field cathodes, Phys. Status Solidi (a) 1(2), 349-355 (1970).
366. G. N. Fursey and P G. Shlyakhtenko, Kinetics of p-type Ge filed emission, Fiz. Tverd. Tela 12(2), 645-647
(1970).
REFERENCES 185
367. G. N. Fursey, N. V. Egorov, and S. P. Manokhin, Transient effects associated withfieldemission of electrons
from silicon, Fiz. Tverd. Tela 14(6), 1686-1690 (1972).
368. S. A. Pshenichnyuk, Yu. M. Yumaguzin, and R. Z. Bakhtizin. Energy distribution of electrons emitted from
diamond like film in a strong electric field, Pis'ma Zh. Tekh. Fiz. 25(15), 46-52 (1999).
369. R. Z. Bakhtizin, Yu. M. Yumaguzin, and S. A. Pshenichnyuk, Field emission properties of diamond thin
films. 9th Int. Vacuum Microelectronics Conf., St. Petersburg, Russia, (July 7-12, 1996), Technical Digest,
p. 658.
370. C. Gazier, Multiple-electron events from field emission, Phys. Lett. A 35(4), 243-244 (1971).
371. M. Hermann, Untersuchung verschiedener Elektronenemission-sprozesse auf Mehrfachemission, Z. Phys.
184, 352-354 (1965).
372. G. N. Fursey, M. M. Mokhasne, N. V. Egorov, V. S. Ponomaryov, and V. N. Shchemelev, Statistics of
filed-electron emission, Fiz. Tverd. Tela 18(2), 631-632 (1976).
373. M. M. Mokhasne, V. S. Ponomaryov, N. V. Egorov, G. N. Fursey, and V. N. Schemelev, Procedure for
investigation of the statistics of field-electron emission, Pribory i Tekhnika Eksperimenta 19(2), 150-152
(1976).
374. N. P. Afanas'eva, N. V. Egorov, A. V. Kocheryzhenkov, and G. N. Fursey, A method of investigation of the
elementary acts number during thefiled-electronemission, Pribory i Tekhnika Experimenta 25(5), 141-142
(1982).
375. G. N. Fursey, N. V. Egorov, and A. V. Kocheryzhenkov, Field emission statistics for various faces of a
tungsten single crystal, Pis'ma Zh. Tekh. Fiz. 7(13), 798-801 (1981).
376. G. N. Fursey, A. V. Kocheryzhenkov, V. I. Maslov, A. L. Shmaev, and L. N. Borisov, Apparatus for study of
multi particle acts of field emission at low temperatures, Prib. Tekh. Ekspen 33(1), 149-151 (1990).
377. G. N. Fursey, V. I. Maslov, and A. V. Kocheryzhenkov, Search for correlated electrons infieldemission from
superconducting niobium, 34th Field Emission Symp., Osaka (July 13-18,1987), (Abstract) p. 67.
378. A. V. Kocheryzhenkov, V. I. Maslov, and G. N. Fursey, Statistics offield-electronemission from tungsten
and niobium at heUum temperatures, Fiz. Tverd. Tela 29(8), 2471-2472 (1987).
379. G. N. Fursey, A. V. Kocheryzhenkov, V. I. Maslov, and A. L. Shmaev, Royal Microscopical Soc. Proc, 36th
Int. Field Emission Symp., Oxford (1989), Vol. 24, art 4, p. 34.
380. V. I. Maslov, G. N. Fursey, and A. V. Kocheryzhenkov, Investigations of the quantity of elementary acts of
field emission with time resolution of 5^JLS-100 ps, COLLOQUE DE PHYSIQUE, CoUoque C8, supplement
au n° 11, Tome 50 (novembre 1989), C8-113-C8-116.
381. L. W. Swanson and L. C. Grouser, Angular confinement of field electron and ion emission, J. Appl. Phys.
40(12), 4741-4749 (1969).
382. L. W. Swanson, Comparative study of the zirconiated and built-up W thermal-field cathode, J. Vac. Sci.
Technol. 12(6), 1228-1233 (1975).
383. A. Tonomura, Electron holography: a new view of the microscopic, Phys. Today 22-29 (April 1990).
384. A. Tonomura, Electron-holographic interference microscopic. Adv. Phys. 41(1), 59-103 (1992).
385. I. L. Sokolskaya, Surface migration of tungsten atoms in electric field, Zh. Techn. Fiz. (USSR) 26(6),
1177-1184 (1956) (in Russian).
386. N. A. Gorbatyi and G. N. Shuppe, Dependence of absorption binding on the metallic crystal surface, Zh.
Tech. Fiz. 24(8), 1364-1375 (1955) (in Russian).
387. I. D. Ventova and G. N. Fursey, Surface self-diffusion in critical build-up, Zh. Tech. Fiz. (USSR) 43(11),
2432-2440 (1973) (in Russian).
388. I. D. Ventova, G. N. Fursey, and S. A. Polezhaev, Formation of microscopic protuberances at the vertex of
a sharp metal tip in a strong electric field. Critical Build-up. U., Zh. Tech. Phys. 47(4), 849-856 (1977) (in
Russian).
389. V. N. Shrednik, V. G. Pavlov, A. A. Rabinovich, and B. M. Shaikhin, Action of strong electric field and
heating on metallic points, Izv. Akad. Nauk SSSR: Ser. Fiz. (USSR) 38(2), 296-301 (1974) (in Russian).
390. V. G. Pavlov, A. A. Rabinovich, and V. N. Shrednik, Field errosion of Mo, Ta, Re, Fiz. Tverd. Tela (USSR)
18(2), 631-632 (1976) (in Russian).
391. V. M. Zhukov and S. A. Polezhaev, Surface evolution of microcrystal on the top of the emitter during the
thermo-field treatment, Zh. Tech. Fiz. (USSR) 57(6), 1133-1136 (1987) (in Russian).
392. M. Drechsler, Kristallstufen von 1 bis 1000 A, L. Electrochemie, 61(48), (1957).
393. I. A. Becker, The use of the field emission electron microscope in adsorption smdies of W on W and Ba on
W, Bell. Syst. Tech. J. 30,907-932 (1951).
186 REFERENCES
394. G. N. Fursey, Critical features offieldemission processes, IVInt. Vacuum Microelectronics Conf., Nagahama,
Japan (1991), pp. 1-2.
395. G. N. Fursey, B.N. Movchan, and V. A. Shvarkunov, Direct electron microscope observations of the formation
of microscopic protuberances on the close-packed faces of single-crystal tungsten in a strong electric field,
Pisma V Zhum. Tech. Fiz. (USSR) 16(20), 4 2 ^ 6 (1990) (in Russian).
396. V. E. Ptitsyn and G. N. Fursey, Effect of field emission on single-crystal point shapes, Izv. Akad. Nauk SSSR:
Ser. Fiz. (USSR) 52(8), 1513-1517 (1988) (in Russian).
397. Yu. A. Vlasov, O. L. Golubev, and V. N. Shrednik, Metal point shape change due to interaction between
electrostatic "capillary" forces, Izv. Akad. Nauk SSSR: Ser. Fiz. (USSR) 52(8), 1538-1543 (1988) (in
Russian).
398. V. G. Butenko, Yu. A. Vlasov, O. L. Golubev, and V. N. Shrednik, Point sources of electrons and ions using
microprotrusion on the top of a tip, Surf. Sci. 266, 165-169 (1991).
399. I. D. Ventova and G. N. Fursey, Build-up of the tip of a microcrystal into polyhedron I, Zh. Tech. Fiz. (USSR)
47(4), 844-848, (1977) (in Russian).
400. V. N. Shrednik, V. G. Pavlov, A. A. Rabinovich, and B. M. Shaikhin, Growth of tips in the directions normal
to close-packed faces by heating in the presence of an electric field, Phys. Status Solidi A, 23, 373-380
(1974).
401. M. Drechsler, Kristallstufen von 1 bis 1000 A. Herstellung der studen in feldemissiosmikroskopen durch
electrishe felder. Nessung der stufenhofen. Eine feldbidungsenergie, Z Electrochem. 61, 48-55 (1957).
402. V. M. Zhukov and S. A. Polezhaev, Surface evolution of point field-emission cathodes during pulse regime
of work, Rad. i Electn (USSR) 33(8), 1741-1747 (1988).
403. J. A. Becker, The use of the field emission electron microscope in adsorption studies of W on W and Ba on
W, Bell. Syst. Tech. J. 30, 907-932 (1951).
404. V. G. Pavlov, B. M. Shaikhin, and V. N. Shrednik, Study of the tips shape evolution in the process of self-
diffusion in strong electric field, I All-Union Conf. Field-Ion Microscopy, Kharkov (1976), pp. 77-80 (in
Russian).
405. D. M. Pautov and I. L. Sokolskaya, Reduction of divergence of an electron beam emitted by a field-electron
cathode, Zh. Tech. Fiz. (USSR) 41(9), 1999-2001 (1971) (in Russian).
406. G. N. Fursey, Destruction of a field emitter initiating a vacuum breakdown, Proc. 3th Int. Symp. Discharges
ElectK Insul. Vacuum, Paris (1968), pp. 113-118.
407. I. P. Zhizhin, G. N. Fursey, S. A. Shakirova, and A. V. Shishatskii, On the problem of obtaining small electron
probes by means of a field-electron emitters and magnetic optics, Izv. Akad. Nauk. SSSR Sen Fiz. 35(2),
302-306 (1971) (in Russian) [Bull. Acad. Sci. USSR—Phys. Ser, 35, 277-281, 1971].
408. G. N. Fursey, R. Z. Bakhtizin, V. E. Ptitsyn, Yu. M. Yumaguzin, and V. A. Shvarkunov, Emission charac-
teristics of field-emission cathodes with emission localized into a small solid angles, Izv. Akad. Nauk SSSR:
Ser Phys. (USSR) 52(7), 1250-1253 (1988) (in Russian).
409. L. W. Swanson and N. A. Martin, Field electron cathode stabihty studies: Zirconium/Tungsten thermal-field
cathode, J. Appl. Phys. 46(5), 2029-2050 (1974).
410. D. Tuggle, L. W. Swanson, and J. Orloff, Application of a thermal field emission source for high resolution,
high current e-beam microprobes, 7. Vac. Sci. Technol. 16(6), 1699-1703 (1979).
411. V. N. Shrednik and G. A. Odisharia, Anisotropy of heat of migration heats and adsorbent-adsorbent binding
energies in a metallic-film systems Zr-W, Hf-W, Hf-Mo, Zr-Nb, Zr-Ta, Izv. Akad. Nauk SSSR: Ser Fiz
(USSR) 33(3), 536-543, 1969 (in Russian).
412. H. Hosoki, T. Yamamoto, H. Todoroko, H. Kawase, and C. Hirai, Field emission cathode and method of
fabricating same. United States Patent No. 4379250 (Publ. April 5, 1983), Hitachi Ltd Tokyo, Japan.
413. V. N. Shrednik, Investigation of zirconium atomic layers on tungsten crystal faces by means of electron and
ion projectors, Fiz Tverd. Tela (USSR), 3(6), 1750-1761 (1961) (in Russian).
414. T. Inoue and M. Nakada, A novel zirconiating technique for high brightness Zr/W emitter, Sutf. Sci. 246,
87-93 (1991).
415. J. E. Wolfe, Electron-beam cathode having a uniform emission pattern, US Patent, 13.04.1982, 4324999
(Burroughs Corp. 30.04.1980, No. 145043).
416. G. N. Fursey, Field emission localization in narrow solid angles, 6th Int. Vacuum Microelectmics Conf,
Newport, USA (1993). Technical Digest pp. 30-31.
417. G. N. Fursey, D. A. Polezhaev, and V. A. Shvarkunov, Development of thermo-field cathode for Auger-
spectrometer (unpubhshed) (1990), pp. 1-42.
REFERENCES 187
418. L. W. Swanson, US Patent 3817592 (1974).

419. T. Inoue and M. Nakada, A novel zirconiating technique for high brightness Zr/W emitter, Surf. Sci. 246,
87-93 (1991).
420. J. Orioff, Thermal field emission for low voltage scanning electron microscopy, J. Microsc. 140(3), 303-311
(1985).
421. G. N. Fursey, Produce of a small emission segment during Zr layer of W forming in electric field (to be
published.)
422. D. Gabor, A new microscopic principle. Nature (London), 161,111-11% (1948).
423. Y. Nakaizumi, K. Kanda, T. Watanabe, M. Yamada, and T. Ohno, Development of Hitachi Model S-806
scanning electron microscope for wafer inspection, Hitachi Instrument News, No. 17, 20-25 (1985).
424. E. Zeitler and M. Ohtsuki, Young's experiment with electrons, Ultramicroscopy 2, 147-148 (1977).
425. C. Py and R. Baptist, Low-energy electrons interferences with Spindt-type microtips. Revue "Le Vide, Les
Couches Minces", Supplement au N 271, pp. 17-20 (1994).
426. V. T. Binh and N. Garcia, On the electron and metallic ion emission from nanotips fabricated by field-surface-
melting technique: experiments on W and Au tips, Ultramicroscopy 42-44, 80-90 (1992).
427. L. A. Chemozatonskii, Z. Ya. Kosakovskaya, Yu. V. Gulyaev, N. I. Sinitsyn, G. V. Torgashov, and Yu. F.
Zakharchenko. Influence of external factor on electron field emission from thin-film nanofilament carbon
structures, J. Vac. Sci. Technol. B 14(3), 2080-2082 (1996).
428. L Brodie and P R. Schwoebel, Vacuum microelectronic devices, Proc. IEEE 82(7), 1006-1034 (1994).
429. J. P. Calame, H. F. Gray, and J. L. Shaw, Analysis and design of microwave amplifiers employing field-emitter
arrays, J. Appl. Phys. 73(3), 1485-1504 (1992).
430. G. N. Fursey, Field Emission in Vacuum Microelectronics. The Proceedings for the 23rd and 24th ISTC
Japan Workshop on Advanced Nanotechnologies in Russia/CIS, November, 2002, Tokyo and 8 November,
Kyoto Japan pp. 63-106 {Appl. Surf. Sci. 215,113-134, 2003).
431. M. E. Crost, K. Shoulders, and M. E. Zinn, Thin electron tube with electron emitters at the intersection of
crossed conductors, US Patent 3500102 (1970).
432. R. N. Thomas, R. A. Wickstrom, D. K. Schroder, and H. C. Naihanson, Fabrication and some appHcations
of large-area silicon field emission arrays, Solid-State Electron. 17,155-163 (1974).
433. D. O. Smith, J. S. Judge, M. Trongello, and P. R. Thomton, Micro-structure field emission electron source,
US Patent 3970887 (1976).
434. C. A. Spindt, I. Brodie, L. Humphrey, and E. R. Westerberg, Physical properties of thin field emission
cathodes with molybdenum cones, J. Appl. Phys. 47, 5248-5263 (1976).
435. A. M. E. Hoeberechts, Field emission device, US Patent 4995133 (1978).
436. C. A. Spindt and W. Abenh, The volcano field ionization source. Proc. 27th Int. Field Emission Symp., edited
by Y Yashiro and N. Igata, (Dept. of Metallurgy and Materials Science, University of Tokyo, Tokyo, Japan)
(1977).
437. C. A. Spindt, C. E. Holland, and R. D. Stowell, Field emission cathode array development for high current
density applications, App/. Surf. Sci. 16, 268-276 (1983).
438. G. Bennig, H. Rohrer, C. Gerber, and E. Weibel, 7 x 7 reconstruction on Si(lll) resolved in real space,
Phys. Rev. Lett. 50, 120 (1983).
439. P. M. Lally, C. D. Mack, and C. A. Spindt, Experiments with a field emission cathode in TWT. Presented at
the Microwave Power Conf. Naval Postgraduate School, Monterey. CA, (1986).
440. H. F. Gray, G. J. Campisi, and R. F. Greene, A vacuum field effect transistor using siUcon field emission
arrays. lEDM Dig. Tech. Papers (Washington, DC), pp. 766-799, (1986).
441. C. E. Holland, C. A. Spindt, I. Brodie, J. Mooney, and E. R. Westerberg, Matrix addressed cathodoluminiscent
display. Int. Display Conf London, UK (1987).
442. V. I. Makhov, BaUistic filed emission devices, 2nd Int. Conf. Vacuum Microelectronics, Bath, UKIOP Conf.
Sen 99 (1989).
443. K. Betsui, Fabrication and operation of silicon micro-field-emitter arrays. Presented at the Japanese Physical
Society Meet (1990).
444. R. B. Marcus, K. Chin, D. Liu, W. J. Orvis, D. R. Diavlo, C. E. Hunt, and J. Trujillo, Formation of silicon
tips with <12nm radius, A/>/7/. Phys. Lett. 56, 236-238 (1990).
445. C. A. Spindt, C. E. Holland, A. Rosengreen, and I. Brodie, Field-emitter arrays for vacuum microelectronics,
IEEE Trans. Electron Devices 38(10), 2355 (1991).
446. N. L Sinitsyn, Yu. V Gulyaev, N. D. Devjatkov, et al.. Carbon nanoclusters structures as a one of materials
of emission electronics in the future, Radiotekhnika (in Russian), No. 2, 9 (2000).
188 REFERENCES
447. C. A. Spindt, Microfabricated field-emission and field-ionization sources, Surf. Sci. 266, 145, (1992).
448. C. A. Spindt. C. E. Holland, A. Rosengreen, and I. Brodie, Progess in field emitter array development for
high-frequency operation, lEDM Dig. Tech. Papers, Washington, DC, (1993), p. 749.
449. Yu. V. Gulyaev, I. S. Nefedov, N. I. Sinitsyn, et al.. Distributed microwave amplifer on field emitter arrays
with a nonhomogeneous energy collector, J. Vac. Sci. Tech. B 13(2), (1995).
450. C. A. Spindt, FEA's state of art and applications, 2nd Int. Workshop on Vacuum Microelectronics, Wroclaw,
Po/fln^, (July 11-13, 1999).
451. G. A. Dyuzhev, G. N. Fursey, A. V. Kotcheryzhenkov, D. V. Novikov, and V. M. Oichenko, Carbon clusters
produced by electric arc evaporation of graphite and their field emission characteristics 11th Int. Vacuum
Microelectronics Conference Proc. IVMC'98, The Grove Park Inn Asheville, NC, USA July 19-24, 1998.
pp.198-199.
452. G. A. Dyuzhev, G. N. Fursey, A. V. Kotcheryzhenkov, D. V. Novikov, V. M. Oichenko, and V. S. Boikov,
Structural and emission characteristics of field emitters based on carbon nanoclasters. 12th Int. Vacuum
Microelectronics Conf, Darmstadt, Germany. Proc. NMC'99 (July 6-9 1999), pp. 316-318; G. N. Fursey,
Explosive emission processes. 2 the Int. Workshop on Vacuum Microelectronics, Wroclaw^, Poland, Proc.
IWVM'99 (July 11-13, 1999), p. 44.
453. G. N. Fursey, D. V. Novikov, and V. M. Oichenko, Alternative field emission cathodes for multi-emitter
systems based on diamond-like films and tubelenes. 13th Int. Vacuum Microelectronics Conf, Guangzhou,
China. Proc. NMC'2000. (August 14-17, 2000).
454. V. I. Makhov, N. A. Duzhev, A. I. Kozlov, B. I. Gorfinkel, and E. N. Petrov, Rat-panel fluorescent display
with sihcon matrixed cold cathode, 1st Int. Conf Vacuum Microelectronics, Williamsburg, VA (1988).
455. M. E. Crost, K. Shoulders, and M. E. Zinn, Thin electron tube with electron emitters at the intersection of
crossed conductors, US Patent 3500102, (1970).
456. J. P. Siberian, A serial addressing method for field emission tip flat panel TV, 1st Int. Conf Vacuum
Microelectronics, Williamsburg, VA (1988).
457. R. Meyer, A. Ghis, P. Rambaud, and F. Muller, Microchip fluorescent display, Proc. Japan Display, (1985),
p. 513.
458. R. Meyer, A. Ghis, P. Rambaud, and F. Muller, Development of a matrix array of cathode emitters on a
glass substrate forflatpanel display applications, 1st Int. Conf Vacuum Microelectronics, Williamsburg, VA
(1988).
459. R. Meyer, Recent development on microtips display at LETI, 4th Int. Conf. Vacuum Microelectronics,
Nagahama, Japan (1991).
460. R. Meyer, Private communication (1993).
461. A. T. Rakhimov, V. A. Samorodov, E. S. Soldatov, et al.. Study of emission and structure characteristic
correlation of diamond films by scanning tunneling microscopy, Poverkhnost 7, 47, (1999) (in Russian).
462. L. M. Baskin, G. N. Fursey, V. I. Ivanov-Omskiy, et al.. The microgeometry and field emission properties
of diamond-like films. Int. Conf Phys. Dielectrics, St. Petersburg (1997) (in Russian).
463. N. I. Sinitsyn, Yu. V. Gulyaev, G. V. Torgashov, et al.. Thin films consisting of carbon nanotubes as a new
material for emission electronics, JAppl. Surf. Sci. I l l , 145, (1997).
464. G. G. Sominski and T. A. Tumareva, Field electron emitter with fuUerene coverages. 12th Int. Vacuum
Microelectronics Conf, Darmstadt, Germany. Proc. rVMC'99, (July 6-9, 1999), p. 304.
465. G. A. I. Amaratunga and S. R. Silva, Nitrogen containing hydrogenated amorphous carbon for thinfilmfield
emission cathodes, Ap/?/. Phys. Lett. 68(18), 2529-2531 (1996).
466. G. A. Dyuzhev, G. N. Fursey, A. V. Kotcheryzhenkov, et al.. Carbon clusters produced by electric arc
evaporation of graphite and their field emission characteristics. 11th Int. Vacuum Microelectronics Conf.
Proc. IVMC'98 The Grove Park Inn Asheville, NC, USA (July 19-24, 1998), pp. 198-199.
467. G. A. Dyuzhev, G. N. Fursey, A. V. Kotcheryzhenkov, et al.. Structural and emission characteristics of field
emitters based on carbon nanoclusters. 12th Int. Vacuum Microelectronics Conf Proc. IVMC'99, Darmstadt,
Germany (July 6-9, 1999), pp. 316-318.
468. J. Robertson M. W. Mechanism of electron field emission from diamond and diamond-like carbon, Proc. 11
A^MC(1998),p.l62.
469. O. M. Kuttel, O. Groning, Ch. Emenegger, E. Mailland, and L. Schlapbach, Electron field emission from
nanotube and other carbon containing films, Proc. 11 NMC (1998), p. 194.
470. G. N. Fursey and L. M. Baskin, Particularly of field emission from semiconductors, Mikroelektronika 26(2),
117-122 (1997) (in Russian).
REFERENCES 189
471. H. Kim, J. W. Huh, J. M. Kim, et al., Development of a diamond-like carbon based field emission display.
11th Int. Vacuum Microelectronics Conf. Proc. IVMC'98, The Grove Park Inn Asheville, NC, USA (July
19-24,1998), pp. 192-193.
472. W. B. Choi, D. S. Chung, S. S. Hong, et al.. Carbon nanotube based field emission displays. 12th Int. Vacuum
Microelectronics Conf., Darmstadt, Germany Proceeding IVMC'99 (July 6-9, 1999), pp. 310-311.
473. W. Stocker, H. W. Fink, and R. Morin, Low-energy electron and ion projection microscopy, Ultramicroscopy
31,379-384(1989).
474. R. Morin, A. Gargani, and F. A Bel simple UHV electron projection microscopy, Microsc, Microanal,
Microstruct. 1, 289-297 (1990).
475. R. Morin and A. Gargani, Ultra-low-energy-electron projection holograms, Phys. Rev. B 48, 6643-6645
(1993).
476. E. S. Snow and P. M. Campbell, Fabrication of Si nanostnictures with an atomic force microscope, Appl.
Phys. Lett. 64(15), 1932 (1994).
477. Kazuhiko Matsumoto, Masami Ishii and Kazuhito Segawa, J. Vac. Sci. Technol. B 14(2), 1331 (1996).
478. K. Tjaden, Chemical mechanical poHshing in FED manufacturing. Chemical Mechanical Polishing
Workshop, ICMCTF (1996).
479. E. I. Givargizov, Field emitters based on diamond-covered silicon tips, Microelectronika 26(2), 102-106
480. E. I. Givargizov and A. T. Rachimov, Field emission cathodes on nanocrystal and nanodiamond films (physics,
technology, application), Uspekhi Fiz. Nauk 170(9), 996-999 (2000) (in Russian).
481. J. Lee and J. Nam, The technologies of FED devices, 2nd Int. Workshop Vacuum Microelectronics, Wroclaw,
Poland(My 11-13, 1999), pp. 2 ^ 3 1 .
482. S. Ahn, N. Lee, H. Jeonge, et al.. Development of 3.5-inch full-color field emission display panels in the
low and high voltage packaging, 12th IVMC, Darmstadt (1999), pp. 28-29.
483. R. Meyer and R. Baptist, Field emission display: the ultimate thin film flat CRT technology. Electronic
Display 94, Germany, (1994).
484. A. R Burden, H. E. Bishop, M. Brieriey, et al.. Field emitting inks for consumer-priced broad-area flat-panel
display, 12th IVMC, Darmstadt, (1999), pp. 62-63.
485. G. A. Mesyats and G. N. Fursey, Explosive electron emission of initial stages of vacuum discharges,
Kholodnie Katodi (Radio, Moskva, 1974), pp. 269-292 (in Russian).
486. G. N. Fursey, Field emission and vacuum breakdown, IEEE Trans. Electrical Insulations EI-20(4), 659-670
(1985).
487. G. A. Mesyats, Ectons in a vacuum discharge: breakdown, the spark, and the arc, - M.: Nauka, (2000),
p. 424.
488. G. N. Fursey and L. A. Shirochin, Explosive emission processes in high electric fields, Uzbek J. Phys. 2(1),
53-62 (2000).
489. M. Benjamin and R. O. Jenkins, The distribution of autoelectronic emission from single crystal metal points,
Proc. Royal Soc.(A} 176, 262 (1940).
490. F. J. Blatt, Physics of Electronic Conduction in Solids (McGraw-Hill Book Company, 1968).
491. E. W. Muller, Weitere Beobachtungen mit dem Feldelectronenmikroskop, Z. Phys. 108, 668 (1938).
492. E. W. Muller, Das Feldionenmikroskop, Z. Phys. 131, 136-142 (1951/52).
493. M. J. Troyon, Energy spread of different electron-beams. 2. Field-emission electron-beams, /. Microscopie
Spectroscopie Electroniques 13(1), 49-64 (1988).
494. B. F. Coll, Present status and future outlook of microtip and CNT based field emission displays. Int. Topical
Meet. Field Electron Emission from Carbon Materials, Moscow July 2-4, (2001).
495. A. P. Barsukov, Displays, casting engineering. Film and TV Engineering No. 11,53-62 (2000) (in Russian).
496. A. Pur, The future of imaging techniques, PC Magazine, 76 (March 25, 1997).
497. D. J. Rose, On the magnification and resolution of thefieldelectron microscope, J. Appl. Phys 27(3), 215-226
(1956).
498. K. Yokoo and H. Ishizuka, RF application using field emitter arrays, 9th Int. VMC, St. Petersburg (1996),
pp. 490-500.
499. B. F. Bondarenko, Yu. L. Ruhbakov, E. G. Sheshin, and A. A. Shuyka, Field emission cathodes and devices
using them. Review of Electronic Engineering, 4,4(814), 2-58 (1981) (in Russian).
500. H. F. Gray, All-silicon gridded FEAs, Presented at 1980 Tri-Services Cathode Workshop, GreffisAFB, Rome,
No. 4 (April 1980).
190 REFERENCES
501. I. K. Yokoo, Vacuum microelectronics, J. Vac. Soc. Japan 32(2), 63-67 (1989).
502. J. K. Cochran, A. T. Chapman, R. K. Feeney, and D. N. Hill, Review of field emitter array cathodes. Int.
Electron Devices Meet. (1980), pp. 462-^66.
503. R V. Philips, R. E. Neidert, C. Hor, C. A. Spindt, Technical Digest ofIVMC-93 (1993), p. 155.
504. A. B. Gurinovich, A. A. Kuraev and A. K. Sinitsyn, The studies of TVT with cathode modulation optimized
on power efficiency, Radiotechnika and Electronika 45(2), 1493 -1498 (2000) (in Russian).
505. I. Brodie, Keynote Address to the First IVMC, June 1988: Pathways to Vacuum Microelectronics, IEEE
Trans. Electron Devices 36(11), 2637-2640 (1989).
506. H. M. Bizek, et al., Gigatron, Trans. Plasma Sci. 16(2), 258-263 (1988).
507. V. A. Isaev, D. V. Sokolov, and D. I. Trubezkov, Lecture on microwave devices with emission modulation,
8th Winter School/Workshop of Engineers (1989), Book 2, pp. 3-36 (in Russian).
508. A. V. Goldetsky, The devices of vacuum super high frequency microelectronics, Microelectronika 26(2),
123-129 (1997) (in Russian).
509. N. I. Tatarenko and A. S. Petrov, Vacuum microelectronics: reality and prospects. Microelectronics
worldwide. Advances the Contemporary Radioelectronics No. 7, pp. 10-31 (1998) (in Russian).
510. E. S. Snow and P. M. Campbell, Fabrication of Si nanostructures with an atomic force microscope, Appl.
Phys. Lett. 64(15), 1932-1934 (1994).
511. Kazuhiko Matsumoto, Masami Ishii and Kazuhito Segawa, J. Vac. Sci. Tech. B, 14(2), 1331 (1996).
512. Y. Sugiyama, Recent progress on magnetic sensors with nanostructures and applications, J. Vac. Sci. Tech.
513,1075-1083(1995).
513. K. Uemura, S. Kanemaru, and J. Itoh, Fabrication of a vacuum-sealed magnetic sensor with a Si field emitter
tip, J. Micromech, Microeng., 11, 81-83 (2001).
514. J. M. Orloff and L. M. Swanson, Study of field-ionization source for microprobe applications, J. Vac. Sci.
Technol. 12(6), 1209-1213 (1975).
515. M. Faubel, W. M. Hebber, and J. R. Toennies, Low background cold cathode ion source for molecular beam
detection, Rev. Sci. Instr 49(4), 449^51 (1978).
516. B. S. Satyanarayana, A. Hart, W. L Milne, and J. Robertson, Field emission from tetrahedral amorphous
carbon, Ap/7/. Phys. Lett. 71, 1430-1432 (1997).
517. D. Hong and M. Aslam, Field emission from p-type polycrystalline diamond films, J. Vac. Sci. Technol. B
13(3), 427-430, (1995).
518. L. M. Baskin, Dynamic of field emission processes in static and super high frequency fields. Thesis of
doctor's dissert., Tomsk, 1990 (in Russian).
519. L. M. Baskin and G. N. Fursey, Decisive role of deep trap states in initiating vacuum breakdown in presence
of dielectric insertions, Proc. 13th ISDEIV, Paris (1998), Part 1, pp. 31-33.
520. Yu. Zakharchenko, L S. Nefedov, G. V Torgashov, N. I. Sinitsyn, and Yu. V. Gulyaev, Calculation of the
electrophysical parameters of electron gas in a carbon nanotube using the quantum well method, 12 IVMC,
(7), Darmstadt, pp. 88-89 (1999).
521. J.-M. Bonard, Field emission from carbon nanotubes: perspectives for apphcations and clues to the emission
mechanism, 12 TVMC, (7), Darmstadt, pp. 300-301 (1999).
522. Y. Chen, D. T. Shaw, and L. Quo, Field emission of different oriented carbon nanotubes, Appl. Phys. Lett.,
76(17), 2468 (2000).
523. N. I. Sinitsyn, G. V. Torgashov, Yu. V. Gulyaev, et al.. Electron field emission from nanocluster carbon-metal
films, 12 IVMC, (7), Darmstadt, 1999, p. 254.
524. W. B. Choi, N. S. Lee, W. K. Yi, et al.. The First 9-inch carbon-nanotube based field-emission display for
large area and color applications, Digest SID 00, pp. 324-327 (2000).
525. I. Brodie, Physical considerations in vacuum microelecronics devices, IEEE Trans. Electron Devices 36(11),
pp. 2641-2644, 1989.
526. D. Nicolaescu, Modehng of the field emitter triode (FET) as a displacement/pressure sensor, 41st IFES,
1994, Rouen, France (publ. In Appl. Surf. Sci. 87/88, 1995, p. 61).
527. J. C. Jiang, P. K. Allen, and R. C. White, Microvacuum diode sensor for measuring displacement and loads.
In 4th Int. Conf. On Vacuum Microelectronics (Nagahama, Japan, 1991).
528. H. C. Lee and R. S. Huang, A novel field emission array pressure sensor. In 6th Int. Conf. On Solid State
Sensors and Actuators (San Francisco, CA, 1991), p. 110.
529. H. H. Busta, J. E. Pogemiller, and B. J. Zimmerman, The field emitter triode as a pressure/displacement
sensor. In 6th Int. Conf On Vacuum Microelectronics (Newport, RI, 1993).
REFERENCES 191
530. G. N. Fursey, A. V. Kotcheryzhenkov, D. V. Novikov, and V. M. Oichenko, Field emission from diamond-like
films and tubelenes, The International Topical Meeting on Field Electron Emission from Carbon Materials,
2-4 July 2001, Moscow, pp. 50-51
531. J. M. Kim, N. S. Lee, J. H. You, J. E. Jung, C. G. Lee, J. W. Kim, K. S. Choi, et al.. Development of carbon
nanotube field emission displays with triode snictures, The International Topical Meeting on Field Electron
Emission from Carbon Materials, 2-4 July 2001, Moscow, p. 29
532. G. N. Fursey, Field Emission in Vacuum Microelectronics, Appl. Surf. Sci. 215, 113-134 (2003).
533. G. N. Fursey, D. V. Novicov, G. A. Dyuzhev, A. V. Kotcheryzhenkov, and P. O. Vassiliev, The field emission
from carbon nanotubes. A/?/?/. Surf. Sci. 215, 135-140 (2003).
LIST OF MAIN NOTATION
A (Je, 7i) function describing nqjlocalized transfer of energy from the

electron subsystem to Ifjf lattice
b(y) "effective" total density of mobile charges
B current density per unit solid angle, or brightness limited by
the Langmiur equatioi to a value i^max
c velocity of light
c(T) heat capacity
C"*", C, (d*^, d) electron generation and annihilation operators
di length of the most probable tunneling path
D(Ex, F) barrier transparency
ei ion's charge
Eo energy of the localized state
£a energy of the lowest state in the metal
Efi, /x Fermi energy
E(k), K(k) complete elliptic integrals of thefirstand second kinds
Eta maximum energy at the normal to the direction of emission
E{y) field strength inside the sample
Ex = pl/2m part of the electron kinetic energy carried by the momentum
component px normal to the surface
g(T) a temperature-dependent Tompson coefficient^"^^
G(r, t) volume density of heat evolution rate determined by the
Joule and Tompson effects
f(Ex, T) Fermi function
F applied electric field
HL unperturbed cathode Hamiltonians
^m barrier height
HR unperturbed anode Hamiltonians
Hj tunneling Hamiltonian
i, / emission current
193
/couple current of paired electrons

j, / field emission (FE) current density.
70 emission current density as defined by the
Fowler-Nordheim equation (Eq. 1.8)
jem emission current density from the conduction band
j (Fs, 7s) emission current density depending on the
emission surface temperature 7$ and the electric
field strength Fs at the surface
Jtr Jakobian of the transition from coordinates
k Boltzmann constant
L length of the tip
m mass of electron
mi ion's mass
m* = mx/m, m* = rriy/m, ratios of the effective mass components to the free
m* = niz/m electron mass.
M = (P/FQU upon a dimensionless parameter, a—the
micro-roughness height
n unit vector normal to the emitting surface
Hi electron concentration in a pure semiconductor
noo current density in the sample bulk
Nc density of states at the bottom of a conduction band
NL = J2k^t^k number of particles operator
p electron momentum
PF momentum of an electron at the Fermi surface
Px/^x = ?^
pj/my = ^2
PJ/rriz = X^
P {£)ds total energy distribution of emitted electrons
relative to the Fermi level.
r radius of the tip endcap
ra anode radius
rk cathode radius
R STM tip of radius
Rj principal radii of curvature of the surface under
investigation at their point of intersection with the
most probable tunneling path
T = T(r, t) emitter temperature
r * = d/2K inversion temperature
= 5.67-lO-VV0-Ky)^
TQ.CQ, Xe electron temperature, electron heat capacity, and
thermal conductivity
To temperature of the tip, far from the apex
Ts a temperature of the emitter apex where r = TQ
Uo potential in a spherical diode in the absence of
space charge
LIST OF MAIN NOTATION 195
U{x) potential function

U{x), ^{x) potential energies of the bottom of the conduction
band and of the electrochemical potential level
V, k velocity of incident particle and its wave vector,
respectively
distance from the interface of the potential maximum
Ti and r2 localized state decay widths in the STM tip
a nonlinear potential function depending on the shape
of the micro-roughness
% polar angle iny, z plane
<t>(r, 0 electric potential
Fermi integral of the order 1/2 defined in the usual
way
function related to ^(7)^"*^
geometric factor
Y surface tension of the emitter tip material
& substance density
spectral function of the electron gas in the metal
S{k)
^p phase shift between the incident and reflected waves
angle in the plane perpendicular to the magnetic field
n plane parallel to the field
work function
V surface emissivity
density of states with the momentum normal to the
PL surface barrier
p resistivity constant
piT) resistivity as a function temperature
PO space charge density
k{T) thermal conductivity
a Stephan-Boltzmann constant
tee time of the electron-electron interaction
carrier mobility
MO carrier mobility infieldEQ
Aep free path length of electron-phonon interaction
^Wigner
= 2R/v-\-(2/vd/dk)[S(k)] time interval between when the centers of mass of the
incident and scattered wave packets crossed the
surface of a sphere of radius R
"^tun tunneling time is assumed to be the mean time of
interaction between a tunneling electron and the
potential barrier
^Hartman time was calculated as the time interval between when
the maximum in the wave packet passes through the
left side of barrier and the formation of the maximum
in the transmitted wave packet at the right side of the
boundary (used by Hartman'^^'^)
Tep characteristic time of the electron-phonon interaction

1^ (jc) degeneration parameter
i^s degeneration parameter near the surface
1^00 value ?>(jc) in the sample bulk
i^ (y), 0 (Y) Nordheim function
Qj a matrix representing the curvature of surface apex of
the S T M probe at the point of its intersection with the
most probable tunneling path
Eg forbidden band width (for emission from the valence
band)
8a acceptor level
ae high-frequency dielectric constant
62, and 6d energies of the acceptor and donor levels
^ = êm — ^cd difference in average energies between the field
emitted electrons (£em) and the conduction band d
electrons (Ecd)
6t activation energy of traps
COL Larmor precession frequency
I;F velocity of electrons on the Fermi surface
X electron affinity
Xe slopes depend on the electron affinity
XT thermal conductivity coefficient
INDEX
Acceptor impurities 77 "Atomically sharp" edges on the W 145

"accumulation" of surface micro-roughness 48 atomically sharp emitters 6, XII
activation energy [migration a.e.] 47, 49, 95, 96 atomically sharp microtips 6
Activation energy of traps 95 Atomically sized object(s) 131
activation energy 47,49, 95, 96 atomically smooth emitters 6
Active element(s) 137,138 attenuator 61
adherent coating 73 Auger spectrometers 115
adiabatic condition 58, 60 Auger-electron spectrometers 136
adiabatic conversions 57, 58, 60 "Auto-epitaxial" 119
Adsorbates (Al,Cr,Ge,Mb,Ti,Si,Zr,Hf) on W Avalanche breakdown 96
and Mo 125 average current 60, 61, 62, 128, 155, 168
adsorbed layers 37
Adsorbed Zr island(s) 124 Backward wave tubes (BWT) 157
Adsorption of zircomium (Zr) on W 54, 128 Band bending changes 83
Allotropic forms of carbon 161,162 band bending 72, 83, 84
Amorphous carbon 163 Bardeen's potential 27, 28
Amplifiers 69, 138, 139, 153, 154, 155, 157, barrier width 7,9,10,11,27
158, 169 Base electrode 146, 154
Amray 117,133,135 Baz's method 59
Anisotropic adsorption of oxygen 101 beam-bunching injectors 69
Annealed emitter tip 124 Bendgap semiconductors 163
Aperture 131 Bias voltage 98
applied voltage 21, 23, 24,48,49, 78, 79, 82, Bi-Sn-Pb alloy 67, 68
96, 128, 163 blade cathode(s) 24
Array cathode 146, 148 Bloch function formalism 12
array of field emitters 69 blunt emitter 30
Arrhenius equation 38 Boguslavskii-Langmuire law 21
atomic scale surface smoothness 6 bombardment of emitter surface 57
atomic sharpness 6, 116 breakdown event 67
Atomic sharpness 6, 116 breakdown, vacuum breakdown 32, 33, 39,
Atomically clean surface 107, 109 47,64
atomically clean 33,49, 72, 98, 101, 103, 107, bright ring 33
109,110,144 Broad-band 96
"Atomically sharp" edges on the Mo 145 Broglie relation its wavelength 142
197
198 INDEX
Build-up (effects) 48 Crystal grain 116

build-up process 55, 118, 123, 128, 131 cry stallographic orientation 116
Built-up emitter tips 123 crystallographic orientations 6,116
Built-up ZrAV 130 Current converters 138
bunches 57 Current kinetics 98
current of a beam collector 61
cut-off potential 66
capillary emitters 67
Capture centers 92
Carbides 157 dc mode 50, 53, 54, 56
Carbon fibers 135, 162 Debye temperature 50
Carbon Nanoclusters 161, 162, 163, 164, 167 deep cooHng 56
carbon nanotubes 137, 161, 163, 165, 166, Deep trap 92,96,153
167, 168 Degeneration criterion 88
Carbon substrate 134 dence plasma 32
carrier concentration 71, 72, 81, 84, 88, 89, 91, Density of pixels 146
92,98 Depletion region 88, 89, 92, 97
Carrier generation 81, 89, 90, 98 desorption in a strong electric field 72
"Cathode deposit" 162, 166, 167 desorption processes 72
cathode instabilities 36 destruction of the semiconductor tip 73
Cathode-ray tube (CRT),(CRTs) 148, 150, deviation from FN theory 9, 29
151, 152, deviation from linearity of the current-voltage
cavity 60,61,66 characteristics 29
Ceramic(s) 111, 112,113,114 deviation-point current 23
CHstrode 155 deviations from FN theory 20
Close-packed crystal facets 119 Diamond-like films 97, 161, 163, 164, 166,
Close-packed layer(s) 125, 126 167, VI, XII
Cluster(s) of atoms 122 Diamond-like structures 140
Coherency 117 Diffusion current 89,90
collector 36, 37, 53, 61, 65, 114, 138, 155, 156, Diffusion-drift equilibrium 98
159,160 directional coupler 61
Color rendition 150 dispersion law 12, 20
conmiutator 61 Display lifetime 148
conditioned electrodes 32 Display panel(s) 138, 139, 144, 149, 152, 167
conducting liquid surface 68 dissipation 39,58,62, 137
conduction band 31, 32, 39, 40, 50, 75, 78, 81, "Double" localization 131
82, 83, 85, 88, 96, 97, 154, 163, XI Dwell time 148
cone-shaped emitter 52 Dyke's Experiments 33
Contaminant(s) 77
contamination removal 73
Contrast 48,73, 75, 125, 138, 151, 152 Effective masses of electrons and holes 86
converters 69, 137, 138 Efficiency of phosphors 150
Correlated electrons 105,107 "ejection effect" 64
Correlation length 14, 111 Electron gun 117
Coverages of Zr 125 Electric arc vaporization o graphite 161
Crests 69,117,119 electrical explosion 32, 34
Critical temperature 47, 111, 113 Electroadsorption 101, 102
CRT-monitors 152 electrochemical etching 67, 111
Cryostat 53,110,111 Electrochemical potential 83, 84, 85
Crystal facets 115, 119 electrohydrodynamic instability 67
INDEX 199
Electrolytic etching 77 energy conversion devices 57

electron affinity 71, 81, 82, 83, 85, 139, 140, energy distribution 8, 9, 11, 12, 13, 40, 41, 57,
161, 163 64, 65, 66, 69, 117, 155, 164, VI, IX
Electron beam lithography 139,158 energy spectrum 4, 6, 12, XI
electron energy spectra 64 evaporation energy 71
Electron energy spread 115 Evaporation field 74, 102
electron holography 6, 20, 35, 39,40, 139, exchange process (energy exchange
V,XI mechanism) 39
Electron lithography machines 115 excitation device 61
electron lithography 115, 138, 140, 158, 161 exitation 33, 60, 61, 67, 68, 106, 160, 161, IX
electron microscopy 57, 101, 121, 122, 134, experimental chamber 60
135, 136, 139, 140, 142, 164, 166, 167, "explosion" of the emitter tip 72
169, V, XI explosive breakdown 34
electron momentum 31, 143 explosive emission 39, 67, XI
Electron optical systems 57, 69, 136 explosive-like evaporation 72
electron pair tunneling 16 Extracting electrode 146
Electron pairs in FE 105 extremely high electric field 41, VI
Electron sources 11, 69, 115, 117, 152
electron spectroscopy 6, 64
electron subsystem 51 facilitating addressing 148
electron tunneling time 15 Fast-acting high frequency devices 137
electron-electron interaction 50,51 fatigue phenomena 72
Electron-electron interactions 105 FE display panels 144
electron-electron scattering 15 FE electron sources 115, 117
electron-electron tunneling time 15 (FE) Field Emission from semiconductors 71,
electron-lattice interaction 51 73, 75, 77, 79, 81, 83, 85, 87, 89, 91, 93,
electron-phonon free path 50, 54 95,97,99, 101, 103, Vn
electron-phonon interaction 50, 62 FE image for nanotubes 168
electrons in pairs 15 FE localization 116, 119, 124, 130
Electrostatic lens 140 FE matrix 138
electrostatic repulsion 25 FE micrographs 120
"Elementary emission events" 105, 114 FE microscope 38, 78, 122, 140, 142, 143,
elliptic Fermi surfaces 32 144, 167
elliptic integrals 3 FE microscopy 33, 60, 101, 119, 138, 141, 143
Emission image 34, 37, 64, 71,78, 79, 80, 81, FE multi-needle cathode 161
82, 142 FE Miller microscope 48, 53
emission images 33, 53, 78, 101, 104, 112, FE pattem(s) 35,48, 73, 74, 102, 108, 111
119, VI FE projector 111, 113
Emission localization to a small sohd angle FE statistics 107, 108, 109, 110, 111, 112,
125, 126 113,114
Emission locaUzation to small solid angles 115 FEtriode 139,153,154
Emission spectra statistics 106 FEAs 138, 139, 140, 149, 151, 152, 153, 156,
emission-included migration 49 157, 158, 160, 166, 169
emitter bulk 43,46,49, 71, 72, 95, 98 Feinman's procedure 59
emitter cone angle 52 Femitron 155
emitter destruction 33, 39, 43, 72, 89 Fermi energy 4, 14, 32,40,41
emitter faces 6 Fermi function 4, 31
emitter instability 72 Fermi integral 86, 88
emitter shape 78, 119 Fermi level 2, 4, 10, 15, 26, 31, 39, 40, 50, 82,
Emitting area 4, 12, 53, 78, IX 85, 87, 88, 89, 154, 163
200 INDEX
Fermi surface structure 12, XI Full color displays 150,151

Fermi surface 12, 31, 32, 51, XI Fullerene(s) 137, 140,161, 162, 163, 164,165,
ferrite filter 61 166, VI
Fiber-optic bundle 111
field (electron) emission from liquid metals 67
Gas lasers 160
field desorption cleaning 73
Gas-filled proportional counter 106
field desorption 72, 73, 74, 100, 101,103, 160
Gate electrode 146,153,154,158
field electron image 60
Gate-to-cathode capacitance 148
(field emission) FE cathode 24, 32, 36, 47, 61,
Ge, Si, GaAs (atomically clean surface) 74,144
64,116, 131,134,136,137,139,148,149,
gedanken experiment 59
150,154, 158,159,168
Generation rate 81,82,95,98
Field Emission (FE) pattern, images 74, 130,
Generators for frequencies 139
164, 165, 167, VI
Geometric factor 8, 79
Field emission arrays (FEAs), FEA 153, 156
geometric quotient 4
field emission current stability 65
Glass faceplate 146
Field emission displays 144, 146, 150, 169
Graphite fibers 159
Field emission localization by a local work
Green's function method 14
function decrease 124
Grid electrode 154
Field emission statistics from metals 108
grids 61,157
Field emitter arrays 146,149,158
Field emitter explosion 32
Field enhancement 67,116, 118,125,163,164 half width of energy distribution 65
Field evaporation 102, 119, 121, 131 Hamiltonian anode 13
Field geometric factor 79 Hamiltonian cathode 13
Field ion imaging 74, 119 Hamiltonian complete 13
Field ion micrograph(s) 75, 123, 124 Hamiltonian tunneling 13
Field ion microscopy 75, 103, 119, 125, 142 Hartree-Fock approximation 27
Field ion picture 141 heat capacity 45, 51
field outgrowths 121 heat localization 47
field penetration 72, 80, 81, 85, 92, 94, 105 Heat resistance of the emitters 137
field stimulated reactions 72 Heterodynes 157
"Field-surface melting" 131 High beam coherence 115
Film of dielectric 139 high brightness 115,131,138,139,148,151,
Flat crystal plane 116 158, 167,168, XI
Flat-panel display 146, 152 high current densities 6,16, 19, 24, 30, 31, 32,
FN coordinates 3, 9, 72, 75, 166 34,39,47,49,56, 114, 131, X, XI
FN formula 3,12,23 High energy tail 105
FN theory 1, 6, 8, 9, 12, 13, 17, 19, 20, 26, 29, high energy tails 15, XI
58, 60, 69 High frequncies 137
Formula of Boguslavskii-Langmuire 20 high power 57,69, 137, 140, 152
Fowlel-Nordheim (FN) coordinates 3, 9, 72, High radiation tolerance emitters 137
75, 166 High resolution television 152
Frame rate 148 High-brightness field emission 115
free carriers 72 High-definition television 148, 150
Free electron model 1,6,12, 31 high-energy electron devices 69
Free-electron lasers 152 High-power electron beams 152
freezing the micro-protrusions 68 high-power traveling wave tubes 69
Frequency converters 137 High-resolution electron probe devices 115
Frequency multipliers 69,138 Hitachi 117
frequency transformation 57 Hole energies 85
INDEX 201
Holograms 135 Larmor precession frequency 59

Honeycomb anode 146 LC-indicators 151
"Honeycomb" control electrod (gate) 138 LC-materials 151
Horisontal microtriodes (vacuum Lead 28,43, 64, 71, 72, 125, 127, 163, 164
transistors) 153 Lifetime 69, 82, 125, 148
"hot" hole 15 Light probe 92, 98
Hydrogen field ion pattern 126 limiting current densities 50, 89
Hysteresis 96, 97, 163 liquid helium temperature 52, 154
liquid metals 61, 67, 69
Liquid nitrogen temperature 155
image potential 9
liquid surface(s), metals 67, 68
Image shrinkage 78, 79
Image stability 152 liquid-metal cathodes 67
initial properties 72, 77, 78, 104 liquid-metal ion sources (LMIS) 67
initiation phase of breakdown 32 LMIS techniques 67
injecting carriers 71 Local radii of curvature 116
input power 60, 61, 65, 67 localized emission 54, 122, 128, 135, X
Integrated circuits 138, 139, 140, 152 localized Green functions 7
interelectrode capacitance 60 long current pulses 33
Interference fringes 131 Lorentz force 25, 158
Interferogram(s) 135 Low electron affinity 139,140,161
intrinsic magnetic field 25 low energy tails 6, 15, XI
Intrinsic noise (level) 155, 156 Low index planes 117
inversion temperature 39,41,43,47 Low voltage displays 150, VI
ion bombardment of emitter 57 low-energy electrons 11,66,136
ion emission microscopy 71 low-energy tails 6, 11, XI
ion emission 67, 71 Low-voltage FED 150
Ion microscope 121,125,142 Luminance 148, 150, 151
ion motion 57, 63
lonpattem(s) 74, 126
Magnetic field gauges 138
ion trajectories 63
Magnetron and laser sputtering of graphite 161
Ionic image 142
magnetron(s) 57, 69, 161
Ionization potential 71,163, 164
Magnification 38,71,78, 133, 136, 140, 141,
Iridium (Ir) FE cathode 159
142, DC
irreversible changes 64, 72
manybody problem 27
many-particle effects 13, 15, 108
Josephson tunneling 13 Mass-spectrometers 140
joule heating 36, 38, 39,43,45,49, 50, X Matrix field emission cathodes 148
maximum brightness 131, 133
maximum current densities 38,49, 62, 125,
Ka-Na alloy 67
135, X
Kernel 46
maximum electron energy at the resonator
Kinks 82,131
output 65
klystron(s) 69
Memory cells 152
metal FE cathodes 159
Langmuir equation 115 Metal gate film 147
Langmuir limit 115 metal-vacuum boundary 19, 57, 59
large micro-protrusions 119 Micro-channel plate intensifier 74
Large-area (LCD) displays 150 microcrystal field emitter 71
Large-pixel-count displays 148 Micrometer size ...dimensions 137
202 INDEX
Micro-protrusion growth 118,119 Niobium (Nb) 12, 35, 108, 123, 158
Micro-ridges 117,123,124 Nokia 151
microtip 8, 32, 63, 130 Nonlinear current-voltage characteristics 75,
Microtransistor 138, 139 76,81,83,97
microwave apparatus 61 nonrelativistic approximation 22
Microwave devices 69, 138, 155,156, 157, "Nonzero current" 88, 92
169, XI Nordheim function 2, 3, 4, 80, 85, 143
microwave field power conversion 57 Nottingham heating 39, 43, 50
microwave power amplification 60 Nottingham mechanism 50
Microwave transit-time devices 157 Nucleation 122
migration process 37 numerical calculation 5, 8, 10
Miller indices 142
Mo 19, 30,43, 48, 50, 108, 109, 125, 131, 145,
one dimensional (problem) 1, 6, 30
157, 165
one-particle approximation 6, XI
Mobility 90, 91, 93, 94, 128, 164
operational lifetime 69
Modulation voltages 148
Optimum crystallographic orientation 116
modulator 61
Ordered array (of metal or semiconductor
Monochrome CRT displays 150
tip) 146
Monochrome display, panels 150, 168
outgassing 73
MuUer microscope 48, 53 Outgrowths 119, 121, 131
multidimensional tunneling of electrons 7 overheated core 46
multidimensional 7, 8, 27 oxygen 67, 101, 103, 128, 129, 149, 164
Multielectron field emission from high
temperature superconducting
ceramics HI Paince-Bohm's formahsm 28
multielectron theory 13 pair electrons 16,105,108
Multi-emitter matrix 149 Pair tunneling 16, 17, 105, 108
Multipeak spectra 108 paired electrons 16, 17
Multiple-beam electron lithography 140 paraxial approximation 65
Multiple-beam electron sources 152 Paraxial electron beam 124
Multiple-beam lithography 158 Particle accelerators 152
Multiple-particle tunneling 114 phase [pre-breakdown phenomena] 32, 35, 36,
Multiple-wall carbon nanotubes 166 37, 38, 58, 65, 131, 138, 140, 143, 160, IX
phase shift 58
Phosphor anode 126
nanoelectronics 6,137, 165, V, XI Photo-field emission 97
nanoemitters 11 Pix Tech Company 150,152
Nano-FE(cathode(s)) 138 Pixel size 150, 152, 158
nanometer emitters 7 Pixel, individual pixel 148, 149
nanometer range 6, 140 Planar (lateral) vacuum FE triode 153
nanometer-scale 17, 54, 137, 139, X Planar addressable matrix 148
Nanometer-size protuberances 143 planar FE cathodes 161
nanosecond duration 56, 62, 63, 113 Planar FE microtriode 153
Nanostructures 131, 140, 158 planar macro-cathode surface 8
nanotip 54 Plasma panel displays 151
Nanotubes micrograph 167 Point FE cathode 139
narrow capillaries 61 Point-cathode diode 20
Near-surface region 13,15, 30, 42, 56, 71, 81, Poisson equation 19,21,22,92
82, 83, 84, 85, 88, 92, 94, 163 polar angle 7,40
Needle 61, 67, 68,153,161 power amphfier 61
INDEX 203
Power consumption 146,150, 151, 152 Scanning electron microscope(s) 121,123,124,

power measuring device 61 135,140,147
pre-breakdown phenomena 33, 34, 35 scanning tunneling microscope 7
pre-breakdown situation 32 scattering theoretic technique 7
precession of tunneling electron spin 59 Schottky barrier 139
pre-explosion phase 38 Schottky emission regime 128
Pressure gauges 138 Screen diagonal 150
prismatic resonator (H03 mode) 61,67 Secondary emission coefficient 105
Probe method 92 Seits-Vasiliev-Van Oostrum 28, 29
probe-hole 36, 37, 38 Selective adsorption 54,124,128
Proportional semiconductor detector 106 Selective chemical etching 139
Protuberances 111, 119, 122, 123, 131, 143 Selective work function reduction 124
p-Type Semiconductors 81, 83, 84, 88,91, self heating 37,44,63
92,95 "Self organizing" systems 137
Pulse amplitude analyzer 106 Semiconductor detector 106,107,110,113
pulse repetition 63 Semiconductor dielectric boundary 139
Pulsed illumination 97 semiconductor emitters 78,81
Pyramidal nanocrystals 165 Semiconductor whiskers 75
Pyramidal protrusion 121 Service life 117,135,150,159
Shadow mask 150,151
Shape of the potential barrier 9,19
quasi stationary conditions 33 sharp edge-plane system 24
quasi-classical (approach) 6 sharp microtips 6, 10
Quasi-Fermi level 87,89 Sharp point(s) 116
quasi-particle operators 13 Sharp protuberances 131
quasistatic regime 61 Sharp riges, edges 145
sharpening 72, 78
Silicon base 147
radii of curvature 7,61, 116
Silicon dioxide 139
radio pulse 63,64,65
Silicon pyramids 139
rearrangement of the emitter surface 47
silicon substrate 153,155,157,158
Reconstruction (of the surface) 135,143, 164
Single atom 7,12,116,131,132
rectangular axially synmietrical barrier 7
Single crystal tip(s) 97,143
rectangular barrier 7, 58, 59
Single-crystal grains 142
relativistic effects 22, 23, 24, 25, 30
Single-electron, character 110,111
Relativistic electrons 30, 152
Si02 dielectric 147
Repetition rate 156
Small electron transport time 137
repulsive potential 69, XI Small micro-protuberances 122
residual gas pressure 53, 63, 100, 159 small-scale field emitters 10
Resist-free electron lithography 158 small-scale objects 6, 10
resistivity 42,45, 75, 76,77,78, 81, 99 small-size emitters 9
resonator(s) 61, 64, 65, 66, 67,157 solid needle substrate 61
Response time 151, 152 solid/vacuum interface 56
retarding potential curves 60, 64, 65, 66 Sony 150, 151
"ring" 33, 34, 35, 36, 37, 38,125,126
Ring structure(s) of Zr 125 Space charge region 78, 80, 82, 88, 92,97
Space-charge limited 88
Space-charge region 78, 80,97
Saturation current 82, 90, 91, 92,95,101 spectral function 14
saturation region 34,76, 80, 81, 83, 90, 91,96, spherical diode model 25
100,101 spherically synmietric model 10
204 INDEX
spin orientation of the incident wave 59 "tight" 12

Spin rotation in a magnetic field 59 tight binding wavefunctions 12
Spind-type field-emitter cathode array 147 time delay 37
Spot size 98, 139, 150 time dependent problem 44, 59
Stabilization by oxygen treatment 128 time of tunneling, traversal time 59
Stable semiconductor field emission cathode 99 Tin 146, 163, 164
static potential 58 Tip apex 78, 98, 134, 141, 142, 143, 148
static regime 65 Tip cathodes 117, 146, 149, 156, 157
Statistical distribution 105,114 tip heating 37,42,43,44
Statistical FE event(s) 106 Tipi profile 53, 121
Statistical processes in FE 106 Titan (Ti) 158
Statistics of FE 105, 108 Tl 157
Sticking coefficient 101 Tonks-Frenkel field 68
STM tip 7 transfer Hamiltonian methods 12
Stratton's theory 82, 83, 85 Transistors 138, 139, 153
stream-liquid mixture 47 transit effects 65, 66
Strip lines 157 transit phase of the emitted electron 65
Submicron dimensions 138,139,156 transit space 66
submicron inhomogeneities 67 transit time effects 64, 66, 69
Subpicosecond range 139 Transition processes 97
super small emission spot 56 transparency 1, 2, 5, 7, 9, 12, 31, 59, 81, 98,
Superconductor(s) 110,111,112 100,101,105,113 '
surface cleaning 71, 72, 73, 101 traps 72, 92, 95, 96, 98
surface diffusion in strong electric fields 71 Travelling wave tubes (TWT) 157
surface impurities 72 Tungsten (W) FE emitters 115
surface migration 37,48, 54, 63, 125, 128, 143 Tunnel microscopy 138
Surface of the dielectric 139 tunneling path 7
Surface self-diffusion 49, 115 tunneling spectroscopy 6
surface tension forces 67 turning points 26
Surface tension 67, 71, 117
surface transport processes 71 Ultra short frame 151
ultra-high-resolution 6
Tantalum (Ta) in the (III) direction 123 Ultrasmall emission site at (100) plane 131
under-barrier friction 58
Tektronix 151
Television modulation frequencies 148
thermal conductivity 42, 44,45, 51 Vacancies 131
thermal destruction 36, 56, 62, X vacuum arc 32, 63
thermal radiation 45 Vacuum condition(s) 63, 100, 106, 117, 124,
"Thermal-field build-up" 115, 116, 118 125, 127, 128
Thermal-field (T-F) emission 4 vacuum gap 4, 47, 65, 82, 98, 137, 149
Thermionic-field emission 5 Vacuum magnetic sensors (VMS) 158
thermoelastic stresses 46 Vapor-liquid-solid (VLS) method 75
thin layers 61 Vertical FE triode 154
Thorium atoms 123 Vertical microtriodes 138,153
three-dimensional (analysis) 44 very high current densities 19, 21, 23, 25, 27,
Three-dimensional microtriodes 153 29,30,31,131
three-dimensional Schrodinger 7 very high electric fields 6, 19, 67
threshold 64, 68, 71, 101, 149, 158, 161, 162, very high resolution electron microscopes 69
163, 164, VI video pulse 63, 65
INDEX 205
Virtual lack 137 X-band cavity 60

VME, Vacuum microelectronics 6, 11, 17, 69, X-ray micro-analyzers 115
71, 136, 137
Voltage drop 78,79,80,81
"Zero-field" approximation 83
Wtip 116,120,123,129,141,142
Zero current approximation 82, 87, 89, 90
W tips oriented in the (100) direction 125
Zr layers on W 126
wave packet 58
wavelengths 66 Zr layers 125, 126
weak magnetic field 59 Zr source 129
Weissner effect 112 ZrAV layers 128
Wide band gap semiconductors 97 ZrAV system 125, 127, 128
Wide band semiconductors 139 ZrOonW 115,125,126
Wide-band ampUfiers 157 ZrOAV layers 128
Wigner model 7 j8-factor 9

Livro - Field Emission in Vacuum Microelectronics Por Fursey

Uploaded by

Copyright:

Available Formats

Livro - Field Emission in Vacuum Microelectronics Por Fursey

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Livro - Field Emission in Vacuum Microelectronics Por Fursey

Uploaded by

Copyright:

Available Formats

Field Emission in Vacuum

Series Editors: Ivor Brodie, Consultant

Recent volume's in this series:

Kluwer Academic / Plenum Publishers

© 2005 Kluwer Academic / Plenum Publishers, New York

feasibility of achieving high current densities, by obtaining stationary current densities of

1 Field Electron Emission from Metals 1

2 Characteristics of Field Emission at Very High Current Densities 19

3 Maximum Field Emission Current Densities 31

3.5.2 The Limiting Current Densities Attained Cooling 50

Field Emission in Microwave Field 57

Field Emission from Semiconductors 71

Statistics of Field Electron Emission 105

7 The Use of Field Emission Cathodes in Electron Optical Systems:

8 Advances in Applications 137

List of Main Notation 193

1.1. FOWLER-NORDHEIM THEORY

Under these assumptions, the current density is given by the equation

where T^(y) is the Nordheim function.

i^(y) = l-^l\\ + (1 - >;2)l/2]l/2 . [^(^) - {1 - (1 - /)l/2}]/^(^), (1.4)

having for an argument

Using (1.3), the FE current density at T = 0 follows the classic FN formula

j = 1.54 . 1 0 - ^ - 2 - — -6.83 . lO^^i^iy) (1.8)

where y = 3.79 • IQ-^^. ^/F/(p, t(y) = i}(y) - (2y/3)(di^(y)/dy).

Figure 1.2. The behavior of i^,t, and 5 as a function of >'.

\nl = /(!/(/)isastraightlineoverawiderangeofemissioncurrentvalues. Here, / = j-S

which, as with t{y), is very close to unity (Fig. 1.2). Then,

1.2. THERMAL-FIELD EMISSION

At nonzero temperatures the energy spectrum of electrons in a metal will contain

D(Ejc,F)=\l-\-Qxp(-2ih~^ j p(x)dxj\ (1.14)

1.3. EXTENDING THE THEORY OF FIELD ELECTRON EMISSION

1.3.1. Deviations from the Fowler-Nordheim Theory and Peculiarities of

m^^= ^exp(-2po^O ^^^^^^

D{E) = ^ ^ exp(-2/7o^i), (1.16)

^(^> = J77^ exp(-2po4), where ix^l. (1.18)

In the case of a planar cathode with a micro-roughness it is correct only where FQ =

U2 = -—- peFox + EF-\-(P (1.20)

U3 = -—- <i>(x) + FF + ^ (1.21)

'^^^ Fermy level

U(x) = - ^ - eFox ( ^ ^ ] +E^-\-(p (1.22)

where FQ is the field strength at the emitting surface.

184 Plain model

•^ 16 Emitter radius Hg=80 A

0.5 1 1.5 2 2.5 3

condition RQ » d, with the emitter radiusfixed,is violated in weakfieldsmore significantly

and more thoroughly. It is important to bear in mind that, as shown experimentally, in

1.3.2. The Effect of Fermi Surface Structure

is described by a linear dispersion of the form E = b(k), where b is a. parameter. The

1,3.3. Many-Particle Effects

where HL = J2k,cr ^kC^Ck and HR = ^^^A^d^dq are unperturbed Hamiltonians,

WT - X ] ^T^k,Cld,a + n,d^aCka). (1-24)

Here, the matrix element Tkq was expressed in the form:^^^

The exponent in the Eq. (1.25) can be approximated by^

obeys the equation

; = -ei([NL. Hj]) = -2e]mYTkq{C-^dq). (1.30)

Expression {C^dq) in an abridged form will be

{C^q) = {S-Ht)Cj^(t)dq(t)S(t)), (1.31)

where Ck(t) = exp(iHLt)Ck expi-iH^t) and

5(0 = exp - / j txp{r)t')Hi{t') dt' .

Figure 1.8. Schematic diagram of FE electron pair tunneling.