(Springer Series in Advanced Microelectronics Vol. 10.1007 - 978!3!319-56561-3) Egorov, Nikolay

Springer Series in Advanced Microelectronics 60
Nikolay Egorov
Evgeny Sheshin
Field Emission
Electronics
Springer Series in Advanced Microelectronics
Volume 60
Series editors
Kukjin Chun, Seoul, Korea, Republic of (South Korea)
Kiyoo Itoh, Tokyo, Japan
Thomas H. Lee, Stanford, CA, USA
Rino Micheloni, Vimercate (MB), Italy
Takayasu Sakurai, Tokyo, Japan
Willy M.C. Sansen, Leuven, Belgium
Doris Schmitt-Landsiedel, München, Germany
The Springer Series in Advanced Microelectronics provides systematic information
on all the topics relevant for the design, processing, and manufacturing of
microelectronic devices. The books, each prepared by leading researchers or
engineers in their fields, cover the basic and advanced aspects of topics such as
wafer processing, materials, device design, device technologies, circuit design,
VLSI implementation, and subsystem technology. The series forms a bridge
between physics and engineering and the volumes will appeal to practicing
engineers as well as research scientists.
More information about this series at http://www.springer.com/series/4076

Nikolay Egorov Evgeny Sheshin
•
Field Emission Electronics
123
Nikolay Egorov Evgeny Sheshin
Saint Petersburg State University MIPT
St. Petersburg Dolgoprudny, Moscow region
Russia Russia
ISSN 1437-0387 ISSN 2197-6643 (electronic)

Springer Series in Advanced Microelectronics
ISBN 978-3-319-56560-6 ISBN 978-3-319-56561-3 (eBook)
DOI 10.1007/978-3-319-56561-3
Library of Congress Control Number: 2017939090
Copyright for the English version is with the authors
Translation from the Russian language edition: Avtoelektronnaya emissiya. Principy i pribory by Nikolay
Egorov and Evgeny Sheshin, © Intellect Publishers 2011. All Rights Reserved.
© Springer International Publishing AG 2017
This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt from
the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, express or implied, with respect to the material contained herein or
for any errors or omissions that may have been made. The publisher remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations.
Printed on acid-free paper
This Springer imprint is published by Springer Nature

The registered company is Springer International Publishing AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Preface
Modern technological applications of electron beam devices require development

and creation of new electron sources with parameters that were not previously
achievable.
The unique quantum mechanics phenomenon of electrons tunneling from con-
densed state into vacuum known as field electron emission (FEE) provides a way of
creating electron sources with emission characteristics of six or seven orders of
magnitude above all the previously known types of emission. This incredibly high
emission current density combined with the fact that one does not require to input
any additional energy for the emission process itself provides exceptional oppor-
tunities for practical applications and, first of all, for development of unprece-
dentedly efficient electron beam devices.
It is also important to note that nowadays the whole science of field emission
goes through a period of renaissance due to the emergence of a new field of
electronics—vacuum nanoelectronics.
This book aims to summarize the theoretical, experimental and technological
advances in modern science of field electron emission that can help in development
of new emission devices based on nanostructured materials.
This book is roughly divided into three basic topics: theoretical models and
formulas for computation of electron gun elements; experimental techniques and
technologies; and basics of practical applications of field cathodes in electronic
devices. Particular attention was paid to the properties of field electron emission out
of nanostructured carbon materials and devices based on them, as in authors'
opinion these materials are the most promising for devices with field cathodes.
Working on this book, the authors did not seek to provide a full review of all
works on the subject or even of a majority of them. The backbone of the book
consists of theoretical, methodical and experimental results obtained by researchers
and teams that the authors led or worked with at department of Vacuum Electronics
of Moscow Institute of Physics and Technology and department of
Electromechanical and Computer Systems Modeling of Saint Petersburg State
University.
v
vi Preface
In the course of their research many original models, techniques and devices
were introduced, and many of the most interesting results are presented in this book.
The authors extend their thanks to their colleagues and students whose collab-
oration over the years, common research and discussion of results were the prin-
cipal motive behind this book.
Chapters 1, 3, 4, 7 are written by N.V. Egorov. The introduction, Chaps. 2, 5, 6,
8 are written by E.P. Sheshin. The conclusion is written by the authors with con-
tributions from K.A. Nikiforov.
St. Petersburg, Russia Nikolay Egorov

Dolgoprudny, Moscow region, Russia Evgeny Sheshin
Acknowledgement
The seventh chapter and conclusion was based on work of N.V. Egorov with the
Saint Petersburg Electrotechnical University “LETI,” Russia. The work has been
funded by Russian Science Foundation (Grant No. 15-19-30022), SPETU LETI.
vii
Contents
1 Basic Principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 1
1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 1
1.2 Fowler–Nordheim Theory of Field Emission from Metals . . . . . .. 11
1.3 Classical Morgulis-Stratton Theory of Field Emission
from Semiconductors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 20
1.4 Basics of Müller Field Emission Microscope Theory . . . . . . . . . .. 27
1.5 Disadvantages of Classical Field Emission Theories
and Limits to Their Applicability . . . . . . . . . . . . . . . . . . . . . . . . .. 35
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 40
2 Experimental Equipment and Technique . . . . . . . . . . . . . . . . . . . . . . 43
2.1 Field Emission Microscopes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
2.1.1 Design of Field Emission Microscopes . . . . . . . . . . . . . . . . 47
2.1.2 Special Designs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
2.2 Field-Electron Total Energy Analyzers. . . . . . . . . . . . . . . . . . . . . . 57
2.3 Field Emission Microscopes of Universal Design . . . . . . . . . . . . . 67
2.4 Electronic Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
2.5 Emitter Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
2.5.1 Metal Tips . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
2.5.2 Machining . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
2.5.3 Foils, Sheets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
2.6 Some Technological Cases of Application of a Field Emission
Microscope. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
2.6.1 First-Type Migration. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
2.6.2 Second-Type Migration . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
2.6.3 Third-Type Migration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
ix
x Contents
3 Modern Developments in Theoretical Research of Field

Emission . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
3.1.1 Emitter-Shape Approximation . . . . . . . . . . . . . . . . . . . . . . . 121
3.1.2 Potential and Field-Intensity Computation Methods . . . . . . 121
3.1.3 Comparison of Theoretical and Experimental Results . . . . . 124
3.2 Thermal-Field Emission (TFE) from Metals . . . . . . . . . . . . . . . . . . 125
3.3 Field Emission Spectroscopy. Energy Distribution of Field
Electrons and Thermofield Electrons Emitted from a Metal . . . . .. 135
3.3.1 Energy Distribution of Thermal-Field Electrons . . . . . . . .. 138
3.4 Phenomenological Theories of Field Emission from
Semiconductors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 141
3.5 Theoretical Aspects of Phenomena and Processes
on the Surface During Field Emission . . . . . . . . . . . . . . . . . . . . .. 155
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 166
4 Simulation of Structure and Parameters of Field Emission
Cathodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 171
4.1 Simulation of a Potential Barrier and Barrier Permeability
by Experimental Data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 171
4.2 Computation of Theoretical Characteristics of a Metal Field
Emission Cathode for a Model Configuration of Its Apex . . . . . .. 181
4.3 Simulation of the Metal Field Emission Cathode Surface
Structure. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 187
4.4 Simulation of Distribution of the Work Function Over
the Field Emission Cathode Surface . . . . . . . . . . . . . . . . . . . . . . .. 194
4.5 Theoretical Studies of the Physical Processes Affecting
the Field Emission Current Density Limits . . . . . . . . . . . . . . . . .. 201
4.6 Special Properties of Field Emission in Presence of Strong
Electric Fields and High Current Densities . . . . . . . . . . . . . . . . .. 214
4.7 Modeling of Elements of Beam Formation and Control
Systems with Controlling Magnetic Field
(High-Precision Beams) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 220
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 225
5 Field Emission Cathodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
5.1 Cardinal Problems of Field Emission Cathodes . . . . . . . . . . . . . . . 229
5.2 Pointed and Multiple-Apex Field Emission Cathodes. . . . . . . . . . . 234
5.3 Blade and Wire Field Emission Cathodes . . . . . . . . . . . . . . . . . . . 249
5.4 Film Field Emission Cathodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255
5.5 Field Emission Cathodes Made of Whiskers . . . . . . . . . . . . . . . . . 263
5.6 Field Emission Nanostructures Based on Nanoporous Anodic
Aluminum Oxide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 273
5.7 Other Field Emission Cathode Types . . . . . . . . . . . . . . . . . . . . . .. 279
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 286
Contents xi
6 Carbon-Based Field-Emission Cathodes . . . . . . . . . . . . . . . . . . . . . . . 295

6.1 Carbon Fibers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295
6.1.1 Polyacrylonitrile Carbon Fibers . . . . . . . . . . . . . . . . . . . . . 295
6.1.2 Carbon Nanofibers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 303
6.2 Carbon Nanotubes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306
6.2.1 Structural and Field-Emission Features . . . . . . . . . . . . . . . . 306
6.2.2 Some Carbon Nanotube Growing Methods . . . . . . . . . . . . 311
6.3 Non-oriented Structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 317
6.4 Carbon Foils . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321
6.5 Large-Size Flat Field Emission Cathodes . . . . . . . . . . . . . . . . . . . . 329
6.5.1 Electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 330
6.5.2 Screen Printing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 334
6.5.3 Structural Materials. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 337
6.6 Emissivity Improvement. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 338
6.6.1 Field Emission Cathode Forming . . . . . . . . . . . . . . . . . . . . 338
6.6.2 Pre-treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 339
6.6.3 Coatings and Doping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 342
6.7 Features of Measurement and Analysis of Current-Voltage
Characteristics of Carbon Materials . . . . . . . . . . . . . . . . . . . . . . .. 345
6.7.1 Ways to Measure Current-Voltage Characteristics
and Their Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 345
6.7.2 Technique of Long-Term Field-Emission Tests
and Experimental Data Analysis . . . . . . . . . . . . . . . . . . . .. 351
6.7.3 Evaluation of Uniformity of Emission from a Field
Emission Cathode Surface . . . . . . . . . . . . . . . . . . . . . . . .. 355
6.7.4 Developed Algorithm of Collection and Preprocessing
of Experimental Data . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 357
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 361
7 Computation of Field-Emission Cathode-Based Electron Guns . . . .. 369
7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 369
7.2 General Problem of Computation of the Optimum Structure
of Electron Guns Based on One-tip and Multi-tip Field
Emission Cathodes, i.e. the Almazov–Egorov Model . . . . . . . . . .. 374
7.3 Mathematical Simulation of Model Triode Electron-Optical
Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 388
7.4 Calculation of Distribution of the Electric Field
in the Forming and Controlling Systems Based
on a Field-Emission Cathode and a Small-Aperture
Focusing Diaphragm System . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 397
7.5 Calculation of Optimum Characteristics in the Forming
and Controlling Systems Based on a Field-Emission Cathode
and a Focusing Diaphragm System . . . . . . . . . . . . . . . . . . . . . . .. 401
xii Contents
7.6 Computation of Electron Paths in a Field-Emission

Cathode-Based System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 411
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 424
8 Field Emission Cathode-Based Devices and Equipment . . . . . . . . . . . 427
8.1 Light Sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 427
8.1.1 Operating Principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 429
8.1.2 Small-Button Glass Cathodoluminescent Lamps . . . . . . . . . 432
8.1.3 Flat Cathodoluminescent Light Sources . . . . . . . . . . . . . . . 439
8.1.4 Cylindrical Cathodoluminescent Light Sources . . . . . . . . . . 445
8.1.5 Spherical Cathodoluminescent Light Sources . . . . . . . . . . . 447
8.1.6 Ultraviolet Lamps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 448
8.1.7 Cathodoluminescent Light-Source Application
Guidelines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 451
8.2 Flat Display Screens . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 459
8.2.1 Display Screens with Pointed Field Emission Cathodes . . . 460
8.2.2 Diode Display Screens . . . . . . . . . . . . . . . . . . . . . . . . . . . . 463
8.2.3 Triode Display Screens . . . . . . . . . . . . . . . . . . . . . . . . . . . . 467
8.2.4 Multielectrode Display Screens . . . . . . . . . . . . . . . . . . . . . . 474
8.2.5 Lateral Display Screens . . . . . . . . . . . . . . . . . . . . . . . . . . . 475
8.2.6 Screens with Reversed Control Electrodes . . . . . . . . . . . . . 477
8.2.7 Thin-Film Field Emission Cathode-Based Displays . . . . . . 479
8.3 Microwave Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 483
8.3.1 Field Emission Microtriodes and Amplifiers . . . . . . . . . . . . 486
8.3.2 Magnetrons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 489
8.3.3 Traveling-Wave Tubes . . . . . . . . . . . . . . . . . . . . . . . . . . . . 492
8.3.4 Reflex Klystrons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 493
8.4 X-ray Tubes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 494
8.5 Electron Guns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 500
8.5.1 General Principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 500
8.5.2 Field Emission Localization Methods . . . . . . . . . . . . . . . . . 501
8.5.3 Low-Power Electron Guns . . . . . . . . . . . . . . . . . . . . . . . . . 503
8.5.4 High-Power Electron Guns . . . . . . . . . . . . . . . . . . . . . . . . . 514
8.6 Other Types of Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 516
8.6.1 Heaters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 516
8.6.2 Sensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 520
8.6.3 Memory Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 521
8.6.4 Ion Sources. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 523
8.6.5 Full-Color Projectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 525
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 527
Conclusion (Prospective Development) . . . . . . . . . . . . . . . . . . . . . . . . . . . 539
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 559
Introduction
Among the most relevant problems of modern electronics, creation of stable field
emission cathodes able to function for extended periods of time in high technical
vacuum (10−6–10−7 mm Hg) occupies an important place. There are well-known
advantages of field electron cathodes compared to other types of free electron
sources. Those are as follows: lack of heating, high emission current density,
resistance against temperature fluctuations, low sensitivity to external radiation,
lack of inertia, and exponential slope of volt-ampere characteristics. Those prop-
erties in conjunction make field cathodes invaluable for use in various electronic
devices such as cathode ray devices, and flat screens.
The main obstacle for creation of stable field emission cathodes is extremely
high sensitivity of field emission process to any changes of cathode geometry or its
surface structure. Function of a field emission cathode in any real device is always
accompanied by various processes occurring on its surface such as: ion bombard-
ment; ponderomotor stress; adsorption and desorption of molecules of residual
gases; and surface migration. Depending on specifics of a particular construction
and mode of exploitation of the cathode, those processes individually or cumula-
tively lead to a series of effects that ultimately change the structure of a cathode or
mode of its operation: cathode dissipation of material, changes in the shape of
emitter surface, changes in number or position of microelevations on emitter sur-
face, changes in work function, heating of the cathode, mechanical stresses, etc.
Field emission is the most conservative type of free electron emission which
allows creation of new efficient electron devices with new consumer properties.
In order to provide better understanding of phenomena and processes emerging
during field emission, this book is separated into theoretical and experimental parts.
Theoretical chapters consider properties of field emission based on modern
physical and mathematical concepts and models.
Experimental chapters are dedicated to experimental techniques and methods of
field emission experiments as well as practical applications of field cathodes. This
part also includes assessment of field cathode manufacturing methods.
A particular attention is paid to perspectives of future development of field
emission electronics.
xiii
xiv Introduction
For readers’ convenience, each chapter has a separate list of references.

The summary of many years of development of field emission science presented
in this book can, as the authors hope, give an impulse to the new breakthroughs in
the study of field cathodes and technologies that employ them.
The authors are fully aware that any of this book's sections can be grown into a
full-blown monograph of its own. Nevertheless, it seems crucial for further progress
of the field emission electronics to summarize the data amassed over the years.
Chapter 1
Basic Principles
Abstract This chapter describes the basic principles of Fowler–Nordheim theory

of field emission from metals and classical Morgulis-Stratton theory of field
emission from semiconductors. Then it discusses the theoretical basics of Müller
field emission microscope as the first experimental device which allowed to test the
main conclusions of the classical theory of field electron emission. Disadvantages of
these classical theories and limits to their applicability are discussed.
1.1 Introduction
Structures comprised of charged particles (electrons or ions), such as solids, liquids

gases or plasma, are all natural emitters of charged particles. Of them, it is easy to
assess the maximum electron current density for metals. As in normal conditions
the electron current onto the surface of metal body has density of j = en0V (where
e = 1.6 10−19 C is the electron charge; n0 = 1022 1023 cm−3 is the concen-
tration of electrons; V—velocity, which for electrons of Fermi energy is equal to
108 cm/s), the maximum possible value of current density is a colossal
j = 1012 A/cm2. But in reality only a small fraction of all the electrons are emitted
out of metal due to significant brake forces that emerge on passing of the electron
through the surface. For classic metals the work function is about 2 6 eV, and
there are not a lot of electrons that have energy exceeding this work function—in
normal conditions, in fact, their amount is negligible. The standard method of
increasing the number of emitted electrons is giving additional energy to conducted
electrons such that they can overcome the work function.
Such additional energy can be applied by some external impact such as: elec-
tromagnetic radiation (photo-electron emission, PEE), bombardment with
high-energy particles (secondary electron emission, SEE) etc. Nowadays the most
widely used method is, however, the method of thermos-electron emission (TEE),
© Springer International Publishing AG 2017 1

N. Egorov and E. Sheshin, Field Emission Electronics, Springer Series
in Advanced Microelectronics 60, DOI 10.1007/978-3-319-56561-3_1
2 1 Basic Principles
i.e. emission of electrons from warmed up solids. There is yet another type of
electron emission that does not require applying of additional energy, which is thus
called field-electron emission (FEE).1
The effect of FEE appears when an extremely strong electric field (with field
strength E * 107 V/cm) is applied to the surface of a solid resulting in transfor-
mation of the potential threshold into a potential barrier which is low and, most
importantly, narrow enough to allow the electron tunneling i.e. the electrons are
able to phase through this barrier and leave the solid without loss of energy.
The phenomenon of FEE was not theoretically explained up until the discovery
of quantum mechanical tunneling effect (e.g., [1, 2]). Soon after this discovery,
about 1928–1929, Fowler and Nordheim have constructed the first well-founded
theory of FEE in metals [3, 4]. Further evolution of FEE theory was mostly based
around development of new models and computation techniques.
As to the theory of semiconductor FEE, one has to admit that up to now such a
theory that would be general enough, logically sound and mathematically and
physically consistent, does not exist [5–7]. The same applies to FEE from carbon
structures [8].
In order to understand the phenomenon of autoelectronemission (same as in case
of thermoelectron and other types of emission) it is necessary to understand the
properties of the system of electrons in solid crystals that serve as emitters. The
most important such properties are the characteristics of the electrons’ energy
spectrum (zonal structure of the crystal), defined by laws of quantum mechanics,
and the energy distribution of electrons in permitted zones that are defined by
statistical laws. These questions are thoroughly studied is a series of excellent
monographs and textbooks (such as, e.g. [1, 5, 6, 9–12]), and thus we are not going
to pursue those problems in too much detail within confines of this book.
Instead we are going to focus mostly on studying and analysis of phenomena on
phase boundaries between metal (or semiconductor) and vacuum and between
metal (semiconductor) and dielectric, that define the phenomenon of FEE itself and
its circumstantial processes.
Basic dependence of emission current density on intensity of the
emission-causing impact (temperature, lighting, electric field intensity etc.) for any
kind of emission can be calculated if we manage to evaluate the integral
Zþ 1 Zþ 1 Zþ 1
j¼e dpy dpz f ðpx ; py ; pz ÞDðpx ; EÞpx m1 dpx ; ð1:1Þ
1 1 1
where e is the electron charge, f is the function of electron distribution in the metal
by pulse components ðpx ; py ; pz Þ; Dðpx ; EÞ is the potential barrier penetrability
factor (penetrability) in presence of an external electric field at the metallic surface
1
Other names for the effect of field electron emission (FEE) include autoelectron emission (AEE),
cold emission, electrostatic emission etc.
1.1 Introduction 3
for the electrons falling on the surface of metal with impulses px along the normal to
the surface coinciding with axis x, and E is electric intensity. If D depends only on
px, then integration of (1.1) over py and pz reduces the problem under consideration
to one-dimensional one and transition from variable px to Ex results in
Zþ 1
j ¼ jej DðEx ; E; UÞnðEx ÞdEx ; ð1:2Þ
0
where nðEx ÞdEx is an electron flux with energies in the interval dEx about Ex from
within the cathode; Ф is the work function known to be caused by the energy spent
to remove an electron from the solid (or the fluid) in the vacuum.
The value of nðEx Þ is calculated from Fermi-Dirac distribution from the ratio

4pmkT E
nðEÞ ¼ ln 1 þ exp ;
h3 kT
where h is Plank constant.

As nðEÞ is given, it is possible to obtain basic emission dependence for the case
of FEE j = j(E) via the integral (1.2) if the value of DðEx ; EÞ is known or has been
determined.
As noted above, electrons are retained inside a solid, metal or semiconductor by
electric forces and cannot leave it by themselves. This situation is formalized by
introduction of a concept of a potential threshold/barrier for the solid–vacuum
boundary.
The potential barrier can be defined as a region of space where the potential
energy of a particle (in this case an electron) assumes a value exceeding that of its
total energy. Putting this in mathematical terms, the potential barrier is represented
as a functional dependence of the potential energy of particle/electron and its
coordinate. Visually the potential barrier can be represented as an arbitrary potential
curve, for an instance the one shown in Fig. 1.1 as an instance. The region where
energy U is distinct from zero is finite. It stretches from x = a to x = b. The region
is referred to as a potential barrier (E\U).
In classical mechanics [13] typically a particle cannot get into the potential
barrier region without some external work function against potential field forces. In
quantum mechanics, however, it is typical for a particle to penetrate a potential
Fig. 1.1 Arbitrary potential

U
barrier
I II U III
0 a b x
barrier even when it does not get any supplementary energy (tunnel effect). Classic
statistical physics often represents thermal energy exchange between particles as an
energy source necessary to overcome the potential barrier. The exchange is similar
to the one taking place, for example, in thermionic emission. If at the “summit” of
the potential barrier E ¼ Umax , the total energy of the particle E is only an
insignificant part of the total energy of the entire closed system in the thermal
equilibrium state, the probability of penetration through the potential barrier is

directly proportional to the Boltzmann factor exp½ðUmax EÞ=kT; where E is the
average energy of the particle, T is the absolute temperature, k is the Boltzmann
constant [14].
In quantum mechanics [15], penetration through a potential barrier can take
place even in a purely conservative system not located in a thermostat nor
exchanging energy with any other objects. In the “classically unattainable” region
U [ E; the wave function of the particle is distinct from zero (although it decreases
exponentially towards the potential barrier region). If the region dimensions are
finite, the particle can penetrate through it. In a one-dimensional case, when the
“classically unattainable” (the so-called subbarrier) region is limited with two
points x1 and x2, whose locations are identified by the equation UðxÞ ¼ E; the
probability of penetration of a particle/electron through a potential barrier
(barrier-transmission factor, barrier penetrability) is defined as
8
x 9
Z 2
< =

D ¼ exp 2 kðxÞdx ;
ð1:3Þ
: ;
x1
where
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2m½E UðxÞ
k¼
h
is the wave number of the particle, m is its mass and ħ is the Planck constant.
Expression (1.3) holds true on condition that within the distance of one particle
wave length from any cusp decomposition of potential energy U into the difference
ðx x1; 2 Þ can be done through linear approximation. Away from the cusps,
quasi-classical approximation of quantum mechanics is applicable everywhere [1].
Given these conditions, the transmission factor D is exponentially small (in clas-
sical mechanics it is zero). For this type of potential barrier, the particle wave
properties can also be significant in the subbarrier region of values E [ U. In
classical mechanics, this region would demonstrate free passage of particles over
the barrier in the direction of their movement and the reflectivity factor R ¼ 1 D
would become zero. In quantum mechanics, it remains finite for any value of E,
though, generally, it is exponentially small in the suprabarrier region.
Let us go back to Fig. 1.1. Obviously, the shape of the potential barrier shown
there is nowhere near the actual situation at the solid-vacuum interface. In absence
1.1 Introduction 5
(a) (b) (c)
Fig. 1.2 Field electron emission. Change the width of the potential barrier with increasing electric
field strength. U(x)—the potential energy of the electron
of an external electric field at the boundary there is a potential threshold rather than
a potential barrier (cf. Fig. 1.2a). One can also imagine that in this case, electrons
are retained in the metal by a semi-infinite potential barrier. Electron emission only
appears when electrons can leave the metal by passing the threshold due to the
energy received before. In particular, the energy can be received as a result of
heating the metal in the same way it happens in thermionic emission. The situation
is radically different if the metal/semiconductor surface is subjected to a strong
electric field. The potential threshold becomes a potential barrier. The higher the
electric field intensity, the narrower is the barrier (Fig. 1.2b, c). As a result, metal
electrons have a theoretical possibility to leave the metal due to the tunnel effect [1].
The triangular potential barrier represented in Fig. 1.2b, c, is closer to the actual
situation, and Fowler and Nordheim even used it in developing their first theory to
evaluate potential barrier penetrability [3]. Nevertheless, it is also highly idealistic.
Just like the electrons in the shell of any atom cannot leave the atom under usual
conditions (i.e. without an external action) due to the forces of attraction to the
nucleus, free electrons of a solid cannot (as we already mentioned) leave the solid
under usual conditions due to the decelerating field forces at the boundaries of the
solid. As established, these forces have a dual origin. Firstly, they are caused by the
fact that the electron leaving a solid polarizes electron gas (the term of polarization
has been used in a general sense, i.e. it includes both electrostatic induction for
conduction electrons and intra-atomic polarization for bound electrons). Secondly,
these forces are caused by double electric layers on clean surfaces of solids. From
the quantum-mechanical point of view, the double layers are explained by
“buckling” of w-functions on the crystal surface. From the classical point of view
double layers are caused by the fact that some free electrons leave the crystal
surface and develop an equilibrium electron cloud near the surface, thus creating a
responding deficiency of electrons on the surface.
Double layers also emerge when alien substances are adsorbed onto the surface
of a solid. Depending on their nature, they can form electric double layers of
various values and signs.
As escape of an electron obviously needs expenditure of energy to deform the
electron gas, there is less energy at the solid-vacuum interface before than after the
escape of the electron irrespective of the real shape of the potential barrier.
The value characterizing the height of the potential barrier at the solid boundary
is the work function U.
Work function is defined as the minimum work necessary to remove an electron
from a metal. In this definition, the work function equals the difference in energies
of two states of the entire crystal. In the initial state, an electrically neutral metal
with N electrons is in its basic state with energy EN . In the final state, one electron
has been removed from the metal and has only electrostatic energy corresponding to
the level of vacuum Evac . The metal with remaining N 1 electrons passes into a
new state EN1 . Hence, the work function can be described as
U ¼ EN1 Evac EN :
For thermodynamic change of the state, the difference EN1 EN can be rep-
resented as a derivative of free energy F with respect to the number
@F of particles at
constant temperature and constant volume V. The derivative @N T;V is an elec-
trochemical potential Г of electrons. At nonzero temperatures it is equal to the
Fermi level EF :

@F
EN1 EN ! ¼C
@N T;V
or

@F
¼ EF
@N T;V
Thus, the work function can generally be expressed as:
U ¼ Evac EF :
The work function from clean surfaces mostly comes down to work against
polarization forces. As for double layers, they are either small (e.g. in alkaline
metals) or compensated by high enough kinetic energy of electrons in the substance.
As for semiconductors or dielectrics, we have not found any discussion of double
layers on their surfaces in the literature yet. From the classical point of view, the
relatively low concentration of “free” electrons makes the double layers very small.
The quantum-mechanical calculations made for metals show that the polariza-
tion forces outside the metal come down to the classical “mirror force” (polarization
force) introduced by Schottky. According to him, the force retaining an electron in a
metal is caused by the fact that the escaping electron induces a distributed positive
charge in the metallic surface, and the field thus generated between the electron and
the induced charge is fully equivalent to the field between the electron and the
positive charge equal to the electron charge and located in the point of the electron
2
image (Fig. 1.3), i.e. F ¼ 4xe 2 : However, according to Schottky considers this only
1.1 Introduction 7
Fig. 1.3 On the calculation

of the mirror image force
(polarization forces): e charge
of the emitted electron, +e its
mirror image
+e 0 -e
x x
holds starting with some distance x a nearly equivalent to the constant of the
metal lattice. At the distances 0 < x < a, the expression does not hold because of
the influence of the atomic structure of the surface. Schottky believes that at dis-
tances 0 < x < a, the polarization force can be regarded as constant and equal to the
image forceR in the point x = a. Therefore, the total work function of the metal
1
U ¼ eu ¼ 0 Fdx turns out to be equal to
e2
U ¼ eu ¼ : ð1:4Þ
2a
This expression cannot be directly applied to semiconductors and dielectrics.

If an electron escapes from some depth of a dielectric or semi-conductor crystal,
then we can show that the force of interaction between the escaping electron and the
positive charge it leaves is negligible compared with the force arising due to
intra-atomic polarization of the substance by the escaping electron. This force is
equivalent to the force of interaction between the electron and its “dielectric
image”. To estimate the work function for this case, we can go along the same line
of reasoning and derive the formula
e2 e 1
U ¼ eu ¼ : ð1:5Þ
2a e þ 1
Here the value of dielectric constant e shows only the electronic part of polar-
ization because the orientation part does not have enough time to manifest itself
within the electron escape period. The expression (1.5) for semiconductors and
dielectrics determines the value of their electron affinity. As it will be clear from the
subsequent, it is this value that matters for consideration of field emission.
If the surface of a metal or a semiconductor is covered with an adatom layer, and
the adatom is polarized and supplied with a dipole moment either due to interaction
between adatoms and the main substance or for some other reasons, then the work
function is changed by the value of potential energy surge caused by the dipole
layer (Fig. 1.4).
In that case, the work function can be expressed as the sum of two summands.
The first summand is connected with the bulk properties of the metal and deter-
mined by the electrostatic potential in the depth of the bulk. In this case, the Fermi
level counts as zero. In terms of a potential diagram (Fig. 1.5), it is U(−∞). The
second summand is a term sensitive to the surface properties and expressed by the
difference of electrostatic potential energies:
Z1
DU ¼ Uð1Þ Uð1Þ ¼ 4pe x½nðxÞ n þ ðxÞdx;
1
Fig. 1.4 The decrease in the

work function eu of the
magnitude D(eu) for the
adsorption of substances that
give a positive dipole
moment. The dashed line
shows the resulting dipole 0
∆(eφ)
eφ
+
D
1.1 Introduction 9
Fig. 1.5 The spatial Distance, Å

distribution of the electrostatic
potential U(x) (dashed line)
and the total effective
one-electron potential Ueff
(solid line) near the surface in
Energy, atomic units

the jelly model for rs = 5 [16]
Energy, eV
Fermi level
Distance x, Fermi wavelenght units
where n(x) is the electron charge density distribution, and n+(x) is the positive
charge density distribution.
This term is referred to as an electrostatic dipole barrier because it corresponds to
the work that has to be done to transfer an electron through the surface dipole layer.
The dipole moment value is a surface characteristic, which changes from surface to
surface. In one and the same metal, closely packed (atomically smooth) faces
usually have a larger dipole moment than loosely packed (atomically-rough) faces
do. It is understandable at the qualitative level if we imagine a rough surface as
containing some frame ions biased to the region of the charged negative layer,
which reduces the dipole moment, and consequently, the work function.
As a rule, the above also means that electron emitter surfaces are not homogeneous
in terms of work function. Inhomogeneity is caused either by the fact that the surface
is formed by different faces of a single crystal, or the different degree of coverage with
adatoms, or the polycrystalline nature of the surface. As a result, the contact potential
difference causes development of patches of accelerating and decelerating fields (spot
fields) on the surface. However, when a strong enough electric field is applied to the
emitter surface, relatively long-range spot fields cease to matter, while the local work
functions of particular surface patches remain significant.
Field emission Field emission is emission of electrons in the case when the emitter
boundary has a strong external electric field applied to it, which accelerates elec-
trons from the surface. A potential wall at the solid-environment interface is turned
into a barrier whose thickness depends on the field applied (the stronger the field,
the thinner is the barrier) (see Fig. 1.1). In this case, non-excited electrons can
tunnel this barrier, generating current whose density is a characteristic of this
process. It is clear that the field emission current density largely depends on the
field intensity at the surface.
Thermal-field emission Strictly speaking, emission of non-excited electrons in the
course of tunneling takes place at zero temperature. If T [ 0, some electrons have
levels of energy E [ E0 For excited electrons, the potential barrier becomes lower
and, respectively, narrower, which increases their barrier-transmission factor. This
leads to growth of field emission current in comparison with the previous case. So
thermal-field emission is caused by both the processes of transmission of electrons
through a potential barrier, and the processes of their thermal excitation.
Photo-electron emission By analogy with thermal-field emission, electrons can be
additionally excited by irradiation of the emitter surface with light quanta. This is
extremely important in case of a semiconductor emitter since this excitation
transfers valence-band electrons to the conduction band. Imposition of the electric
field results in a combined photo-field emission.
Having considered the main types of electron emission, we can now present a
potential barrier model closest to the actual situation and, consequently, most often
used to calculate potential barrier penetrability before calculation of the basic
dependence of field emission j = j(E) [integral (1.2)].
As noted above, electron emission without pre-excitation (field emission) can
only take place as a result of a tunnel effect if a strong electric field is applied to the
surface, whereby the potential threshold (an infinite potential barrier) turns into a
finite-width potential barrier. A triangular potential barrier is based on the speci-
ficity of field emission, i.e. availability of an electric field. However, this barrier
model is too rough. By now, the most complete theory of field emission from metals
Energy
vac
=0 Distance from surface, x
δΦ total potential of the

potential image forces
Φ -e/4x
extenal potential
external
x0
-Eex
F
Φ /eE
metal
Fig. 1.6 The diagram of the potential energy of the electron near the metal surface in the presence
of an external electric field with electric field strength E. The total potential (solid line) is the sum
of the image potential (dotted line) and the potential of the applied field (dashed line). Ф—work
function in the absence of an applied field. Reducing the potential for value dФ because the
Schottky effect. x0—position of the maximum total potential
1.1 Introduction 11
has been developed for the so-called polarization force or electrical image-force
barrier. A potential diagram explaining this model is presented in Fig. 1.6.
In subsequent paragraphs we will take it upon ourselves to develop a strict
quantum-mechanical assessment of the field emission phenomenon. In conclusion
of this paragraph, however, we have to mention that this phenomenon can also be
explained in terms of classical physics.
Actually, the external-field force acting on an electron is accompanied by an
attraction force from its electrostatic image in the metal. Effective potential energy
including interaction with the image is formulated as
e2
Ueff ðxÞ ¼ U0 eEx ;
4x
where x is the distance to which the electron moves from the metallic surface, the
distance between the electron and its image being 2x. The function of Ueff(x) peaks
pffiffiffiffiffiffiffiffiffiffi
at the point x0 ¼ e=4E, the potential energy at the peak being
pffiffiffiffiffiffiffiffi
Umax ¼ Ueff ðx0 Þ ¼ U0 e3 E :
So inclusion of electrostatic-image forces shows that electric field imposition

actually reduces the work function. If we meet the condition that either
EF U0 e3 E
or
e3 E U;
where EF is the Fermi level, U is the work function, then the electrons will be free
to leave the metal. According to classical views, the ratio e3 E U gives us the
electric field intensity at which field emission (e.g. for silver) is E 108 V/cm.
Meanwhile, experimental data (e.g. [17]) show that field emission can be observed
for fields at least an order of magnitude smaller. Hence, classical views cannot
adequately explain the field emission phenomenon.
1.2 Fowler–Nordheim Theory of Field Emission

from Metals
The classic Fowler–Nordheim theory of field emission from metals (e.g. [3–6, 8,
913–15, 17–20, ]) is not only of historical interest. Mainly, this theory describes the
process of field emission from metals adequately in terms of quality and precisely
enough in terms of quantity despite considerable simplifying assumptions included
into the physical and mathematical models of the phenomenon.
Before drilling down the central tenets of the Fowler–Nordheim theory, let us
state that any physical theory is a compromise between a strict physical model,
adequately selected calculation method and a possibility to obtain a result in the
simplest way and an analyzable form and turn our attention to its basic physical
prerequisites and assumptions used to simplify evaluations:
(1) the metal—vacuum interface is considered an ideal plane: the evaluation
problem becomes one-dimensional, the potential function (energy) U(x) de-
pends only on the x-coordinate, the external field is homogeneous;
(2) inside the metal U1(x) = const = − U0; outside the metal, the potential barrier
is caused only by activity of image forces U2(x) = e2/4x, where e is electron
charge;
(3) barrier penetrability D is computed by the semi-classic Wentzel-
Кramers-Brillouin (WKB) method; subsequent improvements of the theory of
field emission from metals were mostly connected to application of penetra-
bility computation methods exceeding the WKB method in validity;
(4) the metal model selected was the Sommerfeld model of free electrons in a
potential box evolving degenerated gas obeying Fermi-Dirac statistics;
(5) the Fowler–Nordheim theory was developed for the temperature T = 0 K, i.e.
for mere field emission.
The current density in the case of field emission can be calculated by the usual
formula that is also suitable for calculating current density for other types of
emission, such as thermionic, photo-electron etc.:
Z1
j¼e nðEx ÞDðEx ; EÞdEx ; ð1:6Þ
0
where nðEx Þ d Ex is the number of electrons with energy ranging between Ex and
Ex dEx falling from the inside of the metal onto the surface unit per second;
EðEx ; EÞ is a potential-barrier transmission factor at the metallic surface equal to the
ratio of intensity of the electronic wave passing through the barrier to the incident
wave intensity. Naturally, D depends on both Ex and the field value E:
DðEx ; EÞ ¼ 1 RðEx ; EÞ; ð1:7Þ
where R is the electron wave reflection factor.

It should also be stated that unlike thermionic emission, when we only have to
count the electrons whose energy exceeds the height of the potential barrier at the
metal-vacuum boundary, in case of field emission, we have to count all electrons
falling on the barrier because theoretically all of them are able to leave the metal,
and it is only probability that is different.
In the general case of arbitrary E and T, evaluation of current density j depending
on the field E, temperature T and work function (Ф = eu) by the formula (1.6) is a
great theoretical challenge. Therefore, it is expedient to break the entire energy
1.2 Fowler–Nordheim Theory of Field Emission from Metals 13
spectrum into several regions and apply particular mathematical methods to each
region. Figure 1.7 shows an energy-level diagram for a metal and a potential barrier
at the metal-vacuum boundary in absence and presence of electric field. When
T = 0, energies of all electrons are concentrated in Region A. T > 0 causes elec-
trons to appear in the Regions B, C and D. Region B electrons take part in field
emission along with Region A electrons. Within Region B, there is a small sub-
group B1, where electron energies get close to the maximum of potential barrier.
Obviously, barrier penetrability for Region B electrons is higher than that for
Region A electrons. The electron emission caused by electrons in Region B is
referred to as thermal-field emission. Region C and D electrons cause thermionic
emission, Region C electrons are responsible for the Schottky effect and Region D
electrons cause emission without an external electric field. (For the sake of sim-
plicity, Figs. 1.7 and 1.8 show a potential energy curve passing through the point
−U at x = 0. Actually, it passes through this point at xk determined from the
2
equality U ¼ Exk þ 4xe k ).
Figure 1.8 shows the image force barrier in presence of an external accelerating
electric field E. Then, for a potential electron energy V in the presence of an external
field, the following ratios are valid
V ¼ U at x\0;
e2 ð1:8Þ
V ¼ 4x eEx at x a:
C (Ee3)1/2
B1
Fig. 1.7 Typical areas of electron energy interval

-eEx
-eEx-e2/4x
-(Ee3)1/2 (x>a)
-W
0 x1 x+ x0 x++ x2
Fig. 1.8 To derivation of the basic formula of field electron emission
In this case, the barrier summit is located below the vacuum level by the value of
pffiffiffiffiffiffiffiffi pffiffiffi
e3 E and at a distance from the metal; x0 ¼ 12 Ee ; x1 and x2 are the distances from
the metallic surface at which the set electron energy E is equal to potential energy.
The Schrödinger equation for the barrier shown in Fig. 1.8 is given by
2
d2w 2m e
þ E eEx w ¼ 0: ð1:9Þ
dx2 h2 g2 4x
By definition, the barrier penetrability is
jat j2 jai j2 jar j2

D¼ 2
or D¼ ; ð1:10Þ
jai j j ai j 2
where ai, ar, at are the amplitudes of incident, reflected and transmitted waves,
respectively.Let
2
e
UðxÞ ¼ k02 E eEx ; ð1:11Þ
4x
where k02 ¼ h2m

2 g2 . The equation
d2w
þ UðxÞw ¼ 0 ð1:12Þ
dx2
can be integrated, and its solution is:

k1 k2
w ¼ meik0 x 1 þ þ 2 þ ð1:13Þ
k0 k0
In this equation, m, x, k1, k2 etc. are functions of x. They can be derived by

substituting the expression (1.13) into (1.12) and then equating the coefficients at
the same powers of k0. If k0 is large, which happens occasionally, we can confine
ourselves to the first term of the solution (1.13), which can be easily shown to have
the form
14 Zx
UðxÞ i 1
w¼ e ½UðxÞ2 dx ð1:14Þ
k02
This limiting form of solution for (1.12) is valid if the barrier (Fig. 1.8) contains
points x+ and x++ meeting the following conditions:
0 1 9
U ðx1 Þ 1 3 >
>
U00 ðx1 Þ jx þ x1 j U0 ðx Þ; >
>
1 >
>
3 =
jU0 ðx þ Þj
jUðx þ Þj2 ;
0 1 ð1:15Þ
U ðx2 Þ 1 3 >
U00 ðx2 Þ jx2 x þ þ j U0 ðx ;>
>
>
2 Þ >
>
0 þþ þþ 2
3 ;
jU ðx Þj
jUðx Þj :
The primes denote differentiation with respect to x.

The conditions (1.15) show that the potential barrier should have a “smooth”
nature, i.e. the potential function should not have any infinitely large first or second
derivatives. Hence, the solution (1.14) is inapplicable to an idealized triangular
barrier (Fig. 1.8). It follows from (1.15) that the higher the potential curve rises
above points x1 and x2, the better are the conditions (1.15) met. In particular, it
directly follows from them that
Zx2
1 4
jUðxÞj2 dx : ð1:16Þ
3
x1
This corresponds to D 7%. The calculation presented above was proposed by

G.F. Vasilev [17]. His method of solution of (1.12) is close to the famous Wentzel-
Кramers-Brillouin (WKB) method [21–23] but not exactly the same. The solution
(1.12) is obtained in form
Rx 1
i ½UðxÞ2 dx
14
w ¼ aUðxÞ e ; ð1:17Þ
where the constant factor a and the lower limit of integration are chosen arbitrarily.
The solution (1.17) is valid given the following conditions:
2
1 U00 5 U0
UðxÞ [ 0;
1: ð1:18Þ
4U U 4 U
It can be shown that for the barrier described by (1.8), this condition is met well.
Calculating the left-hand side, we get:
3a0 1 þ 32 Ex

1: ð1:19Þ
3 e2

8x 1 þ 4 Ex2 3
e
In this equation, a0 is represented with the Bohr radius of a hydrogen atom. It

can be seen that for x 1 Å, the condition (1.19) holds.
So, in the region x > x2, where Ф > 0, it can be stated that:

Rx 1
i U2 dx þ p4
wðxÞ ¼ aUðxÞ4 e
1
x2
: ð1:20Þ
The expression (1.20) differs from (1.17) by the fact that in the former the lower
limit of integration is defined and another wave phase is introduced for conve-
nience: (1.20) represents an electronic wave leaving the metal. It is invalid in the
negative kinetic energy field x1 < x < x2, since Ф < 0 [cf. (1.18)].
According to Kramers [24], in this region, the
individual function grows
R 1
exponentially from x2 to x1 if it behaves as U4 cos U2 dx þ p4 at x > x2 and
1
decreases along the same direction if cos is replaced with sin. So, in the field
x1 < x < x2
2 0 1 0 x 13
Zx2 Z2
14 4 1 1 1
wðxÞ ¼ ajUj exp@ U2 dxA þ i exp@ U2 dxA5: ð1:21Þ
2
x x
It is well-known that at x < 0 inside the metal, the function w has the wave nature
again. Therefore, the expression (1.21) can be represented as follows:
20 x 13
Z
1 p
1 þ e2Q exp4i@ U2 dx þ A5
14 1
wðxÞ ¼ aU eQ
4 4
x1
2 0 x 139 ð1:22Þ
Z =
1 p
þ 1 e2Q exp4i@ U2 dx þ A5 :
1
4 4 ;
x1
where
Zx2
1
Q¼ U2 dx: ð1:23Þ
x1
As inside the metal at a sufficient distance from the surface,
UðxÞ ¼ k02 ðE þ WÞ ¼ v2 ¼ const; ð1:24Þ
then

a 1 1
wðxÞ ¼ pffiffiffi eQ 1 þ e2Q eiðvx þ cÞ þ 1 e2Q eiðvx þ cÞ ; ð1:25Þ
x 4 4
i.e. inside the metal, the function w is the sum of the incident wave and the reflected
wave of almost identical amplitude. By definition,
2
1 14 e2Q e2Q 2Q
D¼1R¼1 2 ¼ 2 e ð1:26Þ
1 þ 14 e2Q 1 þ 14 e2Q
or according to (1.21) and (1.11)

2
rffiffiffiffiffiffiffiffiffiffiffi Zx2 3
2 12
8p 2m e
D ¼ exp42 E eEx dx5: ð1:27Þ
h2 4x
x1
It follows from (1.27) that the barrier penetrability depends on the area of the
potential curve located above this energy level E.
The integral in the exponential curve (1.27) is a complete elliptic integral cal-
culated by Nordheim through its transformation to normal elliptic integrals.
If the expression for D is applicable to the case of a triangular barrier, which is in
general case not necessarily true, then it is easy to get:
pffiffiffiffiffiffi Zx2 pffiffiffiffiffiffi 3
2p 2m pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2 2p 2m jEj2
Q0 ¼ jE þ eExjdx ¼ : ð1:28Þ
h 3 h eE
0
Table 1.1 Nordheim y h(y) s(y) y h(y) s(y)

function h(y) and function s(y)
0 1.0000 1.0000 0.55 0.6351 0.9464
0.05 0.9948 0.9995 0.6 0.5768 0.9366
0.1 0.9817 0.9981 0.65 0.5152 0.9261
0.15 0.9622 0.9958 0.7 0.4504 0.9149
0.2 0.9370 0.9926 0.75 0.3825 0.9030
0.25 0.9068 0.9885 0.8 0.3117 0.8903
0.3 0.8718 0.9835 0.85 0.2379 0.8770
0.35 0.8323 0.9777 0.9 0.1613 0.8630
0.4 0.7888 0.9711 0.95 0.0820 0.8483
0.45 0.7413 0.9637 1 0 0.8330
0.5 0.6900 0.9554
To calculate Q, we introduce function h(y), the so-called Nordheim function:
Q ¼ Q0 hð yÞ; ð1:29Þ
where
e3 E
y¼ ð1:30Þ
jEj
is the ratio of reduction of the work function due to the Schottky effect (i.e. the
mirror forces) to the “work function of an electron with given energy”. Function
h(y) was tabulated by Nordheim (cf. Table 1.1), which makes it easy to calculate Q,
and, thus, D. Graph of h(y) is plotted in Fig. 1.9.
Fig. 1.9 Dependence h(y) of θ(y)

y
1
y
0 0.5 1
Integration over vy and vz results in expression
Z1 2
4pm2 ekT mvx
j¼ ln 1 þ exp E0 =kT Dðvx ; EÞvx dvx : ð1:31Þ
h3 2
0
mv2x
Let E ¼ 2 E0 . When T ! 0, the expression becomes much simpler because

E
E
ln 1 þ e kT ¼ at E\0; ð1:32Þ
kT
and with E [ 0, this expression is equal to zero.

Hence, for the case of T 0
Z0
4pme
j¼ 3 DðE; EÞEdE: ð1:33Þ
h
1
To count the electrons emitted not only from the Fermi level E0 , but also from the
levels close to it, we expand the expression (1.29) into series in terms of E.
Eventually, that results in the expression
D ¼ D0 ekE ; ð1:34Þ
where
" pffiffiffiffiffiffi 3 pffiffiffiffiffiffiffiffi!#
8p 2m u2 e3 E
D0 ¼ exp h : ð1:35Þ
3 h eE U
In this expression, U is the work function for metal and

4p 2mU 2 dh
k¼ h y pffiffiffiffiffi : ð1:36Þ
heE 3 dy y¼ e3 E
U
The bracketed expression is later disregarded because it is very close to one. As a

result of substitution of (1.36), (1.35) and (1.34) into (1.33) and integration of
(1.33), we get:
" pffiffiffiffiffiffi 3 pffiffiffiffiffiffiffiffi!#
e3 E 2 8p 2m U2 e3 E
j exp h : ð1:37Þ
8phU 3 h eE U
The expression (1.37) is the basic formula of field emission, i.e. the Fowler
−Nordheim formula. If work function is measured in eV and electric field in V/cm,
then
" 3 pffiffiffiffi#
6 E2 6:85 107 U2 3:62 104 E
j ¼ 1:55 10 exp h : ð1:38Þ
U E U
Formula (1.38) shows that field emission depends on the electric field in the
same way as thermionic emission depends on temperature, i.e. it is greatly sensitive

to the value of the field. Just as lg Tj2 ¼ f T1 is a straight line in case of thermal

emission, the relation lg Ej2 ¼ f E1 here is an almost straight line as well. Little
deviations from the straight line are caused by presence of function h(y) in the
exponential quantity, but usually for the experimentally studied fields values of
h(y) vary insignificantly.
Field emission currents are already quite considerable at rather small D (ap-
proximately 10−10). When fields E = 6 107 V/cm and U = 4.5 eV,
j 107 A/cm2.
1.3 Classical Morgulis-Stratton Theory of Field Emission

from Semiconductors
The theory of field emission from semiconductors is based on the research by

Morgulis [25] and Stratton [26–28]. In this work we are not intending to describe
the Morgulis-Stratton theory in great detail, instead limiting ourselves to presenting
brief summary of the main results, because this theory will be shown below as
having a limited sphere of applicability. Those wishing to study Morgulis’ and
Stratton’s research papers in more detail we can point to references [5–7, 9, 17,
26–28].
Let us focus on some key differences between field emission from semicon-
ductors and field emission from metals.
The Fermi level in a metal lies mostly by a few electron-volts higher than the
lower boundary of the conduction band. That indicates high concentration of free
electrons. In the semiconductor, the Fermi level lies in the bandgap. Respectively,
the semiconductor conduction band displays a very low concentration of quasi-free
electrons. This is the most significant moment to be taken into account when
developing the theory of field emission from semiconductors.
Along with emission from the semiconductor conduction band, there may also
be emission from the valence band. Its probability increases with decrease in the
width of the bandgap (Figs. 1.10 and 1.11) [6].
The value of field emission current from the conduction band depends mostly on
electron concentration in the surficial region of a space charge. The concentration
1.3 Classical Morgulis-Stratton Theory of Field Emission from Semiconductors 21
, eV Semiconductor Vacuum
4.0
ψ δφ
U0-ρ2e2/4x
U0-eEx
U0-eEx-ρ2e2/4x
0
δR
C
ζn
g
-1.2
υ
V x
x=0
Fig. 1.10 The diagram of the potential energy of an electron near the surface of the
semiconductor in the non-degenerate case fo electron gas. The value Ev ¼ ðE þ Eg Þ—energy,
measured down from the top of the valence band Ev on the surface. Numerical values are given for
silicon
Fig. 1.11 The diagram of the

potential energy of an electron Semiconductor Vacuum
near the surface of a
semiconductor with a
degenerate electron gas
U0-|δR|+ζn
U0
C
ζn δR
F
0
x
x=0
value is affected by both the electric field penetrating into the semiconductor from
the outside and the surface charges. This effect is usually measured by the band dR
bending value. In this case, the distribution function is given by

E þ fn þ dR 1
f ðEÞ ¼ 1 þ exp ; fn [ 0; ð1:39Þ
kT
where fn is the energy-level separation from the Fermi level to the lower boundary
of the conduction band Ec in the bulk of the semiconductor. Both the
Morgulis-Stratton theory and later theoretical papers usually assumed that the bands
bend only in respect of the Fermi level, and the Fermi level remains invariable up to
the semiconductor surface (Figs. 1.10 and 1.11). Band bending is positive when the
bands on the surface are bent upwards, i.e. the electron concentration in the surficial
region of the semiconductor space charge is less than inside the bulk.
If Fermi level lies in the conduction band (Fig. 1.11 shows the degenerate case),
the Fermi distribution function is used (1.39). If Fermi level lies in the forbidden
band up to the surface (Fig. 1.10, non-degenerate case), one can use a simplified
expression (the Boltzmann distribution function):

E þ fn þ dR
f ðEÞ exp ; fn þ dR kT; ð1:40Þ
kT
where energy E is counted from the lower conduction band boundary on the
semiconductor surface. Formulas (1.39) and (1.40) regard dR as a parameter. The
band bending calculation methods have been written up in [6].
In case of semiconductors with a large specific resistance and at high field
emission current densities, thermodynamic equilibrium may be disturbed. In these
cases, we cannot use the distribution function given by (1.39).
Another important issue in development of the theory is selection of expression
for penetrability D in the equation similar to the one used to calculate dependence
j = j(E) for metals (1.2) and taking account of influence of the semiconductor band
structure on field emission.
The Morgulis-Stratton theory has been developed only for n-type low-resistance
semiconductors and for the region of relatively weak fields and currents in the way
quite similar to the one for metals. It is based on the assumption of thermodynamic
stationarity of the emission process and, hence, constancy of the chemical potential
level in the sample. The latter is only valid if just a few electrons falling on the
potential barrier tunnel in vacuum. The summed current through the sample is much
less than the drift electron current to the surface and the diffusion electron current
from the surface into the depth of the sample, so the respective approximation is
referred to as zero current approximation. In particular, zero current approximation
means that voltage drop IR of the cold emitter is zero, I being the total emission
current, R being resistance of the pointed emitter. Let us mention that for metals,
zero current approximation is valid in arbitrary strong fields, but for semiconduc-
tors, the region of its applicability is considerably narrower.
Fig. 1.12 Schematic band

vac
-U0
diagram for the
semiconductor surface, Ф—
work function, v—electron χ
affinity, eVs—band bending,
Ev —the top of the valence Φ
band, Ec —the bottom of the
conduction band, EF —Fermi eVs
level, Ed —donor level
C
With zero current approximation, to solve the problem of the structure of the
surficial region of the space charge and field distribution in the sample means to
solve the Poisson equation with set boundary conditions.
However, when using the ratio (1.39) to calculate field current from the semi-
conductor, one has to bear in mind that the Nordheim function argument is to be
further multiplied by
m ¼ ½ðe 1Þ=ðe þ 1Þ1=2 ;
where e is the static dielectric constant of the semiconductor.

Construction of the theory of field emission from semiconductors should involve
not only the work function concept, but also the electron affinity concept
(Fig. 1.12).
A semiconductor experiences an additional band bending effect and the total
work function is given by
U ¼ v þ eVs þ ðEc EF Þ;
where v is electron affinity, eVs describes band bending, and the term ðEc EF Þ
denotes the difference in energies between the Fermi level and the conduction band
bottom in the bulk. The dipoles caused by adsorbates affect the first two terms,
while semiconductor doping affects the third term.
The integrand in (1.2) has a sharp maximum located either in the neighborhood
of E ¼ EF or near the conduction band bottom. In the neighborhood of this
maximum, the integrand is expanded into the Taylor series and integrated term by
term. Taking into account that electron energy distribution of the conduction band
NðEÞ can be determined in several approximations, Stratton calculation results can
be summed up as follows:
1. If electron gas in the conduction band is not degenerate, penetration of the

external field in the sample is neglected, then
1=2
1 vaf E
NðEÞ ¼ n1 exp ; ð1:41Þ
2pmkT kT
where n∞ is the equilibrium concentration of electrons in the bulk, vaf is the affinity
barrier.
For this approximation, we can get:
" pffiffiffiffiffiffi 3=2 ( )#
kT 1=2 8p 2mvaf e 1 1=2 e3=2 E1=2
j ¼ en1 exp H : ð1:42Þ
2pm 3heE eþ1 vaf
Where H is the Nordheim function, e is the dielectric constant.

2. If electron gas in the conduction band is not degenerate, we should take the
shallow-penetration field into account. To take the field penetration into account,
we should consider not the equilibrium concentration of electrons in the bulk,
but surface concentration n.
We assume that emission current does not practically disturb this distribution,
which then can be given according to the Boltzmann formula:

DR
n ¼ n1 exp ; ð1:43Þ
kT
h ekT 1=2 i
where DR ¼ 2kT arsh eE
2ekT 8phe2 [6]. Similarly to the first case, we derive a
formula for current density, where bulk concentration is replaced with the surface
one.
3. If electron gas in the conduction band is degenerate due to field penetration,
NðEÞ is determined in the same way as for a metal. However, here we deal with
constantly incrementing surface electron concentration due to incrementing
band bending. So we propose to replace the semiconductor work function with
the one determined according to the formula
v0 ¼ vcp bðeEÞ4=5 ; ð1:44Þ
where b is a constant dependent on the semiconductor properties. In this approx-

imation given vaf [ bðeEÞ4=5 we get the following expression for field emission
current density:
" pffiffiffiffiffiffi 3=2 !#

e ðeEÞ2 8p 2mvaf ðeE Þ1=2 j0 bvaf
j¼ exp H 1 1þ2
8ph vaf heE vaf ðeEÞ
" 1=2
#
k0 bvaf
exp 2 :
eE 1=5
ð1:45Þ
4. Stratton also derived the emission current formula with account of surface states.
We are not going to present it in its complexity, but rather illustrate the main
results of the Stratton theory, including those attained with account of emission
from surface states (Fig. 1.13). Figure 1.13 prompts that, firstly, current voltage
characteristics of density of field emission current lg j = f (l/E) from an electron
semiconductor are shown with straight lines in all approximations, except for
the case when surface states have to be taken into account. Figure 1.13 shows a
few calculated current-voltage characteristics for SiC (curve 1 according to
(1.42), curve 2 according to (1.44), curve 3 according to the formula taking
surface states into account). The bending of current-voltage characteristics
arising in the presence of surface states is caused by the following. When fields
are small, field emission becomes lower due to the inner decelerating barrier. In
the process of cancellation of this barrier by a penetrating field, emission begins
growing in an accelerated way. Along with the external field growth, the
emission growth causes a simultaneous increase in both potential-barrier pen-
etrability and electron concentration in the surface layer.
According to (1.42), the field emission current density is proportional to the
electron concentration in this layer, i.e. j * n, if the electron gas of the conduction
band in the surficial layer is non-degenerate.
If field penetration is disregarded, n = n∞. While if we take it into account,
n depends on n∞ according to (1.43), i.e. in a more complex way. According to
(1.43), field emission current also depends on temperature [mainly, due to the
dependence n = f(T)]. For degenerate electron gas according to (1.44), j is not an
explicit function of the temperature.
To conclude this brief summary of the results derived in zero-current approxi-
mation, we should say that for the case of field emission from low-resistance n-
semiconductors and in the field of weak fields and currents, the Morgulis-Stratton
theory is satisfactory for description of experimental facts (linearity of
current-voltage characteristics, absence of photosensitivity and weak thermosensi-
tivity) and allows calculation of emission-current density. However, experimental
research has not identified any dependences lg j = f(1/E), e.g. curve 3 in Fig. 1.13.
The straight current-voltage characteristic slope constructed in coordinates lg j,
E−1 (Fig. 1.13) is determined by affinity to the electron, the bandgap width and
degeneration parameter. In a considerable concentration of surface states, the
current-voltage characteristics of n-type semiconductors can display characteristic
bending (Fig. 1.13). Their effect is that when intensity of external field E is low, the
lgJ
0
2
-1
-2
-3
1
-4
-5
-6
-7
-8
3
-9
-10
2.5 3.0 3.5 4.0 4.5 5.0
0.4 0.35 0.3 0.25 0.2
Fig. 1.13 The calculated current-voltage characteristics for SiC: 1—of formula (1.42), 2—of
formula (1.45), 3—calculation that takes into account the surface states
surface-charge field bends the allowed energy bands upwards, and emission initially
comes from the valence band. As soon as the field reaches the value when the
surface-charge field is canceled, the bands start bending contrariwise, conductance
electron concentration goes up and field emission from the conduction band
becomes prevailing. On the current-voltage curve, emission moves from the valence
band to the conduction band, which is accompanied by upsurge of current.
First attempts to develop a theory of field emission from n-type high-resistance
semiconductors and p-type semiconductors were not made until after it had been
established that current-voltage characteristics in this case are non-linear, and
emission is photo- and heat-sensitive (e.g. [5–7, 29, 30]).
As for the theory of thermofield emission (TFE) from semiconductors, which
was most fully developed by Vasiliev [31], it has no practical sense, as the
experiment shows (e.g. [29]) that heating of a semiconductor field emission cathode
to quite low temperatures * 200 400 °C results in significant distortion of its
initial properties. For this reason, presentation of even some particular elements of
the theory of thermal-field emission from semiconductors is of no special sense,
which is why we are not going to do that.
1.4 Basics of Müller Field Emission Microscope Theory
The main element of a field emission projection microscope, an outstanding

invention by Erwin Müller (e.g. [5, 8, 17, 18, 9, 32–35]) that enabled to verify key
conclusions of the first theory of field emission from metals, is an electron beam
source, viz. a field emission cathode (emitter). E. Müller suggested shaping it as a
sharp tip. That shape of the field emission cathode enabled E. Müller and people
continuing his work not just to achieve intensive electron emission, but also to
examine a surface from tip emitters in an electron microscope projector with res-
olution 20 30 Å (E. Müller’s later invention, a field ion microscope [35],
allowed to increase the resolution up to 2 3 Å) and visually observe the pro-
cesses of adsorption, migration, diffusion etc. within the first and the second
monoatomic layers at relatively low voltages ranging from one-to-two digit
kilovolts.
A wide application of the field emission microscope in studies of surfaces of
metals as well as of other materials (most importantly, semiconductors) is caused by
obvious advantages of the field emission method in comparison with other methods.
(1) In principle, this technique does not apply any restrictions on the working
temperature and enables to study an object at the temperatures ranging from
0 K up to the melting point.
(2) A field emission microscope can directly produce an image of the sample tip
surface. As mentioned above, the resolution reaches 20 30 Å. At present, the
identical resolution power has been achieved in a transmission electron
microscope, as well as in the scanning tunneling and atomic force microscopes.
As compared with field emission methods, those devices are unable to keep
track of kinetics of the processes taking place at the surface, they can only
register some intermediate states of the process. Some detailed surface structure
information can also be received by the low-energy electron diffraction (LEED)
method. A drawback of this method as opposed to field emission technique is
that it can study only one crystallographic plane at a time, and this plane has to
have a long-range order structure.
(3) The field emission method enables to monitor changes of these values due to
exponential dependence of emission current on the work function and the
intensity of the electric field. That makes the field emission microscope espe-
cially advantageous for studies of adsorption and surface migration.
(4) A strong external electric field and high temperature change the emitter shape,
and that is easy to detect both by emission pattern and by change of the
emission current. By the emission pattern one can directly see the exit of
impurities out of the bulk of the crystal.
(5) A field emission microscope can be used to study the phenomena of epitaxy,
crystal irregularities, emergence of dislocations etc.
(6) Field emission microscopy allows to evaluate the work functions, i.e. to make
quite accurate measurements of relative changes in work functions for various
crystallographic planes, to evaluate surface tension factors, activation energy
for migration, energies of adsorption and desorption and to determine
energy-distribution half-widths with high resolution etc.
However, all those advantages of the field emission method do not come without
drawbacks:
(1) Field emission results from the tunnel effect in electric fields of about
107 V/cm. As a rule, the emitter is shaped as a very thin tip with a curvature
radius of 10 100 nm on the top to provide the electric field intensities
necessary for emission. That makes it impossible to take measurements of the
same object by other methods and compare the results.
(2) The phenomena taking place on the field-emitter surface can only be observed
under very strong electric fields. Extrapolation of the results to the case of
infinitely small field is most often incorrect.
Figure 1.14 shows a diagram of a model of the most primitive field emission
microscope (a field emission diode). The anode is almost spherical as the shape of
the anode is of no importance. The emitter tip (terminus) is situated in curvature
center of the anode. Electrons emitted from the termination have very little initial
kinetic energy and their paths are determined by spatial distribution of the applied
acceleration field, this distribution being almost spherical in immediate proximity to
the hemispherical surface of the tip. That is the case for a metal emitter, whose
surface is an equipotential of an electric field and is directed perpendicular to its
power lines. The field value decreases in a close proportion to the squared radial
distance from the terminus. So the electrons acquire their terminal velocity at a
distance of several peak radii from the emitting surface, going further to the anode
along the radial paths [23, 36]. The anode is phosphor-coated, which turns it into a
luminescent screen. Falling electrons activate the screen and it generates a mag-
nified image of the terminus surface. The magnification factor is proportional to the
ratio of the distance between the terminus and the screen to the tip radius; its
practical value is usually in an order of 105.
In Fig. 1.15 one can a simple representation of a field electron motion trajectory
as a continuous curve. Let us note that this diagram is not in a true scale. In reality,
the distance from the emitter terminus to the screen equals several centimeters,
which is almost 105 times as long as the radius of the terminus. The radial and
tangential components of the initial electron velocity are designated with иr and иt.
Since иt is non-zero, the electron reaches the screen in a point biased by the value of
1.4 Basics of Müller Field Emission Microscope Theory 29
Fig. 1.14 The scheme of Fluorescent screen

field emission diode
Collector
Anode
Probe hole
Tip emitter
Fig. 1.15 The trajectory of

the field emission electron
D/2
Ut
rt
δ/2 Ur
D/2 from the radial projection of the point of field-electron escape from the surface
of the cathode. The calculations made by R. Homer [37] showed that
D=2 2hut it; ð1:46Þ
where 〈ut〉 is a mean tangential component of the initial velocity of emitted elec-
trons and t is given as
t xð2eV=mÞ1=2 ; ð1:47Þ
where x is the distance from the terminus to the screen, V is the voltage applied
between the emitting cathode and the anode. 〈ut〉 can be estimated as
hut i ¼ ð2hEt i=mÞ1=2 ; ð1:48Þ
where hEt i is a mean tangential component of field-electron energy on the emitting

surface. Direct calculation of this value (in electron-volts) on the basis of the
Fowler–Nordheim field emission theory gives
1.0
оо ++
оо
0.8
++
0.6
M(θ)/M0 2
β(θ)/β0
0.4
0.2
2 20 40 60 80 100 θ, degrees
Fig. 1.16 Dependence of the local field factor b (circles) and the multiplication M (crosses) of the
azimuthal angle h at h = 0, b = b0, M = M0. The dashed curves are calculated for two models of
the sphere on the orthogonal cone, simulating two different emitter profiles: with a pronounced
(1) and small (2) compression. The solid curve corresponds to the average emitter

hEt i ¼ 0:433Et1 3:79F 1=2 =U U1=14 ; ð1:49Þ
where t is the elliptic function, U is the effective work function, E is the field on the
tip surface (in volts per ångström), which can be found from the ratio
E ¼ bV V=krt ; ð1:50Þ
k 5 at the emitter terminus and grows with increase of the polar angle
(Fig. 1.16).
Figure 1.14 shows that the observer looking at the screen perceives an arbitrary
small segment on the tip surface as stretched by the value of
d=2 ¼ D=2M; ð1:51Þ
where M is a field emission microscope magnification factor. If the emitter terminus

was perfectly spherical, factor M would be equal to x/rt. In practice, presence of the
emitter trunk impresses the field lines towards the termination axis, thereby
reducing the magnification effect. It has been established that for real emitters
M ¼ x=crt ð1:52Þ
where c 1.5. From (1.47) to (1.52) we can find that the resolution of the field
emission microscope (in ångströms) is

1=2
d 2; 62c rt =ktU1=2 ; ð1:53Þ
where rt is the tip radius (in ångströms). It is estimated that for a typical emitter it is
about d 25 Å.
Along with (1.53), there is another condition that constraints the resolution of a
field emission microscope. It is caused by the Heisenberg uncertainty principle,
from which follows that if an electron is emitted from the region of width d0 it has
to have the minimum tangential component of velocity of at least ut ħ/2тd0. This
principle imposes an irremovable limitation on resolution of any microscope.
However, the estimates show that in most cases this limitation can be neglected as
the resolution estimate is not much different from the estimate derived from the
ratio (1.53).
As to identification of the basic crystallographic faces displayed on the lumi-
nescent screen of the field emission microscope (Fig. 1.17), it is carried out with
standard stereographic (or pseudo-stereographic) projections [18]. For the basic
stereographic structures (BCC, FCC, HCP), they can be found in monographs and
textbooks (e.g. [38]) or manuals on crystallography. When deciphering structure
images, they have to be calculated on computer. In the field emission microscope, a
sample (a tip field emission cathode) is shaped and deformed in the process of
surface image acquisition and the deformation nature can be of two kinds, both
elastic and plastic. If the field meeting the conditions of the best mapping of the
sample surface equals E0, then the crystalline lattice of the sample is elastically
extended and deformation
e0 rE0 =K; ð1:54Þ
(a) (b) (c)

112
011 112
301
121 310
111 111
010 102 012
100 211
112
112 310 121
001
011
301 110
031
121 102
012 211
130
111
001 010
112 101 111 121
112
031
112
Fig. 1.17 Examples of typical field emission images of the tip surface of tungsten (a, b) and
niobium (c): a—the surface is smoothed high-temperature heating (T0 ≅ 0.7 Tmelting) without a
field; b, c—surface reconstructed with a slight warming (T0 ≅ 0.2 Tmelting) in the presence of an
electric field
where K is the volumetric modulus of elasticity usually equal to a few percent. The
respective value of the elastic energy accumulated in the sample per atom equals
Eat ¼ Xr2 =2K; ð1:55Þ
where Xa is atomic volume. Since Eat is usually of the order of 0.1 eV, i.e. com-
parable with activation energy of low-temperature diffusion processes, impression
of an electric field in the field ion microscope can considerably change the nature of
pin-hole defect annealing.
Approximating the tip shape with a lathe object having symmetry axis z and
neglecting particular atomic and electronic surface structures acting on the plane
z = z0, we can express force Fz as
Z Z
Fz ¼ ½F 2 =ð8p cos aÞdSe ; ð1:56Þ
where dSe is an element of the tip surface, a is an angle between the normal line to
this element and axis z. Integration is performed over the entire surface, i.e. from the
tip terminus to the chosen section plane. As a result, (1.56) can be represented as
follows:
Zr0 ZF0
Fdr F2
Fz ¼ 2pdl ¼ rdr; ð1:57Þ
8pdl 8p
0 0
where dl is the linear element of the surface, r0 is the distance from the tip centre to
z 0.
Another “concurrent” process is bombardment of the sample surface with
low-energy (several dozen kiloelectronvolts) ions of residual gases. In absence of
an electric field, the tip surface experiences n0 gas atom impacts per time unit
caused by their mere thermal motion:
1=2
n0 ¼ S0 Pg 2pMkTg ; ð1:58Þ
where S0 is the tip surface area; M is the gas particle mass; Pg is the gas pressure
and Tg is its temperature. In the electric field of intensity E, the gas-molecule flow
from the tip surface increases п* times, which is connected with polarization of the
molecules. n* can be approximated from the equation

n pag E 2 = 2kTg ; ð1:59Þ
where ag is gas-molecule polarisability.

As is known [18], the number of impacts nu experienced with the entire tip
working surface per time unit can be estimated by the formula
X X Pi IFE

nu ¼ nu;i ¼ 1; 44 R0 ri ðWÞ ; ð1:60Þ
i i
kT0 e
where Pi is the partial pressure of the residual gas components; k is the Boltzmann
0 is the mean radius of
constant; T0 is the tip temperature; e is the electron charge; R
the tip surface curvature; ri (W) is the section of ionization equal to
Ne pa20 W1
ri ðWÞ ¼ 2
2:66 lnð1:25WÞ; ð1:61Þ
Vion W2
where Ne is the number of electrons in the outer shell of the ionizing atom of the
residual gas, a0 is the first Bohr orbit radius, Vion is the atom ionization potential,
and W ¼ Ee =Vion , where Ee is the electron energy equal to the value of the potential
in the atom ionization point.
The bombardment of the field-emitter surface with low-energy ions of residual
gases can have both negative and positive value. In the practice of operation of
cathode devices on the basis of field emission (in the same way as in the case of
research with field emission microscope), it can result in a change of the surface
structure, pulverization or redistribution of the component (film) activating electron
emission, vacuum breakdown and a total outage of the emitter (sample). At the
same time, the very possibility to irradiate the sample in the field emission
microscope looks undoubtedly fruitful and gives a wide range of opportunities.
Speaking about the bombardment of tip surfaces with low-energy gas ions
functioning in the field emission mode, it has to be noted that such a bombardment
is among the main causes of low-frequency emission noises. As such, let us also
dwell on the issue of the noises a bit.
The current emitted by the sample in the field emission microscope is somehow
labile: in addition to the usual “schrot” noise component natural for any kind of
particle emission, the field emission current has another relatively low-frequency
noise component. The change of the spectral power density W(f) of the latter
component is close to dependence 1/fc, where c ranges between 0.8 and 1.2. That is
the so-called flicker noise of field emission.
As is known, when studying fluctuation of the current, the main values to be
analyzed were the correlation function (auto-correlation) f*(s) and the noise-power
spectral density function W(f):
f ðsÞ ¼ hI ðtÞI ðt þ sÞi;

ð1:62Þ
W ð f Þ ¼ dI 2 =df :
These values are brought together in the Wiener–Khinchin formula
Zþ 1
1
Wðf Þ ¼ f ðsÞ expð2pIf sÞds: ð1:63Þ
2p
1
In addition, the ratio between the mean square of the current intensity fluctuation
ðDIÞ2 , the correlation function and the noise-power spectral density function is
given by the expression
Z1
2
DI ¼ f ð0Þ ¼ Wðf Þdf : ð1:64Þ
0
It is widely agreed that the main cause of field emission flicker noise is the
presence of adsorbed particles on the field-emitter surface and their irregular surface
movement [39].
Another “concurrent” process that has already been mentioned is the
thermal-field change of the tip surface shape. The joint action of heat and the
electric field (the latter manifesting itself through surface atom polarization and
movement of the resulting dipoles in the locally-enhanced field region, whereas
heating of tips enables, facilitates and boosts such mobility) can significantly
change the tip surface shape. This change has to be controlled and taken into
account when studying these or those phenomena in the field emission microscope.
Real surface geometry. Various authors successfully resorted to field emission
microscopy to analyze both the equilibrium shape of the metal microcrystal surface
at different temperatures in and out of an electric field, and to study the proper atom
condensation processes on an approximately hemispherical substrate. Below we’ll
give a consistent consideration of the most important results attained in these
aspects.
To begin with, the field emission microscopic image itself easily enables to
establish the mean curvature radius R 0 of the tip terminus. In case of field emission
microscopy, R 0 is determined by measuring of dependence of field emission current
IFE on tip potential V0. Having expressed the Fowler−Nordheim equation in terms
of IFE and V0 as
" !#
IFE a 0 cV01=2
bR
¼ 2 exp t ; ð1:65Þ
V02 R 0 V0 1=2
R 0
where parameters a, b and c are constant for this material and connected with the
value of the mean work function U, t is a tabulated function, we can see that
2
dependence ln(IFE/V0) of 1/V0 is close to the straight line up to ln a=ðR 0 Þ2 and its
slope equals bR 0 t. At t 1, value R
0 can be approximately determined from the
graph ln(IFE/V20) = f(1/V0) by the line slope. The determined value R0 can be later
used to calculate the argument of function t and then value t can be determined
from the table. Having inserted the new value t in the ratio bR 0 t equal to the
straight line slope, we can determine a more precise value R 0 etc. Precision of this
method of finding R 0 is ±15%. It cannot be improved in view of essential varia-
tions of the work function and, consequently, IFE along the surface of the tip
terminus.
As shown in [40], impact of temperature leads to spheroidisation in tip apex.
Let us note that the mass transfer leading to evolution of the tip terminus profile
is most probably caused [18] by surface self-diffusion. This seems feasible given
the small size of tip apex.
At the same time, we can see an absolutely different pattern of evolution of tip
terminations when they are heated in the presence of an external electric field [18].
In this case, apart from the capillary forces we have to take into account the
ponderomotive ones as well. In this case, the chemical potential of atoms at the
needle tip can be given by:
l ¼ l0 þ xð2c=R E 2 =8pÞ; ð1:66Þ
The minimum value of voltage V0 (tip potential) at which the electric field can
cause a considerable change in the tip shape against the background of capillary
forces can be estimated with the following expression:
V0 ð16pcR=vÞ1=1:4 : ð1:67Þ
Dependence V0* = f(R) means that in the course of needle elongation under the
influence of the field, the termination curvature radius decreases and, consequently,
the difference between the set value of potential V0 and its critical value increases,
which means that elongation of the tip has to be a self-accelerating process.
Analysis of thermal-field and condensation growth patterns of metal crystal tips
by field emission microscopy is presented in more detail in [39–41].
1.5 Disadvantages of Classical Field Emission Theories

and Limits to Their Applicability
Simplicity of presuppositions of the Fowler–Nordheim theory and at the same time

quite satisfactory description of the real phenomenon by this theory caused both
serious efforts to analyze its appropriateness [5–9, 19, 20, , 37, 42, 43] and
numerous attempts to extrapolate it to less idealized models. Essentially, all
researchers involved considered the limits to applicability of the Fowler–Nordheim
theory and possibilities to work around them with some improvements to the
theory. Most papers improving the Fowler–Nordheim theory focus only on the
extremely high current-density region. This region has some specific physical
distinctions, and, at the same time, it is extremely interesting from the standpoint of
some important applications of field emission. Because of this, we are going to
analyze it in greater detail.
In addition, the classical Fowler–Nordheim theory considers only the case of
“pure” field emission (which, strictly speaking, only holds for T = 0 K). It does not
analyses theoretically interesting and practically important cases of thermofield
emission for T > 0 K nor for T 0 K.
Noteworthy is the following often neglected circumstance. The basic formula of
the Fowler–Nordheim theory does not link the voltage and current, but rather the
electric intensity and current density, i.e. basically unmeasurable values, especially
if the tip dimensions are microscopic. The most common workaround for this
probles is to assume that the field intensity value is proportional to voltage and the
current value is proportional to its density, i.e.
E ¼ bV; I ¼ Sj; ð1:68Þ
where b is the so-called field factor and S is the emitting area. The first relation does
not usually cause doubt if values b and E are understood as their local values in an
emitting surface point. As for the second equation, it is more precisely written down
as
Z
I¼ jðx; y; zÞdS; ð1:69Þ
ðSÞ
where S is the total surface of the emitting tip.

This issue is of both academic and quite practical nature because these values
determine functioning of major vacuum-microelectronic devices, e.g. a flat display
[13, 43]. In particular, [44] proposes to determine the line inclination tangent and
the segment cut by the line on the ordinate axis with a line by Fowler–Nordheim
approximation in the coordinates 1/V, ln(I/V2) and then calculate the emission area
with a margin error of at least 15% and the field factor with even lower precision by
the ratios

S ¼ 4 1017 I=V 2 1=V¼0 p2 ; ð1:70Þ
where p = −2.82 109(Ф3/2/b) is a Fowler–Nordheim line inclination tangent.

The current is expressed in amperes, the length in meters and the work function in
electron-volts. The first ratio is based on approximate constancy of value Ф2exp
(9.81Ф−1/2). The second one needs knowledge of value Ф itself, which results in
decrease in precision as the local work function value is measured with great
difficulty and low precision.
Nevertheless, most theories assume strict observance of these equations and
scrutinize only the Fowler−Nordheim equation. The equation is criticized in three
aspects. Firstly, the physical prerequisites of its deduction seem to cause some
1.5 Disadvantages of Classical Field Emission Theories … 37
doubts that are not entirely unbased [45, 46]. Analysis of this type of criticism is
beyond the purpose of our study, though some issues concerning the physical
prerequisites of tunneling calculation with the WKB method will be considered in
further chapters. Secondly, correctness of substitution of the elliptic integrals as
components of the full formulas with their approximate values is open questionable
and improved expressions are being proposed [47, 48]. Thirdly, doubt is caused by
the formula parameter determination methods based on construction of current-
voltage characteristics in coordinates {ln(I/V2), l/V} with approximation by the least
square method [44].
The critics’ arguments are actually quite feasible. Indeed, application of the least
square method is connected with a particular (Gaussian) kind of random error
distribution [49]. It is known that any transformation of variables violates the initial
formula for the distribution function [50]. However, it still remains to be seen how
significant this violation is. This question is even more important if we note that
there are a lot of approximations initially built into the model, which may cause
much more serious errors. It is also important that according to the Neumann-David
theorem, the estimate of the linear-regression coefficient by the least-square method
is asymptotically unbiased and effective irrespective of a particular formula of the
error-distribution function [51]. Certainly, modification of the formula of the
error-distribution function causes change in the estimates of confidence intervals.
However, in this case, we should speak not about the biased estimate of the physical
model parameters, but rather about determination of the confidence interval for this
estimate. We cannot but note that the latter is very seldom, if ever, done.
The papers we know of have not payed enough attention to influence of an error
in the independent variable. The classical least square method assumes that a
random error burdens the measurements only in case of a dependent variable, while
the independent one is considered to be measured in an absolutely precise way.
Certainly, that is not the case. In practice, voltage measurements also bear a random
error comparable with a current-measurement error (of course, in relative terms).
Thus the usual formula of the least square method is mathematically unreasonable
here. What we need is a more complex model that’s still based on minimization of
residual sum of squares but taking into account the errors of both variables. So the
first stage of work should be analysis of applicability of Fowler−Nordheim for-
mulas or their improved analogues, comparative estimate of efficiency in terms of
computer time consumption and, finally, definition of the physical sense of the
values obtained. The last stage is also important if the results attained are not the
end in itself or a reference point for relative comparisons but the initial data for
further mathematical or natural simulation or even practical use.
For such analysis, we developed an information expert system [20] and carried
out a natural experiment on measuring of dependence I(V) in a wide range of
current voltages and densities, the latter exceeding five orders of magnitude. The
results attained were processed in various ways:
1. Construction of a line in coordinates {ln(I/V2), 1/V} by the usual least-square

method [52].
2. Construction of a line in coordinates {ln(I/V2), 1/V} by the least-square method
with account of an error in the independent variable (1/V) [14].
3. Minimization of value v2 in direct approximation of the functional dependence
(this was carried out in Origin software package) of non-linear regression [16].
In this case, the trial function was represented by both the classical Fowler–
Nordheim expression and more complex expressions.
4. Application of nonparametric-regression methods, which do not require from
the error distribution function to have any particular formula [53], except for its
symmetry about zero.
The results are shown in the figures below. Figure 1.18 shows dependence of a
field emission current on voltage in natural coordinates as a semilogarithmic graph
and a non-linear-regression curve optimal in terms of v2 minimization. Table 1.2
shows Fowler–Nordheim line parameters obtained by various methods and meeting
the confidence intervals for 95% significance level. For a direct non-linear
regression, we obtained estimates of coefficients P1 and P2 of the formula
I = P1V2exp(–P2/V). The estimates were optimal in terms of v2 minimization and
we re-calculated them into the respective parameters of the Fowler–Nordheim line
for convenience of comparison. The table proves that it is only non-linear regres-
sion that gives significantly different parameters of the line in terms of statistics.
However, this result is the least satisfactory from the physical point of view.
The matter is that when value v2 is minimized, small values of the dependent
variable cannot compete with large values. Practically, it is obvious that the
regression curve does not take them into account and the differences reach half of
10-3
10-4
10-5
ln(I, A)
10-6
10-7
10-8
6000 8000 10000 12000 14000 16000 18000 20000
V, V
Fig. 1.18 The dependence of the field emission current from the voltage: the experimental points
and the curve of non-linear regression
1.5 Disadvantages of Classical Field Emission Theories … 39
Table 1.2 Regression-curve parameters ln(I/V2) = A(I/V) + B obtained for the available
experimental points by various methods and their confidence intervals with 95% confidence level
Type of regression A DA B DB
Classical −1.158 105 0.007 105 −18.85 0.07
With error V −1.161 105 0.007 105 −18.81 0.08
Non-linear −1.052 105 0.02 105 −19.67 0.09
Non-parametric −1.143 10′ 0.03 105 −19.0 0.1
the order of magnitude. It is natural when a dependent variable varies in a large

dynamic range. That is why the non-linear regression is unsatisfactory in this
application. All other methods are indistinguishable. So it is natural that the one to
be preferred should be the simplest and traditional, for which there are standard
program packages available. However, comparison with the nonparametric-
regression results shows that precision of their estimates is usually overrated.
From the procedure of determination of coefficients S and b, especially popular with
experimenters, it is obvious that all the methods applied result in the values just
slightly different from each other. The maximum difference of the value of emission
area S is no more than 20% (calculation by traditional linear methods gives
5.2 10−16 m2 and non-linear regression results in 6 10−16 m2), i.e. the values
of errors all lye within the margin of error of the model. By the basic rules of
measurement theory, the prospective model accuracy should considerably exceed
the accuracy of determination of its parameters [54].
That means that the experiment does not make a statistically significant dis-
tinction between different mathematical models of the process. Thus, feom this
point onwards we are going to use only the processed Fowler−Nordheim equation
in its usual formulation, i.e. a traditional linear regression in coordinates {ln(I/V2),
1/V}. However, the problem of unreasonable extrapolation of local equation to the
integral characteristics persists and so does the physical essence of the parameters
determined thereby.
Let us turn to discussion of the semiconductor field emission theory described in
previous sections. As a preliminary and in addition to what has already been said in
1.3, let us give a brief summary of the basic assumptions underlying this theory.
(1) For the transmission factor D, we used the expression known from the metal
field emission theory with account of a more general form of the image-force
potential. In this case, we disregarded the possible contribution of relaxation
effects to this potential.
(2) The semiconductor electrons were considered to be quasi-free. Further on, we
assumed the same dependence EðkÞ inside the crystal and on the surface
region, which is closely connected with the issue of preservation of crystalline
symmetry up to the very surface. However, that assumption about function
EðkÞ was not verified, e.g., for germanium and silicon [6, 7].
(3) Everywhere, we proceeded from the assumption that electrons are in thermo-
dynamic equilibrium with the crystalline lattice.
(4) The influence of the surface charges and field penetration on concentration of
electrons in the surface region was described in terms of band bending. Possible
quantization of electron states in the bulk-charge band near the surface was not
taken into account.
(5) We discussed diversions from thermodynamic equilibrium. The results fur-
nished were attained in the assumption that it is only acoustic-phonon scattering
that matters. We considered field emission from n-type semiconductors without
account of surface states. However, the experiments found out hot-electron
emission from p-type semiconductors as well [6, 7].
(6) The influence of effective mass was studied in the assumption that emission
preserves the tangential component of the wave vector. Essentially, the latter
matters when band boundaries in the space of wave vectors k lie at k 6¼0, e.g. in
the case of germanium and silicon. Preservation of the transverse wave-vector
component was admitted both in the semiconductor photoemission theory [9]
and in the general consideration of penetrability. Silicon photoemission
experiments [9] justify such an assumption for at least a large number of
emitted electrons.
For the sake of theory we’d prefer to preserve the transverse wave-vector
component unless the translation symmetry of the crystal is broken on the surface.
That is observed in the case of a flat boundary, when external forces are perpen-
dicular to the surface [5–7].
Silicon field emission experiments [6] indicate possible non-preservation of the
transverse wave-vector component. We gave the following prospective reasons
thereof:
• diffused scattering of electrons on the surface due to its roughness;
• violation of periodicity of the crystal potential on the surface;
• interaction with surface phonons.
Stratton explored this issue and found out that only simultaneous presence of
first and second factors can result in non-preservation of the transverse wave-vector
projection.
It is clear from what has been said that the classical Morgulis-Stratton semicon-
ductor field emission theory, however improved and enlarged in a number of later
works it might be, is still connected with a lot of essential simplifications. Therefore,
it is no wonder that the theory is still unable to explain a lot of experimental results in
terms of quantity. In a lot of cases, qualitative explanation is constrained as well.
References
1. N. Mott, J. Sneddon, Wave Mechanics and Its Applications. M.: Nauka, (1966), 427 p
2. A. Messia, Quantum Mechanics (in 2 volumes). M.: Nauka (1979)
3. R.H. Fowler, L. Nordheim, Electron emission in intense electric field. Proc. Roy. Soc. A. 119
(781), 173–181 (1928)
References 41
4. L. Nordheim, Die Theorie der Elektronemission der Metalle. Physikalische Leitschrift.

(1929). Bd. 30. N7. S.117–196
5. A. Modinos, Field, thermo and secondary electron spectroscopy. M.: Nauka, (1990), 320 p
6. R. Fischer, H. Neumann, Field emission of Semiconductors. M.: Nauka, (1971), 216 p
7. M.I. Elinson (ed.), The cold cathodes. M.: Sov. Radio, (1974), 336 p
8. E.P. Sheshin, Surface structure and field emission properties of carbon materials. M.:
Publishing House of the MIPT. Fizmatkniga (2001), 287 p
9. L.N. Dobretsov, M.V. Gomoyunova, Emission Electronics. M.: Nauka (1964), 364 p
10. M. Green (ed.), The Surface Properties of Solids. M.: Mir, (1972), 432 p
11. A.I. Anselm, Introduction to the Theory of Semiconductors. M.: Nauka (1978), 616 p
12. M. Szilagyi, Electron and Ion Optics. M.: Mir, (1990), 639 p
13. L.D. Landau, E.M. Livshits, Mechanics. M.: “Nauka” (1965), 204 p
14. L.D. Landau, E.M. Livshits, Statistical Physics. Part 1. M.: “Nauka” (1966)
15. Landau L.D., Lifshitz E.M. Quantum Mechanics. Non-relativistic theory. M.: Nauka (1974),
752 p
16. N.D. Lang, W. Kohn. Theory of metal surfaces. Phys. Rev. B. 1(12), 4555–4568 (1970)
17. M.I. Elinson, G.F. Vasiliev, Field emission. M.: Fizmatgiz (1958), 272 p
18. A.L. Suvorov, The structure and properties of the Surface Atomic Layers of Metal. M. :
Energoizdat (1989), 296 p
19. D.A. Ovsyannikov, N.V. Egorov, Mathematical modeling of systems for formation of
electron and ion beams. SPb.: Publishing of St. Petersburg State University (1998), 276 p
20. N.V. Egorov, A.G. Karpov, Diagnostic Information and Expert Systems. SPb., St. Petersburg
State University Publishing House (2002), 472 p
21. G. Wentzel, Zeits. f. Phys. (1926), 38, 516
22. H.A. Kramers, Zeits. f. Phys. (1926), 39, 828
23. L. Brillouhin, Comptes rendus. (1926), 183, 549
24. H.A. Kramers, Zeit. A. Phys. (1926). Vol. 38. p. 516
25. N.V. Morgulis, To a question about the effect Schottky for compound semiconductor
cathodes. ZETP, 16(11), 959–963 (1946)
26. R. Stratton, Field emission from semiconductor. Proc. Phys. Soc. B 68(430B), 746–757
(1955)
27. R. Stratton, Theory of field emission from semiconductors. Phys. Rev. 125(1), 67–82 (1962)
28. R. Stratton, Phys. Rev. 135 (1964)
29. N.V. Egorov, V.R. Tolstyakov, Investigation of the effect of the surface state on the emission
characteristics of semiconductor photo field cathodes. Surface (1996). (9), 10–13
30. N.V. Egorov, V.R. Tolstyakov, The effect of multi-particle tunneling in the field electron
emission from semiconductors. Surface (1996). (9), 10–13
31. G.F. Vasiliev, Radiotehnika i Elektronika, 3(7), 962 (1958)
32. E. Mueller, T Field. Tson. Ion microscopy. M.: Metallurgia (1971), 360 p
33. E.V. Mueller, M. Sauton, D. Brandon et al. Field ion microscopy, ed. by J. Ren,
S. Ranganathan. M .: Mir (1971), 270 p
34. E.W. Mueller, Phys. Z. (1936). 37, 838
35. E.W. Mueller, W. Z. Electrochem, (1957). 61, 43
36. S.A. Fridrihov, S.M. Movnin, Physical Fundamentals of Electronics. M.: Vyshaya Shkola
(1982), 608 p
37. R. Gomer, Field Emission and Field Ionization. (Harvard University Press, Cambridge, MA,
1961)
38. R. Knox, A. Gold, Symmetry in Solids. M.: Nauka, (1970), 424 p
39. A. Brenac, R. Baptist, G. Chauvet, R. Meyer, Caracteristiques energetiques de cathodes a
micropointes a emission de champ. Revue Phys. Appl. 22, 1819–1834 (1987)
40. C.A. Spindt, I. Brodie, L. Humphrey, E.R. Westerberg, Physical properties of thin film field
emission molibdenium cones. J. Appl. Phys. 47, 5248 (1976)
41. K.L. Jensen, E.G. Zaidman, Field emission from an elliptical boss: Exact versus approximate
treatment. Appl. Phys. Lett. 63 (1993)
42. M. Lampert, P. Mark, Injection Currents in Solids. M.: Mir (1973), 416 p
43. R.G. Forbes, Refining the application of Fowler–Nordheim theory. Ultramicroscopy. 79,
11–23 (1999)
44. E. Hantzsche, Beitrage zur Plazma-physik, 22, 325 (1982)
45. G.D. Yakovleva, Tables of Airy functions and their derivatives. M.: Nauka (1969)
46. Y.V. Kryuchenko, V.G. Litovchenko, Computer simulation of the field emission from
multilayer cathodes. JVST. B14, 1934–1937 (1996)
47. M. Hollander, D. Wolfe, Non-parametric statistical methods. M.: Finansy i Statistika (1983),
518 p
48. S.G. Rabinovich, Measurement Errors. M.: Energia (1978), 261 p
49. Y. Linnik. The method of least squares and the foundations of mathematical and statistical
evaluation of parameters. M.: Fizmatgiz (1962), 350 p
50. A. Brunetti. A fast and precise genetic algorithm for a non-linear fitting problem. Comp. Phys.
Comm. 124, 204 (2000)
51. Enrico Fermi. Quantum mechanics. M.: Mir (1965). 242 p
52. Bohr. Atomic Physics and Human Knowledge. M.: GIIL (1961)
53. M. Kendall, A. Stewart, The Theory of Distributions. M.: Nauka (1966), 520 p
54. J.N. David, Probability Theory of Statistical Methods (Cambridge, 1951)
Chapter 2
Experimental Equipment and Technique
Abstract This chapter presents the field emission electron microscope, its tech-
nical characteristics and various modern construction of microscopes. A particular
attention is paid to analyzing full energies of the electrons and universal con-
structions of field electron microscopes that are the main branch of development of
microscopy and spectroscopy. The most popular practical techniques of manufac-
ture of tip field emitters from different materials.
2.1 Field Emission Microscopes
A field emission microscope, invented by Müller in 1937 [1, 2], is a powerful and to
a certain degree one-of-a-kind tool to research emission features of a field-emitter
surface with high-resolution. First of all, those features include adsorption and
desorption kinetics, distribution of field electrons by energies, distribution of an
electron work function over an emitter surface, surface diffusion and eventually,
testing of field emission cathodes.
Some functions of structural research of field emission microscopy have grad-
ually moved to field-ion microscopy (e.g., surface diffusion research), which has
atomic resolution. However, emission and energy characteristics are still the
domain of field emission microscopy.
The basic circuit of a field emission microscope has not changed since the its
invention (Fig. 2.1). Current leads (1) are welded to a molybdenum or tungsten arc
(2), to which, in turn, the sample (3) is attached. This design makes it possible to
heat the sample up to the necessary temperature. The visor window of a glass bulb
has a luminescent screen (5), where the field-electrode image of the emitting surface
of the sample is formed. The luminescent screen is a transparent current-conducting
coating applied to glass with a thin phosphor layer. High positive voltage is applied
to the screen (5) through a current lead (6). The high-energy electrons emitted from
the emitter bombard the phosphor layer, causing its luminescence. Thereby, the
screen displays a pattern of distribution of electrons from sharp bands of the emitter
surface or the regions with a relatively low electron work function. Gas is evacuated

44 2 Experimental Equipment and Technique
Fig. 2.1 Schematic of field

emission microscope.
1 current lead-ins; 2 arch;
3 sample; 4 bulb;
5 luminescent screen; 6 high
voltage current lead-in;
7 exhaust
from the flask through an exhausting tube (7). The microscope bulb can work both
under continuous gas evacuation and as a sealed-off device.
The exponent of Fowler–Nordheim equation has the factor of 6.8 107 in the
numerator. Since the electron work function u of most materials, including
adsorbed systems, belongs to the range of 2–6 eV, it is clear that field emission can
be appreciable if the fields are about 106–107 V/cm. The easiest way to generate
such fields is to make a surface with large curvature, e.g. a needle with a very small
apex curvature radius. In this case, very high intensity of the electric field is
generated at the tip of the tip under moderate voltages. Electric intensity for an
insulated charged sphere is determined as
V
F¼ ; ð2:1Þ
r
where V is the voltage applied; r is the curvature radius.

For an apex emitter, which can be approximated as a sphere on a cone, electric
intensity at the apex can be expressed as
V
F¼ ; ð2:2Þ
kr
where k = 3.5–5, depending on the anode–emitter geometry.

Increase in the field emission microscope is caused by the electrons leaving the
emitter material through the potential barrier with very low energy and flying along
the electric fluxes (at least at the initial stage). Since the tip is mostly a conductor,
2.1 Field Emission Microscopes 45
these fluxes are perpendicular to the apex surface and hence radiating outward.
Hence, the increase is determined as
D
M¼ ; ð2:3Þ
br
where D is the emitter–screen distance, b is image shrinkage factor approximately

equal to 1.5.
It is absolutely clear that influence of the anode is extremely low because
actually all basic changes of the potential manifest themselves at the distance of
several emitter curvature radii, which are essentially less than the anode-emitter
distance. The emitters can be made with radii of less than 100 Å (the emitter
fabrication technique cf. Sect. 2.5). Thereby, the value of D 50–100 mm can
cause an increase of more than 105 and the screen shows a highly-increased
emission card of the apex vertex. Very promptly, the electric intensity wanes within
4.5° of the polar angle from the apex axis [3].
Field emission current is mostly determined by the exponent of the Fowler–
Nordheim equation, which is why it depends on the local electron work function
and electric intensity. If the a tip surface is atomically smooth, the emission current
is determined only by the electron work function and the luminescent screen shows
the image reflecting the distribution of work function of the electrons having dif-
ferent crystallographic directions.
The contrast of field emission images is determined by local changes in the
electron work function and electric intensity or both, which happens more often
than not. E.g. Fig. 2.2 shows the image of a single-crystal tungsten made of
tungsten wire by electrolytic etching and then treated by the temperature flash of
more than 2000 K [3]. Such treatment produces a round-off, almost hemispherical
shape of the apex tip with flat faces corresponding to low-index crystallographic
planes, such as (001), (002) and (112). Atomically flat low-index planes have a high
electron work function, which manifests itself as dark spots on the image; while
atomically rough high-index planes have a lower electron work function and hence
high luminosity. Due to the radial nature of electron paths in the field emission
microscope, the symmetry of the sample is preserved in the image, which can be
connected with the stereographic projection of the original surface. So crystallo-
graphic indexation can be directly done by stereographic projections.
Figure 2.3 shows a stereographic projection of a body-centered cubic lattice of
the orientation [001] corresponding to the image in Fig. 2.2.
Resolution of a field emission microscope is determined by two factors, namely
diffraction and tangential velocity of the emitted electrons. However, the respective
wave length of electrons at the screen where the image is formed, is about
12.5 V−1/2 Å, which gives a very small value for electrons with an energy of a few
kilovolts. Therefore, the total contribution of diffraction is relatively small.
Fig. 2.2 Field emission

image of a pure, thermally
annealed tungsten tip. The
[001] plane is shown in the
center
Fig. 2.3 Stereographic

projection of [001]-oriented
body-centered cubic crystal
with indices of atomic planes
A much more important factor that actually determines the resolution is the
statistical spread of electrons in tangential velocities in relation to the apex-screen
axis. For tangential velocity V at the emitter surface, distance d from the initial
electron motion direction on the screen is
d ¼ 2Vst; ð2:4Þ
where t is the time for an electron to travel from the emitter to the screen.
Factor 2 appears because the tangential motion “throws” the electron away from
the electric flux, so the electron gets tangential acceleration and an additional bias
from the initial electric flux. The electron transit time can be sufficiently well
approximated as
t ¼ Dðm=2VeÞ1=2 ; ð2:5Þ
where V is the voltage applied because the entire electron path meets its total
kinetic energy Ve.
Half-breadth of resolution of a field emission microscope is defined by the
following formula [4]
1=2
d=2 ¼ 1:31 104 b r=kru1=2 ; ð2:6Þ
where r is an insignificantly varying table function [3].

Therefore, the resolution of the field emission microscope mostly depends on the
tip curvature radius and amounts to about 20 Å under ordinary conditions. Very
sharp tips and respectively low running voltages display negligible contribution of
diffraction.
2.1.1 Design of Field Emission Microscopes
The designs of field emission microscopes are numerous because they are not issued
commercially. Which is most important, each experimenter develops the micro-
scope design according to the purpose of his/her experiment and proceeding from
his/her resources.
Figure 2.4 shows classification of the basic structural components of field
emission microscopes. Being relatively conventional, this classification describes
only the most important of them.
As a rule, the design of any field emission microscope contains from one to five
elements specified.
This section gives some examples of designs of field emission microscopes,
disclosing interrelation between the elements aimed at achievement of specific
research goals and giving an idea of general trends in designing of field emission
microscopes.
For a long time, since the very invention of field emission microscope by Müller
in 1937 [1] and nearly up to the late 1920s, most designs of field emission
microscopes were constructed out of glass, as shown in Fig. 2.1. Usually the flask
was made of high-melting glass S53, which is a so-called molybdenum glass, i.e.
the linear expansion factor is almost equal to the molybdenum linear expansion
factor and electrical outputs are made of molybdenum. Since such devices are used
Fig. 2.4 Basic construction parts of field emission electron microscopes
to research apex emitters, the working vacuum in the device should be at least
10−9 mm Hg. Such pressure is achieved by warming to the point 450 °C for more
than 8 h [5] and use of getters that maintain vacuum in the course of operation. For
adsorption research, glass sealed off field emission microscopes are usually sup-
plemented with adsorbate sources (e.g. [6, 7]). With development of vacuum
technique and gas evacuation equipment, fully sealed-off devices were replaced first
by glass metal-flanged devices connected to evacuation systems and then by fully
metallic demountable constructions.
Some most successful and serviceable designs of full-metal field emission

microscopes are shown below.
A field emission microscope with case diameter of 50 mm is shown in Fig. 2.5.
The microscope consists of a case (4) made of X18H10T (Type 321) stainless steel,
one side of which is welded to DN50 flange (9) and the other to a flat observation
window (1) via a kovar ring (2). The kovar ring grooves are to compensate for
temperature deformations and prevent glass cracking during the warm-up. All
cathode and anode parts are mounted on four ceramic-metal current leads (10)
located on the flange (9). A high-melting metal arc (7) with a field emission cathode
welded to it (6) is fixed in two current leads by muffs (8). The anode is a flat
luminescent screen (3) consisting of a flat glass plate with a current-conducting
layer and a luminophor. The screen is secured between two stainless-steel rings. To
improve contact between the current-conducting layer and the rings, the rim of the
current-conducting coating is covered with Aquadag. The anode is secured to the
current leads (10) via a muff (8) and cross-arms (5). Electric insulation of both
the anode and the cathode circuit makes it possible to apply high voltage to the
microscope screen, which permits measuring the emission current at the
low-voltage circuit unit, thereby improving measuring accuracy and operational
safety.
This design is especially good at photographic recording of field emission
images.
For the cases when it is enough to maintain the working vacuum at
10−7–10−6 mm Hg, similar designs with readily removable flanges sealed with
Viton were developed.
For field emission microscopy of samples with a large working area under
increased current draw, the conventional microscope designs are not good because
Fig. 2.5 Small full-metal

field emission electron
microscope. 1 Visor; 2 Kovar
ring; 3 flat luminescent
screen; 4 housing;
5 cross-arm; 6 field cathode;
7 arch; 8 connection muffs;
9 flanges; 10 current lead-ins
Fig. 2.6 The main node of a

strong current field emission
projector; 1 ceramic current
lead-in; 2 DU-50 flange;
3 anode node fixture
cross-arm; 4 luminescent
screen fixation screen; 5 flat
luminescent screen;
6 tungsten lattice; 7 lattice
fixation rings; 8 lattice
stretching rings; 9 lattice
stretching screws; 10 cathode;
11 cathode holder
the heavy current of field emission destroys the conductive coating and the lumi-
nescent layer. Even replacement of the conventional glass with an
yttrium-aluminum garnet plate with an addition of cerium, that has a much better
thermal stability, does not solve the problem. For these purposes, a special
microscope (Fig. 2.6), was designed on the basis of DN50 flange (2).
All elements of the microscope are mounted on four rods of the ceramic current
leads (1). The anode system is a luminescent screen (5) fixed between two rings (4)
and a tungsten grid (6). A reliable electric contact between the conductive coating
and the lower ring (4) is provided by an Aquadag layer. A pre-annealed grid is
strained between the two rings (7). The grid is additionally strained by three screws
(9) and a special ring (8). The distance between the grid and the luminescent screen
is 1.5 mm. The field emission cathode (10) fixed in the holder (11) is at the distance
of 1.5 mm from the grid.
In such a design, the grid assumes more than 90% current draw without essential
distortions in the field emission image of the sample surface.
2.1.2 Special Designs
Tenacity of the emitters and their structural components can be determined with a
special design intended for other purposes as well (Fig. 2.7). The device can hold as
many as four samples at the same time (3). High voltage from the high-voltage
rectifier is applied to the anode (2), which is a transparent disk with a conductive
Fig. 2.7 Field emission projector for stress tests in an electric field: 1 Kovar ring; 2 luminophor
screen; 3 field cathode; 4 nickel holder; 5 ceramic insulator; 6 current lead-ins; 7 Kovar adapter;
8 housing; 9 Viton laying; 10 fixture flange; 11 flange with current lead-ins; 12 contact brush;
13 glass window; 14 Kovar adapter; 15 setup flange
coating and a luminophor. In this design, the role of the anode (2) is different from
that of the observation window (13). The entire anode system is electrically insu-
lated from the chamber case (8) by a high-voltage insulator (5) rated for 30 kV.
Performance of the flange (11) with current leads (6) is improved due to Viton
sealing (9).
Tenacity of structural components of carbon fibers is determined by the method
of ponderomotive electric field loads [8, 9]. In this case, mechanical tension is
calculated by the formula
E2
r¼ ; ð2:7Þ
8p
where E is intensity of the electric field.

These experiments are designed to determine mechanical tenacity of microele-
vations on the effective surface of carbon fibers by negative-voltage supply. The
state of the effective surface of the fiber was controlled by the field emission pattern
(positive voltage).
The research technique is as follows.
After 30-min forming, the field emission current from the sample is set up within
the range of 25–100 µA. The value is selected for each separate case proceeding
from the condition of high-quality luminescence of the luminophor necessary for
photographing.
Then the screen is fed with a negative voltage pulse. After that, a new value of
field emission current and a new field emission image are recorded under direct
voltage with the amplitude being constant during the experiment. A control sample
is put under the microscope. The sample is fed with pulses of only positive polarity
and a constant amplitude. It is intended for accurate differentiation of destruction of
the fibers caused by ionic bombardment under direct voltage and ponderomotive
loads under reverse voltage. The sample tested is fed with a series of pulses of direct
and reverse polarity. The amplitude of the reverse-polarity pulses is increased by
200–400 V per pulse. For every direct-polarity pulse, a new value of field emission
current and a respective field emission image are recorded. Supply of pulses ceases
when the field emission current of the sample reaches the range of 1–5 µA.
The value of the electric field at the microelevation to be destroyed is determined
through the density of the field emission current from the microelevation before the
destruction. The electron work function of the carbon-fiber material with different
temperatures of thermal treatment is approximately the same. It amounts to about
4.6–4.8 eV. The change in the current caused by destruction of the microelevation,
which is identified by field emission images, can be calculated from the flow chart
of the currents as a difference between the currents corresponding to the moments
before and after the destruction.
Then, determination of the microelevation emitting-surface area necessitates
making a number of assumptions. The most important among them is that the size
of the emitting microelevations by the order of magnitude is that of the fibrils. Such
an assumption is quite justifiable because the ionic bombardment results in dis-
closure of the fibril structure of the fiber, and the fibril size is known to amount to
30–100 Å. Knowing the microelevation size, we can express the emitting area for
each particular microelevation as
S ¼ r2 ; S ¼ r2 ;
where r is the typical size of the microelevation.

The Fowler–Nordheim graph determines electric intensity E, while the current
density of field emission from the microelevation and work function u are known.
Then mechanical tension is calculated by the following formula
E2 2
r¼ k ð2:8Þ
8p
where E is the value of the electric field at the microelevation, k is the factor equal
to the ratio of the negative and the positive pulse-voltage polarity amplitudes at the
moment of destruction of the microelevation.
The field emission current is lowly dependent on temperature; however, its value
is among the key operational parameters of electronic devices. Therefore, it is an
urgent problem to study the operating temperatures of field emission cathodes.
The temperature of the apex field emitters welded to the arc of a high-melting
metal (W, Mo) can be measured by change of electrical resistance of the particular
arc section [10]. A more precise temperature value of the emitting apex tip can be
calculated by the design formulas [11]. If the cathode is fixed in a different way or,
which is even more important, more massive cathodes are used, such methods of
measuring temperature are improper.
The temperature of graphite field emission cathodes with a well-developed
effective surface of tens of square millimeters can be determined [12] with a direct
method of measuring by means of thermocouple reference junctions.
The structure of the experimental device (Fig. 2.8) [13] is mounted on DN50
flange (5). The cathode (7) (a rod 12 mm long, cross-section 1.5 1.5 mm) is
mounted at the distance of 0.1 mm from the anode. The cathode is equipped with
two thermocouples (the experiment involved copper constant thermocouples
manufactured from the wires of 30 lm in diameter). One thermocouple (8) is
mounted in the groove filled up with Aquadag at the distance of 0.5 mm from the
effective surface of the field emission cathode, and another one (11), intended to
study distribution of temperature along the length of the cathode, at the opposite
end face of the cathode. The anode (1), as well as the field emission cathode, is
made of MPG-6 graphite. As the anode in our experiments is under high voltage,
Fig. 2.8 Schematic of an experimental temperature measurement device: 1 anode; 2 anode

holder; 3 anode thermocouple; 4 anode holder insulation; 5 flange; 6 thermoconductive ceramic;
7 field cathode; 8 working side thermocouple of the cathode; 9 cathode holder insulator;
10 cathode holder; 11 cathode back end thermocouple
temperature is measured with a thermocouple (3) fixed to the heat-conducting

insulator, in this case made of beryllium ceramics (6), mounted in the groove laid in
the anode at the distance of 1 mm from its effective surface. The effective surfaces
of the anode and the cathode were polished before the fact. The experiments are
carried out in a demountable high-vacuum continuous-evacuation chamber of
X18H10T stainless steel.
Experimentally, the pressure of residual gases in the chamber is 10−6 mm Hg.
The spatial and angular distribution of the field-emitter current is researched with a
device version combining a field emission microscope and a scanner. In this case,
measurements can be made promptly and continuously in the course of both
forming of the field emission cathode and its operation [14]. Such scanners are used
in various data-handling systems, data-reduction units and data-display devices
[15, 16].
An experimental device designed to research field emitters is given in Fig. 2.9.
A tube contains two electron guns located on both sides of a two-sided storage target
(3) with an effective surface of 84 mm. The field emission cathode to be studied is
located on the side of the target substrate at the distance of 30–60 mm from it. The
target records the spatial distribution of the emission current from the field emission
cathode (3). The target substrate is the aluminum film 60 nm thick lying on a
supporting nickel grid (2) with a pitch of 60 lm and transparency of 60%. The
substrate is coated with a layer of high-ohmic semiconductor, which functions as a
storage layer of the target and at the same time has a property of electron-induced
conduction at electron energies of 5–10 keV. The electron-emitting source of such
energy in this device is the field emission cathode. There is a flood gun (5) and a
bowl-shaped collector (4) on the side of the storage layer of the target. The magnetic
fields of coils (6) focus the read-out electron beam into a spot of 1000 lm in
Fig. 2.9 Structure of an experimental device: 1 tip field emission cathode from the recording side;
2 lattice substrate of the target with signal plate; 3 accumulating and semiconductor layers of the
target; 4 collector of secondary electrons; 5 scanning electron projector with field emission
cathode; 6 electron beam focusing and diversion device
diameter and deflect the beam from the target at the velocity of 2000 m/s. The
reading is done by charge exchange with an unmodulated focused beam of electrons
of 1 keV. The tube resolution is 1200 lines per target diameter. The signal-noise
ratio is 10 for the read-out beam current of 0.4 µA. Such device can record signals
under various modes of operation of the field emission cathode: a single pulse,
infrequent and periodic pulses, a continuous current-draw mode. The signals
recorded can be stored and read-out within a period of at least a few tens of seconds,
the erasing time being a few seconds. Independence of operation of the field
emission cathode and the reading gun permits recording in parallel with reading.
Measuring is done at the installation diagrammed in Fig. 2.10. Reading is done
by scanning of the target with an electron beam swept as a standard television raster
Fig. 2.10 Flowchart of measurement of field emitter emission characteristics: PwrSpply1,

PwrSpply2 are power sources of field cathode and target; PreAmp is the preamplifier;
Measuring Pulse Gen is the measurement impulse generator; PwrSpply3, PwrSpply4 are the
power sources of the scanning projector and focusing/diversion system; TV Gauge is a television
indicator; Line Select is the television line selection device; Read Pulse Synchonizer is the
readings timing generator; Oscilloscope is the C1-65 oscillograph
in step with the TV display raster. The signal of reading from the target is amplified
by a preamplifier and supplied to the TV display. The screen of the display shows a
field emission image of the carbon fiber cathode in the course of its forming and
operation. The amplitude of the signal of reading from the selected target region is
measured by a horizontal separator (HS) displaying the information on the screen of
an oscillograph.
Illustratively, Fig. 2.11 shows the oscillograms made in the course of forming of
a field emission cathode from a polyacrylonitrile carbon fiber. The forming usually
takes 30 min. The oscillograms are made with a horizontal separator (HS) along
three parallel lines, perpendicular to the target diameter and dividing it into four
equal parts. The oscillograms show spatial distribution of emission current of a field
emission cathode in selected target regions in arbitrary units. Figure 2.11b oscil-
logram is angular distribution of the field emission cathode current of this direction
after recalculation of linear coordinates into angular ones.
In the central part of the target, measurements are made with angular resolution
2 mrad, where L 50 mm is the distance from the end face of the field emission
cathode to the target, dx = 0.1 mm is the minimum distance along the target set by
the diameter of the focused electron beam of the reading gun.
The oscillograms show that heterogeneity of spatial distribution of the field
emission cathode peaks at the boundaries of emission spots at the beginning of
forming of the field emission cathode. After forming has been over, the hetero-
geneities decrease to the acceptable level in connection with a considerable inter-
ception of electron fluxes from a great number of emission centers.
In this way, the installation described makes it possible to measure the charac-
teristics of spatial and angular distributions with angular resolution up to 2 mrad of
the emission current of the field emitters.
Fig. 2.11 Experimental

dependence of current (in
arbitrary units) along the
selected directions: b along
the horizontal diameter of
television indicator; a, c along
the directions removed from
direction b by half of the
indicator screen radius (field
emitter current is 30 lA, field
cathode voltage is 8 kV)
2.2 Field-Electron Total Energy Analyzers 57
2.2 Field-Electron Total Energy Analyzers
A field-electron total energy analyzer is among the promising field emission devices
used to research the internal structure of a solid body and the phenomena taking
place on the emitter surface [17].
All analyzers share the field emission microscope geometry, which means that
the emission pattern given by the electrons having tunneled from the apex is pro-
jected to the fluorescent screen. The screen has a small hole, the so-called trial hole,
transmitting the electrons subjected to further analysis [18]. To select the surface
region to be analyzed, the emission pattern is positioned so that the image of this
region aligns with the trial hole [19, 20].
There are several ways to position an emission image in respect of the trial hole.
They can be divided into three categories: (a) electromagnetic deflection, (b) elec-
trostatic deflection, (c) mechanical travel.
Irrespective of the deflection type applied, the apex image should be maintained
on the optical axis of the system.
The magnetic deflection system [21] has two pronounced disadvantages when
used in the energy analyzer: difficulty to screen the analyzer from the magnetic field
and a bias of the pattern caused by extraction-voltage changes when volt-ampere
characteristics are read.
The electrostatic deflection system [22, 23] solves both of these problems, but
introduces those of a technical nature. It is necessary to mount small insulated
deflecting plates in the immediate proximity of the sample and apply control electric
signals through the vacuum chamber. In addition, such systems distort the field
emission image of the apex surface.
The third type, the mechanical movement, is rather difficult and cumbersome,
but preferable under certain conditions because it does not distort the field-electron
energy spectrum. The only condition is that the traveling apex tip should not widely
deviate from the analyzer axis.
In measuring energy distribution of field electrons, it is assumed that tunneling
does not change the electron energy. Essentially, there are some regions of accel-
eration and deceleration of electrons in the field-electron analyzer (Fig. 2.12a).
Figure 2.12b shows a potential diagram appropriate to the occasion when electron
emission comes from the Fermi level and the collector is at the vacuum zero level.
If the collector is maintained at the zero potential, the field emitter potential
should be negative. Field electrons reach the collector, if their energy
E [ ucoll þ EF Ucoll ;
where ucoll denotes the collector work function, EF is Fermi energy, Ucoll is the
delay potential.
In this way, research of the emitter-material electron energy distribution by the
moment of emergence of collector current can be supplemented with determination
of the collector work function value.
Fig. 2.12 a Scheme of a field

electron energy analyzer with
a retarding potential.
1 cathode, 2 acceleration area,
3 anode with a diaphragm,
4 retarding area, 5 collector,
b potential diagram of the
analyzer
Fig. 2.13 Construction of an

energy analyzer with a
retarding potential [24, 25].
1 collector, 2 focusing
cylinder, 3 anode,
4 fluorescent screen,
5 1.5 mm hole, 6 glass bellow
assembly, 7 getter lamp, 8 ion
pump
The first design of the field-electron energy analyzer with a retarding potential is
described in [24, 25] (Fig. 2.13). It consists of a flask containing a collector, i.e. a
thin conducting film laid on glass, and a shield electrode. Near the emitter, there is a
screened anode with an hole of 1.5 mm and the focusing cup. The foot is con-
nected with a flexible glass bellow for convenience of positioning of the necessary
emitter face to the hole. For visual observation of the image, the anode is coated
with a phosphor layer (at 106 magnification, resolution is about 10 Å).
The anode design assumes the field in the hole to be zero, i.e. there is no lens
effect. Appropriately focused electron beam converges in the sphere center and then
moves perpendicularly to the spherical collector surface. To improve stability and
eliminate the background, anode voltage is supplied to the microscope from the
battery packages. The retarding-potential source permits changing the delay
potential through a millivolt of up to 1 V. The analyzer design also includes
electrostatic and magnetic screens. Use of thin apices (down to 300 Å) to decrease
anode voltage permits improving the analyzer resolution by diminishing the
influence of geometrical defects and heterogeneity of the collector on the work
function. The analyzer resolution is *30 mV. Figure 2.14 depicts a delay curve
produced by this device.
Collector current is
Z1
Icoll ¼ PðEÞdE;
ucoll u þ Ucoll
after differentiation, we obtain
dI=dUcoll ¼ Pðucoll u þ Ucoll Þ;
i.e. differentiation of the delay curve gives us total-energy distribution proper. The
fall in the delay curve at high voltages of the collector in the structure [24] is caused
by reflexion of electrons from the collector. That is the central disadvantage of the
early design, aggravated by poor sensitivity and long-time constant of the device.
This design serves as the basis for energy analyzers intended to research field
emission properties of semiconductor films [26–32].
Efforts to raise the analyzer resolution resulted in replacement of the spherical
collector with a hemispherical one [21, 33, 34].
Fig. 2.14 Retardation curve

for a tungsten emitter with
21 K, E = 30 106 V/cm,
u = 4.4 eV
These designs permit getting rid of reflexion of electrons, thereby improving

resolution and repeatability of results.
These devices are similar to the previous ones in terms of the operating principle.
The Faraday cup, being under the earth potential, serves as a shield electrode for the
hemispherical collector, and the anode potential is kept constant under the changing
delay potential.
The analyzer design with a retarding potential and a hemispherical collector is
presented in Fig. 2.15 [34]. The key parts of the analyzer are the anode (4) with a
probe hole of 1 mm in diameter with a phosphor layer applied to the anode, a
focusing system, a Faraday cup (9), a hemispherical collector (8). The collector (8)
is made of molybdenum in a form of a hemisphere with high accuracy and surface
finish, which is necessary to improve the analyzer resolution. The focusing system
consists of three diaphragms. The function of the first diaphragm is performed by
the anode (4) with the inlet hole of 1 mm in diameter. The second diaphragm (5)
has a hole of 2 mm in diameter. The third diaphragm is the cover of the Faraday
cup (6) with a hole of 4 mm in diameter. The diaphragms are made of molybdenum
and secured with glass «beads» (7). A field-electron beam emitted from the apex (3)
is defined by the probe hole in the anode (4). It focuses in the center of the
molybdenum hemispherical collector (8) and moves further on along the normal
line to the hemisphere surface. The retarding potential is supplied to the field
emitter (3). The ratio between the anode potential and the lens potential is essential
for operation of the device. Figure 2.16, shows that the ratio of VL/VA determines
the device resolution. In particular, the optimum ratio for design [34] is 0.003.
A millimeter-scale hole in the anode permit getting a ratio of the collector current
to the total current of field emission as just about 10−3, which is especially
important in research of large work function faces. Collector current in the range of
10−9−10−14 A is measured by electrometer.
The delay curve is registered by recorder and then differentiated. Similar designs
are used in [35, 36].
Sensitivity of measuring of collector current can be increased by application of a
collector in a form of a hemispherical grid with subsequent amplification of the
collector current by electronic multiplier [37].
High resolution is more reliably provided by cylindrical retarding-system ana-
lyzers. Such analyzers, which differ from most other designs in having cylindrical
symmetry [20, 38, 39], employ a soft mode, when electrons are gradually retarded
by the analyzer optics along with focusing on the recording element, which permits
improving computational accuracy and diminishing the effect of faulty machining
of the analyzer parts and their assemblage despite some increase in analyzer
dimensions.
The analyzer [38, 39] (Fig. 2.17) is secured on the flanges of the
ultra-high-vacuum chamber (P * 10−9–10−11 Torr) equipped with oil-free
pumping devices. It contains anode (2) with a probe hole and luminophor (1)
applied to the outer side, as well as 11 lenses of different size (3) separated from
each other with L-shaped ceramic inserts (18). The entire lens system is tightened
by three rods (4) through insulated screws (5), which secure the lens system to
Fig. 2.15 Construction of an analyzer with retarding potential and hemispherical collector:
1 measuring wires; 2 arch; 3 tip; 4 anode cylinder with a probing hole; 5 focusing lens; 6 Faraday
cylinder cover; 7 fixation «beads»; 8 hemispherical collector; 9 Faraday cylinder; 10 Faraday
cylinder base; 11 mounting leg; 12 electrode lead-outs; 13 collector lead-out; 14 glass bulb;
15 conductive covering; 16 conductive covering lead-out
foundation (9). The foundation hosts grid (7) manufactured by electroplating and
stretched on a ring (400 lines/cm). The grid is attached to the foundation with three
insulators (8). The grid pitch, negative allowance and equipotentiality of its surface

dependence of half-width of
field electron spread by their
full energies on ratio between
lens voltage VL and anode
voltage VA at T = 79 K. The
minimum is reached at
VL/VA = 0.003
have the most powerful effect on the analyzer resolution. The entire structure is
attached to the chamber flange with three screws (10).
Behind the grid, there is microchannel plate (6), which is used as a collector.
([20] replaces the microchannel plate with an electronic multiplier.) Sample (11) is
welded to arc (12) secured in holders (13) on insulator plate (14). To align the
sample with the analyzer probe hole, plate (14) is secured on bar (15) rigidly bound
with flange (17), which can move with respect to the chamber case by means of
bellow displacement node (16) (Fig. 2.17 shows only the bellow).
The lens potentials are calculated on the computer to optimize the motion tra-
jectories in the retarding and focusing system of the analyzer. The values of lens
potentials and electron trajectories in the analyzer are given in Fig. 2.18. The design
values u1/u0 = 0.11, u2 = 15/17u1, u3 = 13/17u1 etc. The potentials are set with a
divider, which permits changing them by 10% design value to compensate for
lens size deviations from those of the design. The real working value of u1/u0 for
this analyzer is 0.117.
The electrical part of the analyzer is represented in Fig. 2.19. It is assembled on
the basis of a personal computer and contains high-voltage source PwrSrc [38]
controlled by digital-to-analogue converter DAP, voltmeter B7-16 measuring this
voltage 1/1000 divider, controlled delay-potential source DAP X, voltmeter B7-21,
measuring analogue-to-digital voltage converter ADP, sample warm-up unit
SpecimenHeater, electrometer B7-30 measuring current in the analyzer collector
circuit and an x/y-plotter. Automatic functioning of the analyzer is provided by
purposefully-developed programs making it possible to take, digitize and process
delay curves in mathematical terms. The measuring system permits carrying out
research of the energy spectrum in the range from 0 to −12 V with the accuracy
limited only by noise in electric circuits.
To take the delay curve, delay voltage is supplied from digit-to-analogue con-
verter DAP X through range extender BandExtender (scanning ranges from 0 to
5 V) and through smooth bias control unit OffsetFineTuner, which permits smooth
biasing of the initial point of scanning within the range from 0 to −7.5 V.
Fig. 2.17 Construction of a cylindrical analyzer. 1 Luminophore, 2 anode with probing hole and a
protective cylinder, 3 cylindrical lenses, 4 tightening rods, 5 screws, 6 microcanal plates, 7 lattice,
8 insulators, 9 base, 10 fixing screws, 11 the sample, 12 arch, 13 holder, 14 ceramic plate, 15 rod,
16 bellows assembly, 17 flange, 18 ceramic insertions
Fig. 2.18 Potentials on

lenses and electrons inside
retarding and focusing system
of the analyzer at different
levels of deviations of
inbound beam from the
analyzer axis. u0 is the anode
potential
Fig. 2.19 Schematic of the measuring device. A is the analyzer, Ctrllr is the controller, Iface is the
interface, BtryComm-r is the battery commutator
The delay voltage supplied to the cathode is controlled by digital voltmeter

B7-21 within the accuracy of 0.01%. The analyzer collector current corresponding
to each gradation of the DAP X is measured with electrometer B7-30, digitized and
stored in the computer memory. Having covered the entire scanning range, the
curve obtained is processed in mathematical terms and the result is plotted.
The actual resolving ability of the analyzer is determined with experiments on
measuring of the energy spectrum of the electrons emitted from the tungsten face
(111). The resolving ability is determined by the method suggested in paper [40].
At the sample temperature of 300 K and the anode voltage of u0 = 2.5 kV, the
resolving ability of the analyzer is 25 meV. Computation of the electron work
function results in the value of 4.71 eV, which is in good agreement with the
known values.
There is a field-electron energy analyzer version combining the
retarding-potential method and the electron energy dispersion method [41]. Its
updated design is presented in Fig. 2.20. Dispersive analyzers have essentially
lower background currents.
The field emission cathode to be researched is secured in the working chamber
on the four-degree-of-freedom manipulator able to move in x, y and z dimensions
and rotate around the vertical axis on the electron-optical axis of the analyzer. Being
luminophor-coated, the analyzer anode serves as a small screen of the field emission
Fig. 2.20 Electrostatic dispersion analyzer of field electron energies. 1 Anode with a probing hole
and a layer of luminophore; 2–7 cylindrical lenses; 8 insulating rings of quartz; 9 electron
multiplier
microscope. It also permits observing the emission pattern of the field emission
cathode to be researched, controlling of its surface state and determining of the
crystallographic direction. The analyzer anode is supplied with high voltage. The
electron analyzer consists of six electrostatic lenses (2–7) formed by cylindrical
elements electrically isolated from each other by quartz rings (8). The anode and the
lenses are made of molybdenum.
The electrons of the region probed come to the energy analyzer through the input
diaphragm, slow down to initial energies on the first lenses and disperse by energies
by means of chromatic aberration on the last lenses [42]. Through the output
diaphragm, electrons get to the input of the channel secondary-electron multiplier
SEM-6 (9) working in the electron-count mode, which improves analyzer sensi-
tivity. From the output of the secondary-electron multiplier, the electric impulses
come to the input of high-frequency emitter follower EF and move on to broadband
pulse amplifier PA-29. High-speed comparator-inverter HSC converts the amplified
impulses into square impulses of the necessary amplitude coming to the
transistor-transistor logic input of CAMAC-crate counter M401 connected to the
computer.
The energy analyzer element potentials are set by the resistance divider
according to the computations [42]. If R is the total bleeder resistance, the bleeder
resistances fit the following ratios:
R1 =R ¼ 1; R2 =R ¼ 0:263; R3 =K ¼ 0:022; R4 =K ¼ 0:0021:

Energy analyzer characteristics are as follows.

1—input diaphragm diameter din = 0.5 mm;
2—output diaphragm diameter dout = 0.17 mm;
3—resolution 20 meV;
4—analyzed energy range ΔE = 3.5 eV;
5—dynamic range of the signal measured in the electron-count mode—105 el./s.
Further reduction of background currents involves dispersion analyzers with
hemispherical analyzers [43] or 127-degree analyzers [44]. Let us consider the latter
design in more detail. A scaled analyzer diagram is shown in Fig. 2.21.
It includes a retarding-lens system similar to the one described in [43] and two
consecutive 127-degree analyzers. The sequence of two energy analyzers is used to
reduce the number of electrons coming to the analyzer slits and walls as a result of
scattering. The background causes especially serious difficulties for taking field
emission electron energy distribution because the current of the region to be
researched varies in a wide range.
Ideal electron trajectories between deflector electrodes have a curvature radius of
17.5 mm and slits are 1 mm wide. A combination of parallel slits with spherical
apertures of lenses is certainly unusual, but has an advantage that machining of
elements of the 127-degree analyzer is much simpler than that of the hemispherical
analyzer. However, the latter is sure to have best focusing. But a small negative
voltage difference between the front plates (missing from Fig. 2.21) of the
127-degree analyzers and small energy of electrons laced up by analyzers make
Fig. 2.21 Schematic of a

device with two 127-degree
analyzers shown to scale. 1
Two 127-degree analyzers, 2
tip, 3 retarding and focusing
lenses, 4 accelerating lenses.
Voltages on different
elements are as follows:
654 V (internal), 365 V
(external), 355 V (frontal)
divergence of the electron beam in respect of the plane insignificant. The voltage
difference is usually 0.145 V.
The luminescent screen to observe the image in field emission electrons is 6 mm
in diameter. There is a hole of 1 mm in diameter in the center. The distance between
the electron emitter apex and the screen is about 5 cm. The hole in the screen is
covered with a thin wire grid supplied with the same voltage as the screen. As a
result, there is no field in this region and the hole is not functioning as an electron
lens.
For the entire system to fit a small ultra-high-vacuum chamber, it is made quite
compact. All parts of the analyzer are made of electrolytic copper. Insulation and
accurate positioning of the parts was achieved due to rods and rings of aluminum
oxide. The system is mounted on a flange of 20.3 cm in diameter. The system is
protected from magnetic fields and scattered electrons and ions with a screen of
permalloy.
High positive voltage necessary to draw the electrons is applied to the grid, the
first element of the lens and the deflecting plates. In this case, field intensity in the
area of the grid on the screen hole is small, so the field does not practically deflect
the electrons in the grid area. Having entered the lens system through the screen
hole, the electrons are retarded in three stages down to the energy necessary to be
laced up by the analyzer and focused into an almost parallel beam on the entrance
slit. Both analyzers lace the electrons of identical energy. The electrons laced up
through the analyzers are accelerated in two stages and received by the electron
multiplier.
A detailed electric circuit of measuring is given in [44]. The resolution deter-
mined by the Young-Kuyatt criterion [40] is 15 meV at the temperature of 150 K.
There are also modifications of such field-electron total energy analyzers, e.g.
[45, 46].
2.3 Field Emission Microscopes of Universal Design
Under current conditions, most interesting results can come from improved per-
formance of research in the first place and application of universal devices com-
bining several techniques supplementing each other in the second.
The first condition is met by application of field emission microscopes, which
makes it possible to research several samples, e.g. a cassette intended for several
samples, as well as linking devices able to replace the samples without deterioration
in the working vacuum in the primary chamber of the microscope. The linking
devices are widespread in transmission and scanning electron microscopes.
However, they are designed for relatively low vacuum and have no rigid fixation of
samples necessary to provide electric contacts. Therefore, such linking devices
rated at the pressure of at least 10−7 mm Hg are designed on a case-to-case basis or
the available devices are modernized, e.g. [47, 48].
It is most natural to research the surface structure by a combination of a field

emission microscope and a field-ion microscope in one device. In this case, we can
see both the geometrical arrangement and the energy structure of the apex surface.
In this case, the field emission image and the field-ion image can be separated by
an electromagnet [49], reposition of respective screens [50] or simple freezing-out
of residual gases at the temperature of 4.2 K when the imaging gas is not evacuated
by the backing pump [51]. The first design produces the field emission image and
the field-ion image at the same time, and the two other designs produce them one
after another.
The design of a universal field emission-and-field-ion microscope combining
many of the above-listed devices is given below.
A microscope [52] (Fig. 2.22) is a vacuum stainless-steel chamber (1) of
250 mm in diameter and 400 mm of height connected with an auxiliary volume (9).
The chamber has four DN-160 flanges to connect observation windows, a lumi-
nescent screen, a field-electron total energy analyzer and an atom probe, as well as
eight DN-50 flanges to connect sensors and auxiliary devices.
The upper DN-250 flange (3) hosts all sample-reposition drives and their electric
connections. The luminescent screen (2) is mounted on the front lateral DN-160
flange. The four DN-160 flanges of the volume (9) are to connect the evacuation
devices to the microscope chamber.
The chamber kinematics determine a 6-sample cassette, each being connected
with insulated flexible leads through ceramic current leads with external electric
circuits. Having been aligned to the screen, any sample can rotate about the axis
parallel to the chamber axis within 0–360° and perpendicular to it within ±45°,
while the apex-sample vertex is in the point of intersection of these axes. The
chamber is warmed up by four KG-1000-3 quartz lamps mounted in the chamber.
That enables the intrachamber devices to be efficiently warmed up by infrared
radiation. Warm-up with usual heaters takes a long time and is inefficient. In this
installation, the external spiral heaters are used to warm-up the auxiliary volume (9)
and the ducts bearing no internal kinematics.
The vacuum chamber is deplenished with the VN-2MG backing pump (4)
through valve B3. After achievement of preliminary leak-in, valve B3 is closed and
evacuation takes places through valves B4, B6, the nitrogen trap (5) by
turbo-molecular pump VMN-150 (6). Under such conditions, vacuum in the
chamber reaches 5 10−9 mm Hg.
If there is a need in harder vacuum, valve B5 is opened and valve B4 is closed
and the chamber separated from the low-vacuum parts of the installation is
exhausted by pump Trion-150 (7) to the pressure of about 10−10–10−11 mm Hg in
parallel with cooling of the pump (7) and the trap (8) with liquid nitrogen.
Valve B4 has a copper seal with flow section DN-100, and valve B5 has a
liquid-metal seal with flow section DN-160 [53]. After a short-time leak-into the
chamber caused e.g. by sample change, when it is inexpedient to stop the
turbo-molecular pump, the chamber is empties through the bypass line connecting
the backing pump (4) and the chamber (1) through valve B3.
2.3 Field Emission Microscopes of Universal Design 69
Fig. 2.22 Device schematics: 1 vacuum chamber; 2 luminescent screen; 3 upper flange with drive
units for samples; 4 VN-2MG forevacuum pump; 5, 8 nitrogen traps; 6 VMN-150 turbomolecular
pump; 7 Trion-150 magnetic charge pump; 9 additional volume
In the deplenishing system, the total pressure is measured with vacuum gauge
VTB-1, the low and medium vacuum with vacuum gauge VMB-8 and high and
ultra-high vacuum with vacuum gauge VI-14.
The imaging gas necessary to form a field-ion image or process the cathodes in a
special way is puffed to the chamber through leak valve B2, and the chamber gas
components are measured by partial-pressure gage IPDO-2.
High voltage is supplied from the high-voltage (0–40 kV) unit to the lumines-
cent screen of the microscope and measured by a digital voltmeter through a
calibrated voltage divider. The work function change is measured with a stabilized
power source [54]. The potential of the samples is close to the earth potential, which
permits their warming up from the filament voltage unit of a conventional design in
case of need and essentially simplifies sample control kinematics. The emission
current is measured by electrometric voltmeter VK2-16 or VK2-20, depending on
the range of the current to be measured.
The volt-ampere characteristics are recorded with an x/y-plotter or a computer
with built-in ADC plate connected to outputs VK2-16 and a high-voltage unit
divider. Some particular experiments involved different measuring devices as well.
The luminescent screen design is shown in Fig. 2.23. This design [55] is special
in its ability to supply high voltage to the luminescent screen (2) itself by parti-
tioning of functions of the screen (2) and the observation glass (8), which essentially
simplifies the design and improves reliability of the cathode assembly of the
microscope.
The screen (2) is a flat glass plate of 100 mm in diameter having a conductive
coating and a phosphor layer.
The screen (2) is fixed between two rings (3, 4) and separated from the case with
three ceramic insulators (5) attached to the flange (1) through screw pairs (6) able to
relocate the screen, thereby changing the anode–cathode distance, i.e. magnification
of the microscope. High voltage is supplied to the screen (2) through a ceramic
high-voltage current lead (9) protect with the screen (10).
The luminescent screen designed like that can be promptly swapped, and par-
titioning of functions of the observation window and the screen makes it possible to
fabricate screens of larger dimensions irrespective of the observation-glass size.
Microscope kinematics. The universal field emission microscope kinematics
should meet the following conditions:
Fig. 2.23 Construction of a

luminescent screen: 1 flange;
2 luminescent screen; 3, 4
metal fastening rings; 5
ceramic insulators; 6 screw
pair; 7 flange of the chamber
casing; 8 looking glass; 9
ceramic high voltage current
lead-in; 10 protective screen
• sample repositioning in respect of the probe hole without re-configuration of the

electric field;
• bonding able to provide sample heating and current measurement;
• cooling of the sample down to the temperature of liquid helium;
• it is desirable to use a sample cassette to raise productivity of the installation as a
whole.
Various versions of sample re-positioning in the vacuum chamber are discussed
in [56–59]. They are manipulators providing rotation of the sample in respect of
two mutually perpendicular axes. However, these designs provide rotation of the
sample with deflection from the probe-hole center, which changes trajectories of
electrons and ions, thereby downgrading the characteristics of the devices. The
feature affected most is resolution.
To secure the sample vertex in one point of the design described, the sample is
rotated in respect of two mutually perpendicular axes so that the sample vertex is in
the interception of the axes. The entire kinematics are mounted on the upper flange
of the chamber (1) (Fig. 2.24).
Fig. 2.24 Kinematic schematics of the microscope: 1 upper flange of the chamber; 2, 3, 6 pulleys;
4 cross-beam; 5 brace; 7 sample; 8, 9 gears; 10 screw drive; 11 cassette with samples; 12 Dewar
flask; 13 moving and rotation input; 14 electric vacuum inputs; 15 rotation inputs
The design includes a cassette (11) simultaneously holding six samples (7) with
independent electrically insulated leads and general cooling of samples through a
flexible heat conductor from a Dewar vessel (12). Every sample can be placed into
the working position opposite the luminescent screen and turned as stated above.
The sample positioned is rotated about the vertical axis of the chamber with one
of the bellow rotary-motion feedthrough (15) placed on flange DN-50.
A cable transmits rotation from the feedthrough (15) via the pulley (2) to the
pulley (3), which is rigidly connected with the balance beam (4). Gyration of the
balance beam (4) provides rotation of the positioned sample about the vertical axis
within 0–360°. The angle of rotation is only limited by the danger of twisting the
heat conductor and the interconnecting electric wires.
The sample positioned is rotated about the horizontal axis of the chamber by the
rotary-motion and repositioning feedthrough (13).
To perform the rotation, the feedthrough (13) is engaged with the joint box,
which is rigidly connected with the pulley (6). A cable transmits rotation from the
pulley (6) to the pulley set on the worm spindle. The worm wheel (10) is set on the
cramp axle (5), where the sample cassette (11) can run in bearings. The cramp and
the cassette are designed so that the end of the sample (7) lies exactly on the
interception of the axes and the rotation does not change its position in space. The
worm gear can solve the problem of fixing the sample after the rotation has been
over.
Application of a rotary-motion and re-positioning feedthrough is caused by a
necessity of kinematic unlocking of axial rotations because if the interlocking were
rigid, rotation about the vertical axis would transmit a torque to the cramp (5),
which would cause rotation about the chamber axis as well. The samples are
changed by rotation of the cassette (11) with gear wheels (8, 9) and one of the
rotary-motion feedthroughs (15).
The samples are electrically connected through ceramic current leads (14) by
flexible wires (missing from the illustration) tubed in ceramics to prevent short
circuit.
The samples are cooled by a flexible heat conductor from the double Dewar
vessel (12), cooling the sample down to helium temperatures. In this case, the
internal tank is filled up with helium, and the external tank, separated from the
internal tank by a vacuum gap, is filled up with fluid nitrogen.
It is expedient for such large and complex high-vacuum systems as a field
emission-and-field-ion microscope to use cassettes able to load the vacuum
chamber with several samples at once, which essentially reduces the research time.
Operation in ultra-high vacuum necessitated development of a rotary sample cas-
sette [60] (Fig. 2.25). It can hold 6 samples with independent electrical connection
at a time, heat the sample up to the temperature of fusion and cool down to helium
temperature. Every sample can be placed into the working position opposite the
luminescent screen and turned in respect of the two mutually perpendicular axes in
such a way that the sample vertex is always in the interception of the axes, which is
necessary when working with probe holes. It seems optimal to have six samples in
the cassette. A larger number would not only essentially complicate the sample

sample rotation cassette:
1 sample; 2 arch; 3 screws;
4 cross-arm; 5 copper
segments; 6 fastening screws;
7 copper heat transfer;
8 ceramic plate; 9 rotation
input; 10, 11 gears;
12 ball-bearing; 13 bearing
collar; 14 base of the cassette;
15 sapphire plate; 16 hole in
the segment; 17 ceramic
insulation; 18 flexible electric
wires; 20 ball; 21 collet;
22 spring
design, but also have mutual effect on emission characteristics because of a shorter
distance between them.
The samples (1) secured on bails (2) are screwed (3) in the holes (16) of copper
segments (5) separated from each other by vacuum gaps. The segments (5) are
pressed to the heat conductor (19) through a thin sapphirite plate (15) by screws (6).
The screws (6) are separated from the segments with a ceramic plate (8). The heat
conductor (19) is pressed to the foundation (14) by screwing (6). The foundation
can freely rotate in the ball-bearing (12) secured in the holder (13) and, therefore, in
the cramp (4). To reduce heat losses, the heat conductor and the foundation contact
only in the screwing areas (6).
The sample (1) is fixed in the working position by a ball (20) pressed to the
foundation (14) with a spring (22) through a steel insert (21).
The foundation (14) has six grooves located along the same radii with the
samples. The groove size is about 2/3 diameter of the ball, which does not let the
ball slip into the groove, on the one hand, and secures the foundation on the slit, on
the other hand.
To change the samples, the gear wheel (11) rigidly secured on the foundation
axis (14) is engaged with the gear wheel (10) turned by the bellow rotary-motion
feedthrough (9). When the gear wheel (10) rotates, the gear wheel (11) turns
together with the foundation (14), the ball (20) leaves the fixing groove and rolls
along the lateral surface of the foundation to the next groove. The number of
gear-wheel teeth (10) provides rotation of the gear wheel (11) through 60°, i.e. the
value of the shift to the next groove and, therefore, the next sample. This system can
change samples in any direction from the first to the last sample and in the reverse
order.
The samples are electrically connected with external circuits by flexible wires
(18) tubed in ceramics (17). The samples are cooled from the Dewar vessel (po-
sition (12), Fig. 2.24) through a flexible heat conductor (19) pressed into a rela-
tively massive copper heat conductor.
The gear wheels (10) and (11) are only engaged at the moment of change of the
sample. That is why the sample placed into the working position can freely rotate
about the two mutually perpendicular axes. The angles of rotation are only limited
by the wires (18) and the heat conductor (19).
Certainly, that is not a full list of field emission microscopes. This section shows
only the general basic principles of design of such microscopes.
2.4 Electronic Systems
The most important conditions of fruitful experiments include availability of an

appropriate measuring system. For an elementary field emission experiment, it is
enough to have a high-voltage source and a low-current meter, e.g. a galvanometer
[61].
As the high-voltage source is the core of any field emission measuring system,
radioelectronics developed a great number of its circuits and designs, numerous
examples being available in the journals Instruments and Experimental Techniques
and Review of Scientific Instruments. Recently the most promising among them are
the high-voltage sources with a frequency transformer and computer controllability.
They are characterized by small dimensions, high stability and ease of control, e.g.
[62], which makes it possible to carry out new kinds of experiments increasing our
knowledge of field emission processes. A circuit similar but more advanced towards
universality of output voltages (ranging within ±30 kV) and control reliability is
given in [63]. This paper is not to analyze specific details of circuits and designs of
high-voltage power sources, as it is a vast field of experimental technology of
special interest.
Neither are we going to consider current meters because they are mostly standard
factory-supplied devices differing in the ranges of current, accuracy class and
performance.
2.4 Electronic Systems 75
Naturally, improvement of computer facilities has resulted in development of

numerous methods and systems of experimental data processing, e.g. [64, 65]. At
present, practically every experimental field emission research installation should
be computer-assisted. A flow chart of such a system in the version that is the
simplest though able to carry out most of the desired experiments and provide
various techniques is given in Fig. 2.26 [66]. The basic element linking the com-
puter with measuring and controlling devices is the input-output controller pro-
viding reception and delivery of information to necessary addresses. The central
member of the system is a computer-controlled high-voltage stabilizer. The output
voltage is supplied to the experimental device, as well as the digital voltmeter (for
visual voltage control) and the analogue-to-digital converter, which outputs the
code meeting the output voltage of the power source. Conditionally, the emission
current is measured by an amperemeter with the desired number of automatically
switchable ranges. Our experiments cover 6 ranges 100 nA–10 mA. After appro-
priate processing, the outcome of the experiment can be plotted. The boundary
measuring frequency depends on the element base and the circuit design of the
analogue-to-digital converters and the input-output controller.
Naturally, no system can cover the entire variety of situations, especially in
application programs of field emission cathodes. Therefore, particular interesting
circuit designs are to be featured in the respective sections.
Fig. 2.26 Schematic of the field cathode emission characteristic measurement system
2.5 Emitter Technology
As already noted, a necessary condition to develop the values of field emission

currents sufficient for observation or practical use is an intense electric field at the
emitter surface (*106–107 V/cm).
Under acceptable working voltages of approximately 0.1–10 kV, such electric
intensity can be mostly achieved by reduction of the curvature radius of the vertices
of the emitting surfaces, which should look like tips, microelevations etc.
The emitter-tip curvature radii can reach 10 Å. Therefore, high-quality results of
research in the field emission technology largely depend on fabrication of samples of
the necessary shape and structure. In this chapter, we mean the external structure of
the emitter, i.e. the apex curvature radius, the cone angle, selection of the necessary
crystallographic faces at the tip apex etc., depending on the purpose of the experiment.
Research of the material structure, distribution of the electron work function by
single-crystal faces, adsorption-desorption processes, interactions of radiation with
a substance, field emission etc. needs appropriate emitter technologies.
It became clear as long ago as at the beginning of field emission research. Over
the past 70 years, emitter technologies were continuously improved. They go on
developing at present and will be even more sophisticated in the future.
This section provides the main techniques and technologies of field emitters
known for today.
Figure 2.27 shows the main classes of field emitters and the mix of materials
used to fabricate them.
This section features the field emitters fabricated with the emitting-surface
geometry prescribed or close to the prescribed. Practically, they can be fabricated
from any material and used throughout the range of field emission research and in
some practical applications.
2.5.1 Metal Tips
Practically, most well-proven and handy workpieces for metal-apex emitters are
thin wires of 20–200 lm in diameter, though nearly all workpiece shapes will do.
For many metals, they are produced on the industrial scale or fabricated under
experimental conditions, which is to be featured below.
In case of massive materials, a workpiece with a square or round cross-section is
fabricated. Usually it is less than 0.5 mm in diameter and about 1 cm in length. As a
rule, the workpieces are fabricated by ordinary mechanical methods, such as
low-speed diamond sawing, electric spark machining, centerless grinding, wire
drawing, warm forming or extrusion [67].
Selection of the method depends on plasticity and ductility of the material with
account of the effect of the processing technology on the material microstructure.
E.g. wire drawing, warm forming or extrusion are very effective methods to obtain a
2.5 Emitter Technology 77
Fig. 2.27 Review diagram of the main classes of field electron emitters
round cross-section, but they extremely increase dislocation density and change the
grain shape. Polishing removes external layers of the material, thereby eliminating
some injuries caused by the previous machining, but leaving the defects caused by
the operations of a wire-drawing type. If the fabrication method gives a
rectangular-section workpiece, it is important for the sides to be 1:1. Otherwise, the
sample would be shaped as a knife. Right angles usually round in the course of
polishing, resulting in the workpieces with a round cross-section. The samples of a
definite shape or made from particular materials are manufactured with special
methods described in the following sections.
If the materials are not ductile enough for the appropriate operations of drawing
or needed in a metastable state, the wires can be fabricated by rapid-solidification
methods [68].
In fabrication of a liquid-rolled band, the basic drawing process can often give a
by-product of thin metal threads, which can be used as workpieces for field-emitter
samples, despite their rather irregular cross-section [69]. The best method is
drawing of a thread from a melt into a rotating liquid. In this process, the melted
material of the sample is extruded through a hole into a cooling liquid, forming a
continuous wire [70, 71] (Fig. 2.28).
The drum is rotated at a velocity of about 250 rev/min, and the cooling liquid is
introduced into the internal part of the drum, where the material is shaped as a film.
The metal is melted in a high-frequency induction furnace and extruded by pressure
of argon through a quartz or ruby nozzle of 50–300 lm in diameter into the rotating
layer of the coolant. Continuous threads develop when the cooling-surface tan-
gential velocity is 1.15–1.3 times higher than the metal-jet feeding rate. The tan-
gential velocity, the layer thickness and the cooling-liquid rate, the extrusion rate,
the temperature and the metal-jet incidence angle, the nozzle size and the distance
between the nozzle and the cooling liquid, as well as the drum rotation rate are
extremely important experimental parameters [72, 73]. If they are properly selected,
a large amount of a small-diameter wire can be fabricated within one work cycle.
This method also creates fine or amorphous materials due to rapid solidification at
high cooling velocities (103–105 K/s). That is the process to fabricate threads from
the alloys of Fe–C–Si, Pd–Cu–Si, Ni3Al, Cu and Al.
A continuous wire can also be fabricated directly from the liquid phase [74]. The
sample material melts and gradually runs out to the rotating cold cylinder
(Fig. 2.29). Unlike the band fabrication process, the flat cross-section of the cold
cylinder slightly changes in that it has a small prominence where the metal lands.
Fig. 2.28 The process of

pulling a thread out of a melt
into rotating water. a Front
view; b side view. 1 Wire;
2 pressurized argon; 3 electric
oven; 4 liquid melt; 5 wire;
6 water; 7 rotating drum

producing the wire directly
from the melt. Longitudinal
section of the rotating cold
cylinder is shown
The wire section size is usually 40–80 lm. The prominence makes it non-circular.
The method advantage is in production of a great deal of very fine material.
2.5.2 Machining
Many materials, in particular single crystals, molding materials, such as e.g. gra-
phites or high-temperature ceramics, are output as massive workpieces. Therefore, a
high-quality apex comes from a thin workpiece. At the first stage, the machining is
usually mechanical, e.g. by skive, and results in studs of 0.3 0.3 10 mm. The
dimensions are approximate and depend on the nature of the material and the
research objectives. Electric-spark machining is not desirable for blanking because
it can occasionally result in doping of the stud surface with reaction products.
Paper [75] describes the fabrication technology of field tip emitters from Ge and
Si. At the first stage, rectangular workpieces are cut out of Ge single-crystal ingots
oriented in the direction of (100) with the accuracy of ±1° (Fig. 2.30a). Resistance
heating of the emitter is provided by ultrasonic cutting approximately to the middle
of the sample on one of its ends so that the sample is fork-shaped (Fig. 2.30b). After
that, the workpiece is luted to the holder with quartz putty and placed in the chuck
of the jeweler’s lathe. The uncut end of the workpiece is ground with a diamond
wheel to form a cylinder of 0.38 mm in diameter (Fig. 2.30c). The bridge is made
slightly thinner on the fork than on the ends, which enables to heat up the emitter
almost to the melting point, whereas at the clamps, the sample remains cold. That
makes it possible to minimize temperature migration of impurities from the holder
to the apex of the emitter. The processed workpieces are cleaned in methanol under
gentle ultrasonic mixing. After that, the surface pollution is removed by the etching
agent, consisting of 25 parts of smoking HNO3, 15 parts of HF and 15 parts of ice
acetic acid by volume. Emitters are formed by electrolytic etching of cylindrical
workpieces in cold etching agent at the temperature of 10 °C (cf. below). To
preserve the apex shape after the scattering, the cone-apex angle should be about 5°.
Fig. 2.30 Production of field

emitters out of silicon and
germanium. a Blank with a
given axial orientation;
b creating a “fork” with
ultrasound; c the blank after
being lathed; d formed emitter
To provide effective heat transfer to the apex, the emitter is 1 mm long. Formation
of the apex is usually followed by urgent measures of neutralization and removal of
etching residues. The final stage of emitter fabrication is shown in Fig. 2.30d.
2.5.3 Foils, Sheets
A more sophisticated technology can result in making samples of rather thick sheets.
One of such methods is suggested in [76]. It can fabricate sample workpieces of the
Fig. 2.31 An illustration of forming sample blanks out of a thin band. a after the band is
thoroughly cleaned its surface is covered with a thin film of photoresistant material; b the
photoresistant film hardens forming two masks on upper and lower surface, and the construction is
exposed to UV radiation in order to transfer the image to the photoresist; c the masks are removed
and the photoresist is developed by dissolving the films that were not exposed to UV. The
construction is then dried off to harden the photoresist film; d the construction is electropolished to
remove the parts of surface not covered with photoresist; e the electropolishing continues until the
band is etched through and the sample blanks are obtained (1 photoresist; 2 sample; 3 mask)
sheets both as thick as more than 2 mm and as thin as 0.2 mm. Figure 2.31 repre-
sents the operating sequence of sample fabrication with this method.
With this end in view, the sheets of e.g. tungsten or stainless steel are degreased
for 30 min in the concentrated alkali of NaOH, followed by coating with a pho-
toresist. Then they are repeatedly washed with water, dried up and then sunk in the
photoresist solution and re-dried at t = 100 °C for 30 min. This process is repeated
several times to remove possible defects caused by air bubbles or dust present is the
photoresist solution. The unmasked photoresist part is exposed to ultraviolet radi-
ation for 2 min with a high-pressure mercury arc lamp placed 0.14 m away from
the object surface. After unmasking, the sheet is kept in the etching solution to
remove the unexposed photoresist part. To solidify the remaining photoresist, the
sheet is warmed up for 30 min at t = 140 °C. When the photoresist film is abso-
lutely dry, the upper and lower surfaces are covered with photomasks.
The photomask pair is usually made of two films with appropriate printed
drawings. The photomasks should be installed accurately so that the upper and the
lower exposed regions were properly oriented. The lines of the two photomasks can
be located either opposite each other or with a shift. The drawing pitch changes to
meet the band thickness, which results in an approximately square cross-section.
If the photomasks are placed with a shift, the drawing pitch should be approx-
imately equal to two thicknesses of the sample. The photoresist layer is exposed to
UV radiation to fix the print. The photoresist film is developed, solidified and then
removed from the regions that were not exposed to light. In these places, the sheet
is etched through to shape the workpieces. One operation can produce a great
number of the workpieces. This method can give the samples of metal glass [77],
stainless steel and tungsten [78].
The photoresist method is also applied to fabricate the workpieces of emitters
from the surface layer of a thick material (Fig. 2.32) [76]. After the surface is coated
with a photoresist, exposed in ultraviolet through the photomask and developed, the
grooved sides are masked and etching goes on until the grooves on each side of the
photoresist are etched through and a workpiece is formed.
The simplest way to fabricate apex emitters from wire workpieces includes
pointing of the wire in the gas-burner flame [79–81].
This method assumes that the wire material is oxidized at high temperature in the
presence of air oxygen, the oxide being continuously removed. For purely metallic
wires, it is difficult to fabricate high-quality apices in this way. However, this method
is good for the materials containing highly-volatile oxides. E.g. the same method can
shape tips from some materials such as glass carbon [77]. The tip is shaped as a tip of a
vitreous carbon material, e.g. furan, phenol, pyrrole or vinyl resins obtained from
divinylbenzene and carbonization of the solidified and shaped resin in vacuum or an
inert-gas medium. Carbonization takes place under rise in temperature with particular
velocities within particular temperature ranges. Eventually, the apex tip is etched in
the flame of three symmetrical gas burners to obtain the radius of 1000–3000 Å.
Chemical etching was widely used in the earliest days of field emission
(cf. [82, 83]). However, chemical etching is rather difficult to control and the resulting
tips have a wide spread of curvature radii. This method is only up to a highly-qualified
experimenter, which is why it is mostly applied to low-conductance materials.
Fig. 2.32 Illustration on

forming sample blanks out of
a volumetric layer of material
using photoresist method
similar to one shown on
Fig. 2.31. The construction is
electropolished until ditches
form between the areas
covered with photoresist. The
ditches’ sides are then masked
and electropolishing
continues until they etch
through
Electrochemical etching of wire workpieces to fabricate apex samples intended

for field emission research was suggested in the late 1930s [84–86]. Electrochemical
etching in an electrolyte bath is the most widely used and simplest way to fabricate
apex emitters from wire workpieces of 0.05–0.2 mm in diameter [87, 88].
As of today, there are quite a lot of patterns and modifications of apex elec-
trolytic etching systems. However, they can all be grouped according to four
essential features (Fig. 2.33):
1—electrolyte bath type;
2—electrolyte type;
3—electrolyte-feeding method;
4—etching-cessation method.
Fig. 2.33 Classification of methods of electrolytic etching of the tips
Below we exemplify implementation of the main classes of this process. It has to

be noted that the apex/sample fabrication procedure cannot be mechanically
reduced to selection of classification parameters because its major task is to show
the ample possibilities of electrolytic etching, which has to be used to attain the
goals of the research. The fabrication procedure is not trivial even for such more or
less standard materials as tungsten, iridium or platinum. For relatively new mate-
rials, such as e.g. aluminum, silicon, uranium etc., it borders on art and depends not
so much on the fabrication method as on the experimenter’s due care and individual
craftsmanship.
This method makes the tip microscopically smooth due to the electropolishing
effect, i.e. preferential stripping of the prominent places where the concentration of
the electric field is elevated. Change of geometrical configuration of the electrolytic
bath, concentration and chemical composition of the electrolyte, the value and the
form of the applied voltage can largely change the apex curvature radius and the
shape of the tip, e.g. the angle and the form of the cone.
The fifty-years-plus history of development of electrochemical etching methods
have brought a great number of designs of electrochemical waves, electrolyte
compositions and apex-fabrication techniques, e.g. [89–92].
Briefly, the essence of electrochemical etching is as follows [93, 94].
Direct or alternating voltage is applied to the wire workpiece immersed in the
electrolyte and the auxiliary electrode (Fig. 2.34a). Its value depends on the
workpiece diameter, the dimensions and configuration of the device used for
etching, the etching-agent composition and temperature. Therefore, it is expedient
to select its value on the experimental basis for each particular case and use the
values given in this section for reference only.
When voltage is applied, the wire workpiece conducts current and creates a
potential difference UB > UA along its length. As etching-current density is directly
proportional to voltage and inversely proportional to the curvature radius of the
surface stripped, the etching-current density in point B is higher than in point A. In
the course of etching, the current falls in ratio to decrease in the wire diameter, and
resistance of the region stripped increases in inverse proportion to the diameter
squared (pos. 2, 3 in Fig. 2.34b). The drop in voltage along the region immersed in
the electrolyte increases, and the vicinity of point B experiences more intensive
etching. Quicker reduction of the diameter in this place accelerates the etching rate,
which leads to development of a constriction (pos. 4 in Fig. 2.34b).
Fig. 2.34 Schematic of volumetric etching of the tips (a) and steps (1–7) of electrolytic etching of
the tip (b) along with the time curves of etching current and its derivative: 1 arch; 2 tip blank;
3 tub; 4 electrolyte; 5 additional electrode
Further reduction of the constriction diameter necessitates a point of time when

the weight force of the lower part of the workpiece, the Archimedes buoyant force
included, exceeds the wire tenacity in the vicinity of the constriction, i.e.
nd 2 rB
\P; ð2:9Þ
4
where d is the current constriction diameter, rB is the wire-material tensile strength,

P is the weight of the lower part of the workpiece, the Archimedes buoyant force
included. At this moment, the lower part of the workpiece tears off (pos. 5 in
Fig. 2.34b). The upper end of the lower, torn-off part of the workpiece in the
vicinity of the constriction and the lower part of the wire in the vicinity of the
constriction is shaped as an apex with a hemisphere on the end. The samples are
mostly made of the upper part attached to the holder, but the lower part is also
sometimes used if entrapped into special glasses. At the moment when the lower
part of the workpiece tears off, the constriction diameter equals
rffiffiffiffiffiffiffi
qLl
dc ¼ dl ; ð2:10Þ
rB
where q is the specific gravity of the material, the weight force of the displaced
liquid included, dl is the average diameter of the lower part of the workpiece, Ll is
an equivalent length of the lower part, which is determined from the condition that
the volume of the lower part at the tearing-off moment equals
Ll pdl2
Vl ¼ : ð2:11Þ
4
Eventually, the apex-curvature radius equals

rffiffiffiffiffiffiffi
dc dl Ll q
r ¼ : ð2:12Þ
2 2 rB
Estimates show that r can reach the value of *10 Å.

Formula (2.10) is only true if the workpiece-material density exceeds the liquid
density.
If the workpiece-material density is less than the liquid density, this apex etching
method does not work.
However, the etching process does not stop after the lower part has been torn off
because of the contact with the electrolyte, which leads to increase in the cone
angle and the apex-curvature radius (pos. 6 in Fig. 2.34b). Unless the process is
stopped, it goes on up to complete wire stripping in the electrolyte and loss of
electric contact with it (pos. 7 in Fig. 2.34b).
The current dynamics in the course of etching at constant voltage is shown in
Fig. 2.34b. In the process of etching of a thin workpiece, current monotonously
decreases until the lower part of the workpiece tears off (pos. 5 in Fig. 2.34b), after
which the current value slumps. At that moment, the current-derivative diagram
shows a surge of a rather high amplitude. To fabricate an apex with the minimum
curvature radius, it is necessary to stop the etching process either mechanically
(using the design features of the etching systems) or electrically (when the elec-
tronic circuit fixes a current drop or a peak in the current derivative and disconnects
supply, e.g. [95]) from the moment of tearing-off of the lower part of the workpiece.
Efficiency of such systems and, accordingly, the tip curvature radius are determined
by the response time from the moment of tearing-off of the lower part.
Currently, there are quite a lot of patterns and modifications of tip
electrolytic-etching systems.
Below we give three examples of implementation of the main classes of this
process. It is worth noting that the simplest electrolyte for fabrication of
samples/apices from carbon fibers is the 1N solution of NaOH or KOH at the
alternating voltage of *1 3 V.
One of the most conventional and simplest tips etching methods is the method of
volume etching (Fig. 2.34).
When positive voltage is applied to a metal sample and the sample is immersed
in an electrolyte of the opposite polarity, voltage build-up first causes increase in
current, which meets the course of electric etching processes. With further increase
in voltage, current reaches the plateau and the etching process turns into
electropolishing.
At even higher voltages, the sample surface blisters with gas, which can cause
pitting etching of the sample. The electrolyte usually contains an oxidizer and a
solvent to remove the anode-etching products. Electropolishing often produces a
viscous layer denser than that of the electrolyte. It flows down the wire and insu-
lates its lower end. In the course of this process, the material is most effectively
removed from the region located above the wire tip, thereby gradually reducing the
cone angle in the immediate proximity from the developing apex. Under these
conditions, the rising bubbles can prevent further impact on the upper parts of the
sample, thereby causing an undesirable increase in the cone angle.
Plastic deformation changes the mechanical properties of the surface layer near
which the wire is cut. Therefore, tip fabrication follows removal of the damaged
area by electropolishing.
Simple dipping in an electrolyte produces apices of practically any cone angle
depending on the electrolyte depth the workpiece is immersed at. A small cone
angle results from immersion at the depth of 3–5 mm. A large angle results from
immersion at the depth of 1 mm. Such a method can be successfully used to
fabricate tips from easily etchable materials. However, this method is not good at
sample-shape reproducibility and needs much training, which is especially
important.
The most widespread ways to fabricate apex samples include etching in an
electrolyte droplet [79] diagrammed in Fig. 2.35a. A glass tray (1) is filled with a
current-conducting solution which is usually a saturated NaCl solution. The latter
holds an electrode (6) shaped as a plate of the material non-reactive with the
Fig. 2.35 Schematic of electrolytic etching in a drop (a) and wire loop (b): 1 tub; 2 conductive
solution; 3 emitter blank; 4 diaphragm; 5 drop of electrolyte; 6 additional electrode; 7 wire loop
solution, e.g. nickel. A workpiece (3) is put in the center of a small hole in the plate
(4). An electrolyte droplet (5) is pipetted in the hole and retained inside by surface
tension forces. When voltage is supplied, electric current flows through the elec-
trolyte droplet, where it initiates electropolishing.
Another good thing about this method is that etching automatically ceases when
a workpiece is over-etched since its lower part falls off and breaks the electric
circuit.
Formation of tips depends not only on spatial distribution of the concentration
layer consisting of reaction products, but also on the composition, concentration
and temperature of the electrolyte and the sample material [96]. It has also been
demonstrated that gassing that accompanies etching and is to a great extent the
cause of poor reproducibility of the tips shape can be practically eliminated by
adequate selection of the etching mode.
This method prevents the effect of plastic deformation caused by cutting of wire
into workpieces since the wire tip does not belong to the etching area and, con-
sequently, the tip area.
Very thin samples, as well as the apices subject to re-polishing, are etched in an
electrolyte droplet [92] (Fig. 2.35b) filling a loop of 3–5 mm in diameter made of a
thin wire, such as a tantalum or platinum wire of 0.2 mm in diameter.
The loop is filled up with the electrolyte and moved along the needle at the
moment of voltage application usually in a form of a series of manually controlled
impulses lasting for about 0.2–2 s. [97]. This procedure results in desirable shaping
of the needle-apex area by change in duration of polishing or the impulse number.
The desirable accuracy of positioning is obtained by mounting of the sample or the
wire loop on the micromanipulator. The process is usually controlled with a light
microscope. The small volume of the electrolyte necessitates its frequent substi-
tution. As a rule, the electrolytes are the same as at the second stage of usual
electropolishing.
Originally, the micropolishing technique was developed to taper filamentary
crystals [98]. However, it is mostly used to taper the samples that cannot be formed
by usual electropolishing or re-taper the samples blunted as a result of exploitation.
The technique can also be used to fabricate samples from unusually-shaped
workpieces. E.g. tips can be fabricated from singular droplets of fast-curing alloys
[99].
Good reproducibility of polishing of sharp tips during their fabrication is pro-
vided by regulation of the following conditions of polishing: (1) electrolyte solu-
tion; (2) electrolyte temperature; (3) distance between the cathode and the tip;
(4) the value of alternating or direct voltage; (5) frequency of alternating voltage;
(6) number of alternating-voltage periods or duration of constant-voltage impulses.
Paper [100] suggests a power source able to work in the following four modes:
(1) electropolishing with alternating current; (2) electropolishing with direct cur-
rent; (3) automatic polishing with alternating or direct current; (4) charge integra-
tion in any of the three modes.
To reduce the tip-radius spread and the curvature radius down to *100 Å, paper
[101] suggests an advanced method of «a droplet in a loop». Figure 2.36 demon-
strates a simple experimental needle-fabrication device. The metal-wire segment of
a typical thickness ranging from one to several hundreds of micrometers is placed
so that it crosses the region of interface between two liquids. Lower liquid (2) is
denser. It does not conduct current. Usually, that is CCl4. Upper liquid (1) is an
electrolyte. One electrode is wire (6) subject to etching. Electrode (4) has a ring on
the end. When the diameter of the etched wire (6) area reaches a critical value, a
part of wire (5) held in a non-conducting liquid gets detached under action of
gravity. Due to a clear-cut interface between liquids (1) and (2), the detached part of
wire (5) that gets into the SF trap has an extremely sharp tip with the curvature
radius down to *10 nm.
A design similar in principle but without a loop is presented in Fig. 2.37 [102].

device for creating sharp
needles out of various metals
using electrochemical
methods 1 electrolyte; 2
non-conductive liquid; 3, 4
electrodes; 5 sharp needle to
be used as an emitter; 6
remaining piece of the wire; 7
trap; V applied voltage
Fig. 2.37 Etching device. 1 blank; 2 electrolyte layer; 3 hole in the bottom of the tub; 4 receiving
glass; 5 prepared electrode; 6 pipe for supply of the electrolyte for etching; 7 hole for draining the
electrolyte out
Wire workpiece (1) is placed vertically to pass through electrolyte layer (2) and
hole (3) in the bottom of the bath. The hole has a diameter sufficient to hold the
electrolyte due to surface tension. The largest diameter of the hole can be evaluated
by the formula
4c
D¼ ;
nd
where D is the hole diameter; c is the electrolyte surface-tension co-efficient; d is its

specific gravity; nd is the electrolyte-layer thickness. One current-source pole is
connected with the upper part of the workpiece, while another pole is connected
with the electrolyte.
The needle electrode is made of a part of the workpiece located below the
etching layer. Etching goes on until the lower part of the workpiece gets detached
and falls off under action of gravity. When detached, the fabricated electrode ter-
minates the connection to the current source, falls off and promptly leaves the
etching agent.
The tip is protected from damage and pollution by receiving socket (4), where
the fabricated electrode falls off and perks up. The electrode is longer than the
socket diagonal, so the apex does not touch the walls.
In order to fabricate quite similarly shaped apices, etching of workpieces in a
volume involves different electronic circuits switching power supply off upon ter-
mination of etching [103].
Paper [104] suggests localization of etching within the interface between the
electrolyte and the current-conducting liquid.
For some applications, it is quite fruitful to apply a method based on the phe-
nomenon of electroplasticity [92] (Fig. 2.38). The wire workpiece fixed by holder
(2) is immersed into a three-layer bath consisting of fluid contact (6),
non-conducting layer (8) and electrolyte layer (7). Application of voltage between
Fig. 2.38 Making the tips in

three-layer tub: 1 arch; 2 wire
blank; 3 place of etching of
the wire and creation of the
tip; 4 the end of the wire that
falls off; 5 tub; 6 liquid
contact; 7 layer of electrolyte;
8 layer of insulating liquid;
9 weight
the electrolyte and the fluid contact causes etching of the workpiece in the area of
Formation of tip (3). At the same time, voltage is additionally applied between the
fluid contact and the sample holder. Etching of area (3) causes development of a
constriction. The density of the current passing through it increases with decrease in
its diameter and reaches the value at which the electroplasticity effect manifests
itself, usually about several thousand amperes per square mm. Due to dramatic
increase in plasticity of the tip-formation area, falling off of the lower part of the
workpiece causes drawing of the tip with a small-curvature radius. The latter can be
regulated both by the value of voltage applied between the workpiece holder and
the fluid contact, and the length/weight of the part of the wire workpiece immersed
in the electrolyte.
The disadvantages of those methods include overly large droplet or overly thick
electrolyte layer, which results in increase in length of the etching part of the wire
and rather rapid exhaustion of the electrolyte. Localization of the etching area can
be improved by application of the device [105] shown in Fig. 2.39. The electrolyte
is poured between two plates (1) with coaxial holes. The lower end of the work-
piece of field emission cathode (3) inserted from above into the holes of plates (1)
touches current-conducting liquid (4), where the second electrode (6) is held.
Voltage for electrolytic etching comes to electrodes (1) and (6) from source (7).
The cone angle and the length of the cone part of the field emission cathode is
regulated by change of thickness of the current-conducting film of the electrolyte.
Concave area (2) of a plate serves for increase in the total quantity of the electrolyte
between the electrodes, which makes it possible to avoid exhaustion of the elec-
trolyte in the course of etching without increase in thickness of the etching film.
The design of the double bath for automatic etching of tips is shown in Fig. 2.40
[106]. This design is a considerable improvement of two-layer and three-layer baths
[99]. Polarity of voltage between electrodes (1) and (2) shown in Fig. 2.40 induces
positive potential in workpiece (3) in respect of electrode (1). The reason is that
workpiece (3) causes short circuit of the electrolyte area in hole (6) as the workpiece
has a smaller electric resistance in comparison with the electrolyte in a small-section
channel. A certain current between electrodes (1) and (2) makes voltage on
Fig. 2.39 Schematic of the

tip etching device. 1 Two
parallel plates with electrolyte
between them; 2 concave area
on one of the plates; 3 tip
blank; 4 conductive neutral
liquid; 5 electrolyte tub;
6 electrode; 7 power source
Fig. 2.40 Double tub for

automatic etching: 1, 2 ring
electrodes; 3 blank; 4 ring
dielectric screen; 5 pipe for
supply of the electrolyte;
6 hole; 7 intermediary
dielectric vessel; 8 electrolyte
drainage hole; 9 dielectric
body of the tub; 10 electrolyte
workpiece (3) sufficient for over-etching of the wire material in the vicinity of point
A. At that, the lower area of wire AB falls off, causing the upper part of the
workpiece to lose positive potential, and workpiece etching ceases, which provides
safety of the upper tip. Screen (4) and electrolyte pumping are necessary to destruct
the bubbles that intensively stand out in the vicinity of point B and electrodes (1)
and (2). The apex-cone height is mostly determined by the electrolyte-mark height
in vessel (7), electrolyte flow-rate in hole (6) and current between electrodes (1) and
(2). The apex shape is also determined by the dimensions of hole (6). E.g. if the
sectional area of this hole is about 3 mm2, the tip-cone length reaches 2–3 mm.
Some typical materials and conditions of their electrochemical or chemical
polishing are presented in Table 2.1 [67, 107].
Table 2.1 Polishing conditions for some materials

Material Electrolyte Conditions
«Universal» based (1) 25% perchloric acid (70%) in ice-cold Direct current 10–25 V
on acetic acid
Iron (2) 2% perchloric acid in 2-butoxyethanol
Nickel
Manganese
Aluminum
Zirconium
Uranium
Aluminum 2–10% perchloric acid in methanol Alternating current 5–10 V
at −10 °C
Aluminum–lithium 30% HNO3 in methanol Direct current 5–7 V at −30 °C
Aluminum–copper 80% HNO3 in water Alternating current 3 V at 0 °C
Aluminum–zink– 25% HNO3 in methanol <0 °C
magnesium
Beryllium HCl(4) + H2SO4(4) + HNO3(20) + ethyl Direct current 4–8 V
glycol(200)
Concentrated KOH
Boron Concentrated H3PO4, H3 PO4 , KOH or Submerging in hot solution
NaNO3
Carbon/graphite Heating in oxidizing flame or 1.25 n. Direct current 30–50 V
NaOH in water Alternating current 10–20 V, melt
Cobalt-based (1) 25% chromic acid in water Direct current 4.5 V
(2) 10% HCI in water Alternating current 6 V
Copper (1) 70% phosphoric acid in water Alternating current 3.2 V;
stainless steel counter electrode
(2) 100% phosphoric acid Alternating current 1–5 V or
Concentrated H3PO4 Direct current 16 V, copper
counter electrode
Copper–titanium 100 g Na2CrO44H2O H2O in 900 ml of Direct current 1–5 V
acetic acid
Copper–beryllium 10% H3PO4 in water Direct current 8–15 V
Copper–cobalt
Copper–iron
Copper–nickel–iron (1) 70% H3PO4 in water Direct current 4–10 V at 0 °C
(2) 10 g Na2CrO44H2O in 100 ml
of acetic acid
Gallium arsenide 44% H2SO4, 28% H2O2 (30% weight for Direct current 4–7 V at 0 °C
a unit of volume of solution) in water Direct current 8–11 V at 0 °C
Gallium–indium 0.5% H2O2, 0.5% H2SO4 in water Chemical at 60 °C
Gallium phosphide 25% HNO3 in HCI Chemical
30–50% NF in HNO3
Germanium Concentrations may vary Chemical at 60 °C
20 wt% KCN in water Alternating current 50–100 V
Gold 50% HCl + 50% HNO3 Direct current 4–5 V
20% water solution of KCN Alternating current 10 V
30% chromic acid in water Alternating current 10 ! 3 V
(continued)
Table 2.1 (continued)

Iridium Melt of NaNO3 (3) + KOH (7) or NaCl Alternating current 2–10 V
20% water solution of KCN
Concentrated water solution of Alternating current 2–3 V
KCl(100) + H2O(73) + HCl (6)
Melt of NaNO2 Alternating current 2–15 V
HCl(1) + HNO8(1) + H3PO4(2) Fresh solution between −20 and
30% H3PO4, 30 °C
20% of glycerin in water
HNO3 (1) + HCl(1) + H2O(2) Alternating current 12 ! 3 V
1–10% diluted HCl Alternating current 0.5–1 V
30% chrome acid Alternating current 5–15 V
Lanthanum NH4OH water solution (8) + concentrated Alternating current
hexaboride KOH(2) 4V
Concentrated KOH
Iron and steel 20% KCN Direct current 2.5 V
Alternating current
0.5–3 V
Molybdenum 5 n. NaOH in water Alternating current 3–6 V
or 12% H2SO4 in methanol Direct current 5–9 V
(1) 10% perchloric acid, 20% glycerin Alternating current 5 ! 1 V
in ethanol
(2) 2% perchloric acid in 2-butox-ethanol Alternating current 6 V
40% HCl Direct current 6 V
10% HNO3 or HCl Direct current 22 V
Nickel 10% HF in HNO3 Direct current 25 V
Melt of NaNO2
HNO3(9) + HF(1) or Alternating current 3 ! 1 V
H2SO4(17.5) + HF(17.5) + H2O(65) Alternating current 4–5 V
Nichrome alloys 10% HNO3 or HCl Direct current
1–3 V
Niobium HF(1) + HNO3(3) Alternating current 6 V
10% HF in HNO3 Electrolyte cooled in a stainless
HCl(3) + HNO3(7) steel vessel, direct current 0–3 V
Melt of NaNO3(4) + NaCl(1) Direct current 10 ! 3 V
20% KCN
Niobium nitrite HF(1) + HNO3 Direct current 10 ! 3 V
Nb3Sn 10% HF in HNO3 Chemical
Palladium HCl(3) + HNO3 (7) Alternating current 3 V
10% HF in HNO3
Platinum 80% sodium nitrate Fresh solution −20 to −30 °C
20% sodium chloride Alternating current 12 ! 3 V
50% H3PO4 in H2O2 (30% weight per Alternating current 1–5 V
volume unite)
25 KCl or KOH in water Direct current decreasing from
5 to 3 V; melt
(continued)
Table 2.1 (continued)

Platinum–rhodium Same as for platinum or water solution of Alternating decreasing from
Pt3Co HCl or KOH (25% saturated) 9 to 3 V
Rhenium 30–50% HF in HNO3 Direct current 8–9 V
12% H2SO4 Direct current 10 V
Concentrated HNO3 Alternating current the decreasing
48% HF(45) + concentrated HNO3(60) + of 30 to 1.5 V
CH3COOH(20) + Br2(3)
Ruthenium 21 g HNO3 + HF(4 g) (concentrations Alternating current 1.5–30 V
may vary)
45% HF, 22% H2SO4, 22% H3PO4 in Chemical
acetic acid
Silicon 48% HF(2) + CH3COOH(0.5) + Submerging in solution
concentrated H3PO4(1) + 36NH2SO4(1) Alternating current 50–100 V
Tantalum Melt NaNO3 or KOH Direct current 5–15 V
45% HF (48%), 22% H2SO4
22% H3PO4 acetic acid Platinum cathode
Tin 40% HF Direct current 50–60 V at −50 °C
Titan 40% HF Electrolyte hashing
Electropolishing at −30 °C
Titanium carbide 5% H2SO4 in methanol Alternating current 1.6 V or
Direct current 2.5 V
Titanium nitrate 40% HF Alternating current 1–5 V
Tungsten 5% NaOH in water Alternating current 1–5 V
Tungsten carbide 8% H2SO4 in methanol Direct current 12 V at −16 °C
Vanadium 15% H2SO4 in methanol Direct current 6.5 V
YBa2CuO7 and 10% Perchloric acid in 2-butoxyethanol Direct current 10–20 V at <0 °C
related oxides
Uranium 12% Chromic anhydride Direct current 12 V
88% Phosphoric acid
Note Items 1 and 2 refer to the first and the second stages of a two-phase process
Warning A lot of chemical mixtures listed above are toxic, corrodible, combustible or explosive. Their
preparation and application requires appropriate safety measures to be implemented
Sometimes two stages of polishing can be combined into one operation. The first
stage, when an electrolyte layer usually as thin as 6–8 mm floats on the surface of a
denser inert liquid e.g. CCl4 or glycerine, is intended for prompt removal of the
material and development of a constriction in the workpiece center. The inert liquid
should not mix or react with the electrolyte.
At the second stage, polishing proceeds until the weight of the lower part of the
workpiece exceeds the tenacity of the thin constriction. The starting material can be
shaped as a wire of a square or round section. The opposite electrode is usually
made of a precious metal e.g. gold or platinum.
This method is especially good for thick wires or the workpieces cut out of
massive samples e.g. aeroturbine blades or semiconductor single crystals.
Fabrication of apices from thick wires and workpieces can also involve a device
for etching of apex samples in an electrolyte flow (Fig. 2.41 [108]). Wire workpiece
(1) fixed in collet lock (3) is over-etched inside ring electrode (2). The electrolyte
flow-rate appreciably affects the apex-surface quality, the shape and the
apex-curvature radius. E.g. tungsten is electropolished in an alkaline solution in the
alternating-current mode at the optimum electrolyte flow-rate of 0.2 m/s. The
electrolyte is pumped over a closed circuit.
The current used in electropolishing should be controllable to prevent heating or
even melting of the thin area of the sample constriction, which is quite possible if
the constriction is extended enough. When alternating current is used, control of
such parameters as voltage, wave form, phase angle, frequency and cycle number is
useful for fabrication of high-quality samples [109, 110]. It is also possible to
control current and interrupt polishing automatically as soon as the lower half [78]
gets detached.
Fig. 2.41 Device for creating

tips in a stream of electrolyte.
1 blank; 2 ring electrode; 3
collet; 4 electrolyte
The research carried out to determine the ultimate strength of very thin tungsten
wires established dependence between the final cross-section measured with a
transmission electron microscope and the wire-end mass that fell off as a result of
electropolishing. Thus, the method cannot be used to fabricate an unimpaired
sample. It deforms the sample and the field-ion image shows all signs of plastic
deformations and cold working. That takes place even when the end of the wire
immersed in a non-conductive liquid is reduced in diameter and made very short as
a result of preliminary etching to reduce the weight of the lower area of the wire.
The lifting force of gas bubbles flowing to the lower end of the wire just before the
separation can deform the wire by bending the etched part of the sample.
To obtain an undeformed apex with a small cone angle, it is necessary to reduce
etching in the vicinity of the tip apex by elimination of the steep current gradient
through its gradual reduction due to partial screening of the sample apex. After
preliminary electric etching, the sample apex is brought up to the necessary
dimensions by immersion of the wire tip in a protective tube of 1–2 mm in diameter
(Fig. 2.42). For normal circulation and uniform distribution of the dissolved sub-
stance, the tube should be open from both ends and made of a coloured glass for
optimum visual observation. Any necessary cone angle can be obtained without
separation of the lower part of the wire by immersion of the sample at different
depths relative to the upper edge of the tube.
The tip polish quality is usually examined with an optical microscope. For this
purpose, a 500-fold magnification with the aperture of 0.6 is quite enough. The
tip-curvature radius ranges from 100 to 500 Å, which is much higher than the
maximum resolution of an optical microscope. However, achievement of a uniform
change of the cone angle in the immediate proximity of the tip apex is a quite
sufficient criterion of its suitability as emitter samples.
In some cases, it is not expedient to use chemical etchants, especially for more
porous, e.g. carbon-based, materials.
Fig. 2.42 Electrolytic

method of tip manufacture
with partial screening: 1
microscope; 2 sample; 3 glass
pipe; 4 electrode; 5 polishing
solution
It is connected with difficulties of removal of etchants from these materials. That

is why it is sometimes expedient to use the apices fabricated by ionic bombardment.
The method has a number of advantages:
(a) practically all materials can be etched;
(b) the method is not notable for etching selectivity;
(c) etching makes the apex surface clean;
(d) good reproducibility.
However, the treatment causes radiation-induced defects in the sample material,
which are partially removed by annealing.
Below we suggest several versions of such methods.
The diagram of the device fabricating tips by ionic bombardment in a glow
discharge [111] is shown in Fig. 2.43a. A workpiece (1) is sprayed by inert-gas ions
at their partial pressure of 10−3−10−5 Pa. Cone-shaped insulating screens (3) pro-
vide non-uniform spraying along the workpiece, thus forming a preset cone angle.
When spraying, the central part of the sprayed area of the wire grows considerably
thinner and eventually gets detached under the weight of the lower part of the
workpiece. That is the part to be used as a field emission cathode.
A modification of this method is an installation for tapering of carbon fibers in a
discharge of electricity [112]. A field emission cathode workpiece (1) (Fig. 2.43b)
is fixed with Aquadag in a holder (2), which is rigidly connected with a rod (3)
moving due to a micrometer screw (4). Voltage of about 10 kV is applied between
the earthed workpiece (1) and the anode (5). The turn-on voltage generates a
discharge, and the fiber (1) is tapered as affected by ionic bombardment. The
process is controlled by an observer (7) using a microscope (6). The tip shape
depends on the fiber diameter, discharge voltage and electrode-separation distance.
Fig. 2.43 Schematic of tip manufacturing in glow discharge (a): 1 blank; 2 conductive plate;
3 insulating screens. Schematic of sharpening of the graphite fibers (b): 1 field cathode blank
(graphite fiber); 2 holder; 3 rod; 4 micrometric screw; 5 anode; 6 microscope; 7 observer
Variation of voltage or the distance with the micrometer screw (4) can change the
apex shape. For the fiber of 7 lm in diameter, the discharge voltage is about 5 kV
at the electrode-separation distance of 5 mm. The tapering can take place with
exposure to air or in technical vacuum, and voltage can be pulsing if necessary.
Accurate processing, i.e. tapering of new and sharpening of old apices, involves
the systems of focused ion beams of 1–2 mm in diameter [112, 113]. The angle
between the sample axis and the ion beam is usually about 90°.
If an tip workpiece is mounted in the holder of a transmission electron micro-
scope, the end of the apex-fabrication process can be fixed with high accuracy. At
that, fabrication of a uniform cone results from rotation of a sample workpiece
about the axis in the ion flux until the tip radius is reduced to a suitable value.
It has been established that ionic bombardment lasts approximately as long as
usual electropolishing. (Though the name of the paragraph places a priority on ionic
bombardment, any particles e.g. electrons can be used [114].)
The time necessary for ion etching can be reduced by preliminary chemical
etching of the sample cf. [115].
At first a workpiece with the curvature radius of 260 Å fabricated by the elec-
trolytic method is placed in a vacuum chamber. After pumping-out down to the
pressure of 10−5 Pa, the chamber is filled with neon up to the pressure of 10−3 Pa.
Voltage of 800 V is applied between the cathode and the anode. At that, the
electron-current density is 106 A/cm2. At this potential difference, the cathode is
kept for 5 s. Over this period, the field emission current increases due to ion etching
and then stabilizes at the density of 107 A/cm2. After that, the processing voltage is
10% reduced and the cycle is repeated until the running voltage reaches 300 V. At
that, the apex-curvature radius is 100 Å. Reproducibility of emitter parameters is
very high, only about 2% are rejected.
Fabrication of apices from epitaxial layers in the semiconductors of type III–V is
suggested in two techniques [67, 116].
The samples containing epitaxial layers can be made by preliminary cutting of
square-section workpieces from a multicomponent assembly consisting of a sub-
strate material plus a layer. These workpieces are mechanically polished from two
or three sides up to formation of a cone with a apex in the layer. Then the remaining
substrate is chemically removed by selective etching. The resulting conical work-
piece is chemically polished up to the tip shape. In this case, the sample axis is
parallel to the initial surface.
The samples whose needle axis is perpendicular to the surface can be prepared
by surfacing of a ball of about 1 lm in diameter made from polystyrene or quartz in
the ethanol suspension containing a droplet of a surface-active substance followed
by ionic bombardment as shown in Fig. 2.44. A sample with a suitable cone angle
is fabricated with a ball that is more resistant to bombardment than the surface
layer. The material lying under the ball is protected from the bombardment. In the
course of bombardment, the ball diameter is decreased, which gradually opens the
entire layer for ionic treatment. As a result, the sample is tapered.
Most metals can develop thin microelevations on the surface of blunt tips due to
the surface diffusion of tip material atoms. That is usually done by the electric field
Fig. 2.44 Manufacture of field ion microscopy sample out of epitaxial layer
of the field-ion sign with voltage of about 2 V/Å combined with the elevated
temperature of >1000 K. Such microelevations have a size of 40–80 Å [117].
A lot of research projects need formation of a surface layer on the apex.
The most common method is thermal evaporation [118–126], though there is
also activated thermal evaporation [127, 128], chemical reactions [129, 130] and
radio-frequency sputtering [131–134].
There are also some other methods with immersion into a liquid, including
deposition [135], anode treatment or passivation [136, 137] and immersion into a
molten metal [138]. Thickness of the films layered should be controlled because the
sample-curvature radius increases over the homogeneous layer, which brings the
research to increase in running voltage.
Moreover, application of an electric field can strip thick films.
Another type of the surface layer can result from application of biological
molecules [139–142]. These experiments need special techniques to provide opti-
mum application of molecules on an tip.
Tip quality control in a light microscope after preparation and cleaning is a
routine procedure. Though the imaged part of the sample is not resolved at small
magnifications 200–500, a light microscope is helpful in examination of the end
radius, surface processing and selective etching.
A smooth conical rod and presence of diffraction fringes on the needle tip
indicate satisfactory quality of the sample. It is not necessary to use large magni-
fications needing thin immersion oils because a slight increase in magnification is
not enough to obtain more information, whereas the liquid surface tension can
destroy the thin needle. The details of surface processing and selective polishing are
better determined at larger magnification in a scanning electron microscope.
A small apex-tip curvature radius can be examined in a transmission electron
microscope.
Most electron microscopes equipped with a side-entrance stage can be accom-

modated for examination of field-ion samples. There were some reports of gonio-
metric microscope stages (cf. [143–148]). A field-ion sample tip apex is penetrable
for electrons in an electron microscope working at 100 kV. Electron microscopes
with larger running voltages can examine thicker regions of a body.
The main way of securing the resulting sample is immediate or bonded welding
of an apex to an arch wire of a high-melting metal e.g. tungsten, molybdenum.
This way assumes both warm-up of an tip to a high temperature, e.g. degassing
of its surface, and cooling to the temperatures down to those of liquid helium.
Unless spot welding is possible, tips can be secured by soldering or bonding
with a special high-temperature agent, e.g. [149]. The main components of the
agent are the aluminophosphate binder (A12Oa: Pa2O – 1 3 – 20 30 wt%), W
powder with the grain size of 0.5–1 lm (65–70 wt%) and a carbon powder
(5–10 wt%). Warming up and sintering take place at t = 400 °C up to solidification
of the composite (e.g. by passing of current through the arch).
For special research, the arch wire can be replaced with a pin or another holder
depending on the research objective and the microscope design.
It is worth noting that breakage or blunting of the tip while carrying or storage
can be avoided if its vertex is shaped after the apex is secured on the arch or the
holder. The ready samples should be kept in locked cassettes excluding their
breakage or entanglement. A particular design of the cassette is up to the researcher,
who can proceed from his own abilities and needs.
This section features the main ways of fabrication of tip samples from different
materials. The processes aimed at formation of tip are rather complicated, so it is
expedient to select the optimum mode and the processing technique proceeding
from the material used. As stated above, it is sometimes expedient to combine
apex-fabrication techniques e.g. the electrolytic method and the ionic bombardment
etc.
2.6 Some Technological Cases of Application of a Field

Emission Microscope
Below you can find some practical examples of application of field emission
microscopy.
First of all, it is field emission microscopy of different materials and composites,
which has been developing since 1936. Its fundamentals were given in Sect. 2.1.
The maximum brightness of a field emission image is produced by the regions
with a lower value of the electron work function. So crystallographic directions of
single crystals in emission patterns can be decoded on the basis of stereographic
projections [79, 150]. Application of field emission microscopy is a great advance
in comparison with widespread macroscopic diffusion-research methods due to
2.6 Some Technological Cases of Application … 101
visualisation of diffusion processes and huge magnification able to approach the

atomic level [151].
It is well known that the phenomenological theory [152] gives the following
expressions for the surface-diffusion coefficient, which can be used for evaluation
of diffusion parameters and in experiments with a field emission microscope:
x2 =t Q=kT
Ds ¼ ¼ D0e ð2:13Þ
Ds
where x2 is the root mean displacement; t being the time; Q being the activation
energy; T being the absolute temperature; k being the Boltzmann constant; D0 being
the diffusion constant.
According to [153],
D0e ¼ a2 m;
where a is the «jump distance» of the migrating atom, which is approximately equal
to the crystal-lattice constant; m is the frequency factor (*1013 s−1).
The motion of the adsorbed atom in diffusion can be regarded as transition
between the adsorption centers with different binding energy. The centers are
separated with potential barriers, whose height, depending on the geometrical
structure of the surface, can be determined by dependence of the diffusion coeffi-
cient on temperature.
The idea of a potential surface texture is largely substantiated in the «pair-wise
interaction» theory [154], which represents the metal surface as consisting of
densely packed balls featuring neutral atoms pair-wisely interacting with energy
1
u ;
r6
where r is the distance between the ball centers. The energy binding the adsorbed
atom to the surface is calculated by summation of interactions with the nearest and
the farther neighbours, its value being determined by the number of neighbours.
The energy of adsorption on any face can be calculated from the known facet
structure and the atom size.
The energy of migration on different faces is also calculated by change of the
number of the nearest and other neighbours when transiting from the basic state to
the «on-the-seat» state between «wells».
Certainly, the energy of migration differs, depending on the face structure and
the physical properties of adsorbate and adsorbent atoms. Experimental determi-
nation of migration energies for the widest possible set of crystal faces, adsorbates
and adsorbents (if possible) is an important objective in studying of a solid surface.
Solution of this problem will give information on energy «texture» of the surface,
substantiate the pair-wise interaction theory and enable evaluation of the binding
energies of adsorbed atoms in different areas of the surface.
These problems became solvable only with the advent of a field emission
microscope, enabling a researcher to observe movement of an adsorbate on a clean
surface of an adsorbent under the conditions of ultra-high vacuum with the enor-
mous magnification of M 105 times and the resolution of −20 Å. Such resolution
does not differentiate between individual atoms, but allows accurate observation of
the diffusion front, measurement of the rate of its movement and determination of
the diffusion coefficient of Ds and the activation energy of Qm with (2.13).
Visual observation of the diffusion edge is based on the emission contrast
between the surface coated with adsorbed atoms and the clean surface of the
adsorbent, which is caused by their different emissive capacity. Adsorption changes
the transmission factor of surface barrier D (Ex, F) through change of its height (e.g.
change of the work function) or width (e.g. enhancement of the local field), which
significantly affects the field emission current.
If the diffusion front is clear cut, it is not difficult to measure diffusion parameters
to a high accuracy. Otherwise, measured is the time of transition between the two
surface states fixed visually or by the field emission current.
To study diffusion processes with a field emission microscope, the latter is
charged with a source evaporating the adsorbate under investigation to the side of
the apex. The evaporator is usually a tungsten or molybdenum spiral charged with a
substance or a wire holding an adsorbate as a weight. The designs of more
sophisticated sources are also detailed in original articles. The mutual positioning of
the evaporator and the apex should allow the evaporated adsorbate to get only to a
part of the tip surface (Fig. 2.45).
Fig. 2.45 Relative positions

of the source and the tip
during the condensate
evaporation process on the tip
When the tip is heated, the edge between the surface coated with the adsorbate
and the clean part of the surface begins moving. The field emission microscope
flask surface is usually marked to measure the time for the edge to move between
the two marks. As a result, path x in (2.13) remains constant and progression along
this path at different temperatures is timed. Migration energy Qm can be determined
by the Arrhenius equation

1 Qm
t¼ ¼ A exp ; ð2:14Þ
t kT
where t is the diffusion rate, and Qm can be determined by the slope of the line

1 1
lg ¼ f :
t T
To determine diffusion coefficient Ds , diffusion constant D0 and frequency factor

v, it is necessary to know magnification M and hence apex radius r, which can be
found by the approximate formula [155]
U 1=2
r ¼ 0:15 ; ð2:15Þ
u2
where U is the voltage necessary to generate field emission current i = 10−5 A; u is

the work function assumed as 4.5 eV for tungsten.
Exponential dependence of the activation processes on temperature makes
accurate measurement of the latter very important, but rather difficult. Different
temperature intervals are timed with different techniques: pyrometry of the point
where the tip is fixed to the arch with a rated allowance for decrease in temperature
from the bottom to the tip of the apex; measurement of resistance of a small area of
the supporting arch; measurement of dependence of temperature on filament current
etc.
It is very important for the temperature-increase front to be steep enough, which
can be provided with special diagrams.
It is essential for research of diffusion processes to know the value of field F at
which Ds and Qm are measured because both values can depend on F.
Application of a field emission microscope to research of gas-tungsten surface
diffusion on allowed establishment of general patterns of migration processes at the
atomic level and bringing of the diffusion mechanism into focus.
It was found out that gas adsorption at a low temperature heavily reduces
emission from the source, developing a violent emission contrast between the
coated and uncoated parts of the surface. The decrease in emission means that the
layers of hydrogen, oxygen or nitrogen prevent electrons from tunneling from the
metal. Heating of the apex can display migration in a form of a dark layer.
2.6.1 First-Type Migration
If the original coating is significantly thicker than the monolayer, migration takes
place with a clear-cut edge progressing uniformly over the tip surface. The edge
begins moving at T = 20 K for hydrogen [156], T = 27–77 K for oxygen [157],
T = 40 K for nitrogen [158]. Movement of the edge roughly agrees with law (2.13).
If the evaporated portion is insufficient, the edge movement ceases and a part of the
apex surface remains uncoated. Further diffusion becomes possible only at a con-
siderable rise in temperature. Gommer [156] suggests the following mechanism
explaining the first-type migration.
A multi-layer coating resulting from condensation of a sufficient gas portion on a
lateral surface of a cooled apex consists of a chemosorbed first layer and subsequent
poorly bound layers. The chemosorbed layer is motionless at the above-specified
migration temperatures. The upper-layer particles adsorbed physically are the more
poorly bound the further they are from the surface. They can move over the
underlying chemosorbed layer. When they come to the end, they are captured at the
edge and chemosorbed as well, spreading the layer edge and creating an illusion of
its movement. The edge stops moving when the entire material of the second layer
is exhausted (Fig. 2.46).
This mechanism provides filling of all adsorption centers on the surface. The
chemosorbed layer formed in this way is monatomic though the layer of the
roughest surface areas may be significantly thicker than one atom. This type of
migration is often compared with an unfolding carpet or a drawing curtain.
2.6.2 Second-Type Migration
If the original portion is about twice as little as that necessary for full coverage of
the apex in the first-type migration, the edge stops and its movement resumes at
higher temperatures: T = 180–240 K for hydrogen [156], 500–530 K for oxygen
[157], T = 400–650 K for nitrogen [158].
Though in the second-type migration, diffusion also proceeds with a clear-cut
edge, this edge is very anisotropic. All the above-listed adsorbates most promptly
migrate over densely packed faces {011} and channels connecting faces {011} and
{211}. These channels mostly consist of face areas {011}. These adsorbates
migrate more slowly over areas {111} and most slowly over the vicinals around the
faces {001}. Activation energies are different for different areas and average
Fig. 2.46 Multi-layer

diffusion mechanism
5.9 kcal/mole for hydrogen and 25.5 kcal/mole for oxygen and nitrogen.
Undoubtedly, there is migration of the chemosorbed layer atoms in this case. The
clear-cut edge indicates the diffusive capture mechanism theoretically substantiated
in [159].
The clear-cut nature of the edge is caused by the fact that the diffusing atoms are
captured by the adsorption centers, «traps», spaced shorter than the resolution of a
field emission microscope, i.e. less than 20 Å. If there is enough adsorbate, the
diffusing atoms captured by the traps facilitate the diffusion of the subsequent
particles.
2.6.3 Third-Type Migration
If the concentration of adsorbed atoms is very low, the traps are not filled because
the diffusion front can only progress if the adsorbed atoms have enough energies to
escape from the traps. The higher the temperature, the deeper traps become vacant.
The activation energy is determined by the energy of binding of the adsorbed atoms
at the adsorption centers with a high binding energy. The diffusion goes on without
a clear-cut edge because the distance between the similar adsorption centers
exceeds the microscope resolution. The activation energies measured grow with
reduction of coverage h. In particular, hydrogen is reported [160] to have the
following activation energies for diffusion without an edge: 9.5 kcal/mole at
h = 0.1 − 0.3; 16 kcal/mole at h = 0.01 and 64 kcal/mole at h 0.
Papers [156–158], as well as [159, 161–163], dedicated to different gas-metal
systems, give substantial reasons to the point of view that increase in the binding
energy, i.e. desorption heat, with reduction in the concentration is connected not so
much with interaction of the adsorbed particles with each other, as with presence of
various adsorption centers not only on the surface having a number of crystal faces,
but also on a separate face of single crystals, except for the most densely packed
ones [164].
Presence of gas affects the rate and the activation energy of the surface diffusion.
In particular, Müller [162] found that oxygen and water vapors increase the rate of
the tungsten-lattice surface diffusion. Brenner [160] discovered the same effect
when iridium diffused over its own lattice.
At the same time, a number of authors note an increase in the activation energy
for the surface diffusion of adsorbed atoms with gas present. In particular, [165]
shows that oxygen and nitrogen increase the activation energy of copper-tungsten
diffusion from 17 to 25 kcal/mole. It is assumed that gases form compounds both
with the substrate and the atoms of the adsorbate. Depending on the energy binding
the atoms of the adsorbent and the adsorbate with the appropriate gas, it is possible
for the compound to migrate.
It should be pointed out that there is a budding tendency for gases to accelerate
diffusion of metals over their own lattice and decelerate diffusion of adsorbate
atoms over the tungsten surface. As of today, the papers investigating the effect of
gases on the surface diffusion with a field emission microscope are too few to
warrant a particular judgment on this issue.
Some authors [166–169] found an effect of the value and the sign of the external
electric field on the equilibrium concentration of adsorbed atoms at the apex vertex
and on the activation energy of the surface diffusion.
The importance of field emission has increased with enhancement of the applied
component, i.e. development of effective field emission cathodes, especially those
made from carbon materials. In particular, [170] described the experiments with
freshly prepared samples of polyacrylonitrile carbon fibers and stated [170, 171]
that their volt-ampere characteristics plotted on the Fowler−Nordheim coordinates
have breaks/deflections towards smaller currents. A typical characteristic is given in
Fig. 2.47. It also shows the snapshots of field emission images of a carbon fiber
viewed on the screen.
Analysis of the images suggests that the effective surface of the emitter is
obviously irregular with chaotic heaping of emitting microelevations [172].
The breaks of the characteristics can be explained by the fact that manual
measurement of cathode characteristics during voltage boost causes reconstruction
of the surface at heavy currents and, accordingly, voltage, which is well illustrated
by the photos in Fig. 2.47. Such reconstruction shifts the volt-ampere characteris-
tics towards heavier voltages.
Reconstruction of the carbon-fiber surface is accompanied by its destruction, i.e.
removal of poorly bound crystallites, affected by ionic bombardment and pon-
deromotive loads.
In those cases, the characteristics were taken at voltage boost and were not
reproducible.
Fig. 2.47 Emission

characteristic of
polyacrylonitrile carbon fiber
with freshly cleaved side and
field emission images of
different parts of characteristic
As a rule, some operation and formation make the characteristics forward. That
assumes such a range of currents and voltages that do not destroy the fiber. Voltage
boosting does not change the fiber effective-surface geometry, but increases the
emission current due to increase in the field intensity at the working
microelevations.
The emitting-surface area can be evaluated either by field emission images
viewed in a field emission microscope [173, 174] or by volt-ampere characteristics
as suggested by Oostrom [175]. If the average area of the single-fiber end face is
assumed as 4 10−7 cm2, the field emission images in Fig. 2.47 suggest that
growth of the drawn current from 2 10−8 to 5.22 10−5 A causes the
emitting-surface area also to increase from 3.2 10−11 to 1.2 10−9 cm2, i.e.
about 250 times. If each fibril is further assumed to have a diameter of 50 Å and its
end face emits uniformly, the number of the working fibrils increases in this case
from 160 to 4000 pieces, while the density of the current from each fibril increases
from 6 102 to 6.5 104 A/cm2. However, such evaluations of the emitting
surface can be exaggerated due to the electron-optical effect. Evaluation of the
emitting-surface area by its emission characteristics suggests that the
emitting-surface area remains nearly the same and equals 3.5 10−12 cm2
throughout the range of the current drawn. As each fibril end-face area is about
2 10−13 cm2, it is easy to estimate that a stable state of the field emission cathode
involves functioning of up to 15 fibrils, each providing the maximum current
density of 107 A/cm2.
A vast number of standardized fibrils, 106 pieces on the end face of a single
fiber, provide high stability of the field emission cathode emission. Ionic bom-
bardment destroys some fibrils and replaces them with identical new ones, and
stability of such a huge statistical fibrillar system is preserved until shrinkage of the
end face results in decrease in electric intensity and drop of field emission.
In the course of operation, the field emission cathode is exposed to intensive
ionic bombardment, and the ion energy is several orders of magnitude higher than
the thermal energy of atoms (3/2 kT). That is why the bombardment breaks equi-
librium in the adsorption process towards increase in desorption. When the cathode
functions, it prevents the residual-gas molecules from adsorption near the
microelevations on the effective surface of the polyacrylonitrile fiber, where
emission mostly comes from. In the adsorption mode, the residual-gas molecules
are also first adsorbed not on microelevations but in checks and pores of the fiber,
and only after they are filled, the molecules can settle down near the microeleva-
tions. That is why residual-gas adsorption scarcely affects the electron work
function from PAN fibers.
The current step-up taking place after the adsorption mode can be explained by
different mechanisms:
– change of the work function or barrier penetrability;
– change of the surface geometry;
– change of the emitting-surface area.
A cyclically repeated change in the cathode geometry is hardly possible as

analysis of emission images shows that the surface geometry during the experiment
is invariable. Therefore, the first and the third mechanisms are most probable. At
that, it is worth noting that the above-described model of the absorption-desorption
processes proves that these two mechanisms cannot be considered separately from
each other because the adsorption phenomena change the barrier penetrability not
so much at microelevations as in pores, cavities or capillaries, i.e. in those surface
areas that did not emit before that. Therefore, reduction of the potential barrier in
these areas is accompanied by increase in the emitting area.
As a rule, transient characteristics look like descending curves levelling off after
cessation of transient processes for polyacrylonitrile fibers of different processing
temperature [176]. That is why Fig. 2.48 exemplifies one of such characteristics.
Figure 2.48 demonstrates that sufficiently massive reduction of current takes
place in the first minutes and even seconds after the cathode starts working. The
reason seems to lie in the fact that intensive ionic bombardment first tears off the
adsorbed atoms lying near microelevations and therefore having low adsorption
energy, which increases the electron work function from the areas near the
microelevations, and the current promptly decreases. The surface is further peeled
off from the atomic layer at a much slower rate; firstly, due to stronger binding of
adsorbed atoms with the fiber surface; secondly, because the adsorbed molecules
come from capillaries to the surface in exchange of those torn off by the ionic
bombardment. The rate of this process is also determined by the diffusion rate and
seems to have the same order of magnitude. These processes go on until adsorption
and desorption are dynamically balanced.
Analysis of the transient characteristics also shows that the desorption rate
poorly depends on the intensity of ionic bombardment of the cathode.
It follows from Fig. 2.48, as well as from other similar dependences investi-
gated, that the cathode-surface geometry is generally maintained throughout the
trial cycles, which supports the idea that it is not the geometry change that causes
Fig. 2.48 Time diagram of

field emission current of the
cathode made out of a
after 1-day adsorption at
pressure p * 102 mm Hg
and corresponding
photographs of field emission
images
changes in current, it’s the residual gas adsorption. The field emission images
obtained at the beginning of the work period are somehow different from those
obtained at the end of the previous work period; firstly, the image becomes brighter;
secondly, the previous image looks powdery, i.e. it is characterized by magnifi-
cation of the image details, which seems to be explained by additional emission
from other fiber-surface areas. Particularly interesting are the cases when fiber
cathodes are under the pressure of p = 10−1 mm Hg for a long time, usually more
than a month.
Here the original image at the beginning of the next work period after such a
long absorption period is utterly different from that obtained at the end of the
previous and the current period. That may be because the adsorbed molecules are
oriented so that the barrier becomes more penetrable. We believed that residual-gas
adsorption decreases the work function. Actually, it can both decrease and increase.
E.g. a polar molecule of water can be adsorbed by both the positive and the negative
end of a dipole to the outside; in the first case, the work function decreases, in the
second case, it increases.
References
1. E.W. Mueller, Electronenmicroskopishe Beobachtungen von Feldkathoden. Zs. f. Phys. 106,

541–550 (1937)
2. M. Drechsler, Erwin Miller and the early development of field emission microscopy. Surf.
Sci. 70, 18 (1978)
3. R. Gomer, Field Emission and Field Ionization (Harvard University Press, 1961)
4. R. Comer, Field emission and field ionization and field desorption. Surf. Sci. 299/300, 129–
152 (1994)
5. Y. Groshkovsky, Equipment of High Vacuum (M. Mir, 1975)
6. M. Benjamin, R.O. Jenkins, The distribution of autelectronic emission from single crystal
metal points. Proc. Rotal Soc. Ser. A. 18(981), 225–236 (1942)
7. V.I. Makuha, Research of adsorption and electron emission of films of cesium on a tungsten
monocrystal. Phys. Solid Body 9(1), 150–156 (1967)
8. B.V. Bondarenko, V.A. Kyznetsov, Research of mechanical strength of film field emission
cathode from chrome. GTF XLII, 1093–1095 (1972)
9. E.P. Sheshin, N.B. Pogibelskay, in Research of mechanical strength of film field emission
cathode from niobium. Physical Processes in Devices of Electronic Equipment, pp. 15–18
(M. MIPT, 1980)
10. D. Zimmerlan, R. Gomer, Temperature regulator for holders of field emission cathodes.
Devices Sci. Res. 7, 171–173 (1965)
11. M. Pichand, A. Mueller, M. Drecheer, Temperature distribution along metal tips (for field
emission microscopy and the study of surface phenomena). Surf. Sci. 26(1), 14–20 (1971)
12. O.P. Astahov, E.P. Sheshin, others, Direct measurement of temperature of the graphite field
emission cathode. Radio Tech. Electron Electron. 31(5), 1045–1046 (1986)
13. O.P. Astahov, V.I. Makyha, E.P. Sheshin, Temperature condition of graphite field emission
cathode with the developed working surface. Radio Tech. Electron. (USSR) XXXIV(6),
1310–1312 (1989)
14. B.H. Mosienkov, V.A. Seliverstov, E.P. Sheshin, other, System for measurement of field
emission characteristics of the field emitter. PTE 2, 204–206 (1991)
15. S.V. Denbnovetskiy, A.B. Leshinin, G.F. Semenov, Transformation of information on the
new memorable EBT (M.: Energoizdat, 1984)
16. S.V. Denbnovetskiy, B.H. Mosienkov, T.V. Temirazeva, other, Selection and information
transfer. Kiev: Sci. Thought 45, 114 (1978)
17. V.A. Kyznetsov, E.P. Sheshin, Technique and application of analyzers of a power range of
field emission electrons. PTE 5, 7–16 (1972)
18. E.W. Mueller, Z. Physik. 120, 261 (1943)
19. A.L. Suvorov, A.F. Bobkov, B.Y. Kyznetsov, S.B. Zaytsev, Field emission microscope for
research of emission properties of faces of monocrystals. PTE 4, 248–250 (1979)
20. E.W. Plummer, R.D. Yong, Field emission studies of electronic energy levels of absorbed
atoms. Phys. Rev. B1, 2088–2109 (1970)
21. L.W. Swanson, L.C. Crouser, Phys. Rev. 163, 622 (1967)
22. D.R. Bowman, R. Gomer, K. Mutelib, M. Tringides, Surf. Sci. 138, 581 (1984)
23. R. Gomer, A. Auerbach, Surf. Sci. 167, 493 (1986)
24. R.D. Young, E.W. Mueller, Experimental measurement of the total-energy distribution of
field-emitted electrons. Phys. Rev. 113, 115–120 (1959)
25. K. Kowasaki, K. Senzaki, Y. Kumashiro, A. Okada, Energy distribution of field-emitted
electron from TiC single crystal. Surf. Sci. 62, 313–316 (1977)
26. M.I. Elinson, Radio Technician Electron. 4, 140 (1959)
27. A.G. Ghdan, M.I. Elinson, Radio Technician Electron. 6, 671 (1961)
28. A.G. Ghdan, M.I. Elinson, A.B. Sandomirsky, Radio Technician Electron. 7, 670 (1963)
29. I.L. Sokolskay, G.P. Sherbakov, Solid State Phys. 3, 167 (1961)
30. G.P. Sherbakov, I.L. Sokolskay, Solid State Phys. 4, 3526 (1962)
31. I.V. Mileshkina, I.L. Sokolskay, Solid State Phys. 5, 2501 (1963)
32. G.N. Fursey, P.G. Shlahtenko, Solid State Phys. 12, 2645 (1970)
33. F.G.I. Van Oostrom, Validity of the Fowler–Nordheim model for field electron emission.
Philips Res. Rept. 21(1), 1–102 (1966)
34. V.A. Kyznetsov, E.P. Sheshin, Power range of the field emission electrons emitted from
microledges on tungsten, an edge. Radio Eng. Edektronik. 20(7), 1550–1553 (1975)
35. G.L.R. Mair, D.C. Grindrod, M.S. Mousa, Deam-energy distribution measurements of liquid
gallium field-ion sources. J. Phys. D: Appl. Phys. 16, 209–213 (1983)
36. V.A. Ivanov, T.S. Kirsanova, T.A. Tumareva, Field emission spectroscopy of the tungsten
covered with layers of barium and an oxide of barium. Phys. Solid Body 23(3), 664–668
(1981)
37. C.I. Workowski, I.J. Czyzewski, Field emission spectrometer whis an electron multiplier
operating in the phase-sensitive detection system. Acta Physia Polonica. A39(5), 523–529
(1971)
38. A.S. Kypryshkin, L. Rybakov Yu, E.P. Sheshin, The analyzer of total energies of field
emissions electrons with the retarding potential. PTE 1, 151–153 (1990)
39. A.S. Kupryashkin, L. Rybakov Yu, E.P. Shehin, Field-electron total energy analyzer with
retarding potential. Instr. Experim. Tech. 33(1), 155–157 (1990)
40. R.D. Young, C.E. Kuyatt, Resolution determination in field emission energy analysers. Rev.
Scient. Instrum. 39(10), 1477–1480
41. R.Z. Bahtizin, U.M. Ymagyzin, Analyzer of a energy spectrum of field emission electrons.
PTE 3, 212–216 (1984)
42. R.Z. Bahtizin, V.M. Lobanov, K.G. Yu, M. Umagyzin Yu, Calculation of trajectories of
electrons in electrostatistical lenses. Radio Technician Electron. 4, 1556–1558 (1988)
43. C. Kuyatt, E. Plammer, The reflecting analyzer energy for research of field emission
electrons. Devices Sci. Res. 1, 122–126 (1972)
44. N. Kempin, K. Klapper, G. Ertl, The analyzer for removal of energy distributions of field
emission electrons. Devices Sci. Invest. 9, 52–55 (1978)
45. S.I. Shkyratov, S.H. Ivanov, S.N. Shilimanov, Field emission laboratory—the field emission
electronic spectrometer combined with a field ionic/electronic microscope. PTE 4, 126–134
(1999)
References 111
46. C. Oshima, H. Fujii, T. Yamashita et al., Construction of low-temperature gun and

high-resolution field emission spectra from a Nb superconductor. Appl. Surt. Sci. 146,
143–147 (1999)
47. V.I. Maryin, E.G. Kalinychev, N. Lahov Yu, The vacuum lock unit with the AVSh-1
manipulator. PTE 3, 263–264 (1981)
48. A.A. Isaev, Installation for transfer of the holder with samples from a lock on the
manipulator of the ultrahigh-vacuum camera. PTE 6, 185–186 (1987)
49. R.I.-G. Garber, G.I. Dranova, I.M. Mihaylovsky, G.G. Chechelnickiy. A.S. USSR №
410488, of 13.04.1971. Cl N01j 37/285
50. M. Wada, M. Konishi, O. Nishikawa, Binding states of Ga and Sn on W and Mo: structures,
evaporation field and its temperature dependence. Surt. Sci. 100(2), 439–452 (1980)
51. M.P. Ardashev, A.O. Golubok, N.F. Fedorov others, Folding electronic and ion microscope.
PTE 1, 222–223 (1981)
52. V.N. Ilin, E.P. Sheshin, D.A. Shomin, Universal field electron—field ion microscope. PTE
2, 193–197 (1983)
53. A.c. 652403 USSR, Cl. F16K51/02 of 24.06.77. The ultrahigh-vacuum all-metal bakeable
lock valve. K.D. Danilov—It is published 15.03.1979, Bulletin No. 10
54. A.c. 610081 USSR, Cl. G05F1/12 of 08.07.76. Stabilizer of alternating current.
E.P. Sheshin, V.A. Kyznetsov—It is published 05.06.78, Bulletin No. 21
55. A.c. 528637 USSR, Cl. N01 j37/26. The camera of the combined field emission-field ion
microscope. V.A. Kuznetsov, E.P. Sheshin—It is published 15.09.76, Bulletin No. 34
56. M.K. Dib, The precision manipulator for movement of a sample and Faraday’s cylinder in a
low-voltage elektronograf. Devices Sci. Res. 1, 43–48 (1976)
57. V.A. Roer, Installation of the manipulator for movement of a sample in ultrahigh vacuum of
an electronic spectrometer for research of spatial distribution of photoemission. Devices Sci.
Res. 3, 128–130 (1980)
58. H. Dusterhort, B. Vishot, The manipulator for movement of samples in ultrahigh-vacuum
installations. PTE 1, 124–125 (1983)
59. I. Dey, H. Li, F. Dgoyna, The ultra high vacuum three-axis manipulator for research
surfaces. Prib. Sci. Res. 6, 107–111 (1990)
60. Ac. 764007 SSSR, Cl. HO1 J37/285. The turning cartridge for samples of an field ion field
electron microscope. S.M. Borisov, E.P. Sheshin, D.A. Shomin
61. E.W. Mueller, Feldemission. Ergebnisse d. exat. Naturwiss. XXVII, 290–360 (1953)
62. B.V. Bondarenko, C.A. Yu, E.P. Sheshin, The controlled high-voltage power supply. PTE 1,
206–207 (1986)
63. A.S. Kyprashkin, A.G. Shahovskoy, E.P. Sheshin, The stabilized high-voltage power
supply. PTE 4, 238–223 (1991)
64. V.A. Granovsky, T.N. Siraya, Methods of Processing of Experimental Data at
Measurements (L.: Energoizdat, Leningr. otd., 1990)
65. G. Maks, Methods and Technology of Processing of Signals at Physical Measurements
(World, 1983)
66. B.V. Bondarenko, A.U. Cherepanov, E.P. Sheshin, System for measurement of emissive
characteristics of field emission cathodes. PTE 1, 245 (1987)
67. M.K. Miller, G.D.W. Smith, Atom Probe Microanalysis: Principles and Applications to
Materials Problems (M.R.S., Pittsburg, Pennsylvania, 1989) [Translation: M. Miller,
G. Smith, The Probe Analysis in Field Ion Microscopy, M.: World, 1993]
68. S.J. Savage, F.H. Froes, Metals 36(4), 20 (1984)
69. A.J. Melmed, R.J. de Klein, Physique 47(2), 287 (1986)
70. T. Masumoto, I. Ohnaka, A. Inoue, M. Hagiwara, Scripta. Metall 15, 293 (1981)
71. I. Ohnaka, I. Yamauchi, T. Ohmichi, T. Ichiryu, T. Mitsushima, T. Fukusako, in Rapidly
Quenched Metals, ed. by S. Steeb, H. Warlimont (1985)
72. I. Ohnaka, T. Fukusako, T. Ohmschi, T. Masumoto, A. Inoue, M. Hagiwara, in Proceedings
of 4th International Conference on Rapidly Quenched Metals, ed. T. Masumoto, K. Suzuki
(Japan Institute of Metals, 1982), p. 31
73. J. Liu, L. Arnberg, N. Backstrom, S. Savage, Mater Sci. Eng. 98, 21 (1988)
74. R. Maringer, C.E. Mobley, in Proceedings of 3rd International Conference on Rapidly
Quenched Metals, vol. 1, ed. by B. Cantror (Metals Society, London, 1978), p. 49
75. Y. Jashiro, T. Terao, A new method of specimen preparation for FIM and FEM. Surf. Sci. 67
(2), 605–610 (1977)
76. A.R. Bhatti, B. Cantor, D.S. Joag, G.D.W. Smith, Phil. Mag. B. 52, 63 (1985)
77. Pat. USA 414 3292 Cl/ 313-336, 25.06.76. Field emission cathode of glassy carbon and
method of preparation. S. Hosoki, H. Okano
78. R.J. Morgan, Sci. Instrum. 44, 808 (1967)
79. M.I. Elinson, G.F. Vasiliev, Field Electron Emission (GIFML, Moscow, 1958)
80. I.S. Andreev, Gyrn. Tehnich. Fiz. 22(9), 1428–1441 (1952)
81. E.W. Mueller, Zs. F. Phys. 108(9–10), 668–680 (1938)
82. E.W. Mueller, Zs. F. Phys. 102, (11–12), 734–761 (1936)
83. R. Haefez, Zs. F. Phys. 116(9–10), 604–609 (1940)
84. M. Benjamin, R.O. Jenkins, Nature 143, 599 (1939)
85. M. Benjamin, R.O. Jenkins, Phil. Mag. 26(12), 1049–1062 (1938)
86. M. Benjamin, R.O. Jenkins, Proc. Roy. Soc. 176(10), 262–279 (1940)
87. W.J. Tegart, The electrolytic and Chemical Polishing of Metals (Pergamon Press, Oxford,
1959)
88. P.A. Jacquet, Metall. Rev. 1, 157 (1961)
89. B.M. Tsarev, Achievements Phys. Sci. 36(2), 181–209 (1948)
90. M.I. Elinson, V.A. Gorkov, G.F. Vasiliev, Mag. Radio Eng. Electron. 2(2) (1957)
91. M.B. Benyminovich, B.G. Smirnov, G.N. Shyppe, Mag. Phys. Phys. (release 10) 23, 1690–
1699 (1953)
92. A.L. Syvorov, Devices Technol. Exp. 5, 5 (1969)
93. E.P. Sheshin, Structure of a Surface and Field Emission Properties of Carbon Materials
(MIPT-fizmatgiz Publishing House, 2001)
94. G.H. Freyberg, Production of thin field emitters. PTE 6, 176 (1967)
95. H.A. Hubner, Field emitter etching facility with good reproducibility. Optik 63(2), 179–183
(1983)
96. G.N. Freyberg, The improved semiconductor scheme of etching tips. PTE 5, 233–234 (1970)
97. A.J. Melmed, J.J. Carroll, Vac. Sci. Technol. 2, 1388 (1984)
98. A.J.J. Melmed, Chem. Phys. 38, 607 (1963)
99. R.-I.G. Garber, G.I. Dranova, N.A. Mansyrov, I.M. Mihaylovsky, Devices Technol. Exp. 1,
196 (1969)
100. A.S. Kypryshkin, A.G. Shahovskoy, E.P. Sheshin, The stabilized high-voltage power
supply. PTE 4, 238–239 (1991)
101. H. Lemke, T. Goddenbenrich, H.P. Bochem, U. Hartmann, C. Heiden, Improved microtips
for scanning probe microscopy. Rev. Sci. Instrum. 61(10), 2538–2541 (1990)
102. V.Y. Fridman, Production field emission tips by etching. PTE 1, 227 (1974)
103. G.N. Freybrg, The improved semiconductor scheme of etching tips. PTE 5, 233–234
104. A.c. №293515 USSR. Method of production of needle cold emitters. N.I. Komyak, V.G.
Pavlov, A.A. Rabinovich, V.N. Shrednik, Publication in B.I. 1975.31
105. A.c. №568981 SSSR. The device for etching field emitters. D.M. Pautov, A.V.
Kocheryzhnikov. C. HO1 J 9/02 of 04.01.76. Publication in B.I., 1977, No. 30
106. G.N. Freyberg, Double bath for automatic etching tips. PTE 4, 244–245 (1972)
107. E. Mueller, T. Tsong, Field Jon Microcopy (American Elsevier Publishing Company, Juc.,
New York, 1969)
108. V.I. Afanasev, V.M. Tybaev, T.I. Karpenko, T.I. Ivchenko, Electrochemical receiving thin
needles in a stream of electrolyte. PTE 1, 195–196 (1983)
109. H. Morikawa, K. Goto, Rev. Sci. V. Instrum. 59, 2195 (1988)
110. H. Morikawa, K. Goto, F. Iwatsu, T.J. de Terao, Physique 46(6), 589 (1987)
111. A.C. USSR 213200 Kl.HO1 J 1/30 of 8.06.66. A way of production of needle fieldemitters.
R.-I.G. Garber, V.I. Afanasyev, I.M. Mikhaylovsky
References 113
112. J.M. Walls, H.N. Southworth, B.J. Rushton, The preparation of Field electron/field ion
emitters by ion etching. Vacuum 24(10), 475 (1975)
113. M. Hellsing, Mater. Sci. Technol. 4, 824 (1988)
114. M. Drechsler, J.P. Prulhiere, Procedes et dispositifs pour usiner, reaffuter, chanffer et
nettoger des pointes par bombardment electronigue, patent of France No. 2098954, kl.
B 23 K 15/00 of 30.07.70
115. G.I. Dranova, B.V. Kylko, I.M. Mihaylovsky, Method of production of the needle field
emission cathode, a.s. 630669, Cl. HO1 J9/02, of 27.06.77
116. J.A. Liddle, A. Norman, A. Cerezo, C.R.M.J. de Grovenor, Physique 49(6), 509 (1988)
117. V.A. Kyznetsov, E.P. Sheshin, Energy spectrum of the field electrons emitted from
microledges on a tungsten tip. Radio Technician Electron. XX(7), 1550–1553 (1975)
118. T. Gurney, F. Hutchinson, R.D.J. Young, Chem. Phys. 42, 3939 (1965)
119. J.P. Jones, Nature 211, 479 (1966)
120. G.D.W. Smith, J.S. Anderson, Surf. Sci. 24, 459 (1971)
121. W.R. Graham, R.A. Reed, F.J. Hutchinson, Appl. Phys. 43, 295 (1972)
122. J.P. Jones, A.D. Martin, Surf. Sci. 41, 559 (1974)
123. A.P. Janssen, J.P. Jones, Surf. Sci. 41, 257 (1974)
124. G.L. Kellogg, Surf. Sci. 192, 879 (1987)
125. O. Nishikawa, Y. Tsunahima, E. Nomura, M. Wada, S. Horie, M. Shitata, T. Yoshimara,
R. Uemori, Surf. Sci. 126, 529 (1983)
126. T.T. Tsong, S.C. Wang, F.H. Liu, H. Chung, M. Ahmad, J. Vac. Sci. Technol. B1, 915
(1983)
127. E.D.J. Boyes, Comm. Metals 26, 207 (1972)
128. A.L. Emmanue, A.R. Moore, H.M. Pollock, Phys. Stat. Sol. 28a, 511 (1975)
129. R.D. French, M.H. Richman, Phil. Mag. 18, 471 (1968)
130. M.H. Richman, W.D. Sproul, Metallography 2, 149 (1969)
131. N. Ohmae, A. Nakamure, S. Koike, M.J. Umeno, Vac. Sci. Technol. A5, 1367 (1987)
132. S.V. Krishnaswamy, R. Messier, Y.S. Ng, T.T. Tsong, S.B.J. McLane, Non Cryst. Solids. 35
(36), 531 (1980)
133. S.V. Krishnaswamy, R. Messier, S.B. Mclane, Y.S. Ng, T.T. Tsong, Thin Solid Film 79(21)
(1981)
134. S.V. Krishnaswamy, R. Messier, C.S. Wu, S.B. McLane, T.T. Tsong, J. Vac. Set Technol.
18, 309 (1981)
135. K. Rendulic, E.W. Mueller, J. Appl. Phys. 38, 550 (1967)
136. C.C. Schubert, C.L. Page, B. Ralph, Eiectrochima Acta 18, 33 (1973)
137. G. Bauer, M.J. de Keisch, Physique 47(7), 189 (1986)
138. O. Nishikawa, T. Utsumi, J. Appl. Phys. 44, 955 (1983)
139. W.R. Graham, F. Hutchinson, D.A. Reed, J. Appl. Phys. 44, 5155 (1973)
140. E.S. Machlin, A. Freilich, D.C. Agrawal, J.J. Burton, C.L. Briant, J. Microsc. 104, 127
(1975)
141. J.A. Panitz, Rev. Sci. Instrum. 56, 572 (1985)
142. J.A. Panitz, I. Giaever, Ultramicroscopy 6, 3 (1988)
143. H. Norden, K.M. Bowkett, J. Sci. Instrum. 44, 238 (1967)
144. B. Loberg, H. Norden, Phil. Mag. 16, 1147 (1967)
145. B. Loberg, H. Norden, Proceedings of 4th European Regional Conference on Electron
Microscopy, Rome, p. 251 (1968)
146. D.A. Smith, K.M. Bowkett, Proceedings of 4th European Regional Conference on Electron
Microscopy, Rome, p. 261 (1968)
147. K.M. Bowkett, D.A. Smith, Field-Ion Microscopy (North-Holland, Amsterdam, 1970),
p. 215
148. M.J. Goringe, D.A. Smith, Jernkont. Ann. 155, 502 (1971)
149. R.M. Hatapova, L.L. Demskay, V.H. Romanova, Manufacturing techniques of carbon field
emitters. PTE 3, 205-2-7 (1985)
150. G.N. Shyppe, Electron Emission of Metal Crystals (SAGA Publishing House, Tashkent,
1959)
151. I.L. Sokolskay, Application of an field emission microscope for studying of surface diffusion
and self-diffusion. “Surface diffusion and spreading” under the editorship of Y.E. Geguzin,
M. Nauk of pp. 108–148
152. V. Zayt, Diffusion in Metals (M, Il., 1958)
153. J.E. Lennard-Jones, Proc. Phys. Soc. 49, 140 (1937)
154. I.N. Staranski, R. Suhrmann, Ann. Phys. 6 Folge 1, 153 (1947)
155. M. Drechsler, H.-Z. Angew, Phys. 6, 341 (1954)
156. R. Gomer, R. Wortman, R. Lundy, J. Chem. Phys. 26, 1147 (1957)
157. R. Comer, J.K. Hulm, J. Chem. Phys. 27, 1363 (1957)
158. G. Ehrlich, F.G. Hudda, Chem. Phus. 35, 1421 (1961)
159. H. Fujita, J. Chem. Phys. 21, 700 (1953)
160. S.S. Brenner, Surl. Sci. 2, 496 (1964)
161. R. Gomer, R. Wortman, R. Lundy, J. Chem. Phus. 27, 1009 (1957)
162. E.W. Mueller, Z. Elektrochemie 59, 372 (1955)
163. R. Klein, L.B. Leder, J. Chem. Phys. 38, 1866 (1963)
164. R. Gomer. Field Emission and Field Ionisation (Harvard University Press, 1961)
165. A.J. Melmed, J. Appl. Phys. 37, 275 (1966)
166. L.W. Swanson, R.W. Strayer, F.M. Charbonnier-, Surf. Sci. 2, 177 (1964)
167. A.G. Naymovec, Solid State Phys. 6, 2088 (1964)
168. H. Utsugi, R. Gomer, J. Chem. Phys. 37, 1706 (1962)
169. Y.V. Zybenko, I.L. Sokolskay, Radio Technician Electron. 9, 1467 (1962)
170. B.V. Bondarenko, V.Y. Raevsky, E.P. Sheshin, Field electron emission carbon fibers.
Theses of reports of ShM of the All-Union symposium on not heated cathodes, Tomsk,
pp. 47–48 (1980)
171. E.P. Sheshin, in Emission characteristics carbon fibers. Physical Processes in Devices
Electronic and Laser (Tekhnikiyu-M: MFTI, 1980), pp. 6–10
172. E.P. Sheshin, Field emission of carbon fibers. Ultramicroscopy 79, 101–108 (1999)
173. B.V. Bondarenko, Y.L. Rybakov, E.P. Sheshin, Field electron emission of carbon fiber.
Radio Technician Electron. Eng. 27(8), 1593–1597 (1982)
174. B.V. Bondarenko, Y.L. Pybakov, E.P. Sheshin, Field electron and field ion microscopy of
the emitting surface of carbon fiber. Theses of reports of III conference on field ionic
microscopy, Sverdlovsk, p. 30 (1982)
175. F.G.L. Van Oostrom, Validity of the Fowler–Nordheim model for field electron emission.
Philips Res. Rept. 21(1), 1–102 (1966)
176. A.L. Suvorov, E.P. Sheshin, V.P. Babaev, Influence of adsorption of residual gases on
emission properties of carbon materials. ZhTF 66(9), 164–169 (1966)
Chapter 3
Modern Developments in Theoretical
Research of Field Emission
Abstract This chapter discusses modern developments in theoretical research of

field emission from metals and semiconductors. Emitter shape approximation and
methods of electrostatic potential and field strength calculation are considered. It
also presents the theory of thermal-field emission (TFE) from metals. Energy dis-
tributions of field electrons and TFE electrons emitted from a metal are considered
in terms of applications for field emission spectroscopy. Theoretical aspects of
phenomena and processes on the emitter surface during field emission are discussed
and various theories attempting to explain them are presented.
3.1 Introduction
E. Müller’s outstanding invention, a field emission microscope, immediately

entailed purposeful experimental research of field emission from metals. The pri-
mary objective of the research was experimental verification of the basic statements
of the Fowler–Nordheim theory. The focus was on the central formula of the theory,
i.e. the Fowler–Nordheim formula, which can be simplified as follows:

B
j ¼ AE 2 exp ; ð3:1Þ
E
or
j B
lg ¼A ; ð3:2Þ
E2 E
where A and B are known constants [1].

The Fowler–Nordheim theory runs that for a field emission cathode with a set
work function, lg(j/E2) is to be linearly dependent on 1/E. The linear dependence
for lg(j/E2) = f(1/E) resulting from the experiment can be regarded as qualitative
experimental verification of the Fowler–Nordheim field emission theory, whereas
the resulting B value compliant with the theory can be taken as quantitative control.

116 3 Modern Developments in Theoretical Research of Field Emission
However, experimental studies of dependence of j(E) and determination of the

B value involve some complicating circumstances. The reason is that the experi-
ment measures not the density of field emission current j or electric field intensity E,
but total emission current I and potential difference V applied between the field
emission cathode and the anode. j = I/S, where S is the area of the emitting surface
of the apex field emission cathode, and the field intensity at the apex vertex E = bV,
where b is the value dependent on the apex-vertex shape provided that the
space-charge field in the vacuum gap between the field emission cathode and the
anode is negligibly small. So if S and b are constant, experimentally measured
dependence

lg I=V 2 ¼ f ð1=V Þ ð3:3Þ
is also to be linear. However, the E value can only be determined on the basis of
factor b. But this factor would have a constant value of 1/r all over the cathode
surface only for a ball-shaped cathode. For an arbitrarily-shaped apex cathode, b
differs between surface elements, i.e. b = b (n, η), where n and η are apex-surface
coordinates, and function b(n, η) depends on the apex shape. A formula for function
b(n, η) can be calculated from the potential theory only for some
axially-symmetrical apex shapes, namely a rotation hyperboloid, a rotation para-
boloid and a ball on a cone centered in its vertex. Therefore, the apex is usually
approximated by one of these shapes, which is referred to as hyperbolic approxi-
mation, parabolic approximation etc. However, real apices merely approximate
these shapes. Therefore, the error of determination of b(n, η) and, consequently,
E = bV, turns out essential, usually ΔE/E 30%. Moreover, the real apex surface,
even a single-crystal one, reveals different crystal faces having different work
functions Фikl, i.e. Ф = Ф(n,η). Generally speaking, inhomogeneity of the electric
field above the apex surface in different points of this surface and inhomogeneity of
the apex surface by work functions prevent from determination of field emission
current J from the apex by the Fowler–Nordheim formula derived for a cathode
with a definite work function on condition of homogeneity of the electric field
above the cathode surface. J can be determined by summing currents dJ coming
from particular elements dS of the cathode surface. Taking into account that
E = bV, current dJ is
" #
3=2
k 2
ðn; gÞ 0:965 S U ðn; gÞ
dI ¼ B0 V 2 exp 0
dS; ð3:4Þ
U0 ðn; gÞ V bðn; gÞ
consequently,
Z " #
3=2
0 k2 ðn; gÞ 0:965 S U0 ðn; gÞ
I ¼BV 2
exp dS: ð3:5Þ
U0 ðn; gÞ V bðn; gÞ
3.1 Introduction 117
It may seem that (3.4) assumes that dependence lg(I/V2) = f(1/V) cannot be
represented as a straight line and that it is difficult to verify the field emission theory
in such a way. However, a large body of experimental research of field emission has
arrived at linear dependence (3.3) is a wide range of V values and, consequently, the
E value. In particular, Hoffmann [1] measured that dependence (3.3) is linear within
18 orders of field emission current, and only very low and very high electric field
intensities cause appreciable deviations from linearity (Fig. 3.1). So the linear
dependence (3.3) experimentally determined in a wide range of voltages and cur-
rents seem to prove that the Fowler–Nordheim field emission theory is qualitatively
true. The linearity can be attributed to strong dependence of j on the value of Ф3/2/
b, so the measured real current comes from small local areas of the apex, where the
value of Ф3/2/b is minimal. It is this minimum that determines the slope of char-
acteristic (3.3).
Linearity deviations of dependence (3.3) in the area of high current densities, i.e.
at high field intensities, can be attributed to either
(a) effect of the space charge of the field-electron flow, which makes field intensity
E dependent on the current density and, consequently, non-linearly dependent
on potential difference V, i.e. value b is not constant if V changes;
(b) possible deviation of the law of the forces acting on the electron near the
surface at very small distances from the surface, from the electrical-image force
law.
No justified hypotheses of the nature of deviations of dependence lg(I/V2) = f(1/V)
from linearity in the area of small currents has been suggested yet.
Quantitative verification of the theory necessitates accurate determination of the
work function by an experiment. Comparison of the experimental and theoretical
curves shows that their slopes coincide in the order of magnitude. However, this
Fig. 3.1 Current-voltage lgi

characteristic of the field +1
emission current 0
-1
-3
-7
-11
-15
-19
0,5 1 2 3 4 5 6 104/V, V-1
conclusion is insufficient to prove agreement between the theory and the

experiment.
Determination of the work function depends on a possibility to estimate the b
field factor. There has not been any experimental attempt to make such estimation
yet. It may be no accident that all work function measurements by the field emission
method published up to now referred to particular crystalline planes. These values
came from comparison of the current from a particular crystalline plane with the
current from the entire emitting surface with subsequent supposition of the average
work function value for the entire surface.
Elinson et al. [2, 3] used their own theory to suggest a simple method of
determination of work function and field intensity. Their calculations of current
density are based on the integral
Z1 " ( )#1
4pmekT 8p21=2 m1=2 jEj3=2 hðe3=2 E 1=2 Þ
j¼ 1 þ exp
h3 3heE jEj
1
EþU
ln 1 þ exp dE; ð3:6Þ
kT
where E is the electron energy inside the conductor or the semiconductor.

Expression (3.6) is a mere product of the equilibrium E-energy electron current in
the range of E; E þ dE by penetrability of the potential barrier erected on the
surface by the image forces. The authors resorted to numerical calculations of the
current density (3.6) to plot lg j versus 1/E for values u ranging from 2.0 to 4.5 and
for T ranging from 0 to 3000 °C.
This method involves experimental plotting of the emitter volt-ampere charac-
teristic for two temperatures, T1 and T2. Presentation of the measurements as curves
showing lgI1 versus 1/V for T1 and lgI2 versus 1/V for T2 necessitates precise
knowledge of temperatures T1 and T2 and the emitter surface-area. Typical char-
acteristics for T1 = 1000 K and T2 = 2000 K are given in Fig. 3.2a.
The authors plotted theoretical dependences of lg(I2/I1) on U for some particular
values of I1 (Fig. 3.2b). The curve points are taken from the plot of lgj versus 1/
E. Then one arbitrary value, e.g. j1 = 105 A/cm2, is selected and the respective
value j2 is found from Fig. 3.2a. After that, the value of lg(I2/I1) is calculated.
Figure 3.2a shows it equal to 1.07. The intersection of the lg(I2/I1) = 1.07 curve
with the j1 = 105 A/cm2 curve in Fig. 3.2b gives the work function in a direct way.
In this case, it is equal to 3.23 eV.
The authors’ plot enables us to find the value of d(lgj)/d(1/E) if Ф has been
known and estimate the b field factor as the experimental value of d(lgI)/d(1/V) has
been known.
By approximation of the Fowler–Nordheim equation for the coordinates, a
straight line determines the line inclination tangent and the segment cut by the line
on the ordinate axis, which makes it possible to calculate the emission area with a
margin error of at least 15% and the field factor with even lower precision by the
ratios
(a) 10
(b) 1,8
9 1,6
T2=2000 K
8 1,4
I 1=
10
1,2
5
7 lg(I2/I1)
A/
cm
lgI2
) 2
6 1,0
I =1
lg(I2,A/cm2)
lg(I2/I1)
10
lgI1
6
I=
A/
5 0,8
10 7
cm
1
2
A/
)
cm
4 0,6 2
)
T1=1000 K
3 0,4
2 0,2
= 3,23 eV
1
1 2 3 4 5 6 2,0 2,5 3,0 3,5 4,0 4,5
104/V,V-1 eV
Fig. 3.2 Illustration of the method for determining the work function with field emission
measurement data used by Elinson et al. [2, 3]. a Experimental dependences lgI from 1/V for two
temperatures T1 and T2. b Theoretical dependences g(I2/I1) from Ф for three selected current
densities

I
S ¼ 4 1017 p2 ; ð3:7Þ
V2 1=V¼0
where p is the Fowler–Nordheim line slope ratio,
U3=2
p ¼ 2:82 109 ;
b
the current being expressed in amperes, the length in meters and the work function
in electron-volts. The first ratio is based on assuming that value of Ф2 exp
(9.81 Ф−1/2) is roughly constant. For the second one one needs to know value Ф
itself, which results in decrease in precision as the local work function value is
measured with great difficulty and low precision.
The Fowler–Nordheim experimental line determines the emitter parameters,
namely geometrical field factor b for given work function values and the emitting
surface area S.
If the emitter surface is coated with an adsorbate, the expressions (I = jS) and
(3.7) remain true for this new surface. Supposing that adsorption does not seriously
affect the b factor value, causing changes only in the work functions, we can
evaluate this change from formula (3.7):
2=3
Uads pads
¼ ; ð3:8Þ
Ucln pcln
where Ucln and Uads are the work function values before (a clean surface) and after
adsorption, respectively; pcln, pads are the Fowler–Nordheim line slope ratios before
(a clean surface) and after adsorption, respectively [4].
Equations (3.7) and (3.8) result from the assumption that the local electric field
and the work function are constant over the emitter surface and the emission area
does not change with voltage variation. Such assumptions are only true if the
electron emission current is measured from the individual face of the emitter
[5, 6].
The authors [7] have shown that the experimental work function value obtained
with the above-described technique is close to the minimum of the real work
function heterogeneously spread over the surface.
Most research papers are physically unreasonable when assuming the work
function to be constant over the emitter surface. However, the 10% scatter of work
functions over the surface and the 30% scatter of field intensities increase the scatter
of current densities more than twice [8].
On the other hand, ordinary observations of the emitter surface and the
emission area in the field emission microscope show changes of the emission area
even for a small interval of applied voltages V (ΔV/V = 10%) necessary to plot the
Fowler–Nordheim equation. In the course of increase in the applied voltage,
heterogeneity of the work function manifests itself in increase in the emission area
as emission from the areas with high work function becomes non-negligible in
case of strong fields [9]. However, the emission area all the same goes up with
increase in the field due to heterogeneity of distribution of the field over the
emission area near the apex foundation even in a simulation case at Ф = const
over the emitter surface [10].
There is one more method of experimental determination of the emission area
having no those disadvantages. The experimental value of the emission area S of a
tungsten cathode can also be evaluated from the emission image [7] according to
the formula
2
r
S ¼ Sim : ð3:9Þ
rim
Here curvature radius r of the experimental apex vertex is evaluated by the

voltage necessary to generate the total emission current I = 1 10−5 A [11]; rim is
the distance between centers (211) and ð121Þ [or ð211Þ and (121)] of the emission
image measured in pixels; Sim is the number of image pixels whose brightness
exceeds the noise level of the digital camera.
3.1.1 Emitter-Shape Approximation
Theoretical consideration of the problems related to apex emitters is based on the

emitter models good for mathematical calculations. To simplify mathematical
description of field emission, an apex emitter is very often approximated to a sphere
[4, 12–15].
Another approximated model is an orthogonal cone [16, 17]. In comparison with
a sphere, it has serious disadvantages as a cone-shaped model has a special point
that has to be considered separately from the rest of the problem.
At present, apex geometry is described with the models representing axially
symmetrical revolution surfaces of the second order [18–22]. Axial symmetry of such
models saves from extra mathematical difficulties caused by three-dimensionality of
the problem.
According to the authors of the paper [23], who microscopically probed a
considerable number of emitters viewed from different sides at different processing
stages, a real emitter is shaped closely to one of equipotential surfaces of an electric
field generated by a charged orthogonal cone topped with a sphere.
3.1.2 Potential and Field-Intensity Computation Methods
Knowledge of the apex shape is necessary but not enough to determine the field
over its surface. It is difficult to determine the field under real conditions, i.e. for a
real apex shape and with account of the anode geometry and the arm supporting the
apex etc. To find distribution of the field from the potential theory, researchers
usually resort to approximations.
The field calculated for a model shape of the surfaces of the emitter and the
anode is macroscopical and adequate only if the distance from the surface exceeds
*3/4 size of atomic-size surface irregularities [24]. It is about 1–2 nm [23]. The
local electric field immediately at the surface often considerably exceeds the
macroscopical field because of atomic-size surface irregularities. Calculation of the
local field from electrostatics can be inadequate. It is necessary to make
quantum-mechanical calculations of electron-charge distribution. This challenge
was only met for some special cases, which is why there were no theoretical
estimates of the field enhancement factor,
E
c¼ ; ð3:10Þ
Emacro
where E is the electric field at the emitter surface, Emacro is the macroscopic electric
field.
Value c is used as an adjustable parameter to compare the results of simulation
and experiment. It ranges *1–3.
The potential computation methods can be conventionally classified as follows.

1. Solution to the boundary-value problem for the Laplace equation in these or
those curvilinear coordinates, i.e. use of explicit formulas for the potential in the
whole space between the anode and the cathode.
2. Search of charge distribution that would generate the potential solving the
required boundary-value problem. This class of methods is referred to as
potential-theory methods.
3. Numerical solution of the Laplace equation for different apex shapes with
subsequent approximation of the results to some model dependences.
Solution of the Laplace equation with boundary conditions on the revolution
surfaces of the second order in the rectangular coordinate system seems rather
complicated. The mathematical calculations are simplified by introduction of the
orthogonal frame in such a way that emitter surfaces X correspond to the geo-
metrical locus where one of the coordinates is constant. Setting the boundary
conditions on the emitter surface and the surface to be researched, we can derive the
general solution of the Laplace equation.
Let us consider the Laplace equation for an orthogonal frame
X 3 pffiffiffi
1 @ g @V
r V ¼ pffiffiffi
2
¼ 0; ð3:11Þ
g i¼1 @qi gii @qi
where gii stands for the metric-tensor components and

pffiffiffi pffiffiffiffiffiffipffiffiffiffiffiffipffiffiffiffiffiffi
g ¼ g11 g22 g33 ð3:12Þ
is a function depending only on this coordinate in the area limited to two coordinate
surfaces connected with one common curvilinear coordinate qi, i = 1, 2, 3…
As a result, if
V ¼ V1 for q1 ¼ n1 ; 8 q2 ; q3 ð3:13Þ
and
V ¼ V2 for q1 ¼ n2 ; 8 q2 ; q3 ð3:14Þ
for all V1 6¼ V2 and if it is acceptable to represent

pffiffiffi pffiffiffiffiffiffipffiffiffiffiffiffi
g g22 g33
¼ pffiffiffiffiffiffi ¼ Fðq1 Þ f ðq2 ; q3 Þ; ð3:15Þ
g11 g11
then the Laplace equation solution (3.11) depending on coordinate q1 can be rep-
resented as
Z
1
V ¼ C1 dq1 þ C2 ; ð3:16Þ
Fðq1 Þ
where C1, C2 are the constants that can be found from boundary conditions (3.13),
(3.14).
Having found out the electric potential, one can find the intensity of electric field
E as a gradient of function V:
X3
1 @V
E ¼ grad ðVÞ ¼ pffiffiffiffiffi ~
ei ; ð3:17Þ
i¼1
gii @qi
where ~ei are unit vectors.

The electric field intensity in the area limited to two coordinate surfaces [con-
ditions (3.13), (3.14)] can be represented as
C1
E ¼ pffiffiffiffiffiffi ~
e1 : ð3:18Þ
g11 Fðq1 Þ
The Laplace equation solution for potential V and electric field intensity E for the
coordinate surfaces of the second order in the respective curvilinear coordinates is
illustrated by a two-sheeted hyperboloid of revolution.
Potential V and electric field intensity E for a revolution hyperboloid, a revo-
lution paraboloid and flattened and extended revolution ellipsoid [25] in the
respective curvilinear coordinates are specified in Table 3.1.
According to the authors of the paper [22], who microscopically probed a
considerable number of emitters viewed from different sides at different processing
stages, a real emitter is shaped closely to one of equipotential surfaces of an electric
field generated by a charged orthogonal cone topped with a sphere. This equipo-
tential at the very top is well approximated by the sphere. It is followed by a
transitional area, which eventually turns into the orthogonal cone.
Potential distribution for an insulated charged orthogonal cone with a spherical
end is given by expression [22]
Table 3.1 The electric field potential and intensity in curvilinear coordinates. Here C1 is a
constant determined from the respective boundary conditions. It is individual for every revolution
surface
Revolution surface Potential V Field strength E

Two-sheet hyperboloid C1 ln tg g þ C2
2 pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
C1
eg
ch2 lcos2 g0
a sin g0

Elongated ellipsoid C1 ln th l2 þ C2 pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
C1
el
a sh l0 ch2 l0 cos2 g
Paraboloid C1 ln ðlÞ þ C2 pCffiffiffiffiffiffiffiffiffiffi
1 ffie
2 l
l0 2 l0 þ k
Oblate ellipsoid C1 arctg ðsh lÞ þ C2 pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

C1 ffi el
a ch l0 sh2 l0 þ cos2 g
VR n
V¼ ðr a2n þ 1 r n1 ÞPn (cos h). ð3:19Þ
Rn
Naturally, the potential distribution will not change if one equipotential surface
is taken as an emitter and the other as an anode. In (3.19), V is the potential counted
from the cone, r and h are ordinary polar coordinates, R is the distance to the anode,
VR is the applied voltage, a is the sphere radius, Pn is the Legendre first-kind n-order
function of the, n being selected such that if h = a where a is an external semiangle
of the cone, Pn = 0 (n = 0 for a = 0, n = 1 for a = p/2). The equation of the
theoretical anode is derived from the equation U = UR In this case, the summand
a2n þ 1 r n1 r n is neglected and
Rn
rn ¼ : ð3:20Þ
Pn ðcos hÞ
Analyzing the equipotential surface shapes for orthogonal cones with different
angles and different sphere diameters, it is quite easy to select the equipotential that
is good at approximation of the experimental emitter.
3.1.3 Comparison of Theoretical and Experimental Results
As we have already noted, the cathode configurations implemented in practice are

very complex and various, which makes it practically impossible to match them
with a quite simple geometrical figure. Depending on the nature of the problem to
be solved, the cathode can be simulated either by a sphere, or an ellipsoid, or a cone
or a cylinder etc. However simple the models listed might be, the fullest theory of
field emission from metal cathodes is based on an even simpler electron emitter
model in which the cathode is represented as an ideal plane. The experimental
verifications of Fowler–Nordheim theoretical ratios carried out with a Müller field
emission microscope involved only apex structures as test subjects.
Though the technique used to compare the theoretical and experimental results is
generally accepted, it can hardly be considered as quite correct. Apparently, such
comparison has to involve a possibility to pass from values I and V measured in the
experiment to j and E characteristics. However, that can be really implemented only
when independent experiments and calculations have given the real configuration of
the field emission cathode. Certainly, mathematically and physically correct field
emission theory should be built with account of real or at least close to real but not
flat shape of the field emission cathode.
The present stage of research of field emission both from metals and semicon-
ductors began in the 1950s with Dyke group research and has been going on ever
since. We cannot but mention that the scientists of the former USSR republics and,
first of all, Russian researchers, including the research teams and groups of M.I.
Elinson, G.N. Shuppe, I.L. Sokolskaia, G.A. Mesiats, G.N. Fursei, V.N. Shrednik,
B.V. Bondarenko, E.P. Sheshin, N.V. Mileshkina, G.G. Vladimirov, R.Z.

Bakhtizin, N.V. Egorov, V.G. Ivanov, V.M. Zhukov, T.A. Tumareva etc., and
Ukrainian scientists including P.G. Borzyan, A.F. Yatsenko and A.G. Naumovets,
largely contributed to understanding of the field emission phenomenon, as well as
the processes accompanying it on the surface and in the bulk of the field emission
cathode, which developed the field emission theory and its practical applications.
The first successful attempts to improve the field emission theory were made by
Dike et al. [2, 22, 25–34], Murphy, Müller and Good [35, 36, 96], Guth and Mullin
[37], Elinson et al. [2, 38], somewhat later by Christov [39], Itskovich [40, 41],
Modinos [32] etc.
3.2 Thermal-Field Emission (TFE) from Metals
Let us turn to Fig. 1.2. Fowler and Nordheim (cf. Chap. 1) calculated the density of
field emission current from energy band A at T = 0. If T > 0, emission involves not
only these electrons, but also the electrons of bands B, B1 and those higher in
comparison with the case when T = 0. The problem becomes much more compli-
cated: firstly, we can no longer use the limiting expression for distribution of electrons
by energies; and, secondly, for energies of the electrons lying close to the top of the
potential barrier, we can no longer use the expression for transparency D found in
Chap. 1. Gas and Mallin [37] were the first to find an analytical solution of calculation
of D for the case of T > 0. Later Andreev [42] applied the same scheme to make
calculations giving essentially similar, but somewhat simpler expressions for fun-
damental thermal field emission regularities. That was how the first thermal-field
emission theory was developed. Later, more general assumptions excluding the
restrictions accepted in paper [37] enabled numerical solution of the problem [39, 43].
Let us give a brief calculation of transparency of the barrier and the current for
band B1. Schrödinger equations for areas x < 0 and x > 0 can be written down as
d2 w
dx2 þ Ew
¼0 ðx\0Þ;
d2 w ð3:21Þ
dx2 þ EWþ 1
2x þ x
2x20
w¼0 ðx [ 0Þ:
The potential-energy expression is transformed near x = x0. If we limit ourselves to

the members containing x2, we can obtain
1 ðx x0 Þ2
V ¼W þ : ð3:22Þ
x0 2x30
This expression shows quite good approximation of the barrier near the maximum.
Some transformations and subsequent insertion of expression (3.22) into (3.21)
turns the latter into a parabolic cylinder equation. The integration methods for such
an equation are well-known.
Linking the two solutions as usual

dw dw
w þ 0 ¼ w0 ; ¼ ;
dx þ 0 dx 0
and using an ordinary expression for

2
a0
D ¼ 1 ;
an
we eventually get the expression of

1
D ðE; EÞ ¼
1 þ exp ½2paðE W 0 Þ
1 ð3:23Þ
W2 1 exp½paðE W 0 Þcos UB
1 2 2 ;
2 4x0 W f1 þ exp½2paðE W 0 Þg32
suitable for C-band electrons with the energies close to the top of the barrier and the
expression
ð3:24Þ
suitable for B1-band electrons.

Expressions (3.23) and (3.24), as well as the subsequent ones, include the fol-
lowing designations:
3 12 pffiffiffiffiffiffiffiffi
x0 1
a¼ EWþ ; W 0 ¼ W e3 E ;
2 x0
pffiffiffi 1
4 2 2 1 2 W2
UC ¼ x0 1 þ arctg baðW 0 EÞ;
3 W2 4 ð3:25Þ
pffiffiffi 1
4 2 12 2 W2
U B1 ¼ x þ arctg baðE W 0 Þ;
3 0 W 12 4
p6 m3 e9
b ¼ c þ 2 lg 2; c ¼ 0:5772; B ¼ 160 :
ch7
Then we use the expression for n(E)

E E0
n ðEÞ BkT exp
kT
3.2 Thermal-Field Emission (TFE) from Metals 127
true for a high-temperature case, after which it is easy to calculate the density of
emission current from band B1 and band C near the top of the potential barrier.
Density of current from band B1 and C
(
u0
2 kT
X
1
ð2n þ 1Þd
jB1 þ C ¼ BðkTÞ e ð1Þn
n¼0 nðn þ 1Þd2 þ d 1
2
1
2

W W 6
2 cos UC
1 2 4n o12
2kT 4x0 2 ð3:26Þ
pa þ kT1 þ b2 a20
39
>
=
cos UB1 7
þn o 15 :
2 2 >;
2pa 1 þ b2 a2 kT 0
Here U0 ¼ U e3 E; where U is the work function; d = 2pa0kT;
3 12
x
a0 ¼ 0 :
2
In the limiting case of high temperatures and weak fields, the current-density
expression turns into the famous Schottky formula if we neglect a small periodic
member, assume d 1 and use the first member of the sum
0
jB1 þ C ¼ B ðkTÞ2 eU =kT : ð3:27Þ
A better understanding of the temperature effect of field emission can come from the
cases of higher fields (107–108 V/cm) and low and average temperatures (300–
1500 K). Under these conditions, electrons from the B-band, mostly those coming
from its lower part, start playing an increasingly more important role in emission.
The total density of current of bands A and B can be expressed by the final
formula
(
E2 X1
ð1Þn þ 1
j¼ 1:55 106 þ 120; 5T 2
h2 U n¼1
n
!)
1 1
1 þ 1 ð3:28Þ
n þ 8:813 103 hU2 TE n 8:813 103 hU2 TE
" 3
#
6:85 107 U2 h
exp A/cm2 ;
E
where Ф is measured in eV and E in V/cm. This formula is general for thermal-field

emission.
Temperature rise causes a considerable increase in field emission current. In

particular, if T = 1000 K and E = 2 107 V/cm, (3.28) gives the current that is
nearly five times as heavy as that given by the Fowler–Nordheim equation, but if
E = 3 107 V/cm, the current is only 1.7 times as heavy. Large fields cause a less
appreciable temperature effect.
The latest and fullest TFE theory was suggested by Murphy and Good [35].
Let us consider a semi-restricted metal occupying the semi-space ranging from
x = −∞ to x 0. According to Sommerfeld, electrons of the metal are in a rect-
angular potential pit. Let us count energy from the bottom of this potential pit. The
potential energy of an electron in the vacuum near the metal-vacuum boundary is
given by
V ðxÞ EF þ Ue2 =4x: ð3:29Þ
In the presence of external electric field E, potential (3.28) gets an addition eEx.
As a result, the potential of the field that meets the electron in electron-emission
experiments is given by the approximate expression:

EF þ U e2 =4x eEx; x [ xc ;
VðxÞ ¼ ð3:30Þ
0; x [ xc ;
where xc is determined from the condition that V(xc) = 0. Figure 3.3 shows a
potential barrier for typical value EF þ U and external electric field intensity
characteristic for field emission experiments. Please note that potential (3.29) peaks
at
xm ¼ ðe=AEÞ1=2 ¼ 1:9=E 1=2 ; ð3:31Þ
Fig. 3.3 The near-surface V, eV

potential barrier for electrons
at the FEE (solid line);
10
contributions to the potential
of the mirror image force EF Φ
8
(dashed curve) and an
external electric field 6
E = 0.3 V/Å (dashed dotted
curve) 4
2 Metal Vacuum
0
0 10 20
z,Ǻ
if xm is measured in ångströms and E in volts/ångström. This maximum potential in

electron-volts is equal to
1=2
Vmax ¼ EF þ w e3 E ¼ EF þ w3:79E 1=2 : ð3:32Þ
In the version considered, the TFE theory is built as follows. Let us assume
(which is quite reasonable as electric field inside the metal can be neglected) that the
number of electrons falling on the barrier surface from the metal bulk with energy
ranging from W to W + dW does not change during emission and equals its equi-
librium value of
Z1
mdW mkB T W EF
n ðW; TÞ dW ¼ 2 3 f ðEÞdE ¼ ln 1 þ exp dW: ð3:33Þ
2p h 2p2 h3 kB T
W
The probability for every electron to pass through the superficial potential barrier
is equal to D (W). Consequently, emission current density, i.e. the number of
electrons emitted from the unit area per unit time multiplied by electron charge can
as before be expressed by the ratio
Z1
j ðE; TÞ ¼ e n ðW; TÞ D ðWÞ dW; ð3:34Þ
0
where T and E designate temperature and the field applied, as usual. Murphy and
Good showed that one can obtain quite a simple expression for j(E,T) that is true at
any values of E and T.
The potential barrier-transmission coefficient is given by
D ðE;EÞ ¼ fl þ exp½Q ðEÞg; ð3:35Þ
where
Zz2
Q ðEÞ ¼ 2i kðnÞdn; ð3:36Þ
z1
1=2
k ðnÞ ¼ ð2m=h2 Þ ðWEF U þ e2 =4n þ eEnÞ ; ð3:37Þ
and x1, x2 are the roots of equation
k2 ðxÞ ¼ 0: ð3:38Þ
For the case of W < Vmax, roots x1 and x2 are both valid and x1 < x2. For the case
of W > Vmax, roots x1 and x2 are mutually complex conjugate with Im(x1) > 0 and
Im(x2) < 0. Please note that according to (3.37), function k(n) has singularity in a
point n = 0. It does not affect applicability of (3.33) in cases where W < Vmax, but it
cannot be ignored at the energies significantly exceeding Vmax when Re(x1) = Re
(x2) approaches zero. If
pffiffiffi
E [ EL Emax þ ð1 1= 2Þ ðe3 EÞ1=2 ; ð3:39Þ
then (3.35) can turn out unsuitable for application. However, for the energies
W > WL, the transmission coefficient is very close to zero, which is why, for the
current density to be assessed, it can be assumed that
D ðW; EÞ 1: ð3:40Þ
Calculation of D ðW; EÞ at W < WL is reduced to calculation of the integral

determined by ratio (3.36). The result of this calculation is given by
pffiffiffi 1=4
4 2 m2 e5
Q ðWÞ ¼ y3=2 v ðyÞ; ð3:41Þ
3 Eh4
y ðe3 EÞ1=2 jEF þ U W j; ð3:42Þ

( " # " #)
y 1=2 y 1 1=2 y 1 1=2
v ðyÞ 2E þ ðy þ 1ÞK ; y [ 1;
2 2y 2y
ð3:43Þ
( " # " #)
1=2 1 y 1=2 1 y 1=2
vðyÞ ð1 þ yÞ E yK ; y\1; ð3:44Þ
1þy 1þy
where K[k] and E[k] are elliptical integrals
Zp=2
K ½k ¼ ð1 k2 sin2 hÞ1=2 dh; ð3:45Þ
0
Zp=2
E ½k ¼ ð1 k2 sin2 hÞ1=2 dh: ð3:46Þ
0
Please note that y > 1 corresponds to the energies above the barrier maximum
(W > Vmax), a y < 1 corresponds to the energies below this maximum (W < Vmax).
Table 3.2 gives the values of v(y) for y
1 calculated in [35].
Table 3.2 Values of y v(y) t(y) s(y)

functions v(y), s(y) and t(y)
0.00 1.0000 1.0000 1.0000
0.05 0.9948 1.0011 0.9995
0.10 0.9817 1.0036 0.9981
0.15 0.9622 1.0070 0.9958
0.20 0.9370 1.0111 0.9926
0.25 0.9068 1.0157 0.9885
0.30 0.8718 1.0207 0.9835
0.35 0.8323 1.0262 0.9777
0.40 0.7888 1.0319 0.9711
0.45 0.7413 1.0378 0.9637
0.50 0.6900 1.0439 0.9554
0.55 0.6351 1.0502 0.9464
0.60 0.5768 1.0565 0.9366
0.65 0.5152 1.0631 0.9261
0.70 0.4504 1.0697 0.9149
0.75 0.3825 1.0765 0.9030
0.80 0.3117 1.0832 0.8903
0.85 0.2379 1.0900 0.8770
0.90 0.1613 1.0969 0.8630
0.95 0.0820 1.1037 0.8483
1.00 0.0000 1.1107 0.8330
As a result, we get the following general expression of emission current density:
Z1
j ðE; TÞ ¼ e N ðW; TÞ D ðW; EÞdW
0
8W
ZL
emkB T < W EF dW
¼ ln 1 þ exp ð3:47Þ
2p2 h3 : kB T 1 þ exp QðWÞ
0
9
Z1 =
W EF
þ ln 1 þ exp dW :
kB T ;
WL
Under some conditions, the integrals of expression (3.47) can be calculated

analytically. (3.33) shows that rather low temperatures make function n(W,
T) dramatically decreases in the area of W [ EF . On the other hand, the calculated
transmission factor D (W; E) for a typical superficial barrier (U 4 eV,
E 4 107 V/cm) dramatically decreases at W\EF . Under such circumstances,
the second integral in (3.47) makes a negligible contribution to j(E, T), and the
integrand in the first integral disappears from everywhere except for the vicinity of
the Fermi level. So the second integral in (3.47) can be omitted, and the limits in the
first integral can be replaced with −∞ and +∞. In addition, for typical values of U
and E, U = 3 eV, E < 6 107 V/cm:
exp½Q ðVÞ 1; ð3:48Þ
so the one in the denominator of the first integral can be omitted in (3.47). Then we
get
Z1
emkB T W EF
jðE; TÞ ¼ exp½QðWÞ ln 1 þ exp dW: ð3:49Þ
2p2 h3 kB T
1
The main contribution to this integral is given by the vicinity of the Fermi level.
So the exponent of the transmission coefficient can be replaced with the first two
members of Taylor expansion in powers of WEF :
Q ðWÞ ¼ b0 þ c0 ðWEF Þf0 ðWEF Þ2 þ ; ð3:50Þ
where
1=2 3=2
b0 ¼ ð4=3Þ 2m=h2 U =eE vðy0 Þ; ð3:51Þ

b0 ¼ 0:683v 3:79F 1=2 =U U3=2 =E: ð3:52Þ
In (3.52), the value of U is measured in electron volts and E in volts/ångström.

The coefficients
1=2 1=2
b0 2 2m=h2 U =eE tðy0 Þ; ð3:53Þ

b0 ¼ 1:025 U1=2 =F t 3:79E1=2 =U ; ð3:54Þ
1=2
f0 ð1=2Þ 2m=h2 vðy0 Þ ½eEU1=2 ð1e3 E=U2 Þ1 ; ð3:55Þ
f0 ¼ 0:256 ðEU1=2 Þ1 vð3:79E 1=2 =UÞ ðl 14:36E=U2 Þ1 : ð3:56Þ
where
t ðyÞ ¼ v ðtÞð2=3Þ ydv=dy; ð3:57Þ

1=2
y0 e3 E =U; ð3:58Þ
y0 ¼ 3:79E1=2 =U: ð3:59Þ
Table 3.2. shows the numerical values of function t(y). Please note that under the
conditions of field emission, y < 1. Having kept only the first two members in
(3.50) and inserted them in (3.49), we will obtain
Z1
emkB T W EF
j ðE; TÞ ¼ exp ðb0 Þ exp ½c0 ðW EF Þ ln 1 þ exp dW:
2p2 h3 kB T
1
ð3:60Þ
Please note that at c0kBT < 1, the integrand in (3.60) decreases exponentially at
W\EF , behaves in the same way as the initial expression under the integral sign in
(3.49) at W [ EF and coincides with it near the Fermi level. The integral in (3.60)
can be calculated analytically to produce the field emission current density
" #
e3 E 2 pc0 kB T 4 2m 1=2 U3=2
j ðE; TÞ ¼ exp vðy0 Þ : ð3:61Þ
16p2 hUt2 ðy0 Þ sin ðpc0 kB TÞ 3e
h2 E
It can be shown that the expression (3.61) for emission current density holds true
if the following two conditions are met:
1=2 1=4
U e3 E [ p1 h4 = m2 e2 ðeEÞ3=4 þ kB T ð1 c0 kB T Þ1 ; ð3:62Þ
1c0 kB T [ ð2f0 Þ1=2 kB T: ð3:63Þ
Figure 3.4 shows the field emission and TFE bands for U = 4.5 eV determined
by ratios (3.62) and (3.63). For the points lying on the border of the field emission
band, the error in the current-density value (3.61) changes within the range of 15–
40% in comparison with the exact evaluation in (3.47).
At very low temperatures, i.e. when pc0 kB T 1, it is reduced to the
well-known Fowler–Nordheim equation,
j ðEÞ ¼ A0 E 2 exp ðB0 U3=2 =EÞ; ð3:64Þ
where
h 1=2 i1
A0 ¼ e3 16p2 hUt2 e3 E =U ; ð3:65Þ
1=2 3 1=2
B0 ¼ ð4=3eÞ 2m=h2 v eE =U : ð3:66Þ
Fig. 3.4 Areas of the T, K

temperature and applied field
for TEE and FEE at
Ф = 4.5 eV [10]
n
ctro
ele
3000
ion ermo
h
em a of t
iss
Are
2000
1000 Area of field electron

emission
0
0,2 0,4 0,6 0,8 1,0 E, V/Å
Having inserted the respective numerical values for different constants, we can
find that
" ! #
1:537 1010 E 2 0:683U3=2 3:79E 1=2
j ðEÞ ¼ 2 exp v ; ð3:67Þ
Ut ð3:79E 1=2 =UÞ E U
where j(E) is expressed in amperes/square centimeter, U in electron volts and E in

volts/ångström.
According to (3.67), the slope of the Fowler–Nordheim plot is expressed by the
formula of

d ln ðj=E2 Þ 3:79E1=2
SFN ¼ ¼ 0:683s U3=2 ; ð3:68Þ
d ð1=EÞ U
where
s ðyÞ ¼ v ðyÞðy=2Þ dv=dy: ð3:69Þ
The numerical values of s(y) are given in Table 3.2. The table shows that s(y) is
almost constant in the narrow range of variation of the field applied. Consequently,
it is theoretically possible to determine the work function from the slope of the
Fowler–Nordheim plot. In practice, determination of the work function is impeded
by impossibility of sufficiently precise determination of the absolute value of the
local field on the emitting plane. In addition, please note that the factor of pro-
portionality preceding U3/2 in (3.68) depends on the selection of the shape of the
surface potential barrier.
In conclusion of the paragraph, please note that Murphy-Good formula (3.61) for
the TFE-current density can after some transformations be compacted as follows:
ðT6¼0Þ ðT¼0Þ px
jTFE ¼ jTFE ; ð3:70Þ
sin px
where x ¼ 9:3 103 U0:5 T=E is a unitless parameter, U is taken in eV, T is in K

and E in volts/centimeter.
Expanding sin px into a series at small T, ratio (3.70) can be transformed into

ðT6¼0Þ ðT¼0Þ T2
jTFE ¼ jTFE 1 þ 1:40 108 U : ð3:71Þ
E2
(3.71) enables qualitative determination of the features of dependence of field

emission on temperature. For weak fields, the temperature influence is more sig-
nificant when U = const, and when E = const, it is more appreciable for the bands
with larger U. At high temperatures (x 2/3), the approximations making it
possible to derive simple expression (3.70) are no longer valid.
3.3 Field Emission Spectroscopy. Energy Distribution

of Field Electrons and Thermofield Electrons Emitted
from a Metal
A Müller electron microscope (cf. Chap. 1) enables numerous experimental

research of field emission and, first of all, verification of the main conclusions of the
field emission theory, including, primarily, the basic formula of field emission from
metals, the Fowler–Nordheim formula, the ratios obtained by Stratton for field
emission from semiconductors etc. Some other facilities to research field emission
and study the electron structure of metals and semiconductors include the method of
field emission spectroscopy or the method based on research of energy distribution
of field electrons by energies (velocities) [32, 33, 44–47].
The most typical difference of distribution of field electrons by energies from
distribution of thermo- or photoelectrons [44, 45] is accounted for by the field
emission mechanism itself, i.e. the penetration of electrons through the potential
barrier due to tunnel effect. That cannot but shift the distribution curve towards the
negative electron energies (or positive potentials) on the collector by the value of
the collector work function. Figure 3.5 shows the chart illustrating the situation.
As the potential-barrier penetrability at the collector can be neglected in view of
smallness of the electric field, field electrons can only enter the collector when the
fastest (at T = 0 K) electrons emitted from level Гe transit under the surface barrier
at the collector, for which the collector has to be supplied with positive potential uc.
The very first measurements taken by Henderson et al. [48–50] showed presence of
this characteristic shift.
E A C
φe
φc
je φc
jc
Fig. 3.5 The potential distribution in the three electrode system [emitter (E)—anode (A)—
collector (C)] suitable for measuring the distribution of field emission electrons energy
Let us now turn to the theoretical aspects of the method [32, 44, 46, 51, 52].
Energy distribution of field electrons. The number of electrons emitted from the
unit area per unit time with full energy in the range from E to E þ dE can be written
down as J ðEÞ dE. Function J ðEÞ is known as full-energy electron distribution. Let
us designate the full-energy distribution calculated within the framework of the
theory of metals in free-electron approximation as J0 ðEÞ. What we get is
ZE
J0 ðEÞ ¼ N ðE; Ex ÞDðEx ; EÞdW; ð3:72Þ
0
where NðE; Ex Þ and DðEx ; EÞ are determined by the ratios given in Chap. 1 and the
previous sections of this chapter. According to the assumptions made above, the
potential barrier can be given as (3.30) and the barrier-penetrability value can be
found according to (3.35)–(3.41). In the field emission band, the bracketed one in
(3.35) can be omitted according to (3.62) and (3.63) in the same way as has already
been done when calculating the emission-current density. What we will as a result
will be
ZE
J0 ðEÞ ¼ f ðEÞðm=2p h Þ2 3
exp ½QðEx ÞdEx ; ð3:73Þ
0
where Q ðEx Þ is known from (3.41). QðEx Þ can be substituted by the first two
members of Taylor expansion in the vicinity of Ex .
As a result, we have
3.3 Field Emission Spectroscopy. Energy Distribution … 137
Q ðEx Þ bE þ ðEx EÞ = dE ; ð3:74Þ

1 2m 1=2 ðU þ EF EÞ3=2
bE vðyE Þ; ð3:75Þ
4 h2 eE
1=2
2m ðU þ EF EÞ3=2
dE1 2 tðyE Þ; ð3:76Þ
h 2 eE
1=2
yE e3 E =ðU þ EF EÞ; ð3:77Þ
v(yE) and t(yE) are the functions given in Table 3.2. If (3.74) is inserted into (3.73),
the resultant integrand is small everywhere except for the vicinity of point Ex ¼ E.
Therefore, the lower limit of the integral can be substituted with −∞. Then ana-
lytical integration produces
" #
mdE f ðEÞ 4 2m 1=2 ðU þ EF EÞ3=2
J0 ðEÞ ¼ exp vðyE Þ : ð3:78Þ
2p2 h3 3 h2 eE
If expansion in E (3.74) in (3.73) is substituted with expansion in EF, then
QðEx Þ ¼ b0 þ c0 ðE EF Þ f0 ðEx EF Þ2 þ
The expression for full-energy distribution (Young formula [53]) is
j0 expðM=d0 Þ
J0 ðMÞ ¼ ; ð3:79Þ
ed0 1 þ expðM=pd0 Þ
where
M E EF ; ð3:80Þ
p kB T=d0 ; ð3:81Þ
d01 c0 ; ð3:82Þ
and j0 is the emission-current density at the zero temperature as determined by the

Fowler–Nordheim formula. It stands to reason that the Young formula can only be
used for the energies close to the Fermi level (*0.5 eV). However, that is the
energy band that determines the field emission current value. It is obvious that the
emission current density can be expressed in terms of J0 ðMÞ as
Fig. 3.6 Theoretical j, r. u.

dependences of the
1,2
distribution function of the
total energy (DTE) for
1,0
different values of p: 1 0; 2 1
0.048; 3 0.100; 4 0.200; 5
0,8
0.300; 6 0.400; 7 0.500; 8
0.600
0,6
0,4
8
7
0,2 6
5
4
2 3
0
-4 -3 -2 -1 0 1 2 E/d0
Zþ 1
j ðE; TÞ ¼ e J0 ðMÞdM: ð3:83Þ
1
Figure 3.6 gives a set of theoretical curves of the full-energy distribution cal-
culated by Swanson and Krauser [54] on the basis of (3.79) for different values of
parameter p.
3.3.1 Energy Distribution of Thermal-Field Electrons
In this case, penetrability D is to be calculated from ratios (3.35)–(3.42), and the

integral in (3.47) should be evaluated numerically. Such calculations were made in
a number of papers for different values of the field and the temperature [55–57].
Here we are not going to give end full-energy distribution formulas published in
these papers in view of their complexity and insufficient ostensiveness, but limit
ourselves with the characteristic curves plotted on the basis of these formulas (cf.
Figs. 3.7 and 3.8).
In case of weak fields, the main contributor to the current is the narrow energy
band over the top of the barrier. In case of intermediate fields, comparable emission
comes from the area near the Fermi level and in the vicinity of the top of the
near-surface barrier. In case of strong fields, emission comes mostly from the area
near the Fermi level. It is obvious that the full width at half maximum (FWHM) of
the full-energy distribution curve will have a maximum in case of intermediate
fields. It is quite clearly illustrated by Fig. 3.7. The curves shown in that plot were
calculated by Bell and Swanson [57].
Let us consider one more interesting aspect of thermal-field full-energy distri-
bution. According to Gadzuk and Plammer [56], the shape of such distribution near
Fig. 3.7 Theoretical FWHM, eV

dependences of the full width
at half maximum (FWHM) of 2,0
emitted electrons DTE from
the applied field at various
temperatures for U = 4.5 eV
[60] 1,6
1,2
Т = 1800 К
1500 К
0,8
1200 К
900 К
600 К
300 К
0,4 78 К
0 0,4 0,8 1,2 V/Å
Fig. 3.8 Theoretical T=1570 K

dependences of the DTE
function for U = 4.8 eV [59] E=0,41 V/Å
0,31
0,26
0,21
0,16
0,13
0,1
-2 -1 0 1 2 3 , eV
the top of the barrier has to be sensitive to the shape of the barrier itself. For the case
of stronger fields (Fig. 3.8), coordinate xm (3.31) of the barrier maximum counted
from the metal surface is rather small, usually x 3–4 Å. That is why any dis-
crepancy between the theory and the experiment caused by inadequacy of the
image-force potential at small distances should in principle manifest itself as a
divergence from the theoretical full energy distribution curve in the area of the
strong fields. Certainly, such quantitative analysis assumes independent knowledge
of the work function and precise measurements of field E on the emitter surface. At
this point of our reasoning, when we proceed from the data taken from Fig. 3.8, we
can only say that the experimental results correlate with our assumption that the
potential image-force barrier holds true for the distances of 3–4 Å from the
metal-vacuum boundary.
In addition to the energy spread, the electrons approaching the surface barrier
also have some angular dependence. The foregoing consideration assumes energy
E to be the energy connected with the momentum component orthogonal to the
surface. However, it is also necessary to consider the issue of how the results
obtained change with account of angular dependence. The internal discrepancy of
the situation comes from the fact that the surface barrier carries out a kind of energy
analysis by the momentum component orthogonal to the surface, while the spec-
trometer does that by the total energy. If the electron subsystem is slightly over-
heated, the above-described models hold adequate because the angular distribution
of electrons is narrow-beamed to the interface [58, 59].
If the electron subsystem is considerably overheated, i.e. if the electron energy
reaches several tenths of electron-volts, which is equivalent to several thousand
degrees, the particles can be located in an upper valley of the semiconductor. In
particular, when the gallium-arsenide electrons reach the energy of 0.35 eV, they
can quasi-thermalized in the L-valley. This process generates a compact group of
electrons having not only a narrow energy spectrum but also a narrow angular
localisation correlating with the position of the valley in the k-space in terms of the
crystal-lattice direction (111).
However, in this case, the changes of curves of energy distribution in the vac-
uum caused by variation of the work function are to have a prominent feature.
Indeed, let the probability of escape of this-group electrons at a work function is
B. Because of narrow localisation on the scale of energies, the energy interval
occupied by this group in the spectrum is far away from zero. It has already been
mentioned that for gallium arsenide, it is at least 0.35 eV. With decrease in the
work function, probability B goes up. The typical changes on the distribution curves
occur not at the energies of zero or close to it, but in the area separated from the
zero by the interval characteristic for a material.
This raises an extremely interesting question of preservation of the tangential
component of the electron momentum at transition through the semiconductor-
vacuum boundary. Simultaneous observance of the energy and momentum con-
servation laws imposes rather severe restrictions on the cone of electron escape in
the vacuum or the medium with a different effective mass in general (an analogue of
total internal reflection in optics). As this issue is topical in a lot of applications, it is
under lively discussion. Unfortunately, the measurements usually have the afore-
mentioned integrated nature, which is the reason why the results obtained are
sophisticatedly and often inadequately interpreted. What makes the issue even more
sophisticated is that a lot of semiconductor surfaces undergo faceting, which is why
the normal line to the macrosurface can be far from the normal line to the micro-
surface, i.e. the emitting surface. That makes the experimental angular dependences
of electron emission obtained by some authors untrustworthy, either. As a result,
while some authors claim that the results they have obtained can be well explained
subject to tangential momentum conservation law, others have to admit possibility
that it is violated, at least on the incoherent boundary, i.e. the boundary where the
long-range crystallographic order does not hold. Of course, this does not refute one
of the fundamental laws of nature. The point at issue is that when the incoherent
boundary is crossed there may be some exchange of momentum between the
electron and the crystal as a whole, which is perceived as an inconsistency of
particle momentum.
3.4 Phenomenological Theories of Field Emission

from Semiconductors
It has already been mentioned in Chap. 1 that the R. Stratton classical theory of
field emission from semiconductors is true in quantitative and qualitative terms only
as applied to field emission from n-type low-resistance semiconductors, which
follows from comparison of theoretical and experimental volt-ampere characteris-
tics. At the same time, field emission from n-type high-resistance semiconductors
and p-type semiconductors has a number of peculiarities [2, 32, 33, 44, 60–62].
Figure 3.9a shows the most complete experimental non-linear volt-ampere
characteristic of a semiconductor field emission cathode having eight characteristic
sections of change of field current with growth of the running voltage [60].
In this case, the basic features of field emission as compared to the similar
process for metals become clear through analysis of the presented volt-ampere
characteristic for field emission cathodes from p-Si (q = 3000 Ohm cm) [60], as
well as Figs. 3.10, 3.11, and 3.12. These features are listed below.
1. As a rule, the limiting current densities are much less than in case of emission
from metals;
2. The volt-ampere characteristics of lgI = f(l/U) are non-linear, i.e. they can show
the areas of saturation and proliferation of current;
3. In the saturation area, the emission current demonstrates thermo- and
photosensitivity;
4. The emission image in the saturation area tends to shrink [63], which is
indicative of a field-configuration change near the emitter surface (Fig. 3.10);
5. Pulsed emission excitation makes relaxation effects, i.e. the current-pulse shape
depends on the amplitude and duration time of anode voltage (Fig. 3.9b);
(a)
lg I 8
-5 6
-6 5
4 C
-7 3
2
-8
5а
1
-9 B
A
D
1 2 3 4 5 6 7 8 9 104/V, V-1
(b) V
τd
I, А
1
3
Increasing of the action pulse voltage
t,μs 8
Fig. 3.9 Semiconductor field emission current according to the cathode of the p-Si
(q = 3000 Ohm cm). a linear (A) and nonlinear (B—straightforward course, C—reverse)
current-voltage characteristics, taken at T = 300 K; 1–8 portions of presented characteristics; I–III
most commonly recorded portions; D curve plotted at T = 77 K; b the waveform of the current
pulses, registered under the action of the voltage rectangular pulses at the cathode and
corresponding to the points marked on the current-voltage characteristic of arabic numerals;
rectangular pulse duration is 200 ms
6. Energy distribution of escaping electrons broadens with increase in voltage on

the emitter [46], which typically indicates penetration of the electric field into
the sample, heating of electron gas (Fig. 3.11a) and increase in the portion of hot
emission-current electrons. The complex emission-electron spectrum structure
can also indicate that emission goes from different energy bands of the semi-
conductor or from the impurity and surface states (Fig. 3.11b);
3.4 Phenomenological Theories of Field Emission from Semiconductors 143
R1 /R2
0,8
-0,6
0,4
0,2
2 3 4 5 6 7 U, kV
Fig. 3.10 Illustration of the compression effect of the emission images at portions 2, 3 of the
current-voltage characteristics (see. Fig. 3.9a)
(a) (b)
1,0 0,3
f,r. u.
0,2
f,r. u.
0,5
0,1
4,2 4,6 5,0 Uз, V 19 21 23 Uз, V
Fig. 3.11 a energy distribution curve for germanium [68], b energy distribution curve for silicon
with two maxima. The distance between the peaks of 1.1 eV [69]
(a) (b)
N ×103,pulses
1,0
0,5
15 30 15 30 45 , keV
Fig. 3.12 Statistical field emission spectra for different portions of a nonlinear current-voltage
characteristics (see Fig. 3.9a—the curves B and D): 1–4 portions for (a) and the portion 5a for (b)
7. The experimental evidence of heating of the electron gas in the near-surface area
of the space charge of the p-type semiconductor cathode is also the field emission
statistic data from p-Si at the liquid-nitrogen temperature (curve D in Fig. 3.9a)
obtained in paper [62], which leads to the conclusion (Fig. 3.12), that the area of
fields and currents correlating to sect.1–4 in Fig. 3.9a, generates one-electron
emission acts and sections 5 generates many-electron emission acts;
8. It is also necessary to note that it follows from the experiment that electron
emission for the field ranges correlating with the volt-ampere dependence sec-
tions with a large slope (2, 3, 7 in Fig. 3.9a) weakly depends on temperature and
light, i.e. these conditions cause the semiconductor cathode to behave as a
quasi-metal one.
All these features result from relatively low concentration of electrons in the
conductivity band of semiconductors as compared with metals. The external electric
field applied to the sample to accelerate electrons distorts the energy bands of the
semiconductor and near its surface (cf. Chap. 1), shapes the screen layer of the
space electron charge. If penetrability of the potential barrier represented by
intra-crystal and external electric fields is high enough, electrons from the screen
layer tunnel into the vacuum. With growth of anode voltage due to increase in the
barrier penetrability, the tunnel emission-current density also grows up. In this part,
the pattern coincides with the one observed in metals in terms of quality.
But unlike metals, further increase in voltage and the electron-takeoff level makes
density of emission current from semiconductors depend not only on barrier pene-
trability, but also on the rate of electron inflow from the sample bulk to the surface.
For lack of multiplication of charge carriers in the internal crystal field, this flow is
limited. It is determined by the rate of their generation in the near-surface area and
stationary values of drift and diffusion. So the current density grows saturated, i.e.
the volt-ampere characteristic gets a characteristic step (Fig. 3.9a—section 4).
Further increase in anode voltage in the steady state can result in emergence of a
current-surge section (section III in Fig. 3.9a) preceding the explosive rupture of the
emitter. Many authors [46] usually attribute presence of this section to the begin-
ning of avalanche-type multiplication of current carriers in the near-surface space
charge area. When current is taken off in a quasi-steady state, e.g. when reading the
characteristic represented in Fig. 3.9a (reading time 1 s), a more sophisticated
pattern takes place in which an additional volt-ampere dependence section appears.
Analysis of the data represented in Figs. 3.9, 3.10. 3.11 and 3.12 allows the
authors of papers [43, 60–62, 64–67] to conclude that every section of volt-ampere
dependence of the p-type semiconductor cathode or the n-type high-resistance one
meets the classical Fowler–Nordheim linear form lgj = A – B/E. At that, the
electron concentration in the near-surface area of the space charge affects only value
A, and the slope of each section, i.e. value B, is determined by some effective work
function Фeff. So the first section of the volt-ampere characteristic (Fig. 3.9a) is
caused by tunnel electron emission from the surface states; the second one is caused
by tunnel emission from the valence band; the third one, i.e. the first saturation
section, is caused by tunnel emission of “hot” electrons from the conduction band
because of penetration of the electric field into the near-surface area of the space
charge, the rate of generation and recombination of carriers looking approximately
identical; the fourth section is caused by tunnel emission of “chilled” electrons,
wherein penetration of the field into the near-surface area of the space charge results
in the field “throwing” of electrons from the valence band to the conduction band,
i.e. increase in concentration of electrons, and consequent displacement of the field
from the near-surface area; the fifth one is caused by mixed emission: one part of
“hot” electrons penetrates the barrier though tunneling, while the other part goes
that over the barrier; in the sixth section, i.e. the second saturation section, all
electrons emit over the barrier; the seventh section may be caused by thermal
breakdown. It is worth noting that it is not always that thermal breakdown results in
destruction of the field emission cathode. Reduction of time to take the volt-ampere
characteristic causes transition to the eighth section, i.e. the third saturation section,
which is apparently caused by mere thermionic emission.
Elinson model. The above-presented qualitative, hypothetical pattern of field
emission from p-type semiconductors and n-type high-resistance semiconductors
has had no rigorous theoretical substantiation yet though the theoretical research
carried out later than Stratton’s papers had been published was particularly focused
on theoretical volt-ampere dependences similar to the experimental ones at least in
terms of quality.
In our opinion, the papers most physically grounded and mathematically correct
were those by Elinson et al. [2, 38, 65, 66], even though they aimed at solution of a
rather specific problem, i.e. an attempt to answer the question as to why the
experimental field emission characteristics in the area of heavy current densities are
usually steeper than those in the area of low current densities. The authors of the
papers specified explained this phenomenon by deflections from thermodynamic
equilibrium in energy distribution of electrons of the conduction band. These
deflections are more pronounced when internal electric fields are large. With growth
of the internal electric field, the average electron energy grows up, and the prob-
ability of their penetration through the surface potential barrier increased. Further
on, large internal electric fields can increase charge-carrier concentration due to
internal field emission, ionisation by collision etc. Theoretical consideration of field
emission with large current densities is rather difficult. Elinson et al. [65, 66]
attempted to solve this problem, and the way they selected seems to be most
adequate to the actual state of things.
The rudest approximation that could affect correctness of the theoretical con-
clusions was that spatial dependence of electron concentration near the semicon-
ductor surface was not taken into account. This dependence can result in the
respective spatial dependence of electron temperature. If penetration of the field
makes electron concentration in the surface area much higher than that inside the
crystal, it is quite reasonable that electron temperature on the surface is much less
than that calculated for the crystal bulk. This effect was taken into account in paper
[67]. Before featuring the findings of this paper, let us make some comments on the
most precise account of spatial dependence.
The primary goal is to evaluate the function of electron distribution.

Determination of this function is connected with simultaneous solution of
Boltzmann-Poisson equations with account of the emitter geometry. The collision
term of the Boltzmann equation depends on the scattering phenomena considered.
The band structure of the semiconductor is also essential. In addition, the distri-
bution function is a term of the equations for internal and external current density.
Further on, it is necessary to consider possible increase in electron concentration
due to the Frenkel effect [68], internal field emission from the impurity centers of
the valence band to the conduction band [69], ionization by collision [68–78] etc.
The above-mentioned basic equations are also appended with respective boundary
conditions and continuity conditions. The boundary condition for a dielectric flux
density on the semiconductor-vacuum boundary contains intensity of the external
electric field and charge density in the surface states.
It is obvious that solution of such a complex problem encounters great mathe-
matical difficulties.
Let us consider a semi-infinite semiconductor with spherical energy surfaces.
Let all donors be ionized and account be only taken of scattering on ultrasonic
phonons. The charges in the surface states are neglected. If we use the Boltzmann
distribution function for the first term of usual decomposition of the distribution
function over the Legendre polynomials,

nðxÞ E
j ðEÞ ¼ exp ð3:84Þ
½2pmkB T ðxÞ 3=2 kB T ðxÞ
with electron temperature T*(x), then we obtain the following system of equations:
dS 12mn S2
jE þ ndðT TÞ ¼ 0; ð3:85Þ
dx Tl2
dE 4pe0
¼ ðn n1 Þ; ð3:86Þ
dx j

dn n dT
j ¼ ne0 lE þ e0 d þ ; ð3:87Þ
dx 2T dx
where
1=2 1=2
2kB T dT T T
S¼ j 2kB d ; l ¼ l0 ; d ¼ d0 : ð3:88Þ
e0 dx T T
The values used in these equations stand for the following. S is the density of the
energy flow transferred by the electrons, T is the lattice temperature, E is the electric
field intensity, j is the electric-current density, l is the electron free path, n is the
electron concentration, d is the electron diffusion constant, l is electron mobility.
Application of the concept of electron temperature T* assumes presence of the
minimum electron concentration that provides strong mutual interaction between

conduction electrons.
The combined (3.84)–(3.87) should be solved with the following boundary
conditions
• at x = ∞
dT dE
¼ 0; ¼ 0; nð1Þ ¼ n1 ; ð3:89Þ
dx dx
• at x = 0
E ð0Þ ¼ E=k; ð3:90Þ
j ¼ je ; ð3:91Þ
Sð0Þ ¼ S00 : ð3:92Þ
Equations (3.90)–(3.92) express the conditions of electric-induction continuity,

current density and energy-flow density on the surface. Current density je in the
vacuum is equal to the sum of densities of thermoemission and field emission
currents. After introduction of dimensionless values,
x v n T E
n¼ ; g¼ ; m¼ ; s¼ ; x¼ ;
x0 kB T n1 T E0
where
1=2
kB T kB T
E0 ¼ ; x0 ¼ l ;
e0 x0 12mn s2
the equation for je becomes

2 3
Zg
je j T m0 4 g t
ie ¼ ¼ 1=2 s0 exp þ exp D ðtÞ dt5: ð3:93Þ
j0 j0 s s0 s0
0 0
Here

e0 n1 kB T 1=2
j0 ¼ e0 n1 l0 E0 ; jT ¼ ;
2 2mn
m0 ¼ mðx ¼ 0Þ;
s0 ¼ sðx ¼ 0Þ; x0 ¼ xðx ¼ 0Þ; ð3:94Þ
!
tðuÞ
D ðtÞ ¼ exp k1 1=2 :
x0 u3=2
The values to be met in transmission factor D are determined as follows.
dg E 1=2 b
u¼ ; t¼ ; dg ¼ ax0 ; k1 ¼ ;
wkB T t kB T a1=2
3=2 1=2
4c0 e0 2m e0 F0 j1=2
b¼ ; c20 ¼ ; a¼ :
3ðkB TÞ 2
h2 kB T
It follows from condition (3.92) that

ds
s00 ¼
dn 2 x¼0 3
Zg Zg
jT 4 g 1 t t
¼ g exp þ t exp D ðtÞ dt exp D ðtÞ dt5:
2j0 s0 s0 s0 s0
0 0
ð3:95Þ
The combined (3.85)–(3.87) were numerically integrated at l = 10−7 cm,

w = 1.2 eV, s = 5 105 cm/s, k = 5, n∞ = 1015 cm−3, T = 300 K and
E = (5 105 – 1 107) V/cm. The results are presented in Figs. 3.13, 3.14, 3.15
and 3.16.
Area a. We can see almost mere field emission without deflections from ther-
modynamic equilibrium (curve I in Fig. 3.14 at small x0, s 1 and curve III in
Fig. 3.15). This area has a strong spatial dependence of electron concentration and
internal-field intensity (curves I and II in Fig. 3.15). The external field is strongly
screened. The electron concentration on the semiconductor surface is several orders
of magnitude higher than n∞ (curve II in Fig. 3.14 at small x0).
Fig. 3.13 The dependence of 1 c

the current density on the field
strength with a significant ln ia
heating of the electron gas
and emissions above the
-1 b
surface potential barrier
-3
a
-5
-7
0,02 0,03 0,04 ω0-1
Fig. 3.14 The electron III II I

temperature (curve I), the
electron concentration (curve ln ν0' ln ν0 τ0
II) and the electron
temperature gradient (curve 2
III) on the surface of the III
emitter as a function of an 4
0
external electric field
75
-2 3
-4 50 I
2
-6
25
4
II
-8
0 0
25 50 75 100 ω0
Fig. 3.15 The spatial III I II

dependence of the electric τ ω ln ν
field (curve I), the electron
concentration (curve II), and 30
the electron temperature
(curve III) inside the 25
semiconductor at
ie = 7.14 10−5 4
20
III
1,000
3
15
10 2
5 1 II
I
0,900 0 0
2 4 6 8 10 ξ
Area b. In this area, the electron temperature as the external-field function surges
(curve I in Fig. 3.14). What happens is transition to thermoemission, which reduces
electron concentration on the semiconductor surface (curve II in Fig. 3.14).
The field emission characteristic is presented in Fig. 3.13. It clearly shows three
areas a, b and c.
Area c. The current almost reaches saturations. Almost all electrons are emitted
thermally. Electron concentration n = m0 n∞ can be even less than n∞; nevertheless,
electron temperature as the external-field function goes on growing (curve I in
Fig. 3.16 Same as in I III II

Fig. 3.15, but ie = 9.535 ω τ ν
90 III
II
85 2
I
90,600 80
1
90,590 75
90,580 70 0
1 2 3 4 ξ
Fig. 3.14). In this case, spatial dependence of the internal-field intensity is inap-
preciable (Fig. 3.16). Figure 3.16 shows that the enhanced electron concentration
on the surface causes reduction of electron temperature.
Under almost invariable conditions, increase in electron affinity w reduces cur-
rent density, thus causing smaller deflections from thermodynamic equilibrium.
Our opinion is that the theoretical current-voltage diagram (sections a, b, c,
Fig. 3.13) obtained by M.I.Elinson et al. at least qualitatively meets sections 6, 7
and 8 of the experimental current-voltage diagram (Fig. 3.9a).
Theory of A.F.Yatsenko
In his research Yatsenko [76] made an attempt to solve the problem of com-
putation of volt-ampere characteristics of semiconductor emitter that would corre-
spond with the most common observed volt-ampere characteristics (VAC)
consisting of three cardinal areas of change in the current with increase of voltage
(cf. Fig. 3.9a).
Additionally, one of the goals of his paper was to provide a theoretical expla-
nation of experimentally determined high photosensitivity of p-type semiconductor
samples under influence of fields corresponding to area 2 of VAC (cf. Fig. 3.9a).
Probably because of this Yatsenko named his theoretical paper “On modeling of
photo field emission out of p-type semiconductors”.
The object of modeling is lightly doped p-type semiconductor (i.e. the acceptor
concentration Na < 1019 cm−3) at relatively low temperature, but still with enough
heat so that small acceptors would be fully ionized. A flat field emission diode is
approximated with a metal-insulator-semiconductor (MIS) structure (Fig. 3.17) that
is different from the usual in that the insulator prat is represented by vacuum with field
strength (in working mode) up to 107 V/cm. Leaking of charge from the semicon-
ductor into the insulator is an unintended and often harmful phenomenon in most
MIS-devices, but in given case such leakage (i.e. emission) is the base of its function.
The area of volume charge of the emitter can be separated into three parts:
1. Base
2. Area of exhaustion
3. Inverse layer
Fig. 3.17 Structure of photo (a) 5 4 3 2 1

field emission diode (a) and
energy diagram of its cathode
[35] (b). 1 Base with an area
of diffusion with length l0; 2
Area of exhaustion with
length d; 3 Inverse layer. 4
Vacuum gap with length d; 5
Anode δ d
lD
(b)
ty
jes
jed Nd
For each part a problem of field, potential and current distribution is solved.
When considering the states of electrons in inverse layer one should take in
account the quantization of their movement along the field direction. Results of
those computations are presented in Table 3.3, from which one can see that due to
quantization the electron gas in inverse layer is not always degenerated.
The states of electrons of inversion layer were measured in effective mass
approximation.
The charge density in inversion layer is defined by level of balance
dnS nS
¼ gðd d0 Þ ; ð3:96Þ
dt se
Table 3.3 Emission anisotropy for Si and Ge

Silicon
Direction md/m0 m*/m0 gi N Filling of a level Field of degeneration
(E) 10−14 at Esg 107 V/cm
(cm−2 eV−1) Es = 107 V/cm (T = 300 K)
(001) 0.1905 0.9163 2 1.6 34.5 0.52
0.418 0.1905 4 7 7.9 2.27
(011) 0.324 0.3155 4 5.42 10.2 1.76
0.418 0.1905 2 3.5 15.8 1.1
(111) 0.358 0.2594 6 9 61.5 2.9
where se is the electron lifetime in inversion layer relative to emission into vacuum;
d is the depth of exhausted area; g is the total speed of generation of current carriers;
d0 is the Debye radius of screening.
The emission time se is by definition dependent on the potential barrier trans-
parency and thus on field strength.
In stationary case the emission current density can be written as:
enS E
je ¼ ¼ : ð3:97Þ
se 4pðse þ g=Na Þ
In areas with weak fields the emission currents increases with increase of field
with relation close to the Fowler–Nordheim law:
h i
je E 5=3 exp ðw1=2 U3=2 =w2 Þtðw=UÞ : ð3:98Þ
However, the slope of lg j = f(1/E) decreases with increase of the field.

In areas where the field is relatively strong, when
se Na =g; ð3:99Þ
dependence of current density on E follows Ohm’s law:
je ¼ gE=4pNa : ð3:100Þ
With decrease of field strength E, j(g) (e.g. lux-ampere) is saturated, dj/dg ! 0,

which qualitatively corresponds with the experiment (cf. Fig. 3.9a).
The latter relation should not break the inequality
ðmÞ
jðmÞ
e
eNa td ;
ðmÞ
where td is the limit drifting speed of the electrons.
From (3.97)–(3.100) one can see that it is only with sufficiently strong fields
when se = Na/g is emission density considerably increasing with increase of elec-
tron generation speed in the area of exhaustion:
Dje Dg: ð3:101Þ
In weak fields when se > Na/g, je(g) (lux-ampere) is saturated, Dje/Dg ! 0.

Obviously for observations in this models of photo field emission one needs rather
strong fields ensuring that the condition se sil is followed. However, as shown
by estimates for pure silicon with Na 1012–1013 cm−3, the penetration occurs
with fields Ek k(1.5–2)105 V/cm, which is an order of magnitude smaller than

those fields for which se sil . Conditions for observing of photosensitive field
emission (small time se compared to time of creation of inversion layer sil and lack
of penetration) are thus contradictory. In our opinion, this contradiction can be
alleviated by taking into account the process of electron exchange between surface
stated partially screening the volume of the sample from the external field and
valency areas, stimulated by strong electric field.
Consideration of this effect requires two additional parameters: time of capture
of zonal electrons onto the surface states ssr and time of emission out of surface
states ser. Stabilizing effect of surface states is only achieved given a particular
structure of their energy spectrum, in particular when the energy levels of those
states are situated approximately in the middle of forbidden area and their con-
centration Na 1013 cm−2.
Phenomenological theory of A.F. Yatsenko is of considerable interest both as a
new approach and as a theory that correctly qualitatively explains observed
experimental facts. However, if calculated according to the above-described rela-
tions, the emission characteristics differ from experimental data by orders of
magnitude.
Lvov–Baskin–Kaplan model
Physical and mathematical models proposed by theorists from the team of G.N.
Fursey [77, 78] were developed with the same goals as the theoretical studies of M.
I. Elinson et al. and A.F. Yatsenko described in previous sections, i.e. in order to
obtain theoretical VAC that would be similar to ones observed in experiments.
Before presenting the theoretical VAC obtained in those papers we need to note
that their authors (despite making a large number of mathematical assumptions
simplifying the calculation and to a greater degree leading to misrepresentation of
real processes) were only able to theoretically describe the transition from Area I to
Area II (“saturation”) of experimental VAC (cf. Fig. 3.9a). This is mostly due to the
fact that O.I. Lvov et al. were not considering the possibility of multiplication of
carriers because of strong electric field entering the semiconductor, which is prob-
ably the reason of emergence of Area III (rapid grows of current) on experimental
VAC (cf. Fig. 3.9a). There are other approximations as well. For an instance, as
opposed to work by A.F. Yatsenko, they neglect the influence of surface states. It is
also assumed that the quasi levels for electrons and gaps are the same. In reality this
does not necessarily hold. Even the very use of Fermi-Dirac distribution function as
it is used in those papers is not completely legitimate or at least requires additional
justification. Additionally, which is the most important in our opinion, Lvov et al.
formulated and solved the problem using an one-dimensional model that is not
correct with strong influence of field on the semiconductor and cannot be used for
quantitative computations of VAC of real-life emitter.
Due to the above-mentioned facts the Lvov-Baskin-Kaplan model has to be
considered an approximate theory of “saturation effect” during field emission from
semiconductors:
Theoretical VAC calculated in [77, 78] for p-type Ge is shown on Fig. 3.18. The
two curves on this figure correspond to different equations for motility presented in
the caption. The authors of [77, 78] pointed out that the experimental values of
saturation current must be greater that ones calculated due to conical geometry of
real-life emitter which leads to increase of saturation current density up to one or
two orders of magnitude as compared to theoretical estimates (Fig. 3.18). The
following relationship between mobility and the field strength inside the emitter are
assumed:
8
< ln ð0Þ;
> E\102 V/cm;
curve 1 : ln ðEÞ ¼ ln ð0Þð102 =EÞ1=2 ; 102 \E\104 V/cm;
>
:
l ð0Þ 103 E; E 104 V/cm:
( n
ln ð0Þ; E\102 V/cm;
curve 2 : ln ðEÞ ¼
ln ð0Þð102 =EÞ1=2 ; E [ 102 V/cm:
Figure 3.19 presents the results of calculation of length l (defined as the distance
between points where the concentration is twice the minimum value) of an area of
field emission cathode exhausted by carriers depending on E0, performed in those
studies. Lvov et al. also noted that increase of l with growth of E0 in stable current
only emerges due to field entering the semiconductor. Let’s remark that the cal-
culated depth of penetration of the field into the semiconductor reaching 1 mm at
the saturation area is large relative to the tip radius. This means that the
one-dimensional theory would not work very well for quantitative estimates
whenever current saturation occurs. This theory overestimates both length of
exhausted area and length of area of field penetration. One has to note yet another
result of authors of those papers. They found that the concentration of electrons in
the zone of permissibility first slowly increased with increase of E0, and then is
almost constant in a rather wide range of E0, and finally drops abruptly with further
Fig. 3.18 The dependence of j, А/cm2

the emission current density
of the external field for Ge 2
p-type Na = 10l5 cm−3,
T = 300 K 102
101
100
2,2 2,6 3,0 3,4 Е0-1, Å/V
Fig. 3.19 Length of l, cm

exhausted area depending on
external field for p-type Ge 10-1
with Na = 1015 cm−3,
T = 300 K
10-2
10-3
10-4
10-5
0 3,4 3,8 4,2 Е0, 107 V/cm
increase of the field, when the emission current is saturated and becomes practically
stable.
Results similar to ones presented on Fig. 3.19 were also obtained for n-type
silicon. In this case, unlike p-type semiconductors, the saturation current density
was significantly dependent on the concentration of impurities.
3.5 Theoretical Aspects of Phenomena and Processes

on the Surface During Field Emission
Emission properties of field emission cathodes are determined by their geometrical

characteristics and their surface state. Understanding of the regularities affecting
stability and withstandability of cathodes in the field emission mode can come from
studying of the fast processes on the surface, in particular, the migratory processes
taking place when the emitter is affected by a strong electric field, which changes
the surface microgeometry. It is difficult to research the phenomena on the surface
because the surface of the field emission cathode develops extreme conditions:
strong electric fields, considerable current densities, high temperatures etc. The
theoretical problems of this research were covered to the fullest extent possible in
the survey chapter of the monography by Zhukov [79], published in 2007, and the
review by Sokolskaya [80], which has already become a rarity.
It is known that every crystal tends to take the equilibrium shape, when the free
surface energy is minimum for this bulk [81].
The transfer phenomena changing the shape of the heated crystals of pure metals
were theoretically studied by Hering [82, 83] with application of the principles of
thermodynamics. The equations featuring variations in the shape as a result of
surface and bulk diffusion, evaporation-condensation and viscous flow for the
isotropic crystal was solved ad hoc by Mullins [84, 85].
The Mallins equations also give a criterion to evaluate the role of each process in
particular cases. As shown in paper [86], the basic transfer process in field emission
emitters is surface diffusion.
According to Hering, mass transfer by diffusion can only take place if accom-
panied by reduction of the total free energy, which necessitates a gradient of
chemical potential. For a curved surface, Hering derived an expression linking the
difference between chemical potentials of atoms and gaps with the free surface
energy:

@2a @2a
l lh ¼ l0 þ X0 R1 a þ þ R1
a þ P xx ; ð3:102Þ
1
@n21 2
@n22
where µ − µh is the difference of chemical potentials of atoms and vacancies for a

curved surface; µ0 is the chemical potential for a flat surface; R1 and R2 are the main
curvature radii of the curved surface; a is the free surface energy; n1 and n2 are the
directions of normals to the surface in directions R1 and R2; X0 is the bulk per atom;
Pxx is the normal component of the external mechanical stress acting on the surface.
Assuming that the surface-position change rate is proportional to the gradient of
chemical potential, Hering derived an expression determining the flux of atoms per
unit time through the unit-length line perpendicular to the direction of migration for
an arbitrary point of the surface
D0 eQ=kT
JM ¼ ðrlÞM ; ð3:103Þ
A0 kT
where (∇µ)M = ∇(µ − µh) is the gradient of difference of chemical potentials of

atoms and vacancies in point M from (3.102); D0 is the stationary value of the
surface self-diffusion; Q is the process activation energy; T is the absolute tem-
perature of the crystal; A0 is the area per atom.
When the apex gets blunted, the flux of atoms from the vertex to the base travels
over the surface having numerous and different faces. That is why values a for these
faces are averaged and the process is assumed to be isotropic [29], which makes it
possible to neglect terms ∂2a/∂n21 and ∂2a/∂n22 in (3.102). In addition, it is assumed
that R1 = R2 = rc because the apex is usually a figure of revolution. Then the flux of
atoms is expressed as

X20 D0 a cm3
JM ¼ exp ðQ=kTÞrM Pxx ; : ð3:104Þ
A0 kT rc cm s
3.5 Theoretical Aspects of Phenomena and Processes … 157
Consequently, heating in absence of external actions (Pxx = 0) makes the flux of

atoms move from the vertex to the base, which results in shortening of the apex and
increase in its radius. Thereby, free energy decreases (Fig. 3.20).
Increase in the apex radius during the heating was experimentally observed for
the first time by Mueller [87]. It is theoretical and experimental research of this
phenomenon that his paper [29] is dedicated to. It shows that apex-length reduction
rate dz/dt exceeds the radius-increase rate, whereby

dz aX2 D0 C
¼ exp ðQ=kTÞ 3 ; ð3:105Þ
dt Pxx ¼0 A0 kT rc
where C is a dimensionless constant dependent on the apex shape. (3.105)

demonstrates that the apices that get blunted by heating most of all are those with a
small radius (rc < 10−5 cm). Increase in the radius decelerates this process abruptly
[88].
Fig. 3.20 Changing the dz

shape of the tip on heating 1 dV
without the field [29]. 1 and 2
—a sequence of locations of
its surface 2 rM M P
θ
JM
r
A more rigorous theory of heated apex-shape variation as a result of surface

migration was developed by Nichols and Mullins [89], who used numerical
methods to derive dependences of the emitter radius rc(t) and bias of its apex z(t) on
the heating time for different cone-semiangle values at vertex h = a/2. It turned out
that at 0
h
3 °C the apex becomes spheroid (Fig. 3.21a), and a constriction
develops up to h = 8°. The apex profiles for different values of h given in
Fig. 3.21b.
The phenomenological theory by Hering does not consider the surface
self-diffusion mechanism. That is why the process parameters, in particular, acti-
vation energy Q, are introduced formally and disconnected from the real energy and
geometrical structure of the crystal surface.
Effect of the atomic structure of the crystal surface on the surface self-diffusion
processes was taken into account by Frenkel [90]. According to Y.I. Frenkel,
thermodynamic equilibrium causes the crystal surface to be formed not only of
ideal faces with low indices (singular faces), but also of vicinal faces, which can be
regarded as stage terraces developed on the basis of densely packed faces. Singular
surfaces near the absolute-zero temperature form an ideal flat boundary between the
steam and the crystal. When the temperature goes up, such a surface develops
equilibrium defects such as adsorbed atoms and surface vacancies, which expo-
nentially increase in concentration with increase in temperature. So the enhanced
temperature makes the equilibrium crystal faces naturally rough. According to
paper [90], the crystal shape changes due to elongation or shortening of steps as a
result of consecutive attachment or detachment of atoms. The atoms taking part in
this process are not embedded in the regular planes, but located on the edges of the
steps, i.e. in the linear-gas state, or on the plane itself, i.e. in the two-dimensional-
gas state. Support of the diffusion flow making the crystal change its shape
necessitates creation of a particular concentration of planarly and linearly adsorbed
atoms. Obviously, that involves input of the energy Qbind equal to the difference
(a) (b)
Y α=30°
16
10 30
12
10 20
8 15
Y
6
20 10
20695 4 8
10 0,8 2525 10454 2 5
4
3
0 10 20 30 40 50 60 70 80 90 z 0 2 4 6 8 10 12 14 16 z
Fig. 3.21 Sharp profiles obtained by numerical calculation: a a/2 = 2—terminus spheroidization
occurs, b stationary forms with a/2 3°
between the potential energies of the atom embedded into the edge of the flat face
and the planarly adsorbed atom. In the two-dimensional gas state, atoms can only
move upon energization with Qm, which is determined by the difference between
the two potential minima for this face. Consequently, surface self-diffusion will be
an activated process with the activation energy of Q = Qbind + Qm, and the
mass-transfer rate proportional to diffusion constant DS will be determined by the
following expression:

Qbind þ Qm
v DS ¼ D0 exp : ð3:106Þ
kT
It should be expected that experimentally found activation energy Q is actually

the total of some average binding energy Qbind and the energy of migration aver-
aged for different faces Qm .
The diffusion processes taking place when an apex is heated in a strong (E (1–5)
107 V/cm) electric field are of an essentially different nature than those developing
beyond the field. That is accompanied by changes of the emission image, namely
extension of the area of densely packed faces and increase in brightness at their
interfaces. The rounded apex vertex assumes a polyhedron shape. Apparently,
Mueller [91] was the first to observe this phenomenon. Later, it was fundamentally
studied by Benjamin and Jenkins [83], who named it building-up [92].
The built-up stages were subdivided by the faceting depth or extent. It was noted
that the process is independent of the field sign. The quantitative research of the
tungsten apex building-up process in the decelerating electric field was carried out
by Sokolskaya [93].
The theoretical description of building-up is based on the theory suggested by
Hering and developed in his papers [82, 83]. According to Hering, the external
electric field exerts a ponderomotive pressure upon the surface. A normal com-
ponent of the pressure Pxx = E2/8p [cf. expression (3.104)]. Paper [83] derives the
expression determining the shortening rate of the apex heated in the electric field:

dz dz 1 rK E2
¼ 1 ; ð3:107Þ
dt E dt 0 C0 16pa
where C0 0,5 is a unitless coefficient, (dz/dt)0 is the shortening rate beyond the
field determined according to (3.105). From (3.104) one can see that the apex
blunting rate is determined by the ratio of electrostatic pressure and the
surface-tension forces.
The direction of the electric field applied does not matter because of quadratic
dependence. Expression (3.107) makes it possible to determine the field intensity at
which (dz/dt)E = 0 and the vertex point reaches equilibrium
1=2
8pa
E01 ffi : ð3:108Þ
rc
In other surface points, the migratory processes still result in diffusion of atoms
to the periphery, so equilibrium is labile. The field intensity providing equilibrium
in the local area of the hemisphere-shaped apex vertex is determined by the fol-
lowing ratio:
1=2
16pa
E02 ffi : ð3:109Þ
rc
It can be expected that if E > E2, the migratory fluxes will be re-directed.
According to [82], that provides for the processes of ordinary or complete reshaping
based on delivery of the substance to the vertex. I.L. Sokolskaia figured that this
effect can be of secondary importance [80]. In her opinion, an important part in
diffusion in the field is played by the atoms of the plane-adsorbed phase abundant
on the surface of the heated rounded apex. The faces are completed and the faceted
equilibrium shape is reached by the reason that the field pulls the polarized atoms to
the face edges, where the field is the most intense and where they integrate into the
crystal plane (Fig. 3.22).
At that, I.L. Sokolskaia definitely figured that the mentioned bright spots meet
the qualitatively new objects, so-called “microelevations”, which grow on the
emitter surface affected by the threshold field of 1:6 10 V=cm. As the
microelevations easily disintegrated when heated beyond the field, I.L. Sokolskaia
considered them structureless, a “chaotic pile of atoms”. However, I.L. Sokolskaia
also assumed that there are certain conditions, including relatively low temperatures
and the fields that are high but insufficient for desorption by the field, under which it
is possible to grow up irregular bumps resistant to heating at low temperatures
beyond the field.
Fig. 3.22 The proposed F

mechanism of completion of
flat faces in the restructuring
M
It was shown earlier that for the apices subjected to critical reshaping, the field
emission current can be more than doubled and followed by the preexplosive
phenomena, in particular, a spontaneous current growth.
The effect observed can be explained on the basis of the solution of the
thermal-conductivity equation with account of the Nottingham effect [94, 95].
It is possible to consider a non-reshaped rounded emitter as an analogue of the
cylinder (Fig. 3.23a). The problem of heating of such an emitter by the flowing
current was considered in paper [25].
At that, the losses of heat due to radiation or the Nottingham effect were
neglected. Neither was the dependence of physical constants of the emitter material
on temperature taken into account. Then the conductivity equation [25] is given by
@ 2 T mc @T j2 q
¼ ; ð3:110Þ
@l2 k @t k
where m is the emitter-material density, c is the specific hit, k is thermal conduc-

tivity, j = I/pr2 is the field emission current density, q is the specific electrical
resistance. The stationary solution of (3.110) is the expression
T ¼ 0:5al2 þ C 0 l þ C00 ; ð3:111Þ
where a = j2 q/k, C′ and C″ are the integration constants determined from boundary
conditions
Tjl¼0 ¼ 0; ð3:112Þ

@T
¼ 0: ð3:113Þ
@l l¼L
(a) (b) (c)

l 2r l l
j1 : T1
2 1
h h h
2(r-∆)
∆<<r
L j:T L L
j:T
I I
T/Tmax
Fig. 3.23 Profiles of “smoothed” (a) and the “rearranged” (b) emitter and the temperature
distribution (c) at a steady current
Condition (3.112) means that the emitter base is supported at a particular tem-
perature that is assumed to be zero. Condition (3.113) means that there is no heat
flowing through the emitter vertex at l = L (Fig. 3.23a). Under the boundary con-
ditions, the temperature peaks at the emitting vertex [25]:
T max ¼ 0:5aL2 : ð3:114Þ
It has already been noted above that at high temperatures, the Nottingham effect
results in cooling of the emitter vertex. At that, the peak temperature area biases
deep into the apex (Fig. 3.24). Vertex cooling is equivalent to the heat flux through
it.
To take this circumstance into account, we will demand fulfillment of the fol-
lowing condition instead of (3.113):

@T
¼ P; ð3:115Þ
@l l¼L
where П > 0. Using conditions (3.112)–(3.115), let us find constants C′ and C″ of

expression (3.111), C 0 ¼ al P, C″ = 0. We also think that the peak-temperature
area biased deep into the emitter to the distance of h from the vertex (Fig. 3.23a).
But then the following condition is true:

@T
¼ 0:
@l l¼Lh
Fig. 3.24 The temperature T max = 0. 5aL2 .

distribution along the emitter
taking into account various T, K
factors: 1 Nottingham effect
3000
and Joule heating;
2 Nottingham effect;
3 Joule heating
2000 1
1000
2
1 2 3 4
r ×104, cm
Using it, we can find that
P ¼ ah: ð3:116Þ
As a result, we obtain an expression that qualitatively features distribution of

temperature in the emitter for a stationary case with account of the Nottingham
effect
T ¼ 0:5al2 þ a ðL hÞ l: ð3:117Þ
The emitter profile is given in Fig. 3.23a. The qualitative temperature distribu-
tion graph is shown in Fig. 3.23c, curve 1 (cf. with Fig. 3.24).
The emitting vertex temperature is determined by the following expression:
ð3:118Þ
A reshaped emitter is characterized by pitting on the periphery. A model of such

an emitter is given in Fig. 3.23, b. In this case, heating is featured by joint solution
of stationary (3.110), which is true at 0
l
L–h, and the following equation
@ 2 T1 j21 q
¼ ; ð3:119Þ
@l2 k
where (cf. Fig. 3.23b).

I
j1 ¼ ; L h
l
L
pðr DÞ2
At that, it is assumed that field emission current I goes through the vertex of the
cylindrical bump. Strictly speaking, such problem cannot be treated as
one-dimensional, but we neglect this circumstance for qualitative evaluations.
For the combined (3.110) (given that ∂T/∂l = 0) and (3.119), we demand ful-
fillment of the following boundary conditions:
Tjl¼0 ¼ 0; ð3:120Þ

@T
¼ 0; ð3:121Þ
@l l¼Lh
Tjl¼Lh ¼ T1 jl¼Lh ; ð3:122Þ

@T1
¼ a1 h: ð3:123Þ
@l l¼L
Condition (3.120) is similar to condition (3.112). Condition (3.121) assumes that

the peak-temperature area goes on being biased deep into the emitter to the distance
of h. Condition (3.122) follows from temperature-distribution continuity in the
emitter. Condition (3.123) is caused by account of the Nottingham effect and for-
mulated by analogy with the above-stated.
Solution of (3.110) meeting conditions (3.120) and (3.121) has already been
found. It is expression (3.117). Generally, solution of equation (3.119) can be
represented as follows
T1 ¼ 0:5a1 l2 þ C10 l þ C100 ; ð3:124Þ
where
a1 ¼ j21 q=k;
and C1′ C1″ are constant.

Using conditions (3.122) and (3.123), we can find that
C10 ¼ a1 ðL hÞ; C100 ¼ 0; 5ða þ a1 ÞðL hÞ2 a1 ðL hÞ2 ;

ð3:125Þ
T1 ¼ 0; 5a1 l2 þ a1 ðL hÞl 0; 5ða1 aÞðL hÞ2 :
Let us test if conditions (3.122) and (3.123) hold:
Tjl¼Lh ¼ T1 jl¼Lh ¼ 0; 5ðL hÞ2 ;

@T1
¼ a1 h:
@l l¼L
Additionally, it turns out that

@T1 @T
¼0¼ ;
@l l¼Lh @l l¼Lh
i.e. functions T and T1, as well as their derivatives with respect to L, are continuous.
The qualitative graph of distribution of temperature T1 is shown in Fig. 3.23c, curve
2.
If l = L, expression (3.125) gives us the temperature of reshaped emitter apex:
T1a ¼ a1 Lð0; 5L hÞ 0; 5ða1 aÞðL hÞ2 : ð3:126Þ
Having divided expression (3.126) by (3.118) and made respective transfor-

mations, we can find that
T1a h2 a
1
¼1 2 1 \1: ð3:127Þ
Ta L 2Lh a
Ta Ta
given h ! 0 T1 ! 1; h ! L T1 ! aa1 , which has to be the case.
a a
Figure 3.23c, as well as expression (3.127) imply that there are particular modes
when a set full current I supplied to the reshaped field emission cathode apex heats
it less than the filtered current. So the preexplosive effects will manifest themselves
at heavy field emission currents, which is the case in the experiment.
Qualitative reduction of heating of the pitted emitter vertex can be explained as
follows. The field emission current density through the apex has somewhat
increased. So Joule heat can be expected to heat the near-apex area up to a higher
temperature. However, increase in current density intensifies cooling due to the
Nottingham effect. If the apex temperature is higher than the inversion temperature,
this process prevails over heating, which gives this effect.Obviously, the following
condition is to be met:
Qout
1; ð3:128Þ
Qrel
I 2 qsp
where Qrel ¼ p2 r 4 is the volumetric density of the energy released in the near-apex
WIs
area of the emitter (Fig. 3.23a) within the time of flowing of current I; Qout ¼ epr2ph is
the volumetric density of the energy carried away from the near-apex area. W is the
energy carried away during the emission of one electron, e is the electron charge.
Ratio (3.128) can be reduced to
Wpr 2
1: ð3:129Þ
eqhI
The volumetric density of the energy released and carried away from the
near-apex area for the reshaped cathode (Fig. 3.23b) can be determined from the
following expressions:
I 2 qsp WIsp
Q0rel ¼ 4
; Q0out ¼ :
p2 ðr DÞ epðr DÞ2 h
The reshaped cathode functioning in the mode considered is to meet the

condition
Qout
[ 1;
Qrel
where DQrel ¼ Q0rel Qrel ; DQout ¼ Q0out Qout .After transformation we obtain
DQout Wp 1
¼ :
DQrel ehIq ðr DÞ2 þ r 2
As D r,
DQout Wpr 2
ffi 1: ð3:130Þ
DQrel 2ehIq
Comparison of expressions (3.129) and (3.130) shows that the field emission
current making the prevailing Nottingham effect for a reshaped cathode can be fully
twice as high as the current for a rounded emitter. This result is in good agreement
with the experiment.
It is worth adding that the model considered does not give a full description of
non-stationary heating processes of reshaped field emission cathodes. Nevertheless,
its approximate nature does not prevent it from qualitatively and conditionally even
quantitatively correct description of behavior of a reshaped apex cathode as its
vertex is heated up less.
References
1. I.M Hoffmann, Investigation of electrostatic emission of tungsten in a wide range of current

densities. FTT 4, 2005 (1962)
2. M.I. Elinson, F.F. Dobryakova, V.F. Krapivin, On the theory of field and thermionic emission
of metals and semiconductors. Radiotehnika i Electronika. 6(8), 1342–1353 (1961)
3. M. Green (ed.), Surface properties of solids. Mir 432 (1972)
4. L. Wei, W. Baoping, G. Li, Y. Hanchun, T. Yan, Analysis of the emission performance of
field emitter with Laplace interpolation method. Appl. Surf. Sci., 161, 1–8 (2000)
5. A.G.J. Oostrom, Philips Res. Rep. Suppl. 1, 1–162 (1966)
6. G.R. Condon, J.A. Panitz, J. Vac. Sci. Technol. B 16, 23 (1998)
7. J. Plšek, D.V. Zhukov, Z. Knor, The average work function and emission area in the Fowler–
Nordheim equation. Czech. J. Phys
8. G. Binnig, G. Rohrer, Scanning tunnel microscope. World Sci. 10, 26–33 (1985)
9. A. Modinos, Theoretical analysis of field emission data. Solid-State Electron. 45, 809–816
(2001)
10. R.G. Forbes, K.L. Jensen, New results in the theory of Fowler–Nordheim plots and the
modelling of hemi-ellipsoidal emitters. Ultramicroscopy 89, 17 (2001)
11. M. Drechsler, E. Henkel, Z. Angew. Phys. 6, 341 (1954)
12. P.H. Cutler, J. He, N.M. Miskovsky et al. J. Vac. Sci. Technol. B. 11, 387–391 (1993)
13. A. Modinos, J.P. Xanthahis, Energy-broadening of field-emitted electrons due to Coulomb
collisions. Surf. Sci. 249, 373 (1991)
14. Y. Suganuma, M. Tomitori, Analysis of electron standing waves in a vacuum gap of scanning
tunneling microscopy: Measurement of band bending through energy shifts of electron
standing wave. J. Vac. Sci. Technol. B 18(1), 48–54 (2000)
References 167
15. L.N. Kantorovich, A.S. Foster, A.L. Shluger, A.M. Stoneham, Role of image forces in
non-contact scanning force microscope images of ionic surfaces. Surf. Sci. 445, 283–299
(2000)
16. N.M. Miskovsky, S.H. Park, J. He, P.H. Cutler, Energy exchange processes in field emission
from atomically sharp metallic emitters. J. Vac. Sci. Technol. B 11(2), 366–371 (1993)
17. S. Georgieva, D. Vichev, K. Drandarov, Computer simulation of the emission process of
some field emission alloy ion sources. Vacuum 47(10), 1143–1144 (1996)
18. N.V. Egorov, E.M. Vinogradova, Solution of boundary-value problem in bispherical
coordinates. in Proceedings of 3-th International Workshop: BDO-96, St. Petersburg,
pp. 274–278 (1996)
19. G. Mesa, E. Dobado-Fuentes, J.J. Saenz, Image charge method for electrostatic calculations in
field emission diodes. J. Applied Phys. 79(1), 39–43 (1996)
20. Y. Ohkavara, T. Naijo, T. Washio, S. Oshio, H. Ito, H. Saitoh. Field emission properties of
AlZnO whiskers modified by amorphous carbon and related films. Jpn. J. Appl. Phys. 40(12),
7013–7017 (2001)
21. K.L. Jensen, J.E. Yater, Advanced emitters for next generation rf amplifiers. J. Vac. Sci.
Technol. B, 16(4), 2038–2049 (1998)
22. M.S. Yermoshina, Mathematical modeling of electron emission from a point cathode of
complex configuration. Kand. Diss. SPb. 112 p (2004)
23. R.G. Forbes, C.J. Edgcombe, U. Valdre, Some comments on models for field enhancement.
Ultramicroscopy 95, 57–65 (2003)
24. P.J. Birdseye, D.A. Smith, G.D.W. Smith, Analogue investigation of electric field distribution
and ion trajectories in the field ion microscope. J. Phys. D: Appl. Phys. 7 (1974)
25. W.W. Dolan, W.P. Dyke, J.K. Trolan, The field emission initiated vacuum area. II. The
resistively heated emitter. Phys. Rev. 91(5), 1054–1057 (1953)
26. J.R. Barbour, W.W. Dolan, J.K. Trolan et al., Space-charge effects in field emission. Phys.
Rev. 92(1), 45–51 (1953)
27. J.K. Trolan, J.R. Barbour, E.E. Martin, W.P. Dyke, Electron emission from a lattice step on
clean tungsten. Phys. Rev. 100(6), 1646–1649 (1955)
28. W.P. Dyke, F.M. Charbonnier, R.W. Straer et al., Electrical stability and life of the heated
field emission cathode. J. Appl. Phys. 31(5), 790–805 (1960)
29. J.R. Barbour, F.M. Charbonnier, W.W. Dolan et al., Determination of surface tension and
surface migration constants for tungsten. Phys. Rev. 117(6), 1452–1459 (1960)
30. F.M. Charbonnier, J.R. Barbour, L.F. Garret, W.P. Dyke, Basic and applied studies of field
emission at microwave frequencies. Proc. IEEE 51(7), 991–1004 (1963)
31. W.P. Dyke, Field emission. A new practical electron source. IRE Trans. Mil. Electron. 38–45
(1960)
32. A. Modinos, Field, thermionic and secondary electron emission spectroscopy. M. Nauka, 320
(1990)
33. M.I. Elinson (ed.), The cold cathodes. Sov. Radio 386 (1974)
34. W.W. Dolan, W.P. Dyke, Temperature and field emission of electrons from metals. Phys.
Rev. 95, 327–332 (1954)
35. E.L. Murphy, R.H. Good, Thermionic emission, field emission and transition region. Phys.
Rev. 102(6), 1464–1473 (1956)
36. R.H. Good, E.W. Mueller, Field emission. In Handbuch der Physik (ed. By S. Flugge), Bd.
21, 176–231 (Springer, Berlin, 1956)
37. E. Guth, C.J. Mullin, Electron emission of metals in electric field. Phys. Rev. 61(5–6), 339–
348 (1942)
38. M.I. Elinson, The emission of electrons under the influence of strong electric fields. Dokt.
Diss., L: LPI n. MI Kalinin (1961)
39. S.G. Christov, General theory of electron emission from metals. Phys. Stat. Sol. 17(1), 11–26
(1966)
40. F.I. Itskovich, On the theory of field emission from metals. Part I. ZETP, 50(5), 1425–1437
(1966)
41. F.I. Itskovich, On the theory of field emission from metals. Part II. ZETP 52(6), 1720–1735
(1967)
42. I.S. Andreev, The study of electron emission from the metal in its transition from cold to
thermionic. J. Tech. Phys. 22, 1428–1441 (1952)
43. S.C. Miller Jr., R.H. Good Jr., Phys. Rev. 92, 13–67 (1953)
44. M.I. Elinson, G.F. Vasiliev, Field emission. M. Fizmatgiz, 272 (1958)
45. L.N. Dobretsov, M.V. Gomoyunova, Emission elekcronics. M. Nauka, 364 (1964)
46. R. Fischer, H. Neumann, Field emission from Semiconductors. M. Nauka, 215 (1971)
47. V.T. Cherepin, M.A. Vasiliev, Methods and tools for the analysis of material surface
Handbook (Naukova Dumka, Kiev, 1982), p. 400
48. J.E. Henderson, R.E. Badley, Phys. Rev. 38(3), 590 (1931)
49. J.E. Henderson, R.K. Dahlstrom, Phys. Rev. 55(5), 473 (1939)
50. J.E. Henderson, R.K. Dahlstrom, F.R. Abott, Phys. Rev. 41(1), 261 (1932)
51. L. Feldman, D. Meyer, Basics of analysis of surfaces and thin films. M. Mir, 344 (1989)
52. J.D. Caret, B. Feyrbah, B. Heaton et al. The use of electron spectroscopy for surface analysis.
H. Ubach (ed.). Riga Zinatne, 315 (1980)
53. D.A. Orlov, M. Hoppe, U. Weigel, D. Schwalm, A.S. Terechov, A. Wolf, Energy distribution
of electrons emitted from GaAs (Cs, O). Appl. Phys. Lett. 78, 2721 (2001)
54. L.W. Swanson, L.C. Crouser, Phys. Rev. 163, 632 (1967)
55. A.B. El-Karen, J.C. Wolfe, J.E. Wolfe, J. Appl. Phys. 48, 4749 (1977)
56. J.W. Gadzuk, E.W. Plummer. Phys. Rev. 3, 2125.58 (1971)
57. A.E. Bell, L.W. Swanson, Phys. Rev. B. 19, 3353 (1979)
58. V.A. Korablev, Y. Kudinov, MSh Sugainov, T.A. Baranova, Photoelectron spectroscopy with
angular resolution of GaAs with a negative electron affinity. Raditehnika i Elektronika. 32,
321 (1992)
59. G. Vergara, A. Herrera-Gomez, W.E. Spicer, Electron transverse energy distribution in GaAs
negative electron affinity cathodes: Calculation compared to experiment. J. Appl. Phys. 80,
1809 (1966)
60. N.V. Egorov, V.R. Tolstyakov, Investigation of the effect of the surface state on the emission
characteristics of semiconductor photo field cathodes. Surface 8, 23–33 (1996)
61. D.A. Ovsyannikov, N.V. Egorov, Mathematical modeling of systems for formation of
electron and ion beams (Publishing of St. Petersburg State University, Russia, 1998). 276 p
62. N.V. Egorov, V.R. Tolstyakov, The effect of multi-particle tunneling in the field electron
emission from semiconductors. Povrhnoct’ 9, 10–13 (1996)
63. N.V. Egorov, On the possibility of narrow collimated electron beams. ZTP. 52(12), 2440–
2442 (1982)
64. N.V. Egorov, A.G. Karpov, Diagnostic information and expert systems. SPb. (St. Petersburg
State University Publishing House, Russia, 2002) 472 p
65. M.I. Elinson, Effect of internal electric fields in the semiconductor at its field emission.
Raditehnika i Elektronika. 4,140–142 (1959)
66. A.G. Zhdan, M.I. Elinson et al., Raditehnika i Elektronika. 7, 570 (1962)
67. M.I. Elinson et al., Raditehnika i Elektronika. 10, 1288 (1965)
68. Y.A. Frenkel, ZETP. 8, 1893 (1938)
69. W. Franz, Ergeb. exapt, Naturwiss. 27, 1 (1953)
70. W. Franz, Handb. Phys. 17, 155 (1956)
71. R.J. Hodgkinson, Proc. Phys. Soc. 82(58), 1010 (1963)
72. Y.A. Frankel. ZETP. 7, 1069 (1937)
73. R. Stratton, Phis. Rev. 126, 2002 (1962)
74. M. Sanchez, Helv. Phys. Acta. 36, 1 (1963)
75. H. Frohlich, B.V. Paranjape, Proc. Phys. Soc. 69(21), 866 (1956)
76. A.F. Yatsenko, On a model photo-field emission from p-type semiconductors. Phys. Stat.
Solid. 1(2), 169–175 (1970)
77. G.N. Fursei, M. Kaplan, O.I. L’vov, On the theory of field emission from semiconductor of
p-type. Vestnik Leningrad. Universiteta. Ser. Fisiki I Himii. 16, 167–170 (1968)
References 169
78. L.M. Baskin, O.I. Lvov, G.N. Fursey, Generation features of field emission from
semiconductors. Phys. Stat. Sol. B. 47, 49–62 (1971)
79. V.M. Zhukov, Processes on the surface under field emission. SPb (VVM, Russia, 2007), 295
p.
80. I.L. Sokolskaya, Application of a field emission microscope to study the surface diffusion and
self-diffusion. ed. by Y.E. Geguzina, Procedinds of “Surface diffusion and spreading” Nauka,
108–148 (1969)
81. B. Honigman, Growth and shape of the crystals. M. IIL, 224 (1961)
82. C. Herring, Structure and properties of solid surface. ed. By R. Gomer, S.S. Smith (University
of Chicago Press, USA, 1953), pp. 5–72
83. C. Herring, The physics of powder metallurgy. ed. by W.E. Kingston (McGrow Book Co., N.
Y., 1953), p. 143
84. W.W. Mullins, Theory of thermal growing. J. Appl. Phys. 28(3), 333–339 (1957)
85. W.W. Mullins, Flattening of nearly plane solid surface to capillarity. J. Appl. Phys. 30(1), 77–
83 (1959)
86. J.L. Boling, W.W. Dolan, Blunting of tungsten needles by surface diffusion. J. Appl. Phys. 29
(3), 556–559 (1958)
87. E.W. Mueller, Oberflachenwanderung von Wolfram auf dem eigenen Kristallgitter. Z. f. Phys.
126(7–9), 642–665 (1949)
88. R.C. Sanwald, J.J. Hren, Surf. Sci. 52, 697 (1975)
89. F.A. Nichols, W.W. Mullins, Morphological changes of a surface of revolutions due to
capillarity-induced surface diffusion. J. Appl. Phys. 36(6), 1826–1835 (1965)
90. Y.I. Frenkel, On the surface atoms crawling and natural facet roughness. ZTP. 16(1), 39–50
(1947)
91. E.W. Mueller, Weitere Beobachtung mit dem Feldelektronenmicrosko. Z. Phys. 108, 668–
680 (1938)
92. M. Benjamin, R.O. Jenkins, The distribution of autoelectronic emission from single crystal
metal points. I. Tungsten, Molibdenium, Nickel in the clean state. Proc. Roy. Soc. (A) 95,
262–279 (1940)
93. I.L. Sokolskaya, Surface migration of tungsten atoms in an electric field. ZTP. 26, 1177–1184
(1956)
94. V. Zhukov, A.A. Almazov, The spontaneous growth of the field electron emission
(FEE) current to rebuild tip emitters. XXI All-union Conf. Emission Electron. L. 1, 241
(1990)
95. N.V. Egorov, V.M. Zhukov, The effect of increasing the current to rebuild emitting surface of
the field electron cathode. Poverhnost. Phys. Chem. Mech. 3, 48–53 (1995)
96. S.C. Miller Jr., R.H. Good Jr., A WKB-Type approximation to the Schrodinger equation.
Phys. Rev. 91(1), 174–179 (1953)
Chapter 4
Simulation of Structure and Parameters
of Field Emission Cathodes
Abstract In this chapter, simulation of a surface potential barrier is presented and

method of determining the barrier permeability based on experimental data is
proposed. It also presents simulation of current–voltage characteristics, surface
structure and distribution of the work function over the surface of metal
field-emission cathode for a model configuration of its vertices. Theoretical studies
of the physical processes affecting the field-emission current density limits are
discussed towards the end of the chapter.
4.1 Simulation of a Potential Barrier and Barrier

Permeability by Experimental Data
The review methods of computation of D (cf. Chap. 1) used to develop the classical
theories of field emission, i.e. of Fowler–Nordheim for metals and of Stratton for
semiconductors, shows that the mathematical assumptions simplifying computa-
tions do not result in statistically significant deviations. However, it stands to reason
that these assumptions, especially when they are not correct in terms of physics, can
manifest themselves when it is necessary to take “subtle effects” into account.
That is, there is an obvious need to develop new penetrability computation
methods chiefly taking into account a need to use them in solution of the problems
unrelated to defining the potential in the near-surface region. It is a question of
penetrability-function recovery by a set of experimental data, i.e. the energy spectrum
of outgoing electrons obtained as a result of computer or natural experiment. The
small work function variation method to be described below is specific in the respect
that if its conditions are met, penetrability can be found disregarding the barrier shape.
There is one more advantage from the mathematical point of view, as this method is
not bound with the conditions of one-dimensionality or stationarity. According to the
purpose designated, this paragraph will consider the basic results of papers [1–14].
Figure 4.1 shows the barrier-penetrability recovery results.
Dependence of the solution on the regularization parameter was studied with
different values. The pictures specify the bracketed value as recommended. The

172 4 Simulation of Structure and Parameters of Field Emission Cathodes
Fig. 4.1 The D

barrier-penetrability 1.0×10-3
λ=1010
recovered with regularization λ=1013
λ=1016
matrix 8.0×10-4 λ=1019
λ=1022
λ=1025 (λ=4.86×1024)
6.0×10-4 λ=1028
Modelling
4.0×10-4
2.0×10-4
0.0
5.0 5.2 5.4 5.6 5.8 6.0 E, eV
linearity assumption (Fig. 4.1) seems to have failed for the barrier penetrability,
which should have shown exponential dependence.
In our opinion, the best result came from the assumption of quadratic depen-
dence of the solution (Fig. 4.1). A comprehensible result came for a wide range of
regularization parameters.
Let us make a mathematical model of electron transport through a potential
barrier that combines the internal distribution of electrons by energies, a potential
barrier of an unknown shape and detected external distribution of electrons. Let
distribution of electrons inside the metal follows the law of

pffiffiffiffi E
nint ðEÞ ¼ N E exp ; ð4:1Þ
kT
where N is the normalization factor, k is the Boltzmann constant, T is the tem-

perature. In our case, N = 1, T = 500.
Let us define the energy spectrum as
E1 ¼ 0:2 eV; Em ¼ 6 eV; DU ¼ 0:02 eV: ð4:2Þ
Let us assume that in our case, the barrier follows the function Uext(x). Having
calculated the penetrability of this barrier, we can find the external distribution of
electrons next ðEÞ by the formula
next ðEÞ ¼ nint ðEÞDðEÞ:
It is this value that we will regard as the detector observed. Having replaced the
work function and the electron energy with the DФ value, we can find the changed
distribution of electrons n0ext ðEÞ after the outflow into vacuum.
Under the conditions of our experiment, let us employ the expression
~ ðEm Þ ¼ 1:
DðEm Þ 1; or D
4.1 Simulation of a Potential Barrier and Barrier Permeability … 173
Taking the nint ðEÞ function as invariable, we obtain
DðE1 Þ DðEm1 Þ DðEm1 Þ

; ;;
DðE2 Þ DðEm Þ DðEm Þ
~ ðE1 Þ; i ¼ 1; m. Basing on
and recover the function D
next ðEÞ
~nint ðEÞ ¼ ;
~ ðEÞ
D
we can find the distribution n~int ðEÞ. The sought-for characteristics are shown in
Figs. 4.2 and 4.3, where the difference between the values simulated and recovered
as a result of the numerical experiment does not exceed 15%.
Correct experiment simulation seems to need allowance for an error in the data to
be measured, i.e. the so-called “noise”. The small work function variation method
was studied for stability in relation to raise of noise. For this purpose, normally
distributed errors of two types, namely ones with constant and progressing variance,
were introduced into the external (detected) distribution. It was shown that the
Fig. 4.2 The

barrier-penetrability recovery 1,0
with regularization matrix: 1
modeled, 2 recovered with 0,8
numerical experiment
Penetrability
0,6 1
2
0,4
0,2
0,0
0 1 2 3 4 5 6
Energy, eV
Fig. 4.3 Distribution of 2,0×10 -12

electrons inside the emitter:
Distribution of electrons, r.u.
solid line modeled, circles 1,5×10 -12

recovered with numerical
experiment
1,0×10 -12
5,0×10 -13
0,0
0,20 0,25 0,30 0,35 0,40 0,45 0,50

Energy, eV
Fig. 4.4 The 1,2

barrier-penetrability: dashed
curve modeled, circles 1,0
recovered with numerical
experiment without 0,8
Penetrability
measuring errors, solid line
with measuring errors 0,6
0,4
0,2
0,0
0 2 4 6 8 10
Energy, eV
constant-variance error makes only initial bias as compared with the penetrability
recovered without noise. The progressing error behaves in a sufficiently stable way,
which makes it possible to consider the method flexible in terms of measurement
uncertainty. An example of introduction of a measurement error is shown in Fig. 4.4.
An electric field applied to a solid surface causes changes in the shape of a
potential barrier at the interface. If the electric field intensity is high enough, it
induces the current of emitted electrons. That makes it possible to research barrier
penetrability by change of value of the electric field applied. This research tech-
nique has an advantage of a possibility to apply the modulation measurement
technique cardinally reducing the noise by narrow-band recording. It also counts in
favor of this procedure that the field emission current changes very quickly even at
small field-intensity variations [15], which can result in an intolerably large error
when measurements are made under static conditions. When modulation voltage is
applied, the error can be considerably reduced.
Modulation methods are widely used in experimental research. As a rule,
interpretation of the results obtained assumes a linear or a quasilinear relationship
between the parameters under study. However, the foregoing proves that in field
emission studies, the linearity assumption seems unreasonable. This circumstance
necessitates careful analysis of the method suggested by mathematical simulation.
A physical analogue suggested is a semi-infinite metal sample in vacuum with an
electric field applied to the solid surface and a particle energy distribution detector.
Without loss of generality, let us consider a case of an external electric field
changing in time according to the harmonic law of intensity:
EðtÞ ¼ E0 þ E1 cos xt; ð4:3Þ
where E0 is intensity of the electric field, E1 is the modulation amplitude, assuming

that E0 E1, x is the modulation frequency.
Analytical isolation of the harmonic component in distribution of the electrons
that pass to vacuum can be done with a quasi-classical expression for the
potential-barrier penetrability at the cathode boundary [16]. If interaction between
an electron and the cathode surface is neglected, the penetrability can be expressed
as follows.
DðE; E0 ; E1 ; x; tÞ
" #
4ð2mÞ1=2 1=2 1 a
¼ exp ðU þ EF EÞ ¼ exp ;
hjejE0
3 1 þ EE12 cos xt 1þb ð4:4Þ
4ð2mÞ1=2 1=2 E1
a¼ U þ EF E ; b ¼ cos xt:
3hjejE0 E2
The potential barrier transparency (4.4) can be expanded in a Taylor series by b,

b 1 being a series expansion parameter, i.e.

1 2 1 3
D ¼ expðaÞ 1 þ ab þ a a b2 þ a a2 þ a b3 þ : ð4:5Þ
2 6
Density of external distribution of electrons by energies E is determined as
next ðE; E0 ; E1 ; x; tÞ ¼ nint ðEÞ DðE; E0 ; E1 ; x; tÞ; ð4:6Þ
where nint ðEÞ is the density of internal distribution of electrons by energies con-
nected with their velocity components normal to the cathode surface. Insertion of
(4.5) into (4.6) makes it possible to obtain expansion of density of external dis-
tribution of electrons (signal) into degrees of cosine:
E1
next ðE; E0 ; E1 ; x; tÞ ¼ n0ext ðE; E0 Þ þ n1ext ðE; E0 Þ cos xt
E0
2 3
E 1 E1
þ n2ext ðE; E0 Þ cos2 xt þ n3ext ðE; E0 Þ cos3 xt þ . . .;
E0 E0
ð4:7Þ
8 0
> next ðE; E0 Þ ¼ nint ðEÞ exp½aðE; E0 Þ;
>
>
>
< n1 ðE; E0 Þ ¼ nint ðEÞ exp½aðE; E0 Þ aðE; E0 Þ;
ext

>
> n2ext ðE; E0 Þ ¼ nint ðEÞ exp½aðE; E0 Þ 12 a2 ðE; E0 Þ aðE; E0 Þ ;
>
>
: 3
next ðE; E0 Þ ¼ nint ðEÞ exp½aðE; E0 Þ 16 a3 ðE; E0 Þ a2 ðE; E0 Þ þ aðE; E0 Þ :
According to [15], experimental determination of the potential-barrier penetra-

bility needs changing the external electric field intensity by some value. That makes
it possible to obtain the expressions similar to formulas (4.3) and (4.6):
E0 ðtÞ ¼ E00 þ E10 cos xt;

next E; E00 ; E10 ; x; t ¼ n0int ðEÞ D E; E00 ; E10 ; x; t :
In case of a metal cathode, external electric field variations do not make any
appreciable effect on internal distribution of electrons by energies [17]; which
means that
n0int ðEÞ ¼ nint ðEÞ: ð4:8Þ
Having differentiated both parts of (4.6) in terms of time and neglected the
addends of about and higher in (4.7), we can obtain the following ratio based on
equality (4.8):
E0
0 0 0 0 n1ext E; E00 E10
dt D E; E0 ; E1 ; x; t dt next E; E0 ; E1 ; x; t
d d
¼ ¼ 0
: ð4:9Þ
dt DðE; E0 ; E1 ; x; tÞ dt next ðE; E0 ; E1 ; x; tÞ n1ext ðE; E0 Þ EE10
d d
The idea to use a modulating signal is aimed at experimental determination of

function n1ext ðE; E0 Þ. Formula (4.9) gives sufficient condition to determine the
barrier penetrability within the accuracy of a constant multiplier. The density of
energy distribution of electrons in the cathode can be obtained on the basis of
formula (4.6).
Use of quasi-classical approximation [18] for analytical isolation of the harmonic
component of the emission signal has to be justified in virtue of the restrictions
imposed on the WKB method. That can be done by the methods of mathematical
simulation and numerical experiment.
According to the free-electron model [19], the density of distribution of electrons
by energies connected with their velocity components normal to the cathode surface
is given by

4pmkB T E EF
nint ðEÞ ¼ ln 1 þ exp ;
h3 kB T
where kB is the Boltzmann constant, T is the absolute temperature. The potential

energy of an electron at the cathode boundary (potential barrier) is given by
U ðx; tÞ ¼ U þ EF jejE ðtÞx e2 =16pe0 x: ð4:10Þ
Calculation of the barrier penetrability makes it possible to find the density of the
external distribution of electrons next ðE; E0 ; E1 ; x; tÞ by formula (4.6). The variation
of intensity of an external electric field results in modification of the
electron-distribution density next E; E00 ; E10 ; x; t after the escape into vacuum.
Quantum-barrier penetrability can be determined on the basis of solution of the
Schrödinger stationary equation [20] (the time for an electron to tunnel the barrier is
much less than the modulation period [21]):
d 2 w 2m
¼ 2 ½U ðx; tÞ E w; ð4:11Þ
dx2 h
where w is the wave function of the electron. Equation (4.11) can be solved by the
Runge-Kutta-Nyström method for different energy values and different time
moments. That enables numerical integration by e.g. a trapezium method to find
expansion of the potential-barrier penetrability into the Fourier series for a partic-
ular electron-energy value and evaluate the contribution of the addends that we
neglect when deriving formula (4.9).
The machine-assisted implementation of numerical algorithms for the mathe-
matical model presented above in operation [9] employed the cathode parameters
typical for tungsten (Ф = 4.5 eV, EF ¼ 11:7 eV). The density of internal distri-
bution of electrons by energies was calculated for the temperature of 300°K.
Figures 4.5 and 4.6 represent the densities of external distribution of electrons by
Fig. 4.5 Density of external next, r. u.

distribution of electrons (DE =
8,0×1024
0.01): solid line intensity of
electric field E = 0.99 109
V/m, dashed curve E = 1.01
6,0×1024
109 V/m
4,0×1024
2,0×1024
0,0
11,0 11,2 11,4 11,6 11,8 12,0 12,2 12,4

E, eV
Fig. 4.6 Density of external next, r. u.

distribution of electrons (DE =
6,0×1024
0.01): solid line intensity of
electric field E = 0.999 109
V/m, dashed curve E = 1.01
4,0×1024
109 V/m
2,0×1024
0,0
11,0 11,2 11,4 11,6 11,8 12,0 12,2 12,4

E, eV
energies for different values of the modulating-signal amplitudes. Obviously, sen-

sitivity of the device allows of no too-large modulating-signal amplitude, which
meets the procedure requirement of littleness of the parameter DE ¼ E1 =E0 .
Figure 4.7 exemplifies the potential barrier transparency (4.10) found by numerical
solution of the Schrödinger equation and its numerical expansion into the Fourier
series up to the second and third addends. The results show that dependence of the
potential-barrier penetrability on the modulating-signal value can be considered as
linear for the electrons with energies E EF that make the greatest contribution to
the emission current. In virtue of (4.8), the same conclusion can be made about the
density of external distribution of electrons by energies. For electrons with energies
considerably exceeding EF , contribution of the third Fourier component and, hence,
the entire remaining amount, in expansion of the penetrability becomes negligibly
small.
To sum it up, it is necessary to underscore that analytical methods and a nu-
merical experiment can show that the density of external distribution of electrons by
energies can be deemed linear in relation to the modulating-signal value. This
circumstance makes it possible to develop a procedure of determination of the
potential-barrier penetrability at the cathode boundary according to the data of the
natural experiment.
The procedure of optimum-control construction by the successive-approximation
method seems rather inconvenient in our case. So we suggest searching for optimum
control in a class of functions.
Correct solution of the task at hand needs knowing the barrier penetrability for
the widest range of electrons. Actually, a particular value of penetrability can be
matched with any number of barriers of various shape having similar penetrability.
In addition, its continuity means that the energy range should not be narrow because
coincidence will also be quite heavy in it.
Fig. 4.7 Dependence of D

potential barrier transparency
1,2×10-18
on time for full period (DE =
0.001, E0 = 109 V/m, E = 5.0
eV): solid line precise value,
dashed curve two first Fourier 1,0×10-18
components, dashed dotted
curve three first Fourier
components 8,0×10-19
6,0×10-19
4,0×10-19
-100 -50 0 50 100

t, r. u.
Let us consider the function system p0 ðxÞ; . . .pn ðxÞ. Let us represent the potential
as a linear combination of these functions:
X
n
UðxÞ ¼ ai pi ðxÞ: ð4:12Þ
i¼0
Let us attempt to solve an inverse penetrability problem as
next ðEÞ ¼ nint ðEÞDðEÞ:
In this case, we are in the space of linear-combination coefficients and the func-
tional minimum
ZEmax Zb
Q 2
JðuÞ ¼ ðDðEÞ dðEÞÞ2 dE þ uðtÞ U þ Et þ dt
t
Emin a
(here u is the control function, t being the time) should be sought in this space:
Uonm ða; xÞ ¼ Uða; xÞ; ð4:13Þ
where a ¼ ða0 ; . . .
an Þ affords a minimum to the functional J(U) given as
JðUðaÞÞ ¼ min JðUða0 ; . . .an ÞÞ: ð4:14Þ

a0 ;...an
For the system of functions p0 ðxÞ; . . .; pn ðxÞ, physical considerations can help to
parcel out a limited subspace of the linear combination coefficients, where we can
find the minimum quality-functional with descent methods. Naturally, there is a
possibility of hitting a local minimum. This problem can be avoided by step-by-step
scanning of all coefficient values on the desired hypercube. If the minimum point
gets on a hypercube facet, its center can be transferred to that point and the scanning
can go on. However cumbersome this procedure might be, it allows for a solution in
the desired class of functions.
Nevertheless, solution of the task at hand needs additional research of possible
classes of functions suitable for potential approximation. Analysis of the research
findings suggests stopping at the polynomial system as most wide-spread. The
potential reduction for rectangular and triangular barriers results in accurate coin-
cidence because it is a question of the zero- and the first-degree polynomials,
respectively, in this case; the barrier penetrability being considered as known
without an error. In case of a triangular barrier with account of image forces, there is
a necessity to carry out additional research of the possibility of barrier approximation
by polynomials from the point of view of the problem of the barrier penetrability.
The results obtained show that the small work function variation method can
successfully be applied to calculation of the transmission coefficient under the
conditions of a real experiment.
The mathematical model of the method contains the assumptions that are directly
or indirectly justified by the results of the numerical experiment.
Unlike the wide-spread methods of direct solution of the penetrability problem,
the method does not need any information on the kind of the potential electron
function, which is essential because the latter can be written out for a particular
model only with certain assumptions and restrictions.
The method does not stipulate any space dimensionality or time dependence,
which shows its universality.
Introduction of normally distributed errors into the simulation results showed
tolerance of the method to their occurrence.
In conclusion of the paragraph, let us re-write the inverse penetrability problem
in terms of the optimum-control construction problem.
Let there be a system of
v_ ¼ Gðv; u; tÞ; ð4:15Þ
where v denotes the phase coordinates, u is the control function, t being the time.
In our case,
v ¼ Y; u ¼ f ; t ¼ x: ð4:16Þ
Let the coordinate initial-value vector be set as. We have

v1 ðbÞ ¼ expðik3 ðEÞbÞ;
ð4:17Þ
v2 ðbÞ ¼ ik3 ðEÞ expðik3 ðEÞbÞ:
And let there be a set of constraints gj ðv; u; tÞ 0. In our case, such restrictions
can be posed by the inequalities of the following form:
f ðtÞ U 0; DðEÞ 1 0: ð4:18Þ
It is required to find the control u0 ðxÞ that would afford the minimum to the
following functional:
Zb
JðuÞ ¼ pðvðaÞÞ þ G0 ðv; u; tÞdt; ð4:19Þ
a
i.e. JðuÞ
Jðu0 Þ.
Let us construct such a functional for our problem. We know penetrability d for
each point E. At the desired control, the estimated value follows the formula of
k3 ðEÞ 1
DðEÞ ¼ : ð4:20Þ
k1 ðEÞ jAðEÞj2
Wave vectors are known. Therefore, we have to determine value A through the
finite value of the phase coordinates, i.e.

A expðik1 aÞ þ B expðik1 aÞ ¼ x1 ðaÞ;
ð4:21Þ
ik1 ðA expðik1 aÞ B expðik1 aÞÞ ¼ x2 ðaÞ:
Wherefrom

1 1
A ¼ expðik1 aÞ v1 ðaÞ þ v2 ðaÞ : ð4:22Þ
2 ik1
Consequently,
k1 ðEÞk3 ðEÞ
DðEÞ ¼ 4 : ð4:23Þ
k1 ðEÞv21 ðaÞ þ v22 ðaÞ
2
So the first component of functional (4.17) can be written as
ZEmax
2
pðxðaÞÞ ¼ kDðEÞ dðEÞk ¼ ðDðEÞ dðEÞÞ2 dE: ð4:24Þ
Emin
There is also a component part of the functional, namely its time integral. For a
more reasonable control, function G0 can be selected as a diversion from the
theoretically predicted shape, i.e.

Q 2
G0 ðx; u; tÞ ¼ uðtÞ U þ Ft þ : ð4:25Þ
t
We have set all the necessary conditions to establish the optimum control of our
system, which will coincide with the required potential. The notation of the func-
tional makes it possible to apply the V.I. Zubov’s method of successive approxi-
mation [17] to this problem.
4.2 Computation of Theoretical Characteristics of a Metal

Field Emission Cathode for a Model Configuration
of Its Apex
Derived for the flat geometry of a field emission cathode, the Fowler–Nordheim
equation relates emission current density to the external field intensity. It is obvious
that the natural experiment measures the total emission current and the applied
potential difference, but voltage V is assumed to be related to intensity E through

field factor b:
jEj ¼ bV:
Under these circumstances, the total emission current is an integral equation

Z Z Z
B B
I¼ jdS ¼ AjE j2 exp dS ¼ Ab2 V 2 exp dS: ð4:26Þ
jE j bV
ðSÞ ðSÞ ðSÞ
where S is the emitting surface.

For the case of flat geometry of the emitter, when b ¼ d1, d being a constant
distance between the anode and the cathode, we derive dependence of the total
current on voltage by renormalization of factors A and B in expression (4.26):

0 B0
I ¼ A V exp
2
; ð4:27Þ
V
the dependence being linear in the Fowler–Nordheim coordinates. Here,

A0 ¼ ASb2 ; B0 ¼ Bb.
In practice, it is necessary to take into account that the real metal surface is
usually heterogeneous in terms of both chemical and geometrical properties even if
we deal with a flat field emission cathode. A little change in the composition or
presence of microelevations on the surface of the material is enough for the smallest
dust particles to amplify the electric field, generating considerable emission current.
Ohmically heating up, these centers decay, after which the current is generated on
new heterogeneities. So, the experiment shows that emission from flat emitters is
unstable, and the total current is small [22].
That is why flat emitters are very seldom in practice. The up-to-date emission
devices are based on apex-shaped emitters that can be matched with the voltage
values producing high enough current and stable emission [23]. However, that
considerably complicates the mathematical model of field emission. Mathematical
computations are frequently simplified by application of idealized models of
emitters unrealizable in experiment. E.g., many authors approximated apex emitters
as a sphere [24–26], which considerably simplified the mathematical description of
field emission. Actually, real emitting apices have much more complicated con-
figurations [27]. b differs from surface element to surface element; hence, the field
gradient on the apex surface is spatially inhomogeneous. Therefore, the field gra-
dient of a non-planar geometry emitter is the function of coordinates of emitting
apex surface X:
jE ðl; g; uÞj ¼ Vbðl; g; uÞ: ð4:28Þ

4.2 Computation of Theoretical Characteristics of a Metal Field … 183
Then (4.26) is represented as

Z
B
I¼ Abðl; g; uÞ2 V 2 exp dX; ð4:29Þ
bðl; g; uÞV
ðXÞ
and though the squared voltage V 2 can be taken outside the integral sign, we cannot
do the same with the exponent, which means that from the theoretical point of view,
neither dependence IðVÞ can be obtained in the form of (4.27), nor can we speak
about linearity of (4.29) in the Fowler–Nordheim coordinates. However, it has
already been noted that numerous experimental research of field emission lead to a
linear dependence lgðI=V 2 Þ ¼ f ð1=VÞ.
The generally accepted real-emitter shape approximation models include the
second-order rotation surfaces [24–26, 28–32]. The axial symmetry of such models
saves from extra mathematical difficulties in calculation of the emission charac-
teristics connected with tridimentionality of the problem.
In curvilinear coordinates of a prolate ellipsoid of revolution, the emission
current density can be represented as
2 qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi3
AC12 Ba sin g0 ch2 l cos2 g0
j¼ exp4 5: ð4:30Þ
a2 sin2 g0 ch l cos2 g0
2 C1
It is obvious that
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
dI ¼ jdS ¼ 2pa j sin g0 shl sin2 g0 þ sh2 ldl:
2
ð4:31Þ
Then the total electron-emission current from the cathode surface approximated
by a sheet of the two-sheeted hyperboloid of rotation is expressed as
Za pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
sin2 g0 þ sh2 l
2pAC12 shl Ba sin g0
I¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi e C1
dl: ð4:32Þ
sin g0
0 sin2 g0 þ sh2 l
Substitution of the variables s0 ¼ cos g0 , r ¼ ch l, and insertion of the value C1

give
Za pffiffiffiffiffiffi2 1 þ s0 pffiffiffiffiffiffiffi
ffi
r2 s2
2pAV 2 1
aB 1s ln
0 1s0 0
I ¼ pffiffiffiffiffiffiffiffiffiffiffiffi2ffi 2 1 þ s pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffie V dr: ð4:33Þ
1 s0 ln 1s00 r2 s20
1
Total emission current IðVÞ was obtained in numbers for all second-order sur-
faces considered in the paper. Voltage-current characteristics approximated by
straight lines were constructed for a hyperboloid, a paraboloid and the ellipsoids of
rotation. Linear approximation was carried out by the least-squares method, LSM
[33], the correlation factor falling in the range of 0:99999 R 0:99998
(Fig. 4.8). The calculations for all the surfaces of rotation considered were made on
the parameters typical for a classic field emission diode with r0 ¼ 1 µm, d = 1 cm.
What is more, characteristic curves for a hyperboloid and an oblong ellipsoid of
rotation were constructed on the parameters characteristic for a scanning tunnel
microscope, where r0 ¼ 1 nm, d = 2 nm.
Figure 4.9 shows the voltage-current characteristics of all the surfaces of rotation
considered.
A hyperboloid, a paraboloid and an oblong ellipsoid of rotation are suggested as
approximations of apex emitters, while an oblate ellipsoid of rotation is to
Fig. 4.8 Dependence IðVÞ lg(I/V 2), r. u.

for hyperboloid in the -5
Fowler–Nordheim
coordinates: circles
voltage-current -10
characteristics, solid line
approximation by straight
-15
line, U = 4.5 eV
-20
-25
-30
0,0 3,0×10 -5 6,0×10-5 9,0×10-5 1,2×10-4 1,5×10-4 V -1,V-1
Fig. 4.9 Voltage-current hyperboloid

characteristics of the surfaces lg(I/V 2), r. u. paraboloid
of rotation in the Fowler– oblong ellipsoid
-5 oblate ellipsoid
Nordheim coordinates
-10 Φ = 4,5 eV
-15
-20
-25
-30
-1 -1
0,0 3,0×10-5 6,0×10-5 9,0×10-5 1,2×10-4 1,5×10-4 V ,V
approximate edge-type cathodes (emitters of such a shape are also widely used in
microwave electronics [34]).
The graphs in Figs. 4.8 and 4.9 show that in the analyzed range of voltage values
ð5 103 B V 1 105 BÞ, voltage-current characteristics are linear in the
Fowler–Nordheim coordinates, which 1 was also mentioned in many other experi-
mental papers, e.g. [23], as lg V 2 ; V .
I
But from the theoretical point of view, linearity of expression (4.33) in the
Fowler–Nordheim coordinates can only be proved through further mathematical
studies of this area within the framework of the problem in hand.
The mathematical models suggested can be applied to real apex emitters
(Fig. 4.10) [28]. The approximating model for a particular apex was selected with
the least-squares method, LSM.
The apex shown in Fig. 4.10 was suggested to be approximated with two
models: a sheet of a two-sheet hyperboloid of rotation and a paraboloid of rotation.
Analysis of the resulting curve showed that the curve is best of all approximated by
a hyperbola (Fig. 4.11) expressed by
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ðX X0 Þ2
YðXÞ ¼ 1 þ b þ Y0 ; ð4:34Þ
a2
where a ¼ 2:037 107 ; b ¼ 1:62 107 ; X0 ¼ 1:092 105 ; Y0 ¼ 2:49 106 .

The radius of curvature of this real apex was calculated as r0 ¼ 2:56 107 m,
which was why the apex was approximated with the model of a two-sheeted
hyperboloid of rotation. Parameter d, i.e. the distance between the cathode and the
anode, was taken as 2 cm, which is typical for a classical field emission diode.
The constructed mathematical models for different second-order surfaces gave
the emitter parameters as field factor b ¼ jEjmax =V and emission area S as the area
of the approximating surface of rotation within which emission current density falls
no more than 10 times. Expressions for b and S are presented in Table 4.1.
Fig. 4.10 Photo of real 1 μm

emitter by shadow electron
microscopy
Fig. 4.11 Approximation for z, m curve

a particular apex was with the 3,0×10-6 hyperboloid
least-squares method, LSM paraboloid
2,9×10-6
2,8×10-6
2,7×10-6
1,0×10-5 1,1×10 -5 1,2×10 -5

x, m
Table 4.1 The emitter parameters in curvilinear coordinates

Surface of revolution Geometric field Emission surface S
factor
R pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Prolate ellipsoid of revolution C1 2pa2 shl0 sin g sin2 g þ sh2 l0 dg
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
aVshl0 ch2 l0 1
R qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Two-sheet hyperboloid of p
C1ffiffiffiffiffiffiffiffiffiffiffiffiffiffi
aV sin g0 1cos g02 2pa2 sin g0 shl sin2 g0 þ sh2 ldl
revolution
Paraboloid of revolution Cp
1 ffiffiffiffi R qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Vl0 l20 2pl0 k l20 þ k2 dk
R qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Oblate ellipsoid of revolution C1
qffiffiffiffiffiffiffiffiffiffiffi 2pa2 chl0 sin g cos2 g þ sh2 l0 dg
aVchl0 sh2 l0
Here C1 is a constant determined from the respective boundary conditions. It is individual for
every rotation surface
The parameters obtained from mathematical models turned out different from
and determined from the voltage-current characteristic in the Fowler–Nordheim
coordinates by the formulas

I
b ¼ 2:82 109 U3=2 =p; S ¼ 4 1017 p2 ;
V2
V ¼0
1
where p is the voltage-current characteristic slope ratio.

For comparison, Table 4.2 gives the values of b and S obtained by the different
methods. The calculations were made for several types of apex emitters with
involvement of the parameters typical for a classical field emission diode and a
scanning tunnel microscope.
A lot of researchers working in field emission have still been determining emitter
parameters b and S with the procedure developed for an emitter of flat geometry.
Table 4.2 Emitter parameters for some second-order surfaces of rotation

Surface of revolution Curvature radius, m, According to the By volt-ampere
distance, m model characteristics
Prolate ellipsoid of r0 = 1 10−9 b = 1.9 10+9 b = 1.7 10+9
revolution d = 1 10−9 S = 4.2 10−18 S = 3.2 10−18
r0 = 1 10−6 b = 1.9 10+5 b = 1.7 10+5
d = 2 10−2 S = 1.2 10−12 S = 7.2 10−13
Paraboloid of r0 = 1 10−9 b = 1.6 10+9 b = 1.5 10+9
revolution d = 1 10−9 S = 9 10−19 S = 6.7 10−19
r0 = 1 10−6 b = 1.77 10+5 b = 1.71 10+5
d = 2 10−2 S = 1.1 10−12 S = 6.8 10−13
Surface of revolution r0 = 1 10−6 b = 1.89 10+5 b = 1.82 10+5
d = 2 10−2 S = 1.1 10−12 S = 6.9 10−13
Prolate ellipsoid of r0 = 1 10−6 b = 1.99 10+5 b = 1.95 10+5
revolution d = 2 10−2 S = 5.7 10−10 S = 3.9 10−11
The mathematical simulation results we have presented show that the values
determined in that way need substantial adjustment. That may turn out essential
because parameters b and S have not only theoretical, but also technological
importance as variable objectives of vacuum micro- and nanoelectronic devices.
4.3 Simulation of the Metal Field Emission Cathode

Surface Structure
While there is a large number of papers on simulation of field-ion microscopic

images of the apex-emitter surface (cf. [35–38] as an example), similar models for a
field emission cathode are presented in both domestic and foreign literature only by
the publications of the A.L. Suvorov Group [39]. In this single procedure sug-
gested, the cathode shape is ignored, and the apex sample is approximated by an
abstract polyhedron inadequate to the real situation.
Along with [33, 40–51], this paragraph suggests a crystalline-structure surface
and crystallographic-facet building model with different work functions at the
cathode apex. This model can directly calculate the current-density distribution and
the emission area, unlike the traditional approach based on the average-weighted
work function [52].
The chief difficulty arising in quantitative analysis of field emission measure-
ments is connected with an inconvenient emitter shape, as the emitting surface is in
the area of major curvature so that quite small voltages can generate the field
gradient necessary for field emission. As a result, standard physical assumptions
often do not hold, for example, geometrical field factor b and work functions Ф
change over the emitting surface, the emitting area of emitting surface S depending
on the voltage applied.
Though the technique used to compare the theoretical and experimental results
on the basis of the Fowler–Nordheim formula with insertion of constant Ф and S, is
generally accepted, it can hardly be considered as quite correct. Making such
comparison, one should pass from the experimentally measured values I and V to
characteristics j and E, which is suggested to be done by mathematical simulation of
the real configuration of the field emission cathode whose apex surface consists of
flat areas meeting well determined crystallographic facets.
The images obtained in a field-ion microscope reconstruct the atomic structure of
the emitter surface, i.e. spot the position of each surface atom. Sizing up the
reconstructed model and the ideal three-dimensional surface-structure model
(Fig. 4.12), one can identify adsorbed particles and other defects on the surface and
determine the local emitter-surface curvature radii [35].
There are two approaches in this area, namely the “thin shell model” and the
“broken bonds model” also referred to as the local atomic environment model [40].
Basically, similar approaches can be applied to simulation of a field emission
emitter surface as well; however, nobody has done that yet. It is not improbable that
the reason thereof is that computations of this case become rather complicated
because the sizes of a typical metal electron emitter can be one or two order of
magnitude larger than those of an ionic emitter. Papers [37, 39] present the field
emission image models; however, the real shape of the apex sample is actually
ignored and replaced with approximation by an abstract polyhedron.
This paragraph focuses on the apex-emitter surface modeling procedure within
the framework of the thin shell and broken bonds approaches to field emission. The
simulation is meant to analyze distribution of the atomic-packing density and the
Fig. 4.12 “Dark balls” model

for apex surface structure of
tungsten emitter with apex
curvature radius 36 nm. Dark
balls correspond to atoms that
showed in field ion
microscope (according to
“broken bonds” model)
4.3 Simulation of the Metal Field Emission Cathode Surface Structure 189
work function over the surface, determination of arrangement and sizes of the
external single-crystal facets and model-emission imaging. The surface character-
istics listed are critical for the processes of field emission, adsorption, heating, etc.
E.g., the surface-facet sizes indicate not only amplification of the exterior electric
field during field emission, but also the surface area contacting with exterior
molecules or adsorbed particles. In its turn, the density and the atomic-packing
structure determine stability and sustainability of the surface, as well as the most
important emission characteristic, namely the electron work function.
The emitter surface structure is determined by arrangement of surface atoms. An
atom is understood as belonging to the surface if it has at least one broken bond
with an any-order neighbor, i.e. if it has an incomplete number of neighboring
atoms:
Ni Nimax 1 ði ¼ 1; . . .; MÞ; ð4:35Þ
where Ni is the number of the neighboring atoms of the ith order, Nimax is the
maximum number of the ith order neighbors in the body, M is the maximum order
of the neighbors analyzed.
The surface atom coordinates are constructed in two stages. At the first stage, the
thin-shell approach serves to localize surface atoms, i.e. the atoms lying in a
near-surface layer of a set thickness limited to the geometrical surface determining
the shape of the crystal. At the second stage, the broken-bond approach is used to
identify surface atoms, i.e. the number of the closest neighbors of the near-surface
layer atoms are determined and the surface atoms having at least one broken bond
are identified.
Geometrically, the stage 1 algorithm implies representation of the emitter body
as a set of axial crystalline planes/crystal sections perpendicular to the
surface-rotation axis and determination of the sites lying in the near-surface layer of
each plane. The initial data are the crystal-lattice structure, the surface shape and the
crystal orientation.
In the model involved, the geometrical surface limiting the emitter crystal is
approximated by the sphere-on-the-cone model [53] (Fig. 4.13) and the
body-centered cubic lattice structure is considered. However, the algorithm works
on any axially symmetrical crystal shape, as well as the face-centered cubic lattice
structure. It is obvious that the thickness of the thin shell calculated at the first stage
should be at least the length of the link to the neighbors of the Mth order.
The arrangement of the surface atoms determines the crystallographic facets
opening out into the emitter surface. It is necessary to determine the normal line of
the facet containing each surface atom. The problem is solved on the basis of the
fact that the atoms having an identical geometrical environment, i.e. an identical set
of neighbors, are equidistant from the separating plane of the crystal surface [36].
So, a particular surface atom with radius vector rð1Þ and the number of neighbors
ð1Þ
Ni has to be matched with two nearest surface atoms with radius vector rð2Þ and
rð3Þ having the same set of neighbors, and if these three atoms do not lie on one
y, 10-7 m
4
3
2
r
1 θ0 θ
0
a
-1
-2
-3
-4
-10 -8 -6 -4 -2 0 2 x, 10-7 m
Fig. 4.13 The equal potential line of electric field created by charged sphere on the cone (showed
with dashed line). The equal potential line approximates the emitter shape. Symbols are explained
in (4.36): a ¼ 120 109 m, curvature radius of emitter apex r0 ¼ 239 109 m, R ¼ 0:04 m,
n ¼ 0:14 ðh0 177 Þ
straight line, then the plane resting on these atoms is the one determining the
required crystallographic facet and its normal line is calculated by formula
nðrð1Þ Þ ¼ ½ðrð2Þ rð1Þ Þ ðrð3Þ rð1Þ Þ; ð4:36Þ
on condition that
8 ð3Þ ð3Þ
< r r \r r þ r r ;
> ð2Þ ð1Þ ð1Þ ð2Þ
rð3Þ rð2Þ \rð1Þ rð2Þ þ rð1Þ rð3Þ ;

: rð1Þ rð3Þ \rð2Þ rð3Þ þ rð2Þ rð1Þ ;
>
where
ð2Þ
r rð1Þ ¼ min r rð1Þ ; ð4:37Þ
ð1Þ
Ni ðrÞ¼Ni i¼1;...;M
ð3Þ
r rð1Þ ¼ min r rð1Þ : ð4:38Þ
ð1Þ
r6¼rð2Þ N i ðrÞ¼Ni i¼1;...;M
Here Ni ðrÞ is the number of the neighboring atoms of the ith order for the atom with
radius vector r.
Though the identical geometrical environment of the atoms means that the atoms
have an identical number of the nearest neighbors up to the arbitrary order; in
practice, the geometrical environment of an atom can be analyzed only up to some
limited order of neighbors set by finite value M (formulas (4.35), (4.37), (4.38)). In
this connection, we computed the emitter facets with account of the nearest
neighbors of only the first order (M = 1), then only the first and the second order
(M = 2), then etc. up to the 5th order (M = 5). The computations covered
hemisphere-shaped crystals with the radii ranging from 100 to 1000 lattice
parameters and a body-centered cubic structure. For all the radii, the nature of
dependence NNhkl ðMÞ between the relative number of surface atoms of the particular
facet (hkl) and order of the nearest neighbors M taken into account in construction
of the facets was identical, being the number Nhkl of atoms belonging to facet (hkl)
and being the total number of surface atoms N.
Figure 4.14 presents the computation result for the radius of 722 lattice
parameters. The computations show that simulation of hemispheric or nearly
hemispheric crystallographic emitter facets can only be limited to the neighbors of
up to the 5th order inclusive because account of the neighbors of a higher order
makes a correction for less than of atomic number of facets (Fig. 4.14), i.e.

Nhkl ðM þ 1Þ Nhkl ðMÞ

NðM þ 1Þ NðMÞ 0:01; M
5:
So the maximum order of analyzed neighbors this paper operates with is M = 5

(formula 4.35). Accordingly, the thickness of the thin shell for the body-centered
cubic structure in the surface-atom coordinate computation algorithm is assumed to
pffiffiffi
be 3 lattice parameters.
Secondly, dependence between the facet size (radius) and the hemispherical
emitter size (radius) turned out to be linear for all facets. Figure 4.15 shows these
25
M=2
M=3 M=5 (111)
M=4
M=5
20
(211)
N(hkl)/N, %
15 (221)
(411) (321)
(521)
10 (721)
(100)
(310) (210) (110)
5
0
110
321
521
721
421
210
631
543
332
732
510
511
331
831
311
832
654
731
111
921
530
751
951
320
443
211
310
100
411
531
431
542
821
432
652
653
632
641
221
532
741
433
621
541
322
643
754
711
522
941
(hkl)
Fig. 4.14 The relative number of surface atoms on emitter facets calculated with taking in account
neighbors of different order. Facet locations in unit stereographic triangle are significant for five
nearest neighbor’s approximation. Crystal has a body centered cube structure, hemispherical shape
and radius of 722 lattice parameter
Fig. 4.15 Radii of facets in g, 10-9 m

dependence on hemispherical 45
emitter size (radius). They are 40 (110)
calculated with (4.36)–(4.38)
35
(linear dependences) and
(4.39) (root dependences, 30
showed as dashed curve). 25 (210)
Calculations were performed
for tungsten (body centered 20 (310)
cubic lattice, lattice parameter 15 (321)
(411)
−0.316 nm) (110)
10 (211)
(310)
5
(321)
0 (411)
0 50 100 150 200 r , 10-9 m
0
dependences for some most atomically-populated facets. The following root

dependence formula was derived for the sizes of facet (hkl) on the emitter evapo-
rated by the field:
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2 3ar
g¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi; ð4:39Þ
3 h þ k 2 þ l2
2
where g is the average facet radius, a is the lattice parameter, r is the local
surface-curvature radius, h, k, l are Miller indices of the crystallographic facet, the
formula being true for the body-centered cubic structure only if (h + k + l) is even.
So the facet measured by the hemispherical-shaped emitter turn out oversized as
compared with the field-evaporated emitter model (Fig. 4.15).
The simulated and experimental emission images can be compared in both
qualitative and quantitative terms. In quantitative terms, we compare the curves of
photometric measurement of emission images for different crystallographic
directions.
The emission characteristics were studied in a conventional way of observation
of field emission from the apex cathode under the conditions of ultrahigh vacuum.
The studies were carried out in an ultrahigh-vacuum cabinet with continuous
pumping-out at the pressure in the region of 10−9 mm Hg. The apex installed on the
manipulator could move rotationally and transitionally. It was positioned at the
distance of 35–40 mm from the anode. The anode was represented by an earthed
luminescent screen with a diameter of 45 mm. The apex emitter was made of a
polycrystalline tungsten wire of 0.1 mm in diameter by electrochemical etching in
an alkaline medium. The experiment involved generation of projection pictures on
the screen and measurement of volt-ampere characteristics, i.e. dependences of the
total emission current on the voltage applied represented in Fowler–Nordheim
coordinates.
Figure 4.16 shows an emission image model constructed as aforesaid. We can
see qualitative correlation between the calculated (Fig. 4.16, upper right) and the
(a) (b)
(c) (d)
Intensity, relative units Intensity, relative units
60000 70000
50000 60000
50000
40000
40000
30000 30000
20000 20000
10000 10000
0
0
0 100 200 300 400 0 100 200 300 400

pixel pixel
model
experiment
Fig. 4.16 Experimental (a) and model (b) emission images for tungsten emitter, voltage 3885 V,
apex curvature radius r0 = 239 10−9 m; c and d are photometric curves of emission images for
tungsten emitter. Brightness (in relative units) was plotted for dots of (011)–(001) direction (c) and
for dots of (011)–(100) direction (d)
experimental (Fig. 4.16 upper left) emission images of the tungsten emitter. The
images were photometrically measured along the two lines connecting the centers
of facets (011) and (001) (Fig. 4.16, lower left) and the centers of facets (011) and
(100) (Fig. 4.16, lower right).
As the local changes of electric intensity on the emitter surface are only taken
into account on average, by means of integrated parameter c, the maximums of the
photometric curve of the simulated image is somewhat different in value from the
maximums of the experimental curve. The biased position of the maximums of the
simulated curve means that the facets measured by the smooth sphere-on-the-cone
emitter model were oversized.
It is likely that computation of facets by the initial surface structure in a form of
an unsmoothed model of the emitter evaporated by the field can produce more
adequate emission images. Finally, please note that the suggested mathematical
model features a surface structure of the apex-shaped field emission metal emitter,
taking into account the arrangement of different crystallographic facets at the
emitter apex.
The emitter shape is approximated by the smooth equipotential surface, and the
crystalline lattice is assumed to be ideal.
In reality, the ideal emitter surface undergoes changes on adsorption, relaxation,
reconstruction and temperature flattening, and account of these phenomena is a
separate problem basically solvable within the framework of this approach [38].
Disregard of those structural details results in oversizing of the facets as compared
with those of the field-evaporated emitter.
On assumption of the ideal surface structure and application of the integrated
parameter for account of the local changes of the field on microscopic inhomo-
geneities of the surface, the model adequately features the data of the natural
experiment.
The broken bonds approach used to simulate the surface structure makes it
possible to apply the model to the case of adsorption on the emitter involving
account of coordination surface-atom numbers [54].
4.4 Simulation of Distribution of the Work Function Over

the Field Emission Cathode Surface
Among the techniques to study the processes taking place on the surface of metals
and semiconductors, field emission research techniques and especially field emis-
sion microscopy are noted for importance of the information delivered and deep
penetration into the micromechanism of the phenomena studied. The reason is a
possibility of visual observation of the course of adsorption, diffusion and migration
at the resolving ability of 20–30 Å (cf. Chap. 1). Moreover, this technique involves
delicate measurements delivering valuable data on the major surface characteristics
along with a possibility of active influence on the subject of enquiry. It is also
important that probation in a field emission microscope can be carried out in
ultrahigh vacuum along with a possibility of attainment and maintenance of atomic
cleanliness of the surface for the entire period of experimental measurements.
However, studying of the surface processes should not be reduced to qualitative
visual observation, especially as far as comparison of the observational results with
theoretical representations is concerned, which is why quantitative studies should
be based on the data of mathematical simulation of the emitting surface. According
to what has been said, the simulation of distribution of the work function over the
emitting surface is of a doubtless interest.
Let us consider an apex of tungsten wire extended along the crystallographic
direction [001]. Then the cathode apex will have the work functions meeting the
reduced Miller indices. Tungsten was selected for the reason of availability of a
great deal of experimental information.
4.4 Simulation of Distribution of the Work Function Over the Field … 195
The crystallographic plane with Miller indices (hi, ki, li) agrees with a particular
value of work function Фi. This information can be treated as a database filled up as
a result of natural [39, 55] or numerical [52] experiments. Surma’s paper [52] is
taken as a basis because of a high correlation factor with the experimental data.
Normal line ni is found for each work function value from the database by formula

ni ¼ s1 1
hkl ðhi ; ki ; li Þ ¼ shkl hi ex þ ki ey þ li ez ; ð4:40Þ
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
where shkl ¼ h2i þ ki2 þ l2i . All components of vector ni are one-octant-oriented
(x
0, y
0, z
0). In other words, u 2 [0, p/2].
Let us construct normal line n = (nx, ny, nz) in an arbitrary point of a hyper-
boloidal apex surface (l, u):
cos kc shl cos u cos kc shl sin u sin kc chl

nx ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ; ny ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ; nz ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi : ð4:41Þ
sin kc þ sh l
2 2
sin kc þ sh l
2 2
sin2 kc þ sh2 l
Here all components of the normal line are positive, i.e. the cathode apex is
reoriented along axis Oz for convenience. The angle u condition is caused by the
fourth-order symmetry axis in cubic crystals.
So if normal line (4.41) is close to one of normal lines (4.31), e.g. with index i,
the work function value in the (l, u) point can be assumed as equal to Фi.
Let us determine the proximity measure of normal lines (4.40) and (4.41). It can
be a component-by-component comparison or determination of the nearest direction
by the following method. Let us consider the scalar product of

ni ðhi ; ki ; li Þ; n nx ; ny ; nz ¼ cos #;
where # is the angle between the vectors. The smaller the angle between the
vectors, the more cos # tends to one. As a result, the nearest normal line can be
looked for among the minimums of the following expressions:

J1 ðhi ; ki ; li Þ ¼ max jhi nx shkl j; ki ny shkl ; jli nz shkl j ; ð4:42aÞ
2
J2 ðhi ; ki ; li Þ ¼ ðhi nx shkl Þ2 þ ki ny shkl þ ðli nz shkl Þ2 ; ð4:42bÞ

J3 ðhi ; ki ; li Þ ¼ jhi nx shkl j þ ki ny shkl þ jli nz shkl jg; ð4:42cÞ

J4 ðhi ; ki ; li Þ ¼ nx hi þ ny ki þ nz li shkl ; ð4:42dÞ
Equation (4.42a) being the minimax functional; (4.42b) being the congruence;
(4.42c) being the comparison by the sum of the moduli; (4.42d) being angular
proximity.
The functional minimum was searched for by looking through admissible values
in the database. Let us define the acceptable-value concept. As a high-index
crystalline facet can be represented by a set of relatively low-index steps, it is

reasonable to introduce a limit for values h, k and l. It is not least because the
high-index facets have high-surface energy and their occurrence turns out unfa-
vorable in terms of energy. Let us consider one of the three versions:
maxfhi ; ki ; li g imax ; ð4:43aÞ

qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
h2i þ ki2 þ l2i imax ; ð4:43bÞ
hi þ ki þ li imax ; ð4:43cÞ
Equation (4.43a) the normal vector ends get into a limiting cube; (4.43b) the normal
vector ends get into the one-eighth of a ball; (4.43c) the normal vector ends get into
a simplex.
The uniform distribution of points over the cathode surface can be used to
determine the relative fraction of different facets in the surface. For this purpose, it
is enough to construct a histogram of distribution of points over the surface with
assigned Miller-index values (Fig. 4.17). The histogram was constructed by M =
1,118,146 points for the algorithm of N = 106. The total number of points M is
proportional to area S of the hyperboloid sector from l = 0 to l = lmax. As a result,
the histogram gives an idea of contribution of Si into total area S of different
crystalline facets. Let us note absence of several low-index facets, which seems to
be caused by the small value of lmax.
The average electron work function can be determined as follows. Let Mi be the
number of occurrences of a point on the surface. The point has been assigned the
work function value of Фi. Then the average value of U is determined as
1X 1X
U¼ U i Si U i Mi :
S i M i
Fig. 4.17 Histogram of Number of points

distribution of points over the 10 5
4.
35
cathode surface
4.
34
4.
45
5 · 10 4
4.
48
4.
4.
91
42
4.
42
4.
54
4.
4.
29
72
4.
43
4.
74
5.
4.
4.
4.
4.
4.
4.
64
44
41
44
70
68
62
0
12
13
11
11
11
12
01
02
13
22
23
33
00
01
01
01
11
11
12
4
1
2
2
3
3
4
1
1
3
6
Face
Table 4.3 Some work function values for tungsten [52]

Face Experimental Calculated Face Experimental Calculated
index value value index value value
001 4.72 4.67 116 4.29 4.35
011 5.77 5.73 122 4.41 4.39
012 4.55 4.42 123 4.54 4.53
013 4.45 4.52 124 4.35 4.33
016 4.43 4.31 133 4.68 4.34
023 4.42 4.34 134 4.74 4.42
111 4.48 4.59 223 4.70 4.34
112 4.88 4.85 233 4.44 4.45
113 4.48 4.36 334 4.62 4.40
114 4.42 4.44
The natural-experiment database containing 19 records (cf. Table 4.3) gave

value Ф = 4.49961 eV, which approved selection of the average work function
value.
The emission pattern can be simulated in the same way as before.
Proper construction was achieved by development of an algorithm providing a
uniform distribution of points inside the circle representing a projection of the
hyperboloid surface to plane z = 0. The coordinates of these points and the
current-density values corresponding to them were filed.
To begin with, let us set a quite large arbitrary number N. The estimate
interval-decomposition number [0, lmax] is derived by formula
"rffiffiffiffi#
N
Nl ¼ :
p
Let us find the distance between equidistant points:

rmax
rmax ¼ a0 shlmax ; l¼ :
Nl
Sequence li (i = 0.1…, Nl) breaks radius rmax into the segments of identical length
l. At that,
il
ri ¼ a0 shli ; li ¼ Arsh :
a0
Similarly, segment [0, 2p] is decomposed into Nu,i parts by coordinate u:
Nu;i ¼ ½2pi:
The algorithm performance is shown in Fig. 4.18. In this case, number N is a

good estimate of the resultant number of points to be used in drawing the image.
The current-density chart based on the natural-experiment data is given in
Fig. 4.19.
It is obvious that the higher limiting value imax, the heavier is fragmentarity of
the chart. At low values of imax, the resultant image can be treated as the emission
pattern made for a reshaped apex [23] because there are no high-index facets.
Conspicuous is the fact that normal-line component-by-component comparison
functional performance is similar.
Fig. 4.18 Increased

distribution of 103 points
along the circle (N = 100) 100
50
50
100
100 50 0 50 100
Fig. 4.19 The

current-density chart based on
the natural-experiment data
When constructing an emission pattern, it is important how to make it as realistic

as possible. The deterministic approach stated in this paragraph gives images with a
quite distinct silhouette (Fig. 4.20). Real images are more blurred. That can result
from heterogeneity/nonideality of the cathode surface, as well as fluctuation of
temperature, field or current caused by different reasons.
We suggest considering a random component in the normal vector (4.41). That
will enable us to get different work function values in nearby points with a certain
probability, thereby simulating inhomogeneities of the cathode surface. Clearly,
there is an infinitude of versions with a random component in the normal vector. Let
us suggest the following one-parametric model. If the normal line coincides with
axis Oz, the angle of deflection is featured by polar angle urnd and lateral angle #rnd,
which we will count from the axis. To preserve the symmetry of the system, we will
assume angle urnd to be a uniformly distributed random variable. The random
variable causing deflection by angle #rnd should contain the parameter determining
the image blur extent. Then let
urnd ¼ 2pc; #rnd ¼ jarctgðnÞj;
where c is a uniformly distributed random value on segment (0.1), n is a normally

distributed value with an expectation equal to zero and some root-mean-square
deviation r, which is a necessary parameter. The components of the random normal
line look as follows.
xrnd ¼ sinð#rnd Þ cosðurnd Þ;

yrnd ¼ sinð#rnd Þ sinðurnd Þ;
zrnd ¼ cosð#rnd Þ:
Fig. 4.20 Emission images from papers [47] and [55]

In this simulation, angle #rnd assumes the desired values in interval [0, p/2]. If r =
0, the normal line will not vary. The larger value r, the more probable is deflection
of the normal line by an essential angle. The approach suggested results in a bundle
of vectors distributed round axis Oz (Fig. 4.21). Simulation of random variable n
involved the following algorithm:
qffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffi
2 ln R2 2 ln R2
n1 ¼ t 1 R2 ; n2 ¼ t 2 R2 ;
t1 ¼ 2c1 1; t2 ¼ 2c2 1; R2 ¼ t21 þ t22 ;
if R2 > 1, the pair of n1, n2 does not suit us. This approach gives two realisations of
value n by two realisations of random variable c. Simulation of uniformly dis-
tributed value c involved the pseudorandom-number generator Mersenne Twister
MT19937 (period 219937 – 1 106001, uniformity in ℝ623).
The random-component simulation algorithm performance in the normal line is
presented in Fig. 4.21. The normality of distribution of random variable n causes
large-angle deflections to occur much less often and depend on parameter r. Within
the framework of simulation of surface inhomogeneities, that may indicate absence
of sharp turns.
To set the necessary direction corresponding to normal line (nx, ny, nz), the
random vector is turned by normal angles uп and #n:
qffiffiffiffiffiffiffiffiffiffiffiffiffiffi
n2x þ n2y
un ¼ u; #n ¼ arctg
nz
by matrix transformations
0 1 0 1
1 0 0 cos un sin un 0
T#n ¼ @0 cos #n sin #n A; T un ¼ @ sin un cos un 0 A:
0 sin #n cos #n 0 0 1
Fig. 4.21 Simulation of

random component in normal
(50 vectors)
Finally, the simulated normal line with a random component is then of the fol-
lowing form:
~ ¼ ð~nx ; ~ny ; ~nz Þ ¼ Tun T#n ðxrnd ; yrnd ; zrnd ÞT :

n
When a random component is introduced into the normal vector, the emission
pattern blurs because of the noise and visually resembles the images developed as a
result of a real experiment.
4.5 Theoretical Studies of the Physical Processes Affecting

the Field Emission Current Density Limits
The physical processes affecting the field emission current density limits were
thoroughly analyzed in [16, 56, 57]. One of the basic causes limiting the field
emission current density is heating of the field emission cathode with this current.
The spatial source of energy-release in the emitter [27, 56] under field emission
current flow is well-studied Joule heat. Nottingham was the first to identify one
more source of heat release [58]. The phenomenon he noticed is caused by dif-
ference of medium energies of emitted electrons from those replacing them, which
leads to release or absorption of field emission heat. The Nottingham effect at low
emitter temperatures leads to its heating because most of the electrons tunnel from
the energy levels lying below the Fermi level. At rather high temperatures, the
medium energy of the emitted electrons exceeds the energy of the Fermi level, i.e.
the Nottingham effect becomes a cooling factor. In this respect, it is similar to the
so-called calorimetric effect [59] accompanying thermionic emission.
The first attempt to compute the field-emitter temperature mode at limiting field
emission currents was made by Dyke et al. [60]. At that, a real field emission
cathode shaped like a truncated cone with a spherically rounded apex was replaced
with a part of a cone limited with a spherical surface of radius rm orthogonal to the
elements of the cone (Fig. 4.22). Section r = rm was selected for its ability to
provide the maximum current density. The distance from this section to the emitter
apex was approximately equal to the section radius. Thermal constants were rep-
resented by intermediate values corresponding to some medium temperature of the
cathode.
It was also assumed that the temperature at the cone base is always equal to the
initial, preliminarily set temperature T0, and the length of the conical part of the
emitter rl rj, which is only fair for an emitter with small a.
Though paper [60] did not take thermal emission caused by the Nottingham
effect into account aggravated by inadequate emitter geometry, the arbitrary choice
of specific resistance q and specific heat capacity C brought theoretical variables
quite close to experimental values of stationary field emission current density limits.
Fig. 4.22 Model of the

cathode: Tr=rl = T0; a—
hemiangle of cone
rm
rl
A more rigorous computation of field emission current density rupturing the

emitter was made by Gorkov et al. [61]. They allowed for the dependence of
physical constants of the emitter material on temperature and heat losses due to
radiation. In addition, no rl rK condition implied applicability of the solution
obtained to emitters of cone a angles as large as 90°.
The calculated dependences of the emitter temperature on the current flow time
at different densities jf-emis and at jf-emis = 2 108 A/cm2 for different a are given in
Fig. 4.23a, b.
Paper [61] shows lack of steady conditions at current densities jf-emis above some
critical density jcrit even if the emitter-material fusion temperature is not reached at
jf-emis < jcrit. Achievement of fusion temperature was usually considered as a cri-
terion of rupture of the emitting apex.
Computations and experiments showed dependence of the field emission current
density limit on the cone angle.
Papers [62, 63] showed that the Nottingham effect plays an important part in the
thermal conditions of the field emitter. For a cylindrical emitter of a finite length,
the stationary problem of thermal conduction with account of this effect was con-
sidered at the temperatures of the electron free path in the metal l * 5 10−6 cm, at
which sample sizes and shapes make an essential impact on the kinetic properties of
field emission [64].
So analysis of these publications suggests that the basic causes limiting the field
emission current values are the thermal processes determined by Joule heating of
the emitter by its own emission current and the Nottingham effect.
As the field emission density limit current is taken off, the emitter substance is in
extreme conditions. First of all, it is a situation of very strong electric fields, about
4.5 Theoretical Studies of the Physical Processes Affecting … 203
(a) (b) (c)

Tmax×10-3, K Tmax×10-3, K jcrit, A/cm2
7 6 9×108
2,5 5 4
2,0
2,0
3
8
9×10
1,5
5
1,0
1,0
4 A
3 8
2 1×10
1 2 B
1
1,2 2,4
5
t ×10 , s 3 6 t, s 10 50 90 α°
Fig. 4.23 The calculated dependences of the emitter temperature on the current flow time at
different densities jf-emis = const (a) and for different a (b): a jFEE, A/cm2 (1 1.13 108, 2 1.42
108, 3 1.55 108, 4 1.70 108, 5 1.85 108, 6 2.00 108, 7 2.15 108); b a (1 90°, 2 70°, 3
50°, 4 30°, 5 10°); c theoretical (o) and experimental dependences jcrit = f(a)
E ≅ 108 V/cm. Accordingly, there are large, close to the maximum permissible,
tensile stresses caused by ponderomotive forces. Moreover, huge current densities
cause heating of the emitter to high temperatures and emergence of large temper-
ature gradients up to *107 108 K/cm, which essentially changes the nature of the
diffusion processes both in the cathode body and on its surface.
All these factors generate a number of effects observed in experiments, such as
growth of the field emission current with time at the squared voltage impulse and
generation of a bright ring bordering the field emission image (cf. [56] as an
example).
Development of a bright ring on the periphery of an emission image of a metal
field emission cathode (Fig. 4.24) is known to predict its explosive rupture or
transition of field emission into explosive one, which is the same. That is the reason
why the “ring effect” research, going on for several decades, still attracts attention
of the experts interested in conditions of stable operation of a field emission cathode
when limiting emission currents are taken off from it.
W. Dyke et al. were the pioneers of deliberate analysis of the ring effect in their
studies of operating conditions of metal field emission cathodes with removed
currents of extreme densities. Thus, one of the first explanations of generation of a
ring was suggested, assuming movement of the temperature maximum towards the
periphery of the emitting surface of the cathode because of inversion of the
Nottingham effect in this operating mode.
In 1964, Sokolskaia et al. [65] published an article where the ring generation was
attributed to transition of a pure field emission into thermal field one. Unlike Dyke
et al., the authors of publication [65] are notable for registration of ring generation
in the stationary field emission mode, having provided the thermal field emission
Fig. 4.24 Development of a

bright ring on the emission
image of a tungsten field
emission cathode: I = 25 mA,
s = 1 ms, U = 8.9 kV (at U =
11.3 kV—transition into
explosive electron emission)
mode with exterior heating of the cathode. Finally, the recent research attempted to
substantiate the ring effect by wave and diffraction mechanisms [66].
It is obvious that correct interpretation of field emission pre-explosive effects,
including the ring effect, can be based only on consistent concepts disclosing the
physical mechanism of these effects. However, the above-said makes it clear that
there is no single opinion in this respect even when the researchers analyzing the
experimental facts and developing the hypotheses deal with the same experimental
results.
In our opinion, papers [67, 68] suggested a physical analogue and a mathe-
matical model accompanying these phenomenon effects and consistently explaining
the ring effect on the basis of the experimental data on pre-explosive effects.
When constructing the physical analogue, authors [67, 68] proceeded from the
quite definitely established experimental fact that in the pre-explosive operating
mode of the field emission cathode, its surface is heated up by through emission
current to the temperatures close to the surface-melting point, this temperature
being usually 2/3 of the material bulk-melting temperature, which results in a
liquid-metal film on the metal surface (Fig. 4.25a). After development of the film
according to our model, the electric field applied between the emitter and the anode,
i.e. extractor, develops the so-called Taylor cone [68] and draws a nanotip with the
curvature radius of about 10−9 m on the apex from the liquid metal (Fig. 4.25b),
which explodes in the field enhanced by it up to *1010 V/m, thus generating anode
plasma near the emitter apex, which is a source of both electrons and positive ions.
The resulting ions are focused on the cathode apex. They press on the liquid metal
and push it to the periphery of the emitting surface, thus forming a ledge on it
(Fig. 4.25c). This ledge increases electric intensity near its apex, thus amplifying
emission from this area, which cannot but manifest itself in the emission image as a
bright ring.
(a) (b) (c) O
Fig. 4.25 Model of emission surface evolution for metal field emission cathode that explains the
development of a ring on the emission image: a apex with a liquid-metal film; b development of
the Taylor cone and nanotip on the apex; c forming a ledge on emission surface; point O area of
the best ionization
When making a mathematical model adequate to the formulated physical

analogue,
1. the ledge on the periphery of an emitting surface is considered torus-like
(Fig. 4.25c).
2. the torus radius is assumed to be much larger than the ledge curvature radius,
which simplifies the calculations by solution of the flat problem, because when
computing the torus in section can be approximated by the Gauss curve.
3. treatment of the liquid metal as incompressible assumes that the torus radius and
the area outlined by the Gauss curve are constant; hence, the volume of the
metallic liquid is held constant, too.
4. the field ionization current is assumed to be equal to a constant value.
5. increase in intensity between the emitter and the extractor is considered to leave
the torus radius unchanged, i.e. constant, and the effective area of ionization
(point O in Fig. 4.25c) recedes, i.e. the plasma moves from the cathode to the
anode.
The necessary condition of field ionization is preservation of electric intensity E0
≅ Vi/d that ionizes the atoms of the liquid metal, Vi being the metal atom ionization
potential, d being its effective radius.
As mentioned above, the torus section of the ledge is approximated by the Gauss
curve, taking into account that S = const, where S is the area outlined by the Gauss
curve, which can be written down in the following way:
pffiffiffi
y ¼ A a expðax2 Þ; ð4:44Þ
where A is a stationary value depending on the quantity of the liquid metal; a is the
parameter shown further to depend on the value of the voltage applied to the
emitter. As R r, R being the torus radius, we believe that there is only one
curvature radius r, namely the curvature radius of the Gauss curve
pffiffiffi
½1 þ ð2Aa ax expðax2 ÞÞ2 3=2
r¼ pffiffiffi :
2Aa að1 2ax2 Þ expðax2 Þ
The curvature radius at the ledge apex is
1
r0 ¼ rjx¼0 ¼ pffiffiffi : ð4:45Þ
2Aa a
The surface of the liquid metal has to meet the following condition of balance for
pressure forces:
e0 E 2 Y I 0 pa
¼ þ ; ð4:46Þ
2 y eXR2
where E is the electric intensity on the surface of the liquid ledge; Y is the surface
tension; I0 is the field-ionization ion current bombarding the emitter; pa is the
impulse of the ions bombarding the emitter; X is the solid angle of the ionic beam;
R is the torus or crater radius.
In the area of field ionization, the electric intensity should be constant and equal to
Vi U0
E0 ¼ ffi ;
d R
where U0 is the voltage initiating the pre-explosive field emission from the ledge
area; R is the torus or crater radius approximately equal to the virtual-emitter radius.
For the voltage between the anode and the cathode we will assume the following
expression true:
U0
U¼ h; ð4:47Þ
R
where h is the distance from the emitter to the field-ionization area, i.e. the point
O in Fig. 4.25c. The energy-conservation law and ratio (4.47) give us the impulse of
the ions bombarding the emitter:
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
pa ¼ 2eU0 mðn 1Þ;
U ð4:48Þ
n¼ :
U0
The solid angle of the ionic beam is

h
X ¼ 2pð1 cos hÞ ¼ 2p 1 pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
h2 þ R2
or
!
n
X ¼ 2p 1 pffiffiffiffiffiffiffiffiffiffiffiffiffi
n2 þ 1
according to (4.47) and (4.48).

In our opinion, the electric intensity near the apex of a ledge can be calculated if
the ledge is approximated by an infinite cylinder placed into a homogeneous electric
field, i.e.
E1 r1
Effi þ E2 ; ð4:49Þ
r
Here, E1, r1 are the electric intensity and the curvature radius pf the ledge apex at
initial moment of pre-explosive emission, i.e. emergence of the liquid ledge,
respectively; r is the resulting curvature radius of the ledge; E2 = E0n is the electric
intensity at the bottom of the ledge.
With stepping up of the voltage between the cathode and the anode, the
ledge-curvature apex radius becomes smaller, and the electric intensity at the ledge
apex builds up according to (4.49), which leads to current growth according to the
Fowler–Nordheim formula:

109
j ¼ 107 ðE1 r1 KðnÞ þ E0 nÞ2 exp :
E1 r1 KðnÞ þ E0 n
Assuming that stepping up the voltage does not change the area of emission, the
emission current from the ledge apex can be expressed as
I ¼ jðnÞpRr0 :
Finally, it is necessary to note that

1. The physical analogue suggested is based on the assumption of essential
influence of the ion beam generated together with the electron beam by
explosion of the nanotip, i.e. the protruding part of the liquid metal at the vertex
of the main apex made of a high-melting substance (e.g. W), on the processes
taking place in the pre-explosive operating mode of the field emission cathode
and, consequently, on the evolution of the emitting surface in superstrong fields.
2. Despite the estimate nature of the computations made, the resulting theoretical
volt-ampere characteristic constructed in Fowler–Nordheim coordinates is
qualitatively similar to those experimentally observed for the area of superstrong
fields and extremely heavy currents, in particular the theoretical curve, as well as
the experimental one, deviate towards smaller currents.
If the physical analogue suggested is true, the experiment can be expected to
generate a ring on the emission-image periphery under take-off of a considerably
smaller emission currents; firstly, with increase in the pressure of residual gases
(when the vacuum of the experimental device softens); and secondly, during
application of a film of a fusible metal, such as Au, In, Ga etc., on the field emission
cathode made of high-melting metal, such as W, Mo, Ta etc.
Almost direct confirmation of the mechanism suggested was really obtained [56]
by coating of the emitter made of a high-melting material, usually tungsten, with a
layer of a substance with a significantly lower melting temperature. Then emer-
gence of rings caused the layer to change into a liquid state. At that, the initial
temperature of the tungsten apex can be varied.
Let the pulse mode and initial temperature T0 of the tungsten-apex vertex coated
with a layer of a lower-melting material generate the ring at emission current I, and
temperature T00 does the same at current I′. Let us believe that the ring generation
causes the coating to change from a solid to disordered phase due to additional
Joule heating by the flowing emission current. Then disordered-phase temperature
Tliq, and, actually, the thin surface-film melting temperature can be determined as
ðI 0 =IÞ2 T0 T00
Tliq ¼ : ð4:50Þ
ðI 0 =IÞ2 1
For particular computations, let us first take advantage of all data of papers [56,
66]. Their authors saw the ring on the barium- or thorium-coated tungsten apex. The
pulse mode and the apex temperature of T0 = 1000 K generate the ring in the
emission image at current I ≅ 650 µA. Termination of the heating or reduction of
the current make the ring disappear. Without heating, i.e. at the initial temperature
of T0 = 300 K, the ring re-appeared at I ≅ 1300 µA, the coating being thorium.
Insertion of numerical values I, I′, T0 and T00 into expression (4.50) results in Tliq ≅
1233 K. According to [66], the film-melting temperature Tfilm is (2/3) Tbulk, where
Tbulk is the bulk-material melting temperature. For thorium, Tbulk = 2023 K, so Tfilm
≅ 1350 K. Taking into account the inexactitude of values I and I′, coincidence of
Tliq and Tfilm can be considered quite good. Iteration of the experiments in a more
careful way showed the value of the current ratio being I/I′ = 1.78. So Tliq = 1320 K,
which is about 98% coincidence with Tfilm. It is clear that if the initial temperature is
lowered e.g. by 200–300 K, in any ring-generation mode, then the surface layer
cannot change into the disordered phase, and the rings in the emission image have
to disappear, which is really the case.
An obvious way to reduce the self-heating effect of the field emission cathode is
its cooling to ultra-low temperatures, e.g. that of liquid helium.
There is a comparatively inconsiderable number of experimental papers on field
emission of metals at low temperatures that need theoretical substantiation or
evaluation. The research was mostly focused on field emission from the super-
conductive (s) state and detection of the energy gap in transition into the normal
(n) state [69]. According to paper [70], transition from the superconductive into
normal state connected with the change of energy of about kTs, Ts being the
superconductor state transition temperature, could manifest itself in changes of the
field emission current. Dependence of field emission current on voltage was studied
at helium temperatures ranging within 4.4–2.2 K.
It was shown there was no difference in field emission current for a tantalic
emitter of the temperature both above and below Ts within the experimental
accuracy estimated as nearly 0.2%. The Fowler–Nordheim dependence looked in a
way usual for that current density. The tantalic apex of 0.57 µm in radius developed
current density jfe = 80 A/cm2. That was whence we concluded that the potential
barrier on the surface of a superconducting metal and the chemical potential of
conduction electrons do not vary up to the values exceeding kTs.
Measurement of field emission current increment for a niobium emitter in a
superconductive (at 4.2 K) or normal (at temperatures above Ts) state was attempted
in paper [69]. The current increase caused by a rise in temperature having been
excluded, the current increase connected with s–п transition turned out less than the
measurement error. Therefore it was suggested that field emission might make the
field penetrate inside the metal by about 1 Å, which terminates the superconductive
state on the surface.
Septiers and Bergere [71] underscored the importance of the Nottingham effect
in thermal phenomena at low temperatures. A superconducting apex was shown to
change to a normal state with enhancement of energy generated due to the
Nottingham effect at rather small emission current, as small as 90 µA.
Alexeevskiy [72] studied the influence of emission current on the critical
parameters of a superconductor. A Ta-wire cathode was energized with current Js,
after which voltage drop V on the wire was measured. Voltage U was applied
between the cathode and the niobium-disk anode and emission current Je was
running. At Je = 0, the cathode was in a superconductive state and V = 0.
Achievement of some emission current was seen to carry the cathode into a normal
state, value V becoming finite. Further build-up of U increased V up to the maxi-
mum value Vn, leading to the entire cathode being in a normal state. Dependence of
ratio V/Vn on value U at different temperatures is shown in Fig. 4.26. N.E.
Alexeevskiy attributed emergence of the normal-phase area both to local heating
and some other possible reasons. In particular, the latter include emission-caused
reduction of the number of carriers taking part in superconducting current or the
mechanical voltages reducing the critical current.
So, research of field emission from the cathodes that initially had the temperature
below critical in stable experimental environment detected no specificity in
voltage-current characteristics or emission images suggesting that field emission
Fig. 4.26 Voltages: V/Vn

V voltage on the potential 1
conductors, Vn voltage in
normal state
0,5
0
4,5 5,0 5,5 6,0 6,5 7,0 V, kV
current takeoff is provided by the apex emitter in the superconductive state. At the
same time, no data on field emission current density limits at low temperatures is
available in the literature.
As already noted, the effect of increase in j0 at low temperatures is connected
with expansion of the Nottingham heat-release area referred to as “smearing”. At
that, it is necessary to differentiate between the essence of the Nottingham effect and
the mechanism of its implementation. The essence of this effect is known to be
connected with a difference between the average energy of the emitted electrons and
the average energy of the electrons coming from the external circuit, which can
result in both cooling and heating of the field emitter as the case may be. In this
context, let us consider the Nottingham heat-release mechanism coming down to
the following. Field emission injects a hot hole into the cathode (Fig. 4.27a). The
hole scatters the electrons that pass to the states lying above the Fermi level
(Fig. 4.27b). The electrons thermolize over the period see, which is a typical time of
electron-electron interaction. It is worth noting that the excess energy has not been
transmitted to the lattice yet, i.e. the Nottingham effect as a whole has not been
implemented by the time of thermolization yet. To transmit the excess energy to the
lattice, the electrons should interact with phonons, which is really the case as a
result of electron-phonon interaction over the characteristic period of sef and the
characteristic length of kef. Electrons are thermolized only if the condition of see
(a) (b)
λef
− + +
Fig. 4.27 Nottingham heat-release mechanism

sef is true. At low temperatures, the area where energy is transmitted to the lattice
(kef * tФsef, tФ being the velocity of electrons on the Fermi surface) can be as thick
as a few parts per millimeter in pure metals. It also suggests that the Nottingham
heat release is voluminous under these conditions (delocalization of the Nottingham
effect).
Total theoretical analysis of such a complicated situation can only be carried out
within the framework of the kinetic theory. But assuming that the condition of see
sef has been met, metal can be considered as a set of two subsystems, an electron
and a lattice one.
The energy-balance equations for these subsystems looks as follows:
@Te
Ce ¼ rðje rTe Þ AðTe ; Ti Þ þ jf ðrÞ; ð4:51Þ
@t
@Ti
Ci ¼ rðji rTi Þ þ AðTe ; Ti Þ: ð4:52Þ
@t
The Te, Ce, je of these equations are the temperature of electrons, the electron
heat capacity and the thermal conduction, respectively; Ti, Ci, ji are the respective
values for the lattice subsystem; A(Te, Ti) is the function featuring the nonlocal
transmission of energy from the electron subsystem to the lattice. The last term of
(4.51) takes account of Nottingham heat-release in the electron subsystem over
length kee.
Despite some incorrectness of such an approach, appropriate selection of func-
tions A and f can be expected to give adequate representation of the basic features of
the phenomenon of “smearing” of electron-electron interaction over length kee and
electron-lattice interaction over length kef. Functions A(Te, Ti) and f(r) were selected
as
Z
j r xj 3
AðTe ; Ti Þ ¼ a U½Te ðxÞ; Ti ðrÞ exp d x; ð4:53Þ
kef

l
f ðrÞ ¼ b exp ; ð4:54Þ
kee
where U (Te, Ti) is an ordinary local expression of the lattice-energy transmission

velocity; a and b are the constants determined from the condition that if kee ! 0
and kef ! 0, ratios (4.53), (4.54) passed into local ratios; l is the distance from point
r to the emitting surface. The boundary conditions to system (4.51)–(4.54) are
constancy of temperature in the emitter bulk and absence of temperature gradients
of normal n to emitting surface r:
Ti ðr ! 1Þ ¼ Te ðr ! 1Þ ¼ T0 ; ð4:55Þ

@Ti @Te
¼ ¼ 0; f jr ¼ jðTe ; EÞ ð4:56Þ
@n r @n r
with initial conditions
Ti ðt ¼ 0Þ ¼ Te ðt ¼ 0Þ ¼ T0 ; ð4:57Þ
where T0 is the initial temperature, j(Te, E) is the known function of electric

intensity E and Te on the emitting surface.
If we take account of smearing of only electron-electron interaction, assume A to
be A(Te, Ti) = v(Te − Ti), neglect the electron heat capacity for a one-dimensional
cathode, then we can derive
pffiffiffi
B expð cnÞ f
Te T i ¼ pffiffiffi e ; ð4:58Þ
1c c
where c = k2eem/je; n = x/kee; B = keejD/ji;D is the average Nottingham energy per

electron released in the electron subsystem when tunnelling.
If kee (je/m)1/2 d, then
Dj
T e Ti ¼ expðx=kee Þ;
mkee
otherwise, when kee d,
Dj
Te Ti ¼ expðx=dÞ:
md
These ratios show that if kee d, the characteristic electron-gas heating area is
*d. If kee d, the heating area is dramatically expended, and the electron-gas
temperature is much lower than in the first case. Consequently, the lattice-heating
velocity plummets:
Dj t
Ti T0 : ð4:59Þ
Ci kee þ d
So account of only smearing of thermal emission in the electron subsystem over

length kee results in much lower lattice-heating velocity.
Combined (4.51), (4.52) were numerically integrated for a coniform emitter. The
computation was made for T0 = 4.2 and 300 K. Value E was selected so that
transition to fast growth of current took sи = 10 ns.
Figure 4.28 shows the calculated dependences of emission-current density jfiel-em
on time. Their analysis shows that value j0 is 2.3 times larger at T0 = 4.2 K than at
T0 = 300 K. If kef goes up, this ratio somehow increases. But apparently, value kef
Fig. 4.28 The calculated jFEE, A/cm2

dependences of
emission-current density
jfiel-em (rk = 0.2 lm, a0 = 20°)
9
on time: 1 T0 = 4.2 K; 2 T0 = 10
1
300 K
8
10
0 4 8 t, ns
cannot considerably exceed rj/a0, rj being the radius of the emitting surface, a0
being the angle of the emitter cone.
For T0 = 300 K, jfiel-em does not continuously increase over the impulse of 10 ns,
which seems to be caused by heating of the emitter (Fig. 4.28). jfiel-em begins to
grow at the end of the impulse for initial temperature T0 = 4.2 K, which can also be
connected with the fact that the cathode achieved a certain temperature that seems
to be not very much different from 300 K. That is basis for Graph 1 in Fig. 4.28 to
evaluate the ratio of su =su as about 13.6% for sи = 10 ns, which matches the
experimental results of su =su = 13.3%. Once again, this circumstance proves
adequacy of the model suggested.
It is worth noting that the solution of a purely thermal problem shows that for
tungsten the current density limit cannot be increased more than 1.45 times without
account of size effects due to cooling of the emitter.
Increase in current density limit strongly depends on selection of a right emitter
material. The cathode material usually includes a basic chemical element and some
additives, as well as some imperfections/defects of the crystal lattice. At ultra-low
temperatures, it becomes crucially important that electrons are scattered by crystal
lattice defects, namely isolated pin-hole defects, such as vacancies, interstitial
atoms, chemical additives, isotopes etc.
In case of a single-crystal field emitter, the major factor influencing the free-path
length kef of electrons in the cathode at helium temperatures are isolated pin-hole
defects. At that, calculation shows that
kef ¼ a=v; ð4:60Þ
where a is a constant of the crystal lattice of the solid-cathode material, v = Nadd/

Naт (Naт being the number of atoms per metal-bulk unit, Nadd being the number of
additives and inhomogeneities per bulk unit). That means that v determines purity
of the cathode material.
As it was already noted, the density limit of the field emission current taken off
from the apex cathode without its rupture within a certain time interval can be
substantially increased if the heat released in the course of current take-off from the
cathode to destroy the cathode upon reaching a certain value, originated not near the
surface of the cathode, but in the depth of its bulk, the phenomenon being referred
to as smearing of the Nottingham effect. This condition is met if the free-path length
of electrons in the cathode at the working temperature is 10 times longer than the
cathode radius. Actually, this requirement is fair for apex cathodes with very small
cone-vertex angles. Apex cathodes with arbitrary cone-vertex angles should meet a
softer constraint of
rj
kef
; ð4:61Þ
a0
where rc is the apex-cathode radius, a0 is the cone-vertex angle.

So the effect of increase in the field emission current density limit can be
implemented only on condition that a/v
rj/a0. This yields the requirement of
purity of the cathode material as
aa0
v : ð4:62Þ
rj
4.6 Special Properties of Field Emission in Presence

of Strong Electric Fields and High Current Densities
Special properties of field emission in presence of strong electric fields and high
current densities were relatively fully analyzed in the introductory chapter as well as
in book by Zhukov [73] and an overview by Fursey (cf. [56], Chap. 6, Sect. 6.4)
and possible influence of simultaneous influence of spatial charge and proper
magnetic field or spatial charge and external magnetic field on field emission
characteristics was assessed by Egorov theoretically [74] and experimentally [75] as
well as addressed in [17, 76, 77].
The beginning of experimental research of field emission in presence of strong
electric fields and high current densities are connected with work of Dyke et al.
[44–60, 78–80]. In presence of those fields Dyke and Trolan [44] discovered sig-
nificant inconsistencies with the Fowler–Nordheim law in form of much weaker
than predicted dependence of emission current on applied potential difference
(Fig. 4.29).
Dyke and Trolan [44] interpreted the violation of linearity of volt-ampere
characteristics as an effect due to influence of spatial charge of emitted electrons
that weakens the field strength at the cathode surface.
The first attempt to theoretically assess the possible influence of spatial charge
onto field emission was made by Stern, Gossling and Fowler [81]. Calculation of
the theoretical relation lg(j) = f(V) (where j is the current density and V is voltage) is
4.6 Special Properties of Field Emission in Presence of Strong … 215
Fig. 4.29 Experimental lnI(I, A)

dependence ln I = f (V), where E
I is the emission current -4
strength and V is the voltage: C
AC is the segment of linear
volt-ampere characteristic; CE -8
is the area of abnormal
behavior of field emission
-12
A
-16
-20 4 -1
1.0 2.0 3.0 10 /V, V
based on solution of Poisson equation for potential distribution between infinite

planar electrodes. The Poisson equation in this case can be represented as:
d2 U=dx2 ¼ kjU 1=2 ð4:63Þ
where U is the potential; x is the coordinate; k = 2p(2me/e)1/2; j is the current

density. The solution of (4.63) with boundary conditions

dU
Ujx¼0 ¼ 0; Ujx¼d ¼ Ua ; ¼0 ð4:64Þ
dx x¼0
leads to a well-known Boguslavskiy–Langmuir (Schottky–Gleichung) equation:
j ¼ ð4=9kd 2 ÞUa3=2 ; ð4:65Þ
where d is the distance between electrodes.

For field emission the field strength E at the surface of emitter is non-zero and
the boundary conditions (4.64) can be represented as:

dU
Ujx¼0 ¼ 0; Ujx¼d ¼ Ua ; ¼ E; ð4:66Þ
dx x¼0
integrating (4.63) for boundary conditions (4.66) the authors of [81] obtained the
relation between field strength, current density and potential:
h i
E ¼ ðUa =d Þ 1 16=3pðme =2eÞ1=2 jUa1=2 =E 2 : ð4:67Þ
Without taking into account the spatial charge E = Ua/d, were Ua is equal to voltage
on the anode, criterion of smallness of influence of this charge according to (6.21)
can be written as
16=3pðme =2eÞ1=2 jd 2 =Ua3=2 1: ð4:68Þ
In more detail this process is analyzed in [82]. Same as before, the authors based
their conclusions on (4.63). Validity of chosen one-dimensional approximation of
real tip-cathode diode configuration is justified by the fact that the spatial charge
influencing the field close to emitting surface can be viewed as concentrated at a
very small distance from the cathode. Solving (4.67) together with Fowler–
Nordheim equation,
j ¼ AE 2 expðB=EÞ ð4:69Þ
where A = 1.54 10−6/u; B = 6.83 107u3/2t(y). Eliminating the current density j

from (4.67) and (4.69) one can obtain an equation connecting field strength E with
potential U:
4kAUa3=2 expðB=EÞ 3=Ua ¼ 9k 2 A2 E 2 d 2 expð2B=EÞ 3Ed: ð4:70Þ
Together with (4.69) this equation allows to calculate the current density for several
fixed values of E depending on the applied voltage. With small E, (4.70) leads to a
normal interdependence of field and potential for zero current (E = Ua/d). With very
big E the exponential parts become dominant and it’s easy to see that (4.70) is thus
reduced to (4.65).
The parameter d in [82] was chosen as equal to 1/b, where b is the geometrical
factor connecting the field strength with voltage E = bV.
Fig. 4.30 Theoretical

10 D E
dependence of current density
on applied voltage
8 B
C
6
0
A
-2
0 2 4 6
108/βV, cm/V
Figure 4.30 shows a plot of dependence of emission current density on voltage.

The curve ACE is calculated according to (4.69) in absence of spatial charge; ACD
is based on (4.69) and (4.70) with taking the spatial charge in account; BD is the
Boguslavskiy–Langmuir law (4.65). As one can see, for small V the initial part of
the curve corresponds with the Fowler–Nordheim line. With increase of V the curve
starts to tend downwards and eventually asymptotically approaches the
Boguslavskiy–Langmuir curve.
The influence of the spatial charge on field electron emission for a spherical
diode was studied by Eisenberg [83, 84], Kompaneets [64, 85] and Poplavskiy [86].
However, the results of their computation just confirmed the general tendency of
influence of spatial charge discovered via calculations of volt-ampere characteristics
for a flat field emission diode: qualitatively it is full correspondence with minute
quantitative differences.
The results presented below were obtained with assumption that field electrons
in the gap between cathode and anode move at speed v c.
It is rather obvious that movement of electrons in the cathode-anode gap with
relativistic velocities (v c) would have led to a much stronger dampening of
emission. In [87] it was indeed shown that for a planar model of diode the voltage
needed to support the given current density with taking into account the influence of
spatial charge of relativistic electrons grows several times as compared to the case
without accounting for relativist effects.
The approach proposed in [71] seems more strict. It is based on solving the
Poisson system accounting for the energy conservation law and continuity equation
for current density in spherical diode model:
r2 U ¼ 4pqe ; ð4:71Þ
me c2 þ eU ¼ me c2 ð1 v2 =c2 Þ1=2 ; ð4:72Þ
D~j ¼ 0 ð4:73Þ
with boundary conditions

dU
Ujr¼rk ¼ 0; ¼ E:
dr r¼rk
The emission density j3 ¼ jjr¼rk is defined by the Fowler–Nordheim (4.69). Solving

this system leads to
( )
1þg pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 1þg
i ra
DUm ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ln 2ð2 þ gÞ þ ln gð2 þ gÞ ð1 þ gÞ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ;
c gð2 þ gÞ rk gð2 þ gÞ
ð4:74Þ
where rk is the cathode radius and ra is the anode radius; DUm same as (4.70) is an
error taking in account the influence of spatial charge on field emission, but this
time in a relativist conditions and η eErk/mec2.
Equation (4.74) same as (4.70) is obtained assuming that ra rk and its
applicability is limited with condition
jDUm =U0 j 1: ð4:75Þ
where U0 is the potential of spherical anode without accounting for spatial charge, it
is connected with field strength E by relation
U0 ðra Þ ¼ Erk ½1 ðrj =ra Þ: ð4:76Þ
In accordance with (4.69) and (4.74–4.76), for each defined value of E there is a
defined value of current density je and a defined value of anode voltage Va. Using
this relations one can construct non linear volt-ampere characteristics of spherical
diode in coordinates lg(j)(1/Va). Same as before, the approximate character of the
solution limits the construction to the initial part of disconnection from Fowler–
Nordheim line.
Assuming the value of current that differs twofold from the one calculated by
Fowler–Nordheim formula (without accounting for spatial charge and relativist
effects) with the same applied voltage as the criterion of start of disconnection, one
can obtain explicit definition of disconnection current density from (4.74–4.76).
" #
ln 2E 3=2 ð2e=me Þ1=2 1 ra
ln jd ¼ ln ln rk ln ln 4 1 : ð4:77Þ
4pð2 þ Bu3=2 Þ=E 2 rk
Dependence of the disconnection current on emitter radius with two different

values of work function is presented at Fig. 4.31.
The volt-ampere characteristic begin to deviate from the line the earlier the
greater the emitter radius and smaller the cathode material work function is. The
Fig. 4.31 Dependence of the jd, A/cm2

current density for which
there is an observable
influence of electron spatial
charge onto the field emission 10 6
on cathode radius: 1 u = 4.5
eV; 2 u = 2.5 eV
1
10 5
2
10 4
10-4 10-3 rc, cm
computations show that the influence of relativist effects is necessary to take into
account with emitter radius greater that 10−3 cm. It is interesting to note that
whenever a large amount of emission current flows through a spherical diode
I [ 1:7 104 A; ð4:78Þ
solution of the system (4.71–4.73) is greatly simplified and an approximate equa-

tion can be written for U:
U ¼ ðI=cÞ½lnðra =rj Þ 1; ð4:79Þ
from which follows that the amount of flowing current is not dependent on the
cathode material work function. Field emission is fully limited with the spatial
charge.
Volt-ampere characteristics calculated in [71] are presented on Fig. 4.32. As one
can see from the figure there is a rather large area of the volt-ampere characteristics
where growth of the current with increase of the voltage is considerably slower than
on the initial linear area (increase of characteristic slope with large voltages is
connected to the scale effects).
The slowing of the current growth is especially noticeable with rk
10−3 cm
and current densities je > 107 A/cm2. Dampening of emission by spatial charge is so
large in these conditions, that the maximum current densities the given field
emission cathode can potentially provide with small influence of spatial charge
Fig. 4.32 Volt-ampere j, A/cm2

characteristics of a spherical 1
diode with tungsten field 10 8
emission cathode constructed 3 2
with accounting for influence 4
of spatial charge of emitted
10 7
electrons and relativist effect
(u = 4.2 eV): 1 Fowler–
Nordheim line without spatial
charge, 2 rk = 10−4 cm, 3 rk =
10−3 cm, 4 rk = 10−2 cm
10 6
10 5
10 4
0 0.5 1 1.5 108/βV, cm/V
become difficult to achieve (108–109 A/cm2). These current densities were exper-
imentally observed in impulse modes on thin tip cathodes (with rk = 10−4…10−5
cm). For example, for rk = 10−2 cm, in order to obtain je = 108 A/cm2 one has to
increase the voltage by more than an order of magnitude as compared to the value
of voltage computed without taking into account the spatial charge.
As the relation between the magnetic component of the Lorentz force and forces
of electrostatic repulsion of electrons in the beam is of the order v2/c2 and velocities
of electrons with megavolt-scale voltages are comparable to the speed of light,
accounting for the magnetic field can lead to grave corrections to the calculations of
volt-ampere characteristics of real emitter.
4.7 Modeling of Elements of Beam Formation and Control

Systems with Controlling Magnetic Field
(High-Precision Beams)
When constructing a mathematical model of high-precision emission diode it is

necessary to take in account not only the influence of proper spatial charge of
electron beam, but also probable influence of proper magnetic field of emitted
electron beam onto spatial charge. It is also necessary to note (when accelerating the
electrons up to and exceeding the energy of 100 keV) relativistic effects of the
electrons. One has to take into account this possible influence on field emission
characteristics of externally applied longitudinal magnetic field that is usually used
for focusing beams of relativistic electrons.
The problem stated above was solved in [74] for a spherical field diode
(Fig. 4.33): cathode was modeled by a hemisphere of radius r1 and anode with a
hemisphere of radius r2, where r2 r1. It was assumed that the current density
directly from the surface of spherical cathode is defined by the Fowler–Nordheim
equation. As the field strength is connected with potential by a relation E = −∇U,
Fig. 4.33 Model of a x

spherical field emission diode
U=U0
U=0
r1 H
0
z
E
cathode anode
y
4.7 Modeling of Elements of Beam Formation and Control Systems … 221
thus solving this problem can be reduced to solving the equation for potential. In
order to account for the influence of spatial charge in given model case during field
analysis, the Poisson equation in variables of spherical coordinate system is used.
Given the symmetry of the problem, it can be written down as:

1 @U 2 @U qðrÞ
r ¼ ð4:80Þ
r 2 @r @r e0
where e0 is the electric constant with boundary conditions Ujr¼r1 ¼ 0;

Ujr¼r2 ¼ U0 ; the function q(r) is found from the relation
qðrÞ ¼ jðrÞ=v;
and the function for j(r) in given case can be represented as
ðB=E0 Þð1 cos h0 Þ r1 2

jðrÞ ¼ jS0 ¼ AE02 exp : ð4:81Þ
1 cos hðrÞ r
In (4.81) A and B are known constants; h is the half-angle of divergence of the

beam; r1 is the cathode radius; S is the area of surface laying at arbitrary distance r
from the cathode; r and h(r) are the spherical coordinates of a boundary electron; its
velocity v can be found by solving the relativist motion equation that can be written
in vector form:
!
d mv
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ¼ Fðr; vÞ; ð4:82Þ
dt 1 v2 =c2
where F = −e{∇U(r) + [v, H]}; e, m are the charge and rest mass of the electron;
v is its velocity; H is the vector sum of the external and proper magnetic field
strengths of the beam. U(r) is the potential of electric field; c is the speed of light.
After differentiating of the left hand part of (4.82) and several elementary trans-
formations, it can be rewritten as:
@2r
þ AðtÞ ¼ cf; ð4:83Þ
@t2
where
rffiffiffiffiffiffiffiffiffiffiffiffiffi
@r v2 e v2 @v
f ¼ b rU þ ;H ; c ¼ 1 2; b¼ ; AðtÞ ¼ v:
@t c m c2 c2 @t
ð4:84Þ
When calculating the proper magnetic field of a particle beam only the azimuthal
component was taken into account. When finding trajectories of electron motion in
the nodes of mesh enveloping the area through which the beam passes, the spatial
distribution of currents Iij is memorized in it. Based on the assumption that the
current density is constant for each ij-th elementary cell, the formulas for proper
magnetic fields were constructed:
8 lI
>
< 2pð0;5hr Þ2 rT ;
0 1j
rT 0; 5hr1 ;
1
H0 ¼ l I1j r 2 ri1=2
>
: 2pr
0
Ii1;j þ ðIij Ii1;j Þ r2 T r 2 ; rT
0; 5hr1 ;
T i þ 1=2 i1=2
where ri ± 1/2 = 0.5(ri + ri ± 1); hr1 is the step of mesh close to the axis of symmetry;
Iij is the total current, i.e. the sum of all the currents
of all trajectories passing
through the j-th section of DSij ¼ p ri2þ 1=2 ri1=2
2
(close to the axis of symmetry
r 2
DSij ¼ 0:25p h1 ). In the initial moment the coordinates and velocities of a
boundary electron are defined:
rj0 ¼ r1 ; hjt¼0 ¼ h0 ; ujt¼0 ¼ 0;

r_ j0 ¼ v0 ; _ _ t¼0 ¼ 0:
hjt¼0 ¼ 0; uj
Thus solving of our model problem in stationary statement implies solving a

self-consistent system of Poisson equation (4.80), motion equation (4.82) and
equation of flow continuity: ∇j = 0 given the relations (4.83) and (4.84). As strict
solution of given problem is virtually impossible, successive approximation method
had to be employed. The calculations were performed as follows.
At the first step, one has to solve an elementary problem: using the Laplace
equation (i.e. assuming zero charge density), a first approximation of the field
distribution in field emission diode is computed (function U(r)).
The second step is obtaining the numerical solution of the system of motion
equations of a boundary electron exiting the cathode at the angle h0 subject to the
influence of stationary electric and magnetic fields using the Runge-Kutta methods
and finding a family of motion trajectories for the electron. Using the envelope of
this family one can find the spatial distribution of electron densities in the beam
(function q(r)).
At the third step, same as at the first one, one has to define the field distribution
U(r), but this time by analytical solution of Poisson equation with boundary
conditions.
Ujr¼r1 ¼ 0; Ujr¼r2 ¼ U0 ;
where U0 is the initial distribution of potential as found at the first step. For q(z) the
values found at the second step are used.
At the fourth step, same as at the second, we yet again find the family of
trajectories by solving the system of motion equations. This time, however, we use
U(r) (and thus also E(r)) found at the third step in the motion equations.
4.7 Modeling of Elements of Beam Formation and Control Systems … 223
Repeating those steps enough times one can compute the potential distribution U
(r) with any arbitrary precision. Using the known U(r) one can find the electric field
strength E0 at the cathode surface and thus also the current density j0 of the field
electron emission.
Solving the self-consistent problem of integrating of a system of differential
equations (steps 2 and 4) was performed using the Runge-Kutta methods as noted
above. It was assumed that the initial velocities of electrons were equal to the
thermal velocities. The step of integration for each step was connected to the step of
corresponding area of the mesh, and the origin of spherical coordinate system was
positioned at the emitter plane. The process of constructing the trajectory assumes
that for each next trajectory the place of the “zeroth” one in the computer memory is
occupied by the first trajectory, which leads to eventual growth of volume. Thus we
were able to avoid introduction of additional working arrays for storing the infor-
mation on the trajectories being constructed.
Results of computation and numerical experiment. Numerical experiment was
performed for both v(y) = 1 (i.e. tunneling of electrons through the mode triangular
barrier) and for v(y) 6¼ 1 (i.e. the barrier of “imaging forces”).
No qualitative differences between the cases v(y) 6¼ 1 and v(y) 1 were
observed: in both cases there were “pulsating” relationships lg j = f (1/E) (cf.
Fig. 4.34). If in the “limit” case the magnetic field strength was assumed zero, the
“pulsation” disappeared. In the same time there were noticeable quantitative dif-
ferences: for v(y) 6¼ 1 the pulsation amplitudes were significantly larger than for v
(y) 1.
Fig. 4.34 Theoretical 2 1

volt-ampere relationship of a lgj
spherical field emission diode, 3
calculated without taking into
account influences of the
7
spatial charge, proper
magnetic field of the emitted
electron beam or external
electric field (1); with only the
influence of spatial charge (2);
and with accounting for the
influences of spatial charge,
proper and external magnetic 5
fields (3)
3
1 2 E, 102V/cm
This is probably connected to the fact that while considering the values of
Nordheim function not equal to 1, the influence of proper spatial charges of elec-
trons (and thus also the proper magnetic field) in the vacuum cathode-anode gap
begin to manifest themselves at smaller values of emission-exciting electric fields.
In the same time, resulting influences of proper and external magnetic fields onto
the spatial charge distribution for the case of tunneling of the electrons through the
barrier of “imaging forces” must be more efficient.
Analyzis of the obtained results has shown that apart from that, the trajectory
slope and the size of “anode spot” vary dependent on the relation of E and
H. Character of non-monotonousness (e.g. the amplitude of pulsation of current
density) is significantly dependent on the angle h0, i.e. from the side of emitter
surface (or the emission area). There is a certain minimum threshold value of h0
(let’s call it “critical zone”), surpassing which leads to complete disappearance of
magnetic effects: h0 7–9°.
One also has to note that the results of numerical experiment are also dependent
on the chosen numerical method, its precision etc. For example, when integrating
the initial system of equation, due to closeness of all the characteristics of the beam,
it is very easy to cross the line where there will be variance of numerical solutions
from iteration to iteration rather than any “physical” pulsation effect. The other
possible reason of observed pulsations can be instability of the iterated process due
to monoenergeticity of the inbound beam which leads to emergence of defects of
q(r) at the area of virtual cathode. As such, even small fluctuations in this functions
cause large-scale changes in the density of electron flow, i.e. the problem becomes
unstable (i.e. incorrect by Tikhonov: the third condition of correctness in [88] is
broken). Thus all the results of mathematical modeling and numerical experiments
always need experimental confirmation. It is even more important given that in a
mathematical model it is all but impossible to consider all the different phenomena
that take place in real devices based on charged particle beams.
Experimental studies of influence of magnetic fields on field emission from
metals were undertaken in [75–77] and analyzed in great detail in an appendix to
[17]. For an instance, authors of [76] studying the influence of magnetic field on
field emission, claimed based on their results that they were able to observe
multi-electron field emission. However further research showed that “pure” field
emission from metals is always single-electron, i.e. no more than one electron can
be tunneled in one elementary act of field emission, and the multi-electron phe-
nomena observed in [76] are connected to the secondary effects such as superim-
posing of secondary electron and ion-electron emission onto field emission [75].
However absence of real-life effects (such as “pulsation” of field emission cur-
rent) in experimental observations of influence of magnetic field on the emission
process probably has to do with the fact that the magnetic field intensity in those
experiments did not exceed a few tenths of Tesla or sometimes a few Tesla. One
can’t overlook the possibility of emergence of the pulsation effects predicted in [74]
if stronger magnetic fields are to be applied.
References 225
References
1. A.Y. Antonov, V.P. Denissov. Phys. Scr. 59, 235–240 (1999)

2. A.Y. Antonov, V.P. Denisov, N.V. Egorov, Mathematical modeling of the passage of
electrons through a solid-vacuum boundary for small variations in the work function.
Poverhnost. (12), 116–118 (1999)
3. A.Y. Antonov, Mathematical modeling of electron transport across the border solid-vacuum.
The Control Processes and stability, in Proceedings of XXIX Scientific Conference of Applied
Mathematics and Control Processes Faculty of Saint Petersburg State University, St.
Petersburg (1998), pp. 115–118
4. A.Y. Antonov, The dependence of the electron energy states of the shape of the potential in
the semiconductor emitter. The Control Processes and stability, in Proceedings of XXX
Scientific Conference of Applied Mathematics and Control Processes Faculty of Saint
Petersburg State University, St. Petersburg (1999), pp. 203–207
5. A.Y. Antonov, An inverse problem of the transparency of the quantum barrier. The Control
Processes and stability, in Proceedings of XXXI Scientific Conference of Applied Mathematics
and Control Processes Faculty of Saint Petersburg State University, St. Petersburg (2000),
pp. 132–136
6. A.Y. Antonov, V.P. Denisov, Resonant tunneling in processes of electron emission from
semiconductors. Europhys. Conf. Abstr. 23G(ECOSS-18), Th-P-066 (1999)
7. A.Y. Antonov, V.P. Denisov, Potential barrier shape determination from experimental data.
Europhys. Conf. Abstr. 23G(ECOSS-18), We-P-060 (1999)
8. N.V. Egorov, A.Y. Antonov, M.I. Varayun, Determination of transparency of the potential
barrier by the variations of the external field. Poverhnost. (11), 107–110 (2004)
9. N.V. Egorov, A.Y. Antonov, M.I. Varayun, Isolation of the harmonic component of the field
emission signal. Poverhnost. (9), 94–98 (2005)
10. (a) V.P. Denisov. Pis’ma v ZTP. 18(5), 21 (1992); (b) V.P. Denisov, Pis’ma v ZTP. 18(14),
38 (1992)
11. L.I. Antonov et al. Pis’ma v ZTP. 11, 602 (1985)
12. L.I. Antonova, V.P. Denissov. Appl. Surf. Sci. 111, 237 (1997)
13. B.V. Thien, C. Adessi, New mechanism for electron emission from planar cold cathodes. The
solid-state field-controlled electron emitters. Phys. Rev. Lett. 85(4), 864–867 (2000)
14. M.I. Varayun, Mathematical modeling of the tip of the electron emission structures. Thesis for
the degree of Candidate of Physics and Mathematics Sciences, St. Petersburg (2002)
15. M.I. Varayun, A.Y. Antonov, V.P. Denisov, Mathematical modeling of a field emission
cathode, in Proceedings of the X International Meeting: Accelerators -2001. St. Petersburg
State University, St. Petersburg (2001), pp. 346–349
16. E.L. Murphy, R.H. Good, Thermionic emission, field emission and transition region. Phys.
Rev. 102(6), 1464–1473 (1956)
17. A. Modinos, Field, thermionic and secondary electron emission spectroscopy. M.: Nauka
(1990), 320 p
18. L.D. Landau, E.M. Lifshitz, Quantum mechanics. Non-relativistic theory, M.: Nauka (1974),
752 p
19. L.N. Dobretsov, M.V. Gomoyunova. Emission electronics, M.: Nauka (1964), 364 p
20. T. Pang, A numerical method for quantum tunneling. Comput. Phys. 9, 602–606 (1995)
21. K.-Y. Kim, B. Lee, Superlattices Microstruct. 24(6), 389 (1997)
22. P. Niedermann, N. Sankarraman, R.J. Noer, O. Fisher. J. Appl. Phys. 59, 892–896 (1986)
23. V.M. Zhukov, S.A. Polezhaev. Radiotehnika i elektronika. 33(10), 2153–2162 (1988)
24. A. Modinos, J.P. Xanthahis, Energy floating of field-emitted electrons due to Coulomb
scattering. Surf. Sci. 249(1–3), 373–378 (1991)
25. D. Nicolaescu, Physical basis for applying the F-NJ-E relationship to experimental I-V data.
J. Vac. Sci. Technol. B 11(2), 392–395 (1993)
26. A. Seidl, M. Takai, A. Hosono, S. Yura, S. Okuda, Geometry effects arising from anodization
of field emitters. J. Vac. Sci. Technol. B. 18(2), 929–932 (2000)
27. M.I. Elinson, G.F. Vasiliev. Field emission. M.: Fizmatgiz (1958), 272 p
28. N.V. Egorov, V.M. Zhukov, C.A. Polezhaev, Metal tip producing for scanning tunneling
microscopy, in Proceedings of 148. WE-Heraeus-Seminar: STM-Related Spectroscopies of
Semiconductor Interfaces, Bad Honnef (1995), pp. 5–7
29. N.V. Egorov, E.M. Vinogradova, Solution of boundary-value problem in bispherical
coordinates, in Proceedings of 3-th International Workshop: BDO-96, S.-Petersburg (1996),
pp. 274–278
30. G. Mesa, E. Dobado-Fuentes, J.J. Saenz, Image charge method for electrostatic calculations in
field emission diodes. J. Appl. Phys. 79(1), 39–43 (1996)
31. Y. Ohkavara, T. Naijo, T. Washio, S. Oshio, H. Ito, H. Saitoh, Field emission properties of
AlZnO whiskers modified by amorphous carbon and related films. Jpn. J. Appl. Phys. 40(12),
7013–7017 (2001)
32. K.L. Jensen, J.E. Yater, Advanced emitters for next generation rf amplifiers. J. Vac. Sci.
Technol. B. 16(4), 2038–2049 (1998)
33. K.A. Nikiforov, Mathematical modeling of the field electron emission from a metal-dielectric
systems. Dissertation for the degree of Candidate of Sciences, St. Petersburg (2005)
34. C.A. Spindt, C.E. Holland, P.R. Schwoebel, I. Brodie. J. Vac. Sci. Technol. B. 16, 758 (1998)
35. E. Mueller, T. Tson, Ion microscopy. M.: Metallurgy (1972), 360 p
36. J. Rena, S. Ranganathan (eds.), The field-ion microscopy. M.: Mir (1971), 210 p
37. A.L. Suvorov, T.L. Razinkova, V.A. Kuznetsov, Computers in field ion microscopy. Phys.
Stat. Sol. 61A, 11 (1980)
38. A.L. Suvorov, Structure and properties of the surface atomic layers of metal. M.:
Energoatomizdat (1990), 296 p
39. A.L. Suvorov, T.L. Razinkova, V.A. Kuznetsov, Computer simulation of field electro images.
Surf. Sci. 52, 697 (1975)
40. K.A. Nikiforov, N.V. Egorov, Modelling of the surface structure and the numerical
calculation of a current density of field emission metal cathode. Vestnik St. Petersburg State
University, Ser.10. Applied mathematics, computer science, management processes. 2, 39–45
41. I.M. Gribkova, Mathematical modeling of the image field in an electron microscope. Bachelor
work, St. Petersburg (2009), 37 p. I.M. Gribkova, Mathematical modeling of the image field
in an electron microscope. The Control Processes and stability, ed. by N.V. Smirnov, G.S.
Tamasyan, in Proceedings of XXXIX Scientific Conference of Applied Mathematics and
Control Processes Faculty of Saint Petersburg State University, Publishing House of St.
Petersburg state University, St. Petersburg (2008), pp. 116–121
42. N.V. Egorov, On the possibility of narrow collimated electron beams. ZTP. 52(12), 2440–
2442 (1982)
43. M. Drechsler, E. Henkel, Feldemissions-Stromdichten und Oberflachenfeldstarken bei
Feldemissionsmikroskopen sowie Methoden zur Bestimmung des Spitzenrauius, der
Spitzenform, der Vergrosserung und des Auflosungsvermogens. Z. Angew. Phys. 6, 341–
346 (1954)
44. W.P. Dyke, J.K. Trolan, Field Emission: large current densities, space charge and vacuum arc.
Phys. Rev. 89(4), 799–808 (1953)
45. W.W. Dolan, W.P. Dyke, Temperature and field emission of electrons from metals. Phys.
Rev. 95, 327–332 (1954)
46. G.E. Vibrans, Vacuum voltage breakdown as thermal instability of the emitting protrusions.
J. Appl. Phys. 35(10), 2855–2857 (1964)
47. V.M. Zhukov, Stability of the surface of the tip emitters when exposed to thermo and field.
Stability and control processes, ed. by D.A. Ovsyannikov, L.A. Petrosyanm in Proceedings of
the International Conference Volume 1: Section 1–5, 11. St. Petersburg State University,
Research Institute of the VM and PU, OOO VVM, St. Petersburg (2005), pp. 161–164
48. Y.V. Andrianov, V.N. Bazdyrev, D.A. Borisov, V.M. Zhukov, Spontaneous growth of field
emission current in a microwave field. ZTP. 61(9), 183–185 (1991)
References 227
49. V.L. Egorov, Y.V. Chentsov, Field emission cathodes in modern electron microscopes. In:
Proceedings of the State Optic Inst. L. Publishing House of the State Optic Inst, vol. 58,
pp. 68–87 (1985)
50. N.V. Egorov, K.A. Nikiforov, C-C. Shen, Reconstruction of the surface of the field electron
emitter. Poverhnost. X-ray Synchrotron Neutron. (10), 100–106 (2009)
51. K. Nikiforov, Simulation of the crystal structure of the surface of a metal cathode. The control
processes and stability, ed. by N.V. Smirnov, V.N. Starkov, in Proceedings of XXXV
Scientific Conference of Applied Mathematics and Control Processes Faculty of Saint
Petersburg State University, Publishing House of St. Petersburg state University, St.
Petersburg (2004), pp. 244–247
52. V.S. Fomenko, The Emission Properties of the Materials. Directory, 4th edn. Revised and
complementary (Naukova Dumka, Kiev, 1981), 340 p
53. L. Feldman, D. Meyer, Basics of analysis of surfaces and thin films. M.: Mir (1989), 344 p
54. L.A. Rudnicki, On the work function of the electron and the donor-acceptor properties of
non-ideal adsorption particles and coated metal surfaces. Math. USSR Acad. Sci. Ser. Nat. 26,
1240–1246 (1982)
55. V. Linnik, The method of least squares, and the theoretical foundations of mathematical and
statistical processing of observations. L.: Fizmatgiz (1962), p. 352
56. M.I. Elinson (ed.), The cold cathodes. M.: Sov. radio (1974), 386 p
57. T.I. Lewis, Theoretical interpretation of field emission experiments. Phys. Rev. 101(6), 1694–
1698 (1956)
58. W.B. Nottingham, Remarks on energy losses attending thermionic emission of electrons from
metals. Phys. Rev. 59(11), 907–908 (1941)
59. O.W. Richardson, Extraction of electrons from cold conductors in intense electric fields. Proc.
Roy. Soc. 117, 719–730 (1928)
60. W.W. Dolan, W.P. Dyke, J.K. Trolan, The field emission initiated vacuum area. II. The
resistively heated emitter. Phys. Rev. 91(5), 1054–1057 (1953)
61. V.A. Gorkov, M.I. Elinson, G.L. Yakovlev, Theoretical and experimental research in the
fore-arc phenomena of field emission. Radiotehnika i Elektronika. 7(9), 1501–1510 (1962)
62. A. van Oostrom, Dependence of the critical field strength for vacuum breakdown of tungsten
on the field emitting area, in Proceeding III International Symposium on Discharges and
Electrical insulation in Vacuum, Paris (1968), pp. 174–180
63. D. Lee, The influence of Nottingham effect on temperature of field emitting surface, in
Proceedings of Insulation of High Voltages in Vacuum, Boston (1965), pp. 96–98
64. A.S. Kompaneets, Effect of space charge on the field emission. DAN USSR 128(6), 1160–
1162 (1959)
65. Y.V. Zubenko, I.L. Sokolskaya, G.N. Fursey. ZTP. 34(5), 911–912 (1964)
66. N.V. Egorov, V.M. Zhukov, On the mechanism of formation of “rings” on the field emission
cathodes images. Abstracts, in 21st All-Union. Conference for emission electronics, L. (1990),
p. 240
67. N.V. Egorov, A.G. Karpov, Diagnostic Information and Expert Systems (St. Petersburg State
University Publishing House, St. Petersburg, 2002), 472 p
68. N.V. Egorov, B.V. Yakovlev, Evolution of the surface of liquid metal in a strong electric
field. Poverhnost. (8–9), 36–41 (1994)
69. R. Klein, L. Leder, Filed emission from Niobium in the normal and Superconducting State.
Phys. Rev. 124(4), 15–18 (1961)
70. H. Bergeret, A. Septier, Finding of cathode heating owing to Nottingham effect. CR Acad.
Sci. 277(17), 489–492 (1973)
71. L.M. Baskin, V.A. Godyak, O.I. Lvov et al., Effect of space charge of relativistic electrons in
the field emission. ZTP. 42(6), 1282–1287 (1972)
72. N.E. Alexeevskiy, The impact of the emission current on the critical parameters of tantalum.
DAN USSR. 242(4), 816 (1978)
73. V.M. Zhukov, Processes on the Surface Under Field Emission (VVM, St. Petersburg, 2007),
295 p
74. D.A. Ovsyannikov, N.V. Egorov, Mathematical Modeling of Systems of Formation of

Electron and Ion Beams (St. Petersburg State University, St. Petersburg, 1998), 276 p
75. N.V. Egorov, The physical model of the processes occurring in the field emission diode in the
magnetic field. Lett. ZTP 8, 1038–1041 (1982)
76. V.E. Ptitsyn, G.N. Fursei, N.V. Egorov et al., Egorov et al. Effect of magnetic field on the
field electron emission from tungsten. Lett. ZTP 5, 116–119 (1979)
77. V.E. Ptitsyn, G.N. Fursei, N.V. Egorov et al., The temperature dependence of the magnetic
field emission effects in metals. Lett. ZTP 6, 619–622 (1980)
78. P.H. Levine, Thermoelectronic phenomena associated with electron-field emission. J. Appl.
Phys. 33(2), 582–587 (1962)
79. I. Brodie, Temperature of strongly field emmiting surface. Int. J. Electron. 18, 223–232 (1965)
80. L.W. Swanson, L.C. Crouser, F.M. Charbonnier, Energy exchanges attending field emission.
Phys. Rev. 151(1), 327–340 (1966)
81. T.E. Stern, B.S. Gossling, R.H. Fowler, Further studies in the emission of electrons from cold
metals. Roy. Soc. Proc. A. 124, 699–723 (1929)
82. J.R. Barbour, W.W. Dolan, J.K. Trolan et al., Space-charge effects in field emission. Phys.
Rev. 92(1), 45–51 (1953)
83. N.B. Eisenberg, On the role of the space charge in spherical electronic projectors. ZTP. 24
(11), 2079–2082 (1964)
84. N.B. Eisenberg, On the influence of the space charge in the form of characteristics lnJ(1/Va)
of field emission cathodes. Radiotehnika i Elektronika. 9(12), 2147–155 (1964)
85. A.S. Kompaneets, Effect of space charge on the field emission. Radiotehnika i Elektronika.
V.5. 6 y 3JK56. N8, 1315–1317 (1960)
86. R.P. Poplavskiy, The potential distribution in a spherical capacitor in the case of saturation
current. ZTP 20, 149–159 (1950)
87. V.A. Godyak, L.V. Dubovoy, G.R. Zabolotskaya, Calculation of the field emission current
that limited by space charge. ZETP. 57(11), 1795–1798 (1969)
88. A.N. Tikhonov, V.J. Arsenin, Methods of solving of ill-posed problems. M.: Nauka (1974),
286 p
Chapter 5
Field Emission Cathodes
Abstract In the beginning of the chapter the main problems and obstacles to
widespread practical use of field emission cathodes are considered. Then it proceeds
to discuss the features, advantages and disadvantages of various types of field
cathodes known as of today.
5.1 Cardinal Problems of Field Emission Cathodes
The main point of practicality of any cathode as a free-electron source is that of

emission-current stability in time and space and cathode longevity. Field emission
is known to be extremely sensitive to cathode reshaping and surface condition.
A field emission cathode functions in an electronic device along with different
processes on its surface, including but not limited to ion bombardment; pondero-
motive loads; adsorption and desorption of residual-gas molecules; surface
migration etc. Depending on a particular design and an operation mode of the field
emission cathode, these processes separately or cumulatively cause a series of
effects modifying their operation mode: cathode sputtering, emitting-surface
reshaping, microelevation relocation and change in quantity, variation in the elec-
tron work function, cathode heating, mechanical stresses.
These effects can make the key parameters of field emission cathodes, e.g. the
emission-current value and stability, noise bandwidth etc., go beyond the limits
established by specifications for a concrete electronic device, which wrecks the
cathode.
That is why the multidimensional efforts to improve stability of field emission
current are made in the following areas of research:
1. Ultrahigh exhaustion in field emission cathode devices [1].
2. Heating of a field emission cathode to clean its surface and reconstruct its shape
[2].
3. Application of ion traps [3].
4. Use of pulse-operation modes [4].
5. Search of sputter-resistant materials [5].

230 5 Field Emission Cathodes
6. Generation of artificial residual-gas atmosphere [6].

7. Reduction of inter-electrode distance [7, 8].
8. Use of specific emission stability of field emission current of semiconductors
[9, 10].
9. Statistically stable microstructuring of emitting centers [11].
The problem of field emission current stability can be well solved either in the
above-mentioned well-known areas of research or in some radically new ones. In
this connection, it is necessary to analyze the potentials of the above-mentioned
areas of research in terms of field emission current stability improvement.
1. Let us consider the field emission current stabilization potential inherent in the
first, ultrahigh vacuum, area.
When a field emission cathode functions in ultrahigh vacuum, the effects caused
by ion bombardment of its surface, e.g. cathode sputtering, surface migration etc.,
nosedive. In this case, the characteristic dimensions of the field emission cathode,
e.g. the apex-tip radius, change extremely slowly. However, changes of the field
emission current are determined not only by reshaping of the field emission cath-
ode, i.e. its form factor b, but also by changes of its electron work function u.
Therefore, whereas the value of b is almost constant in ultrahigh vacuum for a long
period of time, the value of u continuously varies due to the low sorption rate. In
particular, if q * 10−12, the Tor process of stabilization of the electron work
function takes 300 h. (Duration of stabilization of the electron work function is
expressed through the time necessary for the adsorbed residual-gas atoms to
develop a monomolecular coating of the emitting surface of the field emission
cathode). Elementary estimates show that the tor process of stabilization of the
electron work function in technical vacuum at p * 10−7 takes a few seconds.
However, that is only one side of the coin. Progress in this area of research
aimed at field emission current stability improvement is obstructed by extreme
difficulties of both generation and maintenance of ultrahigh vacuum for a long time.
That is an almost insoluble problem in devices with a glass bulb quite easily
diffusible for helium [12].
The above-said makes it clear that even successful overcoming of technical
difficulties in the ultrahigh-vacuum area of research is not enough to solve the field
emission stability problem.
2. Let us consider the potential of the second field emission current stabilization
area of research: surface migration of atoms in both the field emission cathode
body itself and the residual gases. The research found that a strong electric field
makes the atoms of the field emission cathode body migrate to the apex center,
which thereby tapers.
Residual-gas ions energize and activate the migratory processes at a room
temperature. If the apex is heated without a field; it grows blunt as effected by
surface-tension forces. Alternation of operation and relaxation of a field emission
5.1 Cardinal Problems of Field Emission Cathodes 231
cathode accompanied by heating can extend its service life by emitting surface
reconstruction.
In practice, that operating mode of a field emission cathode is extremely
inconvenient because it involves expansion of the field emission cathode design and
the device in general and deprives it of some advantages over hot cathodes.
Surface migration of adsorbed residual-gas atoms in a strong electric field also
tapers the emitter, which results in growth of E and j up to breakdown. However,
the electrical breakdown of this case is special. It does not result in loss of the field
emission cathode. What the breakdown destroys here is adsorbed-atom micropiles,
which causes field emission current surges in its turn [13].
Migration of adsorbed residual-gas atoms can be excluded by heating of the field
emission cathode to clean its surface. In this case, tungsten has to be heated to
*1920 K [14], rhodium *1320 K and palladium *1120 K [15].
The area of research of field emission current stabilization in heater field
emission cathodes is not prospective because of the above-mentioned inconve-
niences of practical implementation.
3. Let us consider the field emission current stabilization potential coming from
application of ion traps and electrostatic gates.
Such device designs were first suggested by M.I. Elinson et al. and thoroughly
considered in monograph [6]. The research shows that such devices can consid-
erably reduce the number of ions bombarding the working section of a pointed
emitter, thereby extending its service life. However, it is not always possible to
resort to ion traps and gates in practice even if the inter-electrode distances are
relatively large because they significantly complicate the design of the cathode
assembly. Such designs have not been widely used yet.
4. The pulse operation mode considerably reduces average current loading of the
field emission cathode and the thermal loading of the anode. With other things
being equal, the mean power released under the pulse operation mode is sq
times less than under the continuous one, being the pulse length in seconds and
q being the number of pulses per second.
Numerous experiments testify that field emission current is much more stable
under the pulse mode than under continuous one. However, use of field emission
cathodes only under the pulse operation mode prelimits the potential area of their
application.
5. Sputter-resistant substances suitable for making field emission cathodes are
looked for continuously, e.g. [16–21].
Especially careful research covered high-strength metals, usually tungsten,
rhenium, tantalum and metal-like compounds of the carbide and boride group.
The research shows that field emission cathodes made of a series of substances,
e.g. rhenium, lanthanum hexaboride, provide a more stable field emission current
than those of tungsten.
In our opinion, this area of research of current stability improvement is

prospective and has to be extended with search of substances with a stable electron
work function in technical vacuum.
6. M.I. Elinson was the first to point out a possibility of artificial residual-gas
atmosphere, light gases, i.e. with hydrogen and/or helium prevailing, in
sealed-off devices. The light-gas ions are known to be ineffective in terms of
cathode sputtering. «…In this context, a spontaneously generated helium
atmosphere does not seem to be too threatening…» [6].
Generation of such an atmosphere in practice is fraught with extreme difficulties.
Under real conditions, the electrovacuum device parts and their shells slowly, but
continuously gas even at a room temperature, thereby polluting the residual-gas
atmosphere, which is almost uncontrollable in this case. First of all, that pertains to
the inter-electrode space volume with its microclimate determined by gas evolution
and adsorption taking place on the device-electrode surfaces.
7. Let us consider another area of research of field emission current stability
improvement, the one based on reduction of inter-electrode distance
R. Reduction of R in electrovacuum devices causes reduction of the number of
ions generated in the inter-electrode space volume, resulting in decrease in ion
bombardment of the emitter. Moreover, reduction of value R causes decrease in
the U-anode running voltage, which furthers taking down of ion bombardment
due to reduction of ion energy
For different versions of field emission cathodes, the U decrease effect under
reduction of R is far from being equal. It is least effectively manifested in case of
needle field emission cathodes, whose form factor value is very weak, depending on
R according to the logarithmic law; and more effectively manifested in case of edge
and edge-like field emission cathodes, whose form factor depends on R according to
the parabolic law.
The researchers working in microelectronics are very optimistic about this area.
E.g., they know implementable ways to obtain very small values of R reaching
tenths and even hundredths of a micron [7, 8]. Efforts are under way to make active
field emission microelements, i.e. integrated circuits based on field emission.
Apparently, electric vacuum devices of wide practical application have to be
limited to the values of R of several tens of microns. Further reduction of
R generates the difficulties caused by both technical and physical reasons. Even if
we distract from technical difficulties of setting small R, we can see merely physical
difficulties of controlling of the distances of tenths and hundredths of a micron at
temperature variations.
Let us consider that in more detail. Thermal-expansion rates of metals widely
used in vacuum engineering belong to the range of (50–200) 10−7 °C [22]. That
means that if the temperature fluctuates within the range of 100 °C, the linear
dimensions of electrodes change by (50–200) 10–5 of its original value. It is not
difficult to see that if usual linear dimensions of the electrodes are equal to *1 cm
5.1 Cardinal Problems of Field Emission Cathodes 233
(*10−2 m), change of their dimensions can fluctuate within (5–20) µm. Obviously,
field emission current is instable in designs with a very small value of R, in order of
several µm, without thermostating of the electrodes exceeding one centimeter.
Apparently, if R is very small, one can only make the diodes having small linear
dimensions of electrodes. However, development of any current-controlled elec-
tronic device involves additional electrodes to be located in the cathode-anode
space volume. Development of controlled electronic devices needs R of at least a
few tens of microns, which seems to be a reasonable practical limit.
8. Let us consider one more area of research of field emission current stability
improvement. This area is connected with a specific feature of emission char-
acteristic of field emission current for semiconductors [9, 10]. This characteristic
i(u) has a saturation region in external-voltage area u or external-field intensity
E. In this region, field emission current does not depend much on
potential-barrier penetrability, i.e. the field emission cathode surface state. In
other words, the adsorbed-atom layer on the semiconductor field emission
cathode surface resulting from residual-gas adsorption scarcely affects the value
of field emission current, the latter being stabilized thereby.
An author of this book has carried out a great deal of experimental research of
field emission from semiconductor materials, mostly silicon [10]. His papers show
that the prerequisites and conclusions of the semiconductor field emission theory
are in reasonable agreement with the results of experimental research.
Paper [10] points out «… that semiconductor materials can become the basis
for… cardinally new solution of field emission cathode stability improvement».
However, the semiconductor area of research has a series of restrictions.
A relatively high specific resistance of semiconductors entails a considerable Joule
heating of such field emission cathodes with flowing current and limits the possible
values of field emission current. However, the restrictions caused by technical
properties of field emission cathode materials are worsened by a more serious
physical reason constraining development of research in this area. The matter is that
the saturation region makes current uncontrollable or slightly controllable by
external voltage.
That is the critical weakness of semiconductor field emission cathodes: they
practically cannot serve as the basis for current-controlled electrovacuum devices
without a complex cathode assembly.
9. Creation of a statistically stable system of emitting microelevations became
possible with development of carbon field emission research [11]. Such mate-
rials can develop the maximum number of emission centers uniformly spread
over the surface of the field emission cathodes and making nearly the same
contribution to total emission current, i.e. the materials develop an emitting
surface. At that, the internal structure of some carbon materials causes some
emitting microelevations destroyed as a result of ion bombardment to be
replaced with emission centers of similar parameters emerging from the material
structure. The result is a considerable long-term stability of the field emission
cathodes made of such materials under the conditions of high technical vacuum
generated in sealed-off electronic devices. The field emissivity of carbon
materials and designs of field emission cathodes on their basis are featured in
detail in Chap. 6.
So the brief analysis of the identified areas of research of field emission current
stability improvement has shown that the progress achieved in some of them has still
been insufficient to provide field emission cathodes with a stable electron emission in
technical vacuum or to maintain some obvious advantages of field emission cathodes
over hot cathodes, which cannot but constrain the device developers’ interest to such
research. Therefore, it would be expedient to supplement those areas of research with
some new ones supposedly able to improve field emission current stability, which is
the central problem in this area of physical electronics.
The chapter continues with brief overview of the main types of field emission
cathodes and their specific features.
5.2 Pointed and Multiple-Apex Field Emission Cathodes
Pointed cathodes are cone-shaped apices made of different materials in a variety of

methods (cf. Sect. 2.5).
Pointed field emission cathodes made of different materials have nearly identical
weaknesses, the main of them being small values of total current at relatively high
anode voltages and field emission current instability in high technical vacuum.
Hereinafter technical vacuum shall also be understood as vacuum conditions in the
device when pressure of residual gases is *10−6–10−7 Torr.
These weaknesses of pointed field emission cathodes are mostly caused by their
shape. Bombardment of the emitting surface of field emission cathodes with
residual-gas ions appreciably changes its microgeometry as a result of both cathode
sputtering and the surface migration of atoms. That changes the dimensions and
shapes of the emitting surface followed by change of field intensity E and abrupt
change of field emission current. In particular, double increase in the apex radius of
the pointed field emission cathode nearly halves field intensity, but current density
goes down by 4–5 orders of magnitude. The service life of pointed field emission
cathodes in technical vacuum is about one hour at takeoff of field emission current
of a few microamperes, which is obviously insufficient for practical application.
Therefore the main papers of the recent decades were aimed at improvement of
pointed field emission cathode resistance to ion bombardment.
One of them is focused on coating of pointed cathodes with thin films of
different materials. The latter are mostly transition materials [23] and dielectrics
[24–26]. Overwhelmingly, the coating is applied by thermal sputtering in vacuum
and subsequent diffusion on the emitting surface or by simple dipping of the apex
into the appropriate solution [26]. Recent years saw development of some more
5.2 Pointed and Multiple-Apex Field Emission Cathodes 235
sophisticated coating methods, such as application of diamond-like films by the

method of electrophoresis [27, 28].
The main effect of such coatings is reduction of electron work function followed
by reduction of the running voltage at the same value of the emission current.
However, it has already been specified above that pointed cathodes are not good
for high-volume sealed-off devices. So they have been mostly used in expensive
devices, such as electron microscopes, X-ray microanalyzers etc., which can
achieve and maintain ultrahigh vacuum at the level better than 10−9 mm Hg, at
least in the field emission cathode band. Since such devices employ focused
electron beams, it becomes the issue of the day to localize an electron beam in a
small solid angle.
As a pointed field emission cathode has axial symmetry, the emission cone
opening angle is very large, about 60° for emission at a medium current density
(104 105 A/cm2) and 100° for high densities (107 108 A/cm2).
Nowadays, there are several beam-emission localization methods. The angular
limit can be achieved by cutting of the most intensively emitting region of the apex
surface with a diaphragm and increase in its emissivity through restructuring. The
restructuring is carried out by expansion of certain crystallographic planes by
application of back voltage and appropriate temperature processing [29, 30].
A lot of papers has been dedicated to research of adsorption of different stranger
atoms on the surface of tungsten or other materials, which are selectively adsorbed
on certain crystallographic planes, thus reducing the work function in the adsorp-
tion region. As a result, the cathode surface develops an electron-liberation opening
with dimensions determined by those of the adsorption region. System [W, Zr],
which makes it possible to reduce the divergence angle to 0.5 stere, is the best
studied in this respect [31–33].
Local reduction of the work function can be achieved by apex oxidation, tho-
riating, impurity doping or zirconization. Some of thus-produced cathodes have
already been used in industrial devices. In particular, Siemens, FRG, uses oxidized
apices in microscope ST 100F. Electrons are liberated from the apex at 12°.
The electron-beam divergence angle can be reduced by a congruently vaporised
alloy, which can specifically consist of 85–90% Mo and 15–10% Hf. The appro-
priate orientation of the single crystal brings the emitting spot to the top, thus
achieving emission in a small solid angle along the field emission cathode.
Longevity and stability of the field emission cathode are determined by depletion of
the emission-active additive in the course of non-congruent evaporation from the
heater field emission cathode surface.
An apex with a small solid angle of an electron beam for a heater field emission
cathode is usually produced by thermal field processing. In this case, the basis is a
tungsten field emission cathode with the rounding-off radius of about 2 µm and the
axis oriented towards [100]. At first, the field emission cathode is cleaned by
heating up to 2300 K. Then the temperature is reduced to 1900 K and high voltage
is supplied and stepped up until the emission current reaches 1–100 lA, depending
on the emitter rounding-off radius. At the first moment, the cathode emits per-
pendicularly to the facet [310] as shown in Fig. 5.1b. Growth of the voltage triggers
restructuring of the emitter (Fig. 5.1c) and subsequently is reshapes so that the
entire emission takes place only from the facet [100] (Fig. 5.1d).
Emitter restructurization is accompanied by stepping up of the emission current.
At this stage of processing of the cathode, emission is stabilized at constant values
of temperature and voltage [34].
The field emission cathodes produced in this way are capable of a long service
life at T 1200–1700 K with current takeoff of 1–300 µA at residual-gas pressure
of q 10–6 10−5 Pa. Pressure reduction to q 10−9 mm Hg makes it possible
to step down the temperature of the field emission cathode by 570 K.
Paper [35] focuses on application of two-component coatings for field emission
current localization in a small solid angle. Boride field emission cathodes are pro-
duced by consecutive etching in the solutions of sulphuric and phosphoric acids [36].
The limiting current value that can be taken off from a field emission cathode is
limited to heating of the emitter with Joule heat and the Nottingham effect and
ultimately determined by the heat-conducting path conditions [37], the apex
geometry [38] and the operating mode [39]. In practice, the volt-ampere charac-
teristic biases towards current reduction, which limits the emission localization
possibilities. Paper [40] explains this phenomenon by influence of spatial charge.
There are also some other explanations, e.g. those connected with difference
between the true shape of the potential barrier and the classical image-force law as
adopted to derive the barrier-transparency expression [41]. Influence of the space
charge on field emission was studied in paper [42]. The estimates made are in good
agreement with the respective experimental data. So a particular current density of
(1 5) 10−6 A/cm2 causes emission in the localization region to grow much
more slowly than in the surface region unaffected by the space charge. The latter
equalizes emission over the entire emitting surface, practically delocalizing it and
making further reduction of the emitting area inexpedient.
(a) (b)
310
100
310 310
100
310 310
310
(c) (d)
100 100
310 310 310 310
Fig. 5.1 Schematic of rearrangement of emitting surfaces for reduction of the electron beam solid
angle: a initial monocrystal and its Miller indices; b emission in initial state; c beginning of
thermo-field rearrangement process; d final state
Fig. 5.2 Diagram of multi-tip field cathode manufacture out of metal foil
A useful consequence of impact of the space charge is its stabilizing influence on

emission. In the volt-ampere characteristic region, where straightforwardness of the
Fowler–Nordheim graph is disturbed, emission practically does not depend on
external factors such as electron work function and voltage. This circumstance con-
siderably improves robustness and stability of a field emission cathode in this area.
The limited values of the emission currents taken off from pointed cathodes
induced papers on multiple-apex system engineering. Initially, such systems were
separately-fabricated individual field emission cathodes connected into a single
system in this or that way, e.g. [43]. However, serious technical limitations in terms
of accuracy of assembly of particular apices into a single field emission cathode
made this area promptly transform from the multiple-apex field emission cathode
fabrication technology to a single cycle of fabrication of such field emission
cathodes of metal foil by electric spark machining [44, 45].
Subsequently, such blanks are electrolytically etched in a bath in the same way
as single pointed field emission cathodes [46] (Fig. 5.2)
Electrolytic etching causes heterogeneity of field emitters in terms of their
rounding-off radius and height, which causes heterogeneous field emission from
particular field emission cathode emitters.
This phenomenon can be removed, in particular, by variations of the electrolyte
level towards the blanks [47] or heating of apices in the electric field at such
voltages and temperatures when evaporation of the material from the apex surface
prevails over its diffusion to the effective area [48]. In this case, the rounding-off
and the lengths of the cone part of particular emitters somehow equalize.
There was a more technological idea to make a multiple-apex field emission
cathode consisting of numerous thin wires of emitters fastened in a conducting [49]
or dielectric [50] matrix with uniform density. The arrangement of such a matrix is
shown in Fig. 5.3.
Such field emission cathode can contain more than 100 thin metal wires. The
emitting surface of 3 individual emitters can be located both above and below the
surface of field emission cathode 4.
The emission current is controlled with a perforated-plate control electrode
introduced into the design [51].
Additional electrodes [52] generate electron beams of a definite configuration,
e.g. those having a square or a round cross section etc.
Fig. 5.3 Field cathode out of

thin metal wires in dielectric
matrix: 1 metal wires; 2
dielectric; 3 emission surface
of a separate emitter; 4 field
cathode surface
Multiple-apex field emission cathodes often break down as a result of

inter-electrode arcing. With the purpose of protection against this phenomenon,
paper [53] suggests introduction of an additional electrode 2 located on insulator 3
(Fig. 5.4). Selection of potentials on additional electrode 2 makes it possible to
redistribute power lines in the way preventing inter-electrode arcing (Fig. 5.4b, c).
A more technological method of fabrication of multiple-apex metal field emis-
sion cathodes was suggested in [54]. It suggests organic solvent 2 (Fig. 5.5) con-
taining the metal particles of In and Sn (*5000 Å) 3 is applied to glass substrate 1.
A thin hard film 4 of about 100 nm develops on the surface after drying at 50–
280 °C in the nitrogen atmosphere. When hot nitrogen of 280–600 °C is blown
along the film surface, it develops extremely small holes located chaotically on the
film surface. The solvent evaporates through the holes and the stream causes
microelevations consisting of metal particles 5. The shape and the number of the
microelevations depends on the concentration of the microparticles. The higher the
concentration, the larger is the number of holes 6, 7 at microelevations 5.
Microelevations 5 can be as high as 3 µm, and cathode layer 2 can be as thick as
1000 Å. The resulting microelevations have a well-developed emitting surface able
to generate a stable field emission current.
A prospective way to fabricate semiconductor multiple-apex cathodes is the
method of photolithography followed by deep etching. This method of fabrication
of multiple-apex semiconductor structures was initially suggested by researchers of
the Institute of Physics of the Academy of Sciences of the Ukrainian Soviet
Socialist Republic P.G. Borziak, A.F. Iatcenko et al. [55], and then considerably
enhanced by their colleagues from Westinghouse R.N. Thomas, H.C. Nathanson
et al. [56]. Fabrication of multi-apex systems by this method was reduced to cre-
ation of a master system of island regions, i.e. etch-resistant masks, on flat single
crystal plates by the photo lithographic method followed by deep etching of the
Fig. 5.4 Design of multi-tip field cathode with breakdown-proof electrode: a cathode design; 1
tip field cathodes; 2 additional electrode; 3 isolator; 4 contact electrode; b flux distribution between
field cathodes and the protective electrode at zero potential between electrodes; c at positive
potential on electrode 2
semiconductor substrate through the gaps between the island regions up to devel-
opment of microcones with pin-point tips. The rest parts of the mask were removed
from the tips to strip off pointed microstructures.
Etching was done by a chemical-dynamical method in the etchants consisting of
the mixtures of nitric, hydrofluoric and acetic acids. Depending on the
plate-conductivity type, orientation and dimensions of the etched structures, the
equivalence ratio in the etchants varied.
The etching nature was essentially determined by the plate orientation and the
shape of the image elements in the photo mask.
Fig. 5.5 Final stages of manufacturing of a field cathode out of metal particles: 1 substrate; 2
solvent; 3 metal particles; 4 thin film on the solvent surface; 5 microledge; 6, 7 holes in the sides of
microledges
The method of isotropic etching of silicon was used in papers [56–60]. R.

N. Thomas, H.C. Nathanson et al. [59, 60] created large matrices of silicon apices
with a rounding-off radius of about 2500 Å or less by this method. First the silicon
plate was oxidized, and then SiO2-disks matrix was created by projection pho-
tolithography. The disks were as thick as 1000 Å, the diameter was 50 µm and the
distance from center to center was 60 µm. Then the plate was fixed on a Teflon disk
and exposed to etching in the course of revolution in a mixture of nitric, acetic and
hydrofluoric acids in ratio 10:3:0.5, silicon etching being 50 times as fast as SiO2
etching [61]. As a result, oxide island regions underwent less etching and thus a
pointed structure was generated. Etching took place at room temperature for 15–30
and more minutes depending on the dimensions of the oxide pattern applied [60].
Etching resulted in formation of silicon apices under the SiO2 layer. When the
diameter of the apex tip became less than 5000 Å, etching stopped, the apex radius
being controlled visually as SiO2 is transparent. The SiO2 layer was removed by
etching in 10%-hydrofluoric acid [59]. The matrices obtained in paper [56] con-
tained from 1.5 105 to 106 apices/cm2.
The photolithographic method followed by deep etching of a semiconductor
substrate was also applied in the papers by A.I. Klimin, D.A. Sakseev et al. [57] to
fabricate multiple-apex field emission cathodes. The material used was
boron-doped single-crystal silicon of hole-type conductivity with specific resistance
of 7.5 and 500 X m. Multiple-apex mosaic silicon structures were fabricated by
deep etching of facets (100), (110) and (111) of single-crystal silicon polished by
the chemical-dynamical method. The papers describe testing of several types of
masking materials [57, 61].
The geometrical homogeneity achievable with the etching method throughout
the matrix area depended on many factors. The major of them were the uniform
exposure of the photoresist, the target-revolution speed at etching and the etchant
composition. Naturally, some apex-etching cessation skills were also necessary.
A wrong selection of the revolution speed at etching caused the matrix center or
edges to be etched too deeply, which prevented those regions from emission.
Moreover, the method based on liquid chemical etching suggested the resolving
ability limited to the effect of underetching under the masking coating. That method
could not make apex packing at the given height more dense because the initial size,
i.e. that of the etching-resistant mask, and the final size of the resulting apex differed
by more than two orders of magnitude, otherwise, the apices could not achieve the
necessary height. So deviation of the mask dimensions during image transmission
or a local speed variation conduced to geometrical heterogeneity of the resulting
structures. That was why there were some requirements to the etching process.
Nevertheless, chemical etching cannot provide reproducibility of fabrication of
pointed field emission cathodes.
It was suggested to increase effectiveness of multiple-apex cathodes through
apex-packing densification at the given height. That method differs from previous
ones in greater reproducibility of the technological process along with minimum
contamination of the semiconductor substrate surface and provides uniform emis-
sion of surface-structure elements. Ion-plasma etching is included in the flowchart
and the surface decontamination system for all stages of the technological process is
developed. Ion-plasma etching could provide definite-height bumps, which makes it
possible to begin controlling quality of multiple-apex structures as early as at the
initial stage of fabrication.
Papers [62–65] suggest fabrication of multiemitter semiconductor structures by
application of a Ta2O5 film as an etch-resistant mask, the method being now used
for germanium. The emitter-matrix fabrication flowchart is shown in Fig. 5.6.
The germanium substrate decontaminated by magnetron deposition was layered
with tantalum, after which projection photolithography created a tantalum disk
matrix subsequently oxidized to Ta2O5 (stage I). The disks were 10–15 lm in
diameter and 0.1–0.2 lm thick. At stage II, ion-plasma etching created some
definite-height bumps with etch-resistant masks of Ta2O5. Durable oxide Ta2O5
with good dielectric properties makes it possible to reach high etching coefficients
during ion-plasma sputtering at direct current. The resulting bumps were chemically
etched in the mixture of HNO3:CH3COOH:HF = 5:3:3 up to formation of a pointed
structure (stage III). Having good adhesion properties, the Ta2O5 mask stays up to
final formation of apices and remains almost unaffected by the acid mixture.
Etching results in formation of apices under the Ta2O5 layer. It is especially
important to control the structure resulting from etching at the final stage of etching.
The control is visually assisted with an optical microscope because the Ta2O5 layer
is transparent. Finally, the Ta2O5 mask is removed by a high-frequency discharge in
Fig. 5.6 Flow diagram of

germanium field emitter
production
the argon medium (stage IV). Figure 5.7 shows a typical microphoto of the
resulting structures.
Field photocathode matrices were fabricated from single-crystal plates of ger-
manium with specific resistance ranging from 3 to 300 X m, orientation (111), the
area of a few cm2 and thickness of 0.3–0.8 mm.
Homogeneity of emission is largely predetermined by geometrical homogeneity
of the cathode matrix. However, research often identifies some uncontrollable
centers of emission absolutely unrelated to the resulting structure, which used to be
the case for silicon emitters. Auger analysis confirms [66] the earlier-made con-
clusion that the uncontrollable centers are connected with admixture micro inclu-
sions in the initial semiconductor plate, as well as contamination of its surface and
defects in workmanship.
Therefore we removed the etch-resistant dielectric mask by high-frequency
discharge in argon [64], which made it possible to exclude the operation of removal
of the waste photoresist by boiling in monoethanolamine.
It was a large step forward in development of multiple-apex field emission
cathodes when Spindt et al. suggested a field emission cathode fabrication tech-
nique combining a thin-film technology and electron-beam microlithography [67–
69]. The basic fabrication diagram is presented in Fig. 5.8.
Fig. 5.7 SEM micrograph of

GDG-3 (gallium doped)
germanium emitter matrix
shown from (111) face; 500
The initial structure is shown in Fig. 5.8a. Insulating substrate 1 is covered with
a thin layer of metal 2, namely molybdenum, which serves as a cathode electrode.
Over it, there is dielectric layer 3 and control electrode 4, which is also made of
molybdenum. Layer 3 includes cathode cavity 5 as shown in Fig. 5.8a.
Then the revolving substrate is sprayed with an Al2O3 dielectric layer at the
acute angle of about 15°. The angle value was determined by the necessity to
prevent Al2O3 from penetrating into cavity 5 (Fig. 5.8b). When the hole diameter
decreases down to some value determined by the parameters of sputtering, the
molybdenum sputtering source located completely perpendicularly to the structure
surface is turned on (Fig. 5.8c). Due to continuous reduction of the hole diameter,
simultaneous sputtering of Al2O3 and Mo shapes apex 6 on substrate 2.
Some molybdenum that deposits on the Al2O3 layer surface forms mixture 7,
which is chemically dissolved in boiling phosphoric acid and removed. The result is
a thin-film field emission cathode with a control electrode (Fig. 5.8d). The resulting
structure has the following characteristic dimensions, insulating layer 3 as thick as
0.35 1.5 µm, the control electrode as thick as 0.4, 1.5 µm diameter of the hole in
the control electrode, *100 Å rounding-off radius of the resulting pointed emitter.
Such field emission cathodes are shaped in large matrices with the number of
emitters ranging from 5000 [70] to 50,000 [71]. The distance between the centers of
the emitters forming a rectangular matrix range from 12.7 [70] to 4 µm [71]. The
apices are packed as densely as 108 cm−2.
Such analysis of field emission cathodes is given in [70]. Recent achievement
under the pulse mode is currents of 300 mA lasting 100 ls, which is 6 µA per
emitter on average.
Fig. 5.8 Schematic of manufacturing of Spindt field cathodes: 1 substrate; 2 field cathode
electrode; 3 insulating layer; 4 control electrode; 5 hollow part of the cathode; 6 field emitter;
a base blank; b spraying on the protective layer; c forming the tip; d finished structure
Over the years that have passed since invention of this technology, its basic
development has been aimed at densification of apices and involvement of different
materials for the main electrodes of the emitter, the substrate, the control electrodes
etc.
Though the running voltage of such structures is hundreds of volts, short dis-
tances between the cathode and the control electrode cause desorption effects that
can conduce to development of electric arc 2, which puts the field emission cathode
out of order.
So there were attempts to reduce the running voltage by coating of apices with
materials reducing the electron work function and thereby the voltage. E.g., [72]
describes a pointed field emission cathode coated with cesium under the conditions
of the ready device. The amount of the adsorbate should be optimum to minimize
the running voltage and at the same time it should not cover the dielectric layers to
minimize leakage currents. However, such application of cesium prevents from
taking the cathode structure outside, e.g. to mount the devices.
It seems prospective to process the cathode structure with hydrogen plasma [73,
74]. The chemical sorption of hydrogen on molybdenum reduces the electron work
function from *0.3 eV at the processing time of 10 min to *0.92 eV at the
processing time of 100 min. That increases the emission current by 15% under
constant voltage.
Another example is sputtering of ZrC on molybdenum apices [75]. ZrC is well
known to have good chemical stability and resistance to cathode sputtering.
Sputtering took place at the pressure of 10−6 mm Hg. The field emission trials
showed threshold-voltage reduction by about 40% in comparison with pure
molybdenum apices, which is explained by reduction of the electron work function
of the field-emitter material.
A natural desire to simplify fabrication of such field emission cathodes and
improve their operational characteristics, especially the ability to work under the
conditions of serial sealed-off devices, resulted in a considerable number of sug-
gestions in this area.
It is well known that stabilization of the field emission cathode can be improved
by inclusion of a high-ohmic resistor comparable with internal resistance of the
control electrode—field emission cathode gap, in its circuit. An example of
implementation of such a solution is given in [76], which illustrates a standard
approach to this problem. It involves sputtering on the substrate and cathode
contacts of the resistive layer.
Practical purposes are more often attained with well-focused electron beams.
However, the positive potential of the extending electrode causes expansion of the
electron beam. The electrons are drawn to positively-charged ring electrodes. They
will have both radial and axial component of the speed depending on the emission
angle. Therefore use of the second, focusing electrode with a potential lower than
that of the control electrode gives a good focusing effect [77]. This idea was tested
on silicon pointed cathodes located below the extraction electrode by 400–0.7 µm,
respectively [78].
For these dimensions, a negative-resistance region is created on the focusing
electrode at the voltage of 4 12 V. The emission current is stepped up with
voltage increase at the extraction electrode; however, the number of emitted elec-
trons reaching the anode is thereby decreased due to their repulsion by the focusing
electrode. Further increase in extraction voltage switches the anode current off. For
the voltage at the focusing electrode >12 V, increase in the extraction voltage
makes the anode current scale up. That shows absence of negative resistance in the
emission circuit.
It has already been pointed out that a huge number of papers is dedicated search
of alternative materials suitable for Spindt structure. This section is not to mention
all such attempts, but give a few examples. Many metals were tested as a material of
field emission cathodes, e.g. Pt [79], silicon [78] and carbon nanotubes [80, 81].
The technologies of fabrication of metal and silicon apices were featured above.
Carbon nanotubes are deposited on the bottom of the ready structure of each
element by the CVD method, where a metal catalytic agent is preliminarily sput-
tered. The number of the grown-up nanotubes is uncontrollable is this case. To
reduce the number of nanotubes per element, there were attempts to grow them up
on molybdenum apices of Spindt cathode structures. The structure combining the
thin-film technology and growing-up of whiskers is given in [82]. A low-resistant
n-type silicon substrate was sputtered with a SiO2 layer as thick as 1 µm. Then a
layer of molybdenum as thick as 150 nm was applied by electron sputtering. The
control-electrode aperture was made by the usual technology involving pho-
tolithography and reactive ion etching. The sputtered structure was plunged into a
diluted solution of hydrofluoric acid (water:HF = 6:1) to remove oxides and form
an aperture. The end structure had an aperture of 2.5 µm diameter and *1 µm
depth.
Tungsten nanofibers were applied by the electron-induced deposition method in
WF6 vapor. That material was selected for two reasons: (1) its boiling temperature
is 17.5 °C, which mitigates the condensation problem; (2) absence of carbon and
oxygen should conduce to higher purity of the coating applied. The microfiber
radius essentially depends on the conditions of the technological process such as
pressure, current, time etc., but identical parameters of the technological process
result in nearly identical radii. Paper [82] mentions fabrication of tungsten nano-
fibers 1 µm high, 40 nm thick and having the apex diameter of less than 15 nm.
The field emission measurements were made in a vacuum chamber at the pressure
of <6.7 10−5 mm Hg. The emission characteristics were measured with an
earthed control electrode. The anode located 6.7 mm from the structure was sup-
plied with the voltage of +1000 V, and the cathode with the voltage ranging from
−60 to 250 V. The threshold voltage determined as the cathode-voltage value
necessary to generate the emission current of 20 nA was 139 V at the moment of
energization. Then it dropped down to 90 V within a few minutes. That drop of
control voltage at constant current can be accounted for by contamination of the
tungsten apex in the air.
It was discovered that the current of more than 40 nA released by the apex
destroys it, which seems to be connected with very small rounding-off radius of the
apex. At less 20 nA, the emission current remained stable.
The present-day production demands increase in productivity, which in our case
means increase in the number of simultaneously fabricated apices. It is especially
important in fabrication of large display screens. However, there are numerous
technological difficulties in this way. Growing of field emission cathodes of iden-
tical height nearly equal to the thickness of the dielectric layer requires homoge-
neous sputtering over the entire surface of the substrate. For this purpose, an
ordinary sputtering plant needs a large distance between the sputtering source and
the substrate. E.g., the substrate of 10 cm in diameter needs the working distance of
1 m, i.e. the angular spread of the flowing particles to be sputtered not more than
10°. Moreover, reduction of the ratio of the field emission cathode diameter to its
height is accompanied by reduction of the necessary diameter of the hole in the
control electrode from more than 1 µm to less than 0.3 µm without respective
reduction of thickness of the dielectric layer. That also makes it necessary to reduce
the angular spread of the current of particles to be sputtered to prevent early closure
of the cathode cavity and reduction of the cone height. Further increase in the
substrate diameter is impossible without the vacuum chambers of more than 3 m,

which dramatically increases the price of the sputtering equipment. That was why
some new methods to sputter homogeneous nanocones on the substrates of any
dimensions in a relatively small vacuum chamber were suggested. Their technical
details are featured in [83, 84].
The fundamental difficulties in practical implementation of multiple-apex field
emission cathodes stimulate research aimed at expansion of the range of tech-
nologies and materials for their fabrication.
Recent years saw a lot of new research papers dedicated to fabrication of
multiple-apex cathodes of metal oxides, e.g. ZnO, CuO.
Paper [85] suggests a simple method of growing of field emission cathodes on
the basis of ZnO nanowires. Zinc was ground and cleaned in dissolved chloric acid
and water and put in a ceramic tube. The tube is inserted into a suitable tube-type
furnace. The outdoor processing temperature is 350 and 400 °C, which is below the
pure-zinc melting temperature of 419 °C. After a definite period of growth nec-
essary to form ZnO nanowires on the surface of zinc particles, they were taken out
from the ovens and cooled down to room temperature in the open air. The nanowire
diameters range from 20 to 50 nm with the typical diameter of 30 nm.
The research in the raster-type electron microscope showed that nanowires stood
vertically and had almost identical height of about 1 µm. Moreover, they were
absolutely pure and had no inclusions at the ends. The typical value of the threshold
electric field was 6.5 V/µm. In the field of 11 V/µm, a field emission cathode
emitted the current as dense as 2 mA/cm2. All the measurements were made in
ultrahigh vacuum. The characteristics were similar to those of growing CuO
nanowires [86]. The method of fabrication of more stable structures, i.e. the field
emission current fluctuating less than 1.5%, is suggested in [87]. They are triangular
ZnO nanoplates. Briefly, the structure is grown in the following way. An aluminum
crucible with a mixture of metallic Ga and powdered ZnO is put into a quartz tube
and placed in the elevated-temperature band of the furnace. The Si (100) substrate
coated with a 5 nm-layer of gold is located as far as 10 cm from the source further
along the gas-mixture flow. The gas mixture is 95% argon and 5% oxygen. The
process is detailed in [87]. The nanoplate growth direction is [10]. The stoichio-
metric composition is Zn and O. Since this process cannot be implemented without
a golden catalytic agent, it implies growth of ZnO nanoplates by a
liquid-vapor-solid (LVS) method.
In the course of the growth, the quantity of Ga taking part in recovery of ZnO
decreases, which reduces the concentration of Zn vapor and triangulates the
nanoplates. Figure 5.9 shows typical dependences of density of field emission
current j on the macroscopical electric field E when the anode-cathode distance
d changes from 125 to 525 µm. If d = 125 µm, the macroscopical turn-on field,
defined as the field at which j reaches 1 mA/cm2, is 7.0 V/µm. Increase of d makes
both parameters decrease. For d equal to 225, 325, 425, 525 µm, the first field is
4.0; 3.6; 3.3; 3.0 V/µm, the second field is 5.6; 5.0; 4.6; and 4.3 V/µm,
respectively.
Fig. 5.9 Volt-ampere

characteristics of a field
cathode made out of ZnO
nanoplates with different
anode-cathode distances (the
surface area of the cathode is
20 mm2)
The fluctuations of the field emission current measured in ultrahigh vacuum with
pressure <10−10 mm Hg, at the current density of 1.46 mA/cm2 are less than 15%.
In addition to sputtering processes, multiple-apex field emission cathodes can be
formed by pulsed laser bombardment of the silicon surface in the atmosphere of
reacting gases, e.g. SF6 or O2. The bombardments conduce to formation of
self-organized quasi-periodic structures consisting of apices with high aspect ratio
[88]. That is a simple one-step technology to fabricate quasi-periodic structures
without vacuum equipment or contaminant-free rooms. Such structures termed as
black silicon can be sufficiently effective field emission cathodes [89].
Among the pulsing laser parameters, namely wave length, pulse duration, pulse
energy, pulse count, it is pulse duration that is most important for formation of an
apex array.
Paper [90] considers the effect of pulsed-laser bombardment duration on the
surface morphology and field emissivity of silicon irradiated by KrF laser pulses in
the presence of SF6. The KrF laser wave length is 248 nm. The pulse duration used
in experiments is 15; 5; 0.5 ns. The process is detailed in [91]. The field emission
cathode blanks are made of n-type phosphorus-doped Si (100) of 2 2 mm2. The
resulting emitter samples are tested in a vacuum chamber at the pressure better than
10−5 mm Hg in the diode mode. Before electrical tests, the samples are treated in a
10%-water solution of HF to remove oxides from the apex surface. The distance
between the apices and the anode is 150 ± 10 µm for the samples irradiated with
the pulses of 0.5 and 5 ns and 300 ± 10 µm for the pulses of 15 ns. The main
findings of the experiments are given in Table 5.1.
The geometrical parameters are calculated from the cathode-structure image
analyses in a raster-type electron microscope. It is worth noting that the laser pulses
of 15 ns produce the apices as high as 65 µm having the rounding-off radius of
*670 nm. At the same time, reduction of pulse duration reduces the height and the
Table 5.1 Morphological characteristics and field emissivity of the microstructured Si surface
The pulse Typical radii The The average The Threshold
duration of tips height of distance density of field
k = 248 nm curvature tips between apices apices (V/m km)
(nm) (m km) (m km) (1/cm2)
15 ns 670 ± 120 65 ± 5 11 ± 4 4.3 106 2.5
5 ps 470 ± 110 10 ± 2 5.5 ± 1.5 4.33 107 6.0
0.5 ps 530 ± 200 8±1 5.0 ± 1.3 5.53 107 8.0
rounding-off radius of the apices. However, a better combination of the height and
the rounding-off radius in the first case give a higher form factor, which reduces the
threshold field.
5.3 Blade and Wire Field Emission Cathodes
Field emission formulas show that there are only two ways to step up total emission
current:
1. Densification of the field emission current.
2. Increase in the emitting-surface area.
Field emission current can be densified by both reduction of electron work
function u of the cathode material and increase in field intensity E.
A change of the u value of the emitter causes a change of anode voltage
U necessary to obtain a given current value i. E.g., if the given current density is
*105 A/cm2, reduction of u from 4.6 to 4.0 eV reduces the anode voltage by
almost 18%, whereas increase of u from 4.6 to 6.3 eV necessitates increasing the
anode voltage by almost 60%.
It follows that low u-value materials are quite useful in terms of both densifi-
cation of field emission current and stepping-down of anode voltage.
Emission electronics knows a lot of materials having rather low u reaching (1.5–
2.5) eV [92]. However, so low u values are only characteristic for man-made
metal-film systems based on high-melting transition metals and surface films of
atomic cesium, barium, etc. The ideal conditions for long-term existence of such
films are very low temperatures and perfect vacuum. Under real conditions, these
films fail rather rapidly, and their perpetual recovery comes from the reserves of the
active substance in the cathode body, which is only possible at elevated tempera-
tures sufficient to activate thermodiffusion processes.
Application of effective metal-film systems can hardly be successful in case of
the field emission cathodes working at room or even lower temperatures when the
diffusion rates are negligibly low. E.g., the lanthanum hexaboride surface film
cannot to be recovered through diffusion of its atoms from the cathode body at
room temperature. Instability of field emission current of the preactivated
lanthanum hexaboride cathode is obvious in this case. Such and similar cathodes
need heating for normal diffusion processes and recovery of active surface film of
adsorbed atoms. A promising way to solve the problem of stability of field emission
current is search of the materials adsorbing residual gases in the way hardly
affecting the electron work function. E.g., oxygen adsorption on clean surfaces of
niobium and tantalum increases the u value from (4.0–4.2) V, i.e. by as little as
14%.
At that value of u, field emission current can be further densified through
increase in E. However, increase in E is in its turn limited by the mechanical
strength of the field emission cathode material. The theoretical strength of metals is
very high, *(103–104) kg/mm2.
However, the defects of crystal lattices of metals and, first of all, dislocations
reduce these values by several orders of magnitude. Therefore, the data on the
strength of real samples of metals should be treated with care since they charac-
terize only this or that technical sample of metal with its technological history.
It was already mentioned that Joule heating of a field emission cathode is notably
manifested at j 107 A/cm2. This field emission current density value can be
accepted as a benchmark upper limit of this value, and the respective value of E can
be accepted as the electric-intensity upper limit. A particular value of E depends on
the electron work function of the field emission cathode material, the value of
E being less, the less is the value of u.
Actually, such high current-density values can only be obtained in ultrahigh
vacuum in absence of ion bombardment. In technical vacuum, field emission is only
stable at the current density of (102–103) A/cm2, when ion bombardment is rela-
tively small. Current stepping-up conduces to avalanche development of secondary
processes in the inter-electrode gap and the anode of the device [93], resulting in
destruction and loss of the field emission cathode within a few minutes at contin-
uous operation in technical vacuum.
Summing up the discussion of potential increase in the total field emission
current through its densification, we can see that this potential is quite high. The
possible field emission current density values exceed those for hot cathodes by
several orders of magnitude. Nevertheless, they are limited by field emission
cathode material characteristics, such as their electron work function and
mechanical strength. These limitations are of the fundamental physical nature. On
the contrary, the potential for increase in the total current through increase in the
area of the emitting surface is not fundamentally limited. This way of increase in
total current has no physical limitations and seems to be very promising. Blade,
multiple-blade and similar systems enable substantial growth of the emitting surface
area of the field emission cathode up to 10−2–10−1 cm2. Dike suggested increasing
the emitting surface of field emission cathodes with keen blades [94], and E.G.
Shirokov suggested a keen disk chamfered along the perimeter of the disk and
rounded off with the radius of about a micron [95, 96]. A blade field emission
cathode can be exemplified with a multiple-blade field emission cathode assembled
from steel safety-razor blades [97] having a relatively constant rounding-off radius
5.3 Blade and Wire Field Emission Cathodes 251
of the apex from sample to sample. This cathode has shown itself off as a reliable
and convenient emitter for development of electronic devices.
Figure 5.10 presents a design of a prototype diode with a multiple-blade field
emission cathode. The system involved nine elements cut out of the whole blade by
spark cutting. Each element was 10 mm long. The experiments tested different
kinds of standard blades. The best results, both in terms of reproducibility of
voltage-current characteristics and the limiting values of anode current, were shown
by «Perma Super» blades of high-strength steel with the 18°–21° angle of factory
sharpening. The rounding-off radius of such a blade was 1.0 µm. The anode and the
fixing details of the cathode were made of stainless steel. The working surface of
the anode was ground. The insulator was ceramic. The distance of 0.2 mm was set
between the electrodes by measurement probes. The prototype diode was fixed on
molybdenum lead-ins of a dismountable glass device.
The experiments involved continuous pumping-out with an oil-diffusion pump at
the pressure of *10−7 Torr.
The emission characteristics of field emission cathodes were taken in a pulse
mode with a generator of squared high-voltage pulses lasting 60 µs at the recur-
rence rate of 100 pps. Three diodes with voltage-current characteristics that coin-
cided with each other within the possible current and voltage measurement
variations, were fabricated and examined.
The research was limited with the extreme anode-current values not exceeding
10−2 A because increase of this value typically conduced to emission-current
instability, deviation of Fowler–Nordheim voltage-current characteristics in the area
of lower current values and then to vacuum breakdown and loss of the cathode as a
result of destruction of its emitting edge.
Shear blade field emission cathodes turned out more convenient in fabrication
and application. Paper [98] exemplifies fabrication of such field emission cathodes
of tantalum, niobium and copper. The disks of 13 mm in diameter were made of
50 µm foil. They were electrochemically etched in a special installation similar to
the one featured in paper [98]. The disk to be etched was fixed with two Teflon
Fig. 5.10 A model of a diode with a multi-blade cathode. 1 Blade (cathode element); 2 anode; 3
ceramic; 4 base for fastening on the blades; 5 fastening rod (cathode input)
washers on the engine shaft. The axial play of the engine did not exceed 50 µm.
The rotation rate was 100 rpm. The etching was visually controlled with a binocular
microscope.
Tantalum was etched in a mixture of concentrated hydrofluoric acid and nitric
acid concentrated up to 38% with addition of water (15:80:5). The etch-tank
electrode voltage was 1.5 V. The circuital current was 20 mA. The etching time
was 20–80 min.
Niobium was etched in the mixture of concentrated hydrofluoric and nitric acids
in the ratio of 25:70:5 at the voltage of 2 V and the current of 20–40 mA. The
negative electrode used in etching of both metals was an iridium-wire grid of
0.2 mm in diameter. The disk to be etched was immersed at the depth of 1 mm.
Copper disks were etched to the cutting-edge sharpness in a chromic mixture.
The cathode was represented by a lead plate. The disk rotated at the rate of 10 rpm.
The etching time was 30 min. Copper was etched well in case of both the elec-
trochemical method (u = 1.5–20 V) and the chemical method, i.e. without voltage.
After the etching, the blade field emission cathode was formed by heating with
high-frequency currents in vacuum with simultaneous exposure to electric field of
reverse polarity, i.e. the positive of high voltage was supplied to the disk cathode,
and the negative went to the anode. Joint action of surface-tension forces and
ponderomotive forces was supposed to result in leveling and reduction of the radius
along the perimeter.
The field emission cathodes were tested in a sealed-off cylinder-anode device at
the pressure of *10−7 mm Hg. The anode-cathode distance was 0.5 mm. The
current of 100 µA was generated at the voltages of 5 kV for tantalum, 6 kV for
niobium, 7 kV for copper in a steady state.
For all the three field emission cathodes, emission was instable in time and
heterogeneous along the disk edge. It is especially characteristic for a copper field
emitter. In particular, its exposure with 5 minute-current takeoff made the current
drop from 100 to 20 µA.
For the emitters involved, the attempts of large-current takeoff by voltage
increase conduced to transition of field emission into a vacuum arc in some regions
of the blade as a result of which these regions of the emitter failed and melted off.
This phenomenon can be accounted for by the spread of curvature radii in different
regions of the blade and different electron work function of crystal facets opening
out into the metal surface.
Unlike tungsten and molybdenum, which have no stable oxide of their own,
application of silicon turned out most preferable for the field emission cathodes
working in technical vacuum (P * 10−7 mm Hg).
Papers [99, 100] suggest a design of matrix blade field emission cathode of
silicon.
The cathode assembly is a matrix of silicon field emission cathodes containing
6400 pointed cathodes or 3200 blade cathodes spaced 5 µm from each other. The
blade cathode is 5 µm long. The pointed and blade field emission cathodes are
packed in the matrix at the density of 4 106 and 2 106 cm2, respectively.
Fig. 5.11 Single element of a

matrix of silicon blade-shaped
field cathodes
A single matrix element with blade field emission cathodes is presented in

Fig. 5.11. Field emission cathode 2 is fabricated on silicon substrate 1 so that the
cathode tip is at the level of extraction electrode 4. The cathode is 0.65 µm high. The
tip curvature radius is 0.05 µm. Extraction electrode 4 is made of niobium of 0.25 µm
thick and separated from the substrate with dielectric layer 3 of *0.4 µm thick made
of aluminum dioxide. The width of the hole in the extraction-electrode film is 0.8 µm.
The distance from the cathode tip to the extraction electrode is 0.4 µm.
The structure was formed by e-beam or X-ray lithography involving aeolotropic
etching methods standard for silicon planar technology. The technology in opera-
tion creates matrix field emission cathodes of different configuration with
self-reproducible dimensions on nearly any substrate.
The cathode assemblies were tested with matrix field emission cathodes in a
vacuum chamber with residual-gas pressure P * 4 10−4 Pa. The anode was
represented with a conducting plate located 40 µm away from the field emission
cathode. Before the beginning of the vacuum tests, the cathode and the anode were
not exposed to any special treatment.
The tests were carried out at the continuous voltage changing on the extraction
electrode and constant voltage on the anode, the cathode being earthed.
The effective emission area of a silicon field emission cathode with account of its
own tunnel-thick oxide is 2 10−10 cm2.
Research of silicon matrix field emission cathodes points out high stability of
their voltage-current characteristics in technical vacuum of *10−4 Pa, at the run-
ning voltages on the anode below 45 V.
Triode microstructure velocity analysis shows [101] that it is blade field emis-
sion cathode structures that demonstrate the highest speed. A modification of blade
field emission cathodes is thin-wire field emission cathodes, e.g. [102]. Ideally, the
emitting surface is the lateral surface of the wire with a small curvature radius. The
main weaknesses of such cathodes are
– deflections from the axis of the device due to electric field ponderomotive forces
and inevitable misalignment of the design;

wire field cathode. 1 Base rod;
2 metal wire; 3 a layer of
aquadag for fixing on the
emitter
– irregular arrangement of the emitting microelevations over the field emission

cathode surface.
The first problem can be solved with a design presented in Fig. 5.12 [103].
The field emitter consists of rod mount 1 wound round with metal wire 2 and a
layer of Aquadag fixing emitter 3.
If the diameter of the thread would around the rod mount is d, the turn-to-turn
distance is h, the rod-mount diameter is D and the rod-mount length is P and
assuming that it is only 1/3 of the entire surface that emits, the area of the external
emitting surface of such an emitter can be derived from the following formula:
p3 dDl
S¼ h d:
3 h
The effective emitting surface is proportional to the external emitter surface, i. e.

S = kS, where k is the proportionality coefficient.
For a cylindrical thread-shaped emitter of the same length l as the rod mount, the
emitting surface is
Su ¼ pkdl:
assuming that the value of k is identical in the both cases.

So, the increase in the total emitting surface is
Se p D
n¼ ¼ :
Su 3 h
The maximum value of n is reached at h = d. For the wire of d = 10 µm in

diameter wound around the rod of D = 1 mm in diameter, n = 100. If the
turn-to-turn distance is less than 2 down to the limit of h = d, the voltage necessary
to generate field emission current steps up because of mutual shielding. Increase in
the turn-to-turn distance of more than 3 does not practically effect the voltage
necessary for emission but conduces to reduction of the total field emission current.
The working capacity of the field emitter was tested with a prototype emitter
having the following parameters: rod mount (molybdenum, tungsten) D = l mm,
l = 1 cm, d = 50 µm, h = 150 lm.
The helix ends are attached to the rod mount by spot welding. The system is
hardened by coating of the rod-mount surface with a thin layer of Aquadag water
solution, the layer being as thick as 0.5–0.6 diameter of the wire. A smaller Aquadag
thickness value would degrade the mechanical strength of the field emission cathode,
while a larger one would erode field emission current stability. The field emitter is
installed in the diode coaxial system with the anode-cathode distance of 2 mm. The
prototype diode is pumped out up to the pressure of 10–7 mm Hg with heating of the
entire system up to 450 °C. The emitter is heated up to 1500 K by passage of current
through the rod mount. Some aging at the voltage of 15 kV makes it possible to
generate the current of 35 mA.
The field emission current stability is 1.5–2 times as high as the cylindrical field
emission cathode stability with the emitter service-life of 300 h.
The emitter is highly reliable as the field emission cathode does not decay even
after inception of an arc discharge in the diode, while the cylindrical cathode
usually breaks after such a discharge.
Summing up, it may be said that the emission produced by blade field emission
cathodes is extremely heterogeneous.
The spread in the values of the rounding-off radius in different places of the apex
of such cathodes is as wide as *40%, and the emitting surface area is only about
10−4–10−3 of the average surface area of the blade as determined by its geometrical
dimensions.
The wedge shape of blade field emission cathodes is their fundamental weakness
because ion bombardment blunts the keen edge of the blade and reduces field
emission current.
The emitting centers of blade and wire field emission cathodes are the
microelevations on their surface formed as a result of preliminary mechanical and
electrochemical processing of the latter.
5.4 Film Field Emission Cathodes
Film field emission cathodes appeared due to two factors. Firstly, cathode sput-
tering noticeably change the rounding-off radii of the tip of pointed and blade field
emission cathodes in the course of operation, which changes electric field intensity
and field emission current accordingly. Secondly, stepping up of field emission
current necessitates increase in the emitting-surface area. Such field emission
cathodes were suggested in [104, 105]. The potential stability of field emission
current is attributable to the fact that change of form factor b in the course of
operation was much less than this value [106]
db b dR
;
dt 2R dt
where R is the anode-cathode distance.

Stated differently, a change of the form factor of a film field emission cathode is
caused only by a change of the inter-electrode distance since the film thickness
remains constant. In the case of a pointed cathode, there is an increase in the
curvature radius of the emitting surface [107]. Figure 5.13 shows thus-shaped field
emission cathodes coated with a thin metal film of a constant thickness on a
massive substrate. This version of field emission cathodes is referred to as a film
field emission cathode. The working section of this cathode is the keen edge of the
film, i.e. its end face projecting over the substrate.
The extremely simple idea of fabrication of film field emission cathodes includes
the following basic operations. (1) Application of a metal film of a given thickness on
a pretreated substrate surface. (2) Development of the emitting end face of the film.
Fabrication of metal films is well mastered in the today’s technology and
implemented without any fundamentally unsolvable difficulties.
The emitting end face of the film can also be developed by selective chemical
etching.
Other methods are mostly ordinary well-mastered device-fabrication operations.
An obvious advantage of a film field emission cathode over a needle one is that
the emitting surface area of a film field emission cathode exceeds that of a needle
one by 4–5 orders of magnitude, which considerably reduces the load of the film
field emission cathode to the field emission current values acceptable for its stable
functioning.
Fig. 5.13 Schematics of diodes with single-film a and multi-film b field cathodes: 1 field cathode;
2 anode; 3 substrate; 4 an area of the field cathode that is destroyed in the process of cathodic
evaporation; arrows show the directions of field electrons (–e) to the anode and positively charged
ions (+) to the cathode
5.4 Film Field Emission Cathodes 257
In case of a film field emission cathode, there is also a possibility of substantial

increase in the anode working area and respective reduction of specific thermal
loading of the anode in comparison with the respective value for a needle field
emission cathode on condition of identical values of current, the anode-cathode
distance etc.
Here it seems useful to make some reservations. (1) In this case, we examine a
field emission cathode with ideally smooth surfaces in a form of a paraboloid of
revolution and other cone shapes nearly meeting the microgeometry of real needle
field emission cathodes. For blade field emission cathodes, a good example of
approximation is a parabolic cylinder. As for a film field emission cathode, its
working end face looks like a U-shaped cylinder with a rounded-off tip. (2) We
examine field emission cathodes made of an isotropic material, the main process
determining instability of field emission current being the cathode sputtering pro-
cess. Thereby we exclude other processes, e.g. surface migration, ponderomotive
forces, surface-tension forces etc.
Fabrication of such cathodes is exemplified in [108].
Figure 5.14 shows the design of a disk-shaped film field emission cathode,
where 1 is the chromium film; 2 is the copper substrate; h is the film thickness; d is
the cathode diameter; l is the height of the film projecting over the substrate.
The film-cathode fabrication technique is as follows. A copper foil as thick as
0.2 mm was used to cut disks of 13 mm in diameter with the central hole of 3 mm
intended to hold the cathode. The disk surface was cleaned by chemical etching in a
solution of 1000 cm3 of HNO3 and 10 cm3 of 65% HCL for 2 s. Chromium films
were applied on the copper substrates preheated up to 400 °C by thermal evaporation
in vacuum. Special holders fixed the substrate disks on the nickel-chromium tape
surface heated by flow of electric current. The disks stood next to check glass plates.
The evaporators were cone baskets made of a tungsten-rhenium alloy wire of 0.7 mm
in diameter loaded with granulated chromium. The film thickness was regulated by
the chromium-evaporation time. The experimental facility makes it possible to create
the vacuum of 1 10−5–5 10–6 mm Hg under working conditions.
The chromium-oxide film on its surface did not interfere with evaporation
because the rate of atomic chromium diffusion through the oxide film exceeds that
of its evaporation from the clean surface [109].
The chromium-film thickness was measured on the check glass plates with the
interference method. The chromium films on the substrates were usually 5000–
20,000 Å thick. The error in thickness was 15%.
The films fabricated according to this technique uniformly coated the metal
substrates and had metallic appearance characteristic for massive chromium.
Electron-diffraction patterns of the films demonstrated their polycrystalline structure
with the microcrystals as large as 30 Å. Special research has shown that the films of
such a structure had higher and more stable emissive and mechanical properties
than those of a large-block structure sputtered at the substrate temperature of 1000 °
C. A part of the emitting end face of the film was separated from the substrate by
chemical etching in the chromic mixture of (K2Cr2O7 + H2SO4) with a special
installation. The latter consisted of an engine, a shaft bearing a filmed copper disk, a

disk-shaped thin-film field
cathode: 1 a film of chrome; 2
copper base; h is the film
thickness, d is the diameter
and l is the height of the part
of the film protruding over the
substrate
Teflon bath, a bath vertical-shift mechanism and a binocular microscope for visual
control of etching. The disk plunged into the etching solution to the depth of 1 mm.
The etching went on with continuous revolution of the disk at the rate of 60 rpm.
The axial play of the shaft did not exceed 50 µm. The etching took about 30 min.
The result was a chromium film with the end face uniformly projected over the
copper substrate along the entire perimeter of the disk at the height of l = 50–
100 µm.
Such field emission cathodes can be not only disk-shaped, but also flat, tubular
etc. The experimental research of polycrystalline film field emission cathodes made
of chromium [108], titanium [110], niobium [111, 112], hafnium [113] and other
metals confirmed higher stability of field emission current in comparison with that
of the pointed ones at the residual-gas pressure of 10–6 10–7 mm Hg.
Field emission tests were limited with the emission current of 100–500 µA in the
continuous mode and 4–5 mA in the pulse mode [114], which was caused by
overheating of the anode.
Film field emission cathodes open up broad possibilities for fabrication of
electronic devices with microelectronic technologies, which encourages gives a
new impetus to development of such field emission cathodes.
At the initial stage of the research, the film of a metal or a semiconductor was
considered as perfectly smooth, and form factor b was determined by the effective
rounding-off radius of the film end face. Further studies found out that the emitting
surface of the films had some microroughness, which reduces the running voltage, on
the one hand, and queries stability of such field emission cathodes, on the other hand.
The attempts to look into the problem were made in [115] and developed in
[116]. Any device manufactured in the microelectronic technology is based on
reduction of the running voltage, which mostly depends on form factor b.
As for the multifilm field emission cathode [116], running voltage largely
depends on the geometrical macroscopical parameters of the device design, namely
the anode-cathode distance, the cathode-cathode distance, the cathode height, the
film thickness and the field emission cathode tip curvature radius, accordingly.
The experiments were carried out with the system of silicon films formed by the
methods of integrated technology located normally to the monocrystal substrate of
orientation (110) (Fig. 5.15). The samples were boron-doped to the concentration of
*1020 cm−3. The researchers focused on two types of experimental samples of film
field emission cathodes. The first-type films were as high as 5 µm and the distance
between them was 5 µm, too. The second-type films were as high as 1 µm and the
distance between them was 10 µm. Studying of the samples in a raster-type electron
microscope showed that the field emission cathode fabrication technology was able
to produce the films of cylindrical tips with quite the same curvature radius. The
emissivity characteristics were studied with the cathodes filmed with the thickness
ranging from 40 to 400 nm. The samples of d = 0.1 µm showed experimental
dependences of the form factor on the cathode-anode distance (Fig. 5.16). The gap
varied from 20 to 165 µm for the first-type samples and from 0.2 to 165 µm for the
second-type samples.
Fig. 5.15 Field cathode cross

section
Fig. 5.16 Dependence of the

form factor b of a thin-film
field cathode on the gap
R between the cathode and the
anode, with H = 5 lm,
D = 5 lm
Fig. 5.17 Some schematics

of devices with thin-film
fields cathodes: 1 anode; 2
cathode; 3 dielectric substrate;
4 emitting side of the film; 5
dielectric layer; 6 hole; 7
control electrode; 8 edge of
the control electrode
In our opinion, application of film field emission cathodes formed by the

thin-film technology is most effective. A typical example is a series of designs
suggested in [117, 118].
The basic design presented in Fig. 5.17a consists of anode 1 and cathode 2
placed on a dielectric substrate. Level A–A, where the anode is placed, should be
located below cathode level B–B. In this case, the edge of film 4 functions as a field
emitter. Positive voltage in relation to cathode 2 is applied to anode 1. A small

anode-cathode distance and small thickness at the edge of film 4 causes the electric
field intensity sufficient to emit electrons. Electron current is distributed over the
entire surface of anode 1 in the shortest way due to a small distance between emitter
4 and the surface of anode 1. Estimates show that if the anode-emitter distance is
small, the probability of residual-gas ionization is small, which makes reshaping of
the inter-electrode gap improbable, thereby improving stability of emission current
and extending the device service-life. This design makes it possible to reconfigure
anode 1, its material or the anode coating material, as well as the material of
cathode 2, in a wide range.
A small anode-cathode distance reduces the running voltage of the device down
to several hundreds and even tens of volts. A small running voltage makes it
possible to reduce the capacity of the controlling systems and their price,
accordingly.
The design parameters given below can be improved by increase in the emitting
area of the cathode (Fig. 5.17b). In this case, anode 1 and cathode 2 are separated
by dielectric layer 5. Cathode 2 and layer 5 have hole 6 so that the emitting end face
of film 4 is opposite anode 1. This design can produce a more uniform
anode-electron flow, which improves reliability of the device. Moreover, this design
has one more clear advantage, i.e. there is no electron-current defocusing since the
anode area bombarded by electrons is limited by the dimensions of hole 6.
The geometrical dimensions of hole 6 made in dielectric layer 5 slightly exceed
the size of the hole in the cathode film, which conduces to projection of the emitting
end face of film 4 over dielectric layer 5. In this case, the screening effect of
dielectric layer 5 on emitter 4 becomes less, which steps the running voltage down
even more. Moreover, the probability of electrical breakdown between emitter 4
and anode 1 goes down.
The running (anode) voltage can be further reduced by appending the design
with control electrode 7 (Fig. 5.17c). At that, the edge of control electrode 8 is in
immediate proximity from the emitting surface of cathode 4 (0.1–0.3 µm), which
makes it possible to apply the voltage of about 20–30 V to generate the necessary
extraction field.
Paper [118] runs that these designs are appended with an additional electrode
coated with a material of a high secondary-emission coefficient, which makes it
possible to step up emission current and screen-element candlepower in case of a
display screen.
Another step in development of film field emission cathodes is advent of lateral
structures. Such emitters are exemplified in [119–121].
A lateral field emission cathode is diagrammed in Fig. 5.18. Electrode G is
located above the plane bearing emitter E and collector C. It partially overlaps both
the emitter and the collector. With the emitter earthed, the collector was supplied
with positive voltage V, which generated field intensity EGE between the collector
and the emitter. The control electrode generated field EGE ranging from 0 to −5 V/
µm between the collector and the emitter. Dependence of specific emission current
IEC on field intensity EGE is presented in Fig. 5.18 for EQG = 4.5 µA/mm and
Fig. 5.18 Dependence of

emitter-collector current IEC
on the control field strength
EGE
ECE = 3.8 µA/mm. Figure 5.18 shows that total stoppage of current involves almost
twice as little EGE as ECE. Gate-current measurements showed that it is less than
10−9 A/mm at 0 EGE 5 V/µm. The results attained show that the emitter
involved can be used as a basis for a complete analogue of the vacuum valve in
micro- and nano-implementation and, consequently, hardware components for
emission electronics.
A planar field emission cathode and a device on its basis are presented in
Fig. 5.19 [122].
Dielectric substrate 1 develops three-layer conducting structure 2–4 that cannot
but include layer 3 of catalytic materials such as Fe, Co, Ni etc., and upper layer 4
of a non-catalytic material. All end faces of the three-layer structure are coated with
dielectric 5, except for the end face overlooking anode 6. The end face of catalytic
layer 3 develops carbon nanotubes 7. The thickness of layer 3 is selected within the
range of l–20 nm.
The fixed thickness of the catalytic layer limits the diameter of the nanotubes to
be developed. Unlike an ordinary vertical emitter, its flat version produces the tubes
of the diameter not exceeding the thickness of layer 3.
Unlike the vertical design, the flat version provides a dense growth of nanotubes,
which does not severely limit the emissivity of the emitter since the nanotubes form
a thin blade that develops at the length of the nanotubes equal to 10–25 thicknesses
of layer 3, which steps up the electric field intensity near the edge of the blade. To
direct the growth of nanotubes in the plane of the catalytic layer, the latter can be
selectively underetched from the end face. In this case, the developing nanotubes
Fig. 5.19 Construction of a planar field electron cathode based on carbon nanotubes: 1 dielectric
substrate; 2, 4 conductive layers; 3 a layer of cathalytic material (Fe, Co, Ni etc.); 5 a layer of
dielectric; 6 anode; 7 carbon nanotubes; 8 shutter
begin their growth in the plane of the catalytic layer, and the blade projects over the
end face, thereby sticking to the growth direction. This property is preserved if the
nanotube length does not exceed *10 thicknesses of the catalytic layer. If the
nanotube length considerably exceeds this limit, the blade turns into a fringe
consisting of nanotubes. At the catalytic-layer thickness of more than 40 nm and
the nanotube length of *3–10 µm, this fringe is well identified as a set of nan-
otubes coming out of the end face. In this case, it is easy to estimate the tube
diameter with an ordinary electron microscope and compare their diameter with the
catalytic-layer thickness. Structures with the catalytic-layer thickness of *40 nm
and the nanotube length of *5–10 µm were used as test structures to control
observance of the modes of normal formation of carbon nanotubes.
For wider functional use of the emitter, the three-layer structure is coated with
dielectric 5 with subsequent underetching from the open end face before devel-
opment of nanotubes. This is the way to fabricate the vacuum triode. With this end
in view, ordinary electron- or photolithography has to be used to form gate 8 on the
dielectric surface. In the design involved, the anode and the gate can be located at
the distance ranging from a tenth of a micron to a few microns from the blade,
which makes it possible to turn to development of vacuum micro- and nanoelec-
tronic devices.
Four emitters of 500 µm located in the corners of the square with a side of 5 mm
were measured in terms of voltage-current characteristics. The emission currents
spread ±0.25 µA/mm at the average current of 4.25 µA/mm and the applied field
intensity of 5 V/µm. The resulting spread shows that today it is quite realistic to
develop the technology of fabrication of emission integrated circuits in the area of
more than l cm2.
The technological implementation of such a cathode in form of carbon nan-
otubes grown up on a catalytic layer, combined with control electrodes was sug-
gested in [23]. In this version, electrons emitted from the nanotube field emission
cathode move out into the external space, e.g. to the anode of the display screen.
5.5 Field Emission Cathodes Made of Whiskers
Multiple-apex field emission cathodes can be fabricated on the basis of

whisker-growing technologies. Firstly, the resulting apices are long, thin and quite
durable. Secondly, there is a possibility to fabricate the field emission cathodes of a
larger area. Thirdly, the field emission cathodes are fabricated within one work
cycle. Whisker-growing methods are numerous. They are presented in special
monographs, e.g. [123–128]. Below we will exemplify some of them.
Research of the mechanism of nucleation and growth of whiskers turned out
necessary both in terms of their adequate understanding of their structure and
special properties, and in terms of possibility to develop a controllable field
emission cathode manufacturing process.
Whiskers shape in a different and non-equilibrium way, their diameter ranges

from 100 Å to 1 lm, the length/diameter ratio usually exceeding 10, i.e. their
surface-energy fraction is much higher than that of massive crystals. Whisker-shape
uniqueness assumes availability of a specific mechanism of their formation by
unidirectional growth.
The second, diffusion, part of the Sears model is better grounded and viable
because it is applicable to the both mechanisms.
The theoretical and experimental analysis of the diffusion whisker-growth
mechanism was performed in [123].
Numerous experimental observations of whisker growth provided the basis for
the diffusion-growth theory. The result was the basic whisker-growth equation
@n n @2n
¼A I Bn þ Ds 2 ; ð5:1Þ
@t Z @x
where n(x, t), cm2 is the surface concentration of adsorbed atoms on the lateral
surface of the growing whisker in point x at moment t, determined by the exchange
vapor-phase processes, on the one hand, and inflow/outflow of the substance during
surface diffusion, on the other hand; Z is the lattice-point density (cm−2); (I − n/Z)
is a defect fraction, so A − (I − n/Z) is the adsorption rate (A being the
particle-surface collision number (cm−2 s−1); Bn is the desorption rate; Ds ∂2n/∂x2
is the increment of the substance density due to surface diffusion, Ds being the
surface diffusion coefficient (cm−2 s−1).
The equation is solved for a stationary process when ∂п/∂t = 0 and the lateral
surfaces of growing whiskers are free from substance accumulation, which is
possible at small cover degrees and an inconsiderable flow of diffusing adsorbed
atoms. Then (5.1) is reduced to

@2n n A
D s 2 ¼ A I 1 ; n1 0 ¼ ; ð5:2Þ
@x n0 A=Z þ B
where n∞0 is the Langmuir adsorption isotherm, which is the solution of (5.1) at the
whisker length of l ! ∞. This equation admits elementary integration. If we take
into account that the stationary approach assumes the number of adsorbed atoms
getting to the tip and embedding into the lattice to be equal to the difference
between the number of the particles adsorbed and desorbed within the same time in
the region from x = 0 to x = l, then we will get that

dn
¼ n1
0 nl b thðblÞ; ð5:3Þ
dx x¼l
where
rffiffiffiffiffiffiffiffiffiffiffiffi
A 1
b¼ n : ð5:4Þ
Ds 0
5.5 Field Emission Cathodes Made of Whiskers 265
The material embedded into the whisker tip comes from two flows; one of them
is made of the particles condensed directly from the vapor at the end face, while the
other is caused by surface diffusion on the lateral crystal faces. In case of a suffi-
ciently thin apex, the direct condensation at the end face can be neglected; then the
growth rate of the whisker is inversely proportional to its radius, i.e.
dL 2J
thðblÞ; ð5:5Þ
dt rb
where J is the deposition rate during direct condensation on the crystal surface.
At the adsorption balance, the middle of a sufficiently long crystal should contain
the number of adsorbed atoms equal to that of the desorbed ones, i.e. n∞ 0 /ss. In case
of a thin adsorption layer, we have A = n∞ 0 /s s , so
rffiffiffiffiffiffiffiffiffiffiffiffi
I Ds 1 pffiffiffiffiffiffiffiffi
¼ n ¼ Ds s ¼ ks m; ð5:6Þ
b A 0
where ks is the average diffusion-travel length of an adsorbed atom before

re-evaporation.
Here from, it can be derivable for two extreme cases:
(1) for long whiskers l ks, the growth rate is expressed as follows:
dL ks
¼ 3I ð5:7Þ
dt r
and the linear-elongation law:
ks
L ¼ 2I t; ð5:8Þ
r
(2) for short crystals l

ks, the exponential elongation law runs as follows

dL I I
¼ L; L ¼ L0 exp t ; ð5:9Þ
dt r r
where L0 is the crystal length at moment t = 0.

The boundary between these critical areas meets the crystal length of L ks.
The same was observed in experiments with gold and chromium oxide by Nosov
[129, 130]. However, hardly anybody managed to observe the very beginning of the
growth. It can look as the induction period. Actually, the initial stage of growth
involves exchange of the growing crystal with the substrate, and the process moves
on in a more complicated way.
A number of papers employing the whisker-growth observation method sug-

gested in [131] mentions the exponential growth stage. However, there has been no
transition to the linear stage, but sudden stoppage of growth.
Papers [123–131] explain this phenomenon and suggest the following
whisker-growth model. At the initial stage, the atoms getting to the lateral faces
have enough time to diffuse to the tip where their embedding takes place. In this
case, the entire crystal is a rendezvous point for atoms, i.e. the dL/dt elongation rate
is nearly proportional to L and the growth follows the exponential law. The
exponential growth should stop if the atoms deposited at the bottom are unable to
reach the tip. There may be two reasons for that:
(1) at L ks, when the diffusion rate is high in comparison with the
crystal-elongation rate, and the atoms diffusing to the tip re-evaporate; in this
case, the linear growth is possible and it takes place at about the same rate as
that gained by the end of the exponential-growth stage;
(2) if L
ks, the exponential growth can stop because the atoms deposited at the
bottom do not have enough time to catch up with the growing tip. Concentration
of adsorbed atoms at the bottom of the crystal rapidly goes up, and evaporation
is presumed to take place slowly, so the new layers developed on the lateral face
considerably slow the axial growth down. An attempt to take account of the
exchange with the substrate is made in paper [123].
It shows a possibility to use whisker gold crystal-based cathodes in high-current
pulse [132, 133] diode-type and grid-controlled devices. Cathode parameters were
improved by different aging methods [133].
Papers [134, 135] examine this cathode-type instability appearing after contact
of the multiple-apex system with air. This instability is removed by preheating of
the cathodes to high temperatures, which is unacceptable for gold because it is a
very low-melting-point metal. It is worth noting that this type cathodes turned out
inconvenient for operation under steady conditions because the necessary levels of
emission current, in the order of tens of microamperes, are achievable at the
emission-excitation voltages above 1000 V, their emission current continuously
stepping down. Moreover, the papers do not report on technological reproducibility
of parameters of the cathodes suggested.
Condensation of the material from the gas phase through the liquid phase, i.e. the
vapor-liquid-solid mechanism, by the directional crystallization method made it
possible to fabricate multiple-apex systems of semiconductor whiskers on flat
substrates [136, 137]. This method produced low-power systems with satisfactory
emissivity characteristics. That was facilitated by the fact that semiconductors allow
a wide spread of geometrical parameters of individual needles because of the sat-
uration region in the voltage-current characteristic.
Paper [138] examines the fabrication of multiple-apex systems of tungsten
whiskers on a pointed substrate. The whiskers of 100 Å in diameter and 5000 Å
long oriented along the axis are formed during vacuum microdischarges as a result
of destruction of the preliminarily sputtered thin film of WO3. Stationary emission

takes places at the voltages of about 2 kV.
Decomposition of vapor of carbonyls of high-melting metals such as tungsten,
molybdenum and chromium, in strong electric fields conduced to well-developed
systems of whiskers and arborescent masses of these metals [139–145].
The formation of tungsten whiskers and their arborescent masses was featured in
papers [139–141] and presented as a new method of rapid self-controlled whisker
growth. The filaments grow in the substrate regions of the highest electric field
enhancement. The W(CO)6 molecules approaching the substrate surface shaped as a
thin wire or an apex are ionized by field electrons from the substrate and the
electrodes of ionized molecules. The ionized molecules move to the highest electric
field enhancement places of the substrate, deposit on it, decompose and stick to the
outthrusts on the underpan, causing dendritic growth there. The growth causes
development of new emission centers with higher field tension, which facilitates
intensive deposition of metal molecules on them. That is how self-controlled
growth of filaments is established. The optimum substrate temperature-range
facilitating tungsten-dendrite growth has been established as 470–800 °C.
Papers [142–144] apply and extrapolate the above-featured method of formation
during decomposition of carbonyls in discharge in strong electric fields to the
growth of W, Mo and Gr threads on pointed substrates.
Paper [145] made direct observation of growth of dendritic systems by combi-
nation of a reaction chamber with the object space of a transmission electron
microscope (Fig. 5.20), which made it possible to make exhaustive electron-
microscope observation of the system growth kinetics with the formation condi-
tions, as well as clear out the energy of these conditions and its determining role in
the growth mechanism.
The reaction chamber was placed on an EM stage 1. Dendrites were formed on a
pointed W substrate. The apex/cathode 2, attached to the tungsten arc of a wire of
150 µm in diameter, was made by electrochemical etching of a tungsten wire blank
of 100 mm in diameter. The tip-curvature radius of the apices fabricated in this
way was about the tenths of a micron. The anode shaped as a metal center-hole
disk 3 was mounted opposite the apex at the distance of 1 mm. The tungsten
hexacarbonyl vapor flow was supplied to the apex/substrate through this hole along
a microduct 4 of 100 µm in diameter at the outflow. The vapor delivery along the
microduct acting as a microleakage in relation to the microscope vacuum system
did not have a significant impact on the total residual-gas pressure in the column,
which was equal to 10−6 Torr. The apex/substrate was mounted in immediate
proximity from the microduct outflow, which made it possible to assume a cor-
respondence between the vapor pressure at the apex surface and the W(CO)6 vapor
pressure at the microduct outflow. The necessary vapor pressure at the microduct
outflow was set by the temperature of the tank filled up with powdered hexacar-
bonyl according to the known dependence of the W(CO)6 vapor sublimation
pressure on temperature [146]:
Fig. 5.20 Schematic of an

experiment in the slope of the
BS-242 «Tesla» electron
microscope: 1 stage; 2
edge-cathode; 3 anode; 4
microcapillary
lg P ¼ 11:538 3886=T;
where T is the temperature of the vapor source in Kelvin degrees, and P (Torr) is the
W(CO)6 vapor pressure.
When the pressure selected within the range from 10−2 to 10−3 Torr is achieved
at the microduct outflow and a power supply, voltage rising to the discharge
ignition point (Uign), is connected to the electrodes via control resistance (Rctrl),
current is generated in the cathode-anode gap. Growth of the system necessitates an
apex to concentrate the discharge on the apex/substrate, there being a feeble glow in
the apex area. The type of glowing in the apex area suggests that the discharge
seems to be coronal. However, the range of pressures used in this case and strong
dependence of the current on the control resistance in the external circuit are
essentially different from the living conditions of the corona discharge. The main
condition for the discharge to exist in this experiment is availability of a pointed

substrate functioning as a field-electron source. In this case, the W(CO)6 vapor
environment subjected to the pressures we have selected experiences ionic gas
amplification of the field emission current.
Discharge ignition indicates the beginning of growth of the tungsten-dendrite
system. The dendrite-system growth dynamics was continuously observed under
experimental conditions at the electron microscope screen. Along with the imme-
diate observation, the experiment involved photographing of the apex/substrate
surface microrelief variation with periodicity of about 6 shots per minute.
The overall picture of growth of the dendritic structure at different stages is
characterized by the shots presented in Fig. 5.21. In this case, observation was
aimed at ramification and growth of the entire multiple-apex system. Formation was
not accompanied by any additional heating of the substrate (i. e. the initial Tsubstrate
was 300 K). The discharge circuits had the following parameters:
Rcontrol = 0 MX, ignition voltage was 2–3 kV.
The growth begins from formation of microelevations on the apex/substrate
substrate (Fig. 5.21a). The material for the growth is brought to the cathode tip in a
form of metal ions and fragment carbonyl ions like
þn
WðCOÞ6 ! WðCOÞ6n ;
where n = 1−6, formed when thermal field electrons emitted by the pointed sub-
strate make an electron impact on the W(CO)6 vapor molecules. This process can be
represented by the equation
þn
WðCOÞ6 ! WðCOÞ6n þ nðCOÞ;
where n = 1−6; such ionization was established during mass spectrometer analysis
of products of tungsten hexacarbonyl decomposition by an electron impact [147,
148]. The way the material is fed to the growing apices is featured in Fig. 5.22. The
ions formed above the emitting apices/system branches move to them along the
electric field lines and deposit on the apices.
The apices are heated by the current running through them, as well as the
bombarding ions and the crystallization heat released when the metal atoms embed
into the branch. Herewith, the fragment tungsten-carbonyl ions deposited on the
apex/substrate decompose as a result of either thermal dissociation during
adsorption on the heated substrate accompanied by release of kinetic energy, or
bombardment by new ions.
Hence, the underpan forms excess free tungsten continuously fed to the sub-
strate. It can be presented as highly supersaturated metal vapor or a liquid metal
layer on the surface with a large supply of potential energy of crystallization
expressed by the difference of chemical potentials Δl = ll + lcr. That creates
conditions for growth of whiskers of the main trunk or bush branches.
The growth takes place on surface inhomogeneities, i.e. dislocations opening out
into the surface, a foreign-atom group or just microelevations of the substrate.
Fig. 5.21 Images of growth of the system of tips at different points in time (in minutes) a 1, b 2,
c 3, d 4, e 5, f 6, g 7, h 8, i 9, j 10, k 15 min of growth
Here it is necessary to dwell on whisker-growth dynamics, which essentially

differs from that featured earlier [129, 149, 150]. Growth-dynamics analysis wat-
ched at the screen of the electron microscope identified the following microapex
formation stages: (a) the induction stage, including material accumulation, which
took less than ten seconds, and nucleations of an inappreciable size; (b) rapid
growth taking a few split seconds and instantaneous whiskerization, i.e. discharge;
(c) slow growth taking one or more minutes and almost complete growth cessation.
This last stage changes into thickening of the entire whisker with development of
Fig. 5.22 Mechanism of

feeding the material to the
growing whiskers during
discharge
somewhat larger thickening at the tip, which usually generates two new whiskers at
a great rate. Their emergence is similar to original discharge of a bush trunk from
the substrate; however, we cannot determine this rapid growth rate even with
high-speed filming. The phenomenon of abnormally rapid growth rate of a whisker
has nowhere been mentioned before and needs more detailed studying within the
framework of the problem of the multiple-apex whiskerization and arborization
mechanism.
The whisker apices formed in the course of growth conduce to development of
new field emission centers, additional ionization of vapor molecules and domi-
nating deposition of tungsten atoms on these centers. The emitting apices grow in a
linear fashion. After achievement of the length per-selected for the chosen condi-
tions, the linear growth is slowed down and the tip accumulates the excess building
material. As a rule, this aggregation originates crystallization centers, which rapidly
develop into 2–3 new apices.
In their turn, these new apices also begin to emit, which causes the material to be
locally fed to them. At that, the whiskers formed earlier and serving later as a
substrate begin thickening, i.e. their radial growth begins. A trunk with new
branches growing from its tip is formed. The course of this process results in strong
ramification of the developing dendritic system (Fig. 5.21d–f), which in its turn
increments the total emitting area. The finite-size whisker (Fig. 5.21l) is *100 µm
in diameter and *50 µm high.
The effective emitting area calculated by Fowler–Nordheim volt-ampere char-
acteristics in the assumption of the average work function of eu = 4.5 eV grows
from 10–7 µm2 for initial microelevations, to 10–3 µm2 for a completely formed
dendritic structure.
The total current from the dendritic structure reaches 500 µA at the anode
voltage of 400–500 V. Let us note that the total current from this structure is
comparable with the similar parameter of multiple-apex field emission cathodes of
much bigger sizes.
In the course of growth, the electron microscopical image of some thinnest
apices appears washed out (Fig. 5.21), which is most likely connected with their
oscillating motion. Apparently, origination of oscillations can be accounted for by
the electromechanical effect found out for whiskers in [151]. In our case,
mechanical oscillations of dendrites can be observed when current is taken off from
them and they are exposed to the same electric field as in [151]. When the anode
voltage is taken off, the oscillations cease, and the electron-microscopical image
becomes sharp and clear-cut.
It is worth noting that the most intensive emission of 100–200 µA and the
anode-cathode potential difference of 300–400 V, is inherent in the multiple-apex
systems formed under optimum conditions of formation of multiple-apex field
emission cathodes of the dendritic type Rctrl = 10–15 MX; Ip = 200–250 µA;
W(CO)6 vapor pressure of 1–2 10–3 Torr; growth time of 8–10 min.
Low voltage of dendritic field emitters is achieved by making a rather
branched-out bush of thin apices, which is in its turn achieved by the optimum
molecular flow of the building material during formation of microgeometry of
tungsten dendrites. At the same time, the discharge-burning voltage at the end of
formation correlates with the field emission excitation potential during the current
takeoff from the cathode. The reason thereof is that W(CO)6 vapors are ionized by
field electrons from the growing apices and the ions are fed to the most intensively
emitting regions of the dendritic system, providing development of these emitting
whiskers through sprouting of new apex branches, so-called emission centers, and
consequently, their subsequent progressive growth to a certain limit; that is why it is
the power failure in the discharge gap that is the field emission threshold. The
cathode-formation time is determined by the moment of achievement of the voltage
drop to 250–300 V, which amounts to the above-mentioned 8–10 min.
The growth controllability, expressed by direct interrelation of the formation
conditions, i. e. the discharge parameters at the end of the formation, and the
emissivity of the cathodes, provides achievement of high technological repro-
ducibility. Technological reproducibility of emissivity characteristics is provided
under the optimum conditions of formation irrespective of the formation-cell
geometry. The control-voltage spread for constant field emission current of 100 µA
was 10–30 V at the voltage of 350 V, i.e. 5–10%.
The cathode emissivity was tested with the cathode-circuit control resistance of
5–10 MX to prevent development of a vacuum arc. Emission instability, i.e. field
emission current oscillations, when the cathode functions in technical vacuum at the
fixed control voltage of 350 V and the anode current of 100 µA was 3.3%. At this
level, the emission remained steady throughout vacuum testing, *10−6 Torr for
*200 h, further testing being ceased.
For the multiple-apex systems of tungsten dendrites fabricated under optimum

conditions, the emitting area is estimated at about 10−12 cm2; accordingly, the field
emission current density under steady conditions is at least 108–109 A/cm2 [152].
5.6 Field Emission Nanostructures Based on Nanoporous

Anodic Aluminum Oxide
Among the new and intensively developing vacuum-electronic devices, nanos-

tructural field emission matrices are worth special attention. A more detailed review
of achievements in this area is given in monograph [153]. Its authors kindly allowed
to reproduce a small portion of the material in this section.
Porous anodic aluminum oxide with a regular nanoporous structure formed in a
natural way i.e. without submicron lithography, is a cheap and unique building
material for nano-scale field emission devices [154–156]. An idealized structure of
porous anodic aluminum oxide is shown in Fig. 5.23. It shows that porous anodic
aluminum oxide consists of a close-packed oxide cells represented by hexagonal
regular prisms directed to the film surface along the normal line and soldered with
lateral faces [157–159]. Every oxide cell consists of a central pore, oxide walls and
a barrier oxide layer, i.e. the oxide-cell foundation, looking like a hemisphere
protruding into the underlying aluminum layer. The linear dimensions of the oxide
cells and their packing density can be easily controlled by electrochemical alu-
minum anodizing modes, i.e. the selected electrolyte type, its temperature and
anodizing voltage.
It should be noted that the improvements of electrochemical aluminum
anodizing technology suggested in papers [160–171] made it possible to fabricate
porous anodic aluminum oxide films with an almost ideal structure (Fig. 5.23) with
the oxide-cell packing density ranging from 108 to 1012 cm−2 and the pore diam-
eters ranging from 200 to 5 nm, respectively. Paper [154] was the first to suggest
Fig. 5.23 Idealized structure

of porous anode aluminum
oxide out of a layer of
aluminum residue formed on
the substrate in vacuum: 1
oxide cell; 2 pore; 3 wall of
the oxide cell; 4 oxide barrier
layer; 5 a layer of aluminum;
6 substrate
application of the nanoporous anodic aluminum-oxide structure in a thin-film

electrovacuum device as inter-electrode insulation and straight-through traveling
channels from the cathode to the anode. That pioneer was followed by a series of
papers exploring a possibility to use nanoporous anodic aluminum-oxide to fabri-
cate field emission nanoemitter matrices by a non-lithographic method.
The factors influencing the quality of the generated inter-component insulation
based on local porous dense anodic Al2O3 oxides can be conventionally divided
into two groups. The first group includes the factors immediately affecting the
purity and structure of the resulting aluminum films. Such factors are the substrate
material, the vacuum level, the substrate temperature and the aluminum-film
deposition rate. The second-group factors immediately affect the very formation of
porous dense anodic Al2O3 oxides. These factors are the chemical composition of
the electrolyte, the concentration and temperature of the electrolyte, the electrical
porous-anodizing mode, including the anode-current density or formation voltage
and their time variation. The research carried out earlier [172] shows that
high-quality local porous Al2O3 result from the initial aluminum film of a homo-
geneous equilibrium fine-grained structure. Achievement of such a structure of
deposited aluminum films is known to be provided by the vacuum magnetron
sputtering method [172]. The insulation pockets are formed by local
straight-through porous anodizing of the first aluminum layer and the adhesive
tantalum substrate with a photoresistant mask. Therefore one of the major factors
affecting selection of the porous anodizing electrolyte is stability of the photoresist
in this anodizing solution.
The preliminary experiments conducted with different types of electrolytes, as
well as the data available from the research literary [173], showed that no existing
porous-anodizing electrolyte, with a possible exception of an phosphoric acid based
electrolyte, do not make it possible to carry out immediate/direct photoresistant
masking of the necessary regions in the course of porous anodizing of an initial
aluminum film. Paper [173] studied the effect of different concentrations of phos-
phoric acid on the lasting properties of the photoresist and the quality of the
resulting porous dense anodic Al2O3 oxides. At that, the best findings were made in
4 5% water solutions of phosphoric acid. This solution is also used for chemical
polishing/pore enlargement of the resulting porous anodic aluminum-oxide struc-
tures [160–171].
The formation voltage was selected as the factor characterizing the electrical
porous-anodizing mode because the anode-current density is a function of both the
formation voltage and the electrolyte temperature. In addition, maintenance of the
given anode current density necessitates knowing the area of the metal film ano-
dized. At the final stage of the straight-through porous anodizing, this area changes
due to non-simultaneous straight-through anode oxidation of the aluminum film,
which makes it very difficult to determine whether the necessary anode-current
density is maintained. The positive potential supplied to the aluminum film can
cause an anode-current step. Its high voltage may etch the aluminum film through
and deteriorate the formed porous oxide. This effect is excluded by achievement of
the desired formation voltage at the scan frequency of 0.5 V/s.
5.6 Field Emission Nanostructures Based on Nanoporous Anodic Aluminum Oxide 275
Fig. 5.24 Dependence of the

mean pore diameter Dp on the
molding voltage Uf of the
porous anodizing of
aluminum at different
temperatures of electrolyte
Tel: 1 –15 °C, 2 –20 °C, 3 –
30 °C
Figure 5.24 shows that the pore diameter depends on the electrolyte temperature
and linearly grows with the formation voltage in the examined voltage interval from
60 to 210 nm on average.
The idea to fabricate field emission nanoemitters by the galvanic method sug-
gested in papers [174, 175] involves use of porous anodic aluminum oxide with a
remote barrier layer as a mold for embedding of metal nanowires into the
pointed-emitter formation pores.
As porous anodic aluminum oxide is an almost perfect nanostructured dielectric
matrix with vertically-oriented cylindrical nanochannels/pores, a lot of researchers
focused on engineering of carbon nanotube growth in anodic aluminum-oxide
pores. Some of them [175, 176] demonstrated a possibility to grow
vertically-oriented carbon nanotubes in the pores of ordered structures of nano-
porous anodic aluminum oxide without specification of their emissivity
characteristics.
Another researcher [176] synthesized carbon nanotubes in an oxide membrane
with embedded nanowires. The feed material was aluminum foil, where porous
oxide was formed. Thereafter, the barrier anodic aluminum-oxide layer was
removed by dissolution in 20% sulphuric acid. Then the oxide membrane was
separated from the aluminum. After that, one side of the oxide membrane was
covered with a gold layer used as an electrode, and then the membrane was placed
in an electrolytic cell to electro-deposit nickel into the open pores on the other side
of the membrane. I.e., the initial-sample preparation technique was similar to that
mentioned before in [174, 175]. Then the substrate was cleaned in ionized water by
the ultrasonic method. After that, nickel nanowires embedded into the oxide
membrane were used as catalytic agents to grow carbon nanotubes by chemical
Fig. 5.25 Technology of

manufacture of a
nanostructural two-electrode
field emission device
deposition from the vapor phase of the methane-hydrogen mixture reinforced by

application of microwave plasma. The input power of the microwave plasma was
900 W, and the substrate temperature was maintained at about 650 °C. The scan-
ning electron microscopy of the resulting samples showed that the carbon nan-
otubes formed on the surface of the oxide membrane have the diameter of 40–
60 nm, some of them being close to 100 nm. Field emission currents from the
samples were measured in a vacuum chamber at the pressure of 1.33–10−3 Pa with
application of an electrometer. The tests conducted showed that the field emission
current density can reach 1 mA/cm2 at the field of 3 V/µm. At that, the main
contribution to the field emission current is made by the confused nanotubes, their
ends protracting from the samples.
A flowchart of a more advanced nanostructural field emission matrix fabrication
technique is presented in Fig. 5.25.
The nanostructural field emission matrices were fabricated in several basic
stages. Emitting-material layer 2 was deposited on substrate 1 of glass ceramics,
leucosapphire or silicon.
In this case, it was a titanium layer of 200 nm in thickness. It was followed by
deposition of an aluminum layer 3 of 1.5 µm in thickness (Fig. 5.25a). That was
followed by porous anodizing of the aluminum layer to the depth of 1.2 µm, which
resulted in porous anodic aluminum oxide consisting of a close-packed oxide cells
4 represented by hexagonal regular prisms connected with their lateral faces and
directed perpendicularly to the film surface (Fig. 5.25b). Every oxide cell consists
of central pore 5, oxide walls 6 and barrier oxide layer 7 (Fig. 5.25b). The next step
was etching of the formed porous oxidation of the titanium layer, and aluminum
5.6 Field Emission Nanostructures Based on Nanoporous Anodic Aluminum Oxide 277
oxide dissolves on the internal barrier-layer surface, as anodic aluminum-oxide

layer and using of the ordered concavity matrix, i.e. an imprint of the oxide-cell
foundations, developed on the surface of the remaining aluminum layer 3 as a kind
of template to conduct straight-through porous anodizing of the remaining alu-
minum layer 3, thus forming an ordered cellular-porous oxide structure. Growth of
porous anodic aluminum oxide into the bulk of the aluminum film is provided by
equilibration of formation of the oxide on the barrier oxide-cell layer surface and its
dissolution on the internal surface at the pore foundation. So after the top of the
barrier oxide-cell layer reaches the titanium-layer surface, this place begins expe-
riencing local before. After the barrier oxide-layer at the pore foundation com-
pletely dissolves, the anodic titanium-oxide bump embeds into the oxide-cell
foundation followed by growing into the cell pore; and as a result, a regular
nanoscale matrix of titanium-oxide apices 8 embedded into the cylindrical
nanochannels of the dielectric porous anodic aluminum-oxide layer develops
(Fig. 5.25c). The structure was formed as a result of the many-stage electrochemical
anodizing of a double-layer thin-film titanium-aluminum system. For the resulting
nanoscale titanium oxide apices to function as emitters, they were converted into a
conductor structure 9 (Fig. 5.25d) by subsequent thermal vacuum treatment. After
that, some samples were chemically etched to strip of the porous aluminum-oxide
layer and examine the emissivity characteristics directly from the titanium apices.
The pointed structures served as the basis for reliable establishment that the apex
height was constant over the entire surface of the samples under investigation and
determined by the formation voltage at the final stage of anodizing. The visible part
of the pointed structures was 70 nm high determined with the absolute accuracy
of ±5 nm. The number of apices met the number of pores in the oxide and was
about 3.74 1010 apices/cm2 for the samples under investigation, the apex
diameter exceeding that of a pore by about 10–15 nm. The average apex diameter
of the samples under investigation was about 37.5 nm, while the average apex
center-to-center distance was 55 nm. The diameter varied around the average value
by less than ±12.5 nm and was close to the normal distribution.
The nanoscale bump emissivity was studied on the samples subjected to thermal
vacuum treatment and removal of the porous anodic aluminum-oxide layer.
The field-emissivity characteristics were studied in a dismountable vacuum
chamber at the pressure of 10−5 mm Hg in the diode mode.
The sample under investigation always functioned as the cathode. The
anode-cathode gap was about 300 ± 30 µm. The cathode matrix was supplied with
control voltage negative with respect to the anode and represented by rectangular
impulses of 20 µs in duration and 50 Hz in frequency.
The appreciable field emission of such structures begins at the electric field
intensities of about 8 V/µm [177, 178].
A necessity to use high test voltages to initiate field emission is caused by the
fact that the test installation design did not make it possible to reduce the gap
between the cathode and the anode-tester. Coating of the samples with a carbon film
of a few nanometers in thickness halves the threshold voltage initiating appreciable
field emission, which significantly improved the emission stability. The research of
test field emission images shows the effect of the carbon film on the
emission-current density homogeneity over the cathode surface [179]. However, it
is worth noting that despite the inducement to apply carbon films in field emission
microdevice fabrication, their practical use is hindered by high sensitivity of the
carbon films to any subsequent operations, which results in their loss of all the
above-stated properties. Another disadvantage of the carbon films is their high
deposition temperature, generally about 850 900 °C, which determines certain
requirements, most importantly high-temperature stability, to the involved sub-
strates and materials of the functional layers of the field emission devices.
The samples intended for outdoor testing under atmospheric pressure were
subjected to electrochemical polishing aimed at enlargement of the pores by etching
of the incoherent part of the oxide layer surrounding the pores in the oxide cells. As
a result of this operation, the thickness of the dielectric layer of the porous oxide
was reduced to 150 nm. The subsequent vacuum annealing of the surface was
followed by deposition of a double-layer thin-film structure consisting of a tantalum
substrate 10 of (10–20) nm in thickness and an aluminum layer 11 of 1 µm in
thickness. Sputtering of the pores during deposition of the double-layer thin-film
structure on the porous anodic oxide surface generates pointed emitters 12 with the
apex tips directed towards the porous anodic oxide (Fig. 5.25e). The experimental
estimates of the pore-sputtering depth showed that the deposited material runs into
the pores for the distances comparable with their diameter. Then the
aluminum-layer surface is covered with a photoresistant mask meeting the topo-
logical pattern of the upper electrode 13. The aluminum layer unprotected by the
photoresist is etched, the photoresist is removed, and the structure generated in this
way is heat-treated in the atmospheric air. As a result, the tantalum substrate
unprotected by the aluminum layer turns into thermal tantalum oxide 14
(Fig. 5.25f).
This metal-deposition flowchart involved the magnetron vacuum sputtering
method, which can be used along with any other method.
The nanostructural field emission matrices fabricated according to the flowchart
given in Fig. 5.25 were tested outdoors under atmospheric pressure as components
of a nanostructural two-electrode device. When the test device having an active area
of 1 mm2 was under the continuous current loading of 0.5 mA for 10 h, the
observable volt-ampere characteristic did not show any variation. In this case, the
emission-current oscillations at the running voltage of 2.7 V did not exceed 1%.
After outdoor storage under natural, i.e. normal climatic, conditions for one year,
the test device confirmed the initial volt-ampere characteristics.
At the inter-electrode dielectric thickness of 150 nm, the effective field emission
current begins at the running voltages of *2.5 V.
The inter-electrode dielectric resistance measured was about 5 1010 X. As a
result, the conductivity current through the dielectric layer of the device was about
0.2 nA, and its effect on volt-ampere characteristics can be neglected.
Some more details of fabrication of the devices based on field emission cathodes
with anodic aluminum oxide can be found in [153].
5.7 Other Field Emission Cathode Types 279
5.7 Other Field Emission Cathode Types
The simplest application of polymeric structures for fabrication of a field emission

cathode is coating of a single-crystal silicon with a polymeric film/phenylcarbin and
subsequent heating to 800 °C in the Ar atmosphere [180]. Such treatment results in
transformation of the polymer into a carbon film of good emissivity. The experi-
ments showed the threshold-field of 4.3 V/µm at 0.1 mA/cm2 and the current
density of 250 mA/cm2 at the electric field intensity of 10 V/µm.
These data correlate with the microstructural properties of the resulting carbon
film consisting of sp2-bonded nanoparticles. Such coatings make it possible to
fabricate field emission cathodes of a large area.
A large number of papers is dedicated to development of polymer-carbon
nanotube compositions. Such methods do not involve vacuum chambers or
non-standard chemicals, and the composites can be applied by any standard
method.
Paper [181] mentions usage of one-wall carbon nanotubes with the nominal
purity of 60% and poly-3-octylthiophene, P3OT. The films were applied on n+
silicon after treatment in HF. The carbon nanotubes and P3OT were dissolved in
chloroform to form a composite. A composite containing 33 wt% of nanotubes was
prepared by mixing of the respective two solutions along with ultrasonic stirring.
The research needed 3 versions of coatings.
1. The nanotube solution was applied on silicon to form a carpet of nanotubes after
the chloroform had evaporated.
2. The nanotube solution was applied on a pre-dried nanotube polymer composite
layer and dried at room temperature.
The chloroform from the nanotube solution dissolves the polymer on the com-
posite surface, which makes it possible for the nanotubes to open out into the
polymer surface.
3. The nanotube solution was applied on a wet polymer. After that the final
composition was heated at 40 °C for 30 min.
The difference between structures 2 and 3 was that structure 3 had a surface
nanotube layer separated from the silicon substrate with a pure polymer layer, while
structure 2 had an interlayer containing 33 wt% of nanotubes, which considerably
increased its electroconductivity, 10–2 X/m as compared with 10–9 X/m for a pure
polymer. Moreover, the latter structure had a smaller number of nanotubes opened
out into the surface.
The field-emissivity tests were carried out in a vacuum chamber at the pressure
of 10−6 mm Hg in the diode mode. The electrode area was 0.25 cm2 at the
electrode-to-electrode distance of 80 µm.
A simple mixture of carbon nanotubes and the polymer is not field emissive.
Coating of nanotubes with a polymer means that the electric field at the tips of the
nanotubes is too small to generate appreciable field emission. Even treatment of
Fig. 5.26 Schematic of the

Field expansion
surface potential between a
nanotube, polymer and area
vacuum. The lines represent
equipotentials
Vacuum border
Single wall carbon

Polymer nanotubes
such a structure by chloroform does not result in complete removal of the polymer
from the surface layer of the nanotubes, field emission failing to manifest itself even
at the fields of 12.5 V/µm. The chloroform is assumed to dissolve the polymer on
the surface, and then the polymer coats the nanotubes again after drying.
The best results were achieved when the samples were tested according to
process 2. The threshold field was 0.8 V/µm while pure nanotubes resulting from
process 1 generated the threshold field of 2.2 V/µm.
This additional increase in field emission is accounted for by the triple junction
between the nanotube/semi-metal, polymer/semiconductor and vacuum. A similar
mechanism was suggested to account for field emission of diamond films [182] and
nanocrystal graphite [183]. Figure 5.26 the surface potential changes the curvature
at the junction from the nanotube to the polymer due to the difference of the
electron work function of these materials. This non-uniform potential changes the
potential lines directly above the contact of the two materials. This jump in potential
above the junction between the materials results in a very heavy electric intensity.
So this case involves no morphological necessity to create the additional conditions
of increase in the electric field. Moreover, the strong electric field in the immediate
proximity from the surface junction runs into the conducting channels of the matrix,
which causes additional tunneling of electrons.
Paper [184] suggested a triode design using carbon nanotubes and a conductive
polymeric composite. The researchers resorted to electrochemical polymerization of
pyrrole and electrophoretic application of carbon nanotubes. The other elements of
the triode structure were fabricated by standard thin-film processes.
Field emission cathodes were fabricated from one-wall carbon nanotubes of
about 1.6 nm in diameter and 0.5–2.0 µm in length. The cleaned nanotubes were
mixed with pyrrole and lithium perchlorate used as polymeric monomer and
electrolyte, respectively.
Lithium perchlorate intensifies electrochemical polymerization and functions as
a conductive additive in the polymeric matrix. Concentration of carbon nanotubes
Fig. 5.27 Experimental Anode Emission spot

structure cross section
Spacer
Controlling
electrode
from Cr
ITO-
SiO2
cathode
a-Si
was 7.5 10−2 mg/ml, and that of pyrrole and electrolyte was 7 10−4 and
7 10−5 mol/ml, respectively.
The cross-section of the experimental structure is shown in Fig. 5.27.
The structure consists of an emitting spot and a control electrode of Cr. The
diameter of the composite apex made of a nanotube and a polymer, located also on
the conductive cathode, was 20 µm, and the diameter of the hole in the extraction
electrode was 30 µm. The insulator SiO2 is 1.5 µm thick. The amorphous silicon
layer (a–Si) as thick as 2500 Å functions as a resistive layer for the field emission
cathode.
The field emissivity was measured in a vacuum chamber at the pressure of
5 10–6 mm Hg in the pulse mode. At the anode voltage of 1000 V and the
extraction electrode voltage of 60 V, the emission current was 35 µA.
Fabrication of field emission cathodes from vertical tube bundles in a matrix with
cyanoacrylate was suggested in [185].
The carbon nanotube array was synthesized from the ethylene-argon mixture
with the iron catalyst by chemical deposition on the porous-silicon surface. Silicon
was lithographed over the iron-film surface from the electron beam through a mask.
The nanotubes were synthesized in the Ar/C2H4 mixture with the volume ratio of
3:1 at 700 °C. Figure 5.28 shows the sample photographed in a raster-type electron
microscope. The nanotube array was perpendicular to the substrate surface. The
total nanotube array area was about 2.3 cm2, and the nanotube height was con-
trolled by synthesis effect variation. Every cathode contains about 100 nanotube
bundles, the total diameter being about 100 µm (Fig. 5.28b).
Carbon nanotubes grow directly on a substrate to provide electrical contact.
The nanotubes are confused on the bundle surface and oriented on the lateral
surface.
Contact of carbon nanotubes with cyanoacrylate is provided by sticky
cyanoacrylate dripped on top of the nanotube array.
Capillary forces make cyanoacrylate promptly soak throughout the array.
Cyanoacrylate adhesion is not durable. Cyanoacrylate shrinks in the course of
Fig. 5.28 Structure of field

cathode out of carbon
nanotubes, SEM micrograph:
a general view; b bundle of
carbon nanotubes with a
diameter of *100 lm
evaporation and solidification. That is why drying of cyanoacrylate causes the

nanotube bundles to halve their diameter, still keeping the electrical contact with the
substrate.
Essentially, the field emissivity of the resulting structure depends on thickness of
the cyanoacrylate layer on the nanotube surface. If this layer is very thin, the
emitted electrons can easily penetrate through it and the emission current is even
amplified. Otherwise, field emissivity goes down. Unfortunately, there are no
effective methods for fine control of the quantity of the polymer used.
This shortcoming was eliminated and the excess polymer was removed by laser
surfacing. After the laser surfacing, the bundle surface develops well-oriented
nanotubes that essentially improve field emissivity of the resulting structure. This
surfacing effect is accounted for by the fact that cyanoacrylate is easier to be
removed by laser light than carbon nanotubes.
Paper [186, 187] suggested purely polymer field emission coatings without
carbon nanotubes.
The polymeric coating of l–2 µm in thickness was represented by a soluble
imide siloxane copolymer that earlier displayed anomalously high channel con-
ductivity if placed between two metal electrodes [188]. According to [190], syn-
thesis of siloxane imide copolymer by the procedure featured in paper [189] results
in the structure consisting of polyetherimide blocks made of the alternating residua
of dianhydride and aromatic diamine and siloxane imide blocks including the
Fig. 5.29 Copolymer structure
alternating residua of dianhydride and siloxane diamine. The copolymer structure is

shown in Fig. 5.29. The polymer was selected because it forms durable film
coatings due to imide aromatic fragments in the structure unlike polydimethyl-
siloxane, which is viscous at ambient temperature. At the same time, it has flexible
oligodimethylsiloxane and more rigid imide aromatic fragments. Their ratio can be
regulated to affect the supermolecular structure of the resulting polymeric films. The
elastic modulus and the tearing strength of the films of 40 µm in thickness is
40 ± 1 and 6.5 ± 0.3 MPa, respectively, and the temperature of thermal destruc-
tion of imide siloxane copolymer exceeds 400 °C. The copolymer films have good
adhesion to metals.
Thin polymeric coatings of about 1 nm in thickness were applied to the polished
molybdenum or niobic cathodes of 5 mm in diameter by the spout method. With
this aim in view, the cathode was coated with a calculated quantity of 5%
copolymer solution in dry N-methyl-2-pyrrolidone. To remove the solvent, the
electrode with the applied copolymer solution was heated in a drying chamber at the
temperature of 100 °C for l–2 h under atmospheric pressure.
The atomic-power topography of the polymeric-film surface carried out before
research of the field emission effect showed that the surface roughness did not
exceed 30 nm. Thus it can be argued that the polymeric-film surface is quite
smooth. The field emission experiments were carried out in the vacuum of 10−6.
The distance between the flat cathode and the anode grid was 0.5–1.5 mm. The grid
was represented by a shadow mask of a usual kinescope. A flat luminescent screen
image was registered with a television camera with subsequent recording in the
computer memory. The screen was under the grid potential. It was located at a small
distance from the grid, thus quadrupling the image.
The volt-ampere characteristics were measured by 50 Hz high-voltage scanning
both in direct and backward directions. The values of current and voltage were
simultaneously recorded on computer with a two-channel digital oscilloscope.
The lower threshold of electric field intensity when field emission current from
the imide siloxane copolymer coated cathode began to be registered was 4 kV/mm.
However, primary activation of the emitter needed the field intensities exceeding
this value 2.5–3 times. Let us note that the electron emission from the polished
electrodes without a polymeric coating cannot be seen up to 25 kV/mm.
Analysis of atom power topography of the polymer film surface showed that the
polymer surface was considerably modified after studying of the field emission
effect, in particular apices of 400 nm in height appeared. Formation of polymeric
apices in a quite intense electric field can be understood as manifestation of local
electrization of polymer due to self-injection of charge carriers in a quite elastic
polymeric matrix from the metal electrode [191].
A reversing volt-ampere characteristic, VAC, measured at small voltages agrees
with the Fowler–Nordheim theory, which indicates the field emission nature of
work of the polymer-coated cathode. However, the voltage growth up to more than
6000 V reduces the current measured.
According to X-ray research, after an emitter has worked at the field intensity of
E > 10 kV/mm in the inter-electrode gap and the emission currents of I > 1 mA for
a few hours, the polymeric film develops a large number of carboniferous areas due
to polymer thermal degradation. That indicates the local heating of the polymeric
coating by emission currents of more than 700 °C. The researchers detected an
appreciable mass transfer of the carboniferous material to the anode, as well as
intermittent luminosity of the luminescent screen. Some emitting centers stopped
existing, others were formed again.
Paper [192] suggested organic nanowires for field emission cathodes of a larger
area.
They are represented by the organic complexes of Ag TCNQ and Cu TCNQ,
which are good at conducting electric charges (TCNQ = 7, 7, 8,
8-tetracyanoquinodimethane). They have a high density of charge carriers, which
results in high conductivity. The electrical resistance of these materials switches from
high to low within a few nanoseconds by application of the electric field. That
suggests that they can become a new class of materials for field emission cathodes.
The authors of [192] prepared the field emission cathode samples of a large area
(8 3.6 cm2, 14.3 3.4 cm2) on the substrates of silver and copper foil by the
organic vapor-solid body method. This method gave good results and easy uniformity
control.
Practically, the TCNQ powder is loaded into a ceramic boat and placed in the
center of the quartz tube in the horizontal furnace. The copper or silver foil is placed
on the top of the ceramic boat.
The furnace temperature is stepped up to 120 °C for copper and 150 °C for
silver and then maintained for 2 h. The TCNQ is sputtered on the foil surfaces
where the films of Cu TCNQ and Ag TCNQ are formed, respectively. The length of
nanowires is a few tens of microns and the diameter is about 40–150 nm. The
diameter of the resulting nanofibers is one and a half times more at the foundation
than at the top.
Both types of the field emission cathodes have good field emissivity. In par-
ticular, the threshold electric field values necessary to generate the current density
of 10 mA/cm2 were 2.58 V/µm for Ag TCNQ and 3.13 V/µm for Cu TCNQ. The
difference in the threshold field value can be accounted for by different electron
affinity between Ag TCNQ and Cu TCNQ.
According to the authors, an important advantage of the field emission materials

specified above is the low fabrication temperature and a possibility to fabricate the
field emission cathodes of a large area.
The list of organic materials suitable for fabrication of field emission cathodes
can be expanded with the alloyed organic conductors [193], which can be used with
inexpensive technological processes and reusable masks. The alloyed polypyrrole
was mixed with 5% polyvinyl alcohol and dissolved in water in ratio of 1:3.
A polycarbonate membrane with 3-micron holes was placed in a cleaned
deeply-alloyed substrate of P–Si with the dimensions of 2 2 cm. A drop of the
polypyrrole solution applied on the membrane deeply penetrated inside it in the
open air at ambient temperature. Once dry, the membrane was removed from the
substrate surface. The membrane is hydrophilic, and the solution is fixed on the
filter-pore walls in the course of drying, forming conical craters with sharp edges
that greatly amplify the field.
The field-emissivity characteristics were read in the diode mode at the pressure
of 10−8 mm Hg, the anode-cathode distance being 50 µm. However, stability of the
emission current has not been reported. Nevertheless, paper [194] suggested a
design of a polymer-based field emission cathode.
A few words about liquid-metal field emission cathodes.
Potentially, liquid metals make it possible to generate very high electric fields at
their surface due to electrodynamic instability of the surface [195, 196]. E.g., a
typical size of the asperities developed on the surface of a liquid-metal field
emission cathode is about 10−7 cm, which is quite enough to provide acceptable
field emission parameters [197].
The first experiments on liquid-metal field emission cathodes were carried out as
long ago as in 1933. The cathode material was mercury [198]. Much later, these
papers were continued at the up-to-date level. However, ecological factors [199]
and insufficient field emission prevented the field emission cathodes from practical
application.
In addition, the microelevations can be formed on the surface of a liquid con-
ductor by microwave fields causing excitation of microcapillary waves on the
surface [200]. The resulting microelevations have the characteristic dimensions of
about 10−5–10−6 cm, which makes it possible to achieve a sufficient
field-enhancement coefficient to be used for field emission at a quite small average
power of the microwave field in the resonator.
Excitation of field electrons from the liquid-metal cathode by static and micro-
wave fields is given in [200]. It was found out that in a static field, field emission is
not initiated even if the fields exceed those necessary for field emission with
microwave excitation by more than 10 times. Therefore the threshold microwave
field necessary to originate a microstructure on the surface of a liquid metal is less
than the Tonka-Frenkel crystal field [195, 196]. Detailed evaluations of the
threshold field for microwave excitation confirm this conclusion [200].
References
1. E.E. Martin, J.K. Trolan, W.P. Dyke, Stable, high density field emissiom cold cathode.
J. Appl. Phys. 31(5), 782–789 (1960)
2. W.P. Dyke, F.M. Charbonnier, R.W. Strayer, R. Floyd, J.P. Barbour, J.K. Trolan, Electrical
stability and lie of the heated field emission cathode. J. Appl. Phys. 31(5), 790–805 (1960)
3. M.I. Elinson, G.F. Vasiliev, Way of reduction of ionic bombing, emitting electrons metal
more sharply. A. S. USSR. No. 107388 of Cl.H01j 1/30 of 24.09.55
4. W.P. Dyke, Progress in electron emission at high fields. Proc. IRE 43(2), 162–167 (1955)
5. M.I. Elinson, G.A. Kydinceva, Field emission cathodes on the basis of metal similar
refractory connections. Radio Technician Electron. 7(9), 1511 (1962)
6. M.I. Elinson, G.V. Vasiliv, Field emission. GIFML (1958)
7. V.V. Chesnokov, Electron lamps with field emission cathodes. A. S. USSR. No. 314454 of
27.12.65
8. V.V. Chesnokov, V.M. Gayles, N.A. Morozova, Receiving-amplifying lamp with the cold
field emission cathode. A.S. USSR. No. 339986
9. R. Fisher, H. Noyman, Field emission of semiconductors. Science (1971)
10. N.V. Egorov, Experimental dissertation research of field emission of silicon. Abstract of the
thesis. I. LIE. Leningrad (1973)
11. E.P. Sheshin, Structure of a surface and field emission properties of carbon materials.
M. MFTI. Fizmatgiz (2001)
12. D. Alpert, R.S. Burits, Ultra-vacuum. D. limiting factors on the attainment of very low
pressures. J. Appl. Phys. 25(2), 202 (1954)
13. I.L. Sokolskay, Surface migration of atoms of tungsten in electric field. ZhTF 26(6), 1177
(1956)
14. G.N. Fursey, G.K. Karcev, The stability of field emission issue and migratory processes
preparing development of a vacuum arch. ZhTF 40(2), 310 (1970)
15. P. Staib, K. Ulmer, Maintenance in vacuum of clean metallic surfaces. Phys. Stat. Solidi. 32
(2), K163–K165 (1969)
16. G. Krause, D. Stark, Field emission of chromium, cobalt and hafnium (zur feldemission top
cjarom, ilobalt und hafnium), Z. Phys. 201(1), 69 (1967)
17. A.P. Komar, U.N. Talanin, Pictures of field emission of crystals of carbides of tungsten and
molybdenum. Jnt. AS Suur. Ser. fiz. 22(5), 580 (1958)
18. M.I. Elinson, G.F. Vasiliev, Research of field emission of a geksaborid of lanthanum. Radio
Technician Electron. 3(7), 945 (1958)
19. M.I. Elinson, G.F. Vasiliev, Experemental research of field electronic issue of a geksaborid
of lanthanum. Radio Technician Electron. Eng. 2(3), 348 (1957)
20. YuV Zybenko, I.L. Sokolskay, Field electronic emission of carbonized and thoriated
carbonized tungsten. Radio Technician Electron. Eng. 5(8), 1327 (1960)
21. G.A. Kydinceva, Application of refractory connections as field electronic cathodes. Question
Radionics S1 Electron. 3, 75 (1961)
22. V. Espe, Technology of electrovacuum materials, Energy 2 (1968)
23. E.I. Davydova, F.D. Kirichenko,V.A. Shishkin, Stability of field emission tip cathodes
passivated by films of transitional metals. ZhTF 49(11), 2344–2348 (1979)
24. M.I. Elinson, A.G. Gdan, Field electronic emitter. A. S. USSR. №118913. Kl. HOlj 1/30 of
28.02.58
25. M.S. Mousa, Characteristics of tungsten substrate with Al2O3 coatings under UHV
conditions. Vacuum 38(8–10), 835–838 (1988)
26. M.S. Mousa, Field electron emission studies on zinc oxide coated tungsten microemitters.
Surf. Sci. 266, 110–120 (1992)
27. V.V. Zhirnov, W.B. Choi, I.I. Cuomo, I.I. Hren, Diamond coated Si and Mo field emitters:
diamond thickness effect. Appl. Surf. Sci. 94/95, 123–128 (1996)
References 287
28. W.B. Choi, I.I. Cuomo, V.V. Zhirnov others, Field emission from silicon and molybdenum
tips coated with diamond powder by dielectrophoresis. Appl. Phys. Lett. 68(5), 720–722
(1996)
29. A.V. Crewe, Electron gun using a field emission source. Rev. Sci. Inst. 39(4), 576 (1968)
30. N. Shogo, K. Tsukasa, Remolding of tungsten field emitter. Jap. J. Appl. Phys. 11(4), 604–
605 (1972)
31. I.W. Wiggins, I.A. Zubin, M. Beer, High-resolution scanning transmission electron
microscope at Ions Hopkins. Rev. Sci. Instrum. 50(4), 403–410 (1979)
32. G.N. Fursey, S.N. Shakirova, To a question of possibility of localization of an field emission
in small space angles. ZhTF 33(6), 1125 (1966)
33. L.W. Swanson, Comparative study of the zincroniated and built-up W. thermal-field
cathode. J. Vac. Sci. Tech. 12(6), 1228 (1975)
34. T.A. Baker, M.M. Balsiger, K.T. Considine, H.E. Litsjo, Electron-beam tube including a
thermonic-field emission cathode for a scanning electron microscope. Pat. №3809899 (USA)
35. S. Hosoki, S. Yomamoto, H. Todokoro others, Field emission cathode and method of
fabricating the same. Pat. USA. №4379250. CL. 313/336 (HOl l/16) of 5.04.1983
36. E.I. Davydova, A.D. Karpenko, V.A. Shishkin, B.I. Pab, A.c. №439028 (USSR). Way of
production of field emission cathodes. 1974. No. 29
37. G.A. Kudintseva, V.P. Savchenko, B.I. Pab, A.c. №424254 (USSR). Way of production of
needle field emitters. 1974 No. 14
38. E.A. Litvin, A.A. Starobinec, Limit currents of field emission. ZhTF 47(10), 2032 (1977)
39. A.D. Karpenko, U.S. Lobachev, V.A. Shishkin, About limit current of the one tip field
emission cathode. Radio Technician Electron. Eng. 21(1), 207 (1976)
40. M.I. Elinson, V.A. Gorkov, A.A. Yaciopolskay , G.A. Kydinceva, Research of pulse field
emission at the high density of currents. Radio Technician Electron. Eng. 5(8), 1318 (1960)
41. J.P. Barbour, W.W. Dolan, J.K. Trolan, E.E. Martin, W.P. Dake, Space-charge effects in
field emission. Phys. Rev. 22(1), 53 (1953)
42. T.J. Lewis, Theoretical interpretation of a field emission periments. Phys. Rev. 101(6), 1694
(1956)
43. E.G. Shirokov, Field emission of system tungsten tips. Radio Technician Electron. T9(7),
p1320 (1964)
44. A.A. Golov, G.N. Fursey, I.D. Ventova others, Field emission cathode type “comb”. A.S.
USSR. No. 342241. C. HOlj 1/30 of 14.06.72
45. V.M. Gykov, D.M. Paytov, S.A. Polegaev, other, Way of production of the Many tip field
emission cathode. A.S. USSR. No. 1001225. Cl.HOlj 9/02 of 28.02.83
46. P. Genequand, Pat. Switzerland. №603005. C. HOlj 1/30 of 15.08.78
47. I.D. Ventova, L.E. Valuyeva, B.I. Pab, A. S. №416784 (USSR), Way of production of many
tip field emission cathode. 1974, No. 7
48. V.G. Pavlov, A.S. Rabinovich, V.L. Savchenko, V.N. Shrednik, B.I. Pab, A. S. №493834
(USSR), Way of blunting tip cathode. 1975. No. 44
49. B.V. Okylov, I.Z. Gleyzer, L.T. Dronova, Manufacturing techniques of multiemitter cold
cathodes. PTE, (6), 124–125 (1973)
50. J. Shelton, Field effect electron emitter. Pat. USA. №3745402. Cl. 313/309 (HOlj 1/30) of
10.07.1973
51. I. Shelton, Field effect electron gun having at least a million emitting fibers per square
centimeter. Pat. USA. №3783325. C. 313/336 (HOlj 1/02) of 1.01.1974
52. I. Shelton, Edge effect elimination and beam forming designs for field emitting arrays. Pat.
USA. №434181. Cl. 313/309 (HOlj 1/02) of 17.08.82
53. W.L. Less, Method and apparatus for liming-emission current. Pat. USA. №3671798. Cl.
313/336(HOlj 1/16) of 20.06.1972
54. K. Inoue, K. Bersui, T. Nakatani, O. Toyoda, Field emission cathode and process for
producing the same. Pat. USA 2003/0184203 Kl. 313/309 (Hoij 9/04 of 2.10.2003)
55. V.F. Bibik, P.G. Borzak, A.F. Yshenko, Photofield cathode. A. S. USSR. №584667. Cl.
Ho1j 9/09
56. R.N. Thomas, R.A. Wickstrom, D.K. Schroder, H.C. Nathanson, Fabrication and some
applications of large-area silicon field emission arrays. Solid State Electron. 17, 155 (1975)
57. A.I. Klimin, A.A. Mostovsky, I.A. Pystylnik, D.A. Sakseev, L.P. Titova, N.M. Eden, Field
emission research one and many tips silicon catode. Izv. Acad. Sci. USSR. Ser. fiz. 40(8),
1575 (1976)
58. V.F. Bibik, P.G. Borzak, A.F. Ycenko, Germanium and silicon field emission photo cells.
Ukr. Phys. Zh. 13(5), 868 (1968)
59. X.S. Natanson, Y. Goldberg, Matrixes tips for field emission systems.:fizik of thin films, 8
(M.:mir, 1978), p. 264
60. D.N. Schroder, R.N. Thomas, J. Vine, H.C. Hathanson, The semiconductor field emission
photocathodes. JEEE Trans. Electr. Devices 21(12), 785–798 (1974)
61. R.N. Thomas, H.C. Hathanson, Photosensitive Field emission from silicon point arrays.
Appl. Phys. Lett. 21(8), 384–389 (1972)
62. R.Z. Bahtizin, E.K. Ratnikova, V.P. Petrakov, Way of production of the many tips field
emission cathodes. A. S.. USSR No. 1215536. Cl. HOij 1/30
63. R.Z. Bahtizin, E.K. Ratnikova, Edektronny equipment. Ser. 2. Semicond. Device 7(173),
124 (1984)
64. R.Z. Bahtizin, E.K. Ratnikova, G.G. Kylishova, A.F. Yshenko, V.P. Petrakov, Production
germanium many tips structures and research of their emission characteristics. News Acad.
Sci. USSR. 49(9), 1738–1740 (1985)
65. R.Z. Bakhtizin, Field emission from semiconductor point arrays. I. De Physique (Colloque
P. 6). 49(11), 155–160 (1988)
66. P.G. Borzak, E.I. Givargizov,G.G. Kylishova, I.E. Lifshic, A.N. Stepanova, A.F. Yshenko,
News of academy of sciences of the USSR. Ser. Fiz. 40(9), 1571 (1976)
67. C.A. Spindt, K.R. Shoulders, L.N. Heynick, Field emission cathode structures and devices
utilizing such structures. Pat. USA. №3755704. Cl. 313/309 (Hoij 1/16 of 28.08.1973)
68. C.A. Spindt, K.R. Shoulders, L.N. Heynick, Field emission cathode structures and devices
utilizing such structures, and methods of producing such structures. Pat. USA. №3789471.
Cl. 29/25.17 (Hoij 9/02 of 5.02.74)
69. C.A. Spindt, L.N. Heynick, Method of producing field ionizer and field emission cathode
structures. Pat. USA, №3812559 Kl. 29/25.18 (Hoij 9/02 of 28.05.74)
70. C.A. Spindt, I. Brodie, L. Humphrey, E.R. Westerberg, Physical properties of thin-film field
emission cathode with molybdenum cones. J. Appl. Phys. 47(12), 5248–5263 (1976)
71. C.A. Spindt, in Field emission arrays for high-current density, and high frequency
operation. Teth. Digest of 18th. IVNS. (Oxford, UK, 2005), p. 20–21
72. I.M. McCaulay, C.A. Spindt, C.E. Holland, I. Brodie, Field emission cathode array coated
with electron work function reducing material, and method. Pat. USA. №5089292. Kl.
427/78. (BO5D 5/12 of 18.02.1992)
73. M.S. Mousa, C.E. Holland, J. Brodie, C.A. Spindt, The effect of hydrogen and acetylene
processing on microfabricated field emitter arrays. Appl. Surf. Sci. 67, 218–221 (1993)
74. M.S. Mousa, A study of the effect of hydrogen plasma on microfabricated field- emitter
arrays. Vacuum 45(2/3), 235–239 (1994)
75. H. Li, M.Q. Ding, J.J. Feng, X. Li, G. Bai, F. Zhang, in Fabrications and characterization of
spindt-type field emission arrays coated with ZrC thin films. Thech. Digest IVNC. (Oxford,
UK, 2005), p. 237–238
76. C.C. Chang, Field emission display cathode plate with an internal via and the fabrication
method for the cathode plate. Pat/ USA 2002/0105261. Cl. 313/495 H01j 63/04 of 8.08.2002
77. L. Dvorson, A.J. Akinwande, Double-gated Spindt emitters with stacked focusing electrode.
J. Vac. Sci. Tech. B 20(1), 53–59 (2002)
78. L.Y. Chen, A.I. Akinwande, in Negative resistance region in double-gated silicon field
emission arrays. Tech. Digest JVNC. (Oxford, UK, 2002), p. 15–16
79. Y. Gotoh, M. Nagao, D. Nozaki others, Electron emission properties of Spindt-tipe platinum
field emission cathodes. J. Appl. Phys. 95(3), 1537–1549 (2004)
References 289
80. M.Q. Ding, X. Li, G. Bai other, Spindt-type cathodes with aligned carbon nanotube emitters.
JVESC 117–119 (2004)
81. L.R. Baylor, D.H. Lowndes, M.L. Simpson others, J. Vac. Sci. Technol. B. 20, 2646 (2002)
82. X. Yang, Simpson M.L., Rondolph S.J others, Integrated tungsten nanofiber field emission
cathodes selectively grown by nanoscale electron beam-induced deposition. Appl. Phys. Let.
86, 183106 (2005)
83. A. Jankowski, J.P. Hayes, Process system and method for fabricating field emission cathode.
U.S. patent №5746634 of 5.05.1998
84. A. Jankowski, J. Hayes, J. Morse, J. Ferreira, Large area deposition of field emission
cathodes for flat panel displays. Thin Solid Films 355–356, 194–198 (1999)
85. Y-F. Bai, S. Ren, S.Z. Deng, J. Chen, N-S. Xu, in Preparation of ZnO nanowires and there
field emission properties. Tech. Digest of JVNC. (Oxford, UK, 2005), p. 178–179
86. C.Y. Li, X. Liu, J. Chen, S.Z. Deng, N.S. Xu, S. Yang, in Uniform field emission from CuO
nanowires prepared by thermal oxidation method. Tech. Digest JVNC. (Oxford, UK, 2005),
p. 213–214
87. P. Feng, X.Q. Fu, S.Q. Li, Y.G. Wang, T.H. Wang, Stable electron field emission from
triangular-shaped ZnO nanoplate arrays with low local heating effect//Nanotechnology.2007.
V.18.165704
88. A.J. Pedzaza, J.D. Fowlkes, D.H. Lowndes. Appl. Phys. Lett. 77, 3018 (2000)
89. A.V. Karabutov, V.D. Frolov, E.N. Loubnin, A.V. Simakin, G.A. Stapheev. Appl. Phys.
A 76, 413 (2003)
90. V. Zorba, P. Tzanetakis, C. Fatakis others, Silicon electron emitters fabricated by ultraviolet
laser pulses. Appl. Phys. Lett. 88(08), 103 (2006)
91. Zorba V., I. Alexandrou, I. Zergioti others, Thin Solid Films. 453–454, 492 (2004)
92. V.S. Fomenko, Emission properties of materials. Naukova Thought, Kiev (1981)
93. I.N. Slivkov, Electrical insulation and the blunting in vacuum. Atomizdat (1972)
94. W.D. Dyke, Field emission, a newly practical electron source. IRE Trans. Mil. Electron.
MIZ 4(1), 38 (1960)
95. E.G. Shirokov, Field emission cathode. C.S. №180712. in Opening, Inventions, Industrial
Samples, Trademarks, 8, 52 (1966)
96. E.G. Shirokov, Technique of preparation and control emission surface of a tungsten edge of
a disk-shaped form. Изв. Acad. Sci. USSR Ser. Tekh. Nauk. 6(2), 45 (1965)
97. B.V. Bondarenko, V.I. Makyha, The field emission cathode from steel edges. PTE (4), 235–
236 (1965)
98. B.V. Bondarenko, V.I. Makyha, A.S. Gaydarov, Research field electronic edge of emitters of
a disk-shaped form. Radio Eng. Electron. 17(12), 2634–2637 (1972)
99. V.I. Mahov, N.A. Dugev, A.I. in Kozlov, Cathodic Knot With Silicon Matrix Field Emission
Cathodes. The xx vsesoyuzn. konf. on emission electronics: Tez.dokl. Kiyev, (1987) p. 232
100. A.A. Vasenkov, N.A. Dughev, V.I. Mahov, Low-voltage silicon matrix field 1 emission
cathodes. Electron. Indus. (8), 48–49 (1987)
101. N.A. Dughev, A.I. Kozlov, V.I. Mahov, V.M. Serovayskiy, Speed of operation microelec-
tronic triode structures with field emission cathodes. Microelectronics 19(5), 478–485 (1990)
102. V.I. Volosov, V.N. Lazarev, V.E. Teraev, About some features of field emission of
cylindrical cathodes. ZhTF 40(4), 855 (1970)
103. V.A. Kyznetsov,A.N. Kyrochkin, Field e emitter. A. S. USSR №1078492. Cl. H01j 1/30, of
07.03.84
104. V.V. Chesnov, Electronic equipment. Series. (4), 3 (1968)
105. R.L. Akopyn, B.V. Bondarenko, V.I. Makyha, Radio Technician Electron. 16(11), 2236
(1971)
106. B.V. Bondarenko, R.L. Akopyn, Calculation of emission characteristics of film field
emission cathodes. Radio Technician Electron. Eng. 17(5), 1059 (1972)
107. B.V. Bondarenko, Ways of increase of stability of emission and life of field emission
cathodes. Electronic equipment. Series 1(6), 74–82 (1973)
108. B.V. Bondarenko, V.I. Makyha, About durability of the film emission cathode from chrome.
Radio Technician Electron. Engineer. XVII(8), 1683–1686 (1972)
109. E.A. Gulbranson, K. Andern, J. Electrochem. Soc. 99, 402 (1959)
110. B.V. Bondarenko, V.I. Makyha, A.C. Gaydarov, The film emission cathode from the titan.
Radio Technician Electron. 18(10), 2212 (1973)
111. B.V. Bondarenko, L.A. Kirichenko other, Research of field emission foil and films and the
mechanical durability of microinhomogeneities on their surface. Summary of reports of
All-Union conference on issue electronics. Makhachkala (1976)
112. B.V. Bondarenko, L.A. Kirichenko, N.D. Konovalov, Field emission thin foil refractory
metals. Theses of reports of the All-Union symposium on not heated cathodes. Tomsk
(1980)
113. B.V. Bondarenko, V.I. Makyha, A.C. Gaydarov, Research of pulse field electronic emission
of film cathodes from chrome. Radio Eng. Elektronika 17(8), 1771–1772 (1972)
114. V.V. Chesnokov, Ways of production of field emission film electrode system. A. S. USSR
No. 174727. Cl. H01j/30, of 08.07.64
115. N.A. Dughev, A.I. Kozlov, A.A. Kadnikov, V.I. Mahov, I.V. Panaev, Influence of
geometrical parameters of silicon film field emission cathodes on their emission character-
istics. Radio Technician Electron. (11), 2385–2389 (1990)
116. L.D. Karpov, Field-effect emitter device. European patent №0681311. Cl. H01j/30, of
08.11.95
117. L.D. Karpov, Edge emitter display device. Pat. USA. №5965971. Cl. 313/309 (H0Ij 01/02),
of 12.10.1999
118. L.D. Karpov, Edge emitter with secondary emission display. Pat. USA. № 6023126. Cl.
313/310(H01j 1/05, of 8.02.2000
119. S.A. Gavrilov, E.A. Ilichev, E.A. Poltorackiy, N.V. Syetin other. The lateral emitter as basic
Element of integrated emission electronics. Lett. ZhTF 30(11), 48–53 (2004)
120. S.A. Gavrilov, N.N. Dzbanovsky, V.V. Dvorkin et al., in Proceedings of 11th International
Symposium, Nanostructures: Physics and Technology, p. 234–236, St. Petersburg, Russia.
23–28 June 2003
121. S.A. Gavrilov, N.N. Dzbanovsky, V.V. Dvorkin et al. in Abstract International Conference
«Micro-and nanoelectronics-2003», p. 01–27, Moscow-Zvenigorod, Russia, 6–10 October
2003
122. S.A. Gavrilov, E.A. Ilichev, E.A. Poltorackiy, N.V. Syetin, Emitters from carbon nanotubes
for planar emission vacuum micro and a nanoelectronics. Lett. ZhTF 30(14), 75–81 (2004)
123. Givargizov E.I. Growth of threadlike and lamellar crystals from steam. Science, 304c (1977)
124. G.V. Bereghkova, Threadlike crystals. M.: Science, 160c (1969)
125. S.A. Ammer, V.S. Postnikov, Threadlike crystals (Polytechnic University, Voronezh, 1974),
p. 284c
126. V.G. Syrkin, Gas-phase metalliztion through carbonyls. Metallurgy, 264c (1985)
127. V.G. Syrkin., Carbonyls of metals. Khimiya, 200c (1983)
128. B.G. Gribov, G.A. Domrachev, B.V. Ghyk other. Sedimentation of films and coverings
decomposition of metalloorganic connections of M of a.:nauk, (1981), p. 322
129. A.A. Nosov, T.A. Poshehonova, P.V. Poshehonov, About the education mechanism the
filiform of monocrystals on electrodes of electronic devices. RVE, 18(9), 199–1999 (1973)
130. L.V. Poshehonov, D.K. Nosova, A.A. Nosov, in Growth of Filiform Crystals of an Oxide of
Chrome. Materials of the All-Union Conference “Filiform Crystals for New Equipment”.
Voronezh VPK, 73 (1979)
131. R. Gomer, Field emission from mercury whiskers. J. Chem. Phys. 28(3), 457–469 (1958)
132. A.A. Nosov, T.A. Poshehonova, P.V. Poshehonova, Investigation of conditions of
destruction of filiform monocrystals in high-voltage vacuum devices at influence strong
electric fields. 15(10), 2156–2161 (1970)
133. P.V. Poshehonov, V.M. Gennadiev, N.P. Ovsynnikov, Methods of a increasing of efficiency
of training of field emission cathodes on the basis of threadlike monocrystals. RIE 21(12),
2568–2573 (1976)
References 291
134. P.V. Poshehonov, N.P. Ovsannikov other, Investigation of process of production of effective
field emission cathodes when forming many tips system of monocrystals of gold out of the
working volume of the device. In kN.:tex. rep. An all-Union seminar on not heated cathodes,
Tomsk, p. 33 (1977)
135. P.V. Poshehonov, N.P. Ovsannikov, V.M. Genadiev, N.N. Chadaev, Influence of contact
with the gas environment on emission ability of the field emission cathode formed by system
of filiform crystals. Vacuum and gas-discharge electronics, Ryazan, RRTI, p. 73–77 (1981)
136. E.I. Givargizov, G.A. Gydingeva other, Many tips field emission cathodes from a
single-crystal geksaborid of lanthanum. In kN.:tes.rep. Hush Vsesoyuznyu Konferentsii on
emission electronics, M.:nauka, p. 234 (1981)
137. N.G. Borsak, E.I. Givargizov other, Emission properties the many tips of field emission
cathodes six Gei-x from the made crystallization from a steam phase. Izv. Acad. Sci. USSR
40(7), 1571–1580 (1976)
138. F. Okuyama, Vapor-grown tungsten whiskers induced by vacuum discharges. J. Appl. Phys.
45(10), 4239–4241 (1974)
139. H.B. Linden, H.D. Beckey, F. Okuyama, On the mechanism of cathodio growth of tungsten
needles by decomposition of hexacarbonyl tungsten under high field conditions. Appl. Phys.
22(1), 83–87 (1980)
140. F. Okuyama, T. Shibata, I. Yasuda, tungsten needle produced by decomposition of
hexacarbonyl tungsten. Appl. Phys. Lett. 35(1), 6–7 (1979)
141. F. Okuyama, Growth of metallio w wiskers crystals incorporated with field electron
emission. Appl. Phys. Iott. 36(1), 46–47 (1980)
142. F. Okuyama, Cathodio needle growth from Mo (Co)6 and Cr(Co)6 vapors at lower electric
fields. Appl. Phys. 22(1), 39–46 (1980)
143. F. Okuyama, Cathodio needle growth from Mo(Co)6 vapors at higher electric fields. Appl.
Phys. 27(1), 57–64 (1982)
144. F. Okuyama, Growth of Cr needle crystals induced by field electron emission. J. Appl. Phys.
22(2), 245–251 (1984)
145. A.A. Nosov, N.P. Ovsannikov, G.N. Shyppe, Direct electron microscopy supervision of
dynamics of formation of the many tips field emission cathode on the basis of tungsten.
ZhTF 54(2), 372–374 (1984)
146. V.G. Syrkin, Carbonyl metals. M. Metalurgizdat, p. 255 (1978)
147. V.E. Ivanov other, Crystallization the refractory of metals from a gas phase. M. Atomizdat of
1974, p. 212
148. T.A. Razyvaev other, Metalloorganic connections in electronics. M. Sci. 479 (1972)
149. V. Barton, N. Rabrera, F. Frank, Elementary processes of growth of crystals (1959)
150. R. Vagner, Growth of crystals on the PKR mechanism. Monokristalny fibers and materials
reinforced by them. M.:Mir, pp. 42–57 (1973)
151. M.M. Pogorelskiy. Lett, ZhTF 4, 1068 (1978)
152. A.A. Nosov, D.I. Nosova, N.P. Ovsannikov, N.N. Chadaev, Limit density of field emission
current of tip emitter of submicron dimension which are grown up in various ways. Radio
Technician Electron. XXXI(7), 1466–1467 (1986)
153. N.I. Tatarenko, V.F. Kravchenko. Field emission nanostructures and devices on their basis.
M. Fizmatgiz (2006)
154. V.A. Labynov, B.A. Sokol, N.I. Tatarenko, The electrovacuum device. A. S. USSR
No. 713386. H01j 21/10 of 05.10.1979
155. N.I. Tatarenko, in Development of an Integrated Technique for Fabricating Thin-Film Field
Emission Microelectronic Devices. The Third International scientific and technical
conference: Modern technology of hybrid integrated chips, including elements of
superconductor electronics. The Belorussky state univrsitt informatics and radio electronics.
Naroch, Conference Materials, pp. 159–160 (1994)
156. N.I. Tatarenko, Way production of the thin-film vacuum microdevice. Patent USSR
No. 1729243. HO1j 21/10 of 22.12.1991
157. F. Keller, M.S. Hunter, D.L. Robinson, Structural features of oxide coatings on aluminum.
J. Electrochem. Soc. 100(9), 411–419 (1953)
158. J.P. OSullivan, G.C. Wood, The morphology and mechanism of formation of porous anodic
films on aluminium. Proc. R. Soc. London A 317, 511–643 (1970)
159. G.E. Thompson, Porous anodic alumina: fabrication, characterization and applications. Thin
Solid Films 297, 192–201 (1997)
160. A.P. Li, F. Mueller, A. Birner, K. Nielsh, U. Gosele, Hexagonal pore arrays with a 50–
420 nm interpore distance formed by self-organization in anodic alumina. J. Appl. Phys. 84
(11), 6023-6026 (1998)
161. H. Masuda, M. Nakao, T. Tamamura, H. Asoh, Fabrication of ordered nanostructure based
on anodic porous alumina. The SPIE Conference on Micromachining and Microfabrication
Process Technology IV.-SPIE Vol.3511,1998.-P.74–81
162. A.P. Li, F. Mueller, A. Birner, K. Nielsh, U. Gosele, Fabrication and mi-crostructuring of
hexagonally ordered two-dimensional nanopore arrays in anodic alumina. Adv. Mater. 11
(6), 483–487 (1999)
163. D. Crouse, Y.H. Lo, A.E. Miller, M. Crouse, Self-ordered pore structure of anodized
aluminum on silicon and pattern transfer. Appl. Phys. Lett. 76(1), 49–51 (2000)
164. L. Ba, W.S. Li, Influence of anodizing conditions on the ordered pore formation in anodic
alumina. J. Phys. D: Appl. Phys. 33, 2527–2531 (2000)
165. H. Assoh, K. Nishio, M. Nakao, T. Tamamura, H. Masuda, Conditions for fabrication of
ideally ordered anodic porous alumina using pretextured Al. J. Electrochem. Soc. 148(4),
B152-B156 (2001)
166. H. Masuda, M. Yotsuya, M. Asano, K. Nishio, M. Nakao, A. Yokoo, T. Tamamura,
Self-repair of ordered pattern of nanometer dimensions based on self-compensation
properties of anodic porous alumina. Appl. Phys. Lett. 78(6), 826–828 (2001)
167. C.Y. Liu, A. Datta, Y.L. Wang, Ordered anodic alumina nanochannels on
focused-ion-beam-prepattened aluminum surfaces. Appl. Phys. Lett. 78(1), 120–122 (2001)
168. A. Cai, H. Zhang, H. Hua, Z. Zhang, Direct formation of self-assembled nanoporous
aluminium oxide SiO2 and Si substrates. Nanotechnology 13, 627–630 (2002)
169. W.L. Xu, M.J. Zheng, S. Wu, W.Z. Shen, Effects of high temperature annealing on structural
and optical properties of highly ordered porous alumina membranes. Appl. Phys. Lett. 85
(19), 4364–4366 (2004)
170. N.V. Myung, J. Lim, J.P. Fleurial, M. Yun, W. West, D. Choi, Alumina nanotemplate
fabrication on silicon substrate. Nanotechnology 15, 833–838 (2004)
171. N.I. Tatarenko, T.N. Andrushenko, Influence of conditions of anodizing on a microstructure
of porous anode oxide films of aluminum. Prot. мeтaллoв (3), 499–501 (1984)
172. V.A. Labynov, N.I. Danilovich, A.C. Yksysov, V.E. Minaychev, Modern magnetron
sputtering devices. Foreign Electron. Equipment 10, 3–62 (1982)
173. A.N. Govyadinov, S.A. Zakhvitcevich, in Field Emitter Arrays Based On Natural
Seiforganized Porous Anodic Alumina. 10th International Vacuum Microelecronics
Conference. Kyongju, Korea, pp. 735–737 (1997)
174. S. Zakhvitcevich, A. Govyadinov, Nanodimensional field emitter arrays based on porous
anodic alumina. Phys., Chem. Appl. Nanos-truct. 297–300 (1997)
175. E.R. Holland, M.T. Harrison, M. Huang, P.R. Wilshaw, Nonlithographic technique for the
production of large area high density gridded field emission sources. J. Vac. Sci. Technol.
B 17(2), 580–582 (1999)
176. R.R. Greene, H.F. Gray, P-N junction controlled field emitter array cathode. United States
Patent № 4 513 308, Int. Cl. H01L 29/06- H01L 29/34; H01L 27/12, Apr.23,1985
177. N.I. Tatarenko, A.A. Blyablin, N.V. Suetin, in Field Emission Nanostructures Formed By
Self-Organization In Anodic Porous Alumina. 12th International Vacuum Microelectronics
Conference.-Darmstadt, Germany, (1999), pp. 136–137
178. A.A. Blyablin, N.V. Suetin, D.B. Suyatin, E.S. Soldatov, N.I. Tatarenko, in STM u ATM
Study of a Nanoscale Field Emission Array Formed by Self-Organization in Porous Anodic
References 293
Aluminium Oxide. 2nd Europen Field Emission Workshop. Seqovia, Spain, (2000), pp. 176–
177
179. N.I. Tatarenko, A.O. Vorobyev, Volt-ampere characteristics and differential parameters
nanostructural field emission of microdevices. Electromag. Waves Electron. Sistemy 10(9),
68–74 (2005)
180. X.B. Yan, S. Xu, H.W. Liu, S.R. Yang, Field emission properties of polymer-converted
carbon films by heat treatment. Solid State Commun. 133(2), 113–116 (2005)
181. J. Alexandrou, E. Kymakis, G.A.J. Amazatunga, Polymer-nanoyube composites: burying
nanotubes improves their field emission properties. Appl. Phys. Lett. 80(8), 1435–1437
(2002)
182. J.Y. Shim, E.J. Chi, H.K. Baik, K.M. Song. Thin Solid Films 355–356, 223 (1999)
183. H. Busta, D. Furst, A.T. Rakhimov, V.A. Samorodov, B.V. Seleznev, N.V. Suetin, A.
Silzars. Appl. Phys. Lett. 78, 3418 (2001)
184. Y.W. Jin, J.E. Jung, Y.J. Park others, Triode-type field emission array using carbon
nanotubes and a conducting polymer composite prepared by electrochemical polymerization.
J. Appl. Phys. 92(2), 1065–1068 (2002)
185. Y. Liu, S. Fan, Enhancement of field emission properties of cyanoacrylate-carbon nanotube
arrays by laser treatment. Nanotechnology 15, 1033–1037 (2004)
186. A.N. Ionov, E.O. Popov, V.M. Svetlichny, A.A. Pashkevich, Influence of a thin polymeric
covering on field emission properties of flat metal cathodes. Lett. ZhTF 30(13), 77–82
(2004)
187. A.N. Ionov, E.O. Popov, V.M. Svetlichnyi, M.N. Nikolaeva, A.A. Pashkevich, Field
emission from metal/polymer construction. Suf. Interface Anal. 39, 159–160 (2007)
188. A.N. Ionov, V.A. Zakrevskii, V.M. Svetlichnyi, R. Rentzsch. In Abstract of 10th
International Symposium Nanostructure: Physic and Technology. (St.Petersburg, Russia,
2002), June 17–21.. p. 531
189. J. Yilgor, B.C. Johnson. Amer. Chem. Soc. Polym. Prepr. 27, 54 (1986)
190. J. Lowell, A.C. Rose-Innes, Adv. Phys. 29, 947 (1980)
191. H. Liu, Q. Zhao, Y. Li, F. Lu others, Field emission Properties of Large-Area Nanowires of
organic Charge Transfer Complex. J. Am. Chem. Soc. 127, 1120–1121 (2005)
192. J. Kymissis, A.J. Akinwande, Field emission from a patterned organic conduction
composite, Appl. Phys. Lett. 82(14), 2347–2349 (2003)
193. W. Eccleston, G.A.J. Amaratunga, J. Musa, Field emission cathode and field emission
display. Pat.USA. №6853127 of 8.02.2005. Cl. H01j 1/62(313/495)
194. L. Tonks, A theory of liquid surface rupture by uniform electric field. Phys. Rev. 48, 562–
568 (1935)
195. Y. Frenkel. On a theory of liquid surface instability in electric field in vacuum. Zh. Tekh.
Fiz.6, 347–351 (1935)
196. K. Hata, R. Ohya, S. Hishigaki, H. Tamura, T. Noda, Stable field emission of electrons from
liquid metal. Jpn. J. Appl. Phys. 26(6), L896–L898 (1987)
197. J.W. Beams, Field electron emission from liquid mercury, Phys. Rev. 44(15), 803–807
(1933)
198. J. Metterauer, Field emission from thin liquid metal films. Appl Surf. Sci. 94–95, 161–170
199. L.M. Baskin, A.A. Kantonistov, G.N. Fursey, L.A. Shirochin, Peculiarities in explosive
emission of liquid metals in microwave field. Rep. USSR Acad. Sci. (USSR) 296(6), 1352
(1987)
200. A.A. Kantonistov, I.N. Radchenko, G.N. Fursey, L.A. Shirochin, Field emission of liquid
metals in alternating fields. Collogue de Physique, Colloque C8, supplement au Tome 50.
C.8. P.203–207.1989
Chapter 6
Carbon-Based Field-Emission Cathodes
Abstract This chapter reviews the main carbon-based materials that, by the
authors’ opinion, are suitable for use in the manufacturing of field emission cath-
odes. Those are, first and foremost, polyacrylonitrile carbon fibers, carbon nan-
otubes, fullerenes, graphenes etc. The ways to improve the emission characteristics
of carbon-based field emission cathodes and techniques of measurement of
volt-ampere characteristics out of such materials are also discussed.
Recent years have seen a great deal of research dedicated to carbon field emission
and carbon-based field emission cathodes. Monograph [1], issued in 2001, was
supplemented with a lot of new data, so it was decided to dedicate a separate
chapter to carbon-based field emission cathodes to append the data published in [1].
It should be noted that all carbon materials used as field-emission cathodes fall
within the category of nanostructured materials. Field emission of such materials
comes from the micro-elevations sized from 0.1 to 100 nm.
This chapter reviews the today’s areas in development of carbon-based field
emission cathodes.
6.1 Carbon Fibers
6.1.1 Polyacrylonitrile Carbon Fibers
The basic features of field emission of carbon-based, viz. polyacrilonitrile fiber,

field emission cathodes have been widely reviewed in literature [1–6].
So this section considers field emission of both traditional carbon fibers and
some advanced developments in this area.
It is well-known [7] that field emission cathodes of polyacrylonitrile carbon-fiber
bundles can generate the emission current of 1–1.5 mA. However, no attempts to
improve the current draw of such cathodes by increase in the number of the fibers
making up the cathode have not brought the targeted result. The reason is that the

296 6 Carbon-Based Field-Emission Cathodes
bundle fibers are unequally spaced both in terms of their height and the field
emission cathode end-face area. Such inequality in arrangement of the fibers causes
even greater inequality of field-emission current from them due to the exponential
dependence of the current on the electric field intensity, and some fibers carrying
the basic current loading are destroyed. The clouding plasma [8] short-circuits other
emission centers, thereby causing a series of electric discharges on the cathode,
which finally degrades the cathode.
Graphite field emission cathodes with a well-developed emitting surface could
generate the emission current of 10 mA [9] at the average current density of less
than 10−2 A/cm2 from the cathode area of about 1 cm2.
Fiber structure. Polyacrylonitrile carbon fiber is known [10] to consist of
deeply intertwined filamentary fibrils reaching 1 µm in length and 1–5 nm in
diameter. Being the emitting surface of the field emission cathode, the end surface
of fibers is a set of chaotically located microelevations of different curve radii,
height and configuration (Fig. 6.1).
The surface structure results from mechanical cleave of a carbon fiber, when the
fibrils making up the fiber occasionally break off, thus forming a set of microele-
vations. When electrical voltage is applied, the numerous microelevations are sure
to include a few where electric intensity at the vertexes peaks. Such microelevations
are the initial emission centers. With further increase in voltage, electric intensity is
also sufficient for emission from other, less pointed microelevations.
Initial distribution of the emitting centers over the end surface of the fiber is
absolutely arbitrary and does not coincide even for quite closely-located sites of the
same fiber because of the spread of mechanical properties along the fiber axis. This
fact accounts for a wide spread of initial emission characteristics of carbon-based
field emission cathodes, which can amount to several orders of magnitude.
Heat treatment temperature. Field-emission characteristics of carbon materials
essentially depend on heat-treatment temperature because the latter affects the
mechanical characteristics, thermal conduction and electric conductivity of these
Fig. 6.1 Working surface of

the new prepared field
emission cathode out of
6.1 Carbon Fibers 297
Fig. 6.2 Dependence of limit

field emission current with the
polyacrylonitrile carbon fibers
on processing temperature
materials [10]. That accounts for the wide spread of the emission-current values
mentioned in different references, e.g., [11–13].
Measurement of the limiting field-emission current of polyacrylonitrile carbon
fiber, viz. the emission current that the cathode can steadily endure for 1 h, shows
that increase in heat-treatment temperature from 1500 to 2600 °C increases the
limiting field-emission current from 120 to 190 µA [14] (Fig. 6.2). This depen-
dence, based on research of more than 40 samples, is evaluative because some
particular field emission cathodes developed the emission current as high as 230–
300 µA. However, field emission cathodes are only failsafe in operation if the
values are lower than those specified in Fig. 6.2.
Electronic work function. The electronic work function of different carbon
materials, viz. carbon fiber, natural and technical graphite, hovers within
4.75 ± 0.05 eV [15]. Variations of the electronic work function of carbon materials
in a wide range of residual-gas pressures, viz. from ultrahigh vacuum to atmo-
spheric pressure, are relatively insignificant and amount to a few percent of the
initial value; whereas similar variations of the electronic work function of a number
of metals, such as tungsten, amount to a few tens of percent, which causes abrupt
changes in the emission current, reaching several orders of magnitude. The latter is
very important for practical use of carbon materials as field emission cathodes
because it accounts for their stable and long functioning under different vacuum
conditions, including high technical vacuum at the pressures of about 10−7–
10−7 mm Hg. Moreover, the carbon materials studied show that their electronic
work function remains almost the same after the samples are ground and aged in air.
Service life. Service-life testing of carbon-based field emission cathodes shows
their high durability in high technical vacuum. In particular, the operating time of
carbon polyacrylonitrile fiber is 7500 h [16] at the current draw of 60 µA and that
of fiber bundles is more than 500 h in the cyclic mode. Carbon fibers were
repeatedly tested for 1000 h at the currents of 20, 50, 100, 150, 200 µA [17] to
show stabilization of field-emission characteristics of carbon fibers in long-term
operation. Dependence of the instability parameter on field-emission current for

parameter s of field emission
current on time: 1 initial; 2
after 100 h of work; 3 after
500 h of work
different treatment periods is given in Fig. 6.3. The instability coefficient is deter-
mined for 5 min by the formula of
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ffi
N 2 2
r¼ I ðI Þ ð6:1Þ
N1
where I is the average current value; N being the number of measurements.

A similar dependence expressing a drop in instability of field-emission current in
terms of the operating time for fixed current amounts to less than 2% for poly-
acrylonitrile carbon fiber. Generally, experimental research in sealed-off devices
confirms good durability of such field emission cathodes [18].
Forming. According to [12, 19], mere extrapolation of metal field emission
cathode fabrication methods to carbon materials cannot give positive results.
A carbon-based field emission cathode fabrication cycle cannot but include a
cathode-forming operation [20]. Forming is done by step-up of the emission current
drawn from the cathode with time exposure at each step until the cathode surface
has been restructured. The effect of forming on distribution of field-emission current
over the cathode surface is shown in Fig. 6.4 to demonstrate different forming steps.
It is obvious that the emission-center dispersion degree is different, which signifi-
cantly affects stability of the field emission cathode. Otherwise stated, the same
current is achieved by contribution of a different number of emission centers
(Fig. 6.4a–c), which reduces the average current density per emission center, thus
reducing the microelevation destruction probability. So the physical sense of
forming of a carbon-based field emission cathode is to create a maximum number of
emission centers uniformly spread over the surface of the field emission cathode
and making nearly the same contribution to total emission current. A preformed
cathode completely recovers its emissivity [20] after vacuum-chamber depressur-
ization, exposure of the cathode to air and subsequent pumping-out of the chamber
down to the pressure of 10–6 mm Hg.
At the first stage, forming destroys the less durable amorphous filler. As a result,
the durable skeleton of the carbon fiber shows up. Practically, the emitting surface
Fig. 6.4 Emission images of the field emission cathode out of bunches of carbon fibers in process
of molding
of the formed fibers has not changed throughout the experiment, so it has been
impossible to determine the shrinkage rate of the formed-cathode fiber.
Equally, these conclusions pertain to other carbon materials.
Reproducibility of characteristics. Forming is closely connected with one of
the main problems related to practicality of field-emission characteristics in the
system of such cathodes.
Analysis of the carbon-material structure shows that it is almost impossible to
achieve geometrical reproducibility of the form-factor for field emission cathodes of
such materials. So the main efforts should be aimed at achievement of identical
current-voltage characteristics at an uncontrollable form-factor. The characteristics
are achieved by purposeful shift of the current-voltage characteristics to a higher
voltage area up to falling within the tolerance limits. This operation is usually
carried out with supervisory computer control systems by taking a series of
current-voltage characteristics up to the currents exceeding the original value for
forming, followed by reforming of the field emission cathode. On completion, its
current-voltage characteristic practically does not vary in the field of heavy currents
(in Fowler–Nordheim coordinates), but shifts up to falling within the desired tol-
erance in the field of minimum currents (Fig. 6.5).
Parallel switching-on of the field emission cathodes treated in this way causes full
addition of the currents in the achieved multiple-emitter system; i.e., the total current is
equal to the sum of emission currents of each cathode within the limits of fluctuations.
Fig. 6.5 Volt-ampere characteristics of field emission cathodes before molding (a), after
individual molding (b), after shift of characteristics (c)
At that, there is an optimum number of fibers in the bundle. The optimum

number of fibers per bundle is 100–200.
A larger number causes mutual screening of emission centers, and increase in the
number of fibers does not cause increase in emission current with other things being
equal.
Reduction of the number of fibers results in an increase in fluctuation of the
emission current and reduction of reliability of the field emission cathode [21].
Emission current essentially depends on the anode-cathode distance.
In the studied range of distances, current has been established to depend on
distance as 1/R2.
Generally, these dependences are true not only for carbon fibers of different
types [22, 23], but also for other nanostructured materials.
Increase in the total emission current of the field emission cathode can come
from increase in the emitter area. Papers [24, 25] suggest a three-dimensional
woven structure based on polyacrylonitrile carbon fibers impregnated with pyro-
carbon to improve its strength. The cathode assembly is produced so that all the
available filament populations are at an acute angle to the emission direction. In this
case, the emitting surface is a bi-dimensional matrix made of a set of apices, viz. the
fiber ends, projecting by 50–100 µm under the monolithic material of the cathode.
However, these cathodes only work in the explosive-emission mode.
Flat display screens employ mutually insulated emitters. The emitters can be
insulated from each other with special polymers compatible with vacuum production.
This method makes it possible to increase the running current and emission
stability by alternate ageing of some carbon fibers integrated into the dielectric
matrix (Fig. 6.6) [26, 27].
Another way to fabricate the matrices of mutually insulated field emission
cathodes of carbon-fiber bundles is based on their vitrification [28].
This method to fabricate the matrix of a field-emission cathode is to place a carbon
material in the through holes of a glass substrate, the latter being made of block glass
tubes and continuous glass rods. Before the glass tubes are mounted in the block, they
Fig. 6.6 Appearance of the

matrix field emission cathode
from isolated carbon fibers
are pierced with graphite-filament bundles, the filaments being vitrified by heating of
the glass tubes pierced with graphite-filament bundles to the temperature close to the
melting point of glass with subsequent elongation and compression. The carbon
material is arranged in the glass substrate by respective alternation of vitrified
graphite-filament bundles and continuous glass rods, the block being fused and cut
into separate matrices before machining. The matrix is ground and polished until
graphite-filament parts stick up over the matrix surface and the effective area is flat
enough for the matrix to be mounted into the field-emission device.
After the grinding and polishing, the matrix can be etched with, inter alia,
hydrofluoric acid to improve control over the graphite-filament parts sticking up
over the matrix surface.
The graphite-filament bundles passing through the glass tubes can be simultane-
ously fed from one and/or several bobbins into one and/or several glass tubes,
respectively, which not only speeds up fabrication of vitrified graphite-filament bun-
dles, but also provides uniformity of the vitrification mode. The glass selected has the
coefficient of thermal expansion that is close to the coefficient of thermal expansion of
graphite-filament bundles. The purpose is to avoid cracking of the glass and rupture of
graphite filaments during their vitrification, fusing and temperature conditioning.
Carbon-fiber bundles are not particularly rigid, which is why they mix up and/or
break off during attempted insertion into finite-size capillaries. This feature makes it
necessary to pre-spool the bundles, feed them from the spools into the glass tube/billet,
heat up the bottom of the glass tube/billet to a ductile condition, weigh it down and pull
out at a certain rate. Difference in the rates of feeding and pulling generates a vitrified
graphite-filament bundle. The cross-section shape of the resulting item is similar to
that of the initial billet, but has somewhat smaller geometrical dimensions. The
technology suggested makes it possible to vitrify several graphite-filament bundles at
the same time, their number depending, inter alia, on the parameters of the preheater
used in this technology. For this purpose, the graphite-filament bundles are fed from
several spools at the same al respective glass tubes/billets.
The cohesion between the carbon fibers and the glass can be improved, inter alia,
by application of a two-layer glass tube, resulting from insertion of a tube of one
glass type into a tube of another glass type. In this case, the brittle-to-ductile tran-
sition temperature of the internal glass layer is lower than that of the external glass
layer by about 100 °C. Subsequent heating makes the external layer ductile and
shapes the resulting item, while the internal glass layer liquates, which improves the
cohesion between the graphite filaments and the glass after cooling.
The necessary arrangement of graphite filaments in a glass substrate is provided
by formation of a fiber glass block by alternation of vitrified graphite-filament
bundles and continuous-glass rods. The resulting block is fused and cut into matrices.
To provide an equal distance between the cathode and the extraction electrode
when mounting the matrix cathode in the field-emission device, it is necessary to have
a flat effective area of the matrix. For this purpose, the matrix effective area is ground
and polished. To improve connection of the insulated contacts to each of the cathodic
elements, it is also desirable to polish the back surface of the matrix. The macro-
scopical effect of the field near the cathode steps up if the graphite filaments stick up
over the glass surface at constant running voltage. This effect is made by abrasive
machining of the matrix or its selective chemical etching. Polyacrylonitrile
graphite-filament bundles (Fig. 6.7) are coiled on one or several spools 2, one of
which can be seen on Fig. 6.7. Every bundle goes through a respective glass tube
(billet) 3. The most suitable glass type is make C93 because its coefficient of thermal
expansion is close to that of the glass used in device-case assembly. Billet 3 is fixed in
the collet closer of feed mechanism 4, which is ball-screw assembly 5. The ball-screw
Fig. 6.7 Installation for

glassing of bunches carbon
fibers: 1 polyacrylonitrile
carbon fiber; 2 reel with fiber;
3 glass tube (i.e. a blank); 4
feeder; 5 ball-screw; 6
furnace; 7 extension
mechanism
assembly is driven by a worm-gear electric motor (not shown on the drawing). Then
the lower part of the billet is dropped into preheater 6 and heated until the glass
becomes ductile. For the glass used, this temperature is 550 °C. The ductile billet is
weighed down, fed into pulling mechanism 7 and pulled out at a certain rate.
Difference in the rates of feeding and pulling produces an item similar to the initial
billet in the shape of the cross-section, but having somewhat smaller geometrical
dimensions. The pulled-out vitrified graphite-filament bundles are cut into separate
billets of the targeted length, their width and height being determined by the ratio of
the feeding rate to the pulling-out rate of the preheater, and the optimum length being
given by the preheater dimensions. The carbon material is properly arranged in a glass
substrate during formation of the block by pertinent alternation of vitrified
graphite-filament bundles and continuous-glass rods. The block is assembled in a
metal case called “die”. The assembled block is fastened by tying up at the ends with a
thin nickel-chromium thread, followed by removal of the “die”. The laid and fastened
block is put into the preheater and fused under the pressure of 2–3 atmospheres. The
ready block is cut into matrices, their thickness being selected depending on the design
of the vacuum device where they are to be used. The matrices cannot be too thin, i.e.
less than 2 mm, otherwise cohesion between the glass and the graphite filaments can
fail with the result that the graphite filaments can drop out of the matrix. Then the
matrices are ground and polished. For the graphite filaments to stick up over the matrix
surface, the matrix is treated either mechanically or chemically. Difference between
the mechanical rigidity of the graphite filaments and the glass substrate makes the
graphite filaments stick up over the matrix surface as a result of mechanical abrasive
treatment by trying abradants and the rotation rates of the polishing-machine spindle.
The length of the prominent part is finer controlled by chemical etching in a solution
that does not affect the graphite filaments, but dissolves the glass, e.g., hydrofluoric
acid. Duration of the etching is selected on an experimental basis depending on the
glass type and the matrix area.
That makes the method suggested more technological and able to provide any
desired arrangement of the graphite-filament bundles and package density.
Moreover, the matrix produced with this method is quite durable in mechanical
terms and suitable for further treatment if necessary. Use of monolithic graphite
filaments improves stability to ion bombardment of the cathode when working in
technical vacuum and extends the life capability of the device, as well as prevents
premature failure caused by separation of a layer from the emitter. Use of glass to
fabricate the carrier matrix provides electrical isolation of elements of the cathode
and allows to drive each element separately.
6.1.2 Carbon Nanofibers
Recent years saw increasing spread of technologies of direct carbon-fiber growing

on prepared substrates. As a rule, these techniques are somewhat similar to
carbon-nanotube growing methods. The following provides a few examples of this
large area.
Paper [29] suggests field emission cathodes for high-voltage accelerating devices
with voltage of about 70 kV. The carbon fibers of 10 µm in diameter and 0.5 mm in
length were grown with a plate catalyst on a graphite-fabric substrate. The fibers
were grown perpendicularly to the surface consisting of a conducting carbon com-
posite coating the carbon fabric. Paper [30] describes a combination of the Spindt
thin-film technology with carbon fibers. In this case, a fiber-like carbon structure is
vertically sputtered on a molybdenum cone. The growth of carbon nanofibers is
controlled by the emission current. The longer the fiber, the higher is the emission
current. The process comes to the end when the fiber length is a few tens of µm.
However, the technologies of direct carbon nanofiber growing seem to be very
promising. In particular, paper [31] suggests a method of fabrication of
vertically-aligned carbon nanofibers in situ as a field-emission element. The method
is based on the processes recognized in microelectronics, e.g., electronic lithogra-
phy of high-resolution and plasma CVD (PECVD) methods. In brief, the techno-
logical process of fabrication of a field emission cathode structure is shown in
Fig. 6.8. The technological process is detailed in [31].
Nowadays nitrogen-containing carbon nanofibers cause great interest.
Application of carbon-nitrogen materials in field emission cathodes is very
promising because replacement of carbon atoms with those of nitrogen in graphite
layers causes the structural defects facilitating emission of electrons when an electric
field is applied [32, 33]. In addition, a number of theoretical papers shows that the
nitrogen atoms embedded in the carbon nanotube wall graphite layers act as electron
sources, which can also intensify electron emission from such objects [34, 35]. None
the less interesting is the fact that nitrogen-containing carbon nanotubes and nano-
fibers are essentially more inert than pure carbon nanotubes or nanofibers in
chemical terms [36, 37]. That is why the field emission cathodes based on
carbon-nitrogen nanomaterial are less tending to be affected by residual-gas
adsorption-desorption processes, which cannot but improve their functional stability.
The review of research done by now shows that the maximum concentration of
nitrogen in carbon nanofibers does not exceed 5% [35–37]. This concentration of
nitrogen does not produce a high structure-defect density in graphene layers. Nor
does it distort the electron density or essentially increase the emission efficiency.
Paper [38] suggests a new method of synthesis of carbon-nitrogen nanomaterial in a
high-pressure apparatus, as well as a method of fabrication of highly-effective
field-emission cathodes on its basis.
Carbon-nitrogen nanofibers were produced with a high gas pressure apparatus
[32, 39]. The processes of carbon-nitrogen nanofiber synthesis in both nitrogen and
the argon-nitrogen mixture are well developed. Preliminary testing has shown
prospective viability of this material for cold cathodes with the lowest ever initial
electric field intensity [32]. The basic methods of fabrication of field emission
cathodes from this material is screen printing of carbon-nitrogen powder paste,
deposition of carbon-nitrogen nanofibers on the substrate during synthesis in the
working chamber, catalytic growing of carbon-nitrogen material directly on the
substrate.
Fig. 6.8 Stages of the process of manufacturing a field emission cathodic structure out of vertical
carbon fibers: a 4-in. silicon substrate out of n-type Si; b direct electronic lithograph for applying
the plasma CVD catalyst consisting of a 200 Å Ti layer and a 200 Å Ni layer; c drawing an 1 lm
layer of SiO2; d a photolithograph for applying the control electrode; e creation of the image in the
layer of photoresist; f removal of the photoresist coating; g applying of a 1000 Å Cr film; h,
i reactive ion etching in order to remove the layer of exposed resist coating and the SiO2 layer;
j in situ plasma growth of vertically oriented carbon nanofibers, directly in the operating structure
The control samples of 0.5 cm2 were tested in a vacuum chamber at the pressure
of 10−6 mm Hg. At the current density of 10 µA/cm2, the macroscopical threshold
field of turn-on of the field emission cathodes is 1.2–1.3 V/µm. The emission
current density of 1 µA/cm2 is produced by the electric field of 2.3–2.4 V/µm. The
results attained show that carbon-nitrogen nanofibers are very effective emitters
with threshold electric field intensity much less than that of field emission cathodes
based on multiple-wall carbon nanotubes.
The long-term tests were carried out for 80 h in the current stabilization mode.
The current was fixed at 250 µA. At the same time, the dependence of the applied
anode-cathode voltage necessary to attain the preset emission-current value, on time
was measured.
The field emission cathodes produced from carbon-nitrogen nanofibers by screen
printing have a quite high emissivity stability in time. It is worth noting that
long-term stability of field-emission current depends not on nanofiber properties but
on the field emission cathode fabrication method, i.e., the strength of cohesion
between the fiber and the substrate.
6.2 Carbon Nanotubes
6.2.1 Structural and Field-Emission Features
Now a huge number of different modifications of carbon nanostructures has been

known and synthesized, including fullerenes, single-wall and multiple-wall nan-
otubes (SWNTs and MWNTs), graphite filaments, bundles, ropes, rolls, conical
layer nanotubes (CLNT), nanohorns, onions, toroids, graphite nanofibers
(GNF) etc.
Better understanding of structure and properties of carbon nanostructures can
come from consideration of the graphite structure in the first approximation.
Graphite has a layer structure where each layer is made up of a lattice of
hexagons spaced 0.142 nm between the immediate neighbors. The layers follow the
ABAB… pattern (Fig. 6.9), where atoms I are lying just over those in the adjacent
planes, and atoms II are lying over the centers of hexagons in the adjacent areas.
The resultant crystallographic structure is shown in Fig. 6.9, where a and b are unit
vectors in the graphite plane, and c is a unit vector perpendicular to the hexagonal
plane. The distance between the planes in the lattice is 0.337 nm. A unit cell of this
structure is a right-angle prism with a rhombohedral base consisting of 4 carbon
atoms.
As the distance between the layers is larger than that in the hexagons, graphite
can be approximated as a 2D system.
There is a restricted number of schemes intended to construct a nanotube of a
graphene layer. Let us consider points A and A′ in Fig. 6.10. The vector connecting
these points is determined as
6.2 Carbon Nanotubes 307
Fig. 6.9 Crystallographic structure of graphite. The elementary cell is defined by unit vectors a,
b, c
Fig. 6.10 Formation of carbon nanotubes by twisting of the graphene plane and connecting the
points A and A′
ch ¼ na þ mb; ð6:2Þ
where п, m are real numbers, and a, b are unit vectors in the graphene planes.
A carbon nanotube results from folding of the graphene layer and connection of
points A and A′. Then it is only determined by vector ch.
Tube diameter d is equal to
jch j a pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ffi
d¼ ¼ n2 þ nm þ m2 ; ð6:3Þ
p p
pffiffiffiffiffiffiffiffiffiffiffi
where a ¼ 3dcc ; the constant of the lattice, and angle

pffiffiffi m
h ¼ arctan 3 ð6:4Þ
2n þ m
quantitatively determine the chirality of the nanotube. There are two extreme cases
of carbon nanotubes, viz. armchair- and zigzag-type ones. Zigzag is determined by
angle h = 0°, which meets the (n, 0) vector. In this structure, C–C bonds are parallel
to the tube axis. The armchair-type structure is characterized by angle h = 30°,
which meets the (п, п) or (2n, –n) vector. In this group of tubes, the C–C bonds are
perpendicular to the tube axis. Other tube combinations belong to a chiral type with
angles 0°< h < 30°. A single tube is usually termed as a single-wall nanotube,
SWNT. The nanotubes consisting of coaxial graphite cylinders are termed as
multiple-wall nanotubes, MWNT.
We have reviewed the types of ideal 1D-carbon structures. In practice, the
carbon nanomaterials resulting from synthesis have a great variety of structures.
Unlike ideal cases, real carbon structures are characterized by defects of different
types. A failure in the sequence of layers causes packing defects. A large number of
such defects results in a complete disorder of layers about the tube axis. Another
type is lattice-bond defects caused by alien atoms or a change in valence of some
carbon atoms.
Accordingly, there may be compound forms of carbon transitional from nan-
otubes to carbon fibers. They are usually termed as graphene nanofibers [40, 41].
These formations have a structure similar both to nanotubes and carbon fibers, but
the internal walls are not very smooth and usually include graphene layers.
Let us note here that hereinafter the term “carbon nanomaterials” mostly refers to
carbon nanotubes and nanofibers.
Now there are a lot of publications about field-emission properties of different
carbon nanostructures, e.g., review [42]. Almost all the authors mention a dis-
crepancy between the available experimental data about the geometrical shape and
field-emission parameters, viz. emission-current density, threshold field, when
analyzing the simplest model of the emitter with properties characteristic for a metal
apex, e.g., papers [43, 44]. Analyzing the available data based on the regularities for
metal apices, the authors of paper [45] received the geometrical amplification
coefficient exceeding 104, which is only possible if the emitting area is comparable
with an atom in size. According to these researchers, it is only possible if a strong

electric field caused formation of a carbon-atom chain at the end of the carbon
nanotube. Attempts to attribute the above-mentioned mismatch to more careful
account of distribution of the field and the charge density for nanometer-scale
emitters [46] have not so far given any definitive result.
Another possible explanation for this phenomenon can be significant reduction
of the work function on the surface of the carbon nanotubes in comparison with the
work function characteristic for graphite and graphite-like materials [47]. However,
physical interpretation of this possibility seems to be insufficient. So the attempts to
interpret the available experimental data only by modification of its parameters are
obviously unsuccessful within the framework of the Fowler–Nordheim theory.
The complete theory should allow for a possibility of electronic and hole con-
ductivity with account of their effective masses and dispersion dependences e(k).
Moreover, it is necessary to take account of surface states and singularities in
density of nanotube states as reduced-dimensionality systems. Existence of the
latter is indicated by the difference between the spectrum of electrons from the
emitted carbon nanotubes and the field-emission spectrum in the metal [48]. It is
also necessary to recalculate the electron tunneling probability with account of a
possibility of penetration of the external electric field into the nanotube.
By now, this complex of problems is a very difficult theoretical challenge
aggravated by insufficiency of established facts about the electronic structure on the
surface of different carbon nanotubes.
Some progress was achieved in an attempt to model field emission from nan-
otubes by different methods of direct solution of the Schrödinger equation for all
atoms of the carbon nanotube in an external field [49, 50]. What essentially com-
plicates the situation is that the number of atoms in the nanotubes of practical
interest considerably exceeds the up-to-date computing capacities, which necessi-
tates combined approaches.
In addition, recent days have brought some other theoretical interpretations
suggesting a cardinal difference in the emission mechanisms between mostly
nanoscale carbon materials and metals, e.g., papers [51, 52] analyze the field
emission mechanism on the basis of dynamic resonance. According to the authors,
the effect can be attributed to significant modification of the surface potential by the
external field, which is specific for graphite as a material having a low concentration
of high-mobility electrons. The mechanism considered predicts increase in the
current from graphite-like emitters and a current density-external field dependence
different from Fowler–Nordheim one.
Equally interesting is another model of field emission from carbon
nanomate-rials with account of the resonant tunneling phenomenon [53–55].
According to this model, the emitting surface of graphite-like nanostructures, viz.
carbon nanotubes and plate-like nanocrystallites, contains carbon-atom clusters
with diamond-like bonds. Atomic bond transformation modifies electronic prop-
erties on the carbon nanostructure surface. The peripheral carbon atoms forming the
surface layer differ from other atoms in their location in the bending of atomic
planes (Fig. 6.11).
Fig. 6.11 Scheme image of the curved graphene planes arising on edges of nanocrystallites of
graphite
Such structure of vertexes of graphite crystallites and nanotubes seems most

probable. Thermodynamically, when graphite crystallites grow, formation of a
dangling-bond atomic layer on the edge seems less favorable in comparison with
that of bow-shaped structures closing the edges of parallel graphene planes [56]. In
addition, the available data show that formation of bent atomic layers on the edges
of graphite-like structures is much more preferable in comparison with other pos-
sibilities. It is true for a wide class of carbon graphite-like materials ranging from
graphite powders [57, 58] to the surface of massive graphite samples [59], as well
as for the graphite microcrystallites forming carbon fibers [60, 61].
Quantum chemistry methods let us know that such bend of the basic graphite
plane can result in formation of an electronic configuration different from graphite
in its nature. A computation made in [56] by the molecular-orbital method shows
that a bend of atomic graphite planes forms double chains of atoms having an sp3
electronic configuration similar to that of diamond. So it can be assumed that the
electronic energy spectrum of the atoms located on the periphery of graphite-like
carbon layers has a band gap as wide as 4 eV determined by the energy of r–r
bonds. This layer is very thin, amounting to a few lattice constants of some
ångström units.
The small sizes of the layer cause the surface to develop not a band but an
electron-level system. The emitter surface has atoms with a modified electronic
structure, which makes the electrons emitted in vacuum tunnel through two
potential barriers one after another as opposed to only one stipulated by the tra-
ditional Fowler–Nordheim theory. The potential barriers arise at the conductor–
dielectric and dielectric-vacuum boundaries, respectively. It is obvious that a certain
ratio of parameters of these potential barriers and the same applied voltage can
cause the electron tunneling probability to be higher for double barrier than for a
single one. Similar double potential barriers are considered to describe field
emission from cathodes with surface-adsorbed molecules [62], as well as the
cathodes consisting of thin layers of diverse materials [63]. It is specific for carbon
emitters to develop the double potential barrier on their surface from two different
carbon forms with graphite-like and diamond-like electronic properties.
The assumption of resonant tunneling through the surface states of carbon field
emitters is directly proved by the profile of the energy spectrum of the emitted
electrons. In particular, paper [64] shows an additional maximum in the power
distribution of electrons with the energy lower than the basic maximum by 0.45–
0.5 eV when studying field emission from carbon nanocrystals and nanotubes. The
authors establish that the additional maximum is generated by the conditions of the
experiment; in particular, further increase in the emission current eliminates the
additional maximum. The conclusion is that there are surface electronic states at the
Fermi level.
Previously, a similar field-emission model allowing for the effect of the surface
layer of a dielectric or a wide-gap semiconductor was also considered in [63, 65].
As a rule, these models have the surface layer of a dielectric or a wide-gap semi-
conductor that is much thicker than sp2–sp3 transition considered in [53, 54].
However, effective resonant tunnelling necessitates a very thin dielectric layer
approximately equal to the barrier width at the boundary with vacuum.
For metal or semiconductor emitters, production of such thin layers from tra-
ditional materials, viz. W, Mo, Si, etc., is a rather difficult problem in terms of
technology; while for carbon nanostructure based films such a layer is a conse-
quence of bending of atomic layers, which makes them a unique material for
field-emission cathodes.
6.2.2 Some Carbon Nanotube Growing Methods
Nowadays, carbon nanotubes, nanostructures and their modifications can be grown

by various technologies and their modifications. They are thoroughly analyses in
numerous articles and monographs, viz. [66–69]. This section considers just a few
examples illustrating the technological methods and emissivity of these structures.
The basic way to synthesize nanotubes is by chemical vapor deposition, CVD,
with involvement of different catalysts. The main catalysts are nickel, chromium
and iron. As a rule, the current temperature of the process on a silicon substrate is
*700 °C [70]. However, emissivity of the nanotubes grown appears [71] to depend
on catalyst thickness. Paper [72] describes application of Ni spots as thick as 50, 70
and 100 nm to establish the best results in terms of field-emission current, which
turned out to come from nickel as thick as 70 nm at the current density of
6.5 A/cm2. However, it can be attributed to uneven application of nickel to the
substrate.
For many practical applications, especially for flat display screens, the most
accessible substrate for field emission cathodes is glass. Therefore, a lot of efforts
are made to reduce the deposition temperature of carbon nanotubes to 600 °C,
which is compatible with usual cheap sodium glasses. Under such conditions,
nanotubes grow due to application of new-type catalysts or introduction of addi-
tional layers. In particular, paper [73] refers to Fe, Zr and N catalysts of 80, 10 and
10%, respectively. The film is 2.5 nm thick. Nanotubes are deposited at 550 °C in
two-region furnaces by the thermal CVD method. The resulting nanotubes are
about 20 nm in diameter and 1 µm in height. In this case, the threshold electric field
is 3 V/µm and the current density is 0.3 A/cm2 at the field intensity of 4 V/µm.
The author of papers [74–76] managed to achieve further reduction in the
deposition temperature. He introduced an additional layer of CaF2 between the Ni
catalyst and the substrate, which reduced the process temperature to 500 °C, and
produced the field emission cathode samples differing in distribution of the catalyst
over the substrate surface. The experiments show that the cathodes spotted with a
catalyst of 50 µm in diameter have the best field-emission characteristics.
The latest result is of not only theoretical but also great practical importance in
terms of fabrication of field emission cathodes for flat display screens since they
need a large array of insulated field emission cathodes. Some authors suggest
fabrication of matrix field emission cathodes from carbon nanotubes sized
50 50 µm2 [77] and 30 30 µm2 [78]. Paper [77] suggests fabricating field
emission cathodes of 10 10 pixels consisting of multiple-wall nanotubes of
10 102 nm in diameter and a few µm in length at 550 °C, the nanotubes being
interwoven to form a “felt” structure. Though the threshold electric field is 1.8 V/
µm, the total current from 400 emitters is 73 µA at the field of 10 V/µm.
The vertical well-aligned nanotubes are grown on a Si substrate at 800 °C on a Fe
catalyst [78] (Fig. 6.12). This vertical position is caused by Van der Waals interac-
tion between the neighboring nanotubes. The nanotube density is determined by the
density of Fe nanoparticles on the silicon substrate. The current density is indicated to
be 1 MA/cm2 at the field of 2 V/µm and as low as 80 mA/cm2 at 3 V/µm.
Fig. 6.12 The photo in SEM of vertically focused nanotubes on the catalyst from Fe film on Si to
a silicon substrate: a 30 30 lm blocks of nanotubes at distance 125 mm from each other; b the
increased image of the block on which the orientation of tubes is well visible
A disadvantage of growing of direct carbon nanotube on the sputtered-catalyst

particles is insufficient adhesion of nanotubes to the substrate. When a field
emission cathode functions, this circumstance results in tearing-off of the nanotubes
from the substrate under the influence of the field applied. A new approach to
solution of this problem is suggested in [79], and Fig. 6.13 shows a flow-chart of
fabrication of a field emission cathode from carbon nanotubes. The fabrication
involves the processes standard for semiconductor electronics. A silicon substrate is
groove-cut as deep as 3–10 µm, the grooves being intended to grow carbon
Fig. 6.13 Flowchart of manufacturing process of field emitters out of carbon nanotubes
nanotubes. In the groove array, a pixel of 300 300 µm2 square can have different
sizes, namely 5 5, 10 10, 20 20, 30 30 or 40 40 µm2.
The prepared substrate is thermally oxidized to make an insulating layer between
the substrate and the would-be nanotubes (Fig. 6.13a). Then a photoresist and Fe of
25–50 nm in thickness is deposited (Fig. 6.13b). Further on, the photoresist is
removed together with the iron, while the iron film remains where the nanotubes are
grown up (Fig. 6.13c). What follows is carbon-nanotube growing by the thermal
CVD method at 900 °C (Fig. 6.13d). The process is detailed in [80].
The silicon substrate with grown nanotubes is coated with a thin layer of a
composite material based on glass (Fig. 6.13e). Having dried up, it provides a
strong adhesion of the nanotubes to each other and to the substrate. What follows is
mechanical polishing with diamond paste. At the first stage, the structure is graded
(Fig. 6.13f). At the second stage, properly selected diamond powder removes the
excess of glass composite material to release the nanotubes; in addition, it removes
the tops of the closed nanotubes, thereby making them open (Fig. 6.13g).
The emission tests carried out in the diode mode at the pressure of 10−7 mm Hg
show high stability of the emission current for 2 h and a threshold electric field of
*1.7 V/µm.
Another problem of fabrication of field emission cathodes with planar tech-
nologies is current leakage through the electron-electron gap of the field emission
cathode—the extraction electrode. The reason thereof is shortness of the gap. The
gap can be lengthened by sputtering of an additional internal protective layer [81,
82]. This measure alone reduces the leakage current more than 7 times at the
extraction voltage of 100 V.
The temperature of carbon nanotube growth decreases due to reduction of the
catalyst particles down to <10 nm [83]. It is also necessary to remove amorphous
carbon that accompanies deposition of carbon nanotubes at low temperatures.
Vertical position of carbon nanotubes is improved at supply of extraction voltage
between the sample surface and the plasma generator case during the deposition.
Finally, removal of defective nanotubes with gas-mixture plasma gives the most
perfect structure. That is the way to fabricate a field emission cathode with a 1 V/
µm threshold field at *500 °C. A flat display screen with size of 5.4″ produced
with this technology shows a very good uniformity of emission over the screen
area, viz. *3.3%.
Field emission cathodes made out of carbon nanotubes with improved substrate
adhesion produced via methods of powder metallurgy were suggested in [84]. Such
cathodes are made of a mixture of multiple-wall carbon nanotubes and powdered
metal (zinc in case of the paper). Nanotubes of 1–10 µm in length and 50–60 nm in
diameter are made by arc discharge. At first, the zinc powder and the nanotubes are
mixed in a mortar and the resulting mixture is pressed at 0.5 MPa. Fusing takes
place at 420 °C in a vacuum furnace. The final size of the sample is
ø 5 mm 2 mm. Field-emission properties of the surface of the sample fused are
improved by polishing followed by etching in an acid to remove a thin surface zinc
layer. So the ends of carbon nanotubes develop the microelevations effectively
Fig. 6.14 Comparison of

field emission stability of field
emission cathodes
manufactured by methods of
powder metallurgy
(rhombuses) and ones grown
by the CVD method (squares)
aggravating the electric field, while the main part of the nanotubes is in zinc, which
essentially improves their cohesion with the substrate.
It is noted that stability of field-emission current of a field emission cathode
made by powder metallurgy is higher than that of the field emission cathodes grown
by the CVD method (Fig. 6.14).
Further development of this approach leads to application of organotitanium or
organometallic compounds [85], e.g., Ti (O–n–C3H7)4, Al(IC4H9)3, Si(O–n–C4H9)4
and other metals of periodic groups 3, 4, 5 and 6. First comes the paste consisting of
40–95 wt% of an organometallic compound, the rest being multiple-wall carbon
nanotubes. The paste is applied on substrates by the printing method and fused at
the temperature ranging from 250 to 600 °C. The temperature and the treatment
conditions are determined by a necessity of all-burn of the organics, on the one
hand, and the minimum oxidation of nanotubes, on the other hand. During the
fusing, the organic carrier is completely removed with evaporation, and the layer
surface microcracks by 0.1–100 µm forming grooves where the free ends of nan-
otubes are put. The rest of them are securely fixed in the metal layer. A smaller
amount of the organic compound results in worse microcracks, and the content of
more than 95% essentially decreases the emission centers in number.
It has long been known [6, 20] that stable operation of carbon-based
field-emission cathodes needs current forming or aging. Right selection of tech-
niques and modes largely determines the practicality of carbon-based field emission
cathodes in general and carbon nanotubes in particular.
Recent technologies include pre-treatment of the surface of field emission
cathodes made of carbon nanotubes.
Classical ion bombardment is supplemented with laser exposure [86] and glow
discharge [87]. In the former case, carbon nanotubes are applied by a standard
printing, dried at 400 °C for 2 h and are irradiated with a KrF laser of 248 nm in
wavelength in an open air, energy density amounting to 2–4 mV/cm2 or 40–
80 mJ/cm2. The laser bundle of 9 9 mm2 in area is homogeneous. The homo-
geneity of the bundle turns out to result in considerable homogeneity of the
field-emission image. Addition of a glass filler to the original printing composition
and subsequent laser treatment reduce the threshold field to 0.34 V/µm. Paper [87]
suggests an inexpensive method of ageing of field emission cathodes made of
carbon nanotubes in glow discharge in a hydrogen medium. Carbon nanotubes are
grown on a stainless-steel plate by the thermal CVD method with involvement of a
catalyst. The system intended for aging of the resulting field emission cathode and
measuring of its field-emission characteristics was located in a vacuum chamber.
The pilot installation diagram is shown in Fig. 6.15. The metal anode and the field
emission cathode are spaced 100 µm, while the anode and the glow-discharge
cathode are *3 cm apart. For the hydrogen plasma to cover the whole of the field
Fig. 6.15 The scheme of experimental setup for training the field emission cathode out of carbon
nanotubes in abnormal glow-discharge
emission cathode surface, the discharge conditions are similar. The best results are
attained at the hydrogen pressure of less than 1 mm Hg. Under these conditions, the
running voltage of the field emission cathode is seen to go down.
6.3 Non-oriented Structures
Active search of the ways to create new carbon materials is still going on. Most of
them are based on the technologies already developed for nanotubes. The following
summarises the technologies and structures based on them.
Diamond-like carbon is a form of amorphous carbon (a-C) or hydrogenated
amorphous carbon (a-C–H) with a considerable number of sp3-bonds [88]. The
number of sp3-bonds and hydrogen in any diamond-like carbon depends on the
deposition process involved and can be represented by a three-phase diagram
(Fig. 6.16). The most popular method for synthesis of a-C:H is plasma CVD.
However, this technology yields a lot of hydrogen.
The larger number of sp3-bonds in a-C is plasma CVD. However, this tech-
nology yields a lot of hydrogen.
The larger number of sp3-bonds in a-C, which is designated as tetrahedral
amorphous carbon (ta-C), results from use of a highly ionized ion beam of the
respective energy.
Electronic structure a-C depends on the content of sp3-states. The sp3-states
develop only strong bonds, while s sp2-states develop both a bonds and looser p
bonds [89]. p bonds lie close to the Fermi level, thereby determining the forbidden
bandwidth. It has been experimentally found that impurity-doping incorporation of
ta-C and ta-C:H with nitrogen in the course of sputtering does not reduce the
forbidden bandwidth.
Field emission from diamond-like structures is mostly limited by intrinsic
conductivity of the structure. Diamond-like structures are polycrystals, which is
why their grain boundaries function as conducting tracks.
Fig. 6.16 Triple phase-rule

diagram of carbon-hydrogen
system showing types of
diamond-like carbon
Fig. 6.17 Scheme of an element of an address-controlled matrix of field emission cathodes on the
basis of diamond-like structures
Naturally, research of field emission of amorphous diamond-like structures has

led to attempts to create cathode-modulation assemblies on their basis for devices of
different purpose.
The diagram of such a cathode-modulation assembly is shown in Fig. 6.17 [90].
An experimental matrix of 25 field emission cathodes of 5 µm in diameter is made.
First deposited is an oxide film as thick as 500 nm and an aluminum film as thick as
200 nm. Then a standard technology is used to apply a photoresist and make holes,
where tetrahedral amorphous carbon (ta-C) is deposited. The deposition was caused
by a pulsing ArF excimer laser on a continuously revolving target of pyrolytic
graphite. The short-wave irradiation of the ArF laser with size 193 nm has an
obvious advantage over long-wave lasers for two reason. Firstly, the deposited flow
mostly consists of monoatomic carbon clusters (c, c+, c2+) even if the flows are
relatively low. Secondly, the kinetic energy of ions in such lasers is higher than that
of long-wave lasers at any flows or sizes of the focused spot. It is experimentally
proved that the maximum quantity of sp3-fixed carbon in sputtered films is formed
at the kinetic energy of ions c+ 90 eV.
Nanostructural amorphous carbon with the sp2:sp3 bond ratio of 40:60 produces
the initial threshold electric field of 0.6 V/µm [91]. A carbon film coats a thin
titanium film applied by magnetron sputtering to the n-Si (111) treated surface. That
is done with the CVD system with microwave plasma in the H2 and CH4 gas
mixture. The resulting current density is 2.5 mA/cm2 at the electric field density of
3.7 V/µm.
Creation of a triode structure with very small control currents is reviewed in [92].
Figure 6.18 shows a diagram of this structure, which is produced in the fol-
lowing way. First, a considerably alloyed n-silicon substrate is thermally oxidized
to produce an oxide layer as thick as 1 µm. Then the oxidized layer is sputtered
with a chromium layer as thick as 300 nm. Then standard lithographic methods
provide holes in the chromium and silicon dioxide film. This process takes place in
two steps. After that, the reactive ion-etching method produces hollows in the
silicon. Hereupon comes the buffer chromium layer of 30 nm followed by the NiFe
6.3 Non-oriented Structures 319
Fig. 6.18 The scheme of

triode structure with the
immersed cathode
catalyst layer. Then the resist layer is removed and a nanostructured carbon layer of
a methane-hydrogen mixture is applied at the temperature of 680 °C by the thermal
CVD method [93, 94]. It is worth noting that the carbon layer is synthesized only in
the central part of the hollow because the catalyst grains are applied through the
hole of *10 µm. Small control current can result from adjustment of the sticking
up of the modulator electrode over the silicon-dioxide film. Both a simulation and
an experiment show that practically zero current of the modulator electrode can be
achieved if the chromium film is at the same level with SiO2.
At present, the designers of field emission cathodes pay much attention to
amorphous structures based on nanotubes and similar structures. Different authors
give them have different names, viz. felt, mesh, velveteen, velvet etc. However their
essence is the same: chaotically interwoven carbon nanofibers without any common
orientation. Let us give some examples of implementation of these structures
below.
Paper [95] gives an example of such a structure, which the authors termed as
carbon felt. It appearance is given in Fig. 6.19. The total thickness of the layer is
*70 µm. The material, single-wall carbon nanotubes, is 50–70% pure. The
impurities are mostly leftover nickel catalyst in form of nanoscale particles. The
average diameter of the carbon nanotubes in the filaments is *20 nm, which
accounts for a small radius of the field-emission centers.
Fig. 6.19 SEM microphotograph of chaotically intertwined nanotubes (so-called “carbon felt”)
An appreciable field-emission current is observed as early as at the electric field

intensity of 2–5 V/µm.
However, such long operating time as 100 h shows degradation of the emission
current by about 8–10 times. The main reasons for this decrease are heating of the
emitting centers, ion bombardment and internal properties of carbon nanotubes.
A similar result is also attained in [96]. To improve the results, paper [95]
suggests current ageing by residual-gas ions and paper [96] suggests UV-laser
irradiation. A simple method to fabricate amorphous films from carbon multilayer
nanostructures is suggested in paper [97]. The structures in form of nanoscale nets
are planted on an n-type silicon substrate by the electrophoretic method based on
liquid methanol. In this process, the cathode is represented by a graphite plate.
The process runs at the electrolytic current density of about 4–15 mA/cm2, the
voltage of 1000–3000 V and the working temperature of 60 °C.
Raster electron microscopy has established that the nanonet filament diameter is
about 200 nm with a very good purity of the material. The threshold electric field is
1.13 and 3.4 V/µm.
Many types of quite exotic formations, such as nanotubes formed as the lower
part of mangle trees, have been featured in publications. In this process, the
emission centers move away from each other, which decreases screening [98].
It is worth noting that this section does not specify the achievable density of the
emission current because it is often measured with thin wire anodes. This method is
not correct because such an anode collects electrons from a much larger area than
the anode itself.
Another area of amorphous/non-oriented structures is different-type carbon
films; e.g., paper [99] features perpendicularly-oriented nanofilms consisting of 1–5
freely standing graphene layers. They are successfully synthesized on different
substrates, e.g., Fe, Au, SiO2, Al2O3, Ti. Such structures essentially differ from
nanotubes or other similar structures. Firstly, they do not need a catalyst to grow.
Secondly, the electrode pattern can be made after deposition by the standard
photolithographic method. Such nanofilms have lower threshold electric field as
compared with nanotubes. It is 1.1 V/µm at the current of 10 nA. The field
emission cathodes made of vertical nanofilms tend to spontaneously reduce the
threshold field and increase stability of the field-emission current. Pilot experiments
show that the nanofilms as thick as 6 µm develop the current density of
>1 mA/mm2 at the direct current.
Long microblades as high as *6 µm develop on the surface of glass-carbon
field emission cathodes. In addition, the running voltage dips down by several
orders of magnitude, while the current density [100] goes up.
The carbon-film end face in a planar structure can also be used as a source of
field emission cathodes [101]. In this case, the a-C, layer applied by the CVD
method is as thick as 20–30 nm. In particular, this structure can be used to fabricate
light sources and displays.
6.4 Carbon Foils 321
6.4 Carbon Foils
This section understands foils as carbon materials of small thickness and large
length and width, which are not applied on a substrate.
The first material to be used as field emission cathodes were thin pyrolytic-
graphite plates.
Pilot experiments show that pyrographite is quite promising as a field emitter.
The emission centers of field emission cathodes made of pyrolytic graphite are thin
strong microplates forming the pyrolytic graphite structure (Fig. 6.20). So
field-emission properties of foil made of pyrolytic graphite are close to those of
polyacrylonitrile carbon fibers. Further research is aimed at determining the optimum
treatment temperature and thickness of pyrographite plate. To this end, the research
focuses on the emission characteristics of the field emission cathodes made of pyro-
graphite plates of 30 lm in thickness processed at 1400, 2000, 2500 °C, as well as the
plates of 5 and 150 lm in thickness processed at 2000 and 1700 °C, respectively.
The measurements are preceded by a standard forming procedure in the linear
mode up to the maximum possible current of 6–8 mA at the length of the working
edge of 2 mm, though the field emission cathodes made of such materials show
some elements of a self-forming process.
The current-voltage characteristics of these field emission cathodes taken every
10 min during forming determine the dependence of the current at which the
effective cathode area begins restructuring on the heat-treatment temperature of the
pyrolytic graphite as thick as 30 µm and on the thickness of the sample plates at
Ttrtm = 2000 °C (Fig. 6.21).
The diagrams show that the limit current of these field emission cathodes goes
up with increase in the material heat-treatment temperature. The most likely reason
Fig. 6.20 Appearance of a

working surface of a 150 lm
pyrographite field emission
cathode
Fig. 6.21 Dependence of current at which pyrographite field emission cathode surface
reorganization begins on the plate thickness at T = 2000 °C (a) and on temperature at plate
thickness of 30 lm (b)
thereof is that the material tenacity improves with growth of the treatment tem-
perature. A similar dependence also takes place for carbon-based field emission
cathodes [14]. The growth in the sample-plate thickness increases the total number
of emitting microelevations on the cathode surface, thereby increasing the current
value at which the effective area restructures.
The field-emission current instability level is constant for all samples at small
currents, viz. down to 1–10 µA, and goes down with growth of the current at its
higher values, viz. more the 1 mA. At the currents of less than 1 µA, the pyro-
graphite field emission cathodes of 30 lm in thickness showed the lowest insta-
bility value at Ttrtm = 2000 °C (r = 9%), and the highest one at Ttrtm = 2500 °C
(r = 27%). At the currents of 3–6 lA, the instability of all samples is about 1–2%.
Analysis of field-emission current fluctuations shows that the largest number of
the emitting centers with other things being equal belongs to the pyrographite field
emission cathode of 30 lm in thickness and the treatment temperature of 2000 °C.
As the maximum current drawing, long service life and uniformity of field emission
over the cathode surface directly depend on the number of the emitting centers,
pyrographite with such parameters is most preferable for electronic devices.
The main weakness of pyrolytic carbon field emission cathodes is difficulty of
their fixation and fabrication of cathode assemblies or cathode-modulation assem-
blies with reproducible field-emission characteristics, respectively. The reason is
impossibility to achieve field emission from a foil flat surface.
This conclusion refers not only to pyrolytic graphite, but also to any carbon-
based foil.
The following are the data on a successful attempt to fabricate a planar field
emission cathode of graphite foil. In particular, thermally-expanded graphite
(TEG) can be used. The central idea of the method is to form field-emission centers
on the graphite-foil surface by pulsing laser light [102, 103].
The pulsing laser light forms a crater on the TEG foil surface. Figure 6.22 shows
a REM-image of a typical crater. The crater falls into two areas: the central area
Fig. 6.22 SEM image of a

crater on the TEG foil surface
(I) and the peripheral area (II). The central area is a hollow in the foil developed
during intensive evaporation of graphite caused by laser. The peripheral area is
formed by bulged graphite foil layers. The foil is bulged by the forming vapors
penetrating into the pores. This process does not crumple up the initial
microstructure of thermally-expanded graphite but even additionally unfolds it.
The REM-images of the craters show that the crater-surface microrelief has a
characteristic size of 20–60 nm. In the relief structure, the TEG-particles are uni-
formly distributed over the crater surface and the space between the neighboring
particles exceeds their size by an order of magnitude. Moreover, the TEG-particles
are seen to form the crater by orienting themselves at right angle to the crater
surface. In this case, most probable is field emission from the end faces of these
particles; i.e., the emission can proceed from the graphite-layer end faces, i.e. layer
packages positioned at right angle to crystallographic axis c. The bibliographic-data
review made in Chap. 1 mentions that this structure is optimal for a field-emission
cathode. It is also worth noting that the characteristic sizes of the resulting irreg-
ularities are close to the nanotube size. Hence, the crater surface is expected to
produce intensive field emission comparable with that of the carbon-nanotube
cathode surface, which is confirmed by the data given below.
Such emitters have to meet the following requirements:
• the crater has to be formed without a through hole in the TEG-foil;
• the crater has to be as small as possible;
• the crater has to be axially symmetrical;
• the near-crater area has to be free from TEG impurities breaking the crater
symmetry.
The diameter of the resulting crater is determined by studying of a simple model

of crater formation on the TEG-foil surface when a laser beam causes substance
evaporation when getting on the surface.
If the entire pulse energy is spent on evaporation of some substance, the simplest
case is that this area is hemispheric. This situation can be quite easily described with
an energy-balance equation.
It looks as follows
1 4 d3 1 4 d3
E ¼ mL0 ¼ p L0 ¼ p Cq; ð6:5Þ
23 8 23 8
where v is the hemisphere volume, L0 stands for the energy necessary to sublime a
substance volume unit at 0 °C, d is the hemisphere diameter, q is the substance
density, C stands for the energy necessary to sublime a substance volume unit at
0 °C, E is the laser-pulse energy. Then the diameter of the evaporated area can be
expressed by the following formula, viz.
sffiffiffiffiffiffiffiffiffiffiffiffi
3E
d¼2 : ð6:6Þ
3
2pCq
Formula (6.6) is derived empirically; however, it adequately shows dependence

of the diameter on the laser-pulse energy and the substance density. Papers [104,
105] carefully analyse the issues related to making holes by laser light. The phe-
nomenological model generally developed for destruction of opaque materials by
focused laser light with a moderate power density, viz. 106–107 W/cm2, which is
typical for our case, provides the ratios featuring hole formation. In particular, when
the radiation is focused on the material surface and the pulse comes to the end, the
hole diameter is expressed by the following ratio, viz.
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

3 d0 3 3EtgY 3 d0 3 3EtgY
d¼2 þ ¼2 þ ; ð6:7Þ
2 pL0 2 pCq
where d0 is the diameter of the focused laser bundle, Y is the laser-bundle con-
vergence half-angle, C standing for the specific energy necessary to sublime a
substance at 0 °C. Our further calculations are based on Formula (6.7).
Adequate determination of the external size of the crater necessitates consider-
ation of TEG-foil bulging. The crater peripheral-area size depends on how deeply
the graphite vapours generated by laser light have penetrated into the graphite-foil
pores. A characteristic giving an idea of the foil porosity is the ratio of the graphite
density to that of the TEG foil. Accordingly, the linear coefficient of vapor pene-
tration into the pores is proportional to the cube root of this ratio multiplied by
empirical coefficient k.
Proceeding from the assumptions made, the external diameter of the crater,
which depends on the power of laser pulse E and TEG-foil density q, is expressed
by the following formula, viz.
rffiffiffiffiffi 3
q 3 d0 3EtgY
d ¼ 2k 3 0 þ ; ð6:8Þ
q 2 pCq
where k is the empirical coefficient of vapor penetration into pores, q0 is the

graphite density, d0 is the diameter of the focused laser bundle, Y is the laser-bundle
convergence half-angle, C is the specific energy necessary to sublime a substance at
0 °C, the graphite table-value being 40.24 MJ/kg [106].
Coefficient k is found by experiment so that the model results best coincided
with the experimental data. The optimum coefficient found is k = 1.644, whereby
the difference between the model and the experiment does not exceed 5% over the
entire data set. So the conclusion is that Formula (6.8) adequately expresses the
crater forming on the foil surface affected by pulsing laser light of one-millisecond
duration and the power density of 106–107 W/cm2. In other cases, the crater sizes
are determined by different processes.
So the experiments conducted have identified the following optimum parame-
ters, viz. the TEG-foil density of 1.78 g/cm3; the foil thickness of 0.2 mm, selected
for convenience in the technological process; the lens-focus distance of 5 cm; the
laser pulse energy of 0.13–0.35 J. The power less than 0.13 J produces an insuf-
ficient crater size and underdeveloped emitting edges, while the laser-pulse power
of more than 0.35 J produces a too big crater since the graphite material is
excessively carried out beyond the crater bounds. As a result, the TEG foil is
damaged so that graphite flakes open by the angle of more than 90º, thereby
increasing the linear dimensions of the emission center without increase in emissive
capacity.
By now, we have managed to form a crater of about 150 µm in diameter. It has
also become clear that it is not the final limit of reduction of the crater size. It is also
possible to go on reducing the emitter size. However, that needs additional opti-
mization of crater-fabrication parameters, in particular, those of the laser bundle.
The field-emission research involves the samples consisting of one crater and a
field emission cathode consisting of 54 craters spaced 2.5 mm apart. The mea-
surements are made in the diode structure in the continuous and pulse mode. The
characteristics are exemplified in Fig. 6.23.
For the best TEG cathode samples with the cathode being 450 µm away from
the anode the running voltage is 900 V. In this case, the electric field intensity is
2 V/µm. According to the best bibliographical data, it is the same as in the cases of
carbon nanotubes.
The long-term tests are carried out in a sealed-off device in the diode mode.
A 9 11-crater array is produced on the foil surface. The craters are spaced
1.5 mm apart.
Fig. 6.23 Volt-ampere

characteristics of TEG
cathodes in constant mode:
a one TEG crater (with
distance between the anode
and the cathode of 20 lm);
b the cathode consisting of an
array of craters (with distance
of 200 lm)
The total test period is about 800 h. Within the first 70 h, the current is 200 µA,
the current density drawn from the surface of each crater being 0.4 mA/cm2. At the
expiration of the 70 h, the current is 100 µA, the current density being 0.2 mA/cm2.
Figure 6.24 shows the diagrams of current and voltage variation in the device.
Analysis of current-voltage characteristics shows that degradation of emissivity
is caused by reduction of the total number of emission centers on the cathode
surface. The principal cause of the degradation is bombardment of the cathode with
residual-gas ions. The more intensive the ion bombardment, the heavier is the
current drawn. It is also confirmed by the data of the long-term operating time.
Figure 6.24 shows that at the first stage of the operating time, when the cathode
current is 200 µA, the running voltage of the cathode doubles within 70 h, and
when the drawn current is reduced down to 100 µA, the running voltage doubles
within about 700 h.
Research of field-emission properties of flat field emission cathodes based on
graphite foils discovered a ring effect.
Fig. 6.24 Plots of change of

current and voltage on the
device during long-term tests
The rings are a diffusely flared area of the luminophore surrounding the main
field-emission image. The ring diameter often exceeds the size of the main image.
The ring brightness increases with increase in the emission current.
The reason thereof is as follows. The electrons get to the anode and knock
secondary electrons out of its surface. The electric field makes the secondary
electrons come back to the anode and form a ring.
Let us consider a flat-diode model with the anode and cathode spaced L apart
(Fig. 6.25). At the anode voltage of U, the field in the anode-cathode gap is
E = U/L. Considering the motion of the secondary electrons near the anode, we
neglect existence of emission centers on the cathode. They cause local enhancement
of the field near the cathode, but do not essentially affect the electron behaviour near
the anode. Let us consider the motion of the secondary electron having the initial
energy of eS and escaping from the anode surface at the angle of a.
Simple calculations of the electron path in a homogeneous constant electric field
make it possible to find the distance from the secondary-emission point in the anode
plane as
Fig. 6.25 Model of ring forming on a flat diode

eS
r¼2 L sin 2a; ð6:9Þ
eU
in this case, the maximum distance from the anode is

eS
h¼2 Lðsin aÞ2 : ð6:10Þ
eU
As a result, the external radius of the ring formed by electrons with the initial
energy of eS is
eS
rmax ¼ 2L : ð6:11Þ
eU
Assuming that the primary role is played by elastically-reflected electrons, viz.

they produce a ring of the maximum size, we can believe that eS/eU 1. Hence,
the maximum ring diameter is Dmax 4L.
So the ring diameter is determined only by the distance between the cathode and
the anode and does not depend on the cathode current or the anode-cathode voltage.
As the elastically-reflected electrons are only a part of all the secondary electrons
knocked out of the anode and taking into account that the electrons take off from the
surface under different angles, the entire area from the external ring boundary to the
emitter emission image is flared.
Experimental research has completely confirmed these conclusions [107, 108].
Depending on the conditions of application of the field emission cathode, this
effect can be both positive and negative. On the one hand, the ring size limits
resolution in case of a flat display. On the other hand, this effect can be positive in
manufacturing of light sources based on field-emission cathodes, where it is nec-
essary to achieve uniform flare of the luminophor.
The following describes a test structure of a light source employing the ring
effect. The light source consists of two parallel plates, which are anode and cathode.
The anode is made of a glass plate with a current-conducting coating and a lumi-
nophor layer. The distance between the anode and the cathode is fixed at 0.8 mm
with glass spacers. The cathode is made of graphite foil stuck down on a glass plate.
The emission centers are periodically located on the surface. Each center is about
1 mm. The centers are spaced 2.5 mm apart. In this design, the ring diameter is
3.2 mm, and the ring of one center does not overlap another center (Fig. 6.26).
That makes it possible to avoid destruction of the emission center by the
residual-gas ions generated by the neighboring centers. At the same time, all the rings
overlap each other, thereby providing uniform flare of the anode. The structure is put
in a vacuum flask, the residual-gas pressure being less than 10–6 Torr, followed by
field-emission tests. A photo of the working device is given in Fig. 6.27.
6.5 Large-Size Flat Field Emission Cathodes 329
Fig. 6.26 Relative

positioning of two nearest
emission centers
Fig. 6.27 The photo of the

working device using the ring
effect.
6.5 Large-Size Flat Field Emission Cathodes
A lot of applications, such as flat display screens, light sources, some SHF-devices
etc., need large-size field-emission cathodes, i.e., those of the area ranging from a
few cm2 to several hundreds of cm2. Now there is a great number of flat field
emission cathode fabrication methods, some of which we have already considered
in previous sections. They include thermal vapour-phase deposition (CVD) [109,
110], CVD in an electric arc [111], CVD in microwave plasma [112], laser-pulse
sputtering [113].
Each of these methods has a disadvantage of its own; e.g., all kinds of vapor
phase deposition are technologically sophisticated, which results in a high cost price
and a narrow range of suitable carbon materials mostly represented by different
kinds of diamond-like films and carbon nanotubes. Use of such promising carbon
materials as carbon fibers and pyrolytic carbon is thus out of question.
6.5.1 Electrophoresis
Electrophoresis, as well as electroosmosis, is an electrokinetic effect.

Electroosmosis is a flow of fluid arising in a porous medium at application of an
external electric field. Electrophoresis is a motion of particles affected by an electric
field in a motionless fluid. Electrokinetic phenomena arise due to formation of an
interface of an electrical structure. As a contact of two phases neither generates nor
eliminates but only redistributes electric charges between the phases, the adjoining
phases get the charges of an opposite sign but an equal value. What appears is a
system of charges referred to as an electrical double layer, which generates a
potential difference between the solid surface and the fluid. The interface-charge
sign and value depend on the nature of the surface and the fluid.
The mechanism of electrokinetic phenomena becomes quite clear if we imagine
a double layer at the solid surface affected by the force applied tangentially to the
interface. The charged fluid, affected by the external field, tends to move to the
opposite-charge pole. This charged-layer motion is transmitted to the neighboring
uncharged-fluid layers by internal friction. On the contrary, the same field makes
the solid surface with the attached charge move in an opposite direction. Depending
on what is motionless, either the fluid or the wall, it is either the solid phase or the
fluid that moves, viz. there is either electrophoresis or electroosmosis, respectively.
The electrophoresis rate of a free-form particle can be expressed as follows.
e1
Uef ¼ E; ð6:12Þ
4pg
where e is the dielectric permeability of the fluid, η is the fluid viscosity, E is the
electric field intensity, 1 is the electrokinetic potential.
The fundamental value affecting the particle velocity is electrokinetic potential 1,
which is determined by the interphase-potential jump and the double-layer
structure.
There are different theories of the double-layer structure, the mechanisms of its
formation and the effect on the electrokinetic potential. Papers [114–118] feature
and analyze these models in details.
The schematic diagram of electrophoretic application of carbon materials on the
starting-sheet surface is presented in Fig. 6.28 [119].
A glass vessel is filled up with a dispersion medium, viz. isopropyl alcohol, ISA,
with a salt of a high-valence metal dissolved in it and functioning as a charger, and
a carbon powder. In this case, the charger was La(NO3)3. The basic function of the
charger is intensification of application, as well as improvement of adhesion of
carbon particles to the substrate due to formation of a cementing film consisting of
the charger-metal oxide and hydroxide on the surface of the cathode during the
electrolytic reactions. Then the resulting solution assumes electrodes. The anode is
represented by a platinum plate. An advantage of this metal is its inertness in most
chemical processes. Due to this property, platinum is not subject to etching by
Fig. 6.28 Process scheme of electrophoretic application of carbon powder onto the substrate
charger anions during electrophoresis. In the electrophoretic process, the cathode is

represented by the substrate for the would-be field emission cathode. It is a glass
plate with an aluminium film applied. The gap between the electrodes is fixed with
glass spacers. The carbon-fiber particles deposit on the substrate after the electrodes
are supplied with constant voltage. The substrate is supplied with a negative
potential; respectively, the anode is supplied with a positive potential. As a rule, the
invariable parameters are the dispersion medium (isopropyl alcohol), the La(NO3)3
charger type, the carbon-powder concentration of 0.02 g/l, the anode being a
platinum plate.
The main variable parameters are the conducting-film geometry on the substrate;
the charger concentration; the space between the electrodes determined by the
spacer thickness; the application time; the current and the voltage between the
electrodes. The latter two parameters are interconnected and every particular
experiment is performed in a mode where one of them is fixed. These modes are
called current stabilization mode and voltage stabilization mode respectively.
Equation (6.12) determining the particle velocity, suggests that the applied
carbon-layer thickness depends on, firstly, the electrophoresis duration, and sec-
ondly, the carbon-powder concentration in the solution. The concentration value is
to be chosen so that it does not significantly diminish during the process. The
electrophoresis duration is chosen depending on the viscosity of the fluid used, the
electric field value and the electrokinetic potential value, which is rather difficult to
determine.
So the most acceptable technological parameters of the electrophoretic process
have to be determined by experiment.
The electrophoretic method can provide application of a wide range of carbon
materials. In particular, papers [120–123] describe electrophoretic deposition of
finely-divided diamond dust; papers [124–128] dwell upon application of carbon

nanotubes, and milled carbon fibers in [129] are put on different configuration
substrates and surfacing shapes. In particular, earlier papers, viz. [120–122, 130],
deal with diamond-dust deposition on the molybdenum-apex surfaces. Papers [125–
129] describe application on a flat substrate with a continuous conductible coating.
In many cases, the necessary emitting-surface pattern or the pixel system [124, 126]
is achieved by lithography.
The orientation of particles on a substrate is only considered in a few papers.
Paper [124] suggests using multiple-wall carbon nanotubes of 1–5 µm in length and
5–20 nm in diameter, the average diameter being 10 nm, as an emitting material.
The ultrasonic bath disperses the nanotube powder in isopropyl alcohol, after which
the centrifuge removes large particles from the prepared suspension. Then the
starting sheet and the counter electrode are placed in the resulting suspension,
following which electric voltage sufficient to generate a field of 2 103 V/cm
between the electrodes is applied to them.
Paper [129] is dedicated to electrophoretic application of a carbon-fiber powder.
In this case, the powder consists of polyacrylonitrile-carbon fiber segments of 7 µm
in diameter and about a few tens of micrometers in length. The charger is BaBr2.
Deposition covers the substrate with a continuous conductible layer of 10 cm2 at
the voltage of 320 V. The particles are horizontally spaced. Largeness of the par-
ticles and the edge effect prevent from uniform emission from the sample surface.
It is suggested to improve the uniformity and increase the amplitude of the
field-emission current by application of oriented carbon fibers.
The carbon fiber is known to have considerable anisotropy of field emission.
Electric field intensity being equal, the emission current from the carbon-fiber end
face exceeds that from the lateral face by several orders of magnitude. Moreover, if
the carbon fiber is oriented at right angle to the substrate surface, the electric field
enhancement factor goes up considerably. So the increase in effectiveness of the
field-emission cathode necessitates the orientation of powder particles. A way to
orient non-spherical particles is to put the sample in a strong magnetic field when
depositing carbon particles on the substrate [131].
To orient the particles in the necessary way during electrophoresis, a bath
hosting the cathode substrate, the counter electrode and the dispersion medium
containing the dissolved salts of the charger and the particles of ground carbon
fibers is put in a constant magnetic field generated by, e.g., permanent magnets. The
magnetic-field intensity vector should go at right angle to the starting-sheet surface.
The magnetic field vertically orients the elongated particles at right angle to the
cathode substrate. The magnetic field is set within 0.01–10 T depending on the
sizes of the particles applied.
Voltage from the appropriate source is applied between the cathode substrate and
the counter electrode. The voltage ranges from 10 V to 1.5 kV. In this case, the
cathode substrate is usually at negative potential. The ground carbon-fiber particles
adsorb positive metal ions from the electrolytic solution on their surface and
become positively charged. The field makes them move to the negative electrode of
the cathode substrate. The particles attach themselves to the cathode substrate due
to Van der Waals forces and formation of a thin film consisting of the charger
metal, its oxide and hydroxide.
The electrophoresis having been over and the cathode substrate having been
coated with an emitting layer, the cathode substrate is dried.
The above-mentioned operations result in a cathode coated with ground elon-
gated carbon-fiber particles and a given emitter-arrangement geometry, the particles
being oriented at right angle to the cathode surfaces and having a high adhesion to
the substrate.
This cathode-fabrication method provides the vertical orientation of particles on
the substrate, which greatly improves the emissivity of the resulting cathode in
comparison with the one where carbon-coating particles are located on the substrate
at random. Due to anisotropy of field-emission properties of carbon fibers, emission
from the particle end face is much better than from its lateral face. As a result, the
running voltages of the cathode based on vertically-oriented particles on the sub-
strate drop twice or thrice in comparison with those of the cathode based on
horizontally-arranged carbon-coating particles (Fig. 6.29). Moreover, the drop in
the running voltages is provided by increase in the field-enhancement factor at the
ends of the fibers in comparison with such enhancement on polydisposed fibers.
The large value of the emission current is in line with the cathode produced in
presence of the magnetic field. It is nearly five times as large as the current of the
cathode produced without the magnetic field (Fig. 6.29a). As the slope angles of the
curves plotted in Fowler–Nordheim coordinates practically coincide, this difference
cannot be caused by reshaping of the emission centers or the electronic work
function (Fig. 6.29b). It is most likely caused by increase in the number of the
emission centers. The emission centers have increased due to orientation of the
carbon particles affected by the magnetic field during electrophoretic deposition.
Fig. 6.29 Volt-ampere characteristics of the field emission cathodes manufactured by the method
of electrophoretic coating: a in current-voltage coordinates; b in Fowler–Nordheim coordinates; 1
characteristics of the field emission cathode manufactured without magnetic field; 2 characteristics
of the field emission cathode manufactured in presence of a magnetic field
6.5.2 Screen Printing
Another common technique of fabrication of large flat carbon field emission

cathodes is screen printing.
It was quite recently that the screen-printing technique started to be applied to
fabrication of flat field emission cathodes. The technique provides application of the
paste consisting of a carbon powder and an organic binder on a conductive substrate
through a stencil screen. Then the substrate with the paste applied is annealed up to
the temperature at which the organic component burns down. The organic binder is
mostly represented by ethylcellulose dissolved in terpineol.
The technological parameters of screen printing can be optimized in terms of
screen parameters, the optimum concentration of carbon powder in the paste and the
annealing temperature.
In screen printing, one of the main technological elements is the screen through
which the paste containing the carbon powder is applied. The thickness of the
resulting carbon coating on the substrate is determined by the screen type. The
screen type is determined by thickness and transparency. The transparency is meant
as the ratio of the open volume of the meshes to the total volume of the screen
multiplied by 100. Routinely for drawing of carbon paste use a grid with a trans-
parency of 50% and having thickness of 100 µm [132].
The thickness of the applied layer can be found by multiplication of the printing
screen height by its transparency. Taking into account that the temperature treat-
ment results in combustion and removal of the organic binder, the thickness of the
carbon layer on the cathode substrate can be found by multiplication of the
applied-layer thickness by the powder volume contained in the paste. The thickness
of the carbon layer can be determined as follows
d ¼ gbh; ð6:13Þ
where h is the screen thickness, η is the screen transparency, b is the concentration

of the carbon powder in the paste.
The most suitable carbon coating forming the emitting surface of a flat field
emission cathode is considered to be a coating uniformly applied along the entire
surface of the substrate and providing the maximum enhancement of the electric
field. The powder concentration in the paste determines the thickness of the
resulting coating, the extent of its irregularity and uniformity, which are the key
parameters forming the geometrical specificity of the cathode surface. The thickness
of the carbon layer consisting of one or more layers of the particles used ranges
from 2 to 5 µm. Such thickness can result from 4–10 volume percent concentration
of the powder in the carbon paste.
The next stage of fabrication of printed field emission cathodes is determination
of the optimum annealing temperature selected so that the binder is completely
removed from the carbon-coating volume. However, there should not be any
considerable oxidation of the carbon powder during the annealing. The necessary
temperature is established by thermogravimetric/differential thermal analysis,

TG/DTA, of the organic binder. This technique determines the temperature when
the investigated material has completely burnt down. The findings show that the
binder completely decomposes at the annealing temperature of at least 460 °C.
Then the cathodes are printed and annealed at three different temperatures, viz.
460, 500, 520 °C, for 5 min. According to the TG/DTA data, the organic binder is
completely removed from the field emission cathode surface. However, further
increase in temperature cannot but oxidise the carbon-powder particles. Hence, the
resulting field emission cathode surface differs in the oxidation extent of the
carbon-powder particles. Topographic investigations with a raster-type electron
microscope have confirmed that. As the temperature affects the surface topography,
it determines the emissivity of the resulting cathode, other field-emission cathode
(FEC) fabrication parameters being equal.
It has already been mentioned that the high-temperature annealing oxidizes the
carbon powder particles, which essentially worsens their emissivity. The exami-
nation of the surface of the field-emission cathodes produced at different annealing
temperatures shows that the most preferable temperature is 500 °C. The cathodes
exposed to this temperature have the optimal fibrillar structure of the surface.
Hence, such a cathode is expected to have larger form factors and lower running
voltages. Figure 6.30 shows the current-voltage characteristics of field emission
cathodes with three different annealing temperatures. The assumptions based on the
REM of surfaces of the resulting field emission cathodes have proved to be true, the
lowest-voltage sample being the field emission cathode annealed at 500 °C (cf.
Characteristic 2, Fig. 6.30).
The field-emission properties of the cathodes produced by printing are studied as
follows. First the current-voltage characteristic is measured, then the cathode is
aged in the voltage stabilization mode, the initial current being equal to the peak
current measured when taking the first current-voltage characteristic. The aging
Fig. 6.30 Volt-ampere characteristics of the field emission cathodes manufactured at various
temperatures of annealing: a in current-voltage coordinates; b in Fowler–Nordheim coordinates; 1
temperature of annealing is 460 °C; 2 temperature of annealing is 500 °C; 3 temperature of
annealing is 520 °C
time is about 60 min. After that, the second current-voltage characteristic is mea-
sured. The measurement shows the effect of the aging on emissivity of printed field
emission cathodes.
The lowest-voltage characteristic pertains to the cathode annealed at 500 °C
(Fig. 6.30). The characteristic of the field emission cathode annealed at 460 °C lies
in a higher-voltage area. The characteristics plotted in the Fowler–Nordheim
coordinates make it possible to recover the form factor and the area of the emitting
surface. The form factor of the field emission cathode annealed at 500 °C is the
largest among the form f actors of field emission cathodes. The reason is that the
annealing temperature of 500 °C brings out the fiber structure best of all other
temperatures. Apparently, the temperature of 460 °C does not provide complete
evaporation of the binder, which considerably affects the topography of the
resulting cathode. The annealing temperature of 520 °C is uprated anyway as the
carbon-fiber particles experience considerable oxidation. Field-emission research
confirms the assumptions made after the topographic investigation of field emission
cathodes exposed to thermal annealing under different conditions.
A side benefit can be thermal annealing in a special-gas atmosphere; e.g., paper
[133] attempts to anneal the coated plate on the basis of carbon nanotubes in the
argon atmosphere. However, the experiments show deterioration of emissivity as
compared with annealing in air. The reason is most likely the rests of the binder in
case of annealing in the argon atmosphere.
As the effective functioning of field-emission cathodes is widely known to need
their aging, development of field emission cathode aging techniques may be of
crucial importance for the wide practical use of printed carbon-based field emission
cathodes in electronic devices; e.g., paper [134] reports annealing of the applied
paste at 350 °C for 1 h followed by plasma etching of a 20 20 mm2 sample in a
gas mixture, the composition being undisclosed. The wattage is 300, the time being
40 and 160 min. Some increase in the emission current is only observed at increase
in the treatment duration.
According to the authors [135], pulsed plasma treatment triplicates the service
life of the field emission cathode. The treatment is carried out in the diode mode in
the Ne/Xe and Ne/Kr gas mixture at the pressure of *20 mm Hg. The variation of
plasma-treatment parameters, viz. the gas-mixture composition, voltage, pressure
and treatment time, makes it possible to find the optimum treatment of printed field
emission cathodes.
The field-emission properties of the field emission cathodes produced by
printing can be improved by addition of a dilatant material [136]. When carbon is
annealed, the film is fractionated, which increases the form factor, thereby
increasing the emission current.
The special areas that can bring success are a combination of printing with
different technologies. Let us give some examples of such combinations. It is worth
noting at once that such work has just begun. Paper [137] suggests the technology
combining photolithography and ink-jet printing. The flowchart of fabrication of
such a field emission cathode is shown in Fig. 6.31. Photolithography is used to
fabricate an open-hole array structure (Fig. 6.31b). Then the surface is

production of a field emission
cathode array using
lithography and an ink-jet
printing: a application of
thin-film coverings; b creating
perforations; c applying of
water-repellent covering to
the surface; d applying carbon
ink by method of an ink-jet
printing; e drying and ionic
etching in order to control the
gap between the extending
electrode and the cathode
water-proofed. That prevents from carbon-ink bleeding over the substrate and
reduces the leakage current between the modulator and the field emission cathode.
The next stage is carbon ink jet printing and drying (Fig. 6.31d) of the resulting
structure. Finally, the extraction electrode and the field emission cathode are
properly spaced by ion etching of the cathode structure in pure oxygen.
Paper [137] reports of fabrication of a structure containing 51 51 elements,
each of 100 µm in diameter, spaced at a pitch of 200 µm.
6.5.3 Structural Materials
Undoubtedly, the design of large field-emission cathodes produced by elec-

trophoresis, printing etc. potentially has a great practical value. However, the field
emission cathodes made of structural materials can be very useful in some cases.
Certainly, the greatest benefit expected from them is achievement of heavy
field-emission currents. However, some further improvement can make them more
universal.
The first field emission cathodes of this type are those made of high-strength
graphite, e.g., MPG-6 and MPG-7 [138–141]. These cathodes are produced in the
simplest process known that mostly stipulates only machining of suitable billets
[138–140], which is naturally followed by current aging. The microroughness can
also be further improved by other methods, e.g., radiation treatment [141].
Porous glassy carbon is also of interest because field emission cathodes based on
it show promising field-emission parameters [142–144]. This material consists of a
large number of pores widely ranging in size, the manufacturer reporting the
variation from 10 to 100 pores/in. The edges of these pores are sharp blade-like
structures, which potentially makes them good field emitters.
Paper [145] suggests forming nanoporous carbon on a conductive substrate, the
nanopore size ranging from 0.6 to 4.0 nm. Such nanoporous carbon results from
thermochemical treatment of the periodic table group III–IV carbide emitters with
chlorine at 400–1100 °C. Chlorination provides complete removal of the
carbide-forming element from the carbide composition. The chlorination is fol-
lowed by removal of chlorine from the nanoporous carbon lattice by treatment with
inert gases, hydrogen or ammonia at 500–700 °C. That is the way to form almost
any area of a field emission cathode.
6.6 Emissivity Improvement
Due to their structural features, carbon materials cannot provide identical form
factors over the entire surface of the extended field emission cathode or different
specimens of field emission cathodes. In other words, finding the way to provide
reproducibility of carbon field-emission cathodes is crucial for wide practical
application of such field emission cathodes. Now there are several ways to solve
this problem, which have to be closely integrated.
6.6.1 Field Emission Cathode Forming
Forming (ageing) is among the main ways to provide stable field emission from
carbon field emission cathodes [20]. Forming results from the effect of residual-gas
ions and ponderomotive loads of the applied electric field on the effective area of a
field emission cathode. Practically, it results from controlled stage-up or scale-up in
the emission current.
Field emission cathodes are formed on the basis of carbon to create the maxi-
mum number of emission centers uniformly spread over the surface of the field
6.6 Emissivity Improvement 339
emission cathodes and making nearly the same contribution to total emission
current, i.e. to create a well-developed emitting surface.
Apparently, field emission cathodes are formed according to the following
mechanism. Ion bombardment easily destroys the amorphous filler between the
crystallites, which enhances the field effect on them, thereby increasing pondero-
motive loads. That is accompanied by increase in the current from this region and
the density of the ion flow on it. The enhanced field can destroy the
cathode-material conglomerates coupled to the starting sheet in the weakest way.
So the physical meaning of forming of carbon field emission cathodes is
exposure of the internal skeletal structure of the cathode material with formation of
a dynamically equilibrium configuration of the emitting surface. At the same time,
forming to the level of the rated operating current does not result in exposure of the
skeleton in case of low currents. In this case, forming should proceed at increase in
the current up to its threshold value, when the exposure of the skeletal structure has
been over and the emission characteristics of the cathode have become stable.
As the ion concentration of residual gases directly depends on the pressure,
reduction of the pressure can reduce the forming time.
Development of the forming method assumes de-energisation if the vacuum
deteriorates to the threshold value.
The essence of the aging method is as follows. A high-voltage power unit is set
to the current stabilization mode. The value of the current is selected so that
functioning of the cathode over a short period of time deteriorates the chamber
vacuum. That is accompanied by intensive cathode destruction by ion bombard-
ment and forming of the optimum cathode surface structure. However, this process
cannot go on for a long time as increase in gas concentration in the vicinity of the
cathode intensifies bombardment and causes a snowballing process, leading to
disruption of the anode-cathode gap and irreversible destruction of the cathode. To
avoid that, the high voltage has to be switched off at a certain time until the vacuum
has improved; i.e., all evolved gases have been pumped away. After that, the
voltage is switched back on and the cycle is repeated until the cathode has stopped
gassing while functioning.
The value of the current suitable for aging is selected depending on the cathode
area and the method of its fabrication. For a particular cathode, this value is
determined by experiment.
6.6.2 Pre-treatment
Reduction of the forming time often necessitates pre-treatment of the field emission
cathodes produced. That somehow improves their geometrical characteristics and
increases the form factor. Let us give some examples of such a treatment below.
First of all, it is ion and plasma treatment of produced flat field emission cath-
odes, which has been detailed in previous sections. Carbon-fiber bundles can be
plasmachemically treated by corona discharge in air, the method having been
developed and approved. Such treatment can essentially improve emission char-
acteristics of the carbon-fiber cathode [146, 147].
When a field-emission cathode consisting of a carbon-fiber bundle functions,
electrostatic forces have been established to make the peripheral fibers divert. These
diversions result in oscillation of the emission current under pulse operation.
The matter is that the mechanical oscillations can tear off some fibers from the
bundle. The torn-off fibers can short-circuit the interelectrode cathode-modulator
gap. That used to be the case at a large amplitude of emission-current oscillation.
This situation is unacceptable because the purpose of the work done is development
of a reliable design of a cathode-modulator assembly with stable characteristics. So
it is necessary to somehow decrease or completely eliminate the diversion of the
peripheral bundle fibers.
The effect of electrostatic forces diverting the peripheral bundle fibers can be
reduced by shaping the fiber bundle so that it can provide maximally identical
electric field of all fibers in the bundle; i.e., cut some fibers projecting from the
bundle, as well as shorten the peripheral fibers. It is technologically difficult to do
that by machining of the fibers. To achieve this, a new method to treat a cathode
made of a carbon-fiber bundle has been suggested. It is plasmachemical etching of
the cathode carbon-fiber bundle by corona discharge in air as corona discharge is
closest to field emission in its nature.
The effect of the corona discharge on the carbon fibers [148, 149] is that
bombardment of the cathode with oxygen ions O2 causes oxidation of carbon C,
which leads to etching of the cathode material. So the length of a particular fiber
projecting from the fiber bundle decreases until its electric field enhancement factor
is less than or equal to other fibers from the bundle. Then the corona process
involves other fibers as well.
Accordingly, that equalizes the electric field on the surface of the carbon-fiber
bundle. In its turn, the equalization affects functioning of the field-emission cathode
in vacuum. So when the cathode is switched on, a considerable part of fibers begins
emitting because the unetched structure of the bundle begins electron emission from
particular fibers projecting from the overall mass.
Figure 6.32 represents some photos of the cathode carbon-fiber bundle taken
with an optical microscope. The bundle has been etched by corona discharge in air
under the conditions described above.
Figure 6.32 shows that corona-discharge etching makes the carbon-fiber bundle
round, the projecting fibers having been trimmed and the peripheral fibers
shortened.
The carbon-fiber bundle is etched away for a quite long distance (Fig. 6.32a, b),
and it is safe to say that almost all fibers of the bundle have taken part in the corona
process.
Under the above-stated conditions of plasmachemical treatment, the carbon
material of the bundle consisting of *300 fibers is etched in air at the rate of
*5 µm/min.
Fig. 6.32 Bunch of carbon fibers before (a, b) and after (c, d) corona discharge etching in air
The field-emission characteristics show an essential improvement of the

running-voltage behaviour dynamics. It is worth underscoring two positive things
about treatment of the cathode by corona discharge in air.
Firstly, the carbon-fiber bundle treated by corona discharge is less subject to
considerable restructurisation, i.e., redistribution of the fibers in the bundle. In this
experiment, the unetched-cathode behaviour dynamics demonstrates relatively
sharp drops and rises of the running voltage. However, the diagram of the cathode
etched by corona discharge does not practically show it.
Secondly, treatment of the fiber bundle by corona discharge essentially damps
the anodic-voltage local fluctuation. While it is DUA * 400–500 V for an unet-
ched cathode, it is DUA * 200–300 V for an etched one.
The treatment by corona discharge decreases the form factor of the cathode by
*20%, which slightly increases the running voltages; while etching increases the
emitting-surface area of the cathode severalfold. A small change of the form factor
along with a considerable growth of the cathode emitting-surface area means
increase in the number of emission centers on the cathode surface, plasmachemical
treatment by corona discharge making their distribution over the cathode surface
more uniform. That may improve stability of the running voltage in the
field-emission current stabilization mode over the long operating time of the cathode.
For flat nanotube-based field emission cathodes, it is suggested to improve
homogeneity of field emission over the field emission cathode surface by reduction
of carbon-nanotube thickness or density [150]. That can be controlled by the
nanotube growth time and process temperature.
6.6.3 Coatings and Doping
A promising way to improve field-emission properties of carbon-based field

emission cathodes is coating and/or doping of these materials with the elements
lowering the electronic work function and improving resistance of the cathode
material to ion bombardment.
The carbon materials being able to function as both doped and doping. In par-
ticular, paper [151] reports coating of silicon pointed field emission cathode matrix
with diamond-like graphites. The coating methods are CVD and electrophoresis. The
coating involved a synthetic high-pressure and high-temperature diamond, nanodi-
amonds with the crystallites sized 1–10 nm and produced by the method of
explosion. The coating is 0.2–0.4 µm thick. Field-emission research shows that
diamond high-pressure and high-temperature coatings cannot provide good stability.
At the same time, the nanodiamond powder-based coatings demonstrate a low
threshold electric field, a high level of electric current and its stability.
Moreover, the emission current drastically steps up by two or more orders of
magnitude at application of carbon nanoclusters on glass-graphite conic structures
[152]. A coating as thick as 0.2 µm makes it possible to generate the current density
of more than 10 A/cm2.
On the other hand, the coatings of different elements on carbon materials are most
widely spread. In particular, paper [153] reports application of platinum nanopar-
ticles on carbon nanorods. The emission is characterized by the threshold electric
field intensity of 4.17 V/µm and the current density of 45 µA/cm2 at 8 V/µm.
Much attention is paid to doping of carbon materials with different addition
agents that change the carbon structure of the material. Authors [154] have come to
the conclusion that boron-doped carbon nanotubes are more suitable for planar
arrays of field emission cathodes, e.g., flat displays. The electron-beam tubes, which
need high-current densities, are recommended to use [154] nitrogen-doped carbon
nanotubes grown on a tungsten apex.
It is not only carbon materials that can be doped, but also other structural
components of field emission cathodes. Paper [155] refers to H+ hydrogen-ion
doping of a p-Si (100) substrate cleaned in a usual way to avoid sputtering of the
catalyst film. Doping involved the energy of 90 keV and the dose of
5 10−6 cm−3. Then an electron beam sputtered an iron layer as thick as 5 nm.
The nanotube layer is applied in two stages. The first stage is RFCVD. Application
of this layer is preceded by vacuum annealing at 450 °C for 10 min to form
microcavities on the substrate surface. The second stage is treatment of the H2
sample by plasma at 750 °C, acetylene being introduced to form carbon nanotubes.
As a result, the nanotube layer has the holes of about 5 µm in diameter and 30 µm
in depth. This treatment almost halves the threshold field to 0.8 V/µm and densifies
the emission current. The central reason thereof is formation of holes in the nan-
otube layer because the hole edges experience an increase in the value of the
electric field. In addition, the treatment of the substrate makes the nanotubes
straighter.
Particular importance is attached to doping of carbon materials with the metals

reducing the electronic work function.
Now there are several ways to introduce alkaline and alkaline-earth metals in the
graphite structure and produce the interlaminar bonds of carbon [156], e.g., heating
up of the graphite powder with the implanted metal powder in a vacuum-sealed
ampule. This method is unacceptable so far because of the difficulties with pre-
conditioning of the samples for subsequent field-emission tests. It is still unclear if
there are any methods to fabricate a field emission cathode from the resulting
powder without physico-chemical treatment of activator-implanted graphite parti-
cles. So the samples are prepared by impregnation of the graphite plate with
implanted metal salts with subsequent vacuum annealing to the temperature known
to exceed that of decomposition of the salt used. This approach produces an acti-
vator uniformly distributed over the entire body of the sample. Further research
with Auger-spectroscopy and secondary ion mass spectroscopy show that the metal
implanted is in the pores in an almost oxidised state. However, the carbon surface
does not show any trace of the activator. Field-emission tests show that relatively
high voltages improve field-emission properties of field emission cathodes. The
reason is that ion bombardment removes the metal oxide from the pore and
decomposes it. Then the metal atom finds itself on the cathode surface, forming an
active-material monolayer, which reduces the electronic work function. However,
there is another process when ion bombardment destroys the activator monolayer.
That deteriorates the emissivity of the cathode.
During the experiment, some MPG-6 graphite plates are impregnated with
solutions of two different salts, viz. BaBr2 and Ba(ClO4)2. Then they are pulse
annealed in the vacuum of 10−5 Torr up to the temperature of 1000–1100 °C. It is
worth noting that the melting point of the BaBr2 salt is 857 °C; i.e., the tempera-
tures achieved in the experiment are above the melting point. There is no precise
information for Ba(ClO4)2, but the calculations show that the melting point of this
salt exceeds 1300 °C; i.e., the temperature achieved in the experiment is not suf-
ficient to melt it.
So BaBr2 is expected to decompose in vacuum during the annealing, whereby
more volatile Br, the boiling point being 59.2 °C, has to leave the sample and the
melted Ba has to impregnate the graphite matrix.
The further behavior of barium in the graphite sample depends on the sample
temperature and includes three processes. The first process is barium diffusion
through the pores to the sample surface and its further evaporation. The following
process is barium diffusion into graphite crystals. And the third process is formation
of barium oxide in the sample pores. The second and the third processes are the
most preferable because when the field emission cathode functions, barium oxide
decomposes, which results in formation of a mono-layer on the cathode surface.
Analysis of the sample structure after vacuum annealing identifies the effect of
all the three processes. The barium concentration on the sample surface is very low
and crescent with increase in depth. Sometimes barium concentration is much
higher in the crystallites than in the pores, which indicates barium diffusion into the
graphite structure.
Fig. 6.33 Series of

volt-ampere characteristics for
barium doped (1), and pure
(2) MPG-6 field emission
cathodes under identical test
conditions
Field-emission tests of the resulting samples are carried out in a diode structure.
The residual pressure in the chamber is *10−6 Torr. This paper only refers to the
field-emission tests of barium-doped samples. At first, the study focuses on the
features of MPG-6 pure graphite.
Figure 6.33 gives the current-voltage characteristics of one of such samples and
a series of current-voltage characteristics for a pure cathode under the same con-
ditions of field-emission tests.
The Ba-implanted sample demonstrates a considerable bias of the feature into
the area of low voltages, which may indicate increase in effectiveness of the
cathode. Here cathode efficiency is understood as the anode-cathode voltage nec-
essary to attain the preset emission-current value.
Similar results have also come for other barium salts [157], as well as for carbon
nanotubes [158] and graphite foil [159]. The latter case involves impregnation with
BaBr2 barium salt in water for 1 min in a saturated solution. According to the
measurements taken with an atomic-force microscope, the electronic work function
has gone down from 4.32 (cf. points 5–8 in Fig. 6.34) to 3.05 (cf. points 1–4).
Figure 6.34 shows these points with a white line. As a result, the current-voltage
characteristics have considerably biased into the low-voltage area.
Fig. 6.34 Image of an

alloyed cathode
6.7 Features of Measurement and Analysis … 345
6.7 Features of Measurement and Analysis

of Current-Voltage Characteristics of Carbon
Materials
6.7.1 Ways to Measure Current-Voltage Characteristics

and Their Features
In practice, the experimental dependence of field-emission current on applied

voltage is usually approximated by the following type of function.

B0
I ¼ A0 U 2 exp : ð6:14Þ
U
In this case, a dependence of ln(I/U2) on 1/U, viz. the so-called Fowler–

Nordheim coordinates, is represented with a straight line:
I B
ln ¼ Aþ ; ð6:15Þ
U2 U
its slope ratio giving us parameter B = −B′, and its crossing with the vertical axis
indicating parameter A = lnA′.
So in practice, it is coefficients A and B instead of cathode physical properties u,
bmean, Stotal that are really determined. However, the theoretical equation for total
emission current can relate experimentally determined parameters to real ones in the
following way

b2
A ¼ ln A0 ¼ ln Stotal a mean ; ð6:16Þ
u
u3=2
B ¼ B0 ¼ b : ð6:17Þ
btotal
It is worth noting that generally, determination of two approximation coefficients

for the current-voltage characteristic of the FEC in Fowler–Nordheim coordinates
does not provide unequivocal calculation of real parameters, viz. the work function,
the emitting surface area and the form factor. Moreover, it is to be shown below that
the above-mentioned simple analysis of current-voltage characteristics, CVC, is not
quite correct when we move from qualitative to quantitative consideration.
Measurements of current-voltage characteristics of a field emission cathode can
be taken in a slow step-by-step voltage build-up mode. They are the so-called slow
CVC. However, carbon-based FEC are characterised by strong time instability of
the emission current [160], which is why cathode parameters can drastically vary
during the period of measurement of a characteristic. To prevent the variation,
measurements are taken at the so-called turbo speed; i.e., the measurement time
Fig. 6.35 Typical

volt-ampere characteristics
obtained in fast and slow
modes: a normal coordinates;
b Fowler–Nordheim
coordinates
does not exceed the typical time of variation of cathode parameters, viz. Tcat.
Figure 6.35 shows the characteristics taken in different modes.
Unfortunately, turbo characteristics are more sensitive to imperfections of
experimental equipment [161]. There are a few principal causes that affect the CVC
in Fowler–Nordheim coordinates. Let us consider them in more detail.
In practice, the turbo speed of measurement of characteristics is usually
implemented as follows. Right after the control signal has been fed to the power
unit, current and voltage are measured as frequently as possible. However, there is
always a Tcap characteristic measurement time-limit below which we cannot go. It is
connected with availability of capacitive currents in the device. On the other hand,
we are limited by the characteristic cathode-parameter variation time, viz. Tcat, so
measurement time Dt has to be selected from the ratio Tcap Dt Tcat. The time
is usually about a few seconds, so if the number of points per characteristic totals
about one thousand, the measurement period takes some milliseconds. To meet
these requirements, the signal has to be digitised by analogue-to-digital converters,
ADC. As the signal at the ADC output is discrete, there are noises that interfere
with the real characteristic.
The ideal current-voltage characteristic of a field emission cathode in Fowler–
Nordheim coordinates is known to be a straight line (Fig. 6.36a) and follow (6.15).
The digitization of the signal with an ADC discretizes the current signal; i.e., the
current is represented in the integers meeting the number of the ADC counts.
Differently, if the count value is DI, and the current value is Ireal, the measured
current value, viz. IADC, is
Fig. 6.36 Modeling the

influence of ADC on the VAC
in Fowler–Nordheim
coordinates: a the ideal
characteristic; b accounting
for current value
discretization (step of
ADC = 0.1 lA)
Fig. 6.37 Accounting of non-ideality of the device model: a accounting for non-ideality of
contacts; b accounting for leaks in the device
Ireal
IADC ¼ DI : ð6:18Þ
DI
where sign D designates truncation. Taking this fact into account, we can express
the ideal characteristic as shown in Fig. 6.36b.
The ADC noise characteristics studied at the zero current suggest that the ADC
readings can vary within several counts.
In practice, when an anode and a cathode are mounted in a vacuum chamber or a
probe, there are always some imperfections in contacts (e.g. either high contact
resistances, or the cathode is highly resistive in itself) and the device itself (e.g.
there are current leakages between the anode and the cathode). Their effect on the
measuring circuitry can be estimated by adding stray resistance R in the circuit.
Addition of a series-connected resistance describes the situation when there is a
stray resistance in contacts (Fig. 6.37a). Addition of a parallel-connected resistance
describes the situation when there is a current leakage (Fig. 6.37b).
In this case, it is obvious that the real measures of voltage Umeas and current Imeas
differ from the values of current Ipr and voltage Upr of the device.
For serial connection the following holds:

Umeas ¼ Upr þ Ipr ðUpr ÞR;
ð6:19Þ
Imeas ¼ Ipr :
If the dependence of Ipr(Upr) follows the Fowler–Nordheim equation, then the

really measured characteristic is of the form shown in Fig. 6.38a. The diagram
shows that the CVC deviates from the ideal characteristic, this deviation depending
on the value of resistance. This result is also important in the case when the
measuring circuitry is extended with a ballast resistor to prevent damaging a
Fig. 6.38 Modeling the

influence of non-idealities of
the device model on the
aspect of measured
characteristics in Fowler–
Nordheim coordinates: a the
accounting for non-ideality of
contacts (serial connection);
b the accounting for possible
leaks (parallel connection)
high-voltage power unit in the event of short circuit. It is clear that the value of this
resistance should be selected so that it does not essentially distort the characteristic.
In case of a parallel connection, we have

Umeas ¼ Upr ;
ð6:20Þ
Imeas ¼ Ipr ðUpr Þ þ Upr =R:
In this case, the characteristic is computed in the same way as in case of a series
connection. Figure 6.38 shows the really measured characteristic as affected by
leakages.
The diagrams show that the shortcomings of the measuring bench can essentially
distort the current-voltage characteristics, which can later result in incorrect cal-
culation of approximation coefficients A and B in the Fowler–Nordheim coordi-
nates. In turn, the incorrect evaluation of parameters A and B can lead to
misidentification of the reasons of the cathode degradation. So it is necessary to pay
much attention to correctness of the measurements taken.
Research shows that a skill of proper measurement of the field emission cathode
CVC is often insufficient for an adequate evaluation of real field emission cathode
parameters, including the emitting surface area, the form factor and the electronic
work function. The bibliographic-data review shows [162, 163] that calculation of
these parameters on the basis of the measured A and B interpolation coefficients is a
difficult problem since the simple Fowler−Nordheim theory does not take account
of re-shaping or changing the height of the potential barrier as affected by the field,
considering the triangular barrier only. The Murphy–Good theory [164] gives a
more exact expression for field-emission current, viz.
0 pffiffiffiffiffi 1
bU
bv c u 3=2
ab2 U 2 B u C
I ¼ S pffiffiffiffiffi expB
@ C;
A ð6:21Þ
bU bU
t c u u
2
where t, v are Nordheim special functions and c = 3.79469 10−5 m=V eV.
pffiffiffiffiffiffiffi
Let us designate y ¼ c bU =u as the relative decrease in the height of the
potential barrier. In this case, the equation expressed in the Fowler–Nordheim
coordinates takes the following form, viz.

I Sab2 bvðyÞu3=2
ln ¼ ln 2 : ð6:22Þ
U2 t ðyÞu bU
This equation can be written in the previous form (6.15), the only difference
being that now coefficients A′ and B′ depend on the value of the field. However,
these coefficients must not be equated with the values derived as a result of
interpolation of the experimental curve (6.15).
To account for the difference between these parameters, let us consider
Fig. 6.39. It shows a straight line approximating the experimental dependence and
two straight lines with coefficients and A′ and B′ plotted for two values of voltage
U1 and U2, assuming that parameter A′ does not depend on the field because of a
weak dependence of function t on y. The parameter values are clearly seen to differ
greatly. So if the real parameters of a field emission cathode have to be evaluated
precisely enough, the analysis to be carried out is to be more sophisticated.
In practice, the field emission cathode parameters are usually calculated as fol-
lows. As function t(y) weakly depends on the value of the field, this dependence is
neglected. In this case, the equation for coefficient B takes the following form, viz.

bu3=2 1 dv
B¼ vðyÞ y : ð6:23Þ
b 2 dy
Fig. 6.39 The explanation of

distinction of coefficient of
approximation A from
subexponential multiplier in
expression (6.16) at the
accounting of dependence of
coefficient B from field
Let us introduce new function s(y) in the form of
1 dv
sðyÞ ¼ vðyÞ y ; ð6:24Þ
2 dy
pffiffiffiffiffiffiffi
bu3=2 bU
B¼ s c : ð6:25Þ
b /
Knowing the expression for coefficient B, it is easy to find the value of A as

Sab2 bu3=2
A ¼ ln þ ½vðyÞ sðyÞ: ð6:26Þ
t ðyÞu
2 U
So the mathematical treatment of current-voltage characteristics aimed at eval-

uation of the real parameters of a field emission cathode has to involve expressions
(6.25) and (6.29) instead of (6.16) and (6.17).
6.7.2 Technique of Long-Term Field-Emission Tests

and Experimental Data Analysis
At present there are two main ways to study stability of field-emission current over
a long period of time. The first of them [165] is periodic measurement of current at
fixed voltage. However, this technique is not good to study emissivity of
carbon-based FEC because of strong time instability of parameters of such cathodes
[160]. Due to the high transconductance of the current-voltage characteristic, even

minor voltage variations can cause essential current fluctuations.
Another technique [160, 166] is to study the voltage behavior, maintaining the
current constant. This approach provides much less voltage fluctuations, which
facilitates further analysis of the results. During such tests, the cathode works in a
harder mode than during the tests involving maintenance of voltage constant. In
addition, this mode is closer to actual conditions of operation of a field emission
cathode.
However, both of these techniques determine the behavior of a FEC only at fixed
voltage or current. However, that may be insufficient for some applications; e.g.,
such measurements cannot be used to predict the behavior of a device joined up in
the stabilizer circuit with transfer-function I = f(U). In addition, it is not only the
behavior of a field emission cathode in the gap to be measured that is usually of
interest, but also prediction of its further functioning in this or that mode. Below
there are two techniques suggested to predict the behavior of the FEC on the basis
of carbon nanomaterials in a long run.
This technique is based on periodic measurement of current-voltage character-
istics of a field emission cathode all over the operating time. Its main advantage is the
ability to predict the cathode behavior at any current or voltage variation law [167].
In this technique, a FEC is switched on in this or that mode, e.g., at fixed current,
and a series of current-voltage characteristics is measured after a certain period of
time. Then every characteristic is represented in the Fowler–Nordheim coordinates
and approximated by a straight line (6.15) to find coefficients A and B. Over the
operating time, the cathode is affected by a series of physical factors causing its
degradation. This effect makes the device parameters vary to some extent. It is these
changes that are fixed as time dependence of parameters A and B. Then the
approximation coefficients are analyzed to select the main processes accounting for
degradation of the device in this or that work area. In some cases, it is also possible
to carry out quantitative evaluation of the effect of these factors on the cathode
parameters by the mathematical apparatus of the Fowler–Nordheim theory.
Selection of U(t)-meeting points from every current-voltage characteristic gives
dependence I(t). Differently, the period of measurements makes it possible to
predict the behavior of a field emission cathode under any control action I = f(U).
However, it is usually interesting to determine not only the behavior of emission
current in the gap to be measured, but also to predict its further functioning in the
same mode. With this end in view, it is necessary to determine the time variation of
the characteristics. Differently, it is necessary to approximate the dependences of
parameters A and B by some functions. Knowing these dependences, it is possible
to extrapolate them to the periods of time longer than those measured. Then the
characteristics calculated can predict behavior of the device within any period of
time we are interested in.
It is usually impossible to determine coefficients A and B with adequate accu-
racy. The reason is that various noises always make the experimental points scatter
around the straight line. However, if the experimental points scatter, they can be
connected with a straight line to approximate this dependence in the best possible
Fig. 6.40 Emergence of area of acceptable values as a result of dispersion of experimental points:
a experimental points (the best line I and two other lines (II and III) that also describe this
dependence with a smaller coefficient of correlation); b areas of acceptable values A and B (the set
of points corresponding to the approximating lines. Points corresponding to I, II and III are
marked)
way. If the requirements to the correlation factor are mitigated, these points can be
connected with not one, but more straight lines, which more or less precisely
describe this dependence (Fig. 6.40a). The result is that this characteristic is rep-
resented in coordinates A–B with not one point, but an area of acceptable values
(Fig. 6.40b). Then it is clear that if we have a set of such areas, there may be
situations when it is impossible to say unequivocally whether parameters A and/or
B vary in time at all (Fig. 6.40).
In this case, it is expedient to resort to the following approach. Let there be not
much variation in one parameter, e.g., B, during the experiment. That means that the
projections of the areas of acceptable values to axis B intersect. Having fixed this
parameter at an average level, we can recalculate parameter A according to the
formula A ¼ hY iBh X i; i.e., if the time variation of coefficient B is less than the
scattering in its determination, B can be considered invariable and only dependence
of coefficient A can be analyzed. As a result, the two-parametric problem is reduced
to one-parametric.
The authors have tried this testing technique. The test specimen is carbon
nanotube based FEC. The cathode is tested in the following mode (Fig. 6.41). The
testing time totals to about 600 h. The cathode works in a direct-current mode of
100 µA throughout the experiment, the current and voltage being measured at the
device every 10 s. In addition, current-voltage characteristics of the device are
measured every 30 min. The characteristics measured determine coefficients A and
B of the Fowler–Nordheim equation.
To check if the prediction of cathode behavior in the same operating mode
within long-term periods is true, the first 100 h are cut out from the operating time
and the cathode behavior is calculated for the next 500 h according to the technique
Fig. 6.41 Results of

long-term tests of field
emission cathodes out of
carbon nanotubes: a plots of
change of current and tension
in time; b change of
coefficient A; c change of
coefficient of B
suggested. To evaluate adequacy of the technique, the prediction is compared with

real experimental data.
Prediction of cathode work involves approximation of the dependences of
parameters A and B by some functions. The result essentially depends on what
functions are used to describe the cathode-behaviour dynamics. Figure 6.42 shows
the prediction data with involvement of two different approximating functions. In
the former case, the dependences of coefficients A and B are approximated by linear
functions (Fig. 6.41, curve I). However, Fig. 6.42 (curve I) shows that the periods
of time in excess of 300 h have appreciable deviations of the calculated
voltage-time dependence from the real behavior. That can be avoided if we
know/surmise the cathode-degradation mechanism.
Assuming that the main FEC degradation mechanism is a linear emission-center
form-factor variation b; the dependence of coefficients A and B on b being taken
into account, their dynamics in time is described by the following functions
(Fig. 6.41, curve II), viz.
Fig. 6.42 Forecasting the

cathode function. Two
settlement curves
corresponding to various
approximations of
dependences of coefficients
A and B. Equations of
approximating straight lines
are shown:
I A = 14 – 0.005t and
B = −20 – 0.014t; II A = ln
(6 105 − 900t) and
B = 4 104/(2 103 – t)
b0
A ¼ lnða1 a2 tÞ; B¼ ; ð6:27Þ
b1 t
where a1, a2, b0 and b1 are constant, t being the time. Figure 6.42 (curve II) shows
the predicted data. It demonstrates that curve II is in good agreement with the
experimentally observed dependence. So adequate application of the technique
suggested depends on preliminary identification of the main degradation mecha-
nisms of the FEC to be tested.
6.7.3 Evaluation of Uniformity of Emission from a Field

Emission Cathode Surface
In addition to high performance, long service life and large area of planar field
emission cathodes, an important parameter is uniformity of emission-center dis-
tribution over the surface. These FEC are usually applied in design of light-emitting
indicators and information displays, where the problem of uniform flare of one
element/pixel is especially pointed. Moreover, this parameter is also indirectly
connected with stable cathode operation. In particular, if only a few emission
centers work efficiently, destruction of even one of them drastically changes the
characteristics of the field emission cathode. If uniform emitting comes from the
entire surface of the cathode having a great number of centers, destruction of a few
of them does not cause any considerable fluctuations of the emission current.
In practice, uniformity of emission from the surface of a planar field emission
cathode is evaluated with transparent anodic plates coated with a thin luminophor
layer to transform the flow of incident electrons into visible light. Accordingly,
when the field emission cathode functions, the so-called emission image appears at
the anode. It characterizes distribution of emission centers over the field emission
cathode surface, as well as represents their efficiency, i.e. their contribution to the
total current.
Uniformity of distribution of emission centers over the FEC surface is suggested
to be measured with the following procedure. High voltage is applied between the
cathode and the anode to generate the current heavy enough to photograph the
emission image formed at the anode by a digital camcorder in a monochromic, viz.
black-and-white, mode. The resulting digital photo is kept in a personal computer,
and the appropriate software coverts it into a histogram of distribution of luminosity
over the anode surface, an isocandle histogram, according to the following formula
X

hk ¼ yk xi;j ; ð6:28Þ
i;j
where i is the coordinate of the image point in the line, j is the coordinate of the
image point in the column, k is the number showing the light-intensity value,
0 k 255, 0 and 255 being the minimum and maximum luminosity, respec-
tively, hk is the number of image points of luminosity k, xi,j is the luminosity of the
image point with coordinates (i, j), yk is the function of the following type

1 x ¼ k;
yk ðxÞ ¼ ð6:29Þ
0 x¼6 k:
Uniformity of distribution of emission centers over the cathode surface is cal-

culated with the resulting histogram and the following formula
X
255
u¼ hk =M; ð6:30Þ
k¼P
where u is the uniformity coefficient, M is the total number of emission-photo

points.
To illustrate the calculation method, let us consider two emission images made
for carbon nanotube-based field emission cathodes. Figure 6.43 shows the images
themselves, as well as the typical isocandle histograms. Selection of threshold
luminosity P facilitates calculation of uniformity of anode flare.
Despite the apparent simplicity of the technique suggested, there is a series of
difficulties in its practical application. One problem is selection of an adequate
value of the threshold luminosity used in the calculations. In particular, an increase
in the emission current drawn improves the uniformity of the flare spot not only due
to increase in the number of emission centers, but also due to total increase in
luminosity of the luminophor. In the latter case, it often happens that the chamber is
flared, and the particular less bright elements of the image cease to be differentiated
(Fig. 6.43b). The situation is remedied by changing/reducing the chamber aperture
that controls the value of the light flux incident on the CCD-matrix. However, it is
clear that a different size of the chamber aperture causes selection of similar
Fig. 6.43 Example of an assessment of degree of uniformity of distribution of emission centers on

a field emission cathode surface. The left part of the figure shows the images themselves, and the
right one shows histograms of distribution of brightness: a low uniformity of distribution of the
centers; b high uniformity of distribution of the centers
threshold luminosity to give different values of uniformity of the same emission

image. So in practice, it is recommended to compare emission images at the same
extent of luminosity of the phosphor screen.
6.7.4 Developed Algorithm of Collection and Preprocessing

of Experimental Data
This section suggests a universal algorithm of field-emission testing and preliminary

analysis of experimental data. According to the authors, this technique provides the
most complete information about the test specimen. This algorithm can serve as the
first approximation in standardization of field emission cathode testing [168].
The software algorithm of collection and preprocessing of data can be best
accounted for in the following terms.
Let the period of time when voltage of a high-voltage source is maintained
constant be referred to as Implementation. So the entire experiment is a population
of such implementations, the voltage level varying from implementation to
implementation. Such a high-voltage monitoring circuit makes it possible to age the

cathode. Moreover, when each implementation comes to the end, a period of time
when high voltage is not applied to the cathode is added, the additional period of
every implementation lasting in its own way. In other words, the cathode is meant
to have a rest. In its turn, that makes it possible to determine whether or not the
number of switchings on and off is a critical parameter for this cathode; i.e., whether
or not the cathode parameters change after its rest or there is a period of time after
which the FEC works steadily. It is clear that continuous change of parameters after
each switching on or off is unacceptable for a ready device. It is another matter
when there is a period of time before the beginning of steady operation. In this case,
the cathode can be aged as early as in the course of fabrication to improve its
operational stability irrespective of the number of switchings on and/or off.
Now let us consider the structure of implementation itself. Every
Implementation is subdivided into a number of Measurements, the measurements
of an implementation being strictly synchronized for the beginning of a new
implementation to coincide with the beginning of a new measurement and made in
preset periods of time similar for all the implementations. Hence, the total number
of measurements depends on the ratio of the implementation time to the period
between the measurements. It is also necessary to mention that measurements are
made throughout the implementation, including those periods of time when the
cathode is not supplied with high voltage, the time when high voltage is switched
on or off being synchronized with the beginning of a new measurement. All the
above-stated measures make it possible not only to trace cathode degradation in a
stationary area, but also to observe the processes of setting after the applied voltage
has varied in level.
Measurement itself is a population of a number of Frames, viz. individual acts
of reading the value of current and voltage. The number of frames in one mea-
surement is also strictly fixed and identical for each measurement, the frames being
separated from each other by a preset period of time constant for each measurement,
the minimum time being determined by finiteness of the digitization time by one
value of signals from two ADC channels, taking into account the time to change
over from one ADC channel to another. In other words, the frames are synchro-
nized, too, the beginning of a new measurement coinciding with the beginning of a
new frame. However, unlike synchronization of implementations and measure-
ments, which is carried out with a computer-system timer, the frames are syn-
chronized by the LTC crate timer. It is clear that the maximum number of frames is
determined by the ratio of the measurement duration to the period of time between
the shots.
So the arrangement of any experiment in terms of Implementation,
Measurement and Frame is shown in Fig. 6.44.
In addition, a few tests of this algorithm are followed by a decision to add a
possibility to measure the current-voltage characteristics of the cathode at the
moments when high voltage is switched on and off. The number of the charac-
teristics to be measured is set for each implementation, a possibility to set having
been implemented as a number of characteristics of high-voltage switching on and
Fig. 6.44 Experiment scheme. Characteristic values of times are specified: a implementation
scheme; b measurement scheme
off. If it is necessary to study getting to the steady-state operation, measurement of

characteristics is switched off and take measurements according to the plan featured
above. The characteristics are measured at the so-called turbo speed; i.e., when right
after injection of a control signal from the computer to a high-voltage power unit, a
great number of values of current and voltage, usually *5000, are taken within as
short a period of time as possible, usually just a few seconds. This way to measure
CVC makes it possible to avoid considerable variations of cathode parameters
during the measurement period.
The software developed also preprocesses the data obtained. The preprocessing
procedure is as follows. The values of current and voltage frames taken over one
measurement are averaged to find the average value and variance at the moment of
measurement. So every measurement gives 4 figures: the average current and
voltage values and their variance. These data are filed with specification of the
measurement time, the time specified being that returned by the system timer; so
even if there is a measurement bug, every couple of current-voltage values meets
the true time of measurement. Some ordering information is also filed, namely the
implementation number, the number of measurement within the implementation,
the time elapsed from the beginning of the implementation. These data make it
possible to test the software at the stage of development and promptly fix the
defects.
The software also records the files with volt-ampere characteristics, so-called
second-type files, on the hard disk of the computer. The file also keeps the
characteristic-measurement time and the data themselves, i.e., the voltage column
and the respective current column. Before being filed, the data are also reprocessed.
The current values below 5 ADC counts are cut off to eliminate the effect of ADC
noise discretization. Then the characteristic obtained is recalculated in the Fowler–
Nordheim coordinates and approximated by a straight line to find coefficients A and
B. These parameters are also filed. To accelerate the recording, viz. reduce the
errors of data desynchronization in between the implementations, not all the data
are filed but only *500 values uniformly selected from the entire array.
Before the beginning of the long-term test, a test circuit for the particular cathode
is laid. With this end in view, a file is created to contain (1) t1 as the period of time
when high voltage is switched on; (2) t2 as the period of time when high voltage is
switched off; (3) U as the maximum voltage; (4) I as the peak cathode current,
depending on which of these restrictions works up earlier, there is either current or
voltage stabilization; (5) N1 and N2 as the number of volt-ampere characteristics
measured at the beginning and at the end of the implementation, respectively.
The selected experimental-arrangement algorithm makes it possible to carry out
long-term tests of almost any complexity; e.g., it is possible to carry out both simple
long-term tests in the constant voltage or current maintenance mode and sophisti-
cated tests with periodic measurements of cathode characteristics. One more
advantage of this experimental arrangement is the simplicity of adaptation of
measuring to specific targets; i.e., long-term test arrangement can be amended
without a radical change of the control program, the only amendment necessary
being that of the file containing the experimental arrangement.
References 361
References
1. E.P. Sheshin, Structure of a Surface and Autoissue Properties of Carbon Materials (M of a.:
MFTI-fizmatbook, 2001)
2. E.P. Sheshin, Field emission of carbon fibers. Ultramicroscopy 79, 101–108 (1999)
3. E.P. Sheshin, Properties of carbon materials, especially fibers, for field emitter applications.
Appl. Surf. Sci. 215, 191–200 (2003)
4. E.P. Sheshin, Modern ways of formation of field electronic cathodes from carbon materials.
Achievements Mod. Radio Electron. 5–6, 36–40 (2004)
5. Kypryshkin, E.P. Sheshin, A.A. Shyka, Methods of production of field electronic cathodes
from carbon materials. Nano Microsc. Equip. 3, 26–31 (2005)
6. E.P. Sheshin, Possibility of receiving big autoissue currents from autocathodes from carbon
volokn. Electron. Equip. Ser. 4. 2, 58–62 (1988)
7. E.P. Sheshin, L. Rybakov Yu, Field electronic cathodes from carbon volokn. Theses dokl.
xviii Vsesoyuzn. Konf. on issue electronics, pp. 213–214, M.: nauka (1981)
8. V.A. Nevrovsky, V.I. Rahovsky, To a question of time of development of thermal instability
of microledges on the cathode at vacuum breakdown. ZhTF 50, 2127–2135 (1980)
9. B.V. Bondarenko, V.I. Makyha, V. Titov Yu, E.P. Sheshin, Autocathodes with a big
working area. Electron. Equip. (It is Gray. SVCh Electron) 4, 47–51 (1986)
10. A.S. Fialkov, Carbon and Graphite Materials (M.: energiya, 1979), p. 320
11. E. Braun, J. Smith, D. Sykes, Carbon fibers as field emitters. Vacuum 25(9/10), 425–426
(1975)
12. E.P. Sheshin, Emission characteristics carbon fiber, in Physical Processes in Devices of
Electronic Equipment (M.: MFTI, 1980), pp. 6–10
13. R.M. Hatapova, V.H. Romanova, About issue stability of carbon autocathodes in the
unsoldered devices. Theses дoкл. IV Vsesoyuzn. A symposium on not heated cathodes,
Tomsk, ISE, p. 51 (1980)
14. B.V. Bondarenko, V.A. Sileverstov, E.P. Sheshin, Issue properties carbon вoлoкн various
temperature of processing. Radio Technician Electron. Eng. 30(8), 1601–1605 (1985)
15. B.V. Bondarenko, V.I. Makyha, E.A. Tishin, E.P. Sheshin, About work of an exit of
electrons of carbon materials, in The Physical Phenomena in Devices of an Electronic and
Laser Teznika (M.: MIPT, 1983), pp. 13–18
16. B.V. Bondarenko, V.I. Makyha, E.P. Sheshin, Stability of issue and durability of some
options of autocathodes. Radio Technician Electron. Eng. 28(8), 1649–1652 (1983)
17. B.V. Bondarenko, A.G. Shahovskoy, E.P. Sheshin, Stabilization of autoissue characteristics
carbon вoлoкн during the long work, in Physical Yaleniye in Devices of Electronic and
Laser Equipment (M.: MIPT, 1985), pp. 4–9
18. A.S. Baturin, T. Kelly, M.S. Mousa, E.P. Sheshin and others, Lifetime and emission stability
of carbon fiber cathodes. Mat. Sci. Eng. 353, 22–26 (2003)
19. B.V. Bondarenko, V.I. Makyha, E.P. Sheshin, Field electronic cathodes from graphite.
Theses dokl. iv Vsesoyuzn. A symposium but to not heated cathodes, Tomsk, pp. 49–50
(1980)
20. V.I. Makyha, E.P. Sheshin, About possibility of receiving big autoissue currents from
graphite, in The Physical Phenomena in Devices of Electronic and Laser Equipment (M.:
MIPT, 1983), pp. 22–25
21. E.P. Sheshin, A.C. Batyrin, Stability field electron emission uglerodnovolokonnykh of
autocathodes. Mater. The All-Russian symposium on Amis. To electronics, Ryazan,
pp. 141–142 (1996)
22. E.P. Sheshin, A.B. Stolarov, A.V. Anashenko, A.H. Kydravcev, N.B. Dyakonova, Studying
of issue properties and structure carbon вoлoкн various types. Mater. VII meetings
“Radiation physics of a solid body”, Moscow-Sevastopol, pp. 78–82 (1997)
23. E.P. Sheshin, A.V. Anaschenko, S.G. Kuzmenko, Field emission characteristic research of
some type of carbon fibers. Ultramicroscopy 79, 109–114 (1999)
24. D.M. Paytov, V.A. Tkachenko, V.S. Neshpor, E.I. Podolskay, Mnogoostriyny cold cathode.
Ampere-second. USSR, No. 767858. Cl. H01j 1/30 of 14.06.78
25. D.M. Paytov, V.A. Tkachenko, V.H. Fedorov, E.V. Gorbachevskiy, Mnogoostriyny cold
cathode. Ampere-second. USSR, No. 1019518. Cl. H01j 1/30 of 15.01.82
26. A.G. Chaknovskoi, E.P. Sheshin, A.S. Kupryashkin, V.A. Seliverstov, Method of fabrication
of matrix carbon fiber field emission cathode structure for flat-panel indicators. J. Vac. Sci.
Techn. 11(2), 511–513 (1993)
27. T.N. Bredihina, V.K. Esina, A.U. Cherepanov, A.G. Shohovskoy, E.P. Sheshin, Way of
production of the multiemitter autocathode. Ampere-second. USSR, No. 151606. Cl. H01j
9/02 of 15.10.87
28. A.S. Batyrin, V.I. Beloglazov, V.F. Lebedev, N.B. Skibina, A.V. Sherbakov, E.P. Sheshin,
Way of production of the matrichesky autocathode. Stalemate. Russia No. 2183362. Cl.
H01j 1/14 of 04.04.2001
29. R.R. Bessette, M.G. Madeiros, C.J. Patrissi, C.M. Deschenes, C.N. La fratta, Development
and characterization of a novel carbon fiber based cathode for semi-fuel cell applications.
J. Power Sour. 96, 240–244 (2001)
30. T. Keesmann, W.H. Grosse, Field emission cathode using carbon fibers. Pat. USA
2004/0036402. Cl. H01j 1/05 (313/311), of 8.04.2003
31. M.A. Guillorn, M.L. Simpson, G.J. Bordonaro, V.J. Merkulov, L.R. Baylor, D.H. Lowndes,
Fabrication of gated cathode structures using an in situ grown vertically aligned carbon
nanofiber as a field emission element. J. Vac. Sci. Tech. 19(2), 573–578 (2001)
32. V.D. Blank, E.V. Polyakov, D.V. Batov, B.A. Kulnitskiy, U. Bangert, A. Gutierrez-Sosa, A.
J. Harvey, A. Seepujak, Diam. Rel. Mater. 12, 864–869 (2003)
33. Y. Wada, Y.K. Yap, Y. Mori, M. Yoshimura, T. Sasaki, Diam. Rel. Mater. 9, 620–624
(2000)
34. R. Kurt, J.M. Bonard, A. Karimi, Thin Solid Films 398, 193–198 (2001)
35. A.G. Kydashov, A.V. Okotryb, N.F. Udanov, A.I. Romanenko, L.G. Bylysheva, A.G.
Abrosimov, A.L. Chyvilin, E.M. Pazhetov, A.I. Boronin. FTT 44(4), 626–629 (2002)
36. M. Terrones, XV International Winterschool on Electronic Properties of Novel Materials
(Austria, 2001), p. 63
37. C. Kimura, Y. Yamamuro, H. Aoki, T. Sugino, 17th European Conference on Diamond,
Diamond-Like Materials, Carbon Nanotubes, and Nitrides (Portugal, 2006), p. 380
38. V.S. Bormashov, E.P. Sheshin, D.V. Baton, V.D. Blank, S.G. Buga, Novel method of flat
cold cathode production from carbon-nitrogen nanofibers. Technical Digest of 19th IVNC
and 50th IFES, Guilin, China, p. 293 (2006)
39. V.D. Blank, E.V. Polyakov, B.A. Kulnitskiy, A.A. Nuzhdin, L. Alshevskiy Yu, U. Banget,
A.J. Harvey, Thin Solid Films 346, 86–92 (1999)
40. P.M. Ajayan, M. Terrones, A. De La Guardia, V. Hue, N. Grobert, B.Q. Wei, H. Lezec, G.
Ramanath, T.W. Ebbe, Science 296(5568), 705 (2002)
41. N.G. Shang, F.C.K. Au, X.M. Meng, C.S. Lee, I. Bello, S.Y. Lee, Chem. Phys. Lett. 358,
187–191 (2002)
42. Z.Y. Kosakovskay, L.A. Chernozatonskiy, E.A. Fedorov, Lett. ZhETF 56(1), 26–30 (1992)
43. S. Fan, M.G. Chapline, N.R. Franklin, T.W. Tombler, A.M. Cassell, H. Dai, Science 283,
512–514 (1999)
44. A.G. Rinzler, J.H. Hafner, P. Nikolaev, L. Lou, S.G. Kim, D. Tomanek, P. Nordlander, D.T.
Colbert, R.E. Smalley, Science 269, 1550–1553 (1995)
45. Q.H. Wang, T.D. Corrigan, J.Y. Day, R.P.H. Chang, A.R. Krauss, Appl. Phys. Lett. 70,
3308–3310 (1997)
46. J.-M. Bonard, J.-P. Salvetat, T. Stockli, W.A. De Heer, L. Forro, A. Chatelain, Appl. Phys.
Lett. 73, 918–920 (1998)
47. H. Schmid, H.-W. Fink, Appl. Phys. Lett. 70, 2679–2680 (1997)
48. W.A. De Heer, A. Chatelain, D. Ugarte, Science 270(5239), 1179 (1995)
49. V. Gulaev Yu, N.I. Sinicin, G.V. Torgashov, L.A. Chernozatonsky, Z.Y. Kosakovskay, F.
Zaharchenko Yu, Microelectronics 26(2), 84–88 (1997)
References 363
50. V. Filip, D. Nicolaescu, M. Tanemura, F. Okuyama, J. Vac. Sci. Technol. 21, 382–390
(2003)
51. T.C. Choy, A.H. Harker, A.M. Stoneham, J. Phys. Matter. 16, 861–880 (2004)
52. T.C. Choy, A.M. Stoneham, A.H. Harker, J. Phys. Condens. Matter 17, 1505–1528 (2005)
53. A.N. Obraztsov, I. Pavlovsky Yu, A.P. Volkovv, J. Vac. Sci. Technol B. 17, 674–678 (1999)
54. A.N. Obrazcov, A.P. Volkov, I. Palovskiy Yu, A.L. Chyvilin, N.A. Rydina, V.L. Kyznetsov,
Lett. ZhETF 69(5), 381–386 (1999)
55. A.N. Obraztsov, J. Pavlovsy, A.P. Volkov, V.L. Kuznetsov, A.L. Chuvilin, Oriented carbon
nanotubes growth for field emission application. Mat. Res. Soc. Symp. Proc. 558, 111–116
(2000)
56. H. Hiura, T.W. Ebbesen, J. Fujita, K. Tanigaki, T. Takada, Nature 148–151 (1994)
57. F. Salver-Disma, J.-M. Tarascon, C. Clinard, J.-N. Rouzaund, Carbon 37, 1947–1959 (1999)
58. J.K. Huang, H. Yasuda, H. Mori, Chem. Phys. Lett. 303, 130–134 (1999)
59. Y. Omaru, M. Nagamine, J. Appl. Phys. 88, 6389–6377 (2000)
60. K.E. Robinson, D.D. Edie, Carbon 34, 13–36 (1996)
61. S.-H. Hong, Y. Korai, I. Mochida, Carbon 38, 805–815 (2000)
62. M.I. Elinson, Not Heated Cathodes (M.: Soviet Radio, 1971)
63. V.T. Binh, C. Adessi, Phys. Rev. Lett. 85, 864–867 (2000)
64. V.M. Lobanov, ZHTF 75(11), 92–96 (2005)
65. V.D. Frolov, A.V. Karabutov, S.M. Pimenov, V.I. Konov, V.P. Ageev, Diam. Rel. Mater.
10, 1719–1726 (2001)
66. M.S. Dresselhaus, G. Dresselhauss, P. Eklund, The Science of Fullerenes and Carbon
Nanotubes (Academic, 1996)
67. T.W. Ebbesen (ed.), Carbon Nanotubes, Preparation and Properties (CRC Press, 1996)
68. R. Saito, G. Dresselhaus, M.S. Dresselhaus, Physical Properties of Carbon Nanotubes
(World Scientific, 1998)
69. A.V. Eleckiy, Carbon nanotubes and their issue svostvo. UFN 172(4), 401–438 (2002)
70. Y. Zhang, S.Z. Deng, J. Cheng, N.S. Xu, Synthesis and Field Emission Properties of
Large-Area Uniform Carbon Nanotubes Films (Techn Didgest JVNC, Oxford, UK, 2005),
pp. 324–325
71. C. Zhu, C. Lou, W. Lei, X. Zhang, Fabrication and Characterization of High Current
Density Carbon Nanotubes Cold Cathodes (JVESC, Beijing, China, 2004), pp. 237–238
72. N.Y. Huang, J.C. She, J. Chen, S.Z. Deng, N.S. Xu, H. Bishop, S.E. Huq, L. Wang, D.Y.
Zhong, E.G. Wang, D.M. Chen, Phys. Rev. Lett. 93(7), 075501-1 (2004)
73. T. Shiroishi, T. Sawada, A. Hosono, S. Nakata, Low Temperature Growth of Carbon
Nanotubes by Thermal CVD with FeZrN Catalyst (Techn Didgest JVMC, Osaka, Japan,
2003), pp. 13–14
74. A.N. Rodokin, L.V. Malarevich, Receiving carbon nanofibre and nanotubes by method of
ultrahigh heating of vapors of ethanol. Inorg. Mat. 39(4), 433–437 (2003)
75. A.M. Lamanov, R.M. Ibragimov, K.N. Nikolskiy, A.N. Redkin, R.G. Chesov, E.P. Sheshin,
The autoissue cathodes manufactured by method of low-temperature gas sedimentation of
ethanol pa. Nano Microsyst. Equip. 9, 34–37 (2005)
76. A.M. Lamanov, E.P. Sheshin, A.N. Redkin, New Technique of Field Emission Cathodes
Preparation by Low Temperature Deposition From Ethanol Vapor. Hydrogen Materials and
Chemistry of Carbon Nanomaterials (Springer, 2007), pp. 265–268
77. Z. Sun, Y.J. Li, G.Y. Chen, S.P. Lau, B.K. Tay, J.S. Chen, L.K. Chean, Fabrication of
carbon nanotubes film arrays film arrays for field emission flat panel display application.
Surf. Rev. Lett. 8(5), 505–508 (2001)
78. J.I. Sohn, S. Lee, Y.-H. Song, S.-Y. Choi, K.I. Cho, K.S. Nam, Patterned selective growth of
carbon nanotubes and large field emission from vertically well-aligned carbon nanotubes
field emitter arrays. Appl. Phys. Lett. 78(7), 901–903 (2001)
79. G.S. Choi, K.H. Son, D.J. Kim, Fabrication of high performance carbon nanotubes field
emitters. Microelectron. Eng. 66, 206–212 (2003)
80. G.S. Choi, S.S. Cho, S.Y. Hong, J.B. Park, K.H. Son, D.J. Kim, J. Appl. Phys. 91, 3847
(2002)
81. Y.T. Jang, C.H. Choi, B.K. Ju, J.H. Ahn, Suppression of leakage current via formation of a
sidewall protector in the microgated carbon nanotubes emitter. Nanotechnology 14, 497–500
(2003)
82. Y.H. Lee, Y.T. Jang, D.J. Kim, J.H. Ahn, B.K. Ju, Realization of gated field emitter for
electrophotonic applications using carbon nanotubes line emitters directly grown into
submicrometer holes. Avd. Mater. 13(7), 479–482 (2001)
83. S. Kang, C. Bae, W. Son, M. Kim, J. Yi, A. Chang, J. Kim, C. Lee, Low Temperature
Carbon Nanotubes for Field Emission Display (Techn Didgest JVMC, Osaka, Japan, 2003),
pp. 51–52
84. C. Lou, X. Zhang, W. Lei, C. Qi, New method to fabricate field-emission cathode of carbon
nanotubes. Appl. Surf. Sci. 251, 254–257 (2005)
85. J.W. Moon, Electron emission source composition for flat panel display and method of
producing electron emission source for flat panel display using the same. Pat. US
2006/0066200 of 30.03.2006. Cl. 313/311 (H01j 1/14)
86. T. Honda, W. Rochanachirapar, K. Murakami, K. Ohsumi, N. Shimizu, S. Abo, F. Wakaya,
M. Takai, KrF Laser Surface Treatment of CNT Cathode (Tech Digest IVNC, Oxford, UK,
2005), pp. 300–301
87. X.H. Liang, N.S. Xu, I. Chen, S.Z. Deng, Enhanced Field Emission from Carbon Nanotube
Films by Hydrogen Glow Discharge Treatment (Tech Didgest IVNC, Oxford, UK, 2005),
pp. 302–303
88. J. Robertson, Diamond Relat. Mater. 3, 361 (1994)
89. J. Robertson, Amorphous carbon cathodes for field emission display. Thin Solid Films 296,
61–65 (1997)
90. V.J. Merkylov, D.H. Lowndes, L.R. Baylor and others, An addressable field emission array
for E-beam lithography using planar, pulsed-laser deposited amorphous diamond cathodes.
IVMS 178–179 (1998)
91. X. Zhang, Z. Lu, B. Zhang, N. Yao, B. Ma, Y. Zhao, Preparation of Nanostructure
Amorphous Carbon Film and Its Field Emission Properties (Techn Digest IVNC, 2004),
pp. 293–295
92. S. Bae, K.H. Park, S. Lee, K.H. Koh, Triode Emitters with Well-Structure Cathode (Techn
Diest IVNC, 2004), pp. 26–27
93. K.H. Park, K.M. Lee, S. Choi, S. Lee, K.H. Koh, J. Vac. Sci. Tehnol. 19, 946 (2001)
94. K.H. Park, W.J. Seo, S. Lee, K.H. Koh, Appl. Phys. Lett. 81, 358 (2002)
95. W. Knapp, D. Schleussner, Field emission characteristics of carbon buchypaper. J. Vac. Sci.
Tehnol. 21(1), 557–561 (2003)
96. W.J. Zhao, R. Wasu, M. Takai, Field Emission from Carbon Nanotube Material (Techn
Digest IVNC, 2003), pp. 57–58
97. W. Weibiao, X. Yuxue, Z. Chuanping, Electron Emission of Carbon Nanonets (Techn
Digest IVNC, 2003), pp. 155–156
98. H. Busta, Z. Tolt, J. Montgomery, A. Feinerman, Field Emission from Teepee-Shaped
Carbon Nanotube Bundles (Techn Digest IVNC, 2004), pp. 30–31
99. B.C. Halloway, M. Zhu, X. Zhao, J. Wang, R. Outlaw, Milliamp-Class Field Emission
Devices Based on Free-Standing, Two-Dimensional Carbon Nanostructures (Techn
Digest IVNC, 2005), pp. 24–25
100. V. Gulyaev Yu, V.N. Koro, A. Grigorev Yu, G.A. Rehn, Research of the Field Emission of
Fractal Blads Patterns of Large Length (Techn Digest IVNC, 2004), pp. 78–80
101. A. Kastalsky, S. Shokhor, J. Hou, S. Naar, N. Abianshin, B. Gorfinkel, Thin Film
Edge-Emitter Field Emission Flat Panel Display (Techn Digest of SiD, 2001), pp. 201–203
102. S.K. Gordeev, A.M. Lamanov, R.M. Jbragimov, K.N. Nikolskiy, E.P. Sheshin, Pyrolitic
Carbon Cathodes Prepared by Low Temperature Vapor Deposition (Techn Digest IVNC,
2005), pp. 176–177
References 365
103. S.K. Gordeev, S.B. Korchagina, A.M. Lamanov, E.P. Sheshin and other, New technique of
production of cathodes from a graphite foil. Nano Microsyst. тexникa. 12, 33–36 (2005)
104. V.P. Vereyko, M.M. Libenson, A.M. Miluchev, Laser technology. Electronics 68(137), 114
(1970)
105. M.F. Stelmakh, Lasers in technology under the editorship. Energy (1975)
106. V. Polezhaev Yu, F.B. Yurivech, Thermal Protection (1975)
107. K.N. Nikolskiy, A.C. Batyrin, A.C. Kyprashkin, R.G. Chesov, E.P. Sheshin, “Effect of a
ring” at intensive field electronic issue. Microsyst. Equip. 7, 8–10 (2003)
108. K.N. Nikolskiy, A.C. Batyrin, A.I. Knazev, R.G. Chesov, E.P. Sheshin, Formation of rings
round primary field emission image and possibility of their practical use. Mag. Tech. Phys.
74(2), 110–113 (2004)
109. Y.J. Li, S.P. Lau et al., Oriented carbon microfibers grown by catalytic decomposition of
acetylene and their field emission properties. Diamond Relat. Mat. 10(1), 878–883 (2001)
110. L.C. Jin, P. Jeunghee, Growth and field electron emission of vertically aligned multiwalled
carbon nanotubes. Chem. Phys. Latters 326(1), 175–181 (2000)
111. Q.H. Wang, R.P.H. Chang, M. Yam, Flat panel display prototype using gated carbon
nanotube field emitters. Appl. Phys. Lett. 78(9), 1294–1297 (2001)
112. Z. Lan, Z. Binglin et al., A Flat Panel Display Device Fabricated by Using Carbon
Nanotubes Cathode. IEEE, IVMC 2001 University of California, Davis, California, USA,
12–16 August 2001
113. Z. Wei, G. Kochanski et al., Large current density from carbon nanotube field emitters.
Appl. Phys. Lett. 75(6), 873–876 (1999)
114. S. Gleston, Introduction to Electrochemistry (M.: Foreign Literature, 1951)
115. E.M. Lifshic, L.P. Pitaevskiy, Theoretical Physics of T. 10. Physical Kinetics (M.; Science,
1979)
116. K. Fetter, Electrochemical Kinetics (M.: Chemistry, 1967)
117. G.A. Golikov, Guide to Physical Chemistry (M.: The Highest Shkola, 1988)
118. D.S. Dykin, B.V. Deragin, Electrophoresis (M.; Science, 1976)
119. D.A. Kurosov, A.S. Baturin, K.N. Nikolskiy, R.G. Tcheov, E.P. Sheshin, Influence of the
interelectrode distance in electrophoretic cold cathode fabrication of the emission uniformity.
Appl. Surf. Sci. 215, 232–236 (2003)
120. W.B. Choi, J.J. Cuomo, Field emission from silicon and molybdenum tips coated with
diamond powder by dielectrophoresis. Appl. Phys. Lett. 68(5) (1996)
121. V.V. Zhirnov et al., Diamond coated Si and Mo field emitters: diamond thickness effect.
Appl. Surf. Sci. 94(95), 123–128 (1996)
122. A.N. Alimova, N.N. Chubin et al., Electrophoresis of nanodiamond powder for cold cathode
fabrication. J. Vac. Sci. Technol. B. 17(2) (1999)
123. A.F. Pal, N.V. Suetin et al., Emission Properties of Electrophoretically Deposited
Nanodiamond Film. IEEE, IVMC 2001 Davis, California USA, 12–16 August 2001
124. Yoshikazu Nakayama, Seiji Akita, Field-emission device with carbon nanotubes for a flat
panel display. Synth. Met. 117, 207–210 (2001)
125. A.N. Obraztsova, A.P. Volkov et al., Single-Wall Carbon Nanotube Electron Emitters
Formed by Liquid Electron Emitters Formed by Liquid Electrophoresis. ITM-FEECM 2001,
Moscow, Russia, 2–4 July 2001
126. W.B. Choi, Y.W. Jin et al., Electrophoresis deposition of carbon nanotubes for triode-type
field, emission display. Appl. Phys. Lett. 78(11), 12 (2001)
127. H. Zhao, H. Song, Z. Li, G. Yuan, Y. Jin, Electrophoretic deposition and field emission
properties of patterned carbon nanotubes. Appl. Surf. 251, 242–244 (2005)
128. S.J. Oh, J. Zhang, Y. Cheng, H. Shimoda, O. Zhou, Liqid-phase fabrication of patterned
carbon nanotube field emission cathode. Appl. Phys. Lett. 84(19), 3738–3740 (2004)
129. E.P. Sheshin, A.S. Baturin, K.N. Nikolskiy, R.G. Tchosov, V.B. Sharov, Field emission
cathodes based on milled carbon fibers. Appl. Surf. Sci. 251, 196–200 (2005)
130. A.N. Alimova, N.N. Chubin, P.J. Belobrov, V.V. Zhirnov, Electrophoreses of nanodiamond
powder for cold cathode fabrication. Proc. IVMC 222–223 (1998)
131. A.S. Batyrin, D.A. Kyrnosov, K.N. Nikolskiy, E.P. Sheshin, R.G. Chesov, Way of
production of the autoissue cathode. Stalemate. The Russian Federation No. 2225052 from
2002. Cl. H01j1/14
132. E.P. Sheshin, A.S. Baturin, K.N. Nikolskiy, R.G. Tchosov, in Field Emission Cathodes
Based on Milled Carbon Fibers. Proceedings of IVESC, Beijing, China, pp. 114–116 (2004)
133. J.-W. Nam, J.H. Choi, J.-H. Han, J.-B. Yoo, C.-Y. Park, Enhanced Emission from Carbon
Nanotube Paste After Firing (Techn Digest IVNC, Osaka, Japan, 2003), pp. 225–226
134. H. Ren, S.Z. Deng, J. Chen, J.C. She, N.S. Xu, Plasma Etching Treatment for Improving the
Field Emission Properties of Carbon Nanotubes Composite Emitters (Techn Digest IVNC,
Osaka, Japan, 2003), pp. 261–262
135. J.W. Nam, S.H. Cho, S.H. Choi, Y.C. Choi, J.S. Ha, J.H. Park, D.H. Choe, J.B. Yoo, Long
Lifetime of Field Emitters Fabricated Using Carbon Nanotube Paste (Techn Digest of
IVNC, Oxford, UK, 2005), pp. 292–293
136. L.-T.A. Cheng, D.H. Roach, Electron field emitter and composition related thereto.
Stalemate. USA 2004/0017141. 2004 Yu CL. H01j19/06
137. A. Boda, Y. Jshida, K. Hakiai, T. Asano, Fabrication of Gated Carbon Black Field Electron
Emitter Using Inkjet (Techn Digest of IVNC, Oxford, UK, 2005), pp. 42–43
138. B.V. Bondarenko, V.I. Makyha, E.P. Sheshin, Autoelectronic emitters with the developed
working surface. Electron. Equip. (It is gray. 1: Elektronika Microwave Oven) 10, 44–47
(1984)
139. B.V. Bondarenko, V.I. Makyha, L. Rybakov Yu, V.B. Sharov, E.P. Sheshin, Vlyany
roughnesses of a surface of autocathodes on their issue characteristics. Radio Eng.
Elektronika 32(12), 2606–2610 (1987)
140. B.V. Bondarenko, V.I. Makyha, A. Rybakov Yu, E.P. Sheshin, Field Electronic Emission of
Rod Graphite Cathodes. Physical Phenomena in Devices of Electronic and Laser Equipment
(M.: MFTI, 1981), pp. 11–15
141. A.L. Syvorov, E.P. Sheshin, V.V. Protasenko and other, The microrough flat autoissue
cathodes from graphite received in the radiftsionny way. ZhTF 66(7), 156–160 (1996)
142. A.G. Chakhovskoi, C.B. Hunt, Improved Image Uniformity in Light Sources with Carbon
Field Emitters (Techn Digest Lith JVMC, NC, USA, 1998), pp. 190–191
143. A.S. Baturin, E.P. Sheshin, Field Emission Characteristics of Reticulated Vitreous Carbon
(Tecchn Digest of IVNC, Guangzhou, China, 2000), pp. 42–43
144. C.E. Hunt, A.G. Chakhovskoi, Field emission cathode fabricated from porous carbon foam
material. Patent WO 99/43870 of 2.09.99. Cl. C25B3/00. H01j1/30
145. P.G. Gabdyllin, S.N. Davydov, V.V. Korablev, A.E. Kravchik, A. Kykyshkina Yu, V.V.
Sokolov, V.S. Travnikov, Cold emitter of electrons. Stalemate. Russian Federation
No. 2249876 of 6.11.2003. Cl. H01j1/30
146. M. Leshukov Yu, E.P. Sheshin, Plasmachemical Processing of Field Emission Cathodes
Made of Carbon Fibers Bundles (Techn Digest of IVNC, Guilin, China, 2006), pp. 333–334
147. A.S. Leychenko, M. Leshykov Yu, N.V. Lyparev, E.P. Sheshin, in Formation of the
Emitting Surface of Autocathodes from Bunches Carbon вoлoкн the Crown Category. Sb.
Tezisov of dokl. mezhdunarodny Conference “Carbon: Fundamental Problems of Science,
Materials Science, Ekhnologiya”, Moscow, p. 115 (2006)
148. L.B. Loeb, Electrical Coronas (University of California Press, 1965), p. 760
149. R. Morrow, Theory of negative corona in oxygen. Phys. Rev. A. 32, 1799–1809 (1985)
150. A. Cao, X. Zhang, C. Xu, J. Liang, D. Wu, B. Wei, Thinning and diluting aligned carbon
nanotube film for uniform field emission. Appl. Phys. A. 74, pp. 415–418 (2002)
151. V.V. Zhirnov, O.M. Kuttel, O. Gruning, A.N. Alimova, P.Y. Detkov, P.J. Belobrov, E.
Mailadr-Schaller, L. Schlapbach, Characterization of field emission cathodes with different
forms of diamond coatings. J. Vac. Sci. Thechnol. B. 17(2), 666–669 (1999)
152. V. Gilyaev Yu, N.J. Sinitsyn, G.V. Torgashov, S.G. Saveliev, Fundamental and applied
directions of field emission electronics using nanicluster carbon materials. Appl. Surf. Sci.
215, 141–148 (2003)
References 367
153. Y. Choi, W.W. Lee, B.-H. Ryu, Application of Carbon Nanofiber Coated with Pt
Nanoparticles to FED (Techn Digest of Euro Display, Moscow, 2007), pp. 318–320
154. R.B. Sharma, D.J. Late, D.S. Joag, A. Govindaraj Rao, CNR Field Emission Studies of
Boron and Nitrogen Doped Carbon on Pointed and Flat Substrates (Techn Digest of IVNC,
Oxford, UK, 2005), pp. 310–311
155. V.D. Yu, J.H. Zhang, X. Wang, X.M. Li, X.D. Gao, Characterization and field emission
property of aligned porous carbon nanotube film by hydrogen-ion implantation. Appl. Phys.
A. 81, 169–172 (2005)
156. A.C. Batyrin, K.N. Nikolskiy, A.I. Knazev, R.G. Chesov, E.P. Sheshin, Introduction of
shchelochnozemelny metal in structure of graphite for the purpose of decrease in work of an
exit. Mag. Tech. Fiziki 74(3), 62–64 (2004)
157. A.C. Batyrin, A.G. Borisov, V.S. Korsakov, L.A. Plavich, I. Tishin Yu, E.P. Sheshin, Way
of drawing an issue covering on the cathode. The Stalemate of the Russian Federation
No. 2177657. Cl. H01j1/14 of 22.01.2001
158. K.N. Nikolsky, A.S. Batyrin, V.S. Bormashov, E.P. Sheshin, D.V. Shyr, S. Zaginaychenko
Yu, V.E. Myradan and other, Field Emission Researches Dopirovannykh of Carbon
Nanotubes (Techn Didgest of Cont «Hydrogen Materials Science and Chemistry of Carbon
Nanomaterials», Sudak-Crimea, Ukraine, 2003), pp. 824–825
159. P.V. Sherstnev, A.S. Baturin, V.S. Bormashov, K.N. Nikolskiy, E.P. Sheshin, Investigation
of local work function of field emission of field emission cathode from carbon materials
doped by Ba with the help of AFM/SEM/X-ray. Phys. Chem. Solids 62, 159–163 (2004)
160. A. Bobkov, E. Davidov, S. Zaitsev, S. Karpov, M.A. Kozodaev, I.N. Nikolaeva, M.O.
Popov, E.N. Skorokhodov, A.L. Suvorov, N. Cheblukov Yu, J. Vac. Sci. Technol. B 19, 32
(2000)
161. V.S. Bormashov, A.S. Baturin, K.N. Nikolskiy, E.P. Sheshin, Tech Digest of 15th IVMC
and 48th IFES, vol. 2, no. 2, 64 (2002)
162. R.G. Forbes, J. Vac. Sci. Technol. B 17, 534 (1999)
163. R.G. Forbes, J. Vac. Sci. Technol. B 17, 526 (1999)
164. E.L. Murphy, R.H. Good, Phys. Rev. 102, 1464 (1956)
165. J.-M. Bonard, J.-P. Salvetat, T. Stockli, L. Forro, A. Chatelain, Appl. Phys. Lett. 69, 245
(1999)
166. W. Zhu, G. Koshanski, J. Sungho, C. Bower, O. Zhou, Appl. Phys. Lett. 75, 873 (1999)
167. V.S. Bormashov, A.S. Baturin, K.N. Nikolsky, E.P. Sheshin, Appl. Surf. Sci. 215, 178–184
(2003)
168. V.S. Bormashov, A.S. Batyrin, Materials of the international conference «Educational,
scientific and engineering applications in the environment LabVIEW and technologies
National Instruments» Moscow, p. 130 (2003)
Chapter 7
Computation of Field-Emission
Cathode-Based Electron Guns
Abstract This chapter proposes computation models of cathode-based electron

guns along with assumptions that could be helpful for simplifying the computation.
General problem of computation of the optimum structure of electron guns based on
one-apex and multi-apex field-emission cathodes (the Almazov–Egorov model) is
posed and solved. Then mathematical simulation of model triode electron-optical
systems and methods of calculation of distribution of the electric field in forming
and controlling systems based on a field-emission cathode and a small-aperture
focusing diaphragm system are presented. The chapter presents a technique for
calculating the optimum characteristics of the forming and controlling systems
based on a field-emission cathode and a focusing diaphragm system and proceeds to
proposing a computation model of electron trajectories in a field-emission
cathode-based system is proposed.
7.1 Introduction
Practical application of charged-particle beams involves having not only an emitter,

i.e. the element generating an electron beam, but also the elements able to preserve
and, ideally, improve the advantages and neutralize the disadvantages of
charged-particle sources. An electron emitter cannot work by itself because even
creation of conditions initiating field emission or thermal-field ionic emission needs
not only an emitter, i.e. a cathode, but also the second electrode referred to as the
anode, the extractor etc. depending on the destination. Moreover, the second
electrode provides a possibility of transportation of charged particles from their
source. The system of electrodes creating the conditions necessary to excite,
transport, focus and control a charged-particle beam is referred to as a gun.
A system of two continuous electrodes, a cathode and an anode, is an elementary
but analyzable model of the gun. However, it also has a practical importance of its
own, e.g., in development of double-electrode electrostatic accelerators. In more
complex designs, the second electrode normally has a beam port, i.e. it is an
immersion lens/diaphragm. A necessity to at least maintain such emitter

370 7 Computation of Field-Emission Cathode-Based Electron Guns
characteristics as (a) the maximum current density, (b) emission homogeneity,

(c) ability of the emitter to function in a particular medium, (d) the cathode lifetime
under the given working conditions, (e) luminosity, (f) perveance, (g) emittance etc.
can be implemented with a more complex electrode system consisting of an
immersion-lens system, sometimes being a combination of immersion and quad-
rupole lenses. Hence, the gun has to be supplemented with some other devices
making it possible to bring up the beam to the object or the information
receiver/detector analyzing the beam. So the gun is not the only structural element
of the beaming system, the latter being broadly understood as a combination of
electric and magnetic fields, as well as the electrodes and magnetic circuits forming
them and necessary for the electron beams to be properly configured.
Simplifying assumptions. A real beam is a complex physical object. The
mathematical description of properties of any somehow complex object is impos-
sible without particular idealizing assumptions, i.e. introduction of a particular
simplifying model that provides the basic essential properties of the object.
Naturally, the calculation data turn out approximate, the approximation degree
depending on how much the model selected manifests these central essential
properties. An electron or ion beam is a set of charged particles; however, it is
almost impossible to provide scrupulous accounting of discreteness of electric
charges during their interaction. That is why computations of heavy flows assume
that the electrons act upon an individual electron with a force equal to that exerted
by an electrically-charged medium with continuously-distributed charge density.
Another frequent model provision is beam laminarity, i.e. we assume that the
electron paths never cross, which considerably simplifies the problem due to a
possibility of unambiguous determination of the electron velocity at each point. An
important simplifying provision applied to solve almost all problems is path
paraxiality. Paraxiality makes it possible to limit ourselves with the first terms of a
series when expanding the expressions for electric and magnetic fields near the
system axis and consider that the longitudinal velocity of a particle is fully deter-
mined by the potential at the point considered. That simplifies the form and solution
of the equations.
It is extremely important for effective functioning of the device to prevent beam
particles from getting to the forming electrodes. The greatest current-deposition
danger comes from the particles located on the periphery of the beam. That is why
one of the primary goals of beam physics is determination of the beam
contour/profile. Here it seems useful to make an assumption about the concentration
of extreme/boundary electron/ion. The matter is that the beam contour is deter-
mined by the extreme electron path. In our research, we resorted to all approxi-
mations mentioned. The necessary reservations about deviation from these
provisions or introduction of additional assumptions are made in respective places
of the text.
Mathematical models and requirements thereto. In this chapter, our original
results are used to solve some problems related to field emission cathode-based
electron-gun formation and control systems (FCS). It is also necessary to note that
continuously increasing requirements to the quality of designed products result in a
necessity to take account of increasingly more complex elements of simulated

objects and phenomena matching the structure of the objects and their interrela-
tions. With reference to mathematical models, it mostly relates to solution of
tridimensional problems. Though the publications proving urgency of these prob-
lems and successes of physical-field numerical-computation methods are abundant,
solution of tridimensional problems in a quite general formulation still remains a
challenge, especially when it is not a question of consideration of abstract issues,
but particular engineering-design problems. Clearly, the number of practically
working software developments of wide application is insufficient. The reason
thereof is not only high requirements to the computing systems, but also the
complexity of arrangement of the computing experiment.
The priority is to research the properties and configurations of electric and
magnetic fields providing the necessary beam configuration, focusing and trans-
portation. There are both experimental and computational methods to determine
electric fields that can serve as the basis to research the characteristics of
electron-optical systems.
The disadvantages of the experimental field-finding methods include bulkiness
of instrumentation and relatively low precision of measurement. Nowadays, the
onrush of computer technology has almost completely superseded them with
computational, i.e. analytical or numerical, methods [1–38]. The most commonly
applied analytical methods include those of (1) conformal mapping, (2) integral
equation/charge density and (3) variable separation.
The conformal mapping method is applied to compute the fields of
two-dimensional, cylindrical, quadrupole electron-optical systems. It is based on
forming a function of complex variable regular in the area considered and per-
forming its conformal mapping to an area with a known solution of the Laplace
equation for the problem considered. This method is difficult in application because
the constants included into the function performing the conformal mapping depend
on the geometrical parameters of the electron-optical systems and are to be found
when solving the transcendental-equation systems. The primal constraint of the
conformal mapping method is that a function performing conformal mapping to the
canonical area cannot be formed for every area considered. This method is used
both to solve the forming-electrode computational problems providing implemen-
tation of a properly-configured beam and to formulate and solve the inverse
problem, i.e. that of synthesis. In the synthesis problem, it is necessary to calculate
the electron-optical system providing implementation of the beam with some preset
properties, a necessary stage in solving the problem being straightening of the flow
boundary through the adequate selection of conformal mapping, e.g., with appli-
cation of Schwarz-Christoffel transformations.
The integral-equation method is based on space representation of a potential as
an integral over the electrode-system surface with a known kernel and an unknown
density function as integrands, i.e. the potentials of an ordinary/double layer. The
equipotential surfaces of the electrodes are replaced with surface-bound charges
distributed with some density. As the system-boundary potential is known, the
density function is found from the equality of the integral to the potential in the
boundary point system. The first stage of the method is to find the charge distri-
bution over the electrodes, the second being determination of the potential in the
entire space. The basic difficulty of this method is caused by a necessity to solve the
integrated equation linking the boundary potentials with the surface charge-density
distribution. The analytical solution of such integrated equations being impossible
for most lenses, they are replaced with linear algebraic equation systems, which
makes it possible to get an approximate solution. There are also some other surface
charge-density approximation methods. Recently, wide application of the
integral-equation method is caused by a possibility to use high-speed high-volume
computers providing adequate accuracy within quite a short period of time.
The variable-separation method is among the general methods to solve equations
in partial differential coefficients. This method can only find potential distribution
for quite simple boundary conditions when the electrode surfaces belong to coor-
dinate planes, e.g., cylinders or diaphragms, which is an important constraint. If an
adequate coordinate system happens to be selected, the Laplace equation is rela-
tively easy to solve. In this case, it functions as a harmonic-polynomial potential
approximation method.
Numerous electron- and ion-optical devices employ the focusing lens system not
to form an electron-optical image, but make the greatest possible number of par-
ticles escaping from the source at different angles concentrate on the detector. In
particular, examination of a substance surface involves collection of the charged
particles escaping from the surface at as wide an angle as possible, their conduction
along the path including a mass spectrometer and registration with an information
receiver/detector, thus bringing up an issue of functioning of an electron-optical
system with essentially off-axis beams.
In electron-beam devices, electron flows are formed by coordinate-dependent
and time-independent, i.e. static, fields. The position of formed electron beams is
controlled by the fields varying in time. However, the variations within the con-
trolled beam electron transit time are not large, and these fields can be considered as
quasistatic. In paraxial optics, when the particle pathlength from the system axis can
be taken as a small parameter, electron-optical computations usually operate
axisymmetric electron-optical system fields set in a tabular way in a form of axial
distribution of the electric potential calculated by this or that method. At interme-
diate points, the values are usually calculated by the nearest neighbor interpolation
method, which can result in low accuracy of derivatisation of the second-and
higher-order potential by interpolation-polynomial differentiation.
Above-mentioned general methods to compute electron-optical systems are
characteristic for any system. However, computation of such systems encounters
specific difficulties as far as cathode lenses are concerned. It is inherent in cathode
lenses that the field starts forming the electronic image immediately at the cathode
surface. That is why solution of general electron-optical problems through com-
putation of cathodic systems needs special methods. First of all, they are
electrostatic-field computation methods. Each specific cathodic electron-optical
system is usually simplified to a model able to provide an approximate solution of
the problem. When computing the characteristics of pointed systems represented by
cathode lenses, e.g., field-emitter electron guns, the main difficulty encountered is
that the sizes of the emitter and the focusing electrodes of the system differ by a few
orders of magnitude.
As a rule, specific consideration is given to both the cathode area, the approx-
imating functions necessary to determine the field in the apex neighbourhood being
formed, and the focusing system, followed by combination of the solutions arrived
at. Computation of the field near the pointed cathodes involves approximate
replacement of their surfaces with paraboloids or hyperboloids of revolution,
extremely oblong ellipsoids, as well as equipotential surfaces created by a charged
conductive cone with a conductive sphere on its apex, i.e. sphere-on-cone.
Well-developed general methods of theoretical examination of electron-optical
systems have supposedly lost their generality because of a qualitative change of
up-to-date electron-optical problems. The change is accounted for by increase in the
role of the elements whose electron-optical theory cannot rely on the classical
paraxial approximation, e.g., an unconstrained increase in the charged-particle path
slope to the main optical axis of these systems in the cathode neighborhood area.
That is why complication and widening of a circle of electron-optical problems
make it necessary to develop new methods of accurate computation of the field in
electron-optical systems.
This chapter is dedicated to the systems of forming and controlling of electron
guns based on the electron beams generated by pointed field-emission cathodes.
The focus of attention is on the forming and controlling systems based on
immersion electrostatic lenses. The primary reason thereof is that the operational
experience in beam formation and transportation systems shows sufficient effec-
tiveness of these lenses and their superiority over magnetic lenses for the
beam-intensity ranges considered. Moreover, practical implementation of electro-
static lenses is much cheaper than that of magnetic ones.
At the same time, their computation, simulation and numerical experiment for
the systems of forming and controlling electron sources and guns based on
field-emission cathodes is much more difficult. The primary reason thereof is that
specific characteristic of field emission that emission is excited by an electric field,
and current density considerably depends on its intensity directly at the surface of
such cathode. The field can be generated if at least one more electrode in addition to
the cathode is available. Frequently, a charged-particle beam cannot be conditioned
with only one electrode, and the forming and controlling systems have to be sup-
plemented with some other electrodes that are basically able to focus and transport
the beam, as well as control it in an efficient way. Unlike hot cathode-based forming
and controlling systems and guns, even the slightest change of potentials at the
system electrodes based on field emission cathodes can cause drastic changes in the
emissive capacity of the latter. And the change in the emissive capacity of the
cathode causes reconditioning of the emitted-electron beam, which entails a change
in its focusing and transporting conditions. In other words, the problem of com-
putation of field-emission cathode-based guns has to be brought up with account of
the cross-impact of these cathodes and the additional electrodes both on the cathode
emissivity and the characteristics of the forming and controlling systems, which
presents severe difficulties. Apparently, it is these difficulties that account for lack
of essential successes in computation of guns and forming and controlling systems
based on field-emission cathodes.
7.2 General Problem of Computation of the Optimum

Structure of Electron Guns Based on One-tip
and Multi-tip Field Emission Cathodes, i.e.
the Almazov–Egorov Model [25, 27, 28, 39–41]
Let us consider the emission system consisting of a field-emission cathode repre-

sented by an axially symmetrical apex of an arbitrary shape and a system of
electrodes, viz. an anode, a grid etc. (Fig. 7.1).
In such a system, potential U(r, z) meets the Laplace equation of
DU ¼ 0
and the boundary conditions of
UjX ¼ w0 ; ~
UjX~ ¼ w; ð7:1Þ
where Ω is the cathode-apex surface; w0 is the potential at the apex; UjX~ is the
anode surface; w ~ is the potential at the anode. The most important characteristic of
such systems is intensity of field E = –grad U at the apex peak. This value defies
detailed analysis because it depends on the geometrical dimensions of the apex. In
practice, it is thin apices that are most interesting. As a result, there is a small
parameter arising in the problem considered, viz. a ratio of the apex thickness to its
typical longitudinal dimension. To study such apices, it is enough to expand the
field intensity at the apex peak (E0) in an asymptotic series in the vicinity of the
ideally thin apex. A few first terms of this series would enable calculation of the
Fig. 7.1 Model of axially

symmetrical emission system. r
Ω—field emission cathode, ~
~
X—anode Ω ρ(z) Ω
0 z
-L -δ
7.2 General Problem of Computation of the Optimum Structure of Electron Guns … 375
field intensity in a quite precise way. To construct such asymptotic expansion, let us
apply the following approach. Let us represent the sought-for solution as
Uðr; zÞ ¼ U0 ðr; zÞ þ U1 ðr; zÞ þ U2 ðr; zÞ; ð7:2Þ
where r, z are cylindrical coordinates;
ZL
qðz0 Þ
U0 ðr; zÞ ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi dz0 ; ð7:3Þ
d r 2 þ ðz z0 Þ2
functions U1 and U2 meet the following boundary problems, viz.
DU1 ¼ 0; ~
U1 jX~ ¼ w; ð7:4aÞ
DU2 ¼ 0; U2 jX~ ¼ U0 jX~ ; ð7:4bÞ
while function U2 can be represented as
ZL
U2 ðr; zÞ ¼ u2 ðr; z; z0 Þqðz0 Þdz0 ; ð7:5Þ
d
where function u2(r, z; z′) is defined through the Green function of problem (7.3),
viz.
h i1=2
u2 ðr; z; z0 Þ ¼ 4pGðr; z; 0; z0 Þ r 2 + ðz z0 Þ2 ; ð7:6Þ
q(z) being the charge density at the apex axis.

It is obvious that expression (7.2) meets the Laplace equation and boundary
conditions. Our aim is to select charge distribution q(z) in (7.3) for the equipotential
surface of potential U with potential value w0 to coincide with the apex surface.
Having selected q(z) in this way, we will be able to solve starting boundary problem
(7.1) and define value E and emission area Sem we are interested in. As we are only
interested in thin apices, it is enough to study the equipotential surfaces of model
potential (7.2) located near the apex axis. Here let us resort to asymptotic expansion
of the potential near the apex axis, viz.
ZL
2ðL zÞ qðnÞ qðzÞ
Uj ¼ qðzÞ ln qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi þ dn
r!0 j n zj
d z þ r 2 þ ðd zÞ2 d ð7:7Þ
d\z\L
þ U1 ð0; zÞ þ U2 ð0; zÞ:
Taking into account (7.7), the equipotential surface close to the apex axis fol-
lows the ratio of
ZL
2ðL zÞ qðnÞ qðzÞ
qðzÞ ln qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi þ dn
j n zj ð7:8Þ
d z þ r 2 þ ðd zÞ2 d
+ U1 ð0; zÞ þ U2 ð0; zÞ ¼ w0 :
where r0 is the equipotential surface section radius taken as section height function
z. As surface shape r0(z) is known, ratio (7.8) is a charge-density determination
equation. Equation (7.8) includes value d, which is still unknown. Let us define it
from the condition that U(0, 0) = w0, i.e. we believe that apex peak is at the origin
of coordinates. Let us underscore that value d for a thin apex should be small. In the
latter case, condition U(0, 0) = w0 can be represented as
ZL
L qðnÞ qðdÞ
w0 ¼ qðdÞ ln þ dn þ U1 ð0; 0Þ þ U2 ð0; 0Þ: ð7:9Þ
d n
d
In addition, if we put z = d in (7.8), we get
ZL
2ðL zÞ qðnÞ qðdÞ
w0 ¼ qðdÞ ln þ dn þ U1 ð0; dÞ þ U2 ð0; dÞ: ð7:10Þ
r0 ðdÞ nd
d
Comparison of (7.9) and (7.10) demonstrates that equation ln(L/d) = ln2(L − d)

[r0(d)]−1 should be true within the accuracy up to the terms vanishing at d ! 0,
wherefrom we can obtain the equation for determination of value d, viz.
2d ¼ r0 ðdÞ: ð7:11Þ
Let us pay attention to the fact that function r20 for any apex shape can be
represented as

r02 ðzÞ ¼ 2R0 z + 0 r 2 ;
where R0 is the apex-peak curvature radius. Let us assume that the apex peak is
smooth enough. Moreover, we will consider the condition of
r02 ðdÞ ¼ 2R0 ½1 þ 0ðdÞ; ð7:12Þ

as met. With its account, (7.11) can give us an expression to calculate value d, viz.
d ¼ R0 =2: ð7:13Þ
This ratio can be also obtained in a different way, viz. it is necessary to calculate
the equipotential surface curvature radius going through the origin of coordinates
within the range of d ! 0 with involvement of (7.2), (7.3). Using the formula
1
R0 ¼ Uz0 ð0; 0Þ Urr00 ð0; 0Þ
for the equipotential surface curvature radius, let us transform the (7.13) as 2d = R0.
If condition (7.12) is met, (7.8) can be somehow reduced. Let us underscore that
if z d.
2
2ðL zÞ 4zðL zÞ r0 ðzÞ d
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ¼ 1 þ 0 þ 0 ;
d z þ r02 ðzÞ þ ðd zÞ2 r0 ðzÞ
2 z 2 z
and if z and d are of the same order, both expressions, viz.

qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi1
2ðL zÞ d z þ r02 ðzÞ þ ðd zÞ2 ; 4zðL zÞr02 ðzÞ
coincide in the higher order at d ! 0 and are about 2LR1

0 . As a result, (7.8) can be
reformulated as
ZL ZL
4zðL zÞ qðnÞ qðzÞ
qðzÞ ln þ dn þ u2 ð0; z; nÞqðnÞdn ¼ w0 U1 ð0; zÞ:
r02 ðzÞ j n zj
d d
ð7:14Þ
Using (7.14), let us construct the desired expansion for value q(z). To this effect,
it is enough to seek solution of equation (7.14) as an iterative series. Iterating (7.14),
we can obtain the solution in the form of
X
1
qðzÞ ¼ qk ðzÞ; ð7:15Þ
k¼0
w0 U1 ð0; zÞ
q0 ðzÞ ¼ ; ð7:16Þ
ln½4zðL zÞr02 ðzÞ
2 3
1 ZL ZL
4zðL zÞ 4 qk ðnÞ qk ðzÞ dn þ
qk þ 1 ðzÞ ¼ ln u2 ð0; z; nÞqk ðnÞdn5:
r02 ðzÞ jn zj
d d
ð7:17Þ
In fact, the iterative series seems to be asymptotic for thin apices.

Knowing charge density q(z), it is not difficult to calculate the field intensity at
the apex peak. Using (7.2) and (7.3), we can express E0 as
ZL
@U qðnÞ @

E0 ¼ ¼ dn þ ðU 1 þ U Þ
2 : ð7:18Þ
@z z¼0 n 2 @z z¼0
r¼0 d r¼0
Obviously, the main contribution into expression (7.18) for thin apices is given
by the first term. Calculating the integral asymptotically at d ! 0, we find the field
intensity as
E0 ¼ d1 qðdÞ ¼ 2R1

0 qðdÞ: ð7:19Þ
Adjustments to expression (7.19) can be shown to vanish at d ! 0. Inserting

(7.16) and (7.17) into (7.19), we obtain the desired expansion for the field intensity
at the apex peak.
The approach suggested enables determination of emission area Sem. With this
end in view, let us compute the total current, viz.
Z Z
B
I¼ dSjðxÞ ¼ dSAE2 ðxÞ exp ;
EðxÞ
ð7:20Þ
1:54 106
A¼ ; B ¼ 6:83 107 m3=2 #ðyÞ:
mt2 ðyÞ
In (7.20), the integral is calculated over the apex surface, viz. E(x) is the field
intensity at the apex surface at point x. At B E0 integral (7.20) can be calculated
asymptotically by the Laplace method, viz.

2pAE04 B
I¼ 00 ð0Þ
exp :
BErr E0
That is the basis for the emission-area expression, viz.
2pAE02
Sem ¼ IJ01 ¼ 00 ð0Þ
:
BErr
00
Using model potential (7.2), it is not difficult to calculate the value of Err ð0Þ:
However, the respective expression looks cumbersome. It is essentially reduced at
d ! 0, i.e. for a thin apex, and taking into account (7.13) and (7.19)
Sem ¼ 2pR20 E0 =B: ð7:20aÞ
Let us note that the linear dependence of the emission area on the field intensity
was stated in a number of papers, e.g. [42, 43]. However, they did not determine the
coefficient of proportionality for that dependence. Processing the results of the
experiments aimed at determination of field intensity at the apex peak often
involves the Homer formula, viz. E0 = U0/(kR0), where U0 is the potential differ-
ence in the anode–cathode space; R0 is the apex-peak curvature radius, and k is the
geometrical factor amounting to about 5. Application of formulas (7.19) and (7.15)–
(7.17) can make a more accurate calculation of this form factor. For a thin-enough
apex, let us use the first expansion term (7.15). Having inserted (7.16) into (7.19),
we obtain
w0 U1 ð0; 0Þ
E0 ¼ ; ð7:21Þ
kR0
where
1 2L
k¼ ln : ð7:22Þ
2 R0
However, expression (7.22) gives a wrong form-factor value at large L. Actually,

it follows from (7.21) and (7.22) that L ! ∞, E ! 0, though it is obvious that the
range of L ! ∞ should have E ! const 6¼ 0. So construction of a right expression
for value k needs involvement of two terms of expansion (7.15). Having inserted
q0(z) + q1(z) in (7.19) and represented E0 as (7.21), we obtain
2 3
ZL 1
1 2L 4 1 q0 ðnÞ q0 ðdÞ
k ¼ ln 1 dn þ u2 ð0; 0; nÞq0 ðnÞ 5 :
2 R0 w0 U1 ð0; 0Þ n
d
ð7:23Þ
For a thin apex, the integral in square brackets is small because the q0(n) value is
small. Therefore, (7.23) can be represented as
2 3
ZL
1 2L 4 1 q0 ðnÞ q0 ðdÞ
k ¼ ln 1þ dn þ u2 ð0; 0; nÞq0 ðnÞ 5 þ :
2 R0 w0 U1 ð0; 0Þ n
d
ð7:24Þ
Using (7.16), we can transform (7.24) into
ZL

1 2L 4nðL nÞ 1 w0 U1 ð0; nÞ 1
k ¼ ln dn ln þ u 2 ð0; 0; nÞ : ð7:25Þ
2 R0 r02 ðnÞ w0 U1 ð0; 0Þ n
d
As an example, let us consider a system consisting of an apex shaped as a thin

revolution ellipsoid and an anode represented by an infinite plane located at dis-
tance D from the apex peak. Let the anode potential be zero. In this case,
4nðL nÞ 2L
ln ¼ ln ; U1 ð0; nÞ ¼ 0; u2 ð0; 0; nÞ ¼ ð2D þ nÞ1 :
r02 ðnÞ R0
The integral of (7.25) can be calculated accurately. For the form factor, we
obtain
1 4LD
k ¼ ln : ð7:26Þ
2 R0 ð2D þ LÞ
At L D, expression (7.26) turns into (7.22). At L D, the apex form-factor

is shaped as a revolution hyperboloid.
The above procedure to calculate characteristics can also be applied to
multi-apex periodic systems. Towards this end, model potential U has to be taken as
Zhd X 1
U ¼ E1 z dnqðnÞ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi;
h þ d
k jr ak j þ ðz nÞ2
2
qðzÞ ¼ qðzÞ;
where E∞ is the field intensity far from the cathode; h is the apex height; ak is the
vector connecting the apex peak at the origin of coordinates with k-apex peak.
Having performed the computation above, one cannot but be sure that the apex-peak
field intensity is determined by the ratio similar to (7.19) in this case as well, i.e.
2qðh R0 =2Þ
E0 ¼ bE1 ; b¼ ;
R0 E1
However, charge density q(z) meets another equation, viz.

2 3
Z 0 =2
hR
0 X
0 0
4ðh z Þ
2 2
6qðz Þ qðzÞ qðz Þ 7
qðzÞ ln + dz0 4 þ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi5 ¼ E1 z;
r02 ðzÞ jz0 zj
h þ R0 =2
k a2k þ ðz z0 Þ2
ð7:27Þ
where the sum applies to all apices except for the one located in the origin of
coordinates. Integration of (7.27) can help to make up a series similar to series (7.14).
It is not difficult to verify that the individual-apex emission area is given with
expression (7.20a) in both a multi-apex system and a one-apex case. However, it
should be noted that research of the respective asymptotic expansion for (7.27)
when studying multi-apex systems is difficult because, on the one hand, the
asymptotic expansion parameter of (ln4h/R0)−1 can be insufficiently small for such
real systems; on the other hand, even the second term of the series considered is
rather cumbersome, which prevents from efficient research. Therefore, the
multi-apex systems can be conveniently studied directly with (7.27). The approx-
imate solution of this equation seems to be equivalent to the approximate sum-
mation of the asymptotic series generated by this equation.
Let us note that numerical solution of (7.27) needs some caution because its
second term is a singular-nucleus integral, which may cause the computing circuit
to lose its stability.
For numerical computation, we selected a system consisting of semi-ellipsoidal
apices located in the nodes of a flat square grid. The equation was solved with the
Ritz method, the solution having been sought-for as odd polynomials to the 2N − 1
degree. The focus of research was on dependence of reduced field enhancement
factor b/b0, b0 being the intensity enhancement factor of an individual apex, on the
packaging parameter of ah−1, a being the distance between the two next apices, for
the apices with different peak-curvature radii.
The numerical computation showed that the solution stability was lost even for
N = 4 at ah−1 > 0.5. For N > 4, the stability was lost at even less close-packed
arrangement of the apices. But for N = 2 and N = 3 the numerical method worked
well even at ah−1 ≅ 1 and produced the results differing by about 5% at the worst.
The computation results are shown in Fig. 7.2. One can see that the apices begin
affecting each other at ah−1 ≅ 0.3. If the apices are even more close-packed, the
field-enhancement factor plummets down. The computation results suggest that the
apex packaging of more than ah−1 ≅ 3 4 is not effective.
Fig. 7.2 Dependence of β/β0

1
reduced field enhancement
factor b/b0 on the packaging 1.0
parameter of multi apex
2
system
0.8
0.6
0.4
0 4 6 ah-1
To determine field intensity at the apex peaks of the field-emission cathode, the
field potential of the system is represented as
X Z
hd
qðz0 Þdz0
Uðr; zÞ ¼ E1 z qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi;
k jr ak j2 þ ðz nÞ2 ð7:28Þ
h þ d
qðzÞ ¼ qðzÞ;
where E∞ is the field intensity far from the apices, h is the apex height, ak is the
vector connecting the k-apex foundation with the origin of coordinates.
Expression (7.28) meets the Laplace equation everywhere except for the points
lying on the apex axis and the boundary condition of u(r, 0) = 0. If values q(z) and
d meet the following conditions, viz.
d ¼ R0 =2; ð7:29Þ
Zh 0 Z
þh
4ðh2 z2 Þ qðz0 Þ qðzÞ 0 X qðz0 Þdz0 z
qðzÞ ln þ dz þ q ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ¼ ; ð7:30Þ
r02 ðzÞ jz zj
0 h
h
k
h a2k þ ðz z0 Þ2
1. R0 H 1 ¼ 0:01; b0 ¼ 50:1;
2. R0 H 1 ¼ 0:02; b0 ¼ 178:5:
where R0 is the apex-peak curvature radius, rP

0(z) is the apex-section radius taken as
the section-height function, and summation covers all the apices except for the
one located at the origin of the coordinates, then expression (7.28) in the higher
order for r also meets the boundary condition of U(r0(z), z) = 0 on the cathode apex
surface.
In the system of thin periodically located apices, the field follows expression
(7.30) with account of (7.23) and (7.28), wherefrom we can obtain the following
apex-peak field in the same thin-apex approximation, viz.
2qðh R0 =2Þ
E0 ¼ bE1 ; b¼
R0 E1
The total current from the apex peak in the field electronic cathode follows the
expression of
I ¼ Sem J0 ;
where current density J0 follows the Fowler–Nordheim formula, viz.

B
J0 ¼ AE02 exp ;
E0
and emission area Sem is equal to
E0
Sem ¼ 2p R20 :
B
Earlier, we studied a one-apex emission system and attempted to study and

optimise multi-apex systems. Here we are going to set forth the scientific results.
Let us consider a multi-apex system consisting of thin axially symmetrical apices
located on a flat substrate of an arbitrary potential zero (Fig. 7.3). The apices are in
the square-grid nodes spaced apart by a. The system anode is parallel to the cathode
at quite long distance l from it and has a potential of Ua = E∞l, E∞ being the
electric field intensity far from the cathode. The space charge of the generated
bunch of electrons is negligible.
To optimize a multi-apex system, it is necessary to know dependence of field
intensity E0 at the apex peak on the value inverse to the packaging parameter of
ah−1, i.e. ha−1.
Let us consider the different terms of (7.30). To that end, we will introduce some
new notions, viz.

R0 1 R0 1 0
n¼ h z; n0 ¼ h z;
2 2
1
1 R0 h R0 =2
~ðnÞ ¼
q E1 h q; ak ¼ a Sk ; l¼ ha1 :
2 a
Taking into account that for an elliptic apex
4ðh2 z2 Þ 4h
ln ¼ ln ;
r02 ðzÞ R0
Fig. 7.3 Model of multi apex z

system
ak a
x
h
y
(7.27) gives
4h
~ðnÞ ln
q þ K1 ð~
qðnÞÞ þ K2 ð~
qðnÞÞ ¼ n;
R0
where K1, K2 follow the expressions below, viz.
0 Z
þ1
X q~ðn0 Þdn0
K1 ð~
qðnÞÞ ¼ l qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi;
k
1 S2k þ l2 ðn n0 Þ2
Z1
~ðn0 Þ q
q ~ðnÞ 0
K2 ð~
qðnÞÞ ¼ 0 dn :
jn nj
1
Then, we consider the expansion of
1 1 1 ðn n0 Þ2 3 4 ðn n0 Þ4
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ¼ l2 þ l þ Oðl6 Þ;
0 2 jSk j 2 jS j3 8 jS j 5
Sk þ l ðn n Þ
2 2 k k
Given that q is an odd function, i.e. q (−z) = −q (z), and obtain
Z1 Z1
0 0 3 0 0
K1 ¼ C3 l 3
~ðn Þdn C5 l5
nn q ðn3 n0 þ nn 3 Þ~
qðn0 Þdn0 þ Oðl7 Þ;
2
1 1
where C3 and C5 are the coefficients stipulated by the apex-grid configuration only,
viz.
X
0
1 X
0
1
C3 ¼ ; C5 ¼ :
k jSk j3 k jSk j5
These coefficients can be obtained numerically, e.g., C3 ≅ 9.03362,

C5 ≅ 5.09026 for a square grid. It is not difficult to see that K1 is an integral
operator acting on according to the rule in the accepted approximation:
K1 ¼ C3 l3 P1 ðnÞ ~

hP1 ; qi
3 6 2
C5 l 5
~i þ ðP3 ðnÞhP1 ; q
P1 ðnÞhP1 ; q ~i þ P1 ðnÞhP3 ; q
~iÞ þ Oðl7 Þ;
2 5 5
here, h; i is the scalar product; P1(n) = n, P3(n) = 5n3/2 − 3n/2 are the Legendre
polynomials. Here let us focus on the relation of orthogonality, viz.
2 2
hP1 ; P1 i ¼ ; hP3 ; P3 i ¼ ; hP1 ; P3 i ¼ 0:
3 7
Integral K2 is an integral operator whose verifiable eigenfunctions are Legendre

P
polynomials Pn with the eigenvalues of kn ¼ 2 nk¼1 k1 . Therefore, the subspace
covering polynomials P1 and P3 make it follow the rule, viz.
77
~i
K2 ¼ 3P1 ðnÞhP1 ; q ~i:
P3 ðnÞhP3 ; q
6
Then (7.30) is reduced to the form of
4h 77
ln q ~i P3 ðnÞhP3 ; q
~ðnÞ 3P1 ðnÞhP1 ; q ~i þ C3 l3 P1 ðnÞhP1 ; q
~i
R0 6

3 6 2
C5 l5 P1 ðnÞhP1 ; q ~i þ ðP3 ðnÞhP1 ; q
~i þ P1 ðnÞhP3 ; q
~iÞ þ Oðl7 Þ ¼ P1 ðnÞ:
2 5 5
This equation is solved by
~ðnÞ ¼ a P1 ðnÞ þ cP3 ðnÞ;

q
where a and c are stipulated by equations

4h 2 6 6
ln 2 þ C3 l3 C5 l5 a C5 l5 c ¼ 1;
R0 3 5 35
ð7:31Þ
4h 11 2
ln c C5 l5 a ¼ 0:
R0 3 5
Solution (7.31) with the accepted accuracy for µ looks like

2
4h 4h 2 6
a ¼ ln 2 ln 2 þ C3 l C5 l þ Oðl6 Þ;
3 5
R0 R0 3 5
1 1
2 4h 4h 11
c¼ ln 2 ln C5 l5 þ Oðl7 Þ:
5 R0 R0 5
Consequently,
1
4h
~ð1Þ ¼ a þ c ¼ m~
q q0 ð1Þ; ~0 ð1Þ ¼
q ln 2 ; ð7:32Þ
R0
1 1 !
2C3 l3 6C5 l5 4h 4h
m¼1 þ ln 2 þ 3 ln 11 þ Oðl6 Þ:
3 lnð4h=R0 Þ 6 5 R0 R0
Eventually, the b field enhancement factor of a multi-apex system is
2qðh R0 =2Þ ð2h R0 Þ

b¼ ¼ ~ð1Þ ¼ b0 m;
q
R0 E1 R0
b0 being the individual-apex field enhancement coefficient.

Optimization of a multi-apex emission system involves knowledge of the
dependence of current density J of the system (i.e. the current from a
cathode-surface unit) on packing parameter µ. The multiplicative terms having been
taken away, the current of one apex is seen as

B B
I0
E03 exp
m3 ;
E0 bE1 m
and the current density of the system as

1 B
J
2 I0
l m exp
2 3
:
a b0 E1 x
Taking into account that value m is described as
m ¼ 1 d3 l3 þ d5 l5 þ Oðl6 Þ ð7:33Þ
d3 and d5 being obvious from comparison between (7.32) and (7.33), it is not
difficult to obtain
J
l2 ð1 Dðd3 l3 d5 l5 ÞÞ þ Oðl8 Þ; ð7:34Þ
where
B
D¼ þ 3:
b0 E1
Having differentiated (7.34), equated the expression obtained to zero and taken
away the trivial root of µ = 0, we obtain the equation stipulating the optimum
parameter of µ:
2
5d3 l3 7d5 l5 þ Oðl6 Þ ¼ : ð7:35Þ
D
Studies of (7.35) show that it has a single root arbitrary close to zero at D ! ∞.
Other roots are separated from zero at any value of D. Definitely, it is this solution
of the problem that is physically sensible, and other roots appeared due to the
approximate nature of (7.35).
In the higher order, (7.35) produces
1=3
2
l¼ ;
5d3 D
then O(µ6) = O(D−2), and (7.35) looks like
2
5d3 l4 7d5 l5 ¼ þ OðD2 Þ: ð7:36Þ
D
Solution of (7.36) with this accuracy results in

1=3
2 14 d5 1
l¼ þ D þ OðD1 Þ; ð7:37Þ
5d3 D 75 d32
2C3 l3
d3 ¼ ; ð7:38Þ
3 lnð4h=R0 Þ 6
" 1 1 #
6C5 l5 4h 4h
d5 ¼ ln 2 þ 3 ln 11 : ð7:39Þ
5 R0 R0
The procedure proposed enabled to determine the optimum value of

µ = (h − R0/2)/a ≅ ha−1 at the fixed values of the other system parameters (h, R0,
E∞). In the first approximation, aopt, i.e. the distance between the apices, is given
by the expression that easily follows from (7.37) and (7.38), viz.

1=3
5C3 D
aopt ¼ h ; ð7:40Þ
3 lnð4h=R0 Þ 6
B
D¼ þ 3 1: ð7:41Þ
b0 E1
Expressions (7.40) and (7.41) show that the optimum cathode-apex packaging
depends on the operating point, i.e. on E∞ of the emission diode, and we can see
that in high operating voltages in the anode-cathode gap more close-packed
arrangement of apices becomes optimal. The reason is that the curvature of the
volt-ampere emission-diode dependence decreases with field-intensity growth at the
cathode surface, which means that the emission-current density becomes less
sensitive to drop in the field intensity at the apex peak caused by tightening of the
apex structure. Expressions (7.37)–(7.39) give the first two terms of the µ value
expansion (which is equivalent to aopt) in parameter D.
7.3 Mathematical Simulation of Model Triode

Electron-Optical Systems
This paragraph suggests a physical model of a practically significant

electron-optical system, EOS, based on three electrodes, viz. a cathode, an anode
and a modulator.
The physical model of a triode electron-optical system is shown in Fig. 7.4. The
surface of the cathode represented by a thin field apex of length L is a sphere of
radius R0 n a spindle-shaped surface of revolution with foundation radius R The
surface of the anode and the cathode substrate is modelled by radius sphere Ra, and
the modulator surface is simulated by the spherical segment of radius Rm.
The potential at the cathode is equal to 0, the modulator potential being equal to
V1, the anode potential being set by function f(a).
The physical model considered is represented with a mathematical model, which
is to find function Uða; bÞ meeting the Laplace equation and boundary conditions
(7.42), viz.
8
> DUða; bÞ ¼ 0;
>
>
>
< Uða; b0 Þj0 a a0 ¼ 0;
Uða0 ; bÞjb2 \b\b0 ¼ 0; ð7:42Þ
>
>
>
> Uða; b1 Þja1 a a0 ¼ V1 ;
:
Uða; b2 Þj0 a a0 ¼ f ðaÞ:
To solve the problem, let us consider bispherical coordinates. Bispherical

coordinates ða; bÞ are related to cylindrical coordinates ðr; zÞ as [44], viz.
a þ ib
z þ ir ¼ icctg ;
2
0 a p; 1\b\1:
Coordinate surfaces a ¼ const set the family of spindle-shaped revolution sur-

faces, viz.
c 2
ðr c ctgaÞ2 þ z2 ¼ :
sina
7.3 Mathematical Simulation of Model Triode Electron-Optical Systems 389
Fig. 7.4 The physical model

of a triode electron-optical
system β2 V2
β1 α2
V1 0
α1
α0
Coordinate surfaces b ¼ const set the family of non-crossing non-concentric

spheres orthogonal to surfaces a ¼ const, viz.
2
2 c
r þ ðz c cthbÞ ¼
2
:
shb
In the bispherical coordinate system, the surfaces of electrodes of the

electron-optical system considered can be presented as parts of the coordinate
surfaces, viz.
• is the sphere on the apex peak
0 a a0 ; b ¼ b0 ;
• is the apex body
a ¼ a0 ; b2 \b\b0 ;
• is the modulator
a1 a a0 ; b ¼ b1 ;
• is the anode / substrate
0 a a0 ; b ¼ b2 ;
The geometrical sizes of the simulated system electrodes are determined by the
following formulas, viz.
• is the radius of the sphere on the apex peak
1 þ chb0
R0 ¼ c ;
shb0
• is the apex length
1 þ chb0
L¼c ;
shb0
• is the apex foundation radius

a0
R ¼ c ctg ;
2
• is the modulator radius

c
Rm ¼ ;
shb0
• is the anode / substrate radius

c
Ra ¼ ;
shb2
• is the distance from the apex peak to the anode

1 þ chb2 1 þ chb0
S¼c :
shb2 shb0
The potential at the cathode is equal to 0, the modulator potential being equal to
V1, the anode potential being set by function f(a), viz.
(
V 2 ; 0 a a2 ; b ¼ b2 ;
f ðaÞ ¼ ð7:43Þ
0; a2 a a1 ; b ¼ b2 :
The boundary problem (1) is solved by breaking of the entire electron-optical

system area into 3 subareas, viz.
(1) 0 a a0 ; b1 b b0 ;
(2) 0 a a0 ; b2 b b1 ;
(3) 0 a a1 ; b2 b b0 :
For each of the subareas, the potential distribution can be presented as
Legendre-function expansion, viz.
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi X1
shðmn þ 1=2Þðb0 bÞ
U1 ða; bÞ ¼ chb cos a An Pmn ðcosðp aÞÞ; ð7:44Þ
n¼1
shðm n þ 1=2Þðb0 b1 Þ
shðmn þ 1=2Þðb b2 Þ
U2 ða; bÞ ¼ chb cos a An
shðm n þ 1=2Þðb1 b2 Þ
n¼1
ð7:45Þ
þ Bn Pm ðcosðp aÞÞ;
shðmn þ 1=2Þðb1 b2 Þ n
"
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi X 1
P1=2 þ isn ðcos aÞ pnðb0 bÞ
U3 ða; bÞ ¼ chb cos a Cn sin
n¼1
P 1=2 þ isn ðcos a 1 Þ b0 b 2
# ð7:46Þ
X1
shðlm þ 1=2Þðb0 bÞ
þ Dm Pl ðcos aÞ ;
m¼1
shðlm þ 1=2Þðb0 b2 Þ m
where Pk ðcos aÞ is the Legendre function of the first kind; mn ; lm are the roots of
equations, respectively,
pn
Pkn ðcos a0 Þ ¼ 0; Plm ðcos a1 Þ ¼ 0; sn ¼ :
b0 b2
Presentation of the potential as (7.44)–(7.46) meets the Laplace equation.

Obviously, expansions (7.44), (7.45) for functions U1 ða; b1 Þ ¼ U2 ða; b1 Þ provide
continuity of the potential at the boundary between areas 1 and 2. Moreover, as all
the three areas for which expansions (7.44)–(7.46) have been written overlap each
other, the conditions of continuity of the potential and its normal derivatives at the
boundary between the areas are automatically met.
Coefficients Bn ; Dm are computed from boundary conditions (7.44), (7.45) on
surface b = b2, viz.
(
V 2 ; 0 a a2 ; b ¼ b2 ;
Uða; b2 Þj0 a a0 ¼
0; a2 a a1 ; b ¼ b2 :
Let us insert the respective value at the boundary into the potential distribution in
the third area b = b2:
"
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi X 1
P1=2 þ isn ðcos aÞ pn ðb0 b2 Þ
U3 ða; b2 Þ ¼ ch b2 cos a Cn sin ;
n¼1
P1=2 þ is n
ðcos a 1 Þ b0 b 2
#
X1
shðlm þ 1=2Þðb0 b2 Þ
þ Dm Pl ðcos aÞ
m¼1
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi X
1
V2 ¼ ch b2 cos a Dm Plm ðcos aÞ; 0 a a2 ;
m¼1
V2 X
1
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ¼ Dm Plm ðcos aÞ
chb2 cos a m¼1
Using the Legendre-polynomial property, we transfer from summation to inte-

gration with normalizing coefficient
Z1
Pm ðxÞPn ðxÞdx ¼ 1; n ¼ m;
1
we get
Zp pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
V2 ½sin aPln ðcos aÞ= 2ðch b2 cos aÞda ¼ Dm ;
0
replace variables x ¼ cos a and turn to finding coefficients Dm :
Z1 pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
V2 ½Pn ðxÞ= 2ðch b2 xÞdx ¼ Dm ;
1
ð7:47aÞ
Z1
V2 Plm ðxÞ
Dm ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi dx:
Mm ch b2 x
cos a2
Mm being the normalizing coefficient,
Z a1
cos
Mm ¼ ½Plm ðcos aÞ2 sin ada:

0
Coefficient Bn is found from boundary conditions (7.42), (7.43) on surface

b = b2
The second-area potential is represented as follows:

shðmn þ 1=2Þðb b2 Þ shðmn þ 1=2Þðb1 bÞ
U2 ða; bÞ ¼ chb cos a An þ Bn
n¼1
shðmn þ 1=2Þðb1 b2 Þ shðmn þ 1=2Þðb1 b2 Þ
Pmn ðcosðp aÞÞ:
At the outer sphere, the potential is represented as follows:

(
V2 ; 0 a a2 ; b ¼ b2 ;
Uða; b2 Þj0 a a0 ¼
0; a2 a a1 ; b ¼ b2 :
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi X1

U2 ða; b2 Þ ¼ chb2 cos a An þ Bn
n¼1
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi X
1
Pmn ðcosðp aÞÞ; V2 ¼ chb2 cos a Bn Pmn ðcosðp aÞÞ; 0 a a2 ;
n¼1
V2 X
1
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ¼ Bn Pmn ðcosðp aÞÞ:
chb2 cos a n¼1
Using the Legendre-polynomial property, we transfer from summation to inte-

gration with the normalizing coefficient
Z1
Pm ðxÞPn ðxÞdx ¼ 1; n ¼ m;
1
get
Zp pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
V2 ½sin aPmn ðcosðp aÞÞ= 2ðchb2 cos aÞda ¼ Bn ;
0
replace variables x ¼ cos a and turn to finding coefficients Bm
Z1 pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
V2 ½Pn ðxÞ= 2ðchb2 xÞdx ¼ Bm ;
1
ð7:47bÞ
Z1
V2 Pmn ðxÞ
Bn ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi dx;
Nn chb2 x
cos a2
Nn being the normalizing coefficient

Z a0
cos
Nn ¼ ½Pmn ðcos aÞ2 sin ada;

0
If the ratio is x=chb2 1; coefficients Bn ; Dm in (7.47a) can be calculated

approximately:

V2 1 cos a2 mn
Bn pffiffiffiffiffiffiffiffiffiffi Pmn 1 ðcos a2 Þ
Nn chb2 mn þ 1 2chb2 ðmn 1Þðmn þ 2Þ

cos a2 sin2 a2 þ mn cos2 a2
þ Pmn ðcos a2 Þ þ ;
mn þ 1 2chb2 ðmn 1Þðmn þ 2Þ

V2 1 cos a2 mn
Dm pffiffiffiffiffiffiffiffiffiffi Plm 1 ðcos a2 Þ
Mm chb2 lm þ 1 2chb2 ðlm 1Þðlm þ 2Þ

cos a2 sin2 a2 þ lm cos2 a2
þ Plm ðcos a2 Þ þ :
lm þ 1 2chb2 ðlm 1Þðlm þ 2Þ
Coefficients An ; Cn are determined directly from the conditions at the bound-

aries of the areas selected

U3 ða; b1 Þ; 0 a a1 ;
U1 ða; b1 Þ ¼ ð7:48Þ
V1 ; a1 a a0 ;
We insert expressions (7.44)–(7.46) in the right parts of Legendre-function

expansion (7.48), respectively.
shðmn þ 1=2Þðb0 b1 Þ
U1 ða; b1 Þ ¼ chb1 cos a An Pm ðcosðp aÞÞ;
n¼1
"
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi X 1
P1=2 þ isn ðcos aÞ pnðb0 b1 Þ
U3 ða; b1 Þ ¼ chb1 cos a Cn sin
n¼1
P 1=2 þ isn ðcos a 1 Þ b 0 b2
#
X1
shðlm þ 1=2Þðb0 b1 Þ
þ Dm Pl ðcos aÞ ;
m¼1
(
U1 ða1 ; bÞ; b1 b b0 ;
U3 ða1 ; bÞ ¼
U2 ða1 ; bÞ; b2 b b1 :
ð7:49Þ
We insert expressions (7.44)–(7.46) in the right parts of Fourier expansion

(7.49), respectively.
"
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi X 1
P1=2 þ isn ðcos a1 Þ pnðb0 bÞ
U3 ða1 ; bÞ ¼ chb cos a1 Cn sin
n¼1
P 1=2 þ isn ðcos a1 Þ b0 b2
#
X1
shðlm þ 1=2Þðb0 bÞ
þ Dm Pl ðcos a1 Þ ;
m¼1
U1 ða1 ; bÞ ¼ chb cos a1 An Pm ðcosðp a1 ÞÞ;
n¼1

shðmn þ 1=2Þðb b2 Þ shðmn þ 1=2Þðb1 bÞ
U2 ða1 ; bÞ ¼ chb cos a1 An þ Bn Pmn ðcosðp a1 ÞÞ:
n¼1
Using completeness of the respective functions, we get combined algebraic

equations in respect of coefficients An ; Ck
1 X 1 h i
An Ck P1=2 þ isk ðcos a1 ÞP1mn ðcos a1 Þ P11=2 þ sk ðcos a1 ÞPvn ðcos a1 Þ
Nn k¼1
0 b1 Þ
sin pkðb
b b sin a1 1
0 21
¼ ðT1;n þ T2;n Þ;
P1=2 þ isk ðcos a1 Þ s2k þ ðmn þ 1=2Þ Nn
ð7:50Þ
2 X 1
mn þ 1=2 shðmn þ 1=2Þðb0 b1 Þ
Cn An
b0 b2 n¼1 s2k þ ðmn þ 1=2Þ2 shðmn þ 1=2Þðb0 b1 Þshðmn þ 1=2Þðb1 b2 Þ
pkðb0 b1 Þ 2
sin Pmn ðcos a1 Þ ¼ T3;k ;
b0 b2 b0 b2
ð7:51Þ
where
Z a0
cos
Pmn ðxÞ
T1;n ¼ V1 pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi dx; ð7:52Þ
chb1 x
cos a1
X
1
shðlm þ 1=2Þðb0 b1 Þ sin a1 Plm ðcos a1 ÞPmn ðcos a1 Þ
1
T2;n ¼ Dm ; ð7:53Þ
m¼1
shðlm þ 1=2Þðb0 b2 Þ ðlm þ mn Þðlm þ mn þ 1Þ
0 b1 Þ
!
X
1
Pmm ðcos a1 Þ kþ1
sin pkðb
b b
T3;n ¼ Bm sk ð1Þ ðmm þ 1=2Þ 0 2
:
m¼1 s2k þ ðmm þ 1=2Þ2 shðmm þ 1=2Þðb1 b2 Þ
ð7:54Þ
P1k ðcos aÞ is the associated Legendre functions of the first kind.

If the ratio is x=chb1 1; ðcos a1 \x\ cos a0 Þ, coefficients T1;n can be
approximately calculated from (7.52):

V1 1 cos a1 mn
T1;n pffiffiffiffiffiffiffiffiffiffi Pmn 1 ðcos a1 Þ
chb1 mn þ 1 2chb1 ðmn 1Þðmn þ 2Þ

cos a1 sin2 a1 þ mn cos2 a1
þ Pmn ðcos a1 Þ þ ð7:55Þ

1 cos a0 mn
Pmn 1 ðcos a0 Þ :
Thus, this paragraph computes distribution of the electrostatic potential in the

entire area of an axisymmetric electron-optical triode system with a field apex and a
modulator. With the area overlap method, initial problem (7.42), (7.43) is reduced to
finding unknown coefficients in the potential eigenfunction expansion (7.43)–(7.45)
from combined linear algebraic equations (7.50), (7.51), their right parts being cal-
culated from (7.52)–(7.55). All the geometrical sizes, the potentials of the modulator
and the anode are problem parameters. Such systems can serve to simulate a field-apex
electron gun. Finding the potential distribution in such systems is the most difficult
part of the problem as the geometrical sizes of the electrodes differ by several orders of
magnitude, which complicates application of numerical computation methods.
E.g., let us consider a physical model of an electron gun consisting of a cathode
represented by an axial-symmetric apex of an arbitrary shape on a flat metal sub-
strate and an anode represented by an infinite plane and a modulator. The potential
at the cathode and the substrate is
Equal to 0, that of the anode and the modulator being equal to U. Figure 7.5
gives the potential distribution for an apex with a sphere-on-cone surface. This case
involves dimensionless units of measure, viz. the potential at the anode and the
modulator U = 1, the apex length being 0.1, the distance between the apex substrate
and the anode being 1.
U
1.0
0.8
0.6 «test»
0.996
0.4 0.998
0.96
0.2
0.98
0 0.96
0.94
0.995
0.2
0.15
r 0.1 1.3 1.4
1.1 1.2
0.05 0.9 1.0
0 z
0.8
Fig. 7.5 The potential distribution for an apex with a sphere-on-cone surface
7.4 Calculation of Distribution of the Electric Field in the Forming … 397
7.4 Calculation of Distribution of the Electric Field

on a Field-Emission Cathode and a Small-Aperture
Focusing Diaphragm System
The previous paragraph reviewed the mathematical model of an electron gun with a
system of focusing electrodes represented by diaphragms. When the
diaphragm-aperture radii are much smaller than the distances between them, viz.
Ri ðZi þ 1 Zi Þ; ðZi Zi1 Þ; ð7:56Þ
many field-potential distribution formulas become considerably simpler.

Kernels Ki j ðx; tÞ of the combined Fredholm equations in respect of the functions
/i ðtÞ can be represented as series, viz.
" #
X kðZi þ 1 Zi Þ
Kii ¼ 8xt
k¼2;4;6... ððkðZi þ 1 Zi ÞÞ2 þ ðt xÞ2 ÞððkðZi þ 1 Zi ÞÞ2 þ ðt þ xÞ2 Þ
" #
X kðZi Zi1 Þ
þ 8xt
k¼2;4;6... ððkðZi Zi1 ÞÞ2 þ ðt xÞ2 ÞððkðZi Zi1 ÞÞ2 þ ðt þ xÞ2 Þ
X kðZi Zi1 Þ
Kii1 ðx; tÞ ¼ 8xt :
k¼1;3;5... ððkðZi Z1 ÞÞ þ ðt xÞ2 Þ ððkðZi Zi1 ÞÞ2 þ ðt þ xÞ2 Þ
2
ð7:57Þ
On the basis of (7.56) and due to the fact that 0 t; x Ri , kernels Ki j ðx; tÞ
(7.57) assume a very simple form, viz.
1
Kii1 ðx; tÞ8C1 xt ;
ðZi Zi1 Þ3
" # ð7:58Þ
1 1
Kii ðx; tÞ ¼ 8C2 xt þ ;
ðZi þ 1 Zi Þ3 ðZi Zi1 Þ3
where
X 1 X 1
C1 ¼ 1:0517998; C2 ¼ 0:1502571:
1;3;5;...
K3 2;4;6...
k3
In the matrix form, that is

ðE K Þ/ðxÞ ¼ UðxÞ; ð7:59Þ
where E is the identity operator; /ðxÞ ¼ ð/1 ðxÞ; . . .; /N ðxÞÞ is the sought-for
function vector; UðxÞ ¼ ðU1 ðxÞ; . . .; UN ðxÞÞ is the free-function vector. K ¼ ðK ~ i;j Þ
is the matrix operator, viz.
8 RRi1
>
>
>
>
1
Ki i1 ðx; tÞdt; j ¼ i 1;
>
>
2p
>
>
0
>
< RRi
~ i;j ¼ 2p
1
Ki i ðx; tÞ dt; j ¼ i;
K ð7:60Þ
>
>
0
>
> RR
iþ1
>
> 1
Ki þ 1 i ðx; tÞdt; j ¼ i þ 1;
>
>
>
:
2p
0
0; j 6¼ i 1; i; i þ 1:
Obviously, operator K with kernels (7.58) transforms any measured N-dimen-

sional measured function into an N-dimensional symmetric function.
Let us consider function V1 ðr; zÞ the field-potential distribution in the initial
system without account of the apex. Let us introduce the following norm in the

space of CN N-dimensional functions /ðxÞ ¼ ð/1 ðxÞ; . . .; /N ðxÞÞ:

¼ maxðmaxj/ ðxÞjÞ:
/ i
i x

Matrix operator K maps CN in CN. Clearly, if y ¼ K/ðxÞ,
X Z
Rj þ 1
yi ðxÞ ¼ Ki j /j and Ki j /j ¼ Ki j ðx; tÞ/j ðtÞdt:

j
Rj
Then
X Z
Rj þ 1

maxjyi ðxÞj max K:
Ki j ðx; tÞ dt max /j ðtÞ /
x x x
j
Rj
In virtue
of assumption (7.56), stationary value K is less than 1 (K < 1), then
: So the K-operator norm is less than 1, and there is an inverse operator
k yk K /
of (E − K)−1 and
7.4 Calculation of Distribution of the Electric Field in the Forming … 399
ðE KÞ1 ¼ E þ K1 þ K2 þ :
Free terms UðxÞ of system (7.59) are linear functions, viz.

2 Ui þ 1 Ui Ui Ui1
Ui ðxÞ ¼ x:
p Zi þ 1 Zi Zi Zi1
¼ ðbi xÞ. Equations

Let us consider the effect of operator K on linear function /
(7.58) and (7.60) give us
¼ ðdi xÞ;
K/
4 n
di ¼ C1 bi1 ðRi1 =ðZi Zi1 ÞÞ3 þ C1 bi þ 1 ðRi þ 1 Zi þ 1 Zi Þ3
3p h io
C2 bi ðRi Zi þ 1 Zi Þ3 þ ðRi Zi Zi1 Þ3 ; i ¼ 1; N:
So we can derive the solution of system (7.59), viz.
X
1

/ðxÞ ¼ K n U:
n¼0
Accurately to OðRi =ðZi þ 1 Zi ÞÞ6 , we get

/ðxÞ ¼ ðai xÞ; ð7:61Þ
where
( 3 3
1 2 Ri1 Ri þ 1
ai f i þ C1 fi1 þ C 1 fi þ 1
2 3p Zi Zi1 Zi þ 1 Zi
" 3 3 #)
Ri Ri
C2 fi þ ; i ¼ 1; N;
Zi þ 1 Zi Zi Zi1

2 Ui þ 1 Ui Ui Ui1
fi ¼ :
p Zi þ 1 Zi Zi Zi1
System (7.59) can be solved without the successive-approximation method, but

on the basis of combined linear algebraic equations. Really, as is an /ðxÞ—N-

dimensional linear function, we can write that /ðxÞ ¼ ðai xÞ, and get the following
algebraic linear-equation system for ai, viz.
X
N
pi j aj ¼ f i ;
j¼1
8 3
>
> 3p C1 ZiRZ ; j ¼ i 1;
>
>
8 i1
>
>
3
i1
3
>
<
2 þ 3p
8
C2 Zi þR1 iZi þ Zi Z Ri
; j ¼ i;
pi j ¼ i1
>
>
>
>
3
>
> 3p
8
C1 Zi þR1 iZi ; j ¼ i þ 1;
>
:
0; j 6¼ i; i 1; i þ 1;
fi are determined in (7.61). So, the solution of this system is existent and unique.
It can be derived as the approximate analytical solution for a boundary-value
problem of the field-potential distribution in the small-aperture diaphragm system
without account of the apex. The solution is given by
Ui þ 1 Ui
V1 ðr; zÞjZi z Zi þ 1 ¼ Ui þ ðz Zi Þ
Z Z
( i þ 1
iqffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
X1
t þ iak
þ Im ai r 2 þ ðbk i tÞ2 r 2 þ ðak i tÞ2 i ak arcsin
k¼0
r
Ri
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
t þ ibk
þ ibk arcsin ai þ 1 r 2 þ ðb0 itÞ2 þ r 2 þ ða0 i tÞ2
r k k
Ri þ 1 #)
0

t þ iak t þ ibk
þ ia0k arcsin ib0k arcsin ;
r r 0
where
ak ¼ 2k ðZi þ 1 Zi Þ þ ðz Zi Þ;
bk ¼ ð2k þ 1ÞðZi þ 1 Zi Þ þ ðZi þ 1 zÞ;
a0k ¼ 2k ðZi þ 1 Zi Þ þ ðZi þ 1 zÞ;
b0k ¼ ð2k þ 1ÞðZi þ 1 Zi Þ þ ðz Zi Þ:
In particular, for the axial-potential distribution V1 ð0; rÞ, we can write
Ui þ 1 Ui
V1 ð0; zÞjZi z Zi þ 1 ¼ Ui þ ð z Zi Þ
Zi þ 1 Zi
X1
Ri Ri Ri þ 1 Ri þ 1
þ ai ak arctg bk arctg þ ai þ 1 a0k arctg ; b0k arctg 0 :
k¼0
ak bk a bk
7.5 Calculation of Optimum Characteristics in the Forming … 401
7.5 Calculation of Optimum Characteristics

on a Field-Emission Cathode and a Focusing
Diaphragm System
One of the main characteristics of an electron-optical system represented by an

electron gun with a field-emission cathode is the total current generated as a result
of field emission.
For practical purposes, the specialists often introduce a term of a source-uti-
lization coefficient. The source is understood as a real object, viz. an apex, emitting
electrons. The coefficient can be determined at the detector as a ratio of the total
electron flow from a monoenergetic source to the total flow emitted by the source.
Determination of characteristics of a field-emission cathode is difficult because
as far as this kind of emission is concerned, the field generated by all electrodes of
the system directly affects the emission characteristics of the cathode: emission
area, current density and total current. Therefore the beam-focusing problem has to
be solved together with optimization of the emission characteristics of the
field-emission cathode.
Let us understand all geometrical dimensions, viz. the focusing-diaphragm
arrangement and radii, the curvature radius, the apex length, as well as the apex
shape and the electrode potentials, as input system parameters. The requirements to
minimize the deviation of the total apex current from a given value and to increase
the source utilization coefficient at the given running voltages are alternative
because the flow of the electrons emitted by the field emission cathode has a large
semiangle of divergence, viz. from 30 to 45°. There are optimum input parameters
providing minimum deviation of the total apex current from the given value under
the condition that no electrons of the flow deposit on the anodic and focusing
diaphragms, i.e. the source utilization coefficient is about 1. The optimum input
parameters are determined by several reasons. Firstly, the total current emitted by
the source/apex at given parameters can be much closer to the desired value, than at
more or less arbitrary selection of the parameters. Secondly, the current density at
the apex peak cannot exceed some limiting values as researches of pointed
field-emission cathodes have shown that if a cathode is overloaded with its own
field-emission current, it is destroyed by the explosive heat. The apex destruction is
followed by development of an arc between the anode and the cathode. Thirdly,
practitioners always resort to experiments to find the best relations between the
parameters. However, their optimum values are seldom achieved because the
experimental search of the minimum function of many variables is an extremely
laborious and complex task. Knowledge of the design values almost completely
saves from the experimental approach and provides the total-current value desired
in practice.
Strictly speaking, to optimist the input parameters of the electron gun with a field
emission cathode and a focusing-diaphragm system is to get the minimum deviation
of the total-current value from the given one at constant anode voltage and within
the given parametric constraints by proper correlation of the parameters and

selection of the apex shape.
The characteristic criterion of our problem shall be the squared deviation of the
total current emitted by the field-emission cathode from the given value. The best
option is the minimum characteristic exponent.
The basic field-emission dependence, viz. that of the current density on the
intensity of the emission-causing impact represented in this case by the field itself is
computed by the known Fowler–Nordheim formula.
In practice, it is not the current density that is usually measured as a
field-intensity function, but the total field-emission current as a function of the
voltage applied to the field emission cathode, viz.
I ¼ Sem j0 ; ð7:62Þ
Sem being the emission area, j0 being the current density at the apex peak. Earlier
we have given the formulas to calculate the emission area and the density of the
current from the apex peak, viz.
E0 ¼ 2R1
0 qð0Þ; ð7:63Þ
SeH ¼ 2pR0 E0 =B; ð7:64Þ
where
4ð2mÞ1=2 U3=2
B¼ hðaÞ; ð7:65Þ
3he
and qð0Þ being the charge density at the apex axis.

So if we know the current density at the apex peak, we can calculate charac-
teristic criterion I with formulas (7.62)–(7.65).
Addressing the optimization problem shall start with selection of its parameters.
The reason is that computation of a highly-dimensional problem according to the
optimization algorithms is very time-consuming. So there should be as many
parameters as is proper for the optimization problem to be solved within a real
period of time. It is very important to take all basic independent variables into
consideration. Nonetheless, it is important not to “overload” the problem with a
large number of insignificant parameters.
The adequacy of the theoretical designs was tested with a series of numerical
experiments involving development of a software package implementing the sug-
gested mathematical models of an electron gun with a field-emission cathode and a
system of focusing electrodes represented by diaphragms.
Figures 7.6 and 7.7 shows the equipotentials of electrostatic fields. It is worth
noting that researching into the effect of the apex shape involved consideration of
canonic quadratic curves as forming surfaces.
Fig. 7.6 The map of

electrostatic fields for cathode (a) 100
unit: from apex to anode
diaphragm: a 3D map, 80
b equipotential lines U, V
100 60
80 40
60 20
40 0
0.105
20 0.100
0.095 z, cm
0 0.090
r, cm 0.085
0.010
0.008 0.080
0.006
0.004
0.002
0
(b)
r, cm
0.012
78.1374
43.9523
48.8359
53.7195
58.6031
63.4867
68.3703
73.2538
83.0210
87.9046
0.010
0.008
39.0687
0.006
0.004
29.9015 34.1851
0.002
19.5340 24.4179
9.7818
0
0.080 0.085 0.090 0.095 0.100 0.105 0.110
z, cm
The following parameters were taken as given/ known: l = 0.1 cm is the apex
length; R0 = 10−6 cm is the apex-peak curvature radius; z2 = l + 100R0,
r2 = 100R0 are the anode-diaphragm coordinates, viz. the position and the aperture
radius; U2 = 100 V is the potential at the anode diaphragm.
The apex shape depends on the radius of its foundation (rbase), which is an
pffiffiffiffiffiffiffi pffiffiffiffiffiffiffiffiffi pffiffiffiffiffiffiffiffiffi
ellipsoid with lR0 rbase \ 2lR0 , a paraboloid with rbase ¼ 2lR0 , and a hy-
pffiffiffiffiffiffiffiffiffi
perboloid of revolution with rbase [ 2lR0 .
An elliptic apex is the thinnest, so the physical considerations prompt that it is
this apex shape that makes the field change more rapidly. Actually comparing
Fig. 7.7a, b, we can see that the field potential increases more rapidly for the elliptic
rather than hyperbolic apex in the neighbourhood of the peak.
Figure 7.8 shows that the elliptic apex is characterized by the value of
pffiffiffiffiffiffiffi pffiffiffiffiffiffiffi
rbase \ 0:2 103 cm, while the hyperbolic one by rbase [ 0:2 103 cm. It
also shows that increase in the foundation radius from 0.3 10−3 to 5 10−3 cm
Fig. 7.7 Potential

distribution near the apex:
a 3D map for the hyperbolic
apex, b equipotential lines for
(a)
the hyperbolic apex U, V
35
30
25
20
15
10 0.100015
0.100010
5 0.100005
0.100000
0 z, cm
12 0.099995
10
8 6 0.099990
r·10-6, cm 4 2
1 0
(b)
r·10-6, cm
10
8 17.4929
10.4957
13.9943 20.9915 27.9886

6 6.9972
24.4900
3.4986
4
0
0.099990 0.100000 0.100010 z, cm
causes the current value to decrease three times, viz. from 0.75 10−3 to
0.25 10−3A.
The pattern of the constraining electron paths emitted from the apices of different
shape (Fig. 7.9), shows that the revolution-hyperboloid apex is least preferable.
Let us study the effect of other parameters on the value of the total current from
the apex. Figure 7.10 demonstrates some nomograhic charts resulting from a
numerical experiment, viz. the current logarithm—reverse anode voltage curves.
The slopes and arrangement of these straight lines meet the known experimental
volt-ampere dependences, which validates theoretical computations. Table 7.1 give
the values of field intensity at the apex peak (E0 ), current density at the apex peak
(j0 ), emission area (Sem) and total current (I) as dependent on voltage at the anode
diaphragm (U2 ) for work function U = 4.5 eV.
Let us consider the effect of the anode-diaphragm arrangement on the emission
characteristics. Figure 7.11 shows dependence of the total current on the
(a)
I·10-4, A
7
0 1 2 3 4r ·10-3, cm
base
(b)
I·10-4, A
7.40
7.36
7.35
7.28
7.24
3.0 3.4 3.8 4.2 r ·10-3, cm
base
Fig. 7.8 Dependence of full current from rbase for elliptic a and hyperbolic b apex
r· 10-5, cm
18
14
10
8
6
4
2
0.99995 1.00000 1.00005 1,00010 1,00015 z, cm
Fig. 7.9 The pattern of the constraining electron paths: solid line for electron emitted from the
elliptic apex, dashed line for electron emitted from the hyperbolic apex
ln(I/U22)
-5
-15 Φ=2.0 eV
-25 Φ=4.5 eV
Φ=6.0 eV
-35
-1 -1
0.003 0.004 0.005 0.006 0.007 0.008 0.008 U2 , kV
Fig. 7.10 Dependence ln(I/U22) from U−1

2 for three work function values
Table 7.1 Emission characteristics at different anode voltages

U2 (V) E0 107 (V/cm) i0 (A/cm2) Sem 10−12 (cm2) I (A)
100 1.55709 3.22958 0.177319 0.572666
180 2.94974 2.50054 0.387450 0.968834
200 3.10834 8.22359 0.415330 0.341550
225 3.66420 2.37240 0.519888 0.123338
250 4.09674 1.75878 0.610108 0.107304
275 4.51379 8.53180 0.705440 0.601867
300 4.92709 3.15338 0.808786 0.255041
325 5.33913 9.58303 0.922394 0.883933
350 5.75063 2.49093 1.047320 0.260881
375 6.16187 5.73000 1.185780 0.679451
400 6.57298 1.19396 1.340140 0.160008
425 6.98400 2.28392 1.511870 0.345299
450 7.39497 4.09775 1.707140 0.699543
475 7.80592 6.90683 1.926410 0.133054
500 8.21684 1.11109 2.178660 0.242069
I·10-4, A
7.430
7.426
7.422
7.418
7.414
2 4 6 8 r ·10-5, cm
2
Fig. 7.11 Dependence of the total current on the anode-diaphragm radius
anode-diaphragm radius. The effect of the radius can be seen as small and
expectedly decreasing with its reduction.
The effect of the anode-diaphragm arrangement (z2) is shown in Table 7.2.
It shows that the anode-diaphragm arrangement has a great impact on the total
current, viz. its values change by one and a half order of magnitude when the
distance between the anode diaphragm and the apex peak changes from r0 102 to
r0 104 i. e. by two orders of magnitude.
The emission characteristics are greatly effected by introduction of a cathode
diaphragm with coordinates (r1 ; z1 ) at 0\z1 \l, l being the apex length, which is
shown in Tables 7.3 and 7.4. The potential of the cathode diaphragm coincides with
the potential of the apex/substrate. Table 7.3 presents r1 values as selected at the
distance r0 0:1 of from the apex surface. As the cathode-diaphragm aperture is
Table 7.2 Emission characteristics at different anode-diaphragm arrangement

z2 (cm) E0 107 (V/cm) j0 107 (A/cm2) Sem 10−12 (cm2) I 10−4 (A)
0.10010 7.70721 6.124270 1.87134 1.1460600
0.10109 7.54038 4.961780 1.78178 0.8840780
0.10208 7.37814 4.005990 1.69869 0.6804950
0.10307 7.22031 3.222870 1.62143 0.5225660
0.10406 7.06665 2.583120 1.54939 0.4002250
0.10505 6.91685 2.061980 1.48200 0.3055850
0.10604 6.77062 1.645510 1.42012 0.2336820
0.10703 6.62768 1.307360 1.36189 0.1780470
0.10802 6.48774 1.033370 1.30685 0.1350450
0.10901 6.35052 0.812207 1.25470 0.1019070
0.11000 6.21576 0.634435 1.20515 0.0764586
Table 7.3 Emission characteristics at different cathode-diaphragm coordinates

z1 (cm) r1 10−4 (cm) E0 107 (V/cm) j0 (A/cm2) Sem 10−12 (cm2) I (A)
0.0100 3.1470 6.383 8.620 1.2670 0.1092
0.0185 3.1080 6.353 8.161 1.2550 0.1024
0.0270 3.0450 6.313 7.598 1.2400 0.9429
0.0355 2.9570 6.260 6.899 1.2210 0.8427
0.0440 2.8400 6.188 6.026 1.1950 0.7203
0.0525 2.6920 6.085 4.945 1.1580 0.5730
0.0610 2.5060 5.933 3.645 1.1060 0.4033
0.0695 2.2740 5.692 2.197 1.0290 0.2261
0.0780 1.9790 5.280 8.266 0.9055 0.7486
0.0865 1.5870 4.492 7.918 0.7003 0.5545
0.0950 0.9884 2.438 0.1985 0.3038 0.6033
Table 7.4 Total current at different cathode-diaphragm coordinates

z (cm) I 10−5 (A)
r1 = 0.003 cm r1 = 0.01 cm r1 = 0.025 cm r1 = 0.03 cm r1 = 0.05 cm
0.0100 1.0920 1.0900 1.08600 1.0850 1.0970
0.0185 1.0240 1.0210 1.01300 1.0130 1.0300
0.0270 0.9423 0.9385 0.92810 0.9276 0.9553
0.0355 0.8427 0.8372 0.82530 0.8263 0.8727
0.0440 0.7203 0.7134 0.70320 0.7079 0.7856
0.0525 0.5730 0.5645 0.56170 0.5735 0.7023
0.0610 0.4033 0.3932 0.40690 0.4304 0.6419
0.0695 0.2261 0.2170 0.25660 0.2964 0.6464
0.0780 0.7486 0.7258 0.13830 0.2012 0.9293
0.0865 0.5545 0.6183 0.08938 0.2299 3.9530
0.0950 0.6040 0.8434 2.46900 13.3100 122.4000
within the tenth part of the apex-peak curvature radius from its surface, this table
practically shows the dependence of the emission characteristics on the apex length
if expressed as l – z1. Table 7.4 shows dependence of total current (I) on the
cathode-diaphragm coordinates. Table 7.4 demonstrates, firstly, an ambiguous
dependence of the total current on the cathode-diaphragm coordinates, and sec-
ondly, that this diaphragm can change the current within ten orders of magnitude.
Figure 7.12 shows the total-current dependence on the arrangement of the third
diaphragm, viz. the one following the anode diaphragm (r3 ; z3 ). The curves show a
weak dependence of the current on the parameters of the third diaphragm.
Figure 7.13 shows independence of the current values from the arrangement and
the potentail of the fourth diaphragm. So if the anode voltage is given, the most
profound impact on the total-current value is made by the arrangement and the
radius of the cathode-diaphragm, viz. r1 ; z1 . Therefore, let us assume the x1 ¼
r1 ; x2 ¼ z1 as optimization parameters.
Selection of the optimization method should be based on the assumption that we
are dealing with a constrained problem because physical considerations prevent us
from arbitrary/unconstrained variation of the parameters. Moreover, it is expedient
to resort to the methods based on the search only by the criterion function and the
constraints, i e. the direct search methods, because the function itself and the
constraints have a rather complicated dependence on the parameters.
I· 10-4, A
U2/U3=1
7.416
7.412
U2/U3=0.1
7.408
7.404
7.400
0.10 0.15 0.20 0.25 0.30 0.35 (r3, z3), cm
Fig. 7.12 The total-current dependence on the arrangement of the anode diaphragm (the third)
I·10-4, A
7.40
7.39
0.15 0.20 0.25 0.30 0.35 0.40 z4, cm
Fig. 7.13 Dependence of the current values from the arrangement of the fourth diaphragm
Let us stop on the method of complexes [44]. It has a quite effective algorithm
making it possible to apply a direct search by simplex to solution of
inequality-constrained problems.
The general problem of conditional optimization has constraints represented by
inequalities or equalities, as well as higher and lower values of variables.
As the cathode-diaphragm aperture radius should exceed r0 ðz1 Þ, i.e. the distance
from the axis to the apex surface, let us write out the constraints on variables x1 ; x2 :
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ffi
u" 2 #
u z þ a l
x1 \br t
1 z
; 0 x2 l;
az
where
R0 l2 R0 l
az ¼ ; br ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi :
2lR0 r 2 2lR0 r 2
As the argument of Nordheim elliptic functions ranges from 0 to 1, it is nec-

essary to impose a constraint on the value of field intensity at the apex peak, viz.
E\6:96 106 U2 .
Paper [44] suggests implementing the method of direct search by complex with
plotting a set of sampling points P at random and in succession. The pseudorandom
numbers evenly distributed over segment (0,1) are generated within the given
boundaries of variables x1i and x2i , after which the respective coordinates of the point
are determined by the following formula, viz.

xi ¼ x1i þ tt x2i x1i :
Every resulting point is checked for admissibility, and if any constraint is vio-
lated, the point is moved to the center of gravity of already plotted points until the
admissible point results.
The set of points P having been plotted, each of them is calculated for the
criterion function, and the point with the maximum function is rejected. A new
point is obtained by mapping of the excluded point through the center of gravity of
other points. If xR is the excluded point and x is the centre of gravity of other points,
then a new point is determined as

xk ¼ x þ a x xR :
Parameter a sets the mapping distance.

Let us assume that initial strictly admissible point x0, mapping parameter a and
calculation-termination parameters e and d have been determined.
Step 1 Plotting of the initial complex consisting of P admissible points. For
every point p ¼ 1; . . .; P 1, it is necessary to
(a) determine coordinates xpi in a random way;
(b) if xpi is an inadmissible point, find the center of gravity x of already found points
and put xpi ¼ xpi þ 1=2ðx xp Þ; re-apply the procedure until xpi is admissible;
(c) if xpi is an admissible point, re-apply (a) until p = P;
(d) calculate f ðxp Þ for p ¼ 1; . . .; P 1.
Step 2 The complex is mapped as follows.
(a) select point xR meeting the condition of f ðxR Þ ¼ max f ðxp Þ Fmin ;
(b) find center of gravity x and new point xk ¼ x þ aðx xR Þ;
(c) if xk is an inadmissible point and f ðxk Þ Fmax , halve the distance between xk
and x centre of gravity until f ðxk Þ\Fmax ;
(d) if xk is an admissible point and f ðxk Þ\Fmax , move to Step 4;
(e) if xk is an inadmissible point, move to Step 3.
Step 3 Admissibility is provided by the following adjustment, viz.
(a) if xki \x1i , then put xki ¼ x1i ; if xki [ x2i , put xki ¼ x2i ;
(b) if xk is an inadmissible point, halve the distance to the center of gravity;
re-apply the procedure until xk is admissible.
Step 4.
Check of the calculation-completion conditions
(a) Let
X 1X p
f ¼ 1 f ðxp Þ; x ¼ x ;
P P
P P
(b) if ðf ðxp f Þ2 e and ðxp xÞ2 d; stop the calculations; other-
wise, move to Step 2(a).
Paper [44] recommends selecting a = 1.3 and P 2N, N being the problem
order. If the new point is beyond the boundaries of the variables, the respective
coordinate is considered to be equal to the boundary value. Selection of a > 1
compensates for the complex compression caused by halving the distance to the
center of gravity. The large number of peaks is used to prevent degeneration of the
complex, when the search takes place near the boundary of the admissible area.
Results of optimization of the investigated system at different anode voltages and
positions of the anode diaphragm are given in Table 7.5.
7.6 Computation of Electron Paths in a Field-Emission Cathode-Based System 411
Table 7.5 Optimized geometrical parameters of the cathode diaphragm

U2 (V) r1 10−2 (cm) z1 10−2 (cm) z2 (cm) r1 10−1 (cm) z1 10−1 (cm)
410 1.1270 9.5630 0.10002 0.5267 0.2199
430 0.5288 8.2800 0.10006 0.4840 0.1566
440 1.3350 8.6570 0.10010 0.4230 0.1519
460 1.6620 6.9950 0.10015 0.2402 0.1618
480 2.9120 8.3020 0.10020 0.7453 0.3276
490 1.3150 2.9700 0.10030 0.4940 0.2175
510 8.4330 0.5072 0.10100 0.8613 0.2783
7.6 Computation of Electron Paths in a Field-Emission

Cathode-Based System
Let us consider a sheet of a two-sheeted hyperboloid of revolution as the

cathode-surface approximation. The equation of such a surface is conveniently
represented by oblong revolution-ellipsoid coordinates k, l, u. The communication
between the Cartesian coordinate system and the oblong revolution-ellipsoid
coordinate system is expressed by
8
< x ¼ a0 sin kshl cos u;
>
y ¼ a0 sin kshl sin u; ð7:66Þ
>
:
z ¼ a0 cos kchl;
where k 2 [0, p], l 2 [0, ∞) and u 2 [0, 2p).

Let us consider the coordinate system (7.66) (Fig. 7.14). Fixation of coordinates
k = kc gives us a hyperboloid sheet, i.e. a surface intended to approximate the
cathode. The hyperboloid vertex is set by coordinate l = 0. At k = ka = p/2, the
hyperboloid degenerates into a plane intended to approximate the anode surface.
A particular value of coordinate l gives a family of ellipsoids. In the particular case
of (l = 0), the ellipsoid degenerates into segment [−a0, a0] lying on axis Oz.
Having fixed coordinate u, we get a number of half-planes coming out of axis Oz.
The components of the metric tensor of the coordinate system selected look as
follows:
gkk ¼ gll ¼ a20 ðsin2 k þ sh2 lÞ; guu ¼ a20 sin2 ksh2 l:
Other components are zero in virtue of orthogonality of the system selected.

Because of axial symmetry, the given expressions do not depend on coordinate u. It
is important to note that coordinate system (7.66) can be introduced in a different
way, the case considered being referred to as a special system. The advantage of
x
x
t
ns
co
=
0
0 ϕ
=
co
ns
t
a0
0 0
0
z
y
Fig. 7.14 Coordinate system of the oblong revolution-ellipsoid
z
Fig. 7.15 To calculation of
cathode geometric parameters
r0
c
=
= a = 2
0
0
x
special coordinates is that two non-zero components of a metric tensor are equal to
each other. Simulation of the emission system leaves such parametres as
cathode-surface coordinate kc and focal distance obscure a0. They can be calculated
if distance from the anode to the apex peak d and apex-peak curvature radius r0
have been set, d and r0 of a real emission system being determinable by experiment
(Fig. 7.15).
To calculate the hyperboloid-surface curvature, let us write out the coefficients of
the first and second basic quadratic forms, viz.
E ¼ a20 ðsin2 kc þ sh2 lÞ; F ¼ 0; G ¼ a20 sin2 kc sh2 l;

sin kc cos kc sin kc cos kc
L ¼ a0 pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ; M ¼ 0; N ¼ a0 pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi sh2 l:
sin kc þ sh l
2 2
sin2 kc þ sh2 l
The average curvature is distributed over the surface as follows.
1 2 sin2 kc þ sh2 l
HðlÞ ¼ ctgkc : ð7:67Þ
2a0 ðsin2 kc þ sh2 lÞ3=2
Let us equate expression (7.67) at the cathode apex to the reciprocal value of the
curvature radius, viz.
cos kc 1
Hð0Þ ¼ ¼ :
a0 sin2 kc r0
If we assume that kc 2 (0, ka) and consider half-space z 0, the distance from
the cathode apex to the anode appears equal to
d ¼ zjl¼0 ¼ a0 cos kc :
The resulting equations give us

pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
a0 ¼ dðd þ r0 Þ; ð7:68Þ
pffiffiffiffiffiffiffiffiffi
kc ¼ arctg r 0 =d ð7:69Þ
The parameters of the simulated emission system were represented with the
following values: d = 2 cm, r0 = 1 µm (Fig. 7.16). The curvature distribution over
the cathode surface meeting the parameters taken is shown in Fig. 7.17. The cur-
vature can be seen to quite rapidly decrease with growth of curvilinear coordinate.
With the parameters specified, the system is a classical field-emission diode, which
makes it possible to consider the electron-path plotting problem. For its solution
and determination of the electric field intensity near the apex, we have to know the
potential-energy distribution generating the force field in the gap between the
cathode and the anode.
Let us assume that the gap between the electrodes contains such a low volume
concentration of electrons that potential distribution follows the Laplace equation:
Du ¼ 0
with the boundary conditions of

z, cm
z, cm
3 3
2
2
1
0
−1 1
−0. m 0
5 0.5 y,
0 0
0.5 5 −2 −1 2
−0. 0 1
x, 1 1 x , cm
m −
Fig. 7.16 Model of field emission system
1 1
Fig. 7.17 Distribution of H , m H, m
curvature and its radius over 12.5 1.0
the cathode surface
10 0.8
7.5 0.6
5 0.4
2.5 0.2
r0
0 0.0
0 1 2 3 4
(
0; if k ¼ kc ;
u¼
u0 ; if k ¼ ka :
The boundary condition meets the equipotentiality of the cathode and anode
surfaces, i.e. the electrodes are metals. In the coordinate system selected, the
equation is as follows:

@ @u @ @u sin2 kc þ sh2 l @ 2 u
sin kshl þ sin kshl þ ¼ 0:
@k @k @l @l sin kshl @u2
As the boundary conditions do not stipulate any dependence on angle u and

curvilinear coordinate l, the solution should not depend on these variables, either.
Consequently, the equation can be simplified as

@ @u
sin k ¼ 0:
@k @k
The general solution is given by
k
uðkÞ ¼ c1 ln tg þ c2 ;
2
tg k/2 > 0 at k 2 [kc, ka]. The boundary conditions determine the integration
constants
" #
ln tg k2
uðkÞ ¼ u0 1 : ð7:70Þ
ln tg k2c
As a result, the electron potential energy is expressed by U(k) = − eи(k). Its

distribution in the Cartesian coordinates is shown in Fig. 7.18 for the value of
u0 = 20,000 V. It is easy to understand that the equipotential surfaces in our
problem are also sheets of a two-sheeted hyperboloid of revolution.
The gradient of expression (7.70) taken with the negative sign gives the fol-
lowing intensity of electric field E in the interelectrode space:
1 du u0 1
Eðk; lÞ ¼ pffiffiffiffiffiffi ek ¼ k
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ek : ð7:71Þ
gkk dk a0 ln ctg 2 sin k sin2 k þ sh2 l
Fig. 7.18 Potential energy

distribution
0
10
15
20
1
3
0.5 m
0
2 z, c
0.5 1
x, 0
m 1
Fig. 7.19 Equipotential lines z , cm

and force field
2 1 0 1 2
x , cm
The electron that has got into the gap between the cathode and the anode finds
itself in force field − eE (Fig. 7.19; u0 = 20000 V, equipotential lines being drawn
through 1 kV). The direction of the field along vector –ek indicates that the field
lines result from intersection of the planes passing through axis Oz and ellipsoids,
i.e. they are the arcs/quarters of the ellipses with same focal distances a0.
On the cathode surface at k = kc, the intensity module assumes the following
values:
u0 1
EðlÞ ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi : ð7:72Þ
a0 sin kc ln ctg k2c sin2 k þ sh2 l
The electric field intensity on the cathode surface is critical for field emission of
electrons. Its dependence on curvilinear coordinate l is shown by Fig. 7.20 for
different apex-peak curvature radii at the same voltage between the cathode and the
anode. The results show that value E quite rapidly decreases from the cathode peak
to its periphery. The intensity distribution gives an idea of the geometrical factor of
field b.
The result is the mathematical model of a pointed cathode-based field-emission
system. Within the framework of the model, it is relatively easy to get the distri-
bution of the potential energy of an electron in the interelectrode space. The geo-
metrical parameters specified make it possible to consider the electron-path plotting
problem. In addition, we get the distribution of electric field intensity over the apex
surface.
To obtain the particle paths, the Lagrange equations of the 2nd kind is formu-
lated and solved in numerical terms. The solution is based on the Runge–Kutta
method of the 8th order.
Fig. 7.20 Electric field E · 109 , V/m

intensity on the cathode
r 0 =0.5 m
surface r 0 =1.0 m
3.5
r 0 =2.0 m
3.0
2.5
2.0
1.5
1.0
0.5
0 1 2 3 4
2
Fig. 7.21 The paths of the cm

electrons that have left the
cathode at field emission 0 0 0
3 1011 0 0
0 1010 0
0 0 1013
2
0 1.5 m
0.5 1
1 0.5
1.5 0
m
Figure 7.21 shows the paths of the particles that have left the cathode under
different starting conditions, as well as the field line. The starting conditions are
selected so that to show the effect of change of the generalized velocities on the path
behavior, no establishment of the physical sense of these values having been
planned within the framework of the paper. Obviously, the emission-system
parameters considered do not make it possible to approximate the paths with field
lines.
Let us vary the relation of cathode-curvature radius r0 to electrode spacing d to
determine what system parameters make it possible to replace the paths with force
lines. It has already been indicated in the preliminary survey to the paper that it is
important to monitor the intensity on the cathode surface because the metallic
Table 7.6 Error of path d (lm) r0 (lm) r0/d lmax l0 (kV) Error (%)
replacement with force lines
5 50.0 10.0 0.60 16.5 3
5 25.0 5.0 0.57 15.5 6
5 5.0 1.0 0.45 10.9 21
5 2.5 0.5 0.37 8.2 32
bonding in the sample is broken at the values of about 1010 V/m thereby destroying
the sample.
Relation between the electric field intensity at the apex peak and the voltage
between the cathode and the anode is given by formula (7.72). Let us fix the
intensity value at E = 3.5 109 V/m. The parameters used, viz. d = 2 cm and
r0 = 1 µm determine the emission-image area boundary (lmax = 0.005) that meets
the reduction of E by about 1.2 times. Proceeding from this information, we
recalculate voltage u0 and boundary value lmax when varying relation r0/d.
Table 7.6 shows the error of path replacement with force lines calculated as a
relative deviation of the finishing points of extreme path (kc, l, u) at zero initial
velocities. The electrodes are spaced 5 µm apart. The error can be seen to become
admissible only at r0/d = 10.0. Actually, that brings us to flat geometry.
The computations made show that the projections of cathode-surface point (kc,
l, u) to the anode by force lines and the paths differ no more than twice under
rather adverse conditions. Such a result can only be achieved by increase in the
cathode-curvature radius with a consequent necessity to step up the voltage between
the cathode and the anode, which is undesirable. Another way out is to reduce the
space between the cathode and the anode, which reduces the significance of the
path-problem solution since small linear dimensions of the system, in the order of
fractions of micrometer, suggest solving such problem by quantum-mechanics
methods, i.e. in terms of probabilities.
The operation of the path-plotting algorithms suggested can be demonstrated by
simulation of an emission image in a field-electron microscope. Then it is necessary
to have an idea about current-density distribution over the cathode surface. Let the
image brightness be proportional to the current density. Hence, we face the problem
of projection of the current-density distribution to the anode with the paths. Let us
compute the current density by the following formula:
Zþ 1
j¼e mðEf ÞDðEf Þdef :
ðEF þ UÞ
This formula operates a potential-barrier transmission coefficient. Let us simulate

the potential barrier and calculate the probability of its transmission.
We construct the dependence of potential energy U on k by formula (7.70)
(Fig. 7.22). The potential energy can be seen to behave as a homogeneous field in
the neighborhood of kc i.e. at the distances of about a few hundreds of nanometers
from the cathode.
Fig. 7.22 Potential energy U, eV

near the cathode 1 2 3 4 5 6
( – c) 10 –4
0
–50
–100
–150
–200
–250
–300
0 20 40 60 80 100
(d – z ), nm
The distribution of the curvature over the cathode surface with the parameters
selected shows (Fig. 7.17) that the surface segment is considered to be flat for the
tunnel effect. Let us introduce local coordinate f normally counted off from the
cathode surface. Then the potential barrier along the electron path can be described
as the following dependence:
e2 1
UðfÞ ¼ eEf :
16pe0 f
Let us expect smoothness of the potential energy in the neighborhood of the

metal–vacuum interface and neglect the image forces at significant distances from
the interface. Then the potential barrier can be described by the formula, viz.
ð7:73Þ
Here Ф depends on the point where an electron escapes from the cathode sur-
face, viz. its curvilinear coordinates; E is determined by formula (7.72).
Smoothness of the potential energy in point f = 0 is achieved automatically.
Constants f1 and G are derived from the equations, viz.
e2 1 e2 1
EF U þ Gf21 ¼ eEf1 ; 2Gf1 ¼ eE þ :
16pe0 f1 16pe0 f21
Let us introduce the following notation, viz.
e2 EF þ U
Q¼ ; c¼ :
16pe0 eE
f1 is featured by a quadratic equation. It has a real solution under the condition

that
ðEF þ UÞ2
eE\ ;
3Q
which holds true in all physically implemented cases. Even if the total value EF þ U
of makes 5 eV, the field intensity has to be kept below 2 1010 V/m to prevent the
emission from displaying its explosive nature. Then we select the smaller root; the
larger root yields a negative value for G, whereas parabola EF Uðl; uÞ þ Gf2
has to have upward branches (Fig. 7.23). The result is the sought-for values:
"sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi# " #
3Q 1 Q
f1 ¼ c 1 1 ; G¼ eE :
c ðEF þ UÞ 2f1 f21
Distance f2 can be found from the condition of smallness of the polarization

additive

Q=f2

eEf Q=f ¼ e\1:
2 2
Fig. 7.23 Potential barrier U(ζ ), eV
–2 Φ
–4
–3
–6 – ξF – Φ + Gζ 2
ξF
–8
–11
–10 –0.1 0 0.1
0 2 4 6 8 10
ζ , nm
So if the distances from the cathode surface exceed

sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

Q 1
f2 ¼ 1 ;
eE e
the force field for the electron can be considered as homogeneous with a relative
error under e.
Let us note that values f1 and G depend on coordinates l and u; f2 depending on
l and error e.
To get the current-density distribution over the cathode surface, we integrate the
external distribution of electrons, which is intrinsically the product of the delivery
function vint by the transparency of the potential barrier considered. The areas under
the curves being equal to one, the dependence of the normalized external distri-
bution vext on the energy for different field intensities at the cathode E at ambient
temperature and temperature T at intensity E = 3.5 109 V/m is presented in
Fig. 7.24. Normalization was based on the compound formula of trapezoids.
To find the current density, the external distribution of electrons has to be
integrated over all possible values of energy in the same way as in the course of
normalization. But here the integration was based on the Monte–Carlo method. This
method provides a great number of implementations of a random variable formed
so that its mathematical expectation coincides with the exact value of the solution of
the problem at hand. The variance of the probabilistic assessment of the problem
solution depends on the number of statistical tests. In our case, random points were
generated in the area considered and the function values in them were averaged.
Selection of this integration method makes it possible to simulate the
Fig. 7.24 External ν ext, eV –1

distribution of electrons
5 0.5 · 109 V/m
1.0 · 109 V/m
4 3.5 · 109 V/m
5.0 · 109 V/m
10.0 · 109 V/m
3
0
EF Φ
2 0K
293 K
1 1000 K
2000 K
3695 K
0
–10 –8 –6 –4 –2 0
ξ,
Fig. 7.25 Selections of j · 10 6 , A/m 2

current density values
2.4
j =1.31·10 6A/m 2
2.0
1.6
1.2
0.8
0.4
0
10 10 2 10 3
N
current-density fluctuations at the cathode. The formula to compute values j looks

as follows.
ZEmax
eðEmax Emax Þ XN
j¼e mext ðEÞ dE mext ðEi Þ;
N i¼1
Emin ð7:74Þ
Ei 2 ðEmin ; Emax Þ;
Ei ¼ Emin þ ðEmax Emin Þ c; c 2 ð0; 1Þ:
For the current-density values to be random, each of them has to be calculated

with preliminary redefinion of random abscissas Ei of quadrature formula (7.74).
Random variable c was simulated with a Mersenne Twister pseudorandom
number generator [46]. Figure 7.25 shows dependence of the spread of current
densities j on test number N, i.e. the higher N, the closer are all the simulated current
densities to the average value j = 1.31 106 A/m2.
To construct the emission pattern, let us assume that current density j is deter-
mined by the electrons escaping from the cathode-surface point identified by (kc, l,
u). Let all electrons leave the cathode with identical starting conditions. Let us
assume the initial generalized velocities to equate zero.
Coordinate l ranges from 0 to lmax = 0.05, u does so within the applicable
domain, i.e. segment [0, 2p]. The peak value of is selected by the emission-image
luminosity gradation. The distances between the points in curvilinear coordinate
system Dl and Du are also selected so that the points lying within the circle
outlining the emission image are distributed as homogeneously as possible.
Fig. 7.26 Emission image, y, μm j⋅106, A/cm2

N = 103 100 1.2
50 0.9
0 0.6
-50 0.3
-100 0.0
-100 -50 0 50 100
x, μm
The work function in point (kc, l, u) is calculated according to the algorithm

suggested in paper [45, 47]. The electric field intensity is calculated by formula
(7.72). Current density
Zþ 1
j¼e minx ðEf ÞDðEf ÞdEf
ð EF þ U Þ
has already been mentioned as integrated by the Monte–Carlo method.

Barrier-transmission coefficient DðEf Þ (7.73) is determined numerically by solution
of the Schrödinger equation with the Runge–Kutta–Nyström method of the 12th
order.
The result is the emission current-density distribution over the anode surface by
electron paths. Image brightness is considered as a value proportional to the current
density.
Figure 7.26 shows the emission pattern for the case of N = 103. It has a rela-
tively high definition. Real emission images are more blurred. The reason is
presence of imperfections on the cathode surface, such as adatoms etc. Various
defects make the work function and the electric field intensity near the cathode
surface experience drastic changes. In its turn, that affects the current density from
this surface area. Illustratively, Fig. 7.27 presents various emission images with a
different number of statistical tests (N = 10, 102, 103). As expected, the integration
method selected makes it possible to achieve a pattern claiming to be a simulation
of current-density fluctuations as the image becomes more blurred.
Fig. 7.27 Emission images, N = 10, 102, 103
References
1. V.P. Ilyin (ed.), Algorithms And Methods for Calculating the Electron-Optical Systems
(Novosibirsk, 1983), p. 190
2. L.A. Baranova, S.Y. Yavor, The electrostatic electron lenses. M.: Nauka, 192 (1986)
3. V.P. Ilyin , Numerical methods for solving the electro-optics problems. M.:Nauka, 202 (1974)
4. N.D. Devyatkova (ed.), Methods for calculating the electron-optical systems. M.: Nauka, 174
(1977)
5. V.P. Ilyin (ed.), in The Methods of Calculating Electron-Optical Systems: Proceedings of IV
All-Union Sem (Novosibirsk, 1982), p. 202
6. A.P. Drum, V.V. Bulavinov, P.P. Konorov, Non-equilibrium electronic processes in layers on
the surface of silicon. in Molecular and Electronic Processes on the Interphase Boundaries.
(L.: Publishing of Leningrad University, 1989), p. 3–27
7. V.A. Syrovoy, Calculation of forming the electrodes in the optical planar electron beams.
Radiotehnika i Elektronika 39(3), 481–502 (1994)
8. V.A. Syrovoy, Calculation of forming the electrodes in optics rotationally symmetric electron
beams. Radiotehnika i Elektronika 39(4), 666–687 (1994)
9. L.E. Tsyrlin. Selected problem for electric magnetic fields. M.: Soviet Radio 319 (1977)
10. L.A. Neganov, V.A. Syrovoy, V.N. Tskhai, Calculation and experimental investigation of the
electron gun for technological purpose. Radiotehnika i Elektronika 35(10), 2146–2155 (1990)
11. V.V. Plohov, V.A. Syrovoy. On calculation of multi beams injector with a relativistic electron
beam. Radiotehnika i Elektronika 35(12), 2582–2593
12. V.A. Syrovoy, On the synthesis of nonparaxial relativistic electron beams at emission limited
by temperature, and the injection of a non-zero velocity. Radiotehnika i Elektronika 42(3),
348–360 (1997)
13. Y. Xuebiao, X. Zhangchcheng, H. Guoguang et. al. Emission characteristics of the
molybdenum-coated silicon field emitter array. J. Phys. D: Appl. Phys. 29(3), 506–510 (1996)
14. M.I. Yavor, E.V. Strigova, Field distribution and electrical properties of electrostatic conical
slit lenses. Nucl. Instr. Meth. Phys. Res. A 363(1/2), 445–450 (1995)
15. L.N. Baranova, N.S. Ulyanova, S.Y. Yavor, The study of spatial and temporal characteristics
of the charged particle beam focused by axisymmetric electrostatic lens. ZTP 61(2), 144–148
(1991)
16. I.M. Gordion, I.D. Tokman, The problem of electrostatics to an oblate spheroid in the point
charge field. ZTP 67(2), 121–122 (1997)
17. G.Ch. Shushkevich, The electrostatic problem for the disk and torus. ZTP 67(4), 123–126
(1997)
References 425
18. V.V. Rassadin. Calculation of the electric field in periodic structures of linear ion accelerators.
in A system of linear accelerators and the use of beams of charged particles. (M.: Publishing
House of -the Moscow Engineering Physics Institute, 1987), p. 41–45
19. L. Klemar, E. Kasper, On the numerical design of elecnron guns. Optic. 72(1), 29–30 (1985)
20. S.K. Godunov, E.I. Romenskii, G.A. Chumakov, in Construction of numerical grids in
complex domains using quasi-conformal mappings. (Novosibirsk: Nauka, 1990), pp. 75–83
21. A.Taflove, Emerging applications for FD-TD computation. IEEE on Comp. Sci. Eng. 2(4),
24–34 (1995)
22. S. Roques, M. Denizart, F. Sonier, Tetrode field emission guns for electron microscopy.
Optic. 61(1), 51–66 (1983)
23. M.I. Yavor, E.V. Strigova, Field distribution and electrical properties of electrostatic conical
slit lenses. Nucl. Instr. Meth. Phys. Res. A 363(1/2), 445–450 (1995)
24. G.V. Lebedev, N.A. Timchenko, Complex of programs for design of electron-optical systems
“Probe.”. Sci. Instrum. Electron-ion Optics L.: Nauka, 111–116 (1989)
25. N.V. Egorov. Calculation and optimization of the characteristics of the field emission electron
source. in Modeling and Control in the Electrical and Mechanical Systems, ed. by D.A.
Ovsyannikov (St. Petersburg: St. Petersburg State University Publishing House, 2002),
pp. 111–127
26. E.M. Vinogradova, N.V. Egorov. Calculation of the electrostatic field systems with coaxial
axially symmetric electrodes. Radiotehnika i Elektronika 52, 225 (2007)
27. A.A. Almazov, N.V. Egorov, Optimization of multi-tips field source. Vacuum 52, 295 (1999)
28. A.A. Almazov, N.V. Egorov, Optimization of multi-tips emission systems. Radiotehnika i
Elektronika 40(4), 638–643 (1995)
29. E.M. Vinogradova, Calculation of electron-optical systems based on field cathodes, in
Modeling and Control in the Electrical and Mechanical Systems, ed. by D.A. Ovsyannikov
(St. Petersburg: St. Petersburg State University Publishing House, 2002), pp. 99–110
30. E.M. Vinogradova, N.V. Egorov, R.Y. Baranov, in Calculation Of The Electrostatic Potential
In Single Tip And Multi-Tips Emission Systems, 1 (Vestnik of St. Petersburg State University,
Ser. 10, 2007), p. 39
31. E.M. Vinogradova, N.V. Egorov, R.Y. Baranov, Mathematical model of the cathode
assembly of the electron gun field, (Vestnik of St. Petersburg State University. Ser. 10, 2006),
p. 3
32. E.M. Vinogradova, N.V. Egorov, Calculation of the electrostatic field of spherical segments.
ZTP 78(8), 128 (2008)
33. K.A. Krimskaya, Mathematical modeling of the formation of the electron beam on the basis of
the field cathode. Diss. Cand. SPb, p. 85 (2009)
34. N.V. Egorov, E.M. Vinogradova, Mathematical model of electron gun on the field emission
electron cathode basis. Vacuum 57, 267–281 (2000)
35. N.V. Egorov, E.M. Vinogradova, Mathematical modeling of electron beam formatting
systems on the basis of field emission cathodes with various shapes. Vacuum 72, 103–111
(2004)
36. D.A. Ovsyannikov, N.V. Egorov, Mathematical modeling of systems of formation of electron
and ion beams (St. Petersburg State University Publishing House, St. Petersburg, 1998),
p. 276
37. N.V. Egorov, A.G. Karpov. Diagnostic information and expert systems (SPb.: Publishing
house of St. Petersburg University, 2002), p. 470
38. E.P. Shurina, M.U. Velikaya, M.P. Fedoruk, Algorithms for solving Maxwell’s equations on
unstructured grids. Comput. Technol. 5(6), 99–116 (2000)
39. A.A. Almazov, N.V. Egorov, By the method of calculating the field-emission systems.
Radiotehnika i Elektronika 31(12), 2452–2458 (1986)
40. Almazov A.A., Egorov N.V. Mathematical model of field-emission diode. Mathematical
methods of modeling and analysis of controlled processes.( L.: Publishing of Leningrad
University, 1989), pp. 20–28
41. A.A. Almazov, E.M. Vinogradova, Egorov N.V. Mathematical model of the electron gun
with field cathode. Abstracts of 12th All-Union. Sem. on linear particle accelerators. Kharkov,
1991. p.39
42. P.F. Bovey, A 100 kV high resolution analytical STEM. Mirrosc. Spectrosc. Electron. 1,
507–508 (1976)
43. G.F. Vasiliev, Influence of the form of the potential barrier at the emitter-vacuum and electric
field distribution on the surface of the emitter in the form of current-voltage characteristics of
field emission. Radiotehnika i Elektronika 5(11), 1857–1861 (1960)
44. G. Rekleytis, A. Reyvidran, K. Regsdel. Optimization in Technology. M.: Mir 2, 1986
45. N.S. Demchenko, A.Y. Antonov, Construction of the electron trajectories in the emission
system. , 3 (Vestnik of St. Petersburg State University, Ser. 10, 2010), pp. 1–10
46. M. Matsumoto, T. Nishimura, Mersenne Twister: A 623-dimensionally equidistributed
uniform pseudorandom number generator. ACM Trans. Model. Comput. Simul. (Special
Issue on Uniform Random Number Generation) 8(1), 3–30 (1998)
47. I.M. Gribkova, in Mathematical Modeling of the Images in a Field Electron Microscope. eds.
by N.V. Smirnov, G.S. Tamasyan. The Control Processes and stability. Proceedings of XXIX
Scientific Conference of Applied athematics and Control Processes Faculty of Saint
Petersburg State University (Publishing House of St. Petersburg State University, St.
Petersburg, 2008), pp. 116–121
Chapter 8
Field Emission Cathode-Based Devices
and Equipment
Abstract In this chapter the main features of field cathode-based devices are
reviewed. Such devices include various lighting sources, microwave appliances,
X-ray tubes and other electronic devices.
8.1 Light Sources
Light sources are an integral part of our life. The need for artificial illumination and
data-display devices steadily grows. A large number of light systems designed with
account of specific requirements to color rendition, luminosity and illumination
levels are under development. We know a wide class of light sources, viz. con-
ventional incandescent, halogen and fluorescent lamps intended to illuminate pre-
mises and advertisements, high-pressure gas discharge lamps intended to illuminate
large spaces and streets, semiconductor light-emitting diodes, i.e. LED, and organic
light-emitting diodes, i.e. OLED, including illumination systems, video-wall light
elements and cathodoluminescent lamps. Nevertheless, every light source has an
inherent weakness, e.g., an imperfect emission spectrum, a long setup time or an
insufficient efficiency factor.
The latest review of cathodoluminescent light sources it presented in [1].
Table 8.1 gives comparative characteristics of a small-button glass field emis-
sion cathode-luminescent light source (cf. Sect. 8.2.2) compared to industrially
manufactured light sources.
Summary: a field emission light source is comparable with the best up-to-date
light sources in terms of its parameters, but surpasses them in the working-
temperature range, luminescence uniformity, low production price, use of inex-
pensive, available and naturally widespread materials, as well as high ecological
compatibility.

428
Table 8.1 Comparative characteristics of a small-button glass field-emission cathodeluminescent light source
Basic parameters Field Incandescent Luminescent Halogen light bulbs Metal-halogen light bulbs LED lamps
emission light bulbs light bulbs
light source
Lamp type No Incandescent Gas-discharge Incandescent Gas-discharge/arc Semiconductor
incandescent
vacuum
Efficiency (lm/W) 30–60 10–12 40–60 22 80 *30
Energy conversion efficiency 20–30 5 *20 10 *30 20–30
8
(%)
Power-up time (s) 0.0005 0.1 1–3 0.3 3–4 0.0005
Voltage drop sensitivity Low Low High Medium Medium High
Stability of light current High Medium High Medium High High
Environmental damage High −196 to Low Medium Low Medium High −40 to
resistance +150 °C +85
Needs special voltage source Yes No Yes Yes Yes Yes
Lifetime (h) Up to 1000 10,000 180–600 15,000 50,000a
100,000a
Estimated costs for mass Medium Low Medium Medium High High
production
Materials used in production, Glass, steel, Glass, steel, Glass, steel, Glass, steel, tungsten, Glass, steel, tungsten, Indium,
exploitation and utilization carbon tungsten mercury bromine/iodine/chlorine bromine/iodine/chlorine, arsenic,
mercury gallium, glass
a
Lifetime was estimated using the common engineering method
Field Emission Cathode-Based Devices and Equipment
8.1 Light Sources 429
8.1.1 Operating Principle
The operating principle of an electroluminescent light source based on a field

emission cathode is luminous excitation of a luminophore caused by electron
bombardment. As light sources are to have certain properties, let us define some basic
notions adopted in light engineering and accommodate the existent differences [2].
Light flux is the light power, an effective value defined as
dQ
U¼ ;
dt
where Q is the emission energy.

A light-flux unit is a lumen (lm); 1 lm is a light flux emitted in a unit solid angle
by a pointed isotropic source with a luminous intensity of 1 cd, a candela being
defined below. A complex continuous-spectrum light flux is equal to
Z780
U ¼ 683 Uek ðkÞV ðkÞdk;
380
where n is the number of spectral lines; Uek ðkÞ is the spectral-flux density function;
(780–380 nm) is the visible-light wavelength range, nm.
Pointed-source luminous intensity is a space light-flux density,
dU
I¼ :
dX
Candela (cd) is a luminous intensity unit. A candela is equal to the luminous

intensity emitted from the black-body area of 1/600,000 m2 in the perpendicular
direction at the freezing-platinum temperature of T = 2045 K and the pressure of
101,325 Pa.
Brightness of a body or its surface area is the relation of luminous intensity to the
area in the perpendicular direction.
A unit of brightness is cd/m2 or nit, defined as brightness of such a planar surface
that emits luminous intensity of 1 cd from the area of 1 m2 in the perpendicular
direction.
Colorimetry guidelines
In terms of colorimetry, color as a physical value is defined as follows: Color is a
tri-dimensional vector value characterizing a group of emanations visually indis-
cernible under colorimetric observing conditions. This definition already contains
the principles of classification and measurement of colors. The color metric is based
on Grossmann’s laws of color mixture.
430 8 Field Emission Cathode-Based Devices and Equipment
Grossmann’s laws of color mixture are the following:

LAW I: Every color can be made up by tempered mixing of three colors, each
unobtainable by mixing the other two. In other words, color is determined by three
linearly independent variables that form the equality
C ¼ R½R þ G½G þ B½B;
where [R], [G], [B] are the unit quantities of the primary colors of the measurement
system; R, G, B are the fractions of the unit primary colors providing color equality,
i.e. the color coordinates.
LAW II: Color continuously changes with a continuous change of spectral-power
distribution.
LAW III: The mixture color depends only on the colors of blending agents and does
not depend on their spectral distributions, the emission property being termed
metamerism.
Grossmann’s laws generate an idea of the color vector having three components
of the same origin and different directions in space. The aggregate of the tridi-
mensional color vectors makes color space (Fig. 8.1).
Every color corresponds to only one vector in color space. The colors lying in
one line coming from the origin of coordinates, i.e. differing among themselves
only by intensity, are characterized by the same chromaticity. Thus, chromaticity
indicates the direction of the color vector in space and depends on two coordinates.
The chromaticity coordinates are described by a color triangle, which represents
a section of a three-coordinate space passing through the unit colors of the mea-
surement system selected. The chromaticity coordinates are relative values deter-
mining the position of a point in the triangle.
Basically, the existing light-source versions can be divided into two main classes
(Fig. 8.2) [3], viz. those functioning in the mode of transmission and reflection.
Generally, the basic classical version is similar to a usual field emission microscope.
Fig. 8.1 RGB and XYZ

color models
Fig. 8.2 Principles of light source designs: a, b based on clearance; c based on reflection. 1 Core
field emission cathode; 2 modulator; 3 flow electrons; 4 phosphor; 5 transparent conductive
coating; 6 output glass; 7 visible light; 8 aluminum coating; 9 grid; 10 cathode matrix; 11 substrate
It consists (Fig. 8.2a) of a rod-type field emission cathode 1. Such a cathode can be
represented with any configuration, be it a graphite rod, a carbon fiber bundle or
nanostructures at the end face of an appropriate rod. Electrons are extracted with a
metal orifice plate. The cathode-modulator angle configuration is selected so that
three mutually exclusive factors are optimized, viz. the maximum current trans-
mission through the modulator, the minimum control voltage and the maximum
uniformity of electron flow over the screen surface.
The latter is most easily achieved by overlapping of field emission flows from
the sufficient number of emission centers. That is achieved by application of several
rod-type field emission cathodes and respective modulator orifices arranged uni-
formly in a circle to form a round-section light source.
In such classical light sources, luminophore brightness is used only partially (up
to 30%), because some photons go back to the envelope. Better brightness can be
obtained if one would switch the positions of luminophore and conductive coating
and if the conductive coating is made of aluminum (Fig. 8.2b). An aluminum
coating increases brightness up to 70% of maximum possible value.
The maximum efficiency is achieved when the light source is in a reflection
mode [4] (Fig. 8.2c). In this case, electrons 3 travel to luminescent layer 4 deposited
on aluminum mirror 8. In this case, the loss of light is only stipulated by negligible,
viz. less than 10%, light absorption in target glass 6 and transparency of matrix
cathode 10. The latter can be increased up to the level of less than 10%.
The particular technological requirements of practical application stipulate
selection of one of these principles or their modification.
The below covers different light-source versions.
8.1.2 Small-Button Glass Cathodoluminescent Lamps
A small-button glass cathodoluminescent lamp is designed to make the brightest

light source. It can provide a high-voltage electron flow of up to 10–15 kV with
current of up to 1 mA. Such lamps can be more than 100,000 cd/m2 bright.
The arrangement drawing of such a lamp is shown in Fig. 8.3.
The electrons emission by field emission cathode 1 driven by the resultant
electric field of control electrode 2 and anode 3 reach the luminescent screen, thus
causing its glow.
A particular structure of such a source is shown in Fig. 8.4 [5–7].
The light source consists of vacuum envelope 1, anode 3, 4, modulator 5, contact
bumps 12, field emission cathode 9 made of a carbon-fiber bundle coated with a
dielectric, mostly glass, over the length. The field emission cathode is located in the
orifices of alignment disks 8, 10 having grooves along the perimeter tightly holding
canes 6 made mostly of glass and orienting disks so that the centers of their orifices
were coaxial with the modulator orifice, which is provided during the assemblage of
a cathode-modulator unit. Contact unit 11 of the field emission cathode is made of
an electrically conductive substance, Aquadag in particular, deposited on a
shell-free fiber bundle and a shell contacting with the bundle through the lateral
surface fixed to the contact bump of the field emission cathode.
Alignment disks provide cathode centering accuracy and set a distance from the
modulator to the cathode. The alignment disks are tightly connected to the mod-
ulator so that the centers of their holes are coaxial with each other and with the
modulator hole due to glass studs and at the same time connected to the slots of the
alignment disks and the modulator to form a mechanical connection. A field
emission cathode is put into the alignment-disk holes and tightly fixed to the lower
disk. Then the thus-assembled cathode-modulator unit is installed on a mounting
base by the modulator bumps, connected to the field emission cathode with contact
bumps, followed by hermetic sealing and vacuuming of the envelope.
The field emission cathode is a complex multi-tips system consisting of—300
PAN fibers (Fig. 8.5) bundled up and treated in discharge (cf. Sect. 6.2).
A disadvantage of the cathode-modulator unit shown in Fig. 8.4 is considerable
current hogging at the modulator (10–20%), i.e. a large part of the electron flow
from the field emission cathode gets to the control electrode, thereby reducing lamp
efficiency.
Fig. 8.3 Schematic design of

a cathode luminescence
miniature lamp: 1 cathode; 2
control electrode, 3
luminescent anode
Fig. 8.4 a Light source design with the field emission cathode from a bundle of carbon fibers: 1
bulb; 2 high-voltage input; 3 phosphor; 4 layer of aluminum; 5 modulator with a diaphragm; 6
glass rods; 7 emitting part of the bundle of carbon fibers; 8 justified disks; 9 vitrified carbon fiber;
10 adjustment disks; 11 contact unit of the field emission cathode; 12 terminals; b appearance of
the lamp
Fig. 8.5 The image of a

carbon fiber cathode from an
SEM
The working voltage at the modulator ranges from 1.5 to 2.0 kV for this
structure. Similar results are given in [8, 9], where field emission cathodes are also
represented with PAN carbon fibers. The efficiency of the electron gun is improved
by modernization of the cathode-modulator unit of the lamp. The electron gun
(Fig. 8.6b) [10, 11] is redesigned so that current hogging is quite negligible, less
than 1%, unlike the cathode-modulator unit of the previous design, where some
electrons are sure to be retained by the control electrode (Fig. 8.6a).
A series of volt-ampere characteristics taken (Fig. 8.7) for a new design of the
electron gun shows that the anode voltage of +10 kV, which is the operating mode
of cathodoluminescent lamps, causes the maximum modulator voltage of 1200–
1300 V at the cathode current of 100 lA [12].
To increase light-emission power, one has to increase the emission current of the
field emission cathode. As the current from one carbon-fiber bunch is limited for
long-term continuous operation with the value of about 100–150 µA, the current
increase assumes that the field emission cathode contains several carbon-fiber
bunches. Particularly, papers [13, 14] mention a field emission cathode of 10
carbon-fiber bunches arranged in a circle (Fig. 8.8). In this case, the lamp dimen-
sions remain almost similar to those of the one-fiber field emission cathode.
Availability of a large number of fiber bunches in the field emission cathode makes
it possible to increase emission current up to 10 mA. However, such a heavy
current can overheat the anode. Another effect of application of a multi-bunch field
emission cathode is increase in screen glow uniformity. In this design, the modu-
lator is represented by a thin molybdenum grid stretched taut onto a ring.
The most effective design of a vacuum-tube light source based on reflection and
consisting of many carbon-fiber bunches is presented in Fig. 8.9.
Fig. 8.6 Cathode modulator

unit: a a prototype design;
b the new optimized design
Fig. 8.7 Dependence of the cathode current on the voltage on the modulator for the CMU of the
first design (a) and the optimized design (b) at various values of anode voltage: (1) Ua = +10 kV,
(2) Ua = +9 kV, (3) Ua = +8 kV, (4) Ua = +7 kV
Field emission cathode (1) consists of 12 T-300 carbon-fiber bundles manu-

factured by AMOSO Perf Products and located on a hexahedral base. Every fiber
bundle contains 3000 filaments of 7 lm in diameter. The fibers project 5 mm above
the base, whereas the bundles are spaced 3 mm apart. Modulator 2 is represented
with an 80% transparent grid of 0.1 mm-step and spaced 1 mm apart from the
emitting surface of the field emission cathode. Field emission cathode 1 and
modulator 2 are electrically and mechanically separated by two layers of
glass-metal insulators/beads 4. Anode 3 is a polished metal disk with a luminophore
applied. Modulator 2 and anode 3 are spaced 3 cm apart. All parts are assembled
into a rigid structure with three ceramic rods 5. The lamp envelope of 60 mm in
diameter centers the structure with special springs 6.
The structure makes it possible to achieve peak efficiency since there is no
aluminum layer absorbing much energy. Moreover, the photons generated during
electron bombardment of the luminophore are not absorbed by luminophore grains.
Fig. 8.8 The structure

diagram of a light source with
the field emission cathode
with 10 bundles of carbon
fibers: 1 electrical lead of the
multiple-beam field emission
cathode; 2 electrical lead of
the control grid; 3 electrical
lead of the anode; 4
luminescent screen (phosphor
covered with aluminum); 5
vacuum; 6 glass bulb; 7
thin-film getter
Fig. 8.9 The design of a

powerful light source based
on the reflection effect: 1
cathode; 2 modulator grid; 3
anode covered with phosphor;
4 insulator holder grid
cathode; 5 ceramic rack; 6
fixing springs
A large size of the anode and a sufficient quality of carbon fibers make it possible
to generate a heavy emission current and an appropriate light flux.
The light flux can be increased even more by making a convex-concave structure
or a structure in a form of convex collecting lenses on the anode surface. That
increases the area coated with the luminophore and, accordingly, the light flux [15].
Along with carbon fibers as field emission cathodes, light sources employ other
materials as well, e.g., carbon nanotubes. Essentially, such lamps are designed in
the way similar to the one in Fig. 8.8. Nanotubes are laid on flat bases of a few
square millimeters either with the CVD method [13] or the print method [14]. At
the cathode-control grid distance of 0.2 mm, control voltage is 300 V. At the anode
voltage of 10 kV, the anode current is −60% total emission current of the field
emission cathode. Brightness of such a source is 6.3 104 cd/m2 for a green
luminophore (ZnS:Cu) at the anode current of 200 µA, the anode having been
aluminized. Efficiency of such a source is 70 lm/W.
The authors have elaborated skeleton pattern [16] into a high-brightness light
source [17] shown in Fig. 8.10. The field emission cathode is made of special
multiple-wall nanotubes termed as nanofibers fabricated in an electric arc in the
hydrogen atmosphere. One design provides the cathode-current density of
100 mA/cm2 at the grid voltage of 2.9 kV and the cathode-grid distance of about
1.5 mm. Brightness of 1,000,000 cd/m2 is achieved for a green luminophore (ZnS:
Cu) in a continuous mode at the anode voltage of 30 kV and the anode current of
400 lA. The light flux exceeds 1000 lm. The light-source envelope is protected
against Roentgen rays with lead-glass coating 11 and the space between envelope 6
and protection 11 is filled up with cooling shell 10 to prevent the light source from
overheating by the high power released at the anode. Auxiliary electrode 5 placed
around the cathode-grid unit is necessary to protect the field emission cathode from
unexpected high-voltage discharges.
Fig. 8.10 The structure

diagram of the construction
beyond the bright light
source. Source length: 100–
150 mm; diameter: 36.5 mm.
1 Electrical lead; 2 ceramic
and metal terminal; 3
multiwall nanotube field
emission cathode; 4 control
grid; 5 protective electrode; 6
glass cylinder; 7 layer of
aluminum; 8 phosphor; 9
glass plate; 10 cooling cover;
11 protection against X-rays;
12 visible light
Fig. 8.11 The design of a

diode light source with a
cylindrical nanotube field
emission cathode: 1
phosphor; 2 anode lead; 3
protective cylinder; 4 body; 5
field emission cathode; 6
electrical lead; 7 exhaust
opening; 8 aluminized layer
Papers [16, 17] suggest diode light-sources with cylindrical field emission
cathodes made of nanotubes. The circuit of such a light source is presented in
Fig. 8.11. The carbon nanotube paste is applied on the internal surface of
cathode-protective cylinder 3.
No advantages of these designs have been reported so far.
The luminophore can be applied both on one [18], and the two opposite sides of
the envelope [19].
Note should be made of the designs with field emission cathodes made of
nanograhite [20], nanocoils [21] nanotubes turned into carbon filaments [22, 23].
Practically, they coincide with the designs given in Figs. 8.4 and 8.6.
JNA NOV website [24] reports development of tiny light-sources with a field
emission cathode made of carbon nanotubes of 3 mm in diameter.
The section concludes that the few attempts to design lamp sources have not
been a success so far. First of all, that is a necessity to generate quite heavy field
emission current in vacuum in a lamp of about 10−6–10−7 mm Hg.
It is a light source with pointed field emission cathodes of about 10,000 apices
per substrate of 1 mm in diameter [25], as well as a light source with nanowires of
MoO2 [26].
8.1.3 Flat Cathodoluminescent Light Sources
This section considers the general-purpose light sources termed flat.

The basic structural feature of flat light sources is a large area of the
anode/cathode and a small distance between the cathode and the anode in com-
parison with the linear dimensions of the cathode substrate. A flat light source is
made up of the following components (Fig. 8.12), viz. 1 is a luminophore coated
anode; 2 is a glass spacer with thickness ranging from some tens of micrometers to
some units of millimeters; 3 is an exhaust tube; 4 is a getter volume; 5 is a flat field
emission cathode; 6 is a modulator usually represented with a metal gauze.
A proper flat vacuum-device manufacturing technology makes it possible to
develop a light source of more than 500 cm2 in area. A device of such an area can
be implemented by employment of diode and triode structures, which is stipulated
by the finite characteristics of the device. Hyper-luminous light sources result from
a triode structure. Flat light sources of small luminosity, viz. 1000–5000 cd/m2,
result from a diode design, which considerably simplifies manufacture of a vacuum
device. In particular, such devices can be used in LED-backlit displays
(Sect. 8.1.6).
An earlier structure, shown in Fig. 8.13, is based on carbon-fiber bunches [27–
31].
Fig. 8.12 Flat light sources: a diode design of the device; b triode design of the device; 1 anode
covered with a phosphor layer; 2 glass spacer; 3 exhaust stem; 4 getter volume; 5 flat field
emission cathode; 6 modulator
Fig. 8.13 Flat light source

with field emission cathodes
made of bundles of carbon
fibers 1 bundles of carbon
fibers; 2 modulator; 3
isolating support; 4
conductive field emission
cathode base; 5 bottom glass
base; 6 phosphor; 7 top glass;
8 conductive transparent
cover
Carbon-fiber bunches 1 are fixed to metal matrix 4. Modulator electrode 2 is

located in parallel to the emitting plane of the carbon-fiber bunches at the distance
of a few deciles of a millimeter from it.
The modulator-electrode orifices are coaxial with the emitting ends of the fiber
bunches. Cathode 4 and modulator plate 2 are separated by an insulator located
inside the vacuum space bordered by glass plates 5 and 7. Upper plate 7 is covered
with a layer of conducting transparent ITO coating 8 and luminophore 6.
The anode, the modulator and the cathode are connected from the outside via
electrical leads A, B and C, respectively.
Unique TEG cathode properties make it suitable for planar large-area field
emission cathodes reliable when functioning in the diode mode. Such cathodes can
be employed in flat light sources of different area intended for both continuous and
patterned illumination, including signs, digits, marks etc., viz. static information
displays. Such devices have low power consumption. At the same time, they are
quite easy to manufacture.
The technology of emission centers on the TEG-foil surface makes it possible to
create flat field emission cathodes of different area ranging from a few to hundreds
of square centimeters. Moreover, the field emission cathodes fabricated can have
different pattern of the emitting surface.
The design concept of the light source based on the TEG-cathode is presented in
Fig. 8.14. The device is arranged as a diode. A glass plate coated with conducting
ITO layer is attached to TEG foil. Craters are formed on the foil according to the
pattern. The anode is made of glass coated with an ITO luminophore deposited
layer. The distance between the anode and the cathode is set at 1 mm with glass
spacers. As the TEG foil is 200 µm thick and the average height of the crater is
200–250 µm, the space between the emission centers and the anode is about
550 µm. Figure 8.14 shows a ready TEG cathode based light source prototype.
Fig. 8.14 Prototype of a flat light source: a structure diagram; b appearance

The device works in a continuous mode. The anode area is 4.8 cm2. If the
voltage applied is 1200 V, the electric field being 2 V/µm, the anode-current
density is 0.2 µA/cm2. The device is about 3000 cd/m2 bright.
The basic methods of fabrication of large-area field emission cathodes include
screen printing. A field emission cathode based on diamond-dusted carbon nan-
otubes is exemplified in [29, 30]. The printing paste contains carbon nanotubes, a
binder, glass powder, a dispersing agent, an organic solvent and diamond dust at a
rate of 0.1–20 wt%.
After stirring, the paste obtained is applied on the conductive substrate by screen
printing and annealed at 330–470 °C for about one hour. The resulting structure of
the light source is shown in Fig. 8.15.
A mixture of carbon nanotubes and diamond dust produces a relatively high
density of emission current.
Moreover, such a structure is quite advantageous in terms of technology and
service properties of the field emission cathode.
The devices based on the reflection effect and ruled field emission cathodes can
be even more economical. In particular, papers [32, 33] mention such a light source
(Fig. 8.16).
Fig. 8.15 Structure of a flat light source with a field emission cathode printed from carbon
nanotubes and diamond powder. 1 Glass plate; 2 ITO layer; 3 phosphor; 4 diamond powder; 5
carbon nanotubes; 6 power supply
Fig. 8.16 The light source with line field emission cathodes based on the reflection effect: 1
bottom glass plate for light radiation lead; 2 sealant; 3 transparent conductive covering (ITO); 4 top
glass plate; 5 luminescent layer; 6 black cover; 7 aluminum anode; 8 semicylindrical groove; 9 line
cathode (silver paste); 10 nanotubes; 11 getter
Field emission cathode 8, 9 consists of silver-paste bars 8 bearing equally spaced

parallel nanotubes 9 forming the field emission cathode itself. The cathode bars are
centered in grooves 8 shaped as half-cylinders. The grooves are laid in aluminum
plate 7, the latter being the anode. The internal surface of the grooves is covered
with luminophore 5 of appropriate colors, the grooves being parallel and equally
spaced.
Removal of heat from the anode is improved by blackened aluminum layer 6
emitting beyond the device through transparent conductible coating 3 and upper
glass 4. Coating 3 also provides the external electric contact of the anode.
The electrons emitted by the field emission cathodes bombard luminophore layer
5 at anode 7, thus causing its excitation and, consequently, luminescence. The
luminescence reflected from the internal surface of anode grooves 8 passes through
the lower glass plate. The shape of the grooves dissipates and bounces off the light
fluxes, which considerably improves the luminescence uniformity of the light
source.
Sealant 2 makes the device gas-tight, the necessary pressure being maintained by
getter 11.
Another approach to development of ruled field emission cathodes is to use
wire-based microelevations. In particular, paper [34] describes a field emission
cathode represented by wire with some microelevations on it. The substance of
microelevations can be any good field emission material, e.g., ZnO, W, C etc.
Figure 8.17 shows the diagram of such a light source.
The field emission cathodes represented by wires with equally spaced microel-
evations are drawn directly on lower base 4.
The wires are centered in the anode cavity 7. The shape of the cavity can be
different ranging from cylindrical to polyhedral. The internal surface of the curved
section of upper plate 9 is coated with luminophore 8. Rigidity of the structure is
provided by internal transparent spacers 1. The device is pumped out through lateral
exhaust tube 5.
The device can be modified into a structure where the field emission cathode is
between two anodes [34]. In this case, light falls on both side of the device case.
Efficiency of a light source using a ruled-wire cathode can be further improved if
such a cathode is placed/suspended between two electrode plates [35]. The upper
anode plate is flat plate with a conductive coating and a luminophore. The lower
anode plate is flat plate with a luminophore and an aluminum coating. In this case,
Fig. 8.17 Flat source with wire field emission cathodes. 1 Transparent glass spacer; 2 field
emission cathodes; 3 bearing wire field emission cathodes; 4 bottom base; 5 exhaust stem; 6
conductive anode cover; 7 convex surface of the luminous element; 8 phosphor; 9 top plate
emission from both plates adds up, which cannot but improve efficiency of the light
source.
The wires can be replaced with different grids represented by both perforated
plates [36] and mesh grids [37]. The grid is usually placed between two lumino-
phore coated anodes. The grid surface overlooking the anodes is coated with an
emission layer. It consists of numerous carbon nanotubes, metal conductive parti-
cles and a getter powder dispersed in a glass matrix.
The nanotubes are from 5 to 15 lm in length, the diameter ranging from 1 to
100 nm. Metal particles are powdered silver (Ag) or indium-tin oxide (ITO). These
particles provide electric contact between the nanotubes and the base, i.e. the
conductive mesh. The getter powder consists of a non-evaporating getter material,
i.e. the one containing titanium (Ti), zirconium (Zr), hafnium (Hf), thorium
(Th) and their alloys. The average size of the getter powders ranges from 1 to
10 µm.
In addition to nanotubes, field emission cathodes can be made of other materials
both carbon- and non-carbon-based; e.g., nanostructured carbon materials [38] and
carbon nanotube [39], microgranules of 12SaO; 7AL2O3 or 12SrO7AL2O3 with
particle of 0.1–100 µm [40].
The simplest structures employ field emission cathodes made of constructional
graphites [41]. The structure and the milestones of manufacturing of a diode-based
light source using a field emission cathode of constructional graphite are presented
in Fig. 8.18. At the first stage, a blank of 1 1.5 cm is cut out of a lump of
constructional graphite, viz. fine-grain high-density graphite MPG-6. The blank is
grooved 0.5 mm at a pitch of 1.5 mm according to the spark-erosion technology
(Fig. 8.18a). The grooves are added with POLYCOR (or Lucalox) type ceramic
plates as thick as 0.5 mm. The resulting structure is cast in glass-fiber reinforced
cement (Fig. 8.18b).
After the glass-fiber reinforced cement crystallizes in a muffle furnace at 400 °C,
it is ground off together with the ceramics until the graphite surface appears. Then
the surface is ground with an abrasive powder of a rated grain size, viz. *28 µm.
As Lucalox is close to abradants in hardness, while graphite is much softer, the
graphite surface is ground off below the Lucalox plate end-faces by the value close
to the size of the abradant particles. The simple technological means create a small
uniform clearance, viz. 30 µm, over the entire cathode plane (Fig. 8.18c). The
Fig. 8.18 Diagram for production of a diode light source based on constructional graphite:
a milling grooves in the graphite cathode workpiece; b inserting ceramic plates and molding with
glass frit; c completed structure. 1 Graphite field emission cathode; 2 glass frit; 3 ceramic sheets; 4
cathode-anode gap; 5 anode with phosphor and conductive cover
spacers are superimposed with a screen with a luminophore and a conductive

coating.
In addition, graphite field emission cathode surface grinding with the abrasive
powder produces the surface relief pattern necessary for effective field emission.
In addition to fixation of a precise anode-cathode gap, the spacers provide
uniform load on the anodic and cathodic plates to compensate for the atmospheric
pressure and the ponderomotive loads during the operational period of the device.
The prototype device is tested in the constant-voltage mode in a vacuum chamber at
the residual-gas pressure of 10−6–10−7 mm Hg.
The field emission measurements of the resulting diode structure are presented in
Figs. 8.19 and 8.20. An appreciable field emission current starts at the voltage of
500 V. After the first turn-on, the current decreases down to 1 mA and slightly
increases within 200 h. As long-term tests take place in a cyclic mode, i.e. voltage
and pumping-out are turned off at night, every turn-on increases the current, but the
increase disappears in 0.5–1 h.
The field emission image consists of small bright points located almost right up
to each other. The fact of uniform flare spot by the graphite-cathode surface
characterizes high uniformity of microelevations on the graphite surface. Trial
experiments in development of a diode light-source based on a field emission
cathode of constructional graphite prove that this approach is very promising.
Estimates show that a thin-film luminophore and finer abrasive powder used to
Fig. 8.19 Current/voltage

diagram of the diode structure
(anode-to-cathode distance is
30 lm)
Fig. 8.20 Change of current

over time (anode-to-cathode
voltage is 790 V)
grind field emission cathode structures can produce diode structures with a running
voltage of 200–300 V.
One of the most promising directions of development of flat light sources is the
lateral construction, i.e. such arrangement of the device where the field cathode and
its controlling electrode (modulator) lay in the same plane. In Fig. 8.21 one can see
one of the most well-turned examples of such design, a field emission cathode light
source presented in [42]. This construction together with modern developments in
microelectronics allowed to simultaneously achieve rather high level of homo-
geneity and strong light beam.
8.1.4 Cylindrical Cathodoluminescent Light Sources
The conceptual design with an axial arrangement of elements is shown in Fig. 8.22.
This design employs a cathode based on fibers, nanotubes or other field emission
materials.
An important advantage of a cylindrical structure in comparison with a planar
one is that the same applied voltage produces the microscopic electric field value of
E ¼ Ur ‘n Rr, which is more than in the case of a planar structure, where R and r are
the radii of the anode and the cathode, respectively. Consequently, it is possible to
use a diode light-source structure, which is considerably cheaper in manufacturing
than the triode structure.
A cylindrical triode-based light source with field emission cathodes made of
carbon-fiber bunches is presented in Fig. 8.23 [43–45].
This light source has a conventional configuration of anode 1, i.e. a glass
cylinder-shaped base covered with a conductive coating with a luminophore layer.
Modulator 2 is a cylinder of metal gauze or etched metal foil. Field emission
cathode 3 is made of carbon fibers. There are two versions of cylindrical field
Fig. 8.21 Schematic diagram

of lateral-gate triode structure
of field emission device
Fig. 8.22 Design of an axial

e- 2
light source: the 1 central part
with cathode structure, the 2
cylindrical anode with a 1 e-
luminescent covering
Fig. 8.23 Cylindrical light source with field emission cathodes made of bundles of carbon fibers.
a Design of a cylindrical light source; b light source cutting; c diagram of the lamp-chimney brush
field emission cathode d the field emission cathode with a radial arrangement of bundles of carbon
fibers. 1 Anode cylinder with the coating conductive and luminescent layers; 2 modulator; 3 field
emission cathode from carbon fibers; 4 bundles of carbon fibers; 5 wire at the base of the field
emission cathode; 6 clamping disk; 7 bearing disk
emission cathodes. In the first version, carbon fibers 4 are fixed between two or
three twisted wires 5 to form a brush-type structure. The structure is very easy in
manufacturing, but deformation of carbon fibers can result in their breakage and
subsequent extraction from the cathode by the electric field. Moreover, it is very
difficult to maintain the same height of the field emission cathode fibers and
respective equal spacing between the cathode and the modulator. In the second
version of the field emission cathode structure (Fig. 8.23d), carbon fibers are fixed
with conducting glue or soldered between two disks 6 and 7 made of a conducting
material. The assembled units are hafted on a bearing rod of the length desired. This
structure provides a more accurate coaxiality of the field emission cathode and the
modulator, as well as a higher structural efficiency. All those factors improve
uniformity of the field emission current over the surface of the field emission
cathode and reduce the likelihood of cathode-modulator gap bridging.
A lot of papers [46–50] are dedicated to the classical diode structure, where the
functions of a field emission cathode are performed by either a tungsten wire or a
metal rod of 2 mm in diameter coated with carbon nanotubes or ZnO nanoledges
[51]. In all the papers, the length was several centimeters, viz. from 3 to 10, and the
diameter was 15–30 mm. The nanotubes were deposited in tubular reactors made of
quartz tubes at 680–850 °C for 10–30 min. The brightness achieved was about
10,000 cd/m2 at anode voltage of 5.4 kV [46].
Efficiency of such lamps depends mostly on the nature of the luminophore and
the applied voltage and amounts to 21 lm/W for a white luminophore and 37 lm/W
for a green one [49].
Fig. 8.24 The schematic diagram of a cylindrical light source with a reflecting anode [46, 47]. 1
Aluminium semi-cylindrical covering lay phosphor; 2 field emission cathode; 3 glass bulb
Some efficiency improvement can be achieved by changing the anode structure

[52, 53]. Figure 8.24 presents the lamp circuit.
The basic difference from the previous designs is application of reflecting alu-
minum coating 1 to be covered with a luminophore. Cylindrical cathode 2 of 1 mm
in diameter is coaxially placed inside a cylindrical glass envelope of 20 mm in
diameter.
A stronger field emission cathode coating [54] consists of a mix of nanotubes,
metal conductive particles and a getter powder dispersed in a glass matrix. The
composition evenly coats the cathode rod to provide uniform emission over the
field emission cathode surface.
Another approach to development of cylindrical light sources was suggested in
[55, 56]. The circuit of such a source and the photo of its working model are
presented in Fig. 8.25.
The light source is a glass tube of 15–20 mm in diameter and 20–30 cm in
length lined with cathode luminophore from within. Cathode-modulator units are
mounted on the lamp feet welded in the end faces of the tube. The
cathode-modulator units are the same as in small-button glass tubes (Sect. 8.2.2).
The lead-in wire, viz. the anode, is soldered in the middle of the tube and supplied
with the accelerating voltage of 10–15 kV. The electrons emitted by the field
emission cathodes knock out secondary electrons on their way to the anode. As a
result, the lamp glow uniformity considerably increases and its efficiency improves.
The experimental samples demonstrate the efficiency of more than 20 ln/W, the
make-time being less than 0.1 ms.
8.1.5 Spherical Cathodoluminescent Light Sources
Domestic lighting is mostly based on spherical-envelope lamps dissipating light in

all directions within the angle of 300°.
One of the first designs employs 7 isolated bumps with carbon-fiber bunches,
their ends being placed along the envelope radius [57].
In this case, the electron beams cross each other and produce uniform glow. The
anode current is 1–2 mA at the voltage of 4–5 kV.
Fig. 8.25 Cylindrical light source with a face arrangement of field emission cathodes.
a Diagram of the source; b a photo of a working model
Similarly to cylindrical lamps in [58, 59], a spherical envelope can be fitted with
a nanotube wire based field emission cathode. The wire base is waved or
saw-toothed and placed in the middle of the envelope. The nanotubes can be
applied both after and before the wire base is waved or saw-toothed. Application of
a reticle is also envisaged.
Finally, patent [60] finishes publication of a series of designs using a field
emission cathode of glass with metal and getter particles dispersed by nanotubes in
the same way as [37, 54].
8.1.6 Ultraviolet Lamps
Global environmental safety issues have recently accentuated the necessity of

replacing the mercury-containing UV lamps widely employed in medicine, air
clearing at different offices, in dye polymerization facilities, and other fields. Eximer
and deuterium lamps are likewise far from perfect from the standpoint of envi-
ronmental safety. The first attempts to create cathodoluminescent UV lamps inspire
optimism as regards the prospects for solving this important problem. The fol-
lowing additional peculiarities of UV lamps are worth mentioning:
1. practically zero time to turn on (10−8 s);

2. operation in both analog and pulsed modes;
3. wide operating temperature range;
4. generation of UV radiation within virtually any fixed range owing to the
dependence of the emission spectrum on the phosphor composition alone;
5. possibility of producing lamps with a wide range of output powers in various
embodiments (butt type, flat, spherical, etc.);
6. environmental safety at all stages of production, operation, and disposal due to
the absence of mercury and other noxious components;
7. possibility of delivering power into lamps from any line (220, 380 V) or
on-board (12, 24 V) voltage, storage cells and batteries.
The development of novel efficient UV cathodoluminescent lamps depends first
and foremost on the availability of wide bandgap cathodophosphor materials.
Mixtures of oxides (BeO or MgZnO) have a wide bandgap. However, BeO is
highly toxic, while MgZnO production creates considerable difficulties due to the
difference between ZnO and MgO structures. Fluorides are very promising UV
cathodoluminescent materials by virtue of a very wide bandgap [61]. Such fluorides
as KMgF3 and KCaF3 have emissions in a wavelength range from 140 to 220 nm.
KMgF3 possesses a cubic structure and does not need to be doped with rare-earth
elements. The construction of a flat UV lamp with a KMgF3 phosphor is described
in [62, 63]. An oriented MgF2 crystal on which a KMgF3 layer is deposited by laser
evaporation serves as the exit window with an area of 64 mm2. The evaporation
target is prepared by melting KMgF3 (KF:MgF2 = 1:1) in an atmosphere of Ar:CF4
(95:5) at 1220 °C. The resulting structure is cooled to 900 ºC for 1 h and further to
room temperature during the next 48 h.
The anode and the extraction electrode are made from copper mesh with 0.1-mm
cells; the fluoroplastic spacers are 0.3 and 1 mm thick. Cone-shaped nanostructured
carbon field emission cathodes are grown on a glass-carbon substrate by Ar+ ion
bombardment at room temperature. The density, length, and diameter of the
resulting microtips are roughly 5 108 cm−2, 0.3–2.0 and 20 nm, respectively.
The operating pressure in the chamber is 5 10−6 Torr. The parameters of the
system are: extraction voltage of 800 V, accelerating voltage of 1800 V, current of
0.32 mA, and UV radiation power of 2 mW. The KMgF3 emission spectrum has a
maximum at a wavelength of 180 nm; in a system with an MgF2 substrate, the
spectrum has two peaks, at 150 and 180 nm. For UV radiation with longer
wavelengths, it is far more practicable to use readily available and technologically
simple materials for exit windows, e.g., quartz or uviol glasses. Reference [64]
proposes a simple design for a UV lamp with emissions in a wavelength range from
220 to 350 nm (see Fig. 8.26). The use of field emission cathodes from a nanos-
tructured carbon material permits decreasing the threshold electric field strength for
emission current (1–5 V lm−1), increasing the operational life of the field emission
cathode to 50,000 h and thereby prolonging the lifespan of the lamp as a whole.
The start-up time of a vacuum UV lamp is less than 10−8 s, which means that light
is emitted immediately after the power is turned on. The emission spectrum of
Fig. 8.26 Miniature UV lamp: 1 evacuated case made of UV-permeable dielectric; 2 anode; 3 UV
phosphor; 4 aluminum layer; 5 modulator; 6 terminal lead; 7 nanostructured carbon field emission
cathode; 8 alignment disk; 9 contact node; 10 field emission cathode terminal leads
luminescent UV lamps depends on the phosphor chemical composition, which

allows choosing the optimal spectrum for a concrete application by varying the
phosphor composition, e.g., using KL-UV-315 and KL-UV-300 cathode phosphors
with emissions at wavelengths below 350 nm. The aluminum layer deposited onto
the phosphor serves to increase luminous efficiency, because the angular spread of
radiation from phosphor grains is 360° and the aluminum layer plays the role of a
mirror reflecting light into the exterior part of the lamp. The UV lamp is a cylin-
drical evacuated enclosure from a light permeable dielectric material with a field
emission cathode made of a nanostructured carbon material, a modulator with a
hole for passing electron beams, a luminescent screen, an anode, and terminal leads.
Coaxiality is achieved by placing the field emission cathode in the hole of an
alignment disk oriented coaxially with the modulator aperture. To enlarge the
contact surface, the terminal of the field emission cathode can be made of an
electrically conductive material, e.g., aquadag, deposited on the end of the field
emission cathode and a rim adjoining its side surface with which the cathode
terminal is rigidly connected. The luminescent screen of a field emission UV lamp
is covered with a UV phosphor layer and aluminum to increase the light output
from the phosphor. The technical result, namely, efficient conversion of the elec-
trical energy into UV radiation, is attained by using special materials for the field
emission cathode and a phosphor capable of producing efficient UV radiation when
combined. The lamp is characterized by high energy efficiency (at least 15%), a
Fig. 8.27 UV lamp radiation

spectrum
long operational time (50,000 h), almost zero time to turn on, a high resistance to
mechanical vibrations and voltage fluctuations, and the absence of polluting sub-
stances. Phosphors of various chemical compositions may be utilized, depending on
the necessary UV radiation wave-length [65–69]. For example, high-performance
Bin Ym Al2(BO3)4 compositions are suitable for the purpose at a wavelength of
300 nm. The radiation spectrum of a UV lamp with one such phosphor is presented
in Fig. 8.27. The efficiency of these phosphors is high enough for practical appli-
cations (27 mW at an overall power of 900 mW; U = 8 kV, J = 112 mA).
Phosphors of a different composition, e.g., ZnAl2O4, are needed if the wavelength is
to be decreased to 245–260 nm [68–70]. Only the first steps have been taken to
design vacuum luminescent lamps using field emission cathodes and to search for
the optimal chemical compositions of phosphors. There- fore, considerable effort to
continue development of this technology can be expected in the near future.
8.1.7 Cathodoluminescent Light-Source Application

Guidelines
At present, designers intensively develop two areas of application of light sources,

viz. liquid-crystal back-lit display screens and group displays.
Nowadays, liquid-crystal display screens are widely used as information dis-
plays in TV sets, computer monitors etc. They are back-lit by line or point light
sources, such as gas-discharge fluorescent lamps and light-emitting diodes.
However, their basic disadvantages are a complex structure, a sophisticated man-
ufacturing technology, a high production cost and intensive power consumption.
Moreover, conventional light sources are unable to provide uniform glow of the
entire screen in oversize liquid-crystal screens.
Therefore, field emission light sources, which have low power consumption and
more uniform glow, are in increasingly better demand.
Some designs of cathodoluminescent back lights using carbon-based field
emission cathodes are given below. Some of them have something in common with
flat light sources (Sect. 8.2.3). In particular, Fig. 8.28 presents a back light circuit
version. The lamp consists of anode plate 1 bearing a transparent coating layer 2
and a luminophore layer 3. Spacer 4 sets the gap between the anode and the
cathodes located on lower substrate 5.
The parallel bars of transparent conductive ITO coating 6 bear thin metal layers
7 serving as the contacts of field emission cathodes 8, 9. Field emission cathodes 8,
9 are formed by well-known above-described methods, e.g., screen-printing with
the carbon-nanotube paste. It is peculiar that electrodes 8 and 9 of this design can
alternatively function as a field emission cathode and a modulator. This feature
nearly doubles the service life of such a lamp at the same brightness and increases
uniformity of its glow, wherefore they are connected with each other skipping the
nearest one.
Control with the neighboring parallel electrodes in a way similar to [71] is also
applied in [72, 73]. The quality of the backlight is also determined by luminophore
flare uniformity. With this end in view, the modulator is shaped as a ruler
(Fig. 8.29) [74].
As the picture shows, control electrodes 6 are placed on linear insulators 5
located at right angle to cathodic electrodes 4. Cathodic electrodes 4 can be made of
any conducting material; e.g., a metal or a transparent conducting coating (ITO).
Insulators 5 are 3–10 µm high, the distance between the rulers ranging from 10 to
30 µm. Field emission cathodes 8 are made of the carbon nanotube paste by
conventional screen printing. This design has a wider electron-divergence angle,
Fig. 8.28 Backlight with printed field emission cathodes made of carbon nanotubes [54]: 1 anodic
plate; 2 transparent conductive cover; 3 layer of phosphor; 4 spacer; 5 cathodic plate; 6 thin layer
of metal; 7 ITO layer; 8, 9 field emission cathodes manufactured using paste which contains
carbon nanotubes
Fig. 8.29 Module backlight

lamp with a larger divergence
angle of the electron beam: 1
anodic plate; 2 transparent
conductive coating; 3
fluorescent layer; 4 cathodic
electrode; 5 insulating layer; 6
control electrode; 7 base of
the cathode; 8 printed field
emission cathodes
Fig. 8.30 Classic design of a

passive matrix LCD: 1
polaroid; 2 liquid crystal
matrix; 3 light filter; 4 white
backlight
which causes overlapping of electron beams on luminescent screen 1, 2, 3, thereby

improving uniformity and luminance of the light source.
Some authors suggest growing nanotubes directly in the assembled device [75].
For this to be done, the cathode base, represented in this case by the rods placed
between the case plates, is coated with a catalyst. The lamp housing has inlet and
outlet openings for carbon gas, e.g., methane (CH4). In this case, nanotubes are to
be deposited at about 350 °C. As soon as the deposition is over, the operational
openings are closed and the lamp is pumped out.
All the above-described backlit liquid-crystal displays (LCD) are white lamps.
They are intended for the main types of LCD screens, i.e. for passive-matrix LCDs
with a light-filter system. The operating principle of a conventional LCD screen is
shown in Fig. 8.30 [76, 77]. The basic framework is passive gate LCD matrix 2
with a polaroid 1. The matrix is combined with light filter 3 having areas of three
different colors, viz. red (R), blue (B) and green (G). The LCD matrix is lit up with
white-color backlight 4 working in a continuous mode.
Light flux Y generated by backlight Y after the light filter is equal to ŋcf, after the
LCD matrix to Yŋcfŋm, at the front panel of the screen Yŋcfŋmŋp (ŋcf, ŋm, ŋp the
average transparency of the light filter, the matrix and the polaroid, respectively).
Every color element of the light filter passes less than 1/3 of the spectrum. The
exact value is determined by the spectral characteristic of the backlight and the
desired relation between the intensity of red, blue and green at the screen output. As
the average transparency of the light filter is ŋcf * 0.33, a considerable part of the
light flux is wasted in the light filter.
The light efficiency of the screen can be improved by removal of the light filter
and employment of a polychromatic backlight able to switch-on/off the color ele-
ments in a synchronous mode. The arrangement of such a screen is shown in
Fig. 8.31. This screen uses the same matrix 2 and polaroid 1 as the screen in
Fig. 8.30. However, the white backlight is replaced with trichromatic backlight 3
able to switch on each color in an independent way. The basic requirement to a
polychromatic lamp is matrix-lighting uniformity for each color. The sizes of the
backlight color-element do not have to coincide with the matrix-pixel size. At the
given flare non-uniformity criterion, the size of the backlight color-element deter-
mines the gap from the backlight to the LCD matrix.
Every backlight color-element switches on after the LCD matrix develops the
image appropriate to this color. At that, the effective screen resolution, viz. the
number of image elements, trebles as compared with the screen resolution in
Fig. 8.30 because every pixel of the matrix is engaged.
At the frame frequency of 100 Hz, each of the 3 color elements of the backlight
is switched on for not more than 3.3 ms at a power-on time 10 ms. Accordingly, the
lamp-lighting time should be less than 0.1 ms. Fluorescent backlights widely used
now cannot meet these switch-on/off rate requirements. The most acceptable ver-
sion providing high control speeds and high brightness is cathodoluminescent light
sources with a field emission cathode. The overall efficiency and brightness of such
lamps are determined by the luminophore and good luminophore can provide them
at nearly the same level as fluorescent lamps. A field emission cathode used makes
it possible to control the lamp with a frequency from 10 kHz upwards. The theo-
retical limit is determined by the phosphor-persistence time and can last a mere few
microseconds. The trichromatic backlight control is exemplified in Fig. 8.32.

of a passive matrix LCD with
multicolor backlight: 1
polaroid; 2 liquid crystal
matrix; 3 tricolor backlight
Fig. 8.32 Graphic display (right) of the behavior of an LCD matrix and backlight over time: 1
image on the LCD corresponding to the color red (red backlight); 2 image on the LCD
corresponding to the color green (green backlight); 3 image on the LCD corresponding to the
color blue (blue backlight)
Fig. 8.33 Photo of the

module for dynamic LCD
backlighting
The red, green and blue lamps switch on in series, and the liquid-crystal matrix
elements switch on in step with them and in an appropriate configuration with the
video image.
As the sizes of lamp-color elements do not have to coincide with the dimensions
of the LCD matrix, one of the first devices for dynamic light consisted of a set of
small-button glass light sources of three colors (Sect. 8.1d) [77]. A photo of the
dynamic-light unit is presented in Fig. 8.33.
The light sources of each color were switched on in series. The frame rate was
100 Hz.
Cathodoluminescent lamps are to have the switching rate of less than 1 ms. The
switching rate of lamps mostly depends on run-up time and luminophore persis-
tence. For the lamps used, the respective periods meet the requirement specified.
The power consumption of the video-wall unit is within 20 W or 100 mW/cm2
in a dynamic mode at the duty cycle of each color 3. The individual-source
brightness being 104 cd/m2 and the density of their packing in the unit being 2/5,
the white-light unit brightness is 1.3 103 cd/m2 in the dynamic mode and the
image brightness at the LCD screen is 300 cd/m2.
Another promising area to develop color dynamic-light lamps is low-resolution
display screens, i.e. those having the pixels of a few millimeters [78, 79].
Figure 8.34 shows the structure of a trichromatic unit based on carbon-fiber field
emission cathodes.
The anode of a conventional configuration has trichromatic luminophore-coated
sections of 10 10 mm in size. The sections are partitioned to prevent the electron
flows from mixing.
The electron beams are formed by a modulator with appropriate apertures, the
modulator being the same for all the cathodes, and controlled by the control system
through the cathode circuits.
In some cases, the flare-spot uniformity can be improved by holographic dif-
fusers [80]. A holographic diffuser is a surface relief hologram providing more
effective control over the angular distribution of light, (up to 100°) and higher
transmittance (more than 85%), than the conventional methods, e.g., clouded glass.
One of the promising areas of application of light sources is their use in the video
walls of large size and, consequently, low resolution [81]. The video walls can be
used in group displays necessary for many functions such as TV broadcasting,
advertising etc.

of a tricolor module with field
emission cathodes made of
carbon fibers as a
low-resolution display
The units are manufactured in different sizes. Some of them are based on light
sources with field emission cathodes made of carbon-fiber bundles [82] as shown in
Fig. 8.33.
The unit has the following characteristics [83]:
– number of lamps 36,
– power consumed up to 40 W,
– luminous efficiency 18 lm/W,
– brightness up to 12,000 cd/m2,
– viewing angle 180 °C,
– pixel size 40 40 mm2 RGB,
– contrast 200:1.
Every light-emitting element of the unit is controlled by a controller.
Large full-color video walls need the units to be able to be connected into a
screen. Such a unit is to display at least one sign/letter/digit and have a power
supply and control system of its own. It is necessary to streamline the electrical
interconnections and the screen-control system in general. In this case, the unit is
only connected to the mains-supply source and control wires from the computer.
The outward appearance of the unit is presented in Fig. 8.35 [84].
Structurally, the unit is a complete sign space containing 8 8 = 64 full-color
pixels. Each pixel consists of three lamps of red, blue and green. The colors are
mixed up by pulse-width modulation of emission current. The total number of
lamps is 192. The unit structure enables to make up a panel of any size without loss
Fig. 8.35 Exterior of a

full-color video module: 8 8
pixels (192 lamps)
of resolution of the screen. The rated power of the unit when all the lamps work at
the same time is about 200 W. Other parameters are similar to those given above.
Designing and manufacturing of the unit solved the principal challenge of field
emission electron sources, namely identity of emission characteristics [82] in
mass-produced devices.
A new approach to lighting in industrial and office buildings is presented in [85].
In this scheme (Fig. 8.36) the electricity is supplied via a single-wire line S from
a high-frequency resonance Tesla transformer 4 (frequency *1–100 kHz).
Transformer 6 is connected through capacitors 2, 3 from converter 1. As a primary
energy source one can employ the alternating current network or autonomous
energy cells such as batteries. Field emission lamps 6 can be connected to the
ground 8 through either limiting resistors 7 or proper natural capacity (if the lamps
are suspended at some height). A variant of construction of a cathodoluminescent
lamp with a single-wire line is presented on Fig. 8.37.
The lamp on Fig. 8.37 is a transparent glass sphere 1 with electrode holders 2, 3
entering it through welded-in glass “bubbles”. One of them holds (conserving the
electrical contact) one of the field electron emitters 4.
Emitter 4 generates free electrons in one of the light sources 6. The other
electrode 2 is connected to the metallic covering of the luminophore-anode 6. The
other of the electrodes 2 holds, also conserving electrical contract, the emitter 5.
The emitter 5 generated free electrons for another light source 8. Metallic covering
9 of luminophore 10 serves as an anode for light source 8. Anode 9 is connected to
the external electrical circuit through electrode 3. Thus electrode 2 is connected to
electron emitter 5 which serves as the field emission cathode for light source 8 and
metallic covering 7 of luminophore 10 which is the cathode for light source 6. As
such, the light sources in the lamp are connected antiparallelly. So in the scheme of
electric lighting on Fig. 8.37 the cathodoluminescent lamps contain two parallelly
connected light sources 6, 8 (antiparallel relative to power supply) in each of the
bulbs. In one of the half-phases of the current the light is emitted from source 6, and
in the other half phase from source 8.
Fig. 8.36 Single-line coupling diagram for cathode luminescence field emission lamps: 1
converter; 2, 3 capacitors; 4 tesla transformer; 5 single-wire circuit; 6 cathodoluminescent lamp; 7
resistor; 8 ground
8.2 Flat Display Screens 459
Fig. 8.37 Schematic design

of a cathode luminescence
field emission lamp used for
single-line coupling: 1 body;
2, 3 electrode holders; 4, 5
field emitter; 6, 8 light source;
7, 9 anode; 10 phosphor
8.2 Flat Display Screens
The world of today is in need of different video-information display devices. The

requirements to such devices are also becoming increasingly stricter: nowadays to
be competitive the displays are required be compact, reliable, cheap, economical
and provide good color rendition and image sharpness in a wide view angle. At
present, there are three types of displays in the world: liquid crystal displays,
plasma-panel displays and electron-beam displays. Electron-beam displays are
ergonomic, have the best color rendition and image sharpness, but are cumbersome
and uneconomical. Liquid crystal monitors are compact and economical, but have a
number of problems with image quality and go up in price with increase in
dimensions.
Plasma panels are expensive and short-lived in comparison with other displays.
None of the three technologies meets all the necessary requirements. Displays with
field emission light sources or field emission displays have been under development
since 1986. Foreign literature refers to them as Field Emission Displays (FED). It
was already mentioned that the image-acquisition principle of such displays is
partially similar to the electron-beam display principle, viz. the luminophore is
excited by highly-energetic electrons generated due to the field emission effect, i.e.
electron tunneling through a potential barrier on the surface of the material caused
by the electric field. The field emission effect is merely quantum, which means that
it can be close to 100% efficiency factor of such a process, i.e. it can be extremely
economical. A field emission display pixel is a set of three subpixels, viz. field
emission diodes, triodes or quadrodes, whose anode plates are transparent and
coated with luminophores of three primary colors, viz. red, blue and green. The key
difference between field emission displays and electron-beam displays is that the
latter have one electron gun for all the pixels, while the display screens with field
emission cathodes have an individual electron gun for each subpixel. Such an
approach does not involve any cumbersome focusing systems like those in
electron-beam displays and enables us to make the display compact and comparable
in thickness with liquid crystal displays. That makes the technology of flat display
screens based on field emission cathodes look very promising for development of a
perfect flat screen [86].
However, it is only Futaba and PixTech that manufacture FED for commercial
application now. Their competitive but expensive displays are issued in small
quantities for special purposes.
The main reasons thereof are faults of available materials and structural concepts
of such displays. The up-to-date achievements in electronics make it possible to
develop steering circuits for the arrays of field emission cathodes, so the work is
focused on the field emission cathodes themselves and the materials to be used in
them. The materials for field emission displays develop along two main directions:
development of new luminophores with better glow and longer life, and search of
new emission materials.
The guidelines of development of field emission materials considered in this
book can help in designing mass display screens with field emission cathodes. In
terms of internal arrangement, field emission displays fall into several groups. Their
classification is shown in Fig. 8.38 [86].
Each structure solves this or that problem, but none solves all of them. The
below describes the types of structures with particular examples.
8.2.1 Display Screens with Pointed Field Emission Cathodes
Flat display screens with field emission cathodes originate from the field emission
vacuum microtriode suggested by Shoulders [87] and the technology developed by
Fig. 8.38 Classification of

flat display screens based on
field emission cathodes
Spindt [88]. The specific features of the vacuum microdevices with field emission
cathodes are considered in [89].
One of the first displays of 110 90 mm, i.e. 6″ in diagonal size, was
demonstrated by Meyer [90] in 1991. Its brightness ranged from 150 to 300 cd/m2.
A pixel was as large as 0.12 mm2. The apex density was 104. The anode-cathode
gap was 200 lm. The control voltage was 80 V. The anode voltage was 400 V.
Futaba Corporation improved this technology and commercialized the displays
[91, 92]. The operating principle of such a display is shown in black and white in
Fig. 8.39a, while Fig. 8.39b demonstrates a photo of a field emission cathode array.
The control electrode has a lot of apertures of about 1 µm in diameter.
Application of voltage between the cathode and the control electrode generates a
heavy electric field in the vicinity of the top of the case. The field makes the
electrons emit from the cathode and accelerates them on their way to the anode. The
anode is reached by more than 98% of the electrons emitted. The electrons excite
the luminophore, and the light can be seen through the ITO layer.
The arrangement drawing of the display screen is shown in Fig. 8.40 [93].
By now, Futaba Corporation has developed a wide range of both monochrome
and color displays based on conical field emission cathodes. However, their sizes
are not large enough yet. There have already been reports about the display image
size of about 20 cm. The reason is a considerable difficulty to maintain high vac-
uum in the display case. The vacuum problem can be solved if the design is based
on up-to-date field emission materials that are more resistant to ionic bombardment,
viz. carbon nanotubes or graphite [94]. In addition, pointed field emission cathodes
can be controlled by different systems and fabricated by various methods [95–97].
The expertise in designing flat displays based on field emission cathodes has
shown that the wide spreading angle of field electrons can cause a simultaneous
flare of several elements of the image, which reduces the image contrast and makes
a turbidity effect on the image.
Fig. 8.39 Schematic diagram and operating principle of a flat display screen: a diagram and
operating principle; 1 anode substrate; 2 anodic electrode (ITO); 3 phosphor; 4 cold field emission
cathode; 5 control electrode; 6 insulating layer; 7 resistive layer; 8 cathodic electrode; 9 cathode
substrate. b General layout

of a flat display screen with
conic field emission cathodes
(Futaba): 1 cathode substrate;
2 control electrode; 3 cathodic
electrode; 4 phosphor; 5
anode substrate; 6 spacer
The design deficiencies of display screens can be eliminated by additional

structural elements minimizing these effects [98–100]. A display screen based on
pointed field emission cathodes is diagrammed in Fig. 8.41.
This design is distinguished for a more sophisticated structure of the anodic
layer, which includes not only the areas with luminophore 10 of the appropriate
color but also light-absorbing matrix 9. It is also termed as a black matrix since it
mostly consists of carbon. The electron beam from the field emission cathode is
prevented from overlapping one pixel with other pixels by parting barriers 8
Fig. 8.41 The schematic design of a field emission flat display screen with increased contrast
[82]: 1 cathode substrate; 2 insulating layer; 3 control electrode; 4 cold Spindt cathode; 5 cathodic
electrode; 6 focusing electrode; 7 anodic electrode; 8 dividing barrier; 9 matrix is black; 10
phosphor; 11 anode substrate; 12 cathode circuit controller; 13 control circuit controller; 14
focusing circuit controller; 15 anode circuit controller
separating the electron flows from different areas of luminophores. The same
purpose is attained by focusing electrode 6 located somewhat higher than control
electrode 3. This electrode is introduced to completely focus the pixels (P) released
from the field emission cathode on the respective luminophore spot 10.
The overall control of screen electrodes is exercised by respective controllers
12–15 delivering the necessary power and steering signals. When an ITO layer on
the anode is used as anodic electrode 7, it is enough to have accelerating voltage of
+5 kV in respect of cathodic electrode 5. As a rule, light intensity of the screen is
improved by using an aluminum layer as anodic electrode 7. The anode-cathode
plates are spaced about 1 mm apart. In order to increase the reaction speed of the
system, a control system with reduced proper capacitance is proposed in [101].
Chapter 5 has already mentioned that it is not only pointed field emission cathodes
of molybdenum that are under intensive development now but also pointed field
emission cathodes of different materials. The first one worth mentioning is silicon.
Silicon apices can be fabricated both by electrochemical etching [102] and
amorphous-silicon crystallization initiated by laser light [103].
Field emission cathodes for flat displays can also be made of composite mate-
rials; e.g., paper [104] suggests placing a niobium cone on the SiO2 columns of 5–
50 nm in diameter. However attractive, the pointed field emission cathodes have
already been noted above as unable to work in high technical vacuum at the
pressure 10−6–10−7 Torr in the device.
Therefore, most of our further descriptions of displays are dedicated to carbon
field emission cathode-based designs.
8.2.2 Diode Display Screens
The diode structures are simplest and cheapest flat display screens. That is why their
development has still been attracting great attention. Especially active attempts to
develop such screens started with application of the screen-printing technique and
the electrophoretic method in production of field emission cathodes. The
screen-printing technique produced monochrome [105, 106] and color [107, 108]
display diode screens ranging from 4 to 9 inches in size.
In the former case, the screen prototype contained 128 128 graphite cathode
lines and a luminescent anode with the points as large as 300 300 µm. A pixel
was as large as 600 600 µm. In the latter case, there were 256 64 lines of
cathodes and anodes with the points as large as 450 450 µm. The pixels were as
large as 650 650 µm.
In both cases, brightness was 300 cd/m2 at 750 V, being able to reach 500 cd/m2
at 800 V. The displays were fed by impulse signals of 5 ms in duration at the
frequency of 1 kHz, the current density being 80 mA/cm2, the anode–cathode
voltage being 750 V. However, the technical characteristics of the display are noted
to be improvable by optimization and boost of the technological parameters, e.g.,
pica type and formation of surface morphology of the graphite cathode. The greatest
Fig. 8.42 Schematic

structure of the display with
matrix addressing
successes have been noted in design of flat display screens based on finely ground
nanotubes. Paper [109] presents a version of a conventional matrix-addressed diode
display. The principle of design and control is shown in Fig. 8.42. The field
emission cathode is made as follows. A glass was etched to form grooves of
200 µm in width and 100 µm in depth spaced 300 µm apart. The emission material
which is a mixture of nanotubes and epoxy (50/50), was pressed into the grooves.
After solidification, the excess material was removed, and the surface was polished.
A test showed that the bar surface had about two apices per micrometer. The anode
was fabricated in the following way. A transparent conductive layer applied on
glass was cut into bars of 200 µm in width spaced 300 µm apart. Then these bars
were coated with Al2O3 doped with europium by electrophoresis. The anode–
cathode gap was 30 lm. The emission current necessary for the display to glow,
viz. 76 lA/mm2, was reached at the anode–cathode voltage of 230 V. The current
was studied for stability over 12 h at the constant voltage of 300 V. Fluctuations
did not exceed 8%, and no degradation could be seen.
The display was controlled by two impulses of ±150 V fed to the anode and the
cathode, respectively. A pixel was only switched on when the respective anodic and
cathodic plates were fed with simultaneous impulses. The one-pixel current was
limited to 5 lA, that constraint being established by voltage.
In order to increase efficiency of field emission current and reduction of running
voltage, one should prefer designs with nanotubes directed at right angle to the
effective area of the field emission cathode. The easiest way to achieve this is
electrophoretic deposition [110]. In this case, the threshold voltage reaches 180 V.
However, it should be noted that electrophoresis by no means always provides
verticality of the nanotubes along with their good adhesion to the substrate.
Paper [111] suggests achievement of good adhesion of nanotubes to the base by
deposition of the nanotube in the assembled device. Figure 8.43 presents the
arrangement drawing and the industrial process. A field emission display consists of
cathodic 1 and anodic 2 assemblies divided by spacers 3.
The glass base of anode 4 bears conductive coating 5 of ITO or aluminum. The
cathodic glass plate 6 bears conductive layer 7. The layer is made of an alloy with a
melting point decreasing in vacuum. The alloy is made of copper, aluminum, silver
and chromium. The first conductive layer bears nanotube suspension 8.
Fig. 8.43 The schematic design for roasting of nanotubes in the ready field emission display: 1
cathode unit; 2 anode unit; 3 spacer; 4 anodic plate; 5 conductive cover; 6 cathodic plate; 7
conductive cover (low-melting-point alloy); 8 layer of carbon nanotubes; 9 spring clip; 10 vacuum
furnace; 11 heater; 12 contact wires; 13 power supply
The anodic and cathodic structures are connected with a spacer by spring clips 9.
To be sintered, the design assembled is seated in vacuum furnace 10, cathodic plate
1 directly adjoining infrared heater 11, the anode and the cathode being connected
to power supply 13 with flexible refractory-metal conductors 12.
The caking goes as follows. The furnace pressure of about 10−5 mm Hg triggers
the heating element. The furnace temperature is raised to 350 °C and maintained for
10 min. At the same time, layer 7 melts and the nanotubes from layer 8 begin
sinking into the melt. Simultaneously, the anode-cathode gap is energized. If the
anode-cathode gap is 150 µm, the voltage of 300 V is enough to orient the nan-
otubes along the electric lines of force. In a few minutes of soaking, the furnace
goes off, and the ready device is taken out.
The operational stability of the field emission display directly depends on the
residual-gas pressure inside. That is what causes a necessity in different gettering
schemes. Thereto, paper [112] suggests a separate assembly of surface
non-saturated nanotubes functioning as getters.
It has already been noted that polyacrylonitrile carbon fibers have good field
emission properties. That is why we cannot but consider some designs of displays
based on such field emission cathodes.
Figure 8.44 [113] shows a flat field emission imaging device consisting of a
hermetic glass case (unlabeled in Fig. 8.39a) shaped as a parallelepiped with
insulating optically-transparent front panel 1 made, e.g., of glass or another
visible-light transparent dielectric, the internal surface of which bears a series of
anode buses 2 parallel to each other, field emitters 3 spaced about 200 µm apart in
dielectric quartz-glass substrate 4 serving as the back panel of the case and sepa-
rated from front panel 1 with vacuum gap 5. Each of emitters 3 is a carbon
polyacrylonitrile fiber bundle (unlabeled in Fig. 8.44). The crosswise dimensions of
the bundle are about 100 µm. The bundle is extended along its axis and directed
towards corresponding anode bus 2 along the normal line to the internal surface of
Fig. 8.44 Design of a display screen based on PAN carbon fiber. a Screen design; b emitting unit;
1 front panel; 2 anode bus; 3 field emitters; 4 dielectric substrate; 5 interelectron interval; 6
emitting surface; 7 spacer; 8 cathode buses
front panel 1. The bundle is a group of emitting elements 6 (Fig. 8.44b) opening out
into flat end-face of emitter 3, which is its emitting surface overlooking front panel
1. In this display version, emitting elements 6 are the threadlike fibrils extended
along the axis of each fiber. The fibrils are up to 1 µm in length and 1–5 nm in
diameter. The fiber bundles are put into cylindrical glass shells located in the
apertures of dielectric substrate 4 so that their ends are leveled with the shell edges
and form the end face of emitter 3 overhanging substrate 4 by about 40–400 µm,
the space between front panel 1 and back panel 4 being 0.5–1.4 mm. The spaces are
fixed by spacer 7 shaped as a rectangular projection on lateral case walls 8 standing
on both sides of panels 1 and 4 and tightly glued or welded to them. The opposite
substrate-surface 4 furthest from panel 1 bears parallel bus bars 9 of emitter
power-supply 3, each welded to the other end faces of respective emitters 3 and
crossed with anode buses 2 at a right angle. Anode buses 2 represented by con-
ductive aluminum bars as thick as a few hundreds of nanometers have cylindrical
cavities in the areas opposite respective emitters 3. Cavities are of the same size
with the image pixel, viz. the emitter size when the space between the emitter and
the anode bus is small. To make the cavities, bus bars 2 are sputtered first through
the masks round-screened in the place of the cavities and then through the same, but
unscreened masks. As a result, the cavity bottoms, which are as thin as just a few
hundreds of angstroms, become optically transparent. It is only the cavity bottoms
that the luminescent layer of a luminophore is applied to. Other buses 2 are free
from the luminescent layer. Electrical leads to buses 2 represented by a 50-µm thick
aluminum wire (absent from Fig. 8.44) are ultrasonically welded from one side.
The rectangular shape of the device case makes it possible to gang up large-area
screens intended, e.g., for advertising boards.
The device works as follows. After packaging, gas-ion treatment of the emitting
end face of all emitters 3 by delivery of voltage difference to all buses 2 and 9 for
initial microprofiling with uncoated emitting elements 6, the case is sealed by

pumping-out through the exhaust tube (absent from Fig. 8.44) and subsequent
sealing off of the latter. Delivery of running voltage causes emission of electrons
from emitting elements 6 of emitters 3 between respective buses 2 and 8 towards
anode bus 2. Electrons interact with the luminescent layer in the cavities of bus 2
and cause its luminescence. Light passes through the bus 2 cavity bottom and
shows itself glowing in the respective points/pixels of front panel 1 of the device.
Variation of the height of spacer 7 can change the gap between front panel 1 and
back panel 4 and result in a device with the desired running voltage controlling
emission of electrons from field emitters 3. All basic elements of the display,
including the case materials, its front and back panels, the spacer, the carbon fibers
or the pyrolytic graphite columns, the luminescent-layer material, the anode buses
and the emitter power buses, are produced by the industry that has also mastered the
device-fabrication technology.
This example of implementation is not the only one [114]. In particular, sub-
strate 4 can be fabricated by vitrification of carbon fibers of emitters 3 to form their
shells, rimming of emitters 3 to form close packing, heating and caking of the shells
of emitters 3 in a monolithic block.
Subsequent cutting-out of beads to form blanks of the substrate with emitters of
1–2 mm in thickness, grinding and polishing of the surfaces, as well as ionic
treatment/bombardment to deepen and engrave the emitting surface as aforesaid
complete its fabrication. The other end face of each emitter 3 bears a contact pad
connected to the respective power-supply bus 8 (placed on insulators on substrate 4
not directly but through a transistorized element for power-supply voltage adjust-
ment of each emitter 3. The control inputs of such elements meet the control inputs
of the device and serve to give control signals from the control circuit, e.g., a
computer, through the respective interface. This solution makes it possible to
computerize light-distribution flattening over the image field at front panel 1 when
tuning the device by individual power-supply voltage adjustment of each emitter 3
with the appropriate software installed in the computer, thereby eliminating the
disadvantages of the technology of fabrication of substrate 4 with emitters 3, as well
as the possible heterogeneity of their emissivity.
8.2.3 Triode Display Screens
Despite their simplicity, diode screens generally have low screen glow because of a
necessity to reduce control voltage. That is why the greatest efforts are focused on
development of triode display screens. One section does not provide enough space
to cover the whole range of the structures suggested. So we have to specialize our
study with particular structures illustrating different approaches to solution of the
flat display screen problem.
A flat field emission cathode for a display screen based on graphite field
emission cathodes at rather low, viz. 100 V, extraction voltages was implemented
on the basis of a graphite powder [115]. The graphite powder with the particles of
about 10 lm in size was mixed up with an organic binder to a paste-like state. Then
the paste was mechanically pressed into the orifices of a glass plate.
The orifices were made in a photosensitive glass with the ultraviolet pho-
tolithography technology. The orifices were 0.6–1.5 mm apart for the glass of
*1.5 mm in thickness at the diameter of 200–250 lm at the outer and 100 lm at
the inner edges of the orifices. After the paste had dried, the structure was annealed
at the temperature close to that of softening to remove the organic binder. The
protruding graphite material was removed by mechanical abrasion and polishing. In
this case, the modulator was shaped by molybdenum deposition on the structure
obtained with subsequent explosive etching. That resulted in a self-consistent
alignment of the field emission cathodes and the modulator orifices. The distance
between the edges of the graphite cathode and the modulator was about 3–5 lm.
The element of a flat display based on that structure is shown in Fig. 8.45. To
reduce leakage currents between the cathode and the modulator, the glass is etched
to form a cavity of 3–5 lm in depth. The anode used had a conventional structure
of ITO layers and a luminophore. Then a conventional reverse-contact technology
was applied.
The modulator voltage was 20–50 V. The anode voltage was *1000–1500 V.
A specific feature of this structure is that field emission takes place from those
cathode angles where electric intensity is higher. That is why the light spots on the
anode are concentric rings. In addition, the beam divergence angle was 60–90°. To
avoid overlapping of beam couplings of the adjacent cathodes, the cathode-anode
distance was not longer than the distance between the cathodes.
Since it is quite easy to get powders from different graphite materials, fabrication
of flat display screen was further improved by technology simplification.
Fig. 8.45 Design of an

element of a flat display based
on graphite powder: 1 output
window; 2 ITO layer; 3
phosphor; 4 electron flow; 5
molybdenum modulator; 6
cavity in the glass; 7 graphite
cathode; 8 cathode contact; 9
back plate
Paper [116] suggests reducing the space between the emitters. The structure is
fabricated by the process depicted in Fig. 8.46.
At the first stage, the apertures of 30–100 µm in diameter are made in photo-
sensitive glass by ultra-violet photolithography and etching in 10% hydrofluoric
acid, the glass plate being as thick as 250 µm. The space between the apertures can
vary from 100 µm to 1 mm depending on the purpose of the device.
At the second stage, the apertures/microchannels are filled up with standard
photoresist, and then the plate is polished to remove the excess photoresist from its
surface. At the third stage, the surface is sputtered with a thin molybdenum layer of
up to 3000 Å. At the fourth stage, removal of the photoresist forms a control
electrode. The apertures develop an insulating layer followed by filling up with the
nanotube paste. The insulating layer is necessary to insulate the cathode from the
control electrode. The insulating layer is developed by deposition from an appro-
priate solution. The process consists of repeated dipping, drying and annealing. The
Fig. 8.46 Diagram of process of production of triode structures with nanotubes [97]: 1–7 voltage
on control electrode
glass-coating thickness can be controlled by the number of cycles; e.g., the layer
thickness of 4 µm needs 6 cycles.
At the sixth stage, the apertures are filled up with the carbon nanotube paste, and
the plate undergoes heat treatment and polishing down to the metal layer. And
finally, at the seventh stage, the plate is etched in 10% NaOH solution to form a gap
between the emitter and the control electrode. The uniformity of field emission over
the cathode area is improved by microdischarge aging.
The latter involves gradual increase in impulse voltage at the electrode under low
vacuum, viz. 10−5 mm Hg.
The aging was established to have increased the number of working field
emitters from 60 to 95%. The typical parameters of such a structure are given in
Fig. 8.47.
In addition to embrocation of the carbon paste, there are also a few examples of
its application in fabrication of field emission cathodes by printing, viz. [117, 118].
A modification of this method is pulverization of the binder solution [119].
However, this method has not come into wide-spread acceptance because of its
relative complexity and marked heterogeneity of field emission over the field
emission cathode surface.
Screens are also suggested to be fabricated by electrophoresis [120, 121]. This
method is based on an electrolytic bath consisting of 600 g nickel sulphide, about
5 g nickel chloride, 30 g sodium hypophosphite, about 40 g boric acid and about
1 g saccharin per liter of distilled water. This solution is mixed in with about 40 g
carbon nanotubes [121]. Electrophoresis goes on at the temperature of 50 °C.
It has to be noted that this flat-screen fabrication method has not come into
wide-spread acceptance, either.
What attracts implementers best of all is direct carbon nanotube growing.
Such growing is usually done on already prepared structures at the final stage
[122, 123].
Nanotubes are grown according to the following well-known technology,
namely by low-temperature CVD on a pre-sputtered catalyst layer [124]. The
pre-sputtering of the latest version is shown in Fig. 8.48 [125]. Substrate 1 is one by
one coated with the following components, viz. cathodic electrode 2 usually made

anode current of the operating
electrode (current on one
emitter) on the voltage on the
electrode control
(Uanode = 500 V)

of the modulator part of the
display screen with direct
cultivation carbon nanotubes.
1 substrate; 2 cathodic
electrode; 3 resistive layer; 4
insulator; 5 control electrode;
6 buffer layer; 7 catalytic
layer; 8 carbon nanotubes
of Mo or Cr, followed by resistive layer 3 of an amorphous silica film flattening

field emission over the field emission cathode array. Subsequent sputtering pro-
duces a dielectric layer 4 and a metal layer 5. Conventional photolithography and
selective etching form windows in layers 4 and 5. The bottoms of the windows are
first coated with a buffer layer 6 and then with a catalyst bed 7.
The buffer layer of 10–3000 Å is made of a metal well-adhesive to the substrate
material, e.g., Ac, B, Ga, Ju, Ti, Mo, Cr. The catalytic layer as thick as 2–100 Å is
formed of the materials conventional for these purposes, viz. Fe, Co, Ni.
After that, the carbon nanotubes are grown by conventional CVD methods.
A similar but more sophisticated method produces the elements of the display
screen described in a long series of patents, e.g. [126, 127]. Paper [128] features
formation of carbon layers of a field emission cathode and a modulator due to
different growth rate of the layers.
As all the above listed methods of fabrication of flat carbon-based field emission
displays have not led to their practical implementation yet, attempts to improve
these or those parameters of the displays are still going on. Some of these efforts are
considered and exemplified below.
To kill two birds with one stone, viz. improve stability and uniformity of an
electron beam and reduce running voltage, paper [129] suggests using carbon-metal
films as a field emission cathode. Metal powders are selected from the group of Fe,
Co, Ni, Pt, Pd, Ir, Mo, Ti, Cr, W, Ta.
The composite can be sputtered by any known method. The field emission
properties of the composite depend on the metal quantity in the composite.
More stable is the field emission cathode consisting of a mixture of one-wall and
multiple-wall nanotubes [130]. The same purpose can be attained by increase in the
space between the modulator and the field emission cathode surface since the
electric strength of this space is thereby improved [131].
To improve efficiency and controllability of the display, it is necessary to reduce
the leakage current between the field emission cathode and the modulator, as well
as eliminate a possibility of short circuit between these electrodes (Fig. 8.49) [132].
This aim is attained by special resistive layer 4 applied over the entire internal
Fig. 8.49 Design of emission

part of a triode design of the
display with small leaks: 1
substrate; 2 cathode lead; 3
dielectric; 4 resistive layer; 5
control electrode; 6 field
emission cathode from
nanotubes
surface of the cathode aperture and the surface of control electrode 5 and dielectric
3. Moreover, the layer also blocks the generated ultraviolet rays that can cause
leakage current between the field emission cathode and the control electrode.
Short circuits between the field emission cathode elements, e.g., nanotubes, and
the control electrode are eliminated by increase in running voltage [133]. The effect
of this voltage generates an increased current through the defect, which eliminates
the latter; i.e. the nanotube is burnt off. The voltage is increased due to heating of
the ballast resistor, the latter reducing its resistance when heated.
Large-size displays are impossible without a large number of spacers between
anodic and cathodic plates [134].
However, the electron bombardment can charge the spacer surface, which
increases the noise of the device and even puts it out of operation. The disadvantage
is eliminated by mounting of some special electrodes [135] connected to the control
circuit on the lateral surface of the spacers, the latter being as high as 700–
3000 µm. They change the spacer potential in such a way as to liquidate its
charging by electron bombardment.
The display screens considered above had an emitter system well-defined in the
plane. That can cause serious technological difficulties. Paper [136] suggests
forming a pixel or an image element by a large number of chaotically located
cathode-modulator assemblies. The assembly is fabricated as diagrammed in
Fig. 8.50.
First metal layer 2, which serves as the contact of the field emission cathode, is
applied on substrate 1. It is usually made of metals, viz. Ni, Cu, Ag, Co, Fe.
Subsequently, metal layer 2 is sputtered with masking material 6, viz. the par-
ticles of glass, plastic, metal or ceramics. The particles of the masking-material can
be solid or liquid. That is followed by application of the insulating layer and the
second metal layer 4 intended to form a triode structure. Then the masking material
is removed 6. That can be done in water or a solvent with involvement of an
ultrasonic bath. The final process is application of the field emission cathode
material 5 by any known method. The result is chaotically arranged emitters 1
within one pixel 2 (Fig. 8.51).
Fig. 8.50 Scheme of

production of triode unit with
any location of field emitters:
1 substrate; 2 first metal layer;
3 isolating material; 4 second
metal layer (modulator); 5
material of the field emission
cathode; 6 masking material

of a pixel with chaotic
arrangement of emitters: 1
emitter (triode design); 2 pixel
8.2.4 Multielectrode Display Screens
Naturally, introduction of every additional electrode into the screen structure pushes
up the cost. However, the resulting or desired improvement of image brilliance,
brightness and sharpness compensates for these efforts and costs over and above. In
parallel, most structures employ another electrode with orifices coaxial with the
control electrode. It is designed to focus the beam coupling because a wide angle of
divergence of field electrons emitted is well known. The modulator orifices usually
reiterate those of the control electrode and the field emission cathode [137, 138].
The best results can be obtained by focusing with a quadrode system [139]. It
includes three apertures formed by the structural components of the cathode plate.
The focusing electrode can be used not only for its intended purpose, i.e. focusing,
but also to protect the field emission cathode from discharges [140]. In this case, the
aperture of the focusing electrode is smaller than that of the control electrode. In
this case, it is not only focusing that is possible, but also interception of some
electrons by the focusing electrode.
However, it is totally unnecessary to place the control electrode near the field
emission cathode. This electrode can be represented by a focusing one [141]. In
particular, the impulse anode voltage of 1.5 kV, the earthed cathode, the
focusing-electrode voltage of −20 V and the control-electrode voltage of +80 V
produce the electron-beam diameter of 11 µ, the anode-cathode space being
1.1 mm.
The focusing electrode can also be located in the same plane with the control
electrode [142]. In this case, it is expedient to design all the electrodes as bars.
The idea of focusing and protection of the field emission cathode is most
completely implemented in more sophisticated structures, e.g., [143]. Figure 8.52
presents a circuit of such structures. Cathodic I and modulator II electrodes are
fabricated by conventional processing techniques. Thickness of the elements is as
follows: substrate 1 is 0.5–5 lm; cathode lead 3 is 0.1–1.0 µm; the dielectric layer
is 0.5–20 µm; modulator 5 is 0.1–1.2 µm.
A specific feature of this structure is rather sophisticated design of the focusing
electrode consisting of cone-hole metal gauze 6, the internal surface of the holes
being coated with thin dielectric layer 7. Structurally, the diameter of outlet D2 is
from one to three times larger than the total thickness of grid 6 and dielectric layer
7. That screens the field emission of cathode 2 from the effect of the anode
potential. In addition, dielectric layer 7 makes an additional focusing effect.
The difference between diameters D1 and D2 additionally densifies the electron
flow. The parameters of this field emission device are further improved by intro-
duction of a thin-film control transistor into each field emitter [144] (Fig. 8.53).
The display-screen element scheme shown in Fig. 8.53 is clear enough and does
not need any comments. Technicals are detailed in [144].
Fig. 8.52 Schematic design of a display screen with a focusing electrode [124]: Structural
components: I cathode; II control electrode; III focusing system; IV anode. Design elements: 1
cathode substrate; 2 field emission cathode; 3 cathode lead; 4 dielectric layer; 5 modulator; 6 metal
gauze; 7 dielectric layer; 8 anode substrate; 9 conductor layer of the anode. D1 and D2 are
diameters of apertures of the modulator and the focusing electrode, respectively
Fig. 8.53 Diagram of the cathode part of a display screen with control of current of the emitter of
the thin-film transistor: 1 field emission cathode; 2 lead of a source; 3 source of the transistor; 4
control electrode; 5 active layer; 6 drain of the transistor; 7 base; 8 isolating layer; 9 electric drain
terminal; 10 internal isolating layer; 11 dielectric layer; 12 metal gauze
8.2.5 Lateral Display Screens
A number of parameters makes lateral display screens preferable to free-flow ones

(Sects. 8.2.2–8.2.4). The first reason is a possibility to reduce the field emission
cathode-control electrode space. The display design is exemplified in Fig. 8.54
[145, 146]. The basic feature of the screen design is that field emission cathode 8

of a display screen with lateral
design: a overview diagram;
b the enlarged drawing of the
field emission cathode and the
control electrode; 1 cathode
base; 2 electrical lead
channel; 3 spacer; 4
phosphor; 5 anode; 6 anodic
plate; 7 cathode lead; 8 field
emission cathode; 9 control
director electrode; 10
lean-through of the control
electrode; 11, 12
metal-resistive layers; 13 field
emission cathodes
and control electrode 9 are placed on the plane of glass support 1. Their leads 7 and
10 are located on the same plane. The end faces of electrode leads 7, 10 are coated
with resistive materials 11, 12.
In this case, the specific resistance lies within 10−7–10−4 X m. Electric resis-
tance is improved by introduction of some doping substances, e.g., B, P or poly-
tetrafluoroethylene (PTFE), into metal coating (11, 12). The Ni–B–P system can be
based on Ni or P, the concentration ranging from 3 to 40% or from 7 to 40%,
respectively. If doping with Ni involves PTFE, the PTFE concentration is 0.1–30%.
In itself, field emission cathode 13 is carbon nanotubes fixed in metal layer 11.
Control electrode 9 is almost the same as field emission cathode 8 in terms of
structure. That makes it possible to swap the cathode and the control electrode by
simple switching when the field emission cathode becomes less active.
As shown before, lateral designs are characterized by reduction of the
cathode-control electrode gap. Its value can range from 500 µm [147] to µm units
[146, 148]. It is very useful to combine different control circuits. That complicates
the device design, but can seriously improve the image quality. E.g., paper [149]
suggests a display design combining the lateral and the classical method of control
of the emitted electron beam (Fig. 8.55).
Field emission cathodes 2 consisting of carbon nanotubes rest on metal lead-out
bars 1. Control electrodes 6 lie in immediate proximity from the field emission
cathodes and are electrically connected to the external terminals through bus 4. The
electron beams are additionally focused by electrode 5, which works in a con-
ventional way (Sect. 8.2.3). The anode of this design has two elements improving
the image quality. It is, firstly, black mask 14 improving the image contrast; and
secondly, thin aluminum layer 10, which increases the image brightness. The image
Fig. 8.55 The schematic design of a display with a combined focusing system: 1 cathode lead; 2
field emission cathode; 3 bottom (cathode) base; 4 control electrode lead; 5 focusing electrode; 6
control electrode; 7, 8, 9 layers of dielectrics; 10 metal layer (e.g. Al); 11 anode; 12 anode base; 13
phosphor; 14 black matrix
quality can also be improved if the image element is formed with involvement of
either a few field emission cathodes or a few control electrodes [150].
8.2.6 Screens with Reversed Control Electrodes
The position of the control electrode behind the field emission cathode under an
insulator layer sets this structure apart from others. First of all, it excludes a dis-
charge of electricity between the field emission cathode and the control electrode
through the vacuum gap.
One of the first structures of such a screen was presented in papers [151–153].
Figure 8.56a shows the arrangement drawing of the cathode part of the screen,
while Fig. 8.56b demonstrates the distribution of electric potential and the electron
path of such a cathode-modulator assembly.
Glass substrate 1 is coated with an aluminum layer of 1500 Å in thickness, and
then a standard photolithographic process produces control electrodes 2 of 400 lm
in width. Then insulator layer 3 comes. In this case, it is a polyimide layer of 13 lm
in thickness. The latter is coated with a cathode layer of aluminum 4 by
electron-beam vaporization, and the cathode layer of aluminum then develops into
cathode bars of 390 lm in width. The cathode bars are smeared with a paste with
single-walled carbon nanotubes through a metal grid with a mesh of 20 lm.
Subsequent thermal treatment at 350 °C enables the nanotubes to project above the
surface of the field emission cathode coating due to annealing of the binder.
Figure 8.56b shows the estimated data of electric potential and electron paths for
different voltages at the control electrode. The picture shows that the electric field
concentrates in the corners of the cathodic electrode. That is why electrons emit

of a screen with return
operating electrode: (a) and
distribution of electric
potential and trajectories of
the emitted electrons;
(b) [133, 134]: 1 glass
substrate; 2 control electrode;
3 isolating electrode; 4
cathodic electrode; 5 carbon
nanotubes; 6 electron
trajectories; 7 electric field
distribution
mostly from the corners of the cathode rather than its central part. This effect can be
minimized by reduction of width of the cathode assemblies [154].
The problem of edge effects can be solved by application of rounded field
emission cathodes [155]. The arrangement of a display based on such field emission
cathodes is shown in Fig. 8.57.
Glass substrate 1 bears control electrodes 5 represented by ITO buses. Dielectric
layer 2 coats the entire surface of the cathode plate. The dielectric bears leads 3 and
Fig. 8.57 Construction

diagram of a display screen
with return operating
electrode [136]: 1 cathodic
plate; 2 dielectric; 3 cathode
lead; 4 field emission cathode;
5 control electrodes; 6 anodic
plate; 7 anode; 8 phosphor; 9
black layer
field emission cathode 4 laid at right angle to control electrodes. The cathode leads
are made of two layers. The first layer 3a is made of aluminum to provide a good
electric contact and protect the dielectric from cracking. Second layer 3b is made of
chromium to protect the first layer from destruction by possible electrical break-
down and improve reliability of the cathode lead. Field emission cathode 4 is made
of carbon nanotube containing paste and applied by screen printing. Anode 6–9 is
fabricated in a conventional way.
If further focusing is necessary, a special electrode reducing the field-electron
spreading angle can be applied to the dielectric plane [156].
8.2.7 Thin-Film Field Emission Cathode-Based Displays
Thin-film and blade field emission cathodes have an advantage over pointed ones
(Sect. 5.3). That is why implementers wish to design a display screen based on
them. Figure 8.58 presents the design of one screen element based on edge field
emission cathodes [157]. The basic technological stages of such a screen are
– formation of a thin layer of control electrode 2 (Nb);
– application of a photoresist, distribution of control electrodes, etching; further
layers are applied in nearly the same sequence, the dielectric layer consisting of
SiO2 and the field emission cathode being a metal film of Nb or Mo.
This design is simpler in fabrication than that based on pointed emitters. Its
emission characteristics display quite good field emission parameters (Fig. 8.59).
This technology can be assumed as the basis for home television screens of the
color pixel ranging from 0.6 to 1.0 mm, the screen ranging from 20 to 30 in.
The display-screen element design based on thin-film carbon field emission
cathode is shown in Fig. 8.60 [158]. Cathode-assembly glass substrate 1 bears the
layers of cathode leads 2 of molybdenum (Mo). Then the cathode leads are coated
with resistive layer 3 of a–Si, which levels the field emission characteristics of
particular cathodes. The subsequent layer is a-carbon, its end face functioning as a
field emitter. The cathode-anode space being 20–30 µm, field emission can be
initiated with about 600 V. This design can be further developed by introduction of
Fig. 8.58 Design of an element of a flat display screen with a blade field emission cathode
(cathode part) [138]: 1 glass substrate; 2 control electrode; 3 layer of dielectric; 4 field emission
cathode film; 5 passivated layer; 6 field emission cathode

field emission current density
on voltage on the control
electrode
Fig. 8.60 Element of a

display screen with a thin-film
field emission cathode made
of carbon materials [139]: 1
glass substrate of the cathode;
2 cathodic electrode; 3
resistive layer; 4 field
emission cathode film; 5
phosphor; 6 anode; 7 anodic
plate
a control electrode below the level of the field emission cathode [158]. An inter-
esting combination of a thin-film field emission cathode with a reverse anode is
suggested in [159, 160]. The circuit diagram of the device and the pixel structure
are given in Fig. 8.61.
The cathodoluminescent screen has a vacuum shell made of two flat glass faces:
transparent plate and a substrate with electrodes, a flat, thin end-face emitter and an
anode coated with a conducting layer and a luminophore forming a light-emitting
cell, viz. a pixel, located at a small controllable distance from each other in the
parallel planes connected through anode buses and emitter buses with terminals
intended to connect to electric-voltage sources, a dielectric frame hermetically
connecting the face plate and the substrate. In such a screen, the anodes and the
emitters are shaped as two combs enclosed into each another, the emitters with
connectors being located in the upper plane of the substrate, the anodes and the
anode buses being located at the bottom of the grooves of predetermined depth in
the substrate from the side of the upper plane, the free space of the anode-bus
grooves being filled up with a dielectric to form a floor-space area hosting con-
nection emitter buses.
Fig. 8.61 Schematic diagram of the device (a) and structure of the pixel (b): 1 glass base; 2
anode; 3 phosphor; 4 emission surface; 5 cathode lead; 6 anode bus; 7 protective resistor; 8
cathode bus
In addition, the anodes and the emitters can have a form of any geometric figures
enclosed into each other, e.g., in a form of spirals; a cathodoluminescent screen can
be additionally supplied with a protective resistor connecting the emitter bus with
each pixel emitter in series; the selected luminophore has sub-micron scale grains;
the selected low-voltage luminophore has the running voltage of no more than
50–70 V.
To this end in view, the well-known cathodoluminescent screen fabrication
method, including consecutive formation of the anode bus and the emitter bus, the
active pixel structure, installation of the intermediate dielectric stringer, hermetic
sealing of the device along the stringer perimeter, is modified by formation due to
grooving of the substrate by wet etching, coating of the groove bottom with a
delivery layer and filling up of the free space of the grooves with a dielectric, e.g.,
fusible glass, to form a planar structure. Then the entire surface is coated with a
resistive layer followed by formation of the protective-resistor topology by pho-
tolithography. Then the entire surface is coated with emitting and conducting
layers, the latter being used to form emitter buses orthogonal to the former anode
ones, by photolithography. Subsequent photolithography forms an active structure
of anodes and emitters in a pixel by etching in the windows of the conducting layer,
the emitting layer and the glass cavity at the predetermined depth, after which beam
sputtering deposits a conducting layer contacting with the anode bus. Then the
photoresist is removed together with the conducting layer sputtered upon it, and a
layer of low-voltage fine-grain luminophore is applied.
Application of high-precision glass-grooving technology, involvement of
high-precision fine-grain luminophores and consequent small gaps between the
emitter end-face and the anode makes it possible to use low-voltage luminophores
with electron energy below 100 V, thereby providing a long cycle of the device. If
the luminophores used are low-voltage, there is no need in triode structures. Diode
structures at low voltages provide a high resolving ability. Fabrication of
cathodoluminescent displays is streamlined by formation of the anode bus of a

required resistance, viz. its thickness, at the first stage.
In addition, the capacitance between the cathode bus and the anode bus is
reduced by the small width of the anode bus and a thick layer of a dielectric with a
low dielectric permeability, whereby the conducting layers of anodes and emitters
do not cross in the active pixel area, which minimizes the capacitance between the
anode and the cathode down to a negligible value.
In this design of the diode structure, an individual pixel consists of non-crossing
emitters and luminophore-coated anodes; there is no capacitance between the layers
of the emitter and the anode because they do not cross; the number of elements in
the pixel mostly depends on the size of the minimum reproducible element, which
enables an increase in the length of the emitting edge with a possibility of sec-
tionalisation into small areas switched through a current-limiting resistance or a
non-linear thin-film component to the common bus; a high redundancy level in the
pixel, the latter consisting of several emitter sections independent from each other.
To form pixel on one substrate and maintain high resolving ability in absence of
the cross effect, the upper plane of the substrate is a plane of connector-ended
emitters, and the lower plane results from grooving in the substrate at the prede-
termined depth by local etching from the side of the upper anode-bearing plane
formed by a conducting layer coated with a line-connected luminophore. To reduce
input capacitances and the risk of short circuit, the pixel unit cells formed by
emitters and anodes are located so that the electrode planes in the cell are crossed
only in the point of bus crossing, and short circuits between the anode buses and the
emitter buses are prevented by filling up the etched anode-bus grooves with fusible
glass and subsequent planarization of the surface before application of connecting
emitter buses.
To provide redundancy in the glow elements, protective resistance is activated in
series with each emitter segment in the pixel. To fix normalized gaps between the
anode and the emitter, the anodes are precisely coated with a submicron-grain
luminophore. To reduce the running voltages of drivers, considerably cut the cost of
the display unit and extend the life cycle of the screen, the latter employs a
low-voltage luminophore with a running voltage of no more than 50–70 V, when
the luminophore degradation rate slows down, cathode poisoning relents, the life
cycle of the device goes up by an order of magnitude.
The cathodoluminescent screen follows the route, including conducting layer
sputtering (Mo, Cr) of 0.5 µm in thickness; plasma-chemical deposition of amor-
phous carbon of 20–80 nm in thickness; photolithography; etching of amorphous
carbon; etching of the conducting layer; etching of the glass support to the depth of
3–6 µm; electron-beam evaporation of the conducting layer (Al, Cr) of 0.3 µm in
thickness; explosive removal of the photoresist with the sputtered conducting layer;
cataphoretic deposition of luminophores reduced to etching of glass followed by
standard operations of the planar silicon-chip cycle and application of a self-aligned
technology on each transition. The technology is self-aligned, which provides
automatic arrangement of the emitter and the anode one above the other.
To an adequate degree, this idea is also implemented in the design of the

thin-film field emission cathode for a flat display presented in [161].
8.3 Microwave Devices
Fabrication of microwave devices based on field emission cathodes is very difficult

because the physical limits imposed by field emission itself are exacerbated by
specific requirements of microwave engineering [162]. That is why this section is
only going to outline the problem and some areas of its solution [163].
The size of emitters, which is extremely small in comparison with the wave
length in the frequency band of 1–100 GHz, makes it possible to consider them as a
series of localized shunt admittances. In particular, a cone of 5 µm in height is 0.3%
wave length of the frequency of 100 GHz in a SiO2-type dielectric, which makes it
possible to use quasi-static approach.
If we consider a rectangular array of emitters spaced as far as d apart in both
directions, viz. x and z (Fig. 8.62) [164], then the wave propagation in the
Z direction can be represented as
X
1 X
1
2pnx jbmz
Vðx; zÞ ¼ anm cos e ; ð8:1Þ
n¼0 m¼1
d
Fig. 8.62 a Schematic

drawing of a field electron
array of field emission
cathodes; b ladder distribution
of resistances and shunt
capacitances; c simplified
ladder circuit with three
chains used in the analysis
where bm ¼ jc þ 2pm=d and anm ¼ an ðmÞ;

pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
c ¼ ðR þ j-LÞðj-cÞ: ð8:2Þ
R, L, C are the resistance, inductivity and unit-length capacitance of the transmitting
line, respectively. - is the angular frequency, c being the transmission factor.
The wave equation for the transmitting line loaded in a form of a rectangular
matrix of periodic shunt admittances YS is
!
X
1 X
1
r V ðx; zÞ Z Y þ Ys
2
dðx vd Þdðz ld Þ V ðx; zÞ ¼ 0; ð8:3Þ
v¼1 l¼1
where Z ¼ -ðR þ jxLÞ; Y ¼ j-C=-.

Inserting (8.1) in (8.3) and using the orthogonal properties of cosine and com-
plex exponential functions, we get the dependence
" #
d2 ð2pnÞ2 X1 X 1
anm en bm þ
2
þ ZY þ ZYs alv ¼ 0; ð8:4Þ
2 d l¼o v¼1
where en = 2 for n = 0 and e = 1 for n > 0. It represents the equation of a homo-

geneous matrix in the units of unknown coefficients anm. Limitation on the number
of terms to N + 1 in direction X and 2M + 1 in direction Z gives the finite-size
matrix equation. After a few algebraic operations, these operations produce
XN XM
1
1 ¼ 0; ð8:5Þ
b
n¼0 m¼M nm
where
" 2 #
den 2pn
bnm ¼ b2 þ þ ZY
2ZYs m d
" #
d 2 en 2pm 2 2pm 2
¼ jc þ þ þ ZY : ð8:6Þ
2ZYs d d
The numerical solution of this equation gives transmission factor c for some
frequency and the space between the emitters. If the space between the emitters is
much less than the wave length, term 1/B in (8.4) is much more than the others. So
B 1. As a result, the transmission factor is
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
c¼ Z ½Y þ ðYs =d 2 Þ ¼ ðR þ j-LÞ j-C þ Ys -=d 2 : ð8:7Þ
8.3 Microwave Devices 485
Comparison between (8.2) and (8.7) shows that the load connected with the
emitters simply increases the average conductance of the transmitting line per unit
length. If the space between the emitters is not substantially less than the wave
length, the quotient approaching 1 THz, (8.5) and (8.6) are used.
The equivalent shunt admittance caused by conical emitters can be considered as
infinite multiple-link distribution of the resistors shunted by capacities (Fig. 8.62b).
The differential elements of the circuit connected with the differential increase in the
emitter height can be calculated by the finite-difference method.
The capacitance structure is made up of three components (Fig. 8.62c), viz.
ring-type capacitor Cer, cone capacitor Cec and stud capacitor Cep. The values of
these capacitors, their efficient arrangement within the structure can be calculated
by analytical approximations or from the experimental measurements of the
capacitance in scale models. The resistance and inductivity of the cylindrical or
conical part between the efficient positions of the capacitors can be calculated
analytically on the basis of the skin layer in the emitter material. Having found out
these values, one can calculate shunt admittance as
1 1
¼ Zep þ ; ð8:8Þ
Ys Yep þ Zec þ
1
1
Yer
Yec þ
1 þ Zer Yer
where Yer ¼ j- Cer Zer ¼ Rer þ j-Ler etc.

If the emitter has no stud, Yer and Zer are equal to null.
The material in the electrode apertures having been lost, the effective specific
resistance of the control-electrode film goes up. At the same time, the specific
resistance of the substrate remains almost the same. The total resistance per unit
length is approximately described as

1 1 ð6 11:804a þ 16:642a2 17:24a3 þ 6:815a4 Þ 1

R¼ þ ; ð8:9Þ
x r1 d 1 ð6 12a þ 7a2 a3 Þ r 2 d2
where a = 2b/d and b is the radius of the aperture in the control electrode.
In addition, the apertures in the control electrode increase the inductivity per unit
length:
2t3 d1 d2
L ¼ lo þ l1 þ l2 : ð8:10Þ
x ð1 þ f - Þ 2f- - 2-
The unit-length capacitance depends on two factors. Firstly, the capacitance

caused by the parallelism of the planes of transmitting line Cpp decreases only
insignificantly due to the apertures in the control electrode. Expression

Cpp ¼ ðe3 -=t3 Þ 1 pb2 =d 2 ð8:11Þ
provides reasonable approximation of this effect. Secondly, the emitters increase the
capacitance due to complex shunt admittance Ys. Addition of these two effects
result in

- - Cs
C ¼ Cpp þ Im ð Y s Þ jRe ð Y s ÞCpp þ C s j ; ð8:12Þ
- d2 d2 -
where Ys ¼ Gs þ j-Cs :
The emitter losses caused by the distributed capacitance become complex.
The initial estimates of the field emission cathode arrays at 10 GHz are given
below.
A good example is Spindt field emission cathode-based systems, which consist of
molybdenum cones of 0.3 µm in height, a SiO2 dielectric of 1 µm in thickness and a
control electrode represented by an M0 film as thick as 0.3 µm. In addition, let us
take account of the aperture radius of 0.4 µm, the basic rounding-off radius of
0.45 µm and the cone height of 1.3 µm, which provide that the apex peak is above
the control electrode surface. In this case, the parasitic shunt capacitance of this
structure is analytically estimated as Cer = 4.8 10−17 F and Cec = 1.8 10−17 F,
while the resistance is Rer = 0.01 X and Rec = 0.08 X. Parasitic inductance is about
0.2 GHz and negligible at 10 GHz.
8.3.1 Field Emission Microtriodes and Amplifiers
Development of field emission cathode matrix-based microtriodes is the main area

of vacuum microelectronics for many reasons. That is, first of all, because
microtriodes are the simplest possible configuration of amplifying devices based on
field emission cathodes. Furthermore, the entire design can be integrated into one
structure of a very small size.
A self-excited generator based on a microtriode with a field emission cathode is
diagrammed in Fig. 8.63 [165].
The device is quite similar to the conventional triode generator; however,
non-linear resistor R is included in the LC circuit to prevent emitter destruction.
The non-linear volt-ampere characteristic of resistor S(V) is expressed by the
exponent, viz.
I ¼ Io ½expðV=Vo Þ 1: ð8:13Þ
It should be noted that the ideal volt-ampere characteristic of a semiconductor

diode follows formula (8.13), where V0 = kt/e is the thermal potential.
This section does not consider microtriode structures (e.g., [166]) since they are
sufficiently covered in Sects. 8.1 and 8.2.
As a rule, the microtriode of microwave devices, e.g., amplifiers [167], is
designed so that the control electrode also functions as a resonator grid.

diagram of a generator based
on a microtriode with field
emission cathodes: 1 emitter;
2 control electrode; 3 anode;
b analog model of the circuit:
1 linear amplifier; 2 nonlinear
amplifier; 3 external
sinusoidal control signal
Fig. 8.64 Microwave

amplifier with field emission
cathodes: 1 input resonator; 2
microwave generator; 3
output resonator; 4 collector;
5 load; 6 high-voltage source
of direct current; 7 field
emission cathode; 8 square
wave generator
Such as amplifier is diagrammed in Fig. 8.64. The value of the microwave field
and the voltage of the high-voltage source are selected so that the total electric field
intensity is sufficient to excite field emission from the cathode. The field emission
cathode has a ring configuration consisting of 3 concentric rings made of pointed
field emission cathodes, the comb rings being attached to the concentric
piezocrystals connected to the rectangular-pulse generator 8. It controls the position
of the field emission cathodes in the resonator, which contributes to increase in the
average power of the amplifier.
The amplifier works as follows. Low-power microwave generator 2 in input
resonator 1 modulates field emission current from cathode 7 that goes to collector 4.
The amplified signal in the output resonator 3 goes to load 5.
Another promising amplifying device is a klystrode (AKA inductive output

tube), i.e. a high-power microwave amplifier combining the operating principles of
a quadrode and a klystron. In klystrodes, field emission cathodes increase the
operating frequency up to 10 GHz [165] at the minimum. A field emission
cathode-based klystrode is diagrammed in Fig. 8.65. Moreover, field emission
cathodes increase the amplification coefficient and reduce the running voltage.
It seems very promising to use field emission cathodes in microelectronic
modifications of crossed-field microwave amplifiers. The crossed electrostatic E0
and magnetostatic B0 field amplifier is diagrammed in Fig. 8.66.
The upper plane hosts the slow-wave structure 2, while the lower plane hosts a
field emission cathode. The electrons emitted during the accelerating phase of the
electromagnetic wave come back to the cathode surface, where they are adsorbed.
The electrons emitted during the deceleration phase reach the slowing structure
donating their energy to the wave. The main advantages of such amplifiers are
1. There is no need in an extraction grid, therefore a relatively large grid-cathode
capacitance does not limit the operating frequency.
2. Random emitting surfaces such as diamond-like carbon films, carbon-nanotube
films etc., can replace regular field emission cathode arrays.
3. There is no need in focusing systems.
4. Such devices are characterized by high performance and high output power.

of a klystrod with a field
emission cathode: 1 field
emission cathode; 2 output
resonator; 3 collector
Fig. 8.66 Crossing fields

amplifier with a field emission
cathode: 1 field emission
cathode; 2 wave-slowing
structure
8.3.2 Magnetrons
The magnetrons can be triggered by injection of initial electronic current much

lower than that of the magnetron in the operating mode, to the interaction space. Up
to now, only thermionic cathodes providing the necessary initial current of several
milliamperes are used for this purpose. Usually they are effective hot cathodes
whose emission-current density is (101–102) A/cm2.
As for field emitters, they can provide higher values of the current density
reaching as much as (104–106) A/cm2 as shown in Fig. 8.67 comparing the char-
acteristic current-density values of hot cathodes and field emitters.
An idea of a magnetron triggered by a field emission cathode is illustrated in
Fig. 8.68. A magnetron reflector is replaced with a disk field emission cathode.
When anode voltage is supplied, the end face of the field emission cathode gen-
erates a strong electric field that causes field emission current.
These primary field electrons driven by the magnetic field bombard the basic
cathode of the magnetron and cause secondary electronic emission providing the
necessary current of the magnetron in the operating mode.
The magnetron is triggered almost immediately. The triggering time is stipulated
by impulse formation time and amounts to a sub-microsecond value. When hot
cathodes are used, the time required for this type of magnetron to reach full
operation is 150–180 s from the moment of the cathode heat supply. Naturally, field
emission cathodes do not consume heat power, whereas the hot-cathode heat power
sufficient to trigger a magnetron amounts to tens of Watts.
Efficiency of microwave devices needs improvement, and so does their
speed-in-action. Microwave-device heat power is nearly the same or even
exceeding their high-voltage power. In addition, the weight and dimensions of
Fig. 8.67 Work areas of

electronic devices with
thermionic emitting cathodes
I and with field emission
cathodes 2 (j is density of
current, E is intensity of
electric field at the cathode)

explaining how to “instantly”
turn on a magnetron using a
film field emission cathode:
a ordinary magnetron with a
thermionic emitting cathode;
b the same magnetron with an
additional field emission
cathode; 1 cathode; 2
reflector; 3 heater; 4 anode; 5
cathode holder; 6 disk field
emission cathode
filament-supply sources are also comparable or even exceeding the weight and
dimensions of high-voltage sources. Reduction of these parameters is especially
important for independent and portable radio transmitters. Application of field
emission cathodes removes all the heat problems connected with power-supply
sources.
The first magnetron structures were disk edge field emission cathodes. However,
more up-to-date magnetron structures employ disk film field emission cathodes
[168].
The magnetron is diagrammed in Fig. 8.69. Guide core 1 hosts components with
secondary electronic emission (secondary electronic emitters 2) represented by
tungsten-based impregnated materials added with barium, or alloys and compounds
based on platinum-group metals, such as platinum, palladium, iridium, osmium,
having addition agents of a material of alkaline-earth, alkaline or rare-earth (Ba, Ir,
Ca, La) elements. Between emitters 2, there are field emitters 3 projecting above
them by 5–20% interelectrode gap. They are shaped as beads consisting of a
conductive film and dielectric film 4 surrounding it from the sides.
The dielectric film protects the lateral surface of the field emitter not only
mechanically, but also electrically, viz. when treated with positive ions, it is charged
with homogeneous potential that generates a field repelling the inbound ions.
The dielectric film can be away from the lateral emitter surface by the distance
equal to the conductive film end-face radius. According to computations, that
provides additional increase in the resistance to ion-charge drain when the lateral
emitter surface is totally protected against the direct hit of residual-gas ions. If the
gap exceeds the end-face radius of the conductive film, its lateral surface can be

of a magnetron with disk film
field emission cathodes: 1
directing core; 2 secondary
field emitter; 3 field emission
cathode; 4 dielectric films; 5
cylindrical anode
bombarded. If the gap is smaller than the radius, the dielectric film provides total
geometrical and electrical screening of the emitter surface.
Resistance can be increased by elongation of the ion-charge drain path due to
dielectric-film crimping. The computations made by mathematical simulation show
that the ion flow to the field emitter can be totally excluded if the dielectric films
project over the conductive-film surface by the value equal to the radius of the film
end-face.
Increase in the projecting part of the dielectric film by the value exceeding the
radius of the conductive-film end face reduces the field intensity and current at the
emitter.
Film 4 can be made of boride-, nitride-, oxide-type dielectrics, e.g., Si3N4, SiO2,
BN, AlN, Al2O3; or the dielectrics with alkaline and alkaline-earth additives (Ba, K,
Cs, Li, Na). Base 1 with secondary-electron and field emitters is located in cylin-
drical anode 5 with a working clearance, dielectric film 4 of the projecting part of
the field emitter being as far away from the lateral emitter surface as the radius of
the emitting end-face of the conductive film.
Film 4 can project over the field-emitter surface. The number of the triggering
field emitters in a device is determined by the value of the necessary starting
current, i.e. the device parameters. Field emission is provided by a constant field
generated by the anode near the cathode. Electrons collide with residual-gas atoms
in the cathode-anode space to form ions. Positively charged ions accelerate them-
selves to the field emitter along the equipotential power lines and move towards the
lateral field-emitter surfaces, thereby bombarding the dielectric protecting films.
The ions impact the dielectric film, thereby generating a charge equal to the
potential of the falling ions on the surface. As the resulting charge is of the same
potential as the falling ions, it bends the ion path, repelling the ions from the
field-emitter surface.
The main advantage of the device is the extended life cycle of the cathode.
8.3.3 Traveling-Wave Tubes
Traveling-wave tubes provide a long interaction between an electron flow and a

delayed electromagnetic wave. Figure 8.70 diagrams such a device based on a field
emission cathode [169].
The velocities are modulated in the incoming line, which makes it possible for
additional electron beams to get into the drift space, where the amplification grows.
The triode design based on pointed cathodes (Fig. 8.70a) can be improved by
focusing electron beams, which densifies the electron beams emitted from the field
emission cathodes. The electron source is diagrammed in Fig. 8.71 [170]. Field
emission cathodes 3 of the electron source are ultra-thin diamond powder of 5–
1000 nm particles having good adhesion to the material of the cathodic electrode 2.
The insulating layers consecutively sputtered by the conventional photolithographic
method 4 and conductive electrodes 5 create a control-and-focusing system for the
electrons emitted from field emission cathodes 3. A beam can achieve the current
density of more than 10 mA/cm when the electric field intensity is 10 V/lm.
Such a multigrid structure gives high-quality electron beams, which is especially
useful for traveling-wave tubes.

diagram of a travelling-wave
lamp with a field emission
cathode: 1 field emission
cathode distributed along the
lamp; 2 input microstrip line;
3 drift space; 4 output
microstrip line; 5 anode.
b Distribution of electric field
along the lamp at the input
line 1 and the output line 2
Fig. 8.71 Schematic diagram of the electronic part of a travelling-wave lamp with electron beam
focusing: 1 substrate; 2 cathodic electrode; 3 field emission cathode from made of diamond
powder; 4 insulating layers; 5 control electrodes; 6 electron beam; 7 anode
Fig. 8.72 Structure of a

reflective klystron with a field
emission cathode: 1 source of
field emission electrons; 2
resonator; 3 reflective
resonator; 4 grids
8.3.4 Reflex Klystrons
The principle circuit of the field emission cathode-based reflex klystron is dia-
grammed in Fig. 8.72 [171].
The operating principle of this device is based on modulation of the velocity of
the electrons emitted by the electron source 1. After emission, the electrons pass a
pair of metal grids forming resonator 2. Having left the second grid, the electrons
enter the drift region, where they are reflected back along the initial paths. That is
what negatively charged repulsive electrodes do.
Electron-current fluctuations amplify the oscillations of the electromagnetic field
in the resonator, which results in fluctuations in the potential difference between
grids 4. These changes in the potential slow the fast electrons and accelerate the
slow ones when they pass through the resonator, causing formation of electron
beams in the drift zone. If the reflected electron beams come back to the resonator at
the appropriate potential, i.e. when the grid potential is positive, the power is
transmitted to the resonator and the beaming effect goes up. There is feedback under
the circumstances, and the resonance frequency of the resonator causes the variable
power to increase until the state becomes steady, viz. when increase in the gener-
ation power becomes equal to the loss power. Naturally, it is necessary for the
power transmitted from the electron beam to the resonator to be enough to com-
pensate for the losses in the resonator and to transmit the useful power to the
external load. According to such calculations, the devices can used up to
sub-millimeter wavelengths.
8.4 X-ray Tubes
X-ray tubes, requiring high voltage to function, are convenient objects to use field
emission cathodes. Such attempts have been made from long ago [172], the field
emission cathode being the segments of a differently-configured graphite cloth.
Application of pointed molybdenum field emission cathodes in an
ultrahigh-vacuum structure (1010 mm Hg) of a reflective anode X-ray tube shows
the prospects of such field emission cathodes [173]. The cathodic part is a
50,000-apex field emission cathode array of 1 mm in diameter [174].
The voltage being 25 kV and the beam current being 25 mA, the X-ray spot
diameter is 0.3 mm. More than 90% emission intensity is accounted for by the
diameter of 0.2 mm, electrostatic lenses being geometrically adjustable to reduce
the spot diameter to less than 0.1 mm.
The X-ray tube resolution depends on the X-ray source dimensions. That is why
it is very important to design microfocus X-ray tubes. Modern materials for field
cathodes are based on various carbon nanostructures [175, 176]. Nowadays the
most widespread of such materials for X-ray tube field cathodes are carbon nan-
otubes. Figure 8.73 diagrams an X-ray tube based on a field emission cathode made
of a multiple-wall nanotube bundle [177]. The multiple-wall nanotube bundle 2 is
stuck down to the end face of the tungsten wire 1 with a conductor paste. The
potentials at cathode Vc and copper anode 4 Va are −6.5 kV and +6.5 kV,
respectively. So the accelerating voltage is 13 kV. Electrical lens 3 of three parallel
diaphragms is put between the cathode and the anode.
The first diaphragm of the lens is also an extraction electrode. Since the distance
from the end of the field emission cathode to the second electrode is 13 mm and to
the anode 26 mm, increase in the electron-optical system is equal to one. So the
focal dimension of the electron beam at the anode is equal to the dimension of the
nanotube based field emission cathode, its diameter being a few tens of nanometers.
The X-rays of the focused spot are withdrawn through beryl window 5 to the
atmosphere. Then the emission passes sample 6 and records itself on film 7, the
field emission current being 440 lA.
8.4 X-ray Tubes 495
Fig. 8.73 Schematic diagram of a microfocal X-ray tube with a field emission cathode made of a
bundle of multiwall nanotubes: 1 tungsten wire; 2 field emission cathode made of multiwall carbon
nanotubes; 3 electrostatic lens; 4 copper anode; 5 window made of Be; 6 sample; 7 film. Power
supplies of: Vc field emission cathode; Vy extraction electrode; Va anode; Vl control lens

of an electron source with a
field emission cathode made
of carbon nanotubes: 1
substrate; 2 cathode; 3 spacer;
4 control grid; 5 focusing
electrode
To increase the power of the X-ray emitter, one has to increase the emission
current mostly by increase in the area of the field emission cathode. In this case, it is
necessary to use the electrostatic lenses of a more sophisticated design [178].
An X-ray bundle scanning usually involves an anode- or cathode-positioning
mechanism. That considerably complicates the design and increases its dimensions.
To remove these disadvantages, paper [179] suggests using several cathode
assemblies and the respective number of electrical lenses. Paper [179] describes 5
cathode assemblies arranged in a line. The field emission cathodes are controlled by
MOS transistors.
The simple structure of the X-ray tube field emission cathode assembly based on
a carbon nanotube field emission cathode is shown in Fig. 8.74 [180].
Carbon nanotubes 2 are applied on 1-mm thick stainless-steel substrate 1. The
layer of nanotubes is 70 lm thick. Spacer 3 separating the field emission cathode
and control electrode 4 is Micatex of 56 µm in thickness. The functions of a control
electrode are performed by a molybdenum microgrid of 100 lines/in. in trans-
parency and 25 µm in thickness with an orthogonal structure. Focusing electrode 5
is a steel plate of 2 mm in thickness with an aperture of 3 mm in diameter.
The threshold electric field intensity for this cathode is 3 V/µm, and the
emission-current density is about 10 µA/cm2, the emitter area being 1 mm. In this
electron source, the factor of electron-beam transmission through the focusing
electrode is 0.5–0.8, depending on the spacer thickness and the anode voltage. For
the X-ray tube of this design to function, it is enough to have the total power of
about 0.1–0.4 W.
The development of field cathodes based on carbon nanotubes (CNT) led to
multiple inventions in construction and manufacture of miniature X-ray tubes with
CNT-based field cathodes, e.g. [181–186].
Nowadays we can see intensive development in the field of tiny through-target
X-ray tubes. A version of such a tube with a field emission cathode made of a
carbon-fiber bunch is suggested in [187–189] (Fig. 8.75).
The X-ray tube (Fig. 8.75) consists of vacuum envelope 1 containing electron
gun 2, anode 3 shaped as a chamber in this case, and X-ray ejection window 4. As
for electron gun 2, it contains field emission cathode 5, cap 6 with electrical lead 7,
bead 8, field emission cathode connector assembly 9.
Appearance and scheme of the cathode-modulator node of a miniature X-ray
tube is presented on Fig. 8.76.
Carbon-fiber bunch 10 is put into shell 11 to maintain orientation of carbon
fibers, their mechanical mounting and vibration strength. Shell 11 is made of a
conducting or semiconducting material, e.g. a conductor, semiconducting glass or a
metal-coated dielectric, i.e. a material excluding accidental potential shift of the
shell against the carbon-fiber bunch when the X-ray tube is in operation.
Fig. 8.75 Tiny

through-target X-ray tube
with a field emission cathode
made of carbon fibers: 1 glass
bulb; 2 field emission gun; 3
anode; 4 window; 5 field
emission cathode; 6
modulator; 7, 15 electric
leads; 8 dielectric bead; 9
contact unit of the field
emission cathode; 10 bundle
of carbon fibers; 11 fiber
coating; 12 opening in the
modulator; 13
electroconductive paste; 14
metal shell; 16 protective
cavity
8.4 X-ray Tubes 497
Fig. 8.76 Appearance (a) and schematic diagram of the cathodic modulator unit of a miniature
X-ray tube: 1 modulator; 2 cathode lead; 3 slug; 4 lamp stem; 5 carbon fibers; 6 glass insulator
Carbon-fiber bunch 10 projects above shell 11 on the emitter side. For field elec-
trons to transit from the cathode to the anode, the end wall of cap 6 functioning as a
control electrode of the tube has a diaphragm represented by orifice 12.
Contact field emission cathode assembly 9 is made of conducting material 13;
e.g., a conducting paste based on the powder of Ag, Al or other metals or con-
ducting materials. Above-mentioned material 13 is applied on the carbon-fiber
bundle end opposite the emitter and a part of the shell pressed out by metal ring 14,
which can be shaped as a cylinder or a cylinder cut along the plane passing through
its axis. Ring 14 has electrical lead 15 (cathode) from vacuum envelope 1 and is
rigidly connected with dielectric bead 8 with, e.g., glass-fiber reinforced cement,
glue or amber, the central axes of ring 14, carbon-fiber bundle 10 wrapped into shell
11 and the aperture of bead 8 being coaxial with the electron-optical system of the
X-ray tube.
The tube is assembled so that the internal walls of the aperture in bead 8, shell 11
and the upper end face of contact assembly 9 form cavity 16. The cavity is nec-
essary to prevent generation of surface conduction between cathode 5 and control
electrode 6 caused by formation semiconductive and conductive films and paths on
the shells of cathode 11, bead 8 and control electrode 6. The conductive and
semiconductive films emerge from thermovacuum treatment of the X-ray tube, as
well as evaporation of the cathode material exposed to ion bombardment and
breakdowns between the X-ray tube electrodes when aging and operating the X-ray
tube.
To prevent development of these films, cavity 16 can be supplemented with long
or traverse electrically-protective slots in bead 8 to increase the surface electrical
transparency of bead 8 and prevent the generation of surface electrical conductance

between field emission cathode 5 and control electrode 6. The electrical slots
increase the space between cathode 5 and control electrode 6 and prevent formation
of parasitic conductive films.
The X-ray tube works as follows. Supply of anode 3 with high positive potential
with respect to cathode 5 and voltage between field emission cathode 5 and control
electrode 6 from end faces of the carbon fibers projecting above shell 11 results in
field emission. Numerous fibrils, i.e. tetragonal crystalline carbon structures of 250–
1000 Å in length and 20–50 Å in diameter drawn along the fiber axis, project above
the carbon-fiber end-face surface. They are carbon-fiber field emission centers.
Field electrons from cathode 5 go out through orifice 12 made in the end wall of cap
6 and get into an intense accelerating electric field between anode 3 and cap 6. The
stop-down orifice 12 in cap 6 enables field electrons to transit only to anode 3,
preventing cathode field electrons from hitting the dielectric part of shell 1 of the
X-ray tube. Positioning of the cathode assembly consisting of field emission
cathode 5, dielectric bead 8 and field emission cathode connector assembly 9 in the
cavity of cap 6 prevents the electrons from transit from field emission cathode 5 to
the interior face of the dielectric vacuum envelope.
Bombardment of the anode with high-energy electrons generates X-rays whose
beams project above the tube through window 4. Shift in the potential of cap 6
functioning as the control electrode changes the electric field intensity value at the
end face of the emitting carbon fibers of field emission cathode 5, thus adjusting the
X-ray tube current and X-ray intensity.
In the course of operation of the X-ray tube, ion bombardment and evaporation
of carbon from the fibers of the field emission cathode and the shell material can
cause formation of conductive and semiconductive films between the electrodes, on
the control electrode, the shell, as well as on the surface of the dielectric pinning the
electrodes together. The film formation increases the parasitic electrical conduction
between the X-ray tube electrodes, changes their characteristics, lowers electrical
strength, detracts from reliability and reduces life cycle.
The design suggests elimination of these disadvantages in the electron gun by
use of bead 8 vertically or horizontally slotted in the upper part. In addition, field
emission cathode 5 is put in bead 8 so that the central aperture in bead 8 and shell
11, where carbon-fiber bundle 10 is inserted, form cavity 16. It may specifically be
ring-shaped. Moreover, introduction of the bead provides centering and adjustment
of field emission cathode 5 in the optical system, which improves repeatability of
emissive and radiative characteristics of the X-ray tube. Cavity 16 between shell 11
and bead 8, as well as the slots in bead 8, prevent from formation of continuous
conductive and semiconductive films between the cathode and control electrode 6
over the surface of bead 8. The reason is that the operation of X-ray tube is
accompanied by appearance of closed-loop blind spots round the cathode, where
the products of cathode sputtering cannot get in the course of tube operation. As a
result, the conductive film on the surface of bead 8 has closed-loop non-sputtered
spots (ruptures) in cavity 16, which prevent parasitic electrical conductivity
between cathode 5 and control electrode 6.
8.4 X-ray Tubes 499
The tests of the tiny X-ray tube claimed confirm the effectiveness of its structural
solutions. At the supply voltage of 40 kV and the current of 100 lA, the tube has
worked for more than 1000 h without failures, which proves its high reliability and
electrical strength. At the same time, the instability of characteristics of the X-ray
tube has been less than 1%.
A promising approach to manufacture with a shoot-through anode is proposed in
[190]. In this construction of X-ray tube with field cathode the external (incl.
magnetic) electrodes are used for control and focusing of the beam. This allows to
make the technology of manufacturing of X-ray tubes simpler and cheaper.
A scheme of the tube is presented on Fig. 8.77. Its function principle is obvious
from the schematic.

of a field emission X-ray tube
with external control
electrodes: 1 field emission
cathode unit; 2 field emission
cathode; 3 control electrode; 4
focusing electrode; 5 body; 6
output window
A more detailed review of X-ray tubes with low-strength field emission cathodes
is presented in [191].
8.5 Electron Guns
8.5.1 General Principles
Electronic devices mostly employ a narrow directed electron bunch, i.e. an electron
beam. Most devices employ a round-section beam, but sometimes shaped beams,
i.e. ribbon beams, oval-section beams and otherwise shaped beams are applied.
One of the basic units of an electron-beam tube is the device generating an
electron beam, i.e. a focused electron bunch with energy and current density nec-
essary for the tube. Such a device is referred to as an electron gun.
The requirements made to an electron gun can vary over a wide range depending
on the type of the device. E.g. beam electron energy can amount to several kilo-
electronvolt in small oscillograph tubes and some tens or even hundreds of kilo-
electronvolts in electronic microscopes.
In the same way, the beam current of some image pick-up tubes is several deciles
of a microampere, while it can reach several milliamperes in lithographs. However
wide the range of variation of energy and current might be, the beam section in a
particular plane, usually in the receiver plane, i.e. a screen or a target, should be as
small as possible. The beam section is usually estimated by the spot diameter, a spot
being understood as a trace of an electron beam on the receiver surface. In many
types of contemporary devices, the spot diameter is as short as several deciles of a
millimeter or less.
Most devices imply an ability to control the beam current value in a wide range,
from zero, i.e. gun blocking, to the maximum value depending on the purpose of
the device.
The designers of electron-beam devices should also take constructive reasons
into consideration the electron gun should be well agreed with the entire device; the
technology of its fabrication should not be too difficult; the materials of the electron
gun designed should meet the conditions of operation in vacuum. Moreover, the
gun should be quite economic in operation and long-lasting, as its failure can
destroy a sophisticated and usually expensive device.
The electrons emitted by the cathode are accelerated by the electric field gen-
erated in the cathode region by gun electrodes. As the gun should generate not only
accelerated, but also focused beam-forming electron flow, it is expedient to select
the configuration of the accelerating field so that acceleration went along with
electron flow focusing. For this purpose, the cathode is equipped with an acceler-
ating electronic lens. As the accelerating field should reach the cathode surface, the
cathode should be dipped into the lens field.
It has already been mentioned that most guns can control the beam current.
8.5 Electron Guns 501
Similarly to electron tubes, it is expedient to control the current in the cathode

region by changing the electric field at the cathode surface. The field in the cathode
region can be simply changed by change in the potential of the diaphragm located
between the cathode and the first anode or the accelerating electrode of the electron
gun. Such a diaphragm, i.e. a modulator, is similar to the control grid of the electron
tube.
Generally, the cathode-modulator-accelerating electrode/the first anode system
of electrodes can be regarded as similar to the system of electrodes in a conven-
tional three-electrode tube. In this case, the modulation characteristic can be con-
sidered as anode-grid characteristic of the triode, and the out-off voltage as that
similar to the blanking grid voltage of the triode.
However, such analysis of electron guns is rather approximate. The main dif-
ference of the first lens of the triode gun is nonuniform distribution of the current
over the gun cathode. A change in the modulator voltage entails not only a change
in the functioning/control voltage in the modulator plane, but also the dimension of
the emitting cathode surface.
Electron guns based on field emission cathodes have a number of both positive
and negative features.
The positive ones include a possibility to achieve considerable current densities,
speediness, a wide temperature range etc.
The basic disadvantages of such guns are a high initial velocity of the emitted
electrons and a wide angle of their spreading. That seriously complicates focusing
of cold electrons, and the nanodimensions of the emitting centers and their statis-
tical arrangement make exact computation of electron optics impossible.
That is why field emission guns are designed by the methods of approximate
computations and experimental simulation.
8.5.2 Field Emission Localization Methods
There are a lot of applications needing localization of the electron beam in a small
solid angle; e.g., raster-type electron microscopes, X-ray microanalyzers, some
types of electron-beam tubes, microwave devices etc.
As a spike field emission cathode is spherically symmetrical, the emission cone
opening angle is very large, viz. about 60° for emission at an average current
density of 104–105 A/cm2 and about 100° for high densities (107–108 A/cm2).
Nowadays, there are several beam emission localization methods. The angular
limit can be achieved by cutting the most intensively emitting region of the apex
surface with a diaphragm and increase in its emission abilities through restructuring
[192]. The restructuring is carried out by expansion of certain crystallographic
planes by application of back voltage.
A lot of papers have been dedicated to research of adsorption of different
stranger atoms on the surface of tungsten or other materials, which are selectively
adsorbed on certain crystallographic planes, thus reducing the work function in the
adsorption region. As a result, the cathode surface develops an electron-liberation

opening, its dimensions being determined by those of the adsorption region. System
[W, Zr], which makes it possible to reduce the divergence angle to 0.5 steradian, is
the best studied in this respect [193, 194].
The disadvantages of such a design are the need for an additional external source
of the sputtered element and a short life cycle, since ion bombardment of the emitter
surface results in sputtering of a film of the active element.
A local reduction of the work function can be achieved by oxidation, thoriating,
impurity doping or apex zirconization [195–197]. Some of thus-produced cathodes
have already been used in industrial devices and emitted electrons from the apex at
12°.
The divergence angle of the electron beam can be reduced with apices made of a
refractory material doped with an element forming a low work function film on the
surface.
In this case, localization can be done without any additional external source of
sputtering. However, the cathodes of this design have an insufficient life cycle and
electron luminosity.
The emission current is localized by a combination of thermal activation of the
emitting apex surface by an emission-active addition agent dissolved in its body and
delivered to its surface due to diffusion, on the one hand, and thermo-field
restructurization, when the most densely packed facets are profiled in relief, on the
other hand. As a result, the apex surface develops microelevations consisting of an
aggregation of atoms of the emission-active addition agent concentrating mostly in
the places of the highest electric field gradient. The microelevations experience
local electric field enhancement, as well as selective decrease in the work function
due to a hemisorptive interaction of the adsorbate electrically positive to the sub-
strate. The total effect of change of the electric field intensity and electronic work
function for the central facet makes it possible to implement a higher degree of
emission-current localization.
The selective change of the work function of the central apex facet is mostly
affected by the dipole moment of the double layer, depending in its turn on
adsorbate-atom radius ra and the package density of the substrate facet considered
dhkl. The larger ra and the larger dhkl, the larger is P. Among the electropositive
elements, the largest r belongs to the alkali metals of Li, Na, K, Cs. Densely-packed
facets W, tending to emit as a result of thermo-field restructurization, include the
facets of (110), (100), (112). Accordingly, a tungsten-alkali metal system with
orientation to one of the facets specified is developed.
The main difficulty of alloying a refractory metal and an alkali metal is a large
difference in their temperatures of melting and vaporization. However, when
monocrystal W is obtained by the method of electronic zone melting, these metals
can be dissolved in a tungsten matrix with the total concentration of about 0.1%.
The lower element-content boundary, about 0.05 wt%, stipulates the minimum
life cycle of 100 h. The lower content of the elements makes it less than 100 h,
which does not meet the requirements to the field emitter in a microscope.
The alloy can result from zone melting by an electron beam in two passes at the
velocity of 2–4 mm/min. along crystal axis 001. The samples of 0.3 0.3 mm in
cross-section are cut out along this axis by the electric-spark method.
The apex is etched in the solution of 1.5 N NaOH at the voltage of 5 V and the
current of 20 mA.
The field emitter is mounted in a field emission microscope and exhausted down
to the pressure of 10−7 Pa, which is the working vacuum in electron-probe devices.
The field emitter is activated by preheating at 1460 K for 3 min. in the field of
0.2 V/Å, which makes the atoms migrate from the matrix to the apex surface and
form a stable aggregation on facet (001).
An apex with a small solid angle of an electron beam for heater field emission
cathodes is usually produced by thermal field processing [198]. In this case, the
basis is a tungsten field emission cathode with the rounding-off radius of about
2 µm and the axis oriented towards [98]. At first, the field emission cathode is
cleaned by heating up to 2300 K. Then the temperature is reduced to 1900 K and
high voltage is supplied and stepped up until the emission current reaches 1–
100 lA, depending on the emitter rounding-off radius. At the first moment, the
cathode emits perpendicularly to the facet [310] as shown in Fig. 5.1b. Growth of
the voltage triggers the restructuring of the emitter (Fig. 5.1c), which is subse-
quently reshaped so that the entire emission takes place only from the facet [98]
(Fig. 5.1d). The emitter restructurization is accompanied by stepping up of the
emission current. At this stage of processing of the cathode, emission is stabilized at
constant values of temperature and voltage. The field emission cathodes produced
in this way are capable of a long service life at T 1200–1700 K with current
takeoff of 1–300 µA at the residual-gas pressure p of 10−6–10−3 Pa. Pressure
reduction by an order of magnitude down to p 10−7 Pa makes it possible to step
the temperature of the field emission cathode down by 570 K. A face of W(310) is
still considered a classic field emission cathode for high-dimension microscopy, as
evidenced by the latest studies in this field [199].
The limiting current value that can be taken off from the field emission cathode is
limited to heating of the emitter with Joule heat and the Nottingham effect and
ultimately determined by the heat-conducting path conditions [200], the apex
geometry [201] and the operating mode [202]. In practice, the volt-ampere char-
acteristic biases towards current reduction, which limits the emission-localization
possibilities. Paper [203] suggests an explanation of this phenomenon based on the
difference between the true shape of the potential barrier and the classical
image-force law as adopted to derive the barrier-transparency expression.
8.5.3 Low-Power Electron Guns
Low-power electron guns are very widely used in electron probe devices, such as
focused-beam microscopes and transmission-electron microscopes [204, 205].
To present day there has appeared a large number of focused-beam and trans-
mission microscope constructions that employ field emission cathodes for
increasing the image resolution. Some of those designs of previous years are still of
interest at this time.
The first electron gun based on a field emission cathode was developed by
Crewe [206] and calculated by Butler [207] (Fig. 8.78).
Voltage U1 applied between the apex and the first anode determines the total
emission current. Voltage U0 stipulates the energy of electrons. The electrical lens
developing between the anodes fixes the electron beam and forms a real image of
the field emitter in the plane of the sample. This structure was further developed in
[208].
The camera–trap inlet is 1–2 mm away from the field emitter point. It is supplied
with the voltage of 5–8 kV against the emitter. As a result, the gun works with
development of a false image of a fictional electron source. Paper [209] suggests
converting the fictional-source false image generated by the electron gun into a real
one, using an additional magnetic condensing lens for this purpose.
The paper [210] features the lens system of a field emission gun consisting of
three anodes and a magnetic lens. The system is more efficient in terms of current
than the accelerating-lens system of two anodes. Paper [211] features the optical
parameters of such a system in detail.
Magnetic-lens electron guns are featured in [212, 213]. The electron gun consists
of an apex, a reference electrode and an anode put in a magnetic lens, the focal
distance being a few millimeters. Such a system forms a real image beyond the pole

of a field emission electron
gun [178]: 1 tip; 2 ion trap; 3
aperture diaphragm; 4 Butler
accelerating gap; 5 sample

of a field emission cathode
gun with a magnetic lens: 1
anode; 3 magnetic lens
shoes. The reference electrode can control the field at the apex, the emission
conditions being maintained constant in the range of anode voltages from 2 to
10 kV. The gun structure shown in Fig. 8.79, is a diode system where the anode is
inside the magnetic-lens pole shoes. The lens is optimized in respect of the cathode
and forms a real image [213].
Paper [214] shows that the emission rate of a field emission gun is 3–4 orders of
magnitude higher than that of a thermal-emission gun with other things being equal.
Great difficulties in designing field emission cathodes able to work in high
technical vacuum have given rise to preheated field emission cathodes that can
provide thermal dependence or clean the surface of a field emission cathode.
Paper [215] suggests cathode assemblies of electron gun-based transmission
microscopes [216].
Field emission cathode-based electron guns working in ultrahigh vacuum of
10−8–10−7 Pa have a problem of supplementary degassing of accelerating-electrode
surfaces. The reason is that emission from a field emission cathode brings some
emitted electrons to the surface of these electrodes and desorbs the residual-gas
molecules on their surface, thereby increasing the pressure near the apex, which
degrades the electron-beam parameters and reduces the cathode life cycle. It is
suggested to eliminate this disadvantage with an electron gun able to degas the
accelerating electrons [217] (Fig. 8.80).
Degassing is performed by auxiliary ring-shaped thermionic emitter 5 sym-
metrical in respect of field emission cathode 3. First anode 4 is also symmetrical to
it. For second anode 6 to be degassed, the first anode is circumferentially punched
to form apertures 9.
Fig. 8.80 Design of an

electron gun with an electrode
degassing device: 1 vacuum
compartment; 2 insulator with
input leads; 3 cold field
emission cathode; 4 first
anode; 5 auxiliary ring-shaped
thermionic emitter; 6 second
anode; 7, 8 openings in
anodes for the main electron
beam; 9 openings in the first
anode for the auxiliary
electron beam
During the process of degassing, the current passing through auxiliary cathode 5
causes emission of electrons, the accelerating potential between cathode 5 and
anode 4 making the electrons bombard first anode 4. Some emitted electrons pass
through apertures 9 and bombard second anode 6 supplied with a higher positive
potential than the first anode.
To prevent destruction of apex 3, the latter is supplied with a potential equal to
that of auxiliary cathode 5. The symmetrical shape of the electrodes minimizes
perturbation to the emitted electrons.
If the anode surface bombarded by the cathode-emitted electrons is fabricated or
coated with a getter material, e.g., titanium, zirconium, thorium or barium; the
electron beam becomes even more stable and the current fluctuation diminishes
[218]. Before the electron gun starts functioning, the vacuum chamber housing it is
pumped out to the ultrahigh vacuum of 10−8–10−7 Pa. After that, the anode is
heated up to the temperature of 1070 K along with continuous pumping out,
evaporating the residual-gas molecules adsorbed on the anode surface, degassing
the anode surface and activating the getter. After the degassing, very few molecules
are desorbed from the anode surface due to electron bombardment from the field
emission cathode because these molecules are connected by the getter. Moreover,
the residual-gas molecules are forcefully absorbed by the activated getter. As a
result, the cathode, protected from ion bombardment, produces a stable electron
beam.
A two-chamber electron microscope [219] makes it possible to maintain different
pressures in the cathodic and sample compartments and replace the parts of one of
them without breaking vacuum in the other. The two-chamber electron microscope
is diagrammed in Fig. 8.81. Field emission cathode 1 is in compartment 3 of tubular
case 2. High-vacuum pump 4 mounted in the cathodic compartment maintains high
Fig. 8.81 Design of a

two-compartment electron
microscope: 1 cold field
emission cathode; 2 body; 3
cathode compartment; 4
high-vacuum pump; 5 high
voltage unit; 6 heating voltage
unit; 7 focusing and deflecting
coils; 8 aperture diaphragm; 9
wall between compartments;
10 sample stage; 11 vacuum
pump
vacuum of about 10−6–10−4 Pa inside the cathodic compartment. High-voltage unit

5 and filament-voltage unit 6 provide the working capacity of field emission cathode
1. Second compartment 10 contains a beam deflector and a sample stage missing
from the picture. Vacuum pump 11 maintains the second compartment at the lower
vacuum of 10−4–10−3 Pa. Cathodic compartment 3 and sample compartment 10 are
separated with partition 9 equipped with aperture 8 oriented coaxially with emitter
1. Aperture 8 is sized so that high-vacuum pump 4 can maintain a predetermined
pressure in the cathodic compartment despite the leakage through aperture 8.
The necessary pressure difference is usually maintained with the aperture
diameter of 0.5 mm.
The cathodic compartment can be made separately and used in any other device,
e.g., an electron-beam tube. A separate cathodic compartment with a closed
pumped limiter can have an ion pump to maintain the necessary vacuum.
Paper [220] suggests further improvement of the electron gun based on a field
emission cathode. In a conventional two-lens electron gun, some electrons accel-
erated by the first anode pass through the apertures in the anode and the second
anode accelerates them. However, a lot of electrons scatter back from the
first-anode surface. When the potential of the second anode is higher than that of the
first one, a lot of secondary and reflected electrons come to the accelerating field
area. That is caused by joint focusing and acceleration of both anodes. The sec-
ondary beam defies focusing. Moreover, the energy of the beam electrons is dif-
ferent from that of the main-beam electrons, the number of undesirable secondary
electrons being able to reach that of the primary electrons. To be applied, the
primary beam needs high-energy homogeneity of the electrons it comprises. When
scanning, the secondary beam makes spatial noise. The reason is that not only the
desirable, but also the neighboring area is scanned. In addition, a wide energy
spread of secondary electrons degrades the resolving ability of the microscope.
Bombardment of the anode-surface by the emitted electrons not used in the main
beam has been calculated to cause formation of secondary electrons with the energy
ranging from 5 to 50 eV, most of them having the energy of 7–15 eV. After the
passage, the electrons have the energy of e (U3–U1). As the potential difference in
the electron microscope is high, the object should be considered as lit up with the
main beam and the diffused electron beam, which certainly complicates the
research. The secondary-electron suppressor gun is diagrammed in Fig. 8.82.
Emission from field emission cathode 1 is provided by high-voltage source U1, its
voltage being supplied between cathode 1 and anode 2. Emission of electrons
generates two electron beams, viz. main beam 4 and secondary-electron beam 3.
However, the secondary electrons generated on the surface of anode 2 do not reach
second electrode 5 because the voltage negative in respect of anode 2 is supplied
between anode 2 and electrode 5. The voltage generates a field pushing secondary
electrons 3 off from electrode 5, so they do not get to the accelerating field region
between electrodes 5 and 6. Voltage U2 is about 20 V, so most secondary electrons
of 7–15 eV are retained. It is worth mentioning that voltage U2 should be selected
so that it did not essentially affect the primary-beam electrons. Accordingly, it is
necessary for the voltage to be less than the accelerating voltage of U3 by two
orders of magnitude. There are a lot of devices intended to stabilize the current of
field emission guns, e.g., [221].
Paper [222] suggests a design of a field emission stabilization gun.
Figure 8.83a shows the dynamics of the electron-beam current for a conven-
tional electron gun. In this gun, the pointed field emission cathode is firstly cleaned
and somehow smoothed by a flashout and then supplied with high voltage in
vacuum of about 10–7 Pa without heating up the cathode. The beam current
plummets in the first period, stabilities in period B, but grows in period C. The
reason is that infection of the apex surface with adsorbed molecules smooths the
Fig. 8.82 Electron gun with

suppression of secondary
electrons: 1 field emission
cathode; 2 anode; 3 secondary
electrons; 4 primary beam; 5
electrode which decelerates
secondary electrons; 6
accelerating electrode; 7
sample. U1 A high-voltage
source for field emission
initiation; U2 a source of
retarding potenial; U3 a
source of accelerating voltage
Fig. 8.83 Oscillograms of current in an electron gun over time: a change of current for an
ordinary electron gun; b change of current for an electron gun with stabilization; c change of
cathode heating current
surface. After a stabilization period, the current starts increasing because of

microelevations formed as a result of ion bombardment due to desorption of
residual-gas molecules. The stabilization period is about 5–10 min. for the vacuum
of 2 10−7 Pa.
The stabilized current of an electron beam at the vacuum of 2 10−7 Pa is
exemplified in Fig. 8.83b. Initially, the beam current changes as usual, but the point
experiences a filament-current pulse lasting as long as 0.1 s. at the temperature of
about 770 K, the field emission cathode not being deenergized. Then the beam
current remains very stable over the period lasting longer than B.
Figure 8.83c shows an oscillogram of filament-current pulses of the field
emission cathode, s1 being the pulse period and s2 being the heating-pulse duration.
Pulse period s1 is determined by the vacuum degree in the device and service
conditions, but it is usually about 10 s. Pulse width s2 depends on the cathode
heating-temperature, but it is usually less than 1 s.
The temperature varies from 420 to 1270 K. The temperature less than 420 K
cannot clean the cathode surface well enough. On the other hand, heating above
1270 K causes a slump of the beam current.
The implementation of the above-specified method is exemplified by an electron
microscope shown in Fig. 8.84, where beam-return time t3 coincides with
apex-cleaning time U2 (Fig. 8.83c).
A number of circuitries for electron gun current stabilization are suggested in
[223, 224].
Figure 8.85 shows a block diagram for field emission current stabilization.
When passing the aperture of first anode 2, the electron beam is focused by the
electrostatic lens generated by the field between the anode and the accelerating
Fig. 8.84 Example of

application of an electron gun
with current stabilization: 1
cathode heating source; 2
electron gun; 3 sweep
generator; 4 electron beam; 5
focusing lenses; 6 deflector
plates; 7 sample; 8
oscillographic tube; 9 detector
of secondary electrons; 10
oscillogram for the sweep
generator

of an electron gun with
stabilization of field emission
current: 1 cold field emission
cathode; 2 anode; 3
accelerating electrode; 4 arch;
5 control electrode; 6 sample;
7 comparator; 8 resistor; 9
base voltage source; 10
differential amplifier; 11
accelerating voltage source;
12 control unit; 13 control
electrode voltage source; 14
high voltage source; 15
cathode heating unit
electrode. The focused beam gets out of the aperture of the accelerating electrode.
The focusing field depends on the U2/U1 ratio, the voltages applied between the
accelerating electrode and the anode and the cathode, respectively. The surface state
of cathode 1 varies over time. Unless something compensates for these changes, the
emission current is unstable. On the other hand, the field emission current density
depends both on field intensity and the surface state, which makes it possible to
maintain constant emission by variation of the electric field. However, it has already
been stated above that no variation of voltage U1 is allowed to prevent changing the
U2/U1 ratio and consequent variation of the electron-beam focusing. Variation of
U1 makes the beam leaving the aperture of accelerating electrode 3 vary over time
for the above-stated reasons even if the emission is maintained stable.
To avoid that variation, the electron gun is equipped with reference electrode 5
and the appropriate potential source 13. Reference electrode 5 is put not far from
cathode 1, which makes it possible to change the electric field at the cathode by
variation of the voltage of source 13, leaving voltage U1 constant.
Stability is automatically controlled by comparator 7. It consists of resistor 8,
reference voltage source 9 and a difference amplifier.
When sample 6 is lit by the electron beam, the current passes through resistor 8,
causing a drop of potential. The voltage is compared with the reference voltage
from source 9. As a result, the output of difference amplifier 10 develops a dif-
ference signal that comes to control unit 12. Control unit 12 gives out an actuating
signal to source 13, which changes its output voltage until the difference signal
disappears.
A sudden failure of the vacuum in the electron gun rapidly destroys the field
emitter as a result of disruption between the anode and the cathode. Paper [225]
suggests a number of circuits able to protect a field emission vacuum gun against
disruption. The operating principle of all the circuits is based on variation of
resistances of the circuit components, which leads to shunting the discharge gap and
eliminating the discharge.
Paper [226] suggests a design of a field emission gun combining some means of
stabilization and protection of the field emission cathode. Figure 8.86 diagrams a
gun used in a raster-type electron microscope. Field emission gun 1 issues a bright
focused electron beam that lights up sample 2. The focused beam is scanned with

of a scanning electron
microscope: 1 field emission
electron gun; 2 sample; 3
deflecting system; 4 detectors;
5 sweep generator; 6 electron
tube; 7 accelerating electrode;
8 anode; 9 protective
electrode; 10 titanium
evaporator; 11 cold field
emission cathode; 12
additional electrode
deflecting system 3. The information on sample 2 is obtained by detection of the

passed, secondary, reflected etc. electrons by detectors 4. The latter issue a signal
that modulates brightness of synchronously scanning beam 5 of electron gun 6.
What matters for enhancement of the resolution of electron-beam analyzers such
as Auger spectrometers is reduction of the beam diameter without non-negligible
reduction of luminosity. It has already been mentioned that this property is inherent
in field emission electron guns. A field emission cathode introduced in an Auger
spectrometer [227] made it possible to reach the theoretical limit of resolution of
spatial coordinates.
The analyzer showed in Fig. 8.87 generates the primary electron beam by field
emission from spike field emission cathode 1 having the rounding-off radius of
about 5–10−5 lm. Field emission cathode 1 is connected to the filament–supply
source 3 through switch 2.
Chamber 4, which is vacuum-pumped to about 10−8 Pa, houses field emission
cathode 1. Insulator b insulates field emission cathode 1 from anode 5 with an
aperture of about 100 µm. Adjustable source 7 producing 0.5–3.0 kV supplies the
anode with voltage positive if respect of the cathode. High-voltage source 8 pro-
duces 3–10 kV. So the total accelerating voltage supplied to the anode ranges from
3.5 to 13 kV.
Fig. 8.87 The electron beam

analyzer with the field
emission cathode: 1
needle-shaped field emission
cathode; 2 switch; 3 initial
voltage source; 4 cathode
compartment; 5 anode; 6
insulator; 7, 8 controllable
high-voltage source; 9
retarding electrode; 10 bipolar
source of voltage; 11 rejecting
system; 12 target; 13 Auger
electrons; 14 secondary
electrons; 15 spectrometer; 16
first grid; 17 second grid; 18
retarding potential source; 19
accelerating voltage (200 V)
source; 21 transformer; 22
generator; 23 load resistance;
24 capacitor; 25 detector; 26
oscillograph
Source 10 supplies electrode 9 with voltage ranging from −200 to +200 V. As a

result, anode 5 and electrode 9 generate a decelerating electron lens, so the electrons
passing through the aperture in electrode 9 have kinetic energy ranging from 0.5 to
3 keV, which is just enough for optimum excitation of Auger electrons in the
atomic shells of the target substance.
The decelerated electron beam gets to deflecting system 11, which scans the
beam over the surface of target 12. The electron beam falls on the surface of target
12 and excites Auger electrons 13 and back-scattered secondary electrons 14. The
electrons enter spectrometer 15.
The spectrometer consists of first grid 16 maintained at the zero potential and
second grid 17 supplied with the retarding voltage coming from source 18 and
ranging from 0 to 1 kV. As a result, secondary and Auger electrons do not pass the
second grid because their energy is lower than respective voltage 18. The voltage
from source 18 is adjusted to the Auger-electron spectrum line that corresponds to
the element under investigation on the surface. The accelerating voltage from
source 19 makes the electrons pass grid 17 and get to collector 20.
The modulating voltage of about 2 kHz in frequency and 1–10 V in amplitude is
supplied from generator 22 through transformer 21 to target 12. The modulated
signal received at load resistor 23 passes capacitor 24 and gets to detector 25.
Having left detector 25, the signal goes to oscillograph 26, whose scanning is
synchronized with the electron-beam scanning. Respective variation of the voltage
from source 18 results in distribution of the elements over the target surface at the
oscillograph screen.
Paper [228] attempts to design an electron-beam tube producing the beam cur-
rent of up to 10 lA on the basis of a pointed carbon-fiber field emission cathode.
However, the carbon apex is rapidly destroyed even in the ultrahigh vacuum of
10−9 mm Hg, followed by involvement of the entire fiber end face into functioning.
So the effective area must be represented by the end face of the rod carbon-based
field emission cathode [229] especially as applied to electron guns producing the
beam current of 100 lA.
Most applications of electron guns need electron beam focusing. Quite good
results come from application of external solenoids, but high speed of field electrons
makes them too cumbersome. This disadvantage is eliminated by suggestion of an
internal one-band system of magnetic focusing [230], where the magnetic pole is in
immediate proximity from the field emission cathode, which makes it possible to
vary the path of field emission cathodes in the most efficient way.
It seems most interesting to design electron guns with electrostatic focusing and
control because they are most easy to produce and compact.
Statistically resistant space distribution of an electron flow causes difficulties in
parameter computation of the electron guns. That is why the electron guns with a
carbon-based field emission cathode are now designed with the aid of experimental
research.
Some proposals that appeared lately use carbon nanotube based field cathode for
increasing the brightness of scanning and translucent microscopes [231].
8.5.4 High-Power Electron Guns
As the current from one field emitter is very low, generation of high-ampere cur-
rents needs simultaneous functioning of many hundreds and thousands of field
emitters. The core of this problem has not been solved yet. However, the attempts to
meet the challenge are continuously under way.
Some other approaches to development of high-current field emission cathodes
is given in Chaps. 5, 6. In this section, we are going to consider just a few aspects
of this problem.
Classical high-current field emission cathodes can be exemplified with arrays of
pointed molybdenum field emission cathodes (look issue 5.2 and 5.8). Such arrays
employ more than 100,000 apices at the working pressure better than 10−9 mm Hg.
As already pointed out, field emission cathodes are manufactured from
carbon-based materials to reduce the working pressure.
In particular, paper [232] does so with an array of pointed field emission cath-
odes made of glassy carbon manufactured by the photolithography technology and
thermo-chemical etching. The emitters are 14 lm high. The vertex curve radius is
about 1–3 lm. The emitter density is 2.5 105 cm−2. If the cathode-plate diameter
is 3 mm, the emission current is 100 mA. In addition, there are field emission
cathodes of a large area, viz. more than 1 cm2, made of carbon-fiber bunches [233]
or high-tenacity graphite, e.g. fine-grained dense graphite MPG-6 [234] with a
high-transmission control grid positioned near the cathode plane. This electron-gun
structure makes it possible to obtain an electron beam of any configuration
depending on the shape and the size of the effective area of the field emission
cathode.
An example of such a structure is shown in Fig. 8.88.
Fig. 8.88 Electron gun with a graphite field emission cathode: a main circuit of the gun: 1 field
emission cathode made of MPG-6 graphite; 2 work surface of the cathode; 3 extraction electrode (a
tungsten wire gauze with a diameter of 0.01 mm and transparency of 80%); 4 electron beam
boundary; 5 anode. b ribbon-type field emission cathode; c matrix-type field emission cathode
The field-electron gun is field emission cathode 1 with grooved end surface 2. In
close proximity to the effective area of the field emission cathode, there is extraction
electrode 3 represented by a grid made of a high-melting material. Application of an
electric field to the extraction electrode from the microelevations on the effective
area of the field emission cathode causes field emission.
The electron beam, whose configuration depends on that of the cathode, passes
the extraction electrode and gets to anode 5, which collects electrons.
The end surface of the field emission cathode made of a nanostructured
carbon-based material consists of a great number of microelevations statistically
uniformly distributed over the effective area. In the course of emission under
bombardment of the surface with residual-gas ions, the emission centers are
damaged and at the same time new microelevations are exposed to become new
emission centers. After the effective area of the cathode has aged, this process
stabilizes in time, which provides high time stability of the emission current drawn
from such a cathode. Increase in the working area of the cathode by more than 4–
6 mm2 aimed to increase the cross-section of the emitted beam causes spatial
instability of the beam because of heterogeneity of the cathode surface of such size.
A large-section beam can be generated by a set of separate stable electron beams
emitted by the structural elements of the end surface of a small-size cathode
commensurable with the cathode-extraction electrode gap.
If the dimensions of the structural element of the effective area are larger than the
cathode-extraction electrode gap, each element can work as an independent cathode
issuing an almost parallel electron beam. Moreover, such geometry improves
homogeneity of emission from such an element due to more homogeneous electric
field. Almost the same situation is true when the ratio of dimensions is
1 2D. The ratio has been selected experimentally.
When 1 < 2D, the electron beams from the neighboring microelevations
superimpose on one another. Reduction of D is mostly determined by the me-
chanical strength of the grid and the technology of its tensioning, which determines
a possibility to mount the device, less D causing less extraction voltage. These
ratios are assumed for the identical dimensions of the grooves and the projected
parts. The ratio can vary towards increase in the projected part. However, the
increase can reduce the overall performance of each element edges because other
elements can shield them. To prevent the bottom of the grooves from contributing
into the cathode-field emission, i.e. for the electric field intensity at the bottom to be
much lower than it is necessary to excite field emission, the groove depth should
meet ratio 2. The ratio was selected by experiment. h/D = 2 is the minimum
giving a positive effect. Practically, this ratio was obtained as 10–15, which is
mostly determined by the mechanical strength of the material used.
The problem of the extraction electrode is that it should have a high
electron-beam transmission factor, on the one hand, and provide uniformity of the
electric field at the cathode effective area, on the other hand. The structures like thin
foils or orifice plates do not meet the transmission-factor criterion. To provide
uniformity of the electric field of the extraction electrode represented by a wire
mesh of a high-melting conductor, the size of the bridges should be more than 0.02–
0.1D with transparency of 70–90%. Such a grid provides a uniform potential at the
cathode and the maximum transmission of an electron beam. Increase in the size of
the bridge or the wire diameter at invariable D exceeding 0.1D increases
non-uniformity of distribution of the electric field, which results in non-uniform
emission from the surface, the minimum value of 0.02D being determined by the
mechanical strength of the grid material.
The field-emission cathode can be made of different materials, e.g., carbon
nanotubes, carbon foils, constructional graphites etc. [235, 236].
8.6 Other Types of Devices
8.6.1 Heaters
The characteristic features of field emission can substantially improve the opera-
tional parameters of electronic devices. That is why the above-listed guidelines of
development of field emission cathode-based devices do not exhaust the possibil-
ities to use field emission cathodes in a wide range of devices and appliances.
This section outlines some possible types of devices where field emission
cathodes improve the operational parameters.
Some constructional elements of electronic devices need heating when in
operation. They include thermionic cathodes, gas leaks, getter pumps, getter-ion
pumps etc. The operating principle of such heaters is that emitted electrons are
accelerated in the anode-cathode gap and decelerated in the anode material with
heating of the latter.
The optimum heater is a field emission cathode.
When working, the field emission cathode is at the ambient temperature, which
removes the power expenditure due to heat emission by the heated cathode. That
makes the efficiency of the pump nearly two times higher. Unlike hot cathodes,
decrease in the operating currents and absence of a special heater makes it possible
to reduce the number of current collectors to two and cut the power losses due to
heating and heat transfer along them.
A lot of sophisticated electronic devices, especially those equipped with special
glued-in windows, employ embedded tiny getter pumps or getter-ion pumps when
the capacity of conventional flash getters is not enough [237].
As a getter material, such small pumps use lanthanides, e.g., erbium, gadolinium,
etc. This type of hot cathode-based pumps [237, 238] does not provide a short
triggering time and have relatively high power consumption and dimensions.
Figure 8.89 shows the designs of small pumps with field emission cathodes of
fine-grain high-density graphite MPG-6 type based on lanthanides [239].
The operating principle of the pumps is as follows. Voltage is applied between
anode 1 and cathode 2 so that the electrons emitted from the elevations of the field
emission cathode bombard the lateral surface of tantalum cup 6 or thin-walled
tantalum cylinder 7. Dissipation of the field emission electron energy heats the anode
8.6 Other Types of Devices 517
Fig. 8.89 Design of getter pumps with field emission cathodes: a cylindrical cathode with an
internal work surface; b flat cathode; c cylindrical cathode with an external work surface. 1
lanthanide metal; 2 graphite field emission cathode; 3 insulator between the anode and cathode; 4,
5 field emission cathode and anode holders; 6 tantalum cup; 7 thin-walled tantalum cylinder
and sublimes its material. All the three pump designs are essentially identical and
differ only in the heater design and the spreading angle of the getter material. So let
us consider one design in more details and note the distinctive features of the others.
Pump (Fig. 8.89a) consists of an evaporation anode, including cylindrical tan-
talum cup 6 with a pressed-in lanthanide tablet 1. Cup 6 is welded to the bottom of
the structure through supporting arm 5. The functions of graphite field emission
cathode 2 are performed by shape-fitted elevations on the interior of the cathode
cylinder.
The field emission cathode is fixed in back-up ring 4 soldered to insulators
coaxially with anode cup 6. In this design, the cathode envelops the anode, which
reduces the triggering time and the spreading angle of the getter material.
The spreading angle of the material is increased in the design detailed in
Fig. 8.89b.
These two designs can be used in the devices with a limited getter-sputtering
area. An unlimited area makes it possible to sputter round, which provides best
pumping rates (Fig. 8.83c). In this case, cathode cylinder 2 is inside the anode, the
latter being a thin-walled tantalum cylinder 7 with a lanthanide foil either laser- or
contact-welded thereon.
Such pumps have the pumping rate of l l/s. and the volume of 1 cm3, the
designs themselves being sufficiently rigid for vacuum electronic devices.
The field emission cathodes of such pumps should have a number of design
properties. First of all, the effective area of the field emission cathodes is laid down
with elevations of 0.3–1 mm in width. They work effectively if the space between
the emitting microelevations on the effective areas of the elevations and the anode is
substantially smaller than the working part of the elevations. In this case, the
microelevations located on the effective area of the elevations make the best con-
tribution to the field emission. If the space increases, the electric field intensity on
the microelevations decreases and the elevation works as a whole, i.e., mostly by
the perimeter, which substantially weakens the overall performance of the field
emission cathode.
Experiments have established that the effect of the groove bottoms on the field
emission from the elevations is eliminated if the height of the elevation exceeds the
space between the anode and the effective area of the emitting elevations.
Except for the power consumption directly providing field emission, the former
is reduced due to special design properties, e.g., thermal-leakage reduction, parasitic
electron-scattering reduction etc., which are clear enough and do not need any
consideration in this section.
Testing involves a few experimental samples of each of the above-mentioned
designs. The evaporation-anode material is erbium. The tantalum cup is 5 mm in
diameter and 3 mm in thickness. The power consumption is 3 W at the pumping
rate of 1 l/s, which is about 2.5 times less than that of the analogues [201, 238].
Field emission heating is used in a light valve, which can function as a
high-resolution light projector [240], the design and the operating principle being
The device suggested contains a vacuum shell represented by flat glass envelope
1 with entrance window 2 in a form of a plane-parallel plate, target 3 with lobes 4 of
an intelligent material and a system of emissive field emission cathodes 5 as an
electron source (Fig. 8.90a). Target 3 includes a substrate represented by metalized
ceramic bars 6 (Fig. 8.90b) located in one plane in parallel to each other with a gap
in between and a band of an intelligent-material, e.g., nickel-titanium, each having
one series of cut T-shaped lobes 4 connected at base 7 with a film band of the same
shape as bar 6 and fixed on it with an electric contact. Base 7 bears dielectric layer
8. Every lobe 4 is set a reversible memory of two shapes, namely a rectilinear shape
in the target plane and a curved shape at an acute angle to the target, while lobes 4,
coated with a reflective material (missing from Fig. 8.90), form a mirror matrix.
The lobe is as long as the space between bars 6 and wider than the bar, the gap is
wider than the length of the bar, the gap is wider than the length of lobe stem 9, and
lobe body 10 is above the next bar. Field emission cathodes 5 can be made of
fine-grain high-density graphite MPG-6 in a form of combs isolated from each other
and mounted at right angle to bars 6 in the plane parallel to the plane of the
substrate. Comb elevations 11 are located in the gap between the bars opposite lobe
stems 9 at the distance shorter than the linear stem surface dimension, and the
dimensions of effective area 12 of the cathode elevation are smaller than the lobe
stem surface dimensions.
Every base 7 and every comb of field emission cathode 5 is connected to
terminals 13 sputtered on the bottom of envelope 1. The device has projection
optical system 14 consisting of two lenses, viz. 01 and 02, and diaphragm D, the
aperture center of the latter coinciding with the lens focuses.
The light valve works as follows.
When power supply is off, all lobes 4 are rectilinear and pressed to bars 6. As a
result, collimated light beam 15, which has passed through entrance window 2 of
envelope 1, is reflected from the mirror coating applied on the lobes and gets to the
entry to projection optical system 14. Supply of base 7 with a potential positive in
respect of the comb of field emission cathode 5 results in field emission from comb
elevation 11 located at their intersection. An electron beam affects stem 9 of lobe 4
Fig. 8.90 Design and operation diagram of a light-valve device a; its target with a system of field
emission cathodes. 1 Glass bulb; 2 input window; 3 target; 4 lobes made of material with shape
memory; 5 field emission cathodes; 6—metallized ceramic sliders; 7 base; 8 dielectric layer; 9 lobe
stem; 10 lobe body; 11 ledges of the field emission cathode; 12 working surface of a cathode
ledge; 13 leads; 14 optical projection system; 15 light beam
and causes heating and curving of the latter due to a structural-phase transition in
the lobe-stem material. The collimated light beam reflected from the lobe does not
pass through projection optical system 14. When the control voltage is turned off,
the lobe is pressed to the substrate. The entire array of lobes has the information
recorded in the same way.
The intelligent material lobes and data recording with field emission make it
possible for the device to be flat, need no electron beam focusing system or cathode
heating (unlike the prototype), provide the maximum power consumption of an
electron beam, which contributes to downsizing and performance-gaining of the
device. At the same time, it has an increased contrast and vibration resistance.
These advantages scale up application of light valves in data processing systems.
8.6.2 Sensors
Field emission cathodes can essentially increase sensitivity of sensor devices, i.e.
those responding to even a small change in any physical parameter.
In particular, paper [241] describes and rates a high-sensitivity magnetic sensor
based on a lateral field emission device, the arrangement drawing and the electric
connection diagram being given in Fig. 8.91.
Fig. 8.91 Design a and the

connection diagram b of a
vacuum magnetic sensor with
a field emission cathode: 1
control electrode; 3, 4 anodes;
5 quartz substrate
The core of the device is a comb-shaped field emission cathode 1 represented by

a 100–200 nm tungsten film sputtered on a quartz substrate. The optimum ratio of
the period to the width of the emitting part is equal to 3. The field emission cathode
is as wide as 1.4 mm, the emitters being as wide as 3 µm. Field emission cathode 1
and control electrode 2 are spaced 0.5 µm apart.
Anode 3, 4 is a sputtered layer of niobium divided into two parts. The device
works as follows. In absence of a magnetic field, emitted electrons move to anodes
along direct paths, the anode currents being of the same value. Normal application
of a magnetic field to the substrate surface curves the electron paths due to Lorentz
forces and imbalances the currents of anodes Ia1 and Ia2.
The magnetic sensitivity evaluated through the gradient of magnetic dependence
of the differential current is 1000%. It is almost 20 times as much as that for a
magnetoresistor made of a high-mobility material, e.g., GaAs. Such a sensor is
highly sensitive because the velocity of field electrons is 100 times higher than that
of semiconductors.
There have always been attempts to use field emission cathodes as the basis for
vacuum sensors [242], mechanical-impact sensors [243] etc.
8.6.3 Memory Cells
Undoubtedly, field-emission properties cause interest in developers of computer

memory and other devices. The following exemplifies development of these
devices.
Figure 8.92 shows a diagrammatic section of the memory cell based on carbon
nanotube field emission cathodes [244].
Fig. 8.92 a Sectional drawing of a memory element based on field emission cathodes made of
carbon nanotubes. The element consists of two triodes N1 and N2, marked by dashed outlines.
Triodes are assembled separately on two substrates and later connected as a single unit.
b Equivalent electric circuit of an element of a memory element. Cf is the main capacity, Rb is the
ballast resistor, Cbl is stray capacity
This element can be made according to the conventional micro- and nanotech-
nology. Note that the memory cell is made up of an upper part and a lower part
showed on two separate silicon bodies.
Body I is ion-implanted to form, firstly, a lightly-doped region functioning as
limiting resistor RB, and secondly, a heavily-doped layer specified in Fig. 8.92a as a
WL for the memory cell. Installation of vertical carbon nanotubes (CNT) and
reduction of capacitance of the control electrode needs relatively thick, viz.
1 µm, layer of silicon oxide. Subsequently, the oxide layer is sputtered with
metal and the coating is properly configured by the photolithographic method (cf.
the intermittent BL). The next stage is formation of modulator apertures through the
layers of oxide and metal (BL) by electron-beam lithography.
After that, an appropriate diffuse layer, e.g., TiN, and a catalyst, e.g., Ni, Fe, Co,
are sputtered through the modulator aperture, and vertical nanotubes are grown by a
CVD method. The grown-up carbon nanotubes are field emission cathodes of triode
N1. The BL area functions as the control electrode for triode N1 and the anode for
triode N2 (Fig. 8.92b).
Triode N2 nanotubes, i.e. the second part of the memory cell, are fabricated on
substrate II. At first, a thin layer of silicon oxide, viz. 20 nm, is grown up on
substrate II. The metal is sputtered into the oxide and etched away into islands,
forming bias capacitor CF (Fig. 8.92b). This metal layer serves as anode for triode
N1. A thick oxide layer is coated with a metal. This layer is configured by a
photolithographic method (SL line) and the aperture etched in the oxide lets the
triode—N1 electron beam reach the anode. Electron-beam lithography is used to
form an aperture in the control electrode and a place to grow up triode—N2
nanotubes.
It may be noted that substrates I and II can be fabricated both simultaneously and
separately.
Paper [245, 246] suggests field emission solid-state memory cells. The basic
principle applied is that the electric field between the emitter contact and the col-
lector electrode of the condenser-type film becomes high enough for the charge
carriers to get sufficient energy for tunneling through the collector into the insulated
retaining tank below the collector. The basic circuit of the cell is presented in
Fig. 8.93.
Flat substrate 1 can be made of silicon or glass. Silicon is most preferable since
its chemical, mechanical and electrical properties have been well studied.
Insulating oxide layer 2 is applied on substrate 1. It is at least 100 Å thick. This
layer 2 is coated with a layer of polysilicon 3 as thick as 500–100 Å coated with a
layer of oxide or nitride 4. The layer is selected depending on the subsequent
middle-band metal layer 5. The thickness of the layer ranges from 10 to 50 Å.
Metal layer 5 to be sputtered is as thick as a few nanometers. The thickness and
structure of this metal layer can be reproduced in production of semiconductor
chips.
The term of middle band means that the Fermi level of the collector material lies
in the middle between the conduction band and the valence band of the
charge-storage tank, i.e. polysilicon 3.
Fig. 8.93 Schematic diagram of a solid-state field emission element of memory: 1 substrate; 2
isolating oxidic layer; 3 polysilicon; 4 layer of nitride or oxide; 5 metal layer; 6 layer of nitride or
oxide; 7 conductor layer; 8 emitter contact; 9 first electric contact; 10 virtual capacitor; 11 earth
contact
After that, metal layer 5 is sputtered with the second oxide or nitride layer 6
similar to layer 4 in properties. The next stage is sputtering conductive layer 7
consisting of metal or polysilicon. It is as thick as layer 5. All these layers form an
island structure shown in Fig. 8.87.
First electric contact Vp 9 is connected to the middle-band layer 5. Second
emitter voltage 8 is connected to conductive layer 7. Two voltages applied to two
conductive layers through the insulator layer also function as a capacitor. While the
voltage of emitter VE is negative in respect of the collector voltage, the electrons
tunnel from emitter layer 7 to the middle-band metal layer 5. At the voltage of
VE 4 V, electrons reach first nitride layer 4, and some of them overcome the
potential barrier and move to tank 3 to be caught in this way.
Tank 3 has a capacitor connected to substrate 1 specified in Fig. 8.93 as virtual
capacitor 10. Substrate 1 is earthed through contact 11.
When the positive voltage of 6 V is applied to contact 8, the holes are injected
into area 3, which neutralizes the electrons.
A similar version of the memory cell with a silicon field emission cathode is
given in [247].
A very similar principle is the basis for a transistor, which can also be considered
as a memory cell.
8.6.4 Ion Sources
A high initial velocity of field electrons accounts for their good ionization ability.
This feature initiated the first attempts to design them 50 years ago [248], and the
work has still been under way.
Ion sources can be based on both pointed [249, 250] and carbon-fiber [251, 252]
field emission cathodes.
An ion-source structure with field emission cathodes made of carbon fibers is
Fig. 8.94 Low-background ionic source with a field emission cathode from carbon fibers: 1
copper plate; 2 ion lead opening; 3 ion beam; 4 field emission cathode; 5 metal disk; 6 bundles of
carbon fibers; 7 extraction electrode; 8 retarding grid; 9 Faraday cylinder; 10 ion lens; 11
molecular beam; 12 external thermal screen; 13 pipe; 14 liquid helium; 15 radiation screen
Original copper plate 1 has central aperture 2 to transmit ion beam 3 to the mass
spectrometer (missing from the picture). Field emission cathodes 4 rest on metal
disk 5, carbon-fiber bundles of 8 fibers 6 arranged symmetrically around the disk
circle, so that fiber tips were directed to the chamber center. Every bundle is about
1 mm in diameter.
Cylindrical extraction grid 7 is mounted inside field emission cathode 4. The
extraction voltage of 2–4 kV makes the electrons pass accelerating grid 7 and move
to decelerating grid 8. Grid 8 is under the potential of a few hundred volts. The
internal part of Faraday cup 9 is the ionization area where the decelerated electrons
oscillate and ionize gas molecules. A series of diaphragms 10 is an ionic lens
intended to take the ions out of the ionization area.
The left sides of grids 8 and 9 have central apertures intended to transmit
molecular beam 11. Faraday cup a is under the earth potential. Copper plate 1 is in a
good thermal contact with external screen 12 cooled by liquid nitrogen. An
ion-source electrode system is ducted in cylindrical tube 13 of a cryogenic pump or
a cryostat cooled by liquid helium. Tube 13 and other parts of the cryogenic pump
are made of stainless steel and connected with each other by welding.
Sensitivity of a field emission cathode-based ion source is by two orders of

magnitude higher than that of a hot cathode-based one.
Recent years saw further development of carbon nanotube based ionization
sources [253, 254]. In the former case [253], ionization is caused by an electron
beam generated by a large array of carbon nanotubes; in the latter case, the nan-
otubes are grown up on tungsten-apex surface.
8.6.5 Full-Color Projectors
Recently, the suggestions on development of tiny full-color light-emitting

diode-based projectors come from all over the world. The following is the struc-
ture diagram of such a device based on field emission light emitters [255].
The method suggested to form a light flux for a full-color system of mapping of
video information at the external screen is implemented as follows.
A light flux having the properties necessary to form full-color dynamic image
(i.e. homogeneous, collimated, controllable in terms of brightness, having
sequential spectrum of elementary frequencies for red, green and blue) is generated
directly in light source 1 (Fig. 8.95) by amplitude-controlled pulsed excitation of
emission transitions homogeneously distributed in a light-emitting body of the
emitters with a luminous efficacy necessary for comfortable observation of the
image at the external screen, the area of the latter being much larger than the light
modulator area and emitting sequentially collimated light in narrow spectral
intervals corresponding to primary colors.
The necessary collimated light flux is obtained by matrices, i.e. photon crystals
having the necessary set of allowed energy levels. The emitters embedded in such a
matrix emit light quanta in strictly given directions irrespective of the particular
nature of the excited emitter, e.g., electroluminophore. The magnitude variation of
the external exciting impulse excites different energy levels corresponding to the

of a full-color small-sized
projector: 1 light source; 2
light modulator; 3 lens; 4
external screen; 5 control unit
primary spectral colors, thereby changing the wavelength of the light emitted.
Variation of the inter-pulse period changes the brightness of the light flux.
Source 1 of the collimated light flux with a variable emission wavelength can
simultaneously replace a few elements providing an intensive light flux with the
necessary spectrum.
Based on the above principles, variable-wavelength source 1 works in the
highlight pulsed mode of 30,000 cd/m2. Every impulse makes source 1 generate
collimated light emission corresponding to red, green or blue colors. The light-flux
size meets the light-modulator 2 size that forms a polychromatic highlight tiny
image of no more than 20 20 mm projected by lens 3 to external screen 4. The
entire system is controlled from control unit 5.
The light flux emitted should have the maximum intensity in the desired spectral
intervals along with the minimum electrical-energy expenditure. It should also be
incoherent as otherwise there would be light interference destructive towards the
image formed and violating correct image transmission.
The device employs the field sequential color (FSC) system, when a full-color
image frame is formed of three basic components, wherefore the synchropulse
arriving at light source 1 makes the latter emit a light flux sequentially consisting of
three impulses of primary colors, viz. red, green and blue.
As an electron source, light source 1 employs a composite field emission cathode
based on carbon, viz. graphite, carbon, one-wall or multiple-wall carbon nanotubes
etc. Carbon-based field emission cathodes are able to work in high technical vac-
uum of 10−6–10−7 mm Hg achieved in conventional sealed-off devices and provide
high emissive capacity.
Up-to-date luminophores make it possible to achieve high light intensity, viz.
*30,000 cd/m2, which makes it possible to curtail the light-source dimensions by
making its emitting part *20 20 mm, i.e. equal to light modulator 2 in size.
The function of light modulator 2 can be performed by any controllable matrix
of microelements if their variable configuration can modulate the light flux from
light source 1.
Employment of field emission light sources in this type of devices provides the
following advantages:
– contrast improvement in the system based on a pulsing changeable-wavelength
light source;
– increase in the image-forming frequency in the system based on use of the above
light source and, respectively, non-use;
– a necessity to apply a composite light filter with an electromechanical principle
of light-filter displacement;
– reduction of dimensions of the image-forming device caused by lack of a
composite light filter based on an electromechanical principle of light-filter
displacement, a liquid-crystal shutter and a joined optical system consisting of
the first and the second lens arrays, a collimator lens and heat-removal fans;
– reduction of noise and undesirable vibrations connected with lack of elec-
tromechanical light-filter devices and heat-removal fans;
– increase in the efficiency factor of transformation of electrical energy into light

in this method of forming a light flux for image display connected with the
above-listed design properties.
Because of edition-size limits, we have to finish reviewing the devices
employing field emission cathodes. They should be the subject of separate publi-
cation. Let us note that field emission cathodes can be used in different wireless
devices such as noise generators [256] or radio-frequency power amplifiers [257,
258]. In addition, effective field emission photocathodes [259] and powerful elec-
tron guns [260] are under development.
Practical interest to field emission cathodes stretches from displays through cars
[261] to space-vehicle construction [262] and probe tunnel microscopes [263].
References
1. A.S. Bugaev, V.B. Kireev, E.P. Sheshin, A.Y. Kolodyzhnyj, Сathodoluminescent light
sources: status and prospects. Physics-Uspekhi v 58(2), 792–818 (2015)
2. Y.B. Ayzenberg, Lighting Engineering (Energoatom, M., 1983)
3. E.P. Sheshin, Structure of a Surface and Field Emission Properties of Carbon Materials
(MIPT, M., 2001)
4. E.P. Sheshin, A.L. Suvorov, A.F. Bobkov, D.E. Dolin, Light source on the basis of carbon
field electron cathodes: design and parameter, in Abstract of 7th International Vacuum
Microelectronics Conference, Grenoble, France (1994), pp. 423–426
5. A.S. Baturin, V.S. Kaftanov, S.G. Kyzmenko, E.P. Sheshin, Field emission device. Patent R.
F. No. 2180145, Cl. H01J/28 of 15.02.2000
6. M.Y. Leshukov, A.S. Baturin, N.N. Chadaev, E.P. Sheshin, Characterizations of light
sources with carbon fiber cathodes. Appl. Surf. Sci. 215, 260–264 (2003)
7. M.Y. Leshukov, N.N. Chadaev, A.S. Baturin, E.P. Sheshin, Vacuum cathode luminescence
lamps for systems of display of information, in Proceedings of Conference on Modern
Television of Moskow (2004), pp. 36–38
8. M.O. Popov, S.Y. Bulakhon, A.V. Karpov, S.A. Shiriaev, E.N. Skorokhodov, A.L. Suvorov,
Influence of parameters of field emission cathodoluminescence light sources on their
technical properties. Appl. Surf. Sci. 215, 253–259 (2003)
9. E.P. Sheshin, A.Y. Kolodyazhnyiy, A.S. Raufov, Sources of general lighting with field
emission cathodes from carbon materials. Izvestiya Vysshikh Uchebnykh Zavedenii; Seriya
Khimiya I Khimicheskaya Tekhnologiya 58(7), 69–73 (2015)
10. M.Y. Leshukov, E.P. Sheshin, Optimization of an electron-optical system of a cathode
luminescent light source with the field emission from carbon fibers, in Proceedings of
Conference on Stability and Management Processes, vol. 1, St. Petersburg (2005), pp. 202–
208
11. M.Y. Leshukov, E.P. Sheshin, New design of electron gun for field emission light sources
with carbon fibers cathode. Hydrogen Materials Science and Chemistry of Carbon
Nanomaterials. NATO Security Through Science Series A: Chemistry and Biology, Special
issue, XLI (Springer Science, 2007), pp. 255–258
12. M.Y. Leshukov, N.N. Chadaev, E.P. Sheshin, Field emission light source with carbon fibers
bundle cathode. Technical Digest of IVNC, Oxford, UK (2005), pp. 340–341
13. W. Knapp, O.F. Kieler, D. Schleussner, E.P. Sheshin, I.N. Yeskin, Fieldemission von
kohlenstoff-fasern und deren anwendung in einem CRT-lightelement. Konferenzband
Electronic displays, Berlin (2000), pp. 92–98
14. W. Knapp, D. Schleussner, A.S. Baturin, I.N. Yeskin, E.P. Sheshin, CRT lighting element
with carbon field emitters. Vacuum 69, 339–344 (2003)
15. H. Murakami, M. Hirakawa, C. Tanaka, H. Yamakawa, Field emission from well-aligned,
patterned, carbon field emitters. Appl. Phys. Lett. 76(13), 1776–1778 (2000)
16. Y. Saito, K. Hamagychi, R. Mizushima, S. Uemura, T. Nagasako, J. Yotani, T. Shimojo,
Field emission from carbon nanotubes and its application to cathode ray tube lighting
elements. Appl. Surf. Sci. 146, 305–311 (1999)
17. Y. Saito, K. Hata, A. Takakura, J. Yotani, S. Uemura, Field emission of carbon nanotubes
and its application as electron sources of ultra-high luminance light-sours devices. Physica B
B323, 30–37 (2002)
18. Y.-C. Yang, J. Tang, L. Liu, S.S. Fan, Field emission device. Patent USA №US
2008/0007153 of 30.11.2006. Cl 313/497 (2006)
19. Y.-C. Yang, J. Tang, L. Liu, S.S. Fan, Field emission device. Patent USA №US
2008/0007153 of 30.11.2006. cl.313/496 (2006)
20. A.N. Obraztsov, V.J. Kleshch, E.A. Smolnikova, A nano-graphite cold cathode for an
energy-efficient catholuminescent light source. Beilstein J. Nanotechnol. 4, 493–500 (2013)
21. K.-J. Chung, N.N. Pu, M.J. Youh, Y.-M. Liu, M.-D. Ger, W.-K. Huang, Improvement of
lighting uniformity and phosphor life in field emission lamps using carbon nanocoils.
J. Nanomater. 2015, ID373549 (2015) 9 p
22. Y.-C. Yang, Y. Wei, L. Liu, K.-L. Jiang, S.S. Fan, Field emission device. Patent USA №US
2008/0012466 of 16.11.2006, cl 313/495 (2006)
23. Y.-C. Yang, Y. Wei, L. Liu, K.-L. Jiang, S.S. Fan, Pixel tube for field emission device.
Patent USA № US 2008/0030123 of 12.07.2007, cl.313/495 (2007)
24. www.inanov.fr
25. Y. Kondo, J. Matsuura, H. Kimura, Characteristics of field emitter arrays for light source.
Technical Digest IVMC 91, Nagahama (1991), pp. 56–57
26. Y.L. Ke, J. Zhou, J. Chen, S.Z. Deng, N.S. Xu, Luminescent tubes based on MoO2 nanowire
cold cathode. Technical Digest JVNC, Oxford, UK (2005), pp. 362–363
27. V.S. Kaftanov, A.L. Suvorov, E.P. Sheshin, Field emission cathode and methods in the
production thereof. Patent USA. №5588893 cl.445/6(H01 J 9/02) of 6.06.1995
28. V.S. Kaftanov, A.L. Suvorov, E.P. Sheshin, A field emission cathode and methods in the
production thereof. Patent RST № WO 96/25753 of 15.02.1995
29. V.S. Kaftanof, A.L. Suvorov, E.P. Sheshin, Metod vid tillverkning av en faltemissions
cathode samt faltemissions cathode. Patent Sweden №504603 of 02.15.1995 cl.H01 J/30
30. V.S. Kaftanov, A.L. Suvorov, E.P. Sheshin, Field emission cathode and methods in the
production thereof. Patent USA № 5973446 of 15.02.1999 cl.313/310 (H01 J 1/05)
31. G. Forsberg, C.-H. Andersson, Method of producing a field emission cathode and a light
source. Patent RST WOOO/77813 A1 of 10.06.1999
32. W.-H. Tsai, M.-C. Fan, J.-C. Chang, Flat field emission illumination module. Patent USA №
US 2008/0036361 of 14.02.2008 cl.313/496 (H01j 1/62)
33. W-H. Tsai, M.-C. Fan, J.-C. Chang, Flat field emission illumination module. Patent UK GB
2441618, of12.03.2008. cl. H01j 63/02
34. P.J. Chen, P. Liu, L.M. Sheng, Y. Wei, L. Liu, Z.F. Hu, C.L. Guo, S.S. Fan, Light source
apparatus using field emission cathode. Patent USA US 2006/0022574 of 2.02.2006,
cl.313/495 (H01J 9/02)
35. J.H. Choi, A.R. Zoulkarneev, Y.T. Park et al., Cathodoluminescent flat lamp using the
suspend wire structure. Technical Digest of JVNC 2005, Oxford, UK (2005), pp. 350–35
36. K.W. Cheng, S.H. Lee, C.Y. Hsiao, Field emission type planar lamp and method for the
same. Patent USA US 2008/0079348 of 3.04.2008 cl.313/495 (H01 J 1/02)
37. L. Qian, L. Lin, P. Liu, J. Tang, Y. Wei, S.S. Fan, Field emission double-plane light source
and method for making the same. Patent USA US 2007/0222363 of 27.09.2007 cl.313/496
(H01 J 63/04)
References 529
38. A.N. Obraztsov, J.Y. Pavlovsky, A.P. Volkov, Prototype of light emitting device ith thin
film cold cathode. Technical Digest of Euro Display 99, Berlin, Germany (1999), pp. 229–
231
39. A. Gorecka-Drzazga, B.J. Cichy, P. Szczepanska, R. Walczak, J.A. Dziubun, Field emission
light sources for lab-on-a-chip microdevices. Bull. Polish Acad. Sci. Techn. Sci. 60(1), 13–
17 (2012)
40. Y. Kuroiwa, S. Narushima, S. Ito, Electron emitter, field emission display unit, cold cathode
florescent tube, flat type lighting device, and electron emitting material. European patent EP
1876628 29.01.2008 H01 J 1/304 (2008)
41. V.B. Sharov, E.P. Sheshin, A.A. Shyka, Field emission cathodes from graphite for diode
light sources. Nano Microsyst. Equip. 3, 17–19 (2005)
42. M.-J. Youh, C.-L. Tseng, M.-H. Jhuang, S.-C. Chiu, L.-H. Huang, J.-A. Gong, Y.-Y. Li, Flat
panel light source with lateral gate structure based on SiC nanowire field emitters. Sci.
Rep. 5, 10976, c1–34 (2015)
43. V.S. Kaftanov, A.L. Suvorov, E.P. Sheshin, J. Olsford, Field emission cathode and a light
source including a field emission cathode. Patent USA №5877588 of 5.03.1999, cl.313/491
(H01 J 1/30)
44. V.S. Kaftanov, A.L. Suvorov, E.P. Sheshin, J. Olsford, Light source including a field
emission cathode, and a field emission cathode. Patent USA 6008575 of 28.12.1999, cl.
313/484 (H01 J 1/30)
45. E.P. Sheshin, A.L. Suvorov, V.S. Kaftanov, J. Olstors, Field emission cathode and a light
source including a field emission cathode. Patent Canada CA 2293269 cl, H01 J 001/30 of
17.12.1998
46. A.A. Blyablin, A.V. Kandidov, A.T. Rakhimov, V.A. Samorodov et al. Divergence of
electrons emitted from a carbon cold cathode cold cathode. Technical Digest of IVMC 1999,
Darmstadt, Germany (1999), pp. 346–347
47. J.-M. Bonard, T. Stockli, O. Noury, A. Chatelain, Field emission from cylindrical carbon
nanotube cathodes: possibilities for luminescent tubes. Appl. Phys. Lett. 78(18), 2775–2777
(2001)
48. X.X. Zhang, C.C. Zhu, X. Li, W.Z. Cui, The fabrication of novel structure of field emitting
light tube with carbon nanotubes as cathode. Technical Digest of IVNC, Oxford, UK (2005),
pp. 342–343
49. J.X. Huang, J. Chen, S.Z. Deng, J.C. She, N.S. Xu, Optimization of carbon nanotube
cathode for a fluorescent lamp. Technical digest of IVNC, Oxford, UK (2005), pp. 284–285
50. J. Kjellman, M. Lindmark, Light source, and a field emission cathode. Patent USA 6873095
of 29.03.2005 313.336 (H01 J 1/16)
51. Y. Yun, G. Tailiang, J. Yadong. Application of ZnO nanopillars and nanoflowers to field
emission luminescent tubes. J. Semiconductors, 33(4), 043003, 1–5 (2012)
52. A.N. Obraztsov, A.P. Volkov, A.A. Zakhidov, D.A. Lyashenko, Y.V. Petrushenko, O.
P. Satanovskaya, Field emission characteristics of nanostructured thin film carbon materials.
Appl. Surf. Sci. 215, 214–221 (2003)
53. A.N. Obraztsov, Cathodoluminescent light source. Patent USA US 005/0174059 of
11.08.2005 cl. 313/634 (H01 J 63/04)
54. L. Qian, L. Lin, P. Liu, J. Tang, Y. Wei, S.S. Fan, Field emission lamp and method for
making the same. Patent USA US 2007/0228919 of 4.10.2007, 313/310, (H01 J 9/02)
55. A.S. Leychenko, M.Y. Leshukov, N.N. Chadaev, E.P. Sheshin, Effective lamp for
LCD-backlighting with the field emission cathode. Technical Digest IVNC IFEC, Guilin,
China (2006), pp. 383–384
56. A.S. Leychenko, M.Y. Leshukov, N.N. Chadaev, E.P. Sheshin, Field emission lamp
illuminations for l.c. displays, in Proceedings of Modern Television Conference, Moscow
(2006), pp. 30–31
57. A.L. Suvorov, E.P. Sheshin, N.E. Lazarev, N.N. Chubun, Vacuum luminescent light source
with carbon fibres field emission cathode. Technical Digest of IVMC, Portland USA (1995),
pp. 516–521
58. P. Liu, Y. Wei, L.M. Sheng, L. Qian, J. Tang, L. Liu, C.L. Guo, C.L. Du, B.C. Du, S.S. Fan,
Field emission luminescent light source. Patent USA US 2006/0091782, of 04.03.2006
cl.313–496 (H01 J 63/04)
59. K.J. Chung, C.C. Chiang, Y.M. Liu, N.W. Pu, M.D. Ger, The study of fabricating the field
emission lamps with carbon nano materials. Int. J. Chem. Mol. Nucl. Mater. Metall. Eng. 6
(5), 437–439 (2012)
60. L. Qian, L. Liu, P. Liu, J. Tang, Y. Wei, S.S. Fan, Field emission lamp and method for
making the same. Patent USA US 2007/0247071 of 25.10.2007. cl.313/634 (H01 J 61/30)
61. C. Li, K. Song, C. Lan, Field emission excited UV light source structure and preparation
method there of. Patent CN 103 400919 от 20.11.2013 кл.HOIj-063/06
62. S. Ono, T. Suyama, K. Fukura, S. Ishizu, N. Kawaguchi, T. Nagami, A. Yoshikawa, T.
Yanagida, Y. Yokota, Device for emitting vacuum ultraviolet light. Patent W0201127881 от
10.03.2011,кл.Н01о-061/38,Н01j-063/00
63. M. Yanagihara, M.Z. Yusop, M. Tanemura, S. Ono, T. Nagami, K. Fukuda, T. Suyama, Y.
Yokota, T. Yanagida, A. Yoshikawa, Vacuum ultraviolet field emission lamp utilizing
KMgF3 thin film phosphor. APL Mater. 2, 046110 (2014)
64. V.D. Blank, S.G. Buga, I.V. Ekhmenina, N.N. Chadaev, E.P. Sheshin, Patent RF
№2529014, kl.N01j 61/06, H05B33/12 ot 27.09.14. Lampa vakuumnaya ul’trafioletovogo
diapazona spektra
65. I.V. Ekhmenina, E.P. Sheshin, N.N. Chadaev, Problemy sozdaniya ul’trafioletovyh
istochnikov na osnove nanostrukturirovannyh avtoehmissionnyh katodov. ZHurnal
«Nano- i mikrosistemnaya tekhnika» 2, str.39–45 (2010)
66. I.V. Ekhmenina, E.P. Sheshin, N.N. Chadaev, Istochniki izlucheniya na osnove nanostruk-
turirovannyh avtokatodov. Zhurnal «Nano- i mikrosistemnaya tekhnika» 4, str.45–48 (2010)
67. I.V. Ekhmenina, E.P. SHeshin, N.N. CHadaev, Avtoehmissionnyj istochnik ul’trafiole-
tovogo izucheniya s avtokatodom iz nanostrukturirovannogo uglerodnogo materiala. Vestn.
S.-Peterburg, un-ta. Ser.10. Prikl.matem. Inform.Proc.upr. 1, str.3–8 (2011)
68. I.V. Ekhmenina, E.P. SHeshin, Issledovanie vliyaniya razlichnyh faktorov na ehffektivnost’
katodolyuminescencii s cel’yu sozdaniya konkurentosposobnogo avtoehmissionnogo
istochnika izlucheniya N ZHurnal «Trudy Moskovskogo fiziko-tekhnicheskogo instituta»,
tom 5, 1(17), str.36–43 (2013)
69. I.V. Ekhmenina, E.P. SHeshin, Issledovanie harakteristik izlucheniya avtoehmissionnyh
lamp s avtokatodami iz nanostrukturirovannogo uglerodnogo materiala. Izvestiya vysshih
uchebnyh zavedenij. Seriya: Himiya i himicheskaya tekhnologiya. T.56(5), S.74–76 (2013)
70. I.V. Ekhmenina, E.P. Sheshin, N.N. Chadaev, Source of ultraviolet radiation with field
emission cathode made of nanostructured carbon materials.
Springer_Book_BWF_2164446_Carbon Nanomaterials in Clean Energy Hydrogen
System-II (2011), pp. 299–303
71. H.S. Kang, J.T. Han, Y.W. Jin, M.J. Bae, Y.J. Park, Field emission backlight unit, method of
driving the backlight unit, and method of manufacturing lower panel. Patent USA US
2008/0106221 of 8.05.2008 cl.315/334 (H01 J 19/24)
72. B.N. Lin, C.H. Fu, Field emission backlight unit and scanning driving method. Patent USA
US 2008/0100235 of 1.05.2008 cl.315–349 (H05B 37/02)
73. Y.J. Jung, J.H. Park, J.-S. Jeong, J.W. Nam, A.S. Berdinsky, J.B. Yoo, C.Y. Park,
Fabrication and characteristics of flat lamp with CNT based triode structure for back light
unit in LCD. Technical Digest of IVNC 2005, Oxford, UK (2005), pp. 202–203
74. S.H. Park, C.W. Baik, J.H. Lee, Y.W. Jin, Field emission type backlight unit and method of
manufacturing the same. Patent USA US 2007/0229003 4.10.2007 315/324 (H01J 1/02)
75. J.H. Choi, B.G. Song, M.J. Shin, A. Zoulkarneev, D.S. Chung, M.J. Bae, Field emission
backlight device and method of fabricating. Patent USA US 2005/0179363 of
18.08.2005 cl.313/497 (H01J 1/62)
76. A.S. Batyrin, N.N. Chadaev, E.P. Sheshin, N.A. Duzhev, Y.I. Tishin, Development of a
lamp of illumination for liquid crystal screens, in Proceedings of Conference on “Modern
Television” Moscow (2001), pp. 21–23
References 531
77. A.S. Baturin, N.N. Chadaev, M.Y. Leshukov, A.J. Trufanov, E.P. Sheshin,
Cathodo-luminescent dynamic backlighting for color liquid crystal displays, in
Proceeding of Conference on Displays and Vacuum Electronics, Garmisch-partenkirchen,
Germany (2004), pp. 321–323
78. J. Gorog, P.M. Ritt, Liquid crystal display having a field emission backlight. Patent PCT
WO 2008/002321of 03.01.2008 cl. (H01J 9/227)
79. M.Y. Leshukov, N.N. Chadaev, E.P. Sheshin, Three-colored cathode luminescent lamp for
systems of display of information, in Proceedings of Conference on “Modern Television”,
Moscow (2005), pp. 37–38
80. M.Y. Leshukov, N.N. Chadaev, A.S. Baturin, K.H. Nikolskiy, R.G. Chesov, Illumination of
liquid crystal screens cathode luminescent lamps with field emission cathodes from carbon
fibers, in Proceedings of Modern Television Conference, Moscow (2002), pp. 30–31
81. N. Abanshin, A. Ezhkov, P. Ivashkin, S. Kashyrin, A. Vrotov, Y. Timofeev, Cathode
luminescent radiators for display units of collective use. Electron. Compon. 4, 114–116 (2007)
82. A.I. Tryfanov, A.S. Batyrin, M.Y. Leshukov, N.N. Chadaev, E.P. Sheshin, Emission
characteristics of a light source with field emission cathode on the basis of a bunch carbon
fibers. Microsyst. Equip. 3, 32–35 (2004)
83. A.S. Leychenko, M.Y. Leshukov, N.V. Luparev, P.A. Starikov, N.N. Chadaev, E.P. Sheshin,
Element of the big video screen with field emission cathode luminescent light sources as
pixels, in Proceedings of “Modern Television” Conferences, Moscow (2007), pp. 45–46
84. A.V. Kudryashov, E.P. Sheshin, N.N. Chadaev, M.M. Kustikov, The element of a big size
screen based on light emitters with field emission cathodes, in Abstract of International
Vacuum Electron Sources Conference (IVESC), London (2008)
85. D.S. Strebkov, V.Z. Trubnikov, A.V. Pastuhov, E.P. SHeshin, N.N. CHadaev, Sistema
ehlektricheskogo osveshcheniya (varianty). Patent RF №2505 744, ot 27.01.2014, kl.
F2159/00
86. S. Groznov, A.S. Leychenko, E.P. Sheshin, A. Shyka, Flat display screens on a basis the
field emission cathodes. Chip News 7, 21–25 (2008)
87. K.R. Shoulders, Microelectronics using electron-beam-activated machining techniques, in
Advances in Computers, vz, (1961), pp. 135–197
88. S.A. Spindt, A thin film field emission cathode. J. Appl. Phys. 39(7), 3504–3505 (1968)
89. C.A. Spindt, J. Brodie, L. Humphrey, E.R. Westarber, Physical properties of thin film field
emission cathode with molybdenum cones. J. Appl. Phys. 47(12) (1976)
90. R. Meyer, Recent development on «microtips» display at LETI. Technical Digest of IVMC
1999, Hagahama, Sapan (1999), pp. 6–9
91. S. Jtoh, T. Watanabe, T. Yamaura, K. Yano, A challenge to field emission displays. Asia
Displays 95, 617–620 (1995)
92. S. Jtoh, Current status of the field emission display. ASET International Forum on Low
Power Displays, Shinagawa (2000), pp. 59–68
93. S. Stoh, M. Tanaka, T. Tonegawa, Development of field emission display. Technical Digest
of IVMC 2003, Osaka, Japan (2003), pp. 19–20
94. B.E. Russ, J. Barger, J. Wang, Field emission cathode structure using perforated gate.
Patent USA US 2003/0193297 of 16.10.2003, cl.315/169 (H05B 39/04)
95. S. Kubota, K. Kikuchi, H. Sata, Cold cathode field emission device, process for the
production there of, and cold cathode field emission display. Patent USA US 2002/0050776,
of 2.05.2002, cl.313/309 (H01J 9/02)
96. Y. Kamide, S. Kubota, H. Sata, K. Kikuchi, Cathode panel for a cold cathode field emission
display, and cold cathode field emission display, and method of producing cathode panel for
a cold cathode field emission display. Patent USA US 6917155 of 15.07.2005, cl.313/495
(H01J 1/62)
97. Y. Kamide, S. Kubota, H. Sata, K. Kikuchi, Cathode panel for a cold cathode field emission
display and cold cathode field emission display, and method of producing cathode panel for
a cold cathode field emission display. Patent USA US 2005/0236964 of 27.10.2005,
cl.313/495(H01J 1/62)
98. M. Konishi, K. Jida, Cold cathode electric field electron emission display device.
Patent USA US 2006/0087248 of 27.04.2006, cl.315/169, 3 (G09G 3/10)
99. E. Negishi, Method for treating a cathode panel, cold cathode field emission display device,
and method for producing the same. Patent USA US 2008/0012467 of 17.01.2008,
cl.313/495 (H01J 1/62)
100. M. Konishi, Cold cathode electric field electron emission display device. Patent USA US
2005/0082964 of 21.04.2005, cl.313/497 (H01J 31/12)
101. D. Sarrasin, Method of driving a matrix display device having an electron source with
reduced capacitive consumption. Patent USA, 8477156B2, 2.07.2013, cl.345/690
(G09G5/10)
102. M. Yamamoto, K. Koga, A. Shiota, S. Kanemaru, M. Nagao, Field emission electron source.
Patent USA US 2005/0001536 of 6.01.2005, cl.313/497, (H01J 1/02)
103. S.-T. Yan, Field emission display and method for manufacturing same. Patent USA US
2008/0074031 of 27.03.2008 cl.313/496, (H01J 63/04)
104. G.-L. Chen, Field emission display device. Patent USA US 6646282 of 11.11.2003,
cl.257/10 (H01L 29/06)
105. K.S. Choi, S.J. Lee, J.M. Kim et al., FED devices containing a novel graphite cathode
prepared by a screen printing process, in Digest of 12th JVMC, Darmstadt, Germany (1999),
pp. 32–33
106. K.S. Choi, S.J. Lee, J.M. Kim et al., Field emission display devices containing a novel
graphite cathode prepared by a screen printing process. Digest SID 00 (2000), pp. 671–673
107. W.B. Choi, D.S. Chung, J.H. Kang, H.Y. Kim et al., Fully sealed, hight-brightness
carbon-nanotube field emission display. Appl. Phys. Lett. 75(20), 3129–3131 (1999)
108. J.M. Kim, W.B. Choi, N.S. Lee, J.E. Jung, Field emission from carbon nanotubes for
displays. Diam. Relat. Mater. 9, 1184–1189 (2000)
109. Q.H. Wang, A.A. Setlur, J.M. Lauerhaas, J.Y. Dai, E.W. Seelig, R.P.H. Chang, A
nanotube-based field emission flat panel display. Appl. Phys. Lett. 72(22), 2912–2913
(1998)
110. Y. Nakayama, S. Akita, Field emission device with carbon nanotubes for a flat panel display.
Synth. Metals 117, 207–210 (2001)
111. C.C. Kuo, W.S. Hsu, C.Y. Hsao, Sintering method for carbon nanotube cathode of field
emission display. Patent USA US 2006/0009110 of 12.01.2006. cl.445/50 (H01J 9/12)
112. Y.H. Chien, C.P. Peng, C.H. Fu, W.Y. Lin, L.H. Chan, Field emission system and method
for improving its vacuum. Patent USA US 2008/0042547 of 21.02.2008, cl.313/495 (H01J
63/04)
113. E.P. Sheshin, A.M. Kyrnosov, Flat field emission device to display the image. Patent PF ru
2178598 from 20.01.2002 CL.H01J 31/12
114. A.S. Baturin, N.N. Chadaev, K.N. Nikolski, R.G. Tchesov, E.P. Sheshin, I.N. Yeskin,
Application of cathode materials for field mission display, in Proceedings of Conference
Displays and Vacuum Electronics, Garmisch-partenkirchen, Germany (2004)
115. A.Y. Tcherepanov, A.G. Chakhovskoi, V.B. Sharov, Flat panel display prototype using
low-voltage carbon field emitters. J. Vac. Sci. Technol. B 13(2), C.482–486 (1995)
116. Q.H. Wang, M. Yan, R.P.H. Chang, Flat panel display prototype using gated carbon
nanotube field emitters. Appl. Phys. Lett. 78(9), 1294–1296 (2001)
117. L. Yukui, Z. Changchum, L. Xinghui, Field emission display with carbon nanotubes cathode
prepared by a screen-printing process. Diam. Relat. Mater. 11, 1845–1847 (2002)
118. B.E. Russ, J. Barger, Field emission display using line cathode structure. Patent USA US
2003/0193296 of 16.10.2003 cl.315/169.3 (G09 G3/10)
119. A. Okamoto, K. Konuma, Y. Tomihari, F. Ito, Y. Okada, CNT film and field emission cold
cathode comprising the same. Patent USA US 2004/0104660 of 3.01.2004 cl.313/346 R
(H01j 19/06)
References 533
120. W.B. Choi, Y.W. Jin, H.Y. Kim, S.J. Lee et al., Electrophoresis deposition of carbon
nanotubes for triode-type field emission display. Appl. Phys. Lett. 78(11), 1547–1549 (2001)
121. M. Nakamoto, Field emission type cold cathode device, manufacturing method thereof and
vacuum micro device. Patent USA US 2004/0265592 of 30.12.2004 cl.428/408 (B32B
15/04)
122. X. Xu, C.P. Beetz, G.R. Brandes, R.W. Boerstler, J.W. Steinbeck, Carbon fiber-based field
emission devices. Patent USA, №5973444 of 26.10.1999 cl.313/309 (H01J 1/30)
123. M.A. Gruillom, M.L. Simpson, V.J. Merkulov, A.V. Melechko, D.H. Lowndes, Gated
fabrication of nanostructure field emission cathode material within a device. Patent USA US
6858455 of 22.02.2005. cl.438/20 (H011 21/00)
124. Z. Zheng, S.-S. Fan, Method for manufacturing cathode assembly of field emission display.
Patent USA US 2007/0287350 of 13.12.2007 cl.445/24 (H0iJ 9/00)
125. H.J. Kim, J.T. Han, Y.C. Choi, K.S. Jeong, Method of forming a carbon nanotube structure
and method of manufacturing field emission device using the method of forming a carbon
nanotube structure. Patent USA US 2008/0003733 of 3.01.2008 cl.438/197 (H01L 21/8234)
126. M. Muroyama, T. Yagi, K. Jnoue, S. Saito, Electron emitter and its production method.
Patent USA US 2004/0108515 of 10.06.2004, Cl.257/144 (H01L 21/00)
127. S. Kang, C. Bae, J. Kim, Barrier metal layer for a carbon nanotube flat panel display.
Patent WO 2004/064099 of 29.07.2004
128. A.A. Blyablin, A.T. Rakhimov, V.A. Samorodov, N.V. Suetin, Method for producing an
addressable field emission cathode and an associated display structure. Patent USA US
2003/0143321 of 31.07.2003. Cl 427/77 (B 05 D5/12)
129. Z. Sun, B.K. Tay, S.P. Lau, Y. Li, Field emission device and method of fabricating same.
Patent USA US 2005/0077811 of 14.04.2005 Cl.313/495(H01j 1/62)
130. M. Okai, T. Muneyoshi, T. Yaguchi, N. Hayashi, Flat panel display. Patent USA US
2006/0049743 of 09.03.2006, cl. 313/496 (H01j 1/62)
131. F. Ito, Field emission type cold cathode and method for manufacturing the same and method
for manufacturing flat display. Patent USA US 2003/0080663 of 1.05.2003. CL.313/336
(H01j 1/16)
132. H. Lee, S. Lee, S. Park, Triode structure field emission display device using carbon
nanotubes and method of fabricating the same. Patent USA US 2005/040752 of 24.02.2005,
Cl.313/495 (H01j 1/02)
133. K.A. Dean, B.F. Coll, E.M. Howard, L.L. Tisinger, Method for reducing leakage current in a
vacuum field emission display. Patent USA US 2007/0097567 of 03.05.07, cl.361/42
(H02H9/08)
134. W. Gu, W. Lei, X. Zhang, Influence of the dielectric layer on the performance FED.
Technical Digest of IVNC, Oxford, UK (2005), pp. 330–331
135. S.V. Johnson, Discharge of a field emission display based on charge accumulation.
Patent USA US 2008/0048570 of 28.02.2008, cl.315/169.3 (G09 G3/10)
136. M.J. Youh, C.L. Tseng, Y.H.J. Huang, M. Liu, Triode field emission cold cathode device
with random distribution and method. Patent USA US 2005/0104506 of 19.05.2005,
cl.313/496 (Ho1j 1/62)
137. N.S. Xu, Z.S. Wu, S.Z. Deng, J. Chen, High-voltage triode flat-panel display using field
emission nanotube-based thin films. J. Vac. Sci. Technol. B19(4), 1370–1372 (2001)
138. T. Yaguchi, T. Muneyoshi, M. Okai, N. Hayashi, T. Nakamura, Flat panel display device.
Patent USA US 2005/0057178 of 17.03.2005 of 315/169.4 (G 09 G 3/10)
139. T. Oh, Field emission display. Europatent EP 1542258 of 15.06.2005 cl.H01j 31/12
140. K. Oono, Field emission display device and method of manufacturing same. Patent USA US
2004/0239235 of 2.12.2004, cl.313/497 (H01j 1/62)
141. P.S. Ahn, H.W. Lee, Field emission device with focusing control electrode and field
emission display. Patent USA US 2005/0189868 of 1.09.2005. Cl 313/497 (H01j 1/02)
142. J.T. Han, Field emission display and method of manufacturing the same. Patent USA US
2005/0110393 of 26.05.2005, cl.313/495 (H01j 1/62)
143. Y.H. Song, J.H. Lee, K.Y. Kang, Field emission device and field emission display device
using the same. Patent USA US 2006/0290259 of 28.12.2006 cl.313/495 (H01j 63/04)
144. Y.H. Song, J.H. Lee, C.S. Hwang, Field emission display. Patent USA 2005/0248256 of
10.11.2005, cl 313/495 (H01j 1/62)
145. M. Nakamoto, Field emission cold cathode device of lateral type. Patent USA US
2004/0183421 of 23.09.2004 cl.313/309, (H01j 1/62)
146. M. Nakamoto, Field emission cold cathode device of lateral type. Patent USA US
2006/0061257 of 23.03.2006. Cl.313/496 (H01j 63/04)
147. Y. Takeuchi, T. Nanataki, J. Ohwada, T. Horiuchi, Electron-emitting device and field
emission display using the same. Patent USA US 7088049 of 8.08.2006 cl.315/169.1 (G 09
G3/10)
148. H. Kawamura, S. Kafo, T. Maki, T. Kobayashi, Fabrication of planar diamond electron
emitters for flat panel displays. Mat. Res. Soc. Symp. Proc. 558, 155–160 (2000)
149. T.S. Oh, Field emission display. Patent USA US 2004/0222734 of 11.11.2004 cl.313/497
(H01j 1/62)
150. H. Moon, Field emission display device. Patent USA US 2005/0093424 of
5.05.2005 cl.313/495 (H01j 1/62)
151. J.H. Kang, Y.S. Choi, W.B. Choi, N.S. Lee et al., Under-gate triode type field emission
displays with carbon nanotube emitters. Mat. Res. Soc. Symp. Proc. 621, P.R 5.2.1–R 5.2.5
(2000)
152. Y.S. Choi, J.H. Kang, Y.J. Park, W.B. Choi et al., An under-gate structure field emission
display with carbon nanotube emitters. Diam. Relat. Mater. 10, 1705–1708 (2001)
153. Y.S. Choi, J.H. Kang, H.Y. Kim, B.G. Lee et l, A simple structure and fabrication of
carbon-nanotube field emission display. Appl. Surf. Sci. 221, 370–374 (2004)
154. V.P. Mammana, G.E. Mc Guire, O.A. Shenderova, Back-gated field emission electron
source. Patent USA, US 2005/0116214 of 02.06.2005 cl.257/10 (H01 L 29/06)
155. S.H. Ahn, S.J. Lee, Field emission display and method of manufacturing the same.
Patent USA US 2005/0093426 of 05.05.2005 cl.313/496 (H01j 1/62)
156. S.H. Moon, Field emission display device. Patent USA US 2005/019 4880 of
08.09.2005 cl.313/311 (Ho 0j 1/02)
157. M. Taniguchi, M. Kitada, K. Nakamura, S. Kawata, Field emission element. Patent USA US
7067971 of 27.06.2006 cl.313/497 (H01j 1/62)
158. A. Kastalsky, S. Shokhor, F.J. Disanto, D.A. Krusos, B. Gorfinkel, N. Abanshin, Pixel
structure for an edge-emitter field emission display. Patent USA US 2007/0046165 of
1.03.2007.cl.313/309 (H01j 1/02)
159. B.I. Gorfinkel, N.P. Abanshin, Cathodoluminescent screen and its manufacturing method.
Patent PF RU 2152662 of 10.07.2000. cl.H01j 1/62
160. N.P. Abanshin, B.J. Gorfinkel, Thin-film planar edge-emitter field emission flat panel
display. Patent USA 6 590320 of 08.07.2003 cl.313/309 (H01j 1/02)
161. L.D. Karpov, Patent USA №5 965971 of 12.10.1999, cl.313/309 (H01j 01/02), Pat USA №6
023126 of 08.02.2000, cl.313/310 (H01j 1/05)
162. K.L. Jensen, Field emitter arrays for plasma and microwave source applications. Phys.
Plasmas 6(5), 2241–2253 (1999)
163. S.P. Morev, N.P. Abanshin, B.J. Gorfinkel, A.N. Darmaev, D.A. Komarov, A.E. Makeev, A.
N. Yakunin, Electron-optical systems with planar field emission cathode matrices for
high-power microwave devices. J. Commun. Techn. Electron. 58(4), 357–365 (2013)
164. J.P. Calame, H.F. Gray, J.L. Shaw, Analysis and design of microwave amplifiers. J. Appl.
Phys. 73(3), 1485–1504 (1993)
165. A.G. Rozhnev, N.M. Ryskin, D.V. Sokolov, D.J. Trubetskov, S.T. Han, J.I. Kim, G.S. Park,
Novel concepts of vacuum microelectronic microwave devices with field emitter cathode
arrays. Phys. Plasmas 2(2), 4020–4027 (2002)
166. J.C. She, O.H. Ha, N.S. Xu, S.Z. Deng, S. Chen, S.E. Hug, L. Wang, Arrays of vacuum
microdiodes using uniform diamondlike carbon tip apexes. Appl. Phys. Lett. 89, 233518
(2006)
References 535
167. P.A. Ananiev, Y.V. Andrianov, V.A. Bashilov, M.M. Bogatskiy, D.A. Borisov, L.A.
Kontonistov, G.N. Fursey, HFF-amplifier. Patent USSR №1072145, of 07.02.84. cl.H01j
25/00
168. V.I. Mahov, B.V. Bondarenko, M.F. Kopylov, SVCH-devices M-type. Patent RF
№2040821 of 27.07.1995 cl.H01j 1/30
169. N.M. Ryskin, S.T. Han, K.H. Jang, G.S. Park, Theory of the microelectronic traveling wave
klystron amplifier with field emission cathode array. Phys. Plasmas, 14, 093106 (2007)
170. J.E. Graebner, S. Jin, G.P. Kochanski, W. Zhu, Microwave vacuum tube devices employing
electron sources comprising activated ultrafine diamonds. Patent USA, №5796211 of
18.08.1998. cl.315/3.5 (H01j 23/04)
171. J. Garcia-Garcia, F. Martin, R.E. Miles, D.P. Streenson, J.M. Chamberian, J.R. Fletcher, J.R.
Thorpe, Parametric analysis of micromachined reflex klystrons for operation at millimeter
and submillimeter wavelengths. J. Appl. Phys. 92(11), 6900–6904 (2002)
172. T.C. Cunningham, Field emission X-ray tube having a graphite fabric cathode. Patent USA
№3883760, 313/55 (H01J 35/06) of 07.04.71
173. P.R. Schwoebel, Field emission arrays for medical X-ray imaging. Appl. Phys. Lett. 8,
113902 (2006)
174. C.A. Spindt, C.E. Holland, R.D. Stowell, Recent progress in low-voltage field emission
cathode. J. Phys. t.45(Col.C9, suppl. 12), C9-9–C9-278 (1984)
175. B. Diop, V.T. Binh, Quasi-monochomatic field emission X-ray source. Rev. Sci. Instrum.
83, 094704 (2012). doi:10.1063/1.4752406
176. A.N. Bodrov, B.P. Merkulov, Y.V. Nikolyukin, Impul’snaya rentgenovskaya trubka, rat. RF
№ 2521436, ot 27.06.2014. kl.N01j 35/02
177. K. Kawakita, K. Hata, H. Sato, Y. Saito, Development of micro-focused X-ray source by
using carbon nanotubes field emitter. Technical Digest 18th IVNC, Oxford, UK (2005),
pp. 192–193
178. Z. Liu, O.Z. Zhou, S. Lu, Micro-field emission X-ray sources and related methods.
Patent US 2008/0043920, of 21.02.2008, cl.378/138 (H01J 35/14)
179. J. Zhang, G. Yang, Y. Cheng, B. Gao, Q. Qiu, Y.Z. Lee, J.P. Lu, O. Zhou, Stationary
scanning X-ray source based on carbon nanotube field emitters. Appl. Phys. Lett. 86,184104
(2005)
180. W. Knapp, D. Schleubner, S. Bjeoumikhov, H. Wolff, N. Langhoff, X-ray sources with
carbon field emitter cathodes, in IVESC (2004), pp. 309–311
181. S.H. Heo, H.J. Kim, J.M. Ha, S.O. Cho, A vacuum-sealed miniature X-ray tube based on
carbon nanotube field emitters. Nanoscale Res. Lett. 7, 258–265 (2012). doi:10.1186/1556-
276x-7-258
182. Y. Sun, The design and fabrication of carbon-nanotube-based Field emission X-ray cathode
with ballast resistor. JEEE Trans. Electr. Dev. 60(1), 464–470 (2012)
183. J.W. Hwang, C.B. Mo, H.K. Jung, S. Ryu, S.H. Hong, Field emission behavior of carbon
nanotube yarn for micro-resolution X-ray tube cathode. J. Nanosci. Nanotechnol. 13(11),
7386–7390 (2013)
184. Y. Jwai, K. Muramatsu, S. Tsuboi, A. Jyouzuka, T. Nakamura, Y. Onizuka, H. Mimura,
X-ray tube using a graphene flower cloth field emission cathode. J. Appl. Phys. Express 6,
105102 (2013)
185. J.-W. Jeong, J.-T. Kang, S. Choi, J.-W. Kim, S. Ahn, Y.H. Song, Digital miniature X-ray
tube a high-density triode carbon nanotube field emitter. Appl. Phys. Lett. 102, 023504
(2013)
186. G. Phillip, K. Timothy, W. Scotta, A sealing cold cathode X-ray tube for use in small X-ray
sources is provided. Patent USA 2014 0226791, от 14.08.2014. кл.Н01j-035/06
187. E.P. Sheshin, N.N. Chadaev, A.S. Baturin, A.I. Tryfanov, X-ray tube cathodes. Patent RF N
8248643,cl.Ho1J 35/02 of 35/02 of 20.03.05
188. A.S. Baturin, A.I. Trufanov, N.N. Chadaev, E.P. Sheshin, Field emission gun for X-ray
tubes. Technical Digest of Display and Vacuum Electronics, Garmisch-Partenkirchen
(2004), pp. 407–409
189. A.S. Baturin, A.I. Trufanov, N.N. Chadaev, E.P. Sheshin, Field emission gun for X-ray
tubes. Nuclear Instrum. Methods Phys. Res. A, 558, 253–255 (2006)
190. J.W. Jeong, Y.-H. Song, J. Kang, J.-W. Kim, Field emission X-ray tube and method of
operating the same. Patent US 2012/0148027, от 14.06.2012, кл.378/106 (Н05G 1/22)
191. A.S. Bugaev, P.A. Eroshkin, V.A. Romanko, E.P. Sheshin, Low-power X-ray tubes the
current state. Physics-Uspekhi 56(7), 691–703 (2013)
192. A.V. Crewe, Electron gun using a field emission source. Rev. Sci. Jnsx. 39(4), 576 (1968)
193. G.N. Fursey, S.A. Shahirova, To the question of possible localization of the bodily field
emission in small angles-ZhTF t.33(6), 1125 (1966)
194. S. Hosoki, S. Yamamoto, H. Todokoro, S. Kawasa, Y. Hirai, Field emission cathode and
method of fabricating the same. Patent USA N 4379250, cl. 313/336 of 05.04.1983
195. L.W. Swanson, L.C. Grouser, Angular confinement of field electron and ion emission.
J. Appl. Phys. 40(12), 4741 (1969)
196. Z.I. Dranova, V.A. Ksenofontov, V.B. Kylko, B.G. Lazarev, L.S. Lazarev, I.M.
Mihaylovskiy, A method of manufacturing the tip emission field emitter with localized
emission. USSR №828261 cl H01 J9/02 of 07.05.1981
197. V.A. Kyznetsov, B.N. Vasichev, Y.L. Rybakov, Field emitter with localized emission a.s.
USSR №1069029 cl.Ho1J 1/30, 23.01.1984
198. T.A. Bakeer, M.M. Balsiger, K.T. Considine, H.E. Litsjo, Electron-beam tube including a
thermonic field emission cathode for a scanning electron microscope. Patent №3809899
(USA)
199. B. Cho, K. Shigeru, C. Oshima, W(310) cold-field emission characteristics reflecting the
vacuum states of an extreme high vacuum electron gun. Rev. Sci. Instrum. 84, 013305
(2013)
200. E.A. Litvinov, A.A. Starobinec, The limiting currents of field emission. ZhTF 47(10), c.2032
(1977)
201. A.D. Karpenko, Y.S. Lobachev, V.A. Shishkin, On the limiting current of field emission
electronic cathode tip. Radio-electronics t.21(1), s.207 (1976)
202. M.I. Elinson, V.A. Gorkov, A.A. Yasnopolskaya, G.A. Kydinceva, A investigation of
pulsed field emission at high density of currents. Radio Eng. Electron. 5(8), 1318 (1960)
203. J.P. Barbour, W.W. Dolan, J.K. Trolan, E.E. Martin, W.P. Dake, Space-charge effects in
field emission
204. Jeol. J. Electron Microscope
205. http://www.hitachi-hitec.com
206. Patent N3191028 (USA), Scanning Electron Microscope/Crewe A.V
207. J.W. Butler, Digital computer techniques in electron microscopy, in 6th International
Congress Electron Microscopy (Kyoto) (1966), p. 191
208. M. Okai, K. Hidaka, M. Hayashibara, S. Watanabe, Field emission electron gun and electron
beam applied device using the same. Patent USA US 2008/0029700 of 07.02.2008, 250/310
(Ho1J 37/073)
209. D.I. Swan, D. Kynastou, The development of a field Emission SEM, in Proceedings of the
6th Annual Scanning Electron Symposium, N 1 (1973), p. 57
210. K. Kuroda, T. Suzuki, High current efficiency accelerating lens system of field emission
scanning electron microscope. J. Appl. Phys. 46, N1 (1975)
211. K. Kuroda, H. Ebisui, T. Suzuki, Three-anode accelerating lens system for the field emission
scanning electron microscope. J. Appl. Phys. 45(N5) (1974)
212. R.J. Taylor, P.J. Swan, An experimental scanning electron microscope. In: Proceedings of
the Scanning Electron Microscope System and Application Symposium (1973), p. 36
213. H.P. Kue, B.M. Siegel, A field emission probe forming system with magnetic pre-accelerator
lens. In: Proceedings of Electron Microscope Society of America, 34th Annual Meeting
(1976)
214. A.V. Crewe, Electron microscope using field emission source. Surf. Sci. 48(N1), 152–160
(1975)
References 537
215. R.I. Garber, A.K. Malik, I.M. Fishman, Use of needle emitters in electronic microscopy.
PTE 4, 188–190 (1974)
216. A.K. Malik, I.M. Fishman, S.M. Hazan, A.c.№389564 (USSR). Cathodic unit for electronic
gun. №29 (1973)
217. Pat.№1426509 (USSR), Improvements in field emission electron guns
218. Y. Sakitani, Patent №4019077 (USA). Field emission gun
219. T.A. Baker, M.M. Balsiger, K.T. Considine, H.E. Litsjo, Separable-chamber electron-beam
tube including means for puncturing. Pat.№3881125 (USA)
220. L.M. Welter, Field emission gun improvement. Pat №3946268 (USA)
221. V.A. Vasin, V.I. Zaporozhchenko, V.I. Rahovskiy, A.c. №594540 (SSSR). The device for
stabilization of current of an field emission source, №7 (1978)
222. S. Nomura, Patent №3786268 (USA). Electron gun device of field emission type
223. K.C.A. Smith, I.R.A. Cleaver, Electron gun. Patent №3872351 (USA)
224. T. Someya, M. Kobayshi, T. Goto, Feldemissions-Electronenguelle. Patent №2221138
(FRG)
225. R. Aihara, S. Ota, N. Kabayshi, Field emission type electron gun. Patent №3810025 (USA)
226. V.J. Coates, L.M. Welter, Patent №1355365 (England). Improvements in and relating to
electron gun
227. Siemens atteiengesellschaft Co., Electron-beam microanalyzer apparatus. Patent №1389119
(England)
228. R.V. Latham, D.A. Wilson, The development of s high-definition cathoderay tube using a
carbon-fibre field emission electron source. J. Phys. E: Sci. Instrum. 15(10), 1083–1092
(1982)
229. B.V. Bondarenko, V.A. Siliverstov, E.P. Sheshin, Use of field emission cathodes from
carbon fibers in an electronic gun. Theses of Reports of the 5th Symposium to Not Heating
Cathodes, Tomsk, pp. 81–83 (1985)
230. V.I. Tihtilo, E.P. Sheshin, Electrovacuum device. A.c. 1345935 SSSR, Cl.Ho1J 1/30 of
21.10.1985
231. F. Houdeller, L. Knoop, G. Gatel, A. Masseboeuf, S. Manishin, Y. Taniguchi, M. Delmas,
M. Monthioux, M.J. Hytch, E. Snoeck, Development of TEM and SEM high brightness
electron guns using cold-field emission from a carbon nanotip. Ultramicroscopy 151, 107–
115 (2015)
232. Y.A. Grigoriev, A.I. Petrosya, V.V. Penzyakov, V.G. Pimenov, V.I. Rogovin, V.I.
Shesterkin, V.P. Kudryashov, V.C. Semyonov, Experimental study of matrix carbon field
emission cathodes and computer aided design of electron guns for microwave power
devices, exploring these cathodes. J. Vac. Sci. Technol. B v 15(2), 503–506 (1997)
233. A.S. Baturin, I.N. Yeskin, A.I. Trufanov, N.N. Chadaev, E.P. Sheshin, R.G. Tchesov,
Electron gun with field emission cathode of carbon fiber bundle. J. Vac. Sci. Technol. B 21
(1), 1–4 (2003)
234. B.V. Bondarenko, V.I. Makyha, E.P. Sheshin, Y.V. Kydrashov, V.N. Ylacyuk, Y.V.
Petryshenko, Field emission electronic gun. A.c. 1294188 SSSR,Cl.H01J 1/30 of 11.03.85
235. A.G. Chakhovskoi, A.V. Galdetski, A.N. Obraztsov, A.P. Volkov, Experimental investi-
gation of electron gun wih nanostructured carbon cathode. Technical Digest of IVNC,
Oxford, UK (2005), pp. 204–205
236. N.N. Dzbanovskiy, P.V. Minakov, A.A. Pilevskiy, A.T. Rahimov, B.V. Seleznev, N.V.
Syetin, A.Y. Yuriev. High current electronic gun on the basis of an field emission cathode
and a diamond grid. Zh.T.F. t.75(10), 111–114 (2005)
237. V.S. Nazarov, A.G. Gryshnikova, M.B. Silaev, O.G. Varnakova et al., A.c. 490211 SSSR,
Cl. H01J 41/00 of 4.11.72. Ion getter pump/
238. A.c. 1240279 SSSR, Cl H01 J 41/16 of 12.07.84. Ion getter pump/ Varnakova O.G.,
Grinchenko V.T., Konovalov N.D., Pogibelskay N.B
239. A.V. Anashenko, A.S. Batyrin, I.N. Eskin, E.P. Sheshin, A.A. Shyka et al., Field emission
cathodes for heaters of getter-ion pumps. Electron. Ind. 3−4, 80–81 (1998)
240. V.A. Antonov, Y.A. Bykovskiy, A.I. Larkin, V.N. Ylasuk, A.V. Shelakov, E.P. Sheshin,
Light-valve equipment. a.s. SSSR №1609371 of 24.06.1988, Cl. H01J 31/24
241. Y. Sugiyama, J. Itoh, S. Kanemaru, Vacuum magnetic sensor with comb-shaped field emitter
arrays, in The 7th International Conference on Solid-State Sensors and Actuators,
Yokohama, Japan (1993)
242. D. Hong, M. Aslam, Diamond field emitter pressure sensor. Technical Digest of JVMC,
Oregon, USA (1995), pp. 335–337
243. T. Asano, R. Kajiwara, Fabrication of a tunnel sensor with cantilever structure. Technical
Digest of JVNC, Nagahama, Japan (1991), pp. 204–205
244. J.P. Hollingsworth, P.R. Banlaru, Carbon nanotube based nonvolatile memory. Appl. Phys.
Lett. 87, 2333115 (2005)
245. B. Lojek, Solid state field emission charge storage. Patent USA US 2008/0105946 of
2.05.2008.Cl.257/505 (H01L 29/00)
246. B. Lojek, Solid state field emission charge storage. Int. Pat. WO 2008/051675 of 02.05.2008
Cl.H01 L21/336
247. S.S. Furkay, D.V. Horak, C.H. Lam, H.-S.P. Wong, Field emission phase change diode
memory. Patent USA US 2005/0127350 of 16.06.2005, cl.257/4 (H01L 47/00)
248. C. Goodman, Neutron sources. Patent USA №2816242 of 10.12.57 cl.313/61
249. A. Otuka, M. Yoshino, C. Ohima, Application a Spindt emitter to an ionization gauge for an
extremely high vacuum. Tecnical Digest of IVMC, Nagahama, Japan (1991), 202–203
250. B.E. Barrington, A.L. Floras, W.L. Lees, Field-ionization electrodes. Patent USA, 3562881,
of 16.02.71, cl.29/25.18 (H01j 9/16)
251. M. Faubel, W. Hobler, J. Toennies, Field emission cathode. Patent England №2021854 of
5.12.79, cl.H01j 1/30
252. M. Faubel, W.M. Hobler, J.P. Toennies, Electron impact ion source with field emission
cathode. Patent USA №4272699 of 9.06.1981 cl.313/309 (H01j 37/073)
253. P.J. Traynor, R.G. Wright, Carbon nanotube electron ionization sources. Patent RST WO
2005/048290 of 26.05.2005, cl.H01j 49/00
254. O.A. Velikodnaya, V.A. Gurin, V.A. Ksenofontov, I.M. Mihaylovskiy, E.V. Sadanov et al.,
Multiemitter field source of ions on a basis the nanostructure of carbon materials. Letters
ZhTF t.33(13), c.90–94 (2007)
255. G.F. Byrchak, S.P. Chervonobrodov, E.P. Sheshin, Method of forming of a light stream on
the external screen for full-color system of display of a video information and the equipment
for its implementation. Patent RF.RU 2265964 of 05.08.2003, cl.H04 №9/30
256. R.Z. Bahtizin, S.S. Goc, R.F. Zaripov, R.R. Faizov, Noise generator. a.s. USSR №1157642
of 23.05.1985, cl.H03 B 29/00
257. Y.M. Wong, W.P. Kang, J.L. Davidson, V.K. Choi, D.V. Kerns, J.H. Huang, Design and
fabrication of single-chip carbon nanotubes vacuum field emission differential amplifier.
Technical Digest of IVNC, Oxford, UK (2005), pp. 35–36
258. A. Zoulkarneev, J.-H. Choi, Field emission RF amplifier. Patent USA US 2005/0184675 of
25.08.2005 cl.G3/10 (H01j 315/169.3)
259. B.J. Kampherbeek, M.J.-J. Wieland, P. Kruit, Emission photo cathode array for lithography
system and lithography system provided with such and array. Patent USA.US 2003/0178583
of 25.09.2003., cl.250/492.3 (H01j257/10)
260. D.A. Shiffler, Method of making a field emission cold cathode. Patent USA 2004/0202779
of 14.10.2004, cl.427/78 (B05D 5/12)
261. K.W. Cheng, Mirror having a field emission information display. Patent USA US
2008/0012727 of 17.01.2008, cl.340/933 (G08G 1/01)
262. F.G. Rudenauer, Field emission devices for space applications. Surf. Interface Anal. 39,
116–122 (2007)
263. V. Mironov, Fundamentals of the scanning probe microscopy. M. Technosphere (2004)
Conclusion (Prospective Development)
Abstract As previously noted, this book is at its core the corrected edition of a
Russian-language publication from the year 2011. However, although the chapters
1 to 8 of the translated edition are only partially updated and corrected, the con-
clusion chapter was written from scratch. This is mostly due to the fact that in last
few years appeared a significant number of original research papers on the book's
topic that requires a new analysis. Without reviewing them the authors would
consider their mission incomplete.
Given that field emission has indisputable advantages as compared with other free
electron sources, there’s little doubt that work for adapting different devices to
practical usage is to continue. This book lists a number of prospective fields of use
and branches of development for different field emission cathodes and devices based
on them, but below we’d like to mention again some of the most important of them.
Those fields are shown as a diagram on Fig. 1. The directions of development
can be subdivided into three basic categories: further development of traditional
technologies, introduction of new materials and nanotechnology. Different types of
treatment of field cathodes, increasing the homogeneity of electron current and
focusing of electron beam can be used for each of those categories. Naturally, all
these methods are aimed at practical use of field cathodes in various electronic
devices. Qualitative leap in development of field cathodes can be achieved by both
developing new methods and further development of traditional technologies at the
new level of understanding of physical processes.
Development of Traditional Technologies
In this section it is first necessary to note the invention of conical molybdenum

Spindt cathodes. Development of lithographic methods allowed to increase the
density of cathode tips up to 108 per cm2 [1]. In this case the distance between the
tips is about 1 lm (the diameter of control electrode is 0.35 lm). In this

in Advanced Microelectronics 60, DOI 10.1007/978-3-319-56561-3
540 Conclusion (Prospective Development)
Prospects of development of
field emission cathodes
Traditional New materials Nanotechnology

technology
The focusing of Increase the Processing field

the electron beam uniformity of emission
flow of electrons cathodes
Practical applications
of field emission cathodes
Fig. 1 The most important directions of development of field emission cathodes
configuration an ordinary field cathode with diameter of 1 mm contains approxi-

mately 785,000 tips. With median current carrying capacity of 1.5 lA per tip one
can obtain the current (in impulse) of 1 A with density up to 130 A/cm2 which is
more than enough for most electro-vacuum devices.
Taking into account the already well-developed technology of such field cath-
odes (SRJ International, Futaba) where they are used in microwave devices and flat
display screens, they have also a good perspective of taking over the niche of
special devices in which one can obtain a vacuum higher than 10−9 mm Hg.
Such studies are also performed with tip field cathodes out of glassy carbon [2].
As far as carbon material is concerned, there’s a proven influence of adsorbed
residual gases onto the field emission current [3].
Studying of the influence of residual gas adsorption onto field electron emission
out of carbon materials was performed in this research for the case of polyacry-
lonitrile carbon fibers with diameter of about 7 lm, that were annealed beforehand
at the temperatures of 900, 1500, 2000 or 2600 °C. The samples were prepared
using the standard methods. The studies were performed in both unsoldered glass
device and metal disassemble-able field electron microscope. Before beginning the
research, the samples were molded. The maximum molding current was up to
120 lA. For learning the influence of adsorption on field electron characteristics it
is necessary that the cathode surface structure is kept unaltered. For this the samples
were kept at current of 50 lA (i.e. with less voltage) for 10–20 h, which guaranteed
the stability of emission field cathodes. For the maximum decrease of ponderomotor
Conclusion (Prospective Development) 541
stress and ion bombardment intensity (i.e. decrease of probability of alteration of

field cathode work surface structure) the work voltage was further reduced. The
field emission current was kept at the level of 10–30 lA.
The test were performed in cyclic mode, including several sequential measure-
ment cycles. Each such cycle consisted of two parts: work stage and adsorption
stage. Let’s note that those names are just formal, as the processes of adsorption-
desorption are continuous. All the measurements took place in work stages when
there was a high voltage at the anode. At adsorption stages the voltage was removed.
At the initialization of work a pulse of high voltage was sent to the anode, notably its
size was constant for all measurement cycles for each given sample (field electron
cathode). The transitional characteristics (i.e. changes in the field emission current
with time during 1–6 h) were measured directly (s.c. response to voltage “step”).
As a rule, the transitional characteristics look like descending curves tending to a
horizontal segment whenever the transitional processes end (Fig. 2). Those char-
acteristics were qualitatively matching for polyacrylonitrile carbon fibers with
different temperatures of preliminary annealing.
The observed increase of field emission current after the adsorption steps can be
in principle explained by the following factors: changes in electron work function
or transparency of potential barrier; changed in surface geometry; or change of
emitting surface area.
Cyclical changes in field cathode geometry are improbable; analysis of field
electron microscope images (emission patterns) shows that the surface geometry
does not change in any notable way for the duration of the experiment. Thus, as
before, the most attention must be paid to the first and the third reasons. We must
note however that the field emission images at the beginning of a work stage is still
Fig. 2 Transitional characteristics of polyacrylonitrile carbon fiber with a temperature of pre-

liminary annealing of 2600 °C. The periods of adsorption aren’t shown. Hold time: a 12 h, b 24 h;
Pressure: a 5 10−7 Torr, b 10−2 Torr
different from ones obtained at the end of preceding work stage: the image
brightness increases and the previous image seems more “grainy” i.e. its details
look larger. The latter can perhaps be explained by emission from different areas of
the surface of polyacrylonitrile fiber.
According to the model of adsorption-desorption processes described above, the
first and the third cases cannot be considered as separate, as changes of barrier
transparency due to adsorption phenomena occur not so much at the microtips
themselves as in pores, holes, capillary etc, i.e. those areas where there was possibly
no emission at all before. Thus, along with reduction of the potential barrier,
increase of the emitting area also occurs at this areas.
Figure 2 shows the transitional characteristics of a sample annealed at temper-
ature of 2600 °C for different adsorption times. Analysis of these characteristics
leads us to conclusion that the dynamic equilibrium between adsorption and des-
orption in absence of ion bombardment of the cathode (and thus with filling of
capillary) is achieved in about 10 h.
From Fig. 2 one can conclude that a rather large decrease of field emission
current occurs in first minutes or even seconds of cathode function. This can
probably be explained with that that intensive ion bombardment rips off atoms and
molecules that are adsorbed close to emission centers (microtips) and thus have
small adsorption energy. As a result, close to the tips the work function grows and
the current decreases abruptly. Further removal of the adsorbed atom layer goes
significantly slower. This is caused, firstly, by a stronger bond of adsorbed atoms
and molecules of the residual gas at the surface of polyacrylonitrile fiber and,
secondly, the fact that the adsorbed molecules come to the surface out of capillary
and replace ones removed with ion bombardment. Speed if this process is defined
by diffusion speed and thus has the same order of magnitude. Those processes go
on until the dynamic equilibrium between adsorption and desorption is achieved.
Additionally, the analysis of electron characteristics shows weak dependence of
desorption speed on the intensity of ion bombardment of the field electron surface.
Those dependencies are common for all carbon materials. They lead to a
practical conclusion: field cathodes out of carbon materials produce larger current in
impulse mode. Additionally this regime is characterized with increased stability of
field emission current and longevity of the cathode.
As noted above, the increase of overall level of current taken from field electron
cathodes can be achieved by increasing the total emitting surface of cathodes in
case they work in parallel, i.e. with cumulative impact of a large number of separate
emitters. Nevertheless simple increase of emitters quantity does not proportionally
increase of emission current due to large variance of emitter curves in terms if
height and radius as well as because of mutual screening of different emitters.
Increase of working voltage leads to increase of current out of several particular and
the sharpest emitters, and eventually to their evaporation or explosion. This in its
turn leads to emergence of an ark between the anode and cathode that spoils the
vacuum in the interelectrode gap, which leads to further ark etc.
Thus field cathode systems use limiting resistance. Those can come in form of
external resistors or in form of constructive layers with relatively high resistance.
This resistor serves as negative feedback, which stabilized the field emission
current and doesn’t allow disruptions.
With all this in mind we can conclude the optimal construction of field cathodes
is one that involves mutually insulated fragments each with individual limiting
resistors. This technique allows, among else, to straighten the currents of individual
fragments. Finally, there is a way that can be employed already at the current stage.
It involves separation of a large field cathodes into small insulated fragments,
serially to each of which a transistor is added to play a role of controlled limiting
resistor. In this case one can achieve near-perfect homogeneity of emission current
along the field cathode surface. The same transistors can be used to control the
emission current.
New Materials
One of natural of development of field emission technologies is creation and syn-

thesis of new materials and structures with better emission properties. As there’s
only so much pure elements and out of them the time of refractory metals, carbides
and similar substances is mostly behind us, new breakthroughs are to be expected of
different composites and new structures.
For example, let’s consider the carbon nanocomposite material described in
[4, 5]. Its main elements are diamond powder and pyrographite. Such an element
contains particles of diamond surrounded with sp2 conductive carbon matrix. Most
often the diamond powder used had particle sizes of around 5 nm. The powder was
pressed into discs 20 mm large and 1 mm thick. Filling the pores between diamond
particles with pyrographite is performed via a special CVD process. In the same
time, it is rather easy to control the ratio between sp2 and sp3-components and
thickness of the pyrolythic layer (Fig. 3). In experiment [4] the sp2/sp3 ratio varied
between 0 and 0.5 and thickness of pyrographite changed between 0 and 1 nm.
Conductivity of the sample varied between 107 and 0.1 S/m2.
Besides, presence of two phases in the material, i.e. diamond and pyrographite,
considerably decreases the electron work function by adding curvature to the zones
adjacent to thin pyrographite film. Based on this composite there are proposed
constructions of field cathodes, including porous ones [6] and ones with repousse
surfaces of given shape [7].
Fig. 3 The schematic

structure of a composite
pyrographite diamond
Fig. 4 Structure of the composite field emission cathode on the basis of carbon nanotubes: 1 glass
substrate; 2 silver paste; 3 carbon nanotubes
Interesting approaches to creation of carbon-carbon composites were proposed

in [8, 9].
Rakhi et al. [10] proposes a composite material based on multi-wall carbon
nanotubes covered with thin layer of metals. Pt, Pd and Ru are delivered onto clean
nanotubes in form of nanocrystallites by the method of simple chemical reduction
from corresponding salts. Thus prepared billet are then printed onto graphite fabric.
Testing of these field cathodes in diode mode demonstrated relatively high stability
of emission current for 4 h in vacuum of approximately 10−6 mm Hg. A similar
approach is presented in [11].
Figure 4 glass substrate (1) is covered with a thin layer of silver paste (2) in
which the carbon nanotubes (3) acting as electron emitters are inserted. In order to
increase the construction properties the work surface of field cathode is polished
with precision of approximately 0.1 lm.
The other notable achievement in new structures of carbon materials one can
also name the creation of nanoporous carbon [12]. It is obtained by inserting carbon
precursor into quartz porous substrate. After that, carbon is carbonized and the
substrate is removed. As a result, one obtains porous carbon with pores several
nanometers large. Its unit surface area is of the order of 1800 m2/g.
Lobanov et al. [13] proposes yet another new class of carbon materials based on
intercalation of pyrographite microplates with triple carbonate (Ba, Sr, Ca)CO3
which considerably decreases the work function of cathode material.
Nanotechnology
In the last few years the problems of low-field electron emission out of nanos-
tructured materials attract growing attention of researchers all over the world.
Low-field emission occurs out of nano-scale conductive structures (s.c. nano-
phase of a substance) surrounded with insulating phase or vacuum. High emission
properties of such nano-objects are defined not only by geometric factors of electric
field growth but also by lowered potential barrier for electron tunneling into vac-
uum out of those areas. The latter is connected to one of the fundamental properties
of a solid body, namely the difference of physical characteristics of nanoparticles
(electric, magnetic, optical, acoustic etc) from “macroscopic” properties of the
substance.
When quantum-scale effects start to play a dominating role the energy levels of
electrons of the substance become discrete and emitted electrons have rather narrow
energy spectrum. By estimation this usually occurs when nano-object has size
5 nm. When distance between nanoemitters is comparable to the length of
coherency mutual influence of nanoemitters and interference of emitted electron
beams should manifest themselves (at distance of 10 nm or matrix density of
approximately 1012 elements/cm2). Those are so-called quantum points [14].
It is important to note that “self-arranging” carbon nanostructures (e.g.
nanocrystal diamond films) that are to a greater extent non-ordered systems for-
mation of nano-objects on any particular area of surface depends on a number of
uncontrollable factors. This means that creation of electron nanoemitters with given
parameters in this case is all but impossible.
It seems advisable and perspective to solve the problem of creation of
nanoemitters that can serve as monoenergetic electron sources and matrix of
nanoemitters based on ordered carbon nanostructures using a new approach
including directed formation of carbon nano-objects with different given sizes,
compositions, configurations and locations on the substrate.
In current time one can name two technological techniques that can implement
such nanostructures. The first of them is forming nanostructures using nano-scale
modification (i.e. nanophase transformations) of carbon materials. Works in this
field are performed all over the world. In terms of several criteria (composition,
sizes of nano-objects, electric properties) the following types of nanostructures with
pre-defined topological structures are the closest to nanoemitters:
• carbon nanostructures on substrate surface
• nanostructures in diamond and diamond-like films
• nanostructures in polymer films
One of the principal problems of this approach is guarantying stable and re-
producible properties of the created nanostructures. From the point of creation of
nanoemitters, this problem can by formulated based on the following criterif: the
formed body needs to be reliably chemically bonded with the substrate, it must be
conductive and be able to increase its mass by forming new chemical bonds.
The second technological approach is based on growing carbon nanotubes
(CNT) on the substrate surface [15, 16]. Formation of CNT (single or multi-wall)
arrays is an important problem studied in many of leading research centers and
private companies. There are test samples of flat screen displays based on CNT
arrays, in which the elements of cathode matrix include CNTs positioned perpen-
dicular to the substrate surface. The maximum achieved matrix density is about 106
elements per cm2. The main problem with creation of CNT-based nanoemitters is in
creating metals catalytic platforms (Ni, Fe, Co) with sizes of several nanometers for
growth of singular quantum-scale objects i.e. single-wall CNTs on substrate
surface.
Scanning probe nanolithography (SPL) seems to be the most efficient method of
forming a new type of field electron cathodes. SPL is one of the principal fields of
modern nanotechnology methods and is beyond any competition for the task of
creating structures with topological elements of the order of 10 nm and less. It is

with this order of element sizes quantum-scale effects appear in nanostructures.
Such small elements of electron systems as quantum cable or quantum point are
directly related to the given nanostructure type. Modern scanning probe micro-
scope-lithographs provide in situ complex control over formed nanostructures. Thus
the developed approach to creation of a new type of nano-scale electron sources is
based on fundamental principles and real abilities of SPL.
The main nanotechnological technique for forming nanoemitters by nano-scale
carbon modification is local electric influence of the probe onto the initial material.
Analysis of modern achievements in SPL as well as personal experience leads us to
a conclusion that this technique is the most effective for nanophase transformations
of carbon compounds. For an instance, let’s consider the method of growing
nanoemitters with help of scanning probe lithography [17, 18]. In this case it was
performed with Solver R 47 (NT-MDT, Russia) that used platinum-covered can-
tilever. Film of (a-C:H):Si with thickness of 100 nm is transfered onto Si-substrate
by method of CVD.
In the initial step of determination, the surface was scanned in contact mode for
registering the initial structure. Then, a series of voltage impulses of 10 V was sent
between grounded cantilever and the substrate. In the same time the surface
structure and conductivity in the point of contact were tightly controlled.
Immediately after the impulses, a surface topography scan was taken in order to
detect any changes in surface structure.
After this the samples were subjected to vacuum annealing at pressure of about
100 Pa and temperature of 100–700 °C for the duration of one hour. The tip shapes
after annealing at different temperatures are presented on Fig. 5. As one can see
from the figure the conic shapes don't change wafter annealing even at temperatures
as high as 700 °C. In the same time, the shape of tip pedestal changes considerably.
With annealing temperature of 400 °C the pedestal almost completely disappears,
and it 700 °C it becomes concave, i.e. after being subject to temperature the
obtained tip is lowered relative to surrounding field (Fig. 6).
Fig. 5 Section of the

modified area to (1) and after
training by heating to 400 °C
(2) and 700 °C (3). The final
form of an tip is given in an
insert
Fig. 6 Model of the modified

area in initial (a) in final states
(b): 1 pedestal (thermally
unstable area); 2 nanocone;
3 film; 4 substrate. Note Such
approach has big future as
provides ability to integrate
field emission cathodes with
elements of electronic chips
Research of field emission fro graphenes warrant particular interest as the most
thin film field cathode. Despite that this area of research is relatively new [19, 20],
the volume of study is steadily increasing and one can reasonably hope that it would
lead to a new class of field cathodes.
Treatment of Field Emission Cathodes
Almost in every chapter of this book while describing field cathodes we mentioned
various methods of treatment of cathodes after they are already produced. This is
particularly applicable to cathodes out of carbon materials. Methods of treatment
include bombardment with ions of residual gases, plasma, thermal treatment etc.
Paper [21] proposed a method of cathodes by method of radiation.
The cathodes were prepared out of flat billets made of highly resilient fine-grain
graphite MPG-6. There were billets with round and rectangular cross-section with
diameters and sides, respectively, of 1 cm and thickness of 3 mm. After being cut
the billets were positioned onto the target of ion injector. Different samples were
bombarded with ions of Ar+, Xe+ and La+ in technical vacuum (about 10–4 Torr).
Other varied values included ion energies (50, 100 or 200 KeV), radiant exposures
(1016, 1017, 1018 or 1019 ions/cm2 with roughly constant radiation intensity at about
1014 ions/cm2 s) and angles of impact of the beam (angles between the beam axis
and billet surface were 90, 65 or 45°).
As shown by profiling of the manufactured samples, the most developed surface
containing the most potential emission centers is formed at radiant exposures of
about 1017–1018 ions/cm2 and angle of impact of ion beam of 45° to the surface.
Resulting density of emission centers was over 108 cm−1 and their curvature varied
between 10 and 100 nm. There is another notable fact: one can expect that this
method of field cathode production one can pick a mode of radiation that would
reduce height variance of emission centers to no more that 100–1000 nm. This
would allow move the cathode much closer to the fluorescent surface thus con-
siderably reducing working voltage, up to several hundred Volts and less.
Let us also note that in order to assess emission parameters all the produced
cathodes were installed into the vacuum system regardless of CTM image of their
surface. In process of measuring the volt-ampere characteristics of the cathodes
(with cathode-anode distance kept unchanged at about 5 mm) the previous con-
clusion on optimal parameters of radiation was confirmed. It were those cathodes
that allowed to obtain constant electron emission at voltages of 1–10 kV, while the
maximum current achieved the values of several mA. There is also a notable
dependence of cathode longevity on voltages at which they are used. Starting at
certain threshold voltage (usually around 5 kV) cathode lifetime dropped abruptly
from hundreds of hours to several minutes. A logical explanation for this phe-
nomenon is destruction of cathode emission centers due to high outbound current
(and thus heating of the cathode) and intensive ion bombardment of its surface.
Finally, measurements of cathode longevity and stability of emitted electron
current demonstrated that method of bombardment with heavy gas ions shows more
promise for field cathode production than bombardment with ions of metals. This
may be connected to the fact that in the former case we are dealing with emission of
pure carbon materials while in the latter one the carbon (graphite) is partially
metalized and implanted with metals.
This example shows that treatment of field cathodes has an important and often
critical role in defining the functional properties of the whole electronic device.
Thus, development of principles and modes of cathode treatment is a very
promising field.
Increasing the homogeneity of electron currents
Increase of the strength and homogeneity of electron current can be achieved by

inserting electrodes with high secondary emission coefficients into the cathode-
anode gap. This idea emerged long time ago, in times of discovery of microcanal
plates. But their use did not guarantee necessary parameters of electron beam,
because emitted field electrons have high energies and microcanals in the plates
have low angle. As such, efficiency of secondary emission was too low for practical
purposes.
In the last years several new and interesting constructive solutions to this
problem were proposed.
Authors of [22, 23] propose to make secondary emission elements out of glass
plates with slanted holes insides of which were covered with highly secondary
emissive materials. Increasing slant angle to 37.2° increased the secondary electron
output up to 98%. Such slanted holes can be created by etching.
Fig. 7 System of the

secondary emission amplifier
with two plates with inclined
openings: 1 cathodic plate;
2 ITO-on a cathodic plate;
3 field emission cathode
from carbon nanotubes;
4 insulating layers; 5 inclined
openings; 6 spaser; 7 layer
of a phosphor; 8 ITO-on an
anode plate; 9 anode plate;
10 control electrodes
For further increase of characteristics use of two such place with opposite ori-
entation of holes was propose (Fig. 7). In this case the controlling field may be
reduced to 1.5 V/lm. Besides, energy homogeneity of electron current is also
increased which leads to a considerably higher homogeneity of electron current
along the field cathode surface.
Development of this approach led to invention of a system of secondary-electron
electrodes for a flat field emission cathode out of carbon nanotubes [24, 25].
Construction of an element of a flat display is presented on Fig. 8. It’s based on a
funnel-like hole in a glass plate (4) acting like a secondary emission cathode. The
second insulation layer (7) serves for focusing the electron beam. Electronic sys-
tems similar to electron multipliers are proposed in [26, 27].
Fig. 8 The scheme of pixel

with secondary emission
amplifiers in a appearance
funnel-shaped opening:
1 substrate; 2 field emission
cathode; 3 extending
electrode; 4 funneled-shaped
opening; 5 insulator; 6 control
electrode; 7 exhaust outlet;
8 insulator; 9 phosphor;
10 anode
Fig. 9 The scheme of the display with metal secondary emission an amplifier: 1 resistive layer of
the field emission cathode; 2 substrate; 3 cathodic conclusion; 4 insulator; 5 control electrodes;
6 metal plates with big coefficient of secondary emission; 7 insulator; 8 masking layer; 9 layer
ITO; 10 anode plate; 11 phosphor
Figure 9 shows a schematic for one such device. The main distinguishing feature
is use of several electrons made on ferro-nickel or silver-magnesium alloys that are
known to have a high secondary emission coefficient [27]. The other options for the
coating materials include beryllium oxide, strontium oxide, barium oxide and other
metal oxides having high secondary emission coefficients [25]. The other substance
with high secondary emission is diamond [28].
Using this effect it is also possible to develop highly efficient microelectronic
vacuum devices.
Such devices can be based on polycrystalline diamond membranes able to
increase the electron beam of given distribution in a plane perpendicular to the
direction of electron spreading.
One of possible methods of increase of secondary emission is based on use of
porous membranes. In this case the secondary electrons passing through the surface
as in “reflection” scheme pass through the opposite surface through the pores due to
influence of the external field. Given that energies of secondary electrons is small
they can be “pulled in” into membrane pores. Of course, the multiplicative coef-
ficient Kn is considerably smaller that in “reflection” scheme as certain part of
Fig. 10 The image in a

scanning electronic micro-
scope of a porous diamond
film
primary electrons passes through the pores without taking part in birthing of sec-
ondary electrons. In this case Kn = (1 − Sn/S)KD, where Sn and S are total area
covered by pores and total work area of the membrane, respectively; KD is the
coefficient of secondary emission in “reflection” mode.
In order to test this idea experimentally we assessed energies of the secondary
electrons and created porous membranes. Direct measurements show that the
energy of secondary electrons does not exceed 3–5 eV.
We’ve prepared porous membranes with pores of 8 lm and distance between
them of also 8 lm (Fig. 10).
The proposed construction has quite serious practical applications. It all but
negates the dependence of amplification coefficient on membrane thickness and
allows to work in “shoot-through” mode with small energies of primary electrons.
Focusing of the Electron Beam
The problem of focusing for field electrons is still not completely solved due to
great initial energies of electrons and thus dispersion in terms of speeds and
directions.
Reduction of the electron speeds e.g. via transforming them into secondary
emission ones allows to significantly reduce their energies and thus helps to more
efficiently focus them.
This book lists several rather promising constructions of focusing systems.
In our opinion the most efficient system for field electron focusing is reducing
their energies to the order of kT i.e. to the thermoelectron levels. When this is
achieved, it is possible to use already well-developed focusing systems.
However, the technologies of focusing field electrons with their initial energies
are by no means irrelevant or redundant. For an instance, [29] proposes a flat
system of turning off the field electron beam and [30] presents a field electron
display scanning the beam. In the latter case the focusing system consists of three
lenses.
Current Stage of Field Electron Emission Science
The current state of affairs in the science of field electron emission can be char-
acterized as a some transition stage, or according to Forbes, reconstruction phase
[31–33]. Under this representation, the development of fundamental theory of field
electron emission can be viewed as consisting of several phases. Main theoretical
breakthroughs happened in the Fowler–Nordheim phase (1928 to mid-1950s), when
the theory of electron tunneling was initially formulated, and Murphy-Hood phase
(mid-1950s to mid-1990s), when a Fowler–Nordheim-type expression based on
Schottky-Nordheim barrier theory was introduced, which in turn led to develop-
ment of theory of Schottky emission (i.e. emission from the barrier top; later Jensen
[34] would build upon this idea and develop a general theory of field and photo-
electron emission) [33]. The current stage mostly consists of attempts at system-
atization and collating of data to make the mainstream science clearer and to
develop effective theory of specialized field electron emission topics.
This is evidenced by the character of activities of the international associations
and conferences on the topic. The International Field Emission Society holds
regular symposiums (IFES) since 1952. The main focus of attention has since
shifted from electron field emission towards field-ion microscopy and atom-probe
tomography, as well as their various applications. The Society and the symposiums
have always been in a sense practice-oriented. That is exactly why the symposiums
have demonstrated a wide interest in fundamental science of both electron and ion
field emission, surface effects of strong electric field, as well as in development of
new methods and apparatuses based on those effects, understanding of their prin-
ciples of function and their practical use. In the previous years proceedings of the
symposiums included a lot of papers on new sources of field electron emission and
related scientific fundamentals. Other consistent topics of discussions within the
Society included liquid-metal ion sources and many aspects of strong surface field
theory.
Lately the proceedings of the symposium (which, after 2002, takes place bien-
nially) are showing a noticeable decline in the amount of papers on electron field
emission: usage of keywords and references to sources having to do with field
electron emission in the papers of the last few years [35–37] is down about 85%
relative to the previous average. In 2014 the event was altogether renamed into the
Conference on Atom Probe Tomography and Microscopy [37] and its proceedings
contained hardly any mention of field electron emission at all. One can say that the
fundamental theory of electron field emission is waiting for a new theoretical
breakthrough.
Nevertheless, in spite of the absence of recent fundamental discoveries, the

development of systems based on electron field emission continues to advance.
Such advances are inextricably intertwined with vacuum technology, as was thor-
oughly explained in the previous chapters, as well as with other branches of
technology, including material science, micro- and nanotechnology and electronics.
Accordingly, the literature on the development of electron field emission systems
theory is strongly linked with topics of vacuum technology, as can be seen in the
proceedings of regular international conferences on connected topics: on vacuum
electronics (IVEC takes place annually), on vacuum microelectronics (IVMC took
place annually till 2003), on vacuum nanoelectronics (IVNC takes place annually)
and on vacuum electron sources (IVESC takes place biennially).
The mentioned bibliometric indicator (usage of keywords and references to
sources having to do with field electron emission) significantly fluctuates in the
proceedings of IVEC from year to year. The highest value in the whole 21st century
was measured in 2009 [38]. In later conferences this value dropped, sometimes
quite significantly (roughly 50–75%) [39–42], sometimes mildly (20–40%)
[43–45], as in the case of IVESC (16–42%) [46–48]. In any case, it would be
incorrect to use this data to conjecture a halt or decline in practical use of field
electron emission. What is evident, however, is lack of significant theoretical
developments in this field.
Analysis of IVNC proceedings of the later years shows interesting results. This
yearly conference followed the aforementioned trends in modern vacuum micro-
electronics—miniaturization of both particular components of field electron emis-
sion devices and the devices themselves—which led to steady increase of the role
nanotechnology plays in the field. As such, since 2004 the event is called an
international conference on vacuum nano- (rather then micro-) electronics. The
greatest amount of papers in the conference proceedings was observed in 2006 [49],
when IVNC was joined with IFES. However, it wasn't then when the greatest
number of mentions of field electron emission and related references was recorded.
Strange as it may seem, it was referenced much more in later years [50–56]. It is
nowadays that this value increased by about 50–60% compared to the average for
this conference, even though the total number of papers per conference went down
by 30%. This is evidence enough that this area of knowledge is concentrated around
field electron emission in anticipation of serious breakthroughs in this field.
Characteristics of nano-scale field electron emitters have certain peculiarities and
demonstrate new and exceptional properties [57–61], which open up new ways to
improve field electron emission systems based on nanomaterials. These properties
are due to a number of fundamental causes, which, obviously, should be taken into
account when developing a theory of field electron emission: three-dimensional
potential barrier on phase boundaries, changes in interactions between the electron
and charge, quantum confinement [62] etc. as it was written in the previous
chapters.
Field electron emission science needs a coherent intellectual structure that
accommodates basic theory, practical theory, experiment and technology devel-
opment [33]. Amongst other things, this should provide proper physical
understanding of relevant effects; clearer and consistent concepts, terminology and

notation; physical models that are justified by experiment, and have their limitations
known; formulae that are established by transparent mathematical proofs and
(where appropriate) by reference to experiment, with their clearer limitations;
appropriate theory for the interpretation of experiments (where necessary), appro-
priate applied theory that can assist in the development of field emission based
devices [63–71].
In spite of a few systematizing publications that appeared in the latest years
[72, 73], we have to admit that a theory still remains to be developed with better
scientific basis which would mark the end of the reconstruction phase of field
electron emission science.
References
1. C.A. Spindt, Field Emitter Arrays for High-Current, High-Current Density, and High
Frequency Operation (Tech. Digest JVNC, Oxford, 2005), pp. 20–21
2. V.J. Shesterkin, Y.A. Grigoriev, The multiple Beam electron gun with cold field emission
cathode made from glassy carbon, in Proceedings of JVESC-JCEE (2014), p 250
3. A.L. Suvorov, E.P. Sheshin, V.P. Babaev, Influence of adsorption of residual gases on
emission properties carbon properties. J. Technol. Phys. 66(9), 164–169 (1996)
4. A.V. Karabutov, V.D. Frolov, V.J. Konov, V.G. Ralchenko, S.K. Gordeev, P.J. Belobrov,
Low-field electron emission of diamond/pyrocarbon composites. J. Vac. Sci. Tech. B. 19(3),
965–970 (2001)
5. A.V. Karabutov, V.I. Konov, V.G. Ralchenko, V.D. Frolov, S.K. Gordeev, Advanced
diamond/sp2-bonded carbon nanocomposite materials as low threshold field electron
emitters. New Diam. Front. Carbon Technol. 11(4), 265–275 (2001)
6. S.K. Gordeev, A.I. Kosarev, A.N. Andronov, A.Y. Vinogradov, Field emitter of electrons.
Patent R.F. Ru 2149477 of 20.05.2000, cl.Ho1j 1/02, 1/30
7. S.K. Gordeev, V.G. Ralchenko, M.A. Negodaev, A.V. Karabytov, P.I. Belobrov, Field
emitter of electrons and way of its production. Patent RF Ru 2150154, of 27.05.2009, cl.
Ho1j 1/02
8. G.Z. Otarashvili, A.S. Rautov, E.P. Sheshin, Diamond-carbon nanocomposites as material
for cold cathode planar-type emitters. Technology Digest of JVNC (2011), p. 156
9. E.P. Sheshin, A.L. Shornikova, The field emission properties of different carbon-contained
powders, in Proceedings of JVESC-JCEE (2014), p. 252
10. R.B. Rakhi, A.I. Reddy, M.M. Shaijumon, K. Sethupathi, S. Ramaprabhu, Electron field
emitters based on multiwaled carbon nanotubes decorated with nanoscale metal clusters.
J. Nanopart Res. 10, 179–189 (2008)
11. K.-W. Cheng, G.-Y. Hsiao, J.-T. Hsu, Methode of fabricating cathode structure of field
emission display. Patent USA US 2006/0009111 of 12.01.2006, cl.445/51 (H01j 9/04)
12. R. Ryoo, S.H. Joo, M. Krur, M. Jaroniec, Ordered mesoporous carbons. Adv. Mater. 13(9),
677–681 (2001)
13. S.V. Lobanov, J.A. Fedorov, E.P. Sheshin, Researching dependencies of thermonic and field
emission characteristics of intercalated pyrographite cathodes from ingredients composition
and pressing conditions, in Proceedings of XXII conference Vacuum Science and Technology
(2015), pp. 208–210
14. R.-Z. Wang, W. Zhao, H. Wang, B. Wang, Y. Hui, Field emission from AlGaN/GaN
quantum-well nanotip arrays. Technology digest of JVNC (2011), pp. 100–101
15. P. Serbun, S. Putkowski, A. Navitski, N.T. Hong, S. Lee, Field emission from carbon
nanotube (CNT) arrays and triode test of single CNT columns. Technology digest of JVNC
(2011), pp. 38–39
16. F.-G. Zeng, H. Ma, V. Ru, H. Huo, F. Xiang, G. Wang, A configuration of substrate surface
for CNT cold cathode, in Proceedings of JVESC-JCEE (2014), p. 78
17. V.D. Frolov, V.L. Konov, S.M. Pimenov, V.J. Kuzkin, The low-dimensional effect in single
carbon-based nanoemitters of electrons. Appl. Phys. A 78, 21–23 (2004)
18. V.D. Frolov, E.V. Zavedeev, S.M. Pimenov, V.I. Konov, E.N. Loubnin, G.G. Kirpilenko,
Nanocones on (a-C:H): Si composite film: termal stability, growth dynamics and electrical
properties. Diam. Relat. Mater. 16, 1218–1221 (2007)
19. H. Tsurumi, T. Higuchi, Y. Yamada, M. Sasaki, Field emission from dense graphene edges.
Technology digest of JVNC (2011), pp. 226–227
20. Y. Zeng, Z. Chen, Z. Xie, J. Zhang, H. Li, Stable field emission from carbon nanotube grown
on grapheme substrate, in Proceedings of JVESC-JCEE (2014)
21. A.L. Suvorov, E.P. Sheshin, V.V. Protasenko, N.E. Lazarev, A.F. Bykov, V.P. Babaev, The
microrough flat field emission cathodes from graphite received in the radiation method.
ZhTF. T. 66(7), 156–160 (1996)
22. W. Lei, X. Zhang, X. Zhang, X. Zhou, Z. Zhu, C. Lou, Characteristic of the cold cathode
with secondary electron emission. Technology-digest IVESC (2004), pp. 52–54
23. W. Lei, X. Zhang, X. Zhou, Z. Zhu, C. Lou, H. Zhao, Characteristics of a cold cathode
electron source combined with secondary electron emission in a FED. Appl. Surf. Sci. 251,
170–176
24. X. Zhang, W. Lei, M. Liu, L. Zhang, D. Engelsen, X. Zhou, Q. Wang, Analysis of the
transverse energy distribution of HOPFED. Technology digest JVESC (2004), pp. 103–164
25. X. Zhang, W. Lei, M. Liu, L. Zhang, D. Engelsen, X. Zhou, Q. Wang, Analysis of the
transverse energy distribution of hopping electron through a glass funnel. Appl. Surf. Sci.
251, 182–190 (2005)
26. J.M. Chiou, J.L. Kwo, K.J. Chang, S.W. Lai, A.K. Li, H.J. Lai, M. Yokoyama, Field
emission display device. Patent USA US 2005/0099112 of 12.05.2005, cl.313/496 (H01J 21/
10)
27. T.H. Yang, J.M. Chiou, H.Y. Li, C.H. Lin, S.H. Lee, Field emission display device. Patent
USA 2007/0114910 of 24.05.2007, cl.313/495 (H01J 1/00)
28. S.A. Gavrilov, N.N. Dzbanovskiy, E.A. Ilichev, P.V. Minakov, E.A. Poltorackiy, G.S.
Rychkov, N.V. Syetin, Amplification of a flow of electrons by means of a diamond
membrane. ZhTF. T. 74(1), 108–114 (2004)
29. Z. Cui, Simulation of planar deflection system for field emission device. J. Vac. Sci. Technol.
B 19(3), 1077–1081 (2001)
30. H.S. Kim, B.J. Kim, Scanning field emission display. Patent WO 2006/016771 of
16.02.2006, cl.H01J 1/30
31. R.G. Forbes, Progress in establishing field electron emission science, in 10th International
Vacuum Electron Sources Conference, St Petersburg, June–July 2014. doi: 10.1109/
IVESC.2014.6891979
32. R.G. Forbes, Progress in field electron emission fundamentals, in 10th International Vacuum
Electron Sources Conference, St Petersburg, June–July 2014. doi: 10.13140/
RG.2.1.3846.8567
33. R.G. Forbes, Field electron emission theory: Present state and future needs, Lecture, St.
Petersburg, 9 Nov 2015. doi: 10.13140/RG.2.1.1736.2008
34. K.L. Jensen, in Advances in Imaging and Electron Physics, ed. by P.W. Hawkes. Electron
Emission Physics, vol 149 (Elsevier, Amsterdam, 2007)
35. IFES’2010, ed. by Y. Suchorski, G. Schmitz. in 52th International Field Emission Symposia,
Sydney, 2010. Ultramicroscopy, 111(6) (2011)
36. IFES’2012, ed. by B.D. Deconihout, G. Schmitz, G. Thompson. in 53th International Field
Emission Symposia, Tuscalossa, 2012. Ultramicroscopy, 132 (2013)
37. APT&M’2014, ed. by B.D. Deconihout, J. Kreutzer, G. Schmitz. in 1st International

Conference on Atom Probe Tomography & Microscopy, Stuttgart, 2014. Ultramicroscopy,
159(2) (2015)
38. IVEC’2009, in 10th IEEE International Vacuum Electronics Conference, Rome (IEEE,
Piscataway, 2009)
39. IVEC’2010, in 11th IEEE International Vacuum Electronics Conference, Monterey,
California (IEEE, Piscataway, 2010)
40. IVEC’2011, in 12th IEEE International Vacuum Electronics Conference, Bangalore (IEEE,
Piscataway, 2011)
44. IVEC’2013, in 14th IEEE International Vacuum Electronics Conference, Paris (IEEE,
Piscataway, 2013)
45. IVEC’2015, in 16th IEEE International Vacuum Electronics Conference, Beijing (IEEE,
Piscataway, 2015)
46. IVESC-Nanocarbon’2010, in 8th International Vacuum Electron Sources Conference and
Nanocarbon, Nanjing. Ed. by W. Lei, X. Zhang, Zh. Zhao. (IEEE, Piscataway, 2010)
47. IVESC’2012, in 9th International Vacuum Electron Sources Conference, Monterey (IEEE,
Piscataway, 2012)
48. IVECS’2014, in 10th International Vacuum Electron Sources Conference (IVESC), St.
Petersburg. Ed. by N.V. Egorov. (IEEE, Piscataway, 2014)
49. IVNC-IFES’2006, in 19th International Vacuum Nanoelectronics Conference and 50th
International Field Emission Symposium, Guilin (IEEE, Piscataway, 2006)
50. IVNC’2010, in 23rd International Vacuum Nanoelectronics Conference, Palo Alto (IEEE,
Piscataway, 2010)
51. IVNC’2011, in 24th International Vacuum Nanoelectronics Conference, Wuppertal (IEEE,
Piscataway, 2011)
52. IVNC’2012, in 25th International Vacuum Nanoelectronics Conference, Jeju (IEEE,
Piscataway, 2012)
53. IVNC’2013, in 26th International Vacuum Nanoelectronics Conference, Roanoke (IEEE,
Piscataway, 2013)
54. IVNC’2014, in 27th International Vacuum Nanoelectronics Conference, Engelberg (IEEE,
Piscataway, 2014)
55. IVNC’2015, in 28th International Vacuum Nanoelectronics Conference, Guangzhou, ed. by
Q. Zou, J. She, P. Ye, J. Chen, S. Deng. (IEEE, Piscataway, 2015)
56. IVNC’2016, in 29th International Vacuum Nanoelectronics Conference, Vancouver (IEEE,
Piscataway, 2016)
57. Y. Saito (ed.), Carbon Nanotube and Related Field Emitters: Fundamentals and
Applications (Wiley-VCS Verlag GmbH & Co. KGaA, Weinheim, 2010)
58. J. Fan, P.K. Chu, in Engineering Materials and Processes. Silicon Carbide Nanostructures
Fabrication, Structure, and Properties (Springer Science Business Media, Dordrecht, 2014)
59. Y. Li, Y. Sun, J.T.W. Yeow, Nanotube field electron emission: principles, development, and
applications. Nanotechnology 26, 242001 (2015)
60. A. Evtukh, H. Hartnagel, O. Yilmazoglu et al., Vacuum Nanoelectronic Devices: Novel
Electron Sources and Applications (Wiley, Chichester, 2015)
61. M.T. Cole, M. Mann, K.B.K. Teo, W.I. Milne, In Emerging Nanotechnologies for
Manufacturing, in Engineered carbon nanotube field emission devices, ed. by W. Ahmed, J.
M. Jackson (Elsevier, Amsterdam, 2015), p. 125. (Chap. 5)
62. R.G. Forbes, Z. Li, Emission reference level: A missing concept in emission theory, in
IVNC’2011, 24th International Vacuum Nanoelectronics Conference, Wuppertal (2011),
6004592, p. 121
63. R.G. Forbes, Progress in reconstructing field emission theory to make it more transparent, in
IVNC’2015, 28th International Vacuum Nanoelectronics Conference, Guangzhou (2015),
7225514, p. 20
64. R.G. Forbes, A more scientific approach to describing Fowler-Nordheim theory, in
IVNC’2015, 28th International Vacuum Nanoelectronics Conference, Guangzhou (2015),
7225534, p. 66
65. R.G. Forbes, improved methods of extracting area-like information from CFE current-
voltage data, in IVNC’2015, 28th International Vacuum Nanoelectronics Conference,
Guangzhou (2015), 7225536, p. 70
66. R.G. Forbes, Comments on the voltage scaling of field electron emission current-voltage
characteristics, in IVNC’2014, 27th International Vacuum Nanoelectronics Conference,
Engelberg (2014), 6894781, p. 116
67. R.G. Forbes, Progress in establishing field electron emission science, in IVESC-ICEE’2014,
10th International Vacuum Electron Sources Conference, and 2nd International Conference
on Emission Electronics, ed. by N.V. Egorov, St. Petersburg, 6891979 (2014)
68. R.G. Forbes, Development of an integrated theory of field emitter optics, in BDO’2014, 20th
International Workshop on Beam Dynamics and Optimization, ed. by D.A. Ovsyannikov, St.
Petersburg, 6890017 (2014)
69. R.G. Forbes, Conceptual error in the international definition of voltage, and implications for
theories of patch fields and electron emission, in IVESC-ICEE’2014, 10th International
Vacuum Electron Sources Conference, and 2nd International Conference on Emission
Electronics, ed. by N.V. Egorov, St. Petersburg, 6891978 (2014)
70. R.G. Forbes, Use of a spreadsheet to test for lack of field emission orthodoxy, in IVESC-
ICEE’2014, 10th International Vacuum Electron Sources Conference, and 2nd International
Conference on Emission Electronics, ed. by N.V. Egorov, St. Petersburg, 6891980 (2014)
71. A.G. Kolosko, S.V. Filippov, P.A. Romanov, E.O. Popov, R.G. Forbes, Real-time
implementation of the “orthodoxy test” for conformity of current–voltage characteristics
with classical field electron emission theory. J. Vac. Sci. Tech. 34, 041802 (2016)
72. Sh-D Liang, Quantum Tunneling and Field Electron Emission Theories (World Scientific
Publishing, Singapore, 2014)
73. M.A. More, R.G. Forbes, in Encyclopedia of Nanotechnology, ed. by B. Bhushan. Field
Electron Emission from Nanomaterials (Springer Science Business Media Dordrecht, 2015)
Index
A Anodic aluminum oxide, 273, 276, 279

Accelerating electrode, 501, 511 Anodic layer, 462
Accelerating gap, 504 Anodizing electrolyte, 274
Accelerating lenses, 66 Aperture, 495
Activation energy, 104, 158 Apex
Adsorbate, 48 curvature radius, 76, 85
Adsorbed atoms, 231, 265, 266 radius, 157
Adsorbents, 101 shape, 158, 402
Adsorption, 101, 107 surface, 249
Adsorption centers, 105 Aquadag, 50
A flat display, 461 Archimedes buoyant, 85
Algebraic equations, 395 Array of emitters, 483
Alloyed cathode, 344 Atmospheric pressure, 444
Aluminum coating, 442 Auger electrons, 513
Aluminum dioxide, 253 Autoelectron emission, 2
Aluminum film, 54, 274, 318 Axial-potential distribution, the, 400
Amorphous carbon, 317
Amorphous films, 320 B
Amorphous-silicon crystallization, 463 Back light, 452
Amplifier, 487, 551 Balance beam, 72
Analogue-to-digital, 62 Bandgap, 20
Analogue-to-digital converters, 75, 347 Barrier
Analytical approximations, 485 penetrability, 2, 4, 173
Analytical integration, 137 shape, 171
Analytically, 131 transmission, 129
Analytical solution, 372 transparency, 178
Analyzer, 62, 67 Barrier-transmission factor, 4
Analyzer resolution, 59 Beam, 499
Angular dependence, 140 Beam analyzer, 512
Angular distribution, 456 Beryllium ceramics, 53
Angular spread, 247 Binding energy, 105
Anisotropy, 332 Bispherical coordinates, 388
Annealing temperature, 334 Blade, 251, 253, 255, 263, 479
Anode, 29, 50 Body-centered cubic, 46, 191
cathode gap, 465 Boguslavskiy–Langmuir, 215
cavity, 442 Boguslavskiy–Langmuir law, 217
diaphragm, 407, 408 Boltzmann distribution, 146
oxidation, 275 Bombardment, 498
voltage, 406, 408, 410 Boron-doped, 342

in Advanced Microelectronics 60, DOI 10.1007/978-3-319-56561-3
560 Index
Boundary condition, 163, 215, 374, 375, 382, Chemosorbed layer, 104
388, 392, 414, 415 Chloroform, 280
Boundary problem, 375, 391 Chromaticity, 430
Boundary-value problem, 122, 400 Chromium, 259
Brightness, 357, 429, 456, 474 Classical diode, 446
Broken bonds, 188 Coating, 235, 442
Build-up, 345 Coating material, 260
Coaxiality, 446, 450
C Cold emission, 2
Cantilever, 548 Cold emitter, 22
Carbon-based, 296 Collector, 58, 59
Carbon-based material, 515 Color, 430
Carbon-carbon composites, 546 Color displays, 461
Carbon Composite, 546
fiber, 56, 298, 302, 332, 341, 434, 439, 445, Computer-controlled, 75
496 Concentration of impurities, 155
material, 81, 233, 297, 480, 544 Conditional optimization, 409
nanocrystals, 311 Conduction band, 21
nanofiber, 304 Conductive coating, 444
nanomaterials, 308 Cone-apex angle, 79
nanostructures, 306 Cone-shaped, 449
nitrogen, 304 Conformal mapping, 371
polyacrylonitrile, 465 Conical emitters, 485
powder, 331, 334 Contact resistances, 348
Carbon nanotube (CNT), 263, 282, 306, 309, Continuous mode, 441
313, 320, 325, 332, 342, 344, 353, 354, Control, 180
441, 452, 496, 547 electrode, 245, 263, 474
Cartesian coordinate, 411 function, 179
Catalytic growing, 304 signal, 359
Catalytic layer, 262 voltage, 431, 467
Cathode, 116, 142, 214, 352, 354, 373, 390 Controlling systems, 373
array, 461, 494 Copolymer, 283
coating, 477 Copper, 252, 257
current, 435 Corona discharge, 340, 341
degradation, 354 Crater, 323, 325, 326
design, 239 Criterion function, 409
diaphragm, 407 Cross-beam, 71
forming, 298 Crossing fields, 488
lifetime, 370 Crystal
parameters, 358 faces, 105
sputtering, 229, 234, 498 irregularities, 28
stability, 233 plates, 238
surface, 175, 196, 199, 414, 417 tips, 35
Cathodoluminescent, 427 Crystalline facets, 196
Cathodoluminescent lamp, 458 Crystalline lattice, 39
Center of gravity, 410 Crystalline plane, 118
Ceramic insulation, 73 Crystallization, 271
Charge-carrier concentration, 145 Crystallization method, 267
Charge density, 9, 370, 402 Crystallographic
Charge distribution, 122 axis, 323
Charging, 472 faces, 76
Chemical etching, 241, 257, 303 facets, 188, 194
Chemical polishing, 91, 98 planes, 235
Chemical vapor deposition, 311 Crystallography, 31
Index 561
Cubic crystals, 195 Donor level, 24

Current, 327, 546 Doping, 476
density, 12, 36, 117, 119, 131, 133, 146, Dynamic light, 455
147, 165, 197, 198, 202, 210, 213, 217,
218, 224, 386, 401, 404 E
distribution, 151 Ecological compatibility, 427
drop, 86 Effective mass, 151
pulses, 142 Effectiveness, 499
saturation, 154 Efficiency, 447
stability, 229, 231, 232, 509 Efficient system, 553
transmission, 431 Eigenfunction expansion, 396
Current-density distribution, 421 Eigenvalues, 385
Current-density fluctuations, 422, 423 Electrical breakdown, 479
Current-voltage, 360 Electrical lens, 504
Current-voltage characteristic, 25, 37, 142, Electric circuits, 68
335, 345, 353 Electric field, 2, 9, 10, 83, 123, 128, 149, 177,
Curvature distribution, 413 334, 383
Curvature radius, 44, 88, 256, 377, 382, 408, Electrochemical, 91
412 anodizing, 277
Curvilinear coordinate, 123, 183 etching, 84, 192, 268, 463
CVD, 311, 470, 522 polymerization, 281
Cylindrical cathode, 255, 447, 517 processing, 256
Cylindrical structure, 445 Electrode, 91, 373, 418
Cylindrical triode, 445 Electrode coated, 262
Electrode-separation distance, 97
D Electrokinetic potential, 330, 331
Deflecting system, 512 Electrolyte, 84, 86, 88–90, 92, 95
Defocusing, 261 bath, 82, 83
Delocalizing, 236 solution, 88, 332
Density, 177 cell, 276
Desorption, 108 etching, 45, 86, 87
Device, 329 Electromagnetic deflection, 57
Dewar vessel, 74 Electromagnetic wave, 488, 492
Diamond-like films, 235 Electron, 18, 129
Diamond-like structures, 318 affinity, 23
Diamond paste, 314 bombardment, 435, 472
Diamond powder, 441 charge, 165
Diaphragm, 60, 66, 410 concentration, 148–150
Dielectric matrix, 238 current, 1
Diffusion, 160 emission, 2, 144
coefficient, 102, 103 flow, 432
constant, 103 flux, 3
processes, 32, 103, 250 gas, 25, 142
Digital-to-analogue converter, 62 gun, 373, 434, 500, 501, 504, 507, 509, 513
Diode, 151, 327, 438, 439, 459 microscope, 500, 507–509
Diode display, 464 multiplier, 65, 551
Diode structures, 463 path, 419, 405
Dipole layer, 8 temperature, 146, 149, 150
Dipole moment, 9 trajectories, 478
Direct-current, 353 volts, 132
Discharge, 271, 316, 477 Electron beam , 55, 56, 237, 447, 492, 500,
Display screen, 466, 476 503, 520, 553
Distribution, 136 diameter, 474
Distribution of electrons, 176, 177 displays, 459
562 Index
Electron beam (cont.) Emitted electrons, 139

divergence, 235 Emitter, 79, 80
focusing, 511 apex, 194
lithography, 522 area, 496
microlithography, 243 surface, 43
scanning, 513 temperature, 202
vaporization, 477 tip, 28
Electron-diffraction, 258 Emitting
Electron-divergence angle, 452 area, 36, 261
Electron-electron interaction, 212 centers, 501
Electronic devices, 52, 251, 516 material, 276
Electronic systems, 551 microelevations, 322
Electronic wave, 16 surface, 76, 119, 213, 255, 256, 435
Electron-optical, 64, 107 Emitting-surface area, 107
Electron-optical system, 373, 388, 391, 497 Energy
Electroosmosis, 330, 337, 464 analyzer, 57
Electrophoretic deposition, 333 balance, 324
Electrophoretic process, 331 conversion, 428
Electropolishing, 88 diagram, 151
Electrostatic distribution, 23, 57, 115, 142, 143, 176
deflection, 57 source, 458
fields, 402 spectrum, 171, 547
forces, 340, 513 Equilibrium, 159
pressure, 159 Equilibrium shape, 155
repulsion, 220 Equipotential, 123
Ellipsoid, 123, 380, 416 Equipotential lines, 403, 404, 416
Ellipsoid of rotation, 184 Equipotential surface, 194, 376, 415
Elliptic apex, 403, 405 Etching, 80, 96, 241
Elliptical integrals, 130 current, 84
Elliptic integrals, 17 rate, 84
Emission, 546 selectivity, 97
area, 118, 120, 185, 224, 378, 379 Evaporating getter, 443
cathode, 53, 56, 478, 497 Evaporation, 544
centers, 56, 233, 272, 296, 298, 319, 329, Evaporation-anode, 518
355, 356, 549 Exhaust tube, 467
characteristics, 73, 75, 220, 340, 406, 407 Experimental curve, 350
current, 117, 120, 127, 129, 166, 178, 183, Explosion, 544
202, 207, 209, 210, 213, 231, 235, 247, Explosive electron, 204
248 Explosive-emission, 300
device, 474 External distribution, 175
diode, 186, 217, 222, 407 External field, 154
electronics, 250 Extraction electrode, 514, 515
homogeneity, 370
image, 46, 56, 120, 143, 193, 199, 208, F
357, 422, 424 Fabrication method, 481
instability, 273 Facet, 191
mechanism, 309 Faraday cylinder, 61, 524
microscope, 27, 47, 48, 124 Fe nanoparticles, 312
pattern, 198, 543 Fermi-Dirac distribution, 3, 153
properties, 59 Fermi-Dirac statistics, 12
spectroscopy, 115, 135 Fermi level, 8, 20, 23, 133, 138, 210, 522
stability, 300 Fiber, 52, 295
system, 383, 414 Fiber cathodes, 109
Emittance, 370 Fibrils, 52
Index 563
Field, 139 Getter, 465, 516

emitter, 76, 210 Getter pump, 517
enhanced, 204 Glass bulb, 447
enhancement, 267 Glass substrate, 301
penetration, 24 Glass tubes, 300
strength, 152 Graphene plane, 307
Field-electron microscope, 418 Graphite, 301, 304, 517
Field emission, 11, 20, 25, 124, 145, 166, 204, filaments, 302
232, 247, 350, 354 flakes, 325
cathode, 382 foil, 322, 324
diode, 413 Graphite-like materials, 309
spectrum, 309 Grid, 515
Field-enhancement factor, 381 Grossmann’s laws, 430
Film, 257 Growing crystal, 266
Finite-difference, 485 Growing whiskers, 271
First approximation, 387
Flange, 51 H
Flat display, 328, 460, 483 Hafnium, 259
Flow continuity, 222 Half maximum, 138
Fluctuations, 355 Heisenberg uncertainty, 30
Fluorescent layer, 453 Hemisphere-shaped apex, 160
Focal distances, 416 Hemispherical
Focusing, 553 analyzer, 66
Focusing-diaphragm system, 401 collector, 61
Focusing electrode, 373, 462, 477 grid, 60
Focusing system, 63, 475 shape, 191
Form factor, 299, 336, 341, 379, 380 High-index facets, 198
Fourier components, 178 High-melting metal, 208
Fourier expansion, 394 High-power, 514
Fowler−Nordheim High stability, 546
equation, 35, 37, 39, 133, 181, 220, 348 High-vacuum, 72
coordinates, 106, 182, 184, 185, 299, 336, High voltage, 70, 358
345–347, 349, 350, 352, 360 power, 74
formula, 20, 116, 218, 383 source, 74
graph, 237 stabilizer, 75
line, 38, 217, 219 switching, 358
plot, 134 Histograms, 357
theory, 11, 36, 115, 352 Hologram, 456
Fredholm equations, 397 Homer formula, 379
Free electrons, 6 Homogeneity, 541, 550
Fringe, 263 Hydrogen, 317
Full-color image, 526 Hydrogen atmosphere, 437
Full-energy, 136 Hydrogen medium, 316
Full-energy distribution, 137, 138 Hyperbolic apex, 405
Functional, 180 Hyperboloid, 123, 184, 380, 403, 411, 415
Hyperboloid of rotation, 183, 185
G
Gas adsorption, 542 I
Gas ions, 33 IFES, 554
Geometrical environment, 189 Illumination, 440
Geometrical parameters, 416 Image
Germanium, 242 brightness, 543
564 Index
Image (cont.) Lateral-gate triode, 445

brilliance, 474 Lattice, 211
contrast, 461 Lattice-heating, 212
element, 472 Lattice parameter, 191, 192
forces, 7, 13, 118 LCD, 453, 454
potential, 10 Leakage current, 245, 337, 468
Impulse formation, 489 Least-square method, 38, 184, 185
Indium-tin oxide, 443 Legendre-function expansion, 391, 394
Integral-equation, 371 Legendre functions, 395
Integral operator, 384 Legendre-polynomial, 385, 392, 393
Integration, 392, 393 Lenses, 65
Inverse layer, 150, 151 Life cycle, 502
Ion Lifetime, 428
beam, 370, 549 Light
bombardment, 98, 100, 106, 108, 206, 229, emission, 526
303, 342, 502, 506, 542, 544 flux, 429, 454, 525
etching, 337 intensity, 526
injector, 549 source, 427, 436, 439, 440, 441, 446, 448,
sources, 523 526
Ionic gas, 269 Light-emitting cell, 480
Ionic lens, 524 Linear
Ionic source, 524 algebraic equations, 396, 399
Ionization, 260, 523 combination, 179
Ionization potential, 205 function, 399
Ion-plasma etching, 241 regression, 39
Isopropyl alcohol, 330 Liquid contact, 90
Isotropic crystal, 156 Liquid-crystal displays, 453
Isotropic material, 257 Liquid metal, 286
IVEC, 555 Lithium perchlorate, 281
IVESC, 555 Local enhancement, 327
IVNC, 555 Long-term, 327
Long-term stability, 306
J Long-term test, 360
Joule heat, 162, 201, 202, 236 Lorentz forces, 521
Lower-melting material, 208
K Low-index steps, 196
Keen blades, 251 Low vacuum, 470
Kernels, 398 Low-voltage, 481
Kinetic energy, 16 Luminescence, 442
Klystrode, 488 Luminescent layer, 431
Klystron, 488 Luminescent screen, 29, 43, 68, 72, 432
Kovar ring, 49 Luminophore, 49, 435, 454
Luminophore brightness, 431
L Luminophore degradation, 482
Lagrange equations, 416 Luminophor screen, 51
Laplace equation, 122, 123, 374, 375, 388, Luminosity, 356
391, 413
Laplace method, 378 M
Large-area screens, 466 Magnetic field, 222, 333
Laser, 324 Magnetic sens, 520
Lateral Magnetron, 279, 489, 490
design, 476 Magnetron sputtering, 318
emitter, 490 Magnification, 109
structures, 262 Magnification factor, 30
Index 565
Man-made, 250 Monte–Carlo method, 423

Mapping parameter, 410 Morgulis-Stratton theory, 21
Masking material, 473 Multigrid structure, 492
Mass spectrometer, 270 Multi-layer coating, 104
Mass spectroscopy, 343 Multiple-apex, 273
Mathematical model, 185, 224, 416 Multiple-blade, 251
Mathematical simulation, 174, 187, 188 Multiple-wall, 306
Matrix operator, 398 Multiple-wall nanotubes (MWNT), 306, 308,
Maximum function, 409 312, 437
Measurement scheme, 359 Murphy-Good formula, 135
Mechanical polishing, 314 Murphy–Good theory, 350
Mechanical strength, 250, 515
Memory cell, 521 N
Memory element, 521 Nanofilms, 320
Mersenne twister, 200, 422 Nanoporous carbon, 338
Metal/semiconductor surface, 5 Nanowires, 284
Metal, 76, 267 Niobium, 252, 259
Metal foil, 237 Niobium emitter, 209
Metal-jet, 78 Nomograhic charts, 404
Metallic surface, 6 Non-linear regression, 38
Metal particles, 238 Nonparametric-regression, 38
Methane-hydrogen mixture, 276, 319 Nordheim elliptic functions, 409
Microcapillary, 268 Nordheim function, 18
Microchannels, 469 Normally distributed, 180
Microcracks, 315 Nottingham effect, 161–166, 201, 202, 209,
Microdevice, 278 214, 236
Microelevations, 238, 254, 269 Nottingham energy, 212
Micrometric screw, 97 Nottingham heat-release, 211
Micropolishing technique, 88 N-type, 141
Microscope, 96 Numerical
Microscope kinematics, 70 computation, 381
Microscopic image, 35 experiment, 173, 178, 224
Microstructure, 76 method, 224
Microtriode, 486 solution, 125
Microwave
amplifier, 487 O
devices, 483, 486 Oil-diffusion pump, 252
field, 286 One-dimensional, 163
generator, 487 One-dimensional model, 153
plasma, 276, 329 Operating frequency, 488
Migrating atom, 101 Optical system, 518
Migration energy, 103 Optimal parameters, 550
Miller-index, 194, 196, 236 Optimization, 386
Modulation, 457 Optimization problem, 402
Modulator, 388, 432 Optimized design, 435
Molecular flow, 272 Optimum, 386
Molybdenum, 47, 65, 100, 243, 245 Organic complexes, 284
cone, 304 Oscillogram, 509
grid, 434 Oxidation of carbon, 340
conservation, 140 Oxidizer, 86
Monochrome, 461
Monocrystal, 502 P
Monoenergetic source, 401 Packing density, 273
Mono-layer, 343 Packing parameter, 386
566 Index
Parabolic cylinder, 257 Projection, 261, 518

Paraboloid, 123 Prolate ellipsoid, 187
Paraboloid of rotation, 185 Prospective development, 541
Particle beam, 221 Protective cylinder, 63
Penetration of the field, 145 Prototype design, 434
Periodic pulses, 55 Pseudorandom, 422
Phosphor, 433 Pseudorandom-number, 200, 409
Phosphor layer, 43, 450 P-type, 144, 154
Photocathode, 242, 527 P-type semiconductor, 150
Photo-electron emission, 10 Pulse duration, 249
Photo field emission, 152 Pump, 68, 516
Photo lithographic method, 238, 241, 320 Pyrographite, 321
Photolithography, 471 Pyrolytic graphite, 318
Photomasks, 81 Pyrolytic-graphite plates, 321
Photometrically, 193
Photoresist, 81, 314, 469 Q
Photosensitive glass, 469 Quadratic equation, 420
Photosensitivity, 150 Quadratic forms, 413
Physical electronics, 234 Quadrode, 459
Pixel, 454, 457, 461, 480, 482 Quantum confinement, 555
Pixel area, 482 Quantum point, 548
Planar structure, 320 Quantum-scale effects, 547
Plasma, 296 Quasi-periodic structures, 248
Plasmachemical treatment, 340 Quasi-steady, 144
Plasma CVD, 317 Quipotential surface, 377
Poisson equation, 214, 221
Polarization, 7, 420 R
Polarization force, 6 Radial projection, 29
Polishing conditions, 92 Radiation, 161
Polishing solution, 96 Radiation spectrum, 451
Polyacrylonitrile carbon fiber, 106, 295, 321, Radiation treatment, 338
465 Rapid-solidification methods, 77
Polycarbonate membrane, 285 Rare-earth elements, 449
Polychromatic backlight, 454 Rectangular projection, 466
Polymer/semiconductor, 280 Reflected wave, 17
Polymer, 283 Reflection factor, 12
Polymer composite, 280 Reflective klystron, 493
Polymeric coating, 284 Reflex klystron, 493
Ponderomotive, 35, 106, 339 Regularization matrix, 172, 173
Ponderomotive loads, 444 Regularization parameter, 171
Ponderomotive pressure, 159 Relative error, 421
Ponderomotor stress, 542 Relativistic electrons, 220
Pore-sputtering depth, 278 Reproducibility, 273
Porous anodizing, 275 Reproducibility of characteristics, 299
Potential barrier, 2, 4, 15, 108, 118, 127–129, Reproducible properties, 547
134, 136, 172, 174, 419, 420 Residual gas, 33, 68, 260, 505
Potential-barrier transmission, 12 Residual-gas pressure, 328
Potential distribution, 392, 396 Resistance heating, 79
Potential energy, 8, 13, 14, 128, 419 Resolution, 60, 67, 458, 512
Power consumption, 518 Resonant tunneling, 309, 311
Power source, 91 Resonator, 493
Power-supply, 467 Retarding area, 58
Pre-explosive effects, 204 Retarding grid, 524
Probe-hole, 63, 65, 71 Retarding potential, 60
Index 567
Revolution-ellipsoid coordinate, 411 Stabilization period, 509

Ring effect, 203, 204, 326, 329 Stationary solution, 161
Runge–Kutta method, 222, 416 Statistical tests, 421
Rystallographic planes, 501 Steel, 251
Stereographic projection, 45, 46, 100
S Stereographic structures, 31
Sapphire plate, 73 Straight line, 352
Scanning beam, 512 Sub-microsecond, 489
Schottky effect, 13 Sub-millimeter wavelengths, 494
Schottky formula, 127 Substrate material, 471
Schrödinger equation, 14, 178, 423 Successive approximation, 181
Schwarz-Christoffel transformations, 371 Summation, 393
Secondary electrons, 54, 328, 508, 513, 551 Supermolecular structure, 283
Secondary emission, 550–552 Surface, 32, 215
Secondary ion, 343 atom, 189
Self-diffusion, 156 curvature, 33, 188
Self-diffusion mechanism, 158 diffusion, 156, 264
Semiconductor, 20, 95, 98, 144, 233, 241 energy, 155
substrate, 241 geometry, 34
surface, 21, 148 layer, 25, 208
vacuum, 140, 146 migration, 231
Sensor, 520 morphology, 249
Shape memory, 519 phonons, 40
Shape of the tip, 157 state, 155, 25, 26, 153, 233
Sharpness, 474 structure, 296, 548
Shoot-through anode, 499 tension, 206, 230
Silicon, 143, 253, 279, 522 Surface-melting point, 204
Simplifying assumptions, 370 Surface-tension forces, 159
Simulation, 191, 200 SWNT, 308
Single crystal, 95, 242, 279 Symptotic, 378
Single-wall, 306
Small-aperture diaphragm, 400 T
Solid-state memory, 522 Tantalum, 252
Solid surface, 174 Taylor cone, 205
Solid-vacuum interface, 4 Taylor expansion, 132, 136
Solution of equation, 164, 377 Taylor series, 175
Solution of system, 399 Technical vacuum, 230, 505
Solvent, 86 Technological process, 246
Space charge, 116, 117 Temperature, 134, 139
Spark-erosion technology, 443 Temperature deformations, 49
Spatial instability, 515 Temperature distribution, 162, 164
Special coordinates, 412 Test circuit, 360
Spectrometer, 512 Theory of A.F.Yatsenko, 150
Sphere-on-cone surface, 44, 190, 396 Thermal
Sphere on the orthogonal cone, 31 annealing, 336
Spherical, 447 emission, 212
Spherical coordinate, 221, 223 evaporation, 99
Spherical diode, 217 field, 138
Spheroidization, 158 stability, 50
Spindt, 244 Thermal-field electrons, 138
Spindt cathode, 246, 541 Thermal-field emission, 10, 26, 125
Spreading angle, 479 Thermally-expanded graphite, 322, 323
Square-grid, 383 Thermionic cathodes, 489
Stability, 303, 358, 508, 511 Thermionic emission, 201
568 Index
Thermochemical treatment, 338 UV lamp, 449

Thermocouples, 53 UV radiation, 450, 451
Thermoemission, 149
Thin-film, 80, 277, 278, 337, 479, 480, 483 V
luminophore, 444 Vacuum, 68, 128, 339, 432
technology, 246 annealing, 548
transistor, 475 chamber, 71, 247, 298, 348
Thin shell, 188 conditions, 297
Threshold field, 249 device, 439
Time-independent, 372 gauge, 69
Tip apex, 35 microelectronic, 486
Tip-curvature radius, 96 pump, 507
Titanium, 259 sensors, 521
Titanium-aluminum, 277 tube, 434
Titanium film, 318 Vacuum-electronic devices, 273
Topography, 284 Valence band, 21
Total current, 182, 401, 407 Van der Waals interaction, 312
Trajectories, 223 Vapor pressure, 268
Transconductance, 352 Variable-separation, 372
Transition temperature, 302 Video module, 457
Transmission Virtual cathode, 224
coefficient, 132 Virtual-emitter, 206
factor, 39, 148, 484 Voltage-current characteristics, 184
line, 484, 485 Voltage fluctuations, 352
microscope, 504, 505 Volt-ampere characteristic, 208, 219, 279, 346
waves, 14
Transparency, 125 W
Travelling-wave, 492 Wavelength, 451
Travelling-wave tubes, 492 Wave vector, the, 39
Triangular barrier, 15, 179 Wentzel-Кramers-Brillouin (WKB) method,
Tricolor, 456 12, 16, 37, 176
Triode, 389, 439, 459, 467, 472 Whisker, 246, 264, 266, 271, 272
Triode microstructure, 254 Whisker tip, 265
Triode structures, 469 Wide-gap semiconductor, 311
Tungsten, 50, 81, 96, 100, 197, 219, 342 Work function, 1, 3, 6, 9, 18, 19, 28, 29, 45,
dendrite, 269 103, 109, 118, 119, 173, 179, 187, 194,
emitter, 193 197, 199, 218, 297, 344, 405, 502
rhenium, 257 Worm gear, 72
wire, 268
Tunnel effect, 5 X
Tunneling, 103, 145, 280 X-ray intensity, 498
Turbomolecular pump, 69 X-ray tube, 494–496
U Y
Ultra-high vacuum, 67, 69, 72, 230, 247, 494, Yanoacrylate, 282
505 Young-Kuyatt criterion, 67
Ultrasonic cutting, 79
Ultraviolet, 472 Z
Ultraviolet lamps, 448 Zero-current approximation, 26
Ultra-violet photolithograph, 469
Uminescent screen, 436

(Springer Series in Advanced Microelectronics Vol. 10.1007 - 978!3!319-56561-3) Egorov, Nikolay

Uploaded by

Copyright:

Available Formats

You might also like

(Springer Series in Advanced Microelectronics Vol. 10.1007 - 978!3!319-56561-3) Egorov, Nikolay

Uploaded by

Document Information

Original Title

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

(Springer Series in Advanced Microelectronics Vol. 10.1007 - 978!3!319-56561-3) Egorov, Nikolay

Uploaded by

Copyright:

Available Formats

Springer Series in Advanced Microelectronics 60

More information about this series at http://www.springer.com/series/4076

Field Emission Electronics

ISSN 1437-0387 ISSN 2197-6643 (electronic)

Library of Congress Control Number: 2017939090

Copyright for the English version is with the authors

Printed on acid-free paper

This Springer imprint is published by Springer Nature

Modern technological applications of electron beam devices require development

St. Petersburg, Russia Nikolay Egorov

3 Modern Developments in Theoretical Research of Field

6 Carbon-Based Field-Emission Cathodes . . . . . . . . . . . . . . . . . . . . . . . 295

7.6 Computation of Electron Paths in a Field-Emission

For readers’ convenience, each chapter has a separate list of references.

Abstract This chapter describes the basic principles of Fowler–Nordheim theory

Structures comprised of charged particles (electrons or ions), such as solids, liquids

© Springer International Publishing AG 2017 1

where h is Plank constant.

Fig. 1.1 Arbitrary potential

(a) (b) (c)

Thus, the work function can generally be expressed as:

Fig. 1.3 On the calculation

This expression cannot be directly applied to semiconductors and dielectrics.

Fig. 1.4 The decrease in the

Fig. 1.5 The spatial Distance, Å

Energy, atomic units

Distance x, Fermi wavelenght units

δΦ total potential of the

So inclusion of electrostatic-image forces shows that electric ﬁeld imposition

1.2 Fowler–Nordheim Theory of Field Emission

DðEx ; EÞ ¼ 1 RðEx ; EÞ; ð1:7Þ

where R is the electron wave reflection factor.

Fig. 1.7 Typical areas of electron energy interval

Fig. 1.8 To derivation of the basic formula of ﬁeld electron emission

By deﬁnition, the barrier penetrability is

jat j2 jai j2 jar j2

where k02 ¼ h2m

can be integrated, and its solution is:

In this equation, m, x, k1, k2 etc. are functions of x. They can be derived by

The primes denote differentiation with respect to x.

This corresponds to D 7%. The calculation presented above was proposed by

In this equation, a0 is represented with the Bohr radius of a hydrogen atom. It

As inside the metal at a sufﬁcient distance from the surface,

UðxÞ ¼ k02 ðE þ WÞ ¼ v2 ¼ const; ð1:24Þ

or according to (1.21) and (1.11)

Table 1.1 Nordheim y h(y) s(y) y h(y) s(y)

To calculate Q, we introduce function h(y), the so-called Nordheim function:

Fig. 1.9 Dependence h(y) of θ(y)

Integration over vy and vz results in expression

and with E [ 0, this expression is equal to zero.

In this expression, U is the work function for metal and

The bracketed expression is later disregarded because it is very close to one. As a

1.3 Classical Morgulis-Stratton Theory of Field Emission

The theory of ﬁeld emission from semiconductors is based on the research by

Fig. 1.11 The diagram of the

Fig. 1.12 Schematic band

m ¼ ½ðe 1Þ=ðe þ 1Þ1=2 ;

where e is the static dielectric constant of the semiconductor.

m ¼ ½ðe 1Þ=ðe þ 1Þ1=2 ;

" pffiffiffiffiffiffi 3=2 !#

f ðsÞ ¼ hI ðtÞI ðt þ sÞi;

dI=dUcoll ¼ Pðucoll u þ Ucoll Þ;