Elements of Plasma Technology: Chiow San Wong Rattachat Mongkolnavin

SPRINGER BRIEFS IN
APPLIED SCIENCES AND TECHNOLOGY
Chiow San Wong
Rattachat Mongkolnavin
Elements
of Plasma
Technology
123
SpringerBriefs in Applied Sciences
and Technology
More information about this series at http://www.springer.com/series/8884
Chiow San Wong Rattachat Mongkolnavin
•
Elements of Plasma
Technology
123
Chiow San Wong Rattachat Mongkolnavin
Plasma Technology Research Centre, Department of Physics, Faculty of Science
Physics Department, Faculty of Science Chulalongkorn University
University of Malaya Bangkok
Kuala Lumpur Thailand
Malaysia
ISSN 2191-530X ISSN 2191-5318 (electronic)

SpringerBriefs in Applied Sciences and Technology
ISBN 978-981-10-0115-4 ISBN 978-981-10-0117-8 (eBook)
DOI 10.1007/978-981-10-0117-8
Library of Congress Control Number: 2015957048
© The Author(s) 2016

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.
Printed on acid-free paper
This Springer imprint is published by SpringerNature

The registered company is Springer Science+Business Media Singapore Pte Ltd.
Preface
Plasma technology is often referred to as one of the critical technologies of the

twenty-first century. In the developed countries such as the USA and Japan,
intensive scientific effort and large funding are being spent on research to develop
plasma technologies that are useful for industry. In recent years, newly developed
nations such as Korea, China (including Taiwan), and Singapore have also joined
the race in the quest to be at the frontiers in this area of technology development.
The effect of such efforts on the national economy of these nations is obvious.
This monograph covers some fundamental aspects of plasma technology which
the authors believe are important to help researchers who intend to start work in
plasma technology development. These include basic properties of plasma, methods
of plasma generation, and basic plasma diagnostic techniques. The treatments are
more experimental than theoretical, although some discussions on theoretical
background will be given where appropriate. It is also not the intention of the
authors to provide a complete and comprehensive treatment of all the topics cov-
ered. The more advanced readers may find the contents of this monograph too
“elementary,” which is precisely its intended aim. This monograph is for the
beginners, not for the experts.
Also included in this monograph are discussions on several plasma devices with
potential applications in industry including pulsed plasma radiation sources,
low-temperature plasmas such as glow discharge and dielectric barrier discharge.
The principles of the operation of these devices and some results of the studies
carried out by the authors and their teams at the Plasma Technology Research
Centre of University of Malaya and at the Physics Department of Chulalongkorn
University will be reviewed and discussed.
Chiow San Wong

Rattachat Mongkolnavin
v
Acknowledgments
The authors are grateful to University of Malaya and Chulalongkorn University for
providing research grants and facilities for them to carry out research in the area of
plasma technology and also for providing job opportunity to them as well as
supporting their collaboration. They also acknowledge the contribution and coop-
eration of group members, collaborators, and students in all the projects they have
been involved. Particularly, the technical contribution of Mr. Jasbir Singh is most
valuable toward the success of all the projects carried out by them. In addition, they
would also like to thank the American Institute of Physics, Elsevier, Cambridge
University Press and The Japan Society of Applied Physics for permission to use
some figures from their publications as listed below:
Chin OH and Wong CS (1989) “A Simple Monochromatic Spark Discharge Light
Source”. Rev. Sci. Instrum. 60: 3818–3819. © 1989, AIP Publishing LLC.
Wong CS, Woo HJ and Yap SL (2007) “A Low Energy Tunable Pulsed X-Ray
Source Based On The Pseudospark Electron Beam”. Laser & Particle Beams 25:
497–502. © 2007, Cambridge University Press.
Chan LS, Tan D, Saboohi S, Yap SL, Wong CS (2014) “Operation Of An Electron
Beam Initiated Metallic Plasma Capillary Discharge”. Vacuum 103: 38–42. ©
2014, Elsevier.
Wong CS, Choi P, Leong WS and Jasbir S (2002) “Generation of High Energy
Ion Beams from a Plasma Focus Modified for Low Pressure Operation”. Jpn.
J. Appl. Phys. 41: 3943–3946. © 2002, JSAP.
vii
Contents
1 Basic Concepts in Plasma Technology. . . . ............... . . . . 1

1.1 Plasma—The Fourth State of Matter . . ............... . . . . 1
1.2 Collision. . . . . . . . . . . . . . . . . . . . . . ............... . . . . 2
1.2.1 Elastic Collision . . . . . . . . . . . ............... . . . . 2
1.2.2 Inelastic Collision . . . . . . . . . . ............... . . . . 4
1.3 Collision Cross-Section . . . . . . . . . . . ............... . . . . 5
1.4 Fundamental Processes in a Plasma . . . ............... . . . . 6
1.5 Some Consideration of Thermodynamic Properties of Plasma . . . . 7
1.6 Concept of Plasma Potential . . . . . . . . ............... . . . . 9
1.7 Criteria of Plasma . . . . . . . . . . . . . . . ............... . . . . 10
1.8 Effect of Boundary in Plasma . . . . . . . ............... . . . . 11
1.9 Particle Transport Inside Plasma . . . . . ............... . . . . 12
2 Methods of Plasma Generation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.1 DC Electrical Discharge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.1.1 Electrical Breakdown . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.1.2 The I-V Characteristic of Electrical Discharge . . . . . . . . . . 20
2.1.3 The Corona Discharge . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.1.4 The Glow Discharge . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.1.5 Hot Cathode Discharge . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.1.6 The Arc Discharge . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.2 AC (Radiofrequency) Discharge . . . . . . . . . . . . . . . . . . . . . . . . 29
2.3 Microwave Heating of Plasma . . . . . . . . . . . . . . . . . . . . . . . . . . 34
2.4 Pulsed Plasma Discharges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
2.4.1 Pulsed Arc Discharge in High Pressure Flash Lamp. . . . . . 38
2.4.2 Inductive Model of Pulsed Discharge—Shock Heating . . . . 41
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
ix
x Contents
3 Plasma Diagnostic Techniques . . . . . . . . . . . . . . . . . . . . . . . . .... 49

3.1 Electrical Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . .... 49
3.1.1 Pulsed Current Measurement by Using
the Rogowski Coil. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
3.1.2 Pulsed Voltage Measurements . . . . . . . . . . . . . . . . . . . . . 55
3.1.3 Interpretation of the Current and Voltage Waveforms. . . . . 59
3.2 Pulsed Magnetic Field Measurement . . . . . . . . . . . . . . . . . . . . . 61
3.3 Plasma Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
3.3.1 Plasma Radiation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
3.3.2 The Plasma Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
3.3.3 Examples of Population Density Distribution
and Plasma Spectra . . . . . . . . . . . . . . . . . . . . . . . . .... 68
3.3.4 Optical Emission Spectroscopy of Plasma. . . . . . . . . .... 70
3.4 The Langmuir Probe (Electric Probe) . . . . . . . . . . . . . . . . . .... 74
3.4.1 Electron Temperature and Density Measurements
of Steady State Plasmas . . . . . . . . . . . . . . . . . . . . . .... 74
3.4.2 Dynamic Studies of a Flowing Plasma
(or Charged Particle Beam). . . . . . . . . . . . . . . . . . . . . . . 76
3.5 X-ray Diagnostic Techniques. . . . . . . . . . . . . . . . . . . . . . . . . . . 78
3.5.1 X-ray Absorption Filter . . . . . . . . . . . . . . . . . . . . . . . . . 79
3.5.2 X-ray Detectors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
3.6 Neutron Diagnostic Techniques . . . . . . . . . . . . . . . . . . . . . . . . . 91
3.6.1 The Foil Activation Neutron Detector for Absolute
Neutron Yield Measurement . . . . . . . . . . . . . . . . . . . . . . 91
3.6.2 The Detector Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
3.6.3 Time-Resolved Neutron Pulse Measurement . . . . . . . . . . . 96
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
4 Some Examples of Small Plasma Devices . . . . . . . . . . . . . . . . . ... 99
4.1 The Electromagnetic Shock Tube. . . . . . . . . . . . . . . . . . . . . ... 99
4.1.1 Numerical Modeling of the Electromagnetic
Shock Tube Dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . 100
4.1.2 Experimental Measurements of Electromagnetic
Shock Tube Dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . 103
4.2 The Plasma Focus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
4.2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
4.2.2 Characteristics of the Plasma Focus Discharge . . . . . . . . . 105
4.2.3 Design Consideration . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
4.2.4 X-ray Emission from the Plasma Focus Discharge . . . . . . . 108
4.2.5 Neutron Emission . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
4.2.6 Particle Beam Emission . . . . . . . . . . . . . . . . . . . . . . . . . 110
4.3 The Vacuum Spark . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
4.4 Scaled-Down Operation of Vacuum Spark—Flash X-ray Tube . . . 112
4.5 The Pulsed Capillary Discharge . . . . . . . . . . . . . . . . . . . . . . . . . 113
Contents xi
4.6 The 50 Hz Alternating Current (A.C.) Glow

Discharge System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
4.7 Atmospheric Pressure Dielectric Barrier Discharge . . . . . . . . . . . . 118
4.8 Wire Explosion System for Nano Powder Fabrication . . . . . . . . . 119
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
Chapter 1
Basic Concepts in Plasma Technology
Abstract Some of the fundamental basic concepts useful for the understanding of
plasma will be introduced and explained in this chapter. These include particle
collision and the fundamental processes that may occur as a consequence of col-
lision between particles, the concept of Debye shielding, plasma sheath formation at
the surface of object placed inside plasma, particle oscillation. The criteria of
plasma and the particle transport due to electric and magnetic fields as well as
density gradient will be discussed briefly.
Keywords Plasma Plasma properties
1.1 Plasma—The Fourth State of Matter
In the simplest term, we may say plasma is the fourth state of matter. At room
temperature, there are matters that exist in the solid state, some in the liquid state
and some in the gaseous state. Each kind of matter can also exist in all the three
states if it is heated to temperature above room temperature and cooled to below
room temperature. For example, water is in the liquid state at room temperature and
atmospheric pressure. If water is cooled to below 0 °C, much lower than room
temperature, it will turn into ice—the solid state of water. On the other hand, when
water is heated to 100 °C, it turns into its gaseous state—steam (Fig. 1.1).
Now consider what will happen if the steam is heated to temperature much
higher than 100 °C. Some of the water molecules (H2O) may be dissociated into
hydrogen and oxygen atoms. If the temperature is high enough, the hydrogen and
oxygen atoms may even be ionized to form positive ions and electrons. This
“ionized” state of water is the fourth state of matter, or it is called the plasma state if
it satisfies certain criteria. We will discuss about these criteria in more details later.
Inside the plasma, there are electrons, ions of various charge states, neutral atoms
and molecules. These particles move around inside the plasma with kinetic energy.
The particles may exchange energy when they collide with each other. The collision
can be either elastic or inelastic.
© The Author(s) 2016 1

C.S. Wong and R. Mongkolnavin, Elements of Plasma Technology,
SpringerBriefs in Applied Sciences and Technology,
DOI 10.1007/978-981-10-0117-8_1
2 1 Basic Concepts in Plasma Technology
PLASMA
GAS
SOLID
LIQUID
HEAT
Fig. 1.1 Plasma—the fourth state of matter
During an elastic collision, the particles only exchange kinetic energy. The
magnitudes and directions of the velocities of both colliding particles may be
changed after the collision. However, during an inelastic collision, the internal
energy of the colliding particles may be changed. This leads to the occurrence of
various types of processes such as excitation and ionization. In particular, ionization
gives rise to the production of new charged particles and hence causing the number
of charged particles to increase.
1.2 Collision
1.2.1 Elastic Collision
Consider the case of collision between 2 spherical particles with mass m1 and m2
respectively (Fig. 1.2).
1.2 Collision 3
Fig. 1.2 Collision between m1 u1 sinθ

particles
θ u1
v1
u2 = 0
m2
v2
Assume that before collision, m2 is at rest while m1 is moving with velocity u1.
When they collide, at the point of contact, u1 makes an angle of θ with the line
joining the centers of the two colliding particles. After the collision, the velocity of
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
m1 becomes v21 þ u21 sin2 h, while m2 will now start moving at a velocity of v2 in
the direction along the line joining the centers of the two colliding particles.
Considering conservation of momentum before and after the collision,
m1 u1 cosh ¼ m1 v1 þ m2 v2
m1 u1 sinh unchanged
m2
v1 ¼ u1 cos h v2
m1
Similarly, considering conservation of energy before and after the collision,
1 1 1
m1 u21 ¼ m1 v21 þ u21 sin2 h þ m2 v22
2 2 2
v2 2 2m1 cos h 2
ð Þ ¼ð Þ
u1 m1 þ m2
The fraction of energy transferred during the collision:

δ = kinetic energy of m2 after collision/initial kinetic energy of m1
1
m2 v22 m2 v22 m2 4m21 cos2 h 4m1 m2
¼ 12 ¼ ¼ ¼ cos2 h:
2 m 1 u1
2 m1 u1 m1 ðm1 þ m2 Þ
2 2
ðm1 þ m2 Þ2
Consider
ðiÞ m1 ¼ m2 ) d ¼ cos2 h; dmax ¼ 1

ðiiÞ m1 m2 ; ) m1 þ m2 m2 :
m1 4m1
)d4 cos2 h; hence dmax ¼ 1:
m2 m2
Since m1 m2 , the value of δ is always small. This is the case for electron-atom
or electron-ion collision. We see that with elastic collision, the tracks of the electrons
will be deflected when they collide with atoms or ions, with only a small fraction of
their energy is transferred to the atoms or ions. In other word, they will be scattered.
This is what happen for a gas at low or not so high temperature. When the tem-
perature is increased, the particles are moving with higher kinetic energy and they
will collide harder. When the collision is sufficiently hard, energy will be transferred
into the internal energy of the target particle and the collision is said to be inelastic.
1.2.2 Inelastic Collision
For inelastic collision, the momentum is still conserved as in the case of elastic
collision, where
m1 u1 cosh ¼ m1 v1 þ m2 v2
m1 u1 sinh unchanged:
But the energy equation has to be written as
1 1 1
m1 u21 ¼ m1 ðv21 þ u21 sin2 hÞ þ m2 v22 þ DU
2 2 2
where ΔU is the energy transferred into the internal energy of m2 during the
collision.
m2 ðm1 þ m2 Þ v22
ΔU can be shown to be: DU ¼ m2 u1 v2 cos h 2m1 :
Differentiate this with respect to v2 gives
dðDUÞ m2 ðm1 þ m2 Þ
¼ m2 u1 cos h ð2v2 Þ
dv2 2 m1
For ΔU = maximum, let d(ΔU)/dv2 = 0

m1
) v2 ¼ ð Þu1 cos h
m1 þ m2
1 m1 m2
)ðDUÞmax ¼ ð Þ u2 cos2 h
2 m1 þ m2 1
The maximum energy transferred to the internal energy of the target particle is
given by:
m2
d¼ð Þ cos2 h:
m1 þ m2
For m1 m2 ; d ¼ cos2 h; dmax ¼ 1.

1.2 Collision 5
This means that during an inelastic collision, all the energy of the electron
(colliding particle) can be transferred to the internal energy of m2, which may be ion
or atom.
With the increase in the internal energy, the atom or ion may undergo changes in
its electronic configuration leading to processes such as excitation and ionization.
1.3 Collision Cross-Section
In the consideration of collision between electron and atom or ion an important

question to ask is: What is the probability that a collision will occur? And subse-
quently, what is the probability that the collision will lead to a particular process?
This is expressed in terms of the cross-section. For collision between electron and
atom, the cross-section can be simply defined by the size of the atom. Hence the
cross-section for electron-atom collision is given by rea ¼ pa2o , where ao is the
radius of the atom. For collision between charged particles, because of the coulomb
effect, the particles may be able to exchange energy without physical contact. we
may define an impact parameter b which is the distance of closest approach between
the colliding particles. In the particular case when the incoming particle is deflected
at 90° with respect to its original path, the coulombic potential energy between
them (attractive for e-i or repulsive for e-e and i-i collisions) at the point of closest
approach can be shown to be equal to twice the kinetic energy of the incoming
particle at infinity. That is,
Z1 Z2 e2 1
¼ 2 mu21
4peo bo 2
where bo is the distance of closest approach for 90° deflection. This gives
Z1 Z2 e2
bo ¼ ;
4peo ð3kT Þ
where we have taken 12 mu21 ¼ 32 kT. For collision between charged particles, the
cross-section is written as r ¼ pb2o . Hence
2
Z1 Z2 e2
r¼p :
4peo ð3kT Þ
This expression, however, applied only to a slightly ionized plasma. For fully or
strongly ionized plasma,
2
Z1 Z2 e2
r ¼ 5:7‘nKp ;
4peo ð3kT Þ
where ‘nK 10 for fully or strongly ionized plasma.

With reference to the cross-section, we can define the mean free path of collision
which is the average time interval between 2 consecutive collisions as k ¼ n12 r and
the mean free time as s ¼ n2 u11 r where n2 is the number density of the target particles
and u1 is the velocity of the incoming particle.
1.4 Fundamental Processes in a Plasma
As a consequence of collisions between electron and the atom or ion inside the
plasma, various processes may occur. Four of the most elementary processes are:
i. Scattering e þ A ! A þ e
This is caused by elastic collision where the colliding electron will transferred a
small fraction of its kinetic energy to the atom or ion. However, its direction of
motion will be changed.
ii. Excitation e þ A ! A þ e
This occurs when electron of sufficient energy collides at an atom or ion
inelastically. Part of its kinetic energy is absorbed by an inner shell electron of
the atom or ion so that this inner shell electron is raised to a higher energy level
and the atom or ion thus becomes excited.
Most of the excited states have short life time and they will decay back to
their original levels by emitting a photon equivalent to the energy difference.
This process is called de-excitation or relaxation by spontaneous emission.
iii. Ionization e þ A ! A þ þ 2e
With sufficiently high energy, the colliding electron may transfer enough
energy into the internal energy of the atom or ion to release one of the bound
electrons. The atom or ion thus becomes one charge state higher and it is said to
be ionized.
It is through the ionization process that new charged particles (electrons) are
produced in the plasma.
iv. Recombination e þ A ! A þ ðhmÞ
In this case the electron colliding with the ion may be captured and it occupies
the vacancy inside the ion to change its charge state to one level lower than
previously. A photon may be emitted when the electron releases its excess
energy.
The question of which process is probable can be answered by referring to
the cross section of the process which is expressed as probability of its
occurrence at various energy of the colliding electron. An example of the cross
sections for processes (i), (ii) and (iii) for argon gas is shown in Fig. 1.3.
1.5 Some Consideration of Thermodynamic Properties of Plasma 7
Fig. 1.3 Relative cross-sections of scattering, excitation and ionization processes
1.5 Some Consideration of Thermodynamic Properties

of Plasma
Let us consider a gas at sufficiently low temperature when it can be considered to

consist of predominantly neutral particles. The types of species present may consist
of atoms (and/or molecules), ions and electrons. Under such a condition, assuming
thermodynamic equilibrium, the ideal gas law may be applied. If we define αi to be
the fractional population of an ionic specie so that
Ni X
Z
ai ¼ where Nt ¼ Nj :
Nt j¼0
Hence
No
ao ¼
Nt
is the fractional population of neutral particles, and
N1
a1 ¼
Nt
is the fractional population of singly ionized species and so on. For the gas at low
temperature (say near to the room temperature), we expect
a0 1; a1; a2; 0
From kinetic theory, the pressure p (specific) of the gas is related to the tem-
perature T and particle number density n as
N
p ¼ nkT ¼ kT ¼ qRT;
V
where V is the volume of the gas, ρ is the mass density, and R is the gas constant.
This is the ideal gas law. For an ideal gas, the specific heat ratio γ is given by
f þ2 5
c¼ ¼ ;
f 3
since an ideal gas is assumed to have three degree of freedom (f = 3). When the gas
is heated up (for example by electrical discharge), the kinetic energy of the particles
inside the gas increases and their collisions will also be increasingly “hard” and
hence inelastic. This is particularly effective when the colliding particles are elec-
trons, as we have discussed earlier, although their fractional population is relatively
small. This gives rise to transfer of energy into the internal energy of the target
particles which may be atoms or ions. There is thus an increase in the degree of
freedom which means f is getting larger. γ will then drop and approaching unity.
With the occurrence of inelastic collision, the cross section for ionization also
increases and new charged particles (ions and electrons) will be produced. At a
particular temperature, only a certain number (not more than 5 or 6) of species will
be dominant. The distribution of fractional populations of the various species in a gas
heated to elevated temperature can be described by either the Local Thermodynamic
Equilibrium (LTE) Model or the Coronal Equilibrium (CE) Model. The choice of
which model is appropriate depends on the density of the gas. In general, LTE model
is more appropriate for a high density plasma while the CE model is more suitable
for a low density plasma. It is often difficult to decide on what density is high enough
to be considered “high”. We will come back to this topic again when we consider the
radiation emission spectrum of the plasma.
The fact that when a gas is heated to higher temperature, inelastic collisions will
lead to processes such as excitation and ionization (hence increasing degree of
freedom) implies that the gas begins to deviate from ideal gas condition. Hence the
conversion of a gas from the neutral state to the ionized state (plasma) is converting
it into a real gas. For plasma taken as a real gas, the appropriate form of the equation
of state is
1.5 Some Consideration of Thermodynamic Properties of Plasma 9
p ¼ qRTz;
where z is the Departure Coefficient (departure from ideal gas), and R is the gas
constant.
For a plasma with ionic species α0, α1, α2… αi,
X
i
z ¼ 1þ jaj for atomic gas;
j¼1
X
i
z ¼ 1þvþ ½ð2j þ 1Þaj for diatom molecular gas;
j¼1
where χ is the fractional population of molecules.

When the plasma is heated to extremely high temperature until it is fully ionized
(that means all electrons are stripped off from all the atoms), processes such as
excitation and ionization will not occur anymore and the number of degree of
freedom reduced back to 3 only. In this case the plasma consists of only 2 types of
species, fully stripped ions and electrons and it can be considered to be ideal gas
again.
1.6 Concept of Plasma Potential
One of the important characteristics of plasma is that charge neutrality is assumed to

be maintained. This means that there is a balance of positive and negative charges.
If we assume that the charge density is zero everywhere inside the plasma, then
there should be no electrostatic field. However, the potential is not zero inside the
plasma, although it is uniform.
The fact that the plasma potential is not zero is evident from two consequences:
1. When a stray point charge is dropped into the plasma, the charged particles
inside the plasma will re-distribute themselves so as to shield the effect of the
stray charge to wihin a small sphere (the so called Debye sphere) with the stray
charge as the center having radius of a Debye length given by
rffiffiffiffiffiffi
Te
kD ¼ 6:9 cm ½Te : K ; ne : cm3
ne
This effect is called the Debye Shielding. This is determined by the electrons
inside the plasma because it is usually the electrons that will re-distribute to
produce the screening effect. When the stray charge is positive, the electrons
will move towards the charge so that the population of electrons inside the
Debye sphere is higher than that of the ambient. On the other hand, if the stray
charge is negative, the population of electrons inside the Debye sphere will be
lower than the ambient.
2. Imagine now that the stray charge is suddenly removed from the plasma. The
particles will try to re-distribute back to the original situation and as a conse-
quence they may be set into simple harmonic oscillation. Taking analogy of a
mass m hanging on a spring. If the mass is pulled downward to make a dis-
placement and then let go, the mass will tend to go back to its equilibrium
position but it will not stop at the equilibrium position unless the system is
critically damped. It will be set into oscillation (simple harmonic motion)
instead. To better visualize the effect in a plasma we can refer to a static situation
where all the charged particles are aligned in such a way a achieve charge
neutrality. Taking the simplest case of 1-D, we have an infinite array of alternate
positive and negative charges of equal magnitude. If an electron is displaced and
then released, it will start to oscillate about its equilibrium position in simple
harmonic motion with frequency given by
pffiffiffiffiffi
fe 9 103 ne Hz ½ne : cm3
Note that fe is only dependent on the density of electrons.
1.7 Criteria of Plasma
In a very broad term, we say a plasma is an ionized gas. Actually an ionized gas has
to fulfil several criteria in order to be qualified as a plasma.
Criterion 1: kD L
Here L is the “characteristic” dimension of the plasma. This criterion requires that if
the plasma is perturbed by a stray charge, the effect of the stray charge should be
shielded to within a distance of one Debye length which is much smaller than the
characteristic dimension of the plasma.
Criterion 2: ND ¼ 43 pk3D N
1
The number of particles inside the Debye sphere ND must be sufficiently large, at
least much greater than unity, say 100. This requires that the particle number
density of the plasma is sufficiently high.
Criterion 3: xp s [ 1
xp ¼ 2pfe is the angular frequency of electron plasma oscillation while τ is the

electron-atom or electron-ion collision mean-free-time. In other words, the plasma
1.7 Criteria of Plasma 11
oscillation frequency must be higher than the collision frequency. This implies that
between 2 collisions, the electron must have performed many oscillations. This is to
ensure that the particle is able to reach equilibrium after each collision.
1.8 Effect of Boundary in Plasma
So far we have assumed a borderless and uniform plasma. What happen at the
boundary of the plasma such as the chamber wall containing the plasma; or at the
surface of an extended object (not a point charge) placed inside the plasma?
Let’s say we start from an equilibrium situation where the plasma is uniform and
homogenous. When an object (floating and not connected to anyway) is placed
inside the plasma, the plasma will react and since the electrons are much lighter
than the ions, they will reach the surface of the object first. This will cause the
surface of the object to be “charged” up to a negative potential. However, as more
electrons reach at the surface, the negative potential will also increase and tends to
push away any subsequent electrons coming towards the surface. An equilibrium
will soon be reached when the potential distribution in front of the surface becomes
distorted from the plasma potential Vp as shown in Fig. 1.4. The potential at the
surface of the object Vf is negative with respect to the plasma potential and it is
often referred to as the floating potential of the object.
As a first approximation, it can be assumed that the potential will return to the
plasma potential at a distance of λD (the Debye length) from the surface. In other
words, a plasma sheath of thickness λD will be formed at the surface of the object
placed inside the plasma.
The above simplified picture of plasma sheath formation is obtained from a static
charge distribution model. Since the particles are moving around inside the plasma
randomly, the ions will only be accelerated in the direction towards the surface
when they get sufficiently close to the surface. Beyond that point, their velocities
will gradually increase until
Fig. 1.4 Potential

distribution at boundary in Sheath of space charge ρ
plasma—first approximation
Vp
x Vf
Fig. 1.5 Potential x=0

distribution at boundary in
plasma—the Bohm model
Vp
V(0)
PLASMA
Quasi-Neutral
Transition Region
Vf
Positive Charge Region
rffiffiffiffiffiffiffi
kTe
uo [ :
mi
This is the Bohm criterion of the plasma sheath. A more accurate picture of the
plasma sheath structure is shown in Fig. 1.5.
The ions are moving randomly in the unperturbed plasma. When they reach the
quasi-neutral transition region they start to be accelerated towards the surface of the
object. Beyond the point x = 0 in Fig. 1.5, all ions have achieved a velocity
qffiffiffiffiffi
exceeding kT e
mi . This is considered to be the boundary of the plasma sheath formed
on the surface of the object.
1.9 Particle Transport Inside Plasma
So far we have considered charged particles in the plasma to be moving randomly

due to their thermal velocity. How will the particles move under the influence of
externally applied forces?
We will consider here two types of external forces that may influence the motion
of the particles inside a plasma.
i. Motion of charged particles in an electric field
For charged particles (specifically electrons) moving in vacuum under the
influence of an electric field, we can write the equation of motion as (this is
sometime referred to as Lorentz Model):

dve et
me ¼ eE which gives ve ¼ E:
dt me
1.9 Particle Transport Inside Plasma 13
Inside the plasma, due to collision, the motion of electrons will be affected.
Instead of moving straight in the direction of the electric field as in the case of
vacuum, the electron may change direction each time it collides with another
particle. However, the overall direction of the electron is still expected to follow the
direction of the electric field. The electrons are said to be drifting and the electron
drift velocity will be proportional to the electric field. The electron drift velocity can
be obtained by considering its motion between two consecutive collisions to be
same as that of the vacuum case and taking the average effect over many collisions.
Hence the electron drift velocity can be expressed as

esea
ve ¼ E;
me
where τe-a is the electron-atom collision mean

free time, the average time in
esea
between two collisions. The quantity le ¼ me is called the electron mobility.
This is in fact a measure of how the motion of the electron will be affected by
collision. If the frequency of collisions is high, hence τe-a is short, then the electrons
will have low mobility.
If we write the electron current density to be Je ¼ ne eve ¼ ne ele E and compared
to Ohm’s Law for electrical conduction Je ¼ rE, then we see that we can define an
equivalent electrical conductivity of the plasma as
ne e2 sea
r ¼ ne ele ¼ ;
me
Again, more frequent collision means lower conductivity.

Similarly consideration can be given to the ions. The direction of the drift will be
opposite to that of electrons. However, due to the massiveness of the ions as
compared to electrons (mp/me * 103), vi ve under the influence of the same
electric field. Hence in the consideration of particle transport in an electric field, it is
common to assume the ions to be stationary.
ii. Motion of particles due to density gradient
The presence of density gradient inside the plasma may also change the motion
of the particles from random to drifting in a particular direction. This phe-
nomenon is called diffusion. The drift velocity created is directly proportional
to the density gradient ∇n but is inversely proportional to the density n itself.
Hence we write
rn
vd ¼ D :
n
The minus sign is to show that the direction of the diffusion is opposite to the
density gradient, that is from high density region to low density region.
Under the effect of density gradient, both electrons and ions will move in the
same direction. However, due to the massiveness of the ions, they will lag behind
and this results in charge separation. An induced electric field will be produced.
This field is in such a direction to accelerate the ions but retard the electrons.
Imagine that both electrons and ions are originally at the same point and start to
diffuse outwards (spherical geometry) together. First the electrons will move much
faster than the ions but soon charge separation occurs and the induced electric field
will act to slow down the electrons but to accelerate the ions. An equilibrium will
soon be established when the drift velocity of electrons and ions become the same.
This situation is termed ambipolar diffusion.
Chapter 2
Methods of Plasma Generation
Abstract In this chapter, the generation of plasma by gaseous electrical discharge

will be discussed. Townsend Theory of electrical breakdown of gases will be
explained. Various types of discharge, including to corona discharge, glow dis-
charge and arc discharge and the characteristics of the plasmas produced will be
introduced. The electrical power sources used for the generation of these plasma
including DC, AC, RF, microwave and pulsed capacitor discharge are introduced.
Keywords Electrical discharge Pulsed discharge
2.1 DC Electrical Discharge
2.1.1 Electrical Breakdown
In its neutral form, a gas is an insulator and will not conduct electrical current when
an electric field is applied across it, no matter how high is that electric field.
However, some amount of stray charges are always present in the neutral gas. One
such source of stray charges is the ionization of the gas particles by cosmic ray or any
other background radiation from the environment. If the electric field is applied to
the gas using electrodes, another source of stray charges (electrons) may come from
the photoelectric effect at the cathode surface due to the absorption of UV photons.
The presence of stray electrons is crucial in electrical discharge since they can be
accelerated to high energy to produce ionizing collision and hence new charged
particles. Without the presence of stray electrons, electrical discharge may not occur.
The simplest configuration to produce an electrical discharge through a gas is to
apply a potential difference across a pair of parallel electrodes placed inside a
chamber filled with the gas at a suitable pressure as shown in Fig. 2.1.
Consider an electron originated from the cathode due to the absorption of UV
photons. With the presence of electric field, the electron will be accelerated to high
enough energy for excitation or ionization when it collides with an atom. When an
ionizing collision occurs, the colliding electron together with the new electron will
be further accelerated by the electric field and they may produce further ionizing

DOI 10.1007/978-981-10-0117-8_2
16 2 Methods of Plasma Generation
Anode
I
N N N N
High Voltage
Power Supply
N N
V
UV
Cathode
Fig. 2.1 Multiplication of charged particles leading to electrical discharge
collisions. Hence we see that the number of electrons will be multiplied as we

progress from the cathode to the anode. This multiplication of charges (electrons) is
described by the Townsend theory of gas discharge. The number of ionizing col-
lisions per unit distance is called the First Townsend Coefficient, α. The relation of
α with the gas pressure p and the applied electric field E have been established
empirically for various gases. This is expressed as [1]:
rffiffiffiffi
a p
¼ C exp D ð2:1Þ
p E
for atomic gases, and

a p
¼ A exp B ð2:2Þ
p E
for molecular gases.

The constants A, B, C and D are constants of the gas used. For example, for
argon gas C = 29.22 cm−1 torr−1, D = 26.64 (V cm−1 torr−1)1/2; while for air
A = 14.6 cm−1 torr−1, B = 365 V cm−1 torr−1.
Considering the process of ionizing collision by a flux of electrons originated
from the cathode produced by UV absorption, and assuming that α is constant at
constant E and p, Townsend’s theory predicts that the electron flux will grow
according to the simple relation of
FðxÞ ¼ Fo expðaxÞ;
where Fo is the electron flux at the cathode surface expressed in terms of number of
electrons per unit area.
2.1 DC Electrical Discharge 17
For an anode-cathode distance of d, the electron flux finally collected by the

anode is
Fa ¼ Fo expðadÞ:
The anode current density measured is Ja ¼ eFa ¼ Jo expðadÞ. This can also be
written in terms of the total current Ia if we assume that the cross-section of the
discharge current remains the same. That is
Ia ¼ Io ead ð2:3Þ
This is reasonably well obeyed by experimental data, as illustrated in the plot of

ln Ia versus d as shown in Fig. 2.2.
The straight line does not pass through the origin, but intercept the y-axis at (ln
Ia − αdo). This is due to the fact that the electrons must be accelerated for a distance
of at least do before they acquire sufficient energy to produce ionizing collision. In
that case the expression for the anode current density should be corrected as
Ia ¼ Io exp½aðd do Þ. From Fig. 2.2, it can be seen that from the measurements of
Ia at various electrode separation d, the value of α for the gas can be determined
from the gradient of the straight line obtained. Note that for this experiment, when
d is increased, the potential applied across the electrodes V must also be increased in
order to keep E = V/d constant. Hence the value of α obtained is for a particular E.
Figure 2.3 shows an example of the values of α/p as a function of E/p obtained
experimentally for argon.
The plot can be converted to α/p as a function of (p/E × 1000)1/2 as shown in
Fig. 2.4. The values of the constants C and D can then be determined from this plot.
In the above consideration, the electrons produced by ionization in the gas are
assumed to be accelerated to the anode, leaving behind the heavy ions. For ions
near to the cathode surface, they may collide at the surface and for sufficiently high
electric field, the collision at the cathode surface by the ions may be energetic
Fig. 2.2 Plot of lnIe against Current (pA)

inter-electrode distance d e5 Effect of γ
e4
e3
Gradient = α
e2
lnIo - α do
e1
e0
e-1
0 2 4 6 8 10 12
d (mm)
Fig. 2.3 Relationship 10

between α/p and E/p
/p (ionizations/cm/torr)
1
0.1
0.01
1E-3
1E-4
1 10 100 1000
E/p (v/cm/torr)
Fig. 2.4 Plot of α/p against 10

p/E
/p (ionizations/cm/torr)
0.1
0.01
1E-3
1E-4
2 4 6 8 10 12 14 16 18 20
1/2 1/2
[p/E x 1000] [(torr-cm/V)]
enough to overcome the work function of the cathode material and releases elec-
trons from it. The number of electrons released from the surface from each ion
bombardment is called Townsend’s Second Coefficient γ. This acts as an additional
source of electrons which should be considered in Townsend’s theory of gas dis-
charge. The anode current is now given in the form
Io expðadÞ
Ia ¼ : ð2:4Þ
1 cðead 1Þ
Notice that for small d, exp(αd) * 1 so that the denominator will be *1. This is
observed experimentally. For small d, the plot of ‘n Ia against d is a straight line. As
d is increased to sufficiently large value, the straight line will begin to curve
upwards, deviating from the straight line.
The release of electrons from the cathode surface is crucial in gas discharge. It is
responsible for making the discharge “self-sustained” and do not rely on external
source such as UV light.
From Eq. (2.4), if the denominator tends to zero, that is

1 c ead 1 ! 0
then Ia → ∞. This is the condition that an electrical breakdown has occured.

A self-sustained discharge is obtained. The criterion that

1 c ead 1 ¼ 0 or c ead 1 ¼ 1 or c ead ¼ c þ 1
1 ð2:5Þ
or ad ¼ ‘n 1 þ
c
is referred to as the breakdown criterion of electrical discharge.

From the breakdown criterion, substitute for α/p from Eq. (2.1) for atomic gas,
and writing E ¼ VdB at electrical breakdown, it can be shown that the breakdown
voltage can be expressed as
D2 pd
VB ¼
ð2:6Þ
‘n ‘n Cpd
ð1 þ 1cÞ
for atomic gas, and similarly

Bpd
VB ¼
ð2:7Þ
‘n ‘n Apd
ð1 þ 1cÞ
for molecular gas.

γ is actually also a function of E/p, similar to α. It has been found experimentally
that γ is in fact fairly constant for most discharge conditions commonly encoun-
tered. Hence it can be seen that the breakdown voltage is a function of (pd) for both
atomic and molecular gases. This has been demonstrated experimentally. From
experiment, a minimum VB is observed for a fixed value of (pd) for a certain
discharge condition (gas type and electrode material). This minimum breakdown
can be derived by differentiating (2.6) or (2.7) with respect to (pd) to be

7:39‘n 1 þ 1c
D2 1
ðVB Þmin ¼ 7:39‘n 1 þ at ðpdÞmin ¼ ð2:8Þ
4C c C
for atomic gas

2:72‘n 1 þ 1c
B 1
ðVB Þmin ¼ 2:72‘n 1 þ at ðpdÞmin ¼ : ð2:9Þ
A c A
for molecular gas

Fig. 2.5 The Paschen curve for air
The plot of breakdown voltage VB against pd is often referred to as the Paschen

Curve or the Paschen Law. The example for air is shown in Fig. 2.5.
It can be seen from the Paschen Curve that on the left hand side of the VB
minimum, the effect of increasing pd is to lower the breakdown voltage, while on
the right hand side of the VB minimum, the trend is just opposite. Increasing pd will
now makes it more and more difficult for breakdown to occur.
2.1.2 The I–V Characteristic of Electrical Discharge
The variation of the current flowing in the electrical discharge circuit shown in
Fig. 2.1 with applied voltage Vs can be summarized by the I–V characteristic curve
as shown in Fig. 2.6. In this curve, the vertical axis is the voltage drop across the
discharge tube.
The first part of the characteristic is caused by charges produced by background
ionization of the gas either by environmental stray radiation or by photoelectric
effect at the cathode surface due to UV radiation. At low voltage, whatever elec-
trons available may be accelerated towards the anode to constitute to the current. If
no ionizing collision by electron can occur due to the low potential (and hence low
electric field), the maximum current that can be obtained is determined by the total
number of initial electrons available. This current is in the region below
nano-ampere and it increases with applied potential. It reaches a saturation value
corresponding to the maximum number of electrons available.
With increasing applied potential, the electrons may be accelerated to energy
above the excitation and ionization thresholds and these processes will then take
place. New charge particles, both ions and electrons, will be produced by ionization
and this gives rise to an increase in the discharge current. Eventually as the potential
is further increased to reach the breakdown voltage, the discharge current will
Fig. 2.6 I–V characteristic of electrical discharge
increase exponentially and then breakdown will occur and an electrical discharge is
formed. This electrical breakdown will occur at potential VB given by (2.6) or (2.7).
The region of the I–V curve before breakdown is often referred to as the dark
discharge region. It is subdivided into the background current region, the Townsend
region, and the corona region. A corona discharge is maintained by controlling the
current at the micro amperes region, either by external or internal means. We will
look into this again later. The voltage drop across the discharge tube when the
discharge is in the dark discharge region is roughly equal to the applied potential.
After breakdown, the discharge will try to draw infinite current from the power
supply so it is essential to have a current limiting resistor RL in series between the
source and the discharge. The type of discharge that is obtained will depend on the
magnitude of the discharge current, which is controlled by the combined effect of
the limiting resistor RL and the plasma impedance. Ideally, the plasma resistance is
negligible compared to RL after breakdown. This means that the voltage drop across
the discharge tube will be zero and the full voltage will be developed across RL.
However, when RL is adjusted to limit the current to be in the region of mA, there
will be a voltage drop across the discharge tube which is roughly constant when the
discharge current is varied. This is the normal glow discharge region and the
voltage across the electrodes is called glow voltage, Vg. The normal glow region
may be extended down to 10−5 A when the current is reduced gradually from mA.
On the other hand, when the current is further increased to beyond 100 mA, the
voltage across the electrodes will not remain constant but will increase. The glow
discharge is said to become abnormal. When the current is increased to greater than
1 A, the voltage across the electrodes suddenly drops to lower than the glow voltage
and the discharge has changed into the arc discharge. As a summary, the three types
of discharge that can be obtained by controlling the current are:
107 105 A ) Corona discharge

105 1 A ) Glow discharge
[ 1 A ) Arc discharge
2.1.3 The Corona Discharge
After breakdown, if the discharge current can be controlled at the level of several
μA, a corona discharge will be obtained. The potential drop across the electrodes is
still the same as the applied voltage. Corona discharge can also be obtained in sit-
uation where electrical breakdown voltage has not been reached but the electric
field between the electrodes is not uniform. A particular situation is when the high
voltage electrode (can be either anode or cathode) has a sharp profile, such as in the
form of a needle or thin wire. In this case the electric field at the sharp point is
sufficiently high (>30 kV/cm) that the electrons may be accelerated to high enough
energy to produce ionizing collision leading to breakdown within a close distance
from the sharp point. The distance from the sharp point within which electrical
breakdown can occur is called the effective distance of corona discharge.
A phenomenological situation of corona discharge is illustrated in Fig. 2.7. With
such a configuration, the Poisson Equation in spherical coordinate is written as:
1 d 2 q
rE ¼ ðr EÞ ¼ 0 ð2:10Þ
r 2 dr eo
Fig. 2.7 Corona discharge at a sharp point

which gives the radial profile for the electric field between the electrodes expressed
as:
dV a2 Eo
EðrÞ ¼ ¼ 2 : ð2:11Þ
dr r
Integrating with the boundary conditions that at r = a, V = Vo and at r = b, V = 0;

we get,

aðb rÞ
VðrÞ ¼ Vo : ð2:12Þ
rðb aÞ
Hence,
abVo aVo
EðrÞ ¼ 2 when a b: ð2:13Þ
r 2 ðb aÞ r
At the surface of the sharp point with radius a, the electric field is Eo Vao which
can be significantly large for small a. This may also be true for small r.
2.1.4 The Glow Discharge
The glow discharge is one of the most commonly used plasma in industry.
Although it can be produced by a large variety of discharge configurations: DC, RF,
DC or RF magnetron, ECR microwave discharge etc., the basic properties of the
plasmas produced in these discharges are similar. In the classical configuration of
the plane parallel electrodes placed inside a cylindrical glass chamber as shown in
Fig. 2.8, operated in the normal glow discharge mode, the discharge consists of
several bright and dark regions. The most prominent part of the discharge, which is
the plasma proper by definition (zero or at least low electric field, charge neutrality
etc.), is the positive column. In general, the electron temperature of this plasma is in
the range of 1–2 eV while the ions and atoms are near to room temperature, and the
electron density in the range of 106–108 cm−3 in the DC discharge case. With such
conditions, the majority of the species present is in the neutral state, perhaps largely
excited. A small fraction of singly ionized ions, and even smaller fraction of doubly
ionized ions may also be present. According to Scottky’s Diffusion Model where
the production of new electrons by ionization is balanced by the loss of charge
particles due to radial diffusion, the radial electron density profile in the positive
column is expressed in terms of the zeroth order Bessel Function
Fig. 2.8 Characteristic

features of a normal glow
discharge
r
nðrÞ ¼ no Jo 2:405 ; ð2:14Þ
R
where no is the electron number density at the axis of the column and R its radius.
Another region of interest is the cathode fall region consisting of the Aston Dark
Space, Cathode Glow and the Crookes Dark Space. The potential drop sharply
across this region resulting in high electric field. The high electric field is the main
cause of energetic ionic bombardment at the surface of the cathode that results in
electron emission. This is in fact an essential feature of the discharge that makes it
“self-sustained”. The cathode fall region is also the region where the electrons gain
most of their kinetic energies. The thickness of the cathode fall region, dc is rela-
tively small compared to the length of the whole discharge tube. The potential drop
across this region, on the other hand, can be shown to be given be

B 1
Vc ¼ 3‘n 1 þ for a molecular gas: ð2:15Þ
A c
This can be compared with the expression for the minimum breakdown voltage
for a discharge in molecular gas given by Eq. (2.9). This means that the potential
drop across the cathode fall region for a gas discharge is expected to be about the
same as the minimum breakdown voltage for the setup, or vice versa. Similarly, for
an monoatomic gas,

D2 1
Vc 7:39‘n 1 þ : ð2:16Þ
4C c
The other two regions, the negative glow and the Faraday Dark Space, are
regions where the energetic electrons are thermalized before reaching the positive
column. At the anode side, electrons near to the anode may experience slight
increase in acceleration and bombard at the anode surface to produce a layer of
bright region, the anode glow. Separating the positive column and the anode glow is
the anode dark space.
2.1.5 Hot Cathode Discharge
In the glow discharge described in Sect. 2.1.4, no external means of heating is

applied to the cathode. It is thus often referred to as a “cold cathode” discharge,
although the electrodes do become hot, or at least warm during the discharge. We
say it is cold is to contrast it with the case where the cathode is heated to high
temperature when it starts to emit electrons. A tungsten filament is commonly used
as the cathode in this case. Since we now have abundant of electrons emitted from
the cathode, the bombardment of the cathode by energetic ions is no more required
to keep the discharge self-sustained. In fact, this should be avoided since it will
shorten the life of the filament. The schematic setup is as shown in Fig. 2.9.
Since there is now abundant supply of electrons from the hot cathode, a flow of
current can be obtained even if the space between the electrodes is kept at absolute
vacuum.
In theory, the emission of electrons from the filament is governed by its tem-
perature T and the work function ϕ of the material used as the filament. This is
given by the Richardson-Dushman equation:

e/
J ¼ AT 2 exp ; ð2:17Þ
kT
where A is a constant.
Fig. 2.9 Schematic of a hot Hot cathode Anode

cathode discharge
RL V
However, this current density cannot be achieved since the electrons emitted are
not moved to the anode instantaneously and there will be a pile-up of electrons near
the cathode. At equilibrium, a layer of space charge will be formed in front of the
cathode. The distribution potential from the cathode to anode is such that almost all
the potential drop is across this layer of space charge and the potential at the edge of
the space charge is the same as that at the anode. The edge of the space charge layer
thus acts as the virtual anode. The current density is given by
rffiffiffiffiffiffi 3=2
4eo 2e V
J¼ ; ð2:18Þ
9 me d 2
where d is the inter-electrode spacing and V is the applied potential. This is,
however, subjected to a maximum value determined by the Richardson-Dushman
equation given by (2.17).
Note that the expression for J above applied to the vacuum case where the
electrons are assumed to encounter no collisions on their ways to the anode. In the
case when the inter-electrode space is filled with a gas at a fixed pressure, collision
cannot be neglected and electron mobility through the gas should be considered. The
discharge is said to be mobility limited and the discharge current density is given by
9eo le V 2
J¼ ; ð2:19Þ
8 d3
where μe is the electron mobility.
2.1.6 The Arc Discharge
We have seen that when the current of the glow discharge is increased to above 1 A,
the discharge will be transformed into the arc discharge. The potential drop across
the electrodes will drop to a value less than the glow voltage. At low pressure, the
discharge can be highly unstable and it may appear as a glow discharge with
intermittent arcing.
(For practical reason, the transition of glow to arc discharge is seldom observed
experimentally. The reason is that the DC power source used for the glow discharge
is normally rated at a few kV and below 500 mA. To power the arc discharge, the
required current rating is in the >A region. On the other hand, once a stable arc
discharge is formed, only a relatively low voltage is required to maintain it. Hence,
the power supply used for arc discharge is usually in the range of say 100 V, 100 A
rating. The main factor here is, of course, the cost of the equipment.)
The arc discharge is usually operated at higher pressure than that of a glow
discharge. The most common applications of the arc discharge in industry is to use
it as a high temperature heat source, for example in plasma furnace. For these
applications the arc discharge is operated at pressure above 10 torr and it is called
Fig. 2.10 The two regimes of arc discharge—thermal and non thermal
the thermal arc. At high pressure, the rate of collision is sufficiently high that the
electrons and ions/atoms are able to reach thermal equilibrium that they will achieve
a single temperature. For lower pressure operation, the electrons are at thermal
equilibrium corresponding to a temperature higher than that of the ions/atoms. The
two regimes of arc discharge are shown in Fig. 2.10.
Due to the high discharge current employed in the arc discharge, the plasma
column exhibits two distinct features as compared to the glow discharge. The first is
that the high discharge current causes a constriction effect on the plasma column.
Secondly, the electrodes are severely heated and they should be cooled by running
water. The heating effect will eject some of the electrode materials into the plasma
and thus cools the plasma close to the electrode surfaces. As a consequence of this
cooling effect this part of the plasma column is constricted to smaller radius than the
rest of the column.
An important property of the thermal arc discharge operating at high pressure is
its ability to achieve thermodynamic equilibrium. This qualify the plasma formed to
act as a blackbody radiator. The radiation spectrum is expressed by Planck Law as
dME 2phc2 1
¼ W=m2 =m ð2:20Þ
dk k expðhc=kTkÞ 1
5
This is plotted as shown in Fig. 2.11.

The peak emission is at a wavelength of
hc 2500
kMAX ðmÞ Å ð2:21Þ
5kT kTðin eVÞ
Thus for arc discharge at temperature of 1 eV, the peak of the blackbody
radiation is in the UV region. This may give rise to occupational hazard in using the
arc discharge heat source.
dM E
d
Area
MAX (m)
Fig. 2.11 The emission spectrum of a black body radiator
For application purposes, the arc discharge is normally operated in two con-
figurations: the plasma torch and the plasma spray. The plasma torch is used in
applications such as the plasma furnace for solid waste disposal. There are two
modes of plasma torch, one based on the transferred arc concept and another based
on the non-transferred arc concept. These are shown in Fig. 2.12. In the transferred
arc mode, the work-piece becomes one of the electrodes, normally connected to
earth potential. This will work if the work-piece is a conducting material. This
mode of operation is used for metal cutting or melting. In the non-transferred mode,
the arc discharge is formed between the cathode and the anode (ground) which is in
the form of nozzle. The working gas is flown through the nozzle so that the plasma
produced is in the form of a jet of high temperature. This can be used to treat
insulating materials as well.
Since the thermal arc discharge plasma is said to have achieved thermodynamic
equilibrium, the gas temperature (ions and atoms) is expected to be equal to that of
the electrons. A temperature of up to 20,000 K (equivalent to 2 eV) can be
expected.
The plasma spray is a variation of the non-transferred mode plasma torch. It can
be used to spray coat material (for example ceramics) onto a substrate, or to melt
Arc jet
Gas - + Gas -
+
Transferred mode Non-transferred mode
Fig. 2.12 Two possible modes of arc discharge configuration

Fig. 2.13 Plasma spraying using the arc discharge
and rapidly solidify heavy metals such as molybdenum or tungsten or their com-
posites to form power. An example of the setup for plasma spray coating of material
is shown in Fig. 2.13.
2.2 AC (Radiofrequency) Discharge
Gas discharge powered by an AC source can be considered not different from that
maintained by a DC source if the frequency is low. The criterion of “low” fre-
quency is that the characteristic time of the voltage variation (usually taken to be the
periodic time) should be much larger than the transit time of the ions from anode to
cathode. This is normally in the region of below 1 kHz. For high frequency, the
discharge behavior becomes different from the DC case. First, the breakdown
voltage will be lower. In a self-sustained glow discharge, new charged particles are
produced by the ionization of the gas by electron collisions and the secondary
emission at the cathode surface by ion bombardment and this is balanced by the loss
of electrons at the anode. We see from Eq. (2.5) that breakdown will occur when
this balance is achieved. For sufficiently high frequency (probably above 1 MHz)
source, the loss of electrons at the anode will be reduced (or even becomes zero
loss) since the alternating field may reverse some or all of the electrons before they
reach the electrodes. Although there will also be a reduction of electron production
by ion bombardment at the cathode surface, this will be compensated by the
reduction of loss of electrons at the anode and also the increase in ionization in the
gas since the electrons now remain in the plasma for longer time as they are moving
back and forth following the electric field. In this case the main electron loss
mechanism is radial diffusion.
As we can see from the above discussion, one favorable condition for AC
powered gas discharge is to employ sufficiently high frequency to ensure that the
oscillation time of the electrons caused by the alternating electric field is shorter than
their transit time between the electrodes, or ωτ < 1 where ω is the angular frequency
of the field and τ is the electron transit time between electrodes. This is also
dependent on the pressure of the gas as at lower pressure the electron transit time is
expected to decrease. For pressure of a few torr and discharge distance of several cm,
the use of frequency in the radiofrequency range is appropriate. A commonly used
frequency is 13.56 MHz as agreed by international communication authorities.
Under this condition we expect ω < υ where υ is the electron collision frequency.
This means that an electron collides many times within one oscillation of the electric
field so it may be able to transfer the energy it absorbs from the field to other
particles. For lower pressure, say in the region of mtorr, there may not be sufficient
collision for the electrons to achieve equilibrium with other particles so that it is
more effective to heat the electrons collectively. This requires frequency * electron
plasma frequency ωpe which is at the microwave region (>GHz).
In an RF discharge, the plasma will not be “off” between cycles of the electric
field variation which is normally sinusoidal, although the electrons are expected to
travel back and forth between the electrodes as the field changes direction. As a
result, a steady plasma is formed between the electrodes as seen in Fig. 2.14. The
condition of the plasma is expected to be similar to that of the DC discharge except
that the potential distribution between the electrodes may vary during each cycle.
An RF discharge can be produced by using 2 types of configuration: capacitively
coupled or inductively coupled. The capacitive coupling can be implemented with a
set of parallel plates such as those used for the conventional DC glow discharge
shown in Fig. 2.14. For an RF power source, an electrode less discharge can be
obtained by placing the electrodes outside the plasma chamber (Fig. 2.15). This can
eliminate the contamination of the plasma by the electrode materials.
Electrode less discharge can also be produced by using the inductively coupled
configuration. An example of this configuration is to wrap a solenoid of N turns,
Fig. 2.14 Schematic of an Chamber

AC glow discharge
Plasma
Fig. 2.15 The inductive Chamber

mode of AC glow discharge
Plasma
2.2 AC (Radiofrequency) Discharge 31
Plasma
RF Source
Quartz Tube
Coil
Fig. 2.16 Inductively coupled plasma using selonoid
length L and radius r around a glass tube (Fig. 2.16) and a discharge inside the tube
can be produced by the electric field induced by the high frequency current
I flowing through the solenoid.
In this case, there are two components of the induced electric field, one along the
axis of the solenoid (Ez) and one in the azimuthal direction (Eθ) given by:
lo xN 2 pr 2 I lo xNrI
Ez ¼ and Eh ¼
L2 2L
Under normal condition, Ez > Eθ until breakdown in the azimuthal direction

occurs and Ez will be reduced by the field induced by the toroidal plasma current.
Another possible form of inductively coupled plasma discharge system is by
using a planar coil instead of the cylindrical solenoid. The planar coil is placed
outside the discharge chamber, separated by a dielectric window usually made of
quartz. An example of such a system is as shown in Fig. 2.17.
In this configuration, we now have Eθ(r, z) instead of Ez together with Eθ. The
expressions for the capacitive field and the toroidal field are complicated and will
not be discussed here. However, the principle of operation in the two cases
(cylindrical solenoid and planar) are identical. As the input RF power is increased
gradually, the discharge is expected to breakdown due to the capacitive field pro-
duced by the coil. Hence at low RF power, the discharge is in the direction of the
capacitive field and it is said to be operated in E-mode. When the power is raised to
Fig. 2.17 Planar coil

inductively coupled plasma
E ( r, z )
Capacitive Fields from Coils
RF Bias Sheath Fields

sufficiently high level, breakdown in the azimuthal direction will occur. The azi-
muthal discharge is able to draw higher current than that in the axial discharge
because it is acting as the single loop secondary of the transformer with the solenoid
coil as the primary. This is the H-mode operation. The E to H mode transition is
usually abrupt and can be observed distinctly. The plasma produced has density
several order of magnitude higher than that produced in the E-mode operation. This
can be observed experimentally from the spectroscopic measurement of a desig-
nated line radiation of an ionic specie. An example is shown in Fig. 2.18 [2] which
shows the jump in the intensity of the 394.6 nm line of Ar+ in an argon discharge.
The abrupt increase in the intensity of this line is caused by an abrupt increase in the
number density of the Ar+ specie which may be due to a sharp increase in the
electron temperature of the plasma during the E to H mode transition.
A crucial problem in the design of an RF discharge circuit is the efficient
coupling of the RF power source to the plasma. This can be achieved by connecting
the RF power source to the load (parallel plate electrodes as capacitive load;
cylindrical solenoid or planar coil as inductive load) through a matching network.
In general, the plasma load can be considered to be a combination of inductance L,
capacitance C and resistance R. In an RF circuit, we can define the admittance of the
load to be Y = G + jB, where
R
G¼
R2 þ ðXC þ XL Þ2
is the conductance, with XC ¼ xC

1
and XL ¼ xC,
XC þ XL
B¼
R2 þ ðXC þ XL Þ2
is the susceptance.
Fig. 2.18 The sharp increase

spectral line intensity during
E-H mode transition
2.2 AC (Radiofrequency) Discharge 33
RT L
Cp
C
Rp
Fig. 2.19 Matching network for a capacitively coupled discharge
This load can be matched by an RF source with RT ¼ G1 and CT ¼ xB . For a

capacitively coupled discharge, the load can be represented by a capacitor and a
resistor in series as shown in Fig. 2.19. Matching can be achieved with
1
1 1 1 2 1 1 1
C¼ 2 and L ¼ ðRp RT R2p Þ2 þ
x RT Rp RT x xCp
For an inductively coupled plasma, it can be represented by an inductance and a

resistance in series as shown in Fig. 2.20.
For this case, matching is achieved when
1 " #
1 1 1 2 1 1
C2 ¼ and C1 ¼
x RT Rp R2T x xLp ðRp RT R2 Þ12
p
Another type of matching network used for inductively coupled plasma is to use
only C1 instead of both C1 and C2 as discussed above. In this case C and Lp forms an
LC resonant circuit which will achieve maximum current flow when xLp ¼ xC 1
. For
this case, B = 0 and Y = G = 1/Rp. We require RT = Rp for the circuit to match.
However, since Rp is usually varying with plasma condition this is difficult to
achieve. For discharge in E mode, Rp is normally negligibly low so matching has to
be done by balancing XL and XC (varying C) so that the net impedance is equal to RT.
Fig. 2.20 Matching network

for an inductively coupled
discharge
Fig. 2.21 Matching network of an inductively coupled discharge using resonant circuit
This requires RT to be as low as possible if we want to produce discharge current of

as near to the peak of the resonance as possible. The commonlly available RF power
source with 50 Ω output impedance is not suitable in this case. The source impe-
dance suitable for this purpose should be as low as possible, may be 1 Ω or lower.
One possible solution to this problem is to add a step down transformer between the
RF generator and the LC circuit to provide an effective low output impedance for
the generator. This arrangement works well for E mode discharge. However, for the
H mode discharge, Rp may be higher so the matching has now to be achieved by
keeping xLp ¼ xC 1
while the impedance of the source is adjusted to be equal to Rp.
This is difficult to achieve and the circuit in Fig. 2.21 may be more suitable.
Plasma torch powered by RF power source can also be built using the induc-
tively coupled configuration. The most common application of this plasma source is
for elemental analysis by spectroscopic method. The equipment used for this pur-
pose is normally called the “ICP” which actually means inductively coupled
plasma. The plasma is produced by heating the stream of gas (at atmospheric
pressure) flowing through the inside of a solenoid coil. The sample to be analyzed
may be in the gas form or for liquid sample, it will be heated to form vapor to pass
through the torch.
2.3 Microwave Heating of Plasma
The microwave is the region of electromagnetic spectrum with frequency above

1 GHz up to 300 GHz corresponding to wavelength of 30 cm to 1 mm respectively.
This is the range of electron plasma frequency for plasma with electron density in
the range of 1 × 1010 cm−3 to 1 × 1015 cm−3. Hence when the microwave is used for
heating a plasma at such a range of density, the microwave energy will be absorbed
by the electrons collectively and not individually. In contrast to individual particle
2.3 Microwave Heating of Plasma 35
heating (such as DC or RF discharge), the electrons gain energy by absorbing the

microwave energy directly and convert it into kinetic energy.
Assuming that the microwave can be described by the wave equation:
E ¼ Eo exp½iðj:r x tÞ ð2:22Þ
where E is the electric field with magnitude Eo, κ is the wave vector, r is the
position vector, and ω is the angular frequency of the wave. The rate of energy
absorption by electrons can be shown to be given by:
!
ne e2 Eo2 1
P¼ ð2:23Þ
2me m 1 þ ðxmÞ2
where ν is the electron-atom/ion collision frequency, ne is the electron number

density and me is the electron mass. It can be seen that P is maximum when ω ≪ ν.
This requires that the plasma density should not be too low. A schematic of the
typical setup used for this method of plasma heating is shown in Fig. 2.22.
The microwave power can be coupled to the gas inside the chamber by either the
resonant cavity configuration or the multimode cavity configuration. In the resonant
cavity configuration, the dimensions of the cavity, R and d are chosen to be equal to
the wavelength of the microwave, that is R = λ, and d = λ, where λ = 2πc/ω. In the
multimode configuration, the conditions are R > λ, and d > λ. The plasma produced
by the multimode configuration is expected to be more uniform.
For low plasma density where ω > ν, the electron heating by microwave can be
enhanced by applying an external magnetic field. For the case where this external
magnetic field is in the direction parallel to the direction of propagation of the wave,
the wave will become circularly polarized, in both Right and Left directions. The
rate of wave energy absorption by the electron can now be expressed in 2 com-
ponents as: P = PR + PL where
!
ne e2 Eo2 1 1
PR ¼ ðR waveÞ ð2:24Þ
2me m 2 1 þ ðxxce 2
m Þ
Fig. 2.22 Schematic of a

simple system of microwave R
heating of plasma
Waveguide
and Tuner d
Plasma
Window
!
ne e2 Eo2 1 1
PL ¼ ðL waveÞ ð2:25Þ
2me m 2 1 þ ðx þmxce Þ2
Here xce ¼ eB

me is the electron cyclotron frequency of the magnetized plasma.
Hence it can be seen that the absorption of microwave energy by electron will be a
maximum when the frequency of the wave is the same as the electron cyclotron
frequency, ω = ωce. This is known as Electron Cyclotron Resonant (ECR) heating.
Since the magnetic field distribution is not uniform throughout the volume inside
the chamber, the condition of ω = ωce may only be met in certain area and this is
called the resonant surface. It is within this resonant surface that intense plasma
heating will occur. The electron density produced by the ECR heating can be easily
more than an order of magnitude higher than that achieved in the normal microwave
plasma heating. Electron temperature of 5–10 eV can be achieved, as compared to
1–2 eV in most other DC or RF discharges.
2.4 Pulsed Plasma Discharges
The plasmas produced by DC or RF discharge or by microwave heating are at

steady state. The plasma electron temperature of these plasmas is below 10 eV and
the electron density is below 1015 cm−3. In order to produce hotter and denser
plasma, a power source capable of generating higher power density will be
required. This can be achieved by using a capacitor discharge where the energy is
first stored in the capacitor and then discharge through the gas to produce the
plasma. Power density of up to 1018 W/m3 or higher can be generated easily. The
plasma produced is able to achieve condition near to the fusion plasma condition.
The simplest technology used to power a pulsed plasma discharge is the
capacitor discharge system. In this case a capacitor C is first charged up to a high
voltage V so that an amount of energy of 12 CV 2 is stored in the capacitor. This
energy is transferred to the plasma load via a switch. The schematic circuit of this
capacitor discharge system is shown in Fig. 2.23.
In this circuit the plasma is represented by a time varying inductance Lp(t) and a
resistance Rp(t) in series. Lo and Ro are the stray inductance and resistance of the
circuit. Stray inductance and resistance may be contributed by the cables used to
connect the various components together as well as contributed by the switch. The
capacitor or a bank of capacitors used also has an internal inductance. All these
have been lumped together and represented by Lo and Ro. The circuit represents a
typical LCR circuit of which the circuit equation can be written as:
R R
d I dt dI I dt
Vo ¼ ðLIÞ þ IR þ ¼ Lo þ IRo þ þ Vp ð2:26Þ
dt C dt C
2.4 Pulsed Plasma Discharges 37
Fig. 2.23 Schematic of a Lo Ro

capacitor discharge circuit
L p(t)
C
R p(t)
where
L ¼ L o þ Lp ; R ¼ Ro þ Rp
and
d
Vp ¼ ðLp IÞ þ I Rp
dt
is the voltage across the plasma.

Equation (2.26) can be rewritten in the form
d 2 I R dI 1
þ þ ¼0
dt 2 L dt LC
which has a general solution of the form:

R
IðtÞ ¼ ½A expðntÞ þ B expðntÞ exp t ; ð2:27Þ
2L
where
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2
R 1
n¼
2L LC
Define the scaling parameter

rffiffiffiffi rffiffiffiffi
L C
a ¼ R= ¼R ;
C L
then
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

a 2
n ¼ x2 1
2
We can see that n may be imaginary, zero or real corresponding to α < 2, α = 0 or

α > 2.
Depending on n, I(t) may take three distinct forms:
(a) n = imaginary or α < 2. In this case the solution becomes

Vo R
IðtÞ ¼ sin xt exp t :
xL 2L
Here ω ≈ √(−1) × n for α < 2. The current waveform is a damped sinusoid with
angular frequency ω and damping time constant 2L/R. The oscillation is said to
be under damped.
(b) n = 0 or α = 2. In this case the solution becomes

Vo R
IðtÞ ¼ t exp t :
L 2L
This is a waveform which decays with a time constant 2L/R, the shortest time
possible for the current to decay to zero. The waveform is said to be critically
damped.
(c) n = real or α > 2. In this case the solution becomes

Vo R
IðtÞ ¼ sinhðntÞ exp t :
nL 2L
The waveform now decays with a time constant >2L/R. It is said to be over
damped.
In principle, it is ideal to make use of the critically damped discharge for
heating plasma. However, the plasma impedance is time varying so it is dif-
ficult to maintain the matching condition for critically damped discharge. For a
high temperature plasma, the plasma resistance is normally low so the plasma
can be represented by an inductance only. In this case the discharge is expected
to be under-damped. On the other hand, in a high pressure pulsed discharge,
the plasma resistance is sufficiently high that it may be possible to achieve the
condition for critically damped discharge. One such example is the high
pressure flash lamp.
2.4.1 Pulsed Arc Discharge in High Pressure Flash Lamp
The high pressure flash lamp is a high power arc discharge that is commonly used
as a light source. One such application is the use of linear flash lamp for the
pumping of solid state laser (for example ruby laser) or dye laser.
Glass or Quartz
Arc
Electrodes
Fig. 2.24 Schematic of a linear flash lamp
An example of the linear flash lamp is as shown in Fig. 2.24.

The schematic circuit of the flash lamp discharge is similar to that shown in
Fig. 2.23 except that the plasma is now represented by a fixed inductance (un-
known) and a resistance (fixed or time varying).
Here we may combine the fixed inductance of the flash lamp and the stray
inductance of the circuit into L and the stray resistance of the circuit is assumed to
be negligible. Hence the circuit equation can be written as:
R
Idt dI
Vo ¼ þ L þ Vp ð2:28Þ
C dt
where
L ¼ Lo þ Lp
which is constant.
And
dI
Vp ðtÞ ¼ IRp þ Lp :
dt
Equation (2.28) can be normalized and solved for I(t). This can be done by
defining
I t
i¼ ; s¼
Io to
where
rffiffiffiffi
C pffiffiffiffiffiffi
Io ¼ V o and to ¼ LC
L
The normalized form of (2.28) is

Z
di
þ ai þ ids ¼ 1 ð2:29Þ
ds
where
rffiffiffiffi
Rp L
a¼ and Zo ¼
Zo C
This is actually the same as solving the ideal LCR discharge discussed earlier
numerically and by choosing the value of α to be <2, =2 and >2 will give rise to
solution for light damping, critical damping and over damping. The boundary
conditions are: τ = 0, ι = 0, ∫ιdτ = 0, dι/dτ = 1. The solutions for the three cases are
shown in Fig. 2.25.
It is clear that to obtain a single flash with optimum intensity and shortest pulse
width, it is best to operate with critical damping condition. For an under-damp flash
lamp discharge, the output light pulse may have a long tail and several humps in the
light intensity may be observed as shown in Fig. 2.26 [3].
Note that in the above consideration, we have assumed that the resistance of the
flash lamp is constant. In actual situation the resistance has been found to be
dependent on the discharge current. It has been determined empirically that the
pffiffiffiffiffi
voltage across the flash lamp can be expressed as Vp ¼ ko jI j, where ko is often
referred to as the flash lamp constant given by:
15
‘ P
ko ¼ 1:28 ð2:30Þ
d X
where
X a constant of the gas (e.g. 450 for xenon, 805 for krypton as determined
empirically),
‘ the length of the flash lamp,
d the diameter of the flash lamp, and
P the pressure of the gas inside the flash lamp
Fig. 2.25 Current waveforms 0.8 = 0.5

for under-damped (α < 2),
critically damped (α = 2) and 0.6
over-damped (α > 2) LCR
discharge =2
0.4
0.2
=5
0.0
-0.2
-0.4
0 2 4 6 8 10
Fig. 2.26 Example of light

output from a flash lamp
powered by under-damped
LCR discharge (Reprinted
from Ref. [3], Copyright
(1989), with permission from
AIP)
The circuit equation is then written as

R
Idt dI pffiffiffiffiffi
Vo ¼ þ Lo k o j I j ð2:31Þ
C dt
and the normalized form becomes

Z
di
ajij2 þ
1
ids ¼ 1 ½take þ when i ¼ þ ; take when i ¼ ð2:32Þ
ds
ko
where a ¼ pffiffiffiffiffiffiffi
Vo Zo
is now the damping factor. An example of the solutions obtained
for α = 0.2, α = 0.8 and α = 2 are shown in Fig. 2.27. It can be seen that α = 0.8
corresponds to the case of critical damping.
2.4.2 Inductive Model of Pulsed Discharge—Shock Heating
In the flash lamp discharge discussed above, the plasma is heated by the joule
heating effect. In this case the discharge is described by the resistive model where
the plasma is represented by a resistance and its inductance is either assumed to be
constant or negligible. This is due to the fact that the plasma column produced by
the discharge has a fixed geometry. In some cases such as the electromagnetic
shock tube or the pinches, the current is arranged to flow in a thin sheet and it is
driven by the self electromagnetic force (the J ∧ B force) to supersonic speed so
Fig. 2.27 The solutions for 1

under-damped (α < 0.8), = 0.2
critically damped (α = 0.8)
and over-damped (α > 0.8)
= 0.8
flash lamp discharge
=2
-1
0 2 4 6 8 10
that shock heating of the plasma column can occur. The discharge is modeled by
the inductive model where the plasma is represented by a time varying inductance.
The inductance is time varying because its geometry is changing dynamically.
We consider two types of dynamics that may give rise to shock heating: the
electromagnetic shock tube in which the current sheet is moving axially and the
pinch (Z- and θ-pinches) where the current is moving radially towards the axis of
the plasma column. Before we begin with these plasma devices we shall first
discuss the shock heating effect.
2.4.2.1 Shock Heating of Plasma
Shock heating is a phenomenon where the particles in the plasma are driven to
supersonic speed. In general, every medium has a characteristic speed by which
information will propagate. An example is the speed of sound in air. If a piston is
accelerated to a speed faster than the speed of sound, it is said to be supersonic and
the perturbation created by the piston will pile up in front of the piston to form a
shock front as illustrated in Fig. 2.28.
The shock front is the imaginary boundary formed in front of the supersonic
piston which separates the ambient and the shock heated gas or plasma, Particles in
front of the shock front are unperturbed while the particles after the shock front are
being pushed by the piston to the speed of the piston and is said to have been
shock-heated. The layer of shock heated gas in between the piston and the shock
front will grow to be thicker and forms the plasma.
Consider the frame of reference of the shock front (that means you move with
the shock front), the gas on the right hand side is at ambient while the gas on the left
hand side has been shock heated.
vp > vs (characteristic speed)
t=0
piston
t = t1 shock front
piston
t = t2 vp vsf
piston shock front
Fig. 2.28 Schematic illustration of the shock wave formation

q2 q1

P2 ; q2 ; T2 ; h2 P1 ; q1 ; T1 ; h1
where
q1 ¼ vsf
q2 ¼ vsf vp
For this system we can write:
q2 q2 ¼ q1 q1 ðconservation of mass across the shock frontÞ ð2:33Þ
q2 q22 þ P2 ¼ q1 q21 þ P1 ðconservation of momentum across the shock frontÞ

ð2:34Þ

1 2 1
q2 q 2 q2 þ h2 ¼ q1 q2 q21 þ h1 ðconservation of energyÞ ð2:35Þ
2 2
These three equations are called the shock jump equations. Another two relevant
equations are:
R0
P2 ¼ q2 T2 ð1 þ a1 þ 2a2 þ Þ ðequation of stateÞ ð2:36Þ
M
c P2
h2 ¼ ðenthalpyÞ ð2:37Þ
c 1 q2
It is useful to consider strong shock with the conditions that
P2 P1
h2 h1
Then Eqs. (2.34) and (2.35) will be reduced to
q2 q22 þ P2 ¼ q1 q21 ð2:38Þ

1 2 1 2 1 2 1 2
q2 q2 q2 þ h2 ¼ q1 q2 q1 ! q þ h2 ¼ q ð2:39Þ
2 2 2 2 2 1
We now define the density ratio C ¼ qq2 ¼ qq12 and substitute h2 from (2.37)
1
into (2.39),
1 2 c P2 1 2
q þ ¼ q
2 2 c 1 q2 2 1
P2 c1 1 2
) ¼ð Þ ðq q22 Þ
q2 c 2 1
1 c1 2 q2
¼ ð Þq1 ½1 ð Þ2
2 c q1
1 c1 2 1
¼ ð Þq1 ½1 ð Þ2
2 c C
P2 1 c 1 2 C2 1
¼ ð Þq1 ð Þ ð2:40Þ
q2 2 c C2
Also, from (2.38),
q1 q21
P2 ¼ q1 q21 q2 q22 ¼ q2 q22 ½ 1
q2 q22
P2 q2 C1
) ¼ q22 ðC 1Þ ¼ q21 ð 22 ÞðC 1Þ ¼ q21 ð 2 Þ ð2:41Þ
q2 q1 C
Combining (2.40) and (2.41),
1 c 1 C2 1 C1
ð Þð 2
Þ¼
2 c C C2
1 c1
ð ÞðC þ 1ÞðC 1Þ ¼ ðC 1Þ
2 c
2c
C¼ 1
c1
cþ1
Hence, C¼ : ð2:42Þ
c1
This is the first important consequence of strong shock which relates the density
ratio to the specific heat ratio. Also, from the equation of state, we have
P2 Ro T2 z
¼
q2 M
where z ¼ 1 þ a1 þ 2a2 þ is the departure coefficient.

Combining with Eqs. (2.41) and (2.42), we can obtain

M 2ðc 1Þ 2
T2 ¼ q ð2:43Þ
Ro ðc þ 1Þ2 z 1
This relates the temperature of the shock heated plasma to the shock speed.
Write the kinetic energy per unit mass of the shock heated plasma as

1 1 1 1 2
k2 ¼ v2p ¼ ðq1 q2 Þ2 ¼ q21 1
2 2 2 C
and the enthalpy per unit mass from Eq. (2.39),

1 2 1 2 1
h 2 ¼ q1 q2 ¼ q1 1 2
2
2 2 C
Hence the ratio of enthalpy over kinetic energy
h2 1 C12 Cþ1
¼ ¼
k2 1 C1
2 C1
Substitute Γ from (2.42) gives
h2
¼c ð2:44Þ
k2
This means that in a strong shock heated plasma, the enthalpy is given by
specific heat ratio times the kinetic energy.
We consider the example of a strong shock heated hydrogen plasma such that
fully ionized state has been achieved. For this plasma, we have c ¼ 53 ; χ = 0, α = 1,
hence z = 1 + χ + 3α = 4.
For this fully ionized hydrogen plasma, we can assume that the enthalpy is
approximately equal to the ionization potential, which is equal to 15.9 eV (only 1
electron to be ionized per H-atom). While for each H-atom, its kinetic energy when
being pushed by the piston to a speed of vp is
1
k ¼ 1:67 1027 v2p 5:3 109 v2p ðin unit of eVÞ
2

From (2.44), 15:9 ¼ 53 5:3 109 v2p , hence vp = 4.3 × 104 m/s
This means that the hydrogen gas must be driven to a speed of more than
4.3 × 104 m/s for it to be fully ionized.
The shock front speed corresponding to this piston can be obtained from
q1
vsf ¼ q1 ¼ q2 þ vp ¼ þ vp
C
Hence

vp C cþ1
q1 ¼ ¼ vp ¼ vp
1C 1 C1 2
For c ¼ 53 ; q1 ¼ vsf 43 vp ¼ 5:7 104 m=s:

The temperature of the plasma can also be calculated from (2.43) as

M 2ðc 1Þq21 3 M 2
T2 ¼ ¼ q
Ro ðc þ 1Þ2 z 64 Ro 1
2.4.2.2 Pulsed Plasma Systems with Shock Heating as Plasma Heating

Mechanism
There are numerous types of pulsed plasma systems in which shock heating is
operational. These include the electromagnetic shock tube, the Z-pinch and the
θ-pinch.
(a) The electromagnetic shock tube.
In the electromagnetic shock tube, the discharge is initiated at the back-wall
uniformly across a cylindrical insulator between the coaxial electrodes to form
a current sheet as shown in Fig. 2.29.
Fig. 2.29 Schematic of an electromagnetic shock tube
The current sheet acts as the electromagnetic piston which is formed first along
the surface of the insulator. It will be pushed outwards from the surface of the
insulator to the inner surface of the outer electrode by its own electromagnetic
force, or the J ∧ B force. The azimuthal magnetic field Bθ is generated by the
coaxial discharge current itself and is given by
lI
Bh ¼ ;
2pr
which is a function of time and radial position r. It can be seen that its magnitude
will be stronger at the surface of the inner electrode as compared to the inner surface
of the outer electrode. This gives rise to the slanting structure of the electromagnetic
piston as shown in Fig. 2.29. The J ∧ B force is in the direction downstream of the
tube (in the z-direction) and its magnitude can be expressed as
Zb
B2h
2prdr;
2l
a
where a is the radius of the inner electrode and b is the radius of the outer electrode.
This force will drive the electromagnetic piston to supersonic speed so that a shock
heated layer of plasma will be formed. In this way, with discharge current in the
region of 100 kA, piston speed of up to more than 10 × 104 m/s may be achieved,
which is sufficient to produce fully ionized hydrogen plasma.
(b) The Z-pinch and θ-pinch
In the linear Z-pinch, the electromagnetic piston is cylindrical while the J ∧
B force is in the inward radial direction as illustrated in Fig. 2.30.
Here again Bh ¼ 2pr
lI
with r is now time varying and the J ∧ B force is
B2h
Fm ¼ ð2pr‘Þ:
2l
A variation of the pinch configuration is the θ-pinch which in contrast to the

Z-pinch, the discharge current is induced in the θ-direction while its self-magnetic
j
j
B B
j j
j
Primary
current
Z-Pinch -Pinch
Fig. 2.30 Formation of the Z-pinch and the θ-Pinch
field is in the z-direction instead. The combined effect is also a J ∧ B force in the
radially inward direction. This is illustrated in Fig. 2.30, with the Z-pinch config-
uration shown together for comparison.
References
1. Weston GF (1968) Cold cathode glow discharge tubes (Chap. 1). ILIFFE Books Ltd, London
2. Yip Cheong K (1997) Studies on a planar coil inductively coupled plasma system and its
applications. MSc Thesis, University of Malaya
3. Chin OH, Wong CS (1989) A simple monochromatic spark discharge light source. Rev Sci
Instrum 60:3818–3819
General References
4. Grey MC (1965) Fundamentals of electrical discharges in gases, part 1, vol 2. In: Beck AH
(ed) Handbook of vacuum physics. Bergamon Press, Oxford
5. Reece RJ (1995) Industrial plasma engineering. Principles, vol 1. IOP Publishing Ltd, London
Chapter 3
Plasma Diagnostic Techniques
Abstract In this chapter, some of the basic diagnostic techniques that are useful for
studying various types of plasmas are described. These include electrical mea-
surements (discharge current and voltage across the plasma), spectroscopic mea-
surements, the Langmuir (electric) probe, X-ray and neutron measurements.
Keywords Plasma diagnostics
3.1 Electrical Measurements
In many plasma devices such as the plasma focus and the vacuum spark, plasma
heating is achieved by the passage of intense current pulse through the plasma. The
heating mechanism involved may be the magnetic compression (which is inductive
in nature) and/or the joule-heating (which is resistive). In any case, we may con-
sider the plasma as an active element in the discharge circuit with its electrical
properties represented by the combination of a variable resistor and a variable
inductor. This concept is illustrated in Fig. 3.1.
In this circuit, the energy is initially stored in the capacitor C. Upon closing the
switch, the capacitor discharges through the circuit. The discharge current is sig-
nificantly affected by the condition of the plasma that a measurement of the dis-
charge current will be able to reveal information concerning the dynamic changes
of the plasma condition during the discharge. Similarly, the transient voltage across
the plasma is also directly related to the plasma condition. Combined interpretation
of the measured current and voltage waveforms is usually sufficient for basic
dynamics study of a pulsed plasma device.

DOI 10.1007/978-981-10-0117-8_3
50 3 Plasma Diagnostic Techniques
Lo Ro
Lp(t)
C
Rp(t)
Fig. 3.1 Schematics of a pulsed plasma discharge circuit powered by capacitor discharge
3.1.1 Pulsed Current Measurement by Using

the Rogowski Coil
A Rogowski coil is essentially a multi-turn solenoid which is bent into the shape of
a torus encircling the current as shown in Fig. 3.2.
In order to understand the principle of operation of the Rogowski coil, we
assume the ideal situation where the current to be measured I(t) is passing through
the centre of the major cross-section of the torus. Then by Ampere’s Law, the
magnetic field induced by the current at the axis of the minor cross section of the
torus is given by
lo I
BðtÞ ¼ ð3:1Þ
2pa
Then the magnetic flux that threads the minor cross-section of the torus is

lo A
/ ðt Þ ¼ I ðtÞ ð3:2Þ
2pa
I(t)
V(t)
Fig. 3.2 Principle of Rogowski coil for pulsed discharge current measurement. N Number of
turns, A Minor cross-sectional area, a Major radius of coil
3.1 Electrical Measurements 51
The voltage induced across the terminal of the coil is

lo AN dI
V ðt Þ ¼ ð3:3Þ
2pa dt
We can see that the induced voltage is proportional to the rate of change of the
current, not the current itself. In order to obtain the I(t), the coil output voltage must
be integrated. This can be done in two ways.
3.1.1.1 Integration Using Passive Integrator
The coil output voltage can be integrated by a simple passive RC integrator as

illustrated in the equivalent circuit shown in Fig. 3.3.
The Rogowski coil is represented by an inductance L. The circuit equation can
be written as
Zt
d/ di 1
VðtÞ ¼ ¼ L þ iR þ idt ð3:4Þ
dt dt C
0
where i is the induced current flowing in the coil circuit. We assume that R Lx,
then
Zt
1
VðtÞ ¼ iR þ idt: ð3:5Þ
C
0
Further assume that RC is chosen such that it is much larger than the characteristic
time of the expected plasma event, then we obtain

VðtÞ lo AN dI 1
VðtÞ iR ) i ¼ :
R 2pa dt R
Fig. 3.3 Schematic circuit diagram of Rogowsdi coil with RC integrator

The integrated output of the coil is measured across C and hence given by
Zt
1 lo NA
Vo ðtÞ ¼ idt ¼ IðtÞ / IðtÞ ð3:6Þ
C 2paR
0
which is directly proportional to the discharge current itself.

For this mode of operation, the Rogowski coil has a lower frequency limit of
1
f[ ð3:7Þ
RC
while its upper limit is set by the fact that a 50 Ω coaxial cable is normally required
to connect the coil to the RC integrator which is in turn connected directly to the
oscilloscope input. In this case the coaxial cable is shorted to ground via a 50 Ω
resistor before connected to the RC integrator. It is required that ωL < 50. This gives
rise to the upper frequency limit for the Rogowski coil as
50 50
f\ or simply f\ : ð3:8Þ
2pL L
This limit on the frequency response of the Rogowski coil can be quite serious since
if L lH; ) f \50 MHz:
3.1.1.2 The Rogowski Coil as a Current Transformer
The Rogowski coil can be operated as a current transformer by shorting its ter-
minals by a small resistance R as shown in Fig. 3.4.
The circuit equation in this case is
di
V ðt Þ ¼ L þ iðr þ RÞ ð3:9Þ
dt
where r is the resistance of the coil which is normally negligible and L is the coil
inductance. If R is chosen to be small such that ðr þ RÞ xL, then
Coil
Fig. 3.4 Schematic circuit diagram of Rogowski coil operated as current transformer
di d/
V ðt Þ ¼ L ¼ ð3:10Þ
dt dt
and hence
Zt
1 /
iðtÞ ¼ V ðtÞdt ¼ ð3:11Þ
L L
0
with
lo NAIðtÞ
/¼
2paR
and
lo N 2 A
L¼
2paR
R
Vo ðtÞ ¼ iR ¼ IðtÞ ð3:12Þ
N
which is the output voltage measured across R.

In principle, the frequency response of the Rogowski coil operated as current
transformer is only limited by the transit time of electrical signal along the length of
the coil which is given by 2πa. The transit time along the coil is given by
pffiffiffiffi
s ¼ lc 2pa ð3:13Þ
where l is the inductance per unit length while c is the capacitance per unit length of
the coil. This gives the upper frequency limit of the Rogowski coil operated as
current transformer as
1
f\ : ð3:14Þ
s
The lower frequency limit is set by the condition that R < ωL which gives
R
f[ : ð3:15Þ
L
3.1.1.3 Calibration of the Rogowski Coil
For absolute measurement of the discharge current, the Rogowski coil must be cali-
brated. The most straight forward method of calibrating the coil is to use Eq. (3.6) or
(3.12). However, it is difficult to estimate the geometry of the coil (N, A, and a) or the
value of the small resistance used to short the coil (R) accurately. Hence this method of
calculating the calibration factor is seldom practiced. The Rogowski coil calibration is
usually done by the so call in situ method. In this case, an ideal LCR discharge circuit
with light damping is used. The method is said to be “in situ” since the coil can be
calibrated at the position it is being used for the actual experiment. An experimental
condition will be chosen such that an ideal LCR discharge is obtained. An example of
the discharge current waveform for such a discharge is as shown in Fig. 3.5.
This current waveform can be represented mathematically by:
IðtÞ ¼ Io eat sin xt ð3:16Þ
where
rffiffiffiffi
C R 1
Io ¼ Vo ; a¼ ; x ¼ pffiffiffiffiffiffi :
L 2L LC
From the current waveform, we can measure T, V1, and V2. Then
‘nðV2 =V1 Þ
a¼ ð3:17Þ
T
Consider the first peak current at t ¼ T4
I1 ¼ Io eaT=4 ð3:18Þ
which can be calculated since Io, α and T can be obtained. Hence the calibration
factor of the coil can be obtained as
1
I1 2p CVo V2 4
K¼ ¼ : ð3:19Þ
V1 TV1 V1
In most pulsed plasma experiments, the discharge current generated is an under

damped sinusoidal waveform. The main plasma heating mechanism normally
V1
V2
Fig. 3.5 Under-damped sinusoidal discharge current waveform

occurs during the first half cycle of the discharge where the discharge current
waveform is severely deformed from that of an ideal sinusoidal waveform. For
example, in a plasma focus discharge, the plasma heating which occurs first through
the axial acceleration phase and then the radial compression phase during the first
half cycle of the discharge. This implies that the plasma impedance may be
changing and hence the resulting discharge current waveform may deviate from that
of an ideal LCR discharge. For the purpose of Rogowski calibration, the plasma
focus can be shorted at its back wall by a high power resistor acting as the load.
This high power resistor can be constructed by using, for instant, copper sulphate
solution. Alternatively, the plasma focus can be discharged under high ambient
pressure condition so that the current sheath will hardly move during the first few
cycles of the discharge current. The discharge current waveform can be acquired at
a compressed time scale so that several cycles of the waveform can be recorded as
shown in Fig. 3.5.
3.1.2 Pulsed Voltage Measurements
It is informative if the transient voltage across the plasma can be measured. This can
be done by using either the resistive voltage divider or the capacitive voltage
divider.
3.1.2.1 The Resistive Voltage Divider
The principle of resistive voltage divider is very simple, as illustrated in Fig. 3.6.
The high input voltage Vi is divided by R1 and R2 and the output voltage is
measured across the lower resistance R2 which is given by
R2
Vo ¼ Vi :
R1 þ R2
A practical design of the resistive voltage divider is shown in Fig. 3.7.
Fig. 3.6 Schematics of a resistive voltage divider

Fig. 3.7 An example of a resistive voltage probe
While the principle of the resistive divider is rather simple, there are several
technical difficulties in the implementation of the divider to measure the voltage
across the plasma: (i) It may be difficult to connect the divider directly across the
plasma so the voltage measured may not be the true voltage across the plasma but
consists of additional voltage contributed by other stray components of the dis-
charge circuit. (ii) In most cases the voltage divider is connected directly onto the
high voltage point to be measured so it is necessary to use long coaxial cable to
transmit the signal to the oscilloscope which may be placed inside a screened room
or box. This restricts the value of R2 to be 50 Ω which is the impedance of the
coaxial cable and the coaxial cable coupled to the 50 Ω terminator which will act as
R2 as a whole. One possible configuration is to choose R2 to be much lower than
50 Ω, say 1 Ω, but then the attenuation factor will be high since it is required that
R1 + R2 must be sufficiently large so that they will not act to short the plasma load.
For low R2, the upper frequency limit of the divider may also be affected since the
upper frequency limit is determined by
R2
f\
Ls
where Ls is the stray inductance of the probe circuit which is difficult to avoid.
3.1.2.2 Capacitive Voltage Divider
In the capacitive voltage divider, the voltage division is achieved by using capac-
itors instead of resistors as shown in Fig. 3.8.
Here the output voltage is measured across C2 which is of larger value than C1.
This is given by
C1
Vo ¼ Vi :
C1 þ C2
It is however technically difficult to implement the capacitive divider since in order

to measure the voltage across C2 accurately the measuring device must have infinite
impedance. This is not possible since it is normally necessary to use coaxial cable to
connect between the divider and the oscilloscope. This means shorting C2 with
50 Ω and this will defeat the functioning of the capacitive voltage divider.
A possible configuration to implement the capacitive divider is to connect C2
directly to the input of the oscilloscope which has an input impedance of 1 MΩ. In
this case C2 is formed by using a coaxial cable which has a capacitance of 100
pF/m, while C1 is formed between the central conductor of the coaxial cable and the
high voltage point to be measured, separated by insulator. This is shown
schematically in Fig. 3.9 [1]. For this configuration, one important trick is the
oscilloscope input must be set to DC mode to bypass the input capacitor.
The practical usage of this capacitive divider configuration is, however, quite
limited. It is probably only suitable for low power (current of a few kA and voltage
of a few kV) application. For high pulsed power application, a hybrid
Fig. 3.8 Schematics showing voltage division by capacitors

Fig. 3.9 Illustration of the

mounting of a simple
capacitive voltage divider
Fig. 3.10 Hybrid

capacitive-resistive voltage
probe
capacitive-resistive configuration is usually employed. An example of such a

configuration is shown schematically in Fig. 3.10.
The overall voltage division of this high voltage probe is given by:
Vo R2 C1
¼ :
Vi R1 þ R2 C1 þ C2
It is important to choose the values of R1 and R2 so that
1
ðR1 þ R2 Þ :
xC2
This set the frequency limit for the probe to operate in the voltage divider mode
(where Vo / Vi ) at
Fig. 3.11 An example of the voltage signal recorded for the discharge of electromagnetic shock
tube
1
f[ :
ðR1 þ R2 ÞC2
For frequency below this limit, the probe is expected to operate in the differential
mode where
dVi
Vo / :
dt
For frequency f ðR1 þ1R2 ÞC2 , we will get a mixture of the divider and differential
modes which will be difficult to interpret.
3.1.2.3 Calibration of the Voltage Divider
Although it is normally possible to estimate the voltage division of the resistive or

capacitive divider from the values of the components used, it is useful to perform a
calibration of the probe for more accurate measurement. This can be done by using
a standard pulse generator with know amplitude and rise and fall times as the
voltage source. In situ calibration can also be performed depending on the actual
experimental setup. For example, in the case of the electromagnetic shock tube, a
suitably delayed breakdown by lowering the filling pressure will give a signal as
shown in Fig. 3.11.
Calibration of the voltage divider can then be obtained by dividing the discharge
voltage by V measured from this waveform.
3.1.3 Interpretation of the Current and Voltage Waveforms
From the current and voltage waveforms measured, several system parameters of
the electrical discharge circuit as well as information concerning the dynamics of
the plasma can be deduced.
Before beginning a series of experiments, it is necessary to first determine the

basic parameters of the discharge system such as the circuit inductance and resis-
tance. This is done by creating a discharge condition whereby the plasma induc-
tance and resistance are negligible. Referring to the electromagnetic shock tube
again, this means shorting the tube at the back so that it is cut off from the discharge
circuit. Whatever inductance and resistance that are present now will be due to the
circuit external to the tube only. Hence from the discharge current waveform
obtained, which should be that of an ideal LCR discharge with fixed inductance and
pffiffiffiffiffiffiffiffi
resistance, Lo can be estimated from T since T ¼ 2p Lo C . Similarly from the
measurement of the damping factor α, the stray resistance of the circuit Ro can be
Ro
estimated since a ¼ 2L o
.
With the plasma discharge, if we assume the plasma to be inductive and the
plasma resistance is negligibly small such that Rp xLp , then
d
Vp ðtÞ ¼ Lp I :
dt
With I(t) and Vp(t) obtained, Lp(t) can be deduced from

R
Vp ðtÞdt
L p ðt Þ ¼ :
I ðt Þ
Lp(t) is related to the dynamics of the plasma. For example in the electromagnetic
shock tube,

lo b
Lp ðtÞ ¼ ‘n zðt Þ
2p a
so the trajectory of the electromagnetic piston (and hence the shock front) can be
deduced. Similarly, for the Z-pinch discharge, the radial trajectory of the current
sheet can be deduced from

2p Lp ðtÞ
rðtÞ ¼ ro exp :
lo ‘
For a resistive plasma such as the high pressure flash lamp where Lp is fixed,
Rp(t) can be deduced from

Vp ðtÞ Lp dIðtÞ
dt
Rp ðtÞ ¼ :
IðtÞ
Assuming the plasma to be a uniform column with length l and cross-sectional area
A, the resistivity of the plasma can be obtained:
l
g p ðt Þ ¼ R p ðt Þ ¼ 65:3Te5=2 ð‘nKÞZ:
A
With an estimate of Z (the charge state of the plasma) and taking ‘nK 10, the
temperature of the plasma Te(t) can be estimated.
3.2 Pulsed Magnetic Field Measurement
In a pulsed electric discharge, since the current is rapidly changing with time, a
pulsed magnetic field will be induced. The Rogowski coil is in fact a means to
measure this induced magnetic field in order to deduce the discharge current I(t).
Since the Rogowski coil is in the form of a solenoid bent into a torus encircling the
current, it measures the total discharge current and does not provide information
concerning the distribution of the current. In order to obtain information concerning
the distribution, or of greater interest the location of the current path such as the
location of the current sheet in the electromagnetic shock tube, a magnetic pick-up
coil can be used to measure the localised magnetic field.
The magnetic pick-up coil is a small diameter coil of several turns made from
insulated copper wire as shown in Fig. 3.12.
The principle of the magnetic pick-up coil is the same as the Rogowski coil, with
the output voltage induced given by
dBðtÞ
VðtÞ ¼ NA
dt
where N is the number of turns of the coil and A is the cross-sectional area of the
coil. B(t) is the time-varying magnetic field induced by the discharge current at the
location of the probe.
In order to obtain B(t) from the output voltage waveform of the coil, it is
necessary to perform integration as in the case of the Rogowski coil. However, in
this case it will not be possible to operate in the current transformer mode since the
Fig. 3.12 Example of a design of the magnetic pick-up coil. 1 BNC socket. 2 Copper adapter. 3
Perspex holder. 4 Glass tube
Fig. 3.13 Schematic circuit diagram of a magnetic probe
inductance of the coil is very low so the condition R ≪ ωL cannot be satisfied. As

such the magnetic pick-up coil is usually operated with an RC integrator as shown
in Fig. 3.13.
NA
In this case the final output voltage obtained will be Vo ðtÞ ¼ RC BðtÞ. The fre-
Z
quency range for the operation of the coil is RC \f \ L.
1
3.3 Plasma Spectroscopy
In this section, we will discuss briefly on some fundamental concepts and tech-
niques on the measurements of optical radiation emission from plasmas.
3.3.1 Plasma Radiation
The particles (electrons, ions and neutrals) are moving inside the plasma with
kinetic energy. Interactions between these particles, most likely involving collisions
of electrons with the heavier particles will result in various processes. Some of these
processes may lead to emission of photons in a wide range of spectrum from
infrared up to gamma ray depending on the temperature of the plasma. There are
basically three types of processes that will give rise to the emission of radiation
from the plasma, namely Bremsstrahlung, recombination and radiative decay.
3.3.1.1 Bremsstrahlung
In a plasma, electrons are moving under the effect of the electromagnetic field of the
bulk of the particles inside the plasma. They may encounter interaction that causes
them to be retarded. This retardation results in the release of one quantum of energy
(one photon) by the electron. This process is called Bremsstrahlung. Since the
electron emitting the photon is in the free state before and after the interaction, the
transition is often called a “free-free transition”. Bremsstrahlung may give rise to
photon of continuously varying energy and hence contributed to the continuum of
the photon energy spectrum.
3.3 Plasma Spectroscopy 63
The intensity of Bremsstrahlung emission (J per m3 per second) in the frequency

interval of v to v + dv caused by the interaction of Ne electrons/m3 with Ni ions/m3 is
given by
1=2
dEff v hm
¼ CNe Ni Zi2 H gff exp ðW m3 Þ ð3:20Þ
dm kTe kTe
where C = 1.7 × 10−53 J m3, Zi is the charge number of the ions, χH is the ionization
potential of hydrogen, and gff is the free-free Gaunt factor which represents the
departure of the quantum-mechanical calculation from the classical result averaged
over the Maxwell-Boltzmann velocity distribution at electron temperature Te. It is
normally more convenient to express the emission in terms of wavelength, and
Eq. (3.20) will then becomes
1=2
dEff v c hc 1
¼ CNe Ni Zi2 H gff 2 exp : ðW m3 Å Þ ð3:20aÞ
dk kTe k kkTe
If we consider that ionic species with charge states of Zm, Zm+1, … Zn are present
inside the plasma and their fractional abundances are αm, αm+1,…αn then the
average charge state of the plasma can be taken as
Pn
j¼m aj Zj
Zeff ¼ ð3:21Þ
NT
The total Bremsstrahlung intensity can be calculated by using Eq. (3.20a) with
Zi = Zeff. αj’s can be calculated using suitable plasma model such as LTE (Local
Thermodynamic Equilibrium) or CE (Coronal Equilibrium).
3.3.1.2 Recombination
Recombination is a process when a free electron of kinetic energy εe collides with

an ion (atomic species s and charge Zi) is captured in a bound level (principal
quantum number n and ionization potential vni1 ) of the ion of charge (Zi − 1)
resulting in the emission of a photon of energy hm ¼ ee þ vni1 . It can be seen that
the photon energy is a function of the electron energy which is continuous, and only
photons with energy greater than vni1 will be emitted. This results in discontinu-
ities, called recombination edges in the free-bound continuum spectrum.
The intensity of recombination emission (J per m3 per second) in the frequency
interval of v to v + dv caused by the interaction of Ne electrons/m3 with Ni ions/m3
(assuming recombination into the nth shell of a hydrogen like ion of charge i) is
given by
3=2 n 2 n
dEfb v vi1 fn v hm
¼ CNe Ni H gfb exp i1 ðW m3 Þ ð3:22Þ
dm kTe vH n kTe
where χH is the ionization potential of hydrogen atom; vni1 is the ionization

potential of ionic specie at charge state of Zi−1 from bound state with principal
quantum number n; and ζn is the number of vacancies available for recombination
with electron at that bound state. C is the same constant as in expression (3.20).
Expressed in terms of wavelength, it becomes
3=2 n 2 n
dEfb vH vi1 fn c vi1 hc 1
¼ CNe Ni gfb 2 exp k
: ðW m3 Å Þ
dk kTe vH n k kT e
ð3:22aÞ
Similarly, the total intensity of the recombination emission can be obtained by

consideration of the contribution from various ionic species present.
3.3.1.3 Radiative Decay
An atom or ion collided by electron may be excited and subsequently undergoes

radiative decay which results in the emission of a photon. The energy of this photon
corresponds to the energy difference between the two energy levels so it is char-
acteristic of the emitting atom or ion and it has a discrete value. The transition is
between two bound states and it is called bound-bound transition.
The intensity of line emission (J per m3 per second) at the frequency ν due to
transition from level p to level q is given by
hc
Ipq ¼ ðhmÞAðp; qÞNðpÞ ¼ Aðp; qÞNðpÞ ðJ m3 Þ ð3:23Þ
k
where A(p, q) is the transition probability, N(p) is the number density of the upper
energy state.
3.3.1.4 Some Useful Facts About Plasma Emission Spectrum
(i) The shapes of the Bremsstrahlung and recombination emission spectra are
identical except at the recombination edges.
(ii) The peak of the plasma emission continuum occurs at wavelength of
6200
ko ¼ ðÅÞ; ð3:24Þ
Te
where Te is in eV.
(iii) The contribution of Bremsstrahlung towards the continuum will be dominant

at high temperature. That is
dEff dEfb
when kTe [ 3Zi2 vH : ð3:25Þ
dm dm
3.3.2 The Plasma Models
In order to calculate the radiation spectrum emitted from the plasma, it is necessary
to know what are the species present and what is their fractional distribution. The
real plasma state is extremely complicated that no one model is able to give an
accurate and complete description. In this section we will discuss two approximate
plasma models.
3.3.2.1 Local Thermodynamic Equilibrium (LTE)
Ideally, life will be easier if a state of complete thermodynamic equilibrium

(TE) can be satisfied by the plasma. Then the plasma state can be described by a set
of finite number of thermodynamic variables such as the temperature, the pressure
and the concentrations of various elements present while the plasma emission can
be approximated to the blackbody radiation. One necessary condition of this situ-
ation is that the whole plasma is homogeneous and radiation emitted from the
interior of the plasma will be completely re-absorbed by the plasma itself. In reality
this is not the case. The best that can be achieved is to divide the volume of the
plasma into small elements of volume so that the plasma inside each element can be
considered to be homogeneous and yet its dimension is larger than the mean free
path of the particle or photon. This the case of the LTE plasma model.
We consider the ionization of the ith ionization state of the element when it is
collided by electron inside the plasma concerned. If the pressure (and particle
density) of the plasma is sufficiently high, three body recombination will occur and
balance the ionization process:
Ai þ e , Ai þ 1 þ e þ e
Assuming that all species are present at their ground state only, the ratio of the
densities of two species with consecutive ionic states (i and i + 1) is given by the
Saha equation:

Ni þ 1 2 Ui þ 1 2pme kTe 3=2 vi
¼ exp ð3:26Þ
Ni Ne Ui h2 kTe
where U’s are the respective partition functions of the two species and Ei is the
ionization potential [for ionization from ith ionic state to (i + 1)th state]. Within
each specie, the distribution of population densities among the bound levels is given
by the Boltzmann relation:

NðpÞ gp Ep Eq
¼ exp ð3:27Þ
NðqÞ gq kTe
where p and q refer to two energy levels with energies Ep and Eq and statistical
weights gp and gq respectively. For each level, the number density can be written as:

gp Ep
NðpÞ ¼ N exp ð3:28Þ
Ue kTe
where Ue is the electronic partition function of the specie which is given by

X
Ej
Ue ¼ gp exp ; ð3:29Þ
all j
kTe
and N is the total number density given by

X
N¼ NðjÞ: ð3:30Þ
all j
Note that in (3.26), Ui is the electronic partition function of the specie with ith ionic
state.
In addition to the three body recombination that occurs as the inverse of electron
collisional ionization, radiative recombination is also taking place simultaneously.
However, in a high density plasma, the probability for radiative recombination to
occur is very much lower than the three body recombination. This condition can be
shown to be true when
Ne 1:6 1012 Te1=2 Epq

3
cm3 ð3:31Þ
where Epq is the highest energy gap in eV, normally the energy gap between the
first two levels. Te is the electron temperature in K.
3.3.2.2 Corona Equilibrium (CE) Model
In a low density plasma, where collision does not occur with sufficiently high fre-
quency to enable the establishment of thermodynamic equilibrium (not even local),
the plasma must be described by non-thermal or non-LTE model. One of the fre-
quently used model which was first proposed to explain certain features of the solar
corona is the Coronal Equilibrium Model. It is assumed that the changes in the plasma
condition occur sufficiently slowly so that the electrons are always at thermal equi-
librium among themselves so that they will have a Maxwellian velocity distribution.
Again assuming that the ionic species are present in their ground states only, and
electron collisional ionization is always balanced by radiative recombination, then
we may write
Ne Ni SðTe ; iÞ ¼ Ne Ni þ 1 RðTe ; i þ 1Þ ð3:32Þ
where S(Te, i) is the collisional ionization coefficient and R(Te, i + 1) is the radiative
recombination coefficient. Hence the ratio of the densities of two consecutive ionic
species can be obtained from
Ni þ 1 SðTe ; iÞ
¼ : ð3:33Þ
Ni RðTe ; i þ 1Þ
The expressions for S(Te, i) and R(Te, i + 1) are obtained empirically and substi-
tuting them into Eq. (3.33) gives rise to
3=4
Ni þ 1 8 1 kTe vi
¼ 1:27 10 2 exp : ð3:34Þ
Ni vi vi kTe
As we mentioned earlier, it was assumed that the plasma condition changes slowly
so that equilibrium can be achieved. The time for the plasma to achieve this
equilibrium is the atomic relaxation time given by
1012
s : ðsÞ ð3:35Þ
Ne
If the plasma lifetime is shorter than this relaxation time, a time-dependent Corona
Model may be required to describe the plasma. On the other hand, if it is desired to
include effect of electron collision processes causing transition between upper
levels, the Collisional Radiative (CR) model may be appropriate.
In spectroscopic experiments of plasma, it is often required to base the analyses
of data on a particular plasma model. For simplicity, we may decide to choose
between either the LTE model or the CE model. In this case we can base our choice
on expression (3.31). This means that if the density of the plasma is high and satisfy
the condition stated in (3.31), we choose LTE model. Otherwise, the CE model will
be used. However, we should remember that this treatment is not accurate.
3.3.3 Examples of Population Density Distribution

and Plasma Spectra
3.3.3.1 Population Density Distribution
For the following examples, the fraction of ionization is defined as ai ¼ NNit where
P
Nt ¼ all i Ni , with i = 0 denoting neutral. Figure 3.14a, b show the population
density distributions for argon and carbon plasmas according to the CE model.
We can see from the comparison of these two sets of graphs that the argon
plasma is fully ionized at electron temperature above 10 keV while the carbon
plasma is almost fully ionized at an electron temperature of above 200 eV.
Fig. 3.14 a Argon plasma (CE Model). b Carbon plasma (CE Model)
3.3.3.2 Examples of Plasma Emission Spectra
Figure 3.15 shows the spectrum of an argon plasma at an electron temperature of

1 keV and electron density of 1019 cm−3. All three types of emission:
Bremsstrahlung, recombination and line radiation are included. The spectrum is
computed assuming CE model which predicts that at this temperature the most
prominent species are the Ar16+ (Helium-like) and Ar17+ (Hydrogen-like) with
characteristic lines in the X-ray region. The peak of the continuum is expected to be
at about 6 Å. When the temperature is increased to 2 keV, it is clear that the
spectrum has shifted to shorter wavelength, while the recombination edge and the
line radiation are not affected in terms of their wavelengths (Fig. 3.16).
Wavelength (Å)
Fig. 3.15 Simulated spectrum of an argon plasma at electron temperature of 1 keV
Fig. 3.16 Simulated spectrum of an argon plasma at electron temperature of 2 keV

Fig. 3.17 Simulated spectrum of a carbon plasma at electron temperature of 150 eV
Figure 3.17 shows the example of the emission spectrum of carbon plasma at
150 eV computed assuming CE model.
For this plasma, the prominent species are clearly C6+ and C5+ together with
small fraction of C4+.
3.3.4 Optical Emission Spectroscopy of Plasma
In principle, the electron temperature of the plasma can be estimated from the peak
of the continuum if it can be determined with reasonable accuracy. This is normally
difficult to accomplish particularly for low temperature plasma where the spectrum
is recombination and line radiations dominated. One of the most commonly used
method to determine the electron temperature of low temperature plasma is the line
ratio method. This involves the measurement of line radiation emission spectrum of
the plasma (normally in the visible region for low temperature plasma) and the
electron temperature can be deduced from the ratio of a pair of nearby spectra lines.
For a continuous plasma such as the DC glow or arc discharge, a simple setup
consisting of a monochromator (also known as spectrograph) coupled to a single
channel photomultiplier tube will be sufficient. In this case the light emitted from the
plasma is collected and fed to the monochromater through an input (or entrance).
After entering the monochromator, the light is reflected and focussed onto a grating
which acts as the dispersion medium. The dispersed light of various wavelengths
will then be collected at the exit window. The range of wavelengths that will appear
in the exit window is selected by tilting the plane of the grating at different angle. The
overall range of wavelengths that may be detected by the monochromator, as well as
its resolution, will be determined by the groove density (grooves/mm) of the grating.
For the setup using single channel photomultiplier tube to measure the spectral
intensity as a function of wavelength, an exit slit is used to limit the view of the
photomultiplier so that only photons of a particular wavelength will be detected.
Since the width of the slit is finite, this will introduce the second factor of instru-
mental width to the spectral line. The entrance slit gives rise to the first factor of
instrumental spectral width. The mechanism for the tilting of the grating is con-
nected to a dial with reading calibrated to give the wavelength directly. When the
grating is being tilted, each of its angular position is related to the wavelength of the
photon as given by the diffraction formula 2d sin h ¼ nk so that at each angular
position, only the wavelength that satisfies this condition will be “seen” by the
photomultiplier through the exit slit. Hence the spectrum can be scanned by
adjusting the dial for the case of DC plasma. The similar procedure can be followed
for the RF plasma as well.
Recently, with the availability of photo-diode-array (PDA) and CCD, optical
multi-channel analyser (OMA) is often used instead of the single channel PMT to
measure the plasma emission spectrum. Figure 3.18 shows an example of such a
setup used for measuring the glow discharge emission spectrum. In this setup a
range of the spectrum can be measured simultaneously. The range is limited by the
width of the exit window of the monochromater, but more often by the length of the
PDA or CCD. The commonly available PDA and CCD have 1024 pixels, which
mean 1024 points are acquired simultaneously. With smaller pixel size and closely
packed pixels, better resolution can be achieved but we will have to compromise
with a narrower measurable spectral range.
In order to use the setup shown in Fig. 3.18 to determine the electron temper-
ature of, say argon glow discharge plasma (DC or RF), we first consider expression
(3.23) and assume the emission to be isotropic, then the emission per steradian
(W/cm3/sr) is
hcAðp; qÞ
Ipq ¼ NðpÞ: ð3:36Þ
4pk
Fig. 3.18 Schematic setup of

optical emission spectroscopy
of plasma
Lets say we select four lines:
k1 ¼ 8014:79 Å ðAr)
k2 ¼ 8006:16 Å ðAr þ Þ
k3 ¼ 8103:69 Å ðAr)
k4 ¼ 7948:18 Å ðAr þ Þ
For line 1, with the assumptions that the plasma is at LTE and it is optically thin, we
can write
hcA1 1
I1 ¼ N ð3:37Þ
4pk1 o

where No1 is given by expression (3.28) as No1 ¼ No Ug1o exp kT
Eo
e
. The ratio of the
intensities of line 1 to line 2 is given by

I1 A1 k2 No g1 U1 E1 E2
R1 ¼ ¼ exp ð3:38Þ
I2 A2 k1 N1 g2 Uo kTe
where E1 is the energy of the upper level involved in the transition giving rise to
line 1, while E2 is that of line 2. It can be seen that in order to calculate R1 for a
particular electron temperature, we need to know the ratio of No/N1 which has to be
calculated from expression (3.26). Although this is possible, we may also get
around this by taking the ratio of a second pair of lines:

I3 A3 k4 No g3 U1 E3 E4
R2 ¼ ¼ exp : ð3:39Þ
I4 A4 k3 N1 g4 Uo kTe
Taking ratio of R1 and R2 will then give:

R1 A1 A4 k2 k3 g1 g4 E1 E2 þ E4 E3
R¼ ¼ exp : ð3:40Þ
R2 A2 A3 k1 k4 g2 g3 kTe
The values of R for a range of Te can be calculated and plotted as shown in

Fig. 3.19.
Experimentally, the intensity of the four lines will be measured and the value of
R obtained from Eq. (3.40). The electron temperature of the plasma can then be
estimated by using the “calibration curve” in Fig. 3.19.
Fig. 3.19 Optical line ratio plot for electron temperature determination
An example of the spectrum of a high pressure argon spectral lamp in the region
of these 4 lines is shown in Fig. 3.20.
The electron temperature of the argon plasma of the high pressure arc lamp is
estimated to be 0.4 eV.
Fig. 3.20 Time integrated optical lines detected from a low temperature plasma
3.4 The Langmuir Probe (Electric Probe)
3.4.1 Electron Temperature and Density Measurements

of Steady State Plasmas
The Langmuir probe, also often known as the electric probe, is one of the frequently
used diagnostic tools for measuring the characteristics of low temperature plasmas.
The fundamental plasma parameters, such as the electron density, electron tem-
perature, plasma potential and in some cases the electron energy distribution
function can be determined by using the Langmuir probe.
The theory of the flow of charge carriers to an electric probe is extremely
complex. A simplified theory can be derived with the following assumptions:
i. The electron and ion concentrations are equal.
ii. The mean free paths of the electrons and ions are much larger than the probe
dimensions.
iii. The electron temperature is much larger than the ion temperature.
iv. The probe radius is much larger than the Debye length of the plasma.
v. The kinetic energy of electrons and ions obey Maxwell-Boltzmann distribution.
The Langmuir probe can be either cylindrical, planer or spherical in shape. In its
simplest form, it consists of a short length of thin wire inserted into the plasma,
which is biased at a potential with respect to a reference electrode (usually the
electrically grounded electrode) in order to collect electron and/or positive ion
current. As the wire is in contact with the plasma, the dimension of the probe must
be chosen so as to cause minimum disturbance to the plasma. The material used for
the probe construction is such that they do not melt or easily sputtered so as to avoid
introducing impurity to the plasma. The commonly used materials for probe con-
struction include tungsten, molybdenum, tantalum etc. The probe wire of 0.1–1 mm
diameter is embedded in an insulating tube such as alumina or quartz so that it is
insulated from the plasma except for a short length of exposed tip which is
2–10 mm in length. Such a probe is capable of providing a localized measurement
of the fundamental plasma parameters such as electron temperature, electron den-
sity as well as plasma potential.
A simplified schematic of a typical probe circuit and the probe
I–V characteristics are shown in Figs. 3.21 and 3.22, respectively. The probe system
consists of the probe, a variable power supply and an arrangement to measure the
current drawn by the probe. The variable voltage source should be able to provide
output voltage that can be varied from positive to negative range with respect to
ground, either discretely or scan continuously. Depending on the polarity of the
biasing voltage, the probe draws either positive or negative current from the plasma.
Figure 3.22 shows an idealized probe I–V characteristics labelled at various
regions by A, B, C, D, and E. Along the X-axis V (or VP) represents the probe
voltage or the probe bias voltage, while along the Y-axis I represents the total probe
current. The probe current is due to the flow of both ions and electrons drawn from
3.4 The Langmuir Probe (Electric Probe) 75
Fig. 3.21 Schematic setup of

Langmuir probe measurement
of a glow discharge
Fig. 3.22 Typical

I–V characteristic curve of
Langmuir probe (Single)
the plasma in the vicinity of the probe hence it can be expressed as I = Ie − Ii, where
Ie is the electron current and Ii is the ion current. In Fig. 3.22, the ion current to the
probe has been assigned to be negative while the electron current is assigned to be
positive. Region AB is called the electron saturation region, BCD the electron
retardation region and DE the ion saturation region. The floating potential of the
probe is represented by Vf (point D). The floating potential Vf is the voltage at which
the ion and electron currents are equal and hence the net probe current is zero. The
plasma potential VS (point B) is the potential of the plasma at the location of the
probe with respect to the reference electrode.
The electron temperature of the plasma can be determined by plotting ln(Ie)
against V for the electron retardation region (BCD) as shown in Fig. 3.23.
According to the probe theory, the electron current in the electron retardation
region is given by
rffiffiffiffiffiffiffiffiffiffiffiffi
eV k Te
Ie ¼ Ie0 exp ; where Ie0 ¼ A ne e :
kTe 2p me
Here A is the effective probe area, ne is the electron density, Te is the electron
temperature, while e, k and me are respectively the electronic charge, Boltzmann
Fig. 3.23 Analysis of the

I–V probe characteristic for
the determination of electron
temperature
constant and electron mass. Ie0 is called the electron saturation current, which is the
probe electron current at point B. Point B can be identified to be at the point where
the gradient changes, which occurs at the point where the probe potential is Vd,
which is the plasma potential Vs with corresponding current Ie0. The plot of ln(Ie)
against V for the electron retardation region gives rise to a linear graph with gradient
given by e/kTe, hence the electron temperature Te can be obtained. By measuring
the electron saturation current Ie0 at point B of the probe I–V characteristics, and
knowing the probe dimension, the electron density ne can be estimated.
3.4.2 Dynamic Studies of a Flowing Plasma (or Charged

Particle Beam)
The electric probe measurement of IV characteristic and its analysis outlined in

Sect. 3.4.1 can be used to measurement of electron temperature and density of a
steady state plasma with low temperature and density such as the glow discharge
plasma. It is assumed that the energies of the particles inside the plasma, both
electrons and ions, are obeying Maxwell-Boltzmann distribution function. For a
flowing plasma with time varying temperature and density, the electric probe can
still be employed to study the dynamics of the plasma such as those in the shock
tube. For the electromagnetic shock tube, due to the presence of the electric and
magnetic field, the probe signal may become complicated. However, for a free
flowing plasma consisting of charged particles, or a beam of charged particles (ions
or electrons) flowing in a region without electric and magnetic fields, the electric
probe can be used to detect the movement of the flowing plasma or charged
particles.
An example of using the electric probe for the detection of ion beam from the
plasma focus (Refer to Sect. 4.2, Chap. 4) is illustrated in Fig. 3.24 [2]. Two
identical probes are placed along the path of the ion beam and each is biased at a
suitable negative potential, say −100 V. The probes are expected to attract the
3.4 The Langmuir Probe (Electric Probe) 77
Fig. 3.24 Arrangement for time-of-flight measurement of ion beam energy using electric probes
(Copyright 2002 The Japan Society of Applied Physics)
positively charged ions when the beam passes through the probes and produce two
voltage pulses recorded by the oscilloscope. The probe bias circuit used is as shown
in Fig. 3.25. By positioning the probes at a fixed distance apart along the path of the
ion beam, the time taken by the beam to travel from one probe to another (Δt) can
be measured using the oscilloscope, and hence the kinetic energy of the ions can be
estimated. This is illustrated in Fig. 3.26.
Fig. 3.25 The schematic biasing circuit for electric probe used for dynamic study of pulsed ion
beam
1.5
(a) Δt
1.0
Ion collector signal (V)
0.5
0.0
2.0 1.5 2.0 2.5 3.0 3.5
(b)
1.5
1.0
0.5
0.0
1.5 2.0 2.5 3.0 3.5
Time (us)
Fig. 3.26 Typical signals of electric probe measurement of ion beam
3.5 X-ray Diagnostic Techniques
As we have seen in the previous section, the plasma emission spectrum will shift
towards higher photon energy when the electron temperature is increased. At an
electron temperature of higher than 1 keV, the plasma emission is predominantly in
the X-ray region, typically in the wavelength range of 1–100 Å. Hence for high
temperature and high density plasma produced by pulsed discharge such as the
plasma focus and the vacuum spark, the measurement of the X-ray emission will
provide useful information concerning the condition of the plasma. In addition, in
consideration of possible application of these pulsed plasmas as X-ray sources, it is
important to determine its X-ray emission characteristics such as the intensity, the
source size as well as the spectrum.
Figure 3.27 shows the typical arrangement for measuring the X-ray emission
from a plasma source.
Source Filter Detector

Fig. 3.27 Illustration of experimental layout for X-ray detection of plasma
3.5 X-ray Diagnostic Techniques 79
In this arrangement, the detector used is spectral integrated and the filter will
provide exponential filtering. In this case, the X-ray is considered to be generated
from the plasma by Bremsstrahlung and recombination discussed in the earlier
sections. The contribution from the line radiation will be insignificant when we
integrate over the whole spectrum.
3.5.1 X-ray Absorption Filter
The interaction of X-ray photon with the filter material is basically the photoelectric
effect where the atom absorbs the photon energy and results in the removal of one
of the K-shell electrons. Some of the photons may thus be stopped during their
passage through the filter. For lower energy photon (longer wavelength), a limit will
be reached where the photon energy is not enough to remove the K-shell electron.
This is called the K-absorption edge. At further lower photon energy, another limit
will be reached for the L-shell electrons. The K-absorption edge is closely related to
the ionization energy of the electron in the K-shell, which can be taken to be
approximately equal to 13.6 × Z2 eV considering the K-shell ionization to be
“hydrogen-like” and ignoring the screening effect due to the other electrons (which
will cause a lowering of the ionization energy).
The absorption of photons passing through a particular filter can be expressed as
I ¼ Io expðlxÞ ð3:41Þ
where I is the intensity of X-ray after passing through the filter, Io is the X-ray
intensity before passing through the filter, x is the thickness of the filter in cm and µ
is the absorption coefficient in cm−1. A more commonly used form of the absorption
coefficient is the mass absorption coefficient µm (cm2/g) which is equal to µ/ρ where
ρ is the density of the filter material in g/cm3. In this case expression (3.41) will be
written as
I ¼ Io expðlm xm Þ ð3:42Þ
xm = ρx is called the mass thickness expressed in g/cm2. The values of µm for

various materials are determined experimentally only for certain limited and scat-
tered ranges of wavelength. Their values at other wavelengths are either obtained by
interpolation and extrapolation, or by using some empirical expressions. For filter
of composite material XmYn, the resultant X-ray absorption coefficient is written as

m n
lxy ¼ l þ l: ð3:43Þ
mþn x mþn y
As an example, the mass absorption coefficient of mylar (C10H8O4) as a function of

wavelength is shown in Fig. 3.28.
Fig. 3.28 Mass absorption

coefficient of Mylar as a
function of X-ray wavelength
The X-ray transmission through various sets of filters are illustrated graphically
by plotting expression (3.42) with known values of µm as shown in Fig. 3.29. In
particular, we note the following three sets of filters: (1) 149 µm aluminized-Mylar;
(2) 24 µm aluminized-Mylar + 10.5 µm Ti; and (3) 1024 µm aluminized-Mylar. If
these filters are used simultaneously with identical detectors, their signals can be
used to identify the presence of argon Kα line radiation in an argon plasma or an
argon-hydrogen admixture plasma, for example. This is due to the fact that the
argon Kα line radiation which has a wavelength of around 4 Å, will be within the
transmission “window” of Ti. With the above sets of filters, the signals registered by
the detector with filters (1) and (2) should be almost identical if the argon Kα line
radiation is dominant. On the other hand, if copper Kα line radiation is prominent,
the signals of detectors with filters (2) and (3) will be identically weak while that of
detector with filter (1) will be strong. It should also be noted that all materials will
transmit its own Kα line radiation well.
3.5.2 X-ray Detectors
3.5.2.1 Time-Resolved X-ray Detectors
It is very useful to obtain information concerning the time-evolution of the X-ray

emission from the plasma as this can often reveal the dynamic behaviour of the
0.1
Transmission
0.01
24 um Al-Mylar
24 um Al-Mylar + 10.5 um Ti
149 um Al-Mylar
24 um Al-Mylar + 10 mm Cu
974 um Al-Mylar
1E-3
0.1 1 10
WAVELENGTH (A)
Fig. 3.29 X-ray transmission curves for various sets of absorbers
plasma. Several common X-ray detectors used for this purpose include the pho-
tomultiplier tube, silicon PIN diode and the X-ray diode (XRD).
(a) The photomultiplier tube
The photomultiplier tube (PMT) is a sensitive photon detector which is capable of
converting photons into electrons using a photo-cathode and subsequently multiply
them through multiple stages of dynode arrangement. An example of the photo-
multiplier tube is shown in Fig. 3.30. The PMT requires a “base” which normally
Fig. 3.30 Schematics of a typical photomultiplier tube

consists of a resistive divider network to provide a distribution of potential across

the cathode and the first dynode as well as between consecutive dynodes. The
potential differences between the cathode and the first dynode provides the first
acceleration of the photoelectrons and subsequently an almost constant potential is
maintained between two consecutive dynodes until finally the electrons are col-
lected by the anode.
For the detection of X-ray, the PMT is used together with a scintillator. This is
because normally the photocathode is sensitive to photon in the visible to UV
region. The scintillator is capable of absorbing photon (X- and γ-rays) or particles
(α, β and neutron) of sufficient energy and emit in the visible region (blue to violet).
For X-ray, it converts photons with wavelength in the region of 1–100 Å to around
4000 Å which is within the sensitivity region of the PMT’s photocathode.
As we can see, the detection of X-ray using PMT involves several steps of
conversion. First the X-ray photon is absorbed by the scintillator which gives out
blue-violet photon (conversion factor εs). Second, the quantum efficiency (εq) of the
photocathode should be considered. Thirdly, the electron multiplication factors of
the PMT (εm). And finally, the electrons are collected by the anode and the signal
obtained (voltage) is caused by the flow of these electrons in the circuit (current). It
is hence difficult, though not impossible, to calculate the overall sensitivity of the
whole detection system.
Furthermore, due to the transit time required for the electron multiplication
process inside the PMT, its signal is always delayed by a duration of the order of
50–200 ns, depending on the number of multiplication stages. More stages means
higher gain but longer time delay. A delay of around 100 ns can be expected for a 9
stages PMT.
However, despite the fact that the PMT system is normally bulky in size, pro-
duces a delayed signal and it is difficult to determine its absolute response, it is
useful if we need to measure the harder component of the X-ray emission spectrum.
Figure 3.31 shows an example of the hard X-ray signal (lower trace) of a plasma
focus discharge recorded by a PMT system [3]. The upper trace shows the voltage
spike which is the signature of the focussing action. The timing of the PMT signal
has to be adjusted to align with the voltage signal.
Fig. 3.31 Example of the

PMT signal obtained from a
plasma focus discharge
(b) Silicon PIN diode

The silicon PIN diode is a convenient X-ray detector which is compact, fast rise
time and with known response. It has an almost flat response (except for the silicon
absorption edge) between 1 and 20 Å. The PIN diode is basically similar to a PN
junction diode, but with a large intrinsic layer sandwiched between the P-type and
the N-type layer. This thick intrinsic layer is completely depleted and it will absorb
the X-ray photon to form an electron-hole pair.
The structure of the PIN diode is shown schematically in Fig. 3.32 together with
the energy level diagram when a reverse bias voltage is applied across the diode.
Electron-hole pairs will be produced when X-ray photons are absorbed by the
diode, with a response of 1 electron-hole pair produced for each 3.55 eV of photon
energy absorbed. This means electron charge of 0.28 C will be produced for each
joule of X-ray energy absorbed.
To calculate the sensitivity of the diode to X-ray photons, the absorption of
X-ray at the N-type entrance window which acts as filter must be considered. The
electron-hole pair formation will correspond to the amount of photons absorbed by
the intrinsic layer and hence the sensitivity S(λ) of the diode is given by
SðkÞ ¼ 0:28 expðls x1 Þ½1 expðls x2 Þ

ðC=JÞ ð3:44Þ
where μs is the mass absorption coefficient of silicon, x1 is the mass thickness of the
N-type entrance window while x2 is the mass thickness of the intrinsic layer. An
example of the sensitivity curve of a typical X-ray PIN diode is shown in Fig. 3.33.
At the short wavelength region (λ < 1 Å), the X-ray photons may pass through
the intrinsic layer without being absorbed so the sensitivity drops. On the other
hand, for long wavelength region (λ > 20 Å), the X-ray photon will be absorbed by
the silicon entrance layer and hence the sensitivity will also drop. In the region
between this two limits, the sensitivity of the diode is almost flat, except at the
K-absorption edge of Si at 6.73 Å.
The X-ray intensity from the plasma source P(λ, Te) W cm−3 Å−1 actually
detected by the PIN diode with due consideration of the absorption by foil placed
between the source and the detector and the solid angle (assuming point source) is
then given by Pðk; Te ÞSðkÞ expðlxÞ. Figure 3.34 show the example of the
intensity of X-ray emitted by an argon plasma at electron temperature of 1 keV
detected by a PIN diode covered with various sets of filters.
A commonly used biasing circuit for the PIN diode is as shown in Fig. 3.35.
Fig. 3.32 Simplified schematics of a PIN diode

Fig. 3.33 Theoretical 1

sensitivity curve of an X-ray
PIN diode
0.1
Sensitivity (C/J)
0.01
1E-3
0.1 1 10
Wavelength (A)
Fig. 3.34 Calculated X-ray

spectra detected by PIN diode
after various sets of filters
Fig. 3.35 Schematics of a typical PIN diode biasing circuit
In this circuit, the charges produced by the diode due to the absorption of X-ray
energy will flow through the circuit and the output voltage is measured across the
51 Ω resistor.
An example of the X-ray pulse measured is as shown in Fig. 3.36. The amount
of charges produced by the PIN diode due to the absorption of X-ray photons is
given by:
Z R
VðtÞdt
Q ¼ IðtÞdt ¼ :
51
If the source is a monochromatic point source with wavelength λ = λo (line radi-

ation), then the charge Q can be used to calculate the amount of X-ray energy
absorbed:
Q
Ek00o ¼ :
Sðko Þ
Corrected for foil absorption:
Q 1
Ek0 o ¼
Sðko Þ expðlxÞ
V(t) = I(t) × 50
∫ V ( t ) dt
Fig. 3.36 Example of a PIN diode X-ray pulse registered by the oscilloscope
and finally, the total X-ray energy emitted by the point source isotropically into 4π is:
Q 1 4pd 2
E¼ ð3:45Þ
Sðko Þ expðlxÞ A
where d is the source-detector distance, and A is the effective area of detection of

the diode as shown in Fig. 3.27.
Fig. 3.37 The ratios of a pair of identical PIN diode filtered by two sets of filter as a function of
electron temperature
For plasmas such as the plasma focus and the vacuum spark, it is clear that the
spectrum is not monochromatic. If d ≫ dimension of the source which is of the
order of mm (end-on) for the focus and a fraction of mm for the vacuum spark,
the point source assumption may still be maintained. For such cases, we consider
the continuum (Bremsstrahlung + recombination) and we write the rate of charge
production (current) due to the time varying X-ray pulse emitted from the plasma:
Z
dQ VðtÞ A
¼ ¼ Vplasma Pðk; Te Þ expðlxÞ SðkÞdk: ð3:46Þ
dt 51 4pd 2
all k
This can be computed for a fixed Te. For two identical diodes (or any other
detectors) coupled to two different sets of foils, the ratio of the current generated by
the two diodes due to the same X-ray pulse can then be calculated as R = I1/I2. This
can be repeated for a range of temperature, say from 500 eV up to 10 keV, the
values of R as a function of Te can be obtained and plotted as shown in Fig. 3.37.
This can be used as calibration curve for the measurement of electron temperature
by the X-ray ratio method or X-ray foil absorption technique. For this case, two
channels of PIN diode are required.
Alternatively, R for filters of same material (e.g. Al) but varying thickness can be
computed and plotted as R against foil thickness for a particular Te. This can be
repeated for a range of temperature. The set of graphs for Te = 500, 750, 1000,
2000, 5000 and 10,000 eV are as shown in Fig. 3.38. The graph for monochromatic
sources of Ar Kα and Cu Kα are also shown in this figure.
Experimentally, this requires more than 2 channels of detectors. For the case of 5
channels of detectors, with one channel as reference, 4 points can be obtained
experimentally and plotted superimposed onto the ratio graphs in Fig. 3.38. The
experiment is carried out on a plasma focus discharge and the signals of the five
channels PIN diodes are shown together with the voltage signal as shown in
Fig. 3.39 [4].
From this experiment, five X-ray pulses are observed, of which pulse 1 and pulse
2 can be identified to be corresponding to hot plasmas with electron temperatures
between 1 and 2 keV. Pulse 4 and pulse 5 which occurs at time long after the focus
event is over, are found to be close to the Cu Kα line indicating that they consist of
predominantly the Cu Kα line radiation. These are believed to be originated from
dense (but not so hot) plasma jet of plasma vapour.
3.5.2.2 Time-Integrated X-ray Imaging
The time-resolved measurement of X-ray emission from plasmas described above

provides information concerning the time evolution of plasma but it does not tell us
information such as the size and the structure of the X-ray emitting plasma. Such
kind of information must be obtained by imaging the plasma. A very simple setup
Fig. 3.38 The ratios of a series of PIN diodes filtered by aluminium foils of varying thickness
plotted for various electron temperatures. Also plotted on the same graph are the straight line
curves corresponding to mono energetic photons of Ar Kα and Cu Kα line radiations
using time-integrated X-ray film and the primitive pinhole imaging method can be
used to obtain the image of the plasma. Recently, with the availability of gated
micro-channel-plate (MCP), time-resolved image recording can be achieved when
coupled to well designed multiple pinhole system. However, we will illustrate here
only the basic technique of time-integrated X-ray imaging, that is the X-ray pinhole
imaging using X-ray film.
The principle of normal pinhole imaging is straight forward. In this case, the size
of the pinhole is much smaller than that of the object and the image to object ratio is
given by the image-pinhole distance to object-pinhole distance ratio. However, if
Fig. 3.39 Example of X-ray pulses from 5 channels of PIN diode filtered by aluminium foils of
varying thicknesses corresponding to a single plasma focus discharge
Fig. 3.40 Schematics of an

X-ray pinhole imaging setup
with pinhole size larger than
the image size
the size of the object is smaller than the size of the pinhole, the ray diagram is as
shown in Fig. 3.40.
From the geometry, we see that the size of the object can be estimated from

Lo ðq pÞ
x¼ p for x\p
Li
and Lo is the object-pinhole distance

where Li is the image-pinhole distance (=a)
(=b + c). If x > p, we simply have x ¼ LLoi q.
An example of the X-ray pinhole image of the vacuum spark plasma is shown in
Fig. 3.41.
In this image, we can clearly see 2 types of structures: (i) a bright spot with sharp
circular (almost) boundary, and (ii) a diffuse cloud of intensity. The bright round
spot is believed to be caused by the commonly observed hot spot formed in the
vacuum spark plasma, which is smaller than the size of the pinhole which has a
diameter of 300 μm. In fact, the sharp circular structure suggests that it is actually
the projection of the pinhole onto the X-ray film by a source much smaller than the
pinhole size. On the other hand, the plasma cloud is much larger in dimension than
300 μm so this part of the image shows the actual structure of the plasma cloud
observed in the vacuum spark.
Fig. 3.41 Example of the

X-ray pinhole image obtained
from a vacuum spark
discharge
3.6 Neutron Diagnostic Techniques 91
3.6 Neutron Diagnostic Techniques
3.6.1 The Foil Activation Neutron Detector for Absolute

Neutron Yield Measurement
Many nuclides such as 115In, 107Ag, 109Ag and 103Rh have large cross section for
the (n, γ) reaction with thermal neutron (<1 keV). The products of such a reaction
are the gamma ray and the radioactive isotope of the target nuclide. By measuring
the activity of the radioactive isotope for a few half-lives of its decay, the incident
neutron flux can be deduced. The activation foil used is in the form of a thin foil
with less than 1 mm thickness so as to reduce the self absorption of the radiation by
the material.
Consider a piece of foil consisting of NT number of nuclides. When this foil is
exposed to thermal neutrons of flux density ϕ with its plane perpendicular to the
direction of the neutrons, the rate of change of induced radionuclides in the foil at
any instant is given by:
dN
¼ NT ra / kN ð3:48Þ
dt
where
N is the number of radionuclides at any instant,
σa is the average thermal neutron activation cross section,
λ is the decay constant of the radionuclides.
The first term on the right hand side of Eq. (3.48) is the activation term while the
second term represents the decay of the activated radionuclides which occur
simultaneously with the activation process. The number of radionuclides in the foil
after activation for a time of t can be obtained by solving Eq. (3.48). The solution
takes the form:
NT ra /
N ðt Þ ¼ 1 ekt : ð3:49Þ
k
This can be represented graphically as shown in Fig. 3.42.

If t is sufficiently large, the value of N(t) approaches a constant value known as
the saturation activity given by
NT ra /
Ns ¼ : ð3:50Þ
k
Practically, more than 95 % of this saturation value is achieved by activating the foil
for a time of about five half-lives.
We consider neutron activation by a continuous source and a pulsed source.
Fig. 3.42 Variation of

number of neutron activated
radionuclides with exposure
time
3.6.1.1 Steady Source
For a steady and isotropic neutron source with strength S neutrons per second
placed at a fixed distance from the foil, the neutron flux density at the surface of the
foil can be expressed as:
fS
/¼ ð3:51Þ
a
where
f is the fraction of the emitted neutrons reaching the foil, and
a is the area of the foil.
The factor f takes into account the scattering and absorption of the neutrons in
the absorber placed in between the source and the foil (the atmospheric air and the
paraffin block used as moderator) as well as the geometrical factor (solid angle
consideration). Substituting for ϕ in Eq. (3.50) gives:
NT ra fS
N ðt Þ ¼ 1 ek t ð3:52Þ
ka
and the activity of the radionuclides after an activation time of ta is
NT ra fS
Ao ¼ kN ¼ 1 ek ta : ð3:53Þ
a
If the activity of the foil is measured by a counting system of efficiency ε for a

period of ti immediately after activation, the total number of counts obtained is
given by
Z ti
C¼e Ao expðktÞdt
0 ð3:54Þ

ef NT ra S
¼ 1 ekta 1 ekti :
a k
This equation relates the total counts to the strength of the continuous neutron
source. It can be written in the form C ¼ KS if the activation time ta and the
counting time ti are fixed. K is then the calibration constant of the detector setup that
may be calculated from

efNT ra
K¼ 1 ekta 1 ekti ð3:55Þ
ak
provided all the constants can be known with sufficient accuracy. Alternatively,
K can be determined by calibrating the detector against a source with known
strength as
C
K¼ :
S
3.6.1.2 Pulsed Neutron Source
For a pulsed source, the equivalent source strength may be specified as total yield
(n) divided by the pulse duration (τ) assuming square pulse, and the activation time
will be equal to the pulse duration. Hence expression (3.54) will be re-written as

0 efNT ra n
C ¼ 1 ek s 1 ek ti : ð3:56Þ
a sk
Since

1 ek s
! k;
s s!0
we get

0efNT ra n
C ¼ 1 ek ti ð3:57Þ
a k
which relate the counts obtained for exposure of the detector to a pulsed source of
yield n. Comparing (3.57) with (3.54), we can obtain
S C0
n¼ 1 ek ta : ð3:58Þ
k C
Hence the foil activation neutron detector can be calibrated against a constant
source as standard and then used to measure the yield of a pulsed neutron source
provided the same counting time is used for both the calibration and the pulsed
neutron measurement. The factor
S 1
F¼ 1 ek ta
k C
is the calibration factor of the detection system obtained from the calibration against
a constant source of strength S with C as the counts obtained for a counting time of
ti.
An example of the material that may be used for the detection of pulsed neutron
source such as the plasma focus is indium. The activation and decay schemes of
indium are as shown in Fig. 3.43.
Natural indium consists of 95.8 % of 115In and 4.2 % of 113In. During activation,
both components will be activated. However, the activation cross section of 113In is
insignificantly small hence its contribution will be neglected. There are two possible
branches of activation of 115In by thermal neutrons: one has a cross section of about
160 barns and the product radionuclide is 116mIn. Another branch is that of 116In
which has a cross section of about 42 barns, producing 116In. Both these
radionuclides undergo decay to the stable nuclide 116Sn.
Activation schemes: 115 In þ 1 n !116m In þ c ðra ¼ 160 2 barns)

115
In þ n !
1 116
In þ c ðra ¼ 42 1 barns)
116m 116
1/2 = 54.12 min) Sn
2.76
116
1/2 = 14.1 s) 28%
2.50
β- 1.27
Fig. 3.43 Neutron activation and decay scheme of indium

The presence of two branches of decay implies that their contribution to the
radiation activity of the activated indium foil must be considered together.
However, if we choose the activation time and the counting time to be much shorter
than the half-life of the 116mIn branch (τ1/2 = 54.12 min), its contribution can be
made relatively insignificant compared to that of 116In. This will be the case when
we choose ta to be 70 s (*5 half-lives of 116In’s decay) and ti to be 40 s (*3
half-lives of 116In’s decay). Hence for the Indium Foil Neutron Activation Detector
(IFNAD), we may obtain

S
F ¼ 19:4
C
where S is the strength of the standard neutron source used in the calibration and
C is the counts obtained in 40 s immediately after activation for 70 s. If the counts
obtained for the pulsed source in 40 s is C′, then the neutron yield of that pulse is
n ¼ F C0 :
Note that C′ must be counts obtained for the same counting time as C, which is the
counts obtained against the standard calibration source.
3.6.2 The Detector Setup
One possible setup of the neutron foil activation detector is shown schematically in
Fig. 3.44. It consists of an activation foil (indium) attached directly to a plastic
scintillator coupled to a PMT. The PMT is usually mounted onto a base consisting
Fig. 3.44 Schematic setup of a neutron activation detector using photomultiplier tube as the
detector
Fig. 3.45 Schematic setup of neutron activation detector using Geiger Müller tube (GM tube) as
the detector
of a voltage divider circuit and an emitter follower pre-amp (1:1) for impedance
matching. The output pulses of the PMT are counted by a counter through a
integrated discriminator.
Alternatively, a Geiger Müller tube (GM tube) can be used as shown in Fig. 3.45
to measure the β− particles directly. In this case, the indium foil is wrapped around
the GM tube [5].
In both setups, the detector is enclosed inside a hydrogenous paraffin block. The
function of the paraffin is to thermalize the fusion neutrons which are fast neutrons
with energy of 2.45 MeV.
3.6.3 Time-Resolved Neutron Pulse Measurement
This is normally done by using a scintillator-photomultiplier system similar to that

shown in Fig. 3.46, but without the counting circuit. In this case the output of the
photomultiplier is connected to the input of an oscilloscope.
For the measurement of neutrons from high power system, it is necessary to
shield the photomultiplier tube and its supporting circuit against electrical noises.
This gives rise to some difficulties in the implementation of the
scintillation-photomultiplier system. Recently, a system using optical fibre to
couple the scintillator to the photomultiplier has been proposed.
This is shown in Fig. 3.46 [5]. In this system, a cylindrical scintillator (NE102) is
used for the detection of neutrons. The light produced from the scintillation process
is picked up by the fluorescence fibre and its light output is coupled to a glass fibre
optic to transmit the light to the photomultiplier tube which is housed inside a
screened room together with the oscilloscope. In this way, electrical interference
can be minimized.
Fig. 3.46 Time resolved neutron detector setup
Fig. 3.47 Experimental

arrangement for neutron
time-of-flight measurement
Beside providing information concerning the time evolution of the neutron

production, the scintillator-photomultiplier system can also be used to determine the
energy of the neutrons. This is done by utilising the time-of-flight technique. The
experimental setup is as shown in Fig. 3.47. Two or more channels of
scintillator-photomultiplier system are used. With a minimum of 2 channels, placed
at different distances from the plasma neutron source, the time taken for neutrons to
travel a known distance can be measured from their signals as shown in Fig. 3.48.
The speed of the neutrons and hence their energy can then be calculated.
Fig. 3.48 The two well separated neutron pulses obtained for the time-of-flight measurement
References
1. Wong CS (1985) Simple nanosecond capacitive voltage divider. Rev Sci Instruments
56:767–769
2. San Wong C, Choi P, Leong WS, Singh J (2002) Generation of high energy ion beams from a
plasma focus modifie for low pressure operation. Jpn J Appl Phys 41:3943–3946
3. Moo SP, Wong CS (1995) Time resolved hard X-ray emission from a small plasma focus. Laser
Part Beams 13:129–134
4. Wong CS, Moo SP, Singh J, Choi P, Dumitrescu-Zoita C, Silawatshananai C (1996) Dynamics
of X-ray emission from a small plasma focus. Mal J Sci 17B:109–117
5. Yap SL (1998) A study of the temporal and spatial evolution of neutron emission from a plasma
focus. MSc Thesis, University of Malaya
General Reference
6. Huddlestone Richard H, Leonard Stanley L (eds) (1965) Plasma diagnostic techniques.

Academic Press Inc., New York
Chapter 4
Some Examples of Small Plasma Devices
Abstract In this chapter, the principle of operation, experimental setup as well as

some results obtained from research work carried out by the authors and collabo-
rators based on several small plasma devices will be discussed briefly.
Keywords Cost effective plasma devices
4.1 The Electromagnetic Shock Tube [1]
The schematic diagram of an electromagnetic shock tube is shown in Fig. 4.1.

The electrodes are of a coaxial configuration. The electrodes are insulated at the
back-wall by a quartz cylinder. Discharge is initiated along the surface of the
insulator as a surface discharge, forming a current sheet. This current sheet will be
pushed away from the insulator surface by the J ∧ B force until it finally becomes
perpendicular to the electrodes and ready to move in the axial direction.
The initial phase of current initiation along the surface of the quartz insulator and
lift-off is called the breakdown phase. During the breakdown phase, the gas is
mainly heated by the joule heating effect and this is the pre-ionization phase where
a sufficiently ionized plasma will be formed prior to the subsequent axial acceler-
ation phase. This is essential since the effective shock heating of the plasma requires
it to be sufficiently ionized.
During the axial acceleration phase, the current sheet is driven by the electro-
magnetic force which is acting as the axial electromagnetic piston. The azimuthal
magnetic field Bθ is generated by the coaxial discharge current itself and is given by
lI
Bh ¼ ;
2pr
which is a function of time and radial position r. This magnetic field is only present
behind the current sheet. The magnetic field ahead of the current sheet is zero. It can
be seen that its magnitude will be stronger at the surface of the inner electrode as
compared to the inner surface of the outer electrode. This gives rise to the slanting

DOI 10.1007/978-981-10-0117-8_4
100 4 Some Examples of Small Plasma Devices
Fig. 4.1 Schematic diagram of an electromagnetic shock tube
structure of the electromagnetic piston as shown in Fig. 4.1. The J ∧ B force in the
direction downstream of the tube (in the z-direction) can be expressed as
Zb
B2h
2prdr;
2l
a
where a is the radius of the inner electrode and b is the radius of the outer electrode.
This force will drive the electromagnetic piston to supersonic speed so that a shock
heated layer of plasma will be formed. In this way, with discharge current in the
region of 100 kA, piston speed of up to more than 10 × 104 m/s may be achieved,
which is sufficient to produce a fully ionized hydrogen plasma. There is another
component of the J ∧ B force which is acting in the radial direction but no motion is
possible due to the present of the solid inner electrode.
4.1.1 Numerical Modeling of the Electromagnetic Shock

Tube Dynamics
The dynamics of the electromagnetic piston can be modeled by simplifying the

geometry of the shock tube as shown in Fig. 4.2. We assume that (i) the current
sheet starts from zero velocity at the position when it first becomes vertical to the
electrodes, hence the effective length of the electrodes (zo) should be measured from
the edge of the cylindrical quartz insulator; (ii) the current sheet is not slanting but it
is vertical to the electrodes from the surface of the inner electrode to the outer
electrode.
The equation of motion for the piston can be written as:
Zb
d dz B2
½q pðb2 a2 Þ z ð Þ ¼ 2prdr:
dt 1 dt 2l
a
4.1 The Electromagnetic Shock Tube 101
Fig. 4.2 Dynamic model of the electromagnetic shock tube
Substituting the expression for B and re-arranging, we get for the axial accel-
eration of the current sheet:
lo ‘nðb=aÞ 2 dz
½ I ð Þ2
d 2 z 4p2 q1 b2 a2 dt
¼ :
dt 2 z

This differential equation can be solved for the axial current sheet velocity dz dt if
the form of the current as a function of time I(t) is known. Alternatively, the effect of
the dynamics of the current sheet on the discharge current can be taken into con-
sideration by writing the coupled circuit equation and solve it simultaneously with
the equation of motion. The circuit equation is written by referring to the equivalent
circuit of the capacitor discharge system. Assuming the inductive model, where
d
L o þ Lp I I R o þ R p :
dt
Then the circuit equation is reduced to

Rt
d Idt
Lo þ Lp I ¼ Vo 0 ;
dt C
which can be re-arranged to give the rate of change of current as

R
Idt l b dz
Vo C o ‘n I
dI 2p a dt
¼ :
dt l b
Lo þ o ‘n z
2p a
To solve this equation simultaneously with the dynamic equation, it is conve-

nient to first normalize them. This is done by choosing the following normalized
parameters:
z t I
n¼ ; s¼ ; i¼ ;
zo to Io
where
zo effective length of the electrodes.
qffiffiffiffi
Io ¼ Vo C “short circuit” current.
Lo
to ¼ L o C discharge characteristic time.
The normalized form of the equations are

2
d 2 n a2 i2 dn
ds
¼ ;
ds2 n
R
di 1 ids ib dn ds
¼ :
ds 1 þ bn
α and β are scaling parameters which are given by a ¼ ttoa and b ¼ LLao where
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
4p2 ðb2 a2 Þq1 z2o
ta ¼ ;
lo ‘n ba Io2
which is the characteristic dynamic time.

β is the ratio of the maximum tube inductance La ¼ l2po zo
‘n ba and the circuit
inductance.
The solution can be obtained by using the following boundary conditions:
Z
di d 2 n s!0 a
s ¼ 0; n ¼ 0; i ¼ 0; ¼ 1; ids ¼ 0; ! pffiffiffi :
ds ds2 2
The computation will stop when ξ = 1 (z = zo).

An example of the solution obtained is as shown in Fig. 4.3.
These solutions are expected to be not accurate due to the fact that in the actual
plasma discharge, the discharge current may be split to form multiple current sheets
(current shedding effect) and not all the gas particles are swept by the current sheet
(mass shedding effect). In one study [2], it had been shown that the mass swept up
by the current sheet to form the plasma can be as little as less than 30 % of all the
mass available. These two effects are actually affecting the current sheet dynamics
in an opposite manner. While the current shedding effect tends to make the
EMST computation
α = 1.2, β = 0.25
3
ξ
2.5
2
ι, ξ, dξ/dτ
1.5
1
dξ/dτ
0.5
ι
0
0 1 2 3 4
τ
Fig. 4.3 Solutions of the EMST dynamic model
experimentally measured dynamics slower than that predicted by the model, the
mass shedding effect tends to make the experimentally measured dynamics faster.
4.1.2 Experimental Measurements of Electromagnetic

Shock Tube Dynamics
Since the conditions of the plasma produced by the electromagnetic shock tube are
closely related to the dynamics of its current sheet, it is essential in the study of the
device to measure the shock speed or the piston speed produced during the dis-
charge. This measurement is commonly done by using the magnetic pickup coil.
The magnetic pickup coil is made of a few turns of copper wire with mm
diameter and it “picks up” the localized magnetic flux threading it. Hence it can be
used to register the arrival time of the current sheet of an electromagnetic shock
tube. This is illustrated in Fig. 4.4.
Current sheet (piston)

Outer Electrode
Glass insulator
Magnetic Coil
b
Current I(t) a
Inner Electrode
z
zo
Fig. 4.4 Experimental arrangement for measurement of current sheet dynamics using magnetic
probe
Fig. 4.5 Experimental magnetic probe signals at three positions (waveforms 2, 3 and 4) of the
electromagnetic shock tube together with the discharge current (waveform 1)
For the ideal case where the current sheet is assumed to be very thin, the signal
registered by the magnetic pickup coil which is corresponding to B(t) is expected to
have a sharp rising edge which indicates the arrival of the current sheet. However,
the actual current sheet may be diffused and hence the B(t) signal registered may
look like what is shown in Fig. 4.5 instead.
The signals displayed in Fig. 4.5 are (1) the total discharge current measured by
a Rogowski coil; (2) magnetic probe signal at a distance of 2 cm from the back wall
of the shock tube; (3) magnetic probe signal at a distance of 6 cm from the back
wall of the shock tube; and (4) magnetic probe signal at a distance of 10 cm from
the back wall of the shock tube. The times of arrival of the current sheet at various
axial positions are easily deduced from these signals.
An alternative method is to measure the rate of change of the magnetic field dB dt
instead of the magnetic field itself. Experimentally, this means removing the RC
integrator and replaced it by a 50 Ω terminator (or a ×10 resistive voltage divider
with a 50 Ω input resistor) instead. The dB dt signal has significantly larger
Fig. 4.6 The rate of change

of the magnetic probe signal
amplitude as compared to the B signal so it is technically easier to measure. It

should be pointed out that in this case the arrival time of the current sheet is taken to
be corresponding
to the peak of magnetic probe signal, which is the point marked as
“X” of the dB
dt signal as shown in Fig. 4.6.
4.2 The Plasma Focus
4.2.1 Introduction
The plasma focus was proposed in the 1960s as a possible device to achieve
thermonuclear fusion [2]. In 1962, a research team at the Kurchatov Institute in
USSR (Russia) led by Filippov proposed a modification to the linear Z-pinch by
providing a inversed pinch phase as a pre-pinch phase. This pre-pinch phase pro-
vides some preliminary heating of the plasma before the radial pinching action
starts. A similar concept was proposed by Mather and his team in 1964 at the Los
Alamos National Laboratory in USA. The Mather design of the plasma focus was
originated from the electromagnetic shock tube and in addition to the inverse pinch
phase, a second phase of axial acceleration was added as pre-pinch phase. An
obvious difference in the two designs is in their geometrical appearances. The
Filippov design has a short and large diameter inner electrode while the Mather
design has a long and small diameter inner electrode.
Since their respective first reports, the plasma focus device has been studied in
numerous laboratories around the world, initially mostly with emphasis on ther-
monuclear fusion. Currently, the plasma focus is being studied as a possible por-
table pulsed neutron, radiation (EUV to X-ray) and charged particle (electron and
ion) beams source.
4.2.2 Characteristics of the Plasma Focus Discharge [3]
The plasma focus discharge can be described as consisting of three main phases:
(1) the inverse pinch (or lift-off) phase, (2) the axial acceleration phase and (3) the
radial compression (pinch) phase as indicated in Fig. 4.9b. In the Filippov type
plasma focus, the axial phase is practically nonexistence; while in the Mather type,
this phase plays a major role in the focus dynamics.
The plasma focus discharge system is basically the same as the electromagnetic
shock tube described in Sect. 4.1. The initial phases of the discharge are similar to
the electromagnetic shock tube, consisting of the lift-off phase and the axial
acceleration phase. However, when the current sheath reaches the truncated end of
the electrodes, the current sheath will continue to be pushed axially out of the
electrodes thus providing now an additional dimension to be acted upon by the
Fig. 4.7 The signature

voltage spike and current dip
of a focused discharge
electromagnetic force that is radially inwards. This gives rise to what is equivalent
to an axially elongating radially compressing Z-pinch. This is the radial compres-
sion phase of the plasma focus dynamics. The radial compression occurs very
rapidly in a nanoseconds time-scale and the plasma formed at the end of this phase
has electron temperature of several keV and electron density of up to 1019 cm−3.
With such a high temperature and density, the focused plasma is a rich source of
radiation emission including X-ray, electron and ion beam and when deuterium is
used as the working gas, fusion neutrons will be produced.
The most fundamental parameters of the plasma focus discharge measured
experimentally are the discharge current and the voltage across the electrodes. In
particular the discharge current is measured by using a Rogowski coil mounted at
the back of the focus tube around the inner electrode, while the voltage drop across
the electrodes is measured by using a simple resistive voltage divider. An example
of the current and voltage signals for a typical focus discharge is shown in Fig. 4.7.
The distinct voltage spike and the current dip are the signatures of the occurrence of
focusing action during which the hot dense plasma is produced.
4.2.3 Design Consideration
The first step in the design of the plasma focus system may be considered based on
the dynamic model for the axial acceleration phase alone. The main design criterion
is to match the rise-time of the discharge current (tr) to the time of arrival of the
current sheet at the end of the inner electrode (tz) which is actually the time at the
end of the axial acceleration phase. tr is taken to be the quarter period time of
the discharge which is given by
4.2 The Plasma Focus 107
2p Lo C
tr ¼ ;
4
while tz is given by
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
4p2 ðb2 a2 Þq1 z2o
tz 2ta ¼ 2 :
lo ‘n ba Io2
For this two characteristic time to match, we have
64 ðb2 a2 Þz2o q1
¼ 1:
lo ‘n ba C 2 Vo2
Another condition that should be fulfilled is the possible speed that may be
achieved. We may decide that the characteristic speed of the current sheet that is
necessary to achieve reasonable heating of the plasma is *10 cm μs−1. By setting
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

zo lo ‘n ba Io2
Uc ¼ ¼ ¼ 105 ðms1 Þ;
ta 4p2 ðb2 a2 Þq1
the matching condition becomes

p pffiffiffiffiffiffiffiffi
zo ¼ 105 Lo C :
4
We may start with a given capacitor and an estimate of system inductance

including that of the capacitor itself, then the length of the electrode is fixed by the
expression above. For example, if a capacitor of capacitance 30 μF is available. Lets
say we estimate the inductance of the system can be kept at not more than 110 nH,
then an electrode length of 0.14 m may be appropriate.
With C and zo fixed, ρ1, Vo, b and a can then be matched according to
64 ðb2 a2 Þz2o q1
¼ 1:
lo ‘n ba C 2 Vo2
A chart of these parameters that may match this condition can be drawn up with
the constraint that a\b zo . It is probably logical to fix Vo first according to the
capacitor available, then the operating gas pressure (ρ1). This allows us to seek for a
suitable combination of a and b. This procedure can be repeated for other possible
values of Vo and ρ1.
Finally, experimental fine tuning may still be necessary to arrive at a set of
parameters for optimum plasma focus operation.
4.2.4 X-ray Emission from the Plasma Focus Discharge
During the plasma focus discharge, X-rays are emitted due to two possible
mechanisms [4–6]. The first is the hot dense plasma produced at the end of the
radial compression phase. With conditions of electron temperature Te * keV and
density of ne * 1019 cm−3, the plasma is expected to emit photons with energy in
the X-ray region. The peak of the continuum (Bremsstrahlung and recombination)
is expected to be several Å while the line radiations are due to inner shell transition
(Kα, Kβ of highly ionized species) which will be in the X-ray region. The second
mechanism is due to the interaction of high energy electron beam with the inner
electrode surface. Due to the development of MHD instabilities soon after the end
of the radial compression phase, high localized electric field may be induced and
high energy ion beam and electron beam may be produced. The electron beam will
bombard at the surface of the inner electrode giving rise to the generation of hard
X-ray, which is predominantly the Kα line radiation of the electrode material
(in most cases copper). The line radiation may also be due to contamination of
the focused plasma contributed by the electron beam sputtered copper vapor from
the anode [7].
In any typical plasma focus discharge, both these two types of X-ray emission
may be observed as shown in Fig. 4.8. In this discharge, X-ray pulse is observed
both coincide with the voltage spike as well as at time beyond 50 ns after the
voltage spike.
The X-ray emitting region of the plasma focus is observed to have a complicated
structure. Normally the tip of the inner electrode can be seen in the X-ray image
obtained experimentally. The X-ray image of the plasma column is observed to be
above the inner electrode, often with spot-like structures superimposed on it [8, 9].
An example of the X-ray images of the plasma focus is shown in Fig. 4.9.
Voltage spike
Plasma
e-beam target?
Fig. 4.8 Evidence of two types of X-ray emission from a focused plasma
4.2 The Plasma Focus 109
Fig. 4.9 Images of spot-like plasma within the focused plasma
This set of X-ray images are obtained for the same plasma focus discharge but
with different absorption filters: (a) 48 μm aluminized Mylar; (b) 48 μm aluminized
Mylar + 10 μm Al; and (c) 48 μm aluminized Mylar + 20 μm Al. These filters allow
regions of different X-ray hardness to be revealed. It is evident that within the
plasma column in (a), minute spots emitting harder X-ray are present as shown in
(b) and (c). The shape of these X-ray spots is round indicating that their size is
probably smaller than the pinhole used in the pinhole camera setup.
A possible application of the plasma focus X-ray is in X-ray lithography. The
feasibility of such application has been tested in various laboratories. High Z gases
such as Xenon and Krypton are often used for the operation of the plasma focus as a
pulsed X-ray source [10, 11].
4.2.5 Neutron Emission
When deuterium is used as the operating gas, the plasma focus has been shown to
be capable of achieving fusion condition at the end of the radial compression phase.
This gives rise to the generation of neutrons. However, there are evidences that the
neutrons emitted from the focused plasma is not totally of thermonuclear origin.
This is in fact expected since the final temperature achieved in the plasma focus is
only of the order of few keV. Any fusion reaction must be due to deuterons with
kinetic energy at the high energy tail of the energy spectrum of the plasma. A large
percentage of the neutrons measured, probably more than 60 % in the case of a
small plasma focus, are actually produced by the beam target effect. High energy
ion beam, in this case deuteron beam, is generated due to localized electric field
induced by instabilities.
One of the evidences of non-thermonuclear neutron emission is that the energy
of the neutron is 2.8 MeV instead of 2.45 MeV as measured by the neutron
time-of-flight technique as described in Sect. 3.6.3 earlier.
Secondly, the neutron emission angular distribution is found to be anisotropic,
with the ratio of neutron yield measured in the end-on direction (0°) to that in the
side-on direction (90°) to be greater than 1. For example, in a series of measure-
ments, the neutron yields measured end-on and side-on simultaneously give a ratio
of 1.5 [12].
Fig. 4.10 Evidence of two

phases of neutron emission
from the focused plasma
Thirdly, from the analysis of the time-resolved neutron pulse obtained experi-
mentally, it can be shown by using the Gaussian curve fitting technique that
the neutron pulse can be resolved into two Gaussian pulses (Fig. 4.10) indicating
the possibility of 2 neutron components. The first pulse which is more aligned to the
voltage spike is believed to be thermonuclear in nature while the second pulse is
interpreted as produced by the deuteron beam target mechanism [13].
4.2.6 Particle Beam Emission
The existence of energetic electron beam in the plasma focus is believed to be

closely co-related to phenomena that give rise to localized high electric field, such
as the rapid compression during the radial pinch phase and instability. Both
mechanisms are found to be present and indeed two periods of electron beam
generation have been observed. The electron beam observed during the first period
is believed to be associated with the first maximum compression and is measured to
be having energy as high as 180 keV. The second period electron beam, which is
tied to the m = 0 instability, has a much lower energy [14, 15].
Together with the generation of electron beam, the same localized induced
electric field is also responsible for the generation of energetic ion beam [16, 17].
Ion beams with energies of up to MeV have been reported in the literature. The
average nitrogen ion beam measured is about 180 keV, which seems to be con-
sistent with the averaged energy of the electron beam measured. The applications of
the plasma focus generated ion beam to modify the nano-scaled structure of various
material surfaces have been widely reported recently [18–20].
4.3 The Vacuum Spark 111
4.3 The Vacuum Spark
The vacuum spark is known to be a source of intense pulsed X-ray as well as a

source of highly-stripped ions such as Fe-XXVI, Ni-XXVII and Mo-XIII-XVIII. In
view of the close resemblance of its X-ray spectrum with that of solar flares, the
vacuum spark is also being used to simulate the solar flare phenomenon in the
laboratory.
The vacuum spark has a relatively simple configuration as compared to the
plasma focus. It consists of a pair of electrodes kept at a distance of less than 1 cm
apart and placed in vacuum (typically P < 10−3 mbar is sufficient). The electrodes
are connected directly across a capacitor so no switch is used. When the capacitor is
charged to a high voltage V (typically from 10 to 30 kV), breakdown across the
inter-electrode gap is prohibited by the high vacuum since the value of the product
of pressure and inter-electrode spacing, pd, is on the left hand side of the Paschen
curve minimum. Discharge is initiated by injecting a puff of vapor of the anode
material (partially ionized) by one of the following methods: (1) by using a sliding
spark between the cathode and a third electrode to produce some electrons which
will be accelerated by the electric field to bombard at the anode to produce the
anode vapor [21]; (2) by focusing a high power pulsed laser beam to vaporize the
anode material directly [22]; and (3) by utilizing the Transient Hollow Cathode
electron beam to vaporize the anode material [23, 24].
The vacuum spark plasma has been measured to achieve temperature of as high
as 8 keV and density of 1020–1021 cm−3. The structure of the plasma formed may
consist of plasma cloud and minute spots with a source size of 100–200 μm, an
example of which is shown in Fig. 4.11. In this image, the tip of the anode is also
seen to be emitting X-ray, probably due to bombardment by electrons.
The X-ray emission spectrum of the vacuum spark plasma is found to be con-
sisting of a mixture of continuum and line radiations. Its spectral contents can be
classified into 3 categories, namely plasma emission dominated (continuum radi-
ation dominated), electron beam—target emission dominated (line radiation dom-
inated) or mixture of both types of emission. For the plasma emission dominated
discharges, the electron temperatures are estimated to be in the range of 3–10 keV.
For discharges where the X-ray spectrum is dominated by line radiation, the plasma
emission is relatively weak and the X-ray photons are predominantly produced by
(a) (c)
plasma
(b) anode
Time-base: 100 ns/div
Fig. 4.11 a dI/dt signal, b X-ray pulse, c X-ray pinhole image of a typical vacuum spark plasma
electron beam bombardment of the anode which gives rise to strong Kα emission of
the anode material. The intense energetic electron beam may be produced by either
the pre-breakdown transient hollow cathode discharge which is utilized to initiate
the main discharge, or the strong m = 0 instability at the final stage of the pinching
effect of the plasma.
4.4 Scaled-Down Operation of Vacuum Spark—Flash

X-ray Tube
In view of the observation of electron beam target line radiation emission during the
pre-breakdown phase of the vacuum spark induced by the transient hollow cathode
effect, the vacuum spark system can be scaled down in terms of electrical input
energy to operate as a flash X-ray tube [25–27]. The schematic diagram of the flash
X-ray tube is illustrated in Fig. 4.12.
The operation of the flash X-ray tube is exactly the same as the vacuum spark
but the main discharge current is low so the plasma produced will not be hot enough
to produce X-ray. As can be seen from Fig. 4.12, the discharge is powered by eight
ceramic capacitors with capacitance of 2.7 nF each, making up a total capacitance
of 21.6 nF as compared to the 1.85 μF capacitor used in the high energy version.
When discharged at a voltage of 20 kV, the flash X-ray tube input electrical energy
is only 4.32 J as compared to 370 J in the case of the vacuum spark. However, with
Fig. 4.12 Schematics of a flash X-ray tube. Reprinted from Ref. [26], Copyright (2007), with
permission from Cambridge University Press
4.4 Scaled-Down Operation of Vacuum Spark—Flash X-ray Tube 113
0.5
0.4
0.3
dI/dt (A.U.)
0.2
0.1
0.0
-0.1
-0.2
-0.3
-0.4
-7 -7 -6 -6 -6
-5.0x10 0.0 5.0x10 1.0x10 1.5x10 2.0x10
2
X-ray pulse (V)
-5.0x10 -7 0.0 5.0x10

-7
1.0x10 -6 1.5x10 -6 2.0x10 -6
Time (s)
Fig. 4.13 Example of the discharge current waveform and the X-ray pulse of a flash X-ray tube
discharge. Reprinted from Ref. [26], Copyright (2007), with permission from Cambridge
University Press
the same capacitor discharge voltage applied across the electrodes, the intensity of
the electron beam target X-ray emission due to bombardment of the transient
hollow cathode electron beam is expected to be the same in the two cases. An
example of the X-ray pulse produced by the flash X-ray tube is as shown in
Fig. 4.13.
Some of the applications of the flash X-ray tube as X-ray source for radiography
of small biological sample [26] as well as for testing of thermo-luminescence
response of fiber dosimeter has been reported recently [28, 29].
The X-ray emission spectrum of the flash X-ray tube can be easily tuned by
changing the anode material. Figure 4.14 shows the X-ray spectra of (from top to
bottom) Al, Ti, Cu and W [26].
4.5 The Pulsed Capillary Discharge
The pulsed capillary discharge has been shown to be a copious source of EUV and
soft X-ray. Recently, due to the successful demonstration of amplification of the 3 s
1
P1–3p 1S0 line of Ne-like Ar at λ = 46.9 nm in a capillary discharge, this device has
attracted much research interests from various researchers.
The pulsed capillary discharge is basically a Z-pinch discharge restricted to
occur within the small diameter capillary channel. An example of a capillary dis-
charge powered by capacitor discharge [30] is shown in Fig. 4.15a. The discharge
of this system is powered by six 30 kV, 3.6 nF doorknob capacitors connected in
Fig. 4.14 Examples of X-ray emission spectra of the flash X-ray tube. Reprinted from Ref. [26],
Copyright (2007), with permission from Cambridge University Press
parallel to give a maximum discharge input energy of 9.7 J. The capillary is made
of quartz tube with inner diameter of 1 mm and a length of 10 mm. The system is
pumped to a pressure of 10−5 mbar. To perform a discharge, the capacitor is first
charged up to the desired high voltage. A high voltage pulse is then applied to the
triggering pin to initiate the transient hollow cathode discharge in the hollow
cathode region which subsequently leads to the main discharge through the
capillary.
The discharge current waveform together with radiation emission pulses from a
typical 24 kV discharge of this capillary discharge system are shown in Fig. 4.15b.
The discharge current is measured by the magnetic pick up coil (the magnetic
probe), while the X-ray pulse is measured by a PIN diode and the EUV pulse is
measured by a silicon photodiode sensitive to the EUV range. The peak current
reached is about 8 kA.
From Fig. 4.15b, it can be seen that the discharge can be divided into two major
phases: the initiation phase and the main discharge phase. During the initiation
phase, before the main discharge set in, the discharge current is started with a sharp
pulse following by a slow rising edge. Initially the 24 kV high voltage is divided
between the capillary and the external spark gap. Before the occurrence of the full
discharge, due to the voltage held across the cathode and the anode, electrons from
the triggering pulse in the hollow cathode region may be attracted to flow across the
capillary to hit at the anode. This produces the low level X-ray emission observed in
the X-ray pulse. The small sharp current pulse corresponds to the external spark gap
firing which transfers the full voltage across the capillary. This further enhance the
4.5 The Pulsed Capillary Discharge 115
Fig. 4.15 a Schematics of the pulsed capillary discharge device. b The discharge current,
pre-breakdown electron beam induced X-ray pulse and EUV pulse emitted from the plasma.
Figure 4.15a is reprinted from Ref. [30], Copyright (2014), with permission from Elsevier
electron beam activity and eventually leads to avalanche of electrons forming a

beam which bombards at the anode to give a large X-ray signal. Breakdown of the
main discharge gap across the capillary will then occur. The formation of the
capillary discharge plasma is accompanied by the emission of 30 mJ of radiation
energy in the wavelength range of 11–18 nm of the EUV spectrum. The electron
temperature of this pulsed capillary discharge plasma is estimated to be about
10 eV.
4.6 The 50 Hz Alternating Current (A.C.) Glow Discharge

System
There are two major cost factors in any plasma system. First is the requirement of
low or ultra low operating pressure. The creation of high vacuum means large
budget is needed to purchase high vacuum equipments, which consist of the
pumping system as well as the pressure gauges, both of which are expensive. It is
for this reason that the development of atmospheric pressure discharges has
attracted much research efforts in recent years. The second factor is the power
source. Comparing with the RF, microwave or even the DC power sources, the
relatively cheap power source is the 50 Hz AC power source which can be obtained
directly from the domestic power source via a step-up transformer. Another major
advantage of using the 50 Hz AC power source is that no impedance matching
network is needed, which is an important cost reducing factor. For the measurement
of the discharge current, a simple milli-ammeter or a multi-meter can be used. Since
Fig. 4.16 Schematics of the 50 Hz AC glow discharge system
the discharge voltage required is normally in the kV region, a voltmeter coupled

with a resistive voltage divider may be used.
An example of the 50 Hz glow discharge system is shown schematically in
Fig. 4.16. The system consists of the following components: The discharge
chamber, a rotary pump, a step-up transformer and the gas flow control. The
electrodes are two circular stainless steel plates. The operation of the system is
straight forward. After breakdown, the discharge current is controlled by using the
current limiting resistor R to be within the range of glow discharge, normally in the
mA region.
The electron temperature and density of the plasma produced can be measured
by using the Langmuir probe as indicated in Fig. 4.16. In a series of experiments,
argon glow discharge plasma was produced by using a discharge voltage of 16 kV
peak to peak, while the working pressure was varied from 0.3 to 0.6 mbar. Various
sets of I–V characteristics were obtained with the Langmuir probe placed at the
center of the plasma column. The experiment was done with an inter-electrode
distance of 2 cm. An example of the I–V characteristics obtained is displayed in
Fig. 4.17.
The variations of electron density and electron temperature with pressure for
inter-electrode distance 2 cm are shown in Fig. 4.18.
It is seen that there is an increase of electron density and decrease of electron
temperature with increase in pressure from 0.3 to 0.6 mbar. The maximum electron
density and electron temperature are found to be 3.58 × 1016 m−3 and 5.98 eV,
4.6 The 50 Hz Alternating Current (A.C.) Glow Discharge System 117
Fig. 4.17 An example of typical I–V characteristics of Langmuir probe
7 3.8
Ar Plasma, Inter electrode distance=2 cm
Vdis=16 kV (peak to peak)
Electron Temperature
Plasma Density
3.6
6 3.4
Plasma Density (x10 ) m
Electron Temperature (eV)
3.2
5
3.0
2.8
16
4
2.6
-3
2.4
3
0.30 0.35 0.40 0.45 0.50 0.55 0.60
Pressure (mbar)
Fig. 4.18 Variation of electron temperature and density with pressure
respectively. The increase of electron density with pressure is believed to be due to

more ionization. Since increase in pressure increases the neutral particle density, so
the probability of energy loss of the plasma electrons due to the collision with
neutral particle increases. So the energy of the electrons decreases i.e., electron
temperature decreases with increase in pressure. It is noticed that plasma density
increases in case of 2 cm inter-electrode distance compared to that of 1.5 cm for a
particular pressure. On the other hand, electron temperature decreases for a par-
ticular pressure in case of 2 cm inter-electrode distance compared to that of 1.5 cm.
It is believed that when the inter-electrode distance increases, the electrons undergo
more collisions resulting in an increase of plasma density. Because of more colli-
sions, electrons loss energy and that is why electron temperature decreases with an
increase of inter-electrode distance.
The 50 Hz glow discharge plasma has been used successfully to treat
bio-materials such as gelatin and Thai silk fibroin to improve the wettability of their
surfaces [31–35].
4.7 Atmospheric Pressure Dielectric Barrier Discharge
The requirement of a low pressure ambient gas environment for the generation of
plasma is one of the important cost factor in the application of plasma technology in
industry. In view of this consideration, there is much research interest to investigate
the feasibility of using atmospheric pressure plasmas for various processes instead
of low pressure plasmas. One type of device that fit well into this criterion is the
dielectric barrier discharge which is being studied in many small laboratories
worldwide.
The dielectric barrier discharge, or DBD in short, can be constructed in two
types of configuration, either with parallel plate electrodes or coaxial electrodes as
illustrated in Fig. 4.19.
A common feature of these two setups is the present of a dielectric between the
electrodes so that when an oscillating electric potential is applied across the elec-
trodes, intermittent electric current pulses may occur between the surface of the
dielectric barrier and one of the electrodes. These current pulses are normally
occurring in burst at both the positive and negative half cycles of the alternating
voltage as shown in Fig. 4.20 [36, 37]. Each of the current pulse may have a pulse
width of less than 200 ns as shown in Fig. 4.21 [38]. Similar discharge charac-
teristics are also observed with the coaxial DBD.
The cylindrical coaxial DBD has been found to be particularly suitable for
applications as chemical reactor for gases. One example of using the coaxial DBD
as a chemical reactor and it has been tested to be effective for converting oxygen to
ozone [39] and also for dissociating nitric oxide [40, 41].
Fig. 4.19 Two possible configurations of dielectric barrier discharge

4.7 Atmospheric Pressure Dielectric Barrier Discharge 119
Fig. 4.20 Two bursts of current pulses and the corresponding sinusoidal voltage waveform of a
50 Hz dielectric barrier discharge
Fig. 4.21 Single current pulse with *200 ns pulse width
For the application of planar DBD such as treatment of material surface, a more
uniform planar glow discharge type plasma is required. For such an application it
may be necessary to use a voltage source of kHz frequency [42].
4.8 Wire Explosion System for Nano Powder Fabrication
The wire explosion process is a simple technique that can be used effectively to
synthesize nano particles of pure metals or their composites. The pulsed discharge
system is similar to those used for the vacuum spark as shown schematically in
Fig. 4.22. In this particular system, the discharge is powered by a high voltage
capacitor with rating of capacitance 1.85 μF and discharge voltage of up to 50 kV.
Fig. 4.22 Schematics of a wire explosion system
However, for the purpose of the wire explosion process, a nominal discharge
voltage of 15 kV is sufficient. The chamber has a diameter of 9.4 cm and a height of
10.3 cm. Magnetic pick-up coil and high-voltage probe are used to register the
current and voltage signals respectively during the wire explosion, while a PIN
diode (BPX65) can be used to record the light (visible to ultraviolet) emitted during
the wire explosion and the subsequent electrical discharge through the vapor. The
energy deposited into the wire can be calculated from the measured current and
voltage signals. The measured current density through the wire during the explosion
of the wire is found to be of the order of 106 A cm−2. This conduction continues
until the vapors are optically thick and only radiation emitted from the surface will
be detected by the PIN diode. At this stage, a column of ionized plasma is formed.
Subsequently, the plasma begins to expand in the background of the ambient gas
due to the enormous difference in the temperature and pressure between the plasma
and the ambient gas. The expanded plasma particles are rapidly cooled down during
this process of expansion and leading to the formation of a supersaturated vapor,
which will undergo a homogeneous nucleation of nano particles. The ambient gas
type as well as its pressure have been found to have great influence on the nano
particle formation [43–47].
References 121
References
1. Lee S (1985) Experiment 1: electromagnetic shock tube. In: Lee S, Tan BC, Wong CS,
Chew AC (eds) Laser and plasma technology. World Scientific Publishing Co. Pvt. Ltd,
Singapore
2. Bernard A, Bruzzone H, Choi P, Chuaqui H, Gribkov V, Herrera J, Hirano K, Krejci A, Lee S,
Luo C, Mezzetti F, Sadowski M, Schmidt H, Ware K, Wong CS, Zoita V (1998) Scientific
status of plasma focus research. J. Moscow Phys. Soc. 8:93–170
3. Lee S (1985) Experiment II: plasma focus experiment & technology of plasma focus. In:
Lee S, Tan BC, Wong CS, Chew AC (eds) Laser and plasma technology. World Scientific
Publishing Co. Pvt. Ltd, Singapore
4. Choi P, Wong CS, Herold H (1989) Studies of the spatial and temporal evolution of a dence
plasma focus in the X-ray Region. Laser Part Beams 7:763–772
5. Favre M, Lee S, Moo SP, Wong CS (1992) X-ray emission in a small plasma focus operating
with H2-Ar mixtures. Plasma Sources Sci Technol 1:122–125
6. Wong CS, Moo SP, Singh J, Choi P, Dumitrescu-Zoita C, Silawatshananai C (1996)
Dynamics of X-ray emission from a small plasma focus. Mal J Sci 17B:109–117
7. Al-Hawat Sh, Akel M, Wong CS (2011) X-ray emission from argon plasma focus
contaminated with copper impurities in AECS PF-2 using five channel diode spectrometer.
J Fusion Energy 30:503–508
8. Ng CM, Moo SP, Wong CS (1998) Variation of soft X-ray emission with gas pressure in a
plasma focus. IEEE Trans Plasma Sci 26:1146–1153
9. Kulkoulprakarn Titisak, Ngamrungroj Dusit, Kamsing Pirud, Wong Chiow San,
Mongkolnavin Rattachat (2007) X-ray source structures of a small plasma focus device.
J Sci Res Chula Univ 32:55–60
10. Mohammadi MA, Verma R, Sobhanian S, Wong CS, Lee S, Springham SV, Tan TL, Lee P,
Rawat RS (2007) Neon soft X-ray emission studies from UNU-ICTP plasma focus operated
with longer than optimal anode length. Plasma Sources Sci Technol 16:785–790
11. Mohammadi MA, Sobhanian S, Wong CS, Lee S, Lee P, Rawat RS (2009) The effect of anode
shape on neon soft X-ray emissions and current sheath configuration in plasma focus device.
J Phys D Appl Phys 42:045203
12. Yap SL, Wong CS (2007) Development of a 3.3 kJ plasma focus as pulsed neutron source.
J Sci Technol Tropics 3:123–127
13. Yap SL, Wong CS, Choi P, Dumitrescu C, Moo SP (2005) Observation of two phases of
neutron emission in a low energy plasma focus. Jpn J Appl Phys 44:8125–8132
14. Choi P, Deeney C, Wong CS (1988) Absolute timing of relativistic electron beam in a plasma
focus. Phys Lett A 128:80–83
15. Choi P, Deeney C, Herold H, Wong CS (1990) Characterization of self-generated intense
electron beams in a plasma focus. Laser Part Beams 8:469–476
16. Lee CH, Ngamrungroj D, Wong CS, Mongkolnavin R, Low YK, Singh J, Yap SL (2005)
Correlation between the current sheath dynamics in the axial acceleration phase of the plasma
focus and its ion beam generation. J Sci Technol Tropics 1:51–54
17. Lim LK, Yap SL, Wong CS, Zakaullah M (2013) Deuteron beam source based on Mather type
plasma focus. J Fusion Energy 32:287–292
18. Ngoi SK, Yap SL, Goh BT, Ritikos R, Rahman SA, Wong CS (2012) Formation of
nano-crystalline phase in hydrogenated amorphous silicon thin film by plasma focus ion beam
irradiation. J Fusion Energy 31(1):96–103
19. Goh BT, Ngoi SK, Yap SL, Wong CS, Dee CF, Rahman SA (2013) Structural and optical
properties of the Nc–Si: H thin films irradiated by high energetic ion beams. J Non-Cryst
Solids 363:13–19
20. Goh BT, Ngoi SK, Yap SL, Wong CS, Rahman SA (2013) Effect of energetic ion beam
irradiation on structural and optical properties of A–Si: H thin films. Thin Solid Films
529:159–163
21. Lee S, Conrad H (1976) Measurements of neutrons and X-rays from a vacuum spark. Phys
Lett 57A:233–236
22. Wong CS, Lee S (1984) Vacuum spark as reproducible X-ray source. Rev Sci Instrum
55:1125–1128
23. Wong CS, Ong CX, Lee S, Choi P (1992) Observation on enhanced pre-breakdown electron
beams in a vacuum spark with a hollow cathode configuration. IEEE Trans Plasma Sci
20:405–409
24. Wong CS, Ong CX, Moo SP, Choi P (1995) Characteristics of a vacuum spark triggered by
the transient hollow cathode discharge electron beam. IEEE Tran Plasma Sci 23:265–269
25. Wong CS, Lee S, Ong CX, Chin OH (1989) A compact low voltage flash X-ray tube. Jpn J
Appl Phys 28:1264–1267
26. Wong CS, Woo HJ, Yap SL (2007) A low energy tunable pulsed X-ray source based on the
pseudospark electron beam. Laser Part Beams 25:497–502
27. Wong CS, Singh J (2012) Malaysian patent: pulsed plasma X-ray source, MY-145318-A
28. Amin YM, Mahat RH, Donald D, Shankar P, Wong CS (1998) Measurement of X-ray
exposure from a flash X-ray tube using TLD 100 and TLD 200. Radiat Phys Chem 51:479
29. Bradley DA, Wong CS, Ng KH (2000) Evaluating the quality of images produced by soft
X-ray units. Appl Radiat Isot 53:691–697
30. Chan LS, Tan D, Saboohi S, Yap SL, Wong CS (2014) Operation of an electron beam initiated
metallic plasma capillary discharge. Vacuum 103:38–42
31. Wong CS, Lem SP, Goh BT, Wong CW (2009) Electroless plating of copper on polyimide
film modified by 50 Hz plasma graft polymerization with 1-Vinylimidazole. Jpn J Appl Phys
48:036501
32. Prasertsung I, Mongkolnavin R, Damrongsakkul S, Wong CS (2010) Surface modification of
dehydrothermal crosslinked gelatin film using a 50 Hz oxygen glow discharge. Surf Coat
Technol 205:S133–S138
33. Prasertsung I, Kanokpanont S, Mongkolnavin R, Wong CS, Panpranot J, Damrongsakkul S
(2012) Plasma enhancement of in vitro attachment of rat bone-marrow-derived stem cells on
cross-linked gelatin films. J Biomater Sci Polym Ed 23:1485–1504
34. Prasertsung I, Kanokpanont S, Mongkolnavin R, Wong CS, Panpranot J, Damrongsakkul S
(2013) Comparison of the behavior of fibroblast and bone marrow-derived mesenchymal stem
cell on nitrogen plasma-treated gelatin films. Mater Sci Eng C 33:4475–4479
35. Amornsudthiwat Phakdee, Mongkolnavin Rattachat, Kanokpanont Sorada, Panpranot
Joongjai, Wong Chiow San, Damrongsakkul Siriporn (2013) Improvement of early cell
adhesion on thai silk fibroin surface by low energy plasma. Colloids Surf B 111:579–586
36. Tay WH, Yap SL, Wong CS (2014) The electrical characteristics and modeling of a
filamentary dielectric barrier discharge in atmospheric air. Sains Malaysiana 43(4):583–594
37. Tay WH, Kausik SS, Wong CS, Yap SL, Muniandy SV (2014) Statistical modelling of
discharge behavior of atmospheric pressure dielectric barrier discharge. Phys Plasmas
21:113502
38. Tay WH, Kausik SS, Yap SL, Wong CS (2014) Role of secondary emission on discharge
dynamics in an atmospheric pressure dielectric barrier discharge. Phys Plasmas 21:044502
39. Ramasamy Rajeswari K, Rahman Noorsaadah A, San Wong Chiow (2001) Effect of
temperature on the ozonation of textile waste effluent. Color Technol 117:95–97
40. Hashim Siti Aiasah, San Wong Chiow, Abas Mhd Radzi, Hj Khairul Zaman, Dahlan (2007)
Feasibility study on the removal of Nitric Oxide (NO) in gas phase using dielectric barrier
discharge reactor. Malays J Sci 26:111–116
41. Hashim SA, Wong CS, Abas MR, Hj Dahlan KZ (2010) Discharge based processing systems
for nitric oxide (NO) remediation. Sains Malays 39:981–987
42. Bhai Tyata Raju, Prasad Subedi Deepak, Rajendra Shrestha, San Wong Chiow (2013)
Generation of uniform atmospheric pressure argon glow plasma by dielectric barrier discharge.
Pramana 80:507–517
43. Lee YS, Bora B, Yap SL, Wong CS (2012) Effect of ambient air pressure on synthesis of
copper and copper oxide nanoparticles by wire explosion process. Curr Appl Phys 12:199–203
References 123
44. Wong CS, Bora B, Lee YS, Yap SL, Bhuyan H, Favre M (2012) Effect of ambient gas species
on the formation of cu nanoparticles in wire explosion process. Curr Appl Phys 12:1345–1348
45. Bora B, Wong CS, Bhuyan H, Lee YS, Yap SL, Favre M (2012) Impact of binary gas on
nanoparticle formation in wire explosion process: an understanding via arc plasma formation.
Mater Lett 81:45–47
46. Bora B, Wong CS, Bhuyan H, Lee YS, Yap SL, Favre M (2013) Understanding the
mechanism of nanoparticle formation by wire explosion process. J Quant Spectrosc Radiat
Transfer 117:1–6
47. Bora B, Kausik SS, Wong CS, Chin OH, Yap SL, Soto L (2014) Observation of the partial
reheating of the metallic vapor during the wire explosion process for nanoparticle synthesis.
Appl Phys Lett 104:223108

Elements of Plasma Technology: Chiow San Wong Rattachat Mongkolnavin

Uploaded by

Copyright:

Available Formats

Elements of Plasma Technology: Chiow San Wong Rattachat Mongkolnavin

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

Elements of Plasma Technology: Chiow San Wong Rattachat Mongkolnavin

Uploaded by

Copyright:

Available Formats

SPRINGER BRIEFS IN

APPLIED SCIENCES AND TECHNOLOGY

ISSN 2191-530X ISSN 2191-5318 (electronic)

Library of Congress Control Number: 2015957048

© The Author(s) 2016

Printed on acid-free paper

This Springer imprint is published by SpringerNature

Plasma technology is often referred to as one of the critical technologies of the

Chiow San Wong

1 Basic Concepts in Plasma Technology. . . . ............... . . . . 1

3 Plasma Diagnostic Techniques . . . . . . . . . . . . . . . . . . . . . . . . .... 49

4.6 The 50 Hz Alternating Current (A.C.) Glow

Keywords Plasma Plasma properties

1.1 Plasma—The Fourth State of Matter

© The Author(s) 2016 1

Fig. 1.1 Plasma—the fourth state of matter

1.2.1 Elastic Collision

Fig. 1.2 Collision between m1 u1 sinθ

Similarly, considering conservation of energy before and after the collision,

The fraction of energy transferred during the collision:

ðiÞ m1 ¼ m2 ) d ¼ cos2 h; dmax ¼ 1

1.2.2 Inelastic Collision

But the energy equation has to be written as

For ΔU = maximum, let d(ΔU)/dv2 = 0

For m1 m2 ; d ¼ cos2 h; dmax ¼ 1.

1.3 Collision Cross-Section

In the consideration of collision between electron and atom or ion an important

where ‘nK  10 for fully or strongly ionized plasma.

1.4 Fundamental Processes in a Plasma

Fig. 1.3 Relative cross-sections of scattering, excitation and ionization processes

1.5 Some Consideration of Thermodynamic Properties

Let us consider a gas at sufﬁciently low temperature when it can be considered to

is the fractional population of neutral particles, and

where χ is the fractional population of molecules.

1.6 Concept of Plasma Potential

One of the important characteristics of plasma is that charge neutrality is assumed to

Note that fe is only dependent on the density of electrons.

1.7 Criteria of Plasma

xp ¼ 2pfe is the angular frequency of electron plasma oscillation while τ is the

1.8 Effect of Boundary in Plasma

Fig. 1.4 Potential

Fig. 1.5 Potential x=0

Positive Charge Region

1.9 Particle Transport Inside Plasma

So far we have considered charged particles in the plasma to be moving randomly

where τe-a is the electron-atom collision mean

Again, more frequent collision means lower conductivity.

Abstract In this chapter, the generation of plasma by gaseous electrical discharge

Keywords Electrical discharge Pulsed discharge

2.1 DC Electrical Discharge

2.1.1 Electrical Breakdown

© The Author(s) 2016 15

Fig. 2.1 Multiplication of charged particles leading to electrical discharge

collisions. Hence we see that the number of electrons will be multiplied as we

for atomic gases, and

for molecular gases.

For an anode-cathode distance of d, the electron flux ﬁnally collected by the

This is reasonably well obeyed by experimental data, as illustrated in the plot of

Fig. 2.2 Plot of lnIe against Current (pA)

where ‘nK 10 for fully or strongly ionized plasma.