(Optical Engineering (Marcel Dekker Inc.) 21) Cremers, David a._ Radziemski, Leon J - Laser-Induced Plasmas and Applications-M. Dekker (1989)

Laser-Induced
Plasmas
and
Applications
OPTICAL ENGINEERING
Series Editor
Brian J. Thompson
Provost
University of Rochester
Rochester, New York
Laser Advances Editor:

/,ascr l:'ngineering Editor:
Leon J. Radziems ki
Peter K. Oteo
Associate Dean, College of Arts and Sciences
United Technologies Research Center
New Mexico State University
Hartford, Connecticut
Las Cruces, New Mexico
Optical Materials Editor:

Solomon Musikant
Paoli. Pennsylvania
l. Electron and Ion Microscopy and Microanalysis: Principles and Applications,

by Lawrence E. Murr
2. Acousto-Optic Signal Processing: Theory and Implementation, edited by
Norman J. Berg and John N. Lee
3. Electro-Optic and Acousto-Optic Scanning and Deflection, by !lfilton
Gottlieb, Clive L. M. Ireland, and John Martin Ley
4. Single-Mode Fiber Optics: Principles and Applications, by l.uc B. Jeunhomme
5. Pulse Code Formats for Fiber Optical Data Communication: Basic Principles
and Applications, by David J. Morris
6. Optical Materials: An Introduction to Selection and Application,
by Solomon Musikant
7. Infrared Methods for Gaseous Measurements: Theory and Practice, edited by
Joda Wormhoudt
8. Laser Beam Scanning: Opto-Mechanical Devices, Systems, and Data Storage
Optics, edited by Gerald F. Marshall
9. Opto-Mechanical Systems Design, by Paul R. Yoder, Jr.
10. Optical Fiber Splices and Connectors: Theory and Methods, by Calvin M.
Miller with Stephen C. Mettler and Ian A. White
11. Laser Spectroscopy and Its Applications, edited by Leon J. Radziemski,
Richard W. Solarz, and Jeffrey A. Paisner
12. Infrared Optoelectronics: Devices and Applications, by William Nunley
and J. Scott Bechtel
13. Integrated Optical Circuits and Components: Design and Applications,
edited by Lynn D. Hutcheson
14. Handbook of Molecular Lasers, edited by Peter K. Cheo
15. Handbook of Optical Fibers and Cables, by Hiroshi Murata
16. Acousto-Optics, by Adrian Karpel
17. Procedures in Applied Optics, by John Strong
18. Handbook of Solid-State Lasers, edited by Peter K. Cheo
19. Optical Computing: Digital and Symbolic. edited by Raymond Arrathoon
20. Laser Applications in Physical Chemistry. edited by D. K. E~·ans
21. Laser-Induced Plasmas and Applications. edited by Leon J. Radziemski

and David A. Crem ers
LASER HANDBOOKS-Edited by Peter K. Cheo

Handbook of Molecular Lasers
Handbook of Solid-State Lasers
Other Volumes in Preparation
Single-Mode Fiber Optics: Principles and Applications, Second Edition,

Revised and Expanded, by Luc B. Jeunhomme
Infrared Technology Fundamentals, by Irving J. Spiro and Monroe
Schlessinger
Laser-Induced
Plasmas
and
Applications
edited by
Leon J. Radziemski
Department of Physics
Las Cruces, New Mexico
David A. cremers
Chemical and Laser Sciences Division
Los Alamos National Laboratory
Los Alamos, New Mexico
MARCEL DEKKER New York-

Library of Congress Cataloging-in-Publication Data
Laser-induced plasmas : physical, chemical, and biological applications I edited

by Leon J. Radziemski, David A. Cremers.
p. em.
Includes bibliographies.
ISBN 0-8247-8078-7 (alk. paper)
1. Plasma engineering. 2. High power lasers. I. Radziemski, Leon J .,
II. Cremers, David A.
TA2020.L37 1989
620.044--dc20 89-7883
CIP
This book is printed on acid-free paper.
Copyright© 1989 MARCEL DEKKER All Rights Reserved
Neither this book nor any part may be reproduced or transmitted in any form
or by any means, electronic or mechanical, including photocopying, microfilming,
and recording, or by any information storage and retrieval system, without per-
mission in writing from the publisher.
MARCEL DEKKER
270 Madison Avenue, New York, New York 10016
Current printing (last digit):

10 9 8 7 6 5 4 3 2
PRINTED IN THE UNITED STATES OF AMERICA

About the Series
Optical science, engineering, and technology have grown rapidly in the last dec-
ade so that today optical engineering has emerged as an important discipline in
its own right. This series is devoted to discussing topics in optical engineering at
a level that will be useful to those working in the field or attempting to design
systems that are based on optical techniques or that have significant optical sub-
systems. The philosophy is not to provide detailed monographs on narrow sub-
ject areas but to deal with the material at a level that makes it immediately use-
ful to the practicing scientist and engineer. These are not research monographs,
although we expect that workers in optical research will find them extremely
valuable.
Volumes in this series cover those topics that have been a part of the rapid
expansion of optical engineering. The developments that have led to this expan-
sion include the laser and its many commercial and industrial applications, the
new optical materials, gradient index optics, electro- and acousto-optics, fiber
optics and communications, optical computing and pattern recognition, optical
data reading, recording and storage, biomedical instrumentation, industrial ro-
botics, integrated optics, infrared and ultraviolet systems, etc. Since the optical
industry is currently one of the major growth industries this list will surely be-
come even more extensive.
Brian J. Thompson
Rochester, New York
iii
Preface
Plasmas created by the focused radiation from pulsed laser beams were first
reported by Maker, Terhune, and Savage in 1963. Since then, this phe-
nomenon has been studied and applied using a wide range of laser ener-
gies. Much of the early work on laser plasmas and applications has been
described in reviews during the last 25 years. Among these are the classic
work by Y. P. Raizer, Laser-Induced Discharge Phenomena and the review
edited by G. Bekefi, "Principles of Laser Plasmas". The present volume up-
dates and extends those reviews to new subjects. Some overlap of material
exists, but where possible we have departed from the emphasis of earlier re-
views. This is due in part to the development of new areas of interest. As in
other volumes on laser applications in the Optical Engineering series, this
book is both tutorial and review. It will familiarize the reader with the tools
and concepts used in each area discussed, review the state of the subject,
and provide sufficient references for in-depth follow-up .
.As the field of laser plasmas develops, the understanding of the funda-
mental phenomena increases. For this reason, the first two chapters review
the current understanding of the basic physics of plasma initiation, growth,
and decay. In Chapter 1, the physics of plasma initiation is discussed in de-
tail, with emphasis on the two mechanisms for initiating plasmas in gases:
inverse bremsstrahlung and multiphoton ionization. The question of the
"first electron," a problem that varies in importance depending on the laser
wavelength, is discussed. Growth of the electron density, resulting in the
plasma formation, is treated next. Optically-induced breakdown of liquids
and solids is introduced, and some recent results are reviewed.
Following plasma formation, the stored energy is dissipated through a va-
riety of mechanisms. The relative importance of these mechanisms depends
on the parameters of the laser pulse and the medium. In Chapter 2, the
features of dissipating, propagating plasmas are reviewed. Three regimes,
with different absorption characteristics and hydrodynamic behavior, are
covered. These are the laser-supported combustion, detonation, and radi-
ation waves. Modeling codes are described, and applications of the three
different types of propagating plasmas are summarized.
Diagnostics for low-energy plasmas are well known. For this reason,
Chapter 3 deals with advances in diagnostics for high-energy, laser-fusion
v
vi Preface
plasmas, in which laser irradiances are typically 1013 to 1016 W cm- 2 • Where
applicable, connections are made to lower temperature, low-irradiance con-
ditions which prevail in plasmas used for chemical analysis and materials
processing. Also presented are new techniques in x-ray diagnostics and
their application to energy transport, implosion hydrodynamics, and very
dense plasmas. The theoretical foundation of these plasma diagnostics is
discussed.
Although initiation of a laser plasma requires focused energy densities
obtainable only from high-power pulsed lasers, it is possible to sustain a
continuous plasma using the much lower powers available from the cw car-
bon dioxide laser. In Chapter 4 the basic physical processes, plasma char-
acteristics, and implementation of the "continuous optical discharge" are
reviewed. The results of recent experimental and theoretical studies are
presented, along with discussions of some proposed applications such as
laser-based propulsion and spectrochemical analysis.
Over the last 20 years, fusion has been developed as an alternative power
source, using the magnetic and inertial confinement approaches. In Chap-
ter 5, the laser-based inertial confinement technique is reviewed. After a
brief history, the current status of laser fusion research is discussed, includ-
ing detailed treatments of laser-driven ablation and requirements upon ir-
radiation uniformity. The chapter closes with a description of recent and
future implosion experiments.
A new and growing application area for lasers is semiconductor fabrica-
tion, the subject of Chapter 6. This chapter has a slightly different focus
from that of other chapters because of the wide range of laser-based fabri-
cation techniques in use. The review deals not only with ablative procedures
which rely on plasma formation but, for completeness, with irradiation pro-
cedures using lower laser power densities. Examples of the former include
resistor trimming, redundancy and photomask repair, device marking, and
via drilling.
Over the last 25 years, laser plasmas have been used in a variety of spec-
trochemical applications. The most successful of these relate to problems
that are difficult to handle by conventional techniques. These include sam-
pling in situ and in hostile environments. Chapter 7 reviews recent appli-
cations where the laser plasma (sometimes called the laser spark) is both
vaporization and excitation source. Applications to gases, bulk liquids, par-
ticles, and solids are treated. In the latter case, recent progress in solid
and molten metal analysis is reviewed. Also, an operating instrument for
beryllium monitoring is discussed. Chapter 8 describes the considerations
for successful sampling on solid surfaces. As such, it is closely related to
the solid analysis section of Chapter 7. In particular, the relationship be-
tween spectral intensity measurement and elemental analysis is developed
and discussed.
Preface vii
In other spectrochemical applications, the laser plasma has been used

primarily for sample vaporization and atomization, with the actual elemen-
tal analysis being provided by a different, auxHiary technique. This is the
subject of Chapter 9. The first part of the chapter reviews different combi-
nations of the laser plasma with a second spectroscopic method, established
atomic emission or absorption techniques. The second part outlines laser
ablation mass spectrometry, a technique capable of obtaining information
on very small samples.
Some of the newest applications of laser plasmas are likely to be active
topics of investigation for years to come. They are the subject of Chap-
ter 10. The chapter begins with a thorough discussion of laser plasmas
as sources of x-rays and particles. This is followed by an explanation of
how the unique properties of these sources can be applied to diagnostics of
transiently strained crystals and biological samples. The intriguing area of
laser-plasma-based particle accelerators is discussed next, followed by a de-
scription of laser-triggered, pulsed-power switching.
Not all laser-plasma applications have been discussed. 1\vo obvious
omissions are laser welding and cutting, and laser plasma applications in
medicine. Clearly, space limits the number of topics that can be considered.
We wish to acknowledge the competent and imaginative work of the au-
thors. The usefulness of this volume is a result of their key positions in
their fields and of their thoughtful elucidation of the subject matter. We
hope that this volume will serve as a useful reference to the reader and as a
stimulus to further research into laser-plasma applications.
Leon J. Radziemski
David A Cremers
Contents
Preface iii
Contributors XI
1 Physics of Laser-Induced Breakdown: An Update 1

GuyM. Weyl
1.1 Introduction 1
1.2 Creation of Initial Electrons 3
1.3 Electron Growth in Gases 8
1.4 Laser-Induced Breakdown of Solids and Liquids 36
1.5 Concluding Remarks 58
References 59
2 Modeling of Post-Breakdown Phenomena 69

Robert G. Root
2.1 Introduction 69
2.2 Creation of a Propagating Plasma 70
2.3 Absorption Characteristics of Heated Gases 72
2.4 Features of Propagating Plasmas 75
2.5 One-Dimensional Laser-Supported Combustion Waves 77
2.6 One-Dimensional Laser-Supported Detonation Wave 88
2.7 One-Dimensional Laser-Supported Radiation Wave 92
2.8 Transition Regions 93
2.9 Radial Expansion 95
2.10 Thermal Coupling 99
2.11 Other Factors 100
2.12 Summary 101
References 101
3 Introduction to Laser Plasma Diagnostics 105

Allan A. Hauer and Hector A. Baldis
3.1 Introduction 105
3.2 Introduction to Optical Diagnostics 110
ix
X Contents
3.3 Introduction to X-ray Diagnostics 131

References 161
4 Laser-Sustained Plasmas 169

Dennis R. Keefer
4.2 Principles of Operation 172
4.3 Analytical Models 182
4.4 Experimental Studies 189
4.5 Applications of the Laser-Sustained Plasma 196
References 203
5 lnertially Confined Fusion 207

Robert L. McCrory and John M. Soures
5.1 Historical Overview 207
5.2 Laser-Fusion Scaling Laws 211
5.3 Coronal Physics 217
5.4 X-ray Generation by Laser-Produced Plasmas 224
5.5 Laser-Driven Ablation 227
5.6 Hydrodynamic Stability of Ablatively Driven Shells 239
5.7 Irradiation Uniformity Requirements 243
5.8 Implosion Experiments 251
References 260
6 Laser-Based Semiconductor Fabrication 269

Joseph R. Wachter
6.1 Aspects of Semiconductor Fabrication 269
6.2 Applications of Lasers in the Semiconductor Industry 276
6.3 Research Areas 283
6.4 Outlook 290
References 291
7 Spectrochemical Analysis Using Laser Plasma Excitation 295

Leon J. Radziemski and David A. Cremers
7.1 Review 295
7.2 Methods and Properties of Analysis Using Laser Plasmas 296
7.3 Analysis of Gases 302
7.4 Analysis of Bulk Liquids 306
7.5 Analysis of Particles 309
7.6 Analysis of Solids 313
7.7 Advances in Instrumentation 318
Contents xi
7.8 Prognosis 321

References 323
8 Fundamentals of Analysis of Solids by Laser-Produced

Plasmas 327
YongW. Kim
8.1 Chapter Organization 327

8.3 Phenomenology of Laser Heating of Condensed-Phase
Targets 330
8.4 Quantitative Spectroscopy 336
8.5 Intensity Measurements and Elemental Analysis 341
8.6 Summary 344
References 345
9 Laser Vaporization for Sample Introduction in Atomic and

Mass Spectroscopy 347
Joseph Sneddon, Peter G. Mitchell, and Nicholas S. Nagar
9.1 Conventional Solid Sample Introduction for Atomic

Spectroscopy 347
9.2 Laser Ablation of Solid Samples 350
9.3 Laser Ablation for Sample Introduction in Atomic
Spectroscopy 353
9.4 Relative Merits of Laser Ablation for Sample Introduction
in Atomic Spectroscopy 363
9.5 Laser Sources for Mass Spectrometry 365
9.6 Applications of Laser Microprobe 369
9.7 Applications of Laser Desorption and Postionization 372
9.8 Conclusion 376
References 376
10 Current New Applications of Laser Plasmas 385

Allan A. Hauer, David W. Forslund, Colin J. McKinstrie,
Justin S. Wark, Philip J. Hargis, Jr., Roy A. Hamil, and Joseph
M. Kindel

10.2 Applications of Laser-Plasma-Generated X-rays and
Particles 386
10.3 Laser-Plasma Acceleration of Particles 413
xii Contents
10.4 Laser-Pulsed Power Switching 424

References 432
Index 437
Contributors
Hector A. Baldis Senior Research Officer, Division of Physics, Laser and

Plasma Physics Section, National Research Council of Canada, Ottawa, On-
tario, Canada
David A. Cremers Chemical and Laser Sciences Division, Los Alamos N a-

tiona! Laboratory, Los Alamos, New Mexico
David W. Forslund Applied Theoretical Physics Division, Los Alamos Na-

tional Laboratory, Los Alamos, New Mexico
Roy A. Hamil Technical Staff, Laser Projects, Sandia National Laborato-

ries, Albuquerque, New Mexico
Philip J. Hargis, Jr. Technical Staff, Laser and Plasma-Processing Sci-

ences, Sandia National Laboratories, Albuquerque, New Mexico
Allan A. Hauer Acting Deputy Group Leader, Laser Plasma and Fusion
Experiments, Los Alamos National Laboratory, Los Alamos, New Mexico
Dennis R. Keefer Professor, Engineering Science and Mechanics, Center

for Laser Applications, University of Tennessee Space Institute, Tullahoma,
Tennessee
Yong W. Kim Professor, Department of Physics, Lehigh University, Beth-

lehem, Pennsylvania
Joseph M. Kindel Division Manager, Mission Research Corporation, Los

Alamos, New Mexico
Robert L. McCrory Director, Laboratory for Laser Energetics, University

of Rochester, Rochester, New York
xiii
xiv Contributors
Colin J. McKinstrie* Director's Fellow, Center for Nonlinear Studies and

Applied Theoretical Physics Division, Los Alamos National Laboratory,
Peter G. Mitchell Department of Medicine, Medical College of Wiscon-

sin, Milwaukee, Wisconsin
Nicholas S. Nogar Deputy Group Leader, Chemical and Laser Sciences

Division, Los Alamos National Laboratory, Los Alamos, New Mexico
Leon J. Radziemskit Professor, Department of Physics, New Mexico State

University, Las Cruces, New Mexico
Robert G. Root Director, Applied Physics Office, SPARTA, Inc., Lexing-

ton, Massachusetts
John M. Soures Deputy Director, Laboratory for Laser Energetics, Uni-

versity of Rochester, Rochester, New York
Joseph Sneddon Associate Professor, Department of Chemistry, Univer-

sity of Lowell, Lowell, Massachusetts
Joseph R. Wachter Materials Science and Technology Division, Los Ala-

mos National Laboratory, Los Alamos, New Mexico
Justin S. Warkt Research Associate, Laboratory for Laser Energetics,

University of Rochester, Rochester, New York
Guy M. Weyl Principal Research Scientist, Physical Sciences, Inc., An-

dover, Massachusetts
Present Affiliation:
• Scientist, Laboratory for Laser Energetics, University of Rochester, Rochester, New York
t Associate Dean, College of Arts and Sciences
+Royal Society University Fellow, Oxford University, Oxford England
1
Physics of Laser-Induced
Breakdown: An Update
Guy M. Weyl
Physical Sciences Inc.
Andover, Massachusetts
1.1 INTRODUCTION
Soon after the development of the ruby laser, experiments were carried
out that showed that by bringing a laser beam to a tight focus, one could
cause the breakdown of air (Damon and Thomlinson, 1963; Meyerand and
Haught, 1963; Minck, 1964). A bright spot would be visible at focus simi-
lar in appearance to a discharge spark between two electrodes. Since then,
many experiments have been carried out on various gases, solids, and liq-
uids. The purpose of those experiments was to determine the mechanisms
leading to breakdown, to measure the threshold power densities for spark
ignition, and to study the influence of various parameters such as wave-
length, focal diameter, pressure, pulse length, and material on breakdown
thresholds. Many good reviews of the work conducted up to the mid-70s
have appeared (Raizer, 1977; Smith and Meyerand, 1976; Grey-Morgan,
1975, Ready, 1971). Experiments were initially conducted at the wave-
lengths of the high-power lasers available: 10.6 p,m (C0 2 ), 3.8 (DF), 2.7
(HF), 1.06 (Nd, glass), and 0.69 p,m (ruby). The development of dye-laser
systems and excimer lasers and the use of nonlinear crystals to double,
triple, and even quadruple the frequency have resulted in a much broader
wavelength range for experimental studies. Emphasis in more recent years
has been on short-wavelength excimer lasers such as KrF (0.25 p,m) and XeF
(0.35 p,m). The Raman-shifted output ofthese lasers will produce high laser
pulses in the midvisible region.
Laser-induced breakdown can be defined as the generation, by the end
of the pulse, of a practically totally ionized gas (plasma). The experimental
criterion that is generally used is the observation of a glow or flash in the fo-
cal region. Other more quantitative criteria involve the measurement of the
1
2 Weyl
attenuation of the laser beam and the level of ionization in the focal region.
There are two main mechanisms for electron generation and growth. The
first mechanism involves absorption of laser radiation by electrons when
they collide with neutrals. If the electrons gain sufficient energy, they can
impact ionize the gas or solid through the reaction
(1.1)
Reaction (1.1) will lead to cascade breakdown, i.e., the electron concentra-
tion will increase exponentially with time. The situation is very similar to the
dielectric breakdown of solids, liquids, and gases. Two necessary conditions
are (1) that there be an initial electron in the focal volume, and (2) that the
electrons acquire an energy greater than the ionization energy of the gas (or
band gap of the solid). The second mechanism, called multi photon ioniza-
tion (MPI), involves the simultaneous absorption by an atom or molecule of
a sufficient number of photons to cause its ionization (or to eject an electron
from the valence to the conduction band). MPI is described by the reaction
(1.2)
If E1 is the ionization potential (or band gap), the number of photons m must
exceed the integer part of (e1 /hv + 1). The ionization rate in a laser beam
of irradiance" I is proportional to Im and the electron density, for constant
I, increases linearly with time. Multiphoton ionization is important only at
short wavelengths (..\ < 1 pm). Ionization potentials of most gases are larger
than 10 eV. The simultaneous absorption of more than 100 photons from
a C02 laser (hv = 0.1 eV) would be required to ionize most gases, and this
is highly improbable.
Both cascade and multiphoton ionization require high laser irradiances,
usually in excess of lOS W/cm2 • However, breakdown of solids has been
observed at irradiances as low as l<J<i W/cm2 • Ifthe solid is transparent to
the laser radiation, an apparently low breakdown irradiance may be due to
a nonlinear phenomenon called self-focusing that causes the beam, above
a given threshold laser power, to come to a tighter focus than one would
expect from the laws of optics using the low-field index of refraction of the
medium. If the solid is absorbing or has microscopic absorption sites, one
can invoke a third mechanism to explain the low thresholds, which we shall
call thermal runaway. Absorption of radiation by the solid (either bulk or
•The laser community commonly uses the word "intensity" interchangeably with "flux density,"
or "irradiance" that have units of power per unit area. Since intensity is not a well-defined
radiometric unit, we will use irradiance in this chapter and express this quantity in W/cm2 •
Update on Breakdown Physics 3
at impurity sites) causes its vaporization, thus driving a shock into the sur-
rounding air. Absorption of radiation by electrons in the vapor or behind
the shock leads to heating of the vapor or shocked air. This leads to ther-
mal generation of more electrons and concurrently to a higher absorption
rate, i.e., we have a thermal runaway. Thermal runaway is expected to oc-
cur when the vapor or the shocked air contains enough electrons initially to
enable the gas to heat up over the pulse times of interest. All these mech-
anisms will be discussed in more detail below.
1.2 CREATION OF INITIAL ELECTRONS
The initial electron from which a cascade can develop can be readily gener-
ated by multi photon ionization of the gas or solid if the wavelength is short
enough and the laser irradiance high enough. Consider, e.g., the breakdown
of nitrogen at A = 1.06 p,m and p = 1 atm. The threshold irradiance for cas-
cade breakdown is 2 x 1011 W/cm2 for a pulse length of 30 nsec (Armstrong,
et al., 1983). Multiphoton ionization of N2 requires absorption of 12 pho-
tons and the MPI cross section for an incoherent beam has been measured
(Baravian, et al., 1982) to be o-12 = 4 x 10- 142 (W/cm 2 )- 12 sec- 1 • The num-
ber density of electrons generated in 10 nsec by MPI for I= 2 x 1011 W/cm2
is thus calculated to be 4 X 10-142 X (2 X 1011 ) 12 X 10-8 X n = 2 X 10- 15 n,
where n is the number density of nitrogen molecules. At 1 atm pressure,
n = 2 x 1019 cm- 3 and one expects one electron to have been generated by
MPI if the effective focal volume exceeds 2 x w- 5 cm3 • The presence of im-
purities having low ionization potentials, such as organic vapors, can also be
expected to contribute significantly to the generation of initial electrons by
MPI.
The problem of the initial electron becomes more serious at longer wave-
lengths. The cascade breakdown thresholds are much lower (they scale
as A- 2 , see Sec. 1.4) and MPI cannot furnish any electrons. Experiments
conducted in air at 10.6 p,m wavelength showed breakdown to be a spo-
radic event. The probability of initiating a breakdown would increase over
a range of irradiance from 0 to 100%. Definition of threshold was taken to
be that intensity above which the gas would break down on more than half of
the shots (Hill, et al., 1971; Chan et al., 1973). It was discovered that break-
down was initiated by aerosols in the focal volume (Lencioni, 1973; Schlier,
1973; Canavan and Nielsen, 1973). The distribution of aerosols in the atmo-
sphere is shown in Fig. 1.1. Under normal conditions, there are more than
1o4 particles/cm3 larger than 0.1 p.m. These particles would heat up under
laser irradiation and could generate electrons by thermionic emission. In
order to eliminate this sporadic behavior of breakdown, experiments were
performed on partially preionized gas. The laser pulse would be fired a
4 Weyl
E
::!.
/DESERT DUST CLOUDS
0
0
· SMOG
~
z CONTINENTAL
0
i=
0
z
::J
lL
z
0
i= MARITIME
::J (LOW-SPEED WIND)
en
a:
1--
10
(j)
0 McCLATCHEY
w MODEL
N
U5 1o-2
w
....1
0 JUNGE
~ 10-4
<
CL
[(
;;::::
PARTICLE RADIUS (J.lml
Figure 1.1 Solid aerosol distribution functions for various aerosols in the atmo-
sphere, following Reilly et a!. (1977).
certain time 6.t after a discharge was passed through the gas (Chalmeton,
1969; Smith, 1970; Brown and Smith, 1973; Chan et al., 1973). Assuming
a two-body dissociative recombination rate of 10-7 cm3 sec- 1 for N 2+, one
can readily calculate the decay of the electron concentration
dne = -10-7n2
dt e
or
(1.3)
where neo is the initial electron concentration. We assumed in the last step
that ne ~ neo· Equation (1.3) shows that if b.t < 10-3 sec, ne will be larger
than 1010 cm- 3 , a large enough number density to provide initial electrons
in the focal volume.
Experiments carried out by Yablonovich (1974, 1975) at 10.6 J.Lm con-
clusively demonstrated the influence of the initial electron on breakdown
thresholds. He showed that the thresholds would increase many fold when
one eliminated impurities and particles from the focal volume. By filtering
out all particles larger than 0.1 f..lm, he was able to increase the breakdown
threshold at 10.6 f..lm to 1012 W/cm2 • A further tenfold increase in threshold
in a very pure nitrogen gas was obtained by slow evaporation of the liquid.
The gas thus obtained should practically be devoid of aerosols and organic
impurities with low ionization energies.
At sufficiently high laser irradiances, however, the electric field may be
strong enough to pull an electron in an outer shell out of its orbital through
the tunnel effect. A semiclassical treatment of photoionization by tunneling
in a laser field has been developed by Keldysh (1964). Below we present a
much simpler treatment for a rough estimate of the effect.
Consider a bound electron in a square potential well of depth~:. Under an
applied electric field E, the well will be deformed as shown by the dashed
line in Fig. 1.2.
The decay rate of the bound state is given by
(1.4)
where vB is the bounce frequency of the bound electron in the well and T
the transmission coefficient of the barrier. The bounce frequency is related
to the action integral J through (Bohm, 1951)
(1.5)
..
X
Figure 1.2 Electron tunneling in a square potential well.

6 Weyl
where me= electron mass and e =electron charge. The transmission factor
in the WKB approximation is
(1.6)
The upper limitxT is the location where the electron energy becomes pos-
itive, i.e., xT = deE.
Performing the integrals and keeping only first-order terms in the pa-
rameter (eEa/e), we obtain
(1.7)
and
4 (2m )112 e312]

T=exp- [ - e (1.8)
3h e E
Equations (1. 7) and (1.8) are valid for a static field. The justification for
using a static field is that vB -:J> w/2-Tr, i.e., the electron bounces many times
before the field reverses. For example, for e = 1 e V and a = 2 A, we have
vB = 7 x 1014 sec- 1, whereas for a C0 2 laser, w/27r = 2.8 x 1013 sec- 1. For a
linearly polarized ac field, however, one must take a time average ofT. We
write E = E 0 coswt and evaluate
{T)=~ j
rr/2r..~
Tdt (1.9)
7r -rr/2r..~
Now, T oc exp( -AI coswt) with A -;;p 1. Using the method of steepest de-
scent, we obtain {T) = T max x (7rA/2)-[(7rA/2)- 112] 112 and Eq. (1.4) becomes
112
_ (
v- -E- ) 1/2 [ 3he ( -1- ) 112 I-1/2]
-
8a2m 1rmec e 12
3
2 (mc)1/2e312]
x exp- [ 3he ----;- I 1/2 (1.10)
where I = (cE~/81r) is the laser irradiance in cgs units. The second bracket
in Eq. (1.10) is the correction factor due to the ac field.
Letting a = 2 A, e be the ionization energy e1 and expressing e1 in eV and

I in W/cm2 , we find
v = 3.8 x 1011 e112 ( eJI ) 1/4 exp- [ 2.4 x 106 ( j3 ) 112] (1.11)
This tells us that, for a given ionization rate, the laser irradiance must scale
as the cube of the ionization potential.
We have plotted v as a function of the I le] in Fig. 1.3. One sees
that the range of interest for the tunnel effect is between 2 x 109 and
2 x 1010 W/(cm2 ey3). For an impurity to provide an electron during the
pulse time TP' one must require that vVn~_> 1, were Vis the focal volume
and n the number density of impurities. Many inorganic vapors have ion-
ization energies around 10 eV. Assuming a 1 ppm impurity concentration
1Q10 . - - - - - - - - - . - - - - - - - - . - - - - - - - - . .
/
;"
,...,.,"" -
...-""
~1eV
/
/ "'
"'
I
I
I
I
I
I
I
I
I
I
I
I
10---{5 I
1o-IO L - - - - - - - L ._ _ _ _ __.__ _ _ _---J
0 1010 2x1o10 3 X 1Q10

II~} (W/(cm2 eV3))
Figure 1.3 Tunnelling frequency versus I It].

8 Weyl
level, rP = 100 nsec (typical width of a C02 gain-switched spike), E1 = 10 eV,

V = 10-5 cm3, and a total gas density of 1 amagat (2.5 x 1019 cm- 3 ), we
must have v > 4 x 10-2 sec- 1 and, from Fig. 1.3, I> 4 x 1012 W/cm2 • This
is a factor of 2 smaller than the breakdown irradiance in Yablonovich's ex-
periment. For pure nitrogen (E1 = 15.5 eV), we calculate, under the same
conditions, v > 5 x 10-9 sec- 1 and I > 8 x 1012 W/cm2 that is very near the
threshold observed (Yablonovich, 1975).
We close this section by noting that, even under dust-free conditions and
with no impurities present, we can always expect a small amount of ion-
ization induced by cosmic rays and natural radio activity of the earth. The
concentration of ions in the atmosphere at sea level is in the range 102 to
1o3 cm- 3 • There are no free electrons since these, once generated, rapidly
attach to 0 2 in three-body collisions to form 02, which has a binding en-
ergy of 0.36 e V (Huber and Herzberg, 1983). The lifetime of 02 is 10-7 sec
at an intensity of 2 x 109 W/cm2 • Since the long-pulse breakdown threshold
of air at 10.6 pm is 3 x 109 W/cm2 (Brown and Smith, 1973; Lencioni, 1973;
Kroll and Watson, 1972; Canavan et al., 1972) and will be even higher at
shorter wavelengths, it is valid to treat the attached electrons as free for all
experiments carried out at >. < 10.6 pm.
1.3 ELECTRON GROWTH IN GASES

1.3.1 Master Equation
The growth of electron density ne is governed by the following equation:
(1.12)
where W m is the multi photon ionization rate coefficient and vi, va, vR are the
impact ionization, attachment, and recombination rates, respectively, and
D the diffusion coefficient for electrons. The various terms in Eq. (1.12)
are discussed below.
1.3.2 Multiphoton Ionization
The multiphoton ionization rate is proportional to Im, where m is the num-

ber of photons that have to be absorbed simultaneously in order to ionize
the gas
m =integer part of (E1/hv + 1) (1.13)

The coefficient Wm is obtained either by calculation or measurements at

very low gas density. The pioneering calculations of multiphoton ioniza-
tion were carried out in the mid-60s (Keldysh, 1965; Bebb and Gold, 1965;
Gontier and Trahin, 1968; Morton, 1967; Voronov, 1967). Most treatments
involve the use of mth-order perturbation theory in order to evaluate the
transition matrix element (il [(e/c)p ·A]m lf), where (il is the initial (bound)
state, lf) the final (free) state, and A the vector potential of the EM field as-
sociated with the laser beam. The calculations become complex and lengthy
when m is large. Multiple sums over intermediate states have to be calcu-
lated. Bebb and Gold (1965) used approximations for the sum over states
and treat separately those intermediate statesj having energies Ej nearly res-
onant with a multiple of the photon energy, i.e., Ej ~ khw (k < m). This leads
to greatly enhanced multiphoton ionization rates near resonance, where the
order of the process shifts from m to k. Keldysh (1965) used a different,
nonperturbative approach and solved the problem of the hydrogen atom in
a strong radiation field. He derived a general expression for the ionization
rate in terms of a parameter 1 defined by
(1.14)
Since eE/mw is the quiver velocity v of a free electron submitted to an al-

ternating field of magnitude E, one can rewrite Eq. (1.14) as
2
'Y = ( wv = (2wr )1/2
!}(, ) 1/2
T (1.15)
whereXr is the distance over which the electron has to tunnel (see Fig. 1.2)
and r is the tunneling time. When r~w- 1 (i.e., ~~1), the tunneling time
is less than the field reversal time. This corresponds to adiabatic condi-
tions where tunneling calculations are expected to be valid. The general
expression derived by Keldysh reduces for small values of 1 to a tunneling
equation almost identical to Eq. (1.10), but with a slightly different preex-
ponential factor. In the opposite limit 1 ~ 1, Keldysh derived an expression
for the multiphoton ionization rate (e = 2.72)
(1.16)
where A is the factor of order unity. The irradiance wavelength regions

over which electron tunneling and multiphoton ionization occur are shown
in Fig. 1.4.
10 Weyl
ELECTRON
TUNNELING
1'1oB
~
w
()
z
c(
15 MULTIPHOTON
~ IONIZATION
0:
1o12
1011~--~-L~~LU~--------~-L~
0.1 1 10
WAVELENGTH (IJ.)
Figure 1.4 Regions for ionization by tunnelling and by MPI.
When one uses Eq. (1.16) to compare with experimental results on MPI,
one finds that Keldysh's expression underestimates the measured rate by
many orders of magnitude. The intensities that are calculated to yield a
given level of ionization over pulse times of interest are roughly ten times
higher than the experiments indicate.
Another simple expression for MPI has been derived by Grey-Morgan
(1975). He solved a series of coupled rate equations for the virtual states
populated by the absorption of k (k = 1 tom -1) photons. These states
are assumed to decay spontaneously with a lifetime rk given by the uncer-
tainty principle rk = n(knw)- 1 and absorb photons with a cross section a,..
Assuming all a ,.'s to be the same (a,. = a), he finds
(1.17)
The choice a = 10-16 cm2 yields reasonable agreement with the data.
(A) Effect of Resonances

Near resonances with intermediate states will lead to enhanced MPI rates
and a change in the order of the process. The effect of resonances has been
studied extensively both theoretically and experimentally. For a good re-
view of the subject, see Mainfray and Manus (1984), Crance (1984), Lam-
bropoulos (1976), and Delane (1975). Under the strong laser fields where
MPI is observed, the energy levels are both broadened and shifted by the
Stark effect so that the position of the resonances will vary with irradiance.
If the energy of the resonant level is near khw, then for linearly polarized
light, conservation of angular momentum requires that the quantum num-
ber M of the resonant state differ from that of the ground state by the
amount k, k- 2, k- 4, etc. If, however, the radiation is circularly polar-
ized, then the states must differ by D.M = k. Thus, e.g., an excited p state
of an atom that is 5hw (k = 5) above a ground s state will influence the
MPI cross section for linearly but not circularly polarized light. Experi-
ments have been carried out that clearly show the effect of polarization.
(B) Beam Coherence Effects

MPI rates scale as (E 2r. The effect of MPI over a laser pulse is thus related
r
to the integral over the pulse ofE2m. Beam coherence will play an important
role since (E 2 is not the same as (E2m ). If the beam is perfectly coherent,
then (E2m) = (E 2 )m, where the averaging is carried out over many cycles
of the wave. In the opposite limit of a purely incoherent beam (composed of
photons of the same frequency but uncorrelated phases), one can show that
(E2m) = m!((E 2 )r. The MPI rates will thus be m! times larger than the
coherent beam rates. We must note at this point that the theoretical results
previously quoted are for coherent beams.
Q-switched laser pulses of nanosecond duration are usually composed
of many cavity modes, unless a particular mode has been selected by the
use of an etalon in the cavity. The larger the number of modes, the more
incoherent the beam becomes. The variation of MPI rates with the number
of cavity modes is shown in Fig 1.5 (LeCompte et al., 1974). Picosecond
pulses, on the other hand, should be considered coherent insofar as MPI
effects are considered, because there is a correlation of the phases of the
many longitudinal modes of the cavity leading to constructive interference
over the picosecond pulse lengths.
(C) Measurement of MPI Cross Sections and Rates

Although over the last few years, there have been many experimental pa-
pers dealing with MPI of gases, quantitative measurements of MPI rates
are rare. The early quantitative measurement have been carried out in the
Soviet Union (Voronov, 1967) and France (Agostini et al., 1968). Table I of
12 Weyl
(d) (c) (b) (a)
(j)
z
I- 1o2
::J
0
ai 0
a:
<t:
(f)
z
Q
10
1~--~~~~-U~----------~
1 10
AVERAGE LASER INTENSITY T
(ARB. UNITS}
Figure 1.5 Log-Log plot of the variation of the number of ions N; as a function of
the average laser irradiance i (arbitrary units), when the laser operates in (a) one
mode; (b) two modes with visibilityv = 0.6; (c) seven modes; and (d) seven modes
when these modes are locked, from LeCompte et al. (1974).
Grey-Morgan's (1975) review lists the MPI cross section of the rare gases
that have been measured at 1.06 and 0.53 J.Lm. Table VI of Delane's (1975)
review also lists experimental MPI cross sections for the alkali and for Hand
Hg at 1.06 and 0.53 J.Lm. Since these reviews, more work has been carried out
by the French (Normand and Morellec, 1980; eHuillier et al., 1983, 1984,
1985; Lompre et al., 1980, 1985; Baravian et al., 1982). A summary of their
results is shown in Table 1.1. The cross sections in column 4 correspond to
a coherent beam. Measurements are carried out at low pressures. A typical
experimental configuration is shown in Fig. 1.6 (Bavarian et al., 1982). The
Table 1.1 m-Photon Ionization Cross Sections am
,\ CTm Pulse
Gas m (nm) [(W/cm 2)-msec- 1] length References
Cs 2 528 (4.7±1.3) X 10- 13 19 nsec Normand and Morellec (1980)

He* 2S1 2 347 (8.1±0.6) X 10- 13 12 nsec Lompre eta!. (1980)
Hc* 2S3 3 694 (1.2 ± 0.7) X 10- 25 7 nsec Lompre eta!. (1980)
Cs 4 1056 (5.8±2.2) X 10- 34 25 nsec Normand and Morellec (1980)
Xe 6 532 2.1 X 10- 61 50 psec .LHuillier eta!. (1983)
Ne 10 532 2 X 10- 119 50 psec .LHuillier eta!. (1983)
He 11 532 2 X 10-130 50 psec Lompre et a!. (1985)
02 6 532 4 X 10-66 40 psec .LHuillier eta!. (1984)
N2 7 532 10-78 40 psec .LHuillier eta!. (1984)
N2 12 1056 8.7 X 10-151 30 nsec Baravian eta!. (1983)
number of electrons generated is measured by a collector. The pressure is

maintained at a low enough value so that recombination is negligible and
all ions and/or electron generated are collected. The number of MPI events
readily yields the MPI cross section if the temporal and spatial characteris-
tics of the beam are known. As an example, consider a diffraction-limited
laser beam propagating in the Z direction that has a Gaussian radial profile.
The intensity in the focal region is given by (Marcuse, 1972)
1 = 2P(t) exp
7l"W2
(-2r
w2
2)
(1.18)
where
(1.19)
In the above, Z 0 is the location of the focus, wF is the minimum beam radius,
and P(t) is the laser power. The minimum beam radius at focus is related to
the beam radius w 0 at the exit plane of the optics by the diffraction-limited
formula Wp = (.AZ 0 /1rw 0 ). The number of electrons generated in the focal
14 Weyl
f
CHARGE SEPARATION AND ACCELERATING SPACE
CHARGED PARTICLE MULTIPLIER
----------- -------~
CURRENT PROPORTIONAL
:~ 1
TO THE COLLECTED IONS
lC::.:-.::.:..-~..,+·t - r - - -
COrTOR
ELECTRON
----- I
- - - - - - - - .J
~-----
POWER SUPPLY
r_ ___ _ +
Nz
1-------1~._ LASER PULSE
PHOTOCELL
Figure 1.6 Experimental configuration for MPI measurements, from Bavarian

et al. (1982).
volume of a gas of density n is therefore given by
(1.20)
or, in terms of the peak irradiance at focus (r = O,Z = Z 0 ),
(1.21)
where IP = 2P I( 1rw}) and V is the effective focal volume for MPI
V = ?rWft (2m -5)!!

(1.22)
.A 2m-lm(m-2)1
Equation (1.20) or (1.21) can then be inverted to yield the MPI rate coeffi-
cient Wm from the measured valued of Ne.
Figure 1.7 shows the number of ions measured by MPI of 0 2 at 0.53

and 1.06 f.lm (I..:Huillier et al., 1984). The slope of the line log(Ot) vs. I
at 0.53 f.lm is 6 as expected from reaction (1.2) with m = 6. The change in
slope occurring at! = 5.5 x 1012 W/cm 2 corresponds to the saturation of ion-
ization. It corresponds to the irradiance above which complete ionization
on axis will have been reached before the end of the pulse.
(D) Absorption of More Than m Photons

and Multiple Ionization by MPI
The object of recent work in the field of multiphoton ionization has been
to study what happens at above-threshold irradiances. Measurements of
the free-electron energy distribution show that, as the laser irradiance is
105
+
.,oz
1o4 j(··~o·
• ! o8
8
(j)
z
1- I I:
:;) 103
I.& 0
o:i
cr: bo
~
(/)
z
1Q2 1o2
/'
Q I ~~
I
t~ I
Cb •
10 10
~ I
~
I I
1
1012 1Ql3 1o13 1Q14
(a) (b)
LASER IRRADIANCE (W/cm-2)
Figure 1. 7 A log-log plot of the variation of the number of ot

and 0 + ions formed
as a function of the laser irradiance at a) 0.53 f.lm and b) 1.06 f.lm. The vertical dashed
line indicates the saturation intensity Is· A sequential process in 0 + ion production
appears at intensities higher than that shown by the vertical dot-dashed line, from
~Huillier et al. (1984).
16 Weyl
increased, the probability of the absorption of more than m photons in-

creases. One observes peaks in the electron distribution centered at Eo +lhv
(1 = 0, 1, ...) where Eo is a constant. The distribution shifts toward larger 1
as the irradiance is increased. Also, as the irradiance is increased above
threshold, doubly and then triply ionized species are formed (Luk et al.,
1983; I..:Huillier et al., 1983). There is, at present, a controversy as to
whether the multiply ionized species are formed directly by simultaneous
ejection of two or more electrons (Boyer and Rhodes, 1985) or by MPI of
the previous ion (Geltman, 1985; Mu et al., 1986). At sufficiently high fields,
simultaneous ejection of several electrons out of their orbitals should be fa-
vored as compared to successive one-electron events. Lambropoulos (1985)
has argued, however, that in all experiments performed to date, there was
a sufficiently slow rise time of the laser pulse that substantial ionization
should take place before the regime of simultaneous ejection of electrons
is reached.
1.3.3 Electron Cascade Growth
Electrons in a laser field will gain energy through electron-neutral inverse

bremsstrahlung collisions (IB). This is the reverse of the well-known brems-
strahlung process in which high-energy electrons, upon traversing a gas or
solid, emit radiation as they slow down (from the German words "brem-
sen," to slow down, and "strahlung," radiation). Electrons will lose energy
by elastic and inelastic collisions with neutral particles through excitation of
rotational and vibrational degrees of freedom of molecules and excitation
of electronic states of molecules and atoms. Some electrons will be lost by
attachment, but new electrons will be generated by ionizing collisions. If
the laser irradiance is high enough, then a few electrons, in spite of all the
energy loss processes, will gain an energy larger than the ionization energy.
These electrons will generate new electrons by the impact ionization of the
gas, thereby leading to cascade growth.
The growth of the electron density with time can be theoretically ana-
lyzed by solving the Boltzmann equation for the electron distribution func-
tion in velocity space. The distribution function F is split up into an isotropic
part F 0 (E) and an anisotropic part F 1(v) such that JvF1 (v)d 3v is collinear
with the applied E field, E and v being the electron energy and velocity, re-
spectively. At not too large fields, F 1 <t::.F0 and the expansion ofF, keeping
only the first two Legendre polynomials, is valid. When the photon energy
hv is much less than the average electron energy €, absorption can be treated
as a continuous process. The equation satisfied by the distribution function
[if we neglect elastic losses that are small as compared to inelastic losses
(Kroll and Watson, 1972)] is
aF0 _ aE
8t- (2
a 3€312 Q aE F0 )
a El/2 +n:Ej
j (E,e')FoE(
Gj
')d E, (1.23)
where Q is the rate of energy absorbed per electron from the laser field, and
Gj a function that represents the net gain of electrons of energy E due to the
inelastic process j with cross section O"j
(1.24)
The first term in Eq. (1.24) corresponds to electrons generated at energy e

from electrons at energy E1 > E when an excited state of energy Ej is formed.
The second term corresponds to the loss of electrons of energy E when the
excited state is formed. Cross sections for excitation and ionization of many
atoms have been calculated theoretically from first principles. In the case of
molecules, however, one must rely on experiments. The cross sections can
be derived indirectly from electron swarm experiments (Frost and Phelps,
1962) or directly from crossed-beam experiments performed at very low
pressures where the distribution of inelastically scattered electrons is mea-
sured as a function of electron energy. A compilation of cross sections for
many gases can be found in a report by Kieffer (1973).
The heating rate Q can be calculated from the equation of motion of the
electrons under the applied field E = Re (E0 e-i"'1)
dV
m-=-eE-mvV
edt e m (1.25)
where vm is the momentum-transfer collision frequency of electrons with

neutrals. The second term on the right-hand side of Eq. (1.25) represents
the drag on electrons due to collisions. Equation (1.25) is the same equation
that is used in Drude theory for electrical conduction in metals. Solving
Eq. (1.25) for V, one gets
V = Re [ eEoe-iwt ] (1.26)
me(iw-vm)
The power gained from the field is obtained by taking the time average of
(-eE·V)
(1.27)
18 Weyl
We can estimate vm by using the gas kinetic cross section that for most gases
of interest is near 10- 15 cm2 • Under breakdown conditions near threshold,
the average electron energy is one-third to one-fourth the ionization energy
of the gas, i.e., f.~ 3 to 6 e V. We therefore have
where p is the pressure in atmospheres.

At 10.6 pm and shorter wavelengths (w > 2 x 1014 sec- 1 ), we will have
v! «: w 2 unless the gas pressure exceeds hundreds of atmospheres. We then
see from Eq. (1.27) that the heating rate will scale as vmlw 2 , i.e., as n>.. 2 •
When hv ;::: f., the assumption of continuous energy absorption is no
more valid and the absorption of laser photons must be treated quantum-
mechanically. Zel'dovich and Raizer (1965) in their pioneering paper de-
rived a quantum kinetic version of the Boltzmann equation that takes into
account absorption of energy in discrete quanta. The quantum kinetic terms
that replace the first term on the right-hand side of Eq. (1.23) can be writ-
ten as
n[K4 ( e -hv)F0 ( e -hv) -Ke( e)F0 ( e)

-K4 (e)F0 (e) +Ke(e+hv)F0 (e +hv)] (1.28)
where K 4 and Ke are the IB absorption and stimulated emission coefficients

per unit electron and neutral density. K 4 and Ke have units of cm5 • The
four terms are shown diagramatically in Fig. 1.8. K 4 and Ke are related
though detailed balancing, by considering a Maxwell-Boltzmann distribu-
tion of electrons at temperature T that is in equilibrium with a blackbody
source at T. One finds
(1.29)
when e > hv and Ke( e) = 0 when e ~ hv. If one has a Maxwell-Boltzmann

distribution (FB) of electrons at T, then the net absorption coefficient per
unit electron and neutral density averaged over the distribution function is
K= 1 00
[K4 (e)-Ke(e)]FB(e)de
= (1-e-h"/kBT) 1oo Ka(e)FB(e)de (1.30)

e ELECTRON
ENERGY
hu hu
hu
E+h\l
0
Figure 1.8 Absorption and emission channels coupling to energy level t.
Geltman (1973) has calculated the IB absorption coefficient Ka of ar-

gon and other monatomic gases (C, N, 0, ...) in the wavelength range
10 > >. > 0.5 p,m. His quantum mechanical calculations were obtained us-
ing a model atomic potential that was adjusted to fit experimental scat-
tering cross sections. The results that he presents, however, were aver-
aged over a Maxwellian distribution of electrons. Another less exact treat-
ment (Dalgarno and Lane, 1966) relates the IB absorption coefficient to the
momentum-transfer cross section um through an expression of the form
2 112 [e+hv
K a = 347re 2 [2(e+hv)] () _:_ ( h )]
h v 17m e + h v 17m e + v (1.31)
mew m
=6.0x10- 15 >. 2 (e+hvi12 x [(1+ heJum(e)
+ hev um(e +hv)] (1.32)
where in the last step we expressed e and hv in e V, >. in p,m, and Ka

in cm-5 • Equation (1.31) differs from the expression given by Zel'dovich
and Raizer (1%5) and by Kroll and Watson (1972) where the last bracket
is replaced by [2(e +hv)/hv]um(e +hv), differs from an expression derived
by Kas'yanov and Starostin (1965) where the last bracket is replaced by
[(2e+hv)/hv]um(e+hv), and differs from an expression used by Phelps
(1966) where the last bracket is given by [(2e + hv)/hv]um(e + hv/2). Note
that the expressions of Kas'yanov and Starostin (1965), Phelps (1966), and
Dalgarno and Lane (1966) are equivalent when um is independent of e and
20 Weyl
that all expressions go to the same limit when hv «:e. A check of the validity
of the various expressions is best performed by comparing them with data
for a gas, such as argon, where am is strongly dependent on e. Batenin and
Chinov (1972) measured an absorption cross section Ka = 2 x 10-40 ems
in argon at .>.. = 0.454 J.tffi when Te was in the range 0.4 to 0.5 eV. Set-
tinge= 3/2Te = 0.7 eV, one finds Ka = 3.6 x 10-41 , 1.7 x 10-40, 1 x 10-40,
and 5 x 10- 41 ems from the expressions of Dalgarno and Lane, Zel'dovich
and Raizer, Kas'yonov and Starobin, and Phelps, respectively. These cross
sections are all lower than the data indicate, with the cross sections of
Zel'dovich and Raizer (1965) and Kas'ynov and Starobin (1965) being in
better agreement than the other two. More recently, Bremsstrahlung ra-
diation in a drift tube filled with argon was measured at .>.. = 0.5, 0.65, and
1.24 J.tffi (Yamabe et al., 1984). Agreement between the free-free emission
and predictions using Kas'yanov and Starobin's (1965) formula was found
to be very close.
Numerical solutions of the Boltzmann equation have been extensively
used to model the excitation processes in gaseous mixtures in TEA or e-
beam sustained lasers (Nighan, 1970; Lowke et al., 1973). Ionization rates
in air at 10.6 ~-tm (Canavan et al., 1972; Kroll and Watson, 1972; Weyl, 1979)
and in other gases (Phelps, 1966; Weyl, 1979; Weyland Rosen, 1985; Rosen
and Weyl, 1987) have been calculated as a function of laser irradiance. We
show in Fig. 1.9 the computed ionization rate vi in N 2 , H 2 , He, and air at
1 atm pressure as a function of laser irradiance. Near threshold, when vi~
0, practically all the laser power goes into excitation of the gas, with very
little left for ionization. One can define an effective ionization energy Eeff
as the ratio of the power absorbed by a free electron to the ionization rate
e2E 2 v 1
E = m x- (1.33)
eff 2me (w 2 + vm2 ) V·1
eeff is plotted vs. I for various gases in Fig. 1.10. It is seen to decrease from oo
at threshold to a near constant value eeff!::::! 2 to 5e1 at very high fluxes. Many
naive treatments of laser-induced breakdown can be found in the literature
that completely neglect inelastic losses and use Eq. (1.33) with Eeff = e1 in
order to calculate ionization rates. These will be greatly in error especially
near threshold for laser pulses of long duration (rP- 10- 7 to 10-s sec).
Numerical solutions of the quantum kinetic equation have been used
to calculate cascade breakdown at 1.06 J.tffi and shorter wavelengths (Kroll
and Watson, 1972; Weyl and Rosen, 1985; Rosen and Weyl, 1987). At high
enough irradiances, electron impact excitation of states lying within hv or
even 2hv of the ionization continuum will lead to ionization since these
states can readily absorb one or two photons and become ionized (Raizer,
1965). Using a conservative value of 10-18 cm2 for the photoionization cross
'tbr(sec)
10-9
N'
E 1010
0
~
w
0
z<(
0<(
0:
0: 109
108 L-~--~~LL--~~~-L~-L--L-~~--~~~~
107 108 109
Uav(s-1)
Figure 1.9 Ionization rates in H 2 , N2 , and He gases: p = 1 atm, >. = 10.6 J.Lm. The
upper scale is the breakdown time assuming 30 cascades to breakdown, from Weyl
(1979).
section, one estimates a photoionization rate v1
( -I ) (W/cm ) sec- 1
2
v
1
=a hv
= 6I hv (eV)
(1.34)
Thus, at I = 1011 W/cm 2 and hv = 1 eV, one calculates v1 = 6 x 1011 sec-1,

corresponding to a lifetime for the excited state of "' 10-12 sec. Similarly,
for two-photon absorption from the excited state into the continuum, one
estimates as ionization rate (Weyland Rosen, 1985)
(1.35)
where j refers to the excited state, m to a state lying above j that can be
ionized by absorption of a single photon and for which the transitionj --t m
(with/ numberfm) is near resonant. The cross section a in Eq. (1.35) is the
22 Weyl
103
>
~
u.
u.
ww
102
1oto 1012
IRRADIANCE (W/cm2)
Figure 1.10 Effective ionization energy of electrons in air.
cross section for photoionization of state m. Setting u = 10-18 cm2 , fmi =

0.1, hv = 1 eV, hv- Ejm = 0.1 eV, I = 1011 W/cm2 , we find from Eq. (1.35)
v2 = 5 x 1010 sec- 1 • Photoionization by excited states will be important when
v 1rP ~ 1 and v 2rP ~ 1. For long pulse lengths near threshold, one must also
add the condition v 1 1 R > 1 where rR is the lifetime for radiative decay of
the excited state j.
We show in Fig. 1.11 the calculated cascade ionization rates in molecular
nitrogen at 1 atm pressure for a laser wavelength of 0.35 f..L, including only di-
rect impact ionization (vi), impact ionization plus excitation of states within
hv of the ionization edge (v; + v**), or impact ionization plus excitation of
states lying within 2hv of the ionization edge (vi+ v* + v*"'). One sees from
this figure that at a laser irradiance, I= 1012 W/cm2 (vi+ v"' + v**) is more
than 100 times larger than v;.
1.3.4 Attachment, Recombination,

and Diffusion Losses
Many molecular gases (H2 , 0 2 , H 2 0, F 2 , SF6 , •••)can attach electrons. At-

tachment, which is an electron loss mechanism, can occur through two main
mechanisms:
tpxP(atm) (s)
1o-6 10-7 · 10-8 10-9
1o14r---~-----------r-----------r----------,-----,
N'
E 1013
~w
0
z<(
0
<(
cr: 1012
cr:
1011L-~~----~----------~--------~----------~
107 108 109 1010 1011
ui /P (s-1 atm-1)
Figure 1.11 Cascade ionization rates in N2 . The upper scale gives breakdown time
assuming 30 cascades to breakdown, from Weyland Rosen (1985).
Three-body attachment
(1.36)
IWo-body dissociative attachment
(1.37)
The cross section for reaction (1.36) decreases with increasing electron en-
ergy. Reaction (1.37), on the other hand, usually requires a threshold elec-
tron energy, corresponding to the difference between the dissociation en-
ergy ofAB and the attachment energy toA. Both processes occur in 0 2 and
play an important role in fixing its long-pulse breakdown threshold in air
(Kroll and Watson, 1972; Weyl, 1979).
Electron-ion recombination is important in determining the develop-
ment of ionization during the late stages of avalanche buildup, when the
electron density is high. Similarly to reactions (1.36) and (1.37), we will
24 Weyl
have three-body recombination and two-body dissociative recombination
e- +AB+ +X ---+AB +X (1.38)

e- +AB+ ---+A +B (1.39)
Reaction (1.38) has been studied theoretically by Gurevich and Pitaevskii

(1962) when the third body X is an electron, and they obtain a rate
n2
vR = 8. 8 x 10- 27 - e (sec- 1 )
T3.5 (1.40)
e
where Te is the electron temperature in e V and ne the electron density

in cm- 3 .
Diffusion of electrons out of the focal volume may also be an important
loss mechanism for small laser beams. The combined effect of diffusion and
cascade ionization is usually treated in the literature (Grey-Morgan, 1975)
by assuming a top-hat radial irradiance profile and imposing a sink of elec-
trons at the edge. The electron density is then found to grow exponentially
with a reduced rate
2.408D
vAv=v;= a2 (1.41)
where D is the electron diffusion coefficient and a the radius of the beam.
At low electron densities, the electrons diffuse freely and the diffusion co-
efficient, from kinetic theory of gases, is given by
lv v
D=-=- (1.42)
3 3na5
where l, v, and a5 are the mean free path, velocity, and scattering cross sec-
tion, respectively, of electrons and n is the gas density. A rough estimate
of D can be obtained by setting v = (2€/me) 112, € = E1 /3, and using the gas
kinetic cross section for a 5 • We show in Table 1.2 the diffusion coefficient
averaged over the actual electron distribution in several gases, as obtained
from a Boltzmann code under conditions where the cascade ionization rate
is lOS sec- 1 and n = 2.5 x 1019 cm- 3 .
At high electron densities, diffusion is affected by the space charge of ions
remaining in the focal region. If the space charge is large enough, electrons
cannot diffuse without dragging the ions along with them. Diffusion be-
comes ambipolar and the diffusion coefficient is reduced by the square root
of the electron to an ion mass ratio ("' 10-2 ). The transition from free to
ambipolar diffusion occurs when the Debye length .Av = [kTel(41l'nee 2 )] 112
Table 1.2 Free

electron Diffusion
Coefficients at
l amagat Density for
v.I = 108 sec- 1
D
Gas (cm 2/sec)
N2 850
H20 1200
Air 1200
H2 1150
He 5500
Ar 8300
becomes smaller than a. This occurs at a density ne ;::kTel( 47re 2a 2) or
3 l3 Te(eV)
ne(cm- )>5.5x10 [a(p;m)]2 (1.43)
When diffusion losses are so dominant that vAV ~ vi, then the irradiance
at focus will be significantly higher than the threshold irradiance for infi-
nite beam radius. One may question under such conditions the validity of
Eq. (1.41) since laser beams do not have a top hat radial profile and the spa-
tial volume where cascade growth takes place increases with increasing I.
Equation (1.41) is expected to give an upper bound for the diffusion losses.
Another treatment (Rosen and Weyl, 1987) that uses, in order to calculate
cascade growth, the average field seen by an electron as it diffuses away from
the axis during one cascade time gives a lower bound to diffusion losses.
1.3.5 Calculation of Breakdown Thresholds
Breakdown thresholds can be calculated, in principle, by integrating the

master Eq. (1.12) over time and requiring that by the end of the pulse, a
certain degree of ionization of the gas be reached. This is a formidable task
since vi andD need to be derived by solving the Boltzmann Equation (1.23),
which requires a knowledge of many inelastic cross sections. Also, mul-
tiphoton ionization rates have only been measured for a few gases at se-
lected wavelengths. The dependence of breakdown thresholds on wave-
length, pressure, and pulse length, however, can be deduced by looking at
how the various terms in the master equation depend on these parameters.
Such an approach will be useful in predicting breakdown thresholds under a
26 Weyl
given set of conditions based on breakdown measurements under different

conditions.
We will be considering, in order to be definite, a temporal pulse shape
that is Gaussian in time
(1.44)
Such a pulse shape is a good approximation to a Q-switched pulse of a laser

operating in a single longitudinal mode. The pulse full width at half max is
and the average pulse irradiance is related to pulse ftuence ¢>p by
(1.45)
(A) MPI-Dominated Breakdown

At short wavelengths, or at low densities (p ~ 10 Torr in N 2 ) where colli-
sional effects are small, breakdown is dominated by multiphoton ionization,
represented by the second term on the right-hand side of Eq. (1.12). Diffu-
sion losses play no role since electrons are generated by MPI of the neutrals
that remain in the focal volume. The number of electrons generated per
unit volume at the end of the pulse is
(1.46)
where, for the pulse profile given by Eq. (1.44), g(m) = [2(1n2hr) 1f2]m x
[7r/(4mln2)]112 • The factor g(m) decreases from 0.51 form= 3 to 0.13 for
m = 13. One sees from Eq. (1.46) that the irradiance i required to reach a
given degree of ionization! = (neln) by the end of the pulse is independent
of gas density n (or pressure p ). A diagnostic commonly used to determine
breakdown at low pressures, however, is the observation of a visible glow
(Kracyuk and Pashinin, 1972). This requires the formation of a minimum
number Ne of electron-ion pairs in the focal volume. One can estimate this
number by assuming that every recombination event results in the emission
of a visible photon. A dark adjusted eye will be able to "see" a faint glow
when approximately 10 photons hit the retina. If R is the radius of the pupil
and r the distance of the observer's eye from the laser beam focus, then the
number of photons measured is
Setting, e.g., Nph = 10, R = 0.3 em, and r = 102 em, we estimate that a mea-
sured threshold would correspond toNe~ 106 • Since Ne is proportional to
the gas density (or pressure), the observed threshold i will scale as n lim (or
pllm).
(B) Inverse-Bremsstrahlung-Dominated Breakdown

At high pressures (p;::; 100 Torr in N2 ) and long wavelengths (-\ > 1 ttm),
breakdown is usually associated with cascade ionization and the first term
on the right-hand side of Eq. (1.12) dominates. As we did for MPI, we can
define an equivalent square pulse of irradiance i, choosing i as the average
irradiance in the pulse. Pulses will be equivalent when the final develop-
ment of electron density is the same.
Integration of Eq. (1.12) yields
(1.47)
where vAV is given by Eq. (1.41) and nef and nei are the final and initial
electron densities, respectively. For a square pulse of length rP' we have
ln(ne11nei) = vAv(/)rp. For the Gaussian pulse shape given in Eq. (1.44), we
can readily integrate Eq. (1.13) by using a power law approximation to vAV•
v Av(l) = v Av(f)(IJfr'
(1.48)
where g(m') is the same pulse shortening factor as defined in Eq. (1.46).
Equivalently, one can define a square pulse of length rp and irradiance i x
g(m') 11m', which would give the same development in electron density. For
m' = 1, the irradiance would be reduced by the factor 1.22, whereas for
m' = 4, it would be reduced by 1.24.
When diffusion losses are negligible (i.e., in the case oflarge beams), the
power m' is obtained from a solution of the Boltzmann or quantum kinetic
equation. Figure (1.11) shows such a solution for N2 at,\= 0.35 ttm, where
m' is the slope of the curve ln[(vi + v* + v*)] vs. I. The exponent m' is al-
ways larger than unity; it is near 1 for the rare gases and approaches 1 as the
irradiance is increased way above threshold. Many investigators use as cri-
terion for breakdown the condition ln(ne11nei) ~ 40 that would correspond
28 Weyl
to an electron density of 1017 cm- 3 by the end of the pulse if one started
out with nei = 1 cm- 3 .
Equation (1.48) then becomes
(1.49)
where / 0 is some reference irradiance. When w > vm, the electron heating
rate [see Eqs. (1.27) and (1.28)] is proportional to gas density. All terms in
the Boltzmann equation will then scale proportionately with n (or p ), and
we derive from Eq. (1.49) the scaling of breakdown with pressure and pulse
length
(1.50)
When diffusion losses are not negligible, the above scaling changes since
vAVis given by Eq. (1.41). If the diffusion losses are so large that vi~ vAV•
then the breakdown criteria, to a good approximation, become vi~ 2.4D/a 2 •
The diffusion coefficient scales asp-1, whereas vi scales asp. We thus obtain
the following scaling of irradiance with p:
j o:.p-2/m' (1.51)
and, to a first approximation, i will be independent of rr If multiphoton

ionization is capable of producing enough electrons in the beginning of the
pulse so that Eq. (1.43) is satisfied, then diffusion is ambipolar; otherwise,
Dis given by the free-electron diffusion expression, Eq. (1.42).
The number of cascades required for breakdown is expected to be much
less than 40 for several reasons. First, MPI at wavelengths less than 1 p,m
will initially generate a number density of electrons ne much larger than
1 cm- 3 • Second, when ne exceeds 1014 cm- 3 , several new physical pro-
cesses come into play that significantly shorten the cascade time: (1) the
electron-electron collision frequency vee becomes larger than the electron-
neutral inelastic collisional frequency v*, resulting in a more Maxwellian
shape for the distribution function, i.e., more electrons are formed in the
high-energy tail of the distribution function, leading to a larger ionization
rate; (2) the population of excited states increases dramatically, with impact
ionization of these excited states becoming an important ionization process
(Zel'dovich and Raizer, 1965); and (3) electron-ion recombination through
reactions (1.38) and (1.39) often results in the formation of excited states
that can rapidly regenerate the lost electrons by photoionization. Very little
work has been carried out to model the late time stage of the laser-induced
cascade process. This stage is, however, very important, since it is at the end
of the cascade that practically all the energy absorbed by the gas takes place.
A proper treatment would involve introducing in the Boltzmann equation
inelastic and superelastic collision terms with excited states of neutrals (a
superelastic collision involves the transfer of excitation energy of the neutral
to kinetic energy of the electron and is the reverse of an inelastic collision)
and simultaneously solving for the growth of excited state populations.
We show in Fig. 1.12(a) the late time growths of ne and the first excited
state Ar* of argon in the late time breakdown stage of argon (Weyl and
Rosen, 1985). The analysis leading to Fig. 1.12 was greatly simplified by
assuming that vee ~ v* so that a Maxwellian distribution could be main-
tained for electrons and a two-temperature model (electron temperature,
ion temperature) of the partially ionized gas would be valid. Figure 1.12(b)
shows the absorption coefficients due to the various absorption processes.
The most important laser absorption processes just before breakdown are
seen to be ionization of excited states formed by three-body recombination
[process (3)] and photoionization of highly excited states formed by elec-
tron impact excitation from the 4s states (Ar*) of Ar at 11.6 eV [process
(5)]. Figure 1.13 shows the effect of electron-electron collisions on the dis-
tribution function. The features in the distribution function that reflect the
energy thresholds for inelastic processes are seen to be smeared out when
electron-electron collisions are taken into account.
(C) Wavelength Scaling

Wavelength scaling of breakdown thresholds can be derived by looking
at how the electron-neutral inverse bremsstrahlung absorption rate varies
with w. When hv ~ €, Eq. (1.27) can be used and yields the scaling I ex
[w 2 + v~)lvm1· Under steady cascade growth, we can write in Eq. (1.23) that
8F0!8t = viF0 • When hv < €, Eq. (1.27) can be used for Q and we see from
Eq. (1.23) that we have the scaling
(1.52)
where ¢ 1 is a function that depends on the particular gas considered. Since

vm ex p, one expects that each gas will yield a universal curve when vJp is
plotted vs. I /[p 2 (1 + w 2 /v~)]. Figure 1.14, taken from Yablonovich (1973),
demonstrates the validity of this scaling in He and N2 gases for wavelength
longer than 10.6 J.lm.
When w ~ vm, I is expected to increase with decreasing wavelength as
>. - 2 • Buscher, et al. (1967), however, observe a maximum in the threshold
of the rare gases in the wavelength region 0.53 < >. < 1.06. Their results are
not surprising since as>. is shortened below 1 J.lm, photoionization of excited
states and direct MPI of the gas will become increasingly important. The
30 Weyl
10 0
1=1011 3X1o10
10-1
1010
10-2 I
I
z I
0 I
~
a:
I
I- 10-3 I
z I
w
() I
z I
0 I
()
10-4 I
I
10-5
2 3 4 5 6 7 8
TIMEx (1/1010) (1o-7s)
(a)
Figure 1.12 Late time breakdown stage in Argon, I = 1010 W/cm 2 , p = 1 atm.
(a) Growth of electron concentration (-) and first excited 4S state (--). (b) Absorp-
tion mechanisms. 1 photoionization of argon dimer; 2 electron neutral IB; 3 pho-
toionization of excited states formed by three-body recombination; 4 two-photon
ionization of Ar* (4S) states; 5 single photon photo-ionization of argon highly ex-
cited states formed by electron-Ar* collisions; and 6 electron ion IB, from Weyland
Rosen (1985).
breakdown thresholds in N2 and 0 2 for 1 <p < 27 atm were measured to be

five times lower at >. = 1.06 JJm (rP = 10 nsec) than the calculated thresholds
using wavelength scaling of 10.6 JJm results (Stricker and Parker, 1982).
(D) Laser Ionization Based on

Resonance Saturation (LIBORS)
If the laser wavelength is tuned to a resonance line connecting the ground
state of an alkali vapor to an excited state, ionization may result at remark-
®
10-1
I
E
(.)
~
10-2
10-3
10-4
10-SL---------~~~~----~------L---~
4 5 6
TIME (1o-7s)
(b)
ably low irradiances (Lucatorto and Mcilrath, 1976). The breakdown mech-
anism called laser ionization based on resonance saturation or LIBORS
(Measures and Cardinal, 1980) is so effective because one is using the very
large resonant absorption cross section of the atoms to channel the laser
energy via superelastic collisions into the electrons. The explanation of the
process is as follows. The laser beam when tuned to a strong resonant tran-
sition will generate a population of excited states, which is equal (at satu-
ration) to the population of ground states so that a free electron has a 50%
chance of colliding with an excited state. A superelastic collision will make
an electron acquire the resonant energy (hv). If the same electron on its
next collision encounters another excited atom (with a 50% probability), it
can then ionize the atom. LIBORS has been used to break down the alkali
vapors (density ,..._ 1015 to 1016 cm- 3 ) at very low irradiances (106 W/cm 2 )
over short times ("" 100 nsec). The method, however, is not applicable to
32 Weyl
10-1
~WITH e-e COLLISIONS
10-2
1 10-3
2 4 6 8 10 12 14 16 18
ENERGY (eV)
Figure 1.13 Effect of electron-electron collisions on the distribution function of

electrons.
molecular gases because the energy losses to excitation of vibrational and

rotational degrees of freedom are so large.
1.3.6 Recent Measurements of Breakdown in Gases
There are several good reviews of experimental breakdown data in gases

previous to 1976 (Raizer, 1977; Smith and Meyerand, 1976; Grey-Morgan,
1975). We discuss some of the more recent data.
Breakdown measurements in unfiltered atmospheric air using picosec-
ond pulses have been carried out by Williams et al. (1985) at A = 0.53 t-tm
(80 < rP < 140 psec) and by Van Stryland et al. (1981) at A= 1.06 t-tm. The
measurements at 0.53 t-tm showed a strong dependence on pulse length, the
1Qll
•
I
I
E I
"@ .;
0
Q)
1Ql0 I
!!2 I
z J
I
LlJ
I- Nz •
<(
I
•••
[£
z 109
0
i=
<(
4
!::::!
z I
,'
Q
a. 108 I
"-;...
••
;:>
107
103 104 105
EFFECTIVE FIELD Erms/p ~ 1 +<ilium2
(volts I em atm)
Figure 1.14 Ionization rate v;fp as a function of the effective electric field Ermsl
[p(l + w 2/v~) 1 12 ] for helium and nitrogen. The solid line and the dots are DC data.
The crosses are C0 2 laser breakdown data, from Yablonovich (1973).
threshold irradiance / 1h, varying as rP-l [m' = 1 in Eq. (1.50)], indicative of a

cascade-type breakdown with no dependence on focal spot diameter, even
though the diameter was very small (3.4 < 2a < 7.2 pm). At rP =50 psec, the
threshold rms electric field for rP =50 psec was E B = 108 V/cm, correspond-
ing to / 1h = 2 x 1014 W/cm 2 • The lack of focal spot size dependence of / 1h is
surprising, even in the absence of diffusion losses, since the smaller focal ra-
dius a = 1. 7 pm is of the order of three electron mean free paths. Dewhurst
(1978) measured breakdown in N 2 and 0 2 at 1.06, 0.69, and 0.53 pm. We
reproduce in Fig. 1.15 his results for N 2 and also show the results of other
investigators. The low-pressure data at >. = 1.06 pm and A= 0.69 pm show
a weak dependence on rp, indicative of MPI-dominated breakdown. The
steeper slope at A = 0.53 pm is indicative of cascade-type breakdown. If
one takes into account the pulse length dependence of / 1h, one finds that
the results of Dewhurst and Williams et al. (1985) at 0.53 pm agree to within
the accuracy of the data. The fact that breakdown at 0.53 pm is of the cas-
34 Weyl
1Q15
------~=q. --:.:..e ...- ---------------------

L . . .t I ! of-
'·-.... t=18ps
.........
............
t=50ps
1Ql4
1o3 1o4
PRESSURE (Torr)
Figure 1.15 Breakdown threshold intensity variation with pressure in nitrogen at

(a) 1.06 Jlm; (b) 0.69 Jlffi; (c) 0.53 Jlffi. 1 = lline (see text), from Dewhurst (1978).
cade type is surprising in light of the different behavior at the longer wave-
lengths. One would expect MPI effects to be more important at 0.5 11m
than 111m. If we use the scaling / 1h ex:>. - 2 rp-rn' with m' = 1 to predict thresh-
olds at >. = 1 11m from the measured thresholds at 0.53 11m, we calculate
at p = 1000 Torr, / 1h = 4 x 1014 W/cm2 , a factor of 2 lower than the ob-
served threshold of 8 x 1014 W/cm2 . This indicates that photoionization of
excited states may play a role in increasing the cascade breakdown rates at
0.53 11m. Dewhurst (1978) noted in his paper that MPI behavior, when ob-
served, always occurs when the Keldysh parameter 1 [see Eq. (1.15)] is less
than or equal to 1, i.e., in the regime where ionization is expected to occur
by electron tunneling rather than MPI. Baravian, et al. (1982) have mea-
sured the MPI cross section of N2 at 1.06 11m and found it to be 8.7 x 10- 151
(W/cm 2 )- 12 sec- 1 (m = 12). We calculate, using this cross section, complete
ionization by the end of the pulse ( rp = 10 psec) when I;:::: 2.6 x 1013 W/cm2 ,
which is 30 times lower than the observed threshold at p ~ 1000 Torr.
The breakdown thresholds of the rare gases at 0.69 11m and rP = 18 ±
4 psec have been measured vs. gas pressure by Dewhurst (1977). His re-
sults are shown in Fig. 1.16 and are indicative of a cascade-type breakdown
with 1.1 < m' < 1.6. The results do not agree with the previous measure-
ments of Krasyuk et al. (1970) (30 < rP < 100 psec) where a weak pressure
dependence for p < 5000 Torr was observed, indicative of MPI-dominated
breakdown at the lower pressures.
104
1oi5 N'
E
z Kr ~
3 ~
0 Xe UJ
ocn (.)
~0
<(_J 103
z<(
UJo
a: I
0<(
cacn a:
UlUJ Q:;
?:a: 0
I-I
<(I- UJ
_J a:
UJ
a:
1o2 a:
UJ
u...
~
10 103
PRESSURE (Torr)
Figure 1.16 Relative breakdown thresholds in helium, neon, argon, krypton, and
xenon. The gradients of the lines are: 0.90, 0.78, 0.81, 0.61, and 0.63, respectively,
from Dewhurst (1977).
36 Weyl
The wavelength dependence of breakdown thresholds in He and argon

was studied using an 18 nsec dye laser that could be tuned in the 720 and
890 nm band (Byron and Pert, 1979). Although the argon data showed
the expected A- 2 dependency of / 1h, the Helium data was wavelength-
independent. Theoretical analysis of the data indicates that multipho-
ton ionization of excited states plays an important role in determining the
threshold irradiance (Gamal and Harith, 1983). The first excited state of
He needs only about three photons if A< 770 nm, whereas four photons are
required for A > 770 nm. This would explain the slight increase in thresh-
old over the A- 2 curve at A< 770 nm, resulting in a practically fiat behavior
over the limited wavelength band studied.
Measurements of breakdown in argon at p = 1 atm and A = 0.35 J.Lm
were carried out by Weyl et al. (1982, 1985) by using tripled Nd:YAG and
Nd:Glass beams and an XeF laser beam. The laser pulse lengths ranged
from 0.4 to 500 nsec. The measured thresholds were found to agree with
thresholds calculated using multiphoton ionization rates derived the low-
pressure and short-pulse data of Krasyuk and Pashinin (1972) and electron
impact ionization rates obtained from numerical solution of the quantum
kinetic equation. The data give a pulse length-scaling / 1h ex rP- 213 and a
threshold on axis of 3 x 1011 W/cm2 at rP = 10 nsec, a factor of 3 to 5 times
higher than the thresholds reported for similar pulse lengths by Buscher
et al. (1965) and Alcock et al. (1972). The earlier data, however, were taken
at the doubled ruby wavelength. Rosen and Weyl (1987) measured thresh-
olds in N2 and the rare gases at A= 0.53 and 0.35 J.Lm using a Nd:YAG laser
of 15 nsec pulse duration in its second and third harmonics. The gas pres-
sures were varied from 0.2 to 15 atm. J 1h at p = 3 atm was found to be 1012
(5 X 1011 ), 1010 (1 X 1011 ), 5 X 1011 (1012 ), 4 X 1010 (2 X 1010 ) W/cm2 for N2 ,
Ar, Ne, and Xe, respectively, at 0.53 J.Lm (0.35 J.Lm). The p- 1 scaling of J 1h
for Ne, Ar, and Xe and p- 0·6 scaling for N2 indicate that cascade ioniza-
tion rather than MPI dominates the breakdown process. The thresholds
calculated theoretically are in general agreement with the measured values
except for N2 at A = 0.53 J.Lm where the measured threshold was an order
of magnitude higher than the calculated irradiance for 100% ionization on
axis due to MPI alone.
1.4 LASER-INDUCED BREAKDOWN

OF SOLIDS AND LIQUIDS
The thresholds for breakdown of solids and liquids are usually significantly
lower than the breakdown thresholds of gases. Aerosols that are always
present in the atmosphere will initiate breakdown at laser irradiances where
laser beams would normally propagate unattenuated if the air were devoid
of particles. The localized plasmas grow (see the next chapter) until they
coalesce and finally prevent the tail of the beam from propagating any fur-
ther. Because of the aerosols present, there will be a limited fluence per
unit area ¢; that can be delivered to a target when the laser intensity ex-
ceeds a threshold irradiance I. The study of how ¢; depends on particle
distribution, composition, and number density is therefore of great prac-
tical interest. The determination of laser-induced breakdown mechanisms
and thresholds in optical materials, either transparent (lenses, windows)
or reflective (mirrors), is also of great practical interest since plasma igni-
tion will reduce the lifetimes of these materials and, concurrently, limit the
power level that can be extracted from a laser system. Laser damage to opti-
cal materials has been the subject of annual conferences since 1969 (Guen-
ther and Bass, 1969-1985). Many papers presented at these conferences
deal with plasma formation. Finally, lasers are increasingly being used for
cutting, welding, and treating materials. Plasma, once formed, will modify
(usually decrease) the laser coupling to the material surface.
It is useful to separate solids and liquids into two categories insofar as
their response to laser radiation: transparent and absorptive. By transpar-
ent, we mean a material whose absorption length at low irradiances is so
long at the laser wavelength of interest that one must invoke a nonlinear
absorption mechanism that would lead to breakdown at high irradiances.
The main category for absorptive materials comprises metals, which, for
all practical purposes, are surface absorbers and can at short laser pulses
(where thermal diffusion into the material is small) generate a vapor that
becomes rapidly ionized.
1.4.1 Breakdown of 'ftansparent Dielectrics
The breakdown threshold of a transparent dielectric material devoid of im-

purities and defects is expected to be quite high. As in the case of gases,
breakdown will occur when electrons are generated by MPI and grow by
electron impact of neutrals (or ions) in the lattice that result in electrons
going from the valence to the conduction band. The band gap of many di-
electrics is smaller than the ionization potential of most gases so that elec-
tron generation by MPI should be a more important process. Also even
pure materials are expected to have impurities in excess of 1 ppm whose en-
ergy levels lie close to the conduction band and become ionized at low laser
irradiance. If breakdown is due to electron avalanche, then one can use
the wavelength scaling relation based on the heating rate Eq. (1.27) to de-
rive breakdown thresholds at>. for de breakdown thresholds, if the electron
phonon collision frequency is known. We show in Fig. 1.17 the breakdown
thresholds in Si02 and NaCl measured by Aleshkevich et al. (1976). Con-
38 Weyl
•
15
e
~
> 10
6
.....
w
..c:
•4
5
3
•
0
0.250.35 0.53 0.69 1.06
A. (!lm)
Figure 1.17 Experimental data on the breakdown field as a function of the wave-
length of the exciting light (following Aleshkevich et al., 1976): 1) crystalline quartz;
2) fused quartz; 3) crystalline NaCl; 4) data for crystalline NaCI, from Aleshkevich
et al. (1976).
version framE (V/cm) to I(W/cm2 ) is readily made through the formula
(1.53)
Thus, I ~ 1.5 to 7 x 1011 W/cm2 for the quartz and I ~ 4 x 1010 W/cm2
for NaCl. The breakdown field is seen to be practically wavelength-
independent over the wavelength range 0.3 < A< 1 J.tm. This would indicate
that the collision frequency is of the order of 5 x 1015 sec- 1 .
The phenomenology of laser-induced breakdown of dielectrics is similar
to that of gases, with the band gap e8 playing the role of the ionization po-
tential. In order to obtain quantitative results that adequately model the
interaction, one requires an understanding of the source of initial electrons
and a knowledge of the electron-photon-phonon interaction for electron
energies ranging from the Fermi energy eF to ec + e8 , where ec is the en-
ergy at the bottom of the conduction band. This information is, however,
unavailable. The materials most studied both experimentally and theoret-
ically have been the alkali halides and Si02 • A good review of the work
carried out before 1978 can be found in Lee (1978). We discuss below some
of the more prominent theoretical work and present some more recent ex-
perimental results.
Holway and Fradin (1974) used the Frohlich model for the polar con-
tribution to electron-phonon interactions, which is dominant at low elec-
tron energies, and the Seitz model (Seitz, 1949) for the nonpolar contribu-
tion. They numerically solved the Fokker-Planck equation for electron dif-
fusion in energy space. By adjusting the collision frequency to 1015 sec- 1 ,
they were able to obtain good agreement with breakdown data on NaCl
at .A = 10.6 and 1.06 p,m for pulse lengths ranging from 15 to 300 psec.
Analytical solutions to the Fokker-Planck equation under certain limiting
conditions were obtained by Epifanov (1974) and applied to breakdown in
sapphire for pulse lengths ranging from 10-10 to 10-7 sec, but the theory
was unable to take into account the effect of laser frequency. In later pa-
pers, a more general solution of the quantum kinetic equation was obtained
and the results applied to breakdown in NaCl (Epifanov et al., 1976; Gor-
shkov et al., 1977). Qualitative agreement with experimental results were
obtained and the thresholds did show a frequency dependence, with a peak
threshold for a photon energy of 2 eV (Gorshkov et al., 1977). Brauer
(1979) used a phenomenological approach and solved a set of rate equa-
tions that connect the distribution function of electrons at energy e to the
distribution at e ± hv. The theory contains two fundamental parameters:
(1) the rate of absorption of photons r = ul/hv and (2) the rate of emission
of phonons 'Y. Cascade rates are calculated as a function of the ratio r h for
a series of band gaps, represented by the number 8 = Int( eglhv + 1) ranging
from 4 to 20. In the model, 'Y and the cross section u (or r) are taken to
be independent of electron energy. NaCl data at 1.06 p,m are best fit with
u = 1.7 x w- 11 cm2 and a rate of emission of phonons of energy 200 cm- 1
of 1014 sec- 1 • The theory could not account for the fact that the breakdown
threshold of NaCl is practically independent of wavelength.
An attempt at relating the breakdown thresholds of liquids to their
macroscopic properties such as temperature T and viscosity 'f/ was carried
out by Arushanov et al. (1981). They related the relaxation time for the
electron momentum to density fluctuations and showed that the electron
diffusion coefficient in energy space would vary as r,. They predicted
that the breakdown threshold of glycerine, whose viscosity decreases dra-
matically with increasing temperature, would increase with temperature,
whereas the thresholds for water and ethanol would be insensitive to T.
The measurements that they carried out on these liquids using a ruby laser
with rP = 200 nsec were in qualitative agreement with their predictions over
the temperature range 10 to 90°C.
Breakdown thresholds of a variety of optical materials (PMMA, quartz,
ADP, KDP, LiNb03 , Lil0 4 and SrTi0 3 ) were measured at 1.06p,m (rp =
12 nsec, beam diameter= 30 p,m) by Bass and Barrett (1972), who observed
that the probability of damage of all materials could be fit to a relation of
the form p =A exp( -KIE) where A and K depend on the material and E is
40 Weyl
the electric field strength. Such a functional dependence that had already
been observed in de breakdown experiments is indicative of a cascade-type
breakdown (Shockley, 1961). Experiments on laser-induced breakdown of
NaCl crystals at 1.06 pm and rP = 10 nsec were carried out by Fradin et al.
(1973a, 1973b). Breakdown was ascribed to electron avalanche. The theory
of Sparks et al. (1977) that had no adjustable parameters was found to give
results in excellent agreement with the data. More recent experiments on
the same crystals have been carried out by Soileau et al. (1978) at .A = 10.6,
3.7, and 2.7 pm and by Gorshkov et al. (1977) at .A= 0.53, 0.69, 1.06, and
10.6 pm. The effect of crystal temperature and heat treatment was also
analyzed by the last authors, who found that by heating a crystal to a tem-
perature just below its melting point and then rapidly cooling it to ambient
temperature, one could significantly increase its breakdown threshold. A
slower cooling, however, would result in the formation of local clusters that
would enhance the avalanche process for photon energies larger than 1 eV.
The thresholds that Gorshkov et al. (1977) measured are presented in Ta-
ble 1.3. They are significantly higher than the thresholds measured earlier
by other groups.
The dependence of / 1h on pulse length rP in NaCl has been studied at
1.06 pm (Fradin et al., 1973b; Smith, 1978; Van Stryland et al., 1981) with
r;
the result / 1h ex 112 • The strong focal area dependence / 1h ex wF- 4 in the
same experiments indicates, however, that the defects in the crystal play an
important role in the breakdown/damage process. The influence of beam
coherence on breakdown has been studied experimentally by Glebov et al.
(1984). Beam coherence was found to affect bulk damage thresholds but
not damage due to surface defects.
Much lower thresholds are observed for materials that are not of optical
quality either due to bulk impurities or surface defects. Surface defects can
Table 1.3 Breakdown Threshold (1010 W/cm2 ) of

Alkali Halide Crystals
Bandgap >.(p.m) = 10.6 1.06 0.69 0.53

Crystal (eV) rp(nsec) =60 10 10 7
NaCl 8.6 2.3 14 15 13

KBr 7.3 3.0 5 5.7 5
KCl 8.7 2.5 7 8.0
NaF 11.5 14 14.0
LiF 13.6 30 36.0
Source: Gorshkov (1977).

be (Blombergen, 1974) cracks, grooves, spherical pores. There may also be

dust particles or microscopic surface inclusions from the grinding process.
Bulk impurities could be metallic particles such as platinum, coming from
the crucibles used in the forming process. All these defects can cause local
field enhancement that, in effect, lowers the threshold. They can also cause
localized heating that leads to a thermal runaway. Consider an absorbing in-
clusion of radius R and absorption coefficient k. If kR «: 1, then absorption
is volumetric and the power deposited will be proportional to kR 3 • Losses
by thermal conduction to the solid will be proportional to Rt::.T so that the
steady-state temperature rise t::.T due to the inclusion will be proportional
to kR 2 • There will be a distribution of particle sizes in the focal volume,
and the larger particles will be the cause of ignition. The larger the fo-
cal volume, the higher the probability of there being a large inclusion and
the lower the breakdown threshold. Plasma ignition by localized absorp-
tion sites has been analyzed by Anisimov and Melshantsev (1973) and by
Sparks and Duthler (1973). An unresolved issue, however, is the means by
which localized heating can lead to plasma formation and growth. Gener-
ation of thermal electrons requires temperatures of thousands of degrees.
Before a plasma is formed, there would by necessity be formed a hot va-
por pocket, which may or may not continue absorbing the laser radiation.
Mechanisms have been invoked to explain plasma formation such as the
turboelectric process (Dymaev et al., 1983) and band gap narrowing due
to compressive stresses around the heated site (Lysikov, 1978). Extensive
studies of laser-induced damage to polymeric optical materials such as poly-
methyl methacrylate (PMMA), polycarbonate, and polystyrene have been
carried out at the Lebedev Physics Institute. The work is reviewed in the pa-
per by Dymaev et al. (1983). Damage has been ascribed to the presence of
microscopic inclusions that could be eliminated by purification and filtering
of the monomer (Manenkov and Nechitailo, 1980). A factor of 2 increase in
the damage threshold of the monomer has been reported (O'Connell et al.,
1984).
Impurities in the form of microscopic suspended particles also play an
important role in determining the breakdown thresholds ofliquids (Butenin
and Kogan, 1971). One observes, e.g., order-of-magnitude differences in
thresholds in water depending on whether one is dealing with unfiltered
tap water (many particles, low threshold), distilled water (fewer particles,
higher threshold), or doubly distilled water (very high threshold). The pres-
ence of soluble impurities, such as salt, does not seem, however, to affect
thresholds. Since particles play such an important role, one expects a strong
dependence of breakdown threshold on beam diameter, which is indeed the
case. As in gases, breakdown thresholds in subsequent pulses are affected
by breakdown in a previous pulse. This behavior observed in the break-
down of liquid helium has been ascribed to particles ejected from the walls
42 Weyl
(or windows) by the previous pulses, independent of whether breakdown

had occurred (Abrikosova and Boskova, 1965; Abrikosova and Anshukov,
1971). It is interesting to note, however, that the introduction of electrons
in the focal volume before the laser pulse was turned on did not modify the
thresholds observed (Hunklinger and Leiderer, 1971 ). This result is not sur-
prising in light of recent experiments (Migus et al., 1987) that show that free
electrons can only exist in water for 0.1 psec, after which time they become
trapped in shallow localized states, and then into deeper solvated states.
We show in Table 1.4 the measured breakdown thresholds of water and sev-
eral other liquids at different wavelengths. One notes the large discrepancy
in thresholds measured by different groups. Yablonovich and Blombergen
(1971) showed how one could estimate optical breakdown thresholds of
pure dielectrics from measured de thresholds. This is of particular inter-
est for liquids since the dielectric breakdown strength of many liquids has
been measured. In a de discharge, an electron cloud drifts from the cathode
to the anode and grows exponentially with distance X as exp[7J(E)X]. 77(E)
can be related to the rate of ionization vi(E) and the drift velocity V of the
electron through the relation 77(E) = vi(E)IV. Yablonovich and Blomber-
gen (1972) obtained limits on the electron velocity and used results of de
breakdown experiments to predict ac breakdown thresholds. The varia-
tion with wavelength of the breakdown threshold was obtained by use of
Eq. (1.26). The weak link in the derivation, however, is the high field mo-
bility of electrons. There is, at present, no theory from which the mobility
can be derived accurately.
Nonlinear optical effects can affect measured breakdown thresholds. If
the solid or liquid has a nonlinear index ofrefractionn = n 1 +n 2E 2 with n 2 >
0, then above a critical laser power the beam will self-focus in the dielectric
(Pan tell and Putthof, 1969). The nonlinear index n 2 in solids is mainly due
to the electrostrictive effect, i.e., contraction of the solid in one direction,
and self-focusing will occur over a characteristic timet = a/c, where a is the
radius of the beam and c the speed of sound. In the case of liquids, the
nonlinear index is due to the orientation of the molecules in the field when
the molecules have anisotropic polarizabilities (optical Kerr effect).
A derivation of the critical power for self-focusing can be obtained from
simple arguments using Fresnel's laws for construction ofwavefronts. Con-
sider a plane wave traveling in the Z direction. The electric field will vary
as Eei(w/c)nZ ~Eei(w!c)(nt+nzE 2 )Z. The phase difference between the axis of
the beam and the edge (E = 0) is (w/c)n 2E 2 Z. This phase difference will
result in a spherically converging wavefront for n 2 > 0. However, if diffrac-
tion is taken into account, the beam would be diverging with half-angle
8 = 1.22>../D, D being the beam diameter and .A the wavelength. The phase
difference between the axis and the edge of the beam, due to diffraction,
is (w/c)n 1Z(1- cosO)~ (w/c)n 1Z8 2 /2. The threshold power P is reached
c
"C
Q.
a
CD
Thble 1.4 Breakdown Thresholds of Water and Selected Liquids 0
:::s
Intensity ...m
Tp a threshold m
~
Q.
Wavelength Liquid (nsec) (J.tm) W/cm 2 Ref.
~
:::s
l.06J.tm Water ,:r
Doubly distilled 50 30 6.2 X Hf 1 Roach and Davies (1970) '<
Ul
Distilled 7 175 5 X 109 Docchio et al. (1986) n
Ul
Distilled 7 15 3 X 1010 Docchio et al. (1986)
Tap water 25 50 1 X 1011 Pozhidaev and Fatievskii (1981)
(±30%)
Tap water 7 175 2 X 109 Docchio et al. (1986)
Tap water 7 25 4x 109 Docchio et al. (1986)
Saline solution 7 175 4.5 X 109 Docchio et al. (1986)
Saline solution 7 25 2.5 X 1010 Docchio et al. (1986)
Xylene 50 65 5.8 X 1010 Roach and Davies (1970)
Cyclohexane 50 60 2.7 X 1011 Roach and Davies (1970)
Benzene 50 60 4.4 X 1010 Roach and Davies (1970)
0.69J.tm Tap water 21 37 3.8 X 1010 Pozhidaev and Fatievskii (1986)

(±16%)
Tap water 20 X 4 X 108 Ioffe et al. (1970)
.j:>
(,.)
44 Weyl
when both phases exactly balance, i.e.,
or
p = (cE 2 ) 1rD 2 = (1.22)2n 1c..\2 (1.54)

47r 4 32n 2
The threshold power in cs2 is 2 X 1o4 wand 106 win liquid hydrogen. For
a good review of self-focusing in solids and liquids, see Shen (1975).
1.4.2 Breakdown of Absorbing

Dielectrics and Metals
(A) Absorbing Dielectrics
If the dielectric has an absorption coefficient k that is significant at the laser
wavelength, then a layer of thickness k- 1 is heated and will vaporize if a
sufficient energy per unit volume is deposited. The vapor layer may form at
such a high temperature that it will be partially ionized. Subsequent depo-
sition of laser energy by inverse bremsstrahlung absorption may eventually
lead to breakdown in the hot vapor. If the vapor comes off at too low a tem-
perature for significant ionization, as is the case with the vaporization of
liquids, then there is no mechanism for subsequent vapor heating. The va-
por, however, can drive a strong shock into the surrounding air (Steverding,
1974). Absorption in the shocked air that is partially ionized may eventu-
ally lead to breakdown if the shock is strong enough and the laser irradiance
high enough. This is the mechanism of formation of laser-supported deto-
nation waves (Raizer, 1965) that are described in chapter 2.
The laser-induced heating and vaporization of surfaces have been treated
in great detail by Ready (1971). The temperature as a function of time and
depthx is described by the following equation for the one-dimensional case:
8T - -8- ( K8T)
pC - =(1-R)kle -kx (1.55)
8t 8x 8x
where p, C, K, and Rare the density, specific heat, thermal conductivity,

absorption coefficient, and reflectivity of the material, respectively. The
diffusivity of the material is x = K I pC and the scale length l D for diffusion
of heat during the pulse length TP is lD = ..;x;;.
If lD is much less than the
radius of the beam and also much less than the absorption depth k - 1 in the
material, then the one-dimensional Eq. (1.55) that neglects radial diffusion
losses is valid and the temperature rise at the surface is given by
T- To= (1-R)k¢> (1.56)

pC
By setting T = Tv, one obtains the timet and the fluence per unit area ¢>
required to initiate vaporization
¢> = !I d = (Tv- To)PC (1.57)

t k(1-R)
We show below how breakdown can either occur in the hot vapor products
or in the air that has been shock-heated by the vapor acting as a piston.
As an example of an optical material that can break down once it vapor-
izes, consider the case of Vycor glass for which Tv = 3700K, p = 2.2 g/cm2 ,
C(Tv- T 0 ) = 4000 J/g, k = 2000 cm- 1 (at,\= 10.6 flm), and 1-R = 0.96.
Using these values to evaluate ¢>from Eq. (1.57), we find ¢> = 3.4 J/cm 2 • The
thermal diffusivity is (Touloukian, 1967) x ~ 1 x 10-2 cm2 /sec. Thermal dif-
fusion losses can be neglected over pulse lengths shorter than (xk 2 )- 1 =
25 flSec.
One can treat the breakdown of glass as the breakdown of vaporized Si0 2
in which the alkali and alkaline-earth oxides are considered as a perturba-
tion on the vapor properties. Partial ionization of Al, Ca, Mg, Na, ... , that
are constituents of many glasses will create enough electrons in the vapor
that a thermal runaway may occur through IB absorption of laser radiation.
The breakdown threshold can be estimated as follows. The mass flow rate
for the steady-state vaporization of the surface is given by
(1.58)
where c = (8knTJrrmv) 112 , mv is the average molecular weight of a vapor

molecule, and the subscripts refers to surface conditions of density and tem-
perature of the vapor. Equation (1.58) was derived by integrating the dis-
tribution function of vapor molecules over a half-Maxwellian distribution
in velocity space. The mass flow rate is determined by the conservation-of-
energy equation
labs =1(1-R) =rhHv (1.59)
where Hv is the heat of vaporization of the glass. Let the relative concentra-
tion of an easily ionizable alkali or alkaline earth atoms in the vapor phase
be c. The vapor density is related to the surface temperature through the
46 Weyl
Clausius-Clapeyron relation
(1.60)
where B and Bv are constants. We assume that initially the concentration of

ions in the vapor phase is given by the Saba equation
(1.61)
where [M] = cn 8 ; g and g + are the degeneracies of the atoms and ions, re-
spectively; and 81 is the ionization potential inK. If we have only one easily
ionizable specie, then nes = [M + ]. Heating of the vapor is governed by the
equation
(1.62)
where K is the electron-neutral inverse bremsstrahlung cross section per

unit electron and neutral particle density and the subscript v refers to vapor
properties.
As the vapor temperature increases, the electron concentration, the so-
lution of Eq. (1.61), increases and the absorption coefficient Knen on the
right-hand side of Eq. (1.62) increases. This corresponds to a runaway sit-
uation. It can be shown that the time to break down the vapor is simply
obtained by calculating from Eq. (1.61) the increment of temperature rise
6.T required for one e-fold of the electron density and by calculating from
Eq. (1.62) the time required to raise the temperature by 6.T (if we assume
frozen electron concentration). The vapor breakdown time is
(1.63)
Setting nes = [M +] in Eq. (1.61) and differentiating with respect to T, keep-

ing only the leading term, we find
Setting 6.nelne = 1, inserting the corresponding value of 6.T8 into Eq. (1.63)
and using Eq. (1.61) to eliminate ne, we obtain the breakdown time
1.1 x 10- 12 Ts7' 4 58,000

= exp-- (1.64)
J (W/cm2 )c 112 Ts
where 7J is a factor that takes into account the initial expansion of the va-
por across the Knudsen layer (Anisimov, 1968) and 'Y is the ratio of specific
heats.
In the last step, we have used the following values for a vapor with an
alkali impurity: 'Y = CP!Cv = 1.2, K = lQ- 37 cm5 (.A= 10.6 pm), 7J = 0.4,
81 = 55,700K (E1 = 4.8 eV), Hv = 12,500 J/g (8v = 60,300K), corresponding
to a vapor molecular weight mv of 40 g, B = 2.43 x 1014 dyne/cm2 , A = 2.41 x
1015 cm- 3 (K)- 1·5 , (g+fg) = 0.5, andR = 0.05. Equation (1.64) can readily
be evaluated since Eqs. (1.58-1.60) can be solved for Ts as a function of
I, i.e.,
T = (}v = 60,370 (1.65)

s In [ H,m,Bc ] In [8.4 x 1Q14f(ITll2)]
4Ik8 T,(l-R)
The breakdown time is now given by the sum of the time (rH) required to
heat the surface to its vaporization temperature and the time ( rvbr) required
to break down the vapor.
The heating and vapor breakdown times are plotted in Fig. 1.18. The con-
centration of easily ionizable alkali atoms in the vapor for glass is expected
to be in the 0.01 to 0.1 range.
For pulse lengths less than 1 ps, the delay to breakdown is due to the
heatup time of the solid. Breakdown thresholds for pulse lengths in the
10 to 20 psec range are calculated to be around 1.5 to 2.5 x 1o<' W/cm2 •
The calculations can be extended to other wavelengths by scaling the IB
absorption coefficient asK(.A) =K(>.0 )(.AI.A0 ) 2 • This would raise (or lower)
the vapor breakdown time (dashed curve in Fig. 1.18) by the factor (.A/.A 0 ) 2 •
We now turn to the second breakdown mechanism that is applicable to
polymeric optical materials and liquids that vaporize at a temperature too
low for any ionization of the vapor. The strong shock created in air can
drive a laser supported detonation (LSD) wave that is self-sustaining. The
threshold for maintaining an LSD wave in air at p = 1 atm is 2 x 107 W/cm2
(Weyland Shui, 1977). At this intensity, the velocity of the detonation front
is VD = 5 x 105 em/sec, and the velocity and pressure behind the shock are
VDI('Y + 1) = 2.3 x 105 em/sec and 340 atm. If the vapor expansion is initially
driving the wave, then the vapor mass flow rate must be
· _ PvVD _ mvVDP
m- - - - --!....-.!=--
"(+1 (k8 Ty)('Y+1)
48 Weyl
TIME TO STEADY\ 0·1 \ [Na, K] = 0·01

STATE VAPORIZE
\
\
\YJTIMETO
\ ~ ~ BREAKDOWN
~ \ \VAPOR
\ \
------ --------\--"'1'-------- 'tp
1\ ,,
I I
I \
:
I \
\ \
\: \
~ \
:\ \
\ \
\
3400 3600 3800 4000 4200

T 5 (k)
Figure 1.18 Heating and vapor breakdown times for glass, .\ == 10.6 p.m.
The threshold laser irradiance is therefore estimated to be

rhHv mvVdfvP
1th = 1-R- (!+ 1)kBTv(1-R)
Setting mv = 30 g, Hv = 10 kJ/g, Tv = 2000K, 1 = 1.2, R = 0.05, VD = 5 x
105 em/sec, and p = 340 atm, we obtain / 1h = 1.5 x 108 W/cm 2 , which would
correspond to the threshold for an organic polymer at A = 10.6 J-tm. The
same calculation for water (mv = 18 g, Hv = 2000 J/g, Tv = 700K) yields
/ 1h = 5 x 107 W/cm2 , which is a factor of 10 below the observed threshold
(Lencioni, 1974). The much larger observed threshold is probably due to
the fact that as the water is heated to its critical point, the absorption at
10.6 JliD decreases. (Water aerosols absorb strongly at 10.6 JliD, whereas
water vapor absorbs only weakly.)
(B) Metals
Plasma ignition above metal surfaces has been extensively investigated both
experimentally and theoretically. Most of the work has been carried out at
10.6 JliD and the metal the most studied has been aluminum. Experimental
results on aluminum at 10.6 ttm can be found in Maher et at. (1974), Walters
et al. (1974), McMordie and Roberts (1975). Marcus et al. (1976), Golub
and Nemchinov (1980), Bessarab et al. (1978), and Itzkan and Cronberg
(1981). Other metals studied at 10.6 ttm are Ti (Maher et al., 1974; Marcus
et al., 1976), Mg (Barchukov et al., 1974); Cu (Maher et al., 1974; Konov
et al., 1983); Cu, W, Pb, and Zn (Walters et al., 1974); and Mo, Nb, and Ta
(Bondarenko et al., 1979). Metals studied at 1.06 ttm are AI (Kozlova et al.,
1975; Bessarab et al., 1978); Bi (Kozlova et al., 1975); and Cu (Uglov and
Galiev, 1980). Breakdown at the ruby wavelength (>. = 0.69 ttm) has been
studied on AI and Pb (Gnoevoi et al., 1970). Finally, breakdown on Ti and
AI surfaces has been studied (Rosen et al. (1982a, b) at 0.35 ttm using a long-
pulse length (0.5 ttsec) excimer laser. Although not a metal, graphite has
an extremely short absorption length at all wavelengths and good thermal
conductivity. Plasma initiation above graphite surfaces has been studied by
Bessarab et al. (1978) and Danilychev et al. (1981) at the C02 wavelength.
Various experimental means have been used to measure plasma forma-
tion. One can measure the transmission of a probe beam travelling parallel
to the surface or of the main beam by drilling a pin hole through the mate-
rial. Other diagnostics of plasma formation are a dramatic decrease in the
surfaces reflectivity during the pulse, line emission by ionized species, ra-
diometry of the surface, and direct measurement of electron concentration
by electrostatic probes.
Shown in Fig. 19(a) are the ignition data obtained by Walters et al. (1974)
on various metals at >. = 10.6 ttm. Figure 19(b) shows the temporal pro-
file of the gain-switched spike of the laser pulse. As can be seen from
Fig. 19(a), ignition was always found to occur during the rising portion of
the gain-switched spike. Threshold intensities lie in the ranges 3 x 107 to
5 x 108 W/cm 2 •
Plasma formation is always preceded by surface vaporization. Absorp-
tion in metals occurs within skin depth that has a thickness measured in
tens of angstroms. Unless the laser pulse is extremely short, the thermal
diffusion length will be much larger than the skin depth so that the surface
temperature, solution to Eq. (1.55), will be given by (Ready, 1971, p. 73)
(1.66)
where a, K, and x are the metal absorptivity, thermal conductivity, and

thermal diffusivity, respectively. Consider, e.g., the case of aluminum. The
absorptivity at 10.6 ttm of a clean surface of the pure metal is in the range
0.02 < a < 0.04 and increases to 0.07 just below the melting temperature
(Weyl et al., 1981); K = 2.4 W/cm K and x = 1 cm 2 /sec at room tempera-
ture. Using these values, I = 108 W/cm2 , t = 5 x 10-8 sec, and an average
50 Weyl
1.0
>- 0.8
!:::
(j)
z
w 0.6
1-
z
w
>
f= 0.4
<(
_J
w
r:r:
0.2
0
0 100 200 300 400
(a) TIME (ns)
1 o3
(j)
.s
w
:E
i=
z
0
.
i= 1 o2
•• <&
<(
E
~
6 @
! ,. , ,
0 •
r
en
...J
•
1oi
107 108 109
PEAK IRRADIANCE(W!cm2)
(b)
Figure 1.19 Laser induced breakdown above metals at>.= 10.6 J.Lm a) Temporal
profile of laser pulse b) Time to form a plasma in air using the electrostatic probe
method (open symbols) and target voltage method (closed symbols) to measure
electron density, • W, • AI, • Pb, 0 Zn, • Ca, from Walters et al. (1978).
absorptivity a= 0.05, one calculates from Eq. (1.66) T- T 0 = 500K, which

is so low that one would not even expect the metal to melt. Walters et al.
(1978) have shown, however, in their important work, that ignition occurs
on thermally insulated defects (flakes) of thickness d = 0.1 to 0.3 t-tm and
lateral extent R > 3 J.Lm. Since d is much smaller than the thermal diffusion
depth, the temperature rise of a flake is governed by the equation
aft
T=T0 + -
pCd
where p and Care the density (p = 2.7 g/cm3) and specific heat (C = 0.9
Jig K) of the metal (aluminum). Under the same conditions as before and
setting d = 0.2 t-tm, one calculates T- T 0 = 5100K. This temperature rise
is large enough that one would expect vaporization well within the gain-
switched spike.
The generation of a large concentration of electrons (ne ~ 1013 to
1015 cm- 3 ) above the surface due to thermionic emission before or during
vaporization has often been considered as the cause of plasma ignition. The
plasma sheath that one calculates for such high densities is, however, so thin
that electrons leave the surface, reflect off the edge of the sheath, and return
to the surface before they can make any collision with the vapor atoms. It is,
therefore, extremely unlikely that any absorption of laser radiation by elec-
trons in the sheath will occur. A more likely mechanism that can explain the
observed thresholds is absorption by thermally generated electrons in the
vapor. Weyl et al. (1981) have modeled in some detail the breakdown initi-
ated by a flake on an aluminum surface. The characteristic time is the vapor
expansion time texp =RIV, where V ("'lOS em/sec) is the expansion velocity
of the vapor. They found that when I> 3 x 107 W/cm2 , breakdown would
occur in the vapor over a time scale t < texp for a flake radius R = 5 t-tm. The
breakdown time was then dominated by the time required to heat a flake
to its vaporization temperature. When I< 3 x 107 W/cm 2 , the vapor is not
heated during texp; the expansion becomes three-dimensional, with the va-
por density decreasing as 1/r 2 , r being the distance from the flake. There will
be a shock lying above the defect that brings the overexpanded vapor back to
near ambient pressure. Breakdown will occur in the vapor behind the shock
over times (t > 1 J.LSec) that are much longer than the flake heating time. The
first excited state of AI has an energy f* = 3.14 eV that is slightly more than
half the ionization energy E1 ( = 5.98 eV). Electrons can be readily heated
by IB collisions with AI atoms until they reach an energy in excess of f*,
at which point inelastic losses will limit the energy rise. Weyl et al. (1981)
found that the vapor breakdown occurs in two stages. During the first stage,
the population of the first excited state (AI*) increases linearly with time and
the electron concentration remains constant. The second stage starts when
52 Weyl
[AI*]= 1014 cm- 3 , at which point impact ionization of AI* by 3 eV electrons

causes rapid electron growth, leading to breakdown. Other metals such as
Cu, Ti, Fe (but not Zn) have low-lying electronic states. It is expected for
these metals that a two-temperature (bound + free electron temperature
Te and heavy temperature T8 ) model of vapor heating and breakdown is
better suited for describing the physics (Rosen et al., 1982a, b).
Since surface defects play such an important role in determining plasma
ignition at A = 10.6 pm, means of eliminating then in order to increase the
thresholds have been investigated (Itzkan and Cronberg, 1981; Konov et al.,
1983). It was found that the threshold of Cu mirrors would significantly be
raised by cleaning with H 2 0 2 and using C0 2 laser pre pulses to vaporize the
remaining defects (Konov et al., 1983).
Because of the >.- 2 scaling of breakdown time in the vapor, thermally
insulated surface defects play no role in the breakdown process at wave-
lengths shorter than approximately 2 pm. The breakdown data on metal
surfaces at wavelength >. ::::; 1 pm all occur in a region where bulk vaporiza-
tion of the metal occurs. Analysis of breakdown of AI and Ti surfaces at
>. = 0.35 pm have been carried out by Rosen et al. (1982a, 1982b), and a
model based on heating of the surfaces to vaporization that incorporates
measured surface absorptivities, followed by thermal runaway in the vapor,
gave very good agreement with the data. Breakdown thresholds for both
of these metals was 107 W/cm2 for a pulse length of 0.5 psec, with a pulse
length scaling I ex: r- 213 • The effect of pressure on plasma and LSD wave
ignition has been studied on various metal and dielectric surfaces (Maher
et al., 1974) using a C02 laser ( rp = 10 psec) with a small gain-switched spike
(Ipli ~ 1). The measured thresholds ranged from 107 to 4 x 107 W/cm2 with
no apparent pressure dependence for practically all materials over the pres-
sure range studied (0.01 <p < 1 atm). The effect of ambient gas (air, He,
Ar) on plasma initiation at 1.06 pm has been investigated on AI and Be sur-
faces, with the nature of the gas having only a small effect on the measured
thresholds (Kozlova et al., 1975).
1.4.3 Aerosol-Induced Air Breakdown
Solid aerosols are composed of three main ingredients (Voltz, 1972): a

water-soluble element, soot, and mineral dust. The water-soluble element
is composed of salts [(NH4 ) 2 S04 , NaCI] and may represent up to 50% of
the aerosol content in industrial and maritime areas. Over desert areas,
the water-soluble element will be much smaller and will appear in par-
ticles of diameter much less than 1 pm. Soots are present in populated
regions or during and after forest fires. They usually represent less than
10% of the aerosol content. Mineral dust will reflect the mineral struc-
ture of the ground, the most common components being montmorillonite
[(0.5 Ca, Na) 0.7 (Al, Mn, Fe) 4 (Si, Al) 1 0 20 (OH) 4 ], illite [K1 _ 15 Al 4 Si 7 _ 65
A1 1 _1. 5 0 20 (OH) 4 ], kaolinite [(2 A1 2 0 3 , 4 Si 0 2 ) (OH) 8 ], calcite (Ca C03 ),
quartz (Si0 2 ), and metal oxides. The size distribution of aerosols is shown
in Fig. 1.1 and is seen to follow a R- 4 law.
Air breakdown thresholds initiated by solid aerosols are shown in Fig.
1.20, as a function of pulse length and particle size, for .A = 10.6 J.Lm. This
figure represents a compilation of experimental results from several inves-
tigators (Lencioni, 1973; Schlier et a!., 1973; Lowder and Kleiman, 1973).
One sees that a minimum ftuence per unit area of 10 J/cm2 is required to
cause a breakdown and that, for long pulse lengths, the thresholds scale as
R- 1 toR- 2 • In an experiment carried out in unfiltered air, the actual break-
down threshold will be determined by the size of the largest aerosol in the
focal volume. For small focal volumes, the presence or absence of a large
particle in the region of high irradiance will give rise to an erratic behavior
of measured threshold. If one wishes to use Fig. 1.20 to determine, for a
given experimental configuration, the expected breakdown threshold, one
would need to know: (1) the number density n and size distribution[ (r) of
1010
~
.....
O CLEAN AIR THRESHOLD
N' '-..~:;:.1 ec.
~ .... ~""i----------
E
(.) 109 1 11m
~
~~ ~-------o--- 3~-tm
--
w
(.)
z 108 06
<X:
0
<X: ~·'-
~ -----10~-tm
' ' '""'-

C(
~
107
0
u.
_J
0
I
(/)
..o.__ 30 11m
w
C( 106 10 J/cm2
I
I-
10 -9 10 -8 10 -7 10 -6 10 -s 10-4 10 -3
PULSE TIME -TIME TO BREAKDOWN (SEC)
Figure 1.20 Ignition thresholds in aerosol laden air: >. = 10.6 pm, from Reilly et a!.
(1977).
54 Weyl
aerosols and (2) the volume v in the beam where I > I 1h (R 0 ). One could then
calculate the probability of there being at least one particle of size R > R 0
in this volume. The threshold I 1h is given mathematically by the solution of
n r= v(I)f(r)dr;::1
}Ro(I)
(1.67)
where f (r) is the distribution of particle sizes and I 1h is the smallest value
of I for which inequality (1.67) is satisfied. Distributions for various solid
aerosols in the atmosphere are shown in Fig. 1.1.
Aerosol-initiated breakdown has been studied at 3.8 pm (Amimoto et al.,
1984), 2.7 pm (Lencioni et al., 1976), 1.06 pm (Lencioni, 1974; Bunkin
and Savranskii, 1975; Bani and Meskan, 1975), and 0.53 11m (Chylek et al.,
1986). The long-pulse (3 psec) data at 2.7 pm yielded a threshold intensity
that is a factor of 3 higher than at A = 10.6 pm, indicative of a A- l scaling
law, whereas the short-pulse (50 to 100 nsec) data at A= 1.06 pm is more
indicative of a A- 2 scaling law.
All aerosols absorb strongly at 10.6 pm and it can be shown (Reilly et al.,
1977) that when the ftuence per unit area exceeds 10 J/cm2 , a particle will
have absorbed enough energy that it will start vaporizing, if one neglects
conduction losses to the surrounding atmosphere. Smith (1977) explained
the breakdown by an electron cascade process in the vapor in which the
balance of energy gained by IB absorption by electrons and energy lost by
diffusion and by elastic and ionization collisions determined the ionization
rate in the vapor. His results as applied to 10.6 pm, 100 nsec pulses for
Al 2 0 3 and 1.06 pm, 50 nsec pulses for carbon are in reasonable agreement
with the data. We believe that the agreement is, however, fortuitous. The
irradiance levels are too low for cascade ionization of the vapor, which is
composed mainly of molecules (AIO and Al0 2 for the alumina particles,
C3 and C 2 for the carbon particles). Inelastic energy losses will dominate
except perhaps at the high breakdown irradiances observed for submicron
particles. The long-pulse heating and breakdown of Al 2 0 3 vapor products
has been analyzed in detail by Root and Pirri (1979), who found that absorp-
tion was initially dominated by AIO vibration bands in the infrared rather
than by electron-neutral inverse bremsstrahlung. They calculated enough
absorption and heating to obtain breakdown of a 30 pm particle in 15 psec
at 3 x 106 W/cm2 in good agreement with the date.
The steady-state regimes into which an aerosol particle of a given radius
R will evolve as a function of laser irradiance I are shown in Fig. 1.21(a).
They are determined by the energy conservation relation
EMISSIVITY = 0.1
1000
100
E:::1_
(f
10
---- R= LABS
1o-1
1Q-2
1o3 1o4 1os
LASER IRRADIANCE (W/cm2)
(a)
10-2~--------~-----L-L~~----------~L---------~
10-1 10 100
t (sec) xI (W/cm2)
(b)
Figure 1.21 Phenomenology of laser-aerosol (Al 2 0 3 ) with (a) steady state regimes
as a function of intensity and particle radius and (b) times to melt and vaporize,
from Reilly et a!. (1977).
56 Weyl
wherePa is the power absorbed from the laser beam andPc• Pv, andPR are
the powers lost by conduction, vaporization, and radiation. We can write
Pa = 1rR2QaI
where Qa is the Mie absorption efficiency. If R ~ la, where la is the ab-
sorption depth in the aerosol material, then Qa ex R. When R ~ la and
R > A./27r, then Qa approaches 1. At low irradiances, the aerosol heats up
to a temperature T at which heat conduction losses balance heating. In
the free molecular regime (R < 0.1 J.Lm), heat losses are proportional to R 2 ,
whereas in the continuum regime, P c ex R. Since for R < 2 J.Lm, P a ex R 3 ,
the power balance equation (Pc = P a) yields RC:!.T =constant (free molec-
ular) or R 2 C:!.T = constant (continuum) for constant I. At higher irradi-
ances, vaporization losses (Pv ), proportional toR 2 , become dominant. Fig-
ure 1.20(a) shows that radiation losses can be neglected for particles having
R < 1 mm. The experimental breakdown thresholds are seen to occur in
a regime where Al 2 0 3 particles are vaporizing strongly, with a saturated
vapor pressure in excess of 30 atm. The time to reach any of the steady-
state regimes is shown in Fig. 1.21(b), where the curves tm, tm+• tv_, tv+ cor-
respond to the start of melting, complete melting, start of vaporizing, and
complete vaporization of a particle of initial radiusR, respectively. Solution
of the energy equation shows that the vaporization time tv+ scales logarith-
mically with the ratio RJR1, where R; =initial particle radius and R1 = final
particle radius. The condition R;IR1 = 100 was chosen to plot Fig. 1.21(b).
The characteristic time for the vapor expansion is texp = R/V, where Vis the
expansion velocity (approximately the speed of sound). Letting R = 10 J.Lm
and V = 3 x 105 em/sec, we estimate this time to be 3 nsec. In analogy to
the breakdown model developed for breakdown initiated by surface imper-
fections on AI, one would expect that at irradiances in excess of 108 WI cm 2 ,
breakdown would be occurring during the vapor expansion phase. At lower
irradiances, the breakdown will occur in the overexpanded vapor that has
been shocked back to pressure equilibrium with the surrounding air. Heat-
ing will be balanced by thermal and particle diffusion losses with the sur-
rounding air. The observed R- 2 dependence of the long-term breakdown
threshold for particles of radius R > 6 J.Lm in Fig. 1.19 is a consequence of
these losses. Atmospheric dust contains large concentrations of alkali and
alkaline-earths that are partially ionized in the vapor, providing a larger con-
centration of electrons that the pure Si02 and Al 2 0 3 samples studied in the
laboratory. One would expect, therefore, the thresholds for a given particle
size to be lower than those shown in Fig. 1.19. Note, however, that CaC0 3
and other carbonates decompose at low temperatures (llOOK). Breakdown
for these particles may be due to the shock heating of the surrounding air
or the subsequent heating of the solid oxide residue.
300 I I I
t
• BROMOFORM
.& PYRIDINE
e DISTILLED WATER
rt
200 r- -
a:
0
~ r !! !
a:
! 1 -
100 r- II 1
J: •
I I
I • I
• I
0 I I I I
0 10 20 30 40 50
PARTICLE RADIUS r (llm)
Figure 1.22 Particle size dependence of the breakdown threshold ratio R at

0.53 pm for bromoform (n = 1.598), pyridine (n = 1.510), and water (n = 1.333),
from Chylek et al. (1986).
Laser-induced breakdown of water aerosols has been observed at 10.6

p,m (Lencioni, 1973), 3.8 p,m (Amimoto et al., 1984), 1.06 p,m (Mamonov,
1984), 0.69 pm (Pogodaev and Rozhdestvenskii, 1979), and 0.53 pm (Chylek
et al., 1986). The phenomenology is different at 10.6 and 3.8 p,m where wa-
ter is more strongly absorbing the laser radiation than at the shorter wave-
lengths. The breakdown threshold at 10.6 p,m was found to decrease from
3 x 109 W/cm2 to 8 x 108 W/cm2 as the water particle radius increased from
0.5 to 15 p,m. Vaporization of the droplet will generate a strong shock in air
and it is probable that breakdown occurred in the shocked air rather than
the vaporized water. Breakdown at the shorter wavelengths, however, may
occur inside the particles in regions of localized field enhancement that can
be calculated from Mie theory or outside the particle (Hsieh et al., 1987).
The presence of a water particle with 15 < R < 45 p,m was found to lower
the breakdown threshold in air at 0.53 p,m by a factor of 30. Note that from
geometric optics, one calculates that the focusing of paraxial rays will occur
in air on the shaded side of the particle. One expects breakdown for large
particles (R > 0.1 mm) to occur in air rather than inside the particle at a
distanced = [(1- 0.5n )/(n -1)]R behind the shaded edge of the particle.
For water, n = 4/3, and focusing occurs at d ::::::R. The effect of index of re-
fraction, particle size, and particle composition on breakdown thresholds
at 0.53 p,m ( rP = 20 nsec) have been studied by Chylek et al. (1986) for a
58 Weyl
series of organic liquids. The data indicate that the threshold, in general,
decreases as the index of refraction increases and also decreases with in-
creasing radius for r > 10 J.tm. The ratio of clean air breakdown thresholds
to aerosol-induced breakdown thresholds for several aerosols is shown in
Fig. 1.22.
1.5 CONCLUDING REMARKS
We have presented in this chapter the state of our understanding of laser-

induced breakdown in solids, liquids, and gases. The growth of electron
density at long wavelengths (.A> 1 J.tm) occurs mainly by electron avalanche
starting from a few originating electrons. The importance of the seed elec-
trons has been clearly demonstrated by Yablonovich (1974, 1975) in very
pure gases devoid of aerosols. Multiphoton ionization (MPI) will provide,
at shorter wavelengths (.A < 1 J.tm), enough seed electrons that the ob-
served thresholds are insensitive to the purity of the gas considered. The
inverse bremsstrahlung (IB) absorption coefficient that determines cascade
growth rates scales as .A3 [1- exp( -he/.Ak8 T)]. Cascade breakdown thresh-
olds should therefore scale as .A -J or .A- 2 The fact that the observed thresh-
olds of many gases peak near .A = 1 J.tm is a clear indication that MPI and
photoionization of excited states play an important role in lowering the
thresholds at short wavelengths.
An important finding (Rosen and Weyl, 1987) is that breakdown in gases
at near-atmospheric pressure always occurs in a regime where some cas-
cade growth of electrons takes place, even though MPI by itself would cause
complete ionization in the focal volume at flux densities below the observed
thresholds. This may well be due to the diagnostic used for measurement of
breakdown. Cascade breakdown is a catastrophic process that leads to sig-
nificant electron heating and ionization spreading beyond the focal volume.
Therefore, if the diagnostic involves measurement of appreciable absorp-
tion of laser radiation and/or the appearance of a bright flash, one may well
have to be operating in a regime where the cascade growth of electrons is
important. This would explain the extremely high thresholds observed by
Dewhurst (1977, 1978) at 1.06 J.tm with nanosecond and picosecond pulses,
and also the slope near one of log/ vs.logp. Rather than relying on optical
signals to determine the occurrence of breakdown, one should preferably
use an acoustic signal probe to measure energy deposition in the gas and
relate acoustic signal to absorbed energy. This technique is very sensitive
and could be used to obtain information on electron growth mechanisms
below threshold.
Laser-induced breakdown of transparent liquid and solid dielectrics is
not so well understood as the breakdown of gases. The measured thresh-
olds are quite sensitive to material quality and preparation and to the size
of the focal volume. Theoretical work has been limited to few materials
(alkali halides, quartz, ... ) and, even for these materials, not enough is
known about the electron-phonon-photon interactions for quantitative pre-
dictions. The source of the initial electrons is also not well understood. Bet-
ter diagnostics, such as acoustic measurement of energy deposition (Patel
and Tam, 1981 ), could be used successfully to measure the source of the ini-
tial electrons. Jones et a!. (1987), e.g., used an acoustic probe to show that
MPI is the cause of electron generation in NaCl at 0.53 flm, since they mea-
sured an / 4 dependence of energy absorption. More measurements of this
type should be carried out on other materials, with other parameters, such
as pulse length and focal volume, as variables. Solids and liquids present
the added complication that nonlinear effects such as beam self-focusing,
stimulated Raman and Brillouin scattering often occur simultaneously with
breakdown, so that it is not straightforward to separate the effects.
Since MPI plays such as an important role in breakdown at short wave-
length(.\< 1 flm) breakdown experiments, the temporal characteristics of
the beam used will affect the measured thresholds. Also, if several trans-
verse modes are present, one can expect hot spots in the beam near focus,
i.e., regions where the intensity can be up to a factor of 10 times higher than
the spatially averaged intensity at focus (Grey-Morgan, 1975). The pres-
ence of hot spots will result in anomalously low apparent thresholds. In
many of the papers that we have reviewed, not enough information on tem-
poral and spatial beam characteristics is provided. For this reason, caution
should be used in extrapolating the data to different lasers and different
optical configurations.
REFERENCES
Abrikosova, I. and Bochkova, 0. (1%5). Breakdown of liquid and gaseous helium

by a laser beam and observation of stimulated Mandel'shtam scattering in liquid
helium, Sov..Phys. JETP Lett., 9: 167-169.
Abrikosova, I. and Anshukov, B. (1971). Role of self-focusing of laser radiation in
liquid He 4 , Sov. Phys. JETP, 37: 579-581.
Agostini, P., Barjot, G., Mainfray, G., Manus, C., and Morellec, J. (1968). Multi-
photon ionization of hydrogen and rare gases, IEEE J. Quant. Electron., QE-4:
667-669.
Alcock, A and Richardson, M. (1968). Creation of a spark by a single subnanosec-
ond laser pulse, Phys. Rev. Lett., 21: 667-670.
Alcock, A, Kato, K., and Richardson, M. (1972). New features of laser induced gas
breakdown in the ultraviolet, Opt. Commun., 6: 342-344.
60 Weyl
Aleshkevich, V.A., Akhmanov, S., Zhdanov, B., Kuznetsov, B., and Sukhorukov, A.
(1976). Frequency characteristics in optical breakdown of transparent solid di-
electrics by nanosecond laser pulses, Sov. Phys. Tech. Phys., 21: 975-979.
Amimoto, S., Whittier, J., Ronkowski, F., Valenzuela, P., Harper, G., Hofland, R.,
Trusty, G., Cosden, T., and Leslie, D. (1984). DF laser pulse breakdown induced
by maritime aerosols, AIAA J., 22: 1108-1114.
Anisimov, S. (1%8). Vaporization of metal absorbing laser radiation, Sov. Phys.
JETP, 27: 182-183.
Anisimov, S.l. and Melshantsev, B.l. (1973). Role of absorbing inhomogenities in
the optical breakdown process, Sov. Phys. Solid State: 743-745.
Arushanov, S., Kosolobov, V., Koldunov, V., and Mizin, V. (1981). Temperature de-
pendence of optical breakdown thresholds of transparent liquid dielectrics, Bull.
Acad. Sciences, USSR, Series Physics, 45: 183-188.
Armstrong, R., Lucht, R., and Rawlins, W (1983). Spectroscopic investigation of
laser initiated low pressure plasmas in atmospheric gases,Appl. Opt., 22: 1573-
1577.
Baravian, G., Godart, J., and Sultan, G. (1982). Multi photon ionization of molecular
nitrogen by a Neodymium-glass laser, Phys. Rev.,A25: 1483-1495.
Barchukov, A.l., Bunkin, F.V., Konov, V.I., and Lyubin, A.A. (1974). Investigation
of low-threshold gas breakdown near solid targets by C0 2 laser radiation, Sov.
Phys. JETP, 39: 469-477.
Bass, M. and Barrett, H. (1972). Avalanche breakdown and the probabilistic nature
of laser induced breakdown, IEEE J. Quant. Electron., QE-8: 338-343.
Batenin, V. and Chinov, V. (1972). Electron Bremsstrahlung in the field of argon or
helium atoms, Sov. Phys. JETP, 34: 30-33.
Bebb, H. and Gold, A. (1965). Multiphoton ionization of hydrogen and rare gas
atoms, Phys. Rev., 143: 1-24.
Bessarab, A.V., Romanov, V.M., Samylin, V.A. and Funtikov, A.l. (1978). Shielding
time of a surface irradiated by a C0 2 laser, Sov. Phys. Tech. Phys., 23: 995-996.
Bloem bergen, N. (1974). Laser induced electric breakdown in solids. IEEE J. Quant.
Electron., QE-10: 375-386.
Bohm, D. (1951). Quantum Theory, Prentice Hall, N.J. Chap. 12.
Bondarenko, A.V., Golubev, V.S., Dan'shchikov, E.V., Lebedev, F.V., Nastoyashchii,
A.V., and Ryazonov, A.V. (1979). Laser breakdown of air near a target, Sov. Phys.
Tech. Phys. Lett., 5: 87-88.
Boni, A.A. and Meskan, D.A. (1975). Effect of impurities on laser induced break-
down at 1.06 J-Lm, Opt. Commun., 14: 115-117.
Boyer, K. and Rhodes, C. (1985). Atomic inner shell electrons excited by coherent
motion of outer shell electrons. Phys. Rev. Lett., 54: 1490-1493.
Brauer, S. (1979). Phenomenological study of laser damage in insulators, Phys. Rev.,
B20: 3422-3440.
Brown, R.T. and Smith, D.C. (1972). Laser induced gas breakdown in the presence
of preionization, AppL Phys. Lett., 22: 245-247.
Bunkin, V.F. and Savranskii, V.V. (1974). Optical breakdown of gases induced by
thermal explosion of suspended macroscopic particles, Sov. Phys. JETP, 38: 1091-
1096.
Buscher, H., Tomlinson, R., and Damon, K. (1965). Frequency dependence of opti-
cally induced breakdown, Phys. Rev. Lett., 15: 847-849.
Butenin, A and Kogan, B. (1971). Mechanism of optical breakdown in transparent
dielectrics, Sov. J. Quant. Electron., 1: 561-562.
Byron and Pert, G. (1979). Measurement of the wavelength dependence of the
threshold of laser induced gas breakdown, J. Phys. D. AppL Phys., 12: 401-408.
Canavan, G.H. and Nielsen, R.E. (1973). Focal spot dependence of gas breakdown
induced by particulate ionization,Appl. Phys. Lett., 22: 409-410.
Canavan, G., Proctor, W, Nielson, P., and Rockwood, S. (1972). C0 2 laser air break-
down calculations, IEEE J. Quant. Electron., 8: 564-565.
Chalmeton, V. (1969). Etude de la phase initiale du claquage d'un gaz par un laser,
J. de Physique, 30: 687-699.
Chan, C.H., Moody, C.D., and McNight, W (1973). Significant loss mechanisms in
gas breakdown at 10.6 11m, J. Appl. Phys., 44: 1179-1188.
Chylek, P., Jarzembski, M., Chou, N., and Pinnick, R. (1986). Effect of size and ma-
terial of liquid spherical particles on laser induced breakdown,Appl. Phys. Lett.,
49: 1475-1477.
Crance, M. (1984). Calculation of resonant multi photon processes, in Multiphoton
Ionization ofAtoms (S. Chin and P. Lambropoulos, eds.), Academic Press, New
York, pp. 65-108.
Dalgarno, A and Lane, N. (1966). Free-free transitions of electrons in gases,Astro-
phys. J., 145: 623-633.
Damon, E. and Thomlinson, R. (1963). Observation of ionization of gases by a ruby
laser, AppL Opt., 2: 546-547.
Danilychev, V.A., Zvorykin, V.D., Kholin, LV. and Chugunov, A.Y. (1981). Investi-
gation of the dynamics of plasma formation near a target acted by microsecond
C0 2 laser pulses, Sov. J. Quant. Elect., 10: 1518-1521.
Delane, N.B. (1975). Multi photon ionization of atoms, Sov. Phys. Usp., 18: 169-189.
Dewhurst, R. (1977). Breakdown in the rare gases using single picosecond ruby
pulses, J. Phys. D. Appl. Phys., 10: 283-289.
Dewhurst, R. (1978). Comparative data on molecular gas breakdown thresholds in
high-laser radiation fields. J. Phys. D. Appl. Phys., 11: 191-195.
Docchio, F., Dossi, L., and Sacchi, C. (1986). Q-switched Nd: YAG laser irradiation
of the eye and related phenomena: an experimental study. I. Optical breakdown
determination for liquids and membranes, Lasers in the Life Sciences, 1: 87-103.
Dymaev, K.M., Manenkov, A.A., Malsyukov, A.D., Matyushin, G.A., Nechitailo,
V.S., and Prokorov, A.M. (1983). Transparent polymers: a new class of optical
materials for lasers, Sov. J. Quant. Electron., 13: 503-507.
Epifanov, AS. (1974). Avalanche ionization induced in solid transparent dielectrics
by strong laser pulses, Sov. Phys. JETP, 40: 897-902.
Epifanov, A.S., Manenkov, A.A., and Prokhorov, A.M. (1976). Theory of avalanche
ionization induced in dielectrics by an electromagnetic field, Sov. Phys. JETP, 43:
377-382.
Fradin, D.W, Yablonovich, E., and Bass, M. (1973a). Confirmation of an electron
avalanche causing laser-induced bulk damage at l.0611m,Appl. Opt., 12: 700-709.
62 Weyl
Fradin, D.W., Bloembergen, N., and Lettelier, J.P. (1973b). Dependence of laser-
induced breakdown field strength on pulse duration, Appl. Phys. Lett., 22: 635-
637.
Frost, L. and Phelps, A (1962). Rotational excitation and momentum transfer cross-
sections for electrons in H 2 and N2 from transport coefficients, Phys. Rev., 127:
1621-1633.
Gamal, Y. and Harith, M. (1983). Secondary ionization processes in laser-induced
cascade ionization, J. Phys. D. Appl. Phys.: 1901-1906.
Geltman, S. (1973). Free-free radiation in electron-neutral atom collisions, J. Quant.
Spectr. Rad. Trans., 13: 601-613.
Geltman, S. (1985). Multiple ionization of a Hartree atom by intense laser pulses,
Phys. Rev. Lett., 54: 1909-1911.
Glass, A.J. and Guenther, A.H. Proc. of the Annual Conf. on Laser Damage (Amer-
ican Society for Testing Materials, Philadelphia, Pa., 1969), Spec. Publ. 469; Nat.
Bur. Stand. (U.S.) Spec. Publ. 341 (1970), 372 (1972), 387 (1973), 414 (1974),
435 (1975), 462 (1976), 509 (1977), 541 (1978), 568 (1979), 688 (1985).
Glebov, I.B., Efimov, O.M., Petrovskii, G.T., and Rogovstev, R.N. (1984). Influence
of the mode composition of laser radiation in the optical breakdown of solids,
Sov. J. Quant. Electron., 14: 226-229.
Gnoevoi, Y.N., Petrukhin, AI., Pleshatin, Y.E., and Sulyaev, V.A. (1970). Exper-
imental investigation of the appearance of screening in lead and aluminum,
ZhETF Pis. Red., 11: 440-443.
Golub, A.P. and Nemchinov, I. V. (1980). Time of appearance of a plasma due to
action of laser radiation of different wavelengths on aluminum target in air, Sov.
J. Quant. Electron., 10: 1058-1060.
Gontier, Y. and Trahin, M. (1968). Multiphoton ionization of atomic hydrogen in
the ground state, Phys. Rev., 172: 83-87.
Gorshkov, B.G., Danileiko, Y.K., Epifanov, AS., Lobachev, V.A., Manenkov, A.A.,
and Sidorin, A V. (1977). Laser induced breakdown in alkali halides, Sov. Phys.
JETP, 45: 612-618.
Grey-Morgan, C. (1975). Laser induced breakdown of gases, Rep. Prog. Phys., 38:
621-665.
Gurevich, A and Pitaevskii, L. (1962). Recombination coefficient in a dense low
temperature plasma, Sov. Phys. JETP, 19: 870-871.
Hill, G.A., James, D.J., and Ramsden, S.A. (1971). Breakdown thresholds in rare
and molecular gases using pulsed 10.6 micron radiation, J. Phys. D. Appl. Phys.,
5: L97-99.
Holway, L. and Fradin, D. (1975). Electron avalanche breakdown by laser radiation
in insulating crystals, J. AppL Phys., 46: 279-291.
Hsieh, W.F., Eickmans, J.H., and Chang, R.K. (1987). Internal and external laser
induced avalanche breakdown in an argon atmosphere,/. Opt. Soc. Am. B4, 1816-
1820.
Huber, K.P. and Herzberg, G. (1979). Molecular Spectra and Molecular Structure W,
Constants ofDiatomic Molecules, Van Nostrand Reinhold, New York, p. 506.
Hunklinger, S. and Leiderer, (1971). Influence of impurities on the laser-induced
breakdown of liquid He4, Zs. Naturforsch., 26a: 587-588.
Ioffe, AI., Mel'nikov, N.A, Naugol'nykh, K.A, and Upacyshev (1970). Shockwave
from optical breakdown in water. Zh. PrikL Mekh. Tekh. Fiz., 3: 125-127.
Ireland, C. and Grey-Morgan, C. (1974). Gas breakdown by"' 20 ps, 1.06 Jim and
0.53 Jim laser pulses, J. Phys. D. AppL Phys., 7: L87-90.
Itzkan, I. and Cronberg, T.L. (1981). Laser plasma ignition threshold behavior of
prepared aluminum surfaces at 10.6 Jim, AIAAPaper 81-1151, AIAA 16th Ther-
mophysics Conf., Palo Alto, Calif.
Jones, S.C., Shen, X.A, Braunlich, P.F., Kelly, P., and Epifanov, AS. (1987). Mech-
anism of prebreakdown nonlinear energy deposition for intense photon fields at
532 nm in NaCI, Phys. Rev. B 35: 894-896.
Kas'yanov, V. and Starostin, A (1965). On the theory of bremsstrahlung of slow
electrons on atoms, Sov. Phys. JETP, 21: 193-198.
Keldysh, T. (1965). Ionization in the field of a strong electromagnetic wave, Sov.
Phys. JETP, 20: 1307-1314.
Kieffer, L.J. (1973). A compilation of electron collision cross section data for mod-
eling gas discharged lasers, llLA Information Ctr. Rep. 13, Joint Institute for
Laboratory Astrophysics, Univ. of Colorado, Boulder, Colo., Sept.
Konov, V.I., Lavrent'ev, Yu.V., Stepanov, Yu.l., Chapliev, N.l., and Shirkov, AV.
(1983). Dependence of the threshold of formation of an air breakdown plasma
near Cu mirrors on the absorption coefficient of their surface at 10.6 Jim, Sov. J.
Quant. Electron., 13: 1128-1130.
Kozlova, N.N., Petrukhin, AI., and Sysev, V.A. (1975). Experimental investigation
of the onset of evaporation and formation of a plasma layer due to interaction
between laser radiation and metals in different gases, Sov. J. Quant. Electron., 5:
747-749.
Krasyuk, I. and Pashinin, P. (1972). Breakdown in Ar and N2 under the influence of
a 0.35 Jim picosecond laser pulse, JETP, Lett., 15: 333-334.
Krasyuk, 1., Pashinin, P., and Prokhorov, A (1969). Investigation of breakdown in
N2 under the influence of a picosecond ruby laser pulse, JETP Lett., 9: 354-356.
Krasyuk, 1., Pashinin, P., and Prokhorov, A (1970). Investigation of breakdown in
argon and helium produced by a picosecond ruby laser light pulse, Sov. Phys.
JETP, 31: 860.
Kroll, N. and Watson, K. (1972). Theoretical study of ionization of air by intense
laser pulses, Phys. Rev., A5: 1883-1905.
Lambropoulos, D. (1976). Topics on multiphoton processes in atoms, Advances in
Atomic and Molecular Physics, Vol. 12 (D.R. Bates and B. Bederson, eds.), Aca-
demic Press, New York, pp. 87-164.
Lambropoulos, D. (1985). Mechanisms for multiple ionization of atoms by strong
laser pulses, Phys. Rev. Lett., 55: 2141-2144.
LeCompte, C., Mainfray, G., Manus, C., and Sanchez, F. (1974). Experimental
demonstration of laser temporal coherence effects on multiphoton ionization
processes, Phys. Rev. Lett., 32: 265-268.
Lencioni, D.E. (1973). The effect of dust on 10.6 Jim laser induced air breakdown,
Appl. Phys. Lett., 23: 12-14.
Lencioni, D.E. (1974). Laser induced breakdown for 10.6 Jim radiation,Appl. Phys.
Lett., 25: 15-17
64 Weyl
Lencioni, D.E., Petingill, L.C., and DeGloria, D.P. (1976). Air breakdown initiated
by particles in an HF laser beam, Rep. LTP-35, MIT Lincoln Lab., Lexington,
Mass.
I.:Huillier, A and Mainfray, G. (1985). Multiphoton ionization and dissociation of
diatomic molecules irradiated by an intense 40 ps laser pulse, inPhotophysics and
Photochemistry (F. Lahmani, ed.), Elsevier, Amsterdam, pp. 93-101.
I.:Huillier. A, Lompn;, L., Mainfray, G., and Manus, C. (1983). Multiply charged
ions induced by multiphoton absorption in rare gases at 0.53 Jtm, Phys. Rev.,A27:
2503-2512.
I.:Huillier, A, Mainfray, G., and Johnson, P.M. (1984). Multiphoton ionization ver-
sus dissociation of diatomic molecules irradiated by an intense 40 ps laser pulse,
Chern. Phys. Let., 103: 447-450.
Lompre, L.A., Mainfray, G., Mathieu, B., Watel, G., Aymer, M., and Crance, M.
(1980). Absolute determination of two- and three-photon ionization cross sec-
tions of singlet and triplet 2S metastable helium, J. Phys. B., 13: 1799-1816.
Lompre, L., I.:Huillier, A, Mainfray, G., and Manus, C. (1985). Multiphoton ioniza-
tion of He atoms at 532 nm, Phys. Lett., 112A: 319-322.
Lowder, J.E. and Kleiman, H. (1973). Long pulse breakdown with 10.6 ~tm laser
radiation, J. Appl. Phys., 44: 5504-5505.
Lowke, J., Phelps, A, and Irwin, B. (1973). Predicted electron transport and oper-
ating characteristics of C0 2 , N 2 , He laser mixtures, J. Appl. Phys., 44: 4664-4671.
Lucatorto, T.B. and Mcilrath, T.J. (1976). Efficient laser production of a Na+
ground-state plasma column: Absorption spectroscopy and photoionization
measurement of Na+, Phys. Rev. Lett., 37: 428-431.
Luk, T., Pummer, H., Boyer, K., Shakidi, M., Egger, H., and Rhodes, C. (1983).
Anomalous collision-free multiple ionization of atoms with intense microsecond
ultra violet radiation, Phys. Rev. Lett., 51: 110-113.
Lysikov, Yu.l. (1978). Role of thermal stresses in optical breakdown of transparent
dielectrics, Fiz. and Khim. Obrav. Mater. J.: 31-36.
Maher, W., Hall, R., and Johnson, R. (1974). Experimental study of ignition and
propagation of laser supported detonation waves, J. Appl. Phys., 45: 768-781.
Mainfray, G. and Manus, C. (1984). Normal multiphoton ionization of atoms (exper-
imental), Multiphoton Ionization ofAtoms (S. Chin and P. Lambropoulos, eds. ),
Academic Press, New York, pp. 7-34.
Mamonov, V.K. (1984). Air minimization by breakdown in aqueous aerosol drops
acted upon by radiation with>.= 1.06 Jtm, J. Sov. Laser Res., 5: 249.
Manenkov, A.A. and Nechitailo, V.S. (1980). Role of absorbing defects in laser dam-
age to transparent materials, Sov. J. Quant. Electron., 10: 347-439.
Marcus, S., Lowder, J.E., and Moody, D.L. (1976). Large spot thermal coupling of
C0 2 laser radiation to metallic surfaces, J. Appl. Phys., 47: 2966-2968.
Marcuse, D. (1972). Light Transmission Optics, Van Nostrand Reinhold, New York,
p. 232.
McMordie, J.A. and Roberts, P.D. (1975). The interaction of pulsed laser radiation
with aluminum, J. Phys. D. Appl. Phys., 8: 768-781.
Measures, R. and Cardinal, M. (1980). Laser ionization based on resonance satura-
tion-a simple model description, Phys. Rev. A, 23: 804-815.
Meyerand, R.G. and Haught, A.R (1963). Gas breakdown at optical frequencies,
Phys. Rev. Lett., 9: 401-403.
Migus, A, Gauduel, Y., Martin, J., and Antonetti, A (1987). Excess electrons in liq-
uid water: first evidence of a prehydrated state with femtosecond lifetime, Phys.
Rev. Lett., 58: 1559-1562.
Minck, R.W. (1964). Optical frequency electrical discharges in gases, J. Appl. Phys.,
35: 252-254.
Morton, V. (1967). Multiphoton absorption in monatomic gases. Proc. Phys. Soc.
(London) 92: 301-309.
Mu, X.- D., Aberg, T., Blomberg, A., and Crasemann, B. (1986). Production of mul-
tiply charged ions by strong UV laser pulses: Theoretical evidence for stepwise
ionization, Phys. Rev. Lett., 56: 1909-1919.
Nichols, D.B., and Hall, R.B. (1978). Threshold condition for the formation of sur-
face plasma by HF and DF radiation, J. Appl. Phys., 49: 5155-5164.
Nighan, W. (1970). Electron energy distributions and collision rates in electrically
excited N 2 , CO, and C02 , Phys. Rev.,A2: 1989-1999.
Normand, D. and Morellec, J. (1980). Absolute determination of two-and-four-
photon ionization cross sections of caesium atoms, J. Phys. B., 13: 1551-1561.
O'Connell, R., Ronbergh, A, Shaffer, A., Saito, T., Deaton, T., and Seigenthaler, K
(1984). Improved laser-damage-resistant polymethyl methocrylate, J. Opt. Soc.
Am., 1B, 853-856.
Pantell, R.H. and Putthof, H.E. (1969). Fundamentals of Quantum Electronics, Wi-
ley, New York, p. 255.
Patel, C.K and Tam, A.C. (1981). Pulsed optoacoustic spectroscopy of condensed
matter, Rev. Mod. Phys., 53: 517-550.
Phelps, A V. (1966). Theory of Growth of ionization during breakdown, in Physics of
Quantum Electronics (P. Kelley, B. Lax, and P. Tannenwald, eds.), McGraw-Hill,
New York, pp. 538-547.
Pogodaev, V.A. and Rozhdestvenskii, A.E. (1979). Initiation of optical breakdown
in air by weakly absorbing water droplets, Sov. Tech. Phys. Lett., 5: 103-104.
Pozhidaev, V.N. and Fatievskii, AI. (1981). Optical breakdown thresholds in liquid
water and micron size water droplets exposed to single laser pulses, Sov. J. Quant.
Electron., 11: 65-68.
Raizer, Yu.P. (1965). Heating of a gas by a powerful laser pulse., Sov. Phys. JETP,
21: 1009-1017.
Raizer, Yu.P. (1977). Laser Induced Discharge Phenomena, Consultants Bureau, New
York.
Ready, J.R (1971). Effects of High-Power Laser Radiation, Academic Press, New
York.
Reilly, J., Singh, P., and Weyl, G. (1977). Multiple pulse propagation through atmo-
spheric dust at 10.6 microns, AIAA Paper 77-697, AIAA lOth Fluid and Plasma
Dynamics Conf., Albuquerque, New Mexico, June 27-28.
Roach, J.F. and Davies, J.M. (1970). Electric strength of some liquid dielectrics sub-
jected to a Q switched laser pulse, Proc. ofIEEE: 1388-1390.
Root, R. and Pirri, A (1979). Long time laser induced breakdown of particle con-
taminated air, AIAA Paper 79-0248, AIAA 17th Aerospace Sciences Meeting,
New Orleans, La, Jan. 15-17.
66 Weyl
Rosen, D. and Weyl, G. (1987). Laser induced breakdown in N2 and the rare gases
at 0.53 and 0.35 J.Lm, 1 Phys. D. AppL Phys., 20: 1264-1276.
Rosen, D., Hastings, D., and Weyl,G. (1982a). Coupling of pulsed 0.35 J.Lm radiation
to titanium alloys, 1 AppL Phys., 53: 5882-5890.
Rosen, D.l., Mitteldorf, J., Kothandaraman, G., Pirri, AN., and Pugh, E.R. (1982b ).
Coupling of pulsed 0.35 J.Lm laser radiation to aluminum alloys, 1 Appl. Phys., 53:
3190-3200.
Schlier, R.E., Pirri, AN., and Reilly, D.J. (1973). Air breakdown studies, Rep.
AFWL-TR-72-74, AVCO Everett Res. Lab. Everett, Mass.
Seitz, F. (1949). On the theory of electron multiplication in crystals, Phys. Rev., 76:
13}6-1393.
Shen, Y.R. (1975). Self focusing: Experimental, in Progress in Quantum Electronics,
Volume 4 (J.J. Sanders and S. Stenholm, eds.), Pergamon Press, London, pp. 1-
34.
Shockley, W. (1961). Problems related to P-N junctions in silicon, Solid State Elec-
tron., 2: 35.
Smith, D. (1977). Gas breakdown initiated by laser radiation interaction with
aerosols and solid surfaces, 1 Appl. Phys., 48: 2217-2225.
Smith, D.C. (1970). Gas breakdown with 10.6 J.Lm wavelength C0 2 laser radiation,
1 AppL Phys., 41: 4501-4505.
Smith, D.C. and Meyerand, R.G. (1976). Laser radiation induced gas breakdown,
in Principles ofLaser Plasma (G. Bekefi, ed.), Wiley, New York, pp. 457-507.
Smith, W. Lee (1978). Laser induced breakdown in optical materials, Opt. Engineer-
ing, 17: 489-503.
Soileau, M.J., Bass., and Van Stryland, M. (1978). Frequency dependence of break-
down fields in single crystal NaCl and KCl, Nat. Bur. Stand. Spec. Publ. 541:
309-317.
Sparks, M. and Duthler, M.J. (1973). Theory of infrared absorption and material
failure in crystals containing inclusions, 1 Appl. Phys., 44: 3038-3045.
Sparks, M., Holstein, T., Warren, R., Miller, D.L., Maradudin, A.A., Sham, L.J.,
Loh, E., and King, F. (1977). Theory of electron-avalanche breakdown in solids,
in Laser Damage to Optical Materials, (A. Guenther, ed.), NBS-568, Government
Printing Office, Washington, D.C., pp. 467-476.
Steverding, B. (1974). Ignition of laser supported detonation waves, 1 Appl. Phys.,
45: 3507-3511.
Stricker, J. and Parker, G. (1982). Experimental investigation of electrical break-
down in nitrogen and oxygen induced by focused laser radiation at 1.064 J.Lm, 1
Appl. Phys., 53: 851-855.
Touloukian, Y.S. (1967). Thermophysical Properties ofHigh Temperature Solid Mate-
rials, Vol. 4, MacMillan, New York, pp. 1659-1661.
Uglov, A.A. and Galiev, A.L. (1980). Evolution of a laser plasma near a solid surface
in high pressure gases, Sov.1 Quant. Electron., 10: 143-145.
Van Styland, E.W., Soileau, M.J., Smirl, A.L., and Williams, W.E. (1981). Pulse
width and focal volume dependence of laser-induced breakdown, Phys. Res. B,
23: 2144-2146.
Voltz, F.E. (1972). Infrared refractive index of atmospheric aerosol substances,Appl.
Opt., 11: 755-759.
Voronov, G. (1967). Dependence of the probability of multiphotoionization of

atoms on the photon flux intensity, Sov. Phys. JETP, 24: 1009-1011.
Walters, C.J., Barnes, R.H., and Beverly, III, R.E. (1978). Initiation of laser sup-
ported detonation waves, J. Appl. Phys., 49: 2937-2949.
Weyl, G. and Shui, V. (1977). Laser supported detonation waves in gases, Proc. of
11th International Symp. on Shock Tubes and Waves, Univ. of Washington Press,
577-587.
Weyl, G. (1979). Ionization path formation in gases using a laser with a retractable
focus, J. Phys. D. AppL Phys.,'12: 33-49.
Weyl, G., Pirri, A, and Root, R. (1981). Laser ignition of plasma off aluminum sur-
faces, AL4A !., 19: 460-469.
Weyl, G., Rosen, D., Wilson, J., and Seka, W. (1982). Laser induced breakdown of
argon at 0.35 Jlm, Phys. Rev. A26: 1164-1166.
Weyl, G. and Rosen, D. (1985). Laser induced breakdown in argon at 0.35 Jlm. The-
ory and experiments, Phys. Rev.,A31: 2300-2313.
Williams, W., Soileau, M., and Van Stryland, E. (1985). Picosecond air breakdown
at 0.53 Jlm, Appl. Phys. Lett., 43: 352-354.
Yablonovich, E. (1973). Similarity principles for laser induced breakdown, AppL
Phys. Lett., 23: 121-122.
Yablonovich, E. (1974). Self phase modulation of light in a laser breakdown plasma,
Phys. Rev. Lett., 32: 1101-1104.
Yablonovich, E. (1975). Self phase modulation and short pulse generation from laser
breakdown plasmas, Phys. Rev., AJO: 1888-1895.
Yablonovich, E. and Bloem bergen, N. (1971). Avalanche ionization and the limiting
diameter of filaments induced by light pulses in transparent media, Phys. Rev.
Lett., 29: 907-910.
Yamabe, C., Buckman, C.S., and Phelps, A.V. (1982). Measurement of free-free
emission from low-energy electron collisions with argon, Phys. Rev.,A27: 1345-
1352.
Zel'dovich, Y.B. and Raizer, Y.D. (1965). Cascade ionization by a light pulse, Sov.
Phys. JETP, 20: 772-780.
2
Modeling of Post-Breakdown
Phenomena
Robert G. Root
SPARTA, Inc.
Lexington, Massachusetts
2.1 INTRODUCTION
Between the creation of a localized vapor plasma, described in Chap. 1,

and its culmination in a steady-state laser-sustained plasma, described in
Chap. 4, a plasma evolves through several transient phases. These tran-
sient plasmas are most easily investigated in pulsed laser experiments in
which the laser absorption process can be terminated during any stage of the
plasma evolution. The plasmas generated during pulsed laser interactions
with materials are also of interest in their own right, as a method of impart-
ing energy, pressure, and impulse to the target. The nature of the plasma
and the resulting effect on the surface are functions of the ambient atmo-
sphere, irradiance, wavelength, and duration of irradiation. Nevertheless,
the plasmas can be broadly categorized into different types that display dra-
matically different characteristics; each type is associated with a restricted
range of interaction parameters, but together they span the complete inter-
action space. This chapter describes the types of plasmas observed when a
pulsed laser strikes a material.
Since several different types of plasmas are possible, our approach is to
discuss the basic characteristics of each plasma type, the similarities and
differences among them, and the salient features of the plasma interaction
with the target surface. Plasma development has been investigated with
both multidimensional hydrodynamic codes and analytical models. Irre-
spective of the modeling approach employed, the most important element
for representing these plasmas is the inclusion of the appropriate physics.
Since the physics is more transparent in the analytic models, and the dif-
ferent interaction regimes are more readily apparent, we will use analytic
models to illustrate the features of the various plasma types.
69
70 Root
Our emphasis is on the three types of plasmas observed in interactions

in which the ambient atmosphere is capable of sutJporting a propagating
plasma; plasmas restricted to the target products, often termed vacuum
plasmas, are mentioned only briefly. (A special case of the vacuum plasma is
thoroughly discussed in Chap. 5, inertially confined fusion.) Furthermore,
among the three propagating plasmas, we focus primarily on the plasmas
created at relatively low irradiance (a few MW/cm 2) at target surfaces and
that evolve into the laser-sustained plasmas of Chap. 4. Analytic models
for the various plasma types are described briefly since they provide the
basic interpretation for the trends observed as the interaction parameters
are varied; more detailed descriptions of these models and the numerical
simulations are available in the referenced papers.
2.2 CREATION OF A PROPAGATING PLASMA
As discussed in the preceding chapter, the initiation of a plasma over a tar-

get surface begins in the hot target vapor. Absorption generally commences
via electron-neutral inverse bremsstrahlung, but when sufficient electrons
are generated, the dominant laser absorption mechanism makes a transition
to electron-ion inverse bremsstrahlung. Photoionization of excited states
can also contribute for short wavelength interactions. The same absorp-
tion processes are also responsible for the absorption by the ambient gas;
they will be described in more detail later. The change in the dominant ab-
sorption mechanism characterizes the switch from the tenuous absorption
processes, which control the initiation times and threshold irradiance, to
a strongly absorbing plasma, which rapidly attains high temperature. Ex-
cept near the threshold irradiance for creating the surface vapor plasma,
vapor absorption and the resulting temperature increase are so rapid that
significant expansion of the hot vapor does not occur prior to the formation
of a hot totally ionized vapor plasma at a pressure substantially above the
ambient pressure.
The hot vapor plasma interacts with the surrounding atmosphere in the
two ways illustrated in Fig. 2.1: (1) the expansion of the high-pressure vapor
drives a shock wave into the atmosphere, and (2) energy is transferred to the
atmosphere by a combination of thermal conduction, radiative transfer, and
heating by the shock wave. The relative importance of these processes in
determining the subsequent plasma evolution depends on irradiance, size
of the vapor plasma bubbles, target vapor composition, ambient gas com-
position and pressure, and laser wavelength. At low irradiance, conduc-
tion dominates the early stages of plasma development in the surrounding
gas; the vapor plasma is too thin, spatially and optically, to transport en-
ergy efficiently by radiation. At high irradiance, shock heating dominates.
Post-Breakdown Phenomena 71
HOT, HIGH-PRESSURE,
STRONGLY ABSORBING
VAPOR PLASMA
TEMPERATURE
PROFILE
AMBIENT
ATMOSPHERE
SHOCK
WAVE
Figure 2.1 Features of the interaction between the vapor plasma and the ambient
gas.
The vapor plasma may form a continuous sheet of plasma across the target
surface as illustrated in Fig. 2.1, which is typical of interactions at wave-
lengths shorter than 3.8 J-Lm, or it may be restricted to a series of pockets
created from vaporization of small target imperfections, which is commonly
observed over aluminum targets irradiated at 10.6 Jlm (Walters et al., 1978).
Regardless of the details of the initiation process, the atmosphere ad-
jacent to vapor plasma is heated, thereby enabling the gases, which were
initially transparent to the laser radiation when cold, to start absorbing the
laser radiation. Once a critical number of electrons are liberated, the heated
gas layer absorbs strongly and rapidly heats to plasma conditions, following
the same heating history as the vapor. As the atmosphere begins to ab-
sorb a significant fraction of the laser energy, a self-perpetuating absorp-
tion process commences that results in plasma propagation into the sur-
rounding atmosphere; subsequent layers of gas experience the same inges-
tion process-heating initially by energy transfer from the plasma until laser
absorption is initiated in the gas, then rapidly heating by laser absorption to
produce a strongly absorbing plasma. The dominant energy-transfer mech-
anism usually changes as the plasma evolves, transforming from conduction
72 Root
to radiation as the plasma becomes optically thick in the extreme ultravio-

let. By creating absorbing gas layers in front of it, the plasma changes from
its initial confined vapor state to a fully developed propagating absorption
wave in the gas; it absorbs most of the laser energy in its advancing front,
thereby shielding the bulk of the plasma from further direct interaction with
the laser radiation. The absorption wave propagates up the laser beam until
the irradiation is either terminated or reduced to irradiance levels that can
no longer support an absorption wave.
2.3 ABSORPTION CHARACTERISTICS

OF HEATED GASES
The absorption characteristics of gases are responsible for the creation and
propagation plasmas. The key to understanding the nature of the absorp-
tion wave, as well as the transfer of energy from the vapor to the ambient
gas, is the absorption characteristics of the ambient gas. The ambient gas is
transparent to the laser radiation, whereas ionized gas can absorb the laser
energy strongly, at high pressure, ionized gas can absorb the laser radiation
in fractions of a millimeter. This rapid increase in the energy deposition rate
as a function of temperature is reminiscent of the energy-release charac-
teristics of chemical reactions in energetic materials. Consequently, laser-
supported absorption waves have many similarities with chemical combus-
tion and detonation waves.
The absorption coefficient of gases that have no molecular absorption
bands at the laser wavelength increases rapidly as they are heated. These
changes result from the creation of free electrons. In a collision-dominated
hot gas, the electron 1mmber density is determined by equilibrium among
electrons and the various ionic species, including neutral atoms. The Saha
equation relates the electron number density ne to the number densities
ni and ni _ 1 of two ions that differ by only one stage of ionization. The
relationship is
(2.1)
where ui and ui-l are the electronic partition functions of the two ion
stages; me the mass of the electron; h Planck's constant; k Boltzmann's
constant; and Mi the ionization potential of ionic species i, modified for
any lowering of the ionization potential by the high density. The electron
and ion number densities of air for 1 atm pressure and 10 atm pressure are
shown in Fig. 2.2 as a function of temperature to 30,000K. The individual
atomic species are not identified, since the feature of interest is the number
cles. In this temperature regime, where the first stage of ionization is being
populated, the rapid increase in electron density results from the exponen-
tial term in temperature in Eq. (2.1).
Laser radiation is absorbed primarily by inverse bremsstrahlung; pho-
toionization of excited states contributes only for energetic photons imping-
ing on easily ionized gas mixtures (such as aluminum vapor). The two con-
tributions to inverse bremsstrahlung, electron-neutral and electron-ion, are
important during different stages of the heating process. At temperatures
less than lO,OOOK, only a few electrons are present; hence, electron-neutral
inverse bremsstrahlung dominates. Its effective absorption coefficient k;,n
? 10
s
18
()
L.-...J
:>...
~ 10 17
Ul
~
Q)
~
H 10 16
Q)
s
..0
~
z 10 15
1 atm.
10 atm.
10 14wu~~~~~~uw~~~~~~wu~~~~~ww~~
5000 10000 15000 20000 25000 30000

Temperature [K]
Figure 2.2 Electron and ion number densities for air at one and ten atmospheres
as a function of temperature.
74 Root
is given by
k'e,n = [1-exp(-~)]
>..kT ""'Q·n
~ J e n·J
(2.2)
J
where ne and ni are the number densities of the electron and thejth neutral
atomic species; Qi is the average cross section for absorption of a photon
of wavelength >.. by an electron during a collision with species j; k is Boltz-
mann's constant; his Planck's constant; cis the velocity oflight; and Tis the
temperature. The term outside the sum accounts for stimulated emission.
Cross sections of neutral species are of the order of 10-36 cm5 • Values for
air molecular species are given by Taylor and Caledonia (1976).
Although electron-neutral inverse bremsstrahlung starts the absorp-
tion process, when the air is approximately 1% ionized, electron-ion in-
verse bremsstrahlung dominates. The effective absorption coefficient for
electron-ion inverse bremsstrahlung is
(2.3)
where zi and ni are the charge and number density of the ith ionic species,
and gi is the appropriate Gaunt factor that corrects the semiclassical ex-
pression for quantum effects. The Gaunt factor is of the order of 1; the
values for many states have been tabulated by Karzas and Latter (1961).
The absorption characteristics of air at 10.6 pm are shown in Fig. 2.3 as
a function of temperature for pressures of 1 and 10 atm. The temper-
ature regime from 6000 to 30,000K is displayed; it is in this region that
the air becomes singly ionized and the most dramatic increase in absorp-
tion occurs. Subsequent stages of ionization change the absorption coef-
ficient by less than an order of magnitude for a constant pressure. In the
regime of primary interest, where the plasma becomes singly ionized, the
absorption coefficient varies as the square of the electron concentration
that itself is increasing exponentially with temperature until ionization is
almost complete. Furthermore, in the completely ionized stage, the elec-
tron density is proportional to pressure so that the absorption coefficient
varies as the pressure is squared. Thus, atmospheric gases display an abrupt
change from weak absorption to very strong absorption at temperatures of
the order of 10,000K, and the change is magnified when the pressure of
the gas remains high. (For aluminum vapor, which has a low ionization
potential, ionization and absorption occur at temperatures of the order of
SOOOK.)
...
r--1
p 10 atm.
s
I
10
C)
L........J
-+-)
~ 1
Q)
·~
C)
·~
~ p 1 atm.
~
Q) 10 -l
0
u
§ 10 - 2
·~
-+-)
P-t
H
~ 10 -s
,..a
<r:
10 -4wu~~~~~~~~~~~~~~~~~~~~uu~
5000 10000 15000 20000 25000 30000

Temperature [K]
Figure 2.3 Absorption coefficient of air at 10.6 J.Lm for high temperature air at one
and ten atmospheres.
2.4 FEATURES OF PROPAGATING PLASMAS
In the initial stages, the absorption wave is characterized by high tempera-

ture and also by high pressure caused by the shock wave driven off the tar-
get surface during the initiation process. The subsequent evolution of the
plasma follows three major paths depending on irradiance, spot size, and
ambient gas conditions. The history of important quantities such as radia-
tive transfer, surface pressure, plasma velocity, and plasma temperature are
strongly influenced by the nature of the plasma, as is the final steady-state
nature of the plasma.
76 Root
The three major types of laser absorption wave are commonly known as
(1) laser-supported combustion (LSC) waves, (2) laser-supported detona-
tion (LSD) waves, and (3) laser-supported radiation (LSR) waves. The dif-
ferences in the waves arise from the different mechanisms used to propagate
the absorbing front into the cool transparent atmosphere. The character-
istics that figure most prominently in distinguishing the waves are velocity,
pressure, and the effect of radial expansion on the subsequent plasma evo-
lution.
The general configuration of the absorption wave is sketched in Fig. 2.4.
The sketch does not represent a specific class of absorption wave, rather, it
illustrates the zones that must be considered to describe the classes of waves
and their development in time. The zones are the precursor shock, ab-
sorption region, and plasma left behind the propagating region. The shock
and absorption wave are shown propagating away from the target surface,
whereas the plasma behind the wave is expanding radially. Radial expan-
sion of the plasma begins immediately at the edge of the plasma, but the
bulk of the plasma and absorption zone is not affected until the rarefac-
tion waves coalesce in the center of the spot at time ru. Thereafter, radial
RADIALLY
EXPANDING
SHOCK
PRECURSOR
SHOCK
PLASMA LASER BEAM
ABSORPTION
ZONE
Figure 2.4 Features of propagating laser supported plasma created above target
surface.
expansion dominates the evolution of the plasma features at the target sur-
face. Whether or not radial expansion affects the propagating absorption
zone depends on the nature of the propagation mechanism.
Since many measurements and applications are related to the properties
at the target surface, we will discuss both the one-dimensional (early time)
and three-dimensional (late time) plasmas. The major differences among
the absorption waves are most obvious in the one-dimensional plasma.
Ideal representations of the three classes of absorption waves in the one-
dimensional configurations are presented in Figs. 2.5 and 2.6. The distinct
gasdynamic regions, from the ambient atmosphere to the target (wall), are
shown in Fig. 2.5 with the absorption zone represented as rapid change or
jump in properties. Detailed views of the characteristics of the absorp-
tion zones are provided in Fig. 2.6. The variations of gas velocity, pres-
sure, temperature, and density are shown in Fig. 2.5 beneath the sketch
of each configuration; these are highly idealized representations of the ab-
sorption waves. Similarly, below the expanded view of the absorption zones
in Fig. 2.6, the profiles of laser irradiance, plasma radiation, pressure, and
temperature are displayed. These figures will be used as basic references in
the following descriptions of the properties and models for the absorption
waves.
2.5 ONE-DIMENSIONAL LASER-SUPPORTED

COMBUSTION WAVES
At low irradiance, laser-supported combustion waves are produced. Their

configuration, as shown in Fig. 2.5, is comprised of a precursor shock that is
separated from the absorption zone and plasma. The shock wave increases
the gas density, pressure, and temperature; nevertheless, the shocked gas
remains transparent to the laser radiation. The plasma is separated from
the shock wave by a region of shocked gas at constant pressure. The front
edge of the plasma, the laser absorption zone, propagates into the shocked
gas; it is called the laser-supported combustion wave. This configuration
resembles the one-dimensional chemical deftagration (combustion) wave
in a closed tube (Courant and Friedrichs, 1957).
The first analysis of laser-supported combustion waves (Razier, 1970)
examined the long-time propagation at 1 atm. Thermal conduction was
assumed to be the primary propagation mechanism. Subsequent authors
(Boni and Su, 1974; Jackson and Nielsen, 1974; Boni and Su, 1976) sug-
gested that radiative transfer could contribute. One-dimensional propaga-
tion with radiation has been studied with a variety of transport models (Root
et al., 1979; Kemp and Root, 1977). For unconfined plasmas, radial expan-
sion is often important (Keefer, 1982), and for small spots that are optically
LSC WAVE LSD WAVE LSR WAVE .....
CD
AMBIENT
GAS
(a)
~ I I ~s AM:.!sENT PLASMA
~ EXPANSION SHOCK
M .
FANS ABSORPTION
ill A
I I ZONE (LSD WAVE) m z~~VEl J
(b) 8~ vs-us
~
o..__ _ _.,L_ .I '
(c)
0.. pp I
lr-----.....1
--i----__;:::==
I !, '"
'·]'Jf ~ ~ I :
Tp
(d) Tp
To
II '· bb
(e)
I
P,
I pp
!}_·' P,
j1-P,
- i ~ P, I P, I
I
DISTANCE- DISTANCE- DISTANCE -
Fig. 2.5 Qualitative comparison of regions of one-dimensional propagating plasma produced by laser absorption waves:
~
(a) configuration; (b) gas velocity; (c) pressure; (d) temperature; and (e) density.
LSC WAVE LSD WAVE LSR WAVE .,
SHOCKED AMBIENT fVSMA AMBIENT
fVSMA fVSMA
I ABSORPTION
GAS
I ABSORPTION
WAVE
I WAVE GAS
~~~I GAS
/- ;
~
(a)
~
.,::s
::J'
!!l0
~
&II
(b)
(c)
DISTANCE- DISTANCE- DISTANCE-
.....
co
Fig. 2.6 Qualitative comparison of the laser absorption regions of one-dimensional laser absorption waves:
(a) configuration; (b) laser irradiance; (c) pressure and temperature.
80 Root
thin to the plasma radiation, conduction can be the dominant mechanism.

Current descriptions of the propagation of these waves at ambient atmo-
sphere will be described in Chap. 4.
During the transient one-dimensional stage of a pulsed interaction, how-
ever, the absorption wave propagates into a high-pressure gas created by
the precursor shock. The major mechanism causing LSC wave propaga-
tion is radiative transfer from the hot plasma to the cool high-pressure
gas created in the shock wave. The plasma reradiation is primarily in the
extreme ultraviolet; it is generated by photorecombination of electrons
and ions into the ground-state atom. A typical radiation spectrum pre-
dicted for an LSC wave in air is reproduced in Fig. 2.7 (Root and Pirri,
1979) and compared with available data (Mitchell et al., 1978). The spec-
tral region that controls propagation is at wavelengths less than 125 nm.
The cooler portion of the wave that has few electrons cannot absorb the
laser radiation, but the neutral atoms strongly absorb the plasma radia-
tion by photoionization. Both laser heating and plasma radiative trans-
port are much stronger in the high-pressure gas. The laser absorption co-
efficient increases as the pressure is squared; and both the laser heating
rate and the absorption coefficient in the extreme ultraviolet are propor-
tional to the pressure. Thus for the spot sizes typical of pulsed laser in-
teractions, the plasma is optically thick in the radial direction. After a
short transient at the beginning of the plasma development, correspond-
ing to a fluence of a few J/cm2 , the plasma is also optically thick in the ex-
treme ultraviolet in the propagation direction. Radiative transport in the
ultraviolet is typically 10% of the incident laser irradiance, which far ex-
ceeds the energy transport by conduction at irradiance of 1 MW/cm 2 and
above.
The configuration is maintained by the expansion of the hot plasma. The
fresh shocked gas ingested by the LSC wave rapidly heats; its expansion
maintains the pressure that drives the shock wave. The major features of
the LSC wave plasma can be calculated using a quasisteady approximation
in which the configuration is assumed to evolve slowly in time throughout
the one-dimensional plasma development (Pirri et al., 1978). Shock jump
conditions relate the properties of the shocked gas-density Ps, pressure
Ps' enthalpy hs, and gas velocity us behind the shock relative to the ambient
air-to the ambient gas properties p0 , p 0 , and h 0 with us = 0; strong shock
relations yield
Is+ 1
Vs = -2-us (2.4)
Is+ 1 2
Ps = -2-pous (2.5)
"0
fit
m
ii
~
.---,
......
I
E
~
"0
u :r
sr
..... / .I IJC CD
... ::I
t; 0
I 3
N CD
::I
E m
u
~
-Q)
u
c
..,
~ 'l' Air and l%Al
ISll
IIlii:: ..... Laser Intensity 5 MW/cm 2
Temperature 23500K
Pressure = 20 ATM
Slab Thickness 0.5cm
CCEBL Data = )(
,.
Sl
-t ~" I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I
0 2000111 4000111 Ei000111 8000111
Wave number [em -l]

co
....
Fig. 2.7 Comparison of predicted LSC radiation spectrum with measurements at selected wavelengths.
82 Root
'Ys + 1
Ps = 'Ys -1 Po (2.6)
where Vs is the shock velocity and 'Ys the specific heat ratio of the gas be-
tween the shock and the LSC wave (typically 1.4). For an ideal gas, the
enthalpy h is given by
'Y p
h=--- (2.7)
('Y-1)p
which leads to
(2.8)
The general properties of the LSC wave plasma can be estimated by

treating the shocked gas and plasma as two zones having constant prop-
erties with a discontinuity across the LSCwave (see Fig. 2.5). The pressure,
enthalpy, and density in the plasma are represented by the constant values
ofpP, Pp• and hP, respectively. The gas velocity with respect to the target
surface is zero. Employing conservation of mass, momentum, and energy
across the LSC wave discontinuity and the shock wave le.ads to simple re-
lationships for the plasma conditions in terms of the net irradiance IP ab-
sorbed by the plasma and the LSC wave propagation speed Vw. The results
are (Bani et al., 1976; Pirri et al., 1978)
= ('Ys + 1 ) 1/3 [ ('Y- 1 )('Ys - 1 )1p ] 2/3

(2.9)
Ps 2 Po ('Y + W)('Ys - 1- 2W)
U _ [ 2("(- 1)('Ys - 1)1p ] 113

(2.10)
s- C'Ys + 1)('Y + W)('Ys -1- 2W)p0
Pp =1- 2W (2.11)
Ps C'Ys -1)
hp 'Y(1 + W)('Ys -1- 2W)
u; = 2W('Y-1)
(2.12)
where W is the nondimensionalized LSC wave velocity with respect to the

shocked air, W = (Vw -us )Ius, and 'Y is the specific heat ratio in the plasma
(typically 'Y=1.2). The net irradiance is given in terms of the incident laser
irradiance, 10 , the radiative transfer to the targetq70 and the radiative losses
to space qs by 1P =10 -q7 -qs.
The final ingredient that is needed is the velocity with which the LSC
wave propagates into the shocked air. Determining it involves a delicate
balance between the heating of the plasma and the reradiation that causes
propagation. Since significant laser absorption begins only in the ionized
air above some temperature Ti, the LSC wave speed is determined by
the mass flow that can be heated to Ti by the radiative transport qp from
the plasma in the extreme ultraviolet, which is absorbed in the shocked
air.
In the simplest approximation, it is assumed that Ti can serve a dual role
as the "ignition" temperature: At temperatures lower than Ti, all the energy
deposition is from the plasma radiation with no laser absorption; above Ti,
all the heating is from laser absorption with no reabsorption of the plasma
radiation as it passes from the peak tempaature region through the cool-
ing front shown in Fig. 2.6. The plasma reradiation is estimated from the
emissivity of the plasma at the peak temperature. To minimize reabsorp-
tion, the ignition temperature must be selected as high as possible, con-
sistent with the requirement that laser absorption at lower temperatures
can be neglected; temperatures of 10,000 to 14,000K are commonly chosen.
This approach, which mimics an approach used to estimate the propaga-
tion velocity of chemical deflagration waves, provides a first estimate of the
propagation velocity, but detailed calculations indicate that the plasma ra-
diation is significantly altered by reabsorption in the wavefront at tempera-
tures above suitable Ti.
The solution, including the effects of reabsorption, can be obtained nu-
merically by treating the laser irradiance as an eigenvalue and integrat-
ing the energy, momentum, and radiative transport equations through the
LSC wave. In the coordinate system attached to the leading edge of the
LSC wave, with x being positive toward the surface, the equations for the
steady-state one-dimensional LSC wave that ingests a constant mass flow rh
are
pu=rh (2.13)
p + rhu = constant (2.14)
m- u 2 ) +-+2::::-=-
. d ( h+- dl dSi d ( kdT)
- (2.15)
dx 2 dx .dx dx cdx
I
where Si is the radiative flux in the ith spectral band; I the laser irradiance;
and kc the thermal conductivity. These equations are, respectively, conser-
vation of mass, momentum, and energy for a steady-state process. The laser
irradiance obeys
dl = -k}I (2.16)
dx
84 Root
where kJ is the effective absorption coefficient to the laser radiation. In

terms of the optical depth r, defined by
(2.17)
the solution for the irradiance profile is
(2.18)
The key to the obtaining reliable solutions is the fidelity of the plasma ra-
diative transport model. Early models used two radiation bands and treated
the radiation by the conduction approximation if the plasma was optically
thick or by a transparent loss term if the plasma were optically thin (Su and
Bani, 1976; Kemp and Root, 1977). The radiation conduction approxima-
tion is
(2.19)
where the radiation conductivity k 7 i is defined in terms of the effective spec-

tral absorption coefficient ,.;, 1 and the Planck spectral function B as
k .
r,1
= 47r
3
1
band i
B(w, T) dw
K '( w )
(2.20)
and the transparent loss term is
dS· = 47r 1
T K1(w)B(w,T)dw (2.21)
X band i
Subsequent models used more bands and included the differential approx-
imation (Root et al., 1979); namely,
(2.22)
(2.23)
where the average value of"' and the integral ofB over the band are implied.
The additional parameter Ui is the radiant energy density for the spectral
band of interest.
These calculations are similar to the approach used in modeling one-
dimensional LSC waves at 1 atm, which is discussed in Chap. 4. The models
describe the entire plasma profile and need not rely on the simple-step mod-
els described in Eqs. (2.9-2.12). They do, however, use shock jump condi-
tions to determine the properties of the gas into which the LSC wave propa-
gates. To predict accurately radiative transport to the surface, the complete
plasma profile calculation is necessary. Nevertheless, the general character-
istics of the plasma can be obtained from Eqs. (2.9-2.12) and the following
results from the numerical calculations: The combustion wave moves slowly
compared to the shock wave. Typical values of the nondimensionalized ve-
locity Ware 0.0091 213 for irradiance measured in MW/cm 2 • The radiative
losses are typically 15 to 20% of the incident irradiance.
The characteristic velocities in the laboratory frame for LSC waves in
1 atm air are displayed in Fig. 2.8 as a function of irradiance. To illustrate
the differences among the various absorption waves, the characteristic ve-
locities of LSD and LSR waves are also presented in the irradiance regime
where they are applicable. The graphs of pressure and temperature pre-
sented later also compare the characteristics of the three waves. However,
,...----,
r:n 10 6
s
~
()
L-....J
p.... 10 5
.,...,
-+-'
()
0
.....-<
Q)
> 10 4
10 3L_~L-~-LLW~--~-L~~Lll--~-J~~~~--L-~~LW~
1 10 10 2 10 3 10 4
Laser Irradiance [MW/ crn 2 ]

Figure 2.8 Characteristic velocities of LSC, LSD, and LSR waves propagating into
one atmosphere air.
86 Root
the present discussion is limited to the LSC waves, with the LSD and LSR
wave properties discussed in their respective sections.
For the LSC wave, the characteristic velocities are the shock speed, the
LSC wave speed, the velocity of the shocked air, and the velocity of the
LSC wave relative to the gas in which it propagates. An additional veloc-
ity, the "effective mass ingestion velocity" is shown; it is the mass flow rate
through the LSC wave divided by the ambient density and represents the
propagation rate that would be observed in ambient air if there were no
precursor shock. In the idealized LSC wave plasma, the gas is at rest in
front of the shock and in the plasma far behind the LSC wave (at the wall),
but the shocked air prior to ingestion by the LSC wave has a high velocity
imparted on it by the passage of the shock wave. The shocked gas and the
LSC wave propagating into it move at speeds similar to the shock wave (see
Fig. 2.8). The apparent high velocity of the LSC (supersonic with respect
to the ambient gas) wave belies the slow rate at which it is propagating into
the shocked air (subsonic). The propagation velocity of the LSC wave with
respect to the shocked gas increases linearly with irradiance; however, its
value is too small to influence the basic / 113 behavior of V- s, Vw, and us.
The propagation rate in the shocked gas is subsonic; hence, pressure
equilibrium is maintained across the LSC wave [Eq. (2.14)]. Thus, for the
idealized LSC wave (Fig. 2.5), the pressure remains constant between the
shock and region where the plasma radiation is being absorbed and heating
begins [often termed the "foot" of the absorption zone (see Fig. 2.6)]. The
pressure also remains constant in the plasma behind the absorption front;
the pressure drop across the front is just sufficient to decelerate the gas to
zero velocity in the plasma. The shock pressure and plasma pressure (wall
pressure) for the LSC wave are plotted in Fig. 2.9 for an air plasma. The
pressure difference between the wall and shocked gas increases substan-
tially as the irradiance increases.
The temperature profile of the LSC wave is sketched in Figs. 2.5 and 2.6.
The temperature varies slowly with distance in the plasma region near
the wall; thus, it can be characterized by the temperature TP in Fig. 2.5.
However, in the absorption region (Fig. 2.6), the temperature exceeds this
plateau value because the laser absorption occurs more rapidly than the
radiative losses (i.e., the laser absorption coefficient is orders of magni-
tude larger than the plasma radiation absorption coefficient.) The peak
temperature can be estimated by multiplying the enthalpy in Eq. (2.12)
by the ratio of the incident irradiance to the net irradiance. The char-
acteristic temperatures are plotted in Fig. 2.10. The shock initially heats
the gas, but the subsequent passage of the LSC wave results in approx-
imately 200 times as much heating. The peak temperature is approx-
imately 20% higher. Recall that the LSC wave is ingesting mass at a
slow rate compared to the shock wave; this permits the laser energy to
,----,
h
cO 10 3
o:1
L.-....J
Q)
h
;:J
rn
rn
QJ
h 10 2
0..,
Laser Irradiance [W/ cm 2 ]
Figure 2.9 Pressure as a function of irradiance for LSC, LSD, and LSR waves prop-
agating into one atmosphere air.
be channeled into heating a small mass to very high temperature rather

than a large mass to low temperature. It also ensures that the high tem-
peratures needed to propagate the plasma by reradiation are achieved;
indeed, the propagation speed is determined by the balance between
the amount of plasma radiation available to preheat the gas in the foot
of the absorption wave, and the ability of the laser to heat the plasma
high enough to emit the radiation. If the plasma temperature rises too
high, the propagation rate increases, thus causing the plasma tempera-
ture to drop ~ince less laser energy is available per unit mass of gas in-
gested.
The density profile is sketched in the Fig. 2.5. The shock increases the
density approximately a factor of 6 (for a strong shock in air), but the sub-
sequent heating at approximately constant pressure reduces the density to
approximately one-tenth of the ambient value.
88 Root
1'""""""""1 10 6
~
L.--1
Q)
~
;J
10 5
...j..l
(1j
~
Q)
s
0... 10 4
Q)
E-<
10 3
10 2 L-~~~~~~--L-~~~~--~~~~~--~~~~~
1 10 10 2 10 3 10 4
Laser Irradiance [MW/ cm 2 ]

Figure 2.10 Temperature as a function of irradiance for LSC, LSD, and LSR waves
propagating into one atmosphere air.
2.6 ONE-DIMENSIONAL LASER-SUPPORTED

DETONATION WAVE
At intermediate irradiance, the precursor shock is sufficiently strong so that

the shocked gas is hot enough to begin absorbing the laser radiation with-
out requiring additional heating by energy transport from the plasma. The
laser absorption zone follows directly behind the shock wave and moves at
the same velocity, as sketched in Fig. 2.5. This is the analog of the chemical
detonation wave that can be viewed as a shock followed by a strong defla-
gration (Courant and Friedrichs, 1957), as shown in Fig. 2.6. The fluid dy-
namics of shock propagation are primarily responsible for controlling prop-
agation. It has been modeled (Ramsden and Savic, 1964; Razier, 1965) by
analogy to the chemical detonation wave; the detailed behavior is shown in
Fig. 2.6. Mter the shock, the laser heats the gas until the local gas velocity
becomes sonic at the Chapman-Jouget point. The propagation of the LSD
wave is entirely controlled by the absorption of the laser energy between
the shock wave and the Chapman-Jouget point; the behavior behind the
Chapman-Jouget point is uncoupled and has no effect on the propagating
plasma.
Shock wave jump conditions give the initial properties before rapid ab-
sorption. Imposing conservation of energy, momentum, and mass at the
Chapman-Jouget point, where all the laser power has been absorbed, stip-
ulates the LSD plasma conditions. The velocity V5 of the shock wave is
identical to the LSD wave speed; it is given by
(2.24)
The LSD wave properties are characterized by their values at the Chapman-
J ouget point. The density Ps, pressure Ps, and enthalpy h 5 are
p =--p
r+1 (2.25)
s I o
(2.26)
and
(2.27)
where p0 is the density of the ambient gas, / 0 the incident laser irradiance,
and V5 the velocity of the LSD wave. For an LSD wave propagating away
from a wall, the effect of the wall must be included. Behind the Chapman-
Jouget point, expansion fans appear (Pirri, 1972) that permit the gas veloc-
ity to drop to zero at the wall. The plasma properties at the wall are deter-
mined by isentropic expansion through the rarefaction waves. The surface
pressure Pw• density Pw• and enthalpy hw are given by
(2.28)
Pw - Ps ---z::y1) ..,:.1
- ('+ (2.29)
h =h ('+1)2 (2.30)
w s 2,
90 Root
The gas velocity profile of a one-dimensional LSD wave is sketched

in Figs. 2.5 and 2.6. The velocity of the gas jumps as the shock wave is
crossed; further into the narrow absorption region, the gas accelerates rel-
ative to the shock front until the Chapman-Jouget point is reached. (In
the laboratory frame, the gas slows down.) In the expansion fans behind
the Chapman-Jouget point, the gas accelerates further with respect to the
LSD wave, but slows down in the laboratory frame. The velocity of the
LSD wave in the laboratory frame is plotted in Fig. 2.8, together with the
gas velocity at the Chapman-Jouget point with reference to both the LSD
wave and the wall. The LSD wave moves approximately twice as fast as
an LSC wave at the same irradiance. Comparison of the mass ingestion
velocity of the LSC wave and the propagation velocity of the LSD wave,
which processes all the mass that passes through the shock, emphasizes
that the LSD heats approximately 4000 times as much mass as an LSC
wave.
The pressure profile of the wave is illustrated in Figs. 2.5 and 2.6. The
highest pressure is achieved in the shock prior to significant absorption.
During the laser absorption process (the analogy to the chemical defla-
gration zone), the pressure drops to the Chapman-Jouget value. Behind
the Chapman-Jouget point, the pressure continues to drop due to the isen-
tropic expansion to meet the wall conditions. Typical values of the three
pressures are plotted in Fig. 2.9. The shock pressure of the LSD wave is
approximately four times the predicted LSC wave pressure. The surface
pressure is a factor of 5 less than the shock pressure; it is less than the
LSC wave shock pressure and approaches the LSC wave surface pressure
at 20 MW/cm 2 •
The temperature profile is shown in Figs. 2.5 and 2.6. Mter the initial
shock heating, the temperature rises in response to the laser absorption.
Behind the LSD wave, the temperature drops in the expansion fans. Typical
temperatures for LSD waves are plotted in Fig. 2.10. For radiative coupling
to a target surface, which is discussed in a subsequent section, the plasma
temperature at the wall is particularly important. Note that the LSD wave
plasma temperature at the wall is well below the LSC wave temperature for
an equivalent irradiance. The radiation from a uniform planar slab of air
plasma at 40 atm in four spectral regions is plotted as a function of tempera-
ture in Fig. 2.11. The radiation in the extreme ultraviolet (band 1), which is
most effectively for coupling energy to highly reflective targets, is substan-
tially lower at low temperature. Furthermore, because the laser now heats
all the mass that passes through the shock wave, which is a factor of 4000
above the amount that the LSC wave processes under similar conditions,
the peak plasma temperature is lower. Thus, LSD waves are neither effi-
cient radiators (compared to LSC waves), nor effective at transferring the

plasma radiation to the surface.
The LSD wave density profile is sketched in Figs. 2.5 and 2.6. It jumps
at the shock wave, but thereafter it decreases in both the absorption zone
and expansion fans. The density of the shocked gas is typically 6 to 11 times
the ambient density depending on irradiance; the density at the Chapman-
Jouget point is 1.83 times the ambient density, and the density at the wall is
less (0.77) than the ambient density.
10 3
Band 2
10 2
10
0
1
91
i i 1i
1 1
124 177
Wavelength [run]
1
p = 40 atm.
10 -l~~~~~~~~~~~wu~~wu~~~uu~~uu
5000 10000 15000 20000 25000 30000

Temperature [K]
Figure 2.11 Variation of plasma radiation flux in four spectral bands as a function
of temperature for air at 40 atmospheres.
92 Root
2.7 ONE-DIMENSIONAL
LASER-SUPPORTED RADIATION WAVE
At sufficiently high irradiance, the plasma radiation is so hot that, prior to

the arrival of the shock wave, the ambient gas is heated to temperatures
where laser absorption begins. In the idealized configuration, laser absorp-
tion is initiated without any density change, and the pressure profile results
solely from the strong local heating of the gas rather than a propagating
shock wave. This configuration is an example of an overdriven absorption
wave (Razier, 1965), i.e., the transition on the Hugoniot curves goes past
the Chapman-Jouget point. These supersonic waves were modeled numer-
ically by Bergel'son et al. (1975). Their numerical results confirm the basic
structure sketched in Figs. 2.5 and 2.6, once the transient plasma initiation
and formation process is completed so that the quasisteady approximation
is suitable. The steady-state propagation of the ideal waves (constant den-
sity) can also be predicted by numerical integration of the equations for
conservation of energy and radiative transport, Eqs. (2.15-2.23), with the
velocity identically zero and the pressure determined from the equation of
state at constant volume. The simpler approximation of defining an "ig-
nition" temperature has greater validity in this regime and provides good
agreement with detailed numerical calculation. Laser absorption begins
when the temperature climbs above Ti that is approximately 14,000K. Sim-
ilarly, the radiative emission from the hot plasma at temperature TP is ap-
proximated by e.aT:, where e. is the emissivity for the radiation band that
is strongly absorbed in the cool air and a the Stefan-Boltzmann constant.
Since the plasma temperature is of the order of 100,000K, only a few per-
cent of the energy carried by the laser need be used for transport. Similarly,
the loss terms are expected to be only a few percent and can be safely ne-
glected. Although the LSR is named for the plasma radiative transport that
is responsible for its rapid propagation, the reradiated power as a fraction of
the incident laser power is much less for an LSR wave than an LSC wave-
LSR waves are not energy-efficient sources of radiation. Conservation of
energy relates the LSR wave speed Vw and internal energy eP to the inci-
dent laser irradiance 10
(2.31)
where Po is the ambient density. Conservation of energy in the foot preced-

ing the ignition temperature gives
(2.32)
where ei is the internal energy difference between the ambient and ignition
temperatures. For the high temperatures associated with LSR waves, the
gas is multiply ionized and the internal energy varies approximately as TP

where f' is typically 1.6. A separate analysis of the radiation spectrum ( emis-
sivity) completes the set of equations. The major trends can be ascertained
immediately by treating E and ei as independent of irradiance. The scaling
laws for plasma temperature and propagation velocity are
(2.33)
(2.34)
where t and V are constants. Their values for a 1.06 J.lm laser pulse have
been determined by comparison with numerical solutions to be 104,500K
and 1 x 106 em/sec when / 0 is measured in GW/cm 2 • These values are plotted
in Figs. 2.8-2.10. The density remains constant and the pressure profile is
the result of heating at constant density. The LSR wave velocity increases
much more rapidly with irradiance than the LSC and LSD waves do. The
temperature and pressure increase, conversely, is quite slow. This behavior
illustrates that the LSR wave is effective in channeling the absorbed energy
into heating large amounts of gas rather than increasing the local enthalpy.
2.8 TRANSITION REGIONS
In the plots of the plasma properties in Fig. 2.8-10, the transition from one
wave type to the next is deliberately left ambiguous. Just as the transition
from chemical deftagration waves to detonation waves is still not under-
stood completely, the transition from one type of laser-supported absorp-
tion wave to another has not been satisfactorily understood. For the transi-
tion from LSC waves to LSD waves, early analyses (Bani et al., 1976) pre-
dicted that the transition occurs when the shock wave heats the gas suffi-
ciently to enable the absorption zone (shown in Fig. 2.6) to be less than
a predetermined distance. They selected a distance related to the spot
size. For a 10.6 J.tm laser, the transition was typically predicted to occur
20 MW/cm 2 and to be relatively insensitive to wavelength variations. The
best data for testing the change is the surface pressure, which is shown in
Fig. 2.12 where it is compared to the predicted values. It appears that the
transition begins near 5 MW/cm2 and is completed by 15 to 20 MW/cm 2 .
Studies of thermal coupling to targets by radiative transport also suggest
that the transition begins at 4 to 5 MW/cm 2 . However, the shock heating at
5 MW/cm2 is so small that no absorption should occur in the shocked air. A
current speculation is that the plasma absorption region remains similar to
an LSC wave up to 20 MW/cm2 , but that the plasma continues to move away
from the target, creating expansion fans between the target and absorption
94 Root
10 4
,...--,
~
cO x o NRL/AVCO, 10.6~. on metals
o:l "' SRI/AVCO 10.6~ on non-metals
1...-.J o S~Boeing 3. J.Lm on metals and non-metals
tr N , 1.06~ on metals
(])
~ 10 3
:::!
rn
rn
(])
~
P-< tr
(])
C) 10 2
cO
<i-t
~
:::!
rf)
10
1 10 10 2 10 3 10 4
Laser Irradiance [MW/cm 2 ]

Figure 2.12 Comparison of predicted and measured surface pressures for laser
supported plasmas.
zone. This behavior not only reduces the surface pressure, but also cools the
plasma adjacent to the wall so that effective radiative transport is inhibited.
It is not known whether this behavior is a remnant of the plasma initiation
process, which can generate a stronger shock than suitable for quasisteady
LSC wave propagation, or if it is an inherent feature of the propagation
mechanism which has been inadequately modeled.
Data taken on a pulsed 1.06 1-lm laser with an effective pulse duration
of 46 nsec support the transition from LSD waves to LSR waves at high ir-
radiance (McKay et al., 1984). The velocity shows the rapid increase with
irradiance and high-propagation velocities typical of Fig. 2.8, and the pres-
sure data, as shown in Fig. 2.12, saturate at levels consistent with LSR wave
predictions. However, it should be noted, that in the lower irradiance ex-
periments, the total pulse duration was shorter than the time Bergel'son et
al. (1975) found necessary to achieve the quasisteady regime for LSR wave
propagation. Calculations at a constant irradiance of 2.5 GW/cm2 required
65 nsec to approach the quasisteady regime; at 28 nsec the peak tempera-
ture was 1.75 times the final LSR wave temperature, and the apparent ve-
locity of the hot region near the peak temperature was half of the LSR wave
value.
Laser-supported radiation waves have been observed in experiments at
1.06 J.Lm in air (Markovich et al., 1979; French and Zeiders, 1984) and with
C0 2 laser radiation in noble gases (Boiko et al., 1978a; Boiko et al., 1978b)
and air (Boiko et al., 1976). The threshold for the appearance ofLSR waves
was an order of magnitude lower in air for 10.6 J.Lm than for 1.06 J.Lm radia-
tion.
2.9 RADIAL EXPANSION
Although the differences in mechanisms for the absorption waves are best
described in terms of the one-dimensional models, the most dramatic vi-
sual difference appears in long-pulse experiments where the plasma behind
the absorption wave expands radially. Relaxation of the plasma pressure
behind LSD and LSR waves has no influence on their propagation since
the wave propagates at sonic or supersonic velocities relative to the plasma.
Once created, they continue to propagate up the laser beam until the irradi-
ance falls too low to support them. Long pulses produce a narrow column of
plasma extending away from the target in the direction of the incident beam.
In contrast, LSC waves are strongly affected by the presence of the target
and by radial expansion of the high-pressure plasma. The target surface is
required in the initial formative stages to maintain the plasma pressure; it
strongly affects the strength of the initial shock wave. Furthermore, and
perhaps even more important, radial expansion reduces the pressure of the
plasma. After the radial expansion waves reach the center of the plasma,
the LSC wave is unable to maintain sufficient pressure to support the shock
wave. Since most of the apparent velocity of the LSC wave is just the mo-
tion of the shocked gas, the propagation rate decreases and the LSC wave
plasma remains near the target surface even for long pulses. For very long
pulses, the plasma pressure decays to ambient pressure during the pulse
time and the wave resembles the CW LSC wave discussed in Chap. 4.
Integrated features, such as the total energy transferred to the target and
impulse delivered to the target, must account for the evolution of the plasma
beyond the one-dimensional absorption wave. The additional contributions
may occur during irradiation when the plasma expands radially, or after the
irradiation is terminated when the plasma decays.
Consider first the evolution of the plasma for a long pulse. As soon as
the plasma has been created, radial expansion of the high-pressure plasma
commences. A rarefaction wave propagates into the plasma from the edge
at the sound speed of the plasma. After a characteristic time ru that is de-
96 Root
fined as R 5 /ap, where Rs is the laser spot radius and aP the speed of sound in
the plasma, the rarefaction waves coalesce at the center of the spot, and the
plasma pressure can no longer be maintained at the surface. The plasma
assumes the configuration shown in Fig. 2.4 with the pressure, density, and
temperature at the target surface dropping because of the gasdynamic ex-
pansion. The subsequent behavior is a function of the type of absorption
wave.
The temporal behavior of the plasma was first modeled in the LSD wave
using self-similar blast wave approximations (Pirri, 1972). The radial ex-
pansion was represented as a cylindrical blast wave; the pressure history at
the surface is given by
Ps =pld c.~) -l fort> ru (2.35)
where p 1d is the pressure for the one-dimensional configuration. The same

scaling law applies to LSR waves. However, for the subsonic LSC wave,
the entire plasma can response almost instantaneously. The recommended
scaling law (Pirri et al., 1978) is a powered spherical blast wave
t )-~ (2.36)
Ps =pld ( ru fort> ru
If the laser is terminated prior to the onset of radial expansion, a planar

decay law is appropriate
fort> Tz (2.37)
where rzis the time at which axial expansion begins. The definition of Tz is
described below. Finally, when radial and axial expansion occur simultane-
ously, starting at a time r3d, the appropriate spherical blast wave solution
is
fort >r3d (2.38)
where P3d is the pressure at the time r3d.

The decay of plasma has been studied in specific cases with hydrocodes
such as LASNEX (Ferriter et al., 1975), with other numerical simulations
(Tripplet, 1986), and with analytic solutions of the plasma flow behind an
LSD wave (Thomas, 1977). Variable energy blast waves have been modeled
(Dabora, 1972) and studied experimentally (Director, 1978). However, the
10 2
s
r---1
. . . . NRL Data OlcKay and Scbrlempf)
- - l'reaure CcmUnUlty at T• = T•
-+o-:1 ·---- ~ptoUc EzpiUIBlon
<r:
L-.1
- - PrUiiUre Drop At T 0
(])
5 Powered _!lW.st Wave
H 1 Scaling t
;:J
rn
rn
(])
H
10
0...
(]) Spherical ~,Mst Wave
Scaling t
C)
ro
'+-I
H Pulse Time
;:J
r.n
1
10 -l 1 10 10 2
Time/Tzn
Figure 2.13 Comparison of measured surface pressure history for 10.6 ttm LSC
waves with blast wave scaling laws for plasma pressure relaxation.
blast wave solutions given in Eqs. (2.34-2.38) capture the major features of
the decay and have continued to be used routinely for describing the plasma
history. For example, the measured LSC wave plasma pressure history at
the target surface is shown in Fig. 2.13, where it is compared to the blast law
scaling predictions. The time axis is given in units of the two-dimensional
time ru· The pulse time rP of the 10.6 ttm laser is three times longer. The
solid line gives the predictions for r3d = rP' The dashed line illustrates the
improvement obtained by applying refined models of the transition times
and the blast wave scaling laws that are summarized below. Both correla-
tions are satisfactory for characterizing the plasma properties for predic-
tions of the integrated impulse and radiative transport at the target sur-
face for the specific interaction conditions of the experiment; however, the
improved models are necessary for predicting the response for very short
pulses or large spot sizes where the plasma is one-dimensional long after
pulse termination, and also for the very small spot sizes when the pressure
drops close to the ambient pressure. For the interested reader, a brief his-
98 Root
tory of blast wave modeling of the radial plasma expansion follows. For the
details of the models, the reader should consult the original papers.
The first analysis of the surface pressure history for an LSD wave (Pirri,
1973) used Tz = TP and T 2d =DsiVw, whereDs is the laser spot diameter and
Vw the velocity of the leading edge of the plasma. Only three regimes were
used in the model: the one-dimensional LSD wave fort< ( TP' T2d ); cylindri-
cal blast wave scaling fort> T 2d; and planar blast wave scaling if TP < t < T2d.
These scaling laws were used to determine the specific impulse (impulse per
unit area) at the center of the spot and the total impulse delivered to targets
as a function of target area.
The early analysis of thermal coupling (radiative transfer) from LSC
waves (Pirri et al., 1978) used Tz = TP' T 2d = Rslap, and T3d as the larger
of Tz and T2d. These values are used to produce the solid line in Fig. 2.13.
Detailed calculations with LASNEX (Ferriter et al., 1977) for LSD waves
substantiated the main features of the blast wave modeling; however, there
was a delay time between the cessation of the pulse and the start of the drop
in surface pressure. This delay was attributed to the time for the rarefaction
wave, which begins in the laser absorption zone at the end of the pulse, to
reach the surface. Reilly et al. (1979) included the delay time in their blast
wave solutions. They also updated the analysis of the local centerline spe-
cific impulse and the total impulse delivered over the spot size to include the
radial pressure variations caused by progression of the radial rarefactions
to the center of the spot prior to T 2d. Later analyses (Reilly and Bouche,
1980) extended the calculations to include the total impulse delivered to an
infinite target. The results for very short pulses indicated that the specific
impulse on the centerline could be an order of magnitude larger than the
pressure times the pulse duration. The same approach was used to address
temporally varying pulse shapes (Reilly and Rohr, 1984).
Additional analysis (Tripplet, 1986) examined the validity of the pressure
decay law during the one-dimensional regime. A one-dimensional numer-
ical simulation was compared to the scaling laws. Although the onset of
the axial decay was consistent with the time Tz determined by Reilly et al.
(1979), the pressure dropped rapidly after Tz before finally settling into a
decay rate that could be represented by a blast wave scaling originating at
an earlier time.
Some of the models discussed above impose continuity of pressure at
the boundary between the blast wave solutions, as shown in the solid curve
in Fig. 2.13. Others invoke an abrupt drop in pressure, as shown by the
large dashed curve; the size of the pressure drop is determined by a sepa-
rate analysis of the pressure in the expanded plasma that has been validated
by comparison with a specific numerical analysis. An alternative approach
has been developed by Simons (1984) in which the transition time from one
blast wave model to the next is adjusted to ensure that the asymptotic ex-
pansion of the various blast waves be consistent with energy conservation.

The small dashed line in Fig. 2.13 shows the results of combining the asymp-
totic expansions and continuity of pressure at the transitions to determine
where the transition occurs.
Instead of using the asymptotic approach with pressure continuity, the
known pressure history can be used until the transition point, followed by
an abrupt jump to the asymptotic solution, or with an interpolation for-
mula to create a transition to the asymptotic formula. The large dashed
line in Fig. 2.13 shows the pressure predicted for the LSC wave when the
improved axial transient time in the LSC wave is included with a transition
to the asymptotic scaling law. These corrections have only a small influence
on the overall integrated specific impulse for most situations that generate
large impulses, since the majority of the impulse is delivered in the asymp-
totic decay region. Nevertheless, they are an easy method of producing pre-
dictions that have the correct behavior at both long and short time.
A final correction is needed for the interactions in which the target size
is large compared to the laser spot size. Under these conditions, the blast
wave decays to ambient pressure before it spills over the edge of the tar-
get. The plasma pressure at long times must then be modeled as a spherical
blast wave propagation into an ambient pressure. Similarity solutions for
the blast wave scaling laws (Sedov, 1959) predict that the plasma overex-
pands, producing a negative pressure with respect to the background. When
the correct scaling laws for the background pressure are included, the to-
tal impulse Ir generated for interactions in which the plasma expands to
ambient pressure scales as the square root of the total laser energy.
Experiments to substantiate the pressure scaling laws have been per-
formed in the LSC wave regime at 10.6 Jill (McKay and Schriempf, 1979),
and in the LSD wave regime at 10.6 Jill by French and Zeiders (1984) and
at 1.06 Jill by McKay et al. (1984).
2.10 THERMAL COUPLING
The fraction of the incident laser energy that is transferred to the target
is often termed thermal coupling. Much of the original interest in laser-
produced plasmas in air was stimulated by observations that thermal cou-
pling to highly reflective targets can be raised above the intrinsic absorptiv-
ity of the target at the laser wavelength by striking a plasma over the surface
(Marcus et al., 1976). However, other experiments suggested that plasma
shielded the target from the laser. These conflicts were resolved when the
configurations of the LSC wave and LSD waves were fully understood. In
the LSD wave regime, the plasma shields the target from the laser early in
the pulse; furthermore, the plasma temperature in the expansion fans is too
100 Root
low for effective transport by reradiation. Thus, the amount of energy de-
posited at the center of the laser spot on the target is small and the plasma
serves as a shield. However, if the target is much wider than the laser spot,
and the energy deposited in the entire target is mea:mred, the total thermal
coupling (as opposed to the local thermal coupling at the center of the laser
spot) can be of the order of 15%. The additional energy is transferred to
the target beyond the laser spot by the radially expanding plasma. As the
plasma sweeps across the target, thermal conduction across the thin bound-
ary layer can transfer significant energy to the surface. Thermal coupling in
the LSD wave regime can be enhanced or hindered depending on whether
the local or total coupling is measured.
The LSC wave plasmas can enhance the local thermal coupling if the
pulse time is chosen to maintain a one-dimensional plasma. Radiative
transfer in the extreme ultraviolet, which is well absorbed by all targets, can
couple approximately 10% of the laser irradiance instantaneously to the
center of the spot when the LSC wave plasma is in its hot one-dimensional
configuration. However, as the plasma begins to expand radially during the
pulse, radiative transport is quickly curtailed. Thus, high local thermal cou-
pling by radiative transport is obtained by demanding that the laser pulse
length be shorter than the characteristic time for two-dimensional pressure
relaxation. As an additional benefit, the decaying plasma after the pulse is
terminated continues to radiate energy to the target; for a one-dimensional
plasma, the post-pulse radiative transfer can increase the coupled energy
by 50%.
The energy transport to the target can be calculated from the descrip-
tions for the one-dimensional LSC wave plasma and the blast wave expan-
sion laws (Pirri et al., 1978). The available data confirm the predicted drop
in coupling as the pulse time increases beyond the two-dimensional relax-
ation time.
2.11 OTHER FACTORS
In addition to laser irradiance, the absorption waves can be influenced by

the laser wavelength, ambient pressure and gas composition, and angle of
incidence. The models described earlier include gas density explicitly in the
gasdynamic equations and wavelength in the absorption characteristics of
the plasma. The effect of gas composition is also embodied in the models
through the description of the gas state and properties such as enthalpy
and spectral absorption coefficients. The only uncertainty in predicting how
variations in these parameters affect the absorption waves is how they affect
the location of the transition irradiance at which the wave changes from one
form of absorption wave to another.
The angle of incidence is not explicitly included in the models. For LSD
waves, the initial one-dimensional propagation is normal to the target sur-
face (the direction that the initiation process drives the shock wave); how-
ever, at later times when the initial transient from plasma initiation has died
out, the wave propagates up the axis of the beam. Similarly, the LSR wave
initially begins as an LSD wave until the plasma becomes optically thick;
thereafter, the absorption wave races past the initial shock and propagates
up the axis of the laser beam. For the LSC wave, the entire motion is domi-
nated by the target surface until the shock wave has dissipated and the pres-
sure reduced to ambient. The one-dimensional motion is perpendicular to
the target, and the three-dimensional motion is nominally elliptical with
no indication of a preferential direction. Comparison of data and analyses
(Gelman et al., 1979) support this view ofLSCwave propagation for oblique
angles of incidence.
2.12 SUMMARY
The transient phase of plasma development between plasma initiation and

a steady-state laser-sustained plasma at the ambient pressure proceeds
through a stage as a high-pressure laser absorption wave followed by a stage,
or series of stages, where the plasma pressure decays to the ambient value.
The features of the propagating plasma are sensitive to the irradiance level.
The types of plasma can be classified in three groups-LSC, LSD, and LSR
waves-according to the propagation mechanism. The three types of plas-
mas behave very differently. The hot, slow-moving LSC wave is an excel-
lent method of converting laser energy to broadband radiation. The very
rapidly propagating LSR wave is not as efficient at converting laser energy
to radiation, but it creates a very hot plasma at the density of the ambient
gas that emits extremely short wavelength radiation. At intermediate irradi-
ances, LSD waves are produced that provide high-density absorption zones
and propagate into the ambient atmosphere independent of the pressure
relaxation of the plasma behind them. This review describes the qualitative
features of these plasmas to provide the reader with a basic appreciation of
their characteristics and to facilitate determining which plasma type is of in-
terest; detailed models of the plasma types are described in the references.
REFERENCES
Bergel'son, V.I., Loseva, T.V., Nemchinov, LV., and Orlova, T.L (1975). Propagation
of plane supersonic radiation waves, Sov. J. Plasma Phys., 1: 498.
Boiko, V.A., Danilychev, V.A., Zvorykin, V.D., Kholin, LV., Chugunov, A.Yu.
(1976). Sov. J. Quant. Electron., 6: 1065.
102 Root
Boiko, V.A., Danilychev, V.A., Duvanov, B.N., Zvorykin, V.D., and Kholin, LV.
(1978a). Observation of supersonic radiation waves in gases generated by C0 2
laser radiation, Sov. J. Quant. Electron., 8: 134.
Boiko, V.A., Vladimirov, V.V., Danilychev, V.A., Duvanov, B.N., Zvorykin, V.D., and
Kholin, I.V. (1978b). Excitation of supersonic radiation waves in a gas by a C0 2
laser, Sov. Tech. Phys. Lett., 4: 554.
Boni, A.A. and Su, EY. (1974). Propagation of laser-supported defiagration waves,
Phys. Fluids, 17:340.
Boni, A.A., Su, EY., Thomas, P.D., and Musal, H.M. (1976). Theoretical study of
laser-target interactions, Rep. SAI76-722-U, Science Applications Inc., La Jolla,
Calif.
Courant, R. and Friedrichs, K.O. (195'i). Shock Waves and Supersonic Flow.
Dabora, E.K. (1972). Variable energy blast waves,AAIA J., 10: 1384.
Director, M.N. (1978). Private communication.
Ferriter, N., Maiden, D.E., Winslow, A.M., and Fleck, J.A., Jr. (1977). Laser-beam
optimization for momentum transfer by laser-supported detonation waves,AIAA
J., 15: 1597.
French, EW. and Zeiders, G. W, Jr. (1984). Impulse measurements in air with single-
pulse C0 2 and glass lasers, AIAA Paper-84-1587, AIAA 17th Fluid Dynamics,
Plasma Dynamics, and Laser Conf., Snowmass, Colo., June 25-27.
Gelman, H., Pirri, A, Root, R.G., and Wu, P.K.S. (1979). Coupling of nonnormally
incident pulsed laser flux to metals surfaces in an air environment, AIAA Paper
79-0251, AIAA 17th Aerospace Sciences Meeting, New Orleans, La., Janu. 15-
17.
Jackson, J.P. and Nielsen, P.E. (1974). Role of radiative transport in the propagation
of laser supported combustion waves,AIAA J., 12: 1498.
Karzas, WJ. and Latter, R. (1961). Electron radiative transitions in a coulomb field,
Astrophysical!., Supplement Series VI: 167.
Keefer, D.R., Crowder, H.L., and Elkins, R.E. (1981). A two-dimensional model of
the hydrogen plasma for a laser powered rocket, AIAA Paper 82-0404.
Kemp, N.H. and Root, R.G. (1979). Analytical study of laser-supported combustion
waves in hydrogen, J. Energy, 3: 40.
Marcus, S., Lowder, J E., and Mooney, D.L. (1976). Large-spot thermal coupling of
C0 2 laser radiation to metallic surfaces, J. Appl. Phys., 47: 2966.
Markovich, I.E., Nemchinov, I.V., Petrukhin, AI., Pleshanov, Y.E., and Rybakov,
V.A. (1979). Photodetonation waves and supersonic radiation waves in xenon,
Sov. J. Plasma Phys., 5: 560.
Mitchell, R.W., Conrad, R.W, Roy, E.L., Keefer, D., and Mathews, C.W. (1978). The
role of radiative transfer in pulsed laser plasma-target interactions, J.Q.S.R.T.,
20:519.
McKay, J.A. and Schriempf, T. (1978). Personal communication.
McKay, J.A., Weiting, T.J., and French, EW (1984). Pressure and impulse genera-
tion by laser-driven air plasmas at high intensity and short pulse duration, AIAA
Paper 84-1586, AIAA 17th Fluid Dynamics, Plasma Dynamics, and Laser Conf.,
Snowmass, Colo., June 25-27.
Pirri, AN. (1972). Theory for momentum transfer to a surface with a high-power
laser, Phys. Fluids, 16: 1435.
Pirri, A.N., Root, R.G., and Wu, P.K.S. (1978). Plasma en,!rgy transfer to metal sur-
faces irradiated by pulsed lasers,AL4A !., 16: 1296.
Ramsden, S.A. and Savic, P. (1964). Nature, 203: 1217.
Razier, Y.P. (1965). Heating of a gas by a powerful light pulse, Sov. Phys. JETP, 21:
1009.
Razier, Y.P. (1970). Subsonic propagation of a light spark and threshold conditions
for the maintenance of plasma radiation, Sov. Phys. JETP, 31: 1148.
Reilly, J.P., Ballantyne, A., and Woodroffe, J.A. (1979). Modeling of momentum
transfer to a surface by laser-supported absorption waves,AL4A !., 17: 1098.
Reilly, J.P. and Bouche, E. (1984). Private communication.
Reilly, J.P. and Rohr, P.R. (1984), Pressure and impulse production by laser/material
interaction in an air environment, AIAA Paper 84-1590, AIAA 17th Fluid Dy-
namics, Plasma Dynamics, and Laser Conf., Snowmass, Colo., June 25-27.
Root, R.G. and Pirri, A.N. (1979). Theoretical analysis of radiation from laser pro-
duced plasmas, AIAA Paper 79-1489, AIAA 12th Fluid and Plasma Dynamics
Conf., Williamsburg Va., July 23-25.
Root, R.G., Pirri, A.N., Wu, P.K.S., and Gelman, H. (1979). Analysis oflaser target
interaction, Final Rep. PSI TR-170, Physical Sciences, Inc., Woburn, Mass.
Sedov, L.l. (1959). in Similarity and Dimensional Methods in Mechanics (Holt, M,
ed.), Academic Press, New York.
Simons, G.A. (1984). Momentum transfer to a surface when irradiated by a high-
power laser,AAIA J., 22: 1275.
Su, F.Y. and Boni, A.A. (1976). Nonlinear model of laser-supported deflagration
waves, Phys. Fluids, 19: 960.
Taylor, R.L. and Caledonia, G. (1969). Experimental determination of the cross-
sections for neutral bremsstrahlung,J.Q.S.R.T., 9:681.
Thomas, P.D. (1977). Jet flowfield behind a laser-supported detonation wave,AL4A
J., 15: 1405.
Tripplet, J. (1984). Private communication.
Walters, C.T., Barnes, R.H. and Beverly, R.E. III (1978). Initiation of laser-support-
ed-detonation (LSD) waves, J. Appl. Phys., 49: 2937.
3
Introduction to Laser
Plasma Diagnostics
Allan A. Hauer
Hector A. Baldis
National Research Council of Canada
Ottawa, Ontario, Canada
3.1 INTRODUCTION
The interaction of intense laser light with matter has been an active topic in
the fields of plasma and atomic physics for more than 15 years. During this
period, an amazingly rich diversity of plasma conditions has been investi-
gated. As an indication of the interesting range of parameters, the follow-
ing conditions have been encountered: (1) electron temperatures (Te) from
10 eV to many keV; (2) scale lengths, e.g., TeiiVTel, as short as a few mi-
crons; (3) self-generated magnetic fields greater than a megagauss; ( 4) par-
ticles accelerated to energies as high as a megavolt/amu; (5) imploded
compressed densities of> 30 g/cm- 3 (while maintaining Te"' 1 keV); and
(6) ionization stages as high as those of helium-like Kr. In this chapter, we
will describe some of the methods that have been devised for diagnosing
this wide range of plasma and atomic conditions.
A number of excellent review articles (Bekefi et al., 1976; Boiko et al.,
1985; Key and Hutcheon, 1980; Peacock, 1980) and one book (Hughes,
1975) have addressed the subject of laser plasma diagnostics. Because of
this wealth of available background material, we place emphasis on work
that has been done over the last six years. We will also emphasize the un-
derlying physical principles of measurement techniques as opposed to in-
strumentation. Several good reviews of instrumental methods have been
published recently (Richardson et al., 1985; Hall, 1985). In addition, a
somewhat earlier article by Attwood (1978) still is a very valuable source on
105
106 Hauer and Baldls
diagnostic instrumentation and techniques. Finally, because of space lim-

itations, we restrict the present discussion to radiation (optical and x-ray)
diagnostics. This leaves, e.g., the important areas of nuclear product (Cable
et al., 1986; Kacenjar et al., 1982) and ion emission (Sigel, 1980) diagnostics
untouched, and the references should be consulted on these subjects.
We deal primarily with the regime of irradiance that is applicable to
laser fusion (1013 -10 16 W/cm2 ). We attempt, however, whenever possible
to make comments on the application of measurement techniques to the
lower-temperature, low-irradiance conditions that may prevail in other ap-
plications such as materials processing and chemical analysis.
In discussing the diagnostics of LP's (laser plasmas), it is very useful
to consider the spatial structure involved. This structure is illustrated in
Hot e's
I
Ablation I
surface !
t
Critical Expanding
surface sheath and
<ne) plasma blowoff
Figure 3.1 Laser plasma interaction-absorption/transport phase.

Introduction to Diagnostics 107
Figure 3.2 Laser plasma interaction-acceleration/implosion phase.
Figs. 3.1 and 3.2, starting from the absorption and scattering of incident
laser energy and proceeding through the transport of this energy to denser
regions and the resulting ablation pressure, which can be used to drive im-
plosions. The structures indicated in Figs. 3.1 and 3.2 and their characteris-
tic parameters (temperature, density gradient scale lengths, etc.) are evolv-
ing on time scales varying from tens of picoseconds to many nanoseconds.
The basic progression of interaction (from absorption through compres-
sion) is, however, preserved. In recent years, research has revealed a strong
dependence of absorption and scattering processes on the laser wavelength
(Garban-Labaune et al., 1985; Turner et al., 1985). At least with respect
to the laser fusion application, this has led to a considerable emphasis on
short-wavelength laser interaction studies. Typical density and temperature
profiles are shown in Fig. 3.3. The density and temperature gradients be-
come much less steep as the irradiance is reduced.
With reference to Fig. 3.1, laser light is often absorbed by the collisional
process of inverse bremsstrahlung up to a region near the critical surface
(where the local plasma frequency equals the laser frequency), where it is
either reflected or absorbed (or scattered) by parametric processes. Brief
reviews of the possible absorption and scattering processes can be found
in Sec. 3.2 below and Cairns (1983). In the region from the critical sur-
face outward toward the laser, optical diagnostics provide the preponder-
ance of information. Scattered (shifted-frequency) and harmonic light of-
fer a rich source of information on both the absorption processes and the
local plasma conditions. In spanning the range of irradiance from 1012
to 1016 W/cm2 , the background Te varies from a few hundred eV to sev-
eral keV. The critical (electron) density for a given laser wavelength is
given by nc "' 1021 I>. 2 cm3 ( >. is in microns). The aforementioned trend
toward short-wavelength research thus implies investigation of plasma pro-
cesses at a much higher density where collisional effects will be emphasized.
The critical densities for some common laser wavelengths are shown in Ta-
ble 3.1.
In the region from below the quarter-critical density (nc/4) to the crit-
ical density (nc), various absorption and scattering processes can result in
the generation of suprathermal electrons. An example of these processes
is stimulated Raman scattering (SRS) in which the laser wave is coupled to
an electron plasma wave and a shifted electromagnetic wave. The strongly
driven plasma wave can accelerate electrons to very high energies. These
electrons can follow a variety of trajectories, depositing their energy in
dense material or penetrating the sheath potential and escaping. These
suprathermal electrons are very detrimental to efficient laser-driven com-
pression, and their minimization at shorter wavelengths is an important cur-
Te ne Te
Te
2.0
keV
I. 1.0
keV
nc
n0 /4 ---- -----
100 fLm X 200fLm X

(a\ (b)
Figure 3.3 Typical density and temperature profiles for (a) high and (b) low inten-
sity irradiation. nc is the critical density and n, is solid density.
Table 3.1 Critical Densities for Some

Important Lasers
Wavelength nc
Laser (!-lm) (cm- 3 )
C02 10.6 1019

Nd glass (w 0 ) 1.06 1021
Nd glass (3wo) 0.35 9 X 1021
KrF 0.25 1.6 X 1022
rent research topic. They are diagnosed primarily by the x-ray bremsstrah-
lung that they create.
Absorption by inverse bremsstrahlung produces a thermal electron spec-
trum and is the preferred mechanism with respect to laser fusion. The pro-
duction of high-energy particles and x-rays, while deleterious for fusion,
may be useful in other applications (See Chap. 13). In addition to self-
emission, optical probing with an external (short-pulse) laser provides a
wealth of information. Plasma diagnosis using optical emission and probing
is the subject of Sec. 3.2 of this chapter.
In the denser portions of the plasma, such as the thermal conduction
region, opacities are such that only photons with near x-ray energies escape
and usually are employed for diagnosis. X-ray diagnostic techniques are
dealt with in Sec. 3.3.
Even greater densities (and thus opacities) are encountered in laser-
driven implosions, and x-rays must be employed exclusively for diagnosis.
In both optical and x-ray diagnostics, there is an interesting tradeoff be-
tween the use of self-emission and probing with external radiation. Self-
emission is usually simpler to employ and often provides more local infor-
mation. Probing, on the other hand, allows for the following degrees of
flexibility: (1) the ability to probe cool and/or dense plasmas where self-
emission is weak or cannot escape due to large opacity, and (2) the ability
to operate in a "snapshot" mode where the probe radiation is in the form of
a short burst and can be timed with respect to the evolution of the subject
plasma, thus giving time-resolved information.
We have attempted in this chapter to present a heuristic introduction
to the diagnosis of laser plasmas. References have been chosen that the
authors felt would be most instructive to the reader. This is not intended to
be an exhaustive treatment and we regret that because of space limitations,
some important (and even pioneering) work could not be included.
3.2 INTRODUCTION TO
OPTICAL DIAGNOSTICS
Optical plasma diagnostics techniques can be separated into two main cat-
egories: (1) those involving the analysis of the emission spectrum from the
plasma, and (2) those analyzing the changes undergone by radiation intro-
duced into the plasma in the form of a probe beam.
In a laser-produced plasma, we have two main contributions to the emit-
ted spectrum. One contribution consists of the emission from the self-
luminous plasma itself, the continuum, and line radiation. The analysis
of this radiation has been an important diagnostic in laser plasma physics.
However, for LP's with temperatures in the range 100 eV to > 1000 eV (e.g.,
in laser fusion studies), most of the emission will lie in the spectral region
extending from the vacuum ultraviolet to the x-ray region. The analysis of
this emission will be discussed in Sec. 3.3. In applications such as chemical
analysis, temperatures are usually below 100 eV (and are more typically a
few e V). In such studies, the line and continuum emission is in the optical
and ultraviolet regimes. Many of the spectroscopic analysis techniques that
are discussed in Sec. 3.3 in relation to x-ray diagnostics have close analogs in
optical measurements. Some of these will be discussed in Sec. 3.3. A second
contribution to the radiation emitted from the plasma is emission associated
with scattering of the incident laser light by different modes in the plasma.
This emission occurs through parametric processes [Sec. 3.2.1(A)] and the
generation of harmonics and fractional harmonics of the incident laser light
[Sec. 3.2.1(B)].
Another optical diagnostic technique is to use a separate laser beam as a
probe. Information about the plasma can be obtained from changes in the
optical properties of the probe beam. One example is interferometry, by
which the electron density spatial distribution and gradients are commonly
derived, by measuring the change in the phase front of the probe beam.
Diagnostics based on the use of an external laser probe will be discussed in
Sec. 3.2.2.
3.2.1 Scattered light

(A) Parametric Instabilities
Intense laser light can excite a collection of instabilities in the underdense
(corona) plasma (Drake et al., 1974). From the point of view of plasma di-
agnostics, there are two areas of interest related to parametric instabilities.
One is the study of the parametric instabilities themselves. The most widely
used diagnostic for the study of parametric instabilities in LP's has been the
scattering of an electromagnetic wave (the laser radiation itself) from one
of the longitudinal waves resulting from the decay. Typical examples are
the scattered light associated with instabilities such as stimulated Raman

scattering (SRS) and stimulated Brillouin scattering (SBS) (Forslund et al.,
1975). Another example is the 3w0 /2 emission (Karttunen, 1985) resulting
from the scattering of the laser electromagnetic (EM) wave from plasma
waves produced by the decay of the laser light into two plasma waves (two
plasmon decay or TPD) (Goldman, 1966).
A second area of interest in these parametric instabilities is the use of the
scattered light as a diagnostic of the plasma itself. In recent years, this has
become an important diagnostic to characterize the temporal evolution of
plasmas produced by irradiating thin foils with a high-intensity laser (Drake
et al., 1987).
The basic mechanism for the excitation of waves in a plasma is the oscil-
lation of electrons in the electric field E of the laser light, moving across a
variation in the plasma density. An electron density fluctuation 6n is then
driven, which can be expressed as
(3.1)
where xosc = eEw~IMe is the electron quiver amplitude. The variation in

electron density can be due to the overall density gradient (as shown in
Fig. 3.3) or density fluctuations associated with other waves in the plasma.
Given an initial electron density fluctuation associated with one mode of
the plasma, the mechanism leading to the resonant coupling (and instabil-
ity) is a feedback loop (Kruer, 1983). In the case of a radiative instability
(like SRS), the laser electric field couples with the density fluctuation to
drive an oscillating current that, in turn, drives a high-frequency electric
field. The beating of this electric field with the incident EM wave will, in
turn, produce a fluctuation in the radiation pressure that enhances the ini-
tial electron density fluctuation. An instability is then possible as long as
the following matching or resonant conditions are satisfied:
(3.2)
where subscript 0 denotes the incident beam and 1 and 2 represent the
daughter waves. Because all natural modes of oscillation are damped, the
pump amplitude must exceed a certain threshold intensity to drive them
unstable. These modes then grow exponentially as they absorb energy from
the pump, until they are saturated by nonlinear effects. The growth of these
instabilities is favored by plasmas with large scalelengths, defined by
(3.3)
where ne is the electron density. Because long-scalelength plasmas are

expected in laser fusion applications, the study of these unstable plasma
modes has become very important.
What parametric instabilities are possible in LP's? In an unmagne-
tized uniform plasma there are three branches to the dispersion relation:
the longitudinal electron plasma wave (langmuir wave), the longitudinal
ion acoustic wave, and the EM wave. For these three different types of
modes, there are a variety of different ways in which they can be com-
bined to yield a resonant three-wave interaction. Of all possible combi-
nations, only four interactions are of interest in the coupling of a high-
intensity laser beam and a plasma: the decay of the incident EM waves
into (1) an ion acoustic wave and a second EM wave (SBS), (2) an elec-
tron plasma wave and a second EM wave (SRS), (3) two plasma waves
(TPD), and ( 4) an electron plasma wave and an ion acoustic wave. In ad-
dition to these instabilities, the incident EM wave can filament and self-
focus.
A complete review of these instabilities is beyond the scope of this dis-
cussion. Instead, we will summarize briefly some of the characteristics of
the scattered light as encountered in LP's. In particular, we will describe
some of the features of SRS as they relate to diagnostics of these plasmas.
The principal spectral components (not all of them present under the same
plasma conditions) are illustrated in Fig. 3.4. The spectral component near
the incident laser frequency is produced by SBS. Because the scattering has
been done by ion waves, the frequency shift is approximately proportional
FREQUENCY: 2w0 3/2 W 0 Wo 1/2w0

WAVELENGTH FOR:
(a) 10.6fLm LASER 5.3 7.07 10.6 21.2 CfLm)
(b) 1.06fLm LASER 0.53 0.707 1.06 2.12 (fLm)
(c) 0.266fLm LASER 0.133 0.173 0.266 0.53 (fLm)
Figure 3.4 Schematic of scattered-light spectrum for some parametric instabilities

(arbitrary intensities).
lntrodudlon to Diagnostics 113
to the ion acoustic frequency
(3.4)
where k is the wave number of the ion waves, mi the mass of the ions, kB
the Boltzmann constant, Te and Ti the electron and ion temperatures, and
Z the change of the ions, respectively. Although this shift can be used as a
measure of the electron temperature Te, it is an unreliable method due to
uncertainties from the Doppler shifts introduced by the plasma expansion
and also the shift in the ion acoustic frequency due to the presence of the
laser radiation.
Fractional harmonic scattering, shown at 3w0/2 and w 0 /2 in Fig. 3.4, is
connected with instabilities occurring in the plasma near nJ4. The con-
nection of these emissions with instabilities occurring at this density is now
clearly established (Turner et al., 1984; Karttunen, 1985; Seka et al., 1985).
Both 2wP and SRS excite plasmons with frequencies close to w0 /2. The cou-
pling between the incident or reflected photon and a w0 /2 plasmon gives rise
to the emission of radiation with frequencies w0/2 or 3w0/2. These emissions
have been observed in many LP's with different laser wavelengths. Seka
et al. (1985) have recently analyzed the w 0/2 and 3w0 /2 emission from UV
plasmas and evaluated their usefulness as a diagnostic for coronal electron
temperature. The evolution of w 0/2 spectra for different laser irradiation
intensities and target materials is shown in Fig. 3.5. The sharp, slightly red-
shifted peak seen in all spectra can be effectively used as an electron tem-
perature diagnostic. This feature has been identified with Raman down-
scatter processes, which in turn leads to reasonably well-defined frequency
shifts. The 3w0/2 harmonic spectra, on the other hand, are much less suit-
able as a temperature diagnostic, because there does not appear to exist
any effective frequency selection mechanism analogous to those identified
for the w 0!2 spectra. It seems that the spectral splitting on the 3w0/2 emis-
sion depends on not only the electron temperature, but the geometry of the
plasma as well (Seka et al., 1985). The two effects are generally of equal
magnitude and are difficult to separate or distinguish experimentally. The
3w0/2 emission is, nevertheless, a good indicator of high levels of plasma
waves at nc/4, a signature of 2wP and SRS. This feature has been used to
identify filamentation in the plasma corona (Lin et al., 1981).
The broadband emission between w0 /2 and w0 is due to the SRS insta-
bility occurring at densities below nc/4. Because in recent years particular
attention has been given to SRS in connection with laser fusion (as well as
a diagnostic for the plasma corona), we will discuss the source of its char-
acteristics next.
680 700 720 740

A.(nm)
Figure 3.5 Evolution of w 0!2 spectra from uniformly (tangentially) illuminated
spherical targets. Irradiance levels next to material designations are in W/cm 2 (from
Seka et al., 1985 with permission).
Raman scattering is usually identified in experiments by the frequency

spectrum of the scattered light. If we solve the dispersion equation for the
waves involved in Raman scattering, we obtain a relationship between the
wavelength of the emitted light and the plasma (electron) density at which
SRS took place, as long as the temperature of the plasma is known. These
relationships are displayed in Fig. 3.6.
The characteristics of a typical spectrum of SRS are clearly shown in
Fig. 3.7 (Tanaka et al., 1982). Similar results have been obtained in other
laboratories (Phillion et al., 1982). The experiment of Tanaka et al. (1982)
consisted of irradiating a CH target with 351 nm laser radiation at intensities
between 1013 and 1015 W/cm2 • The spectral dependence of the backscat-
tered radiation between 400 and 750 nm is shown in Fig. 3.7. Curve a shows
the scattered spectrum at 4.6 x 1014 W/cm2 (10% above the SRS threshold
for the instability). The signal is strongest at wavelengths that correspond
to Raman scattering from densities somewhere between 10 and 20% of nc.
There is a small signal with frequency near 2w0 ( >. = 703 nm ), which may be
Raman scattering near nc/4, or mode conversion of plasma waves generated
1.1 2.0
1.0 1 .8
0.9 1.6
ws 0.8 1.4 As
wo 0.7 1.2 Ao
0.6 1.0
0.5 0.8
Te=0.1keV
0.4
0.01 0.1 1.0
"e /nc
Figure 3.6 Relation between frequency and wavelength for emitted SRS light as a
function of electron density. The subscripts refers to the scattered radiation.
by the TPD instability. Curve b illustrates the result well above threshold.
A wider spectrum represents a wider range of electron densities at which
SRS is taking place. The cutoff is typically attributed to the suppression of
the SRS instability at low plasma density due to strong Landau damping of
the plasma wave. This observed cutoff can be used to estimate (Seka et al.,
1984) the electron temperature in the low-density plasma. Such estimates
have been found to be in reasonable agreement with hydrodynamic code
calculations and x-ray spectroscopy of thin CH foil targets.
The lack of SRS signal between 610 and 700 nm corresponds to densi-
ties between 20 and 25% nc. Different explanations have been offered for
such a gap (Kruer, 1986). The most popular explanation has been steep-
ening of the electron density profile at these densities due to the pondero-
motive forces generated by the electrostatic plasma wave fields associated
with TPD instability. Another explanation for the gap is the suppression of
the SRS instability by high levels of ion fluctuations produced by either the
nonlinear saturation of TPD (Kruer, 1986) or ion waves from SBS (Aldrich
et al., 1986; Walsh et al., 1984).
The previous discussion illustrates the use of the light scattered by SRS
as a diagnostic tool in LP's. The temporal evolution of exploding thin foil
targets (Drake et al., 1987) has been characterized and compared with the-
oretical models (London and Rosen, 1986) by using the low- and high-
~ 100
~
5
>
t::
(/) 10
z
w
1-
~
z 1
<(
~
<( 400 500 600 700
a:
WAVELENGTH (nm)
Figure 3.7 SRS instability spectra from CH targets (a) near threshold (b) well
above threshold (from Tanaka et al., 1982 with permission).
wavelength cutoffs of the SRS spectrum. This is a relatively new area of

plasma diagnostics in LP's.
One final feature of the generic spectrum, shown in Fig. 3.4, is the emis-
sion near 2w0 • This is the upscatter component of SRS occurring at nc/4.
This feature has only been observed recently (Turner et al., 1986).
(B) Harmonic Emission

When laser light is incident on a target, the reflected light is found to con-
tain the second harmonic of the incident laser light (Babin et al., 1973) and
higher harmonics (Burnett et al., 1977; McLean et al., 1977). More recent
experiments performed with C02 laser light have revealed the generation of
harmonic spectral components up to the 46th order (Carman et al., 1981).
Harmonic generation in LP's is very closely related to the process known
as resonance absorption. The measured efficiency of harmonic production
and the polarization of the harmonic wave can provide insight into the gen-
eral properties of the plasma. Two aspects of the plasma response are pos-
sible contributors to the production of harmonics: (1) the steep density gra-
dient arising from the radiation pressure, and (2) self-focusing originating
from intensity modulation in the laser beam.
In resonance absorption, laser light incident on a plasma density gra-
dient at an oblique angle can excite resonant plasma oscillations near the
critical-density region, as shown in Fig. 3.8. At the turning point, the electric
field of the laser light is parallel to the electron density gradient. Electrons
RESONANT
POINT
FIELD OF
PLASMA WAVE
Q)
>=""
>-
I-
(f)
z
w
0
8 /
~/
/
Figure 3.8 Schematic of resonant absorption process for P-polarized light.
oscillating under this field will be moving into a region of higher and lower
electron density, and their oscillation will not be sinusoidal any longer. Har-
monic components will be superimposed on their oscillations and similarly
on the EM wave they radiate. The number of harmonics will be a function
of the intensity of the laser light and the steepness of the electron density
gradient.
One mechanism to generate profile st'eepening at the critical surface is
the electric-field pressure, or ponderomotive force, of the laser light. Esti-
mates of this effect (Forslund et al., 1977) suggest that for high-power C0 2
lasers, the density profile may be steepened to densities as high as several
hundred times the nc for the laser. The dominant force is due to the res-
onant plasma wave, which accelerates electrons down the density gradient
during resonance absorption (Max, 1982).
A spectrum obtained by Carman et al. (1981) for C0 2 laser interaction
is shown in Fig. 3.9. The observed harmonics include both even- and odd-
integral values, as well as half-integral values. The high order of harmonics
observed in this experiment implies an extremely steep density profile. The
highest harmonic simply corresponds to the wavelength for which its critical
density is the plasma density of the upper self (see Fig. 3.3). The steepening
of the electron density gradient is further enhanced by the self-focusing of
the laser.
Complex second-harmonic emission spectra have been observed using
a short-pulse ("' 20 psec), high-intensity (> 1015 W/cm 2 ), 1-ttm laser light
(Burgess et al., 1984). The spectral features gave information on the local
plasma conditions near nc that was consistent with the predictions of a self-
consistent model of electron density profile modification due to resonant
fields.
Second-harmonic emission has also been used to infer filamentation in
LP's (Stamper et al., 1978 and 1985). Spatially resolved images of this emis-
sion allowed observation of filamented regions in the far underdense plasma
(nc "'0.5nc) at large distances ("' 1 mm) from the critical surface. From
these observations, a relationship was established between the plasma scale
length and the density at which the filamentary structures were observed.
3.2.2 Optical Probing of Laser Plasmas

(A) Optical Probing Basics
Optical probing of plasmas is an important and mature area of plasma di-
agnostics. Optical interferometry, e.g., has been used to measure electron
I I I I I I I
CARBON WIRE 250 p.m diam
I- 8.2x1014 W/cm2
I I
38 36 34 32 30 28 26 37 35 33 31 29 27 25
Figure 3.9 Harmonic spectra from C0 2 -laser-illuminated targets under nominally

identical conditions (from Carman et al., 1981 with permission).
densities since the early days of theta and Z-pinch machines. In more recent
times, Thomson scattering has been one of the most important diagnostics
to characterize the density and temperature of electrons in Tokamak rna-
chines.
The application of optical probing to LP's is more difficult due to the rel-
atively higher densities encountered in these plasmas. Because the probe
pulse is limited to electron densities less than nc, the critical density at which
the plasma frequency equals the laser frequency, a laser probe with a short
wavelength is required to propagate through the high densities encountered
in these plasmas. Because it is desirable in most cases to examine den-
sities greater than the critical density of the laser producing the plasma,
it is necessary to use a probing pulse with a wavelength shorter than this
laser. Probing with optical radiation having a frequency greater than any
characteristic plasma frequency also warrants that the interaction with the
plasma will be weak and the plasma conditions will not be disturbed signif-
icantly.
The diagnostics we describe below require a probing pulse with a wave-
length shorter than the laser producing the plasma and in excellent synchro-
nization with it. In Nd:glass laser interaction studies, the probing pulse is
obtained by frequency doubling and by quadrupling a portion of the main
pulse, thus providing perfect synchronization once the optical paths have
been equalized. If a shorter probing pulse is required, different schemes
have been developed to synchronize two laser oscillators to provide a long
and short pulse simultaneously.
In C02 laser experiments (wavelength 10.6 tlm), the long wavelength of
the laser combined with the practical requirement of a visible probing pulse
imposes the constraint of having to synchronize two lasers of completely dif-
ferent nature. If the C0 2 laser oscillator is actively mode-locked, one tech-
nique for obtaining a synchronized diagnostic pulse has been to employ an
actively mode-locked glass (Tornov et al., 1979) or ruby (Martinet al., 1980)
laser, driven from the same pulse generator that drives the electrooptical
modulator of the C02 laser. Temporal resolution can be obtained by the
short-duration probing pulse or by using a streak camera as a recording de-
vice.
(B) Electron Density Measurements
Optical interferometry is a well-established technique for measuring the

refractivity of plasmas by directly comparing the phase of a wavefront
passing through the plasma with a reference (Jahoda and Sawyer, 1971).
Interferometry can be used to determine the electron density distribu-
tion because in highly ionized plasmas the refractive index p. is related
to the free-electron density. The refractive index in a plasma is given
by
(3.5)
where w is the angular frequency of the probe and we is the electron cy-
clotron frequency. If we retain only the first two terms in the expansion of
p,(r ), the effective changes in the wavefront due to the plasma in units of
fringes N are given by
(3.6)
where .X is the wavelength of the probing light, r is measured along the op-
tical path of the probing beam, and for the plasma conditions of interest
we = 0 (no magnetic field). The electron density distribution is then ob-
tained by Abel inversion (Bockasten, 1961; Sweeney et al., 1976) of the
observed shifts. Cylindrical symmetry has to be assumed if interferome-
try is performed only along one direction. These kinds of considerations
are encountered often in both optical and x-ray diagnostics (for both self-
emission and probing) because the signal received by a detector is an inte-
gration along a path through the plasma where the temperature and density
may vary in complicated ways. Interferometry along multiple lines of sight,
similar to tomography, has been attempted only in some simple plasma con-
figurations.
In LP's, the limited size of the plasma permits the use of a single-beam
interferometer arrangement in which the plasma is placed off axis, thus us-
ing only half of the beam as a probe. This arrangement permits one to use
the other half of the beam as a reference beam. Folding the wave over itself
or splitting the beam in two and shifting one component allows for a com-
pact and stable arrangement. A further improvement is to use holographic
interferometry (Attwood, 1978) because of its reduced requirements on the
quality of some of the optical components and improved spatial resolution.
Because all the phase information is recorded, the focusing of the interfer-
ometer is not crucial. Recently, a multiple-frame interferometer has been
developed by Busch et al. (1985) that permits the study of the temporal evo-
lution of a plasma in a single shot with excellent optical resolution.
One important consideration in utilizing interferometry for density dis-
tribution measurements in LP's is to have a probe pulse that is sufficiently
short in time to freeze the motion of the plasma. Because the plasma den-
sity contours move outward with high velocity, the probing pulse length T
must be, to first approximation,
required spatial resolution

T< ve1oc1ty
. ofd ens1ty
. contour
For greater probe pulse durations, the interferometric fringes will smear,
lose contrast, and eventually disappear. For electron density contour ve-
locities of the order of 106 em/sec and spatial resolution of 1 p:m, a probe
pulse duration of ,...., 100 psec is required.
The most significant limitation to interferometry results from refraction
of the probing beam when there are steep electron density gradients trans-
verse to the direction of the probe beam. The apparent localization of the
fringes can be changed, or if the beam deflection is severe, the fringes can be
masked out. Simple arguments show that in passing through a plasma with
a Gaussian electron density distribution along the axis of the probing pulse
and an exponential distribution normal to it, the probe pulse is refracted
through an angle (} given by
(3.7)
where le corresponds to the 1/e diameter of the plasma, L is the scale length
of the plasma perpendicular to the beam [see Eq. (3.3)], wP is the plasma
frequency at the maximum density, and w is the angular frequency of the
probing pulse. To observe a specific electron density maximum with a steep
profile, large aperture optics are required to accommodate the strong devi-
ation imposed on the probe beam.
Using the holographic system shown in Fig. 3.10, Attwood (1978) studied
the electron density steepening produced by 1.06p:m laser radiation on glass
microballoons (due to radiation pressure). The diffraction-limited spatial
INCIDENT LIGHT,
LENS 1.06 J-Lm
PRISM
PHOTOGRAPHIC
PLA~~~~
GROUND
GLASS
0
2660A
Figure 3.10 Holographic system used for diffraction-limited interferometry (from

Attwood, 1978 with permission).
122 Hauer and Baldis
(2660A, 15 psec, 1 }Jm)
1.06 }Jm, ~ 3 X 10 14 W/cm 2 , 41 }Jm 0

(a)
Figure 3.11 (a) Interferogram at peak irradiation of a sphere, (b) density plot
demonstrating profile steepening, dotted line indicates density profile slope in upper
region (from Attwood, 1978 with permission).
resolution was ~ 1 11m. The probe pulse had a duration of 16 psec at a

wavelength of 0.26 Jlm. An example of a reconstructed interferogram using
a HeNe laser is shown in Fig. 3.11. In this example, a 41 11m diameter glass
microballoon was irradiated with a 30 psec, 1.06 11m laser pulse at an in-
tensity of 3 x 1014 W /cm2 • Figure 3.11(b) shows the electron densities after
Abel inversion. These results demonstrate that the radial density profiles
are indeed steepened by radiation pressure effects.
1.0
>-
1-
CiS,....
Zo
Wo
0'
Q)
~C\i 0.3
a:O
1-:::
(.)
w
...J
w
0 •1 0 6 12
AXIAL DISTANCE FROM
INITIAL WALL, fLm
(b)
The refraction of the probing laser beam by transverse density gradients,

which introduces a limitation on the application of interferometry as a di-
agnostic, is in itself the basis of other diagnostic techniques to detect spatial
variations of the refractive index of the plasma. The schlieren method and
shadowgraphy are two such techniques. The practical distinction between
them is that schlieren detection occurs at the location of the focused image
of the refractive plasma region (Jahoda and Sawyer, 1971). Both schlieren
and shadowgraphy methods can be useful in delineating discontinuities and
indicating overall plasma configurations, although it is quite difficult to ob-
tain accurate quantitative information from these diagnostics.
Benattar and Popovics (1983) developed a technique to deduce the spa-
tial and temporal variation of electron density in LP's. Although this
method cannot be applied when there is a large step in the density profile,
it can be used to deduce the density profile if the density step is moderate.
Benattar et al. (1983) used this technique to deduce the expansion velocity
of the critical layer in an LP created by a C02 laser.
(C) Magnetic Field Measurements

In the presence of strong anisotropies, the polarization vector of a probe
beam can suffer a rotation and be used as a diagnostic itself. If the
anisotropy is produced by strong magnetic fields, the rotation of the electric
field of the probing pulse (Faraday rotation) provides the basis for measur-
ing these fields. The linear polarization is decomposed into two oppositely
rotating circular polarizations. In the magnetized medium, one of these
rotates in the same sense as the electrons, with the other rotating in the op-
posite sense. The dispersion equation yields different phase velocities for
these two wave components, resulting in a net rotation of the linear polar-
ization. The polarization vector is rotated when a component of the prop-
agation vector is aligned with the magnetic field vector. To first order, the
rotation angle ¢> is given by
¢> = 2.6 x 10-25 >. 2 J

ne(r)B(r)dr (3.8)
where ¢>is in radians, >.is in JLm, and the magnetic field B(r) in gauss is the
component of the magnetic field in the direction of the propagation vector.
Note that a knowledge of electron density ne is required to unfold magnetic
field measurements. Rotation of the polarization vector is measured by ob-
serving the intensity variation transmitted by a polarization analyzer of fixed
orientation. Even for large magnetic fields (megagauss), the rotation ¢> is
very small (a few degrees or less).
The interest in magnetic field measurement in LP's is based on theoret-
ical and computational evidence of spontaneous magnetic fields at mega-
gauss levels, which may themselves be useful in applications. Spontaneous
magnetic fields can be generated by the following possible mechanisms:
(1) the flux of charged particles emitted from the plasma, (2) an electron
temperature gradient having a component perpendicular to the density gra-
dient, (3) charge separation during the plasma expansion, and ( 4) momen-
tum deposition by the laser beam, in particular in the presence of irradiation
asymmetries.
Extensive studies have been made (Stamper and Ripin, 1975; Stamper
et al., 1978; Raven et al., 1977; Willi et al., 1981) that confirm the presence
of megagauss-range magnetic fields. Because a determination of the mag-
netic field requires a knowledge of the electron density distribution [see,
Eq. (3.8)], the polarimeter is usually run simultaneously with interferome-
try. Both diagnostics can use the same probe pulse through the use of rela-
tively simple optical arrangements. In a typical experimental configuration,
part of the image beam is split off into a folded-wavefront interferometer
and the remainder is analyzed for polarization rotation using a Wollaston
prism. Figure 3.12 shows an interferogram (a) and a Faraday-rotation im-
age (b) of a planar target irradiated at an intensity of 1014 W/cm 2 (Duncan
et al., 1981). The diagnostic beam consisted of 1 mJ in 50 psec of Raman-
shifted second-harmonic light (0.622 pm). The Raman shifting was nec-
essary due to the strong second-harmonic emission from the plasma itself.
The analyzer for the Faraday-rotated light was set at 80° to the probe beam
polarization, resulting in the dark/bright pattern shown in Fig. 3.12(b). If
a crossed polarizer had been used, two bright regions would have been ob-
100 J.lm
Figure 3.12 (a) lnterferogram and (b) Faraday rotation image from irradiation of
aluminum wire targets (from Duncan, Luckin, and Willi, 1981 with permission).
served, one above and one below the laser focal spot, indicating the pres-
ence of a toroidal self-generated magnetic field.
An alternative technique to measure the magnetic field in a laser plasma
using the Zeeman effect has been proposed and demonstrated by McLean
et al. (1984). Although this technique has the advantage of not requiring
a probe beam, it requires considerable effort to analyze the spectral line
shape.
(D) Thomson Scattering
A laser beam incident on a medium can be scattered by inhomogeneities in

the medium. In a plasma, such scattering is mainly due to electron density
fluctuations. Every charged particle in the plasma is capable of scattering
light, but because the scattering cross section is inversely proportional to the
square of the particle mass, the contribution from the ions is negligible com-
pared with that from the electrons. The spectrum of the electron density
fluctuation in the plasma is impressed on the scattered light, thus providing
information about the collective motion of the electrons as well as the non-
collective motion (velocity distributions). These features make Thomson
scattering a powerful diagnostic, providing direct information about these
two aspects of electron fluctuations. Although only the electron scattering
is detectable, the motion of the ion modifies the electron fluctuations, so
information about the ion motion is also obtained.
In spite of being a powerful diagnostic, Thomson scattering has been lim-
ited to the investigation of relatively long-scale plasmas produced by C0 2
lasers. During the past few years, Thomson scattering has been success-
fully used in C02 LP's to study both the collective and noncollective plasma
fluctuations. The longer electron density scale length encountered in these
plasmas, as well as the relatively low nc (10 19 electrons/cm3 ) of 10.6 f.-LID
laser light, has permitted the isolation of different interaction regions in
the plasma corona, allowing for the first time the direct observation of elec-
tron plasma waves as well as ion acoustic waves. Many reviews have been
written about the application and theory of Thomson scattering (Sheffield,
1975), so only some basic considerations will be discussed below.
Scattering from a collection of charged particles occurs only if there are
spatial fluctuations in the electron density. Conservation of momentum
tells us that the wave number of the density fluctuations must satisfy the
condition k = (ks- k;), where k; and k.s are the wave numbers of the inci-
dent and scattered light beams, respectively. The magnitude of k is given
by
k 411" . ()
= -sln- (3.9)
,\ 2
where 8 is the angle between ~ and~·

The vector k is fixed by selecting a laser wavelength and a scattering an-
gle. The quantity k- 1 is the scale length for scattering and represents the
scale on which the plasma fluctuations are viewed. Fixing the vector k se-
lects one spatial Fourier component of the electron density distribution and
the frequency spectrum that will be observed. Some Fourier components
will correspond to resonant modes in the plasma, appearing as peaks in the
frequency spectrum. However, because plasma resonance can exist only for
modes with wavelengths larger than the Debye length ).D, the kind of scat-
tered spectrum that will be observed can be characterized by comparing k - 1
and ).D. If the scale length k- 1 ~ ).D, the movement of the individual elec-
trons will be observed. If, on the other hand, k- 1 ~ ).D, collective motion of
the electrons and/or ions will be observed. A convenient parameter a can
be defined to characterize the different scattered spectra, where
1
a=-- (3.10)
k>.v
Typical spectra for thermal plasmas are shown in Fig. 3.13 for different val-
ues of the parameter a. For values of a~ 1, the spectrum is approximately
Gaussian, which corresponds to the Maxwellian velocity distribution of the
electrons. The width of the spectrum can then be used to calculate the elec-
tron temperature. If the optical system has been calibrated absolutely, the
integrated scattered light is proportional tone. The Gaussian shape is dis-
torted, as the value of a increases, until for a~ 1, the spectrum consists of
two predominant peaks. The first feature is a well-defined electron satellite
shifted from the incident laser frequency by the Bohm-Gross frequency
(3.11)
where V1h is the electron thermal velocity. The shift of these "sidebands"
is mainly a function of the electron density ne (through wp) and, to a lesser
degree, of electron temperature Te. The electron density is then given by the
location of the satellites. The widths of these satellites are very narrow, and
in practice they cannot be resolved. The second component of the spectrum
consists of the ion feature centered (in the absence of current drifts) at >. 0,
the incident laser wavelength. It consists of two peaks separated from the
center of symmetry by the ion acoustic angular frequency w;a [Eq. (3.4)].
The width of each peak is given by the thermal motion of the ions. The
overall shape of the peaks is then a function of both Te and T;. By a proper
choice of scattering configuration, one can, within certain limits, optimize
the kind of spectra to be obtained. For a given plasma, the choice of >. 0 and
-
X
Q)
-
~
-3
10
10- 2 10 1 1 10
Xe=(w-w0 )/kVe
Figure 3.13 Scattering cross section as a function of frequency difference normal-
ized to the thermal electron velocity in a thermal equilibrium plasma.
()can permit a selection of the parameter a. Fitting the theoretical thermal

spectra can then yield ne, Te, and Ti in most cases.
One important consideration is the power required for the probing laser
pulse, because one of the strongest limitations to the application of this
diagnostic is the low level of the scattered light. Within a certain number of
assumptions (Sheffield, 1975), the ratio of the scattered power to incident
power is given by
(3.12)
where r 0 = 2.82 x 10-13 em is the classical electron radius, l the length of

plasma over which scattered light is collected, and d n the solid angle of the
collecting optics. For ne ~ 1019 cm- 3 (critical density for 10.6 11m light), l"'
0.01 em, and dn"' 10- 3 , we obtainP5 /Pi ~ 10- 11 • Only a small fraction ofthe
incident power is collected for diagnosis. Because of this small cross section,
the two most important problems encountered in scattering are parasitic or

stray radiation and plasma radiation.
Parasitic radiation, which is the radiation reaching the detectors by
routes other than scattering from the plasma, is probably one of the most
important limitations in the application of the Thomson scattering diag-
nostic of inertial confinement fusion (ICF)-type plasmas because reflection
from overdense plasma regions is difficult to eliminate. This problem is
even more severe in short wavelength irradiation experiments, in which the
distance between the solid target and the nc surface is very small. Parasitic
radiation can be partially eliminated by careful design of the focusing and
collecting optics. Because the parasitic radiation is monochromatic, a fur-
ther reduction of its level can be obtained by cascading narrow bandpass
filters in reflection. In general (and true for all types of plasmas), parasitic
radiation is greater for small(} because the collecting optics look nearly into
the source.
The second problem is the plasma radiation itself. The main contribu-
tion to this radiation comes from free-free electron-ion collisions, or brems-
strahlung. Because the emission is proportional to the square of the elec-
tron density, this problem increases in importance as the wavelength of the
laser producing the plasma becomes shorter (because nc "" .A - 2 ). Brems-
strahlung emission can be overcome by increasing the power of the probing
laser beam, as long as it does not perturb the plasma. Because the probing
laser pulse is of a shorter wavelength than the one producing the plasma,
the only severe perturbation is heating of the plasma. Nonlinear effects, as
well as ponderomotive effects, are not important unless the wavelength of
!sPROBE
i'r=~~~::.;;;:.:: __ -<1
SCATTERING
PROBE
I
t~
~SCAT.
\
Figure 3.14 Experimental configuration used to record the k-spectrum of fluctua-

tions on a streak camera (from Baldis and Walsh, 1983 with permission).
k/k 0
0
- 1 ns
1........ -••+-I--TIME --•~
Figure 3.15 Experimental spectrum showing the k-spectrum of electron plasma
waves produced by two-plasmon decay (from Baldis and Walsh, 1983 with permis-
sion).
the probing pulse is a harmonic of the main laser, as is the usual case in
short-wavelength experiments.
Thomson scattering is a natural diagnostic to study plasma waves and ion
waves induced by parametric instabilities. It measures directly the spectrum
of electron density fluctuations in the plasma; and if the conditions are ap-
propriate, one can identify waves associated with specific decay processes.
In the past, Thomson scattering has permitted a unique identification and
detailed study of TPD, SBS, SRS, and parametric decay. A review of these
measurements will be found in Baldis et al. (1986) and references therein.
All these experiments were possible due to the large frequency difference
between the pump and probe lasers. This frequency difference is due to the
requirement for the probe laser to be of much shorter wavelength than the
pump laser in order to minimize refraction of the probe beam and obtain a
suitable scattering vector.
An interesting addition to Thomson scattering has been the incorpora-
tion of wave-vector resolution. By working in the focal plane of the spec-
trograph, one can obtain w-resolved spectra. Alternatively, one can move
the entrance slit to the streak camera back from this focal plane and re-
cover the distribution of scattered light intensities over the face of the col-
lecting lens (Fig. 3.14). For a very-small-angle forward scattering, the an-
gle 8 at which light is scattered from the plasma fluctuation of wavevector
k depends on lk I, and so the procedure allows a k-resolved spectrum of
the fluctuations to be measured. The spectrograph can be set up to look
at either the ion or electron feature of the scattered signal, so the k spec-
trum of either ion acoustic waves or electron plasma waves can be mea-
sured.
A typical result is shown in Fig. 3.15 (Baldis and Walsh, 1983). It shows
the k spectrum of plasma waves generated by the TPD instability. The in-
stability was pumped by a C02 laser in a preformed plasma. The observed
spectrum is in good agreement with theory.
3.3 INTRODUCTION TO X-RAY DIAGNOSTICS
At irradiances above 1013 W/cm 2 , x-rays will be emitted in all portions of

the LP interaction diagrammed in Figs. 3.1 and 3.2. A wide variety of di-
agnostics, both broad and narrow band, have been devised that utilize this
radiation ranging in photon energy from 50 eV to near 1 MeV. X-ray mea-
surements can be divided into two broad (and partially overlapping) cate-
gories: spectroscopic and imaging.
Both broadband continuum and high-resolution atomic-line spectrosco-
PY are of value in LP measurements. Most information is obtained from the
atomic radiation of highly ionized species. As in other branches of plasma
spectroscopy, the goal is to use line intensities (both absolute and relative),
line profiles, and continuum structure as indicators of the atomic radiat-
ing environment. High-resolution line spectra in the range of hv;:::: 1 keY
(::S 12 A) are usually obtained with dispersing crystals (Boiko et a!., 1985),
whereas lower energies are measured with grazing-incidence (or transmis-
sion), diffraction-grating devices (Ceglio et a!., 1981; Richardson et a!.,
1985). Continuum measurements usually employ an array of filtered detec-
tors covering various broadband energy channels [for a description of hard
x-ray continuum systems, see Priedhorsky eta!., (1981); for soft x-rays, see
Pien eta!., (1986)]. At least three characteristics of energetic LP's distin-
guish them from other laboratory sources: the generally high-density condi-
tions (interesting regions almost always havene > 1018 cm- 3 and often much
greater); small spatial extent (usually below 1 mm 2 in cross section and of-
ten much smaller); and the highly transient temporal behavior, often on the
scale of a few tens of picoseconds. Many unique spectroscopic techniques
have been developed to deal with these spatial and temporal characteristics.
Detailed evaluation of the spatial distribution of x-ray emission is also

a very important source of information. As is described in Chap. 5, the
laser fusion application requires very uniform illumination of a target. The
spatial pattern of x-ray emission can provide a mapping of the uniformity
of optical absorption and the resultant plasma heating. Instabilities, both
plasma (in the absorption region, Fig. 3.1) and hydrodynamic (acceleration
and implosion region, Fig. 3.2), can also lead to characteristic spatial struc-
ture in the x-ray emission.
The transient character of x-ray emission from LP's has been dealt with
primarily by the development of x-ray streak cameras (Attwood, 1978;
Stradling, 1982; Richardson et al., 1985). The output of such devices is
essentially a two-dimensional (2D) intensity mapping. One dimension rep-
resents the temporal axis, whereas the other axis can be used either for spec-
tral or spatial information.
The vast majority of x-ray measurement work has employed self-emission
from the luminous laser or compressionally heated plasma. More recently,
absorption techniques have been developed and applied both to spectro-
scopic work and two-dimensional imaging. An external source for absorp-
tion methods can be created by irradiating a separate target with a portion
of the laser energy. In general, this target is of higher atomic number than
the subject and is irradiated at higher intensity. This provides for the back-
lighting radiation to be detectable above the self-emission background from
the subject target.
Due to the complexity of the LP interaction, the interpretation of many
x-ray diagnostic measurements is inevitably model-dependent. Fortunately,
a rather high degree of sophistication has been developed in the modeling
of the overall interaction that permits a prediction of many characteristic
x-ray signatures. The comparison between theory and experiment has, in
turn, produced in many cases important criteria for alteration and exten-
sion of the modeling. The starting point for such modeling is usually a hy-
drodynamic simulation code, such as LASNEX (Zimmerman and Kruer,
1975) that includes a model for (1) laser absorption, (2) electron and ra-
diation energy transport, and (3) hydrodynamic expansion and compres-
sion and many other aspects of the physics. Information from this sim-
ulation is then used by other atomic and radiation codes (Hauer et al.,
1983) to predict specific radiation signatures that are used for diagnos-
tics.
In this section dealing with x-ray measurements, we will first cover the
basic measurement techniques that can be employed using spectroscopic
and imaging methods. We then address applications of these techniques to
specific LP characteristics using the spatial structure pictured in Figs. 3.1
and 3.2 as a guideline.
3.3.1 X-Ray Spectroscopic Measurement Methods
Many of the traditional techniques of plasma spectroscopy (Griem, 1964;

Cooper, 1966; DeMichelis and Mattioli, 1981), such as line ratios and line-
to-continuum ratio measurements, etc., for determination of temperature
and line profile determination of density, have been applied to LP's. Be-
cause of the large temperature and density gradients and rapid temporal
variation, careful attention must be paid to the modeling that is used to in-
terpret the spectral quantities.
(A) Ionization and Excitation Dynamics

With the large variation in temperature and density in LP's, practically all
of the plasma excitation and deexcitation mechanisms are observed. It is
interesting to consider a few of these mechanisms in the context of the spa-
tial and temporal structure of LP's. Roughly speaking, in order to reach an
ionic state with ionization potential E;, the electron temperature must obey
(3.13)
with C 1 varying approximately between 0.1 and 0.3. Also important are the
time scales involved. The hydrodynamic time scales (lifetime of the plasma)
must be long compared to characteristic ionization and excitation times that
are governed either by collision or photon cross sections. Collisional ion-
ization (and excitation) processes are important in all portions of the LP
where Sz represents an ionic species with charge z. Collisional deexcitation

tends to only be important in the higher-density regions from critical inward
(except for t:.n = 0 transitions in the L and M shells).
All three types of recombination, radiative Sz + e-+ Sz_ 1 + hv, three-
body e +e +Sz -+Sz_ 1 +e, and dielectronic e + Sz -+Sz* (where the sym-
bol Sz* represents an ion in a doubly excited state), are observed in LP's.
Radiative recombination is the dominant mechanism for continuum gen-
eration in the less dense regions from the ablation surface outward, but is
also important in the denser regions. Three-body recombination starts to
be important in the denser portions of LP's, such as the compression region
(but does not contribute to continuum). Satellites of spectral lines, due to
the decay of the doubly excited state, are observed in all portions of the LP
but tend to be stronger (relative to the parent resonance line) in the lower-
density regions.
In order to interpret spectral characteristics, suitable models for exci-

tation and ionization must be constructed and linked to an overall model
of the LP interaction. In the higher-density regions from ablation to the
compressed core, local thermodynamic equilibrium (LTE) may be a useful
approximation (Griem, 1964). LTE implies a collision-dominated (CD) en-
vironment with the distribution between ionization states governed by the
Saha equation, and excited state populations are calculated from the Boltz-
mann distribution. For an atomic level in a particular ionic state to be in
CD equilibrium with higher levels, we use the following criterion:
(3.14)
where Z =ionic charge, n =principal quantum number, and EH = ioniza-

tion energy of hydrogen. As an example, the following situation has arisen
in the compressed core of LP's where radiation from argon in its hydrogen-
like state is often used to diagnose the compressed plasma conditions. For
then= 2level to be in CD equilibrium with higher levels, the electron den-
sity ne must be> 4x 1024 cm- 3, which is only reached in the highest compres-
sion experiments. On the other hand, for n = 3, ne must be > 1.1 x 1023 em - 3 ,
which is almost always attained in modern experiments. Similar modeling
considerations apply for low-temperature conditions (produced by irradi-
ance conditions"' 108-1011 W/cm2 ). It was found, e.g., in one experiment
involving chemical analysis (Radziemski et al., 1983; Cremers and Radziem-
ski, 1987) that the low temperatures (Te"' 1-2 eV) and dense conditions
justified use of the LTE approximations.
In the lower-density regions of the LP from critical outward, the coronal
approximation may apply. The coronal approximation assumes that pho-
toexcitation processes are small and that radiative decay rates are domi-
nant over collisional deexcitation. Populations are thus governed by a bal-
ance between collisional excitation and spontaneous emission (McWhirter,
1965). In the steady-state coronal approximation (if we neglect dielectronic
recombination), the ratio of densities between consecutive ionization states
is given by
(3.15)
where Ei(Z) is the ionization potential of the ith stage (in eV). For long-
pulse(> 500 psec) and moderate-intensity irradiation conditions (thus giv-
ing moderate blowoffvelocities and longer time scales), Eq. (3.15) may be
a useful approximation.
Unfortunately, there are many regions in LP's where neither the coro-
nal nor LTE approximations are appropriate. In these cases, a detailed ac-
counting of the excitation and deexcitation processes must be made. This
is generally referred to as a collisional-radiative (CR) model, first inspired
by Bates et al. (1962). This CR formulation for the ionization dynamics is
then incorporated in a comprehensive model that includes radiation, hy-
drodynamic, and plasma effects in a self-consistent manner. In some cases,
a steady-state intermediate between the LTE and coronal bounds may ap-
ply. In other cases, nonsteady-state conditions may necessitate the solution
of time-dependent rate equations. A description of the use of this type of
modeling in LP analysis is given later in this section.
(B) Line Profiles

Line profiles are perhaps the most thoroughly studied x-ray spectral char-
acteristic of LP's. Extensive calculations of hydrogenlike profiles have
been compared with experiments for nearly ten years (Yaakobi et al., 1977;
Kilkenny et al., 1980b; Hauer, 1981). These calculations generally employ
modeling that separates the profile into four basic components (Griem,
1974): (1) quasistatic Stark broadening due to the ion microfie1ds, (2) im-
pact broadening due to electrons, (3) Doppler broadening (natural broad-
ening due to the spontaneous lifetime and fine-structure splitting are often
negligible), and ( 4) opacity broadening [discussed in Sec. 3.3.1(c)]. The situ-
ation is often encountered in LP's where Doppler broadening is a small part
of the totallinewidth. This makes the profile less useful for ion temperature
measurements but simplifies density diagnostics. Typical quantities such as
the half-width (FWHM) are weakly dependent on temperature, but strongly
dependent on density. The dependence of the FWHM of hydro genic lines
can be approximated as
(3.16)
where C 2 = constant, ne = electron density, Z = charge of the emitter,

2/3 <a< 1 for a single-species plasma, and n1 and n 2 are principal quantum
numbers of the transition.
Sophisticated computer codes have been constructed that account for
quasistatic, impact, and Doppler contributions (Griem et al., 1979; Tighe
and Hooper, 1976; Lee, 1978; Woltz et al., 1982). Such codes have been
used extensively for the calculation of H-like profiles and their subsequent
use as density diagnostic tools. The theoretical results most useful to the
experimenter are often presented in the form of extensive tables. For in-
terpolating values not contained in the tables, an n 213 dependence is often
taken as adequate.
The first step in determination of density is to obtain a least-squares fit

of a theoretical profile to an experimental one. An example of such a fit-
ting is given in Fig. 3.16. The experimental spectrum was obtained from
the compressed core of a C02 laser-driven implosion. The points are from
a time-integrated spectrum and thus the fitting is a determination of the
temporally (and spatially) averaged density. In some situations, the use of
temporally averaged spectra as diagnostics yields very useful information.
The conditions that produced the spectral line in Fig. 3.16 are an example
of such a situation. The temperature and density conditions needed for sig-
nificant He-like argon emission lasted for only about 100 psec near peak
compression. The density measurement is thus a sampling that is roughly
indicative of peak compression.
Near the center, or core, of the line profile, the standard approximations
described above are less satisfactory. Here, ion dynamical effects due to
collective ion motion can cause additional broadening of the line core (Lee,
1979). An important recent development in line broadening theory (Stamm
and Talin, 1986) is a sophisticated treatment of ion dynamic effects using
particle-simulation computer codes. Unfortunately, near line-center opac-
ity often masks the ion dynamical contribution and makes detailed mea-
surements difficult.
C/)
(!) 0.30 EXPERIMENTAL THEORETICAL
a:
w POINTS PROFILE FOR
CD ne= s.ox 1 o23
0 0.20 cnf 3
>-
a:
>
z 0.10
-4 -2 0 2 4 6 8
RYDBERGS
Fi,ure 3.16 Fitting of a theoretical calculation to experimental points for argon
ls -ls4p.
Another important contribution since 1980 has been the development of

He-like and Li-like profiles as diagnostic tools. Calculations of the Stark
splitting of the He-like profiles are somewhat more difficult than the more
widely utilized (in LP's) H-like lines. Without the orbital angular momen-
tum degeneracy (of H -like species), the Hamiltonian for a two-electron con-
figuration is no longer diagonal in a single representation that is good for
all field strengths. An individual representation must thus be constructed
for each value of field (Griem, 1974). Despite the difficulties, however, cal-
culations (Kepple and Whitney, 1981) are available whose accuracy is very
close to that of H-like lines. The profile in Fig. 3.16 is an example of such
calculations.
The accuracy of determination of time-average densities (in LP's) with
line profiles has been discussed in several publications (Kilkenny et al.,
1980b; Hauer, 1981). In the compressed cores of LP's, the accuracy has
been shown to be about 50%. The effects of spatial gradients and temporal
variation are discussed later in this section. Line profile analysis has also
been very valuable in the diagnosis of the low-temperature LP's (Cremers
and Radziemski, 1987). The line emission typically comes from neutral,
singly, or doubly ionized species and lies in the optical region of the spec-
trum. The analysis (e.g., in determination of ne) is, however, quite similar
to that described above for x-ray measurements.
(C) Opacity and Radiation Transport

An important consideration for any spectral diagnostic is the transport of
emission from the region of radiation to the observing instrument. In the
case of line profile analysis, the transport essentially adds a broadening
mechanism referred to as "opacity broadening." If the opacity of the in-
tervening plasma is large, it may not be possible to make a unique deter-
mination of the plasma parameters (Delamater et al., 1985; Kilkenny et al.,
1980b). In addition, the large temperature and density gradients in many
LP's complicate the radiation transport. This once again points to the need
for comparison with detailed modeling. It is quite useful when several ex-
perimental observables (several line ratios, profiles, etc.) can be compared
with theory for a single laser experiment.
The radiative transfer equation can be written as
'VI (v,x) =J(v,x)- Kl (v,x) (3.17)
where/ is the intensity of radiation at a given frequency v and at a particular

point x in the plasma, J the emissivity, and K the absorption coefficient.
When LTE prevails, the transfer equation (in a homogeneous plasma) takes
on a simpler form and, e.g., the transported profile of a particular transition
can be written as
I(v) =B(v) { 1- exp [ -~~~v)]} (3.18)
Here, P(v) is the theoretical profile for the transition (labelled 1-2), B(v)
the blackbody function, and To the optical depth at line center, which can
be written
To= M/12 1re2 [n(1)- g 1 n(2)] P(O) (3.19)

me g2
where M is the characteristic spatial dimension of the plasma, n(1) is the

population of the ground state, n(2) is the population of the excited state,
f 12 is the absorption oscillator strength for the transition, andg1 andg2 are
the statistical weights. By fitting optically thick lines, it is possible to use
Eqs. (3.18) and (3.19) to obtain an estimate of the density scale length prod-
uct pM (Yaakobi et al., 1979). The determination of the quantity pM for
compressed cores (Fig. 3.2) is very important for the fusion application. In
fitting optically thick lines, however, care must be taken to understand the
errors involved.
The optical depth is usually the quantity used to define the opacity for
particular lines. In the less dense regions of the LP (near to or outward
from the critical surface), He- and H-like radiation from typical elements
like aluminum or silicon has several members of the series that can be con-
sidered optically thin, To :5 1. In the dense regions, such as the compressed
core, the first few members of the series for elements like argon are very
thick, whereas higher lines are merged with the continuum. This situation
often results in only one line being strong enough to measure and also op-
tically thin.
Even if a line can be assumed to be optically thin, spatial gradients in
the emission region can make diagnosis difficult (Duston et al., 1981; Lee,
1982). Several techniques have been developed to lessen the difficulty of
interpreting transport effects. One method is to seed the primary line-
emitting species into a background of lower-opacity material. For example,
the profile in Fig. 3.16 was obtained by using a target filled with 30 atm of
DT and only 0.2 atm argon. He-r was thus definitely optically thin [see also
Kauffman et al. (1984)]. Another technique (Herbst et al., 1982) is to place
a spatially localized dot, band, etc., of radiating material in a characteris-
tic position on a laser-irradiated target. These techniques have collectively
become known as "tracer dot" methods. These methods can often elimi-
nate the line-of-sight integration (i.e., emission is effectively from a point
source), while at the same time reducing opacity. This, of course, also re-
duces the radiated intensities, but they have been found to be adequate in
many useful experiments, one of which is described in Sec. 3.3.5. A thor-

ough analysis of the spectroscopic aspects of the tracer diagnostics has been
published recently by Apruzese et al. (1986).
(D) Satellite Line, Continuum, and Other

Spectroscopic Measurement Methods
One of the most prominent features of x-ray spectra from LP's is dielec-
tronic satellite lines adjacent to the resonant transitions (see Fig. 3.17), most
often but not exclusively appearing on the long-wavelength side. A great
deal of work has been done on dielectronic satellite radiation from LP's
since 1980 (Lunney and Seely, 1982; Audebert et al., 1984; Boiko et al.,
1982; Duston et al., 1983; Jacobs and Blaha, 1980). The most prevelant
mechanism for the generation of satellites in LP's is the decay of doubly
excited levels of the next lower ionization state from the parent line. As
Boiko et al. (1985) have noted, it is as if the "primary" transition charac-
terized by 1s -nl were taking place in the presence of an additional excited
,....
--
t
(Y)
C\!.
,....
1s 2 -
/1s3p 1s 2 -
1s 2 -
1s5p ,....
,.... 1s4p
t \ -
"'t
- -
C\1
,.... ,....
-- t
"'t
-
C\1
,.... ,.... ,.... ("')
t ( 1)0.. ,....
a. C'?,.... I
(/)a.ll)
-
- --
C\1 C')
"'t ,....
,.... ,.... ,....-
a. C\1
(/) C\1
(/) -C\1
-
("') C\1
,....
~ C\1 ,....
(/)
,....
\ (/)
5.7A 5.6A 5.5A 5.4A 5.3A 5.9 5.8 5.6 5.5

(a) (b)
0
-).(A)
Figure 3.17 Experimental silicon spectrum (a) and theoretical calculation (b) of
satellite lines.
electron n'l'.
lsn'l' -nln'l' transition in a He-like ion satellite to (ls -np) H-like

1s2n'l' -Isnln'l' transition in Li-like ion satellite to ls 2 -1snp He-like
The advantage of satellite lines as diagnostics is that they usually have

small optical depths (owing to the final state being other than the ground
state). Two principal mechanisms are responsible for population of the dou-
bly excited state: dielectronic capture and inner-shell excitation.
Modeling of satellite lines is quite complex and depends strongly on the
density regime involved. Excited-state population distributions vary be-
tween the coronal and LTE limits. Generally, the theory is the best de-
veloped for satellites of Ly.. and ls2 -1s2p. Unfortunately, large opacities
of these resonance lines often cause opacity-broadening overlap of some of
the important satellite transitions. This condition points to the desirability
of using higher n transitions for diagnostics. Use of higher n transitions,
however, increases the complexity of modeling. For example, the complex
of satellite lines adjacent to ls2 -ls3p contains more than 70 separate tran-
sitions.
It has also been shown by Duston and Davis (1980) that opacity effects,
although small in an absolute sense, must be accounted for when using satel-
lite line ratios as diagnostics. For example, they calculated the ratio of two
argon satellites that would be of potential value as a density diagnostic:
(1s2p 3p ---t 2p 2 3P)/(ls2s 3s ---t 2s2p 3P). They found this ratio to vary by
a factor of 2 between 1021 and 1023 cm- 3 (ion density) in the optically thin
approximation while varying by less than 1.4 in the thick case.
Despite these difficulties in modeling, considerable progress is being
made, and satellite lines can presently be regarded as a convenient though
rough indication of density having less accuracy than line profile determina-
tions. In Fig. 3.17, we show a comparison between an experimental silicon
spectrum and theoretical modeling performed by Scofield (1984). The spec-
trum was taken with a new high-collection-efficiency crystal spectrograph
(Hauer et al., 1984). This device allowed a more accurate measurement of
the weaker (but more isolated) satellites to the higher n lines.
Satellite lines are also observed on the short-wavelength side of the par-
ent transition. The inner-shell (K a) transition can, e.g., be shifted by ioniza-
tion of some of the outer electrons. Inner-shell radiation can be an impor-
tant diagnostic of suprathermal electron generation and transport (Hares
et al., 1979).
Measurements of continuum radiation are quite valuable for diagnosis
of temperature and other parameters in LP's. The most familiar is the use
of the exponential shape (as a function of frequency) as a temperature in-
dicator (Matthews et al., 1984). For determinations of the thermal electron

temperature, the continuum is largely due to recombination, whereas the
continuum generated by suprathermal electrons is due to bremsstrahlung.
Both of these techniques are discussed in application examples below.
Two other continuum techniques are available for very rough, but quick
and convenient, determinations of density: Inglis-Teller limit (Cooper,
1966) and edge shifts. As Eq. (3.16) indicates, the breadth of spectral lines
increases with principal quantum number and eventually blends in with the
continuum; this blending is known as the Inglis-Teller limit. The identifica-
tion of the first missing line can be used to estimate the density. Figure 3.18
shows this process of blending into the continuum for a laser-compressed
plasma. The right trace corresponds to a compressed electron density of
about 8 x 1023 cm- 3 •
The free-bound continuum will have discontinuities or edges at the ion-
ization limit for the various stages of ionization. The position of these
edges is density-dependent because of lowering of the continuum by the
free-electron potential. A complete calculation of the structure of absorp-
tion edges bears many similarities to line profile calculations. A complete
theory has, however, not yet been worked out. Edge shifts have, however,
been used for rough density determinations in laser compressions (Lee and
Hauer, 1978).
(E) Non-LTE Modeling

Considerable effort has been devoted to the development of collisional-
radiative modeling of non-LTE radiation and its application to LP diag-
nostics (Lee et al., 1984; Whitney and Kepple, 1982). Rate equations are
written for all of the ionization, recombination, excitation, and deexcitation
processes that are thought to be important and are coupled to the radiative
/LYMAN {3
2
/ 1 s -1s5p
3.86 3.90 3.94

PHOTON ENERGY (keV)
Figure 3.18 Progressive broadening of spectral lines in laser-driven compressions.

transfer equation. In many typical situations encountered in LP's, the com-

plexity is reduced by having to deal with only a few ionization states in detail.
The fast time scales and steep density gradients often cause only a few ion-
ization states to be important over the lifetime of the plasma. Two specific
examples of this situation will be given later in this section.
One of the major issues in such modeling is the choice of rate coeffi-
cients for the various processes. This has necessitated a careful study in
each particular case. Fortunately, a very large amount of x-ray diagnostic
work has involved H- through Li-like states of the Ne through Ar ions, and
a significant amount of work has been done in critiquing the accuracy of
available cross sections and rates for these ions (and calculating new ones).
The accuracy of electron impact ionization and excitation cross sections has
improved steadily over the past two decades as a result of improved theoret-
ical methods coupled to more powerful computers. In order of decreasing
accuracy and computer time, the computational methods range from mul-
ticonfigurational close coupling to the distorted wave and Coulomb-Born
approximations. Recently, there have been a number of attempts to system-
atize the calculations of rates needed in radiation-hydrodynamic modeling
(Clark, 1987 and Nilsen, 1988).
Complex atomic/radiative calculations are usually used to "post-process"
the temperature, density, etc. from the overall plasma/hydrodynamic mod-
eling of the laser interaction (Hauer et al., 1983; Lee et al., 1984). This, of
course, assumes that the detailed radiation processes are not a dominant
factor in the overall plasma evolution. In the case of a small amount of ma-
terial that is placed in a target for diagnostic purposes (e.g., the argon "seed-
ing" that produced the line in Fig. 3.16), this is a good assumption. When
considerable amounts of higher Z materials are present, this assumption
must be reexamined.
Non-LTE modeling has also been performed using an average ion ap-
proximation (Lokke and Grasberger, 1976). With such techniques, an esti-
mate of the non-LTE effects can be made as a calculation that proceeds in
line with the main hydrodynamic simulation.
3.3.2 X-Ray Imaging Measurement Methods
The spatial distribution of x-ray emission can provide important diagnos-

tic signatures of such LP processes as absorption and heating, transport of
energy by thermal and suprathermal particles, and laser-driven compres-
sion. In most cases, x-ray imaging is combined with some degree of spectral
discrimination.
Two-dimensional x-ray imaging is usually provided either with pinhole
cameras or grazing-incidence reflective microscopes. The best resolution
that can be obtained is of the order of a few microns and in more typical ex-
periments is 7 to 12 ~-tm (Attwood, 1978). The simplest type of spectral dis-
crimination is provided by foil filtration. By employing several foil-filtered
channels, two-dimensional images of various parts of a target emitting dif-
ferent lines (or bands) can be obtained. Grazing-incidence reflective optics
provide a high energy cutoff that when combined with foil filtration provides
a bandpass imaging system.
Imaging can also be combined with high spectral resolution, but often at
the sacrifice of one spatial dimension. A slit imaging system in conjunction
with a dispersive crystal is an example of such a system (Yaakobi and Nee,
1976; Boiko et al., 1985).
Two-dimensional x-ray imaging has only recently been combined with
temporal resolution through the development of fast framing cameras
(Dymoke-Bradshaw et al., 1983) and little experimental data have yet
been produced. One-dimensional time-resolved imaging (or selective cuts
through a two-dimensional image) can be obtained with a streak camera.
Spectral discrimination is also quite important from a calibration point
of view. Film is still the principal x-ray imaging detector, and its response as
a function of wavelength can be quite complicated (Henke et al., 1984). A
highly polychromatic image is thus very difficult to evaluate quantitatively.
Unfolding the two-dimensional images to obtain the volume distribu-
tion of emission involves many of the classical techniques of Abel inversion
(Mueller, 1979; Hauer, 1981). In adopting these techniques, account must
be taken of the large density and temperature gradients that exist in LP's. In
Fig. 3.19(a) is a typical time-integrated x-ray pinhole photograph of a laser-
imploded target. The initial position of the heated shell appears as a distinct
ring due to the limb-brightening effect. The compressed core is essentially
a spherical emitting volume. In Fig. 3.19(b), we show a time-resolved streak
photo of the same implosion.
The technique of using spatially localized emitters, mentioned earlier,
also helps to deal with steep plasma gradients and complicated line-of-sight
integrations sometimes encountered in imaging diagnostics. The placement
of small spots or bands of material at strategic positions on a target can be
used as a local diagnostic of plasma conditions. The tracer is chosen so that
its emission, when spatially and or spectrally resolved, will stand out against
emission from other parts of the target. An important consideration in the
use of a tracer/signature layer is the perturbation that the materials cause
to the interaction being studied.
Because of the very high densities that can be produced in laser compres-
sions, x-ray refractive effects may be important and have themselves been
used for diagnostic purposes (Miyanaga et al., 1982).
X-ray imaging diagnostics are often performed by comparison with
comprehensive modeling. In such cases, the aforementioned interac-
(a)
Figure 3.19 X-ray images of a laser driven implosion (a) time integrated (pin-
hole photo) (b) time resolved (streak photo). The target was a gas filled glass shell
and was uniformly irradiated at 0.35 J.lm on the Omega laser at the University of
Rochester.
tion/radiation modeling is used to directly predict the complete imaging-

intensity mapping. This process will be described further in the next sec-
tion.
3.3.3 Comprehensive Modeling of

X-ray Emission Signatures
Considerable progress has been made in using comprehensive laser interac-

tion modeling to predict particular characteristic x-ray spectral and spatial
emission features. A feedback process has thus been established, with ex-
periment and theory reinforcing one another and providing very important
diagnostic information.
Typically, the overall laser/matter interaction is modeled with hydrody-
namic codes such as LASNEX (Zimmerman and Kruer, 1975). A com-
plete description of the physics contained in such large models is out-
side the scope of this chapter [more details are given in the text and
ISODENSITY CONTOURS
700
600
500
-
400 -
VJ
0.
Lll
~
I-
300
200
100
DISTANCE IN TARGET PLANE (mm)

(b)
references of Chap. 5; see also Mead et al. (1984)], but a brief review
will be helpful in establishing the relationship to experimental diagnos-
tics.
The first step is formulation of a model for the laser absorption and scat-
tering process, such as inverse bremsstrahlung and refractive losses. Elec-
tron thermal transport is usually modeled by flux-limited diffusion (which
will be discussed further below), although more complex treatments have
been used. Suprathermal electron transport is typically treated by multi-
group diffusion in the presence of large electric fields.
Radiation transport is often handled by multigroup flux-limited trans-
port, but more sophisticated schemes are also employed. The effects of
departure from LTE on the overall energetics of the target are handled by
using an average ion modeling of the atomic processes. This calculation is
done in real time and the results are fed back into the hydrodynamic calcu-
lation.
Some x-ray diagnostic signatures, such as continuum slopes, can be ob-
tained directly from the overall laser interaction/hydrodynamic modeling.
More often, however, temperature, density, and other information from
the comprehensive model is "post-processed" by a model that treats the de-
tailed atomic processes and radiation transport. Table 3.2 gives a very brief
summary of the methods used to treat some important diagnostic emission
signatures.
The post-processing usually takes two forms. In one case, the plasma
data taken from the hydrodynamic code are evaluated for rough tempo-
rally averaged values of quantities, such as nToT' ne, and Te, which are then
manually entered into the detailed atomic models. This, of necessity, as-
sumes a quasi-steady state and the collisional radiative equilibrium (CRE)
approximation is used. Figure 3.20 shows an example of such modeling as
applied to a laser-driven implosion (Delamater et al., 1986).
In the second form, the detailed temperature and density information as
a function of time and space are automatically output to a file that is read by
the post-processing code (Hauer et al., 1983). This allows for more detailed
and time-resolved modeling.
Comprehensive modeling has also been applied to imaging diagnostics.
Models have been constructed that use input information such as den-
sity, temperature, opacity, as functions of space, and time from a hydrody-
namic simulation code. The models then perform line-of-sight integrations
through the plasma (at various photon energies) to produce a prediction
of the observable output emission. Over the years, numerous options have
been added to these codes to predict many aspects of x-ray images, such as
variation with filtration (spectral bandpass), instrument transfer function,
and line of sight, so that detailed comparison with experiments can be made.
Table 3.2 Summary of Some Techniques Used in Modeling X-ray

Diagnostic Emission Signatures
Plasma
Characteristics Emission Modeling
Diagnosed Signature Technique
1. T. (Thermal and Continuum Available online

suprathermal slopes from the radiation
package that is part
of the hydrodynamic
simulation
2. Thermal transport/ Intensities and Same as above
total radiation temporal behavior
conversion of bands of lines
3. T. (thermal), n. Individual n., T. from overall
atomic-line hydrodynamic code "post-
intensities and processed" by detailed
ratios for atomic/radiation
resonance lines (T.) models
and satellite lines (n.)
4. n.,pM Line profiles Same as above
5. T., n. in cooler Absorption-line Available online
regions intensities (in LTE) from the use of
detailed opacity
tables
6. Spatial distribution Narrow-band n., T. data "post-
of plasma heating two-dimensional processed" by image
x-ray images simulation model
3.3.4 Absorption Region Phenomena
The region between the ablation and critical surfaces in aLP is a very strong
source of x-ray emission. As much as 70% of the absorbed laser light has
been observed in some experiments to be reradiated as soft x-rays. Such
copious emission is important in its own right for many applications in ad-
dition to diagnostics.
Figure 3.21 shows an example of the use of soft x-ray emission in the
diagnosis of the uniformity of laser absorption and resultant plasma heat-
ing. A gold-coated spherical target was irradiated with eight overlap-
FROM HYDRO
SIMULATION
Te = 1100 eV
ne = 1.5 x 10 23 em- 3
3600 3800 4000

ENERGY (eV)
Figure 3.20 Model (solid line) of experimental (dotted line) argon spectrum using
CRE. Spectrum was obtained from a laser driven implosion (from Delamater eta!.,
1986 with permission).
ping C0 2 laser beams, which resulted in relatively uniform illumination

at about 4 x 1015 W/cm2 • A number of explanations have been advanced
for the filamentary structure including (1) thermal or ponderomotive self-
focusing, (2) constrictions of the blowoff due to magnetic field generation,
and (3) subtle nonuniformities in the gold, causing portions to be blown
off preferentially. Similar filamentary structures have been observed using
optical probing diagnostics (Willi and Rumsby, 1981).
Spectroscopic and imaging techniques have been applied to the determi-
nation of the density and temperature profiles in LP's, such as those shown
in Fig. 3.3 (Boiko et al., 1979). Spatially resolved spectral measurements
are used to produce temperature (typically with line ratios) and density
(typically using line profiles) plots. X-ray and optical techniques are com-
plementary for such diagnostics. The optical methods have higher spatial
resolution, whereas the x-ray techniques can be extended to much higher-
density regimes.
The use of localized tracers and seeding has had considerable impact
on such measurements (Burkhalter et al., 1983). A spectral line com-
ing from a tracer is being emitted from a well-defined spatial volume,
and thus it is not necessary to fold in a complex integration (through re-
gions of varying plasma parameters) in order to predict the signal seen
by a particular instrument. The tracer dot stays together as a coherent
entity only in the more collisional (higher-density) regions. At consider-
Figure 3.21 Filamentary structures observed in soft x-ray emission (""' 0.5 ke V).
Gold targets were uniformly irradiated with eight C0 2 laser beams.
able distances from the critical surface, the usefulness of tracers is thus re-
duced.
3.3.5 Energy 'fransport
Once laser light has been absorbed, the next important step in the interac-
tion is the transport of this energy by three principal mechanisms: (1) ther-
mal, (2) suprathermal electrons, and (3) soft x-ray radiation.
The simplest model for thermal electron transport is based on a flux-
limited diffusion picture (Spitzer and Harm, 1959; Shvarts, 1985). In
some laser experiments, transport consistent with large-flux limitation (sig-
nificant departures from classical diffusive transport) is observed. Eluci-
dation of mechanisms that could be responsible for such flux limitation
is of considerable interest. In order for us to model the whole inter-
action, the flux-limited diffusion picture is inserted in the overall hydro-
dynamic simulation code (as described in Sec. 3.3.3). In order to have
confidence in such complex modeling, it needs to be compared with as
many experimental observables (such as the temporal profile of Te and ne
at a particular position in the target) as possible. The temperature and

density profiles, as were schematically indicated in Fig. 3.3, are closely
tied to the thermal transport and represent another important bench-
mark.
X-ray spectroscopic and imaging methods have been the primary source
of information on thermal and supra thermal transport (Yaakobi et al., 1984;
Goldsack et al., 1982; Mead et al., 1984). Very often, a layered or structured
target, such as that shown in Fig. 3.22(a), is used to diagnose transport.
The radiation from a particular layer is used as a signature of the energy
transported to that region. Although transport by thermal and suprather-
mal electrons represent very different phenomena, there is some similarity
in the methods used to diagnose them, especially in the use of layered and
structured targets.
A recent experiment performed at the Rutherford-Appleton Laboratory
(conducted jointly by Imperial College, Rutherford, and Los Alamos) illus-
trates many of the techniques used in x-ray transport (and absorption) di-
agnosis. The type of target used is illustrated in Fig. 3.22( a) and employed
a thin aluminum dot (on other shots other materials were used) coated on a
solid plastic (CH 2 ) sphere and then overcoated further with various thick-
nesses of CH 2 • The x-ray emission from the aluminum is both spectrally and
temporally resolved. The x-ray emission measurement is accurately refer-
enced with respect to the temporal profile of the heating laser. At irra-
diances above 1014 W/cm2 , the onset of emission tends to be fairly sharp
and provides a relatively well-defined signature of the interval needed to
ablate away the plastic overcoating. In Fig. 3.22(b ), we show a streak cam-
era record of the ls 2 - ls4p aluminum line. The interval /}.t in the figure is
related to an average mass ablation rate (g/cm 2 /sec) by
m = m/f}.t (3.20)
where ih is the areal density of the plastic overcoat in Fig. 3.22(b ).

Other x-ray lines from aluminum (or other materials) may yield a slightly
different value for /}.t. Because different energy x-ray lines are excited by
different average plasma temperatures [see Eq. (3.13)], the different delays
yield information on the structure of the thermal heat front.
Figure 3.23 shows an example of modeling the x-ray signatures of trans-
port with the tools described in Sec. 3.3.3 for the particular case of the exper-
iment shown in Fig. 3.22. The onset of aluminum line emission (integrated
over all H-and He-like lines) is calculated using the average ion non-LTE
model that runs online with a hydrodynamic code. This yields an average
m that agrees well with the experiment. Prediction of the emission from
individual AI lines would require the more detailed atomic post-processing
calculation.
AL·DOT (rtf= 75 p.m,

t=0.1p.m)
(a)
TIME
T .._FIDUCIAL
~t
_L !t
T
1
1ns
__ _,..~A
I I I 0
-25 0+25mA
(b)
Figure 3.12 (a) Target configuration for thermal transport measurements; (b) x-ray
streak photo of AI ls 2 - ls4p line.
T= 500 ps T=soo ps
RELATIVE
INTENSITY
1.0 2.0 3.0 1.0 2.0 3.0 1.0 2.0 3.0
h II (keV)
Figure 3.23 Simulation of the onset of aluminum line emission as a diagnostic of
thermal electron transport.
In addition to overall quantities such as the mass ablation rate, more

information is available from the details of the radiation spectrum in
Fig. 3.22(b ). A special high-efficiency, high-dispersion crystal spectrograph
(Hauer et al., 1984) was coupled to an x-ray streak camera to obtain this
spectrum. This coupling enables observation of the details of the line as a
function of time. From this profile, an estimate of the electron density can
be made and compared with theoretical modeling.
Recently (Kephart et al., 1987), layered-target burnthrough/transport ex-
periments have been performed in the low-power regime ("' 1010 W/cm2 ) In
this lower-temperature regime, optical emission was used as the diagnostic
of the thermal transport. Comprehensive hydrodynamic/radiation model-
ing very similar to that used in Fig. 3.23 was employed for analysis.
As was mentioned earlier, some laser absorption processes result in the
production of suprathermal electrons. These electrons are deleterious to
the fusion application, causing preheat of the DT fuel and degrading the
compression. In addition to directly penetrating the dense portions of the
target, these electrons can take circuitous paths, including orbiting the tar-
get (Hauer et al., 1984). As is indicated in Fig. 3.1, the initial blowoff of
electrons and ions can form a plasma sheath that reflects some of the hot
electrons back toward the dense regions. These supra thermal electrons pro-
duce bremsstrahlung and inner-shell line radiation, which are the princi-
pal means for diagnosis. Some further details on these suprathermal ra-
diation processes are given in Chap. 10. Because the radiation is occur-
ring in the dense portions of the target, diagnosis is taking place in a re-
gion quite far separated from the area of generation (typically from nc to
nJ4). The suprathermals penetrate ahead of the thermal front and ex-
cite species that are relatively cold (low ionization states). Spectroscopy
of the inner shell (e.g., K a) and bremsstrahlung radiation can yield infor-
mation on the total energy and spectrum of the suprathermal distribution.

A two-dimensional pattern of suprathermal deposition can be obtained
by filtering an imaging device, such as a pinhole camera, to pass primar-
ily the inner-shell or high-energy bremsstrahlung radiation. An example
of imaging diagnosis is given in Fig. 3.24. In this experiment, C0 2 laser
radiation is incident on a plastic shell with an inner core of solid nickel.
The left-hand image (b) is filtered to block nickel Ko: (>. = 1.54 A), while
the right (c) passes it. Both filtered channels passed x-rays of higher en-
ergy than Ni K o: and the contribution from this radiation has been digi-
tally subtracted. In addition to the spatial deposition pattern, an estimate
of the total suprathermal-deposited energy can be obtained from such pic-
tures.
As was described in the introduction and Sec. 3.2, the laser absorption
process can result in the production of very large magnetic fields. These
fields, in turn, can affect both thermal and suprathermal electron transport
(Yates et al., 1982). A dramatic signature of the effect of magnetic fields
on suprathermal transport is given in Fig. 3.25. In this case, a solid cop-
per sphere has been irradiated by eight high-intensity C0 2 laser beams. A
toroidal magnetic field encircling the axis of each of the beams is gener-
ated. Suprathermal electrons tend to deposit their energy along nodes of
the composite magnetic field (superposition of the eight fields). The pin-
hole camera that produced this image was filtered to pass copper K o: and
some of the high-energy bremsstrahlung produced by suprathermal depo-
sition.
3.3. 7 Implosion Hydrodynamics
The laser ablation process can drive large shocks (> 20 Mbar) and rapid
(> 3 x 107 em/sec) material motion. Diagnosis of laser-driven acceleration
and implosion is, of course, central to the fusion application, but may also be
of interest in other applications such as laser acceleration of macroparticles
and basic materials studies.
In the process of laser-driven acceleration and compression, it has long
been recognized that basic fluid instabilities may lead to the breakup and
distortion of the material being driven. When a light fluid is driving (sup-
porting) a heavier, denser fluid, the possibility exists for growth of the
Rayleigh-Taylor instability (McCrory et al., 1981; Emery et al., 1982). This
growth can happen in the laser case in two more or less distinct phases,
which can be referred to as acceleration and deceleration. The laser abla-
tion and acceleration process is created by a hot tenuous fluid pushing on
the denser material (e.g., a shell or plate), creating the potential for instabil-
ity. When a gas-filled shell is compressed, the shell will eventually stagnate
(b)
Figure 3.24 Use of inner shell x-ray radiation in the diagnosis of suprathermal
electron transport. (a) target configuration: hollow plastic shell with a solid nickel
ball suspended on a thin web inside (laser is incident on the outside of the CH shell);
(b) x-ray image filtered to reject NiK"'; (c) x-ray image filtered to pass NiK"'.
(c)
against the compressed gas, creating an unstable situation during decelera-

tion. One approach that has been taken in diagnosing such phenomena has
been to compare the behavior of targets that have intentionally introduced
perturbations (e.g., rippled surfaces) with nominally perfect ones. The per-
turbations would provide a "seed" for instability growth of a particular spa-
tial scale length.
Almost all instability investigations have involved x-ray techniques (Kil-
kenny et al., 1980a; Grun et al., 1984), usually involving some type of imag-
ing although the use of spectroscopic techniques has also been reported
(Hauer and Gitomer, 1984). One category of diagnostic can be referred to
as two-dimensional interface tracking. A laser-accelerated plate is radio-
graphed (perpendicular to the direction of motion) with an x-ray backlight-
ing source. By irradiating at particular characteristic times, the geometrical
configuration of the growth of perturbations can be observed and compared
with theory. An interesting example (Whitlock et al., 1984) of this kind of
measurement is given in Fig. 3.26. Small perturbations were introduced into
the targets, as illustrated in Fig. 3.26(b). The spike structures (Fig. 3.26(a)]
are thought to be evidence of the growth of Rayleigh-Taylor instability. This
experiment diagnosed the acceleration phase.
Similar diagnostic development work has been performed in imploding
geometry, as is illustrated in Fig. 3.27 (Hauer and Gitomer, 1984). In this
experiment, cylindrical metal shells coated with plastic were imploded with
1NC1DENT
LASER,
BEAMS
MAGNETIC
FIELD
NODAL
UNES
Figure 3.25 Diagnosis of magnetic-field-mediated electron transport with x-ray

imaging (from Hauer et al., 1984 with permission).
C02 laser irradiation. X-ray imaging of self-emission was performed both

along and orthogonal to the cylindrical axis. In some of the targets, small
perturbations (of two different kinds) were manufactured into the targets,
as shown in Fig. 3.27(b). Figure 3.27(c) shows the typical difference be-
tween x-ray images of the implosion of nominally perfect and intentionally
perturbed targets. In this experiments, the pinhole camera was filtered to
emphasize radiation around 1.5 keV. First, one can see that the compressed
core is much larger in the perturbed case, indicating a weaker compression.
In addition, the compressed core is surrounded by a halo of radiating ma-
terial, which has a structured pattern (not readily visible in prints) that may
be indicative of instability behavior.
3.3.8 Very Dense Plasmas
Truly remarkable plasma conditions, with electron temperatures of about 1

keV or greater and densities of at least 30 g!cm3 , can be obtained in laser-
driven implosions. Line profile analysis has played a prominent role in
the diagnosis of such plasmas (Yaakobi et al., 1980; Hauer, 1981; Key and
Hutcheon, 1980). Figures 3.16, 3.18, and 3.20 are examples of line emis-
sion from very dense plasmas. A new technique (Hauer et al., 1986) that
RP BR
==~--sp
__/~-ST
SPIKES BUBBLES
LASER
(a)
~
(b)
Figure 3.26 X-ray radiographic diagnosis of instabilities in the laser acceleration

of a flat plate (from Whitlock et al., 1984 with permission).
has been developed for diagnosis of laser implosions is absorption spec-

troscopy. One configuration that has been used in such studies is shown
in Fig. 3.28. In this case, the absorption spectra are produced by atomic
species in the gas-containing shell or pusher of the target. The pusher is
almost always much colder than the central gas core. Continuum radiation
from the gas passes out through the pusher, which imposes atomic-line ab-
sorption features on it. The absorption coefficient for a particular atomic
transition can be written as
(3.21)
where P(v) is the line profile and/ the absorption oscillator strength. The
attenuated intensity I(v) through a layer of thickness 6.R and density pis
LASER IMPLOSION
\t-
, \
CYLINDRICAL
X-RAY AXIS
IMAGING------.
(a) (b)
WITHOUT WITH
PERTURBATIONS PERTURBATIONS
(c)
Figure 3.27 Diagnosis of cylindrical implosions with x-ray self emission (a) exper-
imental configuration, (b) structure of perturbation (wall thickness "' 40 p.m, per-
turbation amplitude"' 2 p.m, .X"' 25 p.m), (c) images of implosions with and without
perturbations.
related to the incident intensity by
I(v) =/0 exp( -u11 apD..R/M) (3.22)
where a is the fraction of all ions in a particular state andM the atomic mass.
Equation (3.22) must be summed over the contributions from all ionization
states that are present. Finally, we can write an expression for pD..R as
pD..R = M jln(l0 /1 11 )dv (3.23)

u0 a
Cold Pusher
' \ Observed
\ Spectwm
/-...!.._ ~~~~rce
I~ctrum I 1
hv hv
f\ ~/Ar-Lya
~
1s-2p
\~
Features
\
K Absorption
Features
Figure 3.28 Absorption spectrum produced by pusher of laser imploded target.
It is important to note from Eq. (3.23) that only the integral over the line
profile is required for the determination of pM and it is not necessary to
know the details of the line profile.
->Q)
-
.:zt.
.....
...,
w 10-1
()
z
w
::l
...J
u..
>-
<
a:
I
X
PHOTON ENERGY (keV)

Figure 3.29 Simulation of absorption spectra for the experimental case in Fig. 3.28.
(from Hauer, et al., 1986 with permission).
For diagnostic purposes, values of pD.R from Eq. (3.23) can be compared
to (time-integrated) values obtained from hydrodynamic simulation. In ad-
dition, one can use detailed opacity tables online with hydrodynamic sim-
ulation (item 5 in Table 3.2) to predict the complete absorption spectrum.
An example of such a prediction is given in Fig. 3.29, which is a simula-
tion of the experimental spectrum in Fig. 3.28. Through such modeling and
experimental comparison, other plasma parameters (such as the electron
temperature from the ratio of the absorption peaks) can be evaluated.
As even higher compressed densities are reached in the near future, sig-
nificant diagnostic challenges will be posed. Very high pD.R's and resultant
high opacities will make observation of self-emission difficult. Absorption
spectroscopy using external sources (with high photon energy) will hope-
fully extend the useful range of x-ray diagnostics. In addition, nonideal be-
havior of the plasma (More, 1986; Burgess, 1983) will require a reevalua-
tion of the methods (with respect to atomic structure, line profiles, radia-
tion transport, etc.) used to calculate spectral characteristics. On the other
hand, these conditions represent an exciting new frontier for plasma and
atomic physics.
ACKNOWLEDGMENTS
The authors would like to thank the following individuals who carefully
read this chapter and offered many useful and helpful suggestions: J. Davis,
D. Duston, P. Goldstone, H. Griem, J. Bernard, and B. Yaakobi. We also
thank the editors of the present volume for many helpful comments. We
would also like to acknowledge the careful editing and proofreading of
B. Thompson, P. Atencio, K. Serna, and J. Weber, and the excellent artwork
of S. Carlson and C. Boone.
BIBLIOGRAPHY
The reader is referred to Laser-Plasma Interactions 1 (1980) (R. A. Cairns

and J. J. Sanderson, eds.) Scottish Universities Summer School in Physics,
Edinburgh, Scotland. This volume is the first of three monographs that
taken together form an excellent review of the basic physics of laser-plasma
interaction including in-depth discussions of absorption and scattering pro-
cess, transport, hydrodynamics, and many other subjects.
The other two volumes in the series are: Laser-Plasma Interactions 2
(1983) (R. A. Cairns, ed.) and Laser-Plasma Interactions 3 (1986) (M. B.
Hooper, ed.). In the references, the publisher is abbreviated SUSSP Publi-
cations.
Another excellent introduction to the physics of laser-plasma interac-

tions with many useful references is the following: Max, C. E. (1982).
Physics of the coronal plasma in laser fusion targets, in Interactions Laser-
Plasma, Les Houches, (R. Balin and J. C. Adam, eds.), North Holland.
REFERENCES
Aldrich, C. H., Bezzerides, B., DuBouis, D. F., and Rose, H. A. (1986). Langmuir
nucleation and collapse in stimulated laser light scatter, Comments Plasma Phys.
Controlled Fusion, 10: 1.
Apruzese, J.P., Duston, D., and Davis, J. (1986). K-shell aluminum resonance line
ratios for plasma diagnosis using spot spectroscopy, J. Quant. Spectrosc. and Ra-
diation Transfer, 36: 339.
Attwood, D. (1978). Diagnostics for the laser fusion program, IEEE J. Quant. Elec-
tron., QE-14: 909.
Audebert, T. P., Geindre, J. P., Gauthier, J. C., Alaterre, P., Popovics, C., Cornille,
M., and Dubau, J. (1984). Dense plasma effects on dielectronic satellites, Phys.
Rev. A., 30: 1582.
Baldis, H. A. and Walsh, C. J. (1983). Growth and saturation of the two-plasmon
decay instability, Phys. Fluids, 26: 1364.
Baldis, H. A., Walsh, C. J., and Benesch, R. (1983). Temporal and spectrally resolved
measurements at 10.6 micron,Appl. Optics, 22: 2217.
Baldis, H. A., Villeneuve, D. M., and Walsh, C. J. (1986). Plasma waves in laser
fusion plasma, Can. J. Phys., 64: 961.
Bates, D. R., Kingston, A. E., and McWhirter, R. W. P. (1962). Recombination be-
tween electrons and atomic ions. I. Optically thin plasmas, Proc. Roy. Soc. A, 267:
297.
Bekefi, G., Deutsch, C. B., and Yaakobi, B. (1976). Spectroscopic diagnostics of laser
plasmas, in Principles of Laser Plasmas, Wiley, New York, p. 549.
Benattar, R. and Popovics, C. (1983). Time-resolved evolution of laser-produced
plasmas in spherical expansion regime, J. AppL Phys., 54: 609.
Benattar, R., Walsh, C. J., and Baldis, H. A. (1983). Time resolved schlieren study
of highly refracting C0 2 laser produced plasma, Optics Comm., 47: 324.
Bobin, J. L., Decroisette, M., Meyer, B., and Vitel, Y. (1973). Harmonic generation
and parametric excitation of waves in a laser-created plasma, Phys. Rev. Lett., 30:
594.
Bockasten, K. ( 1961 ). Transformation of observed radiances into radial distributions
of the emission of a plasma, J. Opt. Soc. Am., 51: 943.
Boiko, V. A., Pikuz, S. A., and Faenov, Y. (1979). Determination of laser plasma
electron density by Kspectra ofmulticharged ions, f. Phys. B., 12: 1889.
Boiko, V. A., Maigrov, S. A., Pikuz, S. A., Skobelev, 1., Faenov, Y., and Shilov, K. A.
(1982). Strengths of dielectronic satellites of the resonance line of He-like ions
in an optically thick plasma, Opt. Spec., 52: 259.
Boiko, VA., Vinogradov, A. V, Pikuz, S.M., Skobelev, I. Y., and Faenov, Y. (1985).
X-ray spectroscopy of laser produced plasma, J. Sov. Laser Res., 6: 85 (published
in Russian in 1978).
Burgess, D. D. (1983). The spectroscopy of dense laser-generated plasmas, in Laser-
Plasma Interactions 2 (R. A. Cairns, ed.) SUSSP Publications, Edinburgh, Scot-
land, p. 219.
Burgess, M.D. J., Dragila, R. and Luther-Davies, B. (1984). Complex second har-
monic emission spectra from plasmas produced using short pulse, high intensity
one micron laser radiation, Optics Comm., 50: 236.
Burkhalter, P. G., Herbst, M. J., Duston, D., Gardner, S., Emery, M., Whitlock,
R. R., Grun, J., Apruzese, J. P. and Davis, J. (1983). Density and temperature
profiles within laser produced plasmas, Phys. Fluids, 26: 3650.
Burnett, N. H., Baldis, H. A., Richardson, M. C., and Enright, G. D. (1977). Har-
monic generation in C0 2 laser target interaction, Appl. Phys. Lett., 31: 172.
Busch, G. E., Shepard, C. L., Siebert, L. D., and Tarvin, J. A., (1985). Four-frame
holographic probing system for plasma density measurement, Rev. Sci. Instrum.,
56:879.
Cable, M. D., Lane, S. M., Lerche, R. A., Prussin, S. G., and Vieira, R. G. (1986).
Neutron spectroscopy on Nova, Rev. Sci. Instrum., 57: 1729.
Cairns, R. A. (1983). An introduction to the absorption of laser light in plasmas, in
Laser-Plasma Interactions 2 (R. A. Cairns, ed.) SUSSP Publications, Edinburgh,
Scotland, p. 1.
Carman, R. L., Rhodes, C. K., and Benjamin, R. F., (1981). Observation of harmon-
ics in the visible and ultraviolet created in C0 2 -laser produced plasmas, Phys.
Rev. A, 24: 2649.
Ceglio, N. M., Hawryluk, A.M., and Price, R. H. (1981). Space and time resolved
soft x-ray spectra using x-ray transmission gratings SPIE Proc., 316: 134.
Clark, R. W. (1987). Radiation-hydrodynamics and detailed atomic data bases, Proc.
of 6th APS Conf. on Atomic Processes in High Temperature Plasmas; also Los
Alamos Nat. Lab. Pub. LALP-87-30, p. 2A3. (Several other valuable papers on
atomic databases were presented at the cited conference; see, e.g., the paper by
J. Nilsen.
Cooper, J. (1966). Plasma spectroscopy, Rept. Prog. Phys., 29: 35.
Cremers, D. A., and Radziemski, L. J., (1987). Laser plasmas for chemical analysis,
in Laser Spectroscopy and its Applications (L. J. Radziemski, R. W. Solarz, and
J. A. Paisner, eds.), Marcel Dekker, New York, p. 351.
Delamater, N., Hooper, C. F., Joyce, R. F., Woltz, L. A., Ceglio, N. M., Kauffman,
R. L., Lee, R. W., and Richardson, M. C. (1985). Opacity effects on hydrogen-like
lines emitted from laser-driven implosions, Phys. Rev. A., 31: 2460.
Delamater, N., Hauer, A., Hooper, C. F., Garber, Richardson, M., Jaanimagi, P., and
Barnouin, 0. (1986). Spectroscopic analysis of 0.35 micron omega implosions of
Kr/Ar filled plastic shells, BulL Am. Phys. Soc., 31: 1460.
DeMichelis, C., and Mattioli, M. (1981 ). Soft x-ray spectroscopic diagnostics of lab-
oratory plasmas, Nuc. Fusion, 21: 677.
Drake, J. F., Kaw, P. K., Lee, Y. C., Schmidt, G., Liu, C. S., and Rosenbluth, M. N.
(1974). Parametric instabilities of electromagnetic waves in plasmas, Phys. Fluids,
17:778.
Drake, R. P., Young, P. E., Williams, E. A, Estabrook, K., Kruer, W. L., Lasinski, B.
F., Darrow, C. B., Baldis, H. A, Johnston, T. W (1988). Laser intensity scaling in
long scalelength laser produced plasmas, Phys. Fluids 31: 1795.
Duncan, D., Luckin, W. D., and Willi, 0. (1981). Rutherford-Appleton Lab. Annual
Rep., p. 448.
Duston, D. and Davis, J. (1980). Self-absorption of helium-like satellite lines, Phys.
Rev.A,21: 932.
Duston, D., Davis, J., and Kepple, P. C. (1981). Usefulness of x-ray diagnostics in
dense-plasma gradients, Phys. Rev., 24: 1505.
Duston, D., Rogerson, J. E., Davis, J., and Blaha, M. (1983) Dense plasma effects
on K-shell dielectronic satellites lines, Phys. Rev. A, 28: 2968.
Dymoke-Bradshaw, A K. L., Kilkenny, J. D., and Westlake, J. (1983). A gated x-
ray intensifier with a resolution of 500 psec, Proc. lOth European Conf on Laser
Interactions with Matter, London, p. 152.
Emery, M. H., Gardner, J. H., and Boris, J.P. (1982). Rayleigh-Taylor and Kelvin-
Helmholtz instabilities in targets accelerated by laser ablation, Phys. Rev. Lett.,
48:677.
Forslund, D. W, Kindel, J. M., and Lindman, E. L. (1975). Plasma simulation studies
of stimulated scattering processes in laser-irradiated plasmas, Phys. Fluids, 18:
1017.
Forslund, D. W, Kindel, J. M., and Lee, K. (1977). Theory of hot-electron spectra
at high laser intensity, Phys. Rev. Lett., 39: 284.
Garban-Labaune, C., Fabre, E., Max, C., Amiranoff, F., Fabro, R., Virmont, J., and
Mead, W C. (1985). Experimental results and theoretical analysis of the effect
of laser wavelength on absorption and hot electron generation in laser-plasma
interaction, Phys. Fluids, 28: 2580.
Goldman, M. V. (1966). Parametric plasmon-photon interactions, Ann. Phys., 38:
117.
Goldsack, T. J., Kilkenny, J.D., Macgowan, B. J., Cunningham, P. F., Lewis, C. L. S.,
Key, M. H., and Rumsby, P. T. (1982). Evidence for large heat fluxes from the
mass ablation rate of laser-irradiated spherical targets, Phys. Fluids, 25: 9.
Griem, H. R. (1964). Plasma Spectroscopy, McGraw-Hill, New York (available from
University Microfilms).
Griem, H. R. (1974). Spectral Line Broadening by Plasmas, Academic Press, New
York.
Griem, H. R., Blaha, M., and Kepple, P. (1979). Stark-profile calculations for Lyman
series lines, Phys. Rev. A, 19: 2421.
Grun, J., Emery, M. H., Kacenjar, S., Opal, C. B., McLean, E. A, Obenschain, S. P.,
Ripin, B. H., and Schmitt, A (1984). Observation of Rayleigh-Taylor instability
in ablatively accelerated foils, Phys. Rev. Lett., 53: 1352.
Hall, T. A (1985). Diagnostics of laser produced plasmas, in Laser-Plasma Interac-
tions 3 (M. B. Hooper, ed.), SUSSP Publications, Edinburgh, Scotland, p. 341.
Hares, J.D., Kilkenny, J.D., Key, M. H., and Lunney, J. G., (1979). Measurement
of fast-electron energy spectra and preheating in laser-irradiated targets, Phys.
Rev. Lett., 42: 1216.
Hauer, A (1981 ). Diagnosis of high density laser compressed plasmas using spectral
line shapes, in Spectral Line Shapes (B. Wende, ed. ), W DeGruyter, Berlin, p. 298.
Hauer, A., Whitney, K. G., Kepple, P. C., and Davis, J. (1983). Detailed observation
and analysis of radiation from high-density laser-imploded targets, Phys. Rev. A,
28:963.
Hauer, A., Kilkenny, J.D., and Landen, 0. (1984). Torodially curved crystal for time
resolved x-ray spectroscopy, Rev. Sci. Inst., 56: 803.
Hauer, A. and Gitomer, S. (1984). Development of diagnostics of shell break up in
laser driven implosions, Bull. Am. Phys. Soc., 29: 1387
Hauer, A., Goldman, R., Kristal, R., Yates, M. A., Mueller, M., Begay, E, Van Hul-
steyn, Mitchell, K., Kephart, J., Oona, H., Stover, E., Brackbill, J., and Forslund,
D. (1984). Superthermal electron generation transport and deposition in C0 2
laser irradiated targets, in Laser Interaction and Related Plasma Phenomena, Vol.
6 (H. Hora and G. R. Miley, eds.) Plenum Press, New York.
Hauer, A., Cowan, R. D., Yaakobi, B., Barnouin, 0., and Epstein, R. (1986).
Absorption-spectroscopy diagnosis of pusher conditions in laser-driven implo-
sions, Phys. Rev. A, 34:411.
Henke, B., Kwok, S. L., Uejio, J. Y., Yamada, H. T., and Yound, G. C. (1984). Low
energy response of photographic films, J. Opt. Soc. Am., 1: 818, 828.
Herbst, M., Burkhalter, P. G., Grun, J., Whitlock, R., and Fink, M. (1982). Spot
spectroscopy: Local spectroscopic measurements within laser produced plasmas,
Rev. Sci. /nst., 56: 803.
Hughes, T. P. (1975). Plasmas and Laser Light, Wiley, New York.
Jacobs, V. L. and Blaha, M. (1980). Effects of angular-momentum-changing colli-
sions on dielectric satellite spectra., Phys. Rev. A, 21: 525.
Jahoda, E C. and Sawyer, G. A. (1971). Optical refractivity of plasmas, in Methods of
Experimental Physics, VoL 9: Plasma Physics, Part B, (R. H. Lovenberg and H. R.
Griem, eds.) Academic Press, New York.
Kacenjar, S., Skupsky, S., Entenberg, A., Goldman, L., and Richardson, M. (1982).
Direct measurement of the fuel density-radius product in laser fusion experi-
ments, Phys. Rev. Lett., 49: 463.
Karttunen, S. K. (1985). Spectral analysis of half harmonics emission from laser-
plasma, Laser and Particle Beams, 3: 157.
Kauffman, R. L., Lee, R. W., Matthews, D. L., and Kilkenny, J.D. (1984). X-ray
spectroscopy from variable Z laser produced plasmas, J. Quant. Spectrosc. and
Radiation Transfer, 32: 335.
Kephart, J. E, Dingus, R. S., Gitomer, S., and Kopp, R. (1987). Layered target burn-
through experiments using 50 n.s. KrF laser pulses, Proc. of Intemat. Conf. on
Lasers.
Kepple, P. C. and Whitney, K. G. (1981). U.S. Naval Res. Lab. Rep. 4565, Appendix.
Key, M. H. and Hutcheon, R. J. (1980). Spectroscopy of laser produced plasmas, in
Advanced Atomic and Molecular Physics, VoL 16: p. 201.
Kilkenny, J. D., Hares, J. D., Lewis, C. S., and Rumsby, P. T. (1980a). Search for
Rayleigh-Taylor instability in laser irradiated layered thin foil targets, J. Phys.
D., 13: L123.
Kilkenny, J.D., Lee, R. W., Key, M. H., and Lunney, J. G. (1980b). X-ray spectro-
scopic diagnosis of laser-produced plasmas, with emphasis on line broadening,
Phys. Rev. A, 22: 2746.
Kruer, W. L. (1983). Theory and simulation of laser-plasma coupling, in Laser-

Plasma Interactions 2 (R. A. Cairns, ed.), SUSSP Publications, Edinburgh, Scot-
land, p. 185.
Kruer, W. L. (1986). Laser-driven instabilities in long scale length plasmas, in Laser-
Plasma Interactions 3 (M. B. Hooper, ed. ), SUSSP Publications, Edinburgh, Scot-
land, p. 79.
Landen, 0., Kilkenny, J.D., Scofield, J., and Hauer, A. (1984). Observation of satel-
lites to high n resonance lines with a focussing crystal spectrometer, Bull. Am.
Phys. Soc., 29: 1420.
Lee, R. W. (1978). Plasma line broadening of the Lyman-alpha transition including
ion dynamics, J. Phys. B, 11: L167.
Lee, R. W. (1979). Study of ion dynamic effects on Lyman and Balmer hydrogen
lines, J. Phys. B, 12: 1129, 1144.
Lee, R. W. (1982). The effects of gradients on the diagnostic use of spectral features
from laser compressed plasmas, J. Quant. Spectrosc. and Radiation Transfer, 27:
87.
Lee, C. M. and Hauer, A. (1978). Measurements of compressed core density oflaser-
imploded targets by x-ray continuum-edge shift,AppL Phys. Lett., 33: 692.
Lee, R. W., Whitten, B. L., and Stout, R. E. (1984). Spectra a model for K-shell
spectroscopy, J. Quant. Spectrosc. and Radiation Transfer, 32: 91.
Lin, z., Willi, 0., and Rumsby, P. T. (1981). Evidence for beam self-focusing in the
corona of laser-irradiated spherical targets, J. Phys. D, 14: L35.
Lokke, W. A. and Grasberger, W. H. (1976). X S N Q-U-Anon-LTE emission and
absorption coefficient subroutine, Lawrence Livermore Nat. Lab. Rep. UCRL-
52276.
London, R. L. and Rosen, M. D. (1986). Hydrodynamics of exploding foil x-ray
lasers, Phys. Fluids, 29: 3813.
Lunney, J. C. and Seely, J. E (1982). Inner-shell excitation of argon dielectronic
satellite spectral lines in imploded micro ballon plasmas, J. Phys. B, 15: 121.
Martin, E, Grek, B., and Pepin, H. (1980). Synchronized actively mode-locked short
pulse ruby laser, Rev. Sci. Instrum., 51: 1656.
Matthews, D. L., Kauffman, R. L., Kilkenny, J. D., and Lee, R. W. (1984). Electron
energy distributions using time-resolved free bound spectra from coronal plas-
mas,AppL Phys. Lett., 44:586.
Max, C. E. ( 1982). Physics of the coronal plasma in laser fusion targets, in Interaction
Laser-Plasma, Les Bouches, (R. Balian and J. C. Adam, eds.), North Holland.
McCrory, R. L., Montierth, L., Morse, R. L., and Verdon, C. P. (1981). Nonlinear
evolution of ablation-driven Rayleigh-Taylor instability, Phys. Rev. Lett., 46: 336.
McLean, E. A., Stamper, J. A., Ripin, B. J., Griem, H. R., McMahon, J. M., and
Bodner, S. E. (1977). Harmonic generation in Nd:laser produced plasmas,Appl.
Phys. Lett., 31: 825.
McLean, E. A., Stamper, J. A., Manka, C. K., Griem, H. R., Dromer, D. W., and
Ripin, B. H. (1984). Observation of magnetic fields in laser produced plasma
using the Zeeman effect, Phys. Fluids, 27: 1327.
McWhirter, R. W. P. (1965). Spectral intensities, in Plasma Diagnostic Techniques,
(R. H. Huddlestone and S. L. Leonard, eds.), Academic Press, New York.
Mead, W. C., Campbell, E. M., Kruer, W. L., Turner, R. E., Hatcher, C. W., Bailey,
D. S., Lee, P. H. Y., Foster, J., Tirsell, K. G., Pruett, B., Holmes, N. C., Trainor,
J. T., Stradling, G. L., Lasinski, B. F., Max, C. E., and Ze, F. (1984). Characteristics
of lateral and axial transport in laser irradiations of layered-disk targets at 1.06
and 0.35 micron wavelengths, Phys. Fluids, 72: 1301.
Miyanaga, N., Kato, Y., and Yamanaka, C. (1982). X-ray refraction effect and den-
sity determination of steep-gradient, high-density plasma, Opt. Comm., 44: 48.
More, R. M. (1986). Plasma processes in non-ideal plasmas, in Laser-Plasma Inter-
actions 3 (M. B. Hooper, ed.), SUSSP Publications, Edinburgh, Scotland, p. 157.
Mueller, M. M. (1979). Reconstruction of spherically symmetric objects from slit-
imaged emission: Limitations due to finite slit width, Optics Lett., 4: 351.
Nilsen, J. (1988). Radiative-hydro modeling and atomic databases, AlP Conf. Proc.
No. 168 (A. Hauer and A. L. Merts, eds.), Am. Inst. Phys., New York, p. 51.
Peacock, N.J. (1980). Radiation diagnostics, in Laser-Plasma Interactions 1 (B. A.
Cairns and J. S. Sanderson, eds.), SUSSP Publications, Edinburgh, Scotland, p.
711.
Phillion, D. M., Banner, D. L., Campbell, E. M., Turner, R. E., and Estabrook, K. G.
(1982). Stimulated Raman scattering in large plasmas, Phys. Fluids, 25: 1434.
Pien, G. Richardson, M. C., Goldstone, R. D., Day, R. H., Ameduri, F., and Eden
G. (1986). Computerized 3-GHz multichannel soft x-ray diode spectrometer for
high density plasma diagnosis, NucL Instrnm. and Methods Phys. Res. Sect. B, B18:
101.
Priedhorsky, W., Lier, D. Day, R., and Gerke, D. (1981). Hard x-ray measurements
of 10.6-micron laser-irradiated targets, Phys. Rev. Lett., 47: 1661.
Radziemski, L. J., Loree, T. R., Cremers, D. A., and Hoffman, N. M. (1983). Time-
resolved laser-induced breakdown spectrometry of aerosols, Anal. Chern., 55:
1246.
Raven, A., Willi, 0., and Rums by, P. T. (1977). Megagauss magnetic field profiles in
laser produced plasmas, Phys. Rev. Lett., 41: 554.
Richardson, M. C., Gregory, G. C., Letzring, S. A., Marjoribanks, R. S., Yaakobi, B.,
Henke, B. L., Jaanimagi, P. A., and Hauer, A. (1985). Time resolved x-ray spec-
trographic instrumentation for laser fusion and x-ray laser studies, SPIE Proc.,
569:
Scofield, J. (1984). Private Communication. [See also Landen et al. (1984)].
Seka, W., Williams, E. A., Craxton, R. S., Goldman, L. M., Short, R. W., and Tanaka,
K. (1984). Convective stimulated Raman scattering instability in UV laser plas-
mas, Phys. Fluids, 27: 2181.
Seka, W., Afeyan, B. B., Bani, R., Goldman, L. M., Short, R. W., Tanaka, K., and
Johnston, T. W. (1985). Diagnostic value of odd-integer half-harmonic emission
from laser-produced plasmas, Phys. Fluids, 28: 2570.
Sheffield, J. ( 1975). Plasma Scattering ofElectromagnetic Radiation, Academic Press,
New York.
Shvarts, D. (1985). Thermal electron transport, in Laser-Plasma Interactions 3 (M. B.
Hooper, ed.), SUSSP Publications, Edinburgh, Scotland, p. 341.
Sigel, R. (1980). Diagnostics of particle emission, in Laser-Plasma Interactions 1
(R. A. Cairns and T. S. Sanderson, eds. ), SUSSP Publications, Edinburgh, Scot-
land, p. 661.
Spitzer, L. and Harm, R. (1959). Transport phenomena in a completely ionized gas,

Phys. Rev., 89: 477.
Stamm, R. and Talin, B. (1986). Ion-dynamic effects on the line shapes of hydro genic
emitters in plasmas, Phys. Rev. A, 34: 4144.
Stamper, J. A and Ripin, B. H. (1975). Faraday-rotation measurements of mega-
gauss magnetic fields of laser-produced plasmas, Phys. Rev. Lett., 34: 138.
Stamper, J. A, McLean, E. A, and Ripin, B. H. (1978). Studies of spontaneous
magnetic fields in laser-produced plasmas by Farady rotation, Phys. Rev. Lett.,
40: 1177.
Stamper, J. A, Lehmberg, R. H., Schmitt, A, Herbst, M. J., Young, F. C., Gardner,
J. H., and Obenschain, S. P. (1985). Evidence in the second-harmonic emission
for self-focusing of a laser pulse in a plasma, Phys. Fluids, 28: 2563.
Stradling, G. ( 1982). Time resolved soft x-ray studies, Lawrence Livermore Nat. Lab.
Rep. UCRL-5371.
Sweeney, D. W, Attwood, D. T., and Coleman, L. W (1976). Interferometric probing
of laser produced plasmas,Appl. Optics, 15: 1126.
Tanaka, K., Goldman, L. M., Seka, W, Richardson, M. C., Soures, J. M., and Wil-
liams, E. A (1982). Stimulated Raman scattering from UV-laser-produced plas-
mas, Phys. Rev. Lett., 48: 1179.
Tighe, R. J. and Hooper, C. F. (1976). Stark broadening in hot dense laser-produced
plasmas, Phys. Rev. A, 14: 1514.
Tomov, I. V., Fedosejevs, R., and Richardson, M. C. (1979). Actively mode-locked
and Q-controlled ND: glass laser, Rev. Sci. Instrnm., 50: 9.
Turner, R. E., Phillion, D. W, Lasinski, B. F., and Campbell, E. M. (1984). Half- and
three-halves harmonic measurements from laser-produced plasmas, Phys. Fluids,
27:511.
Turner, R. E., Estabrook, K., Kauffman, R. L., Bach, D. R., Drake, R. P., Phillion,
D. W, Lasinski, B. F., Campbell, E. M., Kruer, W L., Williams, E. A, (1985).
Evidence for collisional damping in high energy Raman-scattering experiments
at 0.26 micron, Phys. Rev. Lett., 54: 189.
Turner, R. E., Estabrook, K., Drake, R. P., Williams, E. A, Kornblum, H. N., Kruer,
W. L., and Campbell, E. M. (1986). Observation of forward Raman scattering in
laser produced plasmas, Phys. Rev. Lett., 57: 1725.
Walsh, C. J., Villeneuve, D. M., and Baldis, H. A (1984). Electron plasma-wave pro-
duction by stimulated Raman scattering: Competition with stimulated Brillouin
scattering, Phys. Rev. Lett., 53: 1445.
Whitlock, R. R., Emery, M. H., Stamper, J. A, Mclean, E. A, Obenschain, S. P.,
and Peckerar, M. C., (1984). Observation of Rayleigh-Taylor-like structures in a
laser-accelerated foil, Phys. Rev. Lett., 52: 819.
Whitney, K. G. and Kepple, P. C. (1982). Spectrum diagnostics. The necessity for de-
tailed non-LTE modeling of x-ray emission from dense plasmas, 1 Quant. Spec-
trosc. and Radiat. Transfer. 27: 281.
Willi, 0. and Rumsby, P. T., (1981). Filamentation on laser irradiated spherical tar-
gets, Opt. Comm., 37:45.
Willi, 0., Rumsby P. T., and Duncan, C. (1981). Megagauss magnetic fields on laser
irradiated spherical targets, Opt. Comm., 37: 40.
Woltz, L.A., Iglesias, C. A., and Hooper, C. F., Jr. (1982). Stark broadening in hot,
dense, laser-produced plasmas: A review, 1Q.S.R. T., 27: 233.
Yaakobi, B. and Nee, A. (1976). Spatially resolved and Stark-broadened x-ray lines
from laser-imploded targets, Phys. Rev. Lett., 36: 1077.
Yaakobi, B., Steel, D., Thorsos, E., Hauer, A., and Perry, B. (1977). Direct measure-
ment of compression of laser-imploded targets using x-ray spectroscopy, Phys.
Rev. Lett., 39: 1526.
Yaakobi, B., Steel, D., Thorsos, E., Hauer, A., Perry, B., Skupsky, S., Geiger, J., Lee,
C. M., Letzring, S., Rizzo, J., Mukaiyama, T., Lazurus, E., Halpern, J., Deckmann,
H., Deletterz, J., Soures, J., and McCrory, R. (1979). Explosive-pusher-type laser
compression experiments with neon-filled microballons, Phys. Rev. A., 19: 1247.
Yaakobi, B., Skupsky, S., McCrory, R. L., Hooper, C. F., Deckman, C. F., Bourke,
H. P., and Soures, J. M. (1980). Symmetric laser compression of argon-filled glass
shells to densities of 4-6 g/cm3 , Phys. Rev. Lett., 44: 1072.
Yaakobi, B., Deletterz, J., Goldman, L. M., McCrory, R., Majoribanks, R., Richard-
son, M. C., Schvarts, D., Skupsky, S., Soures, J. M., Verdon, C., Villeneuve, D.,
Boehly, T., Hutchinson, R., and Letzring, S. (1984). Thermal transport measure-
ments in 1.06 micron laser irradiation of spherical targets, Phys. Fluids, 27: 516.
Yates, M. A., Van Hulsteyn, D. B., Rutkowski, H., Kyrala, G. A., and Brackbill, J.
(1982). Evidence for self-generated magnetic fields and remote energy deposi-
tion in laser-irradiated targets, Phys. Rev. Lett., 49: 1702.
Zimmerman, G. B. and Kruer, W. L. (1975). Numerical simulation of laser-initiated
fusion, Comments Plasma Phys. and Controlled Fusion, 2: 51.
4
Laser-Sustained Plasmas
Dennis A. Keefer
Center for Laser Applications
University of Tennessee Space Institute
Tullahoma, Tennessee
4.1 INTRODUCTION
Plasmas created by the radiation from focused laser beams were first ob-
served with the advent of "giant pulse" Q-switched, ruby lasers by Maker
et al. (1963). These plasmas formed spontaneously by gas breakdown at
the focus of a lens and were sustained only for the duration of the laser
pulse. Plasmas were also observed to form on the surfaces of materials ir-
radiated by high-power pulsed or continuous lasers and to propagate into
the incident beam at subsonic or supersonic velocities. With the advent of
continuous, high-power carbon dioxide lasers, it became possible to sustain
a plasma in a steady-state condition near the focus of a laser beam, and the
first experimental observation of a "continuous optical discharge" was re-
ported by Generalov et al. (1970). The continuous, laser-sustained plasma
(LSP) is often referred to as a continuous optical discharge (COD) and it
has a number of unique properties that make it an interesting candidate for
a variety of applications.
The laser-sustained plasma shares many characteristics with other gas
discharges, as explained in detail by Raizer (1980) in his comprehensive re-
view, but it is sustained through absorption of power from an optical beam
by the process of inverse bremsstrahlung. Since the optical frequency of the
sustaining beam is greater than the plasma frequency, the beam is capable of
propagating well into the interior of the plasma where it is absorbed at high
intensity near the focus. This is in contrast to plasmas sustained by high-
frequency electrical fields (microwave and electrodeless discharges) that
operate at frequencies below the plasma frequency and sustain the plasma
through absorption within a thin layer near the plasma surface. This funda-
mental difference in the power absorption mechanism makes it possible to
169
170 Keefer
generate steady-state plasmas having maximum temperatures of 10,000Kor

more in a small volume near the focus of a lens, far away from any confining
structure. A photo of a plasma sustained by a laser beam focused with a lens
is shown in Fig. 4.1(a). The 600 W Gaussian beam from a carbon dioxide
laser was focused by a 191 mm focal length lens into 2 atm of flowing ar-
gon. Fig. 4.1(b) shows schematically how the plasma forms within the focal
region.
Continuous, laser-sustained plasmas have been produced in a variety of
gases at pressures from 1 to 210 atm using carbon dioxide lasers operating at
a wavelength of 10.6 flm and powers from 25 W to several kilowatts. Most
of these experiments were performed in open air or large chambers with
little flow, except for that provided by natural convection, but recent ex-
periments by Gerasimenko et al. (1983), Welle et al. (1987), and Cross and
Cremers (1986) have demonstrated that the LSP can be operated success-
fully in a forced convective flow. Gas discharges that operate in a flowing
environment have been called "plasmatrons" in the Soviet literature, and
the laser-sustained plasma is often referred to as an "optical plasmatron."
These experiments have demonstrated that plasma conditions are strongly
dependent on the position of the plasma within the sustaining beam, and
that the plasma can be controlled within a wide range of conditions us-
ing appropriate combinations of laser power, flow, and optical configura-
tion.
The unique ability to sustain a plasma within a small, isolated volume
at relatively high pressures and temperatures has suggested a number of
potential applications for the laser-sustained plasma. Since the LSP can
operate in pure hydrogen and the power can be beamed remotely, it has
been proposed that the LSP could be used for high specific-impulse space
propulsion. A number of papers have dealt with this application, and it was
the subject of a review by Glumb and Krier (1984). Thompson et al. (1978)
described experiments in which laser energy was converted into electrical
energy using a laser-sustained argon plasma. Cremers et al. (1985) have
suggested the LSP as a source for spectrochemical analysis and given some
experimental results. Cross and Cremers (1986) have sustained plasmas in
the throat of a small nozzle to produce atomic oxygen having a directed
velocity of several km/sec for the laboratory study of surface interactions at
energies and particle fluxes similar to those experienced by satellites in low-
earth orbit. Other applications are suggested by analogy to other plasma
devices including light sources, plasma chemistry, and materials processing.
The physical processes that determine the unique characteristics of the
LSP will be discussed in Sec. 4.2, and the theoretical analyses that have been
used to describe the LSP will be addressed in Sec. 4.3. Experimental results
obtained will be presented in Sec. 4.4 and compared with the theoretical
predictions. Sec. 4.5 will consider some possible applications.
Laser-Sustained Plasmas 171
(a)
(b)
Figure 4.1 (a) Photograph of a plasma sustained by a 600 W carbon dioxide laser
beam focused with a 191 mm focal length lens. (b) Schematic representation show-
ing how the plasma forms within the focal volume.
172 Keefer
4.2 PRINCIPLES OF OPERATION
Plasmas that are created or sustained by lasers can be generated in a variety

of forms, depending on the characteristics of the laser and optical geome-
try used to generate them. High-energy pulsed lasers can generate plasma
breakdown directly within a gas that results in a transient expanding plasma
similar to an explosion. At lower laser intensities and longer pulse times,
plasmas may be initiated at solid surfaces and then propagate into the sus-
taining beam at supersonic velocities as a laser-sustained detonation (LSD)
wave or subsonic velocities as a laser-sustained combustion (LSC) wave.
These transient plasmas have been discussed by Raizer (1980) and will not
be treated here. If the laser is continuous in output power and the optical
geometry, flow, and pressure are favorable, then a steady-state LSP may
be continuously maintained at a position near the focus of the beam. The
intensity that is available from a continuous laser is insufficient to cause
breakdown in the gas, however, and an auxiliary source must be used to ini-
tiate the plasma. A sketch of a steady-state laser-sustained plasma is shown
in Fig. 4.1(b). The plasma may be sustained within a confining chamber to
control the flow and pressure or in open air or a large chamber where the
flow is determined by thermal buoyancy.
In many ways, the laser-sustained plasma is similar to direct current or
low-frequency electrodeless arcs and microwave discharges that are oper-
ated in similar gases and at similar pressures. However, the LSP will gener-
ally be more compact and have a higher maximum temperature than other
continuous arc sources and can be sustained in a steady state well away from
containing boundaries. A fundamental difference in the way in which en-
ergy is absorbed by the plasma is responsible for these unique characteristics
of the LSP.
4.2.1 Basic Physical Processes
In a direct current (de) arc or in an inductively coupled plasma (ICP), en-

ergy is absorbed through ohmic heating produced by the low-frequency or
direct currents flowing in the plasma. The electrical conductivity of an ideal
plasma is given by (Shkarofsky et al., 1966)
_ ne2
(J--
( v-iw) (4.1)
m v 2 +w2
where n is the electron density, e the electronic charge, m the electron mass,
w the radian frequency of the applied electric field, v the effective collision
frequency for electrons, and i the square root of -1. In the de arc (w = 0),
the currents are transmitted through the plasma between electrodes and
the size of the plasma is determined by the size and spacing of the electrode
and the confining boundaries. In the ICP, the currents are induced into
the plasma from alternating currents flowing in a surrounding solenoidal
coil. The arc is sustained within a container that determines the plasma
diameter, whereas the length of the plasma is determined by the length of
the solenoid.
The ICP operates at frequencies well below the plasma frequency
w
P
= (nez )1/2
mE 0
(4.2)
where Eo is the permittivity of free space. In this frequency range, the elec-
tromagnetic field does not propagate as a wave within the plasma, but is
attenuated as an evanescent wave (Holt and Haskell, 1965) over distances
of the order of the skin depth
( 4.3)
where c is the speed of light. Thus, the plasma is sustained by energy ab-
sorbed within a small layer near its outer surface that produces a rather fiat
temperature profile within the plasma and limits the maximum tempera-
tures that can be obtained.
The frequency of the optical fields (28 THz for the 10.6 ;.tm carbon diox-
ide laser) used for the LSP is greater than the plasma frequency, and there-
fore the incident laser beam can propagate well into the interior before
it is significantly absorbed through the process of inverse bremsstrahlung
(Shkarofsky et al., 1966). Since the focusing of the laser beam produced
by a lens or mirror is essentially preserved as the beam propagates into the
plasma, very large field strengths may be produced within the plasma near
the beam focus. It is these large field strengths that lead to peak tempera-
tures in the LSP that are generally greater than those obtained with either
de arcs or the ICP and make it possible to sustain a small volume of plasma
near the focus, well away from any confining walls.
Inverse bremsstrahlung is a process in which the plasma electrons ab-
sorb photons from the laser beam during inelastic collisions with ions, neu-
trals, and other electrons. The collisions between electrons and ions are
the dominant process for the LSP and the absorption coefficient is given by
(Shkarofsky et al., 1966)
a= (7rc)2nSoG (1-e-hw/kT) (4.4)

w kT hw/kT
174 Keefer
where li is Planck's constant divided by 21r, k Boltzmann's constant, and T

the temperature of the electrons. The factor G is the Gaunt factor and the
factor nS0 is given by
(4.5)
where Z is the ionic charge and n+ the ion density. The Gaunt factor is a
quantum mechanical correction to the classical theory, and extensive tables
have been given by Karzas and Latter (1961). For the usual case where
the photon energy is much less than the thermal energy (liw ~ kT), the
bracketed term in Eq. ( 4.4) is nearly independent of w, and the absorption
coefficient is essentially proportional to the square of the wavelength.
The size of the LSP will depend on several factors including the beam
geometry, laser power, and absorption coefficient. The change in intensity
of the laser beam as it propagates within the plasma is given by Beer's law
dl
-=-a/ (4.6)
ds
where s is the distance along the local direction of propagation. The absorp-
tion length 1/a is a dominant length scale for the LSP since it determines
the distance over which the power is absorbed from the beam. For this rea-
son, the dimension of the high-temperature absorbing portion of the plasma
along the laser beam will be of the order of the absorption length. Although
it is the absorption length that determines the length of the plasma along the
beam axis, it is the laser beam diameter that determines the plasma diame-
ter. The plasma expands to fill the beam cone where it is able to absorb
power, then rapidly decreases in temperature outside the beam through
thermal conduction and radiative loss mechanisms.
The position of the LSP relative to the focal point is critical in determin-
ing its structure and the range of parameters for which it can be maintained.
When the plasma is initiated near the beam focus, it propagates into the
sustaining beam and seeks a stable position. The position of stability will be
located where the beam intensity is just sufficient that the absorbed power
will balance the losses due to convection, thermal conduction, and thermal
radiation. A number of factors combine to determine this position of sta-
bility including the transverse profile of the incident beam, the focal length
and aberrations of the focusing lens or mirror, the plasma pressure, and the
incident flow velocity (Keefer et al., 1986; Welle et al., 1987).
The power per unit volume that is absorbed by the plasma is given by
P=al (4.7)
where I is the local irradiance of the laser beam. Since I depends on the
transverse profile of the incident beam as well as the focal length and aber-
rations of the lens, these characteristics will influence the location within
the focal region at which the minimum sustaining intensity is located. For
example, for a small f/number lens, the intensity decreases rapidly with in-
creasing distance from the focus and the plasma will stabilize near the focus.
For a larger f/number system, the intensity decreases less rapidly and the
plasma will stabilize at a position further away from the focus. Indeed, for
sufficiently long focal lengths and high laser power, plasmas have been ob-
served to propagate many meters (Razier, 1980) as "laser-suppo_rted com-
bustion waves" at subsonic velocities.
The detailed spatial structure of the plasma is determined by the interre-
lations between the optical geometry of the sustaining beam, the pressure of
the gas, and the flow through the plasma. At each point within the plasma,
the temperature and flow must adjust to balance the power that is absorbed
from the laser beam with the power lost through convection, conduction,
and thermal radiation. The position in the beam relative to the focal point
at which the plasma stabilizes is very important in determining the structure
of the plasma that, in turn, determines the conditions of power, pressure,
and flow for which a stable plasma can be maintained.
Most of the early experiments with the LSP were carried out inside large
chambers or in open air, where the flow through the plasma was determined
by the effects of thermal buoyancy. Fixed focal geometries were used and
the pressure and laser power were varied to define regions of power and
pressure where it was possible to sustain the LSP in a variety of gases (Gen-
eralov et al., 1972; Kozlov et al., 1979; Moody, 1975). These experiments
indicated that there were upper and lower limits for both laser power and
pressure at which the LSP could be sustained.
Generalov et al. (1972) suggested that the upper limit for power was are-
sult of forming the LSP with a horizontal beam. In this geometry, thermal
buoyancy induces a flow transverse to the optical axis. The induced flow
carries the plasma up and out of the beam when higher laser power causes
the plasma to stabilize farther from the focus. They were unable to estab-
lish an upper power limit when the experiment was operated with the beam
propagating vertically upward. Kozlov et al. (1974) developed a radiative
model for the LSP and explained the upper power limit on the basis that
the plasma must stabilize close enough to the focal point that the geomet-
ric increase of laser beam intensity going into the plasma was greater than
the loss of intensity due to absorption. They speculated that the failure of
Generalov et al. (1972) to observe this limit in a vertical beam was due to
rapid extinction and reignition of the plasma.
It is clear from the experiments of Generalov et al. (1972) that flow can
have a large effect on the range of pressure and laser power that will support
176 Keefer
a stable LSP. Plasmas sustained in the free jet issuing from a nozzle have
been studied by Gerasimenko et al. (1983) who measured the discharge
wave velocity along the beam and ranges for the existence of a steady-state
discharge. Recently, experiments have been conducted in confined tubes
where forced convection dominated the flow (Welle et al., 1987). It was
found that in addition to power and pressure, both the flow and optical ge-
ometry of the beam have a profound influence on the characteristics of the
plasma. It now appears that the earlier experimental results that defined re-
gions of pressure and power for which the LSP could be sustained are valid
only for the particular experimental geometry used to obtain them.
When the plasma is ignited by an auxiliary source near the focal point,
the plasma expands as it absorbs energy from the continuous laser beam and
propagates into the beam with decreasing velocity until it stabilizes. The
plasma becomes stationary at a point where the intensity is just sufficient
that the power absorbed from the beam, given by Eq. (4.7), is balanced by
the convective, conductive, and radiation losses. Since, in general, the in-
tensity is not uniform across the beam, the plasma will adjust in size, shape,
temperature, and flow to satisfy conservation of momentum and energy. If
the flow through the plasma increases, then the thermal convection losses
increase in the upstream region of the plasma and it must move toward the
focus to a region of higher intensity in order to absorb enough power from
the beam to compensate for the increased convective losses.
Thermal radiation plays a significant role in determining the detailed
structure of the plasma. Thermal radiation in the plasma occurs both as
a result of bound-bound transitions, resulting in line radiation and absorp-
tion, and free-bound and free-free transitions that result in continuum radi-
ation and absorption. Over the optically thin portion of the spectrum, this
radiation will not be strongly absorbed by the plasma or surrounding cooler
regions and will simply escape from the plasma. Other portions of the spec-
trum will be strongly absorbed, resulting in a transport of energy within the
plasma. In the optically thick limit, this results in a diffusive energy trans-
port that is similar to thermal conduction, but may be significantly larger.
Detailed calculations of the LSP (Jeng and Keefer, 1986) indicate that this
radiative transport is a dominant factor in the determination of the struc-
ture and position of the LSP. In particular, it is the radiative transport that
determines the temperature gradient in the upstream front of the plasma,
thereby determining the position in the beam for which convection losses
are balanced by absorption.
The position of stability for the LSP also depends on the plasma pres-
sure. The absorption coefficient is a strong function of plasma density, as
seen from Eq. ( 4.4 ). If the pressure is increased and the absorption coeffi-
cient increases, then the plasma can absorb more power from the beam and
will move away from the focus to a lower intensity region in the beam. At the
same time, the plasma length along the beam decreases because of the de-
crease in absorption length, but the diameter increases to fill the larger cross
section of the beam. Thus, for the same laser beam conditions, a higher-
pressure LSP will stabilize at a point farther away from the focal point and
have a smaller length-to-diameter ratio than a lower-pressure LSP.
Incident laser power, as well as the f/number and aberrations of the fo-
cusing optics, will also influence the position at which the LSP stabilizes
within the beam. From the foregoing discussion, it is clear that as the beam
power is increased, the plasma will move up the beam away from the focal
point. The distance that it moves is determined by the f/number (ratio of
focal length to the beam diameter incident on the focusing element) of the
optical system, since the rate of change in beam intensity along the optical
axis decreases with an increase in f/number. Lens aberrations can also have
an effect on plasma position (Keefer et al., 1986). In particular, when an an-
nular beam from an unstable laser oscillator is focused by a spherical lens,
it produces an annular prefocus region before reaching the focal point, and
the intensity in this region may be sufficient to sustain an annular plasma.
From the observations discussed above, it is clear that the position of
the plasma relative to the focal point has a profound effect on the plasma
characteristics. At the upper limits of stability for both laser power and
pressure, it appears that the plasma becomes unstable when it moves too
far from the focal point. This may be due to the fact, as proposed by Kozlov
et al. (1974), that as the plasma moves sufficiently far away from the focus,
the rate of increase of the beam intensity in the direction of propagation
becomes smaller. Since the temperature of the plasma must increase as the
beam propagates into the upstream edge of the plasma, the intensity of the
beam must also increase. At some point, the decrease of the beam intensity
due to absorption is greater than the increase due to focusing, so the plasma
becomes unstable and extinguishes. Recent calculations by Jeng and Keefer
(1987a), however, indicate that there may exist local regions within the LSP
where the beam intensity decreases as it penetrates the plasma.
A considerable degree of control of the structure and position of the LSP
can be gained through both optical geometry and flow, in addition to laser
power and pressure. Utilization of these additional parameters makes it
possible to successfully operate the LSP over a wider range of experimental
conditions, enabling a wider range of potential applications.
4.2.2 Plasma Characteristics
Laser-sustained plasmas have been operated in a variety of molecular and

rare gases at pressures from 1 to more than 200 atm. The resulting plasmas
have characteristics that are similar to arc plasmas operated at similar pres-
178 Keefer
sures, but the peak temperatures in the LSP are usually somewhat higher
than those for the comparable arc. Radiation from the plasma can be a sig-
nificant fraction of the total power input, and radiation transport plays a
major role in determining the structure of the plasma. Continuum absorp-
tion processes are of particular importance in these plasmas since the power
to sustain the plasma is absorbed through these mechanisms.
The continuum absorption process involves both bound-free transitions
(photoionization) and free-free transitions (inverse bremsstrahlung) in
which photons are absorbed from the laser beam. The free-free transitions
involve electron collisions with ions, other electrons, and neutral particles
(Shkarofsky et al., 1966; Griem, 1964). The dominant absorption process
for the LSP is through collisions between electrons and ions, and the absorp-
tion coefficient for this process is given by Eq. ( 4.4). For the usual case in
the LSP, nw ~ kT and the absorption is approximately proportional to the
square of the laser wavelength. Due to this strong wavelength dependence,
all of the reported experimental results for the LSP have been obtained us-
ing the 10.6 J.lm wavelength carbon dioxide laser. Since the length scale
for the plasma is of the order of the absorption length, the length of the
plasma and the power required to sustain it would be expected to increase
dramatically for shorter wavelength lasers. Currently, the only other lasers
that are likely candidates to sustain continuous plasmas are the hydrogen
or deuterium fluoride chemical lasers that operate at wavelengths of 3 to
4J.lm.
Thermal radiation is one of the most important characteristics of the
LSP. Thermal radiation lost from the plasma can account for nearly all
the power absorbed by the plasma when the flow through the plasma is
small and will account for a significant fraction of absorbed power even
when the convective losses are large. The thermal radiation consists of
continuum radiation resulting from recombination (free-bound transitions)
and bremsstrahlung (free-free transitions) as well as line radiation (bound-
bound transitions). Calculation of this radiation is straightforward, al-
though rather tedious, when the plasma is in local thermodynamic equi-
librium (LTE) (Griem, 1964). Local thermodynamic equilibrium is es-
tablished when the electron collisional rate processes dominate the pro-
cesses of radiative decay and recombination. When LTE is established
in the plasma, the density in specific quantum states is the same as a sys-
tem in complete thermal equilibrium having the same total density, tem-
perature, and chemical composition. It should be emphasized that this
does not imply that the radiation is similar to a blackbody at the plasma
temperature. In general, the spectrum of the radiation from the plasma
will have a complex structure consisting of the superposition of relatively
narrow spectral lines and a continuum having a complex spectral struc-
ture.
The absorption coefficient in the plasma depends on the wavelength, and

for the ultraviolet portion of the spectrum below the wavelength of the reso-
nance lines (transitions involving the ground state), the radiation is strongly
absorbed by the plasma and the cooler surrounding gas. This results in a
strong radiative transport mechanism that is important in determining the
structure of the plasma. Often, radiative transport for strongly absorbing
gases is modeled as a diffusive energy transport similar to thermal conduc-
tion. In the strongly ionized regions of the plasma, the radiative transport is
many times larger than the intrinsic thermal conduction and is the dominant
heat-transfer mechanism. This is especially true in the upstream region of
the LSP where the temperature gradient is large, and radiation transport
must offset the convective losses of the incident flow.
In the longer wavelength region above the resonance transitions, the ab-
sorption of the radiation by the plasma and the surrounding gas is much
smaller. The absorption length for this radiation is often large compared
to the characteristic dimensions of the plasma, and much of the radiation
escapes. In this region of the spectrum, the plasma may be considered op-
tically thin, and if the plasma is in LTE, then the escaping radiation can be
used to characterize the temperature within the LSP (Keefer et al., 1986;
Welle et al., 1987).
The temperature within the plasma is far from uniform, as shown in
Fig. 4.2. (The method used to obtain the experimental temperatures shown
in Figs. 4.2, 4.4, and 4.10 is described in detail in Sec. 4.4.2). This figure
shows an isotherm plot of the temperatures measured in an LSP sustained
in 2.5 atm of argon by a carbon dioxide laser operating at a wavelength of
10.6 J-tm. The plasma length and diameter, as determined by the 10,500K
isotherm, are 11 and 4 mm, respectively. Note the steep temperature gradi-
ents that exist in the upstream portion of the plasma and in the radial direc-
tion near the limit of the laser beam. The temperature gradients in these
regions are of the order of 106 K/m. In the upstream regions of the plasma,
the direction of the convective energy transport is the opposite of that due to
thermal conduction and radiation transport, and a strong temperature gra-
dient develops to balance the convective losses with thermal conduction and
radiative transport. The magnitude of the temperature gradient depends on
the flow rate and increases with increasing flow. Strong radial temperature
gradients develop near the edge of the laser beam where the available power
decreases rapidly and are steeper near the focus where large conduction and
radiation transport is required to balance the large powers absorbed from
the beam. The peak temperature in the plasma occurs near focus where
the laser beam intensity is maximum, and the peak temperature has a value
that corresponds closely to the temperature at which the first stage of ioniza-
tion is nearly complete. The correspondence of the maximum temperature
with complete first ionization is to be expected since, for a constant pres-
180 Keefer
---lens
Ill
0
c:
..
....0
1J
0
1J
0
1..
laser power = 728 W
axial distance (nun)
Figure 4.2 Measured isotherms for a 2.5 atm argon plasma sustained by an annular
beam from a carbon dioxide laser focused by a lens of 203 mm focal length. The
horizontal scale was arbitrarily chosen so the focus occurs at 50 mm, and both the
laser beam and the flow are incident from the left. The contour interval is 500K with
the outer contour at 10,500K. (From Welle, R., eta!., 1987. Copyright© American
Institute of Aeronautics and Astronautics; reprinted with permission.).
sure plasma, the absorption coefficient passes through a maximum at that

temperature.
Plasmas have been sustained by carbon dioxide lasers in a variety of gases
including the rare gases xenon, argon, and neon and in several molecu-
lar gases including hydrogen, deuterium, nitrogen, carbon dioxide, and air.
Plasmas in the heavier rare gases (xenon, krypton, and argon) are the easi-
est to sustain since they have relatively low ionization potential, their ther-
mal conduction is relatively small, and no dissociation energy must be sup-
plied. Few experiments have been reported using the lighter rare gases,
but Carloff et al. (1981) have sustained optical discharges in helium and
neon to pressures of 207 atm and Cross (1986) has operated a discharge in
neon with admixtures of oxygen. Relatively higher laser powers were re-
quired with helium to offset the greater ionization potential and relatively
large thermal conductivity. The molecular gases also require greater laser
power since these gases must absorb sufficient energy to undergo dissocia-
tion before they can be ionized. As discussed earlier, the threshold power
required to maintain a plasma depends on the details of the experiment in-
cluding beam quality, optical geometry and aberrations, and the flow rate.
For example, Generalov et al. (1972) found the threshold power for argon
at a pressure of 10 atm to be approximately 90 W, whereas Moody (1975)
found the threshold to be approximately 30 W at the same conditions. For a
single experimental configuration, Kozlov et al. (1979) found the threshold
increased going from argon to nitrogen to air to deuterium to hydrogen.
4.2.3 Practical Implementation
Application of laser-sustained plasmas has not been widespread, and most

of the operational systems have been constructed in research laboratories to
study the plasma itself. The LSP has been used in an experimental attempt
to measure the high-temperature absorption of mixtures of water vapor and
hydrogen (Fowler, 1981). Cross (1986) has used the LSP to generate very
high-speed molecular oxygen beams for the study of satellite surface inter-
actions, and Cremers et al. (1985) have used the LSP as a source for atomic
spectroscopy. Recently, detailed laboratory studies of the LSP have been
conducted by Keefer eta!. (1986), Welle eta!. (1987), and Krier eta!. (1986)
to determine their suitability for laser thermal propulsion.
The experiments with continuous laser-sustained plasmas have been per-
formed using focused beams from continuous carbon dioxide lasers with
output powers between a few tens of watts and several tens of kilowatts.
The simplest way to create the LSP is to focus the beam from the laser into
ambient air using a lens or mirror. If the power is approximately 5 kW or
greater, a plasma can be initiated by a discharge near the focal point that
is created by an arc, a pulsed laser plasma breakdown, or by momentarily
placing a metallic (usually tungsten) surface in the focal volume. The sta-
bility of the LSP is enhanced by operating the beam so that it is propagat-
ing vertically upward toward the focus. This insures that the flow induced
by thermal buoyancy is in the same direction as the beam propagation and
does not tend to convect the hot plasma out of the high-intensity focal re-
gion. Various focal length lenses have been used to focus the beam, with
typical systems having focal lengths from 25 to 300 mm and f/numbers from
1.5 to 5. Plasmas have been sustained using beams of various mode struc-
tures: annular beams produced by unstable oscillators, multimode beams
from stable oscillators, and Gaussian beams. There has been little system-
182 Keefer
atic study of the influence of beam quality on the LSP, but in general the
threshold power required to sustain the plasma decreases with increasing
beam quality. Moody (1975) has sustained plasmas in argon at pressures
greater than 10 atm using only 25 W from a laser having a Gaussian beam
that was focused by a mirror of 52 mm focal length. Various investigators
have reported threshold values that vary by a factor of 10 or more, and many
of these differences are probably due to the variety of beam quality and fo-
cusing geometries that have been employed.
To operate the LSP in gases other than air and at pressures greater than
1 atm, it is necessary to confine the plasma within a chamber. Often, these
have been chambers that were large compared to the plasma, and the flow
in the plasma has been established by the thermally induced, free convective
flow within the chamber. Recent experiments have been conducted (Krier
et al., 1986; Welle et al., 1987) in which argon plasmas were sustained in a
forced convective flow. The forced convective flow velocity was consider-
ably larger than the thermally induced flow velocities, and it was found that
the plasmas could be sustained over a wide range of power and pressure by
adjusting the flow velocity and optical geometry.
4.3 ANALYTICAL MODELS
Soon after laser-induced breakdown was discovered, theoretical models

for breakdown of a gas in an optical field were developed by Zel'dovich
and Raizer (1965). Raizer (1965) then developed a model for the super-
sonic propagation of the absorption wave, and later (Raizer, 1970) devel-
oped a one-dimensional model for the subsonic propagation and mainte-
nance of a laser-sustained plasma. This model was based on an analogy
with combustion, and these subsonic propagating waves, together with the
continuous laser-sustained plasmas, became known in the United States as
"laser-supported combustion waves (LSC)." The closed-form solution for
this model was widely used to interpret experimental observations of prop-
agating plasmas, but was found to greatly underestimate the observed prop-
agation velocity.
The Raizer model was extended by Jackson and Nielsen (1974) to in-
clude the effect of radiative transfer, and they obtained a numerical so-
lution that led to much better agreement between the predicted propaga-
tion velocities and experimental observations. Kozlov and Selezneva (1978)
developed a one-dimensional, time-dependent model and studied numer-
ically the transition of the plasma from the initiation spark to a station-
ary discharge. Kemp and Root (1979) extended the Jackson and Nielsen
(1974) model to predict the characteristics of a hydrogen LSP for laser
propulsion. The numerical methods that were used by Jackson and Nielsen
(1974) and Kemp and Root (1979) led to a divergence of the solution
near the maximum temperature, and Keefer et al. (1985) applied a new
method of numerical solution that provided converged solutions through-
out the solution domain. It was clear from all of the numerical models
that radiative transfer in the forward portion of the propagating plasma
played a dominant role in the determination of the propagation veloc-
ity.
Batteh and Keefer (1974) developed a quasi-two-dimensional model for
the LSP that was based on an extension of the Raizer analysis. They in-
cluded the effect of radial and axial thermal conduction and finite laser
beam diameter, but assumed that the laser beam was collimated and radial
components of the flow could be neglected. The closed-form solutions for
the temperature that were obtained agreed qualitatively with observed plas-
mas. This model was extended by Muller and Uhlenbusch (1982) to include
the important effect of the convergence of the laser beam geometry. They
used a focused Gaussian beam to model the two-dimensional, axisymmet-
ric distribution of laser intensity in the focal region. Approximate solutions
were obtained for the case in which the influence of the containing walls was
neglected and restricted functional forms of the transport properties were
used.
A more elaborate model of the flow in the vicinity of the LSP was de-
veloped by Carloff et al. (1984). They obtained solutions of the Navier-
Stokes equations for the thermally induced buoyant flow within a closed
chamber containing the plasma. The solution was obtained for the flow
within the cooler gas region outside of the high-temperature plasma core
by solving the momentum equations using a temperature field that was cal-
culated separately. The temperature field was obtained from a solution of
the energy equation using an assumed Gaussian beam profile and a veloc-
ity field obtained from the momentum equations using an assumed tem-
perature profile. The equations were not coupled, and it was not clear
whether the radiation and transport properties had an assumed spatial vari-
ation or whether the temperature- and pressure-dependent properties were
used. Their solutions clearly showed the recirculation cells that develop in
the closed chamber and the deflection of the streamlines around the high-
temperature plasma core.
The next step in the development of two-dimensional models was taken
by Glumb and Krier (1985) who included the effect of focusing by assum-
ing a uniform convergent beam focused to an arbitrary spot size and re-
tained the assumption that radial components of the flow could be ne-
glected. However, they obtained numerical solutions using fully coupled
temperature- and pressure-dependent transport properties. This model
produced predicted temperature fields that were similar to those observed
and gave good results for global absorption of laser power.
184 Keefer
Merkle (1984) developed a theoretical model to consider the case in

which the plasma was sustained in a chamber having a converging-diverging
nozzle. He solved the quasi-one-dimensional Navier-Stokes equations nu-
merically using an implicit time-marching method for a plasma mixture of
hydrogen, cesium, and water. Beam convergence was included in an ap-
proximate fashion but radiation transfer was neglected. A model that con-
tained a complete two-dimensional description of the flowfield was devel-
oped by Molvik et al. (1985), but the flowfield was uncoupled from the radia-
tion by the assumption of a constant absorption coefficient. This assumption
was relaxed by Merkle et al. (1985) by assuming absorption into a mixture of
hydrogen, cesium, and water that exhibits strong absorption at relatively low
temperatures. They were unable to obtain convergence for pure gases be-
cause of the strong gradients and nonlinear temperature-dependent prop-
erties.
Jeng and Keefer (1986) developed a full two-dimensional model based
on a Navier-Stokes description of the flowfield, geometric ray-tracing for
laser beams of arbitrary radial profiles, and a radiative transfer model that
divided the radiative transport into optically thick and optically thin regions.
This model was solved numerically using the SIMPLE (Jeng and Keefer,
1986) algorithm for pure argon in a pipe flow where the plasma was sus-
tained by the annular beam from an unstable resonator. These solutions
were found to be in good agreement with detailed measurements, and the
model was extended to obtain solutions for plasmas sustained in hydrogen
by Gaussian laser beams (Jeng et al., 1987). Comparison of the solutions
obtained using the full two-dimensional model with those obtained from the
quasi-two-dimensional model, in which the radial components of the flow
are neglected, clearly show that the full two-dimensional treatment of the
flowfield is required for accurate predictions. Due to the pressure increase
caused by the strong heat addition, the flow is diverted around the outside of
the plasma core and the mass flow through the plasma core may be reduced
by a factor of 5 or more compared with the simpler models (Jeng et al.,
1987). This effect is shown in Fig. 4.3 where the temperature isotherms are
shown in the upper half of the figure, and the local mass flux is represented
by the vectors in the lower half. This calculation was for a 3 atm hydrogen
plasma sustained by a 20 kW, 40 mm diameter Gaussian beam from a carbon
dioxide laser operating at 10.6 pm.
Development of the detailed two-dimensional models has made it possi-
ble to gain considerable insight into the factors that control the LSP. One
factor that was found to be of considerable importance is the characteristics
of the laser beam and the focusing optics. It was found, both experimentally
and from the model, that when an annular beam from an unstable oscilla-
tor was used to sustain the plasma, spherical aberrations caused an annular
prefocus ahead of the focal point, and that this high-intensity region could
laser power = 20 kW
laser beam dia. = 40 mm
velocity = 20 m/s
pressure = 3 atm
10.0 80.0 90.0

axial distance (mm)
Figure 4.3 Theoretical calculation for a hydrogen plasma sustained by a 40 mm di-

ameter Gaussian beam focused by a 203 mm focal length lens to the 50 mm position
on the horizontal axis. The incident flow velocity is 20 m/sec in a cylindrical pipe
22 mm in diameter. The upper part of the figure contains isotherms and the lower
half contains vectors whose length and direction represent the local mass flux (prod-
uct of local density and local velocity). (From Jeng, S. M. and Keefer, D., 1987a.
Copyright © American Institute of Aeronautics and Astronautics; reprinted with
permission.).
sustain a plasma that had significant off-axis temperature maxima. An ex-

ample of this effect is shown in Fig. 4.4. The upper half of the figure shows
the measured isotherms and the lower half shows the isotherms predicted
by the model. It is seen from the figure that the plasma is sustained further
from the focus than that shown in Fig. 4.2, and that off-axis temperature
maxima occur near the annular prefocus indicated by the crossing of the
limiting rays of the annular beam, indicated by the dashed lines.
The models have also shown that the f/number of the focusing lens has
an important effect on the fraction of laser power that is absorbed by the
186 Keefer
laser power= 735 W THEORY

velocity=0.580 rn/s
Ill
0
1:
....0fl
,
,0
0
1..
EXPERIMENT
(Welle, Keefer,
Peters 1987)
60.0
axial distance (rnrn)
Figure 4.4 Comparison between measured and calculated isotherms for an argon
plasma sustained with an annular beam. Both the laser beam and the flow are in-
cident from the left. Off-axis maxima in the temperature appear where the spher-
ical aberration of the lens causes a high-intensity annular prefocus. (From Jeng,
S. M., et a!., 1987. Copyright © American Institute of Aeronautics and Astronau-
tics; reprinted with permission.).
plasma and the fraction of absorbed power that is lost from the optically
thin plasma through radiation. As the f/number of the optical system is in-
creased, the diameter of the plasma will be smaller if there is sufficient inci-
dent flow velocity to sustain the plasma near the focal region. This smaller
volume plasma will lose less power by thermal radiation and can absorb
more power from the beam because it is located in a region of higher laser
intensity. The models predict that in hydrogen plasmas, it should be pos-
sible to sustain a plasma that absorbs virtually 100% of the incident laser
beam power and that radiates no more than 35% of that power out of the
plasma (Jeng et al., 1987).
Since all of the reported experimental data is for carbon dioxide lasers
operating at a wavelength of 10.6 pm, the detailed models provide an op-
portunity to investigate the effects of using different wavelengths to sus-
tain the plasma. An example is shown in Fig. 4.5 (Jeng, 1986). The upper
isotherms are for a plasma sustained using a 10.6 pm wavelength and the
lower isotherms are for a plasma sustained with a wavelength of 3.9 pm
that is typical of chemical lasers. The incident power of 5245 W for the
3.9 pm wavelength has been scaled up from the incident power of 710 W
used to sustain the plasma at 10.6 pm by the square of the ratio of wave-
lengths, to compensate for a similar decrease in the absorption coefficient.
Note that for the shorter wavelength, the maximum temperature is consid-
erably higher, and the plasma shape and size have changed. As discussed in
Sec. 4.2.1, the plasma diameter is about the same for the two wavelengths,
10.6 J.Lm Argon

710W 2 atm.
0.143 m/s
203 mm f.l.
ID
0
c:
...0
"
"0
0
"0
0
L
3.9J.Lm
5245W
40.0
Figure 4.5 A comparison of calculated argon plasmas sustained by lasers of

10.6 J.Lm wavelength and 3.9 11m wavelength. The incident power for the 3.9 11m case
was scaled upwards by the square of the ratio of the wavelengths to compensate for
a similar reduction in the absorption coefficient.
188 Keefer
-
~
~
90
... - absorption
•
....
CD -radiation
31:
0
Q.
....
CD 70
(I)
.!!! laser power = 20 kW
0 beam dia. = 40mm
c
0
pressure = 3 atm
;:
u
.::
CD 50
25 50 75 100
inlet velocity (m/s)

Figure 4.6 Calculated fractions of incident laser power that are absorbed and rera-
diated from a hydrogen plasma at a pressure of 3 atm. The plasma was sustained
by a 20 kW Gaussian beam having a wavelength of 10.6 f-lm. (From Jeng, S. M. and
Keefer, D., 1987a. Copyright© American Institute of Aeronautics and Astronau-
tics; reprinted with permission.).
since this is primarily a function of the beam diameter, but the length of
the plasma has increased for the plasma sustained with the 3.9 f-Ill beam
due to the larger absorption length. A significant amount of laser power is
absorbed beyond the focal point in this plasma that leads to a "dogbane"
shape for the isotherms. The predicted fraction of incident power absorbed
at 3.9 pm was 28%; considerably less than the prediction of75% for 10.6 pm.
The models also indicate that the plasma can be sustained over a wide
range of flow velocities and that it is difficult to "blow out." The flow ve-
locity has a profound influence on the shape of the plasma and the posi-
tion of the plasma within the beam. At low incident flow velocities, the
plasma propagates up the beam becoming larger in diameter and volume
and, therefore, more power is radiated out of the plasma region. When the
incident flow velocity is increased, the plasma moves toward the focus, be-
coming smaller in diameter and somewhat longer in length. The result of
this change in plasma shape is a reduction in the power lost through radia-
tion and an increase in power absorbed from the laser beam. This effect is
shown in Fig. 4.6 where the calculated fraction of the power absorbed from
the incident laser beam and the fraction of the incident power that is lost by
radiation are shown as a function of the incident flow velocity. It is found
that the fractional power absorption increases with incident flow velocity
and, for these conditions, nearly all the incident power is absorbed. The ab-
sorption increases with velocity because the plasma is longer in the higher-
velocity cases, and this increases the absorption pathlength. The fractional
power that is lost from the plasma decreases as the incident flow velocity
is increased. The power radiated from the plasma is proportional to the
plasma volume and, since the diameter of the plasma is smaller for higher
incident flow velocity, the radiating volume of the plasma is reduced.
The development of detailed numerical models has made it possible to
study and interpret the complex interactions among optical geometry, flow,
and pressure that occur within the LSP. The results of these studies indicate
that the characteristics of the LSP can be controlled by a judicious choice of
these parameters to produce plasmas in a wide variety of gases with plasma
properties that can be optimized for a particular application. This detailed
understanding, together with the increasing availability of industrial carbon
dioxide lasers of good beam quality, should greatly expand the practical ap-
plications of the laser-sustained plasma.
4.4 EXPERIMENTAL STUDIES
Experimental studies of the LSP have been limited due to the requirement
for continuous laser power above 1 kW and the lack of clearly identified
applications. Until recently, most of the experimental studies were directed
toward defining combinations of laser power, plasma material, and pressure
for which the LSP could be successfully operated. Within the past few years,
more detailed studies have been made that better elucidate the interactions
among optical geometry, pressure, and flow that control the characteristics
of the LSP.
4.4.1 Operational Parameters

and Plasma Properties
The first systematic study of the LSP was reported by Generalov et al.
(1972). They ignited the plasma using a pulse breakdown in argon or xenon
produced by a pulsed and Q-switched carbon dioxide laser and then sus-
tained the plasma with a continuous carbon dioxide laser having an unspec-
ified beam and focusing configuration. The transition from the pulsed laser
190 Keefer
spark to a continuous plasma was investigated and they measured propaga-

tion velocities of 10m/sec that rapidly decreased to zero. It was found that
the pulsed laser could also extinguish the LSP once it had been established,
if the LSP was operating near threshold. They photographed the plasma
and noted the change in shape and position of the plasma as a function of
pressure and laser power. Stark broadening of the hydrogen H f3 line was
used, together with the Saha equation, to obtain estimates of the plasma
temperature that ranged from 13,000 to 23,000K in argon plasmas.
The principal quantitative results of the Generalov et a!. (1972) study
was a determination of the regions of laser power and pressure for which
continuous plasmas could be sustained in argon and xenon. Their mea-
surements included pressures from 1 to 45 atm and laser power to 300 W.
The threshold power was found to decrease rapidly with increased pressure
in the range from 1 to 10 atm but was nearly constant at higher pressures.
They attributed this pressure dependence of the threshold power to a tran-
sition of the dominant power loss mechanism from thermal conduction to
thermal radiation. When the LSP was operated in a horizontal beam, they
observed an upper pressure limit for existence of the LSP, but when oper-
ated in a vertical beam, no upper limit for pressure was observed. They
attributed this result to the stabilizing effect of the thermally induced buoy-
ant flow. Curiously, they also observed an upper limit for laser power at
the higher pressures when the beam was operated horizontally, but they
did not observe an upper power threshold when the beam was operated in
a vertical orientation. These results for the horizontal beam are shown in
Fig. 4.7. The curves for argon and xenon are similar in shape, but the ar-
gon curve is shifted to higher threshold power and pressure. Note that for
xenon above 10 atm and argon above 18 atm, there was an upper limit of
laser power for which the LSP was observed. The average laser power trans-
mission was also measured, and it was estimated to be approximately 40%
in these experiments. Carloff eta!. (1981) extended the pressure range of
these investigations to 210 atm for argon and helium, using a laser having
good mode structure and focused with mirrors of short (1.5-2.5 em) focal
length.
Franzen (1972) and Moody (1975) conducted similar experiments with
argon plasmas. Moody (1975) made a detailed study of the maintenance
thresholds using a laser with good beam quality, and he further investigated
the extinction of a continuous plasma when it absorbed a high-power laser
pulse. Moody was able to sustain continuous plasmas in argon at pressures
above 10 atm using a focal spot size of 80 ~tm diameter and incident powers
as low as 25 W. Similar studies were extended to the molecular gases hy-
drogen, deuterium, nitrogen, and air by Kozlov eta!. (1979), and the results
were similar to those obtained for the rare gases except that the threshold
power increased for the molecular gases.
PL, watt
300~~----------------r---------------------,--------------------~
OL....--------~--------~------~
0 10 20 30
p, atm
Figure 4. 7 Experimentally measured minimum power required to sustain plasmas

in argon (open circles) and xenon (closed circles). For argon above 18 atm and
xenon above 10 atm there is also a maximum power for which the plasma can be
sustained. (From Generalov, N. A., et a!., 1972. American Institute of Physics;
reprinted with permission.).
Klosterman and Byron (1974) performed experiments on plasmas in air

that were initiated from metallic targets using relatively long pulses from a
gasdynamic laser at powers to 400 kW. These plasmas were not stationary,
but propagated into the beam at subsonic velocities. They measured the
propagation velocity as a function of beam diameter and power, and these
results were widely used to compare the early theoretical models for the
laser-supported combustion wave (LSC). The observed velocities ranged
from 10 to 50 m/sec and were much higher than those predicted by the one-
dimensional model of Raizer (1970) when the radiation transfer was ne-
glected. Similar measurements were also made by Fowler and Smith (1975)
using well-characterized beams from continuous carbon dioxide lasers at
powers to 15 kW, and they also studied the ignition thresholds for aluminum
and steel targets in the presence of a transverse wind velocity to 20 m/sec.
Detailed measurements of the structure of the LSP were obtained by
Fowler and Smith (1975) using interferometric methods, and by Keefer
et al. (1975) using spectroscopic methods. The measurements were ob-
tained for plasmas sustained in ambient air by a vertical beam from a con-
tinuous carbon dioxide laser operating at a nominal power of 6 kW. These
results indicated that the maximum temperature in the plasma was approx-
imately 17,000K with the 15,000K isotherm nearly c0inciding with the outer
192 Keefer
edge of the laser beam. The diameter of the 10,000K isotherm was approx-
imately 10 mm, its length was approximately 14 mm, and the position of the
temperature maximum was approximately 11 mm from the focus. Keefer
et al. (1975) also found spectroscopic evidence that the plasma was not in
LTE in the outer regions of the plasma where the temperature gradients
were large.
4.4.2 Effects of Flow and Optics
Recent interest in the use of the LSP as a means to heat hydrogen for use
in a high specific-impulse laser-powered rocket (Jones and Keefer, 1982;
Caveny, 1984) has led to new and more detailed experimental studies. A
primary objective of these studies has been to determine the absorption
characteristics and stability of the plasma under the influence of forced con-
vective flow.
Conrad et al. (1979) studied LSC waves in hydrogen that were initiated
from a spark discharge. They also described experiments in ambient air
in which a plasma was stabilized at different positions in a large f/number
beam by flowing air under pressure in a direction opposite to the plasma
propagation. They found that the laser power transmitted by the plasma
varied from nearly 100%, near the threshold irradiance of 9 kW/cm2 , to
nearly zero, when the average laser irradiance exceeded 20 kW/cm 2 • These
results, not widely available, were crucial for the propulsion application of
the LSP, since they indicated that virtually 100% of the laser power could
be absorbed by the plasma if the flow configuration was correct.
Measurements of the flow velocity in the cooler regions surrounding the
high-temperature plasma core were obtained by Carloff et al. (1984) using
a laser Doppler anemometer. The plasma was sustained in argon within
a closed cell, and it was observed that a thermal buoyancy driven recircu-
lation cell developed in the chamber. The measured velocity field near
the plasma showed that the streamlines were deflected outward around
the high-temperature core of the plasma. They found that the velocity far
from the plasma core was about 5 em/sec and increased near the plasma to
20 em/sec. A more detailed investigation of the effect of flow on the range
of power and pressure for which the LSP could be maintained was reported
by Gerasimenko et al. (1983) who also established that the discharge wave
velocity exhibited a minimum value.
Experiments were conducted with flowing argon by Krier et al. (1986)
using a 10 kW materials processing laser. The annular beam from the un-
stable oscillator was focused into a 137 mm diameter chamber by a com-
plicated system of mirrors that permitted variation of the f/number of the
focused beam between 2.2 and 3.~. Mass flow rates of 2.3 to 4.6 g/sec pro-
vided bulk flow velocities from 11 to 22 em/sec. Detailed temperature mea-

surements of the outer regions of the flow below 3000K were made using
tungsten-rhenium thermocouples. Measured isotherms were presented for
two pressures, 1.1 atm and 3.3 atm, and two incident laser powers, 3.3 and
6.5 kW. They also measured the plasma absorption using a water-cooled
copper cone calorimeter and found that the fractional absorption of the
laser beam approached 80% at the higher laser powers.
Detailed spectroscopic measurements of the plasma region above
10,000K were obtained by Keefer et al. (1986) for plasmas sustained in a
vertical flow of argon by a 1.5 kW C0 2 laser. The plasmas were sustained
in a quartz tube of 22 mm diameter having a conical entrance section. The
experimental apparatus is shown in Fig. 4.8. Digital images of the plasma
continuum emission in a 1 nm bandpass centered at 626.5 nm, a spectral
region free of line emission, were acquired at video frame rates (60/sec) us-
ing a digital image processing computer. These images were Abel-inverted
to obtain plasma emission coefficients that could be related to the plasma
temperature using an assumption of LTE. Geometric ray tracing through
the measured temperature field provided a spatially resolved measurement
of the power absorption from the laser beam. They found little effect of
flow within the range of 0.8 to 32 em/sec incident flow velocity, which is in
the range of the velocities induced by thermal buoyancy. For this limited
range of incident flow velocity, they found that the flow had little influence
on the characteristics of the plasma. The detailed spatial measurements,
for plasmas sustained with an annular beam focused by lenses of 203 and
305 mm focal length, revealed the importance of spherical aberrations on
the structure of the plasma. They also found that the pressure dependence
of the plasma was considerably different for the two lenses and concluded
that experimentally determined threshold values were specific to the optical
geometry used. Fig. 4.9 shows the pressure dependence of fractional power
absorption observed using focusing lenses of two different focal lengths.
Note that, for the longer focal length lens, the absorbed power decreases
sharply near a pressure of 2 atm. Indeed, it was not possible to sustain a
plasma above 2 atm using the 305 mm focal length lens and a volume flow
of 3.2 standard L/min.
This experimental investigation was extended by Welle et al. (1987) to
include incident convective velocities up to 4.5 m/sec, considerably larger
than those induced by thermal buoyancy. At the higher flow velocities, the
plasma shape, size, and position within the focal volume were significantly
influenced by the forced convection. Fig. 4.10 shows that, in the case of low
flow, the plasma stabilized at a position well away from the focal point and
developed off-axis temperature maxima. The off-axis maxima were created
by the relatively high intensity in an annular ring that resulted from spher-
ical aberrations. With an increase in the incident flow velocity, the plasma
194 Keefer
~Exhaust
Quartz----.:
Jacket
CID Camera
~-~---Quartz
Flow Channel
1.5kW Laser Beam
Figure 4.8 Experimental apparatus used to obtain detailed spectroscopic temper-

ature measurements in a flowing argon LSP. (From Keefer, D., et al., 1986. Copy-
right © American Institute of Aeronautics and Astronautics; reprinted with per-
mission.).
70
-'E
CD
"g
60
·u
.....,
.E
*
•
0 50
:!!.
~
.. •
"g
CD .A
..0
0 40
I
..
Ill
..0
<C
CD
0==
a.. 30
20
1.2 1.4 1.6 1.8
• 2.0 2.2 2.4
Pressure (Atm.)
Figure 4.9 Fractional power absorption as a function of pressure for three sets of
plasmas: 203 mm focal length, 3.2 std. liters/min (.A); 203 mm focal length, 9.6 std.
liters/min (T); 305 mm focal length, 3.2 std. liters/min (+). The incident power
was 710 W. (From Keefer, D., et al., 1986. Copyright © American Institute of
Aeronautics and Astronautics; reprinted with permission.).
stabilized at a position much nearer the focal point. away from the annular
prefocus. The region of maximum temperature was more compact and was
positioned within the region of maximum beam intensity. Because of the
smaller volume of the high-temperature region, the radiation loss decreased
from 53 to 45% ofthe incident radiation, and since the higher-temperature
portion of the plasma stabilized in a higher-intensity portion of the beam,
the power absorption increased from 67 to 78% of the incident beam power.
This effect has significant implications regarding the regions of power and
pressure for which the LSP can exist. For example, although it was impos-
sible to sustain a plasma above 2 atm using the 305 mm focal length lens
and 3.2 standard L/min flow, the plasma could be sustained with this lens at
pressures above 2 atm when the incident flow velocity was increased.
Detailed analyses of a large number of laser-sustained plasmas have re-
vealed the complex interactions between the flow and the energy conversion
processes within the plasmas. These experiments suggest that stable laser-
196 Keefer
1 .8
""'E /
1 .2
/
E /
"'-../
./
OJ 0.6
0
c
0.0
....rl0
-o 0.6
0
-o 1 .2
0
1..
1 .8
laser power= 728 W

3.0 velocity=4.070 rn/s
52.5 55.0 57.5 60.0

Figure 4.10 The effect of incident flow velocity on the measured isothermal con-
tours for two plasmas sustained using a 203 mm focal length lens in argon at a
pressure of 2.5 atm. The upper plasma was sustained in an incident flow velocity
of 0.58 m/sec while the lower plasma was sustained in an incident flow velocity of
4.07 m/sec. The contour interval is 500K, with the outer contour at 10,500K. (From
Welle, R., et al., 1987. Copyright ©American Institute of Aeronautics and Astro-
nautics; reprinted with permission.).
sustained plasmas may be operated over a wider range of conditions than

previously reported by using appropriate combinations of optical geometry,
pressure, and incident flow velocity.
4.5 APPLICATIONS OF THE

LASER-SUSTAINED PLASMA
Although the laser-sustained plasma can produce small, well-controlled,

continuous plasmas in a variety of gases having peak temperatures greater
than those produced by the ICP or de arcs, there have been few reported
applications. This is probably due, in part, to the fact that the high-power
carbon dioxide lasers used to sustain the plasmas are relatively expensive
and have not been widely available. In recent years, this situation has been
changing as the carbon dioxide laser has found increasing acceptance by in-
dustry for various materials processing applications, and it is likely that new
applications will increase in the future.
4.5.1 Laser Propulsion
The concept for using the power beamed from a laser for rocket propul-
sion was proposed by Kantrowitz (1971) and Minovitch (1972). Some early
research on this concept was carried out, primarily by AVCO Everett Labo-
ratories and Physical Sciences, Inc. These applications, however, required
very large laser power (1 to 1000 MW) and interest waned until recently,
when research on the free-electron laser suggested that these large laser
powers were possible. A review of the early work on the laser propul-
sion concept was given by Glumb and Krier (1984) and, in a special vol-
ume edited by Caveny (1984), several chapters were devoted to the use
of beamed laser power to provide propulsion for orbit raising missions.
New experimental investigations of laser-sustained plasmas were begun at
NASA Marshall Spaceflight Center, The University of Tennessee Space In-
stitute, and the University of Illinois. A theoretical study was undertaken at
the University of Pennsylvania. The objectives of the experimental studies
were to develop an understanding of the basic physical processes involved
in the stability and power absorption of the LSP in a flowing gas. The goal
of the theoretical work was to develop numerical methods that could deal
with the strong gradients of density and temperature and the strong cou-
pling of the plasma transport and thermal properties that exist in the LSP.
Many of the results from these recent investigations have been discussed in
the preceding sections.
The use of a LSP to absorb the power from a laser beam and to convert
it to enthalpy in the propellant gas has been called laser thermal propul-
sion (Keefer et al., 1984). This refers to the concept in which power from
a laser is beamed to a vehicle where it is absorbed by the LSP and used to
heat a pure hydrogen propellant to provide thrust at large values of spe-
cific impulse (1000 to 1500 sec). With these high specific-impulse values,
the fractional payload that can be delivered from low-earth orbit (LEO) to
geosynchronous earth orbit (GEO) is significantly larger than that achieved
by chemical propulsion systems. Furthermore, the mass of the propulsion
system can be smaller than other high specific-impulse systems (e.g., nuclear
or electric propulsion) since the source of power is beamed to the vehicle
from a remote site.
198 Keefer
LASER
BEAM>
Figure 4.11 Schematic representation of a rocket motor that operates by laser

thermal propulsion. It is predicted that the specific impulse (lsP) from such a motor
may be several times greater than that obtained from conventional rocket motors.
A schematic representation of the rocket motor is shown in Fig. 4.11.

The incoming laser beam is focused into the absorption chamber where it
sustains a plasma at temperatures of approximately 16,000 to 20,000K. In
contrast to a normal combustion chamber, the high-temperature plasma oc-
cupies only a small fraction of the chamber volume and mixes with the in-
coming propellant to produce a flow having a bulk temperature of approxi-
mately 3000 to 5000K that can be expanded through the nozzle to produce
thrust. Due to the high temperature of the LSP, a substantial portion of
the absorbed power will be radiated from the plasma and absorbed by the
chamber walls. In order to utilize this radiated power effectively, the cham-
ber walls are cooled by the incoming propellant in a regenerative cycle. If
the power radiated from the plasma can be limited to that required to raise
the propellant from storage temperature to an acceptable inlet tempera-
ture, then all of the absorbed power can be utilized to heat the propellant
and no external radiator will be required. Sufficient mixing of the inlet pro-
pellant and the high-temperature region within the LSP must be insured to
produce the bulk temperature required for the nozzle inlet and to minimize
the radial temperature gradients in the nozzle flow.
The fraction of absorbed laser power that is lost from the plasma as
thermal radiation has a profound influence on the performance that can
be achieved by the thruster (Keefer et al., 1987). If the acceptable inlet
temperature limits the enthalpy of the inlet flow to h0 , then the final bulk
enthalpy will be limited to hy = h 0 /a.RAD• where aRAD is the fraction of
the absorbed laser power that is radiated to the walls by thermal radiation.
The final bulk enthalpy and, therefore, the specific impulse of the thruster
will be determined by the radiated fraction of the absorbed laser power that
may be absorbed by regenerative cooling. It is clear that the success of this
concept rests on the ability to sustain a stable plasma in a forced convective
flow that can absorb substantially all of the incoming radiation and to limit
the fraction of the absorbed laser power that is lost from the plasma region
through thermal radiation.
Another factor that will affect the performance of this rocket concept is
the radial temperature distribution at the nozzle throat. Due to the strong
radial temperature gradients associated with the LSP, unless there is suffi-
cient mixing downstream of the plasma with the surrounding colder buffer
flow, there will be a much higher-temperature core near the axis in the
throat. This results in an inefficient utilization of the absorbed power to pro-
duce thrust, with a corresponding loss in specific impulse. Jeng and Keefer
(1987b) have developed a new computational model that spans the subsonic
to supersonic flow regime to study these effects.
Recent, detailed experiments with the LSP in a forced convective flow
have revealed the complex interactions between the flow, pressure, and
laser power that determine the stable operating regimes for these plasmas.
It was found that the position of the plasma within the focal volume of the
laser beam could be controlled with the incident flow velocity. Controlling
the flow, together with the focusing geometry of the laser beam, permits sta-
ble operation of the plasma with high fractional absorption of the incident
laser beam and relatively small thermal radiation losses from the plasma.
Numerical models that incorporate the complex energy conversion pro-
cesses within the plasma have been developed and verified using the de-
tailed experimental measurements. Results from these model calculations
indicate that hydrogen plasmas can be sustained that will absorb essentially
all of the incident laser power and will lose a sufficiently small fraction of the
absorbed power through thermal radiation. The numerical studies for hy-
drogen plasmas indicate that when higher-power Gaussian laser beams are
employed, the chamber walls can be adequately cooled by the propellant
flow in an efficient regenerative cycle. No "fatal flaws" have been discov-
ered by either the experimental or theoretical investigations, and it appears
that a practical, high specific-impulse propulsion system could be based on
the laser thermal propulsion concept, if lasers having adequate power are
developed and suitable window materials can be found.
4.5.2 Atomic and Molecular Beams
A device, similar in configuration to a laser thermal rocket motor, has been

used by Cross and Cremers (1985 and 1986) to produce high-particle flux
beams of xenon, argon, and atomic oxygen. The objective of this research
200 Keefer
HIGH VOLTAGE ELECTRODE
LENS
HIGH TEMPERATURE HOLDER
GAS BEAM 10.6 p.m 70W C0 2
LASER BEAM
TUNGSTEN 0-RING SEAL

NOZZLE
R COOLING
L = ZnSe LENS
( 25.4 mm diameter I fl)
:-~-LASER SUSTAINED PLASMA
..../ 1\'
Po > 1000 torr
P 1 < I torr
Figure 4.12 Schematic of a high velocity molecular beam source that utilizes a
laser-sustained plasma. (From Cross, J. B. and Cremers, D. A., 1985. Copyright ©
American Institute of Aeronautics and Astronautics; reprinted with permission.).
is the production of atomic oxygen beams that can be used to study the in-
teraction of oxygen atoms with spacecraft surfaces at velocities and particle
fluxes similar to those encountered in the orbital environment.
The plasma is sustained in a chamber near a small diameter nozzle as
shown in Fig. 4.12. The gasdynamic expansion of the high-temperature gas
into a vacuum produces a high-velocity molecular beam. Time of flight mea-
surements using xenon and a laser power of 70 W showed that the beam tem-
perature (velocity) increased dramatically as the plasma is moved closer to
the nozzle as shown in Fig. 4.13. The resulting xenon beam had a velocity
of 1.34 km/sec and an absolute intensity estimated at 1019- 20 particles/sec-
sr. A platinum nozzle was developed for use with higher laser powers and
mixtures containing oxygen. This beam was operated at powers to 500 W,
but the time of flight measurements was not reported. Based on their re-
sults with xenon, the authors estimate that velocities of 3.6 km/sec can be
18
17
8000 - 16
15
7000 - 14
13
12
... ,. ...... ,...-;------,

5000 ° , ___ 10
... /
/
9 <l
......
4000 ... /
/
8
7
2000
1000
Figure 4.13 Variation of beam temperature and (t/tJ..t )as a function of the plasma
to nozzle distance for a xenon beam sustained with a 70 W laser. The data were
obtained from an analysis of time-of-flight data. (From Cross, J. B. and Cremers,
D. A., 1986. North-Holland Physics Publishing, Amsterdam; reprinted with permis-
sion.).
obtained using argon, 6 km/sec using neon, and 14.8 km/sec using helium as
the carrier gas.
4.5.3 Spectrochemical Analysis
Various kinds of gas discharge plasmas, including de arcs, ac arcs and

sparks, the ICP, and microwave discharges, have long been used as exci-
tation sources for atomic emission spectroscopy. As described in the pre-
202 Keefer
INJECTION PORT\
:.F
v::J wee_/ -wee
QUARTZ TUBE
I I I
o 5 em
Figure 4.14 Diagram of the small gas cell that was used in analytical experiments
with a static xenon fill. Gas samples were introduced through the injection port
using a syringe. Solid samples were laser ablated from a metal strip located below
the LSP. Here wee is a water cooling coil, G is a pressure gauge, and Vis a vacuum
pump. (From eremers, D. A, et al., 1985. Pergamon Press, Ltd.; reprinted with
permission.).
ceding sections, the LSP produces a continuous high-temperature plasma

that is free of electrode contamination, and it is natural to assume that it
could also be used as an excitation source. The use of the LSP as a source
for emission spectrochemical analysis has been evaluated by Cremers et al.
(1985) who generated the plasma in xenon using a 45 W carbon dioxide
laser.
They used two similar cells, the smaller of which is shown in Fig. 4.14.
A monochromator, together with a chopper and lock in amplifier, was used
to record the spectrum from the plasma. The plasma was initiated using
the laser spark generated by focusing the pulse from a 2 MW, Nd:YAG
laser. It was found that stable xenon plasmas could be maintained from
1150 to 3200 Torr using a laser beam power of 45 W. Detailed spectra of
xenon from 200 to 950 nm were obtained for various conditions of power
and pressure. The electron density was measured by seeding the discharge
with a small amount of hydrogen and using the broadening of theHa. line.
The plasma temperature was determined from a Boltzmann plot using the
emission lines from xenon and krypton that were added to the discharge
gas. Their results were consistent with an assumption of LTE within the
plasma. The stability of the emission was determined, and for the small cell
the background (continuum) emission was found to have a relative standard
deviation (RSD) of 0.3 to 0.6%, and for the 788.74 nm Xe I line the RSD
was 0.42%.
Chlorine and oxygen were added to the xenon discharge and calibration
curves were obtained for 0 and Cl atoms. It was found that as much as 10%
by volume of chlorine or oxygen could be added to the xenon without no-
ticeably affecting the discharge. Several solid compounds were introduced
into the discharge by pulsed laser ablation from a steel substrate located
below the LSP. Approximate detection limits were determined and found
to be higher than those reported for the ICP. The authors conclude that
there are several advantages of the LSP for spectrochemical analysis, but a
more detailed performance evaluation should be carried out using argon as
the carrier gas and operating the LSP as a plasmatr:m.
4.5.4 Other Potential Applications
The LSP is capable of generating a stable, contamination-free plasma at

temperatures and pressures comparable to, or exceeding, that of de arcs in
a variety of gases over a wide range of pressures and flow. In addition, it is
possible to generate small plasmas with absorbed powers that range from
a few watts to tens or hundreds of kilowatts. Electric discharge plasmas
have found numerous applications in chemical synthesis and materials pro-
cessing and, with the increasing commercial availability of suitable lasers,
it is reasonable to expect that similar applications will be found for laser-
sustained plasmas.
REFERENCES
Batteh, J. H. and Keefer, D. R. (1974). Two dimensional generalization of Raizer's

analysis for the subsonic propagation of laser sparks, IEEE Trans. Plasma Sci.,
PS-2: 122.
Carloff, C., Krametz, E., Schafer, J. H., Uhlenbusch, J., and Wroblewski, D. (1981).
Continuous optical discharges at very high pressure, Physica, 103c: 439.
Carloff, C., Gillet, C., Krametz, E., Muller, A., Schafer, J. H., and Uhlenbusch, J.
(1984). Theoretical description and measurement of the flow field of a continu-
ous optical discharge, Arch. Mech., 36: 473.
Caveny, L. H., Ed. (1984). Orbit Raising and Maneuvering Propulsion: Research Status
and Needs, ALAA Progress in Astronautics and Aerona.ltics series, VoL 89, Wash-
ington, D.C.
204 Keefer
Conrad, R. W., Roy, E. L., Pyles, C. E., and Mangum, D. W. (1979). Laser-supported
combustion wave ignition in hydrogen, Tech. Rep. RH-80-1, U.S. Army Missile
Command, Redstone Arsenal, Alabama.
Cremers, D. A, Archuleta, F. L., and Martinez, R. J. (1985). Evaluation of the con-
tinuous optical discharge for spectrochemical analysis, SpectrochimicaActa, 40B:
665.
Cross, J. B. (1986). Private communication.
Cross, J. B. and Cremers, D. A (1985). Ground-based investigations of atomic oxy-
gen interactions with space station surfaces, AIAA Paper 85-1068, AIAA 20th
Thermophysics Conf., Williamsburg, Va.
Cross, J. B. and Cremers, D. A (1986). High kinetic energy (1-10 eV) laser sus-
tained neutral atom beam source, Nucl. Instrum. Meth. Phys. Res., B13: 658.
Fowler, M.C. (1981). Measured molecular absorptivities for a laser thruster.
Fowler, M. C. and Smith, D. C. (1975). Ignition and maintenance of subsonic plasma
waves in atmospheric pressure air by CW C0 2 laser radiation and their effect on
laser beam propagation, J. Appl. Phys., 46: 138.
Franzen, D. L. (1972). CW gas breakdown in argon using 10.6-~.tm laser radiation,
Appl. Phys. Lett., 21: 62.
Generalov, N. A, Zimakov, V. P., Kozlov, G. I., Masyukov, V. A, and Raizer, Y. P.
(1970). Continuous optical discharge, Sov. Phys. JETP Lett., 11: 302.
Generalov, N. A, Zimakov, V. P., Kozlov, G. I., Masyukov, V. A, and Raizer, Y. P.
(1972). Experimental investigation of a continuous optical discharge, Sov. Phys.
JETP, 34: 763.
Gerasimenko, M. V., Kozlov, G. I., and Kuznetsov, V. A (1983). Laser plasmatron,
Sov. J. Quant. Electron., 13: 438.
Glumb, R. J. and Krier, H. (1984). Concepts and status of laser-supported rocket
propulsion, J. Spacecraft and Rockets, 21: 70.
Glumb, R. J. and Krier, H. (1985). A two-dimensional model oflaser-sustained plas-
mas in axisymmetric flow fields, AIAA Paper 85-1533, AIAA 18th Fluid Dynam-
ics and Plasmadynamics and Lasers Conf., Cincinnati, Ohio.
Griem, H. R. (1964). Plasma Spectroscopy, McGraw-Hill, New York.
Holt, E. H. and Haskell, R. E. (1965). Foundations ofPlasma Dynamics, Macmillan,
New York.
Jackson, J.P. and Nielsen, P. E. (1974). Role of radiative transport in the propagation
of laser supported combustion waves,AL4A J., 12: 1498.
Jeng, S.M. (1986). Unpublished calculation.
Jeng, S. M. and Keefer, D. R. (1986). Theoretical investigation of laser-sustained
argon plasmas, JAppl. Phys., 60: 2272.
Jeng, S. M. and Keefer, D. (1987a). Numerical study of laser-sustained hydrogen
plasmas in a forced convective flow, J. ofPower and Propulsion, 3: 255.
Jeng, S.M. and Keefer, D. R. (1987b ). A theoretical investigation of laser-sustained
plasma thruster, AIAA Paper 87-0383, AIAA 25th Aerospace Sciences Meeting,
Reno, Nevada.
Jeng, S.M., Keefer, D., Welle, R., and Peters, C. (1987). Laser-sustained plasmas
in a forced convective argon flow. Part II: Comparison of numerical model with
experiment,AL4A J., 25: 1224-1230.
Jones, L. Wand Keefer, D. R. (1982). NASA's laser-propulsion project, Aeronautics

and Astronautics, 10: 66.
Kantrowitz, A. R. (1971). The relevance of space, Astronautics and Aeronautics, 9:
34.
Karzas, W J. and Latter, R. (1961). Electron radiative transitions in a Coulomb
field,Astrophys. J, Suppl. Series 6: 167.
Keefer, D. R., Henriksen, B. B., and Braerman, W. F. (1975). Experimental study of
a stationary laser-sustained air plasma, J. Appl. Phys., 46: 1080.
Keefer, D., Elkins, R., Peters, C., and Jones, L. (1984). Laser thermal propul-
sion, Orbit-Raisingand Maneuvering Propulsion: Research Status and Needs,AIAA
Progress in Astronautics and Aeronautics series, Vol. 89 (L. H. Caveny, ed. ), AIAA,
Washington, D.C., p. 129.
Keefer, D., Peters, C., and Crowder, H. (1985). A re-examination of the laser-
supported combustion wave,AIAA !., 23: 1208.
Keefer, D., Welle, R., and Peters, C. (1986). Power absorption in laser-sustained
argon plasmas,AIAA J, 24: 1663.
Keefer, D., Jeng, S.M., and Welle, R. (1987). Laser thermal propulsion using laser
sustained plasmas,ActaAstronautica, 15: 367.
Kemp, N. H. and Root, R. G. (1979). Analytical study of laser-supported combus-
tion waves in hydrogen, J. Energy, 3: 40.
Klosterman, E. L. and Byron, S. R. (1974). Measurement of subsonic laser absorp-
tion wave propagation characteristics at 10.6 JJm,/. Appl. Phys., 45: 4751.
Kozlov, G. I. and Selezneva, I. K. (1978). Numerical study of a laser spark and a
continuous optical discharge in a focused laser beam, Sov. Phys. Tech. Phys., 23:
227.
Kozlov, G. I., Kuznetsov, V. A, and Masyukov, V. A (1974). Radiative losses from
an argon plasma and a radiation model of a continuous optical discharge, Sov.
Phys. JETP, 39: 463.
Kozlov, G. I., Kuznetsov, V. A, and Masyukov, V. A (1979). Sustained optical dis-
charges in molecular gases, Sov. Phys. Tech. Phys., 49: 1283.
Krier, H., Mazumder, J., Rockstroh, T. J., Bender, T. D., and Glumb, R. J. (1986).
Studies of CW laser gas heating by sustained plasmas in flowing argon,AIAA !.,
24: 1656.
Maker, P. D., Terhune, R. W., and Savage, C. M. (1963). Optical third harmonic
generation, 3rd Internal. Conf on Quant. Electron., Paris, France.
Merkle, C. L. (1984). Prediction of the fiowfield in laser propulsion devices, AIAA
!., 22: 1101.
Merkle, C. L., Molvik, G. A, and Shaw, E. J. (1985). Numerical solution of strong ra-
diation gasdynamic interactions in a hydrogen-seedant mixture, AIAA Paper 85-
1554, AIAA 18th Fluid Dynamics and Plasmadynamics and Lasers Conf., Cincin-
nati, Ohio.
Minovitch, M. A (1972). Reactorless nuclear propulsion-The laser rocket, AIAA
Paper 72-1095, AIAA/SAE 8th Joint Propulsion Specialist Conf., New Orleans,
La.
Molvik, G.A, Choi, D., and Merkle, C. L. (1985). A two-dimensional analysis of
laser heat addition in a constant absorptivity gas, AIAA J., 23: 1053.
206 Keefer
Moody, C. D. (1975). Maintenance of a gas breakdown in argon using 10.6-~-t CW

radiation, J. Appl. Phys., 46: 2475.
Muller, S. and Uhlenbusch, J. (1982). Theoretical model for a continuous optical
discharge, Physica, 112C: 259.
Raizer, Y. P. (1965). Heating of a gas by a powerful light pulse, Sov. Phys. JETP, 21:
1009.
Raizer, Y. P. (1970). Subsonic propagation of a light spark and threshold conditions
for the maintenance of plasma by radiation, Sov. Phys. JETP, 31: 1148.
Raizer, Y. P. (1980). Optical discharges, Sov. Phys. Usp., 23: 789.
Shkarofsky, I. P., Johnston, T. W., and Bachynski, M. P. (1966). The Particle Kinetics
of Plasmas, Addison-Wesley, Mass.
Thompson, R. W., Manista, E. J., and Alger, D. L. (1978). Preliminary results on the
conversion of laser energy into electricity, AppL Phys. Lett., 32: 610.
Welle, R., Keefer, D. R., and Peters, C. (1987). Energy conversion efficiency in high-
flow laser-sustained argon plasmas,AL4A J., 25: 1093.
Zel'dovich, Y. B. and Raizer, Y. P. (1965). Cascade ionization of a gas by a light
pulse, Sov. Phys. JETP, 20: 772.
5
lnertially Confined Fusion
Robert L. McCrory and John M. Soures

Laboratory for Laser Energetics
Rochester, New York
5.1 HISTORICAL OVERVIEW
Control of the thermonuclear fusion process for the purpose of producing

clean, economically competitive, electrical energy has been the ultimate ob-
jective of a worldwide scientific research effort for the last four decades.
The fusion process involves the nuclear burning of low-atomic-weight ele-
ments such as hydrogen to form heavier elements such as helium with the
subsequent release of energy in the form of energetic neutrons or charged
particles.
To overcome the Coulomb repulsive force that keeps atomic nuclei apart,
fusion fuel must be heated to temperatures in excess of several thousand
electron volts (50,000,000K). In addition, efficient fusion burning requires
that the resulting plasma be kept at a high density for a long enough time
to allow the burnup of a significant fraction of the fusion fuel.
Two general schemes have been investigated in the past 40 years as means
to confine fusion plasmas. Magnetic confinement fusion depends on the
use of high-intensity magnetic fields (in excess of several teslas) to confine
a relatively low-density (1014 /cm3 ) plasma for times on the order of tens of
seconds. This scheme has been pursued as an international research effort
for the last 30 years and is now a mature technology whose future develop-
ment prospects are limited primarily by the size, cost, and complexity of the
required devices (see, e.g., Conn, 1981).
Inertial confinement fusion (ICF) is an alternative approach to con-
trolled fusion that depends on rapid heating and compression of fusion
fuel contained in a spherical target (such as a polymer or glass shell) of
several millimeters diameter to ultrahigh densities (1000 times liquid DT
density ~ 200 gm/cm3 ). At these densities and at temperatures around
207
208 McCrory and Soures
5 ke V, the fusion reaction rate is high enough to allow efficient burning be-
fore the plasma disassembles. Ablation of the outer layers of the fuel pellet
creates the ultrahigh pressures (;:::50 Mbar) required to compress the fuel
to ultrahigh densities. A potential advantage of the ICF approach is the
decoupling of the energy source (particle beam or laser) from the reactor
vessel, significantly reducing the reactor operating costs and minimizing the
amount of material subjected to intense neutron bombardment and activa-
tion compared with a magnetic confinement reactor.
The first demonstration of a man-made inertial confinement fusion de-
vice came with the explosion by the United States of the first hydrogen bomb
in 1952. In the intervening years, a number of concepts were generated
for laboratory-inertial-confinement devices using particle beams or intense
laser pulses by Kidder (1963) and Dawson (1964) in the U.S., and by Basov
and Khrokhin (1964) in the Soviet Union.
In the late 1960s, the availability of large Nd:glass lasers marked the be-
ginning of experimental laser-fusion research. The first experiments result-
ing in measurable neutron generation from laser-heated targets were con-
ducted at the P. N. Lebedev Physics Institute in Moscow by Basov et al.
(1968). By the early 1970s, inertial fusion programs were underway at
the Lawrence Livermore National Laboratory, the University of Rochester
Laboratory for Laser Energetics, Los Alamos National Laboratory, Sandia
National Laboratory, KMS Fusion Inc., and the Naval Research Laboratory
in the U.S.; the Limeil Laboratory in France; the Rutherford Laboratory in
Great Britain; Osaka University in Japan; and the Max Planck Institute at
Garching in West Germany (see, e.g., Floux et al., 1970; Nuckolls et al.,
1972; Yamanaka et al., 1972; Shearer et al., 1972; Soures et al., 1973; Stam-
per et al., 1974; McCall et al., 1973; Charatis et al., 1975; Eidman and Sigel,
1974). As tabulated in the World Survey ofMajor Activities in Controlled Nu-
clear Fusion Research, 1986 edition, there are now laser-fusion and laser-
matter interaction programs operating in 27 institutes in 12 nations using a
total of 59 laser systems of various types.
There are two principal approaches to inertial confinement fusion: di-
rect and indirect drive. In the indirect, radiation or "hohlraum" approach
to fusion, the main approach pursued in the U.S. National Laboratories,
the driver-beam energy is absorbed and converted to x-rays by a case made
of an element of high atomic weight (Nuckolls, 1982); the x-rays are then
used to drive the target's implosion. In the alternative, direct-drive ap-
proach, a short-wavelength, high-intensity, laser pulse directly illuminates
a bare spherical target (see Fig. 5.1). Since no intermediate conversion pro-
cess is required for the direct-drive approach, it is potentially more energy-
efficient than the hohlraum approach; alternatively, with the hohlraum ap-
proach, it may be possible to achieve improved implosion symmetry without
requiring the drive beams to be uniform or symmetrically disposed.
Jnertlally Confined Fusion 209
HOT TARGET
BLOW-OFF
LASER
PULSE
\
\
CRITICAl DENSITY~
(maximum point of
..-:'
laser penetration) ABLATION SURFACE

(region of rocket
action)
Figure 5.1 Schematic diagram illustrating the direct-drive approach to inertial fu-
sion.
The key requirement for either approach to inertial fusion to produce

high gain is the simultaneous achievement of high density and high temper-
ature. A simple argument can be used to demonstrate the need for high
density. The energy released from complete burn of DT is 3.5 x 1011 J/g.
Of course, all of the fuel will not be consumed for the small pellet sizes of
interest for inertial fusion. The burnup fractionfb can be simply estimated
from
dNv
- dt = -NvNr(av} (5.1)
where N D and NT are the ion number densities of deuterium and tritium,
respectively, and (av} is the DT fusion reaction rate (in cm3/sec) averaged
over a Maxwellian distribution. For an equimolar DT mixfb = 1- 2NviN0,
whereN0 is the sum of N D andNr initially. Integrating Eq. (5.1) gives
(5.2)
where r8 is the burn time. If we assume that r8 "'Rl4cs, where R is the

compressed radius of the fuel, cs is the sound speed (see Fraley et al., 197 4a),
then when the ion density is converted to mass density,
pR
(5.3)
fb = pR + [3(T)
where [3(T) = canst. x(csf (av}), and p is the mass density of the com-
pressed fuel. The quantity [3(T) is a strong function of temperature; if
we assume that the ions are isothermal and the ion temperature is much
greater than the electron temperature and use Hively's data (1977) for (av),
f3(T) = 19 g/cm2 at 10 keY and 6.8 g!cm2 at 20 keY. Thus, to obtain a
fractional burn up greater than 30% at 20 keV, a pR greater than 3 g!cm2 is
required. For a given fuel density and pR = 3 g!cm2 , we can estimate the
compressed fuel mass and thermonuclear energy release. Since the com-
pressed fuel mass scales as p- 2 for fixed pR, with fb = 1/3, at solid density
the yield for pR = 3 is 3 x 1014 J, whereas at compressions of 1000 times
solid, the yield is 3 x lOS J. Since reactor engineering constraints limit the
energy release to about 109 J per event, it is evident that densities around
1000 times liquid DT density (200 g!cm3 ) are required for an economical
laser-fusion reactor.
In fact, the need for high density is more than a practical necessity; it is
also an economic and efficiency requirement. Approximately 6 x 108 J/gm
are required to heat the DT fuel to a temperature of 5 keY; if the fuel is
kept cold, however, the energy investment to compress fuel to a density of
1000 times liquid DT density is only 107 J/g; for a Fermi-degenerate DT, the
internal energy is 3.5 x 105 [p (g/cm3 )] 213 J/g. To maximize the energy gain
from an inertial fusion reaction, all of the DT fuel should be compressed to
a high density, but it is only required to heat and compress simultaneously
a small fraction of the total fuel mass. Once ignition of this hot spot is real-
ized, the resulting high-energy alpha particles can redeposit their energy in
the cold, dense fuel and initiate a thermonuclear burn wave. The achieve-
ment of thermonuclear ignition in a small fuel core with modest laser energy
is an important milestone in laser-fusion research.
The key physics issues facing both the direct- and indirect-drive ap-
proaches to inertial confinement fusion are (1) demonstration of efficient
absorption of the drive beam energy by the target; (2) generation of suffi-
ciently high ablation pressures with low fuel preheat; and (3) achievement
of high-density compressions (;:::: 1000 times liquid DT).
In the early days of laser-fusion research, there was much excitement re-
garding the prospects for demonstrating the feasibility of the process with
kilojoule-class laser systems (see, e.g., Brueckner and Joma, 1974) by com-
pressing targets to ultrahigh densities (;:::: 10,000 times liquid DT density).
However, the first experimental results, obtained largely with infrared laser
systems, were disappointing. In these experiments, it was found that a sig-
lnertlally Confined Fusion 211
nificant fraction of the laser energy was converted to suprathermal elec-

trons whose mean free path was long compared with the dimensions of the
target. These electrons cause decoupling of the absorbed energy from the
target and also preheat the target interior, increasing the specific entropy
of the fuel and decreasing the achievable compression for a fixed-driver en-
ergy. The generation of fast electrons was anticipated by Kaw and Dawson
(1969) and Kidder (1971).
By the mid-1970s, it was apparent to a number of researchers that short
wavelength lasers would be required to demonstrate the feasibility of laser
fusion with either the direct or indirect approaches (see, e.g., McCrory and
Morse, 1977a). By the late 1970s, efficient wavelength conversion tech-
niques were developed and demonstrated at the Laboratory for Laser En-
ergetics (LLE) and elsewhere (see, e.g., Seka et al., 1980 and Craxton,
1980). Experiments conducted with frequency-upconverted Nd:glass lasers
by Amiranoff et al. (1979), Mead et al. (1981), Seka et al. (1982), Slater et al.
(1981), and Yaakobi et al. (1981a) have demonstrated efficient coupling of
the laser energy to the target with minimal fast electron preheat.
Progress toward the objective of demonstrating the feasibility of iner-
tial confinement fusion has been rapid since the problem of fast electron
generation was overcome with short-wavelength lasers. At Lawrence Liv-
ermore National Laboratory, Ze et al. (1984) have demonstrated densities
in excess of 100 times liquid DT density. Thermonuclear burn efficiencies
(thermonuclear energy release divided by absorbed laser energy) in excess
of 0.1% have been demonstrated with modern high-energy Nd:glass lasers
operated at the second or third harmonic (see, e.g., Yamanaka et al., 1986;
Richardson et al., 1986b; and Lane et al., 1986). Advanced laser concepts
for future fusion lasers are under development, including the excimer KrF
gas laser (operating at 0.25 pm and being developed at the Los Alamos Na-
tional Laboratory.)
In the remainder of this chapter, we will discuss the basic physics of laser
fusion, review the international effort to address the critical issues facing
inertial confinement fusion research, and attempt to forecast the progress
expected in the next decade.
5.2 LASER-FUSION SCALING LAWS
We can address the requirements for significant thermonuclear energy re-

lease from direct-drive ablatively accelerated laser-driven targets by using
a very simple target model illustrated in Fig. 5.2, a levitated shell of fuel
having density p, assumed to be unity. The p = 1 case may be thought of as
once-shocked cryogenic DT fuel, or a combination of an outer low-Z abla-
tor and an inner layer of fuel. The laser irradiation is assumed to be spher-
ically symmetric and deposited in the outer region of the target. From this
simple model, one may estimate the energy, power, and aspect-ratio (mean
radius of the shell divided by the shell wall thickness) parametric regimes
of interest to directly driven laser-fusion experiments.
In order to achieve an ablative implosion, it is necessary that the shell
thickness, for a given incident laser irradiance, be large compared with the
range of any hot electrons generated in the plasma corona exterior to the
target. Estabrook and Kruer (1978) have suggested the following relation-
ship to estimate the hot electron temperature:
(
T )114 ( ¢> )0.39 ( >. )0.78 (5.4)
That= 8.5 1 ~t 1015 ~/cm2 1.06LJ.Lm keY
where >.L is the laser wavelength and ¢>L the incident laser intensity. The
fast (hot or superhot) (see Fig. 5.2) electron stopping distance Le, which is
DT Shell
(p-1 g/cm3 )
GENERAL REQUIREMENTS:
VIMP~ 3 x 107 em/sec- :
EoT- 109 Joules/g
W > Lb >> Le
Figure 5.2 Simple model of an ablatively driven shell implosion.

proportional to T~ol' then scales as
-~.0.78
L "' :!.1,__ ,\ 1.56 (em) (5.5)
e p L
The advantage of short wavelength lasers in reducing the required mass of

the target while still inhibiting preheat of the inner fuel regime is easily seen
from Eq. (5.5).
From several studies addressing the implosion velocities required for
marginal thermonuclear systems, it is reasonable to take a typical implosion
velocity of 20 to 40 cm/1-'sec (2 to 8 x 107 J/g). In the simplest picture, most
of the fuel is cold and degenerate, and the compression energy 107 J/g must
be furnished by the implosion. More detailed models give a number several
times higher than this. One model of estimating the fuel-energy balance
requirements for ignition is due to Widner (1979); it is similar to the more
elaborate 12-parameter model of Kirkpatrick (1979). With these estimates,
the kinetic energy per unit mass is estimated to be of order lOS J/g.
The hydrodynamic efficiency, defined to be the energy in the imploding
fuel mass divided by the laser energy absorbed, can be estimated to be of
the order of 10 to 20% (Bodner, 1981 and Ripin et al., 1980). It is prudent
to take the lower limit, 10%, to estimate the required absorbed energy to be
a fraction of 109 J/g for this simple system. The use of multishell designs can
reduce this energy requirement since these designs do not require acceler-
ating the entire fuel mass to such high velocities. In keeping with the spirit
of the section, however, it is our goal to use the simplest model possible in
order to map out the parameter space of interest to short-wavelength laser-
driven systems; using 109 J/g for single-shell designs will give us an upper
limit.
Given the target density (p = 1 g!cm3 here), radius, and final velocity,
the energy required, based on the above estimate of 109 J/g, determines the
shell aspect ratio.
For a fixed implosion timer, the absorbed intensity 4>A (in W/cm 2 ) is
given by
(5.6)
where w =wall thickness= RIA where R is the initial shell radius and A the
shell aspect ratio, r = RIV;.mp where V;.mp is the velocity of the imploding
shell, and e0 = 109 J/g is the specific absorbed energy.
In Fig. 5.3(a), we illustrate the absorbed intensity required for given im-
plosion times, 1, 2, and 10 nsec, as a function of shell wall thickness. Also
illustrated in the figure is a dashed line indicating the fast electron stopping
N
E
u
~ 1015
.c
'iiJ
c
c
<ll
D
<ll
.n
0
<rl 1014
.n
<( 1.2 Lb@ 0.35 11m, 1 ns
<(
-e
1.2 Lb@ 1.054 11m, 1 ns
10 100 1000
w -Wall Thickness (11 m)

(a)
1016
N
E
()
~
.ciiJ
c
c
<ll
D
<ll
··•···~
-2 1.2 Lb@ 0.35/lm, 1 ns
0
.n
<f) 1014
<(
1.2 Lb@ 1.054 11m, 1 ns
<(
-e
1013
1 10 100 1000
w -Wall Thickness ( 11 m)
(b)
Figure 5.3 Simple ablative scaling: (a) Pulse length as a function of absorbed in-
tensity and wall thickness; (b) aspect ratio as a function of absorbed intensity and
wall thickness; (c) absorbed energy as a function of absorbed intensity and wall
thickness; and (d) absorbed power as a function of absorbed intensity and wall thick-
ness.
lnertially Confined Fusion 215
w -Wall Thickness (Mm)

(c)
N
E
u
~ 1015
p~
w
2
~
'iii </>A
c
c
<ll
TI
-e
<ll
0
U) 1014
.0 1.2 Lb@ 0.35. I-'m, 1 ns .
<(
<(
'S 1.2 Lb@ 1.054 I-'m. 1 ns
1013
1 10 100 1000
w -Wall Thickness (I-'m)

(d)
distance 4Le from Eq. (5.5) if we assume </>A"' <I>L· The wall thickness used
in this simple model must be to the right of this line.
Figure 5.3(b) identifies several values of the aspect ratio (A = 10, 100,
1000) to illustrate the dependence ofirradiance choice on aspect ratio. The
importance of aspect ratio derives from hydrodynamic stability considera-
tions and is discussed in Sec. 5.6.
If one chooses to fix the input laser energy E, the relationship between
</>A and w is given by
- ervimp(wp)312y'4:;
<I>A- El/2 (5.7)
and is illustrated in Fig. 5.3(c).

We estimate the required laser power P by simply dividingE by r. For a
fixed laser power, the relationship plotted in Fig. 5.3(d) for fixed powers of
10, 30, and 100 TW is given by
(5.8)
If we now consider a simple "rocket" model of ablation, a stationary flow

model of the ablation process determines the ablation rate. From a paper
by Max et al. (1980), the ablation rate m (g/cm2/sec) appropriate in the
presence of inhibited transport is given by
</> )0.48
riz = 2.3 X 1if ( 1014 ~/cm2
R )1.89 ( 1 )0.82
x ( 0.1 em :Lm g/cm2/sec (5.9)
The burn-through depth is then given by
(5.10)
In a simple "rocket" model of ablation, the implosion efficiency has a

local maximum when the "payload" is approximately 20% of the fuel mass.
The locus of optimal thickness of the shell for each irradiance (w = 1.2 Lb)
using Eqs. (5.9) and (5.10) is indicated by the shaded regions in Figs. 5.3(a)
through 5.3(d). In order to operate at optimal hydrodynamic efficiency, it
is desirable to choose a wall thickness near this locus. In order to minimize
the energy requirements of the driver system, while maintaining the lowest
acceptable aspect ratio for hydrodynamic stability purposes, it is shown in
Fig. 5.3 that irradiances from 3 x 1014 W/cm2 to 3 x 1015 W/cm2 are required
for laser powers of the order of 30 TW, for pulse lengths between 1 to 2 nsec,
and energies of the order of 30 to 100 kJ.
Shown for reference in Figs. 5.3(a) to 5.3(d) are the hot-electron stop-
ping distances ( 4 Le) and the optimum shell thickness (1.2 Lb) for 0.35 and
1.05 11m light. Note that the longer wavelength laser pulse requires the use
of lower irradiances (or much larger energies) and higher powers compared
with 0.35 11m light. Note that the hydrodynamic stability problems, as mea-
sured by aspect ratio, are of considerably greater concern for the longer
wavelength drivers.
5.3 CORONAL PHYSICS
At intensities in excess of 1011 W/cm2 , the irradiated areas of the target

undergo rapid ionization and a plasma is formed with density below solid
density for the material. The laser beam energy is absorbed in this coronal
plasma at densities equal to or less than the critical density for the laser
wavelength employed. The critical density is proportional to the inverse of
the square of the laser wavelength and is approximately 1021 electrons/cm 3
for a laser wavelength of l~tm.
Understanding the physics of the corona is important because various
linear and nonlinear laser-plasma interaction processes take place in the
corona; these processes were responsible for the hot electrons present in
early laser-fusion experiments.
Efficient absorption of the laser light by the target is a necessary condi-
tion for the success of laser fusion. Measurements by Seka et al. (1982),
Mead et al. (1981), Garban-Labaune et al. (1982), and Nishimura et al.
(1981) have shown that absorption is significantly higher at short wave-
lengths (see Fig. 5.4). This result is due to the fact that the critical den-
sity increases as the laser wavelength is reduced; at higher plasma densities,
the electron-ion collision rate is higher, leading to more efficient absorp-
tion. Furthermore, the short-wavelength laser experiments show that the
absorbed energy is found predominately in a thermal electron distribution,
whereas at long wavelengths, a large fraction of absorbed laser energy is
converted to suprathermal electrons (see, e.g., Slivinsky et al., 1975; Pried-
horsky et al., 1981; Yaakobi et al., 1976; Hares et al., 1979; and Keck et al.,
1981 ). Predictions based on modeling of the absorption fraction in both
planar and spherical geometries are in reasonable agreement with the mea-
surements.
The underdense coronal region is shown schematically in Fig. 5.5. Some
of the processes responsible for generating supra thermal electrons are iden-
tified in this drawing. Resonance absorption, e.g., occurs near the criti-
c
0
c......
0
~ 40
<t:
20
Laser Intensity (W/cm2)
Figure 5.4 Measured absorption as a function of laser intensity for a number of

laser wavelengths. The data is a summary of results of experiments conducted at
Lawrence Livermore National Laboratory, Los Alamos National Laboratory, the
Laboratory for Laser Energetics, and Ecole Polytechnique.
cal density region and is important for laser wavelengths 1 J.tm or greater
at intensities of interest to laser fusion. Nonlinear parametric interac-
tions that involve the generation of plasma waves, such as the two-plasmon
(2wp) decay instability occurring at the quarter-critical density region, the
parametric-decay instability occurring near critical density, and the stim-
ulated Raman scattering instability (SRS) occurring at densities equal to
or less than quarter-critical density, may also lead to the generation of the
undesired energetic electrons (see, e.g., Kruer, 1981).
SRS is the decay of an incident electromagnetic wave into a scattered
wave and an electron wave. The absolute Raman instability (SRS-A) oc-
curs at a density equal to a quarter of the critical density nc/4 for the incident
laser light. The scattered electromagnetic wave, therefore, has a wavelength
equal to twice that of the laser wavelength. The convective Raman instabil-
ity (SRS-C) occurs at densities lower than nc/4 and results in scattered light
with wavelength shorter than twice that of the fundamental.
Typical spectra obtained with 0.35 ,_,m laser irradiation by Tanaka et al.
(1982), using CH targets, are shown in Fig. 5.6. Two distinct spectral com-
ponents can be observed in the scattered light spectrum at an intensity of
I I
2wp decay resonance
absorption
: SRS
f
SBS
Figure 5.5 Schematic of the coronal plasma region of a laser heated plasma show-
ing various parametric processes and scattered light signatures.
100
....
::1
~
>
1-
u; 10
zw
1-
z
z
ca:
:IE
ca:
a:
1 I
WAVELENGTH (Mm)
Figure 5.6 SRS-A spectra from CH targets. (a) (h = 4.4 x 1014 W/cm2 , (b) (h =
1.2 x 1015 W/cm 2 . The spectra were recorded on Kodak high-speed IR film. The
curves shown have been corrected for film response [from Tanaka et al. (1982)].
1.2 x 1015 W/cm2 ( 450-psec pulses). The narrow line at 0.7 pm corresponds
to wL/2 emission generated at nc/4 by SRS-A. The broad feature from 0.4
to 0.6 pm has been attributed to SRS-C. The width of the spectrum below
0.7 pm indicates that Raman scattering occurs over a considerable region
of the underdense plasma (the plasma density corresponding to 0.4 pm is
1.5 x 10- 2 nc)·
The dependence of the Raman signal on intensity measured in these ex-
periments illustrates the threshold nature of the instability, as shown in
Fig. 5.7. The observed SRS signal grows exponentially as exp(¢>d1015 ),
where ¢>L is the incident laser intensity in W/cm 2 , until saturation occurs.
Both the absolute and convective instabilities showed a saturation in these
••~~~;- ;~:~ t~ ;;~;>ii:::,f(:_~·

./:': [J~?~~~~~op-/B7~b
.~
/-ii-1,-~-c
AVERAGE INTENSITY (1014 W/cm 2)
Figure 5.7 Intensity dependence of SRS. Curves (a) and (b) are absolute backscat-
ter measurements using calibrated photodiodes at 700 nm. The vertical axis repre-
sents the peak fluence in the backscatter light normalized to the incident energy
in J/(J ·sr). Curve (a) is for scattered light polarized parallel to the incident laser,
curve (b) for perpendicular polarization. Curves (c) and (d) are curves similar to (a)
and (b) obtained from spectrographic recordings at 700 and 500 nm respectively.
The vertical axes for these two curves are not absolutely calibrated. Curves (c)
and (d) correspond to backscattering from the convective (SRS-C) and absolute
(SRS-A) Raman instabilities, respectively [from Tanaka et al. (1982)].
experiments at an intensity roughly 1.5 times the threshold intensity. The

total energy backreftected by the Raman process in these measurements
was estimated to be 10- 6 of incident energy. For higher Z targets (Ni and
Au), the threshold intensities are twice as high as those for CH targets. This
result is consistent with a lower-density scale length for the higher Z targets.
Another process that is potentially detrimental to laser fusion is energy
loss through stimulated Brillouin scattering (SBS). Experiments with 1 f.i.m
lasers show significant (:2: 40%) losses of laser light to SBS (see, e.g., Rosen
et al., 1979); however, experiments conducted with short-wavelength laser
pulses show no substantial SBS losses even when the threshold for this in-
stability has been exceeded (Tanaka et al., 1985).
Theoretical research on the spectral characteristics and intensity thresh-
old of this radiation has provided indirect information on the suprathermal-
electron distribution in the corona. As was shown by Simon and Short
(1984) and Simon et al. (1986), enhanced Thomson scattering can some-
times be mistaken for a parametric instability.
1053 nm 351 nm
16 15
l I
15
---
14
Thermal-Electron
14 Thermal- Electron 13 -Component
Component
13
:;- :;- 12
.r:: .r::
"0
w 12 "0
_ _ Hot-Electron w
~
Component
"0 11
Cl 11 Cl
0 0
...I ...I 10
10
Superhot-Eiectron
Superhot-Eiectron
Component 9
9 Component
/
8
/
8
7 7
0 100 200 300 0 100 200
(a) hv (keV) (b) hv (keV)
Figure 5.8 X-ray continuum spectra obtained from spherical targets with (a) the
24-beam OMEGA system at A = 1.053 flm solid CHand (b) the six-beam OMEGA
at A= 0.351 J.Lm [from Richardson et al. (1985)].
High-energy electrons in laser-matter interaction experiments are usu-

ally characterized through x-ray spectra, such as those shown in Fig. 5.8,
obtained with multichannel broadband x-ray detectors by Richardson et al.
(1985). A typical spectrum obtained with 1.053 J.Lm irradiation [Fig. 5.8(a)]
shows a three-temperature distribution: a thermal component around a few
keY, a hot component around 1~ keY generated by processes such as reso-
nant absorption, and a superhot component, which can be as high as several
hundred keY, usually generated by parametric interaction processes. With
0.351 J.Lm irradiation [Fig. 5.8(b)], the hot electron component nearly van-
ishes since the dominant. absorption mechanism is inverse bremsstrahlung
rather than resonant absorption.
The fractional conversion into hot and superhot electrons depends on
the wavelength, laser intensity, and plasma scale length; shown in Fig. 5.9
are the results of 1 and 0.35 J.Lm irradiation experiments conducted with the
2500 J OMEGA laser (at the Laboratory for Laser Energetics), showing sig-
nificantly lower total suprathermal (hot and superhot) electron production
100 10-7
10" 1 "~6s~~~
~ ~"- ::\:l:f~t"EI~c)rQn~" 10-8
1o·2 10-9
+-+--+
u
~
u
c 3w/2 c
...:..-
...... 10"3 10·10 u:r
......
.. /~
uGl
.....
N
3
"ii w
M
w
10"4 10·11
1o-s
flJ:-~ • 351-nm Superhot Elect•ono
o 1053-nm Supe•hol Elect•ono
1o-12
10·6 10·13
0 5 10 15 20
Intensity (101 4 W/cm2)
Figure 5.9 Suprathermal electron conversion efficiency for 1.053- and 0.351 p.m
spherical target experiments. For comparison, the 3w/2 harmonic conversion is also
shown, suggesting that the superhot electrons are generated by the 2wP instability
[from Richardson et al. (1985)].
at the shorter wavelength. This reduced production of high-energy elec-

trons with short-wavelength irradiation is partly due to the increased col-
lisional absorption that lowers the effective laser intensity near the critical
density. In recent experiments on the NOVETTE laser at Lawrence Liver-
more National Laboratory, collisional damping of the parametric instabili-
ties was demonstrated in long-scale-length, high-Z plasmas irradiated with
0.26-pm radiation (Turner et al., 1985). Figure 5.10 shows a summary of
some of the hot electron generation experiments. Note that for the intensity
range of interest to laser fusion (1014-1015 W/cm2 ), the fractional conver-
sion of energy to hot electrons is less than a few percent for laser wave-
lengths less than 0.53 pm.
In summary, the most important result of the coronal-physics experi-
ments is that with laser wavelengths less than 0.53 pm, it is possible to obtain
greater than 50% absorption at intensities up to 1015 W/cm2 • Furthermore,
although many parametric interaction processes are observed, the conver-
e Au@ 10.6 pm
0 Au and Be@ 1.06 pm
& Au and Be@ 0.53 pm
0 CH @ 0.35 pm
Figure 5.10 Suprathermal electron energy dependence on ¢>L>.£ [W/cm 2 (Jlm) 2 ]

[from Mead et al., (1981); Ripin et al., (1980); Keck et al., (1981); and Priedhorsky
et al., (1981 )].
sian of laser energy into suprathermal electrons appears to saturate at a

low value (less than a few percent of laser energy), implying tolerable pre-
heat levels for compression experiments. As greater laser energies become
available, making possible experiments with greater-intensity scale-length
products, it will be important to confirm the favorable scaling (saturation)
observed to date.
5.4 X-RAY GENERATION BY

LASER-PRODUCED PLASMAS
X-ray emission from laser-produced plasmas is not only of considerable in-

terest for inertial-confinement fusion applications, but is also of importance
to a number of other applications.
The increased absorption and reduced supra thermal electron generation
observed with short-wavelength laser radiation has also led to increased
efficiency of conversion of laser light to x-rays in short-wavelength laser-
matter irradiation experiments. In fact, theoretical calculations by Mead
et al. (1987) indicate that the efficiency of conversion of absorbed laser light
to soft x-rays ( < 1 keV) should vary as >..z 2 • At short laser wavelengths, the
increased conversion efficiency is due to the higher level of collisional ab-
sorption and the higher coronal electron density. The physical reason for
Ex
-=---.!!...- (%)
Elaser
SL-----~--------L-----~
1014 3 x 101 4 1015 3 X 1015
Incident Intensity (W/cm 2)
Figure 5.11 Summary of x-ray conversion measurements. E/ELaser is the conver-
sion of laser energy on-target to sub-kilovolt x-ray energy. These results for gold
targets are from the Laboratory for Laser Energetics, Lawrence Livermore National
Laboratory, and Los Alamos National Laboratory.
the increased conversion at low intensities is that under these conditions,

the plasma can radiate the energy fast enough to maintain a relatively cool
coronal absorption region with a long-scale-length absorption and emis-
sion region. At high laser intensities, radiational processes cannot cool the
plasma fast enough and, consequently, much of the energy is "lost" in the
low-density plasma blowing off the target and in energy transport to the
denser regions of the target.
A number of experiments have been conducted to characterize the soft
x-ray emission properties of laser-produced plasmas (see, e.g., Rosen et al.,
1979; Mead et al., 1983; Goldstone et al., 1984; Kauffman et al., 1984;
Mochizuki et al., 1986). In general, these experiments have confirmed the
theoretical expectations of increased x-ray conversion with decreasing laser
wavelength and decreasing laser intensity. A summary of the recent exper-
imental results is given in Fig. 5.11.
In some of the recent experiments (see, e.g., LLE Annual Report, 1985),
the spectral and temporal characteristics of the x-ray emission were mea-
sured. In these experiments, it was shown that at high laser intensities,
((h ~ 1015 W/cm 2 ), the coronal temperature is high enough to lead to in-
creased ionization and increased emission from the M lines (2.5 ke V en-
ergy) on Au targets. Shown in Fig. 5.12 are spectra from Au-coated CH
targets at low (4 x 1013 W/cm2 ) and high intensity ( 4 x 1014 W/cm 2 ), show-
5 24
M-lines
~
c
~
"i: 20
4
.
::I
>-
ca
.:::
..
::I
>-
ca 16
:c.. 3 .:::
:c...
~ ~
12
> 2 >
!:: 1-
8
en iii
zw zw
1-
z
1- 4
z
0 0
5.0 5.5 6.0 6.5 4.5 5.0 5.5 6.0 6.5
(a) WAVELENGTH (A) (b) WAVELENGTH (A)
Figure 5.12 Onset of strong M -line emission (5-6 A) is predicted by LASNEX to

occur at intensities around 4 x 1014 W/cm 2 and is found experimentally on 0.351~tm
irradiation experiments at (a) 4 x 1013 W/cm 2 and (b) 4 x 1014 W/cm 2 [from LLE
Annual Report (1985)].
~
Fig. 5.13 Details of the time-resolved x-ray emission spectrum from spherical Au coated CH targets irradiated with 0.351 JLm light at
an intensity of 4 x 1014 W/cm2 . The Au coating thickness is indicated as f>.r(Au) [from LLE Annual Report (1985)].
ing the onset ofM-line emission at higher laser intensities. High-resolution

time-resolved spectra taken with the LLE streaked elliptical-crystal x-ray
spectrograph (SPEAXS) are shown in Fig. 5.13; they also show the presence
ofM-line emission at high laser intensity. Such data are useful in developing
an understanding of the dynamics of high-Z laser-produced plasmas.
The x-ray conversion data obtained at a number of laboratories may be
summarized as follows:
It is possible to obtain greater than 50% conversion of the absorbed laser
1.
light into soft x-rays (:<:; 1 keY), provided that <PL >..I is less than 10 14 W/cm 2 -
pm2.
2. At an intensity higher than 1016 W/cm 2 , the fractional x-ray conversion is
less than 25% for wavelengths greater than 0.26 pm.
5.5 LASER-DRIVEN ABLATION
The overall efficiency of the laser-fusion process is a time-ordered product

of efficiencies starting with the wall-plug to laser energy and the laser-target-
coupling efficiencies. The laser energy absorbed in the low-density corona
must be efficiently transported to the high-density ablation region of an im-
ploding target in order to generate sufficiently high pressures to compress
the target to the required densities. The target hydrodynamic efficiency is a
measure of the efficiency with which the absorbed laser energy is converted
to hydrodynamic energy in the imploding portion of the target.
For optimum momentum transfer to the imploding shell, it is preferable
to ablate a large mass at a low velocity rather than a small mass at a high
velocity. Short-wavelength laser irradiation is important in this regard as
it results in energy deposition at relatively high densities, leading to higher
ablation rates and therefore higher overall hydrodynamic efficiency than
would be possible with a long-wavelength laser.
McCrory and Morse (1977a) showed that the hydrodynamic efficiency
TJH should scale as
(5.11)
where P = 1.2 for flux-inhibited thermal energy transport and P = 0.5 for
uninhibited transport. Inhibition of energy flow in laser-driven targets oc-
curs because the bulk of the energy is carried by electrons with a mean free
path comparable to (or greater than) the gradient scale lengths in the con-
duction zone, so that the classical theory of conduction of Spitzer and Harm
(1953) is not applicable.
Figure 5.14(a) describes schematically the processes and conditions in
the target that pertain to the thermal electron transport. The absorbed laser
Ablation - - Heat
Surface Front - Corona
-------Te
Laser
Light
I
I
I
--- - - Critical Surface
I
I
I
rc
Radial Direction
(a)
Figure 5.14 Schematic diagrams showing: (a) energy transport in direct-drive laser
fusion targets. In this drawing rc defines the position of the critical density surface,
T. is the electron temperature and n. is the electron density. The energy deposited
by the laser in the corona is transported in the heat front to the ablation region; and
(b) temperature profile for a direct drive laser fusion target. The heat flux tends to
be classical near the ablation radius and flux limited near the critical radius.
energy creates a heat front that progresses through the cold target material;
the corona is almost isothermal because of the high thermal conductivity
and the low density of the blowoff plasma.
Thermal electron transport directly affects the hydrodynamic efficiency
because a reduction in the heat flow from the corona to the ablation region
will reduce the drive pressure and overall hydrodynamic efficiency. Any
thermal energy not transported to the ablation region is wasted in acceler-
ating the coronal plasma. There are also secondary effects of the thermal
electron transport on the absorption and the x-ray emission efficiency of a
laser-driven target.
Because of the significance of the ablation process to overall implosion
energetics, a significant theoretical effort has been devoted to increasing
the understanding of the thermal transport (see, e.g., Bell et al., 1981; Ma-
Flux-Limited Conduction Region
~~--~-~---
/ I i /I
I I y
II }/:
:
/ I
./
I
Classical
Conduction
Region
Radial Direction
(b)
son, 1981; Shvarts et al., 1981; Matte and Virmont, 1982; Albritton, 1983;
Matte et al., 1984; and Delettrez, 1986). Although the understanding of
this process is not yet complete, recent experimental and theoretical efforts
have made significant progress in understanding and characterizing thermal
electron transport in laser-produced plasmas.
Code results show that the flux limiting occurs in the heat front in a re-
gion between the critical surface and about halfway down the heat front [see
Fig. 5.14(b)]. This flux-limiting region controls the amount of heat flowing
into the heat front. The energy that does not enter the heat front raises
the temperature of the plasma in the corona and is dissipated as kinetic en-
ergy in the blowoff plasma. Therefore, it is necessary to know the "correct"
value of the flux limiter in order to predict with confidence reactor-size tar-
get designs, unless another model (see Luciani et al., 1983) is found that is
numerically efficient and describes the heat flux more accurately than flux-
limited diffusion.
5.5.1 Experiments in Thermal

Electron Energy Transport
The electron thermal flux cannot be measured directly but must be inferred
from measured quantities. The quantity of interest that is dependent on the
thermal transport is the mass ablation rate rit in g/(cm2 -sec). The efficiency
of the target drive is directly connected to the mass ablation rate that can be
scaled with the flux limiter in steady-state models. The mass ablation rate
can be obtained from two measurements: the burn-through of the heat front
and the ion velocity spectrum. Burn-through depths are usually measured
in a material in which signature layers are introduced that radiate to signal
the passage of an isotherm of the heat front. The isotherm temperature
depends on the energy of the observed lines emitted by the material. The
mass ablation rate is then calculated from
. p6x
m=-- (5.12)
D..t
where p is the material solid density, 6x the burn-through depth, and D..t
the time interval during which burn-through is occurring. Ion velocity spec-
tra are obtained from charge collector data: Charge collectors measure ion
currents as a function of time. To calculate the mass ablation rate, we as-
sume that the kinetic energy of the ablated mass balances the rate of energy
absorbed
(5.13)
where m is the mass areal density (g/cm 2 ) and v the ablation velocity. As-
suming nearly constant ablation velocity, we can obtain the mass ablation
rate
(5.14)
where EA is the total energy absorbed and ( v2 ) the average of v2 over the
ion spectrum. EA must be corrected for radiation losses and rh for time-
dependent effects such as the temporal shape of the incident pulse. The
measured quantities are compared with the results of hydrodynamic code
simulations with the flux limiter as a parameter.
A less direct way of determining the thermal flux is by simulating implo-
sion experiments with hydrodynamic codes and determining what value of
flux limiter f (f is a dimensionless number, with a value less than unity mul-
tiplying the heat flux carried by electrons free-streaming into a vacuum) is
needed for agreement. The main diagnostics affected by f are the absorp-
tion fraction, implosion time, neutron yield, core and pusher areal densities,
and x-ray imaging. Because these provide indirect measurement of the heat
flux, they can be affected by factors other than the rate of heat transport into
the ablation region, such as irradiation nonuniformities or radiation trans-
port. Therefore, inferences from code simulations must be treated with
care. In fact, as will be discussed later in this section, even direct experi-
ments such as burn-through experiments are affected by factors other than
heat-flux limitation.
The first estimates of thermal transport were obtained from measure-
ment of the fraction of 10 pm laser light transmitted through a thin foil, the
fast ions produced, and the hard x-ray spectrum (see, e.g., Malone et al.,
1975). Other experiments at 10 and 1 pm followed in which hard x-ray
spectra (Young et al., 1977), fast ion spectra (Campbell et al., 1977), and
burn-through depth into a solid (Yaakobi and Bristow, 1977) were com-
pared to code simulations. The heat fluxes inferred were found to be a
small fraction of the free-streaming limit with f varying from 0.03 to 0.1.
However, suprathermal electrons produced by resonant absorption dom-
inate the energy flow at the high intensities and laser wavelengths (1 to
10 pm) used in these experiments and the codes used in the simulations
did not model the fast electron transport. More recent experiments on pla-
nar targets backed by improved modeling have produced conflicting results.
In the first transport experiments at a short laser wavelength (0.35 pm) by
Yaakobi et al. (1981a), the bum-through depth into a plastic-coated glass
substrate was measured over the intensity range 1013-1014 W/cm 2 • Simu-
lations were carried out with a one-dimensional spherical code with a ra-
dius equal to the laser-spot diameter to approximate the two-dimensional
effect of the blowoff. Good agreement with the experimental results was
obtained for f = 0.04. Another set of experiments was carried out at 1.06
and 0.35 pm wavelengths in which the bum-through into a beryllium-coated
aluminum substrate was measured by Mead et al. (1984). Also measured
was the lateral transport. Extensive analysis with one- and two-dimensional
hydrocodes showed that lateral transport was reduced by the presence of
magnetic fields and that the axial transport was best simulated with/ = 0.01
to 0.03 for both laser wavelengths. These values for the flux limiter are
significantly lower than those inferred by Yaakobi and Bristow (1977). This
difference could not be explained by the different methods used in the codes
to calculate the effective thermal flux.
There are serious limitations to using results of planar transport experi-
ments in spherical code simulations and target design. Unless the laser spot
is very large (larger than 500 pm for laser energies of the order of 5 kJ and
pulse lengths of a few nsec ), edge effects will affect the results in many ways.
It was shown by Stamper et al. (1971) and Craxton and Haines (1975) that
magnetic fields produced by the Vn x VTe (where n is the electron density)
sources present near the edge of the beam will reduce the lateral thermal
transport. The divergence of the plasma flow is different: In planar exper-
iments, the effective radius of curvature is set by the laser-spot diameter.
Finally, there will be two types of lateral energy loss: The first, by thermal
conduction, will be small, whereas the second results from the decoupling
of the conduction electrons (with energy two to three times thermal energy)
at the critical surface and below. These electrons, which are collisionless,
will orbit around the azimuthal magnetic field and be deposited in the cold
part of the target (see, e.g., Yates et al., 1982). The resulting reduction in
the populations of conducting electrons will result in a lower axial thermal
flux.
Recently, the availability of more powerful lasers and multibeam lasers
has permitted transport experiments in spherical geometry. One series
of experiments was conducted at the Rutherford Laboratory by Goldsack
et al. (1982), with a six-beam laser system at two laser wavelengths, 1.05
and 0.53 J.Lm. The experiments were carried out at low enough intensi-
ties [4>A .\i : :; 1014 W/cm2 -(J.Lm) 2 where 4>A is the absorbed intensity] so that
suprathermal electrons should not affect the energy transport. The targets
were glass spheres or empty glass shells with a layer of plastic coated with
a thin layer of aluminum. The mass ablation rate was measured from ion
calorimeter traces and from the burn-through time, defined as the time be-
tween the fiducial emission of the AI XII 1s-3p line and the emission of the
Si XIII 1s-2p line. The measured burn-through times of about 600 psec
are smaller than the laser pulse lengths of 1.3 and 1.0 nsec at the two laser
wavelengths so that the mass ablation rate measured is actually the time-
averaged mass ablation rate during the rising part of the laser pulse. A cor-
rection factor was introduced by the authors to take that fact into consider-
ation: They assumed that the scaling of th during the laser pulse is [4>A (t )] 0·7
and averaged over the measured temporal shape of the laser pulse. A sim-
ilar correction was introduced in the calculation of the mass ablation rate
from the charge collector data. From the least-squares fit to the ion data,
a scaling law forth as a function of the absorbed intensity was obtained for
0.53 and 1.05 J.Lm, respectively
4> ) 0.72±0.03
th = 0.85 ± 0.03 x 105 ( 1 ~ 3 g/cm2 - sec (5.15)
and
4> ) 0.86±0.03
th = 0.81 ± 0.03 x 105 ( 1 ~ 3 g/cm 2-sec (5.16)
Mass ablation rates were obtained from the x-ray measurements for a few
shots at high intensities and for .\L = 0.53 J.Lm and show agreement with the
ion data. Code simulations were carried out with a one-dimensional hy-
drocode that included ray tracing in the modeling of the absorption of the
laser light (but not radiation transport). Conclusions were that the experi-
mental mass ablation rate is larger than that given by the simulation with
f = 0.1; m was insensitive to f for f > 0.1.
Three other series of experiments were carried out at the laser wave-
length of 1.05 f-tm. At LLE, parylene-coated targets on substrates of differ-
ent materials were illuminated with the 24-beam OMEGA laser system by
Yaakobi .et al. (1984). The mass ablation rate was obtained both from the
charge-collector traces and the burn-through depth into the plastic as mea-
sured by the emission of particular lines of the substrate material. The sub-
strate materials were Al, Ni, and Ti that gave signatures for the penetration
of different isotherms: Al for temperatures of about 300 to 500 eV and Ti
and Ni for temperatures near 1 keY and 1.5 keY, respectively. The tempo-
ral progress of the heat front was also observed with an x-ray streak camera.
Preheat was measured from Ka line emission from the substrate. Results
show a much larger penetration into anAl substrate with uniform illumina-
tion than into a Ti or Ni substrate, or into anAl substrate with single-beam
illumination (see Fig. 5.15). Code simulations predict a penetration depth
about three times smaller than that observed for f = 0.04 and comparable
for f = 0.65. But simulations with/ = 0.04 predict well the burn-through in
cases with large-Z substrates and with single-beam illumination, as in the
planar experiments discussed previously. The large penetration depth to
the Al substrate cannot be explained solely by the presence of hot spots in
the illumination pattern because it would require that the hot-spot intensity
be nine times the average intensity; such intense hot spots have not been ob-
tained in nonuniformity calculations of the illumination or observed in the
x-ray microscope images of the target. One interpretation of these results
is that the heat front is preceded by a low-temperature (less than 500 eV)
"foot" in the uniform irradiation experiment and that this foot does not exist
in the single-beam experiments because the long mean-free-path electrons
responsible for the larger penetration are lost to the cold material surround-
ing the spot. Figure 5.16 compares the mass ablation rates obtained from
the LLE experiment using various diagnostics and those measured by the
Rutherford Laboratory. At the higher intensities, the mass ablation rate
scales as ¢>~ 5 as compared with the ¢>~ 86 scaling obtained at lower intensi-
ties. Although this observation suggests a break in the scaling curve (see,
e.g., Rosen 1984), it must be remembered that the break occurs between
two sets of results from separate laboratories with different laser systems
and diagnostics. The conclusion would be more acceptable if the break oc-
curred in results by one laboratory in a single series of experiments.
Another series of experiments was conducted at KMS by Tarvin et al.
(1983), with a two-beam clamshell-focusing laser system at 1.05 f-tm. The
KMS experiments were similar to those at Rutherford: The mass ablation
rate was obtained by measuring the burn-through times through a plastic
layer from x-ray signals generated by signature layers. Conditions were
w
!ca:
l-
en
ID
::)
en
:&
0
a:
1&.
z
0
~
Q
cc
a:
0
~ AI
-
:!! Single
0
a: Beam
J:
0
0
z
0
:&
2 4 6 8 10
CH OVERCOAT THICKNESS (J.Lm)
Figure 5.15 Bum-through of parylene coating into various substrate materials in

uniform and nonuniform irradiation of spherical targets with 1.05 J,Lm laser light.
For aluminum substrates the intensi7, of the Lyman-a x-ray line of Al+ 12 and for
titanium and nickel substrates the 2 P-1 1S resonance line of the helium-like ion
(Ti+ 20 at 4.75 keY; N;t" 26 at 7.8 keY) is plotted.
somewhat different from those for the other experiments in that the laser
pulse had a fast rise-time and remained almost constant for about 800 psec
and the targets were smaller, giving rise to smaller gradients. Comparison
with the model of Max et al. (1980) gave a best fit for f = 0.06; this is equiv-
alent to a larger value off obtained from codes that use the harmonic mean
to calculate the effective thermal flux. In a different type of measurement
in the KMS group (Fechner et al., 1984), the electron density profile in the
corona and the excursion of the critical surfaces were measured using in-
terferometry. The density-scale length and excursion of the critical density
are sensitive to the mass ablation rate. Comparison with code simulations
in which a sharp cutoff was used to calculated the effective heat flux showed
that the data were consistent with a flux limiter of 0.03 to 0.06. The mass
Single Beam
6. - LLE (0.35 ~-tm) e
6. • X
+
...'u....... • 6.
• X
X
Cll 0
Ul
<1' 0
l
E 105 oo
x
$
(J Spherical Irradiation
~ o - Rutherford (1.05 J,.tm)
·E 00
0 • - spectra
+ t k LLE
0 - s rea (1.05 J,.tm)
x - collector
1o;o~12~-----------1~o~1a~-----------1Jo~14~----------~1o1s
Absorbed Intensity (W/cm 2)
Figure 5.16 Comparison of mass ablation rates as a function of absorbed intensity.

Results marked by open circles were obtained at the Rutherford Laboratory. The
other data were obtained at LLE with various diagnostics: solid circles from burn-
through curves for Al substrates, triangles from Ti and Ni substrates, crosses from
streak traces, and X's from charge collector signals.
ablation rates deduced from these experiments were close to the values ob-
tained from the charge-collector data shown in Fig. 5.16.
A more recent experiment at the Rutherford Laboratory by Hauer et al.
(1984) used l.0611m laser irradiation and targets that consisted of solid glass
spheres coated with three layers: CH, AI, and CH. The thickness of the CH
layers were such that the burn-through time between the AI layer and the
glass core bracketed the peak of the laser pulse so that the mass ablation rate
was measured at the nominal intensity. This experiment, a joint effort with
the Los Alamos National Laboratory, was simulated with the hydrocode
LASNEX; good agreement was obtained for f = 0.08 ± 0.02. The mass ab-
lation rates deduced overlap the KMS data and the LLE charge-collector
data. Temporal observation of the onset of Si and Ca lines indicates that
the heat front must be steep.
No single conclusion can be drawn from the thermal transport experi-
ments carried out at 1.06 11m. The results obtained at LLE (the large burn-
through depths that can only be modeled with/ = 0.65) were not observed
at the other laboratories, although the first Rutherford series also required
thermal fluxes larger than could be provided by uninhibited thermal trans-
port. Since no measurements were made of the steepness of the front as in
Hauer et al. (1984) and no substrates other than glass were used, one cannot
rule out the possibility that the two series of experiments could have shown
the same behavior. The latest Rutherford results confirm the presence of
a steep temperature profile in contrast to the LLE observation. The small
value of the flux limiter required by KMS to replicate their experimental
results could be due to their use of smaller targets, which would lead to
steeper gradients and thus require smaller flux limiters. If one excludes the
first Rutherford experiment and the LLE spectroscopic data for the burn-
through, most experiments obtain the same scaling for the mass ablation
rate in the range¢A = 0.2 x 1014 to 4 x 1014 W/cm2 , with rh "'<P'}l, which can
be matched by simulations with a sharp cutoff flux limiter between 0.06 and
0.08; in two experiments, the uninhibited heat flux could not simulate the
:.\ = 351 nm
1.00
z
0
u;
rn
j
w
w 0.10
z
~
...+
N
C(
w
>
j:
c(
...1
w
a: 0.01
0 2 4 6 8 10
CH OVERCOAT THICKNESS (!Lm)
Figure 5.17 Experimental burn-through curves for six-beam irradiation of spher-

ical targets by 0.35-JLm laser light; intensity (in W/cm2 ): A-1014, B-5 x 1014 ,
C-1015 • Simulation curves are shown in shaded bands where the left boundary cor-
responds to a flux-limiter f = 0.04 and the right boundary to f = 0.1. Theoretical
bands a, b, c are for the same intensities as A, B, C, respectively.
measured mass ablation rates. The values of the flux limiter are larger than
the value of 0.03 obtained from planar target experiments. This points to
basic differences between the two types of experiments: they may be due
to the presence of magnetic fields or the fact that the electrons responsible
for part of the heat transport are collisionless and are convected out of the
laser spot to the cold surrounding material.
A first series of spherical target experiments at a laser wavelength of
0.35 pm were carried out at LLE by Yaakobi et al. (1985) with six beams.
The experiments were similar to the previous one at LLE at 1.05 pm. Fig-
ure 5.17 shows the measured and calculated AI+ 12 line emission for sev-
eral laser intensities as a function of the CH-overcoat thickness. The burn-
through depths are very large and cannot be modeled by code simulation.
At 1014 W/cm2 , where flux limiting has no effect, there is a factor of 2.5
between the observed and calculated burn-through depths. Such large dis-
crepancies could be caused by hot spots in the beam that may lead to large
penetration depths. Another difference with the 1.05 pm data is that there
are now no differences in the burn-through of CH between uniform or
single-beam illumination and in the burn-through of different isotherms as
measured by different x-ray lines from a titanium substrate. The mass ab-
lation rate obtained from the charge-collector data is larger than that mea-
sured at 1.05 pm by about a factor of 2 for the same absorbed intensity, but
is lower than that obtained from spectroscopy. The mass ablation rate for
0.35 pm irradiation deduced from charge collectors follows the scaling law
rh = 3.8 x lOS(¢>A/1014 ) 0·53 (g!cm2 sec). As in the case of the spectroscopic
data, code simulations at f = 0.1 could not match the larger values obtained
from the charge-collector data.
5.5.2 Modeling of Thermal

Electron Energy Transport
In the hydrodynamic codes used in target designs and experiment simula-

tions, the thermal electron transport is modeled using flux-limited diffusion.
Despite the fact that it has little physical basis, the flux-limited diffusion
model provides a fast and numerically efficient method of calculating the
electron temperature in the plasma. The flux limiter is an ad hoc parameter,
a "knob," whose value needs to be pinned down because it affects the perfor-
mance of targets. Theoretical efforts in the modeling of thermal transport
beyond the flux-limited model can be divided into two groups that are not
mutually exclusive: one group that has attempted to explain why the value
of the flux limiter, as deduced from experiments, is only a small fraction of
the free-streaming value and another that has addressed the solution of the
Fokker-Planck equation in order to understand better thermal transport in
the steep temperature gradients found in laser fusion. One of the goals of
the second group is to find a fast and efficient solution to the Fokker-Planck
equation that can be introduced into the hydrodynamic codes. A survey of
the theoretical work is beyond the scope of this chapter.
5.5.3 Ablation Pressure
Measurements of the ablation pressure generated by the ablating plasma

have also been carried out at a number of laboratories. These experiments
have again confirmed the expected higher ablation pressure with shorter-
wavelength laser irradiation.
We recall from Fig. 5.10 that the fast electron preheat remains at accept-
ably low levels provided that tjJL>..l:::; 1014 W/cm2 -Jl-m2 • In Fig. 5.18, we plot
a summary of ablation pressure measurements as a function of laser wave-
100
-.:-
•
~ 40
!
i
D.
c
0
:;
:a
<( 10
1.5 2.0
Laser Wavelength (~tm)
Figure 5.18 Ablation pressure as a function of laser wavelength in laser-matter

interaction experiments (from Yaakobi et al. (1981), Ripin et al. (1980), Kilkenny
(1981), and Maaswinkel et al. (1984)].
length for a constantvalueof ¢>L>.L = 1014 W/cm2 -t-tm2 (fromLLE, Ruther-

ford, and the Max Planck Institute). We note that only for >.L :::; 0.5 t-tm is
the ablation pressure sufficiently high (30 to 100 Mbar) to effectively drive
targets without substantial fast electron preheat.
5.6 HYDRODYNAMIC STABILI1Y

OF ABLATIVELY DRIVEN SHELLS
High-performance, single-shell, laser-driven targets (see Sec. 5.2) gener-

ally require high-aspect-ratio shells. Unfortunately, Rayleigh-Taylor (R-T)
instability occurs in these systems at (1) the ablation surface when the ac-
celeration is in the same direction as the local density gradient, and (2) later
in the implosion process when the fuel decelerates the pusher.
R- T instability is the instability that occurs at the interface of a heav-
ier fluid superposed over a lighter fluid (see, e.g., Chandrasekhar, 1981).
A modified version of this instability occurs in spherical implosion systems
in inertial confinement fusion targets, with radial accelerations playing the
role of gravity. It has been shown by Fraley et al. (1974b), McCrory et al.
(1977), Montierth et al. (1979), Freeman et al. (1977), Mason (1975), and
Lindl (1977) that the most effective implosion systems are particularly vul-
nerable to this type of instability because they consist of one or more rela-
tively thin spherical shells that accelerate through many times their initial
thickness during implosion.
=
In general, the larger the initial aspect ratio A 0 r 0 / ~r0 (where r 0 and
~r0 are initial shell radius and thickness) of the shells, the better the one-
dimensional performance of the system. Typical desirable values ofA 0 are
in the range of 10 :::;A 0 :::; 100.
R-T instability occurs in these systems when and where acceleration is in
the same direction as the local density gradient, so that, in effect, a heav-
ier fluid is above a lighter one in the local gravitational field. This occurs
first at the outside surface of the outer shell as illustrated in Fig. 5.19. The
role of the lower-density fluid is played by the material that has been heated
directly by thermal conduction from the outer high-temperature corona re-
gion and is being ablated outward from the shell. The inward acceleration
causes ripples in the ablation front region to grow unstably. Later, when the
fuel has reached the center, the interface between the fuel and high-density
pusher shell also becomes unstable, because the pusher is decelerated by
fuel pressure.
Targets that are designed for optimum performance should probably
come as close as possible to failure from R-T instability without failing. The
outside instability causes growth of shell distortions, but should not result in
a complete loss of shell integrity at the inner surface. The inside instability
Heavy
\
~= ~
0
e 'Yt
Figure 5.19 Classical Taylor instability conditions are illustrated by the container
of two fluids in a gravitational field (left). The analogous situation for the ablatively
driven Taylor instability is illustrated on the right [from McCrory eta!. (1981)].
will cause further growth, but should not result in the interface between the
fuel and the pusher shell becoming so distorted that fuel and pusher mate-
rial become intermixed, and fuel compression stopped or ignition precluded
and the thermonuclear burn quenched. Because both the inside and out-
side instability contributes to this failure, one must calculate the behavior
of both in detail to predict and avert failure. However, at present the out-
side instability is in a sense more important: Less is known about it, and it
appears possible that it may not be as troublesome as predicted from linear
analysis (Verdon et al., 1982).
The inside instability behaves nearly classically, since ablation does not
play a significant role. Classical analytic estimates of growth rates, such
as Chandrasekhar's, are therefore approximately correct, and the nonlin-
ear development can be anticipated from previous numerical simulations
of the classical R-T instability that show the bubble and spike structure
(see, e.g., Daly, 1967; Harlow and Welch, 1966; Birkhoff, 1954 and 1955).
Research on linear growth rates by Bodner (1974), Henderson and Morse
(1973), Shiau et al. (1973), Henderson et al. (1974), McCrory et al. (1977),
Emery et al. (1982), and Kull and Amisinov (1986) shows that in cases of
interest, ablation phenomena reduce the linear growth rates significantly at
shorter wavelengths, but that the growth rates are still large enough to cause
serious difficulty. There has also been an indication from numerical simu-
lations (Verdon et al., 1982) that nonlinear saturation mechanisms may in
some cases limit shell distortion enough to prevent failure.
Figure 5.19 shows the classical R-T instability conditions in a container
holding two fluids in a gravitational field g on the left and the analogous
lnertlally Confined Fuston 241
situation on the outside of the shell on the right, which is being given an
inward acceleration a by ablation pressure at the outside surface. The am-
plitude of the ripple of the interface between the two fluids in the container
has the form
(5.17)
wherex is the coordinate parallel to the interface, k = 27r/\ and eo is the

amplitude at t = 0. If the fluids are incompressible and inviscid, the thick-
nesses of the fluid layers are great compared to the ripple wavelength A, and
the heavy fluid is much more dense than the light fluid; then, the growth rate
1 is
(5.18)
To the extent that classical theory is at all applicable to the ablative out-
side surface, the assumption that the heavy fluid is much more dense than
the light fluid is a good approximation at irradiances appropriate for fusion
(<h > 1014 W/cm2 ).
From the classical theory, estimates can be made of the conditions under
which shell failure will occur. This is expected if the surface ripple ampli-
tude is of the order of the shell thickness
(5.19)
where A= r(t)/C:ir(t) is the in-flight aspect ratio of the shell, where &(t)
is defined as the width defined by the positions about the maximum den-
sity in the ablatively accelerated shell that contain approximately 80% of
the shell mass. Here, all quantities are taken not at the initial time but at
the time t1 , when the failure criterion is first met. To obtain numerical es-
timates that indicate the magnitude of the problem, we must choose values
of k that will be most troublesome. From Eq. (5.18), the growth rate in-
creases with k, but the e-folding depth of penetration of the k mode into
the fluid from the interface varies as k- 1 • Nonlinear calculations indicate
that growth slows from exponential when the amplitude reaches this pene-
tration length. Modes with values of k- 1 less than the shell thickness should
not cause the shell to break up but may represent an "energy sink," i.e.,
turbulent energy occurs near the ablation surface that would otherwise be
available to accelerate the shell.
Thus, an estimate of the "worst" k is kw ::::::J &- 1, or since k = 1/r, 1 ::::::JA,
et
and = r/A. This determines the growth rate for the worst mode rw =
(Ag!r) 112 • Using this estimate in Eq. (5.17), atx = 0, we obtain
(5.20)
The larger t can be, i.e., the longer the instability has to grow, the smaller eo
must be to prevent failure, and practical limits on surface finish put lower
limits on eo· An estimate of the time for failure to occur in a fraction of the
implosion time is r1 = (2r 0/g ) 112. As the shell moves toward the origin, its as-
pect ratio first increases rapidly because the shell is made thinner by shock
and acceleration-induced compression. Typically, A increases from A 0 by
a factor of 4 or more. Then, A changes slowly for an intermediate period
>
e
:l.
>
X( ~o~m) X( ~o~m)
Figure 5.20 Time evolution of the unstable 2.5-J.tm-wavelength case prior to burn
through (t = 130 psec) from a DAISY simulation. The amount of mass between
each line is a constant throughout the calculation [from McCrory et al. (1981)].
during which the effects of convergence and ablative mass removal roughly
cancel, then convergence dominates, the shell thickens while the radius de-
creases, and A decreases. It is during the intermediate period when A is
largest and the shell is near its initial radius r0 that the shell is most vul-
nerable to break up by the large 1 (or k) fast-growing modes. Thereafter,
as A decreases, it is expected that the growth of the high 1 modes will have
relatively less disruptive effects of the shell. For these reasons, t1 = r 1 /2,
4A
A = 0, and r = r0 are reasonable choices. Thus, the initial perturbation
amplitude eo that would cause failure is
r exp- [ (4A~")
eo = _o
4Ao _uo_
r0
112 7
2l
..1.. = _o
4
f:!.r exp -(2Ao)l/2 (5.21)
For typical values of interest, r0 = 1 em and A 0 = 25, we find eo =

8.5 X 10-6 em; for Ao = 15, eo = 7 X 10-5 em. These are stringent toler-
ances, and there is little margin for error. Linearized calculations presented
in Fraley et al. (1974b), which include ablative effects, obtain amplifica-
tions smaller than this, mainly because the ablative growth rate is reduced
by about a factor of 2. Emery et al. (1986) have performed calculations
of 0.25 fLID light that indicate a factor of 4 reduction in the growth rate.
With a factor of 4 reduction, for Ao = 25 and 15, eo = 1.17 X 10-3 em and
4.3 x 10-3 em, respectively. These, therefore, are much easier to achieve.
Although a detailed discussion of these effects is beyond the scope of this
chapter, we note in concluding this section that detailed two-dimensional
calculations of ablatively driven shells (McCrory et al., 1981a) furnish ev-
idence for additional optimism. In general, these calculations show that
even under conditions in which linear theory would predict shell breakup,
the inner surface of the shell may retain its integrity. An example of such a
calculation from McCrory et al. (1981b) is shown in Fig. 5.20. Even though
the inner surface in this calculation is distorted, fuel-shell mixing probably
would not occur.
5.7 IRRADIATION UNIFORMITY

REQUIREMENTS
Uniform laser irradiation of spherical targets is essential for obtaining

the high-density compressions necessary for the success of laser fusion.
Nonuniform irradiation can lead to nonspherical implosions, or even com-
plete target breakup by the Rayleigh-Taylor instability as described in
Sec. 5.6, resulting in degradation or failure of the thermonuclear burn. It
is generally accepted that the non uniformity, measured in terms of the rms
deviation of the intensity divided by the mean intensity, can be ±10%, pro-
20 Beams (f/7.9)
32 Beams (f/10.0)
"'
E
c
___ so Beams (f/13.7)
96 Beams (f/17.3)
FOCUS RATIO
(a)
Figure 5.21 (a) Calculated energy-deposition nonuniformity (arms) for various

numbers of beams: all beams together cover a total solid angle of2% of 4rr. (b) De-
pendence on the overall rms energy-deposition nonuniformity on focusing condi-
tions for the 24-beam OMEGA system for a quadratic, for a flat, and for a typical
distribution produced by an LLE laser system. (Tangential focus is defined here as
that focal position in which the target just intercepts 100% of the beam energy.)
vided additional uniformity occurs from transport processes between the

critical surface and the ablation surface of a directly illuminated spherical
pellet (see, e.g., Gardner and Bodner, 1981). Although absolute uniformity
of irradiation on a spherical target is not theoretically possible (Birkhoff,
1923), the required degree of irradiation uniformity can nevertheless be
achieved in principle using multiple-beam irradiation as discussed by Skup-
sky and Lee (1983) or induced spatial incoherence as shown by Lehmberg
et al. (1983). As an example of laser system requirements, the uniformity of
illumination for the 24-beam OMEGA laser system at LLE has been investi-
gated theoretically and nonuniformity values on the order of 5% rms were
found for the optimal focusing of beams whose radial intensity profile is
relatively smooth with a peak-to-valley intensity variation of approximately
20% rms [see Fig. 5.21(a)]. Such symmetry can be maintained during the
5
24 Beams (f/4.0)
4
/n
0 /n
3
~
"'
E
c~
FOCUS RATIO
(b)
full time of illumination for some of the typical, directly driven laser-fusion
targets discussed in Sec. 5.8.
In the present discussion, we consider the upper limit of illumination
uniformity for various beam intensity profiles and varying focal positions.
We give several simulation results for illumination uniformity by the 24-
beam OMEGA laser system to illustrate these considerations.
To estimate the amount of uniformity possible with direct laser irradi-
ation, a ray-trace code has been used to calculate the three-dimensional
energy deposition pattern in a spherically symmetric plasma corona whose
density, temperature, and scale length are taken from the simulation re-
sults of the one-dimensional hydrocode LILAC. The irradiation uniformity

depends on the number of laser beams, their geometrical disposition, the
individual beam intensity profiles, and the beam-to-beam energy balance.
Figure 5.21(a) shows the calculated rms nonuniformity plotted against the
focus ratio for different numbers of symmetrically placed beams, all having
the same quadratic beam profile, and under conditions of perfect beam bal-
ance and perfect beam-target positioning. (The focus ratio is defined as the
ratio of the beam radius in the target plane to the target radius.) As the tar-
get radius decreases during the implosion, the focus ratio becomes larger,
typically sweeping up to 1.5 for an initial ratio of 1.0. Both the 60- and
96-beam systems maintain rms nonuniformities below 1% over this whole
range. The 32-beam system has good uniformity near tangential focus but
would require some thermal smoothing within the target near the peak of
the pulse.
Target performance is affected not only by the magnitude of a nonuni-
formity but also by its spatial wavelength. For instance, short-wavelength
modes may be thermally smoothed within the target after an adequate
plasma scale length has been established (Bodner, 1981 ); prior to that time,
they can imprint themselves on the target surface and "seed" the R-T insta-
bility. Spherical-harmonic decomposition of the irradiation nonuniformity
on the target surface has been performed, with the overall rms nonunifor-
mity arms expressed in terms of components a1
(5.22)
The origins of the contributions from the different spherical-harmonic

modes l have been identified; the main results may be summarized as fol-
lows:
1. Modes with l = 1 to 4 are created by energy imbalance between beams, beam-
target misalignment, and elliptical beam cross sections.
2. Modes with l = 8 to 12 are the characteristic modes arising from beam
overlap in a 20- to 32-beam system. (l = 8 is characteristic of the 24-beam
OMEGA system, and l ""20 is characteristic of a 96-beam system.)
3. Modes withl higher than the values in 2 above are produced by imperfections
in the laser beam.
Calculations similar to those of Fig. 5.21( a) have been carried out for the
24-beam OMEGA system. Figure 5.21(b) shows the overall drive unifor-
mity that can be obtained with three different beam profiles for varying fo-
cusing conditions. A beam profile approaching quadratic leads to optimum
uniformity, with rms variations of order 1% attained over a broad range of
focus ratios.
Another example of a beam containing small-scale imperfections is a
quadratic profile containing 1.2:1 peak-to-valley modulations on various
spatial wavelengths, given by
(5.23)
where r 0 corresponds to the edge of the beam. Results for the rms nonuni-
formity a-rms as a function of the parameter N are shown in Fig. 5.22 for a
32-beam system. The dominant modes of nonuniformity (l) are indicated
on the upper curve. The lower curve indicates the effect of thermal smooth-
ing, assumed here to damp the amplitude a-1 of the lth mode by a factor of
exp( -l/20). (The damping corresponds to a fractional separation 6R/R of
0.05 between the critical and ablation surfaces, which is attained for typical
target designs.) At large values of N, the high-order l modes are strongly
damped, leaving the low-order modes characteristic of the 32-beam symme-
i = 10
i = 12
32 beams (f/20)
Dominant Focus Ratio 0.9
Nonuniform Mode Intensity Profile: Lll/1 = 20%
3.0 =
i 18
1=1 0 (1-r 2 /r~)
x [1 + 0.1 cos (N rr r lr 0 )]
Example: N =3
i = 28
c
(J)
:IE
a:
2.0
i
...
..
...
A
f. . ..
1.0 : •.··)···.... ··.
With Smoothing: ·•·····.••.
.:lR/R =0.05
(a.~- ai e·V20) ··...................' -----
Figure 5.22 Energy-deposition nonuniformity (arms) for beams with a quadratic

intensity profile including 20% ripples on the envelope. The modulation parameter
N determines the number of ripples across the beam profile as shown in the insert.
The l numbers above the curve show which spherical harmonic mode is predomi-
nant for a given N. The lower curve assumes a damping of each model by a factor
exp( -l/20) due to a small amount of thermal smoothing.
Shot- 6391
Energy- 97 J
Pulse Width - 1.0 ns
t-------i
250 J.Lm
CONTOUR PLOT
(a) (b)
1.0
0.8
0.6
0.4
0.2
0.0 ._____.._ __.__ _.__ _,_____..._____,

-0.75 -0.50 -0.25 0.0 0.25 0.50 0.75
RADIAL AVERAGE (mm)
(c)
Figure 5.23 Equivalent-target-plane beam energy distribution obtained on the

one-beam, 1.05 J.Lm, GDL laser at LLE: (a) a contour plot of the energy distribution
with 10% contours; (b) and (c) a radial averaged profile and a three dimensional
representation of the energy distribution.
Multi-Wavelength
Energy-Sensing
System
70% R
UV Pass Filler Etalon
Figure 5.24 Schematic of experimental setup for photographing the equivalent-

target-plane intensity distribution. GR refers to a photodiode filtered for the second
harmonic of the Nd:glass laser (0.527 pm).
try pattern. It is clear from the large enhancement of urms over its value at
N = 0 (no modulation) that particular attention must be paid to the spatial
uniformity of each beam.
The level of performance that is required is characterized by results ob-
tained using the one-beam GDL laser (of the Laboratory for Laser En-
ergetics) at 1.054 pm for the equivalent-target-plane energy distribution
(Fig. 5.23).
Preliminary experimental results have been obtained on UV beam char-
acterization. Accurate determination of the on-target distribution requires
careful analysis of the focusing properties of each of the OMEGA beams.
Since the focusing lenses on OMEGA have very nearly diffraction-limited
performance characteristics, the equivalent-target-plane (ETP) intensity
distribution may be measured using an auxiliary long-focal-length lens
(Fig. 5.24). This allows greatly enlarged images of the ETP intensity dis-
tribution to be photographed and subsequently digitized and analyzed. In-
tensity contour plots (corrected for nonlinear film response) and histograms
may then be obtained. Examples for one beam are shown in Figs. 5.25(a)
Contour Levels: 95%, 75%, 50%, 25%, 1%

600~----------------------------------~
Shot 10964 Spatial Resolution 10 pm
200 400 600
Distance in Target Plane (pm)

(a)
Figure 5.25 (a) Analysis of ETP of one of the OMEGA beams. Intensity contour
plot. This beam suffers from imperfections due to conversion-crystal and laser op-
tics inhomogeneities. Intensity contour levels are 95%, 75%, 50%, 25%, and 1% of
the maximum intensity. (b) Histogram of the ETP intensity distribution. The ordi-
nate gives the fractional energy and the integrated fractional energy as a function
of intensity. The intensities at which 50% and 95% of the energy are included are
indicated.
I nertially Confined Fusion 251
1.0
Shot 109:......-t----
/ 95%
/
/-E(I)
/
I
t
I
c
.2
u 0.5 (arbitrary units)
I!
/so%
II.
I
I
I
I
I
/
o.oi<.:-~----'-------L------L-----'---~
0.0 0.5 1.0 1.5 2.0 2.5
Intensity on Target (10 14 W/cm2)

(b)
and 5.25(b ), respectively (from Soures et al., 1983). The spatial profiles here
are significantly modulated as a result of inhomogeneities in the laser op-
tics and conversion crystals. Although these problems affect the OMEGA
beams at present, we have been able to show that marked improvements in
uniformity are possible if high-quality crystals are chosen and appropriate
precautions taken.
Another area of investigation is the possibility of characterizing the beam
nonuniformities, in a quantitative way, in terms of the features of his-
tograms, Fourier transforms, and contour plots that can be directly related
to specific components of the laser system.
5.8 IMPLOSION EXPERIMENTS
A major portion of the international activity devoted to inertial-fusion re-

search has been dedicated to the study of laser-driven implosions. This ac-
tivity requires the use ofmulti-terawatt laser systems, precision-microscopic
target fabrication, an array of ultra-high-speed diagnostic systems, and so-
phisticated multidimensional hydrodynamic simulation computer calcula-
tions to aid in the design and analysis of the experiments. In this section, we
briefly review some of the laser facilities currently used in laser-fusion ex-
perimental research, examine some of the experimental results from these
facilities, and discuss some of the proposed research leading to the proof
of feasibility of the laser-fusion process. A detailed discussion of the target
fabrication technology is beyond the scope of this publication.
5.8.1 Implosion Experiments and Diagnostics
Some of the world's most sophisticated experimental facilities are currently

dedicated to the study of laser-matter interaction and laser-driven implo-
sions. The majority of these facilities are Nd:glass lasers operating at har-
monics of the fundamental laser radiation. The largest of these lasers is the
NOVA facility at the Lawrence Livermore National Laboratory. NOVA is
dedicated principally to investigation of the feasibility of the radiation drive
approach to laser fusion (see Sec. 5.1).
Much of the U.S. effort on direct-drive laser fusion is centered at the
University of Rochester Laboratory for Laser Energetics using the Omega
24-beam N d:glass laser facility. A schematic of the Omega system is shown
in Fig. 5.26 (see Soures et al., 1983). As shown in Sec. 5.7, the Omega geom-
etry lends itself to uniform target illumination, a prerequisite for effective
direct-drive implosions.
The highest compressed fuel density recorded to date in inertial fusion
experiments is in the range of 10 to 25 g/cm3 • High-density experiments
have been carried out on the Lawrence Livermore NOVA and NOVETTE
lasers, the LLE OMEGA facility, the Gekko XII laser at the University of
Osaka, the Vulcan facility at Rutherford, and the Chroma laser at KMS.
Figure 5.27 (valid as of May 1987) shows a summary of some of the implo-
sion experiments carried out to date in which the fuel temperature and fuel
density could be simultaneously measured (or estimated) using core diag-
nostic techniques.
Reliable measurements of the density (p), density-radius product (pR),
and electron and ion temperatures of compressed targets are very important
in inertial fusion research. Most of these parameters can be measured with
nuclear particle methods if the target compression results in a sufficiently
high thermonuclear yield. X-ray spectroscopic techniques are also useful
for making direct measurements of electron temperature, density, and pR
in situations in which the thermonuclear yield is either unavailable or too
low (see, e.g., Yaakobi et al., 1980).
Stark broadening, line absorption, and continuum absorption are some
of the x-ray techniques that have been applied to high-density diagnostics at
LLE. In Fig. 5.28, the measured and calculated Stark profiles of the Lyman-
~
w
Fig. 5.26 A schematic of the OMEGA, 24-beam Nd:glass laser showing some of its principal features.
100.0
(')
E
l)
10 0
'Ol
?:
(f)
c
Q)
0
Q)
::J
LL
1.0 LLE~ LLE

e radiation dnve
LLE~
0 direct drive
• cannonball """
OSAKA + +LLE
~LLNL
0.1
0.1 1.0 10.0 100 0
Fuel Temperature (keV)
Figure 5.27 Summary of high-density implosion experiments.
f3 line of argon, obtained from the implosion of an argon-filled target, are

compared; the density inferred is 9 g/cm 3 • Calculations show that non-LTE
effects cause the lines to be emitted only near peak compression. Such spec-
tra have given impetus to new theoretical developments in the field of Stark
broadening.
Figure 5.29 shows a different type of x-ray line (emitted by helium-like
rather than hydrogen-like ions). The comparison of the line profile for the
thin-shell (bare) and the thick-shell (plastic-coated) implosions provides a
vivid demonstration that suppression of fast electrons by using a thick shell
leads to higher compression. The higher density is evident through the line
broadening, spectral shift, and appearance of a normally forbidden tran-
sition (ls 2 - ls3d). In these experiments (Yaakobi et al., 1980), the pR of
the fill gas was determined using the optically thick resonance line (Lyman
a); the opacity-broadened line width was found to be consistent with code
predictions, implying the absence of severe layer mixing.
Ne =2.17 x 1024 cm·3

p "'9 g-cm- 3
Ar+ 17
Lyman- {3
-4 -2 0 2 4
.1 E (Rydbergs)
Figure 5.28 Spectrum from an argon-filled glass microballoon imploded with 24

1054 p,m beams of OMEGA at 2.8 x 1016 W/cm 2 . By comparison with a calculated
Stark-broadened profile for the Ar+1 7 Lyman-,Bline, a density of9 g!cm 3 is inferred.
The laser pulse width was 94 ps and the target diameter was 77 p,m.
Various techniques employing nuclear reaction products have been used

to measure the parameters of compressed targets. The two primary tech-
niques implemented at LLE are (1) the analysis of energetic fuel ions pro-
duced by the elastic scattering of DT neutrons and (2) the neutron activa-
tion of target constituents (see Fig. 5.30).
The "knock-on" technique developed at LLE makes use of elastically
scattered fuel ions to measure the fuel pR at burn time in laser-driven tar-
gets (see, e.g., Kacenjar et al., 1982; Skupsky and Kacenjar, 1981; Kacen-
jar et al, 1983; and Kacenjar et al., 1984). With specially constructed tar-
gets containing a mixture of DT and argon, it has been possible to mea-
sure the pR of the DT at burn time using this technique and to compare
this value with an independently derived measure of the pR from argon line
spectroscopy. These two measurements have been found to give consistent
results.
Neutron activation of Si in the glass shell of imploded targets was first de-
veloped as a tamper-pR diagnostic technique at LLNL (see Campbell et al.,
1980) and Lane et al., 1980). A new technique now under development by
LLNL and LLE makes use of krypton activation, allowing direct measure-
ment of the fuel pR by seeding the fuel with a small amount of krypton (see,
e.g., Lane et al., 1986).
- - UNCOATED TARGET
-COATED TARGET
K+1B
1s-2p
ARGON+16
1s2-1s3p
I I I I I I
-60 -40 -20 0 20 40 60
~E (eV)
Figure 5.29 Spectra from a glass target coated with 4.2 1-1m of plastic (heavy line)
and an uncoated target (thin line) irradiated with six 1054-nm OMEGA beams. Pa-
rameters: laser intensity= 2.4x 1016 W/cm 2 , pulse length= 50 ps, target diameter=
51 1-1m, and target wall thickness= 0.5 J-tm.
An alternative technique to measure the fuel areal density in compressed

targets involves the use of secondary fusion reaction products as suggested
by Campbell and Southworth (1974), Gamalii et al. (1975), Blue and Harris
(1981), and Blue et al. (1983). The two measurements that have been con-
sidered and attempted (Miyanaga et al., 1986 and Cable et al., 1986) are
the secondary DT reaction (resulting in a 14.1 MeV neutron) and the sec-
ondary DHe3 reaction (resulting in a 14.7 MeV proton) from a D 2 -fueled
target. For both of these reactions, the yield can be related to the deuterium
fuel areal density at burn time, provided the electron temperature is known
and no significant fuel-ablator mixing takes place. If the fuel is mixed with
ablator material, a simultaneous measurement of these two reactions along
with a knowledge of the electron temperature can lead to a reasonable es-
timate of the mix density.
The highest neutron yields and thermonuclear burn efficiencies have
been measured on directly driven targets at the LLNL, Osaka, and LLE
laser facilities. Generally, the highest neutron yields are produced by irra-
Elastically
Scattered
Neutron
Elastically
Scattered
D or T
(a) (b)
Figure 5.30 Nuclear diagnostic techniques implemented at LLE. The knock-on

technique (a) determines the fuel pR, while the neutron-activation (radiochemistry)
technique (b) determines the pusher pR.
diating large-diameter, thin-wall targets with short-wavelength laser radia-

tion. Figure 5.31 gives a summary of some of the highest-yield experiments
to date.
5.8.2 Future High-Density

Compression Experiments
The objective of the world's principal laser-fusion compression facilities

over the next few years is to demonstrate increasingly high density in direct-
and indirect-drive implosions.
The goal of the 24-beam UV implosion experiments planned for OME-
GA for the next three years is to demonstrate target performance that will
scale to reactor conditions. Reactor targets typically have a convergence ra-
tio (initial fuel radius divided by compressed fuel radius) between 10 and 30,
a compressed fuel density p of several hundred g!cm3, a fuel pR in excess of
1 g!cm 2 , and multi-ke V ion temperatures (T). With the 2 kJ currently avail-
able on OMEGA, the targets under present consideration cannot contain
sufficient fuel to achieve, simultaneously, reactor values of pR, p, and T; it
should, however, be possible to drive these targets to fuel densities in excess
of 20 g/cm2 (100 times liquid density) with convergence ratios above 10 and
temperatures of the order of a few ke V.
Computer simulations have been performed with the one-dimensional
hydrodynamics code LILAC to identify targets that could produce the high-
1014
1013
"'0
Q) 1012
>-
-
c::
0
....
:::J 1011
z
<!)
1010
1.05 1-1m o
109
10 2
Laser Energy (J)

Figure 5.31 Summary of high-neutron-yield implosion experiments. The straight
lines delineate lines of constant thermonuclear yield efficiency, 1JTN (thermonuclear
burn energy/laser energy absorbed).
est compressed fuel densities given OMEGA's expected operating condi-

tions. The conditions assumed were 2.0 kJ of 351 nm energy in a 600 psec
pulse (full-width at half-maximum), focused with f/4 optics at a point 10
target radii behind the target center (corresponding to a focus ratio of 1.2).
This focal position was chosen to provide good uniformity [see Fig. 5.21( a)];
a focus ratio of 1.4 might actually be required and could be used with a small
decrease in total absorption due to increased refractive losses.
Two classes of targets were considered: DT gas-filled and cryogenic DT.
OMEGA experiments will initially make use of gaseous fuel until a cryo-
genic capability is developed. Target shells were assumed to be either pure
CH or CH overcoated on glass. Target radii between 150 and 250 J.Lm were
found to produce both high yields and high densities for cryogenic targets.
For gas-filled targets, optimal results were obtained with 100 J.Lm radius tar-
gets at a fill pressure of 10 atm.
Performance figures of merit for a gas-filled target and cryogenic target
are summarized in Table 5.1, using results from one-dimensional LILAC
simulations. The substantially higher neutron yield and compressed den-
I nertially Confined Fusion 259
Table 5.1 Results from LILAC Simulations of Targets

for High-Density OMEGA Experiments
Near-term Future
gas-filled cryogenics
Target
Target radius 200 ttm 200 ttm
Fuel mass 36 ng (10 atm) 1000 ng (10 ttm)
Shell 4rtm Si02 4ttm CH2
Performance
Neutrons 2 X 10 11 7.5 X 10 11
Pmaxf Pliquid 50 1000
(pR)fuel 7 X 10- 3 g/cm 2 1.3 X 10-l g/cm 2
(pR)shell 3 X 10- 2 g/cm 2 1.4 x 10- 2 g/cm 2
Convergence ratio 15 17
sity achieved by the cryogenic target (with roughly the same convergence
ratio as the gas-filled target) illustrate the importance of a cryogenic-target
capability in laser-fusion experiments.
The effects of laser irradiation non uniformities and target imperfections
are expected to reduce the compressed density of these targets. In one set of
ORCHID simulations, a laser non uniformity in an l = 8 Legendre mode was
applied to the target. (This is the characteristic mode produced by overlap-
ping beams of OMEGA) The resulting degradation in compressed density
and neutron yield for the cryogenic target is shown in Table 5.2, for peak-
to-valley variations of 2 and 4%. The achievable uniformity on OMEGA is
estimated to have 5 to 10% peak-to-valley variations (1 to 2% rms) for l = 8
(Sec. 5.8); an extrapolation of Table 5.2 suggests that the reduction in peak
density and neutron yield should be less than a factor of 12. A much smaller
factor is found for longer-wavelength non uniformities (l = 2 and 4), char-
acteristic of energy imbalance between the beams. The resulting density of
,...., 200 times liquid-DT density would still be among the highest achieved in
laser-fusion experiments. The effects of higher-order modes have not yet
been calculated; they can be smoothed by thermal conduction after an ad-
equate plasma atmosphere has been established, but prior to that they can
imprint a perturbation on the target and seed the Rayleigh-Taylor instabil-
ity.
The primary diagnostics for these targets will be knock-on particles for
the fuel pR and neutron activation (radiochemistry) for the pR of targets
with glass shells. Even with the anticipated degradation in target perfor-
mance, the yield-pR product for these targets should be well above the
Table 5.2 Degradation in Target Perfor-

mance Produced by Irradiation Nonunifor-
mities with l = 8.
Nonuniformity Density Neutron yield

(peak-to-valley) (relative) (relative)
0 1 1
2% 0.83 0.70
4% 0.81 0.66
thresholds for both diagnostics. For some targets, the total pR might ex-
ceed the limit of applicability of the knock-on diagnostic (pR,...., 0.2 g/cm2 ).
For these, a krypton activation technique is being developed. In addition,
a comprehensive set of x-ray diagnostics will be used to determine the dy-
namics of the implosion.
Through these experiments, it is possible to address many of the im-
portant issues related to high-performance targets and obtain an extensive
database to determine the accuracy of multidimensional code calculations.
These experiments, along with the numerical simulations, are expected to
enhance considerably our understanding of the effects of nonuniformities
on ablatively driven, direct-drive, laser-fusion targets.
ACKNOWLEDGMENT
This work was supported by the U.S. Department of Energy Office of In-
ertial Fusion under agreement no. DE-FC08-85DP40200 and by the Laser
Fusion Feasibility Project at the Laboratory for Laser Energetics, which
has the following sponsors: Empire State Electric Energy Research Corpo-
ration, General Electric Company, New York State Energy Research and
Development Authority, Ontario Hydro, and the University of Rochester.
Such support does not imply endorsement of the content by any of the above
parties.
REFERENCES
Albritton, J.R. (1983). Laser absorption and heat transport by non-Maxwell-Boltz-

mann electron distributions, Phys. Rev. Lett., 50: 2078.
Amiranoff, R., Fabbro, R., Fabre, E., Garban, C., Virmont, J., and Weinfeld, M.
(1979). Experimental transport studies in laser-produced plasmas at 1.06 and
0.53 J.tm, Phys. Rev. Lett., 43: 522.
Basov, N.G. and Krokhin, O.N. (1964). Conditions for heating up of a plasma by
the radiation from an optical generator, Sov. Phys. JETP, 19: 123.
Basov, N.G., Kruikov, P.G., Zakharov, S.D., Senatskii, Y.V., and Tchekalin, S.V.
(1968). Experiments on the observation of neutron emission at the focus of the
high-power laser radiation on a lithium deuteride surface, IEEE J. Quant. Elec-
tron., QE4: 864.
Bell, A.R., Evans, R.G., and Nicholas, D.J. (1981). Electron energy transport in
steep temperature gradients in laser-produced plasmas, Phys. Rev. Lett., 46: 243.
Birkhoff, G. (1923). Relativity and Modern Physics, Harvard Univ. Press, Cambridge,
Mass., p. 253.
Birkhoff, G. (1954). Taylor instabilities and laminar mixing, Los Alamos Scientific
Lab. Rep. LA-1862, Los Alamos, New Mexico.
Birkhoff, G. (1955). Taylor instability. Appendices to Report LAS-1862, Los Alamos
Scientific Lab. Rep. LA-1927, Los Alamos, New Mexico (unpublished).
Blue, T.E. and Harris, D.B. (1981). The ratio of D-T toD-D reactions as a measure
of the fuel density-radius product in initially tritium free inertial confinement
fusion targets, Nucl. Sci. Eng., 77: 463.
Blue, T.E., Blue, J.W, Durham, J.S., Harris, D.B., Hnesh, AS., and Reyes, J.J.
(1983). D- 3 He proton yield as a diagnostic for D-T and D 2 filled inertial con-
finement fusion experiments, J. Appl. Phys., 54: 615.
Bodner, S.E. (1974). Rayleigh-Taylor instability and laser-pellet fusion, Phys. Rev.
Lett., 33: 761.
Bodner, S. (1981). Critical elements of high gain laser fusion, J. Fusion Energy, 1:
221.
Brueckner, K.A. and Jorna, S. (1974). Laser-driven fusion, Rev. Mod. Physics, 46:
325.
Cable, M.D., Lane, S.M., Glendinning, S.G., Estabrook, KG., Suter, L.J., Derzon,
M.S., and Prussin, S.G. (1986). Implosion experiments of Nova, Bull. Amer. Phys.
Soc. 31: 1461.
Campbell, H.D. and Southworth, F.H. (1974). Methods of determining pR in laser
fusion experiments using neutron diagnostics, Proc.lst Topl. Meeting on Technol-
ogy of Controlled Nuclear Fusion, San Diego, Calif., April16-18, 1974, American
Nuclear Society, p. 75.
Campbell, P.M., Johnson, R.R., Mayer, F.J., Powers, L.V., and Slater, D.C. (1977).
Fast-ion generation by ion-acoustic turbulence in spherical laser plasmas, Phys.
Rev. Lett., 39: 274.
Campbell, E.M., Hicks, H.G., Mead, WC., Coleman, L.M., Hatcher, C.W, Dellis,
J.H., Boyle, M.J., Larsen, J.T., and Lane, S.M. (1980). Use of a radioactive tracer
to determine the fraction of fusion target debris collected, J. Appl. Phys., 51: 6065.
Chandrasekhar, S. (1981). Hydrodynamic and Hydromagnetic Stability, Oxford,
Chapter X, p. 428.
Charatis, G., Downward, J., Goforth, R. Guscott, B., Henderson, T., Hildum, S.,
Johnson, R., Moncur, K., Leonard, T., Mayer, F., Segall, S., Siebert, L., Solomon,
D., and Thomas, C. (1975). Experimental study of laser driven compression of
spherical glass shells, Proc. of the 5th Internat. Conf on Plasma Physics and Con-
trolled Nuclear Fusion Research, Vol. II, Tokyo, lnternat. Atomic Energy Agency,
Vienna, p. 317.
Conn, R.W (1981). Magneticfusion reactors, in Fusion, Magnetic Confinement, Vol.

1B, (E. Teller, ed.), Academic Press, New York, p. 193.
Craxton, R.S. (1980). Theory of high efficiency third harmonic generation of high
power Nd-glass laser radiation, Opt. Commun., 34: 474.
Craxton, R.S. and Haines, M.G. (1975). Hot spots in laser plasmas, Phys. Rev. Lett.,
35: 1336.
Craxton, R.S. and McCrory, R.L. (1986a). Calculations of the wavelength depen-
dence of absorption, Lab. for Laser Energetics Rept., 108.
Craxton, R.S. and McCrory, R.L. (1980b). Lab. for Laser Energetics Rept., 99.
Daly, B.J. (1967). Numerical study of two fluid Rayleigh-Taylor instability, Phys.
Fluids, 10: 297.
Dawson, J.M. (1964). On the production of plasma by giant pulse lasers,Phys. Fluids,
7:981.
Delettrez, J. (1986). Thermal electron transport in direct-drive laser fusion, Can. J.
Phys., 64:932.
Eidman, K., and Sigel, R. (1974). Backscatter experiments, Laser-Matter Interac-
tion and Related Plasma Phenomena, Vol. 3B, (H.J. Schwarz and H. Hora, eds.),
Plenum Press, New York, p. 667.
Emery, M.H., Gardner, J.H., and Boris, J.P. (1982). Rayleigh-Taylor and Kelvin-
Helmholtz instabilities in target accelerated by laser ablation, Phys. Rev. Lett.,
48:677.
Emery, M.H., Gardner, J.H., and Bodner, S.E. (1986). Strongly inhibited Rayleigh-
Taylor growth with 0.25-J-!m lasers, Phys. Rev. Lett., 57: 703.
Estabrook, K. and Kruer, WL. (1978). Properties of resonantly heated electron dis-
tributions, Phys. Rev. Lett., 40: 42.
Fechner, WB., Shepard, C.L., Busch, G.E., Schroeder, R.J., and Tarvin, J.A. (1984).
Analysis of plasma density profiles and thermal transport in laser-irradiated
spherical targets, Phys. Fluids, 27: 1552.
Floux, F., Cognard, D., Denoeud, L.G., Piar, G., Parisot, D., Bobin, J.L., Delobeau,
R., and Fauquignon, C. (1970). Nuclear fusion reactions in solid-deuterium laser-
produced plasma, Phys. Rev. A, 1: 821.
Fraley, G., Linnebur, E.J., Mason, R.J., and Morse, R.L. (1974a). Thermonuclear
burn characteristics of compressed deuterium-tritium microscopheres, Phys. Flu-
ids, 17: 474.
Fraley, G., Gula, W, Henderson, D., McCrory, R., Malone, R., Mason, R., and
Morse, R. (1974b). In Plasma Physics and Controlled Nuclear Fusion Research,
Vol. 543, Internat. Atomic Energy Agency, Vienna.
Freeman, J.R., Clauser, M.J., and Thompson, S.L. (1977). Rayleigh-Taylor instabil-
ities in inertial-confinement fusion targets, NucL Fusion, 17: 223.
Gamalii, E.G., Gus'kov, S. Yu., Krokhin, O.N., and Rozanov, V.B. (1975). Possibility
of determining the characteristics of laser plasma by measuring the neutrons of
the DT reaction, JETP Lett., 21: 70.
Garban-Labaune, C., Fabre, E., Max, C.E., Fabbro, R., Amiranoff, F., Virmont, J.,
Weinfeld, M., and Michard, A (1982). Effect of laser wavelength and pulse du-
ration on laser-light absorption and back reflection, Phys. Rev. Lett., 48: 1018.
Gardner, J.H. and Bodner, S.E. (1981). Wavelength scaling of reactor-size laser-
fusion targets, Phys. Rev. Lett., 47: 1137.
Goldsack, T.J., Kilkenny, J.D., MacGowan, B.J., Cunningham, P.F., Lewis, C.L.S.,
Key, M.H., and Rumsby, P.T. (1982). Evidence for large heat fluxes from the mass
ablation rate of laser-irradiated spherical targets, Phys. Fluids, 25: 1634.
Goldstone, P.D., Day, R.H., Eden, G., Ameduri, F., Mead, W.C., Goldman, S.R.,
Richardson, M.C., Keck, R.L., Seka, W., Pien, G., Soures, J.M., McCrory, R.L.,
and Knauer, J. (1984). Interaction physics and x-ray emission in high-Z plasmas
at 351 nm, BulL Amer. Phys. Soc., 29: 1318.
Hares, J.D., Kilkenny, J.D., Key, M.H., and Lunney, J.G. (1979). Measurement of
fast-electron energy spectra and preheating in laser-irradiated targets, Phys. Rev.
Lett., 42: 1216.
Harlow, F.H. and Welch, J.E. (1966). Numerical study of large-amplitude free-
surface motions, Phys. Fluids, 9: 842.
Hauer, A., Preidhorsky, W., and Van Hulsteyn, D. (1981). Inner shell satellite lines
in C0 2 laser-produced plasmas,AppL Opt., 20: 3477.
Hauer, A., Mead, W.C., Willi, 0., Kilkenny, J.D., Bradley, D.K., Tabatabaei, S.D.,
and Hooker, C. (1984). Measurement and analysis of near-classical thermal
transport in one-micron laser-irradiated spherical plasmas, Phys. Rev. Lett., 53:
2563.
Henderson, D.B. and Morse, R.L. (1974). Symmetry of laser-driven implosions,
Phys. Rev. Lett., 32: 355.
Henderson, D.B., McCrory, R.L., and Morse, R.L. (1974). Ablation stability of
laser-driven implosions, Phys. Rev. Lett., 33: 205.
Hively, L.M. (1977). Convenient computational forms for Maxwellian reactivities,
NucL Fusion, 17:873.
Kacenjar, S., Skupsky, S., Entenberg, A., Goldman, L., and Richardson, M.C.
(1982). Direct measurement of fuel density-radius product in laser-fusion exper-
iments, Phys. Rev. Lett., 49: 463.
Kacenjar, S., Goldman, L., and Entenberg, A. (1983). Copper activation counter
calibration using solid state track detectors, Rev. Sci. /nst., 52: 831.
Kacenjar, S., Goldman, L.M., Entenberg, A., and Skupsky, S. (1984). (pR) Measure-
ments in laser produced implosions using elastically scattered ions, 1 AppL Phys.,
56:2027.
Kauffman, R.L., Drake, R.P., Turner, R.E., Lasinski, B.F., Tirsell, G., Mead, W.C.,
and Stover, E. (1984). X-ray conversion efficiency from planar Au disks at
0.53 J.Lm and 0.26 J.Lm at Novette, BulL Amer. Phys. Soc., 29: 1183.
Kaw, P.K. and Dawson, J.M. (1969). Laser-induced anomalous heating of a plasma,
Phys. Fluids, 12: 2586.
Keck, R.L., Craxton, R.S., Goldman, L.M., Seka, W., and Soures, J.M. (1981). Con-
tinuum x-ray measurements of laser plasmas generated by 0.35 J.Lm irradiation,
BulL Amer. Phys. Soc., 26: 848.
Kidder, R.E. (1963). LRL Laser Research Program, COTM-63-7 (unpublished).
Kidder, R.E. (1971 ). Interaction of intense photon and electron beams with plasmas,
in Proc. of the Internat. School of Physics, Enrico Fermi, Course XLVIII, Physics
of High Energy Density, (P. Caldirola, ed.) Academic Press, New York.
Kilkenny, J. (1981). Transport and particle emission studies, in Annual Report to
the Laser Facility Committee 1981, Rutherford and Appleton Labo., RL-81-040,
p.434.
Kirkpatrick, R.C. (1979). An overview of design space for small fusion targets, Nucl.
Fusion, 19: 69.
Kruer, W.L. (1981). Laser-plasma coupling in reactor-size targets, Comments on
Plasma Phys. and Controlled Fusion, 6: 161.
Kull, H.J. and Anisimov, S.l. (1986). Ablative stabilization of the incompressible
Rayleigh-Taylor instability, Phys. Fluids, 29: 2067.
Lane, S.M., Campbell, E.M., and Bennett, C. (1980). Measurement of DT neutron-
induced activity in glass-microshelllaser fusion targets, Phys. Lett., 37: 600.
Lane, S.M., Cable, M.D., Prussin, S.G., Glendinning, S.G., Munro, D.H., Hatchett,
S.P., Estabrook, K.G., Suter, L.J., and Richardson, M.C. (1986). High Yield direct
drive experiments on Nova, Rev. Sci. Inst., 57: 2100.
Lehmberg, R.H. and Obenschain, S.P. (1983). Use of induced spatial incoherence
for uniform illumination of laser-fusion targets, Opt. Commun., 46: 27.
Lindl, J.D. (1977). Low aspect ratio double shell targets for high density and high
gain and a comparison with ultra thin shells, Lawrence Livermore Labo. Rep.,
UCRL-79735 (Rev. 1).
LLEAnnual Report (1984). Absorption and hot electron generation measurements
with spherically illuminated targets at A= 351 nm, DOE/DP40124-67, UC-21, p.
11.
LLE Annual Report (1985). X-ray conversion measurements from high-Z targets,
DOE/DP40200-05, UC 21, p. 60.
Luciani, J.F., Mora, P., and Virmont, J. (1983). Nonlocal heat transport due to steep
temperature gradients, Phys. Rev. Lett., 51: 1664.
Maaswinkel, A.G.M., Eidmann, K., Sigel, R., and Witkowski, S. (1984). Compara-
tive study of laser acceleration of thin foils at wavelengths 0.44 I-'m and 1.3 flm,
Opt. Commun., 51: 255.
Malone, R.C., McCrory, R.L., and Morse, R.L. (1975). Indications of strongly flux-
limited electron thermal conduction in laser-target experiments, Phys. Rev. Lett.,
34: 721.
Mason, R.J. (1975). The calculated performance of structured laser fusion pellets,
NucL Fusion, 15: 1031.
Mason, R.J. (1981). Apparent and real thermal inhibition in laser-produced plas-
Matte, J.P. and Virmont, J. (1982). Electron heat transport down steep temperature
gradients, Phys. Rev. Lett., 49: 1936.
Matte, J.P., Johnston, T.W., Delettrez, J., and McCrory, R.L. (1984). Electron heat
flow with inverse bremsstrahlung and ion motion, Phys. Rev. Lett., 53: 1461.
Max, C.E., McKee, C.F., and Mead, W.C. (1980). A model for laser driven ablative
implosions, Phys. Fluids, 23: 1620.
McCall, G.H., Young, F., Ehler, A.W., Kephart, J.F., and Godwin, R.P. (1973). Neu-
tron emission from laser-produced plasmas, Phys. Rev. Lett., 30: 1116.
McCrory, R.L. and Morse, R.L. (1977a). Dependence of laser-driven compression
efficiency on wavelength, Phys. Rev. Lett., 38: 544.
McCrory, R.L., Morse, R.L., and Taggart, K.A. (1977b). Growth and saturation of
instability of spherical implosions driven by laser or charged particles, Nucl. Sci.
Eng., 64: 163.
I nertlally Confined Fusion 265
McCrory, R.L., Montierth, L., Morse, R., and Verdon, C. (1981a). Nonlinear evolu-
tion of ablation-driven Rayleigh-Taylor instability, Phys. Rev. Lett., 46: 336.
McCrory, R.L., Montierth, L., Morse, R.L., and Verdon, C.P. (1981b). Taylor insta-
bility in fusion targets, in Laser Interaction and Related Plasma Phenomena, Vol. 5,
(H.J. Schwartz, H. Hora, and B. Yaakobi, eds.), Plenum Press, New York.
Mead, W.C., Campbell, E.M., Estabrook, K.G., Turner, R.E., Kruer, W.L., Lee,
P.H.Y., Pruett, B., Rupert, V.C., Tirsell, K.G., Stradling, G.L., Ze, F., Max, C.E.,
and Rosen, M.D. (1981). Laser-plasma interactions at 0.53 fLm for disk targets of
varying Z, Phys. Rev. Lett., 47: 1289.
Mead, W.C., Campbell, E.M., Estabrook, K.G., Turner, R.E., Kruer, W.L., Lee,
P.H. Y., Pruett, B., Rupert, V.C., Tirsell, K.G., Stradling, G.L., Ze, F., Max, C.E.,
Rosen, M.D., and Lasinski, B.L. (1983). Laser irradiation of disk targets at
0.53 fLm wavelength, Phys. Fluids, 26: 2316.
Mead, W.C., Campbell, E.M., Kruer, W.L., Turner, R.E., Hatcher, C.W., Bailey,
D.S., Lee, P.H.Y., Foster, J., Tirsell, K.G., Pruett, B., Holmes, N.C., Trainor, J.T.,
Stradling, G.L., Lasinski, B.F., Max, C.E., and Ze, F. (1984). Characteristics of
lateral and axial transport in laser irradiations of layered-disk targets at 1.06 and
0.35 fLm wavelengths, Phys. Fluids, 27: 1301.
Mead, W.C., Coggeshall, S.V., Goldman, S.R., Stover, E.K., Goldstone, P.D., Hauer,
A., Kindel, J.M., Cobble, J., Stradling, G., Barnouin, 0., Jaanimagi, P., Marshall,
F., Marjoribanks, R., Richardson, M.C., Kauffman, R., and Kornblum, R. (1987).
Analysis, modeling, and design of short-wavelength laser-plasma experiments, in
Laser Interaction and Related Plasma Phenomena, Vol. 7, (H. Hora and G. Miley,
eds.), Plenum Press, New York.
Miyanaga, N., Azechi, H., Stapf, R.O., Itoga, K., Nakaishi, H., Shiraga, H., Ya-
manaka, M., Yamanaka, T., Tsuji, R., ldo, S., Sakurai, K., Nishihara, K., Yabe,
T., Takagi, M., Nakatsuka, M., Izawa, Y., Nakai, S., Yamanaka, C., Kobayashi,
K., Kimura, 1., and Morinobu, S. (1986). Radiochemistry and secondary reac-
tions for the diagnostics of laser-driven fusion with plasmas, Rev. Sci. Inst., 57:
1713.
Mochizuki, T., Yabe, T., Okada, K., Hamada, M., Ikeda, N., Kiyokawa, S., and Ya-
manaka, C. (1986). Atomic-number dependence of soft x-ray emission from var-
ious targets irradiated by a 0.53 fLm wavelength laser, Phys. Rev. A, 33: 525.
Montierth, L., Cochran, F.L., and Morse, R.L. (1979). Analysis of the stability of
ablation driven implosions with the steady flow model, Bull. Amer. Phys. Soc., 24:
945.
Nishimura, H., Azechi, H., Yamada, K., Tamura, A., Inada, Y., Matsuoka, F.,
Hamada, M., Suzuki, Y., Nakai, S., and Yamanaka, C. (1981). Experimental study
of wavelength dependences of laser-plasma coupling, transport, and ablation
processor, Phys. Rev. A, 23: 2011.
Nuckolls, J.H., Wood, L., Thiessen, G., and Zimmerman, G. (1972). Laser compres-
sion of matter to super-high densities: Thermonuclear (CTR) applications, Na-
ture, 239: 139.
Nuckolls, J.H. (1982). The feasibility of inertial confinement fusion, Physics Today,
35:24.
Priedhorsky, W., Lier, D., Day, R., and Gerke, D. (1981). Hard x-ray measurements
of 10.6 fLm laser-irradiated targets, Phys. Rev. Lett. 47: 1661.
Richardson, M.C., Craxton, R.S., Delettrez, J., Keck, R.L., McCrory, R.L., Seka,
W, and Soures, J.M. (1985). Absorption physics at 351 nm in spherical geometry,
Richardson, M.C., Keck, R.F., Letzring, S.A., McCrory, R.L., McKenty, P.W, Ro-
back, D.M., Soures, J.M., and Verdon, C.P. (1986a). Neutron diagnosis of com-
pressed ICF targets, Rev. Sci. Inst., 57: 1737.
Richardson, M.C., McKenty, P.W, Keck, R.L., Marshall, F.J., Roback, D.M., Ver-
don, C.P., McCrory, R.L., and Soures, J.M. (1986b). High-aspect-ratio laser-
fusion targets driven by 24-beam UV laser radiation, Phys. Rev. Lett., 56: 2048.
Ripin, B.H., Decoste, R., Obenschain, S.P., Bodner, S.E., McLean, E.A., Young,
F.C., Whitlock, R.R., Armstrong, C.M., Grun, J., Stamper, J.A., Gold, S.H.,
Nagel, D.J., Lehmberg, R.H., and McMahon, J.M. (1980). Laser-plasma interac-
tion and ablative acceleration of thin foils at 1012-1015 W/cm 2 , Phys. Fluids, 23:
1012.
Rosen, M.D. (1984). Evidence of a laser intensity threshold for the onset of inhibited
electron transport, Commun. on Plasma Phys. and Cont. Fusion, 8: 165.
Rosen, M.D., Phillion, D.W., Rupert, V.C., Mead, WC. Kruer, WL., Thomson, J.J.,
Kornblum, H.N., Slivinsky, V.W, Caporaso, G.J., Boyle, M.J., and Tirsell, K.G.
(1979). The interaction of 1.06 J-Lm laser radiation with high Z disk targets, Phys.
Fluids, 22: 2020.
Seka, W, Jacobs, S.D., Rizzo, J.E., Boni, R., and Craxton, R.S. (1980). Demonstra-
tion of high efficiency third harmonic conversion of high power Nd:glass laser
radiation, Opt. Commun. 34:469.
Seka, W., Craxton, R.S., Delettrez, J., Goldman, L., Keck, R., McCrory, R.L.,
Shvarts, D., Soures, J.M., and Boni, R. (1982). Measurements and interpretation
of the absorption of 0.35 J-Lm laser radiation on planar targets, Opt. Commun.
Shearer, J.W., Mead, S.W, Petruzzi, J., Rainer, F., Swain, J.E., and Violet, C.E.
(1972). Experimental indications of plasma instabilities induced by laser heat-
ing, Phys. Rev. A, 6: 764.
Shiau, J.N., Goldman, E.B., and Weng, C.I. (1973). Linear stability analysis of laser-
driven spherical implosions, Phys. Rev. Lett., 32: 352.
Shvarts, D., Delettrez, J., McCrory, R.L., and Verdon, C.P. (1981). Self-consistent
reduction of the Spitzer-Harm electron thermal heat flux in steep temperature
gradients in laser-produced plasmas, Phys. Rev. Lett., 47: 247.
Simon, A. and Short, R.W (1984). New model for Raman spectra in laser-produced
plasma, Phys. Rev. Lett., 53: 1912.
Simon, A., Seka, W., Goldman, L.M., and Short, R.W (1986). Raman scattering in
inhomogeneous laser-produced plasma, Phys. Fluids, 29: 1704.
Skupsky, S. and Kacenjar, S. (1981). Measuring fuel pR for inertial fusion experi-
ments using neutron elastic-scattering reactions, J. Appl. Phys., 52: 2608.
Skupsky, S. and Lee, K. (1983). Uniformity of energy deposition for laser driven
fusion, J. Appl. Phys., 54: 3662.
Slater, D.C., Busch, G.E., Charatis, G., Johnson, R.R., Mayer, F.J., Schroeder, R.J.,
Simpson, J.D., Sullivan, D., Tarvin, J.A., and Thomas, C.E. (1981). Absorption
and hot-electron production for 1.05 J-Lm and 0.53 J-Lm light on spherical targets,
Slivinsky, V.W., Kornblum, H.N., and Shay, H.D. (1975). Determination of supra-
thermal electron distributions in laser-produced plasmas,!. Appl. Phys., 46: 1973.
Soures, J.M., Goldman, L.M., and Lubin, M.J. (1973). Short-pulse-laser-heated
plasma experiments, Nucl. Fusion, 13: 829.
Soures, J.M., Hutchison, R.J., Jacobs, S.D., Lund, L.D., McCrory, R.L., and Rich-
ardson, M.C. (1983). Omega: A short-wavelength laser for fusion experiments,
Proc. of lOth Symp. on Fusion Eng., IEEE, New York, p. 1392.
Spitzer, L., Jr. (1962). Physics offully ionized gases, Wiley-Interscience, New York,
Chap. 5.
Spitzer, L. and Harm, R. (1953). Transport phenomena in a completely ionized gas,
Phys. Rev., 89:977.
Stamper, J.A., Papadopoulos, K., Sudan, R.N., Dean, S.O., McLean, E.A., and Daw-
son, J.M. (1971). Spontaneous magnetic fields in laser-produced plasmas, Phys.
Rev. Lett., 26: 1012.
Stamper, J.A., Barr, O.C., Davis, J., Doschek, G.A., Dozier, C.M., Feldman, U.,
Klein, B.M., Manheimer, W.M., McLean, E.A., McMahon, J.M., Nagel, D.J.,
Tidman, D.A., Whitlock, R.R., Whitney, K., Winsor, N.K., and Young, F.C.
(1974). Laser-matter interaction studies at NRL, Laser Interaction and Related
Plasma Phenomena, Vol. 3B (H.J. Schwarz and H. Hora, eds.), Plenum Press,
New York, p. 713.
Tanaka, K., Goldman, L.M., Seka, W., Richardson, M.C., Soures, J.M., and Wil-
liams, E.A. (1982). Stimulated Raman scattering from UV laser-produced plas-
Tanaka, K.A., Boswell, B., Craxton, R.S., Goldman, L.M., Guglielmi, F., Seka, W.,
Short, R.W., and Soures, J.M. (1985). Brillouin scattering, two-plasmon decay,
and self-focusing in underdense ultraviolet laser-produced plasmas, Phys. Fluids,
28:2910.
Tarvin, J.A., Fechner, W.B., Larsen, J.T., Rockett, P.D., and Slater, D.C. (1983).
Mass-ablation rates in a spherical laser-produced plasma, Phys. Rev. Lett., 51:
1355.
Turner, R.E., Estabrook, K., Kauffman, R.L., Bach, D.R., Drake, R.P., Phillion,
D.W., Lasinski, B.F., Campbell, E.M., Kruer, W.L., and Williams, E.A. (1985).
Evidence for collisional damping in high-energy Raman-scattering experiments
at 0.26flm, Phys. Rev. Lett., 54: 189.
Verdon, C.P., McCrory, R.L., Morse, R.L., Baker, G.R., Meiron, D.I., and Orszag,
S. (1982). Nonlinear effects of multifrequency hydrodynamic instabilities on ab-
latively accelerated thin shells, Phys. Fluids, 25: 1653.
Widner, M. (1979). Fuel energy balance studies of pellet ignition requirements, San-
dia Lab. Rep. SAND79-2454, Albuquerque, New Mexico.
World Survey of Major Activities in Controlled Nuclear Fusion Research, (1986). In-
ternat. Atomic Energy Agency, Vienna.
Yaakobi, B., Pelah, I., and Hoose, J. (1976). Preheat by fast electrons in laser-fusion
experiments, Phys. Rev. Lett., 37: 836.
Yaakobi, B. and Bristow, T.C. (1977). Measurement of reduced thermal conduction
in (layered) laser-target experiments, Phys. Rev. Lett., 38: 350.
Yaakobi, B., Skupsky, S., McCrory, R.L., Hooper, C.E, Deckman, H., Bourke,
P., and Soures, J.M. (1980). Direct measurement of the compression of laser-
imploded targets using x-ray spectroscopy, Phys. Rev. Lett., 44: 1072.
Yaakobi, B., Boehly, T., Bourke, P., Conturie, Y., Craxton, R.S., Delettrez, J.,
Forsyth, J.M., Frankel, R.D., Goldman, L.M., McCrory, R.L., Richardson, M.C.,
Seka, W., Shvarts, D., and Soures, J.M. (1981a). Characteristics of target inter-
action with high power UV laser radiation, Opt. Commun., 39: 175.
Yaakobi, B., Bourke, P., Conturie, Y., Delettrez, J., Forsyth, J.M., Frankel, R.D.,
Goldman, L.M., McCrory, R.L., Seka, W., and Soures, J.M. (1981b). High x-ray
conversion efficiency with target irradiation by a frequency tripled Nd:glass laser,
Opt. Commun., 38: 196.
Yaakobi, B., Delettrez, J., Goldman, L.M., Majoribanks, R., McCrory, R.L., Rich-
ardson, M.C., Shvarts, D., Skupsky, S., Soures, J.M., Verdon, C., Villeneuve,
D.M., Boehly, T., Hutchison, R., and Letzring, S. (1984). Thermal transport mea-
surements in 1.05 J.lffi laser irradiation of spherical targets, Phys. Fluid, 27: 516.
Yaakobi, B., Barnouin, 0., Delettrez, J., Goldman, L.M., Marjoribanks, R., Mc-
Crory, R.L., Richardson, M.C., and Soures, J.M. (1985). Thermal transport mea-
surements in six-beam, ultraviolet irradiation of spherical targets, J. Appl. Phys.,
57:4354.
Yamanaka, C., Yamanaka, T., Sasaki, S., Yoshida, K, Waki, M., and Kang, H. (1972).
Anomalous heating of a plasma by a laser, Phys. Rev. A, 6: 2335.
Yamanaka, C., Nakai, S., Nishimura, H., Uchida, S., Izawa, Y., Norimatsu, T.,
Miyanaga, N., Azechi, H., Nakai, M., Takabe, H., Jitsuno, J., Mirna, K, Nakat-
suka, M., Sasaki, T., Yamanaka, M., Kato, Y., Mochizuki, T., Kitagawa, Y., Ya-
manaka, T., and Yoshida, K (1986). Laser implosion of high-aspect targets pro-
duces thermonuclear neutron yield exceeding 1012 by use of shock multiplexing,
Yates, M.A., van Hulsteyn, D.B., Rutkowski, H., Kyrala, G., and Brackbill, J.U.
(1982). Experimental evidence for self-generated magnetic fields and remote en-
ergy deposition in laser-irradiated targets, Phys. Rev. Lett., 49: 1702.
Young, EC., Whitlock, R.R., Decoste, R., Ripin, B.H., Nagel, D.J., Stamper, J.A.,
McMahon, J.M., and Bodner, S.E. (1977). Laser-produced plasma energy trans-
port through plastic films,AppL Phys. Lett., 30: 45.
Ze, E, Suter, L.J., Lane, S.M., Campbell, E.M., Lindl, J., Rosen, M.D., Phillion,
D.W., Hatcher, C.W., Drake, R.P., Hildum, J.S., and Manes, KR. (1984). Wave-
length scaling of laser-driven implosions, Lawrence Livermore Nat. Lab. Rep.
UCRL-91087.
6
Laser-Based Semiconductor
Fabrication
Joseph R. Wachter
Materials Science and Technology Division
6.1 ASPECTS OF SEMICONDUCTOR

FABRICATION
Research on the use of lasers in the semiconducting industry has been evolv-
ing for more than 20 years. Some of the earliest research of potential benefit
to the industry involved the engraving of alphanumeric symbols on various
materials using focused laser light (Garman and Ponce, 1986). A few years
later, an investigation was undertaken on the use of lasers in the formation
of alloy diodes (Fairfield and Schwuttke, 1968). Still later, research con-
centrated on lasers for altering, repairing, and processing semiconductor
material for device fabrication, e.g., laser-assisted removal of ion implanta-
tion damage sites in integrated circuits (Shtyrkov et al., 1975) and for im-
purity gettering (Pearce and Zaleckas, 1979). Another early use of lasers
in semiconductor device fabrication was for resistor trimming in hybrid cir-
cuits. Here, printed resistor films are micromachined to desired values with
a laser beam. In the 1980s, research has concentrated on methods for di-
rect processing of integrated circuits (IC's), such as defect reduction in epi-
taxial silicon, large grain polysilicon growth, laser-assisted etching and film
deposition methods, and removal of dislocation networks. These and other
topics will be considered in more detail in the following pages.
Currently, lasers are employed in only a limited number of ways in semi-
conductor production facilities. However, trends suggest that more appli-
cations will be found in the foreseeable future because of the advantages
offered by laser processing over conventional techniques. Laser beams can
be focused to small spots and precisely positioned in space to form the small
feature sizes necessary for modern integrated circuitry. They offer a clean
269
270 Wachter
energy source that can treat semiconducting materials without the use of
tools, masks, or chemicals. Lasers are also able to process materials at lower
bulk temperatures than conventional furnace annealing, thereby minimiz-
ing damage to the water. Additionally, the absorption coefficients of many
materials in semiconductor devices can be selectively exploited by lasers of
different wavelengths to provide a variety of printing, repairing, and other
functions. Finally, laser processing may lower manufacturing costs and pro-
vide greater flexibility for production of small quantities of individualized
circuitry.
6.1.1 Device Fabrication
In order to best appreciate the potential impact of lasers in semiconductor

processing facilities, an overview of the major steps involved in the fabrica-
tion of a simplified semiconductor circuit will first be presented. An epitax-
ial oxide layer is initially grown on a silicon substrate prior to incorporation
of doped regions. These active regions are produced by ion implantation
or diffusion processes. The undoped regions of the substrate form the sur-
face on which first, thin resistor films and second, conductive tracts among
circuit elements are deposited. The resistor films are deposited by sputter-
ing, evaporation, or chemical vapor deposition and may range in thickness
from 10 to 200 nm. Prior to deposition of the metal tracks, a contact metal
is set down on the resistor film. This ensures good ohmic contact between
the metal and film and may, in addition, provide electrical contact to one or
more of the active regions. Resistor values are then established by a series of
printing techniques such as etching and photolithography. After patterning
of the resistors, the interconnect metal, generally aluminum, is deposited
and additional printing steps further define the circuit. Furnace annealing
after this step serves to stabilize the structure. To construct multilayer cir-
cuits, insulating layers of an oxide or silicon nitride are first deposited on
all parts of the original circuit; then further dielectric film deposition and
metallization can take place. Finally, the circuit is tested to verify that the
device meets manufacturing specifications.
This outline of the fabrication steps in the production of semiconductor
circuits, although illustrative, does not give full measure to the intricacies
involved in modern IC or very large-scale integration (VLSI) device produc-
tion. Figure 6.1 is a schematic representation of a metal-oxide semiconduc-
tor (MOS) memory device that brings out dramatically the small feature
size, high-device density, and complexity found in modern microcircuitry.
Many of the features in the device have dimensions of 1 Jtm or less, and the
complexity of the circuit elements hints at the precise machining tolerances
needed for its fabrication.
Semiconductor Fabrication 271
p-Si
IIllJ
•
ALUMINUM
ISOPLANARISING OXIDE POL YSILICON
~
~GATE OXIDE NITRIDE TUNNEL OXIDE
~ CVD OXIDE 6'16 PHOSPHOSILICA TE

GLASS
0 3-6~
PHOSPHOSILICATE
GLASS
Figure 6.1 Diagram of an MOS memory circuit. Submicron to several micron fea-
ture sizes are typical in modern IC circuitry (from Boyd, 1984).
6.1.2 Processing Characteristics

of Semiconductor Materials
The response of semiconductor materials to laser irradiation depends both

on the intrinsic properties of the material and the laser parameters. Impor-
tant laser parameters include beam diameter, wavelength, divergence, in-
tensity, timing characteristics, polarization, and angle of incidence. Beam
diameter, wavelength, and divergence establish the spatial distribution of
laser energy on the material surface, and by determining the laser spot size,
primarily control the lateral dimensions of the processing. Laser inten-
sity and timing characteristics (pulse duration, repetition rate) determine
the amount of power delivered to the material and more directly influence
how deeply the beam penetrates the surface. Wavelength, polarization,
and angle of incidence affect the fraction of incident light that will be re-
flected from the surface, and hence the amount of light available for pro-
cessing. Lasers useful for microelectronics applications include those oper-
ated in pulsed and Q-switched modes as well as continuous wave and have
ranged in wavelength from 0.4 to 10 p,m. By far the most widely employed
is the Nd:YAG laser operated at 1.06 p,m. However, research with lasers
at shorter wavelengths (including the frequency-doubled Nd:YAG line at
272 Wachter
0.53 llm) is increasing because modern optical delivery systems permit fo-
cusing of the beam to spot sizes only slightly larger than the wavelength.
Because lasers alter semiconductor materials by raising their tempera-
ture, the absorption characteristics of the material are also important to
an understanding of laser processing. Figure 6.2 shows the absorption co-
efficient for crystalline and amorphous silicon along with the wavelengths
of some lasers used in semiconductor fabrication. At wavelengths shorter
than about 10 flm, silicon absorbs strongly at room temperatures and can
be processed by a variety of laser wavelengths.
10 7 60
Ar+ Nd:VAG
-
(514nm) (1064nm)
RUBY
,... (694nm)
I
-
108 50
E
0 n=5x1o21cm-3
1-
z
~
105
AMORPHOUS Si
l;x1o21c~~ 40
--
?fl.
(J
>-
u:: !::
u.
UJ ,' ;I' >
0 / I ;I' 1-
/, I
(J
(J
z 104 . 30
w
...J
Q "'- "-. / I ./ u.
w
a:
1-
a.
a:
\ "'- ,Y.. ........._/_.,.<= 1 x 1o21 cm-3
0
en
m
103
\ "\-'
\ '-, , " __. /
I
20
<
10 2 10
500 1000 1500 2000
WAVELENGTH (nm)
Figure 6.2 The solid lines indicate the absorption coefficients for crystalline and
amorphous silicon shown as functions of wavelength. The dotted lines indicate the
reflectivity of silicon after p + implantation for three carrier concentrations (n ). The
figure reveals two important considerations for laser processing of semiconductors,
the absorption properties of the material and the change in reflectivity due to doping
(from Brown, 1983).
Differences in the absorption coefficients of two neighboring materials

in an integrated circuit can determine the type of laser to be used because
selective absorption allows heating of particular materials in well-defined
regions of the wafer to take place. For example, localized regions of amor-
phous silicon embedded in Si crystal could be annealed by a Nd:YAG laser
at 1.06 ~-tm because the crystalline material does not absorb as strongly at
this wavelength.
During processing, the absorption properties of the material may change.
As the temperature of the material rises during processing, its absorptiv-
ity will be altered. In addition, surface modification and phase changes
may occur when the material is heated and these factors will also affect
the absorptivity. If vaporization takes place, the emitted particles may
strongly absorb the incident beam, again altering the absorption of light
by the target material. The degree of doping of semiconductors has also
been shown to affect their reflectivity (Miyao et al., 1981). Figure 6.2 also
shows the shift in reflectivity as a function of wavelength for silicon of dif-
ferent doping characteristics. Free carrier excitation processes in which
free electrons or holes are boosted up to higher energy states strongly
influence reflectivity for wavelengths in the infrared. This figure indi-
cates that these excitation processes depend strongly on carrier concentra-
tion.
6.1.3 Laser Processing Systems
Although each laser processing system will differ according to fabrication

requirements and economic considerations, the following parts are typically
included (see Fig. 6.3):
The laser, consisting of the power supply, cooling apparatus, optical modulator
(if appropriate), mirrors, excitation source, active medium, and beam quality-
control monitor.
Beam delivery apparatus containing an optical control head (includes mirrors,
lenses, and motors) for directing the beam to the workpiece. The relative
motion between workpiece and laser beams is accomplished either by move-
ment of the beam or workpiece. The former is used for smaller relative mo-
tion, on the order of a few centimeters, and speeds up to 1000 em/sec have
been realized with these systems. Larger relative motion between the beam
and workpiece generally requires movement of the piece or the entire con-
trol head. Speeds of up to 50 em/sec have been accomplished by this method
(Steffen, 1984). The workpiece and optical control head may be contained in
a vacuum system for some processing tasks to ensure a clean working envi-
ronment where desirable operating conditions can be maintained.
Workpiece positioning system for loading and placement of the workpiece. Po-
sitioning of the workpiece, along with beam delivery and laser settings, can be
274 Wachter
VACUUM SYSTEM
,---------l
I OPTICAL CONTROL I
LASER I HEAD I
LASING MEDIUM
POWER SUPPLY
I LENSES
I
COOLING I MIRRORS
MOTORS I
SYSTEM
OPTICS I I
I
I I I
I
II WORKPIECE POSITIONIN
SYSTEM
G I
I I
0
I I
o
COMPUTER
I
cr I
I
20t::l~ I
I I
I
I
L _ _ _ _ _ _ _ _ _ _j
I
Figure 6.3 Schematic of laser processing system for semiconductor device fabrica-
tion.
automated and/or computer-controlled to best tailor the processing system to

fabrication requirements.
Laser systems can provide flexible, high-quality processing of semicon-

ductor materials. Energy, focusing, and timing characteristics can be
quickly readjusted to handle different processing tasks and materials. The
beam can be subdivided into several beams to perform multiple functions
and directed to parts of the wafer that are difficult to access by other means.
The economic advantages of laser processing systems are determined by
specific application requirements and the relative costs and advantages of
alternate processing systems. For laser systems, investment costs for the
laser and beam delivery apparatus may be substantial; however, operating
costs, such as flash-lamp replacement, are low.
VLSI circuitry may contain up to 10,000 transistors, and as component
density increases to provide improved devices, fabrication costs will increas-
ingly depend on improvements in processing systems to meet these require-
(/)
CD
3
8
:I
c.
1:
Sl
0
...
Thble 6.1 Some Laser Processing Applications and Operating Parameters Used for Semicon- "11
ductor Fabrication Ill
C'
...
Focused Beam &
Irradiance Diameter 0
:I
Application Laser (W/cm 2 ) (pm) Reference
Annealing Nd:YAG, ruby 106 Varies Gotzlich and Ryssel, 1984

XeF, XeCI
Resistor trimming
Thick film Nd:YAG 105-106 25-50 Crawley, 1984
Thin film Nd:YAG 105-106 25-50 Crawley, 1984
Link blowing Nd:YAG 107 5 Swenson, 1983
Photomask repair Nd:YAG,Ar+ 102-105 1 Preiswerk et al., 1986
Wafer marking C0 2 105 Varies Higgins, 1986
Thin-film deposition C0 2 , Varies Varies Boyd, 1984
various excimer
Etching C0 2 , Ar+, ArF 100-1010 1pm-fewmm Houle, 1984
1\)
.....
(11
276 Wachter
ments. Factors that will continue to have a major impact on component

improvements include the following:
Reductions in circuit feature size to sub micron dimensions.
Vertical scaling techniques, such as film deposition methods, to achieve proper
electrical characteristics and physical dimensions in multilayer circuits.
Lower-temperature operations to prevent heat-induced wafer damage.
Advancements in printing methods including maskless circuit writing.
Continued developments in circuit redundancy.
Laser processing may provide solutions to meet some, or all, of these

challenges. A partial listing of laser systems in use or envisioned for semi-
conductor fabrication is shown in Table 6.1. These systems and others will
be described in greater detail throughout this chapter. Section 6.2 covers
current applications of laser processing in semiconductor device fabrica-
tion. These applications will be described along with equipment, operating
parameters, and effectiveness of the laser techniques. In Sec. 6.3, atten-
tion will be directed to research areas for new production applications. Al-
though the economic and engineering advantages of laser production have
not yet been proven for these areas, it is probable that several will find ap-
plication among semiconductor manufacturers. Finally, Sec. 6.4 will briefly
discuss the future of laser processing of semiconductors.
6.2 APPLICATIONS OF LASERS IN THE

SEMICONDUCTOR INDUSTRY
Currently, lasers have found only a limited number of uses in the fabrication
of integrated circuitry. These uses include laser trimming, wafer labeling,
redundancy, and mask repairs. This section will describe these uses in detail
and the products that have been processed with laser techniques. In gen-
eral, laser processing in these areas has proven superior to conventional
techniques because lasers can deliver spatially and temporally controlled
bursts of power to small regions of the workpiece.
6.2.1 Resistor 1iimming
One of the earliest applications of laser processing in the semiconductor

industry was the adjustment, or trimming, of printed film resistors in elec-
tronic circuits. These films, typically rectangular in shape and deposited
on an Si02 substrate, may be connected to other electronic components
such as transistors and capacitors to form hybrid circuits. These circuits are
useful in such devices as miniature switches, potentiometers, and resistor
networks. Many modern analog-to-digital and digital-to-analog convertors

could not be built without these precisely trimmed devices.
There are two main categories of resistor films: thick and thin film.
Thick-film resistors average 'about 25 pm in thickness and are applied as
a paste onto the substrate. They may be found on both single and multi-
layer circuits. Thin resistor films may be deposited by sputtering, evapo-
ration, or chemical vapor deposition and are as little as 0.1 pm thick. Al-
though more expensive to fabricate, thin-film circuits are superior to thick
films in two important respects: electrical stability and accuracy. Only thin-
film circuitry can perform without significant deterioration over a 25-year
lifetime (Crawley, 1984) as required by some military and telecommunica-
tions markets. Accuracy in resistance values to within 0.1% of designed
values has been achieved by thin-film deposition methods. In this regard,
laser trimming techniques have proven particularly valuable for thick-film
adjustments because they are able to cut away precise amounts of film. The
exact resistance values thus obtained can be manufactured at lower cost
than thin films while realizing nearly equivalent accuracies. For both types,
resistance values are determined by three mechanisms: resistor material,
thickness, and area. By using these forms of control, a wide range of values
can be obtained. By controlling the chemistry and thickness of the film, im-
portant parameters such as temperature and drift stability, sheet resistivity,
and conductivity are determined.
Mter a network of resistors and conductors is laid out, parts of the film
are removed to establish final resistance values. The trimming of resistors is
performed by ablating small portions of film from the substrate using pulsed
laser energy. The vaporized material is cut away in fine line segments until
the desired resistance value is realized. For larger changes in resistor values,
links may be cut out of the film or a series of overlapping pulses can be
used to make a continuous cut length. Wide field lasers presently allow
circuit trimming of areas as large as 8 in 2 (Wagner, 1986). Trimming can be
performed in a passive mode while the circuit is inoperative, or in an active
mode when connected with other elements in the circuit to test devices that
measure resistor output under designed load conditions. Table 6.2 indicates
typical laser parameters used in trimming operations.
Laser trimming systems in use in the semiconductor industry are com-
posed of a Q-switched Nd:YAG laser generally operated at 1.06 pm, op-
tics for directing and focusing the beam, a device handler for positioning
the device under the beam, circuit measurement instrumentation, and a
computer controller. Nd:YAG lasers are favored for resistor trimming be-
cause of their proven pulse-to-pulse stability, high-peak power, and reli-
ability. These laser characteristics are necessary for high-cut quality and
stable, defect-free resistor performance. Two movable mirrors under com-
puter control are used for thex,y positioning ofthe beam on the workpiece.
278 Wachter
Thble 6.:Z Typical Laser Beam Attributes

Used for Resistor Trimming (Wagner,
1986)
Attribute Thick film Thin film
Laser type Nd:YAG Nd:YAG

Power(W) 1-10 0.01-0.5
Pulse width (nsec) 70-350 25-200
Spot size (p.m) 25-50 5-25
Both linear translation motors and galvanometers are used for this pur-
pose. The device handler loads and unloads the substrates on which the
film is deposited. Various handlers have been used including carousels, slide
handlers, and step and repeat tables. The handlers must accurately place
substrates rapidly enough to match other operations in the laser trimming
process. Circuit measurement instrumentation measures small changes in
resistance as portions of the film are vaporized by the laser. Measurement
may be undertaken in either the active or passive mode as described pre-
viously. Finally, computers and appropriate software are used to provide
sequence control for trim and test operations, device-edge sensing for loca-
tion of starting and reference points, laser- and beam-positioning parame-
ters, circuit measurement and program modification.
Laser trimming offers several advantages over other trimming methods.
Precise beam focusing and positioning allow spot sizes of less than 4 f.-tiD,
positioning accuracies of 1 f.-tiD (Wagner, 1986), and positioning speeds as
low as 2 sec per function (Mueller, 1986), important qualities for trimming
congested circuits. Laser trimming is a clean process and leaves behind
little material residue. Computer control of laser trimming systems pro-
vides improved trim accuracy, speed, reproducibility, and design flexibility.
Where they have been installed, laser trimmers have proven to be highly
cost-effective compared to alternative methods.
6.2.2 Redundancy Repair
Today's IC circuitry is marked by increased chip complexity. This has been

achieved by shrinking the size of circuit components while increasing the
total chip area. Smaller, more compact circuits are useful because smaller
devices switch faster and provide shorter propagation delay times. Larger
chip areas have allowed the number of components on a single chip to in-
crease. As chip complexity increases, however, the likelihood of finding a
disabling defect in the IC also increases. For example, in initial production
runs of some large memory arrays, more than 90% of the chips were re-
jected because of defects (Swenson, 1983). To overcome these low yields,
spare rows and columns are added to the chip to replace those in which
faults are detected. This redundancy in the circuit has allowed dramatic in-
creases in chip complexity to continue. Today, most 64 and 256K dynamic
random access memories (DRAM's) incorporate some form of redundancy
in their manufacture.
Repair is carried out by first testing the circuit for defective elements in
the array. On determination of a defect, a fuse link is blown either electri-
cally or using a single burst of laser radiation. The laser bursts disconnect
the faulty element by vaporizing the material. Spare links can then be en-
coded in the circuit to replace those that have been disconnected. Links in
memory chips are typically 2 to 3 J.tm in width, are separated from each other
by 5 to 10 J.tm and are made of heavily doped polysilicon, silicides, or polysil-
icon sandwich structures. To accurately decouple links, laser beams must
be focused to approximately the separation distance between links and po-
sitioned to within a few microns of the link center. Alignment requirements
are frequently satisfied by nondestructive measurements of reflected laser
energy. The methods must be capable of positioning the beam to within 2
to 3 J.tm of the target links.
Nd:YAG lasers operated at 1.06 J.tm are generally used for redundancy
repair because the polysilicon link material strongly absorbs at this wave-
length. Since the underlying silicon substrates are more transparent in this
region of the spectrum, alignment requirements are also eased. Use of
shorter wavelengths allows smaller focal spot sizes, but because the sub-
strate better absorbs this radiation, better alignment is required. Some re-
search has occurred using the frequency-doubled Nd:YAG line at 0.53 J.tm
(Parker, 1986). Single Q-switched pulses are used for link blowing as
the short pulse widths are less likely to cause substrate damage. Laser
pulse powers are determined by link thickness and material as well as
the composition of surrounding layers. Generally, a few millijoules per
pulse is sufficient to sever the link without damaging the underlying sub-
strate.
Laser blowing of memory chip links has two important advantages over
electrical methods (Wills and McPherson, 1983). First, smaller chip areas
are affected by the laser process and second, links can be blown under phos-
phosilicate glass (PSG), thereby saving a masking step.
Laser link blowing is now an accepted method of repair for large mem-
ory circuits. Increasingly, the defect-sensitive arrays of MOS transistor and
capacitor cells found in 64 and 256K DRAM's are designed for laser repair.
As manufacturers of memory chips continue to shrink feature sizes and in-
crease circuit complexity, corresponding improvements in beam-spot size
and positioning accuracy will be required.
280 Wachter
6.2.3 Photomask Repair
Since the late 1970s, many semiconductor processing facilities have turned
to small field, step and repeat imaging systems to engrave patterns onto
wafers. In these systems, light is projected through a mask made of glass
and an opaque material such as chrome to form the pattern for a single IC
layer. The opaque material prohibits light from reaching the photosensitive
surface and thereby prevents exposure of specified portions of the wafer to
form the pattern. The step and repeat system then repositions the wafer so
that duplicate images can be formed on new regions of the wafer. These
systems are capable of producing smaller feature sizes and more accurate
registration than techniques such as contact printing or conventional pro-
jection printing. Although they are relatively large, costly, and have lower
throughput than other printing techniques, the production of VLSI circuitry
containing large numbers of circuit elements often requires their use.
Because a single defect in the mask can blemish all the die (an individ-
ual IC produced by separating multiple IC's that have been fabricated on
a wafer) on a wafer, manufacturers are very interested in obtaining defect-
free masks. The masks are generally produced from a photographic master
copy (the reticle) of a single layer pattern by permitting incident light to
expose a resist.
Defects fall into two categories (Parker, 1986); the first consists of
opaque features that appear on transparent parts of the mask. Defects of
this type, which may be due to dirt particles or incomplete etching of the
chrome, can usually be repaired by relatively simple laser vaporization tech-
niques. More difficult to repair are defects in the secondary category, small
holes in parts of the mask that should be opaque. For these defects, a small
amount of opaque material must be deposited on the reticle to cover the
hole. One method of repair for this type of defect is that of laser writing in
photoresist. In this method, the glass substrates are covered with a light-
sensitive resist, then a laser beam is brought to bear on the resist to expose
the pattern. Laser writing in photoresist has been accomplished with cw
argon ion lasers (Preiswerk et al., 1986) to locally repair transparent micro-
faults. The technique is capable of generating submicron gold and platinum
patterns under ambient atmosphere conditions.
Mask repair using laser pattern generation has been shown to be com-
petitive with conventional mask fabrication techniques such as optical pat-
tern generation (nonlaser). In this regard, the versatility oflaser techniques
frequently makes it a more cost-effective alternative. For example, a sys-
tem for fabricating photomasks using both laser ablation and laser writing
in photoresist techniques was recently described by Crowley (1986). The
computer-controlled system uses a Q-switched Nd:YAG laser to microma-
chine thin chrome films and a cw argon ion laser for photoresist exposure.
In addition to its mask fabrication capabilities, the Nd:YAG laser can also
perform resistor trimming functions.
6.2.4 Device Marking
Laser-marking systems have evolved because of improvements in laser reli-

ability and optical steering mechanisms and are now used to imprint a wide
variety of materials including glass, plastic, metals, ceramics, and wood.
They can be used to identify products with trademarks, logos, dot ma-
trix lettering, and bar codes. In the semiconductor and electronics indus-
tries, lasers are used to mark diodes, capacitors, transistors, printed circuit
boards, wafers, hybrids, and IC's. In addition to simply identifying prod-
ucts, laser marking can also be used to prevent process control errors by
including product information with the label.
Pulsed C02 lasers have been used most often for marking semiconductor
devices because the primary wavelength of these lasers is useful for marking
glass. Radiation from N d: YAG lasers, on the other hand, is mostly reflected
by glass and these lasers are less useful. However, C0 2 lasers are generally
not suited for marking on metal. As with other laser fabrication processes,
beam power and repetition rate must be combined with the focusing op-
tics and the thermal and reflectivity properties of the targeted material to
produce optimal marking conditions.
In general, lasers are able to mark materials by a variety of mechanisms:
heating, melting, vaporization, chemical transformation, and burning (such
as localized carbonization of plastics). However, in the semiconductor
industry, laser marking is usually restricted to vaporization and chemical
transformation (Higgins, 1986), the latter being more widely used. For this
case, a laser beam is directed onto the surface of the device or wafer, pro-
ducing a color change in the material. Typically, the mark will penetrate
some tens of microns below the surface. Marking by vaporization is used
more often if the material has weak absorption at the laser wavelength.
With this method, a thin slice of material is removed from the surface by
the focused beam. The new material thus exposed also will be a different
color.
A typical laser-writing system will consist of the laser, optical components
to direct and focus the beam, one or more masks for pattern definition,
device-steering equipment, and component-testing apparatus to classify the
devices prior to marking. A generalized system using multiple spinning disk
masks is shown in Fig. 6.4. Some high-speed laser-marking systems can im-
print devices at rates up to 20,000 parts/hour (Higgins, 1986).
Although the initial cost of a laser-marking system is greater than many
conventional systems, the advantages of laser markers are considerable.
282 Wachter
ROTATING
MASKS
FOCUSING
LENS
DEVICE
COMPONENT
TESTING and
HANDLING
Figure 6.4 Components of a representative laser writing system.
No contact with material. This eliminates tool wear and does not subject the ma-
terial to deformations caused by tool pressure. Also, highly uniform marking
is obtained because there is no tool wear at the mark point.
Messy inks and fluids are eliminated.
Time savings. Parts can be marked at high rates and may even be moving while
the laser imprint is applied.
Irregularly shaped and inaccessible surfaces can be more easily accessed by laser
systems.
Control of the marking process is easily automated.
Increasingly, lasers are required to provide an assortment of labels based
on the results of electronic measurements in the circuitry. The measure-
ments may pinpoint process control errors or determine operating param-
eters of the component. This information can then be included in the label.
Such variable marking has necessitated development of advanced beam de-

livery and computer-control techniques.
Laser marking in the electronic and semiconductor industries has proven
to be a cost-effective alternative to methods using chemicals or tools. In
addition to flexibility in content, it provides durable, highly legible imprints
at high speeds and low operating costs.
6.3 RESEARCH AREAS
In the past two decades, a wide variety of research areas for laser-based
fabrication of semiconductor materials has been explored. The advantages
of laser methods discussed in the previous section also apply to these new
areas; i.e., lasers can provide clean, precisely controlled thermal power to
highly localized wafer regions at lower bulk temperatures than conventional
methods. In addition, larger areas can also be processed with lasers using
beam-scanning or step and repeat techniques.
In the following sections, several new laser-based processing applications
are discussed. Although some of these applications may not prove to be
worthwhile alternatives to existing techniques, several have demonstrated
economic and technical feasibility and may soon be installed in production
facilities.
The first topic to be discussed will be laser-assisted microcircuitry fab-
rication. This topic includes photolytic and pyrolytic film deposition and
etching that provide for localized deposition and removal of materials from
semiconductor surfaces. Next, laser annealing for incorporation of dopant
atoms into semiconductor substrates and recrystallization of polysilicon film
will be examined. This last topic concerns laser melting of small-grain
polysilicon crystals to form the large-grain poly films used for resistor, ca-
pacitor, and gate electrode fabrication. Finally, the use of lasers to form
electrical contacts between layers of multilayer IC's will be reviewed. Also
covered in this section will be the use of lasers for quick repair of random,
nonrepeating defects in semiconductor devices. These areas represent only
a few of the fields of study currently under investigation. Excellent discus-
sions of these and other research efforts can be found in symposium pro-
ceedings or the multitude of technical periodicals on this subject, or in re-
views such as The Industrial Laser Annual Handbook.
6.3.1 Film Deposition
The complex circuit designs in todays IC's and VLSI devices have been
brought about, in part, by reductions in the size of circuit features. Dur-
ing the 1960s, pattern feature sizes generally ranged from 5.0 to 25.0 11m.
284 Wachter
By the 1970s, feature size had been reduced to 3.0 to 6.0 p,m as new print-
ing techniques were developed. Presently, feature sizes as small as 0.5 p,m
are required in integrated circuits. To achieve these reductions in scale, im-
provements in the techniques used to define the interconnections between
transistor blocks in gate arrays, active regions of chips, conduction tracks,
gates, and other features in the chips have been developed. In addition to
reductions in the scale of circuit features, there is also a need to reduce
film thicknesses for multilayer IC's, while faithfully maintaining the wafer's
electrical and structural properties.
Thin films of various materials have been deposited on silicon substrates
by a variety of techniques: evaporation, chemical vapor deposition, sput-
tering, spraying, or spinning. Pattern definition, which sets the electrical
properties in the wafer, then takes place. The earliest method used for pat-
tern definition on IC's involved contact printing of photoresists (a photo-
sensitive resin that, after exposure to light, is used to selectively obscure
certain areas from etching). Contact printing involves first applying the re-
sist, then fastening a mask directly to the wafers. Mterward, the resist is
exposed using highly collimated UV light from a mercury arc lamp. Con-
tact printing techniques are capable of producing resolutions of less than
1 p,m (King, 1981) and are relatively simple and inexpensive. However, an
important drawback in all contact schemes is the large number of defects
resulting from contact between the mask and wafer. Projection printing
techniques, in which the mask and substrate do not come into contact with
each other, have been developed to eliminate these defects. Here, the re-
sist is applied in the same manner and the radiation for exposing the resist
is then directed through a mask and focused with lenses or mirrors onto the
wafer. Printing resolutions of better than 2.0 p,m have been achieved with
projection techniques.
For both projection and contact printing, the resolution of the image is
fixed by the wavelength of the radiation used to expose the photoresist. This
limitation can be overcome with the use of x-ray or electron beam imaging
systems. These two experimental techniques can provide image resolutions
of better than 0.5 p,m (Voltmer, 1981) and both techniques can be useful
for many types of printing operations. In addition to improved resolution,
electron beam processing is also unaffected by distortions in the wafer and
can be positioned with great accuracy. However, high production costs and
low throughputs are limiting features of this technique. X-ray imaging sys-
tems are still at an experimental stage and their utilization will depend on
solving a number of alignment problems.
Lasers may prove useful for satisfying today's thin-film deposition re-
quirements for a wide variety of materials of importance to semiconductor
fabrication. As early as 1965, lasers were used to evaporate refractory mate-
rials for the preparation of thin films (Smith and Turner, 1965). A few years
later nonmetallic films, such as SiO, were deposited with laser-based evap-
oration methods (Hass and Ramsey, 1969). A number of metals, including
Au, Ni, and Cu, have also been deposited onto metal substrates (von Gut-
feld et al., 1979). A partial listing of some other materials that have been
successfully deposited by laser methods is shown in Table 6.3.
There are four generally recognized methods for laser-based thin-film
production: evaporation, electroplating, chemical vapor deposition, and
photodeposition. In laser evaporation and electroplating, lasers aid in re-
moving precise amounts of material from a source that is then deposited to
the desired thickness on a substrate. The deposition process may take place
in vacuum or an inert gas atmosphere. Chemical vapor deposition (CVD)
uses lasers to heat the substrate that is enveloped by a cooler gas (or liquid)
of the material to be deposited. On being heated, the substrate reacts with
the gas to lay down the film. Alternatively, photodeposition involves tun-
ing the laser to an electronic or vibrational level of the gas. The irradiated
material decomposes and the products recombine on the cooler substrate
to form the film. Generally, UV lasers with wavelengths less than 250 nm
are used with photodeposition methods, whereas visible, UV, and IR have
proven useful for CVD because many substrates absorb in these regions.
Film deposition with lasers offers a number of advantages over conven-
tional methods. Perhaps most important, it offers the promise of high res-
olution, maskless writing, and localized microchemistry on semiconductor
devices. A recent investigation of maskless film deposition was performed
by von Gutfeld (1984) who developed a technique for laser-enhanced plat-
ing of Au and Cu on glass substrates. High-density, crack-free films with
2 ttm line widths were deposited at plating rates of up to 10 ttm/sec using
an argon ion laser. This property of laser writing will be of particular im-
portance to small-volume producers of individualized or prototype circuitry
because changes in laser output and beam focusing can be quickly and eas-
Table 6.3 Partial List of Thin Films Deposited by Laser-Assisted

Methods
Material Method Reference
Si0 2 Evaporation Hass and Ramsey, 1969
ZnS Evaporation Cali et al., 1976
Cr2 0 3 Pulsed laser heating Metev et al., 1980
Si Chemical vapor deposition Baranauskas et al., 1980
AI Pyrolysis Rytz-Froideraux et al., 1982
Ga2 0 3 Pulsed laser melting Matsuura et al., 1981
Cd,Sn Photodeposition Deutsch et al., 1979
Au Pyrolysis Houle et al., 1984
286 Wachter
ily programmed into laser systems to meet new production requirements. A

second advantage of laser-based film deposition methods is reduced wafer
damage due to dislocations, redistribution of dopants, and stacking faults.
Reduction in damage comes about because high temperatures can be re-
stricted to small regions of the wafer for brief time periods using lasers.
Laser processing is also a clean process in that unwanted diffusions or chem-
ical reactions on the wafer can be reduced or eliminated. Laser-based meth-
ods may also be used for deposition of films over large regions of the wafer.
For example, thin metal films of Ni, Mo, and W have been uniformly de-
posited by CVD using a pulsed C02 laser (Jervis and Newkirk, 1985).
Although laser-based techniques for deposition of films have not yet
been integrated into semiconductor fabrication facilities on a large scale,
the future is promising. Laser-based methods provide a clean and versatile
manner for deposition of thin films with smooth morphology, good adhe-
sion, 1 to 5 llm linewidths, and well-controlled material and electrical prop-
erties, qualities useful for modern IC and VSLI manufacture.
6.3.2 Etching
Laser light has been used to induce and enhance etching processes in met-
als, semiconductors, and organic materials employed in microelectronic cir-
cuitry. Etching methods have been developed to process such materials as
W, Mo, Al, Ga, As, InP, poly Si, and Si3 N 4 over various substrates. A partial
listing of materials that have been successfully etched using laser processing
is shown in Table 6.4. Both high-resolutior. lithographic and maskless pat-
tern formation have been investigated in liquid and gas ambients. Although
UV light is better suited for submicron pattern etching, efficient etching of
silicon microelectronic components has also been performed with IR and
visible lasers, both pulsed and continuous.
Table 6.4 Partial List of Materials That Have Been Successfully

Etched Using Lasers
Material Etchant Laser source Reference
Al BC13 C0 2 Rockwood and Hudson, 1975
w COF2 ArF Loper and Tabat, 1986
GaAs Cl2 Ar+ Tucker and Birnbaum, 1983
InP CF3I Ar+ Ehrlich et al., 1980
Si3N4 NF3 ArF Loper and Tabat, 1986
Si02 Cl 2 Ar+ Houle and Chuang, 1982
Ceramics Cl2 Ar+ Johnson and Smilgys, 1984
Absorption of the laser energy by the enchant gas or the substrate causes
the etching reaction. At least three different mechanisms have been de-
termined for semiconductor materials fabrication: thermal, photochemi-
cal, and photon-activated surface etching (Brewer et al., 1985). For ther-
mal etching, a laser source is chosen whose wavelength is strongly absorbed
by the surface that then reacts with the etchant. Photochemical etching
involves laser-assisted dissociation of gas (or liquid) molecules near the
surface of the substrate. The dissociated species interacts with surface
molecules and the resultant compound then separates from the solid, leav-
ing an etch pattern. For example, Si has been etched using laser light to
form free Cl radicals that interact with surface molecules to form the etch.
In photon-activated surface etching, a laser wavelength that is strongly ab-
sorbed by the substrate material is selected. Surface material is removed by
evaporation or explosive removal of the surface layers.
Laser-based methods of maskless etching have been demonstrated for a
number of substrates using both liquid and gas etchants. This fabrication
technique creates patterns by steering the beam directly onto the wafer sur-
face without the use of masks. It holds particular promise for the design and
manufacture of customized, small-quantity production runs. Laser methods
have provided faster etch rates ("' 100 ttm/sec) than conventional plasma or
wet-etching techniques and have generated feature sizes of 3 to 4 ttm or less
(Ehrlich et al., 1981 ). Fast etch rates and comparable feature sizes have also
been obtained with laser-based lithographic etching techniques. Here, the
pattern is formed by projecting the laser beam through a mask. Both direct
writing and lithographic etching can bring about anisotropic (image trans-
fer without distortion or undercutting) etch structures useful for the small
feature sizes in VLSI circuits.
Laser-assisted etching provides a number of other advantages in ad-
dition to anisotropy. Some laser-based etching processes are chemistry-
specific, making possible better tailoring of the etchant to the wafer ma-
terial. Laser etching is characterized by fast etch rates and, in some cases,
one-step pattern formation. Also, by using pulsed laser processing with ex-
cimer lasers, submicron structures can be produced without subjecting the
substrate to high temperatures. In this regard, the development of reliable,
high-power UV lasers with improved beam quality will improve prospects
for production-line implementation of laser-based methods.
6.3.3 Annealing
Incorporation of active dopant atoms such as B or P into semiconductor

substrates is important in CMOS (complementary metal oxide semicon-
ductor) technology because these atoms partly determine sheet resistance
288 Wachter
and leakage properties. The dopants are generally inserted into the sub-
strates by ion implantation and thermal diffusion techniques, but these
techniques have not always proved successful in producing well-regulated
dopant doses. Instead of controlled concentrations of dopant atoms con-
fined to a shallow layer of the substrate, conventional techniques may
produce unacceptable concentrations of dopant deeper into the substrate
than desired. The result is the dopant tail shown for the ion implantation
curve of Fig. 6.5. This figure, which details the concentration of dopant
B atoms as a function of depth in a silicon wafer, also shows that the
profile for dopants incorporated by a laser-based process (Holton, 1986)
is considerably less than that from conventional ion implantation. The
high dopant concentrations restricted to shallow regions of the wafer that
are achievable with laser annealing may meet important requirements in
VLSI technology such as reduced defect densities and shallower p -n junc-
tions.
In laser-based doping meth()ds, the beam is used to anneal the surface
and/or photodissociate the dopant gas. For either process, the action of the
laser will bring about diffusion of dopant atoms into the substrate, thereby
forming a shallow region with the desired electrical properties. Both pulsed
and CW lasers have been used to dope Si, GaAs, and InP substrates. Sheet
resistivities of less than 50 n;o and high dopant concentrations have been
obtained with the laser-based methods.
Laser-based annealing techniques may also be useful for the fabrication
of large-grained semiconductor sheets. When impurity-doped polycrys-
1021 .------,-------,-------.------.-------~------,-------,
Figure 6.5 Boron concentration in an Si substrate plotted as a function of depth

for laser doping and ion implantation (from Holton, 1986).
talline films of Si or Ge are deposited on crystalline substrates, the conduc-

tivity of the film may be reduced because of electron scattering at the grain
boundaries. Laser annealing has been shown to form large needle-shaped
grains from amorphous silicon, thereby reducing film resistivity to values
approximating those of a single crystal (Gat et al., 1978). A schematic of
the formation of single-crystal silicon on a crystalline silicon substrate is de-
picted in Fig. 6.6. These annealing techniques may become important in the
fabrication of solar cells, certain kinds of diodes and transistors, gate elec-
trodes, and interconnects on MOS-integrated circuits. In addition, crystal
growth by laser annealing may prove useful in the formation of multilayer
IC's composed of alternating films of crystalline Si and amorphous Si02 •
Annealing techniques using lasers offer a number of benefits over con-
ventional methods. Since the intensity and spatial extent of the beam can be
closely controlled, laser annealing offers the possibility of generating finely
resolved dopant patterns in semiconductor materials. Previous experiments
have shown that the lateral dimensions of laser-generated Si doping pat-
terns can be restricted to 0.3 !-!ill (Osgood et al., 1983). In addition, the
temperature of the annealed material can be well regulated by controlling
the laser intensity and time parameters. This allows the heat treatment to
affect only desired regions of the wafer without degradation of the substrate
or adjacent circuit elements. The reduction of thermal gradients in laser-
assisted recrystallization processes may provide defect-free epitaxial growth
at speeds that are large in comparison to furnace annealing.
LASER BEAM
POL YSILI~ON
ll j l LROWN s;
I
v SILICON
AMORPHOUS
"<4 ~
I r- J
CRYSTALLINE
r""""' SILICON
I
)
Figure 6.6 Regrown single crystal silicon produced from amorphous silicon by
laser annealing (from Brown, 1983).
290 Wachter
6.3.4 Via Drilling and Connections
Another semiconductor fabrication step that may prove amenable to laser

processing is that of via drilling in thick wafers. Vias are electrical con-
nections between metal layers on multilayer IC's. Controlled etching of
these structures in materials such as Si and GaAs is essential for the fab-
rication of a number of microelectronic devices. Presently, wet etching of
single-crystal substrates is the mainstay for conventional via drilling; how-
ever, this method has been limited to making low aspect ratio holes. Laser-
based methods have been found useful for both etching the via structures
and subsequent metallization to form the interconnection because they pro-
vide substantially reduced via dimensions and spacings. Experiments using
a laser-assisted chemical etching technique have shown that they can pro-
vide a fast, convenient method for drilling vias in GaAs wafers (Tucker and
Birnbaum, 1983). In these experiments, the focused CW argon ion laser
beam was used to heat the surface and dissociate the Cl2 etchant gas. Etch
rates of 33 J-lm/sec and hole diameters of 20 J-lm were achieved.
The increased packing density typical of recent integrated circuitry has
required the development of methods for rapidly repairing design errors
and reconstructing random circuit defects. Some of these fault correction
methods were discussed previously with regard to laser ablation and elec-
trical fusing for photomask and redundancy repair. The repairs frequently
require processing restricted to small regions of the wafer. Laser techniques
can fit this requirement and, in addition, produce single-step circuit modifi-
cation. Some candidates for laser repair techniques include incomplete or
poorly controlled etching of conductive tracks, resist failures, or metalliza-
tion shorts (Ehrlich and Tsao, 1984).
Failure identification and isolation in semiconductors may also prove
amenable to laser treatment. Here, individual circuit components can be
disconnected from the balance of the circuitry and tested for faults. An im-
properly etched via hole, e.g., could be detected with an electric probe and
repaired using pulsed laser radiation to modify the contact between layers
(Modlin and Ravich, 1986). Metallized conducting tracks in semiconductor
devices could also be isolated and analyzed for shorts and breaks.
6.4 OUTLOOK
A number of other research areas involving lasers for semiconductor de-

vice fabrication have not been discussed here. Some examples include the
use of lasers for surface cleaning of wafers, inspection of soldered connec-
tions, and as gettering agents for the removal of impurities in silicon lat-
tices. These areas, and those mentioned above, show promise because of
the many processing advantages offered by lasers. Foremost among these

advantages is the ability to deliver specific amounts of thermal power to
highly localized wafer regions. This quality is important for most appli-
cations in the semiconductor processing industry including trimming and
repair operations, marking, film deposition, etching and annealing. The
depth of the thermal treatments can also be well controlled. By proper
choice of laser wavelength, power, and timing parameters, a wide variety
of materials and applications can be addressed. The controlled application
of laser energy to semiconductor devices also results in reduced wafer dam-
age compared to conventional heat treatments because of the low bulk tem-
peratures. This property is increasingly important as smaller feature sizes
are demanded in integrated circuits. Laser processing has additionally been
used to cover large areas of wafers through beam scanning or stepping tech-
niques. For many processes, lasers provide a clean source of energy, thus
eliminating tool wear and masking requirements. These advantages point
toward increased laser usage in the years to come.
Lasers have proven to be very flexible tools in the semiconductor indus-
try. Changes in laser output or beam-focusing requirements provide access
to a broad range of spatial, material, and processing constraints. This flex-
ibility should prove especially important in the development of novel chip
circuits and the production of small quantities of highly specialized IC's.
This aspect of the IC market promises to become increasingly important in
the coming decade. Holton (1986) has indicated that small-quantity, spe-
cialized IC's, which presently form approximately 20% of the total IC mar-
ket, should increase their share to more than 50% by the early 1990s. The
flexibility provided by lasers can lower manufacturing costs and shorten cy-
cle times for small production runs.
Although research into laser fabrication of semiconductor devices has
been ongoing for approximately 20 years, the future is promising. The ad-
vantages and flexibility offered by lasers make them a viable economic al-
ternative for a number of steps in the formation of IC's and semiconductor
devices. Presently, many of the steps in laser processing are only partially
understood, and much investigation of both basic processes and fabrication
techniques remains. However, the potential of the new laser technologies
portend an important role in the semiconductor industry.
REFERENCES
Baranauskas, V., Mammana, C. I. Z., Klinger, R. E., and Greene, J. E. (1980).Appl.

Phys., 36: 930.
Boyd, I. W. (1984). "Laser-Assisted Pyrolytic Growth and Photochemical Deposition
of Thin Oxide Films," in Laser Processing and Diagnostics (D. Bauerle, ed. ), Proc.
292 Wachter
of an lnternat. Conf., Univ. of Linz, Austria, July 1984, Springer-Verlag, New

York, pp. 274-287.
Brewer, P. D., Reksten, G. M., and Osgood, R. M., Jr. (1985). Solid State Techno/.,
28 (4): 273.
Brown, W. L. (1983). Laser Mate1ials Processing (M. Bass, ed.), North Holland, New
York, p. 337.
Cali, C., Danen, V., Orioli, A., and Riva-Sanseverino, R. (1976).Appl. Opt. 15: 1327.
Crawley, M. C. (1984). Industria/Applications of Lasers (H. Koebner, ed.), Wiley,
New York, p. 177.
Crowley, R. T. (1986). Laser fabrication of photomasks for hybrid circuits, Proc.
of SPIE Conf on Laser Processing of Semiconductors and Hybrids, Vol. 611, Los
Angeles, Calif. pp. 18-22.
Deutsch, T. F., Ehrlich, D. J., and Osgood, R. M., Jr. (1979).Appl. Phys. Lett., 35:
175.
Ehrlich, D. J., Osgood, R. M., Jr., and Deutsch, T. F. (1980). Appl. Phys. Lett., 36:
698.
Ehrlich, D. J., Osgood, R. M., Jr., and Deutsch, T. F. (1981).Appl. Phys. Lett., 38:
1018.
Ehrlich, D. J. and Tsao, J. Y. (1984). Laser direct writing applications, in Laser Pro-
cessing and Diagnostics (D. Bauerle, ed.), Proc. of an lnternat. Conf., Univ. of
Linz, Austria, July 1984, Springer-Verlag, New York, pp. 386-395.
Fairfield, J. M. and Schwuttke, G. H. (1968). Sol St. Electron., 11: 1175.
Garman, G. and Ponce, J. H. (1986). The Industrial Laser Annual Handbook, 1986
Edition: (D. Belforte and M. Levitt, eds.), Penwell Books, Tulsa, Okla., p. 121.
Gat, A., Gerzberg, L., Gibbons, J. F., Magee, T. J., Peng, J., and Hong, J.D. (1978).
Appl. Phys. Lett., 33:775.
Gotzlich, J., and Ryssel, H. (1984). Applications oflaser annealing, in Laser Process-
ing and Diagnostics (D. Bauerle, ed.), Proc. of an lnternat. Conf., Univ. of Linz,
Austria, July 1984, Springer-Verlag, New York, pp. 40-49.
Hass, G. and Ramsey, J. B. (1969).AppL Opt., 8: 1115.
Higgins, J. P. (1986). Laser marking of passive components, hybrids, and semicon-
ductors, Proc. of SPIE Conf on Laser Processing of Semiconductors and Hybrids,
Vol. 611, Los Angeles, Calif., pp. 40-47.
Holton, W. C. (1986). Laser processing research for IC manufacture, Proc. of SPIE
Conf on Manufacturing Applications ofLasers, Vol. 621, pp. 56-61.
Houle, F. A. and Chuang, T. J. (1982). J. Vac. Sci. Techno/., 20: 790.
Houle, F. A., Jones, C. R., Baum, T. H., and Kovac, C. A. (1984). Technical Digest
of the Conference on Lasers and Electro-Optics, presentation ThH2.
Houle, F. A. (1984). Laser-assisted chemical etching, Proc. of SPIE Conference on
Laser-Assisted Deposition, Etching, and Doping, Vol. 459, pp. 110-114.
Jervis, T. R. and Newkirk, L. R. (1985). Materials Research Society Fall Meeting,
Dl.2.
Johnson, A. W. and Smilgys, R. V. (1984). Mat. Res. Soc. Ext. Abs. (Houle, F. A.,
Deutsch, T. F., and Osgood, R. M., Jr. eds.), p. 108.
King, M. C. (1981). VLSI Electronics: Microstructure Science, Vol. 1 (N. G. Ein-
spruch, ed.), Academic Press, New York, p. 41.
Loper, G. L. and Tabat, M. D. (1986). "Submicrometer Linewidth Production of

Integrated Circuit Materials by UV Laser Radical Etching," Proc. ofSPIE Conf.
on Manufacturing Applications ofLasers, Vol. 621, pp. 87-91.
Matsuura, M., Ishida, M., Suzuki, A., and Hara, K. (1981). lap. J. Appl. Phys., 20:
L726.
Metev, S.M., Savtchenko, S. K., and Stamenov, K. (1980).!. Phys., D13: L75.
Miyao, M., Matooka, R., Natsuaki, N., and Tokuyama, T. (1981). Laser and Electron
Beam Solid Interactions and Material Processing (J. R Gibbons, L. D. Hess, and
T. W. Sigmon, eds.), North-Holland, New York, p. 163.
Modlin, D. N. and Ravich, G. (1986). ·~pplications of Laser Microsurgery in Wafer
Scale Integrated Circuit Manufacturing," Proc. of SPIE Conf. on Manufacturing
Applications ofLasers, Vol. 621, pp. 76-85.
Mueller, M. J. (1986). "Functional Laser Trimming of Thin Film Resistors on Silicon
IC's, Proc. of SPIE Conf. on Laser Processing of Semiconductor and Hybrids, Vol.
611, Los Angeles, Calif., pp. 70-84.
Osgood, R. M., Jr., Ehrlich, D. J., Deutsch, T. R, Silversmith, D. J., and Sanchez, A.
(1983). "Laser Microchemistry for Direct Writing of Microstructures," Proc. of
SPIE Conf. on Laser Processing ofSemiconductor Devices, Vol. 385, pp. 112-117.
Parker, D. L. (1986). The Industrial Laser Annual Handbook, 1986 Edition: (D.
Belforte and M. Levitt, eds.), Penwell Books, Tulsa, Okla., p. 139.
Pearce, C. W. and Zaleckas, V. J. (1979). J. Electrochem. Soc., 126: 1436.
Preiswerk, H. P., Sheu, C. C., and Allen, S. D., (1986). "Laser Repair of Transparent
Microfaults in IC Photomasks," Proc. of SPIE Conf. on Manufacturing Applica-
tions ofLasers, Vol. 621, Los Angeles, Calif., pp. 71-75.
Rockwood, S.D. and Hudson, J. W. (1975). Chem. Phys. Lett., 34: 542.
Rytz-Froideraux, Y., Salathe, R. P., Gilgen, H. H., and Weber, H. P. (1982). Appl.
Phys.,A27: 133.
Shtyrkov, E. I., Khaibullin, I. B., Zaripov, M. M., Galjautdinov, M. R, and Bayazitov,
R. M. (1975). Sov. Phys. Semicond., 9: 1309.
Smith, H. M. and Turner, A. R (1965).Appl. Opt., 4: 147.
Steffen, J. (1984). Industrial Applications of Lasers (H. Koebner, ed.), Wiley, New
York, p. 209.
Swenson, E. J. (1983). "Laser Redundancy: Past, Present, and Future," Proc. ofSPIE
Conf. on Laser Processing ofSemiconductor Devices, Vol. 385, Los Angeles, Calif.,
pp. 93-96.
Tucker, A. W. and Birnbaum, M. (1983). "Laser Chemical Etching Method for
Drilling Vias in GaAs," Laser Processing of Semiconductor Devices, Proc. SPIE,
385: 131-140.
Voltmer, R W. (1981). VLSI Electronics: Microstructure Science, Vol. 1, (N. G. Ein-
spruch, ed.) Academic Press, New York, p. 1.
von Gutfeld, R. J., Tynan, E. E., Melcher, R. L., and Blum, S. E. (1919).Appl. Phys.
Lett., 35: 651.
von Gutfeld, R. J. (1984). "Laser-enhanced Plating and Etching: A Review," in Laser
Processing and Diagnostics (D. Bauerle, ed.), Proc. of an Internat. Conf., Univ.
of Linz, Austria, July 1984, Springer-Verlag, New York, pp. 323-331.
294 Wachter
Wagner, R. H. (1986). Functional laser trimming: an overview, Proc. of SPIE Conf

on Laser Processing ofSemiconductors and Hybrids, Vol. 611, Los Angeles, Calif.,
pp. 8-17.
Wills, S. and McPherson, J. (1983). Characterization of laser blown molydisilicide
links, Proc. ofSPIE Conf on Laser Processing ofSemiconductor Devices, Vol. 385,
Los Angeles, Calif., pp. 82-86.
7
Spectrochemical Analysis
Using Laser Plasma Excitation
Leon J. Radziemski *
Las Crnces, New Mexico
David A. Cremers
7.1 REVIEW
A laser pulse focused to a sufficiently high irradiance can generate a plasma

useful for spectrochemical analysis (Brech and Cross, 1962). The major-
ity of activity in this area during the two decades following this observa-
tion involved analysis of soiids and measurements of the properties of the
plasma under different conditions. The laser plasma was either used to va-
porize material for analysis by other techniques, or to both vaporize and ex-
cite. Reviews discussing both methods include Anderson (1973), Moenke
and Moenke-Blankenburg (1973), Petukh and Yankovskii (1978), Scott and
Strasheim (1978), Laqua (1979), Adrain (1984a, b), Cremers and Radziem-
ski (1986). Although the number of papers published on laser plasma anal-
yses has declined since the 1960s and 70s, the use of laser plasmas continues
to draw attention for specialized applications. Many of these were discussed
in the review by Cremers and Radziemski (1986). Here, we emphasize anal-
yses ofgases, liquids, particles, and surfaces in which the laser plasma is used
to vaporize and excite a material. Chapter 8 presents a detailed discussion
of the interaction between laser radiation and a solid and some recent an-
alytical results using laser plasma excitation on metals. The use of laser
plasmas as an ablation source is discussed in Chap. 9.
Present Affiliation:
• Associate Dean, College of Arts and Sciences
295
296 Radzlemskl and Cremers
7.2 METHODS AND PROPERTIES OF

ANALYSIS USING lASER PlASMAS
7.2.1 Methods
The principles of laser spark analysis are similar to those of conventional

plasma-based methods of atomic emission spectroscopy (AES) (Torok et
al., 1978). The latter use the hot plasmas produced by the electrode spark,
arc, and the inductively coupled plasma to excite the sample. The mate-
rial to be analyzed is introduced into the plasma where it is vaporized and
the resulting atoms electronically excited. The plasma light is spectrally re-
solved (and sometimes temporally resolved as in the case of pulsed sources)
to identify emitting species. Atomic ions as well as neutral atoms will be
produced in sources having higher temperatures. A measurement involves
integration of the signals from a line or set of lines of the elements of in-
terest over a certain time period. The amount of an element present is de-
termined by constructing a calibration curve of signal vs. concentration by
introducing samples of known composition into the source.
Some advantages of analysis based on AES compared to laser-induced
fluorescence and absorption spectrometry are the simplicity of the method
and the ability to simultaneously monitor all elements in the plasma, pro-
vided they are present in sufficiently high concentrations. Detection sensi-
tivity varies significantly between the different elements and is dependent
on the element and excitation properties of the source. Conventional AES
methods require that the sample be processed before analysis to put it in
a form that can be introduced easily into the plasma. In ICP analyses, the
sample is usually reduced to a solution that is nebulized into the source.
In laser spark analysis, the source is a microplasma formed by focusing
the optical pulses from a sufficiently powerful laser in or on the sample to
be analyzed. This distinguishes the laser spark from the more conventional
AES sources: The sample need not be transported to the source; rather,
the spark can be formed in or on the sample in situ. In addition, because
the spark is formed by optical radiation, the laser spark can be used to an-
alyze gases, liquids, and solids directly without sample preparation. Also,
no physical device need contact the sample to form the spark, thereby pre-
venting contamination.
The laser pulse energies required to form the spark depend on many
factors including properties of the pulse (such as the energy, mode qual-
ity, wavelength, pulse length) and the material [i.e., reflectivity at the laser
wavelength (solids) and density (gases)]. Typically, a 100 mJ/pulse from a
Q-switched laser (15 nsec pulse width) is sufficient to generate the spark
for most types of analyses. This corresponds to unfocused power densities
of about 10 MW/cm2 . Focused power densities can reach 30 GW/cm2 or
more. Immediately after formation, the temperature and electron density
Spectrochemical Analysis 297
can reach 25,000K and 1019 cm- 3 (Radziemski et al., 1983). These excita-
tion conditions are comparable to those of a conventional electrode spark
and significantly greater than many other sources used in AES. At these
high temperatures, matter in the spark volume is vaporized, reduced to its
atomic constituents, and then electronically excited. Emitting species in the
spark are identified by spectrally and temporally resolving the spark light
using detection systems similar to those used in conventional AES.
Because the laser spark is a pulsed source, its excitation characteris-
tics are a function of time. This is easily demonstrated by recording the
spark spectrum during distinct intervals as the plasma decays. The re-
sults are shown in Fig. 7.1 for the spark formed in the air. Within the
first few hundred nanoseconds after spark formation, the spectrum is dom-
inated by a spectrally broad background continuum due to bremsstrahlung
radiation from electron-ion collisions. Emission lines due to once ion-
ized nitrogen and oxygen are observed on top of this background light.
As the plasma cools, the intensities of ion lines and the background light
decrease significantly due to a reduction of the temperature and elec-
tron density, and emission lines due to neutral nitrogen and oxygen atoms
appear. Emission from simple molecular species can also be observed
(e.g., CN emission from the spark formed in air). Some elements dis-
play behavior different from this general scheme. For example, emissions
from once-ionized beryllium and uranium are observed out to several tens
of microseconds after plasma formation. On the other hand, emission
from once ionized fluorine atoms has not been observed (Cremers and
Radziemski, 1983), even at times as short as 100 nsec after plasma forma-
tion.
Because of the small size of the laser microplasma, the amount of sample
required for analysis is very small compared to that for conventional AES
methods. For example, the volume of the plasma produced by 100 mJ of
1.06 11m light from a Nd:YAG laser is about 0.003 cm3 • A typical detec-
tion limit of 1 ppm (w/w) for atoms in air, this corresponds to a minimum
detectable mass of only about 3 ng in each pulse. Similarly, the mass of ma-
terial removed from solid metal samples is only about 1 to 25 11g/pulse using
a Q-switched laser (Thompson et al., 1981; Carr and Horlick, 1982). Even if
the signals from 1000 laser pulses are averaged to perform a measurement,
the total mass required for analysis is still minute.
7.2.2 Properties
A schematic of a general apparatus for performing laser spark analyses is

shown in Fig. 7.2. The main components are the laser, method of focusing
the laser pulses on the sample, sample container, method of collecting the
298 Radziemski and Cremers
0.2fLS 399.499 (N II)
397.326 (0 II)
0.5 fLS
/404.131 (N II)
395.585 (N II)
391.899 (N II) \
394.759
\ 0 I { 394.748
\ 394.730
410.995 (N I)
409.995 (N 1)"-
1
1 1 ' 1 1 1 1 jl 1 1 1 • 1 1 1 1 1 1 1 r • 1 1 1 1 11 1 1 1 • 1 1 t 1 jl r 1 1 ' 1 1 1 1 11 1 1 1 ' 11 1 1 1t 11 r • r t 1 1 I" 1 1 • 11 1
390 400 410 420
WAVELENGTH (nm)
Figure 7.1 Evolution of the spectrum of a laser spark formed in air.

DISPLAY
-
TIME
CONTROL •••
••
ARRAY
SPECTROGRAPH
SPARK
PLASMA
Figure 7.2 Typical apparatus for laser spark analysis of a material. The detector
shown here is a time-gated photodiode array.
spark light, method of spectral resolution, detector, method of time resolu-

tion, and data storage.
The laser used for laser spark analysis must generate pulses of sufficient
power to produce the plasma. Commercially available lasers that meet
this criterion include solid-state lasers (e.g., Nd:YAG and ruby), gas lasers
(e.g., C02 and excimer), and Nd:YAG pumped- and flashlamp pumped-dye
lasers. Each laser has advantages and disadvantages that include properties
of the laser itself such as wavelength, mode quality, and characteristics of
operation. Table 7.1 presents a general comparison of some of these dif-
ferent types of lasers. Presently, the "workhorse," however, is the Nd:YAG
laser because it is a convenient, reliable, and compact source of powerful
Q-switched laser pulses.
The laser pulses are focused on the sample to form the spark using a fo-
cusing lens or mirror. The analysis geometry depends on the sample and will
be discussed below in the appropriate sections. Lenses and mirrors having
focal lengths of a few tens of centimeters are typically used to form the spark
because the high power densities produced in the focal volume reduce the
pulse energies needed for spark formation. Also, the short focal volume
produced by these lenses will highly localize the spark position. In general,
longer focal lengths require higher pulse energies to form the spark. Also,
because of the long focal volume produced by long-focal-length lenses and
Table 7.1 Comparison of Some Lasers Used for Laser Spark Analysis
Wavelength Pulse Pulse
Type (JLm) width (nsec) energy (J) Comments
Nd:YAG 1.06 7-12 0.3-1.0 Compact, low

0.53 maintenance,
reasonably priced,
glass optics
C02 10.6 1-300 JlSec 0.5-500 Simple design, inert
gas, special IR
optics required for
laser pulses
Excimer 0.19~.351 10-30 0.25 Rep rates to 250Hz
UV wavelengths,
toxic gases, quartz
optics required for
laser pulses
the lower powers required to form the spark on a solid surface, particles (in
gases and liquids) crossing the focal volume will induce the spark to form at
different locations along the focal axis on subsequent shots, thereby intro-
ducing noise into the measurements.
The spark light is collected and directed to the detection system using a
lens or fiber optic cable. Maximum light collection is usually achieved us-
ing lens imaging but the position of the spark on the sample in this case is
very critical: Changes in spark position of only a few millimeters can sig-
nificantly reduce the amount of light reaching the detector, especially if a
spectrometer is used for spectral resolution. In addition, with lens imaging
the detection system must be near the spark for optimum light collection.
Light collection using a fiber optic, however, reduces the sensitivity of de-
tection to spark position because of the large acceptance angle of the fiber
(Cremers, 1987). Also, the fiber can be used to transmit the spark light to a
detection system positioned remotely from the spark.
Certain spectral regions of the spark light can be selected for analysis
using (1) narrow pass filters, (2) a monochromator, or (3) a spectrograph.
The composition of the sample and the application determine which device
is appropriate. Narrow pass filters may be used to monitor emission lines
that are sufficiently isolated (i.e., 0.3 nm or more) from each other so that
high resolution is not needed. One filter and detector would be used for
each line to be monitored. More complex samples, composed of elements
having closely spaced emission lines, will require greater selectivity than
can be achieved with filters. If only a single emission line is to be detected,

a monochromator would be used to disperse the light. If several lines are
to be detected and high resolution is needed to separate these from other
spectral features, a spectrograph would be the instrument of choice.
The spectrally resolved light can be detected using either a photomul-
tiplier tube (PMT) or a photodiode array (PDA) (Talmi, 1983). The pho-
tomultiplier tube is used with the filter, monochromator, or spectrograph.
The PMT is positioned behind the filter or the exit slit of the monochro-
mator and the wavelength adjusted to monitor a particular emission line.
Alternately, the wavelength of the monochromator is scanned over the
emission feature of interest as the sample is interrogated by the spark.
Scanning the wavelength requires that the spark be formed reproducibly
on the sample. Either the PDA or PMT's can be used with the spectro-
graph to monitor several emission lines simultaneously. A PMT and slit
would be used for each line. Depending on the size of the spectrograph,
up to several tens of lines can be monitored in this way. Because a PMT
can be positioned anywhere along the focal plane of the spectrograph, the
(discrete) wavelength coverage possible with PMT detection is determined
only by the length of the focal plane and the dispersion of the instrument.
The (continuous) wavelength coverage possible with the PDA is also deter-
mined by the characteristics of the spectrograph but is limited ultimately
by the length of the array, which in present systems is at most 25 mm.
On the other hand, detection versatility is obtained with the PDA because
it provides a continuous record of the spectrum and can record a spec-
trum on a single shot (e.g., Fig. 7.1). This device is most useful for re-
search work because different spectral regions can be monitored by chang-
ing the wavelength setting of the spectrograph. The cost and limited life-
time of this device, however, generally restrict its use to laboratory analy-
ses.
The method of time-resolving the emission signals depends on the detec-
tor. The PMT signal is usually processed by a gated integrator (GI) (such as
that used in a boxcar averager) that only integrates the PMT current during
a time period selected by a gate pulse supplied to the GI circuitry. The gate
pulse is synchronized in time to the arrival of the laser pulse at the sample.
The integrated signal is held by the GI until it is read out to subsequent
electronics, such as an analog-to-digital converter. If the main concern is
discrimination against the strong background light at early times, the high
voltage to the dynode chain of the PMT can be gated so that full gain on the
tube is not realized until several microseconds after plasma formation. An
advantage of this method, compared to postprocessing the PMT signal as
with the GI, is that large tube currents that can change the operating char-
acteristics of the PMT are prevented. The disadvantage of this method is
that the gain of the detector does not remain constant.
Time resolution with the PDA is obtained by time-gating the voltage to

the microchannel-plate image-intensifier in front of the array. In this way,
the gain of the PDA can be controlled by a factor of 106 and time resolution
down to 5 nsec can be obtained.
7.3 ANALYSIS OF GASES
Direct excitation and analysis of gases are probably the simplest analytical
problems addressed by the laser spark technique. Experiments can be per-
formed in an ambient atmosphere (Schmieder, 1983) or in a cell (Cremers
and Radziemski, 1983). A simple static or flowing cell will suffice to hold
the sample. Small cell volumes are desirable when small amounts of sample
are available. A flowing cell can be used to extract decomposition products
and to remove contamination that may arise from outgasing, or that may be
induced by stray laser radiation hitting the cell surfaces. Calibration is rel-
atively easy. Known quantities of pure or mixed gases can be admitted into
the cell and a calibration curve of signal vs. concentration derived. Some
care must be taken with the optics providing the beam focusing. The re-
flection of the focused laser beam from the front window of the cell can
be inadvertently focused onto the front surface of the lens. This results in
gradual pitting of the lens and the loss of laser radiation to the sample cell.
The physics of the laser-induced breakdown process and subsequent
plasma expansion are discussed in Chap. 1 and 2. The breakdown threshold
depends on the gas being analyzed and its purity (e.g., Chap. 1, Fig. 1.16).
However, 100 mJ from a Q-switched Nd:YAG operating at 1.06 11m focused
with a 10 em focal-length lens should be sufficient to cause breakdown for
any gas at atmospheric pressure.
Gases have been analyzed directly by pulsed and continuous wave lasers.
The latter generates the laser-sustained plasma (LSP) that is described
in detail in Chap. 4. Briefly, the LSP is a continuously operating plasma
formed by focusing a cw-C0 2 laser beam in a gas at atmospheric pressure
and above (Generalov et al., 1970). The plasma resides near the focus of
the lens, as a bright ball of light several mm in diameter. Plasma tempera-
tures depend on the gas and range from 9000K for xenon up to 30,000K for
helium. Analytical applications of the LSP were first discussed by Cremers
et al. (1985). In that work, 25 W of C0 2 - laser radiation were used to form
the plasma in a sealed cell containing xenon. Oxygen and chlorine gases
were injected into the plasma and calibration curves constructed that were
linear over almost two orders of magnitude in concentration. Some types
of solid samples could be introduced into the cell by a laser ablation tech-
nique using a Nd:YAG laser. Limits of detection for Na and Cu were found
to be 5.2 and 0.87 ng, respectively, if we assume a measurement time of
1 sec. Because these experiments were performed using a sealed cell, it was
important to eliminate gases or decomposition products that absorbed at
the laser wavelength and hence perturbed or extinguished the plasma. This
problem can be avoided by operating the LSP as a plasmatron, similar to
the inductively coupled plasma, in which the inert gas is flowed through the
plasma axially. In this arrangement, the samples to be analyzed would be in-
troduced into the flow close to the plasma and the decomposition products
carried safely away. Experiments described in Chap. 4 show the feasibility
of constructing a plasmatron and that axial flow can even be used to stabilize
the LSP.
A repetitively pulsed Nd:YAG laser (1.06 J.Lm) was used to analyze efflu-
ent gases from a prototype fixed-bed coal gasifier at the DOE Morgantown
Energy Technology Center (METC). In these tests, gaseous effluents in a
pipe were sampled after a scrubbing process designed to remove alkali el-
ements. The goal was to assure that the alkali concentrations were below
a certain value so the blades of gas turbines would not be corroded by the
gases when they were used to generate electricity. The gas composition was
approximately 50% N2 , 20% C02 , with the balance being 0 2 , H 2 , and H 2S.
The alkali concentrations were at low levels of parts per billion. At the
LASER
POWER SUPPLY
LASER OSCILLATOR a AMPLIFIER
BEAM
DUMP
\I I, I
~
CALROD HEATER i
GAS FLOW ANODE: TO BOXCAR
CATHODE: TO hv SUPPLY
Figure 7.3 Schematic of the apparatus arrangement for generating emission spec-
tra from within a gaseous effluent pipe at the coal gasifier at the Morgantown Energy
Technology Center.
point of sampling, the gas pressure was 10 atm and the temperature 60° C.
Figure 7.3 illustrates the arrangement of the laser plasma analysis equip-
ment. Optical access to the pipe was through nitrogen-purged side ports.
The heater was necessary to keep the laser power supply warm because the
experiments were conducted outdoors at METC in March of 1982 at tem-
peratures close to oo C. Due to the high sample pressure, breakdown was
consistently achieved with a 30 em focal length lens and pulse energies of
300 mJ. Figure 7.4 shows the signal obtained from monitoring the oxygen
atom concentration via 0 I 777 nm. In a similar fashion, signals from H
and S were also observed, the latter via a multiplet at 930 nm. A signal
from alkali atoms was not observed. The experiments demonstrated that
the technique could be used for near real-time monitoring of concentration
changes of major and minor (but not trace) gaseous constituents. Calibra-
tion, however, remained a problem because of the difficulty of introducing
samples of known concentration into the sealed high-pressure system.
Direct laboratory analysis of gases using pulsed lasers was discussed by
Cremers and Radziemski (1983). There, the Q-switched Nd:YAG (1.06pm)
laser-generated spark was used to atomize chlorine- and fluorine-bearing
molecules. This technique is particularly useful for atoms such as those
ON LINE
MONITORING
0 I 777nm
ON LINE
>-
1-
u;
z
w
1-
z
OFF LINE OFF LINE
w GROUND GROUND
>
fi....J
w
a:
0 5
TIME (min)
Figure 7.4 Oxygen atom signal resulting from the apparatus shown in Figure 7.3.
20
0
t=
<t 15
a:
w
0
z
<t 10
c
z
~
m DEFINED TO BE 0.25
<t
-
5 (STOICHIOMETRIC)
0
0
o~~-L--L-~-L~~~~~~._~_.
0 50 100 150 200 250 300
RADIAL DISTANCE (0.001 inches)

Figure 7.5 Relative carbon to oxygen atom abundance in a flame as determined
by laser induced breakdown emission spectra. The ratio was shown to be propor-
tional to the fuel to air ratio (part of Fig. 12, Schmieder, (1983), reproduced with
permission).
with high lying excited states, because the high plasma temperature results
in significant population of these states. Emission from the neutral atoms
was temporally and spectrally resolved to maximize the signal-to-noise ra-
tios. Detection limits for chlorine and fluorine in air were 8 and 38 ppm
(w/w), respectively. The precision for replicate sample analysis was 8% rel-
ative standard deviation. This work also showed that the stoichiometry of
F- and Cl-bearing molecules could be determined. The slopes of calibration
curves were proportional to the numbers of Cl and F atoms in the molecules
(SF6 , CC1 4 , CzC13 F 3), indicating complete dissociation of these species by
the high spark temperatures.
Stoichiometry was also discussed by Schmieder (1983) for 0 and N atoms
in various gases. Other applications discussed in that review paper were
spectroscopic imaging across gasdynamic boundaries and determination of
species concentrations in combustors. By creating a series of sparks along
N 2 LASER TWO HeNe

BEAM, LASER BEAMS
Figure 7.6 Illustration of the use of a laser-induced plasma and probe beam com-
bination to determine velocities in a centrifuge.
a radial direction just above a burner and recording the spectrum for each
radial position, relative intensities of C, N, 0, and H atom lines were ob-
tained. These were analyzed to give fuel/air ratios at a set of positions along
the radius, as illustrated in Fig. 7.5.
A somewhat different application of laser plasmas was advanced by Al-
lison (1985) based on his work on the formation of laser-induced plasmas
in UF6 (Allison, 1983). In the most recent application, it was noted that
plasma formation in UF 6 causes density defects (because of the high tem-
perature) and shock formation. Such perturbations could be traced in space
using probe beams to monitor the passage of perturbations with time. Al-
lison proposed using the method in a gas centrifuge to give mass velocities
in a gaseous flow (Allison, 1985). The principle of the technique is shown
in Fig. 7.6.
7.4 ANALYSIS OF BULK LIQUIDS
The interaction between laser radiation and bulk liquids has been discussed
extensively in the literature (Sigrist and Kneubiihl, 1978; Bunkin and Tri-
bel'skii, 1980; Lyamshev, 1981). A brief discussion of the physics of the
optical breakdown of liquids is presented in Chap. 1. For analytical mea-
surements, the most important result is the dependence of the breakdown
threshold on the presence of microscopic suspended particles. Shot-to-shot
changes in the breakdown threshold can contribute to large variations in the
analytical signal and hence degrade the precision.
Liquids can be analyzed by generating the laser spark on or under the
surface. The advantages of forming the spark in the bulk liquid away from
Figure 7.7 Photograph showing formation of the laser spark on the surface of a
flowing liquid containing dissolved uranium.
the surface are that (1) splashing of the liquid that can contaminate the opti-
cal components and introduce noise into the measurements is avoided and
(2) a stable free surface is not needed. The disadvantages of bulk analy-
sis are (1) the liquid must be transparent at the laser wavelength and the
emitting wavelengths of the species to be monitored and (2) heavy parti-
cle concentrations can interfere with spark formation. The feasibility of
performing bulk analysis on clean solutions, static and flowing, was demon-
strated by Cremers et al. (1984). Detection limits were determined for ele-
ments in aqueous solutions that ranged from 10 ppb up to 1200 ppm. The
lowest limits were obtained for the alkali elements (Li, Na, etc.) due to the
ease with which the strong emission lines of these elements were excited.
Elements such as B and Be were more difficult to excite, resulting in signif-
icantly higher detection limits.
~ 409.013 nm (U II)
-
'iil
r::
:I 20000
.e.
0
>
1-
Ci)
z 10000
w
1-
z
406 408 410 412 414 416 418
WAVELENGTH (nm)
Figure 7.8 Nine sequential spectra obtained by forming the laser spark on the sur-
face of a liquid containing 100 gm/liter of uranium. The strong analytical line of U II
is shown in black. The shot-to-shot differences in the magnitudes of the lines were
due to splashing of the liquid.
The detection sensitivity for some elements in the bulk liquid was in-
creased using a double-spark technique. There one spark was used to va-
porize a small volume of the liquid and a second spark, following closely in
time, was used to excite the gas bubble created by the first spark. The great-
est gain in sensitivity was obtained by spacing the sparks by about 18 p,sec.
By using this method, the detection limit for boron was reduced from 1200
to 50 ppm. Detection limits for other elements difficult to excite with the
single spark were similarly reduced. No gain in sensitivity was obtained for
detection of Li and Na probably because these elements were fully excited
by the single spark.
The detectability of some elements can be increased by forming the spark
on the liquid surface (Wachter and Cremers, 1987). A photograph of the
spark formed on a flowing solution of uranium in 4 molar nitric acid is shown
in Fig. 7.7 and a typical spectrum is shown in Fig. 7.8. Uranium could not be
detected in the bulk liquid at concentrations as high as 300 gm/L, whereas,
by using surface excitation, a detection limit of 0.1 gm/L was obtained. The
lower detection limit was due to the higher temperature of the spark formed
on the liquid compared to the spark generated in the bulk liquid. The advan-
tage of surface excitation is that the solution can be opaque and contain high
particle concentrations without significant perturbation of spark formation.
The disadvantages are splashing of the liquid and the strong dependence of
the analysis on the distance between the liquid surface and focusing lens. By
containing the flowing liquid in the cell design shown in Fig. 7. 7, the splash-
ing problem was eliminated: The liquid flowed off the curved upper surface
of the glass tube. The analysis of individual liquid droplets will be discussed
below in Section 7.5.
7.5 ANALYSIS OF PARTICLES
Particles differ from other solid or liquid state samples in their size, typically
below 100 J.tm in diameter, and in their morphology. Liquids tend to form
microspheres, whereas solids are of irregular shape. These aspects, size
and shape, together with composition lead to interesting sampling problems
and methods of detection. We will review work on particles in air, particles
entrained in flows, and single particles.
The physics of laser-induced breakdown on aerosols is discussed in
Sec. 1.4.3 of Chap. 1. Particularly of interest is the change in breakdown
threshold with particle size, illustrated in Fig. 1.20 of that chapter. It in-
dicates changes in threshold irradiance from 1010 W/cm2 to 106 W/cm 2 as
the diamter-to-wavelength ratio changes. If true for other wavelengths, this
variation will affect shot-to-shot signal levels and degrade the precision of
direct aerosol concentration measurements.
Calibration for aerosols of arbitrary morphology is a difficult task, and to
date few quantitative results have been obtained. As indicated by Cremers
and Radziemski (1985), there is evidence that there is a maximum mass
that a given laser plasma can vaporize. It appears that the technique would
undersample particles greater than 10 J.tm in diameter.
Detection of Pb, Cd, and Zn containing aerosols, generated in the labo-
ratory with a nebulizer/heat chamber, was discussed by Essien et al. (1988).
These experiments were similar to those performed on beryllium aerosols
(Cremers and Radziemski, 1985). Essien used a repetitively pulsed Q-
switched Nd:YAG (1.06 J.tm) to excite particles formed by the nebulization
and drying of sample solutions of appropriate concentrations. The detec-
tion limits in air were 0.20±0.02 J.tg/g (Pb), 0.018±0.002 J.tg/g (Cd), and
0.23 ± 0.14 J.tg/g (Zn). The Pb value is slightly above the recommended av-
erage 8 hr exposure limit, whereas the detection limits for Cd and Zn are 2.5
and 10 times lower than that limit. Hence, the method would be adequate
for health-related real-time monitoring of Cd and Zn, but not Pb.
Other experimental and theoretical studies of laser spectrochemical
analysis of aerosols have been reported with a view to remote analysis.
Belyaev et al. (1979) discussed the process of diffusive evaporation of solid

aerosol particles. Expressions were obtained for the temperature and den-
sity of the vapor around the particle. These were combined with emis-
sion coefficients to predict line emission intensities. Experiments were per-
formed to study Na and Ca emission line intensities as a function of power.
Similar experiments were performed by Godlevskii and Kopytin (1980),
who also studied the blocking of the subsequent beam transmission after the
formation of the plasma. The irradiances for blocking subsequent transmis-
sion (within the same pulse) depended slightly on the aerosol composition,
but were generally on the order of 107 to lOS W/cm2 •
The studies described above culminated in a technique for remote spec-
trochemical analysis of aerosols. Plasma emission signals were retrieved at
distances of from 50 to 150m by a Cassagranian telescope sending-receiving
system. The results were reviewed in the book by Zuev et al. (1985). An ap-
paratus diagram is shown in Fig. 7.9. A 500 J per pulse C02 laser with a
pulse duration of 300 Jl-Sec was used. Spectra from three to seven sparks
were recorded photographically. Alternately, individual spectral lines were
monitored with photomultiplier tubes placed behind narrow slits in the fo-
cal plane of a polychromator. Published spectra show Ca, AI, and Na lines
in addition to the ubiquitous oxygen and nitrogen lines. One interesting
comment bears directly on improving the accuracy of compositions. It was
noted that the ratio of two line intensities is an indirect measure of the mean
temperature of the electron gas in the plasma. This is not quite correct,
because the spectral line intensity ratio will, at most, give an indication of
Boltzmann (excitation) or Saha (ionization) temperatures that are not nec-
essarily in equilibrium with the electron temperature (Radziemski et al.,
1985). However, the former are actually the ones of interest in determining
the degree of excitation in a given laser spark. Hence, correlating spectra
with a given line pair intensity ratio should improve signal to noise and pre-
cision.
In situ elemental analysis of particulate fuels by laser spark spectroscopy
is continuing at the Combustion Research Facility of Sandia National Lab-
oratory (Livermore) (Ottesen et al., 1989). A primary goal is the simulta-
neous measurements of size and composition of individual coal particles.
The size range was generally 30 to 50 11m in diameter, and the shapes were
irregular. Figure 7.10 shows the apparatus for the simultaneous sizing and
spark spectroscopy experiments. Two colinear laser beams were used to
obtain size information via forward scattering. The focus of the frequency-
doubled Nd:YAG laser pulse (532 nm) beam was aligned slightly above the
position of the sizing beams. The laser pulse was triggered by the passage
of a particle through the sizing beams, such that the particle was vaporized
and its elemental constituents excited. These experiments were conducted
on previously analyzed coal. Single-shot spectra allowed identification of all
CASSEGRANIAN
TELESCOPE
GRATING
SPECTROGRAPH
B =BEAM SPLITTER
P =LASER ENERGY MONITOR
Figure 7.9 Apparatus schematic of a spectrochemicallidar using the laser spark.

(adapted from Zuev et al. (1985), reproduced with permission).
constituents with average compositions greater than 100 ppm (AI, Fe, Ca,
Si, K, Mg, Na, Ti, Li). However, Mn and Cr that were at average concentra-
tions below 100 ppm were not observed. Experiments are being extended
to burned coal particles.
Recently, spectroscopy has been performed on laser plasmas initiated
on single, optically levitated microspheres (Biswas et al., 1987, 1988). The
particles were 10 to 20 pm diameter spheres of glycerin or glycerin-water
mixtures. Carbon was followed through the 248 nm C I line and the 251 nm
C II multiplet. Carbon lines were monitored as early in the plasma as time
resolution would permit (20 nsec ), indicating an early vaporization of some
droplet material.
The analysis of isolated nanoliter droplets of a liquid was reported by
Archontaki and Crouch (1986). A stream of droplets was created by a cap-
He-Ne Spectrometer
Laser
PDP 11/34
Computer
Argon lon
Laser
Figure 7.10 Apparatus for simultaneous sizing and laser-spark analysis of coal par-
ticles. The He-Ne and argon ion lasers are used in the size measurement. The sec-
ond harmonic of the Nd:YAG laser is used to form a plasma on the coal particle.
(courtesy of D. K. Ottesen and Sandia National Laboratory (Livermore)).
illary/bimorph droplet generator. The stream traversed the focal volume

of the laser pulse that was fired at the appropriate time to vaporize indi-
vidual droplets. Analytical performance for calcium and sodium solutions
was similar to those reported by others (Radziemski et al., 1983). However,
here the droplet size was much better characterized, a 50 to 80 JLm radius
in these experiments. The well-controlled interaction led to the observa-
tion that the plasma properties were dominated by the air and not by the
droplet, at least at the irradiances used (lOS W/cm2 ). The technique may
have application as an element-selective liquid chromatographic detector.
In related work, Chang and co-workers have observed the formation of
a plasma in and around droplets with spatial and spectral resolution. The
droplets were produced by a vibrating orifice droplet generator. In one ex-
periment (Eickmans et al., 1987), the droplets contained aqueous solutions
of sodium or lithium. The corresponding resonance lines were monitored.
The detector was a gated diode array, which was used in the spatial or spec-
tral mode. In another experiment, shadowgraph pictures of the exploding

droplets were taken with a framing camera (Zhang et al., 1987). Details of
the plasma plume formation and droplet disintegration were deduced.
7.6 ANALYSIS OF SOLIDS
Laser-induced breakdown spectroscopy has found the most widespread use

in the analysis of solid materials, in particular metals. A discussion of laser-
induced breakdown on solids was presented in Chap. 1 and of plasma de-
velopment in Chap. 2. The relationship of the emitted radiant energy to
spectrochemistry is discussed in Chap. 8. The analysis of solids using the
laser spark has been reviewed recently (Cremers and Radziemski, 1986),
and the reader is referred to this work for a detailed discussion and addi-
tional references. The intention here is to summarize the technique, report
recent results, and describe two potential applications.
There are many advantages of the laser spark for solid analysis that do
not occur in the analysis of gases, liquids, and aerosols. Two are (1) the small
focal spot (100 ,_,m or less) obtainable with focused laser beams can provide
a spatially resolved analysis of surfaces and (2) the simultaneous vaporiza-
tion and excitation eliminates sample preparation. In the first laser micro-
probe instruments, the laser pulse was used only to provide a vaporized sam-
ple that was subsequently analyzed by a conventional electrode spark. One
problem associated with the use of the laser spark in the early laser micro-
probe devices was the low repetition rate and poor reliability of the laser.
The former was typically only a few pulses per minute or, at best, one pulse
per second. In addition, the reliability of the laser was poor. Today, very re-
liable solid-state lasers are available that can generate several million shots
at rates up to 30 Hz without requiring maintenance. This opens up the pos-
sibility of completely laser-based analyses because a large number of laser
pulses can be used to collect data rapidly.
The interaction between laser radiation and a solid sample is a compli-
cated process dependent on many characteristics of both. If we assume a
laser is available that generates highly reproducible pulses, characteristics
of the sample will determine the analytical results. These are surface condi-
tion, composition, and homogeneity. Because the laser spark interrogates
only the surface of a material, a clean surface is needed to obtain a repre-
sentative measure of elements composing the bulk sample. In many cases,
however, surface contamination can be removed easily by forming only a
few sparks at the same spot on the sample. This cleaning process is demon-
strated in Fig. 7.11 that shows sequential spectra obtained by monitoring
the light emission from 30 sparks formed at the same spot on a sample. The
aluminum plate was covered with a thick coat of paint so aluminum emis-
392 394 396 398
WAVELENGTH (nm)
Figure 7.11 Sequential spectra recorded by focusing laser sparks (10 Hz) on a
painted aluminum surface. As the paint was ablated off by the laser sparks, strong
aluminum emission lines were observed.
sian was not observed on the first few sparks. As the paint was ablated,
however, strong aluminum emissions were observed.
Irregular surfaces present difficulties because the sampling geometry can
change from shot-to-shot, in particular, the lens-to-sample distance, which
in turn can affect the excitation properties of the plasma. This problem is
especially acute in the analysis of powders because the strong shock waves
set up by the laser spark result in movement of the sample.
The composition of the sample can affect the analysis through the surface
reflectivity at the laser wavelength. For example, high C0 2 -laser powers
(10.6 fLID wavelength) are needed to couple the radiation into the surface
of a clean aluminum sample, whereas Nd:YAG laser (1.06 fLID wavelength)
powers about an order of magnitude less are required to accomplish the
same processing (Schuocker, 1986). In addition, the presence of certain
materials in the sample can affect the emission properties of other elements,
complicating accurate analysis of the material. These so-called "matrix ef-
fects" are observed in all AES methods, but are especially pronounced in
the laser spark because of the high density of matter in the microplasma.
Because of the small masses ablated by the spark and the small size of the
focal spot, sample homogeneity is important if a bulk analysis is desired. To
a certain extent, inhomogeneities can be averaged out by moving the sample

and using many sparks.
Perhaps the easiest solids to analyze are clean metals because the laser
energy couples easily into these materials as indicated by the relatively low
energies required to form a spark. In general, wavelengths below a few
microns couple into the metals better because of the reduced reflectivity
of the surface at these shorter wavelengths. The action of the laser spark
on the surface results in the formation of a small crater and the removal
typically of nanograms of material. For laser spark analysis, high-power
(megawatt) pulses generated by Q-switched lasers are used because these
form the hottest sparks on the material. On the other hand, if the main
concern is only removal of large masses from the sample, as in some of the
laser ablation methods discussed in Chap. 9, then longer pulses having more
energy are preferred.
Based on the extensive literature of laser spark analyses of solids (Ander-
son, 1973; Petukh and Yankovskii, 1978; Cremers and Radziemski, 1986),
a summary of the analytical capabilities of the method can be compiled.
Some representative detection limits are listed in Table 7.2. Note that as
with all methods of AES, the detection limit is strongly dependent on the
element. In fact, elemental masses detectable in a single spark can be quite
low, down to femtograms, although practically, many sparks are needed to
perform analyses. Because the method depends on the condition of the
sample, it is difficult to state analytical figures of merit. In general, how-
ever, with fairly homogeneous samples, accuracy and precision are typically
10%, although with carefully prepared samples of high homogeneity, these
values can be as low as a few percent.
The first applications of the laser spark for solid analysis were directed
to laboratory-based analyses. Recent interest, however, appears to center
on industrial applications. Most of these applications are presently in the
development stage and so have not been reduced to practice. It is instruc-
tive, however, to consider briefly two of these areas: the analysis of molten
metals and the analysis of metals remotely.
For the past several decades, many methods have been investigated to
rapidly determine the elemental composition of molten metals. Molten
steel is currently analyzed by collecting a molten sample, cooling it, and
then transporting it to a nearby analytical laboratory for analysis, a pro-
cess that can require 5 to 10 min. During this time, the composition of the
hot melt (up to 200 tons) can change significantly via vaporization of more
volatile components. It has been calculated that if the elemental analysis
can be carried out in 1 min or less, the cost savings to the U.S. steel industry
alone would be substantial.
The laser spark technique is presently being investigated in several labo-
ratories as a means to provide a rapid analysis of molten metal (Ozaki et al.,
316 Radzlemski and Cremers
Table 7.2 Detection Limits for Elements in Solids
Detection limit
Sample Concentration (ppm) Mass (pg)
Metals in an organic
Matrixa
Hg 0.3
Fe 0.3
Zn 0.05
Mg 0.002
Rare earths in a
NaCl matrixb
Sc 2
La 10
Gd 200
Nd 500
Elements in steele
Cr 200
Ni 570
Si 490
3 Treytl eta!. (1972).
blshizuka (1973).
cFelske eta!. (1972).
1984; Tsunoyama et al, 1985; Cremers, 1984; Cremers et al., 1985). Basi-
cally, these approaches involve the formation of multiple laser sparks on
the molten metal at a free surface formed under the slag layer (which can
be several inches thick). The light from the sparks is collected and analyzed
in the usual way using spectral and temporal resolution. In principle, this
technique should offer the desired analysis speed. Several problems must
be overcome before the method can replace existing techniques. First, it
must be demonstrated that the composition of the free molten surface is
representative of the bulk melt. Changes in the composition of the molten
surface via selective vaporization may be a significant problem. Second, the
method may not have sufficient sensitivity to monitor all elements of inter-
est for steel-making. The elements phosphorus and sulfur, which must be
kept at low concentrations (less than 0.01%) in many steels, are difficult to
detect at these low levels using the laser spark. Third, the method of calibra-
tion may be difficult to implement. For maximum precision, AES methods
require calibration using standards similar to those of the unknown mate-
rial. The development of standards suitable for molten metal analysis may
be a formidable task. For these reasons, development of the laser spark

technique for steel analysis represents a challenging application.
The high accuracy and precision required for steel analysis and which
is attainable using laboratory-based methods may not be necessary in all
applications. In some cases, a semiquantitative or even a qualitative analysis
may be satisfactory. For example, a rough estimate of the elemental content
of ores traveling down a conveyer belt may be useful to determine if the
ore grade is sufficiently high to undergo further processing. Also, a rapid
analysis of metals may be useful to sort scrap metal according to its main
elemental components before recycling (Loree and Radziemski, 1981).
These applications require an analysis method that can be carried out
rapidly and at a distance. Experiments described in Sec. 7.5 have demon-
strated that the laser spark can be generated at significant distances from
the focusing lens and that the spark light can be collected at this distance to
obtain useful results from aerosols. Recent similar experiments have ver-
ified that the main elemental components of metals can be determined at
distances of 10m using laser powers available from compact portable laser
systems (Cremers, 1987; Cremers, 1988a). At distances of a few meters, it
was possible to carry out semiquantitative analyses. The main factor lim-
380 390 400 410

WAVELENGTH (nm)
Figure 7.12 These sequential spectra were obtained by introducing different metal
samples into the laser spark at a rate of 19 Hz. A single spark was used to record
each spectrum. The elements listed at the left of each spectrum indicate the main
elemental component of the corresponding sample.
iting extension of the experiments to much greater distances was the effi-
ciency of collecting light from the spark, which in these experiments could
be increased by at least an order of magnitude by simple improvements of
the detection system. The ability to analyze materials rapidly has also been
demonstrated recently. Figure 7.12 shows nine spectra obtained by intro-
ducing different metal samples into the laser sparks at a rate of 19 Hz. A
single spark was used to collect each spectrum. It is clear from the figure
that under these conditions, the spectra are useful to sort the metals accord-
ing to the elements giving rise to the strongest spectral features. The devel-
opment of high-repetition-rate lasers, fast detection systems, and computer
systems capable of processing the data from such an analysis rapidly, makes
metal sorting feasible.
7.7 ADVANCES IN INSTRUMENTATION
The most extensive use of laser plasmas has been for the analysis of solids.
Laser microprobe instruments that use the plasma to vaporize and excite
solids (laser microprobe) have been manufactured by several vendors in-
cluding Jarrell-Ash (U.S.), VEB Carl Zeiss Jena (Germany), and JEOLCO
(Japan). The availability of these units has declined in recent years due
to competition from other methods that combine the microsampling capa-
bility of the laser pulse with other analytical techniques (e.g., laser ablation
mass spectrometry). The development in recent years of more compact and
reliable laser systems, however, combined with the increased need for field-
deployable instrumentation, has increased interest in laser plasma methods.
The gains made over the last 20 years in laser technology are demonstrated
by the following example. One of the significant results reported in a paper
published in 1968 was the development of a laser system that could be fired
20,000 times over a period of one year without repair or significant degra-
dation in performance (Peppers et al., 1968). Today, Nd:YAG lasers can be
purchased that surpass this performance by several orders of magnitude.
Besides the laser microprobe, there have been few reports of instruments
developed based on laser plasma analysis. A drain ( 1984b) described the
operation of two instruments in a recent review article, but in each device
the laser spark was used only to ablate from a solid surface material that
was analyzed by other means. The feasibility studies being carried out now
in several laboratories, such as the coal gasification experiments and coal
particle analyses described above, however, may eventually result in instru-
ments.
At Los Alamos National Laboratory, an instrument based on the laser
spark has been developed to analyze airborne beryllium collected on filters.
Beryllium is known to be a significant inhalation health hazard to workers
Figure 7.13 Photograph of the mobile beryllium monitor.

2.54cm--J
PMTI
I
I
To Pump To Laser
To Fan Power Supply
• - - - Laser Pulses
- - - - - - - Spark Light
Calibration Light
Figure 7.14 Diagram of the analysis portion of the beryllium monitor. PL =

pulsed laser; HR = high reflector at the laser wavelength; CL = cylindrical lens;
F = filter; RS = rotating stage; L = spherical lens; PMT = photomultiplier tube;
S =spectrograph; M =motor; RM = rotating mirror; W =window; HCL = hollow
cathode lamp; FS =air flow sensor; H =hose.
exposed to dusts of beryllium and its alloys. Current methods of analyzing

airborne beryllium involve (1) collection of the particles on filters, (2) chem-
ical ashing of the filter to produce a solution that is (3) analyzed by atomic
absorption spectrometry. Following collection, the most time-consuming
step is (2), which can require several hours. Ideally, to provide for maximum
worker safety, filter analysis should be carried out as rapidly as possible.
Cremers and Radziemski (1985) demonstrated that the collected beryl-
lium mass can be determined rapidly by scanning repetitive laser sparks
across the filter surface. Because the spark vaporized and excited the beryl-
lium, sample preparation was eliminated, reducing analysis times to a few
minutes. Based on this work, a beryllium monitoring instrument was de-
veloped to automatically analyze filters (Cremers, 1988b). A photo of the
instrument is shown in Fig. 7.13. The basic components are the laser, de-
tection system (spectrograph and PMT's), filter rotating mechanism, and
data acquisition/processing system (electronics and computer). A diagram
of the analysis section of the device is shown in Fig. 7.14.
Filters are analyzed by placing them on a platform that rotates and trans-
lates the filter under the laser sparks. Sparks about 3 mm long are pro-
duced on the filter by focusing the laser pulses with a cylindrical lens. In
this way, about 88% of the filter surface (35 mm diameter) can be sampled
in only 2 min using 1200 sparks. The spark light is spectrally resolved using a
0.125 m spectrograph and detected by a PMT. The PMT signal is processed
by a gated integrator to produce a voltage that is digitized and stored in
computer memory.
The detection limit for beryllium on the filter was about 10 ng, corre-
sponding to a surface mass density of 3.2 ng/cm2 • The calibration curve
showed good sensitivity over the range 10 ng to 1 f.lg. At higher masses,
there was a loss of sensitivity due to saturation of the detection system on
some laser sparks because of high beryllium masses. The dynamic range can
be extended to higher masses by using a second PMT to monitor a split-off
portion of the beryllium signal. Accuracy and precision were somewhat de-
pendent on the mass and typically were between 10 to 20%. Although these
values are larger than the accuracy and precision obtainable with conven-
tional laboratory techniques, they are sufficient to provide for routine mon-
itoring and warning of high airborne mass levels.
7.8 PROGNOSIS
The use of laser plasma-based analysis will probably continue to be limited

to specialized applications. A critique is presented in Thble 7.3. The method
has some advantages and disadvantages common to all AES techniques and
some unique ones. In terms of analytical capabilities, detection limits for
322 Radziemskl and Cremers
Table 7.3 Critique of the Laser Plasma Analysis Method

Advantages
Simple experimental arrangement
Simultaneous multielement analysis
Noninvasive
No sample contamination
Small masses required
Minimal sample preparation
Disadvantages
Cost of laser
Detection sensitivity for some elements
Limited accuracy and precision
Difficulty of calibration for in situ analyses
Affected by sample inhomogeneities because of small plasma size
most elements, and measurement accuracy and precision, the laser plasma
is not competitive with established laboratory-based techniques. There are
several reasons for this, related mainly to shot-to-shot differences in the
plasma properties that affect the magnitude of the element signal. These
properties are plasma size that affects the mass of sample vaporized and
the excitation characteristics (temperature and electron density) that af-
fect both mass of sample and degree of excitation. In addition, compared
to continuously operating sources, such as the inductively coupled plasma,
the integrated signal from the laser plasma is much smaller. For example,
a typical ICP integration period is 10 sec. Each laser plasma produces an
element signal lasting at most 200 psec. The number of laser plasmas re-
quired to match the 10 sec period is 50,000. This requires an analysis time
of 8.3 min for a laser operating at 100Hz.
Based on applications to date, the main advantages of the laser plasma
would seem to be that it can be generated remotely with little or no sam-
ple preparation. The beryllium monitor described in Sec. 7.7 is one exam-
ple. The plasma can sample directly beryllium collected on the filter surface
without special preparation of the filter before analysis. Remoteness means
the laser plasma can be generated in a normally inaccessible, possibly hos-
tile environment, such as a hot gas stream at a coal gasification plant or a
flowing liquid stream in a uranium fuel reprocessing plant, to provide in
situ analysis. Another advantage of the laser plasma, shared with conven-
tional AES methods, is its simplicity, making it amenable to turnkey-type
operations. Many other laser-based methods, such as laser-excited atomic
fluorescence, are not presently feasible for on line analyses because of the
sophisticated laser systems they require. On the other hand, recent devel-
opments in laser and detector technology make a simple method, such as
laser plasma analysis, more attractive.
REFERENCES
Adrain, R. S. and Watson, J. (1984a). Laser microspectral analysis: a review of prin-

ciples and applications,/. Phys. D:App. Phys., 17: 1915.
Adrain, R. S. (1984b ). Some industrial uses of laser induced plasmas, in Industrial
Applications ofLasers (H. Koebner, ed.), Wiley, New York, Chap. 7, pp. 135-176.
Allison, S. W. (1983). Thesis.
Allison, S. W. (1985). Private communication.
Anderson, C. A., ed. (1973). Microprobe Analysis, Wiley, New York, Chap. 12-15.
Archontaki, H. A. and Crouch, S. R. (1986). Laser-induced breakdown spectroscopy
for the analysis of solutions, Abstract in the Final Program of the 13th Annual
Meeting of the Federation of Analytical Chemistry and Spectroscopy Societies,
St. Louis, Missouri, Sept. 28-0ct. 3, 1986; and Crouch, S. R. (1987), private com-
munication.
Belyaev, E. B., Godlevskii, A. P. and Kopytin, Y. D. (1978). Laser spectrochemical
analysis of aerosols, Kvant. Elektron., 5: 2594 (Sov. J. Quant. Electron., 8: 1459).
Biswas, A., Latifi, H., Shah, P., Radziemski, L. J., and Armstrong, R. L. (1987). Time-
resolved spectroscopy of plasmas initiated on single levitated aerosol droplets,
Opt. Lett., 12: 313.
Biswas, A., Latifi, H., Radziemski, L. J. and Armstrong, R. L. (1988). Irradiance and
laser wavelength dependence of plasma spectra from single levitated droplets,
AppL Opt., 27: 2386.
Brech, E and Cross, L. (1962). Optical microemission stimulated by a ruby laser,
AppL Spectrosc., 16: 59.
Bunkin, E V. and Tribel'skii, M. I. (1980). Nonresonant interaction of high-power
optical radiation with a liquid, Usp. Fiz. Nauk 130: 193 [Sov. Phys. Usp., 23: 105
(1980)].
Carr, J. W. and Horlick, G. (1982). Laser vaporization of solid metal samples into
an inductively coupled plasma, Spectrochim. Acta, B37: 1.
Cremers, D. A. (1984). Laser assisted steel sensing program review, Los Alamos,
New Mexico, March 21-22.
Cremers, D. A. (1987). The analysis of metals at a distance using laser-induced
breakdown spectroscopy,Appl. Spectrosc., 41: 572.
Cremers, D. A. (1988a). Unpublished results.
Cremers, D. A. (1988b). Mobile beryllium detector (MOBEDEC) operating man-
ual, Los Alamos National Laboratory, Los Alamos, N.M.
Cremers, D. A. and Radziemski, L. J. (1983). Detection of chlorine and fluorine in
air by laser-induced breakdown spectrometry, AnaL Chern., 55: 1252.
Cremers, D. A. and Radziemski, L. J. (1985). Direct detection of beryllium on filters
using the laser spark,Appl. Spectrosc., 39: 57.
Cremers, D. A, Radziemski, L. J. and Loree, T. R. (1984). Spectrochemical analysis

of liquids using the laser spark,Appl. Spectrosc., 38: 721.
Cremers, D. A and Radziemski, L. J. (1986). Laser plasmas for chemical analysis,
in Laser Spectroscopy and its Applications (L. J. Radziemski, R. W Solarz, and J.
A Paisner, eds.), Marcel Dekker, New York, Chap. 5, pp. 351-415.
Cremers, D. A, Archuleta, F. L. and Dilworth, H. C. (1985). Rapid analysis of steels
using laser based techniques, Proc. 5th Process Tech. Conf. on Meas. and Control.
Instr. in the Iron and Steel Industry, Iron and Steel Society, Warrendale, Pa.
Cremers, D. A, Archuleta, F. L., and Martinez, R. J. (1985). Evaluation of the con-
tinuous optical discharge for spectrochemical analysis, Spectrochim. Acta, B40:
665.
Eickmans, J. H., Hsieh, W-F., and Chang, R. K. (1987). Plasma spectroscopy of H,
Li, and Na plumes resulting from laser-induced droplet explosion,Appl. Opt., 26:
3721.
Essien, M., Radziemski, L. J., and Sneddon, J. (1988). Detection of cadmium,
lead, and zinc aerosols by laser-induced breakdown spectrometry, J. Anal. Atom.
Spect., 3: 985.
Felske, A, Hagenah, W-D., and Laqua, K. (1972). Uber einige Erfahrungen bei
der Makrospektralanalyse mit Laserlichtquellen: I. Durchschnittanalyse met-
allischer Proben, Spectrochim. Acta, B27: 1.
Generalov, N. A, Zimakov, V. P., Kozlov, G. I., Masyukov, V. A, and Raiser, Y.
P. (1970). Continuous optical discharge, Zh. Eksp. Tear. Fiz. Pisma Red., 11: 447
(JETP Lett., 11: 302).
Godlevskii, A P. and Kopytin, Y. D. (1979). Ionization and excitation of emission
spectrum of aerosols by C0 2 laser radiation, Zhumal Prikl. Spektrosk., 31:612 [1
Appl. Spectrosc. USSR, 31: 1229].
Ishizuka, T. (1973). Laser emission spectrography of rare earth elements, Anal.
Chern. 45: 538.
Laqua, K. (1979). Analytical spectroscopy using laser atomizers, in Analytical Laser
Spectroscopy (N. Omenetto, ed.), Wiley, New York, pp. 47-118.
Loree, T. R. and Radziemski, L. J. (1981). The identification of impurity metals by
laser induced breakdown spectroscopy, Proc. Electro Optics/Lasers '81, Cahners
Exhibition Group, Chicago, Ill., p. 28.
Lyamshev, L. M. (1981). Optoacoustic sources of sound, Usp. Fiz. Nauk 135: 637
[Sov. Phys. Usp. 24: 977 (1981)].
Moenke, H. and Moenke-Blankenburg, L. (1973). Laser Micro-Spectrochemical
Analysis, Crane Russak, New York (translation).
Ottesen, D. K., Wang, J. C. F., and Radziemski, L. J. (1989). Real-time laser spark
spectroscopy of particulates in combustion environments, Appl. Spectrosc. (in
press).
Ozaki, T., Takahashi, T., Iwai, Y., Gunji, K., and Sudo, E. (1984). Giant pulse laser
direct spectrochemical analysis of C, Si, and Mn in liquid iron, Trans. ISIJ, 24:
463.
Peppers, N. A, Scribner, E. J., Alterton, L. E., Honey, R. C., Beatrice, E. S.,
Harding-Barlow, I., Rosan, R. C., and Glick, D. (1968). Q-switched laser for
emission microspectroscopic elemental analysis, Anal. Chern., 40: 1178.
Petukh, M. L. and Yankovskii, A. A. (1978). Atomic emission spectral analysis using

lasers, Zh. Prikl. Spektrosk., 29: 1109 (J. Appl. Spectrosc. USSR, 29: 1527).
Radziemski, L. J., Loree, T. R., Cremers, D. A., and Hoffman, N. M. (1983). Time re-
solved laser-induced breakdown spectrometry of aerosols, Anal. Chern., 55: 1246.
Radziemski, L. J., Cremers, D. A., and Niemczyk, T. M. (1985). Measurement of
the properties of a C0 2 -laser induced air-plasma by double floating probe and
spectroscopic techniques, Spectrochim. Acta, B40: 517.
Schmieder, R. W (1983). Techniques and applications of laser spark spectroscopy,
Sandia Rep., SAND83-8618, Sandia Nat. Lab., Livermore, Calif.
Schuocker, D. (1986). "Laser cutting," in Industrial Laser Annual Handbook (D.
Belforte and M. Levitt, eds.), Pennwell Books, Tulsa, Okla., p. 87.
Scott, R. H. and Strasheim, A. (1978). Laser emission excitation and spectroscopy,
in Applied Atomic Spectroscopy (E. L. Grove, ed.), Plenum Press, New York, pp.
73-118.
Sigrist, M. W. and Kneubiihl, F. K. (1978). Laser-generated stress waves in liquids,
J. Acoust. Soc. Amer., 64: 1652.
Talmi, Y., ed. (1983). Multichannel/mage Detectors, ACS Symp. Series No. 102, ACS.
Washington, D. C.
Thompson, M., Goulter, J. E., and Sieper, F. (1981). Laser ablation for the intro-
duction of solid samples into an inductively coupled plasma for atomic-emission
spectrometry, Analyst, 106: 32.
Torok, T., Mika, J., and Gegus, E. (1978). Emission Spectrochemical Analysis, Crane
Russak, New York.
Treytl, W J., Orenberg, J. B., Marich, K. W, Saffir, A. J., and Glick, D. (1972). De-
tection limits in analysis of metals in biological materials by laser microprobe
optical emission spectrometry, Ana/. Chern., 44: 1903.
Tsunoyama, K., Tanimoto, W., Hisada, H., and Asakawa, H. (1985). Direct analysis
of molten iron by laser emission spectrometry, Trans. ISIJ, 25: B41.
Wachter, J. R. and Cremers, D. A. (1987). Determination of uranium in solution
using laser-induced breakdown spectroscopy,Appl. Spectrosc., 41: 1042.
Zhang, J. Z., Lam, J. K., Wood, C. F., Chu, B. T., and Chang, R. K. (1987). Explosive
vaporization of a large transparent droplet irradiated by a high intensity laser,
Appl. Opt., 26: 4731.
Zuev, V. E., Zemlyanov. A. A., Kopytin, Y. D., and Kuzikovskii, A. V. (1985). High-
Power Laser Radiation in Atmospheric Aerosols, D. Reidel, Boston, Mass.
8
Fundamentals of Analysis
of Solids by
Laser-Produced Plasmas
Yang W. Kim
Lehigh University
Bethlehem, Pennsylvania
8.1 CHAPTER ORGANIZATION
This chapter deals with the quantitative analysis of the elemental composi-
tion of a condensed-phase specimen by means of quantitative spectroscopic
measurements of the emission spectra from a plasma plume produced off
the surface of the specimen by focusing a high-power laser pulse. Since
such a laser-produced plasma must satisfy certain critical requirements to
become suitable for elemental composition analysis, a general discussion
of the physics of the interaction of the laser pulse with the target specimen
is first given in Sec. 8.2. A more narrowly focused phenomenology of the
laser-heating mechanisms and the evolution of the plasma are presented
in Sec. 8.3. In Sec. 8.4, the formal basis of the quantitative spectroscopy is
developed, connecting the measured line intensities to the thermodynamic
state of the laser-produced plasma. Measurement of the intensities of the
emission spectra of the laser-produced plasma and a determination of the
elemental abundance of the specimen are treated in Sec. 8.5. A summary is
presented in Sec.8.6.
8.2 INTRODUCTION
The interaction of a high-power laser beam with a condensed phase tar-

get leads to a wide range of nonlinear phenomena (Hora, 1975 and 1981).
Of particular interest in the context of the elemental composition analy-
sis of condensed-phase specimens is the formation of a plasma plume that
327
328 Kim
grows out of the surface when the incident laser irradiance exceeds a cer-
tain threshold value, typically in the neighborhood of 108 W/cm2 (Carroll
and Kennedy, 1981). The high-power lasers used for this purpose can be
any one of the pulsed lasers operating in the ultraviolet to infrared range
with a pulse duration of 5 to 100 nsec. The threshold power density varies
with the operating wavelength of a laser primarily because the absorptance
of the target surface depends on the wavelength of the incident light.
In the laser power density regime where a plasma plume can be produced
from the target surface, the leading edge of the laser pulse is absorbed by
the surface layer and heats it rapidly to vaporize the condensed-phase mat-
ter into gases. If in a solid form, a part of the absorbed laser power must
be expended into the latent heat of fusion, in addition to the latent heat
of evaporation. The latent heat of fusion is not involved in the case of a
liquid. The heating of the surface layer, of course, means a partial loss of
the absorbed power into the bulk as a heat sink by the thermal conduc-
tion process. The early vapor-phase species interact with the incident laser
pulse to be excited internally and ionized. For sufficiently short wavelength
lasers, an ionization event may be a bound-free transition by single-photon
absorption; for long wavelength lasers, multiphoton absorption processes
play a significant role (Eberly and Lambropoulous, 1978). Similarly, free
electrons are also released from the target surface by the single- or multi-
photon photoelectric effect. Collisions among these particles lead to a rise
in temperature and pressure of the gas. Thermal ionization, however, re-
mains a mechanism for the production of charged species at all stages of
laser heating.
The process of evaporation and heating continues until the electron den-
sity of the partially ionized gas becomes high enough for the heating of
the gas by the inverse bremsstrahlung process to begin. The weakly ion-
ized plasma becomes gradually coupled to the middle segment of the laser
pulse, while remaining partially transparent to the laser beam so that di-
rect heating of the target surface continues. The inverse bremsstrahlung
process heats primarily the electrons in the presence of ions, resulting in
free-to-free-state transitions. Such heated electrons increase the plasma
temperature and, consequently, the electron density. At high laser pow-
ers, the electron density of the plasma can attain a value sufficiently high
so that the core of the plasma becomes opaque. In this case, the frequency
of plasma oscillation, as given by (nee 2I 1rm)112, is comparable to or greater
than the laser frequency and the laser beam is no longer able to penetrate
the plasma and reach the condensed-phase target surface. Here, ne is the
electron number density, e the electronic charge, and m the electron mass.
Heating of the target surface can continue only indirectly by thermal con-
duction from the plasma.
Analysis of Solids 329
Laser heating of the plasma continues and the critical region advances
toward the laser, thus resulting in the growth of the plasma plume toward
the laser (Basov et al., 1972). The heating rate of the plasma, of course,
increases sharply at high electron densities by both the linear process of in-
verse bremsstrahlung and nonlinear processes that come into play. Mul-
tiply ionized, high-kinetic energy species are produced (Bykovskii et al.,
1971). These energetic particles escape the plasma essentially ballistically.
Typically, a rapid expansion of the plasma due to the energetic particles
is seen beginning at around the time corresponding to the peak intensity
point of the laser pulse. The expansion velocity is in the neighborhood of
2 x 107 em/sec, quite independently of the ambient gas species. Because of
the above-mentioned mechanism by which the plasma plume develops to-
ward the laser, the energetic particles become increasingly more energetic
in time and more directional as the number of times a particle has been ion-
ized increases. Consequently, the radial expansion of the plasma is charac-
terized by greater concentration of neutral and singly ionized species than
the axial expansion toward the laser that exhibits the prominence of multi-
ply ionized species.
The growth and heating of the plasma at high temperature and high den-
sity continues for the balance of the laser pulse. The plasma up to the end of
the laser pulse is optically thick throughout the entire ultraviolet to infrared
spectral range and, consequently, the observable spectrum is characterized
by the surface temperature of the plasma, which is of the order of 104 to
105K. The strong trapping and diffusion of radiation, together with the high
density and temperature state of the plasma, make the plasma extremely
high in energy density. This fuels the expansion and sustains the luminosity
of the plasma for a long time after the expiration of the laser pulse. De-
pending on the density of the ambient gases and other factors, the postpulse
plasma lifetime may be anywhere from about 300 nsec to more than 40 p,sec
in length.
The high temperature and density of the laser-produced plasma with-
out any external confinement are a direct result of the extremely high rate
of heating made possible by the high-power laser pulse. The expansion of
the main plasma becomes noticeable near the end of the laser pulse with
the launching of a shock wave into the gas medium surrounding the plasma
as a direct result of the expansion. The gas medium may be primarily the
ambient gas that existed before the laser pulse or, in the case of a vacuum
environment, it may be the neutral gas species resulting from the escape of
the fast ions. The region behind the shockwave consists of the shock-heated
ambient gas followed by the expanding plasma. The observed luminous core
of the plasma decreases in size as a function of time because of the cooling
associated with the expansion.
330 Kim
From the standpoint of elemental composition analysis of condensed-

phase specimens, there are two particular segments of time in the full span
of the evolution of a laser-produced plasma that are of critical importance.
The first segment is the very beginning of the interaction of the laser pulse
with the target surface. The critical issue here is the kinetics of the for-
mation of a plasma whose elemental composition is identical to that in the
condensed-phase target. The second segment is the period after the inci-
dent laser pulse is extinguished, when the emission from the plasma under-
goes a transition from a strong continuum emission to line emission charac-
ter as a result of the changing thermodynamic state of the plasma. It is here
that the quantitative determination of the elemental abundances is made.
8.3 PHENOMENOLOGY OF LASER HEATING

OF CONDENSED-PHASE TARGETS
The intensity profile of a typical Q-switched laser pulse has an approxi-

mately trapezoidal shape in time. The leading-edge rise-time and the trail-
ing edge fall-time are different in general, although they are adjustable in-
dividually to some extent by varying the characteristics of the high-voltage
pulse-switching network of the Pockels cell Q switch. The slope of laser-
intensity rise in the leading edge is of the utmost importance in attaining a
plasma plume that is both representative of the condensed-phase specimen
in its elemental composition and robust in its emission characteristics for
quantitative spectroscopy. The leading edge must meet two requirements.
It must be able to heat the target at a rate (1) fast enough to remove the
surface layers at a speed equal to, or greater than, that of the thermal con-
duction front moving into the bulk, and (2) slow enough to allow continued
evaporation without triggering runaway absorption by the plasma plume
prematurely.
In the case of solid specimens, there are two interesting variations in the
response of the target to laser heating. One is that the reflectance of the tar-
get surface at the laser wavelength undergoes a sharp drop at the onset of
the melting transition of the surface layer (Walters and Clauer, 1978). This
brings about a sudden increase in the fraction of the actual laser power ab-
sorbed by the surface and accelerates the process of plasma formation. The
second variation is that the very first laser pulse incident on the surface pro-
duces a plasma that is always different than those produced by subsequent
pulses incident on the same surface area. This is due to the combined ef-
fects of foreign surface substances due to prior exposure to ambient gases
and of preexisting morphological features that are different from the bulk.
The first laser exposure helps eliminate the surface abnormalities, paving
the way for reproducible plasma production.
Figure 8.1 gives a sequence of streak camera photographs taken of a se-

ries oflaser-produced plasmas produced from a solid-steel alloy target. The
Nd:glass laser beam (2 J and 65 nsec full width at half maximum) was inci-
dent normal to the target surface. The plasma plume was viewed through a
linear bundle of optical fibers arranged parallel to the target surface. The
receiving end of the fiber was placed on the image plane of a lens that pro-
jected the side view of the plasma at a magnification suitable to capture all
or a part of the plasma cross section by the fiber bundle. Figures 8.1(a)
through 8.1(h) show the radius-time views of the plasma sampled at dis-
tances from the target surface of 0, 1.2, 2.5, 3.7, 5.0, 6.2, 7.5, and 8.7 mm,
respectively. The exit end of the fiber bundle is placed at the entrance slit
of the streak camera. The individual lines discernible in each photograph
represent the white light from the plasma collected by individual fibers. The
fibers were 200 f..lm in diameter and evenly spaced. The ambient gas was air
at 0.05 Torr. Each streak photograph was taken from an individually pro-
duced plasma under identical conditions of laser power, pulse shape, target,
and ambient gas.
In the figures, a number of the distinct features of the plasma mentioned
earlier are readily observed. First, in Fig. 8.1(a) one can detect the appear-
ance of luminosity from the plasma right at the target surface within about
5 nsec of the start of the laser pulse. No such features are seen in Figs. 8.1( d)
through 8.1(h), indicating spatial localization of the early plasma. Begin-
ning at about 70 nsec, the rapid growth of the plasma due to the ballistic
motion of fast ions and neutral species appears. This coincides roughly with
the maximum intensity point of the laser pulse. The expansion velocity of
the luminous front is approximately 2.5 x 107 em/sec and this corresponds
to the mean kinetic energy of 18.2 ke V of neutral and singly ionized iron
atoms. This rapid expansion phase lasts for about 200 nsec. Figures 8.1(c)
through 8.1(h) show the distinct delayed appearances of the fast expansion
phase with increasing distance from the target, demonstrating its origin at
the plasma that is localized near the surface.
The next feature is the shock front that is seen in the form of an expand-
ing luminous front. It begins at about the tail end of the laser pulse and is
driven by the dense plasma core that is expanding. The shock velocity de-
pends on the ambient gas pressure, and the luminosity is due to both the
shock-heated plasma of the ambient gas species and the original plasma
species. In view of the three-dimensional character of the expansion, the
shock speed decreases in time as can be seen by the decreasing slope of
the front with increasing time. The maximum shock speed is approximately
2.4x 106 em/secor one-tenth the velocity of the fast expansion front that pre-
ceded the shock wave. It is the shock wave that leads to an acoustic pulse at
large distances from the target and also is responsible for driving away the
residual molten layer of the target out of the region of laser illumination.
332 Kim
a) [
0
b) [
0
c) [
0
d) [
0
e) [
0
f) [
0
I I I I I
·2 0 2
r, c:m
Figure 8.1 Streak photographs of a laser plasma produced off a solid steel alloy
target. The radial expansion of the plasma is observed as a function of time at dif-
ferent distances from the target surface: a) 0; b) 1.2; c) 2.5; d) 3.7; e) 5.0; f) 6.2;
g) 7.5; h) 8.7 nm. A2 Joule 1.06 J.Lm laser pulse of65 nsec full width at half intensity
was used.
Finally, one can see a considerably slower evolution of the plasma core,
consisting of an expansion followed by a contraction of the luminous front.
This final phase of the plasma and the shock wave it triggers are seen clearly
separated in time from the first fast expansion front of energetic particles
when one views the plasma at different distances from the target surface
[see Figs. 8.1( c) through 8.1(g)]. The afterglow of the plasma core goes on
for about 1 psec.
We will now examine the spectral characteristics of the emission from
the plasma as a function of position with respect to the focal region of the
laser beam on the target surface and of time. Figure 8.2 shows the emission
spectra from the segment of a plasma located on the beam axis at 1 mm
away from the target. There are altogether 20 spectra, each taken during a
20 nsec interval but separated by 30 nsec successively starting at the onset
of the laser pulse. The intensities are uncalibrated, obtained with a Reticon
6000
5000
4000
.::l
0
3000
X
.
"c
J
2000
0
u
1000
0
2500 3000 3500 4000 4500 5000 5500 6000
Wave~. A
Figure 8.2 Time resolved emission spectra from laser produced plasma. The spec-
tra were taken from the segment of the plasma located 1 mm away from the target
surface along the laser beam axis. Each trace represents an uncalibrated intensity
spectrum integrated over a time duration of 20 nsec, separated by 30 nsec from the
preceding trace. The experimental conditions are the same as those for Fig. 1.
334 Kim
photodiode array detector fitted with a multichannel plate intensifier (for

amplification and time gating) having a multialkali photocathode of Sl-UV-
A responsivity. Each spectrum is a composite of two separate shots covering
the two halves of the spectral region. Such spectra are taken by delaying the
20 nsec optical gate by 30 nsec increments for succeeding shots.
The emission spectrum displays a continuum character right at the onset
of the plasma, indicating an extremely rapid heating. The continuum inten-
sity continues to increase until after the expiration of the laser pulse as a
consequence of the contributions from the bremsstrahlung process, recom-
bination, and line broadening by the Stark effect. Starting at about 150 nsec,
the continuum intensities undergo a decline, leading the way to the emer-
gence of the line spectral character. However, the emission lines cannot
be seen because the gain of the intensifier is kept constant in order to give
an overview of the evolution of the emission characteristics. Self-reversed
lines (Kim and Sincerny, 1985), mostly of the resonance lines of the con-
stituent elements in the target, are seen as depressions in the continuum
intensity profile, and with a decrease of the optical thickness of the plasma,
these grow in time into prominent individual emission lines with little back-
ground continuum intensities. Establishment of the line emission charac-
teristics coincides with the later stages of the plasma afterglow discussed
earlier.
When the plasma is near its peak density and temperature, the laser-
plasma interaction enters into a nonlinear regime. One of the measurable
consequences is a generation of the second harmonic of the incident laser
beam (Banfi and Gobbi, 1983). Starting at 60 nsec, the spectrum shows a
noticeable hump of intensity on the longer wavelength shoulder of the con-
tinuum, which is a blue Doppler-shifted second harmonic of the laser beam.
The Doppler shift by approximately 20 A indicates the expansion velocity of
the plasma core at 1.13 x lOS em/sec, which is consistent with the mean ve-
locity of the fast ions and neutrals at 2.5 x 107 em/sec in the peripheral region
of the plasma as cited earlier. The second-harmonic generation ceases at
about 120 nsec, corresponding to the onset time of the decay of the plasma
afterglow.
Figure 8.3 shows the emission spectra from the segment of the plasma
that extends out to a region 10 mm from the target surface and 10 mm from
the laser beam axis. The gain of the multichannel plate intensifier was in-
creased to 8 x 106 compared to 106 used to obtain the data in Fig. 8.2. The
first detectable emission spectrum appears about 60 nsec after the onset
of the laser pulse and shows some continuum intensity profile, but the line
emission character is quite pronounced. This is entirely consistent with the
earlier observation that the luminosity in this region at this time is due to
the fast ions and neutrals moving ballistically away from the plasma core
near the target surface. After a 40 nsec period corresponding to the time
!BOO
1600
1400
1200
1000
0
0
,; 800
•
"
[
J 600
0
u
400
200
0
2000 2500 3000 3500 4000 4500 5000 5500 6000
Wavelength, A
Figure 8.3 Time resolved emission spectra from a laser produced plasma. The
spectra were taken from the segment of the plasma located 10 mm away from the
target surface and 10 mm away from the laser beam axis. The experimental condi-
tions are the same as those for Fig. 2.
delay between the fast ion front and the shock front arrivals, intense emis-
sion lines emerge. These are due to the shock heating of the ambient gas
and residuals from the earlier expansion of the fast ions and neutrals from
the plasma of the target elements. Finally, the expanding plasma core with
a full display of a rich emission line spectrum is observed.
The details of the emission line spectrum change significantly from one
regime of the plasma to another, largely because the plasma temperature
changes substantially. It is the temporal and spatial changes of the plasma
that present the windows of opportunity for analysis by quantitative spec-
troscopy. Since in the peripheral regions, as observed in Fig. 8.3, the domi-
nant physical process is the expansion of the plasma, the temperature may
become undefined in the following sense. A state of local thermodynamic
equilibrium may not exist if the particle number density is too low and the
collision time is of the order of the time interval during which the plasma
336 Kim
state undergoes a significant change. We will return to this aspect later in

the context of the spectroscopic interpretation of elemental abundances.
8.4 QUANTITATIVE SPECTROSCOPY
Determination of the elemental composition of a condensed-phase speci-

men by means of laser-produced plasmas requires the measurement of the
intensities of those spectral lines that are characteristic of the individual
species present in the specimen. The intensities must then be related to the
number densities of the species present in the plasma. The relationships
turn out to depend on the temperature of the plasma where the species are
present, as we will see in the development that follows.
Consider a plasma of depth D in a state of local thermodynamic equilib-
rium at temperature T. The specific intensity at frequency f of an emission
line is defined as the energy emitted from the plasma per unit time, unit sur-
face area, unit solid angle, and unit frequency. The change in the specific
intensity I (f) along the line of sight is given by the equation of radiative
transfer (Chandrasekhar, 1960)
d~) =K(f)[B(f,T)-I(f)] (8.1)
Here, x is the coordinate axis along the line of sight perpendicular to the
surface area of the plasma, and B(f, T) is the Planck blackbody function.
The absorption coefficient of the plasmaK(f) that includes a correction for
the stimulated emission is given by
(8.2)
h is the Planck constant, ni the number density of the species of interest in

the lower of the two energy levels between which the radiative transition
for the emission line takes place, and k the Boltzmann constant. P(f) is the
line profile function of the emission line whose center frequency is fo and is
normalized so that
11o+oo
fo-oo
P(f)df =1 (8.3)
The Einstein coefficient of absorption Bij is related to the Einstein coeffi-

cient of spontaneous emission or radiative transition probability Aji and of
stimulated emission Bji as follows:
(8.4)
(8.5)
where gi and gj are the degeneracies of the lower and upper energy levels,
respectively, and cis the speed of light in the vacuum.
Equation (8.1) can readily be integrated with respect to x between the
two confining planes of the plasma separated by D to obtain
I (f) =B(f,T)[1-e-Kif)D] (8.6)
The constant of integration has been set to zero because here we have no
external source of radiation outside of the plasma. The product K(f)D is
a measure of the optical thickness of the plasma; the plasma is considered
optically thin if the K (f)D value is small compared with unity and optically
thick if large. In the limit of the K(f)D value growing infinitely large, the
specific intensity is given entirely by the blackbody function, reducing the
plasma to a blackbody surface radiator at temperature T. In the same limit,
Eq. (8.1) indicates that the specific intensity remains constant throughout
the plasma, correctly describing the state of a blackbody as that in which
both the particles and photons exist in a state of complete thermodynamic
equilibrium.
It is customary and convenient to measure the integrated intensity of an
emission line with center frequency fo rather than the specific intensity. The
integrated intensity can be defined by
I= l fo+oo
fo-oo
I(f)df (8.7)
With Eq. (8.6), we have
I~ B(/0 , T) l fo+oo
fo-oo
[1- e-K(f)D]df (8.8)
Here, B(f, T) was replaced by B(f0 , T) and taken out of the integral because
the blackbody function remains essentially constant over the width of a sin-
gle emission line, over which the integration with respect to frequency takes
place.
The integrated intensity as given by Eq. (8.8) grows with the concen-
tration of the species in the plasma through the number density ni of the
species populating the lower energy level designated by i, as contained in
the expression of K(f)D. For a given value of the total number density of
the species, ni is a strong function of temperature and is given by the Boltz-
338 Kim
mann distribution
no15·e-E;IkT
n·
1
= --"'""~---
Q (8.9)
n 0 is the total number density of the species in the same state of ionization
and Ei the energy of the ith level as measured in reference to the ground-
state level of the species in the same state of ionization. The single-particle
canonical partition function Q for the electronic degrees of freedom of the
species in a particular state of ionization is
(8.10)
The sum is over all allowed electronic energy levels. At low temperatures,
the partition function may be approximated by the degeneracy of the ground
state if the first excited state energy is very large compared With kT.
The concentration dependence of the integrated intensity is, in general,
nonlinear. This is because effectively the optical depth of the plasma de-
creases with increasing concentration, reducing the number of radiators
participating in the observable emission intensity below the actual available
number. In the low concentration limit where the plasma becomes optically
thin, the integrated intensity grows linearly with the number density. In this
limit, the integral of Eq. (8.8) can be expanded in a Taylor series and the in-
tegral evaluated by keeping only the lowest-order term. The result is
f1A ..g.n
I = h~':1
()L'· 11 1
e-EjlkT
o (8.11)
Q
where the normalization of the line profile function has been used, together
with the expression for the Planck blackbody function.
Calculation of the intensity vs. concentration relation requires knowl-
edge of the line profile function P(f), which is in general given by a con-
volution integral of two or more line profile functions corresponding to
the line-broadening mechanisms in effect within the plasma (Griem, 1974).
For plasmas that are modestly optically thick and therefore conveniently
suited for quantitative spectroscopy, the degrees of ionization of constituent
species are small and the line broadening by the Stark effect is modest. Then
it is possible to represent, to a good approximation, the line profile function
by the Voigt function
(8.12)
Here 6.fL is the half-width at half-intensity of the Lorentzian profile func-

tion representing the collisional broadening mechanism, which is, in this
approach, generalized to include the contributions from the Stark effect.
Also, 6.fv is the half-width at e- 1 intensity of the Gaussian profile func-
tion arising from the Doppler broadening, and it is given by (2kT/M) 112f 0/c,
where M is the mass of the radiating atom. Operationally, 6.fL is the half-
width at half-intensity of the emission line profile minus the Doppler width,
although in the strict sense of pure collisional broadening, 6.fL is given by
(27rtc)- 1, tc being the time between collisions.
Once the line profile function is given in some form, Eq. (8.8) can be
evaluated to determine the functional dependence of the integrated inten-
sity on the number density of the species of interest. Figure 8.4 shows the
result of such a calculation using a Gaussian function to represent P(f).
Five different values of the ratio of 6.f, the half-width at half-intensity, to
the Doppler width 6.fv are used in order to show the effect of the line width.
The important point to note is the manner in which the intensity vs. density
relationship becomes nonlinear with the increasing optical thickness of the
plasma.
Another factor that affects the optical thickness is the magnitude of the
Einstein coefficient Aji· The larger the value of Aji• the more readily the
intensity-density relation becomes nonlinear.
In summary, it has been shown that the integrated intensity of an emis-
sion line of an elemental species can be expressed in terms of atomic and
thermodynamic properties, including species concentration, given that the
plasma is in the state of local thermodynamic equilibrium. When the in-
tensity is measured and all other properties are known, the number density
of the species can be determined. The ease and accuracy with which such
a determination can be made depend on the extent to which the plasma is
optically thin or how well the plasma properties are known through other
measurements. It is quite evident from the preceding discussions that the
successful analysis of condensed-phase specimens by laser-produced plas-
mas depends on the selection of a favorable region in time and space of the
plasma for quantitative spectroscopic measurements. This has been verified
experimentally as well (Radziemski et al., 1983).
Once the selection is made, there are several ways such an analysis
program can be implemented. One approach is to determine each ele-
mental abundance independently by working with those elemental emis-
sion lines of accurately known transition probabilities and by making ab-
solute measurements of the integrated intensities. A second approach is
to measure the integrated intensities of the elemental emission lines rela-
tive to that of a line of the most abundant element of the specimen. In this
case, the transition probabilities of the emission lines need to be known
only relatively. A third method is to measure the intensities relatively
340 Kim
5 .t>f /.t>f 0 =
10
4 5
3
3 2
2
-5
-6
-7
-8
-9
-10 ~~--~~--~~--~~~~--~~--~~
-4 -3 -2 -1 0 1 2 3 4 5 6 7 8
ln[hf0 8jj 0(1-ehfo/kT)nj~]
Figure 8.4 Integrated line intensity as a function of the number density of the emit-
ting species in the plasma. The effect of the line width is parameterized by the ratio
of the half width at half intensity D.f to the Doppler half width D.fv·
and calibrate their relationships to the respective elemental abundances

by working with specimens whose elemental composition has been ana-
lyzed by other means (i.e., calibration standards). The latter is most of-
ten done in analytical spectrochemistry; however, suitable combinations of
the above three approaches may also prove to be appropriate. We concen-
trate below on the first two methods that are especially useful when cali-
bration standards are not available (as is often the case with in situ sam-
pling).
8.5 INTENSITY MEASUREMENTS

AND ELEMENTAL ANALYSIS
The integrated intensity of an emission line can be measured by a number of

means. The simplest and also quite accurate method is to set the entrance
slit of a spectrograph wide enough so that its width on the image plane is
slightly wider than the spectral width of the emission line. The image plane
slit width is given by the product of the geometrical width of the slit on the
image plane and the linear dispersion of the spectrograph. The resultant
intensity at the line center is indeed the integrated intensity. This comes
about as follows. When the entrance slit is infinitesimally thin, the resul-
tant image of the slit displays the true line profile. A finite slit consists of
an infinitely many infinitesimal slits that are spatially displaced from one
side of the slit to the other. Consequently, a point at the center of the line
on the image plane is illuminated by the successive images of the infinitesi-
mal slits that are continuously displaced from one end of the profile to the
other. The net result is the optical integration of the line profile function.
Whether one has indeed achieved the integrated intensity is readily tested
by the further widening of the slit; the intensity at the line center should re-
main unchanged, provided that the continuum background is either negli-
gible or subtracted out. It is, of course, implicitly assumed that two adjacent
lines are separated wider than either one of the respective line widths.
Such integrated intensities of the emission lines for the plasma can be
measured by any one of several types of detectors, such as photomultipli-
ers, photodiodes, and spectroscopic films. In the case of photoelectric de-
tectors, the resulting signals may be handled as photoelectric current or by a
photon-counting scheme. For films, the signals will be in the form of black-
ening as quantified by the neutral density of the exposed area. In any event,
the integrated intensity can be obtained once the detector sensitivity is es-
tablished, either by direct calibration with a radiation standard or by com-
piling all of the intrinsic properties of the detector's data chain elements.
The external optics must also be analyzed in regard to the light-gathering
solid angle; reflective, absorptive, and scattering losses, and the definition of
the plasma volume contributing to the intensity. The efficiency of the spec-
trograph (or any other dispersive devices employed in the measurement)
enters into the analysis. Another effective approach is to calibrate the en-
tire detection system as a function of wavelength with a calibrated radiation
standard source (e.g., tungsten lamp) placed at the plasma source.
The plasma temperature can be measured by the line intensities to a vary-
ing degree of accuracy depending on the optical thickness of the plasma. In
the optically thin limit, the intensity ratio of a pair of emission lines of one
atomic species in a given state of ionization can be written from Eq. (8.11)
342 Kim
as
(8.13)
All quantities in Eq. (8.13) are known either by measurements or compi-

lations of atomic properties, and therefore Eq. (8.13) can be solved for the
temperature. Since the transition probabilities are known with considerable
variation in accuracy, say, from 1 to 25% (Corliss and Bozman, 1962; Wiese
et al., 1966, 1969; Gray, 1972; Reader et al., 1980) the selection of the emis-
sion lines for this application should be made accordingly. In general, how-
ever, the relative transition probabilities are much more accurately known
than the absolute values and this helps the temperature measurement.
The above measurement must be repeated several times with different
pairs of emission lines from the same atomic species as well as with those of
other species present in the plasma. From the distribution of the resultant
temperature values, a determination can be made of the degree to which
the plasma has established a state of local thermodynamic equilibrium. As
a general rule, a state of local thermodynamic equilibrium exists when the
collision time in the plasma is small compared with the characteristic time
over which the state of the plasma changes significantly (Griem, 1964). For
the laser-produced plasmas of interest here, the characteristic time ranges
from a few nanoseconds in the leading-edge part of the laser pulse to a few
tens of nanoseconds in the post-pulse plasma afterglow.
Once the plasma temperature is established and the optically thin seg-
ment of the laser-produced plasma identified, elemental composition analy-
sis may readily be carried out using the integrated line intensities. Consider
two emission lines: one from an elemental atomic species a and another
from the species b. The intensity ratio is obtained from Eq. (8.11)
(8.14)
The expression for the intensity ratio now contains the ratio of the partition
functions and the number densities of the two species, in addition to those
of the frequencies, transition probabilities, and degeneracies of the upper
levels, and the energy difference between the two upper levels. All ratios
are known either from measurement or compilation of atomic properties.
Alternatively, the lumped factor
(foAjigj )a Qb
(foAjigj )b Qa
can be determined by calibration with a plasma containing known amounts

of the two elemental species. The temperature dependence of the partition
function ratio Qb!Qa is readily and accurately calculated [see Eq. (8.10)],
giving a means of making the necessary adjustment to the factor for its tem-
perature dependence. Once established, Eq. (8.14) can be routinely applied
to measure the elemental abundance of the element a in relationship to b. It
follows that for multielement specimens, it is desirable to choose the most
abundant species as a reference for determination of all other elemental
abundances, thus providing the basis for establishing the percent concen-
tration of each elemental species in the specimen.
Optically dense plasmas are considerably more complex to interpret for
elemental composition analysis than those in the optically thin limit but the
basic approach remains the same. The starting point is the expression for
the integrated intensity, Eq. (8.8). The integrated intensities as well as the
line profile functions may have to be measured. If the plasma has a domi-
nant elemental species, one must determine whether the dominant species
or the minority species control the electron density in the plasma because
the changes in the concentration of the minority species may or may not af-
fect the line profile function depending on its role in controlling the electron
density. An effective way of making headway is to interpret tht: measure-
ments iteratively by making successively more realistic assumptions about
the plasma state, line profile, and the intensity vs. density relation. It is
always helpful to work with emission lines that have small transition proba-
bilities in order to offset the effect of high number densities and reduce the
optical thickness. The line intensities of such alternative lines, however,
must not fall too low so as not to sacrifice a good signal-to-noise ratio for
the intensity measurements.
There is also a consideration of the degrees of ionization of all of the el-
emental species present in the plasma. When each degree of ionization is
not negligibly small, i.e., the number density of singly ionized ions of a given
elemental species is not negligible compared with that of the neutrals, the
number density n0 of Eq. (8.9) becomes an underestimate of the total num-
ber density of the given elemental species. This difference increases with
decreasing atomic ionization potential according to the Saba equation, and
therefore, it introduces errors into the elemental composition analysis. The
Saha equation is a variation of the law of mass action, which is adapted to
ionization as an equilibrium chemical reaction process, and gives the de-
gree of ionization in terms of the thermodynamic variables of the plasma
and the atomic properties of all species appearing in the reaction equation.
For detailed discussions of this point, the Saha equation, and other aspects
of optically thick plasmas, which are outside of the scope of this chapter,
the readers are directed to the rather large body of literature on plasma
spectroscopy (Griem, 1964).
344 Kim
The basic assumption underlying the above development has been that
the laser-produced plasma does indeed have the elemental composition
of the target specimen. The necessary requirement is to choose the laser
power, pulse rise-time, and focal-beam spot size in such a way that the
evaporation front travels into the bulk in pace with, or ahead of, the ther-
mal front. The rationale for this criterion is based on the fact that the
rate of evaporation is a strong function of the elemental species in the
molten state (Kim, 1978; Harding and Kim, 1982; Porter et al., 1982).
It follows that when the heating rate is increased to the point where the
time spent by a volume element of the target specimen in the molten state
approaches zero, the plasma can no longer be constituted preferentially
by more volatile elements present in the condensed phase. The thermal
front moves diffusively over the distance growing as (2at) 112 , where a is
the thermal diffusivity of the bulk solid matter and t the time. Labora-
tory measurement and some numerical simulation calculations show that
the laser energy actually involved in the evaporation of the target surface
layer is typically of the order of 10 ppm of the total incident laser en-
ergy.
For a general review of a number of laser plasma techniques used for
the chemical analysis of gases, liquids, and solid surfaces, the readers are
directed to Chap. 7 and 9 in this book and a considerable body of literature
on the analytical uses of laser plasmas (Cremers and Radziemski, 1987).
8.6 SUMMARY
High-power lasers have been seen to be capable of producing a robust

plasma from a condensed-phase target specimen. Each plasma evolves
through a wide range of electron density, temperature, and pressure in
a time span of the order of 1 11sec. Under sufficiently high irradiance,
the plasma can be produced in the manner truly representative of the
specimen in its elemental composition. It is then always possible to
identify a segment of such a plasma somewhere in the afterglow region
that is accessible for analysis by quantitative spectroscopy. There is con-
siderable latitude in the application of the spectroscopic techniques in
terms of the detectors, intensity measurements, handling of atomic con-
stants, and the optical thickness of the plasma. The optimum avenue
of analysis may be guided by considerations of such aspects as the mor-
phology and stability of the specimens, desired accuracy, reproducibil-
ity, turnaround time, laser characteristics, and data processing capabili-
ties.
ACKNOWLEDGMENTS
The author acknowledges the able assistance of H. Chang, J. Kralik, and

K. Lyu in obtaining in the author's laboratory the experimental data used
in Figs. 8.1 through 8.3. Thanks are also owing to S. Kim for evaluating the
published work on laser-produced plasmas.
REFERENCES
Banfi, G. B. and Gobbi, P. G. (1983). Total second harmonic emission from a laser
produced plasma, Optics Comm., 44: 337.
Basov, N. G., Boiko, V. A, Gribbov, V. A, Zakharov, S. M., Krokhin, D. N., and
Sklizkov, G. V. (1972). Gas dynamics of laser plasma in the course of heating,
Sov. Phys. JETP, 34:81.
Bykovskii, Y. A., Degtyarenko, N. N., Elesin, V. F., Kozyrev, Y. P., and Sil'nov, S.M.
(1971). Mass spectrometer study of laser plasma, Sov. Phys. JETP, 33: 706.
Carroll, P. K. and Kennedy, E. T. (1981). Laser-produced plasmas, Contemp. Phys.
22:61.
Chandrasekhar, S. (1960). Radiative Transfer, Dober Publication, New York.
Corliss, C. H. and Bozman, W R. (1962). Experimental Transition Probabilities for
Spectral Lines of Seventy Elements, Nat. Bureau of Standards Monograph 53,
Washington, D. C.
Cremers, D. A. and Radziemski, L. J. (1987). Laser plasmas for chemical analysis,
in Laser Spectroscopy and Its Applications (L. J. Radziemski, R. W Solarz, and J.
A Paisner, eds.), Marcel Dekker, New York, p. 351.
Eberly, J. H. and Lambropoulous, P., eds. (1978). Multiphoton Processes, Wiley, New
York.
Gray, D. E., ed. (1972).American Institute of Physics Handbook, 3rd ed., McGraw-
Hill, New York.
Griem, H. R. (1964). Plasma Spectroscopy, McGraw-Hill, New York.
Griem, H. R. (1974). Spectral Line Broadening by Plasmas, Academic Press, New
York.
Harding, T. Wand Kim, Y. W (1982). Direct sampling of gas and particulate from
electric arc furnaces, in Physics in the Steel Industry,AIP Con[. Proc. No. 84 (F. C.
Schwerer, ed.) AlP, New York.
Hora, H. (1975). Laser Plasmas and Nuclear Energy, Plenum Press, New York.
Hora, H. (1981). Physics of Laser Driven Plasma, Wiley, New York.
Kim, Y. W (1978). Kinetics of ion production from particle boron, in Shock Tube
and Shock Wave Research (B. Ahlborn, A. Herzberg, and D. Russell, eds.), Univ.
of Washington Press, Seattle, Wash., p. 446.
Kim, Y. W and Sincerny, P. (1985). Reversal line profile of Cu resonance lines in
a non-ideal hydrogen plasma, Proc. of the 17th Internal. Conf. on Phenomena in
Ionized Gases, Budapest, p. 1001.
346 Kim
Porter, J. R., Goldstein, J. 1., and Kim, Y. W. (1982). Characterization of directly

sampled electric arc furnace dust, in Physics in the Steel /ndustry,AIP Conf Proc.
No. 84 (F. C. Schwerer, ed.), AIP, New York, p. 377.
Radziemski, L. J., Loree, T. R., Cremers, D. A., and Hoffman, N. M. (1983). Time-
resolved laser-induced breakdown spectrometry of aerosols, Anal. Chern. 55:
1246.
Reader, J., Corliss, C. H., Wiese, W. L., and Martin, G. A. (1980). Wavelengths and
Transition Probabilities for Atoms and Atomic Ions, Nat. Standard Ref. Data Se-
ries NBS 68, Washington, D. C.
Walters, C. T. and Clauer, A. H. (1978). Transient reflectivity behavior of pure alu-
minum at 10.6 J.Lm,Appl. Phys. Lett., 33: 713.
Wiese, W. L., Smith, M. W., and Glennon, B. M. (1960). Atomic Transition Proba-
bilities, Vol. 1, Hydrogen Through Neon, Nat. Standard Ref. Data Series NBS 4,
Washington, D. C.
Wiese, W. L., Smith, M. W., and Miles, B. M. (1969). Atomic Transition Probabili-
ties, Vol. 2, Sodium Through Calcium, Nat. Standard Ref. Data Series NBS 22,
Washington, D. C.
9
Laser Vaporization for Sampie
Introduction in Atomic and
Mass Spectroscopy
Joseph Sneddon
Department of Chemistry
University of Lowell
Lowe[~ Massachusetts
Peter G. Mitchell
Department of Medicine
Medical College of Wzsconsin
Milwaukee, Wzsconsin
Nicholas S. Nogar
9.1 CONVENTIONAL SOLID SAMPLE

INTRODUCTION METHODS FOR
ATOMIC SPECTROSCOPY
Methods of atomic spectroscopy [atomic absorption spectroscopy (AAS),

atomic emission spectroscopy (AES), and atomic fluorescence spectroscopy
(AFS)] for metal determination in a wide variety of complex samples have
become widely used and accepted techniques. Their advantages include
low detection limits, high sensitivity, wide linear dynamic range, high selec-
tivity, sequential or multielement capability, widespread availability, ease
of operation, relative freedom from interferences, good accuracy, and ac-
ceptable precision. The most widely used techniques are AAS with a flame
and electrothermal atomizer as the atomization-excitation source. AES is
rapidly gaining popularity due to the use of the plasma [inductively cou-
pled plasma (ICP), the direct current plasma (DCP), and the microwave in-
duced plasma (MIP)] as an excitation source. This leads to improved sen-
347
348 Sneddon et al.
sitivity over the traditional flame used in AES. AFS is not a widely used
technique. A comparison of AAS and plasma-AES is available in Routh
(1986), where the relative merits of sample throughput, detection limits,
short-term precision, long-term reproducibility, dynamic range, interfer-
ences, element range, sample compatibility, qualitative analysis, ease of use,
purchase price, and operating costs are discussed. He concluded that AAS
and plasma-AES were capable of performing elemental analysis of solutions
with good accuracy and precision and proposed that the technique used will
depend on the specific analytical requirement.
The area of sample introduction to atomic spectroscopy has created
much interest in recent years and has been reviewed by Browner and Boom
(1984) and Sneddon (1986). Conventional sample introduction techniques
and their advantages and disadvantages are summarized in Thble 9.1. Many
reliable methods for sample introduction to these atomizers and excitation
sources are available for solutions, with the pneumatic nebulizer the most
favored. Solids can be analyzed by reducing them to a solution. Many
solids such as ores, metals, and silicates are difficult to dissolve or digest
and can require extensive and complex dissolution/digestion procedures.
This preparative chemistry is time-consuming and tedious and can intro-
duce contamination from the digestion chemicals. These chemicals may
cause interferences. Dilution of the sample by the digestion method may
take it beyond the analytical capabilities of the atomic spectroscopic tech-
nique.
The introduction of solids into flames has received little attention be-
cause of poor sample delivery and relatively low source temperature, e.g.,
air-acetylene is 2400K, can lead to incomplete atomization. However, Mo-
hamed and Fry (1981) showed the potential of a slurry atomization tech-
nique in which animal tissue was homogenized for 2 min and directly in-
troduced to a flame with a clog-free Babington nebulizer. Many workers
including Headridge (1980), Ebdon and Pearce (1982), and Headridge and
Riddington (1984) proposed the direct introduction of solids to the elec-
trothermal atomizer (EA) for AAS measurements. Although successful in
many cases, this method generally suffers from incomplete atomization and
large background absorbance values due to scatter from particulate matter.
Limited sample size of a few mg may be unrepresentative of the bulk prop-
erties of the sample. EA-AAS is generally limited to a single metal deter-
mination and has poor precision due to the variability in adding the sample.
Hoenig and Van Hoeyweghen (1986) describe an alternative to solid sam-
pling in the EA using liquid-suspended powdered samples and platform at-
omization. Direct introduction of solids and powders to the ICP by Salin
and Horlick (1979), Ng et al. (1984), and Lorbar and Goldbart (1985), and
to the DCP by Derie (1984) has been achieved where incomplete atomiza-
tion is less of a problem due to the higher temperatures of these plasmas
Atomic and Mass Spectroscopy 349
Table 9.1 Conventional Sample Introduction Techniques Used in Atomic Spectroscopy:

Advantages and Disadvantages
Sample introduction
technique Advantages Disadvantages
1. Pneumatic Robustness, good corrosion Minimum volume of 1 mL

nebulization resistance, good stability, per min required, 2 to 10%
widespread availability, ease of nebulization efficiency, low
operation, no memory effect. tolerance for high salt solutions,
wide drop-size range of aerosols
made.
2. Hydride generation Superior detection limits Limited to metals that form
over conventional methods gaseous hydrides (As, Bi, Ge,
(sub-ppb), transport efficiency Pb, Sb, Se, Sn, Te), several
close to 100%, freedom from milliliters required, different
interferences, automation. species of metals may require
chemical pretreatment.
3. Electrothermal Milligram masses and microliter Incomplete vaporization and
vaporization volumes, improved sensitivity memory effects, physical effects
for many metals, preconcen- due to design criteria of the
tration, speciation studies for interface.
the more volatile metals, in situ
pretreatment.
4. Direct introduction No sample preparation re- Incomplete atomization for
quired, microgram masses of many compounds, matrix
solids and powders. effects, large nonspecific
background signals, poor
precision and accuracy in
general.
5. Ultrasonic 30% efficiency, smaller drop Generally unreliable, although
nebulization size with a smaller range of successful in many cases.
sizes, ability to use high-salt
solutions, microliter volumes.
6. Liquid and Gas Speciation studies, microvol- In LC, dilution can lead to
chromatography umes, very good for GC though poor detection limits, band
and flow injection limited to volatile compounds. dispersion, and distortion,
analysis improved interface required.
7. Spark and arc Efficient vaporization, mi- Sample needs to be conducting,
crogram masses, comparable cost.
performance to other systems.
350 Sneddon et al.
(> 4000K). Wagatsuma and Hirakawa (1984) and Ko (1984) have investi-
gated the use of a glow discharge lamp as the emission source for the direct
analysis of solids. The solid was incorporated as a part of the lamp, evac-
uated, and a discharge passed, giving an emission signal characteristic of
the sample. A very stable discharge and low detection limits were obtained.
However, analysis time was long and the sample had to be electrically con-
ducting.
For many analyses, separation of the vaporization and atomization-exci-
tation steps is desirable. This is especially true for solid samples that suffer
from a high continuum background, weak and erratic analyte signal due to
incomplete atomization, and scattering from particulate matter. Separa-
tion of the two processes allows independent optimization of their exper-
imental parameters. Many hybrid techniques have been investigated, in-
cluding electrothermal vaporization-ICP by Sneddon and Bet-Pera (1986),
spark-ICP by Aziz et al. (1984), spark-MIP by Helmer and Walters (1984),
electrothermal vaporization-metastable nitrogen plasma by McCaffrey and
Michel (1983), electrothermal vaporization-flame by Kantor et al. (1979),
electrothermal vaporization-MIP by Van Dalen et al. (1982), electrothermal
vaporization-DCP by Mitchell et al. (1986a, 1986b, 1987), and ICP-ICP by
Allen and Coleman (1986).
A potentially useful technique for sample introduction of solid samples
for atomic and mass spectroscopy is laser ablation. The remainder of this
chapter will deal with recent developments in hybrid atomic and mass spec-
troscopy techniques that use a laser as the ablation device combined with
transportation of the ablated material to a separate analysis stage.
9.2 LASER ABLATION OF SOLID SAMPLES
When laser light is focused and absorbed by the surface of a solid sample,
a variety of phenomena can occur, including surface heating, ablation, va-
porization, dissociation, phase changes, and excitation. The process is ex-
tremely complex and at present is not fully understood. A simplified model
has been proposed by Klocke (1969) in which a portion of the light is trans-
formed into heat energy when light flux is incident on the surface of the solid
sample. The energy level of the electrons on the surface is raised and trans-
ferred to the surrounding surface by electron impact, and the electrons and
subsequent material erupt at high velocity, leaving a crater on the solid sur-
face. The material may be in the form of a gas, liquid, solid, or more likely
a mixture of these phases. Laqua (1979) suggests that the velocity of the
vaporized material is dependent on the irradiance: less than lOS W cm- 2
has a velocity of 1o4 em sec- 1 and greater than 109 W cm- 2 a velocity of
1<Yi em sec- 1 •
Table 9.2 General Results of Different Studies of Laser Ablation
1. Most frequently used lasers are solid-state lasers and have included the ruby
(694 nm), Nd:YAG (1064 nm), Nd:glass (1060 nm), C0 2 (10.6 11m), nitrogen
(337 nm), and dye (220 to 740 nm) laser.
2. Bingham and Salter (1976) compared C0 2 , ruby, and Nd: YAG as sources for ion
production in mass spectrometry using a steel standard. Their results suggested
that the shorter the wavelength, the more material ablated, and subsequently,
the higher the sensitivity.
3. Cremers et al. (1985a, 1985b, and 1989) compared Nd:YAG lasers (acoustoop-
tically Q-switched and electrooptic Q-switched), normal-mode Nd:YAG laser,
and excimer laser for the determination of 12 metals in steel. Their results sug-
gested that the excimer laser produced the most representative particles.
4. Tonan (1974) explained the dependence of laser wavelength on the surface co-
efficient as
411"
K=-J.t
,\
(1)
where K was the surface absorption coefficient, ,\ =wavelength, and J.l = solid-
state absorption index.
5. Fabbro et al. (1982) used a Nd:glass laser that was frequency-doubled and
quadrupled to give wavelengths of 1.06, 0.53, and 0.26 J.tm, respectively, to study
the affect of wavelength and IJ9Stulated the following equation that shows the
mass ablation rate, m (kg/s cm 2 ), dependence on wavelength in 11m and ¢a, the
absorbed flux in W/cm2
.A. 1/3
m = 110-'~'-a_ ,\- 4/3 (2)
1014
They concluded that the mass ablation rate would increase strongly at shorter
wavelengths and subsequently give higher sensitivity.
6. Mitchell et al. (1987) used a low-energy("' 100 mJ per pulse) but relatively high
pulse rate (20 Hz) to ablate powder and steel samples and found precisions of
2 to 10%. They varied the pulse rate at 1, 10, and 20 Hz and concluded that
the higher the repetition rate, the higher the emission intensity, and greater the
sensitivity. However, a near steady-state signal over 6 min could be obtained at
the lower repetition rate.
7. For a given period oftime (duration), the pulse rate (repetition rate) ofthe laser
could allow the calculation of the crater size and depth, and consequently, the
amount of material ablated. In general, the higher the pulse rate, the better the
precision of the analysis (Mitchell et al., 1987). Higher repetition rates increase
the amount of material ablated and minimize sample inhomogeneities and laser-
energy variations between pulses.
352 Sneddon et al.
Table 9.2 (Continued)

8. The irradiance absorbed that causes ablation will depend on the output of the
laser, focusing conditions, and sample surface. The following is proposed:
Irradiance Pulse width Results
104 W cm- 2 Millisecond Heating of surface without ablation.

104 -107 W cm- 2 Several hundred Initially vaporize surface, but
microseconds closely followed by expulsion of
some particles and molten metal.
107 -109 w cm- 2 25 nsec Greater vaporization and fewer solid
and molten particles.
> 109 W em - 2 > 100 nsec Vaporization is major process.
9. Scott and Strasheim (1970) used high-speed photography of ablation process to

study Q-switched pulses and normal-mode pulse and concluded that the latter
mode had a greater degree of vaporization.
10. Carr and Horlick (1982) quantified this observation by finding a Q-switched
pulse generated 25 p,g of material, whereas a normal mode pulse produced
500 p,g. In both experiments, the surface and energy (1.5 J) was the same.
11. In general, the highest pulse irradiance should be used to minimize selective
volatization, and the ablation process is less dependent on surface type. The
Q-switched laser is more reproducible in terms of energy and pulse time com-
pared to the normal mode or semi-Q-switched mode.
12. The sample type surfaces, and relative hardness of sample will effect the abla-
tion efficiency and process. As stated in 8, high irradiances (> 109 W cm- 2 )
can minimize this effect.
13. Sneddon and Mitchell (1986) showed the detection limit was about three times
poorer in a hard steel sample than in a soft powdered sample.
14. Cremers et al. (1985a, 1985b, and 1989) showed that different lasers produced
different composition of ablated material. In this case, the surface type ab-
lated may not be representative of the bulk properties of the sample. A crater,
typically 10 to 300 p,m in diameter and depths of a few microns to over 1 mm,
formed on the surface (depending on experimental conditions).
15. Mitchell et al. (1987) noted that results from the initial ablation were spuri-
ous and irregular and suggested this was due to the surface and recommended
ablation for 2 min before accepting quantitative data.
Several parameters of the laser will affect the ablation process and ulti-
mately the analytical signal obtained in a particular system. These include
wavelength, pulse rate and duration, and irradiance. The sample type and
surface will similarly affect the ablation process. In order to obtain opti-
mum ablation, these factors are most probably interrelated and require an
optimization procedure that will simultaneously consider all factors.
In Table 9.2, general results of different studies on laser ablation are pre-
sented and an earlier review by Dittrich and Wennrich (1984) is available.
9.3 LASER ABLATION FOR

SAMPLE INTRODUCTION IN
ATOMIC SPECTROSCOPY
This section deals with atomic spectroscopy techniques selected and rep-
resentative results and performance when laser ablation of solid samples
is used for sample introduction. A major concern and focus has been the
interface between the laser system and the atom excitation source. In ad-
dition to presenting some analytical results, the section will concentrate on
the components and performance of the systems for laser ablation.
9.3.1 Flame Atomic Absorption Spectroscopy
Kantor et al. (1976) constructed a laser ablation-flame atomic absorption

spectroscopic system that is shown schematically in Fig. 9.1. A Nd:glass
laser with an energy output of 1 J was used. The apparatus operated without
a Q-switch as a "normal" laser and generated pulses of 500 psec duration,
each consisting of approximately 250 laser spikes. A microscope objective
lens was mounted in a glass chamber that was open at the base to allow
the laser beam to exit and for the entry of the sample vapor. It was con-
nected via a plastic tube to the mixing chamber and the flame system for
AAS measurement. With this system, iron in the concentration range 8
to 92% in nickel-base alloys was determined. Absolute detection limits in
the ng range for iron, sodium, cadmium, copper, and zinc, and precision
around 10% were obtained. Kantor et al. (1979) used a passive Q-switched
ruby laser to generate a vapor from a solid sample that was transported into
a mixing chamber for introduction to a flame AAS system. Copper sheets
were electroplated with varying thickness of silver, gold, or nickel to be used
as standards for depth analysis. The amount of material ablated and sub-
sequent flame AAS signals were related to the thickness. The system gave
linear calibration curves for silver from 1 to 10 Jlm, for gold for 0.5 to 8 Jlm,
and for nickel from 1 to 12 flm. In these ranges, precisions were 9.4% (sil-
ver), 8.4% (gold), and 14% (nickel), with detection limits of 0.025, 0.032,
354 Sneddon et al.
Laser
Microscope
Glass Chamber
60 em ( Plastic Tube )
I• •I Ejector
-----=-2--c.._~....--"""""'"\ Mixing
~~ Chamber
Air f.~..l....________.
Acetylene _]
Figure 9.1 Schematic diagram of laser flame combination (Kantor et al., 1976);
reproduced by permission.
and 0.073 Jlm, respectively. The conditions were chosen such that complete
evaporation of a layer by a single shot was obtained. The relatively poor
precision was attributed to the decrease of the energy field of the flash lamp
over extended periods of use that led to a decrease in the mass of sample
ablated. Inhomogeneities in the metal layer further contributed to the poor
precision in the analysis.
9.3.2 Electrothermal Atomization

Atomic Absorption Spectroscopy
Lasers have been used as sample introduction devices to ablate solid sam-
ples for subsequent AAS measurements with electrothermal atomization
(graphite furnace). In this system, the material from the ablated sample
was transported by a flow of inert gas to a hot graphite tube where the ma-
terial was atomized for AAS measurement.
Matousek and Orr (1976) describe the use of a pulsed C02 laser (100 mJ
per pulse, half-width 300 nsec, and peak power of 0.5 MW) interfaced
to a graphite furnace (Fig. 9.2) and conventional AAS system. The fur-
nace temperature was accurately monitored. The C02 laser beam was di-
rected downward into the furnace. It was focused by a 0.2 m radius concave
spherical mirror and a cadmium telluride 38 mm focal-length meniscus lens
r-
3mmdia.
-1
10mm
dia.
J 4mm
dia.
1 ..
12 mm dia. -I
Figure 9.2 Design of graphite furnace for laser ablation-graphite furnace atomic
absorption spectrometry (Matousek and Orr, 1976); reproduced by permission.
through the 3 mm diameter opening in the furnace. The sample was placed
in the hole opposite the opening and vaporized into the already hot furnace
for atomization and subsequent AAS measurement. Silver in copper alloys
was determined with a sensitivity (1% absorption) of 0.0009% and the signal
was of approximately 10 msec duration.
Ishizuka et al. (1977) used a Q-switch ruby laser with a maximum output
of 1.5 J with a 20 nsec pulse width and a laboratory atomization-absorption
cell made of stainless steel walls and quartz windows connected to an AAS
instrument. The system was optimized for flow of argon gas through the cell
and the effect of capacitor voltage on the peak and integrated absorbance.
Aluminum, chromium, copper, iron, manganese, molybdenum, nickel, and
vanadium were determined in brasses, carbon steels, alloys, and aluminum
alloys, with precision in the range 2.1 to 10% for peak height and 1.0 to
12.0% for peak area absorbances. Detection limits ranged from 1.9 ppm
for iron in brass to 32 ppm for copper in aluminum alloy.
Dittrich and Wennrich (1980) investigated laser parameters using a ruby
laser with a passive Q-switch. Silver in solid samples was determined with
the ablated sample carried to the atomizer by a flow of argon. They found
that the analytical signal from silver in geological samples decreased with
subsequent laser shots on the same target. The precision was generally
poor and was attributed to heterogenous evaporation of the analyte from
the sample. Wennrich and Dittrich (1982) constructed a dual-channel AAS
system that simultaneously determined two metals in synthetic powder and
356 Sneddon et al.
standard metal alloys. Variations in the two metals were attributed to het-
erogeneities in the sample with precision in the 15 to 50% range. Large dif-
ferences in the sample mass ablated from different materials were obtained
and detection limits varied with the composition of the matrix.
Sumino et al. (1980) describe the use of a Nd:glass laser having a maxi-
mum pulse energy of 2 J, focused by a microscope to ablate a solid sample
for sample introduction to an electrothermal atomizer. In an application for
histochemistry, localization of cadmium in human kidney cortex was quan-
titatively found. Schron and Bomback (1983) constructed a system for the
rapid analysis of geological samples and showed that silver could be deter-
mined in several different matrices using the same standard curve. Dittrich
and Wennrich (1984) found that at low atomizer temperatures, only ab-
sorption signals from metals with low boiling points could be detected and
that high atomizer temperatures were required for high boiling point metals
(e.g., nickel at a boiling point of 3000K). The residence time that was con-
trolled by the gas flow rate affected the analytical signal. The formation of
diatomic molecules was shown by simultaneously evaporating both metallic
and halogen species from the same sample using two lasers. A decrease in
the absorption signal was observed in the presence of halogens.
9.3.3 Inductively Coupled Plasma

Emission Spectroscopy
Thompson et al. (1981) constructed a laser-ICP system that used a ruby

laser delivering 1 J of energy per pulse to ablate material from solid sam-
ples. The target was accurately aligned using a high-powered binocular mi-
croscope that directed the laser beam into an ablation chamber where a flow
of argon carried the ablated sample to the plasma. The measuring cycle of
the spectrometer was activated by a switch that was sensitive to scattered
light from the laser pulse. Calibration curves for copper, chromium, man-
ganese, molybdenum, vanadium, phosphorous, and sulfur in steel samples
were constructed and detection limits from 15 to 80 pg/g for these metals
were obtained. Precision in most cases was very good (1.8% for iron), al-
though for silicon it was very poor (24.6% ). This was attributed to a sam-
pling error as silicon is known to segregate into a minor phase. The cali-
bration of minor metals was improved by ratioing the analyte emission in-
tensity to the iron signal. Results of analyses of silicate materials were un-
satisfactory due to problems in preparing homogeneous samples. In later
work, Hale and Thompson (1983), Thompson and Hale (1984) and Hale
et al. (1984) used this system to analyze soil samples, stream sediment peb-
ble coatings, and heavy mineral grains in stream sediments. The ICP was
equipped with a 1 m vacuum-grating polychromator that allowed the simul-
taneous determination of up to 36 metals. Precision was similarly improved

by ratioing the analyte emission to a major constituent (iron or manganese).
Calibration of the constituents of heavy mineral grains was obtained byes-
tablishing sensitivity ratios of metals between laser ablation and conven-
tional aqueous sample nebulization.
Carr and Horlick (1982) used a ruby laser to ablate solid metal samples
placed in a sample chamber positioned directly below the plasma torch. Si-
multaneous multichannel analysis and time resolution were obtained using
a photodiode array detector and a spectrograph. Largest signals were ob-
tained using the laser in the normal or free-running mode compared to the
Q-switched mode and were attributed to the much greater mass ablated
(500 11g compared to 25 flg). Calibration curves were determined based
on ratioed intensities of manganese 294.9 nm/silicon 288.2 nm, manganese
294.9 nm/magnesium 285.2 nm, and silicon 288.2 nm/magnesium 285.2 nm
for both free-running and Q-switched modes. Slopes ranged from 0.53 to
0.87. These deviations were attributed to the laser sampling step not being
efficient. The precision for the free-running mode ranged from 1.7 to 9.4%
with the poor precision attributed to sample inhomogeneity. The analytical
signal was quite dependent on the laser focusing.
The effects of laser focusing, Q-switching, and distance between the abla-
tion cell and ICP were investigated by Ishizuko and Uwamino (1983). They
found a linear increase in emission intensity with increasing laser energy in
the 1.1 to 2.3 J per pulse range when the laser was used in the Q-switched
mode. The effect of focusing was less critical, as a constant mass was ab-
lated from metal samples when the Q-switched laser spot size was varied
between 0.5 and 1.3 mm. However, in the normal mode, the amount of
material ablated ranged from 500 pg for a 0.5 mm size to 1 flg for a 3 mm
spot size. They also found that when a Q-switched laser is used, the dis-
tance between the ablation cell and the ICP should be as short as possible,
although they concluded that this factor was not as critical using the normal
mode laser. Detection limits were generally about a factor of 10 better for
a normal mode laser than for a Q-switched mode laser, e.g., aluminum was
20 ppm when Q-switched and 2 ppm in the normal mode. No significant
differences in precisions, which ranged from 3 to 11%, were noted between
Q-switched and non-Q-switched operation.
Kawaguchi et al. (1982) constructed a laser-ICP microprobe system by
using a low-energy Nd:YAG laser (maximum 100 mJ per pulse) and a com-
mercial laser optical system. The amount of metal ablated by a laser pulse,
measured directly by weighing samples before and after the laser irradi-
ance, ranged from < 0.01 flg for gold and silver to 10 flg for lead. The
length and diameter of the particulate carrying tube between the sample
chamber and the ICP torch affected not only the emission signal, but the
reproducibility of the emission intensity. Linear working curves were ob-
358 Sneddon et al.
tained for chromium, manganese, and nickel in low-alloy steels. The de-
tection limits were 0.017, 0.074, and 0.025% for the single pulse mode and
0.004, 0.047, and 0.004% for the multipulse mode, respectively.
Cremers et al. (1985a, 1985b, and 1989) describe the analysis of steel
for 12 different metals using an acoustooptically Q-switched Nd:YAG (AO
laser), electrooptical Q-switched Nd:YAG, normal and excimer lasers to ab-
late samples to the ICP. Precisions and accuracy ranged typically from a few
percent (nickel and chromium) up to 60% (sulfur and phosphorous). They
concluded that the composition of material ablated by the different lasers
was very different from the bulk sample composition for several different
metals. The most representative particles were produced by the excimer
laser.
9.3.4 Direct Current Argon Plasma

A direct current argon plasma has been used as the excitation source for
laser-ablated material by Mitchell et al. (1985, 1986a, 1986b, and 1987),
Sneddon et al. (1986), and Sneddon and Mitchell (1986). The system uti-
lized a relatively low energy ("' 100 mJ per pulse) N d:YAG laser operating at
up to 20Hz. A sample chamber was constructed in which the sample rotated
at 1 rpm and allowed integration over a relatively long period time of 20 sec.
The combination of high pulse rate and a large area of surface sampled
decreased sampling error due to sample microheterogeneities and shot-to-
shot variations in the laser energy. Both National Bureau of Standard-
Standard Reference Material (NBS-SRM) steel samples and solid pow-
dered copper ore samples were analyzed successfully, although it was noted
that with the relatively low energy output, it was difficult to perform trace
analysis. The method of standardization and sample preparation was par-
ticularly important for powdered samples. Matrix effects from particle size,
type of metal compound, and particle composition all affected the analyti-
cal signal. However, with careful attention to these details, it was possible
to determine copper in ores with precision and accuracy of ±5%. Detec-
tion limits for cobalt, chromium, copper, manganese, molybdenum, nickel,
titanium, vanadium, and zirconium were in the ng!g to low pg/g range.
9.3.5 Microwave-Induced Plasma

The microwave-induced plasma is particularly susceptible to being extin-

guished due to solvent overloading when liquid samples are introduced
using commercially available pneumatic nebulization. A laser produces a
mass of dry material and is therefore suited for sample introduction to an

MIP.
Ishizuka and Uwamino (1980) combined a Q-switched ruby laser with an
MIP for the determination of metals in brass, steel, and aluminum alloy.
The laser was directed to the sample that was ablated and carried to the
MIP for analysis in a few milliseconds. The emission signals for aluminum,
chromium, copper, iron, manganese, molybdenum, nickel, lead, titanium,
and zinc were traced with an oscilloscope and were determined using a high-
speed peak holder and integrator. Working curves were constructed for
these metals and were linear over two orders of magnitude, although slopes
for some metals deviated significantly from unity. The detection of limits in
the solid samples ranged from 0.9 ppm (0.7 pg) for zinc in aluminum alloy
to 22 ppm (20 pg) for molybdenum in steel. Precision ranged from 1.2 to
13.8% for peak height and 2.3 to 12.1% for peak area measurements.
Leis and Laqua (1978, 1979) ablated aluminum and zinc alloys using a Q-
switched ruby laser for transport to an MIP for detection. The influence of
the argon pressure on spectral lines and the line-to-background ratio was in-
vestigated. Detection limits for metals found in zinc samples were 250 ppm
(aluminum), 20 ppm (cadmium), 13 ppm (magnesium), 150 ppm (lead), and
150 ppm (copper). Calibration curves were linear over about one order of
magnitude for the metals investigated.
9.3.6 Inductively Coupled

Plasma-Mass Spectroscopy
The use of an ICP as an ion source in mass spectroscopy for metal deter-
mination has shown high promise, and recent experimental work and the
potential of inductively coupled plasma-mass spectroscopy (ICP-MS) has
been reviewed by Houk (1986a and b).
Gray (1985) used a ruby laser to ablate solid samples into an ICP-MS.
When major constituents were quantified, problems with saturation of the
detector system in both the mean-current and pulse-counting detection
modes due to the transient nature of the sample pulse were experienced.
Except when a major peak was saturated, reasonably uniform sensitivity for
most metals across the mass range 7 to 238 m/z was obtained. The detection
limits were 10 ng/g or lower.
Arrowsmith (1986, 1987) has used focused Nd:YAG laser pulses at a
wavelength of 1.064 Jlm and with a 160 mJ pulse energy to ablate NBS-SRM
steel samples. The laser could be used either in the single-pulse mode to
give a transient signal or at 10Hz repetition rate to give a continuous signal
that was constant over a 30 sec time period. The solid sample was contained
in a gas-tight housing mounted on a translation stage as shown in Fig. 9.3.
360 Sneddon et al.
TO SAMPLE INLET OF ICP TORCH VIA 0.5 m

FLEXIBLE TUBE
t Pyrex Cell
Sample
Figure 9.3 Schematic diagram of laser ablation-ICP-MS system (Arrowsmith,

1987); reproduced by permission.
The chamber consisted of a demountable pyrex glass housing clamped to

an aluminum base with an 0-ring seal. The laser beam was transmitted
through a laser damage-resistant optical window. Aerosol transport gas at
atmospheric pressure was introduced into the base tangentially to produce
a circular flow pattern around the sample. The ablated material was trans-
ferred to the ICP via a 0.5 m long plastic tube. The laser beam was incident
on the sample at 45° to reduce the absorption of the beam by the plume
of the ablated material. With this system, analytical curves using standard
steel samples were linear over four orders of magnitude. Estimated detec-
tion limits for arsenic, antimony, bismuth, cadmium, manganese, lead, sil-
ver, and tellurium were in the range 0.04 to 6 11-g/g. Precision and accuracy
were± 5%.
9.3.7 Related Spectroscopic Metal

Determination Techniques
Laser ablation of solid samples has been used for sample introduction to
a number of related spectroscopic metal techniques. Kwong and Mea-
sures (1979) and Measures and Kwong (1979) developed a system called
TABLASER (trace analyzer based on laser ablation and selectively ex-
cited radiation). The principle of the technique involves the laser creat-
ing from the solid sample surface a dense high-temperature plasma that
rapidly expanded into a low-pressure region (l0- 3 to 10-4 Torr). The out-
ward streaming material was interrogated by a 6 nsec pulse of laser radiation
from a dye laser turned to saturate one of the strong resonance transitions
of the metal to be determined. The resulting resonance fluorescence was
detected and could be related to the concentration of the metal in the solid
sample. The system used in this study is shown in Fig. 9.4. It consisted of a
Q-switched ruby laser (10 to 20 mJ per pulse, 30 nsec pulse duration) used
for ablation, a nitrogen laser, pumped dye laser (coumarin 120) used to ex-
cite species selectively, and the ablation chamber to hold the sample. An
optical arrangement coupled with the laser-induced resonance fluorescence
into the photodetection system. With this system, chromium in standard
reference steel, doped skim-milk powder, and doped flour were determined
at ppm levels with no interference from these complex matrices. A linear
calibration curve up to 1% and detection limit of 1 ppm were obtained.
Mayo et a!. (1982) describe a laser ablation and resonance ioniza-
tion spectroscopy (LARIS) system for trace analysis of solids, in partic-
ular detecting sodium contamination in solid silicon on the order of 1011
atoms cm- 3 • In this technique, the vaporized sodium atoms were preferen-
tially ionized, using photons supplied by two counter-propagating dye laser
pulses passing through the ablation region. The electrons generated by pho-
toionization were detected as an indication of sodium atoms because other
atoms were not efficiently ionized by the dye laser wavelengths used.
9.3.8 Collection of Laser-Ablated Material

for Subsequent Determination
In certain cases, interferences can be eliminated or minimized by collecting

the ablated material that is then analyzed in a subsequent step. In this sit-
uation, distinct temporal separation of the vaporization and atomization-
excitation steps can be achieved. The ablation of metal samples inside a
graphite tube produced a condensate on the tube. The tube was heated and
the condensate atomized for AAS measurement. When microprobe analy-
sis using laser sampling and spark excitation is performed, the sample can
be contaminated by erosion products from the spark electrode surface, or
the surface may be struck by the electrical discharge. Several groups have
attempted to improve this situation by collecting the products from the laser
ablation step on electrodes that are subsequently sparked at a later time.
Petukh et a!. (1976) investigated the effects of the parameters such as
laser beam focusing, distance from the electrodes to the surface, and the an-
gle of the electrodes to the surface. They were able to quantitate in metal
362 Sneddon et al.
Tuneable
.---...,....- Dye Laser
~-~---------~
I
I I
I
I ~Split-Field Prism
I
OMoood>"m"'"'
I Laser
I
I
Variable I
Delay I
• Fabry-Perot
I I 1 Etalon
I I
I I
II ~Camera
Sample Target
90° Prism to Deflect Dye

Laser Beam Down Czerny Turner
Through Plume Monochrometer
50/50 Beam Splitter
Ablation
+'
II
II
/I
Chamber
,'I
I I Gate Pulse
:Ln
I I Generator
Fast Ottill=,,.,·
Roby '""' ~
Figure 9.4 Schematic diagram of TABLASER facility (Measures and Kwong,
1979); reproduced by permission.
alloys in the 0.01 to 0.1% range. A comparison of the temporally sepa-

rated two-step method and the simultaneous two-step method was made
by Boitsov and Zil'bershtein (1981). They analyzed impurities in synthetic
oxide monocrystals and found that the condensate on the electrode sur-
face had the same composition as the monocrystal analyzed. Impurities in
the 0.005 to 0.01% range were quantitated with better precision than if the
spark was initiated immediately after the laser shot. The collection effi-
ciency of the laser vaporized products by Rudnevsky et al. (1984) who used
a thin graphite disk with an axial aperture to allow a laser beam to pass on
the sample.
9.4 RELATIVE MERITS OF LASER ABLATION

FOR SAMPLE INTRODUCTION
IN ATOMIC SPECTROSCOPY
Sample introduction to an atomic spectroscopy system for metal determina-

tion will depend on many variables: the physical state of the sample (solid,
Table 9.3 Advantages and Disadvantages of Laser Ablation for Sample Introduc-
tion in Atomic Spectroscopy
Advantages Disadvantages
1. Directly applied to solid samples 1. Standardization is difficult, partie-

with little or no preparation. ularly if the sample to be analyzed
2. Can use conducting and noncon- has an unknown and complex matrix
ducting samples. effects.
3. Can investigate surface hetero- 2. Laser is expensive and leads to a
geneities in solid samples. more complex instrument.
4. Mass of microgram or less required. 3. Detection limits are generally not
5. High irradiance of laser will as good as established solution
produce an efficient vaporization techniques (one to three orders of
stage. magnitude or higher).
6. Separation of vaporization and 4. Precision typically 2 to 10% if
atomization-excitation stage could heterogeneities do not exisHn the
allow these separate experimental sample (compared to 1.0% for
stages to be optimized for improved established solution techniques).
signal and accuracy. 5. Accuracy typically ±5% compared
7. Hard, polymeric, and high- to "' 1.0% for established solution
molecular-weight compounds techniques.
are vaporized. 6. Small portion of sampling may
not be representative of bulk
properties.
364 Sneddon et al.
Table9.4 Detection Limits8 for Various Metals Using Atomic

Spectroscopic Methods with Sample Introduction by Laser
Ablation
Analysis method [Jlg/g (ppm)]
Metal ICPb ICPC DCP MIP ET FLAME ICP-MS
AI 20 2 9.3 11
Ag 0.3
As 2
Bi 0.2
Cd 8 0.6 48
Cr 10 1 31 20
Cu 9 0.3 21 2.4 7.2 10
Mg 1
Mn 3 0.3 18 37 20
Mo 20 2 30 10
Ni 20 1 32 12 24
Pb
Te 0.3
v 20 1
Zn 22 550
Zr 50 0.3
8 These detection limits were taken from the quoted literature. Due to
differences in instrumentation/operating conditions, they should not be
directly compared.
bQ-switched.
0 Normal.
solution, high solid solution, gas, or mixture), complexity of the matrix, level
of metal to be determined, amount of sample availability, and required ac-
curacy and precision. If the sample is in solution or is easily dissolved, then
pneumatic nebulization for sample introduction in atomic spectroscopy will
be the most widely used technique. However, in certain circumstances,
the use of a laser to ablate solid samples for subsequent determination by
atomic spectroscopy may be advantageous. The advantages and disadvan-
tages of laser ablation for sample introduction in atomic spectroscopy are
summarized in Table 9.3.
Detection limits for various metals using atomic spectroscopy with sam-
ple introduction by laser ablation are shown in Table 9.4. In general, detec-
tion limits are in the ppm (pg/g) level that is typically one to three orders
of magnitude higher than that for solution detection limits. Direct compar-
isons between different systems are not valid due to different instrumen-
tation (laser, excitation source), experimental factors (wavelength, geome-

try), or samples (steel, brasses). However, they are useful figures of merit.
Precision was typically 10% or better, if we assume a homogenous sample,
which is about an order of magnitude higher than that obtained using solu-
tions. Accuracy was obtained using National Bureau of Standards-Standard
Reference Materials and found to be typically ±5% or better, which is about
three to five times higher than that obtained with comparable solutions.
Linearity of calibration curves was comparable to solutions. Interferences
were more complex than solutions, but could be corrected using various
chemical and instrumental approaches.
9.5 LASER SOURCES FOR

MASS SPECTROMETRY
Laser volatilization is rapidly gaining popularity as a method of sample in-

troduction for mass spectrometry (Houk, 1986a and b). Since its original
description (Honig, 1963), numerous research papers and review articles
(Conzemius and Capellen, 1980) have appeared on various aspects of laser
mass spectrometry. Pulsed lasers are typically used in these experiments,
and sample vaporization and ionization generally take place in one of two
modes. In one case, relatively high-intensity pulses, I;?: 100 MW/cm2 , are
used to simultaneously vaporize, ionize and in some cases fragment the sam-
ple of interest (Cotter, 1987). This type of instrumentation is often referred
to as a laser microprobe or laser ionization mass spectrometer (Denoyer
et al., 1982). It is available from several vendors as a commercial instru-
ment. In the second case, relatively gentle irradiation, Is; 10 MW/cm2 , is
used to vaporize intact or largely intact molecules. In this type of instru-
mentation, ionization and/or fragmentation occurs by a secondary process,
typically electron impact (Land et al., 1987), chemical ionization (Cotter,
1987), or photoionization by a second laser (Nagar et al., 1985), subse-
quent to volatilization. In the remainder of this chapter, "laser evapora-
tion" will be used as a generic term to include all laser-based photochemi-
cal and photophysical processes that result in the generation of gap phase
atoms, molecules, ions, or radicals from condensed-phase samples.
Both methods share a number of characteristics in common: (1) No back-
ground is introduced due to the bulk heating of the sample. (2) Spatial res-
olution can be very good, limited only by diffraction of the incident beam
(typically ~ 1 ~-tm in diameter) and, at high fluxes, cratering of the sur-
face (typically= 10 ~-tm). This capability can be very important for small
samples and samples containing inhomogenously distributed components.
(3) Little sample preparation is needed. Although conventional wet analyt-
ical samples can be dried on a sample introduction device, non conventional
366 Sneddon et al.
samples can also be introduced directly into the source region of the mass
spectrometer. (4) Sensitivity is excellent, and the detection limit frequently
falls in the femtogram to attagram (absolute) or sub-parts-per-billion range.
(5) The possibility exists for absolute measurements without recourse to
standard samples.
Several topics will be discussed in the remainder of this section. Instru-
mentation for both types of apparatus will be described simultaneously,
since they share many common requirements and attributes. Next, experi-
mental results and applications will be presented, first for the microprobe
system, and second for apparatus requiring postionization. Last, the future
prospects for laser mass spectroscopy will be discussed briefly.
9.5.1 Instrumentation
A variety of laser sources have been used in laser evaporation for mass spec-
trometry. In the vast majority of experiments, pulsed lasers have been used.
Q-switched ruby lasers were used in many early experiments (Vanderborgh,
1977). These instruments produce large pulse energies (Ep 2:: 1 J), but at ex-
tremely low repetition rates,[~ 1Hz. This latter constraint is a distinct dis-
advantage for a process such as laser evaporation, which is highly stochastic
and for which extensive signal averaging may be required.
More frequently, TEA C0 2 lasers (McCreary et al., 1982), Q-switched
Nd +J:YAG (Nagar et al., 1985), and rare gas/halogen excimer lasers (Viswa-
nathan and Hussla, 1986) are used for laser evaporation. All can operate at
repetition rates[ 2:: 10Hz, produce pulse energiesEP 2:: 100 mJ, and produce
relatively short pulses, r ~ 100 nsec.
The wavelength of operation often controls the mode of evaporation.
Long-wavelength (A= 10 p,m) C0 2 laser pulses usually induce a "thermal"
evaporation process, in which the evaporated material exhibits well-defined
translational and internal temperatures in equilibrium with the local sur-
face temperature (Burgess et al., 1984). Commercial C02 lasers are rela-
tively inexpensive and reliable. Nd+ 3 :YAG lasers offer a more expensive
but also more versatile tool for laser evaporation studies. Frequency mul-
tiplication of the output from a Nd+ 3 :YAG laser allows a number of wave-
length regions to be accessed. Operation at the fundamental output fre-
quency, 1.06 p,m, usually results in thermal evaporation processes similar
to those observed with C02 laser initiation. Operation at the fourth har-
monic, 266 nm, or a Raman-shifted variant (Estler and Nagar, 1986), on the
other hand, often initiates processes that are nonthermal in nature. The re-
sult may be non thermal velocity or internal energy distributions, and/or un-
usual fragmentation patterns. Intermediate wavelengths, A = 532, 355 nm,
frequently produce ambiguous results that can be characterized as neither
thermal nor photophysical in nature (Estler et al., 1987). Excimer lasers are
normally operated at 308 nm (XeCl), 248 nm (KrF), or 193 nm (ArF) and
may produce a number of novel and unusual results at short wavelengths
(Yeh, 1986). In addition, in instances where a microplasma is produced, the
wavelength of irradiation determines the fractional absorption. The plasma
resonant frequency vP is given by (Houk, 1986a and b)
vp = (47rn ee 2/m e )112

= 8.9 X 103(ne) 112 (9.1)
where ne is the electron number density (cm- 3) in the plasma, e the electron
charge, and me the electron mass. In order for substantial absorption to
occur in the plasma, v1/vp ~ 1, where v 1 is the laser frequency. Otherwise,
the laser radiation is reflected.
Similarly, the amount of material removed from the surface under con-
ditions of micro plasma formation is given by (Dittrich and Wennrich, 1984)
1/3
m = 110 ( ..!lJL..
1014
) A- 413 (9.2)
wherem is the mass ablation rate (kg sec- 1 cm- 2 ) and ¢>a the absorbed flux
(W cm- 2 ).
Irradiation of the sample may occur either from the "backside" (irra-
diated surface opposite the mass spectrometer flight tube), or from the
"frontside" (laser irradiation falling on the side exposed to the flight tube.
In the former case, thin samples and supports are required, whereas in the
latter case, more latitude is allowed in sample characteristics. An alignment
laser and microscope to view the sample surface are included in many in-
struments. It is often advantageous to have a mechanism for rastering the
sample in front of the laser beam (Wilk and Hercules, 1987) to allow either
for signal averaging in homogenous samples or to measure the distribution
of material in inhomogenous samples (Conzemius et al., 1981).
For experiments in which lasers are also used for postionization, the com-
mon instrument of choice is a pulsed dye laser, often with wavelength ex-
tension capabilities (Moore et al., 1984) such as frequency doubling, fre-
quency mixing, or Raman shifting. The tuning capability allows the use of
resonant intermediate states in a multistep ionization process, thus increas-
ing both the probability of ionization for most atoms (Hurst et al., 1979)
and molecules (Johnson and Otis, 1981), and the selectivity of ionization.
The latter capability can be important in analyzing complex mixtures (Miller
et al., 1982b). In some cases, a flashlamp-pumped dye laser is used, but a
wider tuning range and shorter pulses are usually obtained with N d +3 : YAG
368 Sneddon et al.
or XeCl-excimer laser pumping. The tuning range is typically from the near
IR (::::::: 800 nm) to the VUV (::::::: 200 nm ), with 6 to 15 nsec pulses from 1 to
20 mJ, and repetition rates of 10 to 500Hz. With the use of resonant inter-
mediate states, the ionization can usually be saturated within the laser focal
volume (::::::: 100% conversion of neutrals to ions).
In some instances, a high-power fixed-frequency laser, typically a fre-
quency-quadrupled Nd +3 : YAG (Schueler and Odom, 1987) or KrF-excimer
(Becker and Gillen, 1984), may be used for postablative ionization. Non-
resonant ionization is usually somewhat less efficient than for resonant pro-
cesses (typically 1 to 10% ), although accidental coincidences with autoion-
izing transitions may increase the ionization probability by orders of mag-
nitude (Miller et al, 1982a). On the other hand, the nonresonant process
tends to ionize all species present in the desorbed plume and so yields a
broader range of information.
Only a small fraction of the published work in laser mass spectrome-
try has been performed with single-channel detection instruments, such as
quadrupole and magnetic sector spectrometers. Magnetic instruments are
typically used when superior sensitivity or large dynamic range is required.
By far, the majority of experiments have utilized multichannel detection,
such as is possible in time-of-flight (TOF), and more recently, Fourier trans-
I REFLECTOR
1 : SLOW IONS M+
2: FAST IONS M+
Figure 9.5 Schematic of a reflecting time-of-flight instrument. Superior resolution

is obtained by compressing the ion energy spread in the reflecting optics. High en-
ergy ions arrive first, and penetrate farthest, while low energy ions arrive later, but
travel a shorter distance (Boesl, 1982); reproduced by permission.
form spectrometers (FTICR). In fact, current interest in time-of-flight in-

struments can be largely attributed to laser-based applications. Recent de-
velopments have resulted in conventional-design TOP instruments of un-
precedented resolving power (Yang and Reilly, 1987), limited only by the
temporal pulse width of the ionization laser. In addition, the development
of reflecting TOP instruments has resulted in resolving powers ;::: 104 (Boesl
et al., 1982), Fig. 9.5. A channel electron multiplier, venetian-blind-type
multiplier, channel plate, or "Daly" detector is typically used for ion detec-
tion, while a transient digitizer is usually used to record the data.
FTICR mass spectrometers are extremely well-suited for diagnosis of the
composition of transient vapor pulses (McCreary et al., 1982). FTMS of-
fers the advantage of multiplex detection, superior resolving power (often
;::: lOS), excellent sensitivity (sub pg, absolute), and very high ultimate mass
range (Sherman et al., 1985b). The major drawback is the limited dynamic
range, less than 300, a result both of space-charge restraints on the number
of ions that can be stored in the ion trap and on the method of detection,
which precludes the use of high-dynamic-range electron multiplier tubes.
9.6 APPLICATIONS OF THE

LASER MICROPROBE
9.6.1 Inorganic Analysis and Chemistry
Under the appropriate conditions, the measured analyte signal(s) will be

proportional to the relative concentration of analyte(s) in a solid sample.
This requires that the evaporation and ionization processes be nonselec-
tive, chemical speciation be minimized, and discrimination in transmission
and detection in the mass spectrometer be minimized. This capability has
been demonstrated in selected cases for both metallic (Dittrich and Wen-
nrich, 1984) and nonmetallic analyses (Shankai et al., 1984). Particularly
important areas of application lie in the analysis of discontinuous samples
(Conzemius et al., 1981), inhomogenously distributed impurities and single-
aerosol particles (Wieser et al., 1981), where the high spatial resolution of
the laser microprobe can be put to good use (Fig. 9.6).
In general, however, there exist substantial differences in the sensitivity
of the laser microprobe for various elements; see Table 9.5 for representa-
tive detection limits. These differences are usually matrix-dependent and
particularly sensitive to the presence of low ionization-potential elements.
In addition, the species-dependent sensitivity is often a function of the irra-
diation conditions, i.e., intensity, pulse duration, wavelength, etc. It is thus
normally a requirement to determine independently the relative sensitivity
of the laser microprobe to various elements in the appropriate matrix under
a variety of conditions (Eloy, 1984).
370 Sneddon et at.
II)
:!::: C2Hi c2H2o-
c:
::J Cl-
~
c
cz c2w
....
:!:::
.c
c-/,
.... Cl-
c Br-
0 o-ow (\
c: F-
C)
I
Cl
II)
s-
H
10 20 30 /,Q 50 60 70 80 90 100 110 120 130 m/e -
Particle (about 0.5 pml formed by halogenated hydrocarbons and

sulfur compounds
Figure 9.6 Negative ion mass spectrum obtained by a laser microprobe from a sin-
gle atmosphere aerosol particle (1.4 J.Lm diameter). (Wieser, 1981) reproduced by
permission.
Laser evaporation has been used frequently to generate a variety of gas-

phase ions for subsequent investigation of ion-molecule chemistry (Freiser,
1985). A metallic target is used as the substrate to produce gas-phase atomic
ions (Kang et al., 1985), which are subsequently reacted with gas-phase
species, normally in the trap of an ICR mass spectrometer (Forbes et al.,
1987). The unique trapping qualities of the ICR allow as many as eight
sequential reactions to be monitored.
9.6.2 Organic and Biomolecular Analysis
Considerably more attention has been given to the problems of biomolecu-

lar and nonvolatile organic analysis. The problem to be addressed is usually
one of volatilizing the material of interest without extensive fragmentation
(Posthumus et al., 1978). Laser desorption (LD) is but one phase in the al-
phabet soup of techniques designed to impact this area by "soft" ionization
processes. Alternate methods include SIMS (secondary ion mass spectrom-
etry), FAB (fast ion bombardment) and FD (field desorption), and others
(McCreary and Gross, 1985). Laser desorption is characterized by fast heat-
ing, which preferentially leads to evaporation over thermal degradation.
The characteristic of LD that typically sets it apart from other soft ioniza-
tion methods is that the process of desorption is thermal or quasithermal in
nature, rather than impulsive.
Table 9.5 Representative Detection Limits for Laser Mass

Spectrometry
Method of analysis Sample Detection limit8 Reference
Laser microprobe Liin 2 X 10-19 g Wieser et al.,

sucroseb 1984 1984
Laser microprobec Cdin 1 X 10- 17 g Wieser et al.,
sucroseb 1984
Laser microprobe Modified ~ 1 X 10-9 g Unger et al.,
nucleosides 1981
Electron co 5 X 10-6 Sherman et al.,
postionization monolayer 1985a
Laser postionization 81Kr ~1 X 10-18 g Thonnard et al.,
< 104 atoms 1987
Laser postionization PTpd ~1 X 10- 15 g Hang,
(4 x 10- 17 mole) 1987
a For elemental analysis, fractional sensitivity is typically in the sub-ppm range.
bLow atomic weight, high-volatility elements, such as lithium, can often be de-
tected with exceptionally high sensitivity.
c A gain in sensitivity of 5 times has been reported (Verdun, 1987) atomic transi-
tion.
dProtoporphyrin IX dimethyl ester.
In order to improve ionization efficiency, organic and biochemical sam-

ples are often doped with salt solutions, typically NaCl or KBr (Unger et al.,
1981 ). This procedure leads to dramatically increased yields of quasimolec-
ular ions and fragments of the type (M + Na)+ and (M +K)+. In fact, this
mode of ionization is so strongly preferred that these quasimolecular ions
appear even in the absence of doping, presumably due to the presence of
trance impurities. The mechanism of this process appears to be a combina-
tion of surface and gas-phase effects (Hercules et al., 1982). In the former
case, intense localized heating allows the migration and reaction of alkalai
metal atoms with organic molecules and salts, whereas in the latter case,
ion-molecule reactions (Van der Peyl et al., 1983), possibly driven by the
high temperatures, may produce the same net effect.
Important applications include the analysis of high-molecular-weight
polymers (Brown et al., 1986), as shown in Fig. 9.7. In this work, laser
desorption-Fourier transform mass spectrometry was applied to a variety of
polymers with average molecular weights of up to 6000. Molecular weights
calculated from the observed spectra were quite accurate and exhibited no
mass discrimination in the ionization mechanism. Perhaps more important,
no significant fragmentation was observed.
Figure 9. 7 Mass spectrum of the polymer PEG 3350 doped with KBr, obtained by
laser-desorption Fourier transform mass spectrometry (Brown, 1986); reproduced
by permission.
Substantial progress has also been made in the analysis of biomolecu-

lar species including nucleosides, oligosaccharides, glycosides (McCreary
et al., 1985 and McCreary and Gross, 1985b), and peptides (Tabet and Cot-
ter, 1984). In a number of recent works cationized [M-H]+, alkalai metal
ion cationized and anionized [M-H]- species were observed.
9.7 APPLICATIONS OF LASER

DESORPTION AND POSTIONIZATION
9.7.1 Inorganic, Elemental and Materials Analysis
The use of lasers to desorb neutral atoms, molecules, and fragments fol-
lowed by gas-phase ionization processes can increase sensitivity relative to
laser microprobe analyses. This is a direct consequence of simple thermo-
dynamic considerations that dictate that the largest fraction of desorbed
species will be neutral. For laser desorption, the magnitude and temporal
behavior of the thermal transient can be estimated from (Nagar et al., 1985)
T(t)
f
= [ (~~:c7r)112
] t R(tr112-r) dr +To
Jo (9.3)
where e is the optical absorptivity,~~: (W cm- 1K - 1) the thermal conductivity,

c the heat capacity (J cm- 3 K- 1 ), T 0 the base temperature, andR(t) the tem-
poral profile of the laser. Once the temporal history of the thermal transient
is known, the evaporation rate can be estimated from standard techniques
(Dushman, 1962).
Numerous groups have demonstrated the facility with which laser des-
orption, in combination with resonant multiphoton photoionization, can
be used to detect a wide variety of species. Detection of as few as 105 des-
orbed atoms is practical (Nagar et al., 1985), thus allowing for repetitive
sampling. Applications include the detection of sodium in semiconductor
samples (Mayo et al., 1982), high-resolution spectroscopy (Kronert et al.,
1987) and the development of "atom bunchers" (Hurst et al., 1984).
Materials analysis has also benefited from the application of combined
laser desorption/laser ionization techniques. The detection of submono-
layer coverages by laser desorption followed by electron-impact FTMS has
recently been reported (Land et al., 1987). In other work (Schueler and
Odom, 1987), the microanalysis of high-purity GaAs and HgCdTe has been
reported by nonresonant laser ionization. The elemental ion yields were
significantly greater than for a simple microprobe configuration and pro-
vided a useful measure of the stoichiometry of the starting material. Laser
desorption coupled with resonant laser ionization has been used both to
detect inhomogenously distributed impurities in a recent study of optical
damage (Estler and Nagar, 1988) and to investigate the damage process
itself (Estler et al., 1987).
9.7.2 Organic and Biochemical Applications
A wealth of data has been accumulated recently (Luhman, 1987) on the use
oflaser desorption/laser photoionization for the detection of nonvolatile or-
ganic and biochemical molecules. Two principle modes of operation have
been used. In one case, the molecules of interest are desorbed into a vac-
uum, where they are interrogated by a fixed-frequency UV-laser of suffi-
ciently high intensity to ionize a large fraction of the molecules (Engelke
et al., 1987). This method has resulted in subfemtomole detection limits
(Fig. 9.8). In another case, a monolayer of molecules desorbs into a high-
pressure gas, which then undergoes a supersonic expansion in a vacuum
(Tembreull and Luhman, 1987). This effectively cools the sample and col-
374 Sneddon et al.
206
PTH-Aianine
93
135
,....
....
1/)
0
c:
::J
..ci
....
l1l
._,
PTH-Proline
C/)
w
i= 69 135
en
z
w 0
1-
~
z
0 PTH-Valine
0 5 10
(fLS)
Figure 9.8 Laser desorption TOF mass spectrometry of three different phenyl-
thiohydantoin-amino acids. Approximately 5 pg of material was desorbed by a
200 mJ/cm 2 C0 2 laser pulse. Ionization was affected by ,.__ 1 mJ/pulse from a
frequency-quadrupled Nd+ 3 :YAG laser (Engelke, 1987); reproduced by permis-
sion.
lapses the partition function so that a laser can be used to address a large
fraction of the desorbed species.
Mechanistic processes have been studied by laser mass spectroscopy (Es-
tler and Nagar, 1986), particularly those involving photoablation of poly-
mers used in the electronics industry as photoresists (Yeh, 1986). Also of
particular note are a series of experiments (Bowers et al., 1984, 1986) in
which laser photodissociation of trapped ions was used as an aide in deter-
mining the structure of a gas-phase ion. This work may have substantial
implications in gene sequencing.
9.7.3 Photophysical and Kinetic Studies
The rapid heating rates possible with pulsed lasers (2:: 109K sec- 1 ) allow
a whole new class of surface and material properties to be studied (Hall,
1987). In particular, pulsed laser-induced heating has been used to follow
the course of chemical reactions on well-defined surfaces in real time (Hall
and DeSantolo, 1984). Prototypical is a series of experiments in which laser
desorption was applied in addition to a ramped bulk temperature rise to
follow the course of a reaction as a function of time (temperature, Fig. 9.9).
The temperature rise due to the laser pulse is so fast ( ~ 108 K sec- 1 ) that
desorption takes place before any substantial chemistry can occur, thus
allowing "snapshots" of the surface composition as a function of time (tem-
perature). Similar methods have been used to follow the course of surface
diffusions as a function of temperature (Mak et al., 1987). A small area of
the surface is depleted by a rapid, laser-initiated, thermal pulse. At a vari-
able (selected) delay time, the laser is pulsed again in order to measure the
repopulation of the area. Similar instrumentation has been used to mea-
sure the quantum state distribution of molecules thermally desorbed from
well-characterized surfaces (Kubiak and Zare, 1985).
/\ CO/CH 30H (X 0.1)
I \
I \
400 500 600 700

Temperature (K)
Figure 9.9 Thermal decomposition of methanol on nickel, as measured by laser

desorption/mass spectrometry. As the temperature increases, the amount of des-
orbed methanol decreases, while the amount of product carbon monoxide and hy-
drogen increases (Hall, 1987); reproduced by permission.
376 Sneddon et al.
A number of interesting surface photochemical and photophysical pro-

cesses have also been studied by laser mass spectroscopy. Layer-dependent
sputtering of BaF2 has been reported to be enhanced by the presence of
"surface" optical states (Reif et al., 1986). Tunable-laser mass spectroscopy
has been used to determine the vibrational spectra of surface-adsorbed
molecules (Chuang, 1985). In addition, a number of unusual surface-
photochemical effects have been reported (Bourdon et al., 1984), partic-
ularly for photolysis of CH3 Br on lithium fluoride surfaces (Tabares et al.,
1987).
9.8 CONCLUSION
The use of laser vaporization (in the form of ablation) for sample intro-
duction in atomic spectrochemistry will continue to be investigated. Its
use will be complementary to traditional sample introduction techniques
and will afford the analyst greater flexibility in choosing an appropriate
method for a particular problem. Its use will be extended to areas such
as Fourier transform-mass spectroscopy and liquid chromatography-atomic
spectroscopy.
The field of laser applications in mass spectrometry is still growing.
Laser-based microprobe and materials analysis applications, in particular,
are gaining more widespread acceptance. In addition, it seems possible
that laser vaporization mass spectrometry may soon become the method
of choice for the analysis of polymers, glycopeptides, polysaccharides, and
other biochemical species.
REFERENCES
Allen, G. M. and Coleman, D. M. (1984). Segregated samples and excitation with a

dual inductively coupled plasma, Anal. Chern., 56: 2981.
Arrowsmith, P. (1986). Laser ablation of solid samples for introduction in ICP/MS,
Sciex Application Note 13186A.
Arrowsmith, P. (1987). Laser ablation of solids for elemental analysis by inductively
coupled plasma-mass spectrometry, Anal. Chern., 59: 1437.
Aziz, A., Broekaert, J. A. C., Laqua, K., and Leis, F. (1984). A study of direct analysis
of solid samples using spark ablation combined with excitation in an inductively
coupled plasma, Spectrochern. Acta, 39B: 1091.
Becker, C. H. and Gillen, K. T. (1984). Surface analysis by nonresonant multiphoton
ionization of desorbed or sputtered species, Ana/. Chern., 56: 1671.
Bingham, R. A. and Salter, P. L. (1976). Analysis of solid materials by laser probe
mass spectroscopy, Ana/. Chern., 48: 1735.
Boesl, U., Neusser, H. J., Weinkauf, R., and Schlag, E. W. (1982). Multi photon mass
spectrometry of metastables: Direct observation of decay in a high-resolution
time-of-flight instrument, J. Phys. Chern., 86: 4857.
Boitsov, A. A. and Zil'bershtein, K. L. (1981). Optical emission laser microprobe
analysis of synthetic oxide monocrystals, Spectrochim. Acta, 36B: 1201.
Bourdon, E. B. D., Cowin, J. P., Harrison, 1., Polanyi, J. C., Segner, J., Stanners,
C. D., and Young, P. A. (1984). UV photodissociation and photodesorption of
adsorbed molecules. 1. CH3 Br on LiF(001 ), J. Phys. Chern., 88: 6100.
Bowers, W. D., Delbert, S.-S., Hunter, R. L., and Mciver, R. T. (1984). Fragmenta-
tion of oligo peptide ions using ultraviolet laser radiation and Fourier transform
mass spectrometry, J. Amer. Chern. Soc., 106: 7288.
Bowers, W. D., Delbert, S.-S., and Mciver, R. T. (1986). Consecutive laser-induced
photodissociation as a probe of ion structure, Anal. Chern., 58: 969.
Brown, R. S., Weil, D. A., and Wilkins, C. L. (1986). Laser desorption-Fourier trans-
form mass spectrometry for the characterization of polymers, Macromolecules,
19: 1255.
Browner, R. F. and Boorn, A. W. (1984). Sample introduction: The Achilles heel of
atomic spectroscopy, AnaL Chern., 56: 786A and 875A.
Burgess, D., Jr., Hussla, 1., Stair, P. C., Viswanathan, R., and Weitz, E. (1984).
Pulse laser-induced thermal desorption from surfaces: Instrumentation and pro-
cedures, Rev. Sci. Inst., 55: 1771.
Carr, J. W. and Horlick, G. (1982). Laser vaporization of solid samples into an in-
ductively coupled plasma, Spectrochim. Acta, 37B: 1.
Chuang, T. J. (1985). Photodesorption and adsorbate-surface interactions stimu-
lated by laser radiation, J. Vac. Sci. Tech., B3: 1408.
Conzemius, R. J. and Capellen, J. M. (1980). A review of the applications to solids
of the laser ion source in mass spectrometry, Int. J. Mass Spectrom. Ion Phys., 34
197.
Conzemius, R. J., Schmidt, F. A., and Svec, H. J. (1981). Scanning laser mass spec-
trometry for trace level solute concentration profile, Anal. Chern., 53: 1899.
Cotter, R. J. (1987). Laser Mass Spectrometry: An overview of techniques, instru-
ments and applications,Analyt. Chim. Acta, 195: 45.
Cremers, D. A., Archuleta, F. L., and Dilworth, H. C. (1985a). Direct solid metal
analysis by laser ablation into the inductively coupled plasma, Proc. 1985 South-
west Conf. on Optics, March 4-8, Albuquerque, New Mexico.
Cremers, D. A., Archuleta, F. L., and Dilworth, H. C. (1985b). Rapid analysis of
steels using laser-based techniques, Proc. 6th Process. Techn. Conf. Measure-
ment and Control Instrumentation in Iron and Steel Industry, April14-17, De-
troit, Mich.
Cremers, D. A., Dilworth, H. C., and Archuleta, F. L. (1989). Rapid steel analysis by
laser ablation into the inductively coupled plasma, Anal. Chern. (submitted for
publication).
Denoyer, E., Grieken, R. V, Adams, F., and Natusch, D. F. S. (1982). Laser mi-
croprobe mass spectrometry 1. Basic principles and performance characteristics,
AnaL Chern., 54: 26A.
Derie, R. (1984). Analysis of mineral suspensions by D. C. plasma emission spec-
trometry, Anal. Chern., 166: 61.
378 Sneddon et al.
Dittrich, K. and Wennrich, R. (1980). Atoms absorptions spektrometrie durch laser-

vedampfung mit nachfolgender electrothmischer atomizierung, Spectrochim.
Acta, 35B: 731.
Dittrich, K. and Wennrich, R. (1984). Laser vaporization in atomic spectroscopy,
Prog. Analyt. Atom. Spectrosc., 7: 139.
Dushman, S. (1962). Scientific Foundations of Vacuum Technique, Wiley, New York.
Ebdon, L. and Pearce, W. C. (1982). Direct determination of arsenic in coal by
atomic-absorption spectroscopy using solid sampling and electrothermal atomi-
sation,Analyst, 107: 742.
Eloy, J. F. (1984). Quantitative microanalysis of solid samples with the laser probe
mass spectrometry II, J. Physique Colloq., C2: 265.
Engelke, F., Hahn, J. H, Henke, W., and Zare, R.N. (1986). Determination of phen-
ylthiohydantoin-amino acids by two-step laser desorption/multiphoton ioniza-
tion,AnaL Chern., 59: 909.
Estler, R. C. and Nogar, N. S. (1986). Mass spectrometric identification of wave-
length dependent UV-laser photoablation from polymethylmethacrylate. Appl.
Phys. Lett., 49: 1175.
Estler, R. C., Apel, E. C., and Nogar, N. S. (1987). Laser mass spectrometric studies
of optical damage in CaF2 ,.T. Opt. Soc. Amer., B4: 381.
Estler, R. C. and Nogar, N. S. (1988). A chemical precursor to optical damage de-
tected by laser ionization mass spectrometry,Appl. Phys. Lett. (submitted for pub-
lication).
Fabbro, R., Fabre, E., Amiranoff, F., Garbeau-Labaune, C., Virmont, J., Weinfield,
M., and Max, C. E. (1980). Laser-wavelength dependence of mass-ablation rate
and heat-flux inhibition in laser-produced plasmas, Phys. Rev., Ser. A., 26: 2289.
Forbes, R. A., Tews, E. C., Huang, Y., Freiser, B. S., and Perone, S. P. (1987). Eval-
uation of laser desorbed transition-metal ions as analytical chemical ionization
reagents by pattern recognition, AnaL Chern., 59: 1937.
Freiser, B. S. (1985). Applications of laser ionization/Fourier transform mass spec-
trometry to the study of metal ions in the gas phase, AnaL Chim. Acta, 178: 137.
Gray, A. L. (1985). Solid sample introduction by laser ablation for inductively cou-
pled plasma source mass spectrometry, Analyst, 110: 551.
Hale, M. and Thompson, M. (1983). Laser ablation inductively coupled plasma spec-
trometry for geochemical exploration, Trans. Inst. MetalL, 92B: 23.
Hale, M., Thompson, M., and Wheatley, M. (1984). Laser ablation of stream sedi-
ment pebble coatings for simultaneous multielement analysis in geochemical ex-
ploration, J. Geochem. Explor., 21: 361.
Hall, R. B. and DeSantolo, A. M. (1984). Pulsed laser induced excitation of metal
surfaces; application as a probe of surface reaction kinetics of methanol on Ni,
Swf. Sci., 137: 421.
Hall, R. B. (1987). Pulsed-laser-induced desorption studies of the kinetics of surface
reactions, J. Phys. Chern., 91: 1007.
Hang, J. H., Zenobi, R., and Zare, R. N. (1987). Subfemtomole quantitation of
molecular adsorbates by two-step laser mass spectrometry, J. Amer. Chern. Soc.,
109:2842.
Headridge, J. B. (1980). Determination of trace elements in metals by atomic ab-

sorption spectrometry with introduction of solid samples into furnaces: An ap-
praisal, Spectrochim.Acta, 35B: 785.
Headridge, J. B. and Riddington, I. M. (1984). Determination of silver, lead, and
bismuth in glasses by atomic-absorption spectrometry with introduction of solid
samples into furnaces,Anaryst, 109: 113.
Helmer, D. J. C. and Walters,J. P. (1986). Analysis of the effluent from a high voltage
spark discharge with a microwave-induced plasma,AppL Spectro., 38: 392.
Hercules, D. M., Day, R. J., Balasanmugam, K., Dang, T. A., and Li, C. P. (1982).
Laser microprobe mass spectrometry 2. Applications to structural analysis, AnaL
Chem., 54: 280A.
Hoenig, M. and Van Hoeyweghn, P. (1986). Alternative to solid sampling for trace
metal determination by platform electrothermal atomic absorption spectrome-
try: Direct dispensing of powdered samples suspended in liquid medium, AnaL
Chem., 58: 2614.
Honig, R. E. (1963). Laser-induced emission of electrons and positive ions from
metals and semiconductors,App/. Phys. Lett., 3: 8.
Houk, R. S. (1986a). Mass spectrometry of inductively coupled plasmas, AnaL
Chern., 58: 97A.
Houk, R. S. (1986b). Laser ionization techniques for analytical mass spectrometry,
inAnaryticalApplications ofLasers (E. H. Piepmeier, ed. ), Wiley, New York, 1986.
Hurst, G. S., Payne, M. G., Kramer, S. D., and Young, J. P. (1979). Resonance ion-
ization spectroscopy and one-atom detection, Rev. Mod. Phys., 51: 767.
Hurst, G. S., Payne, M. G., Phillips, R. C., Dabbs, J. W. T., and Lehmann, B. E.
(1984). Development of an atom buncher. J. AppL Phys., 55: 1248.
Ishizuka, T. and Uwamino, Y. (1983). Inductively coupled plasma emission spec-
trometry of solid samples by laser ablation, Spectrochim. Acta, 38B: 519.
Ishizuka, T., Uwamino, Y., and Sunahara, H. (1977). Laser-vaporized atomic ab-
sorption spectrometry of solid samples, AnaL Chem., 49: 1339.
Ishizuka, T. and Uwamino, Y. (1980). Atomic emission spectrometry of solid sam-
ples with laser vaporization-microwave induced plasma system, Ana/. Chern., 52:
125.
Johnson, P. M. and Otis, C. E. (1981). Molecular multiphoton spectroscopy with
ionization detection. Ann. Rev. Phys. Chern., 32: 139.
Kang, H. Kang and Beauchamp, J. L. (1985). Pulsed laser evaporation of solid metal
targets. Implications for studying the gas-phase reactions of laser-generated
atoms and ions, J. Phys. Chern., 89: 3364.
Kantor, T., Polos, L., Fodor, P., and Pungor, E. (1976). Atomic absorption spectrom-
etry of laser-nebulized samples, Talanta, 23: 585.
Kantor, T., Bezur, L., Pungor, E., Fodor, P., Nagy-Balogh, J, and Heincz, G. (1979).
Determination of the thickness of silver, gold, and nickel layers by a laser micro-
probe and flame atomic absorption technique, Spectrochim. Acta, 38B: 341.
Kawaguchi, H., Xu, J., Tanaka, T., and Mizuike, A. (1982). Inductively coupled
plasma-emission spectrometry using direct vaporization of metal samples with
a low energy laser, Bunseki Kagaka, 31: E185.
Klocke, H. (1969). Utersuchungen zum materialabbau durch laserstrahlung, Spec-
trochim. Acta, 24B:
380 Sneddon et al.
Ko, J. B. (1984). New design of glow discharge lamps for the analysis of metals by
atomic emission spectroscopy, Spectrochim. Acta, 39B: 140.
Kronert, U., Becker, S., Hilberath, T., Kluge, H.-J., and Schultz, C. (1987). Reso-
nance ionization mass spectrometry with a pulsed thermal atomic beam, AppL
Phys.,A44: 339.
Kubiak, G. D. and Zare, R. N. (1985). Recombinative desorption dynamics: mole-
cular hydrogen from Cu (110) and Cu(111), J. Chern. Phys., 83: 2538.
Kwong, H. S. and Measures, R. M. (1979). Trace element laser microanalyzer with
freedom from chemical matrix effect, Anal. Chern., 51: 428.
Land, D. P., Tai, T.-L., Linquist, J. M., Hemminger, J. C., and Mciver, R. T. (1987).
Characterization of multilayer thin films by laser-induced thermal desorption
mass spectrometry, Anal. Chern., 59:2924.
Laqua, K. (1979). Analytical spectroscopy using laser atomizers, in Analytical Laser
Spectrometry (N. Omenetto, ed.), Wiley, New York.
Leis, F. and Laqua, K. (1978). Emissionspektral-analysis mit anregung des durch
laserstrahlung erzeugten damfes fester proben in einer mikrowellenentladung-1.
Grundlagen der methode und experimentelle verwirkichung, Spectrochim. Acta,
33B: 729.
Leis, F. and Laqua, K. (1979). Emissionspektral-analyse mit anregung dea durch
laserstrahlung erzeugten damfes fester proben in einer mikrowellententladung-
11. Analytisch andwendungen, Spectrochim. Acta, 34B: 307.
Lorbar, A. and Goldbart, Z. (1985). Convenient method for the determination of
trace elements in solid samples using an inductively coupled plasma, Analyst, 110:
155.
Luhman, D. M. (1987). Multi photon ionization mass spectrometry, Prog.Anal. Spec-
trosc., 10: 529.
Mak, C. H., Koehler, B. G., Brand, J. L., and George, S.M. (1987). Surface diffu-
sion of hydrogen on carbon-covered Ru(001) surface studies using laser-induced
thermal desorption,!. Chern. Phys., 87: 2340.
Matousek, J. P. and Orr, B. J. (1976). Atomic absorption studies of C0 2 laser-
induced atomization of samples confined in a graphite furnace, Spectrochim.
Acta, 31B: 475.
Mayo, S., Lucatorto, T. B., and Luther, G. (1982). Laser ablation and resonance
ionization spectrometry for trace analysis of solids, AnaL Chern., 54:553.
McCaffrey, J. T. and Michel, R. G. (1983). Carbon furnace for sample introduction
into a metastable nitrogen plasma, AnaL Chern., 55: 2175.
McCreary, D. A., Ledford, E. B., Jr., and Gross, M. L. (1982). Laser desorption
Fourier transform mass spectrometry,Anal. Chern., 54: 1437.
McCreary, D. A. and Gross, M. L. (1985). Laser desorption/Fourier-transform mass
spectrometry for the study of nucleosides, oligosaccharides and glycosides,AnaL
Chim. Acta, 178: 91.
McCreary, D. A., Peake, D. A., and Gross, M. L. (1985). Fast atom bombardment
and laser desorption mass spectrometry for determination of alkyltriphenylphos-
phonium salts, AnaL Chern., 57: 1181.
Measures, R. M. and Kwong, H. S. (1979). TABLASER: Trace (element) analyzer
based on laser ablation and selectively excited radiation,AppL Opt., 18: 281.
Miller, C. M., Cross, J. B., and Nogar, N. S. (1982a). Single photon photoionization
mass spectrometry via autoionizing resonances, Opt. Cornrnun., 40: 271.
Miller, C. M., Nogar, N. S., Gancarz, A. J., and Shields, W. R. (1982b). Selective
laser photoionization for mass spectrometry, AnaL Chern., 54: 2377.
Mitchell, P. G. and Sneddon, J. (1987). Direct determination of metals in milligram
masses and microliter volumes by direct-current argon-plasma emission spec-
trometry with sample introduction by electrothermal vaporization, Talanta, 10:
849.
Mitchell, P. G., Ruggles, J. A., Sneddon, J., and Radziemski, L. J. (1985). Direct
determination of copper in solids by laser ablation-direct current argon plasma
emission spectrometry, AnaL Lett., 18 (A14): 1723.
Mitchell, P. G., Sneddon, J., and Radziemski, L. J. (1986a). A sample chamber for
solid analysis by laser ablation/DCP spectrometry,Appl Spectro., 40: 274.
Mitchell, P. G., Greene, B., and Sneddon, J. (1986b). Direct determination of mer-
cury in algal cells by direct current argon plasma emission spectrometry with sam-
ple introduction by electrothermal vaporization, Mikrochern. Acta, 1: 249.
Mitchell, P. G., Sneddon, J., and Radziemski, L. J. (1987). Direct determination of
copper in solids by direct current argon plasma emission spectrometry with sam-
ple introduction using laser ablation,Appl Spectro., 41: 151.
Mohamed, N. and Fry, R. C. (1981). Slurry atomization direct atomic spectrochem-
ical analysis of animal tissues, AnaL Chern., 53: 450.
Moore, L. J., Fassett, J. D., and Travis, J. C. (1984). Systematics of multielement
determination with resonance ionization mass spectrometry and thermal atom-
ization,Anal Chern., 56:2770.
Ng, K C., Zerezghi, M., and Caruso, J. A. (1984). Direct powder injection of NBS
coal fly ash in inductively coupled plasma atomic emission spectrometry with
rapid scanning spectrometric detection, Anal Chern., 56: 417.
Nagar, N. S., Estler, R. C., and Miller, C. M. (1985). Pulsed laser desorption for
resonance ionization mass spectrometry, AnaL Chern., 57: 2441.
Petukh, M. L., Satsunkevich, U. D., and Yankovskii, A. A. (1976). Investigation of
laser sampling of material for spectral analysis, J. Appl Spectro., 25: 828.
Posthumus, M.A., Kistemaker, P. G., Meuzelaar, H. L. C., and Ten Noever de
Brauw, M. C. (1978). Laser desorption-mass spectrometry of polar nonvolatile
bio-organic molecules, AnaL Chern., 50: 985.
Reif, J., Fallgren, H., Nielsen, H. B., and Matthias, E. (1986). Layer-dependent laser
sputtering of BaF2 (111),Appl Phys. Lett., 49: 930.
Routh, M. W. (1987). A comparison of atomic spectroscopic techniques: atomic ab-
sorption, inductively coupled plasma and direct current plasma, Spectro., 2(2):
34.
Rudnevsky, N. K, Tumanova, A. N., and Maximova, E. V. (1984). Preliminary laser
sampling in the microspectral analysis of metals and alloys, Spectrochim. Acta,
39B:5.
Salin, E. D. and Horlick, G. (1979). Direct sample insertion device for inductively
coupled plasma emission spectroscopy, AnaL Chern., 51: 2284.
Schron, W. and Bomback, G. (1983). Schnellverfahren ftammanlosen AAS-bestim-
ming von spurenelementen in geologischen proben, Spectrochim. Acta, 38B:
1269.
382 Sneddon et al.
Schueler, B. and Odom, R. W. (1987). Nonresonant multi photon ionization of the

neutrals ablated in laser microprobe mass spectrometry analysis of GaAs and
Hgo.7sCdo.zz Te,J.Appl. Phys., 61:4652.
Scott, R. H. and Strasheim, A. (1970). Laser induced plasmas for analytical spec-
troscopy, Spectrochirn. Acta, 25B: 311.
Shankai, Z., Conzemius, R. J., and Svec, H. J. (1984). Determination of carbon,
nitrogen, and oxygen in solids by laser mass spectrometry, Anal. Chern., 56: 382.
Sherman, M. G., Kingsley, Hemminger, J. C. and Mciver, R. T., Jr. (1985a). Sur-
face analysis by laser desorption of neutral molecules with detection by Fourier-
transform mass spectrometry, Anal. Chirn. Acta, 178: 79.
Sherman, M. G., Kingsley, J. R., Dahlgren, D. A., Hemminger, J. C., and Mciver,
R. T., Jr. (1985b). Laser-induced thermal desorption utilizing FT-MS detection,
Surf Sci., 148: L25.
Sneddon, J. (1986). Approaches to sample introduction in atomic spectroscopy,
Spectroscopy, 1 (9): 34.
Sneddon, J. and Bet-Pera, F. (1986). Electrothermal vaporization-inductively cou-
pled plasma emission spectrometry, 1i·ends in Anal. Chern., 5 (5): 110.
Sneddon, J. and Mitchell, P. G. (1986). Direct determination of metals in solid sam-
ples by plasma ES with sample introduction by laser ablation,Arner. Lab., 18 (11):
21.
Sneddon, J., Mitchell, P. G., Ruggles, J. A., and Radziemski, L. J. (1986). Direct
determination of metals in solids by laser ablation-direct current argon plasma
emission spectrometry, Proc. Internat. Conf. '85, Dec. 2-6, Las Vegas, Nevada.
Sumino, K., Yamamoto, R., Hatayama, F., Kitamura, S., and Itoh, H. (1980). Laser
atomic absorption spectrometry for histochemistry, Anal. Chern., 52: 1064.
Tabet, J.-C. and Cotter, R. J. (1984). Laser desorption time-of-flight mass spectrom-
etry of high mass molecules, Anal. Chern., 56: 1662.
Tabares, F. L., Marsh, E. P., Bach, G. A., and Cowin, J.P. (1987). Laser photofrag-
mentation of photodesorption of physisorbed CH3 Br on lithium fluoride, 1
Chern. Phys., 86: 738.
Tembreull, R. and Luhman, D. M. (1987). Pulsed laser desorption of biological
molecules in supersonic beam mass spectrometry with resonance two-photon
ionization detection, Anal. Chern., 59: 1082.
Thompson, M., Goulter, J. E., and Sieper, F. (1981). Laser ablation for the intro-
duction of solid samples into an inductively coupled plasma for atomic-emission
spectrometry, Analyst, 106: 32.
Thompson, M. and Hale, M. (1984). Rapid analysis and identification of heavy min-
erals by laser ablation and inductively coupled plasma emission spectrometry,
Prospect. Areas Glaciated Terrain 225.
Thonnard, N., Wilis, R. D., Wright, M. C., and Davis, W. A. (1987). Noble gas count-
ing using RIS and TOF mass spectrometry II, first results, in Resonance Ionization
Spectroscopy (G. S. Hurst and C. G. Morgan, eds.), Inst. Phys., Bristol.
Tonan, G. F. (1974). Ph.D. Thesis No. 107, Univ. of London.
Unger, S. E., Schoen, A. E., Cooks, R. G., Ashworth, D. J., and DaSilva, J. (1981).
Identification of modified nucleosides by secondary-ion and laser-desorption
mass spectrometry, 1 Org. Chern., 46:4765.
Van Dalen, H. P. J., Kwee, B. G., and De Galen, L. (1982). The selective deter-
mination of halogens and sulphur in solution by atmospheric-pressure helium
microwave-induced plasma emission spectrometry coupled to a electrothermal
introductive system, AnaL Chirn. Acta, 142: 159.
Vander Peyl, G. J., lsa, K., Haverkamp, J., and Kistemaker, P. G. (1983). Thermal
aspects of laser desorption mass spectrometry (LDMS), Int. J. Mass Spec. Ion
Phys., 47: 11.
Vanderborgh, N. E. (1977). Laser induced pyrolisis techniques, inAnalytica/ Pyrolisis
(C. E. R. Jones and C. A. Cremers, eds.), Elsevier, Amsterdam.
Verdun, F. R., Krier, G., and Muller, J. F. (1987). Increased sensitivity in laser mi-
croprobe mass analysis by using resonant two-photon ionization processes, AnaL
Chern., 59: 1383.
Viswanathan, R. and Hussla, I. (1986). Ablation of metal surfaces by pulsed ultra-
violet lasers under ultrahigh vacuum, J. Opt. Soc. Arner., B3: 796.
Wagatsuma, K. and Hirakawa, K. (1984). Emission spectroscopic studies of sputter-
ing in a low-power glow discharge, Ana/. Chern., 56: 2024.
Wennrich, R. and Dittrich, K. (1982). Simultaneous determination of traces in solid
samples with laser-AAS, Spectrochirn. Acta, 37B: 913.
Wieser, P., Wurster, R., and Haas, U. (1981). Application of LAMMA in aerosol
research, Fren. Z. Anal. Chern., 308: 260.
Wieser, P., Wurster, R., and Seiler, H. (1984). Laser microprobe mass analysis de-
tection limits and lateral resolution, J. Physique Colloq., C2: 261.
Wilk, z. A. and Hercules, D. M. (1987). Organic and elemental ion mapping using
laser mass spectrometry,AnaL Chern., 59: 1819.
Yang, M. and Reilly, J. P. (1987). A new high-resolution TOF mass spectrometer,
AnaL Instrurn., 16: 133.
Yeh, J. T. C. (1986). Laser ablation of polymers, J. Vac. Soc. TechnoL,A4: 653.
10
Current New Applications
of Laser Plasmas
Allan A. Hauer, David W. Forslund, and Colin J. McKinstrie*

Justin S. Warkt
Laboratory for Laser Energetics
Rochester, New York
Philip J. Hargis, Jr. and Roy A. Hamil

Sandia National Laboratories
Albuquerque, New Mexico
Joseph M. Kindel
Mission Research Corporation
10.1 INTRODUCTION
As the previous chapters have illustrated, the applications of laser-pro-

duced plasma (LP's) have been a wide-ranging field of investigation for
more than two decades. In this final chapter, we will review several ap-
plications that (with one exception) have been active topics of investigation
only for the last few years.
We deal first with the application of laser-generated particles and x-rays.
This topic is largely an outgrowth of laser fusion research and primarily
employs the higher flux regime with irradiances > 1013 W/cm 2 • This topic
demonstrates a very satisfying diversity of practical applications ranging
Present affiliations:
• Laboratory for Laser Energetics, University of Rochester, Rochester, New York
tRoyal Society University Fellow, Oxford University, Oxford, England
385
386 Hauer et al.
from biology to materials science, as well as plasma and atomic physics, that
has been spawned by basic research in laser/matter interaction.
Next, we address the subject of laser LP acceleration. It has been recog-
nized for many years that extraordinary plasma conditions, such as very high
electric fields, are produced at high laser irradiances (> 1015 W/cm2 ). More
recently, it has been suggested that these conditions could be used as the key
element in a new type of high-energy particle accelerator. The field gradi-
ents and other parameters that may be available with these techniques are
such that a whole new area of accelerator technology might be opened up.
Finally, we discuss laser switching and controlling of pulsed power.
10.2 APPLICATIONS OF
LASER-PLASMA-GENERATED
X-RAYS AND PARTICLES
10.2.1 Introduction to the Physical Mechanisms
Involved in the Generation of
X-Rays and High-Energy Particles
The LP can be an efficient, bright, and versatile source of x-rays. As was

noted in Chap. 3, in some situations as much as 70% of the absorbed laser
light can be reemitted as x-rays. In order to have reasonably efficient radi-
ation in the x-ray regime, plasma temperatures in the range of hundreds of
e V or greater must be generated. In order to reach such conditions, laser ir-
radiances > 1012 W /cm2 must be achieved. With variations in irradiance and
wavelength that are allowed by present laser technology, x-ray photon en-
ergies can range across the spectrum up to several hundred keV. The x-ray
emission also has very interesting temporal characteristics. In many cases,
the temporal profile follows the envelope of the laser itself. This, in turn,
implies pulse durations ranging from tens of picoseconds (or less) to tens of
nanoseconds. The LP can thus be a very high peak-power x-ray source. Con-
trollability is another very valuable characteristic of LP x-ray sources. Be-
cause laser-initiated electronic and pulse-powered (see Sec. 10.4) switching
has become highly developed, the x-ray source can be accurately synchro-
nized with respect to a wide variety of instruments (e.g., ultrafast optical and
x-ray streak cameras and shutters). In addition, the laser itself is often used
to stimulate an event (such as a material phase transition), which can then
be probed with an x-ray burst. The spatial distribution of x-ray emission can
be varied from a point source (as small as 10 11m) to distributed emission
(as large as several mm) by simply varying the laser focal conditions.
The physical mechanisms responsible for kilovolt plasma formation and
subsequent x-ray emission show a marked variation with laser wavelength.
Shorter wavelengths (.A< 1 f.liD) are more efficient for producing quasither-
Current Applications 387
mal plasma emission (Yaakobi et al., 1981). Longer wavelength(,\> 10 J.lm)

laser interactions have the additional feature of producing large numbers of
suprathermal electrons that, in turn, produce hard x-ray line radiation and
bremsstrahlung (Hauer et al., 1984).
Equally remarkable is the particle emission from LP's. The thermal por-
tion of LP's produces an ion blowoff that is very valuable for diagnostics but
is less interesting from an application point of view. The aforementioned
suprathermal-electron generation is also the driving source for the produc-
tion of very-high-energy ions. This is usually referred to in the literature as
"fast ion" generation. Both the particle energy and efficiency of generation
are enhanced by going to longer-wavelength laser irradiation, and a C0 2
laser is thus the best choice for such work.
It is useful when we consider the mechanisms for LP generation of x-
rays and particles (as was done in Chap. 3) to refer to the spatial structure
of the interaction, as is illustrated in Fig. 10.1. Laser energy is absorbed,
and a portion of it is conducted to a region near the ablation surface where
the strongest x-ray emission (> 1 keV) occurs. The efficiency of emission is
particularly dependent on the incident laser irradiance and wavelength.
Typical density and temperature profiles are shown in Fig. 10.2. At higher
intensities, the profiles tend to steepen, and the regions of laser deposition
and maximum x-ray emission tend to be of smaller spatial extent.
Both particle and x-ray generation mechanisms can be divided into two
broad categories: thermal and suprathermal. Laser absorption processes
were discussed in more detail in Chap. 3 and 6. It was noted there that
the absorbed laser energy can be channeled into either thermal electrons
that produce thermal plasma heating or suprathermal electrons that tend
to escape or deposit deeply in cold material. We will divide the discussion
into two sections dealing with thermal and suprathermal x-ray generation.
(A) Thermal X-Rays
In assessing the characteristics of the production of thermal x-ray photons of

any particular energy, we must consider the following parameter variations:
1. Laser absorption (for.\< 1 J.lm) increases with decreasing.\. For C0 2 laser

(.\ = 10.6 J.lm) irradiation, there is a special dependence. Typical experimen-
tal data on laser absorption as a function of wavelength and irradiance are
given in Fig. 10.3.
2. The fraction of energy channeled into hot electrons increases roughly lin-
early with the parameter I .\ 2 (where I is irradiance).
3. For a given laser wavelength, the intensity must be increased to reach higher
x-ray photon energies. This is because the rate of energy delivery must be
increased in order to reach higher plasma temperatures and thus higher ion-
ization states.
388 Hauer et al.
Hot e's
Ablation
surface
Critical Expanding
surface sheath and
(nc) plasma blow off
Figure 10.1 Laser plasma interaction-absorption/transport phase.
4. For generating short pulses, ( < 500 psec), short wavelengths should be more
efficient since the higher critical density should lead to a higher collision rate
and more rapid achievement of higher ionization states. Several studies of
these variations have been made (Yaakobi et al., 1981; Matthews et al., 1983;
Phillion and Hailey, 1986), but parameter studies for all the regimes of in-
terest in applications have not yet been completed. In Table 10.1, we try to
give a rough indication of the range of parameters that are reasonably agreed
on in the literature. We define the following efficiency parameter: 'TJ =x-ray
energy in a particular photon energy band (which is noted) into 4rr/incident
laser energy.
In many applications in materials science and biology, a pulse width of
1 to 10 nsec is adequate for the study of transient phenomena. This pulse
width must be balanced against the need to deliver energy fast enough (high
ne Te ne Te
K
/e n Te
ns 2.0 1.0
keV keV
nc nc
nc/4
100 ,urn X 200,um X

(a) (b)
Figure 10.2 Typical density and temperature profiles for (a) high and (b) low den-
sity irradiation. nc is the critical density and n 8 is solid density.
laser power) so that high ionization states can be reached and thus relatively
short-wavelength x-ray radiation generated (before cooling by thermal con-
duction, hydrodynamic expansion, and radiation occurs.) Radiography of
dense plasmas often requires bursts of 100 psec or shorter (see Chap. 3).
This, in turn, requires the use of shorter-wavelength laser irradiation of the
x-ray source.
--
"#. 80
z
0
I- 60
a..
a:
0
en 40
Ill
<(
20
o~~~~~~-L~~~~-L~~~
1013 1014 1015 1016

INTENSITY (Wcm-2)
Figure 10.3 Laser absorption as a function of wavelength and irradiance.
390 Hauer et al.
Table 10.1 Characteristics of Thermal X-Ray Emission from

Laser Plasmas
Characteristics
X-ray ofx-ray generation
emission (for laser wavelength>.
parameters and intensity/) Applications
Broadband For given I considerable Photolithography

subkilovolt increase of "' with
A.v""0.05-1.0 keY decreasing >.
peak efficiencies Surface
>. = 0.35 r,m is "'"" 0.55 studies
at/"" 10 4 W/cm 2
Thermal line >. = 0.35 J.Lm, "'"" 0.02 X-ray radiography

radiation I =5 x 1014 W/cm 2 (e.g., biology)
hv""2 keY
(He-like >. = 1.06 J.Lm, "'"" 0.006
emission for I = 5 x 1014 W/cm 2 Transient
elements like diffraction
AI) Recombination rates from biological
A.v"" 10 eY generally cause x-ray structures
pulse > laser pulse
(for rL::;; 100 psec)
Thermal line >. = 0.35 J.Lm, "'"" 0.006 Transient

radiation hv"" I = 5 x 1014 W/cm2 diffraction
5 keY (He-like in materials
emission for A= 1.06 J.Lm, T/""5 X 10- 4 science and
elements like Ti) I = 5 x 1014 W/cm 2 biology
X-ray pulse usually
closely follows
laser pulse
(B) Suprathermal X-Rays

The generation of supra thermal x-rays is a very different phenomena from
thermal emission (Priedhorsky et al., 1981; Burnett et al., 1984; Hauer et al.,
1984; Forslund and Goldstone, 1985). In essence, the process amounts to
the use of the absorbed laser energy to produce an electron beam, which in
turn has many similarities to a classical x-ray tube. Referring once again to
Fig. 10.1, we see that suprathermal electrons are usually generated in the
region from somewhat below nj4 to nc (where nc is the critical density).

The fraction of absorbed laser energy that goes into suprathermal ("hot")
electrons is proportional to the parameter(/ >. 2 ). For more precise infor-
mation on the scaling laws, the reader should consult the above references,
but with respect to applications, one can assume that for C0 2 irradiation
(>. = 10.6 11m), essentially all ofthe absorbed energy goes into hot electrons.
A simplified picture of hot-electron generation is adequate to understand
the bounds on high-energy x-ray generation. Most of the hot electrons are
generated near the critical surface. This hot-electron energy is resident in a
Maxwellian distribution with a temperature designated T HOT· The depen-
dence ofTHOT on laser wavelength and laser intensity over a wide parameter
range can be very complicated. With respect to applications, however, we
restrict our view to the range 1015 to 1017 W/cm2 and C0 2 laser irradiation.
In this regime, we can use the following relation:
T HOT"' (14.8)/0.53 (10.1)
where I is in units of 1014 W/cm 2 and T HOT in keY. Outward moving elec-
trons create a charge imbalance, and electrostatic potentials of hundreds of
kilovolts can build up. This potential soon forms a sheath of positive ions,
which tends to reflect some of the subsequent electron emission. The hot
electrons lose some of their energy by interacting with this sheath that can,
in turn, produce ion acceleration (described in the next section). The bulk
of the remaining energy is deposited in the dense portions of the target.
The stopping of fast electrons in dense matter is due to ionization and exci-
tation of bound electrons, collective loss to free electrons (Jackson, 1975),
and bremsstrahlung radiation. Bremsstrahlung and inner-shell radiation
excited in this slowing-down process are potentially valuable in applica-
tions. A rough indication of the energy appearing in bremsstrahlung can
be obtained through the use of the following relation (Birkhoff, 1958):
electron energy lost to bremsstrahlung _ EZ

(10.2)
total deposited energy - 800
where E = electron energy in MeV.

The hot electrons also produce inner-shell vacancies that are filled by
transitions, which produce line radiation. For the applications presently
being considered, K-shell radiation (e.g., Ko:, K 13 emission) is the most im-
portant.
Detailed measurements have been made of the absolute yields of both
bremsstrahlung continuum (Priedhorsky et al., 1981) and inner-shell line
radiation. Table 10.2 gives a review of some parts of this data in the context
392 Hauer et al.
Table 10.2 Characteristics of Suprathermal X-Ray Emission

from Laser Plasmas
Characteristics
of x-ray generation
X-ray emission (for laser wavelength>.
parameters and intensity I) Applications
Characteristic A= 10.6 J.Lm Transient

line emission 11 "'4 x 10-5 diffraction
hv~BkeV at/ "'3 x 1015 W/cm 2 from small 2-d
Av"-'5 eV spacing crystals
X-ray pulse
follows laser
pulse for Radiography
rL"' 1 nsec of very dense
plasmas
Bremsstrahlung A= 10.6 J.Lm Transient

continuum I "'3 x 1015 W/cm2 powder
(measured) diffraction
10-500keV 77"' 10-3
for total spectrally integrated Absorption
bremsstrahlung spectra from
very dense
plasmas
Transient
effects with
respect to
radiation
damage in
solids
of potential applications. As can be seen from Table 10.2, suprathermal

x-ray generation is quite inefficient (as is an ordinary x-ray tube). It is, how-
ever, one of the few ways to produce an x-ray source that combines all of
the following characteristics:
1. High photon energy up to at least 30 keY.

2. Small source size ~ 100 J.Lm.
3. Flexibility in changing source configuration by changing focal conditions.
4. Intense fast r :S 1 nsec pulse bursts.
In addition, the technology of C0 2 lasers has progressed to the point

where compact kilojoule, high-repetition-rate (> 0.01 Hz) systems can be
planned.
(C) High-Energy Particles

Hot electrons that are confined to the coronal region by the sheath and
self-generated magnetic fields (Forslund and Brackbill, 1982; Forslund and
Goldstone, 1985) can impart a significant portion of their energy to ions. In
C02 laser experiments at high density ("' 1016 W/cm2 ), as much as 60% of
the absorbed laser energy has been measured in protons with > 100 keV
energy (Hauer et al., 1984). A very simple model of the ion accelera-
tion gives some feeling for the mechanisms involved. The hot electrons
can be thought of as making multiple bounces off the sheath. On each
bounce, they give up a fraction "'2JZ!A(melmp) to ions, where ZIA is
the atomic charge to mass ratio and me, mp are the electron and proton
masses, respectively. As a specific example, in a typical C0 2 experiment
("' 2 X 1015 W/cm 2 with T HOT rv 120 keV), about 40% of the absorbed en-
ergy is observed in fast ions. More detailed calculations indicate (Forslund
and Brackbill, 1982) that about eight effective bounces may occur. For a
gold shell, 8 x 2JZ!A(melmp)= 0.36 or close to the observed fraction.
With respect to applications, the fast ion blowoff has another unique
property. The ions tend to be emitted in a fairly narrow cone or jet. The
self-generated magnetic fields tend to cause a preferential deposition of hot
electrons in high field regions. This, in turn, produces an electric field
(10.3)
where P is the pressure due to the deposition of hot electrons having a den-
sity ne. This accelerating field is much larger than it would be without the
magnetic field and extends over a region much larger than the laser spot
size. This field tends to channel the blowoff into an ion jet normal to the
target surface. In Fig. 10.4, we show typical data on the angular distribu-
tion of fast ion emission. As can be seen, a significant fraction of the ion
energy is contained in a narrow cone angle. Another graphic illustration of
jet formation is given in Fig. 10.5. Here, a composite x-ray and ion emission
image clearly display the formation of an ion jet.
With respect to the applications discussed here, high-energy protons are
the most important. In Table 10.3, we summarize some of the available
experimental information that would be useful in applications and for com-
pleteness include information on heavier ions. At higher laser irradiances,
a. significant flux of particles greater than 1 MeV should be produced at ef-
ficiencies at or greater than those shown in Table 10.3.
394 Hauer et al.
1000
(/)
z
0
LL.
0 • ALL IONS
a:
100 .•/cE>o kev>
w
m
::E
::::::>
z
w 10
> E>500 keV D.
1-
<(
...J
w
a: 1
0 20 40 60
ANGLE FROM
NORMAL (degrees)
Figure 10.4 Angular distribution of fast ion emission.
10.2.2 X-Ray Applications
The unique potential of LP x-ray sources in nonfusion applications has been

recognized for some time (Hauer, 1976; Mallozzi et al., 1979a). Consider-
able improvement in parameters such as repetition rate and reliability for
Nd-glass, C0 2 , and KrF lasers has led to such a substantial amount of work
on nonfusion applications in recent years (Eason et al., 1986; Yaakobi et al.,
1986). Most of the applications make use of the fast intense x-ray burst that
can be produced with LP's. The energy available in a single burst is often
adequate for detection and recording. LP x-ray sources are thus (in con-
trast to synchrotron radiation) uniquely suited to the study of nonrepetitive
or irreversible events (such as some types of material phase changes). Laser
systems suitable as x-ray sources will soon be capable of > 1 Hz repetition
rates, and thus, LP's will even be competitive on a quasi-cw basis. X-ray
fluence available with existing laser systems ranges up to 1016 keV/keV/sr
with photon energies up to 10 ke V or greater.
The applications discussed here fall into three broad categories: (1) x-ray
diffraction from transiently strained crystals, (2) pulsed x-ray radiography
and lithography, and (3) absorption spectroscopy.
Figure 10.5 Composite x-ray/ion photograph showing the formation of an ion jet.
(A) X-Ray Diffraction from

Transiently Strained Crystals
The fast burst of x-ray emission that is available from laser plasma sources
is uniquely suited to transient x-ray diffraction (Forster et al., 1984; Hauer,
396 Hauer et al.
Table 10.3 Summary of the Characteristics of LP Fast Ion

Generation (EL =incident laser energy, E; =ion energy)
Laser irradiance Ion emission
Type of ion W/cm 2 characteristics
Protons E; > 100 keY

Total energy= 0.35EL
Energy in 10° cone = 0.15EL
E;>500eV
Total energy= 0.15EL
Energy in 10° cone= O.OSEL
Protons 5 X 1015 E;>lOOkeV

Total energy = 0.15EL
Taions E;>100MeV
(Z .-v40) Total energy= 0.15EL
1976). Using currently available lasers, investigators can record diffraction

patterns from a wide variety of crystalline materials on a single shot.
There is currently much interest in the use of pulsed lasers to anneal
crystals by transient heating of the crystal surface. A good understanding
of the structural dynamics requires a knowledge of the temperature-depth
profiles within the crystal as a function of time. One means of probing these
profiles is by short-pulse x-ray diffraction.
The basic idea is that when the crystal is irradiated by the laser, it heats
up; the lattice spacing then increases by an amount determined by the co-
efficient of thermal expansion. Any x-ray radiation incident on the crystal
is thus diffracted at a lower angle than that for unheated material. This is
easily seen by simple differentiation of Bragg's law
t::..d
-d = cot0/!:::..0 (10.4)
where d is the spacing between diffraction planes and () the Bragg angle. In
practice, there is a thermal gradient within the crystal, and thus there will
be a range of angles at which diffraction takes place up to a maximum angle
corresponding to the hottest part of the crystal, which is assumed to be at
the surface of the solid phase. The x-rays from a LP allow such a Bragg
diffraction pattern to be recorded on a single-shot basis with a probe time
as low as 100 psec (similar techniques utilizing synchrotron radiation need

multiple shots). ·
Several experiments have been performed in this area using a LP x-ray
source (Lunney et al., 1986; Wark et al., 1986). The experimental setup of
Lunney et al. (1986) is shown in Fig. 10.6. The silicon (111) crystal is irra-
diated with 1.05 Jlm light at a fluence of"' 0.4 J cm- 2 in a pulse length of
1 nsec FWHM. Mter the heating beam, a pulse of x-rays is produced by
tightly focusing another laser beam, synchronous and delayed with respect
to the first, onto a polyvinyl chloride (PVC) target. At such high irradi-
ances ("' 1016 W/cm2 ), He-like chlorine is produced, which has a resonance
line at 4.44 A. Bragg diffraction patterns from the heated and unheated
portions of the crystal were obtained for various delay times between the
heating beam and the x-ray probe beam. The spectrum from the heated
and unheated parts of the crystal can then be deconvolved to yield the rock-
ing curve (reflectivity as a function of angle around the Bragg angle) as a
function of time. The rocking curve of a silicon crystal heated close to the
melting point is two orders of magnitude wider than the natural width.
It is possible to extract the temperature-depth profiles from these rock-
ing curves using dynamical diffraction theory; the technique has been used
for several years to obtain strain-depth profiles from crystals strained me-
chanically, by the introduction of dopants, as well as from induced thermal
gradients. We refer the reader to the work of Burgeat and Taupin (1986)
and Klar and Rusticelli (1973) for a full explanation of the theory. It suf-
fices to say here that the basic idea is to solve Maxwell's equations to first
DIRECT
X-RAY
PVC SHIEL~ H~t~T~~G

TARGET ' ~
ALUMINIZED LIGHT-TIGHT
MYLAR BOX
X-RAY
GENERATING SILICON CRYSTAL
LASER PULSE
Figure 10.6 Experimental configuration used to probe laser annealed crystals with
short LP x-ray bursts (from Lunney et a!., 1986 with permission).
398 Hauer et al.
order within a medium of complex and periodic refractive index (i.e., the
crystal). The strain is introduced via a depth-dependent Bragg angle. A
monotonic strain-depth profile is guessed and the rocking curve calculated.
The calculated rocking curve is compared to that obtained experimentally,
and the guessed strain altered accordingly. Thus, the iterations proceed un-
til the experimental and calculated rocking curves agree within some pre-
determined limits; at this point, the strain profile used in the calculations
is taken to be the actual profile within the crystal. The temperature-depth
profile is obtained from the strain profile using temperature-dependent co-
efficients of expansion and allowing for the one-dimensional nature of the
expansion. The temperature profiles for the experiment described above
are shown in Fig. 10.7. We calculated the curves in Fig. 10.7 for data from
Lunney et al. (1986). Such profiles can be compared to detailed modeling
of the laser/crystal interaction using theoretical techniques, such as those
described in Sec. 3.3 of Chap. 3. The information gained from such com-
parisons will help to optimize the laser annealing conditions. At higher ir-
radiances, this pulsed diffraction diagnosis may help to reveal important
characteristics of the onset of plasma formation (Hauer, 1976).
Transient diffraction studies have also been performed on a subnanosec-
ond time scale (Wark et al., 1986). The first motivation in this work was to
1700
~ 1500
~ 1300
::::>
1-1100
<(
a:
w 900
(L
::lE
w
1- 500
4.5
Figure 10.7 Temperature depth profiles within a laser heated crystal at various
probe times.
investigate the possibility of laser heating as a method for x-ray switching

and shuttering. One concept for doing this is illustrated in Fig. 10.8. To ob-
tain an "on" switch of line radiation, a beam block is placed in the way from
the positive D..(} side, just obscuring the line, as in Fig. 10.8(a). Upon irradi-
ance, the rocking curve broadens to negative D..B, allowing radiation to pass
through. The switching ratio here is high, between two and three orders of
magnitude, but good resolution of the line is lost. An off switch may be made
utilizing a double-crystal arrangement, as shown in Fig. 10.8. The radiation
is double-Bragg reflected off parallel crystals. Upon illumination of the first
crystal, the diffracted radiation that now satisfies the Bragg criterion has
fallen in intensity due to the reduction in peak reflectivity of the first crystal;
i.e., the intensity falls by about a factor of 20. This arrangement would also
be compatible with an extended source ofx-rays, providing the angular size
of the source is small compared with the angle between two diffracted lines.
One configuration used for testing these concepts is very similar to that
used in Fig. 10.6, with two important exceptions. Both the heating and x-ray
source beams were of about 1 nsec duration. The time resolution in this case
was provided by using a streak camera for detection of the diffracted x-rays.
The experimental data are shown in Fig. 10.9. Figure 10.9(a) shows the tem-
porally resolved spectrum of the radiation diffracted from an unirradiated
crystal. The two lines are the resonance and intercombination line of He-
like chlorine. The dielectronic satellite lines, which appear as a broad band
on the higher wavelength side of the other two pictures, are not present
here due to a slight misalignment of the crystal, causing the satellite to miss
the photocathode. Figure 10.9(b) shows the diffracted radiation from an
illuminated crystal with the peak of the heating beam delayed with respect
to the peak of the x-ray producing beam by 500 psec. The lines are seen to
broaden and merge together toward the end of the pulse. In Fig. 10.9(c),
there is no delay between the heating and x-ray-producing pulse, resulting
in broadening early in the pulse. The broadening completely smears out the
PATH OF X-RAYS
AFTER CRYSTAL
~~
HEATING
CRYSTAL UPON HEATING OF THE

FIRST CRYSTAL, X-RAYS
ARE SHIFTED FROM THE
CRYSTAL
ioNSWITCHj IOFF SWITCH I BRAGG ANGLE OF THE
SECOND CRYSTAL
Figure 10.8 Concepts for producing x-ray switching and shuttering.

400 Hauer et al.
4.44 A-
jo.os A
t-
1 ns
Streak from an unirradiated crystal
(a)
(b) Heating pulse 500 ps after x-ray-producing pulse
(c) Heating pulse coincident with x-ray-producing pulse

Figure 10.9 (a) Time-resolved x-ray spectrum from a crystal with no heating pulse,
(b) heating pulse 500 ps after x-ray producing pulse, and (c) heating and x-ray pro-
ducing pulses coincident (from Warr, Hauer, Kilkenny, 1986 with permission).
three original peaks discernible at the start of the pulse. From the angular
separation of the peaks (of order 1000 arc-sec), we gain the approximation
of the width of the rocking curve to be at least 500 arc-sec. This compares
with an original width (in the unperturbed crystal) of the order of 50 arc-sec.
In the absence of a double-crystal arrangement, the rocking-curve width
of the unirradiated crystal was estimated by placing x-ray film in place of the
streak camera (the instrumental broadening of the streak camera was of the
order of 250 arc-sec FWHM). From the structure of the dielectronic satel-
lite features in the resultant spectrum, we estimate the rocking-curve width
to be of the order of 50 arc-sec. The contribution to the spectral broadening
from the width of the 30 pm source was of the order of 30 arc-sec.
A similar theoretical analysis was performed on this data as that de-
scribed above. From the time-dependent rocking curves and unperturbed
crystal parameters, it was found that several hundred arc-seconds away from
the Bragg angle, the reflectivity changes from an order of 10-5 to 5 x 10-2
in less than 200 psec. The usefulness of such switching properties for diag-
nostic purposes in fields, such as laser fusion, may ultimately depend on the
maximum obtainable temporal resolution (in this experiment, the switch-
ing time is less than 200 psec).
Transient diffraction techniques have also been applied to the study of
laser-generated shock waves (Wark et al., 1987). Laser generation of multi-
kilobar shocks has several applications to the study of high explosives (Yang,
1974), high-velocity impacts (Pirri, 1977), the alteration of the mechanical
properties of certain alloys and ceramics (Fairand et al., 1972), and solid-
state phase changes. For the more conventional shock drivers, the shock-
launching region, i.e., the initial layer of the material of interest from which
the shock begins, is obscured by the driving mechanism, whether it be a
flyer plate, tamper, explosive material, etc. For this reason, well-resolved
experimental studies of the shock-launching region have been intractable.
Laser-driven shocks do not present this limitation. Some of the most com-
mon methods of studying such shock waves include interferometric methods
(Sheffield and Fisk, 1983 and 1984) or the use of piezoelectric transducer
gauges (Graham et al., 1965). However, these methods yield little informa-
tion on the dynamics of the near-front face, where the shock wave is initi-
ated. Probing of the surface by short-pulse x-ray diffraction can overcome
this problem. The brightness and duration of the x-ray pulses from LP's
allow unique measurements of the density in the shocked region.
A laser-generated shock is launched into the crystal of interest. At some
point during the shock launching, a short (100 psec) pulse of x-ray-line ra-
diation (produced by a laser beam synchronous and delayed with respect
to the shock-generating beam) is Bragg-diffracted off the front surface of
the compressed crystal. The compression of the crystal changes the Bragg
condition according to Eq. (10.4). Thus, the line radiation is diffracted at a
402 Hauer et al.
higher angle than that for the unperturbed crystal, and this angular shift im-
mediately tells us the change in the interatomic spacing within the shocked
region. As there will be a density gradient at the crystal surface, there will
be a range of angles at which diffraction takes place, as well as an overall
shift. Thus, information about the density gradient, as well as peak den-
sity, can be obtained, analogous to the temperature gradient information in
the annealing experiment. Indeed, it should be possible to extract a density
profile from the data contained in the diffracted pulse.
An experiment of this sort was performed on the JANUS laser system
at the Lawrence Livermore National Laboratory. The shocked targets con-
sisted of 250 Jtm thick (111) silicon wafers 5 em in diameter, the surface of
which had been coated with 1000 A of aluminum and then 25 Jtm of plas-
tic (CH). The motivation for such a target design is explained below. Half
of the target was irradiated with the 1 nsec pulse of 1.06 Jtm laser light at
a ftuence varying from 0.8 to 8 J cm- 2 with a beam diameter on target of
"'4 em. A beam block prevented irradiance of the other half of the target;
diffraction from this unshocked region gave a reference point from which
to measure the changes in Bragg angle. After the shock had been launched
in the silicon crystal, a second laser beam containing "' 10 J of 0.53 to Jtm
light in 100 psec, which was synchronous with known and variable delay with
respect to the shock-launching beam, was focused to a 40 Jtm diameter spot
on a calcium-containing target. The heliumlike calcium lines thus produced
were Bragg-diffracted off the silicon and recorded on x-ray film.
The rationale behind the target design was that the absorption length
of silicon at room temperature to 1.06 Jtm light is several mm; previous
experiments have shown that, under similar irradiance conditions to those
used here, uncoated silicon is simply heated by the laser light, rather than
shocked. The aluminum coating, therefore, acts as an absorber for the
incident radiation, driving the shock into the material. Overcoating the
aluminum with plastic transparent to 1.06 Jtm light causes the expanding
aluminum plasma to be inertially confined between the silicon and plastic,
which in turn increases the strength of the shock launched into the silicon.
Such overcoating techniques have been used to enhance shock pressures in
the range of tens of kilobars for several years (Anderholm, 1970).
A typical diffraction pattern is shown in Fig. 10.10. The three lines in the
unshocked region are the resonance line (3.179 A), intercombination line
(3.196 A), and a group of unresolved dielectronic satellites of He-like cal-
cium (Feldman et al., 1974). It can be clearly seen that in the shocked region
of the crystal, the diffracted radiation is broadened and shifted to greater
angle, corresponding to a reduction in the lattice spacing. Thus, a simple
measurement of the maximum angular shift gives us a direct measurement
of the peak density, and knowledge of the broadening yields information on
the distribution of lattice spacings within the probed region.
I t I
1
1 I f t
t I l
'tooam = 0.5 ns
BRAGG ANGLE (degrees)
(a) (b)
Figure 10.10 Laser-shocked silicon (a) X-ray line spectra diffracted from silicon
shocked at an incident laser energy density of 4 ± 0.3 J em- 2 (an irradiance of 4 x
109 W/cm 2 ) are shown adjacent to reference lines simultaneously diffracted by the
unstrained silicon lattice, from a series of beam delays. Each photograph represents
a separate shot and beam delay. (b) Densitometer scans, taken at indicated locations
through the spectra in (a), are shown for a typical unshocked spectrum, and shocked
spectra at 0.0 ns, 0.5 ns, and 3.0 ns beam delays. ·
The probe depth of the x-rays in such highly strained crystals is deter-
mined by the photoelectric absorption coefficient rather than the extinction
length. This is because x-rays incident at a particular angle only diffract off
that thin region of the crystal at which the lattice spacing is such that the
Bragg condition is satisfied; the rest of the crystal up until that point simply
404 Hauer et al.
acts as an x-ray filter. The calcium He-like lines have a 1/e depth in silicon
of"' 10 JLm. If we take into account the angle of incidence and the dynamic
range as determined by the x-ray film, a maximum probe depth of between
4 and 6 Jlm is estimated.
Diffraction measurements, such as those shown in Fig. 10.10, were made
for a variety of levels of irradiance and delay times, with zero delay being
defined as the point where the peaks of the short and long pulse were coin-
cident. Thus, a measurement of the peak density is a function of time for
those shots where the average fluence was 4 J cm-2 • A maximum change of
the lattice spacing of 3.5% was observed, and the data are consistent with a
pressure pulse with a FWHM of 1 nsec.
It can be seen that at 0 nsec delay, the unperturbed part ofthe crystal can
still be probed, i.e., we still observe diffraction at the original Bragg angle
as well as diffraction from a range of lattice spacings up to a compression
of 2.2%. The obvious physical explanation for this is that at this early time,
the foot of the shock wave has not yet penetrated past the maximum probe
depth, whereas for all data collected after this probe time, when the shock
has proceeded further into the material, we no longer observe diffraction
from the unperturbed crystal. It is interesting to note that even 5 nsec af-
ter the peak of the shock-producing pulse, the surface of the crystal is still
slightly under compression, i.e., the rarefaction wave was not observed. This
is due to the fact that the aluminum plasma is tamped between the silicon
and plastic overcoat; this observation is in agreement with hydrocode pre-
dictions (using modeling techniques described in Chap. 3 and 6).
These recent experiments open up unique possibilities for the study of
transient phenomena within matter. Some of the more exciting prospects
include the time-resolved study of phase changes and the properties of ma-
terials within the elastic-plastic transition, as well as the dynamics of mate-
rials in a high-radiation flux environment.
(B) Time-Resolved EXA.FS

X-ray diffraction is not the only way to obtain information about the inter-
atomic spacings within a material; another method is to observe the EXAFS
(extended x-ray absorption fine structure). In this technique, the x-ray ab-
sorption coefficient of the material of interest is studied as a function of
energy just above the K, L or M absorption edge. As the name suggests,
on the high-energy side of the edge, fine structure in the absorption coef-
ficient exists, and this structure can yield information about the local en-
vironment surrounding an atom. In particular, measurements of the bond
lengths and coordination numbers can be made. The theoretical interpre-
tation of this structure and the means by which physical information can be
extracted were first put forward by Sayers et al. (1971). A few years later,
EXAFS spectra were recorded using a synchrotron as the x-ray source (Kin-
caid and Eisenberger, 1975). Since that time, interest in the technique has
expanded rapidly, and EXAFS spectra are now regularly recorded at many
synchrotron facilities.
The first spectra using aLP as the x-ray source were recorded by Mal-
lozzi et al. (1979b ), and more recently, spectra of higher quality have been
recorded by Eason et al. (1984) and Bradley (1985). These initial mea-
surements were made with a static sample using a 1 nsec probe pulse. The
short recording time afforded by LP's allows measurements to be made on
samples that would normally suffer radiation damage during repetitive syn-
chrotron radiation exposure. However, the real advantage of the LP source
is the availability of high temporal resolution, which allows bond lengths
and coordination numbers to be measured in materials undergoing tran-
sient changes on a subnanosecond time scale.
In the first EXAFS measurement with such time resolution, a 1 nsec pulse
of x-rays was produced by tightly focusing "' 40 J of 0.53 pm light onto a
bismuth target. At such irradiances, the x-ray spectrum of highly ionized
bismuth includes a broad band of emission around the Kedge of aluminum
at 1.5 keY. The x-rays passed through a 3 pm thick aluminum foil and were
diffracted off a flat thallium acid phthalate (TIAP, 2d = 25.9 A) crystal onto
an x-ray streak camera. A sequence of such shots was taken, with the x-ray
source moved closer to the aluminum foil on each successive shot. Thus,
the foil was transiently heated, as well as probed, by the x-rays. Typical data
from this experiment are shown in Fig. 10.11, for which the foil was "'30 mm
away from the x-ray source. Further shots with the foil 3 mm and 4 mm
from the x-ray source show evidence of gross changes in the fine structure
during the pulse. At these distances, calculations show that the foil should
be melted by the x-ray flux. A preliminary analysis of this data by Bradley
(1985) has shown clear evidence of an increase in the interatomic spacing
during the 1 nsec pulse, with an expansion of"' 20% by the end of the pulse.
The use of EXAFS to explore subnanosecond phenomena has been
proven. The two techniques of Bragg diffraction and EXAFS complement
each other well. Almost any structural change that can be synchronized to
the laser can be probed by one or both of these techniques, and thus we
can look forward to a significant increase in our understanding of transient
phenomena over the next few years.
(C) X-Ray Diffraction and Radiography

of Biological Samples
The study of living cells is of obvious fundamental importance in medical

and biological research. Although much can be learned using conventional
microscopy in the optical region, a basic limitation is placed on the max-
EX AF S SPECTRUM
-t_[
_ t +150ps J
_ t +300psl.._.
_ t+450ps~
.,.. TIME -t+600psl.

~
FIDUCIAL
t
Al K-edge ENERGY
Fig. 10.11 Typical streaked EXAFS spectrum recorded using a 1 ns x-ray pulse (from Eason et al., 1984 with permission).
imum spatial resolution by the wavelength of the light used. The resolu-
tion can obviously be improved by probing with shorter wavelengths, i.e.,
electron microscopy or soft-x-ray contact microscopy. However, in electron
microscopy, the sample must be dried and coated before observations can
be made; the cell is thus killed and the sample may appear vastly different
than its living counterpart. Similar problems occur with x-ray microscopy
when a conventional x-ray source is used. A contact print of the sample is
taken by irradiating it with x-rays over a period of several hours. During
this time, the cell suffers severe radiation damage and is killed during the
course of the exposure. If a 1 nsec burst of LP x-rays is used as the source,
this problem is overcome; the radiation damage occurs on a microsecond
time scale; thus, a print of the living cell is obtained before cell death oc-
curs.
The use of LP x-ray sources in this field has been applied by Michette
and co-workers to study a host of biological and botanical specimens. We
refer the reader to their excellent works (Michette et al., 1986; Eason et al.,
1986) for details of the particular results; only the general technique will be
described here. The living cells exist within the necessary nutrients under
atmospheric pressure. A thin Si3 N4 window, transparent to the probing x-
rays, separates them from the surrounding vacuum. The detector in these
experiments is a photoresist in which the image is formed by radiation dam-
age in a plastic material. The radiation damage centers then cause selective
etching when the plastic is exposed to an acid. The cells are in intimate
contact with a copolymer x-ray resist. The LP soft x-rays pass through the
cells and produce a latent image within the resist, corresponding to a two-
dimensional image of the line-integrated mass-absorption coefficient of the
sample. Subsequent etching yields an image of the cell as a relief pattern
within the resist.
To obtain high-contrast images of the specimens, the x-ray wavelength is
chosen to be within the carbon and oxygen K absorption edges (between 44
and 23 A). Thus, the water within the samples is relatively transparent to
the x-rays, whereas the carbon and nitrogen within the proteins have a mass-
absorption coefficient an order of magnitude greater. X-rays within this
so-called water-window can be produced by focusing a laser onto a carbon
target, producing a plasma that emits strongly at the c 5 + (33. 7 A) and c4+
(40.3 A) lines.
Some results of this work are shown in Fig. 10.12. The sample here is
a red blood cell, recorded with a 1 nsec x-ray exposure. The features ob-
served in such images are well below the resolution obtainable with optical
microscopes, and study of several botanical samples has revealed structure
not observable using traditional transmission electron microscopy methods.
Thus, LP x-rays, which can probe living tissue before cell death occurs, show
great potential in furthering our knowledge of the structure of living cells.
408 Hauer et al.
(a)
(b)
Figure 10.12 Replicas of in vivo red blood cell (a), and cells attacked by melittir
(b). (Results of work by M. Clague, University of Essex, U.K.)
We have described above the use of LP x-rays for materials study us-
ing diffraction and for the study of biological systems using contract mi-
croscopy. These two areas of research have been brought closer together by
the work of Frankel et al., who have studied x-ray diffraction from biological
samples. Some of the most interesting work has involved the study of the
purple membrane (PM) of the Halobacterium halobium. This membrane
consists of a hexagonally packed, two-dimensional array of the protein
bacteria-rhodopsin (BR) and is thus ideal for diffraction studies. Frankel
and Forsyth (1979) used the resonance line of He-like chlorine ( 4.44 A) to
create a quasimonochromatic beam to diffract off the PM sample at low an-
gles (the lattice constant of the membrane is at"' 63 A). The resultant pow-
der diffraction pattern showed multiple-order reflections, with the highest
order indicating a resolution of"' 7 A in the plane of the membrane.
The PM is a photobiological system: It has been shown that, upon illu-
mination, the BR protein sets up a transmembrane hydrogen ion gradient,
within which there may be molecular changes. Frankel and Forsyth used
the synchronization and single-shot ability of an LP x-ray source to probe
a PM sample 1 msec after photoactivation. They saw considerable alter-
ations in the resultant diffraction pattern, which they believe cannot simply
be attributed to heating of the lattice but are indicative of photoactivity.
(D) X-Ray Lithography
The last few years have seen an increased interest (mainly by industry) in the
utilization ofx-ray lithography for 1:1 shadowgraphy printing of a patterned
mask in semiconductor circuit manufacturing. This interest results from the
diffraction limitation on the resolution achievable with optical lithography,
as well as problems associated with the small depth of field. X-ray lithogra-
phy is the logical extension of the current optical lithography to submicron
printed circuit manufacturing. For optical systems having very good aberra-
tion correction and employing a high degree of partial coherence, the mini-
mum practical feature size can be as small as 0.6).../NA (where NA is the nu-
merical aperture of the optical system). The depth of focus of a diffraction-
limited system is about 0.8R 2 />. (R is the radius of the effective aperture).
Clearly, for submicron feature sizes the depth of field is not larger than typ-
ical feature heights, necessitating the use of complex multilevel photo resist
techniques. This expression also shows that shorter wavelengths produce a
better combination of resolution and depth of focus, but the realization of
a practical lithography system that utilizes short optical wavelengths is not
easy. A typical high-performance reduction lens might utilize as many as six
different glass types, each with a different index and dispersion to control
the aberrations. Controlling the aberrations in such a lens requires that the
position of each element in the assembly be carefully adjusted. Because the
410 Hauer et al.
index of refraction of glass is also a function of temperature, the entire lens

housing is generally temperature-controlled to a fraction of a degree.
X-ray lithography involves simple one-to-one shadow projection (Spiller
and Feder, 1977) of an x-ray-transparent membrane with absorber struc-
tures onto a resist-coated wafer. Between mask and wafer is a small gap
(typically 20 pm) to protect the mask against mechanical damage. The
resolution is then limited by the penumbral shadowing (except for a col-
limated beam) and also by diffraction. The wavelength region useful for
the purpose ofx-ray lithography is approximately 0.5 to 2.5 nm. The longer
wavelength limit is determined mainly by the high absorption in the thin
mask substrate; the shorter wavelength limit is determined mainly by the
decreasing absorption in the resist layer. A variety of x-ray sources have
been used in lithography development. X-ray tubes are beset by low fiu-
ence, requiring impracticably long exposure times. This is particularly so
because the relatively large focal spot requires a large working distance.
Synchrotron storage rings have the advantages of being powerful and col-
limated, but are exceedingly expensive and complex for industrial applica-
tion. Laser-irradiated targets have the advantages of a very small source
(hence, good resolution even at close distances) and the possibility of scal-
ing to a high repetition-rate industrial laser. Additional characteristics are
the reproducibility of source positioning (to a few pm), almost complete
freedom in choosing the target material (which affects the spectral char-
acteristics), and mechanical stability in single-pulse operation. The typ-
ical shot-to-shot intensity variation of the total dose is about ±10%, but
this can be greatly improved in mode-locked operation with feedback con-
trol.
X-ray lithography using a laser-produced plasma source has been de-
monstrated and studied in the past years (Nagel et al., 1978; Mallozzi et al.,
1979a). More recently, it has been shown (Yaakobi et al., 1983) that a single
1 nsec UV-laser pulse of 35 J energy can produce (shallow) exposure in PBS
and COP resists. The corresponding x-ray dose on the resist was smaller,
by more than a factor of 10, than the nominal threshold exposure of these
resists. The reason for this behavior is that the high-peak irradiance of x-
rays with these short UV-laser pulses is sufficient to induce resist ablation,
so that the pattern is imprinted even without resist development. The shal-
lowness of the ablation pattern can be improved by employing multilayer
resist techniques. An example of relief structures obtained (Frankel et al.,
1987) in single-layer resist by the Hampshire Instruments Company, using
laser-produced x-ray lithography, is shown in Fig. 10.13. The vertical resist
ridges ride over 0.5 pm high silicon dioxide steps on the silicon substrate.
The exposure was obtained with radiation in the range of 8 to 20 A using
a boron nitride mask substrate with a 0.5 pm gold absorber. Submicron
resolution and feature acuity are clearly evident.
Figure 10.13 Relief structure in AZ2400 resist obtained by laser-produced x-ray

lithography (Frankel et al., 1987). The resist ridges ride over 0.5 micron high silicon
dioxide steps on a silicon substrate. (Courtesy of Hampshire Instruments, Inc.)
412 Hauer et al.
10.2.3 Particle Applications
From the data presented in Table 10.3, one can see that a significant portion
of the incident laser energy can be converted into protons and moderately
heavy ions with energies greater than 1 MeV. This is enough energy to cause
proton activation of nuclei and to result in the production of radioactive
isotopes when a secondary target is bombarded.
With a moderate-size laser system producing around 1 kJ of energy in
1 nsec (e.g., one module of the HELlOS laser system at the Los Alamos
National Laboratory), it should be possible to produce 150 J of 1 MeV
protons. An upgrade in laser technology might make possible a 1 Hz fir-
ing rate; this, in turn, would result in a total average current of protons
of about 100 pA. This amperage is comparable to the current available
from cyclotrons. Many advances would be required to create a practical
alternative to cyclotron-produced isotopes. These advances would include
(1) collimation of the ion blowoff to be compatible with the small sample
size needed in isotope work (Obrien, 1987) and (2) protection of the target
from low-energy ion debris and thermal loading.
Another possible application for high-energy ion blowoff is plasma for-
mation in magnetic mirror machines (Mayer et al., 1985). The stability in
magnetic fusion devices is often discussed in relation to the parameter beta,
which is the ratio of plasma particle pressure to magnetic field pressure. The
plasma configuration in mirror machines is unstable at low beta but may be
stable at high beta. It is essential in these systems to have rapid plasma for-
mation and thus rapid achievement of high-beta conditions. One method
of producing the rapid plasma formation could be laser-produced fast ion
blowoff. Ion energies of the order of 100 keY, which can readily be pro-
duced by C02 lasers (as shown in Table 10.3), are required in this appli-
cation. Mayer et al. (1985) have proposed an experiment that would help
to determine the viability of this interesting new application of high-energy
ions.
Finally, ion jets, as typified by Fig. 10.5, may be of value in simulating and
studying astrophysical conditions (Forslund and Goldstone, 1985). The ion
jets are reminiscent of galactic jets that have been identified in radio-wave
photographs. In addition, it has been observed (Forslund and Goldstone,
1985) that there is a striking similarity between the normalized electromag-
netic spectrum for a typical C0 2 laser/target interaction and that emitted by
the Crab Nebula. C0 2 LP's have been observed to emit up to 0.1% of the in-
cident laser energy as microwaves. Such surprising emission efficiency may
be due to an enhancement in the expanding corona of those plasma waves
that couple efficiently to microwave emission. A mechanism of this type
is a strong candidate for explaining microwaves emitted from both galac-
tic jets and solar flares. A more detailed laboratory investigation of these
processes could help us discriminate between the various theories of astro-

physical microwave emission.
10.3 LASER-PLASMA
ACCELERATION OF PARTICLES
10.3.1 Lasers and Plasmas
The maximum accelerating field that can be produced with current accelera-
tor technology is of the order of 1 MeV cm- 1 . At field strengths greater than
this, one encounters the problem of breakdown of the accelerator walls. To
build the next generation of particle accelerators, one has two choices. The
first is to use conventional technology and simply increase the size of the
accelerator, with all the financial and logistical difficulties that this entails.
The second is to develop new accelerator technologies. Ideally, these would
have a larger accelerating field and would cost less per unit particle energy
than today's accelerators.
In this section, one of the proposed new accelerators, namely the plasma
beat-wave accelerator (Tajima and Dawson, 1979; Joshi et al., 1984), will
be described. The basic idea is to use two lasers to generate a longitudi-
nal space-charge wave in a plasma. The phase speed of this space-charge
wave must be close to the speed of light so that a relativistic particle will
stay in phase with the electrostatic field of the wave long enough to be ac-
celerated to high energy. Before describing this scheme in detail, however,
it is instructive to consider why it might be desirable to incorporate lasers
and plasmas in accelerator technology.
The main reason for using lasers and plasmas is the magnitude of the
electric fields that can be produced. Specifically, the peak electric field in a
laser is related to the laser irradiance by
so that an incident irradiance of 1016 W cm- 2 corresponds to an electric

field of approximately 2.7 GeV cm- 1 . This is larger than the maximum ac-
celerating field of conventional accelerators by more than three orders of
magnitude! Unfortunately, however, the transverse polarization of these
electric fields renders them unsuitable for direct particle acceleration. Any
scheme to couple particles to the laser fields necessarily involves transverse
particle acceleration and the resulting radiation losses. This suggests that
waves with longitudinal electric fields should be used. One obvious candi-
date is the longitudinal space-charge wave in a plasma, in which the electro-
static field is generated by periodic density compressions and rarefactions of
the plasma electrons. The natural angular frequency of this "plasma wave"
414 Hauer et al.
is known as the plasma frequency and has a value of
The maximum electrostatic field can be estimated from Poisson's equation

by setting the perturbed density equal to the background density and taking
the plasma wave vector as consistent with the requirement that the phase
speed of the plasma wave is close to the speed of light. The resulting elec-
trostatic field is given by
(10.5)
and is known as the cold wave-breaking field because it is the largest electro-
static field that can be produced before the field amplitude becomes double-
valued (Dawson, 1959). In a plasma with a background density of 1018 cm-3,
the cold wave-breaking field is approximately 0.97 GeV em -1, which is again
larger than the maximum accelerating field of conventional accelerators by
three orders of magnitude. This clearly illustrates one potential advantage
of using plasmas. Another advantage is that a plasma is already ionized and
so no further damage can be done to it by these large electric fields.
10.3.2 Plasma-Wave Generation
In the plasma beat-wave accelerator, the plasma wave is generated by the

beating of two colinear laser beams (Rosenbluth and Liu, 1972). The radi-
ation pressure of the laser induces longitudinal plasma oscillations at the
laser beat-frequency. If the incident frequencies are chosen so that the
plasma wave is resonantly driven, the conservation of energy and momen-
tum is manifested by the frequency and wave vector matching conditions
where the higher-frequency laser is denoted by the subscript 1, the lower-

frequency laser is denoted by the subscript 0, and the plasma wave is de-
noted by the subscript p. It follows that the phase speed of the plasma
wave can be expressed in terms of the incident frequencies and wave vec-
tors as (w 1 - w 0 )!(k 1 -k 0 ). In an underdense plasma, in which the plasma-
wave frequency and wave vector are much smaller than the incident fre-
quencies and wave vectors, this is approximately equal to the group speed
dw1/dk1 = c(1- w~/w[)112 of the light waves.
The plasma-wave amplitude evolves according to the equation
(10.6)
where time is measured in units of w; 1 . AP is the peak amplitude of the

density fluctuation associated with the plasma wave, normalized to the
background density. The corresponding electrostatic field is the product
of this normalized density fluctuation and the cold wave-breaking field
[Eq. (10.5)]. A 1 andA 0 are the peak "quiver" velocities of electrons in the
laser fields (divided by the speed of light) and are related to the incident
laser irradiances by
In the linear regime, the plasma wave grows indefinitely. Eventually, how-
ever, the quiver velocity of electrons in the plasma-wave field becomes so
large that the lowest-order relativistic corrections to the electron mass must
be retained in the equations of motion. This results in a nonlinear reduction
in the natural frequency of the plasma wave by an amount that is propor-
tional to the square of the wave amplitude (Akhiezer and Polovin, 1956).
As a result of this relativistic frequency shift, the plasma wave is driven out
of phase with the beating of the two light waves, and the plasma-wave am-
plitude saturates at the value
(10.7)
If the interaction is allowed to continue, the plasma-wave energy is fed back

into the light waves and the plasma-wave amplitude decreases accordingly.
In solving Eq. (10.6) for the saturated plasma-wave amplitude (10.7),
it was tacitly assumed that the incident laser amplitudes are constant. In
reality, the incident laser pulses have finite pulse length T1 and propagate
through the plasma at their group speed. Consider the interaction of the
light waves and the plasma wave at some arbitrary point ~0 in the plasma.
Initially, the plasma wave has only noise-level amplitude. At some time
T 0 , the leading edges of the laser pulses reach the position ~0 • The plasma
wave then starts to grow according to Eq. (10.6), with laser amplitudes that
now depend on time. This growth continues until the plasma-wave ampli-
tude reaches a value close to that predicted by Eq. (10. 7) or until the trailing
edges of the laser pulses pass by, whichever occurs first. For times later than
To+ T1, the plasma wave oscillates freely with the amplitude it had at time
To+ T1• To maximize the energy in the plasma wave, the laser pulse length
must be tailored to coincide with the maximum plasma-wave amplitude, as
shown in Fig. 10.14. This limits the energy in the incident laser pulses and,
hence, the energy that can ultimately be delivered to the accelerated par-
ticles in one stage. Because the group speed of a plasma wave in a typical
beat-wave plasma is essentially zero, the energy deposited in the plasma
416 Hauer et al.
TURBULENT WAKE / LASER PULSE
\
PLASMA WAVE
Figure 10.14 Maximizing of the energy in the plasma wave through tailoring of
laser pulsewidth.
wave is left behind the laser pulses, which continually propagate into fresh
plasma and continue the process anew.
The saturation time and saturated amplitude for a given pulse shape can
be estimated by using the linear plasma-wave amplitude to determine the
cumulative phase shift due to the plasma nonlinearity. Imposing the con-
dition that the cumulative phase shift equals Ir/2 radians determines the
saturation time and saturated amplitude as a function of the incident laser
intensities and pulse shapes. The incident pulse lengths can then be chosen
to coincide with the maximum plasma-wave amplitude, as discussed above.
For laser-beam amplitudes that grow linearly in time, the saturation time
is given by
Ta [w- 1
p ]~~
20(A 1A 0 )- 213 (10.8)
whereas the saturated amplitude is given by
(10.9)
In the above formulae,A 1 andA 0 denote the peak laser amplitudes.

This simple theory of plasma-wave generation is in good agreement
with the results of computer simulations (Katsouleas et al., 1985; Forslund
et al., 1985). Experimental verification was first obtained by Joshi et al.
(1985). The 9.6 ttm (A 1 ::::::: 0.030) and 10.6 ttm (A 0 ::::::: 0.015) lines of a C0 2
laser were used to resonantly drive a plasma wave in a plasma of density
1.1 x 1017 cm- 3 • The laser amplitudes were modeled as growing linearly in
time for a duration of 1 nsec. For these parameters, Eq. (10.6) predicts a
maximum plasma-wave amplitude AP of approximately 0.08 at the middle
of the laser pulse. By measuring the time-integrated scattered light, an av-
erage amplitude AP of 0.01 to 0.03 was inferred, in reasonable agreement
with the theoretical estimate. The corresponding electrostatic field was 3
to 10 MeV cm-1, which marked the first time a longitudinal field in excess
of 1 MeV cm- 1 had been produced, in a controlled manner.
Behind the laser pulses, the wake of the plasma wave becomes turbu-
lent, as shown in Fig. 10.14. This can be due to the parametric decay insta-
bility, in which the plasma wave decays into a secondary plasma wave and
an ion-acoustic wave (Dubois and Goldman, 1965 and 1967; Silin, 1965) or
the modulational instability (Whitham, 1974; Pesme et al., 1987). These
instabilities both involve ion motion and occur on a time scale longer than
the electron time scale wp- 1 by a factor of (mJZme) 112 , where Z is the ionic
charge. The modulational instability can also occur because of the relativis-
tic nonlinearity in Eq. (10.6). Because the turbulent wake cannot be used
for particle acceleration, it might be inferred that the energy in the wake
of the plasma wave is wasted. This is not necessarily true. In practice, the
presence of these plasma instabilities tends to limit the time for plasma-
wave growth and beam loading (McKinstrie and Forslund, 1987). This. is
discussed in detail in a later section.
In the preceding analysis, the self-consistent evolution of the light-wave
amplitudes was not taken into account. Just as the beating of the incident
light waves produces a resonant plasma wave at the difference frequency,
the beating of the transverse electron quiver-velocities with the plasma-
wave density fluctuation produces oscillating currents at the sum and dif-
ference frequencies. In this way, a spectrum of colinear light waves is gen-
erated, with frequencies and wave vectors that differ from those of the inci-
dent waves by integral multiples of wp and ~' respectively. This nonlinear
interaction can also be viewed as a series of three-wave processes in which
a photon either decays into a lower-frequency photon and a plasmon or re-
combines with a plasmon to produce a higher-frequency photon. Notice
that the total number of photons is conserved and the number of plasmons
is equal to the difference between the number of decay interactions and the
number of recombination interactions. It follows that only a fraction wp/w 1
of the incident laser energy can be transferred to the plasma wave in the
primary three-wave interaction (Manley and Rowe, 1956). To increase the
energy transfer to the plasma wave and ultimately the accelerated particles,
the laser energy must be made to cascade "downward" from the incident
waves to their lower-frequency sidebands. Significant contributions to the
current understanding of this process have been made by Cohen et al. (1972)
and Karttunen and Salomaa (1987). However, the extent to which the inci-
dent energy can be made to cascade to lower frequencies remains an active
and important area of current research. The reader is referred to the review
by McKinstrie and Batha (1987). If one assumes that most of the laser en-
ergy can be transferred to the plasma wave and the rate of energy transfer
is constant, then the pump-depletion length Ld can be estimated from the
requirement that the energy contained in the wake of the plasma is equal
418 Hauer et al.
to the total energy that was originally contained in the laser pulses. Taking
the plasma-wave amplitude to be given by Eq. (10.9) and the laser pulse
lengths to be given by Eq. (10.8) yields
(10.10)
for the case of equal-irradiance lasers. The Lorentz factor'"'' associated

with the phase speed of the plasma wave, is equal to w11wr
10.3.3 Particle Injection
Just as a surfer must paddle to catch an ocean wave, an electron cannot

be trapped and then accelerated by a plasma wave unless it has a certain
minimum velocity parallel to the plasma wave. The corresponding mini-
mum energy is determined by the requirement that in the wave frame, the
electron has insufficient kinetic energy to escape the plasma-wave poten-
tial well. The potential energy of an electron in the plasma-wave field is
shown in Fig. 10.15. Because of the radial fields arising from the finite
width of the plasma wave, only the quarter-wavelength between point 1
and point 2 is both accelerating and focusing to a bunch of electrons (Kat-
souleas, 1986a). Consequently, an electron injected at point 1 must reach
its turning point before it reaches point 2 if it is to be usefully accelerated. In
(\1
()
Q)
E
......
:a,. 0 F-----+-------:-lt--------1--------.- x'
~ 2 ~
-a \ 1,3
defocusing focusing
Figure 10.15 Potential energy of an electron in the plasma-wave field.
terms of laboratory-frame quantities, the minimum injection energy (Kat-

souleas et al., 1985) is related to the corresponding Lorentz factor by
(10.11)
where
is the peak potential energy of the electron in the wave frame, normalized
to the electron rest-mass. This is also the average injection energy for elec-
trons that are injected at random into the field of the plasma wave. The
electron injection energy is plotted as a function of the plasma-wave am-
plitude in Fig. 10.16. For a typical plasma temperature of 200 eV, there is
negligible trapping of electrons, which must therefore be externally acceler-
ated before being injected into the field of the plasma wave. The injection
energy for ions can also be obtained from Eq. (10.11), by multiplying o: by
Zmelmi. Since the resulting value of o: is much less than unity, li must then
be approximately equal to 1¢· This effectively precludes ion acceleration.
15
10
YG> = 100
5
0 ~-L~_L~_L~_L~LLJ-~-LLJ-L~~
0 0.1 0.2 0.3 0.4 0.5
Figure 10.16 Electron injection energy as a function of plasma wave amplitude.

420 Hauer et al.
On reaching point 2, an electron reverses its wave-frame motion and is

further accelerated until it reaches point 3. The final electron energy (Yan,
1986) is given by
(10.12)
From Eqs. (10.11) and (10.12), the energy gain is easily seen to be
(10.13)
The gain in electron energy is plotted as a function of the plasma-wave

amplitude in Fig. 10.17. For the high-gain limit in which a is much greater
than unity, f}.-y is approximately equal to 2-y~r This high-gain limit can
also be derived by a simple physical argument. As an electron moves from
point 1 to point 2, its laboratory-frame energy is increased from "ti to"!¢·
This increase is much smaller than the total increase, which is of the order
of -y~, and can therefore be neglected. As the electron moves from point 2 to
point 3, the large electric field necessary for high gain rapidly accelerates the
electron to relativistic speeds in the wave frame. Consequently, the electron
100 LLJ_~_L~_L~-L~~-L~-L~~~~~
0 0.1 0.2 0.3 0.4 0.5
Figure 10.17 Gain in electron energy as a function of plasma-wave amplitude.

spends an equal amount of time at each point of the wave. This work done
on the electron in the laboratory frame is therefore equal toe (Ep)La, where
(Ep) = 2EP/7r is the average of the electric field taken over one-quarter
wavelength and the acceleration length La is the distance traveled by the
electron in the laboratory frame as it traverses one-quarter wavelength in
the wave frame. If we approximate the speed of the accelerated electron by
the speed of light, the acceleration length is,\{'/[ 4(1- ;1rl>)] ~ ~~>..P/2. Rewrit-
ing the electric field in terms ofAP gives e(Ep)La ~ 2'Y~pmec 2 , in complete
agreement with the high-gain limit of Eq. (10.13).
For easy comparison with the pump-depletion length [Eq. (10.10)], the
acceleration length can also be written as
(10.14)
It follows that the ratio La!Ld is approximately equal to 0.67 Af3 • This ratio
is an upper bound on the LP coupling efficiency "'c• defined as the fraction
of incident laser energy that is transferred to the plasma wave during the
acceleration phase.
Ebrahim et al. (1986) were the first to observe the acceleration of in-
jected electrons. In their experiment, the 9.6 11m (A 1 ~ 0.054) and 10.6 11m
(A 0 ~ 0.060) lines of a C02 laser were used to resonantly drive a plasma
wave in a plasma density 1.1 x 1017 cm- 1 • For these parameters, theory pre-
dicts a maximum plasma-wave amplitudeAP of approximately 0.16 and a
corresponding maximum electrostatic field of approximately 50 MeV cm- 1 .
The electrons were produced by irradiating an aluminum slab with an aux-
iliary high-intensity C0 2 laser. The energy of the electrons obtained in this
manner was 0.5 to 1.0 MeV, which is greater than the minimum injection
energy [Eq. (10.11)] of 0.46 MeV. Measurements indicated that electrons
injected at 0.6 MeV were accelerated to 2.0 MeV. Because the resonant
plasma region was only 0.15 em long, this implied that the average electro-
static field was 10 MeV cm- 1 , in reasonable agreement with the theoretical
estimate.
10.3.4 Beam Loading
In calculating the energy gain for a single electron, the self-generated elec-
trostatic field can safely be neglected. However, in determining the en-
ergy gain of a bunch of electrons, the self-generated field Eb is crucially
important. The standard approach to this problem is to determine the elec-
trostatic field produced by a single electron as it propagates through the
plasma. The electrostatic field produced by a bunch of electrons can then
be obtained by linear superposition.
422 Hauer et al.
Because the group speed of a plasma wave in a typical beat-wave plasma

is essentially zero, there can be no electrostatic field ahead of the electron,
regardless of whether or not it is propagating at a relativistic speed. Behind
the electron, a sinusoidal electrostatic field is produced, with a phase speed
that is equal to the speed of the electron.
Providing the amplitude of the plasma wave is not too close to the cold
wave-breaking limit, the amplitudes of the nonlinear harmonics are much
less than the amplitude of the fundamental and the plasma wave is essen-
tially sinusoidal. If an infinitesimally short bunch of electrons is injected
into the plasma-wave field before the plasma wave becomes turbulent, the
electrostatic field of the plasma wave can be completely cancelled. The laser
energy that would have gone into the wake of the plasma wave is now ab-
sorbed by the electron bunch. By equating the relevant electrostatic fields,
the number of accelerated electrons per unit cross-sectional area of the
plasma wave, denoted by a 0 , is easily found to be
(10.15)
For a plasma of background density 1018 cm- 3 and a plasma-wave amplitude

AP of, e.g., 0.2, this is approximately 1014 electrons per cm2 •
The beam-loading efficiency 1Jb is defined as the plasma-wave energy ab-
sorbed by the electron bunch divided by the plasma-wave energy in the ab-
sence of the electron bunch. The case of an infinitesimally short bunch de-
scribed above corresponds to a beam-loading efficiency of 100%. Unfortu-
nately, however, this is only achieved at the expense of a 100% spread in the
energy of the accelerated electrons. This is because the first electron in the
bunch feels the full accelerating field EP, whereas the last electron in the
bunch feels the superposed field EP + Eb, which is equal to zero. Because
the reduction in accelerating field is linear in the areal number density a of
the bunch, the fractional energy spread is simply equal to a/a0 • The frac-
tion of plasma-wave energy absorbed by the bunch is quadratic in the areal
number density of the bunch and is readily shown to be
(10.16)
Equation (10.16) illustrates the tradeoff between the total number of accel-
erated particles and the fractional energy spread for infinitesimally short
bunches.
In principle, it is possible to reduce the energy spread without signifi-
cantly lowering the beam-loading efficiency by appropriately shaping the
electron bunch (Katsouleas et al., 1987). This is an active area of current re-
search. However, it is not clear whether such precise shaping of the electron
the electron bunch, over distances that are much smaller than the plasma
wavelength, is technologically feasible. A realistic electron bunch has a den-
sity distribution that is approximately Gaussian. Calculations show that, for
particles within one standard deviation of the peak density, the accelerat-
ing field is approximately equal to one-third of the plasma-wave field and
varies by approximately 10%. The corresponding energy gain is therefore
about one-third of the ideal value predicted by Eq. (10.13). It can also be
shown that the number of accelerated electrons is about three-quarters of
the number suggested by Eq. (10.15).
10.3.5 Laser Requirements
The previous sections have shown how lasers can be used to generate a
large-amplitude plasma wave with a phase speed that is close to the speed
of light. The electrostatic field of this plasma wave can be of the order of
1 GeV cm- 1 . Equation (10.13) shows that, in the high-gain limit, the en-
ergy gain of a suitably injected electron is proportional to (w1/wp) 2 • A high
energy gain requires this ratio to be large. For a fixed energy gain, the ac-
celeration length [Eq. (10.14)] is proportional to cwp- 1 • This length can be
minimized by the use of high-density plasmas, which, in turn, requires the
use of short-wavelength lasers. A particular example is a KrF laser (Jensen,
1987), which oscillates at 0.25 f-lm with an inherent bandwidth of about 1%.
The design specifications for a plasma beat-wave accelerator, based on a
KrF amplifier, are given in Table 10.4. The electron energy gain is approx-
imately 1.0 GeV per stage in a distance of approximately 12 em. Thus, by
use of multiple staging, it is theoretically possible to accelerate electrons
to an energy of 10 TeV in a distance of the order of 1 km. This distance is
Table 10.4 Case Study for a

KrFLaser
>.I [!-lm] 2·5 x w- 1

t.>.l [1-lm] 2·5x1o-3
11 [Wcm- 2 ] 1·0 X 1017
r 1 [psec] 9·8
Spot size [~-tm] 25
Laser energy [J] 1·6
n. [cm- 3 ] 1·8x1018
eEP [MeV cm- 1] 360
t.T. [MeV] 1·0x103
La [em] 12
424 Hauer et al.
two orders of magnitude less than that for the proposed superconducting
supercollider (SSC).
There are still many questions to be answered before the plasma beat-
wave accelerator can be considered a practical device. First and foremost
are the open questions of laser efficiency and LP coupling efficiency. From
Table 10.4 and the preceding analysis, any reasonable estimate of laser effi-
ciency leads to a wall-plug efficiency of only a few percent. However, even if
lasers prove to be too inefficient for the next generation of accelerators, they
are an important intermediate step because of their demonstrated ability to
produce large-amplitude, monochromatic plasma waves.
In this brief review, only the basic, one-dimensional theory of the plasma
beat-wave accelerator has been described. For a more detailed account of
this and other plasma-based acceleration schemes, the interested reader is
referred to the articles by Katsouleas (1986b), Joshi et al. (1987), and Mori
(1987).
10.4 LASER-PULSED POWER SWITCHING
Recent progress in understanding the operation and performance of laser-

triggered spark gaps has been made possible by the availability of reliable
pulsed lasers with wavelengths covering the range from the ultraviolet (UV)
to the infrared (IR) (Guenther and Bettis, 1985). At the same time, fun-
damental studies of the interaction of laser radiation with dielectric gases
have been improved through the use of high-bandwidth recorders, such as
streak cameras (Dougal and Williams, 1984), coupled with new optical and
electrical sensors (Chang et al., 1985). The new impetus for developing
laser-triggered spark gaps is based on the need to synchronize numerous
high-voltage events to within several nanoseconds. This need is exempli-
fied by the PBFA-11 light-ion-beam fusion accelerator located at Sandia
National Laboratories in Albuquerque, New Mexico (Turman et al., 1985)
(Fig. 10.18). The PBFA-11 has 36, 5 MV gas dielectric switches that are
routinely triggered with a total timing spread of 10 nsec.
Because of its importance in low-jitter high-voltage triggering, we will re-
strict our attention to UV-laser-triggered SF6 -insulated spark gaps. When
appropriate, comparisons will be made to the performance of similar IR-
triggered switches. The reader is referred to review articles by Guenther
and Bettis (1978, 1985) for in-depth information on other types of laser-
triggered spark gaps.
Laser-triggered switch systems should have the following desirable char-
acteristics:
1. Low jitter.
Figure 10.18 PBFA-II light-ion beam fusion accelerator at Sandia National Lab-
oratories.
2. Minimum sensitivity of switch closure time, i.e., delay, to switch voltage and
laser energy.
3. Low switch prefire rate.
4. Low required laser energy.
5. Low maintenance.
6. High reliability.
7. Efficient laser, beam transport, and optical alignment systems.
Low jitter triggering with minimum change in switch delay with switch volt-
age has been demonstrated with UV lasers in SF6 -insulated switches on nu-
merous occasions (Rapoport et al., 1980; Woodworth et al., 1982; Hamil
and Smith, 1983; Woodworth et al., 1984; Humphreys et al., 1985). The
need for low jitter is obvious when numerous high-voltage switching oper-
ations must be synchronized or the switch needs to be synchronized with
another fast event. The change in the gap closure time with switch voltage
gives the necessary uniformity of switch-to-switch charge voltage for mul-
tiple gas switches to achieve low jitter. The prefire rate is reduced when
the switch is triggered well below its self-breakdown voltage. Switch main-
tenance is also reduced because there is no trigger electrode, typically the
426 Hauer et al.
first component to fail in an electrically triggered switch. However, the op-

tics in a laser-triggered system must be kept free of the debris and caustic
gases produced within the switch. Experience with numerous high-voltage
switches has shown that this requires a proper purge-gas system and an opti-
cal design with the final optical element placed at least 50 em from the laser
breakdown spark (Hamil et al., 1985a and 1985b). The laser, delivery optics,
and alignment system must be economical, simple, reliable, and easy to use.
Rapoport et al. (1980) first described the use of ultraviolet laser radia-
tion from a krypton fluoride (KrF) excimer laser to trigger an SF6-insulated
switch with subnanosecond jitter. In this experiment, SF6 gas was used
because of its low UV-breakdown threshold and its high de-breakdown
strength. A 7 mJ, 20 nsec long pulse at a wavelength of 248 nm from the
KrF laser was focused into the center of a 0.5 em, 80 kV, de-charged, high-
voltage gap. The laser was focused through a 0.25 em diameter hole in the
high-voltage electrode. Subnanosecond jitter at 80% of the self-breakdown
voltage was measured. Empirical studies showed that switch jitter was re-
duced when the incoming laser beam was very close to grazing the entrance
hole in the high-voltage electrode. This effect may be associated with the
particular electrode geometry and the diameter of the laser entrance hole
relative to the gap spacing. In the geometry of Rapoport et al. (1980), the
entrance hole diameter was equal to half the gap spacing. The unillumi-
nated distance from the laser beam to the edge of the hole could therefore
have a significant effect on the measured switch jitter. Unpublished experi-
ments (Hamil and Ramirez, 1987) performed at Sandia National Laborato-
ries with two identical 2.5 MY switches of the Rimfire design (Humphreys
et al., 1985) demonstrated that entrance hole effects can be minimized. In
the Sandia experiments, the nanosecond rms trigger jitter between the two
switches was not affected when one of the laser-trigger beams was moved
from the center to the edge of the entrance hole. In these Rimfire experi-
ments, the ratio of the electrode spacing to the entrance hole diameter was
6, as compared with 2 in the experiments of Rapoport et al. (1980). The
Sandia results illustrate the importance of the gap spacing to entrance hole
diameter ratio in designing laser-triggered switch systems.
Prior to the work of Rapoport et al. (1980), laser triggering experiments
using IR or visible-wavelength lasers had focused the laser near the far elec-
trode in order to generate an electrode-based plasma (Guenther and Bet-
tis, 1978). It has been suggested (Horii et al., 1970; Guenther and Bettis,
1978) that UV radiation from the electrode plasma serves as the source of
the channel that eventually closes the gap. Dougal and Williams (1984 and
1986) showed that the channel that eventually closes the gap preferentially
follows the path of the laser beam. However, subnanosecond jitter is only
achieved with IR and visible-wavelength lasers at pulse energies at least an
order of magnitude greater than those required for UV-laser triggering.
The UV-laser work at Sandia National Laboratories has been aimed at

developing pulse-charged megavolt-class switches that can be triggered with
nanosecond jitter using minimum laser energy. In that effort, a 5 MV mul-
tistage switch was triggered reliably with 10 mJ of laser energy to achieve
1.5 nsec jitter (Humphreys et al., 1985), a 2.5 MV multistage switch was trig-
gered with as little as 5 mJ of laser energy and 1 nsec jitter (Hamil, 1987),
and a 3 MV single-gap switch was triggered with 20 mJ of laser energy to
achieve 2 nsec jitter (Hamil and Smith, 1983).
Several authors (Guenther and Bettis, 1978; Woodworth et al., 1982;
Woodworth et al., 1984; Guenther and Bettis, 1985; Humphreys et al, 1985)
have observed the effect of laser pulse length on switch delay and jitter.
With an IR-laser trigger source, the pulse-length dependency is quite dis-
tinct. The switch delay and jitter become quite large as soon as the switch
delay exceeds the laser pulse length. The UV-laser-trigger data do not
always show this very abrupt increase in jitter and delay when the laser
pulse terminates before switch closure (Woodworth et al., 1984). Recent
UV-laser trigger experiments (Hamil and Smith, 1983) performed at 3 MV
across an 11 em gap showed that laser pulse length effects can be minimized.
These unpublished experiments achieved lower jitter and delay with a short
2 nsec long pulse than with a 20 nsec long pulse of equal energy. Additional
factors, such as laser beam quality, laser pulse shape, spark formation time,
switch geometry, and f-number of the optical system, appear to be impor-
tant in determining the performance of UV-laser-triggered spark gaps.
Dielectric breakdown of the SF6 gas by the incident laser pulse is believed
to influence the performance of IR- and UV-laser-triggered high-voltage
switches. Evidence for this comes from measurements of the temporal de-
velopment of the laser breakdown spark in a flowing gas cell containing SF6
(Hargis, 1983). The incident laser pulse, transmitted laser pulse, and optical
sidelight emitted by the breakdown spark were measured with a temporal
resolution of about 0.5 nsec using fast photodiodes coupled to transient dig-
itizers. A representative single-pulse result for the dielectric breakdown of
lOOOTorr SF6 by a 30 mJlaser pulse at 266nmis shown in Fig. 10.19. Deple-
tion of the incident laser pulse (determined by subtracting the transmitted
and incident laser pulse shapes) and optical emission from the breakdown
spark occur at the same time. The time delay between the leading edge of
the incident laser pulse and leading edge of the breakdown sidelight pulse
increases as the incident laser pulse energy decreases. The total laser en-
ergy required to initiate dielectric breakdown of the SF6 was determined
by integrating the incident laser pulse shape up to the time corresponding
to the onset of the breakdown sidelight pulse. Figure 10.20 shows that the
laser energy required to initiate dielectric breakdown is independent of the
time delay. This result suggests the importance oflaser fluence in determin-
ing the onset of dielectric breakdown in SF6 • Similar data for the onset of
428 Hauer et al.
1.2
1.0 1000 torr SF6 SPARK

.....J 30 mJ/PULSE ~EMISSION
<
z 0.8 ~ SIDELIGHT
(.!)
266-nm SIDELIGHT (A.>300nm)
en 0.6
Q
0.4
w
N
TRANSMITTED~
:::::i 0.2
< 0
~
c:
0 -0.2
z LASER PULSE
-0.4
-0.6
0 4 6 12 16 20
TIME (ns)
Figure 10.19 Temporal development of 266-nm laser spark emission.
--
6
...., 5
E INTENSITY
4 1000 torr SF6 DEPENDENT
~
<(
BREAKDOWN
w (E 8 /T8 • CONSTANT)
. .·...
a:
m
L.U
'.. \
• •• • \ FLUENCE DEPENDENT
BREAKDOWN (Es CONSTANT)
O L _ _ J_ __ L_ _ j __ _L_~L_~_ __ L_ __ L_ _ j __ _
0 2 3 4 5 6 7 8 9 10
TBREAK (ns)
Figure 10.20 Fluence dependence of 266-nm breakdown in SF6 .

dielectric breakdown in SF6 by an IR laser at 1064 nm show that the spark

initiated by theIR radiation develops on a much slower time scale and per-
sists for a longer time than the spark initiated by the UV laser at 266 nm
(Hargis, 1983). A fixed amount of laser energy is also required to initiate
dielectric breakdown with theIR-laser pulse. Because the breakdown spark
initiated by the 1064 nm laser pulse develops on a much slower time scale,
the time at which dielectric breakdown occurs is more susceptible to fluctu-
ations in the laser pulse shape, resulting in increased jitter in the formation
of the laser spark.
Optical emission from the laser-initiated breakdown spark in SF6 pro-
vides information on species playing a role in the laser-induced breakdown
process. This is illustrated in Fig. 10.21 by the time-resolved emission spec-
tra shown of the breakdown spark in 1000 Torr SF6 initiated by a 30 mJ
laser pulse at 266 nm. The SF6 was flowing in a gas cell at a rate sufficient
to change the gas in the breakdown volume between each laser pulse. The
spectra in Fig. 10.21 were recorded on a scanning monochromator and with
a time resolution of 20 nsec. Zero delay corresponds to the incidence of the
266 nm laser pulse in the breakdown spark volume. The most striking fea-
tures of these data are the line-broadened atomic fluorine emission peaks at
wavelengths between 650 and 750 nm. Decreased broadening of the atomic
fluorine lines at late times after the initiation of the breakdown spark is
readily apparent. The spectral resolution in Fig. 10.21 is not sufficient tore-
solve the individual contributions from Stark, pressure, and Doppler broad-
ening effects. Higher-resolution data should make it possible to measure
the temporal dependence of the Stark width of the broadened lines. Re-
cently measured Stark broadening parameters (Mar et a!., 1986) can then
be used to estimate the average electron density in the breakdown spark
volume.
250 375 500 625 750

WAVELENGTH (nm)
Figure 10.21 Laser breakdown spectra in SF6 .
430 Hauer et al.
It is instructive at this point to discuss the dielectric breakdown of SF6

and the related optical sidelight when the breakdown spark is initiated by
two laser pulses at different wavelengths. The time scale for the formation
of the breakdown spark in SF6 is faster in the UV at 266 nm than in theIR at
1064 nm (Hargis, 1983). If both the UV and IR pulses of light are focused
together in a gas cell containing SF6 , the seed electrons produced by the
UV pulse can be heated by the IR pulse through an inverse bremsstrahlung
mechanism (Zel'dovich and Razier, 1965; Morgan, 1975). Figures 10.22(a)
and 10.22(b) clearly show an enhanced emission at wavelengths betwee!}.
300 and 700 nm from the breakdown spark volume when the UV and IR
pulses of light overlap each other temporally and spatially. The enhanced
emission is attributed to IR heating of the seed electrons created by the
UV pulse and/or an increase in the average electron density by the addi-
tion of the IRradiation. Unlike the early results of Evseenko et al. (1979),
which showed a reduced threshold for gas breakdown in a two-frequency
laser field at wavelengths of 1064 and 532 nm, the breakdown threshold at
266 nm is not affected by the presence of the 1064 nm radiation. The very
bright spark created by the addition of the IR radiation was expected to be
far superior to the faint spark produced by the relatively low-energy UV
pulse when used to trigger a high-voltage gap. Figure 10.23 shows the char-
acteristic breakdown curve for multifrequency triggering vs. pure UV-laser
triggering of a 0.5 MV gap. Nate that there is no difference in jitter or delay
even though theIR-laser pulse energy is more than ten times the UV-laser
pulse energy. At this writing, we still do not fully understand the under-
lying physics involved in this phenomenon. The UV pulse alone seems to
determine the performance of the switch.
The results presented in this brief section illustrate the progress made
in understanding the operation and performance of IR- and UV-laser-
266 nm+1064 nm SPARKS OVERLAPPED

j SPARKS
OVER LAPPED
TRANSVERSELY
SPARKS
TRANVERSELY
OVERLAPPED AXIALLY
SPARKS
DISPLACED
AXIALLY
25 50 25 50
TIME (ns) TIME (ns)
Figure 10.22 Multifrequency laser breakdown in SF6 .
0 6 mJ at 266-nm
70 mJ at 1064-nm
o 5.5 mJ at 266-nm
>-
<( 12
_J
w []
0 10 []
w
> 8
0
f-
<(
_J
w 6
a:
4
0
40 60 80 100
PERCENT OF 500 eV SBV

Figure 10.23 Multifrequency versus UV triggering of a 50 kV switch.
triggered spark gaps. New insights provided by fundamental studies of di-

electric breakdown should lead to improved switch designs optimized for
laser triggering. This goal, however, requires the development of compre-
hensive models incorporating multi photon effects in order to more fully un-
derstand the complex phenomena involved in the dielectric breakdown of
SF6 •
ACKNOWLEDGMENTS
The authors would like to thank B. Yaakobi for supplying material for the
section on x-ray lithography and P. D. Goldstone, T. Katsouleas, W. B. Mori,
andY. T. Yan for useful discussions related to laser-particle acceleration.
We would also like to acknowledge the careful editing and proofreading of
432 Hauer et al.
B. Thompson, P. Atencio, K. Serna, and J. Weber and the excellent artwork

of S. Carlson and C. Boone.
REFERENCES
Akhiezer, A I. and Polovin, R. V. (1956). Theory of wave motion of an electron

plasma, Sov. Phys. JETP, 30: 696.
Anderholm, N.C. (1970). Laser-generated stress waves,AppL Phys. Lett., 16: 113.
Bingham, R. and Evans, R. G. (1988). Physics of beat-wave accelerators, Plasma
Phys. and Controlled Fusion (to be published).
Birkhoff, G. D. (1958). The passage of fast electrons through matter, in Handbuch
der Physik XXXIV (S. Flugge, ed.), p. 53.
Bradley, D. K. (1985). Ph.D Thesis, Imperial College of Science and Technology,
Univ. of London (unpublished).
Burgeat, J. and Taupin (1968).Acta Crystallogr.,A 24: 99.
Burnett, N.H. Enright, G. D., Avery, A, Loen, A, and Kieffer, J. C. (1984). Time
resolved K spectra in high intensity laser target interaction, Phys. Rev. ,A29: 2294.
Chang, J., Allen, G., Filter, W F., Lockwood, G. J., and Neyer, B. T. (1985). Photonic
diagnostics and pulsed power research, Digest of Technical Papers, Proc. ofthe 5th
IEEE Pulsed Power Conf, Arlington, Va., June 10-12, 1985 (M. F. Rose and P. J.
Turchi, eds.), IEEE, New York, p.l.
Cohen, B. I., Kaufman, A N., and Watson, K. M. (1972). Beat heating of a plasma,
Dawson, J. M. (1959). Nonlinear electron oscillations in a cold plasma, Phys. Rev.,
113: 383.
Dougal, R. A and Williams, P. F. (1984). Fundamental processes in laser-triggered
electrical breakdown of gases, J. Phys. D: AppL Phys., 17: 903.
Dougal, R. A and Williams, P. F. (1986). Fundamental processes in laser-triggered
electrical breakdown of gases: Unconventional geometries, J. AppL Phys., 60:
4240.
DuBois, D. F. and Goldman, M. V. (1965). Radiation-induced instability of electron
plasma oscillations, Phys. Rev. Lett., 14: 544.
DuBois, D. F. and Goldman, M. V. (1967). Parametrically-excited plasma fluctua-
tions, Phys. Rev., 164: 207.
Eason, R. W, Bradley, D. K., Kilkenny, J.D., and Geaves, G. N. (1984). Improved
laser-EXAFS studies of aluminum foil, J. Phys. C, 17: 5067.
Eason, R. W, Cheng, P. C., Feder, R., Michette, A G., Rosser, R. J., O'Neill, F.,
Owadano, Y., Rumsby, P. T., Shaw, M. J., and Turcu, I. C. E. (1986). Laser x-ray
microscopy, Opt. Acta, 33: 501.
Ebrahim, N. A, Martin, F., Bordeur, P., Heighway, E. A, Matte, J. P., Pepin, H.,
and Lavigne, P. (1986). First observations of acceleration of injected electrons in
a laser-plasma beat-wave experiment, Proc. of the 1986 Linear Accelerator Conf,
Stanford Linear Accelerator Ctr., Stanford, Calif., June 2-6.
Evseenko, V. P., Mitsuk, V. E., and Soldatova, I. V. (1979). Reduced threshold for
optical gas breakdown in a two-frequency field, Sov. Tech. Phys. Lett., 5: 329.
Fairand, B. P., Wilcox, B. A, and Gallagher, W. J., and Williams, D. N. (1972). Laser
shock-induced microstructural and mechanical property changes in 7075 alu-
minum, J. Appl. Phys., 43: 3893.
Feldman, U., Doschek, G. A, Nagel, D. J., Cowan, R. D., and Whitlock, R. R.
(1974). Satellite line spectra from laser-produced plasmas,Astrophys. J., 192: 213.
Forslund, D. W. and Brackbill, J. (1982). Magnetic-field-induced surface transport
on laser-irradiated foils, Phys. Rev. Lett., 48: 1614.
Forslund, D. W. and Goldstone, P. D. (1985). Photon impact, high energy plasma
physics with C0 2 lasers, Los Alamos Science, 12: 2.
Forslund, D. W., Kindel, J. M., Mori, W. B., Joshi, C., and Dawson, J. M. (1985). Two-
dimensional simulations of single-frequency and beat-wave laser-plasma heating,
Forster, E., Goetz, K., Schafer, K., and Zimmer, W. D. (1984). Laser generated
plasma as a source for real time studies in x-ray crystal research, Lasers and Par-
ticle Beams, 2: 167, 187.
Frankel, R. D., Drumheller, J.P., Kaplan, A S., and Lubin, M. J. (1987). Proc. of
Microelectronics Seminar Interface 86, Eastman Kodak, p. 82.
Frankel, R. D. and Forsyth, J. M. (1979). Nanosecond x-ray diffraction from biolog-
ical samples with a laser-produced plasma source, Science, 204: 622.
Graham, R. A, Nelson, F. W., and Benedick, W. B. (1965). Piezoelectric current from
shock-loaded quartz-A submicrosecond stress gauge, J. Appl. Phys., 36: 1775.
Guenther, A H. and Bettis, J. R. (1978). Laser triggering of high voltage switches,
J. Phys. D., 11: 1577.
Guenther, A H. and Bettis, J. R. (1985). Recent advances in optically controlled
discharges, Digest of Technical Papers, Proc. of the 5th IEEE Pulsed Power Conf,
Arlington, Va., June 10-12, 1985. (M. E. Rose and R. J. Turchi, eds.), IEEE, New
York, p. 45.
Hamil, R. A (1987). Private communication.
Hamil, R. A and Smith, D. L. (1983). Laser triggered switch results from a frequency
quadrupled Nd: YAG, Laser, Proc. ofthe 4th IEEE Pulsed Power Conf (M. F. Rose
and T. H. Martin, eds.), IEEE, New York, p. 447.
Hamil, R. A and Ramirez, J. J. (1987). Private communication.
Hamil, R. A, Seamons, L. 0., Schanwald, L. P., and Gerber, R. A (1985a). Optical
system for triggering PBFA II, Digest of Technical Papers, Proc. of the 5th IEEE
Pulsed Power Conf, Arlington, Va., June 10-12, 1985 (M. F. Rose and P. J. Turchi,
eds.), IEEE, New York, p. 803.
Hamil, R. A, Seamons, L. 0., Gerber, R. A, Adams, R. C., Woodworth, J. R., Black-
burn, R. S., and Sweatt, W. C. (1985b ). Laser-trigger system design for PBFA II,
Digest of Technical Papers, Proc. of the 5th IEEE Pulsed Power Conf , Arlington,
Va., June 10-12, 1985 (M. F. Rose and P. J. Turchi, eds.), IEEE, New York, p. 807.
Hargis, P. J. (1983). Spectral and temporal evolution of laser breakdown in SF6 ,
Bull. Amer. Phys. Soc., 28: 186.
Hauer, A (1976). Interaction of x-rays and acoustic waves in solids, Ph.D. Thesis,
University of Rochester, published by University Microfilms, Ann Arbor, Mich.
Hauer, A, Goldman, R., Krista!, R., Yates, M. A., Mueller, M., Begay, F., Van-
Hulsteyn, D., Mitchell, K., Kephart, J., Oona, H., Stover, E., Brackbill, J., and
Forslund, D. (1984). Supra thermal electron generation transport and deposition
434 Hauer et al.
in C0 2 laser irradiated targets, in Laser Interaction and Related Plasma Phenom-

ena, Vol. 6 (H. Hora and G. H. Miley, eds.). Plenum Press, New York, p. 479.
Horii, K., Noguchi, T., and Yano, M. (1970). Spark discharge triggered by a pulse
laser beam, IEEE Internal. Conf on Gas Discharges: Pub. No. 90, IEEE, New
York, p. 6.
Humphreys, D. R., Penn, K. J., Cap, J. S., Adams, R. G., Seamen, J. F. and Turman,
B. N. (1985). Rimfire: A six megavolt laser-triggered gas switch for PBFA II, Di-
gest of Technical Papers, Proc. of the 5th Pulsed Power Conf, Arlington, Va., June
10-12, 1985, (M. F. Rose and P. J. Turchi, eds.), IEEE, New York, p. 262.
Jackson, J.D. (1975). Classical Electrodynamics, (2nd ed.), Wiley, New York, p. 643.
Jensen, R. (1987). KrF for fusion: An overview of laser issues, Fusion Tech., 11:481.
Joshi, C., Mori, W. B., Katsouleas, T., Dawson, J. M., Kindel, J. M. and Forslund,
D. W. (1984). Ultrahigh-gradient particle acceleration by intense laser-driven
plasma density waves, Nature, 311: 525.
Joshi, C., Clayton, C. E., Darrow, C., and Umstadter, D. (1985). Experimental study
of beat-wave excitation of high phase-velocity space-charge waves in a plasma for
particle acceleration, Laser Acceleration of Particles (C. Joshi and T. Katsouleas,
eds.), AlP Conf. Proc. No. 130, AlP, New York, p. 99.
Joshi, C., Katsouleas, T., Clayton, C. E., Mori, W. B., Kindel, J. M., and Forslund,
D. W. (1987). Experimental, theoretical and computational studies of the plasma
beat-wave accelerator at UCLA,AdvancedAcceleratorConcepts (F. E. Mills, ed.),
AlP Conf. Proc. No. 156., AlP, New York, p. 71.
Karttunen, S. J. and Salomaa, R. R. E. (1987). Beat-wave generation of plasma
waves for particle acceleration, IEEE Trans. Plasma Sci., PS-15: 134.
Katsouleas, T. (1986a). Physical mechanisms in the plasma wake-field accelerator,
Phys. Rev. A, 33: 2056.
Katsouleas, T. (1986b). Laser acceleration of particles, High-Intensity Laser Pro-
cesses, SPIE, 664: 2.
Katsouleas, T., Joshi, C., Dawson, J. M., Chen, F. F., Clayton, C. E., Mori, W. B.,
Darrow, C., and Umstadter, D. (1985). Plasma accelerators, Laser Acceleration
ofParticles (C. Joshi and T. Katsouleas, eds.), AlP Conf. Proc. No. 130, AlP, New
York, p. 63.
Katsouleas, T., Wilks, S., Chen, P., Dawson, J. M., and Su, J. J. (1987). Beam loading
in plasma accelerators, Particle Accelerators, 22: 81.
Kincaid, B. M. and Eisenberger (1975). Synchrotron radiation studies of K-edge
photo absorption spectra, Phys. Rev. Lett., 34: 1361.
Klar, B. and Rusticelli, F. (1973). Dynamical neutron diffraction by curved crystals
in the laue geometry, Nuovo Cimento, 13B: 249.
Lunney, J. G., Dobson, P. J., Hares, J.D., Tabatabaei, S.D., and Eason, R. W. (1986).
Time resolved x-ray diffraction from silicon during pulsed laser annealing, Opt.
Commun., 58: 269.
Mallozzi, P. J., Epstein, H. M., and Schwerzel, R. E. (1979a).Advances in X-ray Anal-
ysis, Plenum Press, New York, p. 267.
Mallozzi, P. J., Schwerzel, R. E., Epstein, H. M. and Campbell, B. E. (1979b). Laser
EXAFS: Laboratory EXAFS with a nanosecond pulse of laser-produced x-rays,
Science, 206: 353.
Manley, J. M. and Rowe, H. E. (1956). Some general properties of nonlinear ele-

ments: Part I. General energy relations, Proc. IRE, 44: 904.
Mar, S., Czernichowski, A., and Chapelle, J. (1986). Experimental stark parameters
for some spectral lines in SF6 plasma, J. Phys. D: Appl. Phys. 19: 621.
Matthews, D. L., Campbell, E. M., Ceglio, N. M., Hermes, G., Kauffman, R., Kop-
pel, L., Lee, R., Manes, K., Rupert, V., Slivinsky, R., Turner, R., and Ze, F. (1983).
Characterization of laser produced plasma x-ray sources, J. Appl. Phys., 54: 4260.
Mayer, F. J., Berk, H. L., and Forslund, D. W. (1985). A laser-initiated, energetic ion,
high beta mirror plasma, Commun. on Plasma Phys. and Cont. Fusion, 9: 139.
McKinstrie, C. J. and Batha, S. H. (1987). Energy cascading in the beat-wave ac-
celerator, New Developments in Particle Acceleration Techniques (S. Turner, ed.),
Centre Europeen de Ia Recherche Nucleaire, Geneva, p. 443.
McKinstrie, C. J. and Forslund, D. W. (1987). The detuning of relativistic Langmuir
waves in the beat-wave accelerator, Phys. Fluids, 30: 904.
Michette, A. G., Chen, P. C., Eason, R. W., Feder, R., O'Neill, F., Owandano, Y.,
Rosser, R. J., Rums by, P., and Shaw, M. J. (1986). Soft x-ray contact microscopy
using laser plasma sources, J. Phys. D: Appl. Phys., 19: 363.
Morgan, C. G. (1975). Laser induced breakdown of gases, Rep. Prog. Phys., 38: 621.
Mori, W. B. (1987). Theory and simulations on beat-wave acceleration of relativistic
plasma waves, Ph.D. Thesis, Univ. of Calif. at Los Angeles, published by Univer-
sity Microfilms, Ann Arbor, Mich.
Nagel, D. J., Peckerar, M. C., Whitlock, R. R., Grieg, J. R., and Pecheak, R. E.
(1978). Submicrosecond x-ray lithography, Electron. Lett., 14: 781.
Obrien, H. A. (1987). Private communication.
Pesme, D., Karttunen, S. J., Salomaa, R. R. E., Laval, G., and Silvestre, N. (1987).
Modulation instability in the beat-wave generation, New Developments in Pmti-
cle Acceleration Techniques (S. Turner, ed.), Centre Europeen de Ia Recherche
Nucleaire, Geneva, p. 418.
Phillion, D. W. and Hailey, C. J. (1986). Brightness and duration of x-ray line sources
irradiated with 0.53 micron laser light at 60 and 120 picosecond pulse duration,
Phys. Rev. A., 34: 4886.
Pirri, A. N. (1977). Theory for laser simulation of hypervelocity impact, Phys. Fluids,
20: 221.
Priedhorsky, W, Lier, D., Day, R., and Gerke, D. (1981). Hard x-ray measurements
of 10.6-micron laser-irradiated targets, Phys. Rev. Lett., 47: 1661.
Rapoport, W. R., Goldhar, J., and Murry, J. R. (1980). KrF laser-triggered SF6 spark
gap for low jitter timing, IEEE Transactions on Plasma Science, PS8 (3): 167.
Rosenbluth, M. N. and Liu, C. S. (1972). Excitation of plasma waves by two laser
beams, Phys. Rev. Lett., 29: 701.
Sayers, D. E., Stern, E. A., and Lytle, F. W. (1971). New techniques for investigating
noncrystalline structures: Fourier analysis of the extended x-ray absorption fine
structure, Phys. Rev. Lett., 27: 1204.
Sheffield, S. A., and Fisk, G. A. (1983). In High Speed Photography, Videography, and
Photonics (D. L. Paisley, ed.), Proc. SPIE, Vol. 427, p. 193.
Sheffield, S. A. and Fisk, G. A. (1984). In Shock Waves in Condensed Matted (J. R.
Asay, R. A. Graham, and G. K. Straub, eds.), North-Holland, New York, p. 243.
Silin, V. P. (1965). Parametric resonance in a plasma, Sov. Phys. JETP, 21: 1127.
436 Hauer et al.
Spiller, E. and Feder, R. (1977). Topics in Applied Physics: X-Ray Optics (H. J.
Queisser, ed.), Springer, Berlin.
Tajima, T. and Dawson, J. M. (1979). Laser electron acceleration, Phys. Rev. Lett.,
43: 267.
Turman, B. N., Martin, T. H., Neau, E. L., Humphreys, D. R., Bloomquist, D. D.,
Cook, D. L., Goldstein, S. A, Schneider, L. X., McDaniel, D. H., Wilison, J.
M., Hamil, R. A, Barr, G. W., and VanDevender, J.P. (1985). PBFA II, a 100
TW pulsed power driver for the inertial confinement fusion program, Digest of
Technical Papers, Proc. of the 5th IEEE Pulsed Power Conf, Arlington, Va., June
10-12, 1985 (M. ERose and P. J. Turchi, eds.) IEEE, New York, p. 155.
Wark, J. S., Hauer, A, and Kilkenny, J. D. (1986). Studies of x-ray switching and
shuttering techniques, Rev. Sci. Instrum., 57: 2168.
Wark, J. S., Whitlock, R. R., Hauer, A, Swain, J. E., and Salone, P. J. (1987). The
study of shock launching in silicon using pulsed x-ray diffraction, Phys. Rev. B.,
35: 9391.
Whitham, G. B. (1974). Linear and Nonlinear Waves, Wiley, New York, p. 485.
Woodworth, J. R., Frost, C. A, and Green, T. A (1982). UV laser triggering of high-
voltage gas switches, 1 Appl. Phys., 53: 4734.
Woodworth, J. R., Hargis, P. J., Jr., Pitchford, L. C., and Hamil, R. A (1984). Laser
triggering of a 500-KV gas-filled switch: A parametric study,Appl. Phys., 56: 1382.
Yaakobi, B., Bourke, P., Conturie, Y., Delettrez, J., Forsyth, J. M., Frankel, R. D.,
Goldman, L. M., McCrory, R. L., Seka, W., Soures, J. M., Burek, A J., and Des-
lattes, R. E. (1981). High x-ray conversion efficiency with target irradiance by a
frequency tripled Nd:glass laser, Opt. Commun., 38: 196.
Yaakobi, B., Kim, H., Soures, J. M., Deckman, H. W., and Dunsmuir, J. (1983). Sub-
micron x-ray lithography using laser-produced plasma as a source, AppL Phys.
Lett., 43: 686.
Yaakobi, B., Frankel, R. D., Forsyth, J. M., and Soures, J. (1986). Laser generated x-
ray source for time-resolved biological and materials structures studies, in Struc-
tural Biological Applications ofX-Ray Absotption Scattering and Diffraction, Aca-
demic Press, New York, p. 331.
Yan, Y. T. (1986). Interaction of relativistic electrons and radiation, Ph.D. Thesis,
Univ. of Calif. at Los Angeles, published by University Microfilms, Ann Arbor,
Mich., p. 92.
Yang, L. C. (1974). Stress waves generated in thin metallic films by a Q-switched
ruby laser, 1 Appl. Phys., 45: 2601.
Zel'dovich, Y. B. and Raizer, Y. P. (1965). Cascade ionization of a gas by a light
pulse, Sov. Phys. JETP, 20: 772.
Index
A [Breakdown threshold)
scaling with pulse length, 28--30
Abel inversion, 120, 122, 143 scaling with wavelength, 30-32
Absorption: and surface defects, 51-52
by heated gases, 72-75, 92-93 and thermionic emission, 51
of laser light by a plasma, Breakdown threshold of:
388--389 absorbing dielectrics, 44-48
Absorption coefficient: alkali halide crystals, 39
for inverse bremsstrahlung, 74 aluminum, 49, 51-52
for plasmas, 84, 336 liquid organics, 57
of solids on melting, 330 liquid, 39, 41-43
Absorption wave, 76-77, 100-101 metals, 48--52
Aerosols, 4-5, 52-57 solids by shock waves, 48
spectral analysis of, 309-313 transparent dielectrics, 37-40,
Annealing of semiconductors by 58
lasers, 287-289 Vycor glass, 45
Atomic and molecular beams Bremsstrahlung, 109, 129, 141,
by laser plasma, 170, 181, 152-153,387,391-392
199-201 Bremsstrahlung radiation from
plasmas on solids, 334
B
Bohn-Gross frequency, 127 c
Boltzmann distribution, 134,
337-338 Cascade breakdown, 58
Breakdown threshold: Collisional radiative (CR) model,
acoustic diagnostic, 58 135, 146, 148
aerosol effects, 4-5 Continuous optical discharge
effect of resonance saturation, (COD) (see laser-sustained
32-33 plasma)
measurements,33-36 Coronal approximation, 134-135,
and non-linear effects, 43 140
role of band gap in solids, 38 Coronal plasma:
scaling with pressure, 28-30 generation, 217
437
438 Index
[Coronal plasma] [Electron density]

underdense regime, 217-218 fusion regime, 107-109
Coulomb-Born approximations, in plasmas formed on solids,
142 329,343
Critical density, 217 profile for plasmas, 388
Electron density values:
D from continuum radiation,
140-141
Debye length, 127 from dielectric satellite radia-
Detection limits: tion, 139-140
for elements in solids, 316 from line profiles, 135-137, 152
for laser ablation, 364 from optical interferometry,
for laser evaporation/MS, 371 118-124
Detectors: from stimulated Raman scatter-
gated photodiode array, 299, ing, 114-116
301-302,333-334 from Thomson scattering,
for line intensity measurement, 127-128
341 Electron energy transport model-
photomultipliers, 301 ing, 237-238
Dielectric satellite lines, 139-140,
Electron temperature, 388
399-402
from continuum radiation,
Doppler:
140-141
broadening and width, 135, 339,
fusion regime, 107-109
430
from harmonic scattering,
shift, 113, 334
113-114
from stimulated Raman scatter-
E ing, 113
Einstein coefficients, 336-337, 339 from Thomson scattering,
Electron cascade growth: 127-128
by Boltzmann equation, 16-17 Emission spectroscopy:
measured cross-section, 20 for energy transport in plasmas,
in N 2 , 22 152
numerical solutions, 20-22 for line intensities, 341
by quantum mechanics, 18-19 for solids, 313-318, 330, 336,
Electron cascade losses: 339-340,342-344
by diffusion, 24-25 Energy transport in plasmas:
by three body attachment, 22 optical emission diagnostic, 152
by two-body attachment, 23-24 by soft x-ray emission, 150-152
Electron density: by suprathermal electrons, 150,
critical, 108-109, 119, 387-388, 152-154
390 by thermal electrons, 149-150,
fluctuations, 111, 126-127, 130 153
Index 439
Etching of semiconductors by Inductively coupled plasma (ICP),

lasers, 286-287 172-173,196,201,203
EXAFS, time-resolved: and laser ablation, 356-358
using laser plasma sources, Inertially confined fusion:
404-406 approaches, 208
using synchrotron sources, 405 burnup fraction, 209-210
energy absorption, 217-224
F history, 207-211
hydrodynamic efficiency, 213
Fusion lasers, 252 scaling laws, 211-217
NOVA, 252 temperature, 212
OMEGA, 233-237, 244-250, yields and efficiencies, 256-258
252,257-260 Inglis-Teller limit, 141
Initial electron problem, 3
G Integrated circuit:
fabrication, 270-271
Gaunt factor, 74, 174
laser marking, 281-283
photomask repair, 280-281
H redundancy repair, 278-279
Harmonic emission: resistor trimming, 276-278
for density gradients, 118 Inverse bremsstrahlung, 16-22,
as diagnostic, 108 27-30, 73-74, 108-109, 146,
origins of, 116-118 328-329,430
for plasma filamentation, 118 electron-ion, 74
scattering in fusion plasmas, 113 laser-sustained plasma, 169,
HELlOS laser, 412 173, 178
High-energy particle generation: Ion acoustic frequency, 113
process,387-388,392-393 Ionization:
using laser plasmas, 412-413 multiphoton, 328
Hydrodynamic codes: of atoms in a plasma, 343
LASNEX, 96-98, 132, 144, 235 photoelectric, 328
LILAC, 257-258 thermal, 328
ORCHID, 259 Irradiance, 2
Hydrodynamic instabilities, 153,
154-156 J
I JANUS laser, 402
Implosion diagnostics, 254-256 L

Implosion experiments:
planned,211,246,257-260 Landau damping, 115
summary, 252-254 LARIS, 361
440 Index
Laser ablation: Laser evaporation/mass spec-

detection limits, 364 trometry (MS):
and the DC plasma, 358 applications to analysis, 369-374
and electrothermal atomization detection limits, 371
AA,354-356 instrumentation, 366-369
and flame AA, 353-354 material removal, 367
and the ICP, 356-358 for photophysical studies,
and the ICP mass-spectrometer, 375-376
359-360 with post-ionization, 372-375
and the microwave plasma, review, 365-366
358-359 Laser fusion, (see entries under
and sample collection, 361-363 Implosion diagnostics; Im-
for sample introduction, evalua- plosion experiments; Fusion
tion, 363-365 lasers; Inertially confined
of solids, summary, 350-353 fusion; Laser driven ablation)
Laser-based semiconductor Laser fusion irradiance regime,
fabrication: 106
advantages of lasers for, 269- Laser induced breakdown (see
270,274,276-291 also Breakdown threshold):
annealing by lasers, 287-289 of A1 2 0 3 , 54-56
early uses of lasers for, 269 definition, 1
etching by lasers, 286-287 by inverse-bremsstrahlung,
future developments in, 283 27-30
laser parameters for, 271 liquid aerosol-induced, 56-57
laser processing systems for, mechanisms, 2-3, 70-72
273-276 by multiple photon ionization
material properties, 272-273 (MPI), 26-27
photomask repair, 280-281 solid aerosol-induced, 52-56
redundancy repair, 278-279 of solids, 48-52
resistor trimming, 276-278 threshold definition, 3
thin film deposition, 283-286 threshold irradiance, 3-5
via drilling, 290 Laser induced plasma:
absorption coefficient, 336
Laser desorption, 370--375 effect of pulse shape on forma-
Laser driven ablation (see also tion of, 330
Target irradiation), 216, electron density of, 329, 343
227-239 expansion of, 329, 331-333
ablation pressure, 238-239 formation threshold on solids,
electron thermal flux, 229-232 328,330
hydrodynamic instabilities, lifetime on solids, 329
239-243 nonlinear effects in, 334
mass ablation rate, 229-237 optical depth of, 338
Index 441
[Laser induced plasma] Laser-supported detonation wave:

optically thick, 329, 334, 337- and Chapman-Jouget point,
339,341,343 88-91
optically thin, 338-339, 341-342 pressure profile of, 90-91
for sample introduction, 350- radial expansion of, 96
358 radiation flux, 91
second harmonic generation in, temperature profile of, 90-91
334 theory of, 89
shock waves in, 306,317,331, thermal coupling, 99-100
335 velocity profile of, 90
spectra formed by, 297-298, Laser-supported radiation wave,
314,331-336 92-93
temperature measurements of, Laser-sustained plasma (LSP),
341-342 302-303
temperature on solids, 329, 335 absorption measurements, 181
time resolved studies of solids, absorption processes, 173-181
331-336 atomic and molecular beams,
Laser marking of semiconductors, 170, 181, 199-201
281-283 comparison to other plasmas,
169, 172-173, 177-178
Laser microprobe, 318 in different gases, 170, 177,
Laser plasma diagnostics, review, 180-181, 189-196
105-106 effect of beam orientation, 175,
Laser-supported combustion 181, 190-191
(LSC), 172, 175, 182, 191 electrical energy generation,
170
Laser-supported combustion experimental parameters, 172,
wave: 174-177,181-182,189-
characteristic velocities of, 196
85-86 in flows:
density profile of, 87 experiments, 170, 174-182,
ignition temperature of, 83 192-201
one-dimensional, 77-88 modeling results, 182-189
quasisteady approximation, ignition of, 172, 174, 176, 181,
80-83 189, 191-192
radial expansion of, 97-98 laser propulsion, 170, 181-182,
radiation spectrum of, 80-81 192, 197-199
radiative transport in, 84-85 laser requirements, 169-170,
temperature profile of, 86-88 178, 181, 189-190, 197
thermal coupling, 99-100 models of, 175-176,182-189
Laser-supported detonation radiative processes in, 176-
(LSD), 172 179
442 Index
[Laser-sustained plasma (LSP)] M

for spectrochemical analysis,
170,181,201-203,302-303 Magnetic field, spontaneous in
temperature of: plasmas, 124
from models, 182-189, 199 Magnetic field measurement:
measured, 170, 179-180, from Faraday effect, 123-126
190-196,302 from Zeeman effect, 126
Laser-triggered switches: Mass spectrometers, 368-369
desirable characteristics, 425 Modeling:
jitter in, 426-427 collisional-radiative model, 135
laser energy dependence, 428 of electron cascade growth,
laser pulse length dependence, 16-22
427 of energy transport, 150
laser wavelength for, 427-428, of line shapes, 135-137
430 ofnon-LTE radiation, 141-142
spectral studies on, 428-429 of radial expansion, 96-99
using two laser pulses, 430-431 of satellite lines, 140
Laser use in propulsion, 170, of x-ray emissions, 132-133,
181-182, 192, 197-199 144-147
of x-ray imaging, 143-144
LIBORS, 32-33
Multiphoton ionization (MPI),
Line broadening: 15-16,26-27,328
by Doppler effect, 135, 339 cross sections, 11-15
impact theory, 135 0 2 cross section, 14-15
ion dynamic effects, 136 rate, 8-10
and opacity, 135, 137 coherence effect, 11
by Stark effect, 135-137, 334, by Grey-Morgan, 9-10
338 by Keldysh, 9
Line intensity: near resonances, 10-11
expression for, 337-339 polarization effect, 11
measurement of, 341
Line shapes:
Gaussian, 339
0
Lorentzian, 339 Optical diagnostics:
self-reversed, 334 emission vs. probe methods,
for very dense plasmas, 156 109-110
Voigt profile, 338 for energy transport, 152
as x-ray diagnostics, 135-137, for plasmas, 110-131
156 Optical interferometry:
Local thermodynamic equilibrium holographic, 120-122
(LTE), 134-135, 137, 140, limitations of, 121, 123
143, 178-179, 335-336, 339, to measure electron density,
342 118-124
Index 443
[Optical interferometry] R
Schlieren method, 123
shadowgraphy method, 123 Radial expansion, 96-99
time-resolved, 120-121 Radiation transport equation, 336
Radiative transfer equation, 137
p Rayleigh-Taylor instability, 153,
155,239-243
Parametric instabilities in plasmas, Recombination mechanisms:
110-116,417 dielectronic, 133
Particle accelerators using laser radiative, 133
plasmas: three body, 133
advantages, 413 Redundancy repair of semicon-
applications, 412-413 ductors by lasers, 278-279
beam loading effects, 422-423 Reflectance of solids on melting,
comparisons, 413, 423 330
computer simulations, 416-417 Resistor trimming by lasers,
laser requirements, 423-424 276-278
particle injection, 418-422 Rocking curve, 397-401
s
plasma-wave generation for,
414-418
plasma-wave instability, 417
Saba equation, 72, 134, 343
Partition function, 338, 343
Sample introduction techniques,
PBFA-11 fusion accelerator, 348-350
424-425 Spectrochemical analysis:
Photoelectric effect on solids, 328 by double spark technique, 308
Photoionization: by laser-sustained plasma, 170,
by cosmic rays, 8 181,201-203
bytunneling,5-8 laser-plasma instruments,
Photomask repair by lasers, 318-321
280-281 Spectrochemical analysis of:
Planck blackbody function, 336- aerosols:
337 atmospheric, 310-311
liquid, 311-313
Plasma frequency, 108-119, 173,
with Pb, Cd, Zn, 309
328,367
beryllium, instrument for,
Plasma propagation, by three 318-321
mechanisms, 76, 101 coal gasifier effluent, 303-304
Plasmatron, 170 coal particles, 310-311
Plume: gases, 302-306
effect of laser pulse parameters liquids, 306-309
on,330,334 metals:
formation of, 327-330 solid, 313-315,317-318
444 Index
[Spectrochemical analysis of]: uniformity requirements, 243-

molten, 315-316 245
solids, 313-318 Thermal:
uranium in solution, 308-309 coupling in LSC and LSD waves,
Spectrochemical analysis by laser 99-100
induced plasma: (see also electrons, 109, 141, 149-153,
Spectrochemical analysis; 388-390
Spectrochemical analysis of) x-rays, 388-390
apparatus,297-299 Thin film deposition:
critique, 321-323 conventional methods, 284
detectors, 301 laser-based methods, 285
principles, 296-297 advantages, 285-286
reviews, 295 Thomson scattering:
time-resolution in, 298, 301-302 for electron density, 119, 126-
Stark effect: 128
in plasmas in gases, 430 for electron velocities, 126-128
in plasmas on solids, 334, 338 limitations, 128-129
Stimulated Brillouin scattering in for parametric instabilities, 130
laser fusion plasmas, 111-112, wavevector-resolved, 130-131
115,221 Transient x-ray diffraction:
Stimulated Raman scattering for biological samples, 407-409
(SRS): for crystal annealing, 396-401
for electron density, 114-116 for laser-generated shock waves,
in laser fusion plasmas, 108-116, 401-404
218--221 Transition probabilities:
SRS-A and SRS-C, 218--220 accuracy of, 342
for suprathermal electrons, 108 Einstein coefficients, 336-337,
for temperature, 113 339
Suprathermal electrons, 108, Transition region, irradiance
140-141, 150, 152-154, 211, dependence, 93-95
222-224,387-392 Two-plasma waves instability,
Suprathermal x-rays, 388-390 111-112, 114-115, 130-131
Synchrotron source of x-rays, 394,
405,410 v
T Via drilling of semiconductors by
lasers, 290
TABLASER, 360-362
Target irradiation:
beam characterization for,
w
249-251 Waves:
calculations for OMEGA, ion-acoustic, 112, 131, 417
245-249 plasma, 112-113, 131, 414-424
Index 445
[Waves]: X
x-ray diagnostics:
absorption, 156-160 X-ray generation in laser fusion
continuum methods, 140-141 plasmas, 224-227
emission vs. probe methods, 109 X-rays:
imaging, 142-144 laser plasma generation of:
opacity effects, 138, 140 for biological samples, 407-
overview of methods, 131-132 409
satellite radiation, 139-140 characteristics of, 386, 391-
using soft x-rays, 147-149, 394
150--152 fast switching of, 399-401
using tracers in, 138-139, 143, generation processes, 386-
148-152 392
for lithography, 409-412
for transient x-ray diffraction,
394-404

(Optical Engineering (Marcel Dekker Inc.) 21) Cremers, David a._ Radziemski, Leon J - Laser-Induced Plasmas and Applications-M. Dekker (1989)

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

(Optical Engineering (Marcel Dekker Inc.) 21) Cremers, David a._ Radziemski, Leon J - Laser-Induced Plasmas and Applications-M. Dekker (1989)

Uploaded by

Copyright:

Available Formats

Laser-Induced

Laser Advances Editor:

Optical Materials Editor:

l. Electron and Ion Microscopy and Microanalysis: Principles and Applications,

21. Laser-Induced Plasmas and Applications. edited by Leon J. Radziemski

LASER HANDBOOKS-Edited by Peter K. Cheo

Other Volumes in Preparation

Single-Mode Fiber Optics: Principles and Applications, Second Edition,

MARCEL DEKKER New York-

Laser-induced plasmas : physical, chemical, and biological applications I edited

This book is printed on acid-free paper.

Copyright© 1989 MARCEL DEKKER All Rights Reserved

Current printing (last digit):

PRINTED IN THE UNITED STATES OF AMERICA

In other spectrochemical applications, the laser plasma has been used

1 Physics of Laser-Induced Breakdown: An Update 1

2 Modeling of Post-Breakdown Phenomena 69

3 Introduction to Laser Plasma Diagnostics 105

3.3 Introduction to X-ray Diagnostics 131

4 Laser-Sustained Plasmas 169

5 lnertially Confined Fusion 207

6 Laser-Based Semiconductor Fabrication 269

7 Spectrochemical Analysis Using Laser Plasma Excitation 295

7.8 Prognosis 321

8 Fundamentals of Analysis of Solids by Laser-Produced

8.1 Chapter Organization 327

9 Laser Vaporization for Sample Introduction in Atomic and

9.1 Conventional Solid Sample Introduction for Atomic

10 Current New Applications of Laser Plasmas 385

10.1 Introduction 385

10.4 Laser-Pulsed Power Switching 424

Hector A. Baldis Senior Research Officer, Division of Physics, Laser and

David A. Cremers Chemical and Laser Sciences Division, Los Alamos N a-

David W. Forslund Applied Theoretical Physics Division, Los Alamos Na-

Roy A. Hamil Technical Staff, Laser Projects, Sandia National Laborato-

Philip J. Hargis, Jr. Technical Staff, Laser and Plasma-Processing Sci-

Dennis R. Keefer Professor, Engineering Science and Mechanics, Center

Yong W. Kim Professor, Department of Physics, Lehigh University, Beth-

Joseph M. Kindel Division Manager, Mission Research Corporation, Los

Robert L. McCrory Director, Laboratory for Laser Energetics, University

Colin J. McKinstrie* Director's Fellow, Center for Nonlinear Studies and

Peter G. Mitchell Department of Medicine, Medical College of Wiscon-

Nicholas S. Nogar Deputy Group Leader, Chemical and Laser Sciences

Leon J. Radziemskit Professor, Department of Physics, New Mexico State

Robert G. Root Director, Applied Physics Office, SPARTA, Inc., Lexing-

John M. Soures Deputy Director, Laboratory for Laser Energetics, Uni-

Joseph Sneddon Associate Professor, Department of Chemistry, Univer-

Joseph R. Wachter Materials Science and Technology Division, Los Ala-

Justin S. Warkt Research Associate, Laboratory for Laser Energetics,

Guy M. Weyl Principal Research Scientist, Physical Sciences, Inc., An-

1.2 CREATION OF INITIAL ELECTRONS

PARTICLE RADIUS (J.lml

Figure 1.2 Electron tunneling in a square potential well.

4 (2m )112 e312]

Letting a = 2 A, e be the ionization energy e1 and expressing e1 in eV and

1o-IO L - - - - - - - L ._ _ _ _ __.__ _ _ _---J

0 1010 2x1o10 3 X 1Q10

Figure 1.3 Tunnelling frequency versus I It].

level, rP = 100 nsec (typical width of a C02 gain-switched spike), E1 = 10 eV,

1.3 ELECTRON GROWTH IN GASES

The growth of electron density ne is governed by the following equation:

1.3.2 Multiphoton Ionization

The multiphoton ionization rate is proportional to Im, where m is the num-

m =integer part of (E1/hv + 1) (1.13)

1o-IO L - - - - - - - L ._ _ _ _ . _ _ _---J