INDUSTRIAL

and their applications

James T. Luxon/David E. Parker

This is a flexible introduction to lasers and their industrial applications.

Devices such as photodetectors and modulators are covered, and the
concepts of basic optics pertinent to lasers and their application are
included. The book covers some laser theory for a working understanding
and to dispel the mysticism surrounding the device.

Among its many features, the book:

covers the topic of laser beam optics including propagation, focusing, and
depth of focus in detail for both Gaussian and higher-order mode beams,
includes a chapter on optical detectors, including detector arrays, pre-
ceded by a chapter on semiconductors.
' discusses radiometry, photometry, and optical device parameters,
presents pertinent cases of the interaction.of high-power laser beams and
matter.
devotes separate chapters to industrial applications of low-power and
high-power lasers.
• details specific types of applications with additional theoretical or con-
ceptual material where needed.

PRENTICE-HALL, INC., Englewood Cliffs, N.J. 07632

INDUSTRIAL

.James I Luxon / David JP Part^r

Library of Congress Cataloging in Publication Data

LUXON, JAMES T. (date)

Industrial lasers and their applications.

Includes bibliographies and index.

1. Lasers—Industrial applications. I. Parker,
David E. II. Title.
TA1677.L89 1984 * 621.36'6 84-13385
ISBN 0-13-461369-4

To our wives. Sally and Nancy

Editorial/production supervision and

interior design: Gretchen Chenenko and NanCy DeWolfe
Cover design: Whitman Studio, Inc.
Manufacturing buyer: Gordon Osbourne

© 1985 by Prentice-Hall, Inc., Englewood Cliffs, New Jersey 07632

All rights reserved. No part of this book may be

reproduced, in any form or by any means,
without permission in writing from the publisher.

Printed in the United States of America

1 0 9 8 7 6 5 4 3 2 1

ISBN Q-13-m 3 13t, c ]-M 01

PRENTICE-HALL INTERNATIONAL, INC., London

PRENTICE-HALL OF AUSTRALIA PTY. LIMITED, Sydney
EDITORA PRENTICE-HALL DO BRASIL, LTDA., Rio de Janeiro
PRENTICE-HALL CANADA INC., Toronto
PRENTICE-HALL OF INDIA PRIVATE LIMITED, New Delhi
PRENTICE-HALL OF JAPAN, INC., Tokyo
PRENTICE-HALL OF SOUTHEAST ASIA PTE. LTD., Singapore
WHITEHALL BOOKS LIMITED, Wellington, New Zealand
 Contents

PREFACE VII

Chapter 1 PRINCIPLES OF OPTICS

1-1 Nature of Electromagnetic Radiation 1

1-2 Reflection and Refraction 6
1-3 Mirrors and Lenses 8
1—4 Beam Expanders 13
1-5 Interference and Diffraction 14
1-6 Coherence 19
1-7 Polarization 21
1-8 Fiber Optics 25
1-9 Acousto-Optic Modulators 27
Problems 29
References 30
Bibliography 30

Chapter 2 SOLID-STATE CONCEPTS 32

2-1 Atomic Structure 32

2-2 Energy Bands 35
2-3 pn Junctions 38
2-4 Transistors and Simple Circuits 42
Problems 48
Bibliography 48
 Contents

Chapter 3 RADIOMETRY, PHOTOMETRY,

OPTICAL DEVICE PARAMETERS,
AND SPECIFICATIONS 49

3-1 Radiometry and Photometry 49

3-2 Properties of Light Emitters 51
3-3 Quality Factors for Photodetectors 53
3—4 Spectral Response 54
3-5 Frequency Response 54
Problems 55
Reference 55
Bibliography 55

Chapter 4 LIGHT DETECTION DEVICES 56

4-1 Classification of Detectors 56

4-2 Thermal Detectors 57
4-3 Quantum Detectors 61
4-4 Detector Arrays 70
4-5 Photodetector Circuits 73
Problems 76
Bibliography 77

Chapter 5 INTRODUCTION TO THE LASER 78

5-1 Unique Properties of Laser Light 79

5-2 Requirements for Laser Action 79
5-3 How the Laser Works 81
5—4 Types of Lasers 81
Problems 83
Bibliography 83

Chapter 6 LASER THEORY 84

6-1 Population Inversion 84

6-2 Stimulated Emission 85
6-3 Einstein Probability Coefficients 86
6—4 Amplification 88
6-5 Power Output for CW Operation 90
6-6 Pulsed Operation 91
6-7 Three- and Four-Level Systems 92
6-8 Rate Equations 94
6-9 Optical Resonators 96
Contents

6-10 Line Broadening 98

6-11 Q-switching, Cavity Dumping, and
Mode Locking 101
Problems 103
Reference 104
Bibliography 105

Chapter 7 LASER OPTICS 106

7-1 Nature of the Laser Output 106

7-2 Beam Propagation for Stable Laser
Resonators 108
7-3 Resonator Stability and Beam
Parameters 116
Problems 121
References 122
Bibliography 122

Chapter 8 TYPES OF LASERS 124

8-1 HeNe Lasers 124

8-2 Ion Lasers 127
8-3 Metal Vapor Lasers 128
8-4 Dye Lasers 128
8-5 Ruby Laser 129
8-6 Neodymium-YAG Lasers 131
8-7 Nd-Glass Laser 131
8-8 Diode Lasers 133
8-9 CO2 Lasers 136
Problems 144
Bibliography 144

Chapter 9 LOW-POWER LASER APPLICATIONS I:

ALIGNMENT, GAUGING, AND INSPECTION 145

9-1 Scanning Techniques 145

9-2 Alignment 152
9-3 Triangulation Technique 154
9—4 Diode-Array Camera System 157
Problems 160
Bibliography 162
vi Contents

Chapter 10 LOW-POWER LASER APPLICATIONS II:

INTERFEROMETRY AND HOLOGRAPHY 163

10-1 Mathematical Descriptions of Laser

Light 163
10-2 Laser Velocimetry 167
10-3 Laser Interferometry 169
10-4 Holographic Interferometry 178
Problems 195
References 198
Bibliography 199
Chapter 11 INTERACTION OF HIGH-POWER LASER
BEAMS WITH MATERIALS 200

11-1 Material and Laser Parameters 200

11-2 Basic Heat Transfer Equations 202
11-3 Energy Balance Approximation 203
11-4 Uniform, Constant Irradiance
Model 205
11-5 Gaussian, Circular, and Rectangular
Beam Models 206
11-6 Heating with Melting 208
11-7 Heating with Vaporization 211
11-8 Distributed Sources 214
11-9 CW Welding (Keyholing) 214
11-10 General Approach to Laser
Heating 216
Problems 219
References 221
Bibliography 221
Chapter 12 HIGHER-POWER LASER APPLICATIONS 222

12-1 Surface Hardening 222

12-2 Welding 229
12-3 Cutting 233
12^ Laser Marking 238
12-5 Hole Piercing 239
12-6 Alloying and Cladding 242
12-7 Miscellaneous Applications 243
Problems 243
Reference 243
Bibliography 243
INDEX 245
 Preface

The purpose of this book is to provide the reader with an introduction to lasers
and their industrial applications. To facilitate this objective, such devices as
photodetectors and modulators, which are frequently found in laser applica-
tions, are also covered. And to make the book as self-contained as possible, the
concepts of basic optics that are pertinent to lasers and their applications are
presented. Many engineering students do not cover this material formally in
their course work; moreover, many working engineers and scientists either
have not had training in optics or have been away from it for a long time
and may need a refresher. Some laser theory is presented to provide a work-
ing understanding of the laser and to clear away the mysticism surrounding
the device. When tools are understood, they are used more frequently and
used properly.
The topic of laser beam optics, including propagation, focusing, and depth
of focus, is covered in some detail for both Gaussian and higher-order mode
beams because such information is of practical value to industrial applications
of lasers.
A chapter on optical detectors, including detector arrays, is preceded by
a short chapter on semiconductors, to enhance the understanding of solid-state
optical devices, and by a chapter in which radiometry, photometry, and optical
device parameters are discussed.
It would be impossible to present an exhaustive treatment of the interaction
°f high-power laser beams and matter, but some of the most pertinent cases
are presented in a chapter on laser beam materials interaction. Separate chapters
are devoted to industrial applications of low-power and high-power lasers. Spe-
viii Preface

cine types of applications are presented along with additional theoretical or

conceptual material where required.
This is not a book on lasers but rather a book that is intended to help
prepare engineering students or practicing engineers and scientists for the practi-
cal application of lasers in an industrial manufacturing setting. Thus the book
may be used for a one-semester, junior-senior level course on lasers and laser
applications or by practicing engineers and scientists who need to learn quickly
the essentials of lasers and their applications. Greater depth in the topics on
lasers or materials interactions, for example, can then be obtained from many
more advanced books.
This book can be used in several different ways. The first chapter on
basic optics can be omitted if the reader has a background in optics. The chapters
on semiconductors, parameters, radiometry, and devices can be omitted if the
reader's interests do not lie in these areas; these topics are not essential to the
remainder of the book. \
Readers who have some familiarity with lasers and their properties can
omit the overview chapter on lasers. Any chapters on low-power or high-power
applications may be omitted without loss of continuity.
The authors are greatly indebted to a number of people. We want to
express our appreciation to our families for their patience and encouragement.
We would like to thank many of our students who gave us constructive criticism
and other assistance. We would particularly like to thank Mark Sparschu and
Jim McKinley for reading Chapters 9 and 10 and working the problems. We
also want to thank Ms. Barbara Parker for her skillful proofreading and Ms.
Judy Wing for her patience, skill, and good humor in typing much of the manu-
script.

JAMES T. LUXON
DAVID E. PARKER
 chapter

Principles
of Optics

This chapter is intended to provide the reader with a basic working knowledge
of the principles of optics, including a description of the nature of electromagnetic
radiation as well as geometrical and physical optics. This chapter also provides
a basis for much of what follows in subsequent chapters.

1-1 NATURE OF ELECTROMAGNETIC RADIATION

Electromagnetic radiation exhibits both wavelike and particlelike characteristics,

as does matter when it comes in small enough packages, like electrons. Both
aspects of electromagnetic radiation are discussed and both are relevant to under-
standing lasers. From the point of view of its wavelike characteristics, electromag-
netic radiation is known to exhibit wavelengths from less than 10~13 m to over
1015 m. Included in this range, in order of increasing wavelength, are gamma
rays, x rays, ultraviolet waves (uv), visible light, infrared (ir) light, microwaves,
radio waves, and power transmission waves. Figure 1-1 illustrates the various
parts of the range of electromagnetic waves as a function of wavelength.
The sources of gamma rays (y rays) are nuclear transitions involving
radioactive decay. X rays are produced through electronic transitions deep in
the electronic structure of the atom. Ultraviolet waves result from electronic
transitions involving fairly high energy and overlap the x-ray region somewhat.
Visible radiation extends from about 0.35 jim to 0.75 jim and is due to electronic
transitions, primarily of valence electrons. Infrared radiation results from elec-
tronic transitions at the near visible end and molecular vibrations toward the
 Principles of Optics Chap. 1

x rays Visible Radio

Ultraviolet Infrared Microwaves Power

__ __ I _ | _ | _ | _ | _ | _ | _ | _ |
10-14 10 -12 1Q-10 10-8 1Q -6 10 4 1Q-2 ! 1Q 2- , Q4 , Q6 ^ Q8

Wavelength (meters)

Figure 1-1 Electromagnetic radiation as a function of wavelength.

long wavelength end. Microwaves and radio waves are produced by various
types of electronic oscillators and antennas, respectively.
The term light is used loosely to refer to radiation from uv through ir.
The wavelike properties of electromagnetic radiation can be deduced from
the wave equation presented here in one-dimensional form

where c is the velocity of light and E is the electric field intensity. The wave
equation can be derived from Maxwell's equations, the foundation of all classical
electromagnetic theory. The symbol E in Eq. (1-1) may represent any one of
the various electromagnetic field quantities, but for our purposes the electric
field intensity is of greatest interest.
Another relationship that can be deduced from Maxwell's equations that
is of use to us is Poynting's theorem
S=EXH (1-2)
where S is power flow per unit area, E is electric field intensity, and H is
magnetic field intensity. For a freely propagating electromagnetic wave, it reduces
to
Save = \EH (1-3)

where Save is the average power flow per unit area and E and H are amplitudes.
Light may be thought of as being composed of sinusoidal components
of electric and magnetic fields from the point of view that electromagnetic radia-
tion is a wave. For a simple electromagnetic wave propagating in an unbounded
medium (the electric field varying parallel to a single direction, referred to
as linear polarization), the wave may be schematically represented as in Fig.
1-2.
The electric and magnetic fields are oriented at right angles to each other
and to the direction of propagation z. E, H, and z form a right-hand triad;
that is, E X H gives the direction of propagation. Ordinary (unpolarized) light
contains a mixture of polarizations in all directions perpendicular to the direction
of propagation. Because of the vector nature of the electric field, unpolarized
light can be thought as an equal mix of electric field strength in orthogonal
gec 1_1 Nature of Electromagnetic Radiation

Figure 1-2 Propagation of a plane-polarized electromagnetic wave.

directions, say x and y, with random phase relations between the various contri-
butions to the electric field. The significance of this statement will become appar-
ent later on.
The speed of propagation in free space (vacuum) is approximately 3 X
108 m/s and is equal to 1/V/J-ofo according to classical electromagnetic wave
theory. For an electromagnetic wave propagating in a dielectric medium, the
speed is
1 1

where c is the speed of light in free space and /j,r and er are the relative perme-
ability and permittivity of the medium, respectively. In nonmagnetic materials
v = c/\/Tr. The refractive index of a dielectric medium is defined by
(1-4)

and so it is seen that n = \Xe7 for most dielectrics.

It is possible to show that E = ZH, where Z is called the intrinsic impe-
dance of the medium. This fact can be used to put Eq. 1-3 in the form

= (1-5)
~2 Z
where / is the irradiance (power per unit area). You may recognize the similarity
between Eq. (1-5) and the equation for Joule heating in a resistor, which is
" = F2//?, where P is power, V is voltage, and R is resistance. The one-half
does not appear in the Joule's law heating equation because V is a root mean
square rather than an amplitude.
^__The intrinsic impedance of an unbounded dielectric is Z = \/jI7e, where
Vfj.0 /c0, the intrinsic impedance of free space, is 377fl. Then
377ft 377ft
Z= (1-6)
Or
nonmagnetic dielectrics. Equation (1-5) can therefore be written
 Principles of Optics Chap. 1

2 -
The subject of electromagnetic wave propagation in conductors is beyond
the scope of this book, but a few pertinent facts can be pointed out. The impe-
dance of a good conductor is given by

Z = —e-J (7r/4) (1-8)

where <o is the radian frequency of the light and cr is the conductivity of the
conductor. As can be seen from Eq. (1-8), Z is a complex impedance. E is
very small in a conductor; H is large. When an electromagnetic wave strikes
a conductor, E will go nearly to zero and H becomes large due to large induced
surface currents. The results are considerable reflectance of the incident wave
and rapid attenuation of the transmitted wave. The skin depth, which is a
measure of how far the wave penetrates, is given by

8 = r^—t (1-9)

For frequencies of interest in this book, chiefly visible and ir to 10.6 jam, 8 is
extremely small and absorption can be assumed to take place at the surface
for all practical purposes.
Based on the Drude free electron theory of metals, it can be shown that
the fraction of the incident power absorbed by a metal is given approximately
by
ITT f£~

(1-10)

The reflectance is R = 1 — A, which, for copper with cr = 5.8 X 107(fi —

m)"1 at X = 10.6 p.m, leads to R = 0.985. Actual reflectances may exceed
this value for very pure copper. For highly polished or diamond-turned copper
mirrors the reflectance exceeds 0.99.
The particlelike behavior of electromagnetic radiation is exhibited in many
experiments, such as the photoelectric effect and Compton scattering. It was
in an explanation of the photoelectric effect in 1905 that Einstein proposed
that electromagnetic radiation (light for short) is composed of bundles of energy,
quanta, which are referred to as photons. The energy of each of these photons,
he argued, is hf, where h is Planck's constant (h = 6.6 X 10~34 J • s). Planck
had determined this constant previously in explaining the dependence of black-
body radiation on frequency.
In 1924 DeBroglie proposed a mathematical model for the photon. This
model consists of an infinite sum of waves of different frequencies within a
finite frequency range with an appropriate amplitude function. It was really
Sec. 1-1 Nature of Electromagnetic Radiation 5

nothing more than a Fourier integral representation of a finite pulse, with the
amplitude function chosen to produce a wave packet with minimum uncertainty
products. A schematic representation of such a pulse is given in Fig. 1-3. The
outer solid lines form an envelope of the amplitude of the actual wave. The
uncertainties referred to concern the length of the packet L, its relation to
the uncertainty in its momentum Ap, and the relation between frequency band-
width and the time it takes the photon to pass a given point. These relations
are
h
(1-11)

These relations will be useful later on in discussing coherence of light sources.

DeBroglie proposed a duality of both light and matter; that is, he suggested
that matter should exhibit both wave and particle characteristics. It was later
shown that electrons can be diifracted by crystals and the observed wavelength
agreed with that predicted by DeBroglie. The DeBroglie wavelength can be
deduced by setting Einstein's famous mass energy relationship E = me2 equal
to the energy of a photon hf. Thus
E = me2 = hf
and me = hf/c is the momentum of a photon, and therefore momentum and
wavelength are related by
h
p = me = • (1-12)

This relation holds for both light and matter waves. These results can be used
to show (left as an exercise for the student) that dp/dt = P/c for a total absorption
at a surface, where P is total power in the beam and c is the speed of light.
The radiation pressure is dp/dt divided by the area of incidence. In general,
this pressure can be written
F _(1 + R)P
(1-13)
A cA

figure 1-3 Schematic representation of mathematical model of a photon.

6 Principles of Optics Chap. 1

where R is the fraction of the incident power reflected and A is the area of
incidence. This pressure can be substantial for a focused laser beam.

1-2 REFLECTION AND REFRACTION

The law of reflection states that, for specular reflection of light, the angle of
incidence equals the angle of reflection. This situation is illustrated in Fig.
1—4. The reflected ray lies in the same plane as the incident ray and the normal
to the surface. This plane is referred to as the plane of incidence. A specular
surface is one with a surface finish characterized by rms variations in height
and separation of peaks and valleys (surface roughness) much less than the
wavelength of the light. In other words, a surface that is not a good specular
surface in the visible could be quite specular at longer wavelengths. This is an
important point to remember when working around high power—long waye-
length lasers. i
Most surfaces cause reflected light to contain a portion of specular and
diffusely reflected (scattered) light. The diffuse reflection is the result of random
reflections in all directions due to roughness of surface finish. An ideal diffuser
scatters equal amounts of power per unit area per unit solid angle in all directions.
Hence a perfect diffusing surface is equally bright from all viewing angles. Few
surfaces approach the ideal case.
The law of refraction, or Snell's law as it is commonly called, is given
by Eq. (1-14) and the angles are defined in Fig. 1-5.
n t sin 0i = « 2 sin 02 (1-14)
The law of reflection applies equally to all materials whereas SnelPs law, in
the form given in Eq. (1-14), is valid only for an interface between two dielectrics.
Two phenomena of importance relate simply to Snell's law. The first,
total internal reflection (TIR), occurs when light travels from a medium of
higher refractive index into one of the lower refractive indices. Because n2 <
HI, Snell's law requires that 02 > 0\. At an angle of incidence called the critical
angle 62 becomes 90°. At this angle of incidence 0C, and for all 9\ > 6c, all

Figure 1-4 Law of reflection. Figure 1-5 Refraction at a

dielectric interface.
 1_2 Reflection and Refraction
Sec

• 'dent power is reflected. The critical angle is deduced from Snell's law to
be
2
1,
9aC = sin""
•
— (1-15)

The second phenomenon is Brewster's law, which has significance in many

laser designs as well as other areas. Brewster's law states that when the reflected
and refracted rays are at right angles to each other, the reflected light is linearly
polarized perpendicular to the plane of incidence. The angle of incidence at
which this occurs is called the Brewster angle 0B- Figure 1-6 illustrates this
phenomenon. Because 02 is the complement of 6B, Snell's law gives

n\ sin 0B = n 2 cos 0B
(1-16)

or

This is a reversible phenomenon, unlike TIR, and the Brewster angle from
side 2 to 1 is 90° - 0B or tan'1 (n!/n2).
When light is incident on a surface, a certain fraction of the light is absorbed
or transmitted and the remainder is reflected. The fraction of power reflected,
called the reflectance R, for light normally incident on an interface between
two dielectrics is given by

(1-17)

If light is incident from side 1, there will be a 180° phase shift in the reflected
wave for n 2 > « t but no phase shift if n2 < HI. This factor is important for
antireflection and enhanced reflection coatings on optical components.
For nonnormal angles of incidence, the Fresnel formulas provide the reflec-
tances for parallel and perpendicular polarizations.

igur
Illustration
n oof Brewster's
rewsers law.
aw.
nterface is the yz plane and rays are paral-
M to
<he *
 Principles of Optics Chap. 1

tan . - 02))12
tan(0 2 +0,,)J (1-18)
in (02 - 0Q12
pin
1
[sin ! + 0022)j
sn (0! )J
Note that goes to zero for 02 + 0i = 90°, the Brewster angle condition.

1-3 MIRRORS AND LENSES

In this section the results derive from the application of the law of reflection
to mirrors and Snell's law to thin and thick lenses. Pertinent information concern-
ing spherical aberration is also presented.
The sign convention used here, with regard to lenses and mirrors, follows
that of Jenkins and White, 1976. In this convention, light is always assuined
to be traveling left to right. All object or image distances measured to the left
of a reflecting surface are positive; otherwise they are negative. Object distances
to the left and image distances to the right are positive for refracting surfaces;
otherwise they are negative. Radii of curvature are positive if measured in the
direction of the reflected or refracted light; otherwise they are negative. Lenses
or mirrors that converge parallel rays have positive focal lengths and negative
focal lengths if they diverge parallel rays. An object height measured above
the axis is positive; below the axis it is negative.
Figure 1-7 illustrates the effect of concave (positive) and convex (negative)
spherical mirrors on parallel rays.
Parallel rays are reflected through a point called the focal point F for
the concave mirror. Point C locates the center of curvature and F lies midway
between C and the vertex of the mirror. Hence /, the focal length or distance
from F to the vertex, equals r/2, where r is the radius of curvature of the
mirror. For the convex mirror, rays parallel to the axis are reflected such that
they appear to be coming from the focal point F. For the concave mirror,
rays parallel to the axis are reflected through the focal point. In both cases,

Figure 1-7 The effect of concave and convex mirrors on rays parallel to the axis.
 c 1_3 Mirrors and Lenses

Figure 1-8 Image construction for a

spherical mirror. -=

the rays are reversible. These facts can be used to construct images graphically
for a given object. Graphical image construction is depicted in Fig. 1-8. Analyti-
cally, object distance o, image distance i, and / are related by

V=i
O I f
<i-w>
where all distances are measured from the mirror vertex. Also, it is easy to
see that magnification m is given by

«=-£=-- (1-20)
Y0 0

Equations (1-19) and (1-20) apply equally well to spherical mirrors and thin
lenses if the sign convention is strictly followed. Actually, not all parallel rays
are focused at a common point. This fact leads to what are referred to as
aberrations, the most important of which, for our purposes, is spherical aberra-
tion. Spherical aberration occurs because rays parallel to the axis are focused
closer to the mirror vertex the farther they are from the axis. Also, parallel
rays not parallel to the axis suffer an aberration called astigmatism (nothing
to do with astigmatism of the eye). Strictly speaking, a spherical mirror is
good only for paraxial rays—that is, rays parallel and close to the axis. Parabolic
mirrors do not suffer from spherical aberration; all rays parallel to the axis
are focused at a common point, but astigmatism is more severe.
Spherical mirrors are little used for precisely focusing purposes except
in large telescopes where the focusing is on axis and the spherical aberration
can be corrected. Spherical mirrors are used in lasers as resonator mirrors.Here
spherical aberration is not a problem because the wavefronts produced are spheri-
cal, not plane as we have been tacitly assuming.
Mirrors used for off-axis focusing, such as for high-power lasers, are usually
nonspherical or aspheric. This condition is accomplished by using a sector of
a
parabolic surface as illustrated in Fig. 1-9.

F
figure 1-9 Aspheric off-axis focusing ^ / /'
mirror. ^'
10 Principles of Optics Chap 1

Figure 1-10 Waxicon. Figure 1-11 Axicon.

Many other types of mirrors are used with lasers, such as axicons, so
called because they are specifically designed to produce a particular axial distribu-
tion of light on reflection (axicons may also be transmissive and are used for
spherical aberration correction). A variation, the waxicon (W-shaped axicon),
is used to convert the ring mode output of an unstable resonator to a near-
Gaussian distribution. Such a mirror is depicted in Fig. 1-10. Another type
of mirror focuses the ring mode to a thin ring of finite radius as shown in
Fig. 1-11.
Thin lenses can be treated similarly to mirrors except that the analytical
results are derived from Snell's law instead of the law of reflection. The thickness
of the lens is neglected and two focal points are required. The primary focal
point Fp is defined such that light rays coming from it, or headed toward it,
are refracted parallel to the axis. The secondary focal point Fs is defined such
that rays traveling parallel to the axis are refracted so that they pass through
Fs or appear to be coming from it. Figure 1-12 illustrates these comments
about focal points for converging and diverging lenses. Note that the refraction
is assumed to take place at a single plane referred to as a principal plane.
The focal lengths are also measured from this plane. If the refractive index of
the lens is «' and the refractive indices of the surrounding media are n =
n" = 1, then

(1-21)
f

(b)

Figure 1-12 Primary (Fp) and secondary (Fs) focal point for (a) converging and
(b) diverging thin lenses.
 Mirrors and Lenses 11
Sec 1-3

re r\ and r2 are the radii of curvatures of the first and second surfaces,
actively. The magnification is the same as for the mirrors.
Lenses with spherical surfaces suffer from the same aberrations as mirrors
• several others, including chromatic aberration. Different wavelengths have
different refractive indices; therefore different wavelengths are focused at different
oints. In the context of this book, the monochromatic aberrations, spherical
aberration, coma, and astigmatism are the important ones. Coma, like astigma-
tism, occurs for parallel rays that are not paraxial and causes a lateral smearing
of the image. Coma and astigmatism are minimized by keeping the beam as
nearly parallel to the axis as possible. Curvature of field and distortion are
important in imaging applications.1 The correction for spherical aberration will
be discussed after thick lenses are covered.
In many practical applications, such as focusing of high-power lasers for
materials processing, the thin lens assumption is not valid. Not everything previ-
ously developed need be discarded, however. In fact, by definition of two principal
planes, properly located, the thin lens equation remains valid. The primary
focal length is measured from the primary principal plane and the secondary
focal length is measured from the secondary principal plane. The principal planes
and their relation to the focal lengths and the bending of rays are illustrated
in Fig. 1-13.

Figure 1-13 Thick lens with principal planes H and H".

The lens equation becomes

n =_n=_n
~n (1-22)
w
hich is, in fact, the same equation as for a thin lens. It was assumed before
hat n = n", but it will be instructive to assume n ¥= n" for awhile. Note in
'8- 1-13 where the focal lengths are measured from and where the rays appear
to bend.
In order to determine the focal lengths of a thick lens, the focal lengths
1
See, for example, Fundamentals of Optics by Jenkins and White.
12 Principles of Optics Chap i

V (a)

(b)

Figure 1-14 Refraction at single surfaces.

for refraction at a single surface must be known. The single-surface focal length
equations are given as Eq. (1-23) and relate to Fig. 1-14.
n n n —n
first surface
(1-23)
n{ -n
second surface

Equation (1-24) gives the relationship between the thick lens primary and second-
ary focal lengths in terms of the single-surface focal lengths and the appropriate
refractive indices.
n_n _ n n dri (1-24)
f fa ri fa
J J J\ Jl

Equations (1-25) give the distances of the principal planes (H, H") from the
lens vertices, V\ and V2, as defined in Fig. 1-13.
d ,d_
(1-25)
]
f2 '/I
The procedure for correcting for spherical aberration in spherical lenses
is, in effect, to increase the curvature of both refracting surfaces away from
 c i_4 Beam Expanders 13

Figure 1-15 (a) Partially corrected plano-

convex, (b) optimally corrected meniscus
lens for 10.6-/xm wavelength. (a)

the direction of the oncoming light. An optimum correction exists for all focal
lengths and refractive indices, but spherical aberration cannot be eliminated
for a singleielement lens. Figure 1-15 depicts a partially corrected plano-convex
lens and a fully optimized meniscus lens. The light must enter the lens from
the most curved side; otherwise spherical aberration is increased. Sherman and
Frazier have developed a method for comparing plano-convex and meniscus
lens to determine if the superior correction of the meniscus lens warrants its
higher cost. The spherical-abberation-limited spot diameter dsa (blur of an ideal
focused spot) is given by Eq. (1-26).

dsa=k^ (1-26)

The factor k depends on whether the lens is a plano-convex or meniscus and

the refractive index of the lens D, is the diameter of the beam entering the
lens containing at least 95% of the beam power. Some values of k are listed
in Table 1-1 for X = 10.6 jj,m.
Methods for calculating the diffraction-limited spot size, with which ds&
should be compared, are discussed in a later chapter.

1-4 BEAM EXPANDERS

Because of their importance in laser applications, some discussion of beam expan-

ders is presented here.
Lens beam expanders can be made with two positive or one negative
and one positive lens. These types of transmissive beam expanders, along with

TABLE 1-1 Spherical Aberration Factor

k for Selected Materials at 10.6 ^m

Material Meniscus Planoconvex

Ge 0.0087 0.0295
ZnSe 0.0187 0.0286
GaAs 0.0114 0.0289
14 Principles of Optics Chap.

(a)

'4s

(b)

Figure 1-16 Beam expanders: (a) transmissive, (b) reflective.

two reflective beam expanders, are depicted in Fig. 1-16. In either case, the
lenses have a common focal point. By similar triangles, it is seen from Fig.
1-16 that

(1-27)
Ji
for either lens combination if the absolute values of the focal length of the
negative lens is assumed. It is important to note that the diffraction spreading
of the beams is reduced in proportion to f\ /fi because of the increase in effective
aperture.
The nonfocusing beam expander is practical for most applications since
it can be made more compact. For very high power applications, it is undesirable
to focus the beam in air due to the possibility of breakdown in air or distortion
of the beam due to induced refractive index changes. The focusing beam expander
is used when beam filtering is required. This process basically involves focusing
the beam on a pinhole and filtering (blocking) out the higher-order diffraction
fringes. The beam that exits the pinhole contains much less spatial noise and
hence has a more uniform irradiance distribution.

1-5 INTERFERENCE AND DIFFRACTION

A brief discussion of interference and diffraction phenomena is presented in

this section, particularly as they relate to lasers and laser applications.
When two or more electromagnetic waves having some fixed-phase relation-
ship relative to one another are superimposed, the result is referred to as interfer-
 Interference and Diffraction 15
Sec. 1-5

The electric field intensities must be added rather than simply adding
6
' wers or irradiances as is done when phase relationships are timewise
1 C
dom such as with adding the light from two flashlights. If the beam from
1 r is split into two beams and then these separate beams are recombined
a screen, as depicted in Fig. 1-17, the phase relation between the beams
ill vary from point to point on the screen due to slight path length differences.
This situation can be expressed mathematically as in Eq. (1-28):
! + E2 = Eio sin (2irft) + E20 sin (lirft + 0) (1-28)
where 9 is a phase angle that will vary, depending on location on the screen,
and £10 and £20 are the amplitudes. Since the resultant amplitude is of primary
interest, the electric fields can be treated as phasors (vectors in the complex
plane with time dependence omitted) with an angle 0 between them. Using
the law of cosines produces
E2 = E2 + E2 + 2E10E20 cos 9 (1-29)
Since irradiance is proportional to the electric field amplitude squared, Eq.
(1-29) can be written
/ = /i + /2 + 2^/IJl cos 0 (1-30)
It must be remembered that irradiance is a time-averaged quantity and if 6
varies randomly with time, the last term in Eq. (1-30) goes to zero, leaving
the usual result for addition of noise signals. When 6 does not vary randomly
with time, an interference pattern is produced on the screen with a visibility
V defined by
-* max -« min
V= (1-31)

It is left as an exercise to show that Eq. (1-31) reduces to 2/i/2/(/i + /2).

Clearly 0 < V < 1.
There are numerous ways to produce interference effects with light. Just

Figure 1-17 Split laser beam recombined on screen.

16 Principles of Optics Chap

the propagation of a plane or spherical wave (any shape for that matter) js
the result of continuous interference of an infinity of so-called Huygen's wavelets
that make up the propagating wavefront. Huygen's principle provides a conve-
nient model for the construction of propagating wavefronts given the wavefront
and its direction of propagation at some place and time. According to Huygen's
principle, every point on a wavefront is a source of a spherical wave, with
directionally dependent amplitude, called a Huygen 's wavelet. Each wavelet ex-
pands forward (in the direction of propagation of the original wavefront) at a
rate equal to the velocity of propagation of the original wavefront. The shape
and location of the wavefront after time t are determined by constructing wave-
lets with radius ct and drawing a wavefront tangent to all possible wavelets.
Such a construction is depicted in Fig. 1-18.

New

Figure 1-18 Illustration of Huygen's

principle.

When such a wavefront is interrupted by an aperture or an edge, some

contributors to the propagating wavefront are removed. The result is a diffraction
pattern such as is commonly produced by a knife edge. This situation is schemati-
cally represented in Fig. 1-19. This diffraction pattern can be used for precise
location of a sharp edge.
Let's look at what happens when a laser beam illuminates two thin rectan-
gular slits in a screen. The experimental setup is depicted in Fig. 1-20. This
particular example is referred to as Young's double slit interference. All paralle
rays leaving the slits are focused at a common point in the focal plane of the
lens. The optical path lengths traveled by any two parallel rays starting frorn
a common perpendicular plane are equal. The path length difference between
oblique pairs of rays can be represented as in Fig. 1-20, where a path length
difference of X is indicated for two parallel rays. These rays interfere construc-
tively in the focal plane, because 0 = 360°. If the obliquity is increased unW
 1_5 Interference and Diffraction 17
Sec

ID 4
/-*-

Laser beam Knife edge Screen

(a) (b)

Figure 1-19 Knife edge experiment with laser beam uniformly illuminating the knife
edge.

the difference is 2X or 6 = 720°, constructive interference will again be obtained.

A general equation can be written
m\=d sin 9 (1-32)
where m = 0, 1, 2 Note, also, that y = / t a n 0.
If hundreds or thousands of slits or rulings are provided instead of just
two, it is easy to show that Eq. (1-32) still applies and we have what is called
a diffraction (interference) grating. The number of rulings (lines) per centimeter
is usually given and it equals the reciprocal of d. Because of the multiplicity
of lines, nearly total destructive interference occurs for values of 0 only slightly
different than those given by Eq. (1-32). Hence a diffraction grating can be

Plane
w
avefront

Figure 1-20 Double slit experiment.

18 Principles of Optics Chap

used to resolve (separate) closely spaced wavelengths in a beam of light. Diffrac

tion gratings are frequently used as a means of separating laser lines and in
wavelength selection, or tuning, in lasers.
Another very important diffraction phenomenon is the effect of an aperture
on a beam of light. Figure 1-21 depicts a laser beam incident on a rectangular
slit. In this case, it is not as easy to locate the points of constructive interference-
so we look for the points of destructive interference, which are of more interest
anyhow. The slit is divided into two equal parts of width D/2 and a ray is
taken from the top of each part. If these rays are parallel and differ in path
by X/2 or 180°, then they will interfere destructively as shown. All the rest
of the rays, parallel to these two, will also interfere destructively in pairs. To
find higher-order minima, simply divide the slit into 4, 6, . . . equal parts
and let the path length difference between rays from the tops of adjacent parts
be X/2. A general equation can be developed (left as an exercise)
mX = D sin 6 0-33)
where m = 1, 2, 3, . . . . Note that Eq. (1-33) has the same form as Eq.
(1-32) but a different interpretation. The width of the image line produced on
the screen is taken to be the distance between first minima, 2y. For small angles,

(1-34)

An extremely important case of diffraction'is that caused by a circular

aperture. The derivation for this case is beyond the scope of this book. The
result for the first minimum is
1.22X
sin0 : (1-35)
D

Figure 1-21 Single slit experiment.

 . « Coherence 19
Sec. 1-°

of light made up of plane waves truncated by a circular aperture

A0y
have a minimum divergence angle given by Eq. (1-35), where D is the
mUS
d diameter. Equation (1-35) is frequently applied to laser beams, letting
tfU
h the diameter of the output beam. This is, at best, only a rough estimate
laser beams are not accurately represented by a uniform plane wave.

1_6 COHERENCE

Coherence of electromagnetic waves is a complex subject and no attempt at a

rigorous treatment will be made here. The general nature of coherence is rela-
tively simple, however, and a useful working knowledge can be presented fairly
easily.
Essentially two waves are coherent if there is some fixed-phase relationship
between them. Two pure sinusoidal waves of the same frequency would have
perfect coherence. Unfortunately, physical phenomena can, at best, only be
approximated by pure sinusoids. In reality, there are two manifestations of
coherence, spatial and temporal.
Spatial coherence refers to correlation in phase at the same time but at
different points in space. For the experiments on interference and diffraction
previously described, a high degree of spatial coherence was necessary for inter-
ference or diffraction patterns to occur. Spatial coherence can be produced with
ordinary sources in two ways. One is to place the source far enough away
from the slits, diffraction grating, and so forth so that each emitting point of
the source uniformly illuminates the entire device. The second is to place a
small aperture (pinhole) in front of the source, such as a gas discharge lamp.
Because of the small emitting area of the source viewed by the diffracting device,
substantial spatial coherence is achieved. Lasers emit radiation with nearly per-
fect spatial coherence across the entire beam.
Temporal coherence refers to correlation in phase at the same point in
space at different times. As a wave passes through a given point in space, its
phase will undergo random changes. These changes are unpredictable as to
when they will occur or as to what the change in phase will be. The origin of
ese changes is at the source. The average time between changes in phase for
L lsolated atom is the lifetime of the atom in its excited state, which averages
u s s -pne average emjssion t j me f or iuminous gases, liquids, or solids,
m
in Dorter due to line-broadening mechanisms. Remember, A/Y — 1 accord-
cohe° thC Uncertaintv relations. For lasers operating in a single mode,2 the
bett ' nCe tlmC Tcoh °an ^e much l°nger. approaching that and, in many cases,
enng that of the isolated atom (molecule).
e coherence time is simply the approximate time that it takes a photon

T
j86^ may °Perate in a variety of axial and transverse electromagnetic (TEM) modes.
hev ar»6 dlscui
*ed in
in aalat
later chapter.
20 Principles of Optics Chap

to pass a given point in space. Thus the coherence length Lcoh, which is the
approximate length of the photon, is related to the coherence time by
T f<T (\ "\ f \
i-coh CT c oh U-36)

The Michelson-Morley experiment exemplifies the significance of temporal coher-

ence. This experiment, which uses the Michelson interferometer, is depicted
in Fig. 1-22. Let's assume a laser beam is used as the light source. The beam
is split into two equal parts by the beam splitter. Waves 1 and 2 are reflected
by the corresponding mirrors and are partially recombined at the beam splitter
and observed at the top. In order for interference to be observed over a long
period of time, such as required by the eye, a photographic emulsion, or most
photodetectors, waves 1 and 2 must overlap. Otherwise there will be no fixed-
phase relationship between waves numbered 1 and those numbered 2. Thus
2(L 1 -L 2 )>L c o h
there will be no observable interference. This setup has been used to measure
coherence length or photon length for decades. Before the advent of the laser,
the practical application of the Michelson interferometer was limited to situations
where L\ — L 2 . Lasers may have coherence lengths easily in excess of 100 m;
therefore one leg, the reference leg, can be a few centimeters long whereas the
other leg may be nearly 50 m long. Applications of the Michelson interferometer
and related interferometric techniques, such as holography, will be discussed
in a later chapter.
Some further insight into the concept of coherence and how it can be
measured is obtained from the following simplified approach. A beam may be

,Beam splitter

Laser beam Mirror 2

Y7 Viewer or detector

Figure 1-22 Michelson interferometer.

 21
. -, polarization
Sec. 1-'
f as consisting of a completely coherent part C/ and a completely
thougn 1 — C)7. Because there is no correlation between these parts,
like noise signals:
7 = C / + (1-C)7
x f < i if two such beams are superimposed, the coherent parts
where n0 — . . ., N
w ill add like phasors (coherently):
7COh = (C7) + (C7) + 2VCICI cos 0
whereas the incoherent parts will add like noise signals:
7incoh = (1 - Q7 + (1 - C)7 = 2(1 - C)7
The total irradiance is obtained by simply adding these last two expressions:
=
IT Icoh + lincoh (1~37)

The quantity C could represent the fraction of overlap in a Michelson-Morley

experiment or it could represent the degree of spatial coherence in a double
slit or other diffraction experiment. It is left as an exercise to show that Eq.
(1-37) leads to V = C from the definition of visibility in Eq. (1-30). Thus
the fraction of overlap or degree of spatial coherence is easily measured.

1-7 POLARIZATION

The subject of linear polarization has, of necessity, already been ueau with.
Because electromagnetic radiation is a vector phenomenon, however, other more
general forms of polarization can be obtained from the principle of superposition.
Also, some discussion of how different forms of polarized light may be produced
is in order.
If you imagine superimposing two electromagnetic waves linearly polarized
at right angles to one another but with a 90°-phase difference between them,
e result is circular polarization. Figure 1-23 is a series of head-on views of
ese waves; each point is advanced one-eighth of a wavelength. The tip of
and gSUltant vector is performing uniform circular motion in the plane of E!
and so2 iand haS an amPlitude Er = El = E2. In this view, E2 lags Et by 90°
rotatio C rotation occurs; if the phase relation is reversed, right-hand
in th°H- OCCUrS ' Tlle actual Path of the tip of the resultant vector is a spiral
w
'th ^h lreCtl°n of Propagation. If E± ¥= E2, then elliptical polarization occurs
e
major axis lying parallel to the larger E field. If the phase angle is

Head
'arlv rv.1 '° -°n view of circu-

* W
22 Principles of Optics Char>

Figure 1-24 Head-on view of elliptjca|.

polarized light. ™

not 90°, elliptical polarization occurs with the major axis lying at some angl»
between EI and E2. Figure 1-24 depicts elliptically polarized light from a head-
on view. The phase angle between EI and E2 is given by
sin 8 = Eb/Ea (1-38)
regardless of the relative amplitudes of EI and £2. It is left as an exercise to
prove Eq. (1-38).
Linearly polarized light is produced in many ways. It occurs in gas lasers
when Brewster windows are used to seal the ends of the gas tube or in solid
lasers if the ends of the rod are cut at the Brewster angle. Linear polarization
also arises when repeated reflections occur from mirrors, for the reflectance is
always higher for E± than it is for E\\. Lasers with 45° mirrors to fold the
beam through a number of tubes have this characteristic. Repeated reflection
from a stack of plates oriented at the Brewster angle can linearly polarize an
ordinary beam. Polaroid is used extensively for polarizing light and it consists
of molecules that preferentially absorb light polarized parallel to their long
axis but not their short axis. A sheet of Polaroid is produced by orienting
these molecules with all their long axes parallel. The polarizing direction of
the material is at 90° to the long axis of the molecules.
Figure 1-25 illustrates an important relationship between polarizing direc-
Polarizing
direction

£ cosB

Polarizer

Figure 1-25 Illustration of Malus' law.

 . 7 Polarization 23
Sec. 1-'
adiance called Malus' law. Linearly polarized light is incident on
tion an ^ ^ angle 0 relative to the polarizing direction. If the incident
a
^°^.A* is E the transmitted amplitude is E0 cos 0. Therefore the transmitted
is related to the incident irradiance by
/ = 70 cos2 9 (1-39)
Several other types of devices produce linearly polarized light from ordinary
ht more efficiently and with less attenuation than Polaroid, such as Nicole
A Glan prisms, which are based on the optical properties of transparent solids.
S ace does not permit discussion of all these polarizers. Most standard optics
books describe them in detail.
It is important to discuss briefly some techniques for producing circular
and elliptically polarized light. Many materials exhibit birefringence, which
means two different refractive indices for the perpendicular components of light
passing through the material. Such materials have an optical axis and a unique
refractive index occurs for a ray with the electric field vector perpendicular to
the optical axis—the normal ray. The refractive index for the other ray, the
extraordinary ray, varies from the value for the ordinary ray, depending on
the material, up to a maximum or minimum for the electric field vector parallel
to the optic axis. One such material is the crystal calcite, which is the basis
for the Nicole prism. A ray traveling parallel to the optic axis in calcite travels
faster than one perpendicular to it. Because of the differing refractive indices,
the two polarized beams can be separated. The property of birefringence is
also exhibited by some crystallized polymers and can be induced in a variety
of crystals and liquids by application of a strong external electric field. Examples
are potassium-dihydrogen-phosphate (KDP) and lithium niobate crystals and
nitrobenzene, a liquid. They are referred to as Pockels cells and Kerr cells,
respectively. In both cases, high voltages are required, but the optical axis can
be switched on and off rapidly, in nanoseconds. Besides, the amount of birefrin-
gence (refractive index difference) is a function of applied voltage. Hence the
Phase difference between the two polarizations, after passing through the me-
dium, can be controlled by the applied voltage.
he difference in the refractive index in nitrobenzene is given by
An = k\E2 (1-40)
where k = ? 4d.xv n- 2
Or f- 10 cm/V~ . The electrodes are placed parallel to the direction
live™^agatlon and the optical axis is perpendicular to the electrodes. The refrac-
ex difference in a Pockels cell is given by
kn=pE (1-41)
wl)
ere p = a v m i,
cm/V-i fo . ° cm/V- for deuterated KDP (KD*P) and 3.7 X 1Q-10
1
of r
Pron ° niobate. The electrodes are perpendicular to the direction
Pagation in this type of Pockels cell.
S
consider a device that alters the phase between two rays by 90°;
24 Principles of Optics Chap.

such devices are referred to as quarter wave plates. They could be a sheet
birefringent polymeric material, a Pockels cell, or a Kerr cell. Linearly polarized
light enters the device such that the direction of polarization makes a 45° angi
with the optic axis. Resolving the electric field into components parallel and
perpendicular to the optic axis results in two beams traveling at different speeds
This situation is depicted in Fig. 1-26. Normally a fast axis and a slow axis
are specified in the material. The beam with polarization parallel to the fast
axis advances just one quarter wavelength, 90°, relative to the other beam on
passing through the material. Hence the beam exiting the quarter wave plate
is circularly polarized. If E is not 45° to the optic axis, then elliptically polarized
light is produced. If the thickness corresponds to some phase angle difference
other than 90°, elliptical polarization occurs. A phase change of A/2 or 180°
simply rotates the direction of polarization by 90°. These are important factors
in many laser applications, such as rapid Q-switching of lasers and conversion
of a linearly polarized laser beam to circularly polarized to improve Cutting
ability. ;
An interesting consideration is that of the superposition of two counter-
rotating, circularly polarized light beams. If the beams have equal amplitudes,
the result will be linearly polarized light with twice the amplitude of one of
them. Figure 1-27 will aid in explaining this situation. As the electric field
vectors rotate, their sum remains parallel to the bisector of the angle between
them. If the phase of one of the beams is changed, however, say by moving a
mirror from which it is reflected, the orientation of the resultant E vector will
be changed. A A/8 motion of the mirror causes a A/4, 90° phase change, which
produces a 45° rotation of the resultant E vector. Not only can this be used
to determine how far the mirror has moved, but the direction of rotation of
the resultant E vector indicates which way the mirror has moved.

Figure 1-26 Quarter waveplate. O.A. is the optical axis and t is appropriate thickness
to provide 90° phase shift (A/4) between the orthogonal components.
 1 8 Fiber Optics 25
Sec. 1-°

. !_27 Superimposed counter-rotat-

T circularly polarized beams. (Drawn
slightly different amplitudes for clanty.)
Er is the resultant and is linearly polanzed
with amplitude equal to 2E, for £, = £2.

A final phenomenon of particular interest when it is essential that light

reflected from optical components not be fed back into the laser is the Faraday
effect. Some materials, such as lead glass, become optically active when subjected
to a strong magnetic field parallel to the direction of propagation. Optical activity
means that linearly polarized light entering this material will have its polarization
direction rotated. The amount of rotation is given by
<{> = F51 (1-42)
where B is the magnetic field intensity, 1 is the length of material, and F, the
Verdet constant, has a value of 110 (tesla- meter)"1 for dense flint glass.
If linearly polarized light is passed through a Faraday rotator that rotates
the polarization direction 45°, any light that is reflected back through the rotator
will be rotated an additional 45°, for a total of 90°. This light, traveling in
the reverse direction back toward the laser, is easily attenuated by a polarizer
oriented to transmit the original light beam.

1-8 FIBER OPTICS

he subject of fiber optics is included here because of the growing use of fiber
P ics m optical instrumentation, communication, and laser beam delivery for
surgical and other important applications. Optical fibers may be either glass
s
P astic. When low losses are important, glass fibers are superior, having
ses,f less than l dB/km of fiber length available.s
operat on
' °f simple optical fibers is based on the phenomenon of
^ reflection' If a fiber core is dad with a material of lower refractive
the i rf C°re> tnere wiU always be a critical angle such that rays striking
totally fl°e °f C°re and claddinS at an angle greater than this angle will be
6

This nh ec Thus the fiber becomes a light guide or dielectric waveguide.

men0n 1S i
n H18 28 a ra illustrated in - - .
y is shown entering the face of the fiber at the maximum
and
output enn?SS ,m decibels is defined as 10 log 10 Pm/P^t, where Pin and Pout are the input
nc power, respectively.
26 Principles of Optics Ch
ap.

/
Core Cladding

Figure 1-28 Optical fiber showing entering ray striking cladding at critical angle.

possible angle that will result in total internal reflection of the ray inside the
fiber. Applying Snell's law at the fiber face and assuming the ray strikes the
core-cladding interface at the critical angle, we can show that

Here «i, n 2 , and n3 are the refractive indices of the external medium, core,
and cladding, respectively, 9 is the half-angle of the entrance cone, and NA
stands for numerical aperture.
The cladding in glass fibers is also glass but doped differently than the
core to give it a slightly lower refractive index. For example, if the core and
cladding indices are 1.55 and 1.50, respectively, for a fiber in air, NA = 0.39
and 0! = 23°.
The type of fiber just discussed is a step index fiber. Another type commonly
used is the graded index fiber. Here the refractive index decreases radially from
the center in a parabolic fashion. Rays launched at small angles to the axis
of the fiber continuously change direction toward the axis without ever striking
the cladding. The path is a zigzag combination of straight-line segments between
reflections in a step index fiber whereas the path in a graded index fiber is
essentially sinusoidal.
Losses in glass fibers are primarily due to scattering by impurities and
defects. Losses in plastic fibers are chiefly due to absorption.
Short delay times in optical communications are desirable to minimize
pulse spreading and the resultant degradation of information. There are tw
causes of delay in optical fibers, referred to as modal dispersion and wavelengt
dispersion. Modal delay is caused by the different distances traveled by differen
rays within the fiber. The maximum value for this type of delay in a step m «
fiber can be deduced by considering a ray traveling parallel to the axis an
ray reflecting from the cladding at the critical angle. The time delay pef un
of fiber length is then given by

(1-44)
 Q Acousto-Optic Modulators 27
9
Sec. 1-

to 0 17 ns/m for the fiber example discussed. Actual delay times

This amou ^^ ^ losses are higher for the rays traveling greater distances.
£ fibers are waveguides and, as such, not every imaginable geometri-
n t al
t f r a ray is an allowable mode. In fact, if the fiber is made sufficiently
Pa . .j a f ew micrometers in diameter, only one allowed mode exists.
thm, yV je_mQ(je fleers effectively eliminate modal delay, and picosecond delay
e possible. Graded index fibers minimize modal dispersion because the
f "T r a ray wanders from the axis, the more time it spends in a region of
,ar r refractive index (higher speed). Thus the time delay between axial and
nonaxial rays is reduced.
Wavelength dispersion, or delay due to the variation of refractive index
with wavelength, is minimized by proper material design to reduce this variation
and by the use of narrow bandwidth light emitters, chiefly diode lasers.
It must be recognized that losses also occur due to reflection at entrance
and exit surfaces and coupling of light from emitters into fibers and the coupling
of light out of fibers to detectors. Highly efficient techniques and devices have
been developed, however, for coupling light in and out of fibers and for coupling
light from one fiber to one or several other fibers.

1-9 ACOUSTO-OPTIC MODULATORS

Acousto-optic modulators are used for the production of short, high-power

pulses from lasers, beam deflection, amplitude modulation, and frequency modu-
lation. The variety of laser applications makes a discussion of such devices
pertinent.
An acousto-optic modulator is a device that uses a transparent solid block
of material, such as fused quartz, to which is attached a piezoelectric transducer.
Acoustic waves are produced in the quartz by applying a high-frequency voltage,
say 40 MHz with 0.5 V amplitude, to the transducer. The vibration of the
Piezoelectric transducer sends acoustic waves through the quartz block. This
aTiTn 1S depicted in FiS' l~29- A laser beam is Passed through the block at
froT a"8-le relative to the acoustic wavefronts. Light reflected from the wave-
" S can interfere constructively only if Eq. (1-45) is satisfied,

m\o = 2Xasm6, m = 1, 2, 3, . . . (1-45)

cai
~e, onl 1S},the °ptical wavelength and \a is the acoustic wavelength. With
beam. first-order, m = 1, deflection occurs along with the original
Or
electr °n ~*^ may be recognized as similar to Bragg's law for x-ray
Am"]1 action from crystals, where Xa would be the interplanar spacing.
by varyi ng th ^ ^ Diffracted beam can be varied from zero to a maximum
Xhe f 6 Volta8e applied to the piezoelectric transducer.
requency of the diffracted beam is shifted as a result of the Doppler
28 Principles of Optics a
p. 1

Laser beam

Piezoelectric
transducer

«40 MHz, 1 V
peak-to-peak Figure 1-29 Acousto-optical modulator.

effect. According to relativity theory, when an observer and a light source have
a relative velocity v, the observed frequency is given by
1 + (v/c)
' VI - (vVc 2 ) (1-46)
where /„ is the frequency if observer and source are at rest relative to each
other. In this case, v <^ c and so Eq. (1^-6) may be written
A/= v
fo~~C
(1-47)

The velocity is positive for relative motion toward each other and negative
for motion away from each other. If the light is reflected from a moving mirror,
the Doppler shift is doubled. In the acousto-optic modulator
A/_2vo_sin_0 (1-48)

because the component of acoustic velocity in the beam direction is va sin "•
Combining Eqs. (1-48) and (1-45) for m = 1 gives
A/=±/ s 0'49)

The plus sign is taken if the acoustic wave velocity component is opposite
beam propagation direction and the minus sign is taken if the acoustic veloc
has a component in the direction of beam propagation.
Acousto-optic modulators can be used for deflection (usually less
0.5°) with over 1000 resolvable spots, amplitude modulation, and
modulation.
 Problems 29
]
Chap-
PROBLEMS

1-1. Show that

lutions of Eq. (1-1) and determine the relationship between c, w, and k.

laser beam of 2.5 kW average power is focused to a spot of 50 jum diameter
a medium whose refractive index is 1.5. Determine the irradiance and the
electric field amplitude.
1_3 Copper has an electrical conductivity of 6 X 107 (ohm-m)"1. Calculate the magni-
' tude of copper's intrinsic impedance at the frequency of a HeNe laser (X = 0.6328

1-4. Calculate the skin depth for the data given in Problem 1-3.
1-5. Calculate the theoretical reflectance for highly polished copper for Nd-YAG laser
radiation (X = 1.06 /urn).
1-6. (a) Calculate the momentum of a photon with a wavelength of 10 (mm (CO2
laser).
(b) Calculate the pressure exerted by a 1-terrawatt beam focused to a spot of
150 ^m diameter.
1-7. Calculate the Brewster angle and the critical angle for light traveling from a
medium of 1.5 refractive index (glass) into a medium of 1.33 refractive index
(water).
1-8. Brewster windows are to be attached to the ends of a gas laser tube. The windows
have a refractive index of 1.5 and are to be oriented so that any light reflected
from them is totally polarized perpendicular to the plane of incidence (hence
the name). At what angle relative to the tube axis should these windows be oriented?
1-9. Calculate the pertinent reflectances for light incident normally and at 35° to the
normal for light traveling in air incident on a medium of 1.4 refractive index.
-10. An object is placed on the axis 50 cm from the vertex of a concave spherical
mirror that has a radius of curvature of 30 cm. Locate the image and determine
the magnification. Illustrate with a ray diagram.
peat Problem 1-10 for a convex mirror of the same radius of curvature.
y careful drawing of a ray diagram, demonstrate spherical aberration for a concave
spherical mirror.
• n object is placed on axis 100 cm from a 20-cm focal length positive thin lens.
ae e
j_14 image and calculate the magnification. Illustrate with a ray diagram.
ePeat Pr blem M3 for a
1-15 ° negative lens of -20-cm focal length.
has1*1"* meniscus lens of ZnSe used for focusing high-power CO2 laser beams
r, ~ ^efractive index of 2.4, thickness of 0.92 cm, and radii of curvature of
Cm anC r =
focal 1 * * ^ cm' Locate the principal planes and determine the
e 8th f the lenS assume that the lens is to be used m
C " ° ( air).
at the spherical
10-O fiUm^
1U ! nf O aberration limited spot size of a CO2 laser beam (X =
r 2-cm diameter (95% of the beam power falls in this area) focused
30 Principles of Optics
Chap.

by a 12.5-cm focal length GaAs lens for both meniscus and plano-convex IP
es
(see Table 1-1).
1-17. Design two types of beam expanders, using lenses that will expand a 2 5-
diameter beam to 10 cm. Do not use a lens with a focal length whose absol
value is less than 10 cm.
1-18. Two coherent laser beams interfere with one another on a screen. The irradianc
are 10 mW/cm2 and 15 mW/cm2. Calculate /max, /min> and V. Calculate / at
phase angle difference of 30°.
1-19. Assuming that the diffraction spreading (divergence) of a HeNe laser beam
(A = 0.6328 jum) can be approximated by assuming it is equivalent to the diffraction
caused by a plane monochromatic wave incident on a circular aperture of diameter
equal to the laser beam diameter, calculate the divergence for a 2-mm-diameter
beam.
1-20. The frequency bandwidth of a HeNe laser (0.6328 ju,m) is 100 MHz. Estimate
its coherence time and coherence length.
1-21. A crossed polarizer is placed in front of a plane-polarized light beam. A second
polarizer is placed between the first and the light source such that its polarizing
axis makes an angle of 30° with the E field of the incident light. Assuming 10
mW of light power incident on the second polarizer, how much power would
be detected after the light has passed through both polarizers?
1—22. Calculate the refractive index difference per centimeter in lithium niobate (LiNbOs)
for an applied field of 1000 V/cm. How long must a longitudinal Pockels cell
be to cause a phase change of 90° between the fast and slow rays for ruby laser
light (X = 0.6943 jam)?
1-23. Calculate the numerical aperture and half-angle of the entrance cone for light
entering a step index optical fiber with core refractive index of 1.55 and cladding
index of 1.45. Assume that the fiber is in air.
1-24. Calculate the modal dispersion time delay per kilometer of length for the fiber
in Problem 1-23.
1-25. What angle should a HeNe laser beam make relative to the normal, to the surface
of an acousto-optical modulator operated at 40 MHz with an acoustical wavelength
of 60 /xm? What is the shift in frequency of the optical wave, assuming the optical
wave has a propagation component in the same direction as the acoustical wave.

REFERENCES

JENKINS, J. A., AND W. E. WHITE, Fundamentals of Optics (4th ed). New York
McGraw-Hill Book Co., 1976.

BIBLIOGRAPHY

1-1. Jackson, J. D., Classical Electrodynamics. New York: John Wiley & Sons,
1962. /a
1969-
1-2. Ziock, K., Basic Quantum Mechanics. New York: John Wiley & Sons, Inc.,
 , Bibliography 31
Chap- 1
\4 Ontics and Lasers. Berlin: Springer-Verlag, 1977.
Voun2, i*1'' *
^ and R. A. Phillips, Contemporary Optics for Scientists and Engineers.
Cliffs, NJ: Prentice-Hall, Inc., 1976.
C S and C. A. Becklund, Optics: A Short Course for Engineers and
l 5
' ' Sl&tt New'Vork: John Wiley & Sons, Inc., 1972.
1_6 „Smitn,
-»u w". Tj., Modern Optical Engineering. New York: McGraw-Hill Book Co.,
1966.
v E A., Fiber Optics. Englewood Cliffs, NJ: Prentice-Hall, Inc., 1982.
Mansell D. N., and T. T. Saito, "Design and Fabrication of a Nonlinear Waxicon,"
l
~ ' Optical Engineering, 16, No. 4., July-August 1977, 355-359.
1_9 Arnold, J. B., R. E. Sladky, et al., "Machining Nonconventional-Shaped Optics,"
Optical Engineering, 16, No. 4, July-August 1977.
 chapter

Solid-State
Concepts

The purpose of this chapter is to provide the reader who does not already
have a strong enough solid-state background with an overview of the concepts
required for an understanding of semiconductor optical devices. A short review
of some atomic physics concepts is presented and then the concept of energy
bands in crystals is discussed and related to the pn junction and transistors.
Some simple small-signal amplifier circuits are described.

2-1 ATOMIC STRUCTURE

The Bohr theory of the hydrogen atom (actually any one-electron system), partic-
ularly its successor, the Bohr-Sommerfeld theory, was eminently successful in
providing a model of electronic structure of the atom. Although this theory 1
not valid for systems with more than one electron, the scheme of orbitals an
suborbitals devised for hydrogen serves conceptually as a starting P°'nt ,
building up the electronic structure of all elements of the periodic table,
quantum mechanics later developed by Schrodinger and Heisenberg prove ^
be the correct nonrelativistic theory of the atom but, except for a differ6 ^
in the interpretation of orbital angular momenta, leads to the same resul s
hydrogen as the Bohr-Sommerfeld model. It is the results of the work of Sc
dinger and Heisenberg that are discussed here. The quantum mechanical ana
of the hydrogen atom leads to the conclusion that the electron associated
the nucleus can occupy various energy levels for finite periods of time *i
 , , Atomic Structure 33
Sec. 2-i
. tion when the electron makes a transition from a higher-energy
^ a'^lower-energy level Ei, the energy given off as radiation is
level £2 to
£ JT — Z» f
2 1 — *U

The various states that the electron can occupy are described by a set of
ntum numbers. Three deal with the three dimensions, or degrees of
four q jred t0 describe the electron's motion around the nucleus and
free o ,^ ^^ tQ ^ jntrjnsic angular momentum of the electron. The intrinsic
1 C
1 r momentum of the electron is a relativistic phenomenon and does not
in the Schrodinger or Heisenberg formulations but does in the later
relativistic development due to Dirac. These four quantum numbers are desig-
nated by n, i, mi, and ms. As implied, the first three are a natural result of
the solution of the Schrodinger wave equation for the hydrogen atom and the
fourth, in general, is simply tacked on to account for the relativistic nature of
electrons. The principal quantum number n determines the shells, in chemical
or spectroscopic terminology, and has integer values from one to infinity. Symbol-
ically the numbers n = 1, 2, 3, 4, . . . are represented by K, L, M, N, . . .
etc. The principal quantum number physically relates to the average distance
of the electron from the nucleus and its energy.
The quantum number i can take on values i = 0, 1, 2, . . . n — 1 and
determines the orbital angular momentum of the electron

L = V 1(1 + 1) h
(h = h/2n) and its energy to some extent. Note that the angular momentum
is zero for 1 = 0, suggesting that the electron travels in a straight line back
and forth through the nucleus (in the Bohr-Sommerfeld theory i = 1, 2, 3,
• • • n). It presents no conceptual difficulty, for according to DeBroglie, both
electron and the nucleus behave like waves and can therefore pass through
one another, like electromagnetic waves or water waves, without necessarily
interacting.
he third quantum number, m^, is the magnetic quantum number and
1 68 f mi = lf (1 1}
numb*
(j
" ° projection
~ ~ ~of the
er glves tne • • orbital • (1 ~ 1}>momentum
• ° • • angular L ph sicall this uantum
y onto y anq applied
field f °T external) magnetic field. L can have projections on a magnetic
flgjj • u • ^- The reason L tends to line up with a magnetic
m e ause of tne
ome ° magnetic moment associated with the orbital angular
ass
°ciat ^T °f the electron- Unlike a classical magnet, the magnetic moment
the
magn WUh the electronic orbital motion cannot align itself perfectly with
Ua
^ ntum '1° ^ can ta^e only certain orientations as prescribed by the

^even tho "vT ^Uantum number, ms, sometimes called the spin quantum number
8ives
the'o 8 - therC ^ n° SU°h classical anal°g)> has values of ±\ and physically
ntation of the electron's intrinsic angular momentum relative to
34 Solid-State Concepts ch,

an applied magnetic field. The allowed projections are ±^h. The intrinsic an?
r
momentum of the electron is
1/2

It is easily seen then that the K shell (n = 1) has i = 0, m± = fj, an(1

ms = ±|. At this point we invoke the Pauli exclusion principle, which sav
that no two electrons can occupy the same quantum state. In other words no
two electrons can have the same set of quantum numbers if they are in the
same system. A system might be a single atom, molecule, or an entire crystal
Therefore the K shell can have, at most, two electrons in it. Here we presume
that this quantum number scheme, derived for hydrogen, will apply to atoms
with many electrons, although we do not expect the energy levels to be the
same.
Following this same reasoning, the L shell has 1 = 0, 1, m| = -fl, 0, 1,
and ms = ±| for each distinct set of 1 and m±, providing for a total of eight
electrons. The maximum number of electrons in any shell is 2n2.
In spectroscopic terms, the various values of 1 represent subshells within
each shell and are denoted, for historical reasons, by s, p, d, f, g, . . . for
i = 0, 1, 2, 3, 4, 5 Thus the electronic structure of sodium can be written

Is22s22p63sl

where the numbers represent the principal quantum numbers or shells and the
letters represent the 1 values or subshells. The superscripts are the number of
electrons in each subshell. Sodium has an atomic number of 11; so 11 electrons
should be associated with a neutral sodium atom.
It is the physical nature of atoms, like all natural systems, that they tend
toward their lowest energy configuration which involves achieving full subshells.
Sodium can do so either by gaining another electron or by giving one up, as
it does in NaCl and in its metallic form. In NaCl the chlorine atom picks up
the electron to close its 3p subshell, thus forming an ionic bond with sodium-
Materials with ionic bonds are insulators.
In such metals as sodium or copper the valence electrons are mutua y
shared by atoms in a large region of the material. It is this mutual sharing
electrons that provides the binding forces, the so-called metallic bond.
It is easy to see why sodium and copper are metals: both have unnlie
subshells. It is not so obvious why magnesium, which has closed subshells.
a metal or why .silicon, which at a casual glance does not appear to have
subshells, is an insulator at low temperatures and conducts electricity at ro
temperature. In order to explain metals like magnesium and semiconduc
like silicon, the energy band structure of crystals is introduced.
 Energy Bands 35
Sec 2-2

2,2 ENERGY BANDS

1 the notion of energy bands qualitatively, we will return to sodium

f o dev Imagine 1022 sodium atoms so widely separated that they do
35 an
ct There are 1022 electrons occupying 2 X 1022 separate 3s states.
v '"imagine gradually bringing these 1022 atoms together to form a single
0t

1 with a volume of about 1 cm"3. As they approach each other within a

ry
f atomic radii, the outer energy levels of atoms begin to overlap and become
turbed (this does not mean mad at each other). The perturbation causes
the energy levels to shift, spreading out into a band of 2 X 1022 different, but
losely spaced, energy levels. Figure 2-1 is a schematic representation of this
thought experiment. Here the atoms start out at r = <» and end up at r =
r the equilibrium spacing of the atoms. The original 2 X 1022 3s levels spread
out into a band of 2 X 1022 separate levels shared by the entire assemblage of
atoms. According to the Pauli exclusion principle, this band will be half filled
by electrons. This means that there are 1 X 1022 unoccupied energy states
into which electrons can move to take part in electrical or thermal conduction.
Hence sodium is an electrical conductor (metal in this case) and has an electronic
contribution to its thermal conductivity.
Note that the p levels in Fig. 2-1 also spread out into an energy band
that overlaps the 3s band. Thus sodium would have been a conductor even if
the 3s band had been filled, which explains why the elements of the second
column of the periodic table, like magnesium, are metals.
Materials like diamond, germanium, silicon, and gallium arsenide are more
difficult to explain due to the nature of the covalent bonding in these materials.

Atomic separation
^'gure 2-1 tnep
librinm rgy levels of Na as a function of atomic separation, r0 is the equi-
m atom
'c separation.
36 Solid-State Concepts
Cha P. 2

Covalent bonding is the mutual sharing of electrons between atoms. Con '
carbon, which has atomic number six. We might expect its electronic stru ^
to be Is22s22p2. It turns out that for purely quantum mechanical reasons th &
are beyond the scope of this book, the 2s and 2p levels hybridize into
separate sets of four sp3 levels.1 When the crystal is formed, these sepa
sets of levels form nonoverlapping bands with equal numbers of states in ea t
one. Because the total number of hybrid sp3 states per atom is eight, th
will be four states per atom in each of the bands. If the energy gap betwe
the bands is large, the four electrons per atom will go into the lower enere
band. In diamond the energy bands are separated by a gap of about 6 eV 2
Since the average thermal energy at room temperature is about 0.01 eV, verv
few electrons are able to jump the gap to take part in electrical or thermal
conduction. Therefore, diamond is an excellent electrical insulator. (It happens
to be an excellent thermal conductor, too, but obviously for a different reason)
When the gap between the separate sp3 bands is not so great, the result is a
semiconductor, such as silicon or germanium. Essentially the compound semi-
conductors like gallium arsenide are semiconductors for the same reason.
The distinction between metals, semiconductors, and insulators can be
summarized in a simplified energy band scheme as given in Fig. 2-2. A partially
filled band is indicative of a metal; a material with a relatively small energy
gap will be a semiconductor; a material with a relatively large energy gap is
an insulator. Note, however, that at a high enough temperature, assuming they
do not melt first, all insulators become semiconductors.
Physically the energy gap in a semiconductor or insulator represents the
energy required to break a covalent bond. When sufficient thermal energy is
accumulated to break a bond, the electron moves to the conduction band. There
is now an unoccupied level in the valence band. This unoccupied state, or incom-
plete covalent bond, has a positive charge associated with it and is easily dissoci-
ated from the original site and becomes free to take part in conduction, much
the same as the electron. This positively charged quasi particle is called a hole-
So when a bond is broken, an electron-hole pair (EHP) is produced. Such EHrs
can also be produced by absorption of a photon of sufficient energy, which is
the basis for solar cells and a variety of photodetectors.
Up to this point the nature of pure semiconductors has been discusse •
Such semiconductors are called intrinsic semiconductors because the proper
are intrinsic to the material. Most interesting properties of semiconductors an
from the addition of impurities. The latter semiconductors are referred to
extrinsic. To illustrate the addition of impurities to semiconductors, silicon (
will be used as an example. Silicon forms the diamond structure by shan

1 red to a ^
These hybrid states are denoted by sp3 because one of the 5 electrons is promote
level and four distinct orbitals are formed. . [
2
An eV is the kinetic energy supplied one electronic charge when accelerated tn
volt. 1 eV = 1.6 X ICT19 J.
 37
Energy Bands
Sec. 2-2

- 1 eV

1
T
Semiconductor

(b)

Figure 2-2 Simplified energy band scheme for metals, semiconductors, and insula-
tors.

its four valence electrons with four surrounding atoms in a tetrahedral coordina-
tion. Some Si atoms in the crystal can be substituted for by atoms of similar
size, such as boron (B) or phosphorus (P), which ha^ve valences of 3 and 5,
respectively. When P is substituted, the one electron left over is easily dissociated
from the parent atom to become a free electron in th« conduction band. The
P atom becomes a negative fixed ion and no hole is produced in the process.
Since the number of intrinsic EHPs at room temperature is about 1.5 X 1010
cm"3, the number of atoms about 5 X 1022 cm"3, even a Light impurity concentra-
tion of one part per million (ppm) or 5 X 1016 cm"3 means that there will be
about a million more extrinsic electrons than intrinsic ElHPs. Thus the predomi-
nant charge carrier will be electrons and this extrins-ic material is labeled n
type for negative-charge carriers. Actually, the number of holes diminishes mark-
edly because of the "law of mass action," which requires that
np = nf (2-1)
w er
e n is the number of electrons per unit volume, p the number of holes
er unit volume, and m the number of electrons or holes per unit volume in
nnsic
material. Thus for the example given,
(1.5 X 1010)2
= 4.5 X 104
0.5 X 1016
orders of magnitude less than n. The holes can be ignored.
is doped with B, then a p-type (positive-charge carriers) material is
Val
en of the samp reasoning as for the «-type material. The B atom has a
p e t e ce o com e
bond - P te te our onds wt ts neighbors. The incom-
ion easily
and nr !! dissociated from the parent B atom, making it a negative
C0r a h le in the valence band
»ductio ° capable of electrical and thermal
38 Solid-State Concepts
Cha P. 2

The properties of p-type and n-type semiconductors by themselv

of limited value. However, when combined in pairs—that is, pn junctio ^
myriad of electronic and optical devices becomes possible. ^a

2-3 pn JUNCTIONS

Metallurgical junctions between p- and n-type materials are produced by diff

sion, epitaxial growth, and ion implantation. In all cases, the junction is rna(j
in a single crystal. It is instructive, however, to imagine starting with two separat
crystals, one n type, the other p type, and suddenly joining them to form one
continuous single crystal with the metallurgical junction at the interface. Figure
2-3 depicts the various steps in this thought experiment. At the instant the
materials are joined there is an infinite concentration gradient across the junction
for both holes and electrons. Consequently, holes and electrons diffuse in opposite
directions, producing a large current until sufficient fixed ionic charges are uncov-
ered in the newly created depletion region (devoid of free charge) to produce
a counter-electric field that causes an opposing drift current. The following
current balance holds true at equilibrium.
f elec. — Adrift elec (2-2)
=
holes -/drift holes

Equations (2-2) state that at equilibrium the diffusion current of electrons equals
the drift current of electrons and that the same is true for holes. In terms of
the energy band scheme, this situation can be represented as in Fig. 2-4, where
only the lower conduction band and upper valence band edges are shown. £F

P type n type
, 1 • —
++++++ + 4- + +

n n i n rY - - - ~
1
1 • —
+ + + +' + + 4- + + +

Depletion 1
region

Electrons (~) Negative ions

-f Holes (+) Positive ions

Figure 2-3 pn junction thought experiment. When the junction is formed, ho

and electrons diffuse across the junction, uncovering bound ionic charge in the dep
tion region. The resulting built-in electric field creates drift currents that balanc
the diffusion currents.
 pn Junctions 39
Sec.
2-3

Figure 2-4 Energy band edges for a pn junction at equilibrium.

is called the Fermi energy and for our purposes is a reference energy that lies
about in the middle of the energy gap for intrinsic materials and is shifted
toward the valence band for p-type and toward the conduction band for
n-type materials. Ef is constant throughout a semiconductor at equilibrium;
thus it aids in drawing energy level diagrams.
The electric field built into the depletion region as a result of the uncovered
ions produces a potential energy barrier for electrons and holes, preventing
substantial diffusion of either charge carriers from their majority areas into
the minority regions. The pn -junction device being considered here is a diode
and can be either forward or reverse biased by placing voltages across it as
depicted in Fig. 2-5. When the diode is forward biased, the Fermi level is
shifted upward on the p side, downward on the n side, thus lowering the potential
energy barrier and allowing a large increase in the diffusion of holes and electrons
across the junction. A resistor is placed in this circuit to limit current. The
drift
current is decreased, but this factor has relatively little effect because it

(a)

(b) (c)
p
>gure 2-5
(a) Diode symbol, (b) Forward-biased diode, (c) Reverse-biased diode.
40 Solid-State Concepts Chap. 2

(a) (b)

Figure 3-6 Energy band edges for (a) forward-biased and (b) reverse-biased diodes.

is a current of minority carriers and there are few to begin with. When the
diode is reverse biased, the shifts in the Fermi level are reversed, thereby increas-
ing the potential energy barrier. Although there is a small increase in\the drift
current, it is relatively unimportant because it is a minority current and cannot
be increased to any very large absolute value. In reverse bias the diode acts
like an extremely high resistance. The energy level picture of forward and reverse
biasing is given in Fig. 2-6. The diode current versus diode voltage curve (charac-
teristic curve) is sketched in Fig. 2-7 for Si. Note that the current scales are
different for forward and reverse voltages.
A Si diode is said to "turn on" at about 0.55 to 0.7 V forward bias; 0.6
V will be used here for convenience. Large increases in current occur for voltages
slightly larger than the "turn-on" voltage. The reverse current is relatively con-
stant and usually less than 1 nA until reverse breakdown occurs. This phenome-
non will be discussed subsequently. The characteristic curve, exclusive of the
breakdown region, obeys the following diode equation closely:

Figure 2-7 Diode characteristic curve for Si, showing reverse breakdown.
 2-3 Pn Junctions 41

/= I
°( exp ir~ 1 ) (2 3)
~
here /o is the reverse saturation current, e the electronic charge, V the voltage
ross the diode, k Boltzmann's constant, and T absolute temperature. A conve-
nient number to remember is that kl = 0.026 eV at T = 300°K.
The reverse breakdown is caused by two phenomena. At low voltages it
is caused in extremely heavily doped diodes by quantum mechanical tunneling
of electrons through a very thin depletion layer. At higher voltages in less
heavily doped diodes the cause is an avalanche effect in which electrons passing
through the depletion layer acquire sufficient energy, because of the large electric
field produced by the reverse bias, to ionize neutral atoms in the depletion
layer. At a sufficiently high voltage the result can be an avalanche or multiplying
effect. Neither type of breakdown is inherently harmful. Damage occurs if the
heat generated by a large breakdown current is not dissipated rapidly enough
to prevent melting of the diode. Reverse breakdown is used for voltage control
and as the basis of avalanche photodiodes.
Different types of semiconductor diodes have different turn-on voltages
and reverse saturation currents. Some values of these quantities are listed in
Table 2-1.

TABLE 2-1 Reverse Saturation Currents and "Turn-on" Voltages for Some Important
Semiconductor Diodes

Reverse Saturation Turn-on Voltage

Current (Volts)
Si
<1 nA 0.6
Ge
<1 jj.A 0.3
GaAs _ 12
Ga
As.6P4 _ 18
Gap _ 2.2

In determining the resistor size to use in a circuit in order to limit the

orvvard current through a diode, it is customary to assume a voltage drop
across the diode equal to its "turn-on" voltage. This situation is illustrated in
J
8- 2-8 for a Si diode. The resistance in Fig. 2-8 is given by

„ 12V-0.6V
* = 0.050.4 =288fl

•r t0
^Im't t^le current during reverse breakdown, we would do the same
rig. 2-8 except that the precise value of the reverse breakdown voltage
l be known. This value can be obtained accurately by means of a curve
42 Solid-State Concepts Chap

0.6V

I = 50 ma

12V
— I— Figure 2-8 Silicon diode circuit fur cur-
rent limitation.

2-4 TRANSISTORS AND SIMPLE CIRCUITS

Many different types of transistors are available. In this section the bipolar
junction transistor (BIT) is discussed and the junction field effect transistor
(JFET) and metal-oxide-semiconductor (MOS) are briefly described. The BIT
is referred to as such because two types of charge carriers are responsible for
its properties. It has two back-to-back pn junctions as schematically depicted
in Fig. 2-9 for an npn transistor.
The regions of the BIT are designated E for emitter, B for base, and C
for collector. The base region is very thin, approximately 1 pm. The depletion
regions are shown shaded. The doping levels become progressively lighter going
from emitter to collector with typical values of 1019 cm"3, 1017 cm"3, and 1015
cm~3 for emitter, base, and collector, respectively. The transistor symbol and
an instructive, although impractical, circuit are shown in Fig. 2-10.
Referring to the circuit in Fig. 2-10, the emitter-base junction is forward
biased, thus causing a large flow of electrons into the base region with a relatively
small flow of holes into the emitter because of the lighter doping in the base.
The base region is so thin that most electrons pass directly through it into
the electric field of the reverse-biased base-collector junction, which carries them
on into the collector. A small reverse saturation current occurs at the base-
collector junction. The few electrons that do not make it through the base,

Emitter Base Collector

n P n

1f ? I

\
1
11
i I
w fy
^
Depletion regions
^
1
Figure 2-9 Schematic of npn tran- Figure 2-10 Simple circuit illustrating
sistor. BJT biasing.
 Transistors and Simple Circuits 43
Sec. 2-4

the hole current from the base to emitter, are made up for by a small
P US
of electrons out of the base. The currents shown in Fig. 2-10 are conventional
ents- so they are in the opposite direction of the electron flow. According
° KirchhofFs current law, the three currents must be related as follows:
IE = IB + Ic (2-4)
The dc or small-signal current gain of the transistor is j3 = IC/!B so that

IE — (2-5)

Because /3 usually has a value of 50 to 400, it is frequently sufficient to assume

/ = Ic. pnp transistors behave similarly except that the current directions
and voltage polarities are reversed. Whenever possible, transistors are made
npn to take advantage of better conductivity of electrons compared to holes.
Most silicon BJTs are constructed in a thin epitaxial layer on a silicon
chip and so the actual device looks more like the sketch in Fig. 2-11, although
it is still highly schematic. Contact toE,B, and C is made by means of aluminum
deposited through openings in a SiO2 passivation layer, which is grown directly
on the Si. The n+ symbol represents doping approaching the solid solubility
limit. The buried layer is used to reduce the resistance (collector resistance)
between the collector and collector contact. The drawing, of course, is not to
scale. The lateral dimensions are reduced by at least a factor of ten compared
to the vertical direction. Obviously some transistors, particularly high-power
devices, may look much different than Fig. 2-11, but this diagram is fairly
representative of BJTs used in integrated circuits, low-power applications, and
optoelectronic devices.
Transistor circuits are covered thoroughly in many excellent books. It
may be helpful to the uninitiated reader, however, to present some simple circuits
here.

Collector
Epitaxial layer
n+ buried layer

Figure 2-11 Silicon BJT.

44 Solid-State Concepts Chap

Figure 2-12 BIT used as a switch. Figure 2-13 BJT used as an inverter.

The BJT may be used either as an amplifier or as a switch. A simple

switching scheme is shown in Fig. 2-12. The transistor does noiNfurn on until
the voltage drop across the E-B junction approaches 0.6 volt. In Fig. 2-12
Fin = 0.6 volt + IER and Fout = IER. Another useful configuration in logic
circuits is the inverter shown in Fig. 2-13. When Vin = 5 volt, Fout is near
zero volts and when Vin = 0, Fout = 5 volt. R i can be used to limit the base
current and R% drains stored charge out of the base when the transistor is
turned off to improve the speed of operation. The bleed-off resistor technique
has been supplanted in integrated circuits through the use of Schottky diode
clamped transistors. The Schottky diode is a metal-silicon diode that turns on
at about 0.3 volt and is built from collector to base in such a way that the
B-C junction cannot become forward biased, which is what causes charge to
be stored in the base when the transistor is turned on.
Figure 2-14 contains two simple biasing schemes for small-signal, one-
stage amplification. These circuits are frequently useful with optoelectronic de-
vices. In Fig. 2-14(a), the voltage gain is v out /vin = —Rc/re, where re = 0.026
volt//fi and in Fig. 2-14(b) the voltage gain is v out /v in = -RC/RE. The minus
sign is used because of the phase change seen in the inverter and vout and vm
are small-signal ac voltages. The biasing resistors RI and R2 are chosen to
make /i ~ I2 > 10IB so that

Vn = = 0.6 volt + IERi

R2
where /1( 72, and IE are dc currents. If the output is taken from the ernitte .
the amplifier is referred to as an emitter follower. The voltage gain is slign
less than one, but the output resistance is extremely low, making the erni
follower very desirable as the output stage of a multistage amplifier, such
an operational amplifier (op amp).
The purpose of the biasing resistors RB in Fig. 2-14(a) and Ri,
Fig. 2-14(b) is to provide a dc quiescent or operating point for the tran
 Transistors and Simple Circuits 45
Sec

(a) (b)

Figure 2-14 Simple common emitter amplifiers.

Then the small ac signals input to the base will be amplified at the output.
One of the biasing resistors may be replaced by a photodiode or phototransistor
in optical detection schemes.
A JFET is schematically depicted in Fig. 2-15. When a voltage is applied
between drain D and source S as shown, conduction takes place by means of
a drift current caused by the electric field produced in the channel between D
and S. If VD becomes large enough, or a negative voltage is applied to the
gate G, the channel begins to pinch off as a result of an increase in the size
of the depletion layer between the p and n regions. The current saturates at
this point. BJTs are essentially linear in that when properly biased, changes
m output current are linearly proportional to changes in input current. JFETs
obe
y a quadratic relationship—in this case, ID « FJ, where VG is voltage applied
|° the gate. This nonlinearity makes FETs less desirable for linear amplifiers,
it has many useful applications in communications and optical detection.

^-channel

Figure 2-15 n-channel JFET.

46 Solid-State Concepts Chap

T
'ty/fifM '//%t// y//////
^^?

Figure 2—16 MOS n -channel depletion mode device.

JFETs have high input resistances, for the gate current is always looking into
a reverse-biased diode, relative to the drain. That is, Ia = IG + ID and /G is
extremely small. This makes JFETs useful in the first stage of high-input resis-
tance amplifiers, such as some of the better op amps and the ilrst amplifier
stage in radios.
A MOSFET device is shown in Fig. 2-16. In these transistors conduction
is by field effect—that is, drift current—just as in JFETs. The gate, however,
is completely insulated from the semiconductor. These devices have also been
dubbed insulated gate FETs or IGFETs. In the version shown in Fig. 2-16 a
negative voltage applied to the gate pushes electrons out of the channel, thus
depleting it, and this normally on device is turned off at high enough negative
gate voltage. Like JFETs, the current voltage relation is quadratic. The input
resistance of these devices is extremely high, because the gate is insulated. MOS
devices are used heavily in digital circuits and in some optical devices.
Because of its usefulness in many optical applications, a brief introduction
to the ideal op amp is given here. The reader is referred to any of a large
number of electronics books for more detailed coverage. Some simple optical
applications of op amps will be presented in a later chapter.
The ideal op amp has infinite input resistance, zero output resistance,
and infinite open-loop gain. To a first approximation, any amplifier can be mod-
eled by the circuit shown in Fig. 2-17. The ideal amplifier would then have
Rm = °°, Ro = 0, and A0 = °°. This means that v in could come from a high
resistance source without losing voltage to the source resistance and the outp u
of the amplifier can be connected to a low resistance load without losing voltag
in R0.

O Figure 2-17 Circuit model for amp1

 Transistors and Simple Circuits 47
Sec:. 2-4

Figure 2-18 (a) Inverting amplifier, (b) Noninverting amplifier.

Having v<> = Avi = °° is a convenience that will become apparent as we

look at some simple op amp circuits. The symbol used to represent an op amp
and two useful circuits, an inverting amplifier and a noninverting amplifier,
are presented in Fig. 2-18. Because of the large open-loop gain (this means
no feedback loop) of the op amp, Vi in Fig. 2-18(a) is nearly zero, thus making
the negative or inverting input a "virtual" ground. Also, because of the high
input resistance, the current between the + and — terminals is essentially zero.
Consequently,
V in — 0 0 — Vout

Rf

or . _ Vout _ Rf
^C n (2-6)
Vin Rin

which is the so-called closed-loop gain. The input resistance of this amplifier
is RI and its output resistance is3
RO RI~\- Rf

In
Fig. 2-18(b) the roles of the inverting and noninverting inputs are
rsed. Now the voltage at the inverting terminal is approximately vin. Thus
. _ V in — 0 _ V 0 ut
( — (2-7)
m + Rf
and so

,, D I D
, Vout -<Mn T Kf
Ac = =
Vin Rin

'Do 0 conl
use op amp input and output resistance with amplifier input and output resis-
48 Solid-State Concepts Chap

The gain is a little higher and the input and output voltages are in phas
The input resistance of this amplifier is A0Rin (R^/Ri + J?/), which is generally
very large, and its output resistance is the same as the inverting amplifier.
The input voltage in these circuits may come from a photodetector Or
all or part of the input resistance may be replaced by a photoresistor.
A standard low-cost op amp, the 741, has Rin = 1 to 2 MO, R0 — 75 ^
and a dc open-loop voltage gain of —2 X 105 that decreases with frequency to
—100 at 10 kHz. In order for the relations just given to be valid, the open-
loop gain of the op amp must always be much greater than the closed-loop
gain.

PROBLEMS

2-1. The energy of the ground-state electron in hydrogen is —13.6 eV. In the first
excited state the energy is —3.4 eV. Calculate the frequency and wavelength of
the light given off when an electron undergoes a transition from the first excited
state to the ground state.
2-2. Write out the electronic configurations in spectroscopic notation for Mg, Si, P,
and K.
2-3. A crystal of n-type silicon at room temperature has 5 X 1018 phosphorus atoms
per cubic centimeter. How many electrons and holes are there per cubic centimeter?
2-4. The turn-on voltage for a GaAs light-emitting diode (LED) is 1.2 volts. What
resistance is needed in series with the LED to provide a current of 50 mA if a
12-volt battery is used?
2-5. A transistor has a ft of 200. Calculate the emitter current if the collector current
is 1.5 mA. What is the base current?
2-6. Design a simple inverting op amp circuit to have a closed-loop gain of 15.
2—7. Design a noninverting op amp circuit to have a gain of 10.
2-8. The energy gap in GaAs is 1.4 eV. Calculate the wavelength and frequency of
the photon given off when an electron from the conduction band recombmes
with a hole from the valence band.

BIBLIOGRAPHY

2-1. VanVlack, L. H., Elements of Materials Science and Engineering (4th ed.). Reading.
MA: Addison-Wesley Publishing Co., 1980.
2-2. Kittel, C, Solid State Physics (5th ed.). New York: John Wiley & Sons, I"6"
1976.
2-3. Till, W. C., and J. T. Luxon, Integrated Circuits: Materials, Devices, Fabricat'0"'
Englewood Cliffs, NJ: Prentice-Hall, Inc., 1982.
2-4. Hewlett-Packard, Optoelectronic Designer's Catalog. Palo Alto, CA: 1978.
 chapter
Radiometry,
Photometry,
Optical Device Parameters,
and Specifications

This chapter introduces radiometric and photometric quantities and the parame-
ters and physical characteristics of solid-state light emitters and detectors that
are used to describe the physical properties most frequently considered when
deciding which device is best suited for a given application. The spectral output,
power and/or brightness, spatial distribution of light, current requirements and
limitations, and voltage requirements are discussed for light emitters. Responsiv-
R, detectivity D, detectivity per unit frequency D*, noise equivalent power
P, spectral response, and frequency response are covered for photodetectors.
me
comments about noise considerations in photodetectors are made, but
a
n m-depth coverage of this subject is beyond the scope of this book.

3
~1 RADIOMETRY AND PHOTOMETRY

of . response of the eye is considered in the measurement or specification

e ra
in , . diation, the subject is called photometry as opposed to radiometry,
direct
Physical measurements or specifications are made. The eye re-
CUrve
is given in Fig. 3-1 for what is called photopic or day vision.
Not e Ve 1S ^'fted toward shorter wavelengths for night or scotopic vision.
e> f course
° > has the same eye spectral response, but Fig. 3-1 is
to re resentat ve
the H' P ' - The unit of luminous flux is the lumen, corresponding
ra
dioin lornetric unit, the watt. Incident power per unit area (watts/cm2) in
as
'Hum' 1C terms 's ca^ecl irradiance. In photometric terms it is referred to
ance (lumens/cm2). The units given here are not necessarily SI units
50 Radiometry, Photometry, Optical Device Parameters, and Specificatio
Cha

0.35 0.55 0.75

(Wavelength, /jm)

Figure 3-1 Spectral response curve of the eye (color response.)

but are commonly used combinations. The proper SI units adhere strictly to
the MKS system where applicable, such as in area units.
Before discussing the remaining photometric and radiometric quantities
the concept of solid angle must be introduced. The solid angle H is the three-
dimensional counterpart of the usual planar or two-dimensional angle defined
by the ratio of the arc of a circle to the radius of the circle as illustrated in
Fig. 3-2. A solid angle is defined as the ratio of the area of a portion of a
sphere to the radius squared. This concept is illustrated in Fig. 3-3. The full
planar angle is 2irr/r = 2?r radians whereas the full solid angle is 47rr2/>'2 =
47T steradians. A useful relation between 0 and fl exists for small angles and
circular beams. Refer to Fig. 3-4 for an explanation of this relationship. As
can be seen, fl ~ (4vr)02 in this case.
Luminous intensity and radiant intensity are defined as luminous flux
per unit solid angle and power per unit solid angle, respectively. A lumen per
steradian is called a candela. These quantities define the ratio of the flux emanat-
ing from a source into a cone, defined by solid angle fl to the solid angle

Area = A

^_
,7

Figure 3-2 Defini- Figure 3-3 Definition of solid angle.

tion of planar angle.
 properties of Light Emitters 51
2
Sec 3'

V4 Relation between 02 and 0. for small-divergence angle, circular cross-

sectional beam.

h t cone. A laser emits a beam of very high intensity compared with most
Vrmal light sources because the beam it emits has an extremely small divergence
angle. The radiant intensity of a 1-mW HeNe laser is 1000 W/sr for a full-
angle beam divergence of 1 mrad.
The final quantity of interest is intensity per unit-emitting area. In photome-
try it is called luminance and might have units of candelas/cm2. The radiometric
term is radiance and the units are frequently watts/cm2-sr. Luminance (radiance)
is the flux emitted by an area of the source divided by the emitting area and
the solid angle of the cone into which the flux is emitted. Luminance is often
referred to as brightness.
Consider a 1-mW HeNe laser with a full-angle divergence of 1 mrad
and a beam diameter of 1 mm. The radiance for this laser will be approximately
1.6 X 105 W/cm2-sr. It is informative to compare it with the radiance of the
sun, which is about 130 W/cm2-sr, and a 1000-W mercury arc lamp, which
has a radiance of 1000 W/cm2-sr [Ready 1971]. The radiance (luminance) of
a source cannot be increased by any optical means; it can only be decreased
by stops or attenuation.

3 2
- PROPERTIES OF LIGHT EMITTERS
/-^

as'li ^Unction diodes emit light when forward biased. They are referred
"turn-on" v7 e t tming
i n d'°deS (LEDs)' Like anv other diode' a certain forward
ta§e mUSt be provided and tne curr
k* ''mited b° - ent through the device can
and
maxima * ^"^ resistor LED manufacturers specify the "turn-on" voltage
f r
° LEDs ^ allowable current. Light output versus current is nearly linear
Can
not tolerat ^"^ mEy be suPplied with the other specifications. LEDs
due to
heavy3d a . V£ry large reverse voltage because of low voltage breakdown
^^s em' Pmg tolerable reverse voltage is given in the specifications.
!" ' ngth ba l relatively monochromatic, although incoherent, radiation with
av ele

j!nd 0.55 ^m . Wl dths of 100 A. Typical wavelengths are 0.9 /j.m, 0.66 jam,
r lnfr
ared LED s^ **' K^' an^ green' respectively. The power output in watts
is usually given. The spatial distribution of the power output
52 Radiometry, Photometry, Optical Device Parameters, and Specifications
Cha'P. 3

LED

Figure 3-5 Irradiance versus angle for an

LED.

may be important and is given in terms of a plot of irradiance versus angle

such as in Fig. 3-5.
The output of visible LEDs may be given in terms of intensity\pr brightness.
Many LEDs, particularly those used in displays, approximate lambertian sources.
A lambertian source or diffuser emits radiation with an intensity variation given
by
/ = /o cos 9 (3-1)

where /<> is the intensity normal to the surface and 0 is the angle between the
radiating direction and the normal. Figure 3-6 depicts this situation. A reflecting
or transmitting surface of this sort is considered an ideal diffuser of light. Bright-
ness (radiance) is constant for any viewing angle because the projected area
of the surface as viewed from any direction decreases as cos 0. This factor is
important for displays intended for wide-angle viewing. It is left as an exercise
to show that 70 = TT~I X (total flux from the surface) for a lambertian surface.
So for an approximate lambertian emitter, any two of the three—/o. flux '
or brightness B—can be calculated if the third is given.

Figure 3—6 Variation of / with angle for a Lambertian surface.

 Quality Factors for Photodetectors 53
Sec- 3-3
UALITY FACTORS FOR PHOTODETECTORS

photodetectors that convert incident radiation into a measurable elec-

Soli"'s ^^ voltage are compared by the use of several defined parameters.
tr Ca
" . jest js the responsivity R , which is the ratio of the output signal (in
or volts) to the incident irradiance. The responsivity of a photodiode,
affll?
m n\t, might be 1 ju-a/mW/cm2 compared to 1 ma/mW/cm2 for a photo-
'stor The spectral responsivity R\ is the responsivity per unit wavelength
tra
, . generally plotted as a function of wavelength to provide the spectral
311
onse of the device or the peak value is given along with the wavelength
which it is maximum. For photodiodes, the quantum efficiency 17 is sometimes
ven It is simply the number of EHPs generated per incident photon. Responsiv-
ity is proportional to quantum efficiency.
The detectivity of a device, D, is the ratio of the responsivity to the noise
current (voltage) produced in the detector. Thus it is the signal-to-noise (SN)
ratio divided by irradiance. The units, therefore, are simply cm2/W. This parame-
ter is little used but serves to introduce the following parameter, D*, which is
the detectivity times square root noise frequency bandwidth, (A/)1'2, times square
root detector area, A112. The reason for defining D* in this way
D* = D(A A/)1'2 (3-2)
is to eliminate the dependence on noise bandwidth and detector area. Noise
power is proportional to the frequency bandwidth and detector area and so
the noise current or voltage will be proportional to (A A/)1'2. The units of
D* are cm3 Hz 1/2 /W or cm Hz1/2/W if power instead of irradiance is used in
the definition of R. D£, the spectral form of D*, may be given in addition to
D
* or in place of it.
Noise equivalent power (NEP) is the reciprocal of D and is therefore
simply the irradiance required to produce a signal-to-noise ratio of one. The
most widely accepted figure of merit is D*.
The values of any parameters mentioned depend on how they are measured.
1 e sur
^ ement conditions are usually specified. For example, D*(500°K, 400,
means that D* was measured with a 500° K blackbody source chopped at
'zed t Z ' Using l Hz bandwidth. Actually, the bandwidth is larger but is normal-

6 COns t era t ons

cx ' ' ' ' -- The main sources of noise in semiconductor
flu
ctuatj ectors are thermal noise and generation-recombination noise. Random
s m
by thermal motion of electrons will cause an rms current, given

The .auQo
rtns 1 fluct
cunJ^
iren
uations in generation and recombinations of EHPs cause an
t> given by
54 Radiometry, Photometry, Optical Device Parameters, and Specificatio

in = V4e/G A/
(3-4)
where G is the photoconductive gain, which will be discussed in a later h
A variety of noise sources is loosely referred to as one over / noise h ^'^
all are proportional to -v/A/// and hence are only important at low fre °aUSe
Because noise powers add directly, the total noise current is given by the ^
root of the sums of the squares of the individual noise currents. Note th ^
all cases the noise current is proportional to (A/)1'2. ln

Sources of radiation are also sources of noise. At low light levels stati t
variations in the rate of arrival of photons at the detector cause what is c 11 *
quantum noise. Fluctuations in the interference of waves are a source of ther
noise in a detector. Another source of noise arises from randomly varying canac
tance due to vibrations or breakdown of dielectric films.

3-4 SPECTRAL RESPONSE

The spectral response is generally presented as a variation of responsivity as a

function of wavelength. Spectral response characteristics for specific types of
detectors are described in Chapter 4, but some general comments are appropriate
here. Photodetectors fall into two general categories—thermal detectors and
quantum detectors. Examples of thermal detectors are calorimeters, thermopiles,
bolometers, thermistors, and pyroelectric detectors. In all these devices the inci-
dent radiation is converted to heat, which, in turn, causes a temperature rise
and a corresponding measurable change, such as an emf generated in a thermo-
couple or a current pulse in a pyroelectric detector. Detectors of this type tend
to have a flat spectral response over a wide wavelength range.
Examples of quantum detectors are photomultiplier tubes, semiconductor
photodiodes, and phototransistors. In this type of detector photons produce a
measurable change directly, such as the emission of electrons in photomultip w
tubes or the generation of EHPs in semiconductor photodiodes. Photoemissi«
surfaces for photomultipliers have been designed with fairly flat spectral respo
but have a sharp cutoff at the long wavelength end where the photon ene^
is insufficient to overcome the work function of the material. Semicon
detectors have fairly narrow spectral response, peaked at the wavelengt
spending to photons with energy equal to the band gap energy.

3-5 FREQUENCY RESPONSE

The frequency response of a photodetector refers to its ability to

an incident light beam chopped or modulated at various frequencies.
state detectors behave like lowpass filters and as such their respons
be written
 BibM°8raPhy
Chap- 3
D

=
R T-l i
o 22M1/ 92 (3-5)
(1 + orr ) '
• the responsivity at zero frequency and T is the time constant.
where K ° ^.equency is/ c = (27TT)-1, where R = Ro/^/2. The time constant
The cuto v T = rC, where C is the device capacitance and r is device
ma y be gi ^ c j rcu j t resistance. Clearly an inherently fast device (wide fre-
resistance " can ke sjowecj down by use of a large circuit resistance.

PROBLEMS

vl Calculate the solid angle for a HeNe laser that emits a beam with a full-angle
divergence of 1 mrad if the diameter of the emitted beam is 2 mm.
3-2 The total flux (power) from a lambertian surface is 2 mW. Calculate the intensity
at normal incidence and the brightness (radiance) if the surface is 1 mm square.
3-3. Calculate the noise current in a 100-fl resistor at room temperature for a bandwidth
of 1 kHz.
3-4. The dc responsivity of a solid-state photodetector is 10 mA/mW with a time
constant of 10~7 s. What is the responsivity at 10s Hz?
3-5. Name and briefly discuss the reasons for three types of noise common to photodetec-
tors.
3-6. Name the figures of merit used to qualify photodetectors. Write out the equations
for them and explain.

REFERENCE
EA
DY, J. p., Effects of High-Power Radiation. New York: Academic Press, 1971.

BIBLIOGRAPHY

son, A., Photometry and Radiometry for Engineers. New York: John Wiley
« sons, Inc., 1974
2
Williams W <: A Beck
Scienf ' * '' ' 'und, Optics: A Short Course for Engineers and
Y rk: J hn Wiky & S nS> InC 1972
^ YDriscoli ° ° ° " '
ork°vi ^' °' k*1-)' and w- Vaughn (assoc. ed.), Handbook of Optics. New
3
^ Cha , CGraW-Hi" Book Co., 1978.
v-H^pjx-^l . '

p
rinti ng 'p'' (ed ' )l °Ptoelectronics: Theory and Practice. The Netherlands: Holland
J 5 S Texas Ins
~ RCA, EI ' truments Ltd., 1976.
ectr
1968. °-°Ptics Handbook. Harrison, NJ: RCA Commercial Engineering,
 chapter

Light Detection
Devices

This chapter describes a variety of light detection devices that are frequently
used in laser applications. The discussion includes energy and power measure-
ment as well as a brief description of imaging devices, such as photodiode
arrays and charge-coupled devices (CCDs). Some simple detector circuits are
also briefly described.

4-1 CLASSIFICATION OF DETECTORS

Light detectors can be placed in either of two classifications: thermal or quan •

(sometimes referred to as photon). Thermal detectors are those in which ^
absorbed light energy is converted into heat that produces a temperature
that, in turn, causes some detectable output or variation in a property
detector. Quantum detectors are those in which the absorbed light direct y
a property or produces a measurable output. ,orne-
Examples of thermal detectors are thermocouples, calorimeters, ffli.
ters, and pyroelectric devices. The radiation causes a resistance change m ^^
ters; either a voltage or a current pulse is produced in the others. In
it is a change in temperature that produces the measurable phenorne ^ ^<v
Quantum detectors are exemplified by photoemissive devices, su . tors.
tomultiplier tubes (PMTs), photodiodes, photoconductors, and phototra
A resistance change is directly produced in photoconductors by the
of light; in the others a measurable current or voltage is generate
interaction of the absorbed light with electrons.
 , Thermal Detectors 57
Se c. 4-z
thermal detectors are slower and have a broader spectral re-
Genera y m detectors. There are notable exceptions to this statement,
sponse than ^^ly with respect to the speed of response.
however* pa

4_2 THERMAL DETECTORS

d thermal detectors are based on a change in property or output that

AS
T Vom a temperature increase arising from the absorption of light. In
reSU S
H ices a certain degree of "thermal inertia" is inherent because heat
mos
tion is involved. Consequently, the response time of such detectors tends
COI U
be quite slow unless the mass is kept small, as in some thin film devices.
Most thermal detectors use a blackened absorbing surface and frequently exhibit
relatively flat spectral response from the uv to the far ir. Window materials
placed in front of the absorbing surface and the absorbing coating determine
spectral response.

Resistive heater

TCs or thermisters

Insulation

Figure 4-1 Cone-type calorimeter.

° n 'y a few eters- Although many different types of calorimeters exist,

Sever T6 fly descrit>ed here to elucidate the principles involved.
f r
° high abs ^ °"meter designs use an absorbing cavity that is suitably coated
of
the light e rtPtl°n and mav have a conical shape to produce multiple reflections
r cone
° 'herrnistor
~"~i ll\\^tf\f ii ^ - The
~ --•«-temperature
vv^nij^^i. tinji \,change
x^iiciiicvis
iosensed
k>\^iuv^^j.by
u y thermocouple
m^i im^/^v^iAjjlCS
Ca 1 C d n t le
use of thei i" ' ' absorber, which is usually aluminum or copper
Sente
d in Fig IT86 ttlermal conductivity. Such a device is schematically repre-

sa
ceramic resistor with a large temperature coefficient of resistance.
58 Light Detection Devices
Cha

When light is absorbed by such a device for a period of time th

energy causes a uniform temperature increase in the highly conduct' S°r^e('
Cu. This temperature increase is sensed by the thermocouples or th °r
St0rs
and may be displayed as a voltage on a meter or digital readout c rx'
is carried out by irradiation with a standard source or by supplying J^'0"
quantity of heat through a resistive heater embedded in or attached °*n
absorber. Such calorimeters can measure power by being allowed to rea h i,
mal equilibrium with their surroundings or coolant. The detected stead ^
temperature rise is then a measure of the incident power. Calibration w ^
occur in the same way as for energy measurement.
Because the response time of these devices may be several seconds
possible to measure high power levels by chopping the beam and allowing onl
a small fixed percentage (e.g., 1 to 3%) of the power to enter the calorimeter
In addition, because of the inherent "thermal inertia" of such devices, the output
signal will remain relatively steady. Power levels of thousands of watts can be
measured in this fashion. Figure 4—2 is a photograph of a commercial calorimeter
that uses an absorption cavity for high-power CO2 laser power and/or energy
measurement.
Some calorimeters are water cooled and the difference in the inlet and
outlet water temperature is used as a measure of the power or energy absorbed.
Because heat capacity C is defined as heat absorbed per unit mass per unit

jm>
1K9P
^fip

Allied
Figure 4-2 Absorption cavity calorimeter. (Courtesy of Apollo Lasers, an
Company.)
 59
Thermal Detectors
4-2
hange, Eq. (4-1) gives the power P absorbed in terms of heat
erature C
te mP era ture
change AT, and mass flow rate dm/dt:
->P aCity> HO ^
u (4-1)

" t's nest" calorimeter uses wire, such as copper, randomly tangled
h' hlv reflective container. The tangled mass (or mess if you prefer)
nside a ^ nearly 100% absorption of the incident radiation and the tem-
'se is sensed as an increase in resistance of the wire. Absolute calibration
H' ed by dissipating a known amount of electrical power in the wire.
18 3C
a thermally induced resistance change is used, this device could also
bewailed a rat's nest bolometer.
When thermistors or bolometers are used to sense temperature changes,
it is possible to place two identical calorimeters in a bridge arrangement for
very sensitive power or energy measurements. Ambient conditions are the same
for both calorimeters except that one is irradiated and the other is not. This
technique is depicted in Fig. 4-3. The variable resistance is adjusted for a null
reading on the galvanometer. When the calorimeter is illuminated, the bridge
is unbalanced, thus causing a deflection of the galvanometer that can be calibrated
in terms of power or energy for maximum deflection. The identical dummy
calorimeter tends to provide cancellation of resistance change effects due to
variations in environmental conditions, which should nominally be the same
for both calorimeters.

Thermopiles. A thermopile consists of a number of thermocouples

connected in series. A thermocouple is a junction of two different metals or
semiconductors. Because of the work function difference between the materials,
acurrent is generated when the junction is heated. Thermocouples use a reference
tem° 'On and the voltaSe difference between the junctions is a measure of the
perature difference between them. The reference voltage in modern systems

•4-3
Brid;86 netw
°rk for calorimeter.
60 Light Detection Devices
Ch a
is established electronically so that only one junction, or probe is
e
Such devices are extremely common in temperature measurement. Wh d.
J
tion of dissimilar materials, with an appropriate absorbing coating, js [T j
a temperature increase occurs, thus generating a voltage that is a m '
r
the power absorbed. °f
The voltage produced in metallic thermocouples is only a few mi
per degree change in temperature and so it is common practice to a
several couples in series so that all the junctions of one polarity are irrad' "^
In this way, their voltages add to give a larger response, the reference m n t
being masked from the radiation. Figure 4-4 is a sketch of such a therrno I $
The thermopile may be incorporated as one leg of a bridge circuit with
identical dummy device in the other leg to improve sensitivity and immunit
to environmental fluctuations. Thin film thermopiles can have response times
of the order of microseconds with a flat spectral response.

Bolometers. Bolometers were discussed earlier and so little will be

added here. Such devices are either metallic or semiconductor resistors in which
a thermal change in resistance is induced by absorbed radiation. The change
in resistance causes a change in current in a circuit and hence a voltage change
across the bolometer or another resistor in the circuit. Bolometers, of course,
can be used in a bridge circuit with a dummy bolometer in one leg.

Pyroelectric detectors. Pyroelectric detectors are made from materi-

als that have permanent electric dipole moments associated with them, such
as LiNbO3, triglycine sulfate (TGS), and polyvinylfluoride (PVF). These materi-
als are ferroelectric, meaning that permanent dipole moments are associated
with the unit cell, in the case of crystalline materials, or the basic molecular
unit in the case of polymers. These dipoles are all aligned parallel within domains
analogous to the domains in ferromagnetic materials. The chief difference
tween pyroelectric ferroelectrics and other ferroelectrics is the difficulty of po inj>
pyroelectrics. In order to get the saturated polarization of all domains

Irradiated /
junctions

Figure ' Thermopile.

 Quantum uetectors 61
4-3
Sec
Radiation

t M t t t -Electric
dipoles
t t t t t I

Figure 4-5 Simplified pyroelectric detector scheme.

the same direction, the material must be placed in a strong electric field at
1
elevated temperature and the polarization cannot be reversed at normal
temperatures by application of an electric field. The total dipole moment of
pyroelectric materials is, however, very sensitive to temperature fluctuations.
When absorbed by a pyroelectric detector, light is converted to heat,
thereby causing a temperature rise and hence a change in polarization. Figure
4-5 is a simplified pyroelectric detector scheme. Charges absorbed on the surface
neutralize the polarization surface charge. When the polarization is altered due
to a temperature change, charge must be transferred through the external circuit,
for it cannot be desorbed from the surface quickly. The speed of response for
such devices is very good, in the nanosecond range, because the speed is not
thermal conductivity (thermal inertia) limited. The temperature change causes
a change in dipole moment that responds extremely fast.
In the circuit of Fig. 4-5 R should be less than the resistance of the
elector, which is quite high because it is basically a capacitor. The voltage
developed by the detector is given by

V(t) = AR7 (4-2)

at
c Pyroelectric coefficient or polarization change per unit temperature
p and A is the irradiated area.
detectors have a flat
y
trT spectral response to long wavelengths
f
fre
quencv 6 °r cry°gemc cooling required by other ir detectors and their
a
c
^^ °bCyS Eq' ^~^' is Vety g°°d with bandwidths around
um
iuar e Cen tj incident power is limited to a few tens of watts per

<*ANTUM DETECTORS
.1
res
P°nse as a direct result of the absorption of radiation
V Ces UantUf
' can cW 3 n or photon detectors and, indeed, some photoemissive
tect si
ngle photons.
62 Light Detection Devices

Photoconductors. When a photon is absorbed by a semi

the probability that an electron-hole pair will be produced is quite h' U°tor'
quantum efficiency 17, which is defined as the ratio of conduction i
produced to absorbed photons, is usually 0.7 or higher at the peak of the °tr°ns
response curve. A photoconductor is an intrinsic or extrinsic semico V^5'
in which the change in conductivity resulting from the absorption of r ^^
used as a measure of the incident power. Unlike a bolometer, temperatur h '$
has nothing to do with the conductivity change. Conductivity is given b ^

where n and p are the number of electrons and holes per unit volume res
lively, and /j,e and /j,h are the mobilities of electrons and holes, respectively
The creation of electron-hole pairs in an intrinsic conductor alters both n and
p. The long wavelength cutoff of the device is given by \

c-

where Eg is the width of the energy gap, h is Planck's constant, and c is the
speed of light. The signal current, assuming the dark resistance is very large,
is given by

(4-5)

where A is the irradiated area, A/ is the bandwidth, tc is the carrier lifetime,

and tt is the transit time (time for an electron to traverse the device). The
ratio tc/tt is called the photoconductive gain because 4 can be increased by
designing the photoconductor to have long carrier lifetime and short transi
time. The main source of noise in photoconductors is generation-recombma i
noise, which also increases with tc/tt; so the signal-to-noise ratio is not impr°
by increasing the photoconductive gain. . |V
Narrow bandgap photoconductors, such as HgCdTe, are used extensi <
for ir detection but must be cryogenically cooled to minimize the effect o
mally generated charge carriers.
i nny

Photodiodes. Semiconductor diodes were discussed in chapterductor

so the basic theory of diodes need not be considered here. When a semic ^^
diode is used as a photodetector, an internal current is produced a ,g
of the absorption of photons and the creation of electron-hole Palf;L g nlofe
Many types of photodiodes and modes of operation are possible-
common types are discussed next. . nCtioO |S
ne e
All pn- junction devices respond to incident radiation if t J g{ th
close enough to the surface for light to reach it and if the waveleng , ^
incident radiation is less than the cutoff wavelength. Longer wave
 Quantum Detectors 63
4-3
Sec.

Figure 4-6 Energy level diagram for an unbiased photodiode.

to penetrate farther and short wavelengths are absorbed nearer the surface.
The uv response of photodiodes is poor both because EHPs generated near
the surface tend to encounter surface recombination centers and because most
of the energy of the uv photons is wasted, for only a small fraction is needed
to produce an EHP.
The process by which current is generated can be understood by reference
to Fig. 4-6. It can clearly be seen that an EHP generated in the depletion
region will tend to produce a current as the election and hole slide down their
respective potential hills. If the diode has a resistor connected across it, a current
is produced as long as light is incident on the junction. This situation is depicted
in Fig. 4-7. This is the photovoltaic mode of operation, for a voltage is produced
across the resistor and power Vdld is produced.
Actually, EHPs need not be generated exactly in the depletion region.
EHPs generated within one diffusion length on either side of the junction will
contribute to the current. The diffusion length is the average distance that an
lerTh0" °r h°le travels in the minority regions before recombining. The diffusion
diod f°r h°leS iS substantially less than that for electrons; so most diffused
I5re made witn a thin p-type region on top (toward the incident radiation).
wav Th ldeal resP°nsivity of photodiodes is easily deduced in the following
he
Photocurrent is given by
dn

-tx-
on Photovoltaic mode of opera-
-ww-
6* Light Detection Devices
Ch
ap. 4

where dn/dt, the rate at which photons strike the detector, is given ,
Here Pi is the incident power. The responsivity in amperes per watt 15' ^f.
then
R=-=— = 8.1 X 105X. (A/W)
^ Ac
which yields a typical responsivity of about 0.6 A/W at X = l.o n m
Photodiodes may have areas of 1 cm2 or greater for high sensif
position detection, but this factor increases noise and capacitance. Many h °r
diodes are very small, having an active area of less than 0.01 cm2. The respn ° °
for small-area devices is more conveniently defined as output current r>e ^
irradiance in watts/cm2. A typical responsivity is 1 mA/W/cm2.
The major source of noise in photodiodes is shot noise, which is caus d
by the tendency of electrons and holes to "bunch up" or to cross the depletio
region in a random manner.
The short wavelength response of photodiodes can be enhanced by placing
the junction closer to the surface. Overall response is then sacrificed, of course.
To achieve improved response at all wavelengths in the spectral range of the
device, pin diodes are used. Here an intrinsic layer (actually a very lightly
doped region) is placed between the p and n regions. This step improves quantum
efficiency by providing a larger depletion region and increases speed of response
by decreasing junction capacitance.
Two common silicon photodiode designs are the planar-diffused and
Schottky barrier diodes. Figure 4-8 contains sketches of both types. The planar-
diffused type may be made as a pin or ordinary pn diode, depending on the
spectral response and speed requirements. The Schottky barrier diode has a
superior uv response, due to the fact that the depletion region extends to the
silicon surface. Large-area Schottky diodes are less noisy than their planar-
diffused counterparts but have lower power-handling capability.
The diode equation can be generalized to account for photogenerated cu
rent. If a short circuit is placed across an illuminated diode, the resulting i
current is a reverse current and is called the short-circuit current, /sc ^
resistor is placed across the illuminated diode, then a forward-bias voltage
produced and the diode current is given by
(4-7)
Id = Io(eeVlkT - 1) - /so
Note that the open-circuit voltage, Voc, is given by

„„_*!(,„£+,)
e \ Io / ,
u
Figure 4-9 is a sketch of diode current versus diode voltage for vari° ^ at;n?
light levels. When a diode is operated in the photovoltaic mode, it l* . cird>!l
in the fourth quadrant of the current versus voltage graph. The jgs- ^
current varies nearly linearly with incident power over about six
 Quantum Detectors 65
4-3
Sec-
Schottky barrier
Si02 diode

Gold 1
\ Depletion region /

(a)

Planar diffused diode

SiO,

T
Al

(b)

Figure 4-8 Silicon photodiode designs.

'icident Ordinary
light power diode

4-9
Diode current versus diode voltage for various light levels.
66 Light Detection Devices
Ch,iaP. 4

a circuit that provides essentially zero voltage drop across the diod
v
substantially linear operation. 'de$
Maximum power output in the photovoltaic mode, (F d / d )
by a point determined by inscribing a rectangle of maximum area a g'Ven
in Fig. 4-10. The process involves matching the load resistance to th § °Wn
resistance. Unfortunately, the diode resistance varies with incident lieh
in the photovoltaic mode. Consequently, no single resistance will work f ^
light levels and an optimum load based on operating conditions must be d
mined. Solar cells and nonbattery-powered light meters operate in this fash "
When a photodiode is reverse biased, the operation may be linear
10 or 11 orders of magnitude change in light level and the junction capacit
is reduced, thereby decreasing the response time in some diodes to less th
1.0 ns. Reverse-bias operation is referred to as the photoconductive mode but
it should not be confused with the term photoconductor;\photoconductors are
not even diodes.
Avalanche photodiodes are reverse biased at a precisely controlled voltage
that is slightly less than the breakdown voltage. When light strikes the diode,
creating EHPs, the avalanche process begins and continues as long as the light
is incident on the diode. The advantage of this type of diode is the multiplication
factor. It may be as great as 1000, which provides an internal gain and corre-
spondingly high responsivity. The response time of avalance diodes is also very
good, about 1.0 ns.

Phototransistors. A phototransistor is just what the name implies—

a transistor used as a light detector to take advantage of the internal gain of
the transistor. Light incident on the base-collector junction of a bipolar transistor
produces an internal current that acts as the base current of an ordinary transis-
tor. Consequently, no base lead is required unless external biasing is necessary.
The spectral response of silicon phototransistors is similar to that of silico

Figure 4—10 Maximum power output point.

 , Quantum Detectors 67
- 4-3
. resp0nsivities are about three orders of magnitude larger, typi-
, but /-ly/cm2 The response time of phototransistors is microseconds
cally around 1 A/

a nd slower- totransistors are built in a Darlington configuration to take ad-

^Tthe increased current gain realized from such a design. The light is
vantage o jnge on the collector-base junction of the first transistor and
allowed o nrovides additional gain. The responsivity of these devices is
the
.T?0 A/W/crn' or h.gher.
C 1C
^ Ph to-FET devices are available for applications in which the very high
stance of such transistors is required. Photo-SCRs are also available
for h'ght-activated switching applications, such as fire alarms.

Photon drag. Detectors using the principle of photon momentum

ansfer referred to as photon drag detectors, are not extensively used in manufac-
turing settings, but they are potentially useful for on-line monitoring of CO2
laser power and measurement of the time dependence of CC>2 laser pulses.
The momentum of a photon can be determined by equating the mass-
energy equivalence relation to the photon energy
me2 = hf
so that momentum, p = me, is given by

,-H
In such materials as germanium and tellurium, which are transparent to 10.6-
'im radiation, a sufficient amount of momentum is transferred from the photons
o electrons or holes to cause a measurable open-circuit voltage across a crystal
^ rough which the radiation is passing. This open-circuit voltage is the result
holes or
electrons being pushed or "dragged" along with the radiation.
time may be less than 1.0 ns with responsivities of lO'8 to 10'6

surface e ° °mu'tip''ers- If the energy of photons incident on an absorbing

trie effect0^ ^ W°Fk function2 ofthe electrons in that surface, the photoelec-
by
direct tr* °bserved- The incident photons eject electrons from the surface
a
Photoelec't r °f Cnergy' Collection of these electrons at an anode produces
Photo<;mi .Ven° current that is a measure of the light power incident on the
p. cathode.
are
simply cetn °t0rS that use a sin§le photoemissive surface and a single collector
bUt hav
e lowa phototu
bes or photodiode tubes. Phototubes are quite fast
responsivities because there is no internal gain mechanism. In
2
w
nctlon
'"finity. is the energy required to remove an electron from the surface of the material
68 Light Detection Devices

photomultipliers a series of "dynodes" are used to provide multinr

gain. The electrons ejected from the cathode are accelerated by a 'Ol1 °r
100 to 200 V to a dynode where they, on impact, eject many secondary l ^ of
which are, inin turn
ir»Vi <Arf* turn, Qr*r>£»lf»rQtf»H
accelerated tn to another
Qnr\tnp»r dynode
t\\ir\r\c\(* and
Qnrl so
cr» on.
r\n Tubes
T,,l ma OtK

ten or more dynodes and produce a multiplication factor of one m'lr ^l

more. Response times (transit time of the electrons) of less than 1 n " °r
achieved and single photons can be detected, thus leading to the phrase h ^
counting even though not every photon incident on the emissive surfac ' °n
e
tected. "
A variety of photocathodes covers the spectral range from 0.1 to l
jxm. Examples are (a) Ga-As, which has a fairly flat spectral response f
0.13 to 0.8 urn with a peak responsivity of around 35 mA/W, and (b) As-0
Cs alloy, which covers the spectral range from 0.3 to 1.0 pm, although th
responsivity varies significantly.
Many photomultiplier designs exist. Figure 4-11 is a schematic representa-
tion of one type of design. Photomultiplier tubes are extremely fast and sensitive
but expensive and they require sophisticated associated electronics to operate.

Position detectors. A variety of silicon photodiode position detection

devices is available. Such devices fall into two categories: multidiode units and
lateral-effect diodes.
Multidiode units consist of either two or four separate, but identical, photo-
diodes placed on the same silicon wafer but electronically isolated from each
other. The bicell and quadrant detectors are schematically depicted in Fig.
4-12. When a symmetrical laser beam is centered on either detector, the output
of each diode is identical. If the beam moves more onto one diode than the
other(s), the signal from that diode increases while the signal from the other(s)
decreases. This change in output can be used to measure the displacement o
the laser beam. Measurements of beam displacement of as little as 1.0 fun a
claimed to be possible. One-directional displacements can be measured w
the bicell whereas two-dimensional displacements can be measured with qua

Window

Cathode Dynodes Collector

Figure 4-11 Photomultiplier tube.

 Quantum Detectors 69
4-3

Isolation Quadrant cell

bi-cell
(a) (b>

Figure 4-12 Photodiode position detectors.

ctors The linear displacement of the beam on the detector can aisu uc
easily related to angular displacement of the beam.
The beam profile generally used with position detectors is a filtered Gaus-
sian distribution. Consequently, the output of the diodes is not linearly related
to the amount of beam displacement. Detectors of this type can be made either
by the planar-diffused process as either pn or pin diodes or by the Schottky
barrier process.
The lateral-effect diode position detector avoids the problem of nonlinear
output as a function of beam displacement. Figure 4-13 is a schematic diagram
of a lateral-effect diode. The output current from terminal S is ideally given
by

-(-I) (4-10)

/„ is the actual photogenerated current across the depletion layer at

he
point of incidence of the beam. The farther the beam is from the output

S-L

ure r
Lateral-effect diode. (Adapted with permission of Laser Focus.)
 70 Light Detection Devices c
s.,

I
Figure 4—14 Silicon photodetectors.

terminal, the more electrons cross back over the depletion layer and recombine
at the surface. This type of position detector is slower and noisier than multicell
types, but the change in output with beam position is very linear and the beam
irradiance profile is irrelevant. Beam displacements as small as 0.01 /xm can
be detected with silicon diode position detectors and this factor allows very
small angular changes to be detected.
Figure 4-14 shows a variety of silicon photodetectors that can be used
for power measurement, centering, and position measurements.

4-4 DETECTOR ARRAYS

A thorough discussion of detector arrays is well beyond the scope of this boo
A brief description of the basic types of arrays is included, however, becau
of their usefulness in lower power laser applications. Linear and area arr
using photodiodes, charge injection devices (CIDs), and charge-coupled de
(CCDs) are available. Detector arrays are frequently used in systems design
to measure part dimensions and surface location.

Photodiode arrays. As the name implies, photodiode arrays

of either one-dimensional or two-dimensional arrays of photodiodes. .^
of each diode is proportional to the amount of light incident on it, pa cing
is possible with a resolution dependent on the cell (pixel) size and cen e
of the pixels. In this context, a pixel refers to the basic light detection
The light-sensitive area is about half the total pixel area. These
are usually back illuminated so that there are no contact obstructs •
4-15 is a schematic illustration of a linear photodiode array. ^r aS
type may have from 64 to 2048 elements on center-to-center spacing
 Detector Arrays 71
4-4

Digital shift register

*- Video

-HI M_ H_

- Common

Figure 4-15 Linear photodiode array. (Adapted courtesy of EG & G Reticon Corp.)

as 2.5 jim. Light incident on the silicon photodiodes produces charge which
is stored in the parallel capacitor until a voltage from the shift register is applied
to the gate of the FET. The charge is then dumped onto the video line so
that a sequence of pulses whose amplitudes are proportional to the amount of
light energy incident on the diode during the framing period is read out. The
slower the clock rate, the more charge can be stored and the more sensitive
the device is to light level. Sampling rates may be as high as 40 MHz. These
devices also come with CCD-analog shift registers. In this version a voltage is
a
Pplied simultaneously to the gates of all the FETs, the charge stored in each
capacitor is dumped into the CCD shift register, and the charge packets are
the" fead out serially as with the digital shift register. In both types of arrays
l output is a discrete time analog of the light irradiance distribution across
ne array.

In th WO " dlmenSi0nal di°de arrays with UP to 256 x 256 elements are available.
r a shift
ow by register
egsersequentially
s e q u e n a y applies
a p p e s pulses to the FET gates
CCD sh'f
s' * P'ng *he charge from each diode iin a row into one or two
r two V'H reglsters' which then serially output the charge packets onto one
60 hnes
- Pixel rates of 10 MHz can be achieved.
Ch
°s) cap^6 ° OUpled devices- CCDs use metal-oxide-semiconductor
h t l Th
e ch't0rS dlrectly to store packets of charge produced by the incident
a Par
ticula gC 'Dac'cets> wmch represent an integration of the light incident
° °ie ca Capacitor' are then serially shifted out of the CCD by transference
°r to another. Simplified versions of two schemes for achieving
.
imagers the charge is collected under the gates of
 Light Detection Devices

v////y/////VMtfw//%^^ -SiO,

(a) Three-phase charge-coupled device

SiQ, Al
Polysilicon

(b) Two-phase charge-coupled device

Figure 4-16 CCD linear imagers. (Adapted with permission form Solid State Tech-
nology, Technical Publishing, a company of Dun & Bradstreet.)

capacitors that have a negative gate voltage applied—for instance, PI-

result, a depletion region is created under the 0i gate oxide into which o
generated by incident photons are attracted and stored. The number o
in the packet is proportional to the number of photons incident during
period of time that 61 is negative. The charge packets can then be
to the 02 capacitors by reducing 81 and making 02 large and then to cap
03 by making 0! and 02 small and 03 large and so on until the charge
out as an analog pulse similarly to the output of a photodiode array^ ^.^
tively, the packets stored in the imaging capacitors can be transferre ^ ^
to a parallel line of MOS capacitors that are masked so that light cann ^^
them and then the charge is transferred to the output in the manner ^
To prevent the trapping of electrons by fast surface states at ^^ th«
interface, a "fat zero" (a small number of holes) can be cycled thr ^ ^
capacitors continually to keep these states filled. This process mllUon to pu'
degradation of the signal. Another technique is to use ion implanta 1 ^^
a shallow p-type buried layer below the oxide; doing so pushes t
 Photodetector Circuits 73

the bulk Si away from the surface states and keeps the signal
layer *>wn "j(.°m the surface state traps.
away
r.5rge f er jn the right direction in three-phase devices is determined
Charge
etry in the application of the clock voltages, 0 t , 02, and 03. Charge
•r\ theasy m ^ direction can also be accomplished in a two-phase system
transfer m built-in asymmetry. Figure 4-14(b) illustrates one way of doing
by intro u ^j^ness under the poly-Si gates is less than under the Al gates;
S Th
° h °oxide capacitance for the poly-Si gate capacitors is higher than that
thUS l
Al gate capacitors. Consequently, a smaller voltage occurs in the depletion
6
beneath the Al gate than below the poly-Si gate capacitors. This situation
reglC
T considered in terms of potential wells: the wells under the Al gates are
Tways less deep than the wells under the corresponding poly-Si gates. If 0i is
I ree charge is accumulated in the 0i capacitors. As 0! is decreased and 02
increased, charge can only move to the right because of the potential barrier
that exists on the left due to the shallower well under the Al gate.
Transfer efficiencies, the amount of charge transferred from one MOS
capacitor to the next, of over 99.999% are possible. Center-to-center spacing
of pixels is typically 12.7 urn and a dynamic range3 of 1000 to 1 is common.
The elements in CCD imagers can be made about as small as the limits of
photo or electron beam lithography will allow. Very large linear and area arrays
that can operate at both high- and low-light level conditions are possible. It
has been reported that 200 million pixels per second can be read in large displays
through the use of multiple readout taps.

Charge injection devices. CID imagers are similar to CCDs in that

photon-generated charge is collected and stored in MOS capacitors. In two-
dimensional imagers each pixel consists of two side-by-side capacitors. When
a gate voltage is applied to one, it gathers and stores charge. The voltage is
e transf
a " erred to the second capacitor, which is masked from the light, by
Application of the voltage to it and removal of the voltage from the first. Readout
beenCs°!|npllShed by sec*uentially setting the voltages to zero for a line that has
the v it placing all the second (storage) capacitor voltages high. When
a
detect hi 8°eS t0 Zer°' t'ie Cnar8e 's injected into the substrate, thus causing
^ension \ °Urrent Usin§ this technique, it is possible to address a two-
c
°lurnn and aiTay on an individual basis (random access) by applying appropriate
Ca
Pacitors 1°* Vo'tages- ^ linear array does not require the second set of
unless individual addressing is necessary.

PH
°TODETECTOR C.RCUITS
s fbj- tjj
a
nd Ccr>Peratl°n an^ amPnncati°n of photomultiplier tubes, photodiode
°r CID arrays are beyond the scope of this book. Such circuitry
RaHr\ „/•, .
'ghest to low«t detectable light levels.
74 Light Detection Devices

Relay -~

Figure 4—17 Light-acti- Figure 4-18 Phototran-

vated relay. sistor amplifier.

can be found in the applications notes provided by the manufacturer. A few

simple but useful circuits for photodiodes, phototransistors, or photoconductors
are described, however.
A simple light-activated relay circuit is depicted in Fig. 4—17. Phototran-
sistors can handle sufficient current to activate small relays. When light strikes
the phototransistor, turning it on, the relay is opened or closed, depending on
its normal state. When the light is removed, the relay returns to its normal
state. The diode may be placed in parallel with the relay to protect it from
possible voltage spikes.
A simple phototransistor circuit of the sort shown in Fig. 4-18 can be
used for switching or detection operations and even for light-modulation applica-
tions in which true fidelity of the modulation wave form is not critical. Voice
communication on an amplitude-modulated laser beam, for example, can

Figure 4-19 Transistor amp '

for a photodiode.
 Photodetector Circuits 75
4-5

Figure *-*1
photod,ode operat.on.

lished by taking the output of this phototransistor amplifier and feeding

it into an audio amplifier.
A transistor circuit that can be used to amplify a photodiode output is
given in Fig. 4-19- The photodiode in Fig. 4-19 may be replaced by a photocon-
ductor.
An operational amplifier circuit for photodiodes that provides great linear-
ity and high speed is depicted in Fig. 4-20. In this circuit Ec and R2 may be
zero. EC is provided to reverse bias the diode, which produces faster and more
linear operation. The resistance RI may be chosen to balance offset current
resulting from diode dark current. The output voltage for this circuit is given
by
v0 = -iPRi (4-11)
In applications where large light level changes may be encountered, a
logarithmic amplifier, such as the one given in Fig. 4-21, may be useful. In
is case, the op amp must have an extremely high input impedance. It can
e obtained by using an FET input stage. Here the op amp is used in a noninvert-
mg am
P!ifier arrangement where

ecause of the very high input impedance, Id =* 0 so that

•Ft*,
 Detection

I Figure 4-22 Integrating^,.

enc
O measurement. rgv

which leads to

Thus the output voltage is given by

fcr
v 0 = /,1 +, —-
* AI — iIn I/,1 +, /—«
(4-12)
\ R2/ e \ /„
If energy measurement is desirable, a simple op amp integrating circuit (shown
in Fig. 4—22) may serve the purpose. In this circuit the output voltage is propor-
tional to the integral of the input voltage, which is, in turn, proportional to
the power incident on the photodiode. Therefore the output voltage is a measure
of the energy absorbed by the diode during the time interval of the integration.
Equation (4-13) is the relation between output voltage and input voltage.

v. = — I *dt (4-13)

PROBLEMS

4-1. (a) Name the two classes of photodetectors, describe them, and explain how
work, jjjed
(b) List three types of detector in each class and explain why they are c ass
that way.
4-2. What should the mass flow rate in a water-cooled calorimeter be if 1C0
change is to occur for each 100 W of power absorbed?
4-3. Assuming the specific heat of a pyroelectric detector is C and its mass m
an expression relating output voltage, area, series resistance, pyroelectrtc
the given quantities, and the power P incident on the device.
P
Answer: V = AR7 —
m€
h'<
This assumes all the power is absorbed by the detector and none is los
tion, convection, or radiation. c
_ i X 10
4—4. (a) Calculate the conductivity of an «-type silicon with N — 2 *
 77

late the long wavelength cutoff for silicon.

* £9 = l . l e V
the ideal responsivity in amperes per watt for a photodiode operating

*"5' at^ = °'66 ^m'

the open-circuit voltage at room temperature for a photodiode if I0 =
a
*"*' j /sc = 100 mA.
three different types of semiconductor photodetectors and explain their
operation.
'be two distinct types of diode position detectors and compare them.
o Name two types of detector arrays and briefly explain how they operate.
„ (a) Design a sample op amp circuit that will provide linear output and an output
of —1 volt per microampere of photodiode current.
(b) Calculate the input and output resistances of your amplifier based on the
data for a 741 op amp given in Chapter 2.

BIBLIOGRAPHY

4-1. Driscoll, W. G. (ed.), and W. Vaughn (assoc. ed.), Handbook of Optics. New
York: McGraw-Hill Book Co., 1978.
4-2. Streetman, B. G., Solid State Electronic Devices (2nd ed). Englewood Cliffs, NJ:
Prentice-Hall, Inc., 1981.
4-3. Moss, T. S., G. J. Burrell, and B. Ellis, Semiconductor Opto-Electronics. New
York: John Wiley & Sons, Inc. 1973.
4-4 Stimson, A., Photometry and Radiometry for Engineers. New York: John Wiley
&
Sons, Inc., 1974.
5
- The Engineering Staff of Texas Instruments, Inc., Semiconductor Group, The
Ptoelectronics Data Book for Design Engineers (2nd ed.). Dallas, Tx: Texas Instru-
ments, Inc., 1975.
pp"1?1)0' ° F
'' "Char8e-CouPled Devices," Scientific American, February 1974,

tio»!n\T '' and ° Buss

(eds-)' Charge Coupled Devices: Technology and Applica-
W York: IEEE p
4-g. Gener ress, 1977.
an
d Ra d eCtri° C°'' °Ptoelectromc Systems Operation, CID, Theory of Operation
^9 Stotl " °m ACCeSS DescriPtion- Syracuse, N.Y.: General Electric Co.
D and E J McLellan
«ectors " n tlca ' ' ' "Developments in High-Speed Pyroelectric
' °P > Engineering, 21, No. 3, May-June 1981, 469-471.
 chapter 5
Introduction to the Laser

The purpose of this chapter is to present an overall view of lasers and their
applications. In this way, it is hoped that the reader will have a clearer and
more cohesive understanding of the material presented in subsequent chapters.
It may also enable the reader to be selective, if desired, in the choice of reading
from the remaining chapters. The reader who has some familiarity with lasers
and the various types available may wish to skip this chapter.
The acronym L A S E R is derived from the expression Light Amplification
by Stimulated Emission of Radiation. Used in this sense, the term refers to
electromagnetic radiation anywhere in the spectrum from the ultraviolet ,u
through the infrared (ir), which includes wavelengths from roughly 0.1 to
ju,m. The term M A S E R was originally coined to describe a similar devi^
(using ammonia molecular transitions and a microwave cavity) that op ^ ^
in the microwave region of the electromagnetic spectrum—in this cas ,
wavelength of 1.25 cm. . tjve
The acronym, laser, although slightly redundant, is highly descnp ^
Light and radiation, in this context, are the same thing. The key w u]a
"amplification" and "stimulated emission." It is the ability of light to s ' 1J(jei
the emission of light that creates the situation in which light can be * u f e ,
It may be helpful to point out at this time that a very important e .
most lasers is an optical resonator, usually consisting of two Precl;L •
mirrors, one of which is partially transmitting, to allow an output.
arrangement provides positive feedback. So the laser is basically a p
 Requirements for Laser Action 79

i AS such, it is analogous to electrical positive feedback oscillators

back osc'"at°r' amount of the output is fed back in phase with the input,
*here a Cef a 'Nation at some frequency characteristic of the circuit. In effect,
n m
resulti 8 i cts a frequency component from the noise always present from
the osci"ator . ancj oscinates at that frequency. The laser does essentially
biasing- arnP except that an optical oscillator can operate in many allowed
'^natural resonator frequencies).

5.1 UNIQUE PROPERTIES OF LASER LIGHT

r is basically a light source. The radiation that it emits is not fundamen-

different than any other form of electromagnetic radiation. The nature
f the device, however, is such that some remarkable properties of light are
realized. These unique properties, taken as a whole, are not available from
any other light source to the extent that they are obtained from a laser. The
unique properties referred to are

1. High monochromaticity (small wavelength spread)

1. High degree of both spatial and temporal coherence (strong correlation
in phase)
3. High brightness (primarily due to small beam divergence)
4. Capability of very low (microwatts) to very high (kilowatts) continuous
power output for different types of lasers
5. High peak power (terrawatts) and large energy (hundreds of joules) per
pulse in pulsed output lasers
6 Ca a
P bility of being focused to a small diffraction limited spot size (of
the order of the wavelength of the light)

mav , Pertles are by no means independent of each other and, in fact, some
> * inferred directly from others.

FOR
A nu
llSted
mber of
here a ,C ltlons mus
t be satisfied to achieve lasing action. They are
'" ' ater chant1 arC briefly
discussed at this point with more detail following

r
«qui ' I n some cases
to
achieVpe i •' SUC as tne 's'2 'aser. tne
gain is sufficiently great that feedback is not
lasing action.
80 Introduction lu lne L

1. Population inversion
2. Optical resonator, except in extremely high gain systems
3. Lasing medium
4. Means of excitation
5. Host medium (usually)

The notion of a population inversion refers to a condition in whicW

ensemble of atoms or molecules is in a nonequilibrium situation wh" ^
of these atoms or molecules are in some specifiedexcited energy state (el t"^
or vibrational) than are in a lower energy state> These atoms or mol °T°
undergo a transition to a lower state in which the probability of emissio
photon is extremely high, a so-called radiative transition.
The optical resonator refers to the technique for providing positive feedback
into the system to produce oscillation. This process usually consists of two
parallel mirrors placed some distance apart; one is as nearly totally reflecting
as possible and the other is partially transmitting to obtain a useful output
from the system. The transmittance of the output mirror, which is the ratio
of transmitted power to incident power, ranges roughly from 1 to 60%, depending
on the power level and type of laser.
The lasing medium (called the lasant) refers to the atoms or molecules
that actually emit the light, such as Ne atoms in a HeNe Laser or Cr3+ ions
in a ruby laser.
Some means of excitation is required to achieve the population inversion.
This step is usually accomplished by the electrical discharge or high intensity
light from gas-discharge lamps, such as xenon or krypton gas-discharge lamps
Many other important techniques for excitation exist, however, such as chenuca
reaction, electron beam preionization, nuclear and gas dynamic. In the cne
reaction pumped lasers, two or more reactive gases are mixed in a c a
where an exothermic reaction occurs, thus, producing the energy requir
the population inversion and the lasant as well. In electron beam preioni ^
a broad beam of electrons is accelerated through a high voltage m
from which many more slower electrons are ejected; such electrons ^
through the laser cavity, causing ionization by collision. Excitation i ^
achieved by means of an electrical discharge at a voltage too low ^
ionization but that is more effective at achieving excitation. Nuclear ^
refers to the use of a beam of particles from a nuclear reactor or ^
to cause excitation. Gas dynamic lasers use differences in decay ^^ g
excited states to achieve a population inversion when a high temp
is forced through supersonic nozzles. rllt>y I*-'
A host medium is one in which the lasant is dispersed. In * iurn is'
A12O3 is the host and serves as a matrix to hold the Cr3+ ions. ^^
host in the HeNe laser and is essential to the process of exciting
 ^ Types of Lasers 81

nes have no host; examples are the semiconductor diode laser

like argon and krypton.
and' 0 "

HOW THE LASER WORKS

5,3

ovides a brief qualitative explanation of how lasers work. More

£ information is presented in later chapters.
iic" action begins with the establishment of a population inversion
'tation process. Photons are spontaneously emitted in all directions.
by
' C C tramline through
PhotOnS llaVCiiiife t o
the active
.
medium
. 1
can
.
stimulate
1
excited atoms or
les to undergo radiative transitions when the photons pass near the atoms
m 6
° olecules. This factor in itself is unimportant except that the stimulated
nd stimulating photons are in phase, travel in the same direction, and have
the same polarization. This phenomenon provides for the possibility of gain
or amplification. Only those photons traveling nearly parallel to the axis of
the resonator will pass through a substantial portion of the active medium. A
percentage of these photons will be fed back (reflected) into the active region,
thus ensuring a large buildup of stimulated radiation, much more than the
spontaneous radiation at the same frequency. Lasing will continue as long as
the population inversion is maintained above some threshold level.
The optical resonator, because of its geometrical configuration, provides
for a highly unidirectional output and at the same time, through the feedback
process, for sufficient stimulated radiation in the laser to ensure that most transi-
tions are stimulated. The phenomenon of stimulated emission, in turn, produces
a
highly monochromatic, highly coherent output (some exceptions will be dis-
cussed later). The combined action of the resonator and stimulated emission
^° uces an extremely bright light source even for lasers of relatively low power

M
TYPES OF LASERS

N
° attempt
avail
will h m j to llst a11 the
able M • significant laser types commercially
detai1 mdustrial lasers wil1 be briefl
Presented y described here with additional
at!his
Point ar °° Selected las«rs in a later chapter. The lasers to be discussed
!or e rubv
diode. ' > Nd-YAG, Nd-Glass, CO2, dye, and the semiconduc-
k

'aser with He H!^ 'S by far tlle most common laser in use today. It is a
fred) for
most ^ Ne the lasant The ° utput wavelength is °-6328
i ^er levels r appllcations> although it can emit at two ir wavelengths
y re
liable 1 an,8e fr°m microwatts to about 50 mW continuous. This is
cost laser and is used extensively in such applications as
82 Introduction to the
Laser

alignment, surveying, ranging, displacement measurement, hologr

recognition, communications, surface finish analysis, and flow ^''
The means of excitation in the HeNe laser is a dc glow discha SUre'
Sas
atoms that participate in the lasing process are not ionized; so the M
e e
is a neutral gas laser. ^ las^
The ruby laser has A12O3 (sapphire) as the host and triply J0
mium ions, Cr3+, as the lasant. The output wavelength is 0.6943 n ! ^
is operated in pulsed fashion with a low repetition rate because of' r>o i, ''
properties of A12O3. Many joules of energy can\be realized in a sin 1 ^
however. This energy has been used for drilling diamonds, sending D i
the moon, spot welding, hole piercing, and pulsed holography. Excit f* l°
ls
by means of capacitive discharge through flash lamps.
The Nd-YAG and Nd-Glass lasers are similar in that both have tri i
ionized neodymium (knee-oh-dim-e-um) atoms (Nd3+) as the lasant. YAG stand
for yttrium-aluminum-garnet, a synthetic crystal also used for simulated dia
monds. Its chemical formula is Y3Al5Oi2. The output wavelength for both laser
types is 1.06 jam (near ir). Nd-YAG lasers are operated continuously or at
high repetition rates with average or continuous power output of from a few
watts to 1 kW. Applications of Nd-YAG lasers include welding, cutting, hole
piercing, and excitation of dye lasers with frequency quadrupling. Nd-Glass
lasers are pulsed because of the poor thermal characteristics of glass, but ex-
tremely high peak powers have been achieved. Lasers of this type are pumped
by gas-discharge lamps. Some applications for Nd-Glass lasers are welding,
cutting, hole piercing, and laser fusion experiments.
There are several different basic CO2 laser designs, but, in general, the)
use CO2 as the lasant and also contain N2 and He. The emission wavelength
is 10.6 /o.m, which is in the mid infrared, and the power output is continuous
or pulsed with average power ranging from a few watts to tens of Kil°
Excitation techniques include dc glow discharge, rf, electron beam preioniz
i /rlnVV 0*
chemical and gas dynamic. As far as industrial lasers are concerned, g ^
charge, rf, and electron beam preionization excitation are of greates
Applications for this laser are numerous and include cutting, hole p ^
welding, heat treatment, scribing, and marking. The variety of maten ^^
can be worked is also varied, including paper, plastic, wood, glass, c
ics, and most metals. Q^^
Dye lasers are devices in which the lasant is an organic y • ^ $
tl!
organic dyes fluoresce in a large number of lines (distinct wave ejeng
laser can be tuned to emit at a variety of wavelengths over som ^^
band. By using several dyes, it is possible to obtain thousands of 1S .^le \yP
spaced2 wavelengths in the visible part of the spectrum from
with fairly uniform power output at all these wavelengths. Power
*
2 ucing» c0fl
Line broadening causes the discrete lines to overlap, thus, in effect, pr°
tunable laser.
 83
Bibliography
ChaP- 5
watts. These lasers are pumped by xenon lamps, argon lasers,
.VOID miH'wattS excimer lasers, or frequency-multiplied Nd-YAG lasers. The
pstroge" laserSjasers is normally pulsed. Dye lasers are used in spectroscopy,
!iutput of Hi reaction studies, pollution detection, and surgery.
photochemic . ^uctor diode laser is a distant cousin of the light-emitting
Tne * • tnat it is a semiconductor diode that emits light. The diode
diode (L emits coherent radiation with a much narrower wavelength
laser, ho ' jnconerent emission of an LED. Modern diode lasers are chiefly
3
spread t variety of ternary compounds, such as GaAlxAsi-!. Feedback
compose ^ ^ achieved either by cleaving the ends of the diode parallel to
'" t'other and perpendicular to the diode junction to form mirror surfaces
ea
°h distributed feedback via a diffraction grating etched on top of the chip
Of
llel to the junction. The light is emitted approximately parallel to the junc-
fion3 rather than roughly at right angles to it as in most LEDs. Physically
diode lasers are small, being on the order of 250 jam on a side by 50 ju.m
thick. These devices, however, can emit peak powers of many watts in 100-
to 200-ns pulses with an average power output of several milliwatts. Diodes
operating continuously at room temperature can emit several milliwatts of power.
Wavelengths vary slightly for GaAl^Asi-x but are close to 0.9 /im (near ir).
Applications for diode lasers appear to be in fiber optic communications, pattern
recognition, ir illumination, and pollution detection and control.

PROBLEMS

5-1. List six unique properties of laser light.

2. List five requirements for lasing action to occur and briefly explain what each
means.
• ist our important industrial lasers and suggest a possible type of application
°r , ' alon8 w«h the output wavelength and the range of CW power or energy
per pulse.
"^^ Starting with tu emls
qualitat I sion of a spontaneous photon along the axis of a laser,
SCribe the iaSing acti n in a CW laser Use sketches to
your disT ° illustrate
^ What CUSS1°n and describe the relationship of the radiation to the initial photon.
S the fun
' ction of the mirrors in a laser?

BIBLIOGRAPHY
iea, rj Q w
4 R Callen> and W T
Pplications R ' - Rhodes, Introduction to Lasers and
dln8 MA:
^"gyel, B ' Addison-Wesley Publishing Co., 1978.
Itlc
' !966. A" Itltroduction to Laser Physics. New York: John Wiley & Sons,
:
A

ming three elements in contrast to a binary compound, which contains

 chapter

Laser Theory

The objectives of this chapter are to provide a sufficient theoretical basis so

that readers can pursue additional study on their own and appreciate properly
the material presented in later chapters. A basic understanding of what the
laser is, how it works, and what its unique properties are will help dispel the
mystery that surrounds this useful device. When understood, tools are much
more frequently used and used correctly.
Only a brief theoretical treatment can be presented in this limited space
Many results are given without proof and derivations are kept as simpe
possible without loss of meaning.

6-1 POPULATION INVERSION

nOpU'
One requirement for lasing action is a population inversion, Establis 1 ^.^
lation inversion requires that more lasant atoms or molecules be in s ^.,.
state than in the state to which the most probable transition occurs. ^ ^^
tion must be one in which photons are emitted (radiative transition) ^ )attet
to one in which no photons are emitted (nonradiative transition),
case, the energy ends up as thermal instead of electromagnetic ene
wu _ . M
If only two energy levels were involved, the system «L' .Qll is
two-level system. The only laser that actually resembles this si u ^^ ^f
diode laser1 since the transitions take place between levels near
 Stimulated Emission 85

Inverted population
A/, 5,

Figure 6-1 Two-level energy level scheme.

nductor. Most lasers are better approximated by three- or four-

of the se ^^ actual energy level structure of atoms or molecules is much
leVCl
Complicated than that, however. Nevertheless, much can be learned by
™°KA°\m a hypothetical two-level system. Some discussion of three- and four-
iTnasers will be presented subsequently. With this in mind, we initially discuss
lation inversion, plus a number of other matters, in the context of a two-
level system.
A two-energy-level system is depicted in Fig. 6-1. Here £\ and E2 are
the energies of the respective levels and NI and N2 are the number of atoms
or molecules per unit volume in each level. The ratio of N2 to Ar1 at thermal
equilibrium is given by the Boltzmann equation, which is present as Eq. (6-1).

(6-1)

Here k is the Boltzmann constant and T is absolute temperature. It can be

seen from this relation that N2 can never exceed NI for an equilibrium situation.
A
population inversion refers to a nonequilibrium situation in which N2 >
• i- It occurs as a result of the excitation process and has been referred to as
y n^atlve temperature" situation because Eq. (6-1) can only be made to predict
is ne ativ
res 1 8 e. This was an attempt to fit a thermal equilibrium
0 a nonequilibrium
tion i condition and is not considered further here. A popula-
n 1S absolutel
y necessary for amplification of light to occur. Without
en
«rev fef°-inversion> a beam of light directed through the medium with photon
V ~ E2 - El will be absorbed.

ST||
** *ULATED EMISSION

types
tion, Sp0tranSUlons involving photons in laser operation are of interest:
a
Photon"14060115 emission' and stimulated emission. Absorption occurs
abs
°rbed, thu WUh energy eclual to the energy difference between two levels
^d to a high °aUS'ng an excitation. In electronic transitions the electron
m
°lecule ' er energy level; in molecular transitions the vibrational energy
, 'ncreased. Atoms or molecules2 in excited states eventually
Fre< u
) <mtiy the
erm atom
will be used where either atom or molecule could apply.
86 Laser
Theory

drop to a lower energy state spontaneously; in radiative transit'

is given off with energy equal to the difference in the energy ievei S a Phot0r
The precise theory of the phenomenon of stimulated emission i
the scope of this book, which does not, however, preclude unde ^°n<l
8[
nature and consequences of the phenomenon. Basically what han k
$ S
a photon passing sufficiently close to an excited atom causes it t ' "^
radiative transition before it would otherwise dq so spontaneously Th ^° a
lated photons have the same phase and the same\polarization state Vtuilu'
in the same direction as the stimulating photons. Consequently, gain i
when a beam of light of the right frequency (photon energy) is directed tr,'"6
a medium in which a population inversion has been established.
The fact that the stimulators and stimulatees are in phase with each
produces a high degree of both spatial and temporal coherence in lasers Us ir-
substantially less than perfect temporal coherence is achieved because of rand
fluctuations that occur in the laser due to vibrations and thermal variations
Stabilization techniques can be used to increase temporal coherence greatly
and hence coherence length. The lateral spatial coherence of laser beams is
extremely good.
The occurrence of the phenomenon of stimulated emission is not too sur-
prising with hindsight. The atoms or molecules behave like excited electric
dipoles. When these dipoles come under the influence of the electric and magnetic
field associated with a photon, the dipole experiences a force. If the phase and
frequency of the dipole and photon match (also spatial orientation of the photon
electric field and dipole orientation), this force results in the dipole giving off
radiation (a photon). This interaction has often been referred to as "tickling
the excited atoms with the photon to produce the stimulated emission.

6-3 EINSTEIN PROBABILITY COEFFICIENTS

Based on thermodynamic arguments, Einstein established what is es ^

the theoretical basis for the laser decades before the first one was ^
theory deals with the probabilities for absorption, spontaneous em'SS1he'rates
stimulated emission of photons. Actually, it is necessary to talk abou ^ •$$
of the various types of transitions in order to get a grip on the su ^^
rate of spontaneous emission can be written N^Azi, where Nt utility P*r
of atoms per unit volume in the upper state (2) and A2i is the p un£ j e rgo*
unit time that a given atom in an excited state will spontaneous eo*
transition to state 1. The product N2A2i then gives the number o ab
transitions per unit volume per unit time (rate per unit volume).
and stimulated rates are a little harder to describe because nei
if no radiation field is present. It is reasonable to expect these rateS^^
i2 give*
tional to the radiation energy density. Thus the product MR/ ^sp^1
rate of absorption of photons per unit volume. The energy densi y P
 instein Probability Coefficients 87
6-3
h refore has the units joule/m3Hz. B12 is the probability per
.. ,,^it soectral energy density that a given atom will be excited
time Per u'^e £ Tne umts of B12 are complicated because of the absorption
,,nit state
on p/; they are m3Hz/joule-s. The rate of stimulated emission
e
jape" as tf2pfB2i, where N2 and p/ are the same as previously
^ e*P „ jjas the same interpretation as B12 except that it relates to
^ 'transitions ^Qm state 2 to state 1. At equilibrium, assuming that
' late I transitions occur, these rates are related as follows:
N!pfB12 = N2A21 + N2pfB2l (6-2)
ation simply states that the rate of upward transitions (absorptions)
the rate of downward transitions (emissions) at equilibrium for a two-
system. By using the Boltzmann relationship given in Eq. (6-1), the ratio
can be eliminated from Eq. (6-2), yielding
(6-3)

Setting E2 — £1 = hf, the energy of a photon absorbed or emitted, and

solving for p/ yield

Prior to Einstein's work, Planck worked out a model of blackbody radiation

that correctly predicted the spectral emission and radiation density of a black-
body, the ideal absorber and radiator. The spectral radiation density at frequency
/, according to Planck's theory, is

877-A/3 1
*• '

can c 6CaUSe C)S' (^~*) and (6-5) are equivalent for a two-level system, we
and see what il tells us about the Emstein
only w coefficients. The
right hand
, anA - sides of Eqs. (6-4) and (6-5) can be equal is if B12 =
aA
* is related to Bn by

Ti C
fa (W)
fi!ltne
ans tha" f e a C e d hefe are extremely important. The fact that B12 =
Pr ba
° ble f 0r an ' °r a given radiation density, a stimulated radiation is just as
re> at m ln State 2 as absor
^° the coeffi ° Ption is for an atom in state 1. Further-
1014 1 f B21 In Eq (6 6) iS numericall
rat ^2, which" ° ' ~ y 5-6 x 10~16 (SI units)
fs mea S that in a radiation field of
rad* ° P°ntane " density 1CT6 j/m3Hz the
atio11 US radiation is 101 times
and' - This ° ° smaller than the rate of stimulated
S me
° feedback*CnSUres that when a Population inversion is established
mechanism is included to provide reasonable radiation density
88
Laser'

at the correct frequency, more stimulated emission than sponta

will occur and the device will lase. As noted, all that is need HUS eift's
amplification is a population inversion and a beam of light of th ° ^
Cor
quency, but it generally must come from a laser. rect

6-4 AMPLIFICATION

If there were no losses in a medium with an inverted population th

in radiation would be geometrical.3 This situation is illustrated for t . growtil
linear laser or amplifier arrangement in Fig. 6-2. As can be seen th U$Ua'
photon creates one more; these two, in turn, produce four, which in ^
produce eight and so on. A geometrical growth implies that there will be
fixed fractional increase in the number of photons per unit length of the mediu

Avalanche effect (gain)

due to stimulated
emission

Initial photon

B u i l d u p given by 1 = I0e°* Figure 6-2 Optical amplifier.

Because the irradiance is proportional to the flux of photons, this canbeexpress&>

as
dl

where / is irradiance, z distance, and ft the fractional increase in

per unit length or gain coefficient. Beer's law says that the loss o ^
in a dielectric medium is given by dl/I = —a dz, where a is the « ^
per unit length. Thus the net fractional loss or gain is dl/l — ^
which when integrated is

There are various causes of losses in a laser, such as absorption ^ , u t-

in the lasing medium, absorption by windows used to seal 1 p jete ^
absorption by the mirrors, and, of course, the output. In one c ^
3
We are no longer talking about an equilibrium situation and it shou
a population inversion in a two-level system cannot be achieved by optical p
 ^ Amplification 89

of optical length4 L the round trip gain, ///<> = G, is given by

a
in ^L wjiere J?j and K 2 are the reflectances of the mirrors and all
' a're included in a. This can be written
bt>t losses <"
t1[t1
G = e ( 0- a ' )2i (6-8)

th it will be assumed that all losses are incorporated in a.

e
r ' instructive to derive an expression for the gain coefficient on the
f the population inversion and the Einstein coefficients. Assuming that
^ntaneous emission is negligible, we can write the net transition rate as
(N2pfB2i - NipfBi2)A Az = (N2 - Ni)pfB2lA Az
This transition rate can be equated to the power (irradiance times area) A///4
generated in the volume element A Az divided by the energy per photon hf.
Thus

A Az = (6-9)
. . "J
In the limit as Az — * 0 we then have

^z=(N2-NjpfB2ihf (6-10)

For an unidirectional (linear) beam, // = pfC so that Eq. (6-10) becomes

d!f = (N2 !.„.,-„..,
dz c
*hich integrates to

(6-12)
/ 0 1S' some
initial irradiance. Therefore

SUbStitute f
or Bn in terms of An and get
B_(N2-N^A2l

87r/2
ls
the
^ Probability per unit time that an excited atom will undergo
"sition. Therefore the reciprocal of A21 is the average time
<See
Chapter j
or
'he definition of optical length
90

T that an atom remains in the excited state before making a spom

e Us
tion. Now we can write the gain coefficient as ° trans]

Although we have ignored line-broadening effects (discussed later)

ate energy levels (more than one level with the same energy), some ^tn^-
ere
conclusions can be drawn from Eq. (6-14). sting
The most obvious conclusion from Eq. (6-14) is verification of
that N2 > NT. is required to achieve gain, if N2 < NI loss is incurred^'
second interesting point is that the gain is inversely proportional to the fre *
squared (line-broadening effects make this even worse). This statement "^
that lower gains will be achieved at higher frequencies and therefore high f
quency lasers will probably be harder to develop than low-frequency ones It
appears to be borne out by the low number of lasers in the uv compared with
visible and ir lasers and the total absence of lasers in the x-ray5 or y-ray region
The third very interesting point is that the gain is inversely proportional to
the spontaneous lifetime T. It is often pointed out that the upper lasing level
should be metastable or a long lifetime state. It can be seen from Eq. (6-14)
that a short lifetime is desirable for high gain. Because a finite amount of time
is required to achieve a population inversion, a compromise must be reached.
We would like T as short as possible to provide high gain but sufficiently long
to allow the population inversion to be established and maintained in the case
of continuous output (CW) lasers. It is also desirable that the lifetime of the
upper lasing level be substantially longer than the lifetime of the lower leve ,
although it is not strictly necessary in pulsed lasers.

6-5 POWER OUTPUT FOR CW OPERATION

It is helpful for an understanding of both CW and pulsed laser operation^

look at how CW operation relates to the gain coefficient and optica o ^
we assume some initial power P0 directed along the axis of the laser,
power flux at the same point after one complete round trip would
P0 +Poe(fi~a>2L ^
where L is the optical length of the cavity. The power output o
would be

Trie
where T = 1 — R is the transmittance of the output mirror.

5
Such lasers have been reported, but the jury is still out.
 pulsed Operation 91

this oower is always being generated spontaneously. After an-

because tni* v
trip we have
^ _ jyojl _|_ e(H-a)2L _|_ e 2(/3-a)2L]

that each existing term is amplified and a new TP0 is generated.

it' s see rp<is
re o
i n for a continuously operating laser becomes an infinite
The

P=TP0 2 e n(/3 - a)2L (6-15)

n=0

where n(/3 ~~ a)2L is less than zero (which it must be for stable
Cooperation), this series can be summed and Eq. (6-15) becomes

This result implies that y3 must be at least slightly less than a for CW laser
action, which seems contrary to common sense. However, it must be recognized
that if ft > a, then P would increase indefinitely, which is impractical for
CW operation or any other kind for that matter. To have equilibrium, the
optical power generated must be exactly balanced by optical losses, including
the useful output. The gain coefficient /8 must be slightly less than a because
of the small amount of spontaneously generated optical power. Stated another
way, the apparent deficit due to the fact that the gain is less than the loss is
compensated for by spontaneously produced power. Fortunately, the spontane-
ous power is negligible compared with the stimulated power in lasers.
The value of /3 given by /3th = a is called the threshold value of the
gam
las
coefficient. The gain must closely approach this value in order for continuous
>ng action to occur.

6-6
PULSED OPERATION

In puked
y
alue, oen ,f 10n °^ lasers the gain is raised substantially above the threshold
dis
charge j y means of flash lamps in solid lasers or pulsed electrical
asers
P°Pulation ' - As threshold is exceeded, lasing action begins, but the
ln
creasing D '°n contmues to increase for some time, resulting in a rapidly
P°pu'ation i n v e r -° U t p U t ' This situation quickly leads to a diminution of the
*^T '°oks som h*1 and 3 dr°P m Power- The power output of a pulsed ruby
4 the Periodic b* l at aS Shown in Fi8- 6-3(a)- The power spiking is the result
nTl"16 in Peak U1 and de letion
P °f the population inversion. The gradual
u , arr>Ps Fff- . er ls *he result of the diminishing power output of the
c]p-ir f ' "ctivplv tu i
atl(j j rom £ y> ine laser is alternately lasing and not lasing. The reason
op t j ess than a ,'' @ ^ alternating between values much larger than a
PUtr
>ping H Population inversion alternately increases due to the
decreases due the rapid growth in stimulated emission.
92 Laser
Theory
Cha

Irregular spacing

(a)

Regular spacing

Time

(b)

Figure 6-3 Laser pulses.

The pulsing is random in multimode6 lasers due to mode hopping. In singe

mode operation periodic oscillations occur in the first part of the pulse (relaxa
oscillations) and die out to a smoothly diminishing power output for the re
der of the pulse [Fig. 6-3(b)].

6-7 THREE- AND FOUR-LEVEL SYSTEMS

In reality, lasers are more nearly approximated by three- or four- ev

such as those schematically represented in Fig. 6-4. jgvel !#'
Two common types of lasers are accurately represented as t re ^ pjne
ers—ruby and dye lasers. In both cases, excitation is achieved by optica ^^ ,0
to some level, or band of levels, E2, followed by rapid nonradiat]1V£.J an<) 'h£
some lower level E3. Lasing action then takes place between K

' See Section 6-9 for a discussion of laser resonator modes.

 and Four-Level Systems 93
Three-
6-7

(a) Three-level system

rwv - E,

(b) Four-level system

Figure 6-4 Three- and four-level energy level schemes.

Sl te £l
becau f ; Tne threshold level pump energy is quite high for these lasers
^ e entire population of lasant species must be inverted.
to a levei°Ur"leVel laSCrS> SU°h as Nd'YAG> HeNe, ion, and CO2) excitation is
sl £
3 w'th )band °f leVels EZ' f°llowed by raPid nonradiative transition to
It ,3S h'o smS occurring between that level and another intermediate level
that
te £i rapid relaxation by nonradiative transition to the ground
er S
^ than ' u Potential for CW operation in four-level lasers is much
!ow r e level lasers
« : there i" ~ > for the threshold pump energy is drastically
n y ne nCed tO invert the
° ' cessarv T entire population of lasant species. It is
ptica
( l methods • 3t ^3 ~ N* exceed some threshold value. Excitation is by
^°z.HeNe a 'd S°lld laSetS and usuallv bv electrical discharge in gas lasers.
ti0 °tller ion's n '|netalvaPorlasers various types of energy transfer by collision
'nversion ' ° ecu'es' or atoms play a major role in producing the popula-
94

6-8 RATE EQUATIONS

Primarily for better understanding rather than practical use, the r

for three- and four-level lasers are presenteohere and discussed b ' f
The appropriate equations for the three-level system are
Ni + N3 = Nt
(6-17)
P
dt T (6-18)

where WpNi accounts for the pumping rate, T is the lifetime of the lasin
state, TC is the cavity lifetime of a photon, B = B12 = -B2i is the Einstei
stimulated transition coefficient (probability rate per photon),7 p is the number
of photons in the cavity, and V is the mode volume8 of the active region
Equation (6-17) states that the total number of atoms Nt is shared between
the lasing state and the ground state; N2 is assumed to be negligible because
of the rapid decay from the pump level. Equation (6-18) expresses the rate of
change of atoms in the lasing state as the pump rate minus the stimulated
and spontaneous emission rates. Equation (6-19) expresses the rate of change
of photons in the cavity in terms of the net stimulated emission rate minus
the rate of loss of photons, p/rc .
Letting AN = N3 — NI represent the population inversion, Eqs. (6-17!
to (6-18) can be reduced to the following two equations.

= W,(Nt ~ AN) - 2BP AN -

dt

Equation (6-21) tells us that lasing action will begin at some critica v
population inversion, ANC = \/(rcVB). For AN > AN C) dp/dt > 0 and ^
cation begins. A critical pump rate can be obtained from Eq. (6-2 ) ^ ,-3.
d &N/dt = 0, AN = A7VC and p = 0. Setting p = 0 is justified y^. ^
that the number of photons present in the cavity prior to the onse
is very small. Thus
(Nt
T (Nt -

7
Previously the Einstein coefficients were based on energy density.
8
Mode volume is the volume of the active material actually filled by ra
 QC

Rate Equation
6-8

N3c
(6-23)

4- jVK and A./VC = A^c — A/I C . This last equation simply

N:
using "' .e Qf pump transitions equals the rate of spontaneous transitions
5 tates that ^ mp rate for the onset of lasing action. Actually, Nu ~ N3c
a t the en •mateiy \vpc = 1/r. There must, however, be a large enough
S
° 'TationTnversion to overcome the losses of the laser.
^U dv-state condition for a constant pump rate that exceeds the critical
* will he reached where dp/dt = 0 and d &N/dt = 0. This results
pump rate wl
in

(6-24)

The steady-state population inversion is exactly the same as the critical inversion.
Higher pump rates mean proportionately higher numbers of photons in the
laser. Moreover, because the power output is directly proportional to the number
of photons, the power output is also linearly proportional to the pump rate.
The output power can be calculated from Pout = HfpcT/2L. This conclu-
sion follows from the fact that each photon has an energy hf and all the energy
in the resonator strikes the output mirror in the time 2L/c with T giving the
ac lon em
itted. It does not mean, of course, that more and more power output
an
e achieved by supplying more and more power to the laser. Clearly the
freshP ratC °annot exceed the rate of return of atoms to the ground state. If
a
"d li*10?5 °r molecules can be supplied to the mode volume, as in some gas
rate. Th' ^™' then pnysical 'imitations on flow rates will control the pump
damage A e"ects due to excessive optical pumping in solids can cause severe
!hat
Prod" thUS limit maximum pumping levels. The fraction of power input
''* incre '$ rans'^ons to tfle pump band in gas lasers, frequently decreases
II
is a'"8 °Urrent after some optimum current level is reached.
l =
b
^3 = ° for a four-level laser because of the
y &N ^^ involved. The critical population inversion is once again
^. how ^VBr^ and at steady-state AA' = AA^. The critical pump
e CqUalS AArc/(Ar r) In addit
* critical"' ' - ion, because AATC < Nt, then
tllre
e-level l^s pum P rate for four-level lasers is generally much less than
re
ader tSerS~7m fact- as much as 100 times less. (It is left as a problem
Ce e re W te d Wn the rate ec ua ions for
«ve)
e Oprt
* «ults " ° l t the four-level laser and
ient.) S Stated in this paragraph by analogy with the three-level laser
96 „,aoc
„, 1K
"• r irteory
D R

6-9 OPTICAL RESONATORS

Some form of positive feedback of power is usually necessary if i •

in addition to gain, is to occur. This situation can be achieved b *° act'°n
distributed feedback, such as diffraction gratings, or by instantaneo ^ U$e of
with mirrors. Distributed feedback is used with some diode lasers ***""
as
in cases where lasers are integrated into a system, such as in fibe well;
integrated optical circuits. We will discuss only feedback in which m"optics,pucs Or

used. The fraction of power reflected by the output mirror varies fro °K ^
99% for low-power lasers to less than 50% for some high-power syst °f
Both stable and unstable resonators are used in lasers. A stable r
is one in which, from a geometrical point of view, certain light rays r &
their paths indefinitely inside the resonator. Conversely, in an unstable reson /*
the light rays diverge or "walk off" the axis. Most laser resonator designs are
stable, although unstable resonators have very useful properties for some high-
power systems. Much more detail will be presented in Chapter 7 on the classifica-
tion of laser resonators and their effects on laser beam properties. At this point,
we will consider the various allowable modes of laser resonators.
If all laser resonators consisted of parallel-plane mirrors (Fabry-Perot inter-
ferometer), only axial (longitudinal) modes would need to be considered to
determine allowed resonant frequencies. The existence of these modes is based
on the electromagnetic boundary condition that only those modes will be sup-
ported by the resonator for which the phase change in one round trip, 2L, is
an integral multiple of 360° or, expressed mathematically,
m\ = 2L, m = l,2,3, . . .
Be aware that L is the true optical path length of the laser. In Eq
(6-26) A. can be replaced by c/f. Then
/= mc/2L
We see from Eq. (6-27) that the frequency spacing between adjacent a
r
modes is simply c/2L. resonat° '
When one or more spherical mirrors are included in a stable ^
the wavefronts are spherical, but the mirrors are not necessarily cons a ^^
surfaces. In fact, any mode that is self-reproducing on one comp ^^
trip is an allowable transverse mode. These modes are referred to as ^ jf the
electromagnetic (TEM,P). They can be circular but frequently, esp^ Of th^
laser output is not aperture limited, take a rectangular form. A
modes are depicted in Fig. 6-5. -uorth0^011
The subscripts on TEM simply represent the number of nodes i ^ .fl fig.
directions. Another important mode is the "doughnut" mode *# /-rfMo"
6-6. Actually, this output is the result of a laser operating in the .^
mode with rapid random variation of the orientation of the no
perpendicular to the z axis (propagation axis).
 6

OOO
OOO
TEN!,, TEM21

Figure 6-5 Transverse electromagnetic (TEM) modes.

frequency of a given mode, including axial and transverse effects,

be deduced from electromagnetic theory and is given by

where gi = 1 ~ L/r^, g2 = 1 — i/^2, and PI and r 2 are the radii of curvature

of the resonator mirrors. Radii of curvature are taken as positive if measured
toward the other mirror and negative otherwise. The frequency spacing between
modes due to a difference of one in the transverse mode number is

(6-29)
77

This spacing varies from zero to as much as c/L, which is, in fact, equivalent
to 5/ = 0.
Unstable laser resonators must satisfy the same axial mode condition as
stable resonators. Theoretically unstable systems can produce a variety of trans-
verse modes; in practice, however, the losses associated with all but the lowest-
m th* t,ransverse mode are so ni§n that unstable resonator lasers tend to operate
dim 6 st 0rder
" transverse mode. This mode is in the form of a ring whose
nS depend on the
*alkoff° resonator design. The ring output is due to geometrical
c
°uplin°" beam> not diffraction, as is implied by the expression "diffraction
s
ymmetri ^^ ^ °ften USed to describe this type of resonator. A radially
eff
ects al's^ Substructure in the output ring is caused by diffraction. Diffraction
0
cause the output to be less than predicted geometrically. The ring

Figure 6-6 Doughnut mode. TEMJ 0

98 Lase
' Theory

output of an unstable resonator should not be confused with th

mode of stable systems, which occurs for an entirely different r 8hnu

6-10 LINE BROADENING

Line broadening refers to the fact that some finite bandwidth mu

associated with any electromagnetic radiation regardless of its source
tion from a single isolated atom or molecule has a bandwidth due "
referred to as natural broadening. Natural broadening is a result of th ^ 'S
lifetime of an atom or molecule in an excited state. The average tim th "'^
atom spends in an excited state, prior to dropping to a lower state and ^
off a photon, is called the lifetime of that state. According to a classical D '^
of view (an incorrect point of view that yields essentially correct quantitativ
results in this case), the atom is a linear dipole that radiates for a period of
time equal to the lifetime T, thus producing a radiation pulse of length T in
time. It can be shown by Fourier analysis that a single pulse of time length ^
must contain a continuum of frequencies with a full-width-at-half-maximum
(FWHM) given approximately by
A/T = 1 (6-30)
Thus the shorter the pulse, the greater is the bandwidth. The lifetime, T, is
typically 10~8 s. According to a quantum mechanical point of view (the correct
point of view), Eq. (6-30) follows from the Heisenberg uncertainty relations.
Other broadening mechanisms become important in liquid, gas, or solid
laser media. These mechanisms are referred to as homogeneous and inhomogene-
ous, terms that characterize the physical processes causing the broadening
alternatively, as Lorentzian and Gaussian, respectively. The latter names c a
terize the shape of the lines (power versus frequency curves). ^
Homogeneous broadening is chiefly caused by collisions. In gaseous ^
collisions of the lasant atoms with each other, other species in the mix ^
with the walls confining the gas, will cause perturbations in the ene .
quency) of the photons emitted. By using the Heisenberg uncertainty P ^ ^
the frequency spread can be estimated in terms of the mean time betwe
sions tc. 31 )

• asvs
Frequent collisions result in a large bandwidth and so cooling a ^ ^£
reduce thermal motion will decrease bandwidth. This phenomenon i ^ ,^
to as homogeneous broadening because every atom in the gas is ^e t
same probability of collision as every other atom. Each atom ^ ^
same statistical distribution of frequencies of emission as all the ^ tap^
The mean time between collisions in a gas can be estim jis1
a
the ratio of the mean free pathh due to thermal motion, ith (
 Line Broadening 99
°
collisions) to the average thermal velocity, vth- Equations for
b£tw readily obtained from thermodynamic arguments. Thus
2 kT
1th = --=1 (6-32)
TT pa 2

(6-33)

em is the molecular mass, /> the gas pressure, k the Boltzmann constant,
and a the molecular radius.
Homogeneous broadening in crystalline solids is the result of interactions
between the lasant species and vibrations in the crystal lattice. Lattice vibrations
are quantized and are referred to as phonons. So, in effect, interaction of lasant
species with lattice vibrations is a particle-particle collision and can be treated
just like collisions in a gas if the mean time between collisions, tc, is known.
Inhomogeneous broadening occurs whenever any type of inhomogeneity
causes some lasant species to differ from others in terms of the central frequency
of their statistical distribution of emission frequencies. This situation occurs
in solids as the result of any defects that make the environments of lasant
species vary from point to point.
The most important form of inhomogeneous broadening in gas lasers is
due to the Doppler effect. This effect refers to the fact that the frequency of
h
ght (as with sound) is a function of the relative speed of the source emitting
1
e radiation and observer. The result of the application of the special theory
relat
'vity to this phenomenon is shown as

v is
the° 'S i 6 frequencv measured by an observer at rest relative to the emitter,
*hen the^ atWe speed> and c is tne speed of light. The velocity is positive
when
theyemitter and observer are moving toward each other and negative
(6 35
~ ) may"6 movin8 away from each other. For lasers, v «= c so that Eq.

T, /=(!+-)/o (6-36)
'he
neu__
lasant \ c
~~"-y m th r,molecules in a §as lasgr have, in general, components of
" nbuted a ^ ectlon of the axis of the laser. These velocity components
d!st

ccordmg to Maxwell's velocity distribution

ff \^~~r- \ L™ _ ,.,2 + V2 + V 2)J dVt dVy dVz (6_37)

100
Th eory
Cha
where dN/N is the fraction of atoms or molecules with veloc't
e
and v + dv, m is the mass of an atom, and k is the Boltzrnann eti ,
nStant
is the axis of the laser, integration over the x and y velocity comt> - If j
$
dN

Clearly the largest fraction of atoms per unit velocity component

the z axis occurs for vz = 0. As vz increases, the fraction of atom^"""^ '°
velocity decreases rapidly [a plot of N'^dN/dv)? versus v2 yields a C ^
curve]. The frequency emitted by the atoms is a linear function of their T^"
therefore the line shape will be the same as the velocity distribution th '*'
the emitted power as a function of frequency will have a Gaussian shape cent 'H
around the frequency emitted for an atom at rest, /0.
Doppler broadening is a form of inhomogeneous line broadening because
each value of v2 represents a distinct group of atoms at any instant and these
atoms can only emit at the frequency specified by Eq. (6-36). It can be shown
that the Doppler bandwidth (FWHM) is given by
T
(6-39)
M"
where T is absolute temperature and M is the mass in grams of one mole of
gas.
Figure 6-7 illustrates both homogeneous and inhomogeneous line broaden-
ing. For comparative purposes, the bandwidth (FWHM) is assumed to be the
same for both. The quantity g(/) is called a lineshape factor and is defined
such that /o g(f) df= 1; if / represents the total irradiance, then g(f)l =[('>
gives the spectral irradiance.
It is now possible to modify Eq. (6-14) for the gain (3 to take into accoun
line broadening effects. The gain must be proportional to the lineshape

Figure 6-7 Gaussian and Lorentzian lineshapes compare

 Ciavuy Dumping, and Mode Locking 101

ontaneous power emitted at a given frequency, the greater is

,;ir the m°re,StP frequency. Therefore the spectral gain is given by
^.jin at tn
.,„ „. c2S(f)

bandwidth is always small compared to /, the variation of / in

Becauset e .igibje anc j ($(j) takes the same general shape as g(f). It is
Eq (
t'ne that the maximum value of g(f) is 0.637/A/ and 0.939/A/ for
tzian and Gaussian lines, respectively [Svelto and Hanna 1976]. Thus
lo - n as a f un ction of population inversion can be estimated if
bandwidth and lifetime r are known.

6-11 Q-SWITCHING, CAVITY DUMPING,

AND MODE LOCKING

Techniques for modifying the laser output are discussed here, specifically, tech-
niques for producing, short, high peak power pulses.

Q-switching. This process involves interrupting the optical resonator

in some fashion so that the population inversion process can progress with
little or no feedback present to cause lasing. When the population inversion
has reached some high level, the reasonator interruption is removed and the
aser gives out a short pulse with very high peak power. This process is called
Q-swtching in reference to the Q-value of the resonator.
2-value is defined by
27r(energy stored)
— —
energy lost per cycle
lf
£ . sSt h e energy stored in the cavity and T is the lifetime of a photon in
theJ( y, Eq. C6-4M „,„ u^ ,.,_:*^_
Cdvit C
(Ml) can be written

££*« T
le ra(
io of th°
ne
an
^ type °f Oscillatm8 system can be shown to be given by
oscillator central frequency to the oscillator bandwidth

A/=
i\j*a. LflQoi- TI_
>-aser Theory

which would have been predicted on the basis of the uncertai

This is the ideal resonator linewidth and is much smaller than act ^i ^r'ne'P
ne
due to instabilities from vibrations and thermal fluctuations. *id
g-value can also be related to cavity losses, particularly mirror
Assuming that /?i and R2 are the mirror reflectances and account r Ctatlc
rl e
optical losses, Eq. (6-45) can be written " o
£ _ 1 - R1R2E
TC Tr (6-45)

where TT is the round-trip time for the photon in the resonator. Equ t
45) simply expresses the rate of energy loss in two different ways Th "
e
average number of round trips for a photon is given by
TC
Tr

If we use Eq. (6-46) and r r = 2L/c, Eq. (6-42) then becomes

Equation (6-46) illustrates the fact that Q-values for lasers are quite high, typi-
cally of the order of a million.
Techniques for Q-switching have involved rotating mirrors and mechanical
choppers for pulses on the order of 0.001 s. Acousto-optic modulators are used
for pulses on the order of 10~6 s. Pockels and Kerr cells are used for pulses
on the order of 10~9 s. The last three devices are discussed in more detail in
Chapter 1.

Cavity dumping. Certain types of lasers cannot be g-switched succes

fully due to insufficient lifetime in the lasing state; then a procedure c
cavity dumping may be used. This process literally allows the lasing ^
to build up in the cavity with 100% feedback and then at some inSt ^ n This
feedback is reduced to a low value and a pulse of high energy is emitte ^ js
technique does not yield the high peak power pulses of Q-switc
used effectively, particularly in dye lasers.

Mode locking. Lasers are generally capable of operating J ing i"

of modes, both axial and transverse. Normally when a laser is ^^ tf
several such modes at once, the individual modes (of different r ^ ^y
oscillating independently of each other and tend to have tota lly T. ejia^
relations relative to each other. Mode locking is a process in whic

9
Precisely stated, the uncertainty relation is A/ Ar = l/2ir.
 103
Problems
'6
laced in lock step with each other. The result leads to very
Of mo*165 are ^s one picosecond or less) pulses of extremely high peak power.
-dort (as n technique for accomplishing mode locking is through the use
A corf absorber. The absorber is a material, such as an organic dye,
satura
,i can absorb the radiation in the laser cavity except when all modes
,0rma y^ .fi pnase and in spatial location, and consequently supply
s
are in V' ^e ^yg than it can absorb and dissipate in the time of passage
e
more e" gy ^ , ^ ^ye ^\ (jye molecules are simultaneously in an excited
of the p .^ saturated and briefly becomes transparent, allowing part
StatC
Ise energy to pass through. The length of the cell containing the dye
°f tosen to achieve exactly the right condition for saturation.
'S ° °0nce such a mode-locked pulse occurs, part of it is naturally fed back
the laser and with sufficient gain it saturates the dye each time it passes
through it. Thus once started, the pulse remains in the laser as long as pumping
ontinues. Some laser media are saturable and automatically produce mode
locking. Figure 6-8 schematically depicts the mode-locking process.

Figure 6-8 Mode-locked pulses

The distance between pulses is 2L, the time between the pulses is rr =
2L
/c, and the bandwidth of the pulses is
(6-48)
ere
twg V^ is tne number of modes locked and 8/ is the frequency spacing be-
Pulse 6m A-Ccorc^n§ to tne uncertainty principle, the temporal length of the
c/
(A'8nASPproximately (Ar 8jr ) 1 The spatial length coherence len th is then
-~ '
many as 100 modes ma b
( s )
length i ° y e locked together, therefore picosecond
ses
can be obtained. Both pulsed and CW lasers can be mode locked.

PROBLEMS
i .
>n th" 3te thC rat'° °f the number of atoms in the excited state to the number
/i
is ^1CPVv. r ° Und state at T = 300°K, 700°K, and 1000°K if the energy difference
(b) ^yi
K from fir6 the Wavelen8th and frequency of the light radiated in a transition
eXC ed State t0 the
" Sround state?
& l
ln a
materiri 1 yPes of optical transitions that occur between energy states
 .
6-3. According to Planck's theory, what is the radiation density of
frequency of 1014 Hz at 1000°K? a blac
kbody fo.
6-4. Calculate the ratio of the spontaneous radiation coefficient to the •
S 1Inula
tion coefficient at a wavelength of 0.6328 y.m. ted t^
6-5. If the round-trip gain in a 30-cm-long laser is 1%, calculate the n
6 ln
08-a). * «x^
6-6. The spontaneous power output for a certain CW laser is 1 u\y
power output is 2 mW. If the length of the gain medium is 10 cm '^
tet
gain coefficient. Assume that the only loss is due to 2% transmit! k
at
output mirror. 'he
6-7. Discuss what is meant by three- and four-level lasing systems (ill
diagrams) and compare them. **"
6-8. If the lifetime of a lasing state is 0.5 ms, what is the approximate critic 1
rate? pump
6-9. A solid laser consists of mirrors separated by 50 cm with a laser rod SO cm i
length with a refractive index of 1.5 and a Q-switch with a length of 25 cm
and a refractive index of 1.45. What is the frequency spacing between axial modes
for this laser?
6-10. Calculate the frequency spacing between a TEMni0 and a TEMn2o mode fora
laser with an optical cavity length of 5 m and the following mirror radii:
n = 30 m r2 = 20 m
6-11. Estimate the mean time between collisions and the bandwidth due to collision
broadening for a molecular gas (CO2) at 400°K and a pressure of 30 torr.
6-12. Show that the linewidth for a Doppler-broadened line can be written

= 7.16X
VM
where M is the mass in grams of a mole of gas.
6-13. Calculate the cavity g-value for a CO2 laser of 2-m cavity length and reflectance*
of 100 and 70%.
6-14. (a) What is the average number of round trips for a photon in a laser wit
reflectances of 100 and 40%?
(b) If the optical cavity length is 1.5m, what is the cavity lifetime. _
6-15. A Nd-Glass laser (A. = 1.06 /xm) has an optical cavity length of 1 ^ a^ _
axial modes that are mode locked. What are the frequency band*'ce ,'ength
between pulses, time between pulses, coherence time, and coheren
the ulses?
P . f r = 400*
6-16. Calculate the Doppler frequency linewidth for a CO2 laser n

REFERENCE
sS
i"
m '
SVELTO, O., and D. C. HANNA, Principles of Lasers. New York: Pie""
 105
raphy
Bibli°9
P- 6
BIBLIOGRAPHY

Laser Electronics. Englewood Cliffs, NJ: Prentice-Hall, Inc., 1981.

verdeyen> ' ' An introduction to Lasers and Masers. New York: McGraw-Hill
A- t».
i pn 1y ' '
Book LU., H Dunn> La5gr p^-w New York: John Wiley & Sons,
A., an
lnc 1969.
" M Qptics and Lasers. New York: Springer- Verlag, 1977.
4 Young>
'
co Lasers: Effects and Applications. New York: Academic Press,
w. "-.
1976.
 chapter

Laser Optics

This chapter describes the propagation of laser beams and their focusing proper-
ties. In addition, the relationship of laser beam parameters, such as beam size
and transverse modes, to optical resonator design is covered. Finally, both Gaus-
sian (lowest-order TEM mode) beams and higher-order mode beams are dis-
cussed in connection with the aforementioned properties.

7-1 NATURE OF THE LASER OUTPUT

The beam emitted by a laser consists primarily of stimulated radiation ^

small amount of spontaneous radiation (noise) included. Because the ra
is chiefly stimulated, a fairly high degree of temporal coherence occurs.^ ^
in relatively inexpensive HeNe lasers a coherence length of 1 to ,erence
unusual. Special stabilizing techniques can be used to increase the ^t0tL$
length to 100 m fairly easily (Chapter 8). Spatial coherence across the * ereI)Ct
in gas lasers is nearly perfect. In solid lasers the transverse s P atia £nts withi"
may be poor due to a tendency toward lasing in independent n a
the crystal. Hepje las^
The output of some lasers is linearly polarized. In the case ^ Of in
this situation is the result of Brewster windows attached to t e^ ositi° ,
discharge tube to contain the gas (Chapter 1). Brewster windows a nonnal^
so that the axis of the laser is at the Brewster angle relative to ^^ Of o ^
the window as depicted in Fig. 7-1. Light reflected from any s ^ jncidefl"
of the Brewster windows is polarized perpendicular to the pla
 Nature of the Laser output 107

'
Brewster windows

,, Gas laser with Brewster win-

7-1
do*

h naper in Fig. 7-1). None of the radiation polarized parallel to

° f incidence is reflected. Therefore a large loss is introduced for the
artf ^ comp0nent, resulting in a buildup of the parallel component.
11 the radiation in the resonator ends up with parallel polarization
H minimal (ideally zero) reflection losses occur at the windows. What
reflection does occur is conveniently reflected out of the cavity.
Some crystalline lasers produce a polarized output because of optical aniso-
tropy of the crystal. If ruby, which has a tetragonal crystal structure, is cut so
that its c axis is perpendicular to the cavity a polarized output parallel to the
c axis is produced. If the c axis is parallel to the laser axis, the output is
randomly polarized.
A certain degree of linear polarization occurs in lasers where the beam
path has been extended by folding the beam with mirrors if the angle between
the beam and mirrors is near 45°. This result is clear for dielectric mirrors
from the Fresnel equations presented in Chapter 1. Reflection for the perpendicu-
lar component is greater than for the parallel component (which is true even
though the mirrors may be metal); as a result, this polarization state is preferen-
tially amplified.
In all cases where linearly polarized output is not forced by polarizers,
jewster windows, special crystal properties, or folding mirrors, the output is
timenThne°USly circularly polarized, but the phase fluctuates randomly with
e ct tnat
ti ' the radiation is stimulated results in instantaneous polariza-
erm
t},e , al and mechanical instabilities cause timewise random changes in
the sa polarization. Radiation with random polarization produces
C results
It as unpolarized radiation for most purposes.
of
the laserXtremely important to be aware of the nature of the polarization
refle
ctionsTnm """^ applications- When a laser beam must undergo many
C n
° sequentl * P e' a Iar8e attenuation of the parallel component will occur.
'" the Parallel °^ n° Hght is left if the Output of the laser is linearly polarized
f° n Linear r^T**' ^^ tha" °ne holograPher has run afoul of this phenome-
| 0rtrave lpara°i a i nZatl0n als° causes a difference in cutting speed and quality
^-cutting an 6 ] and perPendicular to the polarization direction in high-power
nee, PP ications. Random and circular polarization exhibit no differ-

'her Words th™ & ^^ expands as though emanating from a point source,
of the TFMaVefr°ntS f°r an axiallv symmetrical beam are spherical
mode. The radius of curvature in the near field changes
108 Laser
Op,ics

Figure 7-2 Propagating laser beam.

from an infinite value, at the waist, to a finite but linearly increasing quantm
as the far field is approached. This situation is shown in Fig. 7-2. '
As can be seen, the radius of curvature at the waist is infinite; as the far
field is approached, the locus of points defining the beam size approaches an
asymptote and from then on the beam appears to have originated from a point
source located at the waist where the asymptotes intersect. As will be seen
later, the far field is generally a long way from the laser and is rarely reached
except in alignment, ranging, and surveying applications.

7-2 BEAM PROPAGATION

FOR STABLE LASER RESONATORS

The output of a stable laser resonator is usually a low-order transverse mo e,

frequently TEMoo, or some higher-order TEM(m mode.1 It is possible for
output to be multimode—that is, a mixture of various TEM modes,
axial modes may be present, but we can ignore them as far as beam prop
is concerned.
f the electf°"
Lowest-order mode beam propagation. Solution ot ^
magnetic wave equation in rectangular coordinates leads to a particu .^
equation for describing beam propagation. Equation (7-1) giyes
as a function of the cross-sectional coordinates x and y and
from a point where the phase front is flat.

, y, z) = /„
(I'

1
Resonator stability is discussed in Section 7-3.
 ornnaaation for Stable Laser Resonators 109
Beam rrup »
7
"
and cr,/ are proportional to the standard deviations of the
V functions in the respective directions. I0 is the irradiance
arn
]Ssian " , O{ at z = 0, which is a point of minimum spot size,
[,e cente Waist, and is normally (but not necessarily) located inside
,ually cal j tj are Hermite polynomials, of which the first few are
;he cavity, am
.nvenn here-
Ho(<7)=l H1(q) = 2q

ful relationships for Hermite polynomials are presented for future refer-
ence.

(7-3)
Hn+i = 2qHn — 2nHn-i
Here the prime refers to differentiation with respect to the argument. The generat-
ing function for Hermite polynomials is given as Eq. (7-4).

Referring to Eq. (7-1), we see that the irradiance for the TEMoo mode
at z = 0 is given by
x2 v2 \
( — +
2)
o-f o-y <7-5>

This may be written

J = / 0 e- r2/ - 2 (7-6)
for a
symmetrical beam.
ltl
e po ^ efln^ion of spot radius may be taken to be a-, which represents
more Pr * W ^e irradiance is one over e of its value at the center. A
a definiti
*ntten ° °n of radius for TEMoo beams results when Eq. (7-6) is

T
^lslhe . / = /oe—2 (7-7)
Value P mt at whlch the 2
at the c ° irradiance has dropped to one over e of its
Feache
d one oventer' This is the point at which the electric field amplitude has
as e f the cent
2 P°t of dj" ° ral value. Over 86% of the total power is contained
(left ameter 2>V Inte ration of
t0tal as an
an
' g the irradiance over a spot of diameter
Povver. SXercise) Yields the following relation between 70 and Pt, the

(7-8)
 The manner in which w(z) varies with z has been deter '
the scalar wave equation for a Gaussian beam and is prese ^ sc
with the equation for the phase front radius of curvature as Eq n
and Li 1966]. ^ 9) tK 0 j

(7--

Here w0 is the spot radius at z = 0 and ZR = TTWJJA is called the Ra

range. Notice that w(z) is symmetrical about z = 0. For very large v i

the divergence angle is given by w(z)/z = 6 = K/TTWO. This is called the far-
field divergence (diffraction) angle. It should be noted that this diffraction angle
is 0.52 times the far-field diffraction angle of the central maximum for a plane
wave passed through an aperture of diameter 2w0. Thus the fundamental (lowest-
order Gaussian) beam propagates with about half the divergence of a plane
wave of the same initial diameter in spite of the fact that the Gaussian beam
is a spherical wave. Also, no higher-order diffraction maxima are associated
with the Gaussian beam as with a plane wave incident on a finite aperture.
Equation (7-9) applies equally to a Gaussian laser beam emanating directly
from the laser,2 where z is then measured from the waist produced by t e
laser optical resonator, or to a focused laser beam. In the case of a loc
beam, z is measured from the point of best focus where the phase ro
planar. If the divergence angle of the beam entering a lens is small, e
is focused approximately at the secondary focal point. Because the conv ^
is rapid (small focused spot size), the lens is in the far field relative
focused spot. Therefore
A/ (?-'•

where w0 is the focused spot size and w± is the spot size at the le
The depth of focus of a focused Gaussian beam may also
from Eq. (7-9). Let w(z) = pw 0 at z = ±d. Then
rM"
TTWf.

2
If the output mirror is not flat the lensing effect of the mirror should oe
 p oaqation for Stable Laser Resonators 111
Beam rwv

re of how much the spot radius has increased at a distance

here p 's a me j sp0t. The depth of focus for a Gaussian beam is substantially
from the to higher-order or plane wave of the same diameter. For p =
'reater than o cgs approximately to d = ±w*/X.
7
•"o5. &• ( ~' '
order modes. Rectangular Hermite-Gaussian higher-order
^ H cussed for two reasons. First, many lasers tend to operate in
modes are^ ^ ^ ^^ operate higher-order. Second, the essential features
order mode beam propagation are contained in this discussion
[II u 6
*"C -' . . - - . . -
Of all ' ""tjiematics are basically straightforward, most having been worked
and ' 6 eo for the quantum-mechanical simple harmonic oscillator. Equation
n is the basic description of the irradiance of a rectangular higher-order
' H beam propagating in the z direction with x and y coordinate axes oriented
rarallel to the rectangular axes of the beam cross section.
William Carter proposed a definition of spot size for higher-order rectangu-
lar mode beams that uses the standard deviation of the Hermite-Gaussian func-
tions [Carter 1980]. Thus the x and y axes are parallel to the sides of the
rectangle defining the spot and the dimensions of the spot are 2a-x(z)m by
2o-,,(z)|.3 Using results of the solution for a quantum-mechanical simple har-
monic oscillator, for which Hermite-Gaussian functions give the probability
function, Carter was able to show that
<r.(z)m = (2m + l)1/2o-s(z) (7-12)
where crs(z)m is the spot size for a mode of order m, s is either the x or y
direction, and crs(z) is the spot size for the corresponding lowest-order mode.
F
'gure 7-3 contains plots of H^ (s/trs) exp (-sVor?) for several values of
nese plots represent the irradiance variations in either the x or y directions.
trrad °e '" the CaSCS shown in FiS- 7~3 that v*™ falls outside the outermost
[Lux n°e ^^ *n ^act> ll may ^e argued that this is true for all values of m
ics °nreasnd Parker 1981]. The expression H*m exp (-sV<r|) in quantum mechan-
goin
8 sim6?8 thC probabilitv (aside from a constant) of locating a particle under-
The
quanr S momc m°tion at position s relative to its equilibrium position.
n m
' is th asm 'S s^own to ^e the classical amplitude of the oscillator. As
Uan
'urn m L1"31111101 number) is increased, the correspondence principle of
fs a
P?roach amCS Predicts that the Peak probability should occur at a value
a!cu
'ated to m§ °'sm' At a value of m = 14 the peak Probabmty has been
ur at 5 = 0.91crsm. Under no circumstances can the peak

•
times the standard deviation of the function
0.2

-8.0 -6.0 -4.0

(a) Order number m - 3

0.2

-8.0
 7_2 Beam Propagation for Stable Laser nesonaiuia
Sec.

0.4 -

0.2 -

-8.0

(c) Order number m = 6

Figure 7-3 (continued)

probability fall outside the classical amplitude. Therefore it may be concluded

by analogy that all irradiance peaks for Hermite-Gaussian beams are contained
within a spot of dimensions 2(Ti| by 2crym. Most energy in the beam falls
within a spot of this size even for relatively low-order modes. Numerical integra-
tion for a TEMU mode shows that nearly 80% of the beam energy falls within
a
spot size defined in this way; 85% of the energy falls within such a spot for

It has been shown that the propagation of a higher-order rectangular

ea
m may be written very simply in Eq. (7-9) for the lowest-order beam. This
is repeated here with the beam size described in terms of cr, since
the
°ne over e2 point has no significance for a higher-order beam.
Xz
(7-13)
In
s
% (7- crs is the spot size for s equal to either x or y and erso is the
Pot Slze
for the focused beam. The modified equation for the spot size <rs(z)m
e
mth higher-order mode is

= o-s0(2m
114 Laser Optics PL
Chap 7

12
and o-s0(2m + I) ' = (rs0m, which is the spot size for the focused b
The far-field divergence angle is given by ^

Z 2TT(TSO ~15)

It is instructive to express the far-field divergence angle in terms of the act

focused spot size, crs0m.
(2m + 1)X
27ro-s0m -<»
Using Eq. (7-16), it is clear that the divergence angle for equal spot size is
actually (2m + 1) times larger for the higher-order beam.
The quantity crsm is a logical mathematical definition of spot size because
all the irradiance peaks and most of the power fall within a spot of area 2<rtm
times 2<ryn . This definition, however, does not lend itself to direct measurement
under any circumstances. It can be determined indirectly by making power
measurements through apertures, a process with many difficulties. There is a
more practical definition, at least for lasers where mode patterns can be readily
viewed, such as burn patterns in acrylic for COa lasers. This definition simply
uses the distance between the centers of the outermost irradiance peaks in
both the x and y directions. For convenience, half this distance will be used
and symbolized by Ds(z)m. Equation (7-14) is easily modified if the relation
between Ds(z)m and crs(z)m is known. If we write

A(z) m =-^-cr s (z) m (7-17)

km

then Eq. (7-14) may be written

A(z)m = Ao(2m + l)i/

The values of km have been numerically computed for m values up to 14 a°

are listed in Table 7-1 [Luxon and Parker 1981].
Equation (7-18) is more useful if it is written entirely in terms o f t e
focused spot size Aom = Ds0(2m + I) 1 ' 2 .

n<^ -n
A(z) m - Aom
f .1 +i
L V27Tfc 2 n A 2 0
The far-field divergence angle may now be written
_ X(2m + 1) (7-20)
\) —

When a laser beam is focused with a mirror or lens, the converge

so rapid that the focusing element is in the far field in terms of distance
 Beam Propagation for Stable Laser Resonators 115

TABLE 7-1 Values of km for

m = 1 through m = 14

1 1.73
. 2 1.42
3 1.30
4 1.24
5 1.20
6 1.18
7 1.16
8 1.14
9 1.13
10 1.12
11 1.11
12 1.11
13 1.10
14 1.10

the focused spot. Assuming that the best focus occurs, to a sufficient approxima-
tion, at the focal point, Eq. (7-19) reduces to Eq. (7-21) for the focused spot
size.

Xf(2m + 1)
(
'
It is worth noting the inverse relationship between spot size and divergence
angle. This leads to the following important result: when a laser beam of any
order is expanded and recollimated, the divergence angle decreases proportional
to
the increase in beam size.
A useful definition of depth of focus for higher-order mode beams can
be deduced from Eq. (7-19) in the same manner that Eq. (7-9) was deduced
rom
% (7-8). Again, let z = d, Ds(d)m = pDs0m, and solve for d:

(7 22)
s -
lmportant to note
foe the inverse effect of the mode number on depth of
focn effect is usually more than compensated for, however, by the larger
Se s ot
T P size associated with high-order beams.
Gau • higher-order mode equations can be applied to circular (Laguerre-
1)_ ^lan) m°des by setting kt = 1 and replacing (21 + 1) by 2(2 p + m +
t^hiir £re P anc* m are tne number of angular and radial nodes, respectively
of polps and Andrews 1983]. The beam radius, in this case, relates to the fraction
beam1 er l^at will pass through an aperture of this radius centered on the
axis.
116 Laser Optics Cha P. 7

7-3 RESONATOR STABILITY AND BEAM PARAMETERS

When referring to the optical stability of a laser resonator, we are talking ab

whether any rays, geometrically speaking, are trapped within the resonate
We can model the optical cavity as an infinite series of lenses with alternatin
focal lengths equal to the focal lengths of the mirrors with spacing betwee
them equal to the optical length of the cavity. If any rays propagating nea
the axis of such an infinite series of lenses are periodically refocused, the system
is stable. If they all diverge, it is unstable. By a purely geometrical analysis, it
can be shown that all stable cavities obey the following relation.
0 < gig2 < 1 (7-23)
Here gi = 1 — L/ri and g 2 = 1 — L/r2, where /-i and r2 are the radii of
curvature of the back mirror and the output mirror, respectively, and L is
the resonator optical length. These g factors were introduced earlier while dis-
cussing the frequencies of TEM modes. A stability diagram is constructed by
plotting gig 2 = 1 in Fig. 7-4. Points lying on the lines or in the shaded region
correspond to stable laser resonators. Points lying on the axes or the hyperbola
are referred to as marginally stable because a slight shift in laser length or
mirror radius of curvature can cause the system to become unstable.
Some points of particular interest on the stability diagram are

1. gi = 1, g 2 = 1, which represents the plane parallel mirror Fabry-Perot

interferometer.

(a) Concentric symmetrical las'*

(b) Confocal symmetrical Figure 7-4 Stability diagram f°r
(c) Plane mirrors resonators.
 7-3 Resonator Stability and Beam Parameters 117

a = 0, g2 — 0> which is a symmetrical confocal resonator. This configura-

tion leads to the lowest diffraction losses in the TEMoo mode compared
with the higher TEM modes.
, gj = — 1, #2 = ~1, which is a symmetrical concentric configuration.

simple rule to determine whether a resonator is stable can be stated. If either

, center of curvature of one mirror or the mirror itself, but not both, falls
hetween the other mirror and its center of curvature, then the reasonator is stable
[Yariv 1968]. If this rule is violated, the resonator is unstable.
Many factors affect the selection of a particular mirror configuration. One
mentioned earlier is whether TEMoo operation is desired. The confocal resonator
exhibits the lowest diffraction losses and the highest selectivity due to a large
differential in diffraction losses for the TEMoo mode and higher-order modes.
This configuration also does a fair job in regard to maximizing mode volume.
Mode volume is the volume of the active medium that is actually filled by
radiation. Several resonator mirror configurations are shown in Fig. 7-5, along
with the approximate mode volume for each.
The confocal mirror configuration is desirable from the standpoint of
TEMoo mode selectivity, but it is quite hard to align because x, y and angular
alignment of the mirrors are critical. The concentric configuration has the same
problem. Both are marginally stable. The problem of marginal stability is not
too difficult because the advantage of the confocal resonator can be nearly
achieved by selecting a configuration that lies safely in the stable region of
Fig. 7-4 but that is close to the origin.
The plane mirror arrangement has the best mode-volume-filling capability
but results in an output that is heavily multimode and it is a difficult system
to align. Furthermore, the TEM modes are degenerate, making low-order trans-
Verse
mode operation impossible.
A popular configuration is part (d) in Fig. 7-5 because of the relative
ase
°f alignment and good mode-filling characteristics. This arrangement is
equently used in high-power lasers, with the radius ri being several times
he 0
Ptical length of the laser.
, . Whether a stable resonator laser operates in the lowest-order mode or
t ,er"or<^er largely depends on the size of the effective aperture in relation
e
cavity length. A gas laser with a long narrow bore tubular arrangement
en
d to operate low order because of the attenuation of the higher-order

la
> Confocal ( b ) Fabry-Perot (c) Concentric (d) Spherical-flat

Figure 7-5 Stable resonator configurations.

118 Laser Optics Cha

modes by the inner walls of the tube. Remember, the higher-order modes
associated with higher diffraction or divergence angles and so are more
attenuated by a limiting aperture.
A measure of the tendency for a stable laser cavity to operate low
higher order is the Fresnel number, given by

where a, the effective aperture radius, should be taken as the radius of the
smallest aperture in the system if it is not too far from one of the mirrors. In
a system where the only limitation is due to the mirrors, assuming thai they
have equal radii (otherwise let a2 = aiO2, where GI and a2 are the radii of
the two cavity mirrors), the Fresnel number represents the number of Fresnel *
zones (interference rings) intercepted by one of the mirrors if a uniform plane
wave illuminates the other mirror. A Fresnel number of one or less generally
results in lowest-order mode (TEMoo) operation.

Stable cavity parameters. Expressions for the waist size and its
location relative to the laser mirrors have been derived for lowest-order Gaussian
and higher-order mode cases [Kogelnik and Li 1966]. For the lowest-order
Gaussian, the waist radius formed inside (or outside) the laser cavity is given
by
L6152V
w0= — 5^J (7_25)
\ T' (gi + 82 ~ 2gig2)1/2
where Wo is the radius measured to the one over e squared irradiance points.
Notice that Eq. (7-25) is an indeterminate expression for certain values of gi
and g2. When gi = g2 = 1 (two plane mirrors), the spot size is aperture limited
and has nothing to do with the resonator parameters. For the gi = £2 =
case (symmetrical confocal resonator), if gi and g2 are replaced by g and soffl*
rearranging carried out, Eq. (7-25) reduces to

T/
The location of the beam waist relative to the output mirror, determin
by purely geometrical means, is given by
g t (l - g2)L v(7_27)
z2 =

Waist location has no meaning for the Fabry-Perot arrangement. The wai ^
located midway between the mirrors for all symmetrical resonators. For ^
with a plane output mirror and spherical back mirror, the waist is l°ca
z 2 = 0 or right at the output mirror.
 Resonator Stability and Beam Parameters 119
Sec.

Based on Carter's results, Eq. (7-23) has been modified by Luxon and
rker for Hermite-Gaussian higher-order modes [Luxon and Parker 1981].
„ using the definition of spot size introduced earlier for higher-order modes
(half the distance between outer irradiance peaks), the waist size becomes

A(0)m = (7-28)
I (gl . - 2glg2)1'2
The relation for the location of the waist is not altered by higher-order mode
operation, for this location depends on purely geometrical factors.

Unstable resonators. Unstable optical resonators are frequently used

in high-power, high-gain lasers in which a relatively low percentage of feedback
is required to achieve lasing operation. Although an infinite number of possible
unstable resonator configurations are possible, Fig. 7-6 typifies the type used
in commercial lasers, particularly multikilowatt CO2 lasers.

Figure 7-6 Unstable confocal resonator.

In an unstable cavity the rays "walk off" the axis and exit past the output
mirror unless some mirror arrangement is used to intercept the rays and deflect
e in som
of ™ e other direction (as is usually the case). In any event, the "walk off"
e ra s
>' leads to a ring-shaped output in the near field with some substructure
aue
to diffraction effects.
just nstak'e resonators are capable of producing a variety of transverse modes,
a s stab
m oo, le resonators do. Yet the losses associated with the higher-order
S
low ° £reat tnat>> m practice,, such lasers nearly always operate inntheir
er
f rder tran
°cus' "° sverse mode. This mode can result in a divergence angle and
the j,!"^Capabuity aPproximately equal to the Gaussian mode. When focused,
hole j ance distribution is similar to that of a focused Gaussian beam.. The
the
center of the beam in the far field is filled in and the irradiance
again is approximately Gaussian.
USCful results can be deduced
from Fig. 7-6 for a confocal unstable
The ratio of ai to a2, the outer and inner radii of the output ring,
120 Laser Optics rk
^ap. 7

is called the magnification tn and is related by similar triangles to the mi

r
radii of curvatures by
_ ai _ ri/2 _ /"i
@2 TI/ 2. /*2 j

where use has been made of the fact that the focal length of a spherical mirr
is half its radius of curvature. Note that a2 is the radius of the output mirr
but a i is the minimum radius of the back mirror.
If it is assumed that the radiation uniformly fills the back mirror up to
a radius a\, the fraction of the power emitted (geometrical loss G) is given bv
ir(a\ — af) _ m2 — 1 _ r\ — r\ _ r\
Traf m2 r\ r\ (7-30)

This provides a reasonable estimate of the radiation geometrically coupled out

of a confocal unstable resonator. It is also useful to note that

i=/i-/. = §-^ (7-31)

where all quantities are taken to be positive.

An equivalent Fresnel number can be defined for unstable resonators and
for the confocal resonator discussed here [Svelto 1976].

#„ = - — ( m - 1 ) (7-32)

For m near 1 Neci < N and excellent mode discrimination occurs for all half-
integer values of Nen.

Effective thin lens equation. S. A. Self has developed a useful equa-

tion that resembles the thin lens equation which relates Gaussian waist locations
to lens focal length and the Rayleigh range [Self 1983]. A thin lens trans-
forms the radius of curvature of a Gaussian beam according to

J_ + _L = i (7-33)
RI R2 f
where RI is radius of curvature entering the lens and R2 is radius of curva
leaving the lens. The sign convention is the same as of o, i, and / for a
lens as given in Chapter 1. Using Eq. (7-33) along with Eqs. (7-9) and t
fact that a thin lens leaves the beam size unchanged immediately on e1
side of the lens leads to
. J
1 -= _
1 (7~34)
 7 Problems 121
Chap- '

h re o is the waist location for light entering the lens and / is the waist
* ation for light leaving the lens. The sign convention on o and / is still the
same as in Chapter 1.
Self also presented an expression for magnification
™ = [ ( l - o / / ) 2 + (Z«//)3]-1/2 (7-35)
hich can be used for spot size determination
It has been pointed out by Luxon et. al. 1984, that Eqs. (7-34) and
n_35) are valid for beams of all modes. For Hermite-Gaussian beams the Ray-
leigh range is given by Eq. (7-36)
,, .,.
w (7-36)
K(2m + 1)\
It should be noted that Z« is purely geometrical and can be calculated
from
~ gig2)]1/2 . ,
gl + g2 — 2glg 2

if the resonator parameters are known.

PROBLEMS

7-1. The total power emitted by a CW laser operating in the TEMoo mode is 10 W.
If the focused spot size is 0.2 mm, calculate the irradiance at the spot center.
7-2. The beam radius for a HeNe laser (X = 0.6328 jam) operating TEMoo is 0.5
mm. Determine the spot size at distances of 1 m and 500 m from the waist.
7-3. If the spot radius for a HeNe TEMoo beam entering a 12.5-cm focal length lens
is 1.5 cm, calculate the focused spot size.
7-4. Calculate the depth of focus of the conditions given in Problem 7-3 for a spot
size increase of 10%.
~'. A CO2 laser (X = 10.6 ju.m) operates in a rectangular TEM64 mode. The beam
dimensions, outer peak-to-peak distance, are 2.5 and 2.3 cm. Calculate the focused
spot size for both spot dimensions for focusing with a 25-cm focal length lens.
Calculate the depth of focus for the conditions given in Problem 7-5 for a 10%
s
Pot size increase.
• A HeNe laser operating TEMoo has a resonator length of 1 m and mirror radii
r =
i 3 m and r2 = 2 m. Locate the waist and determine the spot size at the
w
aist. Also, determine the spot size and divergence angle at the output mirror

A CO2 laser operating rectangular TEM57 has a resonator length of 5 m and

mirror radii /-j = 30 m, r 2 = 20 m. Locate the waist and determine spot dimensions
the waist. Also, determine the spot dimensions and divergence at the output
mirror (r2).
122 Laser Optics Chap. -,

7-9. Prove whether the following mirror configurations are stable and plot them
n
a stability diagram:
(a) L = 2 m, rl = r2 = 2 m
(b) L = 1.5 m, ri = r2 = 0.75 m
(c) L = 5 m, ri = 30 cm, r2 = 20 m
(d) L = 1 m, r-i = 4 m, r-i = —2 m
(e) L = 1.5 m, ri = 3 m, r 2 = —2 m
(f) L = 3 m, n = 2 m, r2 = 2.5 m
7-10. Design a stable laser resonator of 30-cm length that will provide a waist at th
output mirror with a diameter of 1 mm for a HeNe laser operating TEM<x>. What
should the maximum aperture be for this laser to ensure TEMoo operation?
7-11. Design a confocal unstable resonator with a length of 1 m and an output mirror
diameter of 2.5 cm that will provide 40% output coupling. What are the magnifica-
tion and equivalent Fresnel number for this resonator? Do you expect good mode
discrimination for this resonator if X = 10.6 /xm?
7-12. For Problem 7-8 treat the output mirror as a thin lens and determine the new
waist location, size and Rayleigh range. Use this information to locate the waist
and its size produced by a 12.7 cm focal length lens placed 10 m from the output
mirror. Compare your results with the results based on the assumption that the
waist produced by the lens occurs in the focal plane.

REFERENCES

CARTER, W. H., "Spot Size and Divergence for Hermite-Gaussian Beams of Any Order,"
Applied Optics, 19, No. 7, April 1980, 1027-1029.
KOGELNIK, H., and T. Li, "Laser Beams and Resonators," Applied Optics, 5, No. 10,
October 1966, 1550-1567.
LUXON, J. T., and D. E. PARKER, "Higher-Order CO2 Laser Beam Spot Size and Depth
of Focus Determination," Applied Optics, 21, No. 11, June 1, 1981, 1933-1935.
PHILLIPS, R. L., and L. C. ANDREWS, "Spot Size and Divergence for Laguerre Gaussian
Beams of any order," Applied Optics, 22, No. 5, Mar. 1, 1983, 643-644.
SELF, S. A., "Focusing of Spherical Gaussian Beams," Applied Optics, 22, No. 5, March
1, 1983, 658-661.
SVELTO, O., Principles of Lasers. New York: Plenum Press, 1976, 135-142.
YARIV, A., Quantum Electronics. New York: John Wiley & Sons, Inc., 1968, 226.

BIBLIOGRAPHY

7-1. Verdeyen, J. T., Laser Electronics. Englewood Cliffs, NJ: Prentice-Hall, Inc., 1
7-2. Young, M., Optics and Lasers: An Engineering Approach. New York: Sp
Verlag, 1977.
7-3. Siegman, A. E., An Introduction to Lasers and Masers. New York: Me
Hill Book Co., 1971.
 7 Bibliography 123

7^4 Maitland, A., and M. H. Dunn, Laser Physics. New York: John Wiley & Sons,
' Inc., 1969.
7 5 Nemoto, S., and T. Makimoto, "Generalized Spot Size for a Higher-Order Beam
Mode," J. Opt. Soc. Am., 69, No. 4, April 1979, 578-580.
7 6 Luxon, J. T., and D. E. Parker, "Practical Spot Size Definition for Single Higher-
Order Rectangular-Mode Laser Beams," Applied Optics, 21, No. 11, May 15,
1981, 1728-1729.
7_7. Luxon, J. T., D. E. Parker, and J. Karkheck, "Waist Location and Rayleigh
Range for Higher-Order Mode Laser Beams," submitted to Applied Optics, July
1984.
 chapter

Types
of Lasers

Most of the lasers found in industrial applications are discussed in this chapter,
including some, such as the dye laser, that are used primarily in laboratory
facilities. In most cases, only general descriptions of wavelength, output power,
mode of operation, and applications are given. As a specific example, the C02
laser is discussed in some detail.
Laser types can be categorized in a number of ways. In general, there
are gas, liquid, and solid lasers. Subclassifications of neutral gas, ion, metal
vapor, and molecular lasers exist in the category of gas lasers. The only liquid
laser discussed here is the dye laser. Solid lasers may be crystalline, amorphous
(at least in the case of the Nd-Glass laser), or even a semiconductor.

8-1 HeNe LASERS

The HeNe laser is the most common of all the visible output lasers (see rig-
8-1 for a photograph). It emits in the red part of the visible spectrum at 0.63
jLim. The power levels available range from 0.5-mW to 50-mW CW
Most HeNe lasers are constructed with long, narrow bore (about 2 mm diarn
is optimum) tubes, resulting in preferential operation in the TEMoo mode,
though TEMf0 may occur if good alignment of the mirrors is not maintain
The electrical glow discharge is produced by a dc voltage of 1 to 2
at about 50 mA. A starting voltage of several kilovolts is applied to get )]tf
discharge started. The HeNe laser is a low-gain device that requires op
 8 _1 HeNe Laser 125

Figure 8-1 Photograph of a 1-mW HeNe laser with the cover removed.

99% feedback to sustain oscillation. Consequently, the mirrors are multilayer

dielectric on a dielectric substrate for both back and output mirrors. The output
power is 25 mW/m of tube length.
HeNe lasers have a gas mixture with Ne at a partial pressure of 0.1 torr1
and a total pressure of 1 torr. Ne is the lasant, but He plays an important
role in the excitation process. In a four-level process He is excited by collisions
with electrons to a level coincident with an excited state of Ne. During collisions,
the energy is resonantly transferred from He to Ne. The Ne decays to one of
three possible intermediate states with three possible emission lines, one at 0.6328
M*m and two in the ir at 1.15 and 3.39 jam. The mirrors act as selective niters
for the 0.6328-jum wavelength.
Figure 8-2 shows the energy level diagrams of He and Ne and the pertinent
transitions. The excited states of He that are involved are metastable, thus
ensuring a high probability of resonant energy transfer to Ne. Neon atoms
relax to the 1 S level after lasing. This state is metastable and must be depopulated
°y collisions with the wall of the tube. Therefore increasing the bore of the
u
oe past about 2 mm will decrease power output rather than increase it.
Some HeNe lasers are designed to operate in a few low-order modes,
o-called multimode operation. They may also be purchased with either linear
°r random polarization. The mirrors are attached directly to the tube for random
Polarization. At least one Brewster window is used for linearly polarized output.
The bandwidth of HeNe lasers is 1500 MHz because of Doppler broaden-
8- Consequently, these lasers generally operate in several axial modes. Shorten-
8 the cavity to about 10 cm usually ensures single axial mode operation.
In single axial mode operation there will be just one dip in the gain curve
t° "hole burning." This dip can be centered by adjusting the cavity length.

A. torr is approximately 1 mm of mercury.

126 Types of Lasers ch

3.39 jim

0.6328 urn

1.15 /i

1s

He Ne

Figure 8-2 Energy level diagram for HeNe laser.

As the laser frequency drifts from this central value, the power will begin to
increase (a transient effect). The power change can be monitored by a beam
splitter and photodetector. If one mirror is mounted on a piezoelectric crystal,
a servosystem can be used to stretch or contract the crystal slightly (by means
of an applied voltage), thus adjusting the cavity length continuously to maintain
minimum power. In this manner, a frequency stability of one part in 109 can
be achieved.
High stability can be gained by placing a gas in a separate cell with a
coincident absorption line in the laser cavity. I'29 is used with HeNe lasers.
I'29 has a very narrow absorption line at 0.633 jitm. By locking the HeNe laser
on the center of this absorption and saturating the IJ29, frequency stability of
one part in 1012 to 1013 can be achieved.
The diversity of HeNe laser applications is too great to cover exhaustively
and so some representative applications are mentioned here. More detail on
selected low-power applications can be found in Chapters 9 and 10.
HeNe lasers are used in bar code reading, image and pattern recognition,
alignment operations in industry, construction, sewer, and tunnel work, an
surveying. In addition, they have a variety of nondestructive testing (N"
applications, including surface flaw and roughness determination. Measurerne
applications include interferometric displacement measurement, absolute s
measurement, and edge location, using various types of detectors and detec
arrays. Frequently HeNe lasers are used to align other lasers and to l°ca
the point of incidence on a work piece for a high-power laser.
HeNe lasers are also used in line-of-sight communications applicati° '
such as control of construction equipment like earth graders to maintain pr°"
 „ o Ion Lasers 127
SBC- »-*

jg or depth requirements. A major application of HeNe lasers is in the

g
cording and playing back of holograms.

B,2 ION LASERS

Common ion lasers are argon (Ar), krypton (Kr), which has nothing to do
w ith the Superman, and xenon (Xe). The power levels of these lasers vary
from milliwatts to several watts. The major wavelengths available for each are
listed in Table 8-1.

TABLE 8-1 Wavelengths and power ranges

for ion lasers

Laser Wavelength (fxm) Power Range (W)

Ar 0.5145]
0.005-20
0.488 J
Kr 0.6471 0.005-6
Xe 0.995-0.5395 200 pulsed

The ion lasers, as the name implies, use an ionized gas as the lasant in
an electrical discharge in a plasma. Basically they are four-level systems, but
it is the excited states of the ions that are involved in the lasing process. Excitation
is a two-collision process: the first collision ionizes the atom; the second provides
the necessary excitation.
The current densities required in ion lasers are about 1000 A/cm2, which
limits the tube bore diameter to a few millimeters. The ions drift toward the
cathode and so a return path outside the discharge is provided, usually by
means of a segmented BeO tube. An axial magnetic field is applied by means
a
current-carrying coil wrapped around the tube. This step tends to prevent
ectr
j; ons from losing energy through collisions with the wall since the magnetic
e
ld causes them to undergo a spiral motion along the tube axis.
Argon lasers are used extensively in surgery, particularly of the eye. This
lc
e is used to bounce light off the retroreflectors that the astronauts placed
of u moon *°r moonquake and meteor impact studies as well as measurement
! the distance between the earth and moon. The Ar laser, along with the Kr
> is used in spectroscopic work, such as Raman spectroscopy. Both lasers
^ tunable over several wavelengths in addition to the ones listed in Table
^ • ihe Xe laser can be used for some materials processing, such as cutting
in rem°val of thin metallic films. In this type of work the Xe laser is operated
Pulsed fashion with peak powers of around 200 W.
tj hg Ar laser has been used in holographic work to achieve shorter exposure
a
nd thereby minimize the effects of vibration. The development of high-
128 Types of Lasers r-k
Lha
P. 8

speed, high-resolution emulsions for holography has reduced the need f 0r A

r
lasers in holographic work.

8-3 METAL VAPOR LASERS

Many types of metal vapor lasers exist, but the He-Cd and He-Se lasers a
the most important commercial ones. Although essentially ion lasers, they behav
similarly to the HeNe laser in that the metal ions are excited by collisions
with excited He atoms. The excitation process in the He-Cd laser is called
Penning ionization. In this process, an excited metastable He atom collides
with a Cd atom, thereby ionizing it and raising it to an excited ionic state
Lasing action takes place between the excited ionic state and the ionic ground
state of Cd. The He-Cd laser has wavelengths at 0.441 /Am in the blue and
0.325 jj,m in the uv. Excitation in the He-Se laser is by charge transfer. An
excited ionized metastable He atom collides with a neutral Se atom, transfers
the charge to it, and excites it. Emission occurs between excited states of the
Se ion with about 19 lines possible in the visible spectrum.
In metal-vapor lasers the metal is vaporized near the anode and condenses
at the cathode, a process referred to as cataphoresis. Consequently, the metal
is consumed, for example, at a rate of 1 g of Cd per 1000 hours of operation.
The power output of these metal-vapor lasers is in the 50- to 100-mW
range. The current densities required for the one-step excitation process in metal-
vapor lasers are much lower than for single-element ion lasers and so they
can be air cooled and operated from 110-V ac.
Some applications for metal-vapor laser are light shows, ful! color image
generation, spectroscopy, and photochemical studies.

8-4 DYE LASERS

Dye lasers, as the name suggests, use an organic dye as the lasing medium
Commercial dye lasers are pumped by nitrogen, argon, Nd-YAG, and excime
lasers, although Xe lamps can be used. Operation is usually pulsed at 100 pp <
but average powers of 10 mW are attained.
The organic dyes used, such as rhodamine 6G (xanthene dye), have
dreds of overlapping spectral lines at which they can lase. Consequent y, ^
dye laser with a specific dye can, in effect, be tuned continuously over a a
portion of the visible part of the spectrum. By using several dyes, which ^
automatically interchanged, the entire visible portion of the spectrum, as
as some of the ir and uv, can be scanned with relatively equal power ou
at each wavelength. This capability is of great value in certain types of sp .^
scopic investigations in or near the visible wavelengths and also in photocn
try, including uranium isotope separation, and pollution detection.
 g _5 Ruby Laser 129
Sec-

L
Lase

i
Figure 8-3 Energy levels for an organic
dye. Singlet Triplet

Figure 8-3 is a representative energy level diagram for an organic dye.

The diagram is composed of two types of states: the singlet states labled So,
Si: and Sz and the triplet states labeled 7\ and T2.2 Each band of energies
consists of widely spaced vibrational levels and overlapping rotational levels.
Radiative transitions are quantum mechanically forbidden between 5 and T
levels, but transitions between one S level and another or one T level and
another are allowed. The dye molecule is optically pumped to the Si level
and drops to the lowest level in that band. Lasing takes place between that
level and one of the many possible S0 states. Continuous operation is not practical
in most systems because intersystem crossing from Si to TI states, caused by
collisions, results in a buildup of dye molecules in the T\ states and optical
absorption from 7\ to T2 states. Pulse lengths are kept short compared with
fte intersystem crossing time of about 100 ns.
The bandwidth of dye lasers is 1013 Hz, which makes it possible to achieve
Picosecond pulses by mode locking.

8
~5 RUBY LASER

b e ruby laser was the first optical maser (or laser, as it came to be called)
• The first such laser was constructed by Maiman at the Hughes Laboratories,
a rub U) California, in 1960. Figure 8—4- contains a schematic representation of
y laser.
'he active material in the ruby laser is triply ionized chromium, Cr3+,
in (i ^aiglet and triplet states have zero and one, respectively, for the total spin of electrons
olecule and determine the multiplicity of the energy levels.
130 Types of Lasers Chan

Lamp _
, Crystal or
/ glass rod

Lamp

Figure 8-4 Schematic of solid laser — for example, ruby, Nd-Yag, Nd-Glass.

and the host is crystalline aluminum oxide (sapphire), A^Os. C^Oa substitutes
for about 0.05% by weight of the A12O3, providing the crystal with a characteris-
tic light pink color. The deep red color of gemstone ruby is the result of much
higher concentrations of Cr2Os. The output of the ruby laser is at 0.6943 pm
in the red.
Ruby lasers are operated in pulsed fashion with repetition rates the order
of one pulse per second. The energy per pulse may be as high as 100 J but is
usually a few joules. The normal pulse width is several milliseconds long, but
(2-switching can be used to produce pulses of tens of nanoseconds to micro-
seconds and correspondingly higher peak powers.
The ruby laser is a three-level laser and as such requires a high threshold
for lasing action to occur, since over 50% of the Cr3+ must be raised to the
excited state to achieve population inversion. Several joules of energy per cubic
centimeter are required to establish a threshold-level population inversion. A
great deal of the energy from the Xe flash lamps used to excite the laser goes
into heating the laser rod, consequently the duty cycle is limited. This problem
is aggravated by the fact that ruby has a very low thermal conductivity. Cooling
with refrigeration is used, but continuous operation is not practical.
The output of ruby lasers may be randomly or linearly polarized, depending
on how the crystal is cut — that is, whether the c axis is parallel or perpendicular
to the laser axis. The c axis parallel to the laser axis produces random polariza-
tion. When the c axis is perpendicular to the laser axis, linear polarization
occurs.
The pulsed output of a ruby laser consists of random spiking due to neav y
multimode operation. Q-sv/itching is used to produce tens of megawatts pe
power in 10- to 20-ns pulses. Ruby lasers are easily mode locked because
the multimode operation and giga watts of peak power in 10-ps pulses can
achieved. .
The ruby laser was the first to be used for piercing holes in diarnon
for wire-pulling dies. It can be used for piercing small holes in many maten ^
and for spot welding. It is also used in applications where short, g-switc
 g_7 Nd-Glass Laser 131

ulses of light in the visible are required. One such application is in pulsed
holography- By using high-energy submicrosecond pulses from a ruby laser,
holograms of moving objects can be made, thereby eliminating the need for
careful prevention of even the slightest motion during exposure as required
with low-power CW lasers.

8_6 NEODYMIUM-YAG LASERS

The neodymium (knee-oh-dim-e-umm) Nd-YAG laser uses triply ionized

Nd as the lasant and the crystal YAG (yttrium-aluminum-garnet) as the host.
YAG is a complicated oxide with the chemical composition Y3Al5Oi2. The
amount of metal ions replaced by Nd3+ is 1 to 2%. The output wavelength is
1.06 fxm, which is in the near ir.
Nd-YAG lasers are capable of average power outputs up to 1000 W.
YAG lasers are operated continuously up to a few hundred watts, but pulsed
operation, at high repetition rates, is used at the higher power levels.
The YAG laser is a four-level system, which results in a lower threshold
energy to achieve the necessary population inversion. Also, YAG has a relatively
high thermal conductivity; so with cooling CW or high repetition pulsed opera-
tion is practical. CW YAG lasers are Q-switched with acousto-optical modula-
tors to provide pulse rates of up to several thousand pulses per second, although
average power begins to drop above 2000 pps.
Schematically the YAG laser head appears similar to that of ruby. Figure
8-5 is a photograph of a commercial Nd-YAG laser.
YAG lasers can be used for overlapping pulse seam welding, spot welding,
hole piercing (including diamonds and other gemstones), and cutting. Cutting
is accomplished by rapid overlapping hole piercing with a gas jet assist. Through
frequency doubling and higher harmonic generation, YAG lasers are used for
material property, photochemical, and interaction studies in the visible and
ultraviolet ranges.

8
-7 Nd-GLASS LASER

e Nd-Glass laser uses Nd3+ as the lasant, but the host material is glass.
he output is still 1.06 /j,m, although the bandwidth is several times that of
e Nd-YAG laser, which partially explains a much higher pulse energy output.
e large bandwidth results in hundreds of axial modes operating simultaneously.
°de locking puts all these modes in step, timewise and spatially, thereby
suiting in a large bandwidth, about 1012 Hz, and a correspondingly narrow
Pulse width of as low as 10~12 s.
. High energy is achieved by using a physically larger active medium than
lasers, which is practical with glass, for it can be made in excellent
 Flooded housing for lamp
and YAG rod

I TEMOO mode selector

Q-switch

Harmonic generator

Figure 8-5 Nd-YAG laser. (Courtesy of Quantronix Corp.)

 Q 8 Diode Lasers 133
See- o-G

ijty to almost any size. YAG crystals, on the other hand, are quite expensive
d difficu^ to grow in large size. Unfortunately, glass has a low thermal conduc-
• ity as a result, Nd-Glass lasers are limited to low-duty cycle operation of
bout one pulse per second. The energy per pulse can be 100 J or more, making
,tfle j^d-Glass laser suitable for many pulse-welding and hole-piercing applica-
tions.
One phenomenon occurring in solid lasers that can cause excessive beam
divergence and self-focusing results from the fact that in some solids the refractive
index increases with the irradiance of the beam. This creates a situation similar
to that of a graded index optical fiber. The beam is actually guided by the
self-induced refractive index charge. If the irradiance exceeds a threshold level,
the beam is focused to a small spot, usually damaging or destroying the rod.

8-8 DIODE LASERS

Diode lasers, as might be guessed, are diodes that lase, not lasers that diode.
So the phrase laser diode is incorrect (if anyone cares). There are about 20
known compound semiconductor diodes that lase with wavelengths ranging
from 0.33 u,m in the uv to nearly 40 u,m in the ir. Electron beam excitation
is used for all wavelengths shorter than 0.8 /j,m. In most of the others e-beam
or injection current (current in a forward-based diode) can be used. For practical
purposes, only the injection current type is discussed. The most common types
of commercially available diode lasers are based on GaAs and ternary com-
pounds, such as AltGai-^As. The A^Ga^As compounds are discussed here
as an example of diode lasers.
The first diode lasers operated in the early 1960s were simple homojunction
GaAs devices, meaning that there was only one junction between the p- and
"•type GaAs. Figure 8-6 is a schematic representation of a homojunction diode
laser and the corresponding energy level diagram. The cavity is formed by
the cleaved ends of the GaAs chip, which is about 0.5 mm wide and 0.1 mm
hl
gh. Radiation at about 0.9 /Am is emitted from the depletion region. Contact
ls mad
e through gold, Au, deposited on the anode and the Au substrate that
"e chip is mounted on. Notice that the Fermi level is in the bands in the
^nergy level diagram. This condition is indicative of degenerate (extremely heavy)
°Ping approaching the solid solubility limit. Forward biasing of the diode
'is in a large injection current and a population inversion in the depletion
gion. The reason for the population inversion can be seen by inspection of
p §• °~6. Only pulsed operation at liquid nitrogen temperature (77 °K) is possible.
e
s with peak powers of a few watts are possible with repetition rates of
Seve
ral thousand pps.
Modern versions of the GaAs diode laser use single or double heterojunc-
s
With stripe geometry and possibly large optical cavity (LOG) configurations
ec Uce
* the current density and risk of damage due to large radiation fields
134 Types of Lasers Chan

Holes -

Electrons' 'T
(a) (b)

(c)

Figure 8-6 Homojunction diode laser (a) diode, (b) energy bands, unbiased, and
(c) energy bands, forward biased.

in the chip. A heterojunction refers to a junction between GaAs and Alj-Gai-i As.
A double heterojunction device with stripe geometry is depicted in Fig. 8-7.
The purpose of the stripe contact is to improve conduction of heat from the
active region. The ternary compounds of AUGa^j. As have much poorer thermal
conductivity than GaAs. Limiting the active region to a narrow stripe allows
lateral heat conduction in GaAs to greatly reduce the temperature rise in the
active region.
The purpose of the heterojunctions can best be explained in terms of
the energy level diagram shown in Fig. 8-8. Note that the energy gaps are
wider in Alj-Gai-^As than in GaAs. This factor has two effects. First, the
resulting energy barriers reflect both electrons and holes back into the active
GaAs region, thus reducing the current density required to produce stimu lated
Ac
emission. Secondly, there is a refractive index difference between AlrGai-i^
and GaAs—3.4 for x = 0.4 and 3.6 for x = 0—which causes light traveling
at slight angles to the cavity axis to be strongly reflected back into the activ
region so that less radiation need be generated, thereby lowering the require
current density. Devices of this sort can be operated CW at room temperaW
with the aid of a thermoelectric cooler to prevent an excessive tempera
rise. The current density required is 1000 A/cm2. If the bottom heterojunctt°^
is replaced by an ordinary pn GaAs junction, the result is a single heterojunc ^
device that can be operated in pulsed fashion at room temperature. Power
are 10 to 20 W peak with repetition rates of up to 104 pps.
LOG devices are made with two additional layers of AlyGai-j/As on
 8_8 Diode Lasers 135
Sec.

Contact

-Insulator

Ge-doped GaAs

type

p type Zn-doped Al x Ga,_ x A5

p type Si and Zn-doped GaAs 0.5 nm active region

n type Sn-doped Al Ga, As

Substrate1 GaAs

-Contact

Figure 8-7 Double heterostructure diode laser with stripe geometry.

: of the p-GaAs with y <§ x so that the optical cavity becomes much wider,
i of micrometers compared to perhaps 1 fj.m, thus reducing the danger of
aage to the crystal from the radiation. This feature also greatly reduces
diffraction of the beam as it leaves the end of the crystal from an angle of
high as 30° to one of only about 2°. It should be mentioned that diode
TS achieve efficiencies of 4 to 5% for room temperature operation.
Variation of the amount of A.1 in these lasers permits adjustment of the
ssion wavelength from 0.84 to 0.95 pm.

type Al Ga,. As
p-type GaAs

^ 0 0 © n-type AI^Ga, „ As

00000

©00© ©0000X

.E,

Figure 8-8 Energy level diagram for double heterostructure diode laser.
136 Types of Lasers ri,
Uh
ap. 8

The largest application of diode lasers appears to be in the field of cornm •

cations, particularly fiber optics communication. Because of the great sn
with which diode lasers can be switched on and off and their narrow bandwidth
they are excellent candidates for high data rate fiber optic communicatio '
Diode lasers will also undoubtedly find applications in pattern recognition b
code reading, consumer products such as video disks, computer memory, ant1
some ir illumination applications where fairly high-power monochromatic radia
tion is needed.

8-9 CO2 LASERS

The carbon dioxide (CO2) laser is a molecular laser in which molecular vibrations
rather than electronic transitions provide the mechanism for lasing action.
Other lasers in this category are the carbon monoxide and hydrogen fluoride
lasers. The carbon monoxide laser has achieved high efficiency, around 40%,
but the corrosiveness and toxicity of carbon monoxide have prevented scaling
of carbon monoxide lasers to the power levels achieved in CO2 lasers. Hydrogen
fluoride lasers have been primarily of interest in gas-dynamic and chemical
laser work aimed at extremely high-power levels, megawatts, for periods of
time on the order of a second.
Carbon dioxide lasers have achieved success in a wide variety of industrial-
materials-processing applications at power levels from a few watts to 15 kW
with reasonably high efficiency of 10% for commercial models.3 The types of
materials that the CO2 laser is applied to include paper, wood, glass, plastic,
ceramics, and many metals. Industrial processes include heat treating, welding,
cutting, hole piercing, scribing, and marking.
Because of its wide industrial use, the CO2 laser is discussed in some
detail. Some of the CO2 laser designs that are commercially available are qualita-
tively described. Some specific CO2 laser applications, along with material-pi"0'
cessing applications with other types of lasers, are presented in Chapter 12.

General description. Here we consider the interactions that take

place in the CO2 laser. Pure CO2 can be made to lase but only weakly. Early
in the development of CO2 lasers, Patel discovered that a mixture of C0z»
N2, and He greatly increased the power output. A typical ratio of CO2: ^2 •
partial pressures is 0.8 : 4.5 : 10. This will vary somewhat from system to syste •
The use of 10% oxygen mixed with the N2 improves electrode and optics
without noticeably reducing power output. .
The CO2 molecule is a linear triatomic molecule and, as such, has o
three distinct normal modes of vibration. These modes are shown schematic3

3
This is overall efficiency defined as the ratio of the useful power output to the total v
input, not just discharge power.
 CO2 Lasers 137

N
( \^—( Q \~o o o o o o o o"<rV^ Q jTyVoVdVoVo'ij Q j—*-(
v.
Symmetric

( Q V"^ o o u o^o o o o i[ £ VTnrV<Tdo o o o o ^

Asymmetric

o o o o o o o o o i f c ; oooooooooo^l Q . )

Bending

Figure 8-9 Vibrational modes of the CO2 molecule.

in Fig. 8-9. They are the asymmetrical stretching mode at fa = 7.0 X 1013
Hz, the symmetrical stretching mode at fs = 4.2 X 1013 Hz, and the bending
mode at fa = 2.0 X 1013 Hz. The bending mode is degenerate in that it can
occur in either of two perpendicular planes.
Molecular vibrations, like electronic states, are quantized, meaning that
energy can be added to or removed from these vibration states only in discrete
quanta of energy. The amount by which the energy can change in any given
vibration mode is given by A£ = ±hf, where /is the frequency of the particular
vibration mode. This leads to an energy level scheme for molecular vibrations
similar to electronic energy level diagrams. A simplified version of the CO2
energy level diagram is presented in Fig. 8-10, along with the first excited
state of N2. N2 is a diatomic molecule; therefore it has only one normal mode
of
vibration. The abbreviated spectroscopic notation in Fig. 8-10 for CO2 (1,
fn
> n) refers to the number of quanta in the various vibration modes, 1 quanta
m
the symmetrical mode, m in the bending mode, and n in the asymmetrical
Oc
le. Only those levels actively involved in the excitation and lasing process
are
shown in Fig. 8-10.
Note that N2 has a near coincidence between its first vibration state and
at
of the (001) state of CO2. This excited state of N2 is extremely long lived,
"~ly ensuring that it will give up its energy, received by means of collisions
h e
lectrons, to an unexcited CO2 molecule in a resonant or elastic energy
er
i - Direct excitation of CO2 molecules by collision with electrons is also
p0rtan
Or t- Lasing can occur between the (001) and (100) states at 10.6 /u.m
/ the (001) and (020) states at 9.6 ju,m. The probability for the latter transition
n
'y about one-twentieth that of the 10.6 /xm transition and so it does not
Ur
normally unless the 10.6 ju,m transition is suppressed. From an industrial
138 Types of Lasers

9 X 10' 3 Hz

6 X 1013 Hz -

3 X 10U Hz

Figure 8-10 Energy level diagram for CO2 and N2.

applications standpoint there is no advantage to operating at 9.6 ju,m, which

would result in greatly reduced power output.
The vibrational states of the CO2 molecule are split into many closely
spaced energy states due to the quantized rotational states of the molecule.
Quantum mechanical selection rules require that the rotational state either in-
crease or decrease by only one quantum upon a transition from one vibrational
state to another. That is, if / is the rotation quantum number, A/ = **••
This situation leads to a large number of possible lines with two distinct branches,
one for A/ = 1 centered at about 10.2 fxm and one for A/ = -1 centered at
10.6 jum. Because of the slightly lower energy change for A/ = —1, this branch
is favored, leading to a nominal output wavelength of 10.6 jmm. Usually only
one rotational transition operates because the one with highest gain is the mos
populated. Extremely rapid transitions between rotational levels occur due
thermal effects. The bandwidth of CO2 lasers is about 60 MHz as a result o
collisions and the Doppler effect.
The natural lifetime of the (001) state of CO2 is of the order of seconds,
but collision effects in the gas reduce this lifetime to the order of millisecon ^
which is suitable for laser action. The collision lifetime of the lower states a
on the order of microseconds, but thermal effects repopulate the (010) sta '
creating a bottleneck in the return of CO2 molecules to the ground state. Co ^
sions between CO2 molecules in the (100) state and unexcited state can cau
transfer of energy to the (020) state with subsequent collisions with unexci
 g_9 CO2 Lasers 139

Beam

0.05 kW/m

Axial discharge with slow axial gas flow

Figure 8-11 COz laser with axial discharge and slow axial flow.

C02 molecules, causing transfer to the ground state, by collision with He atoms
(which act as an internal heat sink), with the walls of a discharge tube, with
a heat exchanger, or they are simply expelled from the active region. All CC>2
lasers of more than a few watts CW power output use some form of flowing
gas technique to ensure a fresh supply of gas in the active region at all times.
Many designs producing 1 kW or less simply exhaust the gas into the
atmosphere. The total flow rate is around 0.1 SCMH.4 The gas can be recirculated
through a heat exchanger and is generally for higher-power level machines. A
small amount of gas is still continuously or periodically removed and made
up by fresh gas to keep the gas mixture free of decomposition products, particu-
larly carbon monoxide, which absorbs strongly at 10.6 p,m and will quench
lasing action. Such byproducts are also harmful to the optics and electrodes.

C02 laser types. Four basic types of CC>2 lasers are produced today
for operation at 500 W CW or more. The differences in these types are significant
enough that each is briefly described. Certain special laser designs, such as
the TEA (transversely excited atmospheric pressure) laser, the waveguide laser,
ar
>d other sealed tube types, will not be covered.

*• Axial discharge with slow axial flow. This type of laser is schematically
represented in Fig. 8-11. The power capability of this type is 50 to 70
W/m, although efforts are being made to increase this amount through
better heat transfer from the gas to the cooling liquid that surrounds
the laser tube in a separate tube or jacket. Power outputs up to 1.2 kW
a
re achieved by placing tubes in series optically, with electrical discharge
and gas flow in parallel. That is, the electrical discharge and gas flow
a
re independent for each tube that is 2 or 3 m in length. The tubes are
SCMH stands for standard cubic meters per hour.
140 Types of Lasers h
ap. 8

Figure 8-12 One kilowatt axial flow laser. The laser head is in the large cabinet
in back of the work station. (Courtesy of Saginaw Steering Gear, Division of GMC.)

usually mounted on a granite slab or in an Invar rod frame for mechanical

and thermal stability. Figure 8-12 is a photograph of a 1-kW laser of
this type.
Some of these lasers can be electronically pulsed to achieve much
higher peak power without decreasing the average power significantly.
This factor is advantageous in a number of material-removal operations.
The tube diameter of such lasers is usually quite small, resulting in the
TEMoo mode being strongly favored. Somewhat larger bore designs are
available if low-order multimode operation is desired, such as for heat
treating.
2. Axial discharge with fast axial flow. The chief difference between this
design and the previous one is that the gas mixture is blown throng
the tubes at high speed and recycled through a heat exchanger. Eac
CO2 molecule is in the tube only long enough to lase and then pass
through the heat exchanger. The power output of this type of laser
600 W/m. The tubes may be of fairly large bore; consequently, eit e^
unstable or higher-order mode outputs are common. Gaussian °Pe™ ^
is reported for several kilowatts. Such lasers are available from •>
to 6-kW CW power output with the potential for much higher P°w
levels.
 1. Chamber 5. Channel 4. Blowers
14. Access door \ 2. Wind tunnel

11. Output 8. Ionizer /6. Cathode I 10. Return

mirror
9. Fe dback 7. Anode 13. Optics 12. Aerowindow
mirror bench

(a)

(b)

'Sure 8-13 (a) CO2 electron preionization laser with fast transverse flow, (b) Photo-
graph of!5-kW CO2 laser with electron beam preionization. (Courtesy of Avco Everett
Research Laboratory, Inc.)
142 Types of Lasers ru
ha
^ C>. 5

3. Transverse electron beam preionization with fast transverse gas flo^,

turns out that the electron energy required to sustain a discharge j '
CO2 laser is higher than optimum for production of a population inversi *
One effective way to avoid this problem is to ionize the gas via a brn A
electron beam and then use a lower excitation voltage, which is not hi K
enough to sustain the discharge but which provides optimum electr
energy for excitation of the N2 and CC>2 molecules. Figure 8-13 contai
a schematic and a photograph of such a commercially available laser
The power output of this type of laser reaches 10 kW/m and the
commercial models provide up to 15 kW CW power. Because of the very
high power level, these lasers are only operated with an unstable cavity
In the model shown an annular mirror placed around the back mirror
which is actually the output mirror, deflects the ring output past the side
of the other mirror through an aerowindow. The aero window is a chamber
with a vacuum pump attached to it. Air and laser gas entering the chamber
from opposite ends are pumped out, thereby preventing air from entering
the laser cavity.
The electron beam is produced in a high vacuum by the thermal
emission of electrons from a large planar filament and acceleration through
a high voltage. The primary electrons strike a thin metal foil separating
the high vacuum of the electron gun and the much higher pressure in
the laser cavity. Secondary electrons ejected from the foil produce the
ionization in the laser gas that enables the low voltage discharge to be
maintained. Such lasers have produced over 90 kW.

4. Transverse discharge, transverse fast flow. This type of laser uses a fast
transverse gas flow, about 60 m/s, with recirculation, like the previous
design, but with a conventional electrical discharge transverse to the flow
and the beam. The discharge takes place between a hollow water-cooled
cathode or pins and a water-cooled segmented anode with each segment
of the anode individually ballasted by a resistor. Relatively low voltage
sustains the discharge at quite high current because of the short distance
between the cathode and the anode. A power output of about 600 w/
is achieved. The beam is folded back and forth through the discharge
region five to seven times, resulting in an actual power output of af
2.5 kW for a laser with approximately 1.2 m separating the cavity mirr° _
Figure 8-14 shows a schematic diagram and a photograph of one su
laser. n
The outputs of these lasers can vary from nearly Gaussian to -^
high order or may be unstable, depending on the optical configure
of the cavity. Lasers of this type are available with power outputs
1 to 9 kW and much higher power levels are attainable.
 (b)

Figure 8-14 2.5-kW transverse discharge, transverse fast flow CO2 laser, (a) sche-
matic, (b) photograph. (Courtesy of Spectra-Physics, Industrial Laser Division.)
144 Types of Lasers r-
Chk
ap. |

PROBLEMS

8-1. Discuss neutral gas, ion, and metal vapor lasers with particular emphasis on th •
differences. '
8-2. Discuss dye lasers with particular emphasis on the pumping schemes and the rea
for pulsed operation.
8-3. Discuss and compare ruby, Nd-YAG, and Nd-Glass lasers.
8-4. Discuss the diode laser with comparisons between homojunction and heterojunctj
devices and the methods used to reduce injection current density.
8-5. Discuss CO2 laser operation and describe and compare three types of CO2 laser
designs.
8-6. Construct a table of pertinent data, such as wavelength, power or pulse enerev
range, bandwidth, and type of applications, for six of the lasers discussed in this
chapter.

BIBLIOGRAPHY

8-1. Verdeyen, J. T., Laser Electronics. Englewood Cliffs, NJ: Prentice-Hall, Inc., 1981.
8-2. Svelto, O., Principles of Lasers. New York: Plenum Press, 1982.
8-3. Haus, Hermann A., Waves and Fields in Optoelectronics. Englewood Cliffs, NJ:
Prentice-Hall, Inc., 1984.
8-4. Charschan, S.S. (ed.), Lasers in Industry. New York: Van Nostrand Reinhold Co.,
1972.
8-5. Ready, J. F., Industrial Applications of Lasers. New York: Academic Press, 1978.
8-6. Silvast, W. T., "Metal-Vapor Lasers," Scientific American, February 1973.
8-7. Moss, T. S., G. J. Burrell, and B. Ellis, Semiconductor Opto-Electronics. New York:
John Wiley & Sons, Inc., 1973.
8-8. Duley, W. W., CO2 Lasers: Effects and Applications. New York: Academic Press.
1976.
8-9. Demaria, A. J., "Review of CW High-Power CO2 Lasers," IEEE Proceedings,
vol. 61, No. 6, June 1973.
 chapter
Low-Power Laser
Applications I:
Alignment, Gauging,
and Inspection

Industrial systems that use low-power lasers for alignment, gauging, or inspection
generally consist of four basic units. These units are a laser light source, an
optical system to direct and structure the light, an optical system to collect or
image the light after it has interacted with the object or medium being interro-
gated, and a detection system. The detection system may be a human observer
or an optoelectronic device interfaced with an electronic display unit or com-
puter. This chapter describes the basic principles of low-power laser systems
used in industry. Some comments on using a computer to interpret optical
'^formation are made, but basically this topic is beyond the scope of this book.
The applications of low-power laser systems discussed here do not depend
ft the coherent nature of laser light. Most systems are simple in nature and
an
be designed to operate in an industrial plant atmosphere. Low-power laser
ystems that depend on the coherent nature of laser light are discussed in Chapter
these systems tend to be more complex and often require the protection
of a
Moratory to operate in a reliable fashion.

^ SCANNING TECHNIQUES
USe< to scan aser
are ^ ' beams include mirrors and prisms. These devices
4 generally oscillated sinusoidally or rotated with a constant angular velocity.
tj0 'rror scanner is illustrated in Fig. 9-1. In many laser beam scanner applica-
aCc l"e beam must remain parallel to a fixed axis as it scans. This can be
^Plished by using a converging lens or mirror. A scanner system with a
146 Low-Power Laser Applications I: Alignment, Gauging, and Inspection

Laser

Laser beam

Mirror scanner

Figure 9—1 Mirror scanner.

converging lens is illustrated in Fig. 9-2. The scanner is used to scan the laser
beam across a diameter of the converging lens. If the mirror is placed at the
primary focal point of the lens, the beam will emerge from the lens parallel
to the lens axis. Two element lens are often used to reduce spherical aberration
to a minimum. If the lens is designed to meet the Abbe sin condition, the
beam will emerge with a displacement
y = F sin 0 (9-1)

Mirror scanner

Figure 9-2 Beam scanner system with converging lens and mirror scanner at primer"
focal point of lens.
 Scanning Techniques 14
9-1

is the lens focal length and 6 is the angular displacement of the bear
the lens. 6 is the scan angle. The scan angle 9, in terms of the angular
afitcn
of the mirror, is
sed
6 = 2u>t (9-2)
that the angular speed of the beam is two times the angular speed of
mirror. The transverse speed v of the beam emerging from the lens is
dv
v = — = 2coF cos 2<ot (9-3)
dt
The transverse speed of the beam emerging from the lens is called the scan
speed. For small angular displacements, Eq. (9-1) becomes

(9-4)
and Eq. (9-3) reduces to
v = 2a>F (9-5)
Thus by using small angular displacements and rotating the scanner mirror
with a constant angular speed, a constant scan speed can be obtained.
On emerging from the converging lens, the laser beam is not only parallel
to the lens axis but also converges to a waist near the secondary focal plane
of the lens. In many applications a converging beam is undesirable and a small-
diameter, collimated beam is preferred. It can be attained by using a lens or
mirror system or both, with a common focal point located at the reflecting
surface of the scanner. Figure 9-3 is a sketch of a system that collimates the
beam by using a large-diameter lens to focus the laser beam on a mirror scanner,
*here it is then reflected back through the same lens. In this case, the beam

Laser

Scan plane

iV—
Lens axis
/ J
Y

/
Lens
V
\
\
Laser beam
T

Figure 9-3 Scanner system with converging lens.

148 Low-Power Laser Applications I: Alignment, Gauging, and Inspection
nap. (

emerges from the system collimated, the same diameter as the beam dir
from the laser, and the beam scans a plane parallel to the lens axis. ^
Concave and parabolic mirrors can also be used in scanner system
produce a beam that remains parallel to a fixed axis. Parabolic mirrors alimin °
spherical aberration whereas concave mirrors do not. Figure 9-4 is a sket i!
of a scanner system that uses a section of a parabolic mirror. The scanner'
located at the focal point of the parabolic mirror. The light reflected from th
parabolic mirror has a displacement of

y = 2F tan (- (9-6)

where F is the focal length of the parabolic mirror (see Problem 9-4). Xhe
scan velocity is
v = 2o)Fsec 2 (cor) (9_7\
As in the case of the previous system, Eq. (9-7) reduces to Eq. (9-5) for small
angular displacements. In this system a lens with a focal point located at the
scanner is used to collimate the laser beam.

Parabolic mirror

Laser Lens Scanner

Figure 9-4 Scanner system with a parabolic mirror.

Shadow projection technique. Objects should be backlighte

whenever possible for inspection and gauging. Figure 9-5 is a simplified diagr3
of a system that uses a scanning laser beam to backlight an object so tha
shadow can be used for gauging or inspection. This system uses an oft-3
parabolic mirror to eliminate spherical aberration. A converging lens is u ^
to focus the laser beam on a multisided mirror scanner placed at the from
point of the parabolic mirror. The beam is collimated after reflection
the parabolic mirror. By using an off-axis parabolic mirror, the rotating
can be placed so that it will not interfere with the reflected beam. The detect*
consists of a photodiode placed at the focal point of a second off-axis pa
mirror. As the laser beam scans across an object placed between the sc
 Scanning lecnniques 149
Sec-

Paraboiic mirror Parabolic mirror

Top view

Object
I Lens
Laser Scanner Detector

Side view

Figure 9-5 Simplified diagram of laser system designed for gauging or inspection.

and detector, the beam is blocked by the object for a time interval Af. The
shadow width A_y for small angular displacements, is
Ay = Af (9-8)
where L is the horizontal projection of the laser beam length from the scanner
to the parabolic mirror. The output signal from the photodiode can be processed
to determine A f . This result can be electronically multiplied by 2wL and the
product displayed on a digital readout. Accuracy of this system can be increased
°y focusing the beam to a narrow waist at the object position. Doing so reduces
we time interval required for the cross section of the laser beam to sweep
Past an object edge. Figure 9-6 is a photograph of a commercial system being
Us
ed to measure the diameter of a rod.
By using a computer in a laser scanner system, we are not limited to
akmg measurements that require that the scan velocity be constant. Greater
curacy and measurements that require larger angular displacements can be
^ ained by integrating the scan velocity over the time interval that the laser
us j1 1S Blocked t>v tne object. In the previous example a computer can be
ed
to evaluate the integral

A.V = 2wL sec (<at-—\ dt (9-9)

Th
C time interval fi to ?2 is the time period that the laser beam is blocked by
nc s e scan anan e e
is ti anc
*° '
scan 88^° netheeam
beamat at me
timezero.
zero. The scana angle
e scan
e
horizontal projection of the angle between the laser beam incident on
 150 Low-Power Laser Applications I: Alignment, Gauging, and Inspection Chap.,

Figure 9-6 Laser scanner and detector being used to measure the diameter of a
round bar. (Courtesy of Zygo.)

and reflected from the parabolic mirror. The computer can also be used to
run statistical analyses and store data.

Surface inspection. Surface inspection with low-power lasers is done

either by evaluating the specular or diffuse light reflected from the surface being
interrogated. The laser beam used is almost always scanned so that the surfaces
involved can be inspected in the shortest period of time. Evaluation of specular
reflected laser light to detect surface defects has been highly successful for ma-
chined parts where the machined surfaces are specular and surface defects are
filled with black residue from the machining process. Evaluation of diffuse laser
light to determine differences in surface finish has been successful in some cases.
One example is the determination of whether a machined surface has bee
honed.1 Figure 9-7 illustrates a scanner system that is designed to detect surtac
defects and determine the presence of honed surfaces. If the machined surta
is highly specular in nature, most of the laser light energy is reflected to
lens-photodiode detector positioned to receive the specular reflection. Very ' ^
light energy is diffusely reflected to a second detector positioned off to ° ^
side of the first detector. When the laser beam scanning a machined sul7jn
encounters a flaw, most of the energy is absorbed by the residue ernbedde
J.,£gS
1
Honing refers to the machining of a surface with hone stones or emery cloth, which Pr
very fine periodic scratches.
 Scanning Techniques 151
sec.
9-1

Smooth surface Flaw

Honed surface

Flaw Honed surface

analyzer analyzer

Figure 9-7 Laser scanner system designed to detect defects.

the flaw and specular reflected energy decreases. This drop in energy can be
detected by an oscilloscope or electronics interfaced with the photodiode detector
Positioned to receive specular reflected light. To obtain a significant decrease
ln
specular reflected energy, the spot size of the laser beam on the surface
should be smaller than the smallest dimension of the flaw being interrogated.
hen scanned across a honed surface, the specular reflected light drops slightly
and s
°me of the light diffuses to the second detector. This increase of diffusely
e
ected light indicates the presence of a honed surface.
Scanner systems of this type can also be used to make measurements of
e
Q° diameters, object size, and edge locations and to determine the presence
a
bsence of threads. Figure 9-8 is a diagram of a commercial scanner system
^ 'gned to inspect for surface defects on sheets of material several meters wide.
Ij ,re tne detector is mounted above the sheet to observe diffusely reflected
a • The detector consists of a pair of cylindrical lenses to focus the light on
tk SS or P^stic tube. The back side of the tube is treated so it diffusely reflects
refl '8llt that strikes it. The light entering the front side of the tube is diffusely
e
cted at the back surface and then transmitted, because of internal reflection,
152 Low-Power Laser Applications I: Alignment, Gauging, and Inspection Chap. |

PMT
Receiver

Figure 9-8 Laser scanner inspection system. (Courtesy of Intec Corp.)

down the length of the tube to a photomultiplier mounted on one end of the
tube. A computer is used to process the output signals from the photomultipl16
to indicate surface defects, such as pits, scratches, rust, and dark or roug
spots. Transparent sheets can be inspected by mounting the detector so tha
receives the light transmitted through the sheet.
If the reflectance of the object to be inspected can be controlled, o r '
is highly uniform, the inspection system and evaluation of the information
tained is simple and straightforward. Experience has shown that industrial pa '
even of the same type, can have a wide range of reflectance. Under these co ^
tions signal processing is required to interpret properly the information obtai
from the optical system. The systems discussed here are also very £"e
 g_2 Alignment 153
Sec.

lectors of oil, dirt, ink marks, and other contaminants that do not represent
rface defects. These surfaces must be clean in order to be properly inspected.

9_2 ALIGNMENT

Perhaps the most obvious application of a low-power laser is alignment. In

this application the laser beam may simply be used as a pointer to align drainage
tile, ditches, building foundations, or industrial equipment. By using a quadrant
detector, objects can be aligned relative to a laser beam with high accuracy.
A simplified diagram of an alignment system is illustrated in Fig. 9-9. If the
laser beam is incident on the center of the detector, position indicators register
midscale readings. When the beam is not incident on the center of the detector,
the position indicators register a plus or minus reading, indicating to the user
how to reposition the detector so that the beam is again centered.
If alignment is required over long distances, a beam expander can be
used to reduce beam divergence. By expanding the laser beam, which reduces
the irradiance, there is less potential hazard to the human eye. It is extremely
important, even when working with low-power lasers, that current laser safety
standards be followed and that lasers be used safely.
Laser beams can also be scanned to establish reference planes. This tech-
nique is particularly useful in the construction industry. Laser scanner systems
are routinely used to install walls and ceilings. Figure 9-10 is a diagram of
such a system. A beam expander is first used to increase the diameter of the

Vertical position
indicator

Quadrant detector

Horizontal position
indicator

Laser

Figure 9-9 Simplified diagram of a laser alignment system.

154 Low-Power Laser Applications I: Alignment, Gauging, and Inspection

Penta-prism

Detector
- Beam expander

Laser

Figure 9-10 Diagram of a laser scanner system used for alignment.

laser beam. A penta-prism can be used to turn the beam at a right angle. A
penta-prism turns the beam through a right angle even though the prism orienta-
tion is not precisely controlled. Rotating the prism about the axis of the incident
laser beam enables the output beam from the prism to sweep out a flat plane.
Detectors can be used to align objects relative to this plane. If high-precision
alignment is not required, the spot of light produced when the laser beam strikes
diffusive surfaces can be used directly for alignment.

9-3 TRIANGULATION TECHNIQUE

Optical triangulation provides a noncontact method of determining the displace-

ment of a diffuse surface. Figure 9-11 is a simplified diagram of a laser-based
system that is successfully used in many industrial applications. A low-pow£r
HeNe or diode laser projects a spot of light on a diffusive surface. A portion
of the light is scattered from the surface and is imaged by a converging lens
on a linear diode array or linear position detector. If the diffusive surface ha
a component of displacement parallel to the light incident on it from the lase ,
the spot of light on the surface will have a component of displacement para
and perpendicular to the axis of the detector lens. The component of displaceme
perpendicular to the axis causes a corresponding displacement of the lta^
on a detector. The displacement of the image on the detector can be use
determine the displacement of the diffusive surface. to
Many triangulation systems are built with the detector perpendicula
the axis of the detector lens, as shown in Fig. 9-11. The displacement A
the image on the detector in terms of the displacement of the diffusive su
Az, parallel to the incident laser beam, is approximately
 g_3 Triangulation Technique 155
Sec

Linear diode array

or
position detector

Figure 9-11 Diagram of laser-based optical triangulation system.

Ad = Az m sin 9 (9-10)
where m = i/o is the magnification and 6 is the angle between a line normal
to
the surface and the light scattered to the imaging lens. Equation (9-10)
assumes that Az is small and that 6 remains constant as the surface is displaced.
Thus we can use the approximate Eq. (9-10) to design an optical triangulation
Bauge; we cannot, however, assume that this equation is exact. After the gauge
ls
built, it must be calibrated by displacing a diffusive surface by known distances
a
nd determining the actual image displacement on the detector. A computer
er
> can be interfaced with the detector and programmed to determine the
'splacemeiH °f tne diffusive surface from the displacement of the image on
the
detector.
.. The triangulation system illustrated is not designed to keep the image in
Us
as it is displaced on the detector. To keep the image in focus, the detector
st
be orientated at an angle to the lens axis as shown in Fig. 9-12.
Using Eq. (1-21), the image distance / is found to be
oF
i =- (9-11)
o-F
156 Low-Power Laser Applications I: Alignment, Gauging, and Inspection Chan
P- y

Figure 9-12 Diagram of optical triangulation system with detector at an angle relative
to the lens axis.

where o is the object distance and F is the focal length of the lens. If the
diffusive surface is displaced a distance Az, in the direction of the incident
laser beam, the new object distance o' is
o' = o + Az cos 9 (9-12)
and the new image distance /' is
oF + Fhz cos 9 (9-13)
o - F + Az cos 9
The change in image distance is
-F2 Az cos 9 (9-14)
A; = I' — i =
(o - F)2 + (o - F) Az cos 6
For small displacements of the diffusive surface,
F2 Az cos 9 (9-15)
A/=-
(o-F)2
To determine the angle of orientation <f> of the detector, we will assume that
 g_4 Triangulation Technique 157
sec

j. = A1
tan 4> — (9-16)
A;
where A/ is given by Eq. (9-10). That is,
(Az m sin 0)(o - F)2
tan <> = (9-17)
F2 Az cos 0
F
where m =- (9-18)
o- F
Thus

tan 4> = — tan 0 (9-19)

m
. . Ai
and (9-20)
sm V
/sinfl\
or = m I\ -—r
• cf>/
j.1— Az <9~21)
\sm
Again, we cannot assume that Eq. (9-19) and Eq. (9-21) are exact equations.
They are approximations that can be used to design an optical triangulation
gauge. Then once built, the gauge must be calibrated to give maximum accuracy.
Figure 9-13 is a photograph of an optical triangulation detector with a signal
processor.

Figure 9-13 Laser-based optical triangulation system. (Courtesy of Selcom.)

158 Low-Power Laser Applications I: Alignment, Gauging, and Inspection Chap. 9

9-4 DIODE-ARRAY CAMERA SYSTEM

Many optical inspection devices found in industry have a camera lens to focus
the image of an object on a one- or two-dimensional diode array. White-light
illumination is usually preferred in these applications. Laser illumination is often
used when the light must be highly structured or directional or both. Unless
otherwise stated, we will assume that a laser is being used as a light source
As in other applications, backlighting should be used whenever possible and
reflected light used when necessary.
A typical diode-array camera system using laser light to backlight an
object is illustrated in Fig. 9-14. A beam spreader collimates and increases
the diameter of the laser beam. This laser light backlights an object that is to
be inspected or gauged. A camera lens images the sharp edges of the object
on a diode array. The output of the diode array can be displayed on an oscillo-
scope as illustrated in Fig. 9-14. Those diodes that are not illuminated by the
image produce little or no signal and those diodes that are illuminated produce
positive pulses as the diodes are scanned in sequence. By counting the number
of diodes illuminated, or those not illuminated, or both, along with the magnifica-
tion for the lens system, the dimensions of the object can be determined. If
one-dimensional measurements are sufficient, then a linear diode array can be
used. Two-dimensional measurements can be made by using a two-dimensional
diode array. If a system of this type is used for gauging, then the lens must
be designed so that a linear relationship exists between the object and image.
When using a diode-array camera system, it is convenient to predict mathe-
matically the magnification and the location of the image for a given object
distance. The camera lens must be treated as a thick lens. We have found the
Newtonian form of the lens equations particularly useful for this purpose. The
Newtonian form of the lens equation is

Beam expander Camera lens Diode array

Object

Laser
Q

Analyzer

Figure 9-14 Typical optical gauge or optical inspection system designed by using
a diode array camera.
 g_4 Diode-Array Camera System 159

Camera lens

H H'
Object

Figure 9—15 Notation for Newtonian form of lens equation.

xx' = (9-22)
where x and x' are measured relative to the primary and secondary focal planes
as shown in Fig. 9-15. The focal planes are easily found by passing a collimated
laser beam through the camera lens in both directions and finding where the

Analyzer

Figure 9-16 Optical thread checker.

160 Low-Power Laser Applications I: Alignment, Gauging, and Inspection Chap

beam focuses. The focal length of the lens can be obtained from the specification
on the lens. Magnification using the Newtonian form is
_/_ x' _ F
m ——
y — F- — x v(9-2^\
.oj
The resolution with which a measurement of the image on the diode array
can be made is plus or minus the distance between two adjacent photodiodes
Thus the theoretical resolution with which a measurement of the object can
be determined is plus or minus the distance between adjacent photodiodes divided
by the magnification.
Industrial parts can be inspected by frontlighting the parts if sufficient
contrast is present so that the image can be interpreted by the diode array.
Figure 9-16 shows a simplified diagram of an optical inspection system designed
to detect the presence or absence of threads. If the threads are present, the
output of the linear diode array is a sinusoidal waveform as illustrated. If threads
are not present, the output for each diode tends to be the same and will produce
a dc signal. This system could also be used to determine the pitch of the threads
and to detect stripped threads. Rotating the object about the axis of the hole
allows a complete inspection of the threaded hole to be made.
A photograph of a commercially available two-dimensional, diode-array
camera system is shown in Fig. 9-17. Note that this system does not use a

Figure 9-17 Computer vision system. (Courtesy of Machine Intelligence Corpor

tinnl
tion.)
Chap- 9 Pr°blems 161

r for illumination. Laser illumination is undesirable because of diffraction

fleets and speckle due to the coherent nature of laser light. When light of
high irradiance must be directed from long distances into small, confined areas,
however, a laser is often required.

PROBLEMS

9_1. Using y = F6, y = F sin 6, y — 2 F tan (6/2) and assuming that F = 20 cm,
(a) Plot beam displacement versus angular displacement for 0 < 9 < 77/6 on a
single graph.
(b) Find the beam displacement variation from y = F9 for y = F sin 9 and
y = 2F tan (0/2) at 6 = 3°, 6°, 12°, 20°, and 30°.
9-2. Using v = 2(aF, v = 2o>F cos 6, and v = 2&>F sec2 (6/2) and assuming that
F = 20 cm and w = 20007T rad/s,
(a) Plot beam velocity versus angular displacement for 0 < 6 < 77/6 on a single
graph.
(b) Find the beam velocity variation from v = 2<dF for v = 2wF cos 6 and
v = 2wF sec2 (6/2) at 6 = 3°, 6°, 12°, 20°, and 30°.
9-3. The beam scan angle 6 for an oscillating mirror scanner is given by

6 = 0raax sin 2nft

where /is the oscillating frequency and 6mlai is the angular displacement amplitude
of the mirror.
(a) Derive the equations for the transverse beam displacement and speed as a
function of time for a laser beam emerging from a converging lens after being
reflected by an oscillating mirror placed at the focal point of the lens.
(b) Plot the beam displacement and velocity for 0 < t < 120 /us if F = 20 cm, /
= 2000 Hz, and 0max = 12°.
9-4- The equation for a parabola is

where F is the focal length. Assume that a laser beam originates from the focal
point of a parabolic mirror at an angle 6 with respect to the optical axis. Show
that the laser beam is reflected from the mirror so its displacement y from the
optical axis is

s
- Integrate Eq. (9-9) to show that the displacement A^ of a laser beam for a time
interval tl to f 2 for a system configured as shown in Fig. 9-5 is

~ _ 2L sin o)(f 2 ~ 1 1) _
.y _
COS[Q> 12 - (0o/2)] cos [(at!- (00/2)]
Show that this equation reduces to Eq. (9-8) for small angular displacements.
162 Low-Power Laser Applications I: Alignment, Gauging, and Inspection Cha n
"• 9

9-6. Using Eq. (9-21), show that

A<f = m2 Az
as 6 approaches zero and that
A = m Az
for 9 = ir/2.
9-7. Using Eq. (9-21), derive
I'cosfl r F2 cos 0 ~\ .
Az Az
[(o-^cos^J
9-8. An optical triangulation gauge is configured as shown in Fig. 9-12. The object
distance is 12 cm, the image distance is 15.3 cm, and the angle between the
laser beam and the optical axis of the lens is 40°. The detector is a 256 linear
array with photodiodes placed at 25-p.m centers.
(a) Determine the focal length to the lens.
(b) Determine the magnification.
(c) Determine the angle between the diode array and the optical axis.
(d) Determine the maximum displacement Az that can be measured.
(e) Using the distance between two photodiodes as a least count, determine the
minimum displacement Az that can be measured.
9-9. An optical triangulation gauge is configured as shown in Fig. 9-12 except the
laser beam incident on the diffusive surface makes an angle ft relative to a normal
to the surface. Derive the equations

and Az
sm <j> cos ft
which can be used to determine the detector angle 4> and image displacement
Arf for this gauge.
9-10. A diode-array camera gauge is configured as shown in Fig. 9-14. A 1024 linear
array with photodiodes placed on 25 /urn centers is used as a detector. Using
the Newtonian form of the lens equation, determine the following for x = 62.5
mm and x' = 48.4 mm.
(a) The focal length of the lens.
(b) The magnification.
(c) The number of diodes that would not be illuminated if the gauge is used to
measure the diameter of a quarter-inch rod.
(d) Determine the theoretical resolution of the gauge. ,

BIBLIOGRAPHY
:*-
9-1. Jenkins, F. A., and H. E. White, Fundamentals of Optics. New York: McGra'
Hill Book Co., 1976.
9-2. Sears, F. W., M. W. Zemansky, and H. D. Young, University Physics.
MA: Addison-Wesley, 1976.
 chapter
Low-Power Laser
Applications II:
Interferometry
and Holography

The purpose of this chapter is to provide the reader with a basic understanding
of the laser-based optical systems in industry that depend on the coherent nature
of light. An overview of some standard mathematical descriptions of laser light
is presented first and then these descriptions are used to discuss applications
of interferometry and holography.

10-1 MATHEMATICAL DESCRIPTIONS OF LASER LIGHT

One solution to the wave equation, Eq. (1-1), is the equation of a plane wave
Propagating along the z axis
E = iE0 exp [J(kz - of + <f>0)] (10-1)
where £ is the electric field intensity, E0 the electric field intensity amplitude,
=
27T/X. the propagation constant, o> = 2irf the angular frequency, j the
Complex number \/—l, and <f>0 the original phase angle.1 The wavelength of
8ht is X. and / is its frequency. A unit vector i can be used in the equation
. en the light is plane polarized. To find the irradiance / in cases where E
he vector sum of two plane-polarized laser beams, the equation
<E • E*>
/=- (10-2)
2Z
' Boldface type denotes vector quantities and boldface italic type denotes complex vector
164 Low-Power Laser Applications II: Interferometry and Holography Chap

can be used. E* is the complex conjugate of the electric field intensity. 7

angular brackets < > denote the time average

dt

T is the response time of the equipment that measures the ifradiance or th

exposure time for the film that records an image. In applications where th
two laser beams are coherent the time-dependent term in Eq. (10-1) is not
required to describe the light. To separate the time-dependent term, Eq
(10-1) can be written
E = \E0 exp (jkz) exp O'(f>0) exp (—;o>0
Now it is possible to express Eq. (10-4) as

where U is the complex amplitude vector and is

U — \E0 exp (jkz) exp (jtyo) (10-6)
This complex amplitude vector is then used to describe a polarized plane wave
because the time-dependent term is not required. It can be shown that for
coherent linearly-polarized laser beams Eq. (10-2) reduces to

'-
when T » 1/f.
Many other solutions to the wave equation are possible. Figures 10-1
and 10-2 are a summary of some possible solutions in rectangular and radial
coordinates. Each solution can be expressed as a product of an amplitude,
obliquity factor, original phase factor, and propagation factor. The amplitude
term is constant in plane wave descriptions and a unit vector is used if the
wave is linearly polarized. The propagation vector k has a magnitude or
2n/X and is in the direction of propagation of the plane wave. To describe a
plane wave propagating in a direction other than +z directions, cosines It
and yy are used in the rectangular coordinate description and a position vecto
r is used in the radial coordinate description. The amplitude varies inverse y
proportional to r in mathematical descriptions of spherical waves. An obliqui y
factor is often used to describe the fact that laser light radiating from a p"in
usually does not do so uniformly in all directions. The term cos ij> is useo
Figs. 10-1 and 10-2, but it should be realized that it is not the only P°ssi
obliquity factor.
These mathematical descriptions should not be considered complete,
are intended to familiarize the reader with some descriptions that describe la
beams mathematically. To derive these equations is beyond the scope or
book. (For more complete presentations of this material, refer to Gaskill
and Goodman 1968.)
 165
Sec 10-1 Mathematical Descriptions of Laser Light

Factor Equation

Amplitude iE0

Original phase exp (/#„)

Propagation exp (jkz)

U = if0 exp (jkz) exp (/0 0 )

(a) Polarized plane wave propagating in +z direction.

Factor Equation

Amplitude Pfo
Obliquity 1
Original phase exp (/0 0 )
Propagation exp (/k • r)

U = p£0 exp (/k • r) exp (/0 0 )

(b) Polarized plane wave propagating in +k direction.

Factor Equation

A_
Amplitude
r
Obliquity cos (i//)
Original phase exp (/00)
Propagation exp (±/*r)

U - — cos (i//) exp (±jkr) exp (/0 0 )

(c) Spherical wave diverging (positive exponent) or converging (negative exponent)

with respect to origin.

Figure 10-1 Complex amplitude descriptions in rectangular coordinates.

 166 Low-Power Laser Applications II: Interferometry and Holography Chap
10

Factor Equation

Amplitude if0

/
v\ * z
Obliquity
Original phase
Propagation
1
exp (/00)
exp (jkz)

U = \£0 exp (/0 0 ) exp (jkz)

(a) Polarized plane wave propagating in +z direction.

Factor Equation

Amplitude p£"0
Obliquity ' 1
Original phase exp (/<J0)
Propagation exp (jkz0) exp // jfy)

U = pf0 exp (/00) exp (jkz0) exp jk(ix x + yyy)

(b) Polarized plane wave propagating in +k direction.

Factor Equation

A
Amplitude

*- x Obliquity

Original phase exp

Propagation exp

±z0A{exp (±]kz0) exp (/00) exp [±jk(x'2

(zl + x2 + y2}

(c) Spherical wave diverging (positive exponent) or converging (negative exponent)

with respect to the origin.

Figure 10-2 Complex amplitude descriptions in terms of radial coordinates.

 .jO-2 Laser Velocimetry 167
Sec-

_2 LASER VELOCIMETRY
10:

nsider two coherent plane waves with the same amplitude and polarization
hat intersect at an angle of vj/ as shown in Fig. 10-3. In the region of intersection
he interference fringes are bright and dark planes of light that are parallel to
the plane bisecting the angle between the two beams. To derive the expression
f r the irradiance along the x axis, we will use the equation for a polarized
lane wave propagating in the +k direction from Fig. 10-2. By assuming that
__ i = y = 0, this equation can be expressed as
U = ]E0 exp (jkx cos 0) (10-8)
Thus the complex amplitude vectors of the two waves can be described by

U, = j£0 exp Ijkx sin (|) J (10-9)

and t/ 2 = j£o exp | -jkx sin 11 ] I (10-10)

The resultant complex amplitude due to the interference of the two plane waves
is
U = }E0 I exp jkx sin (^ J + exp —jkx sin (-J \ (10-11)

Using Eq. (10-7), the irradiance / is given by

i = ^2 COS 21 kx sin /| JI (10-12)

Figure 10-3 Intersecting coherent plane waves.

168 Low-Power Laser Applications II: Interferometry and Holography Chap

The distance d between two adjacent bright fringes can be found by letting

kd S m
• M
(2J = 7r
Thus

The parallel fringes produced by two intersecting coherent laser beams provide
a nonintrusive technique, called laser velocimetry, to determine the velocity of
small particles moving in a transparent gas or liquid. One common technique
is to seed a gas passing through a wind tunnel or engine with particles that
will move with the same velocity as the gas. As the particles pass through the
fringes, produced by the two intersecting beams, they scatter light proportional
to the irradiance of the fringes. A sketch of a typical system is shown in Fig.
10-4. In this laser velocimetry system a laser beam is first split into two beams
by a beam splitter. These beams pass through two holes in a mirror. Then
they pass through a converging lens that focuses the beams to its focal point
where they intersect, producing interference fringes. As a particle passes through
the fringes, it scatters light back through the converging lens. A portion of
the scattered light is reflected to a second converging lens that focuses this
light on a photomultiplier. The signal from the photomultiplier is periodic and
has a frequency
f=^ (10-15)

where vx is the x component of the particle velocity that is perpendicular to

the interference fringes. Other components of the velocity can be found by

Particle
Mirror Lens

Photomultiplier

Analyzer

Figure 10-4 Laser velocimetry system.

Sec
10-3 Laser Interferometry 169

forming fringes that are orthogonal to the first set. Signal separation is accom-
plished by using laser light of different colors or cross polarizing the light produc-
ing the fringe sets.

LASER INTERFEROMETRY

Industrial applications of laser interferometers include precise measurements

of displacement, velocity, and surface variation between reflective reference and
test surfaces. Long coherence length of the laser allows the reflective surfaces
to be separated by large distances if necessary. In other techniques that do
not use a laser, the test and reference surface must be in intimate contact.
Linear displacement measurements can easily be made with an accuracy of
one-quarter wavelength of the light being used. By using good measurement
techniques and a computer system to compensate for changes in the index of
refraction of air due to atmospheric effects, measurements can be made with
even greater accuracy. As in the section on laser velocimetry, specific applications
are discussed and mathematical descriptions are derived.

Michelson interferometer. The Michelson interferometer illustrated

in Fig. 1-22 can be used to make precise linear displacement measurements
of one mirror relative to the other or to detect and/or measure surface variations
between the two mirrors. We will first consider a Michelson interferometer
system, as shown in Fig. 10-5, that uses a diverging lens to produce a diverging
spherical wave. The beam splitter splits the diverging beam into two beams of
equal irradiance. The two beams are reflected by the corresponding mirrors
and equal portions are recombined by the beam splitter. These recombined

Mirror

Laser with
beam expandei Mirror

^Screen

Figure 10-5 Michelson interferometer with diverging spherical waves.

170 Low-Power Laser Applications II: Interferometry and Holography Chap
10

waves interfere, producing fringes on a frosted glass screen that can be observed
with the human eye.
The diverging wave can be described by the spherical diverging wav
equation in rectangular coordinates from Fig. 10-2. Beam 1 can be described
by the complex amplitude
jkr
= A exp (jkzi) exp (10-16)

and beam 2 by

U2 = A exp (jkz2) exp (10-17)

We will assume that the amplitudes of the two beams at the screen are equal.
Radii of the two beams are Zi and z2, respectively, and r is the radius of the
pattern on the screen and is equal to \/x2 + y2. Figure 10-6 is a better illustration

Figure 10-6 Superposition of two spherical waves with different radii propagating
in the same direction.

of the two recombined beams. The two beams have different radii because of
the different distances through which they travel. Resultant irradiance of the
two interfering beams is

(10-18)

where Az is the difference between the two radii z\ and z2. A photograph or
the interference pattern produced is shown in Fig. 10-7. By observing the center
of the pattern at r = 0, we can reduce Eq. (10-16) to
(10-19)

Bright fringes are observed at r = 0 if

A: Az (10-20)
= rmr

or Az = mX (10-21)
 10-3 Laser Interferometry 171
Sec.

Figure 10-7 Photograph of the interference pattern due to two spherical waves
with different radii propagating in the same direction.

where m = 0,1,2,3,4, . . . . Dark fringes are observed at r = 0 if

k Az
- ITT (10-22)

or Az -(„+{)' (10-23)

Displacing one mirror of the interferometer a quarter wavelength in the direction

of propagation of the light changes the optical path length of that light by a
half wavelength. Thus at r = 0 the fringe would change from bright to dark
or vice versa due to this displacement. That is, the least count for the displacement
01
a mirror in a Michelson interferometer is one-quarter wavelength of the
light being used.
Variation between the two mirror surfaces can also be determined by
using a Michelson interferometer. To check surface variation, the diverging
en
s is removed and a plane wave is used. If both mirrors are optically flat
no one mirror is rotated through a small angle, straight, equally spaced fringes
j-sult as shown in Fig. 10-8. Surface variations are indicated by variation in
e
straightness and spacing of these lines. Figure 10-9 is a photograph showing
variation of these fringes due to a small bump on one mirror of a Michelson
lnte
rferometer.
So that the observer can see the surface being tested as well as the interfer-
fringes, an imaging lens is often added to a Michelson interferometer as
^ °Wn in Fig. 10-10(a). Using this technique, the image of the surface tested
the interference fringes are superimposed on the screen. Interferometers
I ftl
s type are called Twyman-Green laser interferometers and are often de-
172 Low-Power Laser Applications II: Interferometry and Holography Chan

Figure 10-8 Photograph of interference fringes due to the superposition of two

plane waves propagating at an angle to \each other.
\
signed to test curved surfaces as well as flat surfaces as illustrated in Fig.

Fizeau interferometer. A Fizeau interferometer is illustrated in Fig.

10-11. In the next development we will use an optical flat for the reference
surface and a convex mirror for the test surface. The reflection from the reference
surface is a plane wave and the reflection from the test surface is a diverging
spherical wave. The diverging spherical wave can be described by

= E0 exp exp (10-24)

Figure 10-9 Interference fringes due to a small bump on one of the mirrors on
Michelson interferometer.
A
 Laser Interferometry 173
Sec.

Mirror

Screen

System System
for testing for testing
concave mirrors convex mirrors
(b)

Figure 10-10 Schematic of a Twyman-Green laser interferometer.

e
obliquity factor is assumed to be one, the amplitude E0, the original phase
zero, and zi is the radius of the spherical wave. The reflected plane wave
°m ^6 reference surface can be described by
U2 = E0 exp C/<Jx>) exp (Jkz2) (10-25)
ere
atl
is the distance from the focal point of the convex mirror where the
e wave has an original phase of <J>0- Solving for the irradiance by using
q
' (10-7), we have
174 Low-Power Laser Applications II: Interferometry and Holography Chap
10

El
! = -«*

where r is the radius in the xy plane. We can find the radius of curvature of
the test surface by imaging the convex surface on the screen with a magnification
of one. First, we will adjust the interferometer so that the center spot in the
resulting pattern is bright. We will then assume that (k Az — <f>0)/2 is m^
where m0 is the order of the central bright fringe. Bright fringes will occur
when

n =; (10-27)

where n = m — m0 = 0, 1, 2, 3, 4 The focal length of the spherical

surface is zt and thus its radius of curvature R is equal to 2zi. Therefore

\ *=77 0<«8)

Reference
surface Test mirror
Beam
splitter

Screen

Figure 10-11 Schematic of a Fizeau laser interferometer.

As in the case of the Twyman-Green interferometer, the Fizeau interferom-

eter can be designed to test flat surfaces as well as curved surfaces. Two-dimen-
sional diode arrays interfaced with computers can interpret the fringes produced
and quantify the results. A photograph of commercially available Fizeau interfer-
ometer is shown in Fig. 10-12.

Two-frequency interferometry. Another technique for making lin'

ear displacement measurements with an interferometer uses a laser that operate
at two slightly different optical frequencies. A simplified schematic of such
system is illustrated in Fig. 10-13. When neon, the lasant, of a HeNe laser
exposed to a uniform magnetic field, Zeeman splitting takes place and the
 10-3 Laser Interferometry 175
Sec.

Figure 10-12 Fizeau laser interferometer system. (Courtesy of Zygo.)

Stationary
reflector

Moving
Beam reflector
X spli ter /
/ 1 f
4

7 _ A \ , <\ \
:=:r f \\l M i l l 1 /^ 1 1 I I 1 1 I
Laser with > n
v LI \>
e
arn expander — 1 1 1
—

\
- \
~= ^ Polarizing beam
splitter

Differential
amplifier

Figure 10-13 Simplified diagram of an interferometer system designed to measure

''near disnlaremrnt
176 Low-Power Laser Applications II: Interferometry and Holography
10

emits a laser beam of two frequencies, f\ and /2, with opposite circular polari
tions. In the system illustrated, a quarter-wave plate is used to convert th
oppositely circularly polarized signals into two orthogonal linearly-polar^ H
signals. A beam splitter is used to obtain a small fraction of both linearly polarized
signals, which are combined to obtain a reference signal. After being split fi-0
the primary beam, the orthogonal polarized reference signals are given the sam
polarization by a linear polarizer. By letting z\ = z 2 = 0, the reference signals
can be represented by
EI = hE0 exp (—y'coif) (10-29)
and E2 = AE0 exp (-ycu21) (10-30)
where and w2 are the angular frequencies of the two signals. When the
two reference signals are combined with the same polarization, the two signals
beat. The frequency of the resulting signal is (/i + fz)/2. The response time
of light-measuring equipment is not sufficient to measure this high frequency.
The resulting variation in amplitude has a frequency of (/2 — /i)/2 and can
be detected. The resulting variation in irradiance that can be detected, due to
the two signals, is found via Eq. (10-2). By using a high-speed detector, the
time-dependent term can be considered constant over the response time T and
the resulting irradiance is
2/<"2
COS (o>2 — (10-31)

The frequency difference between the two signals is approximately 1.8 MHz.
Irradiance has a maximum value when
I _
(10-32)
2
where mK = 0,1,2,3,4, . . . , or when

The reference signal is incident on a photodetector that produces an electrica

signal that is proportional to the irradiance of the reference light. This electrica
signal is amplified and used to drive a counter.
The primary beam passes through the reference beam splitter to a polarizing
beam splitter. As shown in Fig. 10-13, the polarized light with frequency /J
is transmitted through the polarizing beam splitter and the linearly-polarize
light with frequency /2 is reflected. The transmitted light is reflected back
the polarizing beam splitter by a cube corner reflector attached to an obj
that is experiencing linear displacement. As discussed in Section 1-9, ligh ^
fleeted from a moving reflector, which has a speed v much less than the spe
of light c, is Doppler shifted by
 10-3 Laser Interferometry 177

The plus s'&n l& use(^ wnen ^e moving reflector moves toward the source of
,• ht and *ne ne§ative sign when it moves away.
Light with frequency /2, reflected by the polarizing beam splitter, is re-
flected back to the beam splitter by a second stationary cube corner reflector.
This signal and Doppler-shifted signal are recombined at the polarizing beam
nlitter and then given the same polarization by a linear polarizer. After passing
through the polarizer, and setting z3 = z 4 = 0, the two signals can be represented
by
E3 = nE 0 exp[-/(a>i±A<o)f] (10-35)
£
and 4 = n£0 exp (— ya> 2 0 (10-36)
Xhe resulting irradiance I due to the two superimposed signals is

/ = cos2 - (o>2 — &)j ± Act))? (10-37)

Now the irradiance can be expressed in terms of the optical frequencies /i

and /a-

/ = cos2

Maximum irradiance occurs when

«D=(/2-/i)/±—/ ' 00-39)

c
where mD = 0, 1, 2, 3, 4, . . . . As in the case with the reference light, the
Doppler-shifted light is incident on a photodetector that produces an electrical
signal proportional to the irradiance detected. This signal is amplified and a
counter is used to make a count.
The counts from the reference and Doppler-shifted signal counters are
subtracted, giving

t (10-40)
c
f
counts are made over short time intervals, Eq. (10-40) becomes

A m = ± — Ar (10-41)
c
If
we use/^c = 1A and v Ar = M, Eq. (10-41) becomes

- (10-42)
e
total displacement
acemen iss obtained
o a n e by making the summation

(10-43)
178 Low-Power Laser Applications II: Interferometry and Holography Chap

Figure 10-14 Laser measurement system. (Courtesy of Hewlett-Packard.)

A commercial laser measurement system that uses a two-frequency laser

is shown in Fig. 10-14. Interferometers of this type can be used to measure
linear displacement, velocity, flatness, pitch, yaw, straightness, squareness, paral-
lelism, and perpendicularity.

10-4 HOLOGRAPHIC INTERFEROMETRY

Dennis Gabor published the first papers on holography in 1948. From 1948
through 1960 progress in the use and development of holography was slow
because researchers working in this discipline did not have a light source aval a
that could produce intense light with a long coherence length. Emmett Lei ^
and Juris Upatnieks, working at the University of Michigan Institute of Scie ^
and Technology in the early 1960s, produced the first holograms using a la s ^
In doing so, Leith and Upatnieks transformed holography from an °bsc
concept to a viable scientific and engineering tool. In 1965 Powell and
published the first paper on holographic interferometry. Dennis Gabor
the Nobel Prize in physics for his work in 1972.
For our purposes, a hologram can be thought of as an optical
produced by using photographic techniques and laser light—that is capa
 10-4 Holographic Interferometry 179
Sec.

creating three-dimensional images. The word hologram stems from the Greek
t frolos, which means whole, complete, or entire, and gram, which means
sage- Thus a hologram is a complete record of an optical scene. In conven-
onal photography the light reflected from a scene is focused—using a converging
1 ns system—onto a photographic emulsion. Variation of the irradiance due
to the image being focused on the emulsion is related only to the electric field
intensity amplitude of the light. In holography a photographic emulsion, usually
on a glass plate, is exposed to an interference pattern produced by two coherent
laser beams. One beam, called the object beam, is reflected from an object or
scene to the photographic emulsion. The other beam, called the reference beam,
is reflected directly from the laser to the photographic emulsion by using mirrors.
Using this technique, both the amplitude and the phase information about the
electric field due to light reflected from the scene can be recorded. After photo-
graphic development, the resulting transparency is a hologram. By shining laser
light—in some cases, white light—through the hologram, three-dimensional im-
ages can be produced.
Recording process. Figure 10-15 is a sketch of an optical system
that can be used to produce holograms. All the optical elements used in the
production of optical holograms are mounted on an essentially vibration-free
surface. Usually a heavy table with a steel or granite top, isolated from floor
vibration by air tubes, provides this surface. A photograph of a typical hologra-

Emulsion
Photographic plate

Mirror

Laser beam

Reference beam
Beam splitter

Figure 1(>-15 Basic optical system for producing holograms.

180 Low-Power Laser Applications II: Interferometry and Holography Chap

Figure 10-16 Typical holography system.

phy system is shown in Fig. 10-16. A continuous or pulsed laser operating in

the TEMoo mode at a wavelength within the visible portion of the electromagnetic
spectrum is used for most industrial applications. Helium-neon, argon, and kryp-
ton are continuous-wave lasers commonly used for holography. Pulsed ruby
lasers are used to produce holograms of transient phenomena or when vibration
is a problem.
To produce a hologram, a laser beam is first split into two beams by a
beam splitter as shown in Fig. 10-15. The reference beam is reflected by a
mirror and then spread by a converging lens to illuminate uniformly a photo-
graphic emulsion on a glass plate. The object beam is spread by a converging
lens and then reflected to the object by a mirror so that the object is uniformly
illuminated as seen from the position of the photographic plate. Pinholes are
used in the focal planes of the converging lenses as spatial filters to remove
optical noise from the laser beams. The optical noise is primarily due to light
scattered by dust and flaws on the optical element surfaces. Maximum coherence
of the two beams at the photographic plate is obtained by making the object
and reference beams of equal length.
It is desirable that objects used in holography have diffusive surfaces.
Flat white paint is often used to give industrial parts uniform diffusive surface5'
When illuminated, each point on these surfaces will act as point sources, scatte
ing light in all possible directions. Thus optical information about every P01
on the object facing the photographic plate can be recorded on an infinite nurn
of points on the photographic emulsion.
Consider the complex amplitude of the laser light scattered from a P°|
on the object to a point on the photographic emulsion to be U0 and the comP
amplitude due to the reference beam at the same point on the emulsion to
UR, as illustrated in Fig. 10-17. The resultant complex amplitude of the "%
at the point on the emulsion is the vector sum
. U=UR + U0
 10_4 Holographic Interferometry 181
Sec.

Object

Photographic
plate

Figure 10-17 Recording configuration. Emulsion

Using Eq. (10-7), the irradiance is found to be

1 = -^ (UR ' V* + U0 • U* + U* • U0 + UR • U*) (10-45)

This result can be rewritten

(10-46)

where the first two terms are the irradiances of the reference and object beams,
respectively. The third term depends on the relative phase and polarization of
the beams. If the object and reference beams are cross polarized

/ = IR + I0 (10-47)
a
'l phase information is lost and a hologram cannot be produced. Ideally, then,
wth beams should be linearly polarized in the same direction. We will assume
hat such is the case for the remaining discussion on holographic interferometry.
Exposure g is the product of the irradiance / and the exposure time T.

g= IT (10-48)
er
Development, a typical emulsion used in holography has a light amplitude
ismittance versus exposure before development as shown in Fig. 10-18. To
in t"1 a ^near recording, the emulsion is "biased" to establish a quiescent point
• 6 ear portion of the transmittance versus exposure curve. By making a
^^ recording, it can be assumed that the amplitude transmittance of the
S1
°n after development is a linear function of the irradiance incident on
eir
>ulsion during exposure. That is,

,
t=t0 r- (10-49)
2Z 2Z
182 Low-Power Laser Applications II: Interferometry and Holography

1.0

0.8

I
I 0.4

0.2

Exposure

Figure 10-18 Light-amplitude transmittance versus exposure for typical holographic

emulsion.

where t0 is the "bias" transmittance established by the reference beam and /3

is the product of the slope of the transmittance versus exposure curve at the
quiescent point and the exposure time, [Goodman 1968].

Reconstruction. Holograms produced by using the optical system

diagrammed in Fig. 10-15 can be used to form virtual and real images. To
"play back" a virtual image, a reconstruction beam is used that is identical to
the reference beam used in the production of the hologram. In most industrial
applications a laser is used for reconstruction. When a reconstruction beam,
which is identical to the reference beam, is used to illuminate the hologram,
much of the light passes straight through the hologram. Some light is diffracted
by the fringes recorded in the emulsion due to its exposure to the interference
pattern produced by the object and reference beams.
An observer viewing the diffracted light, as illustrated in Fig. 10-19, wi
see a virtual image of the object in space behind the hologram. This image
three dimensional and in the same location as the object was relative to
looking
emulsion.
v^iiiw-ioivyn. The image will
i ii»». lAiicig,^ appear t^J
ry lii tiLjLs^tii to '-'v*
be i.iiv>
the JM.HJ.X*
same as M.IJ if
1.1 the
v**^ observer
\-T\JLJ***. • —••• were
•- .

at the original object, illuminated by the light used for reconstruction, throug
a window the size of the hologram. j
Assume that a hologram, produced in the manner discussed, is illumma ^
by a reconstruction beam UP. The light transmitted through the hologra

(10-50)
UPt = /3 UpI0
2Z 2Z
It is traditional to write Eq. (10-50) as
Upt = £/! + C/2 U3
 10-4 Holographic Interferometry 183
Sec.

Diffracted light

' Emulsion

Observer

Virtual image

Hologram '

Figure 10-19 Holographic reconstruction of a virtual image.

where t/i = UPt0 (10-52)

C/2 = /8 UpI0 (10-53)

(10-54)
2Z

and (10-55)
2Z
To reconstruct the virtual image,
Up = bUR (10-56)
That is, the reconstruction beam is identical to the reference beam except for
a multiplicative constant b. In this case, it can be shown that the light transmitted
through the hologram distributes itself as indicated in Fig. 10-20.
Note that
(10-57)

U, = bU

°~20 Virtual image reconstruc-

Emulsion Hologram Observer
184 Low-Power Laser Applications II: Interferometry and Holography Cha
P- 10

Thus the light represented by this term is an exact duplicate of the obi
beam except for the multiplicative constant bf$IR. The light represented K
this term is identical to the light reflected from the object to the emulsj
when the hologram was produced!
To reconstruct a real image using holography, it is common practice
collimate the reference beam. This being the case, the reconstruction system
in Fig. 10-21 can be used to reconstruct a real image of the object. In this
playback process, a collimated reconstruction beam is used to illuminate the
hologram from the opposite direction that the reference beam had when the
hologram was produced. This image appears in front of the hologram and is
formed in the same location, relative to the emulsion, as the location of the
object when the hologram was produced.
To reconstruct the real image,
UP = bU*R (10-58)
The reconstruction beam is the complex conjugate of the reference beam multi-
plied by the constant b. For a collimated reference beam, the complex conjugate
of UR represents a collimated beam traveling in the opposite direction of the
original reference beam. (See Problem 10^8.) In this case, the light transmitted
through the hologram distributes itself as indicated in Fig. 10-22. Using this
playback scheme,
Ut = (b/3IR)U* (10-59)
Light represented by this term is an exact duplicate of the object beam except
for the multiplicative constant 6)3 IR and the fact that the light is conjugated.
This conjugated light produces a real image of the object.

Emulsion

Collimated Diffracted
reconstruction light
beam

Observer

rtinn nf a real imaee.

 10-4 Holographic Interferometry 185
Sec.

Emulsion Hologram

Observer

U. + U.

Figure 10-22 Real-image reconstruction.

Real-time holographic interferometry. Figure 10-15 illustrates a

holography system that is frequently used for all types of holographic interfer-
ometry. In real-time holographic interferometry the photographic plate is ex-
posed, developed, and then placed back in the same position it occupied during
exposure. By replacing the developed plate (hologram) and turning the laser
back on, the object beam then illuminates the object, and the reference beam
becomes the reconstruction beam for the hologram. The object is then viewed
through the hologram. The observer will see the object and the virtual image
reconstructed by the hologram. If all the elements in the holography system
have maintained the same position, the virtual image and the object will be
located in the same position. Now if the object experiences a small surface
deformation, fringes will be observed due to the interference of the light reflected
by the object and the light diffracted by the hologram that forms the virtual
lm
age. These fringes are similar to the lines on a topographical map that are
su
perimposed on the object. These fringes can be used to determine surface
displacement.
Surface displacement is most easily determined when a collimated object
ea
m is used and the object beam, surface displacement, and light scattered
0
the hologram are all directed along an axis that is normal to the object
SUf
face as illustrated in Fig. 10-23. The object illustrated in this figure is a
c
ular diaphragm mounted in a rigid frame. The diaphragm is flat when the
to
graphic plate is exposed and then deformed so it forms a convex surface
'Ve to ^e °bject beam after the hologram is returned to the holography
"tern.
Assume that the original object beam is
U0 = a(x, y) (10-60)
e at a
to
to u *^ hologram of the object has been produced and it has been returned
th e holography system so that it can be reconstructed by the reference beam.
186 Low-Power Laser Applications II: Interferometry and Holography Chap

Object beam
Hologram

Observer

Reconstruction
beam

Figure 10-23 Simple holographic interferometry system with object beam normal
to object surface. •*

When reconstructed, the light producing the virtual image of the object can
be expressed as
C/3 = — a(x, y) (10-61)
where the multiplicative constant 6/3IR is assumed to be — 1. The negative
sign is required for most real-time holographic interferometry because the photo-
graphic plates usually used in this work have negative emulsions, making ft
negative. In this reconstruction process, the light reflected from the object and
viewed through the hologram can be expressed as
U'0 = a(x, y) exp [j A<J>(*, y)]
A<|>(x, y) is the phase shift due to the displacement Az. Irradiance due to the
interfering beams is

/<*,)=
*,y) .,
— an" L-
(10-63)

Note that when A<|>(x, y) is zero, the image and object

^ are in the same position
and the irradiance is zero. Change in object beam path length is two tirne
the surface displacement for the configuration being discussed. Hence

vfl (10-64)

Dark fringes correspond to phase changes A<J>(;c, y) = 2-nm, where m ' '
2, 3, 4, . . . . If a section of the object surface experiences no displaced >
dark fringes can be assigned numbers m = 0, 1, 2, 3, where zero
Sec.
10-4 Holographic Interferometry 187

Stationary
region

m=2 m=0

Figure 10-24 Real-time holographic im-

age due to normal displacement of a circu-
lar diaphragm.

assigned to the stationary surface. By counting the number of dark fringes

from the stationary surface, the displacement along the z axis can be calculated
using , ,

Az(x, y) = m- (10-65)

An illustration of the fringes observed using the real-time holography

technique discussed is shown in Fig. 10-24. Recall that the outside perimeter
of the diaphragm was held stationary and that the diaphragm cups away from
the observer as it is displaced. Using Eq. (10-65), it is obvious that the center
of the diaphragm has been displaced 3X/2.
Unfortunately, not all surface displacements can be determined via this
simple technique. Complex object shapes require a vector analysis to determine
surface displacement. Techniques for calculating more complex displacements
are discussed in the next section. It should be realized, however, that the more
generalized techniques discussed in the next section can also be applied to real-
time holographic interferometry.
Example Problem

rind the displacement of the diaphragm center, illustrated in Fig. 10-24, if the blue
lr
-e of an argon ion laser is used.
°lution. We have assumed that the displacement is in the +z direction. Counting the
rlc
fringes starting with zero at the stationary surface, we determine that m is 3.
, e blue line from an argon laser has a wavelength of 488 nm. Using Eq. (10-65), we
have

(3)(488 nm)
Az(0, 0) =
Thu
Az(0, 0) = 732 nm
188 Low-Power Laser Applications II: Interferometry and Holography Cha
P. 10

Double-exposure holographic interferometry. The holograph

system illustrated in Fig. 10-15 can also be used for double-exposure holograph'
interferometry. Here the emulsion is exposed with the object in its original
state, the object is deformed, and then a second exposure is made on the sarn
emulsion. The emulsion is developed after the second exposure. The resulting
hologram can be reconstructed by using the reference beam or its duplicate
Two images will be reconstructed. One image is of the object before deformation
and the second is of the object after deformation. As in the case of real-time
holographic interferometry, fringes will be observed that can be used to determine
the surface displacement that took place between the two exposures.
Fringes observed in double-exposure holographic interferometry are due
to the interference of the light that produces the holographic images. Figure
10-25 can be used to determine the equations needed to determine surface
displacement. Point O is the location of the converging lens focal point in the
object beam. Point P is the original position of a point on the object surface.
Its final position after surface deformation is P' and d is the surface displacement.
Images reconstructed by the hologram are observed through point Q. Because
surface displacement must be small, it can be assumed that the rays represented
by T! and ri are parallel. The same is true for riband r2. ka and k2 are propagation
vectors for the light illuminating P and light scattered to Q, respectively. Phases
of the rays observed in reconstruction are
+ k 2 • r2 (10-66)

and <f>2 = + k2 • r2 (10-67)

Hologram

Object beam

Displaced position

- Original position

Figure 10-25 Ray diagram for evaluating surface displacement.

 10-4 Holographic Interferometry 189
Sec.

is assumed that the original phase of the two rays at point O and the phase
hift due to reflection from the object surface are both zero. The relative phase
of the two rays at point Q is
Acf> = cf>2 - ^ = ki • (ri - TO - k2 • (r2 - r2) (10-68)
Vector differences ri — TI and r2 — r2 are both equal to the surface displacement
d. Thus
A<f> = (k! - k2) • d (10-69)
To demonstrate the use of Eq. (10-69), let us consider a circular diaphragm
with the object beam originating from a point source and at an angle relative
to the z axis as shown in Fig. 10-26. The vector difference
Ak = ki - k 2 (10-70)
is parallel to the bisector of the angle between ki and —k 2 and has a magnitude
of 2k cos (v|//2). Therefore
A<f> = A/c d cos 0 (10-71)

or A<f> = cos cos (10-72)

where 9 is the angle between the vector difference Ak and the surface displace-
ment d and \\i is the angle between ki and —k2. Note that the sensitivity of
the system is determined by both the angle 9 and the vector difference magnitude
hk. For simplicity, 6 and v|//2 are equal in Fig. 10-26. It should be realized

Hologram

Observer

Reconstruction
beam

^'gure 10-26 Simple holographic interferometer configuration with object beam illu-
minating object surface at an angle relative to surface normal.
190 Low-Power Laser Applications II: Interferometry and Holography Chap

that this may not be true for other configurations. Equation (10-72) reduce
to Eq. (10-64) when the surface displacement and the object beam direction
before and after reflection from the object surface are all parallel to the
axis. To determine d with a single observation, the angle 0 must be known.
In completing the analysis of our circular diaphragm, assume that
C/a = —a(x, y) (10-73)
and C/3 = -a(x, y) exp [j A<J>(x, y)] (10-74)
where Ua and t/g are the complex amplitudes of the light reconstructing the
images due to the first and second exposures, respectively. The resulting irradi-
ance due to the interfering light is
,y)
/(*, y) = cos"

Unlike the real-time case, the irradiance is maximum when the relative phase
between the two beams A<f>(x, y) is 277m, where w = 0, 1, 2, 3, 4, . . . .
Figure 10-27 is a sketch of the resulting image that would be observed by
using double-exposure holographic interferometry and the holography system
illustrated in Fig. 10-26. The surface displacement can be calculated by counting
bright fringes from the stationary portion of the object, using

d=- — (10-76)
2 cos (i|//2) cos 0
Recall that 0 is the angle between the vector difference Ak and the surface
displacement d. i|/ is the angle between k! and —k 2 or the angle between the
object light illuminating a displaced point on the object surface and the light
scattered from that point to the observer.
Example Problem
Find the displacement of the diaphragm center illustrated in Fig. 10-27 if a HeNe
laser is used in the holography system and <\i/2 = 6 = 22°.

m=0

Figure 10-27 Double-exposure

graphic image due to normal displacern
m= 1 of a circular diaphragm.
Sec 10-4 Holographic Interferometry 191

Solution- Using Fig. 10-27, m is 3 and A. for a HeNe laser is approximately 633 nm.
Using Eq. (10-76), we have
= 3(633 nm)
~ 2 cos2 22°
Thus
d = 1.10
When the direction of the displacement is not known, then the measure-
ment that can be easily determined is the component of the displacement parallel
to the vector difference Ak—assuming that a stationary portion of the object
exists and can be observed. Obviously, three components of the displacement
can be determined by making three observations from three different angles.
This step can be accomplished by using three different holograms or by making
three different observations through a single hologram. This situation indicates
that the fringe location relative to an object changes as an observer views the
displaced surface from different angles. Such, in fact, is the case. The fringes
can exist in a plane other than the object surface. (A more extensive coverage
of holographic interferometry is given by Vest 1979.)
Displacement measurement in many industrial applications is not as impor-
tant as knowing which portion of an object surface experiences the greatest
displacement between the two exposures of a hologram. Figure 10-28 is a photo-
graph of a double-exposure hologram image that was produced with a holography

Pi
l8ure
10-28 Photograph of recon-
image from a double-exposure ho-
8ram produced using a double-pulsed
y Ias
s er. (Courtesy of Detroit Diesel Alli-
°n Division, General Motors Corpora-
don.)
192 Low-Power Laser Applications II: Interferometry and Holography Chap ir.

system using a double-pulsed ruby laser and reconstructed with a HeNe laser
Here the outside case of a gear box was vibrating at one of its resonant frequencies
at a particular operating condition. It is obvious that the surface area located
on the right side of the gear box is vibrating with maximum amplitude. To
reduce vibration and thus the noise produced by the outer surface of this gear
box, for the operating condition investigated, this portion of the surface must
be treated to reduce its motion.
Vibration analysis is one of the most common industrial applications of
holography. Most other techniques of vibration analysis require that a device
such as an accelerometer, be mounted on a vibrating surface. Holography is a
noncontact technique. When the object to be analyzed is small, holography
offers distinct advantages over other techniques.

Time-average holographic interferometry. Many laboratories do

not have a pulsed laser system that can be used for vibration analysis by double-
exposure holographic interferometry. The alternative is time-average holographic
interferometry, which can be accomplished using a low-power continuous wave
laser.
To simplify our analysis, we will assume, as in the discussion of real-
time holographic interferometry, that a plane wave is used to illuminate the
object. Furthermore, we will assume thaTthe propagation direction of the light
before and after being scattered from the object surface and the surface displace-
ment are both parallel to a z axis that is normal to the object surface as shown
in Fig. 10-23. In this analysis, however, we will assume that the circular dia-
phragm is vibrating sinusoidally such that
z(x, y, t) = D(x, y) cos wDt (10-77)
where D(x, y) is the amplitude of vibration at position (x, y) on the diaphragm
and o)D is the angular frequency of vibration. The corresponding phase change
due to this vibration is
A4>(x, y, t) = 2kD(x, y) cos a>Dt (10-78)
The complex amplitude of the object beam at the hologram will be
U0(x, y, t) = A(x, y) exp [2jkD(x, y) cos o>Dt] (10-79)
We will assume that the period of vibration of the diaphragm is much smaller
than the exposure time. Information recorded by the hologram due to the objec
will be the time average

U0(x, y) =^L IT A(x, y) exp [2jkD(x, y) cos a>Dt] dt 0&-80)

•* J o
where T is the exposure time. This integration yields
U0(x,y) = A(x,y)J<j[2kD(x,yy\ • (10-80
where J0 is a Bessel function of the first kind of order zero.
 10-4 Holographic Interferometry 193

1.0 i i T i i n ^
0.8

0.6

0.4

0.2

0 12 16 20 24

Figure 10-29 Plot of J£(a) versus a.

When the hologram is developed and reconstructed with a reconstruction

beam that is identical to the original reference beam, the complex amplitude
of the light producing the virtual image is
f/a = (bp!R)A(x, y)J0[2kD(x, y)] (10-82)
The irradiance of this light is

c, y)J*[2kD(x, y)] (10-83)

2Z
Figure 10-29 is a plot of J$(d) versus a. Note that the positions on the
object that do not vibrate are easily determined because they are brighter than
other regions where higher-order bright fringes exist. The virtual image of a
sinusoidally vibrating circular diaphragm, vibrating at its fundamental resonant
frequency, is illustrated in Fig. 10-30. The outside perimeter of the diaphragm
is very bright, indicating that this region is stationary. The dark fringes corre-

Stationary
region

m=0

10_30 Virtual holographic image

a
circular diaphragm vibrating sinusoi-
y at its fundamental resonant fre-
luency.
m= 1
194 Low-Power Laser Applications II: Interferometry and Holography Chap

spond to the zero values of the Bessel function. The bright fringes have a decreas
ing irradiance as the order of the fringes increase. That is, the center bright
spot on the diaphragm illustrated in Fig. 10-30 has less irradiance than th
circular bright fringe adjacent to it and each larger diameter fringe is brighter
Example Problem
Determine the amplitude of vibration of the circular diaphragm shown in Fig. 10-3Q
if a HeNe laser operating at a wavelength of 633 nm is used to produce the time-
average hologram. As in the preceding development, assume that the holographic system
used is as illustrated in Fig. 10-23.
Solution. The central bright spot is a third-order bright fringe. From Fig. 10-29 we
can determine that
2kD(Q, 0) = 10.2
Thus
(633 nm)(10.2)
D(0, 0) =

and D(0, 0) = 514nm

Example Problem -'"
Figure 10-31 shows the ray diagram of a time-average holography system and virtual
image produced with this system that is used to determine the mode and amplitude of
vibration of a rod fixed at one end. The rod displacement and direction of the object
beam light scattered to the photographic plate are both along the z axis, which is normal
to the equilibrium position of the rod. The angle between the object beam light illuminating
the rod and light scattered to the photographic plate is 30°. Determine the mode of
vibration and vibration amplitude of the rod tip if a HeNe laser operating at 633 nm
is used to produce the hologram.

Hologram

Rod

Reference beam Virtual image

Object beam

Figure 10-31 Ray diagram of a time-average holography system and virtual image
produced with the system to evaluate a vibrating rod.
 10
ChaP- Problems 195

Solution- The bright rod base indicates that it is stationary. No other areas of equal
1,,-jehtness are present. Therefore the rod is vibrating at its fundamental resonant fre-
auency. A dark fringe is located at the rod tip. Using Fig. 10-29, we determine that
this dark fringe corresponds to a = 14.93. Using Eq. (10-72) and substituting a for
AJ> and D for d, we obtain
14 93X
D =. '
4rr cos (i)//2) cos 6
Both 4*/2 and 6 are equal to 15°. Thus the amplitude of vibration of the rod tip is
806 nm.

PROBLEMS

10-1. Given the expression for the electric field intensity

10 uV /107 , 3 X 1015 , n \
E = —^ cos I — z H t + 0.75 I
m \m s /
find the
(a) Amplitude
(b) Propagation constant
(c) Wavelength
(d) Angular frequency
(e) Frequency
(f) Original phase angle
(g) Direction of propagation
(h) Irradiance in free space by using Eq. (1-7)
10-2. Given the expression for the complex amplitude of the electric field intensity
due to a point source
lOOoV/m
c ... [.fl.22XW \1
U= cos (40 exp ; I m rI
T L \ /J
find the
(a) Amplitude
(b) Propagation constant
(c) Wavelength
(d) Frequency
(e) Original phase angle
(f) Irradiance at 4* = 15° and r = 1 m
"~3. Determine the maximum and minimum distance between adjacent bright fringes
due to two interfering laser beams that can be described as coherent plane waves.
Derive the expression for the irradiance / as a function of z due to two interfering
plane waves
I/i = i£o exp (jkz)
and l/z = i£0 exp (—jkz)
196 Low-Power Laser Applications II: Interferometry and Holography Chap

Plot irradiance / as a function of z and determine the distance between adjace

bright fringes. Repeat the exercise for the two plane waves
Ui = iEtt exp (jkz)
and U3 = jEa exp (—jkz)
10-5. Show that the expression for the radius of the dark fringes shown in Fig.

where m 0 is the order of the central dark fringe and n = m — m0, where tn is
the fringe order. Measure the radii of the dark fringes in Fig. 10-7. Then use
log-log paper or a power curve-fitting computer program to show that the equation
for the dark fringe radius has the form
r = anb
Determine a and b.
10-6. Two laser beams with opposite circular polarizations can be described by
Ei = A(i + ij) exp (—joit)
and EI = A(i — ij) expj|—jo>t)
Describe the beam produced by the superposition EI + £2 and the resulting
irradiance. Repeat the exercise for two beams with opposite circular polarizations
that have slightly different optical frequencies.
10-7. Using the count versus time data given in Fig. 10-32, determine the displacement
of the moving reflector of a two-frequency interferometer. Plot the moving reflector

40

30

20

t
-10
Time (milliseconds)

-20

Figure 10-32 Graph for Problem 10-7.

Chap 10 Problems 197

p
>gure 10-33 Photographs for Problems
10
-13 and 10-14.
198 Low-Power Laser Applications II: Interferometry and Holography Chap
10

displacement versus time. Determine the velocity of the moving reflector f

each time interval and plot a velocity versus time curve.
10-8. Given the expression for the complex amplitude
U — A exp (jkz) exp (y'<f>)
find the direction of propagation of EI = U exp (—ja>t) and £2 = U* exp
(-/cor).

10-9. Given the expression for the complex amplitude

describe this wave and its complex conjugate.

10-10. Describe the results if a hologram is reconstructed with the object beam.
10-11. Find the displacement of the diaphragm center illustrated in Fig. 10-24 if the
angle between the object beam and surface normal before and after it is scattered
from the diaphragm is in the same plane and equal to 30°.
10-12. Using the holography system and virtual image illustrated in Fig. 10-31, determine
the order and type of fringe located at the tip of the vibrating rod if a double-
pulsed ruby laser was used to produce a double-exposure hologram with the
two exposures taken with the rod in positions of maximum positive and negative
displacement relative to equilibrium. ,-
10-13. The series of photographs shown in Fig. 10-33 were taken during a holographic
interferometry experiment. The object is a rectangular steel plate clamped at
one end.
(a) State which is a photograph of the image due to a real-time, double-exposure,
and time-average hologram.
(b) State the reasons for your conclusions.
10-14. The holography system geometry used in the experiment described in Problem
10-13 is identical to the one illustrated in Fig. 10-31, except \\i = 22°.
(a) Determine the maximum displacement of the steel plate by using the double-
exposure and real-time hologram photographs shown in Fig. 10-33.
(b) Determine the maximum vibration amplitude of the steel plate by using the
time-average hologram photograph.

REFERENCES

GASKILL, J. D., Linear Systems, Fourier Transforms, and Optics. New York: John w« e
& Sons, Inc., 1978.
GOODMAN, J. W., Introduction to Fourier Optics. New York: Academic Press,
1968
- ,9
19
VEST, C. M., Holographic Interferometry. New York: John Wiley & Sons, Inc.,
ChaP- 1^ Bibliography 199

BIBLIOGRAPHY

10_1. Caulfield, H. J. (ed.), Handbook of Optical Holography. New York: Academic

Press, Inc., 1979.
10-2. Collier, R. J., C. B. Burckhardt, and L. H. Lin, Optical Holography. New York:
Academic Press, Inc., 1971.
10-3. Fowles, G. R., Introduction to Modern Optics. New York: Holt, Rinehart &
Winston, Inc., 1968.
10-4. Snyder, J. J., "Laser Wavelength Meters," Laser Focus, May 1982, pp. 55-61.
 chapter
^^
Interaction
of High-Power Laser Beams
with Materials

Several basic concepts of energy balance and heat transfer, as they apply to
laser processing, are presented in this chapter. The developments given are
intended to provide understanding and, in some cases, a reasonable approxima-
tion (at least order of magnitude) to reality. The types of applications to which
the concepts developed here are applicable are heat treatment, welding, and
material removal. Alloying, cladding, and glazing, less important industrially
at this time, can also be treated.
The interaction of a laser beam with the workpiece depends on several
laser beam parameters as well as material parameters. These parameters are
discussed prior to a treatment of energy balance and heat transfer concepts.

11-1 MATERIAL AND LASER PARAMETERS

Material and laser parameters that affect laser processing are discussed in this
section.
Material parameters

1. Reflectance R is the ratio of the power reflected from a surface to the

power incident on it. In Chapter 1 it was pointed out that R, for norma
incidence, is related to the refractive index by

R =
Sec. 1 1-1 Material and Laser Parameters 201

where « 2 and «i are the refractive indices of the substrate material and
incident medium, respectively. For conducting or absorbing dielectric ma-
terials, « 2 is a complex number and the square in Eq. (1 1-1) is an absolute
square. The magnitude of the refractive index for good conductors (metals)
is proportional to \/cr/27r/j,/, where cr is electrical conductivity, ju, is
magnetic permeability, and / is the frequency of the light. Consequently,
metals like copper and silver have high reflectances that increase with
decreasing frequency (increasing wavelength). It is found that the reflec-
tance of metals substantially decreases as the temperature nears the melting
point.
2. Absorption coefficient a is the fractional loss of light power per unit dis-
tance for light traveling in a nonmetallic material. Beer's (Lambert's) law
relates power to absorption by
P=P0e-«* (11-2)
where P0 is the power entering the surface and P is the power at depth
z. The absorption coefficient can also be interpreted as the penetration
depth or distance at which the power has dropped to one over e (37%)
of the value entering the surface.
3. Specific heat C is the energy required to raise the temperature of unit
mass one degree. The SI units are joules per kilogram-Celsius degree but
calories per gram-Celsius and BTU (British thermal unit) per pound —
Fahrenheit degrees are also common. A frequently useful variation is the
volume specific heat Cv given by pC, where p is the mass density of the
material. Cv is the energy required to raise the temperature of unit volume
one degree, or energy per unit volume-degree.
4. Thermal conductivity k is the heat flow per unit area per unit thermal
gradient. The units are presented in a variety of ways, but a look at the
one-dimensional heat conduction equation, which relates rate of heat flow
Q to thermal gradient dT/dz,

Q = -kA^ (11-3)
dz
indicates that the SI units for k are watt/m • °C, but watt/cm • °C are
frequently used.
5. Thermal diffusivity K is related to thermal conductivity and volume specific
heat by

and is a measure of how much temperature rise will be caused by a pulse

of heat applied to the material. It also indicates how rapidly heat will
diffuse through the material. Units are generally cmVs. Materials with
202 Interaction of High-Power Laser Beams with Materials Chap i

high thermal diffusivity will experience a relatively small temperature ris

with good heat penetration for a given heat pulse at the surface. A material
with a low thermal diifusivity will undergo a relatively large temperature
rise at the surface with low heat penetration into the material for a given
heat pulse.
6. Latent heat L refers to the amount of heat required to cause a change
of phase of unit mass of material. Lf is the latent heat of fusion or the
energy required to cause melting of unit mass. Examples of units are
cal/g, BTU/lb, and the SI units joule/kg. Lv is the latent heat of vaporiza-
tion and has the same meaning and units as Lf. Latent heats of vaporization
are much larger than latent heats of fusion and vaporization also requires
more energy than that needed to raise unit mass up to the vaporization
temperature. Consequently, when vaporization is involved in processing,
it is a major factor in determining energy requirements.
7. Transformation temperatures refer to the melting temperature Tm, vapori-
zation temperature Tv, and other phase change temperatures Tp, such
as the martensitic transformation temperature on the iron-carbon phase
diagram. The last is particularly important in heat treatment of cast iron
and steel.

Laser beam parameters. Laser parameters relate to both the laser

beam and to the laser power output as a function of time. Normally lasers
are either operated CW or pulsed; the pulses may be produced by (^-switching.
The parameters of the beam that aifect processing are

1. Wavelength
2. Focused spot size
3. Mode structure

Wavelength affects absorption and reflection characteristics and spot size and
mode structure affect average irradiance and irradiance distribution in the spot.
The power output and traverse speed, or dwell time, are clearly important
for CW operation. In pulsed lasers the length of the pulse, energy per pulse,
peak power, and even the shape of the pulse influence the interaction of the
beam and the workpiece.
Pulses from solid lasers usually consist of a substructure or spiking. The
details of this spiking may have a significant influence on the quality of dnlle
holes.

11-2 BASIC HEAT TRANSFER EQUATIONS

The basic heat transfer equations considered are

¥ = -kVT (11-5)
Sec. 11-3 Every Balance Approximation 203

which relates heat flux F to thermal gradient 7" and

(11-6)
T-'Tk
K dt
which applies to non-steady-state heat conduction with A accounting for internal
generation of heat with units of watts/m3 or equivalent. A is zero in materials
where the incident laser power is absorbed at the surface. This statement will
be assumed for all remaining sections of this chapter except Section 11-8. When
the laser beam penetrates the material to any extent, the solution of Eq.
(11-6) requires consideration of a distribution of internal sources and A(z, t)
cannot be neglected.
Equation (1 1-5) is sufficient for steady-state heat transfer problems. Unfor-
tunately, laser applications seldom lead to steady-state situations. For the purpose
of this book, detailed solutions of Eq. (11-6) will not be undertaken, although
solutions that pertain talaser applications will be presented and discussed. Gener-
ally radiation and convection are not significant sources of heat loss in laser
processing and will be assumed negligible. *

11-3 ENERGY BALANCE APPROXIMATION

Simple energy balance approximations frequently provide reasonable ballpark

numbers for many applications. For that reason some discussion of this approach
is presented here.
Figure 11-1 is a schematic representation of a laser beam focused onto
the surface of a workpiece. If it is assumed that the material is heated to a
depth z with cross-sectional area Tra2, then the energy U required is given by
U = (CT+Lf + Lv)pira2z (11-7)

Laser beam

Workpiece

Figure 11-1 Laser beam focused on a workpiece.

204 Interaction of High-Power Laser Beams with Materials Chap
11

where T is used for temperature to avoid confusion with time t and represent
the change in temperature of the part. It has been assumed in Eq. (11-7) that
p and Cv are independent of temperature and are the same for liquid and
solid. Equation (11-7) includes the possibility of melting and vaporization. Onlv
the first term is required for surface hardening without melting; the first two
terms are used for welding; and all three are applicable for material removal
although the first two may be negligible. The depth of the cylinder heated in
Fig. 11-1 may be taken as the thickness of the part, if complete penetration
is desired, or to whatever depth requires heating. The depth is typically 0.25
to 1.3 mm in surface hardening.
Obviously the energy balance approach is crude; it ignores lateral heat
transfer and assumes uniform heating throughout the cylinder. Nevertheless,
it does provide a lower limit to the energy required in a given case and is
fairly accurate for small z or z < a and for t relatively short compared with
the time required for significant heat loss by conduction.
Once U is determined, the length of the pulse, assuming constant power
P, is just t = U/P. Looking ahead somewhat, the solution of Eq. (11-6) for
constant uniform irradiance on the surface of a semi-infinite solid leads to a
concept of thermal penetration depth, z t h, where
Zth = V4icf ^ (11-8)

Equation (11-8) can be used to estimate z in Fig. 11-1 for a given pulse length
or, conversely, it can be used to estimate the pulse length required to achieve

0.1 0.2 0.3 0.4

/4Kt

Figure 11-2 Ratio of temperature T at depth z to the surface temperature Ts versus

z-\/4Kr for uniform, constant irradiance.
Sec. Uniform, Constant Irradiance Model 205

a desired penetration depth. Actually, the temperature at z = \/4i<t is approxi-

mately one-tenth the temperature at the surface, relative to the initial tempera-
ture. Figure 11-2 is a plot of T(z)/Ts against a = z/^/4Kt, where Ts is the
temperature at the surface of the part. A depth of z = \/Kt is frequently
used to estimate thermal penetration depth and is perhaps more reasonable,
for at that depth the temperature change is about 35% of the surface temperature
change.
The energy balance approach can be used with a moving source. Assume
that the source in Fig. 11-1 is moving to the right with speed v (or the part
is moving to the left with speed — v). The time that the beam remains on any
one point is t = 2a/v. Then the energy delivered to a spot of radius a is
U = Pt or

U=^ (11-9)
v
Equation (11-9) can be used in conjunction with Eqs. (11-7) and (11-8) for
problems involving a moving source or part. This treatment is quite good for
nonmetals with high absorption coefficients.

11-4 UNIFORM, CONSTANT IRRADIANCE MODEL

A result of the uniform, constant irradiance model was presented in the previous
section to improve the energy balance approach to laser processing problems.
In this section the complete solution for the uniform, constant irradiance model
is given.
In this model it is assumed that a semi-infinite solid is irradiated with a
constant and uniform power per unit area. In all equations the irradiance 70
at the surface is assumed to be the actual power per unit area absorbed,
(1 — R)It, where /* is the incident irradiance. The solution of Eq. (11-6) for
the temperature, T(z,i), as a function of depth z into the solid and time t is
[Carslaw and Jaeger 1959]
1/2
(11-10)

where

ls
/:•
the error function and erfc (s) is the complimentary error function or
~ erf (s). Tables of the error function are available in most compilations of
Mathematical tables, such as Jahnke and Emde, "Tables of Functions." The
surface temperature of the semi-infinite solid is given by
206 Interaction of High-Power Laser Beams with Materials Chap. ] i

Jo /4KA. 1 ' 2

The graph in Fig. 11-2 was generated from Eqs. (11-10) and (11-11). Equation
(11-11) is useful for estimating the time required to reach a specified surface
temperature for a given irradiance.
In the case of a pulse of length tp, I(f) = IQ, T(z, t) is given by Eq
(11-10) for the duration of the pulse. For f > tp a solution is obtained by the
principle of superposition, thus for t > tp
T(z, t) = [f(t)-f(t-tp)] (11-12)
where f ( t ) is used to represent the right-hand side of Eq. (11-10).
Equations (11-10) and (11-12) are valid when the thickness of the part
exceeds \/4Kt and will give a reasonable estimate of the temperature along
the axis of the beam extending into the part if the beam width exceeds \/4-Kf~
Equations (11-10) and (11-12) are most accurately applied to laser heat treat-
ment when a defocused beam is used, possibly in conjunction with beam scanning
(dithering), rotation, or beam integration, to spread the power out in a more
or less uniform fashion over a fairly large area.

11-5 GAUSSIAN, CIRCULAR,

AND RECTANGULAR BEAM MODELS

Some pertinent results of models that more truly represent the nature of laser
beams are given in this section. More thorough discussions of these models
appear in several of the references, such as Duley 1976; Carslaw and Jaeger
1959; Ready 1971; and Charschan 1972. In all cases discussed in this section,
the irradiance is constant in time, and the substrate is considered to be a semi-
infinite region.

Gaussian beam. Equation (1 1-6) can be solved for constant absorbed

irradiance / distributed in a Gaussian fashion over the cross section of the
beam with 70 the absorbed irradiance at the center. Assuming that the power
is absorbed at the surface, the surface temperature at the center of the spot is
given by (11-7) [Duley 1976]

Circular beam. The solution of Eq. (11-6) for the temperature

the center of a spot of radius a with power P uniformly distributed over t
spot as a function of time and penetration depth z is [Duley 1976]
gsc. 1 1-5 Gaussian, Circular, and Rectangular Beam Models 207

P is the total power absorbed at the surface and ierfc stands for the
integral of the complementary error function and

ierfc(s) = —7= e ~g2 — serfc(s)

VT
The surface temperature at the center of the spot is given by
_,„. p(4Ko i/2 r i i .__ «2
(11-15)

The steady-state temperature as a function of penetration depth may be of

interest. It is given by

z* + a*)"*-z] (11-16)

Therefore the maximum surface temperature that can be attained is

Tmax = —
irak
(ii-i7)
It is important to point out that the surface temperature reaches 75%
of its steady-state value for t = a2/K, which for most metals and focused spots
is a fairly short period of time. A focused spot of 0.025-cm radius and a difFusivity
of 0.1 cmVs, for example, gives t = 6.4 ms. The quantity a2/K is referred to
as the thermal time constant and is a measure of the time required for significant
radial heat conduction loss to occur.
The results for the circular and Gaussian spots do not differ by a great
particularly on the beam axis for values of z > 0. At the surface the
steady-state temperature at the center of a Gaussian spot is about 25% greater
than the temperature at the center of a circular spot for equal total power
and \v = a.

Rectangular beam. A laser operating in a higher-order mode may

Produce a beam whose cross section is approximately rectangular. If x and y
re
half the length and width of the spot, the steady-state maximum temperature
15
given by [Carslaw and Jaeger 1959]
P ( y x\
r m a x =- { x sinh-1 - + y sinh"1 - I (11-18)
2-n-xyk \ x y/
208 Interaction of High-Power Laser Beams with Materials Chap i

The maximum temperature for a square beam with x = y = a is

0.885P
T
* max = •
irak (11-19)

which is about 89% of the maximum steady-state temperature for a circular

beam of radius a.
The constant irradiance model gives approximately the same results as
the three-dimensional models for a dwell time less than the thermal time constant
a 2 /K, which is, as has been shown, typically in the range of milliseconds. The
uniform irradiance model can provide order of magnitude estimates for milli-
second-range welding pulses and for continuous operation if the travel rate is
high. A speed of at least 80 mm/s (189 in./min) for steel would be required
for a spot diameter of 0.5 mm in order to satisfy the criterion that the beam
dwell time on any one spot be less than the thermal time constant. If this
criterion is not met, one of the other models should be used.

11-6 HEATING WITH MELTING

The process of melting a material with the energy supplied by the laser is
relatively simple to analyze if the time to bring the surface to the melting point
is short compared with the thermal time constant. The time to reach melting
at the surface tm can be estimated from the uniform irradiance model by

(11-20)
4K \ 70 J
This is usually much less than a microsecond for typical welding situations.
On the basis of energy balance, the volume of material melted can be estimated
as

V= (11-21)
p(CTm + Lv)
where U is the total energy input. Because conduction losses have been neglected,
U is a lower limit to the energy required. Doubling U usually gives a reasonable
estimate of the actual energy needed. The length of time to reach vaporization
can be estimated from
L-y (n-22)
5
It is typically milliseconds for pulsed welding with 70 on the order otf lO
W/cm2. The parameters K and k are not constant with temperature, but reasona-
ble calculations can be made by selecting an intermediate value for the rang
of temperatures under consideration. Because more time is spent in the molte
if
Sec. 1 1 ~6 Heating with Melting

state than the solid state during heating, the most appropriate values of K and
^ are averages over their values in the molten range of interest.
As an example, consider the welding of steel with a CO2 laser capable
209
1
Of providing a power of 2.5K W at the workpiece. A weld depth of 2 mm is
desired with a weld width of about 1 mm. The appropriate parameters are
rm = 1547°C TV=2752°C
Room temperature is Ta + 16°C, p = 7.87 g/cm3, L/ = 272 J/g, K = 0.21
cmVs, k = 0.75 W/cm°C, and C = 0.46 J/g°C.
According to energy balance, U = pV[CTm + Lf] = 11 J. U should be
doubled to compensate for heat conduction losses. Assuming that half the power
is coupled into the part, tp = U/P = 19 ms. In actual practice, a high-order
CO2 laser beam focused to a spot diameter of approximately 0.5 mm was used
with a traverse speed of 2.54 cm/s. The dwell time of this beam on any one
spot is t = 0.05 cm/(2.54 cm/s) = 20 ms. The close agreement of these times
is coincidental, for the choice of a 1.0-mm wide weld and 50% reflection are
somewhat arbitrary. According to the uniform irradiance model, the time to
reach Tv at the surface is

f
» = T- ~r = 0.04 ms
4K\ I /

Obviously this time is much too short to allow any useful penetration without
substantial material removal occurring. The development of a "keyhole"1 in
the high-power continuous welding, which allows the beam to penetrate unim-
peded deep into the part, is absolutely essential and is what makes possible a
reasonable estimate using the energy balance approach. It is believed, however,
that nearly all the incident energy is coupled into the part during keyholing.
An irradiance of 5 X 105 W/cm2 (the irradiance used in this example) will
tend to drill except in the case of continuous welding where a keyhole develops.
The uniform irradiance solution gives reasonable results in microwelding
°f small, thin parts. If the penetration depth i and 70 are known, the graph
In
Fig. 11-2 can be used to find a value of a at T(z)/Ts = Tm/tv. Then the
Pulse length is obtained from a = i/\/4Ktp.
Another approach, which takes into account the formation of a liquid
r
°nt and its propagation through the workpiece, has been developed by Cohen
aij
d Epperson, 1968. Uniform, constant irradiation for some time t is assumed
ln
this approach; therefore good results may be expected when the spot size
*ceeds the penetration depth and when t < a2/K, where a is a dimension
a
racterizing the spot size, such as the radius for a circular spot. The solution
as
obtained via an analog computer and is expressed in graphical form in
t '8- 11-3. To use this graph, calculate r = 40Tv/Tm and Y = Lf/CTm (tempera-
re
s are relative to ambient). Locate T on the left-hand vertical scale and go
1
The phenomenon of "keyholing" is discussed in more detail in Section 11-9.
210 Interaction of High-Power Laser Beams with Materials Chap i <

Figure 11-3 Cohen and Epperson's technique for determining melt depth. T = 40
Tv/Tm, Z = 8.33 pL/Z//r m , T = tv/tm, Y^Lf/CTm. (Adapted with permission
of the Journal of the Franklin Institute, Vol. 283, No. 4, April 1967, pp. 271-85.)

horizontally across to the appropriate upper curve for Y. Drop vertically to

the corresponding Y curve of the lower set. Dropping vertically to the horizontal
axis from this last point provides r = tv/tm and proceeding horizontally to
the right vertical axis gives Z = 8.33 pLfz/Itm. The times tv and tm are the
times to reach vaporization and melting at the surface, respectively; tm is calcu-
lated from Eq. (11-10). The depth of penetration z is determined from the
expression for the nondimensional depth Z.
An example of the application of Cohen and Epperson's technique is ob-
tained by considering spot welding of two sheets of 0.4-mm-thick steel together
with a beam focused to a spot diameter of 1.25 mm. The most suitable laser
for this application would be a Nd-YAG or Nd-Glass laser. In this application
it is assumed that the pulse length and absorbed irradiance are unknown. The
penetration depth z equals 0.8 mm and the surface temperature Ts equals »
at the end of the pulse. Using the parameters given for the previous examp e
and Eq. (11-11), we have

'•1— = 13 X 103W/cm2
(4)p8.33KL/Z
which means a total power coupled into the piece of I7r(0.063 cm)2 = 1°°
Assuming 50% reflectance, 320 W would be required. The pulse length is
 1
Sec. 1 ~7 Heating with Vaporization 211

75(1526)12
T-
7T
= 29ms = '» = Ttm = 87 ms
Li
4(0.21) [ 13 X 103 J
This time is considerably longer than the thermal time constant, which is (0.063
cm)2/(0.21 cmVs) = 19 ms. Consequently, a fairly large quantity of heat will
be lost to conduction. Much better results are obtained when the weld depth
is much less than the spot diameter. A higher irradiance can be used for thinner
materials, leading to much shorter times to reach vaporization with full melting
through the piece. One point that this example makes is that there is a limit
to weld depth (without keyholing) unless significant material removal is allowed
at the surface. In practice, weld depths in laser or electron beam welding would
be limited to about 1 to 2 mm if it were not for keyholing.

11-7 HEATING WITH VAPORIZATION

At power densities sufficient to produce rapid vaporization, typically 106 W/

cm2 and higher, the time to reach vaporization can be estimated from Eq.
(11-10) because very little heat penetration occurs in this length of time. Using
the data for steel and an irradiance of 106 W/cm2, for example

Once vaporization occurs, a vapor front begins to move into the material, pre-
ceded by a liquid front. This situation is depicted in Fig. 1 1-4. If the material
removal process occurs in a relatively short period of time, the energy require-
ment and pulse time can be reasonably estimated by using the energy balance
approach. The energy required to remove a given volume of material is given
by Eq. (11-7). For a circular spot of radius a and thickness of material removed

Vapor

Liquid

Solid

li_4 Vapor front propagation in

' solid.
212 Interaction of High-Power Laser Beams with Materials Chap i •

z, V = 7ra2z. It has been assumed that the specific heat and density are constant
with respect to temperature and are the same for solid and liquid. The heat
of fusion has been neglected. Letting the depth of material removed be a variable
and taking the derivative of Eq. (11-7) yield
dU
P=—

where v = dz/dt is the speed at which the vapor front moves into the material
The liquid interface between the vapor and solid is extremely thin and is ignored
in this analysis. Because / = P/naz, the speed of the vapor front can be calculated
from
7
v=- (11-24)
p(cTv + Lv)
The time tp required to remove material to a depth z, is then simply
z
(H-25)
V

As an example, consider piercing a 0.5-mm-diameter hole through a 1-mm-

thick sheet of steel with L,, = 6350 J/g. The energy required, using Eq.
(11-21), is
f / = 11.8 J //
If the irradiance is 106 W/cm2, the pulse length is, by Eqs. (11-24) and
(11-25),
tp = 6 ms
Actual drilling times are 0.1 to 1.0 ms in practice.
If instead of piercing a hole it is assumed that the material is being cut,
in the previous example, the cutting rate can be estimated from v = 2a/tp =
0.05 cm/6 ms = 8.3 cm/s (197 in./min). This is in surprisingly good agreement
with experience. Oxygen assist is used to oxidize the metal and to blow away
molten material in most metal-cutting applications. In this technique 60 to
70% of the cutting energy is supplied by the oxidation process and not all
the material removed from the kerf is vaporized. The energy balance approach
assumes that all the kerf material is vaporized and that no conduction losses
occur. These factors are somewhat compensating, which helps to account wf
the accuracy of the crude energy balance calculation.
When material removal involves pulses with variable power, the energy
balance approach is still relatively easy to use if the temporal dependence
the power is known. Using Eq. (11-24), the depth of material removed can
be expressed as an integral

z= f'
p(CTv+Lv))
))00
geC 1 1-7 Heating with Vaporization 213

Figure 11-5 Generalized laser pulse. (Adapted from J. F. Ready, Effects of High-
Power Laser Radiation, with permission of Academic Press.)

This integral can be evaluated for a generalized pulse,2 which applies to a large
number of pulsed or ^-switched lasers. Figure 11-5 is a plot of this generalized
pulse.
The irradiance is given by /i(f) = 5Imax(t/tp) up to t = 0.2tp. And from
t = 0.2tp to t = tp the irradiance is given by 72(0 = 1.25 /max (1 — t/tp).
Putting these expressions into Eq. (11-25) and integrating yield
0.5/ n
(H-27)
P(CT;
Of course, this result, for a simple triangular pulse, could have been written
down immediately, based on the area under the curve in Fig. 11-5. Inherent
in this approach is the assumption that the vaporization temperature at the
surface is reached in a time much less than tp and that full penetration is
reached before / drops too low to maintain a vapor front.
The rate of material removal for irradiances of the order of 108 W/cm2
becomes limited by the rate at which atoms can leave the surface of the vapor
front. The ultimate limiting velocity of atoms in the vapor is the speed of sound
ln
the solid—that is, the maximum speed with which an atom can escape. If
'he rate of energy input is too great, the vapor does not have time to clear
he way for the laser beam and the vapor becomes superheated—that is, T >
*»• Consequently, much of the input energy ends up being absorbed by the
Va
por. In fact, short, high peak power pulses tend to give rise to a vapor plume
that nses above the workpiece and may totally attenuate the laser beam, thereby
asting most of the energy in the pulse. Longer, lower peak power pulses are,
^ fact, superior for most hole-piercing applications. Such pulses are typically
e order of magnitude of 0.1 to 1.0 ms and the energy balance approach

'Suggested by Ready, 1971, p. 75.

214 Interaction of High-Power Laser Beams with Materials Chap. 1,

is reasonable. Because most industrial applications are of this nature, the energy
balance approach is the only approach presented here. More rigorous treatments
are available in more advanced books [Duley 1976, Charschan, et al. 1977].

11-8 DISTRIBUTED SOURCES

A brief discussion of the results of the analysis of internal absorption of a

laser beam is presented in this section.
The basic law governing the absorption of light in a dielectric medium
in Beer's (Lambert's) law.
I = I0e-«* (11-28)
where /o is the irradiance entering the surface of the medium at z = 0 and
a, the absorption coefficient, is the fractional loss of irradiance per unit length.
The reciprocal of a is the depth at which the irradiance has dropped to 1/e
(37%) of its value entering the surface and can be thought of as a measure of
penetration depth. The absorption coefficient for many plastics, such as polycar-
bonate, is 0.33 cm"1 for visible light. ZnSe has an absorption coefficient of
1.0 X 10~3 cm"1 at a wavelength of 10.6 jum. Glass/has an absorption coefficient
of 0.05 cm"1 for visible radiation. Plastics have an absorption coefficient in
the range of 10 to 20 cm"1 at a wavelength of 10.6 pm.
For uniform irradiance, constant in time, the solution of the heat conduc-
tion equation, including a distributed source, leads to the result that the tempera-
ture of the heated material rapidly rises to a value given by [Duley 1976]

(11-29)
ka
to a depth of approximately a"1. The thermal conductivities of plastics range
from 0.001 to 0.004 W/cm°C. This means that less than 100 W/cm2 is required
to raise plastics to their melting point to a depth of 0.5 mm almost instantly
when using a CO2 laser.

11-9 CW WELDING (KEYHOLING)

With the advent of CO2 lasers with 1.0 kW and higher CW power output, 1'
was discovered that the phenomenon of keyholing occurs for continuous sea
welding. A similar phenomenon was known to occur for multikilowatt electro
beam welding. In this phenomenon a hole is produced in the material, alloW'1^
the beam to penetrate relatively unimpeded into the part. Apparently at p°w
inputs of 1 kW and higher the vapor pressure of the molten material becofli
sufficiently high to overcome surface tension and pushes the molten mater1
Sec. 1 1-9 CW Welding (Keyholing) 215

O ut of the way, forming a hole, or cavity, that is virtually 100% absorptive.

f he molten material flows back into the hole after the beam has passed. Generally
the higher the power, the deeper is the cavity that can be formed. This phenome-
non is illustrated in Fig. 11-6.
If it were not for the formation of a keyhole, laser (as well as electron
beam) welding would be limited to penetration depths of about 1 mm by the
thermal properties of metals. Using keyholing, weld depths of several centimeters
have been achieved with very high power CW CC>2 lasers.
Swift-Hook and Gick have modeled the keyholing process by use of a
linear heat source of total power P extending into the metal a distance a [Swift-
Hook and Gick 1973]. Referring to the geometry depicted in Fig. 11-7, their
result for the temperature distribution is
/vx\ /vr\
exp (11-30)
2irak \2Kj K0\2K)
Here v is the weld speed, K0 is the Bessel function of order zero, and r =
V x2 + y2 = x cos <f>. The width of the weld can be determined by letting
T = Tm, cf> = 90° and solving for r = w/2. For high welding speeds, Eq.
(11-31) gives a reasonable estimate of weld width.

w = 0.484 - (11-31)
vakTm
This will apply to most multikilowatt CO2 laser welding. The graph on p.
198 of Duley (11-7) can be used to determine whether Eq. (11-31) is applicable.

Solid weld

Figure 11-6 Schematic of "keyholing."

216 Interaction of High-Power Laser Beams with Materials Chap \ i

High-intensity
beam power w

Molten
pool

Keyhole

Penetration
depth a

Figure 11-7 Linear heat source model of laser keyholingf (Courtesy of Optical Engi-
neering and United Technologies Research Center.)

11-10 GENERAL APPROACH TO LASER HEATING

Many treatments of the general problem of heating a solid with a laser exist.
Considerable success has been achieved if surface absorption of the beam is a
valid assumption, particularly in surface modification applications in which melt-
ing and vaporization are not involved. Complete solutions in closed form are
rarely, if ever, possible for the real cases. Numerical, finite difference and finite
element techniques have been used with success, however. Several references
are listed at the end of the chapter for the reader interested in pursuing this
subject at some length. Most solutions include the assumption of constant ther-
mal constants, which permits the use of the principle of superposition, based
on the linear nature (with this assumption) of the general heat conduction equa-
tion (Eq. 11-2).
Difficulties are involved in the application of any of these approaches.
The variation of thermal constants with temperature is not well known f°r
materials of interest. An absorptive coating must be used in heat treatmen
with CO2 lasers and the absorption efficiency (ratio of absorbed power to inciden
power) and its variation with temperature and coating thickness are not we
understood. To complicate matters, the measurement of laser power is on y
Sec. 11-10 General Approach to Laser Heating 217

fairly accurate for long-wavelength lasers, such as that of the CO2 laser. Neverthe-
less, the qualitative agreement between calculated and measured temperature
profiles is excellent and the quantitative agreement is quite good.
An approach that is straightforward and that provides considerable insight
into heating with a laser will be described in some detail. It is a linear approach
for which the solution of Eq. (11-2) for an instantaneous point source serves
as a basis. This solution is

U
U
T(x, y, z, t) = (11-32)
4pC(pK0 3/2CXP
where U is the energy liberated at time t = 0 at point (x1, y', 0) on the surface
of a semi-infinite medium [Charschan 1972]. Using the principle of superposition
allows solutions for any type of spatial and temporal energy distribution to be

Figure 11-8 Defocused laser beam incident on a part.

obtained. Several solutions have been reported in the literature. Arata et al.
have given results for moving rectangular, circular, and Gaussian heat sources
and have compared these results with experimental findings and with the predic-
tions of the uniform irradiance model [Arata, Marou, Miyamoto 1978]. The
moving Gaussian heat source is the easiest to solve and provides the simplest
result. The results for the Gaussian source will be presented here to illustrate
the approach and to provide understanding of the laser heating process. The
Gaussian model is the least satisfactory of the three for actual surface modifica-
hon applications, but the results for the three models differ only slightly for
Predicted temperatures near the center line of the laser beam.
Figure 11-8 depicts a Gaussian laser beam incident on a part moving in
'he x direction with speed v. It is assumed that the instantaneous energy released
15
given by U = I(x', y') dx' dy' dt', where for a Gaussian distribution
I(x', y') = I0e exp - [x'2/a2 - y'2/b2]
218 Interaction of High-Power Laser Beams with Materials Chap.

To account for the fact that the part is moving, x' must be replaced by
v(t — t'). The temperature distribution for a point source is thus given by
Io [x-x' v(t - t')]2
T(x,y,z,t) = 3/2 [exp -
4pC[iTK(t - r')] - t')

In Eq. (11-33) t has been replaced by (t — t') because the latter represents
the time between the instant that the heat was released and the time of interest,
t; x, y, z are the coordinates of the point at which the temperature value is
desired whereas x' — v(t — t') and y' give the location of the heat source at
time t'. Equation (11-33) must be integrated for a continuous Gaussian source.
The limits on the x and y integrals are, because of the continuity of the Gaussian
function, — oo to +=°. The time integral is taken from — °° to the present time
t = 0. For any fixed point in space, the temperature rapidly approaches a
steady-state value, which is clearly the maximum value of the temperature at
that point. It is the maximum temperature isotherms that are of greatest interest
in surface modification applications.
If the variables are changed, the time integration reduces to an integration
over t' from zero to oo. By completing the squares in'the exponentials, the
integrand can be put in a form that is integrable over the spatial coordinates.
Then after some algebra the steady-state temperature distribution for a Gaussian
heat source is
Pk d?
T(x, y, z) = -x
3 2
(7TK) ' J 0 (?)
(x + vf') 2
4Kt'} (11-34)
2 2
V (b + 4*O (a + 4*0

0.5 1.0 1.5

0.2

0.4

0.6

Figure 11-9 Isothermal lines for a Gaussian beam heated part, z, r, and T are
normalized values. (Adapted with permission of Prof., Dr. Yoshiaki Arata, from
Application of Laser for Material Processing.)
Chap. 11 Problems 219

where P = I0(Trab) is the total power entering the part. The integration in
Eq. (11-34) can be carried out numerically to determine the temperature iso-
therms. Arata et al. present the isotherms in nondimensional form so that they
can be applied to any case regardless of the power, parameter values, or spot
size (see Fig. 11-9) [Arata, Marou, Miyamoto 1978].

PROBLEMS

11-1. (a) Using the energy balance approach, estimate the power required to produce
a spot weld 3 mm in diameter and 1 mm deep in low-carbon steel with a
1-s pulse from a CO2 laser.
p = 7.86 g/cm3, K = 0.7 cm/s
Cv = .7J/g°K, k = 0.4 W/cm°K
Estimated values near melting point are
Tm = 1809°K
Lf = 0.l kJ/mol
Assume 60% reflectance.
(b) Compare the depth given in part (a) with the thermal penetration depth
predicted on the basis of the uniform irradiance model for the power com-
puted. Remember, at this depth the temperature is only about 35% of the
surface temperature.
(c) According to the uniform irradiance model, how long does it take to reach
Tm at the surface for the power deduced in part (a)?
11-2. Repeat Problem ll-l(a), using the uniform irradiance model and calculate the
temperature at the surface at t = 1 s.
11-3. Apply the technique of Cohen and Epperson to Problem ll-l(a) to find the
necessary power and pulse length.
11-4. Compare the surface temperatures at the center of the spot for Gaussian and
circular beams of equal radii of 0.5 mm (1/e2 point for the Gaussian beam)
incident on titanium if the total power in each beam is 500 W and the pulse
length is 1.5 s. Use the following data for titanium.

W/cm°K
K = 0. 1 cmVs
50% reflectance

11-5. Calculate the maximum attainable surface temperature for an 8- by 6-mm rectangu-
lar beam for 2.5 kW incident on steel.
k =0.8 W/cm°K
60% reflectance
1~<>. Low-carbon steel of 5-mm thickness is to be cut with a laser at a speed of 1.25
m/min with a 0.5 mm kerf. Using energy balance, estimate the power that must
be coupled into the material.
220 Interaction of High-Power Laser Beams with Materials Chap. 1j

k = 0.4 W/cm°K
p = 7.86g/cm 3
C = 0.50J/g°K
Lv = 3.50 kJ/mol
Lf = 13.8 kJ/mol
Tv =3135°K
11-7. Mild steel is to be welded with a multikilowatt CO2 laser. Using the model of
Swift-Hook and Gick, estimate the power required to produce a weld of 2 mm
width and 4.5 mm depth at 1 m/min.
K = 0.09 cmVs
/c = 0.4W/cm°K
Tm = 1809°K

11-8. A plastic material is irradiated with a CO2 laser beam with a uniform irradiance
of 20 W/cm2. Calculate the instantaneous temperature rise.

k = 0.003 W/cm°K
a = 20 cm"1

11-9. A 0.4-mm copper wire is to be laser welded to the top oj>a 1-mm diameter, 70
to 30% brass post. For a focused spot size of 0.4 mrn and a melt depth into
the post of 0.5 mm, estimate the required energy and pulse length. Assume
50% reflectance.
The properties of Cu and brass are as follows:

Cu Brass (70-30)
3
p = 8.96 g/cm p = 8.53 g/cm3
C = 0.38J/g°K C = 0.37J/g°K
Tm = 1358°K Tm = 1230°K
Lf = 255 J/gm Lf = 200 J/g
A=4W/cm°K A: = 0.12W/cm°K

11-10. Holes are to be drilled in 0.5-mm-thick nickel with a diameter of 0.13 mm,
using a pulsed ND-YAG laser with a 4000-W peak power output. Use energy
balance to estimate the energy needed and the pulse length, assuming a nearly
square pulse. Show that the time to reach vaporization at the surface is negligible,
based on the results of the uniform irradiance model.

A: = 0.91 W/cm°K
C = 0.44 J/g°K
p = 8.9 g/cm3
Lf = 298 J/g
L»= 6303 J/g
Tm = 1726°K
7i =3187°K
R = 60%
 - 11 References 221

REFERENCES

ARATA, Y., H. MAROU, and I. MIYAMOTO, "Application of Laser for Material Process-
ing—Heat Flow in Laser Hardening," HWDoc. IV-241-78. Japan: Osaka University,
1978.
CARSLAW, H. S., and J. C. JAEGER, Conduction of Heat in Solids (2nd ed.). New
York: Oxford University Press, 1959.
CHARSCHAN, S. S. (ed.), Lasers in Industry. New York: Van Nostrand Reinhold Co.,
1972.
CHARSCHAN, S. S., et al. (eds.), The LIA-Material Processing Guide, Toledo, OH: Laser
Institute of America, 1977.
COHEN, M. I., and J. P. Epperson, Electron Beam and Laser Beam Technology. New
York: Academic Press, 1968.
DULEY, W. W., C(?2 Lasers: Effects and Applications. New York: Academic Press,
1976.
DULEY, W. W., et al., "Coupling Coefficient for CW CO2 Laser Radiation on Stainless
Steel," Optics and Laser Technology, December 1979.
READY, J. F., Effects of High-Power Laser Radiation. New York: Academic Press, 1971.
SWIFT-HOOK, D. T., and A. E. F. GICK, "Penetration Welding with Lasers," Welding
Research Supplement, pp. 492-499, November 1973.

BIBLIOGRAPHY

11-1 Ferris, S. D., H. J. Leamy, and J. M. Poate (eds.), "Laser-Solid Interactions and
Laser Processing-1978," AIP Conference Proceedings No. 50. New York: American
Institute of Physics, 1979.
11-2 Duley, W. W., Laser Processing and Analysis of Materials. New York: Plenum
Press, 1983.
11-3 Bertolotti, M. (ed.), Physical Processes in Laser-Materials Interactions. New York:
Plenum Press, 1983.
H-4 Bass, M. (ed.), Laser Materials Processing. New York: North-Holland, 1983.
 chapter

High-Power Laser
Applications

The purpose of this chapter is to discuss the techniques used in materials process-
ing with lasers and to point out some of the advantages and disadvantages.
The primary concern is with heat treatment, welding, cutting, hole piercing,
scribing, and marking. A brief discussion of laser-assisted machining (LAM),
alloying, and cladding is included. Appropriate theoretical techniques are pointed
out and some quantitative information is presented to provide the reader with
a feel for laser processing. There is, however, no intent to provide a compilation
of laser applications and processing data. Reference may be made to many
excellent books of that nature.

12-1 SURFACE HARDENING

Laser surface hardening (heat treatment) is a process whereby a defocused beam

(generally from a 1.0 kW CW or higher-power CO2 laser) is scanned across a
hardenable material to raise the temperature near the surface above the transfor-
mation temperature. Normally the cooling rate due to self-quenching by heat
conduction into the bulk material is sufficiently high to guarantee hardening-
Steel with over 0.2% carbon and pearlitic cast irons are readily hardened by
this technique.
There are several potential advantages to laser surface hardening:
1. The low amount of total energy input to the part produces a minima
amount of distortion and frequently parts can be used "as is" after
hardening.
r
Sec. 12-1 Surface Hardening 223

2. The depth of hardening is relatively shallow; thus desirable properties of

the substrate metal, such as toughness, are retained.
3. Selective areas can be hardened without affecting the surrounding area.
4. In some cases, the desired wear characteristics can be achieved in a lower-
cost material.

The chief disadvantages to laser hardening are

1. Low-area coverage rate, which limits the process to low production rates
or selective hardening.
2. The high reflectance of metals of 10.6-^.m radiation means that absorptive
coatings must be used.

Hardening with a laser beam does not differ fundamentally from any other
hardening process. In steel or iron the temperature at the surface is raised
above the Martensitic transformation temperature, but is maintained below the
melting point, as the beam scans the surface. The cooling rate near the surface
can easily exceed 1000°C/s; so the result is a hardened zone down to a depth
at which the temperature exceeded the lower Martensitic transformation tem-
perature. Figure 12-1 is an illustration of this process. A defocused beam of
approximate size d produces a hardened track of approximately the same width
and a hardened depth t determined by the laser power and coverage rate.
Most CO2 laser beams are not suitable for heat treating as they come
from the laser due to irregularities in the power density (irradiance) in the
beam cross section. Gaussian and other lower-order modes are not acceptable
because of the highly peaked irradiance distribution even if irregularities are
not present. Single higher-order mode beams are especially suited to hardening

Laser beam

Figure 12-1 Defocused laser beam incident on workpiece for heat treating.
224 High-Power Laser Applications
Cha P. 1
2

because of the large irradiance near the edges of the beam; this fact
to compensate for lateral heat loss and produces a more nearly rect ^^
heat-treat-track cross section. Unfortunately, it is difficult to maintain !! •
quality, high-order mode over a long period of time. Also, the beam m
be symmetrical, which complicates matters if contour tracks are required "^
Some method to produce a nearly uniform average irradiance profil
a specified spot area is used in many production applications. It may be ach' f
by rotating the beam optically, thereby producing an overlapping spiral tra\
or by dithering the beam (rocking the lens or a mirror) perpendicular to th
track, thus producing a zigzag pattern. Figure 12-2 is an illustration of these
techniques. Other techniques, referred to as beam integration, use stationar
optical devices to produce relatively uniform beam images or spots. Some are
basically hollow, polished copper tubes or pipes. The irregular beam enters
one end and a fairly uniform beam exits at the other. Another device, available
from Spawr Optical Company, is a segmented mirror. Each of about 32 segments
independently forms a rectangular image of the light that strikes it. The super-
position of the images formed by the irradiated segments provides a remarkably
uniform rectangular spot in the image plane. Spot size can be varied by the
use of additional optics. Figure 12-3 is a photograph of a Spawr segmented
mirror.
Actual heat-treat-track profiles agree quite well with calculated ones. Nu-
merical integration, finite diiference, and finite element analysis techniques have
all been successful in predicting heat-treat-track profiles for various irradiance
distributions (see Chapter 11). Depth of penetration can be reasonably well
predicted on the basis of the uniform irradiance model, which has an error
function solution. The chief difficulty with any theoretical approach is the uncer-

Beam rotation

Beam dithering

Figure 12-2 Active methods for spreading out a laser beam for surface trea
applications.
 17_1 Surface Hardening 225
Sec- '*

Figure 12-3 Segmented mirror for beam

integration. (Courtesy of Spawr Optical Re-
search, Inc.)

tainty in the value of the absorption coefficient1 and the thermal parameters
of the material.
Commonly used absorption coatings are black spray paint, graphite, phos-
phate, and zinc phosphate. It appears that paint is most effective at the higher-
power levels (~5 kW) because a fairly thick coating is required. Manganese
phosphate (Lubrite) is a common industrial coating and serves as a suitable
absorptive coating for lower-power levels of around 1 kW. Zinc phosphate has
comparable properties to manganese phosphate and graphite sprays work well
but are generally expensive.
The absorption coefficient for uncoated steel or iron may be as low as
10 to 40%, depending on surface condition. A suitable coating raises the absorp-
tion coefficient to 60 to 80%.
It is important that the proper thickness of coating be used. If the coating
18
too thick, too much energy is wasted ablating the coating. If the coating is
to
o thin, it completely burns off before the beam has passed and the bare metal
reflectance occurs for part of the pass. What apparently happens when the
c
°ating thickness is correct is that heat is absorbed by the coating and transferred
0
the metal by thermal conduction at a rate that prevents the coating from
ein
g totally destroyed. In effect, the metal provides a heat sink for the energy
e
nng the coating and the coating is removed at a rate that allows maximum
c
°Upling of beam energy into the part.
Figure 12^4 contains micrographs of heat treat tracks in 1045 steel and
cast iron. Hardness in the Martensitic region is near the maximum
Dieted by the carbon content. A thin overtempered zone may exist at the
er
r °f the hardened area. The ideal cross section for a heat treat track is a
angle to minimize overlapping, variation in hardened depth, and the amount
Ov
ertempered zone at the surface.
Ratio of absorbed power to incident power.
226 High-Power Laser Applications
Chap. 12

(a) Mild steel

(b) Cast iron /

Figure 12-4 Micrographs of heat treat tracks. (Courtesy of Saginaw Steering Gear,
Division of GMC.)

(a) (b)

Figure 12-5 Laser-heat-treated parts, [(a) Courtesy of Saginaw Steering Gear,

sion of GMC. (b) Courtesy Electro-Motive Division of GMC.]
 •''in

12-2 Welding 227

Sec.

Ideally an inert cover gas should be used during laser heat treatment to
:nimize oxidation of the metal and flame formation from the coating. It is
n
Omm° practice, however, to use air for economic reasons. A strong cross
floW should be used to protect the final optics from smoke and spatter from
the coating.
Figure 12-5 contains photographs of production laser heat-treated parts.
Figure 12-6 is a photograph of an industrial heat treatment setup. Figures
12-7 and 12-8 contain plots of penetration depth versus coverage rate for two
different materials.
Optimizing penetration depth in terms of coverage rate is a tradeoff be-
tween rapid and slow heating. If the surface is heated too rapidly, little heat
enters the metal before melting occurs and a shallow depth is obtained. If the
coverage rate is too slow, a substantial amount of heat is lost by thermal conduc-
tion. Generally higher power is required to achieve greater hardened depth.
To some extent, a narrower track can be run to increase power density, but
narrow tracks result in a significant amount of heat being lost by lateral (parallel
to the surface) heat conduction. Typical hardened depths in mild steel are 0.5
mm at 1 kW, 1.0 mm at 2.5 kW, and 2.0 mm at 5.0 kW.

Figure 12-6 Industrial heat treating setup. (Courtesy of Spectra-Physics, Industrial

Laser Division.)
^Prl
228 High-Power Laser Applications
Chap.
12

SAE 1045
Power = 2.5 kW

0.4 -

0.2 -

0.05 0.10 0.15 0.20 0.25

Coverage rate (cm 2 /s)

Figure 12-7 Hardened depth versus coverage rate for SAE 1045 for a 0.12-cm
diameter beam.

0.5

_ 0.4
E
Gray cast iron
power = 2.5 kW

0.3

0.2

0.5 0.7 0.9 1.1

2
Coverage rate (cm /s)

Figure 12-8 Hardened depth versus coverage rate for gray cast iron. Beam diam
approximately 5 mm.
cec. 12-2 Welding 229

l2-2 WELDING

There are basically two types of laser welding, CW and pulsed. Seam welding
is done both by CW CO2 and overlapping pulses with Nd-YAG lasers. Pulsed
Or spot welding is done chiefly with CO2, Nd-YAG, Nd-Glass, and, to a lesser
extent, ruby lasers.
The main advantages of laser welding are

1. Minimum heat input, which results in very little distortion.

2. Small heat-affected zone (HAZ) because of the rapid cooling.
3. Narrow, generally cosmetically good, weld bead.
4. High-strength welds.
5. An easily automated process that can produce very precisely located welds.
6. Weld some metals difficult to weld by other techniques, especially dissimilar
metals.
7. Weld in areas difficult to reach with some other techniques.
8. Frequently faster than other techniques.

Disadvantages of lasers in welding are

1. Extremely hard weld bead in hardenable materials, cold cracking, and

hot cracking may be a problem due to rapid heating and cooling.
2. High capital investment compared with most techniques, but this factor
may be more than offset by improved part quality, high laser up-time
(usually over 90%), and relatively low operating and maintenance costs.

Continuous laser welding. At present, continuous seam welding

is done only with CO2 lasers, usually 500 W or higher-power level. Somewhere
around 300 to 500 W keyholing begins to play a role in continuous welding,
but melting by thermal conduction is predominant up to around 1.0 kW. At
roultikilowatt levels, keyholing is the predominant phenomenon.
Theoretical approaches that give reasonable agreement for conduction lim-
ited welding are: simple energy balance, the uniform irradiance model, and
numerical, finite difference or FEM approaches. The biggest problem with calcu-
•ations in this regime is the uncertainty in the reflectance of the molten metal,
w
hich may approach that of the solid metal.
The method of Swift-Hook and Gick (see Chapter 11) works quite well
°r CW welding when keyholing is predominant. It is known that maximum
Penetration varies approximately as P0-1, where P is power. The energy balance
Pproach will give "ballpark" results if judicious guesses are made regarding
c
°nduction and reflection losses.
Nearly all power is absorbed in the keyhole and very efficient welds can
 230 High-Power Laser Applications rjh
P. 12

be made at high-power levels. Weld efficiency is denned as the ratio

energy used to melt material to the total energy input. A graphic
has been observed in the welding of the same parts at diiferent power 1
In the welding of low carbon to mild steel, for instance, both 2.5 kW H $
kW were used to achieve approximately 5.5-mm penetration. Parts weld
2.5 kW could not be handled with bare hands due to heat loss from the
zone by conduction. Parts welded at 5 kW, at about twice the speed co in
be handled without discomfort.
Cover gases play an important role in welding operations and laser weldin
is no exception. Some unique problems, however, are associated with cove
gases and their method of application in laser welding. Although not unique
to CW CO2 welding, they will be discussed at this point.
The method of application of the cover gas varies with the type of material
being welded and the quality of the weld required. Figure 12-9 illustrates a
method that uses a trailing cup. Gas is brought directly onto the weld through
the nozzle coaxial with the beam to protect the final optical element and also
is brought into the cup to provide coverage of the bead before and after the
beam to prevent oxidation and/or atmospheric contamination. When full pene-
tration is achieved, it may be necessary to provide cover for the underbead as
well. Provision of an inert cover gas is especially important for highly oxidizable
materials, such as aluminum and titanium.
The most frequently used cover gases are He, Ar, and N2. Argon, because
of its high atomic mass, provides the best coverage or protection against oxidation;
however, it interacts with the metal vapor plasma in such a way that more
laser radiation is lost in the plasma than with He. Maximum penetration depths
are substantially lower for Ar and N2 than He. Helium has a much higher
thermal conductivity than either Ar or N2 and is able to cool the plasma and
reduce the beam interaction. The mechanism for the interaction is free electron
absorption, for the photons at 10.6 jLtm are not sufficiently energetic to cause
ionization or atomic excitation. The ionization potential for He is about 1.3
times that of Ar. This factor, coupled with the higher plasma temperature,
results in a much greater beam interaction for Ar than He.

Trailing cup

Figure 12-9 Cover gas arrange-

ment using a trailing cup.
 12-2 Welding 231
S6C-

It is also likely that the molten pool at the surface is minimized due to
the high thermal conductivity of He, thereby allowing better penetration of
the beam into the keyhole. It has been observed that in high-speed welding,
where maximum penetration is not achieved, deeper penetration is obtained
with N2 than He. It is possible that this situation occurs because the beam is
not incident on the plasma as long and the lower thermal conductivity of N 2
allows more conduction-limited welding, combined with the fact that N2 is
better suited for dispersing the plasma because of its higher molecular mass.
Frequently the cover gas is applied by means of a tube aimed directly at
the weld location if weld bead oxidation is not a serious problem. When welding
materials that spatter vigorously, like rimmed steel or high-carbon steel, a strong
cross flow of air is required to protect the final optical element. This cross
flow must be placed sufficiently far above the weld area to avoid disturbance
of the molten bead and dilution of the cover gas.
Figure 12-10 contains a micrograph of a weld bead cross section for mild
steel welded to low-carbon steel. The high aspect ratio and small HAZ are
evident. Figure 12-11 is a photograph of a part welded by CO2 laser continuous
welding.
Pulse welding. Numerous materials can be pulse welded with lasers.
C02 lasers can be used for pulse welding of small parts—so-called microweld-
ing—but are more frequently applied to spot welding of larger parts. This process
can be accomplished by electronic pulsing or simply opening and closing the
shutter on CW lasers. Pulse lengths of a few tenths of seconds will yield spot
welds several millimeters in diameter and weld depths of a few millimeters
for multikilowatt lasers.

Figure 12-10 Micrograph of laser-welded

Part.
232 High-Power Laser Applications rk
'-nap.

§
Figure 12-11 Co2 laser-welded pan.

The greatest bulk of pulsed laser welding is done with Nd-YAG and
Nd-Glass lasers. Nd-Glass lasers are well suited for low-repetition rate spot
welding on small parts, such as wires to terminals and other electronic compo-
nents.
For high-speed spot welding or seam welding of such items as electronic
packages, Nd-YAG lasers are much more suitable. Figure 12-12 is a photograph
of a commercial Nd-YAG welder and Fig. 12-13 shows a Nd-YAG welded
capacitor. Excellent hermetic welds are made by overlapping pulse welding
with Nd-YAG lasers.

Figure 12-12 Commercial Nd-YAG welder. (Courtesy of Apollo Lasers, an All"

Company.)
Sec. 12-3 Cutting 233

Figure 12-13 Nd-YAG welded capacitor. (Courtesy of Laser Inc.)

Basically the setup for welding with Nd-YAG or Nd-Glass lasers does
not differ greatly from CC>2 laser welding. Cover gases should be used whenever
contamination or oxidation is a problem and a cross flow may be required to
protect the optics.
Calculations for ballpark estimates of pulse times, penetration depth, and/
or power requirements can be made by simple energy balance, but the graphical
technique of Cohen and Epperson is superior. Typically pulse lengths are milli-
seconds with power densities of 105 W/cm2.

12-3 CUTTING

Industrial laser cutting is done with CW or pulsed CO2 and high-repetition-

rate pulsed Nd-YAG lasers. The process is a gas-assist technique in which
gas, under pressure, forces molten material from the kerf. Oxygen is used with
oxidizable materials to increase cutting speed.
The advantages of laser cutting are essentially the same regardless of the
type of laser used and may be the same as for other processing applications.
These advantages include

1. Lack of tool contact

2. Ease of automation
3. Small HAZ
4. Narrow, high-precision kerf
5
- Frequently higher speed than other methods
6- Ability to reach difficult-to-reach areas, including working through glass
'• Cuts low-machinability materials with relative ease

Some disadvantages are

1- High capital investment. In systems where three-dimensional contouring

is required, the system cost may be ten times that of the basic laser.
234 High-Power Laser Applications Chap. 12

2. Cut edges may be tapered and serrated when cutting thick sections
3. The cost of oxygen or other cutting gas is substantial, although not ne
sarily greater than other gas-cutting techniques.

Continuous laser cutting. Generally laser cutting is a gas-assist H

process. It is possible to cut some plastics with the CO2 laser, such as acrvl'
without an assist gas, but it is desirable to use the gas assist to help control
the vapors that are produced. There are nearly as many techniques for laser
gas-assist cutting as there are users. All techniques, however, involve the delivery
of a gas under high pressure to the point of incidence of the laser beam in
such a way that the molten material is blown away. Figure 12-14 is a photograph
of sheet steel being cut by a CW CO2 laser. In this case, a coaxial laminar
flow nozzle is used, which means that the beam and gas jet are coaxial. The
nozzle has a special elliptical cross section leading into a straight throat, which
tends to provide a columnar flow of gas on the part without turbulence. The
space between the nozzle and the workpiece is about 0.5 to 1.0 mm and the
pressure in the nozzle is around 35 kilopascals. In cutting oxidizable materials,
such as steel or titanium, oxygen or air is used to7 enhance the cutting rate
and quality by oxidation. Excessive oxygen, however, can cause self-burning,
which will result in discouragingly poor cuts.
Some cutting systems use supersonic nozzles that tend to provide a focusing
effect for the gas (similar to the exhaust plume from a rocket). Such nozzles
are usually off-axis (not coaxial with the beam), although coaxial systems may
be employed. Some systems use a window between the focusing optic and the

Figure 12-14 Sheet-metal cutting system. (Courtesy of Coherent Industrial Group-

sec. 12-3 Cutting 235

nozzle to keep the high pressure in the supersonic nozzle away from the focusing
lens or mirror.
Steel around 5-mm thick is readily cut with CO2 lasers in the 500 W
range with very fine edge quality and thicknesses of low-carbon steel to 15
m rn have been cut at 2.5 kW. Most metals can be cut with CW CO2 lasers,
along with a wide variety of other materials, including composites, glass, quartz,
plastics, ceramics, paper, and wood. Surprisingly straight-sided kerfs can be
achieved when cutting relatively thick materials as a result of a light-guiding
effect due to multiple reflections from the sides of the kerf. Figure 12-15 illus-
trates this point with 2.5-cm-thick plastic sign letters cut at 1 kW. The same
phenomenon occurs to some extent when cutting other materials. The cross
section of a kerf typically looks something like the sketch in Figure 12-16.
The amount of undercutting and taper is controlled, to some degree, by
the location of the focal point. In metal cutting it is usually best to focus at
the surface, but for thick, nonmetallic materials it is frequently beneficial to
focus below the surface. In the case of the plastic letter in Fig. 12-15, the
focal point was approximately one-fourth of the thickness of the part (0.64
cm) below the surface.
The importance of maintaining good control of gas pressure and flow
rate during cutting cannot be overemphasized. Accurate control of the cutting
nozzle relative to the work is vital. In the case of large sheet material or parts
with z-axis contours, some means of positioning the focusing column, even
for very small variations in z-axis, is required. In complicated contouring four-

P|
gure 12-15 Plastic cut by CO2 laser.
236 High-Power Laser Applications Ch

Laser bean

Figure 12-16 Kerf produced by laser cutting.

or five-axis numerical control with automatic location sensing and feedback

control of the focusing column may be needed. In simple cases where relatively
flat stock is to be cut, a two-axis numerical control system with automatic
position sensing and feedback control on the z-axis may suffice. Feedback control
may be accomplished through capacitive coupling to the part, air jet height
measurement, and mechanical or optical position sensing. The type of sensor
that should be used is a function of the nature of the part's shape, material,
and surface condition. A simple technique that is effective on flat material is
to "float" the focusing head mechanically by means of balls that roll on the
sheet.
It is interesting to note that little, if any, damage to focusing optics due
to spatter or smoke occurs during gas-assist cutting. This is true even when
making blind cuts (intentionally or otherwise) because the molten material is
effectively blown through the kerf or out from under the nozzle in the case o
blind cuts. On the other hand, there may be a significant hazard to the opera
tor(s) from vapors or particulates or both given off during cutting, especia ^
plastics, some of which produce highly toxic substances when heated wi
laser.

Pulsed laser cutting. Basically there are two reasons for using t>u
laser cutting. There may simply not be sufficient power in low average
lasers to melt or vaporize the material without pulsing to achieve nig, ^^
power. For example, a Nd-YAG laser with a CW power of only a few
when rapidly pulsed, can be used to cut through thick film resistor
in thick film circuits to adjust resistance at a rate of several resistors per
Sec 12-3 Cutting 237

Lens

Figure 12-17 Trepanning for cutting small holes.

When cutting refractory materials, such as ceramics, it is frequently desirable

to use pulsed cutting to minimize the HAZ and thereby reduce microcracking
at the cut edge.
A leading application of pulsed laser cutting is the manufacture of aircraft
engine parts. The low-machinability, superalloy materials used in aircraft engines,
such as Waspalloy and Hastelloy-X, must be cut and drilled by EDM, ECM,2
or laser. Both turbine blades and combustor parts are being cut and drilled
by using CO2 and Nd-YAG lasers. A laser in combination with a four- to
six-axis numerical control system is a powerful and flexible tool for applications
involving the cutting of thousands of holes and slots required in aircraft com-
bustor parts. Only a program correction is required to make a design modification
°r a new program to process a different part.
Small holes may be cut by a technique referred to as trepanning, which
is accomplished by bringing the beam off center through a rotating lens as
indicated in Fig. 12-17. Such holes have been cut at angles as great as 15° to
the normal to the surface.

Laser-assisted machining (LAM). Laser-assisted machining is a

Process whereby a laser beam is focused on the workpiece ahead of a tool to
Preheat the workpiece and thus increase machinability of superalloys and ceramic
Materials. The tool force required for cutting is reduced and smoother cuts
a
nd higher cutting speed are possible, along with reduced tool wear. An optimum
2
EDM is electrical discharge machining; ECM is electrochemical machining.
238 High-Power Laser Applications /»i.
<~nap. 1 2

cutting speed must be determined for a given laser power level. Hard
6
materials may harden before the tool arrives at the irradiated spot if the
is not high enough. Hardened materials are generally easier to machine '\
W
LAM than without it.

12-4 LASER MARKING

Laser marking is a process whereby serial numbers or other identification, includ

ing logos, is placed on parts by evaporating a small amount of material with
a pulsed CO2 or Nd-YAG laser.
The advantages of laser marking are

1. High-speed, easily automated method.

2. No mechanical contact with the part,
3. Can be done through transparent enclosures and in otherwise inaccessible
areas. /-
4. It is trivial to change patterns or index numbers.
5. Can be tied in with computer inventory control systems.

The major disadvantage is the relatively high capital investment, but it may
be offset by improved inventory control and quality control.
A 10-W, average power Nd-YAG laser 2-switched by an acousto-optical
modulator is capable of marking any material not transparent to the radiation.
Most modern systems average over 50 W. Even transparent materials, such
as acrylic, can be marked if a suitable absorbing material is placed in or on
the part. The anodization can be cleanly removed from anodized aluminum,
for example, leaving bright, undamaged aluminum with 8 W average power
and a pulse rate of 1000 HZ. The marking or engraving speed is about 5 cm/s
with the beam defocused close to 6 mm to widen the lines. Figure 12-18 is a
photograph of laser-engraved, blue-anodized aluminum.
Rapid marking over a field of several centimeters can be accomplished
with computer-controlled galvanometric mirrors. Incorporating part movemen
in the system makes field size essentially unlimited. Figure 12-19 is a photograp
of a Nd-YAG laser marking system. ,
CO2 lasers can be used for marking many materials transparent to •
/o,m radiation and a few watts of average power is all that is required. ru
may be used, but it must be realized that electronically pulsing a C vv ^
laser will produce powers of three to seven or eight times the CW power
Certain types of low-power CO2 lasers, not discussed extensively in this °^
produce very high peak powers and are applicable in marking and sma
piercing applications. lsed
One economical, fast, but less versatile technique for marking witn v ^ ^
, oi-.r.maoh The- nattern to be engrave
iggc. 12-5 Hole Piercing 239

Figure 12-18 Laser-engraved anodized

aluminum.

defined in a reflective masking material, such as copper, and the entire mask
is irradiated and an image is optically produced on the part. If deep engraving
is to be done, as in laser wood engraving, the beam can be scanned across a
full size mask in a raster scan pattern.

12-5 HOLE PIERCING

Laser hole piercing, frequently referred to as drilling for short, is chiefly done
with pulsed CO2 and Nd-YAG, although ruby and Nd-Glass lasers are some-

p
'gure 12-19 Nd-YAG laser marking
Astern. (Courtesy of General Photonics
240 High-Power Laser Applications
Chap. 12

times applied. The applications for laser hole piercing are incredibly v '
ranging from piercing holes in cigarette filter paper, baby bottle nipples
aerosol can nozzles to diamonds and turbine blades for aircraft engines 'H
piercing rates vary from millions of holes per minute for paper to several seco d
per hole in turbine blades.
The advantages of hole piercing with lasers are essentially the same
for cutting or welding. It is a technique in which highly reproducible hoi
can be made rapidly—frequently in materials of low machinability with no
tool contact.
Yet some important disadvantages to laser drilling should be considered
Laser-drilled holes do not have straight, smooth walls. A sketch of a typical
laser drilled in a metal hole is given in Fig. 12-20, In metals dross may be
attached to the underlip of the hole and recast metal may protrude at the
top. The material is rarely completely vaporized in metal hole drilling. Instead,
some metal is vaporized and up to 90% is melted and percussively removed
from the hole. When a gas assist is used, the molten material is blown from
the hole by the gas assist. Molten metal tends to floyi' up the walls under the
influence of the vapor and some resolidifies on the rim of the hole. In addition,
molten material may not be completely ejected from the bottom of the hole,
thus resulting in dross. The dross is usually easily removed by shot blasting
or light grinding, but the recast material on the top is hard and an integral
part of the remaining metal. Proper techniques can minimize or eliminate both
dross and recast metal on the top surface.
Hole taper is an unavoidable problem associated with beam focusing in
fairly high aspect ratio holes. It, too, can be minimized by proper focusing.
In multiple shot drilling the focal point can be adjusted between shots to control
hole shape.
Laser hole piercing in metals has its greatest use in applications in which
numerous holes must be precisely located in a highly reproducible manner.
The holes routinely placed in turbine blades and turbine combustor parts for

Recast material

Dross

Figure 12-20 Sketch of laser-drilled hole.

I
r sec 12-5 Hole Piercing 241

airflow control are an excellent example of this type of application, as can be

seen in Fig. 12-21.
Estimates of hole-piercing parameters can be made on the basis of simple
energy balance and work quite well for vaporization drilling. The pulse time
required to drill a given thickness hole can be estimated by the technique outlined
in Chapter 11 for a propagating vapor front. As noted, however, the development
of a metal vapor plume will seriously attenuate the laser power reaching the
workpiece. Consequently, pulses must be kept short enough to allow the develop-
ing plume to clear the hole. A gas assist or cross flow or both will help reduce
this problem. Typical pulse lengths are tenths of milliseconds at power densities
on the order tens to hundreds of megawatts.
Holes are routinely laser drilled in ceramic materials. For electrical connec-
tions, holes are drilled completely through the boards. Ceramic sheets are scribed
by laser drilling a row of blind holes and then mechanically snapping the sheet
into two or more parts. Both pulsed CO2 and Nd-YAG can be used for ceramic
drilling, although CO2 is faster. Pulses should be kept short to maximize vapori-
zation and decrease heat loss by conduction, which produces thermal stresses
and hence cracking. Also, rapid pulses minimize the amount of resolidified
material in and around the hole, which reduces hole quality. Typically pulse
lengths are one to a few milliseconds with power levels of 50 W for CO2 lasers
and pulse energy of a few joules for Nd-YAG lasers.
It has been found that the nature of Nd-YAG laser pulses can seriously
affect hole quality [Roos 1980]. Normal relaxation pulses, which consist of a
series of initial spikes followed by continuous output, can result in clogged
holes due to the material melted by the continuous portion of the pulse. A
train of spikes produced by loss modulation provides superior holes in thin
materials drilled at relatively low energy per pulse (200 ml or less per pulse).
Proper application of a gas assist also minimizes this problem.
I
I

I (a)

Figure 12-21
(b)

Laser-drilled holes. (Courtesy of Laser Inc.)

242 High-Power Laser Applications Chap

Other materials that are drilled with lasers are rubber and many
drilled with CO2 lasers and diamonds and other gemstones drilled with Nd
YAG lasers.

12-6 ALLOYING AND CLADDING

Two techniques that show great promise for metal surface modification to im-
prove wear, abrasion, corrosion, impact resistance, and strength while retaining
desirable substrate metal properties are laser alloying and cladding. In both
cases, an alloy material, usually a powder, is placed on a metallic surface and
melted with a CO2 laser beam. In alloying a substantial portion of the substrate
material is melted and mixed with the melted powder, thus producing a true
alloy layer on the surface after solidification. In cladding only enough of the
substrate is melted to produce a metallurgical bond between the alloy material
and the substrate surface. This process is also called hardfacing. The technique
is similar to heat treatment in that a defocused or integrated CO2 laser beam
is traversed over the region to be melted. /
Because the spot size is large (—1.0 cm) and melt depths are only on
the order of a millimeter, reasonable calculations can be made by using the
uniform irradiance model or any of the more sophisticated models used in
heat treatment analysis; however, the heat effusion should be taken into account.
Some materials used in this work are cobalt or nickel-based chromium
carbides and tungsten carbide nickel alloys. Methods for applying the powders
prior to laser melting include powder-solvent slurries, mixing the powder with
an organic binder, and plasma spraying. The biggest problem involves applying
the powder with the appropriate thickness and applying it consistently. Mechani-
cal scrapers can be used to remove excess powder to a given thickness. An
automated spraying technique may also be used. In some methods the powder
is sprayed through a small orifice and simultaneously melted onto the surface
by the laser beam.
The advantages of laser cladding and alloying are similar to those found
in laser heat treatment and surface melting with the additional advantages of
reduced alloy material consumption and greater control and uniformity of me
depth. The dilution of the cladding by base metal is negligible in cladding. In
alloying extremely good mixing of the alloying material and substrate me a
occurs due to a vigorous mixing action that takes place in the melt pool.
Surface tension and thermal gradient effects produce strong convection
currents in the melt puddle. These currents rise at the point of incidence o
the beam and flow outwards toward the walls and down in a circulatory
This mixing does cause rippling in the solidified material, both in the
of the track and transverse to it. Generally the region of overlap between a
tracks will not be flat either. So if a smooth, flat surface is required, a
finishing operation is required.
Chap. 12 Bibliography 243

12-7 MISCELLANEOUS APPLICATIONS

There are many other material processing applications of lasers. Some notable
ones are soldering and desoldering with CC>2 and Nd-YAG lasers, wire insulation
stripping with CC>2 lasers, and cutting of silicon for solar cells with Nd-YAG
lasers.

PROBLEMS

12-1. Describe, with the aid of sketches, the process of laser surface hardening. Explain
what the important parameters are and discuss the phenomena of hardening,
quenching, and similar processes.
12-2. Describe the process of laser welding. Explain the use of cover gases and other
important parameters. Briefly discuss both CW and pulse welding with respect
to the type of laser used and the power ranges involved. Explain what keyholing
is, what causes it, and point out what power range it occurs in.
12-3. Describe and discuss pulsed and CW laser cutting and point out what lasers
are used for various types of materials.
12-4. Discuss and describe hole piercing with lasers.
12-5. Discuss and describe the processes of alloying and cladding with lasers.
12-6. Discuss the advantages and disadvantages of using lasers in the processes men-
tioned in Problems 12-1 through 12-5.

REFERENCE

Roos, SVEN-OLOV, "Laser Drilling with Different Pulse Shapes," J. Appl. Phys., Septem-
ber 1980, pp. 5061-5063.

BIBLIOGRAPHY

12-1. Charschan, S. S. (ed.), Lasers in Industry. New York: Van Nostrand Reinhold
Co., 1972.
12-2. Ready, J. F., Industrial Applications of Lasers. New York: Academic Press, 1978.
12-3. Duley, W. W., CO2 Lasers: Effects and Applications. New York: Academic Press,
1976.
12-4. Ferris, S. D., H. J. Leamy, and J. M. Poate (eds.), "Laser-Solid Interactions
and Laser Processing—1978," AIP Conference Proceedings No. 50. New York:
American Institute of Physics, 1979.
12-5. Ready, J. F. (ed.), Lasers in Modern Industry. Dearborn, MI: Society of Manufac-
turing Engineers, 1979.
244 High-Power Laser Applications Chap

12-6. Metzbower, E. A. (ed.), Applications of Lasers in Materials Processing M

6
Park, OH: American Society for Metals, 1979.
12-7. Engineering Staff of Coherent, Lasers; Operation, Equipment, Application
Design. New York: McGraw-Hill Book Co., 1980.
12-8. Ready, J. F. (ed.), "Industrial Applications of High Power Laser Technology "
Proceedings of the Society of Photo-Optical Instrumentation Engineers, Vol 86
Bellingham, WA, SPIE, 1977. ' ' '
12-9. Ready, J. F. (ed.), "Laser Applications in Materials Processing," Proceedings
of the Society of Photo-Optical Instrumentation Engineers, Vol. 198, Bellingham
WA, SPIE, 1980.
12-10. Laser Institute of America, Use of Lasers in Materials Processing Applications
Toledo, OH, LIA.
12-11. Proceedings of the First International Laser Processing Conference. Toledo, OH:
Laser Institute of America, 1981.
12-12. Ream, S. L,, "High-Power Laser Processing," 1981 SME Heat Treating Confer-
ence, MR81-316. Dearborn, MI: Society of Manufacturing Engineers, 1981.
12-13. Gnanamuthu, D. S., "Laser Surface Treatment," Optical Engineering, 19, No.
5, September-October 1980, 783-792. /
12-14. Bolin, S. R., "Bright Spot for Pulsed Lasers," Welding Design and Fabrication,
August 1976, pp. 74-76.
12-15. Miller, F. R., "Advanced Joining Processes," SAMPE Quarterly, October 1976,
pp. 46-54.
12-16. Banas, C. M., "High Power Laser Welding—1978," Optical Engineering, May-
June, 1978, pp. 210-219.
12-17. Bolin, S. R., "Part 2: Laser Welding, Cutting and Drilling," Assembly Engineering,
July 1980, pp. 24-27.
12-18. Benedict, G. F., "Production Laser Cutting of Gas Turbine Components," Society
for Manufacturing Engineers Paper No. MR8Q-851, Dearborn, MI, 1980.
12-19. Engel, S. L., "Laser Cutting of Thin Materials," Society of Manufacturing Engi-
neers Paper No. MR74-960, 1974.
12-20. Terrell, N. E., "Laser Precision Hole Drilling," Manufacturing Engineering, May
1982, pp. 76-78.
12-21. LIA Laser-Material Processing Committee, "Laser Drilling," Electro-Optical Sys-
tems Design, July 1977, pp. 44—48.
12-22. Bolin, S., "Laser Drilling and Cutting," Society of Manufacturing Engineers
Paper No. MR81-365, 1981.
12-23. Jau, B. M., S. M. Copley, and M. Bass, "Laser Assisted Machining," Society
of Manufacturing Engineers Paper No. MR80-846, 1980.
12-24. Jacobson, D. C., et al., "Analysis of Laser Alloyed Surfaces," IEEE Transactions
on Nuclear Science, NS-28, No. 2, April 1981, 1828-1830.
12-25. Bass, M. (ed.), Laser Materials Processing. New York: North-Holland Publishing
Co., 1983.
 Index

Abbe sin condition, 146 Carter, W., I l l , 119

Absorption coefficient (a), 201, 214 Cast iron, pearlitic, 222
Acousto-optic modulator, 27-30, 238 Cavity dumping, 102
Aerowindow, 142 Cavity lifetime, 101
Alignment, 153-54 Characteristic curve, diode, 40
Alloying, 242 Charge-coupled device (CCD), 71-73
AlrGai-zAs, 133-34 Charge injection devices (CID), 73
Amplification of light, 88 Charschan, S. S., 206
Angular frequency, 163 Cladding (hardfacing), 242
Arata, Y., 218 Cohen, M. I., 209-10
Argon laser, 127 Coherence:
Avalanche photodiode, 41, 66 spatial, 19, 86
Average, time, of a function, 164 temporal, 19-20, 86
Axicon, 10 Coherence length, 20
Coherence time, 19-20
Backlighting, 148 Cold cracking, 229
Beam expander, 14 Complex amplitude, 164-66
Beer's law, 201, 214 Complex impedance, 4
Bessel function, 192-93, 215 Complex number, 163
Birefringence, 23 Complimentary error function, erfc (s), 205
Blackbody radiation, 4 Compton scattering, 4
B°hr-Sommerfeld Theory, 32 Covalent bonding, 35
Bohr Theory, 32 Cover gases (Ar, He, N2), 230-31
Bolometer, 59 Critical pump rate, 95
Boltzmann equation, 85
Boron, 37 Darlington phototransistor, 67
jagg's law, 27 DeBroglie, L., 4-5
wewster's law, 7 Depth of focus for a laser:
window, 22, 106 Gaussian, 110
Hermite-Gaussian, 115
Calorimeter, 57 Detectivity (D), 49, 53
rat's nest, 59 per unit frequency (D*), 49, 53
^andela, 50
c
Diamond, 36-37
arbon dioxide (CO2) laser, 81, 93, 114, Dielectric waveguide, 25
136-43, 222, 229, 234-35, 238-39 Diffraction, 14-19
Carslaw, H. S., 206 circular aperture, 18
246
Index

Diffraction (cont.) Heat-affected zone (HAZ), 229, 231

knife edge, 17 Heat flux, 203
single slit, 18 Heat transfer equations, 202-3
Diode array, 70, 158-60 Heat treatment, 222
Diode laser, 133 Helium-cadmium (He-Cd) laser, 128
Doppler effect, 28, 99, 138 Helium-neon (HeNe) laser, 51, 80 93 infi
Doppler line broadening (inhomogeneous), 125 124-37 ' '
Doppler shift, 176-77 Helium-selenium (He-Se) laser, 128
Double heterojunction, 133 Hermite-Gaussian modes, 111, 121
Doughnut mode (TEMf0), 96, 124 Hermite polynomials, 108-9
Drude free electron theory, 4 Hole burning, 125
Duley, W. W., 206, 215 Hologram, 178-79
Dye laser, 82, 128 Holographic:
double-exposure, 188-92
Effective Fresnel number, 120 image, 182-84
Einstein, A., 4-5 interferometry, 185-95
Einstein probability coefficients, 86-87 real, 184
Electrochemical machining (ECM), 237 real-time, 185-87
Electrical discharge machining (EDM), 237 reconstruction, 182-85
Electromagnetic radiation, 1 recording, 179-82
Electron-hole pair (EHP), 36 time-average, 192-95
Emde, F., 205 virtual, 183-84
Energy balance, 203 Homogeneous line broadening (Lorentzian) 98-
Energy bands, 35-36 99
Epperson, J. P., 209-10 Homojunction, 133
Error function, erf (s), 205 Hot cracking, 229
Excimer laser, 128 Huygen's principle:
Exposure, 181-82 wavelets, 16
Exposure time, 164, 181
Illuminance, 49
Fabry-Perot interferometer, 116 Inhomogeneous line broadening (Gaussian), 98
Far-field, 108 Interference, 14-19
divergence angle, 110, 114 Interferometers:
Fiber optic, 25-27 Fizeau, 172-74
Finite difference, 216, 229 Michelson, 20-21, 169-71
Finite element analysis (FEM), 216, 229 two-frequency, 174—78
Fizeau interferometer, 172-74 Twyman-Green, 171
Focal length: Intrinsic impedance, 3
spherical lens, 10 Irradiance, 3, 49, 163-64
spherical mirror, 8
Focal planes, 159 Jaeger, J. C., 206
Focal spot size for a laser: Jahnke, E., 205
Gaussian, 110 Jenkins, F. A., 8
Hermite-Gaussian, 114
Frequency response, detectors, 54-55 Kerr cell, 23-24, 102
Fresnel formulas, 7-8 Keyhole, 209-11, 214, 229
Fresnel number, 118 Krypton laser, 127
effective, 120
Fringes: Laguerre-Gaussian modes, 115
in holography: Lambertain surface, 52
double-exposure, 190-91, 197 Lambert's law, 201, 214
real-time, 186-89, 197 Laminar flow nozzle, 234
time-average, 193-94, 197 Large optical cavity (LOG), 133
interference, 168-75 Laser, acronym, 78 _
Laser-assisted machining (LAM), 222, H
Gabor, D., 178 Laser beam:
Gain coefficient (/3), 88-90 collimated, 184
threshold value, 91 conjugate, 184
Galium-arsenide (GaAs), 38, 133-34 object, 179-82
Gas assisted cutting, 234 reconstruction, 182-84
Gauges: reference, 179-82
diode-array camera, 158-60 Laser beam propagation:
laser scanner, 148-50 higher-order mode, 111-15
Gaussian beam, 110-11, 120, 206, 217 lowest-order mode, 108-10
Gick, A. E. F., 215, 229 Laser resonator, 96, 116
Graded index fiber, 26 Laser surface hardening, 222
Index 247

Laser types: Obliquity factor, 164-66

argon, 127 Operational amplifier (op amp), 46-48, 75
diode, 133 Optical activity, 25
dyelaser, 128 Optical axis, 23
CO2, 81, 93, 114, 136-43, 222, 229, 234-35, Optical resonator, 81, 96
238-39 Optical stability, 116
HeNe, 51, 80, 93, 106, 124-27 Original phase factor, 164-66
metal vapor, 128
Nd-Glass, 81, 131, 229, 232-33, 239 Patel, C. K. N., 136
Nd-YAG, 81, 93, 131, 229, 232-33, 236-39 Pauli exclusion principle, 34-35
ruby, 81, 129-31, 239 Penning ionization, 128
Latent heat: Penta-prism, 154
fusion, 201 Permeability, 3
vaporization, 201 free space, 3
Lateral-effect diode 69-70 relative, 3
Law of mass action, 37 Permittivity, 3
Leith, E. N., 178 free space, 3
Lens equations: relative, 3
Newtonian, 159-60 Phonon, 99
thick, 11-12 Phosphorus, 37
thin, 11 Photoconductor, 56, 62, 75
Light emitting diode (LED), 51 Photodiode, 62-66
Linear polarization, 2, 21, 106-7, 163 array, 70, 158-60
Lineshape factor, g(f), 100 Photoelectric effect, 4
Lithium niobate, 23, 60 Photographic emulsion, 181-82
Lumen, 49 . , Photometry, 49
Photomultiplier tube (PMT), 56, 67-68
Magnification: Photon, 4, 5, 86
lens, 11, 160 Photon drag detector, 67
mirror, 9 Phototransistor, 56, 66-67
Maiman, T., 129 Planar-diffused diode, 64
Planck, M., 4
Malus' law, 23
and blackbody radiation, 87
Manganese phosphate (Lubrite), 225 Plane of incidence, 6
Marou, H., 218 Plane wave, 163
Martensitic transformation, 223, 225 pn junctions, 38
Maser, 78 diode, 39^1
Maxwell's equations, 2 Pockels cell, 23-24, 102
Maxwell's velocity distribution, 99 Polarization:
Metal vapor laser, 128 circular, 21, 196
Michelson interferometer, 20-21, 169-72 elliptical, 21-22
Michelson-Morley experiment" 3, 20 linear, 2, 21, 106-7, 163
Microwaves, 2 random, 107
Mirror: Polaroid, 22
concave, 8, 148 Polyvinylfluoride (PVF), 60
convex, 8 Population inversion, 80, 84, 89
multisided, 148 Position detectors:
off-axis parabolic, 9, 148 bicell, 68
parabolic, 148, 161 lateral effect, 48-69
Miyamoto, I., 218 photodiode array, 70, 157-60
Mobility, of electrons and holes, 62 quadrant, 68, 153
Modal delay, 26 Potassium-dihydrogen-phosphate (KDP,
Mode hopping, 92 KD*P), 23
Mode locking, 102-3, 129 Powell, R. L., 178
Mode volume, 94 Poynting's theorem, 2
Monochromaticity, 79 Principal plane, 110
Propagation constant, 163
Neodymium glass (Nd-Glass) laser, 81,131, 229, Propagation factor, 164—66
232-33, 239 Propagation vector, 164
Neodymium YAG (Nd-YAG) laser, 81, 93, 131, Pyroelectric detector, 60-61
229, 232-33, 236-39
Nitrobenzene, 23 Quadrant detector, 68, 153
Noise equivalent power (NEP), 49, 53 Quantum numbers:
Noise, in photodetectors, 53-54 magnetic, 33-34
Numerical aperture (NA), 26 orbital, 33-34
Numerical methods, 216, 229 principal, 33-34
248 Index

Quantum numbers (cont.) Thermal time constant, 208

spin, 33-34 Thermopile, 59
Quarter wave plate, 24 Thick lens, 11-12
Q-switching, 101, 130-31, 238 Time average of a function, 164
value, 101 Total internal reflection (TIR), 6
Transistor, 42
Radiometry, 49 bipolar junction (BJT), 42^*4
Radio waves, 2 junction field effect (JFET), 42, 45^6
Rayleigh range, 110, 120-21 metal-oxide-semiconductor (MOS), 42, 46
Ready, J. D., 206, 213 Transverse electromagnetic modes (TEM(B) 96-
Reflectance, 4, 7, 200 97, 106-8, 139
Reflection: Triangulation, optical, 154-57
diffuse, 6, 150-51 Triglycine sulfate (TGS), 60
specular, 6, 150-51 Twyman-Green interferometer, 171-73
Refractive index, 3
conductors, 201 Uncertainty relations, 5, 19, 98
Responsivity (R), 49, 53 Unstable resonator, 119-20
Ruby laser, 81, 107, 129-31, 239 Upatnieks, J., 178

Saturable absorber, 103 Velocimetry, 167-69

Scan angle, 147 Visibility, 15
Scanning, optical, 145-52
Schottky diode, 44, 64 Waist location, 118
Schrodinger, E., 32-33 Waist size:
Shadow projection, 148-50 Gaussian, 118
Hermite-Gaussian, 119
Silicon, 35-37 Wave:
Skin depth, 4 amplitude, 163^66
Snell's law, 6 complex amplitude, 164-66
Solid angle, 50 obliquity factor, 164—66
sterandian, 50 original phase factor, 164—66
Spawr Optical Company, 214 plane, 163-66
Specific heat, 201 propagation constant, 163
Spherical aberration, 9, 13, 148 propagation factor, 163-66
Spiking, 201 propagation vector, 164
Step index fiber, 26 spherical, 164—66
Stetson, K. A., 178 Wavelength dispersion, 27
Stimulated emission, 85 Waxicon, 10
Stripe geometry, 133 White, H. E., 8
Supersonic nozzle, 234 Work function, 67
Surface inspection, optical, 150-53
Swift-Hook, D. T., 215, 229 Xenon laser, 127
Thermal conductivity, 201 Young's double slit interference, 16
Thermal diffusivity, 201
Thermal gradient, 203 Zeeman splitting, 174-76
Thermal penetration depth, 204, 214 Zinc phosphate, 225
 INDUSTRIAL

and their applications

James T. Luxon/David E. Parker

This is a flexible introduction to lasers and their industrial applications.

Devices such as photodetectors and modulators are covered, and the
concepts of basic optics pertinent to lasers and their application are
included. The book covers some laser theory for a working understanding
and to dispel the mysticism surrounding the device.

Among its many features, the book:

covers the topic of laser beam optics including propagation, focusing, and
depth of focus in detail for both Gaussian and higher-order mode beams,
includes a chapter on optical detectors, including detector arrays, pre-
ceded by a chapter on semiconductors.
' discusses radiometry, photometry, and optical device parameters,
presents pertinent cases of the interaction.of high-power laser beams and
matter.
devotes separate chapters to industrial applications of low-power and
high-power lasers.
• details specific types of applications with additional theoretical or con-
ceptual material where needed.

PRENTICE-HALL, INC., Englewood Cliffs, N.J. 07632