Professional Documents
Culture Documents
Ebook - Industrial Laser and Their Application
Ebook - Industrial Laser and Their Application
1 0 9 8 7 6 5 4 3 2 1
PREFACE VII
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
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.
Wavelength (meters)
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
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)
= (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
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)
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)
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
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.
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
• 'dent power is reflected. The critical angle is deduced from Snell's law to
be
2
1,
9aC = sin""
•
— (1-15)
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.
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
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
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.
V (a)
(b)
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
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)
Ge 0.0087 0.0295
ZnSe 0.0187 0.0286
GaAs 0.0114 0.0289
14 Principles of Optics Chap.
(a)
'4s
(b)
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.
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)
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
ID 4
/-*-
(a) (b)
Figure 1-19 Knife edge experiment with laser beam uniformly illuminating the knife
edge.
Plane
w
avefront
(1-34)
1_6 COHERENCE
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)
,Beam splitter
Y7 Viewer or detector
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>
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
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-°
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
/
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-
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
1-4. 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.
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
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
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
2-3 pn JUNCTIONS
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
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
„ 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.
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,
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)
Collector
Epitaxial layer
n+ buried layer
Figure 2-12 BIT used as a switch. Figure 2-13 BJT used as an inverter.
(a) (b)
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
T
'ty/fifM '//%t// y//////
^^?
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.
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
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
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
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.
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.
=
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.
Resistive heater
TCs or thermisters
Insulation
sa
ceramic resistor with a large temperature coefficient of resistance.
58 Light Detection Devices
Cha
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.
•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.
Irradiated /
junctions
t M t t t -Electric
dipoles
t t t t t I
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
<*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
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)
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)
T
Al
(b)
'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.
,-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
Window
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)
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.
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.
*- 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,
SiQ, Al
Polysilicon
Figure 4-16 CCD linear imagers. (Adapted with permission form Solid State Tech-
nology, Technical Publishing, a company of Dun & Bradstreet.)
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.
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 *-*1
photod,ode operat.on.
•Ft*,
Detection
which leads to
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
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,
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
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)
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
PROBLEMS
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
Laser Theory
Inverted population
A/, 5,
(6-1)
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
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'
6-4 AMPLIFICATION
Initial photon
A Az = (6-9)
. . "J
In the limit as Az — * 0 we then have
^z=(N2-NjpfB2ihf (6-10)
(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
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
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)
rwv - E,
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
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.
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)
(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
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
(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
• 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
Techniques for modifying the laser output are discussed here, specifically, tech-
niques for producing, short, high peak power pulses.
££*« 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
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
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.
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.
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 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.
'
Brewster windows
'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
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.
, 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>
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--
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
•
times the standard deviation of the function
0.2
0.2
-8.0
7_2 Beam Propagation for Stable Laser nesonaiuia
Sec.
0.4 -
0.2 -
-8.0
= 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)
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)
\) —
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
la
> Confocal ( b ) Fabry-Perot (c) Concentric (d) Spherical-flat
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.
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
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)
#„ = - — ( m - 1 ) (7-32)
For m near 1 Neci < N and excellent mode discrimination occurs for all half-
integer values of Nen.
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
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
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.
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.
3.39 jim
0.6328 urn
1.15 /i
1s
He Ne
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- »-*
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.
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
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.
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
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
-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
Q-switch
Harmonic generator
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.
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
Substrate1 GaAs
-Contact
: 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 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.
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
Asymmetric
o o o o o o o o o i f c ; oooooooooo^l Q . )
Bending
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
Beam
0.05 kW/m
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.)
(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
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
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
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
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
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.
Honed surface
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
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
Vertical position
indicator
Quadrant detector
Horizontal position
indicator
Laser
Penta-prism
Detector
- Beam expander
Laser
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.
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
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
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
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
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
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
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
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.
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
dt
'-
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
Factor Equation
Amplitude Pfo
Obliquity 1
Original phase exp (/0 0 )
Propagation exp (/k • r)
Factor Equation
A_
Amplitude
r
Obliquity cos (i//)
Original phase exp (/00)
Propagation exp (±/*r)
Factor Equation
Amplitude if0
/
v\ * z
Obliquity
Original phase
Propagation
1
exp (/00)
exp (jkz)
Factor Equation
Amplitude p£"0
Obliquity ' 1
Original phase exp (/<J0)
Propagation exp (jkz0) exp // jfy)
Factor Equation
A
Amplitude
*- x Obliquity
Propagation exp
_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
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)
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)
Particle
Mirror Lens
Photomultiplier
Analyzer
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
Mirror
Laser with
beam expandei Mirror
^Screen
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
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)
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.
or Az -(„+{)' (10-23)
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)
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)
\ *=77 0<«8)
Reference
surface Test mirror
Beam
splitter
Screen
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
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 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
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
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
Object
Photographic
plate
(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
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 '
(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
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
Emulsion Hologram
Observer
U. + U.
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)
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
Az(x, y) = m- (10-65)
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
Hologram
Object beam
Displaced position
- Original position
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)
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
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.
1.0 i i T i i n ^
0.8
0.6
0.4
0.2
0 12 16 20 24
Stationary
region
m=0
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) =
Hologram
Rod
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
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
p
>gure 10-33 Photographs for Problems
10
-13 and 10-14.
198 Low-Power Laser Applications II: Interferometry and Holography Chap
10
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
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.
Material and laser parameters that affect laser processing are discussed in this
section.
Material parameters
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
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.
(11-6)
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. *
Laser beam
Workpiece
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
/4Kt
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.
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.
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.
P is the total power absorbed at the surface and ierfc stands for the
integral of the complementary error function and
z* + a*)"*-z] (11-16)
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.
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.)
'•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.
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
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
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
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].
(11-29)
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.
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
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
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.)
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
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
0.2
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.
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- '*
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
Figure 12-4 Micrographs of heat treat tracks. (Courtesy of Saginaw Steering Gear,
Division of GMC.)
(a) (b)
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.
SAE 1045
Power = 2.5 kW
0.4 -
0.2 -
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
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
Trailing cup
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 12-11 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.
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
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.
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
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
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.
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
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.
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
I (a)
Figure 12-21
(b)
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.
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
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