OPTICAL

THIN FILMS
USER HANDBOOK
OPTICAL
THIN FILMS
USER HANDBOOK

James D. Rancourt
Optical Coating Laboratory, Inc.
Santa Rosa, California

SPIE OPTICAL ENGINEERING PRESS

A Publication of SPIE-The International Society for Optical Engineering

Bellingham. Washington USA
Library of Congress Cataloging-in-Publication Data

Rancourt, James D.
Optical thin films: user handbook / James D. Rancourt.
p. cm.
Includes bibliographical references and index.
ISBN 0-8194-2285-1 (hardcover)
1. Optical coatings. 2. Thin films-Optical properties.
3. Optical films. 4. Light filters. I. Title.
TSS17.2.R36 1996
621. 381S'2-dc20 96-28046
ClP

Published by
SPIE-The International Society for Optical Engineering
PO Box 10
Bellingham, Washington 98227-0010
Phone: 360/676-3290 Fax: 360/647-1445
E-mail: [email protected]
WWW: http://www.spie.org/

Copyright © 1996 Society of Photo-Optical Instrumentation Engineers.

Originally published in 1987 by Macmillan Publishing Company, New York.

All rights reserved. No part of this publication may be reproduced or distributed

in any form or by any means without written permission of the publisher.

Printed in the United States of America.

 Contents

PREFACE xi

1 INTRODUCTION TO OPTICAL THIN FILMS 1

1.1. Interference of Light 2
1.1.1. Coherent Addition of Amplitudes 10
1.1.2. Layer Thickness Specifications 11
1.1.3. Nomenclature 12
1.1.4. Index of Refraction and Dispersion 13
l.l.5. Absorbing and Nonabsorbing Filters 13
1.2. Spectral Ranges and Materials Section 15
1.3. Angle of Incidence Effects 16
1.3.1. Snell's Law 16
1.3.2. Optical Invariant 17
1.3.3. Critical Angle 17
1.3.4. Polarization 18
1.3.5. Brewster's Angle 20
1.3.6. Effective Indices 20

v
vi CONTENTS

1.3.7. Birefringence and Differential Phase Shift Coatings 21

1.3.8. Steep Surfaces 23

2 DEPOSITION TECHNOLOGY 24
2.1. Coating Processes 24
2.1.1. Vacuum Evaporation 24
2.1.2. Sputtering 26
2.1.3. Flash Evaporation 30
2.1.4. Nonvacuum Processes 30
2.1.4.1. Organometallic Solutions 30
2.1.4.2. Electroless and Electroplating Solutions 31
2.1.5. Process Conditions 31
2.1.6. Coating Time 32
2.2. Coating Parameters 33
2.2.1. Source Characteristics 33
2.2.1.1. Compositional Variations 33
2.2.1.2. Stoichiometry 33
2.2.1.3. Coating Distribution 34
2.2.2. Deposition Environment 34
2.2.3. Nucleation and Growth 35
2.2.4. Rate and Thickness Monitoring 35
2.2.4.1. Crystal Oscillator 35
2.2.4.2. Optical Techniques 36
2.2.4.3. Atomic Emission Techniques 38
2.2.5. Sticking Coefficient 38
2.3. Coating Equipment 39
2.3.1. Size of Coating Chamber 39
2.3.2. Chamber Geometry and Process Considerations 40
2.3.2.1. Surface to Volume Ratio 40
2.3.2.2. Thickness Uniformity and Coating
Efficiency 41
2.3.3. Substrate Rotation for Coating Uniformity 43
2.3.4. Jigging and Substrate Support 43
2.3.5. Temperature Related Substrate Degradation 44
2.4. Stress Effects 45
2.5. Substrate Cleaning 48
2.5.1. Contacting Cleaning 49
2.5.2. Ultrasonic Cleaning 50
2.5.3. High Pressure Sprays 50
CONTENTS vii

2.5.4. Drying 50
2.5.5. In Situ Processes 52

3 PERFORMANCE OF OPTICAL FILTERS 53

3.1. Reflectance, Transmittance, and Absorption 53
3.2. Absorption and Scattering 54
3.3. Blocking Ranges 56
3.4. Filters in Series: Incoherent Addition of Intensities 60
3.5. Standing Wave Electric Fields 63
3.6. Perfonnance Shifts with Angle of Incidence 68
3.7. Other Parameters that Affect Perfonnance 71
3.7. 1. Water and Humidity 71
3.7.2. Temperature 73
3.7.3. Corrosion 75
3.8. Wavefront Distortions 75

4 DESIGN TYPES 78
4.1. Antireflection Coatings 78
4.1.1. Single-Layer Designs 82
4.1.2. Narrow-Band AR Designs 83
4.1.3. Wide-Band AR Coatings 85
4.1.4. Superwide Infrared AR Coatings 86
4.2. High Reflection Designs 86
4.2.1. Metallic Reflectors 86
4.2.2. All Dielectric Designs 91
4.2.2.1. Nonsymmetric Stacks 91
4.2.2.2. Symmetric Designs 95
4.2.3. Enhanced Metallic Reflectors 96
4.3. Edge Filters 99
4.4. Bandpass Filters 103
4.4.1. Two Edge Filters in Series 103
4.4.2. Fabry-Perot Filter 104
4.4.2.1. Fabry-Perot Filter 105
4.4.2.2. Solid Fabry-Perot Filters 107
4.4.3. Narrow-Band Dielectric Filters 108
4.4.4. Circular and Linear Variable Filters 110
4.4.5. Reflex or Phase Conjugate Filter 112

viii CONTENTS

4.5. Beam Splitters 115

4.5.1. Wide-Band or Neutral Density Beam Splitters 115
4.5.1.1. Slab Beam Splitter 116
4.5.1.2. Cube Beam Splitters 121
4.5.2. Dichroic and Trichroic Filters 121
4.5.3. Filters for Use in Harmonic Generation Experiments 123
4.6. Other Filters 126
4.6.1. Dark Mirrors, Selective Absorbers, and Induced
Absorbers 126
4.6.2. Solar Absorbers 129
4.6.3. Notch Filters 130
4.6.4. Transparent Conductive Coatings 130
4.6.5. Patterned Dielectric Coatings 132
4.6.6. Coatings for High-Power Laser Beam Reflectors 133
4.6.7. Color Correction Coatings 134
4.6.8. Switching Filters 134
4.6.9. Emissivity Control 135
4.6.9.1. Emissivity Reducing Coating 135
4.6.9.2. Emissivity Enhancing Coatings 135

5 FILTER PERFORMANCE AND MEASUREMENTS 139

5.1. Spectrophotometers and Spectrophotometry 139
5.1.1. Reflectance Measurements 141
5.1.1.1. Single-Beam Spectrophotometers 141
5.1.1.2. Double-Beam Spectrophotometers 141
5.1.2. Beam Offset 145
5.1.3. Extra Surface Effects 145
5.1.4. Effect of the Refractive Index of the Incident
Medium 147
5.1.5. Polarization 148
5.1.6. Measurement Cone 149
5.1.7. Nonspecular Samples 151
5.1.8. Nonplanar Surfaces 152
5.1.9. Sample Heating 153
5.1.10. Measuring Absorption by Calorimetry 153
5.1.11. Colorimetry 155
5.2. Environmental Performance and Measurements 158
5.2.1. Use of Witnesses in Environmental Testing 159
5.2.2. Adhesion Testing with Self-Adheshive Tape 160

CONTENTS ix

5.2.3. Abrasion Testing 160

5.2.3.1. Cheesecloth Test 161
5.2.3.2. Eraser Test 161
5.2.4. Humidity Exposure 161
5.2.4.1. Humidity Testing 162
5.2.4.2. Solubility Testing 162
5.2.5. Salt Fog Exposure 162
5.2.6. Other Tests 162
5.2.6.1. Visual Inspections 163
5.2.6.2. Falling Sand Test 163
5.2.6.3. Other Adhesive Tests 163
5.2.6.4. Combined Adhesion and Abrasion 164
5.2.6.5. Combined Temperature, Humidity, and UV
Light 164
5.2.7. Cycled Temperature Testing 165
5.2.8. Combined (Sequential) Testing 165
5.2.9. When to Test 165
5.2.10. Formal Specifications 166
5.3 Aging 167
5.3.1. Shelf Life 167
5.3.2. Material Interactions and Interdiffusion 168
5.3.2.1. Oxidation and Reduction 168
5.3.2.2. Agglomeration 169
5.3.2.3. Interdiffusion 169
5.3.3. Biological Effects 170
5.4. Cosmetic and Other Visual Deficiencies 170
5.4.1. Sleeks, Digs, and Scratches 170
5.4.2. Blotches and Stains 171
5.4.3. Replication of Substrate Defects 171
5.4.3.1. High Reflectors with Opaque Metal Layer 171
5.4.3.2. Low Reflectance Surfaces 17]
5.4.4. Coating Voids and Other Defects 172
5.4.5. Cleaning Coated Parts 173
5.4.5.1. Hard Substrates-Hard Coatings 174
5.4.5.2. Hard Coatings on Soft Substrates, and Soft
Coatings 175
5.4.6. Coating Thickness Uniformity 176
5.4.7. Microscopic Examination 178
5.5. Inspection Lot Size 181

x CONTENTS

6 SUBSTRA TES AND MATERIALS 182

6.1. Substrates in General 182
6.2. Substrate Types 185
6.2.1. Glasses 185
6.2.2. Plastics and Organic Substrates 188
6.2.3. Metals 189
6.3. Safety Glass 190
6.3.1. Lamination 190
6.3.2. Tempering 190
6.4. Substrate Specifications 191
6.5. Blocking Properties 192
6.6. Reststrahl Bands 193
6.7. Island Formation in Thin Films 194
6.8. Thermal Runaway in Semiconductors 195
6.9. Number of Materials in Multilayer Coatings 195
6.10. Toxicity 196

7 FUTURE PROSPECTS 197
7.1. Ion Beams 197
7.2. Nonvacuum Processes 199
7.3. Inhomogeneous Films 200
7.4. Codeposition 200
7.5. X-Ray Coatings 201
7.6. Equipment Advances-Automatic Coating Machines 202
7.7. Materials Research 203

A HOW TO SPECIFY OPTICAL FILTERS 207

B SELECTED UNITED STATES MILITARY

SPECIFICATIONS THAT RELATE TO THIN FILM
OPTICAL FILTERS 210

C BASIC PROGRAM TO CALCULATE TmN

FILM FILTER PERFORMANCE 271

REFERENCES 281

INDEX 287
 Preface

This book is intended for the user of optical thin film products. Its primary aim is
in the area of commercially available products: their preparation techniques and
their characteristics. It is hoped that this will help users to specify filters, and to
be aware of the limitations they might encounter in using them.
The reader interested in a design textbook is directed to a companion volume
in this series written by Alfred Thelen entitled Design of Optical Interference
Coatings.
The vast majority of the discussions in this text are independent of the method
of deposition used-evaporation, sputtering, ion plating, epitaxial growth, or other
scheme of depositing optical films. The emphasis is on pragmatic information that
one needs when using, specifying, or designing into an optical system an optical
interference filter.

A NOTE ON THE STYLE USED

In writing this book, the ubiquity of the computer-handheld, desktop, and main-
frame-models, has led to the decision to minimize the number of tables of com-

xi
xii PREFACE

puted values in favor of curves and formulas. The curves hopefully give the reader
a feeling of the phenomenon described by an equation. If precise values are required,
they can be easily obtained from the equations.

ACKNOWLEDGMENTS

The support given to me by my employer, O.c.L.I., during the effort of writing

this book, is gratefully acknowledged.
I want to acknowledge the help my colleagues and co-workers gave me in
answering my many questions in a friendly and helpful way, and for the many
suggestions they offered. In particular, I thank Shari Powell Fisher who wrote the
BASIC program included in Appendix C, and Jacques Cote, who drew the figures
and then cheerfully revised them as I changed my mind about them.
lowe a debt of gratitude and many thanks to Joseph Apfel for his very
constructive critique of the manuscript.
Any errors, faults, or omissions in this book are not the fault of any of those
mentioned above; they are solely those of the author.
Finally, I thank my family, and especially my wife, Jeannine, for all of her
help, patience, and understanding during the writing of this book.
 Chapter 1

Introduction to Optical
Thin Films

In this introductory chapter, we discuss the origins of interference phenomena. We

then consider the manner in which these effects are combined to produce practical
interference filters. In this book, the tenns filters and coatings are used inter-
changeably, and mean any type of optical coating deposited on a substrate. For
example, an antireflection coating is a filter as is a narrow bandpass coating design.
The coating may consist of one or more layers, the optical thickness of which is
on the order of a wavelength of light.
We confine our discussion to those coatings that are used in optical devices
and for optical purposes. We do not consider nonoptical uses for coatings such as
decorative, protective, an~ electronic.
Assumptions that we make, unless specifically stated otherwise, are:
1. All materials are nonmagnetic
2. The plane of the coatings is infinite in extent so that diffraction may be
ignored
3. Materials are isotropic and homogeneous
The tenn performance is frequently used throughout the text. This is intended as
a generic tenn for any characteristic perfonnance parameter that pertains to a thin
film coating, such as reflectance, transmission, absorption, or phase.
2 CHAPTER 1

1.1. INTERFERENCE OF LIGHT

The origins of interference lie in the wave nature of light and the superposition of
coherent waves. The incident wave may interfere with one or more of the waves
that are reflected from the interfaces formed by the thin film and the substrate. The
phases and the amplitudes of these waves determine whether the resultant sum of
these waves leads to constructive or destructive interference, and an increase or
decrease in the reflectance or the transmittance of the incident light. The incident
waves may vary in wavelength, and there is, in general, a dependence of the
performance on the wavelength. Coatings with a strong wavelength dependence
give rise to useful designs such as narrow bandpass filters. In other situations, one
desires a wavelength insensitive performance such as may be found in a wide band
antirereflection coating. These different performances are attained by varying the
index of the layers, their thicknesses, the order of the layers, and the total number
of layers in a stack. In addition, the indices of refraction of the incident medium
and of the substrate are important.
Figure 1.1 shows schematically a layered structure of thin films deposited on
a substrate. Each layer typically has an index that is different from the adjacent
ones, although there are some instances in which they may be the same. The optical
thicknesses; i.e., the product of the index of refraction and the physical thickness,
of typical layers used in optical filters is of the order of one-quarter of the wavelength
of light.
Light is reflected from, as well as transmitted through, each of the interfaces.
These transmitted and reflected waves interfere with one another. They can also
undergo multiple reflections, and these are taken into account in analytical treat-
ments. The waves are usually all coherent because they originate from the same
incident wave.
Figure 1.2 shows the effects of two simple cases of interference. In the first,
the two waves add to zero, thus leading to complete destructive interference, whereas
in the second illustration, constructive interference takes place.
The normal incidence amplitude reflectance at an interface between two semi-
infinite, nonabsorbing materials of different indices of refraction (Fig. 1.3) is given
by the Fresnel formula:
(1.1 )
where r is the amplitude reflection coefficient, no is the index of refraction of the
incident medium, and nl is the index of refraction of the second medium.
The transmitted amplitude t is given by 1 - r.
Note that nl is complex when the second medium is absorbing. The incident
medium may not be absorbing because the incident wave is assumed to arrive from
infinity with no intervening surfaces. Any absorption in the incident medium would
attenuate the intensity of the incident wave to zero.
Equation 1.1 is also valid at the interface between two thin films if multiple
reflectances are ignored or accounted for separately. In this case, both media may
be absorbing, and the corresponding indices no and nl are complex, as is the result,
INTRODUCTION TO OPTICAL THIN FILMS 3

8, Reflected Energy (R)

Incident Medium (no)

n3

••
•

••
•

Transmitted Energy (T)

Figure 1.1. Structure of a thin film stack. The index of refraction is indicated by n, and the
layers are numbered starting from the incident medium. Reflections occur at each interface,
and the rays are refracted as they cross each interface between layers.

r. The implication of complex values for no or nl is that there will be some absorption
of the energy in the light beam.
The reflection coefficient generally refers to the ratio of the reflected intensity
to the incident intensity and is given by the complex square of r: R = r . r*. For
nonabsorbing substrates, this is
R == (no - nl)2/(no + nl)2,
4 CHAPTER 1

whereas for absorbing substrates it takes on the form,

R = [(no - nl)2 + ki]/[(no + n1f + ki] (1.2)
where the complex index of refraction of the substrate is given by
N = nl - ik l •
In the literature, the real part of the optical constant is almost universally represented
by the letter "n." Some specialties of physics use the dielectric constant, represented
by the Greek letter E, instead of the refractive index. The relations between the
two are
E = + i E2 = (n - ik)2
EI
n = {[EI + (d + d)0.5]12}0.5
k = - {[ - E1 + (d + EDo. 5]/2}0.5
EI = n - ~
2

E2 = -2nk
The imaginary portion of the complex index of refraction (k) given above has a
variety of names including extinction index (Stone, 1963) and extinction coefficient
(Maclet>d, 1969). It also appears in different forms, including the product n K, also
called extinction coefficient or attenuation index (Born and Wolfe, 1970, p. 613).
In this work, only the form N = n - ik is used. Care should be exercised when
using tabulated data to be certain that it is clear which representation of k a particular
author is using.
A value related to k is the absorption coefficient; it is sometimes represented
by a and sometimes by ~. It'is defined as
a = 4 7r klA
where A is the vacuum wavelength.
It has units of reciprocal length, usually inverse centimeters, and is used to
calculate the attenuation of a beam of light in an absorbing material:
1/10 = exp( - ad),
where I is the emergent intensity, 10 is the incident intensity, and d is the thickness
of the material.
The phase of the reflected wave for nonabsorbing substrate is
o if no> nl
<l> =
{7r if no < nl.
A wave reflected at an air/glass interface undergoes a phase shift of 7r (180°), and
a node in the electric field occurs at the interface. On the other side of the glass
plate, at the glass/air interface, no phase shift occurs for the reflected wave.
For absorbing substrates, the situation is more complex; the phase is given by
_ -I 2 nokl
<l> - tan 2
no - nl2 - •.2
Ilf
iL

Time
+
"=f1 Unit

o Units
Destructive Interference

IL

_0--_-->1
1 UNIT

1 UNIT

-
W'UN'TS

Constructive Interference
Figure 1.2. Constructive and destructive interference of light waves. The amplitudes of
two waves with the same frequency (wavelength) add if the two waves are in phase, or
cancel if they have the same amplitude and are exactly out of phase with one another. In
other cases of unequal amplitudes or inexact phasing, the result is a wave with a different
amplitude from the original two interfering waves. More than two waves may interfere.

Incident medium

Substrate or film n1

Figure 1.3. Schematic diagram of an interface. The angle S, of the reflected ray r is equal
to angle Si of the incident ray. The deviation of the refracted ray is dependent on the indices
of refraction of the media on either side of the interface and on the angle of incidence of the
light ray.
5
6 CHAPTER 1

In this case, there is no node at the boundary. In some situations, however, the
indices of refraction can be such as to allow an approximation of a node to exist
at the surface. Silver is one such material where the complex index of refraction
in the infrared portion of the spectrum is approximately (0.05 - 3.0i). The phase
shift for these values is approximately 178 degrees, which is close to the 180
degrees required for a node. Table 1. I contains some complex index values for
this and other materials. This subject is covered in more detail in the section on
electric field standing waves in Chap. 3.
It should be noted that the wavelength does not explicitly enter into the cal-
culation of phase values at an interface, although it does determine which values
of the refractive indices one should use.
It is straightforward to add layers to the initial interface. The Fresnel formula
applies at each interface, and the phases of the individual waves must be included
in the calculation (Fig. 1.4). For a single layer between two semi-infinite regions,
the formula is

, = '1 + '2 exp( - 2i~)

0.3)
1 + '1'2 exp( - 2i~)

where 'i is the amplitude reflectance at the ith interface as given by Eq. 1.1 and
~ is the phase thickness of the film, and is given by
21T
~ ==

T nrdr,

A = wavelength and nr dr = optical thickness of the thin film.

The phase of the reflected wave is given by

""
'I' = tan '1 (Ii +'2(ri
-1 - 1) sin 2 ~
1) + ,iri + 1) cos 2 ~
When the film thickness goes to zero, Eq. 1.3 reduces to Eq. 1.1, the Fresnel
reflection equation for the interface between the incident medium and the substrate.
Figure 1.5 and 1.6 illustrate the behavior of the reflectance as the thickness

Table 1.1 Typical metal indices and normal incidence phase shifts for
normally incident light in air
Phase shift
Metal n k (degrees) Comments

Silver 0.05 30 171> Infrared

Copper 12 60 178 10 11m
Aluminum 23 86 179 9.5 11m
Nickel 2.8 5 163 0.99 11m
Gold 7.6 72 178 10 11m
Gold 1.4 1.9 140 0.45 11m
Dielectrics any 0 180
Most indices are from Ordat at ai., 1983.
INTRODUCTION TO OPTICAL THIN FILMS 7

Figure 1.4. Schematic diagram of the summation of the partial reflectances to get the total
reflectance from two interfaces. If the two interfaces are sufficiently separated so that the
optical path between the surfaces exceeds the coherence length of the wave packets. then
the partial reflected irltensities are added to give the total reflectance. Otherwise. the am-
plitude reflectances add and the phase differences between all of the waves that are intro-
duced by the thickness of the film must be taken into account (Eq. 1.2). The Fresnel reflection
coefficients. rio are indicated at each interface. The indices of the incident medium. the film.
and the substrate are indicated by no. n,. and n•• respectively. The refraction that occurs al
each surface is not shown in this figure.

of a film increases. Two different presentations of this information are used. In

Fig. 1.5, the complex reflectance amplitude is plotted on a polar coordinate rep-
resentation of the complex reflectance plane. The tick marks indicate constant
increments of thickness. Note that these are not equally spaced about the circle. In
Fig. 1.6, the intensity reflectance is plotted against the film thickness. Note that
the shape of the latter curve is not sinusoidal.
Several interesting effects can be gleaned from Eq. 1.3 and be seen in Figs.
1.5 and 1.6. One that has practical application is that the reflectance of the layer
reaches an extremum when the argument of the exponential reaches -rr/2 or 90
degrees and neither the film nor the substrate is absorbing. The optical thickness
that corresponds to this is called a quarter wave. and the layer is said to be a quarter
wave thick. or one QWOT (quarter wave optical thickness). Whether the extremum
at the quarter wave thick point is a minimum or a maximum depends on the relative
values of the index of refraction of the incident medium, the film material, and the
substrate. The extremum will be a minimum if either of the following two conditions
is satisfied:

8 CHAPTER 1

Imaginary Amplitude
Reflectance

0.4

1.6 0.2

-0.6 -0.2 0.2 0.4 0.6

1.0
Real Amplitude
Reflectance
OWOT
0.4 -0.2

-0.4

Figure 1.5. The amplitude reflectance of a film of index 2.35 of increasing thickness plotted
on the complex reflectance plane. The tick marks on the circle are spaced at intervals of
0.1 quarter wave optical thickness. The arrow indicates the direction of increasing film
thickness. The substrate has an index of 1.52. The amplitude of the reflectance is the distance
from the origin to a pOint on the circle, and the angle formed by that vector is the phase
angle of the reflected light. The intensity is the square of the amplitude value. Note that the
spacing of the thickness intervals is not uniform.

where no, nf' and ns are the indices of the incident medium, the film material, and
the nonabsorbing substrate, respectively. Otherwise, the reflectance reaches a max-
imum at the quarter wave point. When the index of the film is the geometric mean
of the incident medium and the substrate, then the minimum is exactly zero:
nf = Vno n ..
If a given material has no dispersion in its index of refraction, then the same peak
value of the reflection is obtained at all odd multiples of the quarter wave optical
thickness.
If the substrate is absorbing, e.g., it consists of a polished metal surface, then
the extremum does not take place at the point where the optical thickness of the
film is exactly a quarter wave, as shown in Fig. 1.7. The nature of the reflectance
for these cases is discussed in detail in in Chap. 3.
As the optical thickness of a nonabsorbing layer approaches twice a quarter
INTRODUCTION TO OPTICAL THIN FILMS 9

wave, i.e., a half wave or a thickness of 180 degrees, the argument of the exponential
tenn in Eq. 1.3 approaches 360 degrees, and the value of the exponential term
becomes unity. The implication of this is that a layer that is one-half wave thick
has no effect on the light passing through it. The complex exponential function is
unity at integral multiples of a half wave, so that the film essentially disappears at
those wavelengths also. At these integral mUltiples of a half wavelength thickness,
the film is described as being absentee, and the reflectance returns to the initial
value of the uncoated substrate.
The above effects are strictly true for homogeneous and non absorbing films.
Films that are either absorbing, inhomogeneous, or both, will not be absentee when

50

Q)
U
C
...
&a
U

-.....
Q)

Q)
a:
tP-

o
o 1 2

Thickness (QWOT)
Figure 1.6. Reflectance (intensity) as a function of thickness for a nonabsorbing film The
film index is approximately 2.35; the index of the substrate is 1.52. The peak reflectance
occurs at the point where the film optical thickness is an odd multiple of a quarter wave. At
even multiples of a quarter wave optical thickness, the film is a multiple of a half wave thick,
and the reflectance returns to that of the substrate. The curve is not a simple sinusoidal
function. Note that if the film index had been less than that of the substrate, the reflectance
would have gone through a minimum at the quarter wave point. The phase content of the
light is not recovered when only the intensity is measured.

10 CHAPTER 1

60~------------------------------------~

CD
U
C
...uca
-...
CD
CD
a:
~

O~--------~~------~--------~---------'
o 2 1

Thickness (QWOT)
Figure 1.7. Reflectance of a dielectric film on a metal substrate. Note the initial drop in
reflectance as the thickness of the thin film increases. The extrema do not occur at the
quarter wave points. The substrate index is 2.0 - 2.Oi and the film index is 2.35. At the half
wave pOints. the reflectance returns to that of the bare substrate.

they are a half wave thick. This information can be useful in evaluating the properties
of an unknown film.
The formulas given previously apply for normal incidence. Nonnormal inci-
dence effects are discussed in section 3.6. A complete treatment of this subject can
be found in Chap. 13 of Born and Wolf (1970).

1.1.1. Coherent Addition of Amplitudes

For interference to occur between waves, they must be coherent. This assumption
was made implicitly in the previous discussion on the interference between the
partial reflectances from the various interfaces of a thin film structure. Light waves
in the useful range of a given filter typically have a coherence length that is larger
than the total thickness of the filter, and the various partial waves can interfere with
each other. This interference alters the reflective and transmissive properties of the
original substrate.
In the discussion of interference, we added the wave from the first surface
with that from the second, and we assumed that the two waves were coherent. We
INTRODUCTION TO OPTICAL THIN FILMS 11

could do this because the coherence length of most light is several wave lengths
long. For a continuous wave (CW) laser beam, the coherence length is several
meters (a pulsed laser with a pulse length of I psec has a coherence length of 300
!Lm). With laser light sources, interference between waves can take place when the
two waves travel paths that are quite different in length. White light from a thermal
source on the other hand, does not have this coherence. In a typical thin film
configuration, where the total thickness of the stack is usually a small number of
waves thick, all of the waves reflected from the interfaces can be added coherently,
i.e., the amplitudes add, and the phases of the waves have to be taken into account
when adding these amplitudes.
The coherence length of light ~1 is approximately
~1 = '!..2/~'!..,

where'!.. is the wavelength and ~'!.. is the bandwidth.

In a monochromator that has a resolution of I nm, for example, this leads to
a coherence length of approximately 0.25 mm for light with a wavelength of 500
nm. A transparent layer with a thickness on this order of magnitude, or thinner,
will produce interference effects. If the wavelength is 5 !Lm, and the bandwidth of
a monochromator is proportionately the same as the preceding case, then the co-
herence length will be 2.5 mm. This is on the order of magnitude of some substrate
thicknesses. Under these circumstances, a substrate acts as an interference layer.
The light waves reflected from the two surfaces interfere and produce a fringe
pattern that is a function of wavelength. When the distance between the two re-
flecting surfaces is large, many interference maxima and minima can occur over a
given wavelength range. This spectral pattern is called a channel spectrum (Jenkins
and White, 1957, p. 284). This channel spectrum can be superimposed on the
spectrum of a filter. This manifests itself as a high frequency ripple on the spectral
curve of a filter. On a wavenumber scale, the ripple frequency will be constant
over the entire spectrum. On a wavelength scale, however, the ripple frequency
will increase as the wavelength decreases.
When the length differences between two paths are greater than the coherence
length, the phases of the two waves are randomized and they cannot interfere with
each other. In such a situation, the two waves behave as two incoherent sources.
Such sources are combined by adding the intensities of the two waves, rather than
their complex amplitudes [see, for example, Stone (1963) Chap. 13, for a derivation
of the addition of incoherent sources].
1.1.2. Layer Thickness Specifications
The thicknesses of the layers in a thin film coating are specified in a number of
ways. The first is the physical or metric thickness of the individual layers, i.e., the
thickness that might be measured by an ultrasensitive micrometer. This thickness
is found frequently in publications, and is practically always used when specifying
absorbing layers.
A more useful way to represent the thickness of a layer is the optical thickness.
The optical thickness is the product nd, where n is the index of refraction of the
12 CHAPTER 1

layer material and d is its physical thickness as described previously. It is this

product that is frequently important, to first order, in optics. Thus, this is a common

A previously mentioned thickness specification that is commonly used in optical

filter circles is called the quarter wave optical thickness, or QWOT for short. It is
defined as

QWOT = 4 nd/A,
where nd is the optical thickness of the film and A is the wavelength at which the
QWOT condition occurs.
Note that this thickness specification implies a wavelength. Although a design
may be designated as consisting of layers with a certain number of QWOTs, it is
not fully specified until the wavelength at which those QWOTs are defined is given.
This measure is useful because many nonabsorbing thin film filters consist of
layers that are one or more quarter waves thick. A very common design is a stack
made up of layers of alternating indices, the thicknesses of which are all one quarter
wave at some given wavelength. A design specified in these terms is instantly
recognizable. If the coating design were given in terms of the physical thicknesses
of the layers, it is not so readily identified, and its performance characteristics would
be(ome obvious only after some amount of computation.
It is important when looking at the design data of a thin film filter to understand
which type of thickness representation is being used.

1.1.3. Nomenclature
The roots reflect, transmit, and absorb can have a number of suffixes such as
-ance, -ivity, and -tion. The distinctions among the suffixes are not universally
accepted, but the current common usages, and the one followed here, are
-ivity: e.g., transmissivity, reflectivity, absorptivity
The fundamental properties of a material, or theoretical or calculated values based
on first principles.
-ance: e.g., transmittance, reflectance, and absorptance
Measured values.
-tion: e.g., transmission, reflection, and absorption
Generally refers to the phenomena rather than actual values. Frequently used in-
terchangeably with the -ance version.
Because quarter wave layers are frequently used in thin film optics, a notation has
evolved that represents such layers, as well as multiples and submultiples of them.
An uppercase letter is used to designate a quarter wave layer with a particular index
of refraction; e.g., H, represents perhaps, a QWOT of a high index material, and
L represents a QWOT of low index material. A basic two-layer period consisting
of a low and a high material would be designated as HL. If more than one period
exists, an exponent is used: (HL)IO represents 10 periods of the basic stack (20
alternating layers). If a layer is less than a quarter wave thick, it may be designated
INTRODUCTION TO OPTICAL THIN FILMS 13

as a fraction of the quarter wave; e.g., 0.5 H or H/2 is an eighth wave thick layer
of the high index material.
1.1.4. Index of Refraction and Dispersion
The index of refraction discussed previously varies as a function of wavelength for
most real materials. In some restricted wavelength ranges, this change may be
negligible. The index of some materials can change as a function of a number of
variables, such as time or UV (ultraviolet) light exposure. Therefore, the common
nomenclature of optical constants of a material is somewhat of a misnomer.
In this text, the theoretical perfonnance curves have no dispersion included in
the calculation unless specified otherwise, so that the primary features in the per-
formance curve can be seen, unobstructed by dispersion effects.
The change in index of refraction as a function of wavelength is referred to
as dispersion. For most nonmetallic materials used in thin films in the visible portion
of the spectrum, the index is typically higher in the blue end of the spectrum, and
lower in the red end. Similarly, in the infrared, the index of most nonmetallic
materials decreases as the wavelength increases. Near absorption bands, however,
this trend may be reversed and the index can increase as the wavelength increases.
Concurrently, the absorption coefficient, k, is generally nonzero.

1.1.5. Absorbing and Nonabsorbing Filters

Filters can be classified as either absorbing or nonabsorbing. In nonabsorbing filters,
the energy conservation fonnula R + T = 1, where R and T are the intensity
reflection and transmission ratios, respectively, applies exactly. Where there is some
absorption, this equation must be modified to R + T + A = 1, where A is the
absorption in the filter. In this discussion, we ignore scattering by assuming that it
is insignificant.
A fundamental theorem of optical thin films is that the transmission of any
coating is always independent of the directions of the arrival of the light [see Macleod
(1969), p. 24 for a proof]. Thus, the transmission of light that is incident from the
substrate side of the coating will be equal to that of light incident from the opposite
side. A change in the index of either the incident medium or of the substrate, or
both, will likely change the transmission value but it will remain independent of
the direction of incidence. If there is no absorption in either the film materials (or
the substrate), then by the law of energy conservation, the reflection must also be
independent of the direction of the arrival of the light.
The reflection of a coating that contains absorbing materials does not follow
the aforementioned rules. The reflection and the reverse reflection do not have to
be the same. Because the transmission is always the same in both directions, and
the reflection depends on the direction of the light's arrival, then in accordance
with the law of conservation of energy, the absorption is also sensitive to the
direction of the light's arrival. This effect is readily shown for a metal film with
the complex index 2 - 2i and thickness 5.0 nm on a glass substrate (n = 1.5).
The reflectances from the air side and the glass side are shown in Fig. I.B. An

14 CHAPTER 1

20,-------------------------------------~

15
Q)
U
C
CO
~ n=2-2i t=8nm

--
U
Q)

Q)
10
n=1.52 1\
a: R1
"iP-
5

... - ................ R1
- ... _----
..
....
------------
O~----------~------------~----~~ ~
400 500 600 700
Wavelength (nm)
Figure 1.8. Reflectance and reverse reflectance of a metal (n = 2, k = 2) film on a glass
substrate (n = 1.5). The reflectance (and absorptance) depends on the direction of incidence
of light. There is no dependency in the transmission on the direction of arrival of the light.

almost perfect antireflection condition can be obtained when the interface is viewed
from the glass side, yet the reflection is moderately high when the coating is viewed
from the air side. Concurrently, absorption of the light arrivbg from the glass side
is higher than that arriving from the air side since the transmission is the same,
irrespective of the light's direction of arrival.
When a filter is measured in transmission, it is not important which side faces
the light source and which, the detector. If the reflectance is being measured,
however, it is important that the proper face of the sample be oriented toward the
light beam if significant absorption exists in any of the layers in the coating in the
spectral range of interest.
Along with absorption, scatter is another mechanism of energy loss from the
specular beam. Scatter sites are generally microscopic volumes, the refractive in-
dices of which are different from that of the surrounding medium. The light dif-
fracted from these sites is distributed in all directions and is, therefore, lost from
the specular beam. If the scatter sites are large enough, the reflection and refraction
may also playa role and contribute to the overall light lost. Each site removes only
a small percentage of the total energy in the beam, but as there can be many such
sites, the cumulative effect can add up to a serious light loss.
INTRODUCTION TO OPTICAL THIN FILMS 15

In optical coatings, surface roughness and local index variations frequently

give rise to scatter. The following are some of the situations that can occur:
1. The substrate can be rough. Because thin films for the most part do not
smooth out a surface, some of the surface defects may replicate to the top
surface.
2. The films can have inhomogeneous properties, such as inclusions or other
defects that have a different index of refraction than the surrounding
material.
3. The interfaces between films can be rough.
4. The surface of the coating may be rough, so that the air/film interface is
the source of the scatter sites. The difference in index ofrefraction between
air and the first film is large, so that this roughness can lead to a significant
light loss.
Many thin film coating materials grow in a columnar structure as they condense
on the substrate. The relatively limited lateral motion of the arriving atoms and
molecules produce little smearing of the underlying structure. As a consequence,
the original structure can be replicated through a large thickness.

1.2. SPECTRAL RANGES AND MATERIALS SELECTION

By far the most difficult problem in the field of thin film engineering is that of
finding the proper materials for a particular application. It is a rather straightforward
task to design a coating that meets just about any arbitrary specification if one does
not have to worry about building it with the materials supplied to us by Mother
Nature. The limitations imposed by real materials can be very severe.
In addition to optical problems, the designer also has to cope with coating
process, mechanical, environmental, and sometimes chemical problems in con-
verting a design from paper to reality. Some of the mechanical problems that are
frequently encountered are stresses in the films that cause the substrate to warp or
the films to delaminate through adhesive or cohesive failure. The process temper·
ature, which will ensure an environmentally stable coating, may be too high for
the substrates. The substrates may react with the coating materials in some ways
that cause degradation. Some materials cannot survive outside the vacuum chamber
because of oxidation or moisture reactions. Less severe cases of this may leave the
film intact, but make it too fragile to be of any practical use in other than a controlled
environment.
Some materials are useful only in certain narrow wavelength regions, whereas
others have much broader utility. A material does not have to be transparent every-
where to be useful; for example, in the blocking of certain wavelengths, absorption
is the property of interest, though usually, some region of transparency is also
generally desired.
Some materials, such as magnesium fluoride, are transparent over very wide
spectral ranges. Mechanical stresses, on the other hand, can severely limit the
maximum thickness attainable without failure. The maximum quarter wave optical
thickness of a magnesium fluoride film that can be typically obtained with a con-
16 CHAPTER 1

ventional evaporation process is on the order of I to 2 ~m. Magnesium fluoride is

an interesting example of a situation in which stress is not necessarily the only
consideration. This material, when evaporated, forms films with high tensile stresses,
yet it is a very usable material because its adhesive strength to glass is extremely
high. The upper limit on the thickness of these films is not governed by adhesive
failure to the substrate, but by cohesive failure in the film when the tensile strength
limit of the film is reached. The failure point depends on the coating processes,
the expansion coefficient of the substrate, and other factors. Magnesium fluoride
is an example of what occurs to some extent in virtually all materials. These problems
of material selection are discussed in more detail in Chap. 6.

1.3. ANGLE OF INCIDENCE EFFECTS

The performance of a filter that uses the absorption properties of materials to obtain
the desired spectral shape does not change markedly as a function of the angle of
incidence. This is readily apparent when a beam of white light is observed through
a colored glass plate: its color does not shift spectrally when the filter is tilted. The
optical density of the filter increases owing to the increase in the path length inside
the absorber. The ratio of the internal transmission at some angle to that at normal
incidence is given by
Ti = exp[ -o.d(sec 82 - I)],
where 0.is the absorption coefficient, d is the thickness of the absorber, and 82 is
the angle the ray makes with the normal inside the absorbing medium.
This equation is Valid for relatively weak absorbers, such as most colored
glasses, which are basically a glass matrix doped with specific ions or atoms. The
complex part of the refractive index of the absorbing filter can be ignored in
evaluating Eq. 1.2. It should be noted that this discussion ignores the effect of the
surface reflectances: at high angles of incidence, these can be substantial.

1.3.1. Snell's Law

A light beam crossing an interface between semi-infinite media with different indices
of refraction changes direction if the angle of incidence is nonnormal. The change
in the direction of the light beam is given by Snell's law:
(1.4)
where ni is the index of refraction in medium i and 8i is the angle the beam makes
with the normal to the interface plane.
The angle of incidence is the angle of the incident beam, whereas the angle of
refraction is that for the refracted beam. The reflected and the refracted beams lie
in the plane defined by the incident beam and the surface normal at the point where
the beam strikes the interface.
In a thin film coating, the interfaces are essentially all parallel, and the product
n sin 8 has a constant value in all layers (see section 1.3.2.). Equation 1.4 applies
INTRODUCTION TO OPTICAL THIN FILMS 17

to both nonabsorbing as well as to absorbing materials. For complex indices, i.e.,

one or both media are absorbing, the angles and their sines are also complex.

1.3.2. Optical Invariant

In a nontilted optical system with plane parallel surfaces, the product n sin e is a
constant throughout the system. This is a restatement of Snell's law. The same
principle holds true at all points in a thin film filter stack. The constant is defined
by the ray entering the coating and the index of the incident medium. The value
is generally determined by the incident angle as measured in air, in which case the
constant is simply sin e (Fig. 1.9).
For a plane parallel glass plate with a coating on the second surface (the surface
away form the incident light beam), the coating is being used effectively immersed
in an index of 1.52. The n sin e product is defined in air, because all of the surfaces
are plane and parallel. Thus, the only effect of the immersion is that due to the
change in the index ratios, and the angle of incidence effect is the same as it would
have been if the coating were in air. Also, it should be noted that in such a situation
the maximum angle of incidence on a coating on the second surface of a plane
parallel glass plate is limited to sin- 1 (lIn), where n is the index of the glass. For
prisms and other nonparallel surfaces, the preceding statement does not hold, and
the rays need to be traced individually to determine the appropriate angles.

1.3.3. Critical Angle

When the index of the incident medium is greater than that of the second medium
and neither is absorbing, there exists an angle in the incident medium beyond which
light will not be transmitted into the second medium. This angle is called the critical

------r-~~--~------~------------~--------~~ X

no Sin 8 0 = n1 Sin 81 = = n3 Sin83

Figure 1.9. Illustration of the n sin 0 invariance. In a plane parallel geometry, this product
is constant throughout the structure.
18 CHAPTER 1

angle. At the critical angle, the angle of the refracted beam is 90 degrees; i.e., the
refracted beam is travelling parallel to the interface. The critical angle 6c is given
by
6c = sin-I (ndno)

where no is the index of the incident medium, nl is the index of the second medium,
and nl < no.
Note that there is no critical angle when the incident medium has an index
that is less than that of the second medium.
Any increase in the angle of incidence beyond the critical angle results in total
internal reflection (TIR), and no light energy enters the second medium. In practice,
it is possible to obtain a reflectance value very close to the theoretical one of 100
percent when TIR is utilized. TIR is used in many prisms. The standard laws of
reflection i.e., the angle of reflection is equal to the angle of incidence, apply to
those cases of reflection beyond the critical angle.
Evanescent waves can penetrate the interface when the angle of incidence is
greater than the critical angle. The amplitude of these waves decreases exponentially
in the low index medium; their short attenuation length is on the order of a few
wavelengths (Jackson, 1962, p. 221). With dielectric media, and in the absence of
an absorbing layer at the interface, this attenuation of the amplitude of the wave
does not remove any energy from a light beam, and the reflectance remains at 100
percent.
An absorbing medium can be located such that it interacts with the evanescent
waves; for example, absorbing particles can be located on the reflecting face of a
prism. This interaction is strong, and energy is removed from the beam that was
undergoing TIR. This has been a fertile field for experimenters who are interested
in the measurement of materials with very low absorption coefficients; [see Harrick,
(1967)]. This technique of measuring absorption also is called frustrated internal
reflection or attenuated total internal reflection. Figure 1.10 shows the relationship
between the critical angle and the index of refraction ratio.
1.3.4. Polarization
The polarization of light is sometimes overlooked in specifying optical filters. In
a complex optical system, some degree of polarization may come about inadvert-
ently. For example, the nonnormal reflection coefficient for a dielectric coating or
metallic mirror depends on the plane of polarization. An unpolarized incident beam
may be partially polarized as it progresses through an optical system, and one
polarization may emerge with a higher intensity than the other. This could be
significant if there are many reflectors in a system. If such is the case, then it is
prudent to do a careful analysis of the system to verify the nature and amount of
the polarization introduced into the beam.
Laser light is frequently plane (linearly) polarized and this can make the analysis
of a laser system simpler. It should be noted that coated mirrors can introduce
differential phase shifts that can lead to circular or elliptical polarization (see section
1.3.7.). When a laser with random polarization is used, a system analysis needs

INTRODUCTION TO OPTICAL THIN FILMS 19

.
......
;;. 60
...
S
...u
~
'1:1
T=O"
U
.E
'0
... 30
'at
c Transmission Possible
c I
I
I
I
I
I
I

o , 2 3 4 5
~
Total Internal Reflection "0 1"1 (Log Scale)
!'!Q.!
Possible (nO >n,)

Figure 1.10. Plot of critical angle versus index ratio (n<ln1) for nonabsorbing materials. In
the areas where transmission is possible, its actual value is determined by the reflection
coefficient for the interface. The transmission is 100 percent for those cases where no = n1
(dashed vertical line), and the value of T decreases on either side of it. The transmission is
zero along the vertical axis.

to be done separately for each plane of polarization. A beam of light can have two
states of polarization. For plane polarized light, these states are specified with respect
to the plane of incidence.
Alternative polarization states are sometimes used when the analysis lends
itself more easily to these different fonns. An example is the use of right and left
hand circular polarization. Details on these topics may be found in many references,
such as Shurcliff (1962) and Clarke and Grainger (1971).
The electric field of a plane polarized electromagnetic wave can be decomposed
into two orthogonal components, commonly designated "s" and "p." In the fonner,
the plane of oscillation is perpendicular to the plane of incidence (The "s" designation
derives from the Gennan word for perpendicular, senkrecht.). In the latter, the
plane of oscillation is parallel to the plane of incidence ("p" state).
In a complex system, each needs to be examined as a beam that has an s
polarization at one interface may, because of the system's geometry, be a p state
at another interface.
The quality of a polarizer is frequently given in tenns of its extinction ratio.
This parameter p is defined by

20 CHAPTER 1

where Tmax is the maximum intensity of light which the filter transmits when its
axis is aligned with the polarization of the probe beam and Tmin is the
minimum intensity of light that the filter transmits when its axis is rotated
90 degrees in the polarized beam (Driscoll and Vaughan, 1978, pp. 10-13).

1.3.5. Brewster's Angle

Brewster's angle at an interface between two nonabsorbing media is given by the
angle of incidence whose tangent is the ratio of the two indices:
tan 68 = n]lnO,
where 68 is Brewster's angle in medium no, no is the index of the incident medium,
and n] is the index of the substrate.
At this angle, no light in the incident beam with polarization parallel to the
plane of incidence (p plane) will be reflected. Conversely, the tangent of Brewster's
angle in the second medium is given by noln]. By conservation of energy, all of
the energy is transmitted into the second medium with 100 percent efficiency. An
interesting aspect of this geometry is that the refracted ray and the reflected ray
have an included angle of 90 degrees. Brewster's angle is the basis of some tech-
niques for determining the index of refraction of films, such as Abeles' method
(1950).
The design of high reflectors becomes difficult as the angle of incidence ap-
proaches Brewster's angle. The reason for this is apparent when one evaluates the
effective indices for both planes of polarization in this angular regime.

1.3.6. Effective Indices

At nonnormal incidence, the Fresnel reflection equation is still applicable, provided
that effective indices are used instead of the normal values. The effective indices
depend on the plane of polarization and are given by
ns = n cos6 and np = nlcos6
where ns and np are the effective indices for the s and p states of polarization,
respectively, and 6 is the angle of the beam of light in a medium with an
index n.
Figure 1.11 shows the effective index for the two planes of polarization as the
angle of incidence in that medium varies from 0 to 90 degrees. The effective index
of two and four times the index of the medium in which the angle of incidence is
measured is also shown. If the incident medium is air, the values of 2 and 4
approximate the indices of ZnS (n = 2.2) and germanium (n = 4.0), two common
infrared coating materials. The effective index changes relatively little for the higher
index materials, whereas it changes markedly for low index materials such as air.
As a result, coatings that consist of higher index materials shift less as the angle
of incidence increases than do coatings with lower average indices.
The curves for the s plane light do not intersect, whereas those for p plane
light do. The angle at which the curves intersect correspond to the Brewster angle

INTRODUCTION TO OPTICAL THIN FILMS 21

4t---------================~~&::
3
neff
no

1 t-====:::::==-~

10 20 30 40 50 60 70 80 90
Angle of Incidence (in medium no)
Figure 1.11. Effective index as a function of angle of incidence. The effective index is
normalized to the index of the medium in which the angle of incidence is defined (no). The
sand p plane values of the effective index are marked. The intersection of the p plane
curves occurs at Brewster's angle for that pair of indices. At index values below no, the
curves have the shape of the unity curves, except for a scaling factor along the abscissa;
the effective index of the s plane component then goes to zero at the critical angle.

for that pair of indices: since the effective indices are equal, the interface disappears
at that angle and there can be no reflectance from it.
Figure 1.12 shows variation in the reflectance as a function of the angle of
incidence at air interfaces with glass (n = 1.5), a poorly reflecting metal (n = 2.0,
k = 2.0), and a highly reflecting metal (n = 0.5, k = 3.0). The reflectance of
light with p polarization is zero at Brewster's angle. At all angles (except nonnal,
of course), bare interfaces act as polarizers to some extent.

1.3.7. Birefringence and Differential Phase Shifting Coatings

Birefringence is a phenomenon related to the optical properties of materials. It
occurs when a crystal or other molecular system is anisotropic so that it has more
than one optical axis. Materials that are ordinarily isotropic can exhibit birefringence
when placed under stress. The strain can break the symmetry of the crystal or
molecular structure of an otherwise symmetric lattice and the strained system will
be birefringent. Thin films are known to be birefringent (Johnston and Jacobs,
1986), but as they are so thin, the effect only rarely manifests itself in an observable
way.
The effects of materials' birefringence and those of the phase shift differences
introduced by thin film coatings can lead to similar polarization changes. Thin film
22 CHAPTER 1

designs with such properties have been caJled birefringent. Strictly speaking, thin
film filters are not birefringent because this term refers to a material property.
Ellipsometric techniques are typically used to measure the differential phase shift.
When the angle of incidence is not 0 degrees, the phase shift for the two planes
of polarization is not equal. In some instances, this is a useful property. By proper
design, a coating can change the relative phase of the two components such that
incident linearly polarized light will be reflected as circularly polarized. In effect,
this type of system functions as a reflective quarter wave plate. These are useful
in spectral regions where transmissive birefringent quarter wave plates are difficult
to obtain. Figure 1.13 shows the performance of a typical design in the infrared
portion of the spectrum. These, and similar designs, are available commercially.
By using more than one surface to achieve the desired phase shift, it is much easier
to fabricate and use these devices, as they are quite wavelength sensitive when the
entire shift is attempted on one reflection. The shifts are additive, so a cascade
system is effective. Several papers have been published on this subject recently
(Apfel, 1982, 1984; Southwell, 1980).

75
G)
U

...r:::uas
G) 50
;:
G)
~
"fP.
25

30 60 90
Angle of Incidence
Figure 1.12. Reflectance at various air/bare substrate interfaces as a function of angle of
incidence. At Brewster's angle for glass, the p component of the light is not reflected. The
reflectances of the p component for metals also have a minima.

INTRODUCTION TO OPTICAL THIN FILMS 23

100 ~--------------------------------------------~

--
en
GI
...
GI
g)

~ 90
en
eI
Q.
e

80 ~----~-------.------~------~------~----~
.925 .950 .975 1.0 1.025 1.050 1.075

Wavelength (pm)
Figure 1.13. Phase difference (<I>p - <1>.) upon reflectance from a mirror as a function of
wavelength. This design has a nominal phase difference of 90 degrees.

1.3.8. Steep Surfaces

In some diffraction limited optical systems, the angle of incidence of a rayon a
lens or mirrors surface can be steep (angles of incidence may be in excess of 30
to 45 degrees). In these severe situations, it is necessary, when doing an optical
systems analysis, to take into account the angular effects, such as polarization, for
each of the different ray angles that strike these elements. A correct analysis uses
a ray tracing system that takes all of the surface polarizations into account, and
then combines the rays according to the rules of diffraction optics in the image
plane. Such an analysis should also take into account the angle shifts of the per-
fonnance of thin film designs, as well as their contributions to polarization phase
shift differences.
 Chapter 2

Deposition Technology

Thin film optical filters can be prepared using a number of different coating tech-
nologies. A few of these methods have found their way into commercial production
processes. In this section we discuss some of the methods currently used for com-
mercial filter production. Although the subject matter is extensive, this section is
necessarily brief. For more detailed information on this subject, the reader is referred
to the text by Maissel and Giang (1970).

2.1. COATING PROCESSES

2.1.1. Vacuum Evaporation

The first thin film interference coatings for optical applications were made by
resistively heated thermal evaporation, and this process is still in use today. A
charge of the material to be evaporated is placed in a crucible of a refractory material.
An electrical resistance heater heats this charge (either by conduction or by radiation)
after the chamber has been evacuated to about 10-6 torr, or one billionth (10-9 ) of
an atmosphere (Table 2.1 contains factors to convert between some common units
used to express pressures and vacuums). As the temperature of the charge rises,
its vapor pressure rises and a significant evaporation rate develops. The vapor

24
DEPOSITION TECHNOLOGY 25

Table 2.1 Conversion Factors Between Various Pressure Units for

Gases and Vacuum
1 Pascal = 10-5 bar = 7.5 x 10-3 torr
1 bar = 0.987 atmosphere = 105 Pa
= 14.5 pounds per square inch
1 micron (of Hg) = 1(J3 torr = 1.3 x 1O-S atmosphere = 0.13 Pa
1 torr = 1 mm of Hg = 133 Pa
Note: The Pascal (Pa) is currently the official metric unij (Systeme Intemationale, SI) for designating
pressures.

pressure-temperature relationship is the primary detenninant of the evaporation rate.

The evaporant then condenses on the cooler substrate. This process of indirectly
heating the charge is still used today, though perhaps with some refinements.
The crucible material may be either a refractory metal or a ceramic container.
In indirect heating, the heating element does not come into direct contact with the
charge. This can be accomplished radiantly by placing a tungsten filament over the
charge or by conduction with a filament or other element that is in contact with the
crucible.
Direct heating takes place when the charge is directly heated by the hot element.
In one method, which is widely used, the charge is placed in a depression on a
refractory metal strip. In some materials available in wire fonn, i.e., aluminum
and gold, short pieces of the wire can be wrapped on a stranded tungsten heater
wire. A current through the refractory metal heats it and the charge. Once the
material in the charge reaches its boiling or sublimation temperature in the vacuum
(lower than the nonnal atmospheric boiling point), atoms and molecules of the
material leave the charge and condense on the cooler surfaces in the chamber. Thus,
a film builds up on the substrate (as well as the rest of the chamber).
The equipment used in the pioneering days of thin films coatings suffered from
a lack of sufficient vacuum pumping capacity as the best high vacuum pumps
available were small mercury diffusion units. As a result, the lowest or base vacuum
that could be attained in a typical coating chamber was not very good by today's
standards. Furthennore, as the evaporation process progressed, the evaporation
source radiantly warmed the chamber, causing the walls to release some of their
adsorbed gases and water vapor (at the worse possible time, i.e., during the actual
deposition of the coating). The coatings thus produced contained entrapped residual
gases from the vacuum system. These soft, porous films had relatively low dura-
bility. On the other hand, this approach is relatively gentle, and can be used for
compounds that have a tendency to decompose.
The technology of vacuum coating improved with time. The development of
high temperature, low vapor pressure silicone oils led to the construction of larger
diffusion pumps with much larger pumping capacity. The elimination of the mercury
also made them safer.
Diffusion pumps operate efficiently only at low pressures at both the inlet and
outlet ports. Mechanical pumps are used to evacuate the chamber to a pressure
26 CHAPTER 2

within the operating range of a diffusion pump. This is called the rough vacuum.
Once the chamber has been roughed, the mechanical pump is used to keep the
outlet port of the diffusion pump at a low pressure, called the backing pressure.
Reliable mechanical pumps of the rotary vane type for roughing and backing of
the diffusion pumps, and of rotary (HRootes") blowers for increasing the mechanical
pump throughput set the stage for modem vacuum systems and coating technology.
In time, more efficient means of heating the evaporant charge were developed,
including the electron beam and the laser beam. This latter method of heating a
charge in a vacuum evaporation chamber was first demonstrated by Smith and
Turner (1960) with a pulsed ruby laser. Hass and Ramsey (1969) used a carbon
dioxide laser to evaporate dielectric and semiconductor materials. Sankur and Hall
(1985) used the laser heating method to produce films with little ionization of the
evaporant. In an electron beam heating system, electrons from a high voltage
(typically 5 to 15 kV), high current (up to several amperes) electron gun are focused
on the surface of the charge that is contained in a water-cooled metal, typically
copper, crucible. This heating process is a very efficient way of depositing energy
into a material in a vacuum. As compared to a resistively heated charge, relatively
little radiant heat, except from the surface of the molten charge, is produced by
the evaporation source in an electron beam heated system. A resistively heated
source with a comparable coating rate radiates from a number of surfaces and this
heats the coating chamber much more. Another major advantage of using the
electron beam heating technique is that it makes it practical to evaporate refractory
materials, the melting points of which can exceed that of tungsten metal. These
materials cannot be evaporated using resistively heated boats because they cannot
be heated to sufficiently high temperatures to obtain good coating rates. In particular,
the oxides of zirconium, tantalum, silicon, titanium, and aluminum became practical
coating materials with the advent of the electron beam heating system. Another
drawback of the resistance method of heating a charge is the fact that some materials
may react chemically or alloy with the hot crucible and cause it to fail.
At room temperature, oxides are geneally good electrical insulators. Under
intense bombardment from a sufficiently energetic electron beam, however, all
insulators and other refractory materials can be heated sufficiently to cause them
to evaporate (unless they first decompose).

2.1.2. Sputtering
Sputtering is another vacuum coating process that is used extensively for depositing
optical thin films. This process is most easily characterized as a momentum transfer
process in which argon or other ions and atoms from a plasma bombard a target
made of the material to be coated. The area of a sputtering target may be an order
of magnitude larger than an equivalent evaporation source. The collision of the
argon atoms and ions with the surface atoms and molecules of the target knocks
off (sputters) the target material, which then forms a deposit on the substrate.
Flowing water keeps the target cool, so thermal evaporation of the target material
does not occur. Figure 2.1 shows the relative scale of a sputtering system's geometry
as compared with that typically used in an evaporation system.

DEPOSITION TECHNOLOGY 27

Planetary Motion

To Pumping System
~ubsir.teCs) SUbstratecsjl ~

\ / ,-----.,

\ /
\ \ I , /

\ \: ;' / Evaporation
Chamber

\ \1 / /
\\111 Source

! I To Pumping S~tem
Substrates Substrates

Target Target

Sputtering Chamber
Figure 2.1. Diagram of sputtering and evaporation chambers on approximately the same
scale. The substrate size capability of these chambers is the same. The evaporation source
is approximately a point source, whereas the sputtering targets are a large area source.
Note that the sputtering system can be made much smaller than a comparable evaporation
system.

The target to substrate distance in typical sputtering systems is short as com-

pared with that found in evaporation systems; this is a consequence of the nature
of the sputtering process. A number of advantages result from this and are outlined
later.
Because the sputtering process does not rely on keeping the target surface at
a high temperature as in an evaporation type of process, it is possible to use relatively
large area targets and to place them close to the substrate without having them heat
the substrate radiatively. This is an important consideration as sputtering processes
yield far fewer atoms and molecules per area of target per second than does an
evaporation source. In particular, oxide targets frequently result in low deposition
rates, and large area targets are required to obtain practical rates.
28 CHAPTER 2

A heavy, chemically inert gas, such as argon, is frequently used as the major
constituent of a sputtering gas. A heavier gas is more efficient in a momentum
transfer process such as sputtering. Few target materials sputter well when light
gases, such as nitrogen and oxygen, are the major components of the sputtering
gas.
The gas pressure in a sputtering system is relatively high compared with that
in an evaporation system 00-3 mbar vs. 10-6 mbar); therefore, the mean free path
of a sputtered particle is three orders of magnitude less than a particle in an evap-
oration system (Fig. 2.2). The sputtered target atoms and molecules are energetic
as a result of the momentum transferred when they were struck by the argon atoms
and ions. For the target atoms to carry this energy to the substrate and not lose
much of it through collisions, the distance between the target and the substrate must
be as small as possible. This is another reason why the substrate must be placed
closer to the target in a sputtering system than it is in an evaporation system.
Because the target is a large area source, as opposed to the approximate point

1000 r-~~----------------------------------~

- 100

--
E
u
.r:.
CO
Q.

..
Q)
Q)

u.
10

c
CO
Q)
~
1

Evaporation Sputtering

0.1 '-!!5~------~4-------~---~2~--~""
10- 10- 10- 3 10- 10-
1
Pressure (mbar)
Figure 2.2. Mean free path versus pressure. The operating ranges of evaporation and
sputtering processes are indicated.
DEPOSITION TECHNOLOGY 29

source characteristics of an evaporation source, there is relatively less of a problem

with nonuniformity in the thickness of the coating with a sputtering system. The
size of an evaporation system is indicated by the need to be far enough away from
the source that the distribution of the evaporant flux is within certain predefined
tolerance limits. This is not the case with a sputtering system.
A short target to substrate distance is advantageous from the standpoint that
there is a better coating material utilization factor in sputtering processes than in
evaporative systems.
There are several variations of the sputtering process. If a conducting target,
such as a metal, is used, then a DC voltage can be applied to it to maintain the
plasma discharge that accelerates the argon ions. This is referred to as a diode
arrangement. Sometimes, a grid is included in the system to help stabilize the
discharge, and these are called DC triode systems. If the target is not conductive,
then it is not possible to use it to carry the direct current. One then uses an alternating
current and couples it capacitively to the plasma. A radio frequency of approximately
13 MHz is used for this purpose. Power levels may range from a few hundred watts
to a few kilowatts.
Reactive sputtering can be used in situations where a nonconductive material
is to be deposited. If a metal target is used, then a DC configuration can be used.
Oxygen is mixed with the argon that is supplied to the plasma, and both gases
become ionized. This highly active oxygen then can oxidize a surface layer of the
target, which then gets sputtered, or it can oxidize the growing film on the substrate.
The exact mechanism that is operative is determined by the geometry of the system,
as well as by the gas mixture and oxidation chemistry. If the oxidation process is
too vigorous, an oxide layer can build up on the target, effectively shutting off the
direct current that excites the plasma, with the consequence that the plasma is
extinguished.
In both the DC and the AC sputtering systems, there are many free electrons
in the plasma. These electrons are highly mobile and are accelerated by the electric
fields present in both types of discharges. The low mass of the electrons allow them
to follow an alternating electric field easily. These electrons bombard the target
and cause it to get hot, thus water cooling lines are included in the target designs
to carry away this excess heat. The substrates are also close to the plasma, and
they too get bombarded by the electrons and become heated. Substrate heating is
one of the limitations of this type of coating system.
Magnetron sputtering is one technique that was developed to reduce the sub-
strate heating by electron bombardment. The concept is that the free electrons can
be trapped by magnetic fields and controlled. The magnetic fields are weak enough
so that the paths of the heavier argon ions and charged sputtered species are not
significantly affected. Thus, the electrons with their small mass to charge ratio as
compared with the ions are confined to the vicinity of the target, and do not bombard,
and thereby heat, the substrate. An extra benefit can be realized from this arrange-
ment: as the electrons are trapped in orbits away from solid surfaces, they can
ionize the gases in the plasma zone more efficiently. Consequently a lower gas
pressure can be used to maintain the plasma. This means that there are fewer
30 CHAPTER 2

collisions between the sputtered atoms and the background gas and plasma as the
sputtered molecule makes its way to the substrate. Magnetron types of systems are
becoming very common today, and are commercially available in AC or DC ver-
sions. Their coating rates are far superior in most cases to what one can achieve
with conventional sputtering.
A recent development in sputtering technology is the ion beam sputtering
technique. In this approach to sputtering, a beam of particles is accelerated in an
ion gun and directed at the target. These particles knock atoms from the surface,
which then coat the substrate. Such a system can operate at the same vacuum levels
as evaporation processes, thus combining the benefits of both evaporation and
sputtering. Unfortunately, this coating process currently has deposition rates that
are far slower than that obtained with standard sputtering techniques. Thus they are
used at present only in research environments and when specialty films are required.

2.1.3. Flash Evaporation

Flash evaporation is a technique used to deposit multicomponent materials that (1)
are incompatible in bulk form at high temperatures and, therefore, cannot be heated
to the evaporation point together or (2) interact with crucibles or other containers
at the evaporation temperature.
In flash evaporation processes, the source material, generally in a powdered
form, is metered slowly onto a surface that is heated to a temperature well above
the evaporation temperature of the most refractory component in the mixture. All
components of the source material evaporate essentially instantly and completely.
Because no liquid pool forms, there is no chance for a chemical reaction to occur.
And as all of the material is evaporated before more arrives, fractionation doesn't
occur and the coating is homogeneous. Thus, the deposit is a homogeneous mixture
of the components that were dropped onto the hot surface. This process finds
relatively limited use as the deposit is not as durable as those produced by con-
ventional evaporation or sputtering.

2.1.4. Nonvacuum Processes

Some nonvacuum techniques are used to produce optical thin film coatings. Two
major categories of these are listed here, though their applications are limited at
present.

2.1.4.1. Organometallic Solutions. A nonvacuum technique for producing

thin optical films is the use of organometallic solutions that are applied to a sub-
strate's surface either by dipping or by spraying. The substrate is then dried and
baked in a furnace to convert the film to a pure oxide while the organic constituents
become volatile oxides. Thickness control is achieved by adjusting the temperature,
viscosity, and other properties of the solution. This technique has been used suc-
cessfully to coat high power laser optical components with antireflection designs
that have extremely high damage thresholds. It is also used to apply multilayer
antireflection coatings to glass substrates. An etch treatment after the baking step
DEPOSITION TECHNOLOGY 31

can produce a coating with a graded index that can result in a surface with extremely
low reflectance over a broad wavelength region (Minot, 1976).

2.1.4.2. Electroless and Electroplating Solutions. These chemical pro-

cesses are used to form opaque metal deposits. Both processes are generally per-
formed at atmospheric pressure, therefore large steel vacuum vessels are not re-
quired. Silver films from electroless processes are commonly used to coat large
astronomical mirrors that will not fit into readily available coating chambers. Many
of the chemicals that are used in these processes are environmentally hazardous,
and they must be disposed of carefully. In a production environment, where the
quantity of chemicals used may be large, the costs associated with their disposal
can become prohibitive, and therefore, the electroless process is not very widespread
commercially. On the other hand, the use of this process on an occasional basis
on a mountain top astronomical observatory could be quite attractive economically
in comparison with the alternatives of acquiring a vacuum chamber, or moving a
large mirror to another location for coating. A fundamental disadvantage of elec-
troless coatings is that they are soft and, therefore, not very abrasion resistant.
The thickness of an electroplated or electroless deposit is not measurable in
real time during the coating process; however, because the coating is typically
opaque, this is not a significant problem.
Metal electroplating is a technique that has been known since the first days of
electrochemical cells. The optical quality as well as the durability of some com-
mercially available gold plate is excellent. *
Soft copper can be plated to a thickness of a few tens of micrometers and then
turned with a single point diamond tool. The plated copper is an excellent material
for diamond turning. As a result of its softness, the turned surface should not be
touched by any solid object lest the surface be damaged.
As far as this author is aware, there are no ultraviolet, visible, or infrared
transparent materials deposited by these techniques.

2.1.5. Process Conditions

The quality of a nonmetallic film is frequently improved when the substrate tem-
perature is elevated above the ambient temperature before and during the vacuum
coating process. This can be achieved with chamber heaters, or, as in the case of
sputtering, from the process conditions themselves. The increase in surface tem-
perature allows the arriving species to move about on the surface prior to their
being incorporated into the growing film. The increased surface mobility increases
the chances that a low energy lattice site will be encountered, which in tum will
lead to a denser, more durable film. In some instances, high substrate temperatures
are required to obtain the proper crystallographic form of a material (Macleod,
1983, pp. 201, 204; Maissel and Giang, 1970, pp. 17-14)

*EpnerTechnology, Inc., of Brooklyn, N. Y" 11222, is one plating company that actively supplies
gold electroplated parts to optical industry types of specifications.
32 CHAPTER 2

In practice, temperatures from 50°C to several hundred degrees are used. The
upper limit on the temperature used is limited by a number of factors. In some
coating equipment, the amount of heating that can be done is limited by the available
heat sinking or cooling, as all of the heat that enters into a system must eventually
be removed. In other situations, the substrates, such as plastic materials, may not
be able to tolerate elevated temperatures (see Chap. 6). Finally, the time it takes
to heat and cool a substrate in a vacuum may be economically prohibitive.
It should be noted that radiation is the primary means of heating and cooling
a substrate in a vacuum chamber. In some special applications, conductive types
of heating techniques can be employed, but these are generally complicated. In
radiative environments, the emissivity of the heater and the substrate are the key
elements. If the substrate is highly reflective in the wavelength range in which the
heaters emit their energy, there will be poor coupling between the two and the
substrate temperature will rise only very slowly. Similarly, the coupling between
the substrate on the chamber walls is important when cooling takes place. A hot
substrate with a infrared reflective coating will cool slowly.

2.1.6. Coating Time

The time required to coat a filter is dependent on a number of parameters, all of
wnich can significantly alter the process time, and all of which depend on the actual
equipment that is used. The pumping speed, the size of the vacuum chamber, and
the state of cleanliness of the chamber, all affect the time it takes to reach base
vacuum, that point from which all subsequent actions are referred. The following
qualitative equation includes most of the important factors:
Coating Time Per Part = (Loading
+ Pump down
+ Heating
+ In situ cleaning
+ Coating
+ Cooling
+ Venting
+ Unloading
+ Machine clean up) / Parts in one run
The coating rate, layer thickness, number of layers in the design, and any special
coating operations enter into the coating time tenn in the equation. Because an
infrared layer may be 10 to 20 times the thickness of a similar layer in a visible
coating simply due to the difference in wavelength, a sizable difference in coating
times can be expected for comparable designs in the two portions of the spectrum.
As previously indicated, the type of coating process, whether evaporative,
sputtering, or other process, significantly affects the coating time. In particular, if
it is necessary to vary the temperature during the coating process, the processing
time increases dramatically as such processes are inherently slow (see previous
DEPOSITION TECHNOLOGY 33

section on heating and cooling in a vacuum). Finally, because cooling rates can
have a significant effect on yields (cooling glass is much more likely to cause
breakage, and broken parts cannot be sold as profitably as whole filters), the coating
temperature and substrate type are additional cost considerations. In summary, it
is not possible to judge the amount of time that may be required to coat a substrate
without intimate knowledge of the actual process that is used.

2.2. COATING PARAMETERS

2.2.1. Source Characteristics

The choice of coating technology depends to a great extent of the behavior of the
coating material. Some materials may decompose when evaporated, whereas others,
such as magnesium fluoride, decompose in a typical sputtering system. Thus, the
selection of the appropriate coating system is dependent on materials related char-
acteristics.

2.2.1.1. Compositional Variations. If the charge in an evaporation system

consists of more than one component, it is unlikely that the two components will
evaporate at the same rate at a given temperature. In that event, one component
will evaporate faster than the other until the concentration of the two components
is such that they evaporate at the same rate. This composition is called a congruently
evaporating one, and the composition of evaporant stream is stable over time until
one of the components is exhausted or the charge is changed. The mixture can
consist of a single generic compound, which has several stable phases, e.g., the
titanium oxide system (Samsonov, 1978), or of two materials that form a nonreacting
mixture, e.g., a mixture of zirconium and tantalum oxides.
A coating source, the composition of which changes with time, leads to a
deposited layer that is inhomogeneous. Such a coating will not perform according
to theoretical calculations based on uniform, homogeneous layers.

2.2.1.2. Stoichiometry. Many oxide materials lose a certain amount of their

oxygen during deposition, and this is especially true in systems in which the charge
is heated with an electron beam. This is not surprising based on the fact that the
chemical binding energy is of the order of a few electron volts, whereas the electron
beam that is supplying the evaporation energy is accelerated by a potential of 5 to
10 kV. It is common practice to admit some oxygen to the coating chamber during
the deposition of these materials to ensure that the deposit approaches stoichiometry.
The oxidation reaction occurs at the substrate surface rather than in the gas phase.
This can be readily understood since the mean free path in a vacuum chamber is
on the order of a meter. Thus it is quite unlikely that a sufficient number of the
evaporant atoms and molecules will strike an oxygen atom before hitting the sub-
strate surface.
For certain materials it is important to obtain good stoichiometry, whereas for
others it is much less important. For example, a slight loss of oxygen in silicon
34 CHAPTER 2

dioxide is generally not a serious problem, because the next lower stable oxide,
SiD, is also rather transparent over a wide wavelength range. On the other hand,
only a very slight oxygen loss from Ti02 results in an absorbing coating. Thus, in
this latter case, it is necessary to pay particular attention to be certain that the titania
is completely oxidized on the substrate.

2.2.1.3. Coating Distribution. The distribution of the evaporant cloud is

more important in an evaporation system than it is in a sputtering system. The
proximity of the target to the substrate averages out most deviations in the distri-
bution in the sputtering process. The large source to substrate distances in evapo-
ration systems, however, make the distribution a concern to the operator of such
a system. Factors that can cause the distribution to change as a function of time
include the level to which the source has been loaded, the gas pressure in the
chamber, and the way in which the evaporation charge is consumed. A change in
distribution can affect the uniformity of the deposit on the substrate, and this in
turn will affect the performance of the resulting filter.

2.2.2. Deposition Environment

The properties of a coating are strongly affected by the temperature of the substrate
during deposition. The crystallinity, intrinsic stress, and transparency of a film are
all functions of the coating temperature. Dielectric films are generally more durable
when coated at higher temperatures.
The addition of a small amount of oxygen into the coating chamber during
the deposition of oxide materials helps to ensure that they are fully oxidized, given
the other process parameters. Suboxides are frequently absorbing, especially near
their short wavelength absorption edges, which frequently occur in the visible
portion of the spectrum. Oxidation reactions typically proceed more readily and
completely at elevated temperatures, therefore, raising the coating temperature
reduces the level of oxygen partial pressure needed to drive the oxidation reaction
to completion. This is desirable as a lower background pressure yields a more
durable deposit.
When the substrate is a common optical glass or a material that can tolerate
high temperatures, there is generally no problem with coating them at elevated
temperatures on the order of or greater than 100°C. Some substrates, however,
cannot be heated very much: plastic substrates can decompose, melt or soften, or
outgas some of their constituents and change their properties. Tempered metal or
glass substrates the surface of which has been accurately figured to a diffraction
limit can warp out of tolerance when heated. For such substrates, only limited
heating can be used, and the accompanying loss in durability is a trade-off that has
to be addressed by the system designer.
Outgasing, the release of gases into a coating chamber either before or during
the coating process, can occur whether or not the substrate is heated. Substrates
can have cavities that trap air, which then slowly leaks out. If the leakage occurs
during the coating process, the gases can disrupt the coating process and cause the
deposited film to be soft, scattering, or otherwise degraded. A blind screw hole is
DEPOSITION TECHNOLOGY 35

one common source of outgasing. This type of gas source is often called a virtual
leak. There are standard design techniques that can alleviate these problems, but
there are no definitive texts currently in print on the subject, though O'Hanlon's
(1980) text and Maissel and GIang (1970) do give some guidelines.

2.2.3. Nucleation and Growth

During the deposition process, films do not always grow uniformly, monolayer by
monolayer, especially in their earliest growth stages. Under some deposition con-
ditions, metallic deposits can form tiny islands in their initial stages of growth
rather than a very thin continuous layer. The atoms arriving at the surface can be
mobile enough to move about on the surface until they encounter a suitable nu-
cleation site. Such a site may be either directly on the substrate or on a part of the
growing film. Eventually, the individual islands become large enough to touch one
another and the layer becomes continuous.
The result of islands forming and then coalescing is an index of refraction that
is dependent on thickness (in this thickness range). Gold and silver are two metals
with well-known thickness dependencies of their complex index of refraction when
they are in the 10 to 200 A thick range.
A similar phenomenon of thickness dependent refractive index occurs with
some dielectrics, but the effect in these materials occurs at larger thicknesses. For
instance, it has been recently reported (Klinger and Camiglia, 1985) that the crys-
talline structure of an evaporated zirconium dioxide film changes from cubic to
monoclinic as it grows beyond 70 nm in physical thickness. The amount of cubic
phase does not change indicating that only the newly arriving material contributes
to the increasing monoclinic phase, whereas the original cubic phase remains stable.
They show that this accounts for the inhomogeneity observed in films of this
material.

2.2.4. Rate and Thickness Monitoring

In a coating process, both the rate of deposition as well as the total thickness of
the coating need to be monitored and controlled. There are several types of mon-
itoring techniques that do this: (1) A quartz crystal oscillator measures the mass of
a deposit as a function of time. (2) An optical monitor measures changes in the
transmission or reflection of a part or witness exposed to the coating flux. (3)
Atomic emission spectroscopy measures the intensity of the light emitted by excited
evaporant atoms.

2.2.4.1. Crystal Oscillator. The most common monitoring technique that

measures both rate and thickness simultaneously is the quartz crystal oscillator. In
this device, an oscillating quartz crystal in a water-cooled holder is exposed to the
flux of material that is coating the substrate. The mass of material that deposits on
the crystal changes its resonant frequency. This change is sensed by an electronics
package and is converted to a display of deposition rate and coating thickness. It
is important to note that this technique produces an output that is a function of the
mass of the deposit, but it provides no optical thickness information.
36 CHAPTER 2

If the deposit on the crystal has excessive stress, the calibration factors of the
crystal can change and cause an error in the output value. The deposit can crack
or delaminate as a result of this stress and an abrupt change in crystal frequency
occurs, leading to an indetenninate result. Modem electronics packages that control
the oscillator can detect an abrupt change in oscillator frequency. Several means
have been developed to get around these problems when the deposition process is
underway and cannot be interrupted:
I. A second crystal can be employed; it is positioned adjacent to the first and
shielded from the coating material by a shutter. At the first indication of trouble,
the electronics in the control package activate a shutter that uncovers the second
crystal.
2. A way of extending the life of a crystal is to reduce the rate of buildup of
thick deposits on the crystal by using a shutter or screen in front of the crystal.
Only a known fraction of the full flux is sensed, thus reducing the amount of
material deposited on the crystal. A shutter that is open intennittently can consist
of a rotating slotted wheel or a blade shutter that opens momentarily from time to
time to sample the rate. It is assumed in these sampling schemes that the source is
sufficiently stable that sampling gives an accurate picture of the total amount of
material deposited and the rate at which this takes place.
Crystal oscillator systems are generally used to monitor metal deposition processes.
There are several reasons for this: (1) Metals are highly absorbing and most metal
films are opaque, therefore, standard optical techniques are not useful except for
very thin films. (2) Metal films are typically not as stressy as dielectric films,
therefore, greater total thicknesses can be deposited on the crystal before it fails.
A quartz crystal can be used equally well with a sputtering or an evaporation
system. In fact, a quartz crystal system is ideal for a sputtering system that has a
large target and close spacing of the substrate to the target. Special crystal holders
are available that are optimized for each type of deposition system.
Sputtering is a very well-controlled process, therefore, in a production envi-
ronment, the thickness of a deposit can be detennined or controlled on the basis
of the target power input and the time of deposition. This is fortunate, as it is rather
difficult to monitor optically the growth of a sputtered film owing to the geometrical
constraints of the close proximity of the target and the substrate. If the coating is
done intennittently by moving either the substrate or the target, then it is possible
to monitor the transmission or reflection between the coating passes.
Standard quartz crystal oscillator systems are available commercially from a
number of suppliers.

2.2.4.2. Optical Techniques. Optical techniques are generally preferred when

coating dielectric films for precision optical designs. The most common situation
where such techniques are useful is in the construction of high reflector coatings
that consists of a stack of layers of alternating index, each layer's being one quarter
wave thick. If one monitors the reflectance from the substrate at the wavelength at
DEPOSITION TECHNOLOGY 37

which the highest reflectance is desired and tenninates the coating of a layer when
the reflectance reaches a peak value, then an error made in one layer will be
automatically compensated in the next layer. This applies to small to moderate
errors, and, of course, is not a substitute for good coating techniques.
Reflection or transmission optical monitoring is possible, and commercial units
are available off the shelf to make these types of measurements. Figure 2.3 shows
the optical layout of such systems. These systems are very useful when one is
coating a stationary part. If the part(s) are in single rotation in the chamber, then
it is still possible to monitor the transmission or the reflection as a given part follows
a relatively simple circular path. This is especially convenient if the part is large
and the geometry is such that a stationary beam is always reflected off the coated
surface as the part(s) rotates. If double rotation or planetary systems are used to
simulate a random motion of the substrate with respect to the sources, then it is

Transmission Detector

Vacuum Window

t:::::::::::1-------f--Monltor Slide

Vacuum Chamber

Evaporation source

Reflectance Detector
Figure 2.3. Optical schematic diagram of reflection and transmission monitors for vacuum
evaporation. A number of variations on this basic arrangement are possible if the coating
machine geometry does not lend itself to this arrangement.
38 CHAPTER 2

difficult to measure the parts directly. In such cases, the optical properties of a
stationary witness part must be employed.

2.2.4.3. Atomic Emission Techniques. A system is commercially available

that monitors the rate in a noncontacting way, and therefore, can run indefinitely.
It is based on ionizing a portion to the evaporating flux of material with a beam of
electrons. The electrons are directed perpendicularly to the direction of the flux,
and at right angles to both of these, a detector evaluates the light emitted by the
excited atoms and molecules in the flux. The emitted light has the characteristic
spectral lines of all of the materials that are being coated. A filter selects the line
that represents the material of interest. The process is excellent for metals, but can
be problematical when compounds, especially oxides, are being deposited in a gas
reaction process. The band spectra obtained from the excited gas molecules obscure
the elemental lines, thus reducing the sensitivity to too Iowa level to be practical.
Further development will probably improve this technique further.

2.2.5. Sticking Coefficient

The sticking coefficient is defined as the fraction of the incident molecules from
the source that actually adhere to the substrate. This coefficient is very dependent
on the material, the temperature of the substrate, the nature of the surface, and so
on. Some materials actually inhibit almost all molecules from sticking to a surface.
It is interesting to speculate what are the dynamic parameters, e.g., kinetic energy,
or velocity vector, after an atom or molecule strikes a surface but does not adhere
to it. For example, is the atom or molecule that strikes the wall once thermalized?
One such material that makes surfaces nonadherent is available commercially
as "Hi Torr"@. Such materials are useful in situations where the coating material
is very valuable, such as gold, and one desires that it coat only on certain areas.
This aids in prolonging the time between vacuum chamber cleanups, as well as
reducing coating material costs. The commercially available material is a fluorinated
hydrocarbon. Unfortunately it is not perfected to the point where it can be used in
chambers where the coating is done at elevated temperatures.
For some materials, the sticking coefficient can be strongly dependent on the
temperature of the substrate. For example, the sticking coefficient of ZnS, when
the substrate is in the range of 150 to 200°C, varies markedly. Small variations in
the substrate temperature produces large variations in coating thickness. At tem-
peratures well below 150°C, the sticking coefficient is close to unity, whereas at
temperatures well above 200°C, it is nearly zero. Temperatures near 200°C yield
different crystalline forms of zinc sulfide, thus there is interest in coating in this
temperature regime, but very careful attention has to be given to temperature control
and the uniformity of the temperature over the entire surface to be coated. Zinc
selenide is another material that shows this sort of behavior. Undoubtedly, other
materials exhibit similar characteristics in some temperature ranges; however, most
DEPOSITION TECHNOLOGY 39

oxides and fluorides, the most common of the coating materials, are quite well
behaved in typical coating temperature ranges.

2.3. COATING EQUIPMENT

Vacuum coaters use a variety of coating equipment to manufacture thin film filters.
The common types of equipment are presented here from the filter user's standpoint.
In particular, we describe the implications of the size of the equipment on filter
performance and cost, and the effect of the substrate support structures in the vacuum
chamber on useful filter aperture.

2.3.1. Size of Coating Chamber

A major factor in the unit cost of producing optical filters is the quantity of parts
in a coating lot. This is no different than many other commercially produced items.
When a large quantity of one item is desired, then the cost may be reduced, given
appropriate capital investment. Thus, the equipment size and the order size need
to be matched. Figure 2.4 shows the unit cost of a standard product produced over
a number of years and in different volumes (Seddon, 1977). Note that a mismatch
in equipment size to order quantity results in an increase in unit cost.
The size of commercially available vacuum chambers varies from 30-cm di-

Learning Curves for Front Surface Mirrors

ID
ID
CD

.o
U

D- 72"
III
C $10 Batch Coater
;::
III
o
o
120"

.....
E
o Carousel
Coater
CD
:::I $1
iii Continuous
...':: ..........
>
........... ~%
Coater

!
"a
............
< I
Acre
150 Acres

10 4 10 5
Cumulative Production (square feet)
Figure 2.4. Learning curves for evaporated front surface mirrors. Each piece of equipment
has a slope of about 80 percent, whereas the overall curve has a 70 percent slope. This
means the the costs drops by 20 percent each time the volume doubles. Note that it is
important to use properly sized equipment for a given size lot-both too small or too large
a chamber will increase costs over the optimum machine. (After Seddon, 1977).
40 CHAPTER 2

ameter glass bell jars used in research facilities and teaching laboratories, to be-
hemoth chambers that are connected in tandem with large glass-making machines
and are used to apply architectural coatings continuously to large glass panes. In
between these extremes, one finds that production chambers are typically in the
0.5- to 1.5-m diameter range. These larger evaporation chambers are frequently
made of stainless steel and are customized for specific applications. Their geometry
may be horizontal or vertical cylinders, or boxlike with fiat sides. Sputtering cham-
bers may have two of their dimensions similar to that of a comparable evaporation
unit, but the third dimension corresponding to the target to substrate dimension
may be much smaller.
Generally, one does not build a production coating chamber to handle both
evaporative processes and sputtering at the same time. The reason for this is that
the geometry of the two techniques is not very compatible.
Nonvacuum processes (see section 2.1.4.) use tanks or vats that correspond
to the size of substrate being coated. The capital cost for this equipment is far less
than that of comparably sized vacuum systems.

2.3.2. Chamber Geometry and Process Considerations

The size of a vacuum coating chamber can have an important effect on several
coating parameters. The surface to volume ratio can affect the residual gas load
and composition and the temperature reached during coating. The distribution of
coating on the part, and the fraction of evaporated material that actually coats the
substrate is related to the size of the chamber with respect to the size of the substrate
or the coating area.

2.3.2.1. Surface to Volume Ratio. The quality of the coating is influenced

by the residual gases in the chamber during the coating process. The source of
these background gases may be intentional, e.g., oxygen is metered into the chamber
to control the stoichiometry of oxides, or it may be unintentional such as leaks or
internal sources. The latter sources of residual gases include desorption of gases
on the chamber walls and other surfaces. The wall area increases as the square of
the average dimension of the chamber, but the volume increases as the cube (an r
vs. an,-3 relationship). If we assume a constant gas evolution per unit area, we find
that the gas load increases with the size of the chamber. The volume that contains
these gases, however, increases even faster yet with the chamber size. Thus, based
on the very simplified model of volume and surface area considerations, larger
chambers are less prone to problems with walJ outgasing. The actual chamber
pressure is determined by a number of other parameters including time and the
pumping capacity on the vacuum system.
The amount of residual gases in a chamber can be reduced by heating the
chamber and its contents during the pumpdown before a coating run to drive off
the adsorbed gases. This process is called a bake.
Considering the upper limit of coater size, one would like an infinitely large
chamber to coat a small part. On the other hand, it would take an infinite amount
DEPOSITION TECHNOLOGY 41

of time to pump out such a volume. The worst case is a chamber just large enough
to hold a part of a given size. Then, the pumping time to base vacuum is very
short; however, unless the chamber were well baked out prior to the start of dep-
osition, the outgasing from the walls and the substrate as the coating process heats
up the chamber may be quite severe.
The preceding remarks do not apply to continuous coating systems, either
evaporation or sputtering, where the coating chamber is not exposed to atmospheric
pressure over the course of days or longer.

2.3.2.2. Thickness Uniformity and Coating Efficiency. The coating uni-

formity obtainable with an evaporation system is dependent on a number of geo-
metric considerations, including the ratio of the size of the area to be coated to the
distance between the source and the substrate. The distribution of the material
leaving an evaporation source can be approximated as a point source with an angular
dependence that closely follows a cosine function with an exponent in the 2 to 4
range, the exact value of which depends on the source type, the background gas
pressure, the source power level, and other parameters. Figure 2.5 shows the
distribution for three values of the power. A small substrate area subtends a smaller
part of this distribution. Preferably, the part is located near the peak of the distri-
bution where the slope is the smallest and there is the least variability in thickness
over the part.
The variation of the coating flux in an evaporator can frequently be modeled
in a manner similar to that used for light intensity calculations in an optical system:
the drop-off in thickness is proportional to cos4 6: one cosine factor comes from
the source distribution (Lambertian), two come from the 1/,-2 increase in distance
between the source and the substrate, and the last factor comes from the change in
the subtended surface area at the substrate. Tilting the substrate so as to make the
surface normal to the flux arrival direction eliminates one of the cosine factors. The
thickness uniformity can be further improved by introducing relative motion between
the substrate and the coating source.
A trade-off must be made between the active coating surface, i.e., the average
substrate area being coated and the coating distance. As a larger fraction of the
distribution, shown in Fig. 2.5, is intercepted by the substrate, the better the
utilization of the coating material. Much of the material evaporated is deposited on
the walls of the chamber. This material is not only wasted, but it contributes to an
increase in the frequency of cleanups of the chamber. The material that coats the
walls is generally porous because the walls are cool and the angles of arrival of
the evaporant are poor. Invariably, the wall deposits are hygroscopic because of
the large surface area and capillaries afforded by the many pores. Thus it serves
as a source of moisture during the pump down and coating procedures. This is
exacerbated in that the walls generally warm during coating, and this tends to drive
out moisture at the worse possible time. A bake before coating, if allowed by
equipment and substrate constraints, helps reduce this problem. This tends to sta-
bilize the process by driving out the adsorbed gases before the process starts. The
chamber may be cooled, if necessary, before the actual deposition starts.

42 CHAPTER 2

Normalized Coating
Flux Density

Figure 2.5. Polar plots of three normalized distribution patterns of a vapor source. The
exponent of the cosine function is indicated on the plot. As the exponent increases, the
source becomes more directional, and a larger fraction of the material deposits on the
substrate. Conversely, the area at the substrate plane over which the uniformity is at a given
level decreases as the exponent increases.

Sometimes, the only practical way to coat certain geometries is with a sputtering
process, e.g., coating the inside of a glass tube with an infrared reflective coating
(Harding et aI., 1979). Sputtering targets, by virtue of the fact that they are solid,
can be arranged to coat in any direction: up, down, or sideways. This flexibility is
not usually found in evaporative systems as many of these sources contain molten
or granular material and are restricted to coating upward. This is not to say that
evaporation sources cannot be designed to coat in directions other than vertical,
but significant source design trade-offs have to be made to obtain this feature. With
sputtering targets, none of the drawbacks apply; therefore, sputtering in the down-
ward direction is as easy as sputtering in the upward direction. One reason that
downward sputtering is not practiced too commonly in high quality thin film man-
ufacturing is that arcs occasionally develop between the target and the plasma, the
substrate, or some other conductor or source of electrical charge. Though the
electronic power supplies are designed to deal with this and extinguish the arc very
quickly, there still may be a shower of fine particles resulting from the arc. If the
substrates are lying face up, they will then be contaminated and the coating degraded
by these particles.
DEPOSITION TECHNOLOGY 43

2.3.3. Substrate Rotation for Coating Uniformity

Historically, the first thin films were made in chambers in which all parts were
stationary. Next, single rotation systems were developed to enable the coating of
larger substrates. Finally, more complex motion systems evolved as economic
production of precision coatings became the dominant factor in thin film technology.
Substrates in a coating chamber generally have a support system that provides
them with a motion with respect to the source. This is done to achieve a more
unifonn deposit over the surface of the substrate. It is conceivable that one could
move the source with respect to a stationary substrate. Even when large substrates
are involved, however, it is simpler to move the substrate rather than deal with the
high current, high voltage, or high power electrical feedthroughs and the cooling
water couplings that are necessary to operate movable evaporation or sputtering
sources.
The relative motion may simply be linear, as in coating glass sheets for ar-
chitectural purposes or the coating of webs of flexible material in a roll coating
machine. In precision coating applications, the substrates are generally rotated in
a planetary fixture. In such a geometry, the parts are mounted in racks that rotate
about one axis (planet), and this axis in tum orbits around a central axis usually
located at the center of the coating machine (sun). The ratio of the gearing that
generates the orbiting motions is chosen such that many orbits are required before
a point on the substrate retraces its path. As long as many orbits are completed
during each layer's deposition, the motion will randomize thickness variations. The
result is a much more unifonn deposit than would be obtained without the complex
motion (Macleod, 1969, Chap. 9). Because the substrate passes over the source
many times during the deposition of any given layer, it is in fact built up of a
succession of very thin deposits. This intennittent coating process has some desirable
characteristics in addition to its improving the unifonnity of the deposit. For ex-
ample, if the coating is an oxide and there is a residual gas of oxygen in the coating
chamber, then the intermittent coating process gives the deposited material a chance
to react before another layer of evaporant molecules arrives.

2.3.4. Jigging and Substrate Support

Periodically, someone will offer an off-the-cuff suggestion that a suction cup be
used to hold the substrates during coating. Such a system is flexible and avoids the
need for customized tooling. Unfortunately, in a vacuum chamber there is an even
better vacuum on the front side of the substrate to pull it off the suction cup! Thus
another means of holding parts is needed.
With a few small parts, it is straightforward to make adapters that mate to
pennanent tooling inside a coating chamber. If the parts have an irregular shape,
then it can be difficult to hold them so that the coating is applied uniformly over
the surface. If there are many parts to be coated, the making of the adapters could
become a significant fraction of the total cost of a coating task. If the quantity of
parts is large enough so that many coating runs are required to complete the task,
then special purpose tooling becomes very attractive from an economic standpoint.
44 CHAPTER 2

The addition of parts to a load usually will not change the cost of producing the
run, therefore, it is an incentive when large quantities have to be coated to spend
more resources on designing tooling that will maximize the load size.
Most part holders in coating systems that coat upward leave a small uncoated
area near the edge of the part. It is here that a small lip in the tooling actually
supports the substrate. The part of the substrate that is hidden does not get coated,
therefore it is desirable to make this zone as small as possible so that it does not
intrude into the active aperture of the optical element. There should be some means
of constraining the substrate from moving laterally in the tooling fixture to prevent
scratching of the front surface by the lip. The tolerance between the side of the
substrate and the holder must include provision for thermal expansion coefficient
differences between the tooling rack material and the substrate.
Substrates with a mass of several hundred kilograms have been coated in double
rotation systems. When the substrates become this large, then the tooling problem
becomes one for mechanical engineers. The vacuum tooling designer is faced with
a number of specialized problems. A major one centers on the tendency for two
similar metals to weld when they come in contact in a good vacuum and some
pressure is applied to them. This makes the design of bearings difficult. Lubricants
have to be selected carefully lest they contaminate the vacuum environment.
It is important to avoid deposits on surfaces that are not intended to be coated.
For example, some coating processes can deposit materials on surfaces that are not
in a line of sight to the source. This happens for a number of reasons, among which
are
1. The gas pressure during the coating process is relatively high and the
coating material is scattered around to the back of the substrate
2. There is decomposition of the source material to gases that diffuse to the
back of the part where they recombine and form a deposit
3. The source material may not stick well on the walls of the chamber, and
thus bounce off and eventually find its way back to the substrate
These unwanted coatings are typically soft, nondurable, and uncontrolled. They
must be avoided, especially if a surface that is susceptible to such coatings has any
value; e.g., it will be subsequently, or has already been, coated in another coating
operation.

2.3.5. Temperature-Related Substrate Degradation

The temperature at which a coating is applied is generally above ambient. This
temperature can have a detrimental effect on certain substrates. In precision optics,
the surfaces of which have been polished so that deviations in them away from the
desired shape are less than a fraction of a wavelength, heating can cause optical
distortions greater than the allowed tolerance. Organic substrates, i.e., plastics,
pose severe coatings problems. As previously mentioned (section 2.2.2.), relatively
low coating temperatures can cause the substrates to melt or otherwise deform
(Agranoff, 1985). In addition, the heat of condensation of the evaporant or the
radiant heat from the source in an evaporation process or the plasma in sputtering
DEPOSITION TECHNOLOGY 45

systems, can easily cause the substrate temperature to rise. This leads to warpage
or outgasing.
A large number of different kinds of plastics are used as optical devices today.
In addition, many of the compositions of these plastics vary from one manufacturer
to another. Each supplier adds his own proprietary combinations of ingredients,
including volatile components such as plasticizers and fillers. The volatile com-
ponents can interfere with the coating process as they diffuse out of the plastic. It
can be expected that the properties of the plastic will also be different after the
coating process and the plastic has lost its volatile components. This limits the
number of materials that can be coated on plastics and the temperature at which
they are laid down. As a result, it is rare to find coatings that are as hard on plastic
substrates as they are on glass and other materials where the coating temperature
can be elevated to over lOOcC.

2.4. STRESS EFFECTS

Although process heating during coating can have an effect on a substrate's optical
figure, stress is an even more significant parameter when the substrate is thin. This
section considers the sources of stress, and the effects of stress in the coating on
both the coating itself and on the substrate.
Stresses in a thin film can have several components. Intrinsic stresses develop
as the film is being formed. The growing crystal lattice is not perfect and stresses
develop as it relaxes. These stresses can be affected to some extent by the nucleation
sites available to the first few monolayers deposited, but after a few layers of atoms,
the film structure tends to some dominant crystallographic form independent of the
initial conditions. Intrinsic stresses are strong functions of the coating temperature,
coating rate, and other coating conditions. Note that no temperature change is
required to obtain intrinsic stresses. Postcoating annealing steps can sometimes
reduce the intrinsic stress in a coating.
A second major source of stress is due to the differences in the coefficient of
thermal expansion of the film and that of the substrate or adjacent layers. The stress
develops when the temperature changes from the deposition temperature. It is
difficult to avoid this type of stress when films are deposited at temperatures above
ambient. Even when films are deposited nominally at room temperature, some
heating of the substrate can occur during the deposition and condensation process.
Finally, temperature variations during use may lead to changes in the stress level.
The sign of the thermally induced stress can change from tensile to compressive,
or vice versa, as it is a function of the differences in the thermal expansion coef-
ficients of the film and substrate materials. For example, if the substrate expansion
coefficient is greater than that of a film that is compressively stressed at some given
temperature, there exists theoretically a higher temperature at which the stress in
the film becomes tensile.
Stress by itself is not necessarily bad as far as the films themselves are concerned
(warpage of the substrate is another matter altogether, however). For years, mag-
46 CHAPTER 2

nesium fluoride films, which have extremely high stress levels (see Table 2.2),
have been used very successfully in forming visible quarter wave thick antireflection
coatings on glass surfaces. The reason for this success is that the adhesion of the
film to the substrate is excellent and the substrates are strong in compression. As
a result, enough of the tensile stress in the film is transferred to a compressive
stress in the surface of the massive substrate that the cohesive strength of the film
is not exceeded. If a magnesium fluoride film is made thick enough, however, the
stresses in the film become high enough to overcome the film's cohesive strength.
A number of fine cracks forms to relieve the tensile stresses; a film with these fine
cracks is said to be crazed. This happens with magnesium fluoride at quarter wave
optical thicknesses greater than about I.S ~m, thus precluding its use as an infrared
coating material.
In some instances, a film's cohesive strength and its adhesive strength to a
substrate can be strong enough to exceed the cohesive strength of the bulk substrate
material. For example, the stresses in a tantala-silica stack on fused silica can be
made large enough so that as the film fails in tension, it tears off small particles
of the silica surface, leaving behind a rough, pitted surface.
The failure mechanism of films which are under compressive stress is delam-
ination of the film from the substructure. The adhesive strength of the film-substrate
bond is the dominant parameter in preventing failure of a film in compression; the
cohesive strength of the film is not a factor.
To first order, the contribution of a material to the total stress in a coating is
proportional to the product of the total thickness and the stress per unit thickness
of the material. Coatings that consist of stacks of periodically repeating assemblies
of layers (periods) have a total stress level proportional to the number of periods
and also to the net stress in a typical period. The net stress of a period depends on
the sum of the stresses (tensile and compressive) in the individual layers.
In a period with a pair of layers, one of which has tensile stress and the other,
compressive stress, there will be only one thickness ratio that will yield a net stress
of zero. The balancing of stresses is only one of the considerations a thin film
designer must take into account in arriving at a producible final design. Although
optical considerations may dictate a particular thickness ratio, stress considerations
may suggest a different ratio of material thicknesses.
The majority of dielectric materials have tensile stress when deposited as a
thin film in conventional processes (simple evaporation and sputtering). The stress
levels of these films have of late been altered by new techniques such as ion beam
bombardment of an evaporated film or ion beam sputtering (see Chap. 7). Fortu-
nately for the thin film engineer, some materials have compressive stresses also.
These materials can be used to compensate for tensile stresses in an adjacent layer.
A typical material pair that has a good net balance between stresses when both
layers have approximately equal optical thicknesses is ZnS and ThF4 (Ennos, 1966).
Another useful material with compressive stress is silicon dioxide. It is the lowest
index oxide material commonly used for durable coatings, and is used to compensate
for the tensile stresses found in the high index materials that are frequently used
in conjunction with it, such as titanium dioxide and zirconium dioxide. There seems
DEPOSITION TECHNOLOGY 47

to be no relation between the stress type (compressive or tensile) and the index of
refraction.
The cohesive and adhesive strengths of thin film materials are not well known
quantitatively. Apart from the difficulty of measuring these quantities for thin film
specimens, the values of these parameters can vary with time as moisture penetrates
the films and as the internal forces relax. Also, as previously mentioned, the internal
stresses and strengths are strong functions of deposition conditions. What is needed
is a complete study of this problem so that it can be reduced to a more firm
engineering practice.
Ennos (1966) took a first step in this direction by publishing a study he made
of in situ stresses. He observed the deflections of a thin substrate while various
films were being deposited in an evaporator. He was able to deduce the stress levels
in films of various materials from his data, and these are summarized in Table 2.2.
In a few instances, a coating with crazing is acceptable. The majority of these
cases involve optical systems that are only collecting energy and are not image-
forming systems. In an energy-collecting system, for example, the key requirement
is to gather energy, either reflected or transmitted and to deposit it on some receiver.
A coating that is crazed loses a small fraction of the incident energy to scattering,
but this lost energy is sometimes insignificant in the overall system performance,
especially when the benefits of the coating are taken into account. The total surface
area occupied by the cracks is very small in comparison to the total surface area
covered by the coating. Diffraction at the edges of the cracks leads to scattering.
An example of a coating in which the crazing is tolerable is in the energy conserving
coating applied to the fused silica envelopes of tungsten-halogen lamps (Rancourt
and Martin, 1986). The scattered light is still useful in lighting applications, and
the coating performs its task of reflecting the infrared radiation back onto the
filament. Other similar applications exist in solar energy collection systems.
Crazed films cannot be used in some designs at all; for example, they could
not be used in an antireflection (AR) coating. An AR coating is intended to reduce

Table 2.2 Measured Stresses In Coated Films

Stress
Thickness

Material (angstroms) (kg/cm2) (lb/lnch 2) (Pa)

Single Layers
Zns 675 +1800 +25,600 +1.77 x 106
Cryolite 1170 -250 -3,500 -2.41 x 107
ThOF 2 1090 -1600 -22,800 -1.57 x 106
PbF2 910 +51 +725 +5.00 x 106
MgF2 2500 -4500 -64,000 -4.43 x 106
Mu"ilayers
ZnS/Cryolite +900 + 12,800 +8.83 x 107
ZnSIThOF2 +100 +1,400 +9.66 x 106
PbF2/Cryolite -800 -11,400 -7.86 x 107
Note: Pos~ive values indicate compressive stress. Negative values indicate tensile stress.
From: Ennos, 1966
48 CHAPTER 2

the reflected light from a surface; a scattering coating will tend to do just the opposite
by increasing the amount of light coming from it.
Stress relaxation by deformation of the substrate occurs when the substrate is
thin enough or flexible enough to bend under the net force produced by the coating.
If the substrate is a thin polymer film, for example, this can make the substrate
curl to the point of rendering it useless. If the substrate is a solid, such as a thin
glass or silicon wafer, the warpage may lead to breakage, or if an optical element,
to unacceptable distortions. Assuming that the coating is thin with respect to the
substrate thickness, the deflection of a flat plate substrate with parallel surfaces can
be approximated by the following formula (GIang et al., 1965):
& = 3[(1- v)/E] (tit;) p2 (J"

where & = the deflection at a distance p, p = the distance from the center of a
circular disk, v = Poisson's ratio, E = Young's modulus, tf = thickness
of the film, ts = thickness of the substrate and, (J" = film stress with di-
mensions of force per unit length.
This formula can be used to estimate the stress in a coating by measuring the
deflection of a substrate of a known material, such as silicon or silicon dioxide.
Once the stress in a film is known, the deflection of other substrates can be estimated.
The sense of the curvature of a substrate can be used to determine whether a
coating has tensile or compressive stress as shown in Fig. 2.6. Stress failures are
easily classified as being caused by tensile or compressive stresses with a micro-
scopic examination of the coating (see section 5.4.7).

2.5. SUBSTRATE CLEANING

Cleaning is one of the most difficult problems in the manufacture of low reflectance
coatings, and it is a prime problem with other coating designs as well. One of the
reasons that cleaning is so difficult is that it is dependent on all of the prior history
of the part: at all stages of handling more dirt, grime, and organics get added to
the surface. If cleaning is not considered from the beginning of the fabrication of

Tensile Compressive
convex concave
Substrate

Concave Coating
Convex
Figure 2.6. Curvature induced by compressive and tensile coatings. if the substrate is thin
or flexible enough, the direction of bowing can be determined with the unaided eye. It is not
always easy. however, to determine which side is coated.
DEPOSITION TECHNOLOGY 49

a part, either on an optical polishing machine, a diamond-turning lathe, or an

injection molding machine for plastics, then it may never be possible to clean the
surfaces properly at a later time. A good example of where this has been found to
be highly important is in the fabrication of diamond-turned metal optics. In this
machining operation, the temperature must be precisely controlled to obtain the
high level of accuracy desired, therefore, an oil shower is used to bathe all of the
work and the tool in a constant temperature fluid. If this oil is not cleaned off the
substrate within minutes of the completion of the turning operation, the oil somehow
combines chemically with the fresh metal surface, and it then becomes impossible
to remove (Buckmelter et aI., 1975). Because the diamond-turned surface is ex-
tremely fragile, contact methods cannot be used to scrub the contaminant off the
optics.
Some optical glasses are rather prone to staining as indicated in the standard
catalogs of optical glasses (e.g., Schott Glass catalog). The cleaning of the glasses
classified by the Schott Glass Company as "Group 5" can be problematical. These
glasses are stained by an exposure of less than 12 minutes to a test solution buffered
at pH 5.6 (only slightly acidic).
The common vigorous cleaning techniques can be divided into several classes:
l. Contacting: Physical contact is made with the surface, including scrubbing
with or without abrasive materials, either manually or automatically.
2. Ultrasonic: The substrate is placed in a cleaning solution that is agitated
with ultrasonic waves.
3. High pressure spraying: A stream of high pressure water or other solvent
is directed at the substrate to dislodge particulates.
4. In situ: Vacuum processes, such as glow discharges or sputter etching,
are used before starting the actual deposition of the coating.

2.5.1. Contacting Cleaning

Scrubbing is the oldest cleaning method. In this technique, a solid object actually
touches the surface that is to be coated. The "solid" object may be as soft as a
cotton ball or as hard as an abrasive particle. When cleaning a diamond-turned
metal surface, even a wet cotton ball is "hard" and can damage the surface. Thus,
"hard" is a relative matter. The more common means of contact cleaning is with
brushes, as are found in automated scrubbers that find use in washing flat glass
sheets. Sometimes an abrasive, such as rouge, is used to enhance the action of the
brushes. The choice of these methods depends on:
l. The state of cleanliness of the incoming substrates
2. The fragility of the substrate surface
3. The strength of the substrate
Sometimes, a technique known as tissue drag is used to clean a surface prior to
coating. The technique is a cross between the contacting methods and the fluid ones
in that the contact between the surface and the cleaning agent is for the most part
a fluid one. The process consists of soaking a lens tissue in acetone or other purified
50 CHAPTER 2

solvent and then dragging the tissue over the surface of the part (Stowers and Patton,
1978).

2.5.2. Ultrasonic Cleaning

Ultrasonic cleaners do a good job of removing most particulate matter from surfaces
as long as they are not too tightly bound to it. A danger exists, however, with this
approach in that if the substrates are left in the active cleaning tanks too long, the
substrate can suffer surface damage in the form of micropits. These defects both
reduce the specular transmission as well as increase the stray light. The solutions
used in ultrasonic cleaning usually contain surfactants to accelerate the process and
to keep the particulates in suspension in the solution. A good review of this technique
has recently been presented by Guenther and Enssle (1986).

2.5.3. High Pressure Sprays

High pressure sprays were originally developed to clean silicon wafers for integrated
circuit production. The Lawrence Livermore National Laboratory applied the tech-
nique to the cleaning of metal and glass optical surfaces (Stowers, 1978). The
technique appears to be ideal for fragile surfaces that cannot be otherwise cleaned.
The cleaning fluid can be a fluorinated hydrocarbon, water, or other suitable solvent.
The pressures u5ed are of the order of 7 x 106 Pa (70 atmospheres or 1,000 lbl
inch 2). Particles with diameters down to 5 1J..m can be removed with this technique.
Such a process is rather complicated and is used only when less complex means
do not work. It should be noted that the fluid remaining on the surface after the
cleaning step needs to be removed without its being allowed to dry on the surface.

2.5.4. Drying
In all cleaning processes where solvents or water are used as the cleaning fluid,
the most important step is likely to be the one where the remaining liquid is removed
from the surface after the cleaning process is complete. Unless the liquid is removed
before it can evaporate, stains known as water marks form where a pool of liquid
remained and evaporated. These cleaning-induced stains are most difficult to re-
move, and likely will not be removed by additional cleaning with the same technique
that put them there in the first place. A high pressure jet of a clean gas, such as
oil-free nitrogen or air, blows the fluid off the surface rapidly; this is a commonly
used technique. The use of a contacting method, such as wiping with a paper or
cloth towel, is not recommended in that the wiping process tends to move dirt from
one area to another.
A vapor degreaser is often used to dry parts. Figure 2.7 shows a schematic
diagram of the equipment used for this. The solvent in a closed (but not sealed)
container boils and fills the container with its vapors. The substrate to be dried is
placed in these vapors, which then condense on it as the substrate initially is cooler
than the boiling point of the solvent. The condensate gives the surface a final rinse
with a pure solvent that has just been distilled. Once the substrate has warmed to
the temperature of the vapors, no more condensation can occur, the part dries, and

DEPOSITION TECHNOLOGY 51

WATER COOLING COILS

TO BE
CLEANED/DRIED

HOT PLATE

Figure 2.7. Vapor degreaser schematic. As cool parts are placed in the hot vapors, freshly
vaporized solvent condenses on the parts. This gives the parts a final rinse. The solvent
may be either an organic or a fluorinated hydrocarbon. If a flammable or toxic material is
used, provisions must be made for adequate venting of escaping vapors.

can then be removed. Solvents typically used are trichloroethane, fluorinated hy-
drocarbons, and alcohols. Naturally, some of these solvents are flammable, espe-
cially when hot, and/or toxic. Care must be exercised when using them. Adequate
ventilation is a must to prevent the accumulation of fumes.

2.5.5. In Situ Processes

There are several in situ techniques to clean surfaces. These are usually used
regardless of what technique was used to clean the substrates before they were
loaded into the vacuum chamber.
In sputtering systems, one frequently has the capability to sputter etch the
surface of the substrate. In essence, this is a reverse of the normal sputtering process,
where the substrate is made to be the target and a small amount of the surface is
removed (on the order of a few atomic layers). This leaves an atomically clean
surface to which the coating can adhere.
In evaporation systems, the common in situ means of cleaning substrates is
the glow discharge. After the chamber has been evacuated to as good a vacuum as
possible in a reasonable amount of time (called the base vacuum), a gas is introduced
into the chamber to raise the pressure in it to about 10-3 mbar. A high voltage,
ranging from about 0.5 to 5 kV, is applied to electrodes in the chamber. The voltage
causes the gas to break down and form a glow discharge. This results in a flood
52 CHAPTER 2

of excited atoms and ions in the chamber. The gas frequently used to form the
discharge is oxygen because it will remove organic contaminants on the surface of
the substrate by oxidizing them. When substrates can be damaged by the oxygen
plasma, another gas, such as argon, is used. The argon atoms bombard the surface
and can remove some material in a manner analogous to sputtering, except that in
a glow discharge cleaning step the material removal process is much less efficient
than a sputter etch one. A glow discharge can heat a coating chamber (Maissel and
GIang, 1970, pp. 6-41; Macleod, 1969, p. 229). If extensive cleaning is done with
this technique, cooling periods may need to be included in the process, lest the
substrate overheat. The more heat sensitive are the substrates, the more care that
needs to be exercised in this type of cleaning process, because the heating can be
nonuniform.
The above-mentioned techniques are frequently used in combination with one
another. The choice depends on the initial cleanliness, on the ultimate cleanliness
required, and on the characteristics of the substrate material and finish. The sub-
strates remain clean for only a short while. For example, a glow discharge is typically
repeated if the coating process does not start within a few minutes (Cox and Hass,
1958). Guenther and Enssle (1986) report that a clean surface remains that way for
less than 3 hours, regardless of how well it is stored.
 Chapter 3

Performance of Optical Filters

3.1. REFLECTANCE, TRANSMITIANCE, AND ABSORPTANCE

Conservation of energy requires an accounting of all of the energy in the incident

light beam. Generally, a consideration of the reflectance and transmission intensities
and of the absorption accounts for most of the incident light. In some instances,
scattering also plays a role; this is not considered here, but is covered in the next
section. If there is no absorption in the spectral range of interest, then reflectance
and transmittance must equal unity, or 100 percent if the values are given as
percentages. In this text, we deal with the ratio form of these values in equations,
unless otherwise noted, and the percentage form is used in the narrative because
of its clearer meaning.
The mathematical derivations of the equations governing the behavior of thin
film interference filters indicate that the transmission of a coating is independent
of the direction of the arrival of the light (Heavens, 1965, p. 77). This is well
established in practice, and it simplifies the measurement of transmission, as the
side of incidence of the filter is unimportant.
There may be reasons, however, for selecting one side of the substrate over
the other on which to locate a nonabsorbing coating. The most common reason for
selecting an inner surface of an optical system is to afford the coating some protection

53
54 CHAPTER 3

from environmental hazards. There may be optical reasons in special circumstances

that dictate which surface is most desirable for a coating. For example, where high
standing wave electric fields may be encountered, i.e., a high-power laser system,
the system designer would prefer to have the electric fields distributed to minimize
the chance for dielectric breakdown. A high-power pulse arriving from one direction
produces damage at a different threshold than a pulse arriving from the other
direction. The actual fields need to be considered in detail to arrive at a valid
conclusion to the question of which side is better (see section 3.5). Such nonlinear
effects do not ordinarily arise with most systems but do need to be considered when
high-power lasers are used.
Nonabsorbing filters are sometimes called dielectric filters to distinguish them
from metallic filters, which are absorbing. It should be pointed out that the term
nonabsorbing is applicable to a restricted spectral region, as all materials are ab-
sorbing at some wavelength. Thus, the term nonabsorbing implies, either implicitly
or explicitly, a wavelength range over which the material is completely transparent.
Nonabsorbing is also a relative term. In high-power laser systems, absorption
on the order of a few tens or hundreds of parts per million may be significant,
whereas in ordinary systems, absorption levels several orders of magnitude higher
than this would barely be measurable, and the level of significance could be another
order of magnitude higher than this. Thus, we must be careful in defining our terms
to be certain that the reader or the listener is calibrated to the same context and
significance level as the originator of the information.
In filter designs with absorbing layers, the reflectance can be asymmetric if
the design is asymmetric. In other words, the reflectance from one side is not equal
to that where the light is incident from the opposite side of the filter. A simple
example of an asymmetric coating is a semitransparent metal layer with a quarter
wave thick dielectric layer on one side, and that is immersed in a medium with an
.index of 1.5. The reflectances from the two sides are different as shown in Fig.
3.1. If the coating design is symmetric, and if the incident medium and the substrate
do not have equal refractive indices, then the reflectance will not be the same from
both sides of the filter. A single metal film on a glass substrate with an air incident
medium will have different reflectances from the substrate side than it will from
the air side, as shown in Fig. 1.8. One characteristic of any filter, either absorbing
or nonabsorbing, is that the transmission is independent of the direction of the
incident light. Because the sum of the reflectance, the transmittance, and the ab-
sorptance must be 100 percent, and as the transmittance is independent of the
direction of incidence of the measurement beam, we conclude that the absorption
also depends on the direction of incidence.

3.2. ABSORPTION AND SCATTERING

It is not possible from specular measurements of transmission and reflection with

a spectrophotometer to determine whether a particular sample is losing light energy
through a mechanism of absorption or one of scattering. Of course, techniques are
available that directly measure any of these parameters and thus one can obtain a
measurement on each one individually.

PERFORMANCE OF OPTICAL FILTERS 55

50~------------------------------------~

R1
40
Dielectric
CI)
u
C
...
n =2.35

...u
co 30 n =1.5 n =1.5

CI)
;:
CI)
Q: 20 R1 R2
~

10
•
Metal
n=2-2i

O+---------~--~~--~-----=~~
400 500 600 700
Wavelength (nm)
Figure 3.1. Reflectance (R 1) and reverse reflectance (R 2 ) for a metal film with a dielectric
film on it. The dielectric film (n = 2.35) thickness is a quarter wave at 400 nm, whereas the
metal layer (A == 2 - 2i) is 40 nm thick. The incident medium and substrate indices are
both equal to 1.5. The reflectance from the one side can be made much higher than that
from the other. The transmittance for this design is approximately 20 percent.

A calorimeter measures an increase in temperature and indicates the absorption

in the coating. The temperature of a thin sample substrate increases as it absorbs
energy from an intense laser beam. The temperature rise can be related to the
absorption coefficient; a certain amount of analysis is required to calibrate these
instruments to take into account the specific heat of the substrate, the heat lost by
radiation, and so forth. Further discussion of this topic is included in section 5.1.11.
A scattermeter is used to measure the scattered light, usually using a laser as
source of the incident light beam. An integrating sphere collects both the specular
and the scattered light. Alternately, detectors can scan over 4 1T steradians to get
the complete scattering function.
The bidirectional reflectance distribution function (BRDF) describes the full
four-dimensional scatter distribution function (two angles to describe the incoming
beam, and two more to describe the scattered beam). Bartell et al. (1980) discuss
the details of converting from the differential form of the mathematical function of
BRDF to the measured quantities of an experiment made with finite-sized detectors.
Total integrated scatter (TIS) is a simple number that describes the amount of
scattered light integrated over an entire hemisphere (Bennett, 1978). This value can
be related to the average size of surface microfeatures:

56 CHAPTER 3

TIS 1 - exp[ - (4 1T 8/A)2]

== (4 1T 8/A)2
where 8 is the rms height of surface micro irregularities and A is the wavelength of
the light.
The assumption is made in deriving this relation that the distribution of the
irregularities is Gaussian. It appears that this is appropriate in practice. In addition,
Bennett shows that the angle of the scattered light can be related to the spatial
frequency of the irregularities of the surface, and the intensity can be related to the
amplitude.
Scatter sites can exist in a number of locations, and each can lead to somewhat
different results (Carniglia, 1979). These sites include:
Substrate bulk (Bennett, 1978)
Substrate surface (insufficient polish)
Film interfaces
Film bulk
The correlation between the sites also is an important parameter.
Scattering can play a surprising role when a coating is used at a wavelength
not originally intended. A case in point is a coating that was designed for use at
3.8 ~m. Unbeknownst to the manufacturer, the system in which this filter was used
was aligned with the help of a 0.633-fLm laser. In one batch of filters, the random
errors normally present in a filter resulted in a filter with low scatter at the alignment
wavelength. In another batch of filters, the random errors were such that the electric
fields in the coating that existed for 0.633-~m light were peaked at some of the
internal interfaces. The result was a filter that was totally opalescent at 0.633 /-Lm,
while being perfectly satisfactory at the specification wavelength of 3.8 ~m.

3.3. BLOCKING RANGES

A filter is said to block when the transmission over a given spectral range is less
than a specified value (Fig. 3.2). Blocking can be obtained from the filter itself,
from the substrate, or from other filters. The transmission characteristics of the
substrate are frequently used to block wavelengths that are far removed from the
spectral zone of interest. It is more common to find materials that will block shorter
wavelengths while passing the longer ones. The electronic band structures of most
materials used in the visible portion of the spectrum have a forbidden energy band
gap that corresponds to a photon energy located in the near ultraviolet (UV) or
violet portion of the spectrum. Short wavelength light with an energy greater than
that of the band gap can cause electronic transitions across the gap. This results in
energy absorption of light with wavelengths shorter than this energy. Most solids
and liquids have this characteristic for some photon energy. Table 3.1 lists the gap
energy for several useful optical materials. Silicon and germanium are two semi-
conductor materials with forbidden gaps in the red or near infrared (IR); they can
block the visible portion of the spectrum while transmitting IR wavelengths (see
section 6.5. on material blocking properties for further details).

PERFORMANCE OF OPTICAL FILTERS 57

-
Q)
ta 10
U
II) (HL)10
C)

--o
Q)
1

-
C
ta
0.1

0.01 Blocking Range

at 0.01% Level

0.001

Wavelength (arb. units)

Figure 3.2. Plot of transmission of a blocking filter. The range for a 0.01 percent blocking
level is shown. The shape of this low transmission band is typical of many blocking filters,
and is slightly asymmetric when plotted on a wavelength scale.

Similarly, some protective covers for silicon solar cells used in space have
cerium oxide doping to block wavelengths shorter than 350 nm. This keeps the UV
light from damaging the cement used to affix the cover to the cell. The absorbed
energy is converted into heat. In some cases, such as low-light level situations,
this absorption is not a concern. In other cases, such as the space satellite usage
just mentioned, the absorption increases the heat load on the satellite and can reduce
the efficiency of the cell and the overall spacecraft.
To get long wavelength blocking, the characteristic (restrahl) oscillation of the
crystal lattice can be used. In the mid- to far-infrared, the wavelengths excite
collective oscillations of the lattice atoms. The vibrations absorb energy from the
incident beam and make the substrate opaque to light of that wavelength. Table
3.2 lists representative data for the long wavelength edge of some common optical
materials.
The materials in a coating can also have absorption characteristics that are
useful for blocking purposes. The absorption edges of the various coating materials
are well established. One function of a thin film designer is to make judicious
choices in this regard.
The relatively fixed absorption characteristics of natural materials limits their
use as blocking filters. In some cases, the shape of the transition from transmitting
to blocking as available naturally is inappropriate for a particular application. To
58 CHAPTER 3

Table 3.1 Values of the Energy Gap Between the

Valence and Conduction Bands of Selected Materials
at Room Temperature
Crystal Energy gap hl.m)
InSb 5.39
PbTe 4.13
InAs 3.76
PbS 3.54
Ge 1.85
Si 1.09
GaAs 0.89
CdTe 0.86
CdSe 0.71
CdS 0.51
ZnSe 0.48
Ti02 0.41
SiC 0.41
ZnS 0.34
MgF2 0.25"
C (diamond) 0.23
AI20 3 0.2
$i0 2 0.15"
"These values are approximate and were obtained from a number of sources.
Adapted; from Kittel (1966), chap. 10, table 1, with permiSSion.

get a precise definition or positioning of the transmission to blocking transition

wavelength, we can use an interference stack and position one of its edges where
the blocking is desired. Figure 3.3 shows the reflectance versus wavelength per-
fonnance of a simple quarter wave stack that consists of layers with alternating
materials. The optical thickness of each of the layers is one quarter wave thick at
a given wavelength. The width of the high reflectance zone, commonly called the
stop band is detennined by the index ratio of the materials that make up the stack.
The total width in percent of the center wavelength of the high reflectance band is
given approximately by

400
.
% total width = 400 Arcsin = -Arcsm (3.1)
'IT 'IT

where na and nb are the refractive indices of the two materials used in the quarter
wave stack, nr = nb1na, the vertical bars indicate the absolute value is to
be used, and Arcsin is the principal value in radians of the arcisine function,
i.e., 0 < Arcsin I x I < 'lT/2.
The behavior of this function is shown in Fig. 3.4. This is the maximum width
PERFORMANCE OF OPTICAL FILTERS 59

Table 3.2 Long Wavelength Limit for some Common

Optical Materials
Material Limit (J.l.m)
Diamond 80
KRS5 40
Ge 23
CdTe 16
PbF 2 16
CdS 16
NaF 15
BaF2 15
Si 15
GaAs 15
CaF2 12
MgF2 7.5
AI 20 3 6.5
Ti02 6.2
Spinel 6.0
Si02 4.5
(From Department of Defense. 1962. Mil-HNDBK-141. Fig.IH3)

obtainable with the given index ratio. Table 3.3 lists the approximate values that
can be obtained with several common coating material pairs.
The spectral location of the stop band is a function of the optical thicknesses
of the individual layers that are periodically repeated in the stack. Finally, the level
of the reflectance achieved is detennined by the accuracy of the individual layer
thicknesses and by the number of periods used in the design. With all of the
parameters at hand, it is a relatively straightforward matter to design a filter that
will block over the range specified. The matter of building it may be entirely different
as, for example, the thicknesses may preclude its construction owing to stress and
other effects. It is at this point that the use of materials properties become critically
important. Frequently, it is possible to make a blocking tilter less expensively with
a combination of interference stacks and the use of absorbing materials, rather than
relying on dielectric stacks to do all of the blocking. Another advantage of absorptive
materials is that the eventual disposition of the energy is known precisely, and can
be accommodated accordingly. This is especially important in systems where a
large amount of energy is being blocked, such as sunlight or laser light, and the
light energy may wann the absorber and distort the system. If the energy were
reflected, it is possible that it may be focused on some nearby object that could be
damaged.
An auxiliary filter can block the light; the spectral perfonnance of this filter
overlaps that of the main filter, so that the pair (or more) of them limit the transmitted

60 CHAPTER 3

100
Design:
Stop
SUba (LH)10 AIR
3d Order
80 A.o=1.0)Jm

C»

..
u
C 60
ca
U Missing
C» 2d Order 1st Order
;:
C» 40
a:
?P.

20

O~----~--~~--~----~----~----~----i
0.25 0.75 1.0 1.25 1.50 1.75 2.00

Wavelength (pm)
Figure 3.3. Reflectance on a linear scale of the design in Fig. 3.2. The figure shows the
stop bands associated with the first and third order harmonics of the design. In this design,
the second order high reflectance band is suppressed. Note that the full width at the half
maximum point (FWHM) are not the same for the first and third orders. The ratio of width
to wavelength is constant, however.

light energy. A series of stop bands can be cascaded to provide the necessary
blocking width. In some instances, the energy blocked may extend from the far
ultraviolet to the radio spectrum. Once again, the auxiliary filter may consist of
either absorbing or nonabsorbing materials, or both. The auxiliary filter can be
placed on the second surface of the substrate that holds the primary filter. This is
commonly done, both for economic as well as for stress reasons.

3.4 FILTERS IN SERIES: INCOHERENT ADDITION

OF INTENSITIES

Care must be exercised when two optical filters are placed in series to block light.
The blocking properties of the filters do not necessarily combine linearly. If there
is some reflection as well as transmission through them, the light that gets reflected
by the second filter can be reflected back by the first and contribute to the transmitted
intensity.
The total reflectance (R tot ) from a pair of parallel mirrors is

T:1S
Rtot = R\ + ------"-
l-R\R 2 (3.2)
PERFORMANCE OF OPTICAL FILTERS 61

where RJ is the reflectance of the first mirror, R2 is the reflectance from the second
mirror, and T) is the transmission through the first mirror.
This formula assumes that there is no absorption in the space between the two
mirrors, though the individual mirrors can have some absorption. If the mirrors do
not have any absorption, the total reflection can be expressed in terms of the
reflectances only:

The total transmission is given by a similar equation:

T) T2
Ttot = 1 - R) R2
Note that when the reflectances R) and R2 are low, the effect of the denominator
is minimal. The denominator becomes more significant as the reflectances both
increase. At high reflectance values, the error in using only the numerator is also
small. When T) = T2 = 1 - R) = 1 - R2, the maximum error in using only
the numerator occurs at R = 0.618 (T = 0.382). The magnitude of the maximum
error in the total transmission is approximately 0.09. Because the denominator is

5.0 .....- - - - - - - - - - - - - - - - - - - - - - - - - . ,

4.0

3.5

0
;: 3.0
III
a:
.c
GI
"tJ 2.5
.E

2.0

1.5

1.0i-----,----~--r_-~~-~--~--~--,_-__i
o 10 20 30 40 50 60 70 80 90
Percent Width of Stop Band
Figure 3.4. Plot of stop band width for a stack of layers, all of which have an optical
thickness of a quarter wave. The stack consists of alternating layers with high and low indices
of refraction. The ordinate of the plot is the ratio of the high index of refraction to the low
one.
62 CHAPTER 3

Table 3.3 Stop Band Width for Some Common

Material Pairs
Material Pair Index ratio Percent stop band width
Ti02 Si02 1.62 9.7
ZnS ThF4 1.47 7.8
Si SiO 1.84 12.2
Ge ZnS 1.82 12.0

always less than unity, the error introduced by using only the numerator will always
be negative, i.e., the estimated transmission will be smaller than the actual value.
In Eq. 3.1, the location of two mirrors does make a difference. To maximize
the total reflection, the mirror with the higher reflection should be placed nearer to
the light source.
A technique that can be used with filters in series is to tilt each filter so that
no surface is parallel to any other reflecting surface. Light that is reflected from
one surface makes its way out of the cavity between the filters and it does not
contribute to the transmission. The tilt must not be so large as to introduce aberrations
in the optical system or to shift the performance of the interference filter, yet it
should be enough to avoid multiple reflections.
A pass band is the inverse of a stop band. In a bandpass filter, the transmission
is low everywhere except in a relatively narrow region called the pass band. High-
reflectance regions (rejection regions) exist on either side of the pass band. With
metallic reflectors, there may be some absorption. The classical Fabry-Perot type
of filter is one such example. A means of identifying the particular stop band is to
number them, starting with the one that occurs at the longest wavelength. The
number corresponds to the harmonics of the design, and is called the order.
The rejection regions on either side of the pass band are often narrower than
desired, and therefore, additional coatings are required to obtain the desired width.
A pass band can be considered as the region between two stop bands. The width
of the pass bands in the lower orders is rather wide, but as the order number
increases, the relative width decreases. This is the inverse of the situation of the
stop band in a quarter wave stack.
A wavenumber scale is commonly used to show the performance of some
filters over large spectral ranges. This scale is also common in infrared work, and
is frequently used by solid state physicists and others as it is a linear photon energy
scale. A common unit for this scale is reciprocal centimeters. The conversion factor
between micrometers and reciprocal centimeters is lQ4. Because the dimensions of
these two units are the reciprocal of each other, it is a matter of dividing into 10,000
the value one has to get the other. One way of visualizing the unit of reciprocal
centimeter is that it represents the number of vacuum wavelengths of light that can
fit into a distance of 1 cm. Table 3.4 lists some convenient points that relate the
wavelength and the wavenumber scales.
A wavelength scale emphasizes the infrared portion of the spectrum by stretch-
PERFORMANCE OF OPTICAL FILTERS 63

ing it with respect to the blue portion. The wavenumber presentation does the
opposite by squeezing the infrared and expanding the ultraviolet.
Figure 3.5 shows the performance on a wavelength number scale of the same
design as that used for Fig. 3.3. The periodicity of the performance of a quarter
wave stack on this scale is apparent in this illustration.
The blocking level of a filter is frequently shown on a logarithmic (base 10)
plot. The attenuation level of some of these filters can reduce the transmitted intensity
six orders of magnitude or more. A measure of the blocking level of a filter is
optical density or O.D. This unit is defined as
Optical Density = - 10g(T)
where T is the transmission of a filter expressed as a ratio (not %).

3.5. STANDING WAVE ELECTRIC FIELDS

Two coherent waves, i.e., having the same frequency or wavelength, traveling in
opposite directions in the same region of space, generate standing waves. In a thin
film stack, the waves reflected at each interface are coherent with the incoming
wave, and therefore, standing waves are set up in the coating. In instances where
high-power lasers are incident on the coating, the existence of these fields can play
a vital role in determining whether the coating will survive the high-power levels
or not.
The absorption of energy in an electromagnetic field is proportional to the time
average of the square of the electric field, represented by (£2), times the product
of the index of refraction (n) and the absorption coefficient (k) at that point:

Table 3.4 Convenient Reference Points Relating

Wavelength and Wavenumber Scales
Wavelength ha.m) Wavenumber (cm-1 )

0.25 40,000
0.40 25,000
0.50 20,000
0.633 15,803
0.70 14,286
1.00 10,000
1.06 9,434
2.00 5,000
2.8 3,571
3.8 2,631
5.0 2,000
10.0 1,000
10.6 943
15.0 667
20.0 500

64 CHAPTER 3

Wavelength (Jim)
0.33 0.5 1.0
100

_ r - 3rd order

80 1 at order -I---

Q)
U

--
C
ca 60
u

-Q)

Q)
a:
?f.
40

20

o
W WwM
40,000 33,000 26,000 19,000 12,000 5,000

Wavenumber (cm- 1 )
Figure 3.5. Reflectance of a quarter wave stack plotted on a wavenumber scale. The
design is the same as that of Fig. 3.2, therefore this figure may be directly compared to Fig.
3.3.

A = C . n . k . (£2) (3.3)
where C is a proportionality constant.
Generally, one considers the value of k constant through the volume of a film,
but recent studies tend to indicate that the absorption is greater at interfaces than
it is in the "bulk" of the film (Donovan et aI., 1980). Consequently, more heat is
deposited at the film interfaces. This is just about the worst place to deposit this
energy as the interfaces are also the weakest part of the coating, thus are the most
likely to fail. It is preferable to reduce the fields at these interfaces. Figure 3.6
shows the standing wave field for light incident from both sides of an interface
with a quarter wave antireflection coating. The fields are sensitive to the direction
from which the light is incident.
The boundary conditions that apply to Maxwell's equations require that a node
or an approximation of one exist at a metallic high reflector/air interface. Figure
3.7 shows the standing wave in front of a silver reflector for both the IR and visible
wavelengths. Relatively little heat is deposited in the metallic mirror as the product


PERFORMANCE OF OPTICAL FILTERS 65

~I = 100% ::-=--,... ~---r______~T.:.=.:98::.:,,7%~......~

== R = 1.3%
Air n = 1.38 n = 1.52

T= 98.7%

Air

T= 95.7%

.. R =4.3%
Air
n =1.52

T= 95.7%

Air

Distance
Figure 3.6. Standing wave (E2) patterns for light incident from both sides of glass/air
interfaces with and without a quarter wave of magnesium fluoride antireflection coating. The
arrows associated with I, R, and T indicate the incident, reflected, and transmitted energy,
respectively.

of the field strength and nk is virtually zero everywhere. If the surface consists of
a metal, such as nickel, which is a relatively poor reflector in the visible, there is
significantly more penetration of the field into the metal surface, and the absorption
is higher (Fig. 3.7c).
In a high-reflectance dielectric mirror, the electric field within the coating has
maxima and minima, as shown in Fig. 3.8. The thin film structure gives us some
design flexibility to adjust the electric field profile to optimize the overall filter

66 CHAPTER 3

A=1.5%
Ag ClR)

1=100%
R=98.2% A =1.8%
Air Ag (Visible)

A =61.5%
1= 100%
Low
Reflectance
Air
Metal
R=38.5%

Distance
Figure 3.7. Plot of (E2) showing the near-node at silver/air (a,b) and nickel/air (c) interfaces.
The wavelengths in the two spectral regions have been scaled to facilitate the comparison.
Some of the electric field penetrates into the metals in the visible portion of the spectrum,
but in the IR, little of the field penetrates into the silver. The metal indices used were
nAg IR = (0.55,15), nAg Vis = (0.055,3.3), nNi vis = (2.0,2.0)

PERFORMANCE OF OPTICAL FILTERS 67

Shaded layers: n:: 2.35

Un.haded layer.: n:: 1.45

n:: 1.52

T::2.3'1\

Distance
Figure 3.S. Plot of (E2) for a OW stack. The wavelength of the incident light corresponds
to the stack center. The field peaks at the interfaces between layers. The amplitude of the
field decays rapidly as it penetrates into the stack.

perfonnance. If the absorption in one of the two materials in a quarter wave stack
is higher than the other, it is possible to minimize the field in the higher absorbing
material, though at the expense of raising the field in the other material. The result
of this is a higher reflecting stack (Carniglia and Apfel, 1980). Another use that
has been made of electric field studies is in the design of laser reflectors where
laser damage is of serious concern. The modification of the thickness of a few of
the outer layers of a quarter wave stack to reduce the peak field in the high index
material has been shown to improve the damage threshold (Apfel, 1977).
A classic example of the way the electric field is involved in an absorbing
design is the Fabry-Perot cavity filter that has a metal, such as aluminum or silver
for the reflector layers. At a reflectance resonance, i.e., the wavelength where peak
reflectance occurs, there are near nodes at the metal surfaces, and the absorption
is relatively low. Away from this resonance, there is a significant amount of ab-
sorption, even with silver films, and the electric field is relatively high in the metal
layers (Fig. 3.9).
The electric fields in dark mirrors (selective absorbers), which consist of
metal/dielectric layer combinations, peak in the absorbing metal layers to increase
the absorption of the incident light. This design technique is used in the so-called
trilayer designs of optical data disks to improve the sensitivity of optical data disks
and to control the reflectance (Bell and Spong, 1978). Figure 3.10 shows the standing
wave fields for the trilayer configuration, a thick layer of the absorber layer by
itself and a thick layer of the absorber material over a high reflector. The absorption
values given indicate the effect of the field on the sensitivity. In addition, the
amount of material required to obtain a given amount of absorption is reduced,
thus reducing the amount of material that needs to be heated to write a bit on the
disk.
The fields in antireflection coatings are relatively low by comparison to those
found in a high reflector design. Figure 3.6 shows the fields for a single layer AR
design, whereas Fig. 3.11 shows the fields present in a multilayer AR design of

68 CHAPTER 3

Dlatane'"

Figure 3.9. Plots of (E2) for a Fabry-Perot filter with silver reflectors at two wavelengths.
The transR'lission is at a peak at the first wavelength (}'1), while it is at a minimum at the
second wavelength (A2). The absorption is at a minimum when the reflection peaks because
the field in the metal layers is also at a minimum. The absorption at A1 is approximately 9
percent; at A2, it is 2 percent.

the quarter-half-quarter (QHQ) type. Here also, the minima and the maxima in the
electric field profile can be moved to accommodate specific requirements. In high-
power laser systems, it is desirable to reduce the absorbed energy to an absolute
minimum to maximize the damage threshold. This approach has been demonstrated
experimentally (Apfel, 1977; Demichelis et aI., 1984).

3.6. PERFORMANCE SHIFTS WITH ANGLE OF INCIDENCE

In section 1.3.6., we discussed the effect of nonnormal angles of incidence on the

effective index of refraction. In this section, we are more specific with respect to
the changes in performance of coatings as the angle changes.
The angle-induced changes in filters can be qualitatively divided into two
categories: (I) the wavelength shift and (2) the polarization effects.
The wavelength shift is straightforward: performapce curves always shift to
shorter wavelengths (in the blue direction) as the angle of incidence increases away
from the normal. At first glance, it may appear that the performance should shift
toward the red because the optical thickness of each of the layers in the design
increases as the angle of incidence increases (by a factor inversely proportional to
the cosine of the angle of refraction inside the layer). The important parameter,
however, is not the total phase shift introduced in the waves by the layers. It is the
phase difference between the waves that reflect from all of the interfaces that enters

PERFORMANCE OF OPTICAL FILTERS 69

(a)

Te
=
A 28.9%

Ag A =5.1%
Distance
(b)

Te
A= 95%
1=100% (

... R=O%
Air
•
~,
~~ ...
(c)

Figure 3.10. The (E2) in optical data disk recording systems. The reflectance and absorption
is shown. In (a). the field is shown at an air/tellurium interface. In (b). the tellurium layer is
deposited over a silver mirror (bilayer). In (c). the trilayer design is seen to be the most
efficient as the maximum amount of energy is absorbed in the least amount of material.

into the summation of the partial reflected waves. This quantity decreases as the
angle increases. Thus, the perfonnance shifts toward the blue as the angle of
incidence increases because the apparent optical phase difference introduced by
each layer is decreasing.
In a bandpass filter, especially narrow ones, this effect can serve as a fine
tuning mechanism to position the pass band at a precise spectral value, if the filter

70 CHAPTER 3

T=100%
1:100%
Air
R=O% n= n= 1.52
1.65

Distance
Figure 3.11. Profile of (E2) for a QHQ antireflection design. In simple AR designs, the field
within the coating does not rise above that of the incident or exit beams, whichever is greater
(over the wavelength range in which they function as AR coatings).

is originally coated such that the pass band is at a wavelength that is longer than
the desired one. The degree of the shift is detennined by the average index of the
materials used to construct the filter. In a two-material, all dielectric Fabry-Perot
bandpass design, the angle shift is less when the higher index material is the half
wave thick spacer layer between the high-reflecting quarter wave stacks. Figure
3.12 shows the perfonnance of two designs with approximately equal perfonnance

100~----------~~~~~~~---------,

\
I \

75
I
,,
I \
\
\
\

,,
CD I \
U I \

-
C

u
co
CD
;;:
50 ,,,
\
,
\

,,
\

,
CD
a:::
, ,, I
\
~
25

, I

I
I ,
I

,,, I \
I
\
\
I \

o+---~~~~------~----------~----~
0.5 1.0 1.5

Wavelength (}1m)
Figure 3.12. Normal incidence reflectance of {Hl)4{lH)4 (solid line) and {lH)4{Hl)4 (dashed
line) as a function of wavelength (nH = 2.2, nL = 1.45). The difference in the performances
of the two designs is due 10 the effect of the index of the outer layer in each deSign.
PERFORMANCE OF OPTICAL FILTERS 71

at nonnal incidence (save for a proper AR treatment in the center of the pass band).
Figure 3. 13 shows the perfonnance of these same designs at a 50-degree angle of
incidence. The shift in center wavelength of the coating with the higher index in
the spacer layer is significantly less, especially relative to the width of the pass
band, than the one for the low index spacer. In addition, the shape of the pass band
stays more constant with the higher index spacer material.
The shape of the perfonnance curves of a coating at nonnonnal incidence may
look quite unlike that at nonnal incidence. Here, as in other situations in optics,
one must examine the perfonnance in the two planes of polarization separately. If
each plane of polarization is examined individually, then the angle effects become
much clearer. The theoretically calculated perfonnance of a quarter wave stack, as
shown in Fig. 3.14, illustrates this well. The stop band for s plane light is wider
than that at nonnal incidence, whereas that of the p plane ois narrower. This is
explained by the ratios' of the effective indices at nonnonnal angles of incidence
being larger for the s plane, whereas the opposite is true for the p plane.
For some incident angles, it is very difficult to design a coating because the
effective indices are all very similar, so that it is not possible to get a reasonably
large index ratio. The effective index varies less for the high index materials than
it does for the lower index values (see Fig. 1.11). A filter that has to operate over
a range of angles with minimum shift in perfonnance should be constructed with
the highest indices possible. The infrared region of the spectrum has a wider range
of available material indices (both high and low) than does the visible portion.
Thus, it is easier to design angle stable filters in the infrared portion of the spectrum
than it is for the visible portion, all other parameters being equal. At angles of
incidence near 90 degrees, the effective indices of the sand p components of all
materials tend toward zero and infinity, respectively. The reflectance, therefore,
approaches 100 percent at grazing incidence. It is difficult to fonnulate good per-
fonnance, wavelength sensitive high reflecting designs at high angles of incidence.
Antireflection coatings are especially difficult to design at high angles.
The effects of angle shift have been measured and reported in the literature
by Baker and Yen (1967). Figure 3.15 is taken from their paper; it shows a range
of wavelength shifts in percent as a function of the angle of incidence of collimated
radiation. These measured results agree with theory.

3.7. OTHER PARAMETERS THAT AFFECT PERFORMANCE

Water, temperature, and other environmental conditions can affect the perfonnance
of thin film filters. The sensitivity of a filter's perfonnance to these conditions is
difficult to specify in a general sense. The actual coating processes have a definite
influence on the details of the interaction between the coating perfonnance and the
environment to which the filter is exposed.

3.7.1. Water and Humidity

It is well known that the perfonnance of optical filters may be sensitive to envi-
ronmental moisture. The degree to which a filter varies depends on the materials
 100'~~=-----------~~---(~~-----_-_-----------'.-
~---, ,~

/
""..........
" " \ I
,,,"
80· I
I
\ I
I
I \ I
, I
~
I \
U I \ I

-
cCU 60
U
~
;:
I
~
I \
\
\
\
\
\
,
,,
I
I
I

~
a: 40 , ,,'
\
~
, ,,
, I

20
\
\ ,
, I

,,''
o+-------~------~--~~~~----~~----~
.75 .8'0 .8'5 .90 .9'5 1.00
Wavelength (}1m)

1001===========~=-------------====~

80 /
"
",,'
~-- .... ..... ,
",' /
//
"" .... --
I/ , I
/ \ I
I \ I
~ I \ I

-
U \ I
C 60 \ I
CU \ I
U
~
;:
\
,
\
I
I
I

, ,
~ 40 \
a:: \
\
I
~ I
\ I
\ I
20 \ I
\ I
\ I
\
\
\ I
, I

0
.75 .80 .85 .90 .95 1.00
Wavelength (}1m)
Figure 3.13. Reflectance at 50 degrees angle of incidence of the same designs as in Fig.
3.12: upper. Low index half wave spacer; lower. High index half wave spacer. Dashed curves
are for the p plane and solid curves are for the s plane polarization. Note the expanded
scale on the abscissa.
72
PERFORMANCE OF OPTICAL FILTERS 73

used in the filter as well as on the processing conditions. Gee and co-workers (1985)
have measured the change in performance of evaporated titanium dioxide and silicon
dioxide filters as they absorbed moisture from the ambient humidity. The general
nature of this effect is that the performance of many coatings can undergo a pro-
nounced change upon removal from the vacuum chamber and exposure to the
atmosphere. For example, the center wavelength of a design shifts to a longer
wavelength as the coating absorbs moisture. If the filter is subsequently dried either
by heating in a warm oven or by leaving it in a dry box for a longer period of time,
the performance shifts back toward the preexposure values; however, it will never
reach the original performance. Thus, there is an irreversible component in the
moisture shifting effect. Subsequent cycling of the filter between wet and dry
environments shows that there is also a reversible component of the moisture effect.
On the basis of this phenomenon, it is apparent that for the most demanding
applications, the measurement of a filter's performance should be made after the
filter has stabilized in an ambient environment that is similar to that in which it
will be expected to operate.

3.7.2. Temperature
Temperature can affect the performance of interference filters because both the
index of refraction and the thickness vary as a function of temperature. Baker and
Yen (1967) measured the actual temperature shifts in typical infrared filters. Figure
3.16 shows their results.
A temperature-induced shift can be especially significant when one is dealing
with a filter with a critical wavelength dependence. The center of a narrow wave-
length structure may shift by a large fraction of its width owing to changes in the
optical thickness of the layers in the coating. The optical thickness of each layer
varies with temperature as a result of (1) the change in index of refraction with
temperature (dn/dT) and (2) the linear coefficient of expansion (dlldT) that affects
the physical thickness.
The indices of refraction of most materials become smaller as the temperature
increases, i.e., dn/dT is negative. On the other hand, the expansion coefficients
dlldT are typically positive, and these lead to corresponding increases in the path
length. Thus, the change in optical thickness change with temperature is a second-
order effect. The result is that for most materials, the performance of a filter shifts
to longer wavelengths as the temperature is increased.
One solution to the problem of temperature shift is to enclose the filter in a
temperature-controlled oven. With this technique, the filter can be tuned over a
small spectral range. This technique is useful, for example, to select a specific line
from a spectrum. The oven approach, however, does leave something to be desired;
it is an active device and it consumes power. In addition, ovens can have a long
time constant, i. e., it cannot be varied rapidly.
An alternate solution of the temperature shift problem is to design the filter
so that it is temperature invariant. This is possible if a material is available whose
index of refraction varies in a way that is opposite to that of the other materials in
a design. If this shift is large enough, one can compensate for the temperature-


100~--------------------~------~----
I
I
__
I
=75
I
I,
M
9 0

O~----------------------~--------~
100~,~,----------------------------'-------------~
I I
I I
I I
I I
I
9 = 60 0

CI)
U

---
c
as
u
CI)
100,-~--------------------------~~----

CI)
a::
~ 9 = 45 0

100,---------------------------------------~

9 =0 0

O~~~~~~~~~~~~~--~
0.25

0.5 0.75 1.0
Wavelength ()1m)

74
PERFORMANCE OF OPTICAL FILTERS 75

induced changes in the refractive index and the thickness so that the overall design
is temperature stable.
Some infrared filters can be designed to be temperature insensitive. A material
that has appropriate dnlD and dl/D in the infrared portion of the spectrum is lead
telluride. Seeley and co-workers (1980) have developed this material and formulated
designs that put this effect to use. It is now possible to obtain filters commercially
that have an extremely small temperature coefficient. These filters have found use
in pollution measurement devices. The filters transmit a selected spectral line near
4 f.Lm, which is emitted by carbon monoxide. Unfortunately, no material has yet
been found that has the proper characteristics to allow the same technique to be
used in the visible portion of the spectrum.

3.7.3. Corrosion
Corrosion can occur in some coatings. In particular, metallic films, such as silver
and aluminum, are subject to degradation as they are attacked by environmental
agents. The films oxidize and lose their metallic properties. The indices of refraction
of oxides are radically different from those of the metallic precursors.
The behavior of metallic silver films is particularly notorious in this regard.
Some recent studies at the University of Arizona may be showing the way to tame
this very useful metal and make it more durable for thin film applications (Song et
al., 1985). Not so well known is the degradation of metallic aluminum film in high
humidity environments. Over a period of time, these films can convert completely
to their oxide. The presence of ultraviolet light greatly enhances the rate at which
this conversion takes place.
Moisture can also lead to the degradation of oxide films. If the oxide is soluble,
then the thin film can be dissolved by the ambient water, leading to a loss of the
filter. Tests are available that will verify that the filter should not fail in normal
use (see Chap. 5).

3.S. WAVEFRONT DISTORTIONS

In most applications, a filter will not degrade the optical performance of the system
in which it is placed. In critical applications, however, the effect that the filter may
have on the wavefront which is passing through it or being reflected off it could
be significant and needs to be considered by the designer. Cases where such con-
ditions may apply are in diffraction-limited optical systems, such as precision camera
lenses or in laser systems that are used in interferometry.
When collimated light traverses a filter on a plane substrate, no aberration is
introduced into the wavefront. A coating may introduce a uniform phase shift into

Figure 3.14. Reflectance of a OW stack as a function of wavelength for various angles of

incidence. The design if (ML)'OM (nm = 1.65, n, = 1.35) with air as the incident medium
and a 1.52 index glass as the substrate. The dashed curve is for the s plane of polarization,
the lower solid line is that for the p plane, and the middle solid line is the average of the
light intensity in the two planes.
76 CHAPTER 3

the wavefront, however, and this may be significant in an optical system in which
phases are being compared. For example, a quarter wave thick film deposited on
a substrate introduces a 270-degree ( - 90°) phase shift to a transmitted beam at the
quarter wave wavelength relative to a bare substrate. When more complicated
designs are used, the result is correspondingly more complex. These phase shifts
are a function of the angle of incidence and the wavelength. The manufacturer of
a filter can supply a calculation of the phase shifts of a given design for use with
the user's modeling programs to calculate the overall system performance.
Both the previously mentioned angle shift (see section 3.6.) and the phase
shift can affect the performance of diffraction limited or interferometric quality
systems in which the lenses have steeply curved surfaces.
On steep surfaces, the thickness of the manufactured coating may vary from
the center to the edge because of process limitations. Careful masking can be used
to eliminate this run-off, if necessary. On the other hand, it should be noted that
even a uniformly thick coating may introduce wavefront distortions because of angle

6~-----------------------------------------'

--
~
CJ)
C
G)
G)
5

>
~
..'"
-..
CD

0
~
Si/SiO

-.--
0
0

Ge/ZnS
~
0
1
~

0
0° 30°

Angle of Incidence
Figure 3.15. Measured wavelength shift of interference filters as a function of the angle of
incidence (from Baker and Yen, 1967). The incident radiation is collimated. The filters were
infrared types. The upper limit corresponds to an index pair of approximately 3.45 and 1.9,
whereas the lower limit pair is approximately 4.0 and 2.2.

PERFORMANCE OF OPTICAL FILTERS 77

+3

+2

..
.....-
.c
+1

.. 0
U)
.c
m
c -1
->
CD
CD
as -2
~
~ -design temperature· =20° C
-3
liquid nitrogen
helium
-4

-5 ~r----~-------r------~------r----~------~------r----~
-250 -200 -150 -100 -50 o +50 +100 +150

Filter Temperature (Oe>

Figure 3.16. Measured temperature shift of infrared interference filters as a function of the
temperature (from Baker and Yen, 1967).

effects. When using off the shelf coated optical elements in interferometry, the user
should pay attention to these details.
In fast and wide angle optical systems, where steep angles of incidence are
commonplace, the operation of a bandpass filter may not be what is anticipated,
based on normal angle of incidence measurements. Specifically, in a system with
a large spread of angles, the performance is the result of an average of all of the
angles incident on the filter. If the filter is especially narrow, this may result in a
significant reduction in the light throughput over what the same filter transmits in
a collimated beam at a normal incidence.
 Chapter 4

Design Types

This chapter catalogs a number of common thin film design types. This list is
intended to be generic in nature to give the reader some idea as to the possibilities
that are available. It cannot be overemphasized that these are only generic designs,
and that they can be, and in most cases should be, adjusted and optimized by a
competent thin film designer for a particular problem. An almost infinite variety of
designs can be made by combining simpler design types. The reader who has an
interest in thin film design techniques is referred to a companion volume in this
series: Design of Optical Interference Coatings by A. Thelen.
The order of the following designs has no particular significance. The more
frequently used ones are listed first and the less commonly used ones follow.

4.1. ANTIREFLECTION COATINGS

Antireflection (AR) coatings were the first coatings to be produced by the thin film
technologist and are the most commonly used coatings today. Undoubtedly the
simplicity and ease of fabrication of a single layer contributed to the early successes
of optical thin films. As discussed in Chap. I, an AR coating for a glass surface
can be made with a single low index layer with an optical thickness of a quarter
wave. A number of materials can be used for this purpose, and they can be deposited

78
DESIGN TYPES 79

quite easily by heating them with a resistive heater in a vacuum chamber. It is

straightforward to judge the thickness of a single low index deposit by eye (Heavens,
1965, p. 243). The precision thus obtained was adequate for the early thin film AR
products.
Since the original quarter wave AR coating, many other different types have
been developed. The goal of all of these is to make a coating the reflectance of
which is as low as possible over as wide a wavelength band as possible. All too
frequently, one can get more width at the expense of the reflectance level, and it
is difficult to get both simultaneously.
AR coatings reduce the reflection that occurs at interfaces between two optical
media. In general, one of these media has an index of unity (air or vacuum), though
it could be water or some other medium. Common soda lime glass, with its index
near 1.5, reflects approximately 4 percent of the incident light. Dense flint glass,
sometimes used in highly corrected optical systems, can have an index as high as
1.8; e.g., Schott Glass type SF6, and a corresponding reflectance of 8 percent. In
an optical system with many surfaces, these losses add up to a significant quantity.
Worse, the reflected light can contribute to veiling glare in the optical system,
which reduces the contrast at the image plane. In infrared systems, with even higher
index semiconductor substrates, the situation is even more critical. For example,
the reflectance from a silicon surface is more than 30 percent, and that from
gennanium is more than 35 percent. These losses are intolerable in just about any
infrared system and must be suppressed.
Because a multilayer interference coating with zero reflectance at all wave-
lengths has yet to be designed, a merit function has to be used to evaluate a given
design, and to compare its performance to other designs. Typically, a weighted
average is used for this calculation. The reflectance is multiplied by a weighting
function and the result is summed over the desired wavelength interval. The weight-
ing function is selected to represent the optical system with which the filter will
operate; it may be proportional to the human eye response, the response curve of
a sensor, or any other function. Thus, the weighting function tailors the optimization
to the specific system. Photopic ally (daylight brightness eye response) weighted
systems function well as antireflection treatments for visual applications.
One area that is not generally included in computer refining of AR coatings
is the effect of color on the visual perfonnance of the coating. In particular, the
purity, the dominant wavelength, as well as the integrated brightness are important.
Also, the variability of these parameters over a substrate, as well as part to part
variability are important. The eye is very sensitive to color shifts, but there is no
widely accepted measure of this effect that can be used in thin film design refining
and optimizing programs. The difficulty arises from the shift in performance with
layer thickness errors. Figure 4.1 shows the variations that can typically be expected
if a 2.5 percent standard deviation in the thicknesses is allowed in a three-layer
QHQ AR design. The indices used in the calculation are shown in the illustration.
Table 4.1 lists the variations in the color space coordinates and the reflected lu-
minance, the dominant color, and the purity of the reflected light. It is clear that
the perceived colors are different for each sample, yet the overall reflectance remains

80 CHAPTER 4

2.0r-------------------------------------------------~

Air
CD
=1.38
..
U n
C
II
n =2.10
~ 1.0 n =1.65
;:
CD
a: 77777777/
l/i.
n. =1.52

500 600 700

Wavelength (nm)
Figure 4.1. Optical thickness tolerance analysis (2.5% standard deviation) or a three-layer
AR QHQ antireflection design. The performance of the nominal design is shown by the
dashed line. The variations in color purity, dominant wavelength, and photopically weighted
integrated reflectance are given in Table 4.1.

Table 4.1 Color Variations as a Result of Layer Thickness Variations

Reflected Dominant C. I. E. (1931) C. I. E. (1976)
Iteration luminance wavelength Purity
number (%) (nm)* (%) x y u' v'
nominal 0.065 565C 16.9 0.318 0.283 0.221 0.442
1 0.066 469 38.1 0.254 0.227 0.195 0.392
2 0.063 472 44.0 0.238 0.218 0.186 0.382
3 0.068 467 44.9 0.242 0.205 0.195 0.371
4 0.070 471 48.6 0.230 0.201 0.186 0.366
5 0.065 556C 43.8 0.328 0.220 0.264 0.397
6 0.068 557C 48.1 0.326 0.207 0.269 0.386
7 0.081 445 22.1 0.295 0.263 0.212 0.425
8 0.061 567C 22.6 0.304 0.263 0.220 0.426
9 0.077 604 18.5 0.392 0.337 0.250 0.484

This data corresponds to the same variations in performance of the design shown in Fig. 4.1.
'A 'C' following the wavelength value indicates the complementary color is given.
DESIGN TYPES 81

low compared with the bare substrate value of 4.3 percent. If an isolated substrate
has some color in reflection, it mayor may not be apparent, depending on how the
gradation from one area to the other occurs. When adjacent panels with different
coating errors (colors) are compared, the differences may be quite noticeable and
unacceptable on the basis of cosmetics.
Generally, AR coatings are fabricated with nonabsorbing materials; as a result,
any reduction in the reflected light results in an increase in the transmitted energy.
It is also possible to reduce the reflectance with the use of absorbing films in the
design. The absorption reduces the transmitted light, but in some applications, the
reduction in the reflected light is more important than keeping the transmitted light
at a high level.
One application of an absorbing AR is the contrast enhancement filter used
on computer terminals with light emitting ("active") displays such as cathode ray
tube, plasma, and electroluminescent displays. (Absorbing AR designs are not useful
with liquid crystal or other "passive" displays because to function, these types of
displays depend on ambient light reaching them.)
Alternative techniques exist that reduce the reflected glare from surfaces, but
there is usually a reduction in optical performance associated with each of these.
For example, the surface of the display may be etched or frosted lightly; this
treatment destroys the specular reflection from the surface by scattering it in non-
specular directions. Although etching and frosting reduce the direct image problem,
it also reduces the sharpness of the screen image, and the dark areas are lighter
than they would otherwise be. In high-contrast, high-visual fidelity systems these
approaches can detract from the other system improvements (see Haisma et al.,
1985).
Given no constraints on indices of refraction, it is possible to design an AR
coating that has zero reflectance at all wavelengths. It is a matter of grading the
refractive index at the interface between the substrate and the other medium. In
other words, the index should vary continuously from that of the substrate to that
of the incident medium so that there is no discontinuity. The actual shape of the
index gradation function is not especially critical, though the thickness of the graded
region should be at least a quarter wave thick, and preferably of the order of a half
wave thick. Such an AR layer has indeed been realized in practice (Minot, 1976;
Yoldas and O'Keefe, 1979). The process begins with a glass, which consists of at
least two separate chemical phases. An etching process dissolves one phase, while
leaving the other untouched. As the etchant attacks a surface, the phase that is
attacked is completely removed near the surface, but as the removal process works
from the surface and proceeds inward, the percentage removed decreases as the
depth increases. The microstructure left behind at the surface is mostly air. Thus
the surface of the layer has an index close to unity, and the index increases with
depth. Because there is almost no discontinuity in the index function at the interface
so treated, the reflection is very close to zero and not a function of wavelength.
Unfortunately, the porous surface consists of many fine capillaries that can retain
liquid contaminants, such as oils from fingerprints, thus making the surface very
difficult to clean.
82 CHAPTER 4

A major challenge exists for the designer of thin film AR coatings. It is to

arrive at a design, the performance of which meets the original specifications, and
that only incorporates indices available in Nature. A number of papers have been
published reporting studies on the index ranges that are most satisfactory for con-
structing AR filters in a number of special cases. The designer has to make com-
promises and use special tools to synthesize the indices needed. The synthesis of
unavailable indices can be done by using combinations of layer thicknesses and
available indices. Epstein (1952) and Herpin (1947) have published papers explain-
ing mathematical formulations that enable a designer to generate a specific equivalent
index. These approaches function adequately over relatively narrow wavelength
ranges, and the challenge is how to make the synthesized index function over ever
wider ranges.
Multilayer AR coatings are highly tuned structures. The design of an AR
coating for a low-index glass will not likely be satisfactory if applied to high-index
glass.
A general rule in noncritical uses of thin film filters is that the performance
does not change markedly with small changes in the angle of incidence; e.g., in
changing from normal incidence to 30 degrees. Reflectance measurements are not
frequently made at exactly normal incidence, but at some small angle away from
normal, often 10 degrees. This small angle will not generally cause a significant
error in the measurement.

4.1.1. Single-Layer Designs

As previously mentioned, the single layer AR coating was the earliest application
of the art of thin film design and fabrication. The pioneer thin film coaters used
cryolite (sodium fluoaluminate, Na3AIF6), chiolite (NaxAlyFz), or magnesium flu-
oride (MgF2) as materials for coating AR layers on glass because they are easily
evaporated from resistance sources and have indices that are suitable to antireflect
glass. The optical thickness required to obtain a minimum reflectance is approxi-
mately a quarter wave. The spectral performance of single layer films is given in
Fig. 4.2. It should be noted that (in the absence of dispersion) the reflectance never
rises above the bare substrate reflectance at any wavelength when a single quarter
wave film with an index lower than the substrate is used. Conversely, when the
film index is greater than that of the substrate, the reflectance never drops below
that of the bare substrate. The half wave points serve as indicators of whether the
film has absorption or inhomogeneities in it. Unless the reflectance returns to that
of the bare substrate, one, or the other, or both of the assumptions of no coating
absorption or inhomogeneity in the layer is violated.
The reflectance of a single quarter wave thick film is given by:

n/ - nofl,]2
R = [ nf2 + noll, '
(4.1)

where nf is the index of the film, no is the index of the incident medium, and n,
is the index of the substrate.
In a film of magnesium fluoride, which has an index of approximately 1.38,

DESIGN TYPES 83

Optical film thickness

2.1\/2 5.1\/4 3.1\/2 771./4

CD
+
• •
..
U
c:
III
U
CD
;:
CD
a:

/ /
\ /
\ / n,<nsub \
\ I/
'-.../ '-'

Wavenumber
Figure 4.2. Reflectance of single layer dielectric coatings with an index n, on a substrate
with an index nsub. The reflectance extrema occur at the points where the films are a multiple
of a quarter wave thick (marked with the respective fraction of a wavelength ),). Note that
the reflectance returns to exactly that of the substrate when the film is an even multiple of
a quarter wave, i.e., a multiple half wave.

deposited on glass with an index of 1.52, the reflectance is reduced from that of
bare glass, 4.26 percent, to 1.26 percent. Equation 3.1 is also applicable when the
index of the film is higher than the index of the substrate.
The minimum reflectance is exactly zero when the index of the film is the
geometric mean of the incident and substrate indices:

(4.2)
In the case of a glass with ns = 1.52, the optimum film index is approximately
1.23. No material exists in nature that satisfies this index requirement and is also
stable in a typical environment. The nearest index material that is durable is mag-
nesium fluoride. Other materials, such as cryolite and chiolite, have lower indices,
but they are to some extent water soluble, and films of these materials will not
survive in humid environments.
Equation 3.1 applies to all odd multiplies of quarter wave thick films.

4.1.2. Narrow-Band AR Designs

A zero reflectance design is the two-layer "V -coat." In fact, for a given combination
of materials that leads to a valid design, there are two such designs. The two designs

84 CHAPTER 4

10.0 . , - - - - - - - - - - - - - - - - - - - - - - - - - _

A: Air B:
t= 126nm n=1.38 t =73nm
7.5 t = 21nm n=2.00 t =116nm
'l7777/ 77777/
ns = 1.5
GI
U

..
c:
III
U 5.0
...
.!
CD
uncoated substrate In = 1.52)
IE:
lP. /
//
2.5 /
//SOlutlon B

/
/
, / /
O.O~---~---_,-~~--~--;'~~,_----_r------~
400 500 600 700

Wavelength (nm)
Figure 4.3. Reflectance of both V-coat solutions. The reflectance is zero at the minimum
for both designs. The shape of the relatively narrow low reflectance band gives this design
its name. The physical thicknesses of the layers in each design are shown in the figure.

for selected indices and a glass substrate are shown in Fig. 4.3, along with the
perfonnances. Occasionally, there will be a single degenerate solution; the thick-
nesses of the layers are both equal to a quarter wave optical thickness. It is not
always possible to come up with a solution for a two-layer design given two films
of arbitrary indices, nor is it always possible to find materials that have the optimum
indices.
The criterion used to select which of the two solutions is to be used is based
on stress considerations, all other things being equal. The difference in the thick-
nesses of the two solutions sometimes causes the stress level in one solution to be
much better balanced than the other solution.
These designs are useful in high-power laser applications. Because they have
few layers, they can have low stress. The materials that are used in the design can
be selected for their low absorption characteristics and high damage thresholds,
rather than their optical index of refraction as index in this type of design is not a
critical parameter.
For many applications, the width of the low reflection region is too narrow
and other types of designs need to be used.


DESIGN TYPES 85

4.1.3. Wide-Band AR Coatings

The limited perfonnance of quarter wave AR coatings led to the development of
more complex designs. The quarter-quarter (QQ) solution consists of two quarter
wave thick films of appropriate indices. The perfonnance of a typical design is
shown in Fig. 4.4. The addition of a half wave thick layer between the two quarter
wave layers results in even better perfonnance over a wider wavelength range. Such
a design is commonly referred to as a QHQ design. Modem AR coatings can easily
achieve an average reflectance of less than 0.5 percent over the visible portion of
the spectrum. Figure 4.4 also shows the theoretical perfonnances of a QHQ design.
A multitude of papers on the selection of indices and the design of such coatings
can be found in the literature.
Source and detector weighting functions become important when evaluating
broad band perfonnances such as these. In particular, these designs have a tendency
to act like the proverbial "toothpaste tube," i.e., when one point is pushed down,
another point rises. If the width of a design is increased, the reflectance tends to
rise. Thus, the appropriate type of weighting function must be used in designing

Bare substrate
4

00 aHa
\ Air
n =1.38 n
Air
=1.38

"c
3
\ n= 1.70
;;»»»)
n= 2.40

..
u n=1.70
III
\ "-0= 550nm
;;;;;J;?
Ao

-
u =525nm

\

.!
II 2
II:
1P. \OHO
\
\
0
400
"
'-
500

Wavelength (nm)
600 700

Figure 4.4. The performance of quarter-quarter (00) and quarter-half-quarter (aHa) an-
tireflection coatings. The reflectance of an uncoated substrate is shown for reference. Note
the additional width that these designs have over the V-coat shown in Fig. 4.3.
86 CHAPTER 4

and evaluating these coatings. When one specifies a wide-band antireflection coating
for visual use, the designer should give consideration to the purity and the dominant
wavelength of the reflected light, as well as the reflected brightness.

4.1.4. Superwide Infrared AR Coatings

With the larger range of material indices available in the infrared as compared with
the visible portion of the spectrum, it is possible to obtain much wider infrared AR
designs. Coatings that span several octaves of the spectrum are not uncommon.
The design of a 4 to I wide AR (~ A/Am in) can consist of a stack of approximately
quarter wave thick layers, the indices of which progressively decrease from the
substrate interface to that of the incident medium. Figure 4.5 shows the perfonnance
of one such design at nonnal incidence. The original reflectance of the bare substrate
is shown for comparison.

4.2. HIGH REFLECTION DESIGNS

High reflectance designs fall into three general classes: (1) metallic, (2) all dielectric,
and (3) enhanced metallic. Each of these is covered in tum.

4.2.1. Metallic Reflectors

Metallic thin film reflectors antedate optical interference thin film coatings by many
centuries. By far the earliest manufactured optical thin films, metal films find use
today in telescope mirrors and in other optical instruments. They are often fragile
and cannot be contact cleaned. The common metals used are silver, gold, and
aluminum. Silver is sometimes used in astronomical telescopes because it can be
deposited by an electroless process on mountain tops without the need for expensive
coating equipment. Silver tarnishes quite readily, therefore, this process has to be
repeated periodically to keep the reflectivity high. Silver has the highest reflectance
of any metal in the visible and infrared portions of the spectrum, and it makes an
excellent material for this application when it is fresh and untarnished. Recent work
be Song et al. (1985) indicates that a solution to the tarnishing of silver may be
near at hand. Such a breakthrough would be extraordinary and a boom to the art
of high reflector manufacture. Silver is not a good reflector in the ultraviolet portion
of the spectrum because it has an absorption band that starts at 350 nm and extends
to shorter wavelengths. This absorption band is caused by electron plasma reso-
nances in the metal lattice, and it is not likely that it can be eliminated as it is a
basic material property.
When light has to be reflected over a large range of angles of incidence and
for a wide band of wavelengths, it is difficult to find a better reflector than a metallic
surface such as silver, gold, or aluminum. Most dielectric stacks have high reflec-
tances over a relatively narrow band of wavelengths. Although several of these can
be used one atop the other to cover a spectral region wider than that obtainable
with a single stack, a point of diminishing returns is reached. It is not possible to
improve the reflectance beyond a point because of stress buildup in the coating,
scatter, or absorption in the coating materials.

DESIGN TYPES 87

50r---------------------------------------------·------~

40
Bare germanium surface

GI Air

..
Co)
C 30 n :1.46
ca
Co) n= 2.04
GI
;: n= 2.86
GI
£C Germanium
?F. 20 11.0 =3.5J1m

10

o

2 4 6 8 10
Wavelength (JIm)
Figure 4.5. Performance of a wide band infrared AR on the surface of a germanium
substrate. All layers are nominally a Quarter wave oplicalthickness in the center of the band.
Each surface of uncoated germanium has a reflectance of 36 percent, and a substantial
improvement in energy throughput is achieved by adding an AR coating to the surface.

Gold is used frequently in optical instruments designed to operate in the near

and far infrared portions of the spectrum. In these regimes, its reflectance is almost
as high as that of silver, and, it does not tarnish. Gold evaporated in a standard
vacuum is not very adherent to glass. Over the years, techniques have been de-
veloped to solve this problem. The most common solution in an evaporation process
involves the use of a thin chrome layer under the gold. The chrome is adherent to
the glass, and the gold adheres well to the chrome layer. The use of an ultrahigh
vacuum coating system or a sputtering process can deposit gold without the inter-
mediate layer. Although these techniques solve the problem of adhesion, they do
nothing for the problem of the fragility of the gold itself. In a sealed optical
instrument where access is limited and the chance of the optics' getting dirty is
minimal, the cleaning and damage considerations are not a drawback.
Both gold and silver mirrors can be overcoated to give them more durability
than they otherwise may have. There are several limitations with this approach.
Overcoats are generally not transparent over the entire range that the metal is
reflective, thus the overcoat limits the spectral range of the mirror. For wideband
instruments and optics, this can be a serious problem. On the other hand, all is not
88 CHAPTER 4

as bleak as it may seem. Figure 4.6 shows the normal incidence standing wave
electric field at the surface of a metallic reflector with a thin overcoat. The node
at the surface limits the amount of energy absorbed in the overcoat. This is illustrated
in Fig. 4.7, which shows the infrared absorption of a free-standing fused silica
layer in air. Figure 4.8 shows the reflection of silver, both bare and overcoated
with a silicon dioxide overcoat, the thickness of which is the same as that in Fig.
4.7. The absorption added by the overcoat is the difference between the two curves
(note the expanded scale). The reflection over most of the spectrum is only slightly
affected by the presence of the coating.
The situation is much different at nonnormal incidence. Figure 4.9 shows the
reflectance at a 60-degree angle of incidence for both planes of polarization. It is
obvious that there is much reduced reflection especially in the 9- to 1O-~m region.
This corresponds to the reststrahlen reflection region of the overcoat material. In
this wavelength region is a resonance frequency: the molecules that make up the
crystal lattice undergo a collective oscillation that is driven by the electromagnetic
field. Thus, the lattice that makes up the film can absorb a significant amount of
the energy in the incident light beam if it is properly coupled into the film. The
electric field plots in Fig. 4.10 show the effect of the angle of incidence on the
standing waves. Although the pattern in the s plane resembles that at normal
incidence; the p plane pattern is rather different. The field in the overcoat is ex-
tremely high, and this leads to a very high absorption level. The vast majority of
the energy is deposited in this layer. That these theoretical plots represent the real
world has been confirmed with measurements made on actual mirrors.
For the overcoat on a metal mirror to be durable, it has to be one or more

Air

•
Distance •
Figure 4.6. E2 plot for overcoated silver in the center of the reststrahlen band (9.3 ....m).
Because there is a node at the surface of the metal, little energy is absorbed in the overcoat.

DESIGN TYPES 89

50 ,-----------------------------------------------.----~

40

•
o 30
Ii
i.
•C~
.a
20
1/1.

10

o ~----~----~----~----~----~~----r_----~----~
8 9 10 11 12

Wavelength (Jim)
Figure 4.7. Absorption of a free-standing silicon dioxide film. The absorption occurs in the
reststrahlen band and is significant when compared to a similar film used as an overcoat on
a high reflecting surface.

visible quarter waves in optical thickness. Thinner overcoats will provide propor-
tionately less protection. Thicker films may have stresses high enough to cause the
entire film system to fail by crazing. It should be noted that metal films do not
have the shear strength of dielectric films, and therefore metal films will fail at
stress levels much less than that which cause the failure of dielectric films. In
addition, the previously mentioned standing wave electric fields have antinodes
inside the protective layer. This enhances the absorption in those regions where
there is some absorption present in the protective overcoat, as discussed previously.
The protective overcoat, which may consist of only one layer, will reduce the
reflectance of the metallic surface if it is nearly a quarter wave thick, as shown in
Fig. 1.7. This problem may be minimized if the thickness is selected to be a quarter
wave at a wavelength remote from the region of interest. An example of this is the
design of Fig. 4.8, where the overcoat has a physical thickness that is only about
I percent of the nominal wavelength of 9 /-Lm. Another solution is to select a film
that is a half wave thick at the wavelength of interest, but once again both stress
and absorption have to be considered. A trade-off needs to be made between several
desiderata to arrive at the best solution.
Aluminum metal mirrors are common because they are relatively inexpensive,
90 CHAPTER 4

100 ,-------------------------------------------------~
Bare allYer

98
Ablorption due to coating
SIlYer with 0.1J1m
5102 oyercoat

II
U 96
..
C
III
U
II
;
II
II:
94
'lP-

92

90~-----.----~----~~----P_----~----_P----~----~
6 9 10 11 12

Wavelength (JIm)
Figure 4.8. Reflectance of a silver substrate with and without a silicon dioxide overcoat at
normal incidence as a function of wavelength. The reflectance scale is expanded. Note the
small amount of absorption as compared to the configuration used in Fig. 4.7.

and can be made with a minimum investment in coating equipment. This may
include a vacuum chamber, a heavy stranded tungsten filament wrapped with alu-
minum wire, and a low voltage AC or DC power source to supply a current to heat
the filament. Aluminum adheres well to glass, therefore, the film does not require
a preparatory adhesion layer as does gold, for instance.
The control of the thickness of aluminum films is important to get the optimum
specular reflectance from them. Overly thick aluminum films, on the order of 100
om or more, begin to scatter light because of surface roughness. This roughness
increases rapidly with thickness beyond this point. On the other hand, films that
are too thin will transmit light, and the reflectance will not be the highest possible.
The optimum reflecting aluminum mirror will just barely transmit light from a
strong incandescent light bulb as viewed with the naked eye.
An aluminum film oxidizes with time, starting immediately upon contact with
air. The oxidation continues, albeit at a reduced rate, for the entire life of the film.
In some situations, this oxidation can proceed at an accelerated rate, as when both
humidity and UV light are present simultaneously. The oxide film does have its
good points. It does slow down the oxidation rate as it gets thicker, and does serve
to protect the remaining metal film. Because aluminum oxide is transparent over a
DESIGN TYPES 91

very wide spectral range, it does not interfere with the metallic reflectance, at least
not in thin thicknesses.
Although aluminum reflectors have less reflectance than silver, they are very
useful because they are good ultraviolet reflectors, whereas silver is quite poor. On
the other hand, there is a dip to 86 percent in the reflectance of aluminum in the
near infrared at about 825 nm. Furthermore, the average reflectance of aluminum
in the visible portion is around 92 percent, as compared with silver's 98 percent.
Thus, when there are many reflecting surfaces in a system, it is obvious that there
are going to be significant light losses if 92 percent reflectance is all that is available:
5 surfaces will reduce the available light by almost 34 to 66 percent. If these 5
surfaces had a fresh silver coating with a 98 percent reflectance each, the combined
reflectance would be about 90 percent.
4.2.2. All Dielectric DeSigns
4.2.2.1. Nonsymmetric Stacks. The alternating high and low index stack of
layers, the optical thicknesses of which are all a quarter wave, is one of the oldest
multilayer designs. The performance of such an arrangement of layers is shown in

100~==============~r=~------~====================~
1"-__ / _-----
I--....... ,~\\ /~"'-----------------::::.-

S Plane----/"

-
I\,-r I

\,\ I"'
I '", ",/ /
75 \\ / I '--'
CI
U

-
C
III
u
CI
;:
CD
a:
50
\\} ,p
\\
,I \
, I
~
I
I
j4AVerag e

,
I
I
II
I
Plane

I ,
?P- I I
I ,
I I
I ,
25 I
I I
,
I ,
I ,
I I
I I
I I
I I
\/
O~------~~-------r--------r-------~--------~
11
7 8 9 10 12
Wavelength ()1m)
Figure 4.9. Reflectances of a silver substrate with a silicon dioxide overcoat on an opaque
silver substrate at a 50-degree angle of incidence. The reflectances for both planes of
polarization are shown along with their average.

92 CHAPTER 4

P plane
1:100%

Air Si02 Ag

S plane

~ ~ 4
1=100% ~
If

Distance
~

Figure 4.10. E2 at 60 degrees for both planes of polarization for a silicon dioxide overcoated
silver mirror. The wavelength and the thickness of the overcoat are the same as that of Fig.
4.6; the two figures can be used to compare the standing wave fields. The penetration of
the s plane field into the overcoat and the substrate is minimal, whereas in the p plane,
there is a great deal of interaction.

Fig. 3.3. There are several parameters that characterize such a design. The index
ratio of the high and low index materials is probably the most important one as its
value determines the width of the high reflection band, and, together with the
number of pair of layers in the design, determines the peak reflectance of the coating.
The width of the high reflectance band is given by Eq. 3.1. This high reflectance
DESIGN TYPES 93

band is sometimes referred to as the stop band. This terminology derives from the
fact that the transmission is blocked (stopped) by the high reflection in this wave-
length band.
The definition of the edges of the stop band is generally given as those wave-
lengths for which the reflectance only increases (theoretically) as the number of
layers in the stack increases. This definition is very practical, in that it specifies
just how wide the design will be in the limit of an infinite number of layers. The
peaks and valleys between stop bands also have an envelope that does not depend
on the number of periods. Macleod (1969, Eq. 5.9) shows that the peak transmittance
at the center wavelength of a quarter wave stack (ns/(LH)p/no) in air is given by

R=

where p is the number of layer pairs, g is the ratio of the high index value to the
low one, and
nH is the index of the higher index material.
A conservative estimate of the number of layers required to attain at a minimum
a given level of transmission can be obtained by dropping the (niI1ns) factor from
this equation. The effect of this term is to increase the reflectance over that calculated
on the basis of the exact equation as the value of this term is greater than unity in
most practical situations. The result of this simplification is

R=

The plot of this function is shown in Fig. 4.11. The importance of the index ratio
is readily apparent by the rapid increase in the reflectance with a relatively small
increase in index ratio.
The reflectance of a quarter wave stack is frequently shown plotted against a
wavenumber or a normalized inverse wavelength abscissa. A typical plot of this
type is shown in Fig. 3.5. The advantage of this type of presentation is readily
apparent: the periodicity of the stop band patterns are very evident and predictable,
therefore, only one order needs to be plotted, and all of the other orders can be
calculated from that one. It should be noted that dispersion in the index and the
absorption coefficient has been ignored in these theoretical calculations. These
parameters can result in significant deviations between the calculated performance
and that obtained in a real coating.
Theoretically, the number of stop bands in a quarter wave design is infinite.
An order number is associated with each stop band. The order of the stop band
that occurs at the longest wavelength, i.e., where all of the layers are a single
quarter wave thick, is 1. Note that there are no stop bands at wavelengths longer
than the wavelength at which the stack layers are a quarter wave.

94 CHAPTER 4

1.0

10- 1

'i
ii
u 10- 2
rn
aI
0
..I
CD 10- 3

..E
U
C
II

III
C 10- 4
II
~

I-

10- 5

10-6i-----------~------------~----------~----------~
1.0 1.5 2.0 2.5 3.0

Index Ratio
Figure 4.11. Approximate maximum transmittance as a function of index ratio at the stack
center wavelength for a stack with quarter wave thick layers. The substrate and incident
medium indices are both 1.0. The numbers next to the curves indicate the number of pairs
of layers in the stack.

A stop band can exist wherever the total optical thickness of a period is equal
to a half wave. Whether it actually exists or not depends on the individual layer
thicknesses in the period. At those wavelengths where the layer thicknesses are all
an integral multiple of a half wave, the entire stack is absentee. At the wavelength
that corresponds to the second order of a quarter wave stack, all of the layers are
a half wave thick. Consequently, no stop band exists at the second order location
for this type of stack. The same is true for all even orders for this type of design;
only the odd orders actually have stop bands. Note that in the plot of the theoretical
performance (Fig. 3.5), only the stop bands with odd order numbers exist.
For a given basic stack period, the width of the stop band is constant on a
wavenumber scale. The width does not depend on the order number. The percent
width decreases as the order increases. On the other hand, on a wavelength scale,
both the width and the percent width decrease as the order increases.
It is straightforward to design stacks that skip any given order, such as skipping
every third one. To do this, one replaces the basic HL period with one that has a
different proportion of one material to another; e.g., a stack with a basic period of
HLH will have the first and second order high reflectance bands present, but every
third stop band will be missing. In practice, because of dispersion that occurs in
most materials, only one order can be made exactly absent by ascertaining that the
DESIGN TYPES 95

optical thicknesses are exactly correct at that one wavelength. At other wavelengths,
the match will not be exactly right, and the order will be more or less evident.
When a basic period other than the simple HL type is used, the peak reflectance
will not be as high as that of the basic HL design, given the same number of periods
and the same kind and amount of coating material.
Thelen (1963, 1973) has published papers indicating how one goes about
designing a filter that has a number of orders suppressed. The reader who is interested
in this type of design is referred to these papers and to the companion volume by
Thelen (1987).

4.2.2.2. Symmetric Designs. In the previous section, we discussed designs

in which the optical thicknesses of all of the layers were equal. The resulting periods
that made up the stack, however, were not necessarily symmetric. For example, a
(HL) stack does not have symmetry, though it has periodicity. In this section, we
treat the case where the basic period is symmetric about its center. The center of
a symmetric period can be either in the middle of a layer, such as in (HLH), or it
can be at an interface, such as in (HLLH). Herpin (1947) and Epstein (1952) have
shown that if the basic stack period is symmetric, then the mathematical represen-
tation of the period is especially simple. The entire symmetric structure can be
reduced to an equivalent single layer with a real equivalent index in wavelength
regions outside the stop band. Such a period is referred to as a Herpin layer, with
a Herpin equivalent index, nH' An equivalent thickness exists for this basic period
that is analogous to the thickness of a single layer. This is a useful concept for
designing thin film filters. Figure 4.12 shows the performance of a symmetric design.
Note the similarity of this performance to that of the quarter wave stack. As
previously noted, the quarter wave stack (all layers are a quarter wave thick) gives
the highest reflectance for a given number of layers. The symmetric stack always
has somewhat less reflectance than a comparable quarter wave one. The major
difference between them, however, is in the transmission zones between the stop
bands. In this wavelength range, the Herpin equivalent index is real, and the
transmission can be high or low, depending on the relative indices of the medium
and the substrate and the equivalent index of the Herpin period. It should be pointed
out that the Herpin equivalent index is dependent only on the thicknesses of the
layers in the basic period and the indices of these layers. It is not dependent on the
number of periods in the stack. On the other hand, the Herpin equivalent thickness
does depend on the number of periods in the coating. Figure 4.13 shows on a
wavenumber abscissa the reflectance of a symmetric stack over a broad wavelength
range. The repeat pattern of the performance is evident. The equivalent index is
different on either side of the stop band.
The regularity seen in Fig. 4.13 for the performance of a periodic stack may
not be observed if the number of periods is small and the index of the substrate
and the incident medium are not equal. Figure 4.14 shows the performance of a
single period of a quarter wave stack and that of a larger number of periods. In
this figure, the index of the incident medium is unity, whereas that of the substrate
is 1.5. It is obvious that the performances of the two designs are quite different.

96 CHAPTER 4

100TA~i\-----r~----------------------------====~

75
L
"0=1.0 ( i H 2L)5 "5=1.0

"-0 = 2.0Jlm
II

..g
U
C
IV

50
;:
II
II:
~

25

O~~~~~~--~~---T------~~~~~----~----~
0.2~ 0.50 0.75 1.00 1.25 1.50 1.75 2.00

Wavelength (Jim)
Figure 4.12. Theoretical reflectance of a (U2 H U2)5 stack on a wavelength scale. There
are no other high reflection bands at wavelengths longer than that of the first order. The
pattern of high reflectance bands repeats as the wavelength decreases.

The dispersion in the indices of refraction of the materials used in a stack can
disrupt the regularity of the stop bands and the high transmission bands. In addition,
a change in the index of refraction can lead to a mismatch in the optical thickness
of some layers. Normally, the thicknesses of the layers of a quarter wave stack are
adjusted so that they are all matched in optical thickness at some wavelength.
Dispersion in the index of refraction can make orders appear where otherwise they
should be suppressed as the layers are not all a quarter optical thickness at wave-
lengths other than the match wavelength.
Orders can be suppressed in symmetric designs in the same way they are in
the previously discussed non symmetric ones. Figure 4.15 shows a design where
the third harmonic has been suppressed, whereas the second harmonic has a high
reflection stop band.

4.2.3. Enhanced Metallic Reflectors

To improve the performance of metallic mirrors, coatings have been developed that
enhance their reflectance. In the simplest case, the design consists of a stack of
layers with alternating high and low indices. With the exception of the first layer
next to the metal, each layer has an optical thickness of a quarter wave at the


DESIGN TYPES 97

wavelength at which the highest reflectance is desired. Figure 4.16 shows the
performance of a design that enhances the reflectance of aluminum in the visible.
The first layer's optical thickness is slightly less than a quarter wave to account for
the phase shift that occurs at the metallic interface.
A trade-off is necessary in as much as the higher reflectance in this type of
design is obtained at the price of bandwidth. At wavelengths outside the stack's
high reflectance band, the reflectance of the combination of metal and coating is
less than that of the bare metal. The electric field profile in such a design is
interesting. Figure 4. 17 shows such a calculation, and it is seen that the electric
field does not reach down to the metal with any significant amplitude, thus indicating
that the resulting mirror will have little absorption. At wavelengths away from the
dielectric stack center, the situation is rather different. Figure 4.18 shows the electric
field at the minimum of the reflectance curve in Fig. 4.16. It is obvious that at this
wavelength much of the field penetrates down to the metal, and that absorption is
enhanced over that which would take place for the bare metal. Some filters are

Wavelength (}1m)

0.25 0.4 0.5 1.0 2.0

100
~ "'\ "'\ n

75

CD
U

..
C
III
U
CD
;: 50
CII
II::
bI?

25

o
40,000

~A 30,000
AI
20,000
lA
10,000
~
o
Wavenumber (cm- 1 )
Figure 4.13. Theoretical reflectance presented on a wavenumber scale of the same (U2
H U2)5 stack shown in Fig. 4.12. Notice that with this representation the stops bands produce
a periodic pattern.
98 CHAPTER 4

Wavelength (JIm)
0.25 0.5 1.0 2.0
100

(~H ~)5
75

nL = 1.35
II

..
U nH= 2.50
C
III
U 50
...
.!!
II
a:
~

25
(~ H~)

o~-L~~~~--~--~~~----~----~~~--~--~~
40,000 35,000 30,000 25,000 20,000 15,000 10,000 5,000

Wavenumber (cm- 1)
Figure 4.14. Reflectance of 1- and 5-period symmetric stacks on a wavenumber scale.
Both the incident medium and the substrate have an index of 1.0. When the number of
periods in a symmetric design is low, the peak reflectance does not occur in the center of
the stop band.

optimized for this feature and are called induced absorbers; they are discussed in
further detail in section 4.6.1.
The advantage of this type of high reflector design is in its economy of layers.
Because a full quarter wave stack does not have to be built up to obtain high
reflectance, the resulting mirror will generally be less costly. As the enhanced metal
design requires fewer layers and total coating material than an equivalently reflecting
design built up only with a stack of quarter-wave thick dielectric layers, the amount
of scatter in the coating can be less. In practice we can, therefore, obtain a higher
absolute level of reflectance from one of these designs. Because most computer
programs that calculate thin film performance do not include scattering in their
mathematical model, this aspect does not show up in theoretical plots.
In addition to the reduced width of the high reflection band, the other consid-
eration that needs to be evaluated is the environmental durability of the coating. In
general, a metallic coating will not be as durable as one with equivalent reflectance
built up of all oxide materials. Thus, if the coating is going to be exposed to a
DESIGN TYPES 99

harsh environment, such as salt fog or high humidity, then perhaps the all dielectric
design would be the preferable choice.
The reduction in the high reflectance zone can be overcome by adding stacks
to the design. The center wavelength of these additional stacks can be located to
optimize the reflection over a broad range of wavelengths.
If absorption and scatter are left out of the computer model, any arbitrarily
high value of reflectance can be obtained. Carniglia and Apfel (1980) have shown
how to maximize the reflectance of a dielectric quarter wave stack when one of the
materials has more absorption than the other.

4.3. EDGE FILTERS

Edge filters are similar to the symmetric period high reflector designs introduced
in the previous section. The region of interest in this case is the edge of the pass
band, rather than the peak reflection and the width of the stop band. In this type

Wavelength (}1m)
0.25 0.5 1.0 2.0
100

75 -
CD

..
U
C
III
u
.! 50
Missing third order
stop band
....
GI
a:
~
l
25

WJiI
o W//
.
40,000 35,000 30,000 25,000 20,000 15,000 10,000 5,000

Wavenumber (cm- 1)
Figure 4.15. Reflectance of a suppressed third order design. Every third order is sup-
pressed in this design, the basic period of which is (L H L). The multiple reflection peaks
between the highly reflecting stop bands can be effectively reduced to allow good trans-
mission between a pair of stop bands.

100 CHAPTER 4

'OO1:============~~------------------------'

95

Bare AI
II
U 90
c
:!u
.!!
"i
II:
85 AI ~ L' H(LH)2 Air
IF dL.H =450nm OWOT
d~ = 350nm OWOT
"L =1.45
80 nH =2.35

75+-------.-----~------~------~------~----~
400 500 800 700
Wavelength (nm)

Figure 4.16. Theoretical reflectance of bare and enhanced AI reflectors. Note that the
ordinate is an expanded scale. Although a higher reflectance can be obtained in one part
of the spectrum with enhancing layers, other portions of the spectrum have a reduced
reflectance.

AI
Substrate
Distance
•
Figure 4.17. Plot of the time average of the electric field in an enhanced mirror. The field
is very close to zero at the metal interface. The wavelength corresponds to that where the
reflectance reaches a peak in Fig. 4.16. The indices of refraction are nL == 1.45 and n" ==
2.35.

DESIGN TYPES 101

Air H L H L H L AI
Substrate
Figure 4.18. E2 for an enhanced AI reflector at the wavelength corresponding to the first
reflectance minimum point. The scales of this plot are the same as that of Fig. 4.17. Note
the relatively large amplitude of the field in the absorbing substrate, which leads to the
enhanced absorption of almost 25 percent. At this wavelength, the coating tends to act as
an absorption inducing or antireflection design.

of filter, it is generally desired that (1) the transition from reflecting (stop band) to
transmitting (pass band) be as sharp as possible and (2) the transmission zone be
as close to 100 percent as possible. Figure 4.19 shows an expanded view of an
edge filter. The basic design in this filter is the same as in the previously mentioned
symmetric high reflector, except that some additiona11ayers may be added to reduce
the reflection losses in the transmitting region. These additional layers can be
designed using antireflection concepts as the high reflector stack can be represented
by real Herpin equivalent indices and thicknesses in this region.
A nomenclature problem sometimes exists with edge filters. The use of terms,
such as short pass and long pass, are to be avoided as they can mean different
things to different people, depending on their background. On the other hand, the
use of terms, such as short wave pass or long wave pass, leaves no room for
ambiguity.
Another terminology exists with these filters. A hot or heat mirror is one that

102 CHAPTER 4

100~--------------------------------------~==__- - - ,

75

.,
U

...,
C
IV
u
50
.,
;;
a::
~

25

510 520 530 540 550 580

Wavelength (nm)
Figure 4.19. Reflectance of an edge filter. Note that the wavelength scale is expanded to
show the shape of the edge. The steepness can be increased by adding periods to the stack.,
but then humps can begin to appear in the transmitting portion.

reflects the infrared and transmits visible light well. Conversely, a cold mirror
reflects visible light and transmits infrared heat. In this latter case, the reflected
light is cold, i.e., it contains relatively little infrared energy.
There is one major distinction between short wave and long wave pass edge
filters. The long wave pass filter that has been designed at the edge of the first
order stop band has no stop bands appearing in the pass band (at wavelengths longer
than the edge). There is no limit to how far the pass band extends in this case.
Dispersion in the real coating materials generally limits the range over which the
coating will be useful. On the other hand, the short wave pass filter has a pass band
constrained by a higher order stop band. Some of these can be suppressed to widen
the pass band (see Thelen, 1963, 1973).
Edge filters are generally made from all dielectric materials, and therefore,
exhibit the typical angle shift described in section 3.6. The shift is dependent on
the average index of the materials that make up the design. As a result of the higher
index materials available in the infrared portion of the spectrum, infrared filters
tend to shift less than equivalent filters for the visible portion of the spectrum,
where available indices are much lower.
In addition to the shift in the performance of a filter as a function of the angle
of incidence, another effect on the performance is introduced by polarization.
Because the width of the stop bands are different for each of the two planes of
DESIGN TYPES 103

polarization at nonnormal angles of incidence, the filter has polarizing properties

near the edge. This can be useful if a polarizing filter is desired. If a nonpolarizing
edge is the requirement, however, it can be built according to a recipe given by
Thelen (1981).

4.4. BANDPASS FILTERS

Bandpass filters allow only a relatively narrow portion of the spectrum to pass
through them, while blocking other portions, either by reflection or by absorption.
Edge filters can be considered bandpass filters of sorts, but generally this terminology
is reserved for filters with transmission bands that are narrower than 5 to 10 percent
of the central pass band wavelength.
The following discussion of these filters is organized along the lines of their
construction, and a given width of pass band may be obtained with more than one
design. Other considerations, therefore, have to be used to determine which filter
is best for a given application. Bandpass filters come in nonabsorbing and absorbing
versions. In either case, it is important for the user to know just how much trans-
mission can be expected from a typical production filter. Even though a filter is
fabricated entirely with dielectric materials, this does not mean that it will have
100 percent transmission. In narrow bandpass filters on the order of 0.25 to 0.5 of
a percent in width, the peak transmission can be well under 50 percent.

4.4.1. Two Edge Filters in Series

When a relatively wide pass band is desired, it can be obtained by placing two
appropriately positioned edge filters in tandem. One of these will be a short wave
pass, whereas the other will be a long wave pass filter. A suitable set of edge filters
enable an experimenter to assemble a bandpass filter to suit the application. Pass
bands in the range of 5 or 10 percent to as little as 2 percent may be obtained with
this method. The primary limitation on the minimum bandwidth that can be obtained
this way is the steepness of the slope of the edge filters.
If custom filters are to be made by placing two bandpass filters in series, there
are several options possible: (1) put both filters on the same side of one substrate,
(2) put the filters on the opposite sides of one substrate, and (3) put the two filters
on separate substrates. Cost considerations enter into this decision: two substrates
cost more than a single one as four sides have to be prepared and polished. If two
substrates are used, then it may be necessary to put an antireflection coating on the
other two surfaces. Thus, the choice of two substrates is not an especially inex-
pensive one. There are exceptions to all rules, and this one is no different. If the
filters are especially difficult to make, i.e., the edge filters have to be precisely
positioned, then it may be less expensive to coat separate substrates to obtain
adequate yields. Additionally, the use of two substrates allows the use of the angle
shift to fine tune each edge separately.
Putting both filters on the same side of the substrate is theoretically the least
costly, as only one surface needs to be cleaned, and only one vacuum cycle is
required to complete the filter (if an AR coating is not required on the second
104 CHAPTER 4

surface). If the substrate is thin and stresses are a problem, this approach may result
in a warped substrate, and the prudent approach is to coat the two-edge filters on
opposite sides of the same substrate.
A bandpass filter made of two edge filters has the highest transmission char-
acteristics because the two halves of the filter can be designed independently to
have good transmission. Transmission values of close to 100 percent can be ex-
pected. On the other hand, the width of the filter cannot be made very narrow.

4.4.2. Fabry-Perot Filter

The classical Fabry-Perot filter, frequently called an etalon, consists of two plane
parallel surfaces with silver reflecting coatings; both of these surfaces require a
slight amount of transmittance. These mirrors are separated by a small air space.
Only those wavelengths that are approximately integral multiples of the air gap
spacing are transmitted by this element. To function properly, the surfaces must
be parallel to very tight tolerances, and therefore, are difficult to adjust. A solid
version of this etalon can be made by replacing the air gap with a transparent thin
film. Although such a device is no longer adjustable as the gap between the two
reflectors is not variable, it does retain the filter performance of its predecessor and
is much more durable.
Table 4.2 gives the prescription for a thin film (solid) Fabry-Perot filter and
Fig. 4.20·shows its performance. The mirror layers consist of silver films that
absorb some light; thus, the transmission does not reach 100 percent in the pass
band. Away from the pass band, the absorption is reduced as can be seen in Fig.
4.21.
The width of the pass band in a Fabry-Perot filter is given by
2
'Y = ~

where

The reflectance of the two cavity mirrors, Rf , is assumed to be equal; the trans-
mission through these mirrors is represented by Tf . The peak transmission through
the etalon is

TmaJ<. =
Tf )2
( 1 - R
f

A final parameter that is useful with this type of filter is the finesse, which is the
ratio of the separation between the peaks to the width of the pass band. This is
given by

F=--.
1TW
2
A more complete discussion of Fabry-Perot filters is given in Chap. 16 of Stone
(1963) and Chap. 7 of Born and Wolf (1970).
 DESIGN TYPES 105

Table 4.2 Design of a Solid

Fabry-Perot Filter·
Incident Medium: n = 1.0

Silver (n = 0.05, k = 3.0) 15 nm

Spacer (n = 2.35) 511 nm
Silver (n = 0.05, k = 3.0) 15 nm

Substrate: n= 1.52
'The performance is shown in Fig. 4.20.

4.4.2.1. Variable Spacer Designs. An air spaced Fabry-Perot filter can be

used as a tunable filter by varying the spacing or index between the mirrors.
Piezoelectric transducers can be used to vary the mirror spacing smoothly, thus
eliminating vibration-induced manual adjustments. The index can be varied by
changing the gas mixture or the gas pressure between the mirrors.
The spacing between orders can be increased by using two cavities in tandem.
These cavities must not be coupled; i.e., there cannot be multiple reflections between

100~--------------------------------------------------,

f.....-T I
I,
I, I
75 I , I
I , I
I , I
I , I
, I I
II
()
I I I
I I

-
C ,
III I
I ,
()
41 50 I
/
;:
I

41
a: , \\ l I
I
If< I I I
I
I
,
,
/
I
, ,,/
/

25 I
I \
\
, I
I
/
'---'"
I \ /

I
/ '--"'"
"
o~------~-------r------~------~------~--------f
400

500

600

700

Wavelength (nm)
Figure 4.20. Theoretical reflectance and transmission of a metallic Fabry-Perot cavity filter
as a function of wavelength. The peaks do not correspond exactly with the optical thickness
of the spacer layer because of the effect of phase shifts that occur at the spacer/metal
interface.
 106 CHAPTER 4

20,-----------------------------------------------~

15

.,
U

...c:
II
0.
~ 10
0
I)
.a
<
iI-

o~------~------~------_,------~~----~~----~
400. 500 600 700
Wavelength (nm)
Figure 4.21. Absorption of a Fabry-Perot filter as a function of wavelength. The design is
the same as that used in Fig. 4.20. The absorption takes place at those wavelengths where
the transmission is high as it is at these wavelengths that the electric fields penetrate into
the coating to the absorbing (metallic) layers.

a mirror in one cavity and a mirror in the other cavity. This can be easily obtained
by introducing a slight tilt between the two etalons. If a pass band of one is chosen
to be at the same wavelength as the other, and they have different cavity spacings,
then many of the other hannonics do not coincide and are blocked by one filter or
the other.
The shape of a Fabry-Perot pass band has a roughly triangular shape; the exact
shape for normal incidence is given by the Airy formula:

Trnax
T = ---"--- 2
+ '" sin E'
where
nd
E=<I>+-
A.o ,
<I> is the phase shift upon reflection from the metal mirror surface, d is the mirror
spacing, n is the index of the spacer medium, and A. o is the wavelength. The use
of thicker gaps only makes the width narrower without changing the overall shape
very much. As the width is narrowed, the peak transmittance decreases also.
Silver is the only practical metal to use to make Fabry-Perot etalons for use
in the visible portion of the spectrum. All other commonly used metallic thin films
DESIGN TYPES 107

have too much absorption. In the infrared, gold and a few other metals can be used.
All of the metals that make good, low loss reflectors are delicate and easily damaged.
This mirror problem, coupled with the alignment sensitivity problem, make air
spaced Fabry-Perot filters with metal mirrors useful in only a few special appli-
cations.
To circumvent the problems that one encounters with metal mirrors, all dielec-
tric mirror coatings have been developed to replace the thin film metal coatings.
Several benefits arise from this: (1) higher reflectivities can be achieved; (2) etalons
can be made that can operate in spectral regions where there are no good metal
film reflectors, such as in the near UV; and (3) the dielectric films are much more
durable than the metal ones and can be cleaned without damage. Of course, the
same limitations that were described in the section on the bandwidth of dielectric
mirrors apply here also. If a wide band of wavelengths is to be covered, then several
stacks are required to get the reflectance to a high value over the entire band.
A problem common to all real etalons, whether they use metal or dielectric
mirrors, is dispersion in the optical constants. Over a small portion of the spectrum
the wavelength is proportional to the space between the reflectors. As the dispersion
in the materials and designs vary, different phase shifts are introduced at different
wavelengths, so that the wavelength that fits exactly in the cavity (resonance) is a
function of the phase shift as well as the mirror spacing. Typically, the only variable
that is considered when one uses a Fabry-Perot filter is the cavity spacing. In many
instances, this is adequate as they are used only over a small wavelength range. If
a large range is to be covered, then calibration of the wavelength versus spacing
function should be undertaken at known wavelengths over the entire spectral range
of interest.

4.4.2.2. Solid Fabry-Perot Filters. As mentioned in the previous section,

the spacer medium between the mirrors in a Fabry-Perot filter does not have to be
air. If the gap is a thin solid film, the performance will be the same with the
exception that the index of the gap is changed. In particular, this has an effect on
the angle sensitivity of the filter. The higher index the spacer layer is, the less the
angle shift, as shown in Fig. 4.22. Of course, these filters do not have the normal
incidence tunability that an air gap Fabry-Perot filter has, but the filters can be
tuned over a narrow range by tilting them. If such a fine tuning technique is to be
employed in a filter, it is necessary that the normal incidence pass band be at a
wavelength longer than that at which one desires to operate the filter, as the angle
shift is only toward the blue.
If more than one Fabry-Perot cavity filter is to be used in tandem, they must
be placed on separate, nonparallel surfaces or different optical elements so they do
not add coherently. If they are placed on the same surface, they will interact by
forming an additional uncontrolled Fabry-Perot filter in which the substrate thick-
ness serves as the cavity spacer.
When narrow band filters are used, the coherence of the light increases. This
can be a source of difficulty if the light is sufficiently coherent that the thickness
of the substrate is less than the coherence length of the light. In such a case, even
coatings on opposite sides of a substrate can interact in a coherent manner and give
 108 CHAPTER 4

30

25

~ 20
IL
i
UJ
~ 15

10

o 10 20 30 40 50 60 70 80 90

Angle of Incidence
Figure 4.22. Peak transmission wavelength shift (given as a percentage of the peak wave-
length at normal incidence) of a Fabry-Perot filter as a function of angle of incidence. The
shift is illustrated for a number of spacer layer indices. The index of the metal reflector layers
is (0.05. 3.0). The direction of the split of the peak positions for the two planes of polarization
depends on the index of the spacer layer and that of the reflector metal.

rise to unexpected effects. Thus, it is prudent, when one is using very narrow
bandpass filters, !o treat the system as if it were illuminated by a laser, and to use
precautions similar to what would be used to eliminate interference effects in such
a system.
4.4.3. Narrow-Band Dielectric Filters
Narrow-band dielectric filters resemble solid Fabry-Perot filters, however, the shape
of the pass band is much different. These filters have squarer tops to their trans-
mission band, thus allowing more energy in the pass band through the filter. Design
techniques allow the shape to be squarer still if three materials are used in con-
structing the filter. Figure 4.23 shows the performance of an infrared square band-
pass design; a comparable simple thin film Fabry-Perot filter is also included in
the illustration for comparison. This type of design is ideal for an application where
maximum energy in a band is to be transmitted and maximum rejection of energy
in an adjacent band is desired. This contrasts with the Fabry-Perot design that may
be a good choice for isolating an atomic emission line, for example, and where the
emission line shape is the dominant factor in determining the filter's performance.
The design of these square-topped filters generally consist of coupled Fa-
DESIGN TYPES 109

bry-Perot cavities, with 3, 5, or more cavities in a single filter. The cavities are
at least one, and frequently more than one, half wave thick. The remaining layers
are usually all a quarter wave thick at the center wavelength. By incorporating a
third material with an appropriate index into the design, the super square types of
filters can be achieved. The details of the methods used to arrive at these designs
is beyond the scope of this book (see Jacobs (1981) for more information).
Square bandpass filters are sensitive to the angle of incidence similar to all
other dielectric interference filters. The performance shifts to shorter wavelengths
as the angle of incidence increases from normal incidence. To minimize this effect,
the designer usually selects a high index of refraction material for the spacer layer,
as this is the layer that affects the shift the most.
Several parameters are required to describe the performance of narrow bandpass
filters. In addition to the transmission at some wavelength, it is common to specify
the width of the filter at specific transmission levels, such as the width at the 10
percent points, or the full width at half maximum (FWHM). Figure 4.24 shows
some common specifications.
The standing wave electric fields set up in these filters by the incident light
can become quite high. Figure 4.25 shows the standing wave fields for a three-
cavity filter. If there is some absorption in the layers with high electric fields, the

100 ~--------------------~__~__, - - - - - - - - - - - - - - - - - - - . ,




Square Band Paas

75


GI
U Fabry Perot
...
C
III

'E 50
•c
III
f::
,
,,,
~
\
\
25 \
I \
I
,/ " " ' ..........
, .... ",'"
------ "
... ...
-----
o +-----------~~------~~--------~~--------~
900 950 1000 1050 1100

Wavelength (nm)
Figure 4.23. Transmission of a square bandpass and a single cavity Fabry-Perot as a
function of wavelength. The design of the square bandpass filter is (LH)2 F (HL)4 (LH)4 F
(HL)2 and that of the Fabry-Perot is (HL)2 F (LH)2. The indices of the high and low layers
are 3.5 and 2.0, respectively.

110 CHAPTER 4

TpEAK

GI
U

..
C
III

'E
I/)
FWHM

.
c
III
t-

~~----------~'----4----~~----------~~A


Wavelength
Figure 4.24. Common specification points for a bandpass filter. Not all points are usually
specified. TAB is the blocking range on the short wavelength side of the pass band, whereas
Teo is that for the long wavelength side. The center wavelength, AO, can be specified on the
basis of an arithmetic average of points on the slopes of the filter. The result will be different
if the abscissa is a wavelength or a wavenumber scale. The width is typically given as the
full width at half of the maximum value (FWHM). The steepness of the slope can be specified
by the width at the 10 percent point (.1 A10%).

total filter absorption becomes quite large owing to the amplification of the ab-
sorption by the field (Eq. 3.2). This absorption reduces the peak transmission
significantly, as shown in Fig. 4.26, where the peak transmission is plotted against
the value of the absorption coefficient for the previously mentioned three-cavity
filter.
The high fields can be used to determine the absorption coefficient in low
index materials. A Fabry-Perot filter is coated and the material to be evaluated
serves as the spacer layer. By varying the spacer thickness and measuring the height
of the transmission peak, the absorption in the spacer layer can be determined (see
Southwell, 1982).
Because the performance of the filters is a result of thickness-sensitive spacer
layers, small changes to these layers can have a significant effect on the overall
performance. For example, if we assume that there will be a 2 percent error in
thickness in the coated spacer iayer, the pass band of the resulting filter can vary
greatly as shown in Fig. 4.27. If the spacer layer is prone to absorb water, which
in turn changes its optical thickness, then, at wavelengths near the pass band, the
transmission will be a function of the relative humidity (Gee et aI., 1985).

4.4.4. Circular and Linear Variable Filters

Although the solid Fabry-Perot filter is not adjustable, except over a fairly narrow
range by angle or by temperature tuning, it is possible to use a wedged film
 DESIGN TYPES 111

configuration to obtain a pass band that varies as a function of the incident beam's
position on the surface of the filter. Such filters have been fabricated in both linear
and circular geometries. The pass band generally has the shape of the classical
Fabry-Perot filter.
A linear version of this type of filter was outlined in a paper by Wirtenson
and Flint (1976). The structure of such a film is shown in Fig. 4.28. As the distance
from a reference edge increases, the thicknesses of the layers increase as well.
Thus, the center wavelength of the pass band varies with position without changing
its shape.
Another variation on the same theme is the circular variable filter (CVF). In
this geometry, the wavelength transmitted by the filter varies as a function of the
azimuthal angle through which the filter is rotated. Complex rotating fixtures allow
linear thickness variations as a function of filter azimuth angle (Apfel, 1965; lllsley
et al., 1969; Thelen, 1965). By proper gearing, the mask and the substrate rotate
at precise speed ratios with respect to each other. This enables the formation of a
well-defined wedge for the spacer layer. The transmission of such a CVF is shown
in Fig. 4.29. These can be made in visible or infrared versions and find uses as

1=100%
T= 100%

R= 0%
Distance.
Figure 4.25. The time average of the square of the electric field in a 2S-layer, 3-cavity
square bandpass filter at the center wavelength of the pass band. The spacer layers consist
of the low index material. The high fields in the spacer layers make the design very sensitive
to any absorption in these layers. The high index layers are shown hatched.

112 CHAPTER 4

100 ~-------------------.~. .~~~~---------------,

75

II

..
Co)
C
CD
k =10- 2
E
III
50
C
...
CD
~
lJI.

25

Oi-----------~-=~------~--------~~----------~
900 950 1000 1050 1100

Wavelength (nm)
Figure 4.26. Transmission versus wavelength for the 3-cavity filter shown in Fig. 4.25. The
k values were introduced into the low index layers only.

wavelength selecting elements in field instruments, such as pollution monitors,

where portability and ruggedness are prime requirements. When silver is used for
the mirror layers, the CVF is frequently laminated with a cover glass and sealed
to improve the environmental durability.

4.4.5. Reflex or Phase Conjugate Filter

The manufacture of a narrow bandpass filter can be quite difficult. For example,
if a filter with a pass band tolerance of 0.25 percent is desired, then the thickness
of the spacer layer must be controlled to this accuracy. This is quite difficult to
achieve in practice, and very narrow bandpass filters are frequently obtained on the
basis of an individual selection process, or are made singly on a custom basis with
the performance of the filter monitored in real time during the coating process to
ensure accurate placement of the transmission band.
In contrast, the phase conjugate filter design allows the pass band deviation
to be determined on the basis of the average error in the entire coating, rather than
the absolute error in a single layer. Thus, it is much more likely that the phase
conjugate filter will be more successful than the standard approach.
The design of a phase conjugate filter was published by Austin (1970). Its
prescription is given in Table 4.3, and its performance is shown in Fig. 4.30. Figure
4.31 shows the variation in performance expected for a standard Fabry-Perot type
 100 _----------~-

75

"
U
C
11'1
....
~

E
0
50

.....
C
11'1

r.f.

25

o+-----------~~------~--------~~~--------~
900 950 1000 1050 1100

Wavelength (nm)
Figure 4.27. Effect on bandpass shape caused by thickness variations in the three half
wave layers. The design is the same as in Fig. 4.25. The standard deviation of the thickness
variations is 2 percent. The dashed curve shows the nominal performance.

no =1.5

ns =1.5
...
Azimuthal angle or poSition along wedge
Figure 4.28. Cross-section of a circular variable filter (CVF) or a linear variable filter. The
vertical scale is exaggerated for clarity. All layers have some degree of thickness gradation.
These filters are often cemented between two glass plates for durability.
113

114 CHAPTER 4

100

Angular Position
80 120 0
90 0 150 0
80 0 175 0
30 0
G)

..
U 50
C 60
ca
·S
CD
C
...ca
~
40

20

o
o
~
2.5
.
\J \./ \
3.0
.)
3.5
\.
. )
4.0
\)
4.5
\) \.
5.0

Wavelength (JIm)
Figure 4.29. Transmission versus wavelength plots for a CVF. The wavelength position of
the pass band depends on the azimuthal rotation or position along the wedge.

of narrow bandpass filter when a 2 percent standard deviation Gaussian distribution

of thickness deviations is applied to the thickness of each layer (including the spacer
layer). Figure 4.31 also shows the same thickness variation factors applied to the
layers of a comparable phase conjugate type of design. It is apparent that the phase
conjugate design has superior stability as compared with the traditional design
approach. A natural characteristic of these designs is to have an extremely narrow
(but stable) pass band. It appears difficult to adjust the bandwidths of these filters
to an arbitrarily wide value for a given index ratio of coating materials.

Table 4.3 Design of a

Phase Conjugate Narrow
Bandpass Filter·
no = 1.00
nL = 1.45
nH = 2.00
n. = 1.00

'The pertormance is shown in Fig. 4.30.


DESIGN TYPES 115

100 ,---------------------------~------------------_,

75

GI
c.I

..
C
III

E 50
•III
C
...
~

?P.

25

o~~========~======::~--------~::======~
600 625 650 675 700

Wavelength (nm)
Figure 4.30. Transmission of a phase conjugate filter as a function of wavelength. The
prescription for this design is given in Table 4.3.

4.5. BEAM SPLITIERS

4.5.1. Wide-Band or Neutral Density Beam Splitters
Beam splitters divide a ray of light into two beams. The performance of wide band
or neutral density designs is relatively independent of wavelength when compared
with that of dichroic types that have a strong wavelength dependence (see section
4.5.2.). Beam splitters are available in two configurations. The simpler one has the
coating exposed to air, whereas in the other, the coating is immersed in a medium
with an index close to that of the substrate. The latter type is frequently in a cube
geometry with the coating deposited on a face diagonal surface. Figure 4.32 shows
typical configurations of both types.
Each type of beam splitter has its advantages. The angle of incidence can be
easily varied and only one surface needs an antireflection coating when a plane
parallel substrate with a beam splitting coating on each side is used. This type of
design is relatively inexpensive when compared with a cube type. In the cube
configuration, the system is symmetric, and this can be an advantage in optical
systems where (I) the path lengths need to be matched or (2) the light returns
through the beam splitter and symmetric splitting is desired. In addition, the coating
is protected against the environment by the cement. Up to four surfaces may require
an AR coating in a cube design.
Figures 4.33a and b show beam splitters designed for 45 degrees, which

116 CHAPTER 4

- Standard Fabry-Perot
spr.ad 2.3nm 0.23%
f- Pha •• conjugate • pr •• d
1.4nm 0.14%

Nominal Nominal
100

75 -

-
E
Gl
U
r::::
III

1/1
50
-- I--
- r--
r::::

~
...
III

IF.

25

o J:iL 'J ~ ~j \. )~~ d

975 1000 1025 1050
Wavelength (nm)
Figure 4.31. Performance variation of two types of bandpass filters when random errors
are present in the construction of the filter. The standard deviation of the thickness errors
for both designs is 2 percent. The center wavelengths of the two designs were slightly offset
so that the performance of the two could be displayed together without overlap. The width
of the spread of bandpass· centers is about 0.23 percent for the Fabry-Perot deSign, and
0.14 percent for the phase conjugate one.

provides for orthogonal reflected and transmitted beams. Other angles can be used
as long as the coating is deSigned properly. Each design type is discussed separately
later.

4.5.1.1. Slab Beam Splitter. Because the surfaces and coatings of these
beam splitters are usually oriented such that the light is incident at nonnormal
angles, polarization parameters of the optical system need to be considered. Figure
4.33 shows the theoretical performance of a 45-degree beam splitter with equal
reflectance and transmittance at about 550 nm for the average of the two planes of
polarization. The coating could consist of a single layer of metal such as silver. At
this wavelength, however, the reflectance does not equal the transmittance for either
plane of polarization. The reflected and the transmitted beams in the p plane are
equal at about 455 nm, whereas the intensity of the reflected s plane beam is almost
2.5 times larger than the transmitted one at that wavelength. At 690 nm, the two
s plane beams are equal, whereas in the p plane, the transmitted component has
DESIGN TYPES 117

about 2.5 times the intensity of the reflected one. These unequal reflectances can
cause difficulties when a laser with a randomly polarized output is directed at such
a device. The beam splitter will convert the fluctuations in the plane of polarization
into intensity noise.
Beam splitters are generally polarizers, and therefore, need to be used with
care in an optical system. Multiple polarizing beam splitters in a system with no
other polarizers can lead to difficulties as the polarizing components are not explicitly
identified.
The coating used in beam splitters can be either a dielectric stack or a metal
layer. Most metal films can be used as beam splitter coatings, though some have
specific properties that are better than others. Figures 4.34 and 4.35 show the
performance of typical metal and dielectric designs, respectively. They both have
approximately the same (relatively low) reflectance at normal incidence at 500 nm.
The metal coating has the flatter curve, indicating that it will function well over a

A. Slab Beam Splitter

'LJL--AR COltlng

$emltranlparant
Coating

B. Cube Beam Splitter

AR COltinga
Semltranaparant COltlng

AR CoaUng

Cement

C. Priem Beam Splitter

Semllranape,ent Coating

AA Coating_

Figure 4.32. Beam splitter configurations. The types shown in illustrations A and B can
produce orthogonal incident, transmitted, and reflected beams. Type C cannot produce
rectilinear beams because of refractive effects at the exit surface of the transmitted beam.
 100

75

II
U

..
C
III

EII)
50

..
C
III
I-
?/l.

25

o
0.4 0.5 0.6 0.7

Wavelength (pm)

100 r-----------------------------------------------------,

75

II
U

..
C
III
u 50
GI
;;
GI
a:
?/l.

25

o L-______________~--------------~----------------J

0.7
0.4 0.5 0.6

Wavelength (pm)

Figure 4.33. Performance of a silver film slab beam splitter as a function of wavelength.
The transmittance (a) and reflectance (b) cannot be equal for the two planes of polarization
at the same wavelength. The angle of incidence is 45 degrees.
118
DESIGN TYPES 119

50

Air
40 7. Cr .
/IT/1.5271l

7o 4nm

~ 30

-
c
II
U
CD
;:
III
s Plane 60·

a: 20
~

Average 60·
10
Normal

p Plane 60·
o
400 500 600 700

Wavelength (nm)
Figure 4.34. Reflectance of a chromium film slab beam splitter film on a glass substrate
as a function of wavelength for 0 and 60 degrees angle of incidence. No dispersion is included
in the calculation. Note that this design makes a very good polarizer.

broad range of wavelengths. At an angle of incidence of 60 degrees, the perfor-

mances of the two designs change dramatically. In the metal coating, we find that
the p plane light reflects poorly or not at all, whereas s plane light reflects well.
Note that in this example the light is incident on the coating at approximately the
Brewster angle.
In the dielectric design, the p plane is relatively flat, whereas there is significant
chromatic variation in the s component. In some applications, such as in high-
power laser systems, the fact that the dielectric layers absorb no energy is a most
desirable characteristic.
A study of the angular performance of slab beams splitters is informative
because they may frequently be used at variable angles of incidence. Figures 4.36a
and b show the transmittance and the reflectance of a low loss metal beam splitter
as a function of the angle of incidence in air. It splits the incident light equally into
the reflected and transmitted beams at normal incidence. At angles above 45 degrees,
the intensity imbalance between the two beams becomes quite large. At angles
greater than 20 degrees, the intensity differences between the two planes of polar-
ization become significant. The losses in this system are small, as shown in Fig.
4.36c.
Figures 4.37a and b show the transmittance and reflectance of another metal
beam splitter coating as a function of angle. The coating material used in this design

120 CHAPTER 4

50r-------------------------------------------------~
Air
nL=1.45

~

40 H nH =2.35
1.52

.."
u 30
c
III
U

"
;:
-- ------------
a:" -- --
.... - ---
20
~
...
60·

10~~~~~ --_.--. p Plane

o~------,_------~------~------~------~~----~
400 500 600 700
Wavelength (nm)
Figure 4.35. Theoretical performance of a dielectric slab beam splitter as a function of
wavelength. Both layers in the design are a quarter wave thick at 500 nm.Note the substantial
variations in reflectance of the two planes of polarization as the wavelength varies.

is much more absorbing and consequently the value at which the reflected and the
transmitted intensities are equal is much lower than that in the previous design.
The average reflectance of the two planes of polarization is equal to the average
transmittance from normal incidence to about 70 degrees. Figure 4.37c shows the
absorptance of this beam splitter. A substantial amount of the incident energy is
lost in this beam splitter. If the beam splitter is turned around so that the incident
light passes through the substrate before encountering the metal layer, the absorp-
tance increases by 10 percent or more. In addition, the two beams will not have
equal intensities when used reversed. Thus, it is important to use these beam splitters
so that the light is incident on the metallic coating from the air side.
On occasion, the beam splitter coating is deposited on the hypotenuse of a
prism (see Fig. 4.32c). This geometry allows the light to strike the coating at any
arbitrary angle. When the coating is a metal, the reflectance and the transmittance
as a function of angle of incidence can have unexpected characteristics. Figure 4.38
shows the transmittance as a function of the angle of incidence. At the critical
angle, the transmitted intensity is reduced to zero, and no light can leave the prism
at angles greater than this angle. The intensity of the reflected beam is quite different,
as shown in Fig. 4.39. There is a range of angles between 50 and 60 degrees where
there is very high absorption of the p plane light. If the prism had no coating on
DESIGN TYPES 121

the hypotenuse, the reflectance for both planes of polarization would increase to
100 percent at the critical angle, and it would remain at that value for all larger
angles. Such components need to be used carefully.

4.5.1.2. Cube Beam Splitters. By their very nature, these beam splitters
are designed to function at a nominal angle of incidence. Many times, optical
systems require equal path lengths in both beams to minimize aberrations or other
differences between the two beams. This is especially true when the beams are
recombined in a second pass through the beam splitter. In those instances, a cube
beam splitter is the best solution.
The cube beam splitter is inherently rugged as the coating is not exposed to
the environment. The symmetry of the cube ensures that there can be no problem
with orientation of the incident light with respect to the beam splitter. The perfor-
mance of this type of beam splitter is very similar to that of a similar slab design.
Most of the comments that were made about the slab beam splitter also are
pertinent to the cube version. The same effects can be observed at high angles of
incidence as are seen with the slab, and the polarization effects are similar.

4.5.2. Dichroic and Trichroic Filters

Dichroic and trichroic filters separate wavelength bands, transmitting certain wave-
lengths while reflecting others. This can be accomplished with edge filters, but
since in color splitters they are frequently used at nonnormal angles of incidence,
they are considered separately here. Because they are used at angles where polar-
ization can be severe, a primary consideration should be the state of polarization
of the incident light beam. If the light is mostly polarized in one plane, then the
designing process can at times be simplified over what may be necessary if the
incident light's polarization state has to be assumed to be random or unpolarized.
If the incident light is polarized in the p plane and the filter has to operate
near Brewster's angle for the interface between the two coating materials, then a
severe design problem exists because, as mentioned in section 3.6, the effective
index ratio is close to unity in this regime. With no index ratio with which to work,
the thin film designer has a very difficult time in conjuring up a filter. During the
layout of an optical system, it is useful to keep in mind the difficulties polarization
may cause so as to avoid a situation in which the desired filter cannot be built.
The most extensive use of dichroic filters is in the separation of infrared (heat)
radiation from visible light so that the reflected beam is cold; this is the origin of
the term cold mirror. Many projector lamps have an integral reflector that has a
coating that is transparent to heat and reflects the visible light onto the film or slide.
The converse of the cold mirror coating is one that reflects the infrared wave-
lengths and transmits the visible light. Such a filter can be included between the
lamp and the film plane to reduce further the heat load on the film. Because these
mirrors reflect heat, they are sometimes called hot mirrors or heat mirrors.
Another application of these filters is in the reflector of a medical or dental
lamp where (1) the heat could be injurious to tissues and (2) patient comfort is
 100 r----------------------------------------------------1
R
Semltranaparent Metal Layer n = 0.05
k = 3.0
I = 18.8nm
n: 1.0

75
T

411

...
()
C

E 50
•.
C

I-
.
~

25

A
o
o 30 60 90

Angle of Incidence (Degrees)

100 r-------------------------------------------------------~
I"
//
/1
75 //
/1 /
/ I
• ",/ I
-.
() ",,/
C
Ave _ - - - /

-•
()

.!
It
lP-
50

-------
1---C5-;::::;::-==-.::=:-::::.:.------------- -
P - _ _ _ ./
/
/1

25

B
o o~--~----_P----~----~--~~--~----~----~--~
30 60 90

Angle of Incidence (Degrees)

Figure 4.3&. The transmittance (A), reflectance (8), and absorptance (e) of a low loss
silver slab beam splitter as a function of the angle of incidence in air. The reflectance and
transmittance are approximately equal at normal incidence. Note the expanded scale on the
absorption plot.

122

DESIGN TYPES 123

10r-------------------------------------------------,

CD
U

...
C
III
ICL
6

o
1/1
,g
c(
4
~

2 -------- p
t-------...~-=:::::-:::-~-::-:.:-:.-=------- __ ~¥e ~
s --.. . - .... , """
,,\
.....
,,~
c
O~--~----~----~----~----~--~----~----~-----
o 30 80 90
Angle of Incidence (Degrees)
Figure 4.36 (Continued)

important. An additional benefit of this type of arrangement is that the apparent

color temperature of the light source can be altered so that the reflected light matches
sunlight more closely, thus greatly reducing the problems associated with meta-
merism (see section 4.6.).

4.5.3. Filters for Use in Harmonic Generation Experiments

A crystal lattice can convert photons from a fundamental laser excitation frequency
into photons that are multiples of this frequency. This phenomenon occurs when a
sufficiently intense excitation energy density drives the electrical field strength into
a nonlinear portion of the dielectric function. Certain crystalline materials and proper
experimental conditions can efficiently generate second, third, and higher harmonics
of the fundamental wavelength. Interference filters are required in these harmonic
generators to separate the various harmonics in the output beam. Specifically, the
detectors used for some low energy content harmonics can be very sensitive to
other, high power, harmonics. Thus, the wavelength separation and blocking be-
come very important. To obtain signals for a wavelength without interfering signals
from other wavelengths, the rejection level of a filter is critical. Currently, the
major application of these filters is in diagnostic subsystems that follow an exper-
imental chamber.
In addition to the directly transmitted light, scattered light may be strong enough


100 r-------------------------------------------------~
R
Metal Layer n: 3.0
k:5.0
t :4.5nm

75
T
CD

..E..
U
C
III

.,--
50
•
..
-------
C p ---"" '\
III
I-
~ \\
~~~------------~~-----


-- ..........
"-
25
'\ \
\\
\
A ~

o
o 30 60 90

Angle of Incidence (Degrees)

100 r------------------------------------------------.

75

CD

..
U
C
III
u
III
50

;;::
CD
a:
~

25

B
o
o 30 60 90

Angle of Incidence (Degrees)

Figure 4.37. The transmittance (A), reflectance (8), and absorptance (C) of a metal (Cr)
beam splitter with equal reflectance and transmittance at normal incidence. The absorption
losses are much greater than that obtained with a silver coating.

124

DESIGN TYPES 125

100

75

II
U
C
...a.
ID

..
0
50
•
J:I
c(
'#

25

o
o 30 60 90

Angle of Incidence (Degrees)

Figure 4.37 (Continued)

to swamp a sensitive detector. Therefore, the coatings need to have very low scatter
in addition to all of the other requirements.
Typical diagnostic separation filters operate with an angle of incidence in the
range of 10 to 40 degrees. Consequently, polarization is a parameter that needs
careful attention. The incident light is often polarized as the conversion efficiency
of crystals is strongly dependent on the orientation of the polarization vector.
In third harmonic generation, diagnostic filters are typically used in pairs. This
makes the experimental setup more flexible and at the same time simplifies the
control of the phase retardance introduced by the mirrors. The phase retardance
must be controlled when ellipsometers and interferometers are a part of the diag-
nostic subsystem. The beam-directing mirrors and wavelength separating filters
should not modify or add a component to the polarization of the beam.
If a laser with a wavelength different than any of the harmonic wavelengths
is used to align the optical system, then there may be some requirements at that
wavelength also. The greatest challenge facing the fabricator of these filters is the
wide range of wavelengths of interest. Because thin film interference filters depend
on the matching of the optical thicknesses of the layers, dispersion in the refractive
index makes it difficult to match the thicknesses at several widely separated wave-
lengths. The lack of a good match reduces the rejection level available at wave-
lengths away from the one at which the optical thicknesses are matched.

126 CHAPTER 4

Angle of refraction in 1.0 medium (degrees)

80 eo
10 20 30 40 50 60 70 Y
100 ,---~--~----~--~~~~--------------------------~

Critical anele
75

.,u
....
C
III
l
E
III
50

..
C
III
I-
~

25

o .-----~----~--~~--~~--~----~----~----_P----~
o 30 60 90

Angle of incidence (Degrees, In n =1.52 Medium)

Figure 4.38. The transmittance of a prism beam splitter as a function of the angle of
incidence on the hypotenuse from within (lower scale) and from outside (upper scale) the
prism. The index of the prism is 1.52. There is no transmission beyond the critical angle.

Because very high-power densities are used in hannonic generation systems,

the damage threshold of these coatings must be high.
The thickness uniformity of the coating must be such that variations in the
optical thicknesses of the deposited layers do not introduce an unacceptable per-
formance degradation. This can be caused by an optical mismatch or a shift in pass
bands or rejections bands away from the center wavelengths as a function of position
over the surface of the filter.
A typical filter set used for third hannonic work may have a transmission of
3 percent at the fundamental frequency, 3 percent of the second hannonic, and 50
percent of the third hannonic.

4.6. OTHER FILTERS

4.6.1. Dark Mirrors, Selective Absorbers,
and Induced Absorbers
Dark mirrors and induced absorbers belong to a class of designs in which a primary
performance parameter is the absorption of light. The difference between the types
DESIGN TYPES 127

is principally a matter of semantics. In dark mirrors, the reflection is low in the

visible portion of the spectrum, but is high elsewhere. Because the substrate is
generally opaque in the visible, all of the incident energy in this spectral range is
absorbed, either in the opaque substrate or within the coating itself. Induced ab-
sorbers are a related type of design, but in these there is no high reflecting region.
Selective absorbers absorb only in selected spectral regions.
In some applications, there is a need for a coating that completely absorbs the
light at one or more wavelengths. These coatings all contain at least one absorbing
layer. Additional dielectric layers enhance the amount of light absorbed for a given
thickness of absorber. Generally, these coatings have metallic layers that may be
susceptible to environmental degradation. It is necessary to use care in designing
these coatings into a system because of the inherent potential for their degradation.
For example, these coatings should be used on surfaces that are not susceptible to
scratching, as such damage leads to light scatter. The dark background of the coating
makes the scratched area highly visible because of the very high contrast ratio.
The reflection from any substrate can be reduced to zero at a single wavelength

Angle of refraction in 1.0 medium (degrees)

80 0
10 20 30 40 50 6070
100

I
I
I
I
75
I, I
/
GI
I, /
..
()
c
III /
()
CD
;:
50 1\
" /
CD
a:
1\ /
?P- '\ /
I\ /
25 I /
/ \ /
s /
------ /

0
p

t\ ,_!:.-/
Crltlca' angl.
./

0 30 60 90

Angle of incidence (Degrees. in n =1.52 Medium)

Figure 4.39. The reflectance of the same prism beam splitter shown in Fig. 4.38. The
behavior of the reflectance is more complex than that of the transmitted beam. There is a
range of angles over which the p plane light is more strongly reflected than s light (the
opposite is more typically the rule). The difference between the reflectance of the two planes
of polarization is relatively small from normal incidence up to nearly the critical angle.
128 CHAPTER 4

with two thin film layers: one layer is an absorber, and the other, a dielectric. The
complex index of refraction of the absorbing layer can be any value. For a dielectric
substrate, the order of the two layers is straightforward: if the index of the dielectric
film is less than the geometric mean of the index of refraction of the incident
medium and that of the substrate, then the absorbing layer is located next to the
incident medium. Otherwise, the absorbing layer is placed next to the substrate.
The latter situation is obviously the more desirable as the metal layer is then protected
to a certain extent from environmental degradation. If the substrate is metallic, the
selection of the dielectric, and the derivation of the design, is more complicated.
These designs are akin to the exact solution one can obtain with the two-layer AR
V-coat. Figure 4.40 shows the reflections from two-layer dark mirrors, one using
a dielectric with an index greater than the square root of the substrate (the incident
medium has an index of unity), and one that is less; the same metal indices are
used for the metal layers in both cases.
Frequently, the need exists that a range of wavelengths be absorbed; this
requires the addition of more layers, usually in the form of added periods of
metal/dielectric pairs.
It is possible to combine designs to obtain certain special effects. For example,

10

A B
Air Air
8
t :97.4nm

II
6

-
U
C
III
U
II
;:
II
a:: 4
~

o+-____________ ~=a ____ aE~----~------------~

400 500 600 700

Wavelength (nm)
Figure 4.40. Reflectance as a function of wavelength for two-layer dark mirrors. When the
index of the dielectric layer is greater than the square root of the substrate, the metal is
placed next to the substrate, whereas the reverse is true when the index of the dielectric
layer is less than that of the substrate. The designs are given in the figure.
DESIGN TYPES 129

it is possible to combine a dark mirror with an edge filter; the reflectance is then
high where the edge filter reflects and very low where it transmits. The transition
between the reflecting and the absorbing regions can be very sharp, and can be
controlled quite precisely with the edge filter transition.
A good example of a use for an induced absorber is in optical data disk recording
media. In write once, read many times (nonerasable) disks, a pulse of heat generated
by a focused laser beam forms a hole or other perturbation in the surface of a
rotating disk. Because the light needs to be absorbed to generate heat, the recording
medium requires a metal or other absorber. Metals have the highest absorption
coefficient per unit thickness. This is an advantage in that only a small amount of
material needs to be heated, thus minimizing the input energy required to write a
data bit. Unfortunately, metal films tend to have a fairly high reflectance, and much
of the incident energy is reflected back toward the source, rather than being absorbed
to heat the recording material.
Bell and Spong (1978) described a recording medium that had a coating that
reflected only enough light to allow the focusing system to function. It consisted
of a metallic reflector, a spacer layer approximately one quarter wave thick, and a
thin absorber layer. The effect of this structure is to position the absorber near the
peak of the standing wave electric field so that the recording layer absorbs the
maximum amount of light possible. This is shown in Fig. 3.10. About 95 percent
of the absorbed energy is deposited in the recording layer, whereas the rest goes
into the reflector. This is a very efficient scheme for absorbing the energy and
depositing the heat in precisely the right location. In terms of the energy absorbed
per unit mass of absorber film, it is far superior to an earlier system in which a
single, relatively thick layer of the absorbing material was used as the recording
medium (Rancourt, 1981).

4.6.2. Solar Absorbers

Solar thermal systems collect the sun's heat and make it available for use in ther-
momechanical systems. The higher the temperature of the working fluid, the more
efficient and economical is the overall system. To absorb and keep as much of the
solar energy as possible, the surface that absorbs the energy must have some fairly
well defined characteristics. Specifically, the absorption spectrum of the surface
should be high for sunlight, whereas the emittance at the operating temperature
should be as low as possible. Because the solar spectrum and the blackbody emis-
sivity spectrum at the system operating temperature do not overlap significantly,
as shown in Fig. 4.41, a coating can theoretically have the desired optimal char-
acteristics as shown by the dotted line in the illustration.
The objective for solar absorbers is to absorb the solar radiation, while losing
as little energy as possible through unwanted conduction, convection, and radiation.
Heat removal is through the hot fluid outlet port. Usually, the absorbing surface is
insulated from convection and conduction losses by a vacuum jacket, therefore
radiation is the primary heat loss mechanism.
A first approximation to this coating is perhaps an oxide of indium and tin
(ITO) coating on black glass. A better, though more complicated solution, may be

130 CHAPTER 4

'O°r---~~----------I~F=-=-=--=-==~~~--~------~
I
I
I
I
I
I
80 I
I
Ideal Profile ----t

II
g eo
!!u
.!
';
II:
'" 40

20

O~~~~~;=~====~--------

0.3 0.4 0.5 0.6 0.8

1 4
____ ____
5 8
~

7 8 9 10
----~

20

Wavelength 'I'm)

FIgure 4.41. Curve for solar absorptance and reflectance of ideal solar collector (from Hahn
and Seraph in, 1978).

a thick layer of silicon over a silver or other infrared reflective material. An AR

coating may be necessary to reduce the first surface losses at the silicon/vacuum
interface. A number of papers have been published on this subject in the solar
energy literature. So far, the major stumbling block has been the economical man-
ufacturing of durable, long-lasting coatings so that the overall system costs can
compete with the other sources.

4.6.3. Notch Filters

Notch filters have a narrow rejection band within a much wider high transmission
band. Such filters are not easily designed since the width of the rejection region is
defined by the index ratio of the materials that make up the coating. Narrow rejection
zones imply a low index ratio, and this in turn implies a large number of layers to
obtain a significant rejection level. In addition, a simple quarter wave stack has
side lobes in the transmission region near the main rejection region. Several papers
(Baumeister, 1981; Thelen, 1971; Young, 1967) show how the side lobes can be
suppressed by varying the thickness of the periods in the stack.

4.6.4. Transparent Conductive Coatings

These coatings are used for a variety of purposes. The more familiar ones are
(1) heating glass surfaces for defogging and defrosting, (2) conducting electrical
signals for transparent displays such liquid crystal display (LCD) units, (3) providing
electrical shielding for electromagnetic or radio frequency interference (EMIIRFI)
suppression, and (4) serving as an infrared reflector and a visible transmitter in
oven windows.
DESIGN TYPES 131

Because these applications usually call for the coating to be on a transparent

substrate that is exposed to all manner of environments, they have to be very durable.
In LCD devices, the coating must have certain etching characteristics so that the
user can fabricate patterns in the coating to operate the device.
A number of materials are used to provide both transparency and conductivity.
Gold and silver are good examples that have been used for a number of years.
From a durability standpoint, gold is the preferred choice between these two metals
as it is less chemically reactive. Both metals, however, are very soft and are easily
scratched unless protected. Other metal and doped semiconductor films can be used,
but their performances do not approach that of the best silver films.
In the past few years, a mixed oxide of indium and tin has been used. The
exact composition depends on the manufacturing process, and it is usually tailored
to the coating's ultimate use. The conduction mechanism is of a semiconductor
nature. The material's properties appear to be consistent with what one would expect
of an indium oxide matrix doped with tin. Many conducting films have patterns
etched in them, and the etch time and other etching characteristics of the film are
of vital interest to the user of conducting films.
The requirements placed on conductive coatings include the usual ones placed
on optical coatings, plus some that are unique to the conductive aspect. Conductivity
is rarely specified explicitly; rather, the sheet resistivity is specified, usually in
terms of ohms per square, since with this form, it is not necessary to specify the
thickness of the layer. The relation between resistivity and sheet resistance is
p
, =-
S t

where 's is the sheet resistance in units of ohms per square,

p is the resistivity of the coating in ohm cm,
and t is the thickness of the coating.
The resistances of transparent conducting films may have a significant tem-
perature coefficients if they consist of semiconductors, such as ITO.
Hall effect measurements of a conducting film can provide useful diagnostic
information on its physical properties by providing data on the number density,
sign, and mobility of the charge carriers in the film.
The optical transmission of the film constrains the maximum thickness of the
film. Thus, the film specifications are given in terms of the functional parameters
of sheet resistance and transmission. This approach to specifying conductive films
take into account the fact that the physical properties of metal films can vary. They
depend on the coating process, especially in the transparent conductive material
most commonly used, indium-tin oxide. The coated film has a composition that
varies, and its conductivity (or resistivity) is determined by the volume density of
the charge carriers in the film, as well as by their mobility. These parameters can
vary significantly, and depend on the composition, the level of oxidation, and the
deposition process parameters.
The transmission and reflection of a thick conductive coating of indium-tin
oxide is shown in Fig. 4.42. The doping level has an especially strong effect on
132 CHAPTER 4

100

80

II
U
C 80

.
II
E
....
0
II 40
A-
11

20 Single Surface
Reflection

0
400 800 1200 1800 2000 2400

W8velength (nm)

Figure 4.42. Measured reflectance and transmission of an ITO film on a glass substrate
in the visible and the near IR. The physical thickness of the film is approximately 325 nm.

the slope in the wavelength region where the material switches from transmitting
to reflecting, i.e., the 1- to 2-~m range.
The shape of the curve in Fig. 4.42 approaches the ideal one for solar thennal
uses. The conductive coating is used as a reflector to return the infrared radiation
back to a receiver, whereas the sunlight passes through the coating to heat the
receiver. Unfortunately, the position of the edge of this coating material is not quite
correct for the kinds of efficiency needed to make this design viable in solar thennal
energy applications. In addition, the infrared reflectance is not as high as might be
desired.
ITO has a high index of refraction when compared to glass. As a consequence,
the reflectance of a coated area appears quite reflective to an observer, as compared
with the bare glass where the ITO was removed. This is frequently objectionable,
and additional thin film layers are used to reduce the contrast in reflected light
between the two areas. These layers afford additional environmental protection to
the conducting films. Fortunately, the electrical conductivity through these overcoats
can be finite, and devices can function with the overcoat over the conducting films.
These dielectric coatings significantly affect the etching rate.

4.6.5. Patterned Dielectric Coatings

On occasion, it is necessary to produce a specified pattern in the thin film optical
coating. One example of such a coating's use is to code the signal from a single
gun vidicon television tube and produce a standard color signal output. A pattern
that accomplishes this consists of fine bands (on the order of lOO/cm) of a color
separation (dichroic) coating separated by clear gaps. Superimposed on this pattern,
but rotated by about 30 degrees, is a second similar pattern. The complete pattern
DESIGN TYPES 133

is placed on a vidicon faceplate with a diameter of about 2 cm. Patterned photoresist

masking was used to delineate the areas that were not to be coated. The coating is
deposited at an elevated temperature; after the coating process is complete, the
resist is stripped off, leaving the desired pattern. The resulting coatings can have
high levels of durability, and photoresists are readily available that are compatible
with the thin film coating process. Other patterns can be prepared in this manner.
Thus, very fine patterns can be produced, and the performance of the resulting
coatings can be made equivalent to those deposited on larger areas (Bartolomei,
1976).
4.6.6. Coatings for High-Power Laser Beam Reflectors
In these coatings, the absorption coefficient must be as low as possible, and the
laser damage threshold as high as possible. This type of design is a subset of the
previously mentioned high reflector types, but with some specifications peculiar to
their use in high-power laser beams. In the visible and near infrared portions of
the spectrum, silicon dioxide is currently the favored low index material because
of its low absorption properties, whereas titanium dioxide is used as the high index
material. The titanium dioxide does not have as Iowan absorption coefficient as
the silicon dioxide, but it has the highest index of refraction in these spectral regions.
The resulting high index ratio means that less material is required because a high
reflectance level can be attained with fewer layer pairs. In addition, the actual
physical thickness of a high index layer is less with the higher index for a given
optical thickness, thus reducing the total amount of material in the coating. The
materials used in other portions of the spectrum are selected on a case by case
basis.
Three types of high reflector coatings are available to reflect high-power laser
beams: (1) protected metals, (2) enhanced metals, and (3) all dielectric.
Protected metal coatings are simple and provide high reflectance over a broad
band. Their durability, although better than unprotected metal layers, is not as good
as that obtained with the other types of reflectors. They are good for continuous
wave laser beams, but do not hold up well under high-power pulse irradiation.
Because of the large bandwidth over which the reflectance is high, the choice of
wavelength with which to align the optical system is also large.
Enhanced metal designs are more durable than protected metal coatings, but
the wavelength region over which the reflectance is high is more limited. This
approach is good for continuous wave lasers also, but fast (picosecond) single pulse
lasers can cause mirror damage. These designs typically have between 2 and 5
periods of reflectance enhancing dielectric layers.
All dielectric reflectors rely entirely on the dielectric layers to provide the
reflecting power of the mirror. These designs have very high reflectance, but the
bandwidth is limited. Between 10 and 15 pairs of layers are used to attain the high
reflectance. The coatings are typically very durable and perform well for both
continuous wave and pulsed situations. Because the bandwidth of these mirrors is
limited, the use of alignment lasers may be problematic if the alignment wavelength
is far removed from the design wavelength. These coatings can also be used on
134 CHAPTER 4

metal or other nontransparent substrates. The penetration of the electric field is

limited to the outer layers of the design so that the substrate does not contribute to
the reflectance (or absorption).

4.6.7. Color Correction Coatings

Color correction reflectors have a high reflectance that is a function of the wave-
length. The functional form of this curve is selected such that when the light of
the incident beam is reflected by this mirror, the spectrum of the light in the reflected
beam is altered to coincide with some predefined shape. For example, the visible
spectrum of the light from an incandescent source can be converted, after reflection
from such a mirror, to that of the solar spectrum.
A commercial use for this type of mirror is in color correction coatings on
medical mirrors. The filament in the light source has a color temperature in the
2600 to 3100K range, and the color correction mirrors can change the apparent
temperature to that of sunlight, which is approximately 6OOOK. The basic design
of these filters is that of a cold mirror (see section 4.3.). These mirrors eliminate
much of the near infrared energy emitted by the filament by virtue of the fact that
they are highly transparent to wavelengths in this spectral range. Thus, these mirrors
improve patient comfort, reduce biological tissue heat damage, and assure good
color matching.
Good color correction eliminates metamerism, the apparent yet false matching
of two different color samples under some lighting conditions. Dentists and other
medical personnel that make color matches in an artificially lighted environment
need this kind of color correction in their lighting systems. The use of a color-
correction mirror is necessary whenever color matching is performed under artificial
lighting conditions.
Similar color or spectral correction filters can function in a transmission mode
as well. One use is in spectral trimming filters for silicon detectors. The response
of the detector can be made to approximate the photopic eye response for use in
photometers. Other uses include trimmers to shape the spectral response on exposure
meters.

4.6.8. Switching Filters

We have mentioned in Chap. 1 how the term optical constant is a misnomer, as
the values change significantly as a function of wavelength. In addition, the con-
stants of some materials vary with temperature (at very high temperature, e.g.,
near the melting point, the indices of most materials change significantly). One
such material, vanadium dioxide, has optical constants that change dramatically
when the temperature exceeds a transition value. The transition is one that involves
a change in the crystal structure, and the optical performance changes from a
transmitting state to a reflective state. This effect can be used as a passive switch.
The incident light heats the filter because there is some slight absorption in the
transmitting state. As long as the power level remains below a critical level, nothing
happens and the filter transmits most of the incident light. When the incoming light
reaches a power density that heats the filter above a critical level, the filter switches
DESIGN TYPES 135

to a reflective state and the transmission of light through it is blocked. Such a switch
can be used to prevent damage to the optics behind the filter. When the light level
decreases and the filter cools, the filter reverts to its original state and transmits the
light once again. Such a scheme is at times superior to a shutter because there are
no moving parts involved.
Additional thin films can be used with the switching material to increase its
sensitivity, in the same way that the dark mirror structure is used to enhance the
sensitivity of the optical data disk recording mechanism (see section 4.6.1.).

4.6.9. Emissivity Control

Coatings are available that can enhance or decrease the emissivity of surfaces. These
coatings are useful especially when radiation is the primary mechanism of heat loss
from a surface, and where conductive and convective heat transfer mechanisms
cannot function effecti vely. Such conditions are found where a surface is surrounded
by a vacuum, as may be found in space or in an evacuated container.

4.6.9.1. Emissivity Reducing Coatings. Emissivity reducing coatings have

been used for many years in Dewars and other vacuum-insulated containers that
are used to contain hot or cold substances; in these situations, the aim is to reduce
the amount of heat transferred between two surfaces.
There are two major types of emissivity-reducing coatings. Opaque layers of
high infrared reflectivity metals, such as gold, silver, copper, and aluminum, are
used when visible transmission through a surface is unimportant. A transparent
conductive coating, such as ITO, is used when visibility through the surface is
required. In a vacuum environment, silver can be used as it will not tarnish. Silver
has high reflectance in the visible as well as in the infrared, and is, therefore, the
preferred material, though gold could also be used. The emittance of the surface
facing the vacuum is the one that is to be controlled. The substrate in these instances
is likely to be glass or a metal. These materials are opaque at the wavelengths
where the blackbody curve peaks for most common temperatures at which tem-
peratures and heat gains and losses need to be controlled. Thus, the coating should
be applied on the vacuum side of the surface. Figure 4.43 shows the blackbody
distribution for a number of temperatures. There are a number of slide rules available
that allow the rapid calculation of various blackbody parameters.
A conductive coating, such as ITO, generally has a fairly high infrared re-
flectance as is shown in Fig. 4.42. As such, it can make a moderately good coating
material to reduce the heat transfer between two surfaces, and yet retain visible
transparency. Semitransparent metal films can also be used. In these metals, how-
ever, the infrared performance is improved over ITO, but at the price of a reduced
visible transmission.

4.6.9.2. Emissivity Enhancing Coatings. An emissivity enhancing coating is used

when it is desired that a surface emit more heat than it would otherwise, i.e., to
couple more closely the surfaces thermally.
Spacecraft have surfaces designed to dissipate heat generated inside the craft
136 CHAPTER 4

100~--~--~--~~--~~----~--~~~--~~--~----~
0.2 0.3 0.4 0.6 0.8 1 2 3 4 8 8 10 20

Wavelength ()1m)

Figure 4.43. Spectral radiant exitance of a blackbody radiator. Curves are for temperatures
of 300K (near room temperature), 600K, 1000K, 3000K (incandescent lamp filament tem-
perature), and 6000K (approximate solar surface temperature).

by the electronics and other internal packages. Photovoltaic solar arrays are tem-
perature sensitive and need to be kept cool to keep their solar to electrical efficiency
as high as possible. For many years, thin fused silica (SiO z) or glass sheets have
been used to increase the emissivity of the silicon cells. The relatively high emis-
sivity of fused silica, as well as the fact that it does not darken under particle
irradiation, was responsible for its selection early in the space program for use in
solar cell covers.
A thennal control surface for temperature control of the body of a spacecraft
consists of a high reflector to reflect the solar radiation placed on the second (inner)
surface of a fused silica sheet. A dielectric reflectance-enhancing stack over a silver
layer is the choice reflector material because it keeps the solar absorptivity to a
minimum. The uncoated (outer) surface of the fused silica serves as the emitting
surface. This arrangement provides a surface from which the spacecraft thennal
energy may be radiated even when it is in direct sunlight.
The hemispherical emissivity of silica is approximately 0.82 at a temperature
of about 300K (Crabb, 1972; Jet Propulsion Laboratory, 1976). Note that it is
necessary to specify a temperature when specifying an emittance value, since the
nonnal emittance is defined as


DESIGN TYPES 137

fA(}") B(}..,T) d}"

e(T) - =---'-'--..:..-..:...-:..-
- fB(}..,T) d}..

where A(}") is the spectral absorptance and

B(}..,T) is the blackbody function for temperature T.
The integral has to be taken over the entire range of wavelengths, though from
a practical standpoint, the blackbody curve will put reasonable limits on the range.
A second integral, one over all angles needs to be calculated to obtain the hemi-
spherical emittance. An angular weighting function is also required to take into
account the directional nature of the emitting surface. In many thermal radiators,
the substrate may be opaque in the infrared region of interest. Thus, in these
instances, the absorptance is equal to 1 - R, where R is the reflectance.
The emissivity of a surface may be increased by the application of coatings
that reduce its reflectance in the range in which the blackbody curve has large
values. For a surface at 300K, blackbody curve reaches a peak between 9 and 10
f.Lm. The reststrablen reflectance of silica also has peaks in this same spectral region,
thus reducing the emittance of this material. These peaks can be reduced with
appropriate coatings that form a dark mirror, which antireflects the reststrahlen

100,---------------------____________________________ ~

Angle of incidence

75

~
GI /\ 1/
U
../ \ 1/
..
C
III
u 50
I
60 0
\ V /1:I:
Ii

/ ," I,
QI
;; / " I:
GI
a: I \ .. I:
-:
?F. 145° :
f I •••.•

f I: :!~~
I ,' :
25
:
I :
I :
J I :
:i
J ,,..':
Oi-~'-'~·~~~------_r------~--------~------~------~
7 g 11 13

Wavelength (pm)
Figure 4.44. Reflectance of bare silica in the reststrahlen reflection region as a function
of wavelength. The performance has been plotted for a number of angles of incidence (shown
labeled in degrees). The reflectance increases monotonically with the angle of incidence at
all wavelengths.

138 CHAPTER 4

1001---------------------------------------------------~

75

Angle of incidence
G)
U

..
C
III
u 50
CII
;;:::
G)
a:
lP.

25 I\eo
1/\\
J/ \ \ ~....... / __
.'./"l' :f:.: \ .._ .................. _ _ _ _ _
\~5 ..........
\.:: ••••••••~::_.::::~ ••C!•••••••••;:,.....--------------::.-
o~~~. .~------~~------~------~--------~------~
7 9 11 13

Wavelength <JIm)

Figure 4.45. Reflectance of silica coated with an emissivity enhancing coating. The per-
formance has been plotted for a number of angles of incidence (shown labeled in degrees).
Most of the curves of the coated substrate lie well below the corresponding curve for the
bare substrate (see Fig. 4.44).

reflection. In this way, the emissivity is increased by a reported 5 to 10 percent

(Rancourt et al., 1984). Calculations show that the reststrahlen peak of the silicon
dioxide widens considerably as the angle on incidence increases. The increase is
considerably diminished with the coating, thus improving the high angle perfor-
mance of these emitters as well. Figures 4.44 and 4.45 show the calculated reflec-
tance of bare and coated silica for a number of angles of incidence.
 Chapter 5

Filter Performance
Measurements

This chapter deals with the functional characteristics, or perfonnance, of typical

thin film filters, and how one measures this perfonnance. The measurement tech-
niques must be up to the task that is required of them. For example, it is not
acceptable to measure the transmittance of a narrow bandpass filter with a spectro-
photometer whose resolution is less than that of the filter. This and other such
problems are covered in this chapter. For further infonnation, the reader is directed
to the excellent book on photometry by Walsh (1958).

5.1. SPECTROPHOTOMETERS AND SPECTROPHOTOMETRY

Spectrophotometers are optical instruments that measure the intensity of light trans-
mitted or reflected by objects as a function of wavelength. They are available in a
variety of types and styles. Often, they are optimized for chemical studies, where
the photometric accuracy, i.e., the precision with which the intensity is measured,
is not necessarily critical. In many cases in optical filter work, however, there is
as much interest in the photometric accuracy of a measuring instrument as there is
in the wavelength accuracy.
The wavelength display of some spectrophotometers is calibrated in a scale
the units of which are actually the reciprocal of wavelength. A common unit for

139
140 CHAPTER 5

reciprocal wavelength is the wavenumber, or reciprocal centimeter, i.e., the number

of wavelengths that occur in the distance of 1 cm.
The wavelength accuracy of the dispersing mechanism can be verified (and
calibrated, if necessary) by using the lines of atomic emission lamps, lasers, stable
absorption filters such as a styrene film for infrared calibration, or other known
standards. The photometric accuracy of an instrument cannot be verified quite so
easily. The photometric linearity can be tested by using known neutral density filters
in the optical path.
The accuracy requirements of the wavelength or wavenumber axis is a well-
understood concept. Photometric accuracy, though often as important as wavelength
accuracy, is frequently not tested and is accepted as correct. A few simple tests
can be done to verify the linearity of the photometric axis, and a few operational
habits ensure proper scaling.
The linearity of an instrument can be verified by placing known filters in the
sample beam and noting if the output of the instrument is proportional to the
transmission of the filter combinations. It is best to use filters whose spectral
transmission curves are relatively flat as a function of wavelength to avoid possible
complications from spectral shifts. By measuring the transmission of a number of
individual filters and then measuring them in various combinations, the linearity of
the instrument can be determined from a plot of the calculated versus the measured
transmittances.
Once the instrument is known to be linear, it is necessary to know the two
endpoints of the range to interpolate accurately. This is accomplished by ensuring
that each measured point lies between two other known points: (1) there should be
a lower value, either zero or some known lower transmission or reflection value
and (2) there should be an upper value that is larger than the one being measured.
For a measurement of transmission, the upper value might be made with no
sample in the beam, whereas the lower value might be made with a well-blocked
sample beam. Specific reflectance measurement standards depend on whether the
reflectance is very low, as in the case of an antireflection coating; medium, as in
a beam splitter; or, high, as in an enhanced metal reflector.
The zero baseline can be obtained with a blocked sample beam. The upper
value needs to be selected so that there is as much expansion of the output scale
as possible.
For a low reflector, the flat face of a wedge of uncoated silica or a glass with
a known index of refraction can be used to obtain a full scale reading in the
neighborhood of 4 to 7 percent. A wedged standard is necessary to eliminate possible
second surface effects. A sample reflectance of 0.5 percent will then produce a
reading of one-eighth full scale with a zero baseline and a 4 percent reflectance
standard. Similarly, a flat piece of silicon or germanium can be used to yield
reflectances in the 30 percent range in the visible portion of the spectrum. Rhodium
and other durable metal reflectors can provide 50 to 70 percent full-scale expansions.
The measurement of very highly reflecting surfaces can be problematic, es-
pecially if the reflectance exceeds that of a calibrated reference, such as freshly
evaporated silver, the metal with the highest reflectance over a large part of the
FILTER PERFORMANCE MEASUREMENTS 141

spectral ranges of interest in optics. In this case, it may be necessary to use ex-
trapolation or rely on the linearity of the gain outside of a range that can be checked.
On the other hand, as the extrapolation is over a small range if a known standard
with a high reflectance is used, the error introduced in the measurement will likely
be small.
The normalization of the performance value to 100 percent can present a
problem because it depends on a number of parameters that cannot be precisely
calibrated in a way similar to those used for wavelengths. In all spectrophotometers,
the output is the result of a ratio: the intensity of the light through the instrument
with the sample in place is divided by that of a calibration measurement. There are
two ways in which this is done by spectrophotometers in current use, and the
instruments can be classified accordingly: (1) single beam and (2) double beam.

5.1.1.1. Single-Beam Spectrophotometers. In the single-beam instrument,

there is only a single optical beam through the instrument, as the name implies.
The baseline scan is recorded for later division into the sample scan. The instrument
must be stable enough so that its response or sensitivity does not change between
the baseline scan and the sample scan. The components that must be stable include
the lamp intensity, the detector response. and all other parameters associated with
the instrument. The time frame during which these parameters must be constant
can range from seconds to hours, depending on the time required to complete a
scan, and the time between scans.

5.1.1.2. Double-Beam Spectrophotometers. Many instruments are of the

dual-beam type, in which the light from the source is alternately directed into a
reference beam and into a sample beam by a chopping wheel or other device. Each
beam travels a different path. Theoretically, the two paths are identical in length,
number of reflections, and so on, except for the sample's being in the sample beam.
If such is the case, the ratio of the intensities of the two beams are proportional to
the actual performance (reflectance or transmittance). The use of alternating beams
reduces the errors that may be introduced by a fluctuation in the intensity of the
light source, absorption in the light path, such as water absorption or mirror atten-
uation, and other similar common mode (common light path) effects as the two
intensities are sampled a fraction of a second apart. Differences that are not common
to the two paths and yet are not related to the filter's performance are the source
of the measurement inaccuracies. An example of this is unequal conditions in the
two optical paths, such as inadequate purging of the beam paths. If the path of one
of the beams has more water vapor in it than the other, one can expect to see a
difference in intensity between the two beams at wavelengths near the highly
absorbing water band near 2.8 j.Lm. Thus, spectrophotometers are frequently purged
with dry nitrogen to ensure that the two paths are equal after a certain purge
stabilization time. Some of the spectrophotometer measurement considerations noted
here also apply to single-beam instruments.
There are other problems that are common to both types of instrument. These
include: (1) substrate scatter that reduces the apparent transmission; (2) beam offset
142 CHAPTER 5

introduced by the substrate (the beam can be vignetted by apertures before it reaches
the detector); and (3) in reflectance measurements where a reference reflector is
used, the uncertainty in knowing its precise reflectance value.
The typical commercially available spectrophotometers were designed prin-
cipally to measure transmission through a relatively small sample. To measure
reflection, attachments have to be added by the user. These attachments lengthen
the path in the two beams, so it is important that the original optical path be
sufficiently flexible that this increase in length does not affect the basic performance
of the instrument.
When the transmission of a sample is measured, it should be tilted a few
degrees so that the optical axis of the instrument is not perpendicular to the surface
of the sample. Otherwise, some of the light that reflects off the sample can travel
back toward the source and upset the calibration of the instrument. This happens
in some instruments when the energy reflected off the sample surface travels back-
ward through the optical system and then enters into the reference beam system.
The dispersing element that is used in a given spectrophotometer is not im-
portant, as long as the overall instrument is adequate. Thus, prism, ruled or hol-
ographic gratings, and Fabry-Perot interferometer units are all acceptable from a
fundamental viewpoint for general work. On the other hand, practical considerations
such as stray light, scan time, ease of operation, or signal to noise ratio may sway
the balance in favor of one type over the others.

5.1.1. Reflectance Measurements

The accurate measurement of reflectance poses a real challenge. Reflectance mea-
surements are difficult to do at exactly normal incidence without a beam splitter or
complex polarization schemes to divert the reflected beam to the detector and to
prevent it from returning to the light source. The introduction of these additional
intensity reducing optical elements into the system can reduce the sensitivity of the
instrument, especially in low signal situations, such as in the measurement of
antireflection coated surfaces. In addition, these auxiliary elements must be scatter-
free to minimize their introducing noise into the measurements.
Because the true normal incidence geometry is so difficult, most measurements
of reflectance are made at some small angle, such as 10 degrees. Usually, this does
not introduce significant errors. At high angles of incidence, the polarization in-
troduced in the optical path by the instrument, in addition to that arising from the
sample, is significant and can distort the results, as discussed in section 5.1.5.
In a dual-beam instrument, to keep the path in the reference beam the same
as in the sample beam, a standard reflector occupies a position in the reference
beam that is identical to that of the sample in the measurement beam. The choice
of reference reflector is arbitrary, though it is desirable that there be no steep
variations in reflectance with wavelength, and that the reflected intensities in the
reference and the sample beams be approximately equal. When a low reflectance
is being measured, a low reflector should be placed in the reference beam to keep
the intensities in the two beams approximately balanced. This reduces the possibility
of stray light from an intense reference beam from interfering with the sample beam.
FILTER PERFORMANCE MEASUREMENTS 143

It also makes the taking of the ratio of the two intensities more accurate as the
divider is working at close to unity ratios. The reflectance of the standard in the
reference beam should be at least as high as that of the sample.
A problem in the measurement of reflectance is that a reference, the reflectance
of which is well characterized, is usually needed. In either a single-beam or a
double-beam instrument, a baseline scan is measured prior to making an actual
measurement. In transmission, this is easily done by leaving the sample out of the
measurement chamber. In measuring the reflectance, a mirror must occupy the
sample position during the baseline scan. This reference mirror is then removed
and the sample inserted in its place for the actual sample measurement. The ratio
of the sample results to the baseline results gives the ratio of the two reflectances
without any contributing factors from the instrument. The result of the ratio is
multiplied by the reflectance of the ref~rence mirror to get the sample reflectance.
Any wavelength variation in the reflectance of the reference mirror must be taken
into account. The problem is thus reduced to that of knowing well the reflectance
of the reference mirror.
In the case of low reflectance values, the reflectance of polished samples can
be used. A polished glass surface can be characterized on the basis of its refractive
index. In addition, such a surface is hard and will not easily scratch. Other reference
reflectors that can be used in the visible and the infrared are the polished surfaces
of silicon and germanium. High reflectance references are more difficult. Typically,
high reflectors can be made from freshly evaporated aluminum and silver for the
visible, and silver and gold for the infrared. These reflectors are all relatively soft
and can be damaged easily. They should be checked frequently. Rhodium metal
films are fairly reflective and quite durable. In any case, the reference mirrors should
be approved by appropriate quality assurance personnel.
Professor John Strong (1938) originated a technique to overcome the previously
mentioned problem of needing a reference for reflectance measurements. His optical
layout is shown in Fig. 5.1. It is sometimes referred to as a V-W configuration as
the optical paths in the two positions look like these two letters. It should be noted
that this arrangement, because it has two reflections off the sample surface, measures
the square of the reflectance, and therefore, it is quite useful for mirrors and beam
splitters with high reflectance. When the reflectance is less than about 30 percent,
the squared value drops below 10 percent, and accuracy can be quickly lost at these
lower values. The fact that one mirror has to be moved from the sample to the
reference positions can pose a problem in the alignment of the optical path. Some
instruments use separate fixed mirrors to redirect the beam, thus avoiding the
alignment problems; the assumption is made that the two mirrors are identical.
Another drawback that is sometimes encountered in this configuration is that the
sample must be large enough to intercept and reflect the two beams that need to
impinge on the sample. In addition, the larger the sample, the greater the need to
ensure that the uniformity of the coating is adequate to ensure good results.
When the reflectance of a sample is low, the problem encountered is frequently
one of low signal levels, combined with possible scattered light from the surface
being measured, or from other sources such as stray light in the monochromator.
Accurate reflectance measurements of samples with very high reflectance levels

144 CHAPTER 5

A.

B.

/,
'/
\ /
\ I
\ I
\ I
""'"7"'";-rr-:-
11111 1
~~~nr,1
11,1",',
1111111111111/11

Original position of M2
Figure 5.1. Diagram of Prof. John Strong's v-w reflectance attachment for the measure-
ment of the reflectance of a sample without having to rely on a standard. Note that the beam
reflects twice off the substrate. This yields R2 as a result of dividing the intensity value from
arrangement B by that from arrangement A. Note that this arrangement may be inaccurate
if the reflectance of the sample is low (R2 is too small).
FILTER PERFORMANCE MEASUREMENTS 145

require the determination of a small difference in intensity between the sample and
the reference reflected intensities, which are large. Sometimes this is a small fraction
of a percent difference between the mirror and 100 percent. In addition, the re-
flectance of the reference needs to be known precisely. For values of reflectance
near unity, it is possible to design a measurement system such that the light is
reflected many times off the sample surface. To first order, this technique increases
the sensitivity of the measurement linearly with the number of reflections the beam
makes on the sample, as can be seen by expanding the difference between unity
and the reflectance losses:
Rn = (I-At = 1 - nA + ...
where R is the value of the high reflectance, A represents the reflectance losses,
and n is the number of reflections the sample beam makes on the sample
surface.
The losses are assumed to be small (the reflectance is high) so that higher
order terms can be neglected. By increasing the number of reflections off the test
surface, we can increase the sensitivity. In such a situation, it is necessary to be
certain that the auxiliary optics that are required to get the multiple bounces from
the sample do not introduce undesirable effects of their own.

5.1.2. Beam Offset

The principle of most double-beam spectrophotometers is that both the reference
and the sample beams traverse equivalent paths at all times during a measurement;
i.e., the path lengths and the mirror reflectances are the same, and the beams
impinge on the detector at the same point. A thick glass substrate or window inserted
into one of the beams offsets it if it is not inserted perpendicular to the optical axis.
The effect of offset can be subtle and difficult to detect. For example, the offset
may cause vignetting of the beam, or cause the beam to strike a different part of
the detector. Detectors can have variable sensitivity over their active surface, both
from the standpoint of wavelength as well as position (Birth and DeWitt, 1971;
Ballik, 1971; Sommer, 1973). In other words, the spectral sensitivity curve shape
varies as a function of position. For precision work, these factors merit careful
consideration.
A number of methods have been devised to reduce the effect of areal non-
uniformities in detectors, including the use of integrating spheres. These approaches
incur a degradation of sensitivity as the walls of the spheres always absorb a certain
amount of light. Another method makes use of a scattering lens that diffuses as
well as focuses the light onto a detector (Anthon, 1982).

5.1.3. Extra Surface Effects

The presence of extra surfaces in the optical path is another cause of measurement
difficulties. For example, in the measurement of the reflectance of one surface of
a thin glass plate that has been coated with an AR coating, the second surface (not
AR-coated) may reflect about 4 percent of the light incident on it. The light reflected
from the uncoated surface dominates that reflected from the treated surface. The
146 CHAPTER 5

exact value of the spurious component is difficult to detennine because the light
reflected from the second surface travels twice the thickness of the substrate, and
will be offset from the nominal beam as well.
There are two frequently used methods to eliminate bothersome second surface
reflectances. A wedge-shaped prism whose index is the same as the substrate can
be placed behind the uncoated surface and an index-matching fluid introduced
between the substrate and a wedge face. The result is to eliminate the optical
interface. The transmitted beam is reflected by a second prism face. The deviation
introduced in the beam, along with the extra path length in the prism, is usually
sufficient to deflect the beam out of the optical train and the collection cone of the
detector. The wedge must not be a right angle prism, as the deviation will be zero
for this configuration and the beam will be returned to the spectrophotometer.
Another technique to reduce the second surface reflection is the use of a
substrate whose second surface has been roughened, either chemically or mechan-
ically. The treatment should reduce the specular reflectance of the surface to a
negligible value.
The thicker the substrate, the less of a problem the second surface reflectance
is as the beam reflected by the second surface will be displaced in proportion to
the thickness of the substrate. The larger the deviation, the less the likelihood of
this offending beam's finding its way to the detector. If the second surface is
nonplanar, then the second surface beam will be defocused before it can get to the
detector, thus further reducing the magnitude of the problem.
When an integrating sphere or other wide acceptance angle detector collects
the light that is reflected by the sample, reduction of the second surface effect can
only be done by absorbing the unwanted light. A paint that has an index equal to
the substrate index and that is completely absorbing at the wavelengths of interest
works nicely. Another similar method of addressing this problem is to attach an
absorbing material, e.g., a piece of black glass, to the back of the substrate with
an index matching oil.
When infrared substrates are used, the second surface reflection may be higher
than that of optical glasses used in the visible. Index matching fluids are not generally
available in this spectral region, therefore, second surfaces have to be treated to
eliminate the specular beam. This has to be done carefully for the following reasons.
1. The substrates are frequently fragile. In the infrared, semiconductor crystals
are frequently used as substrates since they may have very good transparency. In
addition, these substrates are frequently thin, and this contributes to the fragility.
2. Scattering is a strong function of wavelength. The scattered light intensity
can decrease as fast as the fourth power of the wavelength, depending on the type
of scattering mechanism involved. As the wavelength of interest increases, it be-
comes more difficult to attenuate the second surface reflection by roughening it.
Some special purpose instruments have been designed to distinguish between the
first and the second surface reflectances.
In the previous discussion, it is assumed that the light reflected from the second
surface is adding in an incoherent manner to the light from the first surface. This
FILTER PERFORMANCE MEASUREMENTS 147

may not be the case in some infrared, as well as visible, measurements. The two
surfaces of the substrate can serve as the mirrors in a Fabry-Perot cavity with the
result's being a so-called channel spectrum. This appears on a spectrophotometric
scan as an oscillation in the reflectance. On a wavenumber scale, the oscillation
period is constant and related to the thickness of the substrate. The amplitude of
the oscillation in the reflectance depends on the reflectance of the two surfaces: the
closer the reflectances are to being equal, the greater the amplitude of the oscil-
lations. If the j7number (j7#) of the spectrophotometer is similar to that of the
optical system in which the filter will be used, then the oscillations will also be
present in the final system (assuming the filter is not optically bonded to some other
component). If the second surface of the filter has an antireflection treatment, there
should be minimal interference between the beam from the second surface and the
front surface one.

5.1.4. Effect of the Refractive Index of the Incident Medium

Filters that are designed for use in an incident medium other than air may have a
different performance in air, and vice versa. Examples of this are antireflection
coatings applied to silicon photovoltaic cells. One- and two-layer antireflection
coatings are used for this purpose. In air, the performance looks perhaps slightly
better for the single-layer design. When the index of the incident medium is taken
into account, the average reflectance of the single-layer design increases, although

20 ~-------------------------------------------------,

15

GI
U

-
C
1\1 \
u 10 \
GI
;: \
CD \
II: \

--
\
\ "m =1.0
~ \
------
5 '\ \, "
,,'" ... " ......
' ......
\', /

\
\
'"
\ "
,....."... ).(.----
o ~----~-~~~------~~--~----~~--------~------~
400 500 600 700 800 900

Wavelength (nm)
Figure 5.2. Comparison of the reflectance of a single layer (solid lines) and a double layer
(dashed lines) antireflection coating applied to a silicon substrate in incident media of 1.0
and 1.5. Note that after immersion the reflectance of the single layer design increases,
whereas that of the two layer one decreases.
148 CHAPTER 5

the performance of the two-layer design improves. Figure 5.2 shows the individual
performance curves. If the silicon cell is to be encapsulated in a protective organic
material, the reflectance should also be measured immersed in a medium with an
appropriate index of refraction.
To measure accurately the reflectance of the silicon/encapsulant interface, care
has to be taken to consider the first surface reflectance at the air/encapsulant inter-
face, as mentioned in the previous section.
Special purpose instrumentation is required when measuring the reflectance of
an immersed sample at a high angle of incidence because it is difficult to introduce
and extract a beam into the immersing medium at high angles.

5.1.5. Polarization
Unrecognized polarization can cause insidious problems in spectrophotometry. When
a sample is placed in a dual beam spectrophotometer at or near normal incidence,
any intrinsic polarization effects cancel because of the ratioing that takes place
between the two beams.
When the sample is placed at a nonnormal angle with respect to the optical
axis of the spectrophotometer, there will be some polarization effects introduced
into the system since thin film filters behave differently in p as compared with s
plane polarized light. The polarization state of the beams in the typical spectro-
photometer without polarizers is unknown, and it is likely that the intensity of the
two polarization components will be different. The reason for this is that spectro-
photometers have many optical components that cause the light with which they
interact to become partially polarized. Reflection diffraction gratings, for example,
reflect light that is highly polarized. Prisms and mirrors that are used at nonnormal
incidence can also introduce some polarization to the beam. Thus, in a spectro-
photometer that has no specific polarizer, the interaction of the sample beam with
the filter may lead to unpredictable results.
One solution to this problem is to install a polarizer in each beam of the
spectrophotometer when measuring samples at nonnormal angles of incidence. For
angles of incidence less than 30 degrees, however, this may not be necessary.
Sheet polarizers can be used for the visible portion of the spectrum, but these
can introduce problems of their own because for the most part they are not very
good optical elements:
1. They can introduce aberrations into the wavefront unless they are laminated
between glass plates
2. Their spectral range is limited
3. They do not produce complete polarization of the transmitted light
4. They are absorbing, and thus reduce the amount of light available for the
measurement
Their advantages include:
1. They are relatively inexpensive
2. They are compact and simple to use when compared with other techniques
FILTER PERFORMANCE MEASUREMENTS 149

For precision work over broad wavelength ranges, calcite and other crystals are
used to make prism polarizers such as Glan-Taylor and Glan-Thompson. Some of
these crystals are fragile and sensitive to moisture, and the assembled prism can
be large (several centimeters long).
In the infrared portion of the spectrum, wire grids are frequently used as
polarizers in addition to some crystals.
Polarizers are characterized by the amount of light they transmit (how much
less than the theoretical 50 percent) and the attenuation, suppression, or extinction
ratio of the two planes of polarization. The extinction ratio indicates how well the
polarizers discriminate between the two planes of polarization. With this infor-
mation, one can estimate the error that will be introduced in a measurement. This
should be evaluated as a function of the wavelength as the transmission and effi-
ciency of polarizers is not constant. In some portions of the infrared spectrum, the
energy reaching the detector is quite low in the best of situations. Adding a polarizer
that cuts the intensity by at least a factor of 2 makes the situation even worse. The
combination of instrumental polarization and polarizer orientation may reduce the
energy getting to the detector to an extremely low level. In some instances, a
depolarizer or polarization rotator can be of some help.

5.1.6. Measurement Cone

The 11# of the cone of light incident on a sample in a spectrophotometer gives rise
to a range of angles of incidence on the sample surface. The relation between the
f/# and the maximum cone angle is given by
6max :: tan- i 11(2 . f/#),
where 6max is the maximum angle for the cone defined by the f/#. The values
corresponding to this are included in Table 5.1. Table 5.2 lists the approximate f/#
of some spectrophotometers commonly in use today.
Anf/8 cone, for example, has a range of angles from 0 to 3.6 degrees, which
is relatively small as compared with the angular sensitivity of most interference
filters. The results of a performance measurement with such an instrument would
not be skewed except, perhaps, for the narrowest of bandpass filters. An f/4 in-

Table 5.1 FIN umber

and Cone Angles
Half cone
angle
'1# (degrees)

1 26.6
2 14.0
4 7.1
8 3.6
16 1.8
150 CHAPTER 5

Table 5.2 Approximate fin umber of Common

Spectrophotometers
Spectrophotometer fI#
Beckman DK2, DK-2A 10
Beckman IR4, IR-12 10
Cary 14 & 17 8
Cary 90 8
Cary 210,2200,2300 8 to 10
Gamma 180 5
Nicolet FTIR 6000 5
Nicolet FTIR MX1 5
Perkin Elmer 180 10
Perkin Elmer 983G 4
Note: Many instruments have slightly different I/*s in the vertical and the horizontal
planes.

strument, on the other hand, has a maximum cone angle of 7.1 degrees. This could
be the source of some differences between the theoretically evaluated (collimated
light) and the measured spectra. As previously mentioned in this chapter, reflectance
and transmission are almost never measured at exactly 0 degrees angle of incidence.
If a 5-degree tilt of the sample is used for making a transmission measurement,
and this is added to the 7-degree cone angle for an f/4 instrument, then the angle
of incidence on the sample varies over the range from 0 to 14 degrees. This could
be significant for some very narrow band (less than 0.5 %) filters (see Fig. 3. 15).
In addition to the cone angle, the light in a spectrophotometer also has a spread of
wavelengths caused by the finite slit width; this has to be added to the cone angle
to get the total effect of the instrumentation on the measured result.
In narrow bandpass interference filters, the wavelength of the center of the
pass band varies with the angle of incidence caused by the blue shift as discussed
in section 3.6. The measured peak transmittance of a narrow bandpass filter illu·
minated by a cone of light differs from one that is illuminated by collimated light.
The finite cone of light integrates the perfonnance over a range of angles, and the
effect of this is to reduce the total indicated transmitted intensity as compared with
that of a collimated beam. Other types of filters whose measurement may be affected
by this shift of perfonnance with angle are (1) edge filters, where the slope of the
edge may appear to be less steep than it actually is and (2) high perfonnance AR
coatings, e.g., V·coats at the minimum reflectance point, where the low reflection
is not as low as expected based on normal incidence calculations with a O-degree
cone angle.
Reflectance attachments for spectrophotometers are frequently designed for
to·degree angle of incidence or greater. It is rare to find an instrument that can
measure the reflectance at angles much closer than this to the normal. Thus, in
reflection, one can expect that the angles may range from 7 to 13 degrees for an
fl8 instrument.
Some commercially available spectrophotometers have different f/# in the
vertical and the horizontal planes. This complicates an evaluation of the effect of
thef/#.
FILTER PERFORMANCE MEASUREMENTS 151

In summary, the measuring of interference filters in a focused beam is useful

because in many instances that is the way that they are going to be used. In certain
measurement situations, it is simply necessary to account properly for the variations
in the angle of incidence.

5.1.7. Nonspecular Samples

All optical elements have some level of scatter. Some of the scatter is caused by
bulk effects in the substrate, and we are concerned about this type of scatter in this
discussion. Some surfaces, for example, those used for laser gyroscope mirrors,
are nearly perfect in this regard, with surface scatter of visible light measured in
parts per million of the incident intensity. At the other extreme, some samples have
surfaces that are intentionally made scattering. For example, the reflection from the
new generation of frosted fused silica solar cell covers is entirely diffuse and has
no specular component.
Specular generally means that the light beam retains its original distribution
pattern after encountering an optical element. Diffuse means that some of the light
in a specular beam is removed from the beam by the optical element and is redirected
(scattered) in a direction other than the nominal one predicted by the laws of
reflection and refraction for the macroscopic surface. The macroscopic scatter can
be caused by reflection or refraction on a microscopic scale, or by true scatter from
pointlike sites on the surface of the optical element. Note that the terms of specular
and diffuse apply to both reflection and transmission. The effect of scatter is to
reduce the intensity of the specular beam without reducing the total light in the
system. Absorption, on the other hand, reduces the intensity of the specular beam
as well as reducing the total light in the system.
In an ordinary spectrophotometer that measures the intensity of a specular
beam, it is not possible to identify whether an observed intensity reduction is caused
by absorption or scattering-they must be lumped together as one loss mechanism.
Integrating spheres can be used to measure total integrated reflectance and trans-
mittance as it can collect the energy scattered out of the specular beam as well as
that which is specularly reflected from the sample. With such a technique, the only
light unaccounted for is absorption. The many reflections that the light undergoes
inside an integrating sphere before reaching the detector limits the sensitivity, thus
making only relatively high levels of coating absorption in a scattering sample
measurable by this technique.
Appropriately configured integrating spheres can be used to measure absorp-
tance (Edwards et aI., 1961). First, the sphere is calibrated for 100 percent with
the sample removed from the sample beam. Then the sample coating (on a trans-
parent substrate) is suspended so that it is at the center of the sphere and intercepts
the sample beam. Now both the total transmitted and reflected beams are collected
by the sphere, and the difference between the no sample and with sample mea-
surements is the absorption in the sample. *

*Gier Dunkle Instruments, Inc., 1718 21st Street, Santa Monica, Calif. 90404, make such an
attachment, Model SP220, for the Beckman DK-2 spectrophotometer.
152 CHAPTER 5

A narrow band interference coating functions poorly if used in a beam that

has a large diffuse component. The in-band light scattered at the higher angles of
incidence is not passed by the filter, whereas some out of band light arriving at
nonnormal angles of incidence light is transmitted. Note that light with wavelengths
other than the nominal filter is transmitted because of the angle shift to shorter
wavelengths. Similar situations can arise with edge filters that are used to discrim-
inate between closely spaced wavelengths. If a component in an optical system is
scattering, then narrow bandpass or other wavelength sensitive interference filters
should be located so that the beam strikes the filter before it reaches the scattering
element.
Scattered light is a problem in antireflection coatings (transmitting systems).
The scattered light appears as a haze on the surface when viewed by eye (scatter
in other portions of the spectrum may be detected with appropriate instrumentation).
This is especially noxious when a bright light source is located off the optical axis
of the optical system, and a dim object is being viewed through the system. A
similar situation occurs in a reflective system when a mirror is used to observe a
faint source and extraneous light floods the primary or other mirror, which may
then scatter light into the detector.
In precision filters, it is necessary to use the highest quality substrates for peak
performance. Furthermore, the coating should not increase the scatter already pre-
sent in the substrate.

5.1.8. Nonplanar Surfaces

A nonplanar surface, for example, a concave or a convex lens surface, is very
difficult to measure accurately without special tooling or fixtures custom designed
for the purpose. The reason for this is that typical general purpose spectrophotom-
eters have optical systems that assume the sample is flat and will not introduce any
power or other aberration into the optical system. If the surface is not flat, the
sample-introduced aberrations can change the optical system between the reference
beam and the sample beam. This change in the optical system manifests itself as
a change in the transmission or reflection of the part being measured, but which,
in fact, is not related to the performance of the surface being measured.
It is easier to measure the transmission of a nonplanar sample, e.g., a flexible
sheet, than to obtain its reflection. Relatively little aberration is introduced in a
collimated beam by a thin sheet with parallel surfaces, and this is how a transmission
measurement is made. In reflection, however, the nonplanar surface has a first order
effect on the wavefront so that deviations from a flat surface can cause a large loss
in the accuracy of the measurement. A spectrophotometer can be used to obtain a
qualitative sense of the spectral shape of a nonplanar sample, but as some of the
light may be lost after it has been distorted by the sample, the peak value may not
be accurate. An integrating sphere may be helpful in desensitizing the detector to
beam alignments and aberrations.
Some instruments have been available in the past that made spot reflectance
measurements. These were capable of measuring curved parts more accurately than
FILTER PERFORMANCE MEASUREMENTS 153

standard instruments. Currently, there is only one company making a spectropho-

tometer capable of measuring the reflectance of small, nonplanar spots. *

5.1.9. Sample Heating

Care should be exercised during the measurement of narrow bandpass or precision
edge filters to ensure that they are measured at the temperature at which they will
be used. There are two situations here:
1. The spectrophotometer can warm a sample while measuring it
2. The filter is intended for use at a temperature other than ambient
As mentioned in section 3.7, the performance of most filters is a function of
temperature. Many antireflection coatings and other filters with broad, flat spectral
shapes are not temperature sensitive because a small wavelength shift in their
performance would probably go unnoticed. In a narrow bandpass filter, on the other
hand, the spectral position of the pass band may be critically located to accuracies
of better than I percent. In these instances, the shift introduced by a change of
temperature can be significant.
Radiant heating of a sample by the light source in a spectrophotometer may
occur:
1. If a filter is absorbing over the spectral range of the light source in the
spectrophotometer
2. If the wavelength selective device in the instrument is located between
the sample and the detector
It is common in infrared instruments to place the sample between the light source
and the wavelength separator so that some visible light arrives at the sample to
locate the beam; in visible spectrophotometers, the sample is placed after the wave-
length separator. Another source of sample heating is the spectrophotometer light
source via conduction and convection mechanisms. These sources of heat need to
be evaluated when measuring heat sensitive filters.
Filters that are used at cryogenic temperatures should be measured at their
operating temperature. A cryogenic sample holder is required to make these mea-
surements (Fig. 5.3). Care should be exercised to prevent the window from frosting
during the measurement process, otherwise unreliable values are obtained. The filter
appears to change with time as the frost on the windows accumulates. The sample
should be well cooled before measurements are started. After the measurements,
the sample should be examined to ascertain that the low temperatures did not damage
it.

5.1.10. Measuring Absorption by Calorimetry

Calorimetry is a very accurate way of measuring the absorption in a coating. This
technique is accurate because it quantifies the absorption by measuring the tem-

*Model 1001 Spectrophotometer is available from Zygo Corporation, P.O. Box 448, Laurel Brook
Road, Middlefield, Conn. 06455.

154 CHAPTER 5

LN2 Cold Shield 77K

Dewar Wall

3--+---t+-- He Refrigerator 4K

L----t---f~ Cold Finger

-""".~~--

Tilted Windows

Vacuum Sample

Atmospheric Pressure
Figure 5.3. Diagram of a cryogenic sample holder for a spectrophotometer. The windows'
tilts (shown exaggerated) prevent multiple reflections within and between them from contri-
buting to the performance spectrum. Note that the windows must be kept warm or purged
on the outside with a dry gas to prevent condensation of moisture on them. The windows
need to be transparent over the spectral range being measured, and they may require
antireflection coatings.

perature rise in a sample illuminated by a laser beam. The temperature rise is, to
first order, proportional to the power in the incident beam (assuming, of course,
that the power level is not so great as to damage the part being measured). High
sensitivity is achieved by using low thermal mass and low specific heat substrates,
and by designing special fixtures to hold the sample in thermal isolation. With
proper instrumentation, light scatter from the substrate is not a problem. Ultras mall
thermocouples measure the temperature rise of the sample. Mounting the sample
in a vacuum chamber evacuated to a pressure of 133 mbar (100 torr) eliminates
any heat losses from the sample by conduction to the air and by convection. Figure
5.4 shows a schematic diagram of a calorimeter. In addition to the thermal con-

FILTER PERFORMANCE MEASUREMENTS 155

Evacuated Tank

Sample
Support
Tilted windows
with AR coatings

Laser Beam Sample

-...--++

Thermocouple
Thermocouple Output
Figure 5.4. Diagram of a laser calorimeter. The incident laser power can be in the 1- to
10-watt range. The chamber is evacuated at least to a pressure at which the heat loses by
air conduction and convection become insignificant in comparison to other experimental
errors.

siderations, care needs to be exercised to assure that the transmitted beam (for
transparent substrates) is directed properly to a beam dump, lest the reflected beam
return to the sample area, additionally heat the sample, and cause systematic errors.
A stray beam such as this could also heat the measuring thermocouple. The typical
procedure relies on measuring the heating and cooling rate curves of the sample,
and deriving the absorption from these data. Calorimeters can measure absorptances
down to 2 x 10-6 (Allen et aI., 1978).

5.1.11. Colorimetry
The perceived color of the light reflected or transmitted by a filter can be important
in some applications. Colorimetry quantifies the color of a light beam and gives
the results in standard terms. A common means of providing this is based on the
Commission Intemationale de l'Eclairage (C.I.E.) chromaticity diagram, Fig. 5.5.
This chart is based on data gathered in the late 1920s and codified in 1931. Recent
improvements to this chart include the 1976 CIE-UCS system (CIE, 1978), which
displays an improved uniformity of color differences over its predecessor (Keller,
1983). This latter chart is shown in Fig. 5.6.
Two instrumental methods are used to evaluate color. In one system, the
intensity of the light in the visible portion of the spectrum is measured. The color


156 CHAPTER 5

0.8

.51

0.7

0.6 Gr •• n

.50 G•••• llh Y.llow

0.5
Y.llowllh O.e.ge

0.4

0.3

.77


0.2

0.1

.0

L-~----~~----~~~~~----~----~----~~--~~~x

.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7

Figure 5.5. CIE color coordinate diagram (1931) (adapted from a Photo Research chart).
This is the X-V coordinate system with the color regions bounded by Kelley (1943). The
solid line with numbered tics is the locus of the coordinates of the color of a blackbody
emitter at the indicated temperatures. The wavelength in micrometers of the pure colors at
the periphery of the diagram are indicated.

of the beam is quantified by calculating the tristimulus values according to the

following equations:

x = f <I>(A)X(A)dA
Y = f <I>(A)Y(A)dA
Z = f <I>(A)Z(A)dA
where <I> is the spectral radiant flux and x, y, and Z are the red, green, and blue
spectral tristimulus functions, respectively.

FILTER PERFORMANCE MEASUREMENTS 157

y'

820

0.5 • .--~

0.4

0.3

0.2

0.1

~ ______-L________ ~ __ ~~~ ________ ~ ______ ~ ________ ~ __ ~u'

o 0.1 0.2 0.3 0.4 0.5 0.6

Figure 5.6. CIE (1976) color coordinate system (Keller, 1983) (adapted from a Photo
Research chart). This is the u'-v' system. On this representation, equal distances approxi-
mate equal distinguishabilities of colors. The solid line with numbered tics is the locus of the
coordinates of the color of a blackbody emitter at the indicated temperatures. The wavelength
in nanometers of the pure colors at the periphery of the diagram are indicated.

The 1931 eIE chart displays nonnalized fonns x and y of the tristimulus values:
x
x=
X+Y+Z
Y
y
X+Y+Z
In the 1976 fonn of the chart, a different transfonnation is used:
, 4X
u = and
-2X + 12Y + 3
, 9Y
v =
-2X + 12Y + 3'
and the values of u' and v' are plotted on the latter chart.
158 CHAPTER 5

The second method of detennining chromaticity coordinates makes use of

calibrated transmission filters to do these calculations optically. The filters' trans-
mission pass bands match the three functions x, y, and z. One or more detectors
convert the filtered light to electrical signals. Simple circuits can then be used to
provide an indication of the color coordinates. In this approach, the calibration can
be difficult and subject to drift. This is not the case with the technique that uses a
dual-beam spectrophotometer. In addition, the spectral curve from a spectropho-
tometer supplies added infonnation not available from the integrated quantities. On
the other hand, the filter approach can be packaged in a small volume and taken
out of a laboratory and into the field.

5.2. ENVIRONMENTAL PERFORMANCE

AND MEASUREMENTS

An optical filter must be robust enough to survive in the environment in which it

will be used, along with the conditions encountered along the way to that final
location, e.g., shipping and storage. This is a difficult area to define ahead of time,
as these conditions are subject to change with time. In addition, it is difficult to
guarantee that a filter will survive for x years based on 30 days' testing, for example.
Because the mechanisms for degradation are not known precisely, but rather only
in qualitative tenns, no theoretical models have been devised that will predict the
lifetime of optical filters in storage or in use. This section describes the methods
currently available to address these needs from a pragmatic point of view.
Pragmatic and functional tests have been devised over the years to evaluate
filters more or less objectively. These tests have been codified in various govern-
mental specifications; some of these are included in the Appendix B for reference
purposes.
From an optical point of view, the lifetime of a typical coating is essentially
unlimited because it is a passive component in the same way that a lens does not
wear out. In some instances, however, the components of the filter may be attacked
chemically or mechanically by environmental agents. An example of the fonner is
sulfur attacking a silver coating, while the wiping of a coated surface with an
abrasive material is an example of the latter. Obviously, it is not possible to predict
how long a coating will last under all conditions as there are so many variables
that could affect the reaction rate. Parameters, such as temperature and humidity,
are synergistic, and they are not typically well controlled. The only way to evaluate
completely the suitability of a coating is to get a sample of it, expose it to the
environment, and observe its degradation. If it degrades too much or too rapidly,
a different design should be investigated.
The tests mentioned in this section can be done singly on separate coated
substrates or sequentially on one coated substrate. An advantage of separate sub-
strates for each test is that the tests can proceed in parallel, thus cutting down on
the time required to complete the tests. This can be significant, as some tests can
take a week or a month to complete. When this is the case, i.e., the testing time
is long and the available time is short, one can use a combination approach by
making the assumption that the parts will pass the testing. Such an assumption is
FILTEA PERFORMANCE MEASUREMENTS 159

reasonable because the environmental durability of a coating type rarely changes

given a well developed and controlled manufacturing process. In such a scenario,
some witnesses undergo a cursory testing to ascertain in as short a time as possible
that the basic process is sound, whereas others are subjected to the rigorous proof
testing over a longer time period. In this way, it is possible to monitor a production
process and receive timely feedback of information to control the process, without
waiting for the definitive testing. On the other hand, these approaches can use up
many witnesses and valuable space in the coating chamber.
Sequential testing is more rigorous than testing separate substrates. For ex-
ample, a sample may pass a test such as adhesion prior to its exposure to humidity,
and then fail the same adhesion test after the humidity exposure. Care should be
exercised in deciding which tests are required before one specifies the environmental
testing, because this could increase the difficulty of manufacturing the filter. Worse
yet, in some instances, is the fact that it may not be possible to build a filter that
meets the specified environmental durability constraints, or its optical performance
may be reduced to meet them. If these requirements are not operationally necessary
in the first place, then the result is an unnecessary reduction in performance.
In the evaluation of the results of an environmental test, one should be aware
that the level of reflectance of the coating being tested has an effect on the ap-
pearance. Most tests rely on visual examination of the part after the test is complete
to determine whether or not damage has occurred. A coating whose reflectance in
visible light is very low will allow the observation of very low levels of scattered
light, whereas the same physical damage in a coating whose reflectance in the
visible is high may not be visible against the high reflectance background.

5.2.1. Use of Witnesses in Environmental Testing

In the previous discussion, we have referred to the testing of witness parts, rather
than the testing of actual coatings and substrates that may actually be used in some
application. There are a number of reasons why one should consider testing witness
parts rather than the actual deliverable parts.
One or more witness parts must be used when:
1. The actual part is too large to be tested directly
2. The actual part is too expensive to risk in testing
3. The test is for informational purposes only
4. The shape or finish of the coated parts do not lend themselves to simple
measurements
5. The testing is destructive
The measurement of the optical parameters, e.g., the reflectance, of a coating is
generally a nondestructive test, and no damage is done to the coating. High-power
laser damage testing is an exception to this. Environmental testing, on the other
hand, must sometimes damage the substrate, as in abrasion testing that continues
until damage is observed. Environmental tests can also leave a residue on the
surfaces. Thus, a strong case can be made that an actual part should never be tested.
In practice, most environmental testing is done on witness parts.
There are some precautions that must be taken to ensure that the witness is a
160 CHAPTER 5

true representation of the actual part. The use of a witness is acceptable if the results
are really what would be obtained if the actual part were measured. This can usually
be achieved by coating the witness(es) in the same chamber and at the same time
as the actual part is being coated. Temperature, angle of deposition, and coating
thickness are some of the coating parameters that must be the same on the witness
and on the part. One must exercise care in the comparison of the actual and the
witness parts to be certain that the coatings on them are indeed similar in perfor-
mance.

5.2.2. Adhesion Testing With Self-Adhesive Tape

In the vast majority of uses, coatings should not become separated from their
supporting substrate during their lifetime, nor should they fail at the interfaces
between layers. Adhesion tests are perfonned to validate the integrity of a coating
and to demonstrate the absence of a vulnerability to stress failures. A number of
tests have been devised to detennine if the coating adheres well to the substrate.
In this section, we discuss only the test based on the use of self-adhesive tape.
Other adhesion tests are discussed later in section 5.2.6.
The most familiar adhesion test is the Scotch Tape® test, or simply, tape test,
in which a specified pressure sensitive adhesive tape is pressed onto the coated
surface ~d then peeled off again. The test procedure specifies that if no coating
material is removed by this test, then the coating has passed the test. The coating
and the substrate should be clean and dry before this test is perfonned.
An interesting phenomenon occasionally occurs when certain films, such as
metallic lead, are tested this way. The adhesive sticks much more tenaciously to
the lead coating than to its cellophane backing. As a consequence, when the tape
is pulled off the coating, the adhesive remains behind on the coating. The adhesive
residue may be cleaned off with a solvent, and the part examined for damage. If
there is no damage, the part passes the test (but the tape fails?).
The standard tape test is nondestructive if the coating passes the test. On the
other hand, a failure is inherently destructive because some of the coating has to
be removed by the tape.
A more challenging and rigorous adhesion test was adopted by the paint
industry, to test the adhesion of paint to a surface. It is occasionally used to test
optical coatings, though this is by no means a standard in this industry. It is
inherently destructive. It consists of scribing with a razor blade, diamond knife, or
other sharp tool a crosshatched pattern on the coated surface to be tested. Pressure-
sensitive adhesive tape is applied to the scribed area and then peeled off. If any of
the coating is found on the tape after it has been removed, the coating has failed
the test. Naturally, because this is a destructive test, it should be perfonned on a
witness substrate.

5.2.3. Abrasion Testing

Exposed coated surfaces frequently need cleaning, as do all other exposed optical
surfaces. The typical user may wipe the optical surface with a cloth; sometimes,
the cloth may be moistened, at other times, it may not. If the coating is soft, this
FILTER PERFORMANCE MEASUREMENTS 161

treatment will likely damage it. The abrasion test attempts to evaluate and to quantify
the durability of the coating to this type of abuse.
Two tests are in common use for measuring the abrasion resistance of coatings:
the cheesecloth and the eraser tests. The tests are always performed on clean areas
of the coating.

5.2.3.1. Cheesecloth Test. In the cheesecloth test, a pad of cheesecloth

approximately 9.5 mm (3/8 inch) in diameter and 12.5 mm (0.5 in.) thick, is pressed
against a coated surface with a force of 4.4. newtons (1 lb of force). The pad is
moved back and forth 50 times (25 cycles or round trips) over a given coated area.
The substrate is then examined under specified lighting conditions (see Appendix
B) to determine if any damage has been caused by the test. This examination is
frequently a source of disagreement \is the examination conditions are critical to
achieving uniform and repeatable results. The specifications state that no signs of
deterioration will be seen; in marginal cases, the decision on the presence or absence
of sleeks and scratches is somewhat subjective. The interpretation of these tests is
a skill that improves with experience.

5.2.3.2. Eraser Test. The second abrasion test in common use is the eraser
rub test. In this test, a well-specified pencil eraser (type and hardness of the rubber
and the amount of pumice in it are specified) is pressed against a coated surface
with a force of about 9 to 11 newtons (2 to 2.5 lbs. of force) and rubbed 40 times
over a given coated area. As with the cheesecloth test, the coating is then examined
for damage. This test is much more difficult to pass than the cheesecloth test as
the pumice is a much harder material.

5.2.4. Moisture Exposure

A number of humidity tests are extant, and they vary in several ways. In some,
the temperature and humidity are kept constant, whereas in others, one or both of
these parameters vary as a function of time. In the tests in constant conditions, the
specifications give the number of hours of exposure time the sample undergoes. In
time-varying tests, the numer of cycles the witness must withstand is the criterion
specified.
The goal of all these tests is to accelerate the aging process so that some
meaningful data can be obtained in a short amount of time. In the ideal case, the
coating systems behave in a linear manner, i.e., the amount of time in a humidity
chamber is proportional to a certain amount of time in an actual or field environment.
This is not necessarily true in reality. More likely, the degradation mechanism that
takes place in highly accelerated testing is not the same one that causes the longer
term degradation. The tests are used despite these problems because they are the
best techniques currently known to accelerate the aging process simply and re-
peatably. In practice, the testing has evolved such that when a sample can pass a
given test, one can be reasonably confident that the part will last the anticipated
lifetime in some environment. There is no firm scaling factor for any of these
environmental tests. The typical approach is that if a part, for example, passes a
162 CHAPTER 5

humidity test such the 24- or 48-hour test, then the coating is reasonably durable.
The same can be said of the other tests.

5.2.4.1. Humidity Testing. Humidity tests consist of placing a witness part

in a chamber whose temperature and humidity are controlled. Common time and
humidity conditions are specified in Appendix B. The witnesses are placed in the
chamber at specific times during cycled test sequences so that the possibility of
condensation of water on the coatings is minimized.
An example of an effect that occasionally takes place in a humidity chamber
in which both the temperature and the humidity level are cycled is that condensation
sometimes does place, either intentionally or not. The liquid water on the surface
may cause chemical reactions that would not happen if the part were kept free of
liquid water. The fact that this water is essentially distilled increases its reactivity
significantly. This reaction can cause the part to degrade, and thus give the ap-
pearance of a shorter lifetime than would be found in actual practice where the
optic is kept dry.

5.2.4.2. Solubility Testing. Coatings may be required to withstand a number

of solubility tests to demonstrate that they are durable enough for certain applica-
tions. A number of standard tests are available, for example, as given in Mil-C-
48497 (see Appendix B). Some tests require that coatings withstand immersions in
trichloroethylene, acetone, and ethyl alcohol for relatively short times to meet some
specifications. Optional tests include a 24-hour soak in a saline solution or in distilled
water. All of these tests are performed at or near room temperature.

5.2.5. Salt Fog Exposure

Filters that are to be used near saltwater should be tested for their resistance to
degradation by it. A standard test consists of placing the witness parts into a chamber
in which a continuous salt fog bathes them (Mil-STD-81OD, Method 509.2). The
composition of the artificial seawater that is used to make up the fog is also defined
in the test specifications. This is the only test, other than the solubility ones, where
liquid water intentionally comes in contact with the substrate for any length of time.
In general, a dielectric multilayer filter that survives for a given period of time
in a normal humidity chamber will survive the same length of time in the salt fog
test. Metal films, on the other hand, are much more susceptible to the salt fog and
its reactive ions of chlorine and sodium. Filters with metallic layers in them must
be screened carefully if they are to be used in or near saltwater environments.

5.2.6. Other Tests

Additional tests are available to evaluate coated parts. Practically all thin film
interference filters undergo at least some visual examination for cosmetic defects
before they are placed in service. At the very least, they are inspected when they
are removed from the coating chamber, and again when they are installed in service.
Microscopic examination is one test that is extremely sensitive yet not frequently
FILTER PERFORMANCE MEASUREMENTS 163

used, possibly because it is not as quantifiable as other tests. Many times, defects
that are not measurable are readily seen by eye.

5.2.6.1. Visual Inspections. One area where visual examination is the typ-
ical method used is in the examination of parts for edge chips, scratches, and cracks.
Of course, this inspection must be done on the actual parts instead of witnesses.
The chips must be less than a specified size, depending on their depth, and their
number is limited as well. Standards for scratches and digs (pits) are available, *
which allow an inspector to compare visually with the standard the size (width,
length, and depth) of a scratch or a sleek he or she has observed. The inspector
who does the visual examination must have a certain talent and lots of experience
in looking at samples to be repeatable, especially when the parts are only marginally
sleeked.

5.2.6.2. Falling Sand Test. This test is useful in situations where the optical
element will be exposed to an abrasive environment, e.g., spectacles and sunglasses
(on a sandy beach), and coated architectural glass. In this test, sand in a hopper at
a fixed height falls on a slanted substrate. The rate at which the sand falls is
controlled, as is the test time. The sample is examined at intervals to see if the
sand has eroded, abraded, or otherwise damaged the coating. This tests both the
hardness and the adhesion of the coating. A similar test is formalized in Mil-C-
14806, para. 3.11.5.

5.2.6.3. Other Adhesive Tests. Adhesion of a coating to the substrate is of

prime importance to everyone who uses thin film filters. A number of tests have
been devised to obtain a quantitative value for the strength of the bond between a
coating and the substrate. A few of these are discussed here. The reader interested
in more details about adhesion testing than is given here is referred to Mital (1978).
Topple Test. A true measure of adhesive strength between a coating and the
substrate is to attempt to pull the coating off the substrate. This is what the adhesive
tape test attempts to do qualitatively. There is no measure of how well the coating
stuck to the substrate-it is a go/no go type of test.
An instrument is commercially available that measures the tangential force
required to make a metal post fall over. From the geometry of the test fixture, it
is possible to convert the tangential force to a normal force that is related to the
adhesive forces between the substrate and the film, or between some internal film
interfaces, if the failure does not occur at the substrate/coating interface. The post
is attached to the coating with an organic adhesive and allowed to cure. This bond
must be stronger than any of the bonds being tested; otherwise, it will be the one
that fails. The tangential force needed to make the post topple is recorded and then
converted to a normal force. This test is useful in that it provides a quantitative

*Procurement and/or calibration of scratch and dig standards can be obtained from: Armament
Research and Development Center, U.S. Army Munitions and Chemical Command, AMSMC-QAF-I,
Dover, N.J. 07801-5001.
164 CHAPTER 5

value of the adhesive forces. It is assumed that in applying the adhesive to the
coating there is no chemical reaction between the adhesive and the coating. For
some epoxies, this may be difficult to determine. This is a relatively new test and
it is not in widespread use to certify production coatings at this time.
Another test that determines the bond strength between a coating and the
substrate is the measurement of the normal force required to pull off a stud securely
cemented with an epoxy adhesive to a coated surface. * A mechanism applies an
increasing force normal to the surface and provides an indication of the force at
which the bond fails. The stated accuracy of this method is within plus or minus
10 percent. The primary cause of the variation is caused by the error in determining
the size of the contact area between the stud and the coating.
Indentation Tests. This test was first described by Heavens (1950) and later
expanded by Benjamin and Weaver (1959). It involves drawing a smooth point
across the coated part. The load on the point when the film is removed is a relative
measure of the adhesion of the film to the substrate. The absolute value of the
shearing forces can be calculated from the load, the radius of the point, and the
indentation hardness of the substrate. This method was used by Benjamin and
Weaver (1961) to determine the adhesive strengths of a number of metal films on
a glass substrate.

5.2.6.4. Combined Adhesion and Abrasion. A test that evaluates the adhe-
sion as well as the abrasion resistance of a coating is the Tabert abraser test. A
rubber wheel loaded with an abrasive rolls in a circular path on the surface to be
tested. The motion is complex as the width of the wheel is such that some dragging
or sliding takes place along with the rolling. This test is not widely used in evaluating
optical thin films at this time.
No satisfactory correlation has been made among the results of the various
abrasion tests.

5.2.6.5. Combined Temperature, Humidity, and UV Light. A humidity

test sequence that has been found to be very vigorous in attacking coatings is the
combination of a cycled temperature and humidity environment with simultaneous
exposure to ultraviolet light. Some reactions, which otherwise may not be ener-
getically feasible, can take place when the ultraviolet photons contribute their energy
to the chemical reaction. At first glance, this type of test may appear too severe;
however, especially in situations where the filters, in particular those with metal
layers, are exposed to sunlight, this test may be valid. It has found use in the testing
of paints and other similar coatings. No formal test procedure and specifications
exist for use of this type of test for optical coatings.

*An instrument which perfonns this test is called the "Sebastian Adherence Tester," and is available
from Quad Group, 331 Palm Avenue, Santa Barbara, CA 93101.
tModel 503 is made by Teledyne Taber, North Tonawanda, N. Y. 14120.

FILTER PERFORMANCE MEASUREMENTS 165

5.2.7. Cycled Temperature Testing

In some applications, filters are exposed to wide ranges of temperature. Prior to
their being placed in service, they should be tested to ensure that they will survive
the anticipated temperature excursions. The most common upper temperature limit
encountered is 100°C. There are a number lower temperature limits in common
specifications, including O°C, - 40°C (dry ice), 77K (liquid nitrogen), and even
4K. Cyclic temperature testing is frequently required of space satellite components
and those installed in unpressurized portions of high altitude aircraft. The number
of test cycles should be in excess of the anticipated operational number, but there
is no hard and firm rule for the safety factor to use. It is probably safe to say that
the first temperature excursion is the most difficult, and that if the filter survives
the first cycle it will likely survive the succeeding ones.

5.2.8. Combined (Sequential) Testing

Some test procedures can produce hidden damage that will cause subsequent tests
on the part to fail. If the tests were done on separate substrates or in the reverse
order, the conclusions drawn from the results might be different. An example of
this is doing an eraser test on a high reflector prior to a humidity test. The eraser
test may cause minor damage to the coating, but as the coating is a high reflector,
the damage goes unnoticed and the test is deemed a "pass." In the subsequent
humidity testing, the moisture enters through the damage sites and attacks the inner
layers. This could not happen if the surface layers were intact. Thus, it is easy to
see that sequential testing can lead to a more difficult test sequence than individual
or parallel testing of different samples. In addition, it is also obvious that the ordering
of the test sequence may have an effect on the results.
A general rule concerning the order of testing is that the quick tests be done
first to produce results as quickly as possible for process control. Tests, such as
abrasion and adhesion, fit this criterion in that they can be done in a few minutes.
If these tests fail, the information can be given to the production operation in a
timely manner. The longer duration tests, such as humidity exposure, can then
follow. Finally, the comprehensive product testing to verify to the customer's
satisfaction that he or she is getting a quality product can be done.

5.2.9. When to Test

The age of the sample is sometimes important to the testing process. The wavelength
shift of pass band and edge filters due to moisture absorption in the coating has
previously been mentioned in section 3.7. Similar relationships can arise in envi-
ronmental testing. For example, atmospheric moisture can lubricate a surface so
that its coefficient of friction is not as high as one that is freshly removed from the
coating chamber. An eraser abrasion test made on such a part will, therefore, vary
with the age of the coating, as well as on the relative humidity. The rate of change
will be the largest immediately after the sample's exposure to the atmosphere, and
it will decrease to a low value after a few hours, depending on the relative humidity.
166 CHAPTER 5

The adhesion of aluminum films to a glass substrate increases with time (Benjamin
and Weaver, 1961). The conclusion one draws from all of this is that a standard
period of time should pass after removing the parts from the coating chamber before
testing to obtain the maximum amount of test to test repeatability.
Another factor that enters into testing is the cleanliness of the parts being
examined. Abrasive particulate matter on the surface during eraser or cheesecloth
testing may affect the test results. Similarly, particles between the adhesive tape
and a surface could reduce the adhesive forces between the tape and the coating,
thus invalidating the results. It is a standard practice to clean a surface before testing
it. The solvents used for this purpose should be carefully chosen, as some of these
solvents have been known to cause problems of their own. It was discovered some
years ago that when chlorinated hydrocarbon liquids were used to clean parts before
coating, the parts failed humidity testing. Apparently, some hydrogen chloride
residue remained on the surface after the cleaning step, and this subsequently turned
to hydrochloric acid when exposed to moisture. This acid then caused deterioration
of the films.
Filters are sometimes subjected to moderately high temperatures (> 100°C)
during the manufacturing process. This may occur intentionally and be done by the
filter manufacturer to densify the coating or desorb moisture. It can also be done
incidentally, in that it is a part of the manufacturing process after the filter has been
incorporated into another subassembly or component. The filter can shift its per-
formance as a function of wavelength asa result of this heating. Also, the optical
performance can charge because the films in the filter change their oxidation state.
Generally, this means that the coating becomes clearer and less absorbing after
heating. In other cases, however, the coatings become absorbing as oxygen is driven
out from the layers, and then cannot get back in after the heat is removed. In other
situations, the coating materials are simply unstable at elevated temperatures. Wit-
nesses for filters that will be heated in use must also be heated prior to environmental
testing to simulate the life cycle of an actual part. This is especially important if
the filters undergo large temperature excursions during their subsequent processing
or in their use.

5.2.10. Formal Specifications

Both military and civilian agencies have formulated standards by which thin films
can be evaluated. The most common ones in the United States are the military
specifications. The designations of these generally start with "Mil" and are, there-
fore, colloquially called "Mil specs." It is understood that the West German stan-
dards organization is preparing a set of specifications for thin films for international
civilian use.
The original military specifications were for magnesium fluoride antireflection
coatings and for aluminum high reflector coatings. These specifications were sub-
sequently adapted and applied to other types of thin film coatings, to which they
were more or less appropriate. Of late, a new set of Mil specs has been issued that
deals more comprehensively with the specific needs of modem optical coatings,
and interference filters in general. The designations of the new specifications are
FILTER PERFORMANCE MEASUREMENTS 167

Mil-C-14806 Coating, Reflection Reducing, for Instrument Cover Glasses and

Lighting Wedges
Mil-C-48497 Coating, Single or Multilayer, Interference: Durability Require-
ments for
Mil-F-48616 Filter (Coatings), Infrared Interference: General Specification for
Because these are of general interest, they are reproduced in their entirety in Ap-
pendix B. The previous specifications (now obsolete for interference filter appli-
cations, though still referenced) are
Mil-C-675 Coating of Glass Optical Elements (Anti-Reflection)
Mil-M-13508 Mirror, Front Surfaced Aluminized: For Optical Elements
A standard testing specification is
Mil-STD-810 Environmental Test Methods and Engineering Guidelines.
All manner of testing military components and assemblies are described in this
large manual. The majority of these tests are not applicable to optical coatings.
Specific sections of this standard, however, are referenced in other specifications.

5.3. AGING

Optical filters are passive devices, and as such, should have an unlimited lifetime
when properly used. There is nothing to wear out, and nothing in the filter is
consumed. On the other hand, nonoptically related factors can sometimes put a
limit on the shelf life and the working life of a filter. Of course, mechanical shocks
can break some of these filters, whereas too high an energy level, e.g., too powerful
a laser pulse, can damage them. Though most coating materials are inorganic, some
molds and mildews can grow on just about any solid surface, including glass. Thus,
an optical filter, like any other optical component, should be kept in a clean, dry
environment for long-term storage.

5.3.1. Shelf Life

Many of the materials that are used to form coatings are refractory fluorides and
oxides that are insensitive to temperature variations. On the other hand, humidity
is a problem with some of these materials. For example, chiolite and cryolite (sodium
aluminum fluorides with different ratios of sodium and aluminum ions, nominally
approximately Na3AIF6) are two materials used in the earliest vapor deposited
coatings. They were used because they are easily evaporated from resistively heated
sources, and because they have low refractive indices, which make them among
the best choices for single layer antireflection coatings. Unfortunately, these fluo-
rides are soluble in water. When coatings with these materials in them are exposed
to humid environments, they may undergo changes as the ambient humidity diffuses
into the filters. The moisture may simply dissolve the material and remove it. Note
that a layer 1 quarter wave in optical thickness in the visible portion of the spectrum
consists of only 100 to 200 atomic layers. In terms of mass density of the antire-
flection coating, this translates to about 300 ng/cm2 (3 X 10-7 g/cm2 ). Very little
168 CHAPTER 5

moisture is needed to dissolve away this small amount of material. Fortunately,

many of the materials currently used by film manufacturers are nonsoluble com-
pounds, so the problem of moisture's dissolving the film materials is minimized.
In some spectral ranges, however, the manufacturer might still have to use somewhat
soluble materials to meet a specification.
Although the use of non soluble oxides minimizes the problem caused by
moisture, it is by no means completely solved. Condensed moisture is akin to
distilled water, and this can be a surprisingly strong etch ant for thin films. Only a
small number of atomic layers make up a layer, therefore, it is not difficult to see
why freshly condensed moisture, dissolving only a few molecules from the surface
at a time, will eventually change the optical performance. Thus, a coated optical
element exposed to condensing moisture may have a shorter life than one that is
kept free of moisture, or even one that is immersed in less pure water.
Because humidity is such a deleterious component of the environment, various
means have been designed to avoid the problem of humidity degradation. The most
successful one is the hermetic seal, wherein the moisture is prevented from entering
the region where the films are located. The seals can take the form of a solder or
other glass to metal contact, a glass frit seal where a layer of glass is used to prevent
the diffusion of water into the filter, and other similar arrangements. Inorganic
materials are generally less permeable to moisture than organic ones. A rule of
thumb is that water permeates a given thickness of organic material an order of
magnitude faster than an equivalent thickness of inorganic material. Therefore,
inorganic sealants are preferred over organic compounds for hermetic seals. Ex-
amples of coatings that need to be sealed are ones that have thin metal layers, such
as Fabry-Perot cavities with silver reflectors, and optical data disk coatings that
have a thin metal layer as the sensitive recording medium. High-quality filters of
these types are generally supplied sealed.
Most coatings have intrinsic stresses in the individual layers, even though the
net stress in the total coating may be low. As a result of aging, some of these
stresses may relax because of the rearrangement of the atoms or molecules in
individual films. The result of such an annealing over time could be to upset the
original stress balance in the design. A low stress coating is preferable for long-
term stability.

5.3.2. Material Interactions and Interdiffuslon

In some uses, filters are exposed to high temperatures. This environment can make
several degradation mechanisms energetically allowable. This can decrease the
lifetime of a filter.

5.3.2.1. Oxidation and Reduction. A metallic coating can be prone to

oxidation when exposed to high temperatures in air or other oxidizing atmospheres.
Nonoxide compounds can likewise be converted to oxides under appropriate con-
ditions of heat and oxygen. For example, a coating consisting of magnesium fluoride
and zinc sulfide can degrade over a period time at an elevated temperature (> 100°C)
FILTER PERFORMANCE MEASUREMENTS 169

in air. Several chemical reactions can occur, including the conversion of the sulfur,
zinc, and magnesium to their respective oxides. Some of the reaction products are
volatile, and this obviously leads to a change in filter performance. Conversion of
a sulfide or a fluoride to an oxide is generally accompanied by a change in refractive
index, and a corresponding change in filter performance. Similarly, oxide formation
can occur with other compounds under appropriate conditions.
Some molecules, for example, titanium dioxide, can lose a portion of their
oxygen complement at elevated temperatures. A chemically stable suboxide may
form that is stable and that does not return to the fully oxidized stoichiometric state
following removal of the heat source. This may occur because the suboxide is
intrinsically stable or because the oxygen that was released diffused out of the
coating, or was driven out of the thin film structure by a pressure gradient. As the
temperature is reduced, the oxygen does not reenter the coating at a sufficient rate
during the cooling off period. Once the system is back at room temperature, the
reaction kinetics are insufficiently rapid to cause further oxidation to take place.

5.3.2.2. Agglomeration. In metal mirrors, notably silver ones, hillocks can

form when the film is heated in the presence of oxygen (Presland et al., 1972).
The hillocks scatter light, thus destroying the high specular reflector properties of
the surface, while at the same time adding to the veiling glare in the system. If
optically acceptable, overcoats can decrease the tendency for hillock formation and
at the same time improve the overall environmental stability of the design.

5.3.2.3. InterditTusion. The interdiffusion of adjacent thin film layers is

another phenomenon caused by high temperatures. This effect is not ordinarily
observed in films kept at room temperature. Interdiffusion consists of the inter-
mingling of the atoms or molecules of the materials in adjacent layers. A stack
composed of two or more different materials can be degraded as the two materials
interdiffuse. It is also possible for the diffusion to take place between the stack and
the substrate. In the ultimate situation, the coating changes from a system of
alternating homogeneous, discrete layers, to a homogeneous and monolithic single
layer. At intermediate stages, the coating appears as a single inhomogeneous mass.
The optical properties change and the filter no longer functions as intended. On the
bright side of this phenomenon, a small amount interdiffusion at layer interfaces
lowers the stress levels and improves the interlayer adhesion in marginal cases.
It is difficult to predict how quickly interdiffusion effects degrade a coating.
The process depends on the materials, the temperature, the time at temperature,
and other factors. If such a situation in the usage of the filter is anticipated, it should
be brought to the attention of the supplier so that he or she can take appropriate
action to evaluate whether a problem exists or not, and to suggest appropriate tests
that may be perfonned to verify that indeed the problem does not exist in the
delivered product. In addition, the manufacturer's choice of materials could be
influenced by their tendency (or lack thereof) to interdiffuse if this were indicated
to be a potential problem.

170 CHAPTER 5

5.3.3. Biological Effects

Because most filters are constructed from inorganic materials, the general rule is
that one need not be concerned with biological degradation. This is correct in most
instances. In some situations, however, the degradation mode of a filter can be
biological, specifically fungal, in nature. Fungi (mold and mildew) live on just
about any surface, including optical filters. Their secretions can cause etching of
the surface, and eventually, complete failure of the coating. In an operating en-
vironment that is hot and damp, the optical surfaces should perhaps be treated with
an antifungal agent. Once again, the filter supplier should be apprised of the en-
vironment so that some suggestions can be made as to what antifugal treatments
the coatings may and may not survive.

5.4. COSMETIC AND OTHER VISUAL DEFICIENCIES

In this section, we discuss features that may be seen on the surface of an interference
filter. Some defects may be apparent only when a part is examined under the proper
lighting conditions. Other features in a coating cannot be detected with standard
instrumentation, such as spectrophotometers, yet they are readily apparent with the
naked eye. In either case, the fact that a surface defect can be observed does not
necessarily mean that the part is unsatisfactory or otherwise inadequate to serve the
intended function. It indicates that the human eye is very sensitive to certain types
of optical patterns and color changes, and these mayor may not be relevant to the
inspection of a given sample. We give examples of some of these situations.

5.4.1. Sleeks, Digs, and Scratches

A sleek is a very small scratch whose width cannot be resolved with the naked eye.
By comparison to a sleek, a scratch is relatively large. The United States military
specification designated Mil-O-13830, revision L, specifies how to evaluate a scratch
and give it a numeric value. The number given to a scratch is 10 times the width
of the scratch in micrometers (more or less). These dimensions are not precise and
are more a rule of thumb for evaluating the cosmetic appearance of a scratch (Young,
1985, 1986; Young et aI., 1985).
A dig is a scratch whose length is equal to its width, and it is evaluated in the
same way as a scratch.
The design of the inspection station used to examine parts for sleeks, digs,
and scratches is very important in determining the level of sensitivity that can be
attained. The arrangement of the illumination and a dark background influences the
visibility of the light scattered from the defect. A dark background reduces the light
reflected by a high reflector, while it provides a dark background against which to
view an antireflecting coating.
Scratches and digs may originate from a number of causes.
1. Improperly prepared substrates: The raw material for the substrate needs to
be of high quality, homogeneous, and bubble-free. The optical surface must be
properly ground and polished, or otherwise given a smooth surface.
FILTER PERFORMANCE MEASUREMENTS 171

2. Improper handling: Substrates are subjected to a large number of handling

operations, including incoming quality inspections, cleaning, loading into and un-
loading from the coating chamber, and final inspection and packaging. Unless
carefully done, each of these steps may degrade a part.
Tooling is typically configured so that the parts are supported around the periphery
of the coated area. Improper care in tooling design, poor fabrication of the tooling,
or differential thermal expansion between the tooling and the substrate material may
all lead to scratching of the substrates during coating.
In this discussion, we ignore the obvious problems that can occur in improper
substrate handling and packaging.

5.4.2. Blotches and Stains

Coated surfaces sometimes have areas in which the reflected color of the coating
is different from other areas. These have a number of names including discolora-
tions, blotches, or stains. Generally, these are quite faint and can be seen only in
reflection and against a dark background. Even when the stain is quite severe and
is very apparent to the eye, it is frequently very difficult to measure the difference
between the two areas with a spectrophotometer. The reason this is so difficult is
that the human eye is extremely sensitive to color differences in adjacent areas.
The typical stain has quite sharp edges to the discolored area, thus making it easy
to see the stain. From a functional point of view, such a stain will typically not
degrade the optical performance of the filter, though it might be an indication of
the care taken when the substrate was cleaned. On the other hand, it should be
noted that there are some optical glasses that inherently stain easily.
The appearance of a stain is sometimes observed only after a part has been
coated. Even an opaque coating does not always hide a substrate defect. Stains that
are invisible on the bare substrate affect the nucleation and growth of the film,
causing some areas to become delineated and apparent only after coating.

5.4.3. Replication of Substrate Defects

A thin film will not bury defects that exist in the substrate prior to its being coated.
In fact, in many instances, the coating exacerbates the situation as substrate defects
perturb the otherwise planar process of nucleation and growth. Inevitably, the
amount of light scatter increases after the surface is coated.

5.4.3.1. High Reflectors with Opaque Metal Layer. An opaque metal layer
often replicates and enhances substrate defects. An improperly cleaned substrate
causes the metal layer to be hazy. If certain areas are dirty, whereas others are
clean, then some areas will have a highly specular reflectance, whereas other areas
will be much more diffusely reflecting. These areas of differing specular and diffuse
reflectance are readily apparent to the naked eye.

5.4.3.2. Low Reflectance Surfaces. Surfaces that have a low reflectance

readily expose scattering caused by defects. Because they reflect little light, the
172 CHAPTER 5

ratio of scattered intensity to reflected background intensity can be quite high for
these surfaces.
Scattering from low reflectance transparent substrates may be observed by
viewing a brightly illuminated surface against a dark background. Any scattered
light becomes a bright spot on a dark background, and this is readily discerned.
Any dark, opaque substrate is its own dark background, and it is only necessary
to illuminate the surface with a bright light.
A fundamental axiom in the construction of thin film filters is that a thin film
replicates the contour of the surface being coated. This varies in degree for the
various coating materials. For example, a screen of wire mesh with a spatial period
of approximately 2 mm was coated with 2 to 3 cm of aluminum oxide over a long
period of time. The pattern on the surface of the deposit exactly matched that of
the underside where the screen was located. This was a vivid demonstration of the
extent of replication that is possible. The structural morphology in many growing
thin film layers can explain this kind of behavior. The growth is generally columnar,
with the columns perpendicular to the surface of the substrate if the evaporant is
also arriving normally. If the evaporant is arriving at some nonnormal angle, the
angle of growth follows a rule that states that the tangent of the angle of growth
is one-half the tangent of the angle of material arrival (Dirks and Leamy, 1977;
Pulker, 1984, p. 327)
It is possible that a coating can mask surface scratches and sleeks. A germanium
substrate with polishing sleeks was coated with an antireflection coating containing
zinc sulfide. Microscopic examination of the coated surface showed no signs of the
sleeks, yet they were readily visible on the substrate surface. Similarly, electron
microscopy has revealed that some coating materials can smooth out the roughness
of previously deposited layers.

5.4.4. Coating Voids and Other Defects

Coating voids, called pinholes when they are small, are areas ofthe substrate where
there is little or no coating. A void may extend all the way to the substrate, or may
only involve a limited number of layers. These defects have a number of causes.
Particulate matter on the substrate when the coating process starts can cause
poor adhesion in a small area. The coating can suffer local adhesive failure as the
inevitable stresses build up, either during the vacuum coating process, after the
coating is complete, or during the cool down and vacuum chamber venting to
atmospheric pressure. The result is an area without its full complement of the
coating.
Another source of particulate material is the coating source itself. For an
evaporation source, the particles can come from the source, having been ejected
from the molten pool of material. This phenomenon is frequently the reason a
manufacturer chooses one evaporation source type, e.g., electron gun source or
covered or uncovered resistance source, over another.
In sputtering, a momentary arc between the target and some charged surface
or the plasma can cause material to be blown off the target in a small area. This
material can travel in all directions, so that sputtering with the substrate above the
target will not eliminate this problem. When sputtering takes place with the substrate
FILTER PERFORMANCE MEASUREMENTS 173

below the target, the heavy deposits that build up around the target can flake off
and fall on the substrate. This configuration also is subject to contamination of the
substrate from material from any of the walls of the vacuum chamber. As stresses
build up, the materials on the wall may snap off with enough lateral velocity to
reach the substrate.
Typically, pinholes occupy a relatively small area of the total coated surface.
They are frequently apparent because they have a large contrast ratio with the
adjacent coated areas. The significance of pinholes in a filter is highly dependent
on the application. In some, voids are functionally insignificant, whereas in other
instances, they are intolerable. For example, in a coating that will be etched with
a fine pattern, a void could obliterate a part of the pattern, or cause a line to have
a break in it, perhaps causing a portion of a large circuit to fail. Small openings
in a blocking filter, which must have a great deal of attenuation, can allow an
objectionable amount of undesired radiation to pass through the filter.
The observation of pinholes can be a tricky affair, especially when large areas
need to be examined. Visual examination of a small coated area with a microscope
can readily detect a void, either by the difference in intensity or by a difference in
color between the defect area and the normal coating. If the substrate is transparent
and the coating is opaque, the light can be incident from one side of the coating
and the detector can observe from the opposite side. If the coating is an antireflection
coating, then one can search for the pinholes by looking at the reflected light. A
difficult situation is encountered when the coating is semitransparent, as happens
in a single film of indium tin oxide on a glass substrate. Then the contrast between
the coated areas and the pinholes is poor in both reflected and transmitted light.
Thus, the scale-up from scanning a small sample with a microscope to scanning a
large substrate surface area for pinholes is a real challenge.
Inclusions are trapped particulates. They can be caused by particles that are
ejected from the coating material source or from dust or other particles that are
incorporated into the film as it grows. Figure 5.7 shows a scanning electron mi-
croscope photograph of an infrared film with a particle that initiated a conical defect
zone. As the film grew, the disturbed area increased in size, and the amplitude of
the disturbance remained unchanged. The particle continues to make its presence
felt even after it has been buried to a depth many times its diameter. The size of
the disturbance increased in diameter so that the slope error became less, but the
overall volume of the disturbance increased. If the particle detached from the
underlying material, either due to the buildup of stresses in the coating with increased
thickness, or perhaps due to differential expansion from the heating of the film
from the coating source, a void many times the diameter of the particle will form.
Even if the inclusion has the same index of refraction as the adjacent material, a
scattering site is formed because of the displacement it causes to the other layers.

5.4.5. Cleaning Coated Parts

Thin film filters require occasional cleaning to remove dust and other environmental
dirt and grime that may deposit on them. Cleaning them can be simple or difficult,
depending on the type of coating and on the substrate. For example, unprotected
metal layers, such as silver, gold, or aluminum, can be cleaned only with great

174 CHAPTER 5

Figure 5.7. Scanning electron microscope (SEM) photo of a conical defect arising from a
point defect. The white bar is 10 ILm long. As more layers are added to the stack, the
disturbance caused by the defect increases in lateral extent, but the defect height remains
constant. (Courtesy of Phil Swab of Research Dept., O.C.L.I.)

difficulty because the metal surface is soft and easily damaged. On the other hand,
a hard coating of magnesium fluoride on a glass lens is nearly as durable as the
bare substrate. One would not attempt to clean a finely polished glass lens with
ordinary techniques such as might be used on a window pane, and a similar caution
is warranted for coated optics.
Loosely attached particles can be removed from a coated surface by blowing
them off with a gas jet. This is described in the section of cleaning soft coatings.

5.4.5.1. Hard Substrates-Hard Coatings. For our purposes here, a hard

substrate is one with a hardness similar to that of glass. When both the coating as
well as the substrate are hard, it is straightforward to proceed to clean the coated
surface with a solvent and a lens tissue. It is important that no abrasive material
be picked up by the tissue and dragged across the surface, otherwise a scratch or
sleek will result. The solvent could be a soap or detergent solution, an organic
solvent such as isopropanol, or some commercially prepared mixture. It is assumed
here that the coating is indeed inorganic, deposited in some vacuum process, and
is water insoluble. Strong acids, such as nitric and hydrofluoric, and strong bases
should be avoided as they can etch some of the thin film layers as well as damage
FILTER PERFORMANCE MEASUREMENTS 175

some of the optical components. Chlorinated hydrocarbon solvents should be avoided

when metallic layers, especially silver, are present in the coating because the
chlorine can attack them.
The solvent or other cleaning solution should not be allowed to dry on the
surface. There are a number of reasons for this:
1. The solution may have picked up and suspended or dissolved contaminants
that were on the surface. Upon evaporation of the solvent, the contaminants will
be concentrated into the remaining droplets. Thus, the residue from the entire surface
will be concentrated into a few dirty spots on an otherwise very clean surface.
2. For organic solvents that are initially water-free, the evaporation of the
solvent from the surface being cleaned will reduce the temperature locally. This
may cause the moisture in the air to condense in the vicinity of the droplet. Finally,
as the water dries, a mark will be left at the site.
3. If an organic cleaning solvent has been exposed to the atmosphere for a
period of time, it can absorb some of the atmospheric moisture. This contaminates
the solvent; the water will be left behind on the surface when the solvent is used
for cleaning, again causing a water mark.
There are several ways of overcoming the staining problems just mentioned:
1. Do not allow the solvent to dry on the surface. A jet of filtered, dry nitrogen
can be used to blow the solvent off the surface before it has a chance to dry or
evaporate.
2. A clean paper tissue dragged over the surface will absorb and remove the
solvent, so that no droplets are left behind on the critical surface.
3. The solvent selected should have a low volatility so that it will not readily
evaporate. Fluorinated hydrocarbons are one class of such liquids. These solvents
will not dry readily on the surface, nor will they cool the surface and induce moisture
to condense on it. When the cleaning process is complete, the solvent can be removed
by any number of ways, including those mentioned above.
Although the preceding remarks are directed at coated optics, they are applicable
to uncoated surfaces as well. The treatment of substrates before coating is important
to the coating manufacturer. Water marks are just about impossible to see before
a surface is coated. After the coating has been deposited on the surface, however,
the stained areas are readily seen. The coating materials nucleate differently in the
stained areas, and this leads to different properties of the coating. In tum, these
differences lead to the scattering of light, and this is usually a cause for concern.

5.4.5.2. Hard Coatings on Soft Substrates, and Soft Coatings. Although

it is possible to apply a hard coating onto a soft substrate, the lack of a solid support
for the hard coating will cause it to crack and fail when a sharp point force is
applied to it. Short of a self-supporting film, attempts to harden completely a soft
surface are doomed to failure ("You can't armor plate a marshmallow" is a classic
phrase in the thin film industry).
Soft coatings are problematic regardless of whether the substrate is hard or
soft.
176 CHAPTER 5

The ultimate soft coatings are probably the metallic coatings with no overcoats,
such as gold, silver, and, to a lesser extent, aluminum. As previously mentioned,
there is no way that these can be cleaned with a contacting method. Particles can
be blown off the surface with a gas jet, either manually supplied as from a rubber
bulb, or from some other source such as a compressedlliquefied gas in a can, a gas
cylinder, or a gas manifold. When using any gas source to blow off a surface, it
is necessary to ensure that the gas itself is clean and does not contaminate the
surface. If a rubber bulb is used, the bulb should be inspected to ensure that it is
clean. For liquefied gases in aerosol cans, it is difficldt to ensure that none of the
liquid itself comes out of the nozzle; if this happens, it will cause spotting on the
surface. Finally, when the gas comes through pipes and hoses, either from a com-
pressed gas cylinder or from a central gas supply, it is essential to have a final filter
near the nozzle to remove particles that may be in the gas lines. Filters are available
that can remove particles as small as 1 ~m in size.
While blowing off a surface, care should be exercised so as not to touch the
surface with the nozzle. This is particularly important with a manually operated
rubber bulb, as motion is required to squeeze the bulb, and this could result in a
movement of the nozzle toward the surface.
A surface charge of static electricity attracts and retains dust. If the relative
humidity is low, the surfaces can acquire and hold significant static charges from
the cleaning process. The neutralization of surface charge reduces the amount of
dust that collects on a surface. The presence of ions in the environment accomplishes
this, as does a moderate to high ambient humidity level. An accessory that is useful
in areas of low humidity is an electrostatic ion generator. It can be installed near
the tip of the nozzle used to direct the gas that blows off a surface. The ion generator
ionizes some of the gas Howing through the nozzle, and these in tum neutralize
the surface charge. In another form, the generator can be part of the room air
filtration system and can flood an entire room with ions.
A tissue with a solvent can be used to clean surfaces. This is a contact method,
and can remove particles that are more adherent than those that are removed with
a simpl~ gas blast. The force of gravity and the liquid forces are the only ones
acting between the tissue and the surface. The forces applied to the surface by the
wet tissue are small, and therefore, this technique will in general suffice for plastic
and other fragile surfaces. The wet tissue is dragged in overlapping paths over the
entire surface without applying a normal force. The solvent should be removed
before it has a chance to dry (Stowers and Patton, 1978).
Perhaps the ultimate challenge for cleaning substrates are the soluble materials
such as sodium and potassium chloride. Because of their nature, they must be
handled in a room in which the humidity is kept at a low value. In addition, these
materials are excellent insulators, and so acquire and hold an electrostatic charge
very easily.

5.4.6. Coating Thickness Uniformity

In most instances, one desires a coating whose performance is uniform over the
entire useful aperture. The need for thickness uniformity arises from the fact that
FILTER PERFORMANCE MEASUREMENTS

optical filters are generally sensitive to the thicknesses of most of the layers that
comprise it. On flat substrates, this translates to a coating in which the thickness
of all layers is uniform over the useful aperture.
In the simplest situation of a small flat part, the color of the light reflected by
the coating can give a qualitative indication of the variation in the thickness of a
film on a surface. It is fairly easy to measure the reflectance or transmission of flat
parts with a spectrophotometer to determine quantitatively the thickness uniformity
of a coating. On the other hand, if the part has a curved surface, such as a lens or
a dome might have, then a typical spectrophotometer cannot be used to get a good
photometric value.
Good photometry is not necessary in some cases to determine to first order
the uniformity of a coating system. For instance, if the performance of a design
has a feature with a steeply rising slope, the wavelength position of the half max-
imum point can be plotted as a function of position. This technique assumes that
the shape of the curve remains constant as the thickness changes. This might not
be the case if the ratio of the thicknesses of the materials that are used in the coating
changes. This is detectable by comparing the shapes of all of the curves measured
to verify that indeed they are identical except for a wavelength shift.
Some instruments are capable of measuring accurately the reflectance of sur-
faces with optical power (see section 5.1. 8.). They are very useful in measuring
curved surfaces with unknown coatings and for which there are no flat witness
pieces. Of course, special purpose instruments can also be built.
On complex shapes with steep surfaces, such as fast lenses or domes, the
analysis of uniformity is more complicated than that of flats because the angle of
incidence varies as a function of position over the surface. To be more quantitative,
one needs to measure the reflectance (or transmittance) at several points on the
surface at the appropriate angle of incidence. The angle of incidence should not be
neglected on steep parts as this can have a significant effect on the performance
(see section 1.3.). The uniformity is sometimes deliberately adjusted to compensate
for the angle shift on large steep-surfaced optics.
When the parts are small enough to be placed into the sample compartment
of a spectrophotometer, the task of measuring a number of points is limited by the
beam size that gives adequate signal to noise ratio (assuming the focusing problem
is somehow solved).
When the parts are too large to fit in an instrument, the manufacturer has three
choices.
1. Cut the part into pieces small enough to fit on the equipment available
2. Design and build larger measurement equipment that can handle the large
part
3. Install small witness pieces in the coating equipment to stimulate a larger
part, and measure each witness individually after it has been coated .
The assumption is generally made that the witness pieces will fairly represent the
coating that is placed on the actual part.
When many small parts are being coated at one time, the uniformity over an
individual part may be quite good. On the other hand, the thickness may vary from
178 CHAPTER 5

part to part within a single lot, as the various parts sample different portion of the
unifonnity curve.

5.4.7. Microscopic Examination

Microscopic examination is frequently a very revealing technique for evaluating
coating-related problems. There are several reasons why this technique is so val-
uable:
1. Many coating failures can be traced to microscopic phenomena
2. The eye can analyze patterns and features that may give a clue to the
origin of the difficulty
3. Being sensitive to spatial color changes, the eye can detect subtle clues
in the coating
4. The instant feedback of being able to observe the coating first-hand can
lead to successes that cannot be realized in any other way
For example, the reason a coating is scattering might be identified in a matter of
moments with a microscope: stress cracking can be seen and the type of stress
identified, or perhaps there are numerous defects that are causing the problem. We
discuss these examinations below.
The microscope is an excellent method for investigating adhesive failures in
coatings. Microscopic color variations in the coating will frequently indicate whether
the failure has occurred at the film/substrate interface, or within the film stack itself.
If the failure occurred at the substrate interface, the problem may be one of substrate
cleaning, whereas a failure within a stack could indicate that there has been a
buildup of stresses to the point where the stress overcame the adhesive forces.
Another possibility is that some process abnormality, for example, a leak in a
vacuum chamber occurred during the coating run, and this had the effect of weak-
ening the adhesive strength at that point in the design.
The microscopic patterns observed with tensile stress failures are radically
different from those obtained with compressive stress failures. In compressive
failure, the film delaminates from the substructure and buckles, as shown in Fig.
5.8. In tensile failures, the film also fails in a distinctive pattern: a series of cracks,
as shown in Fig. 5.9. The key to the interpretation of this figure is that the lines
only intersect at right angles. Although some of the lines may be somewhat parallel,
at the points of intersection, the two cracks meet at right angles. The explanation
for this phenomenon is that the crack relieves the stress locally in the direction
perpendicular to the crack direction. The stress vectors that are oriented parallel to
the crack are, to first order, unchanged by the rupturing of the film. The further
one moves away from the crack the greater the stresses in all directions. Thus, two
cracks can propagate in roughly parallel directions. As one of the cracks turns to
intersect its neighbor, it finds the stresses in its neighbor's vicinity to be closer and
closer to a one-dimensional stress. At the edge of the neighboring cracks, the stress
is entirely one dimensional, therefore, the new crack intersects at right angles to
the previous one. Depending on the level of the stresses and the amount of relief
that is afforded by the cracking, the point at which the new (second) crack turns

FILTER PERFORMANCE MEASUREMENTS 179

Figure 5.S. Photo of compressive failure as seen through an optical microscope. The bar
is 100 j.l.m long. The film has delaminated from the substructure and buckled upwards.
(Courtesy of T. Tuttle Hart of Research Dept., O.C.LI.)

to align itself in a perpendicular direction to the old (first) crack is a measure of

the stress level remaining in the coating. The higher the stress level, the more
cracks that form in the film. The orientation of the cracks also indicates whether
there is a preferred stress direction in the coating.
A microscope can also help determine the type and the source of dirt and
debris on or in a coating. These particles can be located in three places:
1. Under the coating, which means that they were there before the coating
was applied
2. In the coating, which means that they were a result of the coating process
3. On the surface of the coating
The particles in the third category can usually be removed by carefully cleaning
the coating, as outlined in a previous section. Microscopic examination reveals
what needs to be removed, and what is left after the cleaning process is complete.
Of course, when such work is being done, a clean bench or other clean area should
be used to ensure that extraneous contaminants are not confusing the issue.
Lint and other loose surface dirt are readily apparent under microscopic ex-
amination, as are particles imbedded in the film by source spatter during the coating
process. Pinholes can be identified in a transmitting filter by alternating between
reflected and transmitted illumination: the pinhole area changes color or brightness
with the change in lighting condition. A pinhole in an opaque filter on a transparent

180 CHAPTER 5

figure 5.9. Photo of tensile failure as seen through an optical microscope. The bar is
100 ILfTl long. Note that at the intersection of all lines, the two cracks meet at right angles.
(Courtesy of T. Tuttle Hart of Research Dept., O.C.Li.)

substrate appears bright when examined in transmitted light. A pinhole in a filter

on an opaque substrate may be difficult to identify unless the coating is highly
colored and the area of the hole is dark.
If areas of the coating have failed and the failed material has been removed,
it is frequently possible to determine at just what level in the coating the problem
occurred as long as there are not too many layers in the stack. The interference
colors of the coating change when layers are removed. This is true in visible and
infrared designs. Colors can be seen in infrared designs, as the higher order effects
in these coatings frequently give rise to spectacular colors in the microscope. It is
a simple matter to count the different colors to see which layer from the top was
the last layer left intact. Ultraviolet coatings, and other designs that have very thin
layers, may not produce the vivid colors of coatings that are designed for the longer
wavelengths, but one can still recognize the individual layers by the edges where
the layers are discontinuous.
A number of features simplifies using a microscope for the examination of
coated specimens. First of all, the objectives should be of the metallurgical type.
Biological objectives are designed to function in conjunction with a cover glass
and the specimen immersed in an index matching liquid. If the cover glass is not
present, the image suffers from spherical aberration.
For the proper investigation of thin film coatings, both transmitted and reflected
images need to be examined. It is helpful to have a microscope with an illumination
FILTER PERFORMANCE MEASUREMENTS 181

system in which the light can be made to be incident on the substrate through the
objective. Such illumination systems are standard on metallurgical microscopes.
Another useful microscope technique in the investigation of thin films is the
use of dark field illumination. The objective illuminates the sample with a cone of
light at relatively high angles of incidence. The specularly transmitted or reflected
light cannot be collected by the objective as it is outside of its collection cone.
Only the light scattered into the objective is seen. Features that scatter light such
as dust, voids, edges, and other discontinuities are readily seen, whereas smooth
features of the specimen appear dark. The typical dark field work is done in re-
flection, and this requires the use of special objectives as a well as special microscope
optical system.
An interference or phase contrast, also called Nomarski, attachment to a mi-
croscope is a useful technique for investigating the surface contour of thin films.
The method uses the interference of light to magnify the vertical features on the
surface of the film. Features at different heights are seen as different colors or
shades. Although it is difficult to quantify the extent of the roughness, one can get
a good qualitative feeling for the extent of a problem. This technique produces
colorful displays that can be easily evaluated by eye, but as the effects are caused
by a gradation of colors, it is difficult to place an exact value on the height of a
feature.

5.5. INSPECTION LOT SIZE

When the quantity of coated parts is small, it is not a major effort to scan or
otherwise measure every part. When many parts are being produced either in one
or multiple coating runs, it becomes prohibitively expensive to measure all parts.
A typical approach in these cases is to qualify a vendor by measuring a large fraction
of the initial lots of a large order. As the buyer and seller gain confidence in each
other and in the fact that the process is under control, the requirements for testing
can be reduced to a small sample of each coating run.
When 100 percent inspection is required and the quantities are large, it is
important to automate the measurement process to reduce costs.
 Chapter 6

Substrates and Materials

The type of substrate upon which a filter is deposited generally determines many
of the characteristics of the filter, such as durability, stability, and cost. There are
three major categories of substrates:
1. Glasses and similar crystalline materials
2. Organics and plastics
3. Metals
Substrates can also be classified as either rigid or flexible. The majority of high-
quality optics are fabricated of rigid materials. A large volume of web-coated
substrate is produced each year for decorative and other uses.
In this chapter, we discuss in more detail the influences the various substrate
materials and fonns can have on thin film coatings.

6.1. SUBSTRATES IN GENERAL

The primary driving factor in the selection of a specific material as a substrate for
a thin film coating is the wavelength region over which it must function. Trans-
parency may be the major concern, as coatings can modify only the surface-related

182
SUBSTRATES AND MATERIALS 183

characteristics of the material, such as the reflectance; they cannot alter the internal
transmittance or absorptance of a substrate. Table 6.1 lists the ranges over which
some common optical materials are transparent. This table should be used with
care; the limits shown in it are not absolute as the absorption is a function of the
thickness of the material being considered. The absorption frequently increases
gradually as the wavelength approaches the limits. If the sample is thin enough, it
if possible that the transmittance of a substrate outside the spectral range given in
Table 6.1 may be adequate for some applications.
Substrates should be strong enough to survive the handling that they will
undergo during the coating process. In a typical evaporation process, fixtures support
the parts by contacting them on the coated side only at the edges or outside of the
useful aperture. The fixtures are designed so that they will constrain the substrate
from moving laterally. Typical tooling consists of a milled recess in a metal rack
in which the part is situated as shown in Fig. 6.1. The supporting protrusion, the
jig lip, is usually made as small as practicable to minimize the area of the coated
substrate that is shadowed by it and is, therefore, not coated. This results in an
uncoated area around the periphery of the element. If the coating must extend to
all edges of the surface, then the fabrication process must include steps that coat

Table 6.1 Common Substrate Materials Used for Optical

Filters
Long-
Short wave
wave limit
Material limit (ILm) (ILm)
Si02 0.12 4.5
AI~3 0.14 6.5
MgF2 0.11 7.5
CaF2 0.13 12.0
GaAs 1.0 15.0
Si 1.2 15.0
BaF2 0.25 15.0
Ge 1.8 23.0
C (diamond) 0.25 80.0+
ZnSe (eVD)· 0.5 22.0
IRTRAN - 1 (MgF2) 0.5 7.5
IRTRAN - 2 (ZnS) 0.4 14.5
IRTRAN - 3 (CaF2) 0.13t 11.5
IRTRAN - 4 (ZnSe) 0.5 22.0
IRTRAN - 5 (MgO) 0.23t 9.0
IRTRAN - 6 (CdTe) 0.9 33.0
NoI9s: Transmission limits include surface rellections and are for a T ~ 10 percem level
and for a sample thickness 012 mm. IRTRAN data from Kodak Publication U-72, unless DlhelWise
indicated. IRTRAN is an Eastman Kodak Company trademark.
'CVD Incorporated brochure.
tLichlenstein, 1979.
From Wolle, 1965, p. 327.

184 CHAPTER 6

Substrate

~\\\\\\\\\\\\\\\\\~
r r rrr Motal Flxturo

Coating Material Flux

Figure 6.1. Cross-section of a typical coating fixture supporting a substrate. The substrate
is held in place by protrusions from a rack that usually rotates so that the substrate is coated
with a uniform deposit.

an oversize part that is subsequently trimmed to size. In a sputtering process in

which the part is supported above a source, a similar uncoated area exists on the
substrates. On the other hand, some sputtering systems allow the substrate to be
below the source, thus making the supporting structure very simple and allowing
for complete coverage of the surface. Substrates are frequently covered to protect
those surfaces that are not to be coated.
Coatings can be deposited on substrates that are at ambient or cooler temper-
atures. Elevated process temperatures, however, usually lead to harder, environ-
mentally more durable filters. The range of process temperatures used is frequently
100° to 200° (or higher). A number of factors related to the substrate limit the upper
temperature that can be used: (l) glass transition temperature, (2) recrystallization
temperature, or (3) melting temperature.
If one of these temperatures is exceeded, the substrate may be degraded by
the coating process.
The amount of time required to heat and cool a part is a concern in a production
environment. In general, the cooling rate is more critical than the heating rate for
glasses because the cooling process places the surface in tension. In a vacuum
coating chamber with multiple rotational motions, the only way to effect heat transfer
to the substrates is by radiation, and it is difficult to accelerate the process effectively.
The challenge for the coating supplier is to adjust the process so as not to damage
the substrate.
Most optical coatings in the past have been applied to rigid substrates, and
these have usually been of a glassy nature. Recent advances have enabled the
application of optical coatings of interference quality to nonrigid substrates. In
particular, webs of sheet materials have been coated with optical grade metal and
dielectric layers to make interference filters. The substrate materials can be metal
foils, such as aluminum and stainless steel, or organic polymers, such as polyeth-
ylene terephthalate (PET) and fluorinated hydrocarbons. Glass webs are not flexible
enough to be run through these coating machines at this time.

SUBSTRATES AND MATERIALS 185

6.2. SUBSTRATES TYPES

6.2.1. Glasses

Glasses are the most common substrate materials for thin film filters. One reason
for this is that most optical elements are glass, though plastic lenses are fast becoming
popular in high volume units such as cameras. In both low and high-quality products,
most optical elements are coated to reduce surface reflectance losses, ghost images,
and veiling glare. We also include in the category of glasses some crystal-
line materials, such as silicon, germanium, and calcium fluoride, which, from a
coating standpoint, behave quite similarly to glasses. These crystalline materials
can be heated to typical coating temperatures with no degradation of their optical
properties.
The optical designer has a number of trade-offs to consider when selecting
optical glasses. In addition, some thought needs to be given to how these elements
will be cleaned and coated once fabricated by an optical shop.
The substrate material preferably should not darken under ultraviolet, electron,
or ionized particle irradiation, as a number of coating processes generate such
conditions in the vacuum chamber. In evaporative coating processes, a glow dis-
charge cleaning step is often used, and the glow plasma in which the substrate is
bathed generates ultraviolet light as well as energetic ionized particles and electrons.
A similar plasma may exist during the entire sputtering coating process. If materials
must be used that have poor radiation resistance, the filter manufacturer can tailor
the coating process as necessary.
Polished substrates need to be cleaned at least twice before they are coated.
The optician who polishes the surface must remove the polishing compound im-
mediately after the last polishing step, lest a residue remain on the surface. A similar
cleaning step is required immediately after diamond-turned optics are completed.
Otherwise, a tenacious film of the polishing compound or the diamond turning
lubricant forms, and it cannot be removed at a later time (Buckmelter et al., 1975).
This is true of both conventionally polished and of single point diamond turned
(SPDT) surfaces. The layer that is left behind can then be removed only by additional
polishing. Thus, it is necessary for the optical shop that is preparing the surface to
be certain that they clean off the surface as best they can and as soon after they
are finished with the optical element. Because they are familiar with the properties
of the materials used in fabricating the element, they are in a good position to know
how to remove this surface contaminant. Also, these materials become more ad-
herent as they dry out and age on the surface; thus, time is of the essence in their
removal.
Surface cleanliness is a matter of vital importance. Organic and other residues
from the surface finishing process can leave a layer on the surface. This residue
can act as a release layer for the coating, reducing the adhesive forces between the
substrate and the coating, and possibly causing the coating to separate from the
substrate.
186 CHAPTER 6

A final cleaning of the substrate, done shortly before coating, prevents as much
as possible:
1. Contaminants' reducing the adhesion between the first layer and the sub-
strate; and,
2. Particles' remaining on the surface and leading to scatter and pinholes in
the coating.
Other symptoms of poor cleaning are streaks and wipe marks on the surface.
Unfortunately, these may become visible only after coating, necessitating removal
of the coating, perhaps by repolishing, and then recoating. Thus, it is vitally
important to have good communication between the user and the coating vendor.
The latter must understand the history of the surface so that it may be cleaned
properly.
Some glasses stain more easily than others (see Schott and other glass company
catalogs for these ratings). It is difficult to clean some of these glasses as aggressively
as one would like, lest the cleaning process succeed in cleaning off the contaminants
but stain the glass at the same time. On the other hand, the coating may well
improve the durability of a surface to further degradation from staining by protecting
the glass from the environment. This improvement in resistance to staining cannot
be guaranteed, but it is to be expected that the additional surface material reduces
the substrate's exposure to the active elements in the environment and thereby
protects the surface.
The Schott glass catalog recommends that their glasses not be heated above a
temperature which is 200°C below the glass' transition temperature (Tg - 2(0).
At higher temperatures, the optical figure may deform and the refractive index may
change. Table 6.2 lists some of the glass transition temperatures for selected glass

Table 6.2 Approximate LImit

Temperatures for Various Substrate
Materials
Material Temperature (OC)
Plastic (acrylic) 90-105§'#
Plastic (polycarbonate) 150§'#
Plastic (polyimide) 310-365§'#
Plastic (polyester) 255§'#
BK7 365t
Germanium 936-
BaF2 1280-
CaF2 1360-
silicon 1420-
8i02 (fused Silica) 1710t
AI~3 2030-
Notes: ·Melling temperature. tSoftening temperature. *Max-
imum safe lempefature (Tg - 200).
§Glass transition temperature (Tg ).
#Agranofl (1985).
From WaHe, 1965, p. 330.
SUBSTRATES AND MATERIALS 187

types. It should be noted from this table that there are some glasses for which even
a 200°C coating temperature may be marginal.
One of the considerations the coating manufacturer has to take to account is
the expansion coefficient of the substrate material. If the coating process involves
heating, enough room must be allowed for the expansion of the substrate in the
supporting fixtures. The thermal expansion of the tooling and the substrates must
be matched through the thermal excursion of the coating process. Table 6.3 gives
the expansion coefficients of some typical substrates and tooling materials.
The fabrication of glass substrates into optical components is an art that has
been developed over many centuries, and we will not go into those details here.
We touch upon the dimensional sizing of substrates, however, as this is frequently
something the coating manufacturer is called upon to do.
Optically flat glass sheets can be made flat in two ways:
1. The most expensive of these is the traditional grinding and polishing as
done in an optical shop but on a larger scale.
2. Flat glass can be produced directly from the liquid phase by floating it out
on a pool of molten tin. The liquid tin/liquid glass interface is very flat, as is the
glass/air interface. The two surfaces have slightly different physical and chemical
properties, and these differences may be important in some applications. The tin
and air sides of a float glass sample can be identified by directing the light from a
lamp that emits in the near-UV on the float glass sample: the tin side fluoresces
while the air side doesn't.
Flat glass can be cut to size by scribing, either manually or with a machine, a line
on it with a carbide or a diamond tool and then breaking the glass along the line.
The accuracy that can be obtained with this technique is on the order of + / - 125

Table 6.3 Expansion

Coefficients of Commonly Used
Filter Substrates
Expansion
Coefficient
Material (x1~I"C)

Aluminum 25"
BK7 glass 7.4t
Si02 0.5
PET 65*
Si 4.2
Ge 5.7
AI~3 5.8 (ave.)
CaF2 24
Nate: These values IlRlIor the temperaIuI8 range
normally experienced by optical syslems. 'From Weest,
1974. tFrom Ohara Glass Catalog, 1974. *"9ranof1,
1985
From WoWs, 1965, unless noted.
188 CHAPTER 6

....m (5 mils). This technique works well with most glasses, but it cannot be used
with fused silica or with sapphire substrates as the break line does not necessarily
follow the scribed line for these materials.
Other methods of cutting glass and other hard substrates include (1) loose
abrasive saws that use either carbide or industrial diamonds as the cutting agent
and (2) diamond-tipped saws. Inside diameter diamond saws, which work so well
with semiconductors such as silicon, perform poorly when used to cut fused silica
because of heat buildup as the silica is a poor heat conductor.

6.2.2. Plastics and Organic Substrates

At the present time, organic materials are not very useful as transmitting optical
elements in the ultraviolet. The energy of UV photons can break the chemical bonds
in most plastics, and this causes them to become absorbing. This damage manifests
itself as a yellowing of the material. In addition, many plastics oxidize and become
brittle and crack and craze with age. Chemical additives can delay these reactions,
but it is difficult to stop them completely.
Organic plastics have limited uses in the infrared portion of the spectrum as
a multitude of vibrational bands resulting from carbon-hydrogen bonds occur in
that part of the spectrum. As such, they can make good wavelength calibration
samples, and sheets of polystyrene are frequently used for this purpose.
Plastic substrates are more difficult to handle than glass substrates for several
reasons. First, they are softer, and therefore, more susceptible to damage than glass.
Next, they can be a source of contamination in a vacuum chamber. Certain plastics
contain low molecular weight plasticizing components. When these plastics are
placed in a low pressure environment, the low molecular weight constituents may
diffuse out of the bulk material, especially if the temperature increases as it would
during a coating run. Apart from the fact that the loss of these volatile components
may well change the properties of the plastic, their loss in the vacuum chamber
has the effect of increasing the gas load the pumping system must handle. These
gases may change the actual process gas composition of the coating chamber, and
this may affect the deposited layer. Thus, plastics that do not contain any low
molecular weight components, including unreacted monomers and other precursors,
are preferable to those that do contain these components (from a coating process
standpoint). Finally, most plastics absorb a large amount of water from the air;
when the pressure is reduced, they release it again. This can similarly change the
gas environment during the coating process. Table 6.4 lists some examples of plastic
materials that can be used as substrates without detrimental effects to the coating
process provided care is exercised in selecting process parameters. Most of these
plastics are rigid, or flexible in thin sheets.
For plastic substrates, the maximum allowed process temperature is either the
softening temperature, or a transition or crystallization temperature. The limiting
temperature of multicomponent plastics may be determined by the outgasing of
some of their constituents, such as plasticizers. The temperature limits on plastics
is difficult to determine, and is most often done on a trial and error basis.
Typically, the inorganic materials used in coatings produce more durable de-

SUBSTRATES AND MATERIALS 189

Table 6.4 Plastics Substrates with Good Coating Characteristics

Material Comments
PET Available in web form for roll coating
Polyimide (Kapton") Available in orange or black versions
Polycarbonate (Lexant) Tough but soft
Acrylic (Plexiglas:!:) Low softening temperature; polished surfaces are very soft
'KapIan is a trademark of the Dupont Corporation. tLexan is a trademark of the General Electric Company.
*Trademark 0'
the Rohm and Hass Co.

posits at higher process temperatures, so it is normally desirable to heat the coating

chamber. Because many useful plastics cannot withstand very high temperatures,
the coatings deposited on them are not as durable as similar coatings deposited on
glass substrates at a higher process temperature. Table 6.2 indicates the maximum
operational temperature for several plastics.
A thin sheet of polyethylene plastic can serve as an interference film material
in the very far-infrared portion of the spectrum. The plastic is impregnated with
various materials that have desirable far-infrared properties, and then sheets with
alternating properties are laid up and the entire assembly is treated to make it a
single massive structure. The layered structure may be pressed in various ways to
reduce the overall (and individual layer) thickness. The result is an interference
stack for the far-infrared portion of the spectrum that has remarkably good properties
(McCarthy, 1967). This approach is feasible in the far-infrared portion of the
spectrum since the wavelength is so large. A wavelength of 50 JLm (2 mil) is greater
than the thickness of many commercially available wrapping films made of poly-
ethylene.

6.2.3. Metals
Metal substrates obviously cannot be used for transmitting optics. They can only
serve as mirrors or absorbers. Because metals are generally good thermal conductors
by comparison to dielectrics, they can function in situations where the heat load
on the substrate is high, e.g., high-power laser mirrors. Metal substrates can be
configured with cooling coils or integral channels through which a coolant can flow
to remove heat. Molybdenum and beryllium are two materials that are often used
for mirror substrates. Other metals can also be used to make optical components.
Beryllium is valuable in those instances where lightweight and low thermal
gradients are needed. Molybdenum is useful as it has good thermal characteristics
and can be machined and welded straightforwardly. These substrates are not typ-
ically coated directly. For beryllium substrates, a nickel-based plating (Kanigen)
is applied to the surface prior to figuring. This plating has better polishing char-
acteristics than the bare beryllium metal.
Large molybdenum mirrors can be plated with copper by a number of tech-
niques, including electroplating. The copper coating is then turned on a lathe with
a single-point diamond tool. The copper is much easier to tum than the molybdenum
190 CHAPTER 6

since it is much softer. The highly reflecting coating is subsequently applied to the
figured (turned) surface.
Because a numerically controlled diamond turning machine can produce an
arbitrary surface, it is useful in producing off-axis optics. These find use in instru-
ments such as X-ray telescopes. Such surfaces require coatings with extremely thin
layers to function properly in these wavelength regions. (See Chap. 7 for more
information on these new coatings.)

6.3. SAFETY GLASS

There are two common ways to control glass as it shatters in order to limit the
damage that can be caused by large shards: lamination and tempering. In lamination,
a layer of plastic bonds two conforming sheets of untempered glass together. When
such a glass panel breaks, the plastic interlayer holds the fragments together. Tem-
pered glass, on the other hand, has a high level of controlled internal stresses.
When tempered glass breaks, the result is that the internal stresses cause the pane
to break into small, relatively harmless pieces.

6.3.1. lamination
The plastic used to bond glass sheets is frequently polyvinyl butyral (PVB), and it
keeps the pieces of broken glass together, thus minimizing the potential for injury
or other damage from glass shards. The plastic bonding layer is fused to the two
glass surfaces by placing the unbonded assembly in an autoclave. High isostatic
pressures force the two glass sheets against the intervening plastic. Heat causes the
plastic to melt slightly and bond to the two glass sheets. The PVB plastic layer
may be colored for functional or cosmetic purposes.
PVB plastic can absorb a relatively large amount of water from the atmosphere.
If a bonded panel is to be coated, the absorbed water can be released in the coating
chamber during the coating process and degrade the coating. It is, therefore, gen-
erally recommended that the coating be applied before the lamination of the sub-
strates, if possible. If a coated surface is laminated, the coating must be able to
survive the autoclave bonding process temperatures, which may reach 125°C.
Placing a coating on the interior surface of a bonded assembly gives it additional
environmental protection; e.g., it cannot be scratched, abraded, or otherwise dam-
aged.

6.3.2. Tempering
A properly performed tempering process increases the breaking strength of glass,
even though the tempered glass has a great deal more internal stresses than annealed
glass. Tempered glass cannot be cut after it has been treated; thus, it must be cut
to size before the tempering treatment is performed. Two types of tempering pro-
cesses are in common use: thermal and chemical.
In the thermal tempering process, a hot piece of glass is rapidly cooled so that
the surface develops a compressive stress, while the interior of the sheet is under
SUBSTRATES AND MATERIALS 191

a tensile stress at room temperature. This process works well on glass sheets that
are over 3 mm (l/8 inch) thick; the necessary thermal gradients cannot be set up
when the glass sheet is much thinner than this. The thermal tempering process may
be done either before or after the coating has been applied to the surface. On the
other hand, it should be verified that the coating can survive the tempering process
before a commitment is made to temper coated parts.
In chemical tempering, an ion exchange process replaces the sodium ions near
the surface of the glass with larger ones, such as potassium. The increase in surface
volume causes compressive stresses at the surface, as does heat tempering. Chemical
tempering can be applied to even very thin sheets of glass. The limitations in this
process revolve around the issue of size, as the entire glass sheet must be immersed
in a vat of molten salt. These parts must consequently be cut to shape and then
tempered.
The coating affects the ion diffusion rate, and in some cases can reduce it
markedly. This change depends on the materials in the coating and their total
thicknesses and crystalline form. It is difficult to generalize a tempering process
with the sort of variables involved, therefore, it is generally agreed that the coating
process is the last step in a manufacturing process involving tempered parts.

6.4. SUBSTRATE SPECIFICATIONS

Optical glasses are very tough materials as long as there are no local defects that
can cause stress concentrations that exceed the tensile strength of the glass. Glass
fails mechanically from tensile stresses, not compressive ones. In some circum-
stances compressive stresses can give rise to tensile ones. This phenomenon can
be seen in the conical shape of the hole formed in a glass pane where a small hard
object strikes it. The break occurs where the glass is in tension during the impact.
A piece of glass with an edge crack breaks more easily than one that has a
smooth finish on its edges. Stresses, such as may be generated from uneven heating
from heaters in a vacuum chamber or from a coating source, concentrate at the
edge cracks. The stress at the tip of the crack causes it to propagate very quickly
to the interior portion of the part, and perhaps causes catastrophic failure of the
substrate. Thus, the edges of glass parts are generally carefully inspected for cracks
prior to coating them.
The United States Armed Forces have established specifications for the sizes
of chips and scratches in glass optics. (Section 5.4. describes the relevant speci-
fications.) These defects can scatter light, which is undesirable in a high performance
optical system. Bennett (1978) discusses how the digs and scratches influence the
amount of light scattered by the substrate. Scheele (1977) has measured the forward
and the backward scatter at some common laser wavelengths of some typical infrared
substrate materials. Tables 6.5 and 6.6 show some of the data from his publication.
The surface finish of a substrate is at times a problem for the coating manu-
facturer. A surface that is acceptable from an optical perspective may not perform
well as a substrate for thin films. A surface can have nucleation properties that
192 CHAPTER 6

Table 6.5 Total Forward Scattering

Wavelength (p.m)
Thickness
Sample (mm) 0.633 1.06 3.39 10.6
IRTRAN 1 3.0 0.14 0.027
6.0 0.17 0.43
IRTRAN 2 2.0 0.089 0.012 0.0007
6.0 0.14 0.024 0.0013
IRTRAN 4 2.0 0.045 0.028 0.0025
6.0 0.15 0.11 0.015
AI~3 3.0 4.5 X 10"-5 9.4 x 1~
CaF2 10.0 0.016 0.0044
Ge 4.6 4.7 x 10"""
ZnSe 5.2 0.Q11 0.024 0.0065 9.0 x 10"""
From Scheele, 1sn

cause films to grow in a manner that causes scatter. These surface conditions are
frequently invisible, and show up only after the thin film has been applied. Coating
manufacturers are in the best position to advise those unfamiliar with these problems.

6.5. BLOCKING PROPERTIES

When wide wavelength bands need to be blocked by a filter, it is frequently not

possible or economical to accomplish this solely with interference stacks. It is
sometimes possible to block the transmission over certain ranges by a suitable
choice of substrate material.
All materials have transmission as well as absorption bands at various places
in the electromagnetic spectrum. Table 6.1 gives the approximate ranges over which
a few materials are useful as optically transmitting substrates. The ideal substrate

Table 6.6 Total Backward Scattering

Wavelength (p.m)
Thickness
Sample (mm) 0.633 1.06 3.39 10.6
IRTRAN 1 3.0 0.046 0.0078
6.0 0.066 0.0068
IRTRAN 2 2.0 0.037 o.oon 0.00053
6.0 0.045 0.014 0.0011
IRTRAN 4 2.0 0.017 0.013 0.0017
6.0 0.043 0.037 0.0063
AI~ 3.0 1.6 x 1~ 4.6 x 1~
CaF2 10.0 0.0086 0.0021
Ge 4.6 7.2 x 1~
ZnSe 5.2 0.0024 0.0014 5.6 x 1Q-4 1.5 x 10"""
From Scheele, lsn
SUBSTRATES AND MATERIALS 193

material has high transmittance to the wavelengths of interest and high absorptance
of light energy at those wavelengths that must be blocked.
Absorption filters have no angle shift such as there is in interference filters.
A shift in the transmission level can take place in an absorption filter since the
angle on incidence increases as the path length within the absorbing substrate
increases in proportion to the secant of the propagation angle within the substrate.
Thus, the attenuation increases (transmission decreases) with increasing angle of
incidence. (In calculating the transmission through an absorbing substrate, the
surface Fresnel losses must also be included to get a precise result.)
The attenuation of a light beam is a negative exponential function of the
thickness, and an appropriately absorbing substrate can be very effective at blocking
unwanted wavelengths. It may be quite difficult to get the same level of attenuation
with a reasonable number of thin film layers. An absorption coefficient of about
l40/cm allows a transmission of less than one part in a million through a substrate
1 mm thick. If this same material is used in a thin film design in a layer 1 !Lm
thick, the transmission is of the order of 99 percent (ignoring any interference
effects).
When absorption is used to block energy, the incident energy level should not
be so high that it excessively heats the substrate. If the substrate could overheat,
a reflective coating could be added to the blocking filter to reduce a portion of the
heat load. Otherwise, active cooling of the absorber is required.
The blocking properties of a substrate may be temperature dependent. For
example, gennanium changes its optical properties when the temperature approaches
77K. In addition, the transmission properties of gennanium and other semicon-
ductors can decrease as the temperature is increased. Care should be taken to be
certain that substrates that are to be used as blockers, or as transmitters, for that
matter, at temperatures other than room temperature are carefully checked for their
blocking or transmitting capabilities at the operating temperature.
The use of absorbing blocking filters is frequently a wise choice from a tech-
nical, as well as from an economic, standpoint. The limit on this type of filter is
that the useful blocking bands are fixed by available materials and are not as
adjustable as are interference filters.

6.6. RESTSTRAHL BANDS

Crystalline materials have spectral regions in which the reflectance increases to high
values. Relatively thin sections of the material are essentially opaque in these
regions. These bands are often found in the infrared portions of the spectrum. When
broadband infrared light is reflected by one of these materials, only light with
wavelengths corresponding to these bands remains after several reflections. Thus
the bands became known as reststrahLen reflection bands, from the Gennan word
for remainder or residuaL ray (Kittel, 1968, p. 154; Jenkins and White, 1957, p.
454).
These bands exist in the spectrum of both crystalline and amorphous phases
of a material, though the structure in the reflection peaks is sharper and more distinct
194 CHAPTER 6

in the former. These bands are a manifestation of lattice or interatomic bond res-
onances. The detailed structure depends on the long and short range order of the
material. A good example of this is silicon dioxide; crystal quartz has a much
different reflectance spectrum than does amorphous fused silica, yet they both have
the same chemical composition.
The index of refraction undergoes rapid changes as a function of wavelength
in a reststrahl region. The reflectance at wavelengths in the reststrahl band can
also vary quite rapidly as the angle of incidence varies. Figure 4.44 shows the
theoretically calculated reflectance for a number of angles of incidence for fused
silica.
The reststrahl reflectance can sometimes be used to advantage in designing a
filter. In other instances, it is a hindrance, as when a highly absorbing coating is
desired. Thus, it is wise to know as much about a material as possible when making
a selection.

6.7. ISLAND FORMATION IN THIN FILMS

In the formation of a thin film, the atoms or the molecules that condense on the
substrate surface may have a finite surface mobility. They may move about on the
surface prior to becoming affixed to a particular site. The range of the particles
dePends on the relative abundance of nucleation sites, the surface mobility of the
condensate, the kinetic energy of the molecule or the atom, and so forth. In the
initial stages of film formation, the substrate surface may have a number of nu-
cleating sites on it. The arriving molecules preferentially attach themselves to these
points. In metals especially, the atoms have good surface mobility and may move
a significant distance to find a low energy site. In the former case, material builds
up at the nucleating site at the expense of adjacent areas where there is relatively
much less material. The succeeding molecules that arrive preferentially attach them-
selves to the molecules that are already condensed and therefore, there is a buildup
of material in localized areas. The result of this process is the formation of islands
of the film material. These microscopic islands are the subject of a great deal of
current study, as their properties are not always what one might expect on the basis
of bulk properties of the material. Effective media theories provide means of treating
discontinuous island films as homogeneous films.
The optical properties of spherical particles immersed in a dielectric (massive)
medium have been studied for many years. J. C. Maxwell-Garnett (1904) wrote
one of the first treatises on the subject. A number of theories have been proposed
that can approximate these films with equivalent refractive indices. Depending on
the exact shape of the metallic particles, the resulting equivalent complex index of
refraction can take on a variety of values. Thus, it is quite difficult to model
accurately the behavior of these films, especially when the index of the surrounding
medium varies.
This phenomenon is the basis of some practical coatings. Optical data disk
designs have been patented that are based on the transformation of one type of
island film into another (LaBudde et al., 1983). The sensitivity of very thin metal
SUBSTRATES AND MATERIALS 195

films to elevated temperatures is used to transform the index of the film. Initially,
the metal islands are very small. The islands in a spot heated with a focused laser
beam coalesce and form much larger spherical particles. The difference in index
of refraction between these two regions allows the data bits to be read optically.
A film composed of noncontiguous metal islands has little thermal conductivity in
the plane of the film. This reduces the amount of power that needs to be deposited
in the film to raise its temperature to the point where an irreversible change in its
optical properties occurs.
The transmittance of visible light through a growing thin metal film is inter-
esting. For example, the transmission of a glass substrate starts at approximately
96 percent and decreases as silver is deposited on it. At the point where the islands
begin to coalesce with one another to form a continuous film, the transmission
begins to increase, until, at the point where most of the film is a single continuous
sheet, the transmission again begins to drop. It is at this point that the index of
refraction of the film begins to become stable. The exact thickness at which this
occurs varies with the nucleation properties of the underlayer. A similar behavior
of the index is observed for noble metals. Metals, such as nickel and molybdenum,
do not show this property quite as distinctly as the noble metals. This may be
caused in part by the fact that the noble metals form more distinct islands.

6.8. THERMAL RUNAWAY IN SEMICONDUCTORS

Germanium, with its high index of refraction in the infrared, is a very useful material
for thin films designs. It is generally not a good material to use in high-power laser
systems. At room temperature, germanium has sufficient absorption that a high-
power laser beam can heat it. Germanium, like all semiconductors, becomes more
absorbing as its temperature increases. This results in more heating from the laser
beam, and more absorption. This vicious cycle is termed thermal runaway. This
avalanche process increases until the laser beam is turned off, the optic is damaged,
or the transmission goes to zero. In some instances, this effect has been used as a
passive switch, but it is generally a drawback in most optical systems. The use of
germanium should be avoided when the temperature is expected to exceed 55°C
(Savage, 1985). Cooled germanium substrates cannot tolerate power densities greater
than 88 w/cm2 • Other semiconductor materials can show this effect also, but it is
most evident with germanium as it has a small band gap. Materials with a wider
gap, such as silicon, do not have this problem as their intrinsic absorption at room
temperature is much smaller.

6.9. NUMBER OF MATERIALS IN MULTILAYER COATINGS

Although in theory there is no upper limit to the number of materials that may be
used in a thin film design, most manufacturers try to limit designs to as few materials
as possible. A trade-off is sometimes possible between the number of materials and
the number of layers in a design, and the manufacturer will select the optimum one
given the various constraints.
196 CHAPTER 6

Designs with two materials are far easier to control because the thickness ratio
of the two materials is a major parameter in the performance of a filter. Generally,
the number of layers and the number of materials in a design are of little interest
to the user, but of significant interest to the manufacturer.

6.10. TOXICITY

Manufacturers naturally prefer to produce, and users prefer to purchase filters with
materials that are not toxic. At times, it is necessary to use less desirable toxic
materials in a design because they are the only ones that allow the filter to meet
the user's specifications. These materials can frequently be incorporated safely into
a design with no adverse effect on the user or the environment.
A good example of a toxic material that can be safely used and is quite valuable
from an optical point of view is thorium fluoride. This material is radioactive because
it contains thorium, an emitter of soft alpha particles. Optically, it is an excellent
material for thin film coatings as it has an optical transparency from the near
ultraviolet to the far-infrared (350 nm to 12 ,...m). In designing a coating that
incorporates thorium fluoride, some material other than thorium fluoride is generally
used as the outer layer. With a sufficient thickness of material over the thorium
fluoride, the thorium's alpha particles can be stopped before they leave the coating.
Thorium fluoride is perhaps one of the best toxic materials available because it is
easily detected. A Geiger counter with an appropriate window, or one of its more
modem counterparts, can easily detect the presence of the material, whereas most
other commonly used toxic materials require elaborate chemical analyses to deter-
mine their presence. Thus, one can easily check a coating for stray radiation, and
can check the area around a coating machine for residual thorium fluoride without
elaborate equipment.
Although a fairly large data base exists on the effects of various bulk materials
on the human body, there is very little information available on these and other
materials when they are in thin film form. Thus, one has to assume the worst case
situation for the materials one uses in fabricating thin films. This undoubtedly leads
to excessive caution in some instances to be on the safe side. Fortunately, the
quantity of materials involved in a coating is relatively small in terms of total mass
involved in a given filter.
 Chapter 7

Future Prospects

The technology of fabricating optical thin films has been very dynamic over the
years, and the prospects for the future are of more exciting advances to be expected.
Many coating problems are being addressed in universities and industrial laboratories
alike. Also, new technological advances, some of which are fallout from the space
program, can be expected. In this section, we review a few of the current areas
that hold some promise of changing the state of the coating art and some of the
implications that these advances have for the users of these coatings.

7.1. ION BEAMS

The conditions under which a coating is deposited determine the quality of the thin
film coating. The surface mobility of the arriving species is usually an important
parameter in the physics of the deposition and nucleation process. At low temper-
atures, the arriving atom or molecule does not move very far from the initial point
of contact with the substrate before it loses its kinetic energy. If the substrate is
hot, the atom or molecule can migrate over a larger area before becoming immobile,
thus increasing its chances of finding a nucleation site with the lowest possible
energy state. Another important parameter in deposition physics is the kinetic energy

197
198 CHAPTER 7

of the arriving species: atoms with higher kinetic energy move about on the substrate
surface more than those that arrive at low speed with less energy. This explanation
accounts for some of the differences observed between sputtered and evaporated
films. Sputtering processes impart much more energy to each atom or molecule
that is knocked out of the target than do conventional evaporation processes in
which the evaporant has the energy associated with the evaporation temperature.
Unfortunately, sputtering processes frequently have a significantly lower deposition
rate than do evaporative ones for many oxide materials, and some materials are not
sputterable at all. Thus, a great deal of development effort is being expended to
understand better the deposition physics to get the best of both worlds: high quality
films formed at high deposition rates at low substrate temperatures.
In evaporative processes, the substrate is frequently heated to enhance the
surface mobility of the evaporant (Lyon, 1946). Higher temperatures in the chamber
can yield better, more durable, denser (higher index) films. Unfortunately, these
higher temperatures are costly in terms of energy consumption, process time to get
the substrate to the coating temperature and to cool it down again (without thermally
fracturing it), and coating chambers and equipment that will handle these temper-
atures. In addition, some substrates degrade at high temperatures. In particular,
most common organic (plastic) materials degrade at temperatures above 70° to
100°C. Some glass and metal precision optics may warp (on a submicrometer scale)
at temperatures that yield durable films.
One would like to have a method whereby the energy that depositing species
need to be able to move freely on the surface prior to becoming immobile is supplied
by a means other than heat. Several schemes have been proposed to accomplish
this. The most successful ones have involved the use of particle beams that flood
the substrate surface with low energy ions « 1000 eV, and frequently < 100 eV)
or neutral atoms. The energy of these atoms should not be so high that they cause
a loss of stoichiometry or other coating degradation. For example, titanium dioxide
films are particularly susceptible to a loss of some oxygen, which in turn makes
them absorbing in the visible and the UV portions of the spectrum. Finally, there
should not be a net electrical charge in the particles arriving at the surface, lest the
surface becomes charged and repels the activating beam.
The ideal beam source has a large flux of activating species with a low (by
ion gun technology standards) energy per atom or ion. The problem with today's
ion sources, such as many of those that have been developed under the auspices
of the NASA space program, is that the particle flux that can be drawn out of them
at low energies is limited. Thus a central theme in contemporary research on ion
sources for thin film deposition use is to find a way to achieve high currents at low
voltages. In particular, as Ohm's law applies, what is needed is a source with a
very low impedance so that a large current can be obtained at low voltages.
Research on sputtering sources is also proceeding apace. It has been recently
reported that magnesium fluoride has been successfully deposited with sputtering
equipment. This is a significant breakthrough as magnesium fluoride is a low index
material that is very useful in making wide band antireflection coatings in the visible
portion of the spectrum. Prior to this announcement, fused silica was the lowest
index material that was sputterable. The goal of sputter research is to enable the
FUTURE PROSPECTS 199

process to deposit materials at higher rates than previously possible. In addition,

in some situations, it is desirable to sputter at lower pressures than allowed by
conventional systems. This lower pressure operation reduces the gas that may be
trapped in the film. Some low-pressure sputter sources are beginning to find their
way into production environments.
Sputtering processes are prone to arcing between the plasma and the metal
objects. This can result in a deposit of debris on the substrate. Once the physics
of the process is better understood, we can expect significant advances in sputtering
technology.
The results of this research will be more durable, more stable coatings and
filters that can be deposited at lower temperatures (McNally et aI., 1986). One
finding is that ion beams can influence the stresses in coatings. Materials with high
tensile stresses, in particular, will benefit from deposition in the presence of an ion
beam. Thus, the kind of coatings that are today only possible on glass substrates
could one day be found on plastics and other materials that cannot withstand high
temperatures.

7.2. NONVACUUM PROCESSES

The processes used in the commercial manufacturing of optical thin films are
predominantly based on vacuum technology. The sources of the deposited species
are sputter targets or evaporation sources. Other technologies are now being de-
veloped that may play some role in the fabrication of the optical filters of the future.
One promising new coating technology consists of using organometallic com-
pounds as the source of the metallic ion. After the compound has been applied to
the surface of the substrate, the molecule is oxidized during a bake in an oxidizing
atmosphere. The oxidized hydrocarbon components are all volatile, consisting for
the most part of carbon dioxide and water. The metal ions form an oxide that
remains as a film on the substrate.
There are two basic forms of this process. In the sol-gel technique, the molecule
containing the metal ion is suspended in an organic solvent solution, whereas in
the other technique, the solution consists of the organometallic molecule dissolved
in an organic solvent. Either of these liquid solutions can be sprayed onto a substrate,
but the more common usage involves a dip process, wherein an immersed substrate
is withdrawn from the solution at a controlled rate. The thickness is determined by
a number of parameters, including viscosity, temperature, withdrawal rate, angle
of withdrawal, and surface tension of the solution.
The equipment required for the dip process is relatively simple when compared
with the capital investment required for an equivalent vacuum-coating process. On
the other hand, there is still a great deal of work ahead for the developers of this
technology. A major difficulty at present is the high level of stresses in the film
after baking.
There is a great deal to be gained in pursuing this technology. If the problems
can be overcome, the cost of coatings could decrease significantly, and thus open
new areas where thin film coatings might be used.
200 CHAPTER 7

High-pressure (by vacuum-coating standards) chemical vapor deposition (CVO)

is a standard coating process in the fabrication of semiconductors. This technology
may one day be applied to the manufacture of large quantities of thin film filters.

7.3. INHOMOGENEOUS FILMS

Inhomogeneous films have been produced since the earliest days of the making of
optical thin films. For the most part, they have not been desirable as their index
profiles were unpredictable. In addition, with the exception of simple antireflection
coatings, there were few workable systems for designing complex filters with in-
homogeneous index layers incorporated into them.
Some recent breakthroughs have given inhomogeneous coatings new impor-
tance:
I. Recent improvements in computer algorithms have enabled the theorist to
calculate the complex inhomogeneous index profiles required for novel filters (Dob-
rowolski, 1986; Southwell, 1985).
2. The rate control of coating sources is far superior today than it was just a
few years ago.
3. Today's computers can evaluate the performance of a growing inhomoge-
neous film in real time. This enables the coating technician or system to monitor
the index profile as the film grows, and make adjustments or corrections as nec-
essary.
Inhomogeneous films have several advantages over discrete layers. The lack of
discrete interfaces between layers eliminates one of the weakest mechanical points
in a thin film stack, the interlayer boundaries. By eliminating these, the durability
of the coating should be improved. In addition, the elimination of the interface may
reduce some of the absorption in the coatings (see section 3.5.). This concept can
be extended to discrete layer designs by grading or blending the interfaces between
layers.
The use of a continuously variable index introduces additional degrees of
freedom for the thin film designer to use in the design process. With inhomogeneous
layer design techniques, it may become possible to design filters whose performances
are far superior than can be achieved with discrete homogeneous layers. These
advances may well be reflected by the emergence of inhomogeneous films in the
commercial marketplace.

7.4. CODEPOSITION

Thin films can be made by codepositing two materials with different refractive
indices. The index of the resulting deposit usually is intermediate between the two
indices of the initial materials, and the specific value depends on the the composition
of the coating. By varying the relative coating rates of the different sources, the
index can be adjusted from one limit to the other. Inhomogeneous films can be
obtained by varying the deposition rates according to a prescribed function of time
FUTURE PROSPECTS 201

or film thickness. This coating technique, being dynamic, requires good stability
of the coating sources and a monitoring system adequate to the task of ensuring
good control of the index profile.
Several problems must be overcome if a uniform index profile is to be obtained
over a substantial substrate area. In evaporation, the two simultaneously operating
sources cannot be located at precisely the same point, and different parts of the
coating plane or surface will get different ratios of the two coating materials. A
common technique to obtain homogeneous coevaporated films is to use two sep-
arated sources and rotate the substrate above them so that the substrate is exposed
alternately to one and then the other source's maximum coating flux. Although the
index profile has a high frequency modulation component, on the average, it closely
approximates the desired shape. If the theoretical profile must be matched even
more closely, then the evaporation rates of the two sources can be slowed to the
point where the substrates get coated with a monoatomic layer or less during each
revolution. Then, the two components will be very well intermixed. This latter
approach is necessary to coat a filter with a carefully controlled index profile over
a large area, if current technology is used. Thus, if coevaporation is to become
commercially viable, new deposition techniques require development.
The geometry of a sputtering system lends itself to different solutions to the
uniformity problem. One means of obtaining a mixed deposit is by sectioning the
target so that some portions are of one composition, whereas others are from the
second component. The ratio of the two areas determines the compositional ratio
of the deposit. A second approach is to move the substrate between two targets in
a manner similar to what is done in evaporation systems. The profile can be varied
by changing the coating rate, and this is easily accomplished in sputtering by simply
adjusting the power. It is not clear that an inhomogeneous film with an arbitrary
index profile can be deposited over large areas in a sputtering process. One exception
to this exists in reactive sputtering of pairs of materials in which there is a common
constituent that comes from the target, and the other component comes from a
variable gas mixture.
Some materials that may not be compatible in bulk form can be made to coexist
in a film form by codeposition. The basic effect is a solid state condensation taking
place at a relatively low temperature so that a chemical or a physical reaction cannot
take place. The two materials may exist in a metastable state for a long time as
energy is not available to overcome the potential barrier that prevents a reaction
between them. The atoms are frozen into place as they condense, and they do not
have the energy to undergo a phase separation. This approach to making new
materials has not yet resulted in many materials being manufactured in this way.
It is possible as technology develops and new requirements arise that materials of
this type may become economically viable. Cermets are one class that may benefit
from this type of technology.

7.4. X-RAY COATINGS

Researchers are pushing the frontiers of optics to ever shorter wavelengths (Un-
derwood, 1986). Current research focuses on wavelengths in the 1 to 300 A range.
202 CHAPTER 7

As in conventional optics, these new spectral regions require high reflectors, an-
tireflection coatings, and bandpass filters. In the past, X-ray imaging involved the
use of mirrors at glancing angles of incidence. This approach limits the available
aperture, results in large aberrations in the image, and limits the resolution of the
system. At present, we are just beginning to understand how to design and construct
multilayers that can function efficiently at these extremely short wavelengths.
There are many problems associated with the making of coatings that function
in the extreme ultraviolet and the X-ray portion of the spectrum. The most severe
problem is that of finding materials that have appropriate optical constants. At these
wavelengths, most solid materials are absorbing, and the real parts of their refractive
indices are close to unity. In addition, the energy of the X-rays is such that the
entire electron cloud around an atom is involved in determining the optical response,
so that changes in the chemical composition have little effect on the refractive index.
Thus, the material choices are reduced to a consideration of the properties of the
elements. Design schemes are being developed that take advantage of these physical
constraints, or that work around them.
Another problem being addressed is that of making these filters. Some layers
may have to be as thin as 3 atomic layers to form a quarter wave thick interference
film. Monitoring such thin layers during the deposition process becomes crucial
and may involve the installation of X-ray sources in the coating chamber (Spiller,
1981). In addition, stability of the films to both diffusion and chemical attack may
be difficult to obtain.
Optical elements need to have surfaces that are smooth to a small fraction of
a wavelength of the light with which they will function. At far-UV and X-ray
wavelengths, the substrate surfaces must be far smoother than that needed for visual
optics. At least two current technologies are available that can produce surface with
the requisite low surface roughness. Semiconductor wafers have surface finishes in
the 3-A rms range, but unfortunately their surface figures are not generally good
enough for imaging work. Laser gyroscope mirror substrates are prepared with good
surface finishes as well as good figure control. The scatter is the coatings needs to
be improved, and this is one area of study that should prove fruitful for all spectral
ranges.
It is just a matter of time before production techniques for X-ray coatings will
no longer be state of the art and become routine and commonplace. The result will
be complex optical systems with multiple reflecting surfaces that yield quality
extreme ultraviolet and X-ray images.

7.6. EQUIPMENT ADVANCES-

AUTOMATIC COATING 'MACHINES

Significant advances in the area of coating equipment has been taking place in the
past few years. Microprocessors have been added to coating chambers, and new
process monitors make interfacing them to the actual deposition process even sim-
pler. As a result, it is now possible to purchase a totally automated vacuum-coating
system.
FUTURE PROSPECTS 203

The state of the art has not yet progressed to the point where it is possible to
buy a turnkey system that is guaranteed to produce a particular coating, but these
will probably be available in a few years. The systems will be set up to manufacture
one specific coating design, and it will be difficult for the novice to alter the process
to coat successfully another type of design. The dedicated coating manufacturers
will continue in business, and their customers will continue to return to them for
most of their coating requests.

7.7. MATERIALS RESEARCH

The area of thin film technology that will advance the field the most in the next
few years is undoubtedly materials science. Fundamental data in the thin film area
is still lacking at the most basic levels. As an example, Ennos' paper (1966) on
stresses is still the only readily available publication on the subject. The work was
performed 20 years ago!
It is necessary to undertake fundamental studies of the basic physical parameters
of matter in thin film form. Many of the physical properties of materials in thin
film form are unknown: stresses, thermal conductivity in both the film plane and
perpendicular to it, and Young's modulus and other mechanical constants. Even
the refractive index of most compounds is poorly known except in a few wavelength
regions. The advent of lasers, synchrotron light, and other light sources in unex-
plored portions of the spectrum will mean new requirements for coatings and coating
materials in these wavelength regions. It will be necessary to know the optical
constants of materials in these wavelength regions before adequate filters can be
designed and constructed.
The chemistry of thin films is not the same as that of bulk materials. When
the density of a material is 70 to 90 percent of that of the bulk, one can be assured
that there is a large surface area within the film. The effect of this on the chemistry
of the material is poorly known. Silver is a very good example of what can happen
to a material when it is in thin film form. In bulk form, silver is relatively inert,
save for the fonnation of the sulfide. In thin film fonn, it seems to participate in
many different reactions in the presence of moisture. Thus, this seems to be a fertile
field for study, and the results of these studies will be of great benefit to those
working in this field.
 Appendix A

How to Specify
Optical Filters

By far the most important thing one must do when one specifies a filter's performance
is to determine one's absolute "needs" and to distinguish them from one's "wants"
or desires. The needs relate to the functional aspects without which the filter and
the rest of the optical system will not function properly. These needs do not nec-
essarily have to be optical in nature: the tolerance on the size of the filter or the
environmental parameters might be very important needs.
The wants, on the other hand, are additional features that add value to the
system but are not critical to its fundamental operation. A bandpass filter may be
adequate with a certain tolerance on its bandwidth, but the system's performance
might be significantly better if it were made narrower: this would be classified as
a want.
There are always gray areas that need to be resolved subjectively on an in-
dividual basis. Hopefully, these guesses are minor when compared with the total
number of decisions required in the system. Usually, more information or a better
understanding of the system in which the filter will be used can help to clear up
ambiguities.
Cost considerations are always of paramount importance in commercial pro-
curements. On the other hand, in high technology and military purchases, the
primary considerations are likely to be optical and environmental performance.

207
208 APPENDIX A

To get the maximum benefit from a filter supplier's expertise, it is wise to

involve that person in the design of the system. The function and the location of
the filter may have an effect on the system's performance, and the supplier may
be able to make some suggestions that will have the effect of reducing the total
system cost.
The filter manufacturer should be brought into the design picture at the earliest
possible stage. Filters are an integral part of an optical system, and if a system
change needs to be made late in the design cycle, the costs of the design task may
escalate. Polarization effects are one area where a small change can sometimes
make a large difference in the difficulty of producing an acceptable filter. A simple
rotation of the plane of incidence or of the filter can simplify the design by changing
the plane of polarization with respect to the plane of incidence on the filter.
Coatings are not separable from the performance of the optical system in
diffraction limited systems. Filters should not be left to the end of a project; an
adequate filter may not be available off-the-shelf.
The following questions are typical of those that are likely to be asked by the
representative from a coating house in response to a call for a filter. Of course,
they may not all be appropriate in any given instance.
1. What type of filter is desired? What is the spectral range over which the
coating must function? What performance (transmission, reflection, and ab-
sorption) is required at each point?
a. Questions Specific to Long-Wave Pass Filters
What attenuation level is required over what wavelength range? At what
wavelength is the edge located? How precisely must this be positioned?
Is there a requirement on the steepness of the slope?
What must be the minimum and average transmission in the high trans-
mission region? Over what range does this specification apply?
b. Questions Specific to Short-Wave Pass Filters
At what wavelength is the edge located? How precisely must this be
positioned?
What passband width is required?
What minimum and average transmission levels are desired over this
range?
Is there a requirement on the steepness of the slope of the edge?
c. Questions Specific to Bandpass Filters
What is the center wavelength?
What is the desired bandwidth? How is this specification defined?
What is the minimum or average transmission in the pass band?
Is any blocking outside of the pass band required? How low must it be?
2. Will the filters be used at normal incidence?
Will there be a range of angles of incidence on the filter?
Will the light be collimated, or will there be a steeply convergent beam of
light in the filter plane?
3. If the angle of incidence on the filter is nonnormal, is the incident light
polarized?
4. What are the tolerances on all numbers specified?

HOW TO SPECIFY OPTICAL FILTERS 209

S. In what kind of environment will the filter be used?

What environmental tests should be performed on the filter to verify adequate
durability?
6. Will the coating be exposed to high-power beams? Will the light be coherent
or incoherent? Will this energy be in-band or out-of-band?
Will the coating be exposed to other energy sources, such as ultraviolet light,
ionizing radiation, or high energy particles?
7. If the filter has absorbing layers in it and is not symmetric, is there a require-
ment on the reflectance from the second surface (reverse reflectance)?
8. Is scattered light a problem in the optical system? If so, to what level must
scatter be controlled in the filter? How should this be measured to verify that
the desired level of performance has been attained?
9. Will the filter be exposed to elevated temperatures (> 100°C) in storage or in
use? If so, what level will the temperature reach?
Will the filter be exposed to temperatures below O°C?
How long will the filter be exposed to these temperatures?
How rapidly will the temperatures change?
10. What is the incident medium? What is its index of refraction?
11. Of what type of material(s) will the substrate be made? If glass, what type
and what is its index of refraction? Similar questions will be asked for plastic
and infrared materials.
12. What is the physical size and shape of the substrate? Will it be round, square,
or some peculiar shape? What will be the aspect ratio (ratio of the largest
dimension to the smallest)?
Who will supply the substrate? If the filter manufacturer is to fabricate the
substrate, who will supply the raw material for the substrate? What are the
specifications for this?
13. Are there any special precautions that need to be taken with the substrates;
e.g., is the substrate material prone to staining or sensitive to heat?
14. How many substrates will be required? Over what period of time? What will
the lot size be? How many extra pieces can be supplied for testing purposes?
15. When will the customer want the first parts delivered?
When will the coating manufacturer receive the first shipment of substrates
(if customer supplied)?
16. Does the customer have any special cleaning requirements or recommenda-
tions? This is especially important if the coating manufacturer is unfamiliar
with a specific substrate material.
17. Are there any requirements to assemble the filter into an assembly or sub-
system?
18. Are there any other requirements or specifications that the filter must meet
that might not be obvious to the supplier?
 Appendix B

Selected United States

Military Specifications
that Relate to Thin Film
Optical Filters

A number of United States Armed Forces specifications are used to qualify optical
coatings. An early specification, Mil-C-675, was specifically designed for single
layer antireflection coatings of magnesium fluoride on glass optical elements. In
fact, it even calls out the range of thicknesses the coatings can have. In addition,
however, it specified that the coated element should pass certain humidity, solubility,
abrasion, and adhesion tests (Table B.t).
These environmental tests have been applied to a much broader class of coatings
than originally intended by the authors of this specification.
Another specification, Mil-M-13508, was written specifically to deal with
aluminum mirror coatings, which are typically less durable than a hard magnesium
fluoride AR coating.
A third specification, Mil-STD-81O, covers material testing in general and is
not specific to optics. Some sections of this rather long document, however, codify
tests that are suitable for performing accelerated degradation tests on optical coat-
ings. Sections 507.2 ("cycled humidity") and 509.2 ("salt fog") are very rigorous
tests that cause degradation in all but the most durable coatings.
Specifications -675 and -13508 are now obsolete. These have been replaced
by three specifications that are more appropriate to the wide variety of modem
coatings. The new documents are

210

SELECTED UNITED STATES MILITARY SPECIFICATIONS 211

Table B.1 Tests for Coated Elements

C-675 M-13508 STD-810 C-14806 C-48497 F-48616

Revision C C D A.2 A Ammend.1

Date Sept. 10 Mar. 19 July 19 May 1 Sept. 8 Aug. 22

1982 1973 1983 1974 1980 1980

Abrasion Severe: 50 rub none 20 rub 50 rub 50 rub

40 rub cheesecloth eraser cheesecloth cheesecloth
eraser

Moderate: Severe: Severe:

50 rub 20 rub 20 rub
cheesecloth eraser eraser
optional optional

Adhesion
(tape) fast slow none none fast fast

Humidity
Method A
Temp (OC) 48.9 48.9 30-60 82-160 48.9 48.9
Temp (OF) 120 120 86-140 27.8-71.1 120 120
% ReI. Hum. 95-100 95-100 85-95 85-95 95-100 95-100
Hours 24 24 24(cycle) 240 24 24
Conditions Constant Constant Variable Variable Constant Constant

Method B
Temp (OC) 48.9
Temp (OF) 120
% ReI. Hum. 95-100
Hours 24
Conditions Constant

Method C
Temp (OC) 82-160
Temp (OF) 27.8-71.1
% Rei. Hum. 85-95
Hours 48
Conditions Variable

Salt Fog/Salt Spray

Time (hr) 24 24 48 48 none 24(option)
Temp. (OC) 48.9 35 35 35 ? ?
Temp. (OF) 120 95 95 95 ? ?

Solubility (hr)
Water 24(opt.) 24(opt.)
Saltwater 24 24(opt.) 24(opt.)

Other 10 min in each:

trichloroethylene,
acetone, and ethanol

212 APENDIX B

Table B.1 (Continued)

M-13508 C-14806 C-48497 F-48616
Temperature
(OC) -62 to 71 -54 to 71 -62 to 71 -62 to 71
("F) -80 to 160 -65 to 160 -SO to 160 -80 to 160
Time at
each temp. (hr.) 5 48 2 2

Shock
(dTldt) <5 min to <4°F/min <4°F/min
cover range
Room Tto
160, hold 4
hr, to -80,
hold 4 hr, to
160, hold 4
hr ("F)

Dust 28 hr in fine
sand in 2000
ftlmin
airstream

Fungus 28 days in a
mixed spore
suspension

Notes Test sequence (order) and delay time

between tests is specified

These specifications include scratch

and dig values

Mil-C-14806 Coating, Reflection Reducing, for Instrument

Cover Glasses and Lighting Wedges
Mil-C-48497 Coating, Single or Multilayer, Interference:
Durability Requirements for
Mil-F48616 Filter (Coatings), Infrared Interference:
General Specification for
This last specification also includes a number of supplements. These are intended
to be specific to certain types of optical filters.
As a result of the availability of these specifications, which are directed at
multilayer thin film interference technology, it is now possible to specify tests that
are directly applicable to the product being tested. Mil-C-67S is directly replaced
by Mil-C-14806, and therefore, the latter should be now used exclusively. Similarly,
the other two specifications cover the general field and the cases of specific infrared
coatings.
All of these specifications are subject to change at any time. The specifications
SELECTED UNITED STATES MILITARY SPECIFICATIONS 213

given here were correct as of their release date. Revisions are designated by a suffix
letter, starting with the letter "A." Users should assure themselves that they are
referring to the latest edition of a specification, or know the changes that have been
made to a given edition.
Some of the newer specifications explicitly define the order in which the test
procedure is to be conducted on a coating sample. The change in procedure has
two significant implications: (1) the test is much more severe than that when the
same tests are run in parallel on separate substrates, and (2) the time that it takes
to perform all of the tests is much longer than when tests are run in parallel,
especially if there are some long duration humidity tests to be performed before
abrasion and adhesion tests.
In most tests, the criterion for passing the test is that after the test is complete
no degradation due to testing be observed. Any degradation visible under the
specified examination conditions indicates that the coating has failed the test.

214 APPENDIX B

Hll.-C-14S06A
Incl. Amendment 2
1 May 1974
SliPERSl:DINr.
tIlL-C-14806
27 Hay 1968

HILUARY SPl.CIFlCA1'ION

COATI~G, REFLECTION RE~UCIKG, FOR

INSTRIJI-lIJ'il COVER GLASSJ::5 AND LImiTING WJ::DGES

This specification is malldatory for use by all Departments and Agencies

of the Department of Defense.

1. SCOPE

1.1 This specification covers reflection reducin~ coatin~s applied to

substrate material bavinE indices of refraction within the rany.e of 1.47 to
1.55 to be uaed as instrument cover glasses and li,-htin~ wedres herein
referred to as optical elements.

2. APPLlCAIlLE DOCIJI-ll.Nl"S

2.1 The following documents, of the issue in effect on date of invitation

for bids or request for proposal, form a part of this specification to the
extent specified herein.

SPJ::CIFlCATIONS

Hilitary

HlL-E-12397 J::raser, RuLber-Pumice for Testing

Coated Optical Uements
HlL-Q-13830 Optical Components for Fire Control
Instruments; General Specification
Governing the Manufacture, Assembly
and Inspection of

STANDARDS
Kil1tary

HIL-STD-I05 Samrling Procedures aDd Tables for

Inspection by Attributes
MIL-STll-SIO Environmental Test Hetbods
HIL-STD-l24l Optical Terms and Definitions
(Copies of specifications. standards. drawin~s. and publications required by
suppliers in connection with specific procurement functions should be obtained
from the procuring activity or as directed by the contracting officer).

FSC 6650

SELECTED UNITED STATES MILITARY SPECIFICATIONS 215

HIL-C-14806A

3. REQUIREMENTS

3.1 Optical terms and definitions.- Terms and definitions peculiar to

the general field of optics as used herein are defined in HIL-STD-124l.

3.2 Coated are'a.- Unless otheNise specified, optical elements shall

be coated over their entire effective aperture except for an all~able
uncoated holding area as shown in the table bel~.

Diameter or Maximum Diagonal

Dimension of Optical Element Max Width of Uncoated Area

0- 1.5" 0.030"
1.5- 2.5" 0.040"
2.5 - 3.5" 0.050"
3.5 - 5.0" 0.060"
5.0 - 7.0" 0.100"

3.3 Cemented and bonded surfaces.- An optical element swrface or

portion thereof which is to be cemented or bonded to another s~rface or
portion thereof shall not be coated on the area to be cemented or bonded.
3.4 Specular reflectance.- When applied to substrate materials having
indices of refraction within the range of 1.47-1.55 the specular reflectance
from each coated surface shall not exceed the following limits for energy
incident on the surface at an angle within the range 0-15 degrees inclusive
and at an angle of 30 degrees:
Maximum Reflectance (%)
Wavelength Range (at any wavelength within
(Nanometers) specified spectral range)
00 to 15 0 incl. 30 0

440-460 0.55
460-660 0.45
660-675 0.55
425-460 l.0
460-620 0.7
620-665 l.0

3.5 Light 10ss.- Within the vsvelength range 425 to 700 nanometers,
light loss (absorbance plus (+) diffuae reflectance) in the coating ahall
not exceed the following limits per surface unless otherwiae specified
(see 6.2):
Maximum average loaa 0.5%
Maximum absolute 10.. 2.0%
3.6 Coating gua1ity.- The coating shall be uniform in quality and
condition, clean, smooth, and free from foreiRn materials, and from physical
imperfections .nd optical imperfections sa follaws:

3.6.1 The coating sball shaw no evidence of flaking, peeling or blistering.

3.6.2 The coating shall not contain blemishes such sa discolorations,

stains, smears and streaka.

2

216 APPENDIX B

!lIL-C-1480bA

3.6.3 The coat~ng ahall show no evidence of a cloudy or hazy appearance.

3.6.4 Unlesa otherwise Ipecified, the coating shall meet the followin~
scratch aDd dig requirementa. CoatiDi defects ahall be counted in addition
to those allowed by the aub.trate specification.

Lighting vedges or light pipiDi alass - 40-20 per !lIL-o-13830

Cover glas. for lighted instruments - 60-20 per !lIL-o-13830
Cover gla.. for unlighted instruments - 8O-SO per !lIL-o-13830

Coating spatter ahall be counted aa a dig or inclusion.

3.6.5 Pinholes in the coated surface shall not eKceed tvice the allowable
dig aize. Unles. othervile lpecified, pinhole distribution ahall be ignored.

3.7 Abraaion resiatance.- There shall be no visible damage to the coated

Bur face vhen rubbed vith an eraaer conforming to !lIL-E-12397 under a force
of 2 to 2 1/2 lbs.

3.8 Blllllidity - "a".- COIIt1l1UOU8 upolure for 24 ....rs in an a~osphere

controlled at 116-' to 124·' and 95 to 100% relative humidity.

3.9 Salt fog.- A S% .alation at a temperature of 95·' for 48 hours.

3.10 Humidity - "C".- Continuous eKposure for 48 hours in a controlled
a~aphere vith a relative humidity ranging betveen 85 and 95% at temperatures
ranging from 82·' to 160·'.
3.11 Envir01llllental.- Coated optical elements ahall withstand the followiDi
environmental expoaures.

3.11.1 Bigh temperature.- Temperature of 160·' for not lesa than

48 hOllrs.

3.11.2 Low temperature.- Temperature of -65·' for not les. than

48 hours.

3.11.3 Temperature shock.- Temperature of 160·' for 4 hours and transferred

vithin 5 minutes into a temperature of -65·' for 4 hours and retllrned vithin
5 .tnutes to 160·' for 4 hour••
l.11.4 ll.a1dity - "A".- Continuous exposure for Z40 houra in • controlled
a~sphere vith a relative humidity ranging betveen 85 and 95% at temperatures
rangiDi from 82·, to 160·,.

3.11.) Dust.- A fine aand-laden air with. maximum velocity of 2000 feet
per minute for 28 hOllrB.
3.11.6 Fungus.- A mixed apore auspension for 28 days.

3
SELECTED UNITED STATES MILITARY SPECIFICATIONS 217

HIL-C-14S06A

4. QUALITY ASSURANCE PROVISIONS

4.1 Responsibility for inspection.- Unless othe~ise specifiea in tne

contract or purchase order, the supplier is responsible for the performance
of all inspection requirements as specified herein. Except as othe~ise
specified, the supplier may utilize his own facilities or any other facilities
suitable for the performance of the inspection requirements specified herein,
unless disapproved by the Government. The Government reserves the right to
perform any of the inspections set forth in the specification where such
inspections are deemed necessary to assure supplies and services conform to
prescribed requiremeuts.

4.2 Certification.- Unless actual performance of the tests is specified

iu the contract (see 6.2), the supplier shall certify product conformance to
the requirements of 3.11 for each shipping lot.

4.2.1 Certification testing.- When such testing is reqUired, each of

six test parts shall be subjected to the sequence of tests in Table I. Upon
completion of the test sequence, the test parts shall be assembled into an
appropriate instrument or test device and evaluated. The ability to read
the instrument without a substantial reduction in clarity shall be considered
evidence of successful camp let ion of the testing. Test procedures are
specified iu HIL-STD-S10.

4.2.2 Certification test samples.- Unless othe~ise specified in the

contract (see 6.2), certification test samples shall be crown glass of
refractive iudex 1.523 and Abbe number of 58.6, and shall be of such a size
and coufiguration that they can be assembled into an instrument or suitable
test device for evaluation.

TABLE I - CERTIFICATION TESTS

Characteristic Reguirement Test Method

Light Loss 3.5 4.4.5

Temperature - High 3.11.1 KIL-STD-S10, Method 501-1
Temperature - Low 3.11.2 KIL-STD-S10, Method 502-1
Temperature - Shock 3.11.3 HIL-STD-8l0, Method 503-1
Humidity "A" 3.11.4 HIL-STD-S10, Method 507.1
Salt Fog 3.9 HIL-STD-S10, Method 509.1
Dust 3.11.5 HIL-STD-S10, Method 510.1
Fuugus 3.11.6 HIL-STD-S10, Method 5OS.l

4.3 luspectiDD prgyis1ons.

4.3.1 Final acceptance inapection.- The classification of defects in

Table II constitutes the minimum inspection to be performed by th~ supplier
prior to government acceptance or rejection by item or lot.

4.3.1.1 Acceptance and rejection.- Rejected lots shall be screened for

all defective characteristics. Removal or correction of defective units and
reaubmittance of rejected lots shall be in accordauce with "Acceptance and
Rejection" as specified iu HIL-STD-IOS.

4

218 APPENDIX B

KIL-C-14806A

4.3.1.2 Submission of product.- Unless otherwise specified by the

contracting officer, inspection lot size, lot formation, and presentation of
lots shall be in accordance with "Submission of Product" provisions in
KIL-STD-105. A lot will be considered to be a shippin~ lot (see 6.5).

TAHLE II - CLASSIFlCAIION OF DEFECTS

Level II of Table I with Sampling Plan Table ItA of KIL-STD-10s

Requirement Test Paragraph

Critical: lIone

Kajor: AQL 1.0% .efective

101. Coated Area 3.2 4.4.1

102. Cemented and Bonded Surfaces 3.3 4.4.2
103. Coating Quality 3.6 4.4.3.1, 4.4.3.3,
4.4.3.4, 4.4.3.5
4.3.2 Final acceptance test.- The classification of defects in Table III
shall constitute the minimum tests to be performed by the aupplier prior to
government acceptance or rejection by item or lot.

4.3.2.1 Test samples.- Unleas otherwise specified (aee 6.2), two (2)
test witness pieces of a similar refractive index and aurface finish as the
parts to be coated and coated in the aame evaporation lot (aee 6.4),. as the
parta, ahall be subjected to each individual test of Table III. No two (2)
test witness pieces need be aubjected to more than one test. The test
wienas pieces shall be placed in the coating chamber such tha,t. they repr.sent
the whole evaporation lot.

4.3.2.2 Acceptance arut rejection.- Failure of any test ahall cauae the
evaporation lot to be rejected. Rejected lots ahall be screened for all
defective characteristics. Removal and correction of defective units and
raubmittance of rejected lots ahaU be in accordance with "Acceptance and
Rejection" as apecified in KIL-STD-10S.

TABLE III
Characteristic Reguirement Test Parasraph
201. Humidity ftBIt and 3.8 4.4.6, 4.4.3.1, 4.4.3.2
Abrasion 3.7 4.4.7, 4.4.3.1
202. Reflectance 3.4 4.4.4
203. Salt Fog 3.9 4.4.8, 4.4.3.1
204. Humidity "c" 3.10 4.4.9, 4.4.3.3
4.4 Teat _thuda and procadurea

4.4.1 Coated area.- Coated area shall be viaually examiped to determine

contormance to 3.2.
SELECTED UNITED STATES MILITARY SPECIFICATIONS 219

KIL-C-14806A
4.4.2 Cemented and bonded surfaces.- Coated srea shall be Visually
examined to determine conformance to 3.3.

4.4.3 Coating quality.- Coated surfaces shall be examined for conformance

to 3.6 as follows:

4.4.3.1 Tbe coating ahall be examined by both transmission and reflection

for evidence of flaking, peeling or blistering. The test method is as shown
in Figure I-b. Such evidence aball be cause for rejection.

4.4.3.2 The coating aball be examined by both transmission and reflection

for evidence of discoloration, stains, streaks, or smears. nle test method
is as shown in Figure I-b. When such evidence is found, the blemished area
shall be evaluated by specular reflectance. Parts exceeding the requirements
of 3.4 sball be rejected.

4.4.3.3 The coatings sball be evaluated as shown in Figure I-c for evidence
of cloudy or hazy appearance. Sucb evidence shall be cause for rejection.

4.4.3.4 The coating shall be examined by transmission as shown in

Figures I-a and I-b for pinhole, and scratcb and dig requirements. Parts
exreeding the requirements specified in 3.6.4 and 3.6.5 shall be rejected.

~.4.3.5 The coating sball be examined by tbe transmission method"for

evidence of discoloration, atains, streaks and smears. The test method is
as s~own in Figure I-b. Such evidence shall be cause for rejection.

4.4.4 Specular reflectance.- To determine conformance to the requirements

of 3.4, the test aball be conducted with a calibrated spectrophotometer or a
low level reflectometer capable of meaauring within the 400 to 700 nanometer
wavelength range. The measuring equipment shall be suitably equipped to
measure specular reflectance at the angles specified.

4.4.5 Light Loss.- To determine conformance to the requirements of

3.5, tbis test ahall be conducted using a calibrated spectrophotometer capable
of measuring witbin the 400 to 700 nanometer wavelength range. Several
elements are scanned in aeries witb tbe spectrophotometer. The elementa may
be skewed in the beam, precoating and post ~oatinr measurements are compared,
reflection losses are extracted and the resulting absorption loss per surface
determined.

4.4.6 Humidity - "8".- Tbe test aamples shall be placed into an

enVironmentally controlled teat cbamber at a controlled temperature of
120 + 4°F and 95% + 5% relative humidity and beld for 24 hours minimum.
After completion of this teat, the parts shall be evaluated for conformance
to 3.6.1 and 3.6.2. Following this evaluation, the parts shall be subjected
to the abrasion teat of 4.4.7.

6

220 APPENDIX B

MIL-C-14B06A

4.4.7 Abrasion.- The coating shall be tested by rubbing the coated

'urface with a standard eraser conforming to KIL-E-12397 mounted in the
holding device. A force of between 2 and 2 1/2 lbs. shall be applied.
Strokes of approximately one inch shall be made if the size of the element
permits; smaller atrokes for smaller elements. All atrokes shall be made
on one path for twenty strokes. After completion of this test, the parts
ahall be evaluated for conformance to 3.6.1.

4.4.B Salt fog.- The coatings shall be tested as specified in

KIL-STD-BlO, Method 509-1. After completion of this test, the parts shall
be evaluated for conformance to 3.6.1.

4.4.9 Humidity - "C".- The coating shall be tested as specified in

KIL-STD-8l0, Kethod 507-1, except that the test time shall be 48 hours
(2 cycles). Upon completion of these tests, the parts shall be evaluated
for conformance to 3.6.3.

s. PREPARATION FOR DELIVERY

5.1 Packaging shall be accomplished in a manner to ensure that the

coated glasses during shipment and storage, viII not be damaged by chipping,
abrading or ataining.

6. NOTES

6.1 Intended use.- The coatings covered by this specification are

applied to the surfsces of optical elements primarily to reduce the intensity
of reflected surface light and increaae light transmission.

6.2 Ordering data.- Procurement documents should specify the following:

a. Title, number and date of this specification.

b. Performance of certification teats if required (see 4.2) and

qusntity of teat parts required (if different than 4.2).

c. Type and aize of certification test parts (if different

than 4.2).

d. Light loaa permitted (if different than 3.5).

e. Number of test aamples (if different than 4.3.2.1).

6.3 Abrasion testing device.- A device conforming to Frankfo~d Arsenal

Drawing No. D7680606 may be used for the abrasion testing of 4.4.7.

6.4 Evaporation lot.- An "evaporation lot" shall be defined ar the

group of parts which has the coatings applied at the s~e time and in the
same evaporation chamber.

7
SELECTED UNITED STATES MILITARY SPECIFICATIONS 221

MIL-C-14806A
6.5 Shipping lot. - A "shippine lot" shall consist of units manufactured
under essentially the same conditions and at approximately the Bame time.

6.6 International standardization. - Certain provisions (paragraphs

3.4, 3.5, 3.6, 3.7, 3.8 and 3.9) of this specification are the subject of
international standardization agreement (ASCC AIR STD 10/61, Coating,
Reflection Reducing for Aircraft Station Displays). When amendment,
revision, or cancellation of this publication is proposed which will
affect or violate the international agreement concerned, the preparing
activity will take appropriate reconciliation action through international
standardization channels including departmental standardization offices,
if required.

Custodians: Preparing Activity:

Army - HU Army - HU
Navy - AS
Air Force - 11

Jeview Activities: Civilian Agencies

Interest-DK
Army - HU, AV
Navy - AS, OS
Air Force - 11, 17, 26

User Activities: Project No. 6650-0035

Army - WC, KI
Navy - SII
Air Force - None

8

222 APPENDIX B

IIIL-C-14806A

C_.rcld fixtuu cDIltaiD-

i", CWO U-VI tt fluorUclllt
lapa (I. . lot. 3)

e
e Tut I ~T.at
e
d _ t ' /Il_t
aurfacI quality for r ...
ltaDdard (1.1 fl.ctiDll.

Tut el_llt
(a" .Di:~ 1) e Surface Quality
Standard
lot. 2)

(a) (b)

r:--n..
IOTES:

2.
Icratch•• 011 the tut
d.ant ahall bl aUIllad paral-
lel to cho.. af che cDllpln1all
.ubc. quality .taDdlrd.

rDr . . .uri", pillhal...

~
L16"1 G
"

Tut d_llt

• cracchu. aDd dial. boch chI

• taDdard aDd chI .1• •1lt lurface
are .,1-..1 uailll t r _ i ttad
11&bt.

3. fb. DIll7 f.ll..l1ll1t1oll ill thl

t .. t uu IbIll b. frae che
11&bt .aurCI uaeel far ten.

4. rar ..,a1uatiDII of flak1ll1.

and hi...
, ..lilll. b1l1terl.... bum..bee.
el_to ahall 11.
cb.cked .1 chaut chI u•• af
'era tch lad 41& • taDdlrd••
(c)

,
fiGUU 1
SELECTED UNITED STATES MILITARY SPECIFICATIONS 223

MIL-C-48497 A
8 September 1980
SUPERSEDING
MIL-C-48497 (AR)
27 June 1974

MILITARY SPECIFICATION

COATING, SINGLE OR MULTILAYER, INTERFERENCE: "DURABILITY REQUIREMENTS FOR

This specification is approved for use by all Departments and Agencies of

the Department of Defense.

1. SCOPE

1.1 Scope. This specification establishes mlnlmum quality and dura-

bility requirements for single layer and multilayer interference coatings
that;are primarily used within the protective confines of sealed optical
systems. (See 6.1).

2. APPLICABLE DOCUMENTS

2.1 The follOl<ing documents, of the issue in effect on date of

invitation for bids or request for proposal, form a part of this specification
to the extent specified herein.

SPECIFICATIONS

Federal

L-T-90 Tape, Pressure-Sensitive, Adhesive,

(Cellophane and Cellulose Acetate)
O-A-Sl Acetone, Technical
O-E-760 Ethyl Alcohol (ethanOl), Denatured
Alcohol, and Proprietary Solvent
O-T-634 Trichloroethylene, Technical
CCC-C-440 Cloth, Cheesecloth, Cotton Bleached and
Unbleached
Military

MIL-E-12397 Eraser, Rubber Pumice, for Testing Coated

Optical Elements
MIL-E-13830 Optical Components for Fire Control
Instruments; General Specification Govern-
ing the Manufacture, Assembly and Inspection of
MIL- I -45607 Inspection Equipment, Acquisition,
Maintenance and Disposition of

Beneficial comnlents (recommendations, additions. deletions., and any pertinent data

which may be of usc in improving this document. should be addressed to: Commander.
US Anny Armament Research and Development Command, ATTN: DRDAR-TST-S,
Dover, New Jersey 07801, by using the self-addressed Standardization Document
Improvement Proposal (DD Form 1426). appearing at the end of this document, or
by letter.

.... u s. GOVERN~ENT PRINTING OFFICE. 1990-703·023,4247 FSC 66S0


224 APPENDIX B

MIL-C-48497A

STANDARDS

Military

MIL-STD-lOS Sampling Procedures and Tables for

Inspection by Attributes
MIL-STD-l09 Quality Assurance Terms and Definitions
MIL-STD-124l Optical Terms and Definitions

DRAWINGS

U. S. Army, Armament Research &Development Command

7641866 Surface Quality Standards for Optical

Elements
7680606 Coating, Eraser Abrasion Tester

(Copies of specifications, standards, drawings and publications required

by suppliers in connection with specific procurement functions should be
obtained from the procuring activity or as directed by the contracting
officer. )

3. REQUIREMENTS

3.1 Optical terms and definitions. Terms and definitions peculiar

to the general field of optics as used herein are defined in MIL-STD-124l.

3.2 Coated area. Optical components shall be coated over their

entire clear aperture.

3.3 Coating quality. The coating shall be uniform in quality and

condition and shall conform to the following:

3.3.1 Physical. The coating shall have no evidence of flaking, peeling,

cracking or blistering.

3.3.2 Cosmetic. When specified the coating shall conform to the cosmetic
requirements (stains, smears, discolorations, streaks, cloudiness, etc.)
stated on the component drawing or other documents.
3.3.3 Environmental and solubility blemishes. The coated surface shall
be free of blemishes such as stains, smears, discolorations, streaks, cloudi-
ness, etc., that would cause non-conformance to the coating's spectral
requirements as stated on the component drawing or other documents.

3.3.4 Spatter and holes. Coating spatter and holes in the coating
shall be considered as a dig and shall not exceed the allowable dig size
and quantity stated on the componet drawing or other documents.

2

SELECTED UNITED STATES MILITARY SPECIFICATIONS 225

MIL-C-48497A

3.3.5 Surface defects (scratch and dlg). Coating scratches and digs
shall not exceed the values specified for the substrate on the component
drawing or other documents. Coating scratches and digs shall be considered
separate from the substrate scratch and dig requirements.

3.3.5.1 Transparent coated surfaces. Scratch and dig requirements for

transparent coated surface shall be specified by two (2) numbers separated
by a hyphen (i.e. 60-40) in accordance with MIL-O-13830.

3.3.5.2 Opaque coated surfaces. Scratch and dig requirements ~r

opaque coated surfaces shall be specified by two letters separated by a
hyphen (i.e. F-C). The first letter of the pair is the maximum scratch
value; the second letter is the maximum dig value.

3.3.5.2.1 Scratches. The scratch letter defines the width of the

scratch in accordance with the following table:

SCRATCH LETTER SCRATCH WIDTH(MICRONS) SCRATCH WIDTH (INCHES)

A 5 .00020
B 10 .00039
C 20 .00079
D 40 .00157
E 60 .00236
F 80 .00315
G 120 .00472

3.3.5.2.1.1 Density of maximum size scratches. The accumulated length

of all maxlmum size scratches present shall not exceed 1/4 the average
diameter of the coated surface.

3.3.5.2.1.2 Density of all scratches. When a maximum size scratch

is present, the sum of the products of the widths designated by the scratch
letters times the ratio of their length to the diameter of the coated
surface shall not exceed one half the width specified by the scratch letter.
When a maximum size scratch is not present, the sum of the products of the
widths designated by the scratch letters times the ratio of their length
to the diameter of the coated surface shall not exceed the width specified
by the scratch letter.

3.3.5.2.2 Digs. The dig letter specifies the average diameter of the
dig in accordance with the following table:

DIG LETTER DIAMETER (MM) DIAMETER (INCHES)

A 0.05 .0020
B 0.10 .0039
C 0.20 .0079
0 0.30 .0118
E 0.40 .0158
F 0.50 .0197
G 0.70 .0276
H 1.00 0.394

3

226 APPENDIX B

MIL-C-48497A

The permissible number of maximum size digs shall not exceed one per each
20 millimeters of (.80 inches) diameter or fraction thereof on any single
coated surface. The sum of the diameters of all digs shall not exceed
twice the diameter of the maximum size specified by the dig letter per 20
millimeters of diameter.

3.4 Coating durability.

3.4.1 Environmental and pyhsical durability. The coated optical

surface shall meet the following service conditions in the order specified:
3.4.1.1 Adhesion. The coated optical surface shall show no evidence of
coating removal when cellophane tape is pressed firmly against the coated
surface and quickly removed at an angle normal the the coated surface.

3.4.1.2 Humidity. After exposure in an atmosphere of 1200 ~ 40 F and

95 to 100\ relative humidity, the coated optical surface shall meet the
requirements of 3.3.1 and 3.3.3.
3.4.1.3 Moderate abrasion. The coated optical surface shall show no
signs of deterioration such as streaks or scratches when abraded with a dry,
clean cheesecloth pad.

3.4.2 Thermal and cleaning durability. The coated optical surface

shall meet the following conditions:

3.4.2.1 Temperature.
The coated optical surface shall be exposed to
~emperatures of -80DF and +160 0 F for 2 hours at each temperature.The rate
of temperature change shall not exceed 40 F per minute. Subsequent to these
exposures, the coated optical surface shall meet the requirements of 3.3.1
and 3.4.1.1.

3.4.2.2 SOlubilit* and cleanability. After immersion in trichloroethYlene,

acetone and ethyl alco 01 and wiping with cheesecloth, the coated optical
surface shall show no evidence of coating removal or scratches and shall
meet the requirements of 3.3.1 and 3.3.3.

3.4.3 ~;;.,;.~~~~~~r..,.:r:.;e~u~i'i-r.:;em~e~n:o't::.;s~.. When specified on the component

t e following requirements shall be

3.4.3.1 Severe abrasion. Abrasion by an eraser conforming to MIL-E-

12397 shall not cause deterioration such as streaks or scratches on the
coated optical surface.
3.4.3.2 Salt solubility. After immersion in a saline solution, the
coated optical surafce shall meet the requirements of 3.3.1 and 3.3.3.

4
SELECTED UNITED STATES MILITARY SPECIFICATIONS 227

MIL-C-48497A

3.4.3.3 Water solubility. After immersion in distilled water, the

coated surface shall meet the requirements of 3.3.1 and 3.3.3.

3.5 Optical. The coated component shall conform tothe pertinent

optical requirements specified on the component drawing or other documents
(See 6.2).

4. QUALITY ASSURANCE PROVISIONS

4.1 Responsibility for inspection. Unless otherwise specified in the

contract or purchase order, the supplier is responsible for the performance
of all inspection requirements as specified herein. Except as otherwise
specified, the supplier may utilize his own facilities or any other facilities
suitable for the performance of the inspection requirements specified here-
in, unless disapproved by the Government. The Government reserves the right
to perform any of the inspections set forth in the specification where such
inspections are deemed necessary to assure that supplies and services con-
form to prescribed requirements (See 6.2).

4.1.1 General provisions. Definit;ans of inspections terms shall be

as listed in MIL-STD-109.

4.1.2 Witness piece. Unless oth~rwise specified, witness pieces as

defined in 4.1.3 or coated componets may be used to test the optical and
durability requirements of the coated component. (See 4.2, 4.3 &6.2).
The witness pieces shall be positioned in the coating chamber such that
they represent the optical and durability characteristics of the whole
evaporated lot (See 6.4). The Government reserves the right to test the
actual coated component with the same test to which the witness pieces
were subjected. Should a component fail, even though the witness pieces
pass the test, the lot shall be rejected.

4.1.3 Characteristics of the witness piece. When witness pieces are

used to test the optical and durability requirements of the coated component
they shall exhibit the following characteristics:

a. The witness piece shall have the same refractive index and absorption
coeffecient as the componets to be coated.

b. The witness piece shall have a surface finish similar to that of

the component to be coated.

c. The witness piece shall be such that it presents no difficulty in

measuring and testing the optical and durability requirements of the coating.
d. Where transmission characteristics are required the witness piece
shall simulate the axial thickness of the component by either computational
methods or using a witness piece that has the same thickness as the coated
component.

5
228 APPENDIX B

MIL-C-48497A

4.2 First article (initial production) approval. The requirement for

first article approval and the responsibility (government or contractors) for
first article testing shall be as specified in the contract. (See 6.2).
Unless otherwise specified, the sample for first article approval tests
shall consists of five (5) coated components plus ten (10) coated witness
pieces (See 4.1.2 and 4.1.3). 'he 5 coated componets shall be tested,
as specified herein, for all the requirements of 3.2, 3.3.1, 3.3.2, 3.3.4
and 3.3.5. Five coated witness pieces shall be tested, as specified herein,
for all the requirements of 3.4.1 and 3,5. The other 5 coated witness
pieces shall be tested as specified herein for all the requirements of
3.~.2 and 3.5, The government reserves the right to subject the coated
components to all the tests specified here. The sample shall be caated
in the same manner, using the same materials, equipment, processes, and
procedures as used in regular production. All parts and materials including
packaging shall be obtained from the same source of supply as used in
regular production.

4.2.1 Government testing. When the Government is responsible for

conducting first article approval tests, the contractor, prior to submitting
the sample to the government, shall inspect the sample to insure that it
conforms to all the requirements of the contract and submit a record of
this inspection with the sample, including certificates of conformance for
materials, as applicable.

4.2.2 Contractor testing. When the contactor is responsible for

conducting first article approval tests, the sample shall be inspected by
the contractor for all the requirements of the contract. The sample and a
record of this inspection, including certificates of conformance for materials,
shall be submitted to the government for approval. The government reserves
the right to witness the contractor's inspection.

4.3 Inspection provisions.

4.3.1 Submission of product.

4.3.1.1 Inspection lot size. The inspection lot size shall consist of
all components or subassemblies (unit of product) coated within one (8 hour
or 10 hour) work shift.

4.3.1.2 Lot formation and presentation of lots. The unit of product

shall be submitted for inspection on a moving inspection lot basis where
the components or subassemblies are continuously offered for inspection in
the order produced.
NOTE: Each evaporation lot forming a part of a moving inspection lot
shall be identified for subsequent evaluation, if required.
4.3.2 Examination and tests.
4.3.2.1 Components and subassemblies. All coated componets and sub-
assemblies shall be inspected in accordance with this specification and

6
SELECTED UNITED STATES MILITARY SPECIFICATIONS 229

MIL-C-48497A

the inspection prov1s10ns contained in Supplementary Quality Assurance

Provisions (SQAP) listed in the Technical Data Package (TOP). Examination
and tests related to Section 3 herein shall be performed on a class basis
in accordance with MIL-STD-I05 and the sampling plans specified in Tables
I and II herein. The tabulated classification of defects in Tables I and
II shall constitute the minimum inspection to be performed by the supplier
after first article approval and prior to Government acceptance or rejection
by item or lot.

TABLE I - CLASSIFICATION OF DEFECTS FOR COMPONENTS

CLASS CHARACTERISTIC REQUIREI>lENT TEST PROCEDURE

CRITICAL: NONE DEFINED

MAJOR: AQL 1.0%

101. Coated area - ,

~.~ 4.5.1
102. Coating quality, physical 3.3.1 4.5.2.1
103. Coating quality, cosemtic 3.3.2 4.5.2.2
104. Coating quality. spatter & holes 3.3.4 4.5.2.4
105. Coating quality. scratch & dig 3.3.5 4.5.2.5
MINOR: NONE DEFINED

NOTE: The inspection for the characteristics in Table I shall be conducted

- - at a temperature between +60 o F and +90 o F.

4.3.2.2 Acceptance and rejection. Rejected lots shall be screened for

all defective characteristics. Removal of defective units and resubmittal
or rejected lots shall be in accordance with "Acceptance and Rejection" as
specified in MIL-STD-I05.

4.3.2.3 Special sampling.

4.3.2.3.1 Environmental and h sical durabilit. A minimum of one (J)

coated witness piece (or one coated component, when required) shall be
selected from each evaporation lot to form a minimum of five (5) samples for
each inspection lot. The samples shall meet the requirements and tests in
Table II.
230 APPENDIX B

MIL-C-48497A

TABLE II - ENVIRONMENTAL & PHYSICAL DURABILITY

NO. CHARACTERISTIC REQUIREMENT TEST PROCEDURE

30l. Adhesion 3.4.11 4.5.3.1

302. Coating quality, physical 3.4.1.2, 3.3.1 4.5.3.2, 4.5.2.1
(post. humidity)
303. Coating quality, blemishes 3.4.1.2, 3.3.3 4.5.3.2, 4.5.2.3
(post. humidity)
304. Moderate abrasion 3.4.1.3 4.5.3.3
305. Optical 3.5 4.5.6

NOTE: The inspection for the charact.eristics in Table II shall be conducted

at. a temperature between +60 0 F and +90 0 F.

4.3.2.3.2 Failure of sample. Should anyone item of a special sampling

fail to meet the specified test requirements, accept.ance of the represent.ed
inspection lot shall be suspended by the Government until necessary correc-
tions have been made by the contract.or and the resubmitted it.ems have been
approved. (See 4.3.2.2).

4.4 Inspection equipment. Except. as otherwise provided for by t.he

contract., the contract.or shall supply and maint.ain inspection equipment in
accordance with the applicable requirements of MIL-I-45607.

4.4.1 Government furnished inspection equipment. Where the cont.ract

provides for Government furnished test equipment, supply and maint.enance of
test equipment shall be in accordance with applicable requirements specified
in MIL-I-45607.

4.4.2 Contractor furnished inspection equipment.

4.4.2.1 Government design. All inspection equipment. specified by

drawing number in specifications or SQAP forming a part of the contract
shall be supplied by the contractor in accordance with technical dat.a
included in the contract.

4.4.2.2 Contractor equipment. The contractor shall supply inspection

equipment. compatible wit.h t.he "Test. Met.hods and Procedures" specified in
4.5 of this specification. Since tolerance of test equipment is normally
considered to be within 10\ of t.he product tolerance for which it is intended,
this inherent error in the test equipment must be considered as part. of the
prescribed product tolerance limit. Thus, concept, construct.ion, mat.erials,
dimensions and tolerances used in the test. equipment. shall be so selected and
controlled as to insure that the test. equipment. will reliably indicat.e
acceptability of a product which does not. exceed 90\ of the prescribed
tolerance limit., and permit positive rejection when non-eonforming.
Construct.ion shall be such as to facilit.at.e routine calibrat.ion of t.est
equipment.

8
SELECTED UNITED STATES MILITARY SPECIFICATIONS 231

MIL-C-48497A

4.5 Test methods and procedures.

4.5.1 Coated area. The coated area of the component or subassembly
shall be examined with standard measuring equipment (See 6.4). The coated
area shall conform to the requirements of 3.2.

4.5.2 Coating quality.

4.5.2.1 Physical. The coating shall be visually examined by reflection,

with the unaided eye, for evidence of flaking, peeling, cracking or blistering.
The examination shall be performed using two 15 watt cool white fluorescent
light tubes as the light source. The viewing distance from the coated surface
to the eye shall not exceed 18 inches. The coated surface shall be viewed
against a black matte background. The only illumination in the test area
shall be from the light source used for testing. This method of examination
is as depicted in Figure 1. The coating shall conform to the requirements
of 3.3.1.

4.5.2.2 Cosmetic. The coated component shall be examined using the

test method specified in 4.5.2.1 for evidence of discoloration, stains, smears,
streaks, cloudiness, etc. The coating shall conform to the requirements of
3.3.2.
4.5.2.3 Environmental and solubility blemishes. The coated witness
samples (or coated components, when required) shall be examined using the
test method specified in 4.5.2.1 for evidence of discoloration, stains, smears,
streaks, cloudiness, etc. The coating shall conform to the requirements of
3.3.3 When evidence of blemishes is found, the blemished area shall be
examined (See 4.5.6) for compliance to the spectral requirements specified
on the component drawing.

4.5.2.4 Spatter and holes. The coated components shall be examined

for dig size and quantity using the applicable test method specified in
4.5.2.5. The coating shall conform to the requirements of 3.3.4.

4.5.2.5 Surface defects (scratch and dig).

4.5.2.5.1 Transparent coated surfaces. Transparent coated surfaces
shall be examined utilizing the technique specified in 4.5.2.1, except that
(hole) in the coating shall be evaluated by comparison with the Surface
Quality Standards for Optical Elements (scratch &dig) Drawing 7641866.
The visual appearance of the scratch, and the diameter of the digs in the coating
shall conform to the requirements of 3.3.5.1. The lengths and density of all
scratches, and the diameters and density of all digs shall conform to the
requirements of MIL-O-13830.
4.5.2.5.2 Opaque coated surfaces. Opaque coated surfaces shall be
examined utilizing the lumination, reflection and viewing background technique
specified in 4.5.2.1. The length and width of scratches, and the dig (hole)
diameters shall be determined by use of interferometry, microscopic measuring

9

232 APPENDIXB

MIL-C-4849iA

CO~1MERC IAL
FIXTURE CONTAINING
r---- TWO COOL WHITE lS-WATr FLUORESCENT
LAMPS (SEE NOTE l)

~-----l8" foIAX-----=---••

BLACK MATrE
~CKGRO\JND

TEST COMPONENT
FOR REFLECTION
(SEE NOTE 2) ---

NOTES:
1. THE ONLY ILLUMINATION IN THE TEST AREA SHALL
BE FROM THE LIGHT SOURCE USED FOR TEST.
2. TILT AT AN APPROPRIATE ANGLE TO SEE THE
COATED SURFACE.

FIGURE

10
SELECTED UNITED STATES MILITARY SPECIFICATIONS 233

MIL-C-4B497A

devices, calibrated preclslon comparator~,or similar applicable preclsIon

measuring devices. The width, length and density of all scratches in the
coating shall conform to the requirements of 3.3.5.2. The diameter and
density of all digs (holes) in the coating shall cORform to the requirements
of 3.3.5.2.

4.5.3 Environmental physical durability tests. The coated components

or witness samples shall withstand exposure to the following test conditions:

4.5.3.1 Adhesion. The coated component or witness samples shall be

subjected to an adhesion test using 1/2" wide cellophane tape conforming to
Type I of L-T-90. Press the adhesive surface of the cellophane tape firmly
against the coated surface and quickly remove at an angle normal to the
coated surface. A visual inspection shall be made for conformance to
3.4.1.1. Subsequent to this test, the coated component or witness sample
shall be subjected to the test of 4.5.3.2.

4.5.3.2 Humidity. The coated componen~or witness samples shall be

placed into an environmentally controlled test chamber and exposed to a
temperature of 120 0 F and 95% to 100% relative humidity. rhe coating shall
be exposed for a minimum of 24 hours. Subsequent to this exposure, the coat-
ing shall be cleaned and evaluated for conformance to 3.3.1 and 3.3.3. The
coated compone~or witness sample shall then be subjected to the test
specified in 4.5.3.3.

4.5.3.3 Moderate abrasion. Within one hour after the humidity test
of 4.5.3.2, the coated component or witness sample shall be subjected to a
moderate abrasion test. It shall consist of rubbing a minimum of SO strokes
across the surface in straight lines. The abrader shall be a 1/4" thick by
3/8" wide pad of clean dry, laundered cheesecloth conforming to CCC-C-440.
The bearing force shall be a minimum of 1 pound and shall be applied approxi-
mately normal to the coated surface. The actual test apparatus shall be the
Eraser Abrasion Coating Tester of Drawing 76B0606 except that the eraser
portion shall be completely covered with cheesecloth and the cheesecloth
shall be secured to the shaft of the tester with an elastic band. Subsequent
to this test, the coating shall be evaluated for conformance to 3.4.1.3.
Following the tests of 4.5.3.1, 4.5.3.2 and 4.5.3.3 the coated component or
witness sample shall be evaluated for conformance to the provisions of 4.5.6.

4.5.4 Thermal and cleaning durability. The witness samples shall with-
stand exposure to the follGwing test conditions:

4.5.4.1 Temperature. The witness samples shall be subjected to

temperatures of _BouF • 2uF and +160 0 F +2 0 F for a period of 2 hours at each
temperature. After each exposure the coated componets or witness samples
shall be stabilized at an ambient temperature between +600 and +90 0 F and
subjected to the test specified in 4.5.3.1. Subsequent to this test, the
coating shall be evaluated for conformance to 3.4.1.3 and 3.3.1. The rate
of temperature change between any two temperature levels shall not exceed
40 F per minute.

11

234 APPENDIXB

MIL.C-48497A

4.S.4.2 Solubility and cleanability. The witness samples shall be

immersed, in sequence, in the following solutions maintained at room
temperature (+600 to +900 F): trichloroethylene conforming to O-T-634;
acetone conforming to O-A-Sl; and ethyl alcohol conforming to O-E-760.
The immersion time in each solution shall be a minimum of ten minutes.
Upon removal from each solution, the solvent shall be allowed to evaporate
to dryness without wiping or forced drying before proceeding to the next
solution. Upon removal from the alcohol solution, and after drying any
resultant stains on the coated surface shall be removed by wiping the coating
to a clean, stain-free condition with an ethyl-alcohol-mositened cheesecloth.
Subsequent to this cleaning, the coating shall be evaluated for conformance
to 3.4.2.2, 3.3.1, 3.3.3 and 3.S.

4.S.S Optional durability tests. The coated componets or witness

samples shall withstand exposure to the following test conditions when
specified in the contract, purchase order or on the componet drawing.
(See 6.2).

4.S.S.l Severe abrasion. The coated componets or witness samples

shall be subjected to a severe abrasion, rubbing the coated surface with a
standard eraser conforming to MIL-E-12397 mounted in the Eraser Abrasion
Coating Tester of Drawing 7680606. The bearing force shall be between
2 and 2 1/2 lbs. Applied approximately normal to the coated surface. A
total of 20 strokes shall be made along a straight line path. Subsequent
to this test the coating shall be evaluated for conformance to 3.4.3.1 and
then subjected to the test specified in 4.S.3.

4.S.S.2 Salt solubility. The coated componen~or witness samples

shall be immersed for a period of 24 hours in a solution of water and
sodium chloride (salt). The mixture shall be 6 ounces of salt per gallon
of water at room temperature (+60 to +90 oF) • . Subsequent to this immersion,
the coated componets or witness samples shall be removed from the solution,
dried with clean, laundered, cheesecloth, and then evaluated for conformance
to 3.3.1 and 3.3.3.

4.S.S.~ Water solubility. The coated componets or witness samples

shall be immersed for a period of 24 hours in distilled water at room
temperature (+60oF to +90 oF). Subsequent to this immersion, the coated
componets or witness samples shall be removed from the solution, dried with
clean laundered cheesecloth, and then evaluated for conformance to 3.3.1
and 3.3.3.
4.S.6 Optical. Appropriate optical instruments such as spectrophoto-
meters, radiometers, etc. shall be used to test conformance to the optical
requirements of 3.S.
S. PACKAGING
This section is not applicable to this specification.
SELECTED UNITED STATES MILITARY SPECIFICATIONS 235

MIL-C-48497A

6. NOTES

6.1 Intended use. This specification extablishes minimum quality and

durability requirements for single or multi-layer interference coatings
applied to components in optical systems. It is applicable to coatings
utilized on a wide variety of substrate materials for such items as lenses,
prisms, mirrors, beam splitters, reticles, laser elements, windows, filters,
etc.

Coatings produced to~e requirements of this specification may be

exposed without damage to those handling and cleaning procedures associated
with optical instrument assembly practice and will generally be employed
within the protective confines of a sealed instrument.

Certain applications may necessitate the use of additional or expanded

quality and durability requirements, some of which are included within the
optical durability section of this specification. (See 3.4.3). Other
quality and durability requirements with the appropriate test methods must
be defined when they are applicable. All aHditional requirements must be
specified on the applicable drawing or other documents. Examples of coating
applications where adaitional requirements must be specified include laser
systems where resistance to laser energy damage is necessary; also when
coatings are used outside the confines of a sealed instrument, resistance
to such environments as salt fog, fungus, extended humidity, etc. may be
necessary.

This specification is not intended as a substitute for other coating

specifications, e.g., MIL-C-675, MIL-C-14806, etc. It is intended that
this specification provide the durability requirements for coating
applications not covered by other documents and in some cases to serve as a
supplement, by reference, in other coating specifications.

6.2 Ordering data. Purchasers should exercise any desired options

offered herein, and procurement documents should specify the following:

a. Title, number and date of this specification.

b. Optional abrasion, salt solubility, and water solubility
requirements (see 3.4.3 and 4.5.5)
c. Optical requirements of coating (See 3.5)
d. Responsibility for performance of inspection requirements
(See 4.1)
e. Mandatory use of coated componets for tests (See 4.1.2)
f. Requirement for first article approval (See 4.2, 4.2.1 and
4.2.2)

236 APPENDIX B

MIL-C-48497A

6.3 First article sample. When the government is responsible for

conducting first article approval tests the required article sample
specified in 4.2 shall be forwarded to:

Commander
US Army Armament Research &Development Command
Ai"j'I: Dil.DAR-QAF
Deller, NJ 07801

6.4 Definitions.
6.4.1 Evaporation lot. An evaproation lot is defined as the group
of parts which has the coating applied at the same time and in the same
chamber.
6.4.2 Standard measurin e ui ment (5MB). Standard measuring
equipment is e in as the common measuring devices which are usually
stocked by commercial supply houses for ready supply (shelf items) and
which are normally used by an inspector to perform dimensional inspection
of items under procurement. This category also includes commercial testing
equipment such as meters, optical comparators, etc.

Custodian: Preparing Activity:

Army - AR Army-AR
Air Force - 11

Project No. 6650-0108

14


SELECTED UNITED STATES MILITARY SPECIFICATIONS 237

MIL-F-486l6
SUPPLEMENT 1
29 ·JULY 1977

MILITARY SPECIFICATION

FILTER (COATINGS), INFRARED INTERFERENCE:

GENERAL SPECIFICATION FOR

This supplement forms a part of Military Specification MIL-F-48616

Specification Sheets

MIL-F-48616/l00 Coating, Low Reflectance, Ordinary Requirements

MIL-F-48616/10l Coating, Low Reflectance on Germanium 8.0 to 11.S

M1crometers, High Durability, Special Requirements
for

MIL-F-486l6/102 Coating, Ultra-Low Reflectance on Germanium 7.S to

12.3 Micrometers, Special Requirements for

MIL-F -48616 /200 Coating, High Reflectance, First Surface, Ordinary

Requirements for

MIL-F-48616/300 Filter, Wide Bandpass, 7.7 to 11.7 Micrometers, Ordinary

Requirements for (not issued, reserved for future
application)

MIL-F-486l6/30l Filter, Wide Bandpass, 7.7 to 11.7 Micrometers,

Special Requirements for

MIL-F-48616/400 Filter, Narrow Bandpass, Ordinary Requirements for

(not issued, reserved for future application)

MIL-F-486l6/S00 Coating, Long Wavelength Pass Filter, Ordinary

Requirements for

MIL-F-486l6/600 Coating, Short Wavelength Pass Filter, Ordinary

Requirements for

Preparing activity:

Artily - MU

FSC-66S0
238 APPENDIX B

MI1-F-48616/100
29 JULy 1977
MILITARY SPECIFICATION SHEET

COATING, LOW REFLECTANCE, ORDINARY REQUIREMENTS FOR

The complete requirements for procuring the low reflectance coating described
herein shall consist of this document and the latest issue of Specification
MI1-F-48616.

This Specification is approved for use by all Depart-

ments and Agencies of the Department of Defense.

Spectral reflectance requirements for low reflectance coatings are given in

Table I. Maximum reflectance values for each surface are tabulated for
ranges of substrate index of refraction and for several wavelength ranges.

TABLE I

LOW Reflectance Coatings - Maximum Percent Ref1ectance* per Surface

Index of Refraction Wavelength Radio Reflectance Dash No.

of Substrate A2/Al

1.30 to 1. 75 1.0 0.3 1

1.30 to 1. 75 1.0 to 1.2 0.4 2
1.30 to 1.75 1.2 to 1.5 0.6 3
1.30 to 1. 75 1.5 to 2.0 1.5 4

1. 75 to 3.00 1.0 0.3 5

1.75 to 3.00 1.0 to 1.2 0.4 6
1.75 to 3.00 1.2 to 1.5 1.0 7
1. 75 to 3.00 1.5 to 2.0 4.0 8

3.00 to 4.20 1.0 1.0 9

3.00 to 4.20 1.0 to 1.2 1.5 10
3.00 to 4.20 1.2 to 1.5 3.0 11
3.00 to 4.20 1.5 to 2.0 5.0 12

*For 'A2/'Al > 1.0 maximum average reflectance between Ai and 'A2

NOTES ON TABLE I:
1. The wavelengths Ai & 'A2 define the spectral region of interest.

2. The values given in this table apply at an angle of incidence ranging

from 00 to 150 and ambient temperature. Reflectance measurements a~~ to
0
be performed at an angle of incidence between QP and 15 (See 6.4b and
6.4c) of MIL-F-48616. When measurements are made at larger angles, the
data shall be corrected to normal incidence.

FSC-6650
SELECTED UNITED STATES MILITARY SPECIFICATIONS 239

KIL-F-48616/100

3. For A2/Al = 1.0 spectral resolution required to verify the requirements

in Table I shall be 0.5% or less of the specified wavelength. For A2/Al
>1.0, the spectral resolution shall be 1.0% or less of (A2 + Al)/2.
Scan speed and time constant shall be set such that the spectral
features of interest can be resolved within the accuracy of 4.4.2.3.2.2
of KIL-F-48616.

Ordering Information:

Anti-reflection coatings are used primarily to reduce reflections within

optical systems. Performance is strongly dependent on substrate material
and wavelength region. For applications for which these coatings are
used to enhance transmission throught an optical element by reducing
reflection losses at substrate surfaces, the user should specify the
required transmission. The user should be aware of the possibility of
absorption within these coatings and within the substrate material, and
should therefore specify the absorption requirement if applicable.

Ordering Data:

At a minimum, the substrate (See 6.2d) (nominal refractive indeK value)

to be coated and the wavelength region of interest (See 6.2m) (Al & A2)
and the dash number (See 6.2b) shall be specified by the procuring
activity (See 6.2).

Preparing activity:

Army - MlJ

Project No. 6650-0091-1

2

240 APPENDIX B

MIL-F-48616/101
29 JULY 1977

MILITARY SPECIFICATION SHEET

COATING, LOW REFLECTANCE ON GERMANIUM, 8.0 to 11.5, MICROMETERS,

HIGH DURABILITY, SPECIAL REQUIREMENTS FOR

The complete requirements for procuring the low reflectance coating described
herein shall consist of this document and the latest issue of Specification
MIL-F-48616.

This specification is approved for use by all Depart-

ments and Agencies of the Department of Defense.

Spectral requirements exclusive of substrate absorptance for the low

reflectance coating on germanium are given the follOWing paragraphs:

1. The average transmittance through two coated surfaces shall be greater

than 937. over the 8.0 to 11.5 micrometer region.

2. The minimum transmittance through twO coated surfaces at any paint

between 8.0 and 11.5 micrometers shall not be less than 907..

3. The average reflectance shall not be greater than 3.07. per surface
over the 8.0 to 11.5 micrometer region.

4. The maximum reflectance at any point between 8.0 and 11.5 micrometers
shall not be greater than 5.0% per surface.

Additional Durability Reguirements:

In addition to the minimum acceptable durability requirements of MIL-F-48616

the coated element shall meet all of the optional durability requirements
stated in 3.4.2.3.

!!Ql!:

Spectral resolution required to verify the above requirements shall be 0.1

micrometer or less. Scan speed and time constant shall be set such that
scan time is two (2) minutes or greater for transmittance measurements
and ten (10) minutes or greater for reflectance measurements.

Ordering Data:

Not applicable.

Preparing activity:
Army-MU
Project No. 6650-0091-2

FSC-6650
SELECTED UNITED STATES MILITARY SPECIFICATIONS 241

MIL-F -48616 /102

29 JULY 1977

MILITARY SPECIFICATION SHEET

COATING, ULTRA-LOW REFLECTANCE ON GERMANIUM,

7.5 TO 12.3 MICROMETERS, SPECIAL REQUIREMENTS FOR

The complete requirements for procuring the ultra-low re£lectance coating

described herein shall consist of this document and the latest issue of
MIL-F-486l6.

This specification is approved for use by all Depart-

ments and Agencies of the Department of Defense.

Requirements exclusive of substrate absorptance for the ultra-low reflectance

coating on germanium are given the following paragraphs:

1. The average transmittance through two coated surfaces shall be equal

to or greater than 98% over the 7.5 to 11.5 micrometer region.

2. The average reflectance shall not be greater than 0.3% per surface
over the 8.0 to 11.5 micrometer region.

3. The average reflectance shall not be greater than 0.9% surface over
the 7.5 to 8.0 micrometer region and 1.3% per surface over the 11.5 to
12.3 micrometer region.

!QI!:

Spectral resolution required to verify the above requirements shall be 0.1

micrometer or less. Scan speed and time constant shall be set such that
scan time is two (2) minutes or greater for transmittance measurements
and ten (10) minutes or greater for reflectance measurements.

Ordering Data:

Not applicable.

Preparing activity:
Army - MlJ
Project No. 6650-0091-3

242 APPENDIXB

MIL-F-486l6/200
29 JULy 1977

MILITARY SPECIFICATION SHEET

COATING, HIGH REFLECTANCE, FIRST

SURFACE, ORDINARY REQUIREMENTS FOR

The complete requirements for procuring the first surface high reflectance
coating described herein shall consist of this document and the latest
issue of Specification MIL-F-486l6.

This specification is approved for use by all Depart-

ments and Agencies of the Department of Defense.

Requirements for first surface high reflectance coatings are given in

Table I. The table provides the minimum allowable reflectance for
several wavelength ratios.

TABLE I

First Surface High Reflectance Coatings - Minimum Reflectance per Surface

Dash
Wavelength Radio A2/ Al Minimum Allowable Reflectance. No.

1 to 1.05 99 1

1.05 to 1.12 98 2

1.12 to 2 97 3

2 96 4

*For },2/},I, 1.0 maximum average reflectance between},l and A2

Nons F'JR TAlILE I:

1. The wavelengths 'AI and A2 are the extreme values of the wavelength
region of interest.

2. The values given in this table apply at an angle of incidence ranging

from 00 to 150 and ambient temperature. Reflectance measurements are to
be performed at an angle of incidence between CO and 150 (See 6.4b and
6.4c) of MIL-F-486l6. When measurements are made at larger angles, the
data shall be corrected to normal incidence.

Ordering Data:

The wavelength region of interest (See 6.2m) (AI & A2) and the dash
number (See 6.2b) shall be specified by the procuring activity.

Preparing activity:

Army - MlJ

Project No. 6650-009l~

FSC-6650
SELECTED UNITED STATES MILITARY SPECIFICATIONS 243

MIL-F-486l6/30l
29 JULY 1977
MILITARY SPECIFICATION SHEET

FILTER, WIDE BANDPASS, 7.7 to 11. 7 MICROMETERS,

SPECIAL REQUIREMENTS FOR

The complete requirements for procuring the wide bandpass filter described
herein shall consist of this document and the latest issue of MIL-F-486l6.

This s ecification is approved for use by all Depart-

ments and Agencies of t e Department 0 De ense.

Requirements for the wide bandpass filter (coated substrate) are given
in the follOWing paragraphs:

1. Average transmittance shall be equal to or greater than 80% from 8.3

to 11.1 micrometers.

2. Absolute 45% transmittance points shall be at wavelengths than 7.7

micrometers and less than 11.7 micrometers.

3. Absolute 5% transmittance points shall be at wavelengths than 7.5

micrometers and less than 12.1 micrometers.

4. Out of band transmittance shall be:

a. Less than 0.1% average from 1.0 to 7.0 micrometers and from 12.6
to 15.0 micrometers.

b. Less than 1.0% absolute from 1.0 to 7.0 micrometers and from 12.6
to 15.0 micrometers.

c. Less than 5% absolute from 7.0 to 7.5 micrometers and from 12.1 to
12.6 micrometers.

~:

1. Spectral resolution required to verify the above requirements shall be

0.03 micrometer or less.
2. Scan speed and time constant shall be set such that scan time is ten
(10) minutes or greater fom 7.0 to 12.6 micrometers and twenty (20) minutes
or greater total time for the two regions from 1.0 to 7.0 micrometers and
12.6 to 15.0 micrometers.

3. The spectral requirements stated above apply at an angle of incidence

ranging from 00 to 50 and at ambient temperature. Measurements will be
made at an angle of incidence within this range ( See 6.4b and 6.4c).

FSC-6650
244 APPENDIX B

MIL-F-48616/301

Ordering Data:

Not applicable.

Preparing activity:

Army-MU

Project No. 6650-0091-5

2
SELECTED UNITED STATES MILITARY SPECIFICATIONS 245

MIL-F-486l6/500
29 JULy 1977
MILITARY SPECIFICATION SHEET

COATING, LONG WAVELENGTH PASS FILTER,

ORDINARY REQUIREMENTS FOR

The complete requirements for procuring the long wavelength pas~ filters
described herein shall consist of this document and the iatest issue of
MIL-F-486l6.

This specification is approved for use by all Departments

and Agencies of the Department of Defense.

The follOWing figure illustrates the terms used for describing long
wavelength pass filters.

i ,_, I
-~
-- ~/- - -:- ' - + - ' . \ /_. Average
Transmittance
I j'

5%

AC ~l = 1.07A<;.
Wavelength
NOTES FOR FIGURE:

1. The cut-on wavelength (A c ) is the wavelength at which the ahsolute

transmittance is 5%.

2. Average transmittance is calcuated between AI, 6. A2 where U =

1.07 Ac and A2 is specified by the user.

FSC-6650

246 APPENDIX B

KIL-F-486l6/500

Long Wavelength Pass Filter Requirements

TABLE I

Cut-on Longest Wavelength Minimum Average Dash

Wavelength of Transmitting Region Transmittance No.

0.75 to 1.1 um 85% 1

0.7 to 1.1 UBI 1.1 to 2.2 um 80% 2
2.2 to 3.5 um 75% 3

1.18 to 1. 8 um 85% 4
1.1 to 1.8 um 1.8 to 3.0 um 80% 5
3.0 to 5.0 um 75% 6

1.93 to 3.0 um 80% 7

1.8 to 6.0 um 3.0 to 5.0 um 75% 8
5.0 to 8.0 um 70% 9

6.45 to 12.0 um 80% 10

6.0 to 14.0 um 12.0 to 14.0 um 75% 11
14.0 to 20.0 UBI 65% 12

15.0 to 22.0 UBI 70% 13

14.0 to 25.0 um 22.0 to 30.0 um 60% 14
30.0 to 50.0 um 50% 15

26.9 to 30.0 um 60% 16

25.0 to 45.0 UBI 30.0 to 50.0 UBI 50% 17

NOTES FOR REQUIREMENTS TABLE:

shall be located to within t 2%

1. The 5% absolute transmittance point Ac
of the specified wavelength.

2. Transmittance shall be equal to or less than 0.1% absoulte from 0.3 um

to 0.93 Ac.

3. The values in this table apply at an angle of incidence ranging from 00

to 50 and at ambient temperature. Measurements will be made at an angle of
incidence within this range (See 6.4b and 6.4c).

2
SELECTED UNITED STATES MILITARY SPECIFICATIONS 247

MIL-F-486l6/500

4. Spectral resolution to verify the above requirements in the cut-on

and transmitting regions shall be 0.5% or less of the cut-on wavelength.
Resolution shall be 10% or less of the cut on wavelength in the attenuation
region. Scan speed and time constant shall be set such that the spectral
features of interest can be resolved within the accuracy of 4.4.2.3.2.2
of MIL-F-48616 •

Ordering Data:

At a minimum, the cut-on wavelength (See 6.2m) ( A ) and the longest wave-
length o~ the transmitting region (See 6.2m) (A )c and the dash number
(See 6.2b) shall be specified by the procuring ictivity.

Preparing actitity:

Army - MIJ

Project No. 6650-0091-6

3

248 APPENDIX B

MIL-F-48616/600
29 JULY 1977

MILITARY SPECIFICATION SHEET

COATING. SHORT WAVELENGTH PASS

FILTER. ORDINARY REQUIREMENTS FOR

The complete requirements for procuring the short wavelength pass filters
described herein shall consist of this document and the latest issue of
MIL-F-4861"P.'

This specification is approved for use by all Departments

and Agencies of the Department of Defense.

The following figure illustrates the terms used for describing short wave-
length pass filters.

%T

-. '--.'-"'~/
Average
Transmi t tance

I
YI
I
I
I
5'7. I
0.6). c 0.93 ~c " c 1. ~c

Wavelength

NOTES FOR FIGURE:

1. The cutoff wavelength (~c ) is the wavelength at which the absolute
transmittance is 5%.

2. Average transmittance is determined between 0.6 ~c and 0.93 ~c •

FSC-6650


SELECTED UNITED STATES MILITARY SPECIFICATIONS 249

MIL-r-48616/600

Short Wavelength Pass Filter Spectral Requirements

TABLE I

Cut-Off Minimum Average Dash

Wavelength (A C ) Transmittance No.

0.9 to 15.0 um 75% 1

15.0 to 25.0 um 70% 2
25.0 to 35.0 um 60% 3
35.0 to 50.0 um 50% 4

NOTES ON SPECTRAL REQUIREMENTS:

1. The 5% absolute transmittance pOint shall be located to within 12%

of the specified wavelength.

2. Transmittance shall be equal to or less than 0.1% absolute from

to 1.4 AC •
1.07 AC

3. The values in this table apply at an angle of incidence ranging from

00 to 50 and at a~inet temperature. Measurements will be made at an
angle of incidence within this range (See 6.4b and 6.4c).

4. At wavelengths longer than 1.4 AC short wavelength pass filters may

have high transmittance.

5. Spectral resolution to verify the above requirements in the trans-

mitting and cut off regions shall be 0.5% or less of the cut off wave-
length. Resolution shall be 10% or less of the cut off wavelength in
the attenuation region. Scan speed and time constant shall be set such
that the spectral features of interest can be resolved within the
accuracy of 4.4.2.3.2.2 of MIL-F-48616.

Ordering Data:

At a minimum the cut off wavelength (See 6.2m) ( AC ) and the dash number
(See 6.2b) shall be specified by the procuring activity.

Preparing activity:

Army - MU

Project No. 6650-0091-7

250 APPENDIX B

HIL-F-48616
29 JULy 1977
MILITARY SPECIrlCA1'ION

FILTER (COATINGS), INFRARED INTERFERENCE:

GENERAL SPECIrICATION FOR

This specification is approved for use by all Departments and Agencies of the
Department of Defense.

1. SCOPE

1.1 ~.- This specification establishes general performance and

durability requirements for thin film coated optical elements which are used
in the spectral region from 0.7 am to 50.0 am (See 6.1).

2. APPLICABLE DOCUMENTS

2.1 The following docu=ents, of the iSlue in effect on date of invitation

for bids or requests for proposal, form a part of this specification to the
extent specified herein.

SPECIFICATIONS

Federal

L-T-90 Tape, Pressure-Sensitive Adhesive Cellophane and

Cellulose Acetate
CCC-C-440 Cloth, Cheesecloth, Cotton Bleached and Unbleached

Military

MIL-E-12397 Eraser, Rubber-Pumice for Testing Coated Optical

Elements
MIL-I-4S607 Inspection Equipment, Acquisition, Kaintenance and
Disposition of

STANDARDS

MIL-STD-IOS Sampling Procedure and Tables for Inspection by

Attributes
MIL-STD-I09 Quality Assurance Terms and Definitions
MIL -STD -1241 Optical. Terms and Definitions
Beneficial comments (recommendations, additions, deletions) and any
pertinent data which may be of use in imprOVing this document should
be addressed to: Commander, U. S. Army Armament Research & Development
Command, ATTN: DRDAR-TST-S, Dover, R. J. 07801 by using the aelf-addressed
Standardization Document Improvement Proposal (DD Form 1426) appearing
at the end of this document or by letter.

FSC 66S0
SELECTED UNITED STATES MILITARY SPECIFICATIONS 251

KlL-F-48616

DRAWINGS

U. S. ~, Frankford Arlenal

D7680606 Coatinl, Eraaer Abralive Telter

2.2 Other puhlicationa.- The follovinl documents form a psrt of this

specification to the extent specified herein. Unless otherwise iDdiceted,
the issue in effect on date of invitation for hids or request for proposal
shan apply:

AMERICAN SOCIETY FOR TESTING AND MATERIALS

Bl17 - Standard Method of Salt Spray (fog) Testing

(Applications for copies should be addressed to the American Society for

Testing and Materials, 1916 Race Street, Philadelphia, Pa. 19103. Technical
society and technical association specifications and standards are senerally
available for reference from libraries. They are also distributed among
technical groups and using Federal Asencies.)

3• REQUIREMENTS

3.1 Specificstion sheets.- Infrsred filters and coattals ahall comply

with the requirements of this specification and the applicable specification
sheet except as further defined in applicable dravinss or procurement document.

3.1.1 Optical terms and definitions.- Terms and definitions are as

used herein and defined in MIL-STD-l241 and 6.3 herein.

3.1.2 Detail specifications.- Specification eheets are categorized as

shown in the supplement. As additional specification sheets are senerated
they will be placed in the appropriate class el. the first additional
specification sheet applicable to the class '~ilter, Wide Bandpaa., 7.7 to
11.7 Kicrometers, Special Requirements for" shan be inserted as KIL-F-
48616/301. Additional class deSignations, when necessary. shall be assigned
outside this initial allocation in ascandinl order i.e. MIL-F-48616 - 700
to 799 would be allocated to the first class addition required.

3.2 Spectral.- The minimum spectral requirements for infrared filters

and coatings are specified in the specification sheets. The spectral values
for the wavelength. specified herein are given a. mea.ured in vacuum. (See
6.4). When specified on the component dravinl or the procurement document
(See 6.2c) the coated elament shall meet the spectral requirements after
environmental and durability testing.

2
252 APPENDIX B

MIL-F-48616

3.3 Substrate.- As a minimum, when the aubstrate material is indicated

on the component drawing or procurement document (See 6.2d), the requirements
of 3.3.1, 3.3.2 and 3.3.3 shall be specified. When the lubstrate material
and quality (internal and surface defects) are Dot lpecified, the aubstrate
shall be chosen so that the optical component ahall meet the requirements of
3.2 and 3.4.

3.3.1 Internal defects.- Prior to coating, .ubltrate materials which

are transparent or semi-transparent to visible radiation Ihall be evaluated
for internal defects, such as bubbles and other defects which are essentially
round in nature. These defects shall not exceed the dig requirements
specified for surface defects specified in 3.4.1.3. The aubstrate materials
shall also be evaluated for other internal defects auch a. striae, fractures,
incluSions, etc. The limits on tbese defects ahall be specified by the
procurement agency (See 6.2e). For lubltrate material a which do not transmit
visible radiation, the method for evaluation of the internal quality of the
materials (See 6.4d) ahall be apecified by the procurement agency.

3.3.2 Surface defects.- Surface defects on the subatrate shall be

such that the coated component does not have defects in excess of the
surface quality requirements of 3.4.1.2 and 3.4.1.3.

3.3.3 Dimensions.- The substrate ahall meet the dimenaional require-

ments of the applicable drawings or procurement document prior to coating
(See 6.4e).

3.4 Coated component.- AS a minimum, the coated component shall meet

the following requirements.

3.4.1 Surface qualitx.

3.4.1.1 Coating.- The coating shall show no evidence of flaking,

peeling, cracking, fingerprints, brush marka, out gaaaing, blistering,
back-coating, crazing, etc. Spatter and holes on or in the coating shall
be considered as a dig and shall not exceed the allowable dig size and
quantity. (See 3.4.1.3).

3.4.1.2 Scratches.- Surface scratches (coating and substrate) ahall

not be in excess of the values specified On the component drawing or procure-
ment document (See 6.2f). Scratches are permissible provided the Width does
not exceed that specified by the scratch letter. The accumulated length
of all maximum scratches shall not exceed 1/4 of the average diameter of
the element. The scratch letter and correaponding width are shown in
Table 1.

3.4.1.2.1 Integrating scratches.- Where the acratches (coating and

substrate) do not exceed the requirements of 3.4.1.2, each aurface shall
be evaluated further by integrating scratches. All scratches of widths
less than or equal to the maximum allowable scratch width, and greater
than or equal to the minimum scratch width to be conSidered, (See Table 1)
shall be includec in the integration. The length of each scratch shall be

3
SELECTED UNITED STATES MILITARY SPECIFICATIONS 253

MIL-F-486l6
multiplied by the scratch width. These products are to be added and the
sum divided by the average diameter of the element. If a maximum scratch
is present, this resulting value shall not exceed 1/2 the maximum allowed
scratch width. If no maximum scratch is present, this value shall not
exceed the maximum allowed scratch width.

3.4.1.3 Digs.- Surface digs (coating and substrate) shall not be in

excess of the values specified on the component drawing or procurement
doc~nt (See 6.2g). Digs are permissible on a surface provided the average
dia~ter does not exceed that specified by the dig letter and no more than
(1) maximum size dig occurs in any 20an (O.BO") diameter circle on the
substrate. The dig letter and corresponding sverage diamter are shown in
Table II.

3.4.1.3.1 Integrsting di8s.- Where the digs (coating snd substrate)

do not exceed the requirements of 3.4.1.3, each surface shall be evaluated
further by integrating digs. All digs of diameters less than or equal to
the maximum allowable dig diameter and greater than or equal to the minimum
dig diameter to be conSidered (See Table II) shall be included in the
integration. All digs shall be accumulated such that the sum of the
diameters does not exceed twice the diameter of the maximum allowed dig
for any JrOrrm (O.80") diameter circle on the surface. All digs of size B
or small shall be separated by 1.0II1II (0.04") minimum. the measurement of the
distance" between digs shall not be required for surfaces where digs larger than
size B are allowed.

TABLE I

SCRATCH IDENTIFICATION
(See 6.3.1)
Oisre8ard Scratch
Scratch Letter scratch Width Widths less than
Millimeters ~ Mill imeters .ill!!!!.
.005 .00020 .0010 .00004
A .00010
B .010 .00039 .0025
C .020 .00079 .0050 .00020
.040 .00157 .0100 .00039
0
E .060 .00236 .0100 .00039
.080 .00315 .0200 .00079
F
G .120 .00472 .0200 .00079

4
254 APPENDIX B

KIL-F-486l6

tABLE II

DIG IDENtIFICAtION
(See 6.3.1)
Disregard Digs
AveraBe Dig Diameter Su11er than
Dig Letter Millimeters ~ Millimeters Inches

A .05 .0020 .010 .0004

B .10 .0039 .025 .0010
C .20 .0079 .050 .0019
D .30 .0118 .050 .0019
E .40 .0157 .100 .0039
F .50 .0197 .100 .0039
G .70 .0276 .200 .0079
H 1.00 .0394 .250 .0099

3.4.1.4 Cosmetic.- No blemishes (coating and substrate) such as

streaks, smears, stains, blotchiness, discoloration, etc. shall be permitted
On an optical component lying in a focal plane. Unles. otherwise specified
on the component drawing or procurement document (See 6.2h). blemishes on
a component which lies outSide the foeal plane in an optical system shall
be acceptable when it can be shown that these blemishes do not impair the
spectral performance and durability requirements.

3.4.1.5 Coated area.- Optical components shall be coated over their

entire clear aperture. In those instances where the clear aperture is not
specified on the component or procurement document, the following allowable
.uncoated area shall apply:

Diameter or Maximum
Diagonal Dimension of Maximum Width of
Optical Element Uncoated Area

Up to 2" 0.040"
Greater than 2" 0.040" pluS additional
width of 0.015" for each
inch in diagonal greater
than 2"
or or

Up to 5CM 1MM
Greater than 5CM lMM plus additional width
increasing at the rate of
0.15MM far uch CM over
5CM
tooling marks are measured perpendicular from the edge of the components.
In those instances where the clear aperture is specified, areas outside the
clear aperture may be coated at the discretion of the contractor.

3.4.2 Surface durability (coating and substrate).

5
SELECTED UNITED STATES MILITARY SPECIFICATIONS 255

MlL-F-48616

3.4.2.1 Environmental and physical durability.- The coated optical

surface shall meet the following service conditions in the order specified:

3.4.2.1.1 Adhesion.- The costed optical surface shall show no evidence

of coating removal when cellophane tape is pressed firmly against the coated
surface and quickly removed at an angle normal to the coated surface:

3.4.2.1.2 Humidity.- After exposure in an atmosphere at 1200 t 40F

(49 0C) and 95 to 100% relative humidity, the coated optical surface shall
meet the requirements of 3.4.1.1 and 3.4.1.4.

3.4.2.1.3 Moderate abrasion.- The coated optical surface shall show

no signs of deterioration such as streaks or scratches when abraded with a
dry, clean cheesecloth pad.

3.4.2.2 Thermal and cleaning durability.- The coated optical surface

shall meet the following conditions:

3.4.2.2.1 Temperature.- The coated optical surface shall be exposed

to temperatures of -800F and +1600Ff~oc and 71 0C) for 2 hours at each
temperature. The rate of temperature change shall not exceed 40F (2°C)
per minute. Subsequent to these exposures, the coated optical surface
shall meet the requirements of 3.4.1.1 and 3.4.2.1.1.

3.4.2.2.2 solubility and cleanability.- After immersion in trichlo-

roethylene, acetone and ethyl alcohol and wiping with cheesecloth, the
coated optical surface shall show no evidence of coating removal or
scratches and shall meet the requirements of 3.4.1.1 and 3.4.1.4.

3.4.2.3 Optional durability requirements.- When specified on the

component drawing or other documents (See 6.2j) the following require-
ments shall be added or substituted:

3.4.2.3.1 Severe abrasion.- There shall be nO viSible damage, such

as evidence of abrasion or coating removal, to the coated surface when
abraded by an eraser conforming to KIL-E-12397.

3.4.2.3.2 Salt solubility.- After immersion in a saline solution,

for a period of 24 hours, the coated optical surface shall meet the
requirements of 3.4.1.1 and 3.4.1.4.

3.4.2.3.3 Water solubility.- After immersion in distilled water,

for a period of 24 hours, the coated surface shall meet the requirements
of 3.4.1.1 and 3.4.1.4.
3.4.2.3.4 Salt spray fog.- After exposure to a salt spray fog for
a continuous period of 24 hours, the coated surface shall meet the require-
ments of 3.4.1.1 and the applicable abrasion resistance requirements of
3.4.2.1.3 or 3.4.2.3.1.

6

256 APPENDIX B

KIL-F-48616

4. QUALITY ASSURANCE PROVISIONS

4.1 Responsibility for inspection.- Unless otherwise specified

in the contract or purchase order, the supplier is responsible for the
performance of all inspection requireaents as specified herein. Except
as otherwise specified. the supplier may utilize his own facilities or
any other facilities suitable for the performance of the inspection require-
ments specified herein. unless disapproved by the Government. The Govern-
ment reserves the right to perform any of the inspections set forth in the
specification where such inspections are deemed necessary to assure that
supplies and services conform to prescribed requirements.

4.1.1 General provisions.- Definitions of inspection terms shall

be as listed in KIL-STD-l09 and 6.3 herein.

4.1.2 Witness piece.- Unless otherwise specified, witness pieces

representing the actual coated componentM'may be used for spectral and.
environmental testing. The witness piecea ahall be positioned in the
costing chambe~ such that they represent the whole evaporation lot
(See 6.3.2). The Governaent reaerves the rigbt to test the actual coat-
ed component. Should a component fail, even though the representative
witness pieces pass the test, the lot shall be rejected.

4.1.~ Characteristics of the witness piece.- The witness piece shall

be of the same material and have a surface finish similar to that of the
component to be coated. The witness piece ahall be auch that it presents
nO difficulty in measuring and testing the spectral requirements of the
coating. Spectral performance of coating applied to small, thick, or
curved COmponents may be verified on convenient Sized flat, thin (approx-
imately 1.0mm, or 0.40 inch) witness piece.

4.2 First article (initial production) approval.- The requirement

for first article approval and the responsibility (Government or contrac-
tor) for first article testing shall be as specified in the contract
(See 6.2k). Unless otherwise specified, the sample for first article
approval tests shall consist of five (5) coated components plus ten (10)
coated witness pieces (See 4.1.2 and 4.1.3). The five (5) coated components
shall be tested, as specified herein, for all the requirements of 3.4.1.
Five (5) coated witness pieces shall be tested, as specified herein, for all
the requirements of 3.4.2.1, then 3.2. The other five (5) coated witness
pieces shall be tested, as specified herein, for all the requirements of
3.4.2.2, then 3.2. When an optional durability requirement is specified
in the contract or order an additional five (5) coated witness pieces shall
be tested, as indicated herein, for each optional durability requirement
specified. In addition to the foregOing, when the substrate material is
identified by component drawings or procurement documents, five (5) un-
coated substrates shall be provided and tested as specified herein for
all the requirements of 3.3. The Government reserves the right to subject
the coated components to all the tests specified here. The sample shall
be coated in the same manner, using the same materials, equipment, processes,

* and coated in the same eval'oratiOl~ lot

SELECTED UNITED STATES MILITARY SPECIFICATIONS 257

MIL-F-486l6
and procedures as used in regular production. All parts and materials
including packaging shall be obtained from the same source of supply as
used in regular production.

4.2.1 Government testing.- When the Government is responsible for

conducting first article approval tests, the contractor, prior to submitt-
ing the sample to the Government, shall inspect the sample to insure that
it conforms to all the requirements of the contract and submit a record of
this inspection with the sample, including certificates of conformance for
materials, as applicable.

4.2.2 Contractor testing.- When the contractor is responsible for

conducting first article approval tests, the sample shall be inspected by
the contractor for all the requirements of the contract. The sample and
a record of this inspection, including certificates of conformance for
materials, shall be submitted to the Government for approval. The Govern-
ment reserves the right to witness the contractor's inspection.

4.3 Inspection provisions.

4.3.1 SubmisSion of product.

4.3.1.1 Inspection lot size.- The inspection lot size shall consist
of all components or subassemblies (unit of product) coated within one
(8 through 12 hour) work shift.

4.3.1.2 Lot formation and presentation of lots.- The unit of product

shall be submitted for inspection on a moving inspection lot basis where
the components or subassemblies are continuously offered for inspection
in the order produced.

NOTE: Each evaporation lot (coated components and witness pieces) form-
ing a part of a moving inspection lot shall be identified for sub-
sequent evaluation, if required.

4.3.2 Examination and tests.

4.3.2.1 Substrate material.- Uncoated substrate material, when

specified on the component drawings or procurement document, shall be
inspected in accordance with this specification, the inspection provisions
contained in Supplementary Quality Assurance Provisions (SQAP), or other
procurements documents listed in the Technical Data Package (TDP).
Examination and tests related to Section 3 herein shall be performed on
a class basis in accordance with MIL-STD-105 and the sampling plan
specified in Table Ill. The tabulated classification of defects in Table
III shall constitute the minimum inspection to be performed by the
supplier after first article approval and prior to Government acceptance
or rejection by item or lot.

8

258 APPENDIX B

MIL-F-486l6

TABLE III - CLASSIFICATION OF DEFECTS

~ CHARACTERISTIC REQUIREMENT TEST PROCEDURE

CRITICAL: NONE DEF11IED

~: ~L 0.651 DEFECTIVE

101. Internal defecta 3.3.1 4.6.4

102. Surface defecta 3.3.2 4.6.5
103. Dimensions 3.3.3 4.6.6

!rn!Q!: NONE DEFINED

NOTE: The inspection for the Characteristics in Tsble III shall be conducted
- - at a temperature between ~Oo and -+90oF (l6 oC and 32 oC).

4.3.2.2 Components, subassemblies and witness pieces.- Costed

components subassemblies and witness pieces shall be inspected in accordance
with this specification, the inspection provisions contained in Supplementary
Quality Assurance Provisions (SQAP), or other procurement documents listed
in the Technical Data Package (IDP). Examination and tests related to
Section 3 herein shall be performed on a class basiS in.accordance with
MIL-SID-lOS and the sampling plans specified in Tables IV and V herein. The
tabulated classification of defects in Tables IV and V shall constitute the
minimum inspection to be performed by the aupplier after firat article
approval and prior to Government acceptance or rejection by item or lot.

TABLE IV - CLASSIFICATION OF DEFECTS

CHAllACTERISTIC REQUIllEMENT TEST PROCEDURE

CRITICAL: NONE DEFINED

!!MQ!: AQL 0.65X DEFECTIVE

104. Surface quality, coating 3.4.1.1 4.6.7.1
105. Surface quality, 8cratchea 3.4.1.2 4.6.7.2
106. Surface quality, digs 3.4.1.3 4.6.7.2
107. Surface quality, cosmetic 3.4.1.4 4.6.7.3
108. Coated area 3.4.1.5 4.6.7.4

MINOR: NONE DEFINED

NOTE: The inspection for the Characteristics in Table IV ahall be conducted
at a temperature between ~OOp and -+9oDp (16 oC and 32oC).

9
SELECTED UNITED STATES MILITARY SPECIFICATIONS 259

MIL-F-48616
4.3.2.3 Acceptance and rejection.- Rejected lots shall be screen-
ed for all defective characteristics. Removal of defective units and
resubmittal of rejected lota ahall be in accordance with "Acceptance
and Rejection" as specified in MIL-STD-I05.

4.3.2.4 Special sampling.- A minimum of three coated witness pieces,

or coated componenta (when required), ahall be selected as samples from
each evaporation lot. Each sample shall meet the requirements and tests
in Table V. The sequence of test shall be as delineated in Table V.

TABLE V - CLASSIFICATION OF DEFECTS

NO. CHARACTERISTIC REQUIREMENT TEST PROCEDURE

301. Spectral 3.2 4.6.3
302. Adhesion 3.4.2.1.1 4.6.8.1
303. Surface quality, coating 3.4.2.1.2 4.6.8.2,
(post humidity) 4.6.7.1
304. Surface quality, cosmetic 3.4.2.1.2 4.6.7.3
(post humidity)
305. Moderate abrasion 3.4.2.1.3 4.6.8.3

~: The inspection for the Characteristics in Table It shall be conducted

at a temperature between +600 and +900 F (16 0 C and 32 0 C).

4.3.2.5 Failure of sample.- Should anyone item of a special

sampling fail to meet the specified test requirements, acceptance of the
represented inspection lot shall be suspended by the Government until
necessary corrections have been made by the contractor and the resub-
mitted items have been approved. (See 4.3.2.3).

4.4 Inspection equipment.- Except as otherwise provided for by the

contract, the contractor shall supply and maintain inspection equipment
in accordance with the applicable requirements of KIL-I-45607.

4.4.1 ~overnment furnished inspection equipment.- Where the contract

provides for Government furniahed test equipment, supply and maintenance
of test equipment shall be in accordance with applicable requirements
specified in KIL-I-45607.

4.4.2 Contractor furnished inspection equipment.

4.4.2.1 Government design.- All inspection equipment specified by

drawing number in specificationa or SQAP forming a part of the contract
shall be supplied by the contractor in accordance with technical data
included in the contract.
4.4.2.2 Contractor eguipment.- The contractor shall supply inspec-
tion equipment compatible with the "Test Methoda and Procedures" specified
in 4.6 of thia specification. Since tolerance of test equipment is nor-
mally considered to be within 10l of the product tolerance for which it

10

260 APPENDIX B

KlL-F-48616

is intended, this inherent error in the test equipment must be considered

as part of the prescribed product tolerance limit. Thus, cOncept, con-
struction, materials, dimensions and tolerances used in the test shall
be so selected and controlled as to insure that the test equipment will
reliably indicate acceptability of a product which does not exceed 90%
of the prescribed tolerance limit, and permit positive rejection when non-
conforming. Construction shall be such as to facilitate routine cali-
bration of test equipment.

4.4.2.3 Inspection equipment design requirements.

4.4.2.3.1 Spectrophotometric eguipment.- The spectrophotometer

utilized for spectral measurements of finished components, or witness pieces,
shall have an optical system collimated to f/4.0, or a larger "f/-"
number.

4.4.2.3.2 Spectrophotometric measurement accuracies.

4.4.2.3.2.1 Wavelength resolution, scanning speed and time constant.-

The wavelength resolution of the spectrophotometer shall be better than
that required by the applicable detailed specification sheet. The scan-
ning speed and time constant shall be compatible with the resolution require-
ment of the item under test.

4.4.2.3.2.2 Wavelength and photometric accuracy.- When not specified

in the applicable detailed specification sheet, the spectral parameters shall
be measured to the following accuracies in the region from 0.7 um to 50.0
urn:

a. Wavelength accuracy.- Plus or minus (i) 0.5% of the wave-

length being measured, or in the case of narrow bandpass filters, t 20.0%
of the filter's half-band width.

b. Transmittance (T) accuracy.

TRANSMITTING REGION ACCURACY

(T ~ 1.0% absolute) t 1.5% of full scale

ATTENUATED REGION ACCURACY

(T~ 1.07. absolute) '!" 0.1%

11
SELECTED UNITED STATES MILITARY SPECIFICATIONS 261

MIL-F-48616

c. Reflectance (R) accuracy.

HIGH REFLECTANCE REGIO~ ACCURACY

(R ~ 80.01) ± 1.51 of full scale

KID-REFLECTANCE REGION ACCURACY
(5.01< R <: 80.07.) t 2.0% of full scale
LOW REFLECTANCE REGION ACCURACY

(1.51<' R ~ 5.01) ! 0.11

(R ~ 1.51) ~ 0.05%

4.5 Inspection documentation.

4.5.1 Spectral.- A copy of the spoctral test results shall be

supplied with each coating lot and shall contain sufficient data to show
compliance with the spectral requirements for the parts being shipped.
This data should include as a minimum the following:

a. Type of spectrophotomer used.

b. Properly labelled axes with spectral reference lines

(e.g. where appropriate, zero reference and/or full
scale reference lines).

c. Test conditions.

d. Actual measured values plus any theoretical or

measured corrections to the measured values.

4.5.2 Physical.- Inspection records shall be available which show

complisnce with sll physical requirements.

4.5.3 Environmental.- Inspection records shall be available which

show compliance with all environmental durability requirements.

4.6 Test methods and procedures.

4.6.1 Cleaning.- Before and after subjecting a coated sample (com-

ponent or witness piece) to any inspection or test, the coated sample shall
be thoroughly and carefully cleaned to remove dirt, finger marks, smears,
etc. The cleaning solution shall be aeetoDS, ethyl alcohol, isopropyl
alcohol, or mixtures thereof. FollOWing the cleaning, the coated sample
shall be carefully dried with lens tissue or a soft clean cloth. The
temperature of the cleaning solution shall not exceed +BOoF (27oC).

12

262 APPENDIXB

HIL-F-48616

4.6.2 Tempereture conditions.- Unless otherwise specified the

testing equipment and coated components, or witness pieces, shall be
stabtl ized at a temperature between ~oo end + 90~ (l6 0 C and 32°C), al'd
maintained at that temperature during all spectral testing.

4.6.3 Spectral.- Perform the test for spectral requirements

utilizing the test equipment specified in 4.4.2.3. The measured spectral
performance of infrared filters and coatings ahall conform to the require-
ments of 3.2. Unless otherwise specified the measured spectral values
for the applicable wavelengths shall be recorded as .. aaured in vacuum.

4.6.4 Internal defects.- Substrate materials which are transparent,

or semi-transparent to visible rediation, shall be visually examined by
transmission with the unaided eye for the requirements of 3.3.1. The
examination shall be performed using two 15 watt cool white fluoresce.nt
light tubes as the light source. The viewing distance from the substrate
surface to the eye shall not exceed 18.0 inches (45.7 cm). The substrate
material shall be viewed against a black matte baCkground. The only
illumination in the inspection area Shall be from the light source used
for examination. This method of examination is as depicted in Figure I.
The quality of the substrate material shall not exceed the dig tolerance
specified in 3.3.1. The method of inspection and the evaluation of
transparent and semi-transparent material for internal defects, other
than defects which are essentially round in nature, shall be as
specified by the procurement agency. The overall internal quality
of the substrate material shall conform to the requirements of 3.3.1.
The method of inspection to evaluate the internal quality of substrate
materials which do not transmit visible radiation shall be as specified
On the component drawing, or procurement document (See 6.2e).

4.6.5 Substrate surface defects.- The component substrate shall

be examined for scratches and digs by reflection or transmission, where
applicable using the inapection technique specified in 4.6.4. Magnifi-
cation shall be used as needed in these examinations. The length and
width of scratches, and the dig diameters shall be determined by use of
interferometer, microscopic measuring devices, calibrated precision
comparators, or Similar applicable precision measuring devices. The
width, length and density of all scretches on the substrate surface
shall conform to the requirements of 3.4.1.2 and 3.4.1.2.1. The diameter
and density of all digs in the substrate shall conform to the requirements
of 3.4.1.3 and 3.4.1.3.1. The quality of each substrate surface shall
meet the requirements of 3.3.2.

4.6.6 Substrate dimensions.- The substrate shall be inspected by

use of standard measuring equipment (See 6.3.3). The substrate dimensions
shall conform to the requirements of 3.3.3.

4.6.7 Surface quality of coated components.

4.6.7.1 Coating.- The coating on the component surface shall be

examined for evidence of flaking, peeling, cracking, etc by reflection
or transmission, where applicable, using the inspection technique

13
SELECTED UNITED STATES MILITARY SPECIFICATIONS 263

MIL-P-48616

specified in 4.6.4. The coating shall conform to the requirements of

3.4.1.1.

4.6.7.2 Scratch and dig.- The coated substrate shall be examined

for scratches and digs by reflection or transmission, as applicable,
using the inspection technique specified in 4.6.4. }!agnification shall
be used as needed in these examinations. The length and width of scratches,
and the dig (hole) diameters shall be determined by use of interferometer,
microscopic measuring devices, calibrated precision comparators, or similar
applicable precision measuring devices. The width, length and density of
all scratches in the coating or substrate shall conform to the requirements
of 3.4.1.2 and 3.4.1.2.1. The diameter and density of all digs (holes)
in the coating or substrste shall conform to the requirements of 3.4.1.3
and 3.4.1.3.1.

4.6.7.3 Cosmetic.- The coating and substrate shall be examined by

transmission or reflection, where applicable, using the inspection
technique specified in 4.6.4. Blemishes on the coating or substrate shall
conform to the requirements of 3.4.1.4. Coatings or substrate that exhibit
evidence of discoloration shall be inspected for conformance to the
applicable spectral requirements (See 4.6.3) specified on drawing, contract
or purchase order. The blemished area on the coated substrate shall then
be subjected to the moderate abrasion test of 4.6.8.3. The coating and
substrate shall conform to the requirements of 3.4.1.4.

4.6.7.4 Coated area.- The coated area of the component shall be

inspected by use of standard measuring equipment. The coating of the
component shall conform to the requirements of 3.4.1.5.

4.6.8 Environmental and physical durability.- The coating on the

component, or witness piece shall withstand exposure to the following
conditions in the order specified.

4.6.8.1 Adhesion.- The coated component or witness piece shall

be subjected to an adhesion test USing 1/2" (12.7mm) wide cellophane
tape cbnforming to Type I of L-T-90. Press the adhesive surface of the
cellophane tape firmly against the coated surface so as to cover the
stained area, then quickly remove it at an angle which is normal to the
coated surface. Immediately follOWing the removal of the adhesive tape,
the coated surface of the component, or witnes8 piece, shall be examined
by reflection using the inspection technique specitied in 4.6.4. The
coating shall conform to the requirements of 3.4.2.1.1. Subsequent to
this test the coated component, or witness piece, shall be subjected
to the test in 4.6.8.2.
4.6.8.2 Humidity.- The coated component, or witness piece, shall
be placed into an environmental controlled test chamber and exposed to
a temperature of + 1200 t 40 p (49 0 t 20 c) and 95X to 100% relative
humidity for a minimum of 24 hours. Subsequent to this exposure the

14
264 APPENDIX B

MIL-F-48616

coated component, or witness piece, shall be removed from the test chamber,
cleaned (See 4.6.1), dried, and then subjected to the examinations specified
in 4.6.7.1 and 4.6.7.3. The coating shall conform to the requirements of
3.4.2.1.2. The coated component, or witness piece, shall then be subjected
to the test specified in 4.6.8.3.

4.6.8.3 Moderate abrasion.- Within one hour after the humidity

test of 4.6.8.2 the coated component, or witness piece shall be subjected
to a moderate abrasion by rubbing the coated surface with a 1/4 inch
(6.4mm) thick by 3/8 inch (9.5mm) wide pad of clean dry, laundred cheese-
cloth conforming to CCC-C-440 affixed to an abrasion tester that conforms
to Drawing D7680606. The cheesecloth pad shall completely cover the
eraser portion of the tester and be secured to the shaft with an elastic
band. The cheesecloth pad shall be rubbed across the coated surface from
one point to another over the same path for 25 complete cycles (50 strokes)
with a minimum force of 1.0 pound (0.45kg) continuously applied. The
length of the stroke shall be approximately equal to two widths of the
cheesecloth pad when the diameter or area of the component permits. The
abrasion tester shall be held approximately normal to the surface under
test during the rubbing operation. Subsequent to the rubbing operation
the compOnent, or witness piece, shall be cleaned (See 4.6.1), dried snd
then subjected to an examination by reflection using the inspection
technique specified in 4.6.4 for evidence of phySical damage to the coat-
ing. The coating on the component, or witness piece, shall meet the
requirements of 3.4.2.1.3.

4.6.9 Thermal and cleaning durability;- The coated component, or

witness piece, shall Withstand exposure to the following test conditions:

4.6.9.1 Temperature.- The coated component, Or witness piece,

shall be subjected to temperatures of -800r ± 2~ (-62 0 ± 10C) and + 1600r
t 20f (71 0C T 10C) for a period of 2 hours at each temperature. After
each exposure the coated components or witness pieces shall be stabilized
at an ambient temperature between +600 and +900r (16°C and 320C) and
subjected to the examination and test specified in 4.6.7.1 and only the
adhesion test of 4.6.8.1. The coating shall conform to the requirements
of 3.4.2.2.1.
4.6.9.2 Solubility and cleanability.- The coated component, or
witness piece, shall be immersed, in sequence, in the following solutions
maintained at room temperature (+600 to +90Of): trichloroethylene; acetone;
and ethyl alcohol. The itmlE!rsion time in each solution shall be a minimum
of ten minutes. Upon removal from each solution, the solvent shall be
allowed to evaporate to dryness without wiping or forced drying before
proceeding to the next solution. Upon removal from the alcohol solution,
and after drying any resultant stains on the coated surface shall be
removed by wiping the coating to a clean, stain-free condition with an
ethyl-alcohol-moistened cheesecloth. Subsequent to this cleaning, the
coating Shall be subjected to the examinations specified in 4.6.7.1 and

15
SELECTED UNITED STATES MILITARY SPECIFICATIONS 265

mL-F-486l6

".6.7.3. The coating shall conform to the requirements of 3.4.2.2.2.

4,6.10 Optional durability tests.- The coated cOmpOnents or

witness samples shall withstand exposure to the following test conditions
when specified in the contract, purchase order or on the cOmponent
drawing (See 6.2j).

4.6.10.1 Severe abrasion.- The coated components, or witness piece,

shall be subjected to a severe abrasion by rubbing the coated surface with
a standard eraser conforming to KIL-E-12397 mounted in a eraser abrasion
coating tester that conforms to the requirements of Drawing 07680606. The
eraser shall be rubbed across the surface of the component, or witness
piece, from One point to another over the same path for 10 complete cycles
(20 strokes) with a force of 2.0 to 2.5 pounds continuously applied. The
length of the stroke shall be approximately equal to 3 diameters of the
eraser when the diameter or area of the compOnent, or witness piece,
permits. The eraser abrasion tester shall be held approximately normal
to the surface under test during the rubbing operation. Subsequent to
the rubbing operation the component or witness piece, shall be cleaned
(See 4.6.1), dried and then subjected to an examination by reflection
using the inspection technique specified in 4.6.4 for evidence of physical
damage to the coating. The film coating on the component, or witness
piece, shall meet the requirements of 3.4.2.3.1.

4.6.10.2 Salt solubility.- The coated component, witness piece,

shall be immersed for a period of 24 hours in a solution of distilled
water and sodium chloride (salt). The mixture shall be 6 ounces (170 grams)
0
of salt per gallon (3.8 liters) of water at room temperature (16 0 to 32 C).
Subsequent to this immersion the coated component, witness piece, shall
be removed from the solution and ~ntlY washed, or dipped in clean
running water not warmer than 100 F (3Soe) to remove salt deposits. The
coated component, witness piece, shall then be cleaned (See 4.6.1), dried,
and then subjected to the examination specified in 4.6.7.1 and 4.6.7.3.
The coated component, or witness piece, shall conform to the requirements
of 3.4.2.3.2.

4.6.10.3 Water solubilitX.- The coated components, or witness

piece, shall be immersed for a period of 24 hours in distilled water at
room temperature (~aoF to +9()OF). Subsequent to this i.-rs ion, the
coated components, or witness piece, shall be removed from the solution,
cleaned (See 4.6.1), dried, then subjected to the examination specified
in 4.6.7.1 and 4.6.7.3. The coating on the component, or witness piece,
shall meet the requirements of 3.4.2.3.3.

4.6.10.4 Salt spray (f08).- The coated component, witness piece,

shall be subjected to a salt spray fog teat in accordance with ASTMB-117-
73, for a continuous period of 24 houra. Subsequent to the salt spray
fog exposure the coated component, or witness piece, shall be washed and
dried as specified. The coated component or witness piece, shall then be
cleaned (See 4.6.1), dried, and then subjected to the examination specified
in 4.6.7.1. Subsequent to the visual examination for physical defects

16

266 APPENDIX B

MIL-F-48616

the coated component or witness piece, shall be subjected to the test in

4.6.8.3 or 4.6.10.1, as specified in the contract or order. The coated
component or witness piece, shall meet the requirements of 3.4.2.3.4.

5. PACKAGING

5.1 Packaging and packing.- In the absence of specified packaging

data in the contract, or purchase order, each thin film coated optical
element shall be individually wrapped in a chemically inert paper that
will not scratch, leave a residue or corrode the coated element (See 6.4g).
The wrapped coated element ahall be immobilized in a unit container that
provides adequate protection during handling and shipment.

5.2 Marking.- Each packaging and ahipping container shall be mark-

ed to show the following:

a. Lot or batch number.

b. Contract or purchase order number.
c. Part or drawing n~mber.
d. Quantity of parts in container.
e. Manufacturer's name or trademark.

5.3 Shipping container.- Shipping containers shall be marked With

words ''DELICATE OPTICAL COMPONEnS REQUIRING SPECIAL HANDLING". The
appropriate side of the container shall be clearly marked to indicate
"TOP" or "OPEN THIS SIDE".

6. NOTES

6.1 Intended use.- This document is intended to be used for all

types of infrared interference filters and coatings. It is anticipated
that most of the coatings/filters covered by this document will not
include specialized characteristics such as fluorescence, high energy
handling properties, c~nduction, etc. However, specialized characteristics
as those mentioned, or those in which more exacting performance cap-
abilities are needed can be imposed by the generation of additional
specification sheets that include the desired specific properties. The
primary requirement on the specification sheet is the spectral charac-
teristic of the filter or coating. However. requirements in addition
to spectral requirements may be added, if needed. The specification
sheet in conjunction with the general document will then adequately
specify the required filter and/or coating.

6.2 Ordering data.- Purchasers should exercise any desired options

offered herein, and procurement documents should specify the following:

a. Title, number and date of this specification.

b. Number and revision of the specific detailed specification

sheet and dash number, if applicable.

c. Post enVironmental spectral evaluation (See 3.2).

d. Substrate material (See 3.3).

17
SELECTED UNITED STATES MILITARY SPECIFICATIONS 267

MIL-F-48616

e. Internal defect requirement (See 3.3.1).

f. Scratch value (See 3.4.1. 2) •

g. Dig value (See 3.4.1.3).

h. Cosmetic requirement (See 3.4.1.4).

j. Optional durability requirements (See 3.4.2.3).

k. Requirement for first article approval (See 4.2, 4.2.1

and 4.2.2).

m. Supplemental spectral requirements (See specification

sheets).

6.3 Definitions.

6.3.1 Scratch and dig identification.- This is specified by

two letters separated by a hyphen (i.e. F-F). The first letter of the
pair is the maximum scratch letter. The second letter is the maximum
dig letter.

6.3.2 Evaporation lot.- An evaporation lot is defined as the group

of parts which has the coating applied at the same time and in the same
chamber.

6.3.3 Standard measuring equipment (SME).- Standard measuring

equipment is defined 9S the common measuring devices which are usually
stocked by commercial supply houses for ready supply (shelf items) normally
used by an inspector to perform dimensional inspection of items under
procurement. This category also includes commercial testing equipment
such as meters, optical comparators, etc.

6.4 Informational notes:

a. Spectrophotometer measurements: Spectrophotometers are

calibrated for wavelength uSin~tandard air (pressure
a760mm Kg and temperature· 15 C). The purpose is to
clearly identify the index of refraction of the medium in
which the calibration is to apply and eliminate possible
fluctuations in poorly deftned conditions. For man
applications, such details in calibration are inconse-
quential since the change in oper~ting conditions affects
measuring only slightly. For example, the difference
between wavelengths measured in vacuum and the same wave-
lengths measured in standard air amount to about 0.025%.
(Reference: Commission On Molecular Structure and
Spectroscopy. International Union of Pure and Applied
Chemistry, Table of Wavenumbers for a Calibration of
Infrared Spectrometers. (Butterworths, 1961).

18
* standards evaluated either in a vacuum or in
268 APPENDIX 8

MIL-F-48616

b. Temperature effects on coatings:

If the application for these coatings is for other than

normal room temperature, the user should request from the
manufacturer specific performance information. Spectral
performance is temperature dependent and allowance for
temperature effects can frequently be included in filter
design. Coating durability is also temperature dependent.

c. Angle of incidence effects on coatings:

Normally, as the angle of incidence of the radiation

increases from 00 (normal incidence to the optical
element) the spectral features of the thin film coating
shift to shorter wavelengths with changes in transmit-
tance and reflectance also occurring. The same holds
true as the radiation changes from parallel to convergent
beam. Allowance for non-normal angles of incidence and
convergency angles ~an be included in filter design. 0
For angles of incidence greater than approximately 20 ,
spectral performance is influenced by the polarization
state of the incident radiation. For a particular
application, the user should request pertinent information
from the manufacturer.

d. Internal defects of substrates:

Internal defects of spectral filter substrates can

seriously deteriorate the related system performance.
Techniques of various degrees of precision exist for
evaulation~~isually opaque substrates for internal defects.
Some of these are image spoiling property test, snooper
scope evaluation, infrared microscope test, laser scanning,
X-ray, etc.

The higher the resolution and lower the tolerance of the

system to scattered radiation, the more exacting the tests
that should be selected.

Improperly oriented birefringent substrates can seriously

deteriorate higher resolution performance. A favorable
orientation should be specified.

e. Physical thickness of coatings:

Coatings in the infrared spectral region can be several

thousandths of an inch thick. Consequently, allowance
must occasionally be made for coating thicknesses as well
as substrate thickness when calculating focal length and
designing filter :lOluers.

19


SELECTED UNITED STATES MILITARY SPECIFICATIONS 269

}!IL-F-48616

f. Test conditions:

Should in-use conditions differ from these test conditions,

the customer should request from the manufacturer the per-
tinent adjusted parameters either by theoretical calculation
or change in the test conditions. Adjustments that may be
required shall include the effects of temperature, angle of
incidence, convergence angle of the incident radiation, and
polarization of the incident radiation. The optical
component or witness may absorb energy during spectro-
photometric tests thus causing a temperature change.
Controlled temperature during testing should be specified
by the customer.

g. l~rapping tissue:

Papers which contain certain chemicals, or are fabricated

of bleached wood pulp, may cause deterioration of optical
coatings. Care must be exercised to insure that chemically
inert paper made from unbleached wood pulp is used to
wrap optics.

Custodians: Preparing activity:

Army - AR Army - AR
Navy - SH Project No. 6650-0091
Air Force - 99

20

270 APPENDIX B

KIL-F-48616

COMMERCIAL FIXTURE CONTAINING

'lWO COOL WHITE 15-WArr FLUORESCENT
LAMPS (SEE rmE 1)

18" MAX ~6" MU

(45.'eM' \ 1 (15 • ...,

BLACK KArrE
BACKGROUND

~,-----

SEE JIOTE 2

NOTES

1. THE ONLY ILLUKINATION IN THE TEST AREA SHALL

BE FROM THE LICHT SOURCE USED FOR TEST.

2. RELATIVE POSITION OF COMPONENT FOR INSPECTION BY REFLECTION.

3. RELATIVE POSITION OF COMPONENT FOR INSPECTION BY ~Q;_NSHISSION.

FIGURE 1

21
 Appendix C

BASIC Program to Calculate

Thin Film Filter Performance

Since Berning's landmark work and review (1963) on the mathematics of thin film
filter performance calculations, there have been many computer programs written
to carry out the numerical drudgery. Baumeister (1986) has recently published an
overview of currently commercially available software packages that perform these
calculations. DeBell published a review of calculational algorithms in 1978. The
majority of the modern programs use the matrix formulation as their mathematical
basis.
Liddell (1981) published a monograph on computer-aided thin film design
calculations. She includes a listing of a FORTRAN program that evaluates the
performance of a given design.
This Appendix includes a listing of a computer program written in the BASIC
language for calculating the transmission, reflection, and absorption of a thin film
design. It is compatible with both the mM PC and the Apple Macintosh classes of
computers. The comments in the listing explain the functioning of the program.

271
~
~ REM Program name THIN FILMS by S.P. Fisher

REM This program calculates the following properties of a stack of thin

REM film coatings at a given wavelength:
REM reflection (R),
REM transmission (T),

REM phase shift upon reflection (rho),

REM and phase shift upon transmission (tau).

REM It uses the matrix computational method outlined by Peter H. Berning

REM in IIPhysics of Thin Films" Vol. 1, p.87-92, Academic Press, 1966.
REM
REM This program is intended to be a starting point for more generalized

REM cases of thin film coatings that would add absorption in each film and
REM oblique angles of incidence.
REM
REM Below is an outline of the THIN FILMS program and the variables used in each step
REM
REM 1. Data entry
REM 1.1 Set constant(s), (pi)
REM 1.2 Enter the index of the incident medium, (No)

REM 1.3 Enter the index and physical thickness of each layer of thin film, (Nx,Hx)

REM 1.4 Enter the complex index of substrate, (index=Ns + jKs)

REM 1.5 Enter the wavelength range and increment, (wv11,wv12,inc)

REM 1.6 Print all the input data to system printer for verification.
REM
REM Repeat steps 2 through 6 for each incremental wavelength throughout
REM the wavelength range of interest.
REM
 REM 2. Calculate the characteristic matrix for the thin film stack

REM 2.1 Initialize the characteristic matrix.

REM 2.2 Find the characteristic matrix for each layer, (Y)

Note: Matrix notation and definition is as follows:
REM
REM I Yl+jY2 Y3+jY41 cosO (j INx) sinO 1
REM Y =I 1= I
REM I Y5+ jY6 Y7+jYSI (j*Nx) sinO cosO I

REM
REM where 0 2*pi*Nx*Hxlwavelength
REM
REM 2.3 Multiply the individual layer matrices to

REM obtain the characteristic matrix for the stack, (M)
REM
REM M [Y(layerl)]*[Y(layer2)]* ... * [Y(layer L)]
REM
REM
REM
REM 3. Calculate the reflected and transmitted electric field amplitudes, (Eo and Ho)

REM I Eo I 11 I

REM I I = [M] *1 1
REM I Ho I I Ns I
REM
REM
N
~
Col
N
~ REM 4. Calculate the reflection (R), the transmission (T), the phase shift upon

REM reflection (rho), and the phase shift upon transmission (tau) for the thin film

REM stack

REM I Eo - (HolNo) \**2 4 Ns
REM R = I --------- I T =
REM
I Eo + (HolNo) I No*IEo + (HolNo) 1**2
REM
REM Eo - (HolNo) 1
REM rho = arg , ------------ tau arg
REM Eo + (HolNo) Eo + (HolNo)
REM
REM
REM 5. Determine the correct quadrant for rho and tau.
REM
REM 6. Print out R,T,A,rho, and tau.
REM

REM 1. Data entry

DIM N(99) ,h(99),label(70)
CLS
pi = 3. 141593

PRINT "Enter comments, up to 70 characters"

INPUT "comments =?", comments$
PRINT "Enter index of incident medium, e. g. air = 1. 0"
INPUT "No = ? " , No
L=1
Layers:
PRINT "Enter the index of layer"; L; ". Enter N=O to stop input"
 INPUT "N=?", N (L)

IF N(L) =0 THEN GOTO Substrate
PRINT "Enter the physical thickness of layer";L
INPUT "h=?", h (L)
L=L+l

GOTO Layers
Substrate:
L=L-l
IF L=O THEN PRINT "No film entered. The calculation will be for the uncoated
surface. "

PRINT "Enter the substrate index, e. g. glass=1.52"

INPUT "NS=?", Ns
PRINT "Enter the substrate absorption, e. g. glass=O.O"

INPUT "Ks=?", Ks
PRINT "Enter the beginning wavelength, the ending wavelength, and the wavelength
increment. "
INPUT "wvll =?, wvl2 =?, wvlinc=?", wvll, wvl2, wvlinc
LPRINT "COMMENTS"

LPRINT" ";comments$

LPRINT
LPRINT "INPUT PARAMETERS"
LPRINT USING" Incident index: No=###.###";No

FOR i=l TO L
LPRINT USING" Layer###, index: N =###.### Phys.Thk:h=###.####";i,N(i),
NEXTi
LPRINT USING" Substrate index: Ns=###. ### Ks=###. ####";Ns,Ks

Sl LPRINT
Ks= -Ks
I\)

Ol FOR wvl =wvl1 TO wvl2 STEP wvlinc

REM 2. Calculate the characteristic matrix for the thin film stack.

REM Initialize the characteristic matrix.

MI=I

M2=O

M3=O

M4=O

M5=O

M6=O

M7=1

M8=O

FORi=1 TO L

REM X is set equal to M before finding the matrix for each layer, then X is used

REM to find the new product of M times Y for i number of layers.
XI=MI

X2=M2

X3=M3

X4=M4

X5=M5

X6=M6

X7=M7

X8=M8

REM Calculate the characteristic matrix for each film.

Phi = 2*pi *N (i) *h (i) Iwvl

YI =COS (Phi)
Y7=YI
temp=SIN (Phi)

Y4 = templN ( i )

Y6 = temp*N (i)
REM These are zero because the films are not absorbing.

Y2=O

Y3=O

Ys=o

Y8=O
REM Calculate the product of the characteristic matrices for each layer

MI = Xl *YI- X4*Y6

M2 =X2*YI +X3*Y6

M3 =X3*Y7 -X2*Y4

M4 = Xl *y 4 + X4 *Y7

MS =XS*YI-X8*Y6

M6 =X6*YI + X7*Y6

M7 =X7*Y7 -X6*Y4

M8 = XS *y4 + X8 *Y7
NEXTi

REM 3. Calculate the reflected and transmitted electric field amplitudes, Eo and Ho

Eol =MI +M3*Ns -M4*Ks

Eo2 = M2 + M3 *Ks + M4 *Ns

HoI =MS +M7*Ns-M8*Ks

Ho2 =M6 +M7*Ks +M8*Ns

REM 4. Calculate R,T,A,rho and tau

REM n is a temporary complex numerator,

REM d is a temporary complex denominator,

REM dsqr is the magnitude of d,

~


N
~ REM crl,cr2,ctl, and ct2 are complex components resulting from complex division

nl =Eo1-HollNo
n2 =Eo2 -Ho2lNo

dl =Eol + Ho1INo
d2 =Eo2 + Ho2INo
dsqr = dl1\2 + d21\2

crl= (nl*dl+ n2*d2)/dsqr

cr2 = (n2*dl- nl *d2) /dsqr

ctl == dl /dsqr
ct2 = - d2/dsqr

REM Results for R,T,rho,and tau

R= (crlI\2+cr21\2) *100

T= (4*Ns/(No*dsqr» *100

IF Ks<>O THEN T=O

rho=ATN (cr2/crl) *180/pi

tau=ATN (ct2/ctl) *180/pi

REM 5. Determine the correct quadrant for rho and tau

REM re-calculate rho as necessary

quad=l

IF crl<O THEN IF cr2<0 THEN quad=3 ELSE quad=2 ELSE IF cr2<0 THEN quad=4

IF quad=2 OR quad=3 THEN rho=rho+180

IF quad = 4 THEN rho=360+rho

REM recalculate tau as necessary

quad = 1
IF ct1<O THEN IF ct2<0 THEN quad=3 ELSE quad=2 ELSE IF ct2<0 THEN quad =4

IF quad = 2 OR quad=3 THEN tau= 180 + tau

IF quad=4 THEN tau=360+tau

IF Ks<>O THEN tau=O
 REM 6. Print out R,T,rho, and Tau.
IF wvl =wvll THEN LPRINT "PERFORMANCE RESULTS"
IF wvl=wvll AND L=O THEN LPRINT" Note: Performance is for the uncoated substrate

IF wvl =wvll THEN LPRINT" wvl Refl rho Trans tau"
LPRINT USING"#####. ### ###. #### ###. ## ###. #### ###. ##"; wvl, R, rho, T, tau
PRINT

NEXT wvl
LPRINT
LPRINT
LPRINT
PRINT "Program ended. "
END

N
~
 REFERENCES

An excellent bibliography on thin film technology is given at the end of the chapter
written by Dobrowalski on optical filters in the Optical Society of America's Optics
Handbook (Driscoll and Vaughan, 1978).

Abeles, F. (1950). La determination de l'indice et de l'epaisseur des couches minces trans-

parentes. 1. Phys. Roo., 11, 310.
Agranoff, Joan (1985). Modern Plastics Encyclopedia 1985-1986, vol. 62, no. lOA, McGraw-
Hill, New York.
Allen, Thomas H., Joseph H. Apfel, and C. K. Camiglia (1978). A 1.06 micrometer laser
absorption calorimeter for optical coatings. In Glass, Alexander J., and Arthur H.
Guenther, eds.: Laser Induced Damage in Optical Materials: 1978. Washington, D.C.,
National Bureau of Standards Special Publication #541.
Anthon, Erik (1982). Spatially uniform detector assemblies. In Seddon, R. I., ed.: Optical
Thin Films. Proceedings of the SPIE, volume 325, Bellingham, Wash., 98227.
Apfel, Joseph H. (1965). Circularly wedged optical coatings. II. Experimental. Appl. Opt.,
4(8):983.
Apfel, Joseph H. (1977). Optical coating design with reduced electric field intensity. Appl.
Opt., 16(7):1880.
Apfel, Joseph H. (1982). Phase retardance of periodic multilayer mirrors. Appl. Opt., 21(4):733.
Apfel, Joseph H. (1984). Graphical method to design internal reflection phase retarders.
Appl. Opt., 23(8):1178.

281
282 REFERENCES

Austin, R. (1970). Narrow band interference light filter. United States Patent #3,528,726,
September 15.
Baker, Martin L., and Victor L. Yen (1967). Effects of the variation of angle of incidence
and temperature on infrared filter characteristics. Appl. Opt .. 6(8):1343.
Ballik, E. A. (1971). Area and wavelength sensitivity of a photomultiplier. Appl. Opt.,
10(3):689.
Bartell, F. 0., E. L. Dereniak, and W. L. Wolfe (1980). The theory and measurement of
bidirectional reflectance distribution function (BRDF) and bidirectional transmittance
distribution function (BTDF). In Radiation Scattering in Optical Systems. Proceedings
of the SPIE, vol. 257, SPIE, Bellingham, Wash.
Bartolomei, Leroy A. (1976). Striped dichroic filter with butted stripes and dual lift-off
method for making same. U.S. Patent #3,981,568, September 21.
Baumeister, Philip (1981). Theory of rejection filters with ultranarrow bandwidths. J. Opt.
Soc. Am., 71(5):604.
Baumeister, Philip (1986). Computer software for optical coatings. Photonics Spectra, p.
76, July.
Beauchamp, W. T. (1986). Critical issues for coating on large high power laser components.
In Optical Component Specifications for Laser-Based Systems and Other Modern Op-
tical Systems. Proceedings of the SPIE, vol. 607, SPIE, Bellingham, Wash.
Bell, Alan E., and Fred W. Spong (1978). Antireflection structures for optical recording.
IEEE Journal of Quantum Electronics. QEI4(7):488.
Benjamin, P., and C. Weaver (1959). Adhesion of metal films to glass. Proc. Roy. Soc. A,
254:177.
Benjamin, P., and C. Weaver (1961). The adhesion of evaporated metal films on glass.
Proc. Roy. Soc. A, 261:516.
Bennett, H. E. (1978). Scattering characteristics of optical materials. Opt. Eng. 17(5):480.
Berning, P. H. (1963). Theory and calculations of optical thin films. In Hass, G.,
ed.: Physics of Thin Films. vol. 1. Academic Press, New York and London, pp.
69-121.
Birth, G. S., and D. P. DeWitt (1971). Further comments on the areal sensitivity of end-
on photomultipliers. Appl. Opt .. 10(3):687.
Born, Max, and Emil Wolf (1970). Principles of Optics. 4 ed. Pergamon Press, Oxford.
Buckmelter,1. R., T. T. Saito, R. Esposito, L. P. Mott, and R. Strandlund (1975). Dielectric
coated diamond turned mirrors. In Laser Induced Damage in Optical Materials. Na-
tional Bureau of Standards Special Publication 435.
C. I. E. (1978). Supplement No.2 to C. l. E. Publication No. 15 (1971), Bureau Central
de la C. I. E., Paris.
Camiglia, C. K. (1979). Scalar scattering theory for multilayer optical coatings. Opt. Eng.,
18(2):104.
Camiglia, C. K., and Joseph H. Apfel (1980). Maximum reflectance of multilayer dielectric
mirrors in the presence of slight absorption. J. Opt. Soc. Am., 70(5):523.
Clarke, D., and J. F. Grainger (1971). Polarized Light and Optical Measurement. Pergamon
Press, Oxford.
Cox, J. T., and G. Hass (1958). Anti-reflection coatings for germanium and silicon in the
infrared. J. Opt. Soc. Am .• 48(10):677.
Cox, J. T., J. E. Waylonis, and W. R. Hunter (1959). Optical properties of zinc sulfide in
the vacuum ultraviolet. J. Opt. Soc. Am., 49(8):807.
Crabb, R. L. (1972). Evaluation of cerium stablized micro sheet covers lips for higher solar
cell outputs. Conference Records of the Ninth IEEE Photovoltaic Specialists Confer-
ence.
REFERENCES 283

DeBell, G. W. (1978). Computational methods for optical thin films. In Optical Coatings
Applications and Utilization II. Proceedings of the SPIE, vol. 140, SPIE, Bellingham,
Wash.
Demichelis, F., E. Mezzetti-Minetti, L. Tallone, and E. Tresso (1984). Optimization of
optical parameters and electric field distribution in multilayers. Appl. Opt., 23(1):165.
Dept. of Defense (1962). Mil-HNDBK-141. Defense Supply Agency, Washington, D. C.
Dirks, A. G., and H. J. Leamy (1977). Columnar microstructure in vapor-deposited thin
films. Thin Solid Films, 47:219.
Dobrowolski, J. A. (1986). Comparison of the Fourier transform and flip-flop thin-film
synthesis methods. Appl. Opt., 25(12):1966.
Donovan, T. M., P. A. Temple, Shiu-Chin Wu, and T. A. Tombrello (1980). The relative
importance of interface and volume absorption by water in evaporated films. In Bennett,
H. E., A. J. Glass, A. H. Guenther, and B. E. Newnam, eds.: lAser Induced Damage
in Optical Materials, 1979. National Bureau of Standards Special Publication no. 568,
Washington, D. C.
Driscoll, Walter G., and William Vaughan, eds. (1978). Handbook 0/ Optics. McGraw-
Hill, New York.
Edwards, D. K., J. T. Gier, K. E. Nelson, and R. D. Roddick (1961). Integrating sphere
for imperfectly diffuse samples. J. Opt. Soc. Am., 51(11):1279.
Ennos, Anthony, E. (1966). Stresses developed in optical film coatings. Appl. Opt., 5(1):51.
Epstein, L. Ivan (1952). The design of optical filters. J. Opt. Soc. Am., 42(11):806.
Giang, R., R. A. Holmwood, and R. L. Rosenfeld (1965). Determination of stress in films
on single crystalline silicon substrates. Rev. Sci. Instr. 36, no. I, p. 7 (Jan.).
Gee, John R., Ian J. Hodghinson, and H. Angus Macleod (1985). Moisture-dependent
anisotropic effects in optical coatings. Appl. Opt., 24(19):3188.
Guenther, K. H. (1981). Nodular defects in dielectric mUltilayer and thick single layers.
Appl. Opt., 20(6):1034.
Guenther, K. H., and H. Enssle (1986). Ultrasonic precision cleaning of optical components
prior to and after vacuum coating. In Conference on Thin Film Technologies. Pro-
ceedings of the SPIE, vol. 652, SPIE, Bellingham, Wash.
Hahn, R. E., and B. O. Seraphin (1978). Spectrally selective surfaces for photothermal solar
energy conversion. In Hass, Georg, and Maurice H. Francombe, eds.: Physics o/Thin
Films. Academic Press, New York.
Haisma, Jan, Jan H. T. Pasman, Johan M. M. Pasmans, and Pieter van der Werf (1985).
Wide-spectrum tint-free reflection reduction of viewing screens. Appl. Opt., 24(16):2679.
Harding, G. L., B. Window, D. R. McKenzie, A. R. Collins, and C. M. Horwitz (1979).
Cylindrical magnetron sputtering system for coating solar selective surfaces onto batches
of tubes. J. Vac. Sci. Technol., 16(6):2105.
Harrick, N. J. (1967). Internal Reflection Spectroscopy. Interscience-Wiley, New York.
Hass, G., and J. B. Ramsey (1969). Vacuum deposition of dielectric and semiconductor
films by a CO 2 laser. Appl. Opt .. 8(6): 1115.
Heavens, O. S. (1950). Some factors influencing the adhesion of films produced by vacuum
evaporation. J. Phys. Rad., 11:355.
Heavens, O. S. (1965). Optical Properties o/Thin Solid Films. Dover, New York.
Herpin, Andre (1947). Calcul du pouvoir reflecteur d'un systeme stratifie quelconque. Comptes
Rendus, 225:182.
Hill, Russell J., ed. (1976). Physical Vapor Deposition. Temescal, Berkeley, Calif.
Holland, L. (1956). Vacuum Deposition a/Thin Films. Chapman & Hall, London.
IIIsley, Rolf F., Alfred 1. Thelen, and Joseph H. Apfel (1969). Circular variable filter. United
States Patent #3,442,572, May 6.
284 REFERENCES

Jackson, John David (1962). Classical Electrodynamics. John Wiley & Sons, New York.
Jacobs, Carol (1981). Dielectric square bandpass design. Appl. Opt., 20(6):1039.
Jacobson, M. (1986). Deposition and Characterization of Optical Thin Films. Macmillan,
New York.
Jenkins, F. A., and H. E. White (1957). Fundamentals of Optics. 3 ed., McGraw-Hili, New
York.
Jet PropUlsion Laboratory (1976). Solar Cell Array Design Handbook. Report SP 43-38,
Table 7.14-1 in volume 2, JPL, 4800 Oak Grove Drive, Pasadena, Calif., 91103.
Johnston, S. C., and S. F. Jacobs (1986). Some problems caused by birefringence in dielectric
mirrors. Appl. Opt., 25(12):1878.
Keller, Peter (1983). 1976 CIE-UCS chromaticity diagram with color boundaries. Proceed-
ings of the Society for Information Display, 24(4):317.
Kelley, K. L. (1943). Color designations for lights. National Bureau of Standards Research
Paper RP1565, Journal of Research of NBS, 31:271; also see the EIA Publication of
the JEDEC IGC, Optical characteristics of cathode ray tubes, 1955.
Kittel, Charles (1968). Introduction to Solid State Physics, 3 ed., John Wiley, New York,
p. 150 et seq.
Klinger, R. E., and C. K. Carniglia (1985). Optical and crystalline inhomogeneity in evap-
orated zirconia films. Appl. Opt., 24(19):3184.
LaBudde, Edward V., Robert A. LaBudde, and Craig M. Shevlin (1983). Theoretical mod-
eling, calculations, and experiments characterizing the laser-induced hole-formation
mechanism of an in-contact optical disk medium. In Optical Data Storage. Proceedings
of the SPIE, vol. 382, SPIE, Bellingham, Wash.
Leybold-Heraeus Vacuum Products, Inc. (1984). Product and Vacuum Technology Reference
Book, p. 77.
Li, H. H. (1976). Refractive index of alkali halides and its wavelength and temperature
derivatives. J. Physical and Chemical Reference Data, 5(2):329.
Li, H. H. (1980). Refractive index of silicon and germanium and its wavelength and tem-
perature derivatives. J. Physical and Chemical Reference Data 9(3):561.
Liddell, Heather M. (1981). Computer-aided Techniquesfor the Design of Multilayer Filters.
Adam Hilger Ltd., Bristol.
Lyon, D. A. (1946). Method for coating optical elements. United States Patent #2,398,382,
April 16.
Macleod, H. A. (1969). Thin-Film Optical Filters. American Elsevier, New York.
Macleod, H. A. (1983). Design of switching filters. J. Opt. Soc. Am., 73(12):1879.
Maissel, Leon I., and Reinhard GIang, eds. (1970). Handbook of Thin Film Technology.
McGraw-Hili, New York.
Martinu, L., H. Biderman, and L. Holland (1985). Thin films prepared by sputtering mag-
nesium fluoride in an RF planar magnetron. Vacuum, 35(12):531.
Maxwell-Garnett, J. C. (1904). Phil. Trans. Roy. Soc., A203:385 [also see A205:237
(1906)].
McCarthy, Donald E. (1967). Black polyethylene as a far infrared filter. J. Opt. Soc. Am.,
57(5):699.
McNally, James J., Ghanim A. Al-Jumaily, John R. McNeil, and B. Bendow (1986). Ion
assisted deposition of optical and protective coatings for heavy metal fluoride glass.
Appl. Opt., 25(12):1973.
Merck (1976). The Mercklnda. ged. Martha Windholz, ed., Merck and Co., Inc., Rahway,
N.J.
Minot, Michael Jay (1976). Single layer, gradient refractive index antireflection films ef-
fective from 0.35 to 2.5 micrometers. J. Opt. Soc. Am., 66(6):515.
REFERENCES 285

Mittal, K. L. (1978). Adhesion Measurement of Thin Films, Thick Films, and Bulk Coatings.
American Society for Testing and Materials, Philadelphia.
O'Hanlon, John F. (1980). A User's Guide to Vacuum Technology. Wiley-Interscience, John
Wiley & Sons, New York.
Ordal, M. A., L. L. Long, R. J. Bell, S. E. Bell, R. R. Bell, R. W. Alexander, Jr., and
C. A. Ward (1983). Optical properties of the metals AI, Co, Cu, Au, Fe, Pb, Ni, Pd,
Pt, Ag, Ti, and W. Appl. Opt., 22(7):1099.
Presland, A. E. B., G. L. Price, and D. L. Trimm (1972). Hillock formation on
thin silver films. Surface Science, 29:424; and The role of microstructure and surface
energy in hole growth and island formation in thin silver films. Surface Science,
29:435.
Pulker, H. K. (1984). Coatings on Glass. Elsevier, Amsterdam.
Rancourt, 1. (1981). Design and production of tellurium optical data disks. In Advances in
Laser Scanning Technology. Proceedings of the SPIE, vol. 299, SPIE, Bellingham,
Wash.
Rancourt, J. D. (1984). Anti-halo coatings for cathode ray tube faceplates. In Proceedings
of the Society for Information Display, vol. 25, no. I, p. 43.
Rancourt, 1., W. T. Beauchamp, V. Foster, and I. Sachs (1984). Emissivity enhancement
of fused silica for space applications. Proceedings of the Seventeenth IEEE Photovoltaic
Specialists Conference, p. 201.
Rancourt, James D., and Robert L. Martin, Jr. (1986). High temperature lamp coatings. In
Optical Thin Films l/: New Developments. Proceedings of the SPIE, vol. 678, SPIE,
Bellingham, Wash.
Samsonov, G. V., ed. (1978). The Oxide Handbook, 2 ed. IFlJPlenum Press, New York,
p.429.
Sankur, Haluk, and R. Hall (1985). Thin-film deposition by laser-assisted evaporation. Appl.
Opt., 24(20):3335, 3343.
Savage, J. A. (1985). Infrared Optical Materials and their Antireflection Coatings. Adam
Hilger, Ltd., Bristol and Boston, p. 144.
Sax, N. Irving (1984). Dangerous Properties of Industrial Materials, 6 ed. Van Nostrand
Reinhold, New York.
Scheele, S. R. (1977). Scattering from infrared transparent materials. In Stray Light Pro-
blems in Optical Systems. Proceedings of the SPIE, vol. 107, p. 48, SPIE, Bellingham,
Wash.
Schott Optical Glass, Inc., Schott Glass Catalog, Duryea, Penn.
Seddon, R. I. (1977). Opportunities in thin films to meet energy needs. In The Business
Side of the Optical Industry III. Proceedings of the SPIE, vol. III, SPIE, Bellingham,
Wash.
Seeley, J. S., R. Hunneman, and A. Whatley (1980). Temperature-invariant and other
narrow-band IR filters containing PbTe, 4-20 micrometers. In Symposium on Contem-
porary Infrared Sensors and Instruments. Proceedings of the SPIE, vol. 246, p. 83,
SPIE, Bellingham, Wash.
Shurcliff, William A. (1962). Polarized Light Production and Use. Harvard University Press,
Cambridge.
Smith, Howard M., and A. F. Turner (1962). Vacuum-deposited thin films using an optical
maser. J. Opt. Soc. Am., 52(11):1319.
Sommer, A. H. (1973). Stability of photocathodes. Appl. Opt., 12(1):90.
Song, Dar-Yuan, R. W. Sprague, H. Angus Macleod, and, Michael Jacobson (1985).
Progress in the development of a durable silver-based high-reflectance coating for
astronomical telescopes. Appl. Opt., 24(8): 1164.
288 REFERENCES

Southwell, William H. (1980). Multilayer coating design achieving a broadband 90° phase
shift. Appl. Opt., 19(16):2688.
Southwell, William H. (1982). Using deposited thin-film Fabry-Perot cavities to measure
thin-film absorption. J. Opt. Soc. Am., 72(12):1731.
Southwell, William H. (1985). Coating design using very thin high- and low-index layers.
Appl. Opt., 24(3):457.
Spiller, Eberhard (1981). In Attwood, D. T., and B. L. Henke, eds.: Topical Conference
on Low-Energy X-ray Diagnostics. AlP Conference Proceedings 75, p. 124, New York.
Stone, John M. (1963). Radiation and Optics. McGraw-Hill, New York.
Stowers, Irving F. (1978). Advances in cleaning metal and glass surfaces to micron level
cleanliness. Journal of Vacuum Science and Technology, 15(2):751.
Stowers, Irving F., and Howard G. Patton (1978). Cleaning optical surfaces. In Optical
Coatings Applications and Utilization II. Proceedings of the SPIE, vol. 140, SPIE,
Bellingham, Wash., p. 16.
Strong, John (1938). Procedures in Experimental Physics. I ed. Prentice-Hall, New York,
p.376.
Thelen, Alfred (1963). Multilayer filters with wide transmittance bands. J. Opt. Soc. Am.,
53(11):1266.
Thelen, Alfred (1965). Circularly wedged optical coatings. I. Theory. Appl. Opt., 4(8):977.
Thelen, Alfred (1971). Design of optical minus filters. J. Opt. Soc. Am., 61(3):365.
Thelen, Alfred (1973). Multilayer filters with wide transmittance bands D. J. Opt. Soc. Am.,
63(1):65.
Thelen, Alfred (1981). Non-polarizing edge filters. J. Opt. Soc. Am., 71(3):309.
Thelen, Alfred (1987). Design of Optical Interference Coatings. Macmillan, New York.
Turner, A. F., and P. W. Baumeister (1966). Multilayer mirrors with high reflectance over
an extended spectral range. Appl. Opt., 5(1):69.
Underwood, James H. (1986). Multilayer mirrors for X-rays and the extreme UV. Optics
News, 3:20.
Vasicek, A. (1960). Optics of Thin Films. North Holland, Amsterdam.
Walsh, John W. T. (1958). Photometry. 3 ed. Dover, New York.
Weast, Robert C. (1974). CRC Handbook of Chemistry and Physics. 55 ed. CRC Press,
Cleveland.
Wirtenson, G. Richard, and Douglas Flint (1976). Large aperture wedge filter for contrast
enhancement. J. Opt. Soc. Am., 66(2):16l.
Wolfe, W. (1965). Handbook of Military Infrared Technology. Office of Naval Research,
Department of the Navy, Washington, D.C.
Yoldas, B. E., and T. W. O'Keeffe (1979). Antireflective coatings applied from metal-organic
derived liquid precursor. Appl. Opt., 18(18):3133.
Young, Leo (1967). Multilayer interference filters with narrow stop bands. Appl. Opt.,
6(2):297.
Young, Matt (1985). The scratch standard is only a cosmetic standard. Laser Focus/Electro
Optics, 21(11):138.
Young, Matt (1986). Scratch-and-dig standard revisited. Appl. Opt., 25(12):1922.
Young, M., E. G. Johnson, and R. Goldgraben (1985). In Measurement and Effects of
Surface Defects and Quality of Polish. Proceedings of the SPIE, vol. 525, SPIE,
Bellingham, Wash., p. 70.
Young, P. A. (1971). Thermal runaway in germanium. Appl. Opt., 10(3):638.
 Index

abrasion resistance, 160, 161 beam splitters, 115-l2l

abrasion testing, 160, 161, 164, 165 biological effects, 170
absentee layers, 9 birefringence, 21, 22
absorption, 14, 16, 54, 153-155, 183 blocking, 192, 193
absorption coefficient, 4 blocking range, 56-60
adhesion, 46, 160, 163, 164, 166, 180 blotches, 171
agglomeration, 169 BRDF,55
aging, 167-170 Brewster's angle, 20, 21
angle of incidence, 16
angle shift, 68-71 calculations, thin film performance,
annealing, 45 271-279
antireflection coating, 8, 14, 65, 67, 70, calorimetry, 153-155
78-86 channel spectrum, 11
atomic emission monitor, 38 cheesecloth test, 161
automatic deposition, 202, 203 chemical reactions, 168, 169
cleaning, 48-52, 173-176, 185
band gap (table), 58 cleaning, glow discharge, 52
band pass filter, narrow band, 108-114 cleaning, in situ, 52
band pass filter, phase conjugate, 112, cleaning, spray, 50
114 cleaning, ultrasonic, 50
bandpass filters, 103-114 coating efficiency, 41, 43
beam offset, 145 coating equipment, 39-41
287
288 INDEX

coating temperature, 31, 32, 38, 39, 44, heat mirrors, 121
45 Herpin index, 95, 101
coating time, 32 Herpin layer, 95
codeposition, 200, 201 Herpin thickness, 95
coherence length, 10, 11 high power coatings, 133, 134
cohesion, 46 high reflector, 86-99, 171
cold mirrors, 121 high reflector, metallic, 66
color, 79-81, 134, 171, 178, 180 hot mirrors, 121
color correction, 134 humidity, 167, 168
colorimetry, 155-158 humidity testing, 161, 162
conductive coatings, 129-132
corrosion, 75
incidence direction, 13, 54
cosmetic inspection, 163, 170
critical angle, 17 incident medium index effect, 147
inclusions, 173
cryostat, 153, 154
indentation tests, 164
crystal monitor, 35, 36
index of refraction, 4, 6, 13
cycled testing, 165
index of refraction, variable, 134, 135
induced absorbers, 98, 126-130
dark mirrors, 126-129 inhomogeneous films, 200, 201
decomposition, 33 inspection lot size, 181
diamond turning, 32, 185, 189, 190 integration sphere, 146
dichroic filters, 121 intensity addition, 60
dielectric function, 4 interdiffusion, 168, 169
dielectric reflectors, 91-96 interference, 2, 5, 10
digs, 170 ion beam processes, 197-199
dip coating processes, 199, 200 island films, 194, 195
discoloration, 171 110 coatings, 129, 131, 132, 135
distribution, 41, 42
drying, 50
jigging, 43, 44, 171, 183, 184
edge filters, 99-103
effective index, 20 laser coatings, 133, 134
electric field standing waves, 54, 63-70, learning curve, 39
88, 92, 97, 111
electroless processes, 31 magnetron sputtering, 29
electroplating, 31 materials research, 203
emissivity, 38 mean free path, 28, 33
enhanced metalIic reflectors, 96-99 measurement cone, 149-151
eraser test, 161 metal films, 194, 195
evanescent waves, 18 metal reflectors, 86-91, 171, 176
evaporation, flash, 30 microscope, 178-181
evaporation, vacuum, 24-26, 39-41 microscope objectives, 180
expansion coefficient, 187 microscopy, dark field, 181
extinction coefficient, 4 microscopy, Nomarski, 181
microscopy. phase contrast, 181
Fabry-Perot filter, 67, 68, 70, 104-108, mil specs, 166, 167, 210-270
110, ll2, 168 mildew, 170
filter specifications, 207-209 mold, 170
filters in series, 60 multiple surfaces, 145-147
fungus, 170
nonplanar surfaces, 152, 153
glass lamination, 190 nonvacuum processes, 199, 200
glass tempering, 190, 191 notch filters, 130

INDEX 289

optical constants, 13 solubility, 167, 168

optical data disk, 67, 69, 168, 194 solubility testing, 162
optical density, 63 spatter, 172
optical invariant, 17 spectrophotometry, 139-153
optical monitor, 36-38 sputtering, 26-30, 42, 198, 199, 201
optical thickness, 7, 11 stains, 171, 175
optimization, 79 static charge, 176
organometallic processes, 30, 31 steep surfaces, 23, 177
oxidation, 168, 169 sticking coefficient, 38
stoichiometry, 33
particulates, 172 stop band width, 58-62, 94
pass band, 62 stress, 15, 16, 36, 45-48, 178, 203
patterned coatings, 132, 133 stress crazing, 46, 47
performance, 1 substrate degradation, 44, 45
phase shift, 21-23 substrates, 182-194
pinholes, 172, 173 substrates, metal, 189, 190
plastic substrates, 44, 45 substrates, plastic, 188, 189, 198
polarization, 18-20, 22, 23, 119, 148, switching filters, 134, 135
149
pressure chart, 25 tape test, 160
temperature effects, 73, 75
quarter wave stack, 67, 71, 91-96 temperature coating, 166, 184, 186-189,
QW high reflector, nonsymmetric, 91-95 198
QW high reflector, symmetric, 95, 96 tensile stress, 178
QWOT,7,12 testing, environmental, 158-167
testing, sequential, 159, 165
reduction, 168, 169 thermal runaway, 195
reflectance, measurement, 142-145 thickness monitor, 35-38
reflectance, amplitude, 7 tissue cleaning, 176
reflectance, Fresnel, 2, 3 tissue drag, 49
reflectance, interface, 4, 5, 6 tooling, 43, 44, 171, 183, 184
reflectance, phase, 4, 6 topple adhesion test, 163
reflectance, QW film, 82 total internal reflection, 18
reflector, metallic, 6 toxicity, 196
replication, 171, 172 transmission, 16
resistance, sheet, 131 trichroic filters, 121
RestStrahl reflection, 137, 138, 193, 194
Reststrahl reflection band, 88-91 uniformity, thickness, 41-43, 176-178
roughness, surface, 15
V-coat (AR), 83, 84
salt fog testing, 162
vapordegreaser, 50, 51
sample heating, 153
visual inspection, 163
sand test, 163
voids, 172
scatter, 14, 15, 55, 56, 151, 171, 172
scatter, total integrated, 55
scattering, 14, 15, 55, 56, 151, 171, 172 wavefront distortion, 75
scratches, 170, 191 wavenumber, 62, 63, 94, 139, 140
semiconductor films, 195 weighting function, 79
shelf life, 167, 168 wide band AR, 85, 86
sleeks, 170 witnesses, 159, 160, 177
Snell's Law, 16, 17
solar absorbers, 129 X-ray coatings, 201, 202