The Art of Radiometry - J.M. Palmer B. G. Grant

The
Art
Radiometry
James M. Palmer
Barbara G. Grant
SPIE
PRESS
Bellingham, Washington USA
Library of Congress Cataloging-in-Publication Data
Palmer, James M.
Art of radiometry / James M. Palmer and Barbara G. Grant.
p. cm. -- (Press monograph; 184)
Includes bibliographical references and index.
ISBN 978-0-8194-7245-8
1. Radiation--Measurement. 1. Grant, Barbara G. (Barbara Geri), 1957- II.
Title.
QDl17.R3P352009
539.7'7--dc22
Pub lished by
SPIE
P.O. Box 10
2009038491
Bellingham, Washington 98227-0010 USA
Phone: + 1 360.676.3290
Fax: + 1 360.647.1445
Email: Books@spie.org
Web: http: //spie.org
Copyright 2010 Society of Photo-Optical Instrumentation Engineers (SPIE)
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.
The content of this book reflects the work and thought of the author(s).
Every effort has been made to publish reliable and accurate information herein,
but the publisher is not responsible for the validity of the information or for any
outcomes resulting from reliance thereon.
Printed in the United States of America.
On the cover: A Crooke radiometer and the equation of radiative transfer.
SPIE
Contents
Foreword ....... .... ....... .. ...... .... ... .. ... ... ........ .. .... .. ..... ............. .. ... ... ..... .. .. ....... xi
Preface ....... ................ ........ ..... ... ....... ..... ..... .......... ...... ... ......................... xiii
Chapter 1 Introduction to Radiometry 11
1.1 Definitions ................... ... .... ......................................... ......................... 1
1.2 Why Measure Light? ............. ................................. ............................ . 2
1.3 Historical Background ............ ................................. .... ......................... 4
1.4 Radiometric Measurement Process ... .......... .. .... ......... .. ...... .. ...... ........ 5
1.5 Radiometry Applications ...... ... ... .. .... .. ... .. .. .. .. .... .......... ...... ............. ..... . 7
References .. .... ..... ..... .. ..... ... .... ...... .. ... ... ..... ....... ..... ..... ... .. ..... ........... ....... .. 9
Chapter 2 Propagation of Optical Radiation 111
2.1 Basic Definitions ....................................... .............. ......... .................. 11
2.1 .1 Rays and angles ..... ............................................................... 11
2.1 .2 System parameters ................................ ... .... ... ... ..... ............. 19
2.1 .3 Optical definitions .. .. .............. .. ............. ........ .. ....................... 23
2.2 Fundamental Radiometric Quantities .... ................ .... .... ......... ........... 24
2.2.1 Radiance .. .. ..... ... ... ........... .... ..... .. .. ............. .... ....... ...... .... .. .... 24
2.2.2 Radiant exitance .. ... .... .. ...... ...... ........ ...... ... .... ... ... .. .... .......... .. 26
2.2.3 Irradiance .... ........ ... ....................... .......... ..... .. ... .................... 28
2.2.4 Radiant intensity .. ... ...................................... ... ...................... 29
2.3 Radiometric Approximations ...... ........................................................ 30
2.3.1 Inverse square law ....................................... ......................... 30
2.3.2 Cosine
3
law .......... ..................................... ............................ 31
2.3.3 Lambertian approximation ..................................................... 32
2.3.4 Cosine
4
law ..... ..... .. .......................... ......... .... ..... ................... 33
2.4 Equation of Radiative Transfer ..................... .. ... .......... ....... .... ........... 36
2.5 Configuration Factors .. ...... ...... .. .... ... ..... .... .... .. ..... ..... ... ....... .... .. .... .. .. 38
2.6 Effect of Lenses on Power Transfer ...... ..... ...... ...... ...... .............. ... ... . 40
2.7 Common Radiative Transfer Configurations ............ .. .......... ......... .... 42
2.7.1 On-axis radiation from a circular Lambertian disc .... .. ........... 42
2.7.2 On-axis radiation from a non-Lambertian disc ...................... 43
2.7.3 On-axis radiation from a spherical Lambertian source .......... 44
2.8 Integrating Sphere .................................................... ....... .................. 46
2.9 Radiometric Calculation Examples .............................. ............ .......... 48
2.9.1 Intensities of a distant star and the sun .. ..... .... .................... .. 48
v
vi Contents
2.9.2 Lunar constant. .......... .. ................... .... ................ .. ................. 50
2.9.2.1 Calculation .......... ................... .................. .. .................... 50
2.9.2.2 Moon-sun comparisons ........... .................. ................ ... . 51
2.9.3 "Solar furnace" ...... ....... ..... ....... .. .... .. ............. .. .... .. ............... .. 52
2.9.4 Image irradiance for finite conjugates .. .. .................... .. .. .. ..... 53
2.9.5 Irradiance of the overcast sky ...... .. ................ .... .. .. ............ .. . 55
2.9.6 Near extended source .................. .... ............... .. .................... 55
2.9.7 Projection system ...... ... .. ... .......... ..... .. ................ .. ................. 56
2.10 Generalized Expressions for Image-Plane Irradiance .. .. .. .... ... ........ 57
2.10.1 Extended source ........ .. ......................................................... 57
2.10.2 Point source ............ ... ................ .. ..... .......... ......... ................. 58
2.11 Summary of Some Key Concepts ........ .. ................ .. .. .. ................... 58
For Further Reading ...... ............ .... ..... ......... .. .... ... .. ............. ..... .............. . 59
References ..... ...... ............... .. ...... ..... ............ .... .............. ........................ . 59
Chapter 3 Radiometric Properties of Materials I 61
3.1 Introduction and Terminology .. .......................................................... 61
3.2 Transmission .. ... ... ..... ......... .. .. .. ................ ................ .... .................. ... 62
3.3 Reflection ........ ... ... ... ............ ..... .... ..... ...... .. .................. ................ ... .. 63
3.4 Absorption ......... ... ................. ... ................ .................. .... ................... 69
3.5 Relationship Between Reflectance, Transmittance, and
Absorptance ...... .. .................... ................................... ....................... 69
3.6 Directional Characteristics ... .. ....... ....... ... ..... ..... ..... ...... ... ...... ..... ..... ... 69
3.6.1 Specular transmittance and reflectance .............. .... ...... ....... . 69
3.6.2 Diffuse transmittance and reflectance ................................ .. . 73
3.7 Emission ................................. .. ..................................... ... ................. 76
3.8 Spectral Characteristics .................................................................... 77
3.9 Optical Materials Checklist .. .... ..... ... ..... ... .. .... ... ... ...... .. .... .... .... .......... 79
For Further Reading .. ... ................ ...... .................................. ............ ....... 80
References ................. .. ............. ..... ................ .. .................. .. ................ ... 80
Chapter 4 Generation of Optical Radiation I 83
4.1 Introduction ............. ... .... ........ ..... ..... ... .. .... ..... ..... ....... ...... ...... ......... ... 83
4.2 Radiation Laws .................................................................................. 84
4.2.1 Planck's law ............. ... ................................... ... .. ................... 84
4.2.2 Wien displacement law ............... .. .. .................... .. .............. .. . 86
4.2.3 Stefan-Boltzmann law ..... .. ........... .... ... ............. ... .. .............. .. 89
4.2.4 Laws in photons ...... .......... ..... ..... .. ................... .............. .. .... . 89
4.2.5 Rayleigh-Jeans law ..................... .. ........................................ 92
4.2.6 Wien approximation .... ... ................. ................. .. .................... 93
4.2.7 More on the Planck equation ...... ...... ................. .. .. ... ............. 93
4.2.8 Kirchhoff's law ...... .... ... .......... ..... ... .... ..... .... ... ..... ..... ... ......... .. 97
4.3 Emitter Types and Properties .. ....... .......... .. ... ............... ........... ... ..... 102
4.3.1 Metals ...................... .......................................... ... ............... 102
..
Contents vii
4.3.2 Dielectrics ... ........... .............................................................. 102
4.3.3 Gases .................................................................................. 103
4.4 Practical Sources of Radiant Energy ............................................... 104
4.4.1 Two major categories .......................................................... 104
4.4.2 Thermal sources ........... .... ............. ......................... ............. 105
4.4.2.1 Tungsten and tungsten-halogen lamps .................. ...... 105
4.4.2.2 Other metallic sources .................................................. 108
4.4.2.3 Dielectric thermal sources ............................................ 108
4.4.2.4 Optical elements ........................................................... 109
4.4.2.5 Miscellaneous thermal sources ... ....... ............ ....... ....... 109
4.4.3 Luminescent sources ..... ......................... ............................ 110
4.4.3.1 General principles ........................................................ 110
4.4.3.2 Fluorescent lamps ........................................................ 115
4.4.3.3 Electroluminescent sources ......................................... 117
4.4.3.4 LED sources ...... ... ..... .... ...... .......... ........... ..... .. ............. 117
4.4.3.5 Lasers .......................................................................... 118
4.4.4 Natural sources ................................... ................................ 119
4.4.4.1 Sunlight ........................................................................ 119
4.4.4.2 Skylight, planetary, and astronomical sources ............. 120
4.4.4.3 Application: energy balance of the earth ....... .... ........... 121
4.5 Radiation Source Selection Criteria ................................................. 121
4.6 Source Safety Considerations ......................................................... 123
4.7 Summary of Some Key Concepts .................. ................................. 123
For Further Reading .............................................................................. 123
References ................... ....... ........... ...... ................ .. ................. ... ...... ..... 124
Chapter 5 Detectors of Optical Radiation 1127
5.1 Introduction ....................................................... ............................... 127
5.2 Definitions ... .. ......... ... .. .. .. ........... .... .. .............. ..... .... .. .......... ........ .. ... 128
5.3 Figures of Merit ................................. ................... ...................... ..... 131
5.4 #N$O%&I*S@E-A ........................................................................... 133
5.4.1 Introduction to noise concepts ............................................. 133
5.4.2 Effective noise bandwidth .................................................... 136
5.4.3 Catalog of most unpleasant noises ... .. ... ................... .. .. ..... . 137
5.4.3.1 Johnson noise ......... ..................................................... 137
5.4.3.2 Shot noise .................................................................... 139
5.4.3.3 1 If noise ................... ..................................................... 139
5.4.3.4 Generation-recombination noise .................................. 140
5.4.3.5 Temperature fluctuation noise ...... .. ............... .. .. ........... 141
5.4.3.6 Photon noise ................................................................ 141
5.4.3.7 Microphonic noise ........................................................ 142
5.4.3.8 Triboelectric noise ........................................................ 142
5.4.3.9 CCD noises ................................. .................... ............. 142
5.4.3.10 Amplifier noise ......... .. ..... ....... ..... .. ......... .... .... ... .......... .. 143
5.4.3.11 Quantization noise ......................................... ............... 143
viii Contents
5.4.4 Noise factor, noise figure, and noise temperature ............... 143
5.4.5 Some noise examples ........................ .. ............................... 144
5.4.6 Computer simulation of Gaussian noise .............................. 147
5.5 Thermal Detectors ........................................................................... 147
5.5.1 Thermal circuit ...................................... ............................... 147
5.5.2 Thermoelectric detectors ......................... .... ........................ 150
5.5.2.1 Basic principles ...................................... ...................... 150
5.5.2.2 Combinations and configurations .. .. ................... .. ........ 153
5.5.3 Thermoresistive detector: bolometer .... .. ............................. 155
5.5.4 Pyroelectric detectors ....... ..... .... ........... ... ............................ 157
5.5.4.1 Basic principles .............................. .. ............................ 157
5.5.4.2 Pyroelectric materials ................................................... 160
5.5.4.3 Operational characteristics of pyroelectric detectors ... 162
5.5.4.4 Applications of pyroelectric detectors ........................... 162
5.5.5 Other thermal detectors ....................................................... 163
5.6 Photon Detectors ..... .... .................................................................... 164
5.6.1 Detector materials .................. ........................ .. .. .. .. .... .... .... . 164
5.6.2 Photoconductive detectors .... .. .... .. ...... .... .. .. ... .. .. .... .... .... .... . 169
5.6.2.1 Basic principles ...... ....... .. ... ......... ... .................... .... .... .. 169
5.6.2.2 Noises in photoconductive detectors ........................... 173
5.6.2.3 Characteristics of photoconductive detectors .............. 174
5.6.2.4 Applications of photoconductive detectors ................... 175
5.6.3 Photoemissive detectors ................... .... .............................. 175
5.6.3.1 Basic principles ............................. ............................... 175
5.6.3.2 Classes of emitters ....................................................... 176
5.6.3.3 Dark current ................................. .... .. .......................... 181
5.6.3.4 Noises in photoemissive detectors .. ... ......... .. .... ........... 182
5.6.3.5 Photoemissive detector types ...... ... ........ .. ................... 183
5.6.4 Photovoltaic detectors ........................ ........ .. ....................... 185
5.6.4.1 Basic principles ............................................................ 185
5.6.4.2 Responsivity and quantum efficiency ........................... 195
5.6.4.3 Noises in photovoltaic detectors .. .. .............................. 196
5.6.4.4 Photovoltaic detector materials and configurations ...... 198
5.7 Imaging Arrays ......... ... .................................. .... .............................. 199
5.7.1 Introduction .... ....................................... ............................... 199
5.7.2 Photographic film ................................... .... .. .. .... .. ................ 199
5.7.2.1 History ...... ...................... ... ..... .. ... ....... ......... .. ........ .. .. ... 199
5.7.2.2 Physical characteristics .... ... ... .... ... ... ............ ........ ... .... . 201
5.7.2.3 Spectral sensitivity .... ................. ..... ............................. 201
5.7.2.4 Radiometric calibration ................................................. 201
5.7.2.5 Spatial resolution .......................................................... 202
5.7.2.6 Summary ....................................... .. ............................. 202
5.7.3 Electronic detector arrays ....................... ............................. 203
5.7.3.1 History ......................................... .. .. ............................. 203
5.7.3.2 Device architecture description and tradeoffs .............. 203
Contents ix
5.7.3.3 Readout mechanisms ......................... ......................... 204
5.7.3.4 Comparison ..... .. ..................................... ...................... 207
5.7.4 Three-color CCOs ... ............................................................ 207
5.7.5 Ultraviolet photon-detector arrays ........... ............................ 208
5.7.6 Infrared photodetector arrays ..................... ... .... ....... ... ........ 209
5.7.7 Uncooled thermal imagers ........ .................. ... ............... ...... 210
5.7.8 Summary .... ....... ..... ... ..... ... ... ............. .. .... ........ ..... .. ....... .. .... 211
For Further Reading ... ..... ..... .... ........... ..... ... ... ..... .. ....... ..... ... .............. ... 211
References ......... ........ ...... .. .... ..... ....... ....... ......... ..... ......... ........... .... ...... 213
Chapter 6 Radiometric Instrumentation I 215
6.1 Introduction ........................ .................................. .. ..... ..................... 215
6.2 Instrumentation Requirements ................................ ...... .. ................ 215
6.2.1 Ideal radiometer .. ......................................... ....................... 215
6.2.2 Specification sheet ............................ .... ... ... ........................ 215
6.2.3 Spectral considerations ..... .... .. .... .. .... ...... ... .... .. ... ........... .. ... 216
6.2.4 Spatial considerations .. ..... ... .. ........ ... ... ...... .... ... .. ............. .. . 217
6.2.5 Temporal considerations ............. .. ..... ....... ... ... ... ..... ...... ...... 217
6.2.6 Make or buy? .. ... .... ................................... .... ...................... 218
6.3 Radiometer Optics ....... ..... ..... ............... ................... ... ..................... 218
6.3.1 Introduction .......... ................................................................ 218
6.3.2 Review of stops and pupils .................................................. 218
6.3.3 The simplest radiometer: bare detector. .... ... ..... .................. 219
6.3.4 Added aperture ........ ........... .......... ............. ..... ...... ............... 219
6.3.5 Basic radiometer .. ... ... ... ............................ ... ...... ..... .. .......... 221
6.3.6 Improved radiometer ... ..... .. ....... ......... ... ...... ... ... ... ......... ...... 223
6.3.7 Other methods for defining the field of view .... ...... ........ .. .... 224
6.3.8 Viewing methods .... ............................... .. ... ..... ...... .............. 224
6.3.9 Reference sources ... .. ........... ....................... ... ..... ............... 226
6.3.10 Choppers .. ...... .. .. ....... ............................ ....... .... ................... 226
6.3.11 Stray light ........ .... ... ...................................... ....................... 227
6.3.12 Summing up ........ ...... .......................................................... 228
6.4 Spectral Instruments .. ........... ........................... ......... ... ................... 228
6.4.1 Introduction .... .. .... ................................. ..... .......................... 228
6.4.2 Prisms and gratings ............... ..................... .. .... .... .. .......... ... 230
6.4.3 Monochromator configurations .. .. ..... ... .. ......... .. .... ... ........ ... . 231
6.4.4 Spectrometers ...... ... ... ...... .. .. ............ ... ...... ...... ....... ............ . 234
6.4.5 Additive versus subtractive dispersion ..... .. ..... .. ..... ......... .... 235
6.4.6 Arrays ............... ............................................ ... ... ................. 236
6.4.7 Multiple slit systems ... ................................... ............. ......... 236
6.4.8 Filters ............... ..... .. ..................................... ... ..................... 236
6.4.9 Interferometers .... .. ..... ......................................................... 237
6.4.10 Fourier transform infrared ............ ........ ....... .......... ............... 237
6.4.11 Fabry-Perot ..... ... .... ................................ .. .... ............. .......... 238
x Contents
For Further Reading ............................................................................. . 240
References ................................ ................... ......................................... 240
Chapter 7 Radiometric Measurement and Calibration 1241
7.1 Introduction .... ... ................. ......... ..................................................... 241
7.2 Measurement Types .... ....... ....... ...................................................... 241
7.3 Errors in Measurements, Effects of Noise, and Signal-to-Noise Ratio
in Measurements ................ .. ................. .......................................... 241
7.4 Measurement and Range Equations ... ....................... ..................... 250
7.5 Introduction to the Philosophy of Calibration .... ............. .... .... ..... .... . 253
7.6 Radiometric Calibration Configurations ........................................... 257
7.6.1 Introduction .......................................................................... 257
7.6.2 Distant small source ............................................................ 258
7.6.3 Distant extended source ............. .................... ..................... 260
7.6.4 Near extended source ............. .................... .... .... ... .... .... ..... 261
7.6.5 Near small source .. ...................... ...... ................................. 262
7.6.6 Direct method ...................................................................... 262
7.6.7 Conclusion ........................................................................... 263
7.7 Example Calculations: Satellite Electro-optical System .......... ........ 263
7.8 Final Thoughts .................... .. ... ............. ....... ...... ........ ... .... ............ ... 267
For Further Reading ............................................................................. . 268
References .......... .. .............. .................................................................. 268
Table of Appendices 1 269
Appendix A: Systeme Internationale (SI) Units for Radiometry and
Photometry ......... .............................................................. 271
Appendix B: Physical Constants, Conversion Factors, and Other Useful
Quantities ........... .............. .. ............................................... 275
Appendix C: Antiquarian's Garden of Sane and Outrageous
Terminology ...................................................................... 277
Appendix D: Solid-Angle Relationships ........ .................... .. ................. .. 283
Appendix E: Glossary ............................................................................ 285
Appendix F: Effective Noise Bandwidth of Analog RC Filters and the
Selection of Filter Parameters to Optimize Signal-to-Noise
Ratio ................................................................................. 297
Appendix G: Bandwidth Normalization by Moments ........ .. ................... 305
Appendix H: Jones Near-Small-Source Calibration Configuration ........ 309
Appendix I: Is Sung lint Observable in the Thermal Infrared? .............. 313
Appendix J: DocumentaiY Standards for Radiometry and Photometry 321
Appendix K: Radiometry and Photometry Bibliography ........................ 341
Appendix L: Reference List for Noise and Postdetection Signal
Processing ......... .... .................. .. ................ ..... ................. . 357
Index 1361
Foreword
The material for this book grew out of a first-year graduate-level course,
"Radiometry, Sources, Materials, and Detectors," that Jim Palmer created and
taught at the University of Arizona College of Optical Sciences for many years.
The book is organized by topic in a similar manner, with the first five chapters
presenting radiation propagation and system building blocks, and the final two
chapters focusing on instruments and their uses. Chapter 1 provides an overview
and history of the subject, and Chapter 2 presents basic concepts of radiometry,
including terminology, laws, and approximations. It also includes examples that
will allow the reader to see how key equations may be used to address problems
in radiation propagation. Chapter 3 introduces radiometric properties of materials
such as reflection and absorption, and Chapter 4 extends that discussion via a
detailed consideration of sources. Point and area detectors of optical radiation are
considered in Chapter 5, which also includes thermal and photon detection
mechanisms, imaging arrays, and a discussion about film.
In Chapter 6, the focus shifts to instrumentation. Concepts introduced in
Chapter 2 are here applied to instrument design. Practical considerations relating
to radiometer selection are detailed, and a "Make or Buy?" decision is explored.
Several monochromator configurations are examined, and spectral instruments
are discussed. Proceeding from instruments to their uses, Chapter 7 details types
of measurements that may be made with radiometric systems and provides a
discussion of measurement error. The philosophy of calibration is introduced,
and several radiometric calibration configurations are considered.
The material in the appendices covers a variety of topics, including
terminology, standards, and discussions of specific issues such as Jones source
calibration and consideration of solar glint. Due to Jim's attention to detail and
the length of time over which he accumulated material, the long lists he provided
here may be viewed as comprehensive, if not current by today's standards.
The level of discussion of the material is suitable for a class taught to
advanced undergraduate students or graduate students. The book will also be
useful to the many professionals currently practicing in fields in which
radiometry plays a part: optical engineering, electro-optical engineering, imagery
analysis, and many others.
In 2006, Jim Palmer was told that he was terminally ill, and he asked me to
complete this work. I was humbled and honored by the request. I'd met Jim as a
graduate student in optical sciences in the late 1980s, and he had served on my
thesis committee. My career after graduation had focused on systems engineering
and analysis, two areas in which radiometry plays a significant role. For nearly
the last ten years of Jim's life, I'd been able to receive mentoring from the master
simply by showing up at Jim's office door with a question or topic for discussion,
but I never anticipated that our discussions would one day come to an end. Upon
Jim's death, I sought to weave his collection of resources and narrative together
xi
xii Forward
with newer material and discussion in a manner I hope will be both informative
to read and valuable to reference. The preface that follows was written by Jim
before he died and has been left as he wrote it.
I am grateful for the assistance of many. First is William L. Wolfe, Jim's
professor and mentor, who offered helpful comments on each chapter and
adapted Chapter 6 on radiometric instrumentation. Others for whose help I am
grateful, all from or associated with the University of Arizona College of Optical
Sciences, are Bob Schowengerdt, who contributed the narrative on film; Anurag
Gupta of Optical Research Associates, Tucson, Arizona, who adapted the
appendix material; and L. Stephen Bell, Jim's close friend and colleague, who
revised the signal processing discussion that appears in that section and provided
a complete bibliography on the subject. A special note of thanks goes to Eustace
Dereniak, who provided office space for me, helpful discussions, and hearty
doses of encouragement. I also wish to thank John Reagan, Kurt Thome (NASA
Goddard Spaceflight Center, Greenbelt, Maryland), Mike Nofziger, and Arvind
Marathay for review, discussion, and helpful insights. In addition, I am grateful
for the assistance of Anne Palmer, Jim's beloved sister, and University of
Arizona College of Optical Sciences staff members Trish Pettijohn and Ashley
Bidegain. Gwen Weerts and Tim Lamkins of SPIE Press have my gratitude for
the special assistance they provided to this project. I also gratefully acknowledge
Philip N. Slater, my professor in optical sciences, who selected me as a graduate
student and trained me in remote sensing and absolute radiometric calibration
from 1987 to 1989, and Michael W. Munn, formerly Chief Scientist at Lockheed
Martin Corporation, who instilled the value of a systems perspective in the
approach to technical problems.
Finally, I am grateful to my family for providing financial support; to Ralph
Gonzales, Arizona Department of Transportation, and Sylvia Rogers Gibbons for
providing professional contacts; and my friends at Calvary Chapel, Tucson,
Arizona, whose donations and prayers sustained me as I worked to complete this
book.
Barbara G. Grant
Cupertino, California
October 2009
Preface
This volume is the result of nearly twenty years of frustration in locating suitable
material for teaching the subject of radiometry and its allied arts. This is not to
say that there is not a lot of good stuff out there-it's just not all in one place,
consistent in usage of units, and applicable as both a teaching tool and as a
reference. I intend this book to be all things to all people interested in radiometry.
The material comes from teaching both undergraduate and graduate-level courses
at the Optical Sciences Center of the University of Arizona, and from courses
developed for SPlE and for industrial clients. I have unabashedly borrowed the
tenor of the title from the superb text The Art of Electronics by Paul Horowitz in
the hope that this volume will be as useful to the inquisitive reader.
I gratefully acknowledge the contributions of my mentor, William L. Wolfe,
Jr., and the hundreds of students whose constant criticism and occasional faint
praise have helped immeasurably.
This book is dedicated to the memory of my mother, Candace W. Palmer
(1904-1996) and my father, James A. Palmer (1905- 1990). She was all one
could wish for in a Mom, and he showed me the path to engineering.
James M Palmer
1937-2007
xiii
Chapter 1
Introduction to Radiometry
1.1 Definitions
Consider the following definitions a starting point for our study of radiometry:
radio- [<L. radius] a combining form meaning ray, raylike
-metry [Gr. -metria < metron] a terminal combining form meaning the process,
art, or science of measuring
radiometer (ra'de-om'i-ter) n. [radio- + -meter], an instrument for detecting
and measuring the intensity of radiant energy, by exposing to sunlight a set of
vanes blackened on one side and suspended on an axis in a vacuum and
measuring their speed of rotation (i.e., the mechanical energy into which the
radiant energy has been converted) (See Fig. 1.1)
radiometry (ra'deom'i-tri), n. the use of the radiometer: the measurement of
radiation!
These definitions are taken from Webster's New World Dictionary, and may
be satisfactory for the general nonscientist. The definitions are not satisfactory,
however, for scientists and engineers pursuing the art of radiometry. So let's get
technical:
radiometry (ra'deom'itri) n. the measurement of optical radiant energy
The practical electromagnetic spectrum extends from dc to frequencies
greater than 10
20
Hz. The optical portion of the spectrum covers the five-decade
frequency range from 3 x 10
11
to 3 X 1O!6 Hz, corresponding to the wavelength
range from 10 nm to 1000 flm, as shown in Fig. 1.2. This range includes the
ultraviolet, visible, and infrared regions. Shorter wavelengths are called x rays
and gamma rays, while longer wavelengths are microwave and millimeter radio
waves.
2 Chapter 1
Figure 1.1 Classic vane radiometer, commonly called the Crooke radiometer.
1
[Reprinted
by permission from Webster's Third New Intemationa/ Dictionary, Unabridged 1993 by
Merriam-Webster, Incorporated (www.Merriam-Webster.com)].
The optical radiation spectrum will be treated in this text, including the
ultraviolet, visible, and infrared regions. The visible portion of the optical
spectrum covers a rather narrow band of wavelengths between 380 nm and
760 nm; the radiation between these limits, perceivable by the unaided normal
human eye, is termed "light." Measurements within this region may be called
"photometric" if the instruments used incorporate the response of the eye. The
short wavelength (ultraviolet) limit of radiometric coverage is about 200 nrn,
approximately the shortest wavelength that our atmosphere will transmit. The
longest wavelength (infrared) treated in this book is about 100 Ilm. This
wavelength range includes 99% of the energy (95% of the photons) from a
thermal radiator at 0 C (273.16 K).
1.2 Why Measure Light?
But why measure light in the visible, ultraviolet, or even infrared region? What
are these measurements good for? Let's look at some historical perspectives:
I often say that when you can measure what you are speaking
about, and express it in numbers, you know something about it;
but when you cannot measure it, when you cannot express it in
numbers, your knowledge is of a meager and unsatisfactory
kind; it may be the beginning of knowledge, but you have
scarcely, in your thoughts, advanced to the stage of science,
whatever the matter may be.
Lord Kelvin
/.. Optical -I
Visible

: !.: iI Microwave Radio
Gamma I x -rays ,UltraViolet I : Infrared' A'--__ -= __ =--_ __=__
Rays : : : : : 'EHF SHF UHF VHF HF MF LF VLF'
I I [[ I J I I I I I I I
0.11 11 IOl 1001 Ip.m 10p.m loop.m O.lcm lem IOcm I m 10m 100m I km 10 km IOOkm
/ .....
3,,10
19
3,,10
18
3x[(P3xid6 3,,10
15
3xKf' 3xlO
l3
3x/012 3-;10" 3x10
'0
3x10
9
3xl0
8
3xlO
1
3,,10
6
3xlO
5
3xl0
4
3xl0
3
/ --
/ -.....
/ -
/ --
/ --
/ -
/ --
/ --
/ --
UV
,
I Middle Infrared:
I I
, I
Far Infrared
1 Extreme
: Visible
iVBGYOR
Near Infrared
I
WAVELENGTH
FREQUENCY, Hz
0.4 0.6 0.8 I 1.5 2 3 4
25,000 10,000 5000 2500
6 8 10
1000
15 20
500
30 WAVElENGTH, /LITI
WAVENMBER; cm-
I
Figure 1.2 The electromagnetic spectrum.
2
[Reprinted by permission of author from Optical Radiation Measurement series,
Vol. 1, F. Grum and R. J. Becherer, Radiometry, p. 1 (1979)].
:J
a
g-
o
a-
::J
o
&?

3
(l)

VJ
4 Chapter 1
... nobody will object to an ardent experimentalist boasting of his
measurements and rather looking down on the "paper and ink"
physics of his theoretical friend, who on his own part is proud of
his lofty ideas and despises the dirty fingers of the other.
Max Born
If you are really doing optics, you get photons under your
fingernails.
James M Palmer
Measurement is the point at which the rubber meets the road. Hypotheses,
uncorroborated by measurement, cannot fulfill the same function. And if rubber
doesn't meet the road, the car cannot move.
The measurement of light is often critical in transitioning from theory to the
development of systems and techniques. Although instrument and system design
may be based on theory, performance evaluation and system improvement
require that accurate radiometric measurements be applied. When calibrated
measurements are needed, that is, when field or laboratory measurements must
be correlated with specific values presenting the relationship between measured
phenomena and an absolute standard, radiometric measurements take on even
greater significance.
1.3 Historical Background
Scientists and engineers have been involved in the measurement of light since the
early experiments and instruments described by P. Bouguer in 1729 and by J. H.
Lambert in 1760. Exploration into other spectral regions began with the
discovery of the infrared region by W. Herschel in 1800 and the ultraviolet
region by J. W. Ritter the following year. Table 1.1 shows some of the significant
events in the history of radiometry and photometry.
Year
?
1666
1729
1760
1800
1801
1830
1837
1839
1859
Table 1.1 Some significant events in radiometry.
Event
... and then there was light!
Investigation of the visible spectrum
Inverse square law
Cosine law, exponential absorption
Discovery of the infrared region
Discovery of the ultraviolet region
Radiation thermopile (first practical
detector)
Calorimetric detector
Photoelectric effect
Relation between absorption and emission
Principal investigator
from Genes is
Newton
Bouguer
Lambert
Herschel
Ritter
Nobile, Melloni
Pouillet
Becquerel
Kirchhoff
Year
1860
1879
1880
1892
1893
1900
1900
1905
1910
1931
1936
1938
1948
1954
1955
1960
1961
1970
1975
1977
1980
1983
1984
1985
Table 1.1 (Continued.)
Event
Standard lamp fueld by sperm whale oil
Incandescent lamp (carbon filament)
Bolometer
Integrating sphere (theory)
Absolute thermal detector (pyrheliometer)
Blackbody radiation theory
Integrating sphere (reduction to practice)
Photoelectric effect
Tungsten lamp
Adoption of colorimetric standards
Photomultiplier tube (multistage)
Pyroelectric detector (theoretical)
Adoption of platinum blackbody for
standard candela
Silicon photodiode
Pyroelectric detector (reduction to practice)
Invention of light amplification by
stimulated emission of radiation
(LASER)
Tungsten-halogen lamp
Laser calorimetry
Fourier transform spectrometer
Photometry relegated to subset of
radiometry
Self-calibrated silicon detector
Trap detector
Definitive measurement of the Stefan-
Boltzmann constant
Cryogenic absolute radiometer
1.4 Radiometric Measurement Process
5
Principal investigator
British Metropolitan
Gas Act
Edison
Langley
Sumpner
Angstrom, Kurlbaum
Planck
Ulbricht
Einstein
Collidge
International
Commission on
Illumination (CIE)
Zworykin, Morton &
Malter
Ta
Consultative
Committee on
Photometry and
Radiometry (CCPM)
Chapin, Fuller &
Pearson
Chenowyth
Maiman
Zubler & Mosby
West
Vanesse
CCPM
Zalewski & Geist
Zalewski & Duda
Quinn & Martin
Martin, Fox, & Key;
Foukal
This book describes the many facets of optical radiation measurement, from
radiation sources to detectors and signal processing. To fully understand and
appreciate a radiometric measurement, we must understand the processes of
6 Chapter 1
generation, transmission, and detection of optical radiation. In addition, we must
possess a firm grasp of the underlying mathematics and what is loosely called
"measurement science."
A generic radiometric configuration is shown in Fig. 1.3. A target, or object
of measurement interest, can be either active, emitting radiation by virtue of its
temperature or some form of atomic excitation, or passive, reflecting radiation
from a different active or passive illuminator. Examples of active sources include
the sun, tungsten or fluorescent lamps, lasers, and any nontransparent object with
a temperature greater than 0 K. Passive sources include the entire range of natural
and artificial reflective surfaces that make up our environment. An additional
source of optical radiation can be classified as background, the radiation that may
be in our instrumental field of view along with the target. Also included is the
intervening medium, the atmosphere, which includes both radiation sources and
sinks, acting through the mechanisms of absorption, emission, and scattering.
The myriad of small arrows in Fig. 1.3 represent scattered, absorbed, emitted,
and reflected radiation.
After traversing the atmosphere, the rays from the target (and possibly the
background as well) reach our instrument, whose parameters define the ranges
over which the spatial, spectral, temporal, and radiometric characteristics of
incoming radiation will be accepted. This is accomplished through the use of
lenses and mirrors, choppers and apertures, prisms, gratings, filters, attenuators,
Figure 1.3 Generic radiometric configuration.
2
[Reprinted by permission of author from
Optical Radiation Measurement series, Vol. 1, F. Grum and R. J. Becherer, Radiometry, p.
7 (1979)].
Introduction to Radiometry 7
polarizers, optical fibers, etc. The optical radiation transmitted through the
instrument is finally incident on one or an array of detectors, transducers which
convert the incident optical radiation to a more tractable form of energy.
Detectors may be thermal (thermoelectric, bolometric, and pyroelectric) or
photon (photoemissive, photoconductive, and photovoltaic) in character; other
viable detectors include the human eye and photographic film.
The final block in the diagram involves signal processing. Most of the
detectors in common use generate electrical signals. Postdetector processing may
include filtering, linearization, and background suppression before the processed
result is recorded and displayed. Even the eye and film detectors include
processing steps, such as the filtering and interpretation of information by the
brain and the photographic development process for film.
1.5 Radiometry Applications
The fields in which radiometric instruments and techniques are used are very
diverse. Table 1.2 lists some of the more common applications.
Table 1.2 Common applications of radiometry.
Appearance measurement
Astrophysics
Atmospheric physics
Clinical medicine
Colorimetry
Diagnostic medicine
Remote-sensing satellites
Electro-optics
Illumination engineering
Laser measurements
Materials science
Meteorology
Military systems
Night-vision devices
Optoelectronics
Photobiology
Photochemistry
Photometry
Radiative heat transfer
Solar energy
Television systems
Visual displays
Vision research
8 Chapter 1
Most books on radiometry begin with a vast and often confusing array of
terms, definitions, etc. In this work, detailed listing of terminology is relegated to
the glossary in Appendix E, and radiometric terms will be introduced as they are
needed.
Radiometry and photometry are applied to a variety of phenomena whose
output occurs over many orders of magnitude. Tables 1.3 through 1.5 illustrate
the ranges of illumination encountered. "Luminance" is power per unit area and
unit solid angle weighted by the spectral response of the eye; its units are lumens
per square meter per steradian (lmlm2/sr), or candelas per square meter (cd/m2).
"Illuminance" is power per unit area weighted by the same function; its units are
lumens per meter squared (lmlm
2
) .
Table 1.3 Luminances of astronomical sources.
Source
Night sky, cloudy, no moon
Darkest sky
Night sky, clear, no moon
Night sky, full moon
Clear sky 0.5 hr after sunset
Clear sky 0.25 hr after sunset
Cloudy sky at sunset
Gray sky at noon
Cloudy sky at noon
Moon
Average clear sky
Clear sky at noon
Solar disk
Lightning
Luminance (cd/m2)
10-4
4 X 10-4
10-
3
10-
2
0.1
1
10
10
2
10
3
2.5 x 10
3
8 X 10
3
10
4
1.6 X 10
9
8 x 10
10
For more thorough discussions of photometry, see J. T. Walsh, Photometry, Dover, New York
(\958); P. Moon, The Scientific Basis o/Illuminating Engineering, Dover, New York (1961); and
R. McCluney, Introduction to Radiometry and Photometry, Artech House, Boston (1994).
Table 1.4 Luminances of practical sources.
Source
Minimum visible level (human)
Scotopic vision valid (human)
Photopic vision valid (human)
Green electroluminescent
T8 cool white fluorescent
Acetylene burner
60 W inside frosted lamp
Candle
Sodium vapor lamp
High-pressure Hg vapor lamp
Tungsten lamp filament
Plain carbon arc crater
Cored carbon arc crater
Atomic fusion bomb
Luminance (cd/m2)
3 X 1 ~ 6
< 0.003
>3
25
10
4
10
5
1.2 X 10
5
6 X 10
5
7 X 10
5
1.5 X 10
6
8 X 10
6
1.6 x 10
8
10
9
10
12
Table 1.5 Illuminances of various sources.
Source
Absolute minimum Mv = 8
Typical minimum Mv = 6
o Mv star outside atmosphere
Venus
Full moonlight
Street lighting
Recommended reading
Workspace lighting
Lighting for surgery
References
Illuminance (Im/m2)
1.6 x 10
9
1 O ~ 8
2.54 x 10-6
1.3 x 10-
4
1
10
10
2
10
2
to 10
3
10
4
9
1. Webster's Third New International Dictionary, Unabridged, Merriam-
Webster, Inc. (1993).
2. F. Grum and R. J. Becherer, Radiometry, Vol. I in Optical Radiation
Measurements, F. Grum, Ed., Academic Press, New York (1979).
Chapter 2
Propagation of Optical
Radiation
2.1 Basic Definitions
2.1.1 Rays and angles
A ray is represented by a vector: a straight line indicating the magnitude and
direction of propagation. A wavefront is a notional surface locally normal to a
ray. Thus, a wavefront could be a plane (all rays parallel, as if from infinity) or a
curved surface (indicating diverging rays, as if emanating from a point). Figure
2.1 illustrates rays and wavefronts for both cases.
A beam can be defined by two separated elements of area as shown in Fig.
2.2. It is thus the locus of possible rays that pass through the two areas separated
by distance d. As dA 1 and dA
2
approach zero, the beam approaches a single ray.
Since there is a small, but nevertheless finite, cross-sectional area associated with
a beam, it is capable of carrying power without the flux density (power per unit
area) approaching infinity.
(a) (b)
Figure 2.1 (a) Plane and (b) diverging wavefronts. with arrows indicating the direction of
radiation propagation.
11
12 Chapter 2
d
c = : J . . . . . . . . . . . . . . . . . . . . . . . . . ~
Figure 2.2 A beam between two area elements.
The speed of light is the rate at which light propagates through a vacuum. It
is represented by c and is a constant equal to about 3.00 x 10
8
ms-
I
, a faster rate
of travel than in any other medium. The index of refraction of a medium is the
ratio of the speed of light in vacuum to the speed of light in the medium:
where
c
n=- ,
v
n = index of refraction of the medium, and
v = velocity of light in the medium.
(2. 1)
Because c must be greater than or equal to v in Eq. (2.1), a medium's
classical index of refraction is constrained to values equal to or greater than one.
Examples of nominal indices of refraction of various media are shown in Table
2.l.
Snell's law allows one to calculate the angle change in propagation direction
upon refraction:
(2.2)
Note from Fig. 2.3 that the refracted ray is closer to the normal to the
boundary between media when the index of refraction is higher; that is, when
n2 > n j. Also note that a cone of light becomes narrower in a higher-index
medium.
Table 2.1 Indices of refraction of various media.
Medium n
Vacuum 1
Air 1.0003
Water 1.33
Quartz 1.45
Glasses 1.5- 1.9
CaF2 1.42
Ab
0
3 1.75
ZnSe 2.4
Si 3.4
Ge 4.0
Propagation of Optical Radiation 13
_ .. _ .. _ .. _.
Figure 2.3 Illustration of Snell's law.
Projected area is defined as the rectilinear projection of a surface of any
shape onto a plane nonnal to the surface's unit vector. The differential fonn is
dAp = cose dA , where e is the angle between the line of observation and the local
surface normal n. Integrating over the surface area we obtain
(2.3)
Some common examples of projected area are shown in Table 2.2. Figure 2.4
depicts the relationships between surface and projected area for a circle and a
sphere.
Plane angles and solid angles are both derived units in the SI system. Figure
2.5 depicts the plane angle e, with I the arc length and r the circle radius. The
solid angle <0 extends the plane-angle concept to three dimensions. It is the ratio
of the element of spherical area dA
sph
to the square of the sphere radius ? Figure
2.6 illustrates a solid angle.
The unit of the plane angle is the radian, defined as:
The radian is the plane angle between two radii of a circle that cut off on
the circumference an arc equal in length to the radius. I
Table 2.2 Shapes and corresponding surface- and projected-area formulas.
Shape Surface area Projected area
Rectangle A =LW
Ap=LWcose
Flat circle A = 1t? = [1t(f14)] Ap = 1t? cose = [1t(f14)] cose
Sphere A = 41t? = 1tf A =A14 = 1tr2
p
14 Chapter 2
n

-----------
---------------
(a)
(1
.
... ... 1 ... ... --------------

' ", --------------
(b)
Figure 2.4 Surface- and projected-area relationships for a (a) circle and (b) sphere.
e = fir
Figure 2.5 Planar-angle relationships.
dA
sph
ro = dA
sPh
7
Figure 2.6 Solid-angle relationships.
Propagation of Optical Radiation
d<l>
:::=.:y---
---
e
rsine
rsined<l>
rde
dA
sph
rsined<l>
Figure 2.7 An element of solid angle in spherical coordinates.
2
15
In other words, I = r defines one radian. Since there are 2n radians in a circle, the
conversion between degrees and radians is I rad = (360/2n) = (180/n) degrees.
The steradian (sr) is defined in an analogous manner:
The steradian is the solid angle that, having its vertex in the center of a
sphere, cuts off an area of the surface of the sphere equal to that of a
square with sides of length equal to the radius of the sphere. I
Dividing the entire surface area of a sphere by the square of its radius, we
find that there are 4n sr of solid angle in a sphere, and 2n sr in a hemisphere.
Figure 2.7 depicts the solid angle in terms of the planar angle e and the
rotational (azimuthal) angle <1>, where r is sphere radius. The small element of
area dA
sph
lies on the surface of the sphere. The element of solid angle subtended
by dA
sph
is expressed as:
dAs h
doo=-f- = sineded<l>.
r
(2.4)
To determine the value of the solid angle, integrate over e and <1>:
00= ffsineded<l>.
(2.5)
 8
16 Chapter 2
Figure 2.8 Right circular cone, with 8
112
the cone half angle.
In the most general case, a solid angle can subtend a surface of any shape.
However, in optical systems, which typically have circular apertures, the
envelope of the solid angle is a right circular cone, as shown in Fig. 2.8.
In Fig. 2.8, 81/2 is the cone half angle. It is the plane angle between the
centerline of the cone and anywhere on the edge of the cone. It is related to the
solid angle of the cone (J) by:
21t 0112
(J) = f d<j> f sin SdS ,
(2.6)
o 0
which results in
(J) = 21t(l- cos 8
1/ 2
) (2.7)
Application of a simple trigonometric identity provides the equivalent
expreSSIOn
8
(J) = 41tsin
2
-ill..
2
(2.8)
A feel for the magnitudes of various solid angles can be obtained by inspecting
Table 2.3. Table 2.4 facilitates conversion from steradians to other units.
Table 2.3 Some objects and corresponding solid angles.
Object
Dime at 1 km
Jupiter (mean)
Cone in human eye
Sun, moon at earth's surface
Tennis ball at I m
Plane
Sphere (from the inside)
Linear subtense
0.065 arcsec
0.65 arcsec
1 arc min
32 arcmin = 0.53 deg
3.7 deg
Solid angle (J)
3.1 X 10-
13
sr
3.1 x 10-
11
sr
2.7 x 10-
7
Sf
6.8 X 10-
5
sr
0.0033 sr
21t
41t
Table 2.4 Steradian conversions.
1 sr = 1 rad
2
1 sr = 3282.8 deg
2
1 sr = 1.1818 X 10
7
arcmin
2
1 sr = 4.2545 X 1010 arcsec
2
17
Both plane angles and solid angles are dimensionless quantities, and their use
can lead to confusion when attempting dimensional analysis. For example, the
simple inverse square law of irradiance (to be discussed in detail in Sec. 2.3.1),
E = J/d
2
, appears dimensionally inconsistent. The left side has units W/m2, while
the right side has W/m
2
sr. It has been suggested that this equation be written E =
I o.)d
2
, where 0.
0
is the unit solid angle, 1 sr. Inclusion of the term no will render
the equation dimensionally correct, but no will far too often be considered a free
variable rather than a constant equal to 1, which leads to erroneous results.
Current practice suggests that another type of solid angle, the projected (or
weighted) solid angle, is more useful. The symbol for a projected solid angle is
0., and the units are also steradians. It is defined as the solid angle 00 projected
onto the plane of the observer, as shown by the defining equation:
dO. = doocos8 = sin8cos8d8d<l>. (2.9)
It involves an additional cosine term. This is depicted graphically in Fig. 2.9.
dO)
---
---.............
'\
dO
Figure 2.9 Projected solid-angle relationships. [Adapted from Radiometric Calibration:
Theory and Methods, C. Wyatt, p. 21 (1978)].
18
To determine the projected solid angle, integrate over e and <1>:
Q= ffsinecoseded<1>.
q, e
Chapter 2
(2.10)
As before, the example most relevant to optical systems is the right circular
cone depicted in Fig. 2.8, for which the integral may be expressed as
n = f" d<1> 1'12 sin e cos e de, (2.11 )
resulting in
(2.l2)
where 8 1/2 is the cone half angle, as before. The plane area subtended by the cone
is Qr2, and the spherical surface area subtended is ror2. A couple of special cases
are worthy of mention. For a hemisphere, Q = 1t sr, while an entire sphere
subtends 21t sr of a projected solid angle.
In addition to the dimensional concern raised above, there is another good
reason to employ two definitions of a solid angle. For many radiometric
problems, the emitter or receiver is flat, and the projected solid angle n is the
proper choice as it requires the inclusion of the cose term relating to the
projected area of the surface. In other cases, the emitter or receiver approximates
a point, emitting uniformly in all directions or responding to incoming radiation
equally from all directions. The solid angle ro is appropriate under these
conditions.
For a right circular cone at a half angle 81/2 of 90 deg, the projected solid
angle is 1t according to Eq. (2.12), and ro is 21t according to Eq. (2.7). At the
other extreme, when 8 1/2 is equal to 0 deg, both ro and n are zero. For small
angles, the solid angle and projected solid angle differ only by a cosine, are
nearly identical, and are therefore interchangeable in numeric value, if not in
concept. The error incurred by using ro rather than n is given in Table 2.5 (note
that ro> Q for angles greater than 0 deg).
The definitions and symbols presented here have not been universally
applied in the past. One must be very cautious when reading the literature, as
different investigators use the terms and symbols solid angle ro and projected
solid angle Q interchangeably, incurring predictable confusion and potentially
incorrect results.
Table 2.5 Percentage error when not using a projected solid angle.
@1/2 (in deg)
10
16
25
35
48
2.1.2 System parameters
Error using 0) rather than n
< 1%
< 2%
< 5%
< 10%
<20%
19
So far, we have defined only the terms necessary to describe the angles important
to a study of propagation. Now, let's define quantities specific to optical systems.
The f-number (f/#) of an optical element or system is most often seen as the
ratio of the effective focal length f (object at infinity) of the optical element or
system to the diameter D of the entrance pupil:'
(2.12)
Good optical systems fulfill the Abbe sine condition with a spherical
wavefront converging to the focal point, and the preferred definition ofjl# is
f/#= .1
2sme
l / 2
(2.13)
The numerical aperture (NA) of a system is defined as the sine of the vertex
angle (half angle) of the largest cone of meridional rays that can enter or leave an
optical system or element, multiplied by the refractive index of the medium in
which the vertex of the cone is located. It is generally measured with respect to
an object or image point, and will vary as that point is moved. The defining
equation is
NA = nsine
l / 2

Bothjl# and numerical aperture can be related to projected solid angle:
For an informative discussion ofP#, see D. Goodman, "The f-word in optics," Optics &
Photonics News, p. 38, April (1993).
(2.l4)
20
Q = 1t
4(J/#)2'
and
Q = 1t(NA)2
2 '
n
and may be related to one another by
J/#=_n_.
2NA
Chapter 2
(2.15)
(2.16)
(2.17)
Throughput, also known as geometrical extent or etendue, is the product of
the cross-sectional area of a beam and its projected solid angle
T=AQ. (2.18)
Figure 2.10 shows the interaction between two differential area elements dA I
and dA
2
a distance d apart. The area dAI subtends a solid angle (OJ, having apex at
dA
2
, while dA
2
subtends a solid angle ffi2, whose apex is at dA,. The surface
normals are shown in the figure by boldface n.
The throughput relationships become
dAI cose1dA
2
cose
2
= d(O
l
dA
2
cos 8
2
= -..:..---"----"----=-
d
2
dA
z
cos 8
z
dA
I
cos 8
1
= dffi
2
dA
1
cos 8
1
=
d
2
By inspection, 7;--->2 = T2--->1 = T. The throughput is invariant.
W2
............... .
................................
.....................................
d
(2.19)
Figure 2.10 Area- and solid-angle relationships used to define throughput. (Adapted from
a figure courtesy of William L. Wolfe.)
F
Q
05
A
5
L
..................................
A
o
Figure 2.11 Invariance of throughput for a case in which the source image fills the
detector.
So far, our discussion has focused on theoretical constructs. Now, let's
introduce some system elements as we proceed with demonstrating the
invariance of throughput in an optical system.
Consider Fig. 2.11. It represents a special case in which the image of the
source exactly fills the detector. This configuration consists of a source, a lens,
and a detector. In the figure, As is the area of the source, Ao the area of the optics,
and Ad the area of the detector. The projected solid angles are defined thus: Qos is
the angle the optics subtend at the source, Q
so
the angle the source subtends at the
optics, Q
do
the angle the detector subtends at the optics, and Q
od
the angle the
optics subtend at the detector. For this case, the following equalities hold:
(2.20)
Since these pairings are equal, any of the above pairs can be chosen for
calculation purposes. If the image of the source does not exactly fill the detector
area, care must be taken to determine the proper AQ product to use. The author's
(Palmer) personal preference is the most often-used pair AoQ
do
, inasmuch as the
entrance aperture size Ao and the field of view of the system Q
do
are determinable
characteristics of a radiometer. The next most often used pair is AdQod, as the
detector size and the jl# of the optics are also measurable characteristics of the
radiometer.
Basic throughput is the name given to the quantity conserved across a
lossless boundary between two media having different indices of refraction. It
can be written as
(2.21)
and the relationship between medium 1 and medium 2 is
22 Chapter 2
(2.22)
where the subscripts denote the respective quantities in media and 2. Like
throughput, basic throughput is invariant. Most optical systems have both the
object and the image located in the same medium, typically air, so basic
throughput is not often used.
Figure 2.12 depicts the appropriate and inappropriate area- and solid-angle
pairings used to define throughput. The correct area-solid angle pair is shown in
Fig. 2.12(a), the incorrect angle pair is shown in Fig. 2.12(b). Because the
definition of throughput includes two (projected) areas and the distance between
them, a correct pairing has the apex of its solid angle located at the correct area.
The incorrect pairing uses one area twice and ignores the other. The maxim "no
ice cream cones" should be applied.
Let's look at some examples of the An product. First, a spectrometer: Fig.
2.13 shows the area- and solid-angle product of the entrance slit (typically 1 x 10
mm) and the projected solid angle the collimating lens subtends at the slit, n
r
,.
The An product of a spectrometer is usually very small, and narrow spectral
bandwidths are typically employed. Therefore, it is difficult to get much light
through a spectrometer.
A different example of throughput may be found in the common camera. It is
related to the j/# of the lens and the size of the film. In this case, the detector
(film) size is not the overall dimension of the image, but the size of an individual
film grain. The smaller the j/#, the "faster" the camera and the greater the
throughput. Similarly, "fast" film has a larger grain area, permitting a higher
throughput and a shorter exposure time than "slow" film with a smaller grain
area.
(a) (b)
Figure 2.12 Right and wrong area-solid angle combinations for throughput determination.
---
Slit,
Area A
23
Lens
Figure 2.13 Example of the correct area (A) and solid angle (n/
s
) product used to
determine throughput in a spectrometer. (Adapted from a figure courtesy of William L.
Wolfe.)
2.1.3 Optical definitions
Some optical quantities are relevant to a study of radiometry, and they are
defined here. For more detailed treatment, the reader is advised to consult a text
on geometrical optics. t
Figure 2.14 depicts the location of object and image planes, along with some
key rays. In Fig. 2.15, the chief ray in an optical system originates at the edge of
the object and passes through the center of the entrance pupil (NP). It passes
through the center of the aperture stop (AS), the edge of the field stop (FS), and
the center of the exit pupil (XP), to the edge of the image, defining the image size
(height) and the lateral (transverse) magnification. There may be several
intennediate pupil planes in a complex optical system.
Image
Figure 2.14 Chief and marginal rays in an optical system, shown schematically.
t Several excellent texts exist; a recent one is E. L. Dereniak and T. D. Dereniak, Geometrical and
Trigonometric Optics, Cambridge University Press, Cambridge (2008).
24 Chapter 2
CHIEF RAY
..........................
. -
Figure 2.15 Optical system stops and pupils.
The marginal (rim) ray in an optical system is the ray from the object that
originates at the intersection of the object and the optical axis and passes through
the edge of the entrance pupil. It touches the edge (rim) of the aperture stop,
proceeds through the center of the field stop, and the edge of the exit pupil. The
marginal ray intersects the optical axis at the center of the image, defining the
location of the image and the longitudinal magnification. There may be several
intermediate image planes in a complex optical system.
The entrance pupil is the image of the aperture stop in object space (as seen
from the object), while the exit pupil is the image of the aperture stop in image
space (as seen from the image).
Of particular importance to radiometry are the aperture and field stops. The
aperture stop determines how much light may enter the system, while the field
stop determines the system's angular field of view. In a simple system consisting
of a lens and detector at the rear focal point, the lens is the aperture stop, and
both the entrance and exit pupils are located at the lens, with the same size as the
lens. In more complicated systems, the stops may be internal and separated from
the pupils, as shown in Fig. 2.15.
2.2 Fundamental Radiometric Quantities
2.2.1 Radiance
The study of radiometry begins with fundamental units. Radiant energy has the
symbol Q and has as its unit the joule (J). Radiant power, also known as radiant
flux, is energy per unit time (dQldt), has the symbol <1>, and is measured in watts
(W). These definitions give no indication of the spatial distribution of power in
terms of area or direction. Radiance is the elemental quantity of radiometry,
power per unit area, and per unit projected solid angle. It is a directional quantity;
it can corne from many points on a surface that is either real or virtual; and
because it is a field quantity, it can exist anywhere. The symbol for radiance is L
and the units are W Im
2
sr.
The defining equations are
(2.23)
-
where e is the angle between the normal to the source element and the direction
of observation as shown in Fig. 2.16.
Radiance is also associated with a source, either active (thermal or
luminescent) or passive (reflective), as discussed further in Chapter 4. Because
radiance may be evaluated at any point along a beam, it is associated with
specific locations within an optical system, including image planes and pupils.
Other radiometric units may be derived from radiance by integrating over
area andlor solid angle. Integration over solid angle yields irradiance (arriving at
a location, such as a sensor) or radiant exitance (leaving a location, such as a
source), both of which are expressed in W/m
2
Integration over area yields
radiant intensity expressed in W Isr. Integration over both area and solid angle
yields radiant power in watts.
If rays are traced across a lossless boundary between two materials having
different indices of refraction as shown in Fig. 2.3, the solid angle changes
according to Snell's law. Taking this change into account, the quantity Ln-
2
is
seen to be invariant across the boundary. This quantity is called basic radiance. It
is useful for calculations when an object and its image are located in spaces with
different indices of refraction.
In the absence of sources or sinks along the path of a beam, power along a
beam is conserved. Since it was previously demonstrated that throughput is
conserved in an optical system, the radiance must also be invariant in order for
conservation of power (energy per unit time) to be obeyed. The results of this
invariance of radiance are significant:
(1) The radiance of the image at the detector plane of a camera (film or array
device) is the same as the radiance of the scene if there are no
transmission losses due to atmosphere and optics; and
(2) The radiance at the focal plane of a radiometer (imaging or point) is the
same as that of the target, if there are no transmission losses due to
atmosphere and optics.
Note that the transmission of atmosphere and optics is not likely to be unity
(perfect transmission); however, results (1) and (2) greatly simplify radiometric
calculations.
dW\
d
---
---
---
---
---
---
--------------------f-
n
Figure 2.16 Radiance from area element dA, tilted at angle e from surface normal, n.
26 Chapter 2
The defining equation for radiance can be inverted and integrated over area
(and in the most general case, a projected solid angle) to determine the power in
an optical system:
 = JJL dAsdro
s
cos8 = JJLdA
s
dO..
(2.24)
2.2.2 Radiant exitance
Radiant exitance is radiation that exits a source. It is defined as power per unit
area radiated into a hemisphere (dl dAs). The symbol for radiant exitance is M
and the units are W/m
2
Its defining equation is
M = lim (L\ .
M , -+O M s dAs
(2.25)
The radiant exitance of a source is obtained by integrating radiance over the
projected solid angle of a hemisphere:
(2.26)
Integrating radiant exitance in W 1m
2
over area results in radiant power in
watts:
=fMdA.
JA s
(2.27)
Figure 2.17 further illustrates the concept of radiant exitance. The relationship
between radiant exitance and radiance is complex, depending on the angular
distribution of radiance L(8,<1 . This is illustrated in Fig. 2.18 and described
mathematically in the following equations:
2 L(8,<1dA] c o s d ~
d= 2 '
r
(2.28)
where r is the radius of the hemisphere, as shown in Fig. 2.18, dA] is an element
of the emitting area, dA
2
an element of the receiving area, and 8 is the angle
between the surface normal of dA] and the direction of propagation.
Making a substitution for dA
2
using polar coordinates,
d ~ = r2 sin 8d8d<1>
results in
d
2
 = L(8, <1> )dA] sin 8cos 8d8d<1> . (2.29)
..
Figure 2.17 Radiant exitance from a source.
Applying limits for a hemisphere and constructing an integral results in
f21t f" / 2
=.b d<l>.b L(8, <IA1sin8cos8d8 .
(2.30)
Assuming that L is independent of direction (that is, the source at A I is
Lambertian), then
(
2 J"/2
 = 2rrLAl i ~ 8 0 (2.31 )
Substituting the equality M = <l>IA and solving, we obtain
M=rrL. (2.32)
The error most often committed in radiometry is in this conversion. Don' t
make the frequent mistake of forgetting the sine and using 2n!
Figure 2.18 Relationships between emitting and receiving areas used in deriving the
relationship between radiance and radiant exitance.
28 Chapter 2
2.2.3 Irradiance
Irradiance (radiant incidence) is the opposite of radiant exitance, in that it is the
power per unit area that is incident on a surface. Its symbol is E and its units are
again W/m
2
:
E = lim = d .
Ms dA,
(2.33)
lrradiance on a surface is obtained by integrating radiance over the projected
solid angle of a hemisphere:
(2.34)
Integrating irradiance over area results in incident power
 = fEdA. (2.35)
A
This is shown conceptually in Fig. 2.19.
The most common example of irradiance is the solar "constant." It is the
irradiance from the sun incident upon the earth's atmosphere at the mean earth-
sun distance. Its numeric value is nominally Eo = 1368 W/m2, but it varies
slightly over the years, primarily due to sunspots. Also, the number 1368 is not
exact, and other values are often quoted, usually within 0.1 % or so. The solar
irradiance at the earth's surface, on the other hand, varies widely due to
atmospheric effects; a nominal value of 1000 W 1m
2
is often used.
Figure 2.19lrradiance upon a surface.
pc
2.2.4 Radiant intensity
Radiant intensity is power emitted per unit solid angle in a specific direction.
Radiant intensity is frequently used when describing the radiation of an isotropic
(same in all directions) point source. Its symbol is I and its units are W/sr. Figure
2.20 conceptually illustrates the radiant intensity of a point source and a surface
illuminated by it. Equation (2.36) defines radiant intensity mathematically:
I = lim ( d<l> ) = d<l> .
dCO,-+O dOO
s
doo
s
(2.36)
Intensity is derived from radiance by integrating over area
(2.37)
and radiant power can be derived from intensity by integrating over solid angle
00. (Note that the definition of intensity does not include an area term; therefore,
the integration is performed over solid angle, without a cosine projection.)
(2.38)
In the case of the isotropic point source, the radiant power is 41t times the
intensity. Conversely, intensity may be found by dividing the total emitted power
 by 41t.
A word of warning: intensity is the most problematic radiometric quantity,
because it means different things to different people in different but related
fields. Laser physicists are prone to use it as the square of the electric field
strength, with units W/m
2
Atmospheric scientists and heat transfer engineers use
the term "specific intensity," frequently omitting the "specific," to mean
W/m
2
sr. Some scientists and engineers equate intensity simply to power, and a
few use it to describe spectroscopic line strengths. Which usage is correct?
~ . . ~
Figure 2.20 Radiant intensity from a point source.
30 Chapter 2
4A
A V
i ..................................... .
, ........................... ? ..... .
d ~ 4 d ~ I
Figure 2.21 Illustration of the inverse square law.
In answering the question, we might invoke precedent. Atmospheric scientists'
usage goes back to Chandrasekhar's Radiative Transfer (l950t The physics
community usage goes back as far as Drude' s Lehrbuch der Optik (1900).
5
Usage
in photometry and radiometry dates back to at least 1760 (Bouguer). But such
citing leads only to pointless argument. There is one compelling reason to
standardize our usage: intensity is an SI base quantity. Any other use of the term
intensity is incompatible with the SI and should be deprecated and abandoned.
6
,7,8
2.3 Radiometric Approximations
2.3.1 Inverse square law
There are a number of approximations in radiometry, most of them having to do
with radiation geometry. The inverse square law ofirradiance is perhaps the best
known approximation. It states that the irradiance from an isotropic point source
varies inversely with the square of the distance from the source.
As shown Fig. 2.21, the rays are straight lines diverging from a source
having intensity 1. At distance d, they fill area A. At a distance 2d, the length of
each side has doubled and the area subtended has increased by a factor of four.
Assuming a lossless medium, the amount of radiant power at the second surface
is the same as that at the first surface, due to the conservation of power. At the
second surface, power is spread over a larger area, and irradiance decreases. The
relationship between irradiance and distance is given by
(2.39)
The relationship expressed by Eq. (2.39) presumes that the area shown in
Fig. 2.21 lies normal to the optical axis. If this is not the case and the surface is
tilted at an angle e to the optical axis, a cosine e factor must be included in the
equation:
(2.40)
This situation is illustrated in Fig. 2.22.
p
I
e-._._._._._._._._._._._._.-
d 8
n
Figure 2.22 Area A tilted at angle 8 from the axis.
Inverse square laws pervade all of physics, beginning with Newton's
universal law of gravitation. Our radiometric application works well for small
sources observed at great distances. But how well maya real source approximate
an isotropic point source?
To avoid the question entirely, one might measure the source only in a well-
defined, specified direction. This approach will allow application of the inverse
square law, but results may not be repeatable if measurements are made in a
different direction. (In photometric terminology, source output under this
condition is termed "directional candlepower.")
Another approach involves use of a "small" source rather than a point
source-once we determine what "small" is. If the maximum dimension of the
source is less than dllO, i.e. , one-tenth of the distance from the source (equivalent
to about a 3-deg half angle), the inverse square law may be applied with an error
of less than 1 %. If this is not good enough, a stricter criterion may be applied. If
the maximum source dimension is less than dl20 (1.5-deg half angle), the error in
applying the inverse square law is less than 0.1 %.
2.3.2 Cosine
3
law
The cosine
3
law is applied to calculate the irradiance on a plane surface from an
isotropic point source, for example, the distribution of irradiance over a floor
from a bare light bulb suspended above. It is often convenient to describe the law
in terms of 8 rather than distances along the surface, as the expression is
somewhat simpler.
Referring to Fig. 2.23, the irradiance at a point x directly below the source
with intensity I may be calculated according to the inverse square law as
I
EX=-2 .
D
At position y, the distance from the source has increased such that d=Dlcos8.
If the target at y is perpendicular to the vector between source and target, the
irradiance at y is:
32 Chapter 2
y X
Figure 2.23 The cosine
3
law.
If the target is now rotated so that it is parallel to surface x-y (the floor), an
additional cos8 term is introduced, reSUlting in
or more generally,
(2.41 )
which is the cosine
3
law. Note that two cosines were introduced to account for
the increased distance from the source; the third accrued from the projected area
of the target.
2.3.3 Lambertian approximation
Before discussing the final important approximation, another look at Lambertian
radiation is needed. A Lambertian source is one whose radiance L is independent
of direction:
L (8,<1 = constant.
Thus, in the equation
d
2
cI> = L(8,<1dAcos8dQ,
the power received from a Lambertian radiator is proportional only to cos8, the
angle of observation, and by extension, the projected area of the source.
....
The Lambertian approximation is frequently used to describe the angular
distribution of radiation from a source. It is powerful because it significantly
reduces the complexity of the mathematical solutions to radiative transfer
problems, allowing them to be simplified so that calculation focuses only on
radiation geometry.
Fortunately, the Lambertian approximation is surprisingly good. Most
conventional blackbody radiation simulators (often used as infrared sources) and
integrating spheres (often used as visible sources) are very nearly Lambertian
over a wide range of angles. Reflective sources, also called "flat" or matte
reflectors (flat paint, matte white paper, etc.) are quite Lambertian out to nearly
grazing angles. Most natural surfaces, with the exception of still water, are highly
Lambertian. Surprisingly, polished metal surfaces, which are highly polarized in
emission at angles away from normal, are Lambertian out to 50 deg or so when
the total radiation is taken into account.
By contrast, specular reflecting surfaces, such as glossy paints, mirrors, still
water, etc., do not lend themselves to the Lambertian approximation.
It is common practice to utilize the Lambertian approximation, but make sure
to test the assumption with measurement whenever possible, and to keep track of
its effect on radiometric calculations.
2.3.4 Cosine
4
law
The cosine
4
law of irradiance is similar to the cosine
3
law, except that a small
Lambertian source with radiance L replaces the isotropic point source. Such a
source might be the combination of a single lamp with a flat diffuser panel
oriented parallel to the ground, as shown in Fig. 2.24.
Since the basic geometrical layout is the same as that in the previous
example, start with the cosine
3
law:
E = I cos
3
e
D2
y x
Figure 2.24 The cosine
4
law.
34 Chapter 2
Recall that the projection of the target onto the floor added the third cose
term; the projection of the source area in the direction of point y adds the fourth.
For a Lambertian source, the relationship between intensity (point source)
and radiance ( extended source) is I = LA
proj
. Substituting,
which is the cosine
4
law.
LA cos
4
e
E=--'-P---
D2
(2.42)
This fourth-power reduction in irradiance (falloff) applies to many, but not
all, sources that are extended and Lambertian. The effect is a noticeable dimming
of brightness near the image edges, often seen in images from projection systems
and wide-angle cameras. The effects of cosine
3
and cosine
4
falloff are shown
numerically in Table 2.6 for a variety of angles.
There are some cases in which a source may be "too extended" for the
cosine
4
law to correctly predict irradiance. Consider the configuration shown in
Fig. 2.25, in which a circular source (shown edge on) radiates downward to the
observation plane. The angle e is the angle of incidence to the surface, and angle
01/2 is the half angle of the system viewing the source, related to the system's jl#
by Eq.(2.13). Values of01J2 for a variety ofcommonjl#s are shown in Table 2.7.
Table 2.6 Fractional irradiance at observation point as a function of cosine factor.
Cosine
Angle 9 (degrees)
factor
0 10 20 30 40 50 60 70 80
cos e 0.98 0.94 0.87 0.77 0.64 0.50 0.34 0.17
cos
3
e 0.96 0.83 0.65 0.45 0.27 0.13 0.040 0.005
cos
4
e 1 0.94 0.78 0.56 0.34 0.17 0.063 0.014 0.001
Circular source viewed edge on
8
112
Figure 2.25 Foote's formula geometry.
....
Table 2.7 System half angles as a function of f/#.
fl# <Il1l2 (deg.)
16 1.8
8 3.6
4 7.2
2.8 10.3
2 14.5
1.4 20.9
1 30
An expression for the irradiance at the sensor for this particular geometry has
been derived by Foote (1915).9,[0 Foote's formula is shown below, and the
corresponding plot using j!# as a parameter is shown in Fig. 2.26. For this
particular geometry, irradiance is least likely to follow the cosine
4
law in lower
j!# systems. Note that at normal incidence (9 = 0), the equation reduces to
E = nLsin
2
9.
w
(J
z

0.8

g;:

f= 0.4
:5
w
0::
0.2

o 10 20 30
ANGLE (degrees)
40 50
Figure 2.26 Foote's formula results compared to Lambertian.
60
36 Chapter 2
2.4 Equation of Radiative Transfer
The equation of radiative transfer is the most important equation in radiometry.
All axioms discussed in the previous section depend on it. It is, in differential
form:
(2.43)
The physical situation described by the equation is shown in Fig. 2.27, where
dA, and dA
2
are the differential area elements, d is the distance between them,
and 8, and 8
2
are the angles between the normals to the area elements and the
optical axis.
As the equation indicates, the radiant power received at surface 2 depends
directly on the radiance L(8,$), the area elements, the angles between area
normals and the optical axis, and inversely on the square of the distance between
the surfaces. If the distance squared is much larger than the largest area, i.e.,
d
2
(A, or A
2
), the differential areas may be replaced with the actual areas, and
Eq. (2.44) can be applied to calculate power:
(2.44)
Because d
2
is much greater than the size of any area element, the variations
in 8 and $ are small, and the term L(8,$) need be evaluated at only one particular
set of angles. If either area is appreciable in relation to d
2
, Eq. (2.44) cannot be
used.
The integral form of the equation of radiative transfer looks much more
formidable:
(2.45)
Figure 2.27 Radiation transfer geometry.
To utilize this integral form, the following factors must be taken into
account:
(1) The angles 8
1
and 8
2
may vary from one part of area A 1 or area A2 to
another.
(2) The distance d may also vary from one part of area A 1 or area A2 to
another.
(3) The angular variation in radiance L(8,<j)) may be significant.
An additional implied assumption is that the principle of superposition is
applicable when adding up the small elemental contributions to radiant power; in
other words, the source providing the radiation is incoherent and interference
effects do not occur in the beam.
If the source is Lambertian, the radiance is independent of 8 and <j), and the
equation of transfer becomes much simpler. It is
 =Lffcos8IcOS82dAdil.
1-2 d2 1 H 2
A, A,
(2.46)
In this case, the transfer equation can be thought of as the product of a
radiance term L and the geometric term expressed by the double integral.
The transfer equation can be simplified even further if certain assumptions
can be made regarding the radiation geometry: first, that the square of the
distance d
2
is much larger than either area A 1 or A
2
, and second, that both areas
are on axis, 8
1
and 8
2
are zero, and their cosines are therefore unity. If these
assumptions can be made in conjunction with the Lambertian approximation, the
result is the ultimate simplification:
(2.47)
A corollary expression to calculate flux density, power-per-unit area in a
system, is
E=LQ. (2.48)
These simple equations are the logical starting points for all radiometric
engineering calculations, as tbey provide first-cut, back-of-the-envelope answers.
In many instances, they are all we need. In order to fully understand a particular
application, assumptions must be tested and the errors incurred by their use
assessed.
The Lambertian approximation is relatively simple to verify if we possess
hard data about the directionality of the source. The on-axis assumption is also
easy to verify. As discussed in Sec. 2.3, the distance between the two area
elements must be at least lOx the maximum linear dimension of the largest
38 Chapter 2
element in order for the inverse square law to be good to 1 %. If the distance
between the two is increased to 20x the maximum dimension, the uncertainty in
applying the law reduces to 0.1 % or less.
2.5 Configuration Factors
The concepts of radiative transfer are used not only in optics, but also in related
fields, such as thermal and illumination engineering. Equation (2.46) shows the
separation of the equation of radiative transfer into a radiance term and a
geometry term, assuming the Lambertian approximation may be applied. Other
names for the geometry term are view factor, interchange factor, shape factor,
form factor, and the term we will utilize, configuration factor. The symbol F is
used to denote this quantity, defined by Eq. (2.49):
F = <1> 1->2
 '
1
(2.49)
where <1>1 is the power leaving surface 1 and is the power reaching surface
1---> 2
2 from surface 1. Both power terms are dimensionless. The radiant power terms
are further defined as
where MI is the radiant exitance in W 1m2 leaving surface 1, and A 1 its area; and
 = MI flcOS 9
1
cos 9
2
dA dA
1--->2 1t d2 1 2 ,
where the radiance term outside the integral is obtained from Eq. (2.32), itself
dependent upon the Lambertian approximation. The fraction of radiant power
leaving surface 1 that arrives at surface 2 is
F. = _1_ ficas 9
1
cos 9
2
dA dA
12 1tA d2 1 2'
1
which is the configuration factor. The power transferred from surface 1 to surface
2 then becomes
<1>1 --->2 =

=
pu
Because the radiant power incident on surface 2 equals LIA 10.
21
by Eq.
(2.47),
and
To restate: in engineering calculations, using the assumptions given above,
(2.50)
and
(2.51)
The advantage of using configuration factors is that numerous solved
geometries appear throughout the literature. Some relevant to optics are shown in
Fig. 2.28. II Further information on configuration factors may be found in Refs.
12, 13, and 14.
F = 1
d1 - 2 (hi r)2 + 1
( )
2
r
F --
d1- 2 - h
h
1
(a)
(b)
Figure 2.28 Configuration factor examples: (a) Planar element parallel to circular disk,
and (b) planar element to sphere. (Adapted from Ref. 11.)
40 Chapter 2
2.6 Effect of Lenses on Power Transfer
Radiometer configurations will be discussed in detail in Chapter 6, but the effects
of lenses on power transferred to a detector will be introduced here. Look first at
Fig. 2.29. Two configurations are shown: Al and A2 are the areas of the stops; D
is the distance from source to stop at A I; and S is the distance from A I to A2 (at
the detector d). The difference between the two configurations is solely the
presence of the lens at A I in the second. Expressions for power at the detector in
each case, for both point and extended sources, will be formulated. In both cases,
it will be assumed that the transmittance of the atmosphere between the source
and the detector is unity.
For the point source case, the irradiance at the detector in Fig. 2.29 is
expressed by the inverse square law, Eq. (2.39):
where d is the distance between source and detector. Without a lens, [Fig.
2.29(a)], assuming no transmission losses in the intervening medium, the power
at the detector is simply the irradiance multiplied by the available sensitive area:
(2.52)
Expressed in terms of intensity, the power is
(2.53)
where (D + S) is the source-detector distance. Note that in this configuration, A2
acts as the aperture stop, defining how much flux is collected, while AI , the field
stop, defines the detector's field of view.
__ ___ A 2 _ ~
o I ~ 5 .1
(a)
(b)
Figure 2.29 Configurations (a) without and (b) with a lens.
....
Adding a lens, as in Fig. 2.29(b), yields a different set of equations. In this
case, the power at the detector may be expressed as
(2.54)
or, more specifically, as
(2.55)
where 'tl
ens
is the lens transmission.
In terms of intensity, it is:
(2.56)
Note that in this case, the aperture that limits the flux into the system is A},
the aperture stop, and that A2 is the field stop, limiting the detector's field of
view. The size of A2 is unimportant as long as it does not vignette the source's
image at the detector.
The difference in received power between the two expressions in terms of
irradiance is expressed as
G = 'ttensAJ
~
(2.57)
where G is the gain of power on the detector. To maximize G, make the lens
transmission and the area ratio as large as possible, while not vignetting the
source Image.
To determine the effect of a lens on the same instrument configuration with
an extended source, begin with Eq. (2.47):
At area A = A
2
, the appropriate solid angle is subtended by area A I. This solid
angle is expressed as in Eq. (2.12) by
where 81/2 is the cone half angle. Assuming that it is small, the approximation
AIls'- may be used for n
12
, so that
(2.58)
42 Chapter 2
and
(2.59)
The same approach can be pursued with the other area-solid angle
combination, that is, with A I and 0
21
In that case, the solid angle is
approximately A2/S, and dis again obtained by Eq. (2.59). Inserting a lens at Al
limits the power by the transmission of the lens, so that
(2.60)
The radiance-area-solid angle relationship holds true regardless of whether
the first or second area-solid angle combination is used to calculate throughput.
Inserting a lens yields no net gain in detector power for an extended source.
Rather, the power is less due to the nonunity (in the real world) transmission of
the lens.
2.7 Common Radiative Transfer Configurations
2.7.1 On-axis radiation from a circular Lambertian disc
This case is shown in Fig. 2.30. Assuming a lossless optical system, the flux
transferred from source to detector is given by Eq. (2.47), where L is the
Lambertian disc radiance. The area-solid angle pair we will use in this case is the
area of the detector Ad (in the figure) and the solid angle the source subtends at
the detector Osd, which may also be expressed by Eqs. (2.12) and (2.15) as
no . 28 1t
:''',d = 1tsm 112 = / 2'
. 4(f #)
(2.61)
wherefl# is defined in Eq. (2.13). Considering the geometry in the figure,
2
. 28 a
SIn 1/ 2 = 2 2'
(a +b )
We can now substitute in Eq. (2.48) to provide several equivalent
expressions for the irradiance at the detector:
Ed = nLsin
2
8
1/ 2
= LO d = nL 2a
2
2 = nL = nL(NA)2
S (a +b) 4(f/ #)2 .
(2.62)
-
Figure 2.30 On-axis Lambertian disc, irradiance measured at detector of area Ad.
If the distance b is far greater than 2a, the linear dimension of the source,
then the inverse square law holds and Ed may be approximated as rrLa
2
b-
2
The
error incurred using a
2
b-
2
rather than a
2
(a
2
+ b
2
r l is less than 1% when the
diameter-to-distance ratio is less than 0.1. Under these conditions, source
intensity Is may be substituted for TtLa
2
(the radiance times the area of the source)
so that
Table 2.8 summarizes the relationships between source-detector distance and
irradiance at the detector for a variety of cases.
Table 2.8 Irradiance at detector as a function of source distance for a Lambertian disc.
Distance
b 2a
b =2a
b=a
b=O
Half-angle 9
112
(deg)
very small
26.5
45
90
Irradiance Ed
TtLa
2
b -2= Ib -2
rrL/5
TtL/2
rrL
To determine the irradiance on the detector from an annulus (ring) rather
than a disc, calculate irradiances from discs having both outer and inner radii,
and subtract the latter from the former.
2.7.2 On-axis radiation from a non-Lambertian disc
In this case, source radiance is not independent of observation direction, and an
integration must be performed. The source's radiant exitance may be obtained by
44 Chapter 2
Figure 2.31 On-axis Lambertian sphere, irradiance measured from surface.
integrating Eq. (2.26) over a hemisphere, and the irradiance at the detector
calculated as a function of the half angle. To illustrate, take the relatively simple
example in which Ls = LacosS. In this case, the radiant exitance is
where dO. is taken from Eq. (2.9). The resulting integral is then
which results in
M = 2rrLa .
s 3
Assuming a lossless medium, the irradiance at the detector is
(2.63)
(2.64)
(2.65)
(2.66)
In general, closed-form solutions are not readily available, and numerical
methods must be employed.
2.7.3 On-axis radiation from a spherical Lambertian source
If the disc is replaced with a Lambertian sphere of the same radius, as in Fig.
2.31, Eq. (2.62) may still be used, except that the sine squared of the half angle
now becomes
(2.67)
F
Table 2.9 Irradiance at the detector as a function of source distance for a Lambertian
sphere, measured from a surface.
Distance Half-angle @112 (deg) Irradiance Ed
= = ~ ~ ~
b 2a Very small TtLa
2
b-
2
=Ib-
2
b = 2a 19.5 nL/9
b = a 30 nLl4
b=O ~ TtL
==============================
and the expression for irradiance at the detector is
2
E =TtL a
d a
2
+ 2ab + b
2

(2.68)
Table 2.9 summarizes irradiance at the detector for a variety of cases. Note
that when b 2a, the inverse square law applies and the irradiance from the
sphere is the same as that from the disc, above.
If the source-detector distance is measured from the center of the sphere, as
shown in Fig. 2.32, the sine of the half angle is always a/b. The irradiance at the
detector is therefore
(2.69)
Equation (2.69) reveals an interesting result: the inverse square law holds for any
sphere and at any source-detector distance, as long as the surface is Lambertian
and the distance is measured from the center of the sphere. This counterintuitive
result simplifies calculation; results are shown in Table 2.10.
Figure 2.32 On-axis Lambertian sphere, irradiance measured from the center.
46 Chapter 2
Table 2.10 Irradiance at the detector as a function of source distance for a Lambertian
sphere, measured from the center.
Distance Half-angle @1/2 ( deg)
b2a
b=2a
b=a
b=O
2.8 Integrating Sphere
very small
30
90
Irradiance Ed
rtLa
2
b -2=Ib-
2
rtLI4
rtL
The integrating sphere, invented by British scientist W. E. Sumpner in 1892, and
fully described by German scientist R. Ulbricht a few years later, is a device that
provides a spatially uniform source of radiance. It is depicted in Fig. 2.33, with
two elements of area inside the sphere labeled dA I and dA
2
, the linear distance
between them d, and the sphere radius R.
To analyze the sphere's behavior, we begin with the differential form of the
equation of radiative transfer, Eq. (2.43):
By inspection, 8
1
= 8
2
= 8 and cos8 = d(2Rr
1
Also assume that dA I = dA
2
=
dA. If the interior of the sphere is Lambertian, i.e., coated with material having
Lambertian properties, then
(2.70)
Figure 2.33 The integrating sphere. (Adapted from Ref. 15 with permission from John
Wiley & Sons, Inc.)
pa
This result means that irradiance within the sphere, for any area element dA,
is independent of position e within the sphere and is dependent only on sphere
radius and radiance L. In other words, irradiance is constant over the sphere. This
fact makes the integrating sphere useful as a uniform radiance source.
If a source with power is placed into the sphere (through a "port" in the
sphere), the radiance of the sphere wall L (assumed to be Lambertian) can be
determined as
L=Ep
, (2.71)
1t
where p is a property of the sphere coating material called its reflectance (to be
discussed in detail in Chapter 3). Combining Eqs. (2.70) and (2.71) and solving
for dE, we obtain
(2.72)
This is the irradiance on an infinitesimal element of sphere area dA.
Integrating over the area to produce sphere irradiance is complex, as it must take
into account multiple reflections within the sphere. The bottom line is
(2.73)
This result is interesting because as the reflectance approaches unity, the
irradiance approaches infinity, as all the input power remains in the sphere. Real
sphere coatings are nonideal, however, with non-Lamberti an surfaces and
reflectances less than one. Real spheres are fitted with ports and baffles, the
purpose of the latter to prevent "first pass" (unreflected) radiation from reaching
the detector. A useful equation for the radiance in a practical sphere is 16
L = p
1tA.
ph
[I - p(l- f)] ,
(2.74)
in which f is the ratio of the total port area to that of the sphere. (A sphere may
have several ports.) Thus, real spheres are not particularly efficient unless
reflectance is high and the total port area is kept small. Table 2.11 details some of
the many uses of integrating spheres.
48
Table 2.11 Some integrating sphere applications.
Uniform light sources
Measurement of transmission
Depolarization
Light source mixing
Uniform detection systems
Measurement of reflectance
Cosine receiver
Color mixing
2.9 Radiometric Calculation Examples
2.9.1 Intensities of a distant star and the sun
Chapter 2
Figure 2.34 depicts a simplified configuration in which a distant star is viewed by
a telescope. Assuming that there are no atmospheric or optical system
transmission losses (i.e., that the mirror reflects 100% of the incident radiation),
that all power collected by the mirror is relayed to the detector, and that 10-{i W
are incident on the detector, we can determine the irradiance at the detector. We
can then use detector irradiance to calculate the star's intensity.
As seen in Fig. 2.34, the system is j12 with a focal length of 1 m. The mirror
diameter D isfl(f/#), or 0.5 m. The area of the mirror (assumed circular) is 7tIY14.
The irradiance on this (perfectly reflecting) mirror is
The inverse square law applies due to the source distance, and as the source
is on axis, no cosine term is required. Inverting Eq. (2.39) to calculate intensity,
we have
A DET
"Q. ............ _ .......... __ .......... _ ..... _ .............. / \ / ................................. .

fl2,
f= 1
Figure 2.34 Hypothetical distant star and system used to measure irradiance.
-
Now, the intensity of our sun can be approximated by a spherical blackbody
source at 5750 K. Its radiance is given by the following expression (which will
be discussed in detail in Chapter 4):
(2.75)
and its value is 2 x 10
7
W/m
2
sr. Consider the geometry in Fig. 2.35, where Ap is
the projected area of the sun. The sun's diameter is 1.4 x 10
9
m, so according to
Table 2.2, its projected area is (1td
sun
2
)/4, or 1.54 x 10
18
m
2
At earth-sun distance
d = 1.5 x 10 II m, the solid angle subtended by the sun at the detector is nsd = 6.8
x 10-
5
sr. Noting also that the sun subtends approximately 32 minutes of arc
(arcmin), n sd may also be calculated as 1tsin
2
(16 arcmin), which produces the
same result. The irradiance at the detector, area A in the figure (assumed to be
placed at the top of the atmosphere, therefore no atmospheric transmission loss),
1S
(2.76)
Note that the diameter-to-distance ratio is substantially less than 0.1, and the
mverse square law may be applied. Calculating intensity as in the example
above:
(2.77)
I
d
Figure 2.35 Source-detector geometry for solar irradiance calculation.
50 Chapter 2
Irradiance may also be obtained in another way. Consider that the power
delivered to the detector with area A may be represented by
(2.78)
where O ds is the solid angle subtended by the detector, and that
 LA
E =_d =-_P =1368 W/m
2
.
d A d
2
(2.79)
Applying the inverse square law as in Eq. (2.77), we obtain I sun = 3.08 X 10
25
W Isr. Note that though the numbers are not identical, they are very close. The
value of 1368 W/m
2
is referred to as the solar constant, and is specifically defined
as the irradiance falling upon a I m
2
unit surface (hypothetical surface) at the
mean earth- sun distance. The solar constant has wide application in fields
including remote sensing. It is often given the symbol Eo.
Note that the total intensity of the sun has to do with the power radiated into
41t sr, the solid angle of a sphere as referenced in Table 2.3. Table 2.12 provides
relevant calculations related to solar power and intensity.
2.9.2 Lunar constant
2.9.2.1 Calculation
This concept is analogous to that of the solar constant, whose 1368 W/m
2
are
incident upon the moon as well. If the moon is assumed to be Lambertian, with a
reflectance of 0.2, its radiance is
L = EaP = 87 W/m
2
sr.
mi)()n 1t
Table 2.12 Solar quantities and their values.
Quantity
Solar area
Total solar radiant exitance
Total solar power
Total solar power (alternative)
Intensity
Intensity (alternative)
Formula
Asun = 4Ap
M=1tL
 = MA
 = 41tI
1= /41t
1= Eaef
Value
6.16 x 10
18
m
2
6.28 x 10
7
W/m2
3.87 x 10
26
W
3.87 x 10
26
W
3.08 x 10
25
W/sr
3.08 x 10
25
W/sr
(2.80)
-
At the earth' s surface, the angular subtenses of the moon and the sun are the
same, approximately 32 arcmin. This means that QME, the solid angle subtended
by the moon at the earth, is equal to Qsd, above, with a value of 6.8 x 10-
5
sr.
Neglecting the relatively minimal distance between the top of earth's atmosphere
and its surface, the irradiance produced by the moon at the top of earth's
atmosphere is
Emoon. TOA = E = L moonQME = 5.9 x 10-
3
W/m
2
.
pE
(2.81)
Note that in the above equation, A pE is the projected area of the earth,
analogous to the projected area of the sun discussed earlier.
2.9.2.2 Moon-sun comparisons
Comparing irradiances from the sun and moon, we have
E sun = ~ J [ 1368 '!;lm2 2) = 2.3 x l0
5
.
Emoon E o, moon 5.9xl0 W/m
Also, comparing radiances we find
The numbers are the same because the solid angles subtended are the same.
Assuming an atmospheric transmission of 0.75, the solar irradiance at the
earth's surface is this factor multiplied by the solar constant
E eort" = 'tE o = 1026 W/m
2
,
and assuming an earth reflectance of 0.2 along with the Lambertian
approximation, the earth' s radiance is
L = E earthP = 65 W/m
2
sr
eorth 7t .
Applying an atmospheric transmittance of 0.75 to the moon' s radiance at the
top of the atmosphere, we obtain the moon's apparent radiance; that is, its
radiance when viewed from the ground:
L'moon = 'tL
moon
= 65 W/m
2
sr
52 Chapter 2
This interesting result means that the radiance of an "average" sunlit scene
t
is
the same as the apparent radiance of the moon. It also means that in photography,
the same exposures can be used to photograph both. Exposure parameters should
be set during the daytime and applied to night photography. If the moon is
photographed through a telescope, exposures should be increased to compensate
for transmission losses within the instrument. In addition, given the factor of
2.3 x 105 difference between solar and lunar irradiances, photographing a
moonlit scene requires significantly longer exposures than are needed for
daylight illumination.
A point about assumptions should be made, specifically, that the moon is not
a strict Lambertian surface. It is somewhat retrorefiective, as though covered
with Scotchlite. Simple measurements made by Palmer with a silicon detector
indicate that the apparent intensity of the full moon is more than ten times that of
the quarter moon. When viewed with a telescope or binoculars, the edge appears
a bit brighter than the rest. The lunar surface is dusted with small glassy
spheroids, ejecta from meteorite collisions. Its reflectance is approximately 0.08,
somewhat less at shorter wavelengths and somewhat more at longer wavelengths.
2.9.3 "Solar furnace"
This example concerns a "solar furnace" operated in space, delivering power to a
collector just outside the earth's atmosphere, but the equations are valid for any
source located at a large distance from a collector. What is the irradiance
delivered to the target E
t
?
Consider Fig. 2.36, in which the sun is represented by the vertical bar at the
left-hand side with radiance [Eq. (2.75)] of 2 x 10
7
W/m
2
sr. The power from the
sun to the collector is <l>se = LsAcn se, using the area of the collector and the solid
angle the sun subtends at the collector. The irradiance at the collector is then
E = <l> s->c = L n
c A s SC'
C
where n se is the solid angle the sun subtends at the collector, 6.8 x 10-
5
sr.
Choosing a target diameter (or linear dimension) and system focallengthfso that
nsc = n/e, we have
t Eastman Kodak has shown through extensive research that the reflectance of an average scene is
18%; all exposure meters are calibrated using this assumption (1. M. Palmer, 2005).
F
Ac 0
tc
(
... I'"
......... . .......... ...
------------ ~ ~
L
- 2 10
7
W -2 -1 6............ " i ~ . ~ .
S
- X m sr ....... I ;
.' ~
o - ;,; t
sc- tc ;
f-.... I
Figure 2.36 The "solar furnace."
Assuming no transmission loss between collector and target, the irradiance at
the target E
t
is expressed as
Therefore, the target irradiance is the product of the source radiance and solid
angle the collector subtends at the target. That solid angle may also be
characterized [Eq. (2.15)] as
Q = n
ct 4(/ / #)2 '
so that
E = nLs
t 4(/ / #)2 '
(2.82)
which provides a way of characterizing target irradiance in terms of both source
radiance and thejl# system parameter, for a configuration such as this one.
2.9.4 Image irradiance for finite conjugates
The definition of jl# presented earlier was for an object at infinity; however,
many systems operate at finite conjugates. Figure 2.37 depicts such a system, in
which neither image nor object is at infinity. In such cases, a "workingjl#," often
symbolized as jl#', is used.
17
54 Chapter 2
"
..... : ..... - . - . - . - . ~ . .:-:=.::-=:::. =+-=f:=-=::::. ~ . : : . : . - . - . - . - ...... ,.:"!
L
Figure 2.37 Finite conjugates.
A workingjl# is defined as
(2.83)
where magnification m is the ratio of image height to object height, and has
values between 0 and infinity. The term mp is pupil magnification, the ratio ofthe
diameter of the exit pupil to the diameter of the entrance pupil, and has values
between 0.5 and 2. For a single lens or mirror, it is always 1. Substitutingjl#' for
jl# in Eq. (2.82) we obtain
(2.84)
If mp = 1, as it frequently does, Eq. (2.84) becomes the camera equation
(2.85)
Table 2.13 shows two important cases.
Table 2.13 Target irradiance using the camera equation.
Case I - Object at infinity Case II - Equal conjugates
m=O m= 1
E= TtL
I 4(/ /#)2
E= TtL = TtL
I 4(/ /#/ (1 + 1)2 16(/ /#/
....
The expressions for irradiance at the target show us that image irradiance
decreases as the in-focus object is moved closer to the camera. In order to
maintain focus, the detector must be moved backward, which decreases the solid
angle of the lens as seen from the detector n et by a factor of four.
2.9.5 Irradiance of the overcast sky
A reasonable value for the radiance of the overcast sky is 50 W/m
2
sr, somewhat
less than the 65 W Im
2
sr calculated above for a typical sunlit scene. Assuming
that the sky radiance is constant, with no brightening at the horizon, the
irradiance from the sky at the earth' s surface is
 21[ 1[/ 2
E ear' h = sky4 earlh = L Sky n sky-ear,h = L sky J d<l> J sin a cos ada
A earlh 0 0
E earlh = rrLsky sin
2
(90 deg)
E earlh = 157 W/m
2

This is a factor of 6 or 6.5 less than the irradiance received from the sun on a
clear day (1000 W/m2), which explains why flat-plate solar collectors continue to
function well on a cloudy day (provided that the clouds are "conservative"
scatterers.) By comparison, on a clear day, the diffuse solar irradiance (excluding
the direct beam) can be as high as 50 to 100 W/m
2
due to scattering.
2.9.6 Near extended source
A near extended source such as the one shown in Fig. 2.38 may be found in the
laboratory. It provides a nice way to calibrate a radiometer, because:
(1) If the image of the extended source overfills the field of view of the
detector with area Ad, the distance d is unimportant;
(2) If the source is Lambertian, the angle between source and optical axis is
unimportant; and
(3) If the detector or radiometer with area Ad is not placed exactly at the
focal distance f, it doesn't matter.
The power <l>don area Ad is calculated according to
which equates to
56 Chapter 2
8
~ d ~
,--.. 1
Figure 2.38 Near extended source.
2.9.7 Projection system
Figure 2.39 depicts two different designs for projection systems. The Abbe
projector was invented first, and has significant disadvantages. As can be seen
from the diagram, the source is imaged onto the slide, which is then imaged onto
the screen. Hot spots can occur at the slide, resulting in smoke.
The Koehler system is superior. The source is imaged into the projection
lens, a pupil location rather than an image location. The slide is positioned in an
area of relatively uniform brightness, allowing for a more uniform image on the
screen.
CONDENSER
CONDENSER
PROJECTION LENS
. -::':"'6 J'-"-"_o,
........
CR ........
SCREEN
ABBE PROJECTOR
PROJECTION LENS
,'-:' -t'[oo_"-,,
CR -- --
SCREEN
KOEHLER PROJECTOR
Figure 2.39 Two projection systems.
,..
The equation for illuminance on the screen resembles the camera equation,
Eq. (2.85), with the addition of a cos
4
e term to account for the off-axis angle to
the screen as seen from the projector:
E = 'tartLv cos
4
e
v 4(//#)2(1+m) 2'
(2.86)
where Ev is illuminance. The transmission of the optical system 'to appears also.
To maximize irradiance for a given magnification, there are only two
possibilities: minimize the j/# or maximize the radiance of the source. Candidate
sources with high radiance values include tungsten lamps, tungsten-halogen
lamps, carbon arcs, xenon arcs, metal-halide lamps, and high-brightness
phosphor screens.
2.10 Generalized Expressions for Image-Plane Irradiance
2.10.1 Extended source
To provide a more general expression for the irradiance at the image plane from
an extended source, several factors must be added to the expression in Eq. (2.85).
First is a cos
n
term, accounting for the reduction in irradiance as we look off axis.
Its value is typically 4 to account for projections of the source and target areas,
but good optical designers can reduce this factor to 3.
18
Next, losses in the optical
system due to transmission, reflection, and scattering may be combined into the
general term 'to as discussed above. A term to account for vignetting, lv, the
reduction in the cross-sectional area of the beam as the off-axis angle is
increased, applies as well. Finally, to account for the presence of a central
obscuration in a system such as a Cassegrain, the factor (1 - A2) is applied, where
A is the ratio of the diameter of the central obscuration to the diameter of the
primary mirror. (If there is no central obscuration, this factor can be eliminated.)
Considering the above terms and using the most general expression for
source radiance, the expression for image-plane irradiance from an extended
source becomes
E = 't
o
1tfv(1- A
2
)L(e,<jcos
n
e
I 4(/ /#)2 (1 + m)2 .
(2.87)
Note that this expression does not take into account the spectral nature of the
radiation, to be discussed in greater detail in Chapter 4, nor does it account for
the transmission of the atmosphere.
58 Chapter 2
2.10.2 Point source
Ifwe assume an isotropic source, apply the factors for vignetting, etc., mentioned
above, and begin with Eq. (2.41) cosine
3
law, we obtain
E = 'tal: (1- A2)1 cos
n
e
I d2 '
(2.88)
where n is 3 for an isotropic source.
2.11 Summary of Some Key Concepts
This chapter has presented a number of concepts fundamental to an
understanding of radiometry; a short summary of some appears below.
First, the basic equation of radiative transfer in differential form:
Second, the integral form of this equation:
Handy simplifications can be utilized, provided that their underlying
assumptions are met. These are
 = LAO, ,
and
E=Lo..
Further, the Lambertian approximation is okay for most emitters, but angles
must be considered when applying it to metals. It is alright for matte reflectors,
and no good at all for specular reflectors.
The choice of solid angle for radiometric calculations is an important one.
When the source or receiver is isotropic, solid angle ro may be used. When the
source or receiver is Lambertian, projected solid angle 0. is the correct choice.
A final comment on sources deserves mention. These "notes" (lyrics by Jon
Geist and Ed Zalewski, NIST, ca. 1982) are to be sung to the theme song from
the ancient television series, "Mr. Ed:"
A source is a source, of course, of course
And no one can make a point of a source
Unless, of course, it's the sort of source
That only exists in your head!
For Further Reading
F. C. Grum and R. J. Becherer, Radiometry, Vol. 1 in Optical Radiation
Measurements series, F. Grum, Ed., Academic Press, New York (1979).
R. McCluney, Introduction to Radiometry and Photometry, Artech House,
Boston (1994).
W. L. Wolfe, Introduction to Radiometry, SPIE Press, Bellingham, Washington
(1998).
E. F. Zalewski, "Radiometry and Photometry," Chapter 24 in Handbook of
Optics Vol. 2: Devices, Measurements, and Properties, Second Edition, M.
Bass, Ed., Optical Society of America, Washington, D.C. (1994).
References
1. Guide for the Use of the International System of Units (SI), NIST SP811,
U.S. Government Printing Office (1995).
2. F. E. Nicodemus, Self-Study Manual on Optical Radiation Measurements:
Part I-Concepts, NBS Technical Note 910-1, p. 68, NBS, U.S. Government
Printing Office, Washington, D.C. (1976).
3. C. Wyatt, Radiometric Calibration: Theory and Methods, Academic Press,
New York (1978).
4. S. Chandrasekhar, Radiative Transfer, Clarendon Press, Oxford (1950),
reprinted Dover (1960).
5. P. Drude, Lerhbuch der Optik (1900), translated into English and reprinted as
The Theory of Optics, Longmans, Green, New York (1902).
6. B. N. Taylor, The International System of Units (SI), NIST SP330, U.S.
Government Printing Office, Washington D.C. (1991).
7. J. M. Palmer, "Getting intense on intensity," in Metrologia 30(4), pp. 371-
372 (1993).
8. J. M. Palmer, "Intensity," in Optics & Photonics News, p. 6, February
(1995).
9. P. D. Foote, "Illumination from a radiating disc," Bulletin of the Bureau of
Standards, NBS, 12, p. 583 (1915).
10. R. Kingslake, Optical System Design, Academic Press, New York (1983).
11. B. T. Chung and P. S. Sumitra, "Radiation shape factors from plane point
sources," J. of Heat Transfer 94(3), pp. 328-330 (1972).
12. E. M. Sparrow and R. D. Cess, Radiation Heat Transfer, Brooks/Cole,
Belmont, California (1970).
13. P. Moon, The Scientific Basis of Illuminating Engineering, McGraw-Hill,
Dover, New York (1936).
60 Chapter 2
14. M. Donabedian, "Cooling systems," Chapter 15 in The Infrared Handbook,
W. L. Wolfe and G. L. Zissis, Eds., u.s. Government, Washington, D.C.
(1978).
15. R. W. Boyd, Radiometry and the Detection of Optical Radiation, Wiley &
Sons, New York (1983).
16. "A guide to integrating sphere theory and applications," Labsphere Inc., at
http://www.labsphere.comitecdocs.aspx (2006).
17. R. Kingslake, Optical System Design, Academic Press, New York (1983).
18. P. N. Slater, Remote Sensing: Optics and Optical Systems, Addison-Wesley,
Reading, Massachusetts (1980).
Chapter 3
Radiometric Properties of
Materials
3.1 Introduction and Terminology
When radiant flux is incident upon a surface or medium, three processes occur:
reflection, absorption, and transmission. A fraction of the beam is reflected,
another fraction is absorbed, and the remainder is transmitted. Transmittance 't is
the ratio of transmitted power to incident power. Reflectance p is the ratio of
reflected power to incident power. Absorptance a is the ratio of absorbed power
to incident power.
Figure 3.l shows an ideal geometric case, where the transmitted and reflected
components are either specular (regular, in the mirror direction) or diffuse
(scattered into the hemisphere). Figure 3.2 shows the transmission and reflection
for real surfaces. Both spectral and directional properties are important.
INCIDENT
BEAM
SPECULAR
REFLECTION
RETROREFLECTION
DIFFUSE
REFLECTION
SPECULAR (REGULAR)
TRANSMISSION
Figure 3.1 Idealized reflection and transmission.
61
62
STRONG DIFFUSE
REFLECTION
STRONG SPECULAR
REFLECTION
STRONG RETI'10-
REFLECTION
DIFFUSE RANSMISSION
Figure 3.2 Generalized reflection and transmission.
Chapter 3
A continuing dialog over terminology has taken place, particularly over the
suffixes -ance and -ivity. I.2.3.4.5 The usage here reserves terms ending with -ivity
(such as transmissivity, absorptivity, and reflectivity) for properties of a pure
material, while the suffix -ance is used when the characteristics of a specific
sample are described. One can then distinguish between the reflectivity of pure
aluminum (as calculated from the complex index of refraction nand lC) and the
reflectance of a particular specimen of 6061 aluminum with surface structure
associated with rolling or machining and with a natural oxide layer.
The adjective spectral refers to a characteristic at a particular wavelength and
is indicated as a function of wavelength A, i.e., 'teA), peA) or a(A). For example,
spectral transmittance 'teA) is often plotted against wavelength A for a colored
filter. The absence of "spectral" implies integration over all wavelengths,
weighted by a source function.
3.2 Transmission
Transmission is the process by which incident radiant flux leaves a surface or
medium from a side other than the incident side (usually the opposite side). The
spectral transmittance 'teA) of a medium is the ratio of the transmitted spectral
flux <PAt to the incident spectral flux <P
Ai
:
't(A)= <PAt .
<PAi
(3.1)
Total transmittance 't is the ratio of the total transmitted flux <P
t
to the total
incident flux <Pi:
Radiometric Properties of Materials 63
<P r 't(A.) <P).i dA.
't = _I = :f:: 1 't(A.)dA..
<Pi r <PI-i dA.
(3.2)
Note particularly that the total transmittance is not the integral over
wavelength of the spectral transmittance; it must be weighted by the incident
source function <PI-i.
The transmittance may also be described in terms of radiance as follows:
(3.3)
where LAi is the spectral radiance L
A
i
(A.;8
i
,<j>;) incident from direction (8
i
,<j>i), LAI is
the spectral radiance L
A
t
(A. ;8
t
,<j>t) transmitted in direction (81)<j>1), and dO. is the
elemental projected solid angle sin8 cos8d8d<j>.
Geometrically, transmittance can be classified as specular, diffuse, or total,
depending upon whether the specular (regular) direction, all directions other than
the specular, or all directions are considered. The bidirectional transmittance
distribution function (BTDF, symbol It and units sr-
1
) relates the transmitted
radiance to the radiant incidence (irradiance) as
(3.4)
This descriptor facilitates specification of the angular dependence of all (specular
and diffuse) transmitted radiation.
3.3 Reflection
In reflection, a fraction of the radiant flux incident on a surface is returned into
the hemisphere whose base is the surface containing the incident radiation. The
reflection can be specular (in the mirror direction), diffuse (scattered into the
entire hemisphere), or a combination of both. Table 3.1 shows a wide range of
materials that have different goniometric (directional) reflectance characteristics.
Spectral reflectance is defined at a specific wavelength A. as
(3.5)
64 Chapter 3
Table 3.1 Goniometric classification of materials.
6
(Reprinted with permission of CIE.)
Material
Scatter
"(
Structure Example
classification (de2)
Exclusively
none 0
=0
none Mirror
reflecting
materials
weak :::;0.4 :::;27
mIcro Matte aluminum
macro Retroreflectors
1=0 none Lacquer & enamel
coatings
strong > 0.4 > 27
mICro Paint films, BaS04,
PTFE
macro Rough tapestries, road
surfaces
Weakly transmitting,
Sunglasses, color filters,
strongly reflecting none 0
=0
none
materials
cold mirrors
mICro Matte-surface color
weak :::;0.4 :::;27 filters
1:::; 0.35 macro Glossy textiles
none Highly turbid glass
strong > 0.4 >27 micro Paper
macro Textiles
Strongly none 0
=0
none Window glass
transmitting
materials none Plastic film
weak :::;0.4 :::;27
micro Ground glass
1 > 0.35 macro Ornamental, prismatic
~ l s s
none Opal glass
micro Ground opal glass
strong > 0.4 >27 macro Translucent acrylic
plastic with patterned
surface
while the total reflectance p is the ratio of the reflected flux <Pr to the incident
flux <Pi:
(3.6)
As in the case of transmittance, above, the integrated reflectance is not the
integral over wavelength of the spectral reflectance; it must be weighted by the
incident source function ).i.
Reflectance factor R or RCA) is defined as the ratio of (spectral) radiant flux
reflected by a sample to the (spectral) radiant flux which would be reflected by a
perfect diffuse (Lambertian) reflector under the same irradiation conditions.
While reflectance (in the absence of luminescence) cannot exceed unity,
reflectance factor can assume values from zero to nearly infinity. Since the
reference is a perfect diffuse reflector, reflectance factor is only useful as a
descriptor for diffuse surfaces. Equations for nine types of reflectance factor
appear in Table 3.2.
Some notes on Table 3.1 are in order:
(I) "Structure" refers to the nature of the surface. In a micro scattering
structure, the scatterers cannot be resolved with the unaided eye. The
macrostructure scatterers can be readily seen.
(2) Sigma, cr, is a diffusion factor, the ratio of the mean of radiance
measured at 20 deg and 70 deg to the radiance measured at 5 deg from
the normal, when the incoming radiation is normal.
cr = [L(20) + L(70)] / [2L(5)]. It gives an indication of the spatial
distribution of the radiance, and is unity for a perfect (Lambertian)
diffuser.
(3) Gamma y is a half-value angle, the angle from the normal where the
radiance has dropped to one half the value at normal. Its value is 60 deg
for a perfect (Lamberti an) diffuser.
(4) It is suggested that the diffusion factor is appropriate for strongly
diffusing materials and that the half-value angle is better suited for
weakly diffusing materials.
No single descriptor of reflectance suffices for the wide range of possible
geometries. The fundamental geometric descriptor of reflectance is the
bidirectional reflectance distribution function (BRDF) f,.. It is defined as the
differential element of reflected radiance dL
r
in a specified direction per unit
differential element of irradiance dEi, also in a specified direction.
7
It carries the
unit of sr-
I
.
(3.7)
As shown in Fig. 3.3, the polar angle e is measured from the surface normal,
z. The azimuth angle is measured from an arbitrary reference in the surface
plane, most often the plane containing the incident beam. The subscripts i and r
refer to the incident and reflected beams, respectively.
66 Chapter 3
z
,
/j
Y
<l>i
X
<1>,
Figure 3.3 Geometrical definitions for BRDF (Adapted from Ref. 7).
Nicodemus et a1.
7
integrated over various solid angles and applied the earlier
work of Judd
8
to obtain nine goniometric reflectances and nine goniometric
reflectance factors. These are listed in Tables 3.2 and 3.3. In these tables, the
term "directional" refers to a differential solid angle dO) in the direction specified
by (8,<1. "Conical" refers to a cone of finite extent centered in direction (8,<1; the
solid angle Q of the cone must also be specified. The reflectances are illustrated
in Fig. 3.4.
Table 3.2 Nomenclature for nine types of reflectance factor.?
Bidirectional
reflectance factor
Directional-conical
reflectance factor
Directional-
hemispherical
reflectance factor
Conical-directional
reflectance factor
Biconical
reflectance factor'
=
r
= (8i ,
p
8r,r )dQ;
I
t
I r
-
Conical-
= hemispherical R(oo;;21t)
reflectance factor

I
Hemispherical-
= !nf(8;<\>;;8
r
<\>r)dQ; directional R(21t;8
r
,<\>,)
reflectance factor
Hemispherical-
= !n 1, f(8;,<\>;;8r,<\>,)dQrdQ; conical reflectance R(21t; 00,)
factor
.
r
B ihemispherical
R(21t;21t)
11 f(8p<\>;; 8
r
' <\>r)dQ
r
dQ;
reflectance factor 1t It It
Configurations that are measurable in practice.
Table 3.3 Nomenclature for nine types of reflectance.
7
Bidirectional
dp(8p<\>; ; 8
r
' <\>r) = f(8p<\>;;8
r
'<\>r)dQr
reflectance
Directional-conical
p(8;, <\>;; OOr) = 1, f(8;,<\>;;8
r
,<\>r)
dQ
r
reflectance
Directional-
= !nf(8;,<\>;; 8
r
,<\>,)
dQ
r hemispherical P(8p<\>i; 21t)
reflectance
Conical-directional
dp( 00;; 8
r
, <Pr)
- dQ
r
1 .
reflectance
- fr ; f(8;,<\>p8
r
' <\>r)dQ;
I
Biconical
p( 00;; 00,) = 1, f(8;,<\>;;8r,<\>r)dQrdQ;
reflectance

I
Conical-
= !nf(8i'<\>;;8r,<\>,)dQrdQ; hemispherical p( 00;; 21t)
reflectance
I
Hemispherical-
- dQ
r
1 .
directional dp(21t;8
r
,<\>,) - - f(8;<\>i' 8
r
<\>,)dQ;
reflectance
1t It
Hemispherical-
11 f(8p<\>;;8
r
,<\>,)dQ
r
dQ; conical p(21t; OOr)
reflectance

1t It ,
Bihemispherical
p(21t;21t)
11 f(8;, <\>;; 8r,<\>,)dQrdQ;
reflectance 1t It It
Configurations that are measurable in practice.
In both Tables 3.2 and 3.3, configurations containing a directional term are
considered theoretical, as dQ.-:;O.
Lt wel ' Hr ' lnn;

DlRfCTIOIW.
INCIDENT
OIRECT1OIIAL
COI.l.ECTBl
BIDIRECTIONAl REFLECTANCE

CONICAL
INCIDENT COUECTBl
CONICAlDIRECTIONAl REFLECTANCE

DlRECTlOIW.. COftCH. DIRECTIONAL HEMISPHERICAL
INCIDENT COllECTED INCIDENT COlLECTED
DIRECTIONAL-{;ONICAl REFLECTANCE DIRECTIONAlHEMISPHERICAl REFLECTANCE
!! !'i)
BICONICAL REFLECTANCE CONICAL-HEMISPHERICAL REFLECTANCE
r71. P;
" y -c z,.. 3'
Q
HBOSI'HERICAI.
IICIlENT
DIRECTIONAl.
CCUECTED
......... ERJCAl
INCIOBIT
CONICAl
COlLECTED
;
HEMISPHERICAL
IHCfDHT
tBIISPHERICAL
COllECTED
HEMISPHERICAl.DIRECTIONAL REFLECTANCE HEMISPHERICAl-{;ONICAl REFLECTANCE BIHEMISPHERICAL REFLECTANCE
Figure 3.4 Nine reflectance configurations.
en
00
9

(b
..,
W
. me" anent r r Yftttr l 'ftY? 111m .. tnt 'j'
'#'s '" . "bY "H ""u 'f r*1"" "n ' C' r 'W' it "ft ', ' '}' t-swtc'; w-iffrettt "' t6 t ," ,"., , ,
--
3.4 Absorption
Absorption is the process in which a fraction of the incident radiant flux is
converted to another form of energy, usually heat. Absorptance is the fraction of
incident flux that is absorbed. Spectral absorptance is defined at a specific
wavelength A as
(3.8)
with the subscripts denoting absorbed and incident power, respectively. Total
absorptance is defined as
(3.9)
Note the analogy with Eqs. (3.2) and (3.6).
Absorption removes power from a beam; directional characteristics such as
direct absorption and bulk scattering are not often taken into consideration.
3.5 Relationship Between Reflectance, Transmittance, and
Absorptance
Because energy is conserved, the sum of the transmission, reflection, and
absorption of flux incident on a surface is unity, or
't+P+U=1. (3.10)
The above statement assumes integration over all wavelengths and directions. In
the absence of wavelength-shifting effects (such as luminescence or Raman
scattering), this relationship is also valid for any specific wavelength:
'teA) + peA) + u(A) = 1. (3.11 )
3.6 Directional Characteristics
3.6.1 Specular transmittance and reflectance
The specular transmittance and reflectance for a single surface can be calculated
via the Fresnel equations using the complex index of refraction n + iK. The
simplest case is that of no absorption, i.e., K = O. The general equations are
70 Chapter 3
(
, )2
n cos8 - ncos<l>
P p = n' cos8 + ncos<l>
2 '
= (ncos8 - n' cos <1
Ps ,
ncos8+ n cos<l>
(3.l2)
and
( )
'
2n cos 8 - n' cos 
't p = n' cos 8 + n cos (n cos 8 )
(3.13)
(
2ncos8 )2 (n' cos <1 ,
't
s
= ncos8+n'cos<l> ncos8
where the subscripts p and s represent the two polarization states, nand 8 are on
the incident side of the interface, and n' and are on the transmitted (or reflected)
side. The total transmittance and reflectance for unpolarized light is the average
of the two polarized components
and (3.14)
Figure 3.5 shows reflection and transmission curves for a single surface of a
nonabsorbing optical material with an index of refraction of 2. The three curves
represent s- and p-polarization states as well as total polarization. To compute the
curves for absorbing media, substitute n iK for n in Eqs. (3.12) and (3.13),
where K = a'A/4n . Since refractive index n is wavelength dependent, the
calculated reflectance and transmittance are also.
In Fig. 3.6, we see a partially transparent plane slab of an optical material.
Reflection, transmission, and absorption are all present.
The Fresnel equations are greatly simplified at normal incidence, in which
8 = = O. For a nonabsorbing material, the reflectance and transmittance at a
single surface are
(
' )2
n -n
Pss = n' + n
4nn'
't =------::-
.<s (n'+n)2
(3.15)
The internal transmittance 't; of a piece of optical material describes only the
absorption component and neglects the reflectance losses. The exponential
absorption law, often referred to as the Lambert-Bouguer-Beer law, is
--
UJ
U
z

f-
U
UJ
...J
L1.
UJ
c::
(a)
UJ
U

L1.
c::
:::J
rJl
UJ
...J
Cl
z
iii
UJ
u
z

f-
I-

<J)
z

(b)
I-
UJ
u

L1.
0::
:::J
<J)
UJ
...J
Cl
z
iii
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
--
1
--r--
- __ 1- __
- - r- - -
- __ 1- __
1
- - r- - -
10
- - - - - - - - - - - - - - - - - - - - _1- ___ , _______ I _. _____ , __
I I I I ,
__ i ____ -i- ____ -,- ____ -:- ____ -:- _____ 1 _____ y
- _1- _________ , ____________ : ______ : ______ : ___ -/_:-}
--i--- -- -----,------1---- --:------:------:-/--11
--:- -----;------;- -----;------: ------: -----/f ---f: 1- -
1 1 1 v' 1/_''-
--I------!------ -----:------1------1-7---
1
--- -[--
-:- - - - -- - - - - - - - - - - -- - - - - -:- -5 ./ - - - - -:/- - f: --
_ _ __ ,- __ __ _ 1_ ____ ___ 1/_ -I - -
20 30 40 50 60 70 80 90
ANGLE FROM NORMAL (deg)
- -
,
r
-- - r --
r
- - -- - - -- - - --
1 1
- - f- --
-v--
1
- - 1--
1 __ , l ___
,-
- - f- - - - - f- --
--\\-
1 1
-- - 1- -- - - -,- - -

_ _ L __ __ L __
1
--;---
- - r- --
OL-_L __ __ __ __ __ __ __ __ __ __
o 10 20 30 40 50 60 70 80 90
ANGLE FROM NORMAL (deg)
Figure 3.5 (a) Reflection and (b) transmission of a single surface, n = 2.
'tJA) = e-a'(A)X ,
(3.16)
where a'(A) is the spectral absorption coefficient (cm-
I
) at the specified
wavelength, and x is the thickness (cm). Different units can be used for thickness
and absorption coefficient (i.e., m, km, mm, /lm) as long as they are the same for
both the absorption coefficient and the thickness. The product of the absorption
coefficient and the thickness x is often called the optical thickness 'to. These units
and symbols are used differently in different fields; be careful!
External transmittance is the quantity that is ordinarily measured, and
includes the Fresnel reflection losses and the absorption. Equation (3.17)
describes the transmittance of a parallel slab at normal incidence with single-
surface Fresnel reflection pss> absorption coefficient a', and thickness x:
--_._-------
72
<1>,
Figure 3.6 Transmitted, reflected, and absorbed rays_
If the slab is nonabsorbing (a = 0), then 1; = 1, and Eq_ (3.17) reduces to
2n
1=-z- '
n +1
Chapter 3
(3.17)
(3.18)
Similar equations can be derived for reflectance and absorptance. For
reflectance:
(
1- )z -2a'x (1_)2 2
<Pr Pss Pss e Pss Pss 1j
P
- - - P + - P +
- <P - ss 1- P 2 -Za'x - ss 1- 2 2
j ss e P ss 1j
(3.19)
For absorptance:
a = <Pa = = (l-Ps.)(1 - 1;) .
<Pj I-p
ss
e-
ax
I-PSS1j
(3.20)
When the optical thickness is large enough, the material becomes opaque and
the transmittance goes to zero. In this case, the reflectance P approximates the
single surface reflectance pss, and the absorptance a approaches 1 - p. For the
opposite case, in which the optical thickness approaches zero, the material
becomes transparent and the following relationships hold:
P
=p +Pss (l-p"f
ss 1 Z '
- Pss
(3.21)
(3.22)
As an example, an ordinary transparent glass (n = l.5) has a single-surface
reflectance Pss of 0.04. The total reflectance P is 0.077 (rather than 0.08), and the
total transmittance 't is 0.923.
3.6.2 Diffuse transmittance and reflectance
A large part of the optical radiation in our environment is the result of reflections,
with sources including sunlight and artificial tungsten and fluorescent lamps. For
nonspecular reflection or transmission, the BRDF or its analog, the BTDF, may
be characterized for any combination of incident and reflected beams. A full
BRDF or BTDF specification is very complex, particularly when the reflectance
and transmittance lie out of the plane that includes the incident and specular
beams. Figure 3.7 illustrates this complexity.9
The set of polar diagrams in this figure for a partially diffuse surface is
specific to just one incidence angle. A complete characterization would require
data at other incidence angles as well, as the pattern is variable. Surfaces become
more specular as the angle of the incoming beam increases towards grazing
incidence. In addition, the diffuse scatter from rough surfaces diminishes with
increasing wavelength, i.e., the surface becomes more specular.
The BRDF characteristics of a surface are normally plotted as BRDF (sr-l)
versus angle as measured from the specular beam. This method places the
specular beam on the left ordinate, regardless of incidence angle. Positive and
negative angles as measured from the specular beam are typically shown on the
same graph, and angles as large as 175 deg are seen (for an 85-deg angle of
incidence). The angle axis can be linear for diffuse surfaces, but a log plot is
better suited for specular surfaces as shown in Fig. 3.8
For some materials, it has been found that the diffusely reflected radiation is
symmetrical about the specular beam. In this case, a single plot suffices for each
wavelength. The materials with this characteristic are spatially uniform (called
isotropic in the literature, but different from isotropic as applied to a point
source) and are either nearly specular or nearly perfectly diffusing. Insight into
their BRDF characteristics can be gained through a plot of the natural log of the
BRDF versus a special parameter, (B - Bo), where B is the sine of the scattered
beam angle and Bo is the sine of the specular angle. This places the BRDF into
direction cosine space. Figure 3.9 shows such a BRDF plot for a perfectly diffuse
reflector (fr = lht) and for a very good specular reflector. The plot for the perfect
specular reflector would be a delta function of infinite height at (B - Bo) = 0, and
would not be seen on this plot. What is actually seen is the instrument function
that is primarily the convolution of the incident beam profile with the detected
beam profile.
74 Chapter 3
10'"
Figure 3.7(a) BRDF of rough aluminum at incidence angle of 33 deg. (Reproduced from
Ref. 9 with permission.)
po
Figure 3.7(b) BRDF of rough aluminum at incidence angle of 33 deg. (Reproduced from
Ref. 9 with permission.)
1E2
1EO
L:'
1E-2
~
LL
0
0:::
1 E-4
co
1 E-6
1 E-8
0.1
I I I,
I " "
: : : :: INSTRUMENT
, . , , , P ~ O F l l ~
1 10 100
ANGLE FROM SPECULAR (degrees)
Figure 3.8 Conventional BRDF plot of perfect mirror ("instrument") compared with a
perfect diffuse reflector and two mirrors with scatter.
76
-
L..

u..
0
0:::
III
1 E2
1EO
1 E-2
1 E-4
1 E-6
1 E-8
0.01 0.1
sin(ANGLE)-sin(SPECULAR)
Chapter 3
Figure 3.9 Special parameter versus BRDF plot of perfect mirror ("instrument") compared
with a perfect diffuse reflector and two mirrors with scatter.
A perfect specular reflector has a BRDF equal to p/Q; where p is the
reflectance and Q ; is the solid angle subtended by the source. For example, a
piece of opaque glass with a reflectance p of 0.05 in sunlight (Q; = 6.8 x 10-
5
sr)
has a BRDF of 14706 x P = 735.3. To determine the radiance associated with this
specular reflection, multiply the BRDF by the irradiance. The BRDF for a
hemispherical source is pin, the same as for a perfectly diffuse (Lambertian)
reflector. In the confines of an integrating sphere, diffuse and specular samples
having the same reflectance are indistinguishable.
3.7 Emission
So far, we have considered the radiometric properties of materials with respect to
incoming radiation. In fact, all materials above 0 K radiate, so the emission of
radiation by a material is an important property, as well. The Infrared Handbook
defines emissivity as "the ratio of the radiant exitance or radiance of a given body
to that of a blackbody."l o Its symbol is E. Emissivity may be considered a
"quality" factor, indicating the capability for thermal radiation by a material. It
has both spectral and directional properties, it is dimensionless, and its values are
between 0 and 1. As with the material properties already discussed, emissivity
refers to the characteristics of a pure substance, while emittance refers to the
properties of a specific sample.
Spectral emittance E(A.) is defined as emittance at a given wavelength, and it
is not a derivative quantity. In the case that a radiator is neutral with respect to
wavelength, with a constant spectral emittance less than 1, it is called a graybody.
In that case, the spectral emittance is the ratio of the radiance of that source at
that wavelength to the radiance of a blackbody at that wavelength:
--
Radiometric Properties of Materials
L
(A)=_A ,
LABB
77
(3.23)
where LABB is the value of the Planck function at that wavelength for a
blackbody. Further discussion of this function and the radiometric characteristics
of sources follows in Chapter 4.
As reflectance, transmittance, and absorptance are related, as indicated by
Eq. (3.l0), so too are reflectance, transmittance, and emittance related. At
equilibrium, the power emitted by a body to its surroundings must equal the
power absorbed by the body from its surroundings. More succinctly, the body's
absorptance must equal its emittance:
(x=' (3.24)
This is Kirchhoffs law, to be discussed in more detail in Chapter 4. As a
consequence, emittance may be substituted for absorptance in Eqs. (3.10) and
(3.11), so that
=1-'t-p, (3.25)
and spectrally,
(A) = 1- 'teA) - peA) . (3.26)
If the body is opaque ('t = 0), then
=1-p, (3.27)
and
(A) = 1 - p(A) . (3.28)
3.8 Spectral Characteristics
The radiometric properties of materials of interest all share one spectral
characteristic: the property of interest is not independent of wavelength. Many
material characteristics may be "flat" over a portion of the spectrum, but at other
wavelengths may differ significantly. Since these properties are weighting
functions, the values are irrelevant in those spectral regions where the source
function is insignificant. For example, ordinary window glass has a transmittance
of about 0.92 (clean, normal incidence) in the visible part of the spectrum but
drops to zero in the infrared where the eye is nonresponsive.
78
1or-.--.-----,--------------------"

w 0 .6
o
z
WHITE NONMETALS

w
0 . 2
POLISHED METALS
o I 2 5 15
WAVELENGTH l}"l
Figure 3.10 Spectral emittance for several generic surfaces.
11
Chapter 3
These materials are often used for temperature control of spacecraft. Heating
is by absorption of sunlight for wavelengths shorter than 311m, and cooling
results from thermal radiation for wavelengths longer than 311m. If a material has
high reflectance at shorter wavelengths, it will not absorb much of the incident
radiation. If the reflectance is low at longer wavelengths, the absorption and
consequently the thermal emission will be high. The surface will be cold. If the
spectral regions are reversed, the surface will become hot. These surfaces are
known as selective surfaces, and a wide range of surface temperatures have been
achieved.
Designers of these materials often utilize the ratio alE to describe the value of
absorptance in one spectral region (usually solar) relative to the emittance value
in another region (usually infrared.) Figure 3.10 shows the spectral emittance
(1 - spectral reflectance) for several generic surfaces; some are selective.
The Infrared Handbook provides detailed examples of the radiometric
properties of both natural and artificial sources. An example of the spectral
reflectances of several metals is shown in Fig. 3.11.

---r' - ,.. ------
........ /AI _---
, ':....--
80 ) .- Rh
r-"
leu
/' I .
60, ._--T:..:.'---
20
Wavelength (j.1m)
Figure 3.11 Spectral reflectance of films of silver, gold, aluminum, copper, rhodium, and
titanium.
12
-
3.9 Optical Materials Checklist
Finally, Table 3.4 provides an "optical materials checklist" that includes several
properties useful to the designer selecting materials for an optical design effort.
Table 3.4 Optical materials checklist.
Optical properties
Transmission (function of wavelength, temperature,
direction)
Index of refraction (function of wavelength,
temperature, direction)
Dispersion, partial dispersion
Surface reflectance
Scatter (surface & bulk)
Absorption (bulk)
Homogeneity
Birefringence, stress coefficient
Fluorescence
Anisotropy
Electro-optic and/or acousto-optic coefficients
Mechanical properties
Young's modulus
Yield point
Hardness
Optical workability
Coating compatibility
Density, specific gravity
Thermal properties
Thermal conductivity
Specific heat, heat capacity
Coefficient of linear thermal expansion
Softening point, melting point
Environmental properties
Solubility in H
2
0, other solvents
Surface deterioration, devitrification
Radiation susceptibility (UV, hard particle)
Other factors
Availability
Safety factors, toxicity
Cost
Compiled by James M. Palmer 02/2 1189
-- -------------
80 Chapter 3
For Further Reading
"Standard practice for angle resolved optical scatter measurements on specular
and diffuse surfaces," ASTM Standard E1392-90, ASTM International,
Philadelphia (1990).
"Standard practice for goniometric optical scatter measurements," ASTM
Standard E2387-05, ASTM International, Philadelphia (2005).
1. C. Stover, Optical Scattering: Measurement and Analysis, SPIE Press,
Bellingham, Washington (1995).
References
1. A. G. Worthing, "Temperature, radiation, emissivities and emittances," in
Temperature: Its Measurement and Control in Science and Industry,
Reinhold, New York (1941).
2. 1. C. Richmond, "Rationale for emittance and reflectivity," Applied Optics
21(1), pp. 1-2 (1982).
3. J. C. Richmond, W. N. Harrison, and F. J. Shorten, "An approach to thermal
emittance standards," in Measurement of Thermal Radiation Properties of
Solids, NASA SP-31, 1. C. Richmond, Ed., NASA, Washington, D.C. (1963).
4. W. L. Wolfe, "Proclivity for emissivity," Applied Optics 21(1), p. 1 (1982).
5. J. C. Richmond, J. J. Hsia, V. R. Weidner, and D. B. Wilmering, Second
Surface Mirror Standards of Spectral Specular Reflectance, NBS Special
Publication SP 260-279, U.S. National Bureau of Standards, Washington,
D.C. (1982).
6. Radiometric and Photometric Characteristics of Materials and Their
Measurement, CIE Publication 38 (1977).
7. F. E. Nicodemus, 1. C. Richmond, J. J. Hsia, I. W. Ginsberg, and T.
Limperis, Geometrical Considerations and Nomenclature for Reflectance,
NBS Monograph 160, U.S. National Bureau of Standards, Washington, D.C.
(1977).
8. D. B. Judd, "Terms, definitions, and symbols in reflectometry," J. Opt. Soc.
Am. 57(4), pp. 445-450 (1967).
9. F. E. Nicodemus, "Directional reflectance and emissivity of an opaque
surface," Applied Optics 4(7), cover picture (1965.)
10. W. L. Wolfe, "Radiation theory," Chapter 1 in The Infrared Handbook, W.
L. Wolfe and G. J. Zissis, Eds., pp. 1-28, U.S. Government, Washington,
D.C. (1978).
11. J. C. Richmond, "Coatings for space vehicles," in Surface Effects on
Spacecraft Materials , F. 1. Clauss, Ed., Wiley & Sons, New York (1960).
12. W. L. Wolfe, "Optical materials," Chapter 7 in The Infrared Handbook, W.
L. Wolfe and G. J. Zissis, Eds., pp. 7-81, U.S. Government, Washington,
D.C. (1978).
Chapter 4
Generation of Optical
Radiation
4.1 Introduction
From a discussion in the previous chapter on the interaction of radiation with
materials, we now tum to the subject of how radiation is generated, and the roles
that emission, reflection, and other processes play.
Sources of optical radiation can be classified in a variety of ways. Active
sources emit optical radiation due to their temperature (thermal sources) or as a
result of atomic transitions (luminescent sources). Passive sources reflect optical
radiation from active sources or from other passive sources. Passive sources can
also be classified as thermal or luminescent, depending upon the process that
generated the radiation initially. Examples of thermal sources include blackbody
radiation simulators, tungsten-filament lamps, gases, the sun, the moon, and you
and 1. Examples of luminescent sources include lasers, fluorescent lamps,
mercury arcs, sodium lamps, electroluminescent panels, LEDs, and gases. Some
sources combine both thermal and luminescent mechanisms, and some may be
both active and passive, reflecting in one spectral region and emitting in another.
Other means of classification have also been used. Some authors distinguish
between artificial (man-made) and natural sources. Lamps are artificial sources,
whereas the earth, the sun, and stars are natural sources. Still another practice is
to divide sources according to their output spectral characteristics. Continuous
sources have a spectral radiance that is slowly varying with wavelength, typical
of thermal radiation, while line sources emit in narrow, well-defined spectral
regions. Yet another attempt to distinguish sources is by their degree of spatial
and/or temporal coherence.
Thermal radiation has been extensively studied since the late nineteenth
century. Stefan was the first to experimentally examine the relationship between
radiation and temperature in 1879. He analyzed data from Tyndall and found that
the total radiation is proportional to the fourth power of temperature r.
Boltzmann derived this r
4
relationship from the Camot cycle in 1884. In 1891
Wien derived his displacement law, which relates the peak radiation to
83
84 Chapter 4
temperature, and the r relationship for the magnitude of the peak. In 1896, Wien
derived an equation for the spectral distribution of thermal radiation based on
thermodynamic arguments. In 1900 Rayleigh derived another equation for
spectral distribution based on equipartition, and Jeans in 1905 independently
repeated this derivation. Planck in 1901 published an empirical equation
involving the notion that energy exists as discreet "packets" that fit experimental
data better than either the Wien or the Rayleigh-Jeans equation. This proved
monumental, as it was later accepted as the birth of quantum mechanics.
Verifications of the Planck equation and the physical constants continued up to
1982 with precise measurements of the spectral distribution of blackbody
radiation, determination of the Stefan-Boltzmann constant, and confirmation of
the thermodynamic temperature scale.
4.2 Radiation Laws
4.2.1 Planck's law
The Planck expression for blackbody radiation is at the heart of all thermal
radiation equations. Any object at a temperature above absolute zero (0 K) will
radiate as a modified Planckian radiator. A full derivation of the Planck equation
will not be presented here, as it can be found in many physics texts. A brief
outline of the derivation will show its important features. It requires the
determination of the number of discrete frequency modes in a cavity. The total
mode density in a cavity (in vacuo) as a function of frequency v is:
The average energy per mode looks like
hv
q = hvlkT l'
e -
(4.1)
(4.2)
The energy density (energy per unit volume per unit frequency interval) is
the product of the number of available modes and the average energy of each
mode. It is expressed as:
8rchv
3
1
u = ------,----
v c
3
(ehVl kT -1) .
(4.3)
The energy exits a cavity at velocity c into 4rc sr. The behavior is described by
Planck's law, expressing radiance per unit frequency interval (spectral radiance):
1
I
j
!
1
,
I
1
I
j
,
1

Generation of Optical Radiation 85
(4.4)
Frequency is fundamental (independent of the medium), but wavelength is
preferred to frequency in radiometric applications due to ease of measurement.
Making use of the identity
and the relationship between wavelength and frequency 'A = c/v, the spectral
radiance of blackbody radiation can be expressed in wavelength terms as
L = 2hc
2
1
A 'A
5
ehcl'AkT -1 .
(4.5)
This is one form of Planck's equation, whose units are W/m
2
srm.
The above equation implies that the process is taking place in vacuum where
the refractive index is unity. A more general expression utilizes n. The most
frequently encountered and useful form of the Planck equation is
(4.6a)
Because blackbody radiation is Lambertian, the spectral radiant exitance may be
expressed as
(4.6b)
In this pair of important equations,
2
CI = 21thc = 3.74177107(29) x 10-
16
Wm
2
(7.8 x 10-
8
) (first radiation
constant)
C2 = hc/k = 1.4387752(25) 10-
2
m-K (1.7 x 10-6) (second radiation constant)
h = 6.62606876(52) xl 0-
34
Js (7.8 x 10-
8
) (Planck constant)
C = 299792458 mls (exact) (velocity oflight)
k = 1.3806503(24) x 10-
23
JIK (1.7 x 10-6) (Boltzmann constant)
n = index of refraction (1 for vacuum, ::::1.00028 for air)
86 Chapter 4
The above notation is a conventional way to present uncertainties in the
physical constants. The number in parenthesis that is attached to the value is the
absolute uncertainty (one sigma, or lcr) in the last digit(s) of the constant. For
example, C2 = 0.014387752 0.000000025 mK. The last number in parenthesis
is the relative uncertainty in the value of the constant. It is the absolute
uncertainty divided by the value of the constant.
The complete Planck equation as shown above is valid in any media having
index of refraction n. In a vacuum (n = 1), the n can be dropped, A = Ao, and the
Planck equation is most often seen this way. If in air (nair "" 1.00028), this
correction is usually ignored except for extreme low-uncertainty applications.
It is more convenient for optical measurement work to use micrometers (!-lm)
or nanometers (nm) for wavelength to get W/m
2
sr!-lm or W/m
2
srnm. The
radiation constants then become
C, = 3.74177107 x 10
8
W!-lm
4
/m
2
(wavelength in !-lm),
C, = 3.741771 07 x 10
20
Wnm
4
/m
2
(wavelength in nm),
C2 = 14387.752 !-lmK = 1.4387752 x 10
7
nmK.
The two curves in Fig. 4.1 show the form of Planck's equation as a function
of wavelength, with temperature as a parameter. They are strongly peaked, with
the form governed by the A - 5 term for wavelengths longer than the peak and by
the exponential term for shorter wavelengths. Note that only a limited range of
temperatures can be shown on a single linear plot such as these, as the ordinates
are highly nonlinear, varying over many orders of magnitude. The plots also
show a dashed line, the locus of the wavelength of peak spectral radiance, having
a characteristic hyperbolic shape. As will be seen, these dashed lines represent
the Wien displacement law.
If the spectral radiance curves are plotted on log-log axes as in Fig. 4.2,
several interesting things are seen. First, this form allows for a wide range of
temperatures and wavelengths on a single plot. Second, the locus of the
wavelength of peak radiance is a straight line on a logarithmic plot, indicating
hyperbolic behavior. Finally, note that all of the curves have an identical shape
when the logs are plotted. Since the shape of the curve is independent of
temperature, one could construct a nomogram by tracing a single curve and the
straight line locus of maxima onto a transparent sheet and use it as an overlay.
Slide the overlay along the straight line to display the blackbody radiation curve
for any temperature.
4.2.2 Wien displacement law
The Wien displacement law describes this line, the locus of the peak wavelength
of the blackbody curve as a function of temperature. It is determined by taking
the derivative of OJ... with respect to wavelength (in the medium) and setting it to
zero to find the maximum. The result is
nAmax T= 2.897 7686(51) x 10-
3
m-K (1.7 x 1O--<i). (4.7)
A rough but useful approximation for the peak wavelength (in setting n = 1)
IS
Table 4.1 lists the peak wavelengths for several common sources.
4.OE+07
1
\1
E 3.0E+07
I I
i :::1=
0.5 1.5
WAVELENGTH (micrometers)
(a)
(b)
2
(4.8)
Figure 4.1 Spectral radiance of blackbody radiation for (a) high temperatures and (b)
lower temperatures.
88
Chapter 4
10'
''- I / II / ,/'6_.A
0
l'-_..L I
7 I' II Ii I V L,;()
I: 1/ II / :/
:J j / / II' V
II I: V .,1 / i
10-
4

0.1 \. 0.5 1.0 5 10 50 io
WOYll,nvth A.l'm
Figure 4.2 Log-log plot of blackbody spectral radiance as a function of wavelength and
temperature.
Table 4.1 Peak wavelengths for several common sources.
Source
Temperature Peak wavelength
(K) (flm)
Sun

Tungsten lamp

Typical hot IR source

Typical IR test source

Room temperature

Liquid nitrogen 77

To find the value of LA at the peak wavelength, solve the blackbody equation
using the peak wavelength (in the medium):
(4.9)
where cr' = 1.286 x 10-
11
W/m
2
K\lm. The radiance at the peak wavelength varies
as the fifth power of the temperature.
4.2.3 Stefan-Boltzmann law
To determine the total radiance of blackbody radiation, integrate the Planck
equation over all wavelengths. The result is the Stefan-Boltzmann law,
(4.10)
where cr is the Stefan-Boltzmann constant, 2n
S
ej15c
2
h
3
=5.670xlO-
8
W/m
2
K4 .
The Stefan-Boltzmann law is most often seen in terms of radiant exitance with
n = 1:
(4.11)
Insert room temperature (300 K) into the Stefan-Boltzmann equation to find
that one square meter of a black surface emits ",,460 W. Therefore the walls,
ceiling, and floor of a typical 6 x 6 x 2.5-m classroom have a total surface area of
132 m
2
and emit over 60 kW into the room. For the moment, we ponder why we
don't bake under this onslaught.
4.2.4 Laws in photons
It is frequently advantageous to work with photons rather than watts, particularly
in cases where light levels are low. Thus, a photon-based analysis may be helpful
for applications in which photomultiplier tubes provide photon counting or CCDs
are employed to image a faint target. The energy associated with a photon, in
joules, is given by:
c
Q =hv=h-
A '
o
(4.12)
where h is again Planck's constant. For energy in units of electron-volts (eV),
divide by the electronic charge q (1.602 x 10-
19
C): At a (vacuum) wavelength
Ao of I ~ m it takes 5.034 X 10
18
photons per second to equal one watt, and more
at longer wavelengths. Conversely, a single photon at a (vacuum) wavelength Ao
of 1 m has an energy of 1.986 x 10-
19
joules (W's), or 1.239 eV, and has
proportionally more energy at shorter wavelengths .
More accurate values for the constants that appear in this section are given in Appendix B; higher
accuracy is sometimes required.
90 Chapter 4
The Planck equation rewritten for photons (n = 1) is
(4.13)
where Clq = 2nc = 1.883 x 10
9
mls. For wavelength in flm, Clq becomes 1.883 x
10
27
m
2
/flms. Thus, Lq'A is expressed in photons per second per area per unit
wavelength.
Plots of spectral photon radiance are shown in Figs. 4.3 and 4.4. The curves
appear similar to the previous radiance curves, but the range of ordinate values is
not so extreme. The equation is subjectively the same; the exponential term is
identical but the wavelength in the denominator is only raised to the fourth
power.
1.0E+26
E
=>
:i; B.OE+25
N
E
.:,
~
6.0E+25
::::.-
w
U
~
4.0E+25
~
0:::
Z
0
I-
2.0E+25
0
I
a..
O.OE+OO
0 0.5 1.5
(a)
2
2. OE+ 21 . - . . . , . . . . . . , - - , - - . , . , . . - , - - , . . . . . . . . . , . . . . , - - , - - ~ - , - - - - , - - - - , - - ~ ~ - , - - , . . - , - . . . , . . . . . . , - - , - - ~
E 1.8E+21 t';''f ;
:;; 1.6E+21
N
E l.4E+ 21 ,:.', C'+-C !'-., . ...;. . .,.t.""'._c.c'i.: .. +-.: ....
~ 1.2E+21 f' ;:;+/';;;\;; .... ;;.'l .....;:: +:; :
w
u 1.0E+21
~
8.0E+20
~
0::
6.0E+20
z
0
I-
4.0E+20
0
I
c...
2.0E+20
O.OE+OO
0 5 10 15 20
(b)
25 30
Figure 4.3 Photon spectral radiance versus wavelength for (a) high temperatures and (b)
lower temperatures.
....
Generation of Optical Radiation
91
/

/

v/
V

1 I /

I
I /
/'

1'.1::"
I
-"","

I'r-
r .... i'"

i
II
I
V
1/
V"
p;
r.::

I
I
I
II /
.....
I L
:fl
I
//
/
v ....
'-0.

I
I
_L J I /
11
I
I
/
/

I
i
. I

I
I
I
I
......
/1 I
/1
,,'"
-.
['-.i't"
I
If

. i"
:
I
.L
I
/'"
t"-l .. r-
I I
1/
1
I
Ii
V

I
1
I
/

,
I
;
L
II
/
"
,
\
.:.1
,
j I
I
Ii / .1.
4 0.80.81.0. t 4 10 l o 60 80
0..2 0. 100 "'0..1
Figure 4.4 Log-log plot of spectral photon radiance as a function of wavelength and
temperature.
The Wien displacement law for photons is derived as described earlier; the
result is
Aq, max T = 3 .6696986( 62) x 1 0-
3
m . K .
(4.14)
If the wavelength is expressed in micrometers, a useful approximation is
A maxT "" 3700 IlmK .
q ,
(4.15)
The photon spectral radiance at the peak wavelength is
(4.16)
where O'q = 2.101 X lOll s-l
m
-
2
K-4/l
m
-l.
The total photon radiance is the integral of the Planck equation for photons:
(4.17)
92 Chapter 4
where Oq = 1.520 x 10
15
S-l m-
2
K-
3
. At a temperature of 300 K, one square meter
of a black surface emits about 4 x 10
22
photons per second. Therefore there are
more than 5 x 10
24
photons per second being emitted by the walls, ceiling, and
floor in our typical classroom.
4.2.5 Rayleigh-Jeans law
There are two common approximations to the Planck equation. The Rayleigh-
Jeans equation for blackbody radiation was derived independently in 1900 by
Rayleigh and in 1905 by Jeans from the thermodynamic principle of
equipartition. It successfully predicted the shape of the spectral curve at long
wavelengths, but was clearly inappropriate at short wavelengths. It predicted
infinite energy at A = 0, and the integral did not converge, giving rise to the
labeling of this equation as the "ultraviolet catastrophe." Assume that n = 1 in Eq.
(4.6) and begin the derivation using the Planck equation for spectral radiance:
The exponential term can be expanded:
c, /AT -1 ~ (c
z
/ AT)z (c
z
/ AT)3
e - + + + + ... .
AT 2! 3!
If c2iAT < < 1 (corresponding to a large value of An, drop all of the higher-order
terms. Then,
Rearranging terms and substituting into the Planck equation, the result becomes:
(4. 18)
This is the Rayleigh-Jeans law.
This expression is valid with less than 1 % error if AT > 0.778 moK (i.e., long
wavelengths and/or high temperatures). This approximation is not particularly
useful, even for far-infrared work, as less than 0.1% of the output ofa blackbody
is located at AT values larger than 0.8 mK, and the wavelength where it becomes
valid is some 250x the peak wavelength. This means for sunlight (6000 K) and
room temperature (300 K), the minimum wavelengths for which the
approximation is valid are 130 and 2600 ~ m respectively.
4.2.6 Wien approximation
To derive the Wien approximation, begin with the Planck equation (n = 1).
Assume that e(c,f"A.T> 1. Then, the 1 can be dropped and the equation
becomes:
(4.19)
The Wien approximation is valid with less than 1% error if AT < 3000 J.lffi"K
(short wavelengths and/or low temperatures). It is quite useful for a great deal of
radiometric work as it is valid for blackbody radiation at all wavelengths shorter
than the peak. Figures 4.5 and 4.6 depict the curves and the ranges of validity for
the two approximations.
4.2.7 More on the Planck equation
A career can be spent fiddling with the Planck equation, presenting it for various
frequencies, wavenumbers, etc. We will keep it simple and just do some
normalization. Divide both sides of the Planck equation by r:
LA. c
j
rs = 1t(AT)5 e
C
2
1A.T -1 .
0.01 L - _ ~ _ ~ ~ ~ ~ ~ _ - - - - - - - . . . L J - _ ~ ~ ~ ~ ~ ~ .
1E3 1E4
WAVELENGTH-TEMPERATURE (umK)
1E5
Figure 4.5 Wien and Rayleigh-Jeans blackbody approximation curves.
(4.20)
94 Chapter 4
z
Q 0.05 f-!
I-

o

z
o
i=
o
-:----:--- --:--:--: -
, , , ,
-0. 05 _+_' . ..;-I---7--7-;-;-';--:-;--{-1
U.
-0.1
1E3 1E4 1E5
WAVELENGTH*TEMPERATURE (um-K)
1E6
Figure 4.6 Errors associated with Wien and Rayleigh-Jeans approximations.
The result is an expression which is now a function of a new variable,
CA 1). The result of this normalization is a single curve as shown in Fig. 4.7 for
Planckian radiation for any wavelength and temperature.
The Planck function can also be normalized by dividing by L
Amax
such that
the resulting curve peaks at unity. This curve is denoted asj('AD. An additional
very useful curve shows the cumulative (integrated) radiance from 0 up to 'AT
divided by the total radiance at temperature T. This cumulative curve is labeled
F('AD. The defining equations for these two functions are:
and
feAT)= =L"CAT)
O"maJ) cr' T5
1t
" "

F('AT) = - -,,- 0 __ _
JL" ('AT)d'A T4

(4.21)
(4.22)
5E-12
_ I ___ !... _ _ _ _ _ _ 1__ _ _ _ 1 __
_ L_ _ _L_ _ _ L__ _ _ J __
1 I I I
4 E-12 t--- ---;- ' -- -- .,-----r---,---t----;---.,------,-: -- -- -,---\
--T--T -- T--T -- -- ,- -,- - ,- -, --
- - T - -T -- T--T - - -- , - - ,- - ,-- , --
< -- - T -- -t ---r -- -i- - - --1 --"1- -
1
--""1 --
tii 3 E-12 t---------4----I-----'-\:-f------+- + -- -- ..... ---- ..... - ---..... -; -- -- -4 -; ------j
"
Z __ .1. _ _ 1. __ ..1 __ -.1 _ ___ --1 _ _ j _ _ -l ___ .J __
< __ L __ l __ _ _ ' _ _
o 2E-12
$
-. - I - - I - - -I - - -I - -
- - I - - "l - -I - - --I - -
1E-12
- - -; ---1- ----1 -- -1 --
- -4 ----1-- - --1----1 --
'"-- --1 ----1 - --1 -- --1 ---

o -- --
o 5000 10000 15000
WAVELENGTH x TEMPERATURE (um-K)
Figure 4.7 Normalization of the Planck curve.
Since these definitions are ratios, exitance M can readily be substituted for
radiance L. These two functions are graphed in Fig. 4.8.
To use these curves to determine radiance in a narrow wavelength interval
< O.05A
c
), first select T for the blackbody radiation and the desired center
wavelength Aco Determine the radiance at the peak using the equation for
Finally, use the functionj{U) from the graph and the wavelength interval to
arrive at the result
A, +fY..12
f (AT)dA = . (4.23)
\ .. - 6'1..12
If the wavelength interval is large, typically greater than O.05x the center
wavelength, use the other function F(AT) to determine the radiance in a finite
wavelength interval. Again select T and the two desired wavelengths, AJ and
A2. From the graph, read F(\ T) and F(A2T) and compute the total radiance using
the Stefan-Boltzmann law. The result is
A,
f LA (AT)dA = cr T4 [F(A2T) - F(A)T)] A2 > AI.
A, 1t
(4.24)
Figure 4.8 also shows the corresponding curves for photons, h(AT) and
Fq(An. The defining equations are
96
and
w
=:l
-'
~
w
>
~
w
0:::
Chapter 4
0.01 LL-_____ - - - _ ~ ~ ~ ~ ~ _ _ _ _ _ __ _ _ _ _ _ _ _ ~ ~ . . . . . . . . . . . . . ,
1E3 1E4
Figure 4.8 Curves of f(A) and F(A) for watts and photons.
1E5
(4.25)
(4.26)
The application of the hand Fq photon curves is identical to the curves for
energy.
In some applications, it is desirable to maximize the radiation contrast
between a target and background of similar temperature. What wavelength might
one choose for this task? The problem occurs regularly in the infrared where both
the target and the background radiate near 300 K. Take the second derivative
ciLt!(dTaA), and set it to zero. The result is:
"'contrastT = 2411 /lm . K .
(4.27)
This equation implies that the best "visibility," or contrast with the
background, occurs at a wavelength somewhat shorter than the peak wavelength,
at "'contrast = 0.832Amax. Since the result is on the short wavelength side of the
peak, the Wi en approximation is valid and it simplifies the calculus considerably.
Thus, if your target and background temperature were 305 K, the peak
wavelength is 9.50!lm and the wavelength of maximum radiation contrast A contrast
is 7.9 !lm. The wavelength for maximum photon contrast is 2898 IlmK, the same
as the peak wavelength for energy.
We often need to know how the spectral radiance LA. changes with
temperature; this can be determined by differentiating the Planck function with
respect to temperature. The result, shown in differential form, is
M.. xe
X
!1T
4= eX -IT'
(4.28)
he e
where x=-- = _ 2 .
AkT 'AT
We define 2 == xe
x
For a small change in temperature !1T such that the
eX -1
change in x is also small, the change in L'J... with temperature at any wavelength is
(4.29)
If eX is significantly greater than unity ('AT <3000 !lmK, the Wien
approximation), then 2 = x and M'J.../L = x!1TIT. Inspection of the blackbody
curves and this equation shows that 2 approaches infinity as the wavelength
approaches zero, and 2 approaches unity at longer wavelengths. Figure 4.9 shows
2 plotted against 'AT; the curve is identical for watts and photons. The region
where the Wien approximation is valid is 'AT < 3000 K (2 = 5).
4.2.8 Kirchhoff's law
Blackbody radiation exists in any closed cavity at thermal equilibrium. As such,
it is idealized, because whenever a hole is made into the cavity to allow radiation
to exit, the conditions for blackbody radiation are altered. As noted in Chapter 3,
emissivity is defined as the ratio of the radiance of an object to the radiance of a
blackbody at the same temperature. The directional spectral emittance of a
specific sample, e('A;8,<), is
e(A;8,<) = 4 (8,<) .
~
(4.30)
98

(J)
c
0
+-'
0
.c
a.
L..
0
(J)
t:
(IS

N
__ -__ -__ ___ __
10
: :::::: : :!=
::i:
------- ----'--- --'-
_______ ----1...-- __ , _
=1: ::::::::!:::: :
t j-:::::::t::: ::t: : :ttt
" , , ".
" , , ". _I .J ________ , ______ -' ___ .1..-'.1...
" , , '"
" , , " .
" , , ". _ ,_ -' _________ , ______ -'. ___ 1...-'_ ....
" , , '"
" , , '"
" , , ".
,. . . '"
" , I '"
___________ ... ____ __ 0 .... _________ , ______ .... ____ ........ _ ....
'" , . ,.,
". , , '"
, '" , . '"
, ". , , '"
, '" , , '"
, '" , , '"
, '" I I '"
, .,' , , '"
, .,' , , '"
___ ..o. __ '- _ ... _.J. l _,_ _ _ _ _ _,_ -' ________ ,_ _ _ _ _ _ -' _ _ _ .1.. -'.'-
, , , , '" " " , , '" - - _ .. _ - _ .. _ .. - -0_ 0. _ _ _ _ _ _ ___ .... _ _ _ _1_ _1_ .... ________ .1 _ _ _ _ _ .... _ _ _Oo .. _ ..
:::Lt:j:j:j:i:t :::: :!: L::::+:: ::j:: : :tj:t
, , , , '" , '" , , ,,'
, , , , '" "'" , , ," ---"---""-"-"-"-'-T -------,,----,...- __ , __ ,_ ., ___ ______ , ______ ., ____ ,... .. _,...
, , , , '" ,,',' , , ,',
, , , , ,,' """ , , '"
, , , , '" '" " , , '" ----'---""-"--'-"-'-T -------,,----,...--,.- -,- ., _________ , ______ ., ____ ..... _ ..
, , , , ", """ , , '"
, , , , '" """ , , '"
, , , , '" '" '" , , '"
, , , , '" '" '" , , ,,'
-- --i- --i- -;- -:-: -:- t --- ----: ----i- --t --:- -:- -:-: , -- ---. --:-- -- -- -:- - -
, , "" '" , , '" , , '"
, , "" '" , , '" , , '"
, , "" ,,' , , ,', ,,',
, , "" '" , , '" ""
, , "" '" , , '" ""
1
1E2 1E3 1E4
Figure 4.9 Parameter Z as a function of AT.
1E5
Chapter 4
This equation differs slightly from Eq. (3.23), as it includes the directional
component.
Total emissivity is the integral of spectral directional emissivity over all
angles and wavelengths, also weighted by the Planck function, which introduces
a temperature dependence if the spectral emissivity is not uniform (that is, gray):
JJJ(9,<\>;A)L
ABB
sin 9cos 9d9d<\>dA

J / n)T4
(4.31 )
Since (A;9,<\ is not a derivative (per unit wavelength interval) quantity, it
makes no sense to integrate it alone:
:#: J(9,<\>;A)dA. (4.32)
A
Kirchhoff's law was stated in Eq. (3.24):
a=. (4.33)
To illustrate, Fig. 4.10 shows a small body within a large isothermal
enclosure. At equilibrium, in the absence of other sources or sinks, the source and
enclosure must be the same temperature. The power absorbed by the small body
(by the definition of absorptance) is a = a<l>;. The power emitted by the body
(by the definition of emittance) is <l>e = <1>88. At equilibrium, the power emitted
-
by the body to the enclosure must equal the power absorbed by the body from the
enclosure:
a<l> i = E BB
As the incident and emitted power are the same (no other sources or sinks), the
conclusion is a = E.
This seems too simple; is it really true? No! Look at the following equation
for the equilibrium temperature of a flat plate in space, facing the sun and
insulated on the back side. It is a simple statement of conservation of power,
equating radiant incidence (absorbed) to radiant exitance (emitted):
(4.34)
Here Eo is the solar constant (1368 W/m2). The units on both sides of the
equation are W 1m
2
Solving for T, we find that it is a function of the ratio (alE).
But Kirchhoff's law says that a = E, and these terms therefore cancel. The
equilibrium temperature is therefore a function only of Eo. The implication is that
a white (allegedly reflective) car and a black (supposedly absorptive) car have the
same equilibrium temperature after sitting out in the sun all day. People from
Arizona and Florida know better than that, and often buy white cars to keep
cooler in the summer! So what's wrong?
The answer is that Kirchhoffs law does not apply in all situations. Table 4.2
indicates the applicability in terms of spectral and directional conditions.
Specifically, a(A.;9,<\ = E(A.;9,<\; the absorptance equals the emittance at a single
wavelength in a single direction. After integrating over wavelength and
geometry, test for the stated restrictions before applying Kirchhoff's law. The
Figure 4.10 Illustration of Kirchhoff's law.
1
[Reprinted with permission of author from
Optical Radiation Measurement series, Vol. 1, F. Grum and R. J . Becherer, Radiometry, p.
98 (1979).]
100 Chapter 4
simple equation above attempts to relate the absorbed energy from the sun, which
is concentrated in the spectral range between 0.3 and 3 ~ m to the thermal
radiation located in the wavelength range between 3 and 30 ~ m . White and black
paints have completely different reflectances in the solar region, but their
emittances at 300 K are very similar.
The thermal radiative properties of any material can be given in terms of its
temperature T and its spectral directional emittance e(A;8,<j. The spectral
dependency is of primary importance, and the directional properties and
temperature coefficients are usually of lesser concern.
The blackbody equation is highly nonlinear with both wavelength and
temperature, and is thus not particularly tractable. Fortunately, there are many
computational and visual aids to help. The Infrared Handbook gives several
calculator programs for now-obsolete HP and TI calculators, and some early
calculators were available with plug-in cards to do blackbody calculations. The
latest incarnation, The Infrared and Electro-Optical Systems Handbook,
substitutes an extensive set of programs in BASIC to do the computations. There
have been several slide rules (remember them?) that do calculations on
blackbody radiation. The GE Radiation Calculator is a plastic rule that
occasionally surfaces at a reasonable price. Cardboard knock-offs are often given
away by vendors. For a little more money, Electro-Optical Industries at one time
sold a high-quality metal rule. One side of my venerable GE calculator is shown
in Fig. 4.11.
Back when computers were scarce, tables generated by mainframes were
commonly used to do precise blackbody calculations. Several tables may still be
found in musty libraries.
2
,J ,4
Today, simple computer programs can easily be written. Spreadsheets are
sufficiently powerful to do these calculations with relative ease and provide
superior graphics as well. Tools like Mathcad, MATLAB, and Mathematica also
work well.
Figure 4.11 The venerable GE radiation calculator-J. M. Palmer's own!
-
Table 4.2 Summary of absorptance-emittance relations.
1
[Adapted with permission of
author from Optical Radiation Measurement series, Vol. 1, F. Grum and R. J. Becherer,
Radiometry, p. 115 (1979).]
Directional spectral
Required conditions:
None, other than thermal equilibrium
Directional total
a(A;9,,1) = (A;9,, 1)
a(9,, 1) = (9,<1>,1)
(1) Spectral distribution of incident energy proportional to blackbody at T, or
(2) a(A;9,, 1) = (A;9,, 1) independent of wavelength
Hemispherical spectral
a(A, 1) = (A, 1)
(1) Incident radiation independent of angle, or
(2) a(A;9,, T) = (A;9,, 1) independent of angle
Hemispherical total
a(1) = (1)
(I) Incident energy independent of angle and spectral distribution proportional to
blackbody at T, or
(2) Incident energy independent of angle and a(A;9,, 1)
= (A;9,, T) independent of wavelength, or
(3) Incident energy at each angle has spectral distribution proportional to
blackbody at T and a(A;9,, 1)
= (A;9,, 1) independent of angle, or
(4) a(A;9,, 1) = (A;9,,T) independent of angle and wavelength
102 Chapter 4
Spectral emittance was defined as the ratio between the spectral radiance of
an object and the spectral radiance of a blackbody at the same wavelength and
temperature. In general, it is also a function of angles 8 and <1>. Using this
definition, any isothermal object has a spectral radiance given by
L (8 th) = [(A;8,<1C
1
]( 1 )
'J... ,,+, 1[1.,5 ec,lAT -1 '
(4.35)
and if (1.,;8,<1 and T are known, so is the spectral radiance. As (1.,) increases
towards unity, blackbody radiation is approached. Blackbody radiation
simulators attempt to do just this, and many are commercially available. Since
the spectral radiant exitance can be described by this simple equation, blackbody
radiation simulators are most often used as standard sources of optical radiation
in the calibration laboratory.
4.3 Emitter Types and Properties
4.3.1 Metals
An important factor in the behavior of a material is its electrical conductivity.
Metals have a high conductivity, meaning large quantities of free electrons are
available to interact directly with the radiation field.
Since metals are generally good reflectors (a direct consequence of their high
conductivity), they are poor emitters (conservation of energy). The emittance is a
slowly varying function of wavelength, as shown by a simple empirical equation
given by Hagen and Rubens:
5
resistivity
"" constant A (A > 2 !lm)
(4.36)
This equation, which is generally valid at wavelengths longer than 2 !lm, shows
that increasing resistivity (decreasing conductivity) increases emittance, and that
emittance decreases at longer wavelengths. Figure 4.12 gives examples.
The directional properties can be derived from Maxwell's equations. The
radiation is highly polarized at angles off normal, as shown in Fig. 4.13. Note,
however, that the total radiation, the sum of both polarizations, is quite constant
with angle (i.e., Lambertian) to within a few percent out to nearly 60 deg from
speCUlar.
4.3.2 Dielectrics
Dielectrics and gases have much lower conductivity than metals. Their electrons
are more tightly bound to their parent nuclei and require specific atomic
interactions with the radiation field. This implies that dielectrics tend to radiate in
specific, fairly well-defined spectral regions, and not elsewhere.
--
,...
:t:
,..
.;;;
...
'ji
..
-
c::

'j
...

.;:
..
..
Q,
en
1.0r--------------------.
0.8
0.6
-Copper
_. - Graphite
-+-Iron
- -- Silver
-0- Tungsten
-0- Aluminum
10
Wayelength, microns
Figure 4.12 Spectral emissivity for some metals,
6
20
103
The emitting properties of dielectrics are closely associated with the complex
index of refraction n + iK where K, the extinction coefficient, is in tum related to
the absorption coefficient a', as discussed in Chapter 3. As noted in that chapter,
the product of the absorption coefficient and the thickness of a material is the
optical thickness 'to. A material is considered optically thin when 'to < 0.1
(transmission high) and optically thick when 'to> 2 (transmission low). Optically
thin materials approach transparency and have low emittance; for optically thick
materials, the normal emittance is (1 - reflectance) and depends on the index of
refraction as determined by the Fresnel equation at normal incidence as
illustrated in Fig. 4.14. Emittance at other angles also comes from the Fresnel
equations and is polarized.
4.3.3 Gases
Gases are optically thin over wide wavelength ranges and may be transparent
over long paths. Therefore, their emittance is essentially zero at these
wavelengths. However, there are specific spectral regions where absorption, and
therefore emission, occur. Each species has its own absorption characteristics,
correlated with its atomic and molecular structure and energy levels. These
characteristics take the form of a series of spectral lines at regular locations in the
spectrum. They are occasionally seen as discrete lines, but more often as a series
of overlapping lines called bands.
104
LU
0.8

0.6
...J
<t:
z
o 0.4

LU
0:::
15 0.2

o 15 30 45 60 75 90
ANGLE FROM SPECULAR
Figure 4.13 Directional emittance as a function of angle from specular.
Chapter 4
The important species identified with our own atmosphere include "fixed"
gases O
2
, N
2
, and CO
2
Those species classified as variable include water, ozone,
and methane (H
2
0, 0
3
, and CH
4
.) It must be stressed that the energy absorbed by
a gas is dependent upon the concentration and absorbing characteristics of the
gas, and the energy emitted by the gas depends upon the temperature and spectral
emittance of the gas. The processes of absorption and emission are not
independent, but occur simultaneously.
4.4 Practical Sources of Radiant Energy
4.4.1 Two major categories
Table 4.3 presents a division of sources into the two categories of thermal and
luminescent. Several of these sources will be discussed in this section. Natural
sources of radiation, which may be thermal or luminescent, will also be
discussed.
w
o
Z 0.9 --_ .

W
....J

::2:
0.8 - -
I
I
0:: 0.7 _ ... __ .-
o
--- .. ------"
Z
I
0.6 L-_ _ L--__ L--__ L--_ _ L--_ _ L-_--'
1 2 4
INDEX of REFRACTION
Figure 4.14 Normal emittance versus index of refraction.
-
Table 4.3 Practical radiation sources.
Thermal sources
Natural radiators
Direct radiators
Earth (longwave)
Sun
Stars
Atmospheric gases
Reflective natural objects
Earth (shortwave)
Moon, planets
Skylight, clouds zodiacal light
Atmospheric scatter
Artificial radiators
Incandescent substances
Tungsten lamps
Hot metals
Silicon carbide rods
Ceramic tubes
Nernst glower
Electrical transmission
equipment
Machinery
Personnel
Carbon arc
Optical elements (lenses,
mirrors, etc.)
Open coal, wood, oil fires
4.4.2 Thermal sources
Luminescent sources
Natural sources
Glow worms, fireflies
Aurora, airglow
Artificial sources
Semiconductor devices
LEDs,IREDs
Diode lasers
Phosphors
CRT tube (TV, computer
display, oscilloscope)
Electroluminescent panels
Lasers
Gas (He-Ne, CO
2
, lon, N
2
)
Solid-state (Ruby, Y AG)
Metal vapor (He-Cd)
Liquid (dye, cyanide, chelate)
Chemical (HF, DF)
Metal-vapor lamps
Sodium vapor lamp
Mercury vapor lamp
Fluorescent lamp
Gas discharge lamps
Plasma display
Neon lamp
Glow discharge
Xenon arc and flashtube
Hot gases
Welsbach mantle
Bunsen burner
Exhaust gases
4.4.2.1 Tungsten and tungsten-halogen lamps
The most common household source of optical radiation (other than fire) has
been the tungsten-filament lamp. Equation (4.35) may be used to describe the
spectral radiance of a tungsten lamp; it is the equation for blackbody radiation
with the emittance term included. Tungsten's spectral emittance has been
extensively studied. The most frequently used data are from DeVos, shown in
106 Chapter 4
Fig. 4.15. He measured the emittance of a flat ribbon of tungsten, but the results
are also applicable to a round wire. Most lamps are made by drawing the
tungsten into a round wire and then tightly coiling it, enhancing the emittance by
making a partial cavity. If this tight coil is then further loosely coiled, emissivity
is enhanced further.
Tungsten lamps are designed to operate at a nominal voltage and current. The
design compromise is between light and lifetime. If more light is needed, voltage
(or current) may be increased, but not for very long. Figure 4.16 shows light
(lumens), efficiency (lumens/watts), and lifetime as a function of operating
voltage.
Tungsten-filament lamps decay due to evaporation of the filament, leaving
brown deposits on the inside of the lamp envelope. An uneven rate of
evaporation creates "hot spots" which cause the lamp's overall rate of decay to
increase. Additionally, filaments may crystallize and become brittle particularly
when the lamp is operated on dc, mechanically weakening the filament and
making it susceptible to breakage from mechanical or thermal shock.
Tungsten-halogen lamps are better suited than tungsten for most applications,
as they either eliminate or delay the onset of both decay mechanisms. The
addition of a halogen such as bromine or iodine creates a regenerative cycle, in
which the evaporated tungsten combines with the halogen rather than plating on
the envelope. A hot (minimum 250
0
C) envelope, usually of fused silica, is
required. The resulting halide compound decomposes at a rate proportional to
temperature. This decomposition occurs preferentially at the "hot spot," causing
tungsten to plate back onto the filament.
.50
.40
>-
~ 3 5
>
iii
~
~ 3 0
.25
.20
WAVELENGTH
Figure 4.15 Spectral emittance of tungsten. (Reprinted from Ref. 7 with permission from
Elsevier.)
180
160
LUMENS
EFFICIENCY
(lUMENS/WATTS)
........ ",
u.,Il.fO . . . ~
.... .......... .... WATTS
::J . . ~
~ ..... ~
ffi 120 ~ ~
~ /-
~ ~ ~
~ l 0 0 ~ - - - - - - - - - - - - - - - - - - - - - - ~ ~ - - - - - - - - - - - - - - - - - - - -
~
~ 80
. ~
20;-__ ____ =-__ ~ __ __ 7 - __ ____ ~ ~ ____ __ __
m ~ ~ 00 ~ ~ D m
120
% NORMAL VOLTS
125
107
Figure 4.16 Tungsten lamp characteristics as a function of percent normal operating
voltage.
The advantages of the tungsten-halogen lamp are many. They have a
significantly longer lifetime-at least a factor of ten-than the tungsten filament
lamp alone for a given filament temperature, and they can be operated at much
higher temperatures. They exhibit superior long-term stability because the
envelopes don't darken. Because they require a high envelope temperature, they
are physically of much smaller size than the simple tungsten lamp for a given
wattage.
There are disadvantages to this approach as well. The hot quartz envelope
requires greater care, both in fixture design and in handling. Fingerprints are
verboten; in calibration applications, fingerprints and grease on the lamp can
severely impair measurement accuracy. If the temperature of the envelope is not
hot enough, the regenerative cycle fails, and the lamp will fail sooner. The lamps
have high internal pressure, and the increased ultraviolet content requires that
users take safety measures. Finally, their purchase price is considerably higher
than tungsten lamps alone.
It should be noted that at the time of publication (2009), incandescent
tungsten lamps have fallen out of favor, largely due to energy efficiency
considerations in comparison to alternatives such as compact fluorescent bulbs
and light-emitting diode (LED) lighting. In fact, sales of incandescent light bulbs
108 Chapter 4
are being phased out in the European Union, Australia, the United States, and
other nations.
4.4.2.2 Other metallic sources
Metallic sources have some general properties worth noting. Since they are fair-
to-good conductors of electricity, they can be heated to the operating temperature
by simply applying an electric current. The power dissipated is given by PRo The
resistance of metals increases with increasing temperature, meaning that a current
surge can be expected when a device is turned on. This is particularly noticeable
with tungsten, where the starting current required is an order of magnitude
greater than the operating current. Other materials with lower operating
temperatures show less of this effect.
Refractory (high-temperature) metals that may find occasional application
include tantalum and molybdenum. They have lower melting points than
tungsten and higher vapor pressures. They are more reflective than tungsten,
hence at normal operating temperatures their emittance is lower. They also are
more susceptible to oxidation. Other than the above, they work well.
Noble metals, those that remain in pure form, include platinum, palladium,
and iridium. They also have lower melting points and lower emittances than
tungsten, but are resistant to oxidation and will therefore operate in air.
Certain alloys have been developed for use in heating applications. The most
often-used alloy is 80Ni-20Cr (one trademark is Nichrome), which is the element
in toasters, space heaters, and other devices. The resistance ofNi-Cr is essentially
constant with temperature, simplifying the power supply design. To season it, run
it at an elevated temperature in air to maximize the oxidation, which enhances the
emittance.
4.4.2.3 Dielectric thermal sources
Nominal characteristics of dielectric sources include a high emittance in those
spectral regions where they are optically thick, and a negative resistance versus
temperature characteristic. The resistance at room temperature is usually
sufficiently high that extreme voltages would be required to generate sufficient
heat to get them started. Large amounts of ballast are needed to ensure an overall
positive resistance versus temperature characteristic. Alternatively, they can be
indirectly heated by a conventional resistive heater.
The globar is a rod fabricated from silicon carbide (SiC). It is typically 5 mm
in diameter and 50 mm long and requires about 200 W to reach the operating
temperature of 1000 to 1500 K. Its emittance is approximately 0.75 out to a
wavelength of 15 Ilm. This source can be directly heated, but requires an
auxiliary ballast to overcome the slight negative resistance characteristics of the
SiC. Establishing electrical contact is difficult and is usually accomplished with
water-cooled silver electrodes.
The Nernst glower is a small (l-mm diameter by 10 mm long) ceramic rod
comprising mixed oxides of Zr, Y, Ce, Th, Be and the like. It dissipates about
200 Wand operates at 1500 to 2000 K. The emittance is about 0.75 out to a
useful wavelength of 30 ~ m It is rather fragile and has a lifetime of 200 to 1000
hours. Its resistance at room temperature is extremely high, requiring an auxiliary
platinum heater coil to get it started. Nernst glowers are most often found in
infrared spectrometers where their shape is ideally suited to image onto a narrow
slit.
A popular low-cost IR source consists of an aluminum oxide (Ah03) ceramic
tube heated with a coaxial nichrome wire. This source is often used in
inexpensive infrared spectrophotometers and operates at about 1200 K with an
emittance of about 0.8. The power source is a simple line-operated transformer.
The Welsbach mantle is a woven fabric mesh impregnated with refractory
oxides such as thorium oxide. It is heated with propane or white gas to a
temperature up to 2400 K. The emittance is high for A > 1 0 ~ m This source is
commonly known as the "Coleman" lantern.
The carbon arc is a valuable source for announcing grand openings, detecting
hostile aircraft, simulating solar radiation, etc. The radiation comes from the
plasma-heated carbon at its sublimation temperature of 3800 K. By placing
refractory oxides in the carbon rod, higher temperatures commensurate with the
evaporating points of these oxides can be achieved, in the range 5000 to 8000 K.
The high brightness comes from the small (about 10-mm diameter) size of the
carbons. The disadvantages are lack of stability, mechanical issues involving
continuous feeding of carbons, the power required (usually a noisy, smelly,
unregulated motor generator) and the need for ventilation (hydrogen cyanide is
generated).
4.4.2.4 Optical elements
The components in our optical systems also radiate in spectral regions where they
are not perfectly transparent or reflective. Mirrors are coated with thin metallic
films for high (p > 0.9) reflectance. Their emittance is consequently low in the
infrared ( = 0.02 is a reasonable approximation). Even a surface with this low
emittance still generates copious quantities of photons per second unless cooled.
Refractive elements are transparent at wavelengths where they are normally
used; the emittance is low. At wavelengths where the material is absorptive,
however, the emittance is high. If our detectors are responsive to wavelengths
where the windows and lenses emit, they will respond to such radiation,
contributing nothing but noise.
Choppers are used to impart modulation to a beam. One must analyze what is
being seen when a chopper is used. If the chopper is reflective, you must
determine what the system sees in reflection. If the chopper is nonreflective
(black), then it has self emission governed by its own temperature and emittance.
4.4.2.5 Miscellaneous thermal sources
Flames emit large quantities of radiation, principally in the infrared. There are
two components to radiation from flames. The first is from particulate matter,
110 Chapter 4
like the hot soot from burning hydrocarbons with insufficient oxygen. Examples
include the oxyacetylene torch before the oxygen is turned on, and flames from
solid rocket motors. These particles are effectively "black," emitting as small
blackbody radiators. The other component is radiation from the gases, including
impurities. For example, the yellow color from a Bunsen burner is due to atomic
emission from sodium in the supply gas.
Exhaust gases emit radiation in spectral regions where they are optically
thick. The important gases (from a radiometric standpoint) are the hot
combustion products CO, C
2
0, and H
2
0. Particulates, "black" if their size is
> 30 !lm, are often present, particularly from diesel exhaust and muzzle flash.
If a gas both emits and absorbs at the same wavelength, how can we see it?
Our initial reaction is that the absorption lines from these gases block the
emission from being seen. In fact this is not the case. A phenomenon known as
"line reversal" takes place wherein hot gases have slightly different spectral
profiles than their unheated counterparts, broader because of increased
temperature and pressure. Absorption will occur at the center wavelength of the
profile, but the hot gas may be observed in two wavelength bands bordering the
center.
You and I make good infrared sources. Our temperature is about 300 K and
our emittance is nearly unity for all wavelengths longer than about 5 !lm. We are
all black in the thermal infrared, regardless of skin pigmentation in the visible.
Appliances and conveyances also make interesting sources. Cars, trucks,
trains, tanks, and aircraft all have different temperatures than their surroundings
when at work. Even such little things as insulators on a power transmission line
get hot when leakage occurs, rendering them observable targets in the infrared.
Table 4.4 lists the total (integrated over wavelength) directional (normal
incidence) emittance for several materials at the indicated temperatures (in
degrees kelvin).
4.4.3 Luminescent sources
4.4.3.1 General principles
The term "luminescence" refers to emission of light from materials that receive
energy from various sources. Table 4.5 details some of the types of luminescence
and the sources of energy for each.
As we have seen, temperatures must be high to get significant radiation from
thermal sources in the visible and ultraviolet regions of the spectrum.
Luminescence can occur at much lower temperatures, with light generated as
atoms decay to a lower energy level from a higher level. In this process, photon
emission occurs at a specific wavelength, but emission lines may be spectrally
broadened by Doppler (Gaussian) and pressure (Lorentzian) effects. In practice,
this means that the emission occurs in a narrow band, rather than being confined
to a single wavelength.
Table 4.4 Emittances of some common materials.
Material Temperature (K) Emittance
Metals
Aluminum, polished 300-900 0.04-0.03
Aluminum, heavily anodized 373 0.84
Brass, polished 373 0.1
Chromium 310-l370 0.08-0.4
Copper, polished 300 0.02
Copper, black oxidized 300 0.8
Gold 300 0.02
Nickel, polished 300 0.05
Nickel, oxidized 500-1500 0.4-0.8
Silver 300-800 0.01-0.03
Stainless steel, polished 300 0.16
Stainless steel, oxidized 1000 0.85
Steel 300 0.08
Tungsten 300-3000 0.03-0.39
Dielectric Materials
Alumina 300-1000 0.96-0.6
Brick, red 300 0.93
Carbon, lampblack 300 0.95
Concrete, rough 300 0.93
Glass, plate 300 0.94
Ice 270 0.97
Magnesium oxide 400-750 0.69-0.55
Oil, 0.001 in. thick on nickel 300 0.27
Oil, thick on nickel 300 0.82
Paint, oil 350 0.94
Paint, laqcuer 350 0.97
Skin, human 300 0.98
Snow 270 0.82
Soil, dry 300 0.92
Soil, wet 300 0.95
Water 273-373 0.96
At low pressures (below 100 mbar), the Doppler effect predominates. The
shape of the curve is Gaussian, and a typical equation takes the form:
0.47S _ln2[(V:;)']
k(v)=--e , (4.37)
a
where k(v) is an extinction coefficient at frequency v, S is a line strength, a is an
absorption coefficient related to the halfwidth of the line, and v 0 is the line
112 Chapter 4
Table 4.5 Luminescence processes and associated energy sources.
Process
Bioluminescence
Cathodoluminescence
Chemoluminescence
Electroluminescence
Photoluminescence
Radioluminescence
Sonoluminescence
Thermoluminescence
Triboluminescence
Energy source
Organic chemical reactions
Electron beam
Inorganic chemical reactions
Electric field
Optical radiation
Charged particles
Acoustic energy
Mild heating below that required for
incandescence
Energy from friction or pressure
center frequency. This shape is characteristic of low-pressure discharge sources
and atmospheric absorption and emission lines at elevations greater than 80 km
in the stratosphere. The key element in determining this profile is a long mean-
free path.
At higher pressures, where the mean-free path is short, the lines are further
broadened by collisions among the molecules. This pressure broadening gives a
Lorentzian shape of the form
(4.38)
The Lorentzian shape characterizes the emission of high-pressure discharge
sources and atmospheric absorption and emission lines at elevations less than
50 km in the troposphere.
At intermediate pressures, the line shape takes on a hybrid shape, a
convolution of the Gaussian and the Lorentzian curves. Figure 4.17 compares the
shapes of the two curves.
The comparatively narrow, Doppler-broadened linewidth of low-pressure
discharge is exploited in a number of sources. Because of the narrow width, there
is little power in each of the lines; sources are limited to low-power applications
involving sharp lines, such as wavelength calibration of spectrometers, excitation
of fluorescence spectra, and interferometry. The means of excitation include
alternating current and radio frequencies. In the latter case, unknown gases
placed in a sealed tube are excited by an external radio frequency (RF) field and
produce emission spectra which allow their identification. Common gases
identified in this manner include mercury, helium, neon, sodium, potassium, zinc,
cadmium, and cesium. These gases have further application in colorful signs and
illumination. An example of the emission lines of mercury and argon appears in
Fig. 4. 18.
-
3
2
Doppler
-t 0
Figure 4.17 Doppler and Lorentzian line shapes.
113
The high-pressure regime is entered when the pressure is increased to the
point where a low-resistance arc is observed. The line shape becomes Lorentzian,
with broad wings extending to either side of the line center. These wings
combine to provide a continuum. Arc lamps are efficient and extremely bright.
Common types include mercury, xenon, mercury-xenon, and sodium. They can
be modulated and flashed, as in the photographic strobe lamp. They are useful for
illumination, ultraviolet exposure for photoresist applications, projectors,
searchlights, and solar radiation simulators. Figure 4.19 depicts the spectral
irradiance of xenon arc sources.
5000
4000

3000

of 2000
c(
1000
200 300 400 500 600 700 800 900 1000
Wavelength (nm)
Figure 4.18 Emission lines for Hg 600 nm) and Ar (> 600 nm).8 (Copyright held by
Ocean Optics, Inc. Reproduced with permission.)
100
"'"e
c:
'1'e
3:
..s
e
It)
c:i 10
t(
w
U
Z
<C
Ci
<C
a::
g;
200 500 1000
10
200
1500
WAVELENGTH (nm)

300 400 500 600 700 800
2000 2400
Figure 4.19 Spectral irradiance of xenon arc lamps. (Permission to use granted by Newport Corporation. All rights reserved.)
-"
-"

Q

(b
...,
.:..
l ...... "",...... nv' " . ' 'S4"," 0 ow" " .... " . , d O 0 ' n "" _ . _ . " "eo _'_, _.""" ........... ....... .. ... <.-.... ...." .,= .".... '''-.. "... .." . _ ..
Phosphors exhibit luminescent mechanisms and are highly temperature
dependent. They represent a variety of colors and persistences (decay time
constants greater than 20 ms). They are used in fluorescent lamps, cathode ray
tubes (CRTs), x-ray and gamma-ray screens, UV detectors, charged particle
detectors, and flat-panel displays, to name a few applications. Figure 4.20 depicts
the spectral characteristics of several phosphors.
4.4.3.2 Fluorescent lamps
The modern fluorescent lamp is a combination of a moderate-pressure (about 1
atm) mercury discharge lamp and a phosphor. The phosphor is excited by
radiation from the mercury discharge, predominantly from the 254-nm Hg line.
The phosphor, coated on the inside of the cylindrical glass envelope, reradiates at
longer wavelengths in the visible spectrum. Many phosphors are available with
varying color characteristics. Several types of daylight simulation are available,
from the rather harsh but inexpensive "daylight" to more subdued "warm white."
Several pale colors are available for special effects, and a plant growth phosphor
is common for hothouse use. This latter phosphor emits in the red and blue
portions of the visible spectrum, and lacks green, since plants reject green light
by reflection. The mercury lines may be prominent, particularly when the more
transparent phosphors are used.
General characteristics of fluorescent lamps include a long lifetime and
relatively high illumination efficiency, as their output is concentrated in the
visible. They are a low-radiance extended source, principally suited for
illumination. They operate with dc, ac, and RF excitation. The discharge needs
help in getting started, as the mercury must be vaporized. Be careful in disposal
of spent lamps, as both the mercury and some of the phosphors are toxic.
4 ~ ~ ~ - - - - ~ - - ~ - - - - - r - - - - r - - - - r - - - - ~ - - - - - - - ,
WAVELENGTH - NANOMETERS
Figure 4.20 Phosphor spectral characteristics.
9
(Reprinted by permission of Pearson
Education, Inc., Upper Saddle River, NJ.)
116 Chapter 4
The compact fluorescent lamp (CFL) finds increasing use in today' s
environment, particularly as a replacement for tungsten incandescent bulbs
common in household use. The efficiency of these lamps is high, primarily due to
triphosphors and high-frequency solid-state electronic ballast. They are available
in several colors, and their liftetime extends to 12,000 hours. Compact
fluorescents are more expensive to buy than incandescent lightbulbs, but their
longer lifetime has rendered them a cheaper choice than the latter for many uses.
Figure 4.21 shows the components ofa compact fluorescent lamp.
I I
II I I
II
Glass
Discharge
Tube
Tube
Retention
Clip
Mounting
Plate
j.
Electronic
Ballast
Components
Polycarbonate
Housing
Edison

Base
Figure 4.21 Components of the comract fluorescent lamp. (Reprinted with permission
from the IESNA Ughting Handbook, gt Edition, by the Illuminating Engineering SOciety of
North America.)10
4.4.3.3 Electroluminescent sources
Electroluminescent sources are activated by an ac field, typically at 60 Hz. These
lamps are constructed by layering the phosphor on a metallic electrode, covered
by a semitransparent electrode and glass. They draw very little power and are
quite efficient. Their radiance is low, and their geometrical characteristics are
nearly Lambertian. The color selection is limited to available phosphors. Large
areas and complex shapes are possible. Their output depends upon voltage,
temperature, and excitation frequency. They are stable and exhibit long lifetimes.
They find uses for nightlights, emergency signs, backlighting for portable LCD
computer screens, and calibration sources for x-ray film. Other forms of
excitation are available, including biological (fireflies and glowworms), chemical
(plastic-encapsulated light sticks), and atomic (tritium-excited watch dials), for
example.
4.4.3.4 LED sources
A forward-biased p-n junction is an excellent emitter of optical radiation. The
emission comes from recombination radiation at the band edge excited by the
high current density. The wavelength region of emission is narrow,
approximately 5% of the center wavelength. These light-emitting diodes (LEDs)
are small, consume little power, and have lifetimes in excess of 10,000 hours.
The radiance from an LED is high due to the small emitting area, but the total
radiated power is low. They may be modulated to frequencies up to 1 GHz. They
are useful for displays and back-illuminating liquid crystal displays. They are
also used extensively in airline illumination for exit signs, floor and overhead
lighting, and cockpit lighting.
Traffic and roadway lighting applications are increasingly benefiting from
LED technology. Many older traffic lights are being replaced by LED sources in
a matrix configuration, which allows the light to continue to function although
several individual LEDs burn out. LEDs are also seen increasingly in
architectural displays. Table 4.6 lists several materials used for LED lighting
with their corresponding wavelengths or wavelength bands.
OLEDs (organic light-emitting diodes) are a special class of LED. They are
created from thin organic emitters sandwiched between a transparent anode and a
metallic cathode. Enhancement of efficiency and color control are provided by
doping emissive layers with highly fluorescent molecules. The structure of
organic layers and the choice of anode and cathode materials are designed to
maximize the recombination process in the emissive layer, thus maximizing light
output.
Due to the emitters' thin width (a few micrometers), OLED display designs
can replace LCDs in many applications, including television monitors and
automotive displays. Figure 4.22 shows the OLED physical structure.
118 Chapter 4
Table 4.6 LED materials and corresponding wavelengths.
Material Wavelength (nm)
GaN 350
SiC 465
GaP 565
ZnTe 620
CdTe 855
InP 1000
ZnCdTe 530 - 830
GaAsP 550 - 900
InGaAsP 550 - 3500
GaAlAsSb 580 - 1800
AlGaAs 620 - 900
InGaAs 850 - 3150
InPAs 910 - 3150
4.4.3.5 Lasers
The LASER (light amplification by stimulated emission of radiation), was first
demonstrated in 1960, and represented a radical departure from conventional
sources known at the time. The combination of stimulated (rather than
spontaneous) emission to provide gain and a frequency-selective feedback
mechanism gives the laser its unique properties. The stimulated emission results
from a population inversion, where there are more electrons at a higher energy
level than at a lower energy level. This population inversion can be generated
using one of several different energy sources, including electrical, optical,
magnetic, chemical, or nuclear. The frequency-selective feedback mechanism
EI.EC1RON
TRANSPORT
lAYER
ORGANIC
EMITTERS
UGHTOUTPUT
2 -IOYDC
ANODE
GlASS
SUBSTRATE
Figure 4.22 OLED structure.
11
(Reproduced by permission of Silicon Chip.)
....
most often takes the form of a resonant cavity with mirrors. For frequencies
having overall gain greater than unity, oscillation is possible.
Lasers differ from incoherent sources in many ways. Since the amplification
process depends upon strict phase relationships, the radiation produced is
coherent, i.e., the waves are in phase with each other. Since we are dealing with
atomic transitions and a resonant cavity, the output is very nearly
monochromatic, with a very narrow bandwidth. Because of the requirements
placed on the resonant cavity, the beam diameter and the beam divergence angle
are both small.
Lasers come in all forms, shapes, and sizes. Some generate unmodulated
light [continuous wave (CW)] with power levels from milliwatt to megawatt and
above, while others generate pulses. Pulse widths can vary from several
femtoseconds (10-
15
s) to several milliseconds with repetition rates varying from
gigahertz to millihertz. Size can be as small as a TO-18 transistor to as large as a
full-size laboratory.
A wide variety of materials have been found useful as gain media. Gas lasers
include He-Ne and CO
2
Examples of ion lasers are argon and krypton. Solid-
state lasers are exemplified by ruby and Nd:YAG. A semiconductor laser is
similar to an LED but incorporates the resonant cavity needed for laser operation
within the semiconductor structure. Tunable lasers are available using organic
dyes and special solid-state crystals such as alexandrite. Erbium can be doped
into a fiber to provide in situ gain.
4.4.4 Natural sources
4.4.4.1 Sunlight
As Table 4.3 shows, natural sources appear in both thermal and luminescent
categories.
The most prominent natural source of radiation, indeed the most important, is
our sun. The best estimate for the solar constant is 1368 W/m
2
with an
uncertainty and drift of about 0.2%. Since the orbit of the earth around the sun is
elliptical, there is an additional diurnal variation of 3.5%, with the maximum
experienced in January.
Things are somewhat different at the surface of the earth. Our atmosphere
selectively attenuates by two primary means: scattering by molecules (Rayleigh
scatter) and aerosols (Mie scatter), and absorption by molecules (H
2
0, CO
2
, and
0
3
) . The term "air mass" is frequently used to indicate how much atmosphere the
solar radiation is traversing; it is approximately equal to the secant of the solar
zenith angle. t Figure 4.23 shows extraterrestrial solar spectral irradiance in the
visible and near infrared, along with transmitted solar spectral irradiance for
several air masses.
t At or near sea level. As the altitude increases significantly, and air pressure goes down, the air
mass decreases significantly.
120 Chapter 4
2Z!iO
i;
:I.
2000 0, 0.35 cm(NTP)
III
e W 2cm
!
1150
f3 0.1
a 1.:\
...
1500 I..>
z
ct
0
ct
12!i0
II:
!!:
AIR MASS
...J
I
ct 1000
II:
l-
I..>
...
1!iO ..
CII
...J
ct
::i 500
a:
0
z
l-
I..>
...
II:
Q
0
0. 29 O. lI 1.0 UI 2.0 2.5
WAVEI.ENGTH
C"ml
Figure 4.23 Spectral characteristics of direct sunlight,12
4.4.4.2 Skylight, planetary, and astronomical sources
Skylight is of some importance, as it contributes to both heating and illumination.
It arises from both Rayleigh and Mie scattering. The sky appears blue during
daytime because of the strong A -4 wavelength dependence of Rayleigh scattering
at shorter wavelengths. In the morning and evening, with airmass high, longer-
wavelength oranges and reds appear. Figure 4.24 shows some typical spectral
irradiance curves. Note the dependence on solar zenith angle (z) and albedo
(surface reflectance.)
Il.1O O.J5 Q,4Q Q,4a NO _ _ .... I/oIIQ HQ
.....,..,.," ..... '
Figure 4.24 Irradiance from skylight,13 [Reprinted from Daylight and its Spectrum, S. T.
Henderson, p. 113 (1970).].
,',
-
Astronomers and planetary scientists are greatly interested in stellar and
planetary radiation. Optical radiation from the planets in our solar system has two
components: reflected solar radiation appearing like a 5700 K blackbody, and
emitted thermal radiation corresponding to the temperature of the planet.
Radiation from stellar bodies is even weaker, requiring ever larger telescopes to
capture enough radiation to be measured. Astronomers use a number of systems
to classify stars. Visual magnitude is a logarithmic system, wherein a difference
of 1 magnitude equals an irradiance ratio of 2.512. It has been determined that a
zero-magnitude star produces an illuminance of2.65 x 1O-
6
Imlm
2
. The sun has a
magnitude of about -27 (the scale is upside down.)
Progressing to even weaker diffuse sources, we encounter the aurora, the
airglow, and zodiacal light. Both the aurora and airglow arise from solar
electrons and protons exciting atomic species in the upper regions of our
atmosphere. The earth's magnetic field dictates that most of this activity occurs
in polar regions. The aurora is transient and related to intense solar activity, while
the airglow is a constant background. Both are characterized by spectral line
structure. Zodiacal light arises from solar radiation scattered by dust in the
ecliptic plane.
4.4.4.3 Application: energy balance of the earth
The temperature of the earth is determined by the combination of the energy
absorbed from the sun and the energy radiated to space. The globally averaged
earth albedo weighted by the solar spectrum is 35%. This means we absorb 65%
of the incident irradiance, or about 900 W/m
2
To maintain thermal equilibrium,
we must emit about 225 W/m2 to space (recall that the ratio of the total area to
the projected area of a sphere is four). This irradiance corresponds to a blackbody
equivalent temperature of about 250 K. The radiation from the earth does not
appear as blackbody radiation because of the spectral characteristics of the
atmosphere. In the atmospheric windows (regions of relative transparency), the
radiation comes from the surface. In those regions where the atmosphere is
opaque, the radiation comes from the cold upper atmosphere. Figure 4.25 shows
the spectral radiance of a 272 K blackbody for reference (dashed curve) and the
outward radiance of the earth as seen from space (solid curve). The dotted curve
is the radiance seen when looking upward from the surface of the earth, which
includes upwelling earth radiance and whose primary atmospheric component is
emission at these wavelengths.
4.5 Radiation Source Selection Criteria
Several factors must be considered when selecting a source of optical radiation
for experimental or design use. Table 4.7 details selection criteria in the form of
questions, the first eight of which were posed by A. G. Worthing in 1937.
122
i
j
\J
Chapter 4
f\
: .

5 10 15 20 25
WAVELENGTH (I'm)
Figure 4.25 Earth radiance in the thermal infrared region. Dashed curve: blackbody at 272
K; solid curve: radiance of earth from space; dotted curve: radiance seen from earth's
surface, looking Up.14
Table 4.7 Source selection criteria.
Does it supply energy at such a rate or in such an amount as to make
measurements possible?
Does it yield an irradiation that is generally constant or that may be varied with
time as desired?
Is it reproducible?
Does it yield irradiations of the desired magnitudes over the areas of the desired
extent?
Has it the desired spectral distribution?
Has it the necessary operating life?
Has it sufficient ruggedness for the proposed problem?
Is it sufficiently easy to obtain and replace, or is its purchase price/construction
cost reasonable?
What is its physical size? Will it fit into the allocated space?
What are its operational limitations (cooling, shock & vibration, etc.)?
Where do you plug it in?
"'----,
4.6 Source Safety Considerations
Table 4.8 provides tips on the safe handling oflight sources.
Table 4.8 Necessary considerations for safe handling of light sources.
Ultraviolet eye damage is possible with UV sources; shield the lamp and wear
protective goggles.
There is an explosion hazard from high-pressure lamps. Use an approved
housing, clean quartz envelopes thoroughly, handle lamps with gloves, and
wear eye protection.
A fire hazard exists with powerful lamps; keep flammable materials away from
them.
Ozone is generated by ultraviolet lamps; use ozone-free lamps (envelopes that do
not transmit UV) or provide proper ventilation.
There are electrical hazards from lamp power supplies; take normal precautions,
particularly with high-voltage starters for arc sources and capacitor banks for
pulsed lasers.
Careful handling and disposal of CFLs is a must! See http://www.energystar.gov
for more information.
4.7 Summary of Some Key Concepts
All objects above 0 K radiate according to Planck's law modified by the
directional spectral emittance term, Eq. (4.35):
For blackbody radiation (as opposed to blackbody radiation simulator and
graybodies), the emittance term can be neglected.
Metals are poor emitters.
Opaque dielectrics are good emitters.
Transparent dielectrics are poor emitters.
Gases and glasses radiate only in absorption bands and are transparent
elsewhere.
For Further Reading
F. E. Carlson and C. N. Clark, "Light sources for optical devices," Chapter 2 in
Applied Optics and Optical Engineering, Vol. 1, R. Kingslake, Ed. ,
Academic Press, New York (1975).
124 Chapter 4
J. E. Eby and R. E. Levin, "Incoherent light sources," Chapter 1 in Applied
Optics and Optical Engineering, Vol. 7, R. R. Shannon and J. C. Wyant,
Eds., Academic Press, New York (1975).
D. Kryskowski and G. H. Suits, "Natural sources," Chapter 3 in Sources of
Radiation, G. 1. Zissis, Ed., Vol. 1 of The Infrared & Electro-Optical
Systems Handbook, SPIE Press, Bellingham, Washington (1993).
A. J. La Rocca, "Artificial sources," Chapter 2 in Sources of Radiation, G. 1.
Zissis, Ed., Vol. 1 of The Infrared & Electro-Optical Systems Handbook,
SPIE Press, Bellingham, Washington (1993).
A. J. La Rocca, "Artificial sources," Chapter 10 in Handbook of Optics, Vol. 1,
Part 4, Optical Sources, McGraw Hill, New York (1995).
H. Z. Malacara and A. Morales, "Light sources," Chapter 5 in Geometrical and
Instrumental Optics, Academic Press, New York (1988).
J. B. Murdoch, Illumination Engineering-from Edison 's Lamp to the Laser,
Macmillan, New York (1985).
M. S. Rea, Ed., Lighting Handbook: Reference and Application, 9
th
Ed.,
Illuminating Engineering Society of North America, New York (2000).
J. C. Richmond and F. E. Nicodemus, "Blackbodies, blackbody radiation, and
temperature scales," Chapter 12 in NBS Self-Study Manual on Optical
Radiation Measurements, Part 1, U.S. Government (1985).
R. Siegel and J. R. Howell, Thermal Radiation Heat Transfer, Hemisphere
Publishing Corp., New York (1981).
W. T. Silfvast, "Lasers," Chapter 11 in Handbook of Optics, Vol. 1, Part 4,
Optical Sources, McGraw Hill, New York (1995).
R. H. Weissman, "Light emitting diodes," Chapter 12 in Handbook of Optics,
Vol. 1, Part 4, Optical Sources, McGraw Hill, New York (1995).
References
1. F. Grum and R. J. Becherer, Radiometry, Vol. 1 in Optical Radiation
Measurements, F. Grum, Ed., Academic Press, New York (1979).
2. M. A. Bramson, Infrared: A Handbookfor Applications, Plenum, New York
(1966).
3. M. Pivivonsky and M. Nagel, Tables of Blackbody Radiation Functions,
4. M. Czerny and A. Walther, Tables of the Fractional Function of the Planck
Radiation Law, Springer-Verlag, Berlin (1961).
5. E. Hagen and H. Rubens, "Uber die beziehung des reflexions und
emissionsvermogens der metalle zu irhrem elektrischen leitvermogen," Ann.
d. Physik, 4(11), pp. 873-901 (1903).
I
-'
6. J. A. Jamieson et aI, Infrared Physics and Engineering, McGraw-Hill, New
York (1963).
7. 1. C. DeVos, "A new determination of the emissivity of Tungsten ribbon,"
Physica, 20, p. 690 (1954).
8. "HG-2 mercury argon calibration source," Ocean Optics, Inc.
http://www.oceanoptics.com/productsihg I.asp.
9. I. Csorba, Image Tubes, Howard W. Sams & Co, Inc., Indianapolis, (1985).
10. IES Lighting Handbook, 9th Ed., Illuminating Engineering Society of North
America, New York (2000).
11. P. Smith, "OLED displays: better than plasma or LCD," Silicon Chip Mag.
179 (1 Aug. 2003). http://www.siliconchip.com.auicms/A_30650/article.html
12. S. L. Valley, Handbook of Geophysics and Space Environments, Office of
Aerospace Research, USAF (1965).
13. S. T. Henderson, Daylight and its Spectrum, Elsevier, New York (1970).
14. F. Kneizys, et aI., "Optical properties of the atmosphere," Air Force
Cambridge Research Laboratory (1972).
-
Chapter 5
Detectors of Optical Radiation
Section 5.7.2 contributed by Robert C. Schowengerdt
5.1 Introduction
Optical radiation detectors are transducers that transform optical radiant energy
into a different form of energy that is more readily measured. Electrical energy is
typically used for this purpose, as electrical measurement technologies are well
established. Both thermal and photon detectors convert incident optical energy
into electrical signals; in the thermal detector, the initial output takes the form of
heat before conversion. Either detector type may be a "point" or an "area"
detector. The former are single-element detectors, designed to respond to incident
energy. The latter are one- or two-dimensional arrays used particularly for
imaging, and include mechanisms to read out the signal on the array. Table 5.1
gives examples of photon and thermal detectors, while Table 5.2 lists differences
between them.
Point
Area
Thermal
Low detectivity
Table 5.1 Detector types.
Thermal
Bolometer
Thermocouple
Pyroelectric vidicon
Linear thermoelectric array
Microbolometer array
Photon (quantum)
Photodiode
Photoconductor
CCD array
CMOS array
UV photo diode array
Table 5.2 Some important differences.
Photon
Slow response time
Do not require cooling
High detectivity
Fast response time
Typically require cooling for
IR operation
127
128 Chapter 5
R(A.) R(A.)
R(A.) = constant

Wavelength A. Wavelength A.
Figure 5.1 Characteristic response curves of (a) photon and (b) thermal detectors.'
[Reprinted from Optical Radiation Measurement series, Vol. 1, F. Grum and R. J.
Becherer, Radiometry, p. 177 (1979).].
The parameters in Table 5.2 will be discussed in more detail later in this
chapter. Figure 5.1 illustrates the generic differences in spectral response
between photon and thermal detectors. Note that the thermal detector response is
wavelength independent, and that the photon detector response is a function of
several parameters including wavelength and quantum efficiency. A specific
example of the latter is the commonly used silicon detector, whose wavelength of
peak response is 950 nm and whose cutoff wavelength is near 1100 nm.
5.2 Definitions
Several terms commonly used to describe detector parameters are not often used
outside the field of optics. Recall that the term wavelength refers to the optical
regime below 100 flm. The term frequency is used to describe the direct current
(dc) to 10
12
Hz audio/radio regime. The following defintions will aid our study of
detectors.
Spectral responsivity 9\(A) is the ratio of the output of a detector or
radiometer to that of a monochromatic source of optical radiant power. The
detector ouput is typically a current (amperes, A) or a voltage (volts, V), while
the incoming optical quantity is power, measured in watts. It is also a function of
wavelength, in that it is measured at specific wavelengths, and is correctly
reported in AlW or V /W at a specific wavelength. It can also vary as a function
of detector temperature and input power level.
Responsivity 9\ is the ratio of the output of a detector or radiometer to the
incoming optical radiant power, integrated over the spectral range which is
common to the source and detector. The input and output signals are those
described above. Since this parameter is the result of the quotient of two integrals
as shown below, it is source dependent for nonmonochromatic or nontlat sources.
9t= fR(A)(A)dA
f(A)dA
129
(5.1 )
On occasion, one will also find responsivities related to other radiometric
quantities, such as irradiance responsivity 9t
E
in amps (or volts) per W/m
2
, or
radiance responsivity 9t
L
in amps (or volts) per W /m
2
sr. These responsivities are
more likely to characterize radiometric instruments than detectors alone.
Photon spectral responsivity 9tiA) is the ratio of the output of a detector or
radiometer to the monochromatic optical radiant photon flux incident upon the
detector or radiometer. The output is typically a current (A) or a voltage (V),
while the incoming optical quantity is photon flux (photons/s). Therefore
9tq(A) carries the units (A-s) or (Vs). This parameter is wavelength dependent for
most detectors and also can vary with temperature and input power level. Like
spectral responsivity, photon spectral responsivity is a function of wavelength.
Photon responsivity 9t
q
is the ratio of the output of a detector or radiometer
to the incoming photon flux, integrated over the spectral range which is common
to the source and detector. The input and output signals are those described
above. Since this parameter is the result of the quotient of two integrals, it is also
source dependent for nonmonochromatic or nonflat radiation sources.
The electrical signal that is output by a detector is governed by the electrical
characteristics of the detector (resistance, capacitance, etc.) as well as its
associated circuitry. A Bode plot is a convenient way to depict a system's
electrical frequency response, with gain or phase plotted on the ordinate and
frequency on the abscissa. For our purposes, voltage or current gain are often
depicted. Both axes are logarithmic, as seen from the generalized Bode plot in
Fig. 5.2. With the proper choice of axes, slopes associated with single time
constants (6 dB/octave, 20 dB/decade) plot as 45-deg lines, and the cut-on and/or
cutoff frequencies are at the intercepts of the straight-line asymptotes.
w
(J)
z

(J)
w
Ir
PlOT
0.1 '-_--'_--'---'----'----'--'-'-..!.-.L _ _
1 10
FREQUENCY
Figure 5.2 Bode plot.
100
130 Chapter 5
Time constant 't is the time required for a signal to achieve 63% of its final
output, given a step input. For a simple single-section resistance-capacitance
(RC) circuit, 't = RC. The rise time is the time the signal takes to rise from 10%
of its final value to 90%. For a single time-constant circuit, its value is
2.2't. Figure 5.3 illustrates this concept.
Cutoff frequency Ic is the frequency at which the voltage or current response
of a detector or circuit falls to IJ2, or 0.707x its dc or midband maximum value.
It is related to the time constant by Ic = 1/(27t't). In terms of power, it is the
frequency where the power drops to half of the dc or midband maximum value.
This is also called the 3-dB frequency.
The root-mean-square (rms) value of a quantity (voltage in this example) is
defined as
1 T 2
Vrms = - fV (t)dt,
To
(5.2)
where T is a single or integer multiple period for periodic waveforms. The rms
values discussed in this chapter may be values of voltages or currents and are
often referenced in descriptions of noise.
Signal is the component of the output voltage or current from a detector that
arises from a specific radiometric input. Signal is what you want to work with. It
is an rms current or voltage and is denoted Vs or is. The signal is the integral of
the product of the spectral responsivity and the radiometric input such that:

SIGNAL = fR(A)<l>(A)dA.
o
RISE TIME
I
__ L __
__ _ _ 1 __ _ _
I
I
I _
.1
I
--bl+
0.8 H-_ --: _:-_ -. -..
i ::: .. :-:.-: .. :. .. ..
-{/
I
- - - 1- -- .- - - - - T - - -
,
I
.. . ---l-- - --1---- - - +
0.5 1.5 2 2.5 3
TIME (time constants)
Figure 5.3 Signal output as a function of time constant.
(5.3)
.,
1
J
I
j
l
l
1
!
'-
I
1
....
Detectors of Optical Radiation 131
Noise is the component of the output voltage or current from a detector that
arises from random fluctuations in the detector circuit, in the incoming radiation,
or from a number of other sources discussed in this chapter. It is characterized by
an rms current or voltage and denoted Vn or in.
Signal-to-noise ratio (SNR) is the ratio of the rms signal current is to the rms
noise current in. Voltage may be substituted for current when calculating SNR,
which is dimensionless.
Impedance Z is the slope of the current versus the voltage curve of a device
at the designated operating point:
dV
Z=-(ohms) .
dI
(5.4)
Linearity is the proportionality of output to input for a detector or an
instrument. Linearity implies that the responsivity of a detector or instrument is
constant over a defined range of input power, irradiance, etc. Expressed
differently, it is the region in which the slope of a plot of output versus input is
equal to a constant.
5.3 Figures of Merit
Several figures of merit have been defined for optical radiation detectors. Among
the more common are the following:
Responsive quantum efficiency (RQE, 11) is the number independent output
events per incident photon. This definition eliminates gain terms that could mask
the fundamental detection mechanism. It is frequently expressed as electrons per
photon, which, in the case of a detector with internal gain (i.e., a photomultiplier
or an avalanche photodiode) gives misleading information. RQE is between 0
and 1 and is often simply called "quantum efficiency."
Detective quantum efficiency (DQE) is the ratio of the square of the output
SNR to the square of the input SNR. It is a measure of the SNR degradation
caused by the detector and has value between 0 and 1. DQE may be expressed
mathematically as
(5.5)
Noise-equivalent power (NEP) is the incoming signal, in watts, that produces
a signal-to-noise ratio of 1. It is therefore the ratio of the rms noise current (or
voltage) to the responsivity. Like responsivity 9\, this term results from
integration over wavelength.
132 Chapter 5
i i
NEP = ...!!.... = ...!!.... or
9\ is '
(5.6)
NEP= vn =<1> Vn.
9\ Vs
Note that NEP is a useful figure of merit only when the limiting noise is
inherent in the detector, and not due to the input signal.
Spectral NEP(A) is the NEP for a monochromatic single-wavelength input
signal. Again, this is not a derivative quantity, per-unit wavelength, but merely a
value at a specific wavelength.
Noise-equivalent photon flux NEq is the incoming signal in photons/s that
produces a signal-to-noise ratio of unity. It is therefore the ratio of the rms noise
current (or voltage) to the photon responsivity.
Detectivity D is the reciprocal of NEP, originally defined as a term because
"bigger is better." The unit is WI. It is conveniently thought of as the SNR for a
1-W input. Multiply D by the input power to get the SNR. Like responsivity 9\
and NEP above, this term is a result of integration over wavelength:
D=_l_= SNR.
NEP <l>input
(5.7)
Specific (or normalized) detectivity D* is pronounced "dee-star." Most
detectors display noise that is proportional to (A,;J)II2, where Ad is the detector
area and B is the noise bandwidth (to be defined later). D* permits a useful
comparison of detectors of different materials, unequal areas, and different noise
bandwidths. Its units are cm HZ
Il2
W-
I
Derived from responsivity 9\ and NEP,
this term is again a result of integration over wavelength. Its defining equations
are
r;4jj r;4jj v
D* = D[4;B = _,,_ fidd
D
_ = -"- fidd
D
----. cmHz
Il2
/W.
d NEP vn
(5.8)
In the expression above, signal and noise currents may replace signal and
noise voltages. To obtain the SNR from D*, multiply D* by the input power and
divide by (A,;J)II2.
This term is called blackbody D* if the incident power comes from a
blackbody radiation simulator. The notation used in this case is D*(T,!c,B), where
T is the blackbody radiation simulator temperature in degrees kelvin (500 K is
common), !c is the chopping frequency in Hz, and B is the effective noise
bandwidth, by convention, 1 Hz.
Spectral D*(A) is D* at a specific wavelength. It is defined as
D * (A) - _ ~ - - - _ A - - d B _ cmHz
l12
/W.
NEP(A)
(5.9)
A notation often seen is D*(Ap,fc,B), where Ap is the wavelength at the peak
spectral responsivity (Ilm),.fc is the chopping frequency (Hz), and B is the
effective noise bandwidth.
D**, pronounced "dee double star," is a further normalization of D* to a
1t steradian, hemispheric, projected-solid-angle field of view. The term applies
only to what we will later define as background-limited detectors. There is a
spectral D** as well. D** is defined as
.. . rn ..
D =D v-; =D sm81/2 '
(5.l0)
since n = 1tsin
2
8 1/2, where 8
112
is the system or detector' s half-angle field of
vIew.
Photon D* is the specific or normalized detectivity in terms of photons:
he .J AJ] i 112
D* =D*-=--...!... cmHz /(phot/s)
q 'I -. ,
I\, n In
(5.11)
where n-bar in the denominator is the photon flux in photons per second.
BLIP is an acronym for background-limited infrared photodetector, the
condition in which the limiting noise in a detector output arises from background
photons. A BLIP detector' s internal noise has been reduced to the point where it
is not significant. This ideal condition allows SNR to be easily calculated.
RA product is the product of the detector resistance and area and is a
constant for many materials. D* is proportional to (RAi
i2
for many photovoltaic
detectors; thus the RA product can be used as a figure of merit for material
compansons.
5.4 #N$O%&I*S@E-A
5.4.1 Introduction to noise concepts
Noise places a fundamental limit on the detection process. Even if all noises
intrinsic to the detector could be reduced to insignificance, the noise associated
with the random arrival of incident photons will still limit the amount of power
that can be detected. Noise is of three types:
(1) intrinsic (i.e., noises from physical processes in detectors, associated
preamplifiers, and signal processing circuits)
(2) anthropomorphic (man-made noises, i.e. from motors, radio, TV, etc.)
(3) environmental (i.e., lightning or ionospheric effects).
134 Chapter 5
Assuming that (2) and (3) can be minimized using good engineering
practices (shielding, grounding, proper component and circuit layout), we
concentrate on (1). Figure 5.4 is a plot of noise spectral density versus frequency
that includes several of the above noise mechanisms.
Noise comes in many forms and colors. The most prevalent model of noise is
white noise, having a power spectrum independent of frequency. Of course real
noise is never white, as that would imply infinite total power. Noise density
actually goes to zero above 10
12
Hz because of the finite mass of electrons. In
Fig. 5.4, the region at frequencies greater than about 50 Hz is white
(superimposed on spikes from fixed frequency sources). Pink noise has a power
spectrum that increases with decreasing frequency at a nominal rate of 3 dB per
octave, proportional to the inverse of the frequency (lif). It occurs at low
frequencies, below about 5 Hz. Noise increasing at a rate of 6 dB per octave
(proportional to lif2) at low frequencies is referred to as red (or brown) noise.
There are special names for relatively rare noises that increase with increasing
frequency. Blue noise has a power spectrum that increases 3 dB per octave with
increasing frequency (proportional to./), and purple noise has a power spectrum
that increases 6 dB per octave with increasing frequency (proportional to f\
10'
TEMP
CHANGE OF
CLASSES, WORK,
SHIFTS, ETC
~
POWER LINE
60 Hz
" "-
180 Hz
~
~ W H I T E NOISE "
10-8 10-6 1
FREQUENCY (Hz)
"
10'
10
4
TV
10'
Figure 5.4 Noise spectral density as a function of frequency. (Reproduced from Ref. 2
with permission of Wiley-Blackwell.)
j
I
1
i
i
1
j
1
J
J
j
I
J
!
j
I
J
p
Table 5.3 Gaussian noise distribution characteristics.
Peak-to-peak value
2 x rms (1cr)
4 x rms (2cr)
6 x rms (3cr)
6.6 x rms
8 xrms
10 x rms
12 x rms
Fraction of time
p-p e x e ~ d e d
0.32
0.046
0.0027
0.001 (1000 ppm)
60 ppm
0.6 ppm
2 x 10-
9
ppm
Most noise can be accurately modeled as white. Also, many forms of noise,
such as thermal noise within detectors, can be effectively modeled as Gaussian.
There is no necessary correlation between the white noise model and the
Gaussian noise model. The oscilloscope trace in Fig. 5.5 shows the characteristic
appearance of Gaussian noise.
White noise without a dc component has a zero average over time, and its
peak is infinite (though with zero probability). The preferred way to characterize
this noise is by its rms value denoted by cr, which is also called the standard
deviation. Table 5.3 shows peak-to-peak values of Gaussian noise, along with the
fraction of the time that the peak-to-peak (p-p) value is exceeded.
A good estimate of the rms amplitude of the noise can be obtained by
estimating the difference between the maximum and minimum amplitudes in an
oscilloscope trace and dividing this difference by six: rms amplitude = (peak
max
-
peak
min
)/6. This can also be done by looking at amplitudes in the data set.
Noise can be expressed as a voltage, a current, or a power, and is best
expressed as a spectral quantity if its magnitude depends upon frequency. The
expression for a mean square noise voltage (equivalent to a noise power) looks
like
Probability of
instantaneous value
e of voltage exceeding e 1
Time
Figure 5.5 Gaussian noise oscilloscope trace and accompanying probability distribution.
(Reprinted from Ref. 3 with permission from John Wiley & Sons, Inc.)
136 Chapter 5
Its units are volts squared. In the expression, Vi is the instantaneous voltage,
v
avg
is the average voltage, and T is the observation time. The rms noise
is vn = R . Current may also be used according to Ohm's law, v
2
= ji XR2.
If the various noises in a system have different origins, they may be
considered independent or uncorrelated. The noise powers then add algebraically,
while the noise voltages (or currents) add in quadrature:
(5.13)
If the noises are partially correlated, emanating from the same source, a
correlation coefficient must be included in the equation, and the resulting noise
will be greater. In any case, the total noise will not exceed the algebraic sum of
the noise voltages or currents.
5.4.2 Effective noise bandwidth
The effective noise bandwidth (ENB, but more often seen as B or !l.f), first
mentioned in the definition of D*, is defined by the equation
(5.14)
where G is the power gain and fa is the frequency where G is a maximum.
Spectrally flat white noise is often assumed, simplifying the equation to
B = -l-fG(f)df .
G(fo) 0
(5.15)
The ENB is an equivalent square-band bandwidth and differs from the
conventional 3-dB bandwidth, as illustrated in Fig. 5.6. It is realized by an
electrical filter operating in the audio- to low-radio-frequency range, 20 Hz to
several MHz, or possibly by the detector itself. For a typical single-section (one-
time constant) filter, the relationship between these two bandwidths is B = (nI2) x
.!3dB. A combination of white plus 1ifnoise increases B, whereas pure lifnoise or
a combination of white plus generation-recombination (G-R) noise decreases B.
These noises will be discussed in more detail later in this chapter.
For multipole electrical filters, the ENB and the 3-dB bandwidth converge as
the number of poles increases, and the filter response becomes more rectangular.
For a common two-pole low-pass filter with equal time constants (Bessel filter),
B = 1.22 X hdB. For other filter types, e.g. Butterworth or Chebyshev, the
pd
Frequency
Figure 5.6 Effective noise bandwidth. (Adapted from Ref. 4.)
relationship differs and must be derived via integration. If the noise is nonwhite,
integration must be performed. ENBs can also be specified in terms of time
constants. The common two-pole low-pass fitter with equal time constants
(Bessel fitter) has B = 1I(8't), and the simple single-pole fitter has B = 1/(4't).
Note that the ENB is always stated in frequency space at frequencies below
1 a
d2
Hz in the electronic realm. Do not confuse this bandwidth with the passband
of an optical fitter operating at frequencies greater than 10
12
Hz, whose center
and passband are specified in units of wavelength, nm or Ilm.
5.4.3 Catalog of most unpleasant noises
Most texts give the noise equations including the ENB term. The power spectral
densities for each noise will be given first here, and then the total noise
expression that results from integration over frequency will be presented.
5.4.3.1 Johnson noise
Johnson (Nyquist, thermal) noise arises from the random motion of carriers in
any electrical conductor. Both the amplitude and frequency distribution are
Gaussian. The power spectral density of Johnson noise is given by the equation
where
k = Boltzmann's constant, 1.380658 x 10-
23
JIK,
R = the resistance of the conductor (0.), and
T = the absolute temperature (K).
(5.16)
Note that since there is no explicit frequency term, Johnson noise is "white"
at least to 10d2 Hz. Integrating over frequency to obtain the mean-square noise
voltage:
138 Chapter 5
~ =4kTRB,
(5.17)
where B is the effective noise bandwidth, defined above. The rms Johnson noise
voltage is the square root of the mean square noise voltage, or
Vj = -J4kTRB .
(5.18)
As an example, a 10
6
Q resistor generates 4 Jl V rms for B = 10
3
Hz, T = 290
K. Since v
2
/R is power, the noise power per unit bandwidth is simply 4kT,
independent of resistance. Invoking Ohm's law, the mean-square Johnson noise
current is given by
.2 _ 4kTB
1.---.
} R
(5.19)
If the bandwidth is increased, it appears as if the noise voltage approaches
infinity. Not so; at extremely high frequencies, a quantum correction must be
applied. The energy per degree of freedom kT in the noise expressions must be
replaced:
hi
kT (hflkT)
e -1
(5.20)
Then,
"2 x
v. =4kTRB--,
} eX -1
(5.21)
where x = hjlkT.
It is easy to show that Eq. (5.21) reduces to Eq. (5.17) by applying the
approximation e
S
'"' 1 + s when s is small, that is, much less than 1. This
substitution is only necessary for frequencies greater than 10
12
Hz, so it is not
significant for most applications. This expression also indicates the linkage
between thermal noise and blackbody radiation.
In most practical applications, the noise bandwidth is established by an RC
circuit time constant. Under these circumstances, the mean-square Johnson noise
is
"2 ~ f S(f)dl
Vj = 0 1 + (21tjRC)2 '
(5.22)
F
which integrates to
2 kT
v. =-,or
J C
Vi =.JkT / C.
139
(5.23)
Note that the expressions in Eq. (5.23) are independent of resistance R, but
dependent upon circuit capacitance C. As an example, if T = 300 K and
C = I picofarad (pt), then vr 64 I! V rms. Using the identity q = CV, the noise can
also be expressed as 6.43 x 10-
17
coulombs (C), or about 400 electrons.
5.4.3.2 Shot noise
Shot noise appears due to the discrete nature of electronic charge and occurs
whenever current flows across a potential barrier. Its power spectral density is
Ss(f) = 2ql
dc
'
(5.24)
where q equals the charge on an electron, 1.60217733 x 10-
19
C, and Ide equals
the direct current flowing across the potential barrier.
The mean square shot noise current is
(5.25)
Like Johnson noise, shot noise is also spectrally flat ("white"). As an
example, if 2 x 10-
5
amps flow across a barrier, then the noise current in a 10
3
-Hz
bandwidth is 8 x 10-
11
amps, or 80 pA. If this noise current were to flow through a
10
5
Q resistance, then the noise voltage is 80 I! V rms. Contrary to some claims, a
potential barrier must exist in order for shot noise to be present, as it is in
photovoltaic detectors that will be discussed later in the chapter. In a purely
resistive conductor, the flow of electrons is highly correlated, not independent.
Shot noise follows Poisson statistics: the events all have the same amplitude
at low photon-arrival rates. At high rates, the amplitude distribution becomes
Gaussian.
5.4.3.3 11f noise
llf (contact, modulation, excess, flicker) noise is an empirical nOIse having
various sources. A typical power spectral density equation is
(5.26)
140 Chapter 5
where K = a "constant" for the particular technology/noise process, 1.25 < ex, < 4
(usually 2), and 0.8 < ~ < 3 (usually 1).
Note the spectral frequency dependence of this noise, which has been
demonstrated to occur at frequencies as low as 10-
5
Hz and below. This is a
particularly nasty noise, ultimately limiting dc and low-frequency detection and
amplification. This form of fluctuation is ubiquitous, found throughout nature. A
few of the manifestations of lifnoise are seen in Table 5.4.
The low-frequency form of lifnoise is often called "drift." Other names are
"pink noise" ~ = I), "red" or "brown noise" ~ = 2), "excess," "flicker," and
"contact." Causes in semiconductor detectors, for example, include nonohmic
contacts and surface impurities.
This noise is particularly insidious, as the noise power is constant in each
frequency decade. The total noise is the integral of the power spectral density
over frequency, which is proportional to In(fhig,//tow) for pure lifnoise. Evaluation
is easy for ac-coupled systems, but in direct-coupled systems, we must choose
some number other than zero for /tow. Unless a system has a particularly long
integration time, /tow of 0.1 Hz is probably sufficient. While integration of
"white" noise over time reduces white-noise effects, integration fails to decrease
lifnoise due to the decrease in/tow with increased integration time.
Table 5.4 Where does 1lfnoise occur?
Waves on a beach
Fluctuation in axon membrane
Earthquakes
Economic variables
Ecological time series
Self-organizing systems
Fluctuations in human heart rate
Photon counting
Most music (not Metallica!)
Frequency of rotation of earth ~ = 2)
Feedback controls in nuclear reactors
Base arrangement of DNA sequences
Traffic flow (both vehicular & network)
Motion of man standing on one foot
5.4.3.4 Generation-recombination noise
Generation-recombination (G-R) noise arises from fluctuations in the rate at
which charge carriers are generated and recombined in semiconductor devices. It
includes variations in the carrier lifetime. The expression for G-R noise can be
complicated, and the specific form depends upon the detector configuration. An
example of the power spectral density of G-R noise for an extrinsic
photoconductor (to be described later in this chapter) is
where
N = the mean number of carriers,
i = the average photocurrent due to N,
R = the resistance,
1:/ = the carrier lifetime, and
f= frequency.
(5.27)
Note that G-R noise is frequency dependent (nonwhite). Its Bode plot looks
like that of a low-pass filter.
5.4.3.5 Temperature fluctuation noise
Temperature fluctuation noise is found only in thermal detectors. This noise has
to do with microfluctuations in the temperature of thermal detectors and is the
limiting noise for this detector class. The equation is related to statistical noise in
blackbody radiation and is given by an equation of the form
where
K = thermal conductance (W / deg),
H = heat capacity (J/deg), and
k = Boltzmann's constant.
(5.28)
Thermal conductance and heat capacity define a thermal time constant 'tr =
HIK that limits the temperature fluctuations at higher frequencies (nonwhite).
The effect of this noise must be determined via the responsivity of the particular
detector.
5.4.3.6 Photon noise
Photon noise occurs because in a beam of light, photons do not propagate in an
orderly fashion. They obey Poisson statistics, giving the result that the mean
142 Chapter 5
square fluctuations in the photon arrival rate are equal to the arrival rate. The rms
photon noise equals the square root of the number of photons:
(5.29)
where n is the number of photons.
The assumption of independence breaks down at extremely high photon
arrival rates ("photon clumping") and requires a quantum correction:
(5.30)
There is still some controversy in the literature about the relationship
between photon noise and other noises. The controversy stems from the carrier
generation mechanism and its associated noise. If the generation of carriers in a
photodetector is a random process, independent of the randomness of the photon
arrival rate, then shot noise or G-R noise can exist independently from the noise
in the photon stream. Careful measurements have not yet resolved this dilemma,
but high-level measurements indicate that the noise is essentially that given by
the shot or G-R equations. The implication is that the generation process is
instantaneous, i.e., when a photon is absorbed, the associated carrier is generated
instantaneously.
5.4.3.7 Microphonic noise
Microphonic noise, acoustically generated, has two primary manifestations. The
most important is found in pyroelectric detectors, which are by nature
piezoelectric and act like microphones. A more subtle form occurs when
unanchored wiring is allowed to vibrate within a vacuum dewar, causing small
but noticeable changes in circuit capacitance.
5.4.3.8 Triboelectric noise
Triboelectric noise is a curious electrostatic noise originating from charges that
are built up in dielectrics, typically in coaxial cables used to prevent the entry of
external noises. It is particularly nasty when the cable is permitted to flex, and the
charge may take some time to dissipate.
5.4.3.9 CCO noises
CCD noises are those specific to the charge transport and readout mechanisms in
charge-coupled device (CCD) arrays, including charge-transfer efficiency (CTE)
variations, readout noise, KTC reset noise, and others. CCDs will be discussed in
more detail in Sec. 5.7.
5.4.3.10 Amplifier noise
Amplifier noise is a combination of the noises typically found in the low-level
circuits that directly couple and signal-condition the detector. There are both shot
noise components due to the flow of current through p-n junctions in diodes and
transistors and Johnson noise components because of finite resistances in the
circuit. There may also be "popcorn" noise, found in some integrated circuit
amplifiers due to instabilities in the forward current gain. Amplifier noise is nasty
in that it is a form of lifnoise with a typical exponent/. Other forms of lifnoise
are also present in amplifier noise and may be seen in data sheets.
5.4.3.11 Quantization noise
Quantization noise occurs in systems requiring digitization of analog signals. It
arises from the discrete nature of the digitization process and is related to the
number of bits n in the digital word by
LSB = SIGNALmax
2
n
'
where
n = number of bits,
SIGNAL
max
= the full-scale signal (amps, volts, or electrons), and
LSB = the magnitude of the least-significant bit.
(5.31)
Since the quantization noise is proportional to the LSB, the larger the number
of bits, the lower the quantization noise.
Johnson noise, G-R noise, and lif noise are usually uncorrelated and
therefore add in quadrature as shown in Fig. 5.7. This figure is typical for a
photoconductive detector. The three frequencies labeled Ji, 12, and h are the
"corner" frequencies on the curve. In particular, since a plot of responsivity 9\
versus frequency looks like the G-R noise curve, the region where the D* is
nominally flat extends fromJi toh and is the typical choice for operation.
5.4.4 Noise factor, noise figure, and noise temperature
Noise factor and noise figure are terms often seen in the electrical engineering
literature. Noise factor F is the ratio of the actual noise to the theoretical
(Johnson-limited) noise:
F = real noise power
ideal (Johnson) noise power
SNR
in
1
SNR
out
- DQE .
(5.32)
144 ChapterS
TOTAL
0.1
JOHNSON
0.01 '------------"--------------'''-------'
1EO 1E1 1E2 1E3 1E4 1E5
FREQUENCY (Hz)
Figure 5.7 Combination of several significant noises.
Noise figure NF is the log (base 10) of the noise factor, given by
NF = lOXIOg[ real noise power ] dB.
ideal (Johnson) noise power
(5.33)
Noise figure is often seen in the specification of preamplifiers and is a valid
measure of the additional noise introduced by such a device. One can attempt to
artificially lower the noise figure by the simple maneuver of increasing the circuit
resistance, thereby increasing the Johnson (theoretical) noise. Only one comment
about this: don't do it!
Another quantity often encountered is noise temperature Tn. It is the
temperature of a thermal source that provides a signal power level equal to the
noise power level. It is defined as:
2 .2R2
T = Vn +Zn
n 4kRB
(5.34)
5.4.5 Some noise examples
It is beneficial to see what some noise sources look like on an oscilloscope. The
figures below are taken from Motchenbacher.
3
Figure 5.8 shows scope traces of
noise that is white over three finite bandwidths, listed in Table 5.5.
As can be seen in Fig. 5.8, reducing bandwidth has an effect on the peak
amplitude as well as the rms noise value. Figure 5.9 presents traces for 1/j noise
within three different bandwidths, listed in Table 5.6.
p:aaz
Figure 5.8 Noise that is white over three different bandwidths. (Reprinted from Ref. 3 with
permission from John Wiley & Sons, Inc.)
Table 5.5 White-noise bandwidths.
Curve
Upper
Center
Lower
Bandwidth
200 KHz
20KHz
2KHz
Figure 5.9 11f noise at three different bandwidths. (Reprinted from Ref. 3 with permission
from John Wiley & Sons, Inc.)
146
Table 5.6 1lfnoise bandwidths.
Curve
Upper
Center
Lower
Bandwidth
2KHz
200Hz
20Hz
Chapter 5
Unlike the case for white noise, limiting bandwidth does not proportionately
reduce peak amplitude for lifnoise:
Finally, Fig. 5.10 presents a comparison of sinusoidal signals within white
and lif noise, respectively. In both cases, the signal-to-noise ratio is
approximately I. As seen in the figure, white noise has a "furry" or "fuzzy"
quality in the trace, while lifnoise is 'Jumpy."
Figure 5.10 (a) White noise and (b) 11f noise for SNR = 1. (Reprinted from Ref. 3 with
permission from John Wiley & Sons, Inc.)
From the frequency plane perspective, the smoothing shown is the result of filtering out higher-
frequency terms. From the time-plane perspective, rapidly changing signals are smoothed out by
the energy storage elements in the filter (capacitors and inductors).
1
1
j
l
1
I
i
I
:l
5.4.6 Computer simulation of Gaussian noise
Those considering computer simulations of noise should note that the random
number generator found in most programming languages and spreadsheets
generates a uniform distribution rather than a Gaussian distribution.
Consideration of the central limit theorem leads to a simple algorithm
~ i -O.S
Y=L.,; ,
i =1 Jk112
(S.3S)
where x = RND(l). Use of this series generates random numbers y with mean = 0
and standard deviation cr = 1. The starting point is k uniformly distributed
numbers Xi, where X is between 0 and 1. A value of k = 12 is suggested, which
will give maximum values for y of 6 at the 3cr point. If we wish to generate a new
random variable y' with mean m and standard deviation cr, we form the
expression
y' = m+cry. (S.36)
5.5 Thermal Detectors
5.5.1 Thermal circuit
Thermal detection of optical radiation is a two-step process. Incident optical
radiation is absorbed by the receiving surface of the detector, giving rise to an
increase in the temperature of the surface. The rise in temperature can then be
detected by one of several means, which will be discussed in this section.
All thermal detectors follow the same equation relating a change in
temperature I':1T = Td - To to incident radiation and the thermal properties of the
detector and its surroundings:
(S.37)
where
I':1T= the rise in temperature of the receiving element over a local reference
temperature,
Td= the detector temperature,
To = the temperature of the heat sink,
ex = the absorptance of the receiver,
RT = thermal resistance between the receiving element and the heat sink,
H = the heat capacity of the receiving element (J/deg),
148 Chapter 5
0) = radian frequency of the incoming signal, and
<p = the incident radiant power, defined as <po(!<lY to allow for either steady-
state or modulated-input radiation.
Because the thermal time constant 'trmay be written
(S.38)
Eq. (S .37) may also be written as
(S.39)
The thermal circuit is illustrated conceptually in Fig. S.11.
A receiver having ex as close as possible to 1 (a perfect absorber) provides the
best detector performance, and several black coatings provide good absorption
characteristics. For example, Lampblack (carbon) is quite black in the visible and
near IR, but becomes somewhat transparent at longer wavelengths. Metallic
blacks (gold, platinum, etc.) formed by evaporation in a poor vacuum are of very
low density (low Rr) and very porous, and are black due to multiple reflections.
They are also extremely fragile, sintering at modest temperatures (-6S
0
C for
gold black). Black paints can be effective absorbers, yet their heaviness yields
rugged but slow (high Rr) detectors. Newer etched and anodized coatings and
conversion processes that yield extremely black coatings with less mass are also
available.
As seen from Eq. (S.37), the thermal resistance Rr must be as high as
possible for maximum sensitivity (high I1T.) However, high values of Rr come at
the expense of speed and ruggedness. If speed and sensitivity are equally
important, H must be small. This requirement necessitates a very small structure,
a lightweight substrate, and a black coating. The ultimate limit of Rr is achieved
H
HEAT SINK
Figure 5.11 Thermal circuit.
1
!
.,
~
;
I
l
i
j
j
I
F
when conduction and convection to the surroundings are minimized. One way to
achieve this goal is to place the detector in a vacuum, support it with low-
conductivity materials, and utilize small thin connecting wires. It can be shown
5
that under these circumstances
(5.40)
where Ad is the area of the detector.
The thermal conductance K expressed in W/deg is the reciprocal of R
T
:
(5.41)
Under the conditions described, the Kin Eq. (5.41) is the only conductance
between the detector and its surroundings. The input power producing a
temperature change IlT may be expressed as
cI>=KIlT. (5.42)
Setting the signal-to-noise ratio equal to one so that cI>=NEP, and assuming
the limiting case in which the noise is due only to fluctuations in the incident
power, we obtain:
(5.43)
The mean square value of the temperature fluctuations in the incident beam
may also be expressed as
(5.44)
where B is the noise bandwidth.
6
Equation (5.42) may be rewritten as
(5.45)
Substituting for conductance from Eq. (5.41) and rearranging terms, we
obtain
(5.46)
bearing in mind that (X= E by Kirchhoff s law for a system at thermal equilibrium.
150 Chapter 5
The maximum D* that can be achieved for a 300 K detector (Ad = 1 cm
2
,
B = 1) is thus 1.8 x 10
10
Cm-HZ1l2/W. This is called the Havens limit for a thermal
detector. If the detector element is cooled, this limit increases dramatically,
reaching a D* of nearly 10
15
cm-Hz
1l2
/W at 4 K and close to 10
19
cm-Hz
1l2
/W at
0.1 K. At these low temperatures, other noises predominate; we cannot reach the
theoretical maximum.
5.5.2 Thermoelectric detectors
5.5.2.1 Basic principles
Thermoelectric detectors operate via the thermoelectric effect, in which a
temperature difference produces a voltage difference and vice versa. These two
conditions are described below. The thermoelectric effect was first used for
optical radiation measurements (solar radiometry) by Nobili and Melloni in 1835,
with subsequent applications including early investigation of infrared spectra.
Two dissimilar metals connected in series form a thermocouple, which may
be used in a circuit as shown in Fig. 5.12. If the junctions between the wires are
at different temperatures such that T2 > T}, a current will flow around the loop in
the direction indicated. The current's magnitude will be proportional to i1.T =
(T2 - T
I
); its exact value depends upon the resistance of the circuit and the
difference between the thermoelectric powers of the two metals. This
phenomenon is called the Seebeck effect, named after its discoverer, T. 1.
Seebeck (1821).
Opening the circuit, as shown in Fig. 5.13, and measuring the voltage results
in the following relationship:
i1.V
S=- (V/deg),
i1.T
(5.47)
where S = the Seebeck coefficient, or, alternatively, thermoelectric power. The
open-circuit voltage is Si1.T.
METAL 1
J
METAL 2
Figure 5.12 Thermoelectric circuit.
6
[Reprinted from Optical Radiation Measurement
series, Vol. 4, W. Budde, Physical Detectors of Optical Radiation, p. 101 (1983).]
METAL 1
METAL 2 J3@T3 J4@T3 METAL 2
METAL 3 METAL 3
Figure 5.13 Open-circuit thermoelectric pair.
6
[Reprinted from Optical Radiation
Measurement series, Vol. 4, W. Budde, Physical Detectors of Optical Radiation, p. 101
(1983).]
In this case, an additional pair of metal wires represents a voltage-measuring
instrument inserted into the circuit. These wires are typically made of copper.
Two additional junctions, J
3
and J
4
, are formed, whose thermoelectric
contributions cancel out if they are at the same temperature, according to the law
of intermediate materials commonly applied in mechanical engineering. In
addition, if the wires are homogeneous, the voltage depends only on the
temperatures of the junctions, and not on temperature distributions along the
WIres.
In 1834, Jean C. A. Peltier experimented with thermoelectric circuits by
passing currents through them. He noted that one of the junctions became warm,
while the other cooled. The coefficient describing the magnitude of this effect is
called the Peltier coefficient. It is expressed as:
IT = ~ d Q )
I dt '
(5.48)
where dQldt is heat flow and I is current. This effect is exploited in
thermoelectric coolers, which are used to cool detectors, laser diodes, small
refrigerators, dew-point sensors, and many other things. The Peltier coefficient is
related to the Seebeck coefficient by the second kelvin relationship, IT = T x S,
where T is absolute temperature in degrees kelvin. This second-order effect
occurs in opposition to the Seebeck effect; that is, a current flow due to a change
in temperature causes a reduction in that temperature change. It is thus
undesirable in radiation detectors, in which it is desirable to maximize AT for
optimum sensitivity. Hence, such detectors are usually operated with little or no
current flow.
Signals from thermal detectors can be increased by placing several junction
pairs in series, connecting alternate junctions to blackened receivers exposed to
incoming radiation, and connecting the other junctions to heat sinks. Devices
152 Chapter 5
using multiple thermocouple junctions are called thermopiles and are used in
most thermoelectric transducers.
Some of the theory behind thermoelectric detectors can help to understand
their operation. Going back to the thermal equations common to all detectors, we
have
!1T = a14 (dc case) (5.49)
and
aR
!1T = T (ac case).
1 + 00
2
1 2
T
(5.50)
Neglecting Peltier cooling, the responsivities are given by
9\=aS14
(dc case) (5.51)
and
aSR
9\(00) = T (ac case).
1+0021 2
T
(5.52)
If current is permitted to flow, the Peltier effect causes a reduction III
sensitivity. In that case, !1 T is given by
(5.53)
where unsubscripted R is the electrical resistance. Since thermocouples are
purely resistive, the limiting noise is Johnson noise in the resistance R.
Performance can thus be enhanced somewhat by cooling. It can also be enhanced
by reducing the value of R; however, this results in decreased RT as well. The
overall effect is to decrease thermal sensitivity !1 T
A better solution is to use materials having low electrical resistance and high
thermal resistance, but for most metals this is not much of an option. Instead, we
choose materials for their thermoelectric properties and then optimize either
speed or sensitivity, depending upon application. Table 5.7 lists common
thermoelectric materials, along with their thermoelectric power values.
Table 5.7 Thermoelectric materials and thermoelectric power values.
Material S (JiV/oC) Material
Al -0.5 Bi -60
Cu +2.7 Sb +40
Ag +2.9 Si -400'
Fe +16 Ge +300'
Constantan 38

variable, depends on doping
5.5.2.2 Combinations and configurations
There are many material combinations that furnish useful thermoelectric pairs,
the majority of which are designed for industrial thermometry over wide
temperature ranges. Constantan, for example, is an alloy of copper and nickel
developed specifically for thermoelectric measurement. It is paired with copper
or iron to form the popular type J and T thermocouples, respectively. The type T
thermocouple has an output voltage of approximately 40 I.L V for a 1
0
C difference
between the hot and cold junction. Seebeck worked with bismuth-antimony (Bi-
Sb) pairs. These have the highest thermoelectric power for any of the commonly
used metals, 100 II V 1C for the pair. Specially doped silicon and germanium
yield the highest output but are difficult to fabricate.
Four distinct configurations for thermocouplelthermopile radiation detectors
are shown in Fig. 5.14. All early thermopiles were fabricated by soldering or
welding fine wires (usually Bi-Sb), culminating with the Coblentz designs [Fig.
5. 14(a). The linear models are still used for large spectroscopic detectors and the
circular for laser power meters and radiometry. The Schwarz design [Fig.
5.l4(b)] features two pins of doped Si and Ge, connected by a 0.3-11m-thick
blackened gold foil. They are extensively used for small spectroscopic detectors.
A wirewound thermopile [Fig. 5.14(c)] can be fabricated by winding many turns
of Constantan wire onto a thin insulator and electroplating silver on half of it.
The junctions are not very efficient, but they are easy to fabricate.
More recent thermopile designs have been realized by vacuum evaporation
of alternating layers of bismuth and antimony onto a substrate [Fig. 5. 14(d). One
example utilizes a thin MyiarTM substrate placed over an insulating channel. The
thickness of the Mylar determines the speed/responsivity tradeoff for a given
application. Another example utilizes micromachined silicon with a thin oxide
layer (vanadium oxide for example) or aluminum to create monolithic structures
used for thermal imaging. Typical characteristics for both wirewound and
evaporated thermopile detectors are shown in Table 5.8.
I
i
(
I
154 Chapter 5
o
LACK GOLD FOIL
10
(a)
(c)
Figure 5.14 Thermopile configurations: (a) Coblentz, (b) Schwartz (c) wirewound, and (d)
evaporated.
7
[Reprinted from Semiconductors and Semimeta/s series, Vol. 5, N. B.
Stevens, "Radiation Thermopiles," pp. 300-304 (1970).]
Table 5.8 Characteristics of wirewound and evaporated thermopile detectors.
Parameter Wirewound Evaporated
Active area 1 x 3 mm to 1 x 10 mm 0.5 x 0.5 mm to 4 x 4 mm
NEP 0.1 to 1 n' W/HZI/2
0.3 to 1 n'W/Hz
1I2
D* lOB to 10
9
cm'Hz
Il2
/W 1 to 3 x lOB cm'Hz
1l2
/W
Time constant 4 to 400 ms 25 to 100 ms
Responsivity 0.1 to 10V!W 10 to 50 V/W
Resistance
10 n to 2 kn 2 to 20 kQ
Spectral range 0.3 to 100 /lm 0.3 to 30 /lm
Window materials Si0
2
, KRS-5 BaF 2, CaF, KBr
Filter types Long-wave pass, bandpass
-
Table 5.9 Some applications of thermopile detectors.
Passive intrusion alarms
Spectral gas analyzers
Flame detection
Radiometry-laser, spectral, and
broadband
IR thermometry
155
Thermopile detectors are extremely versatile due to their small size, low cost,
and wide wavelength range of operation at dc and room temperature. They can be
ruggedized to survive space applications such as horizon sensing and earth
radiation budget measurements. Some terrestrial uses of thermopile detectors are
shown in Table 5.9.
5.5.3 Thermoresistive detector: bolometer
When a thermo resistive material absorbs incident radiation, it becomes warmer
and its electrical resistance changes. The resistance change can be sensed using a
device called a bolometer, first invented by S. P. Langley in 1880. The bolometer
is a resistor that possesses a high temperature coefficient of resistance, often
abbreviated as TCR and symbolized here as ~ , with units of K-
1
Bolometers are
fabricated from metals (the classical approach) and semiconductors (modem).
General characteristics of resistance as a function of temperature for both
material types are shown in Fig. 5.15.
The equation for resistance as a function of temperature is
R(T) = Ro (1 + ~ T , (5.54)
where ~ = (l/R)(dR/dn and Ro is the resistance at some nominal temperature,
often 25 C.
R
METAL
SEMI
T
Figure 5.15 General characteristics of resistance as a function of temperature for metal
and semiconductor materials used in bolometers.
156 ChapterS
As can be seen from Eq. (S.S4), the change in resistance M is proportional to
R o ~ ~ T where ~ T may be obtained using Eq. (S.37). Note that the resistances
referred to above are electrical, not to be confused with R
T
, a material's thermal
resistance described earlier in the chapter.
For metals, ~ is quite small, on the order of O.S%/oC. These bolometers are
rather insensitive. For the semiconducting materials, mixed oxides called
thermistors, ~ is proportional to ]2, making them potentially quite sensitive.
In order for a resistance to be measured, we must force a current through the
circuit and measure a voltage drop. Although many circuit types may be used, the
half-bridge circuit depicted in Fig. S.16 is the most popular. A voltage source
drives two resistors in series, with the upper resister the load resistor, and the
lower resistor the bolometer itself. A coupling capacitor (C) is frequently used to
block the dc signal across the sensitive component, R
B
This is necessary due to
the sensitivity of the component to dc voltage. As a result, modulated beams are
required for successful operation of the circuit.
The expression for signal voltage measured across the terminals is
(S.SS)
Taking a page from electrical engineering, we invoke the maximum power
transfer theorem to set RL equal to R
B
. Then,
(S.S6)
Figure 5.16 Bolometer half-bridge circuit.
-
Substituting for !:IT from Eq. (5.50) and recognizing that for small !:IT, R- Ro,
the bolometer voltage responsivity 9\v is
(5.57)
Inspection of this equation shows that to maxImIze the responsivity, we
should make RT as large as possible. However, by Eq. (5.38), this also increases
'fT, making the device slower. Thus, we have a tradeoff between responsivity and
speed. Because responsivity varies linearly with bias voltage VB, we may
consider increasing the bias voltage. This also increases the current through the
bolometer, which heats it. Increased heating is typically not a problem in metal
bolometers, but can lead to thermal runaway and burnout in thermistor
bolometers. A constant current bias can prevent burnout, but the additional
electronics add noise to the circuit, making this approach unattractive unless wide
variations in ambient temperature are encountered. A better solution is to make
the load resistor a matched bolometer element and shield it from the incident
radiation.
Because the bolometer is a resistive device, Johnson noise predominates. If
the electrical contacts are less than perfect, we can also get IIf noise. The
ultimate performance limit for the device, in which temperature fluctuation noise
predominates, is rarely seen until the device temperature approaches 4 K.
Typical room temperature bolometers exhibit D* values on the order of 10
9
cm'Hz
I/2
/W, with responsivities varying between 10 and 10
4
V/W depending on
the material and temperature, with time constants varying between 1 and 100
milliseconds, depending upon the size of the bolometer element. (Uncooled
microbolometer arrays, part of a relatively recent development in thermal
imaging, feature thermal time constants of a few milliseconds.) Cooling the
bolometer to cryogenic temperatures increases the D* to 10
12
cm'Hz
l12
/W at 2 K
and 10
16
cm'Hz
1l2
/W at 0.1 K. Bolometers used at these temperatures are typically
fabricated from germanium (the Low bolometer) or composite materials. Another
interesting variant is the superconducting bolometer, which operates at the
superconductor transition temperature. It is extremely sensitive but has limited
dynamic range, even with active bias control.
5.5.4 Pyroelectric detectors
The pyroelectric detector is unusual in that it is capable of high-speed operation
and responds only to changing signals. The pyroelectric effect was first suggested
for radiation detection by Yeou Ta in 1938. Certain ferroelectric materials with
asymmetric crystal orientations display the pyroelectric effect, which is a change
in surface charge (spontaneous polarization) with temperature. The effect is not
158 Chapter 5
observed at constant temperature as mobile charges within the material align to
maintain neutrality. The pyroelectric coefficient p is the change in electric
polarization per change in temperature:
dPs C/ 2K
p= dT cm ,
(5.58)
where dP
s
is the change in polarization.
The pyroelectric coefficient increases with temperature as shown in Fig. 5. 17
until the Curie temperature is reached, when it abruptly drops to zero with
attendant loss of response. Fortunately, this is not usually a permanent condition,
and the device can often be reactivated.
The equivalent circuit is shown in Fig. 5 .l8(b) as a current generator in
parallel with a capacitor and a shunt (or load) resistance R
L
A change in
temperature I:!J.T produces a charge Q such that
(5.59)
where Ad = the sensitive area of the detector and p = the pyroelectric coefficient.
The pyroelectric current ip is the product of radian frequency ro and charge:
100 t ~ L_
tf
lli
---
T
-'
:::==:= ' I
70
60
~ o . - ..' _--L ;-- \-
, \
40 ' - - - - ' - - - - -----r.-- \
J
I ,
30 f--t-.-_ . -H
I
- i ! I
I . ! ! I
20 --l- --. ' .----1=--+-- 1
10 -- --L.-J-- - --+--;
0 - L_ _ __ LJ________ _1_ .1
- 40 -30 -20 -10 0 10 20 30 40 50
TEMPERATURE roc]
(5.60)
Figure 5.17 Pyroelectric coefficient versus temperature.
6
[Reprinted from Optical
Radiation Measurement series, Vol. 4, W. Budde, Physical Detectors of Optical Radiation,
p. 129 (1983).]
j

I
ip Cp
J
RA C
A
V
I
.,
Detector
Amplifier
(a)
input
---01'

t
t
I
ip C
R
V
I

:
7_
Equivalent
circuit
(b)
Figure 5.18 (a) Initial circuit and (b) equivalent circuit of a pyroelectric detector, with the
current generator in parallel with a capacitor and load resistor.
8
[Reprinted from
Semiconductors and Semimeta/s series, Vol. 5, E. Putley, "The Pyroelectric Detector"
(1970).]
The current responslvlty for the pyroelectric detector may be found by
inserting the expression for ~ T in Eg. (5.37) and substituting the expression for'tT
in Eg. (5.38):
<;R . = = o p A d ~ T = oopAdcxRT _ oopAdcxRT
I ~ + 00
2
RT
2
H2 ~ + oo
2
't
T
2
'
(5.61)
where 'tT is the thermal time constant.
To determine the output voltage across a load resistor, RL in Fig. 5.18, recall
that the output voltage signal is the product of current and impedance.
9
In this
case, the output voltage is given by
(5.62)
where RLC = the circuit's electrical time constant 't and ip is given by Eg. (5.60).
Therefore, the expression for voltage may be rewritten as
(5.63)
160 ChapterS
w
Figure 5.19 Voltage responsivity as a function of radian frequency.8 [Reprinted from
Semiconductors and Semimetals series, Vol. 5, E. Putley, "The Pyroelectric Detector"
(1970).].
Substituting for I1T from Eq. (5.37) and applying the definition of thermal
time constant, the voltage responsivity may be expressed as
9\ = aoopAdR
L
RT
v .Jl + 00
2
'[2 ~ + oo2'[T
2

(5.64)
In this case, the responslvlty increases from zero to a flat region, then
decreases at even higher frequencies. The width of the flat region depends upon
the separation of the thermal and electrical time constants. The generic voltage
responsivity behavior is shown in Fig. 5.19, while Fig. 5.20 depicts voltage
responsivity as a function of electrical frequency with load resistance as the
parameter.
5.5.4.2 Pyroelectric materials
Several pyroelectric materials and their properties are listed in Table 5.10. Tc is
the Curie temperature and FM is a figure of merit defined as
where
p = the material's pyroelectric coefficient,
E = the material's dielectric constant, and
C' = the product of the material's specific heat and density.
(5.65)
;
J

RL -Id'n
RL -Kl'n

'"
RL'lo"A
V
-.......;

RL-I0
7
A
V

RL-Id'A
V

RL-IO'A
l'\.
V

RL-lo"A

V
"
I I I lin I 111m
I I 111111 11111111 11111111 11111111 '\. I I 111111
FREQUENCY
Figure 5.20 Typical voltage responsivity curve for pyroelectric detectors.
6
[Reprinted from
Optical Radiation Measurement series, Vol. 4, W. Budde, Physical Detectors of Optical
Radiation, p. 131 (1983).]
Table 5.10 Pyroelectric materials and their properties.
Material Tc (0C) P (C/cm
2
K) FM D* (cmHz
1l2
/W) Remarks
TGS 49 4 x 10
8
5200 10
9
HighD*
Doped TGS 60 5 x 10
9
Best D*
LaTa03 618
2 x 10-
8
1500 6 x 10
8
Bulletproof
SrBaNb0
3
60 to 200
6 x 10-
8
750 5 x 10
8
Fastest
PVF
2
80
2 x 10-
8
1300 2 x 10
8
Cheap,
flexible
TGS is the acronym for the organic compound triglycine sulfate. It depoles
(loses its internal charge) readily, requiring the periodic or even continuous
application of an electric field to maintain operation. It is also sensitive to
moisture and needs protection. The D * can be enhanced by doping the material
with L-alinine.
SBN is strontium barium niobate, a mixture of the general form
SrxBal-xNb03. Polyvinylidene fluoride, PVF
2
, is a plastic film (tradename Kynar)
162 Chapter 5
which can be cut and formed into custom configurations. Other pyroelectric
materials include lead zirconate titanate (PZT), ceramic, barium titanate, and
barium strontium titanate (BST).
5.5.4.3 Operational characteristics of pyroelectric detectors
The outstanding feature of the pyroelectric detector is its uniform wavelength
response (depending on crystal absorption or external blackening) coupled with
high sensitivity attainable without cooling. These detectors are small and require
no external bias source. Several are available with built-in FET preamplifiers and
matched load resistors. The impedance of the device is extremely high such that
restoration of charge equilibrium is a relatively slow process.
The crystal fitted with electrodes behaves like a capacitor with a variable
dielectric. It can be used in either a current or voltage mode. Although the
response of the pyroelectric is slow compared to photon detectors, the
pyroelectric can be used in current mode for higher-speed applications when
sufficient input power is available.
The noises inherent in pyroelectric detectors include Johnson noise, thermal
fluctuation noise, and microphonics. Pyroelectric detectors are piezoelectric and
act as microphones; they must be isolated from the surrounding acoustic
environment for successful operation. Preamplifiers provide an additional noise
source.
In terms of performance, D* values greater than 10
9
cm-Hz1l2/W have been
achieved, as shown in Table 5.5 above, with voltage responsivities on the order
of 10
4
V /W. If the Curie temperature is exceeded, the following procedure may
be used to restore operation of the device:
(1) Heat the pyroelectric element to a temperature slightly above the
Curie temperature.
(2) Apply a bias voltage across the electrodes.
(3) Slowly lower the temperature back to ambient with the bias voltage
applied.
Some pyroelectric detectors will spontaneously depole, albeit slowly, at room
temperature. These detectors require periodic application of a poling bias, or
even a constant bias, to maintain proper poling.
5.5.4.4 Applications of pyroelectric detectors
Perhaps the most pervasive use of pyroelectric detectors is in infrared motion-
detection systems, in which two detectors are placed behind a Fresnel lens array
to monitor a designated area. At the equilibrium, or no motion condition, there is
no output signal from either detector. Motion of any thermally radiating object
across the detectors' fields of view causes a change in received power in one or
both detectors, setting off an alarm or triggering the lighting of an area. A
separate photocell within the system inhibits operation during daylight hours.
Pyroelectric detectors are also employed in:
(I) tympanic thennometers
(2) laser power meters and beam pro filers
(3) thennal imaging systems
(4) fire detection
(5) pollution monitoring/gas analysis systems.
163
Pyroelectric detectors are available as single elements and in one- and two-
dimensional arrays. Some two-element arrays are differential, with the elements
wired in opposition. This scheme enhances sensitivity for motion detection
systems. Other linear arrays are available with 128 and 256 elements. Two-
dimensional arrays have been made to 320 x 240 elements.
5.5.5 Other thermal detectors
The Golay cell was invented by M. J. E. Golay in 1947 for use in what was at
that time the relatively new field of infrared spectrophotometry. It remains one of
the most sensitive thennal detectors ever built. Figure 5.21 shows its
constructi on.
In operation, incoming radiation is absorbed by the blackened membrane
which heats the gas within the cell and distends the membrane, which is silvered
on the opposite side. Light from a lamp illuminates a Ronchi screen which is
imaged on the membrane. The reflected beam passes through another Ronchi
screen, and the photocell detects the modulation when the membrane moves.
The Crooke radiometer is an early example of an optomechanical detector. It
is pictured in Fig. 1.1 and consists of a rotating element with four flat vanes.
Each vane is absorptive (black) on one side and reflective (polished) on the other.
Theoretically, the rotation would be such that the reflective side would recede
when radiant energy is incident. The momentum of the incident photons would
be absorbed on the black side, and would recoil from the shiny side. This mode is
never observed, as a very hard vacuum and virtually perfect bearing would be
required. In practice, a small amount of residual gas is present inside the glass
envelope, and the gas is heated by the black side of a vane. The local gas pressure
is slightly higher at the black face than at the shiny face, resulting in a slight
pressure differential. This causes the black face to recede from the incoming
radiation.
(GRID
FLEXIBLE MIRROR) --.---

I I - ------
I - -----___
I CHAMBER __ j<...j-_...Y
\ ABSORBING FILM
DETECTOR
Figure 5.21 Golay-cell-detector schematic.
6
[Reprinted from Optical Radiation
Measurement series, Vol. 4, W. Budde, Physical Detectors of Optical Radiation, p. 133
(1983).]
164 ChapterS
Other optomechanical detectors include the liquid-in-glass thermometer, a
bimetallic strip, and a piezoelectric bimorph.
5.6 Photon Detectors
This section describes the most common photon detection schemes: the
photoconductive, photoemissive, and photovoltaic. First, we need a brief review
of some simple semiconductor physics.
5.6.1 Detector materials
Intrinsic semiconductors are nearly pure materials with a minimum concentration
of impurities. Most have a well-defined energy gap. Photons with energies less
than the energy gap value are not absorbed, but are either reflected or
transmitted. For a photon to be absorbed, it must have a minimum energy and
will then elevate an electron from the valence band (bound state) up into the
conduction band (free state) such that it is available for conduction under an
applied electric field. Figure 5.22 demonstrates this concept schematically, while
Eq. (5.66) expresses the high wavelength cutoff beyond which electrons will not
have the energy to reach the conduction band:
"l _ he _ 1.2398
/\,c------,
Eg Eg
(5.66)
where Ac = the high wavelength limit (cutoff wavelength) in micrometers, and
Eg = the gap energy, or the minimum energy required to elevate an electron into
the conduction band. It is expressed in electron volts (eV).
Thermalization loss
Ep = Photon
energy
Conduction band
Eg = Gap energy
E
f
= Fermi
level
Valence band
Figure 5.22 Illustration of valence and conduction bands, with Eg the energy necessary to
promote an electron from the former to the latter.
Another way to think about the concept is simply to recall that when Ep =
he/A is greater than or equal to E
g
, for a given material, electrons generated by
photons at wavelength A will possess the energy necessary to elevate into the
conduction band.
In Fig. 5.22, the excess energy Ep - Eg appears as heat. In the case of a
photovoltaic cell (to be discussed in greater detail later) this thermalization loss
causes the cell's voltage and power to decrease. It is one of the two primary loss
mechanisms responsible for the fairly low peak theoretical efficiency of about
28% for simple photovoltaic cells.
A number of interesting intrinsic semiconductor materials are candidates for
optical radiation detection, as shown in Table 5.11. For example, silicon is seen
to require a maximum wavelength of 1.1 Ilm, 1100 nm, and this leads to the
dramatic falloff on the right side of the photon detector curve seen in Fig. 5.1.
Most of these materials may be categorized in a straightforward manner, but the
last combination deserves special mention. Mercury telluride (HgTe) is classified
as a semimetal, with a small negative energy gap. This means that HgTe is a
conductor at room temperature, albeit a rather poor one. Cadmium telluride
(CdTe) is a semiconductor which has been exploited for visible radiation
detection. When the two tellurides are combined as Hg1-xCdxTe, then the energy
gap depends on x, the fraction of CdTe in the mix. A mixture where x = 2 is
common and yields a detector with response out to about 12 Ilm. The longest
wavelengths that are practical with this trimetal detector are about 25 Ilm, and
cold temperatures are required for effective operation.
Table 5.11 Semiconductor materials useful for optical radiation detection.
Material Energy gap (e V)
Cutoff A l m ~
SiC 3.0 0.41
CdS 2.4 0.52
CdSe 1.74 0.71
GaP 2.25 0.55
GaAs 1.4 0.89
InP 1.25 0.99
Si 1.12 1.1
Ge 0.68 1.8
PbS 0.37 3.35
PbSe 0.26 4.8
lnAs 0.33 3.8
InSb 0.23 5.4
CdTe 1.6 0.78
HgTe -0.3
Hg1_xCdxTe variable 1 to 24
HgCdTe {x = 2} 0.1 12.4
166 Chapter 5
Table 5.12 Intrinsic carrier concentrations for Si and InSb.
Material T(K) r Eg(eV)
e-
EglkT
ni
Si 300 2.7E7 1
3.T19
1.34
10
InSb 300 2.7E7 23
1.4-4 2.i7
InSb 77 4.6E5 23
8.8-
16
9.010
The relationship between intrinsic carrier concentration and temperature is
due to thermal activity, and is expressed as:
2 3 -E IkT
nj = constantxT Xe g ,
where
ni = the concentration of carriers,
constant 2 x 10
31
cm-6K-
3
for most materials, and
kT= 0.02585 at 300 K with units ofeV.
(5.67)
Figure 5.23 shows the intrinsic carrier concentration plotted versus
temperature, with energy gap as the parameter. As seen from the figure, an
increase in temperature produces an increase in carrier concentration, and
materials whose energy gap is large have lower carrier concentrations. Table 5.12
shows these effects for silicon and indium antimonide (InSb).
INTRINSIC CARRIER CONCENTRATION
100 150 200 250 300
TEMPERATURE (K)
350 400
Figure 5.23 Intrinsic carrier concentration as a function of temperature for several Eg
values.
...
Looking at the two temperatures for InSb in Table 5.12, it is apparent that the
T in the exponent in Eq. (5.67) has more influence on the calculated value of n;
than does the yJ term. Since we want sufficient carriers available for photon
excitation even at low values of n;, we must cool low Eg detectors that are
designed for long-wavelength operation. This fact explains the need, for
example, to cool HgCdTe detectors to temperatures of 77 K or lower. Detectors
should be cooled so that
E 600
kT <-.L or T <--,
25 Ac
(5.68)
with Ac given by Eq. (5.66).
Extrinsic semiconductors are those in which a small amount of a selected
impurity (called a dopant) is introduced into an intrinsic material. Dopants create
additional energy levels within the intrinsic energy gap. The most common host
materials have been silicon (Si) or germanium (Ge), both from group IV of the
periodic table of the elements. Germanium has historical precedence, and was for
many years the only photon detector material available for wavelengths longer
than 5 11m. Silicon-based detectors are used extensively in visible and near-
infrared applications; their ease of integration with other monolithic circuit
components is a strong element in their favor.
Donors are atoms from group V of the periodic table which have five
electrons in their outer shell. Only four are needed to match up with Si or Ge,
however. The fifth electron is rather loosely bound to the host atom. It resides at
the donor level at 0 K, but may be thermally elevated to the conduction band at
an elevated temperature. Figure 5.24 shows this schematically.
Acceptors are atoms from group III which are short an electron when
compared with the host material. These holes (absence of an electron) are at the
acceptor level at 0 K. When heated, electrons are thermally elevated from the
E
f
OK VALENCE BAND 300 K
Figure 5.24 Energy band structure for donor semiconductors.
168 ChapterS
o K CONDUCTION BAND 300 K
E,
000 ACCEPTOR LEVEL
VALENCE BAND
Figure 5.25 Energy band structure for acceptor semiconductors.
valence band to acceptor levels, leaving behind a hole for conduction, as shown
schematically in Fig. 5.25. Table 5.13 lists some of the impurity dopants and
their levels and cutoff wavelengths.
The maximum practical doping for extrinsic materials is about 1 ppm, or
about 6 x 10
17
atoms/cm
3
At room temperature, they are nearly all used up. As
the temperature increases, the material reverts to an intrinsic conductor.
There are a number of other interesting semiconductor materials, among
them PbxSnl-xTe, that have made good photovoltaic detectors. Many alloys can
be formed from combinations of materials in groups III to V or II to VI in the
periodic chart. The materials in Table 5.14 below have been successfully used in
the fabrication of heterostructure alloys for solid-state sources and detectors.
Some are useful in the ultraviolet, others in the visible, and many others in the
infrared.
Table 5.13 Donor and acceptor levels for germanium and silicon.
in Germanium in Silicon
Dopant eV
Ac (J1m)
eV
Ac (J1m)
Au 0.15 A 8.3 0.54A 2.3
Cu 0.041 A 30 0.24 A
Zn 0.035 A 35 0.26 A
Hg 0.087 A 14 none
Cd 0.055 A 22 0.3 A
Ga 0.011 A 112 0.0723 A 17.8
B 0.045 A 27.6
Al 0.0685A 18.4
In 0.011 A 0.155 A 7.4
S 0.187 D 6.8
As 0.013 D 0.054 D 23
Sb 0.0096 D 129 0.039 D 32
Note: "A" denotes an acceptor and "D" denotes a donor.
-
Table 5.14 Semiconductor materials used for solid-state sources and detectors.
AlP
AlAs
AISb
GaN
GaP
GaAs
GaSb
InP
lnAs
InSb
5.6.2 Photoconductive detectors
Intrinsic and extrinsic photoconductive detectors have been discussed in the
previous section. To recap, intrinsic photo conductors are typically found at
shorter wavelengths than extrinsic photoconductors, whose doping with an
impurity provides a longer-wavelength response. Figure 5.26 shows the general
layout of a photoconductive detector, with L the distance between electrodes.
The conductivity of a slab of semiconductor material is
where
q = unit of electric charge on one electron or one hole,
n = electron concentration (electrons/cm\
p = hole concentration (holes/cm
3
) ,
f.!n = electron mobility [cm
2
/(Volts)], and
f.!p = hole mobility [cm
2
/(Volts)].
INCIDENT PHOTONS
Figure 5.26 Photoconductive detector structure.
(5.69)
170 ChapterS
In an extrinsic semiconductor, one carrier is present. An n-type extrinsic
semiconductor will be assumed to simplify the analysis, so that
cr
e
(5.70)
In the presence of incoming light, the conductivity is
(5.71)
where !:J.n = change in carrier concentration due to incident radiation.
The relative change in conductivity due to incident radiation is
(5.72)
To obtain an expression for !:J.n, recall the definition of responsive quantum
efficiency, RQE (symbolized as 11), from Sec. 5.3 as the ratio of independent
output events per incident photon. In this case, the RQE is the number of
electrons elevated to the conduction band per incident photon. The photon
irradiance on the detector is E
q
, expressed in photons/scm
2
, and the carrier
lifetime is 't/. The physical processes occur in a detector of thickness z. The
change in carrier concentration due to incident radiation is
(5.73)
Equation (5.72) can now be rewritten as
!:J.cr
e
't,
(5.74)
cr
e
crez
The relative change in conductivity can also be written in terms of detector
resistance R: \0
!:J.cr -M
__ e = __
cr
e
R
(5.75)
with the negative sign indicating that the relative change in resistance has
opposite slope to the relative change in conductivity.
We wish to derive an expression for the voltage responsivity 9\v at
wavelength t... To do so, we consider the placement of a photoconductive detector
in a circuit, as shown in Fig. 5.27.
+
VB
Figure 5.27 Photoconductive detector bias circuit.
The signal voltage out Vs is
where
VB = bias voltage,
RL = load resistance, and
RD = detector resistance.
In Fig. 5.27, C(opt) is an optional capacitor in the circuit.
171
(5.76)
The optional capacitor placed in the circuit does not factor into the analysis,
but is included because values of V, can be very large, and VB can get as high as
200 V. Placing a capacitor in the circuit allows a modulated signal. Note that the
device is symmetrical, and that the polarity of the applied bias in Fig. 5.27 is
unimportant.
We need an expression for !lV
s
, the change in output signal voltage due to a
change in resistance. Differentiating Eq. (5.76), we obtain
(5.77)
Note that VB /(RL+RD) is Ide, the dc current flowing through the detector, so that
(5.78)
Eq in Eq. (5.74) is the photon irradiance on the detector of area Ad. At a
particular wavelength, it is <PA- (hcAdr
l
Making this substitution in Eq. (5.74)
172 Chapter 5
and applying the results of Eqs. (5.75) and (5.78) to the definition of voltage
responsivity, we obtain
~ V 5 = IdcRLRDATJ.q)ln '[I
 hczAd(je(RL + R
D
)
(5.79)
According to the maximum power transfer theorem of electrical engineering,
the output signal V, is maximized when RL = RD. For small signals, we can call
them both R. In that case, Eq. (5.79) becomes
9\ = v = IdcRATJ.q)l'[1
v 2hczA
d
(j e '
(5.80)
where /l, the carrier mobility, may be /In as above, or more generally, /In + /lp.
For good photoconductivity, we want high values for TJ., '[I, and /l, and low
values for (je and z. As in most other practical applications, compromises and
tradeoffs are required. If '[I is large, the device will have a slow response. If the
device is thin, small z, the RQE TJ. will be reduced because of incomplete
absorption.
If a carrier in transit comes too close to a nucleus having a vacancy, it may
recombine. In that case, an electron will return to the valence band or top the
relevant donor or acceptor level. The carrier lifetime '[I is the statistical time
between generation and recombination. If a photogenerated carrier arrives at one
of the electrodes without recombining, another carrier leaves the opposite
electrode in order to maintain charge neutrality. In this way, more than one
carrier may exist for each absorbed photon, a situation that constitutes a gain.
The photoconductive gain G is defined as the ratio of the carrier lifetime to the
carrier transit time
(5.81)
where the transit time, in tum, is defined as
(5.82)
The gain is increased by decreasing the spacing between electrodes or applying a
large electric field. The price paid here is an increase in response time.
We can also derive an expression for voltage responsivity in terms of gain.
The signal photocurrent is
= 'I'lq A G = 'I'lq A . 't//lV
B
.
is ' I he ' I he /2
(5.83)
Applying Ohm's law to the circuit in Fig. 5.27, with RL = RD as above, we
obtain
v = (A<I('t//lVB R)
s llq he /2 2 '
(5.84)
and
(5.85)
Note that the current responsivity 9\j may be obtained directly from Eq.
(5.83) by dividing the signal current by the power term, such that
9\, = llq(AA1e)G NW.
We can maximize 9\v by increasing VB, but if Joule heating occurs and the
PC detector heats up, decreasing its resistance, we will burn it out!
5.6.2.2 Noises in photoconductive detectors
The noises commonly found in photoconductive detectors are Johnson, G-R, and
1/j. A typical noise expression is
In' =4q llq - G
2
+qG
2
N' +--+- B.
[ (A) kT kT]
he qRD qRL
The terms inside the bracket, according to their order, are
(1) G-R noise from incident photons (signal and background),
(2) dark current noise due to N' thermally generated carriers,
(3) Johnson noise in the detector resistance R
D
, and
(4) Johnson noise in the load resistor R
L

(5.86)
The limiting noise is G-R noise from radiation (signal plus background).
When G-R noise overpowers all other noises, the rms signal current is
(5.87)
while the rms noise current is expressed as
174 ChapterS
(5.88)
where B is the effective noise bandwidth.
Recall from Eq. (5.8) that D* may be expressed as
where Ad is the detector area. Noting that = EA
d
, and rearranging terms in Eqs.
(5.87) and (5.88), D* for the photodetector becomes
(5.89)
This quantity is called D*BLIP because, as discussed in Sec. 5.3, the limiting
noise arises from incident photons. Equation (5.89) expresses a quantity
referenced to a particular wavelength and having a specific modulation frequency
f Expressing this irradiance in terms of photon incidence Eq we also obtain
D*BLlP fl.
Eq
(5.90)
The background can be reduced by appropriate shielding and filtering using
cold filters when necessary. A detector can be background limited for a 300-K
background, but may become Johnson-noise limited for cold backgrounds.
Extrinsic detectors require more cooling for a given cutoff wavelength than do
intrinsic detectors, and their physical thickness is typically much greater, due to
the low absorption coefficient of the host material. A typical infrared
photoconductor installation consists of a liquid nitrogen-cooled dewar with an
appropriate window, and an optional cold shield and cold filter, as illustrated in
Fig. 5.28.
5.6.2.3 Characteristics of photoconductive detectors
In the infrared, PbS, PbSe, InSb, and HgCdTe intrinsic photoconductors along
with doped silicon and germanium extrinsic photo conductors are common. Each
is optimized for a different wavelength region. HgCdTe is unique in that it is an
alloy of CdTe (E
g
"" 1 eV) and HgTe (E
g
< 0 eV, a semimetal). By selecting the
composition of the alloy, almost any bandgap and therefore almost any peak
pz
LIQUID
NITROGEN
Figure 5.28 Detector in vacuum dewar.
175
wavelength can be obtained. The doped silicon detectors are currently in favor
for focal plane applications as on-chip signal processing can be accomplished
with conventional silicon technology. The lead-salt detectors will operate at room
temperature, albeit poorly, but do much better when cooled to -193 C. Most
InSb and HgCdTe detectors prefer 77 K while the extrinsic detectors based on Si
and Ge require even lower temperatures. The lead-salt detectors have somewhat
slower response times than the others.
5.6.2.4 Applications of photoconductive detectors
In the visible portion of the spectrum, CdS, CdSe, and mixtures thereof are the
most common photoconductive detectors. They are reliable and rugged if
protected from the environment, fairly sensitive, though not very linear. They can
handle large amounts of power. They are most often used in industrial and
commercial photoelectric controls. In the infrared, photo conductors find use as
point and array detectors, particularly beyond 5 /lm where photovoltaic detectors
are either unavailable or do not perform as well.
5.6.3 Photoemissive detectors
The process of photoemission relies on an external photoeffect, in which an
electron receives sufficient energy from an incident photon to physically escape a
photosensitive material called a photocathode. Photoemission can take place
from virtually any surface given sufficient incident photon energy. It was first
176 Chapter 5
observed in metals in the UV by Heinrich Hertz (1887). Einstein received the
Nobel Prize in physics in 1921 for his explanation of photoemission, originally
published in 1905.
Photoemission is a three-step process:
(l) a photon is absorbed; the result is a "hot" electron;
(2) the electron moves to the vacuum interface; and
(3) the electron escapes over the surface barrier to the vacuum.
In practical devices, the electron is attracted to and collected by a positively
charged anode.
The energy required for an electron to escape the surface barrier is known as
the "work function," symbolized as <1>. It is material specific. In metals, it can be
used to calculate the cutoff wavelength of an incident photon, beyond which an
electron will not obtain the necessary kinetic energy to escape the barrier:
(5.91)
where Ac is the cutoff wavelength in nm, and is in eV.
5.6.3.2 Classes of emitters
Figure 5.29 schematically depicts photoemission in metals. Because no two
electrons may occupy the same energy state according to the Pauli exclusion
principle, energy states within the metal are separated into a number of closely
spaced levels. \0 The lines to the left in the figure denote energy levels, the highest
being E
j
, the Fermi level. Typical work functions are shown in Table 5.15, along
with the corresponding long-wavelength cutoffs.
Metallic photocathodes are used in the ultraviolet as they are stable and
"blind" to photons having lower energies. However, due to metallic reflection
and internal absorption and electron scattering, the quantum efficiency is very
low (11 "" 0.001).
E
METAL VACUUM
Figure 5.29 Photoemission in metals.
Table 5.15 Work functions for metals.
Metal
{eV} Ac (nm}
Pt 6.3 200
W,Cu 4.5 275
Mg 3.67 340
Ca 2.71 460
Na 2.29 540
K 2.24 555
Cs 2.14 580
Higher quantum efficiency and longer wavelength operation can be achieved
with the "classical" photocathodes fabricated from alkali metals, chiefly cesium,
and with semiconductor photocathodes. In semiconductors, the photon must
impart enough energy to an electron in the valence band to reach the conduction
band and have enough energy left to escape the material into the surrounding
vacuum. The energy difference between the bottom of the conduction band and
the vacuum level is called the electron affinity Ea. Recall that the energy required
to elevate an electron from the valence band to the conduction band is called the
gap energy E
g
Their sum defines the work function for semiconductors,
 = Eg + Ea. This is the total energy required to escape the potential barrier. In
practice, Eg should be larger than Ea to minimize scattering losses within the
material.
As was the case for metals, the cutoff wavelength can be calculated
according to Eq. (5.91). Figure 5.30 schematically depicts photoemission in
semiconductors. The electron behavior described above, for both semiconductors
and metals, may be described according to Fermi-Dirac statistics:
1
P(EJ = 1 (E -Ef)lkT '
+e "
(5.92)
E
VACUUM
CONDUCTION
TEa BAND

Eg
E
f
VALENCE
BAND
Figure 5.30 Photoemission in semiconductors.
178
0.8

:::; 0.6
iii

CD
Chapter 5
FERMI-DIRAC FUNCTION
0.4 I- ' _____ 1_______ ..----.--'---.-- --.---. j ___ c._____ .-- --
a.
0.2 ----"--------------'-- ---
_ l
o 0.5 1
ENERGY
1.5 2
Figure 5.31 Fermi-Dirac function, plotted with temperature as a parameter.
where
En = the energy of the nth state,
PeEn) = the probability that state n is occupied, and
k = Boltzmann's constant.
At En = E
j
, the probability of a state being occupied is 0.5. This statistical
function is plotted as Fig. 5.31.
With semiconductor materials, quantum efficiencies up to 0.3 can be
achieved at wavelengths as long as 800 nm, and a quantum efficiency of 0.01 is
found at wavelengths out to 1.2 Table 5.16 gives the pertinent parameters.
The higher quantum efficiencies result from lower reflection losses and less
scattering. More recent materials based on photoconductive single-crystal
semiconductors with a thin Cs-based surface demonstrate even higher quantum
efficiency and response out to 1.65 Table 5.17 shows some of those
materials.
Table 5.16 Photocathode energies and cutoff wavelengths.
Material Eg(eV) Ea(eV) t,=Eg+Ea (eV) AADm)
LiF 12 I 13 95
Csl 6.3 0.1 6.4 195
GaAs 1.4 4.1 5.5 225
Si 1.1 4 5. 1 245
Ge 0.7 4.2 4.9 255
Cs
2
Te 3.5 350
K
2
CsSb (bi-alkali) I 1.1 2.1 590
CsSb (S-II) 1.6 0.45 2.05 605
Ag-Bi-O-Cs (S-lO) 0.7 0.9 1.6 775
Na2KSb:Cs (S-20) 1 0.55 1.55 800
Table 5.17 Nominal composition and characteristics of various photocathodes.
1o
(Reprinted by permission of Burle Technologies.)
Nominal PC
JETEC Conversion Luminous
composition type
response (lumen/W responsivity
designatiou
at A...ax) (/lAllm)
Ag-O-Cs 0 S-1 92.7 25
Ag-O-Rb 0 S-3 285 6.5
Cs
3
Sb 0 S-19 1603 40
Cs
3
Sb 0 S-4 1044 40
Cs
3
Sb 0 S-5 1262 40
Cs
3
Bi 0 S-8 757 3
Ag-Bi-O-Cs S S-10 509 40
Cs
3
Sb S S-13 799 60
Cs
3
Sb S S-9 683 30
Cs
3
Sb S S-l1 808 60
Cs
3
Sb S S-21 783 30
Cs
3
Sb 0 S-17 667 125
Na2KSb S S-24 758 85
K
2
CsSb S 1117 85
Rb-Cs-Sb S 767 120
Na2KSb:Cs S 429 150
Na2KSb:Cs S S-20 428 150
Na2KSb:Cs S S-25 276 160
Na2KSb:Cs S ERMA II 220 200
Na2KSb:Cs S ERMA III 160 230
GaAs:Cs-O 0 116 1025
GaAsP:Cs-O 0 310 200
In.06Ga.94As:Cs-0 0 200 250
In. 12Ga.88As:Cs-0 0 255 270
In. 18Ga.82As:Cs-0 0 280 150
Cs
2
Te S
CSI S
CuI S
K-Cs-Rb-Sb S 672 125
S = semitransparent
0= opaque
180 Chapter 5
Wavelength Responsivity Quantum
Dark
Nominal emission
composition
of maximum
at Amax
efficiency
at 25 C
response (nm) (mAIW) at Amax(%)
{fAlcm2}
Ag-O-Cs 800 2.3 0.36 900
Ag-O-Rb 420 1.8 0.55
Cs
3
Sb 330 64 24 0.3
Cs
3
Sb 400 42 13 0.2
Cs
3
Sb 340 50 18 0.3
Cs
3
Bi 365 2.3 0.77 0.13
Ag-Bi-O-Cs 450 20 5.6 70
Cs
3
Sb 440 48 14 4
Cs
3
Sb 480 20 5.3
Cs
3
Sb 440 48 14 3
Cs
3
Sb 440 23 6.7
Cs
3
Sb 490 83 21 1.2
Na2KSb 420 64 19 0.0003
K
2
CsSb 400 95 29 0.02
Rb-Cs-Sb 450 92 25 1
NazKSb:Cs 420 64 19 0.4
NazKSb:Cs 420 64 19 0.3
Na2KSb:Cs 420 44 13
Na2KSb:Cs 530 44 10.3 2.1
Na2KSb:Cs 575 37 8 0.2
GaAs:Cs-O 850 119 17 92
GaAsP:Cs-O 450 61 17 0.01
In. 06Ga. 94As:Cs-0 400 50 15.5 220
In. 12Ga.88As:Cs-0 400 69 21 40
In. 18Ga.82As:Cs-0 400 42 13 75
CSzTe 250 25 12.4 0.0006
CSI 120 24 20
CuI 150 13 10.7
K-Cs-Rb-Sb 440 84 24
These conversion factors are the ratio of the radiant responsivity at the peak
of the spectral response characteristic in amperes per watt (A/W) to the luminous
responsivity in amperes per lumen (Allm) for a tungsten lamp operated at a color
temperature of 2856 K.
A newer class of photoemitters known as negative electron affinity (NEA)
materials feature a special surface treatment on a p-type semiconductor substrate
to "bend" the band structure. In extreme cases, the vacuum level is below the
bottom of the conduction band. The advantages of this NEA photocathode
include longer wavelength operation and higher quantum efficiency. Figure 5.32
schematically depicts photoemission from these materials.
CONDUCTION
BAND
E
VALENCE ""
BAND
VACUUM
Figure 5.32 Photoemission from NEA materials.
5.6.3.3 Dark current
181
Dark current is a limiting factor in photocathodes, and is indistinguishable from
photocurrent. The principal source is thermionic emission from the photocathode,
characterized by the Richardson equation:
(5.93)
where
J = current density in Alm
2
,
C = a constant, approximately 1.2 x 10
6
, and
k = Boltzmann's constant, 8.617385 x 10-
5
eV/K.
The temperature dependence of dark current is illustrated in Fig. 5.33.
This graph also shows that cooling is required to achieve best possible
performance, especially with photocathodes designed for longer-wavelength
operation (i.e., low work function). For intrinsic semiconductor photocathodes,
substitute Ea + (E
g
I 2) for <1> in the Richardson equation.
Other sources of dark current are thermionic emissions from the dynodes,
leakage current between the anode and other structures, photo current from
scintillation from the envelope or electrode supports, field-emission current,
ionization from residual gasses in the envelope, and currents caused by cosmic
rays, environmental gamma rays, and radioisotope radiation from the envelope
and support structures.
182 ChapterS
~
l-
z
UJ
0::
0::
::>
u
~
0::

Cl
1E-6
1E-8
- -
I
- - - -
I
-
,- ,- -
1E-10
- _ I--
- - - - '- - -
I
1E-12
I I
.
- f- -
-. - -
r
- -
1E-14
- _ L
- -
1E16
1E18
I
-- _. r - -
RICHARDSON EQUATION
- - - - --
I
I
- - I-
-
. -
I
- - - - -
I
I
-
_L
- -
I
-
-r
- -
I
. -- -- 1---
__ _ -- 1- --
I I
- - ,- - - -- - 1' - - -
__ '- __ __ 1- ____ 1-- _ _ _ __ '- _ ___ '- __
I I
I I
--r-- --r-- -- r-- -- r - - - - r - -
I I I I
_ _ L _ _ __ L ____ L ____ '- _ __ _ 1- __
I I I I
--r-- --r - - --r-- -- r -- - r
I
1E_22L--L __ ~ __ ~ __ -L __ L--L __ ~ __ ~ __ __ L ~
200 220 240 260 280 300 320 340 360
TEMPERATURE (K)
Figure 5.33 Richardson equation expressing dark current as a function of temperature.
5.6.3.4 Noises in photoemissive detectors
Noise sources in photoemissive detectors include the following:
(1) shot noise from signal photocurrent,
(2) shot noise from background photocurrent,
(3) shot noise from dark current, and
(4) Johnson noise from the load resistor.
A noise expression may be developed by considering the quantities that make
up these sources.
The signal current from a photoemissive detector is
i = T!q<1> = T\ q<l>!:.. .
s q he
(5.94)
If signal current flows through load resistor R
L
, then the signal voltage is
(5.95)
Applying Eq. (5.95) and the results from Eqs. (5.18) and (5.25), the noise
voltage is
(5.96)
The three terms in the inner bracket are as follows:
(1) shot noise due to dark current i
d

(2) Shot noise due to signal + background current,
(3) Johnson noise in load resistor R
L
.
The signal-to-noise ratio is therefore
(
A) 11q<l> -
SNR= he
[(
. 2 A 4kT) ]1/2.
2q
l
d + 2q 11<1> he + RL B
183
(5.97)
The ultimate limit is achieved when the dark current shot noise and the
Johnson noise from the load resistor can be reduced, leaving only the signal-
dependent shot noise. Under these conditions, the SNR is
SNR = .
2heB
5.6.3.5 Photoemissive detector types
(5.98)
Photomultiplier tubes. The impact of an electron onto a secondary emlttmg
material releases several secondary electrons. The gain is defined as the number
of secondary electrons per incident electron; its symbol is o. Values are in
MgO, in Cs
3
Sb, and variable at 50/keV for GaP:Cs. Special structures called
electron multipliers arrange a series of these secondary emitting materials such
that electrons can be accelerated towards the next electrode (dynode) which has a
more positive potential. The total electron multiplier gain is On, where n is the
number of dynodes. The gain also depends upon applied voltages.
In a photomultiplier tube (PMT), a photosensitive photocathode is combined
with an electron multiplier. In operation, a photoelectron is ejected from the
photocathode and accelerated towards the first dynode. Several electrons are
released and accelerated towards the second dynode, the third, and so on. There
are many interesting designs for electron multiplier structures, yielding up to 14
stages of gain.
There is some additional noise introduced in the multiplication process. A
noise factor (NF) may be calculated as
(5.99)
For large values of 0, Eq. (5.99) becomes
184 Chapter 5
NF=_O_.
(0-1)
(5.100)
This noise factor is quite small, typically less than 1.2. The gain of the
electron multiplier is essentially noise free. Table 5.18 lists some of the positive
and not-so-positive characteristics of photomultiplier tubes.
Photomultiplier tubes have found a number of different uses in areas
including photon counting, spectroradiometry, and imaging. In the latter, many
PMT -based devices have been replaced with solid-state imagers.
Microchannel plates. Microchannel plates (MCPs), useful in many UV, visible,
and x-ray applications, are disks built up from millions of microchannels, small
glass tubes whose diameters may range from 10 to 40 !lm. They provide an
electron multiplication function and form the core of many image intensifier
systems, with each channel of the disk (plate) contributing one picture element
(pixel) to the resulting image. Typical MCP disk sizes range from 18 to 75 mm in
diameter with lengths between 0.5 and 1 mm.
Table 5.18 Photomultiplier tube characteristics.
PMT characteristics (good) PMT characteristics (not so good)
Large number of photocathode spectral FRAGILE! Most are made of glass
sensitivities
Detectors with S-numbers are
"classical" photocathodes
Newer NEA photocathodes described
by base semiconductor material
Very fast, limited by transit time
Crossed-field version confines electron
paths via a magnetic field
Quantum efficiencies from 0.01 to 0.5
Can be physically large
Require stable high-voltage power
supply ~ 1 kV)
Voltage divider string required
Require shielding from electrostatic
and magnetic fields
May require light shielding to prevent
photons from getting to dynodes
Residual response to cosmic rays,
radioactive materials in tube
Can be physically large
Phosphorescence in window
Photocathode memory and fatigue
Photocathode spatial nonuniformity
Photocathode stability (particularly S-l)
F
1
I
1
e
-f.:'
~
"';- }
jtXLX,
J \.,.:.,.1
(a) (b)
1
;mUIUi]
""\--r.r.r",
f'-'-' -'-'-
(c)
(d)
00
..
00-
(e)
(f)
Figure 5.34 Photomultiplier dynode arrangements: (a) circular-cage type, (b) box-and-grid
type, (c) linear-focused type, (d) venetian blind type, (e) fine mesh type, and (f)
microchannel plate.
11
(Reprinted by permission of Hamamatsu Photonics K. K.)
The principle of operation of the microchannel plate is very similar to that of
the photomultiplier tube, with the difference that the microchannel replaces a
series of dynodes as the vehicle for amplification. The microchannel's inner
surface is coated with a high-resistivity material having good secondary emission
characteristics. In operation, a primary electron entering from a photocathode
strikes the wall and causes secondary emission; this process continues until a
high number of electrons have been accelerated toward the positive electrode at
the other end of the tube.
7
Fig. 5.34 shows the dynode arrangement for several photomultipliers. There
are a number of photocathode spectral sensitivities from which to choose; a
representative sample is shown in Fig. 5.35.
5.6.4 Photovoltaic detectors
The photovoltaic detector is a popular detector whose operation relies upon an
internal potential barrier with an electric field applied. A p-n junction in a
semiconductor material is typically used to provide this condition. The potential
barrier is formed by doping adjacent regions such that one is an n-type (donor)
region and the other a p-type (acceptor).
186
<
s

::;
i=
Cii
z
UJ
CfJ
I-
Z
<
15
<
II:
UJ
o
o
J:

()
o
b
J:
a..

<
. .s

::;
i=
Ui
Z
UJ
(/)
I-
Z
<
15
<
II:
UJ
0
0
J:
I-
<
8
I-
0
J:
a..
100
80
60
40
20
10
8
6
4
1.0
O.B
0.6
04
0.2
ChapterS
REFLECTION MODE PHOTOCATHODE

!"'" 552U ;_ ,-552S _
..

:ri

,
,.; ..- I ' . "-
;
V
I,
' ,L

ill

f'l\ 350K
,
1'\ !

V

--

J I
i---I
f.-
\"/o.;:z 551S
-.

+\551U
,

-

,J.--

=\650t"= i

! \ '
:

...... I 1-1. .........
:
.----r- I

t-z .'i
1
"
i

I
-+-I i
I I

I

;
I I

! i
'7SoJ I

l--
I i I..--t::
,
0;;'* r:::
..
..
1----+- , - t-o.
r-

,
r--r
,
.
;
l..--r
I
:
,d---:-"1 f , 1
11
1----1
1 Jd::J
t--
1 1
I
I I , ,
--
I I I
I
J--t-t
f--
I
I I
i
!

j I
,
I
!
I
!
I
,
I I I
0 1
100 200 300 400 500 600 700 800 1000 1200
WAVELENGTH (nm)
(a)
TRANSMISSION MODE PHOTOCATHODE
100
80
60
40
20
10
8
6
4
2
1.0
0.8
t:::- :

r,pio J----::-;-
500S
I
,
..
. :
k-'I If7.
f""" J
t- p;
rL
5OOKo..:::: l
_\.\."'- ,,, l::::i i :

i
V I-
r A' lJ'

l%1
3OOK -
8 200M
400U
-
..
:cf

r '

N 'Y
i : r ---
; . ,
h...J--1 i H

.J.-
!--'l
t:l 1!
1
A

'
! ! l--
rr
1
f-
H
n
...-,'t
f.--' -t-'

06
.--
501K
..:: 700K:
,
0.4
-
,
-
"-
", -::: 1
'-
..t-.
\
0.2
-----+-1 i
,
I-
....
:
.. +-
I
1
-,-
i
!
r-
iI ' \
, I
0.1
100
f ld-t
200
1 :
1
300 400 500 600 700 800 1 000 1200
WAVELENGTH (nm)
(b)
Figure 5.35 Typical spectral responses of common photocathode materials.
11
(Reprinted
by permission of Hamamatsu Photonics K. K.)
During the process of junction formation, the following events occur:
(1) Free electrons in the n region are attracted to the positive charge in the p
region and drift over.
(2) Free holes in the p region are attracted to the negative charge in the n
region and they drift over.
(3) Carrier drift leaves the n region with a net positive charge and the p
region with a net negative charge. The crystal stays neutral with no net
carrier gain or loss.
In the n-type material (dopants are As, Sb, and P), the electrons are the
majority carriers and the holes are the minority carriers. In the p-type material
(dopants are AI, B, In, and Ga), the holes are the majority carriers, and the
electrons are the minority carriers. Majority carriers are far more mobile than
minority carriers, and they are the primary contributors to current flow.
The barrier height depends upon the donor and acceptor levels and
concentrations. This is shown schematically in Fig. 5.36. The region between the
nand p regions is called the depletion region, and there is an electric field across
it.
The barrier height is calculated as
where
n" = electron concentration in the n-region (majority carriers),
Pp = hole concentration in the p-region (majority carriers), and
n/ = the intrinsic carrier concentration.
NTYPE
E, ------ DONOR lEVEL_
VALENCE BAND
DEPLETION
REGION
CONDUCTION BAND
ACCEPTOR lEVE
PTYPE
Figure 5.36 Energy levels in a p-n junction.
(5.101)
Eg
188 Chapter 5
CONDUCTION BAND
N TYPE
/
E,
- DONOR LEVEL
1\
Eg
I-
Vf
....
I--- ACCEPTOR LEVEL",!
./
!cp..Vf
VALENCE BAND
DEPLETION
PTYPE
REGION
1IIIr
Vf
Figure 5.37 Application of forward bias to a p-n junction.
Since ni is an extremely strong function of temperature, so is <1>. As T
increases, decreases. In the equilibrium junction, the Fermi level is constant.
If we apply an external bias across the junction, we can change the energy-
level structure. Application of a forward bias to the p-type region reduces the
barrier height by the amount of applied voltage ( V
f
in in Fig. 5.37). The positive
terminal of the bias source attracts carriers from the other side of the junction (n-
type) and vice versa. The consequence is a high current flow due to conduction
by majority carriers. Lowering the barrier height reduces the depletion region.
Application of a reverse bias (Vr in Fig. 5.38) to the n-type region increases
the barrier height by the amount of applied voltage. The positive terminal of the
NTYPE
Vr
DONOR LEVEL
VALENCE BAND
DEPLETION
REGION
CONDUCTION BAND
ACCEPTOR L E V E ~
PTYPE
$+Vr
- - - - - - - - ~ + 11
1
11---------'
Vr
Figure 5.38 Application of reverse bias to a p-n junction.
bias source repels carriers from the other side of the junction (n-type) and vice
versa. The consequence is a low current flow due to conduction by minority
carriers. Increasing the barrier height widens the depletion region.
The equation expressing the 1- V characteristic of a p-n junction diode is
derived from a continuity equation:
I =1 [e;; -1)
do'
(5.102)
where
q = electronic charge,
k = Boltzmann's constant,
T = absolute temperature in degrees kelvin,
V = applied voltage,
~ = a "constant" to make the equation fit the data, sometimes called the
"ideality" factor, and
10 = reverse saturation current.
At 300K, q/kTis equal to 38.7. The "constant" ~ varies with applied voltage.
It is typically I, but can be as high as 3.
Curves for the I-V equation are shown in Fig. 5.39 for various values of 1
0
, A
large 10 yields a large reverse current and a small forward voltage drop and vice
versa.
CURRENT vs. VOLTAGE for VARIOUS 10
1.00E-07 r------,----,----,-----,---- -,--,..,.-.,..,-'TT---.,
5.00E-08
O.OOE+OO t = = = ~ ~ = = = = = = = t = = = = f = = = ~ ~ ......
-S.00E-08 '-------''------'------'------'-----'-----'-------'
-0.5 -0.4 -0.3 -0.2 -0.1 o 0.1 0.2
VOLTAGE
Figure 5.39 Current-voltage curves for a p-n junction with various ,
0
.
190 Chapter 5
The expression for the reverse saturation current 10 is made up of terms
involving the minority carrier concentrations, the Einstein diffusion constants,
the minority carrier diffusion lengths, and the minority carrier lifetimes:
where
q = electronic charge,
np = minority carrier (electrons) concentration in the p-region,
Dn= Einstein diffusion constant for electrons,
Ln = minority carrier (electrons) diffusion length in the p-region,
pn = minority carrier (holes) concentration in the n-region,
Dp = Einstein diffusion constant for holes,
Lp = minority carrier (holes) diffusion length in the n-region, and
Ad = detector area.
The Einstein diffusion constant D is defined as
D = kT ).L,
q
(S.103)
(S. 104)
and has units of cm
2
/s with Il being carrier mobility. Like D, it may be
sUbscripted with n or p to specify electrons or holes. The minority carrier
diffusion length is
REVERSE SATURATION CURRENT VS. TEMP
0.000 1 F; ..'1".. .. .... r i ..,...."i ..I .., ..F .. ' .... Ti T ....r ..i .. .....i..... ...r....F .....'1' ..., , .. . ,., .., ...... , ....... , ........ . - ,
1E-06
1 E-09 . ..-...J.---'--'
260 270 280 290 300 310 320 330 340 350 360 370 380 390 400
TEMPERATURE
(S.10S)
Figure 5.40 Reverse saturation current versus temperature for a typical p-n junction.
.....
where 1[ is carrier lifetime. The reverse saturation current is strongly temperature
dependent as shown in Fig. 5.40.
Note that Figs. 5.39 and 5.40 combine to show that as the temperature
increases, the saturation voltage decreases. Thermalization losses will heat the
diode or PV cell. The result is that the voltage goes down, the cell current goes
up slightly, and the power goes down.
Incident optical radiation generates a current through the diode, which adds
to the dark current and shifts the I-V curve downward, as seen in Fig. 5.41. The
expression for the total current becomes
I = I (e;:;' -lJ -I
o g'
(5.106)
where Ig is the photogenerated current, expressed as
A
I = llq<P = llq-<P .
g q he
(5.107)
As seen from Eq. (5.107), the current generated is directly proportional to
photon flux and reduced by the detector's quantum efficiency. At a given
wavelength, the current is also directly proportional to the incident power, and
also directly reduced by quantum efficiency. A typical set of I-V curves for
various incident power values is shown in Fig. 5.41.
There are several ways in which one can operate a photovoltaic detector. The
first, and one of the most important for radiometry and photometry, is called
short-circuit-current mode. In a practical sense, this is accomplished by
connecting the detector to a transimpedance amplifer, which is easy to make with
1E-05
SE-OB
BE-OS
4E-06
I- 2E-OS
z
w
a: 0
a:
::::>
() -2E-OS
-4E-OS
-6E-06
-SE-OS
-lE-05
--
DARK
J
UL LITE
MEDLITE
LOTSA LITE )
-2.00 -1.75 -1.50 -1.25 -1.00 -0.75 -0.50 -0.25 0.00 0.25
VOLTAGE
Figure 5.41 I-V curve for a photodiode at several light levels.
192
<J
0
>
Chapter 5
for Ig 1
0
, it's logarithmic
0.15 r--------,------.......,.------,--------,
0.1
0.05
,
PERFECT
LOGARITHMIC
o L = = = ~ ~ ~ ~ L ~
0.01 0.1 1 10 100
RATIO (/gllo)
Figure 5.42 Output voltage versus current for Ig 1
0
,
an op amp and feedback resistor. This mode is represented mathematically by
setting V= 0 in Eq. (5.106). The result is thatl= -Ig, in which current generated
is a linear function of incident radiant power. This linearity is easily
demonstrated to seven decades in radiometric quality silicon photodiodes, though
unsubstantiated claims place the number of decades at 14.
Open-circuit voltage is another operating condition for a photovoltaic
detector. This is represented mathematically by setting 1= 0 in Eq. (5.106) and
solving for V:
(5.108)
If Ig 10, which is almost always the case, then Vac is logarithmic with
radiant power as seen in Fig. 5.42. If 19 10, then Voc becomes linear with
incident power as shown in Fig. 5.43. If 10 ~ 1
9
, operation is intermediate between
linear and logarithmic.
Photovoltaic detectors are also commonly used as power generators. Placing
a load resistor RL directly across the detector causes the 1- V curve to enclose an
area, as seen in Fig. 5.44. Short-circuit current, graphically depicted as the
location where the 1- V curve crosses the current axis (V = 0), and open-circuit
voltage, where the 1- V curve crosses the voltage axis (1 = 0), form the two
extreme points of the (inverted) 1- V curve; in between, both current and voltage
are available simultaneously. This is the requisite condition for power generation,
and devices operating in this region are commonly called solar cells. The lower
curve in Fig. 5.44 is the power versus voltage characteristic for a particular cell,
in which the maximum power is achieved at about 0.48 V.
for Ig 10, it's linear
PERFECT
LINEAR
0.1 ___ --+--------t--""""""-"7"''''-,...-t---== __ --=j
0.001\---""7""C-.....-+----_---t-_----,...-t------.........J
0.0001 '--_____ -L-_____ -'-_____ --'-_____ --'
0.01 0.1 1 10 100
RATIO (Igllo)
Figure 5.43 Output voltage versus current for Ig ,
0
,
193
The open circuit voltage Vac is a result of the forward-biased rolloff seen
generically in Fig. 5.39. For silicon, this voltage is close to 0.6 V as in Fig. 5.44.
The short-circuit current Ig is proportional to the incident radiation as shown in
Eq. (5.107), and the radiation is in tum proportional to the area of the cell. The
cell in Fig. 5.44 with an Ig of 50 rnA is quite small with an area of about 1 cm
2

Cells for typical commercial solar panels or modules have an area of around 100
or more cm
2
, producing a short-circuit current of around 5 A. A typical panel has
perhaps 72 cells connected in series to produce an open-circuit voltage near 40 V
and a short-circuit current near 5 A. The ratio of the maximum power to the
product VaJ
g
is called the "fill factor," (not to be confused with fill factor in
CCDs, below) which is typically on the order of 0.75 or so.
SOLAR CELL OPERATION
60.--,--,---,--,--.--,,--.--.--,---.--.--.
I
- - "- r _. _. - - I - - - - - - --. - - - - - - - - 1- - - - - - - r- - - - - - - T - - -
I I
__ __
---"---

- - - - - - - - - - - - - - - - - - - / - - - - - - -\ - - -

---4--- /-----:---- --v---
20
/ ___ '- ______ , __ _
..
o 0.1 0.2 0.3
VOLTS
0.4 0.5 0.6
Figure 5.44 Photovoltaic solar cell operation (lower curve) and I-V characteristic (upper
curve).
194 Chapter 5
2E-06
.
,
1.5E-06
,
.'
....
,
..
1E-06
, ... .
,
. '
.

I-
5E-07
, . ..
z
w
0::
0
0::
:::>
u
-SE-07
-1E-06
, ... '.
.... Vb DARK
.........
,
r---.:..
.' .
,
.
.

,
,
.
,
,
-1.5E-06
,
LITE

,
LOAD LINE "'-.
'-......
.'.
,,' 1:
1
-2E-06
-10.00 -9.00 -8.00 -7.00 -6.00 -5.00 -4.00 -3.00 -2.00 -1.00 0.00 1.00
VOLTAGE
Figure 5.45 Reverse-biased photodiode /- V behavior.
The photovoltaic detector may also be operated in reverse-bias mode, in
which 1 = I
g
. This mode of operation requires an external voltage source and
a load resistor, both of which contribute to noise. An 1- V curve for this mode of
operation is seen in Fig. 5.45, showing the load line of the resistor and the bias
source VB. In dark conditions, the diode resistance is very high, and most of the
voltage drop occurs across it. With the addition of light, the diode resistance
drops, and a higher fraction of the applied voltage is found across the load
resistor. Note, however, that the curves in Fig. 5.45 are idealized, and that a real
diode will eventually break down for sufficiently negative voltages due to the
Zener effect. There are advantages and disadvantages of operating the
photodiode in reverse-bias mode. Some of these are detailed in Table 5.19.
Most manufacturers refer to reverse-bias operation as operation in
"photoconductive mode." This term is confusing, as photovoltaic and
photoconductive detectors operate very differently. More accurate terms are
"reverse-biased" or "back-biased."
Table 5.19 Advantages and disadvantages of reverse-bias operation.
Advantages Disadvantages
Better long-wavelength response due
to less recombination
Increased speed due to E-field, which
sweeps carriers out
Increased speed due to lower junction
capacitance C=const
< n <
Presence of 10 and its temperature
dependence
Requires relatively stable, quiet voltage
source and low-noise RL
Lower SNR due to noises from RL and
VB
A common approach to operating a photovoltaic detector in the short-circuit
mode is to connect it to a transimpedence amplifier, which is easy to make using
an op amp and a feedback resistor.
5.6.4.2 Responsivity and quantum efficiency
Figure 5.46 depicts the ideal and real current responsivities for a photovoltaic
detector. Recalling that the current responsivity is equal to the current out of the
detector divided by the input power, or equivalently 9\ .= 11qA(hcr
l
, it is apparent
,
that non ideal (less than 1) quantum efficiency 11 drives the nonideal response.
There are a number of reasons why the quantum efficiency is less than 1,
including:
(1) Unabsorbed photons of wavelength A> Ac (i.e. photons at longer
wavelength than the cutoff wavelength)
(2) Unutilized electron-hole pairs created beyond diffusion length (depletion
region can be widened via reverse biasing the photo diode )
(3) Surface recombination of carriers (can be reduced with a dielectric
coating)
(4) Optical losses due to transmission and reflection (can reduce with
antireflection coating)
(5) Heating of the device due to the fact that most photons have more energy
than needed to create an electron-hole pair
(6) All efficiencies are less than one.
I
I I I I
- - - 1- - - - - - - -1- - - - - - - - I - - - - - - 1 - - - - - - T - - -
I
0.8 f-------'--t----'---t---'--+---'---+--'--'"7"i
I I
.. 1- - - ,- - - - - 1- - - - - - - -I - - -
I I

0.6
:> - - -:- - - - - - - 1- - -
Ui I
B 0.4 __ '-----1
c..
If)
w
n:
I
- - -1-"'
0.2 f-------'--t..,...71"<:::...--t---'--+---'---+--Yr-I
I
- - -.,---
o 0.2 0.4 0.6 0.8
WAVELENGTH (urn)
Figure 5.46 Ideal (straight line) and actual (curved line) current responsivity versus
wavelength.
196
5.6.4.3 Noises in photovoltaic detectors
Noise sources in photovoltaic detectors include the following:
(1) shot noise from current flow across the potential barrier,
(2) Johnson noise, and
(3) lifnoise.
Chapter 5
The most important noise source is shot noise. Shot noise current arises from the
signal, background, and the device's own dark current. The rms shot noise
current is
. [2 (.. . )B]1I2
Ishol = q Is + Iback + Id
(S.109)
Johnson noise can be minimized by choosing a detector with high dynamic
resistance Rdyn which varies with voltage. Cooling the detector reduces Johnson
noise due both to the decrease in temperature and the increase in R
dyn
with
decreasing temperature:
i ~ t k T B .
Rdyn
(S.110)
The dynamic resistance is low at forward bias and high at reverse bias. It is
calculated as
(S.lll)
At the important zero-bias operating point it becomes
R = dV = ~ k T
dyn d'l 1
q a
(S.l12)
Although it appears that Rdyn is linear with temperature, the exponential
nature of I as a function of temperature predominates, and the dynamic resistance
increases with decreasing temperature. Longer wavelength photodiodes have
higher la, lower R
dyn
, and are noisier.
The insidious lifnoise is troublesome, as always. It cannot be avoided due to
signal current, but can be minimized by operation at a higher frequency due to
the following relation:
. _ s(constant) (Idcdf)
IIff -
[ ]
1/ 2
f
197
(5.113)
To calculate D*BLlP for the photovoltaic detector, we need an expression for
signal current. Recall that is 9t where 9\ . is the current responsivity.
, ,
Expressed in tenns of power,
t..,
is =l1Q-.
he
(5.114)
Noise current, which is shot-noise dominated as stated above, may be
expressed as a function of the signal current, in the visible portion of the
spectrum:
. (2' B)112
lshot = qls ,
(5.115)
or as a function of the background, in the infrared portion of the spectrum:
(5.116)
Inserting the expression for the signal current and the shot-noise current Eq.
(5.116) (for the infrared) in Eq. (5.8) for D* results in
D* (t.., f) - 11t..,
BLIP , - 2E h '
back e
(5.117)
which is the background-limited expression, applicable in the infrared, where
Eback refers to the background irradiance. Similar logic may be followed for
photon flux, beginning with Eq. (5.114). In that case,
D
BLIP
(/1.,/) = - .
'I t . . ~
he 2E
q
,back
(5.118)
Comparing Eqs. (5.89) and (5.117), and (5.90) and (5.118), we see that the
photovoltaic detector is better (higher D*) than the photoconductive detector by a
factor of Ji. The physical reason for the difference is the absence of G-R noise
in the photovoltaic detector. Recombination takes place within the
photoconductive detector itself; in the photovoltaic detector, it occurs in an
external ohmic circuit where carriers are not statistically correlated. An exception
to this rule arises when photoconductive detectors are operated in "sweepout"
198 Chapter 5
mode. In this case, carriers are swept out of the detector before they recombine,
and the D* is equal to that of a photovoltaic detector.
5.6.4.4 Photovoltaic detector materials and configurations
Table 5.20 lists some materials that have proven useful for photovoltaic
detectors. There are a number of configurations for photovoltaic detectors. For
high-speed operation, a conventional photodiode requires a large reverse bias to
decrease device capacitance and create a large drift field. Table 5.21 lists several
common configurations; more detailed information about these may be found in
the references at the end of the chapter. Moreover, photovoltaic materials are
increasingly finding application in solar-cell technology. Some of these materials
are listed in Table 5.22.
Table 5.20 Some photovoltaic detector materials.
Material Bandgap (eV) Center wavelength (pm)
GaP 2.4 0.52
GaAs 1.4 0.93
Si 1.12 I.l
InGaAs 0.73 1.7
Ge 0.68 1.82
InAs 0.28 3.5
InSb 0.16 5.5
HgCdTe variable variable
PbSnTe variable variable
Table 5.21 Common photodiode configurations.
Configuration
p-i-n
Avalanche
Schottky barrier
Inversion layer
Ultraviolet
Infrared
Position sensing
photodiodes (PSPD)
Comments
Built-in depletion region, with an intrinsic layer of Si
between p and n materials
Operate under a large reverse bias; fabricated from Si, Ge,
and InGaAs. Can be cooled to 77 K and biased beyond
breakdown point.
Created by depositing a thin semitransparent metal electrode
on top of a semiconductor material. Particularly useful for
large-area UV detectors.
Created by doping top layer on p-type silicon with a material
having a positive charge to form an n-type material.
Quantum efficiency approaches 1 at short wavelengths.
Include Schottky barrier, front-illuminated PtSi, and
AlxGax_1N
Include InGaAs, InAs, HgCdTe, and PbSnTe. Require
cooling beyond 3 !lm.
Output as a function of position on the detector, often used in
tracking applications.
jiIW"
Table 5.22 Photovoltaic materials for solar cells.
Material
Monocrystalline silicon
Poly- or multicrystalline silicon
Ribbon silicon
New structures based on silicon
Cadmium telluride (CdTe)
CIGS
CIS
GaAs multijunction
Amorphous silicon
5.7 Imaging Arrays
5.7.1 Introduction
Comments
nla
nla
Formed from flat thin films of molten silicon
Special silicon arrangements to improve efficiency
Large-scale manufacturing and delivery possible
Copper indium gallium diselenide
CuInSe2
High efficiency; have been used in satellites
nla
Many applications require spectral and spatial information that the use of a
single-element detector cannot reasonably provide. In order to obtain this
information, some imaging systems use single detectors along with scanning
optics and other components. More commonly, linear or area (2D) arrays of
detectors are used, with the area array in widest use. A comparison of the
functional differences of single and multiple detectors is shown in Table 5.23.
Array imagers are relatively recent attempts to emulate human vision. Other
innovations throughout history are listed in Table 5.24.
This section will review concepts important to array detection, including
history, basic array parameters, device architecture, and specific array types
along with their applications. Due to the large and growing amount of material on
this subject, the reader is strongly encouraged to investigate the references at the
end of the chapter.
5.7.2 Photographic film
5.7.2.1 History
One of the earliest man-made optical detectors was photographic film. The
treatment here is brief because film has been supplanted in most scientific and
consumer applications by solid-state electronic detectors, as described in this
chapter. It is still used, however, in some specialized applications. For more
information on photographic film, the reader is referred to two thorough books
on the application of photography in science and engineering (Ray, 1999) and the
theory and technology of photography (Stroebel et aI, 2000).
200 ChapterS
Table 5.23 Comparison of single- and multiple-detector capabilities.
Single detectors
Gather information from one
wavelength or field of view at a time
Provide instantaneous measurement
Require a simple electronic interface
Need complex scanning scheme due to
the presence of a single element
One detector, one response
Multiple detectors
Gather simultaneous measurements in
multiple spectral bands, multiple
fields of view, or both
Utilize integration over time for all
pixels in array
Require more complex electronic
interface (serial or parallel)
Simple or no scanning possible; array
with multiple elements can be
utilized as a staring array
Magnitude of response varies due to the
presence of multiple detector
elements and nonuniformities in the
array
Table 5.24 Vision emulation efforts.
Imaging mechanism
Film photography
Mechanical scanning systems
Early television
Remote sensing systems
Imaging vacuum tube devices
Solid-state array imagers
Time period
19
th
to 20
th
century
20
th
century
1930s and onward
1970s and onward
Johann Heinrich Schulze discovered that silver nitrate (AgN0
3
) darkened
when exposed to light and published a scientific paper describing the
phenomenon in 1727. The technologies behind photographic film were
developed over several decades in the 19
th
century. The earliest "permanent"
recorded film image is generally credited to Joseph Nicephore Niepce in 1827.
The photograph required an eight-hour exposure to sunlight to form. Various
pioneers, including Louis Jacques Mande Daguerre and Henry Fox Talbot,
further developed the technology, eventually allowing image recordings onto
paper. Toward the end of the 19
th
century, George Eastman invented roll film and
factory processing, thus commercializing the technology for routine consumer
use. Photographic film was the most prevalent image recording technology for
the next 100 years or so, when electro-optical detectors began to dominate in
most fields of science and consumer applications.
5.7.2.2 Physical characteristics
Film is a "sandwich" of materials, whose primary layers are a stable base of
plastic or glass topped by a gelatin emulsion layer containing silver-halide
crystals. The base thickness is in the range 50 to 100 11m, and the emulsion layer
is about 5 11m thick for black and white film and about 15 11m thick for color
film. Various refinements are included by additional layers to protect the
emulsion, adhere the emulsion to the base, and reduce backscattering from the
rear surface of the base. The film is exposed to light by a camera shutter, and
subsequent chemical processing (development) converts the silver-halide crystals
to silver, thus darkening the image in proportion to the amount of exposure (for a
negative film).
The spectral sensitivity characteristics are controlled by the chemical
characteristics of the emulsion. The radiometric calibration characteristics of the
final image are determined in part by the emulsion, but also by the chemical
processing of the exposed film.
5.7.2.3 Spectral sensitivity
Black and white (panchromatic) films typically respond to light wavelengths of
about 0.4 to 0.65 11m. The sensitivity for a particular type of film depends on the
details of its construction, chemical compounds, and processing chemistry.
Kodak has developed films sensitive to near-IR wavelengths out to about 0.9 11m;
such films are able to record the high reflectance of vegetation in the near IR.
Color films have three different sensitive layers in the emulsion that respond
to red-, green-, and blue-light wavelengths. Development of the exposed film
results in red, green, or blue transparent dye in the corresponding layer of a
positive (slide) film, or cyan, magenta, or yellow in the corresponding layer of a
negative film.
The fairly broad spectral response of film cannot be easily refined by
chemical modifications, so lens spectral filters are commonly used to limit the
spectral response. For example, a yellow (haze) filter is used to remove blue
scattered skylight, or a blue-green blocking filter is used to restrict IR film to the
red and near-IR regions.
5.7.2.4 Radiometric calibration
A scientific procedure for calibration of film was established by Hurter and
Driffie1d in 1890. They proposed a series of regular exposure steps (total
exposure is irradiance multiplied by exposure time) onto film and measurement
of the resulting optical density D, defined as the negative logarithm of the film
transmittance T:
D=-log(T) . (5.119)
A plot of D for each step versus the logarithm of the corresponding exposure
10g(E) is known as the "D-logE' or "H-D" (for Hurter Driffield) curve. The H-D
202 Chapter 5
curve essentially shows output (optical density) versus input (log exposure)
calibration. It depends not only on film type, of course, but also on processing
conditions (chemical temperatures, processing time), exposure spectral
characteristics, density measurement procedures, and many of the variables
associated with photography. The main characteristics of the H-D curve are a
saturation toe at low exposure (the "base + fog" level, typically a D of about 0.1),
a linear segment for moderate exposure, and a saturation shoulder at high
exposure (typically a D of about 3.0). The equation of the linear segment is
D = ylog(E) + Do, (5.120)
where Do is the projected intercept (usually negative) on the density axis (not the
"base + fog" level), and gamma y is the gain of the film, synonymous with its
contrast; i.e. a low gamma means a low-contrast film. The H-D curve describes
"macrocalibration" for relatively large-area measurements of a millimeter or
more. If calibration of film is attempted for "micro" conditions of measurement
of a tenth of a millimeter or less, other factors come into play, for example, the
adjacency effect where nearby exposure affects the density at a point of interest.
5.7.2.5 Spatial resolution
The spatial resolution of film is not defined geometrically, as with electronic
detector arrays. Film resolution is affected by light scattering within the
emulsion, size of the developed silver grains, thickness of the emulsion, etc.
Consequently, it has been traditionally measured by exposure of a specific target
(for example, three high-contrast bars in a series of decreasing spacing and
width) and determination of the smallest target that can be visually identified.
A modulation transfer function (MTF) of film can be measured by exposing
a series of sine wave targets of varying frequency and contrast. However, the
intrinsic nonlinear characteristics of film are problematic for the application of
MTF, which assumes a linear imaging system. For example, the adjacency effect
causes the MTF to go above one for low- to midrange spatial frequencies.
5.7.2.6 Summary
Film remains a unique image recording mechanism in that the recorded and
developed image is the archival medium itself, i.e. no additional processing is
necessary to save the original image for long periods of time. It is also an
efficient detector in terms of its combination of large format and high resolution.
For example, a large-format aerial photograph 10 x 10 in. with a resolution of 10
)lm contains some 645,160,000 resolution cells, or pixels. At the time of this
writing, such large monolithic electronic detector arrays are not possible
(although mosaics of individual arrays can achieve this size).
However, the inconvenience, delay, and cost of chemical processing and
conversion to a digital format by scanning and the associated quantitative
difficulties in image measurements have seriously disadvantaged film relative to
--
electronic detectors for applications requiring quantitative measurements. Steady
improvements in the spatial density of detector arrays and computer processing
capabilities continue to erode the role of film in scientific imaging.
5.7.3 Electronic detector arrays
5.7.3.1 History
One of the earliest array devices was the linear photo diode array marketed by
EG&G Reticon. It used silicon photodiodes in the photon flux integrating mode
coupled with field-effect transistor (FET) switches. These were driven by
clocked shift registers and stored electrons in the device capacitance until
readout. During the readout phase, electrons were transferred to an output signal
amplifier. Similar linear-array devices today find application in spectroscopy,
astronomy, grocery and department store scanners, and many other products.
The first large-area imaging arrays were built in the late 1960s using x-y
addressable photodiodes, phototransistors, and photoconductors. They were not
particularly successful due to responsivity nonuniformities and spatial noise. In
addition, only instantaneous readouts were employed, and the signal could not be
integrated over time. Advantages were good area utilization and random access
to pixels within the array.
Charge-coupled devices (CCDs) allowed integration of photosensitive
elements with complex readout mechanisms. Photodetectors within a CCD
operate in integration mode, with the outputs serially clocked at high rates to a
single readout circuit. CCDs having on the order of 100 megapixels (MP) have
been fabricated; this number will increase with technology development over
time. Some of the parameters characterizing imaging detector arrays are listed in
Table 5.25.
5.7.3.2 Device architecture description and tradeoffs
Although CCD technology has achieved focal plane arrays with high pixel
counts, high fill factors, high sensitivities, and charge-transfer efficiencies greater
than 0.99999 (the fraction of stored charge transferred out of the array),
complementary metal oxide semiconductor (CMOS) technology competes
effectively with CCDs in the large-format array market. This is due primarily to
the ability of CMOS manufacturing to take advantage of already-existing
fabrication lines developed for commercial microprocessors. As a result, the cost
of manufacture is lowered and the production yield of CMOS-based imagers is
improved, relative to imagers based on CCD technology.
CMOS devices can be active pixel sensors, employing signal processing
functions, including amplification, on chip. This configuration requires more
hardware, however, and reduces the fill factor from approximately 80% (in a
passive-pixel CMOS) to between 30 and 50%. To increase the amount of light
204
Table 5.25 Array parameters.
Number and size of individual detector elements
Overall size of array
Format (aspect ratio) of array (e.g. 4:3, 16:9)
Instantaneous field of view (IFOV) and array field of regard
Fill factor (portion of an array sensitive to light)
Scanning or staring array
Spatial resolution (MTF)
Responsivity
Spectral characteristics
Dark current
Integration time
Noise amount and type (fixed pattern, readout, etc.)
Dynamic range
Linearity
Pixel data rate, frame rate, bandwidth
Chapter 5
Noise-equivalent temperature difference (NETD) for infrared arrays
Pixel response uniformity
Information capacity
Charge transfer efficiency (for CCDs)
Long-term storage potential of detector elements
Auxiliary equipment necessary to provide maximum utilization of
information
detected, and minimize the possibility of light falling onto nonphotosensitive
areas of the chip, arrays of micro lenses that focus the incoming radiation are
often used.
Charge-injection devices (CIDs) are also finding increasing use in large-
format arrays. They've been around since the 1970s, but exploitation of their
unique properties has only recently gained momentum. One of these properties is
the capability to stare at scenes for long exposures without spillover of charge
into adjacent pixels ("blooming").
CID pixels are individually addressable, with charge dumped (injected) to the
substrate after each read. By contrast, CCD readouts involve transfer of charge
from site to site during the readout process.
5.7.3.3 Readout mechanisms
CCD. CCDs can utilize one of several readout mechanisms; the three described
below are most common. In a full-frame architecture, the simplest to make and
operate, serial and parallel shifts move the charge through the array to an output
amplifier. This is shown schematically in Fig. 5.47. Rows of charge are shifted
Parallel
clocks
Serial clocks
II
SERIAL
REGISTER
FULL
ARRAY
Parallel f
shift direction
Analog
Outpul
Figure 5.47 Full-frame readout.
205
upward to the serial-shift register in time with the parallel clock signals depicted
at left. Upon reaching the serial-shift register, charge is read serially into the
output amplifier. The process is repeated until all rows are read out. It requires
some form of shuttering of the array, so that light collection does not occur
simultaneously with charge transfer.
Frame transfer architecture, depicted schematically in Fig. 5.48, is similar to
the full-frame architecture with the inclusion of a storage array. As in the case of
the full frame, rows of charge are shifted toward the serial register, where charge
from each element is read out sequentially in accord with serial clock signals.
The storage array seen in Fig. 5.48 is not photosensitive, and light collection and
integration in the image array can occur simultaneously with integration in the
storage array. As a result, there is no need for a shuttering scheme.
A weakness of this mechanism is that integration within the image array is
still occurring while transfer to the storage array takes place, resulting in image
smear. Because twice the silicon is used as in the full-frame device, a frame-
transfer device costs more. On the other hand, higher data rates are achieved.
In the interline transfer architecture, shown schematically in Fig. 5.49,
photosensitive areas and readout registers are arranged in successive columns.
After detection and integration takes place in the photosensitive area, charge is
transferred to column registers, which are then clocked to the serial register as in
the previous two architectural techniques. This architecture allows very fast
response time, but is comparatively difficult to fabricate. Although image smear
as in the frame transfer device is not entirely eliminated, the amount of smear is
reduced.
206
Parallel
clocks
(storage)
Parallel
clocks
(image)
Serial clocks
"
Parallel t
shift direction
Figure 5.48 Frame-transfer readout.
ChapterS
C/D. Pixels are individually addressable in a CID. The charge remains intact in
each pixel after its signal level has been determined, making for a nondestructive
readout. Rowand column electrodes are shifted to ground to make way for
collection of the next image frame, and the charge is "injected" to the substrate.
This capability for individual pixel control is valuable in many imaging
applications. For instance, long exposures to low-light-level sources can allow
optimum exposure of a particular target, such as in astronomical use. If other
objects of interest within the sensor's field of view appear during the long
integration, the pixels containing such images can be read out while the low-
light-level target continues to integrate on the array. At brighter target levels,
blooming is less likely to occur than in the CCD, due to the fact that charge
overload is confined to a single pixel.
Para
cloc
:::
lIel
ks
Photosensit
area
ive
--
Serial clocks
U
SERIAL
REGISTER
u
RR
rr
I-
t al09
utpul 0
lin terline
mask
Parallel 1
shift direction
Figure 5.49 Interline-transfer readout.
CMOS. Like CIDs, CMOS devices are x-y addressable, allowing individual pixel
information to be read out without reading out the entire array. A disadvantage of
CMOS is high readout noise unless an amplifier is part of the configuration of
each pixel (that is, unless the device is an active-pixel sensor) which, as stated
above, reduces the device fill factor. Additional signal processing circuitry
providing functions such as thresholding, edge detection, and motion detection
can be incorporated into the CMOS chip, allowing CMOS to compete on a
performance level with scientific CCDs.
In addition to the common monolithic CMOS structure (detector and readout
together on one chip) CMOS devices are available in hybrid form (two chips
bump-bonded together). The latter structure allows a 100% fill factor in the
visible spectrum.
5.7.3.4 Comparison
When presented with a choice of a camera based on one of these three
technologies-CCD, CID, or CMOS--one must consider one' s application
relative to the technology and its cost. As in most radiometric system and
measurement problems, there are tradeoffs to be made. While not purporting to
offer the final word on technology selection, Table 5.26 presents a comparison of
some of the characteristics of these technologies. This comparison is by necessity
general; many specific tradeoffs can be made to enhance performance and reduce
cost, depending upon application.
5.7.4 Three-color CCDs
CCD imagers employing three-color arrays having red, green, and blue filters can
provide color imagery in the visible portion of the spectrum. Though this
technique is not as effective as film photography in rendering true color images,
it finds wide utility in digital photography, in which color images may be quickly
available. Typical array formats are large, though not nearly as large as standard
CCDs: Kodak, for example, has developed a 16.6-MP array camera, and larger
arrays feature 25 MP. 12
A four-square "checkerboard" filter pattern is often used on the focal plane to
provide the red, green, and blue response channels. These schemes typically
include two out of four pixels green, and one each blue and red. The sensor
therefore gathers 50% of the green light over the area of the filter pattern, and
25% each of blue and red light. Substantial post processing is required to provide
all three colors at all pixel locations; arguably the most famous technique
employed for this purpose is the Bayer algorithm (U.S. patent No. 3971065)
which may also be utilized in cheaper CMOS-based cameras in addition to other
devices.
As one might expect, radiometry performed on data that have undergone
significant processing will be complex due to the number of error sources. At this
time, absolute radiometry performed using three-color CCDs is problematic at
208 ChapterS
Table 5.26 CCO, CIO, and CMOS comparisons.
CCD CID CMOS
Array fill Can reach 100% 80% or better Can range from
factor 30% to 100%
depending on
configuration
Sensitivity Highest High when Not as high as CCD
successive
frames are
integrated
Pixel-to- High Not as high as Can be high if
pixel CCD signal processing
uniformity is on chip
Blooming Can be a problem Excellent Needs improvement
control
Fabrication Requires specialized Uses standard Takes advantage of
processmg CMOS existing
processmg semiconductor
fab techniques
Advantages Well-established Can be used in low- Many on-chip
technology, well and high-light- signal-processing
characterized, level applications; features possible
many scientific excellent
uses blooming control
Problems Blooming control Noisy Readouts can be
nOIsy
best. The situation is changing, however, as digital cameras are being developed
in competition with film cameras for photogrammetric purposes. Because of the
application, these digital cameras would have to offer better radiometric
sensitivity than that obtained using digitized film images and would provide the
capability for self calibration. 13
Alternatively, three separate CCDs may be used to provide color imaging,
with filters designed to pass radiation in the red, green, and blue portions of the
spectrum. This approach involves more hardware than the one discussed above,
but will become more cost effective as the price of this technology comes down.
5.7.5 Ultraviolet photon-detector arrays
Ultraviolet detector arrays typically utilize hybrid architectures. They commonly
employ arrays of photodiodes, though this is not always the case. Many arrays
utilize aluminum gallium nitride (AIGaN) as the detector material due to its
sensitivity in the UV and insensitivity at longer wavelengths. Flame detection,
astronomy, and undersea communications are some applications that benefit from
these devices. Current technical challenges include tailoring the cutoff
wavelengths of AIGaN p-i-n photo diodes by controlling alloy composition,
allowing AlxGal_xN arrays to cut off between 227 nm and 365 nm.
5.7.6 Infrared photodetector arrays
Infrared photodetector arrays are utilized over all regions of the infrared
spectrum, providing detection in the near infrared (NIR), short-wave infrared
(SWIR), and long-wave infrared (L WIR). L WIR devices have been successfully
used for many years in sensitive military applications, as have SWIR-based
systems.
Indium gallium arsenide (lnGaAs) arrays are used in the NIR to provide
alternatives to more traditional night-vision devices based on image intensifier
technology, as well as many other applications. InGaAs arrays may be comprised
of p-i-n photodiodes or avalanche photodiodes (APDs.) Varying the chemical
composition of the alloy (InxGal_xAs) allows optimization at a desired
wavelength for maximum signal strength.
Detectors made of lead selenide (PbSe) and lead sulfide (PbS) operate in the
1 to and 3 to portions of the spectrum; these detectors typically
utilize a monolithic architecture. Another monolithic detector is platinum silicide
(PtSi), operating in the 3 to region. PtSi arrays are highly uniform, even if
relatively low response compared to other detectors, and are often used to image
objects with high spatial variation.
Mercury cadmium telluride detectors (HgCdTe) are used in 3 to and
particularly in 8 to regions. They typically employ a hybrid architecture
and have high sensitivity and fill factor. They exhibit substantial nonuniformity
when tailored for the L WIR and require cryogenic cooling to 77 K, increasing the
size, weight, cost, and power consumption of devices. Indium antimonide (lnSb)
arrays are also used in hybrid devices and for scientific and defense applications
in the SWIR and midwave infrared (MWIR).
Signal processing in the element (SPRITE) detectors, commonly used in
forward-looking infrared (FUR) modules of British manufacture, utilize
HgCdTe, but do not have the many leads and preamps that predominate in
HgCdTe two-dimensional arrays. SPRITEs are typically linear arrays that allow
integration of the incoming signal by scanning along the array at the same rate as
the carrier drift. SPRITE-based FUR imagers have been deployed on helicopters,
aircraft, and other platforms for tactical military uses in the 8 to region.
Quantum-well infrared photo detectors (QWIPs) do not have the
nonuniformity problems associated with HgCdTe. Based on GaAs and InGaAs
technology, their manufacture takes advantage of established wafer productions
processes.
210 Chapter 5
Table 5.27 Initial approach to uncooled imaging detector array development.
14
Staring focal plane arrays based on thermal detection mechanisms
Focal plane response at video rates (30 Hz in the U.S.)
Noise-equivalent temperature difference (NETD) of 0.1 K
Resistive micro bolometer staring arrays (Honeywell) and hybrid pyroelectric
detector staring arrays (Texas Instruments)
Use of modem integrated circuit electronic processing technology to enable
development of low-cost multielement arrays
5.7.7 Uncooled thermal imagers
Arrays of thermal detectors operating according to the detection mechanisms
discussed in Sec. 5.5 are finding increasing use in a wide variety of applications.
They are cheaper to manufacture than infrared photon arrays and do not require
cooling.
In the early to mid-1990s, formerly classified research into uncooled thermal
arrays became public. This research, sponsored by the U.S. Defense Advanced
Research Projects Agency (DARPA), the U.S. Army's Night Vision and
Electronic Sensors Directorate (NVESD), and the U.S. Naval Air Warfare Center
(NA WC) provided an approach to the development of uncooled thermal imaging
detector array-based systems. Two U.S. companies were at the forefront of
efforts to develop these systems: Honeywell Corporation in Plymouth,
Minnesota, and Texas Instruments in Dallas, Texas. In addition, several
organizations worldwide, including GEC Marconi of Great Britain, participated
in developing uncooled infrared imaging detector arrays. Table 5.27 lists some
factors important to the initial approach to uncooled detector development.
Thermoelectric array development was also pursued; linear arrays based on
this technology have been used in several radiometric applications including in-
motion inspection of railroad train wheels.
While early 2D arrays provided 320 x 240 focal planes with individual pixel
elements of 50 /lm in size and NETD of 0.1 K (jll, 30 Hz), developments over
the past decade have significantly improved performance. Sensitivity of the
micro bolometer arrays has been improved through improving pixel thermal
isolation and array fill factor. Microbolometer arrays based on vanadium oxide
detector material in a 640 x 480 format are now available, having NETDs of 50
mK or less (jll, 30 Hz) with element sizes of approximately 25 /lm. Packaging
improvements are also in process.
In fact, device-production strategy, effective use of existing fabrication
methods, and development of low-cost fabrication techniques all contribute to
how well a technology will be received in the marketplace. This reception, in
tum, drives research. More "exotic" expensive technology is not always better.
As in so many facets of radiometry, a technology user is wise to trade off costs,
benefits, and level of technical performance required before purchasing a system
or building a product. Table 5.28 lists some applications of uncooled thermal-
imaging devices.
Table 5.28 Applications of uncooled thermal detector array-based systems.
Night-driving systems in automobiles
Roofing inspection
Leak detection from pipes and other fluid and gas delivery/storage
structures
Law-enforcement investigations of buildings, trucks, and other
possible concealment locations
Quality control along industrial process lines
Detection of "hot spots" in electrical wiring to determine possible
locations of malfunction
Aircraft inspection to determine weak points in structure based on
thermal information
Border-patrol surveillance
5.7.8 Summary
211
The above descriptions of array detectors, their uses, and some comparisons are
only an overview. Trading off parameters for a specific application is often
necessary, and there is no particular "right" answer to many detector selection
problems. A thermographer examining a housing structure for sources of heat
leakage can purchase an uncooled thermal instrument for far less cost than a
cooled device and yet receive adequate imagery for that task. An astronomer
observing a distant galaxy or low-light-level astronomical source may find a
cooled photodetector array to be his only reasonable option in the infrared, but
might he choose a CID over a CCD for observations in the visible?
Questions such as these are important for the student or professional wishing
to become adept at systems analysis and systems engineering, two of the most
important disciplines making use of the art of radiometry.
For Further Reading
D. F. Barbe, Ed., Charge-Coupled Devices, Springer Verlag, Berlin (1980).
J. D. E. Benyon and D. R. Lamb, Charge-Coupled Devices and Their
Applications, McGraw-Hili, London (1980).
L. Biberman and S. Nudelman, Photoelectronic Imaging Devices, Plenum, New
York (1970).
I. P. Csorba,Image Tubes, Howard W. Sams, Indianapolis (1985).
G. C. Holst and T. S. Lonheim, CMOS/CCD Sensors and Camera Systems, JeD
Publishing and SPIE Press, Bellingham, Washington (2007).
R. J. Keyes, Ed., Optical and Infrared Detectors, Springer-Verlag, Berlin (1980).
Photomultiplier Tubes: Principles and Applications, Philips Photonics (1994).
Sidney F. Ray, Scientific Photography and Applied Imaging, Focal Press, Oxford
(1999).
G. H. Rieke, Detection of Light: From the Ultraviolet to the Submillimeter,
Cambridge (1994).
212 ChapterS
R. A. Smith, F. E. Jones, and R. P. Chasmar, The Detection and Measurement of
Infra-red Radiation, Oxford, London (1957).
A. H. Sommer, Photoemissive Materials, John Wiley & Sons, New York (1968).
Reprinted by Krieger (1980).
L. Stroebel, J. Compton, I. Current, and R. Zakia, Photographic Materials and
Processes, Second Edition, Focal Press, Boston (2000).
J. D. Vincent, Fundamentals of Infrared Detector Operation and Testing, John
Wiley & Sons, New York (1990).
A. T. Young, Photomultipliers, Their Causes and Cures, Chapter 1 in Methods of
Experimental Physics: Astrophysics, Vol. 12, N. Carleton Ed., Academic
Press, New York (1974).
Several books in the series Semiconductors and Semimetals, R. K. Willardson
and A. C. Beer, Eds., are relevant. Vo1s. 5 and 12 both treat infrared detectors,
and Vol. 11 deals exclusively with solar cells. Vol. 47 deals with uncooled
infrared imaging array systems.
Similarly, the series Advances in Electronics and Electron Physics is rich in
pertinent articles. Numerous volumes deal with imaging detector conference
proceedings, while others have significant feature articles. Most notable are Vols.
34 and 55.
A number of books have one or more chapters dealing with optical radiation
detectors. These include the following:
R. W. Boyd, Radiometry and the Detection of Optical Radiation, John Wiley &
A. Chappell, Optoelectronics: Theory and Practice, McGraw-Hill, New York
(1978).
R. D. Hudson, Infrared System Engineering, John Wiley & Sons, New York
(1969).
D. Malacara, Physical Optics and Light Measurements, Academic Press, New
York (1988).
Several major handbooks offer articles and/or chapters dealing with optical
radiation detectors. These include the following:
J. S. Accetta, and D. L. Shumaker, Eds., Infrared and Electro-Optical Systems
Handbook, Vol. 3, SPIE Press, Bellingham, Washington (1993).
M. Bass, Ed., Handbook of Optics Vol. 1, McGraw-Hill, New York (1995). Parts
5, 6, and 7 contain Chapters 15 through 25, all pertinent to detectors and
detection.
W. L. Wolfe, Ed., Handbook of Military Infrared Technology, Chapters 11 and
12, Office of Naval Research, Washington D.C. (1965).
W. L. Wolfe and G. Zissis, Eds., The Infrared Handbook, Chapters 11-16, ERIM
and SPIE Press (1978).
F
Another rich source for detector infonnation, and probably the best for
assessment of the current state of the art, is documented in the Proceedings of the
various conferences of SPIE. SPIE sponsors several major conventions per year,
each having one or more conferences on detectors. Compilations of the best
Proceedings papers, along with other critical papers, are gathered in SPIE's
Milestones series http: //spie.org/x649 .xml.
References
1. F. Grum and R. J. Becherer, Radiometry, Vol. I in Optical Radiation
2. D. M. Munroe, "Signal to noise ratio improvement," in Handbook of
Measurement Science, Vol. 1, P. H. Sydenham, Ed., Wiley & Sons, New
York (1982).
3. C. D. Motchenbacher and J. A. Connelly, Low Noise Electronic System
Design, Wiley & Sons, New York (1993).
4. H. W. Ott, Noise Reduction Techniques in Electronic Systems, Wiley &
5. E. L. Dereniak and G. D. Boreman, Infrared Detectors and Systems, Wiley &
6. W. Budde, Physical Detectors of Optical Radiation, Vol. 4 in Optical
Radiation Measurements series, F. Grum and C. J. Bartleson, Eds. , Academic
7. N. B. Stevens, "Radiation thennopiles," in Infrared Detectors, Vol. 5 in
Semiconductors and Semimetals series, R. K. Willardson and A. C. Beer,
8. E. Putley, "The pyroelectric detector," in Infrared Detectors, Vol. 5 in
Semiconductors and Semimetals series, R. K. Willardson and A. C. Beer,
9. E. L. Dereniak and D. G. Crowe, Optical Radiation Detectors, Wiley &
10. R. W. Engstrom, RCA Photomultiplier Handbook, Burle Technologies,
Lancaster, Pennsylvania (1980).
11. Photomultiplier Tube: Principle to Application, H. Kume, Ed. , Hamamatsu
Photonics K. K. (1994).
12. A. Rogalski, Opto-Electron. Rev. 12(2), p. 235 (2004).
13. F. Leberl, R. Perko, M. Gruber, and M. Ponticelli, "Novel concepts for aerial
digital cameras," ISPRS Commission I Symposium, ISPRS, Denver (2002).
14. P. W. Kruse and D. D. Skatrud, Uncooled Infrared Imaging Arrays and
Systems, Academic Press, New York (1997).
....
Chapter 6
Radiometric Instrumentation
Adapted by William L. Wolfe
6.1 Introduction
Radiometric instruments vary in what they are intended to measure, how they do
it, how complicated and expensive they are, how rugged, and in a number of
other ways. In this chapter, the simplest of radiometers is considered, the
components of radiometers are described, and spectral radiometers are covered.
6.2 Instrumentation Requirements
It surely comes as no surprise that different instruments have different
requirements. For instance, the required accuracy and repeatability vary greatly
with the application. A device meant for routine maintenance in the factory might
have a radiometric uncertainty of about 20%, yielding a temperature uncertainty
of a few degrees, whereas several different bureaus of standards have strived for
uncertainties of 0.01%.
6.2.1 Ideal radiometer
It is hard to describe one ideal radiometer, but, in general, they should all
measure the radiometric property accurately, precisely, and repeatedly. They
should be simple to use, with easily interpreted outputs; of course, the ideal
radiometer should be free of defects. Some need to be quite rugged for field use,
but all should be able to withstand the environment or environments to which
they are exposed. They should be sensitive and have a wide dynamic range, or
they should at least be sensitive enough for the task at hand and have sufficient
dynamic range for their use.
6.2.2 Specification sheet
In the fervent hope that you can obtain the ideal radiometer, the following
specification sheet is presented. It may not be complete for all applications, and
not all applications need all of these specifications, but it is a good start and a
good guide.
215
216
Table 6.1 Radiometer specification sheet.
Radiometric properties
Noise-equivalent irradiance, NEI
Noise-equivalent temperature difference, NETD
Accuracy, precision
Linearity, dynamic range
Spatial characteristics
Entrance aperture size, A
Instantaneous field of view, W
Fixed or variable focus
Single detector, linear array, or 2D array
Spectral characteristics
Overall bandwidth range
Number of spectral channels
Spectral bandwidth for each channel
Continuous tuning or discrete wavelengths
Optical characteristics
Out-of-band rejection
Out-of-field rejection
Polarization sensitivity
Temporal characteristics
Spectral scan rate
Spatial scan rate
System response time and/or bandwidth
Operational requirements
Size, weight
Power, cooling requirements
Scanning requirements
Motor drives or manual
Environmental considerations
Interfaces
Mounting
Output presentation
6.2.3 Spectral considerations
Chapter 6
Values
These can be divided into three main categories: total radiation, broadband
radiation, and narrowband radiation. The first category involves detecting as
much of the total radiation as you can by using a spectrally flat detector. The
applications are things like measuring the earth's radiation budget and radiation
temperatures.
....
Radiometric Instrumentation 217
Broadband applications include making background measurements in the
infrared spectral bands for the design of infrared cameras and similar devices-in
the 3 to 5-llm and 8 to l2-llm bands, for instance. Some measurements are made
in the visible spectrum over the photopic response of the eye; this is called
photometry. Others are radiometric measurements over the visible region, while
still others are trichromatic wherein the three colors are used to obtain color
renditions. In the ultraviolet, both UV-A and UV-B are measured to assess such
things as skin cancer susceptibility.
In these cases, the spectral passband is usually defined by one or more
absorption or interference filters, the detector, the window, or a combination of
all of them. It is important to measure the spectral response of radiometers of this
type.
Narrowband measurements are usually made with some kind of
spectrometer, or in this case, a spectroradiometer. Typically, a spectrometer is a
device that measures the spectral distribution of radiation, but not its absolute
magnitude, whereas a spectroradiometer does both. In some cases narrowband
filters can be used to provide spectral selection. These measurements are usually
made in situations where the radiation varies rapidly with wavelength or a certain
amount of species analysis is desired. They also apply when brightness or ratio
temperatures are to be determined. Since the bands are narrow, relatively intense
sources are necessary to provide appropriate signal strength.
6.2.4 Spatial considerations
There are two primary types of spatial coverage-large area and small area. It
seems fairly obvious that one designs or purchases a radiometer to cover the area
of interest. This may be a spot on an electric generator or it may be an average
over an entire house. The instantaneous field of view (IFOV) is determined by
the angular subtense of the detector, that is, the size of the detector divided by the
focal length, assuming the detector is at the focus. Its size is typically on the
order of milliradians (mrad) for many IR instruments. If the instrument is a
scanning radiometer, then the field of regard, the entire scanned field, should also
be considered. Some instruments have zoom optics, a variable focus, or
interchangeable optics. These allow both small and large fields. Some have linear
or 2D arrays. These have the resolution of the individual pixels but the field of
regard of the entire array. The array must be calibrated pixel by pixel and at
different levels of irradiance.
6.2.5 Temporal considerations
One of the primary considerations in radiometric measurements is the fidelity of
the data, and this relates directly to the signal-to-noise ratio (SNR) obtained by
the instrument for a given source. In tum this is essentially the reciprocal of the
relative uncertainty. So, to obtain a relative uncertainty of, say 5%, one needs an
SNR of 20, and 1% requires an SNR of 100. These requirements are much more
severe than simple detection and tracking.
218 Chapter 6
Aliasing is an artifact in the data that can be caused by inadequate sampling.
To avoid these false signals, the bandwidth must be at least twice the highest
frequency in the signal. These frequency components can be generated by natural
fluctuations or by the use of a chopper, or perhaps turbulence.
If a chopper or some other device is used to generate square pulses as the
signal, then the bandwidth must be at least equal to the reciprocal of twice the
pulse width. If just a chopper is used, the chopper frequency must be three times
the highest frequency in the signal.
For an uncertainty of 5%, the settling time must be two times the response
time of the detector. Similarly, for 1 %, five times, and for 0.1 %, seven times.
Since all radiometers have a finite bandwidth, there is noise in that
bandwidth. Section 5.4 addresses noise in detail. To restate, white noise has a flat
power spectrum, and effective noise bandwidth B is defined by Eq. (5.15). For l/j
noise, the expression in Eq. (5.14) is used.
6.2.6 Make or buy?
The decision to make or buy is one of availability of devices and of money. If
there is a radiometer available that meets your needs or comes close enough and
it fits within your budget, buy it. If the most useful unit needs too much
adaptation, don't buy it. If you need to train students, don't buy it. If you do buy
it and need to modify it, do so with as many commercial items as possible. For
example, do not make the monochromator of a spectroradiometer; there are units
for sale.
6.3 Radiometer Optics
6.3.1 Introduction
Chapter 2 introduced basic optical concepts relevant to radiometry as well as
some simple radiometer configurations. They will be reviewed here in
preparation for more detailed discussion of instrumentation. The main issues of
radiometer optics are the placing of stops and pupils, possible reimaging, use of a
field lens, and the reduction of scattered light.
6.3.2 Review of stops and pupils
There are two types of stops: aperture stop and field stop. The aperture stop
controls the amount of light that is accepted; the field stop controls the field of
view from which it is collected. Pupils are images of stops. The entrance pupil is
the aperture stop as projected into object space (by all the optical elements that
are between the aperture stop and object space). The exit pupil is the aperture
stop projected into image space. The entrance and exit windows are the
corresponding images of the field stop. These are illustrated in Fig. 6.1. Notice
that the second lens L2 is the aperture stop because its image in object space is
smaller than that of the first lens. The field stop is C, and its image out in object
space is the entrance window.
...................
A
B -------
~ ' I L2
,
, .
B ------
A
.-
.- LI
Figure 6.1 Stops and pupils.
6.3.3 The simplest radiometer: bare detector
A bare detector, shown in Fig. 6.2, can be used as a radiometer with its
associated electronics. Its angular field of view is almost 180 deg, but it is
probably not uniform over that range. Ideally it varies as the cosine of the angle,
but most detectors have some other dependency as a result of coatings, surface
imperfections, partial specular reflectivity, etc. If the detector has a polished
specular window, the angular variation may be calculable from the standard
Fresnel equations for reflection. Again, and it cannot be said enough, check it
out! The bare detector can be used for relatively imprecise quasi-hemispherical
measurements or for a distant source.
6.3.4 Added aperture
A slight improvement is the addition of an aperture, as shown in Fig. 6.3. With
almost no expense, just a case with a hole in it, one can define the field of view.
As a matter of possible interest, one can make a thermocouple by welding two
dissimilar metals, then punch a hole in a number ten tin can to make the business
end of a radiometer. The electronics could be one of those little multi meters you
can buy at the electronics store for about five dollars. The detector could be a
solar cell for visible spectrum use.
Incoming
Radiation
Detector
Figure 6.2 Bare detector.
220 Chapter 6
The Can!
Aperture
Detector
Figure 6.3 One-aperture system.
In this case the detector is the aperture stop, and the hole is the field stop.
Since there are no lenses, there are no pupils. The flux on the detector from a
point source is given by
(6.1)
where Is is the source intensity, Ad is detector area, and d is the source-detector
distance. For an extended source (one that is larger than the image of the
detector), the flux on the detector is given by:
(6.2)
where Ls is the radiance of the source and Aap is the area of the aperture.
The field of view is "fuzzy": that is, different parts of the detector view
somewhat different parts of the field of view because the detector is the field
stop. Note that if you make the detector a little smaller, the field of view is a little
smaller, so the outer portion of the detector views a different portion of the field
than the inner portions.
A reasonable improvement is obtained by the use of two apertures, as shown
in Fig. 6.4. The detector is no longer the field stop. For a point source, and A2
smaller than the active area of the detector, A2 is the aperture stop, A I is the field
stop, and the flux on the detector is given by:
(6.3)
and for an extended source,
(6.4)
where d is the distance between the two apertures. The field is still a little fuzzy.
A2
, -I
~ Detector
Figure 6.4 Two-aperture system.
One example of such a radiometer that uses a single detector and two
apertures and baffles but no lenses, is a pyrheliometer (fire + sun, measuring the
sun's fire) used by the World Meteorological Organization and shown
schematically in Fig. 6.5. The angles made with the detector by A, B, and C
should be about 1, 2, and 4 deg, respectively.
6.3.5 Basic radiometer
Figure 6.6 shows a basic radiometer, just a lens and a detector. A mirror can be
used in place of a lens, and often is, but the diagrams get more complicated than
necessary. The lens is the aperture stop and the detector is the field stop. In this
A Be
, , I
, ,
, I
" I
~
Detector
Field Stop
Baffle
Baffle
Figure 6.5 Pyrheliometer.
222 Chapter 6
Lens
Detector
Figure 6.6 Basic radiometer.
arrangement, the field is well (not fuzzily) defined, but the detector still "sees"
radiation coming at different angles from different parts of the field of view. This
is the case as long as the detector is the field stop. The advantages are that this is
a simple system---one lens, one detector, and associated electronics. The
disadvantage is that the detector averages over different parts of the field of view.
In this system the irradiance on the entrance pupil (the lens) from a point
source is given by
Is't
atm
-;T'
(6.5)
where Is is the source intensity, 't
atm
is the atmospheric transmittance and d is the
distance from the source to the entrance pupil. The flux on the entrance pupil is
just the lens area A/ens times this, and the flux on the detector is just the lens
transmittance 't/ens> times that:
(6.6)
The corresponding quantities for an extended source are
(6.7)
If the source is extended, the distance from the instrument to the source IS
irrelevant; it is the focal length that counts.
Aperture
stop
Field
lens
Figure 6.7 Improved radiometer.
6.3.6 Improved radiometer
223
The improvement includes the addition of a field lens, as shown in Fig. 6.7. The
field lens images the aperture A onto the detector D, thereby making the flux on
the detector uniform. Although it is not yet obvious, this also helps control stray
light. The radiometry of this radiometer is not much different, but it must be
separated, as before, into the point-source case and the extended-source case.
For an extended source, where the detector is overfilled with the object, the
expression for the flux on the detector is
(6.8)
where 'to is the transmission of the optics, Ls is the radiance of the source, Ad is
the detector area, and nftdis the solid angle the field stop subtends at the detector.
(6.9)
where d is the distance from the detector to the field stop. Thus, one can also
write the power on the detector as
(6.10)
If the image of the objective underfills the detector, then
(6.11 )
224 Chapter 6
If the image of the objective exactly fills the detector, either of Eqs. (6.10) or
( 6.11) is accurate.
6.3.7 Other methods for defining the field of view
The two other principal methods for defining the field of view are the use of
diffusers and of integrating spheres. One may think of an integrating sphere as a
special diffuser. A simple diffuser is just a good scattering material. In the choice
of such a material, the grit and the spectral characteristics need to be considered.
Usually a good absorbing material, a diffuse black, is preferred. And remember,
just because it looks black does not mean it is black in the infrared or ultraviolet.
The black can be spread on the detector by an appropriate method, or it could be
used to reflect to the detector, in which case, it should be a white! There is no
substitute for measuring its spectral and angular responses. Such a diffuser is
usually a good depolarizer as well as a spatial averager.
The integrating sphere is coated with a highly reflecting diffuse material. It
has an entrance aperture (hole) and an exit aperture. The light enters the first and
exits the second to the detector. The multiple reflections inside the sphere are
what make the exiting light both uniform and unpolarized. Again, the spectral
characteristics need to be considered. There are several good references on the
calculations of the performance of integrating spheres. Try entering the search
criteria "integrating spheres" on the Internet. There are several good references
that give the throughput equations and offer a variety of types and sizes for sale.
Spheres provide an excellent cosine response, which is nice for
hemispherical measurements, but they are notoriously inefficient.
6.3.8 Viewing methods
It is usually very helpful to be able to see exactly what is being measured. Thus,
people have developed a variety of viewing schemes to accompany the
measurement instrumentation. One simple scheme is to put a telescope on top or
to the side of the radiometer. This scheme is simple, straightforward, and
removable, but it suffers from parallax: the two fields of view will not coincide at
all distances. Coaxial methods are preferable from this standpoint. One way to do
this with obscured systems is to use the folding mirror, as shown in Fig. 6.8.
Another popular way is to use a reflecting chopper. Then the fields are coaxial
and alternating in time. The persistence of the eye takes care of the interruptions.
Such a scheme is shown in Fig. 6.9. The mirror of Fig. 6.9 could be a large
chopper in the incoming beam, and then the eye and the reference can be in either
position. Other schemes may incorporate a fiber-optic pickoff someplace in the
system, but these do not usually work out very well, mostly because they are not
coaxial.
.....
Mirror
Detector
.,.
1\
I \
J ~
-- \
............ .J.
I ~
Eye
Figure 6.8 Viewing with an obscured (Newtonian) optical system.
Detector
Reflective
chopper
Reference
225
I
~ ~ = ; ~
-- ,
- ......
I ~
_ - .4'f---+-----------------------------
-'
,-
Mirror
Eye
Figure 6.9 Viewing with a chopper.
226 Chapter 6
6.3.9 Reference sources
Although all radiometers must be calibrated before and after use, it is also useful
to have an internal reference source. In radiometers with choppers, covered in the
next section, the measurement signal is the difference between the two detector
outputs: when the detector views the object and when the detector views the
reference source. The ideal situation is when there is no difference between the
signals. So the reference source should be designed to give approximately the
same flux on the detector as the object to be measured. This is a "null" reading.
The fundamental requirement is that the reference must be known. Sometimes it
is a blackened chopper, but this can be difficult since the spinning chopper may
vary in temperature and may have temperature gradients. Of course the chopper
temperature and emissivity must be known at the time of the measurement. A
good alternative is for the chopper to be highly reflective. Then it reflects an
internal reference to the detector. The internal reference can then be a well-
controlled source with high emissivity. Its temperature can be measured
continuously, and, since it is stationary, it is likely to be more stable. Most of
these internal references are blackbody simulators. They are usually conical in
shape with a relatively small aperture. Extensive research has been done on the
shapes and sizes to get the highest possible emissivity. The critical thing is that
the interior length should be much larger than the aperture. However, the ultimate
is not necessary for these internal references since they are secondary calibration
sources. Other shapes have included waffles, cylinders, spheres, and the like.
Basic design techniques have been published by Gouffe' and DeVos
2
and are
also described in The Infrared Handbook. 3 Many commercial units can be found
by surfing or using the Photonics Spectra catalogs (on disk or in hard copy).
These are not usually designed to fit into equipment but to be laboratory
standards. You can surf for "blackbody simulators," and find many commercial
devices.
6.3.10 Choppers
Most good radiometers use choppers, or radiation modulators. Of course there
are advantages and disadvantages. The ac signal from a chopped radiometer
provides a discriminant against a static background. It allows the use of drift-free
ac amplifiers. It avoids the low-frequency part of the lifnoise region. It provides
the ability to use a synchronous detector, and it provides an a-b type of
comparison measurement. On the other hand, choppers reduce the available flux
by a little over 50%; they can be noisy both electrically and acoustically, they can
have reliability problems, and there can be phase noise if things are not exactly
right.
Most choppers are just spinning blades, driven by an electric motor. In this
form they do not use much energy. There are many types available on the market.
Some can be resonant devices that oscillate in and out of the beam, perhaps
tuning forks. These, too, are readily available from companies listed on the
Internet. Browse for "optical choppers"; both types are advertised.
...
6.3.11 Stray light
Stray light on the detector is a false signal. Therefore, stray light should be
eliminated or reduced to a level that makes it insignificant. There are basically
two types of stray-light problems: those that come from stray light that is in the
field of view of the radiometer (background light) and those that come from
outside the field of view. The latter can be the sun or another intense source that
is near the edge of the field of view. It can be greatly reduced by careful design
of stops and baffles. The proper use of baffles can attenuate these out-of-field
sources by many orders of magnitude.
The stray light from out-of-field sources can be reduced by the following
procedures. First, place the field stop or its image as close to the front of the
system as possible. Place the aperture stop or its image as close to the detector as
possible. Avoid obscured systems like Newtonians and Cassegrains (and this
may make a viewer more difficult). Make any baffles angled and black. There is
still an argument as to whether they should be specular or diffuse. The diffuse
baffles have better attenuation; the specular ones control the direction of the light
better. Obscured systems will provide their own scatter, but they can be used.
Figure 6.10 shows how to design a simple baffle. The baffle is designed to
protect against a source, like the sun, that is at an angle e off axis. The first step
is to make the baffle long enough to achieve this. Then draw line ab at that angle
from the tip of the baffle to past the mirror tip. Then draw line cd parallel to
- -
ab and to the end of the baffle tube. Then draw ef to intersect that line; that is
where the first vane is inserted, as shown. Then draw j to intersect at f Place the
next vane where f intersects the line of baffle tips. This line is parallel to the tube
at the edge of the mirror. Now repeat the process for the other vanes.
Figure 6.10 Baffle design.
228 Chapter 6
The final step is to analyze the system with one of several programs that can
calculate the amount of radiation on the detector from out-of-field sources as a
function of their angle out of field. And there is no substitute for a measurement
if the application warrants it. A number of stray-light-analysis systems are listed
on the Internet. Search for "stray light optical techniques." The reduction of stray
light is an iterative technique that is shortened by experience. The analysis
programs tell you how you are doing.
The Lyot stop is a special kind of stop useful in controlling scatter. It was
used by Lyot to measure the solar corona. He imaged an interior disk inside the
system to the front. The image of the disk matched the solar disk, allowing him
to see the light outside the main disk of the sun-the corona and flares. The
image of the disk became the entrance pupil, an image of the aperture stop. This
is one application of using an entrance pupil to control stray light.
6.3.12 Summing up
The design and use of radiometric instruments is simple in principle, but difficult
in practice. It is a science of precision and care. One of the maxims is to think of
everything. One way to do this is to write the radiometric equation for the
radiometer:
v = R(A, t, x,y,z, a, ~ moon, RH, T, ... ). (6.12)
This means that the output voltage (or other electrical signal) is a function of
wavelength, time, spatial coordinates, angular coordinates, the phase of the
moon, the relative humidity, the temperature (of the radiometer and the source
and the background), and everything else you can think of. Then test the
radiometer against all of these variables. The devil is in the details-and so are
good results!
6.4 Spectral Instruments
6.4.1 Introduction
Spectral instruments include both those that make relative measurements and
those that make absolute measurements. The first type measures how much
radiation there is in each part of the spectrum, but only on a relative basis. They
are useful for chemical analysis, for instance. The absolute instruments, generally
called spectroradiometers, measure the spectrum, but also give information about
how many watts (or equivalent radiometric quantity) are at each wavelength.
They are calibrated and more difficult to use. This section is about various types
of spectroradiometers. Another way to view them is that they are radiometers
with several, or many, narrow spectral bands. The different types vary from each
other according to the way they obtain the spectrum. That can be with prisms,
F'"
gratings, filters, by scattering, interference, and, in general, any phenomenon that
is spectrally dependent.
A typical spectroradiometer (or spectrometer) is shown in Fig. 6.1l.
Foreoptics on the left bring the light to the entrance slit of a monochromator (the
device that generates the spectrum). The output of the monochromator goes to a
detector, and then to electronics, and then to some kind of display or recording.
The monochromator singles out a specific, usually narrow, wavelength band at a
time, and usually then sequences through several bands. With multiple detectors
all wavelengths can be sensed at the same time.
The monochromator consists of an entrance slit, collimating optics, a
disperser (usually a prism or grating that spreads the light of different
wavelengths), focusing optics, and an exit slit. Although the optics shown here
are lenses, most monochromators use mirrors to avoid problems with chromatic
aberration.
Spectroradiometers are characterized by how much light they collect, and
therefore their sensitivity, and by how narrow their spectral bands are. The
amount of light they collect is specified by their throughput (also called optical
extent, etendue, and An product.) The narrowness of the spectral band is usually
specified by the resolving power, that is, the central wavelength divided by the
spectral width of the band. This is the same as the Q of an electronic filter. In
some instances the descriptor is the resolution, which is just the width of the
spectral band, usually specified as the full width at half maximum (FWHM). The
free spectral range is a description of the spectral region over which there is no
interference of other spectra, perhaps by overlapping of orders. Other descriptors
include the multiplex advantage and the throughput advantage. These latter two
are also called the Felgett and the Jacquinot advantages. The throughput
advantage relates to how much light you can get through the system. The
multiplex advantage refers to the measurement of many wavelengths of light at
one time, rather than sequentially. If there is a multiplex advantage, the
bandwidth can be narrower and the sensitivity therefore greater.
Entrance
Disperser
Foreoptics slit slit
Exit
~ ~ w
Source
Collimating Condensing
lens lens
Figure 6.11 Typical spectrometer.
Focusing
lens
230 Chapter 6
To get the optimum response from a spectrometer, one should make the
width of the entrance and exit slits the same. One should also just fill the entrance
slit with the foreoptics. Underfilling robs you of input; overfilling gives no
advantage and even increases stray light.
6.4.2 Prisms and gratings
Spectroradiometers based on prisms and gratings are probably the most common.
Each has its advantages and disadvantages. The resolution of a grating
spectrometer is constant across the spectrum and usually more than that of a
prism system. The throughput of the grating is also a little larger than that of the
prism. The resolving power of a grating is generally higher than that of a prism.
The grating equation is
(6.13)
where m is the order number, an integer, A is the wavelength of the light, and
a
i
and ad are the angles of incidence and diffraction, respectively. For a given
wavelength and angle of incidence there are many maxima at various diffraction
angles that correspond to different values for m and ad. The oth order includes all
wavelengths and is just regular reflection (or transmission). For a given incidence
angle, the first order, m = 1 for a wavelength A, has the same diffraction angle as
the second order and halfthe wavelength. This is overlapping of orders and limits
the free spectral range.
The resolving power Q is given by
A
Q= dA =mN ,
(6.14)
where m is the order number and N is the number of grating lines illuminated.
The throughput is the slit area times the projected area of the used portion of
the grating divided by the focal length of the optics:
Astir A graring cos a
T=An=------"----::-"---
j 2
(6.15)
The free spectral range of the prism is unlimited, while that of the grating is
limited by multiple orders, the different maxima corresponding to different
values of m. Gratings can be transmissive, reflective, and even concave to
incorporate some of the focusing properties of the system. The resolving power
of a grating is approximately constant across the spectrum, whereas the prism's
Q varies. The resolving power of a prism is given by
pas
(6.16)
where b is the length of the base of the prism (actually the length of the beam
near the base) and dn/d)... is the change in refractive index of the prism with
respect to wavelength.
The throughput is dictated by the slit area and the focal ratio of the focusing
optics:
(6.17)
Although it is a generalization, gratings are generally better: they have
greater throughput and resolving power than prisms.
6.4.3 Monochromator configurations
Since the monochromator is the heart of generating spectra and measuring them,
several variations are described here. They can be divided generally into prism
devices and grating devices, according to the type of disperser that is used. Prism
systems consist of several different mounts: Littrow, Wadswoth, and Amici.
There are more, but these are the main ones. The Littrow, as shown in Fig. 6.12,
has a mirror at an angle behind the prism. Thus the light enters the prism, exits to
the mirror, and returns through the prism, thereby creating a double pass for
twice the dispersion. The mirror can be adjusted for the degree of separation of
the input and exit beams.
The Wadsworth, as shown in Fig. 6.13, is a single pass. The light is refracted
at minimum angle and exits the mirror parallel to the input beam. The mirror is
an extension of the prism base.
The Amici, which is shown in Fig. 6.14, allows the light with the central (or
any chosen wavelength) ray undeviated and undisplaced (with careful design). It
requires the use of prisms of at least two different materials (with different
dispersions). More than three prisms can be used at the designer's discretion.
Figure 6.12 Littrow mount.
232 Chapter 6
Figure 6.13 Wadsworth prism mount.
Figure 6.14 The Amici prism.
Mirror systems have more variability, but generally there is an entrance slit,
an exit slit, one or two mirrors, and a grating. The Czerny-Turner, Fig. 6.15, was
developed by Marianus Czerny and Francis Turner, a onetime student of
Czerny's and later a researcher at Eastman Kodak and professor at the College of
Optical Sciences, The University of Arizona. There are two versions; one might
be called straight and the other crossed, as in Fig. 6.16. The crossed version uses
the two paraboloids closer to on axis and should therefore have smaller
aberrations. Of course, if aberrations are not a problem, the paraboloids can be
replaced by spheres. In that case, and depending upon the speed of the optics,
spherical aberration rather than coma will be dominant. It is shown with a
reflective grating, but it is not hard to imagine how this setup can be used with a
transmissive grating. The essence of the design is that both the collimating optic
and the focusing optic are off-axis paraboloids. However, they are off axis in
opposite directions so that the comatic aberrations offset each other.
The Fastie Ebert mount is shown in Fig. 6.17. It has the advantage of using
only one mirror, although the mirror has to be larger than either of the mirrors of
the Czerny-Turner system. Again, the mirror can be either a sphere or a
paraboloid, and again the comatic aberrations tend to offset each other.
Obviously, this cannot be used with a transmissive grating.
z
ps
Entrance
slit
Grating Exit
slit
Figure 6.15 Czerny-Turner (laid out) configuration.
Grating
Figure 6.16 Alternate (cross) Czerny-Turner configuration.
Entrance
slit
Mirror
Grating Exit
slit
Figure 6.17 Fastie-Ebert configuration.
233
234
-
Concave
grating
Entrance Exit
slit slit
-
Figure 6.18 Seya-Namioka configuration.
Chapter 6
The Seya-Namioka is shown in Fig. 6.18. It has the advantage of having no
mirrors at all. Thus, it is very compact and usually cheap. The focusing properties
are all in the convex reflective grating. However, this grating can be difficult to
make so that is has enough concavity and also enough lines.
These configurations are also discussed in the first of the online references
below.
6.4.4 Spectrometers
These devices come in several different varieties. They can be single or double
pass through the dispersers, like the Wadsworth (single) or the Littrow (double).
They can be single or double beam, as shown below. The double beams were
developed to get an automatic referencing system. In some older single-beam
systems, one had to wrap a cord around the screw that rotated the prism in just
the right way. This was to "program" the slit to account for the variation in
intensity from the source. One also had to run a calibration run, and then a
sample run, and manually do the ratioing. The double-beam system eliminates all
this. There are many ingenious variations on the several simple systems shown
here. One beam passes through the sample, whatever it may be. The other is the
reference beam and is just in air in the spectrometer. The output of the sample
beam is divided by the output of the reference beam to give the transmission of
the sample. There are several ways to do this, but the best way is to use the same
detector for both beams, perhaps with an appropriate chopper. This avoids the
obvious problem of detector matching, initially and repeatedly. The generic basic
two-beam instrument is shown in Fig. 6.19. The output coming from the exit slit
of the monochromator is collimated and divided into two beams by a divider. It
can be a (semitransparent) beamsplitter or a bladed chopper. The one beam goes
straight through; the other is diverted by two mirrors and then combined with the
Monochromator
exit slit
Mirror
Beam
divider
Beam
combiner
Figure 6.19 Generic double-beam spectrometer.
235
Detector
first beam by a beam combiner just like the beam divider-a synchronized
chopper or a plane-parallel plate. If the divider and combiner are choppers, the
electronics can just take the ratio of signals. If beamsplitters are used, some
technique must be used to "tag" each of the beams. The beams need not be
collimated, but could be relayed by lenses for compactness.
Other commercial instruments can be found on the internet by entering the
search term "spectrometers." Many will show up with diagrams, prices, and
advertising!
There is one subtlety about double-beam systems that I (Wolfe) encountered
quite by accident. I was participating in a military study devoted to the detection
of land mines. Someone suggested that soil was transparent around 4 /--lffi. I
scoffed, but I went home to prove my point. I took one inch of certified backyard
dirt into our double-beam spectrometer. Lo and behold, there was a transmission
peak at 4.3 Ilm! This could not be! So I took one inch of aluminum plate and
made the same measurement. Same result. I pondered this for a while. Clearly
aluminum is not transparent at 4.3 /--lm. The answer was in the reference beam.
The atmosphere is very absorbent at 4.3 /--lm due to carbon dioxide. What I was
measuring was the absorption in the reference beam that translated to apparent
transmission in the overall measurement. I have since seen this phenomenon in
other double-beam measurements. If there is extra transmission at 4.3 /-!ill in the
scan of an optical material you have ordered, be skeptical. This illustrates at least
one principle of radiometry: do not simply believe whatever you measure; make
sure it makes sense. By the way, this problem can be obviated by filling the
reference tube with nitrogen.
6.4.5 Additive versus subtractive dispersion
Additive dispersion is just that: two prisms or two gratings operating in series to
increase the dispersion of the light. The Littrow-mounted prism gives additive
dispersion. Additive dispersion can be accomplished with multiple dispersers or
multiple passes or both. Subtractive dispersion combines the dispersions in
236 Chapter 6
opposite directions, thereby combining the light and eliminating or canceling the
spread of the spectrum.
6.4.6 Arrays
Monochromators can also be used with detector arrays. The arrangement
includes the foreoptics described above, an entrance slit, collimating optics, and
the array, which functions as a set of exit slits. If the array is linear, the system
can be viewed simply as a monochromator with many exit slits, each one sensing
a different wavelength. This system obviously has the multiplex advantage, but
does not have the throughput advantage. The detector elements can usually be
sized correctly for the proper operation, but apertures can also be used with a
concomitant loss of signal.
6.4.7 Multiple slit systems
Several different schemes have been developed to give a multiplex advantage by
using several entrance slits and several exit slits at the same time. Perhaps the
first of these was invented by Marcel Golay;4 perhaps the most popular, the
Hadamard transform, was developed by Harwit.
5
Although these are interesting
designs that incorporate the multiplex advantage, they have not supplanted the
other prism, grating, and interferometric spectrometers.
6.4.8 Filters
Filters come in a variety of types. They can be based on absorption, interference,
and even scattering. Absorption and interference filters are the main candidates in
spectroradiometry. Almost any kind of bandpass can be generated by a proper
thin-film design. They can be narrowband, broadband, angle tolerant, multiple
bandpass, etc. Then they can be put in a filter wheel to obtain a spectrum of a
sort, although not a continuous spectrum. They have good throughput but do not
have as good resolving power as prisms and gratings. They can come in
segmented wheels or circular or linear variable filters. These latter devices are
interference filters with layer thickness variation around the circumference or
along the length. They therefore have a spectral band that varies with either angle
or position.
Their characteristics vary by design. The circumference of a circular variable
filter (CVF) is given by
c = /).'A D
d'A 0'
(6.18)
where /).'A is the spectral range, dA is the resolution and Do is the diameter of the
aperture stop, where it is placed. The diameter of the CVF is this value divided
by n. A representative filter has the following characteristics: the spectral range
is 2.5 to 14.1 ~ m Q is 67, FWHM is 1.5%, and resolution varies from 0.03 ~ m at
p
2.5 /lm to 0.2 /lm at 14.1 /lm. The CVF is divided into three segments that cover
from 2.5 to 4.3, 4.3 to 7.7, and 7.7 to 14.1 /lm. It can be operated at the focal
plane, as the manufacturer recommends, but it should be operated at a field stop
rather than the focal plane. Since it would then be in converging light, the
resolution will be affected. Of course other varieties are available, semicircles for
instance. Typically the spectral range is an octave, the resolving power is about
50, and the diameters are about 4 cm. Specific devices can be found on the
Internet by searching for "optical filters" or "variable optical filters." If you leave
out "optical," all sorts of electrical filters will appear in the search results.
Optical filters are optically and mechanically simple, cheap, rugged, and easy
to automate, but they have poor stray-light suppression and limited resolving
power.
6.4.9 Interferometers
The main interferometers used for spectral analysis are the Michelson and the
Fabry-Perot, or as some insist, the Perot-Fabry. The former is a two-beam
instrument, the latter a multiple-beam device. The mUltiple-beam system has
greater resolving power, but smaller free spectral range.
6.4.10 Fourier transform infrared
The Fourier transform infrared spectrometer is a Michelson, or actually a
Twyman-Green, as shown in Fig. 6.20. (The Twyman Green is a Michelson with
collimated light). The light from the source is collimated by the first lens. The
beam is then divided into two by the beamsplitter. One beam goes up, the other
continues to the right. The mirrors then reverse the direction of both beams and
they are combined by the same beamsplitter, which is now acting as a combiner.
The beams interfere, and the intensity of the interfering radiation is sensed by the
Mirror
Beam divider/combiner
Source
Mirror
............ Lens
Figure 6.20 Twyman-Green interferometer.
238 Chapter 6
Detector, whose operation can be understood by considering first a single
monochromatic input, then, successively, added inputs of different wavelengths.
When the interferometer is set so that the two arms have exactly the same optical
length, light from the two beams will interfere constructively. Then, as one
mirror is moved, the beams gradually go out of phase until they reach destructive
interference, and as the mirror is further moved, they gradually reach
constructive interference again. So there will be a sinusoidal output based on the
motion of the scanning mirror. If a beam of somewhat different wavelength is
inserted it will do the same, but for somewhat different positions of the mirror. A
third beam, a fourth, and so on will add to the complexity, but each will
contribute a sinusoidal component to the output signal. At any position of the
scanning mirror, the output is the sum of all waves, each at a different point in
phase. The sum total of the scan is called an interferogram-the interference
pattern of whatever the input beam was. The Fourier transform of the
interferogram is the spectrum of the input beam. Because this device is used
mostly in the infrared, it is called the Fourier transform infrared spectrometer. It
has good throughput and the so-called multiplex advantage; i.e., it senses all
wavelengths of light at the same time (as opposed to prism and grating
instruments that see one small sample of the spectrum at a time).
The resolving power is given by
Q _ 'A _ a _ a _ 5000
- d'A - da - 20 - .
(6.19)
The expression a/da is the wavenumber equivalent. The final expression
incorporates the fact that the wavelength is usually given in 11m while the
wavenumber is in cm-
l
.
These systems are complex, susceptible to vibration, have a limited spectral
range, and require considerable computing power to perform the transform
operation, but they have the multiplex advantage and good throughput. They are
limited to the infrared, but they are very useful for what they do.
6.4.11 Fabry-Perot
This interferometer is generally used for high-resolution spectroscopy. It is
essentially a pair of plane-parallel plates, as shown in Fig. 6.21. The light from
the source is collimated by the first lens, and passes through the first plate after a
little lost reflection. Light reaches the second plate and is reflected back to the
first, which reflects it back to the second, which reflects it back to the first. This
interferometer does have high resolution because of the multiple-beam
interference. The basic equation for the transmission of the Fabry-Perot is
(6.20)
jiiii"
and
(6.21)
where 'to is the transmission of the plates, p is their reflectivity, n is the refractive
index of the medium between the plates, d is the plate separation, 8
i
is the angle
of inclination of the incident beam, which is usually 0, and the <\>i are the phase
changes on reflection from the plates. It can be shown that the maximum
transmission is when the reflectance is highest (in the limit, 1), which is a rather
strange result. The resolving power is given by
0" A JP
Q=-=-=--m1t
dO" dA I-p ,
the throughput is given by
and the free spectral range is given by
1
dO"=-.
2d
(6.22)
(6.23)
(6.24)
The Fabry-Perot has a relatively poor throughput because of the requirement for
collimation between the plates. But it is better than a prism or grating.
Final note: Acousto-optical tunable filter (AOTF) devices for spectroscopy of all
kinds are relatively new. "Classical" spectrometers employing prisms and
gratings have seen many improvements. There are pros and cons for both. They
are a tool that should be in the radiometrist's kit and are described further in
Wolfe and in Chang, below.
Source Detector
Figure 6.21 Fabry-Perot Interferometer.
240 Chapter 6
For Further Reading
P. Baumeister, Optical Coating Technology, SPIE Press, Bellingham,
Washington (2004). Good section on filters.
1. C. Chang, "Acousto-optic devices and applications," Chapter 12 in Handbook
of Optics, Vol. II, M. Bass, Ed., McGraw-Hill, New York (1999).
D. S. Goodman, "Basic optical instruments," Chapter 4 in Geometrical and
Instrumental Optics, D. Malacara, Ed., Academic Press, New York (1988).
F. Grum and R. Becherer, Radiometry, Sec. 7.3, Vol. 1 in Optical Radiation
G. R. Harrison, R. C. Lord, and J. R. Loofbourow, Practical Spectroscopy,
Prentice-Hall, New York (1948).
R. Kingslake, "Dispersing prisms," Chapter 1 in Applied Optics and Optical
Engineering, Vol. 5, R. Kingslake, Ed., Academic Press, New York (1969).
R. Meltzer, "Spectrographs and monochromators," Chapter 3 in Applied Optics
and Optical Engineering, Vol. 5, R. Kingslake, Ed., Academic Press, New
York (1969).
W. J. Potts, and A. L. Smith, "Optimizing the operating parameters of infrared
spectrometers," Appl. Opt. 6, 257 (1967).
D. Richardson, "Diffraction gratings," Chapter 2 in Applied Optics and Optical
Engineering, Vol. 5, R. Kingslake, Ed., Academic Press, New York (1969).
R. A. Sawyer, Experimental Spectroscopy, Dover, New York (1963).
R. Willey, Practical Design and Production of Optical Thin Films, Marcel
Dekker, New York (2002). Good reference on filters.
W. L. Wolfe, Introduction to Imaging Spectrometers, SPIE Press, Bellingham,
Washington (1997).
J. Workman and A. W. Springsteen, Applied Spectroscopy, Academic Press, New
York (1998).
References
1. A. Gouff6, "Corrections d'ouvertures des corp-nois artificels compte tenu des
diffusions multiples internes," Revue d' optique 24( 1) (1945).
2. J. C. DeVos, "Evaluation of the quality of a blackbody," Physica 20, p. 669
(1945).
3. A. J. LaRocca, "Artificial sources," Chapter 2 in The Infrared Handbook, W.
L. Wolfe and G. J. Zissis, Eds., u .S. Government, Washington, D.C. (1978).
4. M. Golay, "Multi slit spectroscopy," J Opt. Soc. Amer. 39, pp. 437-444
(1949).
5. M. Harwit and N. Sloane, Hadamard Transform Optics, Academic Press,
New York (1979).
'S
F'
Chapter 7
Radiometric Measurement and
Calibration
7.1 Introduction
This chapter deals with the numerous measurements for which we use
radiometers and spectroradiometers. This is the true meaning of the word
radiometry, the measurement of radiant energy. First, we describe the types of
measurements made. Next is a discussion on errors, their sources, and treatment.
The generalized measurement equation and several derived range equations
follow. An introduction to the philosophy of radiometric calibration is presented
next, and a discussion of calibration configurations completes the chapter.
7.2 Measurement Types
Radiometric measurements may be classified into four general types. They are:
(1) detector and radiometer characterization, (2) optical radiation source
measurement, (3) material properties measurement, and (4) temperature
measurement. A fifth measurement type is calibration, which will be discussed
later. Table 7.1 subdivides the first four categories into what is not an exhaustive
list.
7.3 Errors in Measurements, Effects of Noise, and Signal-to-
Noise Ratio in Measurements
A measurement of any kind is incomplete unless accompanied with an estimate
of the uncertainty associated with that measurement. The term error implies a
difference or deviation from a "true" value, while the term uncertainty means an
estimate characterizing the range of values within which the true value of the
measured quantity lies, including all sources of error. Errors come in two primary
flavors, random and systematic, as we will see through consideration of Fig. 7.1.
Systematic (type B) errors are readings that vary in a predictable, hopefully
detectable, way. Systematic errors are repeatable and consistent, with a fixed
bias, the difference between the measured value of x (mean of N measurements)
and the true value of x.
241
242
Table 7.1 Types of radiometric measurements.
Detector and radiometer characterization
Relative spectral responsivity
Absolute spectral responsivity
Noise properties
Detective properties (NEP, D*, D**)
Field of view, out-of-field response
Linearity
Frequency response
Polarization response
Wavelength characterization (for spectroradiometers)
Passband characterization
Measurement of optical radiation sources
Active (self-radiating) and passive (reflective) sources
Source intensity
Source radiance or brightness
Source power or total flux
Light (photometry)
Ultraviolet and infrared sources
Source temperature
Collimated (laser) sources
Measurement of radiometric properties of materials
Specular reflectance
Diffuse reflectance
Transmittance
Scattering properties, BRDF, and BTDF
Indirect measurements of absorptance and emittance
Direct measurement of emittance
Color
Measurement of temperature
Radiation temperature using entire spectrum
Brightness temperature using one wavelength
Ratio temperature using two wavelengths
Color temperature using chromaticity
Temperature using multiple wavelengths
Distribution and correlated color temperature
Chapter 7
Examples of a systematic error are an incorrect setting of a calibration
potentiometer of a voltmeter, or a slipped or improperly installed temperature
dial on a blackbody radiation simulator. Systematic errors are not revealed by
repeated measurements. If detected, they may often be corrected or at least taken
into account. The term accuracy is often applied to systematic errors, implying
small systematic errors. This term should be replaced by inaccuracy, inasmuch as
-
Radiometric Measurement and Calibration 243
f(x)
Systematic error
Random error
L ~ ~ ~ ~ ~ ~ x
True value of x Measured value ofx
Figure 7.1 Systematic and random errors.
a voltmeter with only 1 % accuracy is far less desirable than one with 99%
accuracy. The systematic error in a measurement is closely related to calibration
of the apparatus used to conduct the measurement.
Random (type A) errors are those that vary in an unpredictable manner when
the same quantity is repeatedly measured under identical conditions. They are
revealed only by multiple measurements. Precision is a term often associated
with random errors; a measurement is considered precise if it is repeatable. We
may employ several methodologies to reduce random errors, enhancing the
measurement precision. First, we can use a finer scale division, i.e., more bits, to
reduce granularity in the measurement. Second, we can reduce some of the
inherent noise in the measurement process by filtering, cooling, shielding, etc.
The most important tool to reduce random errors is statistical analysis. We take
multiple readings and perform analysis to reduce the effects of "noise" and
increase our confidence in the measurement.
To understand the total uncertainty in a measurement, we must consider both
the systematic and the random error components. Figure 7.2 shows a set of
"measurements" with various combinations of accuracy and precision. The
average (x and y) of the third pattern lies very close to the center of the target.
Errors can be specified as absolute, the magnitude of the error in the appropriate
engineering units, or as a relative or fractional error, usually in percent.
HIGH PRECISION
HIGH ACCURACY
HIGH PRECISION
LOW ACCURACY
LOW PRECISION
HIGH ACCURACY
Figure 7.2 Precision and accuracy possibilities.
LOW PRECISION
LOW ACCURACY
244
Table 7.2 Signal-to-noise ratios and corresponding uncertainties.
SNR (single measurement)
1
10
100
1000
Uncertainty
1 (100%)
0.1 (10%)
0.01 (1%)
0.001 (0.1 %)
Chapter 7
Errors may also be categorized as multiplicative or additive. A multiplicative
error, often referred to as a scale or gain error, is proportional to an instrument
reading. An additive error, often referred to as an offset, yields an absolute
uncertainty that is independent of reading. The error on a typical instrument
specification might read, "O.5% of reading O.2% full scale." These are the
multiplicative and additive errors, respectively. Note that the symbol is
redundant; a 0.5% deviation can clearly go in either direction.
The fundamental error limit in a measurement is random noise. If all other
error sources were reduced to zero, the remaining noise would be Gaussian, most
likely Johnson or shot noise associated with our detector. We certainly desire to
have the limit to uncertainty in a system dependent upon noise, as this indicates
that the systematic errors are understood and under control. In such a system
where noise is the predominant error term, the measurement uncertainty is
inversely related to the signal-to-noise ratio (SNR), or
M
. 1
easurement uncertamty = --.
SNR
(7.1)
Thus, an SNR of 10 implies a 10% uncertainty (lcr) in a single measurement,
i.e. , 1. Table 7.2 shows uncertainties for several values of SNR.
We wish to take repeated measurements to enhance the SNR. With multiple
measurements, the signal is additive, and the average of the noise tends towards
zero. The resultant SNR is proportional to the square root of the number of
independent measurements, (SNR DC IN). We then statistically analyze the data
and make the following assumptions:
(I) The data possess a "normal" (Gaussian) distribution of random errors.
(2) The individual measurements are statistically independent (uncorrelated).
(3) The quantity being measured is stationary.
We then apply the analytical tools listed in Table 7.3.
Radiometric Measurement and Calibration
Table 7.3 Mathematical tools for dataset analysis.
Tool
Mean
Variance (sample)
Standard deviation
Standard deviation of mean
Formula
LXi
X = -i-,N=numberof
N
measurements
L(xi-xY
a 2= -",-. __ _
N-l
a=N
a
ax = ..IN
We further explore the three assumptions listed above.
(1) Assurance of a normal distribution is approached by taking a large
amount of data (central limit theorem). Verification of the normal
distribution can be accomplished using a Chi-square (X
2
) test.
245
(2) The measurements must be independent (uncorrelated). In reality, it is
impossible to achieve complete independence because of the exponential
nature of signal changes with time. Spacing the data sampling interval by
one time constant improves the SNR by (N/2)1I2. Note that there is no
prohibition to faster sampling, but the correlation of such measurements
reduces the apparent gain in SNR. The maximum SNR improvement is
proportional to the square root of the observation time and is independent
of the sampling rate.
1
(3) If there is a low frequency drift (lifnoise) in the data, normal statistical
analysis may be compromised and further analysis is mandatory. You
may be able to low-pass filter the data to reduce the random noise, then
use regression analysis (linear, exponential, power, etc.) to fit a curve to
the remaining low-frequency (drift) component. A more sophisticated
method is the analysis of Allan variances, developed for drift assessment
of atomic clocks. Data sets are analyzed over different time frames. The
Allan variance is given by:
(7.2)
Compare with the classical variance
(7.3)
246 Chapter 7
If (J2 (N) / (1 + 1/ IN), then white (classical) noise predominates
over drift (lij), and we can increase the sampling window 'to and therefore the
number of samples N. Several examples of this situation are shown in Fig. 7.3.
The bottom curve is the white-noise case, where the SNR improves as
1 / IN . The other three curves show the effects of varying drift rates. As the drift
rate climbs (i.e., progressively higher lifnoise than white noise) the sampling or
integration time shortens, resulting in a lower SNR.
An important contribution to random error (precision) is the granularity of
the instrumentation. If the markings on the analog meter movement are too
widely spaced, interpolation between divisions may be difficult. Reading a ruler
or a micrometer may be difficult for the same reason. Most electrical
measurements are now taken with digital instrumentation, where quantization
takes place via an analog-to-digital (AID) converter. Quantization noise was
defined in Eq. (5.31) and is
LSB = SIGNALmax
2
n
'
(7.4)
where LSB is the least significant bit and SIGNAL
max
is the full-scale reading.
If this quantization error is significant in a measurement, the use of more bits
(larger n) is indicated. If the presence of sharp quantization levels is apparent in
your data and you cannot get more bits, you can artificially smooth the data by
adding noise. This is called dithering, used to "smear" and mask the appearance
of discrete levels.
EFFECT OF DRIFT ON DATA AVERAGING
- - - - ---: ----- - - - - ---: -- - - - -
, ,
, ,

06 f-+ _\-- __ __ -- ___ --+-_1 _-- ___ -+-_ ___ __ --_. _-- _ _ _ ;-: _ __ -- __ -+--r __ -- ___ : __ -- __ _-7 __ :_-- __ .="" ..
: : : :
0.4 : ___ f..--"'T"
- - - -- - - - - -:- - - - -- -- --;- - -- - - -- -- - -'- - ---- - - - - - -' - -- -
0.2
-
-- --- -- -- ---

o 20 40 60 80 100
NUMBER OF DATA POINTS
Figure 7.3 Drift effects on data averaging for a number of cases.
Table 7.4 Chauvenet's criteria for data rejection.
N umber of readings
4
6
8
10
15
20
25
50
100
200
500
1000
Data point deviation
Standard deviation
1.54
1.73
1.86
1.96
2.13
2.24
2.33
2.57
2.81
3.02
3.29
3.48
247
One is often tempted to reject data that appears to be "out of line." The
general recommendation is DON'T. The data has warts, leave them alone. If you
really must, there are reasonable guidelines for data rejection. Don't make the
mistake of expunging data from your data set, just set the outliers aside. After all,
there might be something really interesting there. The most popular tool is the
Chauvenet criteria for rejection of outliers. Here we reject a data point if the
probability of observation of the suspect point is less than ~ where N is number
of observations. Then we can recompute mean and standard deviation. The
criteria are shown in Table 7.4 above.
Application of Chauvenet's criteria affects standard deviation more than
mean. It is interesting to test the dataset with and without data rejection to see if
it really matters! Usually it doesn't. A final note: even if tempted, do not apply
this technique more than once. If you do, the data will probably be skewed. If one
keeps rejecting data, eventually one will have only a single data point remaining.
To minimize systematic errors, calibrate frequently with suitable apparatus,
as noted above. Table 7.5 lists several sources of systematic error.
The evaluation of systematic error is, in practice, a judgment call, based upon
an investigator's familiarity with previous data, instrument behavior,
manufacturers' specifications, handbook reference data, and calibration. Current
recommended practice is to estimate systematic errors at the 10' (68%
confidence) level.
Systematic errors in optical radiation measurements tend to be large because
the quantities involved are a function of everything in the world; the following
factors are the most common:
(1) Wavelength (broadband or monochromatic)
(2) Power or energy level
(3) Linearity
248
Table 7.S Sources of systematic error.
Environmental factors (temperature, humidity,
pressure, electromagnetic interference, dust and
dirt, power line fluctuation)
Dynamic (response time and slew rate) errors
Offset error (additive)
Gain or scale error (multiplicative)
Hysteresis
Nonlinearity
Faulty procedures
Personal bias
Calibration frequency
Calibration standards
(4) Modulation frequency or pulse characteristics
(5) Position, direction
(6) Polarization, diffraction, coherence, "phase of the moon."
Chapter 7
There are two secondary error categories: illegitimate errors and model
validity. Illegitimate errors include blunders, mistakes, computational errors
(roundoff, etc.), and chaotic errors. Model validity has to do with our
preconceived ideas about the results of a measurement, which influences the way
we conduct the measurement. While these errors may be significant or even
overwhelming in a particular situation, none are essential contributors to the
limits of radiometry and can be eliminated with proper care.
When all of the various error sources have been identified, they must be
combined into an overall uncertainty estimate. Errors propagate according to
well-known formulae as shown in the following tables. Table 7.6 presents the
formulas for single-valued functions.
For combinations of functions, the worst-case ultraconservative treatment is
to directly add the errors. This method would be correct if all of the errors were
correlated. Table 7.7 presents the formulas for functions in which independent
variables are combined.
Table 7.6 Error expressions for single-valued functions.
Constant Reciprocal Power Exponential Logarithmic
(q = Ax) (q = lIx) (q = xn)
q=e
wc
q = In(ax)
a
q
_ ax a
q
_ ax a a a
q
a
q
_
1
ax
jqf-g jqf-g Iql I n l l ~
jqf=aax
jqf aln(ax) Ixl
....
Table 7.7 Error expressions for functions of several variables.
Addition and subtraction (q = x + y - z)
Multiplication and division (q = xy/z)
Most often, the errors are independent or uncorrelated. Under these
conditions, we use the root sum square (RSS) technique shown in Table 7.8.
When all of the random and systematic uncertainties have been assessed, we
combine them into a single uncertainty estimate using the propagation formulas
above. The combined standard uncertainty is the standard deviation of the
measurement. We assess the random and systematic uncertainties independently
at the lcr level, then add the standard deviations in quadrature (that is, we use
RSS) to get the combined lcr uncertainty:
(7.5)
The 1 cr level implies a confidence level that is only 68%. To give more
confidence in the measurements, we multiply sigma by a coverage factor k. A k
of2 is 2cr, where cr is the standard deviation of the measurement. Use a suggested
confidence factor k of 2, which gives a confidence of approximately 95%.
A recommended format for reporting results shows the measured value with
the standard (1cr) uncertainty and a coverage factor k as follows:
(45.26 0.03) W/cm
2
(2cr) or (k = 2) . (7.6)
In standards work, authors frequently split the random and systematic uncertainty
components and report both. For further details, consult Taylor?
Table 7.8 Root-sum-square formulas.
Addition and subtraction (q = x + Y - z)
Multiplication and division (q = xy/z)
250 Chapter 7
7.4 Measurement and Range Equations
The general form of the measurement equation relates the observed SIGNAL
from a radiometric instrument or detector having a responsivity 9\ to an input
radiance L. In differential form, it is
(7.7)
which is not particularly useful. We must integrate over area, solid angle, and
wavelength to get the integral form:
SIGNAL = ffJ9\(A)L" (A)dAdndA.
In both equations above,
SIGNAL = the output radiometric signal,
L;...(A) = the input spectral radiance, and
9\(1..) = is the radiometric system's power responsivity.
The "world" (measurement environment) consists of:
e,<j> angular dependences
x,y position dependences
A wavelength
t, v time, frequency
s,p polarization components
diffraction
nonlinearities
"phase of the moon"
(7.8)
These equations are valid only for incoherent radiation; another layer of
complexity is added when speckle and interference effects are added.
There are many alternate forms of the measurement equation. For a single
detector with spectral responsivity 9\(1..), it simplifies to:
(7.9)
where <1>",(1..) = spectral radiant power incident on the detector and 9\(A) = the
radiometer spectral power responsivity.
For a detector with spectral responsivity 9\(1..) and a narrow-band source or
filter:
d
po
(7.10)
where 'teA) = spectral transmittance of the filter and J:..A = bandwidth of the source
or filter.
For a laser line where the spectrallinewidth is so small that the responsivity
of the detector/radiometer can be considered constant, then
SIGNAL = c I > . (7.11 )
Different radiometric configurations have different measurement equations.
For example, the signal from a distant small object that underfills the radiometer
field of view is described as
where
SIGNAL = the radiometric signal,
h(A) = the spectral radiant intensity of the source,
d = the distance to the source, and
~ A ) = the radiometer spectral power responsivity.
(7.12)
Many other forms can be developed depending upon the chosen measurement
configuration.
Similar to the measurement equation is the range equation, which gives the
distance at which a source will generate a specified SNR. It is useful to visualize
which system parameters are important in particular applications. A general form
is provided by Hudson:
3
{D*)li
(
1 )li
SNRJB
(7.13)
TERM 1 2 3 4
where terms I through 4 are defined as:
Term 1: Target parameters: intensity Is and atmospheric transmission 'ta
Term 2: Optical system parameters:j7#, optics diameter Do, FOV fl, and
transmission 'to
Term 3: Detector parameter: D*
Term 4: Signal processing parameters: required SNR and noise bandwidth B.
252 Chapter 7
This range equation can take many specific forms to determine SNR or range
for various source/radiometer spatial and temporal configurations. Note, also,
that the term intensity implies a point source.
There are several alternate means of assessing the performance of a
radiometric system based on what are called "noise equivalent" quantities, whose
input values produce an SNR of 1 (signal equals noise). Noise-equivalent power
(NEP), for example, was defined in Chapter 5 as
i i
NEP =...!!.... = ...!!...or
9\ is '
v v
NEP = = or
9\ vs'
(7.14)
fAB
NEP = _,,_ fldd
ll
-.
D*
It is usually applied to detectors, but may also be useful for systems designed
to measure radiant power.
We can also define a noise-equivalent irradiance or flux density (NEI or
NEFD), which is frequently used to characterize systems for detection of distant
small sources:
=NEP.
SNR Vs Ad
(7.15)
Note that the term NEI does not reflect our symbol for irradiance E.
For an unresolved target that underfills the FOV, Hudson
3
gives the noise-
equivalent irradiance (W/m2) as
4(f /#)0.
1/2
BI /2
NEI = ---"'---'---''-----
rcDoD * 'to
which bears a striking resemblance to the range equation.
(7.16)
The noise-equivalent temperature difference (NETD or NELll) is used to
predict the performance of thermal radiometers and infrared cameras. It
characterizes a system by its ability to distinguish a small temperature difference
between a resolved target (underfills FOV) and the background. For a single
small detector Holst gives:
4
NETD=
,
C4 S- 't . (A) aM (A, T
B
) D * (A)dA
"fld OpllCS aT
A,
(7.17)
pas
where Tn in the equation is the temperature of the background.
Numerous variations of these system parametric equations can be found in
the literature, each specific for a given application. Each experimenter must
model his or her own system to generate one of these equations for accurate
assessment. Once the basic equation has been generated and tested, we can do
further modeling to include the effects of other variables, such as scene
parameters, environmental influences, degradation, etc.
7.5 Introduction to the Philosophy of Calibration
What is calibration and why do we do it? The format of the raw data from a
radiometric instrument is usually in the form of a digital count or data number for
digital instruments or a voltage, current, or resistance for analog instruments.
These numbers are quite meaningless, inasmuch as the units of radiance are not
volts, and irradiance is not given in digital counts. Therefore, the primary
purpose of calibration is to assign absolute values in engineering units to
measured data according to an accepted standard. A secondary but still important
purpose of calibration is to estimate uncertainties of the acquired data.
Several formal definitions of calibration are provided below.
calibration n: (1) The set of operations which establish, under
specified conditions, the relationship between values indicated
by a measuring instrument and the corresponding known values
of a standard (NASA EOS). (2) The measurement of some
property of an object that yields as an end result a number that
indicates how much of the property the object has (Webster New
Collegiate). (3) The comparison of a measurement standard or
instrument of known accuracy with another standard or
instrument to detect, correlate, report, or eliminate by adjustment
any variation in the accuracy of the item being compared (MIL-
STD-45662A.) (4) A set of operations, performed in accordance
with a definite documented procedure that compares the
measurements performed by an instrument to those made by a
more accurate instrument or standard for the purpose of
detecting and reporting, or eliminating by adjustment, errors in
the instrument tested (Fluke, Calibration). (5) The process of
assigning engineering units and uncertainties to meter
deflections, digital counts, etc., such that an instrument reading
conforms to a recognized standard (Palmer's definition).
The last of these five definitions is the most useful, as it adds reality to the
measurements that we conduct. The common thread in the definitions of
calibration is the involvement of a standard. A valid measurement is inextricably
linked to the calibration process and therefore to physical standards. How well
we can measure is closely related to the quality of the standards we employ, and
254 Chapter 7
future improvements in our measurements will necessitate better standards. A
measurement or physical standard can be defined as an accepted object, artifact,
material, instrument, experiment, or system that stores or provides a physical
quantity that serves as the basis for measurements of the quantity. It is used as a
reference for establishing a unit for the measurement of the physical quantity.
There are several types of standards used by the community. A primary
standard is one that has the highest metrological qualities. It may be realized
from first principles, calculable, or built to plan with no other measurements
required. This type of primary standard is also known as an intrinsic standard. An
example is the degree, which is based upon the triple point of water. Primary
standards may also be established by international agreement as an artifact
standard. An example is the kilogram, a particular artifact stored at the
International Bureau of Weights and Measures (BIPM) in Paris. A secondary
standard is designed to carry and transport a calibration scale. It must be as
repeatable and stable as possible, and is calibrated with reference to a primary
standard. Secondary standards are used to disseminate a scale for widespread
distribution. Finally, a working standard is similar to a secondary standard but is
one generation further removed from the primary standard. These are the
generation that we usually purchase from secondary suppliers of calibration
equipment and standards and use for routine calibrations of our instrumentation.
Note that as one gets further from the primary standard, the uncertainties increase
due to the inevitable errors present in the transfers.
Other terminology is also applied to standards. An international standard is
one which has been adopted based upon an international agreement. A transport
standard is one which has been designed to maintain its calibration through the
rigors of transport via common carrier. A consensus standard is one used by
consenting parties when no suitable standard is available.
u.S. Department of Defense requirements for the use of physical standards
are spelled out in MIL-STD-45662A, which states that they must be:
(1) certified as traceable to the National Bureau of Standards [now National
Institute of Standards and Technology (NIST)), or
(2) derived from accepted values of natural physical constants, or
(3) derived by ratio type of self-calibration techniques.
Traceability is defined as the ability to relate individual measurements to
national standards or nationally accepted measurement systems through an
unbroken chain of comparisons (MIL-STD-45662A). We have a problem here
with requirement (1) because NIST says:
NIST does not define nor enforce traceability except in its NVLAP
laboratory accreditation program. Moreover, NIST is not legally
required to comply with traceability requirements of other federal
agencies; nor do we determine what must be done to comply with
another party's contract or regulation calling for such traceability.
--
However, NIST can and does provide technical advice on how to
make measurements consistent with national standards.
5
There are several possible solutions to this dilemma:
255
(1) We can keep on doing what we have been doing for many years,
claiming our calibrations and measurements are traceable to an agency
that disavows the word. Standards maintained in this way are called Type
I standards.
(2) There have been other attempts to define traceability, and as the above
quotation implies, they have not been entirely successful. Perhaps
another try is in order:
Traceability n: the demonstration that an instrument or artifact
standard has been either calibrated by NIST (or equivalent) at
appropriate intervals or has been calibrated against other
designated standards via an unbroken chain of comparisons. The
designated standard may be a national standard, an international
standard, or a standard based upon fundamental physical
constants.
This new definition allows us to:
(3) Go offshore to another recognized national laboratory with a reputation
for low-uncertainty measurements.
(4) Purchase or generate a Type II standard given sufficient expertise, time,
and funding. Example: freezing-point blackbody radiation simulator
(requires certified pure material), or electrical substitution radiometry
(requires measurement of electrical power). Here we generate our own
standards from first principles (i.e., standard of length by counting
fringes). NIST is moving in this direction as well, conducting research
and providing information that will allow our self determination.
(5) Implement a Type III standard using ratio and self-calibration
techniques.
There is another type of standard in regular use called a procedural or
documentary standard, also called a protocol in Europe. It is a document
outlining operations and processes to be performed in order to achieve a
particular end. These are generated and maintained by several organizations,
including the American Society for Testing and Materials (ASTM), the
International Electrotechnical Commission (lEC), and the International
Commission on Illumination (ClE), to name a few. A prime example is the
aforementioned MIL-STD-45662A.
Over the years I (Palmer) have developed a calibration philosophy which
may be of interest, consisting of the principles in Table 7.9.
256 Chapter 7
Table 7.9 The Palmer philosophy of radiometric calibration.
(1) Calibration is the process of assigning absolute engineering units (i.e.,
radiance, temperature, etc.) to acquired data (volts, digital counts); i.e., it is
the determination of the instrument transfer function.
(2) Calibration requirements are driven by science and engineering goals.
(3) For any instrument, include calibration requirements and methodology in the
initial design phase. Among these considerations is the feasibility of
including an on-board calibrator.
(4) Apply the following general principles in the design of the calibration
exercise:
A. make the calibration independent of the specific instrument;
B. calibrate the sensor in the configuration it will be used;
C. take into account every factor that may influence the calibration: the more
that Principle B is violated, the more important Principle C becomes;
D. calibration involves a comparison with primary or secondary standards;
select appropriate standards.
(5) Conduct an error assessment during the calibration planning phase to allow
estimation of uncertainties on acquired data. Give special attention to model
error.
(6) An end-to-end calibration is preferable to summation of individual
component-level calibrations.
(7) Vary relevant external environmental parameters (temperature, pressure,
humidity, etc.) to determine their influence on the transfer function.
(8) Determine the transfer function over the entire dynamic range of the
instrument.
(9) To maximize confidence in the calibration, use several calibration
configurations and compare the results for consistency.
(10) Prior to the final calibration of a flight instrument, conduct the entire
calibration procedure on a dummy, prototype, engineering model, or
whatever is available, to uncover and fix any problems with the calibrator
and/or procedures:
(11) Inspect and interpret the results early, while the device undergoing
calibration is still in the test position; this allows for an immediate reality
check and timely fix if needed.
(12) Calibration is the last step on the PERT/CPMlGANTT chart prior to
delivery. Because of this precarious position, it is most susceptible to the old
squeeze play, so plan ahead.
(13) Above all, adhere to the KISS (Keep It Simple, Smarty!) principle.
-This was done with imager and sounder prior to the launch of GOES-8 and proved invaluable
(BGG).
-
Statement 4 is a response to the "de-embedding problem"; i.e., that no
system, device, or component has a unique identity outside its environment.
Embedded items will always differ from their de-embedded counterparts due to
mutual interaction/parasitics among various components. Following the
principles listed in the statement will help minimize, although never completely
eliminate, these interactions.
An example: one of the authors (Grant) had the experience of witnessing
electro-optical instrument calibration and test results and comparing them with
results of similar tests after the instrument was integrated to its platform.
Different results were observed due to the different environments. One take-away
lesson is that when developing instrument specifications at the beginning of a
design process, do not fail to include consideration of the manner/platform in
which the instrument will ultimately be used! Add a little extra margin for
postintegration performance, if you can, or at least identify the factors involved.
Further, in order to approach the requirements of calibration philosophy
Statement 4, Principle B, one must choose between a standard source and a
standard detector. If at all possible, choose a standard source when the unknown
quantity is radiance ( extended source) or intensity (point source). Place the
standard next to the unknown and view them sequentially. Examples of standard
sources include blackbody radiation simulators, calibrated tungsten or deuterium
lamps, and calibrated integrating sphere sources. Choose a standard detector
when the unknown quantity is irradiance or a power. Interchange the standard
detector with the radiometer being calibrated. Examples of standard detectors
include electrical substitution radiometers and light-trapping quantum detectors.
The selection of an appropriate configuration to conduct a calibration is
governed both by the desired measurements and by the availability of appropriate
standards. There are several calibration configurations that parallel the
measurement configurations shown in Chapter 6. One of them should suffice for
any radiometric instrumentation calibration. If at all possible, use two or more
configurations to gain additional confidence in the calibration.
7.6 Radiometric Calibration Configurations
7.6.1 Introduction
The selection of an appropriate configuration to conduct a calibration is governed
both by the desired measurements and by the availability of appropriate facilities
and standards. There are a number of different ways to set up a calibration or
comparison source and a radiometer. The five basic calibration configurations
that follow are the most common combinations of radiometer aperture-stop and
field-stop considerations and source distance and size. One of them should
suffice for any radiometric instrumentation calibration. If at all possible, use two
or more configurations to gain additional confidence in the calibration.
258
----------
-----
----
--
---
I
I
I
I
I
FIELD ANGLE 80
I
"" ... ..---------- S
I
--------...... APERTURE
STOP
Figure 7.4 DSS calibration configuration.
7.6.2 Distant small source
Chapter 7
In the distant small source (DSS) configuration shown in Fig. 7.4, the source is
placed at a distance where the inverse square law is valid. The radiometer need
not be focused, but the image of the source must be entirely contained within the
field stop. Then,
9\ - (SIGNAL) S2 ,
E - I
where
9\ = Irradiance responsivity in SIGNALI(W/m
2
),
S = distance from source to radiometer (m), and
I = source intensity (WI sr).
(7.18)
The primary advantage of this configuration is that almost any calibration
source can be used as long as the distance is sufficient to meet the inverse square
law. There are several disadvantages: (1) the signals are typically small, (2) there
may be an intervening atmosphere, (3) one must know S (the error in 9\ is twice
the error in S due to the square term), and (4) a background is present because the
source image does not typically fill the field stop. Examples of DSS sources
include small blackbody radiation simulators and small tungsten lamps at suitable
distances, and the sun and stars.
The DSS configuration can be significantly improved if the small source is
placed at the focal point of a collimator, as shown in Fig. 7.5. As before, the
image of the source must fall within the field stop. The size of the image is equal
to the size of the collimator source multiplied by the ratio of the radiometer focal
length to the collimator focal length. Then
9\ - (SIGNAL) j2 ,
E - I
(7.19)
-
COLLIMATED BEAM
Figure 7.5 Eccentric pupil parabola used in DSS calibration.
where f = focal length of the collimator (m) and 1 = source intensity (W /sr) = LAs.
The result is again irradiance responsivity 9\E in SIGNALI(W/m
2
). The
advantages to this approach include: (1) a controllable atmosphere (vacuum
chamber, if necessary), (2) a controllable background, (3) the distance S is not in
the equation, (4) you can use almost any small source, and (5) the radiometric
signals tend to be larger. The disadvantages include (1) the need to know the
focallengthf(typically a one-time measurement), and (2) the collimator/chamber
hardware can get very expensive. Examples include laboratory calibrators and
low-background test chambers. As to collimators, their basic types are refractive
and reflective.
The basics of a refractive collimator are shown in Fig. 7.6. Its advantages
are: (1) relatively simple alignment due to unfolded path (in the visible, only),
and (2) a simple setup, particularly if off-the-shelf components are used.
The disadvantages include: (1) the wavelength range is limited by the 'teA) of
the lens material, (2) reflection losses occur due to refractive index, (3) ghost
images arise due to reflections from optical elements, (4) antireflection coatings
are wavelength dependent, (5) chromatic aberration from refractive optical
elements occurs, and (6) difficulties in alignment occur if the lens is not
transmissive. However, most of these disadvantages can be dismissed if
operating at a wavelength for which the collimator components are optimized.
Figure 7.6 Basics of a refractive collimator.
260 Chapter 7
Figure 7.7 On-axis reflective collimator.
Disadvantages of the reflective on-axis collimator include: (1) a central
obscuration, (2) diffraction from the secondary mirror mount, (3) a direct path for
stray light from the source, and (4) difficulties in baffling. On the other hand, an
advantage is that wide fields of view are possible with this collimator, shown
schematically in Fig. 7.7.
Several of these disadvantages can be eliminated by use of an off-axis
collimator, shown schematically in Fig. 7.8. For example, stray light is
minimized and an off-axis parabola minimizes aberrations. The reflective off-
axis design has a narrow field of view, however. As in so many other topics
mentioned in this book, tradeoffs and choices must be made.
7.6.3 Distant extended source
In the distant extended source (DES) configuration shown in Fig. 7.9, the distant
source subtends a larger angle than the radiometer field of view, overfilling it.
For this configuration,
9\ L =
SIGNAL
L
where 9\L = radiance responsivity in SIGNALI(W/m
2
sr).
Figure 7.8 Off-axis reflective collimator.
(7.20)
SOURCE
8
FIELD ANGLE
APERTURE
Figure 7.9 DES calibration configuration (adapted from Wolfe and Zissis) .6
261
FIELD
STOP
The advantages of this configuration include: (1) the distance between the
source and the radiometer is not important, and (2) there is no background due to
the fact that the source overfills the field of view. Disadvantages include: (1) an
intervening atmosphere, and (2) need for a large uniform source. Source
examples include White Sands, New Mexico, and a lake of known surface
temperature for remote-sensing applications, and a large integrating sphere,
blackbody radiation simulator, or a white diffuse panel in the laboratory.
7.6.4 Near extended source
In the near-extended-source (NES) configuration shown in Fig. 7.10, an extended
source is placed directly in front of the radiometer undergoing calibration.
Radiation from the source (out of focus) must completely fill the field stop. In
this configuration, Eq. (7.20) applies and the radiance responsivity is in
SIGNALI(W/m
2
sr) as before.
Advantages of this configuration include: (1) the distance between the source
and the radiometer are not important, (2) there is no background, and (3) there is
minimal atmosphere. On the down side, you need a rather large uniform source.
Examples include a large-area blackbody radiation simulator, an integrating
sphere, or a transmission or reflection diffuser used with a standard lamp.
SOURCE
ANGLE
APERTURE
STOP
FIELD
STOP
Figure 7.10 NES calibration configuration (adapted from Wolfe and Zissis).6
262
z
Optic Axis
Lens and
Aperture Stop
Chapter 7
Figure 7.11 NSS (Jones) calibration configuration (adapted from Wolfe and Zissis).6
7.6.5 Near small source
Also called the "Jones method" (as seen in Fig. 7.11) after its ubiquitous
inventor, R. Clark Jones, the near-small-source (NSS) calibration provides
radiance responsivity calculated according to
(7.21 )
where A = aperture area (m
2
) and As = source area (m
2
).
In this approach, the source must be contained within the region bounded by
XZ and YZ; both segments make the angle eo with the optical axis. The chief ray
angle is also eo, which defines the field of view. This is simply a scaling of areas,
and the radiometer is focused at infinity. The radiance responsivity 9\L has units
SIGNALI(W/m
2
sr).
The advantages of the Jones method include (1) minimal atmosphere, and (2)
the possibility of using a small calibration source. The primary disadvantage is
that you must account for background radiation. An example is the use of a small
blackbody radiation simulator that provides radiation to a system having a large
entrance aperture. Appendix H provides additional information on this method.
7.6.6 Direct method
In the direct-method approach, seen in Fig. 7.12, we use a small narrow beam
that underfills both aperture and field stops (for example, a laser). The beam
power is measured with a calibrated detector or laser-power meter. The beam is
then pointed toward the radiometer and the output is measured. The result is a
power responsivity 9\<1> in SIGNALIW.
The primary advantage of this method is its extremely simple setup. The
disadvantages include: (1) visible background radiation, and (2) no accounting
for aperture and field-stop nonuniformities. In addition, lasers can be quite noisy.
...
INCOMING BEAM
LENS AND
APERTURE
STOP
Figure 7.12 Direct-method calibration configuration.
263
To mInImIZe effects from laser drift and noise, use a beam-power stabilizer
(expensive) or a beamsplitter and another stable detector to characterize the
beam-power fluctuations during the measurement. In addition, you must ensure
that saturation of either detector does not occur.
7.6.7 Conclusion
The above calibration configurations yield different types of responsivities, but
under many circumstances we can use the simple equation for transfer of radiant
power in an optical system, Eq. (2.47):
 = LAn .
The An product (T, throughput, etendue) of a radiometer is usually
characterized by the area of its entrance pupil and its field of view. If the
radiometer has both a well-defined aperture A and field of view n, we may
convert from one form of responsivity to another using
9\ = 9\ = 9\L
 A An
(7.22)
These conversions permit the use of such a well-defined radiometer to
measure one quantity using a calibration derived from a different calibration
configuration.
7.7 Example Calculations: Satellite Electro-optical System
An example will help to illustrate some of the equations presented in this chapter.
Consider a satellite in space located 200 km from a spherical source of I-m
diameter that radiates as a blackbody of 2000 K against a background of cold
black space. This is depicted in Fig. 7.13. The satellite contains an electro-optical
system having parameters listed in Table 7.10.
264
o
1

200km

Figure 7.13 Source-satellite configuration.
(1) What is the detector noise-equivalent power, NEP?
Chapter 7
As we are not given specific information regarding voltage or current, it is best to
use the third of Eqs. (7.14):
NEP= = .Jlcm
2
1Hz
D * 10
10
cmHzl /2 /W
=1 X 10-
10
W. (7.23)
(2) Does the source represent a point source for this configuration?
The first thing to figure out is if we are dealing with a point source or an
extended source, as we do not wish to use system performance equations
indiscriminately. We know that the source is "small," given its distance to the
sensor, but we need to determine its relationship to our detector's size.
Table 7.10 Satellite system parameters.
Primary mirror diameter, Do 0.2 m
jl# of optics jl3
Detector active area Ad 1 cm
2
Detector D* 10
10
cm Hz1l2/W
Nominal wavelength range of operation 811m to 12 11m
Electrical bandwidth B I Hz
To do so, we determine the "diffraction-limited spot size" on the focal plane:
using a wavelength of 10 f..lm (center wavelength of our band of interest):
D
b1ur
= 2.441.,(//#) = 2.44(10xl0-
6
m)3 = 7.32xl0-
5
m. (7.24)
Because the diameter of the (diffraction-limited) source image is less than the
dimension of our detector, 1 cm or 0.01 m, this source qualifies as a point source.
(3) What is the system's noise-equivalent irradiance, NEI?
Noise-equivalent irradiance, as stated above, is often used to characterize a
system for its ability to detect distant small sources. From Eq. (7.15), the NEI of
this system may be calculated as
(4) How does this value compare to the NEI obtained from Eq. (7.16),
above?
(7.25)
First, we have to determine the solid-angle field of view of the sensor. We
need to calculate system focallengthf
/ = (//#)xD
o
=3x0.2 m =0.6m. (7.26)
Next, the solid-angle field of view is determined as
A 1 cm
2
1 4
n=_d = =--=2.78xl0- sr.
/2 (60cm)2 3600
(7.27)
If we assume that the transmittance is unity (there is no atmospheric
component, and we will assume unity optical transmission), then NEI becomes
( )
112
4(3) 2.78xl0-4 1112Hz
NEI = 10 1/2 I = 3.183 X 10-
13
W/cm
2
.
1t(20cm)lxlO cmHz W-
(7.28)
Note, although this is not a text on system design, the size of the diffraction-limited "blur" plays a
role in sizing a system's detector(s). Further information on system design is obtained in
references, some listed below and others in the appendices.
--------------------------------_._ ... _ . .. _ .... _-
266 Chapter 7
This result differs from that in Eq. (7.25) by three orders of magnitude.
Rechecking calculations, we find no error causing a discrepancy this large; what
could the problem be?
To answer, look again at Eq. (7.25). NEI was calculated according to the area
of the detector-not the area of the system entrance pupil Ao. If we repeat that
calculation using the correct surface, we obtain
NEI = NEP = lxlO-
1O
W = 3.185xlO-
13
W/cm2
Ao 314cm
2
(7.29)
which is much better. Conclusion: Make sure you are addressing problems with
the mathematical expression which corresponds to the setup/configuration you
are analyzing.
(5) What is the expected signal-to-noise ratio?
The answer to this question requires an inversion of the range equation, Eq.
(7.l3). It is:
(7.30)
It also requires that we know source intensity within the particular spectral band.
This is obtained through
(7.31)
where Ap is the source projected area. From a blackbody radiation calculation
program, Ls = 5.04 x 10
3
W/m
2
sr in the 8- to 12-llm band. The projected area of
the spherical source of I-m diameter is 0.785 m
2
, so
I = 5.04xl0
3
xO.785 = 3956 W/sr.
..
Substituting into Eq. (7.30), assuming unity transmittances and a I-Hz electrical
bandwidth,
SNR = 3956 W/sr It rad(20cm)(lOlO cmHz"2 /W) :=:: 31.
4(3).J2.78 x 10-
4
sr(4xlOl4 cm
2
)l HZl /2
(7.32)
Whether or not a SNR of 31 is adequate depends, of course, on the particular
application.
-
Equation (7.32) may appear, on first glance, to be dimensionally inconsistent,
as it appears to reduce to units of steradians in the denominator. But look at Eq.
(7.30) again, in this way:
SNR= \ .... .
R .... .
Discussing irradiance in Chapter 2, we noted that E=IIR2 can sometimes be
confusing, as to units, and that a different way to consider the expression is
where no may be thought of as the "unit solid angle," having value 1 sr.
Applying this notion to Eq. (7.32) allows for dimensional consistency.
In conclusion, while real-world problems may be different from those
described in this section-for example, they may include source/target
parameters that change with time and sources that differ from blacklgraybodies-
these equations are basic, yet powerful enough to provide the engineer or analyst
a starting point from which to develop solutions.
7.8 Final Thoughts
1 often say that when you can measure what you
are speaking about, and express it in numbers,
you know something about it; but when you
cannot measure it, when you cannot express it in
numbers, your knowledge is of a meager and
unsatisfactory kind; it may be the beginning of
knowledge, but you have scarcely, in your
thoughts, advanced to the stage of science,
whatever the matter may be.
- Lord Kelvin
A measurement of any kind is incomplete
unless accompanied with an estimate of the
uncertainty associated with that measurement.
- James M. Palmer
268 Chapter 7
For Further Reading
Y. Beers, Introduction to the Theory of Error, Addison-Wesley, Reading,
Massachusetts (1957).
P. R. Bevington and D. K. Robinson, Data Reduction and Error Analysis for the
Physical Sciences, 2
nd
Edition, McGraw-Hill, New York (1992).
A. Daniels, Field Guide to Infrared Systems, SPIE Press, Bellingham,
Washington (2006).
Calibration: Philosophy in Practice, 2
nd
Edition, Fluke Corporation, Everett,
Washington (1994).
J. Mandel, The Statistical Analysis of Experimental Data, Dover, New York
(1984).
S. L. Meyer, Data Analysis for Scientists and Engineers, John Wiley & Sons,
New York (1975).
J. R. Taylor, An Introduction to Error Analysis, 2
nd
Edition, University Science,
Mill Valley, California (1997).
J. D. Vincent, Fundamentals of Infrared Detector Operation and Testing, John
H. D. Young, Statistical Treatment of Experimental Data, McGraw-Hill, New York
(1962).
References
1. S. J. Wein, "Sampling theorem for the negative exponentially correlated
output of lock-in amplifiers," Appl. Opt. 28,4453 (1989).
2. B. N. Taylor and C. E. Kuyatt, Guidelines for Evaluating and Expressing the
Uncertainty of NIST Measurement Results, NIST Technical Note 1297,
(1994). Available at http://physics.nist.gov/Pubs/guidelines/contents.html.
3. R. D. Hudson, Infrared Systems Engineering, Wiley & Sons, New York
(1969).
4. G. C. Holst, Common Sense Approach to Thermal Imaging, JCD Publishing,
Winter Park, Florida and SPIE Press, Bellingham, Washington (2000).
5. NIST SP-250 Appendix, U.S. Government Printing Office, Washington D.C.
(1998).
6. G. J. Zissis, "Radiometry," Chapter 20 in The Infrared Handbook, W. L.
Wolfe and G. J. Zissis, Eds., U.S. Government, Washington, D.C. (1978).
Table of Appendices
Adapted by Anurag Gupta
Appendix A: Systeme Internationale (SI) Units for Radiometry and
Photometry .................................................................... 271
Appendix B: Physical Constants, Conversion Factors, and Other Useful
Quantities ....................................................................... 275
Appendix C: Antiquarian's Garden of Sane and Outrageous
Terminology ................................................................... 277
Appendix D: Solid-Angle Relationships .............................................. 283
Appendix E: Glossary ......................................................................... 285
Appendix F: Effective Noise Bandwidth of Analog RC Filters and the
Selection of Filter Parameters to Optimize Signal-to-Noise
Ratio .............................................................................. 297
Appendix G: Bandwidth Normalization by Moments .......................... 305
Appendix H: Jones Near-Small-Source Calibration Configuration ..... 309
Appendix I: Is Sunglint Observable in the Thermal Infrared? ........... 313
Appendix J: Documentary Standards for Radiometry and
Photometry .................................................................... 321
Appendix K: Radiometry and Photometry Bibliography ..................... 341
Appendix L: Reference List for Noise and Signal Processing,
by L. Stephen BelL ......................................................... 357
269
--
--
Appendix A
5ysteme International (51)
Units for Radiometry and
Photometry
Table A.1 SI* base units.
Base quantity
Length
Mass
Time
Electric current
Thermodynamic temperature
Amount of substance
Luminous intensity
Name
meter
kilogram
second
ampere
kelvin
mole
candela
Table A.2 Selected SI-derived units.
Quantity Name Symbol
Plane angle radian rad
Solid angle steradian sr
Energy joule J
Power watt W
Frequency hertz Hz
Electric charge coulomb C
Luminous flux lumen 1m
Illuminance lux Ix
Luminance candela per square meter cd/m
2
Radiant intensity watt per steradian W/sr
Radiance watt per square meter W/(m
2
sr)
steradian
Symbol
m
kg
s
A
K
mol
cd
Equivalent
Nm
J/s
S-l
A-s
cdsr
Irnlm
2
lrnlm
2
sr
* Complete SI information is available on the World Wide Web at www.bipm.fr and at
physics.nist.gov/pubs/sp81Ilsp811.html.
271
272 Appendix A
Table A.3 Si prefixes.
Factor Prefix Symbol Factor Prefix Symbol
10 yotta Y
10-
1
deci d
10
21
zetta Z
10-
2
centi c
10
18
exa E
10-
3
milli m
10
15
peta P
10-6
micro

10
12
tera T
10-
9
nano n
10
9
glga G
10-
12
pico
P
10
6
Mega M
10-
15
femto f
10
3
kilo k
10-
18
atto a
10
2
hecto h
10-
21
zepto z
10
1
deka d
10-
24
yocto
Y
The following tables show radiometric and photometric quantities, symbols,
definitions, and units.
Table A.4 Radiometric quantities.
Quantity Symbol Definition Units
Radiant energy Q joule [J]
Radiant power (flux)
<f>
dqldt watt [W]
Radiant intensity I d<f>l dO) W/sr
Radiant exitance M d<f>l da
W/m
2
Irradiance E d<f>lda
W/m
2
Radiance L d
2
<f>/(da cosadO)) W/m
2
sr
Table A.5 Photon quantities.
Quantity Symbol Definition Units
Photon power (flux)
<f>q
dn/dt Is
Photon intensity I dnldm ISf"s
Photon exitance M dnlda Im
2
s
Photon irradiance Eq dnlda Im
2
s
Photon radiance Lq ctn/(da cosadO)) Im
2
sf"s
n = photon number.
Systeme International (SI) Units for Radiometry and Photometry 273
Spectral Quantities
Spectral quantities are derivative, per unit wavelength with the additional
dimension m-
I
, and are indicated by a subscript A (e.g., spectral radiance LAwith
units W/m
3
sr). Nonspectral quantities that are wavelength dependant are
indicated as (A); e.g., transmission 'teA).
Photometry is the measurement of light (optical radiant energy as above, but
weighted by the response function of the human eye). The symbols used are the
same as radiometric quantities with the subscript v (for visual) added.
Table A.6 Spectral quantities.
Quantity Symbol Units
Luminous power
<l>v
1m
Luminous exitance Mv lmlm
2
Luminous incidance Ev lmlm
2
Luminous intensity (SI base unit) Iv
lmlsr = cd
Luminance Lv
lmlm
2
sr = cd/m
2
------------------------------------------------------------------------------------------------
..
Appendix B
Physical Constants,
Conversion Factors, and Other
Useful Quantities
Table 8.1 1998 CODATA recommended values of the fundamental physical constants.
Quantity Symbol Value Units
Relative
uncertainty
Speed of light C, Co 299,792,458 m/s exact
(vacuum)
Permeability of
/lo
4n x ]0-7 N/A2
exact
vacuum
Permittivity of vacuum eo 1I1.V
2
= 8.854 187 ... xlO-
12
F/m exact
Planck constant h 6.62606876 (52) x ]0-34
Js
7.8 X ]0-8
Electronic charge q,e 1.602176462 (63) x ]0- 19 C
3.9 X ]0-8
Boltzmann constant k 1.3806503 (24) x 10-
23
J/K
1.7 x ]0-6
Boltzmann constant k 8.617342 (15) x ]0-5 eV/K 1.7 x 10-6
Stefan-Boltzmann (J 5.670400 (40) x 10-
8
W/m2K4
7.0xl0-6
constant
First radiation constant
CI
3.74177107 (29) x ]0-16 Wm
2
7.8 x 10-
8
(2nhc
2
)
First radiation constant
CIL 1.191042722 (93) x 10-
16
Wm
2
/sr 7.8 x ]0-8
for L).
Second radiation C2 1.4387752 (25) x 10-
2
m-K 1.7 x 10-
6
constant
Wien displacement b 2.8977686 (51) x ]0-3 mK 1.7 x ]0-6
law constant
These are the 1998 CODATA recommended values of the fundamental physical
constants. Adapted in part from P. J. Mohr & B. N. Taylor, "The fundamental
physical constants," J. Phys. Chern. Ref Data 28, 1713 (1999), and Rev. Mod.
Phys. 72, 351 (2000). These constants are also available in Physics Today, 54
(Part 2), BG6 (2001), reprinted yearly, and from http://physics.nist.gov/constants.
275
276 Appendix B
Here are some useful conversion factors:
hc = 1.986445 X 10-
25
Jm = 1.986445 x 10-
19
Jllm = 1.986445 x 10-
16
Jnm
hc/q = 1.23984 eYllm
kT/q = 0.025852 Y at 300 K
1 eY = 1.602176462 x 10-
19
J
1 astronomical unit (AU) = 1.495 x lOll m
"'-maxT= c2/4.96511423 ...
Appendix C
Antiquarian's Garden of Sane
and Outrageous Terminology
Perhaps the most difficult task in both teaching and learning about radiometry
and photometry is learning and conveying an appropriate and sensible system of
symbols, units, and nomenclature. This can be a formidable task because of the
enormous extent of these found in the literature. I have attempted to be
consistent with the accepted units in this text and have addressed the situation
with regard to intensity as well . The following is a collection of terms, symbols,
and units that I have gathered with little effort. Perhaps you can add some more
to this list. Some are still current and some are long obsolete.
Photometry
Perhaps in no scientific field is the language more obtuse than in photometry.
This is in large measure because of the tortuous path of the development of
suitable standards.
Luminous intensity
The SI base unit of luminous intensity is the candela (cd).
1 Bougie decimale = 1.02 cd.
1 Bougie nouvelle = 1 cd.
1 International candle = 1.01937 cd (Average of candle standards of the
U.S., U.K., and France).
I new candle = I cd.
1 Carcel = 10 cd.
1 Carcel unit = 9.79613 cd (The measure of a Careel lamp burning calza
oil).
1 hefnerkerze = 0.903 cd (German measure of luminous intensity from
1884 to 1940 = 0.903 cd or 0.92 cd. (Replaced by the candela).
1 violle = 20.4 cd.
1 Pentene candle = 1 cd.
I English sperm candle = 1 cd.
277
278 Appendix C
Table C.1 Some interesting numbers from Phillips Lighting Company.
Bicycle headlamp without reflector, in any direction
Bicycle headlamp with reflector, center of beam
Incandescent reflector lamp PAR38E Spot 120 W,
center of beam
Lighthouse, center of beam
Luminous power
2.5 cd
250 cd
10,000 cd
2,000,000 cd
The (derived) SI unit ofluminous flux (power) is the lumen (1m).
1 1m = 1 cdsf.
A light watt is a unit of radiant power weighted by human-eye response. One
light watt is the power required to produce a perceived brightness equal to that of
light at a wavelength of 555 nm and luminous power of 683 1m. Symbol for light
watt is <l>v.
830
<l>v =Km f <l>Y(A)dA .
360
Mechanical equivalent of light is 1/683 W 11m.
Illuminance
The (derived) SI unit of illuminance is the lux (Ix = Imlm
2
).
1 footcandle (fc) = 1 1m per square foot.
1 lux (Ix) = 1 Irnlm
2
= 1 meter-candle.
1 phot (ph) = llrnlcm
2
= 10
4
Ix.
1 milliphot (mph) = 10-
3
Imlcm
2
.
1 nox = 1 millilux = 10-
3
Ix.
1 sea-mile candle = 1 cd @ 1 nautical mile (6,080 ft) = 2.9 x 10-
7
Ix.
1 pharosage = 11mlm
2
.
Luminosity L is expressed in Imlfe.
Luminance
The (derived) SI unit of luminance is the nit (cd/m2).
1 nit = 1 candela per m
2
= 1t apostilb = 0.2919 foot-lambert (fL).
1 stilb (sb) = 1 candela per cm
2
.
1 nit = 10
4
Bougie-Hectometre-Carre.
Antiquarian's Garden of Sane and Outrageous Terminology 279
Note: Several luminance units are related to the illuminance units by assuming a
perfect (p = 1) diffuse (Lambertian) reflector. This "simplification" leads to:
I foot-candle (fc) of illumination 1 fL ofluminance.
1 lambert (L)= 1 Imlcm
2
= (lin) cd/cm
2
.
1 footlambert (fL) = (l In) cd/ft2.
1 apostilb (asb) = (lin) cd/m
2
= (lin) nit.
1 skot = 10-
3
(lin) cd/m
2
= 10-
3
apostilb.
1 millilambert '" 1 fL.
1 equivalent phot = 1 L.
I equivalent lux = 1 blondel = I asb.
1 equivalent footcandle = 1 fL.
The unit bril is used to express the "brilliance" or subjective brightness of a
source of light:
bril = 10gL + 100 .
log2
The scale is logarithmic: an increase of I bril means doubling the luminance
emitted by the source. A luminance of 1 lambert (L) is defined to have a
brilliance of 100 brils.
Luminous energy
Luminous energy is radiant energy weighted by the visual response of the eye.
The (derived) SI unit of luminous energy is the talbot (lms).
I talbot = 107lumergs.
10
7
erg = 1 Ws.
1 phos = 1 talbot.
Vision Research
troland: 1. Retinal illuminance produced by luminance of 1 cd/m2 if
entrance pupil of eye is 1 mm
2
, corrected for the Stiles-Crawford effect;
formerly called the photon. 2. The external illuminance that produces
retinal illumination of 0.002 Ix.
280 Appendix C
Ultraviolet
E-viton is erythemal effectiveness equivalent to 10 IlW at 296.7 nm.
1 Finsen = 1 E-vitonlcm
2

1 erythemal watt = 10
5
E-vitons.
1 EU = 1 E-viton = 1 erytheme.
Floren is UV flux equivalent to 1 m W between 320 and 400 nm.
Bactericidal microwatt is weighted by bactericidal action spectrum.
Ultraviolet microwatt or UV watt is evaluated at 253.7 nm.
MPE (minimum perceptible erythema) = 0.025 erythemal W/cm
2

1 MPE = 2500 finsens = 2.5 x 10
5
erg/cm
2
at 296.7 nm.
One minimum erythemal dose (MED) is the dose required to produce a minimum
redness on sun-sensitive skin. Its value is dependent on skin type. For the most
sensitive skin type it is 200 J/m
2
, weighted by the standard erythemal action
spectrum. For less-sensitive skin types, it rises to 1000 J/m
2

1 MED = 2 SED (standard erythemal dose).
At the wavelength of maximum sensitivity for production of erythema (295 nm),
the MED is 50 J/m
2
. [Br. J. Dermato!' 82,584 (1970).]
Shade number is a unit of light transmission for the protective glasses used in
welding. If T is the fraction of visible light transmitted, the shade number is
1 + 7(-loglOn/3. For example, if 1 % of the light is transmitted, the shade
number is 4.
Astronomy
1 Jansky (Jy) = 10-
26
W/m
2
Hz (spectral irradiance).
1 W/cm
2
f.1m = 3 x 10
16
/1..1 Jy.
1 Solar flux unit (s.f.u.) = 10
4
Jy.
Visual magnitude zero = 2.65 x 10-6 Ix outside atmosphere (Infrared
Handbook, pp. 3-23).
Visual magnitude zero = 2.54 x 10-6 Ix outside atmosphere (radiometry
and photometry in astronomy).
Visual magnitude zero = 2.09 x 10-6 Ix outside atmosphere.
....
Antiquarian's Garden of Sane and Outrageous Terminology 281
1 Rayleigh = 10
6
photons/cm
2
s.
I Rayleigh = (1I4n) x 10
6
photonslc
2
ssr.
I SIO = 1.23 X 10-
12
W/cm
2
sq.tm at 0.55 /lm (equivalent to the number of
10
th
magnitude stars per square degree).
1 SIO = l.899 X 10
6
photons/s cm
2
sr-/lm at 0.55/lm.
Color and Appearance
Reciprocal megakelvin (MKr
l
= 106/Tc , where Tc is color temperature, also
known as mirek or mired (microreciprocal kelvin or microreciprocal degree).
Miscellaneous
1 microeinstein (/lE) = 6.022 x 10
17
photons = 1 micromole.
angstrom (A) is obsolete unit of wavelength = 10-
10
m.
Kayser is waves per centimeter.
Gillette is a measure oflaser energy, sufficient to penetrate one standard
razor blade.
Microflick is a unit of spectral radiance /lWlcm
2
sr/lm.
Spectrallamprosity is in youngs per watt.
Lamprosity (y) is in youngs per radiated watt.
1 = 1 lumens per input watt.
Photosynthetic photon flux density (PPFD) is measured in mol/m
2
s.
Spherical photosynthetic photon flux density (SPPFD) is measured in
mol/m
2
s.
Photosynthetically active photons (PAP) is measured in mol/m
2
.
Spherical photosynthetically active photons (SPAP) is measured in
mollm
2
.
Irradiation = radiant pharosage = radiant incidance (W/m2).
Radiosity = radiant pharosage = radiant exitance (W /m
2
).
Phengosage = spectral pharosage.
Radiant pharos = radiant power (W).
1 W/m
2
= 0.317 BTU/ft2hr.
1 langley = 1 gmcal/cm2.
1 langley/minute = 697.3 W/m
2
.
1 pyron = 1 calorie/cm2min = 697.633 J/m
2
s used to measure heat flow
from solar radiation.
Radiant phos is exposure (Ws).
Radiant helios is radiance (hershel).
282
I pharos = I lumen.
I helios = I blondel.
I heliosent = blondel/m.
Radiant heliosent = path radiance (hershel/m).
Luminous efficiency is luminous flux/radiant flux.
Luminous efficacy is luminous flux/electrical input power.
Appendix C

Appendix D
Solid-Angle Relationships
e (deg) e (rad) ro (sr) o (sr) mlO jl#
NAln*
0.573 .0100 .00031 .00031 1.000 50.00 0.010
1.000 .0175 .00096 .00096 1.000 28.65 0.017
1.146 .0200 .00126 .00126 l.000 25.00 0.020
l.719 .0300 .00283 .00283 1.000 16.67 0.030
2.000 .0349 .00383 .00383 l.000 14.33 0.035
2.292 .0400 .00503 .00502 l.000 12.50 0.040
2.865 .0500 .00785 .00785 1.001 10.00 0.050
3.000 .0524 .00861 .00861 l.001 9.554 0.052
4.000 .0698 .0153 .0153 l.001 7.168 0.070
5.000 .0873 .0239 .0239 l.002 5.737 0.087
5.730 .1000 .0314 .0313 l.003 5.008 0.100
10.00 .1745 .0955 .0947 1.008 2.880 0.174
11.46 .2000 .1252 .1240 l.01O 2.517 0.199
15.00 .2618 .2141 .2104 1.017 l.932 0.259
17.19 .3000 .2806 .2744 l.023 l.692 0.296
20.00 .3491 .3789 .3675 l.031 1.462 0.342
22.92 .4000 .4960 .4764 l.041 1.284 0.389
25.00 .4363 .5887 .5611 l.049 1.183 0.423
28.65 .5000 .7692 .7221 1.065 1.043 0.479
30.00 .5236 .8418 .7854 l.072 l.000 0.500
34.38 .6000 1.097 1.002 1.096 0.886 0.565
40.11 .7000 1.478 1.304 l.l33 0.776 0.644
45.00 .7854 1.840 1.571 1.172 0.707 0.707
45.84 .8000 l.906 l.617 1.179 0.697 0.717
51.57 .9000 2.377 1.928 1.233 0.638 0.783
57.30 l.000 2.888 2.224 1.298 0.594 0.841
60.00 l.047 3.142 2.356 1.333 0.577 0.866
71.63 1.250 4.269 2.830 1.521 0.527 0.949
85.95 1.500 5.839 3.126 1.868 0.501 0.997
90.00 1.571 6.283 3.142 2.000 0.500 1.000
"To obtain the numerical aperture NA, numbers in this column must be
multiplied by the index of refraction n of the local media.
Adapted from F.E. Nicodemus et at., Self-Study Manual on Optical Radiation
Measurements, NBS Technical Note 910-01, National Institute of Standards and
Technology, Washington, D.C. (1976).
283
-
Appendix E
Glossary
lifnoise
An product
Bandwidth normalization
Blackbody radiation
simulator
Background-limited
infrared photodetector
(BLIP)
Bode plot
Weird, ubiquitous noise from many familiar and
strange sources, inversely proportional to
frequency (Pink or red noise). An approximation
. -:z ( const )( I:C
B
) .
IS III f = , where a IS between 1.25
f
and 4 (typically 2), and is between 0.8 and 3
(typically 1). Also called flicker, contact, excess,
modulation, etc. A major pain!
Symbol T, units m
2
sr; geometrical term relating to
amount of power that can get through a system;
Also throughput, etendue.
Determination of an equivalent responsivity using
a rectangle; the areas under the curve and
rectangle are set equal.
An object that simulates blackbody (Planckian)
radiation via careful cavity design and
temperature measurement.
One whose noise is predominantly due to the
noise in the incident photon stream, not intrinsic
to the detector.
Plotting log(signal or noise) versus
log(frequency); shows many orders of
magnitude, asymptotes of linear plots map to
straight lines on Bode plot.
285
286
Bidirectional reflectance
distribution function
(BRDF)
Brightness temperature
Bidirectional transmission
distribution function
(BTDF)
Charge transfer efficiency
(CTE)
Chopping factor
CIE chromaticity diagram
Cold filter
Cold stop
Collimator
Color temperature
Appendix E
A directional quantity that denotes output radiance
as a function of direction and irradiance. A
perfectly diffusing reflector has a BRDF of p/1t,
while a perfectly specular reflector has a BRDF
p/Q; P is reflectance and Q is the projected solid
angle of the source. Units: sr-
1

The brightness temperature of an object is the
temperature of blackbody radiation that has the
same spectral radiance as the object.
The angular distribution of transmitted radiance
around the normal transmitted beam. Units: sr-
1

Fraction of charge that is successfully transferred
from one CCD charge storage element (potential
well) to the adjacent charge storage element.
Ratio of the rms amplitude of the fundamental
frequency component of a modulated signal to
the peak-to-peak amplitude of the unmodulated
signal. Equal to 0.45 for square wave chopping.
A horseshoe-shaped diagram showing the gamut
of all possible colors in terms of hue and
saturation. The horseshoe is the spectrum locus
and white is at the center where x = y = z = 113.
A filter that passes desired bandpass and is cooled
to minimize self radiation at other wavelengths.
An aperture placed in front of a detector to limit
the field of view and cooled to minimize the
stop radiance. Improves SNR.
An optical system designed to make a near small
source appear as if it were located at infinity.
The color temperature of an object is the
same chromaticity (color) as the object.
Glossary 287
Conduction calorimeter A device to measure laser power and energy by
calorimetric means, i.e., heating of an absorber.
Contrast sensitivity function The visual acuity of the eye as a function of both
spatial frequency and contrast. Determined by
looking at variable frequency cos
2
wave patterns
that have contrast decreasing from bottom to
top.
Correlated color The temperature of a blackbody having a
temperature chromaticity (color) as close as possible to the
chromaticity of the source in question.
Cosine response The response curve desired for an instrument
designed to measure irradiance from a
hemisphere. Related to projected area.
D Detectivity, the reciprocal of noise equivalent
power (NEP), the input power for which the
SNR is 1. Unit: W-
I
.
D* Specific or normalized detectivity. It is detectivity
D = NEP-
1
normalized for bandwidth and area.
D*= (AB)1/2INEP. Unit: cmHz
I/2
1W. It is the
SNR per watt for a l-cm
2
detector with a
bandwidth of 1 Hz. Allows a fair comparison
between detector types.
D** A normalized detectivity, taking into account
detector area, noise bandwidth, and field of view
(FOY). It is the SNR per incident watt for a
l-cm
2
area, I-Hz noise bandwidth, and n sr ofa
projected solid angle.
D**= (ABQ./n)
II2
INEP. Unit: cm'Hz
1/2
1W. A
further normalization for field of view:
D*(Q.ln) 1/2,
D*BLlP Background-limited infrared photodetector, the
best SNR you can get when photon noise from
the background limits detection.
Decade, octave Frequency ratio of 10 and 2, respectively,
288
Decibel (dB)
Detective quantum
efficiency (DQE)
Diffuse reflectance
Distribution temperature
Effective noise bandwidth
Electrical substitution
radiometer
Appendix E
A ratio of two voltages, currents, or powers.
12 Pz
dB = 2010g
lo
- = 2010g
lo
- dB = lOloglO-

It is a relative measure. There are several ways of
denoting absolute values.
dBv refers to I-V rms.
dBm refers to 1 m W with stated impedance
(75 Q).
To compare unlike waveforms, such as a sine
wave to Gaussian noise, use the power
formulation.
A relative measure of the amount of noise added
by a detector. Detective quantum efficiency is
like RQE but includes noise, (SNR
ou
i l(SNR;n)2,
unitless, 0 < DQE < I .
Ratio of radiation reflected into a hemisphere
(whose base is the reflector) to the incident
radiation. Excludes specular component.
The distribution temperature of an object is the
temperature of a blackbody radiator that has the
same (or nearly the same) relative spectral
distribution over a substantial portion of the
spectrum as the object.
The equivalent square-band power bandwidth,
used for evaluation of noise. Given for "white"
noise by
ENB=_1_2' [[ A (/)Td/ .
IAol
The equivalent "brick-wall" rectangular
passband. Alternate symbols are Band 11f, units
are Hz.
A radiometer based upon a thermal detector with
provisions for injecting a known power via
electrical means for the purpose of calibration.
Glossary
Full well capacity
Generation-recombination
nOise
H-D curve
Hemispherical reflectance
Iluminance
Intensity
Irradiance
Isotropic point source
Johnson noise
Jones calibration
configuration
289
The number of electrons (signal + noise + dark
current) that a potential well in a CCD structure
can hold.
Noise due to the generation of carriers by photon
absorption and by recombination random
motion of carriers (electrons) in a resistive
material. Spectral power density depends on
frequency.
The characteristic curve of photographic film
(named after Hurter and Driffield) which plots
density log(l/t) versus log of the exposure
(product of irradiance and time).
Directional reflectance integrated over an entire
hemisphere.
Luminous flux per unit incident on a surface from
a hemisphere; units Imlm
2
= Ix. Analagous to
irradiance in W 1m
2

Symbol I, watts per unit solid angle, often from an
isotropic "point" source W/sr.
Symbol E, units W/m2, watts per unit area
incident on a surface.
Small (relative to distance) source where intensity
is independent of direction.
Noise due to random thermal agitation of carriers
(electrons) in a resistive material. Spectral
power density independent of frequency (i.e.,
- -
white). v ~ = 4kTRB or i ~ = 4kTB I R .
Also known as near small source. The radiometer
is focused at infinity. A small calibration source
is placed within a cone whose base is the
entrance aperture and whose half angle is the
chief ray angle. It fills a fraction of the entrance
aperture. The calibration equation is
9\ = ( SIG;AL )( ).
290
Lambertian source
Laser calorimeter
Luminance
Luminous intensity
Measurement equation
Moments normalization
Noise-equivalent photon
flux
Noise-equivalent power
(NEP)
Noise-equivalent
temperature difference
(NETD, NEllI)
Passband normalization
Appendix E
Source where radiance is independent of
direction.
A device to measure laser energy, particularly for
short pulses, by calorimetric means, i.e., heating
of an absorber. Output proportional to time
integral of power, i.e., energy.
Photometric equivalent of radiance. Measure of
visible power per unit-projected area per unit
solid angle. Unit: candela m-
2
= Imlm
2
sr.
Measure of visible power per unit solid angle.
Unit: candela (cd) = Imlsr. One of the seven SI
base units. Analogous to radiant intensity, W /sr.
An equation that relates the output signal from a
detector or radiometer to a function of the
receiver and source spectral parameters. An
example:
SIGNAL = An r LA ~ A ) d A .
A normalization based on a moments analysis of a
spectral responsivity. The center wavelength is
the centroid and the bandwidth, and cut-on and
cutoff wavelengths are computed from the
vanance.
The photon flux incident on a detector which
gives rise to a signal-to-noise ratio of one. Units:
- I
s .
The power incident on a detector which gives rise
to a signal-to-noise ratio of 1. Units: W.
The temperature difference between the target and
the background that produces an rms signal
equal to the rms noise.
Normalization method wherein the band limits of
the equivalent rectangle are assigned at fixed
response points (50%, 10%, lie, etc.). The
normalized responsivity is then related to the
area under the response curve.
Glossary
Peak normalization
Photoconductive gain
Photon noise
Photopic
Photopic visibility curve
Precision
Projected solid angle
Quantum efficiency
Quantum trap detector
Radiance
Radiance temperature
Radiant exitance
291
A normalization where the band responsivity is
set to the peak of the actual response curve. The
bandwidth is calculated by matching the area of
the equivalent rectangle to the area under the
response curve.
The ratio of the transit time to the carrier lifetime
for a photoconductive detector. A measure of
the number of electrons a single absorbed
photon can generate.
Noise due to the random arrival of photons;
manifest as shot noise or the G component of G-
R noise.
Pertaining to light-adapted vision.
The relative spectral responsivity of the
standardized light-adapted human eye (cones).
Symbol is V(A); dimensionless.
A measure of the repeatability of a measurement.
Comes from granularity, noise, etc. Determined
and enhanced by repeated measurements.
Solid angle x cos9, projected onto flat surface
dO. = dro cos9. Symbol 0., units sr.
see Responsive quantum efficiency.
A multiple-detector array where detectors are
placed in series optically and in parallel
electrically. Has a quantum efficiency
approaching unity.
Fundamental quantity of radiometry, "brightness."
Symbol L, units W/m
2
sr.
The radiance temperature of an unknown object is
the temperature of a blackbody that has the same
spectral radiance as the unknown object.
Radiant power per unit area leaving a source into
a hemisphere. Symbol M, units W/m
2

292
Radiant intensity
Radiation reference
Radiation temperature
Range equation
Ratio temperature
Reflectance factor
Responsive quantum
efficiency (RQE)
Responsivity
Retroreflectance
Appendix E
Radiant power per unit solid angle. Symbol I,
units W/sr.
A comparison source for a radiometer; zero-based
for short-wave radiometers, a known thermal
source for long-wave radiometers.
The radiation temperature of an object is the
same total (integrated over all wavelengths)
radiance as the object.
An equation that gives the distance from a source
that one can detect with a stated SNR.
The ratio temperature of an object is the
same ratio of spectral radiances at two
wavelengths as the object.
Ratio of flux reflected from a sample to the flux
that would be reflected from a perfect diffuse
reflector (Lambertian, p = 1).
The number of independent output events per
incident photon. Dimensionless, between 0 and
1. Symbols: 11, RQE.
Ratio of the output of a detector to its input. Units:
NW or V /W. Symbol is 9\. The result of an
integral over wavelength:
9\ = f9\(A )<1\ dA
f<l>A dA
A reflectance wherein the reflected beam retraces
the path from the source back to the source.
--
Glossary
rms (root mean square)
Saturation exposure
Shot noise
Solid angle
Spectral directional
emissivity
Spectral radiance
Spectral radiant intensity
Spectral reflectance
293
A measure of the equivalent heating effect of a
voltage or current v"ms = ~ f v
2
(t )dt , where t
is time, usually a single or integer multiple of
periods for periodic waveforms. For a sine
wave, the rms voltage is 1 I J2 times the peak
voltage (measured from 0), or 1/2J2 times the
peak-to-peak voltage. Electric power companies
deliver 117 Vrms. which is 330 V
p
p
For a square
wave, the rms amplitude is the same as the peak
amplitude.
The exposure (product of irradiance and time)
necessary to saturate an integrating detector.
Used for CCD, CID, and CMOS array detectors.
Noise associated with current flow across a
potential barrier, due to discrete nature of
electrons. Power spectral density is independent
of frequency (i.e., white). Mean-square shot-
noise current, i; = 2qJ de B.
Projected area + (distance?, given by
21t(1-cos 9
112
) for right circular cone. Symbol OJ,
units, sr.
The ratio of radiance at a specified wavelength in
a particular direction to that of a blackbody, at
the same wavelength and in the same direction.
Symbol L')..., watts per unit area per unit-projected
solid angle per unit-wavelength interval;
fundamental unit of radiometry.
Symbol h, units W 1sT" m watts per steradian per
unit wavelength, often from an isotropic "point"
source.
Ratio of reflected power at a specific wavelength
to the incident power at the same wavelength.
294
Spectral responsivity
Throughput
Time constant
Total hemispherical
emissivity
Transimpedance amplifier
Trap detector
Type A error
Type B error
Appendix E
Responsivity as a function of wavelength. Symbol
9\("-), units: amps (or volts, etc.) per watt. Not a
derivative quantity, per-unit wavelength
interval.
Ratio of radiation reflected in the mirror direction
to the incident radiation. Excludes diffuse
(scattered) component.
Geometrical term relating to amount of power that
can get through a system. The product of area
and projected solid angle AQ. Symbol T, units
m
2
sr.
A measure of the speed of response of a device.
The time required to reach (I-lie) or 0.632 of
the final value in response to a step input.
Symbol 't, units s. For a simple RC circuit,
't=RC.
Integral of spectral directional emissivity over the
entire spectrum and over an entire hemisphere.
Also, current-to-voltage converter, used to
interface with current sources.
R
E=iR
A mUltiple detector array where detectors are
placed in series optically and in parallel
electrically. Has a quantum efficiency
approaching unity.
Older term: precision. Also, random error. A
measure of the repeatability of a measurement.
Comes from granularity, noise, etc. Determined
by repeated measurements.
Older term: accuracy. Also, systematic error or
bias. A measure of the difference between the
mean reading and "truth." May be corrected by
careful calibration and characterization.
Glossary
Uncertainty
Wien's displacement law
Wien approximation
295
The total estimated difference between a
measurement and "truth"; includes both random
and systematic error terms and a confidence
parameter.
Product of wavelength and temperature constant.
An equation representing blackbody radiation at
short wavelengths and/or low temperatures.
Form is:
Appendix F
Effective Noise Bandwidth of
Analog RC Filters and the
Selection of Filter Parameters
to Optimize Signal-to-Noise
Ratio
Introduction
By James M Palmer
Revised by L. Stephen Bell, May 2009
Engineering calculations involving noise and signal-to-noise ratio need to use the
effective noise bandwidth (ENB) in order to calculate noise properly. Often the
conventional (-3-dB voltage, -6-dB power) bandwidth is used, leading to
erroneous results. When the difference between -3-dB bandwidth and ENB is
recognized, it is often oversimplified by attempting to relate ENB to the -3-dB
bandwidth. Table F.1 shows several of these relationships found in the open and
corporate literature.
These discrepancies, while not extremely serious, are disconcerting,
particularly for the two-section filter, which is readily realizable with a single
operational amplifier and a handful of Rand C components. So I set off to find
out which values are correct. This was accomplished by means of simple
spreadsheet analysis and BASIC computer programs to do the necessary
integrations.
There are many multiple-pole filter types in the literature. This appendix is
limited to simple Butterworth RC filters (maximally flat-frequency response),
where each section has the same Rand C and therefore the same cutoff
frequency. The primary emphasis is limited to "white" noise with a uniform
power spectral density. We conclude with a simple bandwidth optimization to
maximize the SNR ofa single-frequency signal in the presence of white noise.
297
298 Appendix F
Table F.1 Ratio of ENS to -3-dS bandwidth for low-pass filters.
Sections A,B C,D E
1 1.57 1.57 1.57
2 1.22 1.11 1.12
3 1.15 1.05 1.08
4 1.13 1.025 1.06
5 1.11 1.05
A: C. D. Motchenbacher and J. A. Connelly, Low Noise Electronic System
Design, Secs. 1-4, John Wiley & Sons, New York (1993).
B: H. W. Ott, Noise Reduction Techniques in Electronic Systems, Table 8-2,
John Wiley & Sons, New York (1976).
C: P. Horowitz and W. Hill, The Art of Electronics, Sec. 7.21 , Cambridge Univ.
D: "Measuring noise spectra with variable electronic filters," Ithaco Application
Note IAN-I02, Appendix I, Ithaco Corp., Ithaca, New York (July 1983).
E: E. L. Dereniak and G. D. Boreman, Infrared Detectors and Systems, Table
5.1, John Wiley & Sons, New York (1996).
Definitions
The signal bandwidth of a low-pass filter is defined as the frequency where the
signal voltage falls to -3 dB (1I.J2 = 0.707) of the transmission at dc (typically
unity). This frequency is also referred to as the half-power point, where the
power transmission has dropped to -6 dB (0.5). For a bandpass filter, the same
terms are used relative to the peak transmission of the filter. The bandwidth is
often referred to as full-width at half-maximum (FWHM) for power.
The signal bandwidth is not appropriate for characterization of noise. The
effective noise bandwidth (ENB, more often seen as B or !:::.J) is defined as the
equivalent brick-wall (rectangular) filter, having the same area under the power
transmission curve. The equation is
B = 1 r G(j)V(j)2 dj ,
G(fo) ~
(F. 1)
where G(f) is the power gain as a function of frequency, and fa is the frequency
where Glf) is a maximum. v (/)2 is the power spectrum for the noise under
consideration, and v 0
2
is the noise power at the peak frequency. For "white"
(spectrally flat) noise, Eq. (F.1) simplifies to
B=_(1 ) r G(j)dj,
Gfo
(F.2)
Effective Noise Bandwidth of Analog RC Filters ... 299
and this format is the most often seen. In electro-optical systems, this bandwidth
may be limited by the frequency response of a detector or its associated
electronics, or by the insertion of an electrical filter operating in the audio to low-
radio frequency range, 20 Hz to several MHz.
The signal-to-noise ratio (SNR, SIN) of a system is the ratio of the rms signal
current i
sig
to the rms noise current in. Voltage may be substituted for current.
SNR is dimensionless. Use of rms is indicated as it is the only measure
appropriate for the characterization of random noise.
Low-Pass Filters
The voltage transmission of a single-section (first-order) RC low-pass filter is
(F.3)
where (0 is the radian frequency equal to 2nf in Hz, and 't is the circuit time
constant, the RC product. At low frequencies (ffi't 1), A is unity. The
transmission curve drops to -3 dB in amplitude at the point where ffi't = 1, and
falls at 6 dB/octave (20 dB/decade) at higher frequencies (ffi't 1). The cutoff
frequency is defined as the -3 dB (0.707) point, although there is substantial
transmission for higher frequencies. The -3 dB point is at 1I't rad/s or 1I(2m) Hz.
The ENB of this simple filter is readily found by integrating the square (power
transmission) ofEq. (F.3) in closed form; the result is n/(2't) in rad/s, or I/(4't) =
(rc/2}fo expressed in Hz. The ENB is therefore n/2 or 1.571 times the
-3-dB bandwidth, and everybody in Table F.I agrees!
For higher-order filters, the closed-form integration becomes more complex.
It must be noted that whenever two or more RC sections are cascaded, the -3-dB
point for the new composite filter shifts to lower frequencies; thus, one does not
use the - 3-dB point for a single RC section. In addition, multiple sections are
buffered (isolated from each other) so that subsequent sections present no loading
to preceding sections). The results of the higher-order integrations are shown in
Table F.2.
Table F.2 Low-pass Butterworth filter characteristics, matched sections.
0l.3 dB
BSF
F-3dB
ENB ENB Ratio
Sections (rad/s) (Hz) (rad/s) (Hz) ENBIB_
3dB
1 1I't
1.0 1/2m nl2't 1I4't nl2 = 1.571
2 0.644/'t 0.644 0.102/'t n/4't 1I8't
1.220
3 0.5IlI't
0.511 0.0811't 0.589/'t 0.094/'t
1.155
4 0.436/'t
0.436 0.069/'t 0.492/'t 0.078/'t
1.130
5 0.386/'t
0.386 0.061/'t 0.4301't 0.068/'t
1.115
6 0.350/'t
0.350 0.056/'t 0.387/'t 0.062/'t
1.106
300 Appendix F
These results confirm that the values given by Motchenbacher and Connelly
and by Ott in Table F.1 are correct. Note that the third column represents a
bandwidth shrinkage as sections are cascaded. This is also termed bandwidth
shrinkage factor (BSF) when normalized to 1.0 for a single section.
The formula for the effective (shrunk) bandwidth ratio of cascaded identical
sections is
BSF = ,h(lln) -1 ,
(FA)
where BSF is the net bandwidth ratio due to shrinkage by cascading sections. For
design purposes, the bandwidth of the individual stages must be set to the
reciprocal of the BSF.
Bandpass filters with matched time constants
The simple RC bandpass filter with matched time constants cascades a single
high-pass filter with a single low-pass filter, each with the same time constant
1 = RC. The voltage transmission of a single-section RC high-pass filter is
(F.S)
where (0 is the radian frequency, equal to 21t/ in Hz, and 1 is the circuit time
constant, the RC product, in seconds. At high frequencies, A approaches unity,
and the curve falls at 6 dB/octave (20 dB/decade) at lower frequencies. The
cutoff frequency is defined as the -3-dB (0.707) point, even though there is
substantial transmission at lower frequencies. The -3-dB point, where (0212 = 1,
is 1/1 radls or 1/(21t1) Hz. The peak transmission for these simple bandpass filters
is no longer unity, and the -3-dB-bandwidth points must be evaluated with
respect to the peak transmission of the composite filter, not unity. For the single-
section bandpass filter, a peak transmission of 0.5 (-6 dB) is found, located at the
coincidence of the -3-dB points of the high-pass and low-pass sections. The
-3-dB passband of the bandpass filter is determined at 3 dB below this level, or
at - 9 dB = 0.3535. The closed-form integration of this single-stage bandpass
filter gives an ENB of 1tfc. Results for filters with one to four sections are shown
in Table F.3. Note that the bandwidth shrinkage factor (BSF) matches the same
ratios as for cascaded Butterworth low-pass filters.
It should also be noted that the cut-on and cutoff frequencies are disposed
about the center frequency in a geometric sense, i.e., JLp x /HP= fc2, where fc is
the center frequency of the filter. The location of/LP is given by
(F.6)
where B is either the -3-dB bandwidth or the ENB.JH is thenfc2/fi.
Table F.3 Butterworth bandpass filter characteristics with matched cascaded sections.
Sections 1 2 3 4
Peak transmission 0.5 0.25 0.125 0.0625
B_
3
dB/fC =lIQ 2 1.287 1.019 0.871
BSF 1.0 0.644 0.511 0.436
/LP/fc
0.414 0.546 0.613 0.655
/Hp/fc 2.414 1.833 1.632 1.526
ENB/fc 1t 1t/2 1.177 0.982
fip/fc 0.291 0.486 0.572 0.623
/HP/fC
3.432 2.057 1.749 1.605
ENBIB_
3dB 1t/2
1.220 1.155 l.127
Bandpass Filters with Different Time Constants
Here we discuss bandpass filters constructed using a single high-pass section in
series with a single low-pass section. Each section has a different time constant.
First we note that the minimum bandwidth is achieved with the low-pass cutoff
and high-pass cut-on frequencies matched. If the low-pass cutoff is chosen at a
lower frequency than the high-pass cut-on frequency, the resulting bandwidth
remains the same as the case where the two frequencies are identical. The only
thing achieved is a reduction in filter transmission.
A simple BASIC program was written to find the ENB, the peak
transmission, the -3-dB points, and the -3-dB bandwidth for these filters. Again
the integrations were done for several selections of low-pass cutoffs and high-
pass cut-on frequencies, and the -3-dB points were located with respect to the
composite transmission. The results are shown in Table F.4 for a single-section
bandpass filter. These calculations were done holding/HP constant at unity.
Table F.4 ENB for bandpass filters where hp fHP.
jiP//HP T peak
B_
3
dB ENB
I 0.5 2 3.14
1.5 0.6 2.5 3.91
2 0.667 3 4.70
3 0.75 4 6.26
4 0.8 5 7.82
5 0.833 6 9.39
6 0.857 7 10.96
10 0.909 II 17.20
20 0.952 21 32.78
302 Appendix F
From these data we find some surprisingly simple relationships. The
maximum transmission is given by
T = / IP
peak I' + I' '
J LP J HP
(F.7)
and if/LP /HP, the peak transmission approaches unity as expected. The -3-dB
bandwidth is given by
(F.8)
so that if/LP /HP, the -3-dB bandwidth approaches that of the low-pass filter
alone. The effective noise bandwidth is given by
(F.9)
Figure 7.63 in Horowitz and Hill shows a simple equation to relate the ENB
for this case to the individual cut-on and cutoff frequencies:
(F. IO)
Ifwe were to use Eq. (F.IO) to determine the ENB in the special case where
/LP = /HP, it shows that the ENB is nl4 times the 3-dB bandwidth, which is less
than the -3-dB bandwidth, a surprising result. This analysis shows that the ENB
is (nI2) times the 3-dB bandwidth. It can be seen that whenfL /H, the results of
Eq. (F. I 0) converge to the correct solution, just that of the low-pass filter alone.
A comparison between these new calculations for ENB and the results predicted
using Eq. (F.IO) is given in Table F.5. It is apparent that Eq. (F.lO) gives
erroneous results.
Table F.5 Comparison with Horowitz and Hill.
IdlH
ENB Eg. (F.I0) Ratio
1
n nl4
4
2 4.70 2.09 2.24
5 9.39 6.55 1.43
10 17.20 14.28 1.2
20 32.78 29.9 1.1
Filter Selection to Optimize SNR
We now address the selection of the optimum filter to maximize the signal-to-
noise ratio. For many applications, we can ignore preservation of the signal
waveform to achieve better SNR. We will recover just the fundamental with
maximum SNR. Our discussion will be limited to "white" noise (uniform power
spectral density).
For the low-pass filter, the choice of cutoff frequency fc was determined
using a short BASIC program to iterate fc for a fixed signal frequency of 100
(arbitrary units). The results show that for the single-section filter, the best SNR
is obtained whenfc is equal to the signal fundamental frequency.
The results differ little for multisection low-pass filters. The double-section
filter achieves its peak SNR at a frequency slightly higher (1.11 times the
composite -3-dB point) than for the single-section filter. If we choose the
composite -3-dB point, the SNR is 0.991 times the maximum achievable SNR.
Three- and four-section filters showed similar results: the peak SNR was realized
at about a 15% higher cutoff frequency, but the SNR at the composite -3-dB
point was within 1 % of the maximum SNR. Satisfactory results will be achieved
by using the composite -3-dB frequency for one- to four-section low-pass filters.
For the simple bandpass filter, where the high-pass and low-pass sections are
set to the same frequency, the optimum SNR is achieved by setting both -3-dB
points at the signal frequency, as that will maximize signal transmission. It was
hypothesized that if the bandpass were increased somewhat, the transmission at
the peak may increase faster than the noise, which is proportional to the square
root of the ENB. To test this hypothesis, further spreadsheet work was done to
implement the calculations. I again chose to keep the geometric mean of the low-
pass and high-pass sections equal to the center frequency, i.e.,/LPx/HP =/c
2
. The
optimum SNR is achieved when the passband is defined where JL = /H. It can also
be seen that setting the low-pass section to a lower frequency than the high-pass
section is futile, yielding the same SNR, with both signal and noise attenuated
equally. These above results are valid only for "white" noise.
Every practical circuit has at a minimum a single RC time constant
associated with it, defining an ENB and a corresponding SNR. If additional noise
filtering is necessary, we can add a simple active filter using a single operational
amplifier with two RC sections. It is then interesting to consider the simple
bandpass filter (one section each matched high-pass and low-pass) and one- and
two-section low-pass RC filters. The bandpass filter has an ENB of 1t and a peak
gain of 0.5, whereas a single-section low-pass filter has an ENB of 1t/2 and a gain
of unity. The noise transmission, the product of the peak transmission and
ENB
I /2
, is 1.253 for the low-pass filter and 0.886 for the bandpass filter. The
signal transmission for the low-pass filter is 0.707, whereas it is 0.5 for the
bandpass filter. The SNR for each is therefore the same. However, for a two-
section low-pass filter with the signal at the (composite) -3-dB frequency, the
ENB is 1.220 x /C, the noise transmission is 1.105, the signal transmission is
304 Appendix F
0.707, and the SNR is therefore improved by some 13.5% over the simple
bandbass and single-section filters.
Conclusions
Several recommendations can be made:
(1) A double-section (second-order) low-pass filter is sufficiently better in
the reduction of white noise to warrant the use of two additional
components.
(2) In the absence of lifnoise, a two-section low-pass filter will outperform
a bandpass filter with the same number of components.
(3) When cascading (identical) sections of Butterworth-type filters, the BSF
must be taken into account.
(4) When noise content of the (linear) signal channel is other than white
noise, additional noise suppression techniques may be needed. Transient
suppression and lifnoise effects are not covered in this discussion. In the
case of system nonlinearities (such as pulsed signals or in certain laser
applications), judicious addition of certain types of noise can actually
improve the SNR. This is called stochastic resonance, and it is beyond
the scope of this book.
(5) The above discussion relates only to the linear-circuit analog domain.
Once the filtered signal becomes digitized, quantization noise and
antialiasing effects must be taken into account. A good rule of thumb to
follow is that use of a first-order filter requires a sample rate of at least
five times the cutoff frequency of the filter. Only then will the filter fully
meet the Nyquist criteria (to minimize aliasing in the digital output of the
AID converter).
Appendix G
Bandwidth Normalization by
Moments
General measurement equation:
If9\(A) can be represented as a rectangle,
I = 9\/1 f.' <1>1. dA .
I
Let the source function <l>A be described as a second-degree polynomial:
<1>1.. = A + BA + CA?
(G.1)
(G.2)
(G.3)
Substitute Eq. (G.3) into Eq. (G.1), divide both sides by ffl(A)dA, and multipl y
both sides by (A2 - AI) to get:
(GA)
Next, integrate Eq. (G.3) between the limits Al and A2:
Note the similarities between Eqs. (GA) and (G.S). If the following conditions
are applied,
305
306 Appendix G
A2 +A[ _ [A9t(A)dA A; +A2A[ +A; _ [A
2
9t(AdA)
2 - [9t(A)dA ' 3 - [9t(AdA)
(G.6)
and
l.' A
dA
= I(A2 -A[) .
, [9t(A)dA
(G.7)
Assume that area of response curve = area of equivalent rectangle, i.e.
[9t(A)dA. (G.8)
Then,
r ' 
1, A 9t '
n
(G.9)
and we have a band-limited power <l>in-band.
Now we proceed to determine AI, A2, and 9t
n
Substitute:
[A9t(A)dA
M[ = -="------
[9t(A)dA
[A
2
9t(A)dA
M2=-=-----
[9t(A)dA
(G.lO)
Then
(G.lI)
M[ is the first moment divided by the area (Oth moment) and is the centroid of the
response curve, the effective or center wavelength Ac. M2 is the second moment
divided by the area, which is related to the square of the radius of gyration.
Solution of simultaneous Eqs. (G. I I ) with the substitution M = Ac yields
(G.12)
showing the bandpass limits Al and A2 are symmetrically disposed about the
center wavelength Ac.
The quantity (M2 - A/ ) is recognized as the variance d. The bandwidth
between wavelength limits Al and A2 is:
= A2 -AI = 2J3cr ,
and the short- and long-limit wavelengths are then
(G.13)
Bandwidth Normalization by Moments 307
(G.14)
The bandwidth-normalized responsivity is
9\n= [9\(A)dA.
2V'3cr
(G.15)
Now we have our three parameters, 9\n, AI, and A2. Note that the coefficients
A, B, and C of the second-degree source polynomial [Eq. (G.3)] have vanished.
The implication is significant:
Any source that can be represented by a second-degree polynomial can
be characterized between the wavelength limits Al and A2 (which are
determined solely by the radiometer) without error.
There is no ambiguity in any of the normalization parameters; they are
all uniquely determined from only the spectral responsivity curve.
The errors are related to the deviation of the source function from a
quadratic.
Moments Normalization Summary
This is the step-by-step procedure for accomplishing a moments normalization.
The starting point is absolute spectral responsivity 9\(A).
Zero'th moment
First moment
Second moment
Center wavelength (centroid)
Variance
Short wavelength limit
Long wavelength limit
Bandwidth
Normalized responsivity
Mo = [9\(A)dA
MI = [A9\(A)dA
M
z
= [Az9\(A)dA
A =M
I
c
M
o
cr
Z
= M
z
_A
z
M c
o
AI =Ac -J}cr
A2 =Ac +J}cr
f1A = 2J}cr
9\ = Mo
n 2J}cr
-
Appendix H
Jones Near-Small-Source
Calibration Configuration
The following method is adapted from an obscure application note further
describing the near-small-source method of radiometric calibration.
How to calibrate a radiometer
A simple method of calibrating a radiometer using a small blackbody source
close to the radiometer aperture has been described by Dr. R. Clark Jones of the
Polaroid Corporation. The principles involved in this method are briefly
reviewed here with special emphasis on the application to the optitherm-
radiometer Cassegrain system.
The essential point in the method is that a small radiation source close to the
aperture of the radiometer will uniformly irradiate an area in the focal plane of
the radiometer. The radiation on the detector (with the radiometer focused at
infinity) is then given by
where = radiation on detector (watts, <1
L = source radiance (W/m
2
sr, Ls)
Ad = detector area (m
2
)
(Or = solid angle of radiometer field of view (sr) at the focal plane
= source area I (focallengthf
(H.I)
Thus, the responsivity of a detector in a radiometer can readily be calibrated
by dividing the signal voltage output of the instrument by the radiation on the
detector.
An important aspect of the method is that the radiation on the detector is
independent of the radiometer aperture and source location, providing that a
uniformly irradiated area covers the detector. This restriction places limits on the
source size and location as given below.
309
310 Appendix H
Focal plane
I
Op; (Entrance pupil of radiometer)
t
0,:
Image, I tl
Point source, S
t
p
)
(
Q
Figure H.1 Point source near a radiometer. The detector surface is located at the focal
plane.
Consider a point source near a radiometer (Fig. H.I). For a point source, the
diameter of the uniformly irradiated disc D
j
in the focal plane is given by Eq.
(H.2) (after using the thin-lens imaging formula):
(H.2a)
where D
j
, f, P, and Dpi are defined in Fig. H. I. The size of the detector must be
less than D
f
It is easy to prove Eq. (H.I) for the configuration shown in Fig. H. !. For a
small Lambertian source, flux at the entrance pupil of the radiometer is given by
(H.3)
For a detector of size D
j
at the focal plane of the radiometer, the flux detected is
given by
<l>D = L AD 0) .
f S f r
(H.4)
But we know that D . = D So, from Eqs. (H.2), (H.3), and (H.4), we get
P' f
(H.5)
Asssuming uniform irradiance at the focal plane, Eq. (H.4) can be written as Eq.
(H.I).
-
Jones Near-Small-Source Calibration Configuration 311
Similarly, if the radiometer optical system is of the Cassegrain type, an
obscured area will result, and the uniformly irradiated area will then be an
annular ring of width X and mean diameter Dm:
(H.6a)
where DP2 = the diameter of an obscured disc produced by secondary mirror.
(H.7)
If, as is always the case, the source has a finite diameter Us (see Fig. H.2),
the edges of the disc or ring will be vignetted by an amount if Us/P), divided
equally on either side of the unblurred edge. The remaining uniformly irradiated
disc or annular ring width is given by
(H.2b)
(H.6b)
The mean diameter of the ring remains the same as given in Eq. (H.7).
Image
I
I
I
~
Q
D x = uniformly irradiated disc
Figure H.2 Thin-lens-equivalent layout for a Cassegranian optical system.
312 Appendix H
In calibrating a radiometer, a convenient source aperture and distance are
first chosen. Then the size of the uniformly irradiated area must be checked to
see whether it covers the detector. For example, if a Cassegrain system is used,
the detector must fit in the ring of width z and mean diameter D as computed
from Eqs. (H.6b) and (H.7). The source must be placed off axis such that the
annular ring of width x falls on the detector.
Several other precautions must be observed, particularly if a thermal detector
is used. Since the calibration source fills only a small part of the field of view, the
detector will "see" other radiation as well. If the source is chopped, the
radiometer will respond only to it. However, in many cases the detector is
chopped, and hence it responds to all radiation in its field of view. One way of
separating the response to the calibration source from the background is to record
the difference in response when the source aperture is opened and when it is
closed.
-
Appendix I
Is Sunglint Observable in the
Thermal Infrared?
The short answer is yes. Orbiting spacecraft like GOES, AVHRR, ASTER,
MODIS, and ATSR all have spectral observation windows in the thermal
infrared. They all have to deal with sunglint at least some of the time. It is not
just a minor nuisance, but a major effect that saturates their sensors. A saturated
sensor gives no indication of the actual magnitude; all information from that
direction is lost. The directional characteristics of sunglint have been adequately
treated by Maurice Cox. I This report deals with the radiometric aspects, the
magnitude of the sunglint in comparison with the magnitude of normal target
thermal emission.
Background
Every object at a temperature above absolute zero emits electromagnetic
radiation (EMR) as a result of molecular motion. This emission as a function of
wavelength is described by the Planck equation
(
2hC
2
) ( 1 )
L,. = (1..) Y ehclAIcT -1 '
(1.1)
where LA is spectral radiance in W/m
2
sr, (1.) is spectral emissivity, a wavelength-
dependent quality factor, h is Planck's constant, c is the velocity of light, A is
wavelength, k is Boltzmann's constant, and T is the absolute temperature. This
equation gives the spectral radiance in W/m
2
srJ.lm at any wavelength for an
object at temperature T. If (1..) is 1 at all wavelengths, the emission is called
blackbody radiation. For temperatures on the order of room temperature ~ 3 0 0 K,
80 F) this radiation peaks near 10 J.lm, and the region surrounding 10 J.lm is
known as the thermal infrared (TIR).
When a beam of EMR encounters an object, three things can happen. The
object can:
1. absorb the EMR,
2. transmit the EMR, or
3. reflect the EMR.
313
314 Appendix'
In fact, all three processes happen simultaneously: an object reflects some of
the incoming EMR, absorbs some of it, and transmits the balance. All of these
three are relative quantities, the fraction reflected, absorbed, or transmitted, and
have values between 0 and 1. Since EMR must be conserved, the sum of these
three is 1:
A+T+R=1. (1.2)
This means that if an object is highly reflective (R-71, like a mirror), the sum
of the absorption and transmission must necessarily be low. Similarly, if the
object is transparent (T -71, like a window), it must have both low absorption and
low reflection. Kirchhoff related the absorption and emission of an object, stating
that if an object is a good absorber, it must also be a good emitter. With
Kirchhoff's law in mind, objects that are highly reflective or highly transmissive
cannot be good emitters.
The TIR is an interesting place to observe things. An infrared camera or
forward-looking infrared (FUR) "sees" a picture of the distribution of radiance
in the observed scene, which usually includes an object of interest, a target. The
observed radiance is comprised of three factors, the direct emission from the
target, the radiance of whatever is behind the target multiplied by the
transmittance of the target, and the radiance of whatever surrounds the target
multiplied by the reflectance of the target. These cameras do not observe at all
wavelengths, but are limited by the transmission of their optics and that of our
atmosphere. Two "windows" are commonly used, between 3 and 5 fim and
between 8 and 12 fim. In general, the 3- to 5-fim window is better suited for
hotter targets and the 8- to 12-fim window is best for objects near room
temperature. To determine the radiance within the wavelength region defined by
an atmospheric window, we integrate Eq. (1.1) over the TIR wavelength range.
The integration can be done with standard mathematical computer programs (i.e.,
MathCad, Mathematica, etc.), a custom computer program, or even brute force
using a spreadsheet.
Another issue with which we must deal is the atmosphere itself. It has a
transmission less than one, and it also radiates in the thermal infrared according
to Eq. (1.1). The problem is extraordinarily complex, and massive computer
programs have been devised to determine the characteristics of our atmosphere.
For this work, the program PCTRAN@ (www.ontar.com). a commercial version
of the USAF program LOWTRAN7, has been used to determine the relevant
parameters. To keep the analysis quite general, the preprogrammed atmosphere
known as MID-LATITUDE SUMMER was chosen as characteristic of a broad
range of characteristics in the CONUS over at least half of the year. The altitude
was taken at sea level, and the solar zenith angle (as measured from directly
overhead) was 25 deg. At this angle, the calculated transmission, averaged over
the TIR (8- to 12-fim band), is 0.57, and the in-band radiance of the sky in this
direction is 12.5 W/m
2
sr. The radiance was also determined over a range of
..
Is Sunglint Observable in the Thermal Infrared? 315
angles and integrated to determine the total in-band irradiance on the ground
from the entire sky; it is 51.2 W/m
2

Direct Emission
Most objects seen by FUR devices are at ambient temperature, here taken as 300
K (80
0
F). A 300-K blackbody has an in-band (8- to 12-llm) radiance of
38.5 W/m
2
sr. For real objects, we multiply by the emissivity; for this analysis the
emissivity is assumed to be constant over the band. We consider four materials:
earth, water, aluminum, and glass. Table 1.1 gives the radiative characteristics.
The difference between the aluminum and the other materials is that aluminum is
a good reflector and therefore a poor emitter. This is a characteristic of metals,
which are good electrical conductors.
Reflected Radiation
If these materials have an emissivity less than unity, the sum of their reflectance
and transmittance must make up the difference. Since all of these materials are
opaque (T = 0) in the TIR, the reflectance is (1 - emissivity). The directional
properties of reflectance are highly varied, from perfectly diffuse (flat, matte) to
perfectly specular (like a perfect mirror). We shall inspect the two limiting cases.
Since the materials are reflecting something we must determine what that
something is. For an object out-of-doors, lying horizontally on the ground, the
background consists of three parts: the sun, the sky, and nearby objects that are
above grade. Because of the variability of above-grade objects, we shall ignore
them, assuming our object of interest is out in the open. The radiance of the sky,
obtained via PCTRAN, was previously stated to be 12.5 W/m
2
sr at 25-deg
zenith angle. The sun is assumed to be a blackbody at 6000 K and subtends a
solid angle of 6.8 x 10-
5
sr. The in-band radiance from a 6000 K blackbody is
19840 W Im
2
sr, so the irradiance outside the atmosphere is the product of the
radiance and the solid angle and is 1.35 W/m
2
After transmission through the
atmosphere, the irradiance is reduced to 0.77 W/m2, a rather feeble amount.
Table 1.1 Direct TIR emission from selected materials (300 K).
Material
Emissivity Radiance
(8- to 12-J.lm band) W/m
2
sr
Earth 0.95 36.6
Water 0.985 37.9
Aluminum 0.1 3.85
Glass 0/96 37.0
316 Appendix I
Table 1.2 Reflected radiance from diffuse materials.
Material
Earth
Water
Aluminum
Glass
Diffuse
reflectance
0.05
0.015
0.9
0.04
Diffuse reflectance
Reflected sun
(W/m
2
sr)
0.012
0.004
0.22
0.010
Reflected sky
(W/m
2
sr)
0.81
0.24
14.67
0.65
For diffuse objects, we invoke the Lambertian approximation, which states that
radiance is independent of direction and is the product of the irradiance and the
reflectance, divided by 1t, or
(1.3)
This equation may be used directly for a source like the sun. For a uniform
hemispherical diffuse source like the sky, the irradiance E is 1t times the source
radiance, so the reflected radiance is simply the source radiance multiplied by the
surface reflectance. The sky is not really uniform, but for a diffuse reflector, we
can use the total irradiance from the sky, multiply by the reflectance, and divide
by 1t. Table 1.2 shows the reflected radiance of each of these four materials from
the direct sunlight and from the sky.
Sunlight reflected from a diffuse surface in the 8- to l2-I..Im band is very
small. The total radiance from all of these objects is the sum of the direct and
reflected components, as shown in Table 1.3. Note that with the exception of the
aluminum, they are comparable, and will look much the same on a thermal
image. The aluminum, being reflective rather than emissive, has a significantly
lower radiance, and will appear dark on a typical FUR display ("white-hot,"
where a more radiant object appears white and a less radiant object appears
black). Since we do not know a priori the nature of the objects in the scene, dark
objects will be interpreted as cooler than white objects.
Reflection from smooth surfaces is called specular, and straightforward laws of
geometry apply. The incident and reflected beams and the normal to the
reflecting surface all lie in the same plane, and the angles of both the incident and
reflected beams with respect to the normal to the reflecting surface are equal. For
an outdoor scene viewed from above, this means that at one specific angle, which
must follow the above rules, the sun will be seen. At all other angles, the sky will
be seen.
Table 1.3 Total radiance from diffuse materials.
Material
Direct radiance Reflected sun Reflected sky Total radiance
(W/m
2
sr) (W/m
2
sr) (W/m
2
sr) (W/m
2
sr)
Earth 36.6 0.012 0.81 37.42
Water 37.9 0.004 0.24 38.14
Aluminum 3.85 0.22 14.67 18.74
Glass 37.0 0.010 0.65 37.66
Window method
One way to look at the problem is to treat the reflecting surface as a window
through which the sun or the sky can be viewed. This window has the same size
and orientation as the reflecting surface, and its transmission has the same value
as the reflection of the reflecting surface. In this model, the effective radiance is
the product of the surface reflectance and the radiance of whatever is seen in
reflection. Table 1.4 gives the effective radiance of the three materials that are
capable of specular reflectance for both sunlight and skylight. Recall that the
radiance of the sky in the 8- to 12-)lm band (zenith angle 20 deg) is 12.5 W/m
2
sr,
and the transmittance of the atmosphere is 0.57. The in-band radiance of the sun
(before atmospheric losses) is 19840 W/m
2
sr, so the effective radiance, including
the atmospheric transmission, is 11310 W 1m2 sr.
The radiance of the object from a uniform diffuse sky is seen to be the same
whether the reflector is diffuse or specular. When compared with the typical total
radiances for diffuse surfaces, it can be seen that the specularly reflected sun
(sunglint) is a factor of 5 to nearly 300 times the total from a diffuse surface
~ 3 7 . 5 W/m
2
sr).
BRDF method
Another way of looking at reflection is through the concept of bidirectional
reflectance distribution function (BRDF). This term is the ratio of the radiance of
a surface, as a function of direction, to the irradiance incident upon the surface. It
carries the units lIsr. The BRDF is the surface reflectance divided by the
projected solid angle of the irradiating source. The radiance of a surface is then
the product of the irradiance E and the BRDF, or L = E x BRDF. For a diffuse
Table 1.4 Reflected radiance from specular materials.
Material
Specular Reflected sun Reflected sky
reflectance (W/m
2
sr) (W/m
2
sr)
Water 0.015 170 0.19
Aluminum 0.9 10180 11.25
Glass 0.04 452 0.50
318 Appendix I
Table 1.5 Reflected radiance from specular materials via BRDF.
Material
Specular BRDF Irradiance from Reflected sun
reflectance from sun sun (W/m2) (W/m
2
sr)
Water 0.015 221 0.77 170
Aluminum 0.9 13235 0.77 10190
Glass 0.04 588 0.77 453
surface, the BRDF is pi n and the radiance L is then pEl n. For a hemispherical
diffuse source like the sky, the irradiance E is the source radiance L
sky
multiplied
by the projected solid angle of the source, which is n. Then the object radiance is
just the source radiance L
sky
times the reflectance p. For a specular surface, the
BRDF is the surface reflectance divided by the projected solid angle of the
source, or pin. The sun has a projected solid angle of 6.8 x 10-
5
sr, so the BRDF
becomes 14706 p. Using these values, Table I.5 gives the reflected radiances.
The feeble 0.77 W/m2 irradiance from the sun turns into an overwhelming
radiance upon specular reflection. The results of the BRDF computation of the
glint radiance are identical to the results obtained using the "window" method.
Several factors can alter these results. First, the surface must be flat and
clean. The presence of a scattering overlay, like a layer of dust, will remove a
portion of the specularity and add a diffuse component. Second, the polish on the
surface must be fairly good, though at these longer wavelengths, the surface
roughness can be over ten times as great as that of a mirror designed for
reflection in the visible.
The effect of this large radiance on a FUR depends upon several factors.
First, our sensor must be positioned such that it is (1) located within the specular
beam, and (2) looking in the proper direction. The instantaneous field of view
(lFOV) of the sensor is the angle that a single detector subtends when observing
a target and is measured in milliradians (rnrad). Detector considerations (number
and diffraction-limited size) and overall instrument constraints (aperture size,
focal length, and these combined as j7#) place typical values of IFOV in the
neighborhood of 0.1 to 0.2 mrad. If the detector is square, the footprint, or target
area seen by the detector, is also nearly square, depending upon the viewing
angle. The observed area of the target is then (lFOV x R)2 where R is the range,
the distance from the target to the sensor. For example if the range is I km (3280
ft), the area seen by a detector with an IFOV of 0.1 rnrad is 0.01 m
2
, and the
detector footprint is a square 0.1 m (about 4 in.) on a side. Larger IFOV and/or
larger range will cause an increase in the size of the detector footprint. If we are
looking at an angle from nadir (which is vertical, straight down) then we divide
the footprint by the cosine of the angle from nadir. The increase is small for low
values of nadir angle; a 10% increase in effective area is seen if the nadir angle is
25 deg.
The power on a detector from glint depends upon the size of the reflecting
object compared with the detector footprint. The diameter of the sun subtends
nearly 9 mrad, sufficiently large to overfill any properly oriented detector. The
total radiance as seen by a detector is the sum of the earth radiance and the glint
radiance, each weighted by its fraction of the subtended area. For example, a
single square piece of glass (p = 0.04) that is 1 cm on a side has an area of
1 ~ m
2
If the detector footprint is 10-
2
m
2
, the fraction subtended by the glass is
0.01 (1%). The remaining earth has a fractional area of 0.99 (99%). Taking the
radiances from above and weighting them according to these fractional areas, we
get
0.01 x 450 + 0.99 x 36.6 = 40.7 W/m
2
sr.
Specular reflection of sunlight from a small piece of glass results in an 11 %
increase in radiance over the bare earth at 300 K (26.8 C, 80 F). This increase
in radiance would be interpreted as a temperature difference of 12 F, a rise from
80 F to 92 F. An IFOV that includes the same size piece of shiny aluminum
(square, 1 cm on a side) will have a total radiance of
0.01 x 10200 + 0.99 x 36.6 = 138.2 W/m
2
sr.
This radiance is likely to saturate a detector. If a detector does have sufficient
dynamic range to respond linearly to this radiance, it will indicate a temperature
of 409 K (136 C, 276 F), which is totally unrealistic.
Conclusion
We have demonstrated here that sunglint can be a significant factor in imaging
systems in the thermal infrared. It will be observed if the sensor is looking at a
target with a sung lint component, and the geometrical angles are such that the
sun can be seen in reflection. The only way to avoid it is to look in some other
direction.
Works Consulted
1. M. Cox, Sun Reflection Geometry, (1999). Unpublished.
2. A. W. Cooper, E. C. Crittenden, E. A. Milne, P. L. Walker, E. Moss, and D.
J. Gregoris, "Mid- and far-infrared measurements of sunglint from the sea
surface," in Proc. SPIE 1749, 176-185 (1992). Thefirstfive authors are/rom
the Naval Postgraduate School, Monterey, CA. Measurements were made
with an AGA Thermovision 780 dual-band camera operating in the 2- to
5.6-fim and 8- to 12-fim bands.
-
Appendix J
Documentary Standards for
Radiometry and Photometry
Numerous agencies prepare and disseminate documentary (protocol) standards
for radiometry and allied fields. Those listed here are either U.S. or international
standards organizations. The European Committee for Standardization is called
CEN,i whose mission is to promote voluntary technical harmonization in Europe
in conjunction with worldwide agencies and its European partners. The purposes
include the lowering of trade barriers and the promotion of common technical
understanding. European standards (e.g., BSI from UK and DIN from Germany)
may be accessed via CEN. The Japanese Standards Association
ii
also has
extensive listings.
ANSI
American National Standards Institute. The U.S. representative for ISO, the
International Standards Organization. iii
C78.40-1985
C78.180-1989
C78.375-1991
C78.386
C78.387
C78.388
C82.1-1985
C82.I(a-c) 1990
C82.3-1989
C82.4-1985
i www.cenorm.be/
ii . .
www.Jsa.or.Jp
Specifications for mercury lamps
Specifications for fluorescent lamp starters
Guide for electrical measurements of fluorescent lamps
Mercury lamps-measurement of characteristics
Metal-halide lamps-measurement of characteristics
High-pressure sodium lamps-measurement of
characteristics
Specifications for fluorescent lamp ballasts
Specifications for fluorescent lamp ballasts (supplement to
C82.1)
Specifications for fluorescent lamp reference ballasts
Specifications for intensity discharge and low-pressure
sodium amp ballasts
iii (www.ansi.org) ANSI/IES RP-16, American National Standard Nomenclature and Definitions for
Illuminating Engineering (1986).
321
322
C82.5-1990
C82.6-1985
ANSIINCSL
540-1-1994
ASTM
Appendix J
Specifications for high-intensity-discharge lamp reference
ballasts
Methods of measurement of discharge lamp ballasts
American National Standards for Calibration--calibration
Laboratories and Measuring and Test Equipment-general
requirements
The American Society for Testing and Materials
iv
maintains an extensive
collection of protocol or documentary standards, including practices,
specifications, guides, procedures, and test methods for a large range of materials
and instrumentation for their characterization. Some 10,000 standards are
published annually in a 73-volume set, of which one volume is the index alone!
The ASTM Book of Standards occupies nearly two meters of shelf space. These
documents are subject to periodic review and revision by the committees that
were responsible for their generation and maintenance. Always check at ASTM's
website to see if a later version is available. A number of these standards are the
result of the efforts of committee E12 on Color and Appearance, committee E20
on Temperature Measurement and committee E37 on Thermal Measurement.
Special collections in specific areas are also published, for example, ASTM
Standards on Color and Appearance Measurement, 8th Ed. (2008). This book
contains 130 ASTM standards as well as ISO and ISO/CIE standards used in
appearance analysis for a variety of materials and products, including all the
standards listed above. It includes a CD/ROM with even more information. The
following listing shows a selection of relevant standards to radiometry,
photometry, and colorimetry. Another useful guide is "Nomenclature and
Definitions Applicable to Radiometric and Photometric Characteristics of
Matter," ASTM Special Technical Publication 475 (1971).
"Guide for recommended uses of photoluminescent safety markings," ASTM
E2030 (1999).
"Standard practice for preparation of pressed powder white reflectance factor
transfer standards for hemispherical geometry and bi-directional geometries,"
ASTM E259 (1998).
"Recommended practice for goniophotometry of objects and materials," ASTM
El67 (1996).
"Specification for daytime pedestrian visibility enhancement," ASTM E1896
(1997).
"Standard guide for describing and specifying the spectrometer of an optical
emission direct-reading instrument," ASTM E1507 (1998) .
iv www.astm.org
Documentary Standards for Radiometry and Photometry 323
"Standard guide for designing and conducting visual experiments," ASTM
E1808 (1996).
"Standard guide for establishing spectrophotometer performance tests," ASTM
E1866 (1997).
"Standard guide to evaluation of optical properties of powder coatings," ASTM
D5382 (1995).
"Standard guide for examining electrical and mechanical equipment with infrared
thermography," ASTM E1934 (1999).
"Standard guide for modeling the colorimetric properties of a visual display
unit," ASTM E1682 (2001).
"Standard guide for preparation, maintenance, and distribution of physical
product standards for color and geometric appearance of coatings," ASTM
D5531 (1999).
"Standard guide for quality assurance of laboratories using molecular
spectroscopy," ASTM E924 (1994).
"Standard guide for quantitative analysis by energy-dispersive spectroscopy,"
ASTM E1508 (1998).
"Standard guide for Raman shift," ASTM E1840 (1996).
"Standard guide for selection of geometric conditions for measurement of
reflection and transmission properties of materials," ASTM E179 (1996).
"Standard guide for use of lighting in laboratory testing," ASTM E1733 (1995).
"Standard guide to evaluation of optical properties of powder coatings," ASTM
D5382 (1995).
"Standard guide to properties of high visibility materials used to improve
individual safety," ASTM F923 (2000).
"Standard method for calibration of reference pyranometers with axis vertical by
the shading method," ASTM E913 (1999).
"Standard practice for angle resolved optical scatter measurements on specular or
diffuse surfaces," ASTM El392 (1996).
"Standard practice for calculating solar reflectance index of horizontal and low-
sloped opaque surfaces," ASTM E 1980 (2001).
"Standard practice for calculating yellowness and whiteness indices from
instrumentally measured color coordinates," ASTM E313 (2000).
"Standard practice for calculation of photometric transmittance and reflectance of
materials to solar radiation," ASTM E971 (1996).
"Standard practice for calculation of weighting factors for tristimulus
integration," ASTM E2022 (2001).
"Standard practice for calibrating thin heat flux transducers," ASTM el130
(2001).
324 AppendixJ
"Standard practice for calibration of ozone monitors and certification of ozone
transfer standards using ultraviolet photometry," ASTM D5110 (1998).
"Standard practice for calibration of the heat flow meter apparatus," ASTM
CI132 (1995).
"Standard practice for calibration of transmission densitometers," ASTM E1079
(2000).
"Standard practice for calculating yellowness and whiteness indices from
instrumentally measured color coordinates," ASTM E313 (2000).
"Standard practice for color measurement of fluorescent specimens," ASTM
E991 (1998).
"Standard practice for computing the colors of fluorescent objects from bispectral
photometric data," ASTM E2152 (2001).
"Standard practice for computing the colors of objects by using the cie system,"
ASTM E308 (1999).
"Standard practice for describing and measuring performance of dispersive
infrared spectrometers," ASTM E932 (1997).
"Standard practice for describing and measuring performance of fourier
transform mid-infrared (FT-MIR) spectrometers level zero and level one
tests," ASTM E1421 (1999).
"Standard practice for describing and measuring performance of laboratory
fourier transform near-infrared (FT -NIR) spectrometers: level zero and level
one tests," ASTM E1944 (1998).
"Standard practice for describing and measuring performance of ultraviolet,
visible, and near-infrared spectrophotometers," ASTM E275 (2001).
"Standard practice for describing and specifying inductively-coupled plasma
atomic emission spectrometers," ASTM E1479 (1999).
"Standard practice for describing photomultiplier detectors in emission and
absorption spectrometry," ASTM E520 (1998).
"Standard practice for describing retroreflection," ASTM E808 (1999).
"Standard practice for determining the steady state thermal transmittance of
fenestration systems," ASTM E1423 (1999).
"Standard practice for electronic interchange of color and appearance data,"
ASTM E1708 (2001).
"Standard practice for establishing color and gloss tolerances," ASTM D3134
(1997).
"Standard practice for evaluating solar absorptive materials for thermal
applications," ASTM E744 (1996).
"Standard practices for general techniques of ultraviolet-visible quantitative
analysis," ASTM E169 (1999).
"Standard practice for goniophotometry of objects and materials," ASTM E167
(1996).
"Standard practice for identification of instrumental methods of color or color-
difference measurement of materials," ASTM E805 (2001).
"Standard practice for measuring colorimetric characteristics of retroreflectors
under nighttime conditions," ASTM E811 (2001).
"Standard practice for measuring photometric characteristics of retroreflectors,"
ASTM E809 (2000).
"Standard practice for measuring practical spectral bandwidth of ultraviolet-
visible spectrophotometers," ASTM E958 (1999).
"Standard practice for near infrared qualitative analysis," ASTM E1790 (2000).
"Standard practice for obtaining bispectral photometric data for evaluation of
fluorescent color," ASTM E2153 (2001).
"Standard practice for obtaining colorimetric data from a visual display unit
using tristimulus colorimeters," ASTM E1455 (1997).
"Standard practice for obtaining spectrophotometric data for object-color
evaluation," ASTM E1164 (1994).
"Standard practice for obtaining spectroradiometric data from radiant sources for
colorimetry," ASTM E1341 (2001).
"Standard practice for preparation of pressed powder white reflectance factor
transfer standards for hemispherical geometry and bi-directional geometries,"
ASTM E259 (1998).
"Standard practice for preparation of textiles prior to ultraviolet (uv) transmission
testing," ASTM D6544 (2000).
"Standard practice for qualifying spectrometers and spectrophotometers for use
in multivariate analyses, calibrated using surrogate mixtures," ASTM E2056
(2000).
"Standard practice for reducing the effect of variability of color measurement by
use of multiple measurements," ASTM E 1345 (1998).
"Standard practice for selecting and calibrating sources for the visual assessment
of object colors," ASTM Z6606Z.
"Standard practice for selection and use of portable retroreflectomers for the
measurement of pavement marking materials," ASTM E 1743 (1996).
"Standard practice for solar simulation for thermal balance testing of spacecraft,"
ASTM E491 (1999).
"Standard practice for specifying and matching color using the colorcurve
system," ASTM E154l (1998).
"Standard practice for specifying and verifying the performance of colorimeters,
spectrocolorimeters and goniospectrocolorimeters," ASTM Z6899Z.
"Standard practice for specifying color by the munsell system," ASTM D1535
(2001).
"Standard practice for specifying color by using the optical society of america
uniform color scales system," ASTM E1360 (2000).
326 AppendixJ
"Standard practice for specifying the geometry of observations and
measurements to characterize the appearance of materials," ASTM E1767
(1995).
"Standard practice for the periodic calibration of narrow band-pass
spectrophotometers," ASTM E925 (1994).
"Standard practice for testing fixed-wavelength photometric detectors used in
liquid chromatography," ASTM E685 (2000).
"Standard practice for testing variable-wavelength photometric detectors used in
liquid chromatography," ASTM E1657 (2001).
"Standard practice for transfer standards for reflectance factor for near-infrared
instruments using hemispherical geometry," ASTM El791 (2000).
"Standard practice for validation of multivariate process infrared
spectrophotometers," ASTM D6122 (1999).
"Standard practice for visual appraisal of colors and color differences of
diffusely-illuminated opaque materials," ASTM D 1729 (1996).
"Standard practice for visual color evaluation of transparent sheet materials,"
ASTM E1478 (1997).
"Standard practice for visual evaluation of metamerism," ASTM D4086 (1997).
"Standard practices for general techniques of ultraviolet-visible quantitative
analysis," ASTM El69 (1999).
"Standard practices for infrared multivariate quantitative analysis," ASTM
E1655 (2000).
"Standard Practices for Internal Reflection Spectroscopy," ASTM E573 (2001)
"Standard Solar Constant and Zero Air Mass Solar Spectral Irradiance Tables,"
ASTM E490 (2000).
"Standard specification for infrared thermometers for intermittent determination
of patient temperature," ASTM El965 (1998).
"Standard specification for nighttime photometric performance of retroreflective
pedestrian markings for visibility enhancement," ASTM El50l (1999).
"Standard specification for photoluminescent (phosphorescent) safety markings,"
ASTM E2072 (2000).
"Standard specification for physical characteristics of nonconcentrator terrestrial
photovoltaic reference cells," ASTM EI040 (1998).
"Standard specification for retroreflective sheeting for traffic control," ASTM
D4956 (2001).
"Standard specification for silvered flat glass mirror," ASTM C1503 (2001).
"Standard specification for solar simulation for terrestrial photovoltaic testing."
ASTM E927 (1997).
"Standard tables for references solar spectral irradiance at air mass 1.5: direct
normal and hemispherical for a 37 tilted surface," ASTM G 159 (1998).
"Standard terminology of appearance," ASTM E284 (2001).
"Standard terminology relating to molecular spectroscopy," ASTM E 131 (2000).
"Standard terminology relating to photovoltaic solar energy conversion," ASTM
E 1328 (1999).
"Standard terminology relating to solar energy conversion," ASTM E772 (1993).
Standard test method for 20-deg specular gloss of waxed paper," ASTM D1834
(2000).
Standard test method for 45-deg specular gloss of ceramic materials," ASTM
C346 (1998).
Standard test method for 60-deg specular gloss of emulsion floor polish," ASTM
01455 (1997).
"Standard test method for calibration of a spectroradiometer using a standard
source of irradiance," ASTM G 138 (1996).
"Standard test method for calibration of heat transfer rate calorimeters using a
narrow-angle blackbody radiation facility," ASTM E638 (1992).
"Standard test method for calibration of narrow- and broad-band ultraviolet
radiometers using a spectroradiometer," ASTM G 130 (1995).
"Standard test method for calibration of a pyranometer using a pyrheliometer,"
ASTM G 151 (2000).
"Standard test method for color and color-difference measurement by tristimulus
(filter) colorimetry," ASTM E1347 (1997) (formerly E97).
"Standard test method for determining the linearity of a photovoltaic device
parameter with respect to a test parameter," ASTM E1143 (1994).
"Standard test method for estimating stray radiant power ratio of
spectrophotometers by the opaque filter method," ASTM E387 (1995).
"Standard test method for field measurement of raised retroreflective pavement
markers using a portable retroreflectometer," ASTM E1696 (2001).
"Standard test method for haze and luminous transmittance of transparent
plastics," ASTM D 1 003 (2000).
"Standard test method for identifying fluorescence in object-color specimens by
spectrophotometry," ASTM El247 (2000).
"Standard test method for luminous reflectance factor of acoustical materials by
use of integrating-sphere reflectometers," ASTM E1477 (1998).
"Standard test method for measuring and calculating emittance of architectural
flat glass products using spectrometric measurements," ASTM E1585 (1993).
"Standard test method for minimum detectable temperature difference for
thermal imaging systems," ASTM E1311 (1999).
"Standard test method for noise equivalent temperature difference of thermal
imaging systems," ASTM E1543 (2000).
328 Appendix J
"Standard test method for obtaining colorimetric data from a visual display unit
by spectroradiometry," ASTM El336 (1996).
using tristimulus colorimeters," ASTM E1455 (1997).
"Standard test method for obtaining spectroradiometric data from radiant sources
for colorimetry," ASTM E1341 (1996).
"Standard test method for photopic luminance of photoluminescent
(phosphorescent) markings," ASTM E2073 (2000).
"Standard test method for radiation thermometer (single waveband type),"
ASTM E1256 (1988).
"Standard test method for reflection haze of high-gloss surfaces," ASTM D4039
(1999).
"Standard test method for reflectance factor and color by spectrophotometry
using bidirectional geometry," ASTM E1349 (1998).
"Standard test method for solar absorptance, reflectance and transmittance of
materials using spectrophotometers with integrating spheres," ASTM E903
(1988).
"Standard test methods for solar energy transmittance and reflectance (terrestrial)
of sheet materials," ASTM E424 (1993).
"Standard test method for solar absorptance, reflectance, and transmittance of
materials using integrating spheres," ASTM E903 (1996).
"Standard test method for spectral bandwidth and wavelength accuracy of
fluorescence spectrometers," ASTM E388 (1998).
"Standard test method for specular gloss," ASTM D523 (1999).
"Standard test method for specular gloss of paper and paperboard at 75," ASTM
D 1223 (1998).
"Standard test method for total luminous reflectance factor by use of 30/t
integrating-sphere geometry," ASTM E 1651 (1999).
"Standard test method for transfer of calibration from reference to field
radiometers," ASTM E824 (1994).
"Standard test method for transmittance and color by spectrophotometry using
hemispherical geometry," ASTM E1348 (1996).
"Standard test method for transparency of plastic sheeting," ASTM D1746
(1997).
"Standard test method for visual evaluation of gloss differences between surfaces
of similar appearance," ASTM D4449 (1999).
"Standard test method for total hemispherical emittance of surfaces from 20 to
1400 C," ASTM C835 (2000).
"Standard test method for brightness of pulp, paper, and paperboard (directional
reflectance at 457 nm)," ASTM D985 (1997).
"Standard test method for calculation of color differences from instrumentally
measured color coordinates," ASTM D2244 (2000).
"Standard test method for calibration of pyrheliometers by comparison to
reference pyrheliometers," ASTM E816 (1995).
"Standard test method for calibration of a pyranometer using a pyrheliometer,"
ASTM G167 (2000).
"Standard test method for calibration of a spectroradiometer using a standard
source ofirradiance," ASTM G138 (1996).
"Standard test method for calibration of narrow- and broad-band ultraviolet
radiometers using a spectroradiometer," ASTM G130 (1995).
"Standard test method for calibration of primary non-concentrator terrestrial
photovoltaic reference cells using a tabular spectrum," ASTM El125 (1999).
"Standard test method for calibration of reference pyranometers with axis tilted
by the shading method," ASTM E941 (1999).
"Standard test method for calibration of silicon non-concentrator photovoltaic
primary reference cells under global irradiation," ASTM E 1 039 (1999).
"Standard test method for calorimetric determination of hemispherical emittance
and the ratio of solar absorptance to hemispherical emittance using solar
simulation," ASTM E434 (1996).
"Standard test method for coefficient of retroreflection of retroreflective sheeting
utilizing the coplanar geometry," ASTM E810 (2001).
"Standard test method for color and color-difference measurement by tristimulus
(filter) colorimetry," ASTM E1347 (1997).
"Standard test method for color of liquids using tristimulus colorimetry," ASTM
D5386 (2000).
"Standard test method for conducting aqueous direct photolysis tests," ASTM
E896 (1997).
"Standard test method for detecting delaminations in bridge decks using infrared
thermography," ASTM D4788 (1997).
"Standard test method for determining solar or photopic reflectance,
transmittance, and absorptance of materials using a large diameter integrating
sphere," ASTM E 1175 (1996).
"Standard test method for determining the linearity of a photovoltaic device
parameter with respect to a test parameter," ASTM El143 (1999).
"Standard test method for diffuse light transmission factor of reinforced plastics
panels," ASTM D1494 (1997).
"Standard test method for electrical performance of photovoltaic cells using
reference cells under simulated sunlight," ASTM E948 (1995).
"Standard test method for evaluating color image output from color printers and
copiers," ASTM F1206 (2000).
330 AppendixJ
"Standard test method for evaluation of color for thermoplastic traffic marking
materials," ASTM D4960 (1998).
"Standard test method for evaluation of visual color difference with a gray
scale," ASTM D2616 (1996).
"Standard test method for haze and luminous transmittance of transparent
plastics," ASTM DI003 (2000).
"Standard test method for hiding power of paints by reflectometry," ASTM
D2805 (1996).
"Standard test method for identifying fluorescence in object-color specimens by
spectrophotometry," ASTM E1247 (2000).
"Standard test method for linearity of fluorescence measuring systems," ASTM
E578 (2001).
"Standard test method for luminous reflectance factor of acoustical materials by
use of integrating-sphere reflectometers," ASTM E 1477 (1998).
"Standard test method for measuring total-radiance temperature of heated
surfaces using a radiation pyrometer," ASTM E639 (1996).
"Standard test method for measurement and calculation of reflecting
characteristics of metallic surfaces using integrating sphere instruments,"
ASTME429
"Standard test method for measurement of high-visibility retroreflective-c1othing
marking material using a portable retrorelectometer," ASTM E1809 (1996).
"Standard test method for measurement of retroreflective pavement marking
materials with cen-prescribed geometry using a portable retroreflectometer,"
ASTM El710 (1997).
"Standard test method for measurement of retroreflective signs using a portable
retroreflectometer," ASTM E1709 (2000).
"Standard test method for minimum detectable temperature difference for
"Standard test method for minimum resolvable temperature difference for
"Standard test method for noise equivalent temperature difference of thermal
imaging systems," ASTM E1543 (2000).
"Standard test method for normal spectral emittance at elevated temperatures,"
ASTM E307 (1996).
"Standard test method for normal spectral emittance at elevated temperatures of
nonconducting specimens," ASTM E423 (1996).
by spectroradiometry," ASTM E1336 (1996).
"Standard test method for opacity of paper (15 diffuse illuminant A, 89%
reflectance backing and paper backing)," ASTM D589 (1997).
"Standard test method for photoelastic measurements of birefringence and
residual strains in transparent or translucent plastic materials," ASTM D4093
(1995).
using hemispherical geometry," ASTM E1331 (1996).
using bidirectional geometry," ASTM E1349 (1998).
"Standard test method for reflection haze of high-gloss surfaces," ASTM D4039
(1999).
"Standard test method for relative tinting strength of aqueous ink systems by
instrumental measurement," ASTM D6531 (2000).
"Standard test method for relative tinting strength of white pigments by
reflectance measurements," ASTM D2745 (2000).
"Standard test method for retroreflectance of horizontal coatings," AS TM D4061
(2000).
"Standard test method for solar absorptance, reflectance, and transmittance of
materials using integrating spheres," ASTM E903 (1996).
"Standard test method for solar photometric transmittance of sheet materials
using sunlight," ASTM E972 (1996).
"Standard test method for solar transmittance (terrestrial) of sheet materials using
sunlight," ASTM E1 084 (1996).
"Standard test method for spectral bandwidth and wavelength accuracy of
fluorescence spectrometers," ASTM E388 (1998).
"Standard test method for specular gloss," ASTM D523 (1999).
"Standard test method for specular gloss of glazed ceramic whitewares and
related products," ASTM C584 (1999).
"Standard test method for total luminous reflectance factor by use of 30/t
integrating-sphere geometry," ASTM E 1651 (1999).
"Standard test method for transmittance and color by spectrophotometry using
hemispherical geometry," ASTM E1348 (1996).
"Standard test method for transparency of plastic sheeting," ASTM D1746
(1997).
"Standard test methods for continuous measurement of ozone in ambient,
workplace, and indoor atmospheres (ultraviolet absorption)," ASTM D5156
(1995).
"Standard test methods for measurement of gloss of high-gloss surfaces by
goniophotometry," ASTM E430 (1997).
"Standard test methods for measuring and compensating for emissivity using
infrared imaging radiometers," ASTM E1933 (1999).
"Standard test methods for measuring and compensating for reflected
temperature using infrared imaging radiometers," ASTM E1862 (1997).
332 Appendix J
"Standard test methods for measuring and compensating for transmittance of an
attenuating medium using infrared imaging radiometers," ASTM E1897
(1997).
"Standard test methods for measuring optical reflectivity of transparent
materials," ASTM E1682 (1996).
"Standard test methods for measuring spectral response of photovoltaic cells,"
ASTM E1021 (1995).
"Standard test methods for measuring total-radiance temperature of heated
surfaces using a radiation pyrometer" ASTM E639 (1990).
"Standard test methods for measurement of gloss of high-gloss surfaces by
goniophotometry," ASTM E430 (1997).
"Standard test methods for minimum detectable temperature difference for
thermal imaging systems" ASTM E1311 (1993).
"Standard test methods for minimum resolvable temperature difference for
thermal imaging systems" ASTM E1213 (1992).
"Standard test methods for noise equivalent temperature difference of thermal
imaging systems" ASTM E1543 (1994).
"Standard test methods for radiation thermometers (single waveband type),"
ASTM E1256 (1995).
"Standard test methods for solar energy transmittance and reflectance (terrestrial)
of sheet materials," ASTM E424 (1993).
"Standard test methods for total normal emittance of surfaces using inspection-
meter techniques," ASTM E408 (1996).
BIPM
The BIPM (Bureau International des Poids et Mesuresf is an international
institute operating under the supervision of the Comite International des Poids et
Mesures (CIPM). It is charged with the establishment and maintenance of
reference standards, the organization of international comparisons and carrying
out of calibrations, and fundamental investigations that may result in better
reference standards or measurement techniques. The prototype kilogram is
located here. Some of their publications include:
Principles Governing Photometry (1983).
The International System of Units (SI), BIPM, 7th Edition (1998).
"Radiation Thermometry," Chapter 6 in Supplementary Information for the
International Temperature Scale of 1990, BIPM (1990).
International Vocabulary of Basic and General Terms in Metrology, joint BIPM,
IEC, IFCC, ISO, IUPAC, IUPAP, OIML Standard, BIPM (1993).
Vwww.bipm.fr
Guide to the Expression of Uncertainty in Measurement, joint BIPM, IEC,IFCC,
ISO, IUPAC, IUPAP, OIML Standard, BIPM (1993).
CIE
The Commission Intemationale de L 'Eclairage vi has numerous technical
committee reports that are relevant. Check into their web page for more
information. Division 1 is involved with vision and color, while Division 2 deals
with the measurement oflight and radiation. The U.S. National Committee of the
ClE is at http://www.cie-usnc.org.Somepertinentreports.primarily from
Division 2, are:
13.3 Method of Measuring and Specifying Colour Rendering of Light Sources
(1995).
15.2 Colorimetry, 2nd Ed. (1986).
17.4 International Lighting Vocabulary, 4th Ed. (Joint publication IECiCIE)
(1987).
18.2 The Basis of Physical Photometry, 2nd Ed. (1983).
38 "Radiometric and photometric characteristics of materials and their
measurement," (1977).
41 "Light as a true visual quantity: principles of measurement," (1978).
44 "Absolute methods for reflection measurements," (1979).
46 "A review of publications on properties and reflection values of material
reflection standards," (1979).
53 "Methods of characterizing the performance of radiometers and photometers,"
(1982).
54 "Retroreflection: definition and measurement," (2001).
59 "Polarization: definitions and nomenclature, instrument polarization," (1984).
63 "The spectroradiometric measurement oflight sources," (1984).
64 "Determination of the spectral responsivity of optical radiation detectors,"
(1984).
65 "Electrically calibrated thermal detectors of optical radiation (absolute
radiometers)" (1985).
69 "Methods of characterizing illuminance meters and luminance meters:
Performance, characteristics and specifications," (1987).
70 "The measurement of absolute luminous intensity distributions," (1987).
75 "Spectral luminous efficiency functions based upon brightness matching for
monochromatic point sources, 20 and 100 fields," (1988).
vi www.cie.co.atJ Write to Thomas Lemons, TLA- Lighting Consultants Inc., 7 Pond St., Salem,
MA 01970 or contact Tel: (508) 745-6870, Fax: (508) 741-4420 for a listing of current
publications and pricing.
334 AppendixJ
76 "Intercomparison on measurement of (total) spectral radiance factor of
luminescent specimens," (1988).
78 "Brightness-luminance relations: classified bibliography," (1988).
81 "Mesopic photometry: history, special problems and practical solutions,"
(1989).
84 "Measurement of luminous flux," (1989).
85 "Solar spectral irradiance," (1989).
86 "CIE 1988 2 spectral luminous efficiency function for photopic vision,"
(1990).
87 "Colorimetry of self-luminous displays: a bibliography," (1990).
96 "Electric light sources, state of the art-1991 ," (1992).
1 05 "Spectroradiometry of pulsed optical radiation sources," (1993).
114 CIE Collection in Photometry and Radiometry (1994).
121 "The photometry and goniophotometry of luminaires," (1996).
125 "Standard erythema dose, a review," (1997).
127 "Measurement ofLEDs," (1997).
130 "Practical methods for the measurement of reflectance and transmittance,"
(1998).
141 "Testing of supplementary systems of photometry," (2001).
IES
The Illuminating Engineering Society of North America (IESNAf
ii
has an
extensive list of publications dealing with illumination. Their IESNA Lighting
Handbook, <jh Ed. is the definitive reference. The following are procedures
dealing with photometric measurements of various lamps and luminaires:
LM-9
LM-IO
LM-l1
LM-20
Electrical and Photometric Measurements of Fluorescent Lamps
Photometric Testing of Outdoor Fluorescent Luminaires
Photometric Testing of Searchlights
Photometric Testing of Reflector-Type Lamps
LM -31 Photometric Testing of Roadway Luminaires
LM-35 Photometric Testing of Floodlights Using High-Intensity Discharge
Lamps or Incandescent Filament Lamps
LM-41 Photometric Testing ofIndoor Fluorescent Luminaires
LM-44 Method for Total and Diffuse Reflectometry (1985)
LM-45 Electrical and Photometric Measurements of General Service
Incandescent Filament Lamps
vii www.iesna.org
LM-46 Photometric Testing of Indoor Luminaires Using HID Discharge or
Incandescent Filament Lamps
LM-50 Photometric Measurement of Roadway Lighting Installations
LM-51 Electrical and Photometric Measurements of High-Intensity Discharge
Lamps
LM-52 Photometric Measurement of Roadway Sign Installations
LM-54 Lamp Seasoning
LM-55 Measurement of Ultraviolet Radiation from Light Sources
LM-58 Spectroradiometric Measurements
LM-59 Electrical and Photometric Measurements of Low-Pressure Sodium
Lamps
LM-63 Standard File Format for Electronic Transfer of Photometric Data
LM-64 Photometric Measurements of Parking Areas
LM-66 Electrical and Photometric Measurements of Compact Fluorescent
Lamps
LM-68 Photometric Evaluation of Vehicle Traffic Control Signal Heads
LM-70 Near-Field Photometry
LM-72 Directional Positioning of Photometric Data
RP-16 Nomenclature and Definitions for Illuminating Engineering (ANSI
Approved)
ISO
The International Standards Organization. viii
ISO 2470: Brightness for Fluorescent Materials (1999).
ISO/CIE 10526: CIE standard illuminants for colorimetry (CIE S005/E-1998)
(1999).
ISO/CIE 10527: CIE standard colorimetric observers (CIE S002, 1986) (1991).
ISO/CIE 15469: Spatial distribution of daylight-CIE standard overcast sky and
clear sky (CIE S003, 1996) (1997).
ISO/CIE 16508: Road traffic lights-Photometric properties of 200 mm round
signals (CIE S006) (1999).
IS017166:
ISO 11475:
viii www.iso.ch
Erythema reference action spectrum and standard erythema
dose (CIE S007) (1999).
Paper and board-Determination of CIE whiteness, D65/ 10
degrees (outdoor daylight) (1999).
336 Appendix J
ISO 11476: Paper and board-Determination ofCIE-whiteness, C/2 degrees
(indoor illumination conditions) (2000).
ISO 8599: Optics and optical instruments-Contact lenses-Determination
of the spectral and luminous transmittance (1994).
ISO 9845-1: Solar energy-Reference solar spectral irradiance at the ground
at different receiving conditions-Part 1: Direct normal and
hemispherical solar irradiance for air mass 1,5 (1992).
ISO 9022-9: Optics and optical instruments-Environmental test methods-
Part 9: Solar radiation (1994).
ISO 9022-17: Optics and optical instruments-Environmental test methods-
Part 17: Combined contamination, solar radiation (1994).
ISO 9050: Glass in building-Determination of light transmittance, solar
direct transmittance, total solar energy transmittance and
ultraviolet transmittance, and related glazing factors (1990).
ISO 9059: Solar energy-Calibration of field pyrheliometers by
comparison to a reference pyrheliometer (1990).
ISO 9060: Solar energy-Specification and classification of instruments
for measuring hemispherical solar and direct solar radiation
(1990).
ISO 9488: Solar energy-Vocabulary (1999).
ISO 9846: Solar energy-Calibration of a pyranometer usmg a
pyrheliometer (1993).
ISO 9847: Solar energy-Calibration of field pyranometers by comparison
to a reference pyranometer (1992).
ISO/TR 9901: Solar energy-Field pyranometers-Recommended practice for
use (1990).
ISO 6: Photography-Black-and-white pictorial still camera negative
film/process systems-Determination of ISO speed (1993).
ISO 2240: Photography-Colour reversal camera films-Determination of
ISO speed (1994).
ISO 8478: Photography-Camera lenses-Measurement of ISO spectral
transmittance (1996).
ISO 12232: Photography-Electronic still-picture cameras-Determination
of ISO speed (1998).
lEe
Founded in 1906, the International Electrotechnical Commission (IEC)ix is the
world organization that prepares and publishes international standards for all
;x lEe, 3, Rue de Varembe, PO Box 131, 1211 Geneva 20, Switzerland. Tel: +412291902 II ,
Fax: +41229190300, Website: http://www.iec.ch.
electrical, electronic, and related technologies. The IEC was founded as a result
of a resolution passed at the International Electrical Congress held in St. Louis,
Missouri in 1904. The membership consists of more than 50 participating
countries, including all the world's major trading nations and a growing number
of industrializing countries.
IEC
61966-2-1
Ed. 1: 1999
IEC 61966-3
Ed. I: 2000
IEC 61966-4
Ed. 1:2000
IEC 61966-5
Ed. 1:2000
IEC 61966-8
Ed. 1:2001
IEC 61966-9
Ed. I: 2000
NVLAP
Multimedia systems and equipment --Colour measurement and
management Part 2-1: Colour management-Default RGB
colour space-sRGB
management Part 3: Equipment using cathode ray tubes.
management Part 4: Equipment using liquid crystal display
panels.
management Part 5: Equipment using plasma display panels.
Multimedia systems and equipment -Colour measurement and
management Part 8: Multimedia colour scanners.
Multimedia systems and equipment -Colour measurement and
management Part 9: Digital cameras.
The National Voluntary Laboratory Accreditation Program.
x
Established in 1976
and administered by the National Institute of Standards and Technology (NIST),
NVLAP is an unbiased government-based third-party system for accrediting
calibration laboratories and testing laboratories found competent to perform
specific tests or calibrations. Criteria for NVLAP accreditation are published in
the Code of Federal Regulations (Title 15, Part 285) and encompass the
requirements of ISO/IES Guide 25 and the relevant requirements of ISO 9002.
NVLAP accreditation is available to commercial laboratories, manufacturers'
in-house laboratories, university laboratories, federal, state, and local government
laboratories, and foreign-based laboratories.
"Procedures and General Requirements," NIST Handbook 150, National
Voluntary Laboratory Accreditation Program (NVLAP), C. D. Faison et aI,
Eds. (February 2006).
x National Institute of Standards and Technology, National Voluntary Laboratory Accreditation
Program, 100 Bureau Drive, MS 2140, Gaithersburg, Maryland 20899-2140, Telephone: 301-975-
4016, Fax: 301-926-2884, E-mail: NVLAP@nist.gov, Website: http://www.ts.nist.gov/nvlap.
338 AppendixJ
"Energy Efficient Lighting Products," NIST Handbook 150-1, National
Voluntary Laboratory Accreditation Program (NVLAP), C. C. Miller and L.
I. Knab, Eds. (March 2008). xi
NVLAP identifies its accredited laboratories in a published directory, NIST
Special Publication 810, and on their website.
SAE
SAE International, formerly the Society of Automotive Engineers,xii is a
nonprofit educational and scientific organization dedicated to advancing mobility
technology to better serve humanity. Nearly 70,000 engineers and scientists who
are SAE members develop technical information on all forms of self-propelled
vehicles including automobiles, trucks and buses, off-highway equipment,
aircraft, aerospace vehicles, marine, rail, and transit systems. SAE disseminates
this information through meetings, books, technical papers, magazines,
standards, reports, professional development programs, and electronic databases.
Here is a selection of their relevant standards, mostly dealing with lighting and its
measurement.
HS-34 SAE Ground Vehicle Lighting Standards Manual (1999).
J387 "Terminology-Motor Vehicle Lighting" (1995).
11330 "Photometry Laboratory Accuracy Guidelines" (1994).
J575 "Test Methods and Equipment for Lighting Devices and Components
for Use on Vehicles Less than 2032 mm in Overall Width" (1992).
J 1383 "Performance Requirements for Motor Vehicle Headlamps" (1996).
J2217 "Photometric Guidelines for Instrument Panel Displays that
Accommodate Older Drivers" (1991).
TAPPI
The Technical Association of the Pulp and Paper Industry.xiii TAPPI is the
leading technical association for the worldwide pulp, paper, and converting
industry. TAPPI provides its members rapid access to: (1) the largest
international group of technically experienced people in the industry, (2) the most
comprehensive collection of reliable technical information and knowledge in the
industry, and (3) the highest quality products and services created to meet the
needs of people who solve technical problems in the industry. Among their
documentary standards are the following:
T-425 Opacity of paper (15/d geometry, Illuminant N2, 89% reflectance
backing and paper backing) (1996).
xi Handbook 150 and 150-1 are available on the NVLAP website at http://www.ts.nist.gov/nvlap.
xii www.sae.org
xiii http://www.tappi .org
T-452
T-480
T-560
T-562
T-1212
T-1213
T-1214
T-1215
T-1216
T-1217
T-1218
Brightness of pulp, paper and paperboard (directional reflectance at
457 nm) (1998).
Specular gloss of paper and paperboard at 75 degrees (1999).
CIE whiteness and tint of paper and paperboard (diO geometry, C/2
illuminantJobserver) (1996).
CIE whiteness and tint of paper and paperboard (45/0 geometry, C/2
illuminantJobserver) (1996).
Light sources for evaluating papers, including those containing
fluorescent whitening agents (1998).
Optical measurements terminology (related to appearance evaluation
of paper) (1998).
Interrelation of reflectance; reflectivity; T APPI opacity; scattering, s;
and absorption, k (1998) .
The determination of instrumental color differences (1998).
Indices for whiteness, yellowness, brightness, and luminous
reflectance factor (1998).
Photometric linearity of optical properties instruments (1998).
Calibration of reflectance standards for hemispherical geometry
(1998).
-
Appendix K
Radiometry and Photometry
Bibliography
In 1969 Fred Nicodemus authored a paper "Optical resource letter on
radiometry" (JOSA 59, 243). It was reprinted in an AlP Radiometry-Selected
Reprints collection in 1971. This bibliography is intended to update and
supplement the 1969 version. The topic breakdown is more in line with the order
of topics in this book than the original. Individual papers are for the most part
bypassed in favor of books, significant book chapters, monographs, or reprint
collections.
Symbols, Units, and Terminology
American National Standard Nomenclature and Definitions for Illuminating
Engineering. ANSI, Standard ANSIIIESNA RP-16 96 (1996).
International Lighting Vocabulary, CIE Publication 17 (1970).
"Quantities and units-Part 6. Light and related electromagnetic radiations," ISO
Standards Handbook, Quantities and Units 389.15 (1993).
Symbols, Units and Nomenclature in Physics, International Union of Pure and
Applied Physics (1987).
C. L. Wyatt, V. Privalsky, and R. T. Datiu, Recommended Practice: Symbols,
Units and Uncertainty Analysis for Radiometric Sensor Calibration, NIST
Handbook 152 (1998).
General
R. W. Boyd, Radiometry and the Detection of Optical Radiation, John Wiley &
Sons, New York (1983). Good for the theoretical aspects of radiation
geometry, along with several other topics.
M. A. Bramson, Infrared: A Handbook for Applications, Plenum, New York
(1966).
F. C. Grum, and R. J. Becherer, Optical Radiation Measurements 1. Radiometry,
Academic Press, New York (1979). The best all-around book on radiometry to
date.
341
342 Appendix K
A. Hadni, Essentials of Modern Physics Applied to the Study of the Infrared,
Pergamon, Oxford Univ. Press, New York (1967).
M. R. Holter, S. Nudelman, G. H. Suits, W. L. Wolfe and G. J. Zissis,
Fundamentals of Infrared Technology, Macmillan, New York (1962).
J. A. Jamieson, R. H. McFee, G. N. Plass, R. H. Grube, and R. G. Richards,
Infrared Physics and Engineering, McGraw-Hill, New York (1963).
P. W. Kruse, L. D. McGlaughlin, and R. B. McQuistan, Elements of Infrared
Technology, John Wiley & Sons, New York (1962).
W. R. McCluney, Introduction to Radiometry and Photometry, Artech House,
Boston (1994). A recent entry, again slightly slanted to the author's (Palmer)
specialty of solar energy applications. Both elementary and thorough.
I. J. Spiro and M. Schlessinger, Infrared Technology Fundamentals, Marcel
Dekker, New York (1989).
W. L Wolfe and G. Zissis, The Infrared Handbook, Office of Naval Research,
Washington, D.C. (1978). The most bang for your buck! At $60, everybody
should have one! Available from SPIE.
W. L. Wolfe, Introduction to Radiometry, SPIE Press, Bellingham, Washington
(1998). Uneven and terse in spots, still recommended. He was my (Palmer)
mentor!
E. F. Zalewski, "Radiometry and Photometry," Chapter 24 in Handbook of
Optics, Vol. II, McGraw-Hill, New York (1995). An excellent review.
Geometric Radiation Transfer
A. Gershun, "The Light Field," Moscow, 1936, trans. P. Moon and G.
Timoshenko, J. Math. Phys. 18, 51-151 (1939).
H. C. Hottel and A. F. Sarofim, Radiative Transfer, McGraw-Hill, New York
(1967).
J. R. Howell, A Catalog of Radiation Configuration Factors, McGraw-Hill, New
York (1982).
F. E. Nicodemus et aI., Self-Study Manual on Optical Radiation Measurements,
NBS Technical Note 910-xx. NIST, Washington (various dates). A series that
is rather theoretical with increasing coverage. Some may be out of print. Check
GPO for availability.
R. Siegel and J. R. Howell, Thermal Radiation Transfer, Hemisphere,
Washington, D.C. (1981). Intended for mechanical and heat transfer engineers
(uses non-SI units like BTUs, degrees Rankine, and the wrong intensity) and
highly detailed, this book is the best compendium on radiative transfer.
E. M. Sparrow and R. D. Cess, Radiation Heat Transfer, Brooks/Cole, California
(1970).
J. H. Taylor, Radiation Exchange, Academic Press, New York (1990).
Radiometry and Photometry Bibliography 343
Radiosity and Ray Tracing
I. Ashdown, Radiosity: A Programmer's Perspective, John Wiley & Sons, New
York (1994).
M. F. Cohen and J. R. Wallace, Radiosity and Realistic Image Syntheses,
Optical Radiation Sources
F. E. Carlson and C. N. Clark, Light Sources for Optical Devices, Chapter 2 in
Applied Optics and Optical Engineering, Vol. 1, R. Kingslake, Ed., Academic
K. L. Coulson, Solar and Terrestrial Radiation, Academic Press, New York
(1975).
J. E. Eby and R. E. Levin, Incoherent Light Sources, Chapter 1 in Applied Optics
and Optical Engineering, Vol. 7, R. R. Shannon and J. Wyant, Eds., Academic
A. S. Green, Ed. The Middle Ultraviolet: Its Science and Technology, John Wiley
& Sons, New York (1966).
F. Grum and R. J. Becherer, Radiometry, Vol. 1 in Optical Radiation
Measurements, Academic Press, New York (1979).
H. Hewitt and A. S. Vause, Lamps and Lighting, Elsevier, New York (1966).
M. Iqbal, An Introduction to Solar Radiation, Academic Press, New York (1983).
P. A. Jacobs, Thermal Infrared Characterization of Ground Targets and
Backgrounds, SPIE Press, Bellingham, Washington (1996).
D. Kryskowski and G. H. Suits, "Natural sources," Chapter 3 in Sources of
Radiation, G. J. Zissis, Ed., Vol. 1 of The Infrared & Electro-Optical
Handbook, SPIE Press, Bellingham, Washington (1993).
A. J. LaRocca, "Artificial sources," Chapter 2 in Sources of Radiation, G. J.
Zissis, Ed., Vol. 1 of The Infrared & Electro-Optical Handbook, SPIE Press,
A. J. LaRocca, "Artificial sources," Chapter 10 in Optical Sources, G. J. Zissis,
Ed., Vol. 1, Part 4 Handbook of Optics, McGraw Hill, New York (1995).
M. Luckiesh, Applications of Germicidal, Erythemal and Infrared Energy, Van
Nostrand, New York (1946).
D. Malacara, "Light sources," Chapter 5 In Geometrical and Instrumental Optics,
Academic Press, Boston (1988).
P. Moon, The Scientific Basis of Illuminating Engineering, McGraw-Hill, New
York (1936), reprinted by Dover, New York (1961).
J. B. Murdoch, Illumination Engineering: From Edison 's Lamp to the Laser,
344 Appendix K
RCA Electro-Optics Handbook, RCA, Lancaster PA (1974). Now available from
Burle Industries.
M. Rea, Ed., Lighting Handbook: Reference and Application, 8
1h
Ed.,
Illuminating Engineering Society of North America (1993).
J. C. Richmond and F. E. Nicodemus, "Blackbodies, blackbody radiation, and
temperature scales," NBS Self-Study Manual on Optical Radiation
Measurements, Part 1, Chapter 12, GPO (1985).
W. T. Silfvast, "Lasers," Chapter 11 in Handbook of Optics, Vol. 1, Part 4.
Optical Sources, McGraw-Hill, New York (1995).
G. H. Suits, "Natural sources," Chapter 3 in The Infrared Handbook, ERIM
(1982). Distributed By SPIE.
R. H. Weissman, "Light emitting diodes," Chapter 12 in Handbook of Optics,
Vol. 1, Part 4. Optical Sources, McGraw-Hill, New York (1995).
G. J. Zissis and A. J. Larocca, "Optical Radiators and Sources," Chapter 3 in
Handbook Of Optics, McGraw-Hill, New York (1978).
Detectors
A. Ambroziak, Semiconductor Photoelectric Devices, Gordon & Breach, New
York (1969).
D. F. Barbe, Ed., Charge-Coupled Devices, Springer-Verlag, New York (1980).
C. J. Bartelson, F. C. Grum, Optical Radiation Measurements Vol. 5: Visual
Measurements, Academic Press, New York (1984).
1. D. E. Beynon and D. R. Lamb, Charge-Coupled Devices and their
Applications, McGraw-Hill, New York (1980).
L. Biberman and S. Nudelman, Photoelectronic Imaging Devices, Plenum, New
York (1970). Two volumes.
R. M. Boynton, Human Color Vision, Holt, Rinehart & Winston, New York
(1979).
W. Budde, Optical Radiation Measurements Vol 4: Physical Detectors of Optical
Radiation, Academic Press, New York (1983). The best practical book on
detectors and their characterization, slanted a bit towards the visible spectrum.
C. Buil, CCD Astronomy, Willman-Bell, Richmond, Virginia (1991).
A. Chappell, Optoelectronics: Theory and Practice, McGraw-Hill, New York
(1978).
I. P. Csorba, Image Tubes, Howard W. Sams Engineering-Reference Book
Series, Indianapolis, Indiana (1985).
P. N. J. Dennis, Photodetectors, Plenum, New York (1986).
E. L. Dereniak and G. Boreman, Infrared Detectors and Systems, John Wiley &
E. L. Dereniak and D. Crowe, Optical Radiation Detectors, John Wiley & Sons,
New York (1984). Good general book, slanted towards the infrared.
S. Donati, Photodetectors: Devices, Circuits and Applications, Prentice Hall,
New Jersey (2000). Concise, accurate, highly recommended.
M. J. Eccles, M. E. Sim, and K. P. Tritton, Low Light Level Detectors in
Astronomy, Cambridge Univ. Press, New York (1983).
R. W. Engstrom, RCA Photomultiplier Handbook, RCA (1980). Available from
Burle Industries (http://www.burle.comlcgi-binibyteserver.pllpdf/Photo.pdf).
G. C. Holst, CMOS/CCD Sensors and Camera Systems, SPIE Press, Bellingham,
Washington (2007).
R. D. Hudson and 1. W. Hudson, Eds., Infrared Detectors, Hutchinson and Ross,
Stroudsburg, Pennsylvania (1975). A collection of the more important papers
to date.
1. R. Janesick, Scientific Charge-Coupled Devices, SPIE Press, Bellingham,
Washington (2001).
R. J. Keyes, Ed., Optical and Infrared Detectors, Springer-Verlag, Bellingham,
Washington (1980).
I. S. McLean, Electronic Imaging in Astronomy, John Wiley & Sons, New York
(1997).
Photomultiplier Tube: Principle to Application, H. Kume, Ed., Hamamatsu
Photonics K. K. (1994).
S. O. Flyckt and C. Marmonier, Photomultiplier Tubes: Principles and
Applications, Philips Photonics, Brive, France (2002).
G. H. Rieke, Detection of Light from the Ultraviolet to the Submillimeter,
Cambridge Univ. Press, New York (1994).
A. Rose, Vision: Human and Electronic, Plenum, New York (1973).
M. Ross, Laser Receivers, John Wiley & Sons, New York (1966).
C. H. Sequin and M. F. Tomsett, Charge Transfer Devices, Academic Press,
New York (1975).
R. A. Smith, F. E. Jones, and R. P. Chasmar, The Detection and Measurement of
Infra-red Radiation, Oxford Univ. Press, New York (1957).
A. H. Sommer, Photoemissive Materials, John Wiley & Sons, New York (1968).
Reprinted by Krieger, Huntington, New York (1980).
1. D. Vincentm, Fundamentals of Infrared Detector Operation and Testing, John
Wiley & Sons, New York (1990). Separate chapters on cryogenics, vacuum
systems, and detector electronics. Answers many of those nagging questions.
Highly recommended.
A. T. Young, "Photomultipliers, their causes and cures," Chapter 1 in Methods of
Experimental Physics: Astrophysics, N. Carleton Ed., Vol. 12, Academic Press,
New York (1974).
-------------------
346 Appendix K
The major handbooks offer one or more chapters dealing with optical radiation
detectors. These include the following:
J. S. Accetta and D. L. Shumaker, Eds., Infrared and Electro-Optical Systems
Handbook, SPIE Press, Bellingham, Washington (1993). See Vol. 3, Chapter 4.
M. Bass, Ed. Handbook of Optics, Vol. 1, McGraw-Hill, New York (1995). Parts
5, 6, and 7 contain chapters 15 through 25, all pertinent to detectors and
detection.
W. G. Driscoll and W. Vaughn, Handbook of Optics, McGraw-Hill, New York
(1978). Chapter 4.
W. L. Wolfe, Ed., Handbook of Military Infrared Technology, ONR, Washington
(1965). Chapters 11 & 12.
W. L. Wolfe and G. Zissis, Eds., The Infrared Handbook, ERIM and SPIE
(1978). Chapters 11 through 16.
Several books in the series Semiconductors and Semimetals, R. K.
Willardson and A. C. Beer, Eds., are relevant. Volumes 5 and 12 both treat
infrared detectors, and Vol. 11 deals exclusively with solar cells. Volume 47
deals with uncooled infrared imaging array systems."
Similarly, the series Advances in Electronics and Electron Physics is rich in
pertinent articles. Numerous volumes deal with imaging detector conference
proceedings, while others have significant feature articles. Most notable are Vols.
34 and 55.
Another rich source for detector information, and probably the best for
assessment of the current state of art, are the Proceedings of the various
conferences of SPIE (see p. 355), available on SPIE's digitallibrary.i They have
several major conventions per year, each having one or more conferences on
detectors. Compilations of the best proceedings papers along with seminal papers
are often gathered in their "Milestones" series.
Several detector manufacturers publish manuals, collections of applications
notes, etc. Particularly attractive are the photomultiplier tube books by Burle
(formerly RCA), Hamamatsu, and Philips, and the CCD books by EG&G
Reticon and Fairchild. Data books by Dalsa, EG&G Vactec, Texas Instruments,
Thompson CSF, etc., are also quite helpful.
Noise, Electronics, and Signal Processing
D. A. Bell, Noise and the Solid State, John Wiley & Sons, New York (1985).
E. R. Davies, Electronics, Noise and Signal Recovery, Academic Press, New
York (1983).
P. J. Fish, Electronic Noise and Low Noise Design, McGraw-Hill, New York
(1994).
i http://spiedigitalIibrary.aip.orgJ
T. M. Frederiksen, Intuitive Ie Op Amps, National Semiconductor (1984).
J. G. Graeme, Photodiode Amplifiers: Op-Amp Solutions, McGraw-Hill, New
York (1996).
M. S. Gupta, Ed., Electrical Noise: Fundamentals and Sources, IEEE (1977).
P. Horowitz and W. Hill, The Art of Electronics, Cambridge Univ. Press, New
York (1989). The best single book available for the broad field of electronics,
my personal favorite!
W. G. Jung, IC Op-amp Cookbook, Howard W. Sams (1980). More tutorial than
Stout, with fewer useful circuits.
C. D. Motchenbacher and J. A. Connelly, Low-Noise Electronic System Design,
H. W. Ott, Noise Reduction Techniques in Electronic Systems, 2
nd
Ed., John
R. A. Pease, Troubleshooting Analog Circuits, Butterworth-Heinemann, Newton,
D. F. Stout and M. Kaufman, Handbook of Operational Amplifier Circuit Design,
McGraw-Hill, New York (1976). Excellent single-purpose reference book.
A. Van der Ziel, Noise in Measurements, John Wiley & Sons, New York (1976).
C. A. Vergers, Handbook of Electrical Noise: Measurement and Technology, 2
nd
Ed., Tab Books, Inc., Pennsylvania (1987).
1. Williams, Analog Circuit Design: Art, Science and Personalities, Butterworth-
Heinemann, Newton, Massachusetts (1991).
1. Williams, The Art and Science of Analog Circuit Design, Butterworth-
Heinemann, Newton, Massachusetts (1995).
T. H. Wilmshurst, Signal Recovery, Adam Hilger, Bristol (1990).
A. Yariv, Optical Electronics in Modern Communications, Oxford Univ. Press,
New York (1997). Chapter 10 treats noise in optical detection and generation.
Radiometric Instruments
B. N. Begunov, N. P. Zakaznov, S. 1. Kiryushinand V. 1. Kuzichev, Optical
Instrumentation: Theory and Design, MIR, Moscow (1988).
R. D. Hudson, Infrared System Engineering, John Wiley & Sons, New York
(1969). Dated but still valuable.
J. F. James and R. S. Sternberg, Design of Optical Spectrometers, Chapman &
Hall, London (1969).
K. Seyrafi, Ed., Engineering Design Handbook on Infrared Military Systems,
United States Army Materiel Command, Pamphlet AMCP 706-127 (1971).
Similar to Hudson but in MIL-format. Later modified and self published as
Electro-Optical Systems Analysis, Electro-Optical Research Co, Los Angeles
( 1985).
348 Appendix K
W. L. Wolfe, Introduction to Infrared System Design, SPIE Press, Bellingham,
Washington (1996).
W. L. Wolfe, Introduction to Imaging Spectrometers, SPIE Press, Bellingham,
Washington (1997).
W. L. Wolfe, Infrared Design Examples, SPIE Press, Bellingham, Washington
(1999).
C. L. Wyatt, Radiometric Systems Design, Macmillan, New York (1987). More
for the designer than the calibrator, but very useful. Later available as Electro-
Optical System Design for Information Processing, McGraw-Hill, New York
(1990). More for the system designer, but very useful.
Radiometric Measurements and Errors
D. C. Baird, Experimentation: An Introduction to Measurement Theory and
Experiment Design, Prentice-Hall, New York (1962).
Y. Beers, Introduction to the Theory of Error, Addison-Wesley, Reading,
P. R. Bevington and D.K. Robinson, Data Reduction and Error Analysis for the
Physical Sciences, 2
nd
Ed., McGraw-Hill, New York (1992).
Calibration: Philosophy in Practice, 2
nd
Ed., Fluke Corporation, Everett,
Washington (1994). Available directly from them.
J. P. Holman, Experimental Methods for Engineers, McGraw-Hill, New York
(1978).
J. Mandel, The Statistical Analysis of Experimental Data, Dover, New York
(1984).
S. L. Meyer, Data Analysis for Scientists and Engineers, John Wiley & Sons,
New York (1975).
S. Rabinovich, Measurement Errors: Theory and Practice, AlP Press (1995).
B. N. Taylor & c. E. Kuyatt, Guidelines for Evaluating and Expressing the
Uncertainty of NIST Measurement Results, NIST Technical Note 1297 (1994).
Available on the web at http://physics.nist.govlPubs/guidelines/contents.html.
J. R. Taylor, An Introduction to Error Analysis, 2
nd
Ed., University Science
(1997).
H. D. Young, Statistical Treatment of Experimental Data, McGraw-Hill, New
York (1962).
Measurement of Radiometric Quantities
G. Bauer, Measurement of Optical Radiations, Focal Press, London (1965).
W. E. Forsythe, Ed., Measurement of Radiant Energy, McGraw-Hill, New York
(1937).
H. J. Kostkowski, Reliable Spectroradiometry, Spectroradiometry Consulting, La
Plata, Maryland 20646. Self-published lessons learned over some 40 years.
Authoritative!
D. Malacara, Physical Optics and Light Measurements, Academic Press, New
York (1988). Chapter 6.
Photometry
W. E. Barrows, Light, Photometry and Illuminating Engineering, McGraw-Hill,
New York (1951).
P. Bouguer, Essai d'optique sur fa gradation de la lumiere, Paris (1729). It all
started here! Translated into English by W. E. K. Middleton, The Gradation of
Light, Toronto (1961).
C. DeCusatis, Handbook of Applied Photometry, AlP Press (1997). Authoritative,
with pertinent chapters written by technical experts at BIPM, CIE, and NIST.
Primarily for visible radiation, mostly excellent.
C. Fabry, "Photometrie," Revue d'Optique, Paris (1927).
H. A. E. Keitz, Light Calculations and Measurements, Cleaver Hume, London
(1955), MacMillan, New York (1971).
J. H. Lambert, Photometrie, Augsberg (1760). A beautiful English translation
from the original Latin with copious notes by D. DiLaura was published in
200 I, available from IESNA.
Y. Ohno, "Radiometry and Photometry Review for Vision Optics," Chapter 14 in
Handbook of Optics Ill, McGraw-Hill, New York (2001).
J. M. Palmer, "Radiometry and Photometry: Units and Conversions," Chapter 7
in Handbook of Optics III, McGraw-Hill, New York (2001).
A. Stimson, Photometry and Radiometry for Engineers, John Wiley & Sons,
New York (1974).
J. W. T. Walsh, Photometry, 3
rd
Ed., Constable (1958), reprinted by Dover, New
York (1965). The standard reference since 1924.
Astronomical Measurement of Light
I choose not to include these in the photometry section to avoid confusion with
the strict definition of the word, the measurement of light.
A. A. Henden and R. H. Kaitchuk, Astronomical Photometry, Van Nostrand
Reinhold, New York (1982).
J. L. Hopkins, Zen and the Art of Photoelectric Photometry, HPO Desktop
Publishing (1990).
350 Appendix K
Color and Appearance
A. Berger-Schunn, Practical Color Measurement, John Wiley & Sons, New
York (1994).
F. W. Billmeyer and M. Salzman, Principles of Color Technology, Wiley-
Interscience, New York (1981).
R. W. Burnham, R. M. Hanes, and C. J. Bartleson, Color: A Guide to the Basic
Facts and Concepts, John Wiley & Sons, New York (1963).
G. J. Chamberlin and D. G. Chamberlin, Colour: Its Measurement, Computation
and Application, Heyden & Sons, London (1980).
Committee on Colorimetry, The Science of Color, ~ S A Crowell (1953). A
second edition is due any day now!
R. M. Evans, An Introduction to Color, John Wiley & Sons, New York (1948).
B. Fortner and T. E. Meyer, Number by Colors, Springer-Verlag, New York
(1997).
F. C. Grum, and C.J. Bartelson, Optical Radiation Measurements II: Color
Measurement, Academic Press, New York (1981).
A. C. Hardy, Handbook of Colorimetry, MIT Press, Cambridge (1936).
R. S. Hunter, The Measurement of Appearance, John Wiley & Sons, New York
(1987).
G. KortUm, Kolorimetrie, Photometrie und Spektrometrie, Springer-Verlag,
Berlin (1962).
D. B. Judd, Contributions to Color Science, D. L. MacAdam, Ed., NBS Special
Publication SP-545 (1979). A collection of the significant papers by Judd, the
Deane of colorimetry!
D. B. Judd and G. Wysczecki, Color in Business, Science and Industry, John
W. D. Wright, The Measurement of Colour, Macmillan, New York (1969).
G. Wyszecki, and W. S. Stiles, Color Science, John Wiley & Sons, New York
(1967); 2
nd
revised Ed. John Wiley & Sons, New York (1982). This classic has
been reprinted in a low-cost paperback format by John Wiley & Sons, New
York (2000). A must!
Radiometric Properties of Materials
H. H. Blau and H. Fischer, Radiative Transfer from Solid Materials, Macmillan,
New York (1962).
C. Burgess and K. D. Mielenz, Advances in Standards and Methodology in
Spectrophotometry, Elsevier, New York (1987).
F. J. Clauss, Material Effects in Spacecraft Thermal Control, John Wiley & Sons,
New York (1960).
po
G. G. Gubareff, J. E. Janssen, and R. H. Torberg, Thermal Radiation Properties
Survey, Honeywell, Minneapolis (1960).
B. Hapke, Theory of Reflectance and Emittance Spectroscopy, Cambridge Univ.
Press (1993).
J. Kamler, Luminescent Screens: Photometry and Colorimetry, Illife, London
(1969).
S. Katzoff, Ed., Symposium on Thermal Radiation of Solids, NASA SP-55
(1965).
T. G. Kyle, Atmospheric Transmission, Emission and Scattering, Pergamon,
Oxford (1991).
E. P. Lavin, Specular Reflection, Elsevier, New York (1971).
R. Mavrodinuenu, J. I. Schultz, and O. Menis, Accuracy in Spectrophotometry
and Luminescence Measurements, NBS Special Publication 378 (1973).
K. D. Mielenz, Ed., Optical Radiation Measurements III: Measurement of
Photoluminescence, Academic Press, New York (1982).
J. C. Richmond, Ed., Measurement of Thermal Radiation Properties of Solids,
NASA SP-31 (1963).
G. H. Schenk, Absorption of Light and Ultraviolet Radiation: Fluorescence and
Phosphorescence Emission, Allyn & Bacon, Boston (1973).
J. C. Stover, Optical Scattering: Measurement and Analysis, SPIE Press,
R. Tilley, Colour and the Optical Properties of Materials, John Wiley & Sons,
New York (2000).
Spectroscopy
R. B. Barnes, R. C. Gore, U. Liddel, and V. Z. Williams, Infrared Spectroscopy:
Industrial Applications and Bibliography, Reinhold, New York (1944).
W. R. Brode, Chemical Spectroscopy, John Wiley & Sons, New York (1943).
J. R. Edisbury, Practical Hints on Absorption Spectrometry, Plenum, New York
(1967).
N. J. Harrick, Internal Reflection Spectroscopy, Harrick Scientific, Ossining,
New York (1967).
G. R. Harrison, R. C. Lord, and J. R. Loofbourow, Practical Spectroscopy,
Prentice-Hall, New Jersey (1948).
J. F. James and R. S. Sternburg, The Design of Optical Spectrometers, Chapman
& Hall, London (1969).
G. KortUm, Reflectance Spectroscopy, Springer-Verlag, Berlin (1969).
J. R. Lakowicz, Principles of Fluorescence Spectroscopy, Plenum, New York
(1983).
G. F. Lothian, Absorption Spectrophotometry, Hilger & Watts, London (1958).
352 Appendix K
E. J. Meehan, Optical Methods of Analysis, Interscience, New York (1964).
R. A. Sawyer, Experimental Spectroscopy, Prentice-Hall, New Jersey (1951).
Steams, E. I., The Practice of Absorption Spectrophotometry, John Wiley &
J. E. Stewart, Infrared Spectroscopy, Marcel Dekker, New York (1970).
W. W. Wendlandt and H. G. Hecht, Reflectance Spectroscopy, Interscience, New
York (1966).
J. Workman and A. Springsteen, Applied Spectroscopy: A Compact Referencefor
Practitioners, Academic Press, New York (1998).
Laser Power and Energy
H. G. Heard, Laser Parameter Measurements Handbook, John Wiley & Sons,
New York (1968).
Temperature
D. P. DeWitt and G. D. Nutter, Theory and Practice of Radiation Thermometry,
H. Kaplan, Practical Applications of Infrared Thermal Sensing and Imaging
Equipment, SPIE Press, Bellingham, Washington (1999).
J. M. Lloyd, Thermal Imaging Systems, Plenum, New York (1975).
T. Quinn, Temperature, 2
nd
Ed., Academic Press (1990).
J. C. Richmond and D. P. Dewitt, Eds., Applications of Radiation Thermometry,
ASTM (1985).
W. P. Wood and J. M. Cork, Pyrometry, McGraw-Hill, New York (1941).
Radiometric Standards and Calibration
F. Hengtsberger, Absolute Radiometry, Academic Press, New York (1989).
Broad coverage with emphasis on room-temperature electrically calibrated
thermal radiometers.
C. L. Wyatt, Radiometric Calibration: Theory and Methods, Academic Press,
New York (1978). Good, fairly thorough treatment, somewhat slanted towards
the author's (Palmer) specialty, which is far-infrared cryogenic radiometry.
Compilations
A. J. Drummond, Ed., Precision Radiometry, Vol. 14 in Advances in Geophysics,
H. K. Hammond ITI and H. L. Mason, Selected NBS Papers on Radiometry and
Photometry, in Precision Measurement and Calibration, NBS Special
Publication 300, Vol. 7 (1971).
R. B. Johnson and W. L. Wolfe, Selected Papers on Infrared Design, 2 vols.
SPIE Press, Bellingham, Washington (1985).
I. Nimeroff, Selected NBS Papers on Colorimetry, in Precision Measurement and
Calibration, NBS Special Publication 300, Vol. 9 (1972).
I. J. Spiro, Ed., Selected Papers on Radiometry, SPIE Press, Bellingham,
Washington (1990).
Optical Design
R. E. Fischer and B. Tadic-Galeb, Optical System Design, McGraw-Hill, New
York (2000).
R. Kingslake, Lens Design Fundamentals, Academic Press, New York (1978).
R. Kingslake, Optical System Design, Academic Press, New York (1983).
P. Mouroulis and J. Macdonald, Geometrical Optics and Optical Design, Oxford
Univ. Press, New York (1997).
D. C. O'Shea, Elements of Modern Optical Design, John Wiley & Sons, New
York (1985).
R. R. Shannon, The Art and Science of Optical Design, Cambridge Univ. Press,
New York (1997). Bob was my (Palmer) professor for optical design.
G. H. Smith, Practical Computer-Aided Lens Design, Willmann-Bell, Richmond,
Virginia (1998).
W. J. Smith, Modern Optical Engineering, 3
rd
Ed., McGraw-Hill, New York
(2000).
Miscellaneous
P. C. D. Hobbs, Building Electro-Optical Systems: Making It All Work, John
Wiley & Sons, New York (2000). This publication is a pretty useful book.
C. A. Poynton, A Technical Introduction to Digital Video, John Wiley & Sons,
New York (1996).
Technical Organizations in Radiometry and Photometry
CIE
The Commission Intemationale de L'Eclairage (CIE) has numerous technical
committee reports that are relevant. Contact Thomas Lemons, TLA-Lighting
Consultants Inc., 7 Pond St., Salem, MA 01970, Tel: (508) 745-6870, Fax: (508)
741-4420, for a listing of current publications and pricing. Check into their web
page (www.cie.co.at)formoreinformation.Somepertinentreports.primarily
from Division 2, are listed in Appendix H.
354 Appendix K
CORM
The Council for Optical Radiation Measurements (CORM) was founded over
twenty years ago to promote optical radiation measurement science and
engineering and foster cooperation among the many government agencies,
industrial firms, and universities, and to formulate and transmit national needs to
NIST. They meet annually in May and publish Optical Radiation News
biannually. Contact: CORM Treasurer, 1043 Grand Ave. #312, St. Paul,
Minnesota 55105 (www.corm.org). CORM documents are listed in Appendix H.
NEWRAD
The NEWRAD conference series is an outgrowth of a meeting organized by
Peter Foukal and the papers were published by Cambridge University Press,
Massachusetts, in 1985. The proceedings of the first meeting were a private
publication. The second meeting took place at the National Physical Laboratory
in London in 1988, and the proceedings were published as New Developments
and Applications in Radiometry, N. Fox and D. Nettleton, Eds., by lOP
Publishing, London (1989). The next four meetings were held in Davos (1990),
Baltimore (1992), Berlin (1995) and Tucson (1997). The proceedings of these
meetings were published as special issues of Metrologia (Elsevier), Volumes
28(3), 30(4), 32(6), and 35(4), respectively. The seventh conference was held
October 25-27, 1999, in Madrid. Check out http://newrad.metrologia.csic.es.
NIST
The U.S. National Institute of Science and Technology (NIST, formerly National
Bureau of Standards, NBS) has a number of valuable special publications that
describe their calibration services and procedures. Appropriate ones include the
following. They may be purchased from NTIS.
NBS Measurement Services: Spectral Radiance Calibrations, J. H. Walker, R. D.
Saunders, and A. T. Hattenburg, NBS, Spec. Publ. 250-1 (1987).
NBS Measurement Services: Far Ultraviolet Detector Standards, L. R. Canfield
and N. Swanson, NBS, Spec. Publ. 250-2 (1987).
NBS Measurement Services: Radiometric Standards in the Vacuum Ultraviolet, J.
Z. Klose, J. M. Bridges, and W. R. Ott, NBS, Spec. Publ. 250-3 (1987).
NBS Measurement Services: Regular Spectral Transmittance, K. L. Eckerle, J. J.
Hsia, K. D. Mielenz, and V. R. Weidner, NBS, Spec. Publ. 250-6 (1987).
NIST Measurement Services: Spectral Reflectance, P. Y. Barnes, E. A. Early, and
A. C. Parr, NIST Spec. Publ. 250-8 (1987, revised 1997).
Photodetector Spectral Response Calibration Transfer Program, E. F. Zalewski,
NBS, Spec. Publ. 250-17, 45 (1988).
NIST Measurement Services: Photometric Calibrations, Y. Ohno, NIST, Spec.
Publ. 250-37 (1997).
-
NIST Measurement Services: Radiance Temperature Calibrations, NIST, Spec.
Publ. 250-43 (1997).
NBS Measurement Services: Spectral Irradiance Calibrations, J. H. Walker, R.
D. Saunders, J. K. Jackson, and D. A. McSparron, NBS, Spec. Publ. 250-20
(1987).
NIST Measurement Services: Spectroradiometric Detector Measurements: Parts
I and II- Ultraviolet and Visible to Near Infrared Detectors, T. C. Larason, S.
S. Bruce, and A. C. Parr, NIST, Spec. Publ. 250-41, (1998).
NIST Measurement Services: Spectroradiometric Detector Measurements: Part
III-Infrared Detectors, A. L. Migdall and G. Eppeldauer, NIST, Spec. Publ.
250-42, (1998).
SPIE
Proceedings of SPIE are published unrefereed papers as presented at conferences
of SPIE. Several conferences and resulting proceedings have been devoted to
radiometry, photometry, and closely allied topics. These Proceedings volumes
are available on SPIE' s digital library: http://spiedigitallibrary.org. They include:
196 Measurements of Optical Radiations (1979)
499 Optical Radiation Measurements I (1984)
888 Laser Beam Radiometry (1988)
1109 Optical Radiation Measurements II (1989)
2161 Photometry (1993)
2815 Optical Radiation Measurements III (1996)
7298 Infrared Technology and Applications XXXV (2009)
7299 Thermosense XXXI (2009)
7300 Infrared Imaging Systems: Design, Analysis, Modeling, and Testing
XX (2009)
7410 Optical Modeling and Measurements for Solar Energy Systems III
(2009)
7419A Infrared Detectors and Focal Plane Arrays X (2009)
Open Literature
Numerous archival and trade journals also offer significant papers on radiometry
and detection of optical radiation. They include:
Applied Optics (OSA)
IEEE Transactions on Electron Devices (IEEE)
Infrared Physics and Technology (Elsevier)
Journal of the Optical Society of America A (OSA)
356
Journal of Scientific Instruments (lOP)
Laser Focus World (PennWell Corp.)
Lightwave Technology (IEEE)
Optical Engineering (SPIE)
Photonics Spectra (free)
Review of Scientific Instruments (AlP)
Publications Available on the Internet
Appendix K
All you ever wanted to know about the SI is contained at BIPM (www.bipm.fr)
and at NIST (physics.nist.gov/cuuJ). Available publications (highly
recommended) include:
The International System of Units (SI) i
h
Ed. (1998), direct from BIPM. The
official document is in Prench; the English translation is available in PDP
format.
NIST Special Publication SP330, "The International System of Units {SI)." The
U.S. edition of the above BIPM publication. Available in PDP format.
NIST Special Publication SP81 I, "Guide for the Use of the International System
of Units {SI)." Available in PDP format.
Papers published in NIST Journal of Research since 1995 are also available on
the web in PDP format.
-
Appendix L
Reference List for
Noise and Postdetection Signal
Processing
Books and articles
Supplied by L. Stephen Bell
May 7, 2009
D. P. Blair and P. H. Sydenham, "Phase sensitive detection as a means to recover
signals buried in noise," J. Phys. E: Sci. Instrum. 8, p. 621 (1975).
E. R. Davies, Electronics, Noise and Signal Recovery, Academic Press, London
(1993).
A. De Sa, Principles of Electronic Instrumentation, Arnold, London (1981). See
Chapter 10, "Principles of signal recovery."
H. Doherty, "Techniques of low level light measurement," Lasers &
Applications, pp. 41-45, July (1983).
P. J. Fish, Electronic Noise and Low-Noise Design, McGraw-Hill, New York
( 1994).
1. Graeme, Photodiode Amplifiers, McGraw-Hill, New York (1995).
Fundamentals of photodetection circuits. Good coverage of noise-reduction
techniques for circuits and discusses effects from power supplies and external
sources.
P. Horowitz and W. Hill, The Art of Electronics, 2
nd
Ed., Cambridge, New York
(1989). One of the best single books on electronics. Look in Chapter 7 on
"Precision Circuits and Low-noise Techniques" and Chapter 15 on
"Measurements and Signal Processing," which has sections on reducing
bandwidth and spectral analysis.
G. Lawday, D. Relann, and G. Edlund, A Signal Integrity Engineer's Companion,
Prentice Hall, New Jersey (2008). Real-time test and measurement and design-
simulation techniques to systematically eliminate signal integrity problems.
357
358 Appendix L
M. L. Meade, "Advances in lock-in amplifiers," J. Phys. E: Sci. Instrum. 15, p.
395 (1982).
C. D. Motchenbacher and J. A. Connelly, Low-Noise Electronic System Design,
D. M. Munroe, "Signal-to-noise ratio improvement," Chapter 11 in Handbook of
Measurement Science, Vol. 1, P. H. Sydenham, Ed., John Wiley & Sons, New
York (1982). Includes sections on noise and noise bandwidth, signals and
SNR, preamp selection, grounding and shielding, bandwidth reduction, lock-in
amplifiers, signal averaging, correlation, and photon counting.
H. W. Ott, Noise Reduction Techniques in Electronic Systems, 2
nd
Ed, John
R. A. Pease, Ed., Analog Circuits-World Class Designs, Elsevier, Burlington
Massachusetts (2008). A top-notch, thorough coverage of analog amplifiers,
filters, and analog-to-digital converters. The filter sections cover low pass, high
pass, and bandpass, along with the operational amplifier requirements to make
the filters meet their desired specifications.
R. G. Lyons, Understanding Digital Signal Processing, Prentice Hall, New
Jersey (2004). This is another easy-to-understand introduction to DSP.
D. C. Smith, High Frequency Measurements and Noise in Electronic Circuits,
Kluwer Academic Publishers, New York (1993).
J. Williams, Ed., Analog Circuit Design-Art, Science, and Personalities,
Butterworth-Heineman, Newton, Massachusetts (1991). Per reviewer Phil
Hobbs: "Design ... is an art; there's always more than one way to do it, and the
individuality of the designer has a strong influence on the way the design turns
out: hence Art, Science, and Personalities. What it means is that for designs
that are not routine, the designer's personality has a lot to do with how it comes
out." I recommend this book to anyone who wants to become a better analogue
designer and is not easily put off by whimsy in technical writing.
J. Williams, Ed., The Art and Science of Analog Circuit Design, Butterworth-
Heineman, Woburn, Massachusetts (1998). This is Jim William's second book.
Per reviewer Phil Hobbs: "The emphasis of this volume is growing good
analog engineers, by teaching the rhythm of the insight, design, prototype,
debug iteration as practiced by the best. If you have circuits to design, this
book will pay for itself in about 5 minutes, and you'll be a more confident and
adventurous designer."
T. H. Wilmshurst, Signal Recovery from Noise in Electronic instrumentation, 2
nd
Ed, Adam Hilger, Bristol (1990).
Reference List for Noise and Postdetection Signal Processing
Software and Commentary, Downloadable Resources, and
E-books
359
Numerical Recipes: The Art of Scientific Computing, 3
rd
Ed., Cambridge Univ.
Press, New York (2007). The book includes commented full listings of more
than 400 unique C++ routines that can be downloaded in machine-readable
form for inclusion in users' programs. Information is available from
www.m.com.
Matlab software has extension packs ("toolkits") for signal processing, image
processing, and wavelets. It is available from www.mathsoft.com. Academic
and student versions are available for discounted pricing.
R. Pratap, "Getting Started with Matlab 7," A quick introduction for scientists
and engineers, Oxford Univ. Press, New York (2006).
V. Ingle and 1. Proakis, Digital Signal Processing Using Matlab, 2
nd
Ed. , CL-
Engineering (2007). This book is part of a large library of Matlab
information via: http://engineering.thomsonlearning.com. ISBN-13: 978-0-
495-07311-6.
This is the big list of books for Matlab signal processing:
http://www.mathworks.comlsupportlbooks/index _ by _ categorytitle.html?category
= 11 &sortby=title
The main Matlab repository of contributed code is found at:
http://www.mathworks.comlmatlabcentral.
Wavelab is a software library of Matlab routines for wavelet and other signal
analyses. The library is available free of charge over the Internet from Stanford
University:
MathCad software has extension packs ("toolkits") for signal processing, image
processing, and wavelets. It is available from: PTC (formerly Mathworks) :
http: //www.ptc.comlproducts/mathcadl. Academic and student versions are
available for discounted pricing.
Mathematica software also has capabilities for signal processing, image
processing, and wavelets via its "Wavelet Explorer." It is available from
www.wolfram.com. Academic (discounted) versions are available.
Stephane Mallat, A Wavelet Tour of Signal Processing, Academic Press, New
York (1999). Information is available from the site: http://www.cmap.

This is the wavelet toolbox called "LastWave," which is written in the "c"
language. Available from
http: //www.amara.comlcurrentlwavesoft.html This site also includes links to
wavelet code for use with Mathematica.
FFTW is a "c" subroutine library for computing the discrete Fourier transform
(DFT) in one or more dimensions, of arbitrary input size, and of both real and
complex data (as well as of even/odd data, i.e. the discrete cosine/sine
360 Appendix L
transforms or DCT/DST). FFTW is free (source code in "C") software
available from: www.fftw.org.
S. Smith, The Scientist and Engineer's Guide to Digital Signal Processing,
California Technical Pub. (1997). Available as a book and freely distributed
over the internet in electronic form: http://www.dspguide.com/
Other online resources for signal processing and/or FFT -related questions are
available from http://www.dsprelated.com/and from Usenet: comp.dsp (digital
signal processing), sci.math.num-analysis (numerical analysis and scientific
computing).
R Project for Statistical Computing. This is free GNU-licensed software from
http://www.r-project.org.
This website provides a peer-reviewed encyclopedia for many topics in science:
http://www .scholarpedia.org/.
A specific discussion on the phenomena of stochastic resonance regarding SNR
is available from http://www .scholarpedia.org/article/Stochastic Jesonance.
D. L. Instruments (Ithaco) App Note IAN-102, "Measuring noise spectra with
variable electronic filters," Available from the website:
http://www.dlinstruments.com/. 233 Cecil A. Malone Drive, Ithaca, New York
14850. The site has many other "Technotes" on use oflock-in amplifiers.
Index
IIf, 139
lifnoise, 143, 145, 157, 196,226,
245,246,304
2D array, 216
3-dB bandwidth, 136
3-dBfrequency, 130
absorptance, 69, 72, 77, 98
spectral, 69
absorption, 6, 61, 69, 79, 104,314
absorption coefficient, 71
accuracy, 242, 243
acousto-optical tunable filters, 239
active sources, 6
additive dispersion, 235
air mass, 119
airglow, 121
aliasing, 218
Allan variance, 245
Amici,231
amorphous silicon, 199
amplifier noise, 143
angle of observation, 32
aperture stop, 23, 24, 40, 41, 218,
220
apparent radiance, 51, 52
area array, 199
array
2D,I27
detector, 175
imagers, 199
atmosphere, 6, 28, lO4, 235, 314
atmospheric transmission, 51
loss, 49
aurora, 121
361
background, 6, 96
test chambers, 259
baffles, 227
bandpass filter, 301
simple, 303
bandwidth shrinkage factor, 300
barrier height, 187
basic
radiance, 25
radiometer, 221
Bayer algorithm, 207
bias voltage, 157
bidirectional reflectance distribution
function, 65, 73, 317
bidirectional transmittance
distribution function, 63, 73
blackbody, 77, 121
curve, 86
equation, 100
radiation, 87, 95, 138
simulator, 33, 83, 132, 226,
242,255,261
source, 49
spectral radiance, 88
BLIP, 133
blooming, 204
Bode plot, 129
bolometer, 155, 157
Boltzmann's constant, 85, l78, 181,
189,313
Bouguer, 4, 30
Butterworth RC filters, 297
calibrate, 247
a radiometer, 309
362
calibration, 102, 253, 256
configurations, 241, 256, 257
philosophy, 257
camera equation, 54, 57
carbon arc, 109
carrier lifetime, 170, 172
CCD, 89,193
noise, 142
central obscuration, 57, 260
charge injection device, 204, 208,
211
charge-coupled device, 203, 204,
208,211
charge-transfer efficiency, 142, 203
Chauvenet criteria, 247
chief ray, 23
chopper, 109,218,225,235
reflecting, 224
circular variable filter, 236
classical variance, 245
collimators, 259
color films, 201
compact fluorescent lamp, 107, 116,
123
complementary metal-oxide
semiconductor, 207
conduction band, 164, 177
cone half angle, 16, 18, 41
configuration factor, 38
cosine
3
1aw, 31, 32, 33, 58
cosine
4
law, 33-35
Crooke radiometer, 163
cross-sectional area, 20
Curie temperature, 158, 160
current responsivity, 195
cutoff
frequency, 130
wavelength, 128, 177
Czerny-Turner, 232
D*, 132, 150, 154, 162, 174, 197
photon, 133
D**, 133
D*BLlP, 174, 197
dark current, 181, 191
detection, 5
Index
detective quantum efficiency, 131
detectivity, 132
detector, 7
array, 236
D*,264
field of view, 55
noise-equivalent power, 264
silicon, 128
thermal, 128
dielectrics, 102, 123
sources, 108
diffuse, 61, 227
materials, 317
reflector, 75
surface, 65, 73, 316
diffusers, 224
dimensional analysis, 17
direct-method approach, 262
distant extended source
configuration, 260
distant small source configuration,
258
Doppler
Gaussian, 110
effect, III
dynamic resistance, 196
earth
projected area, 51
radiance, 122
reflectance, 51
effective focal length, 19
effective noise bandwidth, 136, 137,
174,218,297
Einstein diffusion constants, 190
electrical
resistance, 152
time constant, 159
electroluminescent sources, 117
electron-hole pairs, 195
electro-optical
instrument calibration, 257
Index
system, 263, 299
emission, 6, 76, 83, 104, 314
emissivity,76
emittance, 76, 98
spectral, 76, 78
emitter, 18
energy gap, 164, 166, 167
entrance pupil, 19,23,24,54,218,
222,263,266
entrance slit, 22
equation of radiative transfer, 36
basic form, 58
differential form, 46
integral form, 36
error, 241, 244
assessment, 256
etendue, 20, 229, 263
exhaust gases, 110
exit pupil, 23, 24, 54, 218
extended source, 34, 222, 223, 264
extrinsic
photoconductor, 169
semiconductor, 167
eye damage, 123
fi#,19,21,22,35,264
system parameter, 53
Fabry-Perot, 237, 238
Fastie Ebert, 232
Fermi level, 176
field of view, 6, 220, 222, 251, 263
instantaneous, 217, 318
of the detector, 55
field stop, 23, 24, 218, 220, 258
filters, 236
flat-plate solar collectors, 55
FUR devices, 315
fluorescent lamp, 115
flux density, 11
Foote's formula, 35
forward-looking infrared, 209, 318
devices, 315
Fourier transform infrared
spectrometer, 237
frame transfer, 206
architecture, 205
free spectral range, 230, 237, 239
frequency, 134
Fresnel
equation, 103
reflection losses, 71
full frame, 205
architecture, 204
gain of power, 41
gases, 103, 123
363
Gaussian, 110, 111, 112, 137, 139
distribution, 147,244
noise, 135
GE radiation calculator, 100
generation, 5
generation-recombination noise,
141, 143, 173, 197
geometrical extent, 20
glint, 319
Golay cell, 163
half-power point, 298
Havens limit, 150
H-D curve, 202
heat capacity, 141, 147
hybrid architecture, 208
illegitimate errors, 248
illuminance, 8, 57
illumination engineering, 38
image irradiance, 55
impedance, 131
incandescent light bulbs, 107
incoherent
radiation, 250
sources, 119
index of refraction, 12, 79
infrared,2
photodetector array, 209
sources, 110
364
Infrared Handbook, The 76, 78,226
instantaneous field of view, 217,
318
integrating sphere, 5,46,47,224,
261
intensity, 31, 40, 48, 266
interferometer, 237, 238
interline transfer, 206
architecture, 205
International Bureau of Weights and
Measures, 254
intrinsic
photoconductor, 169
semiconductor, 164
invariance of throughput, 21
invariant, 20
inverse square law, 37, 45, 48, 258
of irradiance, 17, 30
inversion layer, 198
irradiance, 25, 28, 30, 31,121,201,
267,316
at the detector, 42, 48
inverse square, law of, 17, 30
on the detector, 43, 171
reduction in, 57
responsivity, 129
isotropic source, 58
1- V curve, 191, 194
1- V equation, 189
Johnson noise, 137, 143, 152, 157,
162,173,182,196
Jones method, 262
Jones, R. Clark, 262, 309
Kirchhoff, 4
law, 77, 98, 99,149,314
KTC reset noise, 143
Lambert, 4
Lambert-Bouguer-Beer law, 70
Index
Lambertian, 27, 37
approximation, 33, 37, 316
disc, on-axis, 43
disc radiance, 42
source, 32, 310
sphere, 44
sphere, on-axis, 45
Langley, S. P., 155
large-area
blackbody radiation simulator,
261
imaging array, 203
laser, 118, 262
laser-power meter, 262
least-significant bit, 143
lens transmission, 41
light-emitting diode (LED), 83, 105,
117
linear
array, 216
photodiode array, 203
linearity, 131, 247
Littrow, 231, 234
long-wave infrared, 209
Lorentzian, 110, 112, 113
lossless
medium,44
optical system, 42
LOWTRAN7,314
lumens, 8
luminance, 8
luminescence, 69, 110
luminescent sources, 83
Lyot stop, 228
magnification, 54, 57
majority carriers, 187
Marcel Golay, 236
marginal ray, 24
matte reflectors, 58
maximum power transfer theorem,
156
of electrical engineering, 172
-
Index
mean
earth-sun distance, 50
square noise voltage, 135, 137
measurement, 4
measurement equation, 250
metallic sources, 108
metals, 102, 123, 156, 176
Michelson, 237
microchannel plates, 184
microphonic, 162
noise, 142
Mie scatter, 119
minority carriers, 187
concentrations, 190
model validity, 248
modulation transfer function, 202
moments normalization, 307
monochromator, 218, 229, 231, 234
moon and sun
angular subtense, 51
multiple reflections, 47
muzzle flash, 110
narrowband measurements, 217
national standard, 255
near infrared, 209
near-extended-source, 55
configuration, 261
near-smaIl-source calibration, 262
negative electron affinity material,
180
Nernst glower, 108
Nicodemus, 66
noise, 131, 134, 243
l/f, 139, 143, 145, 157, 196,
226,245,246,304
amplifier, 143
bandwidth, 132, 149
CCD,142
factor, 143, 184
Gaussian, 135
generation-recombination, 141,
142, 143, 173, 197
365
Johnson, 137, 143, 157, 162,
173,182,196
microphonic, 142
power, 298
quantization, 143,246
shot, 139, 182, 196
temperature, 144
fluctuation, 141
thermal, 135
thermal fluctuation, 162
triboelectric, 142
white, 135, 146,246, 303
noise-equivalent
flux density, 252
irradiance, 216, 252, 265
photon flux, 132
power, 154,252
temperature difference, 210,
216
normal
distribution, 245
incidence, 35, 103
numerical aperture, 19
object at infinity, 53
off-axis collimator, 260
Ohm's law, 136, 138, 173
OLEDs, 117
optical
axis, 24, 36
element, 19
radiant power, 128
radiation, 73
detectors, 127
system, 21, 23, 48
systems, 16, 19
thickness, 71, 72
transmission, 265
organic light-emitting diodes. See
OLEDs
overlapping of orders, 230
366
passive sources, 6
Pauli exclusion principle, 176
PCTRAN@, 314
Peltier
coefficient, 151
effect, 152
phosphors, 115
photocathode, 181, 185
photoconductive, 7, 164
detector, 169, 171, 173, 175
gain, 172
mode, 194
photodiode, 5
photoemission, 176
photoemissive, 7, 164
photoemissive detector, 182
photo emitter, 180
photographic film, 199,200
photometric, 2
photometry, 191, 217
photomultiplier tubes, 89, 183
photon, 128
0*, 133
detectors, 127
flux, 129
irradiance, 170
noise, 142
radiance
spectral, 90
responsivity, 129
photopic response of the eye, 217
photovoltaic, 7, 164, 191
cell, 165
detector, 175, 185, 191, 192,
195, 198
solar cell, 193
physical standards, 253
Planck, 5, 84
constant, 89,313
equation, 84, 86, 89, 91, 92, 93
expression for blackbody
radiation, 84
function, 77
plane angle, 13, 16, 17
p-n junction, 185, 187
Index
point
detector, 175
source, 29, 30, 31, 34, 220,
222,252,310
polarization, 248
sensitivity,216
states, 70
polarized, 102, 103
polished metal surfaces, 33
power
gain, 136, 298
generator, 192
responsivity,262
spectrum, 134
precision, 243
pressure (Lorentzian), 110
primary standard, 254
principle of superposition, 37
projected
area, 13,32
solid angle, 17, 18, 19, 20, 26
projection systems, 56
pure material, 62
pyrheliometer, 221
pyroelectric
coefficient, 158, 160
detector, 157, 159, 161
quantization noise, 143,246
quantum efficiency, 176, 178, 195
quantum-well infrared
photodetectors, 209
RA product, 133
radian, 13, 15
radiance, 24,26, 27,32, 52,63,314
apparent, 52
of the source, 223
responsivity, 129
spectral, 63, 87
radiant
energy, 24
exitance, 25, 26, 27, 38, 44
...
Index
flux, 61,63
intensity,25
power,24,25,29,36,39
radiation
contrast, 96, 97
geometry, 30
radiative transfer, 36, 38
basic equation, 58, 263
radiometer, 21, 215, 241
optical system, 311
radiometric
calibration, 17, 241
configuration, 6, 251
equation, 228
instruments, 215, 228, 250
measurements, 217
system, 252
radiometry, 1,24,191,211,223,
248
Raman scattering, 69
random
errors, 243
noise, 244
uncertainty,249
range equation, 251, 252, 266
ray, 11
Rayleigh, 84
scatter, 119
Rayleigh-Jeans, 93
equation, 84, 92
law, 92
RC
bandpass filter, 300
circuit, 130
low-pass filter, 299
receiver, 18
reflectance, 47, 64, 70, 315
factor, 65, 67
spectral, 63, 78
reflecting chopper, 224
reflection, 61, 63, 83
reflective natural objects, 105
refractive
collimator, 259
elements, 109
367
index, 19,85,231,239
resolving power, 229, 230, 238
responsive quantum efficiency, 131,
170
responsivity, 128, 132, 157
reverse saturation current, 189, 190,
191
reverse-bias mode, 194
Richardson equation, 181, 182
right circular cone, 18
root mean square, 130
root sum square, 249
rough aluminum, 75
scattering, 6
Schottky barrier, 198
secondary standard, 254
Seebeck
coefficient, 151
effect, 150
semiconductor
detectors, 140
photocathodes, 177
Seya-Namioka, 234
short-wave infrared, 209
shot noise, 139, 182, 196,244
signal processing, 7
signal-to-noise ratio, 131, 149,217,
244,266,297,303
silicon
amorphous, 199
detector, 128
simple bandpass filter, 303
simple equation for transfer of
radiant power, 263
single-element detectors, 127
Snell's law, 12,25
solar
cell, 192, 198,219
collectors
flat-plate, 55
constant, 28, 50
irradiance, 49
panels, 193
368
spectrum, 121
zenith angle, 119, 314
solid angle, 15-17, 25, 66, 223
field of view, 265
source image, 21
spatial resolution, 202
specification sheet, 215
spectral
absorptance, 69
absorption coefficient, 71
bandwidth, 216
directional emissivity, 98
emittance, 78, 102, 106
linewidth, 251
photon radiance, 90
radiance, 63, 87,250,313
blackbody, 88
radiometer, 215
reflectance, 63, 78
response, 201
of the eye, 8
responsivity, 128, 130,250
sensitivity,201
transmittance, 62, 251
spectrometer, 22, 217, 230, 235
spectroradiometer, 217, 228, 241
specular, 61, 227, 316
reflecting surfaces, 33
reflection, 319
reflectors, 58
surfaces, 73
speed of light, 12
sphere coatings, 47
SPRITE detector, 209
standard, 254-255
detector, 257
source, 257
statistical analysis, 243
Stefan-Boltzmann constant, 84
Stefan-Boltzmann law, 89, 95
steradian, 15
stray light, 227, 228, 260
subtractive dispersion, 235
sun
projected area, 51
sunglint, 313
sunlit scene, 52
surface
normal,13
reflectance, 31 7
systematic
error, 241, 247
uncertainty,249
systems
analysis, 211
engineering, 211
target, 96
temperature coefficient of
resistance, 155
Index
temperature fluctuation noise, 141
thermal, 7, 104, 127
conductance, 141, 149
detection of optical radiation,
147
detector, 128, 141, 147,312
emission, 313
fluctuation noise, 162
infrared, 3 19
noise, 135
resistance, 147, 148
time constant, 141, 148, 159
thermistors, 156
thermocouple radiation detectors,
153
thermoelectric
detector, 150
effect, 150
thermopile radiation detectors, 153
three-color CCDs, 207
throughput,20,229,230,239,263
time constant, 130, 154
total emissivity, 98
traceability, 254
transmission, 5, 61, 62, 79
loss, 40, 48, 53
of the lens, 42
of the optics, 223
Index
transmittance, 70, 72, 315
total,62
triboelectric noise, 142
tungsten lamps, 105,258
tungsten-filament lamp, 105-106
tungsten-halogen lamp, 106-107
Twyman-Green, 237
type A errors, 243
type B errors, 241
Ulbricht, R., 5, 46
ultraviolet, 2
detector array, 208
uncertainty, 241, 244, 253
uncooled thermal-imaging devices,
210
uniform radiance source, 47
unit solid angle, 17
unpolarized light, 70
valence band, 168, 177
vector, 11
vignette, 41
vignetting, 57
visible, 2
voltage responsivity, 160
Wadsworth, 231, 234
wavefront, 11
Welsbach mantle, 109
white
diffuse panel, 261
369
noise, 134, 135, 146,246,297,
303
Wien, 84, 93
approximation, 93, 97
displacement law, 86
work function, 176
working standard, 254
Zener effect, 194
Zodiacal light, 121
James M. Palmer (1937-2007) was
a research professor emeritus in the
College of Optical Sciences,
University of Arizona. He received
his AB in physics from Grinnell
College in 1959, and his MS and
PhD degrees in optical sciences in
1973 and 1975, respectively, from
the University of Arizona,
specializing in radiometry and
infrared systems. Prior to attending
the University of Arizona, he worked in industrial positions at Hoffman
Electronics Corporation and Centralab, Semiconductor Division of Globe Union,
Inc. Over a career spanning more than 40 years, he authored or coauthored more
than 60 technical papers on many aspects of radiometry and photometry, and he
was named Fellow of SPIE in 2003. Other awards include a NASA Group
Achievement award for his work on the Pioneer Venus Mission (1979), a Tau
Beta Pi Teacher of the Year Award (1992), and a Non-Traditional Student
Teaching Award from the University of Arizona (1993). He taught numerous
short courses at SPIE conferences, CIE meetings, and conferences of the Optical
Society of America. He served as a consultant on commercial and government
projects. Dr. Palmer was a brilliant lecturer whose former students, worldwide,
have expressed gratitude for the knowledge they gained under his tutelage.
Barbara G. Grant received her BA in mathematics
from San Jose State University in 1983, and her MS in
optical sciences from the University of Arizona in
1989, where her graduate research focused on the
absolute radiometric calibration of spaceborne imaging
sensors. She was subsequently employed at Lockheed
Missiles and Space Company, Sunnyvale, California,
where supported by excellent management, she pursued
problems in infrared sensor calibration and postflight
data analysis of electro-optical payloads. She also
worked as a NASA contractor, overseeing integration
and test of imager and sounder payloads on the GOES weather satellite, for
which she received two NASA awards. She is the author of two book-length
volumes of market research for process spectroscopy instruments. Since 1995,
her consultancy, Lines and Lights Technology, has addressed problems in
systems engineering, infrared imaging and data analysis, UV measurement, and
spectroradiometry, among other areas.

The Art of Radiometry - J.M. Palmer B. G. Grant

Hochgeladen von

Dokumentinformationen

Originalbeschreibung:

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

The Art of Radiometry - J.M. Palmer B. G. Grant

Hochgeladen von

Copyright:

Verfügbare Formate

The

Das könnte Ihnen auch gefallen