Virendra N Mahajan - Optical Imaging and Aberrations PDF

Virendra N.
Mahajan
The Aerospace Corporation

and
College of Optical Sciences, The University of Arizona
Bellingham, Washington USA

Library of Congress Cataloging-in-Publication Data
Mahajan, Virendra N.
Optical imaging and aberrations / Virendra N. Mahajan.
p. cm.
Includes bibliographical references and index.
ISBN 0-8194-4135-X
1. Aberrations. 2. Imaging systems. 3. Geometrical optics.
I. Title.
QC671.M36 1998
621.36—DC21
97-7721
CIP
Published by
SPIE
P.O. Box 10
Bellingham, Washington 98227-0010 USA
Phone: +1 360 676 3290
Fax: +1 360 647 1445
Email: books@spie.org
Web: http://spie.org
Copyright © 2001 Society of Photo-Optical Instrumentation Engineers
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.

Second printing 2004.
The author was an adjunct professor at the University of Southern California when the book was
first published. He has been an adjunct professor with the University of Arizona since 2004.
In loving memory of my parents:
mother Shrimati Sushila Devi
and father Shri Ram Chand

FOREWORD
Three years ago Vini Mahajan published a book on the subject of Geometrical
Images in the presence of aberrations. In this book, Mahajan extends this work to include
the effect of wave optics. He continues his thorough tutorial on image formation with a
detailed look at the approaches to calculating the form of images. Anyone interested in
understanding the methods of predicting the light distribution to be expected in real
imaging situations will find this book of interest.
The book begins with an exhaustive development of the basics of diffraction image
formation. Mahajan covers the issues associated with the calculation of point-spread
functions and discusses the accuracy of such calculations. He introduces the Optical
Transfer Function as the Transform of the Point Spread Function and reviews the
procedures involved in calculating the OTF. Asymptotic and approximate evaluations of
the OTF are included, as are several examples throughout the book.
These approaches are then applied to some real examples of circular and annular
apertures. In this discussion, Mahajan carries out in detail many of the classical
computations for various image descriptors. This is a topic that is generally treated only
lightly in most texts on the subject. Such issues as edge response and line spread function
and encircled power are carefully considered. A good discussion of optimal balancing of
aberrations is also provided. The treatment of aberration balancing and tolerances in
annular pupils is unique in its completeness in this book.
Mahajan includes a detailed discussion of the effect of imaging with a Gaussian

weighted aperture, a topic of much interest today. Finally the always-interesting subject
of imaging through turbulence is discussed for both circular and annular pupils.
The completeness of the discussions regarding the calculation of image structure is

thorough and detailed in this book. Mahajan includes in his discussion many of the
“classical” computations of image quality functions that form the basis of the field today,
but are seldom encountered in most texts on the subject. The student will have the
opportunity to learn the details and limitations of the process. The experienced worker
will find this volume useful as a reference in carrying out diffraction image calculations
in a wide range of optical systems.
Tucson, Arizona R. R. Shannon

April 2001
vii
TABLE OF CONTENTS
PART II. WAVE DIFFRACTION OPTICS

Preface ..............................................................................................................................xv
Acknowledgments ......................................................................................................... xvii
Symbols and Notation.................................................................................................... xix
CHAPTER 1: IMAGE FORMATION ............................................................

1.1 Introduction ............................................................................................................................ 3
1.2 Rayleigh-Sommerfeld Theory of Diffraction and Huygens-Fresnel Principle ................. 5
1.2.1 Rayleigh-Sommerfeld Formula ................................................................................ 5
1.2.2 Fresnel and Fraunhofer Approximations ................................................................. 9
1.2.3 Transfer Function of Free Space ............................................................................ 12
1.3 Gaussian Image .................................................................................................................... 12
1.4 Diffraction Image ................................................................................................................. 14
1.4.1 Pupil Function ........................................................................................................ 14
1.4.2 Diffracted Wave ..................................................................................................... 17
1.4.3 Incoherent PSF and Shift-Invariant Imaging of an Incoherent Object ................... 22
1.5 Physical Significance of PSF ............................................................................................... 24
1.6 Optical Transfer Function (OTF) ....................................................................................... 27
1.6.1 General Relations ................................................................................................... 27
1.6.2 Physical Significance of OTF ................................................................................ 31
1.6.3 Properties of OTF ................................................................................................... 33
1.6.4 OTF Slope at the Origin ......................................................................................... 35
1.6.5 OTF in the Limit of Zero Wavelength ................................................................... 40
1.6.6 Geometrical OTF ................................................................................................... 41
1.6.7 Comparison of Diffraction and Geometrical OTFs ............................................... 44
1.6.8 Determination of OTF ............................................................................................ 45
1.6.9 Significance of PTF................................................................................................ 45
1.7 Asymptotic Behavior of PSF ............................................................................................... 45
1.7.1 Point-Spread Function............................................................................................ 46
1.7.2 Encircled Power ..................................................................................................... 47
1.8 PSF Centroid ........................................................................................................................ 50
1.8.1 Centroid in Terms of OTF Slope ........................................................................... 50
1.8.2 Centroid Related to Wavefront Slope .................................................................... 51
1.8.3 Centroid Related to Wavefront Perimeter .............................................................. 52
1.9 Strehl Ratio ........................................................................................................................... 53
1.9.1 General Relations ................................................................................................... 53
1.9.2 Approximate Expressions for Strehl Ratio............................................................. 56
1.9.3 Determination of Strehl Ratio ................................................................................ 58
ix
1.10 Hopkins Ratio ....................................................................................................................... 59
1.11 Line- and Edge-Spread Functions (LSF and ESF) ........................................................... 61
1.11.1 Line-Spread Function ............................................................................................. 61
1.11.2 Edge-Spread Function ............................................................................................ 64
1.11.3 LSF and ESF in Terms of OTF .............................................................................. 64
1.12 Shift-Invariant Imaging of a Coherent Object .................................................................. 67
1.12.1 Coherent Point-Spread Function ............................................................................ 67
1.12.2 Coherent Transfer Function ................................................................................... 69
1.13 Summary of Theorems......................................................................................................... 71
Appendix A: Fourier Transform Definitions ............................................................................... 74
Appendix B: Some Frequently Used Integrals ............................................................................. 75
References ........................................................................................................................................ 76
Problems........................................................................................................................................... 78
CHAPTER 2: OPTICAL SYSTEMS WITH CIRCULAR PUPILS ...................... 79

2.1 Introduction .......................................................................................................................... 81
2.2 Aberration-Free System....................................................................................................... 82
2.2.1 Point-Spread Function............................................................................................ 82
2.2.2 Encircled Power ..................................................................................................... 87
2.2.3 Ensquared Power.................................................................................................... 88
2.2.4 Excluded Power...................................................................................................... 90
2.2.5 Optical Transfer Function....................................................................................... 93
2.2.6 PSF and Encircled Power From OTF..................................................................... 96
2.3 Strehl Ratio and Aberration Tolerance.............................................................................. 97
2.3.1 Strehl Ratio............................................................................................................. 97
2.3.2 Primary Aberrations ............................................................................................... 98
2.3.3 Balanced Primary Aberrations ............................................................................... 99
2.3.4 Comparison of Approximate and Exact Results .................................................. 101
2.3.5 Rayleigh’s h 4 Rule ........................................................................................... 102
2.3.6 Strehl Ratio for Nonoptimally Balanced Aberrations .......................................... 103
2.4 Balanced Aberrations and Zernike Circle Polynomials ................................................. 105
2.5 Defocused System ............................................................................................................... 110
2.5.1 Point-Spread Function.......................................................................................... 111
2.5.2 Focused Beam ...................................................................................................... 113
2.5.3 Collimated Beam.................................................................................................. 119
2.6 PSFs for Rotationally Symmetric Aberrations ................................................................ 121
2.6.1 Theory .................................................................................................................. 121
2.6.2 Numerical Results ................................................................................................ 124
2.6.3 Gaussian Approximation...................................................................................... 134
2.6.4 Summary of Results ............................................................................................. 135
x
2.7 Symmetry Properties of an Aberrated PSF ..................................................................... 136
2.7.1 General Theory ..................................................................................................... 137
2.7.2 Symmetry About the Gaussian Image Plane ........................................................ 138
2.7.3 Symmetry of Axial Irradiance.............................................................................. 141
2.8 PSFs for Primary Aberrations .......................................................................................... 142
2.8.1 Defocus ................................................................................................................. 142
2.8.2 Spherical Aberration Combined with Defocus..................................................... 142
2.8.3 Astigmatism Combined with Defocus ................................................................. 144
2.8.4 Coma..................................................................................................................... 148
2.8.5 2-D PSFs .............................................................................................................. 150
2.8.6 Comparison of Diffraction and Geometrical PSFs .............................................. 157
2.9 Line of Sight of an Aberrated System .............................................................................. 159
2.9.1 PSF and its Centroid ............................................................................................. 159
2.9.2.1 Wavefront Tilt ...................................................................................... 162
2.9.2.2 Primary Coma ...................................................................................... 162
2.9.2.3 Secondary Coma .................................................................................. 165
2.9.3 Comments ............................................................................................................. 168
2.10 Diffraction OTF for Primary Aberrations ....................................................................... 169
2.10.1 General Relations ................................................................................................. 169
2.10.2 Defocus ................................................................................................................. 172
2.10.3 Spherical Aberration............................................................................................. 173
2.10.4 Astigmatism ......................................................................................................... 173
2.10.5 Coma..................................................................................................................... 175
2.11 Hopkins Ratio ..................................................................................................................... 182
2.11.1 Tolerance for Primary Aberrations ...................................................................... 182
2.11.2 Defocus ................................................................................................................. 182
2.11.3 Hopkins Ratio in Terms of Variance of Aberration Difference Function............ 185
2.11.4 Variance of Aberration Difference Function for Primary Aberrations ................ 186
2.12 Geometrical OTF................................................................................................................ 187
2.12.1 General Relations ................................................................................................. 188
2.12.2 Radially Symmetric Aberrations .......................................................................... 189
2.12.3 Defocus ................................................................................................................. 189
2.12.4 Spherical Aberration Combined with Defocus..................................................... 190
2.12.5 Astigmatism Combined with Defocus ................................................................. 190
2.12.6 Coma..................................................................................................................... 191
2.13 Incoherent Line- and Edge-Spread Functions ................................................................. 191
2.13.1 Theory .................................................................................................................. 192
2.13.1.1 LSF From PSF .................................................................................. 192
2.13.1.2 LSF From Pupil Function ................................................................. 192
2.13.1.3 Struve Ratio and Aberration Tolerances ........................................... 193
2.13.1.4 LSF From OTF ................................................................................. 196
2.13.1.5 ESF From OTF ................................................................................. 198
xi
2.14 Miscellaneous Topics ......................................................................................................... 205
2.14.1 Polychromatic PSF ............................................................................................... 205
2.14.2 Polychromatic OTF .............................................................................................. 208
2.14.3 Image of an Incoherent Disc................................................................................. 209
2.14.4 Pinhole Camera .................................................................................................... 218
2.15 Coherent Imaging............................................................................................................... 222
2.15.1 Coherent Spread Function .................................................................................... 222
2.15.2 Coherent Transfer Function ................................................................................. 223
2.15.3 Coherent LSF ....................................................................................................... 224
2.15.4 Coherent ESF ....................................................................................................... 229
2.15.5 Image of a Coherent Disc ..................................................................................... 234
2.15.6 Use of a Lens for Obtaining Fourier Transforms ................................................. 238
2.15.7 Comparison of Coherent and Incoherent Imaging ............................................... 241
References ...................................................................................................................................... 253
Problems......................................................................................................................................... 257
CHAPTER 3: OPTICAL SYSTEMS WITH ANNULAR PUPILS ..................... 259

3.1 Introduction ....................................................................................................................... 261
3.2 Aberration-Free System..................................................................................................... 261
3.2.2 Encircled Power ................................................................................................... 265
3.2.3 Ensquared Power.................................................................................................. 265
3.2.4 Excluded Power.................................................................................................... 266
3.2.6 Optical Transfer Function..................................................................................... 272
3.3 Strehl Ratio and Aberration Tolerance............................................................................ 281
3.3.1 Strehl Ratio........................................................................................................... 282
3.3.2 Primary Aberrations ............................................................................................. 283
3.3.3 Balanced Primary Aberrations ............................................................................. 283
3.3.4 Comparison of Approximate and Exact Results .................................................. 284
3.4 Balanced Aberrations and Zernike Annular Polynomials ............................................. 291
3.5 Defocused System ............................................................................................................... 298
3.5.2 Focused Beam ...................................................................................................... 299
3.5.3 Collimated Beam.................................................................................................. 303
3.6 Symmetry Properties of an Aberrated PSF ..................................................................... 305
3.7 PSFs and Axial Irradiance for Primary Aberrations ..................................................... 308
3.8 2-D PSFs .............................................................................................................................. 311
xii
3.9.2.1 Wavefront Tilt ...................................................................................... 323
3.9.2.2 Primary Coma ...................................................................................... 324
3.9.2.3 Secondary Coma .................................................................................. 327
References ...................................................................................................................................... 330
Problems......................................................................................................................................... 331
CHAPTER 4: OPTICAL SYSTEMS WITH GAUSSIAN PUPILS.................... 333

4.1 Introduction ........................................................................................................................ 335
4.2 General Theory.................................................................................................................. 336
4.3 Systems with Circular Pupils ............................................................................................ 337
4.3.1 Theory .................................................................................................................. 337
4.3.2 Aberration-Free System ....................................................................................... 338
4.3.2.1 PSF ....................................................................................................... 338
4.3.2.2 OTF ...................................................................................................... 341
4.3.3 Strehl Ratio and Aberration Tolerance ................................................................. 343
4.3.4 Balanced Aberrations and Zernike-Gauss Circle Polynomials ............................ 344
4.3.5 Defocused System ................................................................................................ 348
4.3.5.1 Theory .................................................................................................. 348
4.3.5.2 Axial Irradiance .................................................................................... 349
4.3.5.3 Defocused Distribution......................................................................... 350
4.3.5.4 Collimated Beam.................................................................................. 352
4.3.6 Weakly Truncated Gaussian Circular Beams....................................................... 353
4.3.6.1 Introduction .......................................................................................... 353
4.3.6.2 Irradiance Distribution and Beam Radius ............................................ 354
4.3.6.3 Imaging of a Gaussian Beam................................................................ 359
4.3.6.4 Aberration Balancing ........................................................................... 362
4.3.7 Symmetry Properties of an Aberrated PSF ......................................................... 365
4.4 Systems with Annular Pupils............................................................................................. 366
4.4.1 Theory .................................................................................................................. 367
4.4.2 Aberration-Free System ....................................................................................... 368
4.4.3 Strehl Ratio and Aberration Tolerance ................................................................. 370
4.4.4 Balanced Aberrations and Zernike-Gauss Annular Polynomials ......................... 371
4.4.5 Defocused System ................................................................................................ 374
4.4.5.1 Theory .................................................................................................. 374
4.4.5.2 Axial Irradiance .................................................................................... 374
4.4.5.3 Defocused Distribution......................................................................... 376
4.4.5.4 Collimated Beam.................................................................................. 376
4.4.6 Symmetry Properties of an Aberrated PSF .......................................................... 378
xiii
4.5.2.1 Wavefront Tilt ...................................................................................... 380
4.5.2.2 Primary Coma ...................................................................................... 381
4.5.2.3 Secondary Coma .................................................................................. 382
4.6 Summary ............................................................................................................................. 382
References ...................................................................................................................................... 385
Problems......................................................................................................................................... 386
CHAPTER 5: RANDOM ABERRATIONS................................................. 387

5.1 Introduction ........................................................................................................................ 389
5.2 Random Image Motion.......................................................................................................389
5.2.1 General Theory ..................................................................................................... 390
5.2.2 Circular Pupils...................................................................................................... 391
5.2.2.1 Theory .................................................................................................. 391
5.2.2.2 Gaussian Approximation...................................................................... 392
5.2.2.3 Numerical Results ................................................................................ 393
5.2.3 Annular Pupils...................................................................................................... 393
5.2.3.1 Theory .................................................................................................. 393
5.2.3.2 Numerical results ................................................................................. 397
5.3 Imaging Through Atmospheric Turbulence .................................................................... 401
5.3.1 Long-Exposure Image .......................................................................................... 402
5.3.2 Kolmogrov Turbulence ........................................................................................ 407
5.3.3 Circular Pupils...................................................................................................... 413
5.3.4 Annular Pupils...................................................................................................... 417
5.3.5 Phase Aberration in Terms of Zernike Polynomials ............................................ 421
5.3.6 Short Exposure Image .......................................................................................... 430
5.3.6.1 Near-Field Imaging .............................................................................. 430
5.3.6.2 Far-Field Imaging ................................................................................. 436
5.3.6.3 Short-Exposure Images ........................................................................ 439
5.3.7 Adaptive Optics .................................................................................................... 441
Appendix: Fourier Transform of Zernike Polynomials ........................................................... 443
References ...................................................................................................................................... 445
Problems......................................................................................................................................... 447
BIBLIOGRAPHY ......................................................................................................................... 449

REFERENCES FOR ADDITIONAL READING ..................................................................... 451
INDEX ............................................................................................................................................ 459
xiv
PREFACE
In Part I of this book, we discussed imaging based on ray geometrical optics. The
aberration-free image of an object according to it is the exact replica of the object, except
for its magnification. The aberration-free image of a point object is also a point. In reality,
however, the image obtained is not a point. Because of diffraction of the object wave at
the aperture stop, or equivalently, the exit pupil of the imaging system, the actual
aberration-free image for a circular exit pupil is a light patch surrounded by dark and light
rings. Geometrical optics is still assumed to hold from the object to the exit pupil in that
rays are traced through the system to determine the shape of the pupil and the aberration
across it. The determination of the characteristics of the diffraction image of an object
formed by an aberrated system is the subject discussed in Part II. The emphasis of this
part is on the numerical results on the effects of aberrations on a diffraction image and not
on the formalism, exposition, or a critique of the variety of diffraction theories proposed
over the years. It is a compilation of the works of masters like Nijboer, Hopkins, Barakat,
and Fried with a sprinkling of my own work.
In Chapter 1, the diffraction theory of image formation is discussed. Starting with a

brief account of the Rayleigh-Sommerfeld theory from a Fourier-transform standpoint,
we derive the Huygens-Fresnel principle from it. We show that the diffraction image of a
point object, called the point-spread function (PSF), is proportional to the modulus square
of the Fourier transform of the complex amplitude across the exit pupil. It is shown that
the diffraction image of an isoplanatic incoherent object is equal to the convolution of its
Gaussian image and the diffraction PSF. The optical transfer function (OTF) of an
imaging system, which is the Fourier transform of its PSF, is also discussed. The spatial
frequency spectrum of the image of an isoplanatic object is shown to be equal to the
product of the spectrum of the object and the OTF. The OTF based on geometrical optics
is also considered and an approximate expression valid for low spatial frequencies is
given. A brief comparison of imaging based on diffraction and geometrical optics is given
in terms of both the PSF and the OTF. The line of sight of a system is identified with the
centroid of its PSF which, in turn, is obtained in terms of the OTF or the aberration
function. The line- and edge-spread functions are introduced and obtained in terms of the
PSF or the OTF. Expressions for Strehl, Hopkins, and Struve ratios, useful for obtaining
aberration tolerances, are derived. Finally, imaging of coherently illuminated objects is
discussed. It is shown that the image of an isoplanatic coherent object is equal to the
convolution of its Gaussian amplitude image and the coherent PSF of the imaging system.
This chapter forms the foundation of Part II in that the fundamental relations derived in it,
and stated as theorems, are used in the succeeding chapters to obtain some practical
results for imaging systems with circular, annular, and Gaussian pupils. However, a
reader need not read all of this chapter before attempting to read others.
Chapter 2 on systems with circular pupils starts with the aberration-free PSF and its
encircled and ensquared powers. The effects of primary aberrations on its Strehl ratio are
discussed and aberration tolerances are obtained. Balanced primary aberrations are
xv
considered and identified with Zernike circle polynomials. Focused and collimated beams
are discussed and the concept of near- and far-field distances is introduced. Aberrated
PSFs and their symmetry properties are discussed, and a brief comparison is made with
their counterparts in geometrical optics. The line of sight of an aberrated system,
identified with the centroid of the PSF, is determined for primary aberrations. The OTF
for these aberrations is also discussed, phase contrast reversal is explained, and aberration
tolerances for a certain value of Hopkins ratio are given. Expressions for the geometrical
OTF for primary aberrations are also given. Both incoherent and coherent line- and edge-
spread functions are discussed, and aberration tolerances for a certain value of the Struve
ratio are given. A brief comparison of incoherent and coherent imaging is given with
special reference to the Rayleigh criterion of resolution. The Fourier-transforming
property of wave propagation is illustrated in altering the image of an object by spatial
filtering in the Fourier-transform plane.
Systems with annular pupils are given a cursory look at best in books where imaging
is discussed. Our Chapter 3 is written in a manner similar to Chapter 2 where, for
example, the balanced aberrations are identified with the Zernike annular polynomials.
Although the propagation of Gaussian beams is discussed in books on lasers, their
treatment is generally limited to the weakly truncated aberration-free beams. In Chapter
4, we consider the effect of arbitrary truncation of beams with and without aberrations.
The balanced aberrations in this case are identified with the Zernike-Gauss polynomials.
It is shown, for example, that the pupil radius must be at least three times the beam radius
in order to neglect beam truncation without significant error.
Finally, random aberrations are considered in Chapter 5. The effect of random image
motion is considered first, and expressions and numerical values of time-averaged Strehl
ratio, PSF, and encircled power are given for systems with circular and annular pupils.
The random aberrations introduced by atmospheric turbulence when a wave propagates
through it, as in astronomical observations by a ground-based telescope, are discussed;
and expressions for time-averaged PSF and OTF are obtained. The aberration function for
Kolmogorov turbulence is expanded in terms of the Zernike polynomials, and auto-
correlation and cross-correlation of the expansion coefficients are given. The atmospheric
coherence length is defined, and it is shown that the resolution of a telescope cannot
exceed that of an aberration-free telescope of this diameter. Both the short- and long-
exposure images are considered.
As in Part I, each chapter ends with a set of problems. It is hoped that they will
acquaint the reader with application of the theory in terms of some practical examples.
References for addtional reading are given after the Bibliography. These references
represent the author’s collection as the editor of Milestone Series 74 entitled Effects of
Aberrations in Optical Imaging, published by SPIE Press in 1993.
xvi
ACKNOWLEDGMENTS
It is a great pleasure to acknowledge the generous support I have received over the
years from my employer, The Aerospace Corporation, in preparing this book. My special
thanks go to the senior vice president Mr. John Parsons, for his continuous interest and
encouragement in this endeavor. I thank Mr. John Hoyem for preparing the figures and
Mr. Victor Onouye for some figures as well as the final composition. My thanks also go
to Dr. Rich Boucher for computer generating the 2-D PSFs and Mr. Yunsong Huang for
numerical analysis and computer plotting. The Sanskrit verse on p. was provided by
Professor Sally Sutherland of the University of California at Berkeley.
I am grateful to Dr. P. Mouroulis and Professor Emeritus D. Schroeder for their

careful review of the manuscript. I also benefited from the comments made by many
friends and colleagues over the years. Included among them are Dr. Bill Swantner,
Professor A. Walther, Professor A. Marathay, and Professor J. Harvey. Of course, any
shortcomings or errors in the book are my responsibility. I also thank Professor R. R.
Shannon for writing the Foreword to both Parts of this book.
I thank my sister Pushpa and brother Devinder for encouraging me to study phyiscs. I
can not say enough about the constant support I have received from my wife Shashi over
the many years it has taken me to complete this two-part book. I dedicate this book to my
departed parents who brought and nourished me in this world.
Once again, it has been a pleasure to work with the SPIE Press staff in bringing this
book to its completion. My thanks to them, especially, Sharon Streams, for their
cooperative spirit and quality support.
El Segundo, California Virendra N. Mahajan

April 2001
xvii
SYMBOLS AND NOTATION
a pupil radius r0 atmospheric coherence diameter

r
ai aberration coefficient ri image point position vector
r
A amplitude rp pupil point position vector
Ai peak aberration coefficient R radius of reference sphere
Bd defocus coefficient Re real part
Bt tilt coefficient S Strehl ratio
D pupil diameter Sen area of entrance pupil
Ᏸ structure function U complex amplitude
F focal ratio Sex area of exit pupil
ESF edge transfer function W wave aberration
H Hopkins ratio x, y rectangular coordinates of a point
I irradiance z optical axis, axial distance
Im imaginary part Z nm (r, q) Zernike circle polynomial
k wavenumber
r image spatial frequency vector
vi
l distance, log amplitude
r object spatial frequency vector
vo
LSF line spread function v normalized spatial frequency
mp pupil magnification o optical transfer function
M magnification ^ phase transfer function
MTF modulation transfer function e polar angle of a position vector
OTF optical transfer function q polar angle of frequency vector
P object point ⑀ obscuration ratio
Pv Gaussian image point b(u) Dirac delta function
Pex power in the exit pupil d transmission factor
Pi image power \ phase aberration
P(u) pupil function r, e polar coordinates of a point
PSF point-spread function h optical wavelength
PTF phase transfer function j, d normalized frequency coordinates
r radial coordinate mW standard deviation of wave aberration
rc radius of circle Rnm (l) Zernike radial polynomial
xix
Anantaratnaprabhavasya yasya himam
. na saubhagyavilopi jatam
Eko hi doso
. gunasannipate
. ˙ .
nimajjatindoh. kiranesvivankah
.
The snow does not diminish the beauty of the Himalayan mountains
which are the source of countless gems. Indeed, one flaw is lost
among a host of virtues, as the moon’s dark spot is lost among its rays.
Kalidasa Kumarasambhava 1.3

CHAPTER 1
IMAGE FORMATION
1.1 Introduction ..............................................................................................................3

1.2 Rayleigh-Sommerfeld Theory of Diffraction and Huygens-Fresnel Principle... 5
1.2.1 Rayleigh-Sommerfeld Formula ................................................................... 5
1.2.2 Fresnel and Fraunhofer Approximations ..................................................... 9
1.2.3 Transfer Function of Free Space................................................................12
1.3 Gaussian Image ......................................................................................................12
1.4 Diffraction Image ................................................................................................... 14
1.4.1 Pupil Function............................................................................................14
1.4.2 Diffracted Wave ........................................................................................17
1.4.3 Incoherent PSF and Shift-Invariant Imaging of an Incoherent Object ......22
1.5 Physical Significance of PSF ................................................................................. 24
1.6 Optical Transfer Function (OTF)......................................................................... 27
1.6.1 General Relations ......................................................................................27
1.6.2 Physical Significance of OTF ....................................................................31
1.6.3 Properties of OTF ......................................................................................33
1.6.4 OTF Slope at the Origin ............................................................................35
1.6.5 OTF in the Limit of Zero Wavelength ......................................................40
1.6.6 Geometrical OTF ....................................................................................... 41
1.6.7 Comparison of Diffraction and Geometrical OTFs ................................... 44
1.6.8 Determination of OTF ............................................................................... 45
1.6.9 Significance of PTF ................................................................................... 45
1.7 Asymptotic Behavior of PSF ................................................................................. 45
1.7.1 Point-Spread Function ............................................................................... 46
1.7.2 Encircled Power......................................................................................... 47
1.8 PSF Centroid ..........................................................................................................50
1.8.1 Centroid in Terms of OTF Slope ............................................................... 50
1.8.2 Centroid Related to Wavefront Slope ........................................................51
1.8.3 Centroid Related to Wavefront Perimeter ................................................. 52
1.9 Strehl Ratio ............................................................................................................. 53
1.9.1 General Relations ......................................................................................53
1.9.2 Approximate Expressions for Strehl Ratio ................................................56
1.9.3 Determination of Strehl Ratio....................................................................58
1.10 Hopkins Ratio ......................................................................................................... 59
1.11 Line- and Edge-Spread Functions (LSF and ESF) ............................................. 61
1.11.1 Line-Spread Function ................................................................................61
1.11.2 Edge-Spread Function ............................................................................... 64
1.11.3 LSF and ESF in Terms of OTF..................................................................65
1
2 IMAGE FORMATION
1.12 Shift-Invariant Imaging of a Coherent Object ....................................................67

1.12.1 Coherent Point-Spread Function................................................................67
1.12.2 Coherent Transfer Function ......................................................................70
1.13 Summary of Theorems ..........................................................................................71
Appendix A: Fourier Transform Definitions ................................................................74
Appendix B: Some Frequently Used Integrals..............................................................75
References ........................................................................................................................76
Problems ........................................................................................................................... 78
Chapter 1
Image Formation
1.1 INTRODUCTION
In Part I of Optical Imaging and Aberrations, 1 we showed how to determine the
location and size of the image of an object formed by an imaging system in terms of the
location and size of the object and certain parameters of the system. We discussed the
relationship between the irradiance distribution of the image and the radiance distribution
of the object, including the cosine-to-the-fourth-power dependence on the field angle and
vignetting of the rays by the system. We discussed the ray distribution of the aberrated
image of a point object, called the geometrical point-spread function (PSF) or the spot
diagram. We showed how to design and analyze imaging systems in terms of their
primary aberrations. We also showed how to calculate the primary aberrations of a
multisurface optical system in terms of their radii of curvature, spacings between them,
and the refractive indices associated with those spacings. We pointed out that the image
obtained in practice differs from that predicted by geometrical optics. For example, when
the system is aberration free, all of the rays from a point object transmitted by the system
converge to its Gaussian image point according to geometrical optics. In reality, however,
the image obtained is not a point. Because of diffraction of object radiation at the aperture
stop of the system or, equivalently, at its exit pupil, the actual aberration-free image for a
circular exit pupil is a light patch surrounded by dark and bright diffraction rings. The
determination of the characteristics of the diffraction image of an object formed by an
aberrated system is the subject of Part II.
In this chapter, we first describe the diffraction theory of image formation of

incoherent objects, i.e., objects for which the radiation from one of its parts is incoherent
with that from another. We start with a brief account of Rayleigh-Sommerfeld theory of
diffraction from a Fourier transform standpoint and derive the Huygens-Fresnel principle
from it. First, we consider the Gaussian image of an object, i.e., the aberration-free image
according to geometrical optics, and show that it is an exact replica of the object scaled
by its (transverse) magnification. Next, we consider the diffraction image based on the
Huygens-Fresnel principle. We introduce the concept of a diffraction PSF, i.e., the
diffraction image of a point object, and show that it is proportional to the modulus square
of the Fourier transform of the complex amplitude across the exit pupil, called the pupil
function, of the system.
Geometrical optics is assumed to hold from the point object to the exit pupil in that
rays are traced through the system to determine the shape of the pupil and the aberration
across it for the point object under consideration, as discussed in Section 3.2 of Part I.
The amplitude associated with a ray is obtained by consideration of the transmission and
reflection characteristics of the imaging elements of the system. In many applications, the
amplitude variation across the pupil is negligible. There are instances, however, when the
transmission across the pupil is varied to alter certain characteristics of the diffraction
3
4 IMAGE FORMATION
PSF. The pupil is said to be apodized in these cases. Another important example of
variable amplitude across the pupil is that of laser beams for which the amplitude has a
Gaussian form. The irradiance distribution of the diffraction image of an incoherent
object is obtained by adding the irradiance distribution of the images of its infinitesimal
elements. In particular, we show that the diffraction image of an isoplanatic object, i.e.,
one for which the pupil function of the system is independent of the position of a point on
it, is equal to the convolution of its Gaussian image and the (diffraction) PSF of the
system.
We also discuss the optical transfer function of an imaging system, which is equal to
a Fourier transform of its PSF. It describes the amplitude and phase of the sinusoidal
image of a sinusoidal object as a function of its spatial frequency. Thus, it relates the
spatial frequency spectrum of the object to that of the image. We show, for example, that
the slope of its real part at the origin is independent of aberrations. We also compare the
diffraction OTF with the corresponding OTF based on geometrical optics. The asymptotic
behavior and centroid of an aberrated point-spread function are related to the slope of the
real and imaginary parts of the OTF at the origin, respectively. It is shown that the
centroid of the diffraction PSF is identical with that of the geometrical PSF.
Next, image quality based on the Strehl ratio of an aberrated imaging system is
discussed, and approximate expressions in terms of its aberration variance are obtained. A
brief discussion of how the Strehl ratio of a system may be determined is also given. The
image quality based on Hopkins ratio in the spatial frequency domain is also discussed.
Whereas the Strehl ratio represents the ratio of the central value of the irradiance
distribution of the aberrated image of a point object and the corresponding aberration-free
value, the Hopkins ratio represents the ratio of the contrasts of the aberrated and
aberration-free images of a sinusoidal object of a certain spatial frequency. The Strehl and
Hopkins ratios provide simple means of determining the aberration tolerances. The line-
spread function (i.e, the image of a line object) and the edge-spread function (i.e., the
image of an edge or a step object) are discussed in terms of the PSF and related to the
OTF. The Struve ratio of a system, representing the ratio of the aberrated central value of
the LSF and the corresponding aberration-free value, is also considered and aberration
tolerances are derived from it.
Finally, imaging of a coherently illuminated object is discussed. The complex

amplitude distribution of the image is obtained by adding the complex amplitude
distributions of the images of its infinitesimal elements. We show that the coherent point-
spread function of an imaging system is proportional to the inverse Fourier transform of
its pupil function, and its coherent transfer function is proportional to its pupil function.
Accordingly, the complex amplitude image of a coherent isoplanatic object is equal to the
convolution of its Gaussian amplitude image and its coherent spread function; or the
spatial frequency spectrum of the amplitude image is given by the product of the
spectrum of the Gaussian amplitude image and the coherent transfer function. The
incoherent PSF of an imaging system is proportional to the modulus square of its
1.2 Rayleigh-Sommerfeld Theory of Diffraction and Huygens-Fresnel Principle 5
coherent spread function, and its incoherent transfer function is equal to the
autocorrelation of its coherent transfer function.
The fundamental relations derived in this chapter are stated as theorems and used in
the succeeding chapters to obtain some practical results for imaging systems with
circular, annular, and Gaussian pupils.
1.2 RAYLEIGH-SOMMERFELD THEORY OF DIFFRACTION AND THE

HUYGENS-FRESNEL PRINCIPLE
Many scalar (as opposed to vector) theories of diffraction have been proposed over
the years to explain what is and might be observed when an optical system forms the
image of an object.2 We discuss only one of them, namely, the Rayleigh-Sommerfeld
theory,3 and obtain the Fresnel-Huygens principle from it. The concept of aberrations of a
diffracted wave is introduced, their Fresnel and Fraunhofer approximations are discussed,
and the transfer function of free space is obtained. This theory is adequate for what we
need to discuss in terms of the diffraction theory of image formation, including the
effects of aberrations of an imaging system. The following two theorems are derived in
this section.
Theorem 1. As a wave propagates, its disturbance according to the Huygens-

Fresnel principle is given by the superposition of secondary spherical wavelets weighted
by the amplitudes at the points where they originate on the wave.
Theorem 2. Under certain approximations, the propagation of a wave is described

by a Fourier transform of its complex amplitude modified by a quadratic phase factor in
the Fresnel approximation, or without modification in the Fraunhofer approximation.
1.2.1 Rayleigh-Sommerfeld Formula

Consider an optical wave of wavelength and wavenumber k = 2 propagating
in the z direction, as illustrated in Figure 1-1. Suppose we are given the complex
r r
amplitude U ( r ; 0) of the wave in the plane z = 0 , where r = ( x, y) is the 2D position
r
vector of a point in a plane defined by the z value. The complex amplitude U ( r ; z ) in a
plane at a distance z can be determined by solving the Helmholtz scalar wave equation for
free-space propagation, namely,
r
( 2
)
+ k 2 U (r ; z) = 0 , (1-1)
where 2 is the 3D Laplace operator, i.e.,
2 2 2
2 = 2 + 2 + . (1-2)
x y z 2
r r
In Figure 1-1, r is used in the z = 0 plane to distinguish it from the r in the z plane, as
the two planes become the source and observation planes, respectively.
6 IMAGE FORMATION
x
x
P (r; 0)
l
P (r; z)

r
l0
O
y r
z = 0 Plane
z
O z
y z - Plane
Figure 1-1. Geometry of wave propagation for determining the complex amplitude
r r
U ( r ; z ) in a z plane from its knowledge U ( r ; 0) in the z = 0 plane. O P = l 0 ,
P P = l and O O = z .
r r
To relate U ( r ; z ) to U ( r ; 0) , we decompose both of them into plane waves. Thus,
for example, we write
r r r r r
U ( r ; 0) = A( v ; 0) exp (
2 ir v ) d v , (1-3)
r
where A( v ; 0) is the amplitude of a plane wave propagating with direction cosines
r
(, , ) such that the spatial frequency v is given by
r 1
v = ( , ) (1-4)

and
2 + 2 + 2 = 1 . (1-5)
r r
r(
A v ; 0) is also referred to as the spectral component of U ( r ; 0) with a spatial frequency
v . Written in the form A( , ; 0) , it is also called the angular spectrum of
r
U ( r ; 0) . [Equation (1-4) may be obtained by comparing
r r the exponentialrin Eq. (1-3) with
[(
the standard form of a plane wave exp i t
k r , where and k are its angular
frequency and propagation vector, and t is time.]
)]
r r
Multiplying both sides of Eq. (1-3) by exp (2ir v ) and integrating, we obtain
r r r r r r r r r r

U ( r ; 0) exp (2 ir v ) dr = A( v ; 0) d v exp 2 ir ( v
v ) dr [ ]
r r r r
= A( v ; 0) ( v
v ) d v
r
= A( v ; 0) ,
or
r r r r r
A( v ; 0) = U ( r ; 0) exp (2 ir v ) dr , (1-6)
r
where () is a Dirac delta function. It is clear from Eqs. (1-3) and (1-6) that U ( r ; 0) and
r r
A( v ; 0) form a 2-D Fourier-transform pair; U ( r ; 0) is the inverse Fourier transform of
r r r
A( v ; 0) and A( v ; 0) is the Fourier transform of U ( r ; 0) . (For a definition of the Fourier
transform, see the Appendix.) Similarly, we may write
r r r r r
U ( r ; z ) = A( v ; z ) exp (
2ir v ) d v , (1-7)
r r
where A( v ; z ) is the amplitude of a plane-wave component of spatial frequency v in the
z plane given by
r r r r r
A( v ; z ) = U ( r ; z ) exp (2ir v ) dr . (1-8)
r
Since U ( r ; z ) satisfies the wave equation, which is linear, each component of the
plane waves that comprise it must also satisfy it. Hence, substituting
r r r
A( v ; z ) exp (
2ir v ) for it into Eq. (1-1), we obtain
r
r 2 A( v ; z )
k A( v ; z ) +
2 2
= 0 . (1-9)
z 2
Solving Eq. (1-9), we find that

r r
A( v ; z ) = A( v ; 0) exp (ik z ) , (1-10)
showing how the amplitude of a plane wave component changes as it propagates. We

note that only its phase changes. Substituting Eq. (1-10) into Eq. (1-7), we obtain
r r r r r
U ( r ; z ) = A( v ; 0) exp (ik z ) exp (
2 ir v ) d v . (1-11)
r
Let h( r ; z ) be the inverse Fourier transform of exp (ik z ) , i.e.,
r r r r
h( r ; z ) = exp (ik z ) exp (
2 ir v ) d v , (1-12)
so that
r r r r
exp (ik z ) = h( r ; z ) exp (2 ir v ) dr . (1-13)
Substituting Eq. (1-13) into Eq. (1-11), we obtain

r r r r r r r r
[
U ( r ; z ) = h( r ; z ) dr A( v ; 0) exp
2i ( r
r ) v d v]
r r r r
= h( r ; z ) U ( r
r ; 0) dr , (1-14)
where we have used Eq. (1-3). By change of variables, Eq. (1-14) containing the
convolution integral can also be written
r r r r r
U ( r ; z ) = h( r
r ; z ) U ( r ; 0) dr . (1-15)
r
The integrals in Eqs. (1-14) and (1-15) giving the wave field U ( r ; z ) atthe z plane
8 IMAGE FORMATION
r r
represent a convolution of h( r ; z ) and the wave field U ( r ; 0) in the z = 0 plane.
r
If a point source of unit amplitude is placed at a point ro in the z = 0 plane, i.e., if
r r r
U ( r ; 0) = ( r
ro ) , (1-16)
r
then Eq. (1-15) shows that the complex amplitude at a point r in the z plane is given by
r r r
U ( r ; z ) = h( r
ro ; z ) . (1-17)
r r
Thus, h( r ; z ) represents the complex amplitude at a point r in the z plane due to a point
r
source of unit amplitude located at r = 0 in the z = 0 plane. It is called the complex
amplitude point-spread function (or impulse response) of free space. Carrying out the
integration in Eq. (1-12), it can be shown that4
r 1 1 z exp (ikl0 )
h( r ; z ) =
i , (1-18)
kl0 l0 l0
where
( )
1/ 2
l0 = z 2 + r 2 (1-19)
is the distance OP between the origin in the z = 0 plane and the observation point
r
( r ; z) . Equation (1-18) is a mathematical expression for a Huygens’ spherical wavelet
diverging from the point source, and it describes the complex amplitude point-spread
r r
function of free space. We note from Eq. (1-12) that h( r ; z ) (r ) as z 0 , i.e., it
becomes the point source, as expected. Accordingly, Eq. (1-15) is a mathematical
description of Huygens-Fresnel principle, namely, that the complex amplitude in the z
r r
plane is a linear superposition of Huygens’ secondary spherical wavelets h( r
r ; z )
r
weighted by the amplitudes U ( r ; 0) of the wave where they originate (Theorem 1). The
diffracted wave field described by Eq. (1-15) is shift invariant (or isoplanatic) in that a
r r r r
spherical wavelet at a point r due to a source point at r depends on r
r , i.e., the
form of a spherical wavelet is independent of the location of its origin in the z = 0 plane,
except for a shift in the center of the distribution.
r 1 r 1 z exp (ikl ) drr ,

U (r ; z) = U ( r ; 0)
i (1-20)
kl l l
where (see Figure 1-1)
[ r r 2
]
12
l = z2 + r
r (1-21a)
r r 1 r r 4
= z+
1 2
2z
( )
r + r 2
2 r r
3 r
r + ...
8z
(1-21b)
r r
is the distance P P between a source point (r ; 0) and the observation point (r ; z ) .
Equation (1-20) is the Rayleigh-Sommerfeld formula describing propagation of a wave
from one plane to another.
1.2.2 Fresnel and Fraunhofer Approximations

For large values of z, kl >> 1, l ~ z (so that the obliquity factor z l , representing the
cosine of the angle P P makes with the z axis, may be assumed to be unity) except in the
exponent where we retain additional terms according to Eq. (1-21b) to within a fraction
of a wavelength (since they are multiplied by k). Hence, Eq. (1-20) may be written
r
i r2 ikr 2 r 2 i r r r
U (r ; z) = exp ik z + exp U ( r ; 0) exp
z r r dr ,
z 2 z 2 z
r r 4
for z 3 >> k r
r max 8 ( Fresnel) , (1-22a)
and
r
i r2 r 2 i r r r
U (r ; z) = exp ik z + U ( r ; 0) exp
r r dr ,
z 2 z z
for z >> krmax

2 2 ( Fraunhofer ) . (1-23a)
Thus, depending on the value of the distance z relative to the extent of the regions of the
source field and observation, the complex amplitude in a z plane is proportional to the
inverse Fourier transform of the complex amplitude in the z = 0 plane with or without
modification by a quadratic phase factor kr 2 / 2 z (Theorem 2). Equations (1-22a) and
(1-23a) represent diffraction integrals in the Fresnel and Fraunhofer approximations,
respectively. It should be noted that the Fresnel condition of large distance z in Eq. (1-
22a) is a sufficient but not a necessary condition. What is necessary is that the sum of the
neglected terms be small so that their contribution to the diffraction integral is negligible.
The integrals in Eqs. (1-22a) and (1-23a) are referred to as representing the Fresnel
r
and Fraunhofer diffraction patterns of the distribution U ( r ; 0) . The region of space
satisfying the condition z >> krmax
2 / 2 is called the Fraunhofer or the far-field region of
diffraction. The condition itself is called the Fraunhofer or the far-field condition. The
r r 4
region of space satisfying the Fresnel condition z 3 >> k r
r max / 8 but not the
Fraunhofer condition is often referred to as the region of Fresnel or near-field diffraction.
The region of very small z values is referred to as the Rayleigh-Sommerfeld region of
diffraction. It should be evident, though, that the Rayleigh-Sommerfeld integral in Eq. (1-
20) will yield accurate results (within the range of its validity) regardless of the value of
z. Similarly, the Fresnel integral will yield accurate results in the Fraunhofer region as
well. However, calculations in the Fraunhofer region are simpler and more common in
imaging applications (due to cancellation of the quadratic phase factor of free-space
propagation by the focusing quadratic phase factor provided by the imaging system).
10 IMAGE FORMATION
Once z is large enough to satisfy the Fraunhofer condition, a larger z value changes only
the scale of the irradiance distribution without changing the distribution.
A better approximation of the Rayleigh-Sommerfeld diffraction integral applicable

for large angles of diffraction may be obtained by expanding l in terms of l0 instead of z,
i.e., by writing
r r
r r
( )
2
r 2 r r r 2
r r
l = l0 +

+ ... . (1-21c)
2l0 l0 8l0 3
Accordingly, Eqs. (1-22a) and (1-23a) representing the Fresnel and Fraunhofer
approximations are replaced by
r
i z exp(ikl0 ) ikr 2 r 2 i r r r
U (r ; z) = exp U ( r ; 0) exp
r r dr (1-22b)
l0 l0 2l0 l0
and
r
i z exp(ikl0 ) r 2 i r r r
U (r ; z) = U ( r ; 0) exp
l r r dr . (1-23b)
l0 l0 0
It should be evident that exp(ikl0 ) l0 represents a spherical wave originating at the origin
r
of the z = 0 plane, and ( r ; z ) l0 represents the direction cosines of the position vector of
the observation point. Since l0 varies with the point of observation, a uniform array of
points in the observation plane becomes nonuniform when divided by l0 , thereby making
the numerical calculations of the Fourier transform cumbersome. In Eq. (1-22b), there is
an additional difficulty due to the quadratic phase factor under the integral. These
difficulties may be overcome by making observations on a hemisphere of radius l0 , since
the points in that case are uniformly distributed in the direction cosine space. The
obliquity factor z l0 reduces to unity in that case. In imaging, however, the observations
are made in a plane. Also, the fields of view of high-quality imaging systems are
generally small, and Eqs. (1-22a) and (1-23a) representing diffraction at small angles
suffice. Moreover, imaging systems generally have aberrations that increase rapidly with
the field of view and impact the diffracted field a lot more significantly than any error
made in assuming small angles of diffraction.
To get an idea of the distances of the Fresnel and Fraunhofer regions from a
diffracting aperture, we consider a circular aperture with a diameter of 1 cm (so that
rmax = 0.5 cm ) illuminated by a collimated beam of light with a wavelength of 0.5 mm.
r4
The Fresnel condition z 3 > k rmax
/ 8 for on-axis diffraction is satisfied for z > 10 cm. Of
course, the larger the distance, the better the approximation. However, for the Fraunhofer
r2
condition z > k rmax
/ 2 to be satisfied, observations must be made at z > 157 m. In the
nomenclature of optical aberrations, a phase factor varying as r 4 is referred to as a
primary spherical aberration. Similarly, a phase factor varying as r 2 is referred to as a
defocus aberration. They are discussed in detail in Section 1.4.2. At z = 10 m, for
example, the amount of defocus wave aberration rmax 2 2 z = 2.5 , but the spherical
4 8z 3 is only onthe order of 10-7 l. Even at z = 1 m, the defocus aberration
aberration rmax
is 25 l but spherical aberration is only on the order of 10-4 l and, therefore, negligibly
small. At shorter distances, it increases rapidly and the Fresnel approximation does not
hold. Since the only difference between the diffraction integrals represented by Eqs. (1-
22a) and (1-23a) is the presence of the quadratic phase factor in the former, Fresnel
diffraction may be referred to as the defocused Fraunhofer diffraction.
In the Fresnel approximation, Eq. (1-18) for the point-spread function reduces to
r i r2
h( r ; z ) =
exp ikz 1 + 2 . (1-24)
z 2 z
Here, kz represents the phase delay of all components in propagating from one plane to
another a distance z apart. The factor of
i indicates that the diffracted wave is out of
phase with the incident wave by / 2 . The inverse dependence on z represents the
inverse-square law of wave intensities. Equation (1-24) shows that a Huygens’ spherical
wavelet is replaced by a parabolic one in the Fresnel approximation. The parabolic
surface shifts laterally as the source of the secondary wavelet shifts in the z = 0 plane,
and the diffracted field is shift-invariant. The shift-invariant property may also be seen by
r r r
letting U ( r ; 0) equal a Dirac delta function ( r
r0 ) in Eq. (1-22).
r
If we consider a Huygens’ wavelet centered at r , Eq. (1-24) is replaced by
r r 2
r r i " ( r
r ) $"
h( r
r ; z ) =
exp !ikz 1 + % . (1-25)
z 2 z 2 "
#" &
In the Fraunhofer approximation, the term in r 2 in the exponent on the right-hand side
of Eq. (1-25) is neglected so that it reduces to
rr r r
i r 2
2r r
h( r ; r ; z ) =
exp ikz 1 + . (1-26)
z 2z 2
If we ignore the quadratic phase factor varying as r 2 in Eq. (1-26) (because it is small for
large z and it does not impact the irradiance distribution of the Fraunhofer pattern), Eq.
(1-26) shows that in effect the Huygens’ spherical wavelet has been replaced by a planar
one. The direction cosines of its surface normal are given by (, , ) , where
r
(, ) = (r / z) . Physically, this is understandable since a small portion of a large
spherical surface can be approximated by a plane. As the source of the secondary wavelet
r
shifts in the z = 0 plane, thereby changing r , the plane surface representing it is tilted.
Hence, the shift-invariant property of the diffracted field is lost. This may also be seen by
r r r
letting U ( r ; 0) equal a Dirac delta function ( r
r0 ) in Eq. (1-23). However, this loss
is only in a mathematical sense. In reality, the quadratic phase factor kr 2 / 2 z (which is
12 IMAGE FORMATION
negligibly small in the Fraunhofer region) is nonzero, and the diffracted wave field
remains shift invariant.
1.2.3 Transfer Function of Free Space

r
The Fourier transform exp (ik z ) of the point-spread function h( r ; z ) , where
( )
12
= 1
2 v 2 , is called the transfer function of free space since it relates the spectral
r r
component A( v ; z ) in the z plane to the corresponding component A( v ; 0) in the z = 0
plane according to Eq. (1-10). In principle, the diffracted wave field in a z plane can be
obtained by inverse Fourier transforming the product of the Fourier transforms of the
wave field in the z = 0 plane and the transfer function of free space. It should be noted
that the waves with spatial frequencies v > 1 or 2 + 2 > 1 correspond to an
imaginary value of . Such waves decay exponentially to negligible values in a short
distance as they propagate, and are called evanescent waves. Hence, the region of
integration in the (, ) plane is limited to 2 + 2 ' 1.
The free-space transfer function, namely, exp (ik z ) may be written
[ ( )]
1/ 2 $
H (, ; z ) = exp !ik z 1
2 + 2 % ( Rayleigh
Sommerfeld) . (1-27)
# &
For small values of the direction cosines a and b, i.e., for small angles of diffraction with
z axis, Eq. (1-27) reduces to
# 2
(
H (, ; z ) = exp !ik z 1
2 + 2
1
)$%& ( Fresnel and Fraunhofer ) . (1-28)
It is easy to show that it represents the Fourier transform of the point-spread function in
the Fresnel approximation given by Eq. (1-24)[see Problem 1]. Hence, the Fresnel
approximation is synonymous with diffraction in the paraxial approximation. It should be
noted that, mathematically, there is no transfer function associated with Fraunhofer
diffraction since the shift invariant property of the diffracted field is destroyed in the
Fraunhofer approximation. However, as explained above following Eq. (1-26), the
Fresnel transfer function can be used in this approximation as well.
1.3 GAUSSIAN IMAGE

In this section we derive a theorem relating the radiance distribution of an object and
the irradiance distribution of its Gaussian image formed by an imaging system.
Theorem 3. The Gaussian image of a small object lying at a large distance from an
imaging system is an exact replica of the object, except for its magnification.
The object and the image lie in mutually parallel planes that are perpendicular to the
optical axis of the system. The aperture stop of the system and its images in the object
and image spaces, namely, the entrance and exit pupils, respectively, also lie in planes
that are parallel to the object and image planes.
1.3 Gaussian Image 13
r
Consider, as illustrated in Figure 1-2, an object of radiance B ( ro ) (in W m 2 sr ),
r
where ro is the position vector of a point in the object plane lying at a distance zo from
the plane of the entrance pupil. The position of the Gaussian image of a point object
r
located at ro is given by
r r
rg = M ro (1-29)
in the Gaussian image plane, where M is its transverse magnification. An object area
r r
element * ro centered at ro is imaged as an area element
r r
* rg = M 2 * ro (1-30)
r
centered at rg . We assume that the object lies at a very small angle from the optical axis
of the system so that its cosine is approximately equal to unity. We also assume that the
size of the entrance pupil is small compared to zo , so that all points on it lie at
approximately the same distance zo from the object element. Accordingly, the solid angle
subtended by an entrance pupil of an area Sen at the object element is Sen zo2 . The power
entering the pupil from the object element is given by
r r
( )
Pen = Sen zo2 B ( ro ) * ro . (1-31)
If , is the transmission factor of the system representing the fraction of the incident
power transmitted by it, the corresponding power exiting from the exit pupil is given by
Pex = , Pen
r r
( )
= , Sen zo2 B ( ro ) * ro . (1-32)
EnP ExP

Po( ro) Pg( rg)
Oen OA Oex
Object Optical Gaussian

plane system image plane
(–)zo zg
Figure 1-2. Schematic of Gaussian imaging by an optical system with its optical axis
OA, entrance pupil EnP, and exit pupil ExP. A point object lies at Po with a position
r
vector ro in the object plane at a (numerically negative) distance zo from the
r
entrance pupil. Its Gaussian image lies at Pg with a position vector rg in the
Gaussian image plane at a distance z g from the exit pupil.
14 IMAGE FORMATION
r
(r )
This power is contained in the image element * rg . If Ig rg is the irradiance (in W m 2 )
of this image element, then
(r )
Pex = Ig rg * rg
r
. (1-33)
Comparing Eqs. (1-32) and (1-33), we obtain
(r ) r
( ) r
Ig rg * rg = , Sen zo2 B ( ro ) * ro
r
. (1-34)
Substituting Eqs. (1-22) into Eq. (1-30), we obtain
(r ) (
Ig rg = , Sen zo2 M 2 B rg M) (r ) . (1-35)
Equation (1-35) describes the irradiance distribution of the Gaussian image, i.e., the
aberration-free image according to geometrical optics. It shows that, except for its
magnification, the Gaussian image of a small object lying at a large distance from the
imaging system is an exact replica of the object (Theorem 3). For systems with large
fields of view, the irradiance actually decreases according to the fourth power of the
cosine of the chief ray angle (i.e., the angle of the ray passing through the center of the
aperture stop, and therefore through the centers of the entrance and exit pupils), as
discussed in Section 2.6 of Part I. There is an additional decrease beyond a certain angle
due to vignetting of the rays by one or more elements of the system, resulting in an
effective aperture with a reduced area compared to that for an on-axis point object (see
Section 2.2.4 of Part I). Although the aberrated image according to geometrical optics can
be calculated, we will not do so, since what is observed in practice is determined by
diffraction, which we discuss next.
1.4 DIFFRACTION IMAGE

Now we determine the irradiance distribution of the diffraction image, i.e., the image
based on diffraction of the object radiation at the exit pupil of the imaging system. From
the Huygens-Fresnel principle, we derive the following two theorems regarding the
images of a point object and an incoherent isoplanatic object.
Theorem 4. The diffraction PSF of an imaging system is proportional to the modulus

square of the inverse Fourier transform of its pupil function.
Theorem 5. The diffraction image of an isoplanatic incoherent object is equal to the

convolution of its Gaussian image and the PSF of the imaging system.
1.4.1 Pupil Function

r
Consider a system imaging a point object Po located at ( ro ; zo ) , as illustrated in
Figure 1-2. A diverging spherical wave, with its center of curvature at the object point, is
incident on its entrance pupil. If the system is aberration free, a spherical wave of the
form exp(
ikR) R , where R is its radius of curvature, emerges from its exit pupil
r
( )
converging to the Gaussian image point rg ; zg . As illustrated in Figure 1-3, a spherical
1.4 Diffraction Image 15
surface S passing through the center of the exit pupil with its center of curvature at the
Gaussian image point Pg is called the Gaussian reference sphere. A spherical wave
emerging from the exit pupil implies that the object rays transmitted by the system travel
equal optical path lengths when propagating from the object point to the reference sphere.
It should be evident that the reason for considering diffraction of the object radiation
at the exit pupil is that it represents the limiting aperture of the ray bundle converging to
the image. However, for point objects with sufficiently large field angles, different
portions of a ray bundle may be limited by different apertures (including imaging
elements) of the system. In such cases, the complex amplitude in the image plane may be
determined by considering diffraction at each of the limiting apertures. In practice,
however, the shape (perimeter) of the exit pupil is determined in the plane of the axial
exit pupil and the image is determined by considering diffraction at it.
If the optical path lengths of the rays are not equal to each other, a distorted spherical
wave emerges from the exit pupil and the system is said to be aberrated. If rays from a
point object are traced through the system so that they travel an optical path length equal
to that of the chief ray up to the exit pupil, the surface passing through their end points is
called the system wavefront for that point object. The optical deviation (i.e., the
ExP

Pi (ri; zi)
Pig
Pg(rg; zg)
R
O l s

Pp (rp; 0) W
S
Defocused Gaussian
image plane image plane
zi
zg
Figure 1-3. Elements of diffraction imaging. The reference sphere S of radius of

curvature R passing through the center of the exit pupil is centered at the Gaussian
r
image point Pg . Pp is a point in the plane of the exit pupil with a position vector rp .
r
W is the aberrated wavefront and ri is the position vector of a point Pi in a
defocused image plane at a distance z i from the exit pupil. Pig is the point of
intersection of the line joining the center Op of the exit pupil and the Gaussian
image point with the defocused image plane.
16 IMAGE FORMATION
geometrical deviation times the refractive index ni of the image space) of the wavefront
from the Gaussian reference sphere along a ray is called the wave aberration. It
represents the difference between the optical path lengths of the ray under consideration
and the chief ray (passing through the center of the exit pupil) in propagating from the
point object to the reference sphere. Thus, the wave aberration of a ray is numerically
positive if it travels a longer optical path length compared to the chief ray in reaching the
reference sphere. Since the optical path lengths of the rays from the reference sphere to
the Gaussian image point are equal, the wave aberration of a ray is also equal to the
difference between its optical path length and that of the chief ray in propagating from the
point object Po to the Gaussian image point Pg .
r (r r )
For an aberrated system, let W rp ; ro be the wave aberration of a ray passing
r
through a point rp in the plane of the exit pupil. The complex amplitude at rp due to an
r r
object element * ro centered at ro may be written
(r r ) (r r )
*Uex rp ; ro = P rp ; ro exp (
iks) , (1-36)
where
xp

Pp(rp)
p xi
rp

Pi (ri)
Op
Pig ri i xg
an il
pl up
e
P
zi
R Pg(rg)
yp
Oi g
rg
e ion
an at
pl erv
zg
bs
O
Og
yi
z
pl n
e
e sia
an
ag us
im Ga
yg
Figure 1-4. Right-hand coordinate system in the pupil plane, and defocused and
Gaussian image planes. The optical axis of the system is along the z axis.
pg = p
g and pi = p
i .
r r2 12
s = zg2 + rp
rg (1-37a)

1 r r 2 1 r r 4
= zg + rp – rg
rp – rg + .... (1-37b)
2 zg 8zg3
1
( 1
) (r )
2
= zg + rp2 + rg2
2 rp rg cos pg
3 2
p + rg2
2 rp rg cos pg + .... (1-37c)
2 zg 8z g
is the distance between the pupil point Pp and the Gaussian image point Pg (see Figure
1-4) and
(r r ) (r r ) [
P rp ; ro = A rp ; ro exp ikW rp ; ro (r r )] , inside the exit pupil
= 0 , outside the exit pupil , (1-38)
r r
is called the pupil function of the system. The amplitude function A rp ; ro , called the ( )
apodization function, represents the variation of amplitude of the wave across the exit
pupil and accounts for any nonuniform transmission of the system. The inverse-square-
law dependence of irradiance on the distance s is contained in the amplitude function. It
r r
should be evident that rp = rp , rg = rg , and pg is the angle between the position
r r r
( )
12
vectors rp and rg . Note that s = R = zg2 + rg2 when rp = 0 .
r
The irradiance at a point rp in the plane of the exit pupil due to the object element
under consideration is given by
(r r ) (r r )
2
I p rp ; ro = *Uex rp ; ro (1-39a)
(r r )
2
= P rp ; ro (1-39b)
(r r )
= A 2 rp ; ro . (1-39c)
The corresponding power in the exit pupil is obtained by integrating the irradiance across
the pupil, i.e.,
r r r r
P (r ; r )
2
Pex ( ro ) = p o d rp (1-40a)
r r r
= A ( rp ; ro ) d rp
2
. (1-40b)
1.4.2 Diffracted Wave

Now we consider propagation of the wave from the pupil plane to an observation
plane at a distance zi from it. Substituting Eq. (1-36) into Eq. (1-20), neglecting 1 kl
(compared to unity) and the obliquity factor zi l , and replacing l by zi (except in the
exponent) for zi much greater than the size of the exit pupil, we obtain
18 IMAGE FORMATION
r r r r r
*Ui ( ri ; ro ; zi ) =
i
zi
( ) [
P rp ; ro exp ik (l
s) d rp ] , (1-41)
where
r r 2 1 2
l = zi2 + rp
ri (1-42a)

1 r r 2 1 r r 4
= zi + rp
ri
rp
ri +K (1-42b)
2 zi 8zi3
1
( 1
) (r )
2
= zi + rp2 + ri2
2 rp ri cos pi
3 2
p + ri2
2 rp ri cos pi + .... (1-42c)
2 zi 8 zi
r
is the distance of the observation point Pi from a pupil point Pp , ri = ri and pi is the
r r
angle between the position vectors rp and ri (see Figure 1-4). We may add that if the
obliquity factor is not close to unity, it may be replaced by z R for a better
approximation.
Since the quantity l

s in the exponent is multiplied by a large number k, it must be
evaluated to a small fraction of . Subtracting Eq. (1-37c) from Eq. (1-42c), we obtain
ri2 rg2 1 r r zi r 1 1 1 2 1 ri4 rg

4
(
l
s = zi
z g + ) 1
2 z
i

zg zi
r r
p i

r +
zg 2 zi zg
g rp

8 zi3 zg3
1 1 1 4 1 3 ri rg

3
3 rp + 2 rp 3 cos pi
3 cos pg
8 zi zg zi zg
1 2 ri2 rg2 1 r 2 rg2

rp 3 cos 2pi
3 cos 2pg
rp2 i3
3
2 zi zg 4 zi zg
1 ri3 rg3
+ rp 3 cos pi
3 cos pg + ... . (1-43)
2 zi zg
Each term on the right-hand side of Eq. (1-43) depends on the location of the
observation and Gaussian image planes and represents a phase term when multiplied by k.
The first term represents a piston aberration. It is simply the distance between the two
planes. Similarly, the second term represents a quadratic phase factor. These two terms do
not affect the irradiance distribution in the defocused iamge plane. Hence, they have no
impact on the image of an incoherent extended object since it is linear in irradiance. The
third term, when exponentiated, forms a Fourier kernel in Eq. (1-41). The quantity
r
( )
zi zg rg in this term represents the position vector of a point Pig where the line joining
the center Op of the exit pupil and the Gaussian image point Pg intersects the defocused
image plane (see Figure 1-4). The fourth term represents the usual defocus aberration due
to the observation plane being different from the Gaussian image plane. The fifth term is
a higher-order piston aberration, which also does not affect the irradiance distribution in
the observation plane. The sixth term has the form of a primary spherical aberration, as
may be seen by comparing it with Eq. (3-34) of Part I. The next four terms have two
components each, but only one of them represents a classical aberration. The other
component in each case is only deceptively similar to a classical aberration, but it
depends on the coordinates of the observation point (rather than the Gaussian image
point). For example, the component varying as rp3ri cos pi may be called axial coma in a
manner somewhat similar to the axial coma of a perturbed system (see Section 7.2.1 of
Part I), and the component varying as rp3rg cos pg may be called Seidel coma.
The quantity
r r r 1 r r z r
( ) zi

Wd rp ; ri ; rg = l
s + rp ri
i rg
zg
(1-44)
is the aberration of the diffracted wave.5 It should be noted, however, that this aberration
is related to the propagation of the wave and has nothing to do with the characteristics of
the imaging system (other than its dependence on the location of the Gaussian image
point). It would be convenient to call the last five terms of Eq. (1-43) spherical aberration,
coma, astigmatism, field curvature, and distortion of the diffracted wave, respectively.
The aberrations that are radially symmetric in pupil coordinates vanish when the
observation plane coincides with the Gaussian image plane, i.e., when zi = zg . Hence,
they are aberrations of the diffracted wave associated with defocus. Only the first four
terms are retained in the Fresnel approximation, and only the first three are retained in the
Fraunhofer approximation. These approximations limit the region of observation space
that accurately satisfies the Fourier-transform relationship to near the optical axis and
away from the pupil. When the observation plane coincides with the Gaussian image
plane, the Fraunhofer region of diffraction reduces to the Gaussian image plane.
In practice, the aberrations of the diffracted waves are negligibly small.6 For
example, the depth of focus will be determined by the amount of defocus aberration that
can be tolerated. For systems with large Fresnel numbers, e.g., photographic systems, the
depth of focus for 8 defocus aberration tolerance is F 2 . Here, F = zg D is the focal
ratio of the image-forming light cone, where D is the diameter of the exit pupil. In that
case, the corresponding spherical aberration term in Eq. (1-43) is approximately equal to
( )
3 128 F 2 , which is negligibly small. The spherical aberration is also small for a
system with a small Fresnel number. This may be seen by considering, as discussed in
Section 2.5, a laser transmitter with an aperture diameter of 25 cm focusing a beam of
wavelength 10.6 mm at a distance of 1.47 km so that its Fresnel number is unity. Based on
the defocus aberration, a Strehl ratio of 0.8 is obtained at a distance of 3 km, showing a
large depth of focus. The magnitude of the spherical aberration term in this case is only
about 10 -2 -m . For the angle-dependent terms, such as coma, it is important to note that
the region of interest in the image of a point object is where there is a significant
r
( )
illumination. This region corresponds to ri , pi values that lie, say, within the Airy disc
20 IMAGE FORMATION
(discussed in Section 2.1) that is centered at the Gaussian image point. Thus, even for
rg >> F , we have ri . rg and pi differs from gi by no more than approximately D.
Hence, the coma term for a field angle = rg zg is approximately equal to 16 F 2 , ( )
which is also negligibly small. The value of the axial coma is equal to 16 F 2 .
Similarly, the astigmatism, field curvature, and distortion terms are approximately equal
( ) ( ) ( )
to 2 8F , 3 2 16 , and 3 4 , respectively. Their axial values are extremely
( ) ( )
small and equal to 8zg , 16zg , and ( 2 D) . Thus, the aberrations of a
2
diffracted wave are negligibly small. The expressions for their values are summarized in
Table 1-1. It should be noted that in high-quality imaging systems, the value of is often
quite small (few degrees or less) since it is limited by the aberrations of the system itself.
We may neglect all terms in Eq. (1-43) beyond the defocus term if
1 1 4
3
3 a < 8 , (1-45)
zi zg
where a is the maximum value of rp , i.e., if the peak aberration contributed by the
primary spherical aberration is < / 8 . We will see, for example, in Section 2.3 that such
an aberration has a negligible effect on the quality of an image. Hence, Eq. (1-41) for the
complex amplitude may be written
r r " 1 r 2 rg $"
2
*Ui ( ri ; ro ; zi ) =
i
zi
( )
exp !ik zi
zg + i
%
"# 2 zi zg "
&
r r 2 i r r z r r

( )
× P rp ; ro ; zi exp
zi

rp ri
i rg d rp
zg
, (1-46)

where
Table 1-1. Aberrations of a diffracted wave for an on-axis and an off-axis point
object. The field angle of the image point is indicated by the angle .
Aberration Axial Value Value for field angle

>> / D
Defocus /8 /8

Spherical (3 128 F ) 2
(3 128 F ) 2
Coma 16 F 2 ( 16 F ) 2
Astigmatism ( 8zg ) ( 8F)

2
Field Curvature ( 16zg ) (3 16) 2
Distortion ( 2 D)2 ( 4)

3
r r r r ik 1 1
(
P rp ; ro ; zi ) ( )
= P rp ; ro exp
rp2
2 zi zg
(1-47)
is the defocused pupil function signifying that the image is observed in a plane other than
the Gaussian image plane, i.e., zi / zg .
So far we have been concerned with the phase variations across the exit pupil for a
wave from a given point object and ignored the optical path length of the chief ray from
the object point to its Gaussian image point. There is an additional phase factor that needs
to be considered when comparing the waves from different point objects. It represents the
difference in the optical path lengths of their chief rays from the object points to their
Gaussian image points. The difference in the optical path lengths of a chief ray for an
r
object point located at ro at a distance zo from the entrance pupil and that for an axial
point object, up to the quadratic approximation (see Figure 1-2), is given by
(r r )
W ro ; rg = ( Po Oen + z )
Oex Pg
zg ( )
2 r2
~
ro + g , (1-48)
2 zo 2 zg
where zo is numerically negative according to our sign convention given in Section 1.3.2
of Part I. It is assumed that the optical path lengths of the chief rays from the entrance
pupil to the exit pupil are the same. This must, of course, be verified for the system under
consideration by ray tracing, or the phase must be adjusted if the assumption is not valid.
Equation (1-48) is evidentally valid in the case of a thin lens when its aperture stop and,
therefore, its entrance and exit pupils are located at the lens. Adding the phase
contribution from Eq. (1-48), Eq. (1-46) is modified to
r r " 1 r 2 r 2 $"
*Ui ( ri ; ro ; zi ) =
i
zi
( )
exp !ik zi
zg + i
o %
"# 2 zi zo "
&
r r 2 i r r z r r

(
× P rp ; ro ; zi exp

)
rp ri
i rg d rp . (1-49)

z zg
i
The additional phase factor does not impact imaging of an incoherent object, but does
impact coherent imaging, which is discussed in Section 1.12.
In obtaining Eq. (1-49), we have in effect propagated the wave from the exit pupil to
the observation plane in the Fresnel approximation, as may be seen by substituting Eq. (1-
36) into Eq. (1-22a) and retaining only the first two terms on the right-hand side of Eq.
(1-37b). Indeed, for a thin lens of focal length f, Eq. (1-49) can also be obtained by
considering Fresnel propagation of the spherical wave from a point object to the entrance
pupil, its modification by the lens, i.e., introduction of the quadratic phase factor
( )
exp
ikrp2 2 f , and Fresnel propagation to an observation plane (see Problem 4).
22 IMAGE FORMATION
However, we have demonstrated that such an approximation is quite valid in describing

image formation. Accordingly, diffraction described by Eq. (1-49) is known as Fresnel
diffraction. The approximations of l ~ zi and retaining terms up to the defocus term in
Eq. (1-43) are known as Fresnel approximations. The conditions that the distances zi and
zg be much greater than the sizes of the pupil and the image, and the condition of Eq. (1-
45) under which these approximations are valid may be referred to as Fresnel conditions.
The integral in Eq. (1-49) is called the Fresnel diffraction integral of the pupil function
r r
( )
P rp ; ro and represents the diffraction pattern of the aberrated pupil in a defocused
image plane. Similarly, when zi = zg , it is called the Fraunhofer diffraction integral and
represents the Fraunhofer diffraction pattern of the aberrated pupil in the Gaussian image
plane. Since the only difference between the two diffraction integrals is the effect of the
defocus aberration, Fresnel diffraction may be regarded as defocused Fraunhofer
diffraction. It may be emphasized that the focusing aspect of the imaging system
collapses the vast region of Fraunhofer diffraction to its Gaussian image plane. The
irradiance distribution in a defocused observation plane on either side of the Gaussian
image plane represents the Fresnel diffraction of the aberrated pupil.
1.4.3 Incoherent PSF and Shift-Invariant Imaging of an Incoherent Object

For an incoherent object, the irradiance distribution of its image is obtained by a
linear superposition of the irradiance distributions of its image elements. From Eq. (1-49),
the irradiance distribution in the image of an object element in the defocused image plane
may be written
r r r r
* Ii ( ri ; ro ; zi ) = * Ui ( ri ; ro ; zi )
2
r r r
= Pex (ro ) PSF ( ri ; ro ; zi ) , (1-50)
where
2
rr r r 2 i r r z r r
PSF ( ri ; ro ; zi ) = r
1
Pex ( ro )2 zi2
( )
P rp ; ro ; zi exp
zi

rp ri
i M ro d rp
zg
(1-51)

is called the incoherent point-spread function of the imaging system, as we shall see in
the next section. It is proportional to the modulus square of the inverse Fourier transform
r
of the defocused pupil function of the system (Theorem 4). Hence, substituting for Pex (ro )
from Eq. (1-32) into Eq. (1-50) and integrating over the object, we obtain
r rr
Ii ( ri ; zi ) = * I ( ri ; ro ; zi )
r r r r
(
= , Sen zo2 ) B ( ro ) PSF ( ri ; ro ; zi ) d ro
object
, (1-52)
where we have neglected the dependence of Sen and the solid angle it subtends at an
object element on the location of the element. Thus, the object is assumed to be small, far
from the system, and at a small angle from the axis of the system so that the cosine-to-
the-fourth-power dependence on the chief ray angle in the object space is negligible, and
vignetting of its pupil for the off-axis object elements (discussed in Section 2.6 of Part I)
is either zero or negligible.
Now, we make additional approximations to obtain a shift-invariant PSF, i.e., a PSF

that shifts in the image plane without changing its functional form as the object element is
shifted in the object plane. In imaging systems, only a small amount of longitudinal
defocus can be tolerated; therefore, zi ~ zg . (This may not be true in some laser beam
illumination systems, as discussed later in Section 2.5.2. It is shown there that the Fresnel
number of the exit pupil as observed from the center of the reference sphere determines
the depth of focus. The larger the Fresnel number, the smaller the depth of focus. The
imaging systems generally have a large Fresnel number and laser beam illumination
systems have a small Fresnel number.) Moreover, high-quality systems can tolerate only
small amounts of aberration and, therefore, often afford only small fields of view. For
such systems, z ~ R .
r (r r
)
The pupil function P rp ; ro ; zi represents the actual complex amplitude at a point
r r
rp in the plane of the exit pupil due to an object element * ro centered at ro . Evidently, it
r
depends on the location ro of the object element. We now introduce a relative pupil
r r
( )
function G rp ; ro ; zi that represents the complex amplitude per unit amplitude of the
object element. The two pupil functions are related to each other according to
(r r
P rp ; ro ; zi ) r
(r
= U ( ro ) G rp ; ro ; zi
r
) , (1-53)
r r
where U ( ro ) is the amplitude at a point ro . It should be evident that the two pupil
r
functions have identical dependence on rp . Substituting Eq. (1-53) into Eq. (1-40a), we
obtain
r r r r r
G (r ; r ; z )
2
Pex ( ro ) = U ( ro )
2
p o i d rp . (1-54)
For a small object, we may neglect any changes in the nonuniform transmission or
r r r
aberrations introduced by the system as a function of ro . Hence, G rp ; ro ; zi may be
r ( )
assumed to be independent of the position vector ro . [The assumptions of a small field of
view and a small object are the same as those made in obtaining the Gaussian image
distribution of Eq. (1-35).] An object of small extent for which the relative pupil function
r r
( )
G rp ; ro ; zi may be considered independent of the position of a point on it is called
isoplanatic (or the system is said to be isoplanatic for the small object under
r r r
( )
consideration). Of course, the pupil function P rp ; ro ; zi will vary as U ( ro ) varies with
r
ro for such an object.
Substituting Eqs. (1-53) and (1-54) into Eq. (1-51), neglecting the dependence of
r r r
( )
G rp ; ro ; zi on ro , and letting zi ~ zg ~ R , for an isoplanatic object imaged by a system
with a small field of view, we obtain
24 IMAGE FORMATION
r 2 i r r r r
2
r r
( )
G rp ; zi exp
R rp

( ri
M ro ) d rp

PSF ( ri ; ro ; zi ) = r r . (1-55)
( )
2
2 R 2 G rp ; zi d rp
We note that the integral in the numerator on the right-hand side of Eq. (1-55) depends on
r r r
the position vectors ri and rg (which is equal to M ro ) of the observation and Gaussian
r r
image points, respectively, through their difference ri
rg only, and we may replace
r r r r
PSF ( ri ; ro ; zi ) by PSF ( ri
M ro ; zi ) . Thus, the PSF is shift invariant in the sense that its
form does not change as the object point is shifted; only its location changes by virtue of
r
it being centered at rg . Accordingly, Eq. (1-52) for the irradiance distribution of the image
of an isoplanatic incoherent object may be written
r r r r r
(
Ii ( ri ; zi ) = , Sen zo2 ) B ( ro ) PSF ( ri
M ro ; zi ) d ro
object
, (1-56a)
(
= , Sen zo2 M 2 ) B (rr M ) PSF (rr
rr ; z ) d rr
g i g i g (1-56b)
r r r r
= Ig ( rg ) PSF ( ri )

rg ; zi d rg , (1-56c)
where we have made use of Eq. (1-35). Thus, the diffraction image of an isoplanatic
incoherent object is equal to the convolution of its Gaussian image (which is a scaled
replica of the object) and the PSF of the imaging system (Theorem 5). It should be
evident that the magnification M of the image is assumed to be constant across the
isoplanatic region of the object. Of course, a nonuniform magnification will result in a
distortion of the image also.
Multiplying the numerator and denominator of the right-hand side of Eq. (1-55) by
r
U ( ro ) , we may also write the PSF in the form
2
r r 2 i r r r
2
PSF ( ri ; zi ) =
1
2 2 P rp ; zi exp
Pex R R
( )
rp ri d rp

, (1-57)
r r
where we have suppressed any explicit dependence on ro and the position vector ri is
r
written with respect to rg . The irradiance distribution of the image of a particular object
element is obtained by multiplying the PSF by Pex for that element.
1.5 PHYSICAL SIGNIFICANCE OF INCOHERENT PSF

To understand the physical significance of the PSF defined by Eq. (1-51), we
r
consider the image of a point object. A point object of intensity B j (in W/sr) located at rj
in the object plane may be described by
r
(r
B ( ro ) = B j ro
rj
r
) , (1-58)
r r

where ( ) is a Dirac delta function. Its Gaussian image point is located at rg = M rj .
1.5 Physical Significance of PSF 25
Substituting Eq. (1-58) into Eq. (1-52), we obtain the irradiance distribution of its image,
namely,
(r r
Ii ri ; rj ; zi ) (r
= Pex PSF ri ; rj ; zi
r
) , (1-59a)
where
(
Pex = , Sen zo2 B j ) (1-59b)
represents the total power from the point object in the exit pupil of the imaging system.
Substituting Eq. (1-51) into Eq. (1-59a), we obtain
2
rr r r 2 i r r z r r
(
Ii ri ; rj ; zi ) 1
= 2 2
zi
(
P rp ; rj ; zi exp
zi
)
rp ri
i M rj d rp
zg
. (1-60)

For an isoplanatic object, we substitute Eq. (1-58) into Eq. (1-56a) and obtain
(r r
Ii ri ; rj ; zi ) (r
= Pex PSF ri
M rj ; zi
r
) , (1-61)
or substituting for the PSF from Eq. (1-57),
rr 1 r 2 i r r r
r
2
(
Ii ri ; rj ; zi ) R
(
= 2 2 P rp ; zi exp
R
rp ) ( ri
M rj d rp

) . (1-62)
It is evident from Eq. (1-62) that the images of two point objects lying within an
isoplanatic patch are identical except that they are centered at their respective Gaussian
image points. Such an imaging is referred to as shift-invariant imaging, and comes about
because of the invariance of the relative pupil function with the position of an object
point on the isoplanatic object. Integrating both sides of Eq. (1-62), we find that
r r r 1 r r 2 i r r r r
2
(
i i j i R
)
I r ; r ; z d ri = 2 2 d ri P rp ; zi exp
R
rp ( ) ( )
ri
M rj d rp

r r r r r r
= d rp P ( rp ; zi ) P * ( rp) ( rp
rp ) d rp
r r
P (r ; z )
2
= p i d rp
= Pex . (1-63)
We obtain the same result if we integrate both sides of Eq. (1-61). It represents the
conservation of power; i.e., the total power in the image is equal to the total power in the
r r r
exit pupil. It also shows that the integral of PSF ri ; rj ; zi over all values of ri in the ( )
image plane is unity. Hence, PSF given by Eq. (1-51) or Eq. (1-57) represents the
r
irradiance distribution of the defocused image of a point object located at ro per total
26 IMAGE FORMATION
power in the image. It is called the diffraction point-spread function (PSF) of the system.
Note that its dimensions are m
2 .
Unapodized Pupil
A system with a pupil that is uniformly illuminated is said to be unapodized. For
r
( )
such systems, G rp is a constant varying inversely with the distance zo of the entrance
r
( )
pupil from the object. For an unapodized pupil, let A rp be equal to a constant, say, A0 .
r
If we redefine ri , as illustrated in Figure 1-5, as the position vector of an image point Pi
r r
with respect to the Gaussian image point rg = M rj when the image is observed in the
r
Gaussian image plane, or with respect to the corresponding point zi zg rg (where the ( )
line joining the center of the exit pupil and the Gaussian image point intersect the
defocused image plane) if it is observed in a defocused image plane at a distance zi from
the plane of the exit pupil, Eq. (1-62) becomes
r 2 i r r r
2
Ii ( ri ; zi )
I r
[ ( )]
= 2 0 2 exp i kW rp ; zi exp
R R
rp ri d rp

, (1-64)
where I0 = A02 is the irradiance at the exit pupil. If, in addition, the pupil is aberration
free and the observation is made in the Gaussian image plane so that there is no defocus
r
( )
aberration, i.e., if W rp ; zg = 0, then Eq. (1-64) reduces to
r 2 i r r r
2
I (0)
Ii ( ri ) = i 2 exp
Sex R
rp ri d rp , (1-65)
where
Pex Sex
Ii ( 0 ) = (1-66)
2 R 2
r
is the aberration-free irradiance at the Gaussian image point ri = 0 . Of course,
Pex = I0 Sex in this case. The effect of aberrations and or amplitude variations on the
central irradiance is discussed in Section 1.9.
yi yg
Pi Pi

ri ri
Pig Pg
xi xg
Oi Og
(a) Defocused image plane (b) Gaussian image plane
r
Figure 1-5. Redefinition of ri as the position vector of an image point Pi in the (a)
defocused image plane with respect to Pig and the (b) Gaussian image plane with
respect to Pg .
1.6 Optical Transfer Function (OTF) 27
1.6 OPTICAL TRANSFER FUNCTION (OTF)

In Section 1.4, we have shown that the irradiance distribution of the diffraction
image of an incoherent object formed by an imaging system is equal to the convolution of
the irradiance distribution of its Gaussian image and the PSF of the system under certain
conditions. Instead of decomposing an object into object elements and determining the
image as a sum of the corresponding image elements, as was done in obtaining Eq. (1-
56c), we now discuss imaging in the spatial-frequency domain. We decompose the object
into its sinusoidal spatial frequency components and show that each component is imaged
by the system with a reduced contrast and a phase change that depend on its spatial
frequency. For example, if the spatial frequency of an object component is too high, the
contrast of the corresponding image component may be zero, i.e., an image component
with a uniform distribution. Thus, those spatial frequencies in the object that are imaged
with zero contrast will be absent from the image. This is another way of saying that
certain details present in the object will be missing from its image. We introduce the
optical transfer function (OTF) of an imaging system as the Fourier transform of its PSF.
The physical significance of the OTF is discussed, and certain properties including its
slope at the origin are derived.
1.6.1 General Relations

We derive the following three theorems related to the OTF.
Theorem 6. The spatial-frequency spectrum of the diffraction image of an

isoplanatic incoherent object is equal to the product of the spectrum of its Gaussian
image and the OTF of the system, which by definition is equal to a Fourier transform of
its PSF.
Theorem 7. The OTF of an imaging system is also equal to the autocorrelation of its
pupil function.
Theorem 8. The PSF and the OTF of a system with a radially symmetric pupil
function form a zero-order Hankel transform pair.
A relationship between the spatial-frequency spectra of the object and image may be
obtained by taking a Fourier transform of both sides of Eq. (1-56c). Let
r r r r r
B˜ ( vo ) = B ( ro ) exp (2 i vo ro ) d ro (1-67)
and
r r r r r
I˜i ( vi ) = Ii ( ri ) exp (2 i vi ri ) d ri (1-68)
be the spectral components of the object and image corresponding to the object and image
r r
spatial frequencies vo and vi , respectively, where
r r
vi = vo M . (1-69)
28 IMAGE FORMATION
Substituting Eq. (1-56c) into Eq. (1-68) and suppressing the zi dependence of the PSF
and the OTF, we obtain
r r r r r r r r
I˜i ( vi ) = d ri exp (2 i vi ri ) Ig ( rg ) PSF ( ri
rg ) d rg
r r r r r r r r r r
= d rg Ig (rg ) exp (2i vi rg ) PSF (ri
rg ) exp [2 vi (ri
rg )] d ri
r r
= I˜g ( vi ) 0 ( vi ) , (1-70)
r
where Ĩg ( vi ) is the spectral component of the Gaussian image corresponding to a spatial
r
frequency vi and
r r r r r
0 ( vi ) = PSF ( ri ) exp (2 i vi ri ) d ri (1-71)
is called the optical transfer function (OTF) of the imaging system at this frequency.
Equation (1-71) states that the OTF, which is dimensionless, is the Fourier transform of
the PSF. From Eq. (1-35), it is evident that
r r
(
I˜g ( vi ) = , Sen zo2 B˜ ( M vi ) ) (1-72a)
r
(
= , Sen zo2 B˜ ( vo ) . ) (1-72b)
r
Equation (1-70) states that the spatial-frequency spectrum Ĩi ( vi ) of the diffraction image
r
of an isoplanatic incoherent object is equal to the product of the spectrum Ĩg ( vi ) of its
Gaussian image (which in turn is equal to a scaled spectrum of the object) and the OTF
r
0 ( vi ) of the imaging system (Theorem 6). It is a consequence of the fact that if a function
is equal to the convolution of two functions, as in Eq. (1-56c), its Fourier transform is
equal to the product of their Fourier transforms.7
Substituting Eq. (1-57) into Eq. (1-71), the OTF can also be written
r r r r r 2i r r r
2
0 ( vi ) =
1
Pex 2 R 2

( )
d ri exp (2 i vi ri ) P rp exp
R
rp ri d rp

r r
=
1
Pex 2 R 2
d rp P rp ( ) d rrp P * (rrp) d rri exp 2Ri rri (rrp
rrp + R vri )
r
= Pex
1 d rp P rp (r ) d rrp P * (rrp) (rrp
rrp + R vri )
r r r r
= Pex
1 P ( rp ) P 1 ( rp
R vi ) d rp (1-73a)
r r r r r 2 r
P ( rp ) P * ( rp
R vi ) d rp P ( rp ) d rp
=
(1-73b)
r r r
( ) (
= Pex
1 A rp A rp
R vi exp iQ rp ; zi d rp ) [ (r )] r
, (1-73c)
(r )
where the pupil function P rp is given by Eq. (1-38) when the image is observed in the
Gaussian image plane and by Eq. (1-47) when it is observed in a defocused image plane,
and
(r r
Q rp ; vi ; zi ) (r ) (r
= 2 rp ; zi
2 rp
R vi ; zi
r
) (1-74)
is a phase aberration difference function defined over the region of overlap of two pupils,
r r r
one centered at rp = 0 and the other at rp = Rvi . The phase aberration 2 is related to
the wave aberration W according to
2 = (2 ) W . (1-75)
Equation (1-73b) shows that the OTF of an incoherent imaging system is also equal to the
normalized autocorrelation of its pupil function (Theorem 7). [The autocorrelation of a
function f ( x ) is equal to its convolution with the function f * (
x ) .] The region of
integration in the numerator of the right-hand side of this equation is the overlap region of
r
the two pupils. We may note here that the region of overlap is maximum for vi = 0,
giving a value of unity for 0 (0) . Because of the finite size of the pupil, the overlap region
r
reduces to zero at some frequency vc , called the cutoff frequency, and stays zero for
r r r
larger frequencies; i.e., 0 ( vi ) = 0 for vi 3 vc . If the image is observed in a defocused
image plane such that zi is significantly different from R, then R in Eq. (1-73) is replaced
by zi , as may be seen by using the PSF obtained from Eq. (1-60). Problem 4.2 is an
example of this type, where the defocused OTF of a weakly truncated Gaussian pupil is
considered.
Taking an inverse Fourier transform of both sides of Eq. (1-71), we obtain

r r r r r r r r r r
0 (vi ) exp (
2 i vi ri ) d vi = PSF ( ri ) d ri exp [2 i vi ( ri
ri )] d vi
r r r r
= PSF ( r i) ( ri
ri ) d ri
r
= PSF ( ri ) , (1-76)
r r
which shows simply that the functions PSF( ri ) and 0 ( vi ) form a 2-D Fourier transform
pair. The value of the PSF at the origin may be written
r r
PSF (0) = 0 ( vi ) d vi (1-77a)
r r
= Re 0 ( vi ) d vi , (1-77b)
r r
where Re 0 ( vi ) is the real part of 0 ( vi ) . Since the PSF at any point is a real quantity,
only the real part of the complex OTF contributes to the integral. The integral of its
imaginary part must be zero.
Consider a system with a radially symmetric pupil function, i.e., one for which
30 IMAGE FORMATION
(r )
P rp = P rp ( ) . (1-78)
If we let
r
( )
rp = rp cos p , sin p , 0 ' p < 2 , (1-79a)
and
r
ri = ri (cos i , sin i ) , 0 ' i < 2 , (1-79b)
it follows from Eq. (1-57) that

2
2
r 2i
P (rp ) rp drp ( )
1
PSF( ri ) = exp
R rp ri cos p
i d p . (1-80)
Pex 2 R 2
0
Noting that8
2
exp (ix cos ) d = 2 J0 ( x ) , (1-81)
0

where J 0 ( ) is the zero-order Bessel function of the first kind, Eq. (1-80) reduces to
( ) P ( r ) J (2 r r )
2
PSF(ri ) = 4 2 Pex 2 R 2 p 0 p i R rp drp . (1-82)
Thus, the PSF of a system with a radially symmetric pupil function is also radially
symmetric. Multiplying both sides of Eq. (1-82) by Pex , we obtain the radially symmetric
irradiance distribution of the image of a point object. Similarly, Eqs. (1-71) and (1-76)
reduce to
0 (vi ) = 2 PSF(ri ) J 0 (2 vi ri ) ri dri (1-83)
and
PSF(ri ) = 2 0 (vi ) J 0 (2 vi ri ) vi dvi , (1-84)
r
where vi = vi . Thus, the PSF and the OTF of a system with a radially symmetric pupil
function form a zero-order Hankel transform pair (Theorem 8). This result is a
consequence of the fact that the Fourier transform of a radially symmetric function is a
zero-order Hankel transform. From Eq. (1-83), we note that, since the PSF is a real
function, the radially symmetric OTF is also a real function.
By multiplying PSF(ri ) with Pex in Eqs. (1-83) and (1-84) we obtain relationships
between the irradiance distribution of the image of a point object and the OTF of a system
with a radially symmetric pupil function. For example, the irradiance distribution may be
obtained from the OTF according to
Ii (ri ) = Pex PSF(ri )
= 2 Pex 0 (vi ) J 0 (2 vi ri ) vi dvi . (1-85)
The encircled power of the image, i.e., the power contained in a circle of radius rc
centered at the Gaussian image point, is given by
r r
Pi (rc ) = I ( r ) d ri
rr ' r i i
. (1-86)
i c
For a radially symmetric image, Eq. (1-86) reduces to

rc
Pi (rc ) = 2 Ii (ri ) ri dri . (1-87)
0
Or, substituting Eq. (1-85) into Eq. (1-87) and noting that8
z0
J0 (az ) z dz = ( z0 a) J1 (az0 ) , (1-88)

0

where J1 ( ) is a first-order Bessel function of the first kind, we obtain
Pi (rc ) = 2 rc Pex 0 (vi ) J1 (2 vi rc ) dvi . (1-89)
Equation (1-89) gives the encircled power of the image in terms of the radially symmetric
OTF of the system.
1.6.2 Physical Significance of OTF
To understand the physical significance of the OTF defined by Eq. (1-71), we derive
the following theorem.
Theorem 9. Whereas the Gaussian image of a sinusoidal object is sinusoidal with

the same modulation and phase as the object, the diffraction image is also sinusoidal but
with a reduced contrast and changed phase depending on the spatial frequency. The
modulus of the OTF represents the factor by which the contrast changes and its phase
gives the change in the phase.
r
Consider the image of a sinusoidal object of spatial frequency vo , as illustrated in
Figure 1-6:
r r r
[
B ( ro ) = Bo 1 + m cos (2 vo ro + 4)] , (1-90)
where Bo is a constant in units of radiance, m is a dimensionless quantity representing the

modulation or the contrast of the object, and 4 is an arbitrary phase constant. Following
Eq. (1-35), the Gaussian image of the object can be written
32 IMAGE FORMATION
Bo(1+ m)
B(ro)

1/ "6o "
Bo(1– m)
0
xo
lo(1+ m)
lg(ri)

lo(1– m)
0
xi

1/ "6i "
7
lo(1+ m "0 ")
li(ri)

lo(1– m "0 ")

1/ "6i "
0
xi
Figure 1-6. Image of a sinusoidal object shown along the x axis. (a) Object, (b)
Gaussian image, and (c) Diffraction image.
r r r
[
I g ( ri ) = Io 1 + m cos (2 vi ri + 4)] , (1-91)
where
(
Io = , Sen zo2 M 2 Bo ) (1-92)
is the average irradiance of the image. We note that the image is also sinusoidal but with
r r
a frequency vi = vo / M , where M is its magnification. Its modulation and phase are the
same as that of the object, respectively.
Substituting Eq. (1-90) into Eq. (1-56a), we obtain the irradiance distribution of the
diffraction image, which may be written
r r r r r r
r r r r
{ r[
Ii ( ri ) = Io M 2 [ PSF( ri
M ro ) d ro + m Re exp
i (2 vi ri + 4 )
r
]
[
× PSF( ri
M ro ) exp 2 i vi (ri
M ro )] d ro }]. (1-93)
Now, the first integral on the right-hand side of Eq. (1-93) is
r r r r r 2i r r r r

PSF ( ri
M ro ) d ro =

1
Pex 2 R 2
d ro P rp exp
R
( )
rp ( ri
M ro ) d rp

r 2 i r r r r

( )
× P * rp exp
R
rp ( ri
M ro ) d rp

(
= 1 Pex 2 R 2 ) d rr P (rr ) d rr P * (rr )
p p p p
r 2 i r r
× d ro exp
R
rp
rp ( ) (rri
M rro )
(
= 1 Pex M 2 ) d rr P (rr ) d rr P(rr ) (rr
rr )
p p p p p p
( ) P (rr ) r
2
= 1 Pex M 2 p d rp
= 1 M2 . (1-94)
Following Eq. (1-71), the second integral on the right-hand side of Eq. (1-93) may be
written
r r r r r r r
PSF ( ri
M ro ) exp [2 i vi ( ri
M ro )] d ro = M
2 0 ( vi ) . (1-95)
Writing the OTF in the form

r r r
[
0 ( vi ) = 0 ( vi ) exp i 7 ( vi ) ] , (1-96)
r r
where 0 ( vi ) and 7( vi ) are its modulus and the phase, called the modulation and phase
transfer functions (MTF and PTF) of the imaging system, respectively, and substituting
Eqs. (1-94) and (1-95) into Eq. (1-93), we obtain
r r r r r
{
Ii ( ri ) = Io 1 + m 0 ( vi ) cos 2 vi [ ri + 4
7 (vi )] } . (1-97)
Thus, like the Gaussian image, the diffraction image of a sinusoidal object of spatial
r r r
frequency vo is also sinusoidal with a spatial frequency vi = vo M , and its average
r
irradiance is Io . However, the modulation of this image is m 0 ( vi ) ; i.e., its modulation
r
is different from that of the object or its Gaussian image by the MTF factor 0 ( vi ) . It is
r
shown later [see Eq. (1-106)] that 0 ( vi ) ' 1. Thus, the modulation of the diffraction
image is always less than or equal to that of the object at any spatial frequency. The phase
of the sinusoidal image is also different from that of the object or its Gaussian image by
r r
the PTF 7( vi ) . Hence, the OTF 0 ( vi ) of an incoherent imaging system corresponding to
r
an image spatial frequency vi is a complex function whose modulus when multiplied by
r
the modulation of the corresponding sinusoidal object (of spatial frequency vo ) gives the
modulation of the sinusoidal image, and whose phase gives the phase of the diffraction
image relative to that of the object or its Gaussian image, as illustrated in Figure 1-6
r r r r

(Theorem 9). Note that vo ro = vi ri .
1.6.3 Properties of OTF
Now we derive certain properties of the OTF from its definition.
Theorem 10. (a) The OTF is a complex symmetric function with a value of unity at
the origin. (b) Its magnitude, the MTF, is unity or less at any other spatial frequency. (c)
The aberrated MTF at any frequency is less than or equal to the corresponding
aberration-free value.
34 IMAGE FORMATION
From Eq. (1-73b), we note that its value at the origin is unity, i.e.,
0(0) = 1 . (1-98)
This result can also be obtained from Eq. (1-71) since

r r
PSF ( ri ) d ri = 1 , (1-99)
as may be seen from Eqs. (1-61) and (1-63). Equation (1-98) represents the fact that the
OTF at zero spatial frequency is unity, i.e., the contrast of an image is zero for an object
of zero contrast. From Eq. (1-71), we also note that
r r
0 ( vi ) = 0 * (
vi ) , (1-100)
i.e., the OTF is complex symmetric or Hermitian (Theorem 10a). Therefore, its real part
is even and its imaginary part is odd; i.e.,
r r
Re 0 ( vi ) = Re 0 (
vi ) ,
(1-101)
and
r r
Im 0 ( vi ) =
Im 0 (
vi ) . (1-102)
From Eq. (1-73a) we note that

r r r r r
0 ( vi) = Pex
1 P ( rp ) P * ( rp
R vi ) d rp . (1-103)
r r
For two arbitrary but well-behaved functions f ( r ) and g( r ) , Hölder’s inequality 9
states that
r r r r r 1n
r n ( n
1) r ( n
1) n
f ( r ) g( r ) d r ' f ( r ) d r g( r ) dr
n

. (1-104)
It reduces to the more familiar Schwarz’s inequality when n = 2 . Letting f rp = P rp ,

r r r (r ) (r )
( ) ( )
g rp = P * rp
R vi , and n = 2 in Eq. (1-104), Eq. (1-103) may be written
r r r r r r 12
0 ( vi ) ' Pex
1
( ) ( ) d rp
2 2
P rp d rp P rp
R vi . (1-105)

Since both integrals are equal to Pex , we obtain

r
0 ( vi ) ' 1 . (1-106)
Thus, the MTF at any spatial frequency is less than or equal to unity (Theorem 10b).
(r ) (r ) (r
Using Hölder’s inequality with f rp = P rp P * rp
R vi , g( r ) = 1 , and n = 1,
r
) r
Eq. (1-103) yields
r r r r r
0 ( vi ) ' Pex
1 P (rp )P * (rp
R vi ) d rp
r r r r
= Pex
1 P ( rp ) P * ( rp
R vi ) d rp
r r r r
= Pex
1 A( rp ) A( rp
Rvi ) d rp
r
= 0 ( vi ) w = 0 ; (1-107)
i.e., the aberrated MTF at any spatial frequency is less than or equal to the corresponding
aberration-free MTF (Theorem 10c). The ratio of the aberrated and aberration-free MTFs
at a certain spatial frequency is called the Hopkins ratio for that frequency, as discussed
later in Section 1.10. Equation (1-107) shows that Hopkins ratio is less than or equal to
one.
The MTF of an aberration-free system with a nonuniform amplitude across its exit
pupil can be higher or lower than that for a uniform amplitude, depending on the spatial
frequency and the nautre of the amplitude distribution. For example, as discussed in
Section 4.3.3.2, the MTF of a Gaussian pupil is higher for low frequencies and lower for
the high. However, if the amplitude increases from the center of the pupil toward its edge,
then the MTF is lower for low frequencies and higher for the high.
We now summarize the results represented by Eqs. (1-98), (1-100), (1-106), and (1-
107) (Theorem 10).
i. OTF at the origin is unity.
ii. OTF is complex symmetric or Hermitian.
iii. MTF at any spatial frequency is less than or equal to unity.
iv. Aberrated MTF at any spatial frequency is less than or equal to the
corresponding aberration-free MTF; i.e., the Hopkins ratio is less than or
equal to one.
1.6.4 OTF Slope at the Origin

Theorem 11. The slope of the real part of the OTF at the origin is independent of the
aberration and it is equal to the corresponding slope of its MTF.
r
We consider the spatial frequency vector vi in rectangular coordinates (8, ,) and
polar coordinates (vi , 9) , where
( )
12
vi = 8 2 + ,2 (1-108)
and
9 = tan
1 ( , 8) . (1-109)
Thus, vi is the magnitude of the spatial frequency of a periodic image and 9 gives its
orientation, corresponding to rectangular frequency components 8 and ,. It should be
evident that the OTF corresponding to a two-dimensional periodic image of spatial
36 IMAGE FORMATION
frequencies 8 and , is equivalent to the OTF corresponding to a one-dimensional

periodic image of spatial frequency vi oriented at an angle 9 with the 8 axis.
Following Eq. (1-73a), we write the OTF corresponding to a spatial frequency (8, ,)
in the form
( ) (
0 (8, ,) = Pex
1 P x p , y p P * x p
R8, y p
R , dx p dy p ) , (1-110)
r
( )
where x p , y p are the rectangular coordinates of a pupil point with a position vector rp .
To determine, the OTF slope at the origin we write the integral in Eq. (1-110) in a (U, V )
(
coordinate system whose origin is shifted from the x p , y p coordinates system by )
( )
( R 2)(8, ,) but whose axes are parallel to the x p , y p axes, as indicated in Figure 1-7.
Thus, letting
xp = U + R8 2 (1-111a)
and
yp = V + R, 2 , (1-111b)
Eq. (1-110) can be written
0 (8, ,) = Pex
1 P(U + R8 2, V + R, 2) P * (U
R8 2, V
R, 2) dU dV . (1-112)
Equation (1-38) [or Eq. (1-47) with zi dependence suppressed] for the pupil function may
be written
yi
v
q

6i
6i–1 p
(R/2)(8,,)
9
yp
9
xi
u
xp
(–R/2)(8,,)
-m /0/)
I (1
oo
R6:
0/)
-m/
Ioo(1
(a) (b)
Figure 1-7. (a) Sinusoidal object yielding image of spatial frequency (vi , 9) or (8, ,) ,
mean irradiance I0 , and modulation m 0 . (b) Geometry for evaluating the OTF of
an optical imaging system. The centers of the two pupils are located at (0, 0) and
( )
R (8, ,) in the x p , y p coordinate system, m ( R 2) (8, ,) in the (U , V ) coordinate
system, and m ( R 2) (vi , 0) in the ( p, q ) coordinate system, where vi = 8 2 + ,2 ( )
12
and 9 = tan
1 ( , 8) . The shaded area is the overlap area of the two pupils.
( ) ( ) [ (
P x p , y p = A x p , y p exp i 2 x p , y p )] ,
(1-113)
or, utilizing Eqs. (1-111),
[
P(U + R8 2, V + R, 2) = A(U + R8 2, V + R, 2) exp i2(U + R8 2, V + R, 2) ] .
(1-114)
Substituting Eq. (1-114) into Eq. (1-112), the real part of the OTF may be written
[ ]
Re 0 (8, ,) = Pex
1 I (U , V ; 8, ,) cos Q(U , V ; 8, ,) dU d V , (1-115)
where
I (U, V ; 8, ,) = A (U + R8 2 , V + R , 2) A(U
R8 2, V
R , 2) (1-116)
and
Q(U , V ; 8, ,) = 2 (U + R8 2 , V + R , 2)
2 (U
R8 2 , V
R , 2) . (1-117)
The region of integration in Eq. (1-115) is the overlap area of two pupils centered at
m ( R 2)(8, ,) in the (U , V ) coordinate system. Expanding 2(U ± R8 2 , V ± R , 2) in
a Taylor series about the point (U , V ) , e.g.,
2 (U , V ) 2 (U , V )
2 (U + R8 2 , V + R , 2) = 2 (U, V ) + ( R 2) 8 +,
U V
1 2 2(U , V ) 2 2(U , V ) 2 2(U , V )

2
(1-118)
+ ( R 2 ) 2 8 2 + 2 8, + , + ... ,
2! U 2
U V V 2

we find that
2(U , V ) 2(U , V )
Q(U , V ; 8, ,) = R 8 +,
U V
3

( R 2)3 8 + , 2(U, V ) + ... .
1
+
3! U V (1-119)
Substituting Eq. (1-119) into Eq. (1-115) and noting that
x2 x4
cos x = 1
+
K , (1-120)
2! 4!
we may write
[ ( )]
Re 0 (8, ,) = Pex
1 I (U , V ; 8, ,) 1 + O 8 2 , ,2 , 8, dU d V . (1-121)
Thus, Re 0 (8, ,) does not consist of any aberration-dependent terms that depend linearly
on 8 or , alone. Hence, we obtain
38 IMAGE FORMATION

8 Re 0 (8, ,) = Pex
1 8 I (U , V ; 8, ,) dU d V , (1-122)
8 = , = 0 + 8 = , = 0 +
and a similar equation for the derivative with respect to h. Thus, the derivative of the real
part of the OTF with respect to a spatial frequency component x or h evaluated at the
origin is independent of the aberration, irrespective of the shape of the pupil. It does,
however, depend on the amplitude variations across the pupil. If the amplitude is
uniform, then the integral on the right-hand side reduces to the power contained in the
overlap region of the two pupils whose centers are separated by R(8, ,) from each
other. When divided by the total power Pex , it yields the fractional overlap area of the
two pupils.
From Eq. (1-101), we note that
Re 0 (8, ,) = Re 0(
8 ,
,) . (1-123)
Hence, the derivative of Re 0 (8, ,) with respect to 8 or , evaluated at the origin is

discontinuous; its sign depends upon whether the origin is approached from the first or
the third quadrant of the (8, ,) plane. In Eq. (1-122), the plus sign on the zeros indicates
that the origin is approached from the first quadrant. Note that the imaginary part
Im 0 (8, ,) of the transfer function depends on the sine of the aberration difference
function Q(U , V ; 8, ,) which, when expanded in a power series, will contain linear terms
in 8 and ,. Consequently, its derivative with respect to x or h evaluated at the origin
will depend on the aberration. As shown later in Section 1.9, this derivative determines
the centroid of the PSF. However, the derivative of the MTF 0 (8, ,) evaluated at the
origin is also independent of the aberration (Theorem 11). Indeed, the derivatives of
Re 0 (8, ,) and 0 (8, ,) evaluated at the origin are equal to each other. This may be seen
by evaluating the derivative of both sides of
0 (8, ,) [
= Re 0 (8, ,)] + [Im 0 (8, ,)]
2 2 2
(1-124)
at the origin and noting from Eq. (1-110) that Im 0(0, 0) = 0 .
Now we consider the OTF in polar coordinates (vi , 9) . We let ( p, q ) be a coordinate

system whose origin coincides with that of the (U , V ) coordinate system but whose axes
are rotated by an angle 9 with respect to it, as illustrated in Figure 1-7. Thus, in the
( p, q) coordinate system, the centers of the two pupils are located at m (R / 2) (vi , 0 ) .
(
Let P( p, q) be a function obtained from the pupil function P x p , y p by replacing x p )
with p cos 9
q sin 9 and y p with p sin 9 + q cos 9 . Then, the transfer function 0 (vi , 9)
may also be written
0 (vi , 9) = Pex
1 P( p + Rv i 2 , q ) P * ( p
Rvi 2, q ) dp dq . (1-125)
If we write P( p, q) in the form

[
P( p, q ) = A( p, q ) exp i2( p, q ) ] , (1-126)
the OTF may be written

r r
[ ]
0 (vi , 9) = Pex
1 I ( p, q; vi ) exp iQ( p, q; vi ) dp dq , (1-127)
where the mutual irradiance function is given by

r
I ( p, q; vi ) = A( p + Rvi 2, q ) A( p
Rvi 2, q ) (1-128a)
and the phase aberration difference function is given by

r
Q( p, q; vi ) = 2( p + Rvi 2 , q )
2( p
Rvi 2 , q ) . (1-128b)
Its real part can be written

r r
[ ]
Re 0 (vi , 9) = Pex
1 I ( p, q; vi ) cos Q( p, q; vi ) dp dq . (1-129)
Expanding 2( p ± Rvi 2 , q ) in a Taylor series about the point ( p, q) , Eq. (1-128b)

yields
r
[
Q( p, q; vi ) = 2 ( Rvi 2) 2 ( p, q ) + (1 3!) ( Rvi 2) 2 ( p, q ) + K
3
] , (1-130)
where the primes on 2 indicate partial derivatives of 2( p, q ) with respect to p. Using

the power series expression for a cosine function, we can write
r
[ ]
cos Q( p, q; vi ) = 1
1
2
[
( Rvi )2 2 ( p, q) ]2 + O (vi4 ) . (1-131)

r
Re 0 (vi , 9) = Pex
1 I ( p, q; vi ) [1 + O (vi 2 )] dp dq . (1-132)
Taking the derivative of both sides of Eq. (1-132) with respect to vi and evaluating at
vi = 0, we obtain
r
Re 0 (vi , 9) = Pex
1 I ( p, q; vi ) dp dq . (1-133)

iv vi = 0
iv vi = 0
That is, the derivative of the real part of the transfer function with respect to the
radial frequency v evaluated at the origin is independent of the aberration, irrespective of
the shape of the pupil. It does, however, depend on the amplitude variations across the
pupil. This result is to be expected from the component derivatives given by Eq. (1-122).
As discussed later in Section 1.7, the asymptotic behavior of the irradiance distribution of
the image of a point object depends on the value of this derivative. Once again, the
imaginary part of the transfer function depends on the sine of the aberration difference
40 IMAGE FORMATION
which, when expanded in a power series, will contain a linear term in vi . Consequently,
its derivative with respect to vi evaluated at vi = 0 will depend on the aberration.
However, the derivative of the modulus of the transfer function evaluated at vi = 0 is also
independent of the aberration. Indeed, the derivatives of the real part and the modulus of
the transfer function evaluated at vi = 0 are equal. This may be seen by evaluating the
derivative of both sides of 0 = ( Re 0)2 + (Im 0)2 at vi = 0 , and noting from Eq. (1-
2
125) that the imaginary part Im 0 is zero at vi = 0.
We summarize the results of this section in that the derivative of the real part of the
OTF of a system with respect to a spatial frequency component 8 or , or a radial
frequency vi evaluated at the origin is independent of its aberration irrespective of the
shape of its exit pupil. The value of this derivative is equal to the corresponding
derivative of the MTF of the system.
1.6.5 OTF in the Limit of Zero Wavelength

Now we derive a theorem that the image based on geometrical optics is a limiting
case of the diffraction image.
Theorem 12. The diffraction OTF approaches the geometrical OTF as the
wavelength approaches zero.
Substituting Eq. (1-114) into Eq. (1-112), the diffraction OTF can be written
{ }
0 (8, ,) = Pex
1 I (U , V ; 8, ,) exp iQ(U , V ; 8, ,) dU d V , (1-134)
where the region of integration is the overlap area of two pupils centered at
m( R 2)(8, ,) . Substituting Eq. (1-75) into Eq. (1-119), we obtain
W (U , V ) W (U , V )
Q(U , V ; 8, ,) = 2 R 8
U
+,
V
+ O 2 ( ) , (1-135)
( )
where O 2 consists of terms with third and higher odd-order derivatives of W (U , V ) .
Substituting Eq. (1-135) into Eq. (1-134) and letting 0, we obtain
W (U , V ) W (U , V ) $
[0 (8, ,)] 0 = Pex
1 A 2 (U , V ) exp !2 iR 8
U
+,
V &
% dU d V , (1-136a)
#
or
[0 (8, ,)] 0 (
= Pex
1 I x p , y p )
" W x p , y p
× exp !2i R 8 +,
( )
W x p , y p( ) $" dx
% p d yp , (1-136b)
"# x p y p "
&
where
( ) (
I x p , y p = A2 x p , y p ) (1-136c)
( )
is the irradiance at a pupil point x p , y p , and the region of integration is the area of the
exit pupil. Equation (1-137) may also be obtained directly from Eq. (1-110) in a similar
manner. However, the equation corresponding to Eq. (1-135) in that case will consist of
terms with odd and even-order derivatives of W x p , y p . ( )
We show in the next section that the OTF obtained according to geometrical optics is
also given by Eq. (1-136b). Hence, it yields the result that the diffraction image reduces
to the geometrical image as the wavelength approaches zero (Theorem 12). It should be
noted that, in practice, the limit of zero wavelength is equivalent to a very large
aberration. This is why the diffraction PSF may be expected to be similar to the
geometrical PSF for large aberrations.
1.6.6 Geometrical OTF10

Now we consider the OTF based on geometrical optics and show that it is equal to
the diffraction OTF in the limit of zero wavelength. The aberration-free image based on
geometrical optics is a point or a Dirac delta function, regardless of the amplitude
distribution across the pupil. Hence, the corresponding OTF is unity for all frequencies.
We obtain an approximate expression for the aberrated OTF that is valid for low spatial
frequencies, and show that its slope at the origin is zero.
The geometrical PSF (not the Gaussian image) is given by [see Eq. (4-10) of Part]

1
( x, y)
(
I g ( x, y) = I x p , y p ) , (1-137)
(
x p , yp )
( )
where x p , y p are the coordinates of a ray in the plane of the pupil and ( x, y) are its
coordinates in the image plane with respect to the Gaussian image point. Its Fourier
transform for unit total power gives the geometrical OTF:
[
0 g (8, ,) = Pex
1 Ig ( x, y) exp 2 i (8 x + ,y) dx d y ] (1-138a)

1

( x, y)
= Pex
1 (
I x p , yp ) [ ]
exp 2 i (8 x + ,y) dx d y

(
x p , yp )
( ) [
= Pex
1 I x p , y p exp 2 i (8 x + ,y) dx p d y p ] . (1-138b)
Note that it is properly normalized since 0 g (0, 0) = 1. Like the diffraction OTF, it is also
Hermitian.
Substituting for the ray aberrations ( x, y) in terms of the slope of the aberration
function [see Eq. (3-11) of Part I], namely,
42 IMAGE FORMATION
( x, y) = R
( ,
)
W x p , y p W x p , y p ( ) , (1-139)
x p y p

we obtain

" W x p , y p ( )
W x p , y p ( ) $" dx
0 g (8, ,) =

( )
Pex
1 I x p , y p exp !2 i R 8
x p
+,
y p
%
"
p d yp ,
"# &
(1-140)
which is the same as Eq. (1-136b). The refractive index of the image space is assumed to
be unity here, which is generally the case in practice. If this is not true, then the right-
hand side of Eq. (1-139) is divided by the refractive index. Also in that case, the
wavelength of the object radiation is divided by the refractive index in all of the
diffraction equations.
To obtain an approximate expression for the geometrical OTF for small spatial
frequencies, we write Eq. (1-138b) in terms of the centroid ( xc , yc ) of the aberrated PSF
discussed later [see Eq. (1-177)] in the form
[
0 g (8, ,) = Pex
1 exp 2 i(8 xc + ,yc ) ]
( ) { [
× I x p , y p exp 2i 8( x
xc ) + ,( y
yc ) dx p d y p ]} . (1-141)
Expanding the exponential under the integral in a power series and retaining only the first
three terms for small values of (8, ,) (the second term vanishes upon integration), we
obtain
0 g (8, ,) ~ Pex
1 exp[2 i(8 xc + ,yc )]
(
× I x p , yp ) {1
2 2 [8( x
xc ) + ,( y
yc )] 2 } dx p d y p . (1-142)
Writing 0 g in terms of its modulus 0 g and phase 7g in the form
0 g (8, ,) = 0 g (8, ,) exp i7g (8, ,) [ ] (1-143)
and comparing Eqs. (1-142) and (1-143), we obtain
~ 1
2
2
0 g (8, ,) I ( x p , y p ) [8( x
xc ) + ,( y
yc )]
2
dx p d y p (1-144a)
Pex
and
7g (8, ,) ~ 2 (8 xc + ,yc ) . (1-144a)

We note that the geometrical PTF depends on the pupil irradiance only through the
centroid of the PSF. The PTF for a symmetric aberration is zero, as expected, since the
centroid for such an aberration lies at the origin.
It is evident from Eq. (1-144a) that the slope of the geometrical MTF at the origin is
zero:
0 g (8, ,) 0 g (8, ,)
= 0 = . (1-145)
8 ,
8 = 0 = , 8 = 0 = ,
This result can also be obtained directly from Eq. (1-138b). As shown in Section 1.6.4,
the slope of the diffraction MTF at the origin, which is equal to the corresponding slope
of the real part of the OTF, is nonzero.
If we write the wave and ray aberrations in the ( p, q ) coordinate system, Eq. (1-
138b) reduces to
( )
0 g (vi , 9) = Pex
1 I ( p, q) exp 2 ivi x9 dp d q , (1-146a)
where
W ( p, q )
x9 = R (1-146b)
p
is the component of the ray aberration along the p axis. It may also be obtained from the
ray aberration ( x, y) according to
x 9 = x cos 9 + y sin 9 . (1-147)
Equation (1-150) may also be written
( ) [ (
0 g (vi , 9) = Pex
1 exp 2 ivi x 9 I ( p, q ) exp 2 ivi x9
x9 )] dp d q , (1-148a)
where
x9 = Pex
1 I ( p, q ) x9 dp d q (1-148b)
is the mean value of x 9 . For small values of the radial spatial frequency vi , we expand
the exponential under the integral in Eq. (1-146a) and retain only the first three terms (the
second term vanishes upon integration) and obtain
( )
2
0 g (vi , 9) ~ 1
2 2 vi2 x9
x9 (1-149a)
and
44 IMAGE FORMATION
7(vi , 9) ~ 2 vi x9 . (1-149b)
Thus, the variance of the ray aberration along the p axis determines the reduction of the
MTF along that axis, and its mean value determines the PTF. It is evident from Eq. (1-
149a) that the slope of the MTF 0 g (vi , 9) with respect to the radial frequency vi
evaluated at the origin is zero.
1.6.7 Comparison of Diffraction and Geometrical OTFs

There are three significant differences between the diffraction and geometrical OTFs.
First, the integration in Eq. (1-127) or Eq. (1-134) for the diffraction OTF is over the
r
common region of two pupils with their centers separated by Rvi , but it is over the
whole pupil in Eq. (1-143) for the geometrical OTF. Second, the mutual irradiance
( )
I (U , V ; 8, ,) given by Eq. (1-116) is different from the pupil irradiance I x p , y p , unless
the amplitude across the pupil is uniform. Third and finally, the exponents under the
integrals in the OTF equations are different from each other, unless the third- and higher-
order derivatives of the aberration function are zero, as in the case of defocus and or
astigmatism. Thus, we expect the diffraction OTF for defocus and or astigmatism for very
low spatial frequencies (so that the overlap region is approximately equal to that of the
pupil) to be approximately equal to the geometrical OTF. The difference between the two
OTFs is expected to be small for low spatial frequencies even for the higher-order
aberrations, because terms beyond the first-order derivatives are less significant and the
areas of integration less different compared to the case of high frequencies. However,
since the slope of the geometrical OTF at the origin is zero while that of the diffraction
OTF is not, the former will overestimate the true value of the OTF at very low
frequencies. Photographic systems have relatively large aberrations and, therefore, very
low MTF at high frequencies. Hence, from a practical standpoint, one is interested
primarily in the OTF for such systems at low frequencies. Accordingly, the calculation of
the geometrical OTF, which is simpler than the corresponding diffraction calculation, is
beneficial, at least in the early stages of the design of such an optical system. Of course,
the geometrical OTF does not have a cutoff frequency but the diffraction OTF does.
It should be noted that the diffraction OTF depends strongly on the wavelength of the
object radiation. The cutoff frequency, for example, is inversely proportional to the
wavelength. Since the wave aberration of a ray as an optical path length error in units of
wavelength will be different for different wavelengths, the aberrated OTF will also vary
with the wavelength. However, the geometrical OTF depends on the ray aberrations,
which are independent of the wavelength as long as the optical path length errors are
independent of it. If the path length errors of a system vary with the wavelength, i.e., if it
has chromatic aberration, then the ray aberrations will also vary with it. Thus, for
example, the diffraction OTF of a system consisting of mirrors will vary with the
wavelength, but the geometrical OTF will not. In a system using lenses and suffering
from chromatic aberration, there will be some variation of the geometrical OTF as well
with the wavelength.
1.7 Asymptotic Behavior of PSF 45
1.6.8 Determination of OTF

A direct measurement of the OTF of a system requires sinusoidal objects of varying
spatial frequencies. Such objects are hard to fabricate in practice; square wave objects are
easier, for example. However, since OTF is the Fourier transform of PSF, an obvious
approach is to measure the latter and calculate its transform. In practice, the PSF is quite
small in its extent and its measurement requires an array of very small and sensitive
photodetectors. Even so, the noise of the detectors may be larger than the PSF values in
the region outside its central bright spot. An alternative is to measure the aberration
function of the system and calculate its OTF by autocorrelation of its pupil function. One
may also measure the line-spread function (discussed in Section 1.11) for various angular
orientations of a line object and calculate the OTF profiles by Fourier transforming them
according to Eq. (1-228). It may also be obtained from a measurement of the edge-spread
function according to Eq. (1-229).
1.6.9 Significance of PTF

Although the OTF is a complex function with real and imaginary parts, or a modulus
MTF and a phase PTF, often only the modulus is measured in practice and the phase is
ignored. This is adequate for an aberration-free system, or for systems with symmetric
pupil functions, as in the case of spherical aberration or astigmatism, since the OTF in
those cases is real (see Problem 5). However, the PSF and the LSF and ESF of a system
(discussed in Section 1.11) with a complex OTF, e.g., a system aberrated by coma, cannot
be obtained from its MTF. Both the real and imaginary parts of the OTF are needed to
determine these spread functions. The real part is needed to determine the Strehl ratio
(discussed in Section 1.9), i.e., the value of the PSF at the origin. However, the imaginary
part is needed to determine the value of the ESF at the origin. The centroid of the PSF
(discussed in Section 1.8) is determined by the slope of the imaginary part of the OTF at
the origin. While the MTF determines the contrast of a sinusoidal component of a certain
spatial frequency in the image, the PTF determines the phasing of those components as
they are added to obtain the net image. Hence, a knowledge of the PTF is essential to
fully characterize the image.
1.7 ASYMPTOTIC BEHAVIOR OF PSF11

Now we discuss the asymptotic behavior of the PSF of an imaging system, i.e., its
value at large distances from its center. We show that, up to the first order, the PSF at
large distances depends on the slope of the OTF at the origin. Since, as shown above, this
slope is independent of the aberration of the system, therefore, up to the first order, the
asymptotic value of the PSF is also independent of the system aberration. Similarly, the
asymptotic value of the encircled power is independent of the aberration. We also show
that, for a system with an unapodized (i.e., a uniformly illuminated) exit pupil, the
asymptotic behavior of encircled power is related simply to the ratio of the perimeter
length and the area of the exit pupil.
46 IMAGE FORMATION
1.7.1 Point-Spread Function

Now we derive the following theorem on the asymptotic behavior of the PSF.
Theorem 13. The PSF of a system with a radially symmetric pupil function behaves
asymptotically as the inverse cube of the distance from its center, independent of the
aberration.
For an imaging system with a radially symmetric pupil function, its PSF may be
obtained from its OTF according to [see Eq. (1-84)]
PSF (ri ) = 2 0 (vi ) J 0 (2 ri vi )vi dvi , (1-150)
where 0 (vi ) is a real function. Willis12 has shown that for a function f ( x ) whose
derivatives exist
;
1 f (0) 3 f iv (0)
m f ( x ) J 0 ( mx ) dx ~ f (0)
2 + 2 K (1-151)
2 m 2 2! m 4
0
for large values of m, where
[
f ( n ) (0) = n f ( x ) x n ] x=0
. (1-152)
Letting x = vi , f (vi ) = vi 0 (vi ) and m = 2 ri , Eq. (1-151) can be written for large values
of ri :
0 ( 0 ) 30 (0)
PSF (ri ) ~
2 3 +
... . (1-153)
4 ri 32 4 ri5
Equation (1-153) gives an asymptotic representation of the PSF in terms of the properties
of its Fourier transform, the OTF, at the origin.
For large values of ri , we may neglect higher-order terms and write
PSF (ri ) ~
0 (0) 4 2 ri3 . (1-154)
However, as shown in Section 1.6.4, 0 (0) is independent of aberration. Hence, for large
values of ri , the PSF is independent of an aberration and varies with ri as ri
3 (Theorem
13). Tatian’s results for circular pupils show that Eq. (1-154) holds even for rotationally
nonsymmetric aberration.13 This result is plausible, because the effect of an aberration is
to reduce the irradiance at the center of the image and increase it at nearby points. This
does not imply that the minimum value of ri above which the PSF is (approximately)
described by an ri
3 dependence is the same whether or not the imaging system is
aberrated, but that for large enough values of ri , the point-spread function depends on ri
through ri
3 even when the system is aberrated.
1.7.2 Encircled Power

Now we derive the following theorems on the asymptotic behavior of encircled
power.
Theorem 14. The encircled power for large circles is independent of the aberration.
Theorem 15. The asymptotic behavior of encircled power of an unapodized system is

determined by the ratio of the perimeter length and the transmitting area of its exit pupil.
The encircled power given by Eq. (1-86) may be written

rc 2
Pi (rc ) = I (ri , i ) ri dri d i , (1-155)
0 0
where I (ri , i ) is the irradiance at a point (ri , i ) in the image plane. Following Eqs. (1-
61) and (1-76), we may substitute for I (ri , i ) in terms of the corresponding transfer
function 0 (vi , 9) , and thus write Eq. (1-155) in the form
rc 2 2
Pi (rc ) = Pex ri dri d i vi dvi
0 0
0 (vi , 9) exp [
2 iri vi cos (i
9)] d9 . (1-156)
0
Changing the order of integration, the integration over i gives 2 J 0 (2 ri vi) , and noting
that
x0
J ( mx ) x dx = x 0 J ( mx ) ,
0 1 0 (1-157)
m
0

the integration over ri gives (rc vi ) J1 (2 rc vi ) , where J1 ( ) is the first-order Bessel
function of the first kind. Thus, Eq. (1-156) reduces to
2
Pi (rc ) = Pex rc d9 0 (vi , 9) J1 (2 rc vi ) dvi . (1-158)
0
Since the left-hand side is real, the integral over the imaginary part of 0 (vi , 9) must be
zero. Hence, we may consider only the real part of 0 (vi , 9) in Eq. (1-158) and write it in
the form
2
Pi (rc ) = Pex rc d9 Re 0 (vi , 9) J1 (2 rc vi ) dvi .
0 (1-159)
Willis12 has given another formula

;
f ( 0 ) 1 f (0) 1 3 f v (0)
m f ( x ) J1 ( mx ) dx ~ f (0) +
+
K (1-160)
m 2 m3 2 2 2! m 5
0
48 IMAGE FORMATION
for large positive values of m. Therefore, for large values of rc , Eq. (1-159) can be
written
2
" Re 0 (0, 9)
Pi (rc ) ~ Pex d9 ! +
Re 0 (0, 9)

[ ] [
Re 0 (0, 9)
+ ...
$"
% ,
] (1-161)
" 2 4 2 rc 32 3r 3c "&
#
0
where, for example,
[Re 0 (0, 9)] (n) = [( n

)
vin Re 0 (vi , 9) ]vi = 0
(1-162)
is the nth radial derivative of Re 0 (vi , 9) evaluated at the origin. Equation (1-162) gives
an asymptotic representation of the encircled power in terms of the radial derivatives of
the real part of the OTF at the origin. Since the transfer function is normalized to unity at
the origin, the first term on the right-hand side is Pex . Note that the equation is properly
normalized since Pi (rc ;) = Pex . For large values of rc , we may neglect higher-order
terms and write
2
Pi (rc ) (
~ Pex 1 + 1 4 2 rc

) [Re 0 (0, 9)] d9 . (1-163)
0
Since the integrand in Eq. (1-163) is independent of aberration [see Eq. (1-133)], we
conclude that, up to the first order, the encircled power for large circles is also
independent of an aberration (Theorem 14).
(r )
We now show that, for an unapodized system [i.e., one for which A rp = A0 across
its exit pupil], the integral on the right-hand side of Eq. (1-163), and therefore the asymp-
totic behavior of the encircled power, is related simply to the ratio of the perimeter length
and the transmitting area of the exit pupil.14 From Eq. (1-73b), we note that the OTF of an
r
unapodized aberration-free system corresponding to a spatial frequency vi is equal to the
r
fractional area of overlap of two pupils whose centers are separated by R vi . Thus, from
Figure 1-8, we note that the OTF may be written
r
0 ( vi ) = 1
( R 2 Sex ) vi cos ds , (1-164)
s
where ds is a differential length segment along the perimeter of the pupil and is the
r
angle between vi and the normal to the segment. The factor of 2 accounts for the fact that
the nonoverlap area is counted twice in the integral in Eq. (1-164). For very small values
of vi (since we are interested in the slope of the OTF at the origin), it may be considered
constant and brought outside the integral. Hence, for such values, we may write
r
0 ( vi ) = 1
( Rvi 2 Sex ) cos ds . (1-165)
s
,
ds cos
n^
O
R6 i 9
8
O
Figure 1-8. Geometry for calculating the OTF of a system for very small spatial
r
frequencies. The displaced pupil is centered at O , whose position vector is Rvi
making an angle 9 with respect to the pupil centered at O. Note that n̂ is a unit
vector along the normal to the pupil at the point where the line joining O and O
intersects it. The overlap area of the two pupils is shown shaded.
Since d = d9 , we obtain
2 2
0 ( vi , 9) d9 = 2
( RLvi 2 Sex ) cos d
0 0
= 2
2 RLvi Sex , (1-166)
where L is the length of the perimeter of the exit pupil. Taking the derivative of both sides
with respect to vi , we obtain for very small values of vi and in the limit vi 0 :
2
2 RL
0 (0, 9) d9 =
S . (1-167)
0 ex
Since the slope of the real part of the OTF at the origin is independent of aberration, Eq.
(1-167) is valid for aberrated but unapodized systems as well. Substituting this equation
into Eq. (1-163), we obtain
(
P (rc ) ~ Pex 1
RL 2 2 rc Sex ) . (1-168)
Thus, we see that the asymptotic behavior of encircled power depends on the ratio of the
perimeter length L of the exit pupil and its transmitting area Sex (Theorem 15). The
effect of an aberration is to increase the value of rc for which Eq. (1-168) is valid. The
larger the aberration, the larger the value of rc required for the validity of Eq. (1-168).
50 IMAGE FORMATION
1.8 PSF CENTROID15

In this section, we derive expressions for the centroid of the light distribution of the
image of a point object formed by an aberrated imaging system in terms of its PSF, OTF,
and aberration function. We derive the following theorems.
Theorem 16. (a) The centroid of the diffraction PSF is given by the slope of the
imaginary part of the its diffraction OTF at the origin. (b) It is the same as the centroid of
the ray geometrical PSF. (c) The centroid of an aberration-free system lies at the
Gaussian image point regardless of the amplitude variations across its exit pupil. (d) For
a system with an aberrated but unapodized exit pupil, the centroid can be obtained from
its aberration only along its perimeter.
1.8.1 Centroid in Terms of OTF Slope

From Eq. (1-62), the irradiance distribution of the image of a point object may be
written
2
2i
Ii ( xi , yi )
1
(
= 2 2 P x p , y p exp
R R
) (
x p xi + y p yi dx p dyp

) , (1-169)
where ( xi , yi ) are the coordinates of a point in the image plane with respect to the
( )
Gaussian image point, and x p , y p are the coordinates of a point in the plane of the exit
pupil. By dividing both sides of Eq. (1-169) by Pex we obtain the function PSF ( xi , yi )
for the system.
By definition, the coordinates of the centroid of the image are given by
xi = Pex
1 xi Ii ( xi , yi ) dxi dyi (1-170a)
and
yi = Pex
1 yi Ii ( xi , yi ) dxi dyi . (1-170b)
From the definition of the PSF of an imaging system, it should be evident that its centroid
is synonymous with that of the irradiance distribution of the image of a point object.
From Eqs. (1-61) and (1-71), the OTF may be written
0 (8, ,) = Pex
1 Ii ( xi , yi ) exp [2 i (8 xi + ,yi )] dxi dyi , (1-171)
r
where (8, ,) are the rectangular components of a spatial frequency vector v in the image
plane. Differentiating both sides of Eq. (1-171) with respect to x and evaluating the result
at 8 = , = 0 , we find that
1 0
xi = . (1-172a)
2 i 8 8 = , = 0
1.8 PSF Centroid 51
Similarly,
1 0
yi = . (1-172b)
2 i , 8 = , = 0
However, since xi and yi are real, only the slope of the imaginary part of the OTF at
the origin contributes to the centroid. Thus, we may write
1 Im 0
xi = (1-173a)
2 8 8 = , = 0
and
1 Im 0
yi = . (1-173b)
2 , 8 = , = 0
Thus, the centroid of the PSF of an optical system is given by the slope of the imaginary
part of its OTF at the origin (Theorem 16a). It can not, for example, be obtained from a
knowledge of only the MTF of the system.
1.8.2 Centroid Related to Wavefront Slope
The OTF is also given by Eq. (1-73a), which may be written
( ) (
0 (8, ,) = Pex
1 P x p , y p P * x p
R8 , y p
R , dx p dy p ) . (1-174)
Substituting Eq. (1-174) into Eqs. (1-173), we obtain
R P * x p , y p ( ) dx
xi =
Im P x p , y p
2 Pex
( x p
)
p dy p (1-175)

and a similar equation for yi . The pupil function given by Eq. (1-38) or Eq. (1-47) may
be written in the form
( ) ( ) [
P x p , y p = A x p , y p exp (2 i ) W x p , y p ( )] . (1-176)
W x p , y p ( )
xi =
R
Pex

(
Ip x p , yp )
x p
dx p dy p , (1-177a)

( ) ( )
where I p x p , y p = A 2 x p , y p is the irradiance at a pupil point x p , y p . Similarly, ( )
W x p , y p ( )
yi =
R
Pex

I x ,y
p p p ( y p
) dx p dy p . (1-177b)

52 IMAGE FORMATION
( ) ( )
From Eq. (1-142), R W x p and R W y p represent the ray aberrations, i.e., the
( )
image-plane coordinates of a ray passing through the pupil point x p , y p . Hence, Eqs.
(1-177) show that the centroid of the PSF according to wave diffraction optics is identical
with that according to ray geometrical optics (Theorem 16b).
From Eqs. (1-177), we also note that amplitude variations across the pupil affect the
centroid only if it is aberrated. In the absence of aberrations, the PSF centroid lies at (0,
0); i.e., it lies at the Gaussian image point where the center of curvature of the spherical
wavefront lies, regardless of the shape of the pupil and/or the amplitude distribution
across it (Theorem 16c). This may also be seen from Eqs. (1-169) and (1-176). We note
( )
from these equations that if W x p , y p = 0 , then Ii ( xi , yi ) = Ii (
xi ,
yi ) . Hence, the
symmetry of the aberration-free PSF yields its centroid at the Gaussian image point.
Similarly, since in that case the aberration-free OTF is real [see Eq. (1-174)], Eqs. (1-173)
also give the centroid at the Gaussian image point.
Equations (1-170), (1-173), and (1-177) give the centroid in terms of the PSF, OTF,
and the aberration function, respectively. In practice, given an imaging system, the most
convenient expression to use would be Eqs. (1-170), since the PSF can be measured by
using a photodetector array. In optical design and analysis, the simplest way to obtain the
centroid would be to use Eqs. (1-177) since the aberrations must be calculated even if the
other two expressions were used. Thus, one may trace rays all the way up to the image
plane and determine the centroid of the ray distribution in this plane with appropriate
( )
weighting I p x p , y p of each ray.
1.8.3 Centroid Related to Wavefront Perimeter
Unapodized (or Uniform) Pupil

If the pupil is unapodized, e.g., if the amplitude
( )
A x p , y p = A0 (1-178)
so that the irradiance
( )
I p x p , y p = A02
(1-179)
= Pex Sex ,
then Eqs. (1-177) reduce to
=
R

(
W x p , y p )

xi dx p dy p (1-180a)
Sex x p

and
=
R

(
W x p , y p )

yi dx p dy p . (1-180b)
Sex x p

1.9 Strehl Ratio 53
Using Stokes theorem,16 the surface integrals in Eqs. (1-180) involving the derivative of
the aberration function can be written in terms of its line integral along the curve
bounding the surface. Thus, we may write
r
W ( x , y ) xˆ d s
R
xi = p p p (1-181a)
Sex
and
r
yi
( )
= ( R Sex ) W x p , y p yˆ p d s , (1-181b)
r
where x̂ p and ŷ p are unit vectors along the x p and y p axes, respectively, and d s
represents an element of arc length vector along the perimeter of the pupil. It is evident
from Eqs. (1-181) that, in the case of an aberrated but an unapodized pupil, the centroid
of the PSF can be obtained from the value of the aberration function only along the
perimeter of the pupil (Theorem 16d). Accordingly, in that case, to calculate the centroid
the knowledge of the aberration across the interior of the pupil is not needed.
1.9 STREHL RATIO17-21

It is evident from the foregoing that the imaging properties of a system are
determined by its PSF or, equivalently, by its OTF. In this section, we derive some
general results on the effects of nonuniform amplitude, called apodization, and
nonuniform phase, called aberration, at the exit pupil (i.e., across the reference sphere)
on the irradiance at the center of the reference sphere. We show that, for a fixed total
power going into the image of a point object, maximum central irradiance is obtained for
a system with an unapodized and unaberrated pupil. It is also shown that the peak value
of an unaberrated image lies at the center of curvature of the reference sphere regardless
of the apodization of the pupil. For a given total power in the pupil (and, therefore, in the
image), the value of the central irradiance for an apodized and aberrated pupil relative to
its value for an unapodized and or unaberrated pupil is called the Strehl ratio of the image
and provides a measure of its quality. Approximate but simple expressions are obtained
for the Strehl ratio in terms of the variance of the aberration across the pupil. A brief
discussion of how one may determine the Strehl ratio of an image in practice is also
given.

We derive the following theorems related to the Strehl ratio.
Theorem 17. (a) The central irradiance for an apodized-aberrated system is less
than or equal to the corresponding value for an unapodized-unaberrated system. (b) For
a given total power, any amplitude variations reduce the central irradiance and any
phase variations further reduce it. (c) The peak value of an unaberrated PSF lies at the
center, regardless of the amplitude variations across the pupil.
54 IMAGE FORMATION
The irradiance distribution of the (defocused) image of a point object is given by Eq.
r
(1-60). For simplicity of notation, we let ri be the position vector of the observation point
r
( )
with respect to the point zi zg rg lying on the line joining the center of the exit pupil
and the Gaussian image point in the observation plane at a distance zi . Thus, the
irradiance distribution given by Eq. (1-60) may be written
2
r 1 r 2 i r r r
I ( ri ; zi ) = 2 2 P rp ; zi exp
zi zi
(rp r d rp

) , (1-182)
(r )
where P rp ; zi is the defocused pupil function given by Eq. (1-47). The irradiance at
the center of the distribution for an apodized and aberrated pupil is obtained by letting
r
ri = 0 . Thus,
r r
P ( rp ; zi ) d rp
1 2
Iaa (0; zi ) = . (1-183)
2 zi2
Similarly, for the same total power Pex in the image, the central irradiance at the
Gaussian image point for an unapodized and unaberrated system may be obtained from
r r r
( )
Eq. (1-183) by letting zi zg , A rp = A0 and W rp = 0 , so that P rp = A0 . Thus, ( ) ( )
we may write
(
Iuu 0; zg = ) Pex Sex
2 zg2
(1-184)
r r
( )
Sex 2
= P rp ; zi d rp . (1-185)
2 zg2
The ratio of the central irradiance in the defocused plane for an apodized and
aberrated pupil, and in the Gaussian image plane for the unapodized and unaberrated
pupil may be written
Iaa (0; zi )
2
zg
= Saa , (1-186)
(
Iuu 0; zg ) zi
where
r r
P ( rp ) d rp
2
Saa = r r (1-187)
( )
2
Sex P rp d rp
is the corresponding Strehl ratio. It represents the ratio of the central irradiances in the
defocused and Gaussian image planes, except for the inverse-square-law effect
( )
2
represented by the factor zg zi . This factor is unity if the aberrated irradiance is also
observed in the Gaussian image plane. It is practically equal to unity in imaging systems
because of their small depth of focus. However, it can be significantly different from
1.9 Strehl Ratio 55
(r ) (r ) (r )
unity in laser transmitters. Letting f rp = P rp , g rp = 1, and n = 2 in Hölder’s
inequality (1-104), we find that
Saa ' 1 . (1-188)
Thus, the Strehl ratio for an apodized-aberrated system compared to an unapodized-

unaberrated system is less than unity (Theorem 17a). Equation (1-186) shows that when a
beam of light is focused at a certain distance zg , the central irradiance at a distance
zi < zg is higher due to the inverse-square law effect, but lower due to the Strehl ratio.
Thus, the inverse-square law and the nonconstructive interference of the Huygens’
spherical wavelets compete with each other in determining the irradiance value.
Accordingly, the defocused irradiance for zi < zg can be higher than the corresponding
focal-point value. This is indeed the case in laser transmitters, as discussed in Section
2.5.2.
When the system is apodized but unaberrated, i.e., if
(r )
P rp = A rp (r ) , (1-189)
the central irradiance is given by
( ) ( ) [ A (rr ) d rr ]
2
Iau 0; zg = 1 2 zg 2 p p . (1-190)
The corresponding Strehl ratio is given by
(
Sau = Iau 0; zg ) Iuu (0; zg )
[ A (rrp ) rd rrp ]r .
2
= (1-191)
Sex A 2 ( rp ) d rp
Following the same argument as for Eq. (1-187), we find that

Sau ' 1 . (1-192)
Thus, any amplitude variations across the pupil of an aberration-free system reduce the
central irradiance. For example, as discussed in Section 4.3, a pupil with a Gaussian
illumination across it yields a smaller value of the central irradiance than a pupil with a
uniform illumination, the total power being the same in the two cases.
Comparing Eq. (1-191) with Eq. (1-187), and using Hölder’s inequality (1-104) with
r r r
( ) ( ) ( )
f rp = P rp , g rp = 1, and n = 1, we find that
r r 2
Saa P ( rp ) d rp
=
[ A (rrp ) d rrp ]
2 (1-193a)
Sau
' 1 . (1-193b)
56 IMAGE FORMATION
From Eqs. (1-192) and (1-193b), we note that amplitude variations reduce the central
irradiance, and phase variations (i.e., aberrations) further reduce it. Note, however, that
an irradiance reduced by phase variations alone does not necessarily reduce any further
if amplitude variations are also introduced (Theorem 17b). In fact, amplitude variations
can even increase this irradiance. The maximum value of central irradiance is obtained
when the system is unapodized and unaberrated.18 Hence, the maximum value of
irradiance at a certain point in the image space is obtained when a beam exiting from the
system is focused at it with uniform amplitude and phase at its exit pupil. Any variation
in the amplitude or phase across the exit pupil reduces the value at the chosen point.
The peak value of the aberrated irradiance distribution of the image of a point object
does not necessarily occur at the center of the reference sphere. However, the peak value
of its unaberrated image does occur at this point regardless of the apodization. The
Huygens’ spherical wavelets emanating from the spherical wavefront are equidistant from
this point. Hence, they interfere constructively, producing a maximum possible value at
r r
( ) ( )
this point. Mathematically, this may be seen by letting zi = zg , and P rp = A rp in
r r
Eq. (1-182), and comparing the irradiance at a certain point ri with that at ri = 0 . Thus,
we may write
r r r
f ( rp ) d rp
2
Iau ( ri )
=
Iau (0)
[ (r ) r
]
2
f rp d rp
' 1 , (1-194)
where
r r 2 i r r
( ) ( )
f rp = A rp exp
zg
rp ri .

(1-195)
Hence, the peak value of an unaberrated image occurs at the Gaussian image point
regardless of the apodization of the system (Theorem 17c).
1.9.2 Approximate Expressions for Strehl Ratio

For a given amplitude distribution across the wavefront at the exit pupil of an
imaging system, we now consider the effect of an aberration on the image formed by it
and derive the following theorem.
Theorem 18. For small aberrations, the Strehl ratio of an image is determined by
the variance of the phase aberration across the exit pupil.
The ratio of the central irradiance at a distance zi with aberration and that at the
Gaussian image point without aberration is given by
1.9 Strehl Ratio 57
Iaa (0; zi )
2
zg
= S , (1-196)
(
Iau 0; zg ) zi
where
r r r
A (rp ) exp [ i2 (rp )] d rp
2
S =
[ (r ) r (1-197)
]
2
A rp d rp
is the Strehl ratio of the image. The right-hand sides of Eqs. (193a) and (1-197) are
identical. Hence,
S ' 1 . (1-198)
This Strehl ratio gives a measure of the irradiance reduction due to aberration in the
system and or due to defocus. It can be written in an abbreviated form
2
S = exp (i 2) , (1-199)
where the angular brackets L indicate a spatial average over the amplitude-weighted
pupil, e.g.,
r r r
2 =
A ( rp ) 2 ( rp ) d rp
[ A (rrp ) d rrp ]
2 . (1-200)
r
Since 2 is independent of rp , Eq. (1-199) can be written
[ )]
2
S = exp i ( 2
2
= cos (2
2 ) + sin (2
2 )
2 2
3 cos (2
2 )
2 (1-201)
,
equality holding when 2 is zero across the pupil, in which case S = 1. For small
aberrations, expanding the cosine function in a power series and retaining the first two
terms, we obtain the Maréchal result generalized for an apodized pupil
(1
? 2)
2
S >~ 2
2 , (1-202)
where
? 22 = (2
2 )2 (1-203)
is the variance of the phase aberration across the amplitude-weighted pupil.

58 IMAGE FORMATION
For small values of ? 2 , three approximate expressions have been used in the
literature:
~ (1
? 22 2)
2
S1 , (1-204)
S2 ~ 1
? 22 , (1-205)
and
S3 ~ exp (
? 22 ) . (1-206)
The first is the Maréchal formula, 19 the second is the commonly used expression obtained
when the term in ? 2 4
in the first is neglected, 20 and the third is an empirical expression
giving a better fit to the actual numerical results for various aberrations21 as we shall see
in Sections 2.2.4 and 3.2.4. The simplest expression to use is, of course, S2 , according to
which ? 22 gives the drop in the Strehl ratio. We note that the Strehl ratio for a small
aberration does not depend on its type but only on its variance across the apodized pupil
(Theorem 18). For a high-quality imaging system, a typical value of the Strehl ratio
desired is 0.8, corresponding to a wave aberration standard deviation of ? w = 14 .
1.9.3 Determination of Strehl Ratio

The Strehl ratio of an optical imaging system can be determined in a number of
different but equivalent ways. First, given its pupil function, its Strehl ratio can be
calculated by using Eq. (1-197). The amplitude distribution across its pupil may be
determined by measuring the irradiance distribution and taking its square root. The
aberration function may be determined by using some interferometric method, e.g., a
Twyman-Green interferometer discussed in Section 3.6.2 of Part I.
Second, the Strehl ratio may be determined from the PSF of the system. We calculate
its unaberrated PSF from its pupil shape and amplitude distribution, and normalize it so
that its central value is unity. Next, we integrate it in its plane to give the total power.
Finally, we measure the aberrated PSF and normalize it so that its integral has the same
value as for the aberration-free case. This step insures that the total power in the aberrated
image is the same as in the aberration-free image. The central value of the aberrated PSF
normalized in this manner gives the Strehl ratio.
Third, the Strehl ratio may be determined from the OTF of the system. Since its PSF
and OTF form a Fourier transform pair [see Eq. (1-77)], its Strehl ratio may be written
S = PSFa (0) PSFu (0)

r r r r
= Re 0 a ( v ) d v 0 u (v ) d v , (1-207)
where the subscripts a and u refer to an aberrated and an unaberrated system,

respectively. As explained following Eq. (1-77), only the real part of the aberrated OTF
contributes to the integral; since the PSF at any point is a real quantity, the integral of the
imaginary part of the OTF must be zero. Thus, the Strehl ratio may be calculated by
1.10 Hopkins Ratio 59
integrating the real part of the measured aberrated OTF over all spatial frequencies and
dividing it by a similar integral of the calculated unaberrated OTF. In any of these three
approaches for determining the Strehl ratio, one must take into account the effect of the
inverse-square law, i.e., the effect of zi being different from zg , unless they are
practically equal to each other (see Section 2.5).
1.10 HOPKINS RATIO

An aberrated system with a Strehl ratio of 0.8 forms the image of an object with a
quality that is only slightly inferior to the corresponding quality for an aberration-free
system, regardless of the spatial frequencies (or the size of the detail) of interest in the
object. However, systems having much larger aberrations form good-quality images of
objects in which the size of the detail is much coarser than the limiting resolution of the
system. Accordingly, we now consider aberration tolerances based on a certain amount of
reduction in the MTF of the system corresponding to a certain spatial frequency. In
particular, we derive a theorem that gives the ratio of the aberrated and unaberrated MTFs
in terms of the variance of the difference of aberration functions for two pupils separated
by a parameter related to the spatial frequency.
Theorem 19. (a) For small values of the phase aberration difference function, the
r
Hopkins ratio for a certain spatial frequency vi is determined by the variance of the
difference function across the overlap area of two pupils displaced with respect to each
r
other by Rvi . (b) The mean value of the phase aberration difference function represents
the corresponding PTF.
r
Following Hopkins,22 we define a modulation ratio H ( vi ) as the ratio of the MTFs
r r r
0 ( vi ) and 0 u ( vi ) of a system at a spatial frequency vi with and without aberration, i.e.,
r r r
H ( vi ) = 0 ( vi ) 0 u ( vi ) . (1-208)
r
For obvious reasons, we call H ( vi ) the Hopkins modulation (or contrast) ratio. From Eq.
(1-107), this ratio is less than one. From Eq.(1-73c) it can be written
r r r
r A(r ) A(r
p p ) {[ (r )

Rvi exp i 2 rp
2 rp
R vi (r r
)] } d rrp
H ( vi ) = r r r r . (1-209)
( ) (
A rp A rp
Rvi d rp )
In the rotated ( p, q ) coordinate system shown in Figure 1-4, Eq. (1-209) may be written
r r
r
H ( vi ) =
I ( p, q; vi ) exp [iQ ( p, q; vi )] dp dq
r , (1-210)
I ( p, q; vi ) dp dq
r
where the mutual irradiance function I ( p, q; vi ) and the phase aberration difference
r
function Q( p, q; vi ) are given by Eqs. (1-128). As in Eq. (1-125), a function in the ( p, q )
coordinate system is obtained from the corresponding function in the x p , y p coordinate ( )
system by replacing x p with p cos 9
q sin 9 and y p with p sin 9 + q cos 9 . As discussed
( )
in Section 3.5 of Part I, the aberration function 2 x p , y p for a rotationally symmetric
60 IMAGE FORMATION
imaging system depends on x p and y p through x 2p + y 2p and x p , where

r
x p = rp cos , y p = rp sin , and is the angle the vector rp makes with the x p axis
(which lies in the tangential plane containing the optical axis and the object point).
( )
Hence, 2( p, q ) is obtained from 2 x p , y p by replacing x 2p + y 2p by p 2 + q 2 and x p
by p cos 9
q sin 9 .

r
H ( vi ) = {
exp i [Q
Q ]} , (1-212)
where the angular brackets indicate an average across the overlap region of the two
apodized pupils. For example,
r r
Q =
I ( p, q; vi ) Q(rp, q; vi ) dp dq . (1-213)
I ( p, q; vi ) dp dq
For small values of Q
Q , we may retain only the first three terms in the expansion of
the exponential in Eq. (1-212), one of which reduces to zero upon averaging. Thus, we
obtain the result
r ~ 1 2
H ( vi ) 1
?Q , (1-2150
2
where
2
? Q2 = Q 2
Q (1-215)
is the variance of the phase aberration difference function across the overlap region of the
r
[
two apodized pupils. Multiplying both sides of Eq. (1-208) by exp i 7( vi ) , we obtain]
[see Eq. (1-96)]
r r r r
[ ]
H ( vi ) exp i 7( vi ) = 0 ( vi ) 0 u ( vi )
{
= exp i [Q
Q ]} exp (i Q )
~ 1
? Q2 exp (i Q ) .
1
(1-216)
2
Comparing Eqs. (1-214) and (1-216), we find that

r ~
7 ( vi ) Q ; (1-217)
i.e., Q represents approximately the phase transfer function.
We noted in Section 1.9.2 that, for small aberrations, the Strehl ratio of an aberrated
system depends on the variance of the aberration function across its pupil and not on the
type of the aberration. Similarly, we note from Eq. (1-214) that, for small values of the
phase aberration difference function, the Hopkins ratio for a certain spatial frequency
1.11 Line- and Edge-Spread Functions 61
depends on the variance of the difference function across the overlap region of two
displaced pupils (displacement depending on the spatial frequency) and not on the type of
the aberration (Theorem 19a). Similarly, Eq. (2-217) shows that the mean value of the
phase aberration difference function represents the PTF (Theorem 19b).
As in the case of Strehl ratio, a better approximation to Hopkins ratio is obtained by

using the exponential relation 23
r ~ Ê 1 ˆ
H (v i ) exp Á - s Q2 ˜ . (1-218)
Ë 2 ¯
1.11 LINE- AND EDGE-SPREAD FUNCTIONS

In this section, we discuss the images of line and edge objects and show how they
can be obtained from the image of a point object. We also relate them to the OTF of a
system.
1.11.1 Line-Spread Function

We derive the following theorems.
Theorem 20. (a) The aberration-free line-spread function is symmetric about its
center, and the aberrations reduce its central value. (b) The line-spread function is equal
to the derivative of the edge-spread function.
Consider an imaging system with a point-spread function PSF( xi , yi ) , representing

the irradiance distribution of the image of a point object located at (0, 0) in the object
plane for unity total power. For a point object located at (0, yo ) , its Gaussian image is
( )
located at 0, yg , where yg = M yo for an image magnification of M. The corresponding
PSF is given by PSF( xi , yi < yg ) . The image of an infinitely long incoherent line object
lying along the yo axis may be obtained by considering it as a collection of an infinite
number of points and linearly superimposing their PSFs. Thus, the image of the line
object, called the line-spread function (LSF), may be written
'
LSF( xi ) = 0 PSF( xi , yi < yg )dyg
<'
'
= 0 PSF( xi , yi )dyi . (1-219a)
<'
For a line object along the yo axis, the LSF depends only on xi , i.e., its variation with xi
is the same regardless of the value of yi , as expected for an isoplanatic line object. Since
r
( ) ( )
its Gaussian image Ig rg = b x g , its diffraction image can also be obtained by applying
Theorem 5 as the convolution of its Gaussian image and the PSF, i.e.,
' '
<' <'
( )
LSF( xi ) = 0 0 b x g PSF( xi < x g , yi < yg ) dx g dyg . (1-219b)
It should be evident that the LSF can be obtained from the PSF by scanning it with a long
and narrow slit. If the slit is parallel to the yi axis, the LSF thus obtained is LSF( xi ) .
62 IMAGE FORMATION
The LSF of a system can be written in terms of its pupil function by substituting for
its PSF from Eq. (1-57) into Eq. (1-219) . Thus
' GG 2 /i
LSF( xi ) =
1
(
2 2 0 dyi P x p , y p exp ³ <
Pex h R < ' H H h R
) (
xi x p + yi y p µ d x p dy p

)
GG 2/i
HH
(
× P * x vp , y vp exp ³
h R
) (
xi x vp + yi y vp µ d x vp dy vp

)
GG £ 2 /i ¥ GG £ 2 /i ¥
HH
( )
= P x p , y p exp ² <
¤ h R
xi x p ´ d x p dy p P * x vp , y vp exp ²
¦ HH ¤ h R
( ¦
)
xi x vp ´ d x vp dy vp
'
G 2 /i
×
1
Pex h2 R 2
H
exp ³
h R
yi y vp < y p µ dyi

( )
<'
2
G £ 2 /i ¥
=
1
Pex h R H
(
0 dy p P x p , y p exp ² <
¤ h R
)
xi x p ´ d x p
¦
, (1-220)
where in the last step we have used the fact that the integral over yi is equal to
( )
h Rb y vp < y p and thereby carried out the integration over y vp . Hence, the LSF
normalized by its aberration-free value at the origin, is given by
2
G £ 2 /i ¥
H
(
0 dy p P x p , y p exp ² <)
¤ hR
xi x p ´ d x p
¦
LSF( xi ) = . (1-221a)
[0 ( ) ]
2
0 dy p P x p , yp d x p
The aberration-free LSF is symmetric about xi = 0 . The central value of the LSF
normalized by its aberration-free value is called the Struve ratio. It is given by
( )
2
0 dy p 0 P x p , y p d x p
LSF(0) = (1-221b)
0 dy p [ 0 P( x p , y p ) d x p ]
2
)1 , (1-221c)
where in the last step we have used Hölder’s inequality (1-104) with f x p = P x p , ( ) ( )
( )
g x p = 1, and n = 2 . Thus, aberrations reduce the central value of the LSF, as in the case
of the PSF.
If the line object lies along a yov axis (or parallel to it), making an angle ` with the
yo axis, as illustrated in Figure 1-9, then the LSF( xiv) is obtained by writing the pupil
( )
function in the coordinate system x vp , y vp by replacing x p by x vp cos ` < y vp sin ` and y p
by x vp sin ` + y vp cos ` .
yo
yov
xov
xo
(a) Object plane

yp
ypv
xpv
xp
(b) Pupil plane

yi
yiv
xiv
xi
(c) Image plane
Figure 1-9. Coordinate system in various planes for determining the LSF. (a) Object
plane. (b) Pupil plane. (c) Image plane. The line object lies along the yov axis or
parallel to it, making an angle ` with the yo axis.
64 IMAGE FORMATION
1.11.2 Edge-Spread Function

The image of a line object located at a distance x o parallel to the yo axis is given by
( )
LSF xi < x g , where x g = M x o for an image magnification of M. The image of an edge
object passing through the origin and lying parallel to the y axis in the left half of the
space can be obtained by considering the object, which is the right half of the space, as a
collection of line objects and linearly superimposing their LSFs. Thus, the image of an
edge object, called the edge-spread function (ESF), can be written
'
(
ESF( xi ) = 0 LSF xi < x g dx g
0
)
xi
= 0 LSF( x ) dx , (1-222a)
<'
xi '
= 0 dx 0 PSF ( x , y ) dy , (1-222b)
<' <'
where in the last step we have substituted for the LSF from Eq. (1-219a). It is evident
from Eq. (1-222a) that the LSF is the derivative of the ESF, i.e.,
d
LSF( xi ) = ESF( xi ) . (1-223)
dxi
Moreover, it is evident from Eq. (1-222b) that ESF(') = 1, as may be seen by the use of
Eqs. (1-61) and (1-63). It may also be seen by noting that the integral for xi = ' is equal
to o(0, 0) , which, in turn, is equal to unity.
Introducing the step or the edge function representing the edge object,
H ( x ) = 1 for x * 0
(1-224)
= 0 for x < 0 ,
we may write Eqs. (1-222) in the form

'
ESF( xi ) = 0 LSF( x ) H ( xi < x ) dx (1-225a)
<'
'
= 0 PSF( x, y) H ( xi < x ) dx dy . (1-225b)
<'
Thus, ESF( xi ) is equal to a 1-D convolution of the edge object H ( xi ) with LSF( xi ) , or a
2-D convolution with PSF( xi , yi ) , as expected for an isoplanatic extended object.
In practice, a PSF can be measured by scanning it with a photodetector with a

pinhole mounted on the detector. However, even with a bright point source, the signal on
the detector may be too small, especially at points outside the central bright spot of the
PSF. The PSF may also be measured using a two-dimensional array of small
photodetectors. But, again, precise measurements of the PSF outside the bright spot may
be difficult owing to their low values. The LSF can be obtained by scanning a
corresponding PSF with a long and narrow slit. The difficulty of a low signal can be
overcome by using a line object and measuring the LSF by scanning its image with a long
and narrow slit. The slit is parallel to the line object and its scan direction is normal to it.
By removing one side of the slit used as a line source, the ESF can be measured in a
similar manner.
1.11.3 LSF and ESF in Terms of OTF
Theorem 21. (a) The LSF and the OTF form a 1-D Fourier pair. (b) Similarly, the
ESF and the OTF divided by a spatial frequency are related to each other by a 1-D
Fourier transf orm.
Substituting for PSF( x, y) in terms of its Fourier transform, the OTF o (j, d) [see
Eq. (1-76)], namely,
PSF ( x, y) = 00 o (j, d) exp [< 2 /i (j x + dy)] dj d d , (1-226)
where, for simplicity, we have dropped the subscript on the image coordinates ( xi , yi ) ,
Eq. (1-219) may be written
'
[ ]
LSF ( x ) = 0 dy 00 o (j, d) exp < 2 / i (j x + d y) dj d d
<'
= 00 o (j, d) exp ( < 2 / i j x ) dj d d 0<'' exp ( < 2 / i dy) dy
= 00 o (j, d) exp ( < 2 / i j x ) b( d) dj d d
= 0 o (j, 0) exp ( < 2 / i j x ) dj . (1-227a)
Thus, LSF( x ) , representing the image of a line object lying along the yo axis, is the 1-D
inverse Fourier transform of o (j, 0) (Theorem 21a). Since LSF( xi ) is a real function,
only the real part of the integrand contributes to the integral; the imaginary part yields
zero upon integration. Hence, Eq. (1-227a) may be written
LSF( x ) = 0 Re o(j, 0) cos(2 /j x )d j < 0 Im o(j, 0) sin(2 /j x )d j . (1-227b)
Similarly, by considering a line object along the x axis, we can show that its image
LSF( y) is the inverse Fourier transform of o (0, d) , i.e.,
LSF( y) = 0 o (0, d) exp ( < 2 / i d y) dd . (1-228 )
The LSF for a line object making an angle ` with the yo axis is given by
66 IMAGE FORMATION
LSF( x v) = 0 o (j v, 0) exp ( < 2 / i j v x v) d j v , (1-229)
where x v and j v are parallel to each other and make an angle ` with the x axis in the
image plane, and o (jv, 0) is simply the OTF along the direction j v . It should be clear
from Eqs. (1-227) through (1-229) that the LSF represents the variation of irradiance in
the image of a line object along a direction that is normal to the direction of the object.
Now we consider an edge object that is parallel to the yo axis. To relate the ESF to
the OTF, we Fourier transform both sides of Eqs. (1-225). Since the integrals on their
right-hand sides represent convolutions, their Fourier transforms are equal to the product
of the Fourier transforms of the convolving functions. From Eq (1-226), the Fourier
transform of LSF( x ) is given by
0 LSF( x ) exp (2 /i j x )d x = o(j, 0) . (1-230)
The Fourier transform of H ( x ) is given by24
0 H ( x ) exp (2 /i j x )d x = b(j) 2 < (1 2 /ij) . (1-231)
Hence, Fourier transformation of Eq. (1-225a) yields
0 ESF( x ) exp (2 /i j x )d x = b(j) 2 < o(j, 0) 2 /i j , (1-232)
since o(j, 0)b(j) = o(0, 0) = 1 . Fourier transformation of Eq. (1-225b) also yields Eq. (1-
232). The left-hand side of Eq. ( 1-232) represents the spatial-frequency spectrum of the
image of an edge object lying parallel to the yo axis, and its right-hand side represents the
product of the spatial-frequency spectrum of the edge object (or its Gaussian image) and
the system OTF for d = 0 , as expected for an isoplanatic object according to Eq. (1-70).
Inverse Fourier transforming Eq. (1-232), we obtain the ESF in terms of the OTF:
1 1 G o( j , 0 )
ESF( x ) = < exp ( < 2 /i j x ) d j . (1-233)
2 2 /i
H j
For an incoherent object, ESF( x ) represents the irradiance distribution of the image of an
edge object. It is, therefore, a real function. Hence, only the imaginary part of the
integrand in Eq. (1-233) contributes to the integral; the real part yields zero upon
integration. Accordingly, Eq. (1-233) may be written
1 G sin(2 /j x ) G cos(2 /j x )
ESF( x ) = + Re o(j, 0) d j < Im o(j, 0) dj . (1-234)
2 H 2 /j H 2 /j
Equations (1-232) and (1-233) show that the ESF( x ) and o(j, 0) j are related to each
other by a 1-D Fourier transform (Theorem 21b). Letting x = 0 in Eq. (1-234), we obtain
1 G Im o(j, 0)
ESF(0) = < dj , (1-235)
2 H 2 /j
1.12 Shift-Invariant Imaging of a Coherent Object 67
i.e., the value of the ESF at the origin is determined solely by the imaginary part of the
OTF o(j, 0) . If the OTF is real, e.g., for a symmetric aberration, then ESF(0) = 1 2
independent of the aberration. If the edge object makes an angle ` with the yo axis, then
the ESF is given by an equation similar to Eq. (1-233), except that x and j axes are
replaced by x v and j v axes as in Eq. (1-229).
Equations (1-227) and (1-233) for the LSF and ESF of a system in terms of its OTF
are also applicable to coherent imaging, provided the OTF is replaced by the
corresponding coherent transfer function (CTF) discussed in Section 1.12.2. This is done
in Section 2.15 for systems with circular pupils.
1.12 SHIFT-INVARIANT IMAGING OF A COHERENT OBJECT

So far we have considered imaging of incoherent objects. The irradiance distribution
of the image was obtained by adding the irradiance distributions of the images of the
object elements. For a coherently illuminated object, the complex amplitudes from its
elements bear a time-independent relationship among them (as opposed to randomly
varying relationships in the case of an incoherent object). Hence, the complex amplitude
distribution of its image is obtained by superimposing the complex amplitude
distributions of its image elements. We show that the coherent point-spread function is
proportional to a Fourier transform of the relative pupil function and, therefore, the
coherent transfer function is simply a scaled pupil function. Accordingly, the incoherent
point-spread function is proportional to the modulus square of the coherent spread
function, and the incoherent transfer function is equal to the autocorrelation of the
coherent transfer function. We now establish the following theorems.
Theorem 22. The coherent point-spread function of an imaging system is

proportional to the inverse Fourier transform of its pupil function. Accordingly, the
coherent spread function.
Theorem 23. (a) The complex amplitude image of an isoplanatic coherent object is
equal to the convolution of its Gaussian amplitude image and its coherent spread
function. (b) Accordingly, the spatial frequency spectrum of the diffraction amplitude
image is given by the product of the spectrum of the Gaussian amplitude image and the
coherent transfer function.
Theorem 24. The coherent transfer function of an imaging system is proportional to

its pupil function. Accordingly, its incoherent transfer function is equal to the
1.12.1 Coherent Spread Function

(r r )
If we let G rp ; ro be the relative pupil function of an imaging system representing
the complex amplitude in the plane of its exit pupil due to a unit amplitude of an object
r
element centered at ro per unit area of the object element at a distance zo from the
68 IMAGE FORMATION
entrance pupil, then the complex amplitude distribution in an observation plane at a

distance zi from the plane of the exit pupil is given by [see Eq. (1-49)]
r r r r
Ui ( ri ; ro ; zi ) = 0 6Ui ( ri ; ro ; zi )
i £ r2 ¥ G r r £ r2 ¥
= < exp ³ik ² zi < zg + i ´ µ d ro Uo ( ro ) exp ² < ik o ´
h zi ³ ¤ 2 zi ¦ µ H ¤ 2 zo ¦
G r r 2 /i r £ r z r ¥ r

(
× G rp ; ro ; zi exp ³<
³ h zi
) ¤ zg ¦ µ
u
rp ² ri < i rg ´ µ d rp , (1-236)
H
r r r r
where Uo ( ro ) is the amplitude at the object point ro , rg = M ro is the position vector of
the Gaussian image point, zg is the distance of the Gaussian image plane, and
r r r r ik £ 1 1¥
(
G rp ; ro ; zi ) ( )
= G rp ; ro exp ³ ² < ´ rp2 µ
³ 2 ¤ zi zg ¦ µ
(1-237)
is the defocused pupil function.

r
For a point object located at rj with an amplitude U j , i.e., for
r
Uo ( ro ) = U j b ro < rj (r r
) , (1-238)
Eq. (1-236) reduces to
(r r
Ui ri ; rj ; zi ) = U j CSF ri ; rj ; zi(r r
) , (1-239)
where
r r ¨« 1 £ r 2 rj ¥ ¬«
2
(
CSF ri ; rj ; zi ) = <
i
h zi
( )
exp ©ik ³ zi < zg + ² i + ´ µ
«ª ³ 2 ¤ zi zo ¦ µ «
®
G r r 2 /i r £ r z r¥ r

(
× G rp ; rj ; zi exp ³<
³ h zi
) ¤ zg
u
rp ² ri < i M rj ´ µ d rp
¦ µ
(1-240)
H
is the coherent (point-) spread function of the system.
Now, we make certain approximations so that the amplitude distribution in an

observation plane is equal to the convolution of the corresponding distribution of the
Gaussian image and the CSF. As in the case of incoherent imaging, we assume that the
object is small enough that the relative pupil function does not vary significantly with the
location of an object element on it; i.e., we assume that the system is isoplanatic for the
r r r
( )
small object under consideration. Thus, we replace G rp ; ro ; zi by G rp ; zi for a small ( )
isoplanatic object. In practice, the Fresnel number of imaging systems is quite large,
resulting in a very small longitudinal defocus. Hence, we may replace zi by zg , except in
1.12 Shift-Invariant Imaging of a Coherent Object 69
determining the defocus aberration, which varies as rp2 . For systems with small fields of
view, we may replace zg by R, the radius of curvature of the Gaussian reference sphere
with respect to which the aberration is defined.
For coherent imaging, we must consider another approximation, namely, that the
r
variation of the quadratic phase factor depending on ro is negligible.25 We will see that
the magnitude of the CSF is significant only in a small region (on the order of the Airy
disc discussed in Section 2.2) surrounding the Gaussian image point. Therefore, at a
r
particular point of observation ( ri ; zi ) , the contributions of the quadratic phase factor will
r r r
come from a small region of the object. Hence, we may replace ro = rg M by ri M in
the object-dependent quadratic phase factor. Any phase factors that do not depend on the
location of the object point can be dropped if it is the irradiance that is measured in the
observation plane.
In view of the above approximations, Eq. (1-240) reduces to
(r r
CSF ri ; rj ; zi ) = <
i G

hR H
r
( 2 /i r r
G rp ; zi exp ³<)
hR
r r
u(
rp ri < M rj µ d rp

) . (1-241)
r r
We note that the integral on the right-hand side depends on the position vectors ri and rg
of the observation and Gaussian image points, respectively, through their difference
r r r r r r
( ) (
ri < rg only. Hence, we may replace CSF ri ; rj ; zi by CSF ri < M rj ; zi . Thus, the CSF )
is shift invariant in that its form does not depend on the location of the object point; the
r r
whole distribution shifts by virtue of being centered at rg = M rj . Accordingly, for an
isoplanatic coherent object, Eq. (1-236) for the amplitude distribution of its image may be
written
r
(r ) (r
Ui ( ri ; zi ) = M <1 0 Uo rg M CSF ri < rg ; zi d rg
r
) r
r r r r
= 0 Ug ( rg ) CSF( ri < rg ; zi ) d rg , (1-242)
where
r i G r £ 2 /i r r ¥ r
CSF( ri ; zi ) = <
hR
H
( )
G rp ; zi exp ² <
¤ hR
rp ri ´ d rp
¦
u (1-243)
and
(r )
U g rg = M <1 Uo rg M (r ) (1-244)
is the amplitude distribution of the Gaussian image. Thus, the coherent spread function is
proportional to a Fourier transform of the pupil function (Theorem 22), and the image
amplitude distribution for an isoplanatic coherent object is equal to the convolution of the
Gaussian amplitude image (which is an exact replica of the object amplitude except for its
magnification M) and the coherent spread function (Theorem 23a). Comparing Eq. (1-55)
70 IMAGE FORMATION
with Eq. (1-242), we find that the incoherent and coherent point-spread functions are
related to each other according to
r
CSF( ri )
2
r
PSF( ri ) = r r . (1-245)
0 ( )
2
G rp ; zi d rp
1.12.2 Coherent Transfer Function

Fourier transforming both sides of Eq. (1-242), we obtain
r r r r r r r r
u ( )
U˜ i ( vi ) = 0 d ri exp (2 / i vi ri ) 0 U g rg CSF ri < rg d rg ( )
r r r r r r r r r r
= 0 d rgUg (rg ) exp (2/i vi u rg ) 0 CSF (ri < rg ) exp [2/ vi u (ri < rg )] d ri
r r
= U˜ g ( vi ) CTF( vi ) , (1-246)
where
r r r r r
U˜ g ( vi ) = 0 Ug ( rg ) exp (2/i vi u rg ) d rg (1-247)
is the spectrum of the Gaussian amplitude image and

r r r r r
CTF( vi ) = 0 CSF ( ri ) exp (2 / vi ri ) d ri u (1-248)
r
is the coherent transfer function of the system corresponding to a spatial frequency vi in
the image plane. Substituting Eq. (1-243) into Eq. (1-248) and dropping the constant
phase factor < i, we obtain
r 1 G r r rG r £ 2 /i r r ¥ r
CTF( vi ) =
hR H H
u
d ri exp (2 / vi ri ) G rp exp ² <
¤ hR
( )
rp ri ´ d rp
¦
u
r
r r G r r £ r rp ¥
=
1
hR 0
d rp G rp ( )
³ ¤
u
d ri exp ³2 /i ri ² vi < µ
h R ´¦ µ
H
r
r r £ r rp ¥
=
1
hR 0 ( )
d rp G rp b ² vi <
¤ h R ´¦
r
= G(h Rvi ) . (1-249)
Thus, the coherent transfer function is simply a scaled version of the relative pupil
function (Theorem 24). If the system is diffraction limited, i.e., if it is aberration free, and
r
( )
its transmission across the pupil is uniform so that G rp is equal to a constant across the
pupil (whose value varies inversely with the distance zo of the entrance pupil from the
r
object), then the normalized CTF is unity for those frequencies for which h Rvi lies
inside the pupil.
1.13 Summary of Theorems 71
r
Fourier transforming both sides of Eq. (1-245), substituting for CSF( ri ) from Eq. (1-
243), and proceeding as we did in obtaining Eq. (1-73b), we obtain
r r r r r r r
0 G (r ) G * (r ) 0 G (r )
2
o ( vi ) = p p < h R vi d rp p d rp . (1-250)
Thus, the incoherent transfer function is the (normalized) autocorrelation of the coherent
transfer function. Since the pupil functions P and G are proportional to each other, the
OTF given by Eq. (1-250) is the same as that given by Eq. (1-73b).
Fourier transforming both sides of Eq. (1-242), we obtain

r r r
U˜ i ( vi ) = U˜ g ( vi ) CTF( vi ) , (1-251)
where the spectra of the amplitude diffraction and Gaussian images are given by
r r r r r
u
U˜ i ( vi ) = 0 Ui ( ri ) exp (2 / vi ri ) d r (1-252)
and
r r r r r
U˜ g ( vi ) = 0 Ug ( ri ) exp (2 / vi u ri ) d r , (1-253)
respectively. Thus, the spectrum of the diffraction image is given by the product of the
spectrum of the Gaussian image and the CTF (Theorem 23b).
1.13 SUMMARY OF THEOREMS

Theorem 1. As a wave propagates, its disturbance according to the Huygens-
Fresnel principle is given by the superposition of secondary spherical wavelets weighted
by the amplitudes at the points where they originate on the wave.
Theorem 2. Under certain approximations, the propagation of a wave is described
by a Fourier transform of its complex amplitude modified by a quadratic phase factor in
the Fresnel approximation, or without modification in the Fraunhofer approximation.
Theorem 3. The Gaussian image of a small object lying at a large distance from an
imaging system is an exact replica of the object, except for its magnification.
Theorem 4. The diffraction PSF of an imaging system is proportional to the modulus

square of the inverse Fourier transform of its pupil function.
Theorem 5. The diffraction image of an isoplanatic incoherent object is equal to the

convolution of its Gaussian image and the PSF of the imaging system.
Theorem 6. The spatial-frequency spectrum of the diffraction image of an

isoplanatic incoherent object is equal to the product of the spectrum of its Gaussian
image and the OTF of the system, which by definition is equal to a Fourier transform of
its PSF.
72 IMAGE FORMATION
Theorem 7. The OTF of an imaging system is also equal to the autocorrelation of its
pupil function.
Theorem 8. The PSF and the OTF of a system with a radially symmetric pupil
function form a zero-order Hankel transform pair.
Theorem 9. Whereas the Gaussian image of a sinusoidal object is sinusoidal with

the same modulation and phase as the object, the diffraction image is also sinusoidal but
with a reduced contrast and changed phase depending on the spatial frequency. The
modulus of the OTF represents the factor by which the contrast changes and its phase
gives the change in the phase.
Theorem 10. (a) The OTF is a complex symmetric function with a value of unity at
the origin. (b) Its magnitude, the MTF, is unity or less at any other spatial frequency. (c)
The aberrated MTF is less than or equal to the corresponding aberration-free value.
Theorem 11. The slope of the real part of the OTF at the origin is independent of the
aberration, and it is equal to the corresponding slope of its MTF.
Theorem 12. The diffraction OTF approaches the geometrical OTF as the
wavelength approaches zero.
Theorem 13. The PSF of a system with a radially symmetric pupil function behaves
asymptotically as the inverse cube of the distance from its center, independent of the
aberration.
Theorem 14. The encircled power for large circles is independent of the aberration.
Theorem 15. The asymptotic behavior of encircled power of an unapodized system is

determined by the ratio of the perimeter length and the transmitting area of its exit pupil.
Theorem 16. (a) The centroid of the diffraction PSF is given by the slope of the
imaginary part of the its diffraction OTF at the origin. (b) It is the same as the centroid of
the ray geometrical PSF. (c) The centroid of an aberration-free system lies at the
Gaussian image point regardless of the amplitude variations across its exit pupil. (d) For
a system with an aberrated but unapodized exit pupil, the centroid can be obtained from
its aberration only along its perimeter.
Theorem 17. (a) The central irradiance for an apodized-aberrated system is less
than or equal to the corresponding value for an unapodized-unaberrated system. (b) For
a given total power, any amplitude variations reduce the central irradiance and any
phase variations further reduce it. (c) The peak value of an unaberrated PSF lies at the
center, regardless of the amplitude variations across the pupil.
Theorem 18. For small aberrations, the Strehl ratio of an image is determined by
the variance of the phase aberration across the exit pupil.
1.13 Summary of Theorems 73
Theorem 19. (a) For small values of the phase aberration difference function, the
r
Hopkins ratio for a certain spatial frequency vi is determined by the variance of the
difference function across the overlap area of two pupils displaced with respect to each
r
other by h Rvi . (b) The mean value of the phase aberration difference function represents
the corresponding PTF.
Theorem 20. (a) The line-spread function is symmetric about its center and the
aberrations reduce its central value. (b) The line-spread function is equal to the
derivative of the edge-spread function.
Theorem 21. (a) The LSF and the OTF form a 1-D Fourier pair. (b) Similarly, the
ESF and the OTF divided by a spatial frequency are related to each other by a 1-D
Fourier transform.
Theorem 22. The coherent point-spread function of an imaging system is

proportional to the inverse Fourier transform of its pupil function. Accordingly, the
coherent spread function.
Theorem 23. (a) The complex amplitude image of an isoplanatic coherent object is
equal to the convolution of its Gaussian amplitude image and its coherent spread
function. (b) Accordingly, the spatial frequency spectrum of the diffraction amplitude
image is given by the product of the spectrum of the Gaussian amplitude image and the
coherent transfer function.
Theorem 24. The coherent transfer function of an imaging system is proportional to

its pupil function. Accordingly, its incoherent transfer function is equal to the
74 IMAGE FORMATION
APPENDIX A: FOURIER TRANSFORM DEFINITIONS

An examination of the literature shows that there is no universal definition of the
r
Fourier transform. We have used the following definition: The Fourier transform F(@) in
r
the spatial frequency domain of a function f (r ) in the space domain, is given by
r r r r r

F(@) = f (r ) exp(2 i @ r ) dr . (A1)
r r
The inverse Fourier transform f (r ) of F(@) is accordingly given by
r r r r r

f (r ) = F(@) exp(
2 i @ r ) d @ . (A2)
Thus, we use a + i in the Fourier transform integral and a

i in the inverse Fourier
transform integral. Accordingly, since there is a
i in the exponent in Eq. (1-51), the
incoherent PSF is proportional to the modulus square of the inverse Fourier transform of
the pupil function. Similarly, the OTF, which by definition is the Fourier transform of the
PSF, has a + i in the exponent in Eq. (1-71). It is also equal to the autocorrelation of the
pupil function, as in Eq. (1-73b).
Our definition of a Fourier transform is similar to that by Born and Wolf2 , but is
different from that by the others. For example, Hopkins 22 uses a + i in the PSF integral,
and a
i in the OTF definition which he calls an inverse Fourier transform. Welford26
uses a
i in the PSF and a + i in the OTF, just as we have in Eqs. (1-51) and (1-71).
However, he refers to OTF as the inverse Fourier transform of PSF. Goodman2 and
Gaskill27 both use
i in both the PSF and the OTF. According to them, PSF is related to
the Fourier transform of the pupil function and OTF is the Fourier transform of PSF. The
OTF according to them, as the autocorrelation of the pupil function, is the complex
conjugate of the corresponding integral given in Eq. (1-73b).
r r
This ambiguity does not arise if f (r ) is radially symmetric, i.e., if f (r ) = f (r ) ,
r
where r = r , since in that case its Fourier transform is real and reduces to a zero-order
Hankel transform:
; 2
F(@) = f (r ) exp(2 ir@ cos ) rdrd
0 0
;
= 2 f (r ) J 0 (2 r@) rdr , (A3)
0
where J 0 is the zero-order Bessel function of the first kind. Similarly,
;
f (r ) = 2 F(@) J 0 (2 r@) @ d@ . (A4)
0
According to our sign convention, the wave aberration associated with an object ray
is numerically positive if it travels an extra optical path length to reach the reference
sphere compared to the corresponding chief ray. Born and Wolf2 and Goodman2 have the
Appendix B: Some Frequently Used Integrals 75
same sign convention, but Welford’s26 is opposite to ours. A change in the sign of the
wave aberration changes the sign of the imaginary part of the pupil function, and the
coherent as well as the incoherent transfer functions. It also rotates the incoherent PSF by
p as, for example, in the case of coma.
It is quite common in the optics literature to consider a point object lying along the y
axis. When using polar coordinates of a point in the plane of the exit pupil, the polar
angle q in that case is defined as the angle made by the position vector of the point with
the y axis, contrary to the standard convention as the angle with the x axis. We choose a
point object along the x axis so that, for example, coma aberration is expressed as
( ) ( )
x x 2 + y 2 and not as y x 2 + y 2 . A positive value of our coma aberration yields a PSF
that is symmetric about the x axis (or symmetric in y) with its peak and centroid shifted to
a positive value of x with respect to the Gaussian image point.
APPENDIX B: SOME FREQUENTLY USED INTEGRALS

Dirac delta function:
;
( )
exp ± 2i8 x d8 = ( x ) , (B1)

;
;
f ( x ) ( x ) dx = f (0) , (B2)

;
1
( ax ) = ( x ) . (B3)
a
Convolution of two functions f ( x ) and g( x ) :

;
f ( x )A g( x ) = f ( x ) g( x
x ) dx . (B4)

;
Crosscorrelation of two functions f ( x ) and g( x ) :
;
f ( x ) A g * ( x ) = f ( x ) g * ( x'

x ) dx . (B5)

;
Autocorrelation of a function f ( x ) :
•
f ( x) ƒ f * ( - x) = Ú f ( x ¢ ) f * ( x ¢ - x ) dx ¢ , (B6)
-•
Zero-order Bessel function of the first kind and its radial integral:
2
exp(ix cos ) d = 2 J 0 ( x ) . (B7)
0
r0
r0
J 0 ( mx ) rdr = J1 ( mr0 ) . (B8)
0 m
76 IMAGE FORMATION
REFERENCES
1. V. N. Mahajan, Optical Imaging and Aberrations, Part I: Ray Geometrical Optics
(SPIE Press, Bellingham, WA, 1998).
2. For various theories of diffraction, the reader may refer to M. Born and E. Wolf,
Principles of Optics, 7th Ed. (Cambridge University Press, New York, 1999); J.
W. Goodman, Introduction to Fourier Optics, 2nd ed. (McGraw-Hill, New York,
1996).
3. A. Sommerfeld, Optics, Vol 4 (Academic, New York, 1972); also J. W.

Goodman, Introduction to Fourier Optics, 2nd ed. (McGraw-Hill, New York,
1996).
4. E. Lalor, “Conditions for the validity of angular spectrum of plane waves,” J. Opt.
Soc. Am. 58, 1235–1237 (1968).
5. J. E. Harvey and R. V. Shack, “Aberrations of diffracted wave fields,” Appl. Opt.

17, 3003–3009 (1978); also, J. E. Harvey, “Fourier treatment of near-field scalar
diffraction theory,” Am. J. Phys. 47, 974–980 (1979).
6. V. N. Mahajan, “Aberrations of diffracted wave fields I. Optical Imaging,” J. Opt.

Soc. Am. 17, 2216–2222 (2000); and “Aberrations of diffracted wave fields II.
Diffraction Gratings,” J. Opt. Soc. Am. 17, 2223-2228 (2000).
7. R. Bracewell, The Fourier Transform and Its Applications (McGraw-Hill, New

York, 1965).
8. For properties of Bessel functions, the reader may refer to G. N. Watson, Treatise
on the Theory of Bessel Functions (Cambridge University Press, New York,
1944).
9. G. A. Korn and T. M. Korn, Mathematical Handbook for Scientists and Engineers

(McGraw-Hill, New York, 1968), p. 118.
10. H. H. Hopkins, “Geometrical optical treatment of frequency response,” Proc.

Phys. Soc. (London) B70, 1162–1172 (1957). Also, K. Miyamoto, “On a
comparison between wave optics and geometrical optics by using Fourier
analysis. I. General Theory,” J. Opt. Soc. Am. 48, 57–63 (1958); “II. Astigmatism,
coma, spherical aberration” 48, 567–575 (1958); “III. Image evaluation by spot
diagram,” 49, 35–40 (1959); “Wave optics and geometrical optics in optical
design,” Progress in Optics, Vol. 1, 31–65 (1960).
11. V. N. Mahajan, “Asymptotic behavior of diffraction images,” Can. J. Phys. 57,

1426–1431 (1979).
12. H. F. Willis, “A formula for expanding an integral as series,” Philos. Mag. 39,
455–459 (1948).
References 77
13. B. Tatian, “Asymptotic expansions for correcting truncation error in transfer

function calculations,” J. Opt. Soc. Am. 61, 1214–1224 (1971).
14. P. P. Clark, J. W. Howard, and E. R. Freniere, “Asymptotic approximation to the

encircled energy function for arbitrary aperture shapes,” Appl. Opt. 23, 353–357
(1983).
15. V. N. Mahajan, “Line of sight of an aberrated optical system,” J. Opt. Soc. Am. 2,
833–846 (1985).
16. Reference 9, p. 164.
17. K. Strehl, “Ueber Luftschlieren und Zonenfehler,” Zeitschrift fur

instrumentenkunde 22, 213–217 (1902).
18. V. N. Mahajan, “Luneburg apodization problem I,” Opt. Lett. 5, 267–269 (1980).
19. A. Maréchal, “Etude des effets combines de la diffraction et des aberrations

geometriques sur l'image d'un point lumineux,” Revue d'Optique 26, 257–277
(1947).
20. B. R. A. Nijboer, “The diffraction theory of optical aberrations. Part I: General

discussion of the geometrical aberrations,” Physica 10, 679–692 (1943); “The
diffraction theory of optical aberrations. Part II: Diffraction pattern in the presence
of small aberrations,” Physica 13, 605–620 (1947); and K. Nienhuis and B. R. A.
Nijboer, “The diffraction theory of optical aberrations. Part III: General formulae
for small aberrations: experimental verification of the theoretical results,” Physica
14, 590–608 (1949). The second paper gives a derivation of the Strehl ratio.
21. V. N. Mahajan, “Strehl ratio for primary aberrations in terms of their aberration
variance,” J. Opt. Soc. Am. 73, 860–861 (1983).
22. H. H. Hopkins, “The aberration permissible in optical systems,” Proc. Phys. Soc.
(London) B52, 449–470 (1957).
23. S. Szapiel, “Hopkins variance formula extended to low relative modulations,”

Optica Acta 33, 980–999 (1986).
24. A. Papoulis, Systems and Transforms with Applications in Optics (McGraw-Hill,

New York, 1965), p. 100.
25. H. H. Hopkins, “Image formation with coherent and partially coherent light,” Soc.
Photo. Scient. Eng. 33, 980–999 (1986).
26. W. T. Welford, Aberrations of Optical Systems (Hilger, Philadelphia, 1989).
27. J. D. Gaskill, Linear Systems, Fourier Transforms, and Optics (Wiley, New York,
1978).
78 IMAGE FORMATION
PROBLEMS
1. Free-Space PSF and Transfer Function: Show that the transfer function of free-
space propagation in the Fresnel approximation given by Eq. (1-28) is the Fourier
transform of the corresponding PSF given by Eq. (1-24).
2. Angular Spectrum: (a) Determine the angular spectrum of a circular aperture of

diameter D for a plane wave of wavelength incident normally on it. (b) Repeat
the problem for a circular opaque disc of diameter D.
3. Aberrations of a Diffracted Wave: Determine the primary aberrations of the

diffracted wave when a plane wave is incident on an aperture and observations are
made on a hemisphere. Repeat the problem when a spherical wave is incident on
the aperture and observations are made on a hemisphere passing through the focus
of the incident wave with its center of curvature lying at the center of the aperture.
4. Imaging by a Thin Lens: Consider imaging of an object by a thin lens of focal

length f lying at a distance zo from it. Using Fresnel propagation from the object
plane to the lens, quadratic phase approximation of the lens resulting from focusing
by it, and Fresnel propagation from the lens to an observation plane at a distance zi
from it, show that the complex amplitude in the observation plane is given by Eq.
(1-49).
5. Show that the PSF is symmetric and the OTF is real for a system with a symmetric
r r r r
( ) ( )
pupil function, i.e., show that if P
rp = P rp , then PSF (
ri ) = PSF ( ri ) and
r *r
0(v ) = 0 (v ) .
6. Edge-Spread Function: The spectrum of the image of an isoplanatic incoherent

object is given by the product of the spectrum of the object and the OTF of the
imaging system. Consider an edge object and determine its image in terms of the
OTF by inverse Fourier transforming the spectrum of its image.
7. Image Sharpness: Consider an extended but isoplanatic incoherent image with an

r r r
irradiance distribution I (r ) . Show that its sharpness defined as I 2 (r ) dr is
maximum when it is aberration-free.
8. Rectangular Pupil: Consider an imaging system with a rectangular pupil of half-

widths a and b. Determine its PSF and OTF with and without defocus. Determine
its cutoff frequency.
CHAPTER 2
OPTICAL SYSTEMS WITH CIRCULAR PUPILS
2.1 Introduction ............................................................................................................81

2.2 Aberration-Free System ........................................................................................82
2.2.1 Point-Spread Function ............................................................................... 82
2.2.2 Encircled Power......................................................................................... 87
2.2.3 Ensquared Power ....................................................................................... 88
2.2.4 Excluded Power ......................................................................................... 90
2.2.5 Optical Transfer Function ..........................................................................93
2.2.6 PSF and Encircled Power From OTF ........................................................96
2.3 Strehl Ratio and Aberration Tolerance ............................................................... 97
2.3.1 Strehl Ratio ................................................................................................97
2.3.2 Primary Aberrations ..................................................................................98
2.3.3 Balanced Primary Aberrations ..................................................................99
2.3.4 Comparison of Approximate and Exact Results......................................101
2.3.5 Rayleigh’s l 4 Rule ............................................................................... 102
2.3.6 Strehl Ratio for Nonoptimally Balanced Aberrations ............................. 103
2.4 Balanced Aberrations and Zernike Circle Polynomials ................................... 105
2.5 Defocused System ................................................................................................. 110
2.5.1 Point-Spread Function ............................................................................. 111
2.5.2 Focused Beam..........................................................................................113
2.5.3 Collimated Beam ..................................................................................... 119
2.6 PSFs for Rotationally Symmetric Aberrations..................................................121
2.6.1 Theory ......................................................................................................121
2.6.2 Numerical Results....................................................................................124
2.6.3 Gaussian Approximation ......................................................................... 134
2.6.4 Summary of Results ................................................................................135
2.7 Symmetry Properties of an Aberrated PSF....................................................... 136
2.7.1 General Theory ........................................................................................137
2.7.2 Symmetry About the Gaussian Image Plane ........................................... 138
2.7.3 Symmetry of Axial Irradiance ................................................................. 141
2.7.4 Symmetry in Aberration Coefficient ....................................................... 141
2.8 PSFs for Primary Aberrations ............................................................................142
2.8.1 Defocus ....................................................................................................142
2.8.2 Spherical Aberration Combined with Defocus ........................................142
2.8.3 Astigmatism Combined with Defocus..................................................... 144
2.8.4 Coma ........................................................................................................148
2.8.5 2-D PSFs ..................................................................................................150
2.8.6 Comparison of Diffraction and Geometrical PSFs ..................................157
2.9 Line of Sight of an Aberrated System ................................................................159
2.9.1 PSF and its Centroid ................................................................................159
2.9.2.1 Wavefront Tilt ........................................................................162
2.9.2.2 Primary Coma ........................................................................162
79
80 OPTICAL SYSTEMS WITH CIRCULAR PUPILS
2.9.2.3 Secondary Coma ....................................................................165

2.9.3 Comments ................................................................................................168
2.10 Diffraction OTF for Primary Aberrations ........................................................169
2.10.1 General Relations ....................................................................................169
2.10.2 Defocus ....................................................................................................172
2.10.3 Spherical Aberration ................................................................................173
2.10.4 Astigmatism............................................................................................. 173
2.10.5 Coma ........................................................................................................176
2.11 Hopkins Ratio ....................................................................................................... 182
2.11.1 Tolerance for Primary Aberrations ..........................................................182
2.11.2 Defocus ....................................................................................................182
2.11.3 Hopkins Ratio in Terms of Variance of Aberration Difference
Function ................................................................................................... 185
2.11.4 Variance of Aberration Difference Function for Primary Aberrations ... 186
2.12 Geometrical OTF ................................................................................................. 187
2.12.1 General Relations ....................................................................................188
2.12.2 Radially Symmetric Aberrations ............................................................. 189
2.12.3 Defocus ....................................................................................................189
2.12.4 Spherical Aberration Combined with Defocus ........................................190
2.12.5 Astigmatism Combined with Defocus..................................................... 190
2.12.6 Coma ........................................................................................................191
2.13 Incoherent Line- and Edge-Spread Functions ..................................................191
2.13.1 Theory ......................................................................................................192
2.13.1.1 LSF From PSF ......................................................................192
2.13.1.2 LSF From Pupil Function..................................................... 192
2.13.1.3 Struve Ratio and Aberration Tolerances ..............................193
2.13.1.4 LSF From OTF ..................................................................... 196
2.13.1.5 ESF From OTF ..................................................................... 198
2.14 Miscellaneous Topics ........................................................................................... 205
2.14.1 Polychromatic PSF ..................................................................................205
2.14.2 Polychromatic OTF ................................................................................. 208
2.14.3 Image of an Incoherent Disc ....................................................................209
2.14.4 Pinhole Camera........................................................................................218
2.15 Coherent Imaging ................................................................................................222
2.15.1 Coherent Spread Function ....................................................................... 222
2.15.2 Coherent Transfer Function..................................................................... 223
2.15.3 Coherent LSF........................................................................................... 224
2.15.4 Coherent ESF........................................................................................... 229
2.15.5 Image of a Coherent Disc ........................................................................234
2.15.6 Use of a Lens for Obtaining Fourier Transforms ....................................238
2.15.7 Comparison of Coherent and Incoherent Imaging ..................................241
References ......................................................................................................................253
Problems ......................................................................................................................... 257
Chapter 2
Optical Systems with Circular Pupils
2.1 INTRODUCTION
In this chapter, we consider the imaging properties of a system with a circular exit
pupil by applying the general formulas derived in Chapter 1. We derive expressions for
its point-spread function (PSF), optical transfer function (OTF), and encircled, ensquared,
and excluded powers. The effect of aberrations on a PSF is first studied in terms of its
central value, i.e., its Strehl ratio. An exact expression is obtained for the Strehl ratio and
the results for primary aberrations are compared with those obtained from the
approximate expressions based on the phase variance of the aberration across the pupil
derived in Chapter 1. The tolerance for primary aberrations are given for a certain value
of the Strehl ratio. Balanced primary aberrations are also discussed and they are identified
with the Zernike circle polynomials. A defocused PSF is considered next and it is shown
that systems with large Fresnel numbers, such as photographic systems, have a very small
depth of focus. However, systems with small Fresnel numbers, e.g., a laser transmitter,
have a large depth of focus. Focused and collimated beams are discussed, and the concept
of near- and far-field distances is introduced. It is shown that if the Fresnel number of the
beam focused on a target is small, the axial irradiance is maximum at a point that is closer
to the focusing optics. However, maximum central irradiance on a target at a given
distance is obtained when a beam is focused on it, even though a larger irradiance is
obtained at a closer point.
Aberrated PSFs for rotationally symmetric aberrations are considered, and numerical
results are obtained that show that the size of their central spot is practically independent
of the aberration. Symmetry properties of the aberrated PSFs are considered next. Full
PSFs for primary aberrations are shown graphically and pictorially. A brief comparison
of the diffraction and geometrical PSFs is also given. The line of sight of an optical
system is identified with the centroid of its PSF and discussed for aberrated systems. It is
shown that only coma-type aberrations change the line of sight. The OTFs for primary
aberrations are discussed, and the concept of contrast reversal is described. The OTF
aberrated by coma is complex, and we give its real and imaginary parts as well as its
modulus and phase. Aberration tolerances are given for a certain value of the Hopkins
ratio. The geometrical OTF is also discussed and approximate expressions for it are
obtained for the primary aberrations. It gives reasonably accurate results in the region of
very low spatial frequencies, a region of practical interest when the MTF of an aberrated
system at high frequencies is practically negligible. It is shown that the MTF averaged
over all angular orientations of a spatial frequency vector is maximum when the variance
of the ray aberration, or the root mean square (rms) radius of the PSF with respect to its
centroid, is minimum. The diffraction OTF is shown to approach the geometrical OTF in
the limit of zero wavelength. A summary of the differences between the diffraction and
geometrical OTFs is also given. The incoherent line- and edge-spread functions are
81
discussed, and numerical results for both the aberration-free and the aberrated systems are
given. The aberration tolerances for a certain value of the Struve ratio are also given.
A few miscellaneous topics are considered next. The changes in the monochromatic
PSF as the spectral bandwidth increases are discussed. It is shown that the diffraction
rings do not disappear until the relative bandwidth approaches almost unity. The
aberration-free polychromatic OTF is also considered, and it is shown that it decreases at
every spatial frequency as the spectral bandwidth increases. Next, the image of a
uniformly radiating or illuminated disc is considered as an example of an extended
incoherent object. As a corollary, we show that an object whose Gaussian image size is
less than or equal to one quarter of the Airy disc can be treated as a point, thereby
establishing a criterion for the size of a pinhole. A pinhole camera is also considered, and
a relationship between the radius of the pinhole and camera length is obtained from
different points of view.
Finally, coherent imaging is discussed briefly, and expressions for the point-spread,
transfer, line-spread, and edge-spread functions are given. Both aberration-free and
aberrated systems are considered. The image of a coherently illuminated disc is
considered as an example of an extended coherent object. It is shown that the image
obtained is distorted compared to that of an incoherent disc. The Fourier-transforming
property of wave propagation is considered, and it is shown how the image of an object
illuminated by a plane wave can be altered by spatial filtering of its spectrum obtained in
the focal plane of the imaging lens. The chapter ends with a brief comparison of coherent
and incoherent imaging. In particular, Rayleigh criterion of two-point resolution is
discussed.
2.2 ABERRATION-FREE SYSTEM

In this section, we discuss the point-spread function, encircled, ensquared, and
excluded powers, and the optical transfer function of an aberration-free system. The
results of this section provide the basis for studying the effects of aberrations discussed in
the following sections. The entrance and exit pupils of an imaging system that is
rotationally symmetric about its optical axis are circular for an on-axis point object. For
an off-axis point object, the pupils will deviate from the circular form due to pupil
aberrations and or vignetting, as discussed in Sections 5.15 and 2.2.4 of Part I (see Figure
2-3c of Part I). However, such pupils can be approximated by an ellipse which, in turn,
can be converted to a circle by coordinate scaling. For systems with small fields of view,
the deviations of the pupils from a circular form may be negligible.
2.2.2.1 Point-Spread Function

Consider, as indicated in Figure 2-1, an optical imaging system with a circular exit
pupil of radius a, diameter D = 2 a , imaging a point object of intensity Bo (in units of
r
W/sr) radiating at a wavelength l , and located at ro in an object plane at a distance do
from its entrance pupil (not shown) of area Sen . The Gaussian image of the point object is
r r
located at a point Pg with a position vector rg = M ro in the image plane, where M is the
2.2 Aberration-Free System 83
xp xg
xi
Æ
ri
Æ
rp ri qi
qp Æ
rg
rp R Pg
z
Op Og
Guassian
image plane
yi
Exit
yp pupil yg
zg
Figure 2-1. Geometry of imaging indicating an optical system with a circular exit
pupil of radius a forming an image in a plane at a distance z g ~ R , where R is the
radius of curvature of the spherical wave emerging from the exit pupil with its
r
center of curvature at the Gaussian image point Pg having a position vector rg . The
r
position vector of the observation point in the Gaussian image plane is ri with
respect to Pg . (For a complete schematic of imaging, see Figure 1-2.)
magnification of the image. The solid angle subtended by the entrance pupil at the point
object is given by Sen d02 . Let h be the transmission factor of the system for light
propagation from its entrance pupil to its exit pupil.
Assuming that the exit pupil is uniformly illuminated and aberration free, a spherical
wave originating at the point object and incident on the system emerges from its exit
pupil as a spherical wave converging to the Gaussian image point Pg . The irradiance at a
r r
point ri in the Gaussian image plane with respect to the Gaussian image point rg is given
by Eq. (1-65):
2
r Û Ê 2p i r rˆ r
Ii (ri ) = Ii (0)[ 2
Sex ] Ù exp ÁË - l R rp
ı
◊ ri ˜¯ d rp , (2-1)
where
Ii (0) = Pex Sex l2 R 2 (2-2a)
= p Pex 4l2 F 2 (2-2b)

r
is the irradiance at the Gaussian image point ri = 0 ;
(
Pex = h Sen do2 Bo ) (2-3)
is the total power in the exit pupil, and, therefore, in the image; R is the radius of
curvature of the spherical wavefront at the exit pupil passing through its center and
centered at the Gaussian image point;
Sex = p a 2 (2-4)
r
is the area of the exit pupil over which the integration in Eq. (2-1) is carried out; and rp is
r
the position vector of a point in its plane such that rp £ a . The quantity F in Eq. (2-2b)
is given by
F = R D (2-5)
and represents the focal ratio (f-number) of the image-forming light cone exiting from the
exit pupil. The aberration-free properties of an optical system, such as the irradiance
distribution given by Eq. (2-1), are often referred to as its diffraction-limited properties. A
uniformly illuminated pupil is called an unapodized pupil. If we let
r
(
rp = rp cos qp , sin qp ) , 0 £ rp £ a , 0 £ qp < 2 p , (2-6a)
and
r (2-6b)
ri = ri (cos q i , sin q i ) , 0 £ q i < 2 p ,
Eq. (2-1) may also be written

a 2p 2
Û Û È 2p i ˘
[
Ii (ri , q i ) = Ii (0) 2
Sex ] Ù Ù
ı ı
exp Í-
Î lR
( )
rp ri cos q p - q i ˙ rp drp d qp
˚
. (2-7)
0 0
For simplicity of equations as well as numerical analysis, we use normalized

quantities,
r r
r = rp a (2-8a)
(
= r cos q p , sin q p ) , (2-8b)
r r
r = ri l F (2-9a)
= r (cos q i , sin q i ) , (2-9b)
and
r r
I (r ) = Ii (ri ) Ii (0) , (2-10)
where 0 £ r £ 1. Using normalized quantities, Eq. (2-7) may be written

1 2p 2
Û Û
[ (
I (r, q i ) = p -2 Ù Ù exp - pir r cos qp - q i r dr d qp
ı ı
)] . (2-11)
0 0
Noting that1
2p
Û
Ù exp (ix cos a ) da = 2 p J0 ( x ) , (2-12)
ı
0
◊
where J 0 ( ) is the zero-order Bessel function of the first kind, Eq. (2-11) reduces to
1 2
ÈÛ ˘
I (r ) = 4 ÍÙ J 0 ( p r r) r dr˙ . (2-13)
ÍÎı ˙˚
0
Noting further that

a
Û a
Ù x J 0 (bx ) dx = J1 ( ab) , (2-14)
ı b
0
◊
where J1 ( ) is the first-order Bessel function of the first kind, Eq. (2-13) becomes
[
I (r ) = 2 J1 ( p r ) p r ]2 . (2-15)
We note that the irradiance distribution is radially symmetric about the Gaussian image
point r = 0 ; i.e., the irradiance at a point (r, q i ) in the image plane is independent of q i .
This is a consequence of the fact that the pupil function for the aberration-free circular
exit pupil is radially symmetric. Indeed, it is uniform across the pupil. From Eq. (1-59a),
we note that
r r
PSF(ri ) = Ii (ri ) Pex . (2-16)
Figure 2-2a shows the radially symmetric irradiance distribution in the image plane
as described by Eq. (2-15). It is also shown in a pictorial form in Figure 2-2b. This figure
is referred to as the Airy diffraction pattern for a circular aperture.2 We note that the
principal maximum of the irradiance distribution occurs at the Guassian image point r =
0. Since the system is aberration free, all of the Huygens’ spherical wavelets originating
at the spherical wavefront at the exit pupil arrive in phase at this point. Therefore, they
interfere constructively, thereby yielding a principal maximum at this point. Note that
È 2 J ( x) ˘
Limit Í 1 ˙ = 1 . (2-17)
x Æ 0 Î x ˚
The minima of the irradiance distribution are zero at the positions given by the roots of
1.0
0.8
P
I(r), P(rc)
0.6
0.4
0.2 I
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
r, rc
(a)
(b)
Figure 2-2. (a) Irradiance and encircled power distributions for an aberration-free
system with a circular pupil. (b) 2-D PSF, called the Airy pattern.
J1 ( p r ) = 0, r π 0 . (2-18)
Noting that
d È J1 ( x ) ˘ J (x)
= - 2 , (2-19)
dx ÍÎ x ˙˚ x
the positions of secondary maxima are given by the roots of
J2 ( p r ) = 0, r π 0 , (2-20)
◊
where J2 ( ) is the second-order Bessel function of the first kind. The positions of several
minima and maxima of the irradiance distribution are given in Table 2-1. For increasing
values of r, the separation between two successive minima or two successive maxima
approaches a value of unity.

The encircled power, i.e., the amount of power contained in a circle of radius rc
in the image plane centered on the Gaussian image point, is given by
rc
Û
Pi (rc ) = 2 p Ù Ii (ri ) ri dri . (2-21a)
ı
0
Substituting Eqs. (2-9a) and (2-10) into Eq. (2-21a) and defining a normalized or
fractional encircled power
P (rc ) = Pi (rc ) Pex , (2-21b)
Table 2-1. Irradiance and encircled power corresponding to the maxima and
minima of the PSF. The irradiance is normalized by the central value
Ii (0) = Pex Sex l 2 R 2 , and the encircled power is normalized by the total power Pex
in the exit pupil and, therefore, in the image. r and rc are in units of l F.
Max/Min r, rc I(r ) P( rc )
Max 0 1 0
Min 1.22 0 0.838
Max 1.64 0.0175 0.867
Min 2.23 0 0.910
Max 2.68 0.0042 0.922
Min 3.24 0 0.938
Max 3.70 0.0016 0.944
Min 4.24 0 0.952
Max 4.71 0.0008 0.957
we obtain
rc l F
Û
(
P (rc ) = p 2 2 ) Ù I (r ) rdr .
ı
(2-22a)
0
Or, if we let rc be in units of l F ,

rc
Û
P (rc ) ( 2
)
= p 2 Ù I (r ) r dr .
ı
(2-22b)
0
Substituting Eq. (2-15) into Eq. (2-22b) and noting that

xc
2
Û 2 J1 ( x )
Ù dx = 1 - J 02 ( xc ) - J12 ( xc ) , (2-23)
ı x
0
we obtain3
P (rc ) = 1 - J 02 ( p rc ) - J12 ( p rc ) . (2-24)
Since, according to Eq. (2-18), the dark rings (minima of zero irradiance) correspond to
J1 ( p r ) = 0 , we note that the powers inside and outside an m-th dark ring of radius rm (in
units of l F ) are given by
Pin (rm ) = 1 - J 02 ( p rm ) , (2-25a)

and
Pout (rm ) = J 02 ( p rm ) , (2-25b)
respectively.
Figure 2-2a also shows how the encircled power P (rc ) varies with rc . The relative
irradiance and encircled power corresponding to maxima and minima of the irradiance
distribution are given in Table 2-1. We note that the radius of the first dark ring is 1.22 (in
units of lF ), and it contains 83.8 percent of the total power. The central bright spot of
radius 1.22 is called the Airy disc of the Airy pattern. The first bright ring contains 7.2
percent, the second bright ring 2.8 percent, and the third bright ring 1.4 percent of the
total power.
2.2.3 Ensquared Power

Because of the widespread use of detector arrays with square detector elements, it is
of interest to know the amount of power contained in a square region in the image plane.4
The ensquared power in a square region of half width rs centered on the Gaussian image
point in the image plane is given by
ÛÛ
Pi (rs ) = ÙÙ Ii (ri ) ri dri d q i , (2-26)
ıı
det
where the integration is carried over the square region of the detector. Integration over q i
yields
2 rs
Û
Pi (rs ) = Ù Ii (ri ) q i (ri ) dri , (2-27)
ı
0
where q i (ri ) is the angle subtended by detector points lying at a distance ri from its
center. For ri £ rs , it is evident that q i (ri ) = 2 p , as in the case of a circular detector.
However, as indicated in Figure 2-3, only some of the points for which rs < ri £ 2 rs lie
on the square detector. For a given value of ri , the angle subtended by these points is
made up of four equal but discontinuous parts, one in each quadrant. Each part
contributes an amount ( p 2) - 2 cos -1 (rs ri ) to the angle q i ( ri ) . Hence, this angular
function may be written
q i (ri ) = 2 p , 0 £ ri £ rs (2-28a)
= 2 p - 8 cos -1 (rs ri ) , rs < ri £ 2 rs . (2-28b)
Substituting Eqs. (2-28) into Eq. (2-27), we obtain
yi
ri
qi(ri)/4
xi
rs
Figure 2-3. Angular function q i ( ri ) for a square detector of half-width rs . q i ( ri ) 4

as shown represents the angle subtended by those detector points in its first
quadrant which lie at a distance ri from its center.
2 rs 2 rs
Û Û
Pi (rs ) = 2 p Ù Ii (ri )ri dri - 8 Ù Ii (ri ) cos -1 (rs ri ) ri dri . (2-29)
ı ı
0 rs
The first term on the right-hand side of Eq. (2-29) represents the image power contained
in a circle of radius of 2 rs , and the second term gives the power contained between a
circle of radius 2 rs and a square of half-width rs . Substituting Eqs. (2-9), (2-10), and
(2-15) into Eq. (2-29), the ensquared power may be written
2
8Û du
Ps (rs ) = Pc ( )
2 rs - Ù J12 ( p rs u) cos -1 (1 u)
pı u
, (2-30)
1
where
Ps (rs ) = Pi (rs ) Pex (2-31a)
and
Pc (rc ) = P (rc ) (2-31b)
are the fractional ensquared and encircled powers, u = ri rs , and rs is in units of l F .

Figure 2-4 and Table 2-2 show the relative encircled and ensquared powers in graphical
and tabulated forms, respectively. We note that both Ps and Pc start at zero and their
difference becomes maximum and equal to 7.3 percent of the total power when
rc = rs = 0.5 . As I (r ) fluctuates undergoing maxima and minima, so does Ps - Pc .
2.2.4 Excluded Power

We have calculated above the included power, encircled or ensquared, depending on
whether the detector is circular or square. It is also useful to define the excluded power,
i.e., the image power contained outside a certain area,
Xi = Pex - Pi , (2-32)
where Pi is the corresponding included power. Equation (2-25b) gives the excluded
power for an m-th dark ring of the Airy pattern. For arbitrary sizes, the excluded power
can be calculated quite accurately in a closed form, if the included area is large enough so
that Xi £ 0.1 Pex .
For large arguments, we can use the asymptotic expression for Bessel functions,
namely, that1
J1 ( z ) ~ (2 pz )1 2 sin ( z - p 4) . (2-33)
Thus, for large values of r, Eq. (2-15) can be written

1.0
I(r)
Ps
Pc
I (r), Pc (rc), Ps (rs) 0.8
0.6
10(Ps – Pc)
0.4
0.2
0.0
0 1 2 3 4
r, rc, rs
Figure 2-4. Encircled and ensquared power distributions for a circular pupil. The
irradiance distribution and the difference between ensquared and encircled power
distributions are also shown.
I (r ) ~ (8 p 4 r 3 ) sin 2 [p (r - 1 4)] . (2-34)
Noting that the average value of a sine square is half, the average irradiance (indicated
by a bar) for large values of r is given by
I (r ) ~ 4 p4r 3 . (2-35)
For a large circular detector of radius rc , following Eq. (2-22b), the excluded power
in units of Pex may be written
•
Û
Xc (rc ) ~ ( 2
)
p 2 Ù I (r ) r dr
ı
rc
= 2 p 2 rc . (2-36)
Similarly, for a large square detector of half-width rs , the excluded power is given by
Table 2-2. Irradiance and encircled and ensquared powers for a circular pupil. r is
the radial distance of a point in the image plane from the Gaussian image point in
units of l F. rc and rs represent in units of l F the radius and half-width of a
circular and square detector, respectively, centered in the image plane at the
Gaussian image point. 4
r , rc , rs I(r ) Pc ( rc ) Ps ( rs )
0.0 1.0000 0.0000 0.0000

0.1 0.9756 0.0244 0.0309
0.2 0.9053 0.0940 0.1178
0.3 0.7975 0.1989 0.2444
0.4 0.6645 0.3248 0.3889
0.5 0.5209 0.4559 0.5290
0.6 0.3806 0.5775 0.6475
0.7 0.2558 0.6785 0.7351
0.8 0.1544 0.7532 0.7910
0.9 0.0803 0.8011 0.8209
1.0 0.0328 0.8264 0.8339
1.2 0.00018 0.8378 0.8434
1.4 0.00846 0.8417 0.8603
1.6 0.01729 0.8623 0.8839
1.8 0.01355 0.8896 0.9020
2.0 0.00457 0.9064 0.9112
2.2 0.00008 0.9099 0.9161
2.4 0.00148 0.9110 0.9217
2.6 0.00390 0.9180 0.9291
2.8 0.00361 0.9287 0.9360
3.0 0.00141 0.9359 0.9402
3.5 0.00103 0.9394 0.9471
4.0 0.00061 0.9513 0.9548
4.5 0.00048 0.9534 0.9591
5.0 0.00031 0.9607 0.9638
5.5 0.00026 0.9621 0.9666
6.0 0.00018 0.9671 0.9698
6.5 0.00016 0.9681 0.9718
7.0 0.00012 0.9717 0.9741
7.5 0.00010 0.9724 0.9756
8.0 0.00008 0.9752 0.9773
Xs (rs ) ~ Û Û
Ù dx Ù dyI (r )
ı ı
x > rs y > rs
= 4 2 p 3rs , (2-37)
12
(
where r = x 2 + y 2 ) . From Eqs. (2-36) and (2-37) we note that
Xs (rc ) = 0.9 Xc (rc ) . (2-38)
When rc or rs is greater than 1.6, Eqs. (2-36) and (2-37) give the excluded power
accurate to within 1 percent of the total power, as may be seen by comparing the results
given in Table 2-2.
The approximate result of Eq. (2-35) and those that follow from it, although obtained
for the aberration-free case, are valid even when aberrations are present in the system.
[This may be seen by substituting Eq. (2-46) given later into Eq. (1-154) and considering
the normalizations used in these equations.] It should be noted, however, that the value of
r for which Eq. (2-35) is valid increases as aberrations are introduced into the system.
Moreover, if the aberration is rotationally nonsymmetric, the irradiance distribution in an
image plane is not radially symmetric, but the first term of its asymptotic expansion is
symmetric and given by Eq. (2-35). Similarly, Eq. (2-36) may be used to obtain the
excluded power even when aberrations are present. However, the value of rc for which it
is valid increases as aberrations are introduced into the system.
2.2.5 Optical Transfer Function

For an optical imaging system with a uniformly illuminated aberration-free circular
exit pupil of radius a, Eq. (1-73b) for its OTF reduces to
r Û r r r r
ı
( ) (
t (vi ) = Pex-1 Ù A rp A rp - l R vi d rp ) , (2-39)
where
(r )
A rp = ( Pex Sex )
12
,
r
rp £ a ,
= 0 , otherwise . (2-40)
It is evident that the OTF is radially symmetric and represents the fractional area of
overlap of two circles, each of radius a, separated by a distance l Rvi .
From Figure 2-5, we note that the area of overlap is given by four times the
difference between the area of a sector of radius a and cone angle b and the area of the
triangle OAB, i.e.
a
b
O
A O¢
lRni
Figure 2-5. Aberration-free OTF as the fractional area of overlap of two circles
whose centers are separated by a distance l Rvi .
Û r r r r Ê b 1 ˆ
ı
( ) ( )
Ù A rp A rp - l Rvi d rp = 4( Pex Sex ) ÁË
2 p
p a 2 - OA ◊ AB˜
2 ¯
= (2Pex p)(b - sin b cos b) , (2-41)
where
cosb = OA OB
= l R vi 2 a
= vi (1 l F ) . (2-42)
Let
v = vi (1 l F ) (2-43)
be a normalized radial spatial frequency. Thus, cos b = v . Hence, substituting Eq. (2-41)
into Eq. (2-39) and writing b in terms of v, we obtain
1 2˘
ÍÎ (
t (v) = (2 p) Ècos -1 v - v 1 - v 2 ) ˙˚
, 0 £ v £1 ,
= 0 , otherwise . (2-44a)
It may also be written in the form
t (v) = (1 p) [2b - sin 2b] , 0 £ v £ 1 . (2-44b)
The spatial frequency v = 1 or vi = 1 l F is called the cutoff frequency of an

incoherent imaging system, since the OTF for v ≥ 1 is zero. Figure 2-6 shows how the
OTF given by Eq. (2-44) varies with v. Its numerical values are given in Table 2-3.
1.0
0.8
0.6
t
0.4
0.2
0.0
0.0 0.2 0.4 0.6 0.8 1.0
n
Figure 2-6. Aberration-free OTF.
Table 2-3. Numerical values of the aberration-free OTF.
v t (v )
0 1.000
0.1 0.873
0.2 0.747
0.3 0.624
0.4 0.505
0.5 0.391
0.6 0.285
0.7 0.188
0.8 0.104
0.9 0.037
1.0 0.000
According to Eqs. (1-83) and (1-84), the PSF and the OTF of an optical system with a
radially symmetric pupil function are related to each other by a zero-order Hankel
transform. It can be shown that the PSF given by Eq. (2-15) and the OTF given by Eq. (2-
44) for a system with a circular pupil are related to each other by such a transform. We
also note that
1
Û
Ù t (v) v dv = 1 8 . (2-45)
ı
0
The slope of the OTF at the origin is given by

t ¢( 0 ) = - 4 p . (2-46)
As pointed out in Section 1.6.4, the slope given by Eq. (2-46), although obtained from the
aberration-free OTF given by Eq. (2-44) is aberration independent. Equation (2-46) can
also be obtained from Eq. (1-167) by noting that t ¢(0, f) is independent of f , v is in
units of 1 l F , L = 2 p a , and Sex = p a 2 .
2.2.6 PSF and Encircled Power From OTF

For an aberration-free system with a circular exit pupil under consideration, the PSF
is given by
PSF(r ) = I (r ) Pex , (2-47)
where I(r) is given by Eq. (2-15). The corresponding OTF is given by Eq. (2-44). To
relate the two, we write, for example, Eq. (1-83) in terms of the irradiance distribution.
Thus,
Û
t (vi ) = (2 p Pex ) Ù Ii (ri ) J 0 (2 pvi ri ) ri dri . (2-48)
ı
Substituting Eqs. (2-9a), (2-10), and (2-43), Eq. (2-48) may be written
Û
t (v) = ( p 4) 2 p Ù I (r ) J 0 (2 pvr ) rdr . (2-49a)
ı
Similarly,
Û
I (r ) = ( 4 p) 2 p Ù t (v) J 0 (2 pvr )v dv . (2-49b)
ı
The functions I (r ) and t (v) given by Eqs. (2-15) and (2-44) are related to each other,
according to Eqs. (2-49), by slightly modified zero-order Hankel transforms. These
modifications consist of the factors of p 4 and 4 p on the right-hand side of these
equations, because of the particular normalizations used; namely, I (0) = 1 and t (0) = 1 .
Substituting Eq. (2-49b) into Eq. (2-22b), we obtain

rc
Û Û
P(rc ) 2
= 4 p Ù t (v)v dv Ù J 0 (2 pvr ) rdr
ı ı
0
Û
= 2 p rc Ù t (v) J1 (2 prc v) dv . (2-50)
ı
Equations (2-49) and (2-50) also hold for a radially symmetric aberration, as may be seen
by comparing with Eqs. (1-83), (1-85), and (1-89), respectively.
2.3 Strehl Ratio and Aberration Tolerance 97
2.3 STREHL RATIO AND ABERRATION TOLERANCE

In this section, we determine the tolerance values for primary aberrations using
approximate expressions for the Strehl ratio in terms of the aberration variance developed
in Section 1.9. The approximate results are compared with the exact results to determine
the range of their validity. Rayleigh’s quarter-wave rule is discussed, and it is shown that
the tolerance determined in terms of the standard deviation of an aberration gives a better
approximation to the exact results. It is shown that the aberrations balanced to give
minimum variance yield a higher Strehl ratio only for small aberrations. Aberration
tolerances based on the Hopkins ratio in the spatial frequency domain are discussed in
Section 2.11.
2.3.1 Strehl Ratio

Following the normalizations and notation of Section 2.2.1, Eq. (1-64) for the
irradiance distribution of the image of a point object formed by an aberrated optical
system with a uniformly illuminated circular pupil may be written
1 2p 2
Û Û
[ ( )] [
I (r, q i ) = p -2 Ù Ù exp i F r, qp exp - pi rr cos qp - q i r dr d qp
ı ı
( )] , (2-51)
0 0
( ) (
where F r, q p is the phase aberration of the system at a point r, qp in the plane of its )
exit pupil. By definition, the Strehl ratio of the image or the system is given by the ratio
of the irradiances at the center r = 0 with and without aberrations. The aberration-free
central irradiance is unity in units of Pex Sex l2 R 2 [as may be seen from Eqs. (2-2a) and
(2-51)]. Hence, the central irradiance obtained from Eq. (2-51) is the Strehl ratio, i.e.,
1 2p 2
Û Û
S = p -2
ı ı
[
Ù Ù exp i F(r, q) r dr dq ] , (2-52)
0 0
where we have dropped the subscript p on the angle q p for simplicity. Approximate
expressions for the Strehl ratio when the aberration is small are given by Eqs. (1-204)
through (1-206), i.e.,
2
S1 ~ (1 - s 2F 2) , (2-53a)
S2 ~ 1 - s 2F , (2-53b)
and
S3 ~ exp (- s 2F ) , (2-53c)
where
s 2F = < F 2 > - < F > 2 (2-54)

is the variance of the phase aberration across the uniformly illuminated pupil. The mean
and the mean square values of the aberration are obtained from the expression
1 2p
Û -1 Û
< F > = p Ù Ù F n (r, q) r dr dq ,
n
(2-55)
ı ı
0 0
with n = 1 and 2, respectively.
2.3.2 Primary Aberrations

As discussed in Section 3.5 of Part I, the aberration function of a rotationally
symmetric imaging system consists of terms containing integral powers of one or more of
three rotational invariants h 2 , r 2 , and hr cos q , where h is the object height and (r, q)
are the polar coordinates of a pupil point. (Strictly speaking, q is the relative angle of the
position vector of the pupil point with respect to the position vector of the object point.
However, without loss of generality, the object point may be assumed to lie along the x
axis, in which case q is simply the polar angle of the pupil point.) The degree of each
term in the object and pupil coordinates is even. Thus, a power series expansion of the
aberration function consists of terms of degree 4, 6, 8, etc. The corresponding aberrations
are referred to as the primary, secondary, tertiary aberrations, etc. For a given object
point, the dependence of an aberration term on the object height may be suppressed by
containing it in the aberration coefficient.
Table 2-4 gives the form as well as the standard deviation s F of a primary (or a
Seidel) aberration, where an aberration coefficient Ai represents the peak value of the
aberration. It also lists the tolerance, i.e., the value of the aberration coefficient Ai , for a
Strehl ratio of 0.8. This tolerance has been obtained by using the Strehl ratio expression
S2 , according to which the standard deviation for a Strehl ratio of 0.8 is given by
sF = 0.2 (2-56a)
or
s w = (l 2 p) 0.2 = 0.07l = l 14.05 , (2-56b)
where s w is the standard deviation of the wave aberration. The aberration tolerance
listed in Table 2-4 is for the wave (as opposed to the phase) aberration coefficient, as is
customary in optics. We have used the symbol Ad for the coefficient of field curvature
aberration, which varies quadratically with the angle a point object makes with the optical
axis of the system. However, to avoid confusion, we will use the symbol Bd for
representing the defocus wave aberration, which is independent of the field angle but has
the same dependence on pupil coordinates as field curvature. Similarly, we have used the
symbol At for distortion, which varies as the cube of the field angle. However, we will
use the symbol Bt to represent wavefront tilt, which is independent of the field angle but
has the same dependence on pupil coordinates as distortion.
Table 2-4. Standard deviation and aberration tolerance for primary aberrations.
Aberration F (r, q) sF A i for S = 0.8
Spherical As r 4 2 As As l 4.19
=
3 5 3.35
Coma Ac r3 cos q Ac Ac l 4.96

=
2 2 2.83
Astigmatism Aa r2 cos 2 q Aa l 3.51

4
Field Curvature Ad r2 Ad Ad l 4.06

=
(defocus) 2 3 3.46
Distortion (tilt) At r cos q At l 7.03

2
2.3.3 Balanced Primary Aberrations

In Chapter 4 of Part I, where we discussed ray aberrations, we mixed one Seidel
aberration with another in order to minimize the size of the ray spot in an image plane.
For example, in the case of spherical aberration, the circle of least confusion was
determined to be in a plane 3/4 of the way from the Gaussian image plane to the marginal
image plane. The radius of the circle of least confusion was found to be 1/4 of the spot
radius in the Gaussian plane. Similarly, in the case of astigmatism, the circle of least
confusion was determined to be in a plane lying midway between the planes containing
the sagittal and tangential line images. This circle had a diameter equal to half the length
of the line images.
Based on diffraction, the best image for small aberrations is one for which the Strehl
ratio is maximum. Since, according to Eq. (2-54), the Strehl ratio is maximum when the
aberration variance is minimum, the best image plane is one that minimizes the variance
of the aberration. Thus, for example, we balance spherical aberration with defocus and
write it as
F(r) = As r 4 + Bd r2 . (2-57)
The defocus aberration is introduced by making an observation in a plane at a distance z

[see Eq. (2-81b) for a relationship between Bd and z]. We determine the value of defocus
Bd such that the variance s 2F is minimized; i.e., by letting
∂ s 2F
= 0 . (2-58)
∂ Bd
We thus find that the optimum value is Bd = - As , or z - R = 8 F 2 As for z ~ R . The

corresponding standard deviation is As 6 5 , which is a factor of 4 smaller than that for
Bd = 0 . Since the standard deviation has been reduced by a factor of 4 by balancing

spherical aberration with defocus, the optical tolerance has been increased by the same
factor. Moreover, since Bd = 0 and Bd = - 2 As correspond to the Gaussian and marginal
image planes, respectively, we note that, based on diffraction, the best image is obtained
in a plane lying midway between them. This is different from the plane containing the
circle of least confusion based on geometrical optics, which corresponds to Bd = -1.5 As .
Figure 2-7 shows how the aberration varies with r for various values of Bd . It is evident
that Bd = - As yields the minimum aberration variance.
Astigmatism and coma aberrations can be treated similarly. Table 2-5 lists the
form of a balanced primary aberration, its standard deviation, and its tolerance for a
Strehl ratio of 0.8 according to Eq. (2-53b). Also listed in the table is the location of the
diffraction focus, i.e., the point with respect to which the aberration variance is minimum
so that the Strehl ratio is maximum at it. We note that in the case of coma, the balancing
aberration is a wavefront tilt with a coefficient that is minus two-thirds of the coma
coefficient. Thus, the maximum Strehl ratio is obtained at a point that is displaced from
the Gaussian image point by 4 FAc 3 but lies in the Gaussian image plane. It is shown in
Section 2.9.2.2 that this is valid for Ac < ~ 0.7 l . For larger values of Ac , the peak
irradiance lies at a point that is closer to the Gaussian image point than that given by
4 FAc 3 . The standard deviation for balanced coma is 1/3 of its corresponding value
when not balanced. In the case of astigmatism, the best Strehl ratio is obtained in the
plane of the circle of least confusion, i.e. at a distance 4 F 2 Aa from the Guassian image
plane. The standard deviation of balanced astigmatism is only 1/1.225 of its
corresponding value for classical astigmatism (i.e., without any balancing).
1.00
0.75 W(r)
= r4 +(Bd /As)r2
As
0.50
Bd
=0
0.25 As
W(r)
As
0.00
–1
– 0.25
–1.5
– 0.50
–2
– 0.75
– 1.00
0.0 0.2 0.4 0.6 0.8 1.0
r
Figure 2-7. Variation of spherical aberration with r when balanced with various
amounts of defocus.
Table 2-5. Balanced primary aberrations, and corresponding diffraction focus,

standard deviation, and aberration tolerance.
Balanced Diffraction A i for

Aberration F (r, q) Focus* sF S = 0.8
As
Spherical (
As r 4 - r2 ) (0, 0, 8F A )
2
s
6 5
0.955l
Ac
Coma (
Ac r3 - 2r 3 cos q) (4 FAc 3, 0, 0 )
6 2
0.604l
Aa
Astigmatism (
Aa r2 cos 2 q - 1 2 ) (0 , 0 , 4 F A )
2
a
2 6
0.349l
= ( Aa 2) r2 cos 2q
*The diffraction focus coordinates are relative to the Gaussian image point.
2.3.4 Comparison of Approximate and Exact Results

Substituting the expressions for the various primary and balanced primary
aberrations into Eq. (2-52) and carrying out the integration, we obtain the exact
expressions for the Strehl ratio.5 These are listed in Table 2-6. We note that in the case of
coma, the integration must be carried out numerically. In the case of balanced
astigmatism, only the first few terms of the infinite series need to be considered for
adequate precision.
Figure 2-8 shows how the Strehl ratio of a primary aberration varies with its
standard deviation. Approximate as well as exact results are shown in this figure. The
curves for a given aberration and for the corresponding balanced aberration can be
distinguished from each other by their behavior for large s w values (near 0.25 l ). For
example, coma is shown by the evenly dashed curves; the lower curve being for balanced
coma. The same holds true for astigmatism. The curves for spherical and balanced
spherical aberrations are identical since the Strehl ratio for a given value of s w is the
same for the two aberrations. The following observations may be made from Figure 2-8.
i. For small values of s w , the Strehl ratio is independent of the type of aberration. It
depends only on the value of its variance, as illustrated by Eqs. (2-53).
ii. The expressions for S1 and S2 underestimate the true Strehl ratio.
iii. The expression for S3 underestimates the true Strehl ratio only for coma and
astigmatism; it overestimates for the other aberrations. The error, defined as
100 (1 - S3 S ) , is £ 10 percent as long as S > 0.3 .
Table 2-6. Primary aberrations, their peak and peak-to-valley values, and standard
deviations; and exact expression for the Strehl ratio and values of the aberration
coefficent, peak absolute aberration, and peak-to-valley aberration for a Strehl ratio
of 0.8.
S = 0.8
Aberration F(r, q) sF S Wp Wp–v Ai Wp W p -v
2 As †
Spherical As r 4
3 5
1È 2
b ÍÎ
C ( b ) + S ( b )˘˙˚
2 As As 0.25 0.25 0.25
Balanced As †
spherical As (r 4 – r 2 )
6 5
1È 2
b ÍÎ
C ( b ) + S ( b )˘˙˚2 As / 4 A s / 4 1 0.25 0.25
2
Ú ( ) ˚˙
Ac È J 2 2s x dx ˘
1 3/2
Coma Ac r 3 cos q 0 0 F Ac 2Ac 0.21 0.21 0.42
2 2 ÎÍ
2
Balanced Ac ÏÔ 1 Ê 3/2 2 1/2 ˆ ¸Ô
Ac (r 3 – 2r / 3) cos q Ì Ú0 J 0 [6 2s F Á x – x ˜ ]dx ˝ Ac / 3 2 Ac / 3 0.63 0.21 0.42
coma 6 2 ÔÓ Ë 3 ¯ Ô˛
Aa Aa Aa 0.30 0.30 0.30

Astigmatism Aa r 2 cos 2 q J 02 (2s F ) + J12 (2s F )
4
2
Balanced È • ˘
A a r 2 (cos 2q - 1 / 2)
astigmatism
= ( Aa / 2)r 2 cos2 q
Aa
2 6
Í 2 ÂJ
Í 6 s k = 0 2k +1 ( 6s F ˙
˙ ) Aa / 2 Aa 0.37 0.18 0.37
Î F ˚
b b
† b = 3 5 s F / p , C (b) = Ú cos(p x
0
2 / 2)dx, S (b) = Ú sin(p x
0
2 / 2) dx
iv. S3 gives a better approximation for the true Strehl ratio than S1 and S2 . The reason is
4
that, for small values of s w , it is larger than S1 by approximately s F 4 . Of course,
4
S1 is larger than S2 by s F 4 .
v. The Strehl ratio depends strongly on the standard deviation of an aberration but
weakly on its detailed distribution over a wide range of Strehl ratio values and not
just for large values of it.
2.3.5 Rayleigh’s l 4 Rule

Rayleigh6 showed that a quarter wave of primary spherical aberration reduces the
irradiance at the Gaussian image point by 20 percent; i.e., the Strehl ratio for this
aberration is 0.8. This result has brought forth Rayleigh’s l 4 rule; namely, that a Strehl
ratio of approximately 0.8 is obtained if the maximum absolute value of the aberration at
any point in the pupil is equal to l 4 . A variant of this definition is that an aberrated
wavefront that lies between two concentric spheres spaced a quarter-wave apart will give
a Strehl ratio of approximately 0.8. Thus, instead of Wp = l 4 , we require Wp - v = l 4 ,
where Wp is the peak absolute value and Wp - v is the peak-to-valley value of the
aberration. From Table 2-6, we note that a Strehl ratio of 0.8 is obtained for
Wp = l 4 = Wp - v for spherical aberration only. For other primary aberrations, distinctly
different values of Wp and Wp - v give a Strehl ratio of 0.8. In Table 2-6, Wp and Wp - v
are also given in terms of the aberration coefficient Ai . Thus, we see that it is
advantageous to use s w for estimating the Strehl ratio. A Strehl ratio of S > ~ 0.8 is
<
obtained for s w ~ l 14 .
1.0
0.8
0.6
S
0.4
S3
0.2
S2
S1
0.0
0.00 0.05 0.10 0.15 0.20 0.25
sw
Figure 2-8. Strehl ratio for primary aberrations as a function of their standard
deviation s w in units of optical wavelength l . For large values of s w , coma and
astigmatism give a higher Strehl ratio than the corresponding balanced aberration.
The Strehl ratios for spherical and balanced spherical aberrations for the same
value of s w are identical. s F = (2 p l ) s w . Spherical or balanced spherical....,
Coma– – – –, Astigmatism __.__.__.
2.3.6 Strehl Ratio for Nonoptimally Balanced Aberrations

When a Seidel aberration is balanced with other aberrations to minimize its variance,
the balanced aberration does not necessarily yield a higher or the highest possible Strehl
ratio. Only for small aberrations is a maximum Strehl ratio obtained according to any of
the Eqs. (2-53) when the variance is minimum. For large aberrations, however, there is no
simple relationship between the Strehl ratio and the aberration variance.
As an example, let us consider spherical aberration balanced with an arbitrary

amount of defocus, as in Eq. (2-57). Substituting it in Eq. (2-52), it can be shown that the
Strehl ratio is given by
{[
S = ( p 2 As ) C(b+ ) + C(b- ) ]
2
[
+ S(b+ ) + S(b- ) ]
2
} , (2-59a)
where
12
b± = ( As ± D ) (2 p As ) , D = As + Bd , (2-59b)
◊ ◊
As and Bd are in radians, and C( ) and S( ) are the Fresnel integrals defined in Table 2-6.
It is evident that the Strehl ratio is independent of the sign of D, the deviation of
defocus from its optimum (in the sense of minimum variance) value. Thus the axial
irradiance of a focused wave aberrated by spherical aberration is symmetric about the
axial point with respect to which the aberration variance is minimum. Figure 2-9 shows
how S varies with D in the case of circular pupils for several typical values of As . It is
seen that, for large Strehl ratios, S is maximum when D = 0 , i.e., a minimum variance
leads to a maximum of Strehl ratio. For small Strehl ratios, however, minimum variance
gives a minimum of Strehl ratio. The Strehl ratio is maximum for a nonoptimally
balanced aberration. For example, when As = 3l , the optimum amount of defocus is
Bd = - 3l , but the Strehl ratio is a minimum and equal to 0.12. The Strehl ratio is
maximum and equal to 0.26 for Bd ~ - 4l or - 2l . For As < ~ 2.3l , the axial irradiance
is maximum at a point with respect to which the aberration variance is minimum.
As a second example, we consider the case of coma balanced with an arbitrary

amount of tilt; i.e., we consider an aberration given by
W (r, q) = Ac r 3 cos q + B t r cos q , (2-60)

1.0
0.8
0.6
S
As=1l
0.4
2l
0.2
3l 4l
0.0
0 1 2 3 4
D
Figure 2-9. Strehl ratio for circular pupils aberrated by spherical aberration as a
function of the deviation of focus from its optimum balancing value ( D = As + Bd )
for several values of the aberration coefficient As . The curves are symmetric about
the origin.
2.4 Balanced Aberrations and Zernike Circle Polynomials 105
where Bt is the peak value of the tilt aberration. We note from Table 2-6 that
Bt = - 2 Ac 3 for minimum variance. Thus, the point with respect to which the aberration
variance is minimum lies in the image plane at a distance of 4 FAc 3 from the origin,
where Ac is in length units. As discussed later in Section 2.9.2.2, we find that maximum
irradiance occurs at this point only if Ac < ~ 0.7l , which in turn corresponds to
S >
~ 0 . 76 . For larger values of Ac , the distance of the point of maximum irradiance does
not increase linearly with its value, and even fluctuates in some regions. Thus, only for
large Strehl ratios, is the irradiance maximum at the point associated with minimum
aberration variance. Moreover, it is found that for Ac > 2.3l , the Seidel coma gives a
larger Strehl ratio than the balanced coma; i.e., the irradiance at the origin is larger than at
the point with respect to which the aberration variance is minimum. An examination of
the axial irradiance for astigmatism (discussed later in Section 2.8.3) shows that the peak
irradiance lies at the point Bd = - Aa 2 corresponding to minimum aberration variance
only for Aa < l .
When secondary spherical aberration and secondary coma are balanced with lower-
order aberrations to minimize their variance, it is found that a maximum of Strehl ratio is
obtained only if its value comes out to be greater than about 0.5. Otherwise, a mixture of
aberrations yielding a larger-than-minimum possible variance gives a higher Strehl ratio
than the one provided by a minimum variance mixture.7
2.4 BALANCED ABERRATIONS AND ZERNIKE CIRCLE POLYNOMIALS

Nijboer used Zernike circle polynomials to describe the diffraction theory of
aberrations.8 The phase aberration function of a system with a circular exit pupil for a
point object at a certain angle from its axis can be expanded in terms of a complete set of
Zernike circle polynomials Z nm (r, q) introduced in Section 3.5.3 of Part I, that are
orthonormal over a unit circle in the form
• n
F(r, q) = Â Â c nm Z nm (r, q) , 0£r£1 , (2-61a)
n =0 m =0
where cnm are the orthonormal expansion coefficients that depend on the field angle of
the object, n and m are positive integers including zero, n – m ≥ 0 and even, and
[
Z nm (r, q) = 2( n + 1) (1 + d m 0 ) ]1/ 2Rnm (r) cos mq . (2-61b)
Here, d ij is a Kronecker delta, and
(n - m) / 2
(-1)s (n - s)!
Rnm (r) = Â rn - 2 s (2-61b)
n+m ˆ Ên-m ˆ
s=0 s! Ê -s ! -s !
Ë 2 ¯ Ë 2 ¯
is a radial polynomial of degree n in r containing terms in rn , rn -2 , K, and rm. The

radial circle polynomials Rnm (r) are even or odd in r, depending on whether n (or m) is
even or odd. Also, Rnn (r) = rn , Rnm (1) = 1 , and Rnm (0) = d m 0 for even n 2 and - d m0 for
odd n 2 . The polynomials Rnm (r) obey the orthogonality relation
1 1
Ú Rn (r) Rn ¢ (r) r dr =
m m
d . (2-61c)
0 2(n+ 1) nn ¢
The orthogonality of the angular functions yields

2p
Ú cos mq cos m¢q dq = p (1 + d m 0 ) d mm ¢ . (2-61d)
0
Therefore, the polynomials Z nm (r, q) are orthonormal according to
1 1 2p m
Ú Ú Z (r, q)Z n ¢ (r, q) r dr d q = d nn ¢ d mm ¢
m¢
. (2-61e)
p0 0 n
The orthonormal Zernike expansion coefficients are given by
1 1 2p
Ú Ú F(r, q)Z n (r, q) r dr d q ,
m
c nm = (2-62)
p0 0
as may be seen by substituting Eq. (2-61a) and utilizing the orthonormality of the
polynomials.
The Zernike circle polynomials are unique in that they are the only polynomials in
two variables r and q, which (a) are orthogonal over a circle, (b) are invariant in form
with respect to rotation of the coordinate axes about the origin, and (c) include a
polynomial for each permissible pair of n and m values.9
The orthonormal Zernike polynomials and the names associated with some of them
when identified with aberrations are listed in Table 2-7 for n £ 8. The polynomials
independent of q are the spherical aberrations, those varying as cos q are the coma
aberrations, and those varying as cos 2q are the astigmatism aberrations. The number of
Zernike (or orthogonal) aberration terms in the expansion of an aberration function
through a certain order n is given by
2
n
Nn = Ê + 1ˆ for even n , (2-63a)
Ë2 ¯
= (n + 1)(n + 3) 4 for odd n . (2-63b)
Consider a typical Zernike aberration term in Eq. (2-61a):
n (r, q) = c nm Z n (r, q) .
Fm m
(2-64)
Unless n = m = 0 , its mean value is zero; i.e.,

1 2p 1 2p
n (r, q) = Ú Ú F n (r, q) r dr d q Ú Ú r dr d q
Fm m
0 0 0 0
= 0 , n π 0, m π 0 . (2-65a)
For m = 0 , this may be seen with the help of Eq. (2-61c) and the fact that R00 (r) = 1 is a
member of the polynomial set. The orthogonality Eq. (2-61e) yields the result that the
Table 2-7. Orthonormal Zernike circle polynomials and balanced aberrations.
n m Orthonormal Zernike Polynomial Aberration Name*

12
È 2(n + 1) ˘
Í ˙ Rnm (r) cos mq
Î 1 + d m0 ˚
0 0 1 Piston
1 1 2r cos q Distortion (tilt)
2 0 (
3 2r 2 - 1 ) Field curvature (defocus)
2
2 2 6 r cos 2q Primary astigmatism
3 1 (
8 3r3 - 2r cos q ) Primary coma
3 3 8 r3 cos 3q
4 0 (
5 6r 4 - 6r2 + 1 ) Primary spherical
4 2 (
10 4r - 3r 4 2
) cos 2q Secondary astigmatism
4
4 4 10 r cos 4q
5 1 (
12 10r5 - 12r3 + 3r cos q ) Secondary coma
5 3 12 (5r 5
- 4r 3
) cos 3q
5 5 12 r5 cos 5q
6 0 (
7 20r6 - 30r 4 + 12r2 - 1 ) Secondary spherical
6 2 (
14 15r6 - 20r 4 + 6r2 cos 2q ) Tertiary astigmatism
6 4 14 (6r 6
- 5r 4 cos 4q)
6 6 14 r6 cos 6q
7 1 ( )
4 35r7 - 60r5 + 30r3 - 4r cos q Tertiary coma
7 3 4 (21r - 30r + 10r ) cos 3q
7 5 3
7 5 4 (7r - 6r ) cos 5q
7 5
7 7 4 r7 cos 7q
8 0 (
3 70r8 - 140r6 + 90r 4 - 20r2 + 1 ) Tertiary spherical
*The words “orthonormal Zernike” are to be associated with these names, e.g., orthonormal Zernike primary
astigmatism.
mean value of Rn0 (r) is zero. This indeed is the purpose of the constant term in Rn0 (r) ; it
makes the mean value zero. When m π 0 , the average value of cos mq is zero. Similarly,
the mean square value of the aberration is given by
1 2p 1 2p
2 Û Û 2 Û Û
[ F nm (r, q)] = Ù Ù F nm (r, q)
ı ı
[ ] r dr d q Ù Ù r dr d q
ı ı
0 0 0 0
2
= cnm . (2-65b)
Hence, its variance is given by
m 2 2
s 2nm = (F ) n - F nm
2
= cnm , n π 0, m π 0 . (2-66)
Thus, each orthonormal expansion coefficient, with the exception of c00 , represents the
standard deviation of the corresponding aberration term. The variance of the aberration
function is accordingly given by
2 • n
s 2F = F 2 (r, q) - F(r, q) 2
= Â Â cnm . (2-67)
n =1 m = 0
Unless the mean value of the aberration < F > = 0, s F π F rms , where F rms = < F 2 >1 / 2
is the root-mean-square (rms) value of the aberration. Substituting Eq. (2-67) into any of
Eqs. (2-53) yields the Strehl ratio for small aberrations.
A balanced aberration represents an aberration of a certain order in the power series

expansion of the aberration function in pupil coordinates mixed with aberrations of lower
order such that the variance of the net aberration is minimized. The balanced primary
aberrations can be identified easily with the corresponding Zernike polynomials. For
example, for n = 4 and m = 0 , Eq. (2-64) becomes
F 04 (r, q) = 5c40 R40 (r)
= (
5c40 6r 4 - 6r2 + 1 ) . (2-68)
Comparing this with the balanced spherical aberration given in Table 2-5,
F bs (r) = As r 4 - r2 ( ) , (2-69)
we note the following. The aberration F 04 contains a constant (independent of r and q)

term. This term does not change the standard deviation of the balanced aberration or the
Strehl ratio corresponding to it. In Eq. (2-68), as in Eq. (2-69), the spherical aberration is
balanced with an equal and opposite amount of defocus. Comparing the coefficients of
r 4 term in these equations, we find immediately that the standard deviation of the
balanced spherical aberration is given by
s bs = c40
= As 6 5 , (2-70)
in agreement with the result given in Table 2-5.
When n = 3 and m = 1, Eq. (2-64) becomes
F13 (r, q) = 2 2 c31 R31 (r) cos q
(
= 2 2 c31 3r3 - 2r cos q . ) (2-71a)
We note that this polynomial represents balanced coma
2
F bc (r, q) = Ac Ê r3 - rˆ cos q (2-71b)
Ë 3 ¯
for which the standard deviation is given by
s bc = c31 = Ac 6 2 . (2-71c)
For n = 2 and m = 2 , Eq. (2-64) becomes
F 22 (r, q) = 6 c22 R22 (r) cos 2q
(
= 2 6 c22 r2 cos 2 q - 1/2 ) . (2-72a)
This polynomial represents balanced astigmatism
(
F ba (r, q) = Aa r2 cos 2 q - 1 2 ) (2-72b)
s ba = c22 = Aa 2 6 . (2-72c)
F 20 (r, q) = 3c20 R20 (r)
= (
3c20 2r2 - 1 ) . (2-73a)
It represents defocus aberration (except for a constant term)
F d (r) = Bd r2 (2-73b)
with a standard deviation of
s d = c20 = Bd 2 3 . (2-73c)
For n = 1 and m = 1, Eq. (2-64) becomes
F11 (r, q) = 2c11 R11 (r) cos q
= 2c11 r cos q . (2-74a)
It represents a wavefront tilt
F t (r, q) = Bt r cos q (2-74b)
s t = c11
= Bt 2 . (2-74c)
Finally, for n = m = 0 , Eq. (2-64) becomes
F 00 (r, q) = c00 , (2-75)
which represents a uniform (piston) aberration. Obviously, its standard deviation is zero
and has no effect on the Strehl ratio of a system with a single exit pupil. (This will not be
true in the case of a multiexit pupil system, as in a phased array system.)
Thus, we see that Zernike polynomials can be identified with balanced aberrations;
that, in fact, is their advantage. Here we have discussed only the primary aberrations. In
general, the aberration function of an optical system may consist of higher-order
aberrations. Moreover, in a system without an axis of rotational symmetry, the aberration
function will consist of terms not only in cos mq but in sin mq as well. This is true in the
case of aberrations introduced by atmospheric turbulence (see Chapter 5).
2.5 DEFOCUSED SYSTEM10-12

Defocus is the simplest and most common aberration. It is introduced when an
imaging system forms the image of an object in some plane but it is observed in another.
Field curvature of a system represents defocus aberration that varies quadratically with
the field angle. Defocus is also introduced when a beam is focused at a certain distance
but observed at some other.
We have already seen that a quarter wave of defocus aberration yields a Strehl ratio
of 0.8 (see Table 2-4). How this defocus aberration translates into longitudinal defocus, or
depth of focus, depends on the Fresnel number of the focused image or beam. In this
section, we show that the depth of focus for a system with a large Fresnel number, such
as a photographic camera, is very small. However, it is quite large for a system with a
small Fresnel number, e.g., a laser transmitter focusing a beam on a distant target. We
also show that for such a system, maximum central irradiance is not obtained at the
geometrical focus but at a defocused point that is closer to the system. However, this does
not mean that a beam should be defocused to obtain maximum central irradiance on a
target at a given distance. Maximum central irradiance on a target is still obtained when a
2.5 Defocused System 111
beam is focused on it, even though a larger irradiance is obtained at a defocused point
that is closer to the focusing system. A far-field distance of the system is defined such
that a beam focused beyond it behaves practically like a collimated beam.

Now we examine how Eq. (2-1), which gives the irradiance distribution of the image
of a point object in the Gaussian image plane at a distance R from the plane of the exit
pupil, is modified if the irradiance distribution is observed in a plane at a distance z from
the pupil plane. [For convenience, we drop the subscript i on z used in Eq. (1-51).] For an
aberration-free image in this observation plane, the wavefront at the pupil must be
spherical with its center of curvature lying in the plane. Such a spherical wavefront now
forms the reference sphere with a radius of curvature z . Accordingly Eq. (2-1) is
modified in two ways. First, the quantity R in the exponent is replaced by z . Second, an
aberration corresponding to the optical path difference between two spheres, one of
radius of curvature R (with respect to which the aberration function is defined) and the
other of radius of curvature z (with respect to which the aberration must now be
considered) is introduced. This additional aberration is approximately equal to the
difference in sags of two spheres of radii of curvature z and R. It is given by
1 Ê1 1 1 1 1
( )
Wd rp = - ˆ rp2 + Ê 3 - 3 ˆ rp4 + º , (2-76)
2Ëz R¯ 8Ëz R ¯
where the first term on the right-hand side represents a defocus aberration and the second
term represents a fourth-order spherical aberration. For a circular exit pupil of radius a, if
z is large enough that
Ê 1 - 1 ˆ a4 £ l ;
Ë z 3 R3 ¯
i.e., if
1 1 l
£ 3 + 4 , (2-77)
z3 R a
then the term in rp4 and those of higher order in Eq. (2-76) may be neglected. A spherical
aberration of l 8 gives a Strehl ratio of 0.946. When this aberration is balanced with an
equal and opposite amount of defocus, the Strehl ratio increases to 0.996. Assuming the
validity of Eq. (2-77), Eq. (2-76) may be written
1 Ê1 1
( )
Wd rp = - ˆ rp2 . (2-78)
2Ëz R¯
Moreover, the quantity l0 in the Huygens’ spherical wavelet of Eq. (1-18) or l in Eq.
(1-20) is now approximately equal to z . Hence, the irradiance distribution in a defocused
image plane at a distance z from the pupil plane may be written
r r 2 pi r r ˆ r
2
Ii (ri ) = (P
ex Sex l z 2 2
) Û
{[ ( )
Ù exp i F rp + F d rp
ı
( )] } exp Ê - ◊r r dr
Ë lz p i ¯ p
,(2-79)
r
where Ii (ri ) represents the irradiance at a point whose position vector in the observation
r
plane is ri with respect to the center of curvature of the new reference sphere, and
( )
F d rp = (2p l )Wd rp ( ) (2-80)
(r )
is the defocus phase aberration. It should be clear that F rp is the phase aberration at a
r
point rp in the plane of the exit pupil with respect to a reference sphere centered at the
r r
Gaussian image point. Dividing Ii (ri ) by Pex , we obtain PSF(ri ) . Equation (2-79) may
r
also be obtained from Eq. (1-51) by regarding ri as stated above and multiplying both
sides by Pex .
Following the procedure used to obtain Eq. (2-11), we find that Eq. (2-79) for the
defocused irradiance distribution may be written for an otherwise aberration-free
r
[( ) ]
F rp = 0 system
1 2
Û
I ( r; z ) = (2 R z )
2
(2
)
Ù exp i Bd r J 0 ( p rr R z ) r dr
ı
, (2-81a)
0
where
R
Bd ( z ) = p N Ê - 1ˆ (2-81b)
Ëz ¯
is the peak value of the defocus phase aberration and
N = a2 l R (2-81c)
is the Fresnel number of the exit pupil as observed from the focal point. As in Eq. (2-11),
r
r = a -1 rp , r is in units of l F , and the irradiance is in units of the aberration-free central
irradiance Pex Sex l2 R 2 . In Eq. (2-81a), we have already carried out the angle integration
following Eq. (2-12).
From Eq. (2-81a), we note that the irradiance distribution is asymmetric about the
Gaussian image plane; i.e., the irradiance distributions in observation planes located at
z = R ± D , where D is a longitudinal defocus, are not identical. There are three reasons
for this asymmetry. First, the inverse square law dependence on z increases I (r ) for
z < R and decreases it for z > R . Second, Bd is asymmetric since
Bd ( R + D ) π - Bd ( R - D ) . (2-82)
◊
Third, the argument of the Bessel function J 0 ( ) depends on z .
However, if N is very large (>> 10), Bd becomes large even for very small
differences in z and R. In that case, the defocus tolerance dictates that z be
approximately equal to R. Hence, Eq. (2-81b) may be written
( )
Bd = p 4l F 2 ( R - z ) , (2-83a)
which, in turn, yields
Bd ( R + D ) = - Bd ( R - D ) . (2-83b)
Now, according to Eq. (2-81a), the irradiance distribution is independent of the sign of
Bd . Hence, for z ~ R , the distribution is symmetric about the Gaussian image plane.
For small N (£ 10), z can be much different from R for Bd to achieve a significant
value. In this case, therefore, all three factors contribute to the asymmetry of the
irradiance distribution about the Gaussian image plane. One consequence of this is that
the irradiance on and near the axis can be higher for z < R than that for z > R .
2.5.2 Focused Beam

Consider an optical system focusing a beam of power Pex distributed uniformly
across its exit pupil of radius a. If the beam is focused at a distance R, its irradiance
distribution in a plane at a distance z from the exit pupil will be given by Eq. (2-81a). If
we let r = 0 in this equation, we obtain the axial irradiance of the beam
2
I (0; z ) = ( R z ) S , (2-84a)
or
( )
Ii (0; z ) = Pex Sex l2 z 2 S , (2-84b)
where
2
[
S = sin ( Bd 2) ( Bd 2) ] (2-84c)
is a Strehl ratio defined as the ratio of the axial irradiance at a distance z when the beam is
focused at a distance R to that when it is focused at a distance z . The axial irradiance is
minimum and equal to zero when Bd is an integral multiple of 2p; i.e., when the peak
defocus wave aberration is an integral multiple of l. The corresponding z values are
given by
R z = 1 + 2n N , (2-85)
where n is a (nonzero) positive or negative integer according to whether z < R or z > R .

These z values correspond to n waves of defocus as an aberration at the outer edge of the
pupil. Note that when N £ 2 , the axial irradiance is nonzero for z > R , except, of course,
in the limit z Æ • . When N is very large, as in a photographic camera, the axial
irradiance is zero when z is different from R by an integral multiple of ± 8l F 2 . Thus, the

depth of focus corresponding to a l 4 defocus wave aberration in such cases is only
± 2l F 2 .
The maxima of axial irradiance can be determined by equating to zero the derivative
of Ii (0; z ) of Eq. (2-84b) with respect to z . We find that they occur at z values given by
the solutions of the transcendental equation
tan ( Bd 2) = ( R z ) Bd 2 , z π R . (2-86)
Although z = R is a solution of this equation, it must be excluded from consideration,

because the derivative of Ii (0; z ) with respect to z at z = R is nonzero and equal to
- 2 Pex Sex l2 R 2 . Equation (2-86) can be solved graphically by determining the
intersection of the curves tan ( Bd 2) and ( R z ) Bd 2 vs z . A maximum of axial
irradiance lies between two of its adjacent minima. The principal maximum lies between
the geometrical focus and the first minimum adjacent to it for which z < R ; i.e., it lies at
z p , where R (1 + 2 N ) < z p < R .
We note from Eqs. (2-84a) and (2-84c) that, since S £ 1, I (0; z ) £ 1 for z ≥ R .
However, for z < R , an increase in axial irradiance due to the inverse square law
competes with a decrease due to the defocus aberration (nonconstructive interference of
Huygens’ spherical wavelets). Which effect dominates depends on the Fresnel number N.
When N is very large, a slight amount of longitudinal defocus produces a large amount of
defocus aberration Bd and correspondingly a small Strehl ratio S . Hence, for z < R , an
increase in axial irradiance due to the inverse square law is overcome by the loss due to
the destructive interference of Huygens’ spherical wavelets. Accordingly, in such cases,
I (0; z ) < 1 even when z < R . When N is very small, z can be much different from R
2
before Bd becomes significant. Hence, over a considerable range of z < R , ( R z ) > S -1
so that I (0; z ) > 1 . For example, S = 0.8 is obtained at R z = 1.52 when N = 1, giving
I (0; z ) = 1.84 at this z value. When N = 10 , S = 0.8 is obtained at R z = 1.05 so that
I (0; z ) = 0.88 . Similarly, when N = 100 , R z = 1.005 for S = 0.8 , and, accordingly,
I (0; z ) is practically the same as S.
The range of distance z over which I (0; z ) > I (0; R) depends on the Fresnel number.
Let z0 be the distance at which I (0; z0 ) = I (0; R) , so that for z0 < z < R ,
I (0; z ) > I (0; R) . It is evident that z0 lies between the focus and the first minimum to the
left of it; i.e., 1 < R z0 < 1 + 2 N . As N increases, z0 Æ R ; i.e., the range of the distance
over which I (0; z ) > I (0; R) reduces to zero. For N < 0.25 , i.e., for R > D2 l , a focused
beam propagates very much like a collimated beam, as discussed later in Section 2.5.3.
Figure 2-10 shows how z0 R varies with N . For N ≥ 10 , 0.975 £ z0 R £ 1. It also

shows how the distance z p of the principal maximum of axial irradiance and the
irradiance at this point vary with N . Just as the principal maximum occurs significantly
far from the geometrical focus when N < ~ 3 , the axial irradiance is greater than or equal
to the focal-point irradiance over a significant range of z values when N < ~ 3 . We note
1.0 10
zp/R
0.8 8
z0/R
0.6 6
z0 /R, zp /R
I(zp)/I(R)
0.4 4
0.2 2
I(zp)/I(R)
0.0 0
0 2 4 6 8 10
N
( )
Figure 2-10. Variation of z 0 , z p , and I z p with N. z 0 is the minimum distance of an
axial point from the aperture so that I ( z 0 ) = I ( R) = p 2 N 2 I0 , where I0 is the pupil
irradiance. z p is the distance of the location of the principal maximum. I z p is the ( )
irradiance at this point.
that z0 R <
~ 0.92 when N <
~ 5. When N <
~ 3 , z0 R < ~ 0.8 . As N Æ 0 , corresponding
( )
to a collimated beam, R Æ • . Therefore, I 0; z p I (0; R) Æ • , z0 R Æ 0 , and
zp R Æ 0 .
Figure 2-11 shows how the axial irradiance varies for beams with N = 1, 10, and
100. We note that it is highly asymmetric about the focal point when N = 1, but it
becomes more and more symmetric as N increases. Figure 2-12, which shows the axial
irradiance for N = 1 labeled as S ¢ , also shows how the peak defocus aberration
Wm = (l 2 p) Bd ( z ) varies with z. It is evident that Wm , shown in units of l , is not
symmetric about the focal point. In particular, we note that when N = 1, the principal
maximum of axial irradiance lies at z = 0.6 R and not at the focal point z = R . The value
of Wm for this point is only l 3 and reduces the irradiance to 68 percent of the focal-
point value. However, the inverse square law increases the irradiance by a factor of
approximately 2.78. Hence, the net irradiance at this point is 1.9 times the focal-point
irradiance Pex Sex l2 R 2 .
The dotted curve in Figure 2-12 shows how the axial irradiance at a distance z varies
with z if the beam is focused at this distance. Considering that the irradiance in this figure
2
is in units of Pex Sex l2 R 2 , this curve represents the variation of ( R z ) as a function of
z R. The ratio of the solid and dotted curves gives the Strehl ratio S shown in this figure
as a function of z R. It should be evident that S = 1 when z = R .
2.0 2.0 2.0

N=1 N = 10 N = 100
1.5 1.5 1.5
lu
l (0; z)
1.0 1.0 1.0
lg
0.5 0.5 0.5
0.0 0.0 0.0

0.0 0.5 1.0 1.5 0.5 1.0 1.5 0.8 0.9 1.0 1.1 1.2
z/R z /R z /R
Figure 2-11. Axial irradiance of a uniform circular beam focused at a distance R

with Fresnel number N = 1 , 10, and 100. The irradiance is normalized by the focal-
point irradiance Pex Sex l 2 R 2 . The dashed curves are for a Gaussian beam,
discussed in Chpater 4.
2.0 1.00
S¢
S
1.5 0.75
S¢,ÔWmÔ, (R/z)2
1.0 0.50
(R/z)2
0.5 0.25
ÔWmÔ
0.0 0.00
0.0 0.5 1.0 1.5 2.0
z/R
Figure 2-12. The axial irradiance S ¢ of a circular beam focused at a distance R in

units of its focal-point irradiance. The minima of irradiance occur at z R = 1 3,
1 5 , 1 7 , º etc. The varation of wave aberration Wm and Strehl ratio S with z,
and the inverse square law dependence on z are also shown.
Now, we ask the question: Given a target at a distance z , how should a beam be
focused on it so that the central irradiance on it is maximum? Thus, we would like to
know the optimum value of R. The answer is that we should choose R = z , as may be
seen by differentiating Eq. (2-84b) with respect to R and equating the result to zero. It is
evident that the central irradiance on a target when the beam is focused on it is
Pex Sex l2 z 2 , which is 1.47 times the corresponding irradiance when the beam is focused
at a distance R = 0.6 z .
Figure 2-13 illustrates how the central irradiance on a target at a fixed distance z
varies when the beam is focused at various distances R along its axis. The irradiance in
this figure is in units of Pex Sex l2 z 2 . The quantity
Nz = a2 l z (2-87)
represents the Fresnel number of the circular exit pupil as observed from the target. We
note that as N z increases, the curves become symmetric about the point R = z . The
encircled power in a defocused image plane is given by
rc
Û
( ı
2
)
P(rc ; z ) = p 2 Ù I (r; z ) rdr . (2-88)
0
Numerical calculations of integrals in Eqs. (2-81) and (2-88), are discussed in Section
2.6.
Figure 2-14a shows for N = 1 the irradiance distribution and encircled power in the
focal plane ( z = R) and in the planes z = (1 ± 0.4) R . We note that the power is
concentrated most in the plane z = 0.6 R . However, for 0.65 < rc < 1.10, the power in
thisplane is smaller than the corresponding value in the focal plane. The irradiance and
1.0 1.0 1.0
0.8 lu 0.8 0.8

lg
lz (0; R)
0.6 0.6 0.6

Nz = 1 Nz = 10 Nz = 100
0.4 0.4 0.4
0.2 0.2 0.2
0.0 0.0 0.0

0.5 1.0 1.5 2.0 0.5 1.0 1.5 2.0 0.8 0.9 1.0 1.1 1.2
R/z R/z R/z
Figure 2-13. Central irradiance (in units of Pex Sex l 2 z 2 ) at a distance z from the
plane of the exit pupil when a beam is focused at various distances R. The quantity
N z = a 2 l z represents the Fresnel number of the exit pupil as observed from the
target. The dashed curves are for a Gaussian beam.
2.0 1.00
1.5 0.75
P(rc)
l(r)
1.0 0.50
z=R
z = 0.6 R
z = 1.4 R
0.5 0.25
0.0 0.00
0.0 0.5 1.0 1.5 2.0
r, rc
(a)
1.00
rc = 1.5
0.75 1.0
0.8
P(z;rc)
0.50
0.6
0.25
0.4
0.2
0.00
0.4 0.6 0.8 1.0 1.2 1.4
z /R
(b)
Figure 2-14. (a) Focused and defocused irradiance distributions and corresponding
encircled powers for a circular focused beam with N = 1 . The irradiance is
normalized by the focal-point irradiance and encircled power is normalized by the
total power Pex . The units of r and rc are l F. (b) Encircled power in a circle of fixed
radius rc for a circular focused beam as a function of the axial distance z from the
exit pupil.
power distributions when the beam is focused at a distance z = 0.6 R is the same as for
z = R , except that the horizontal scale is changed by a factor of 0.6 so that r and rc are in
units of 0.6 l F . Thus, 84 percent of the power is contained in a radius of 0.6 l F instead
of l F , i.e., the power is much more concentrated.
Figure 2-14b shows how the power in a spot of certain radius rc (in units of lF )
varies with the axial distance z in the vicinity of the geometrical focus. Although the axial
irradiance has a principal maximum at z = 0.6 R , where its value is nearly twice the
focal-point irradiance, the encircled power is maximum at this z value only for small spot
sizes. For moderate spot sizes (0.3 £ rc £ 1), the maximum of encircled power occurs at a
z value that varies with rc but lies between the point of principal maximum and the focal
point. It is evident from this example that where the diffraction focus of a converging
spherical wave lies depends on the criterion used to define it. Is it the maximum of
irradiance or maximum of encircled power? Maximum encircled power on a target at a
given distance is obtained when the beam is focused on it. Thus, for a moving target, the
beam must be actively focused on it to obtain the maximum possible encircled power.
However, if the encircled power is adequate when the beam is focused on a target at a
certain distance so that the Fresnel number of the beam aperture as observed from the
focus is small < ( )
~ 5 , it is more than adequate over a considerable range of the target
distance without any active focusing of the beam.
2.5.3 Collimated Beam

We have seen that the principal maximum of the axial irradiance of a focused beam
with N = 1 has a value that is 1.9 times the focal-point irradiance. This ratio increases as
N decreases and approaches infinity as N Æ 0 , corresponding to diffraction of a
collimated beam or a plane wave.
The results for a collimated beam can be obtained from those for a focused beam by
letting R Æ • . Thus, for example, the peak defocus aberration given by Eq. (2-82)
becomes
Bd = p a 2 l z
= Sex l z . (2-89)
It represents the peak phase aberration of a plane wavefront with respect to a reference
sphere centered at a distance z from the exit pupil and passing through its center.
Equation (2-84b) for axial irradiance reduces to
(
I (0; z ) = 4 I0 sin 2 p a 2 2 l z ) , (2-90)
where
I0 = Pex Sex (2-91)
is the irradiance at the exit pupil. The axial irradiance is maximum and equal to 4 I0 at z
values given by
z = a 2 l (2 n + 1) , n = 0 , 1 , 2 , K . (2-92)
It is minimum and equal to zero at z values given by
z = a 2 2 l (n + 1) , n = 0 , 1 , 2 , K . (2-93)
These z values for the location of maxima and minima correspond to those axial positions
at which the circular exit pupil subtends an odd or an even number of Fresnel’s half-
wave zones, respectively.
For z > a 2 l , the axial irradiance decreases monotonically to zero. For
z ≥ D2 l , (2-94)
it decreases approximately as z - 2 . For z satisfying Eq. (2-94), a collimated beam gives

an axial irradiance at a distance z that is ≥ 0.95 times the irradiance at this point if the
beam were focused at it, i.e., S ≥ 0.95 . This is illustrated in Figure 2-15, where the axial
irradiance of a collimated beam, focal plane irradiance of a beam focused at a distance z
2
[given by ( p 4 z ) from Eq. (2-84b)], and their ratio S are plotted as a function of z. The
distance z in this figure is in units of D2 l , which is called the far-field distance of the
exit pupil. It is evident that a collimated beam yields practically the same irradiance on a
target lying in the far field of the exit pupil as a beam focused on it; in other words, beam
focusing does not significantly increase the power concentration on the target.
4.0 1.00
3.0 0.75
I(0;z), (p/4z)2
I(z)
2.0 0.50
(p/4z)2
1.0 0.25
0.0 0.00
0.00 0.25 0.50 0.75 1.00
z
Figure 2-15. Axial irradiance of a collimated circular beam normalized by the exit
pupil irradiance I0 . The axial distance z is in units of the far-field distance D 2 l .
2.6 PSFs for Rotationally Symmetric Aberrations 121
The irradiance distribution in a plane at a distance z can be obtained from Eq. (2-81)
by letting R Æ • and noting that the units of irradiance in this equation are Pex Sex l2 R 2
and those of r are l F = l R D . Thus, for a collimated beam, we may write
1 2
Û
ı
( )
I (r; z ) = 4 Ù exp i Bd r2 J 0 ( p rr) r dr , (2-95)
0
where the units of irradiance are Pex Sex l2 z 2 and those of r are lz D. The irradiance
distribution given by Eq. (2-95) is called the Fresnel diffraction pattern of a circular exit
pupil. For z ≥ D2 l , since the aberration is negligibly small ( Bd £ p 4 or l 8), Eq. (2-
95) reduces to the aberration-free result
I ( r; z ) ~ [2 J1 (p r ) p r]2 , z ≥ D2 l . (2-96)
Equation (2-94) is the far-field condition for a circular exit pupil and Eq. (2-96)
represents its far-field or Fraunhofer diffraction pattern. Except for the units of irradiance
and r, Eq. (2-96) is the same as Eq. (2-15).
( )
The irradiance distribution for small z values z < D2 l , called the near-field
diffraction pattern, is equivalent to a defocused Fraunhofer distribution. Therefore, using
our approximate model discussed in Section 2.5, we can quickly estimate the encircled
power for small z values. Thus, for example, for z > a 2 l , corresponding to S > 0.4 , the
encircled power for rc < 0.8 (in units of lz / D ) is within 8 percent of S times the
encircled power for a beam focused at a distance z , i.e., it is given by
P(rc ) ~ S [1 - J02 (p rc ) - J12 (p rc )] . (2-97)
For z > 2
~ D 6l , corresponding to S > ~ 0.1 , Eq. (2-97) gives the encircled power for
rc £ 0.5 with an error of < 5 percent. The axial irradiance and, therefore, S are both zero
when z = D2 8l . Thus, over a considerable range of z values, the encircled power can be
quickly estimated. Note that z = a 2 l is the maximum value of z for which the axial
irradiance is a maximum (with a value of 4 I0 ).
2.6 PSFs FOR ROTATIONALLY SYMMETRIC ABERRATIONS13, 14
Now, we consider aberrated PSFs for rotationally symmetric aberrations. Numerical

results are obtained for the fourth-, sixth-, and eighth-order balanced spherical aberrations
represented by the corresponding Zernike polynomials.13 It is shown that the normalized
PSFs are practically identical with the aberration-free PSF within the Airy disc.
Defocused PSFs are also considered in a similar manner. A simple Gaussian model is
developed that can be used to calculate the aberrated PSFs and encircled power within the
Airy disc. The only parameter needed is the standard deviation of the aberration, which is
used to estimate the Strehl ratio. A modification to the simple result is also given when
the aberration is not represented by a Zernike polynomial.14
2.6.1 Theory
Consider an optical imaging system with an exit pupil of radius a forming a Gaussian
image in a plane at a distance R from the plane of its exit pupil. The irradiance
distribution in a plane at a distance z is given by Eq. (2-81). The effect of defocus as an
aberration can be separated from the scaling effects due to z being different from R, by
comparing the irradiance distributions (or PSFs) observed at a distance z when the image
is formed at a distance R and when it is formed at a distance z . Thus, if we measure r in
units of lz D (instead of l R D ) and I in units of Pex Sex l2 z 2 , Eqs. (2-81) become
2
1
Û
ı
[ ]
I (r ) = 4 Ù exp iF 2 (r) J 0 ( p r r) r dr , (2-98)
0
where
F 2 (r) = Bd r2 (2-99a)
Sex Ê 1 1
= - ˆ r2 (2-99b)
l Ëz R¯
represents the defocus phase aberration. Equation (2-98) does not change if we replace
F 2 (r) by the orthonormal Zernike circle polynomial representing defocus, i.e., if we let
F 2 (r) = 3 R20 (r) s F (2-100a)
= ( )
3 2r 2 - 1 s F , (2-100b)
where
Sex 1 1
sF = - (2-101)
2 3l z R
is the standard deviation of the aberration over the exit pupil.
In a similar manner, we can show that for a system aberrated by a rotationally (or
radially) symmetric phase aberration F(r) , Eq. (1-82) giving the irradiance distribution
of the image of a point object, formed in a plane at a distance R from the plane of its exit
pupil, reduces to
1 2
Û
ı
[ ]
I (r ) = 4 Ù exp iF(r) J 0 ( pr r) r dr , (2-102)
0
where r is now in units of l F = l R D and the irradiance is in units of Pex Sex l2 R 2 .

Equations (2-98) and (2-102) are similar to each other except that the units of r and
irradiance are different. Keeping this in mind, we will use Eq. (2-102) in what follows.
According to this equation, the aberration-free central value of irradiance I (0) is unity.
For numerical analysis we now write the right-hand side of this equation as a product of
two complex conjugate integrals
1 1
Û Û
ı
[ ı
]
I (r ) = 4 Ù exp iF(r) J 0 ( p r r) r dr Ù exp [ - iF (s)] J 0 ( p rs) sds . (2-103)
0 0
Combining the two exponentials and noting that, since I (r ) is a real quantity, the
imaginary part on the right-hand side of Eq. (2-103) must be zero, we may write it in the
form
1 1
Û Û
[ ]
I (r ) = 4 Ù Ù cos F(r) - F (s) J 0 ( p r r) J 0 ( p rs) rs dr ds .
ı ı
(2-104)
0 0
The Strehl ratio of the image is the central value I (0) ; i.e.,
1 1
Û Û
S = 4 Ù Ù cos F(r) - F(s) rs dr ds .
ı ı
[ ] (2-105)
0 0
The encircled power, i.e., fraction of total power in a circle of radius rc (in units of
lz D for the defocused image and l R D for the aberrated images) is given by [see Eq.
(2-22b)]
rc
Û
P(rc ) (
= p 2 2 Ù I (r ) rdr .
ı
) (2-106)
0
Or, substituting Eq. (2-104) into Eq. (2-106),

1 1
Û Û
P(rc ) [
= 2 p Ù Ù cos F(r) - F( s) Q (r, s; rc ) rs d r ds ,
ı ı
2
] (2-107)
0 0
where
rc
Û
Q(r, s; rc ) = Ù J 0 ( p rr) J 0 ( p rs) rdr
ı
0
( )[
= rc2 2 J 02 ( p rrc ) + J12 ( p rrc ) ] if r = s , (2-108a)
[ (
= rc p r2 - s 2 )] [r J (p rr ) J (p sr ) - sJ (p sr ) J (p sr )] if r π s .(2-108b)
1 c 0 c 1 c 0 c
For small values of rc , J 0 ( p rrc ) ~ 1 and J1 ( p rrc ) ~ 0 , so that Q(r, s; rc ) ~ rc2 2.

Therefore, utilizing Eq. (2-105), Eq. (2-107) yields
P(rc ) ~ (p rc 2) 2 S , (2-109)
which simply amounts to assuming that the average irradiance in the image plane across a
circle of small radius rc is approximately uniform and equal to its central value S.
2.6.2 Numerical Results

We now apply the above equations to obtain some numerical results on the Strehl
ratio, PSF, and encircled power for imaging systems aberrated with rotationally
symmetric aberrations. In particular, we consider defocus, fourth-, sixth-, and eighth-
order spherical aberrations given by orthonormal Zernike polynomials n + 1 Rn0 (r) , i.e.,
aberrations
F 2 (r) = ( )
3 2r 2 - 1 s F , (2-110a)
F 4 (r) = ( )
5 6r 4 - 6r2 + 1 s F , (2-110b)
F 6 (r) = (
7 20r6 - 30r 4 + 12r2 - 1 s F ) , (2-110c)
and
(
F 8 (r) = 3 70r8 - 140r6 + 90r 4 - 20r2 + 1 s F ) . (2-110d)
Each polynomial represents a classical or a Seidel aberration of a certain order optimally

balanced with one or more classical aberrations of lower order in the sense of minimum
variance over a unit circle. Note that
< Fn > = 0 (2-111a)
and, therefore,
< F 2n > ∫ s 2F . (2-111b)
Thus, the coefficient s F of each aberration represents its standard deviation as well as its
root mean square value. How these aberrations vary with r is shown in Figure 2-16 for
s F = 1. Since 0 £ r £ 1, it is evident that n 2 is the number of roots of the aberration
F n (r) ; i.e., F n has a value of zero at n 2 different values of r.
The integrals in Eqs. (2-104), (2-105), and (2-107) for the aberrations given above
may be evaluated by the Gauss quadrature formula,15 according to which, for a function
f (r, s) ,
3.0
F8
2.0
F2
1.0
F(r)
0.0
–1.0
F4
–2.0
F6
–3.0
0.0 0.2 0.4 0.6 0.8 1.0
r
Figure 2-16. Variation of an aberration Fn with r . s F = 1 for each aberration in

this figure.
1 1
M
Û Û M
Ê 1 + xi 1 + x j ˆ
Ù Ù f (r, s) dr ds = (1 4) Â Â w iwj f Á
Ë 2
, ˜ , (2-112)
ı ı i =1 j =1 2 ¯
0 0
where M is the number of points of a 1-D quadrature, xi is the i-th zero of the M-th order
Legendre polynomial PM ( x ) , and w i are the weight factors given by
-2
2 dPM ( xi )
wi = . (2-113)
1 - xi2 dx
Since the integrands in these equations are symmetric in r and s, Eq. (2-112) reduces for
our application to
1 1
M
Û Û 2 Ê 1 + xi 1 + xi ˆ
Ù Ù f (r, s) dr ds = (1 4) Â w i f Ë 2 , 2 ¯
ı ı i =1
0 0
M i -1
Ê 1 + xi 1 + x j ˆ
+ (1 2) Â Â w iw i f Á , ˜ . (2-114)
i =1 j =1 Ë 2 2 ¯
Thus, because of the symmetry of the integrand, we have to consider only M ( M + 1) 2

instead of M 2 terms. The values of xi and w i for Gauss quadrature of different points
are given in Table 2-8.
Table 2-8. Zeros xi of Legendre polynomial Pn ( x ) and weight factors w i for an n-

point Gauss quadrature [from M. Abramowitz and I. Stegun, Eds., Handbook of
Mathematical Functions with Formulas, Graphs, and Mathematical Tables, (Dover,
New York, 1970), p. 916.]
n
+1
Ú–1 f ( x )dx = Â w i f ( xi )
i =1
xi are the zeros of Legendre polynomials and wi are the weight factors
±xi wi ±xi wi
n =2 n =8
0.18343 6424 5650 0.36268 7833 8362
0.57735 2691 9626 0.00000 0000 0000 0.52553 4099 6329 0.31370 6458 7887
0.79666 4774 3627 0.22238 0344 3374
n =3 0.96028 8564 7536 0.10122 5362 0376
0.00000 0000 0000 0.88888 8888 8889
0.77459 6692 1483 0.55555 5555 5556
n= 4 0.00000 0000 0000 0.33023 3550 1260
0.33998 0435 4856 0.65214 1548 2546 0.32425 4234 3809 0.31234 0770 0003
0.86113 3115 4053 0.34785 8451 7654 0.61337 4327 0590 0.26061 6964 2935
0.83603 1073 6636 0.18064 1606 4857
n =5 0.96816 2395 7626 0.08127 3883 1574
0.00000 0000 0000 0.56888 8888 8889 n = 10
0.53846 3101 5683 0.47862 6704 9366 0.14887 3389 1631 0.29552 2247 4753
0.90617 8459 8664 0.23692 8850 6189 0.43339 3941 9247 0.29626 7193 9996
n =6 0.67940 5682 9024 0.21908 3625 5982
0.23861 1860 3197 0.46791 9345 2691 0.86506 3666 8985 0.14945 3491 0581
0.66120 3864 6265 0.36076 5730 8139 0.97390 5285 7172 0.06667 3443 8688
0.93246 5142 3152 0.17132 4923 9170 n = 12
0.12523 4085 1469 0.24914 0458 3403
n =7 0.36783 4989 8180 0.23349 5365 8355
0.00000 0000 0000 0.41795 1836 3469 0.58731 9542 6617 0.20316 4267 3066
0.40584 1513 7397 0.38183 0505 5119 0.76990 6741 4305 0.16007 3285 3346
0.74153 1855 9394 0.27970 3914 9277 0.90411 2563 0475 0.10693 3259 5318
0.94910 9123 2759 0.12948 9661 8870 0.98156 6342 6719 0.04717 3363 6512
±xi wi
n = 16
0.09501 25098 37637 440185 0.18945 06104 55068 496285
0.28160 35507 79258 913230 0.18260 34150 44923 588867
0.45801 67776 57227 386342 0.16915 65193 95002 538189
0.61787 62444 02643 748447 0.14959 59888 16576 732081
0.75540 44083 55003 033895 0.12462 89712 55533 872052
0.86563 12023 87831 743880 0.09515 85116 82492 784810
0.94457 50230 73232 576078 0.06225 35239 38647 892863
0.98940 09349 91649 932596 0.02715 24594 11754 094852
n = 20
0.07652 65211 33497 333755 0.15275 33871 30725 850698
0.22778 58511 41645 078080 0.14917 29864 72603 746788
0.37370 60887 15419 560673 0.14209 61093 18382 051329
0.51086 70019 50827 098004 0.13168 86384 49176 626898
0.63605 36807 36515 025453 0.11819 45319 61518 417312
0.74633 19064 60150 792614 0.01093 01198 17240 435037
0.83911 69718 22218 823395 0.08327 67415 76704 748725

0.91223 44282 51325 905868 0.06267 20483 34109 063570
0.96397 19272 77913 791268 0.04060 14298 00386 941331
0.99312 85991 85094 924786 0.01761 40071 39152 118312
n = 24
0.06405 68928 62605 626085 0.12793 81953 46752 156974
0.19111 88674 73616 309159 0.12583 74563 46828 296121
0.31504 26796 96163 374387 0.12167 04729 27803 391204
0.43379 35076 26045 138487 0.11550 56680 53725 601353
0.54542 14713 88839 535658 0.10744 42701 15965 634783
0.64809 36519 36975 569252 0.09761 86521 04113 888270
0.74012 41915 78554 364244 0.08619 01615 31953 275917
0.82000 19859 73902 921954 0.07334 64814 11080 305734
0.88641 55270 04401 034213 0.05929 85849 15436 780746
0.93827 45520 02732 758524 0.04427 74388 17419 806169
0.97472 85559 71309 498198 0.02853 13886 28933 663181
0.99518 72199 97021 360180 0.01234 12297 99987 199547
By substituting Eqs. (2-110) into Eq. (2-105), we can calculate the Strehl ratio for the
aberrations under consideration. In the case of defocus and fourth-order spherical
aberration, the Strehl ratio can be obtained analytically from Eq. (2-102) by letting r = 0
and carrying out the integration. The result obtained is
2
S2 = [(sin 3 sF ) 3 sF ] (2-115)
and
(
S4 = C 2 b + S 2 b b , ) (2-116)
where
b = 3 5s F p , (2-117a)
◊ ◊
and C( ) and S( ) are the Fresnel integrals given by
b
Û
C(b) = Ù cos p x 2 2 dx
ı
( ) (2-117b)
0
and
b
Û
(
S(b) = Ù sin p x 2 2 dx .
ı
) (2-117c)
0
Note that S2 gives the ratio of central irradiances at a distance z when the focused image
lies at a distance R and when it lies at a distance z. The Strehl ratio S4 for fourth-order
spherical aberration was considered in Section 2.3.4 also; see Table 2-6.
Figure 2-17 shows how the Strehl ratio varies with the standard deviation s w (in
units of the optical wavelength l ) of the various aberrations under consideration. The
wave and phase aberrations are related to each according to s w = (l 2 p) s F . S2 , S4 , S6 ,
and S8 , represent the actual Strehl ratios corresponding to aberrations F 2 , F 4 , F 6 , and
F 8 , respectively. For comparison, the Strehl ratios obtained by using the approximate
2
( ) ( )
expressions 1 - s 2F 2 – Maréchal formula, exp - s 2F – Gaussian approximation Sg ,
( )
are also plotted in this figure. It is evident that exp - s 2F approximates the Strehl ratio
very well and better than the other two expressions. We note that S2 deviates the most
from Sg . However, the difference between the two is less than 10 percent for S2 > 0.38.
The values of s F , accurate to the fourth decimal place, for Strehl ratios of 0.1 (0.1) 1 are
given in Table 2-9. We note that, except for very small Strehl ratios, their difference for
various aberrations for a given Strehl ratio is negligible. However, we used these exact
values to calculate and assess the aberrated PSFs and encircled powers.
1.0
0.8
0.6
S
0.4
Sg
0.2 1– sF2
Sm
0.0
0 0.05 0.10 0.15 0.20 0.25
sw
Figure 2-17. Strehl ratio as a function of wave aberration standard deviation

s w . S2 , S4 , S6 , and S8 represent Strehl ratios corresponding to aberrations
2
(
F2 , F4 , F6 , and F8 , respectively. S2, ––; S 4 …; S6,– –; S 8 , –. –. ; Sg = exp - s F
2
)
and
(
Sm = 1 - s F 2
) 2 represent the Gaussian and Maréchal approximations to the
Strehl ratio, respectively.
Table 2-9. Standard deviation of rotationally symmetric aberrations for various

Strehl ratios.
s F (Radians)
S F2 F4 F6 F8
1.0 0 0 0 0
0.9 0.3228 0.3233 0.3236 0.3238
0.8 0.4671 0.4687 0.4695 0.4702
0.7 0.5865 0.5899 0.5913 0.5928
0.6 0.6964 0.7025 0.7044 0.7071
0.5 0.8034 0.8139 0.8159 0.8203
0.4 0.9129 0.9303 0.9314 0.9383
0.3 1.0302 1.0598 1.0570 1.0680
0.2 1.1646 1.2203 1.2037 1.2216
0.1 1.3385 1.5014 1.3986 1.4312
Substituting Eqs. (2-110) into Eqs. (2-104) and (2-107), we can calculate the
defocused PSFs and encircled powers. The PSF and encircled power curves for S ∫ S2 =
0.8, 0.6, 0.4, 0.2, and 0.1 are shown in Figure 2-18. Comparing the actual aberrated
encircled power with the aberration-free encircled power scaled by the actual Strehl ratio,
we find that for S ≥ 0.4, rc £ 0.8, the difference between the two is < 8 percent. For S ≥
0.1 and rc £ 0.5, the difference is < ~ 5 percent. Note that a given value of s F and,
therefore, a given value of S2 is obtained for two values of z, one for which z < R and the
other for which z > R . Since r and rc in Figure 2-18 are in units of lz D, the spread of
the irradiance distribution for a given value of S2 depends on whether z < R or z > R . It
should be clear, however, that the central irradiance and the concentration of power on a
target at a distance z are maximum when the beam is focused on it, i.e., when R = z .
The aberrated PSFs corresponding to Strehl ratios of 0.8, 0.6, 0.4, 0.2, and 0.1 for the
fourth-order spherical aberration are shown in Figure 2-19. For comparison, the
aberration-free PSF (corresponding to a Strehl ratio of 1) is also included in this figure.
The corresponding curves for encircled power are also shown in this figure. As expected,
the irradiance distribution decreases inside the Airy disc (of radius 1.22) and increases in
the bright ring around it. However, the location of the first minimum remains practically
unchanged. Moreover, its value remains close to zero. Thus, the Airy disc remains
distinct and practically unchanged in size as a fourth-order spherical aberration is
introduced into an otherwise aberration-free optical system.
1.0
Pg(rc)
0.8
F2
l(r)/S, P(rc)/S
0.6
0.4
0.2
lg(r)
0.0
0.0 0.5 1.0 1.5 2.0
r, rc
Figure 2-18. Defocused PSFs normalized to unity at the center and encircled power
corresponding to various values of S. Aberration-free ( S2 = 1) curves are included
for comparison. A Gaussian approximation of the PSFs is also included. S2 = 1, ___;
S 2 = 0.8, _ _ _ S2 = 0.6, _. _. _; S2 = 0.4; - - -; S2 = 0.2, ...; S2 = 0.1.
1.0
Pg S=1
0.8
0.8
0.6
l(r), P(rc)
0.6
0.4
0.4
lg(r)
0.2 0.2
0.1 F4
0.0
0.0 0.5 1.0 1.5 2.0
r, rc
Figure 2-19. Aberrated PSFs and encircled power for F4 corresponding to Strehl
ratios of 0.8, 0.6, 0.4, 0.2 and 0.1. Aberration-free ( S = 1) curves are included for
comparison. The PSF approximated by a Gaussian function, is also included. The
parameters r and rc are in units of l F , where F = R D is the focal ratio of the
image-forming light cone. The Airy disc radius is 1.22 in these units.
An even more dramatic result is obtained if the aberrated PSFs of Figure 2-19 are all
normalized to unity at the center as in Figure 2-20. It is evident from this figure that,
within the Airy disc, the normalized aberrated PSFs are practically identical with the
aberration-free PSF. Thus, the aberration reduces the irradiance distribution quite
uniformly by the Strehl ratio at points in the region, say, r <
~ 1. The encircled power in
this region is also correspondingly scaled by the Strehl ratio, as may be seen from the
normalized encircled power curves shown in Figure 2-20. Note that the total normalized
power is given by S -1 instead of unity. Therefore, the normalized power curves should
be taken seriously only for rc <
~ 1. Similar results are obtained for the sixth- and eighth-
order spherical aberrations, as shown in Figures 2-21 and 2-22, respectively. For these
two aberrations, only the normalized curves are shown.
We have shown that a PSF aberrated by a rotationally symmetric aberration

represented by a Zernike polynomial may be approximated by
2
È 2 J ( pr ) ˘
I (r ) = S Í 1 ˙ , 0 £ r £1 . (2-118)
Î pr ˚
If the aberration is not represented by a Zernike polynomial, e.g., if the phase aberration
is given by
1.0
Pg(rc)
0.8
l(r)/S, P(rc)/S F4
0.6
0.4
lg(r)
0.2
0.0
0.0 0.5 1.0 1.5 2.0
r, rc
Figure 2-20. Aberrated PSFs for F4 normalized to unity at the center and
corresponding normalized encircled power. The curves correspond to Strehl ratios
indicated in Figure 2-19. The Gaussian PSF and encircled power are also included.
1.0
Pg(rc)
0.8
l(r)/S, P(rc)/S
0.6
F6
0.4
0.2
lg(r)
0.0
0.0 0.5 1.0 1.5 2.0
r, rc
corresponding normalized encircled power. The curves correspond to Strehl ratios
1.0
Pg(rc)
0.8
l(r)/S, P(rc)/S
0.6
F8
0.4
lg(r)
0.2
0.0
0.0 0.5 1.0 1.5 2.0
r, rc
corresponding normalized encircled power. The curves correspond to Streh ratios
F(r) = a0 + a1r2 + a2 r 4 + a3r6 + ... , (2-119a)
then the central portion of the PSF spreads somewhat more than the Airy disc.
Szapiel14 has shown that a better approximation in such cases is obtained if the right-
hand side of Eq. (2-118) is multiplied by a factor
{
f (r ) = exp [ g1 F(r )] - g2 F(r )
2
} , (2-120)
where
[
F(r ) = 1 - (1.22 r )
2
] , (2-121a)
g1 = a1s1 + a2 s2 + a3 s3 + ... , (2-121b)
and
g2 = b1s1 + b2 s2 + b3 s3 + ... . (12-121c)
Here, bi ’s are the coefficients of the terms in F 2 , i.e.,

F 2 (r) = b0 + b1r2 + b2 r 4 + b3r6 + ... , (2-119b)
where
b0 = a02 , b1 = 2 a0 a1 , b2 = 2 a0 a2 + a12 ,
b3 = 2 a0 a3 + 2 a1a2 , b4 = 2 a0 a4 + 2 a1a3 + a22 ,
and
b5 = 2 a0 a5 + 2 a1a4 + 2 a2 a3 , b6 = 2 a0 a6 + 2 a1a5 + 2 a2 a4 + a32 . (2-122)
The coefficient a0 has been chosen to be given by
Êa a a ˆ
a0 = - 1 + 2 + 3 + ... , (2-123a)
Ë2 3 4 ¯
so that
F =0 (2-123b)
and
s 2F = F 2
b1 b2 b3
= b0 + + + + ... . (2-123c)
2 3 4
The values of the coefficients si are given in Table 2-10.
Table 2-10. Value of coefficients si in Eq. (2-121).
i si
1 0.272 443
2 0.247 985
3 0.209 272
4 0.177 535
5 0.153 009
6 0133 952
7 0.118 879
8 0.106 727
9 0.096 753
10 0.088 439
2.6.3 Gaussian Approximation

Figure 2-17 shows that, unless the value of the Strehl ratio is very small, it can be
calculated quite accurately from a knowledge of the aberration variance using the
relationship
Sg ~ exp ( – s 2F) . (2-124)
The error is < 10 percent as long as S > ~ 0.3 . The percent error is defined to be
( )
100 1 - Sg S . Note that this error is negative for the aberrations considered.
Now we approximate the PSF with a Gaussian function. The exact aberration-free
PSF and encircled power are given by Eqs. (2-15) and (2-24), respectively. According to
these equations, the PSF has a central value of unity and its total power is also unity. A 2-
D Gaussian function having a central value of unity and a total power of unity is given by
[
Ig (r ) = exp – ( p r 2)
2
] . (2-125)
The corresponding encircled power, obtained by substituting Eq. (2-125) into Eq.
(2-106), is given by
2
[
Pg (rc ) = 1 - exp ( - p rc 2) ] . (2-126)
Since, for small values of r, the aberrated PSF and the corresponding encircled power
are simply scaled by the Strehl ratio, their Gaussian approximations may be written
[
Ig (r; S ) = S exp ( - pr 2)
2
] (2-127a)
and
{ [
Pg (rc ; S ) = S 1 - exp - ( p rc 2)
2
]} , (2-127b)
respectively. For very small values of rc , Eq. (2-127b) reduces to Eq. (2-109).
Substituting for S from Eq. (2-124) into Eqs. (2-127), we obtain
( ) [
Ig (r; s F ) = exp - s 2F exp - ( p r 2)
2
] (2-128a)
and
( ){
Pg (rc ; s F ) = exp - s 2F 1 - exp - ( p rc 2) [ 2
]} , (2-128b)
respectively. It is evident, as expected, that Eqs. (2-128) are not valid for large values of r
( )
or rc . For example, although the total power is unity, Pg (rc ; s F ) Æ exp - s 2F as rc Æ • .
Equations (2-124) and (2-128) give a complete description of our Gaussian model for
aberrated systems. Of course, Eqs. (2-124) and (2-128b) can be obtained from Eq. (2-
128a). For example,
S = Ig (0; s F ) Ig (0; 0) . (2-129)
Thus, Eq. (2-128a) is the basic equation for the Gaussian model.
Figure 2-23 shows how the aberration-free exact results given by Eqs. (2-15) and (2-
24) compare with their corresponding Gaussian approximations given by Eqs. (2-125)
and (2-126). We note that the two sets of results are practically identical for r and
rc <
~ 0.5 . As r and rc increase, the differences between the two also increase.
The Gaussian PSF and encircled power for an aberrated system are shown in Figures
2-18 to 2-22. It is evident that the results for the aberration-free case scaled by the Strehl
ratio approximate the results for an aberrated case better and over a wider range than the
corresponding results obtained by the Gaussian approximation.
2.6.4 Summary of Results

For a radially symmetric aberration represented by a Zernike polynomial, we find
that although power flows out of the Airy disc and into the outer regions, the first dark
ring remains practically dark. Thus, the Airy disc remains distinct. Moreover, the
1.0
Pg(rc)
0.8
2 2
1 – J0(prc) – J1 (prc)
0.6
l(r), P(rc)
0.4
2
[2J1 (pr)/pr]
0.2
lg(r)
0.0
0.0 0.5 1.0 1.5 2.0
r, rc
Figure 2-23. Aberration-free PSF, encircled power, and their Gaussian

approximations.
aberrations reduce the irradiance distribution quite uniformly within the Airy disc scaled
simply by the Strehl ratio of the system. Accordingly, the calculation of encircled power
(
for small circles rc < )
~ 1 reduces to a knowledge of the Strehl ratio. In the case of
spherical aberrations, for S ≥ 0.4 and rc £ 1, the aberration-free encircled power scaled
by the exact Strehl ratio is in error by < 2 percent. For S ≥ 0.1, and rc £ 1, the error is <
~
10 percent. In the case of defocus, for S ≥ 0.4 and rc £ 0.8 the error is < 8 percent. For
S ≥ 0.1 and rc £ 0.5 , the error is <
~ 5 percent. The error is generally larger for a smaller
Strehl ratio and a larger value of rc .
To evaluate the integrals involved in the expressions for the Strehl ratio, PSF, and
encircled power, adequate accuracy is obtained even with an 8-point Gauss quadrature.
For example, for S ≥ 0.4 , r £ 1, the encircled power is in error by at most a few percent
(<
~ 5 percent). Using a 12-point quadrature, the error is reduced to < 2 percent for Strehl
ratios as small as 0.2 and rc as large as 3. The error generally increases as S decreases and
as rc and the aberration order increase. If the aberration is not represented by a Zernike
polynomial, the PSF in its central region spreads more than the Airy disc, and a simple
correction can be applied to account for it
The calculations can be simplified if the aberration-free PSF is approximated by a

Gaussian function. We have considered a Gaussian PSF that has the same irradiance
value at the center and the same total power as the actual PSF. The aberrated PSFs are
then approximated by a Gaussian function scaled by the Strehl ratio. The scaled Gaussian
PSF approximates the true PSFs quite well for r £ 0.5 . Use of the complete Gaussian
model, Eq. (2-128a), gives an error of < 10 percent even when rc is as much as unity.
Thus, for radially symmetric aberrations, the aberration-free results scaled by the Strehl
ratio give a better approximation over a wider range than the Gaussian approximation.
Moreover, an advantage of approximating a PSF with a Gaussian function is that the
average effect of the random motion of an image described by a Gaussian probability
distribution can be calculated very simply. As discussed in Chapter 5, the average PSF is
given by the convolution of the motion-free PSF and the probability distribution
describing the image motion. The convolution of two Gaussian functions is also a
Gaussian function whose variance is the sum of the variances of the two convolving
Gaussian functions. Hence, the average PSF is also given by a Gaussian function that is
wider than the motion-free PSF. Its central value is determined from a consideration of
power conservation.
2.7 SYMMETRY PROPERTIES OF AN ABERRATED PSF16
In Section 2.2.1 we considered the PSF of an unaberrated imaging system with a

uniformly illuminated circular exit pupil. We showed that the PSF in the Gaussian image
plane was radially symmetric about its origin. (The origin lies at the center of curvature of
the reference sphere with respect to which the aberrations are defined; and the PSF is
measured in a plane normal to the z axis, which is along the line joining the center of the
exit pupil and the center of curvature of the reference sphere.) In Section 2.5.1, we
2.7 Symmetry Properties of an Aberrated PSF 137
showed that the PSF of an unaberrated system is asymmetric about the Gaussian image
plane unless the Fresnel number N of the exit pupil as observed from the Gaussian image
point is very large. In this section, we discuss the symmetry of the irradiance distribution
of the aberrated mage of a point object about the axis of the image-forming light cone as
well as in and about the Gaussian image plane.
2.7.1 General Theory

Consider a PSF aberrated by an aberration represented by a Zernike polynomial
given by Eq. (2-64). For simplicity, we write this aberration in the form
Fnm (r, q) = Anm Rnm (r) cos mq , (2-130a)
where Anm is the aberration coefficient. We will refer to the plane in which the center of
the reference sphere lies as the Gaussian image plane. If the image is observed in another
plane that lies at a distance z from the exit pupil, then the aberration becomes
F(r, q) = Anm Rnm (r) cos mq + Bd r2 , (2-130b)
where Bd is the coefficient of defocus aberration given by Eq. (2-81b).
For an unapodized system, Eq. (2-79) for the irradiance distribution of the image of a
point object in a defocused plane at a distance z from the plane of the pupil may be
written
1 2p 2
2
ÏR¸ Û Û È R ˘
Ó pz ˛ ı ı
[ Î
]
I (r , q i ; z ) = Ì ˝ Ù Ù exp i F(r, q) exp Í - pi r r cos (q - q i )˙ r dr dq
z ˚
, (2-131)
0 0
where r is in units of l F and the irradiance is in units of the aberration-free central

irradiance. Noting that
•
exp (iz cos q) = J 0 ( z ) + 2 Â i s Js ( z ) cos sq , (2-132)
s =1
we may write
Ï R ¸
exp Ìi ÈÍF (r, q) - p r r cos (q - q i )˘˙˝
ÓÎ z ˚˛
• •
() [
= 4 Â ¢ Â ¢ i s ( -i ) s ¢ exp iBd r2 Js Anm Rnm (r)
s = 0 s¢ = 0
]
¥ Js ¢ (p r rR z) cos (msq) cos [s¢(q - q )] ,i (2-133)
where a prime on the summation sign indicates, for example, that terms with s = 0 are to
be taken with a factor of 1/2. Substituting Eq. (2-133) into Eq. (2-131) and noting that
2p
Û
ı
[ ]
Ù cos ( msq) cos s ¢(q - q i ) dq = p cos (s ¢q i ) d ms, s ¢ , (2-134)
0
where d ij is a Kronecker delta, we obtain
•
2 ( m - 1) s
I ( r, q i ; z ) = ( 4 R z ) Â ¢( - i ) cos ( msq i )
s=0
1
2
Û
( ) [
¥ Ù exp i Bd r2 J s Anm Rnm (r) J ms ( p r rR z ) r dr
ı
] . (2-135)
0
The irradiance distribution for the aberration-free case can be obtained by letting
Anm = 0 . Since Js ( 0 ) = d s, 0 , only the s = 0 term in Eq. (2-135) is nonzero. This term is
associated with a factor of 1/2. Hence, Eq. (2-135) reduces to Eq. (2-81) when the system
is aberration free. Since
È 2p ˆ ˘
cos Íms Ê q i + j = cos ( msq i ) , j = 1, 2, K, m ,
Î Ë m ¯ ˙˚
Eq. (2-135) shows that the irradiance distribution in any observation plane (normal to the
z-axis) is m-fold symmetric about the z axis. It also shows that the tangential plane z x and
all planes containing the z axis and making angles of pj m with the tangential plane are
planes of symmetry. For m = 0 the irradiance distribution does not depend on q i . This is
to be expected; for a radially symmetric aberration, the irradiance distribution in any
observation plane is also radially symmetric. These symmetry properties are also
possessed by the aberration given by Eq. (2-130a), and they are not affected by the
inclusion of the defocus aberration as in Eq. (2-130b). It is shown in Section 2.7.2 that the
distribution in the Gaussian image plane is 2m-fold symmetric when m is even, although
the aberration is only m-fold symmetric.
2.7.2 Symmetry About the Gaussian Image Plane

The irradiance distributions in two defocused planes located symmetrically about the
Gaussian image plane are not identical for reasons discussed already in Section 2.5.1.
This asymmetry of the irradiance distribution is especially large for exit pupils of small
Fresnel numbers. When the Fresnel number of the exit pupil is large, Bd becomes
significant even when z is only slightly different from R. Hence, for systems with large N,
we may limit our discussion to defocused planes in the vicinity of the Gaussian image
plane; i.e., planes for which z ~ R . Therefore, Eq. (2-135) reduces to
• ( m -1) s
I (r, q i ; z ) = 16 Â ¢( - i ) cos ( ms q i )
s=0
1
2
Û
( ) [
¥Ù exp iBd r2 J s Anm Rnm (r) Jms ( p r r) r dr
ı
] , (2-136)
0
where
pa2
Bd = ( R - z) . (2-137)
l R2
Note that z dependence is now contained in the defocus coefficient Bd only.
For m = 0 (spherical aberration), Eq. (2-136) reduces to

1 2
•
sÛ
I (r; z ) = 16 Â¢i ı
( 2
) [ 0
]
Ù exp i Bd r J s An 0 Rn (r) J 0 ( p r r) r dr (2-138a)
s=0
0
1
2
Û
ı
{[
= 4 Ù exp i An 0 Rn0 (r) + Bd r2 ] } J (p r r) rdr
0 . (2-138b)
0
Two planes located symmetrically about the Gaussian image plane at distances z = R + D
and z = R - D from the exit pupil, where D is a small distance along the z axis,
correspond to Bd = - D 8 F 2 and Bd = D 8 F 2 , respectively. It is evident from Eq. (2-
138b) that if we change the sign of Bd , we obtain a different distribution. Hence, the
irradiance distribution is not symmetric about the Gaussian image plane. It is shown in
Section 2.7.3 that the axial irradiance is symmetric about this plane. However, we note
from Eq. (2-138b) that the irradiance distribution does not change if the sign of An0 is
also changed along with that of Bd . In that case, the total aberration function of Eq. (2-
130b) (with m = 0 ) changes sign. Thus, the irradiance distribution in two symmetrically
defocused image planes are identical provided they are for spherical aberrations of equal
magnitude but opposite signs.
( m -1) s
When m is an odd integer (e.g., for coma), the quantity ( - i ) in Eq. (2-136) is
real; it is either equal to 1 or –1. Since Bd changes sign as we go from one defocused
plane to another located symmetrically about the Gaussian image plane, the values of the
integral in Eq. (2-136) for the two defocused planes are complex conjugates of each
other. Hence, the irradiance distributions in these two planes are equal; i.e., the
distribution is symmetric about the Gaussian image plane. Accordingly, the axial
irradiance is also symmetric about this plane. Note that the total aberration function
according to Eq. (2-130b) is different for the two symmetrically defocused planes.
When m is an even integer (e.g., for astigmatism)
(- i)( m -1)s = ± 1 when s is even

= ± i when s is odd
and
[ ]
cos ms(q i + p m) = cos ( msq i ) when s is even
= - cos ( msq i ) when s is odd .

Therefore,
*
( – i)( m –1)s cos (msqi ) = {(-i)( m -1)s } cos [ms(qi + p m)] .
Hence, the irradiance at a point in a defocused plane is equal to the irradiance at a point
resulting from a reflection in the Gaussian image plane and a rotation through an angle
p m about the z axis. Since the axial irradiance is unaffected by this rotation, it is
symmetrc about the Gaussian image plane. The total aberration functions corresponding
to two symmetric planes, of which one is rotated with respect to the other by p m , have
equal magnitude but opposite signs. It follows from Eq. (2-136) that the irradiance
distribution in the Gaussian image plane is 2m-fold symmetric about the z axis when m is
even, although the aberration function is only m-fold symmetric. This, of course, is true
for small values of N also.
In the Gaussian image plane, z = R and Bd = 0 , and Eq. (2-136) for the irradiance
distribution reduces to
2
• 1
( m -1) s
I (r, q i ; R) = 16 Â ¢( - i )
s=0
cos ( ms q i ) Ú
0
[
J s Anm Rnm (r)]J ms (prr)rdr . (2-139)
An interesting property on the effects of aberrations in this plane can be obtained from
◊
the above equation by expanding J s ( ) in a power series according to
• ( x / 2) s + 2 n
J s ( x ) = Â ( -1) s .
n=0 n ! ( n + s )!
Thus, we may write

2
1 1
I (r, q i ; R) = 4 Ú J 0 ( prr)rdr - 2(i ) m -1 cos ( m q i ) Anm Ú Rnm (r) J m ( prr)rdr + O Anm
0 0
2
( ) .
(2-140)
m -1
For odd values of m, since ( - i ) is real, the term depending linearly on the aberration
coefficient Anm is real [All of the terms are real in this case]. However, for even values of
m -1
m, since ( - i ) is imaginary, this term is also imaginary. Hence, we obtain the result
2
È 2 J ( pr ) ˘
I ( r , q i , R) = Í 1 ˙ + O ( Anm ) , m odd (2-141a)
Î pr ˚
2
È 2 J ( pr ) ˘
= Í 1
Î p r ˚
2
˙ + O Anm ( ) , m even . (2-141b)
Thus, for example, a first-order amount of aberration produces a first-order change in the
irradiance distribution of the aberration-free image in the case of coma, but a second-
order change in the case of astigmatism. The ring structure of the Airy pattern near its
center is affected much more strongly in the case of coma than in the case of astigmatism.
2.7.3 Symmetry of Axial Irradiance

The axial irradiance may be obtained from Eq. (2-136) by letting r = 0 . Thus, for
m π 0 , since Js ( 0 ) = d s, 0 , only the s = 0 term is nonzero, and we obtain
2
1
Û
( ) [
I (0; z ) = 4 Ù exp i Bd r2 J 0 Anm Rnm (r) r dr
ı
] ,m π 0 . (2-142)
0
It is evident from Eq. (2-142) that the axial irradiance does not depend on the sign of Bd ;
i.e., it is symmetric about the Gaussian image plane.
When m = 0 , the axial irradiance according to Eq. (2-138b) may be written

2
1
Û
( ) [
I (0; z ) = 4 Ù exp i Bd r2 exp i An 0 Rn0 (r) r dr
ı
] . (2-143)
0
Since
(
R20n (r) = Pn 2r2 - 1 ) , (2-144)
◊
where Pn ( ) is a Legendre polynomial of degree n, then letting
2r 2 - 1 = x ,
we may write Eq. (2-143) in the form
1 2
Û
I (0; z ) = (1 4)
ı
[
exp (i Bd 2) Ù exp (i Bd x 2) exp i An 0 Pn 2 ( x ) dx ] . (2-145)
-1
For spherical aberration of any order, n 2 is even, and, therefore, Pn 2 ( - x ) = Pn 2 ( x ) .

Accordingly, if we change the sign of Bd and change x to –x, we obtain the original
result. Hence, the axial irradiance is symmetric about the Gaussian image plane. In
Section 2.5.1 we obtained this result for defocus (n = 2 , m = 0) , and in Section 2.3.6 for
primary balanced spherical aberration, which is equivalent to a Zernike polynomial with
n = 4 and m = 0 . Hence, the axial irradiance is symmetric about the Gaussian image
plane for any aberration expressed by a Zernike polynomial. (If a wavefront tilt is
present, then the symmetry holds along the tilted axis.)
2.7.4 Symmetry in Sign of Aberration Coefficient

Since
J s ( – z ) = ( –1) s J s ( z ) , (2-146a)
we may write for even values of m,

*
(- i)( m - 1)s Js [ Anm Rnm (r)] = [
(- i)( m - 1)s ] J [- A
s
m
nm Rn (r)] . (2-146b)
Hence, we note from Eq. (2-139) that a change in the sign of the aberration coefficient
Anm for even values of m (including zero) has no effect on the irradiance distribution in
the Gaussian image plane. (This is true for systems with small values of N as well.) This
may be seen more easily from Eq. (2-141b) when m = 0. Thus, the sign of such an
aberration cannot be determined from the Gaussian image-plane distribution. In a
defocused plane, the total aberration function changes when the sign of the aberration is
changed, and so the distribution also changes. However, for even m, distributions in two
symmetrically defocused planes are the same, provided the aberration coefficient for one
has a sign opposite to that for the other. Similarly, for odd m, the distributions in any
observation plane for aberration coefficients with opposite signs are the same, provided
one is rotated with respect to the other by an angle p. In each case, the total aberration
functions are the same except for a difference in their signs.
From the general discussion given above, it is easy to obtain the symmetry properties
of the PSFs aberrated by primary aberrations. These are discussed in the next section.
2.8 PSFS FOR PRIMARY ABERRATIONS

We now describe briefly the full PSFs for varying amounts of primary aberrations
and show them graphically. In particular, we show the 2-D PSFs for various values of a
primary aberration, as would be seen in practice when observing the aberrated image of a
point object. It should be noted that the Zernike aberrations (e.g., spherical, astigmatism)
considered in Section 2.7 contain a defocus term. Hence, the term Gaussian image plane
used there would be different from the one used in this section.
2.8.1 Defocus Bd r2
As explained in Section 2.5.1, the irradiance distribution is symmetric about the
Gaussian image plane when the Fresnel number is large. Figure 2-24 shows the defocused
PSFs for several values of the peak defocus aberration Bd , such that the aberration-free
central value is unity. As Bd increases, the central value, i.e., the Strehl ratio, decreases.
The minima of the aberration-free PSF are no longer zero. The central value is zero when
Bd is an integral number of waves, as may be seen from Eq. (2-84c). Since the
corresponding PSFs have very low values, they have been multiplied by a factor of 10 in
the figure. The PSF for Bd = 0.64l is included here because, as will be discussed in
Section 2.10, the OTF is negative for certain spatial frequency bands for values of Bd
larger than this.
2.8.2 Spherical Aberration Combined with Defocus As r4 + Bd r2 ( )

The irradiance distribution is rotationally symmetric about the z axis. Accordingly, it
is radially symmetric in any observation plane (normal to the z axis), regardless of the
value of the Fresnel number. For large Fresnel numbers, the irradiance along the z axis is
2.8 PSFs for Primary Aberrations 143
1.00
Bd = 0 Defocus
W(r) = Bdr2
0.75
0.25
I (r; Bd)
0.50 1
0.5
10 ¥
0.25
3 10 ¥
0.64
2
10 ¥
0.00
0 1 2 3
r
Figure 2-24. PSFs of a defocused system. Bd represents the peak value of defocus
aberration in units of l . The curves for Bd = 1, 2, and 3 have been multiplied by ten.
symmetric about the defocused point corresponding to Bd = - As . The irradiance

distributions in two observation planes located symmetrically about this point are not
equal to each other unless they are for aberration coefficients with opposite signs. The
distribution in the z = R plane does not change if we change the sign of the aberration
coefficient As .
Figure 2-25 shows the axial irradiance for As = 1l, 2 l, and 3 l . It is evident that
the axial irradiance is symmetric about the defocused point corresponding to Bd = - As ;
i.e., about the plane with respect to which the aberratioan variance is minimum.
1.0
W(r) = Asr4
1
0.8
As = 0
0.6
I(0; Bd)
2
0.4
3
0.2
0.0
–6 –5 –4 –3 –2 –1 0 1 2 3
Bd
Figure 2-25. Axial irradiance of a beam aberrated by spherical aberration As in

units of l . The arrow on the curve for As = 3 is located at the point of symmetry.
The axial irradiance is shown in a different form in Figure 2-9.
Figure 2-26 shows the PSFs for spherical aberration combined with defocus. Figure
2-26a shows the PSFs in defocused image planes Bd = - As corresponding to minimum
aberration variance. The PSF for As = 2.2 l is also included here because, as discussed
later in Section 2.10, the OTF is negative for certain spatial frequency bands for values of
As larger than this. We note that the radius of the central bright spot does not change as
the aberration is increased. Figure 2-26b shows the PSFs for As = 1 l and Bd As = 0 , –1,
–1.5, and –2, corresponding to the Gaussian, minimum-aberration-variance, circle-of-
least-confusion, and marginal image planes, respectively. The PSFs in the Gaussian and
marginal image planes have been multiplied by a factor of 10 in this figure because of
their low values. We note that the central irradiances in these two planes are equal to each
other, showing that the axial irradiance is symmetric about the Bd = - As plane, as
pointed out in Table 2-11, where the symmetry properties are summarized. It is also clear
from the figure that the best image is obtained in the plane corresponding to minimum
aberration variance. The PSFs in other defocused image planes illustrating symmetry of
axial irradiance are shown in Figure 2-26c.
2.8.3 Astigmatism Combined with Defocus Aa r2 cos 2 q + Bd r2 ( )

The irradiance distribution for astigmatism in any observation plane is symmetric
about two orthogonal axes, one of them lying in the tangential plane, regardless of the
value of N. It is four-fold symmetric in the plane Bd = - Aa 2. For large values of N, the
axial irradiance of a beam aberrated by astigmatism is shown in Figure 2-27. It is
symmetric about the defocused point corresponding to Bd = - Aa 2. Although the
symmetry point corresponds to minimum aberration variance, the peak value of
irradiance does not lie at it for Aa ≥ l . The irradiance distributions in two observation
planes located symmetrically about the plane corresponding to Bd = - Aa 2 are equal to
each other, provided one is rotated with respect to the other by p 2. Figure 2-28 shows
the PSFs in defocused planes Bd = - Aa 2 corresponding to minimum-aberration-
variance, along the directions q i = 0 and p 4 .
Figure 2-29 shows the PSFs for Aa = 1l and Bd Aa = 0, – 1/2, and – 1,

corresponding to sagittal-line, minimum-aberration-variance [circle-of-least (astigmatic)-
confusion], and tangential-line image planes, respectively. (As discussed in Section 4.3.3
Table 2-11. Symmetry properties of aberrated PSFs.

Aberration General Symmetry Symmetry of Symmetry in Symmetry in
Axial Irradiance* Defocused Images* Coefficient Sign
None Rotational about z axis About Bd = 0 About Bd = 0 Not applicable
Radial in any z plane
Spherical Rotational about z axis About Bd = – As About Bd = – As In Bd = 0 plane
As r 4 if As Æ – As
Coma About tangential plane About Bd = 0 About Bd = 0 If rotated by p about
Line symmetry in any z z axis
Ac r 3 cos q plane about x axis
Astigmatism Line symmetry about x About Bd = – Aa / 2 About Bd = – Aa / 2 In Bd = – Aa / 2 plane
and y axes
Aa r 2 cos 2 q if rotated by p / 2 or
4-fold in Bd = – Aa / 2 if Aa Æ – Aa
plane
*For large values of the Fresnel number N.

1.00
0.25
0.5 As = 0 Balanced Spherical

W(r) = As(r4 – r2)
0.75
1
I (r; As)
0.50 (a)
2
0.25 2.2
3
0.00
0 1 2 3
r
1.00
Bd = 0 Spherical and Defocus

W(r) = Asr4 + Bdr2
0.75 As = 1
10 ¥
I (r, Bd)
–1
0.50 (b)
10 ¥
0.25 – 1.5 –2
0.00
0 1 2 3
r
1.0 0.20
Aberration free W(r) = Asr4

0.8 As = 1 0.16
Bd = – 1
0.6 0.12
I(r; Bd)
– 0.5
0.4 0 (5 ¥) 0.08
– 1.5
0.2 – 2 (5 ¥) 0.04
0.0 0.00
0 1 2 3 4
r
Figure 2-26. PSFs for spherical aberration. (a) Various values of As and defocused
image planes Bd = - As corresponding to minimum aberration variance. (b) and (c)
Fixed value of As and various image planes. Bd and As are in units of l and r is in
units of l F. The curves for Bd = 0 and - 2 have been multiplied by 10 in (b).
Similarly, they have been multiplied by 5 in (c), or ignore the factor of 5, but use the
right-hand scale for these curves.
1.0
1 Aa = 0
W(r,q) = Aar2cos2q
0.8
5
I(0; Bd)
0.6
2
0.4
3
5
0.2
5
0.0
–6 –5 –4 –3 –2 –1 0 1 2 3
Bd
Figure 2-27. Axial irradiance of a beam aberrated by astigmatism Aa . An arrow on

a curve locates its point of symmetry corresponding to Bd = - Aa 2. For clarity, the
irradiance values have been multiplied by five, except when Aa = 0 .
1.00
Aa = 0 Balanced Astigmatism
W(r,q) = Aa(r2 cos2q - r2/2)
0.25
0.75 qi = 0
I (r, Aa)
0.5
0.50 (a)
0.25 2 10 ¥
1
3 10 ¥
1.28
0.00
0 1 2 3
r
1.00
Aa = 0 Balanced Astigmatism
W(r,q) = Aa(r2 cos2q - r2/2)
0.25
0.75 qi = p/4
I (r, Aa)
0.5
0.50 (b)
2
10 ¥
0.25
1
3 10 ¥
0.00
0 1 2 3
r
Figure 2-28. PSFs for various amounts of astigmatism in defocused image planes
corresponding to minimum aberration variance along the directions. (a) q i = 0 and
(b) p 4. Aa represents the peak value of astigmatism in units of l and a defocused
image plane is represented by Bd = - Aa 2 .
Bd = 0 Astigmatism and Defocus

W(r,q) = Aar2 cos2q + Bdr2
0.2 Aa = 1
qi = 0
I (r; Bd)
– 0.5
(a)
0.1
–1
0.0
0 1 2 3
r
(a)
Astigmatism and Defocus

0.2 W(r,q) = Aar2 cos2q + Bdr2
Aa = 1
qi = p/4
I (r; Bd)
Bd = – 0.5 (b)
0.1
0, –1
0.0
0 1 2 3
r
(b)
Figure 2-29. PSFs for astigmatism with Aa = 1 in image planes Bd Aa = 0 , – 1 2 ,

and –1, corresponding to sagittal-line, minimum-aberration-variance, and
tangential-line image planes, respectively, along the directions. (a) q i = 0 and (b)
p 4.
of Part I), the two line-image planes are those in which the astigmatic focal lines based on
geometrical optics are obtained.) We note from Figure 2-29a that the central irradiance of
the line images are equal to each other, showing that the axial irradiance is symmetric
about the Bd = - Aa 2 plane, as pointed out in Table 2-11. The distributions of the two
line images are rotated with respect to each other by p 2. Accordingly, the solid curve in
Figure 2-29b corresponds to both Bd Aa = 0 and – 1.
(
2.8.4 Coma Acr3 cosq )
Figure 2-30 shows the axial irradiance for coma with Ac = 1l . The axial irradiance
for coma without any tilt is along the z axis. For coma optimally balanced with tilt, the
axial irradiance shown is along an axis that is parallel to the z axis but passing through the
diffraction focus (4/3, 0). The irradiance is symmetric about the Gaussian image plane. It
is evident from the figure that the central value of the PSF is much larger at the
diffraction focus than at the Gaussian focus.
The PSF for coma is symmetric about the tangential plane z x . Thus, it has a line
symmetry in any observation plane, the line lying in the tangential plane, i.e. the PSF is
symmetric about the xi axis. The distribution in two observation planes located
symmetrically about the z = R plane are identical. The PSF is also symmetric about the
Gaussian image plane. A change in the sign of the aberration coefficient Ac produces a
rotation of the distributions by p about the axis. The centroid of the PSF does not lie at
the Gaussian image point due to its asymmetric distribution. Its location is discussed in
Section 2.9.
Figure 2-31 shows the PSFs for coma along the directions for q i = 0, p 4 , p 2,
where q i = 0 corresponds to the xi axis. Similarly, q i = p 2 corresponds to the yi axis.
The diagonal corresponding to q i = p 4 is indicated by xi¢ in Figure 2-31b. We note that,
as pointed out in Section 2.3.3, the peak value of an aberrated PSF does not lie at the
Gaussian image point (r = 0) . Its location is discussed in Section 2.9.2.2. Figure 2-31c
shows the symmetry of the PSFs in yi , i.e., about the xi axis.
0.6
W(r,q) = Acr3cosq
0.5 Ac = 1
0.4 Through
(4/3, 0)
I(Bd)
0.3
0.2
On axis
0.1
(0, 0)
0.0
–4 –3 –2 –1 0 1 2 3 4
Bd
Figure 2-30. Axial irradiance through the Gaussian image point and the diffraction
focus for coma with Ac = 1 .
1.00
Ac = 0 0.25 Coma
W(r,q) = Acr3 cosq
qi = 0
0.75 0.5
I (xi; Ac)
1
0.50 (a)
0.25
2
3
0.00
– 2.5 0.0 2.5 5.0
xi
1.00
Ac = 0 Coma
W(r,q) = Acr3 cosq
0.75 qi = p /4
0.25
I (x⬘i ; Ac)
0.50 (b)
0.5
0.25
3
1
2
0.00
– 2.5 0.0 2.5 5.0
xi⬘
1.00
Ac = 0 Coma
W(r,q) = Acr3 cosq
0.75 0.25 qi = p /2
I (yi; Ac)
0.50 (c)
0.5
0.25
3 2
1
0.00
– 2.5 0.0 2.5 5.0
yi
Figure 2-31. PSFs for coma along the directions (a) q i = 0, (b) q i = p 4 , and
(c) q i = p 2 . Ac is the peak value of coma aberration in units of l . Symmetry of the
PSF about the xi axis is evident in (c).
2.8.5 2-D PSFs

The PSF Figures 2-24 through 2-31 are useful for quantitative assessment of the
effect of aberrations. However, they do not lend themselves easily to what the PSFs may
look like when observed in practice. This is especially true when the PSF is not radially
symmetric. For this reason, we give computer-generated 2-D pictures of the PSFs in this
section for various values of a primary aberration. The emphasis of these pictures is on
the structure of a PSF, i.e., on the distribution of its bright and dark regions, and not on its
irradiance distribution. Some of the symmetry properties of the aberrated PSFs discussed
above are evident from these pictures. It should be clear, however, that a random mixture
of various aberrations will only lead to a complex PSF.
Figure 2-32 shows how the aberration-free PSF, or the Airy pattern, appears when
observed in a defocused image plane. The central irradiance approaches zero when a
defocus aberration of integral number of waves is introduced; hence the dark spot in the
center when Bd = 1, 2, or 3 l . The PSFs aberrated by one wave of spherical aberration
( As r 4 with As = 1l) are shown in Figure 2-33 when observed in defocused image planes
that are often considered in geometrical optics (see Figure 4-3 of Part I). These planes are
the Gaussian ( Bd = 0 ), minimum variance ( Bd = - As ), circle of least confusion
( Bd = - 1.5 As ), and marginal ( Bd = - 2 As ). The PSFs for an increasing amount of
spherical aberration observed in a corresponding defocused image plane Bd = - As , so
that the variance of the aberration is minimum, are shown in Figure 2-34. There is no
dark spot at the center of the PSFs aberrated by spherical aberration.
The PSFs aberrated by astigmatism and observed in the Gaussian or sagittal

( Bd = 0 ), minimum variance ( Bd = - Aa 2 ), and tangential ( Bd = - Aa ) image planes
are shown in Figure 2-35. As expected from both diffraction and geometrical optics, the
sagittal and tangential images are identical to each other except for a rotation of p 2 of
one with respect to the other. See, for example, the PSFs for Aa = 1l and Bd = 0 and
- 1l . Asymmetry of the aberrated images about the Gaussian image plane is also
illustrated, as may be seen by comparing the PSFs for Aa = 1l and Bd = ± 0.5 l . As the
aberration increases, the PSFs begin to look more like the line images of geometrical
optics, as exemplified by the PSF for Aa = 3 l . Similarly, the elliptical pattern of
geometrical optics is illustrated by the example of Aa = 3 l and Bd = 1.5 l . The bilateral
symmetry of the PSFs is evident from Figure 2-35. A PSF becomes 4-fold symmetric
when observed in a defocused image plane Bd = - Aa 2 to yield minimum variance (or
circle of least confusion in geometrical optics), as shown in Figure 2-36. The PSFs in two
planes located symmetrically about the plane for balanced astigmatism are identical to
each other except for a rotation of p 2. The sagittal and tangential images are a special
case of this general result.
The PSFs aberrated by coma are shown in Figures 2-37. Their symmetry about the
xi axis is evident. Moreover, it is easy to see that their centroid does not lie at the
Gaussian image point. As the aberration increases, they begin to take the conical shape of
the geometrical PSFs. The locations of peak irradiance and centroid are discussed in
Section 2.9.2.2.
Bd = 0 Bd = 0.5
Bd = 1 Bd = 1.5
Bd = 2 Bd = 3
Figure 2-32. Defocused PSFs. Bd represents the peak value of defocus wave
aberration in units of l . The central value of the PSF is zero when Bd is equal to an
integral number of wavelengths.
(a) Bd = 0 (b) Bd = - 1
(c) Bd = - 1.5 (d) Bd = - 2
Figure 2-33. PSFs aberrated by one wave of spherical aberration ( As r4 with

As = 1 l ) observed in various defocused image planes indicated by Bd in units of l .
(a) Gaussian. (b) Minimum variance. (c) Least confusion. (d) Marginal.
As = 0.25 As = 0.5
As = 1 As = 2
As = 3
[ ( )]
Figure 2-34. PSFs for balanced sphercial aberration As r4 - r2 . Thus, a PSF is
observed in a defocused image plane corresponding to Bd = - As . The aberration
coefficient As is in units of l .
Aa = 1, Bd = 0 Aa = 1, Bd = - 0.5
Aa = 1, Bd = - 1 Aa = 1, Bd = 0.5
Aa = 3, Bd = 0 Aa = 3, Bd = 1.5
Figure 2-35. PSFs aberrated by astigmatism observed in various image planes.

Bd = 0, - Aa 2, and - Aa represent the Gaussian or sagittal, minimum-variance or
circle-of-least (astigmatic)-confusion, and tangential image planes. The aberration
coefficient Aa is in units of l .
Aa = 0.25 Aa = 0.5
Aa = 1 Aa = 2
Aa = 3
( )
Figure 2-36. PSFs for balanced astigmatism Aa r2 cos 2 q - r2 2 . Thus, Bd = - Aa 2,
and the PSFs are 4-fold symmetric. The aberration coefficient Aa is in units of l .
Ac = 0.25 Ac = 0.5
Ac = 1 Ac = 2
Ac = 3
( )
Figure 2-37. PSFs aberrated by increasing amount of coma Acr3 cosq . They are
symmetric about the horizontal ( xi ) axis. The peak and the centroid of the PSFs do
not lie at the Gaussian image point. The aberration coefficient Ac is in units of l .
2.8.6 Comparison of Diffraction and Geometrical PSFs17

It is a common practice in lens design to look at the spot diagrams for qualitative
assessment of the quality of a design. The more closely packed the rays are in the image
of a point object, the better the design. The distribution of the density of rays in an image
is the geometrical PSF of a system for the point object under consideration. The
analytical geometrical PSFs for primary aberrations are discussed in Chapter 4 of Part I.
As the design improves, a designer begins to look at the diffraction PSF or the OTF as the
last step for a quantitative assessment. It was shown in Section 1.7.5 that the diffraction
image reduces to the geometrical image as the object wavelength approaches zero. In
view of this, we compare the geometrical and diffraction PSFs and establish a golden rule
of optical design, similar to Rayleigh’s l 4 rule for aberration peak value, or Maréchal’s
criterion of l 14 aberration standard deviation, that a design is close to its diffraction
limit if the ray spot radius is less than or equal to the radius of the Airy disc.
For an aberration-free system, the wavefront exiting from the exit pupil is spherical,
and all of the object rays transmitted by the system converge to its center of curvature
where the Gaussian image point lies. Accordingly, the irradiance or the density of rays is
infinity at this point and zero elsewhere in the Gaussian image plane. Thus, the
aberration-free geometrical PSF is a Dirac delta function. The corresponding diffraction
PSF is the Airy pattern consisting of a bright central Airy disc of radius 1.22l F
containing 83.8% of the total power P surrounded by alternating dark and bright rings.
The central irradiance is finite and given by p P 4l2 F 2 . The radius of the (aberration-
free) central bright disc increases linearly with the wavelength l , and the central
irradiance decreases as l- 2 . The aberrated geometrical PSFs increase linearly in size as
the aberration increases. They are completely independent of the wavelength except for
any dependence of the aberration on it. The diffraction PSFs, on the other hand, undergo
dramatic changes as the aberration increases.
A defocused but otherwise aberration-free geometrical PSF has a uniform irradiance

of P / p ( 4 FBd )2 across a circle of radius 4FBd . As the peak value Bd of defocus wave
aberration increases, the spot radius increases linearly with it, and the value of the
uniform irradiance across it decreases as Bd- 2 . The defocused PSF based on diffraction
does not have a uniform irradiance (see Figure 2-24). Its central irradiance, for example,
[ ]
varies as sin( p Bd l ) ( p Bd l ) 2 . Thus, the axial irradiance of a focused beam is zero at
points for which Bd is an integral multiple of l .
The aberrations reduce the central irradiance of the Airy pattern and increase it at the
other points. In the case of spherical aberration, the radius of the central bright spot does
not change significantly as the aberration increases. The diffraction PSFs are not bounded
in size as the geometrical PSFs are. Moreover, the diffraction PSF is zero at certain points
compared to the geometrical PSF, which is nonzero everywhere inside the image spot.
The presence of zeros in a diffraction PSF is due to the destructive interference of
Huygens’ spherical wavelets exiting from the exit pupil, a phenomenon that is absent in
geometrical optics. Note also that the irradiance in a diffraction PSF is never infinity
unlike at some points in a geometrical PSF. For small aberrations, the central irradiance
( )
of the diffraction PSF varies approximately as exp - s 2F , where s 2F is the variance of
the phase aberration across the exit pupil.
In the case of coma, the irradiance of the geometrical PSF is infinite along the lines
of a (bounded) cone of semiangle 30 ˚ with its vertex at the Gaussian image point. The
peak value of the diffraction PSF does not occur at the Gaussian image point, unlike the
geometrical PSF which has a value of infinity at that point. For small aberrations, the
peak value occurs at the point [( 4 / 3) FAc , 0] , where Ac is the peak value of the coma
wave aberration. For large aberrations there is no simple relationship for obtaining the
peak-value point. The centroid of the diffraction PSF is identical with that of the
geometrical PSF. The two PSFs begin to resemble each other qualitatively as the
aberration increases (Compare Figure 2-33 with Figure 4-8 of Part I).
The diffraction PSFs for astigmatism are completely different from the
corresponding uniform geometrical PSFs. The general elliptical spot image in the case of
astigmatism also has a uniform irradiance. The sagittal and tangential line images, which
are special cases of the elliptical image, have infinite irradiance. As the aberration
increases, the central bright spot of the Airy pattern is replaced by a thick line image
corresponding to the sagittal and tangential line images of the geometrical PSFs.
In geometrical optics, we balance spherical aberration with defocus in order to

minimize the spot radius or its rms value. The balanced aberration giving the smallest ray
spot is As [r 4 - (3 / 2) r2 ] . Similarly, the balanced aberration that gives the smallest rms
spot radius is As [r 4 - ( 4 / 3) r2 ] . Aberration balancing in this manner has been
considered for improving the modulation transfer function for low spatial frequencies.13-
1 5 Based on diffraction, the balanced aberration that yields minimum variance and
therefore maximum Strehl ratio for small aberration is As [r 4 - r2 ] , similar to the Zernike
circle polynomial R40 (r) .
Measuring the centroid of a PSF with respect to a point other than the Gaussian
image point is equivalent to introducing a wavefront tilt in the aberration function. Thus,
the balanced coma aberration that yields the minimum rms spot radius is given by
( )
Ac r3 - r cos q . The balanced aberration for minimum variance is
[ ]
Ac r3 - (2 / 3) r cos q , similar to the Zernike circle polynomial R31 (r) cos q , consistent
with the peak value of the PSF being at the point [( 4 / 3) FAc , 0] for small aberrations. In
the case of astigmatism, the smallest geometrical spot (the circle of least confusion) and
minimum aberration variance are obtained when Bd = - Aa 2; i.e., the balanced
(
aberration based on both geometrical and diffraction optics is Aa r2 cos 2 q - r2 2 , )
similar to the Zernike polynomial R22 (r)cos 2q.
It is a common practice in lens design to look at the spot diagrams in the early stages
of a design, in spite of the fact that they do not represent reality. Just as in the diffraction
treatment an optical system is considered practically diffraction limited if it yields an
image with a Strehl ratio of 0.8, or the peak (or peak-to-valley) aberration is less than
2.9 Line of Sight of an Aberrated System 159
l 4 (Rayleigh's quarter-wave rule), or the standard deviation of the aberration across the
exit pupil is less than l 14 (Maréchal’s criterion), similarly there is a golden rule in
optical design that a system is close to its diffraction limit if the ray spot radius is less
than or equal to the radius of the Airy disc. We note, for example, that this holds for
spherical aberration in the Gaussian image plane if As £ l 6 .6 , although a larger value of
As is obtained in the other image planes. Considering that the long dimension of the
coma spot is 6FAc and the line image for astigmatism is 8 FAa long, the aberration
tolerance for the spot size to be smaller than the Airy disc is Ac < 0.4 l and Aa < 0.3 l ,
respectively. The aberration tolerances based on the spot size are summarized in Table 2-
12. These tolerances are roughly consistent with Rayleigh’s quarter-wave rule. Hence, it
is reasonable to use the size of the spot diagrams as a qualitative measure of quality of
the design until it becomes smaller than the Airy disc. Thus, as the design improves, a
designer strives for spot diagrams of a size smaller than or equal to that of the Airy disc,
and then analyzes the system by its aberration variance and diffraction characteristics
such as the PSF or the MTF, since only a diffraction PSF represents the actual image
quality of a system. Key results of the comparison of the geometrical and diffraction
PSFs are summarized in Table 2-13.
2.9 LINE OF SIGHT OF AN ABERRATED SYSTEM18
In this section, we discuss the line of sight (LOS) of an aberrated system in terms of
the centroid of its PSF. It is shown that only coma terms of an aberration function affect
the LOS. Moreover, terms with cosq dependence affect the x-LOS and those with the sinq
dependence affect the y-LOS. Hence, in considering the tolerances for coma terms, their
impact on the LOS of the system must be taken into account. Numerical results are given
showing that it may be advantageous to define the LOS in terms of the peak of the PSF
rather than its centroid.
2.9.1 PSF and its Centroid

The irradiance distribution of the image of a point object formed by an aberrated
system in the Gaussian image plane is given by Eq. (2-79):
Table 2-12. Aberration tolerance based on the ray spot size
Aberration Spot ‘radius’ in Tolerance for

Gaussian image plane near diffraction limit*
Spherical 8FAs As £ l / 6.6
Coma 3FAc Ac £ 0.4l
Astigmatism 4FAa Aa £ 0.3l

*Based on the golden rule of optical design
Table 2-13. Comparison of geometrical and diffraction PSFs

Parameter Geometrical PSF Diffraction PSF
Aberration free Dirac delta function Airy pattern
Size Finite Infinite*
Infinite irradiance Yes No
Zero irradiance No Yes
Uniform irradiance Yes No
Best image Smallest spot Max Strehl ratio
(or rms radius) (for small aberrations)
Aberration increase Spot size increases Image structure changes
Balanced aberration
Spherical r 4 - 1.5 r2 r 4 - r2 or R40 (r)
Coma (r3 – r) cos q [r3 – (2 / 3)r] or R31 (r) cos q
Astigmatism r2 cos 2 q – r2 / 2 r2 cos 2 q – r2 / 2 or R22 (r) cos 2 q
Centroid Same Same
Wavelength dependence No Yes
* In practice, the diffraction PSF will also have a finite extent beyond which its value will be too low to be measurable.
r r 2 pi r r ˆ r
2
Û
Ii (ri ) = ( Pex Sex l2 R 2 Ù exp i F rp exp Ê -
) [ ( )] r r dr ◊ . (2-147)
ı Ë lR p i ¯ p
The PSF is obtained from Eq. (2-147) by dividing both sides by Pex .
The LOS of an aberration-free optical system coincides with the center of its PSF.
For an aberrated system, let us define its LOS as the centroid of its aberrated PSF. Let the
aberration function of the system in terms of Zernike circle polynomials be given by
• n 12
W (r, q) = Â
n=0 m=0
[
Â 2 (n + 1) (1 + d m 0 ) ] Rnm (r) (cnm cosmq + snm sinmq) , (2-148)
where cnm and snm are the orthonormal Zernike aberration coefficients representing
standard devi-ation of a corresponding aberration term, with the exception of c00 and s00 .
For a system with a uniformly illuminated circular exit pupil of radius a, the centroid of
its aberrated PSF, following Eq. (1-181), is given by
2p
Û
xi , yi = ( R Sex ) Ù W ( a, q) (cosq, sinq) a dq , (2-149)
ı
0
where W ( a, q) is obtained from Eq. (2-148) by letting r = 1, and we have used the
relations
( xˆ p , yˆ p ) = (cos q, sin q) , (2-150a)
and
r
(
d s = x p , y p dq . ) (2-150b)
Noting that Rnm (1) = 1 , Eq. (2-149) reduces to

•
xi , yi = 2 F Â ¢ 2(n + 1) (cn1 , sn1 ) , (2-151)
n =1
where a prime indicates a summation over odd integral values of n. We note that only
those aberrations contribute to the LOS that vary with q as cosq and sinq. Aberrations
varying as cosq contribute to xi , and those varying as sinq contribute to yi . This
may also be seen from the symmetry properties of the aberrations. Since Rn1 (r) consists
of terms in rn , rn -2 , K, and r, therefore Rn1 (r) cos q , for example, is symmetric in y p
but not in x p . Hence, the PSF is symmetric in yi , as may be seen from Eq. (2-147).
Accordingly, yi = 0 for this aberration. For a given value of cn1 or sn1 , an aberration
of a higher order gives a larger LOS error. Thus, two Zernike aberrations with m = 1 but
different values of n having the same standard deviation give different LOS errors, even
though they give (approximately) the same Strehl ratio. A higher-order aberration, i.e.,
one with a larger value of n, yields a larger LOS error. Hence, in tolerancing an optical
system, one should be careful in allocating equal standard deviation to two aberration
terms that also contribute to the LOS error.
If we consider an aberration of the form
W (r, q) = Â¢ Wn rn cos q , (2-152)

n
where Wn is in units of l, and use normalized quantities of Eqs. (2-8) through (2-10), Eq.
(2-147) may be written
2
1 2p
Û Û Ï È ˘¸
I (r, q i ) = p -2 Ù Ù exp Ìp i ÍÂ¢ 2Wnrn cos q - rr cos(q - q i )˙ ˝ r dr dq , (2-153)
ı ı Ó Î n ˚˛
0 0
Integration over q can be carried out if we let
Â¢ 2Wnr cos q - rr cos(q - q i ) = B cos(q - y ) ,

n
(2-154)
n
where
2
Ê ˆ
B2 = Á Â¢ 2Wnrn - rr cos q i ˜ + (rr sin q i )
2
(2-155)
Ën ¯
and
Ê ˆ
tan y = - rr sin q i Á Â¢ 2Wn rn - rr cos q i ˜ . (2-156)
Ën ¯
Thus, Eq. (2-153) reduces to
2
1 2p
Û Û
I (r, q i ) = p -2 [ ]
Ù Ù exp p i B cos (q - y ) r dr dq
ı ı
,
0 0
2
È1 ˘
ÍÛ
= 4 Ù J 0 ( p B) r dr˙ , (2-157)
Íı ˙
Î0 ˚
where we have used Eq. (2-12).
Substituting Eq. (2-152) into Eq. (2-149), we obtain the centroid of the aberrated
PSF,
xi = 2 F Â ¢ Wn , (2-158)
n
where it is understood that yi = 0 . Thus, the LOS error depends on the value of the
peak aberration Wn but not on n. We note that for n = 1, the aberration is a tilt and for
n = 3 it is primary coma, but they both give the same LOS error if W1 = W3 , even though
the corresponding PSFs are completely different. The reason for the same LOS error is
that for a uniform circular pupil, the centroid depends only on the aberration along the
perimeter of the pupil, which depends on Wn but not on n. From Eq. (2-153), it is evident
that the centroid of a PSF aberrated by balanced coma represented by a Zernike
polynomial Rn1 (r) cos q does not lie at the origin. However, if the aberration is balanced
for minimum spot radius (see Table 4-3 of Part I), then the centroid lies at the origin; in
other words, such an aberration minimizes the spot radius with respect to the centroid.

2.9.2.1 Wavefront Tilt
The aberration corresponding to a wavefront tilt is given by
W (r, q) = Wt r cos q . (2-159)
The PSF simply shifts such that its peak and centroid locations move from (0, 0) to
< x > = 2Wt . (2-160)
Note that, whereas x is units of l F , Wt is in units of l .
2.9.2.2 Primary Coma

The primary coma aberration is given by
W (r, q) = W3r3 cos q . (2-161)

The aberrated PSF may be obtained from Eq. (2-157) by substituting Eq. (2-156) into it
with n = 3. If we let q i = 0 , then y = 0 ; and Eq. (2-157) giving the PSF along the x axis
may be written
2
È1 ˘
Û
I ( x ) = ÍÙ J 0 (p B) dt ˙ , (2-162)
Íı ˙
ÍÎ 0 ˙˚
where
t = r2 (2-163)
and
B = (2tW3 - x ) t 1 2 . (2-164)
Figure 2-38 shows how I ( x ) varies with x for several typical values of W3 ∫ Ac (in
units of l ) varying from 0 to 3. Equation (2-158) shows that the centroid of the PSF is
given by
< x > = 2W3 . (2-165)
1.0
W3 = 0
0.8 0.5
0.6
l(x, 0)
0.4
1.5
0.2
2
0.0
–3 –2 –1 0 1 2 3
x
Figure 2-38. PSF for several values of coma W3 in units of l .

For small values of W3 , the peak value of the aberrated PSF occurs at a point such that, if
the aberration is measured with respect to a reference sphere centered at this point, the
variance of the aberration across the pupil is minimum. From the properties of Zernike
polynomials, we find that the polynomial R31 (r) cos q gives the optimum combination of
r3 cos q and r cos q terms leading to a minimum variance. Since
R31 (r) = 3 r3 - 2r , (2-166)
we note that, for small values of W3 , the peak value of the aberrated PSF occurs at
x m = ( 4 3) W3 , (2-167)
where the subscript m refers to the point corresponding to minimum aberration variance.
From the form of the aberration, it is evident that ym = 0 . Thus, primary coma shifts the
centroid and the peak of the PSF by different amounts, the movement of the peak being
two-thirds of the shift of the centroid.
Figure 2-39a shows how the irradiance I m at x m , the peak irradiance I p and the
irradiance Ic at < x > vary with W3 . Figure 2-39b shows how x m , x p (the point at which
the peak irradiance occurs) and xi vary with W 3 . Several typical values of x m , I p , and
< x > and the corresponding irradiances Im , I p , and Ic are noted in Table 2-14. The
aberrated central irradiance I(0) is also given in this table. The irradiance values I(0) and
I m are the Strehl ratios for primary and balanced primary coma, respectively. We note
that the peak of the PSF lies approximately at x m , the point corresponding to
W1 = (2 3)W3 only when W3 < ~ 0.7l . For larger values of W3 , the peak occurs closer to
1.0
0.8
10
8
0.6
l
6
<x>
x
0.4
xm
4
0.2 xp
lp
2
lm
lc
0.0 0
0 1 2 3 4 5 0 1 2 3 4 5
W3 W3
(a) (b)
Figure 2-39. (a) Variation of Im , I p , and Ic with W3 . (b) Variation of x m , x p , and

< x > with W3 .
Table 2-14. Typical values of xn , x p , and < x > and corresponding irradiances Im ,
I p , and Ic for circular pupils aberrated by primary coma.
W3 xm xp <x> Im Ip Ic I(0)
0 0 0 0 1 1 1 1
0.5 0.67 0.66 1.00 0.8712 0.8712 0.6535 0.3175
1.0 1.33 1.30 2.00 0.5708 0.5717 0.1445 0.0791
1.5 2.00 1.80 3.00 0.2715 0.2844 0.0004 0.0618
2.0 2.67 1.57 4.00 0.0864 0.1978 0.0061 0.0341
the origin than the point corresponding to minimum aberration variance. For W3 > ~ 1.6l ,
the distance of the peak from the origin does not increase monotonically, but fluctuates as
W3 increases. Since, according to Eq. (2-165), the distance of the centroid increases
linearly with W3 , it is clear that the separation between the locations of the centroid and
the peak increases as W3 increases.
2.9.2.3 Secondary Coma

The secondary coma aberration is given by
W (r, q) = W5r5 cos q . (2-168)
Its standard deviation is given by s W = A5 2 3 . Following Eq. (2-157), we find that the
aberrated PSF along the x axis is given by Eq. (2-162) where
( )
B = 2t 2 W5 - x t 1 2 . (2-169)
Figure 2-40a shows how I(x) varies with x for several typical values of W5 . According to
Eq. (2-158), the centroid of the PSF is given by
< x > = 2W5 . (2-170)
From Eqs. (2-165) and (2-170), we note that the primary and secondary coma with the
same peak value yield the same centroid < x > , although the corresponding aberrated
PSFs are different. This may be seen from Figure 2-41, where the PSFs for W3 = 5 = W5
are shown.
The standard deviation of the aberration given by Eq. (2-168) is reduced by a factor
of 2 if an aberration W1r cos q , where W1 = - 0.5 W5 , is introduced. Accordingly, for
small values of W5 , the peak of the corresponding aberrated PSF occurs at x m = W5 ,
although its centroid lies at < x > = 2W5 . The values of x m , x p , and < x > and the
corresponding irradiances Im , I p , and Ic for the values of W5 considered in Figure 2-40
are noted in Table 2-15.
1.0
W5 = 0
0.8
0.5
0.6
l(x)
0.4 1
1.5
0.2
2
0.0
–3 –2 –1 0 1 2 3
x
1.0
W5 = 0
1
W3 = – 1.2 W5
xm = – 0.6 W5 2
0.8
0.6
l(x)
0.4
0.2
0.0
–3 –2 –1 0 1 2 3
x – xm
Figure 2-40. (a) PSF for several typical values of secondary coma W5 in units of
l .(b) PSF for combined primary and secondary coma aberration given by Eq. (2-
172). Note that the horizontal coordinate is x - x m .
(a) (b)
Figure 2-41. PSF aberrated by (a) primary coma with W3 = 5l and (b) secondary
coma with W5 = 5l .
Table 2-15. Typical values of x m , x p , and < x > and corresponding irradiances Im ,
I p , and Ic for circular pupils aberrated by secondary coma given by Eq. (2-168).
0 0 0 0 1 1 1 1
0.5 0.50 0.49 1.00 0.8150 0.8153 0.4114 0.4955
1.0 1.00 0.83 2.00 0.4464 0.4664 0.0025 0.2332
1.5 1.50 0.81 3.00 0.1685 0.3237 0.0098 0.1873
2.0 2.00 1.11 4.00 0.0420 0.2523 0.0073 0.1389
The variance of the aberration r5 cos q is reduced even further if an appropriate

amount of r3 cos q aberration is also introduced. For a given value of W5 , the
appropriate amounts of W3 and W1 that give minimum variance may be obtained from
the radial Zernike polynomial R51 (r) , where
R51 (r) = 10r5 - 12r3 + 3r . (2-171)
The standard deviation of the aberration given by Eq. (2-168) is reduced by a factor of 10
if we introduce r cos q and r3 cos q aberrations with W1 = 0.3 W5 and W3 = - 1.2 W5 .
Hence, the peak value of a PSF aberrated by a small value of W5 and W3 = - 1.2 W5
occurs at x m = - 0.6 W5 . According to Eq. (2-158), the corresponding centroid occurs at
< x > = - 0.4W5 . Therefore, the separation between the peak and the centroid is 0.2 W5 .
For large values W5 , the minimization of variance with respect to W3 and W1 does not
lead to a maximum of the PSF.
As an example, we consider the PSF aberrated by an aberration
(
W (r, q) = W5r5 + W3r3 cos q ,) (2-172)
where
W3 = - 1.2W5 . (2-173)
According to Eq. (2-171), the point in the image plane with respect to which the
aberration variance is minimum is given by
x m = - 0.6W5 . (2-174)
The corresponding centroid is given by
< x > = - 0.4W5 . (2-175)
The aberrated PSF along the x axis is given by Eq. (2-162), where
(
B = 2t 2W5 + 2t W3 - x t 1 2 ) . (2-176)
Figure 2-41 shows the aberrated PSF I(x) for several values of W5 with W3 given by Eq.
(2-173). The values of x m , x p , and < x > and the corresponding irradiances Im , I p , and
Ic are given in Table 2-16. Note that x m , x p , and < x > are all negative. Moreover, their
magnitude for the values of W5 considered is very large. Therefore, in Figure 2-41, the
horizontal coordinate is chosen to be x - x m .
2.9.3 Comments
The results given here are applicable to both imaging systems, e.g., those used for
optical surveillance, as well as to laser transmitters used for active illumination of a
target. In both cases, the LOS of the optical system is extremely important. A LOS error
of a surveillance system will produce an error in the location of the target. In the case of a
Table 2-16. Typical values of x m , x p , and < x > and corresponding irradiances Im ,
I p , and Ic for primary and secondary coma given by Eq. (2-172).
0 0 0 0 1 1 1 1
1.0 – 0.60 – 0.59 – 0.40 0.9676 0.9682 0.8763 0.3721
2.0 – 1.20 – 1.18 – 0.80 0.8765 0.8784 0.5870 0.0030
3.0 – 1.80 – 1.77 – 1.20 0.7429 0.7459 0.2981 0.0014
4.0 – 2.40 – 2.37 – 1.60 0.5886 0.5914 0.1173 0.0465
2.10 Diffraction OTF for Primary Aberrations 169
laser transmitter, a large LOS error may cause the laser beam to miss the target
altogether. Whereas for static aberrations we may be able to calibrate the LOS, for
dynamic aberrations it is the analysis given here that will determine the tolerances on
aberrations of the type rn cos q and rn sin q .
Although we have defined the LOS of an optical system in terms of the centroid of
its PSF, it could have been defined in terms of the peak of the PSF (assuming that the
aberrations are small enough so that the PSF has a distinguishable peak). For an
aberration-free PSF, its peak value and its centroid both lie at its origin, regardless of the
amplitude variations across its pupil. The two are not coincident when aberrations are
present. The precise definition of the LOS will perhaps depend on the nature of the
application of the optical system. Moreover, in practice, only a finite central portion of
the PSF will be sampled to measure its centroid, and the precision of this measurement
will be limited by the noise characteristics of the photo-detector array used for such a
measurement.
2.10 DIFFRACTION OTF FOR PRIMARY ABERRATIONS19-25
Now we discuss the full OTFs for primary aberrations. The MTFs, PTFs, and the real
and imaginary parts of the OTFs are considered for various values of a primary
aberration. The phenomenon of contrast reversal is pointed out and discussed.

Assuming a uniformly illuminated pupil, Eq. (1-125) for the OTF in a rotated ( p, q )
coordinate system (see Figure 2-42) may be written
r -1 ÛÛ r
t(vi ) = Sex [ ]
ÙÙ exp iQ( p, q; vi ) dp dq
ıı
(2-177a)
-1 È ÛÛ ÛÛ ˘
= Sex Í ÙÙ cos Q dp dq + i ÙÙ sin Q dp dq ˙ , (2-177b)
Î ıı ıı ˚
where
r
Q ( p, q; vi ) = F ( p + l Rvi 2, q ) - F ( p - l Rvi 2, q ) (2-178)
is the phase aberration difference function and the integration is carried over the overlap
region of two pupils centered at ( m l Rvi 2, 0) . The aberration function F ( p, q ) in the
( )
rotated coordinate system is obtained from the nominal function F x p , y p by replacing
x p with p cos f - q sin f and y p with p sin f + q cos f . For a rotationally symmetric
system, there is no loss of generality if we assume that the point object for which the
aberration function is under consideration lies along an axis parallel to the x p axis. The
aberration function in that case depends on terms with integral powers of x p and
x 2p + y 2p . Hence, the aberration function F ( p, q ) can be obtained from the function
( )
F x p , y p by replacing x 2p + y 2p by p 2 + q 2 and x p by p cos f - q sin f .
q
p
yp
xp
(0,0)
ni
lR
Figure 2-42. Geometry for evaluating the OTF. The centers of the two pupils are
( )
located at (0, 0) and l R (x, h) in the x p , y p coordinate system and m (l R 2) (vi , 0)
12
in the ( p, q ) coordinate system, where vi = x 2 + h2 ( )
and f = tan -1 ( h x) . The
shaded area is the overlap area of the two pupils. When normalized by the pupil
radius a, the centers of the two pupils of unity radius lie at m v along the p axis.
Let the primary aberrations be given by
(
Ïbd x p 2 + y p 2 ,
Ô
) Defocus
Ô 2 2
( 2
Ôas x p + y p
F( x p , y p ) = Ì
) (
+ bd x p 2 + y p 2 , Spherical + Defocus) (2-179)
(
Ôaa x p 2 + bd x p 2 + y p 2 ,
Ô
)
Astigmatism + Defocus
(
Ôac x p x p 2 + y p 2 .
Ó ) Coma
Letting
(x p , yp ) = ar(cos q, sin q) , (2-180)
they may be written
Ï Bd r2 , Defocus
Ô
Ô As r 4 + Bd r2 , Spherical + Defocus
F(r, q) = Ì (2-181)
2 2 2
Ô Aa r cos q + Bd r , Astigmatism + Defocus
Ô 3
Ó Ac r cos q , Coma
where the two types of aberration coefficients are related to each other according to
Bd = a 2 bd , (2-182a)
As = a 4 as , (2-182b)
Aa = a 2 aa , (2-182c)
and
Ac = a 3 ac . (2-182d)
Writing the aberration function in the ( p, q ) coordinate system by replacing

x p + y p 2 by p 2 + q 2 and x p by p cos f - q sin f , and substituting into Eq. (2-178), we
2
obtain the aberration difference function Q for the primary aberrations:
Ï4 Bd pv Defocus
Ô
r ÔÔ s ( )
8 A pv p 2 + q 2 + v 2 + 4 Bd pv Spherical + defocus
Q ( p, q; v ) = Ì (2-183)
( 2
)
Ô4 Aa v p cos f - q sin f cos f + 4 Bd pv Astigmatism + defocus
Ô
( )
ÔÓ Ac v 6 p 2 + 2 q 2 + 2 v 2 cos f - 4 Ac pqv sin f , Coma
where ( p, q ) are now normalized by the pupil radius so that p 2 + q 2 £ 1 and the spatial
frequecy v is normalized by the cutoff frequency 1 l F so that v £ 1. In the normalized
coordinates, the OTF given by Eq. (2-177a) becomes
r 1 ÛÛ r
t( v ) = ÙÙ
p ıı
[ ]
exp iQ( p, q; v ) dp dq . (2-184)
We note from Figure 2-42 that in the ( p, q ) coordinate system, the two unit circles
are centered at mv along the p axis and the limits of p and q are m ÊË 1 - q 2 - vˆ¯ and
m 1 - v 2 , respectively. Their overlap region, which forms the region of integration in the
above integrals, is symmetric in p and q. Hence, if Q is an odd function of p and/or q, the
imaginary part of the integral vanishes. This is true for defocus, spherical aberration, and
astigmatism. It is also true for coma for a spatial frequency parallel to the y p axis, i.e., for
f = p 2 . In such cases the OTF is real and, depending on whether its value for a certain
spatial frequency is positive or negative, its phase, the PTF, for that frequency is zero or
p, respectively. A phase of p is sometimes associated with a negative value of the MTF.
It represents contrast reversal, i.e., for example, bright regions in the object appear as
dark regions in the image. Similarly, the dark regions of the object appear as bright
regions in the image.
That the OTF is real for a symmetric aberration can be seen even more easily by
virtue of the corresponding symmetry of the PSF, namely that
PSF( - x, - y) = PSF( x, y) , (2-185)
as in the case of astigmatism or spherical aberration. By definition, the OTF is given by

the Fourier transform of the PSF; i.e.,
• •
[
t (x, h) = Ú Ú PSF( x, y) exp 2 pi(x x + hy) dx dy
-• -•
]
• •
[
= Ú Ú PSF( - x, - y) exp 2pi(x x + hy) dx dy
-• -•
]
• •
[ ]
= Ú Ú PSF( x ¢, y ¢) exp - 2pi(x x ¢ + hy ¢) dx ¢ dy ¢
-• -•
= t * (x, h) . (2-186)
Hence, t (x, h) is real for a symmetric aberration.
2.10.2 Defocus
Figure 2-43 shows how the OTF of a defocused system varies with the spatial
frequency. We note that it is real and radially symmetric; i.e., its value depends on the
value of v but not on the value of f. For Bd <~ 0.64l , the OTF is positive for all spatial
frequencies. However, for larger values of Bd it becomes negative, corresponding to a
PTF of p, for certain bands of spatial frequencies. It becomes negative for smaller and
smaller spatial frequencies as the amount of defocus Bd increases.
To illustrate the significance of the OTF and, in particular, the contrast reversal, we
consider, as shown in Figure 2-44a, a 2-D object that is sinusoidal along the vertical axis
with a spatial frequency that increases linearly in the horizontal direction. The maximum
frequency in the object is chosen to equal the cutoff frequency of the aberration-free
system. This frequency is normalized to unity. The aberration-free or the diffraction-
1.00
Defocus
Bd = 0 W(r) = Bdr2
0.75
0.50 0.25
t
0.5
0.25
2 1 0.64
3
0.00
– 0.25
0.00 0.25 0.50 0.75 1.00
n
Figure 2-43. OTFs of a defocused system. Bd represents the peak defocus

aberration in units of l .
(a)
(b)
(c)
1.0
0.8
0.6
Bd = 0
t
0.4 Bd = 2l
0.2
0.0
– 0.2
0.0 0.2 0.4 0.6 0.8 1.0
n
(d)
Figure 2-44. Aberration-free and defocused images of an object. (a) Object. It is

sinusoidal along the vertical axis with a spatial frequency that increases linearly in
the horizontal direction. The Gaussian image is identically the same, except for any
magnification. (b) Aberration-free image. (c) Defocused image with Bd = 2 l . (d)
Aberration-free and defocused OTFs.
limited image of the object is shown in Figure 2-44b. The monotonic reduction in
contrast with increasing spatial frequency is quite evident from this figure. A defocused
image corresponding to Bd = 2l is shown in Figure 2-44c. It is clear that the contrast in
the image reduces with frequency rapidly to zero, reverses its sign back and forth as the
frequency increases, with practically zero values for frequencies v ≥ 0.3 . As a
convenience, the aberration-free and defocused OTFs are shown in Figure 2-44d to
illustrate the regions of zero and near-zero contrast as well as the regions of contrast
reversal.
The OTF is independent of the sign of Bd . Hence the OTFs corresponding to two
image planes symmetrically defocused about the Gaussian image plane are identical
when the Fresnel number of the system is large. However, when the Fresnel number is
small, the two symmetrical planes correspond to different magnitudes of the defocus
aberration and, therefore, yield different OTFs. Moreover, the cutoff frequency in that
case is D l z instead of D l R , as may be seen by replacing R with z in Eq. (1-73b).
2.10.3 Spherical Aberration

Figure 2-45 shows how the OTF of a system aberrated by spherical aberration As r 4
varies with spatial frequency. As in the case of defocus, it is real and radially symmetric,
and independent of the sign of As when Bd = 0 or when the sign of Bd is also changed.
Figure 2-45a shows the OTF for a defocused image plane Bd = - As , corresponding to
minimum aberration variance across the pupil, for various values of As . We note that the
OTF is positive for all spatial frequencies for As £ 2.2 l . For larger values of As , it
becomes negative for certain bands of frequencies. Figure 2-45b shows the OTFs for
As = 1l in various image planes. As noted in Section 2.3.3, Bd As = 0 , – 1, – 1.5, and –
2 correspond to images observed in the Gaussian, minimum-aberration-variance, circle-
of-least-confusion, and marginal image planes. We note that, except at very high spatial
frequencies, the OTF values in the minimum-aberration-variance plane are higher than
the corresponding values in other planes.
2.10.4 Astigmatism
The OTF aberrated by astigmatism Aa r2 cos 2 q is real and independent of the sign
of Aa when Bd = 0 or when the sign of Bd is also changed. Figure 2-46 shows it for a
defocused image plane Bd = - Aa 2, corresponding to minimum aberration variance
across the pupil, for various values of Aa . The OTF in this case is four-fold symmetric,
with one axis lying in the tangential plane. We note from the figure that it becomes
negative for smaller values of Aa when f = p 4, compared to when f = 0 . For example,
the OTF for f = 0 is positive for Aa £ 1.28l . However, it is negative for f = p 4, when
Aa = 1 l and v > 0.35 . Figure 2-47 shows the OTF for Aa = 1l in various image planes.
As discussed in Section 4.3.3 of Part I, Bd Aa = 0, – 1/2, and – 1 correspond to image
planes in which the tangential focal line, circle-of-least (astigmatic)-confusion, and
sagittal focal line, respectively, are observed. The OTF has a biaxial (or inversion)
symmetry, except for observations in the plane of least confusion, in which case it has a
1.00
Balanced spherical
As = 0 W(r) = As(r4 – r2)
0.75
0.25
0.5
0.50
1
t
(a)
0.25 2
2.2
3
0.00
– 0.25
0.00 0.25 0.50 0.75 1.00
n
1.00
Spherical and defocus

W(r) = Asr4 + Bdr2
0.75
As = 1
Bd = – 1
0.50
t
– 1.5 (b)
0.25
0
0.00
–2
– 0.25
0.00 0.25 0.50 0.75 1.00
n
Figure 2-45. OTFs for spherical aberration. (a) Various values of As and image
planes Bd = - As corresponding to minimum aberration variance. (b) Fixed As and
various image planes. As and Bd are in units of l .
four-fold symmetry, as stated earlier. The PTF is either 0 or p depending on the spatial
frequency. For f = 0 , the OTF in the image plane Bd = - Aa is the same as the
aberration-free OTF, as may be seen from the third of Eqs. (2-183). Similarly, the OTF
for Bd = 0 is the same as the defocused OTF in Figure 2-43 for Bd = 1 l . For f = p 4,
the OTFs in the image planes orresponding to Bd Aa = 0 and – 1 are equal. The OTF for
f = p 2 is independent of astigmatism, as may be seen from Eq. (2-183).
1.00
Balanced astigmatism
Aa = 0 W(r,q) = Aa(r2cos2q – r2/2)
0.75
f= 0
0.25
0.50
t
0.5
(a)
0.25 1
1.28
2
3
0.00
– 0.25
0.00 0.25 0.50 0.75 1.00
n
1.00
W(r,q) = Aa(r2cos2q – r2/2)
0.75
f = p/4
Aa = 0
0.50
t
0.25
0.5 (b)
0.25
1
2
3
0.00
– 0.25
0.00 0.25 0.50 0.75 1.00
n
Figure 2-46. OTFs for various amounts of astigmatism in defocused image planes
corresponding to minimum-aberration variance. As indicated in Figure 1-7, f
represents the angle the spatial frequency vector makes with the x axis. Aa
represents the peak value of astigmatism aberration in units of l and a defocused
image plane is represented by Bd = - Aa 2 . (a) f = 0 and (b) f = p 4 .
2.10.5 Coma
The OTF aberrated by coma Ac r3 cos q is a complex function, except for f = p 2,
i.e., when the radiance of the sinusoidal object varies along a direction that is normal to
the axis about which the PSF is symmetric. A conjugate function is obtained when the
sign of Ac is changed. The OTF dependence on the spatial frequency is shown in terms of
its MTF and PTF in Figures 2-48 and 2-49, respectively. Figure 2-48 shows the MTF for
several values Ac when f = 0 , p 4 , and p 2. The corresponding PTF is shown in Figure
1.00

0.75
Bd = – 1 Aa = 1
f= 0
0.50
– 0.5
t
(a)
0.25
0
0.00
– 0.25
0.00 0.25 0.50 0.75 1.00
n
1.00

0.75
Aa = 1
f = p/4
Bd = – 0.5
0.50
t
(b)
0.25
0, – 1
0.00
– 0.25
0.00 0.25 0.50 0.75 1.00
n
Figure 2-47. OTFs for astigmatism with Aa = 1 l in image planes Bd Aa = 0, - 1 2 ,

and - 1 , corresponding to sagittal-line, minimum-aberration-variance, and
tangential-line image planes, respectively. (a) f = 0 and (b) f = p 4 .
2-49, except when f = p 2, in which case the PTF is zero. The value of Y in this figure
is in radians. Note that if a wavefront tilt is added to the coma aberration to minimize its
variance across the pupil, as discussed in Section 2.3.3, it only introduces a PTF varying
linearly with the spatial frequency. The real and imaginary parts Ωt(v)Ωcos Y (v) and
1.00
Coma
W(r) = Acr3 cosq
Ac = 0
0.75 f= 0
ÔtÔ 0.25
0.50 0.5 (a)
0.25 2
3
0.00
0.00 0.25 0.50 0.75 1.00
n
1.00
Coma
W(r) = Acr3 cosq
Ac = 0
0.75 f = p/4
0.25
ÔtÔ
0.50 0.5 (b)
1
2
0.25
3
0.00
0.00 0.25 0.50 0.75 1.00
n

r rq
f p

t

n
Figure 2-48. MTFs for coma. A c represents the peak value of coma aberration in
units of l . (a) f = 0 , (b) f = p 4, and (c) f = p 2 for which the OTF is real.
Ωt(v)Ωsin Y (v) , respectively, of the OTF are shown in Figures 2-50 and 2-51 for f = 0
and p/4, respectively. As discussed in Section 1.6.4, the slope of the real part at the origin
is independent of the aberration and is equal to that of the MTF. Moreover, the slope of
the imaginary part at the origin, which yields the centroid of the PSF (see Section 1.8.1),
increases as the aberration increases. Finally, as discussed in Section 1.9.3, the Strehl
ratio is obtained by integrating the real part, and the integral of the imaginary part is zero.
40
Coma
W(r) = Acr3 cosq
30 f= 0
Ac = 3
Y
20 (a)
2
10 1
0.5
0.25
0
0.00 0.25 0.50 0.75 1.00
n
30
Coma
W(r) = Acr3 cosq
f = p/4
20
Ac = 3
Y
(b)
2
10
1
0.5
0.25
0
0.00 0.25 0.50 0.75 1.00
n
Figure 2-49. PTFs for coma. A c represents the peak value of coma aberration in
units of l . (a) f = 0 and (b) f = p 4.
1.00
Coma
W(r) = Acr3 cosq
0.75
f= 0
Ac = 0
0.50
Re t
0.25
(a)
0.5
0.25
1
2
3
0.00
– 0.25
0.00 0.25 0.50 0.75 1.00
n
1.00
Coma
W(r) = Acr3 cosq
0.75
f = p/4
0.50 Ac = 0
Re t
0.25 (b)
0.5
0.25 1
3 2
0.00
– 0.25
0.00 0.25 0.50 0.75 1.00
n
Figure 2-50. Real part Rett of the OTF of a system aberrated by coma. (a) f = 0 and
(b) f = p 4.
1.00
Coma
3 W(r) = Acr3 cosq
0.75
2 f= 0
1
0.50 0.5
Im t
(a)
0.25 Ac = 0.25
0.00
–0.25
0.00 0.25 0.50 0.75 1.00
n
1.00
Coma
W(r) = Acr3 cosq
0.75
3 f = p /4
2
0.50 1
Im t
0.5 (b)
0.25
Ac = 0.25
0.00
–0.25
0.00 0.25 0.50 0.75 1.00
n
Figure 2-51. Imaginary part Imtt of the OTF of a system aberrated by coma.
(a) f = 0 and (b) f = p 4.
2.11 HOPKINS RATIO

r
As defined in Section 1.10, the Hopkins ratio H (v ) of an imaging system is the ratio
r
of its MTFs at a spatial frequency v with and without aberrations. In this section, we first
give the tolerance for a primary aberration for a Hopkins ratio of 0.8. Next, a power series
expression is obtained for the defocused OTF from which an approximate expression for
the Hopkins ratio is derived and defocus tolerance for a ratio of 0.8 is obtained. Finally,
expressions for the aberration difference function and its variance for primary aberrations
are given, which can be used to obtain the aberration tolerance for a Hopkins ratio as
small as 0.4 with a reasonable precision.
2.11.1 Tolerance for Primary Aberrations

Based on numerical analysis for primary aberrations, Hopkins26 has shown that
H (v) >
~ 0.8 for v <
~ 0.1 , provided their coefficients obey the following conditions:
Bd <
~ ± l 20 v , (2-187)
Aa <
~ ± l 10 v in the plane Bd = - Aa 2 , (2-188)
Ac < Ê 0.071 + 0.16ˆ with Y v = m 0.89 + 0.48v when f = 0 ,

~ ± lË ¯
( ) (2-189)
v
< Ê 0.123 + 0.19ˆ with Y v = 0 when f = p 2 ,

~ ± lË ¯
( ) (2-190)
v
and
Ê 0.106 + 0.33ˆ in the plane B = - 1.33 - 2.2 v + 2.8v 2 A

As <
~ ± lË
v ¯ d ( ) s , (2-191)
where Y(v) is the PTF at a spatial frequency v. As in Table 2-4 Bd , Aa , Ac and As

represent the peak coefficients of defocus, astigmatism, coma, and spherical aberration,
respectively. We note that the amount of balancing defocus in the case of spherical
aberration is different from its corresponding value for optimizing the Strehl ratio.
Moreover, it depends on the magnitude of the spatial frequency at which the MTF is
optimized. The balancing defocus Bd = - ( 4 3) As in the limit v Æ 0 is the same as that
obtained from the geometrical OTF for low spatial frequencies [see the statement
following Eq.(2-230)]. An approximate expression for the Hopkins ratio for defocus is
derived next, from which the defocus tolerance given by Eq. (2-187) is obtained.
2.11.2 Defocus
Equation (2-187) for defocus tolerance can be obtained as follows. Substituting the
aberration difference function Q( p, q; v) = 4 Bd pv for the defocus phase aberration
( )
Bd p 2 + q 2 into Eq. (2-184), and noting that the limits of integration for p and q are
m ÊË 1 - q 2 - vˆ¯ and m 1 - v 2 , respectively, we obtain
1- v 2 1- q 2 - v
1 Û Û
t (v) = Ù dq Ù exp( 4iBd pv) dp
p ı ı
- 1- v 2 Ê ˆ
- Á 1- q 2 - v˜
Ë ¯
1- v 2
1 Û
= sin È4 Bd v ÊË 1 - q 2 - vˆ¯ ˘ dq . (2-192a)
p Bd v Ù
ı ÍÎ ˙˚
0
It is evident that t (v) is independent of the sign of Bd , i.e., it is symmetric about the
Gaussian image plane.
The OTF may be written in a different form by first integrating over q and then over
p. The limits of integration for p and q in this case are different. They may be obtained
from Figure 2-42 by noting that the centers of the two unit circles lie at mv along the p
axis. Thus, the OTF may be written
1- v 1- ( p + v ) 2
1 Û Û
t (v) = dp exp ( 4iBd pv) dq
p Ù
Ù
ı ı
- (1- v )
- 1- ( p + v ) 2
1- v
4 Û 2
= 1 - ( p + v) cos ( 4 Bd pv) dp
p Ù
ı
0
1
4Û
= Ù 1 - s 2 cos [ 4 Bd v(s - v)] ds , (2-192b)
pı
v
where we have kept only the real part since the OTF is real and we have let p + v = s in
the last step.
Letting q = sin a , where 0 £ a £ b with cosb = v (see Figure 2-5), Eq. (2-192a)
yields the aberration-free result of Eq. (2-44b) if we let Bd Æ 0 . Expanding the sine
function in a power series, we may write
1- v 2
4 Û ÈÊ 8 32
t (v) = Ù ÍË 1 - q 2 - vˆ¯ - Bd2 v 2 1 - q 2 ( ) + ...˘˙ dq
p ı Î 3 ˚
0
Ê b ˆ
1Á 8 2 2Û
= 2 - sin 2 - Bd v Ù cos d + ...˜
b b 4
a a
pÁ 3 ı ˜
Ë 0 ¯
1
=
p
(
2b - sin 2b - 4 Bd2 v 2b + ... )
2 È -1 12
=
p ÎÍ
(
cos v - v 1 - v 2 ) - 2 Bd2 v 2 cos -1 v + ...˘ .
˚˙
(2-193)
The first two terms in Eq. (2-193) correspond to the aberration-free OTF.
Noting that
p
cos -1 v =
2
- v + O v3 ( ) , (2-194)
we obtain
4v
t (v) = 1 -
p
( )
- 2 Bd2 v 2 + O v 3 . (2-195)
As expected, the slope of the defocused OTF at the origin is - 4 p , same as that of the
aberration-free OTF. The Hopkins ratio for small values of v may be written
t (v; Bd )
H (v) =
t (v; Bd = 0)
~ 1 - 2 Bd2 v 2 . (2-196)
For small values of v, the Hopkins ratio based on the geometrical OTF is also given by
Eq. (2-196) [see Eq. (2-228)]. Of course, since the aberration-free geometrical PSF is a
point (or a Dirac delta function), the corresponding OTF is a constant independent of the
spatial frequency. Hence, the geometrical Hopkins ratio for a certain spatial frequency is
simply the geometrical OTF for that frequency.
The defocus tolerance for a Hopkins ratio of 0.8 is given by
Bd = 0.1 v rad
or
Bd = l 20 v
= 1 20 Fvi . (2-197)
The corresponding tolerance for the longitudinal defocus is given by
Dz = 8 lF 2 20 v
= 0.4 F v i , (2-198)
which, like the aberration coefficient Bd , is independent of the wavelength.

2.11.3 Hopkins Ratio in Terms of Variance of Aberration Difference Function

For small values of the variance s Q2 of the aberration difference function, the
Hopkins ratio is approximately given by [see Eq. (1-214)]
H (v) ~ 1 - 1 s Q2 . (2-199)
2
where
2
s Q2 = Q 2 - Q . (2-200)
A mean value in Eq. (2-200) is given by

r
n ÚÚ Q n ( p, q; v ) dp dq
Q =
ÚÚ dp dq
r
ÚÚ Q n ( p, q; v ) dp dq
= , (2-201)
2b - sin 2b
where cosb = v . The corresponding PTF is given by

r
Y( v ) = Q . (2-202)
We noted in Section 2.3 that a maximum possible value of the Strehl ratio of a
system for a particular aberration is obtained by a minimization of its variance across the
exit pupil, provided the Strehl ratio thus obtained is not too small. The approximate Eq.
(2-199) shows that the Hopkins ratio of a system at a particular spatial frequency is
maximum when the variance of the aberration difference function across the overlap
region of two displaced exit pupils is minimum, the displacement of the pupils being
dependent on the spatial frequency under consideration. An investigation27 of balancing
secondary spherical aberration with primary spherical aberration and defocus, and
r
secondary coma with primary coma, shows that unless H (v ) > ~ 0.r7 , a decrease in the
value of s Q may be associated with a decrease in the value of H (v ) for v £ 0.1. Thus,
the variance of the aberration or that of the aberration difference function is a useful
criterion of image quality, provided the corresponding Strehl or Hopkins ratio obtained is
relatively high. For spatial frequencies v > 0.1, it is more appropriate to use the Strehl
ratio as the criterion of image quality and aberration balancing.
We also noted in Section 2.3 that the Strehl ratio of an aberrated system depends
approximately on the variance of the aberration function across its pupil and not on the
type of the aberration. Similarly, we note from Eq. (1-214) that the Hopkins ratio at a
certain spatial frequency depends approximately on the variance s Q2 of the aberration
difference function across the overlap region of two displaced pupils (displacement
depending on the spatial frequency) and not on the type of the aberration. As in the case
of Strehl ratio, a better approximation to the Hopkins ratio is obtained by using the
exponential relation
H (v) ~ exp Ê - 1 s Q2 ˆ . (2-203)

Ë 2 ¯
r
Szapiel27 has shown that for primary aberrations, when the true value of H (vi ) exceeds
0.4, the approximate relation of Eq. (2-203) estimates it to within 0.06.
2.11.4 Variance of Aberration Difference Function for Primary Aberrations

Substituting for Q for a primary aberration from Eq. (2-183) into Eq. (2-200), we can
obtain s Q2 . We illustrate this by considering defocus (for which Qd = 0):
1- v 2 1- q 2 - v
16 Bd2 v 2 Û Û
s 2d (v) = Ù dq Ù p 2 dp
2b - sin 2b ı ı
- 1- v 2 - 1- q 2 + v
1- v 2
64 Bd2 v 2 Û Ê 2 ˆ
3
= Ù Ë 1 - q - v¯ dq
3(2b - sin 2b) ı
0
1
4v 2 ÈÍ 1 + 4v 2 2b - 13 + 2 v 2 sin 2b˘˙
( ) ( )
Î 3 ˚
= . (2-204)
2b - sin 2b
The variance of the aberration difference function for other aberrations can be
calculated in a similar manner. For the primary aberration function
F(r, q) = Bd r2 + As r 4 + Aa r2 cos 2 q + Ac r3 cos q , (2-205)
it is given by28
{ ( )
s Q2 (v) = 4v 2 4 P1 Bd2 + 4 Aa2 P6 cos 2 f + P1 sin 2 f sin 2 f + 8 P1 Bd Aa sin 2 f
[ (
+ Ac2 4 P3 cos 2 f + 4 P4 - P52 sin 2 f )]
+ 16 P2 As2 + 16 P7 Bd As + 128 P4 Aa As } , (2-206)
where
am 2b - bm sin 2b
Pm = , (2-207)
2b - sin 2b
1
a1 = 1 + 4v 2 4 , b1 = Ê13 + 2 ˆ 12 ,
( ) (2-208)
Ë 8v ¯
2.12 Geometrical OTF 187
1 17 2 29 139 2 292 4 28 6
a2 = + v + 8v 4 + 8v 6 , b2 = + v + v + v , (2-209)
8 6 24 10 45 45
1 1 2 9 7 2 1 4
a3 = + v , b3 = + v - v , (2-210)
24 4 40 90 90
3 17 2 101 83 2 17 4
a4 = + v + 4v 4 , b4 = + v - v , (2-211)
8 4 40 15 30
(
a5 = 1 + 4v 2 , b5 = 11 + 4v 2 3 , ) (2-212)
5 1 2
a6 = 1 4 , b6 = - v , (2-213)
12 6
and
1 7 25 2 2 4
a7 = 1 + 2 v 2 + 2 v 4 , b7 = + v + v .
6 6 9 9 (2-214)
2.12 GEOMETRICAL OTF
A comparison of the diffraction and geometrical OTFs was given in Section 1.6.7.
Since the aberration-free image of a point object in geometrical optics is a point, or a
Dirac delta function, the corresponding OTF is unity for all spatial frequencies. There is
no cutoff frequency as there is in the diffraction OTF. Of course, it would be unreal to
consider the geometrical OTF beyond the diffraction cutoff frequency. Whereas the
diffraction OTF depends strongly on the wavelength, the geometrical OTF is independent
of it, except for any variation due to chromatic aberration.
In this section, we consider the aberrated geometrical OTF. It was shown in Section
1.6.6 that the slope of the geometrical MTF at the origin is zero, unlike that of the
diffraction MTF. For low spatial frequencies, a simple but approximate expression is
given. It is shown that the MTF averaged over all angular orientations of a spatial
frequency vector is maximum when the variance of the ray aberration, or the root mean
square (rms) radius of the PSF with respect to its centroid, is minimum. The OTF
reduction for a radially symmetric aberration is directly related to the rms radius of the
PSF. The general formulas derived here are applied to obtain the OTFs for primary
aberrations. A closed-form exact expression is obtained in the case of astigmatism and/or
defocus. Generally, the geometrical OTF agrees with the diffraction OTF for very low
spatial frequencies when the aberration is larger than about a wavelength.21-23, 29-32 This is
a region of practical interest since the MTF at high frequencies for such aberrations is
negligibly small. Gostick, for example, has shown that the geometrical MTF for very
small spatial frequencies obtained according to Eq. (1-150) in the case of secondary
spherical aberration or secondary coma gives an excellent approximation to the
diffraction MTF.31

For a uniformly illuminated pupil, the geometrical OTF given by Eq. (1-138b)
reduces to
(
t g (x, h) = 1 p a 2 ) Ú Ú exp [2pi (x x + hy)] dx p d yp , (2-215)
where ( x, y) are the ray aberrations (i.e., they are coordinates of a ray in the image plane
with respect to the Gaussian image point). Using polar coordinates for a pupil point; i.e.,
letting
(x p , yp ) = ar(cos q, sin q) , (2-216)
the ray aberrations are given by [see Eq. (4-6) of Part I]
Ê ∂W sinq ∂W ∂W cosq ∂W ˆ
( x, y) = 2 F Á cosq – , sinq + ˜ . (2-217)
Ë ∂r r ∂q ∂r r ∂q ¯
Writing the OTF in the form
t g (x, h) = t g (x, h) exp iYg (x, h) [ ] , (2-218)
the MTF and the PTF for small spatial frequencies are approximately given by [see Eqs.
(1-144)]
t g (x, h) ~ 1 - 2 p2 Ú Ú [x( x - xc ) + h( y - yc )]2 dx p d y p (2-219)

a
and
Yg (x, h) ~ 2 p(x xc + hyc ) , (2-220)
respectively. Since the aberration-free MTF is unity for all spatial frequencies, the
aberrated MTF given by Eq. (2-219) represents the geometrical Hopkins ratio.
Using polar coordinates for a spatial frequency,
(x, h) = vi (cos f, sin f) , (2-221)
Eq. (2-219) becomes

2
t g (vi , f) ~ 1 - 2 p2vi Ú Ú [( x - xc ) cos f + ( y - yc ) sin f]
2
dx p d y p . (2-222)
a
The MTF averaged over all values of the angle f is given by
2p
1 Û
t g (vi ) = t g (vi ) df
2p Ù
ı
0
2.12 Geometrical OTF 189
2
~ 1 - pv2i 2
Ú Ú [( x - xc ) + ( y - yc )] dx p d y p . (2-223)
a
Comparing Eqs. (2-222) and (2-223), we note that the average MTF is also equal to the
the MTF for f = ± 45o . 29 Equation (2-223) shows that, for small spatial frequencies, the
average MTF is maximum when the variance of the ray aberration (or the rms radius of
the geometrical PSF with respect to its centroid) is minimum. Balanced aberrations in
terms of linear combinations of Zernike polynomials yielding minimum rms radius are
given in Section 4.4 of Part I.29, 31 The centroid of the PSF for an aberration balanced in
this manner lies at the origin, i.e., xc = 0 = yc for such an aberration.
2.12.2 Radially Symmetric Aberration
For a radially symmetric aberration, xc = 0 = yc , and t g (x, h) is real, given by
t g (x, h) ~ 1 - 2 p2 Ú Ú (xx + hy)2 dx p d y p . (2-224)

a
Substituting for the ray aberrations
∂W
( x, y) = 2 F (cos q, sin q) , (2-225)
∂r
and using polar coordinates for a spatial frequency and a pupil point, we obtain
1 2
2ÛÊ ∂W ˆ
t g (vi ) ~ 1 - 8(pFvi ) Ù Á ˜ r dr (2-226a)
ı Ë ∂r ¯
0
2
= 1 - ( pvi rirms ) , (2-226b)
where rirms is the rms radius of the geometrical PSF. It is evident that the OTF for any
frequency is maximum when the rms spot radius is minimum.
2.12.3 Defocus
The defocus aberration is given by
W (r) = Bd r2 . (2-227)
Substituting into Eq. (2-226a), we obtain
t g (vi ) ~ 1 - 8(p FBd vi )2 . (2-228)
If we write the aberration coefficient Bd in radians by multiplying it by 2p l and

replace vi by v l F , Eq. (2-228) reduces to Eq. (2-196) for the diffraction Hopkins ratio.
2.12.4 Spherical Aberration Combined With Defocus

For spherical aberration combined with defocus,
W (r) = As r 4 + Bd r2 ; (2-229)
the rms spot radius is given by
8
2
rirms = 8 F 2 Ê 2 As2 + Bd2 + As Bd ˆ . (2-230)
Ë 3 ¯
Its minimum value is given by rirms = ( 4 3) FAs for Bd = - ( 4 3) As [This value of Bd is

the same as that for the diffraction Hopkins ratio in the limit v Æ 0 , as may be seen from
Eq. (2-191)]. Substituting the minimum value of rirms into Eq. (2-226b), we obtain the
optimum value of the OTF:
t g (vi ) ~ 1 - (4 p FAs vi 3)2 . (2-231)
2.12.5 Astigmatism Combined With Defocus

Consider astigmatism combined with defocus:
W (r, q) = Aa r2 cos 2 q + Bd r2 . (2-232)
In this case, a closed-form expression for the exact OTF can be obtained. Substituting for
the ray aberrations
[
( x, y) = 4 Fr ( Aa + Bd ) cos q, Bd sin q ] , (2-233)
and using polar coordinates for a spatial frequency, Eq. (2-215) yields
1 2p
1Û Û
[
t g (vi , f) = Ù Ù exp 8pFGivi r cos(q - y ) r dr dq
pı ı
]
0 0
1
Û
= Ù 2 J 0 (8pFGivi r ) r dr
ı
0
2 J1 (8pFGvi )
= , (2-234a)
8pFGvi
where
12
G = [( A + B ) cos f + B sin f]
a d
2 2 2
d
2
(2-234b)
and
2.13 Incoherent Line- and Edge-Spread Functions 191
Bd
tan y = tan f . (2-234c)
Aa + Bd
For small frequencies, noting that
x x3
J1 ( x ) = - + ... , (2-235)
2 16
we obtain
t g (vi , f) ~ 1 - 8 (pFGvi )2 , (2-236)
a result that can also be obtained by substituting Eq. (2-233) into Eq. (2-219) and noting
that xc = 0 = yc for the aberration under consideration. The smallest image spot is
obtained when Bd = - Aa 2, in which case it is circular with a uniform irradiance. In that
case, G reduces to Aa 2, y Æ - f , and t g becomes independent of f , as expected for a
uniform circular image. This is not true of the diffraction image, which is nonuniform and
results in an OTF that depends on f , as was seen in Figure 2-46.
2.12.6 Coma
Finally, we consider coma given by
W (r, q) = Ac r3 cos q . (2-237)
The ray aberrations and the PSF centroid are given by
( x, y) = 2 FAc r2 (2 + cos 2q, sin 2q) (2-238)
and
( xc , yc ) = (2 FAc , 0) . (2-239)
Substituting into Eqs. (2-219) and (2-220), we obtain
t g (vi , f) ~ 1 - 4 (pvi F Ac )2 (1 + 2 cos 2 f) , (2-240a)

3
and
Yg (vi , f) ~ 4 pFAc vi cos f . (2-240b)
The OTF given by Eqs. (2-240) can also be obtained by determining the ray aberration
along the p axis and substituting its mean value and variance into Eqs. (1-155) (see
Problem 8).
2.13 INCOHERENT LINE- AND EDGE-SPREAD FUNCTIONS33-35

The general relations for the incoherent line- and edge-spread functions were given
in Section 1.11. In this section, we give an expression for the LSF in terms of the pupil
function which, in turn, is used to obtain an approximate expression for the Struve ratio.
The balanced aberrations and their tolerance are given for a Struve ratio of 0.8. The
expressions for the LSF and ESF are also given in terms of the OTF. Plots of numerical
results for primary aberrations are provided.
2.13.1 Theory
2.13.1.1 LSF From PSF
The line-spread function (LSF) of an imaging system is, by defintion, the image of a
line object. For an object lying along the yo axis, it is given by [see Eq. (1-219)]
LSF( xi ) = Ú-•• PSF( xi , yi ) dyi . (2-241)
For a radially symmetric PSF, we may write

•
Û
LSF( xi ) = 2Ù PSF ÊË xi2 + yi2 ˆ¯ dyi ,
ı
0
12
(
or, substituting ri = xi2 + yi2 ) , we obtain
•
Û PSF ( ri )
LSF( xi ) = 2Ù 12 ri dri , (2-242)
ı
xi
(r i
2
- xi2 )
i.e., the LSF is the Abel transform36 of the PSF. Substituting Eq. (2-15), the aberration-
free LSF is given by
•
8 Û J12 ( pri )
LSF( xi ) = 2Ù dri . (2-243)
p ı r r2 - x2 1 2 ( )
i i xi i
2.13.1.2 LSF From Pupil Function33

The infinite limits over the PSF integration can be avoided by writing it in terms of
the pupil function, as was done in obtaining Eq. (1-221a). For a circular pupil of radius a,
the normalized LSF in terms of the pupil function may be written
2
a a 2 - y 2p
Û Û Ê 2 pi ˆ
Ù dy p
ı
Ù
ı
[ (
exp iF x p , y p exp Á -
Ë
)]
l R
xi x p ˜ dx p
¯
-a
- a 2 - y 2p
LSF ( xi ) = ,
a
È a 2 - y 2p ˘
Í ˙
Û Í Û ˙
Ù dy p Ù dx p
ı Í ı ˙
-a Í- a 2 - y 2p ˙
Î ˚
or
2
1 1- y n2
3 Û Û
LSF ( x ) =
16 Ù
ı
dyn Ù
ı
[ ]
exp iF( x n , yn ) exp ( - pi xx n ) dx n , (2-244)
-1 - 1- y n2
( )
where F is the phase aberration function, ( x n , yn ) = x p , y p a are the coordinates of a
pupil point normalized by the pupil radius, and x = xi l F is the distance of an image
point in units of l F . The LSF is normalized by its aberration-free central value.
For a line object along a yo¢ axis (or parallel to it) making an angle b with the yo
axis (see Figure 1-9), the LSF( xi¢) is obtained by writing the pupil function in the
( )
coordinate system x ¢p , y ¢p by replacing x p by x ¢p cos b - y ¢p sin b and y p by
x ¢p sin b + y ¢p cos b . For a rotationally symmetric imaging system, since the aberration
function for a point object along the x o axis consists of terms with integral powers of x p
and x 2p + y 2p , the aberration function in the rotated coordinate system is obtained by
replacing x p by x ¢p cos b - y ¢p sin b and x 2p + y 2p by x ¢p2 + y ¢p2 .
2.13.1.3 Struve Ratio and Aberration Tolerances

For small aberrations, the Struve ratio, i.e., the central value of the LSF normalized
to unity in the absence of any aberration, is given by
2
1 1- y n2
3 Û Û
LSF (0) =
16 Ù
ı
dyn Ù
ı
[ ]
exp iF( x n , yn ) d x n
-1
- 1- y n2
2
1 1- y n2
3Û Û
= Ù dy n
16 ı
Ù
ı
{[
exp i F( x n , y n ) - F( y n ) ]}d x n
-1 - 1- y n2
2
1- y n2
3Û
1
Û Ô [
Ï1 + i F( x n , y n ) - F( y n ) ] ¸
Ô
= Ù dy n Ù Ì 1 2 ˝d x n
16 ı
-1
ı
- 1- y n2 Ó 2
[
Ô - F( x n , y n ) - F( y n ) ] + ...Ô
˛
1
2
3Û
~
4Ù
ı
(
dyn 1 - yn2 ) ÈÍÎ1 - 12 s ( y )˘˙˚
2
F n
-1
1
3Û
~
4Ù
ı
( )[
dyn 1 - yn2 1 - s 2F ( yn ) ]
-1
1
3Û
4ı
(
= 1 - Ù 1 - yn2 s 2F ( yn ) dyn ) , (2-245)
-1
where
-1
È 1- yn2 ˘ 1- y n2
Í Û ˙ Û
F ( yn ) = Í Ù dx n ˙ Ù F( x n , yn ) dx n
Í ı ˙ ı
ÍÎ- 1- yn2 ˙˚ - 1- y n2
1- y n2
1 Û
= Ù F( x n , yn ) dx n (2-246)
2 1- yn2 ı
- 1- y n2
is the mean value of the aberration function along the x n axis for a given value of yn and
2
s 2F ( yn ) = F 2 ( yn ) - F( yn )
(2-247)
is the corresponding aberration variance. We have already shown [see Eq. (1-221c)] that
this ratio is less than or equal to 1.
Now we outline the derivation of Struve ratio for defocus and astigmatism and give
tolerance for a Struve ratio of 0.8 for some aberrations.
Defocus
(
F( x n , yn ) = Bd x n2 + yn2 ) , where Bd is in radians.
1- y n2
Bd Û
F ( yn ) =
2 1- yn2
Ù
ı
(x 2
n )
+ yn2 dx n
- 1- y n2
Bd
=
3
(
1 + 2 yn2 ) .
1- y n2
Bd2 Û 2
F 2 ( yn ) =
2 1- yn2
Ù
ı
(x 2
n + yn2 ) dx n
- 1- y n2
1 4 2 8 4ˆ 2
= Ê + y + y B .
Ë 5 15 n 15 n ¯ d
4 Bd2 2
s 2F ( yn ) =
45
(
1 - yn2 ) .
1
Û 128 Bd2
ı
( 2 2
)
Ù 1 - yn s F ( yn ) dyn = 1575 .
-1
LSF(0) = 0.8 for Bd = 0.29l .
Astigmatism combined with defocus
F( x n , yn ) = Aa x n2 + Bd x n2 + yn2 ( ) .
b = 0 : The line object is parallel to the yo axis.
4 2
s 2F ( yn ) =
45
(
1 - yn2 ) (A a + Bd )
2
= 0 if Bd = - Aa .
Hence, a line object parallel to the yo axis is not affected by astigmatism if the image is
observed in a defocused plane Bd = - Aa .
b = p 2 : The line object is parallel to the x o axis.
(
F( x n¢ , yn¢ ) = Aa yn¢ 2 + Bd x n¢ 2 + yn¢ 2 ) .
16 2
s 2F ( yn¢ ) =
45
(
1 - yn¢ 2 ) Bd2 .
Hence, a line object parallel to the x o axis is not affected by astigmatism if its image is
observed in the Gaussian image plane ( Bd = 0 ).
b = p 4:
1
F( x n¢ , yn¢ ) =
2
2
Aa ( x n¢ - yn¢ ) + Bd x n¢ 2 + yn¢ 2 ( ) .
1 4 2
s 2F ( yn¢ ) =
45
(
1 + 13 yn¢ 2 - 14 yn¢ 4 Aa2 +
45
)
1 - yn¢ 2 ( ) (B2
d + Bd Aa ) .
s 2F ( yn¢ ) is minimized if Bd = - (1 2) Aa , in which case
Aa2
s 2F ( yn¢ ) =
3
( )
1 - yn¢ 2 yn¢ 2 .
1
Û 16 Aa2
ı
( 2 2
)
Ù 1 - yn¢ s F ( yn¢ ) dyn¢ = 315 .
-1
LSF(0) = 0.8 if Aa = 0.365l .

The balanced aberrations and tolerance for a Struve ratio of 0.8 are given in Table 2-
16. Comparing with the Strehl ratio tolerances given in Tables 2-5 and 2-6, we note that
the tolerance for defocus is slightly larger than that for a Strehl ratio of 0.8; the tolerance
for spherical aberration is about the same; the tolerance for coma is about the same when
b = 0 , but it is less severe for b = p 2; the tolerance for secondary coma is more severe
for b = 0 and less severe for b = p 2 . Note that when b = 0 , the line object and its
Gaussian image lie along the y axis, which is perpendicular to the symmetry axis of the
coma PSF.
2.13.1.4 LSF From OTF34, 35

The infinite limits over the PSF integration can also be avoided by writing the LSF in
terms of the OTF, as was done in Eq. (1-226). Thus, the LSF may be written
LSF( x ) = Ú t (x, 0) exp ( - 2 p i x x ) dx . (2-248a)
The LSF along any other direction may be obtained by using the corresponding profile of
the transfer function. Thus, for example, the LSF for a line object along a direction yo¢ (or
parallel to it) making an angle b with the yo axis is given by [see Eq. 1-229)]
Table 2-17. Balanced aberrations and their tolerance for a Struve ratio of 0.8.
3 1 Tolerance for a
Aberration Balanced aberration (1 – yn2 ) s w2 ( yn ) dyn
4 Ú
–1
Struve ratio of
0.8 in units of l
Defocus Bd r 2 2.4063 Bd2 0.29
Primary 4
As (r – 0.984 r ) 2 0.1835 As2 1.09 in the plane
spherical Bd = – 0.984As
As r4
Secondary Aa yn2 cos q for b = p / 2 0.0124 W62 4.02
spherical
6
W6 r
Astigmatism Aa yn2 cos q for b = 0 0 No dependence on Aa

in the plane B d = – A a
Aa x n2 = Aa r2 cos 2 q
Aa yn2 cos q for b = 0 0 No dependence on Aa
in the Gaussian image
plane Bd = 0
Aa x n¢ yn¢ cos q for b = p / 4 1.504 Aa2 0.365 in the plane

Bd = – A a/ 2
Primary Ac (r 3 – 0.657r)cos q for b = 0 0.5962 Ac2 0.579

coma
Ac r 3cos q Ac r 3 cos q for b = p / 2 0.2674 Ac2 0.865
Secondary W5 (r 5 – 1.19 r 3 + 0.29 r) cos q for b = 0 0.036 W52 2.36

coma
0.020 W52 3.13

W5 r5 cos q W5 (r5 – 1.17 r 3 )cos q for b = p / 2
r 2 = x n2 + yn2 , x n = r cos q, yn = r sin q, x n = x n¢ cos b – yn¢ sin b, yn = x n¢ sin b + yn¢ cos b

LSF( x ¢) = Ú t (x ¢, 0) exp ( - 2 p i x ¢x ¢) dx ¢ , (2-248b)
where x ¢ and x ¢ are parallel to each other and normal to y ¢ and t (x¢, 0) is the OTF
along the direction x ¢ .
For the aberration-free case, the OTF is real given by Eq. (2-44). Hence, noting that
-1 £ x £ 1 , Eq. (2-248a) reduces to
1
Û
LSF( x ) = 2Ù t (x, 0) cos (2 p x x )dx , (2-249)
ı
0
where [see Eq. (2-44a)]
2 È -1 1 2˘
t (x, 0) =
p ÍÎ
(
cos x - x 1 - x 2 ) ˙˚
, -1 £ x £ 1 . (2-250)
Its value at the center is given by
1
Û
LSF(0) = 2Ù t (x, 0) dx = 8 3p . (2-251)
ı
0
Accordingly, the LSF normalized by its central value may be written33
1
Û
LSF( x ) = (3p 4) Ù t (x, 0) cos (2 p x x )dx (2-252a)
ı
0
3
= H1 (2 p x ) , (2-252b)
8p x 2
where H1 ( x ) is the first-order Struve function.37 Its behavior for small values of x may be
obtained by expanding cos (2p x x ) in a power series and integrating term by term. Thus,
1
Û
ı
[ 2
LSF( x ) = (3p 4) Ù t (x, 0) 1 - (2 p x x ) + O x 4 dx ( )] (2-253a)
0
1
3p
= 1- (2 p x)2 Û
Ù t (x, 0) x dx + O x
2 4
( ) . (2-253b)
4 ı
0
For a symmetric aberration (e.g., spherical aberration or astigmatism), t (x, 0) is real

and the integral of the first term in Eq. (2-253a) yields the central value of the LSF, i.e.,
the Struve ratio:
1
Û
LSF(0) = (3p 4) Ù t (x, 0) dx . (2-254)
ı
0
For large values of x, we use a formula by Willis38 in Eq. (2-252a), namely,

•
Û f ¢(0) f ¢¢¢(0)
Ù f ( x ) cos mx dx ~ - + + ... , (2-255)
ı m2 m4
0
and obtain
È t ¢ (0, 0) ˘
LSF( x ) ~ (3p 4) Í 2 ˙+O x
ÍÎ (2 p x ) ˙˚
( )
-4
= 3 4p 2 x 2 , (2-256)
where we have used Eq. (2-50) in the last step. Since t ¢ (0, 0) = - 4 p is independent of
the aberration, so is (approximately) the asymptotic behavior of the LSF.
The cumulative LSF is given by

x0
Û
L( x 0 ) = Ù LSF( x ) dx . (2-257)
ı
- x0
Substituting for the aberration-free LSF from Eq. (2-249), we obtain
1 x0
Û Û
L( x 0 ) = 4Ù dx t (x, 0) Ù cos(2 px x ) dx
ı ı
0 0
1
Û sin(2 px x 0 )
= 2 Ù t (x, 0) dx . (2-258)
ı px
0
By Dirchlet’s integral formula39
L( x 0 Æ • ) = 1 . (2-259)
2.13.1.5 ESF From OTF35

The edge-spread function represents the image of an edge, i.e., an object that covers
one half of the object plane. Its derivative gives the LSF. The image of an object lying
parallel to the y axis in the left half of the object plane is given by Eq. (1-233). For a
system with a radially symmetric aberration, the OTF is real. Hence, Eq. (1-233) reduces
to
1
1 Û sin(2 px x )
ESF( x ) = + t( x , 0 ) dx (2-260)
2 Ù
ı px
0
1
=
2
[1 + L( x )] . (2-261a)
We note from Eq. (2-260) that
ESF( ± x ) = [1 ± L( x )] 2 . (2-261b)
Moreover, since L(0) = 0 , it is evident from Eq. (2-261a) that, regardless of the value or
the type of a symmetrical aberration, the central value of the ESF is equal to 1/2, i.e.,
ESF(0) = 0.5 . (2-261c)
Also, since L(•) = 1, ESF( - •) = 0 and ESF(•) = 1.
If the aberration is not radially symmetric as, for example, in the case of coma, then
the ESF will depend on the orientation of the edge. For an object along a direction yo¢
making an angle b with the yo axis (see Figure 1-9), Eq. (1-233a) may be written
1
1 1 Û t( x ¢ , 0 )
ESF( x ¢) = - exp( - 2 p ix ¢ x ¢) d x ¢ . (2-262a)
2 2 pi Ù
ı x¢
-1
From Eq. (1-233b), the ESF at the origin is given by

1
1 1 Û Im t(x ¢, 0)
ESF(0) = - d x¢ . (2-262b)
2 pÙı x¢
0
If Im t(x¢, 0) = 0 , as for coma when b = p 2, then ESF(0) = 1 2 .

The aberration-free LSF normalized to unity at the center is shown in Figure 2-52,
both on a linear and a log scale. Its minima are not zero, unlike the corresponding PSF.
Figure 2-53 shows the defocused LSF. It also shows the LSF aberrated by primary
spherical aberration. Both the classical and balanced (for a point object) aberrations are
considered. The irradiance at the center decreases, and it increases in the surrounding
region as the aberration increases. Of course, the aberration-free LSF and those aberrated
by a radially symmetric aberration are symmetric about the origin. The LSF aberrated by
primary coma is shown in Figure 2-54 for various orientations b of the line object. For
b = p 2 (corresponding to a line object along the symmetry axis of the coma PSF), the
peak value lies at the origin. For small aberrations and b = 0 , it lies (very nearly) at the
same point where the peak of the PSF lies, i.e., at x ¢ = ( 4 3) Ac . The aberration-free ESF
is shown in Figure 2-55. It has a value of 0.5 at the origin. The value of the ESF aberrated
by primary coma at the Gaussian image of the edge as a function of the peak value Ac of
coma is shown in Figure 2-56. When b = p 2 , this value is equal to 0.5 regardless of the
value of the aberration. The ESF for primary coma is shown in Figure 2-57.
0.8
0.6
LSF(x)
0.4
0.2
0
0 0.5 1 1.5 2 2.5 3 3.5 4
x
(a)
100
10–1
LSF(x)
10–2
10–3
0 0.5 1 1.5 2 2.5 3 3.5 4
x
(b)
Figure 2-52. Aberration-free LSF normalized to unity at the center. (a) Linear scale.
(b) Log scale. x is in units of l F. The radius of the Airy disc is 1.22 in these units.
0.8
Bd = 1/4
0.6
LSF(x)
(a)
1/2
0.4
0.2
0
0 0.5 1 1.5 2 2.5 3
x
0.8
As = 1/4
0.6
LSF(x)
(b)
1/2
0.4
0.2
0
0 0.5 1 1.5 2 2.5 3
x
0.8
0.6
As = 1
LSF(x)
(c)
0.4
2
0.2
0
0 0.5 1 1.5 2 2.5 3
x
Figure 2-53. Aberrated LSF. (a) Defocus. (b) Primary spherical aberration As r4 . (c)
( )
Balanced spherical aberration As r4 - r2 . Bd and As are in units of l and x is in
units of l F.
1
p/2 Ac = 1.0
p/4
0.8
b=0
0.6
LSF(x) (a)
0.4
0.2
0
–3 –2 –1 0 1 2 3
x
1
Ac = 1
0.8 p/2
p/4
0.6
LSF(x)
b=0 (b)
0.4
0.2
0
–3 –2 –1 0 1 2 3
x
Ac = 1.5
0.8
0.6
LSF(x)
(c)
p/2 p/4
0.4
b=0
0.2
0
–3 –2 –1 0 1 2 3
x
Figure 2-54. LSF aberrated by coma Acr3 cosq , where Ac is in units of l and x is in
units of l F. The line object lies along a direction making an angle b with the yo
axis and the variation with x shown here is along a direction that is normal to it.
0.8
0.6
ESF(x)
0.4
0.2
0
–4 –3 –2 –1 0 1 2 3 4
x
Figure 2-55. Aberration-free ESF. The Gaussian image of the edge has a value of
zero for x < 0 and unity for x ≥ 0 . The vertical dashed line indicates the Gaussian
image of the edge object and x is in units of l F. The edge object lies along a
direction making an angle b with the yo axis and the variation with x shown here is
along a direction that is normal to it.
0.5
0.4
ESF(0)
0.3
0.2
p/4
0.1
b=0
0.0
0 0.5 1.0 1.5 2.0 2.5
Ac
Figure 2-56. ESF value at the Gaussian image of the edge as a function of the peak
value Ac of primary coma Acr3 cosq in units of l .
1.0 Ac = 0.5
0.8
p/4
0.6
ESF(x)
0
0.4 b = p/2
0.2
0.0
–8 –6 –4 –2 0 2 4 6 8
x
1.0
Ac = 1.0
0.8
p/4
ESF(x)
0.6
0
0.4 b = p/2
0.2
0.0
–8 –6 –4 –2 0 2 4 6 8
x
1.0
Ac = 1.5
0.8
p/4
0.6
ESF(x)
0
0.4 b = p/2
0.2
0.0
–8 –6 –4 –2 0 2 4 6 8
x
Figure 2-57. ESF for primary coma Acr3 cosq for various values of Ac in units of l .
2.14 Miscellaneous Topics 205
2.14 MISCELLANEOUS TOPICS

Now we discuss a few special topics. First, we consider how a monochromatic PSF,
namely the Airy pattern, changes as the spectral bandwidth increases. In particular, the
diffraction rings disappear as the relative bandwidth approaches unity. The polychromatic
OTF is also considered briefly. As an application of the convolution technique or its
counterpart in the spatial frequency domain, we consider the image of an incoherent disc
as an example of an extended object and thereby define a point source. Next, we consider
a pinhole camera as an example of imaging by a circular aperture without a lens. The
Rayleigh criterion is applied to determine the tolerable defocus in order to obtain an
approximate relationship between the pinhole radius and the camera length.
2.14.1 Polychromatic PSF

Equation (2-15) represents the aberration-free PSF for an object emanating
monochromatic radiation. For example, it gives the irradiance distribution of a laser beam
of radius a and wavelength l focused at a distance R. The monochromatic PSF, as
discussed in Section 2.2.1, consists of a bright spot surrounded by dark and bright
diffraction rings with well-defined maxima and minima. The radius of the bright spot and
the location of the diffraction rings depend on the wavelength of the object radiation. For
an object radiating in a wide spectral band, a diffraction pattern is formed at each
wavelength with the result that the diffraction rings loose their sharpness.
In this section, we develop an expression for the polychromatic PSF (PPSF) such that
parametric numerical analysis can be carried out in terms of the mean wavelength and the
spectral bandwidth of the object radiation.40, 41 Although the object radiance will
generally vary with wavelength (e.g. as in the case of a tungsten lamp or solar
illumination), we assume it to be constant for simplicity of numerical calculations. The
primary objective of such calculations is to see how the diffraction rings loose their
sharpness as the spectral bandwidth is increased, and to determine the bandwidth for
which they completely disappear. In practice, of course, the spectral transmittance of the
imaging system and spectral sensitivity of the observation medium, e.g., a film or a solid-
state detector array, must also be taken into account to determine the measured PPSF.
For an extremely narrow spectral bandwidth dl as, for example, in a light-emitting

diode, the PSF is nearly monochromatic. From Eq. (2-15), the quasi-monochromatic PSF
(QPSF) in this case may be written
2
p p dl È 2 J1 ( p ri l F ) ˘
I (ri ; dl ) = Í ˙ , (2-263)
4l2 F 2 Î p ri l F ˚
where p is the image power per unit spectral bandwidth ( p dl is the total image power),
and l is the mean wavelength. Assuming for simplicity that the polychromatic object
radiates uniformly as a function of wavelength, the PPSF is obtained by integrating Eq.
(2-263) over the spectral bandwidth Dl = l 2 - l1 , where l1 and l 2 represent the

minimum and the maximum wavelengths of the spectral band. Thus, the PPSF is given by
l2 2
p p Û È 2 J1 ( p ri l F ) ˘ 1
I (ri ; Dl ) = Ù Í ˙ dl . (2-264)
4 F 2 ı Î p ri l F ˚ l2
l1
Normalizing the PPSF with the central value p p dl 4l2m F 2 of the QPSF, we may write
l2 2
l2 Û È 2 J ( p r l F ) ˘ 1
In (ri ; Dl ) = m Ù Í 1 i ˙ dl , (2-265)
dl ı Î p ri l F ˚ l2
l1
where the mean wavelength l m = (l1 + l 2 ) 2 .
The central value of the normalized PPSF is given by

l2
l2 Û 1
In (0; Dl ) = m Ù 2 d l
dl ı l
l1
Dl 1
= , (2-266)
dl 1 - ( Dl 2 l m ) 2
where we have substituted for l1 and l 2 in the form
Ê Dl ˆ
l1 = l m Á1 - ˜ (2-267a)
Ë 2l m ¯
and
Ê Dl ˆ
l 2 = l m Á1 + ˜ , (2-267b)
Ë 2 lm ¯
respectively. For the quasi-monochromatic object, let dl l m = 10 - 4 . This choice is

somewhat arbitrary and merely helps to establish the central value of the PPSF relative to
the central value of the QPSF. Hence, Eq. (2-266) becomes
Dl 10 4
In (0; D l ) = 2 . (2-268)
lm Ê Dl ˆ
1- Á ˜
Ë 2l m ¯
Figure 2-58a shows the variation of the normalized central irradiance as a function of the
normalized spectral bandwidth.
Letting r = ri l m F and l n = l l m , Eq. (2-265) reduces to

107
106
105
In (0)
104
103
102
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
Dl/l m
(a)
106
104
1 Dl / lm = 1.99
102
In (r; Dl)
0.25
0.1
100
10–2
10–2
10–4
–4
10
–6
10
0 1 2 3 4 5 6 7 8 9 10
r
(b)
Figure 2-58. (a) Aberration-free central irradiance of the PPSF normalized by the
central value of the QPSF as a function of the relative spectral bandwidth D l l m .
(b) Aberration-free PPSF for various relative spectral bandwidths. The units of r
are l m F.
Dl
1+
2l m
4 ¥ 10 4 Û
In ( r; D l ) = Ù J12 ( p r l n ) d l n , (2-269)
p2r 2 ı
Dl
1-
2l m
which is suitable for parametric numerical analysis. Figure 2-58b shows the PPSF for
various values of the relative spectral bandwidth Dl l m . The diffraction rings begin to
lose their identity as Dl l m increases. However, the ring structure does not disappear
until Dl l m = 1, i.e., until the spectral bandwidth equals the mean wavelength.
2.14.2 Polychromatic OTF

Equation (2-44) gives the aberration-free OTF of a system imaging a monochromatic
object. If the object radiates at wavelengths varying from l1 to l 2 , then the
polychromatic OTF of the system can be obtained by integrating the monochromatic OTF
over the spectral range.42 Neglecting for simplicity the variation of spectral radiance of
the object, spectral transmittance of the system, and spectral sensitivity of the image
recording medium, the aberration-free polychromatic OTF can be written
l2
1 Û
t p (vi ) = t(vi ) d l
Dl Ù
ı
l1
l2
2 Û È -1 1 2˘
=
p Dl Ù
ı Í
Î
cos (l Fvi ) - l Fvi 1 - l2 F 2 vi2 ( ) ˚˙
dl . (2-270)
l1
Or,
q
2 Û È -1 12
t p (v) = 2 2
Ù cos ( xv) - xv 1 - x v
p (q - 1) ı ÎÍ
( ) ˘dx , 0 £ v £1 ,
˚˙
(2-271)
1
where x = l l1 , q = l 2 l1 , and v = vi (1 l1 F ) is a normalized spatial frequency. In

practice, q £ 3, since a given detector does not respond to a very wide spectral
bandwidth. The cutoff spatial frequency is given by 1 l1 F . Hence, the shortest
wavelength l1 yields the largest cutoff frequency, and the longest wavelength l 2 yields
the smallest cutoff frequency. Since the monochromatic OTF decreases monotonically
with frequency from a maximum value of unity at the origin to a value of zero at the
cutoff frequency (see Figure 2-6), the polychromatic OTF is lower at each spatial
frequency compared to the monochromatic OTF for l1 . Hence, the polychromatic OTFs
obtained are as shown in Figure 2-59, as the spectral bandwidth is increased. The quantity
q is related to the relative spectral bandwidth Dl l m considered in the discussion of the
polychromatic PSFs according to
0.8
q=1
0.6
2
tp (n)
4
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1
n
Figure 2-59. Polychromatic OTF of an aberration-free system.
2 + Dl lm
q = . (2-272)
2 - Dl lm
The monochromatic OTF corresponds to q = 1.
2.14.3 Image of an Incoherent Disc 43-46
As an example of the image of an incoherent extended object, we consider the image

of a uniformly radiating or illuminated disc of radius ho and radiance B lying at a
distance zo from the entrance pupil of area Sen of an imaging system. The irradiance
distribution of the image may be obtained by convolving the Gaussian image with the
PSF of the imaging system43 or by inverse Fourier transforming its spectrum.44 The latter
approach is simpler for numerical evaluations of systems with radially symmetric
aberrations since it involves only a one-dimensional integration. As a corollary to this
problem, we determine the size of an object that can be treated as a point. It is shown that
if the radius of the Gaussian image of the object is approximately one quarter of the
radius of the Airy disc, then the object is practically a point. This helps to define the size
of a pinhole that can be treated as a point source.
Gaussian Image
The total power entering the system is given by
(
Pen = p ho2 Sen zo2 B . ) (2-273)
The total power in the exit pupil and, therefore, in the image is given by Pex = hPen ,
where h is the transmission factor of the system, as in Eq. (2-3). This power is uniformly
distributed in the circular Gaussian image of radius hg = Mho , where M is the
magnification of the image. Hence, the irradiance of the Gaussian image is given by
r
( )
Ig rg = Pex p hg2
(
= h Sen zo2 M 2 B )
r
= Ig for rg £ hg
= 0, otherwise , (2-274)
r
where rg is the position vector of a point on the Gaussian image.
Diffraction Image
The diffraction image of the disc is given by Eq. (1-56c), where the PSF for an
aberration-free system is given by Eq. (2-16). Hence, the irradiance distribution of the
aberration-free diffraction image may be written
hg 2 p
2
r Ig Sex Û Û È 2 J ( p s l F ) ˘
Ii (ri ) = 2 2 Ù Ù Í 1 ˙ rg drg dq g , (2-275)
l R ı ı Î ps l F ˚
0 0
r
where ri is the position vector of a point in the image plane and
r r
s = ri - rg
12
[ (
= ri2 + rg2 - 2 ri rg cos q i - q g )] . (2-276)
The irradiance distribution is radiallly symmetric since the integral does not depend on
q i ; indeed we may let q i = 0 in Eq. (2-276). Letting
r = ri l F, r g = rg l F, and bg = hg l F , (2-277)
Eq. (2-275) reduces to

bg 2 p
2
p Ig Û Û È 2 J1 ( p t ) ˘
Ii ( r ) = Ù Í ˙ r g dr g dq g , (2-278)
4 Ù ı ı Î pt ˚
0 0
where
12
(
t = r 2 + r2g - 2 ri r cos q g ) . (2-279)
The irradiance at the center of the image is given by

bg
2
Û 2 J1 pr g
Ii ( 0 ) = I g Ù
( )
dr g
ı rg
0
[ ( )
= Ig 1 - J 02 p bg - J12 p bg ( )] . (2-280)
The variation of Ii (0) with bg is similar to the variation of the encircled power of the
Airy pattern [see Eq. (2-24) and Figure 2-2]. It approaches Ig asymptotically.
The diffraction image can also be obtained by working in the spatial frequency
domain, i.e., by inverse Fourier transforming its spectrum, which itself is equal to the
product of the spectrum of the Gaussian image and the OTF of the imaging system. Since
the Gaussian image is radially symmetric, its spectrum (equal to its Fourier transform) is
given by its zero-order Hankel transform:
hg
r Û
(
Ĩg (vi ) = 2 p Ig Ù J 0 2 pvi rg rg drg
ı
)
0
( ) (
= Ig hg vi J1 2pvi hg ) . (2-281)
Hence the spectrum of the aberrated diffraction image is given by

r r r
I˜i (vi ) = I˜g (vi ) t(vi )
r
( ) (
= Ig hg vi J1 2pvi hg t(vi ) , ) (2-282)
r
where t(vi ) is the aberrated OTF. The irradiance distribution of the image is obtained by
r
inverse Fourier transforming the spectrum, i.e., Ii (ri ) is given by
r r r r r
Ii (ri ) = Ú I˜i (vi ) exp( - 2pivi ◊ ri ) dvi
r r r r
( )
= Ig hg Ú (1 vi ) J1 2 pvi hg t(vi ) exp( - 2 pivi ◊ ri ) dvi . (2-283)
The OTF of a system with a radially symmetric aberration is also radially symmetric
r
[see Eq. (1-83)] and, therefore, we may replace t(vi ) by t(vi ) . Accordingly, the image
spectrum is also radially symmetric. Its inverse Fourier transform, which gives the
irradiance distribution of the diffraction image, is equal to its zero-order Hankel
transform, i.e.,
Ii (ri ) = 2 p Ú I˜i (vi ) J0 (2 pvi ri ) vi dvi
( )
= 2 p Ig hg Ú J1 2 pvi hg J 0 (2 pvi ri ) t(vi ) dvi . (2-284)
Or, letting r = ri l F and v = vi (1 l F ) ,
( )
Ii (r ) = 2 p I g bg Ú J1 2 pvbg J 0 (2 pvr ) t(v) dv . (2-285)
The value of the irradiance at the center of the image is given by
(
Ii (0) = 2 p Ig bg Ú J1 2 pvbg t(v) dv .) (2-286)
For an aberration-free system, the integration can be carried out in a closed form [see
Eqs. (2-50) and (2-24)] with the result
[ ( )
Ii (0) = Ig 1 - J 02 p bg - J12 p bg ( )] , (2-287)
in agreement with Eq. (2-280).
For a system with a radially symmetric aberration, the power in a circle of radius rc
(in units of l F ) is given by
rc
2 Û
P (rc ) = 2 p (l F ) Ù Ii (r ) r dr . (2-288)
ı
0
Substituting for Ii (r ) from Eq. (2-285), we obtain
rc
Û
P (rc )
2
( )
= (2 pl F ) Ig bg Ú dv J1 2 pvbg t(v) Ù J 0 (2 pvr ) r dr
ı
0
( ) ( )
= 2 Pex rc bg Ú J1 2 pvbg J1 (2 pvrc )t(v) (1 v)dv . (2-289)
Letting rc = bg , we obtain the power in a circle whose radius is equal to that of the
Gaussian image:
( ) ( )
P bg = 2 Pex Ú J12 2 pvbg t(v) (1 v)dv . (2-290)
( )
It is easy to see from Eq. (2-289) that the power P rc ; bg in a circle of radius rc due to a
( )
source of radius bg and the power P bg ; rc in a circle of radius bg due to a source of
radius rc are related to each other according to
( ) (
bg2 P rc ; bg = rc2 P bg ; rc ) . (2-291)
For a very small disc, i.e., for very small values of bg , e.g., a pinhole,
(
J1 2pvi hg ) ~ pvi hg , (2-292)
and Eq. (2-283) reduces to

r r r r r
Ii (ri ) = Pex Ú t(vi ) exp ( - 2 pivi ◊ ri ) dvi
r
= Pex PSF(ri ) . (2-293)
Thus, as expected, it reduces to the irradiance distribution of the image of a point object.
This result may also be obtained from Eq. (2-285). In that case, the PSF is the zero-order
Hankel transform of the OTF because of the radial symmetry.
It is much simpler to calculate the encircled power from Eq. (2-289) using the
transfer function, than, for example, by first calculating the irradiance distribution of the
extended image and then integrating it over a circle, as was done by Goldberg and
McCulloch45 for an aberration-free system. Substituting the value of t(v) in Eq. (2-285),
we can calculate the irradiance distribution of the diffraction images.
Numerical Results
Numerical results for the disc images can be obtained from Eqs. (2-285) and (2-289),
where the aberration-free t(v) is given be Eq. (2-44) The defocused results can be
obtained by substituting for t(v) from Eq. (2-192b). Figure 2-60 shows the irradiance
distribution of the aberration-free images of a disc of various sizes, normalized to unity at
the center. We note that when the radius bg of the Gaussian image is equal to one quarter
(in units of l F ), the image is approximately the same as the Airy pattern, i.e., such an
object can be treated as a point. Similarly, a pinhole whose Gaussian image radius is one
quarter can be treated as a point source.
The encircled power of the aberration-free image of a disc is shown in Figure 2-61a
for several disc sizes. Again, the curve for bg = 1 4 closely resembles the encircled
power of the Airy pattern, confirming the size of a pinhole that can be treated as a point
source. The power in a circle whose radius is equal to that of the Gaussian image of a disc
is shown in Figure 2-61b as a function of the disc size. It increases monotonically as the
disc size increases. For small values of bg , this circle contains only a small fraction of the
total power since it is spread by diffraction. It contains a significant fraction of the total
power when its size is significantly larger than the Airy disc, since then the relative effect
of diffraction is small.
Figure 2-62a shows how the central irradiance varies as a function of bg for several
defocus values. It increases according to Eq. (2-286) and approaches unity (actually Ig )
asymptotically. The axial irradiance of a disc of various sizes is shown in Figure 2-62b.
For bg = 1 4, it is approximately the same as for a point source [see Eq. (2-84c)]. For
example, it reduces to a value of nearly zero for integral values of Bd (in units of
wavelength). As the disc becomes larger, the axial irradiance does not reduce to zero for
such values values of Bd .
0.8
2
0.6
Ii(r)
0.4 1
1/2
0.2
bg = 0
1/4
0
0 0.5 1 1.5 2 2.5 3
r
Figure 2-60. Irradiance distribution of the aberration-free image of an incoherent

disc normalized to unity at the center. The radius of the Gaussian image is bg . The
dotted curve for bg Æ 0 is the Airy pattern. The disc can be treated as a point
source when bg £ 1 4 . Both bg and r are in units of l F.
Figure 2-63 shows the defocused images of a disc of various sizes. We note that the
image generally resembles the object. In particular, it is bright in the central region and
dim in the outer region. As defocus increases, the irradiance decreases in the central
region and increases in the outer region. As stated earlier, if the disc is small and the
defocus is large, the irradiance at the center may be smaller than that in the outer region;
see for example, Figure 2-63a for Bd = 1l . This behavior is similar to that for a point
source, and it disappears as the disc size increases.
In Section 2.15.5, we consider the image of a coherently illuminated disc and show
that, unless the disc is very small, the central irradiance of the image can be much lower
than that in the surrounding region, whether or not the image is defocused. Moreover,
defocus can increase the central irradiance. As shown in Figure 2-77, the irradiance at the
center is lower than that in the surrounding region, for example, when bg = 2 and
Bd = 0 . The central irradiance increases significantly as defocus is introduced; i.e., for
Bd π 0 .
0.8 bg = 1/4
1/2
1
2
0.6
P(rc)
0.4
0.2
0
0 0.5 1 1.5 2 2.5 3 3.5 4
rc
(a)
1.0
0.8
0.6
P(bg)
0.4
0.2
0
0 0.5 1 1.5 2 2.5 3 3.5 4
bg
(b)
Figure 2-61. (a) Encircled power of the aberration-free image of an incoherent disc
normalized by the total power Pex . rc is in units of l F. (b) Power in a circle of
radius equal to that of the Gaussian image of the disc as a function of its size.
Bd = 0
0.8
1/4
1/2
0.6
I(0; bg)
0.4
0.2
0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
bg
(a)
0.8
4
0.6
2
I(0; Bd)
1
bg = 1/4
0.4
0.2
0
0 0.5 1 1.5 2 3 4
Bd
(b)
Figure 2-62. (a) Central irradiance of the defocused image of an incoherent disc. Bd
is in units of l . (b) Axial irradiance normalized to unity at the center. The actual
values (normalized by I g ) in increasing order of bg are 0.14, 0.825, 0.91, and 0.94.
bg = 1
0.8
Bd = 0 Gaussian
image
1/4
0.6
Ii(r)
(a)
1/2
0.4
0.2
0
0 0.5 1 1.5 2 2.5 3 3.5 4
r
Bd = 0 bg = 2
0.8 1/4
Gaussian
image
1/2
0.6
Ii(r)
(b)
0.4
1
0.2
0
0 0.5 1 1.5 2 2.5 3 3.5 4
r
Bd = 0 bg = 2.3
0.8 1/4
Gaussian
image
1/2
0.6
Ii(r)
(c)
0.4 1
0.2
0
0 0.5 1 1.5 2 2.5 3 3.5 4
r
Figure 2-63. Irradiance distribution of defocused images of an incoherent disc. The

Gaussian image is also indicated in each figure. (a) bg = 1. (b) bg = 2 . (c) bg = 2.3 .
2.14.4 Pinhole Camera

A pinhole camera (or camera obscura) has been the subject of many investigations
including those by Petzval and Rayleigh and those that describe it in terms of coherence
theory. 47-53 Not only is it simple (a pinhole on one side of a box and the film on the
other), but it is distortion free with an infinite depth of field and a very wide field of view.
In this section, we obtain a relationship between the pinhole radius a and its distance L
from the image plane.
Based on geometrical optics, the image of a distant point object in the absence of a
lens will be approximately the same size as the pinhole if the pinhole is large. Reducing
the pinhole size reduces the image size until diffraction by the pinhole spreads it. Petzval
obtained the relationship by minimizing the image spot radius representing the sum of the
geometrical and diffraction contributions to it. The image radius based on diffraction (see
Figure 2-64) is approximately equal to the distance from the center of the image at which
a disturbance from the center of the pinhole cancels that from its edge. Thus, the
disturbance at the point P in the image plane at a distance L from the plane of the pinhole
is zero when the difference in optical path lengths between the disturbances from the
points O and A is equal to l 2, i.e., when
OP - AP = OC = l 2 . (2-294)
The diffraction spot radius is given by
rd = L q
= L(OC a) (2-295)
= Ll 2 a
Adding the geometrical and diffraction contributions, we obtain the image spot radius as
ri = a + L l 2 a . (2-296)
rd
A
a
q q
O C
B
L
Figure 2-64. Diffraction spot radius based on the cancellation of disturbances from
the center and the edge of a pinhole of radius a. The camera length L >> a .
The smallest spot radius is obtained by differentiating with respect to a and equating the
result to zero. This yields the result that
12
aP = ( L l 2) . (2-297)
Substituting Eq. (2-297) into Eq. (2-296), we obtain the minimum spot radius
ri = 2 aP (2-298a)
12
= (2 Ll ) . (2-298b)
The difference between a pinhole camera and a regular camera is that the former
does not have a lens to form the image. The lens in a regular camera converts a diverging
spherical wave from a point object into a spherical wave converging to a point in the
image plane. The absence of a lens in a pinhole camera may be thought of as a lens of
zero power. It implies that, except for diffraction by the pinhole, the diverging spherical
wave continues as a diverging wave toward the image plane. The corresponding defocus
aberration is given by
1Ê 1 1ˆ 2
( )
W rp = Á - ˜ rp
2 Ë Li Lo ¯
, (2-299)
where rp is the radial distance of a point in the plane of the pinhole from its center and
Lo and Li are the object and image distances from it (where Lo is numerically negative
according to our sign convention). It represents the sum of the sags of two spherical
wavefronts passing through the center of the pinhole with their centers of curvature lying
at the object and image points. The image will be practically diffraction limited according
to the Rayleigh criterion, if the peak value of the aberration is less than or equal to l 4 .
For a distant object ( Lo Æ - •), letting rp = a , Eq. (2-299) yields the Petzval result of
Eq (2-297). The image spot for a point object is approximately the Airy disc with a radius
of 0.61l Li a , or 0.61 times the value estimated by Petzval.
Another approach to optimize the pinhole radius is to maximize the central irradiance
of the diffraction pattern of a circular aperture. The axial irradiance for a circular aperture
is given by Eq. (2-84), where P is the power entering the pinhole and Bd is the peak
defocus phase aberration. For a point source of intensity Io , the power entering the
pinhole is given by the solid angle pa 2 L2o that it subtends at the source multiplied by
Io . Hence, the axial irradiance is given by
4 Io Ê p a 2 Li - Lo ˆ
I (0) = sin 2 Á ˜ . (2-300)
( Li - Lo )2 Ë 2 l Li Lo ¯
For a given object distance Lo , the maximum central irradiance is obtained at an image
distance Li given by
∂I (0)
= 0 , (2-301)
∂Li
or
È pa2 Ê1 1 ˆ˘ pa2 Ê1 1 ˆ Lo
tan Í Á - ˜˙ = - Á - ˜ (2-302)
ÍÎ 2 l Ë Li Lo ¯ ˙˚ 2l L
Ë i Lo ¯ Li
or
p p N
tan ÈÍ ( Ni + No )˘˙ = ( Ni + No ) i , (2-303)
Î2 ˚ 2 No
where No = a 2 l Lo and Ni = a 2 l Li are the Fresnel numbers of the pinhole as

observed from the object and image planes, respectively. For a distant object
( Lo Æ - •) , the central irradiance is maximum in the image plane Li = a 2 l , which is
equivalent to Ni = 1. By Taylor's expansion, an approximate solution of Eq. (2-302) or
Eq. (2-303) is given by
1
-
1
~ 1 ÈÍ1 + 42 No + Ê1 - 82 ˆ No2 ˘˙ , (2-304)
Li Lo f Î p
•
Ë p ¯ ˚
or
Ni ~ 1 - Ê1 - 42 ˆ No + Ê1 - 82 ˆ No2 , (2-305)
Ë p ¯ Ë p ¯
where f• = a 2 l is the focal length of the pinhole (i.e., the image distance when the
object distance is infinity).
Unless the object distance is very small, we may write Eq. (2-304) approximately as
1 1 1
- = , (2-306)
Li Lo fe
where
1
~ 1 + 0.4 No (2-307)
fe f•
is the effective focal length of the pinhole. We note that fe > f• and the maximum
irradiance is obtained at a point closer to the pinhole as the object moves closer to it, or
the pinhole subtends more than one Fresnel zone. This is in contrast to the case of a thin
lens for which the real image of a real object moves in the same direction as the object.
Rayleigh 48 also investigated Eq. (2-300) and, based on experimental observations,
12
concluded that the best image was obtained when a R = 0.95 ( Li l ) , which is equivalent
to the pinhole intercepting only 90% area of the first Fresnel zone.
Figure 2-65 shows how Ni varies with No according to Eq. (2-305). We note that
Ni decreases as the object moves closer to the pinhole (so that No increases), reaches a
minimum value, and then increases monotonically. The value of a as a function of Lo is
shown in Figure 2-66 for several values of Li . All the quantities are in units of l in this
figure. We note that, for a given value of Li , a does not change much as Lo increases.
2.5
2
Ni
1.5
0.5
0 1 2 3 4 5
No
Figure 2-65. Variation of image-side Fresnel number Ni with object-side Fresnel

number No .
350
L i = 100,000
300
70,000
250
a
200 40,000
150
10,000
100
50
105 106 107 108
ΩLoΩ
Figure 2-66. Optimum value of pinhole radius a as a function of object distance for
several values of camera length Li . All three quantities are in units of wavelength.
A relationship between a and L may also be obtained by a consideration of the OTF

of a pinhole. In order that the Hopkins ratio be greater than 0.8 for spatial frequencies less
than or equal to 0.1 of the diffraction cutoff, the peak defocus aberration must be less than
or equal to l 2, according to Eq. (2-187). Thus, for a distant object, we obtain the same
result as from Eq. (2-300).
Since the focal length of the pinhole camera, approximately equal to a 2 l , depends
on the wavelength, it suffers from chromatic aberration. Similarly, since the pinhole
appears to be elliptical from an off-axis point object, its focal length for an object in the
horizontal plane differs from that in a vertical plane. Hence, it suffers from astigmatism.
However, it is free of distortion.
2.15 COHERENT IMAGING

In this section, we briefly discuss aberrated imaging of a coherently illuminated
object.54 Since the distinction between a coherent and an incoherent object disappears as
the object becomes vanishingly small, there is no difference between a coherent and an
incoherent point-spread function. The difference between the two types of objects comes
about for an extended object. In the case of an incoherent object, the irradiance
distribution of its image is obtained by adding the irradiance distributions of the images
of its infinitesimal elements. But in the case of a coherent object, the complex amplitude
distribution of its image is obtained by adding the complex amplitude distributions of the
images of its elements. The irradiance distribution of the image is obtained by taking the
modulus square of the amplitude distribution thus obtained. Hence, whereas we speak of
the incoherent PSF as the irradiance distribution of the image of a point object, we speak
of the coherent PSF as the complex amplitude distribution of the corresponding image.
For an isoplanatic coherent object, the spatial frequency spectrum of its amplitude
image is given by the product of the spectrum of its Gaussian amplitude image and the
coherent transfer function of the imaging system. We give expressions for the aberration-
free point-, line-, and edge-spread functions. The transfer function is shown to be simply
a scaled version of the pupil function. The effect of the aberrations on the line- and edge-
spread functions is also discussed. As an example of an extended object, the image of a
coherent disc is considered. It is shown that, unless the disc is very small, defocus does
not necessarily decrease the central irradiance of its image.
It is shown that a Fourier transform of an object is obtained in the focal plane of an

imaging lens, thus offering the opportunity to alter the characteristics of the image by
spatial filtering in the focal plane. A qualitative comparison of coherent and incoherent
imaging is also given. In particular, a two-point resolution is compared, showing a sharp
difference between the images of coherently and incoherently illuminated points.
2.15.1 Coherent Spread Function

The coherent spread function representing the image of a point object observed in the
Gaussian image plane is given by Eq. (1-243):
2.15 Coherent Imaging 223
r 1 Û r Ê 2 pi r r ˆ r
CSF( ri ) = Ù
lR ı
( )
G rp exp Á -
Ë lR
rp ri ˜ d rp
¯
◊ , (2-308)
where we have omitted the factor of - i in the front of the right-hand side of the
equation. If a system with a circular exit pupil of radius a is aberration free, then its
relative pupil function is constant, say Go , that is inversely proportional to zo , and may
be written
(r )
G rp = Go
r
, for rp £ a ,
= 0 , otherwise . (2-309)
For such a system, Eq. (2-308) reduces to

a 2p
r 1 Û Û È 2p i ˘
CSF( ri ) =
lR Ù
ı
rp drp Ù exp Í-
ı Î l R
rp ri cos q p - q i ˙ d qp
˚
( )
0 0
a
2p Û Ê 2 p rp ri ˆ
= Ù J0 Á ˜ rp drp
lR ı Ë lR ¯
0
p a 2 2 J1 ( p r )
= , (2-310)
lR pr
where r = ri l F is the radial distance of a point in units of l F in the Gaussian image

plane from the Gaussian image point. Here, F = R D is the f-number of the image-
forming light cone. When the CSF is multiplied by the amplitude of the Gaussian image,
we obtain the amplitude distribution of the image of a point object. The CSF, given by
Eq. (2-310) but normalized by p a 2 l R so that it is unity at the center, is illustrated in
Figure 2-67. Its square is the Airy pattern shown in Figure 2-2, representing the irradiance
distribution of the image of a point object.
2.15.2 Coherent Transfer Function

The coherent transfer function of an imaging system is given by Eq. (1-249):
r r
CTF( vi ) = G(l Rvi ) . (2-311)
r r
( )
If F rp is the phase aberration function where rp £ a , and if the amplitude across the
pupil is uniform, then
r r
[
CTF( vi ) = exp iF(l Rvi ) ] . (2-312)
Thus, for an aberration-free system with a circular exit pupil of radius a, all frequencies
r
vi £ 1 2l F are transmitted by the system without any amplitude or phase distortion.
0.8
0.6
CSF (r)
0.4
0.2
– 0.2
0 1 2 3 4
r
Figure 2-67. Aberration-free coherent spread function normalized to unity at the

center. r is in units of l F. The Airy pattern shown in Figure 2-2a is the square of
the CSF shown here, and represents the irradiance distribution associated with the
coherent ESF.
r
The spatial frequency vic = 1 2l F is called the cutoff frequency of the system in the
sense that object frequencies corresponding to image frequencies greater than 1 2l F are
not transmitted by it. We note that the cutoff frequency for coherent imaging is half of
that for incoherent imaging.
2.15.3 Coherent LSF

The coherent LSF can be obtained from Eq. (1-229) provided the OTF is replaced by
the CTF, where - x c £ x £ x c and x c = 1 2 l F . The LSF for a line object along a
direction yo¢ making an angle b with the yo axis may be written
1
1Û
LSF( x ¢) =
2Ù
ı
{[
exp i F(x ¢, 0) - px ¢ x ¢ dx ¢ , ]} (2-313)
-1
[ ]
where F(x ¢, 0) is the aberration function and exp iF(x ¢, 0) is the CTF along the
direction x ¢ , x ¢ is in units of l F , x ¢ is in units of 1 2l F, and division by 2 on the
right-hand side normalizes the aberration-free LSF to unity at the origin. For an even
aberration such as defocus or astigmatism, the LSF may be written
1
Û
[ ]
LSF( x ¢) = Ù exp iF(x ¢, 0) cos( px ¢ x ¢) dx ¢ .
ı
(2-314a)
0
It is complex and symmetric about x ¢ = 0 . For an odd aberration such as coma, the LSF
is real and given by
1
Û
[
LSF( x ¢) = Ù cos F(x ¢, 0) - px ¢ x ¢ dx ¢ .
ı
] (2-314b)
0
For an aberration-free system, we obtain
sin p x
LSF( x ) = , (2-315)
px
independent of the orientation of the line object. It is symmetric about x = 0 , as shown in

Figure 2-68. Its value is zero for x = ± n , where n is a nonzero integer. This is different
from the incoherent LSF which is nonzero for these values of x (see Figure 2-52).
Figure 2-69 shows the irradiance distribution associated with a defocused LSF, i.e.,
when F(x) = Bd x 2 , where Bd is in radians. However, the value of Bd given in the figure
is in units of wavelength. The curve for Bd = 0 is the aberration-free case and represents
simply the square of the LSF shown in Figure 2-68. A defocus of Bd = l corresponds to
a PSF with a central value of zero. Figure 2-70 shows the irradiance distribution
associated with the LSF aberrated by one wave of spherical aberration balanced with
defocus, i.e., for W (r) = As r 4 + Bd r2 with As = l . The optimum balancing defocus for
imaging of a point object is Bd = - As , but it is given by Bd = - 0.86 As for a line object.
The irradiance distribution associated with the LSF aberrated by one wave of coma is
shown in Figure 2-71, i.e., for W (r, q) = Ac r3 cos q , where Ac = l , or
F(x ¢, 0) = Ac x ¢ 3 cos q , - 1 £ x ¢ £ 1 , and Ac is in radians. When b = p 2 , the CTF is
unity, since it is equivalent to letting q = p 2 in the aberration function. Hence, the LSF
associated with this orientation is the same as the aberration-free LSF.
1
0.8
0.6
0.4
LSF (x)
0.2
– 0.2
– 0.4
–6 –4 –2 0 2 4 6
x
Figure 2-68. Aberration-free coherent LSF normalized to unity at the center. It is

symmetric in x about x = 0. x is in units of l F.
1
Bd = 0
0.8
0.6 1/4
Re LSF (x)
1/2
0.4
(a)
0.2 1
0
– 0.2
– 0.4
0 0.5 1 1.5 2 2.5 3 3.5 4
x
0.6
1/2
0.4
Im LSF (x)
1/4
0.2
1 (b)
0
Bd = 0
– 0.2
– 0.4
0 0.5 1 1.5 2 2.5 3 3.5 4
x
Bd = 0
0.8
1/4
0.6
ΩLS F (x)Ω2
(c)
0.4
1/2
0.2
0
0 0.5 1 1.5 2 2.5 3 3.5 4
x
Figure 2-69. Defocused coherent LSF. (a) Real part. (b) Imaginary part. (c)
Irradiance distribution associated with the LSF. The curve Bd = 0 represents the
aberration-free LSF. x is in units of l F.
0.8
As = 1
0.6 Bd = – 0.86
Re LSF (x)
–1
0.4
(a)
0.2
– 0.2
0 0.5 1 1.5 2 2.5 3 3.5 4
x
0.6
0.4 As = 1
–1
Im LSF (x)
0.2
Bd = – 0.86
0
(b)
– 0.2
– 0.4
– 0.6
– 0.8
0 0.5 1 1.5 2 2.5 3 3.5 4
x
3.5
As = 1
3
Bd = – 0.86
2.5 – 1
ΩLS F (x)Ω2
(c)
1.5
0.5
0
0 0.5 1 1.5 2 2.5 3 3.5 4
x
Figure 2-70. Coherent LSF aberrated by one wave of spherical aberration balanced
with defocus (a) Real part. (b) Imaginary part. (c) Irradiance distribution associated
with the LSF. The optimum defocus for the LSF is Bd = - 0.86 As as opposed to
Bd = - As for optimizing a PSF. x is in units of l F.
1
Ac = 1
p/2
p/4
0.5
b=0
Re LSF (x)
– 0.5
–4 –3 –2 –1 0 1 2 3 4
x
(a)
1
p/2 Ac = 1
p/4
0.8
0.6
LSF2 (x)
b=0
0.4
0.2
0
–4 –3 –2 –1 0 1 2 3 4
x
(b)
Figure 2-71. Coherent LSF aberrated by one wave of coma (a) LSF which is real. (b)
Irradiance distribution associated with the LSF. The line object makes an angle b
with the yo axis and the variation with x shown here is along a direction that is
normal to it. The aberration-free case is obtained when b = p 2.
2.15.4 Coherent ESF

Replacing the OTF by the CTF in Eq. (1-233), the coherent ESF for an edge object
making an angle b with the yo axis is given by
1
1 1 Û 1
ESF( x ¢) = -
2 2 pi Ù
ı
{[
exp i F(x ¢, 0) - px ¢ x ¢
x ¢
dx ¢ . ]} (2-316)
-1
For an even aberration, Eq. (2-316) can be written in terms of its real and imaginary parts
as
ESF( x ¢) = Re ESF( x ¢) + i Im ESF( x ¢) , (2-317)
where
1
1 1Û 1
Re ESF( x ¢) = + Ù cos F(x ¢, 0) sin( px ¢ x ¢) dx ¢ (2-318a)
2 pı x¢
0
and
1
1Û 1
Im ESF( x ¢) = Ù sin F(x ¢, 0) sin( px ¢ x ¢) dx ¢ . (2-318b)
pı x¢
0
We note that
Re ESF( x ¢) = 1 - Re ESF( - x ¢) (2-319a)
and
Im ESF( x ¢) = - Im ESF( - x ¢) . (2-319b)
Moreover, Re ESF(0) = 0.5 and Im ESF(0) = 0, regardless of the value of the aberration.
For an odd aberration function, the ESF is real given by
1
1 1Û 1
ESF( x ¢) =
2 pı
[
- Ù sin F(x ¢, 0) - px ¢ x ¢
x¢
dx ¢ . ] (2-320)
0
For an aberration-free system, Eq. (2-316) reduces to

1
1 Û sin( px x )
ESF( x ) = + dx , (2-321)
2 Ùı px
0
independent of the orientation of the edge object. The ESF is real and obeys Eq. (2-319a).
Equation (2-321) can also be obtained by integrating the LSF from - • to x, as may be
seen from Eq. (1-228). The aberration-free ESF is shown in Figure 2-72. Its maxima and
minima occur at integral values of x. Its value at x = 0 is 0.5, and its principle maximum
lies at at x = 1 with a value of 1.09. It approaches unity as x Æ • and 0 as x Æ - •.
The irradiance distribution associated with the aberation-free ESF is shown in Figure 2-
73. It has a value of 0.25 at x = 0 and a value of 1.19 at the principle maximum at x = 1.
It is evident from the figure that the edge appears to be shifted toward the bright region.
There are strong fluctuations in the irradiance distribution on the bright side of the
Gaussian image of the edge. Such fluctuations are absent in the incoherent ESF shown in
Figure 2-55. Moreover, the spatial period of the fluctuations on the dark side (shadow
region) is half of that on the bright side. It results from the squaring operation of the
positive and negative fluctuations of the amplitude ESF on the dark side.
Figure 2-74 shows the real and imaginary parts and the irradiance distribution
associated with a defocused ESF. As expected for an even aberration, the real part has a
value of 0.5 at x = 0 and the imaginary part is zero, regardless of the value of defocus.
Accordingly, the value of irradiance at the Gaussian image of the edge is 0.25
independent of the defocus value. The irradiance distributions do not bear any symmetry
about x = 0 , unlike the incoherent ESF (compare with Figure 2-55). The ring pattern
shifts more and more to the right as defocus increases. The ESF aberrated by half wave of
coma is shown in Figure 2-75a. It is real, as expected for an odd aberration. The
corresponding irradiance distribution is shown in Figure 2-75b. We note that the ring
pattern is strongly dependent on the orientation of the edge. When b = p 2, for example,
an aberration-free image is obtained.
0.8
0.6
ESF (x)
0.4
0.2
–6 –4 –2 0 2 4 6
x
Figure 2-72. Aberration-free coherent ESF. x is in units of l F. The dotted curve

represents the Gaussian image of the edge object.
1.2
0.8
ESF2 (x)
0.6
0.4
0.2
–6 –4 –2 0 2 4 6
x
(a)
100
–1
10
ESF2 (x)
–2
10
–3
10
–4
10
–6 –4 –2 0 2 4 6
x
(b)
Figure 2-73. Irradiance distribution associated with the aberration-free coherent

ESF. x is in units of l F. (a) Linear scale. (b) Log scale.
1.2
1/2 1
Bd = 0
1
0.8
Re ESF (x)
0.6
(a)
0.4
0.2
– 0.2
–3 –2 –1 0 1 2 3 4 5 6
x
0.3
0.2
0.1
Im ESF (x)
Bd = 0
0
(b)
– 0.1
– 0.2 1/2 1
– 0.3
– 0.4
–3 –2 –1 0 1 2 3 4 5 6
x
1.4
1/2 1
1.2
Bd = 0
1
ΩES F (x)Ω2
0.8
(c)
0.6
0.4
0.2
0
–3 –2 –1 0 1 2 3 4 5 6
x
Figure 2-74. Defocused coherent ESF. (a) Real part. (b) Imaginary part. (c)
Irradiance distribution associated with the ESF. The curve Bd = 0 represents the
aberration-free ESF. Bd is in units of l and x is in units of l F.
1.2
p/2
1 p/4
b=0
0.8
Ac = 0.5
0.6
ESF (x)
0.4
0.2
– 0.2
– 0.4
–3 –2 –1 0 1 2 3 4 5 6
x
(a)
1.4
Ac = 0.5
1.2 p/2
1
p/2
ESF2 (x)
0.8 b=0
0.6
0.4
0.2
0
–3 –2 –1 0 1 2 3 4 5 6
x
(b)
Figure 2-75. Coherent ESF aberrated by coma Acr3 cosq , where Ac = 0.5 l . (a) ESF,
which is real. (b) Irradiance distribution associated with the ESF. The edge object
lies along a direction making an angle b with the yo axis and the variation with x
shown here is along a direction that is normal to it. The aberration-free case is
obtained when b = p 2.
2.15.5 Image of a Coherent Disc

As an example of an extended incoherent object, we considered the image of an
incoherent disc in Section 2.14.3. We showed that the image generally resembles the
object. It is bright at and near the center and becomes dimmer away from it. The central
irradiance of the image increases as the size of the disc increases. As defocus increases,
the central irradiance decreases. Now we consider the same problem except that the disc
is coherently illuminated. We show that, unless the disc is very small, the central
irradiance of the image does not necessarily increase as the size of the disc increases.
Moreover, defocus may increase the central irradiance compared to the corresponding
aberration-free value. Thus, with or without defocus, the image may be distorted.
Diffraction Image
The Gaussian image of a disc is independent of its coherence. Let this image be of
radius hg with a uniform amplitude Ag . Its spatial frequency spectrum is given by
hg
r Û
(
Ũ g (vi ) = 2 p Ag Ù J 0 2 pvi rg rg drg
ı
)
0
(
= Ag hg vi J1 2pvi hg ) ( ) . (2-322)
r
The spectrum Ũi (vi ) of the diffraction amplitude image of an extended coherent object is
given by Eq. (1-252). Accordingly, the amplitude distribution of the diffraction image is
given by
r r r r r
Ui (ri ) = Ú U˜ i (vi ) exp( - 2pivi ◊ ri ) dvi
r r r r r
= Ú U˜ g (vi ) CTF(vi ) exp( - 2pivi ◊ ri ) dvi . (2-323)
r
Substituting for Ũ g (vi ) from Eq. (2-322) and assuming a radially symmetric aberration
( )
F rp and, therefore, a radially symmetric CTF, Eq. (2-323) reduces to
1 lF 2p
Û Û
Ui (r ) = Ù Ú U˜ g (vi ) CTF(vi )vi dvi Ù exp( - 2 pivi ri cos q) d q
ı ı
0 0
1
Û
( )
= p Ag bg Ù J1 pvbg J 0 ( pvr ) exp[iF(v)] dv ,
ı
(2-324)
0
where in the last step we have used normalized quantities r = ri l F , bg = hg l F , and

v = vi (1 2l F ) . The irradiance distribution of the image is given by
2
Ii ( r ) = Ui ( r ) . (2-325)
The central irradiance as a function of the disc radius is given by

2
1
Û
ı
( )
Ii (0) = p 2 Ag2 bg2 Ù J1 pvbg exp[iF(v)] dv . (2-326)
0
For very small values of bg , e.g., a pinhole, J1 pvbg ( ) ~ pvbg 2 and Eq. (2-324) reduces
to
1
Û
( 2
Ui ( r ) = p 2 ) Ag bg2 Ù
ı
J 0 ( pvr ) exp[iF(v)] v dv . (2-327)
0
The integral on the right-hand side is proportional to the zero-order Hankel transform of
the CTF(v) [see, for example, Eq. (2-49b)]. Hence, Eq. (2-327) may be written
Ui (r ) = Ag bg2 CSF(r ) . (2-328)
If we take the modulus square of both sides, we obtain the incoherent PSF, as expected,
since the distinction between a coherent and an incoherent disc disappears as it gets
smaller. Thus, a pinhole, whether coherently or incoherently illuminated, is a source of
coherent illumination.
Numerical Results
[
Figure 2-76a shows the central irradiance of the defocused image F(v) = Bd v 2 of a ]
coherent disc as a function of its radius. The irradiance is normalized by Ag2 . We note
that for small discs bg £ 1.3 , the aberration-free central irradiance is higher than a
corresponding defocused value. However, for larger discs, the aberration-free value is not
necessarily higher than a defocused value. Their relative values depend on the disc radius.
Figure 2-76b shows the axial irradiance of a disc of various sizes. When the disc is very
small, e.g., bg = 1 4, it behaves like a point source. Note that the central value in this case
is normalized to be 4 to accentuate the secondary maxima. For a small disc, the irradiance
is maximum at the center of its image. For a large disc, the maximum occurs at a
defocused point, as evidenced by the curves for bg = 2 and 4.
Figure 2-77a shows the irradiance distribution of the defocused images for bg = 1.
The aberration-free image looks more like the Gaussian image compared to the defocused
images. Figure 2-77b shows the image distributions for bg = 2 . The aberration-free
distribution has a dip in the middle of the image and does not resemble the Gaussian
image. It is also clear that the central value for a defocus of one wave is much larger than
the corresponding aberration-free value, contrary to the case of an incoherent disc
[Compare with Figure 2-63b]. It is a coincidence that, for the disc size under
consideration, the image distribution for a half wave of defocus is similar to that for one
wave of defocus. They are quite different, for example, for bg = 2.3 , as may be seen from
3.5 1
2.5 1/2
Ii (0, bg)
2
Bd = 0
1.5
1/4
0.5
0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
bg
(a)
4.5
bg = 1/4
3.5
3 2
Ii(0; Bd)
2.5
1
1.5
4
1
0.5
0
0 0.5 1 1.5 2 2.5 3 3.5 4
Bd
(b)
Figure 2-76. (a) Central irradiance of the defocused image of a coherent disc as a
function of its radius. The radius bg of the Gaussian image is in units of l F and the
defocus coefficient Bd is in units of l . (b) Axial irradiance of a coherent disc of
various sizes. The central value in the case bg = 1 4 is normalized to 4 to accentuate
the secondary maxima. The actual central values in increasing order of bg are 0.02,
1.70, 0.61, and 0.71.
1.8
bg = 1
1.6 Bd = 0
1.4
1.2 1/4
1 Gaussian
Ii (r)
image (a)
0.8
1/2
0.6
0.4
0.2
1
0
0 0.5 1 1.5 2 2.5 3 3.5 4
r
3
bg = 2
2.5 1/2
1
2
Ii (r)
1.5 (b)
1/4
1
Gaussian
image
Bd = 0
0.5
0
0 0.5 1 1.5 2 2.5 3 3.5 4
r
3.5
bg = 2.3
1
2.5
2
Ii (r)
1/2
(c)
1.5
1/4
1
Gaussian
image
0.5 Bd = 0
0
0 0.5 1 1.5 2 2.5 3 3.5 4
r
Figure 2-77. Irradiance distribution of the defocused images of a coherent disc. (a)
The disc radius is bg = 1 in units of l F. (b) bg = 2 . (c) bg = 2.3 . The Gaussian
image is also indicated in each figure.
Figure 2-77c. It is evident that, whereas the image of an incoherent disc resembles the
object, the image of a coherent disc can be quite different from it.
This may all be explained qualitatively as follows. If we divide the disc into radial
zones (similar to the Fresnel zones), the parts of a given zone contribute to the central
amplitude in phase. However, different zones do not contribute in phase (unless the phase
difference between two zones happens to be 2p ). The complex amplitudes contributed
by different zones can partially cancel each other. This cancellation may be reduced by
the defocus aberration, resulting in a larger defocused central irradiance. The relative
phase of a zone does not matter in the case of an incoherent disc, because it is the
irradiance contributions from different zones that are added to obtain the net irradiance.
Hence, the central irradiance increases as the radius of an incoherent disc increases.
2.15.6 Use of a Lens for Obtaining Fourier Transforms

In Section 1.2.2, we showed that a Fourier transform of a certain complex amplitude
distribution of a propagating wave is obtained in its Fraunhofer region. Thus, if an object
transparency of radius a is illuminated by a plane wave, a Fourier transform of its
amplitude transmittance function is obtained in the region z > ka 2 / 2 . The larger the
value of z, the better the approximation of obtaining a Fourier transform. Because of the
large value of k, the distance z beyond which a Fourier transform is obtained can be quite
large. This distance can be reduced by the use of a lens. The Fourier transform is obtained
in the focal plane of the lens, where the plane wave illuminating the object transparency
is focused by it. It should be clear, though, that the Fourier transformation is a
characteristic of the wave propagation and not a property of the lens. The lens merely
collapses the Fraunhofer region to its focal plane. Of course, an image of the transparency
is obtained in the image plane according to Gaussian optics. However, this image can be
altered by spatial filtering of the complex amplitude representing the spatial-frequency
spectrum of the object in the focal plane.
r
If an object of transmittance Uo ( ro ) illuminated by a plane wave of uniform
amplitude is placed against a lens of focal length f, then the amplitude of the wave
r r r
( )
transmitted by it may be written Uo ( rl ) exp - ikrl2 2 f , where rl ∫ ro in the case of a
thin lens, or a multielement lens system if diffraction within it is neglected. We assume
that the lens is aberration free in that it focuses the plane wave without any aberrations.
According to Eq. (1-22a), the amplitude at a distance z in the Fresnel region is given by
r -i È Ê r2 ˆ ˘ Û r È ikr 2 Ê 1 1 ˆ ˘ Ê 2 pi r r ˆ r
U (r ; z) = exp Íik Á z + ˜ ˙ Ù Uo ( rl ) exp Í l Á - ˜ ˙ exp Á - r ◊ rl ˜ drl .
lz ÍÎ Ë 2 z ¯ ˙˚ ı Î 2 Ë z f ¯ ˚ Ë lz ¯
(2-329)
If we let z = f , we obtain
(r ) r
-i È Ê rf2 ˆ ˘ Û Ê 2 pi r r ˆ r
U rf ; f = exp Íik Á f + ˜ ˙ Ù Uo ( rl ) exp Á - rf ◊ rl ˜ drl , (2-330)
lf ÍÎ Ë 2 f ¯˙ ı Ë lf ¯
˚
r
where rf is the position vector of a point in the focal plane of the lens. We see that except
for a phase factor outside the integral, the amplitude in the focal plane of the lens is the
Fourier transform of the object.
Similarly, if we place an object transparency in a converging spherical wave, such as

formed by a lens when focusing a plane wave, the amplitude of the wave transmitted by
r
( )
the object is given by Uo ( ro ) exp - ikro2 2 d , where d is the distance of the center of
curvature of the incident wave from the object plane. According to Eq. (1-22a), the
amplitude of the wave at a distance z is given by
r -i È Ê r2 ˆ ˘ Û r È ikr 2 1 1 ˘ Ê 2 pi r r ˆ r
U (r ; z) = exp Íik Á z + ˜ ˙ Ù Uo ( ro ) exp Í o Ê - ˆ ˙ exp Á - r ◊ ro ˜ dro .
lz ÍÎ Ë 2 z ¯ ˙˚ ı Î 2 Ë z d ¯ ˚ Ë lz ¯
(2-331)
If we let z = d , we obtain
r -i È Ê r2 ˆ ˘ Û r Ê 2 pi r r ˆ r
U ( rd ; d ) = exp Íik Á d + d ˜ ˙ Ù Uo ( ro ) exp Á - rd ◊ ro ˜ dro , (2-332)
ld ÍÎ Ë 2 d ¯ ˙˚ ı Ë ld ¯
r
where rd is the position vector of a point in the observation plane at a distance d from the
object plane. Again, except for a phase factor outside the integral, the amplitude in the
plane of convergence of the spherical wave is given by the Fourier transform of the
object.
We can also obtain the Fourier-transform relationship from the imaging equation by
determining the amplitude in a defocused plane. As illustrated in Figure 2-78, consider an
object illuminated by a plane wave and imaged by a thin lens. The Fourier transform of
the object is obtained in the focal plane of the lens, while its image is obtained in the
image plane determined by Gaussian optics. If the object is illuminated by a spherical
wave, the Fourier transform is obtained in a plane where the focus of the spherical wave
is imaged by the lens. We also illustrate how the image can be altered by spatial filtering
certain frequency components in the Fourier-transform plane.
From Eq. (1-236), the complex amplitude in the focal plane of a lens of focal length f
is given by
r i È Ê rf2 ˆ ˘ Û r r Ê r2 ˆ
(
U f rf ; f = - )
lf
Í
exp ik Á f - zg +
ÍÎ Ë
˜ ˙ Ù d ro Uo ( ro ) exp Á - ik o ˜
2 f ¯˙ ı Ë 2 zo ¯
˚
r r
Û r r Ê ik 2 ˆ È 2 pi r Ê rf M ro ˆ ˘ r
Ù
( )
¥ Ù G rp ; ro exp Á - rp ˜ exp Í-
Ë 2 zo ¯ ÍÎ l
rp Á -
Ë f
◊ ˙ d rp
zg ˜¯ ˙
, (2-333)
ı ˚
where
1 1 1
- = (2-334)
z g zo f
(–)zo zi
Object plane Lens Fourier-transform Image plane

plane
Figure 2-78. Imaging of an object illuminated by a plane wave. Fourier transform of

the object lies in the focal plane of the lens.
and the magnification of the image is given by
zg
M = . (2-335)
zo
Note that according to our sign convention (see Section 1.3.2 of Part I), zo is numerically
negative. If the system is aberration free and the lens is quite large, we may replace
r r
( )
G rp ; ro by a constant that is inversely proportional to the distance of the object zo and
extend the region of integration in Eq. (2-333) to infinity. Therefore, the integral in this
equation becomes the (inverse) Fourier transform of an imaginary Gaussian function
r r
( )
evaluated at the spatial frequency rf f -1 - ro zo-1 l . Hence, within a constant, Eq. (2-
333) reduces to
r 1 È ik Ê 2
zo ˆ r f ˘ Û r Ê ik r r ˆ r
(
U f rf ; f ) =
f
exp Í Á + ˜ ˙ Ù Uo ( ro ) exp Á - rf ◊ ro ˜ d ro
ÍÎ 2 Ë
1
f ¯ f ˙˚ ı Ë f ¯
, (2-336)
where we have dropped the unessential phase factors. Thus, the complex amplitude in the
back focal plane of the lens is proportional to a Fourier transform of the object. The phase
factor varying as rf2 vanishes when the object is placed in the front focal plane of the lens
(so that zo = - f ). The image is formed at infinity in this case, and a second lens must be
used to observe the image in its focal plane. The finite size of the pupil causes some
vignetting, which can be avoided by placing the object directly against the lens. (Based
on the angular spectrum concept, vignetting is equivalent to the lens missing certain high
spatial frequencies in the object.) One must contend with the quadratic phase factor in
that case, as may be seen by letting zo Æ 0 in Eq. (2-336) (unless the transform is
received on a spherical surface). A measurement of the irradiance distribution yields the
power spectrum of the object in either case. It should be noted that in order to obtain
high-quality Fourier transform and image of an object, the lens must be aberration free
for forming the image of the source illuminating the object (e.g., aberration-free focusing
of the plane wave) and for forming the image of the object; i.e., it must be aberration free
for two conjugates.
To demonstrate how the image of an object can be altered by spatial filtering its
Fourier transform, we consider a wire mesh as an object illuminated by a plane wave, as
illustrated in Figure 2-78, and first explained by Porter55 nearly a century ago. A
diffraction pattern or a Fourier transform of the wire mesh is seen in the focal plane of the
lens, as illustrated in Figure 2-79a. Each spot in the focal plane is actually surrounded by
the diffraction rings characteristic of the mesh aperture. As expected, the image of the
wire mesh in the image plane consists of both vertical and horizontal lines. Figures 2-
79b-e show how the image changes if a mask is used in the focal plane to spatially filter
out certain parts of the Fourier transform. If the mask transmits only the central horizontal
spots (which are generated by the vertical wires of the mesh), as in Figure 2-79b, then the
image consists of only vertical lines. Similarly, if only the central vertical spots are
transmitted, as in Figure 2-79c, then the image consists of only the horizontal lines. If the
central spot is also blocked, then the image undergoes a contrast reversal; the wire images
appear as bright horizontal or vertical lines with wide dark spaces between them. If the
central spot and only the second spot on each side of it are transmitted, then the spacing
of the image lines is reduced by a factor of two. If only the spots on a central diagonal are
transmitted, as in Figure 2-79d, then the image consists of diagonal lines that are absent
in the object. Similarly, if the orthogonal central diagonal spots are transmitted, as in
Figure 2-79e, the image consists of orthogonal diagonal lines, which are also absent in the
object. If both sets of diagonal spots are transmitted, as in Figure 2-79f, the image appears
as if the object has been rotated by 45o .
Lenses used to form and observe Fourier transforms of objects are called Fourier
lenses, and the process of selectively masking portions of a Fourier transform is called
spatial filtering. Indeed, this field has come to be known as Fourier optics.
2.15.7 Comparison of Coherent and Incoherent Imaging
Now we briefly compare coherent and incoherent imaging. We have already seen
how different the images of a line, an edge, and a disc are under coherent and incoherent
illuminations (compare Figure 2-52 with 2-69; 2-55 with 2-73; and 2-63 with 2-77). Here
we compare the frequency spectra and two-point resolution of the coherent and
incoherent images of an object. We illustrate that we can not a priori assert that one type
of image is always better than the other.
Frequency Spectra of Images
We have seen that whereas the coherent transfer function (CTF) of an imaging
system is simply a scaled version of its pupil function, its incoherent transfer function
(OTF) is equal to the autocorrelation of its pupil function. Accordingly, the cutoff
frequency of the CTF is half that of the OTF, regardless of the amplitude or phase
(a)
(b)
(c)
(d)
(e)
(f)
Fourier-transform Image plane

plane
Figure 2-79. Spatial filtering of the spectrum of an object consisting of a wire mesh
and the corresponding changes in its image. A mask selectively transmits a certain
portion of the object spectrum and filters the rest. (a) No mask. (b) Central
horizontal spots transmitted, giving an image consisting of vertical lines. (c) Central
vertical spots transmitted, giving an image consisting of horizontal lines. (d)
Diagonal spots transmitted, giving an image consisting of diagonal lines absent in
the object. (e) The orthogonal diagonal spots, again giving an image with orthogonal
lines that are absent in the object. (f) Both central diagonal spots transmitted, giving
an image as if the object has been rotated by 45o .
variations across the pupil. Thus, the bandwidth of frequencies transmitted by a system is
twice as wide for an incoherently illuminated object as for the one illuminated coherently.
However, an examination of the spectra of the observed irradiance images of coherently
and incoherently illuminated objects shows that it is not a simple matter to decide which
image is better.
Coherent imaging is linear in complex amplitude and, as illustrated by Eq. (1-242),

the complex amplitude distribution of the image of a coherently illuminated object is
given by the convolution of the amplitude of the Gaussian image and the CSF:
r r r
Ui ( ri ) = U g ( ri ) ƒ CSF( ri ) , (2-337)
where ƒ indicates a convolution. The irradiance distribution of the image is given by
r r r 2
Ii ( ri ) = U g ( ri ) ƒ CSF( ri ) . (2-338)
Fourier transforming both sides, we obtain the image spectrum:

r r r r r
[ ][
I˜i ( vi ) = U˜ g ( vi ) CTF( vi ) * U˜ g ( vi ) CTF( vi ) ] , (2-339)
where * represents an autocorrelation.
Incoherent imaging is linear in irradiance and, as illustrated by Eq. (1-56c), the

irradiance distribution of the image of an incoherent object is given by the convolution of
the Gaussian image and the PSF:
r r r
Ii ( ri ) = I g ( ri ) ƒ PSF( ri )
r 2 r 2
= U g ( ri ) ƒ CSF( ri ) . (2-340)
Fourier transforming both sides, we obtain the spectrum of the image:

r r r r r
[ ][
I˜i ( vi ) = U g ( vi ) ƒ U g ( vi ) CTF( vi ) ƒ CTF( vi ) ] . (2-341)
Comparing Eqs. (2-339) and (2-341), it is evident that the spectra of the coherent and
incoherent irradiance images are different from each other. However, we can not
conclude that one image is better than the other.
As an example of incoherent and coherent images of an extended object with a given

imaging system, we show diffraction-limited images of the Taj Mahal in Figure 2-80. For
clarity, a magnified image of its boxed portion only is shown. The incoherent image,
corresponding to a solar-illuminated Taj, is definitely better than the coherent image
corresponding to a laser-illuminated Taj. The edge ringing associated with coherent
imaging is quite evident in the coherent image. If a diffuser is placed between the laser
and the object, the image breaks up into small random spots called speckles.
Object
Incoherent Image
Coherent Image
Figure 2-80. Incoherent and coherent images of an extended object. The object is the
Taj Mahal built by the Mughal Emperor Shah Jahan. For clarity, a magnified
image of its boxed portion only is shown. The edge ringing associated with coherent
imaging is quite evident in the coherent image.
Two-Point Resolution
A measure of the imaging quality of a system is its ability to resolve closely spaced
objects. For astronomical objects (stars), Rayleigh defined a criterion of resolution that
two point objects of equal intensity are just resolved if the principal maximum of the Airy
pattern of one of them falls on the first zero of the other.56 Thus, two incoherent object
points are just resolved if their angular separation is 1.22 l D, or the separation between
their Gaussian images is 1.22 l F . Let the Gaussian images be located at x = ± 0.61 . The
irradiance distribution of the image along the x axis is given by
2 2
ÏÔ 2 J [ p( x - 0.61)] ¸Ô ÏÔ 2 J1[ p( x + 0.61)] ¸Ô
I( x) = Ì 1 ˝ +Ì ˝ , (2-342)
ÔÓ p( x - 0.61) Ô˛ ÔÓ p( x + 0.61) Ô˛
where x is in units of l F . This distribution, which is symmetrical about x = 0 , is shown

in Figure 2-81a. We note that there is a dip in the irradiance at the center. The central
value is 0.73, compared to a maximum value of unity at x = ± 0.61 . The 2-D image of
the two-point object is shown in Figure 2-81b.
For two coherently illuminated point objects separated by the Rayleigh resolution,
the corresponding irradiance distribution is given by
2
2 J1[ p( x - 0.61)] 2 J1[ p( x + 0.61)]
I( x) = + exp(id ) , (2-343)
p( x - 0.61) p( x + 0.61)
where d is the phase difference between the two object points. This distribution is also
symmetrical about x = 0 . It is shown in Figure 2-82 for three different values of d . We
note that, since J1 (1.22 p) = 0 , I ( ± 0.61) = 1, regardless of the value of d . There is no dip
in the irradiance when the point objects are in phase; hence, they can not be resolved.
However, when their phases are opposite of each other, then the irradiance at the center is
zero, making it easier to resolve them. When d = p 2 , then the irradiance distribution is
the same as for two incoherent point objects. Thus, the ability to resolve points of a
coherently illuminated object is strongly dependent on the phase relation among them.
Now we consider the effect of defocus on the images and show that it is possible to
obtain an asymmetric image of the two coherent point objects. 57, 58 From Eq. (2-308), the
complex amplitude distribution of the defocused image of a point object can be written
1
Û
[ ]
A(r ) = 2 Ù exp iF(r) J 0 ( p r r) r dr ,
ı
(2-344)
0
where F(r) = Bd r2 is the defocus phase aberration, Bd being its peak value in radians.
Hence, the irradiance distribution of the defocused image of two incoherent point objects
0.9
0.8
0.7
0.6
I (x)
0.5
0.4
0.3
0.2
0.1
0
–4 –3 –2 –1 0 1 2 3 4
x
(a)
(b)
Figure 2-81. (a) Irradiance distribution along the x axis of the image of two
incoherent point objects separated by the Rayleigh resolution of 1.22ll F . The
central value is 0.73. (b) 2-D image.
1.5
d=0
p/2
I (x)
0.5
0
–4 –3 –2 –1 0 1 2 3 4
x
Figure 2-82. Irradiance distribution of the image of two coherent point objects
separated by the Rayleigh criterion of resolution 1.22ll F with a relative phase of d .
When d = p 2 , the distribution is the same as for two incoherent point objects. The
central values are 1.47, 0.74 and 0. The principal peaks have a value of 1.11 at x =
0.78 when d = p .
along the x axis becomes
È 1 2
ÍÛ
[ ]
I ( x ) = 4 Í Ù exp iF(r) J 0 [ p( x - 0.61)r] r dr
Í ı0
Î
2˘
1
Û ˙
[ ]
+ Ù exp iF(r) J 0 [ p( x + 0.61)r] r dr ˙ .
ı
(2-345)
0
˙
˚
The distribution is symmetric about x = 0 , and it is shown in Figure 2-83 for defocus
wave aberrations of Bd = 0 , l 4 , and l 2. The peak value is given by I ( ± 0.61) = 1,
0.84, and 0.48 for these values of Bd . The irradiance in the central region decreases and
increases in the sourrounding region. Figure 2-83b shows that when normalized to the
same peak value, the central dip lowers slightly with increasing value of Bd .
The corresponding irradiance distribution of the defocused image of two coherent

point objects with a phase difference of d between them is given by
1.0
Bd = 0
0.8
1/4
0.6
I (x)
0.4
1/2
0.2
0
–4 –3 –2 –1 0 1 2 3 4
x
(a)
0.8
0
1/4
1/2
0.6
I (x)
0.4
0.2
Bd = 1/2
1/4
0
0
–4 –3 –2 –1 0 1 2 3 4
x
(b)
Figure 2-83. (a) Irradiance distribution of the defocused image of two incoherent
point objects separated by the Rayleigh criterion of resolution 1.22ll F . The central
values in increasing order of defocus are 0.73, 0.61, and 0.33. The principal peaks
have a value of 1, 0.84, and 0.48 at x = ± 0.61. Bd is in units of l . (b) The
distributions are normalized to unity at the principal peaks.
È 1
Û
I ( x ) = 4 ÍÍ Ù exp iF(r) J 0 [ p( x - 0.61)r] r dr
[ ]
ı
ÍÎ 0
2˘
1
Û ˙
[ ]
+ exp(id ) Ù exp iF(r) J 0 [ p( x + 0.61)r] r dr ˙ .
ı
(2-346)
0
˙
˚
This distribution is shown in Figure 2-84. It is symmetric for d = 0 and p . The central
value decreases with defocus in Figure 2-84a, and the side peaks decrease in value in
Figure 2-84b (the central value is zero in this case). However, the defocused distribution
is quite asymmetric when d = p 2 . Although the presence of two point objects can be
inferred from this figure, one would incorrectly infer that they are of unequal intensity. If
the defocus aberration is negative, as when z > R , then the object appearing dimmer in
Figure 2-84c will lie on the right-hand side of the origin.
The 2-D defocused images of two incoherent and coherent point objects separated by
the Rayleigh resolution are shown in Figures 2-85 and 2-86. An example of an incoherent
two-point object is a double star observed by a telescope. In principle, one should be able
to resolve the image of two points that are closer to each other than the Rayleigh
resolution. However, in practice, the aberrations of the imaging system and the noise in
the image recording device (e.g., a film or an array of photodetectors) will limit the
resolution of a system whether the objects are coherent or not. 59
1.5
d=0
Bd = 0
1/4
1
I (x)
(a)
1/2
0.5
0
–4 –3 –2 –1 0 1 2 3 4
x
Figure 2-84. Irradiance distribution of defocused images of two coherent objects of

equal intensity and a phase angle d . (a) d = 0 . (b) d = p . (c) d = p 2.
1.4
d=p
1.2
Bd = 0
1
0.8 1/4
I (x)
(b)
0.6
1/2
0.4
0.2
0
–4 –3 –2 –1 0 1 2 3 4
x
1.4
d = p/2
1.2
0.8 Bd = 0
I (x)
(c)
0.6
1/4
0.4
0.2
1/2
0
–4 –3 –2 –1 0 1 2 3 4
x
Figure 2-84. Irradiance distribution of defocused images of two coherent objects of

equal intensity and a phase angle d . (a) d = 0 . (b) d = p . (c) d = p 2. The central
values in (a) in increasing order of defocus are 1.47, 1.22, and 0.67. The central value
in (b) is zero. The principal peaks have a value of 1.11, 0.95, and 0.58 at x = ± 0.77 .
In (c), the principal peaks have a value of 1.13 at x = 0.60 , 1.00 at x = 0.61, and 0.82
at x = 0.59 ; and 1 at x = - 0.61 , 0.55 at x = - 0.66 , and 0.15 at x = - 0.80 . The
minima near the center have values of 0.74 at x = 0, 0.50 at x = - 0.28 , and 0.13 at
x = - 0.45 .
Bd = 0 Bd = 0
Bd = 1 4 Bd = 1 4
Bd = 1 2 Bd = 1 2
(a) Incoherent points (b) Coherent points in phase
Figure 2-85. Aberration-free and defocused images two object points separated by
the Rayleigh resolution. (a) Incoherent points. (b) Coherent points in phase ( d = 0 ).
Bd = 0 Bd = 0
Bd = 1 4 Bd = 1 4
Bd = 1 2 Bd = 1 2
(a) (b)
Figure 2-86. Aberration-free and defocused images of two coherent point objects
with a phase difference of (a) d = p . (b) d = p 2.
References 253
REFERENCES
1. For properties of Bessel functions, the reader may refer to G. N. Watson, Treatise on
the Theory of Bessel Functions (Cambridge U. P., New York, 1944).
2. G. B. Airy, “On the diffraction of an object-glass with circular aperture,” Trans.

Camb. Phil. Soc. 5, 283–291 (1835).
3. Lord Rayleigh, Phil. Mag. (5)11, 214 (1881); also his Scientific Papers (Dover, New
York, 1964), Vol. 1, p. 513.
4. V. N. Mahajan, “Included power for obscured circular pupils,” Appl. Opt. 17, 964–
968 (1978).
5. V. N. Mahajan, “Strehl ratio for primary aberrations: some analytical results for
circular and annular pupils,” J. Opt. Soc. Am. 72, 1258–1266 (1982), Errata, 10, 2092
(1993); “Strehl ratio for primary aberrations in terms of their aberration variance,” J.
Opt. Soc. Am. 73, 860–861 (1983).
6. Lord Rayleigh, Phil. Mag. (5) 8, 403 (1879); also in his Scientific Papers (Dover,
New York, 1964) Vol. 1, p. 432.
7. W. B. King, “Dependence of the Strehl ratio on the magnitude of the variance of the
wave aberration,” J. Opt. Soc. Am. 58, 655–661 (1968).

discussion of the geometrical aberrations,” Physica 10, 679–692 (1943); “The
diffraction theory of optical aberrations. Part II: Diffraction pattern in the presence of
small aberrations,” Physica 13, 605–620 (1947); and K. Nienhuis and B. R. A.
Nijboer, “The diffraction theory of optical aberrations. Part III: General formulae for
small aberrations: experimental verification of the theoretical results,” Physica 14,
590–608 (1949); also, V. N. Mahajan, “Zernike circle polynomials and optical
aberrations of systems with circular pupils,” Appl. Opt. 33, 8121–8124 (1994); and
“Zernike polynomials and optical aberrations,” Appl. Opt. 34, 8060–8062 (1995).
9. A. B. Bhatia and E. Wolf, “On the circle polynomials of Zernike and related
orthogonal sets,” Proc. Cambridge Philos. Soc. 50, 40–48 (1954).
10. V. N. Mahajan, “Axial irradiance and optimum focusing of laser beams,” Appl. Opt.
22, 3042–3053 (1983).
11. Y. Li and E. Wolf, “Focal shifts in diffracted converging spherical waves,” Optics
Commun. 39, 211–215 (1981).
12. D. S. Burch, “Fresnel diffraction by a circular aperture,” Am. J. Phys. 53, 255–260
(1985).
13. V. N. Mahajan, “Aberrated point spread functions for rotationally symmetric

aberrations,” Appl. Opt. 22, 3035–3041 (1983).
14. S. Szapiel, “Aberration-variance-based formula for calculating point-spread

functions: rotationally symmetric aberrations,” J. Opt. Soc. Am. 25, 244–251 (1986).
15. M. Abramowitz and I. A. Stegun, Eds., Handbook of Mathematical Functions with

Formulas, Graphs, and Mathematical Tables, Dover, New York (1970), p. 887.
16. V. N. Mahajan, “Symmetry properties of aberrated point-spread functions,” J. Opt.

Soc. Am. A11, 1993–2003 (1994).
17. V. N. Mahajan, “Comparison of geometrical and diffraction point-spread functions,”

in International Conference on Optics and Optoelectronics '98, K. Singh, O. P.
Nijhavan, A. K. Gupta, and A. K. Musla, eds., Proc. SPIE 3729, 434-445 (1998).
18. V. N. Mahajan, “Line of sight of an aberrated optical system,” J. Opt. Soc. Am. A2,
833–846 (1985).
19. H. H. Hopkins, “The frequency response of a defocused optical system,” Proc. Roy.
Soc. A231, 91–103 (1955).
20. W. H. Steel, “The defocused image of sinusoidal gratings,” Optica Acta 3, 65-74
(1956).
21. M. De, “The influence of astigmatism on the response function of an optical system,”
Proc. Roy. Soc. A233, 91–104 (1955).
22. N. S. Bromilow, “Geometrical-optical calculation of frequency response for systems

with spherical aberration,” Proc. Phys. Soc. 71, 231–237 (1957).
23. A. S. Marathay, “Geometrical-optical calculation of frequency response for systems

with coma,” Optica Acta 10, 721–730 (1963).
24. R. Barakat, “Computation of the transfer function of an optical system from the
design data for rotationally symmetric aberrations. Part I. Theory,” J. Opt. Soc. Am.
52, 985–991 (1962); “Part II. Programming and numerical methods,” J. Opt. Soc.
Am. 52, 992–997 (1962).
25. R. Barakat and A. Houston, “Transfer function of an optical system in the presence
of off-axis aberrations,” J. Opt. Soc. Am. 55, 1142–1148 (1965).
(London) B52, 449–470 (1957).
27. W. B. King, “Correlation between the relative modulation function and the
magnitude of the wave aberration difference function,” J. Opt. Soc. Am. 59, 692–697
(1969).
28. S. Szapiel, “Hopkins variance formula extended to low relative modulations,” Optica
Acta 33, 981–999 (1986).
References 255
29. H. H. Hopkins, “Geometrical-optical treatment of frequency response,” Proc. Phys.

Soc. B70, 1162–1172 (1957).
30. W. Lukosz, “Der Einfluss der Aberrationen auf die optische Uebertragungsfunktion
bei kleinen Orts-Frequenzen,” Optica Acta 10, 1–19 (1963).
31. R. W. Gostick, “OTF-based optimization criteria for automatic optical design,” Opt.
Quant. Elect. 8, 31–37 (1976).
32. J. Braat, “Polynomial expansion of severely aberrated wave fronts,” J. Opt. Soc. Am.
A4, 643–650 (1987).
33. H. H. Hopkins and B. Zalar, “Aberration tolerances based on the line spread
function,” J. Mod. Opt. 34, 371–406 (1987).
34 R. Barakat and A. Houston, “Line spread function and cumulative line spread
function for systems with rotationally symmetry," J. Opt. Soc. Am. 54, 768–773
(1964).
35. R. Barakat and A. Houston, “Line spread and edge spread functions in presence of
off-axis aberrations,” J. Opt. Soc. Am. 55, 1132–1135 (1965).
36. R. Bracewell, The Fourier Transform and Its Applications (McGraw-Hill, New
York, 1965), p. 262.
37. M. Abramowitz and I. A. Stegun, Eds., Handbook of Mathematical Functions with

Formulas, Graphs, and Mathematical Tables (Dover, New York, 1970), p. 495.
38. H. F. Willis, “A formula for expanding an integral as series,” Philos. Mag. 39, 455–
459 (1948). There is a minus sign missing on the right-hand side of the formula
corresponding to Eq. (2-255).
40. W. S. Kovach, “Energy distribution in the PSF for an arbitrary passband,” Appl. Opt.
13, 1769–1771 (1974).
41. H. S. Dhadwal and J. Hantgan, “Generalized point spread function for a diffraction-
limited aberration-free imaging system under polychromatic illumination,” Opt. Eng.
28, 1237–1240 (1989).
42. L. Levi, “Detector response and perfect–lens–MTF in polychromatic light,” Appl.

Opt. 8, 607–616 (1969).
43. W. Weinstein, “Images of incoherently illuminated bright and opaque disks,” J. Opt.
Soc. Am. 45, 1006–1008 (1955).
44. R. Barakat and A. Houston, “Image of an incoherently illuminated disk,” J. Opt. Soc.
Am. 55, 881–883 (1965).
45. I. L. Goldberg and A. W. McCulloch, “Annular aperture diffracted energy

distribution for an extended source,” Appl. Opt. 8, 1451–1458 (1969).
46. J. Otterman, “Diffraction-limited resolution for geoscene imagery,” Appl. Opt. 8,

1887–1889 (1969).
47. J. Petzval, “Bericht über dioptrische Untersuchungen,” Wien Ber. XXVI, 33-90
(1857).
48. Lord Rayleigh, “On pin-hole photography,” Phil. Mag. XXXI, 87-99 (1891); also in
his Scientific Papers Vol I, 429–440 (Dover, New York, 1965).
49. M. Young, “Pinhole optics,” Appl. Opt. 10, 2763–2767 (1971).
50. X. Jiang, Q. Lin, and S. Wang,“Optimum image plane of the pinhole camera,” Optik,
97, 41–42 (1994).
51. K. Sayanagi, “Pinhole imagery,” J. Opt. Soc. Am. 57, 1091–1099 (1967).
52. G. Reynolds and J. H. Ward, “Coherence theory solution to the pinhole camera,” J.
Soc. Photo. Instr. Eng. 5, 3–8 (1966).
53. R. E. Swing and D. P. Rooney, “General transfer function for the pinhole camera,” J.
Opt. Soc. Am. 58, 629–635 (1968).
54. R. G. Barakat, “Diffraction images of coherently illuminated objects in the presence

of aberrations,” Optica Acta 16, 205–223 (1969).
55. A. B. Porter, “On the diffraction theory of microcopic vision ,” Phil. Mag. 11, 154–
166 (1906).
56. Lord Rayleigh, “Investigations in optics, with special reference to the microscope:
Resoling, or separating, power of optical instruments,” Phil. Mag. VIII, 261-264
(1879); also in his Scientific Papers Vol I, 415–418 (Dover, New York, 1965); “On
the resolving-power of telescopes I,” Phil. Mag. X, 116-119 (1880); also in his
Scientific Papers Vol I, 488–490 (Dover, New York, 1965).
57. T. S. McKechnie, “The effect of defocus on the resolution of two points,” Optica
Acta 20, 253–262 (1973).
58. D. K. Cook and G. D. Mountain, “The effect of phase angle on the resolution of two
coherently illuminated points,” Optical and Quan. Elec. 10, 179–180 (1978).
59. A. J. den Dekker and A. van den Bos, “Resolution: a survey,” J. Opt. Soc. Am. 14,
547–557 (1997).
Problems 257
PROBLEMS
1. Consider a parallel beam of light of wavelength 0.5 mm focused by a lens of

diameter 4 cm and an f-number of 10. Assume that a flux of 1 W is transmitted by
the lens. (a) Calculate the focal-point irradiance. (b) Determine the irradiance at
and the location of the first secondary maximum of the irradiance distribution in
the focal plane. (c) Determine the amount of light in the first bright ring. (d)
Determine the locations of the points along the axis of the beam at which the
irradiance is zero or half of the focal-point irradiance.
2. Consider two point sources with an angular separation of 3 l D imaged by a

system. Let the ratio of their intensities be 10. Determine their image powers on a
pixel of angular size l Dcentered at the image of the dimmer source.
3. Calculate the mean and mean square values of spherical, coma, and astigmatism
aberrations. Show that their standard deviations are given by as in Table 2-4.
Determine the amount of balancing aberration to minimize the variance of each
aberration.
4. Consider spherical, coma, and astigmatism aberrations. Determine their tolerance

(i.e., the value of Ai) for a Strehl ratio of 0.6 [Use the approximate expression S3 for
Strehl ratio]. For each aberration, determine the location and value of the maximum
irradiance when a parallel beam of light is focused by a lens as in Problem 1.
5. (a) Determine the depth of focus of a photographic camera with a lens of diameter 2
cm and an f-number of 5 forming an image at a wavelength of 0.5 mm with a Strehl
ratio of at least 0.8. (b) Now consider a laser transmitter focusing a beam of
diameter 25 cm and a wavelength of 10.6 mm on a receiver at a distance of 1.47
km. Determine the depth of focus for a Strehl ratio of 0.8. Also, determine the point
and the value of maximum irradiance for a transmitted power of 1 kW. Is the
receiver in the near or the far field of the transmitter?
6. (a) Compare the axial irradiance of a beam focused at a distance of D2 l with that
of a corresponding collimated beam. Determine the location and the value of the
principal maximum in the case of the focused beam.
(b) Show that the PSF aberrated by a certain amount of coma rotates by p when
the sign of its aberration coefficient is changed.
7. Consider a point source radiating at a wavelength of 0.5 mm with an intensity of

1 W/sr imaged by a system lying at a distance of 10 m with an entrance pupil of
diameter 20 cm. The exit pupil of the system has a diameter of 10 cm and the focal
ratio of the image-forming light cone is 10. If the system has a quarter wave of
astigmatism, (a) determine the irradiance distribution in the Gaussian image plane
according to geometrical optics (irradiance is proportional to the density of rays).
(b) Describe the irradiance distribution according to diffraction. In particular,
calculate the irradiance at the center of the image. (c) Compare the symmetry
properties of the irradiance distribution obtained according to geometrical and
diffraction optics. (d) Do we obtain identical or different irradiance distributions in
two planes symmetrically defocused from the image plane? Explain. (e) How is the
irradiance distribution in the Gaussian image plane affected if we change the sign
of the aberration?
8. (a) Compare the Strehl ratios, peak irradiances and their locations, and centroid
irradiances and locations of the images of a point object formed by a system with a
half wave of primary coma with those for a half wave of secondary coma. (b)
Estimate the Strehl ratio if both aberrations are present simultaneously. Calculate
the position of the centroid also.
9. Show that a power series expansion of the OTF aberrated by astigmatism in the
presence of defocus is given by
4v
t (v, f) = 1 -
p [ ( ) ] ( )
- 2 Bd2 + Aa2 + 2 Aa Bd cos 2 f v 2 + O v 3 .
Determine the corresponding Hopkins ratio for small values of v in the Gaussian,
sagittal, and tangential image planes.
10. Using Eqs. (1-155), show that the coma-aberrated geometrical OTF for small
spatial frequencies is given by Eqs. (2-240).
11. Transillumination of a system by a pinhole: Consider a pinhole illuminating a

microscope objective, as in scanning microscopy. Let the pinhole be illuminated by
an extended source with a condenser lens. The pinhole images the condenser on the
objective acting as a lens of zero power. (a) Determine the radius of the pinhole so
that it can be treated as point source. (b) Show that the ratio of angular subtenses of
the condenser and the objective on the pinhole must be approximately three to
obtain a central illumination on the objective of approximately 80% of the
maximum possible value. (c) Compare the central illumination thus obtained with
its value when the angular subtenses are equal. (d) What should the ratio be for
maximum illumination if the pinhole is replaced by a large aperture? [See P. N.
Slater and W. Weinstein, “Light transmitted by very small pinholes,” J. Opt. Soc.
Am. 48, 146–149 (1958).]
12. Consider two point objects of equal intensity, phase angle d , and degree of
coherence g located symmetrically about the axis of an imaging system. Obtain the
conditions under which the irradiance distribution of a defocused image along the
line joining their Gaussian image points is symmetric about the optical axis (see
Reference 57).
CHAPTER 4
OPTICAL SYSTEMS WITH GAUSSIAN PUPILS
4.1 Introduction........................................................................................................... 335

4.2 General Theory ....................................................................................................336
4.3 Systems with Circular Pupils............................................................................... 337
4.3.1 Pupil Illumination and Transmitted Power ..............................................337
4.3.2 Aberration-Free System........................................................................... 338
4.3.2.1 PSF ............................................................................................338
4.3.2.2 OTF ........................................................................................... 341
4.3.3 Strehl Ratio and Aberration Tolerance ....................................................343
4.3.4 Balanced Aberrations and Zernike-Gauss Circle Polynomials ............... 344
4.3.5 Defocused System ................................................................................... 348
4.3.5.1 Theory ....................................................................................... 348
4.3.5.2 Axial Irradiance......................................................................... 349
4.3.5.3 Defocused Distribution ............................................................. 350
4.3.5.4 Collimated Beam....................................................................... 352
4.3.6 Weakly Truncated Gaussian Circular Beams ..........................................353
4.3.6.1 Introduction ............................................................................... 353
4.3.6.2 Irradiance Distribution and Beam Radius ................................. 354
4.3.6.3 Imaging of a Gaussian Beam ....................................................359
4.3.6.4 Aberration Balancing ................................................................362
4.3.7 Symmetry Properties of an Aberrated PSF..............................................365
4.4 Systems with Annular Pupils ............................................................................... 366
4.4.1 Theory ......................................................................................................367
4.4.2 Aberration-Free System........................................................................... 368
4.4.3 Strehl Ratio and Aberration Tolerance ....................................................370
4.4.4 Balanced Aberrations and Zernike-Gauss Annular Polynomials ............371
4.4.5 Defocused System ................................................................................... 374
4.4.5.1 Theory ....................................................................................... 374
4.4.5.2 Axial Irradiance......................................................................... 374
4.4.5.3 Defocused Distribution ............................................................. 376
4.4.5.4 Collimated Beam....................................................................... 376
4.4.6 Symmetry Properties of an Aberrated PSF..............................................378
4.5 Line of Sight of an Aberrated System ................................................................. 379
4.5.2.1 Wavefront Tilt ........................................................................... 380
4.5.2.2 Primary Coma ........................................................................... 381
4.5.2.3 Secondary Coma ....................................................................... 382
4.6 Summary................................................................................................................382
333
334 OPTICAL SYSTEMS WITH GAUSSIAN PUPILS
References ......................................................................................................................385
Problems ......................................................................................................................... 386
Chapter 4
Optical Systems with Gaussian Pupils
4.1 INTRODUCTION
In Chapters 2 and 3, we have considered optical systems with uniformly illuminated
(i.e., with uniform amplitude across) circular and annular pupils, respectively. Systems
with nonuniform amplitude across their exit pupils are referred to as apodized systems.
Often, the transmission of a system is made nonuniform by placing an absorbing filter at
its entrance or exit pupil in order to reduce the secondary maxima of its PSF. The word
apodization in Greek means "without feet" implying without or at least reduced
secondary maxima. The purpose in reducing the secondary maxima is to improve the
resolution of the system.
In this chapter, we consider optical systems with Gaussian apodization or Gaussian

pupils, i.e., those with a Gaussian amplitude across the wavefront at their exit pupils,
which may be circular or annular. 1, 2 The discussion of this chapter is applicable equally
to imaging systems with a Gaussian transmission (obtained, for example, by placing a
Gaussian filter at its exit pupil) as well as laser transmitters in which the laser beam has a
Gaussian distribution at its exit pupil. In the case of an imaging system, we are interested
in its PSF, i.e., the irradiance distribution of the image of a point object. In the case of a
laser transmitter focusing a beam on a target, we are interested in the irradiance
distribution in the target plane or the focal plane of the beam.
It is evident that whereas a Gaussian function extends to infinity, the pupil of an

optical system can only have a finite diameter. The net effect is that the finite size of the
pupil truncates the infinite-extent Gaussian function. It is shown that the Gaussian
illumination of the pupil broadens the central disc and reduces the secondary maxima of
the Airy pattern obtained for a uniform pupil. The corresponding OTF is higher for low
spatial frequencies and lower for the high. As in the case of uniformly illuminated pupils,
here too the principal maximum of axial irradiance of a focused beam with a small
Fresnel number lies at a point that is inside and not at the geometrical focus. However,
maximum central irradiance on a target at a fixed distance is still obtained when the beam
is focused on it. If the Gaussian function is very narrow (i.e., its standard deviation is very
small) compared to the radius of the pupil, it is said to be weakly truncated. In essence,
the pupil can be assumed to be also infinitely wide with the result that a Gaussian beam
exiting from the pupil remains Gaussian as it propagates. The standard deviation of an
aberration for a Gaussian pupil is smaller compared to that for a uniform pupil. This is
due to the fact that the wave amplitude decreases as a function of the radial distance from
the center of the pupil but the aberration increases, i.e., the amplitude is smaller where the
aberration is larger. Accordingly, the Strehl ratio for a Gaussian pupil for a given amount
of a primary aberration is higher than that for a uniform pupil, or the aberration tolerance
for a given Strehl ratio is higher for a Gaussian pupil.
335
4.2 GENERAL THEORY

We now give a brief outline of the theory and then apply it to systems with circular
and annular pupils. The pupil function for a system with a Gaussian pupil may be written
(r ) ( ) [ (r )]
P rp = A rp exp i F rp , (4-1)
where
( )
A rp = A0 exp È- rp w ˘ , ( )
2
(4-2)
ÎÍ ˙˚
r
with A0 as a constant describes the Gaussian amplitude at the exit pupil and rp = rp .
The quantity w, called the Gaussian radius, represents the radial distance from the center
of the pupil at which the amplitude drops to e - 1 of the amplitude at the center. The
r r
aberration function F( rp ) represents the phase aberration at a point rp in the plane of the
exit pupil. The amplitude at its center is determined from the total power according to
r r
Pex = Ú A (rp ) d rp
2
. (4-3)
For an aberration-free system, the wavefront at its exit pupil is spherical passing
through its center and with a center of curvature at a distance R from it. This distance
represents the distance between the planes of the exit pupil and the Gaussian image in the
case of an imaging system. In the case of a laser transmitter, it represents the distance at
which the beam is focused at from the plane of the exit pupil. The following discussion is
in the context of a laser transmitter, although, as said above, the results are equally
applicable to an imaging system with a Gaussian pupil.
The irradiance distribution of the diffracted beam normalized by its total power in a
plane at a distance z is given by Eq. (1-62):
r 1 Û r Ê 2 pi r r ˆ r
2
l R ı
(
Ii ( ri ; z ) = 2 2 Ù P rp ; z exp Á -
Ë lR
)
rp ri ˜ d rp
¯
◊ , (4-4)
r
where ri is the position vector of a point in the focal plane with respect to the focal point,
(r ) (r )
P rp ; z = A rp exp iF rp ; z [ (r )] (4-5a)
and
(r )
F rp ; z = F rp + (r ) p Ê1 1ˆ 2
-
l Ë z R¯
rp . (4-5b)
It is assumed here that z ~ R so that their difference enters only through the defocus
aberration. In particular, the variation of the irradiance due to inverse-square law
dependence is negligible. Now we apply the above equations to systems with circular and
annular pupils.
4.3 Systems with Circular Pupils 337
4.3 SYSTEMS WITH CIRCULAR PUPILS

First we consider systems with circular pupils. We show that the Gaussian
illumination broadens the central bright spot of the Airy pattern for a uniform pupil and
reduces the power in its diffraction rings. For a given total power, the central irradiance
for a Gaussian pupil is smaller than that for a uniform pupil. The effect of aberrations on
the Strehl ratio is also smaller in the case of a Gaussian pupil. We discuss balancing of
aberrations to minimize their variances and relate the balanced aberrations to the Zernike-
Gauss circle polynomials whose derivation is outlined. As in the case of a uniform beam
of a small Fresnel number, the principal maximum of axial irradiance of a focused
Gaussian beam also lies at a point that is closer to its pupil than the focal point, but
maximum irradiance on a target at a given distance is obtained when the beam is focused
on it.
Gaussian beams that are very narrow so that their truncation by a lens is negligible,
are shown to remain Gaussian as they propagate. Such a beam also yields maximum
irradiance on a target when it is focused on the target, although its waist and therefore the
principal maximum of axial irradiance lie in a plane that is closer to the lens. It is shown
that an aberration-free beam may be considered weakly truncated when the beam radius
is half of that of the pupil. However, the radius of an aberrated beam must be smaller than
one third of the pupil radius in order that it may be treated as a weakly truncated beam.
Considering the waist of a beam incident on a lens as the object, an imaging equation is
developed in which the waist of the transmitted beam acts as the image. Aberration
tolerances for the weakly truncated beams are shown to be significantly larger due to their
narrowness. Unlike a uniform focused beam, the axial irradiance of a Gaussian focused
beam aberrated by spherical aberration is not symmetrical about any point. However,
their symmetry properties for astigmatism and coma are similar.
4.3.1 Pupil Illumination and Transmitted Power

Let the radius of the exit pupil be a and let the power transmitted by it be Pex . We let
r r
(
r = rp a = r cos qp , sin qp ) (4-6)
and
g = (a w )
2
, (4-7)
where 0 £ r £ 1 and 0 £ q p < 2 p . Thus, Eq. (4-2) may be written
A(r) = A0 exp - g r2 ( ) . (4-8)
[
Pex = A02 ( Sex 2 g ) 1 - exp( - 2 g ) ] , (4-9)
where Sex = p a 2 is the area of the exit pupil. If Pinc is the total power incident on the
pupil, then the fraction transmitted by it is given by
Pex
Ú (
exp - 2 gr2 r dr )
= 0
• = 1 - exp( - 2 g ) . (4-10)
Pinc
Ú (
exp - 2 gr2 r dr )
0
Squaring both sides of Eq. (4-8) and substituting for A02 from Eq. (4-9), we may write the
pupil irradiance distribution in units of Pex Sex :
I (r) = 2 g exp - 2 g r2 ( ) [1 - exp (- 2 g )] . (4-11)
The parameter g characterizes the truncation of a Gaussian beam by the exit pupil. Its
square root represents the ratio of the pupil and beam radii. Large values of g represent
narrow Gaussian beams. Similarly, small values represent wide beams. As g Æ 0 ,
I (r) Æ 1; i.e., a Gaussian beam reduces to a uniform beam. The pupil irradiance
[
decreases from a maximum value of 2 g 1 - exp( - 2 g ) at its center to a minimum ]
[ ]
value of 2 g exp(2 g ) - 1 at its edge. Figure 4-1 shows the pupil amplitude (normalized
to unity at the center) and irradiance distributions for g = 1, 2, and 3. The transmitted
power is 86.47% when a = w , 98.89% when a = 1.5w , and 99.97% when a = 2w .
4.3.2 Aberration-Free System
4.3.2.1 PSF
For simplicity of equations and numerical analysis, we use normalized quantities
r r
r = ri l F = r (cos q i , sin q i ) (4-12)
and
r r
(
I (r ) = Ii (ri ) Pex Sex l2 R 2 ) , (4-13)
where F = R 2 a is the focal ratio of the beam. Note that in Eq. (4-13), the irradiance is
normalized by the focal-point irradiance of a uniform beam of the same total power.
1.0 18
g =1 16
0.8 14
g =3
12
0.6 2 10
A
8
0.4 3 2
6
0.2 4
1
2 0
0.0 0
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
r r
(a) (b)
Figure 4-1. Amplitude and irradiance distributions at the exit pupil for
g = a w = 1 , 2, and 3. (a) Amplitude. (b) Irradiance. The amplitude is normalized
to unity at the center, but the irradiance is in units of Pex Sex .
Substituting Eqs. (4-6), (4-12), and (4-13) into Eq. (4-4), the irradiance distribution
for an aberration-free system may be written
2
1 2p
I ( r; q i ; g ) = p -2
Ú Ú
0 0
[ ( )]
I (r) exp -pirr cos q p - q i r dr dqp . (4-14a)
Carrying out the angle integration by using Eq. (2-12), we obtain

2
È1 ˘
I ( r; g ) = 4 Í Ú I (r) J 0 ( prr) r dr˙ . (4-14b)
ÍÎ 0 ˙˚
The amount of power in the focal plane contained in a circle of radius rc centered at the
focal point is given by
rc
Pi (rc ; g ) = 2 p Ú Ii (ri ; g ) ri dri . (4-15)
0
Substituting Eqs. (4-12), (4-13), and (4-14b) into Eq. (4-15) and defining a normalized
power
P(rc ; g ) = Pi (rc ; g ) Pex , (4-16a)
we obtain
rc
(
P(rc ; g ) = p 2 2 )Ú I (r; g ) rdr , (4-16b)
0
where rc is in units of l F.
Figure 4-2 shows the irradiance and encircled-power distributions for g = 1. For
comparison, the corresponding distributions for a uniform beam are also shown. The
subscripts u and g refer to uniform and Gaussian beams, respectively. For clarity, the
irradiance distributions are also plotted on a logarithmic scale. At and near the focal
point, a uniform beam gives a higher irradiance than a Gaussian beam. Thus, Iu > Ig for
r < 0.42 . For larger values of r, Ig > Iu , except in the secondary rings, where again
Iu > Ig . The encircled power Pu >< Pg for rc <> 0.63 Of course, as rc Æ • , Pu Æ Pg Æ 1.
The Gaussian illumination broadens the central disc but reduces the power in the
secondary rings.
The positions of maxima and minima and the corresponding irradiance and
encircled-power values are given in Table 4-1. Comparing with the corresponding results
for a uniform beam given in Table 2-1, it is evident that the corresponding maxima and
minima for a Gaussian beam are located at higher values of r than those for a uniform
1.0
0.924
Pg
0.8
Pu
l(r; R); P(rc ; R)
0.6 1
10 – 1
0.4
l(r; R)
lg
10 – 2
0.2
lu 10 – 3
0.0 10 – 4
0 1 2 3 0 1 2 3 4 5
r; rc r
Figure 4-2. Focal-plane irradiance and encircled-power distributions for uniform

and Gaussian ( g = 1) beams of a given total power Pex . The irradiance is in units of
Pex Sex l 2 R 2 and the encircled power is in units of Pex . r and rc are in units of l F.
The irradiance distribution is also shown on a log scale to highlight the differences
between the secondary maxima of uniform and Gaussian beams. The dashed curves
are for a Gaussian beam.
Table 4-1. Maxima and minima of focal-plane irradiance distribution and

corresponding encircled powers for a Gaussian circular beam with g = 1.
Max/Min r, rc I(r) P(rc)
Max 0 0.924 0
Min 1.43 0 0.955
Max 1.79 0.0044 0.962
Min 2.33 0 0.973
Max 2.76 0.0012 0.976
Min 3.30 0 0.981
Max 3.76 0.0005 0.983
Min 4.29 0 0.985
Max 4.75 0.0002 0.986
beam. For example, the radius of the central bright spot is 1.43 for the Gaussian beam
compared to 1.22 for a uniform beam. Moreover, whereas the principal maximum for a
Gaussian beam is only slightly lower (0.924 compared with 1), the secondary maxima are
lower by a factor > 3 compared with the corresponding maxima for a uniform beam.
Letting r = 0 in Eq. (4-15), we obtain the focal-point irradiance for an aberration-

free system
[
I (0; g ) = tanh ( g 2) ( g 2) ] . (4-17)
For large values of g,
I (0; g ) Æ 2 g . (4-18)
Figure 4-3a shows how I (0; g ) varies with g. It has a maximum value of unity for g = 0 ,
i.e., for uniform illumination, and decreases monotonically as g increases. It proves Eq.
(1-192) by way of an example that, for a fixed total power in the pupil, any amplitude
variations across it (but without any aberrations) reduce the central irradiance. Now the
increase in g can be due to an increase in a or a decrease in w. If we consider the
unnormalized value of the focal-point irradiance, we note that it increases quadratically
with the pupil radius a in the case of a uniform beam, but for a Gaussian beam, it
increases only slightly (< 10 percent) with a for g ≥ 3. Its variation with a is shown in
Figure 4-3b. It should be evident that Iu (0) ~ a 2 and Ig (0) ~ tanh ( g 2) .
4.3.2.2 OTF
From Eq. (1-125) and Figure 2-42, the OTF for an aberration-free Gaussian pupil in
the ( p, q ) coordinate system can be written
(
t (v ) = a 2 Pex )Ú Ú A( p + v , q) A( p - v , q) dp dq , 0£ v£1
1- q 2 - v
( )
1- v 2
8g exp -2 gv 2
=
[
p 1 - exp( -2 g ) ]
Û
Ù dq
ı
Û
Ù
ı
[ (
exp -2 g p 2 + q 2 dp ,)] (4-19)
0 0
1.0 1.0
0.9
0.8
0.8
tanh(g/2)
I(0;g)
0.7 0.6
0.6 0.4
0.5
0.2
0.4
0.3 0.0
0 1 2 3 4 5 0 1 2 3 4 5
g g
(a) (b)
Figure 4-3. (a) Focal-point irradiance in units of Pex Sex l R for a Gaussian beam
2 2
as a function of g . (b) Variation of unnormalized focal-point irradiance with exit

pupil radius a for a fixed value of Gaussian radius w.
where v is a radial spatial frequency normalized by the cutoff frequency 1 l F , ( p, q ) are

the coordinates of a pupil point normalized by the pupil radius a, and the integration is
over the overlap region of two pupils whose centers are separated by a distance v along
the p axis.
For large values of g (e.g., g ≥ 4 ), the contribution to the integral in Eq. (4-19) is
negligible unless v = 0 , in which case it represents the Gaussian-weighted area of a
quadrant of the pupil, and the equation reduces to
t (v ) = exp -2 gv 2 ( ) , 0 £v £1 . (4-20)
This result may also be obtained by zero-order Hankel transforming the PSF for z = R
from Eq. (4-56b). The corresponding defocused OTF is considered in Problem 4.2.
Figure 4-4 shows how the OTF varies with v for several values of g . We note that
compared to a uniform pupil (i.e., for g = 0 ), the OTF of a Gaussian pupil is higher for
low spatial frequencies, and lower for the high.3 Moreover, as g increases, the bandwidth
of low frequencies for which the OTF is higher decreases and the OTF at high frequencies
becomes increasingly smaller. This is due to the fact that the Gaussian weighting across
the overlap region of two pupils whose centers are separated by small values of v is
higher than that for large values of v. If we consider an apodization such that the
amplitude increases from the center toward the edge of the pupil, then the OTF is lower
for low frequencies and higher for the high. Thus unlike aberrations, which reduce the
MTF of a system at all frequencies within its passband [see Eq. (1-107)], the amplitude
variations can increase or decrease the MTF at any of those frequencies.
0.8
1
0.6
t(n; g)
0.4
G 2
0.2
0
0 0.2 0.4 0.6 0.8 1
n
Figure 4-4. OTF of a Gaussian pupil. A uniform pupil corresponds to

g = 0 , and a large value of g represents a weakly truncated pupil.
4.3.3 Strehl Ratio and Aberration Tolerance

Using normalized quantities of Eqs. (4-6), (4-7), (4-12), and (4-13), the irradiance
distribution in the focal plane of an aberrated system given by Eq. (4-4) reduces to
2
1 2p
I ( r; q i ; g ) = p -2
Ú Ú
0 0
[ ( )] [ (
I (r) exp i F r, q p exp - pirr cos q p - q i r dr dqp ) ] , (4-21)
( )
where F r; q p is the the pahse aberration of the system at a point r, q p in the plane of ( )
its exit pupil. By definition, the Strehl ratio of the system is the ratio of the irradiances at
the focal point r = 0 with and without aberration. The aberration-free focal-point
irradiance is given by Eq. (4-19). From Eq. (4-21), the Strehl ratio is given by
2 2
1 2p È1 2p
˘
S = Ú Ú A(r) exp[i F(r, q)] r dr dq Í Ú Ú
ÍÎ 0
A(r) r dr dq ˙
˙˚
0 0 0
2
1 2p
{g p[1 - exp(- g ) ] } Ú Ú exp(- gr ) exp[i F(r, q)] r dr dq

2
= 2
, (4-22)
0 0
where we have dropped the subscript p on q p . Approximate expressions for Strehl ratio
when the aberration is small are given by Eqs. (1-204) through (1-206), e.g.,
S3 ~ exp(- s 2F ) , (4-23)
where
s 2F = < F 2 > - < F > 2 (4-24)
is the variance of the aberration across the Gaussian-amplitude weighted pupil. The mean
1 2p 1 2p
Ú Ú A(r) [F(r, q)] Ú Ú A(r) r dr dq

n
<F > =n
r dr dq (4-25)
0 0 0 0
with n = 1 and 2, respectively. Following the same procedure as for a uniformly

illuminated circular pupil, we can obtain the balanced primary aberrations and their
standard deviation.1, 4 Table 4-2 gives the aberrations and their standard deviations for
g = 1, i.e., when a = w . Comparing these results with those given in Tables 2-4 and 2-5
for uniform circular pupils, it is evident that the standard deviation of an aberration for a
Gaussian pupil is somewhat smaller than the corresponding value for a uniform pupil.
This is due to the fact that while an aberration increases as r increases, the amplitude
decreases, thus reducing its effect compared to that for a uniform pupil. Accordingly, for
a given small amount of aberration Ai , the Strehl ratio for a Gaussian pupil is somewhat
higher than that for a uniform pupil. Similarly, for a given Strehl ratio, the aberration
tolerance for a Gaussian pupil is somewhat higher than that for a uniform pupil. We also
Table 4-2. Primary aberrations and their standard deviations for optical systems
with Gaussian circular pupils and g = 1.
Aberration F(r, q) sF
Spherical As r 4 As 3.67
Balanced spherical (
As r - 0.933r
4 2
) As 13.71
Coma Ac r cos q
3
Ac 3.33
Balanced coma (
Ac r - 0.608r cos q
3
) Ac 8.80
Astigmatism Aa r2 cos 2 q Aa 4.40
2
(
Aa r cos q - 1 2 2
) Aa 5.61
Defocus Bd r 2
Bd 3.55
Tilt Bt r cos q Bt 2.19
note that, although aberration balancing in the case of a uniform pupil reduces the
standard deviation of spherical aberration and coma by factors of 4 and 3, respectively,
the reduction in the case of astigmatism is only a factor of 1.22. For a Gaussian pupil, the
trend is similar but the reduction factors are smaller for spherical aberration and coma,
and larger for astigmatism. They are 3.74, 2.64, and 1.27, corresponding to spherical
aberration, coma, and astigmatism, respectively.
4.3.4 Balanced Aberrations and Zernike-Gauss Circle Polynomials1, 2
We have shown in Sections 2.4 and 3.4 that when a classical aberration of a certain
order is combined with aberrations of lower order to minimize its variance across the
pupil, the balanced aberration thus obtained can be identified with a corresponding
Zernike polynomial. These polynomials are orthonormal over the circular or the annular
region of the uniformly illuminated exit pupil. The corresponding polynomials
orthonormal over an exit pupil with a Gaussian illumination can be obtained from the
polynomials for uniform illumination by the Gram-Schmidt orthogonalization process.5
The phase aberration function of a system with a circular exit pupil can be expanded
in terms of a complete set of Zernike-Gauss circle polynomials Z nm (r, q; g ) that are
orthonormal over a unit circle weighted by the Gaussian amplitude in the form
• n
F(r, q; g ) = Â Â c nm Z nm (r, q; g ) , (4-26a)
n =0 m =0
where cnm are the orthonormal expansion coefficients, n and m are positive integers
including zero, n – m ≥ 0 and even, and
[ ]1/ 2Rnm (r; g ) cos mq

Z nm (r, q; g ) = 2( n + 1) (1 + d m 0 ) . (4-26b)
The polynomials are are orthonormal according to

1 2p 1 2p
Ú Ú Z n (r, q; g ) Z n ¢ (r, q; g ) A(r) r dr d q Ú Ú A(r) r dr d q = d nn ¢ d mm ¢
m m¢
. (4-26c)
0 0 0 0
The radial polynomials are given by
È ( n - m) 2 ˘
Rnm (r; g ) = M nm Í Rnm (r) - Â (n - 2i + 1) Rnm (r) Rnm- 2i (r; g ) Rn - 2i (r; g )˙ , (4-27a)
ÍÎ i ≥1 ˙˚
where the angular brackets indicate an average over the Gaussian pupil; i.e.,
1 1
Rnm (r) Rnm- 2i (r; g ) = Ú Rnm (r) Rn - 2i (r; g ) A(r) r dr Ú A(r) r dr . (4-27b)
0 0
The normalization constant Mnm is chosen such that the radial polynomials obey the
orthogonality relation
1 1
1
Ú Rnm (r; g ) Rnm¢ (r; g ) A(r) r dr Ú A(r) r dr = d . (4-28)
0 0
n + 1 nn ¢
Note that except for the normalization constant, the radial polynomial Rnn (r; g ) is
identical to the corresponding polynomial for a uniformly illuminated circular pupil
Rnn (r) ; i.e.,
Rnn (r; g ) = Mnn Rnn (r) . (4-29a)
The radial polynomial Rnm (r; g ) is a polynomial of degree n in r containing terms in rn ,

rn -2 , ..., and r m , whose coefficients depend on the Gaussian amplitude through g; i.e., it
has the form
Rnm (r; g ) = anm rn + bnm rn - 2 + K + dnm r m , (4-29b)
where the coefficients anm , etc., depend on g. The number of Zernike (or orthogonal)
aberration terms in the expansion of an aberration function through a certain order n is
given by Eq. (2-63), as in the case of circle polynomials. The orthonormal Zernike-Gauss
expansion coefficients are given by
1 2p 1
c nm = Ú Ú F(r, q; g )Z nm (r, q; g ) A(r) r dr d q 2 pÚ A(r) r dr , (4-30)
0 0 0
as may be seen by substituting Eq. (4-26a) and utilizing the othonormality of the
polynomials. The Zernike-Gauss circle polynomials Z nm (r, q; g ) are similar to the Zernike
circle polynomials Z nm (r, q) , including their number through a certain order n, except that
they are orthonormal over a unit circle weighted by the Gaussian amplitude.
Consider a typical Zernike-Gauss aberration term in Eq. (4-26):
n (r, q; g ) = c nm Z n (r, q; g ) .
Fm m
(4-31)
Unless n = m = 0 , its mean value is zero; i.e.,

1 2p 1 2p
n (r, q; g ) = Ú Ú F n (r, q; g ) A(r) r dr d q 2 p Ú Ú
Fm A(r) r dr
m
0 0 0 0
= 1 . (4-32a)
For m = 0 , this may be seen with the help of Eq. (4-26c) and the fact that R00 (r; g ) = 1 is
a member of the polynomial set. The orthonormality Eq. (4-26c) yields the result that the
mean value of Rn0 (r; g ) is zero. When m π 0 , the average value of cos mq is zero.
Similarly, the mean square value of the aberration is given by
1 2p
[F (r, q; g )] [F (r, q; g )] A(r) r dr d q
1
2 p Ú A(r)r dr
2 2
m
n = Ú Ú m
n
0 0 0
= c nm
2
. (4-32b)
(F mn ) 2
2
s 2nm = - Fm
n
= c nm
2
, n π 0, m π 0 . (4-32c)
Thus, each orthonormal expansion coefficient, with the exception of c00 , represents the
standard deviation of the corresponding aberration term.
From Eqs. (4-31) and (4-32c), we note that the standard deviation of an aberration
can be obtained immediately by comparing its form with the corresponding orthonormal
aberration represented by Eq. (4-31), without having to calculate the integrals in Eq. (4-
32). The variance of an aberration consisting of two or more terms of the form of Eq. (4-
31) is given by the sum of the variances of each of the aberration terms. The variance of
the aberration function is accordingly given by
s 2F = F 2 (r, q; g ) - F(r, q; g )
2
• n
= Â 2
Â c nm . (4-33)
n =1 m = 0
A classical aberration of a certain order when combined with aberrations of lower

order to minimize its variance across the Gaussian-weighted circular exit pupil may be
represented by a corresponding Zernike-Gauss circle polynomial. The radial polynomials
Table 4-3. Radial polynomials representing balanced primary aberrations for

Gaussian beams. Polynomials for special cases of g = 1 and g = 0 (corresponding to
uniform beams) and weakly truncated Gaussian beams are also given.
Aberration Radial Gaussian* Gaussian Uniform Weakly Truncated

Polynomial g =1 g =0 Gaussian
Piston R00 1 1
1 1
1.09367r
R11 a11r r g / 2r
Distortion (tilt)
2.04989r2 – 0.85690
Field curvature R20 a20r2 + b20 2r2 – 1 ( gr2 – 1) / 3
Astigmatism R22 a22r2 1.14541r2 r2 ( g / 6 )r2
Êg ˆ
Coma R31 a31r3 + b31r 3.11213 r3 – 1 ◊ 89152r 3 r3 – 2 r g / 2 Á r3 – r˜
Ë2 ¯
Spherical aberration R40 a40r4 + b40r2 + c40 6.12902r4 – 5.71948r2 + 0.83368 6 r4 – 6 r2 + 1 ( g 2r4 – 4 gr2 + 2) / 2 5
1
*a11 = (2 p2 ) –1/2 , a 20 = [3( p4 – p22 )] –1/2, b20 = – p2 a 20 , a 22 = (3 p4 ) –1/2 , a 13 = ( p6 – p42 / p2 ), b31 = – ( p4 / p2 ) a 13 ,
2
{ }
–1/2
a 40 = 5 [ p8 – 2 K 1 p6 + (K 12 + 2 K 2 ) p4 – 2 K 1 K 2 p2 + K 22 ] , b40 = – K 1 a 40 , c40 = K 2 a 40 ,
p s = < r s > = (1 – expg ) –1 + ( s / 2 g ) p s – 2 , s is an even integer,
p 0 = 1, K1 = ( p6 – p 2 p 4 ) / ( p 4 – p 22 ), K 2 = ( p 2 p6 – p 42 ) / ( p 4 – p 22 ) .
corresponding to balanced primary aberrations are listed in Table 4-3. Thus, the balancing
defocus for spherical aberration As r 4 is given by Bd r2 , where
(
Bd = b40 a40 As ) . (4-34)
Or, for g = 1, Bd = - 0.933As so that the balanced spherical aberration becomes
(
F bs (r; 1) = As r 4 - 0.933r2 ) . (4-35)
The standard deviation of the balanced aberration is given by
s bs = As 5 a40 , (4-36)
which for g = 1 is s bs = As 13.705 . Similarly, the balancing tilt in the case of coma
Ac r3 cos q is given by Bt r cos q , where
(
Bt = b31 a31 Ac . ) (4-37)
Or, for g = 1, Bt = - 0.608 Ac so that the corresponding balanced coma is given by
F bc (r, q; 1) = Ac r3 - 0.608r cos q .( ) (4-38)

Its standard deviation is given by
s bc = Ac 2 2 a31 , (4-39)
which for g = 1 is s bc = Ac 8.802 .
In the case of astigmatism Aa r2 cos 2 q , the balancing defocus is given by
Bd = - Aa 2 . (4-40)
Thus, balanced astigmatism is given by
F ba (r, q; 1) = (1 2) Aa r2 cos 2q . (4-41)
We note that its form is independent of the Gaussian illumination. The standard deviation
of the aberration, of course, does depend on the illumination. It is given by
s ba = Aa 2 6 a22 , (4-42)
which for g = 1 is s ba = Aa 5.609 .
4.3.5 Defocused System

Now we consider a defocused system and show that, as in the case of a uniform
pupil, maximum central irradiance on a target at a given distance is obtained when a
beam is focused on it even though a larger irradiance is obtained at a point that is closer
to the pupil.
4.3.5.1 Theory
In this section, we consider a defocused but otherwise aberration-free system with a
Gaussian pupil. The axial irradiance, the irradiance distribution in a defocused plane, and
the case of a collimated beam are discussed with numerical examples. The differences
from similar results for a uniform pupil are pointed out.
Following Eq. (2-81), the irradiance distribution in a defocused plane at a distance z

from the plane of the exit pupil for an aberration-free beam may be written
1 2
2
Ê 2R ˆ
I (r; z; g ) = Á
Ë z ¯
˜ Ú ( 2
)
I (r) exp iBd r J 0 (prr ) r dr , (4-43)
0
where, as in Eq. (4-15), the irradiance is in units of Pex Sex l2 R 2 . However, r is in units
of lz D and r = 0 lies in the defocused plane at the point where the axis of the beam
intersects it. Equation (4-43) represents the modified form of the defocus free Eq. (4-15)
due to the defocus aberration. The quantity Bd in Eq. (4-43) represents the peak defocus
phase aberration. It is given by
Bd = p N Ê - 1ˆ
R
, (4-44a)
Ëz ¯
where
N = a2 l R (4-44b)
is the Fresnel number of the exit pupil as observed from the focus.
4.3.5.2 Axial Irradiance

If we let r = 0 in Eq. (4-43), we obtain the axial irradiance of the beam
2 Ê 2g ˆ Ê cos Bd ˆ
I (0; z; g ) = ( R z ) Á 2 2˜ Á coth g - ˜ . (4-45)
Ë Bd + g ¯ Ë sinh g ¯
By equating its derivative with respect to z to zero, we obtain the positions of its maxima
and and minima as the solutions of
Ê lz B ˆ
2Á - 2 d 2 ˜ (cosh g - cos Bd ) = - sin Bd . (4-46)
Ë Sex Bd + g ¯
They occur at approximately those z values at which the pupil subtends an odd or an even
number of Fresnel zones, respectively.
Figure 4-5 shows how the axial irradiance of a focused Gaussian beam differs from
that of a focused uniform beam when g = 1 and the Fresnel number N = 1, 10, 100. We
note that the principal maximum is higher for the uniform beam compared with that for
the Gaussian beam. However, the secondary maxima are higher for the Gaussian beam.
Moreover, whereas the axial minima for the uniform beam have a value of zero, the
minima for the Gaussian beam have nonzero values. We note that the curves become
symmetric about the focal point z = R as N increases.
It should be noted that even though the principal maximum of axial irradiance does
not lie at the focus, maximum central irradiance on a target at a given distance from the
pupil is obtained when the beam is focused on it. This can be seen by equating to zero the
derivative of the axial irradiance, Eq. (4-45) with respect to R. When doing so, the
normalization factor Pex Sex l2 R 2 should be substituted in these equations with the
consequence that the R 2 factor in front of its right-hand side disappears. Figure 4-6
illustrates how the central irradiance on a target at a fixed distance z varies when the
beam is focused at various distances R along the axis. The irradiance in this figure is in
units of Pex Sex l2 z 2 . The quantity N z = a 2 l z represents the Fresnel number of a
circular pupil as observed from the target. Note that the maximum irradiance values for
uniform and Gaussian ( g = 1) beams are 1 and 0.924, respectively. We also note that as
N z increases, the curves become symmetric about R = z for reasons discussed in Section
2.5.2.
2.0 2.0 2.0
N=1 N = 10 N = 100
1.5 1.5 1.5
lu
l (0; z)
1.0 1.0 1.0
lg
0.5 0.5 0.5
0.0 0.0 0.0

0.0 0.5 1.0 1.5 0.5 1.0 1.5 0.8 0.9 1.0 1.1 1.2
z/R z /R z /R
Figure 4-5. Axial irradiance of a circular beam focused at a fixed distance R with a
Fresnel number N = a 2 l R = 1, 10, 100. The irradiance is in units of the focal-point
irradiance for a uniform circular beam. The Gaussian beam results shown in this
figure by the dashed curves are for g = 1. The minima of axial irradiance of a
uniform beam have a value of zero and are located at z R = 1 3 , 1 5 , 1 7 , K.
1.0 1.0 1.0
0.8 lu 0.8 0.8

lg
lz (0; R)
0.6 0.6 0.6

Nz = 1 Nz = 10 Nz = 100
0.4 0.4 0.4
0.2 0.2 0.2
0.0 0.0 0.0

0.5 1.0 1.5 2.0 0.5 1.0 1.5 2.0 0.8 0.9 1.0 1.1 1.2
R/z R/z R/z
Figure 4-6. Central irradiance on a target at a fixed distance z from the plane of the
circular exit pupil when a beam is focused at various distances R. The quantity
N z = a 2 l z represents the Fresnel number of the pupil as observed from the target.
The irradiance is in units of Pex Sex l 2 z 2 . The Gaussian beam results shown in this
figure by dashed curves are for g = 1.
4.3.5.3 Defocused Distribution

Since irradiance is a real quantity, following Eq. (4-43), the irradiance distribution in
a defocused plane can be written
1 1
I ( r ; z; g ) = ( 2 R z )
2
Ú Ú [ I (r) I (s)]
0 0
12
[ (
cos Bd r2 - s 2 )] J (prr) J (prs) rs dr ds . (4-47)
0 0
If we let r = 0 and note that J 0 (0) = 1 , we obtain a different form of the expression for
axial irradiance, namely,
1 1
I (0; z; g ) = (2 R z )
2
Ú Ú [ I (r) I (s)]
0 0
12
[ ( )]
cos Bd r2 - s 2 rs dr ds . (4-48)
The encircled power (in units of Pex with rc in units of lz D ) is given by
1 1
P(rc ; z; g ) = 2 p 2
Ú Ú [ I (r) I (s)]
0 0
12
[ ( )]
cos Bd r2 - s 2 Q(r, s; rc ) rs dr ds , (4-49)
where Q(r, s; rc ) is given by Eqs. (2-108). The integrals in Eqs. (4-47) through (4-49) can
be evaluated by the Gauss quadrature method as discussed in Section 2.6.
An example of a defocused distribution is illustrated in Figure 4-7 for a Gaussian

beam with g = 1 and a large Fresnel number (so that the inverse-square-law variation is
negligible). The amount of defocus Bd = 2.783 rad (or 0.443l) is such that the central
irradiance for a uniform circular beam is reduced to half of the corresponding focal-point
irradiance. (The defocused distributions shown can also be interpreted as the distributions
on a target at a fixed distance z when the beam is focused at a distance R, such that
Bd = 2.783 rad . In this case, the irradiance would be in units of Pex Sex l2 z 2 , and r and
rc would be in units of lz D.) We note that, as in the case of focal-plane distributions,
the central irradiance for a Gaussian beam is lower than that for a corresponding uniform
beam. It decreases from 1 for a uniform beam to 0.500 and from 0.924 to 0.483 for a
Gaussian beam. The encircled power is slightly higher for a uniform beam for small
values of rc compared with that for a Gaussian beam.
0.6 1.2
0.500
0.5 0.483 1.0
Pg
0.4 0.8
P(rc)
l(r)
Pu
0.3 0.6
0.2 0.4
lg
0.1 0.2
lu
0.0 0.0
0.0 0.5 1.0 1.5 2.0 2.5
r; r0
Figure 4-7. Defocused irradiance and encircled-power distributions for uniform and
Gaussian ( g = 1) beams. The amount of defocus aberration Bd = 2.783 rad
l ) is such that it gives a central irradiance of 0.5 for a uniform circular
(or 0.443l
beam of a large Fresnel number N. The units of r, rc , I(r), and P(rc) are the same as
in Figure 4-2.
4.3.5.4 Collimated Beam

letting R Æ • . Thus, for example, Eq. (4-43) for the irradiance distribution in a plane at
a distance z from the pupil reduces to
2
1
I ( r; z ) = 4 Bd2 Ú (
I (r) exp i Bd r 2
) J (p rr) r dr
0 , (4-50)
0
where
Bd = Sex l z (4-51)
represents the phase aberration of a plane wavefront with respect to a reference sphere
centered at a distance z and passing through the center of the pupil. In Eq. (4-50), the
irradiance in both the pupil and the observation planes is in units of the pupil irradiance
Pex Sex for a uniform circular beam. As in Eq. (4-43), r is in units of lz D.
In the far field, i.e., for z ≥ D2 l, the phase aberration Bd £ p 4 (corresponding to

a wave aberration of less than or equal to l / 8) and may be neglected. Hence the
irradiance distribution and, correspondingly, the encircled power distribution in a far-field
plane is similar to the focal-plane distribution discussed earlier. The only difference is in
scaling of the diffraction pattern. Similarly, in the near field, i.e., for z < D2 l, the
irradiance and encircled-power distributions correspond to defocused distributions
discussed earlier. The only significant difference is in the definition of Bd .
If z is in units of the far-field distance D2 l and we let r = 0 in Eq. (4-50), we

obtain the axial irradiance (in units of Pex Sex )
I (0; z ) = 2 g { [1 + (4g z p) ] } [coth g - cos(p 4z) sinh g ]

2
. (4-52)
Its maxima and minima are located at z values given by
{2(4 z p) g [1 + (4g z p) ] } [cosh g - cos(p 4z)] = - sin(p 4z)

3 2 2
. (4-53)
Figure 4-8 illustrates how the axial irradiance of collimated uniform and Gaussian
beams vary with distance z from the pupil. With reference to Figure 4-5, Figure 4-8
corresponds to N = 0 . Unlike the principal maximum of a focused beam in Figure 4-5a,
the maximum farthest from the pupil has a lower value than those closer to it in the case
of a Gaussian beam. Moreover, whereas the maxima of axial irradiance of a collimated
uniform beam have the same value of 4 and minima have a value of zero, the maxima in
the case of a Gaussian beam are higher and the minima are nonzero. Because of their
different amplitudes, the Fresnel zones do not cancel completely at the location of the
minima.
N=0
4
3
l(0; z)
2 lg
1
lu
0
0.0 0.1 0.2 0.3 0.4 0.5
z
Figure 4-8. Axial irradiance of a collimated Gaussian circular beam with g = 1

compared with that for a corresponding uniform beam. The irradiance is in units of
the pupil irradiance Pex Sex at a uniform pupil. The distance z is units of the far-
field distance D 2 l for a uniform pupil.
4.3.6 Weakly Truncated Gaussian Circular Beams

4.3.6.1 Introduction
When the radius a of the exit pupil is much larger than the Gaussian radius w of the
beam, it is said to be weakly truncated. For such a beam, the upper limit of integration in
the diffraction integrals across the pupil can be replaced by infinity with negligible error.
The integrations for the aberration-free case in any defocused plane can be carried out in
a closed form, and simple results are obtained by virtue of the fact that the Fourier
transform of an (untruncated) Gaussian function is also Gaussian. Thus, a weakly-
truncated Gaussian beam remains Gaussian as it propagates. When such a beam is
focused at a certain distance, it is shown that its radius is not minimum in the focal plane.
Instead, it lies in a plane that is closer to the pupil than the focal plane. Its axial irradiance
is accordingly maximum in that plane. The distance between the plane of minimum
radius and the focal plane increases as the Fresnel number of the beam decreases. The
minimum radius of the beam is called its waist. The focusing effects of a lens on a
Gaussian beam incident on it are also considered. In particular, the imaging relationships
between the waists of the incident and transmitted beams are derived by considering them
as an object and an image, respectively. Interesting results are obtained that have no
counterparts in conventional imaging. A collimated beam is considered as a limiting case
of a focused beam.
The aberration tolerance of a weakly truncated beam is much higher owing to the
pupil irradiance being significant only near its center where the aberration is small.
Accordingly, the tolerances for primary and balanced primary aberrations are discussed in
terms of their value at the beam radius rather than at the edge of the pupil. Based on the
aberration analysis, it is shown that the pupil radius must be at least three times the beam
radius in order that the beam may be treated as weakly truncated with a negligible error.
4.3.6.2 Irradiance Distribution and Beam Radius

In this section, we consider a weakly truncated Gaussian circular beam, i.e., one for
which the aperture radius a is much larger than the beam radius w. When a > > w ; i.e.,
for a large value of g, Eq. (4-11) for the pupil irradiance distribution may be written
(
I (r) = 2 g exp - 2gr2 ) , (4-54a)
or
( ) ( )
I rp = 2 Pex p w 2 exp È- 2 rp w ˘ . ( )
2
(4-54b)
ÍÎ ˙˚
In Eq. (4-54a), I(r) is in units of Pex Sex as stipulated in Eq. (4-11). In Eq. (4-54b), no
such normalization is used.
When g is large, the upper limit of integration in Eq. (4-43) may be replaced by
infinity with negligible error. Hence, if we let b = pr and a = g - i Bd and follow
•
Ú (- ar ) J0 (br) r dr = (1 2a ) exp - b 2 4a ( ) , Re a > 0 ,

2
(4-55)
0
Eq. (4-43) for the defocused irradiance distribution reduces to
I ( r; z ) = ( R z )
2
[2 g ( B 2
d )] [ (
+ g 2 exp - g p 2 r 2 2 Bd2 + g 2 )] , (4-56a)
or
( ) (
I (r; z ) = 2 Pex p w 2z exp - 2 r 2 w 2z ) , (4-56b)
where
w 2z = (l z p w ) + w 2 (1 - z R)
2 2
. (4-56c)
is the beam radius at a distance z. In Eq. (4-56a) the irradiance is in units of Pex Sex l2 R 2
and r is in units of lz D, as was the case in Eq. (4-43). In Eq. (4-56b) these quantities
are not normalized. Comparing Eqs. (4-54) and (4-56), we note that, if the truncation of
the beam by the pupil is neglected, a Gaussian beam remains Gaussian as it propagates.
The corresponding OTF is considered in Problem 4.2.
If we let r = 0 in Eqs. (4-56a) and (4-56b), we obtain the axial irradiance
I (0; z ) = ( R z )
2
[2 g ( B 2
d + g2 )] , (4-57a)
or
I (0; z ) = 2 Pex p w 2z . (4-57b)
Of course, for large values of g, Eq. (4-45) also reduces to Eq. (4-57a), as expected.
If we let z = R in Eqs. (4-56), we obtain the focal-plane irradiance distribution
(
I (r; R) = (2 g ) exp - p 2 r 2 2 g ) , (4-58a)
or
( ) (
I (r; R) = 2 Pex p w 2R exp - 2 r 2 w 2R ) , (4-58b)
where
wR = lR pw (4-58c)
is the beam radius in the focal plane. The focal-point irradiance is given by
I (0; R) = 2 g , (4-59a)
or
I (0; R) = 2 p Pex w 2 l2 R 2 , (4-59b)
a result already obtained in Eq. (4-18).
If we equate to zero the derivative of Eq. (4-57) with respect to z, we obtain the z
value at which the axial irradiance is maximum. It can be shown that this value is given
by
[
z p R = 1 + (g p N ) ]
2 -1
, (4-60a)
or
[
z p R = 1 + (w R w ) ]
2 -1
. (4-60b)
Substituting Eq. (4-60) into Eq. (4-57), we obtain the peak value of axial irradiance
( ) (
I 0; z p = (2 g ) + 2 g p 2 N 2 ) , (4-61a)
or
( )
I 0; z p = 2 Pex p w 2z p , (4-61b)
where
[
w 2z p = w 2 1 + ( w w R ) ]
2 -1
. (4-61c)
Comparing Eqs. (4-59a) and (4-61a), we note that the peak axial irradiance is higher than
the corresponding focal-point irradiance by 2 g p 2 N 2 (in units of Pex Sex l2 R 2 ) or by
2 Pex p w 2. Equations (4-60a) and (4-61c) can also be written in the forms
- 2˘-1
ÎÍ
(
z p R = È1 + p N g ) ˚˙
(4-62a)
and
-1
(
w 2z p = w 2 È1 + p N g ˘ )
2
, (4-62b)
ÍÎ ˙˚
respectively, where N g = w 2 l R is the Gaussian Fresnel number. It represents the

number of Fresnel zones in the pupil plane within the Gaussian-beam radius, as observed
from the focus, just as N represents the number of zones within the full pupil. Since the
axial irradiance is maximum at z p , the beam radius is minimum at this position. This
may also be seen by equating to zero the derivative of Eq. (4-56c) with respect to z. The
minimum beam radius w z p is generally referred to as the beam waist.
It is evident that z p < R ; i.e., the peak of axial irradiance does not occur at the focal
point but at a point between it and the pupil. Moreover, w z p < w ; i.e., the waist of the
diffracted beam is smaller than the beam radius at the pupil. Note, however, that, as
discussed earlier, even though the peak axial irradiance and the beam waist are not
located at the focal point ( z = R) , the smallest beam radius and maximum central
irradiance on a target at a fixed distance z are obtained when the beam is focused on it.
The encircled-power distribution in an observation plane is given by
[
P(rc ; z ) = 1 - exp - g p 2 rc2 2 Bd2 + g 2 ( )] , (4-63a)
or
{
P(rc ; z ) = Pex 1 - exp - 2 rc2 w 2z [( )] } , (4-63b)
where rc is in units of lz D.
The complex amplitude of an apertured converging spherical wave of radius of

( )
curvature R varies as exp - ip ri2 l R , where ri is the radial distance of a point in the
pupil plane from its axis; i.e., its radius of curvature is given by the inverse of the
coefficient of - ip ri2 l in the exponent of its complex amplitude representation. The
complex amplitude of the diffracted Gaussian spherical wave consists of two factors that
(
contain iri2 dependence in the exponent. One of these is exp i p ri2 l z , as may be seen )
from Eq. (1-49). The other comes about when the integral on the right-hand side of
Eq. (4-43) is evaluated utilizing Eqs. (4-54a) and (4-55). Noting that r is in units of lz D
in b = p r , we find that the radius of curvature Rz of a diffracted Gaussian spherical
wave at a distance z from the plane of the exit pupil is given by
z Rz = ( Sex l z ) Im(1 a ) - 1
[ (
= Sex Bd l z Bd2 + g 2 - 1 )] (4-64a)
1-z R
= - 1 . (4-64b)
( )
2
(1 - z R)2 + lz p w2
At the waist position z p , Rz p = • , implying a plane wave. Moreover, at the focal plane,
Rz = - R . A negative value of Rz indicates a diverging spherical wave. For z > R , the
beam continues to expand as it propagates.
The results for a weakly truncated collimated Gaussian beam can be obtained from
those for a focused beam by letting R Æ • . Thus, for example, Eqs. (4-56) reduce to
{ [
I ( r; z ) = 2 g 1 + ( 4 g z p )
2
] } exp {- 8g z r [1 + (4g z p) ] }
2 2 2
, (4-65a)
or
( ) (
I (r; z ) = 2 Pex p w 2z exp - 2 r 2 w 2z ) , (4-65b)
where
(
w 2z = w 2 È1 + l z p w 2 ˘ . )
2
(4-65c)
ÎÍ ˙˚
Similarly, Eq. (4-64b) reduces to
(
Rz = - z È1 + p w 2 l z ˘ . )
2
(4-66)
ÎÍ ˙˚
In Eq. (4-65a) the irradiance is in units of Pex Sex , r is in units of lz D (z is not

normalized here), and z is in units of D2 l. We note from Eq. (4-65c) that w z increases
monotonically as z increases; i.e., the beam expands as it propagates.
In this section we have written equations in two equivalent forms. Equations (a) are
written in a normalized form so that they can be investigated parametrically. Equations
(b) are written without any normalizations, and they are more suitable for evaluating
results when the specific parameters involved are known. Incidentally, as may be shown,
all equations in this section reduce to the corresponding equations in the pupil plane if we
let z Æ 0 .
Figure 4-9 shows the aberration-free focal-plane irradiance and encircled-power
distributions for g = 2 . The solid curves have been obtained by using Eqs. (4-14b) and
0.5 1.0
g =2
0.4 0.8
Pg
0.3 0.6
Pg(r;R)
lg(r;R)
0.2 0.4
0.1 lg 0.2
0.0 0.0
0 1 2 3
r; rc
Figure 4-9. Focal-plane irradiance and encircled-power distributions for a Gaussian

beam with g = 2 . The irradiance and encircled power are in units of
Pex Sex l R and Pex ,respectively. The radial distance r or rc in the focal plane is in
2 2
units of l F. The focal point is at r = 0 . The solid curves represent the exact results
and the dashed curves represent their corresponding approximations by neglecting
the beam truncation.
(4-15) and the dashed curves represent their corresponding approximations given by
Eqs. (4-58a) and (4-63a), respectively, with Bd = 0 . We note that the approximate results
agree well with the true results. The maximum difference, which occurs at the focus, is
less than 4%. For larger g, the agreement is found to be even better. We will see in
Section 4.3.6.4 that when the beam is aberrated, a larger value of g , namely, g = 3, is
required for the validity of weakly truncated approximation.
Figure 4-10 shows how the axial irradiance of a focused Gaussian beam varies when
g = 2 and N = 1, 10, 100. Once again, the solid curves in this figure have been
obtained by using Eq. (4-45); the dashed curves represent their corresponding
approximations given by Eq. (4-57a). It is evident that Eq. (4-57a) represents the true
axial irradiance quite well. The only significant difference occurs when N = 1, in that the
true results show secondary maxima and minima, but the approximate result shows only
the principal maximum. For larger values of g ; e.g., g = 2.5 , the secondary maxima
and minima disappear and the true and approximate results overlap each other at the scale
of Figure 4-10. Hence, we conclude that the truncation of an aberration-free Gaussian
beam by a pupil has a negligible effect on the irradiance distribution as the beam
propagates when g = 2 .
2.0 1.0 1.0

g =2 g =2 g =2
N=1 0.8 N = 10 0.8 N = 100
1.5
lg (0; z)
0.6 0.6
1.0
0.4 0.4
0.5
0.2 0.2
0.0 0.0 0.0

0.0 0.5 1.0 1.5 0.5 1.0 1.5 0.8 1.0 1.2
z /R z /R z /R
Figure 4-10. Axial irradiance of a Gaussian beam with g = 2 focused at a distance

R with a Fresnel number N = 1, 10, 100. The irradiance is in units of Pex Sex l 2 R 2
The solid curves represent the exact results and the dashed curves represent their
corresponding approximations by neglecting the beam truncation.
4.3.6.3 Imaging of a Gaussian Beam6, 7

We now consider a Gaussian beam incident on a lens of focal length f and determine
the characteristics of the beam transmitted by it. Let the waist of the beam have a radius
w 0 lying at a (numerically negative ) distance z0 from the lens. By the definition of the
waist, the radius of curvature of the beam wavefront is infinity. The beam expands as it
propagates towards the lens. Letting R = • in Eqs. (4-56c) and (4-64b), the Gaussian
beam incident on the lens has a width w l and a wavefront of radius of curvature Rl- ,
given by
(
w l2 = w 20 È1 + z0 / z R ˘ )
2
(4-67)
ÍÎ ˙˚
and
(
Rl- = z0 È1 + z R / z0 ) ˘ ,
2
(4-68)
ÍÎ ˙˚
where
z R = pw 20 / l (4-69)
is called the Rayleigh range and represents the distance from the plane of the beam waist
to a plane in which the beam radius has increased by a factor of 2 and, therefore, the
axial irradiance has decreased by a factor of 2. Rl- has a minimum value of - 2z R for
z0 = - z R , and an approximate value of z0 , like a spherical wave, for z0 >> z R .
The wavefront exiting from the lens has a radius of curvature Rl+ , given by
1 1 1
+ = - + . (4-70)
Rl Rl f
The beam transmitted by the lens first converges and then expands as it propagates to
theright of the lens. Its width w ¢ at a distance z ¢ from the lens is given by Eq. (4-56c)
with appropriate substitutions:
2 2
Ê lz ¢ ˆ 2Ê z¢ ˆ
w¢ 2
= Á ˜ + w l Á1 - + ˜
Ë pw l ¯ Ë Rl ¯
(w 0 z ¢ ) 2
2
z02 + z R2 Ê 1 1 z0 ˆ
+ (w 0 z ¢ )
2 (4-71)
= Á z¢ - f - 2 ˜ .
z02 + z R2 2
zR Ë z0 + z R2 ¯
If we let z ¢ = f , Eq. (4-71) reduces to
lf
w¢ = . (4-72)
pw 0
Thus, the beam radius in the back focal plane of the lens is independent of the location z0
of the waist.
The waist of the transmitted beam according to Eq. (4-61c) is given by
È Ê 2˘ -1
w ˆ
w ¢02 = w l2 Í1 + Á l ˜
˙
Í Áw + ˜ ˙
ÍÎ Ë Rl ¯ ˙˚
-1
È Ê pw l2 ˆ ˘
2
= w l2 Í1 + Á +˜
˙
ÍÎ Ë l Rl ¯ ˙
˚
w 20 f 2
= .
( z0 + f )2 + z R2 (4-73)
It occurs at a distance z0¢ from the lens which, according to Eq. (4-60b), is given by
-1
z0¢ È Ê lR + ˆ 2 ˘
= Í1 + Á l2 ˜ ˙
Rl+ ÍÎ Ë pw l ¯ ˙˚
or
z0¢ =
[
f z0 ( z0 + f ) + z R2 ] . (4-74)
( z0 + f ) 2
+ z R2
Considering the incident and transmitted waists as an object and its corresponding
image formed by the lens, Eq. (4-74) relating their distances from the lens may be written
1 1 1
- = , (4-75)
z0¢ z0 + z R ( z0 + f )
2
/
f
which reduces to a conventional imaging equation in the limit z R f Æ 0 . If we plot

z0¢ f as a function of z0 f using z R f as a parameter, as illustrated in Figure 4-11, the
curves pass through an inflection point ( - 1, 1) with maxima and minima lying at
( ) (
1 - f 2 z R , 1 - z R / f and 1 + f 2 z R , 1 + z R / f , respectively, as may be seen by )
differentiating Eq. (4-75) with respect to z0 and equating the result to zero. Thus, when
the waist of the incident beam lies in the front focal plane of the lens, the waist of the
transmitted beam lies in its back focal plane, and not at infinity as in conventional
imaging.
In conventional imaging of a real object by a positive lens forming a real image,

there is a minimum separation of 4 f between an object and its image, which corresponds
to - z0 = z0¢ = 2 f and a unity magnification. However, there is no minimum separation
between a real object waist and a real image waist. For example, if the incident waist lies
at the lens, i.e., if z0 = 0 , then Eq. (4-71) reduces to
1/ 2
È 1 Ê 1 1ˆ2˘
w ¢ = w ¢0 z ¢ Í 2 + Á - ˜ ˙ . (4-76)
ÍÎ z R Ë z ¢ f ¯ ˙˚
The radius and position of the corresponding transmitted waist are given by Eqs. (4-73)
and (4-74) according to
w0
w ¢0 = (4-77)
[1 + ( z )2 ]
1/ 2
R f
and
f
z0¢ = . (4-78)
1 + ( f zR )
2
4
4
3
zR
= 0 1/4
f
2 3 zR
1/2 =0
1 f
1 2
z0 / f
0 1/4
2
–1
1/2
–2 1
1
–3
2
–4 0
–4 –3 –2 –1 0 1 2 3 4 –4 –3 –2 –1 0 1 2 3 4
z0 /f z0 / f
(a) (b)
Figure 4-11. Incident and transmitted waist relationships. (a) Object and
corresponding image distances. (b) Transverse magnification.
From Eq. (4-73), the waist magnification is given by
M = w ¢0 / w 0
1
= , (4-79)
[ ( )] (
1/ 2
Ï
) ˝˛
2 2¸
Ì 1 + z0 / f + zR / f
Ó
which reduces to conventional image magnification as z R f Æ 0 , provided the negative

sign associated with an inverted image is ignored. The maximum value of magnification
is f / z R , which occurs when the object waist lies in the front focal plane, i.e., when
z0 = - f . Hence, unity magnification in this case occurs only if f = z R .
The Rayleigh range of the image beam is given by
z R¢ = pw ¢02 l = M 2 z R . (4-80)
Differentiating Eq. (4-75) and utilizing Eq. (4-73), the longitudinal magnification of the
waist is given by
[ ( )]
2
∂ z0¢ z 2 - 1 + z0 / f
2 R
Ml = = - M , (4-81)
∂ z0
[ ( / f )]
2
z R2 + 1 + z0
which reduces to the expression for conventional imaging as z R Æ 0.
4.3.6.4 Aberration Balancing

When a Gaussian circular beam is weakly truncated, i.e., when g is large, the quantity
ps in Table 4-3 reduces to
ps = < rs >
= (s 2 g ) ps - 2
= (s 2) ! g - s 2 . (4-82)
As a result, we obtain simple expressions for the radial polynomials representing

balanced primary aberrations. They are also listed in Table 4-3. If we normalize rp by w
(instead of by a), then g disappears from these expressions. The standard deviation of an
aberration can be obtained by comparing its form with the corresponding orthonormal
aberration of Eq. (4-26).
The standard deviation of primary aberrations and the corresponding balanced

aberrations for a weakly truncated Gaussian beam is given in Table 4-4. In this table,
r¢ = g r (4-83)
Table 4-4. Primary aberrations and their standard deviations for optical systems
( )
with weakly g ≥ 3 truncated Gaussian circular pupils.
Aberration F(r¢, q) sF Ai¢ for S = 0.8
Spherical As¢ r¢ 4 2 5 As¢ l 63
Balanced spherical (
As¢ r¢ 4 - 4 r¢ 2 ) 2 As¢ l 28
Coma Ac¢ r¢ 3 cos q 3 Ac¢ l 24
Balanced coma ( )
Ac¢ r¢ 3 - 2 r¢ cos q Ac¢ l 14
Astigmatism Aa¢ r¢ 2 cos 2 q Aa¢ 2 l 10
Balanced astigmatism (
Aa¢ r¢ 2 cos 2 q - 1 2 ) Aa¢ 2 l 7
Defocus Bd¢ r¢ 2 3 Bd¢ l 24
Tilt Bt¢ r¢ cos q 3 Bt¢ l 20
is a radial variable in the pupil plane normalized by the beam radius w, and the aberration
coefficients Ai¢ represent the peak values of the Seidel aberrations at r¢ = 1. These
aberration coefficients are related to the coefficients Ai according to
As¢ = As g 2 , Ac¢ = Ac g 3 2 , Aa¢ = Aa g , Bd¢ = Bd g , Bt¢ = Bt g . (4-84)
The reason for defining the primed aberration coefficients in this manner is as follows.
Since the power in a weakly truncated Gaussian beam is concentrated in a small region
near the center of the pupil, the effect of the aberration in its outer region is negligible.
Accordingly, the aberration tolerances in terms of the peak value of the aberration at the
edge of the pupil (r = 1) are not very meaningful. We note from Table 4-3 (or Table
4-4) that the point with respect to which the variance of an aberration is minimized is
given by
Bd = - ( 4 g ) As (4-85a)
= - 4g As¢ , (4-85b)
Bt = - (2 g ) Ac (4-86a)
= - 2 g Ac¢ , (4-86b)
and
Bd = - (1 2) Aa (4-87a)
= - ( g 2) Aa¢ , (4-87b)
in the case of spherical aberration, coma, and astigmatism, respectively. From Table 4-4
we note that the balancing of a primary aberration reduces its standard deviation by a
factor of 5 , 3 , and 2 in the case of spherical aberration, coma, and astigmatism,
respectively. These reduction factors are listed in Table 4-5 for the uniform ( g = 0) , e - 2
truncated ( g = 1) and weakly truncated (large g) Gaussian beams. It is evident that as g
increases, the reduction factors decrease for spherical aberration and increase for
astigmatism. The amount of balancing aberration decreases as g increases in the case of
spherical aberration and coma, but it does not change in the case of astigmatism. For
example, in the case of spherical aberration, the amount of balancing defocus for a
weakly truncated Gaussian beam is ( 4 g ) times the corresponding amount for a uniform
beam. Similarly, in the case of coma, the balancing tilt for a weakly truncated Gaussian
beam is (3 g ) times the corresponding amount for a uniform beam. For large g , the
aberration tolerance in terms of the coefficients Ai¢ is given in Table 4-4 for a Strehl ratio
of 0.8.
Table 4-6 gives the reduction factors that relate the peak value Ai of a primary
aberration at the edge of a circular aperture and the standard deviation of its
corresponding balanced aberration for various values of l. In the case of balanced
aberrations, these numbers are given by 5a40 , 2 2 a31 , and 2 6 a22 for spherical
aberration, coma, and astigmatism, respectively. For example, for spherical aberration
As r 4 , the standard deviation of the corresponding balanced spherical aberration when
g = 2 is equal to As 18.29 .
Comparing the standard deviation results given in Table 4-6 with those for a weakly
truncated Gaussian beam given in Table 4-4, we find that they agree with each other with
negligible difference for g > 3. This provides a convenient definition for a weakly
truncated Gaussian beam, namely, that a > 3w . Some authors have assumed that g ≥ 2
Table 4-5. Factor by which the standard deviation of a Seidel aberration across a
circular aperture is reduced when it is optimally balanced with other aberrations.
Reduction Factor
Weakly Truncated
Aberration Uniform ( g = 0) Gaussian ( g = 1) (
Gaussian g ≥ 3 )
Spherical 4 3.74 2.24
Coma 3 2.64 1.73
Astigmatism 1.22 1.27 1.41

* The numbers given in this table represent the factor by which the peak aberration coefficient Ai must be
divided by in order to obtain the standard deviation.
Table 4-6. Standard deviation factor for primary aberrations for a Gaussian
circular beam with various values of g.
g Balanced Spherical Balanced Coma Balanced Astigmatism
0 13.42 8.49 4.90

0.5 13.69 8.53 5.06
1.0 13.71 8.80 5.61
1.5 14.90 9.74 6.81
2.0 18.29 12.21 9.08
2.5 26.33 17.62 12.82
3.0 43.52 27.57 18.06
3.5 75.78 42.96 24.51
4.0 128.09 64.01 32.00
* The numbers given in this table represent the factor by which the peak aberration coefficient Ai must be
divided by in order to obtain the standard deviation.
provides a sufficient condition for the validity of the aberration analysis of a weakly
truncated Gaussian beam given here. When g = 2 , the standard deviation of balanced
spherical aberration according to the weakly-truncated beam assumption is given by
As 8, whereas the true value, as stated above, is given by As 18.29 , which is
significantly different. When g = 3, the corresponding standard deviations are given by
As 40.50 and As 43.52 , which are nearly equal to each other. The difference between
the true and approximate results is even less for g > 3.
When g = 2 , even though the true focal-plane distribution obtained from

Eq. (4-15) agrees quite well with that obtained from Eq. (4-58a), the true and the
approximate standard deviations of primary aberrations are significantly different, as
pointed out above. The reason for the discrepancy in the case of an aberrated beam is
simple. Even though the irradiance in the region of the pupil w a £ r £ 1 is quite small
compared with that at or near its center, the amplitude in this region is not as small.
Moreover, the aberration in this region can be quite large and thus have a significant
effect on the standard deviation. In the case of spherical aberration, it increases as r 4 . In
the case of coma and astigmatism, it increases as r3 and r2 , respectively. Hence, we
require a larger value of g, namely g ≥ 3, for the aberrated-beam analysis of this section
to be valid. This is also true of defocus, which varies as r2 .
4.3.7 Symmetry Properties of an Aberrated PSF8

The symmetry properties of a PSF for a Gaussian pupil aberrated by a Zernike-Gauss
circle polynomial can be obtained in a manner similar to those for a uniform pupil
discussed in Section 2.7. All equations in that section remain the same, except that the
( )
integrand in all integrals is multiplied by A(r) ~ exp - g r2 . Of course, the radial
polynomial is also different. Equation (2-131) describing the aberrated irradiance

distribution is modified for a Gaussian pupil to
2
I ( r , q i ; z, g ) = Ê ˆ
R
Ë p z¯
2
1 2p
Û Û Ï ¸
I (r) exp Ìi ÈÍ F(r, q; g ) - p r r cos (q - q i )˘˙ ˝ r dr dq
R
¥ Ù Ù , (4-88)
ı ı ÓÎ z ˚˛
0 0
where
F(r, q; g ) = Anm Rnm (r; g ) cos mq + Bd r2 , (4-89)
with Bd given by Eq. (4-44) and the irradiance is normalized by Pex Sex / l2 R 2 . However,
the reasoning used to determine the symmetry properties and the results obtained are the
same, except that the axial irradiance for spherical aberration (expressed by a Zernike-
Gauss polynomial) is not symmetric about the Gaussian image plane even for large
Fresnel numbers. Letting z ~ R for a system with a large Fresnel number, the axial
irradiance for spherical aberration may be written
2
1
Û
I (0; z; g ) = 4 Ù
ı
( ) [ ]
I (r) exp iBd r2 exp i A n 0 Rn0 (r; g ) r dr . (4-90)
0
The radial polynomials Rn0 (r; g ) are no longer given by the Legendre polynomials and do
not possess their symmetry property, unlike the polynomials Rn0 (r) . Moreover, the
arguments made following Eq. (2-145) do not hold. For example, for uniform pupils
aberrated by primary spherical aberration As r 4 , the axial irradiance is symmetric about
the axial point corresponding to Bd = - As (see Figure 2-25). For Gaussian pupils, this is
no longer true. Several examples of axial irradiance for a Gaussian pupil with g = 1
aberrated by spherical aberration are shown in Figure 4-12, where I0 is the irradiance at
the center of the pupil and the aberration-free central value of 0.924 has been normalized
to unity. The lack of its symmetry about any point is evident from this figure. From the
form of the Zernike-Gauss circle polynomial for astigmatism, we note that the point of
axial symmetry for astigmatism Aa r2 cos 2 q is the same, namely, Bd = - Aa 2, for
uniform and Gaussian pupils.
4.4 SYSTEMS WITH ANNULAR PUPILS

Now we consider systems with annular pupils. The effect of a central obscuration in
the pupil is shown to be opposite to that of the Gaussian apodization. It reduces the size
of the central bright disc and increases the power in the rings of the diffraction pattern.
Accordingly, as the obscuration increases, the difference between the diffraction effects
of uniform and Gaussian beams decreases. The balancing of aberrations is discussed in a
manner similar to that for circular pupils, and the balanced aberrations are identified with
4.4 Systems with Annular Pupils 367
1.0
W(r) = Asr4 As = 0
1
I(0; Bd) 0.8 I(r) = I0exp(– 2r2)
0.6
2
0.4
0.2 3
0.0
–6 –5 –4 –3 –2 –1 0 1 2 3
Bd
Figure 4-12. Axial irradiance for spherical aberration As r4 with various values of
As and g = 1. The aberration-free central value of 0.924 has been normailzed to
unity.
the Zernike-Gauss annular polynomials whose derivation is outlined. The symmetry

properties of aberrated annular Gaussian beams and their line of sight in terms of their
centroids are discussed.
4.4.1 Theory
We now consider systems with annular Gaussian pupils with inner and outer radii of
a and a, respectively, where 0 £ < 1 is the linear obscuration ratio of the pupil. For a
fixed total power Pex transmitted by the pupil, regardless of the value or g , the
irradiance distribution at the pupil may be written
I (r; ) = 2 g exp - 2 g r2 ( ) [exp (-2 g ) 2

- exp ( - 2 g ) ] . (4-91)
( )
-1
As in Eq. (4-11), the irradiance is in units of Pex Sex . As g Æ 0 , I (r; ) Æ 1 - 2 , as
expected for a uniformly illuminated pupil. Of course, Eq. (4-91) reduces to Eq. (4-11) as
Æ 0 . Figure 4-13a shows how the irradiance across a pupil with g = 1 and = 0.5
varies compared to a corresponding uniform annular pupil. It varies from a maximum
value of 2.57 at the inner edge of the pupil to a minimum value of 0.57 at its outer edge,
compared to a value of 1.33 for the corresponding uniform pupil. The ratio of the peak
values of the pupil irradiance for Gaussian and uniform annular beams is given by
Ig ( ) Iu ( ) = 2 g 1 - 2 ( ) {1 - exp [- 2 g (1 - ) ] }
2
. (4-92)
The variation of this ratio with is shown in Figure 4-13b for g = 0.5, 1, 2, and 3. It is
( )
evident that as g 1 - 2 increases, the ratio approaches a value of 2 g 1 - 2 . ( )
3 20
2.57 = 0.5
g =3
15
2 lg
lg / lu
l
10
lu 2
1
5
0.57 1
0.5
0 0
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
r
(a) (b)
Figure 4-13. (a) Pupil irradiance distribution for uniform and Gaussian ( g = 1)
beams of a given total power Pex across an annular pupil with = 0.5 . (b) Ratio of
peak values of pupil irradiance as a function of for several values of g . The units
of pupil irradiance are Pex Sex , where Sex = p a 2 is the area of a circular pupil.
4.4.2 Aberration-Free System

Following the same procedure as for circular pupils, we find that the irradiance and
encircled-power distributions in the focal plane of a focused annular beam are given by
2
È1 ˘
I (r; g ; ) = 4 Í Ú I (r; ) J 0 ( p rr ) r dr˙ (4-93)
ÍÎ ˙˚
and
rc
(
P (rc ; g ; ) = p 2 2
)Ú I (r; g ; ) r dr . (4-94)
0
They differ from the corresponding Eqs. (4-14b) and (4-16b) for a circular pupil, in that
I (r; ) is different from I (r) and that the lower limit of integration is instead of zero.
Figure 4-14 shows the irradiance and encircled-power distributions for an annular
pupil with = 0.5 . Both Gaussian ( g = 1) and uniform annular pupils are considered in
this figure. Since Gaussian illumination broadens the central disc and reduces the power
in the secondary rings, and obscuration reduces the size of and power in the central disc
thereby increasing the power in secondary rings, we note that the difference between the
focal-plane distributions for Gaussian and uniform annular pupils are less compared with
those for corresponding circular pupils. However, note that the values of secondary
maxima decrease monotonically for a Gaussian beam; not so for a uniform beam. We
1.0
0.8 0.750
l(r; R); P(rc ; R) 0.717
0.6 Pg
0.4
Pu
= 0.5
0.2
lg
lu
0.0
0 1 2 3
r; rc
10 – 1
= 0.5
l(r; R)
10 – 2
10 – 3
10 – 4
0 1 2 3 4 5
r
Figure 4-14. Focal-plane irradiance and encircled-power distributions for Gaussian

g = 1 and uniform annular beams of a given total power Pex and = 0.5 . The
irradiance and encircled power are in units of Pex Sex l 2 R 2 and Pex , respectively.
The radial distances r and r c in the focal plane are in units of l F. Irradiance
distributions are also shown on a log scale. The dashed curves are for a Gaussian
beam with g = 1.
also note that the difference Pu - Pg between the encircled powers changes its sign from
positive to negative to positive, as rc increases.
The focal-point irradiance for an annular Gaussian beam may be obtained from
Eq. (4-93) by letting r = 0 . It is given by
I (0; g ; ) = (2 g ) tanh g 1 - 2 2 [( ) ] . (4-95)
The corresponding value for a uniform beam is equal to 1 - 2 . The ratio of the two is
given by
h = I (0; g ; ) I (0; 0; )
[(
= tanh g 1 - 2 2 ) ] [g (1 - ) 2] 2
. (4-96)
Figure 4-15a shows how h varies with g for several values of . It is evident that h
decreases as g increases, regardless of the value of . However, as shown in Figure 4-15b,
for a given value of g, h increases as increases. Note that for large values of g,
h Æ 2 g 1 - 2 ( ) . (4-97)
4.4.3 Strehl Ratio and Aberration Tolerance

For an annular pupil, since £ r £ 1, Eqs. (4-22) and (4-25) for the Strehl ratio and
mean values of F n are replaced by
2
1 2p
S = {g p [exp(- g ) - exp(-g )] } Ú Ú A(r) exp[iF(r, q)] r dr dq
2
0
(4-98)
1.0 1.0 0
1
0.8 0.8
0.6 0.6
h
0.75
2
0.4 0.5 0.4
=0
0.2 0.2 g =3
0.0 0.0
0 1 2 3 4 0.0 0.2 0.4 0.6 0.8 1.0
g
(a) (b)
Figure 4-15. Focal-plane irradiance ratio for Gaussian and uniform beams as a
function of g and .
and
1 2p
<F > = n
{g p [exp (- g ) - exp(- g)] } Ú Ú A(r)[F(r; q)]
2
0
n
r dr dq , (4-99)
respectively.
4.4.4 Balanced Aberrations and Zernike-Gauss Annular Polynomials1, 2

For systems with annular Gaussian pupils, a classical aberration of a certain order
when combined with lower-order aberrations to minimize its variance can be identified
with a corresponding Zernike-Gauss annular polynomial. These polynomials can be
obtained from Zernike annular polynomials given in Section 3.4 by the Gram-Schmidt
orthogonalization process5 in exactly the same manner as the Zernike-Gauss circle
polynomials are obtained from the Zernike circle polynomials, except that now the lower
limit on the radial variable r of integration is instead of zero. The radial polynomials in
this case may be written Rnm (r; g ; ) .
The phase aberration function of a system with an annular exit pupil can be expanded
in terms of a complete set of Zernike-Gauss annular polynomials Z nm (r, q; g; ) that are
orthonormal over a unit annulus weighted by the Guassian amplitude in the form
• n
F(r, q; g; ) = Â Â c nm Z nm (r, q; g; ) , £ r £ 1 , (4-100a)
n =0 m =0
where cnm are the orthonormal expansion coefficients, n and m are positive integers
including zero, n – m ≥ 0 and even, and
[
Z nm (r, q; g; ) = 2( n + 1) (1 + d m 0 ) ]1/ 2Rnm (r; g; ) cos mq . (4-100b)
The polynomials are orthonormal according to
1 2p 1
Ú Ú Z n (r, q; g; ) Z n ¢ (r, q; g; ) A(r) r dr d q 2 pÚ A(r) r dr = d nn ¢ d mm ¢
m m¢
. (4-100c)
0
The radial polynomials are given by
È ( n - m) 2 ˘
Rnm (r; g; ) = M nm Í Rnm (r; ) - Â (n - 2i + 1) Rnm (r; ) Rnm- 2i (r; g; ) Rn - 2i (r; g; )˙ .
ÍÎ i ≥1 ˙˚
(4-101)
The angular brackets indicate an average over the annular Gaussian pupil; i.e.,
1 1
Rnm (r; ) Rnm- 2i (r; g; ) = Ú Rnm (r; ) Rn - 2i (r; g; ) A(r) r dr Ú A(r) r dr . (4-102)

The normalization constant Mnm is chosen such that the radial polynomials obey the
1 1
1
Ú Rnm (r; g ; ) Rnm¢ (r; g ; ) A(r) r dr Ú A(r) r dr = d . (4-103)

n + 1 nn ¢
It should be noted that the radial polynomial Rnn (r; g ; ) is identical to the corresponding
polynomial for a uniformly illuminated annular pupil Rnn (r) , except for the
normalizationconstant, i.e.,
Rnn (r; g; ) = M nn Rnn (r; ) . (4-104)
The radial polynomial Rnm (r; g ; ) is a polynomial of degree n in r containing terms in

rn , rn -2 , ..., and r m whose coefficients depend on the Gaussian amplitude through g,
i.e., it has the form
Rnm (r; g ; ) = anm rn + bnm rn - 2 + K + dnm r m , (4-105)
where the coefficients anm , etc., depend on g and . The number of Zernike (or
orthogonal) aberration terms in the expansion of an aberration function through a certain
order n is given by Eq. (2-63), as in the case of circle or annular polynomials for
uniformly illuminated pupils. The orthonormal Zernike-Gauss expansion coefficients are
given by
1 2p 1
c nm = Ú Ú F(r, q; g; )Z nm (r, q; g; ) A(r) r dr d q 2 pÚ A(r) r dr , (4-106)
0
as may be seen by substituting Eq. (4-100a) and utilizing the othonormality of the
polynom . The Zernike-Gauss annular polynomials Z nm (r, q; g; ) are similar to the
Zernike-Gauss circle polynomials Z nm (r, q; g ) , including their number through a certain
order n, except that they are orthonormal over a unit annulus weighted by the Gaussian
amplitude.
n (r, q; g; ) = c nm Z n (r, q; g; ) .
Fm m
(4-107)
Unless n = m = 0 , its mean value is zero, i.e.,
1 2p 1
n (r, q; g; ) = Ú Ú F n (r, q; g; ) A(r) r dr d q 2 pÚ A(r) r dr = 0 .
Fm m
(4-108)
0
For m = 0 , this may be seen with the help of Eq. (4-100c) and the fact that Z 00 (r; g; ) = 1
is a member of the polynomial set. The orthonormality Eq. (4-100c) yields the result that
the mean value of Rn0 (r; g ; ) is zero. When m π 0 , the average value of cos mq is zero.
1 2p
[F (r, q; g; )] [F (r, q; g; )] A(r) r dr d q
1
2 pÚ A(r) r dr = c nm
2 2
m
n = Ú Ú m
n
2
. (4-109)
0
(F mn ) 2
2
s 2nm = - Fm
n = c nm
2
, n π 0, m π 0 . (4-110)
Thus, each expansion coefficient, with the exception of c00 , represents the standard
deviation of the corresponding aberration term.
From Eqs. (4-107) and (4-110), we note that the standard deviation of an aberration
can be obtained immediately by comparing its form with the corresponding orthonormal
aberration represented by Eq. (4-107), without having to calculate the integrals in Eqs. (4-
108) and (4-109). The variance of an aberration consisting of two or more terms of the
form of Eq. (4-107) is given by the sum of the variances of each of the aberration terms.
The variance of the aberration function is accordingly given by
• n
s 2F = F 2 (r, q) - F(r, q)
2
= Â 2
Â c nm . (4-111)
n =1 m = 0
The radial annular polynomials Rnm (r; g ; ) representing balanced primary

aberrations are given by the same expressions as for circle radial polynomials in Table
4-3 except that now
ps = < rs >
Ë {
= Ê s exp g 1 - 2 [( )] - 1} {exp [g (1 - )] - 1}ˆ¯ + (s 2 g ) p
2
s-2 . (4-112)
Using these expressions, numerical results for the coefficients of the terms of a radial
polynomial for any values of g and can be obtained.
The coefficients for g = 1 and = 0, 0.25, 0.50, 0.75, and 0.90 are given in Table
4-7. For comparison, the coefficients for a uniformly illuminated pupil, i.e., for g = 0 , are
given in parentheses in this table. An increase (decrease) in the value of a coefficient anm
of an orthogonal aberration Rnm (r; g ; ) cos mq implies a decrease (increase) in the value
of s F for a given amount of the corresponding classical aberration. This, in turn, implies
that for small aberrations, the system performance as measured by the Strehl ratio is less
(more) sensitive to that classical aberration when balanced with other classical
aberrations to form an orthogonal aberration. Thus, as increases, irrespective of the
value of g, the system becomes less sensitive to field curvature (defocus) and spherical
aberration but more sensitive to distortion (tilt) and astigmatism. In the case of coma, it
first becomes slightly more sensitive but is much less sensitive for larger values of . As
g increases, i.e., as the width of the Gaussian illumination becomes narrower, the system
becomes less sensitive to all classical primary aberrations. Although the results for g = 0
and g = 1 only are given in Table 4-7, the coefficients for 0 £ g £ 3 show that the
Table 4-7. Coefficients of terms in Zernike-Gauss radial polynomials Rnm (r; g ; ) for
g = 1. The numbers given in parentheses are the corresponding coefficients for
uniform illumination.
a 11 a 20 b20 a 22 a 13 b31 a 40 b40 c40
0.00 1.09367 2.04989 – 0.85690 1.14541 3.11213 – 1.89152 6.12902 – 5.71948 0.83368
(1.00000) (2.00000) (– 1.00000) (1.00000) (3.00000) (– 2.00000) (6.00000) (– 6.00000) (1.00000)
0.25 1.04364 2.18012 – 1.00080 1.08940 3.01573 – 1.84513 6.95563 – 6.98197 1.25153
(0.97014) (2.13333) (– 1.13333) (0.96836) (2.94566) (– 1.97099) (6.82667) (– 7.25333) (1.42667)
0.50 0.92963 2.70412 – 1.56449 0.93620 3.14319 – 2.06618 10.79549 – 13.08900 3.46706
(0.89443) (2.66667) (– 1.66667) (0.87287) (3.11400) (– 2.17980) (10.66667) (– 13.33333) (3.66667)
0.75 0.80827 4.59329 – 3.51548 0.74439 4.55179 – 3.57767 31.47560 – 48.77879 18.39840
(0.80000) (4.57143) (– 3.57143) (0.72954) (4.53877) (– 3.63858) (31.34694) (– 48.97959) (18.63265)
0.90 0.74453 10.53581 – 9.50324 0.63890 9.60573 – 8.69629 166.33359 – 300.66342 135.36926
(0.74329) (10.52632) (– 9.52632) (0.63679) (9.60023) (– 8.72012) (166.20500) (– 300.83102) (135.62604)
differences between the coefficients for uniform and Gaussian illumination are small, and
they decrease as increases and increase as g increases. This is understandable because
as increases or g decreases, the differences between the two illuminations decreases.
4.4.5 Defocused System
4.4.5.1 Theory
The irradiance distribution in a defocused image plane given by Eq. (4-43) is now
replaced by
1 2
I (r; z; g ; ) = (2 R z )
2
Ú (
I (r; ) exp i Bd r 2
) J (p rr) r dr
0 . (4-113)

4.4.5.2 Axial Irradiance
By letting r = 0 in Eq. (4-113), we obtain the axial irradiance of the beam
I (0; z; g ; ) = ( R z ) 2 g
2
[ (B 2
d + g2 )] {coth [g (1 - )] 2
[ (
- cos Bd 1 - 2 )] [(
sinh g 1 - 2 )] } . (4-114)
The positions of its maxima and minima are obtained by equating to zero its derivative
with respect to z. Thus, they are given by the solutions of
Ê lz
2Á
B ˆ
{ [(
- 2 d 2 ˜ cosh g 1 - 2 - cos Bd 1 - 2
Ë Sex Bd + g ¯
)] [ ( ) ]}
( ) [ (
= - 1 - 2 sin Bd 1 - 2 )] (4-115)
Figure 4-16 shows how the axial irradiance of a focused Gaussian beam differs from
that of a focused uniform beam when g = 1, = 0.5 , and Fresnel number N = 1, 10, 100.
We note that, as in the case of a circular beam, the principal maximum is higher for the
uniform beam compared with that for the Gaussian beam. However, the secondary
maxima are higher for the Gaussian beam. Moreover, whereas the axial minima for the
uniform beam have a value of zero, the minima for the Gaussian beam have nonzero
values. For a given value of , the locations of maxima and minima, except the principal
maximum, are nearly the same for the two beams. The effect of the obscuration is to
reduce the irradiance at the principal maximum but to increase it at the secondary
maxima. Also, the maxima and minima occur at smaller z values for an annular beam.
These z values correspond approximately to those axial points at which the annular pupil
subtends an odd or an even number of Fresnel’s half-wave zones, respectively. We note
that the curves become symmetric about the focal point z = R as N increases.
Once again, even though the principal maximum of axial irradiance does not lie at
the focus, maximum central irradiance on a target at a given distance from the pupil is
obtained when the beam is focused on it. Figure 4-17 illustrates how the central
irradiance on a target at a fixed distance z varies when the beam is focused at various
distances R along its axis. The irradiance in this figure is in units of Pex Sex l2 z 2 . The
quantity N z = a 2 l z represents the Fresnel number of a circular pupil as observed from
the target. As in Figure 4-13, the maximum irradiance values for uniform and Gaussian
( g = 1) beams are 0.750 and 0.717, respectively, when = 0.5 . We note that as N z
increases, the curves become symmetric about R = z .
2.0 2.0 2.0
N=1 N = 10 N = 100
1.5 = 0.5 1.5 = 0.5 1.5 = 0.5
l (0; z)
lu
1.0 1.0 1.0
lg
0.5 0.5 0.5
0.0 0.0 0.0

0.0 0.5 1.0 1.5 0.5 1.0 1.5 0.8 0.9 1.0 1.1 1.2
z/R z/R z/R
Figure 4-16. Axial irradiance of an annular beam focused at a fixed distance R with
a Fresnel number N z = a 2 l R = 1 , 10, 100. The irradiance is in units of the focal-
point irradiance for a uniform annular beam. The Gaussian beam results shown in
this figure by the dashed curves are for g = 1.
1.0 1.0 1.0

Nz = 1 Nz = 10 Nz = 100
0.8 lu = 0.5 0.8 = 0.5 0.8 = 0.5
lz (0; R)
lg
0.6 0.6 0.6
0.4 0.4 0.4
0.2 0.2 0.2
0.0 0.0 0.0

0.5 1.0 1.5 2.0 0.5 1.0 1.5 2.0 0.8 0.9 1.0 1.1 1.2
R/z R/z R/z
Figure 4-17. Central irradiance on a target at a fixed distance z from the plane of the
annular exit pupil when a beam is focused at various distances R. The quantity
N z = a 2 l z represents the Fresnel number of the pupil as observed from the target.
The irradiance is in units of Pex Sex l 2 z 2 . The Gaussian beam results shown in this
figure by dashed curves are for g = 1.
4.4.5.3 Defocused Distribution

The irradiance and encircled-power distributions are given by Eqs. (4-47) and (4-48),
respectively, except that now the lower limit is instead of zero and I (r) is replaced by
I (r; ) . An example of a defocused distribution is shown in Figure 4-18 for
Bd = 2.783 rad , as in Figure 4-7. Comparing these two figures, we note that the defocus
aberration does not reduce the central irradiance for the annular beam as much as it does
for the circular beam, so much so that for the amount of defocus aberration considered
here, the defocused central irradiance for the annular beam is higher than that for the
corresponding circular beam. For the uniform and Gaussian circular beams, the central
irradiance decreases from 1 and 0.924 to 0.500 and 0.483, respectively. For annular
beams, it decreases from 0.750 and 0.717 to 0.514 and 0.497, respectively. This indicates
that the tolerance for a radially symmetric aberration, such as defocus, is higher for an
annular beam than that for a circular beam. Whereas for a circular beam, the encircled
power is higher for a uniform beam for small values of rc compared with that for a
Gaussian beam, the difference in encircled power for the two types of beam changes from
positive to negative to positive depending on the value of rc .
4.4.5.4 Collimated Beam

letting R Æ • . Thus, for example, Eq. (4-113) for the irradiance distribution in a plane at
a distance z from the pupil reduces to
1 2
I (r; z; g ; ) = 4 Bd2 Ú ( 2
)
I (r; ) exp i Bd r J 0 ( p r r) r dr , (4-116)

where the defocus coefficient is given by

0.6 1.2
0.514 = 0.5
0.5 0.497 1.0
0.4 0.8
P(rc)
l(r)
0.3 0.6
Pg
0.2 0.4
Pu
0.1 0.2
lg
lu
0.0 0.0
0.0 0.5 1.0 1.5 2.0 2.5
r; rc
Figure 4-18. Defocused irradiance and encircled-power distributions for uniform

and Gaussian annular beams with = 0.5 and g = 1.
Bd = Sex l z (4-117)
and represents the phase aberration of a plane wavefront with respect to a reference
sphere centered at a distance z and passing through the center of the pupil. In Eq. (4-116),
the irradiance in both the pupil and the observation planes is in units of the pupil
irradiance Pex Sex for a uniform circular beam.
In the far field, i.e., for z ≥ D2 l , the irradiance and encircled-power distributions
are similar to the focal-plane distribution. The only difference is in scaling of the
diffraction pattern. Similarly, in the near field, i.e., for z < D2 l , these distributions
correspond to defocused distributions. The only significant difference is in the definition
of Bd .
If z is in units of the far-field distance D2 l , and we let r = 0 in Eq. (4-116), we

obtain the axial irradiance
{ [1 + (4g z p) ] }
I (0; z; g ; ) = 2 g
2
¥ {coth [ g (1 - )] - cos [ p(1 - ) 4 z ]

2 2
[(
sinh g 1 - 2 ) ]} . (4-118)
Its maxima and minima are given by

{2 (4z p) g [1 + (4g z p) ] } { cosh[g (1 - )] - cos [p (1 - ) 4z]}

3 2 2 2 2
= - (1 - ) sin [ p (1 - ) 4 z ]} .
2 2
(4-119)
Figure 4-19 illustrates how the axial irradiance of collimated uniform and Gaussian
annular beams varies with distance z from the pupil for = 0.5 . It is similar to Figure 4-8
for a circular beam. Note, however, that the maxima are higher and minima are lower for
the annular beam.
4.4.6 Symmetry Properties of an Aberrated PSF8

The symmetry properties of a PSF for a Gaussian annular pupil aberrated by a
Zernike-Gauss annular polynomial are similar to those for a Gaussian circular pupil
discussed in Section 4.3.7. Equations (4-88) through (4-90) describing the irradiance
distribution of an aberrated Gaussian beam for circular pupils are modified for annular
pupils to
I (r, q i ; z, g ; ) = ( R p z )
2
2
1 2p
Û Û Ï ¸
I (r; ) exp Ìi ÈÍ F(r, q; g ; ) - p r r cos (q - q i )˘˙ ˝ r dr dq
R
¥ Ù Ù , (4-120)
ı ı Ó Î z ˚ ˛
0
where
N=0
5
= 0.5
lg
4
l(0;z)
lu
2
0
0.0 0.1 0.2 0.3 0.4 0.5
z
Figure 4-19. Axial irradiance of a collimated annular Gaussian beam with = 0.5
and g = 1 compared with that for a corresponding uniform beam.
F(r, q; g ; ) = Anm Rnm (r; g ; ) cos mq + Bd r2 , (4-121)
and
1 2
Û
I (0; z; g ; ) = 4 Ù
ı
(
I (r; ) exp iBd r exp i
2
) [ An 0 Rn0 (r; g ; )] r dr , (4-122)

respectively. As in Eq. (4-90), the Fresnel number in Eq. (4-122) is assumed to be large
so that z ~ R . Once again, the axial irradiance for spherical aberration is not symmetrical
about any point. This is illustrated in Figure 4-20 for primary spherical aberration As r 4
for several values of As , = 0.5 , and g = 1. Moreover, from the form of the Zernike-Gauss
polynomial for astigmatism, the point of axial symmetry for astigmatism Aa r2 cos 2 q is
the same as for circular pupils, namely, Bd = - Aa / 2 .
Now we discuss briefly the line of sight of an aberrated system with a circular or an
annular pupil in terms of the centroid of its PSF. As pointed out in Section 1.8, the
centroid for a Gaussian pupil can not be obtained from the aberration along its perimeter.
Accordingly, unlike in the case of a uniform pupil, the centroids of the PSFs for a circular
pupil aberrated by primary and secondary coma having the same peak value are not
identical. Of course, in the case of an annular pupil, the centroids are different even when
the illumination across it is uniform, as was shown in Section 3.9.1.
1.0
1
W(r) = Asr4 As = 0
I(r) = I0exp(– 2r2)
0.8 2
= 0.5
I(0; Bd)
0.6
3
0.4
0.2
0.0
–6 –5 –4 –3 –2 –1 0 1 2 3
Bd
Figure 4-20. Axial irradiance of an annular Gaussian beam with = 0.5 and g = 1
aberrated by spherical aberration. The quantity I0 represents the coefficient of
( )
exp - 2r2 on the right-hand side of Eq. (4-91) for g = 1, and the aberration-free
central value of 0.717 has been normalized to unity.

The aberrated PSF is given by Eq. (4-120), where we let z = R . For a radially
symmetric pupil illumination, Eqs. (1-177) for the PSF centroid may be written
1 2p
Ê ∂W sin q ∂W ˆ
xi = (2 F p ) Ú Ú I (r; ) ÁË cos q ∂r -
r
˜ r dr d q
∂q ¯
(4-123)
0
and
1 2p
Ê ∂W cos q ∂W ˆ
yi = (2 F p ) Ú Ú I (r; ) ÁË sin q ∂r -
r
˜ r dr d q ,
∂q ¯
(4-124)
0
where W (r, q; g ; ) is the aberration function. Expanding the aberration function in terms
of Zernike-Gauss polynomials, we may write
•
Â Â [2(n + 1) (1 + d m 0 )]
n 12
W (r, q; g ; ) = Rnm (r; g ; )
n=0 m=0
¥ (cnm cos mq + snm sin mq) , (4-125)
where cnm and snm are the aberration coefficients. Substituting Eq. (4-125) into
Eqs. (4-123) and (4-124), we obtain
1
∂
Ú I (r; ) ∂r [r Rn (r; g ; )] r dr
•
xi , yi = (2 FSex Pex ) Â ¢ 2(n + 1) (cn1 , sn1 ) 1
. (4-126)
n =1

We note that, as in the case of systems with uniformly illuminated circular and annular
pupils, the only aberrations that contribute to the centroid are those with m = 1.
Aberrations that vary with q as cos q contribute to < xi > , and those that vary as sin q
contribute to < yi > . This may also be seen from the symmetry of these aberrations in a
manner similar to the discussion following Eq. (2-151).
We now consider aberrations given by Eq. (2-152). Substituting this equation into
Eq. (4-4) and following the procedure of Section 2.9.1, it can be shown that the irradiance
distribution of the aberrated image of a point object may be written
2
È1 ˘
I (r; q i ; g ; ) = 4 Í Ú I (r; ) J 0 ( p B) r dr ˙ , (4-127)
Î ˚
where B is given by Eq. (2-155).

For a wavefront tilt aberration given by W1 r cos q , the PSF simply shifts such that
its peak and centroid locations move to x p = < x > = 2 W1 .

For primary coma W3 r3 cos q , Eq. (4-127) yields the irradiance distribution along
the x axis according to
2
È1 ˘
I ( x; g ; ) = 4 Í Ú I (r; ) J 0 ( p B) dt ˙ , (4-128)
ÍÎ 2 ˙˚
where B = (2tW3 - x ) t 1 2 , as in Eq. (2-164). Substituting Eq. (4-91) and the aberration
W = W3 r3 cos q into Eq. (4-123), we obtain
Ï
Ô 1
= 4 W3 F Ì +
[ (
2 exp 2 g 1 - 2 - 1 ¸Ô
˝ .
)]
[ ( )]
xi (4-129)
ÔÓ 2 g exp 2 g 1 - 2 - 1 Ô
˛
From the form of the radial polynomial R31 (r; g ; ) given in Table 4-3, the point x m in the
image plane with respect to which the aberration variance is minimum is given by
2W3 F p4 p2 , where ps is given by Eq. (4-112). Thus,
Ï
Ô
= 2 W3 F Ì(2 g ) +
g 4 exp g 1 - 2 [( )] - 1 ¸Ô
˝
) [( )] - (1 + g ) Ô˛
xm . (4-130)
ÔÓ (
1 + g 2 exp g 1 - 2
For small values of W3 , the peak value of the aberrated PSF occurs at this point.
Figure 4-21 shows how I ( x ) varies with x for several values of W3 when g = 1 and
= 0 or 2 = 0.5 . The values of x m , x p , and < x > and the corresponding irradiances
Im , I p ,and Ic for these values of W3 are given in Table 4-8. The values of I (0) are also
included in this table. The irradiances given here are normalized by the aberration-free
central irradiance for the annular Gaussian pupil.
1.0 1.0
W3 = 0 W3 = 0
= 0 2 =
0.5 0.5
0.8 g= 1 0.8 g= 1
0.5
1
0.6 0.6
l
1
1.5
0.4 0.4
2
1.5
2
0.2 0.2
0.0 0.0
–3 –2 –1 0 1 2 3 –3 –2 –1 0 1 2 3
x x
Figure 4-21. PSF for a Gaussian pupil with g = 1 aberrated by primary coma W3 (in
units of l ).
Table 4-8. Typical values of x m , x p , and < x > in units of l F and corresponding
irradiances Im , I p , and Ic in units of the aberration-free central irradiance for
Gaussian pupils with g = 1 aberrated by primary coma.*
0 0 0 0 1 1 1 1
(0) (0) (0) (1) (1) (1) (1)
0.50 0.61 0.60 0.69 0.8805 0.8806 0.8670 0.4567

(0.76) (0.76) (1.42) (0.9288) (0.9288) (0.1126) (0.0602)
1.00 1.22 1.15 1.37 0.6013 0.6062 0.5590 0.1708

(1.51) (1.51) (2.84) (0.7435) (0.7435) (0.1273) (0.0348)
1.50 1.82 1.40 2.06 0.3205 0.3672 0.2479 0.1199

(2.27 (2.24) (4.25) (0.5112) (0.5122) (0.0014) (0.0033)
2.00 2.43 1.46 2.75 0.1305 0.2947 0.0624 0.0733

(3.03) (2.93) (5.67) (0.3005) (0.3065) (0.0399) (0.0000)
* The numbers without parentheses are for = 0 , and those with parentheses are for 2 = 0.5 .
The aberrated PSF along the x axis in the presence of secondary coma of the type
W5 r5 cos q is given by Eq. (4-128), where B is now given by Eq. (2-169). Substituting
for the aberration and Eq. (4-91) into Eq. (4-123), we obtain
= 6W F
( 4
) [ ( )] (
+ 2 g -1 + 1 2 g 2 exp 2 g 1 - 2 - 1 + g -1 + 1 2 g 2 ).
exp [2 g (1 - )] - 1
xi 5 (4-131)
2
Figure 4-22 shows how I ( x ) varies with x for several values of W5 when g = 1 and
1.0 1.0
W5 = 0 W5 = 0
= 0 2 = 0.5
0.8 g= 1 0.8 g= 1
0.5 0.5
0.6 0.6
l
1 1.5
1
0.4 0.4
2
1.5
0.2 0.2 2
0.0 0.0
–3 –2 –1 0 1 2 3 –3 –2 –1 0 1 2 3
x x
Figure 4-22. PSF for a Gaussian pupil with g = 1 aberrated by secondary coma W5
(in units of l).
4.6 Summary 383
= 0 or 2 = 0.5 . The values of x p , < x > , I p , Ic , and I (0) are given in Table 4-9. It is
evident from the data given in Tables 4-8 and 4-9 that the centroids of two PSFs aberrated
by equal amounts of primary coma and secondary coma are different. For example, when
W3 = W5 = 1l , < x > 3 = 1.37 and < x > 5 = 1.12 in units of l F . In the case of a uniformly
illuminated circular pupil, < x > 3 = < x > 5 for W3 = W5 , as may be seen from Tables 2-14
and 2-15. In the case of an annular pupil, the centroid values < x > 3 and < x > 5 are
different whether or not the pupil is uniformly illuminated.
4.6 SUMMARY
In this chapter, we have compared the effects of diffraction, obscuration, and
aberrations on the propagation of uniform and Gaussian beams of a fixed total power.
The following general conclusions can be drawn from the discussion given here.
i. The mirrors in a high-power Gaussian beam optical train are illuminated unevenly
and must withstand higher flux densities than those in a corresponding uniform beam
train. This will cause thermal distortions of the mirrors, thereby introducing aberrations in
the beam.
ii. The focal-point irradiance for a focused Gaussian beam is smaller than the
corresponding value for a uniform beam of the same total power.
iii. The encircled power for small circles is higher for a uniform beam, but for large
circles it is higher for a Gaussian beam.
Table 4-9. Typical values of x p and < x > in units of l F and corresponding
irradiances I p and Ic in units of the aberration-free central irradiance of PSFs for
Gaussian pupils with g = 1 aberrated by secondary coma.*
W5 xp <x> Ip Ic I(0)
0 0 0 1 1 1
(0) (0) (1) (1) (1)
0.50 0.41 0.56 0.8452 0.8105 0.6322

(0.59) (1.57) (0.8451) (0.0123) (0.2253)
1.00 0.64 1.12 0.5659 0.4161 0.3793

(1.14) (3.14) (0.5144) (0.0026) (0.0025)
1.50 0.63 1.68 0.4541 0.1147 0.3083

(1.49) (4.70) (0.2595) (0.0075) (0.0011)
2.00 0.74 2.24 0.3824 0.0075 0.2476

(1.67) (6.27) (0.1892) (0.0043) (0.0084)
* The numbers without parentheses are for = 0 , and those with parentheses are for 2 = 0.5 .
iv. The effect of the Gaussian apodization of the pupil is to increase the size of the
central bright disc of the PSF and to decrease the power in the rings of the diffraction
pattern.
v. The effect of a central obscuration in the pupil is opposite to that of the Gaussian
apodization. It reduces the size of the central disc and increases the power in the rings of
the diffraction pattern. Accordingly, as the obscuration increases, the difference between
the diffraction effects of uniform and Gaussian beams decreases.
vi. The minima of axial irradiance for a uniform beam have a value of zero, and those
for a Gaussian beam have nonzero values. Whereas the principal maximum of axial
irradiance for a Gaussian beam has a smaller value than the corresponding value for a
uniform beam, the secondary maxima for a Gaussian beam have higher values. Even
though the principal maximum does not necessarily occur at the focus, maximum central
irradiance and encircled energy are obtained, on a target at a given distance from the
pupil, when the beam is focused on it.
vii. For a < w, the Gaussian beams are somewhat less sensitive to aberrations than
uniform beams. Accordingly, the aberration tolerance is somewhat higher for the
Gaussian beams. However, this tolerance increases rapidly as a becomes increasingly
larger compared with w. This is understandable since, for a > > w , the power in the pupil
is concentrated in a small region near its center; and therefore the aberration in its outer
region has little effect on the irradiance distribution. For a ≥ 2w , the truncation of an
aberration-free Gaussian beam by the pupil has negligible effect on its propagation.
Accordingly, it remains a Gaussian beam as it propagates. However, when the beam is
aberrated, a ≥ 3w is required in order to neglect the effect of its truncation by the pupil.
Moreover, when a ≥ 3w , it is more appropriate to define the aberration coefficients as the
peak aberrations at the beam radius w rather than at the pupil edge a, since the power in
the beam is concentrated in a small region near the center of the pupil, and the effect of
an aberration in its outer region is negligible. With the aberration coefficients defined in
this manner, the beam becomes most sensitive to spherical aberration and least sensitive
to astigmatism, rather than being most sensitive to coma and least sensitive to spherical
aberration.
viii. The equation for imaging of a Gaussian beam by a lens, where its beam waist is the
object, is different from that of conventional imaging. For example, when the incident
beam waist lies in the front focal plane of the lens, the image waist lies in the back focal
plane instead of infinity.
ix. Whereas equal amounts of primary and secondary coma yield the same centroid in
the case of a uniform circular beam, they yield different centroids in the case of a
Gaussian beam.
References 385
REFERENCES
1. V. N. Mahajan, “Uniform versus Gaussian beams: a comparison of the effects of
diffraction, obscuration, and aberrations,” J. Opt. Soc. Am. A3, 470–485 (1986).
2. V. N. Mahajan, “Zernike annular polynomials for imaging systems with annular

pupils,” J. Opt. Soc. Am. 71, 75–85, 1408 (1981); 1 685 (1984); “Zernike annular
polynomials and optical aberrations of systems with annular pupils,” Appl. Opt.
33, 8125–8127 (1994); also, “Zernike-Gauss polynomials for optical systems with
Gaussian pupils,” Appl. Opt. 34, 8057–8059 (1995).
3. C. S. Chung and H. H. Hopkins, “Influence of nonuniform amplitude on the

optical transfer function,” Appl. Opt. 28, 1244–1250 (1989). Unlike our Eq. (4-
19), the corresponding OTF equation in this paper is not properly normalized to
unity at the center, although the numerical results are.
4. S. Szapiel, “Aberration balancing techniques for radially symmetric amplitude

distributions; a generalization of the Maréchal approach,” J. Opt. Soc. Am. 72,
947–956 (1982).

6. L. D. Dickson, “Characteristics of a propagating Gaussian beam,” Appl. Opt. 9,

1854–1861 (1970).
7. S. A. Self, “Focusing of spherical Gaussian beams,” Appl. Opt. 22, 658–661

(1983).
8. V. N. Mahajan, “Symmetry properties of aberrated point-spread functions,” J.

Opt. Soc. Am. A11, 1993–2003 (1994).
9. V. N. Mahajan, “Line of sight of an aberrated optical system,” J. Opt. Soc. Am.

A2, 833–846 (1985).
PROBLEMS
1. Consider a parallel beam of light of wavelength 0.5 mm incident on a lens of

diameter 4 cm and a focal length of 40 cm transmitting a flux of 1 W. Assume that
the transmission of the lens has a Gaussian shape such that its irradiance value at
the edge of the lens is e - 2 of its value at the center. (a) Determine the amount of
power incident on the lens. (b) Calculate the focal point irradiance for an
aberration-free lens. (c) Repeat problem (b) if the lens has one wave of primary
spherical aberration. Determine the point of maximum axial irradiance and the
value of irradiance at this point. What is the value of Strehl ratio? (d) Repeat
problem (b) if the lens has a quarter wave of coma. Determine the location and the
value of maximum of irradiance in the focal plane, and the centroid of the
irradiance distribution. Give the value of the Strehl ratio.
2. Consider a system with a Gaussian pupil of negligible truncation, i.e., with a large
value of g . Show that its defocused OTF representing the normalized
autocorrelation of its pupil function is given by
{[ ] }
t (v ) = exp - (2 g + (2 g ) Bd ) v 2 , 0 £ v £ 1 ,
or
(
t(vi ) = exp - p 2 w 2z vi2 2 ) , 0 £ vi £ D l z ,
where v = vi (l z D) is a spatial frequency normalized by the cutoff frequency

D l z and w z is the Gaussian radius in an image plane at a distance z from the
pupil. This result may also be obtained by zero-order Hankel transforming the PSF
obtained from Eq. (4-56b).
3. (a) Show that the variance of a primary aberration in a system with a circular pupil
and Gaussian illumination with g = 1 is minimized when balanced as follows:
(
F bs (r) = As r 4 - 0.933r2 ) ,
(
F bc (r, q) = Ac r3 - 0.608r cos q , )
and
(
F ba (r, q) = Aa r2 cos 2 q - 0.5r2 ) .
(b) Repeat problem (a) for g = 4 . (c) Repeat problem (a) for an annular pupil with
= 0.5 . Use the results given in Table 4-7. (d) Compare the results obtained in (a)
and (c) with those for a uniformly illuminated circular pupil.
CHAPTER 3
OPTICAL SYSTEMS WITH ANNULAR PUPILS
3.1 Introduction ......................................................................................................... 261

3.2 Aberration-Free System ......................................................................................261
3.2.2 Encircled Power....................................................................................... 265
3.2.3 Ensquared Power ..................................................................................... 265
3.2.4 Excluded Power ....................................................................................... 266
3.2.6 Optical Transfer Function ........................................................................272
3.3 Strehl Ratio and Aberration Tolerance ............................................................. 281
3.3.1 Strehl Ratio ..............................................................................................282
3.3.2 Primary Aberrations ................................................................................283
3.3.3 Balanced Primary Aberrations ................................................................283
3.3.4 Comparison of Approximate and Exact Results......................................284
3.4 Balanced Aberrations and Zernike Annular Polynomials ............................... 291
3.5 Defocused System ................................................................................................. 298
3.5.2 Focused Beam..........................................................................................299
3.5.3 Collimated Beam ..................................................................................... 303
3.6 Symmetry Properties of an Aberrated PSF....................................................... 305
3.7 PSFs and Axial Irradiance for Primary Aberrations ....................................... 308
3.8 2-D PSFs ................................................................................................................311
3.9 Line of Sight of an Aberrated System ................................................................322
3.9.2.1 Wavefront Tilt ........................................................................... 323
3.9.2.2 Primary Coma ........................................................................... 324
3.9.2.3 Secondary Coma ....................................................................... 327
References ......................................................................................................................330
Problems ......................................................................................................................... 331
259
Chapter 3
Optical Systems with Annular Pupils
3.1 INTRODUCTION
In this chapter, we discuss the imaging properties of a system with an annular pupil
in a manner similar to those for a system with a circular pupil. The two-mirror
astronomical telescopes discussed in Chapter 6 of Part I are a typical example of an
imaging system with an annular pupil. The linear obscuration ratios of some of the well-
known telescopes are 0.36 for the 200-inch telescope at Mount Palomar, 0.37 for the 84-
inch telescope at the Kitt-Peak observatory, 0.5 for the telescope at the McDonald
Observatory, and 0.33 for the Hubble Space Telescope.
Expressions for the PSF, OTF, and encircled, ensquared, and excluded powers are
given. The Strehl ratio of an aberrated system is considered and tolerances for primary
aberrations are discussed. Symmetry properties of aberrated PSFs are discussed, and
pictures of the PSFs for primary aberrations are given as examples. The line of sight of an
aberrated system is discussed in terms of the centroid of its PSF. Numerical results are
given and compared with the corresponding results for systems with circular pupils
wherever possible and appropriate.
3.2 ABERRATION-FREE SYSTEM

We start this chapter with a discussion of the PSF, encircled, ensquared, and
excluded powers, and the OTF of an aberration-free system. Equations are developed in a
way that the results for a circular pupil can be obtained as a limiting case of the annular
pupil. It is shown that the obscuration in an annular pupil not only blocks the light
incident on it, but it also reduces the size of the central disc and increases the value of the
secondary maxima of the PSF. It also increases the OTF value at high spatial frequencies
while reducing it at the low frequencies.

Consider, as illustrated in Figure 3-1, an aberration-free optical system imaging a
point object with a uniformly illuminated annular exit pupil having outer and inner radii
of a and a , respectively, where is the linear obscuration ratio of the pupil. The
r
irradiance at a point ri in the image plane with respect to the Gaussian image point is
given by Eq. (1-65), that is
r Ê 2 pi r r ˆ r
2
Û
[(
Ii ( ri; ) = Ii (0; ) Sex
2
)]
( ) ◊
Ù exp Á- l rp ri ˜ d rp
ı Ë R ¯
, (3-1)
where
( )
Sex ( ) = p 1 - 2 a 2 (3-2)
261
262 OPTICAL SYSTEMS WITH ANNULAR PUPILS
Defocused
image plane
ExP
Gaussian
a image plane
a
Figure 3-1. Imaging by a system with an annular exit pupil of inner and outer radii
a and a, respectively.
is the clear area of the obscured exit pupil. The quantity 2 in Eq. (3-2) is sometimes
referred to as the area obscuration ratio. The optical wavefront at the exit pupil is
spherical with a radius of curvature R and center of curvature at the Gaussian image
point. The central irradiance is given by
Ii (0; ) = Pex Sex ( ) l2 R 2 (3-3a)
(
= pPex 1 - 2 ) 4l2 F 2 . (3-3b)
where Pex is the total power in the exit pupil, and, therefore, in the image. For an object
of intensity Bo radiating at a wavelength l at a distance z0 from the entrance pupil of
area Sen ( ) , the total power is given by
[ ]
Pex = h Sen ( ) z02 Bo , (3-4)
where h is the transmission factor of the system for light propagation from its entrance to
its exit pupil. The quantity F in Eq. (3-3b) is given by
F = R D , (3-5)
where D = 2 a is the outer diameter of the exit pupil. It represents the focal ratio
(f-number) of the image-forming light cone exiting from the exit pupil. The integration in
r
Eq. (3-1) is carried over the clear area of the exit pupil such that the position vector rp of
r
a point in its plane satisfies a £ rp £ a .
As in Section 2.2.1, we express the position vectors of points in the pupil and image
planes in polar coordinates according to
r
( )
rp = rp cos q p , sin q p , a £ rp £ a , 0 £ q p < 2 p , (3-6)
and
r
ri = ri (cos q i , sin q i ) , 0 £ q i < 2 p . (3-7)
Substituting Eqs. (3-6) and (3-7) into Eq. (3-1), we obtain
2
a 2p
È 2ip ˘
[
Ii (ri , q i ; ) = Ii (0; ) Sex
2
( ) Û Û
]
Ù Ù exp Í- (
rp ri cos q p - q i ˙ rp drp dq p ) . (3-8)
ı ı Î l R ˚
a 0
Comparing Eqs. (2-7) and (3-8), we note that the significant difference between the
two lies in the lower limit of the integration over rp ; in Eq. (2-7), the lower limit is 0,
indicating an unobscured pupil; in Eq. (3-8) it is , indicating an obscured pupil. The
values of Sex are different in the two equations by a factor of 1 - 2 . The values of Pex
would also be different by this factor if the pupil irradiance were the same in both cases.
For simplicity of equations as well as numerical analysis, we use normalized

quantities
r r
r = rp a (3-9a)
(
= r cos q p , sin q p ) , (3-9b)
r r
r = ri l F (3-10a)
= r (cos q i , sin q i ) , (3-10b)
and
r r
[
I ( r ; ) = Ii ( ri ; ) Pex Sex (0) l2 R 2 ] . (3-11)
Note that in Eq. (3-11), we have normalized the irradiance by the central irradiance for a
system with a circular pupil. Using normalized quantities, Eq. (3-8) may be written
2
1 2p
-1 Û Û
[ (
I (r, q i ; ) = p 2 1 - 2 )] [ ( )]
Ù Ù exp - pir r cos q p - q i r dr dq p
ı ı
. (3-12)
0
Integrating over q p by using Eq. (2-12), we obtain
È1 ˘2
Í Û ˙
[ (
I ( r; ) = 4 1 - 2 )] ÍÙ J ( prr) r dr ˙
ı 0
. (3-13)
ÍÎ ˙˚
Carrying out the radial integration by using Eq. (2-14), we finally obtain
2
1 È 2 J1 ( pr ) 2 J ( pr ) ˘
I ( r ; ) = Í pr - 2 1 . (3-14)
(1 - ) 2
Î pr ˙˚
We note that the irradiance distribution is radially symmetric about the Gaussian image
point r = 0 , as may be expected for a radially symmetric (annular) pupil function. It is
not normalized to unity at the center. Its central value is given by 1 - 2 . Except for a
normalization factor, Eq. (3-14) also gives the PSF of the system. It follows from Eq. (1-
61) that
r r
PSF ( ri ; ) = I i ( ri ; ) Pex (3-15)
We note that as Æ 0 , Eq. (3-14) for the annular pupil reduces to Eq. (2-15) for the
circular pupil. In order that the total power be the same for the two pupils, the irradiance
across the annular pupil must be higher than that for a circular pupil by a factor of
2 -1
(1 - ) . For a given total power Pex in the exit pupil, the central irradiance I (0;) is
smaller by a factor of 1 - 2 compared to that for a circular pupil. However, if the pupil
irradiance is the same in both cases, as in astronomical observations, then Pex () is also
smaller than Pex (0) by a factor of 1 - 2 . Hence, I (0;) will be smaller than I (0; 0) by a
2
(
factor of 1 - 2 . )
The principal maximum of the image irradiance distribution lies at the Gaussian
image point r = 0 , since all the Huygens’ spherical wavelets originating at the spherical
wavefront in the exit pupil arrive in phase at this point and, accordingly, interfere
constructively. From Eq. (3-14), we note that the image irradiance is zero at those values
of r for which
J1 ( p r ) = J1 ( pr ) , r π 0 . (3-16)
These values of r locate the minima of the irradiance distribution. Noting Eq. (2-19), we
find that the secondary maxima lie at those values of r that satisfy
J2 ( p r ) = 2 J2 ( pr ) , r π 0 . (3-17)
The irradiance distribution for a system with a very thin annulus pupil ( Æ 1) may
be obtained from Eq. (3-13) by noting that r ~ 1, and the variation of J 0 ( prr) is
negligibly small over the variation of r that it can be replaced by J 0 ( pr ) . Hence, Eq. (3-
13) reduces to
I (r; Æ 1) = J 02 ( p r ) (3-18)
when normalized by the central irradiance Pex Sex () l2 R 2 .


The amount of power in the image plane contained in a circle of radius rc centered at
the Gaussian image point is given by
rc
Û
Pi (rc ; ) = 2 p Ù Ii (ri ; ) ri dri . (3-19)
ı
0
Substituting Eqs. (3-10a) and (3-11) into Eq. (3-19) and defining a normalized or
fractional encircled power
P(rc ; ) = Pi (rc ; ) Pex , (3-20)
we obtain
rc l F
Û
ı
(
P(rc ; ) = p 2 Ù I (r; ) rdr .
2
) (3-21)
0
If we let rc be in units of lF and substitute Eq. (3-14) into Eq. (3-21), we obtain
È 1 ˘
1 Í Û du ˙
P(rc ; ) = P(rc ) + P ( rc ) - 4 Ù J1 ( p rc u) J1 ( prc u)
2
, (3-22)
1 - 2 Í ı u˙
ÍÎ 0 ˙
˚
where
P(rc ) = 1 - J 02 ( p rc ) - J12 ( p rc ) (3-23)
is the encircled power for a system with a circular exit pupil.
3.2.3 Ensquared Power

The ensquared power in a square region of half-width rs centered on the Gaussian
image point in the image plane is given by1
ÛÛ
Pi (rs ; ) = ÙÙ Ii (ri ; ) ri dri dq i , (3-24)
ıı
rs
where the integration is carried over the square region. Following the same procedure as
in the case of circular pupils (see Section 2.2.3), Eq. (3-24) reduces to
2
Û 8 2 du
Ps (rs ; ) = Pc ( 2 rs ; - ) ( Ù
p 1 - 2 ı ) [ J1 (p rsu) - J1 (prsu)] cos -1 (1 u)
u
, (3-25)
1
where
Ps (rs ;) = Pi (rs ) Pex (3-26)
and
Pc (rc ) = P(rc ) (3-27)
are the fractional ensquared and encircled powers, respectively, and rs and rc are in units
of l F . The first term on the right-hand side of Eq. (3-25) represents the image power
contained in a circle of radius 2rs . The second term gives the image power contained in
a region lying between a circle of radius 2rs and a square of full-width 2rs . Both of
these terms require numerical integration.
3.2.4 Excluded Power

The excluded image power Xi , i.e., the power contained outside a certain area in the
image plane can be calculated quite accurately in closed form if the included area is large
enough so that Xi £ 0.1 Pex . For large arguments we can use the asymptotic expression
for Bessel functions; namely, Eq. (2-33). Thus, for large values of r, Eq. (3-14) can be
written
8 2
I ( r ; ) =
p r (1 - )
4 3 [sin p (r - 1 4) -
2 ]
sin p (r - 1 4) . (3-28)
Noting that the average of a sine square is half and the average of the product of two sines
with different arguments is zero, the average irradiance (indicated by a bar) for large
values of r may be written
4
I (r; ) ~ . (3-29)
p r (1 - )
4 3
Hence, the excluded encircled power is given by

•
Û
Xc (rc ; ) ~ ( 2
p 2 Ù
ı
) I (r; ) r dr
rc
2
= , (3-30)
p 2 rc (1 - )
and the excluded ensquared power is given by
Xs (rs ; ) ~ Û Û
Ù dx Ù dyI (r; )
ı ı
|x| > rs | y| > r s
4 2
= , (3-31)
p 3rs (1 - )
12
(
where r = x 2 + y 2 ) and the subscripts c and s on X indicate a large circular or a square
region of exclusion centered on the Gaussian image point. From Eqs. (3-30) and (3-31),
we note that
Xs (rc ; ) = 0.9 Xc (rc ; ) . (3-32)
The factor of 0.9 between Xs and Xc is independent of . We also note that excluded
-1
power for an annular pupil is larger by a factor of (1 - ) compared to that for a
corresponding circular pupil.
The approximate result of Eq. (3-29) and those that follow from it, although obtained
for the aberration-free case, are valid even when aberrations are present in the system.
This may be seen by substituting Eq. (3-42) given later into Eq. (1-154) and considering
the normalizations used in Eq. (3-29). It should be noted, however, that the value of r for
which Eq. (3-29) is valid increases as aberrations are introduced into the system.

Figure 3-2 shows the irradiance distribution and encircled power for several values
of including zero. The irradiance is normalized to unity at the center in Figure 3-2b.
The values of r for the first several minima are given in Table 3-1 for = 0 (0.1) 0.9 . We
note that the radius of the central bright disc (first dark ring) decreases monotonically as
increases. As Æ 1, this radius approaches a value of 0.76 [first zero of J 0 ( p r ) ]
compared to a value of 1.22 [first zero of J1 ( p r ) ] when = 0 . Moreover, the secondary
maxima become higher as increases. For example, when = 0.5 , the first secondary
maximum has a value of 9.63 percent of the principal maximum compared to a value of
1.75 percent for a circular pupil. The radius of the second dark ring first increases with ,
achieves its maximum value for = 0.4 , and then decreases. The radius of the third dark
ring first decreases, then increases, achieving its maximum value for = 0.5 , and then
decreases again. The radius of the fourth dark ring first increases, then decreases,
increases again, and finally decreases as increases.
Figure 3-3 shows the irradiance distribution given by Eq. (3-14) for = 0 , 0.25, 0.5,
and 0.75 normalized to unity at the center. This figure and Table 3-2 also show how
Pc , Ps and Ps - Pc vary with rc in graphical and tabulated forms. The value of Pc in a
given dark ring decreases or increases with in a manner similar to how its radius varies,
although a peak or a valley in its variation is not achieved for the same value of . For
small values of , the first maximum of Ps - Pc is the highest. However, for large values
of one of the secondary maxima is the highest. As increases, the secondary maxima
of irradiance become increasingly more significant, and therefore, Eqs. (3-30) and (3-31)
give accurate results for increasingly larger values of rc and rs , respectively. For
example, if the difference between actual results (given in Table 3-2) and those obtained
from Figure 3-2 is to be less than 2.6 percent of the total power, rc must be larger than
1.7, 1.8, and 3.8 when is equal to 0.25, 0.50, and 0.75, respectively. A general rule of
thumb is that for accurate results (less than 1 percent error), rc must be large enough so
1 1
10 – 1 10 – 1
= 0 = 0.5
l(r; R)
l(r; R)
10 – 2 10 – 2
10 – 3 10 – 3
10 – 4 10 – 4
0 1 2 3 4 5 0 1 2 3 4 5
r r
Figure 3-2a. Irradiance and encircled power distributions for an annular pupil. is
the obscuration ratio of the pupil. The example of a circular pupil is shown for
comparison. The dashed curves are for a Gaussian beam discussed in Chapter 4.
1.0
0.9 I
P
=0
0.8
0.7 0.25
I (r), P(rc)
0.6
0.5
0.50
0.4
0.3 0.75
0.2
0.1
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
r; rc
Figure 3-2b. The irradiance and encircled power distributions for various values of
.
that X is less than 10 percent. It is interesting to note that although a square region of a
certain width has 27 percent more area than a circular region of the same diameter, the
difference between the ensquared and encircled powers is less than 9 percent of the total
power, regardless of the value of . (A maximum difference of 8.97 percent occurs when
= 0.54 .)
Table 3-1. Positions r of PSF maxima and minima for an annular pupil in units of
l F, and the corresponding irradiance and the encircled power.
0.0 0.1 0.2 0.3 0.4
Max/
Min r, rc I(r) P(rc) r, rc I(r) P(rc) r, rc I(r) P(rc) r, rc I(r) P(rc) r, rc I(r) P(rc)
Max 0 1 0 0 1 0 0 1 0 0 1 0 0 1 0
Min 1.22 0 0.838 1.21 0 0.818 1.17 0 0.764 1.11 0 06.82 1.06 0 0.584
Max 1.63 0.0175 0.867 1.63 0.0206 0.853 1.63 0.0304 0.818 1.61 0.0475 0.766 1.58 0.0707 0.702
Min 2.23 0. 0.910 2.27 0 0.906 2.36 0 0.900 2.42 0 0.899 2.39 0 0.885
Max 2.68 0.0042 0.922 2.68 0.0031 0.914 2.69 0.0015 0.904 2.73 0.0011 0.902 2.77 0.0033 0.893
Min 3.24 0 0.938 3.18 0 0.925 3.09 0 0.908 3.10 0 0.904 3.30 0 0.903
Max 3.70 0.0016 0.944 3.70 0.0024 0.936 3.68 0.0037 0.926 3.64 0.0028 0.916 3.66 0.0007 0.905
Min 4.24 0 0.952 4.32 0 0.949 4.37 0 0.947 4.22 0 0.929 4.04 0 0.907
Max 4.71 0.0008 0.957 4.71 0.0004 0.951 4.74 0.0004 0.949 4.75 0.0016 0.938 4.66 0.0028 0.922
Min 5.24 0 0.961 5.15 0 0.953 5.16 0 0.951 5.42 0 0.949 5.31 0 0.939
Max 5.72 0.0004 0.964 5.71 0.0008 0.959 5.69 0.0006 0.955 5.73 0.0001 0.950 5.79 0.0008 0.944
Min 6.24 0 0.968 6.35 0 0.965 6.23 0 0.959 6.07 0 0.950 6.43 0 0.950
Max 6.72 0.0003 0.970 6.73 0.0001 0.966 6.74 0.0004 0.962 6.67 0.0006 0.955 6.72 0.0001 0.950
Min 7.25 0 0.972 7.14 0 0.967 7.35 0 0.966 7.27 0 0.961 7.03 0 0.950
Max 7.73 0.0002 0.974 7.72 0.0003 0.970 7.72 0.0001 0.967 7.77 0.0003 0.963 7.65 0.0004 0.954
Min 8.25 0 0.975 8.34 0 0.974 8.11 0 0.967 8.38 0 0.966 8.22 0 0.958
Max 8.73 0.0001 0.977 8.74 0.0001 0.975 8.72 0.0003 0.971 8.72 0.0000 0.966 8.77 0.0004 0.962
Min 9.25 0 0.978 9.16 0 0.975 9.38 0 0.974 9.06 0 0.967 9.46 0 0.966
Max 9.73 0.0001 0.979 9.72 0.0001 0.977 9.75 0.0000 0.975 9.70 0.0002 0.970 9.78 0.0000 0.966
Min 10.25 0 0.980 10.30 0 0.979 10.16 0 0.975 10.32 0 0.973 10.13 0 0.966
0.5 0.6 0.7 0.8 0.9
Max/
Min r, rc I(r) P(rc) r, rc I(r) P(rc) r, rc I(r) P(rc) r, rc I(r) P(rc) r, rc I(r) P(rc)
Max 0 1 0 0 1 0 0 1 0 0 1 0 0 1 0
Min 1.000 0 0.479 0.95 0 0.372 0.90 0 0.269 0.85 0 0.172 0.81 0 0.082
Max 1.54 0.0963 0.618 1.48 0.1203 0.512 1.41 0.1395 0.389 1.35 0.1527 0.256 1.28 0.1600 0.124
Min 2.29 0 0.829 2.17 0 0.717 2.06 0 0.560 1.95 0 0.376 1.85 0 0.184
Max 2.76 0.0124 0.859 2.69 0.0306 0.784 2.58 0.0533 0.649 2.47 0.0734 0.456 2.35 0.0861 0.229
Min 3.49 0 0.901 3.39 0 0.873 3.22 0 0.761 3.05 0 0.554 2.90 0 0.284
Max 3.78 0.0004 0.902 3.84 0.0045 0.886 3.74 0.0192 0.808 3.57 0.0401 0.619 3.40 0.0566 0.328
Min 4.12 0 0.903 4.52 0 0.902 4.38 0 0.865 4.16 0 0.695 3.95 0 0.379
Max 4.50 0.0009 4.80 0.0001 0.903 4.86 0.0050 0.880 4.68 0.0218 0.741 4.46 0.0404 0.421
Min 5.05 0 0.910 5.11 0 0.903 5.52 0 0.899 5.27 0 0.795 5.00 0 0.468
Max 5.66 0.0022 0.923 5.58 0.0004 0.905 5.91 0.0005 0.901 5.78 0.0110 0.824 5.51 0.0299 0.507
Min 6.30 0 0.938 6.00 0 0.906 6.47 0 0.903 6.37 0 0.857 6.05 0 0.549
Max 6.81 0.0008 0.943 6.61 0.0016 0.916 6.72 0.000 0.903 6.87 0.0048 0.872 6.56 0.0224 0.584
Min 7.50 0 0.950 7.19 0 0.925 6.97 0 0.903 7.47 0 0.889 7.10 0 0.622
Max 7.79 0.0000 0.950 87.75 0.0013 0.943 7.53 0.0004 0.905 7.95 0.0016 0.894 6.61 0.0169 0.652
Min 8.12 0 0.950 8.40 0 0.944 7.98 0 0.906 8.57 0 0.901 8.16 0 0.685
Max 8.62 0.0001 0.951 8.87 0.0004 0.947 8.58 0.0010 0.913 8.98 0.0003 0.902 8.67 0.0127 0.711
Min 9.05 0 0.952 9.53 0 0.950 9.13 0 0.919 9.58 0 0.903 9.21 0 0.739
Max 9.68 0.0004 0.957 9.80 0.0000 0.950 9.69 0.0011 0.927 9.83 0.0000 0.903 9.72 0.0094 0.761
Min 10.31 0 0.962 10.11 0 0.950 10.28 0 0.935 10.10 0 0.903 10.26 0 0.784
1.0 1.0
In In Ps
Ps
0.8 0.8 Pc
Pc
l(r), Pc(rc), Ps(rs)

0.6 0.6
= 0.25 = 0.50
0.4 0.4 10(Ps – Pc)

10(Ps – Pc)
0.2 0.2
0.0 0.0
0 1 2 3 4 0 1 2 3 4
r, rc, rs r, rc, rs
1.0
In Ps
0.8 Pc
= 0.75
0.6
10(Ps – Pc)
0.4
0.2
0.0
0 2 4 6 8
r, rc, rs
Figure 3-3. Encircled and ensquared power distributions for an annular pupil. The
irradiance distribution and the difference between ensquared and encircled power
distributions are also shown.
An interesting observation comes about when the irradiance distribution is

considered for large values of r and large values of . Figure 3-4 shows the distributions
for = 0 , 0.5, 0.8, and 1. We note that for a circular pupil, the distribution consists of
maxima and minima indicating a bright central disc surrounded by dark and bright rings.
The successive maxima decrease in value monotonically. However, for an annular pupil,
the distribution consists of not only the bright and dark rings but also of a periodic ring
group structure. The number of rings in a group is given by2
n = 2 (1 - ) , (3-33)
Table 3-2. Encircled and ensquared powers for a centrally obscured circular pupil
with a linear obscuration ratio of .
= 0 = 0.25 = 0.50 = 0.75
rc , rs Pc Ps Pc Ps Pc Ps Pc Ps
0.0 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000

0.1 0.0244 0.0309 0.0228 0.0290 0.0182 0.0231 0.0106 0.0134
0.2 0.0940 0.1178 0.8787 0.1099 0.0696 0.0869 0.0400 0.0496
0.3 0.1989 0.2444 0.1851 0.2270 0.1450 0.1766 0.0817 0.0984
0.4 0.3248 0.3889 0.3007 0.3586 0.2317 0.2730 0.1271 0.1469
0.5 0.4559 0.5290 0.4191 0.4833 0.3161 0.3582 0.1675 0.1847
0.6 0.5775 0.6475 0.5262 0.5851 0.3869 0.4205 0.1969 0.2073

0.7 0.6785 0.7351 0.6120 0.6565 0.4372 0.4572 0.2132 0.2170
0.8 0.7532 0.7910 0.6719 0.6984 0.4658 0.4736 0.2188 0.2209
0.9 0.8011 0.8210 0.7068 0.7181 0.4770 0.4804 0.2193 0.2280
1.0 0.8264 0.8339 0.7221 0.7260 0.4786 0.4889 0.2218 0.2448
1.2 0.8378 0.8434 0.7260 0.7418 0.4897 0.5399 0.2553 0.3114

1.4 0.8417 0.8603 0.7412 0.7817 0.5499 0.6355 0.3319 0.3920
1.6 0.8623 0.8839 0.7868 0.8334 0.6538 0.7332 0.4150 0.4478
1.8 0.8896 0.9020 0.8429 0.8736 0.7544 0.7977 0.4618 0.4770
2.0 0.9064 0.9112 0.8826 0.8940 0.8126 0.8273 0.4699 0.5033
2.2 0.9099 0.9161 0.8979 0.9003 0.8282 0.8403 0.4773 0.5423

2.4 0.9110 0.9217 0.8998 0.9018 0.8293 0.8529 0.5177 0.5918
2.6 0.9180 0.9291 0.9006 0.9042 0.8407 0.8696 0.5848 0.6399
2.8 0.9287 0.9360 0.9032 0.9083 0.8641 0.8862 0.6425 0.6754
3.0 0.9359 0.9402 0.9045 0.9128 0.8867 0.8970 0.6665 0.6975
3.5 0.9394 0.9471 0.9126 0.9289 0.9015 0.9031 0.6948 0.7573

4.0 0.9513 0.9548 0.9384 0.9430 0.9031 0.9061 0.7916 0.8168
4.5 0.9534 0.9591 0.9418 0.9468 0.9051 0.9131 0.8050 0.8439
5.0 0.9607 0.9638 0.9488 0.9515 0.9100 0.9212 0.8576 0.8756
5.5 0.9621 0.9666 0.9501 0.9546 0.9163 0.9323 0.8712 0.8862
6.0 0.9671 0.9698 0.9524 0.9580 0.9356 0.9418 0.8884 0.8971

6.5 0.9681 0.9718 0.9546 0.9623 0.9383 0.9455 0.8978 0.9006
7.0 0.9717 0.9741 0.9639 0.9665 0.9468 0.9499 0.9005 0.9002
7.5 0.9724 0.9756 0.9657 0.9680 0.9497 0.9519 0.9027 0.9048
8.0 0.9752 0.9773 0.9666 0.9695 0.9500 0.9547 0.9029 0.9187
which is equal to the ratio of the outer diameter and the width of the annulus, provided
that n is an integer. The group minima are the lowest ring maxima and correspond to ring
numbers that are multiples of n, e.g., 10, 20, 30, etc., for = 0.8 . The radius of a ring
group is also a multiple of n (in units of l F ) since the spacing between two successive
maxima or minima is approximately unity. The central bright spot or the first dark ring of
radius 1.22 contains 83.8 percent of the total power in the image when = 0 . For
= 0.8 , as may be seen from Table 3-1, the first dark ring has a radius of 0.85 and
contains only 17.2 percent of the total power. However, the central ring group in this case
has a radius of 10.10 and contains 90.3 percent of the total power. When n is not an
integer, the distribution becomes complex. For example, for = 0.7 , n = 6.67 , and the
distribution has a double periodicity with the number of maxima in the two periods equal
to 6 and 7 (two integers closest to n).
3.2.6 Optical Transfer Function

From Eq. (1-73b), the OTF of an optical system with a uniformly illuminated
aberration-free annular exit pupil may be written
r Û r r r r
ı
( ) ( )
t ( vi ; ) = Pex-1 Ù A rp A rp - l R vi d rp , (3-34)
where Sex () is given by Eq. (3-2) and
(r ) 12 r
A rp = [ Pex Sex ( )] , a £ rp £ a
= 0, otherwise . (3-35)
The OTF is radially symmetric and represents the fractional area of overlap of two annuli
whose centers are separated by a distance l Rvi . Figures 3-5 show the overlap region of
the two annuli for various separations for >£ 1 3. We note from the figure that to obtain
the OTF, we need expressions for the overlap area of two circles of equal radii whose
centers are separated by a distance l Rvi and two circles of different radii also separated
by l Rvi . An expression for the former is given by Eq. (2-41). For the latter, we refer to
Figure 3-6 showing two circles of radii a1 and a2 separated by a distance d, and let
(r ) 12
A1 rp = [ Pex Sex ( )] ,
r
rp £ a1
= 0, otherwise (3-36)
and
(r ) 12
A2 rp = [ Pex Sex ( )] ,
r
rp £ a2 . (3-37)
10 0 10 0
10 –1 = 0.0 10 –1 = 0.5
10 –2 10 –2
10 –3 10 –3
I (r,)
10 –4 10 –4
10 –5 10 –5
10 –6 10 –6
10 –7 10 –7
0 10 20 30 40 50 0 10 20 30 40 50
10 0 10 0
10 –1 = 0.8 10 –1 = 1.0
10 –2 10 –2
10 –3 10 –3
I (r,)
10 –4 10 –4
10 –5 10 –5
10 –6 10 –6
10 –7 10 –7
0 10 20 30 40 50 0 10 20 30 40 50
r r
Figure 3-4a. Irradiance distribution for systems with circular ( = 0) and annular
( π 0) pupils. The case Æ 1 represents the limiting case of a totally obscured
pupil. In practice, it approximates the PSF for a system with a very thin annular or
a ring pupil.
=0 = 0.5
= 0.8 Æ1
Figure 3-4b. 2-D PSFs for systems with circular ( = 0) and annular ( π 0) pupils.
0 < n< n=
< n < (1 – )/2 n = (1 – )/2
(1 – )/2 < n < (1 + )/2 n = (1 + )/2
(1 + )/2 < n < 1 n= 1
Figure 3-5a. OTF of an aberration-free system with an annular pupil as the

fractional area of overlap of two annuli whose centers are separated by a distance
l Rvi . There are three cases of interest: >£ 1 3 . (a) = 0.25 .
n n
< n
n

Figure 3-5b. OTF of an aberration-free system with an annular pupil as the

l Rvi . (b) = 1 3.
0 < n < (1 – )/2 n = (1 – )/2
(1 – )/2 < n < n=
< n < (1 + )/2 n = (1 + )/2
Figure 3-5c. OTF of an aberration-free system with an annular pupil as the

l Rvi . (c) = 0.5 .
a1 a2
q1
q2
O A O¢
Figure 3-6. Overlap area of two circles of radii a1 and a2 separated by a distance d.
The overlap area shown shaded in Figure 3-6 is given by
Û r r r r
ı
( ) ( ) 2
[( 2
◊ ) (
Ù A1 rp A2 rp - d d rp = [ Pex Sex ( )] q1a1 - OA AB + q 2 a2 - AO¢ AB ◊ )]
(
= [ Pex Sex ( )] q1a12 + q 2 a22 - da1 sin q1 ) , (3-38)
where
d 2 + a12 - a22
cos q1 = (3-39a)
2 da1
and
d 2 + a22 - a12
cos q 2 = . (3-39b)
2 da2
In Figure 3-5, d = l Rvi , a1 = a , and a2 = a . Hence, the OTF representing the

fractional overlap area in this figure (since the amplitude across the pupil is uniform) is
given by
1
t (v; ) =
1 - 2
[ ]
t (v) + 2 t (v ) - t12 (v; ) , 0 £ v £ 1 , (3-40)
where t (v) is given by Eq. (2-44) and represents the OTF of the system if there were no
obscuration, v is a normalized radial spatial frequency defined by Eq. (2-43), and
t12 (v; ) = 2 2 , 0 £ v £ (1 - ) 2 (3-41a)
( )
= (2 p) q1 + 2 q 2 - 2 v sin q1 , (1 - ) 2 £ v £ (1 + ) 2 (3-41b)
= 0, otherwise . (3-41c)
In Eq. (3-41b), q1 and q 2 are given by
4v 2 + 1 - 2
cos q1 = (3-41d)
4v
and
4v 2 - 1 + 2
cos q 2 = , (3-41e)
4 v
respectively. It is evident that the cutoff frequency, v = 1 or vi = 1 l F , is the same as that

for a circular pupil. Moreover, we note from Eq. (3-40) that at least for spatial frequencies
1+
< v < 1 , t (v; ) > t (v)
2
-1
( )
by a factor of 1 - 2 . It may be seen from Figure 3-5 that the overlap area in this
frequency range is independent of , but the fractional area is larger owing to the smaller
area of the obscured exit pupil. For a thin annular pupil, as Æ 1, a sharp peak near the
cutoff frequency is obtained. The peak frequency represents the spatial frequency of
fringes obtained in a two-dimensional analog of a Young’s double-slit aperture.3
How t (v;) varies with v is shown in Figure 3-7 for various values of , including
zero. We note that an annular pupil gives a higher OTF at high frequencies but a lower
OTF at low frequencies, compared to the OTF for a corresponding circular ( = 0) pupil.
This is the frequency domain analog of smaller radius of the central bright ring and
higher secondary maxima of the PSF for an annular pupil compared to those for a circular
pupil. Table 3-3 gives numerical values of t (v;) for = 0 (0.05) 0.95 and v = 0 (0.05) 1.
As pointed out in Section 2.2.5, the slope of t (v) at the origin is equal to - 4 p .
From Eq. (3-40) we find that the slope of the OTF for an annular pupil at the origin is
given by
t ¢(0; ) = - 4 p (1 - ) . (3-42)
As pointed out in Section 1.6.4, this slope does not change when aberrations are
introduced into the system. Equation (3-42) may also be obtained from Eq. (1-167) by
noting that t (v;) is radially symmetric, v is in units of 1 l F , L = 2 p (1 + ) a , and
( )
Sex = p 1 - 2 a 2 . We also note that
1.0
0.8
= 0
0.6
t (n; )
0.25
0.4 0.50
0.75
0.2
0.0
0.0 0.2 0.4 0.6 0.8 1.0
n
Figure 3-7. OTF of an aberration-free system with an annular pupil of obscuration

ratio .
1
Û
Ù t (v;) vdv = 1 - 8 .
ı
2
( ) (3-43)
0
Since, according to Eqs. (1-83) and (1-84), the PSF and the OTF of a rotationally
symmetric system are related to each other by a zero-order Hankel transform, Eq. (3-40)
may also be obtained by taking such a transform of Eq. (3-14). This was done by
O’Neill 3 who obtained an expression for the OTF similar to Eq. (3-40) except that he
used an angle y instead of the angles q1 and q 2 of Eq. (3-41b), where, as may be seen
from Figure 3-6, y is the angle OBO¢ so that
y = p - (q1 + q 2 ) (3-44a)
and
1 + 2 - 4v 2
cos y = . (3-44b)
2
To show that his expression for the OTF is equivalent to the one given by Eq. (3-40), we
note from Figure 3-6 that
AB = a1 sin q1 = a2 sin q 2 .
Therefore,
= a2 a1 = sin q1 sin q 2
and
Table 3-3. Aberration-free OTF t (v; ) of optical systems with annular pupils.
v\ 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45
0 1 1 1 1 1 1 1 1 1 1
0.05 0.936 0.934 0.930 0.925 0.921 0.915 0.909 0.909 0.894 0.884
0.10 0.873 0.870 0.869 0.852 0.842 0.831 0.819 0.805 0.789 0.769
0.15 0.810 0.807 0.798 0.782 0.766 0.749 0.731 0.709 0.685 0.656
0.20 0.747 0.744 0.734 0.718 0.695 0.670 0.645 0.616 0.583 0.544
0.25 0.685 0.682 0.672 0.655 0.630 0.597 0.563 0.526 0.484 0.435
0.30 0.624 0.620 0.610 0.592 0.566 0.532 0.488 0.441 0.389 0.345
0.35 0.564 0.560 0.549 0.531 0.504 0.468 0.422 0.374 0.334 0.299
0.40 0.505 0.501 0.490 0.470 0.442 0.413 0.381 0.347 0.307 0.272
0.45 0.447 0.443 0.431 0.417 0.400 0.380 0.357 0.330 0.298 0.260
0.50 0.391 0.390 0.385 0.378 0.367 0.354 0.337 0.316 0.291 0.261
0.55 0.337 0.338 0.340 0.340 0.336 0.328 0.317 0.303 0.284 0.261
0.60 0.285 0.285 0.288 0.291 0.297 0.297 0.294 0.286 0.274 0.257
0.65 0.235 0.235 0.237 0.240 0.245 0.251 0.258 0.260 0.257 0.247
0.70 0.188 0.188 0.190 0.192 0.196 0.201 0.207 0.214 0.224 0.227
0.75 0.144 0.144 0.146 0.148 0.150 0.154 0.159 0.164 0.172 0.181
0.80 0.104 0.104 0.105 0.106 0.108 0.111 0.114 0.119 0.124 0.131
0.85 0.068 0.068 0.069 0.070 0.072 0.073 0.075 0.078 0.081 0.085
0.90 0.037 0.037 0.038 0.038 0.039 0.040 0.041 0.043 0.045 0.047
0.95 0.013 0.013 0.013 0.014 0.014 0.014 0.015 0.015 0.016 0.017
1 1 0 0 0 0 0 0 0 0 0
v\ 0.50 0.55 0.60 0.65 0.70 0.75 0.80 0.85 0.90 0.95
0 1 1 1 1 1 1 1 1 1 1
0.05 0.873 0.859 0.841 0.818 0.788 0.745 0.682 0.576 0.364 0.163
0.10 0.746 0.718 0.683 0.637 0.577 0.492 0.365 0.254 0.164 0.080
0.15 0.621 0.579 0.526 0.457 0.367 0.289 0.224 0.165 0.108 0.054
0.20 0.498 0.441 0.371 0.310 0.258 0.211 0.167 0.124 0.082 0.041
0.25 0.377 0.327 0.283 0.243 0.206 0.169 0.134 0.100 0.066 0.033
0.30 0.306 0.270 0.236 0.204 0.173 0.143 0.113 0.084 0.056 0.028
0.35 0.266 0.236 0.207 0.179 0.151 0.125 0.099 0.074 0.049 0.024
0.40 0.241 0.213 0.187 0.161 0.136 0.112 0.089 0.066 0.044 0.022
0.45 0.227 0.199 0.173 0.149 0.126 0.104 0.082 0.061 0.040 0.020
0.50 0.224 0.192 0.166 0.141 0.119 0.097 0.077 0.057 0.037 0.019
0.55 0.231 0.194 0.163 0.138 0.115 0.093 0.073 0.054 0.036 0.018
0.60 0.235 0.207 0.169 0.138 0.114 0.091 0.071 0.052 0.034 0.017
0.65 0.233 0.213 0.186 0.148 0.117 0.092 0.071 0.052 0.034 0.016
0.70 0.222 0.211 0.194 0.167 0.128 0.097 0.073 0.952 0.034 0.016
0.75 0.192 0.195 0.190 0.175 0.151 0.111 0.078 0.055 0.035 0.017
0.80 0.139 0.149 0.163 0.166 0.157 0.135 0.194 0.061 0.037 0.017
0.85 0.091 0.098 0.106 0.118 0.134 0.136 0.120 0.077 0.042 0.019
0.90 0.050 0.054 0.058 0.065 0.073 0.085 0.104 0.102 0.060 0.023
0.95 0.018 0.019 0.021 0.023 0.026 0.030 0.037 0.048 0.070 0.041
1 0 0 0 0 0 0 0 0 0 0
1+ sin q1 + sin q 2

= -
1- sin q1 - sin q 2
q1 + q 2 ˆ q - q2 ˆ
= - tan Ê cot Ê 1 .
Ë 2 ¯ Ë 2 ¯
Or,
p q - q2 ˆ 1+ p q + q2 ˆ
tan Ê + 1 = tan Ê - 1
Ë2 2 ¯ 1- Ë 2 2 ¯
giving
1 Ê1 + yˆ
(p + q1 - q 2 ) = tan -1 Á 1 - tan 2 ˜ .
2 Ë ¯
Hence, we may write
1 1
q1 + 2 q 2 = p 2 -
2
( ) ( )
1 + 2 y + 1 - 2 ( p + q1 - q 2 )
2
1 Ê1 + yˆ
= p 2 -
2
( ) ( )
1 + 2 y + 1 - 2 tan -1 Á
Ë1 -
tan ˜
2¯
. (3-45a)
We also note from Figure 3-6 that
d AB = (OA + O¢A) AB ,
or
2 v sin q1 = sin (q1 + q 2 ) = sin y . (3-45b)
Substituting Eqs. (3-45) into Eq. (3-41b), we obtain
È 1 Ê1 + yˆ˘
2
( ) ( )
t12 (v; ) = (2 p) Íp 2 - 1 + 2 y - sin y + 1 - 2 tan -1 Á
Ë1 -
tan ˜ ˙ .
2 ¯˚
(3-46)
Î
In Eq. (3-40), if we replace Eq. (3-41b) by Eq. (3-46), we obtain O’Neill’s expression for
the OTF, an expression that was obtained by Steel4 earlier using the convolution
approach. However, it seems that Eq. (3-41b) is simpler than Eq. (3-46).
3.3 STREHL RATIO AND ABERRATION TOLERANCE5

Now, we discuss how the obscuration affects the Strehl ratio. It is shown that the
standard deviation of some aberrations increases as the obscuration increases, while for
others it decreases. Correspondingly, the tolerance for an aberration decreases or
increases depending on its type. The balanced primary aberrations giving minimum
variance are discussed and identified with the corresponding Zernike annular
polynomials.
3.3.1 Strehl Ratio

The irradiance distribution of the image of a point object formed by an aberrated
optical system is given by Eq. (1-64). Following the normalizations and notation of
Section 3.1.1, this equation for a system with a uniformly illuminated annular pupil may
be written
2
1 2p
-2 Û Û
[(
I (r, q i ; ) = p 1 - 2 )] [ ( )] [ ( )]
Ù Ù exp iF r, q p ; exp - pi r cos q p - q i r dr dq p
ı ı
, (3-47)
0
( )
where F r, q p ; is the aberration of the system. By definition, the Strehl ratio of the
image or the system is given by the ratio of the irradiances at the center r = 0 with and
without aberrations. The aberration-free central irradiance is unity [as may be seen from
Eq. (3-47)] in units of Pex Sex () l2 R 2 [see Eq. (3-3a)]. Hence, the central irradiance
according to Eq. (3-47) gives the Strehl ratio, i.e.,
2
1 2p
-2 Û Û
[(
S = p 1 - 2 )] [
Ù Ù exp iF(r, q; ) r dr dq
ı ı
] , (3-48)
0
where we have dropped the subscript p on the angle q p for simplicity. Approximate
expressions for the Strehl ratio when the aberration is small are given by Eqs. (1-204)
through (1-206), i.e.,
2
S1 ~ (1 - s 2F 2) , (3-49a)
S2 ~ 1 - s 2F , (3-49b)
and
S3 ~ exp (- s 2F ) , (3-49c)
where
s2F = < F2 > - < F > 2 (3-50a)
is the variance of the aberration across the uniformly illuminated annular pupil. The mean
1 2p
-1 ÛÛ
n
< F > = p 1- [( 2
)] Ù Ù F (r, q; ) r dr dq ,
ıı
n
(3-50b)
0
with n = 1 and 2, respectively.

3.3.2 Primary Aberrations

Table 3-4 gives the form as well as the standard deviation s F of a primary
aberration. For small aberrations, the tolerance, i.e., the value of the aberration coefficient
Ai , for a certain Strehl ratio may be obtained by using any of the Strehl ratio expressions
given above. As in Chapter 2, we will use the symbols Ad and Bd for field curvature and
defocus, respectively. Although their dependence on the pupil coordinates is the same,
the former varies quadratically with the object field angle, while the latter is independent
of it. Similalry, we will use the symbols At and Bt for distortion and wavefront tilt,
respectively. Their dependence on the pupil coordinates is the same, but the former varies
as the cube of the field angle, while the latter is independent of it.
3.3.3 Balanced Primary Aberrations

According to diffraction, the best image is one for which the Strehl ratio is
maximum. Since, according to Eqs. (3-49), the Strehl ratio is maximum when the
aberration variance, assuming it to be small, is minimum, the best image plane is one that
minimizes the variance of the aberration. Thus, for example, we balance spherical
aberration with defocus and write it as
F(r; ) = As r 4 + Bd r2 . (3-51)
We determine the amount of defocus Bd such that the variance sF2 is minimized; i.e., we
calculate sF2 and let
∂s 2F
= 0 (3-52)
∂Bd
Table 3-4. Primary aberrations and their standard deviations for a system with a
uniformly illuminated annular pupil.
Aberration F(r, q) sF
Spherical As r 4 12
(4 - 2
- 6 4 - 6 + 4 8 ) As 3 5
Coma Ac r3 cos q 12
(1 + 2
+ 4 + 6 ) Ac 2 2
Astigmatism Aa r2 cos 2 q 2 12
(1 + ) Aa 4
Field curvature (defocus) Ad r2 (1 - ) A

2
d 2 3
Distortion (tilt) 2 12
At r cos q
(1 + ) At 2
to determine Bd . Proceeding in this manner, we find that the optimum value is

2
( )
Bd = - 1 + 2 As . The corresponding standard deviation is 1 - 2 As 6 5 . ( )
Astigmatism and coma aberrations can be treated similarly. Table 3-5 lists the form
of a balanced primary aberration and its standard deviation. Also listed in the table is the
location of the diffraction focus, i.e., the point with respect to which the aberration
variance is minimum so that the Strehl ratio is maximum at it. We note that in the case of
coma, the balancing aberration is a wavefront tilt whose amount depends on . Thus,
maximum Strehl ratio is obtained at a point that is displaced from the Gaussian image
point but lies in the Gaussian image plane. In the case of astigmatism, the amount of
balancing defocus is independent of .
Figure 3-8 shows how the standard deviation of an aberration, for a given value of
the aberration coefficient Ai , varies with the obscuration ratio of the pupil. In Figures
3-8a and 3-8b, the amounts of defocus and tilt required to minimize the variance of
spherical aberration and coma, respectively, are also shown. We observe from these
figures that the standard deviation of spherical and balanced spherical aberrations and
defocus decreases as increases. Correspondingly, the tolerance in terms of their
aberration coefficients As and Bd , for a given Strehl ratio, increases. Thus, for example,
the depth of focus for a certain value of the Strehl ratio increases as increases. The
standard deviation of coma, astigmatism, balanced astigmatism, and tilt increases as
increases. The standard deviation of balanced coma first slightly increases, achieves its
maximum value at = 0.29 and then decreases rapidly as increases. The factor by
which the standard deviation of an aberration is reduced by balancing it with another
aberration is reduced in the case of spherical aberration, but increases in the case of coma
and astigmatism, as increases.
3.3.4 Comparison of Approximate and Exact Results5

Substituting the expressions for the various primary and balanced primary
aberrations into Eq. (3-48) and carrying out the integration, we obtain the exact
expressions for the Strehl ratio. These are listed in Table 3-6. The coefficient Ai of an
◊ ◊
aberration is in radians. The quantities C( ) and S( ) are Fresnel integrals given by
Table 3-5. Balanced primary aberrations, their standard deviation, and diffraction
focus
Aberration F(r, q; ) sF Diffraction Focus
Balanced
spherical [ ( )
As r 2 - 1 + 2 r 2 ] 1
6 5
1 - 2( )
2
As [0,0,8(1 + )F A ]
2 2
s
Balanced 2 1 + 2 + 4 4 12
coma
Ê
Ac Á r3 -
ˆ
r˜ cos q (1 - ) (1 + 4 + )
2 2
Ac Í
(
È 4 1 + 2 + 4 ) ˘
FAc , 0, 0 ˙
Ë 3 1 + 2 ¯
6 2 (1 + ) 2 12
Î (
Í 3 1+ 2
) ˙
˚
Balanced
astigmatism a
(
A r 2 cos 2 q - 1 2 ) 1
(1 + 2
+ 4
12
) Aa (0, 0, 4 F A )
2
a
2 6
0.30 0.12 1.2 0.12

Spherical Balanced
0.25 0.10 1.0 0.10
sf /Ac (coma), balancing tilt

coma
sf /Ac (balanced coma)

Balanced defocus
0.20 0.08 0.8 0.08
sf /As
(1 + 2) 2(1 + 2 + 4)/3(1 + 2)

0.15 0.06 0.6 0.06
0.10 0.04 0.4 0.04

Coma
Balanced
0.05 spherical 0.02 0.2 0.02
0.00 0.00 0.0 0.00

0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0

(a) (b)
0.40 0.30
0.25
Defocus
0.35
0.20
sf /Ad
sf /As
0.30 0.15
Astigmatism 0.10
0.25
Balanced 0.05
astigmatism
0.20 0.00
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0

(c) (d)
0.75
0.70
0.65 Tilt
sf /At
0.60
0.55
0.50
0.0 0.2 0.4 0.6 0.8 1.0

(e)
Figure 3-8. Variation of standard deviation of a primary and a balanced primary

aberration with obscuration ratio . Variation of balancing defocus in the case of
spherical aberration and tilt in the case of coma are also shown. (a) Spherical
aberration, (b) coma, (c) astigmatism, (d) defocus, and (e) tilt.
Table 3-6. Exact expressions for Strehl ratio for primary aberrations*.
Aberration F ( r, q ; ) S
4
Spherical As r [ p / 2 As (1 – 2 )2 ]
{
¥ [C( 2 As / p ) – C( 2 As / p 2 )]2 + [ S( 2 As / p ) – S( 2 As / p 2 )]2 }
Balanced spherical As [r4 – (1 + 2 )r2 ] {
[2 p / As (1 – 2 )2 ] C 2 [( As / 2 p)1/ 2 (1 – 2 )] + S 2 [( As / 2 p)1/ 2 (1 – 2 )]}
2
È 1 ˘
Coma Ac r3 cos q (1 – 2 ) –2 Í Ú J0 ( Ac x 3 / 2 ) dx ˙
Î 2 ˚
È 2 (1 + 2 + 4 ) ˘ Ï
2
Ac Ír3 – r˙ cos q Ô 1 È Ê 2 1 + 2 + 4 1/ 2 ˆ ˘ ¸Ô
3 (1 + 2 ) (1 – 2 ) –2 Ì J0 Í Ac Á x 3 / 2 –
Balanced coma ÍÎ ˙˚
ÔÓ
Ú
2
ÍÎ Ë 3 1 + 2
x ˜ ˙ dx ˝
¯ ˙˚ Ô
˛
Astigmatism Ad r2 cos2 q (1 – 2 ) –2 [ H 2 ( Aa / 2) + 2 H 2 ( 2 Aa / 2) – 2 2 H ( Aa / 2) H ( 2 Aa / 2)
¥ cos[(1 – 2 )( Aa / 2) – a( Aa / 2) + a( 2 Aa / 2)]}
2
ÏÔ • ¸Ô
Balanced astigmatsm Aa r2 (cos2 q – 1 / 2)
ÔÓ
2
k =0
Â
Ì[ 4 / Aa (1 – )] J2 k +1 ( Aa / 2) – J2 k +1 ( Aa / 2)˝
2
Ô˛
2
È sin[ Ad (1 – 2 ) / 2] ˘
Curvature of field (defocus) Ad r2 Í ˙
ÍÎ Ad (1 – ) / 2 ˙˚
2
2
È 2J (A ) 2 J ( At ) ˘
Distortion (tilt) At r cos q (1 – 2 ) – 2 Í 1 t – 2 1 ˙
ÍÎ At At ˙˚
* A i is the coefficient of aberration in radians, C (b ) = Ú0b cos(p x 2 / 2) dx, S (b ) = Ú0b sin(p x 2 / 2) dx,
H (b ) = [ J 02 (b ) + J12 (b )], a (b ) = tan –1[ J1 (b ) / J 0 (b )].
b
Û
C(b) = Ù cos p x 2 2 dx
ı
( ) (3-53a)
0
and
b
Û
(
S(b) = Ù sin p x 2 2 dx ,
ı
) (3-53b)
0
◊
respectively. The quantities H ( ) and a( ) are given by ◊
12
[
H (b) = J 02 (b) + J12 (b) ] (3-54a)
and
tan a(b) = J1 (b) J 0 (b) , (3-54b)
respectively. We note that in the case of coma the integration must be carried out
numerically. In the case of balanced astigmatism, only the first few terms of the infinite
series need to be considered for adequate precision.
Figures 3-9a and b show how the Strehl ratio of a primary aberration varies with its
standard deviation for = 0.5 and 0.75, respectively. Approximate as well as exact
results are shown in these figures. The curves for a given aberration and the
corresponding balanced aberration can be distinguished from each other by their behavior
for large s w values (near 0.25l). For example, coma is shown by the evenly dashed
curves; the higher dashed curve is for coma and the lower is for balanced coma. The same
holds true for astigmatism. The following observations may be made from Figures 3-9.
i. For small values of s w , the Strehl ratio is independent of the type of aberration. It
depends only on its variance.
ii. The expressions for S1 and S2 underestimate the true Strehl ratio.
iii. The expression for S3 overestimates the true Strehl ratio for >~ 0.5 . It gives the
>
Strehl ratio with an error of less than 10 percent for S ~ 0.4. For smaller obscurations,
the error is less than 10 percent for S > ~ 0.3. The percent error is defined to be
100(1 - S3 S ) .
iv. S3 gives a better approximation for the true Strehl ratio than S1 and S2 .
The reason is that, for small values of s F , it is larger than S1 by approximately sF4 4 .
4
Of course, S1 is larger than S2 by s F 4.
v. The Strehl ratio depends strongly on the standard deviation of an aberration but
weakly on its detailed distribution over a wide range of Strehl ratio values and not just for
large values of it.
Using S1 to estimate the Strehl ratio, Figure 3-10 shows how the aberration
coefficient Ai (in units of l ) of a primary aberration for 10 percent error varies with the
obscuration ratio. It is evident that this coefficient increases with obscuration in the case
of spherical, balanced spherical, and balanced coma, but decreases in the case of
astigmatism, balanced astigmatism, and coma. Thus, S1 estimates the true Strehl ratio
with a small error for a larger aberration as increases in the case of spherical, balanced
spherical, and balanced coma, but a smaller aberration in the case of astigmatism,
balanced astigmatism, and coma.
When the aberration coefficient Ai of an aberration is equal to a quarter wave, the

variation of the corresponding Strehl ratio with is shown in Figure 3-11. It is evident
that a Strehl ratio of 0.8 is obtained only for spherical aberration and = 0. Otherwise, a
smaller or a larger value is obtained depending on the type of the aberration. Comparing
this figure with Figures 3-9a and b, we again conclude, as in the case of circular pupils,
that it is advantageous to use the standard deviation of an aberration instead of the
aberration coefficient to estimate the Strehl ratio. For example, a Strehl ratio of 0.8 is
obtained for any aberration with a standard deviation of s w = l 14 . On the other hand,
this value of Strehl ratio is obtained for different values of the aberration coefficient for
different aberrations.
1.0
= 0.5
0.8
0.6
S
0.4 S3
0.2
S2
S1
0.0
0.00 0.05 0.10 0.15 0.20 0.25
sw
Figure 3-9a. Strehl ratio for annular pupils with = 0.5 as a function of the
standard deviation s w of an aberration in units of l . The Strehl ratio for a given
value of the standard deviation for classical coma is practically the same as that for
balanced coma. For large values of s w , the Strehl ratio for classical astigmatism is
larger than that for balanced astigmatism. Spherical...., Coma----, Astigmatism–.–.
1.0
= 0.75
0.8
0.6
S
S3
0.4
0.2
S2
S1
0.0
0.00 0.05 0.10 0.15 0.20 0.25
sw
Figure 3-9b. Strehl ratio for annular pupils with = 0.75 as a function of the
standard deviation s w in units of l . For large values of s w , the Strehl ratio for
balanced coma is higher than that for coma. The opposite is true for astigmatism.
Note that the curves for coma and astigmatism are practically identical.
Spherical...., Coma----, Astigmatism–.–.
1.0 5.0
Spherical, Balanced spherical, Balanced coma

0.8 4.0
Astigmatism, Balanced astigmatism, Coma

Balanced
spherical
Balanced
astigmatism
Ai 0.6 3.0
Astigmatism
0.4 2.0
Coma
Balanced
coma
0.2 1.0
Spherical
0.0 0.0
0.0 0.2 0.4 0.6 0.8 1.0

Figure 3-10. Variation of a primary aberration coefficient Ai (in units of l ) with

for 10 percent error when S1 is used to estimate the Strehl ratio.
1.0 0.75
BS
BC
0.9 0.65
BA
0.8 0.55
S
S
0.7 0.45
C
0.6 0.35
0.5 0.25
0.0 0.2 0.4 0.6 0.8 1.0

Figure 3-11. Strehl ratio for Ai = l 4 as a function of . S, Spherical; BS, balanced

spherical; C, coma; B C , balanced coma; A, astigmatism; and BA, balanced
astigmatism. The right-hand side vertical scale is only for coma.
3.4 Balanced Aberrations and Zernike Annular Polynomials 291
3.4 BALANCED ABERRATIONS AND ZERNIKE ANNULAR POLYNOMIALS6

The phase aberration function of a system with an annular pupil for a point object at
a certain angle from its axis can be expanded in terms of a complete set of Zernike
annular polynomials Z nm (r, q; ) that are orthonormal over a unit annulus in the form
• n
F(r, q; ) = Â Â c nm Z nm (r, q; ) , £ r £ 1 , (3-55a)
n =0 m =0
where cnm are the orthonormal expansion coefficients that depend on the field angle of
the object, n and m are positive integers, n - m ≥ 0 and even, and
[
Z nm (r, q; ) = 2( n + 1) (1 + d m 0 ) ]1/ 2Rnm (r; ) cos mq . (3-55b)
The polynomials Z nm (r, q; ) are orthonormal according to
1 1 2p
m¢
Ú Ú Z n (r, q; ) Z n ¢ (r, q; ) r dr d q = d nn ¢ d mm ¢
m
. (3-55c)
(
p 1- 2
)
The annular polynomials Z nm (r, q; ) are similar to the circle polynomials Z nm (r, q)
discussed in Section 2.4, except that they are orthogonal over an annular pupil. Thus, they
are also unique in that they are the only polynomials in two variables r and q, which
(a) are orthogonal over an annulus, (b) are invariant in form with respect to rotation of the
coordinate axes about the origin, and (c) include a polynomial for each permissible pair
of n and m values. They can be obtained from the corresponding circle polynomials by
the Gram-Schmidt orthogonalization process.7
The radial polynomials Rnm (r; ) are given by
È (n - m) 2 ˘
Rnm (r; ) = Nnm Í Rnm (r) - Â (n - 2i + 1) < Rnm (r) Rnm- 2i (r; ) > Rnm- 2i (r; )˙ , (3-56a)
Î i ≥1 ˚
where
2 1 m
< Rnm (r) Rnm¢ (r; ) > = Ú Rn (r) Rn ¢ (r; ) r dr
m
(3-56b)
1 - 2
and Nnm is a normalization constant such that the radial polynomials satisfy the
1 1 - 2
Ú Rn (r; ) Rn ¢ (r; ) r dr =
m m
d . (3-56c)
2(n+ 1) nn ¢
A radial polynomial Rnm (r; ) is of degree n in r containing terms in rn , rn -2 , K, and r m

with coefficients that depend on . it is even or odd in r depending on whether n (or m) is
even or odd.
For m = 0 , the radial polynomials are equal to the Legendre polynomials Pn (◊)
according to
R20n
È 2 r2 - 2
(r; ) = Pn Í - 1
˘
˙ .
( ) (3-57)
Í 1 - 2 ˙
Î ˚
Thus, they can be obtained from the circle radial polynomials R20n (r) by replacing r by
[(r 2
- 2 ) (1 - )] 2 12
; i.e.,
ÈÊ r2 - 2 ˆ 1 2˘
R20n (r; ) = R20n ÍÁ 2 ˜
˙ . (3-58)
ÍÎË 1 - ¯ ˙˚
It can be seen from Eqs. (3-56a) and (3-55b) that

12
Ê n 2i ˆ
Rnn (r; ) = r n
ÁÂ ˜
Ë i=0 ¯
(3-59a)
12
= r n 1 - 2 {( ) (1 - 2(n +1) )} . (3-59b)
Moreover,
Rnn - 2 (r; ) =
nrn - (n - 1) 1 - 2 n [( ) (1 - ( ) )] r
2 n -1 n-2
12 . (3-59c)
-1
Ï 1 - 2
Ì
Ó
( ) È 2
ÎÍ (
n 1- 2 ( n +1
)
) - (n - 1)(1 - ) (1 - ( ) )˘˚˙¸˝˛
2 2n 2 2 n -1
It is evident that the radial polynomial Rnn (r; ) differs from the corresponding circle
polynomial Rnn (r) only in its normalization. We also note that
Rnm (1; ) = 1, m = 0
π 1, m π 0 . (3-60)
The orthonormal Zernike expansion coefficients are given by
1 1 2p
Ú Ú F(r, q; )Z n (r, q; ) r dr d q ,
m
c nm = (3-61)
(
p 1- 2
) 0
as may be seen by substituting Eq. (3-55a) and using the orthonormality of the
polynomilas. The Zernike annular radial polynomials for n £ 8 are listed in Table 3-7.
The number of Zernike (or orthogonal) aberration terms in the expansion of an aberration
function through a certain order n is given by Eqs. (2-63), as in the case of a circular
pupil.
n (r, q; ) = c nm Z n (r, q; ) .
Fm m
(3-62)
OPTICAL SYSTEMS WITH ANNULAR PUPILS 293
Table 3-7. Zernike annular radial polynomials.
n m Rnm (r; )
0 0 1
1 1 (
2r 1 + 2 )1 2
2 0 (2r2 - 1 - 2 ) (1 - 2 )
12
2 2 r2 (1 + 2 + 4 )
3 (1 + 2 ) r3 - 2 (1 + 2 + 4 ) r
3 1 12
(1 - 2 ) [(1 + 2 ) (1 + 42 + 4 )]
12
3 3 r3 (1 + 2 + 4 + 6 )
4 0 [6r4 - 6 (1 + 2 ) r2 + 1 + 42 + 4 ] (1 - 2 )2
4r4 - 3 [(1 - 8 ) (1 - 6 )] r2
4 2 12
Ï 2 -1 È 8 2 6 ˘ ¸
Ì(1 - ) Í16 (1 - ) - 15 (1 - ) (1 - )˙ ˝
10
Ó Î ˚ ˛
12
4 4 (
r 4 1 + 2 + 4 + 6 + 8 )
( ) ( ) (
10 1 + 4 2 + 4 r5 - 12 1 + 4 2 + 4 4 + 6 r3 + 3 1 + 4 2 + 10 4 + 4 6 + 8 r )
5 1 12
(1 - ) [(1 + 4 + ) (1 + 9 + 9
2 2 2 4 2 4
+ 6 )]
5 r - 4 [(1 - ) (1 - )] r
5 10 8 3
5 3 12
Ï1- 2 -1 10 2
Ì(
Ó
) ÈÎÍ25 (1 - ) - 24 (1 - ) (1 - )˘˚˙¸˝˛
12 8
12
5 5 (
r5 1 + 2 + 4 + 6 + 8 + 10 )
6 0 [20 r 6
( ) (
- 30 1 + 2 r 4 + 12 1 + 32 + 4 r2 - 1 + 9 2 + 9 4 + 6 ) ( )] (1 - )
2 3
( ) (
15 1 + 4 2 + 10 4 + 4 6 + 8 r6 - 20 1 + 4 2 + 10 4 + 10 6 + 4 8 + 10 r 4 )
6 2
( + 6 1 + 4 + 10 + 20 + 10 + 4 +
2 4 6 8 10 12
)r 2
12
(1 + ) [(1 + 4
2 2
)(
2
+ 10 4 + 4 6 + 8 1 + 9 2 + 454 + 656 + 458 + 9 10 + 12 )]
6r - 5 [(1 - ) (1 - )] r
6 12 10 4
6 4 12
Ï1- 2 -1 12 2
Ì( ) ( ) (
È36 1 - - 35 1 -
) (1 - )˘ ¸˝
14 10
Ó Í
Î ˚˙ ˛
12
6 6 (
r6 1 + 2 + 4 + 6 + 8 + 10 + 12 )
Table 3-7. Zernike annular radial polynomials.
n m Rnm (r; )
7 1 a17 r7 + b71 r5 + c17 r3 + d71 r
7 3 a73 r7 + b73 r5 + c73 r3
7 5
7r7 - 6 1 - 14 [( ) (1 - )] r 12 5
12
Ï 1 - 2 -1 È
Ì
Ó
( ) ÍÎ (
49 1 - 16 ) - 48 (1 - ) (1 - )˘˙˚¸˝˛
14 2 12
12
7 7 (
r7 1 + 2 + 4 + 6 + 8 + 10 + 12 + 14 )
8 0
( ) ( )
70 r8 - 140 1 + 2 r6 + 30 3 + 82 + 34 r 4 - 20 1 + 6 2 + 6 4 + 6 r2 + e80 ( )
2 4
(1 - )
(
a17 = 35 1 + 9 2 + 9 4 + 6 ) A71
(
b71 = - 60 1 + 9 2 + 154 + 9 6 + 8 ) A71
(
c17 = 30 1 + 9 2 + 254 + 256 + 9 8 + 10 ) A71
(
d71 = - 4 1 + 9 2 + 454 + 656 + 458 + 9 10 + 12 ) A71
3 12 12
(
A71 = 1 - 2 ) (1 + 9 2
+ 9 4 + 6 ) (1 + 16 2
+ 36 4 + 16 6 + 8 )
(
a73 = 21 1 + 4 2 + 10 4 + 20 6 + 10 8 + 4 10 + 12 ) A73
(
b73 = - 30 1 + 4 2 + 10 4 + 20 6 + 20 8 + 10 10 + 4 12 + 14 ) A73
(
c73 = 10 1 + 4 2 + 10 4 + 20 6 + 358 + 20 10 + 10 12 + 4 14 + 16 ) A73
2 12 1 2
( ) (1 + 4 + 10 + 20 + 10 + 4 + )
A73 = 1 - 2 2 4 6 8 10
12
¥ (1 + 9 + 45 + 165 + 270 + 270 + 165 + 45
2 4 6 8 10 12 14
+ 9 16 + 18 )
e80 = 1 + 16 2 + 36 4 + 16 6 + 8
Unless n = m = 0 , its mean value is zero, i.e.,
1 2p 1 2p
Û Û Û Û
F nm (r, q) = Ù Ù F nm (r, q) r dr d q Ù Ù r dr d q
ı ı ı ı
0 0
= 0 , n π 0, m π 0 . (3-63a)
For m = 0 , this may be seen with the help of Eq. (3-55b) and the fact that R00 (r; ) = 1 is
a member of the polynomial set. The orthonormality Eq. (3-55c) yields the result that the
mean value of Rn0 (r; ) is zero. When m π 0 , the average value of cos mq is zero.
1 2p 1 2p
2 Û Û 2 Û Û
[ F nm (r, q; )] = Ù Ù F nm (r, q; )
ı ı
[ ] r dr d q Ù Ù r dr d q
ı ı
0 0
2
= cnm . (3-63b)
2
2
s nm = (F mn ) 2 - Fm
n
2
= cnm , n π 0, m π 0 . (3-64)
Thus, each expansion coefficient, with the exception of c00 , represents the standard
deviation of the corresponding aberration term. The variance of the aberration function is
accordingly given by
2 • n
s 2F = F 2 (r, q; ) - F(r, q; ) = Â 2
Â c nm . (3-65)
n =1 m = 0
Just as in the case of circular pupils, we identified balanced primary aberrations with
low-order Zernike circle polynomials (see Section 2.4), we now identify balanced
primary aberrations of annular pupils with Zernike annular polynomials. A balanced
aberration represents an aberration of a certain order in the power series expansion of the
aberration function in pupil coordinates imixed with aberrations of lower order such that
the variance of the net aberration is minimized.
The balanced primary aberrations discussed in Section 3.3.3 can be easily identified
with the corresponding Zernike annular polynomials. For example, for n = 4 and m = 0 ,
Eq. (3-62) becomes
F 04 (r, q; ) = 5 c40 R40 (r; )
= [ ( )
5 c40 6 r 4 - 6 1 + 2 r2 + 1 + 4 2 + 4 ( )] (1 - )
2 2
. (3-66a)
Comparing this with the balanced spherical aberration given in Table 3-5, namely,
[ (
F bs (r) = As r 4 - 1 + 2 r2 ) ] , (3-66b)
we note the following. The aberration F04 contains a constant (independent of r and q)
term. This term does not change the standard deviation of the balanced aberration or the
Strehl ratio corresponding to it. In Eq. (3-66a), as in Eq.(3-66b), the spherical aberration
is balanced with an amount of defocus that is - 1 + 2 times the amount of the ( )
aberration. Comparing the coefficients of the r 4 term, we find immediately that the
standard deviation of the balanced spherical aberration is given by
s bs = c40
2
(
= As 1 - 2 ) 6 5 , (3-66c)
in agreement with the result given in Table 3-5.
When n = 3 and m = 1, Eq. (3-62) becomes
F13 (r, q; ) = 2 2 c31 R31 (r; ) cos q
=
[( )
2 2 c 31 3 1 + 2 r3 - 2 1 + 2 + 2 r cos q ( )]
12 . (3-67a)
(1 - ) [(1 + ) (1 + 4
2 2 2
+ 4 )]
We note that this polynomial represents balanced coma
Ê 2 1 + 2 + 4 ˆ
F bc (r, q; ) = Ac Á r3 - r˜ cos q , (3-67b)
Ë 3 1 + 2 ¯
s bc = c31
12 12
( )(
= Ac 1 - 2 1 + 4 2 + 4 ) (
6 2 1 + 2 ) . (3-67c)
F 22 (r, q; ) = 6 c22 R22 (r; ) cos 2q
12
(
= 2 6 c 22 r2 cos 2 q - 1 2 ) (1 + 2
+ 4 ) . (3-68a)
This polynomial represents balanced astigmatism
(
Fba (r, q; ) = Aa r2 cos 2 q - 1 2 ) , (3-68b)
sba = c22
12
(
= Aa 1 + 2 + 4 ) 2 6 . (3-68c)
F 20 (r, q; ) = 3 c20 R20 (r; )
= (
3 c20 2r2 - 1 - 2 ) (1 - )
2
, (3-69a)
which represents the defocus aberration (except for a constant term)
F d (r; ) = Bd r2 (3-69b)
sd = c20
(
= Bd 1 - 2 ) 2 3 . (3-69c)
For n = 1 and m = 1, Eq. (3-62) becomes
F11 (r, q; ) = 2c11 R11 (r; ) cos q
12
= 2c11r cos q 1 + 2 ( ) , (3-70a)
which represents a wavefront tilt
F t (r, q; ) = Bt r cos q (3-70b)
s t = c11
12
(
= Bt 1 + 2 ) 2 . (3-70c)
Finally, for n = m = 0 , Eq. (3-62) becomes
F 00 (r, q; ) = c00 , (3-71)
which represents a uniform (piston) aberration. Obviously, it has no effect on the standard
deviation or the Strehl ratio of a system with a single exit pupil. In a multiexit pupil
system, any relative piston errors among the subpupils will cause some destructive
interference.
Thus, we see that Zernike polynomials can be identified with balanced aberrations;
that, in fact, is their advantage. Here we have discussed only the primary aberrations. In
general, the aberration function of an optical system may consist of higher-order
aberrations. Moreover, in a system without an axis of rotational symmetry, the aberration
function will consist of terms not only in cos mq but in sin mq as well.
3.5 DEFOCUSED SYSTEM8

In this section, we consider a defocused system and show that, as in the case of
systems with circular pupils, the axial irradiance is symmetric about the geometrical
focus only if its Fresnel number is large. A far-field distance is defined beyond which a
focused beam behaves practically like a collimated beam.

We now consider the irradiance distribution in a defocused plane at a distance z from
the pupil plane. Following the discussion on defocused systems with circular pupils, z
values satisfying the condition of Eq. (2-77) imply neglecting a spherical aberration of
l/8, which corresponds to a Strehl ratio of more than 0.946 when > 0. When this
aberration is balanced with an appropriate amount of defocus, the Strehl ratio increases to
more than 0.996. Assuming that Eq. (2-77) is satisfied, the defocused irradiance
distribution for an otherwise aberration-free system may be written
1 2
2 Û
[ (
I (r; z; ) = 2 R z 1 - 2 )] ı
( 2
)
Ù exp i Bd r J 0 ( p r rR z ) r dr , (3-72)

where the irradiance is in units of Pex Sex ( ) l2 R 2 and r is in units of l F ,
R
Bd = N Ê - 1ˆ (3-73)
Ëz ¯
is the peak defocus phase aberration (relative to a zero aberration at the center of the
annular pupil), and
N = a2 l R (3-74)
is the Fresnel number of the exit pupil as observed from the focal point if there were no
obscuration. The corresponding Fresnel number of the annular pupil is given by
(
N = N 1 - 2 ) . (3-75)
If we let z = R , Eq. (3-72) reduces to Eq. (3-13), except for the difference due to
different normalizations in the two equations; the irradiance is normalized to values of 1
and 1 - 2 at the Gaussian image point in these equations, respectively. As in the case of
systems with circular pupils, the irradiance distribution is asymmetric about the Gaussian
image plane for the same three reasons, unless N is very large so that z ~ R .
3.5.2 Focused Beam

Consider an optical system focusing a beam of power Pex distributed uniformly
across its annular exit pupil. If the beam is focused at a distance R, its irradiance
distribution at a distance z from the exit pupil will be given by Eq. (3-72). If we let r = 0
in this equation, we obtain the axial irradiance of the beam
2
R
I (0; z; e) = Ê ˆ S , (3-76a)
Ë z¯
or
Pex Sex ( )
Ii (0; z; ) = S , (3-76b)
l2 z 2
where
2
{ [ (
S = sin Bd 1 - 2 2 ) ] [ B (1 - ) 2]}
d
2
. (3-77)
Equation (3-77) differs from the corresponding Eq. (2-84c) for systems with circular
( )
pupils in that the quantity Bd in the latter has been replaced by Bd 1 - 2 . It represents
the peak defocus phase aberration at the outer edge of the annular pupil relative to its
value at the inner edge. Also, the area Sex () given by Eq. (3-2) and as a result the focal-
point irradiance given by Eq. (3-3) are smaller for an annular beam by a factor of 1 - 2
compared to a circular beam.
( )
The axial irradiance is minimum and equal to zero when Bd 1 - 2 is an integral
multiple of 2p. The corresponding z values are given by
( )
R z = 1 + 2 n N 1 - 2 , n = ± 1 , ± 2 , K . (3-78)
When N is very large (>>10), the axial irradiance is zero when z is different from R by an
( )
integral multiple of ± 8l F 2 1 - 2 . Accordingly, the depth of focus of a system with an
( )
annular pupil is larger by a factor of 1 1 - 2 , compared to a corresponding system with
a circular pupil. The maxima of axial irradiance, obtained by equating the derivative of
Eq. (3-76b) with respect to z equal to zero, are given by the solutions of
[ ( ) ] (
tan Bd 1 - 2 2 = ( R z ) Bd 1 - 2 2 , z π R . ) (3-79)
Figure 3-12a shows how the axial irradiance of an annular beam with = 0.5 varies
for N = 1, 10, and 100. Comparing it with Figure 2-11, we note that the effect of the
obscuration is to reduce the irradiance at the principal maximum, but to increase it at the
secondary maxima. Also, the maxima and minima occur at smaller z values for an annular
aperture. As in the case of circular beams, the axial irradiance of annular beams also
becomes symmetric about the focal point z = R as N increases.
2.0 2.0 2.0
N=1 N = 10 N = 100
1.5 = 0.5 1.5 = 0.5 1.5 = 0.5
l (0; z)
lu
1.0 1.0 1.0
lg
0.5 0.5 0.5
0.0 0.0 0.0

0.0 0.5 1.0 1.5 0.5 1.0 1.5 0.8 0.9 1.0 1.1 1.2
z/R z/R z/R
Figure 3-12a. Axial irradiance of an annular beam with = 0.5 focused at a distance
R. The minima of irradiance occur at z R = 3 11, 3/19, 3/27, when N = 1. The
irradiance is in units of the focal-point irradiance of a corresponding circular beam
with the same total power. Accordingly, the focal-point irradiance in this figure is
1 - 2 = 0.75 . The axial irradiance becomes symmetric about the focal point as N
increases. The dashed curves are for a Gaussian beam with g = 1, as discussed in
Chapter 4.
1.0 1.0 1.0

Nz = 1 Nz = 10 Nz = 100
0.8 lu = 0.5 0.8 = 0.5 0.8 = 0.5
lz (0; R)
lg
0.6 0.6 0.6
0.4 0.4 0.4
0.2 0.2 0.2
0.0 0.0 0.0

0.5 1.0 1.5 2.0 0.5 1.0 1.5 2.0 0.8 0.9 1.0 1.1 1.2
R/z R/z R/z
Figure 3-12b. Central irradiance (in units of Pex Sex l 2 z 2 ) at a distance z from the
plane of the exit pupil when a beam is focused at various distances R. The quantity
N z = a 2 l z represents the Fresnel number of the exit pupil as observed from the
target. The dashed curves are for a Gaussian beam.
Although the principal maximum of axial irradiance does not lie at the geometrical
focus, the maximum central irradiance on a target at a fixed distance z is obtained when
the beam is focused on it, i.e., when R = z , as may be seen by differentiating Eq. (3-76b)
with respect to R and equating the result to zero. It is evident that the central irradiance on
a target when the beam is focused on it is Pex Sex ( ) l2 z 2 . Figure 3-12b illustrates how
the central irradiance on a target at a fixed distance z varies when the beam is focused at
various distances R along its axis. The irradiance in this figure is in units of
Pex Sex ( ) l2 z 2 . The quantity N z = a 2 l z in this figure represents the Fresnel number of
a circular exit pupil as observed from the target. We note that as N z increases, the
curves become symmetric about the point R = z .
To calculate the irradiance distribution in a defocused plane we write the right-hand

side of Eq. (3-72) as a product of two integrals and retain only its real part, since
irradiance is a real quantity, i.e.,
1 1
2 Û Û
[ (
I (r; z; ) = 2 R z 1 - 2
)] ı ı
2
Ù Ù cos Bd r - s[ (
2
)] J (prr) J (prs) r s dr ds
0 0 . (3-80)

If we let r = 0 and note that J 0 (0) = 1 , we obtain a different form of the expression for
axial irradiance, namely,
1 1
2ÛÛ
[ (
I (0; z; ) = 2 R z 1 - 2
)] 2 2
[ (
Ù Ù cos Bd r - s r s dr ds .
ı ı
)] (3-81)

The encircled power (in units of Pex with rc in units of lz D) is given by

1 1
Û Û
2
ı ı
[ ( )]
P (rc ; z; ) = 2 p Ù Ù cos Bd r2 - s 2 Q (r, s; rc ) r s dr ds , (3-82)

where Q (r, s; rc ) is given by Eqs. (2-108). The integrals in Eqs. (3-80) through (3-82)
may be evaluated by the Gauss quadrature method, according to which, for a function
¶(r, s)
1 1
M M i -1
Û Û
ı ı
[
Ù Ù f (r, s) d r ds = (1 - ) 2 ]2 Â w i2 f (ri , si ) + 2 Â Â ( )
w i w j f ri , s j , (3-83)
i =1 i=2 j =1

where, as in Eq. (2-118), M is the number of points of a 1-D quadrature, w i are the
weight factors, and
ri = si
[
= 1 + + (1 - ) xi ] 2 , (3-84)
xi being the i-th zero of the M th-order Legendre polynomial. Note that by letting Bd = 0
in Eqs. (3-80) through (3-82), we can calculate the focal-plane distributions as well.
Equations (3-80) through (3-83) are generalizations of Eqs (2-104) through (2-107) for
circular beams to annular beams with F (r) = Bd r2 . Note that with slight modification,
Eqs. (3-80) through (3-82) can be applied to diffraction calculations involving any
radially symmetric amplitude and phase distributions at the exit pupil. For example, if
spherical aberration As r 4 were present, the cosine factor in these equations would
[ ]
become cos F (r) - F (s) , where F (r) = Bd r2 + As r 4 .
An example of a defocused distribution is illustrated in Figure 3-13 for an annular

beam with = 0.5 and a large Fresnel number (so that the inverse-square-law variation is
negligible). The amount of defocus Bd = 2.783 rad (or 0.433l) is such that the central
irradiance for a uniform circular beam is reduced to half of the corresponding focal-point-
irradiance. (The defocused distributions shown can also be interpreted as the distributions
on a target at a fixed distance z when the beam is focused at a distance R such that
Bd = 2.783 rad . In this case the irradiance would be in units of Pex Sex/l2 z 2 of r and rc
would be in units of lz D .) Note that the defocus aberration does not reduce the central
irradiance for the annular beam as much as it does for the circular beam, so much so that,
for the amount of defocus aberration considered in Figure 3-12, the defocused central
irradiance for the annular beam is higher than that for the corresponding circular beam.
Whereas for a circular beam, the central irradiance decreases from 1 to 0.500, for the
annular beam, it decreases from 0.750 to 0.514. This indicates the well-known fact that
the tolerance for a radially symmetric aberration, such as defocus, is higher for an annular
beam than for a circular beam.
0.6 1.2
0.514 = 0.5
0.5 0.497 1.0
0.4 0.8
P(rc)
l(r)
0.3 0.6
Pg
0.2 0.4
Pu
0.1 0.2
lg
lu
0.0 0.0
0.0 0.5 1.0 1.5 2.0 2.5
r; rc
Figure 3-13. Defocused irradiance and encircled-power distributions for an annular

beam with = 0.5 . The amount of defocus aberration Bd = 2.783 rad (or 0.433 l ) is
such that it gives a central irradiance of 0.5 for a circular beam of large Fresnel
number N. The dashed curve is for a corresponding Gaussian annular beam, which
is discussed in Chapter 4.
3.5.3 Collimated Beam

letting R Æ • . Thus, the defocus coefficient becomes
Bd = p a 2 l z . (3-85)
It represents the peak phase aberration of a plane wavefront with respect to a reference
sphere centered at a distance z from the exit pupil. Equation (3-76b) for the axial
irradiance reduces to
[ ( )
I (0; z; ) = 4 I0 sin 2 p a 2 1 - 2 2 lz ] , (3-86)
where
I0 = Pex Sex ( ) (3-87)
is the irradiance at the exit pupil. The axial irradiance is maximum and equal to 4 I0 at z
values given by
(
z = a 2 1 - 2 ) l(2 n + 1) , n = 0, 1, 2, K . (3-88)
It is minimum and equal to zero at z values given by
(
z = a 2 1 - 2 ) 2 l(n + 1) , n = 0, 1, 2, K (3-89)
These z values for the location of maxima and minima correspond to those axial positions
at which the annular exit pupil subtends an odd or even number of Fresnel’s half-wave
zones, respectively.
( )
For z > a 2 1 - 2 l, it decreases monotonically to zero. For
(
z ≥ D2 1 - 2 l , ) (3-90)
( )
corresponding to a defocus aberration of £ l 8 1 - 2 , it decreases approximately as
z -2 . For z satisfying Eq. (3-90), a collimated beam gives an axial irradiance at a distance
z that is ≥ 0.95 times the irradiance at this point if the beam were focused at it, i.e.,
S ≥ 0.95. This is illustrated in Figure 3-14, where the axial irradiance of a collimated
beam is plotted as a function of z. The distance z in this figure is in units of
( )
D2 1 - 2 l , which may be called the far-field distance of the annular exit pupil. The
axial irradiance is normalized by the exit pupil irradiance Pex Sex (0) for a circular beam
of the same power. The pupil irradiance and, therefore, the maxima of axial irradiance are
-1
(
higher for an annular beam compared to a circular beam by a factor of 1 - 2 . It is)
evident that a collimated beam yields practically the same irradiance on a target lying in
the far field of the annulaar exit pupil as a beam focused on it.
N=0
5
= 0.5
lg
4
l(0;z)
lu
2
0
0.0 0.1 0.2 0.3 0.4 0.5
z
Figure 3-14. Axial irradiance of a collimated annular beam normalized by the exit
pupil irradiance Pex Sex (0) for a circular beam of the same power. The pupil
irradiance and, therefore, the maxima of axial irradiance are higher for an annular
-1
beam compared to a circular beam by a factor of 1 - 2 . The axial distance z is ( )
(
in units of the far-field distance D 2 1 - 2 l . The dashed curve is for a )
corresponding Gaussian annular beam, which is discussed in Chapter 4 .
The irradiance distribution in a plane at a distance z can be obtained from Eq. (3-72)
by letting R Æ • and noting that the units of irradiance in this equation are
Pex Sex ( ) l2 R 2 and those of r are l F = l R D . Thus, for a collimated annular beam, we
may write
1 2
4 Û
I (r; z; ) =
2 2
( 2
)
Ù exp i Bd r J 0 ( p rr) r dr , (3-91)
(1 - ) ı

where the units of irradiance are Pex Sex ( ) l2 z 2 and those of r are lz D. For
( )
z ≥ D2 1 - 2 l , since the effect of the defocus aberration is negligibly small, Eq. (3-
91) reduces to the aberration-free result, Eq. (3-14)
2
1 È 2 J1 ( p r ) 2 J1 ( pr ) ˘
I (r; z; ) ~ Í - 2 ˙ . (3-92)
(1 - )2 2
Î pr pr ˚
Equation (3-90) is the far-field condition for an annular exit pupil and Eq. (3-92)
represents its far-field or Fraunhofer diffraction pattern. Except for the units of irradiance
and z, Eq. (3-92) is the same as Eq. (3-14). Hence, the discussion of Section 3.2 on
encircled power, is applicable to this equation.
3.6 SYMMETRY PROPERTIES OF AN ABERRATED PSF9

Now we consider the symmetry properties of a PSF aberrated by a Zernike annular
polynomial. These properties can be obtained in a manner similar to those obtained in
Section 2.7 for circular pupils. Let the Zernike aberration be given by
F nm (r, q) = Anm Rnm (r; ) cos m q , (3-93a)
where Anm is the aberration coefficient. We will refer to the plane in which the center of
the reference sphere of radius of curvature R lies as the Gaussian image plane. If the
image is observed in another plane that lies at a distance z from the exit pupil, then the
aberration becomes
F(r, q) = Anm Rnm (r; ) cos m q + Bd r2 , (3-93b)
where Bd is the peak value of defocus aberration. It is given by
p a2 Ê 1 1 ˆ
Bd = - (3-94a)
l Ë z R¯
R
= p N Ê - 1ˆ . (3-94b)
Ëz ¯
As in Eq. (2-81c), N = a 2 l R is the Fresnel number of the pupil without obscuration as

observed from the Gaussian image point (or the focal point of a focused beam).
The aberrated irradiance distribution in a plane at a distance z normalized by the

central value Pex Sex ( ) l2 z 2 is given by
2
1 2p
I ( r, q i ; z ) =
( R z )2 Û Û Ï ÈF r, q - p R rr cos q - q ˘ ¸ r dr dq
2 Ù Ù expÌi ÍÎ ( ) ( i )˙ ˝ . (3-95)
[p (1 - )]
2 ı ı
0
Ó z ˚˛
Proceeding exactly as in Section 2.7, we can show that:
i. When N is small, the irradiance distribution is asymmetric about the Gaussian image
plane whether or not the system is aberrated. The distribution in any plane normal to the z
axis is m-fold symmetric. Moreover, the tangential plane and all planes containing the z
axis and making angles of p j m with the tangential plane, where j = 1, 2, ..., m, are
planes of symmetry. When m = 0 , the irradiance distribution in any observation plane is
radially symmetric. The aberration function given by Eq. (3-93) also possesses these
symmetry properties.
ii. For large values of N, we may let z ~ R in which case Eq. (3-95) reduces to
2
1 2p
1 Û Û
I ( r, q i ; z ) = 2 {[
Ù Ù exp i F(r, q) - p rr cos(q - q i ) ] } r dr d q , (3-96)
[(
p 1- 2
)] ı ı
0
where
pa2
Bd = ( R - z) . (3-97)
l R2
We note from Eq. (3-97) that the magnitude of Bd for z = R + D is the same as for
z = R - D , but its sign does change. Thus, for large Fresnel numbers, two planes located
symmetrically about the Gaussian image plane correspond to Bd values that are equal in
magnitude but opposite in sign.
The irradiance distribution is not symmetric about the Gaussian image plane when
m = 0 , i.e., for spherical aberration. However, the irradiance distributions in two planes
located symmetrically about the Gaussian image plane are identical if they are for
spherical aberrations of equal magnitude but opposite signs. The two aberration functions
in this case are different from each other only in their signs.
When m is odd, as for coma, the irradiance distribution is symmetric about the
Gaussian image plane, even though the corresponding aberration functions are not equal
to each other. When m is even, as for astigmatism, the distribution in the Gaussian image
plane is 2m-fold symmetric although the aberration function is only m-fold symmetric.
The irradiance distributions in two planes located symmetrically about the Gaussian
image plane differ from each other by a rotation of p m about the z axis. The
corresponding aberration functions have equal magnitude but opposite signs.
iii. The axial irradiance for large Fresnel numbers [obtained by letting r = 0 in Eq. (3-
96)] is symmetric about the Gaussian image plane. This is evident from the symmetry
properties of the distribution when m is not equal to zero. When m is equal to zero,
although the distribution is not symmetric about the Gaussian image plane, the axial
irradiance is. To see this, we let r = 0 in Eq. (3-96), which for spherical aberration may
be written
1 2
4 Û
I (0; z ) =
2 2
[ ] ( )
Ù exp i An 0 Rn (r; ) exp iBd r r dr
0 2
. (3-98)
(1 - ) ı

Now, since
R20n
È 2 r2 - 2
( ) = Pn Í
r;
( ˘
- 1˙ ,
) (3-99)
Í 1- 2
˙
Î ˚
◊
where Pn ( ) is a Legendre polynomial of degree n, we let
[(
x = 2 r2 - 2 ) (1 - )] - 1
2
(3-100)
in Eq. (3-98) and obtain

1 2
1 Û
I (0; z ) =
4 Ù
ı
[ ( ) ] [
exp i Bd x 1 - 2 2 exp i An 0 Pn /2 ( x ) dx ] . (3-101)
-1
For spherical aberration of any order, n 2 is even, and therefore, Pn / 2 ( - x ) = Pn / 2 ( x ) .

Accordingly, if we change the sign of Bd and change x to -x , the integral in Eq. (3-101)
does not change. Hence, the axial irradiance is symmetric about the Gaussian image
plane.
As an example, we consider the axial irradiance for primary spherical aberration. We

let
F(r) = As r 4 + Bd r2 . (3-102)
Substituting Eq. (3-102) into (3-98), we obtain the Strehl ratio
) {[ }
2 2 2
ÎÍ
(
S = Èp 2 As 1 - 2 ˘ C(b+ ) + C(b- )
˙˚ ] + [S(b+ ) + S(b- )] , (3-103)
where
b± = [(1 - ) A ± d] (2p A )
2
s s
12
, (3-104)
and
(
d = Bd + 1 + 2 As ) . (3-105)
It is evident that the Strehl ratio as a function Bd represents the axial irradiance. Hence,
the axial irradiance of a system aberrated by spherical aberration As is independent of the
(
sign of d. Now d = 0 corresponds to a defocus value Bd = - 1 + 2 As , which, in turn, )
represents its optimum value giving minimum variance of the aberration given by Eq. (3-
102). Indeed, for this defocus value, if we add an appropriate value of piston aberration,
the aberration of Eq. (3-102) takes the form of Zernike annular polynomial R40 (r;) .
Hence, the axial irradiance for primary spherical aberration is symmetric about the axial
point with respect to which the aberration variance is minimum, or it is symmetric about
the Gaussian image plane when the aberration is represented by the Zernike annular
polynomial R40 (r;) . It should be noted that the point of axial symmetry in the case of
astigmatism Aa r2 cos 2 q corresponds to Bd = - Aa 2, which is independent of the value
of . Similarly, in the case of coma Ac r3 cos q , the axial irradiance is symmetric about
the Gaussian image plane.
iv. A change in the sign of the aberration coefficient Anm when m is even has no effect
on the irradiance distribution in the Gaussian image plane. Thus, the sign of such an
aberration cannot be determined from this PSF. However, the irradiance distribution in a
defocused image plane does change when the sign of the aberration coefficient is
changed. For even values of m, the irradiance distributions in two symmetrically

defocused planes are identical (when N is large), provided the aberrated coefficient for
one has a sign that is opposite to that for the other. Similarly, for odd values of m, the
distributions for aberration coefficients with opposite signs are different from each other
by a rotation of p.
The symmetry properties of the PSFs aberrated by a primary aberration are

summarized in Table 3-8.
3.7 PSFs AND AXIAL IRRADIANCE FOR PRIMARY ABERRATIONS

It should be noted that the Zernike aberrations (e.g., spherical, astigmatism)
considered in Section 3.6 contain a defocus term. Hence, the term Gaussian image plane
used there would be different from the one used now in the discussion of a classical
primary aberration. The PSFs for a system with = 0.5 and aberrated by one wave of
spherical aberration are shown in Figure 3-15 for various defocused image planes. The
planes Bd = 0 and Bd = - (1 + 2 ) As represent the Gaussian and minimum-aberration-
variance planes, respectively. We note that only the central irradiances in two planes
located symmetrically about the minimum-aberration variance plane are equal to each
other, illustrating the symmetry of axial irradiance about this plane. The PSFs are
otherwise different from each other. Figure 3-16 shows the axial irradiance for various
values of As . Symmetry about the defocused point given by Bd = - (1 + 2 ) As is evident
from the figure.
The axial irradiance for astigmatism is shown in Figure 3-17. It is symmetric about
the point Bd = - Aa 2 , as indicated by an arrow on each curve. Figure 3-18 shows the
axial irradiance for one wave of coma. The irradiance for coma without any tilt is along
the z axis. For coma optimally balanced with tilt, the axial irradiance is along an axis that
is parallel to the z axis but passing through the diffraction focus (1.4, 0). The PSFs for
coma are discussed in Section 3.9.
Table 3-8. Symmetry properties of a PSF aberrated by a primary aberration.
Symmetry of Symmetry in Symmetry in

General Symmetry
Axial Irradiance* Defocused Images* Coefficient Sign
None Rotational about z axis About Bd = 0 About Bd = 0 Not applicable

Spherical Rotational about z axis About Bd = – (1 + 2 ) As About Bd = – (1 + 2 ) As In Bd = 0 plane
As r 4 Radial in any z plane if As Æ – As
Coma About tangential plane About Bd = 0 About Bd = 0 If rotated by p about
Ac r 3cos q Line symmetry in any z z axis
plane about x axis
Astigmatism Line symmetry about x About Bd = – Aa / 2 About Bd = – Aa / 2 In any plane
Aa r2 cos 2 q and y axes in Bd = 0 if rotated by p / 2
plane, 4-fold in or if Aa Æ – Aa
Bd = – Aa / 2 plane
*Only for large values of N.

3.7 PSFs and Axial Irradiance for Primary Aberrations 309
1.0 0.20
Aberration free W(r) = Asr4
0.8 Bd = – 1(1+2) As = 1 0.16
= 0.5
0.6 0.12
I(r; Bd)
– 0.5(1+2)
0.4 0.08
– 1.5(1+2)
0 (5¥)
0.2 – 2(1+2) (5¥)) 0.04
0.0 0.00
0 1 2 3 4
r
Figure 3-15. PSFs for = 0.5 and one wave of spherical aberration in various
defocused image planes. The right-hand side vertical scale is for PSFs for Bd = 0
( )
and - 2 1 + 2 l , since they have been multiplied by 5.
1.0
1 W(r) = Asr4
= 0.5
0.8 2
As = 0
0.6
I(0; Bd)
0.4
0.2
0.0
–6 –5 –4 –3 –2 –1 0 1 2 3
Bd
Figure 3-16. Axial irradiance for = 0.5 and various values of spherical aberration.
It is symmetric about the point Bd = - As 1 + 2 . ( )
1.0
W(r,q) = Aar2cos2q Aa = 0
0.8 = 0.5
0.6
I(0; Bd)
1
0.4
0.2
2
3
0.0
–6 –5 –4 –3 –2 –1 0 1 2 3
Bd
Figure 3-17. Axial irradiance for = 0.5 and various values of astigmatism. It is
symmetric about a point where an arrow is indicated on a curve and corresponds to
Bd = - Aa 2 .
0.6
W(r,q) = Acr3cosq
0.5 Ac = 1
= 0.5
0.4
Through
(1.4, 0)
I(Bd)
0.3
0.2
On axis
0.1
(0, 0)
0.0
–4 –3 –2 –1 0 1 2 3 4
Bd
Figure 3-18. Axial irradiance for = 0.5 with one wave of coma. It is symmetric
about the Gaussian image plane Bd = 0.
3.8 2-D PSFs 311
3.8 2-D PSFs

The PSFs for spherical aberration shown in Figure 3-15 or for coma shown in
Section 3.9 are useful for quantitative assessment. However, as stated in Section 2.8.5
regarding the PSFs for circular pupils, they do not lend themselves easily to what they
may look like when observed in practice. This is especially true when the PSF is not
radially symmetric. Accordingly, we give computer-generated 2-D pictures of the PSFs in
this section for various values of a primary aberration. The emphasis of these pictures is
on the structure of a PSF, i.e., on the distribution of its bright and dark regions, and not
on its irradiance distribution. Some of the symmetry properties of the aberrated PSFs
discussed above are evident from these pictures.
Figures 3-19 and 3-20 illustrate the defocused PSFs for = 0.5 and = 0.8. From
( )
Eq. (3-77) the axial irradiance is zero when Bd 1 - 2 is equal to an integral number of
wavelengths. Thus, when = 0.5, the central irradince of the PSFs observed in image
planes corresponding to Bd = ± 4 3 , ± 8 3, etc., is zero. Similarly, when = 0.8, the
central irradiance is zero in planes corresponding to Bd = ± 2.78 , ± 5.56 , etc. Hence,
there is a dark spot at the center of these PSFs. The PSFs aberrated by spherical
( )
aberration are shown in Figures 3-21 and 3-22. Both the classical As r 4 and balanced
[ ( ) ]
As r 4 - 1 + 2 r2 aberrations are considered. There is no dark spot at the center of the
PSFs for these aberrations.
(
Figures 3-23 and 3-24 show the PSFs aberrated by classical Aa r2 cos 2 q and )
( )
balanced Aa r2 cos 2 q - r2 2 astigmatism, respectively for = 0.5. The balancing
defocus is independent of the value of . For a given value of Aa , the PSFs in planes
Bd = 0 and Bd = - Aa are the sagittal and tangential images. They are identical to each
other except for a rotation of one with respect to the other by p 2. This is a special case
of the general result that the PSFs in two planes located symmetrically about the plane
Bd = - Aa 2 for balanced astigmatism are identical except for a rotation by p 2. For
example, when Aa = 3 l , the PSFs in defocused planes Bd = 1.5 l and Bd = - 4.5 l
differ from each other only by the rotation. The PSFs have bilateral symmetry. The PSF
in a defocused plane Bd = -1.5 l corresponds to balanced astigmatism and has a 4-fold
symmetry. Similar results are obtained when = 0.8. The sagittal and tangential images
for a large value of deviate considerably from the line images of geometrical optics
even when the aberration is large. See, for example, the sagittal image for Aa = 3 l , i.e.,
in the plane Bd = 0 . The central bright line appears broken near the center when
compared with the corresponding image for = 0 in Figure 2-35.
The irradiance distribution for coma is symmetric about the tangential plane. Thus, it
has a line symmetry in any observation plane, the line lying in the tangential plane. The
distribution in two observation planes located symmetrically about the Gaussian image
plane are identical. A change in the sign of the aberration coefficient A c produces a
rotation of the distributions by p about the axis. The PSFs aberrated by coma are shown
in Figures 3-27 and 3-28. The conical shape of the PSFs for large aberration when = 0
is lost for large values of .
Bd = 0 Bd = 4 3
Bd = 0.5 Bd = 8 3
Bd = 1 Bd = 16 3
Figure 3-19. Defocused PSFs for = 0.5 . The central irradiance of a PSF is zero
( )
when Bd 1 - 2 = ± nl , where n is an integer.
3.8 2-D PSFs 313
Bd = 1 Bd = 2.78
Bd = 2 Bd = 5.56
Bd = 3 Bd = 8.33
Figure 3-20. Defocused PSFs for = 0.8 . The central irradiance of a PSF is zero
( )
when Bd 1 - 2 = ± nl , where n is an integer.
As = 0.5, Bd = 0 As = 1, Bd = - 1.25
As = 1, Bd = 0 As = 2, Bd = - 2.5
As = 2, Bd = 0 As = 3, Bd = - 3.75
Gaussian image plane ( )

Defocused plane Bd = - 1 + 2 As
Figure 3-21. PSFs for = 0.5 and spherical aberration.

3.8 2-D PSFs 315
As = 0.5, Bd = 0 As = 1, Bd = - 1.64
As = 1, Bd = 0 As = 2, Bd = - 3.28
As = 2, Bd = 0 As = 3, Bd = - 4.92
Gaussian image plane (
Defocused plane Bd = - 1 + 2 As)
Figure 3-22. PSFs for = 0.8 and spherical aberration.
Aa = 1, Bd = 0 Aa = 1, Bd = 0.5
Aa = 3, Bd = 1.5 Aa = 3, Bd = 0
Figure 3-23. PSFs for = 0.5 and astigmatism observed in various image planes.
3.8 2-D PSFs 317
Aa = 1 4 Aa = 1 2
Aa = 1 Aa = 2
Aa = 3
( )
Figure 3-24. PSFs for = 0.5 and balanced astigmatism Aa r2 cos 2 q - r2 2 . Thus,
Bd = - Aa 2 and the PSFs are 4-fold symmetric.
Aa = 1, Bd = 0 Aa = 1, Bd = 0.5
Aa = 1, Bd = - 1.5
Aa = 3, Bd = 0 . Aa = 3, Bd = 1.5
Figure 3-25. PSFs for = 0.8 and astigmatism observed in various image planes.
3.8 2-D PSFs 319
Aa = 1 4 Aa = 1 2
Aa = 1
Aa = 2 Aa = 3
( )
Figure 3-26. PSFs for = 0.8 and balanced astigmatism Aa r2 cos 2 q - r2 2 . Thus,
Bd = - Aa 2 and the PSFs are 4-fold symmetric.
Ac = 1 4 Ac = 1 2
Ac = 1 Ac = 2
Ac = 3
Figure 3-27. PSFs for = 0.5 and coma Acr3 cosq.

q They are symmetric about the
horizontal axis.
3.8 2-D PSFs 321
Ac = 1 4 Ac = 1 2
Ac = 1 Ac = 2
Ac = 3
Figure 3-28. PSFs for = 0.8 and coma Acr3 cosq . They are symmetric about the
horizontal axis.

Finally, we discuss the LOS of an aberrated system in terms of the centroid of its
PSF. Unlike a system with a circular pupil, the LOS error depends on the order of the
coma aberration. Moreover, for a given amount of aberration of a certain order, the error
is larger for an obscured pupil.

The aberrated PSF is given by Eq. (3-96). The line of sight (LOS) of an aberration-
free system coincides with the center of its PSF. For an aberrated system, let us define its
LOS as the centroid of its aberrated PSF, as in Section 2.9. Let the aberration function of
the system in terms of the orthonormal Zernike annular polynomials be given by
• n 12
W (r, q; ) = Â
n=0 m = 0
[
Â 2 (n + 1) (1 + d m 0 ) ] Rnm (r; ) (cnm cos m q + snm sin m q) , (3-106)
where cnm and snm are the orthonormal Zernike aberration coefficients. For a system
with a uniformly illuminated annular pupil of inner and outer radii a and a ,
respectively, the centroid of its aberrated PSF, following Eqs. (1-181), is given by
2p
Û
< xi , yi > = [ R Sex ( )] Ù
ı
[W (a, q; ) - W (a, q; )] a (cos q,sin q) d q , (3-107)
0
where we have made use of Eqs. (2-150). Substituting Eq. (3-106) into Eq. (3-107), we
obtain
•
[ (
< xi , yi > = 2 F 1 - 2 )] n =1
1
[ 1
]
Â ¢ 2(n + 1) Rn (1; ) - Rn ( ; ) (cn1 , sn1 ) , (3-108)
where a prime on the summation sign indicates a summation over odd integral values of
n. Thus, the only aberrations that contribute to the LOS error are those with m = 1.
Aberrations of the type Rn1 (r;) cos q contribute to < xi > and those of the type
Rn1 (r;) sin q contribute to < yi > . This follows from the symmetry of the aberrations.
We note that, whereas two aberration terms with m = 1 but different values of n give
(approximately) the same Strehl ratio if their coefficients are equal in magnitude, their
contribution to the LOS error is different for a different aberration.
Since a radial polynomial Rn1 (r;) consists of terms in rn , rn -2 , ..., and r, with their
coefficients varying with , there is no loss of generality if we consider aberrations of the
type rn cos q , where n is an odd integer, to determine their contribution to the LOS error.
Thus, we consider, for simplicity, an aberration
W (r, q) = Wnrn cos q , (3-109)
where Wn is its peak value (at the outer edge of the annular pupil relative to a value of
zero at its center). Substituting Eq. (3-109) into Eq. (3-107), we obtain
( n -1) 2
< x > = 2Wn Â 2i , (3-110)
i=0
where Wn is in units of l , < x > is in units of l F , and it is understood that < y > = 0 .
As discussed in Section 2.9, when = 0 , the LOS error depends only on Wn , but not on
the power n of r in Eq. (3-109). However, when π 0 , the LOS error does depend on n.
The reason for this is that, even if the aberration value along the outer perimeter of an
annular pupil is the same for different values of n, its value along the inner perimeter is
different. Hence, for a given value of Wn an annular pupil gives aberrated PSFs with
different centroids for different orders of the aberration. Of course, these observations
also hold for aberrations represented by Zernike annular polynomials, as may be seen by
noting that whereas Rn1 (1; 0) = 1 , regardless of the value of n, Rn1 (1;) and Rn1 (; ) do
depend on n.
Following the procedure used in Section 2.9.1, it can be shown that for an
aberration
• n
W (r, q) = Â ¢ Wnr cos q , (3-111)
n =1
the irradiance distribution of the image of a point object formed by an aberrated optical
system with an annular exit pupil may be written
2
È1 ˘
ÍÛ J ( p B) r dr ˙
[ (
I ( r , q i ; ) = 4 1 - 2 )] ÍÙı
0 ˙
, (3-112)
ÍÎ ˙˚
where B is given by Eq. (2-155) and Wn is in units of l. The centroid of the distribution
is given by
• ( n -1) 2
<x> = 2 Â ¢Wn Â 2i . (3-113)
n =1 i=0
The aberration corresponding to a wavefront tilt is given by
W (r, q) = W1r cos q . (3-114)
The PSF simply shifts such that its peak and centroid locations move from (0, 0) to
< x > = 2W1 . (3-115)


The primary coma aberration is given by
W (r, q) = W3r3 cos q . (3-116)
The PSF along the x-axis may be written

2
È 1 ˘
2 -1 ÍÛ ˙
(
I ( x ; ) = 1 - ) ÍÙı
J 0 ( p B) dt ˙ , (3-117)
Í 2 ˙
Î ˚
where
B = (2t W3 - x ) t 1 2 . (3-118)
Figure 3-29 shows how I(x) varies with x for several values of W3 (in units of l) varying
from 0 to 2 and 2 = 0.5 . The centroid of the PSFs is given by
(
< x > = 2W3 1 + 2 ) . (3-119)
Thus, for a given value of W3 , the centroid for an annular pupil shifts by an amount that
( )
is larger by a factor of 1 + 2 than that for a circular pupil.
1.0
W3 = 0
2 = 0.5 0.5
0.8
0.6
l
1.5
0.4
0.2 2
0.0
–3 –2 –1 0 1 2 3
x
Figure 3-29. PSF for several typical values of primary coma W3 in units of l and
2 = 0.5 .
For small values of W3 , the peak value of the aberrated PSF occurs at a point such
that, if the aberration is measured with respect to a reference sphere centered at this point,
the variance of the aberration across the annular pupil is minimum. From the properties of
the Zernike annular polynomials, we find that the polynomial R31 (r;) cos q gives the
optimum combination of r3 cos q and r cos q terms that lead to a minimum variance.
Since
R31 (r; ) =
( ) (
3 1 + 2 r3 - 2 1 + 2 + 4 r ) , (3-120)
12
(1 - ) [(1 + ) (1 + 4
2 2 2
+ )]
4
we note that, for small values of W3 , the peak value of the aberrated PSF occurs at
(
x m = 4 W3 1 + 2 + 4 3 1 + 2 ) ( ) , (3-121)
where the subscript m refers to the point corresponding to minimum aberration variance.
From the form of the aberration, it is understood that ym = 0 . Thus an amount W3 of
primary coma shifts the centroid and peak of the PSF by different amounts, the
2
( ) (
movement of the peak being 2 1 + 2 + 4 3 1 + 2 of the movement of the centroid. )
Figure 3-30 shows how the irradiance Im at x m , the peak irradiance I p , and the
irradiance Ic at < x > vary with W3 . Figure 3-31 shows how x m , x p (the point at which
the peak irradiance occurs) and < x > vary with W3 . The observations made above about
the PSFs aberrated by primary coma are evident from these figures. Several typical values
of x m , x p , and < x > and the corresponding irradiances Im , I p , and Ic are noted in
Table 3-9 for 2 = 0.5 . We note that the peak value lies approximately at x m for small
values of W3 . For large values of W3 , it occurs at a point that is closer to the origin than
x m . The distance of the peak from the origin does not increase monotonically but
fluctuates as W3 increases. Since according to Eq. (3-110), the distance of the centroid
increases linearly with W3 , it is clear that the separation between the locations of the
centroid and the peak increases as W3 increases.
Table 3-9. Typical values of x m , x p , and < x > and the corresponding irradiances
Im , I p , and Ic for annular pupils aberrated by primary coma* when 2 = 0.5 .
0 0 0 0 1 1 1 1
0.5 0.78 0.78 1.50 0.9283 0.9283 0.0524 0.0403
1.0 1.56 1.55 3.00 0.7410 0.7412 0.1357 0.0319
1.5 2.33 2.32 4.50 0.5063 0.5064 0.0139 0.0010
2.0 3.11 3.07 6.00 0.2936 0.2946 0.0160 0.0000
*The aberrated central irradiance I(0), i.e., the Strehl ratio, is also given here.
1.0
2 = 0.5
0.8
0.6
l
0.4
lc
0.2
lp
lm
0.0
0 2 4 6 8 10
W3
Figure 3-30. Variation of Im , I p , and Ic with W3 for 2 = 0.5 .
30
2 = 0.5
20
x
<x>
xm
10
xp
0
0 2 4 6 8 10
W3
Figure 3-31. Variation of x m , x p , and < x > with W3 for 2 = 0.5 .


The secondary coma aberration is given by
W (r, q) = W5r5 cos q . (3-122)
The aberrated PSF along the x-axis is given by Eq. (3-117), where
(
B = 2t 2 W5 - x t 1 2 ) . (3-123)
Figure 3-32 shows how I ( x;) varies with x for several values of W5 and 2 = 0.5 .
Following Eq. (3-110), the centroid of the PSF is given by
(
< x > = 2 W5 1 + 2 + 4 ) . (3-124)
(
The variance of the aberration r5 cos q is minimized if - W5 1 + 2 + 4 + 6 1 + 2 )( )
amount of r cos q aberration is introduced. Accordingly, the point x m with respect to
which the aberration variance is minimum is given by
(
x m = W5 1 + 2 + 4 + 6 ) (1 + ) 2
. (3-125)
The values of x m , x p , and < x > and the corresponding irradiances Im , I p , and Ic are
noted in Table 3-10 for the values of W5 considered in Figure 3-32 and 2 = 0.5 .
1.0
W5 = 0
2 = 0.5
0.8
0.5
0.6
l
1
0.4
0.2 1.5
0.0
–3 –2 –1 0 1 2 3
x
Figure 3-32. PSF for several typical values of secondary coma W5 in units of l and
2 = 0.5 .
Im , I p , and Ic for annular pupils aberrated by secondary coma* when 2 = 0.5 .
0 0 0 0 1 1 1 1
0.5 0.63 0.62 1.75 0.8400 0.8402 0.0760 0.1768
1.0 1.25 1.21 3.50 0.4948 0.4966 0.0282 0.0002
1.5 1.88 1.74 5.25 0.2003 0.2196 0.0130 0.0009
2.0 2.50 1.71 7.00 0.0573 0.1478 0.0074 0.0065
*The aberrated central irradiance I(0), i.e., the Strehl ratio, is also given here.
The variance of the aberration r5 cos q is reduced even further if an appropriate

amount of r3 cos q aberration is also introduced. For a given value of W5 , the
appropriate amounts of W3 and W1 that give minimum variance may be obtained from the
Zernike annular radial polynomial R51 (r;) , where
R15 (r; ) =
( ) ( ) (
10 1 + 4 2 + 4 r 5 - 12 1 + 4 2 + 4 4 + 6 + 3 1 + 4 2 + 104 + 4 6 + 8 r ) (3-126)
12
(1 - ) [(1 + 4 + ) (1 + 92 + 94 + 6 )]
2 2 2 4
As an example, we consider the PSF aberrated by an aberration
(
W (r, q) = W 5r 5 + W 3r 3 cos q , ) (3-127)
where
(
W3 = - 1.2 W5 1 + 4 2 + 4 4 + 6 ) (1 + 4 2
+ 4 ) . (3-128)
According to Eq. (3-126), the point in the image plane with respect to which the
aberration variance is minimized is given by
(
x m = - 0.6 W5 1 + 4 2 + 10 4 + 4 6 + 8 ) (1 + 4 2
+ 4 ) . (3-129)
Substituting Eqs. (3-127) and (3-128) into Eq. (3-113), we obtain the centroid
(
< x > = - W5 0.4 + 2 2 + 7.2 4 + 2 6 + 0.48 ) (1 + 4 2
+ 4 ) . (3-130)
Substituting Eqs. (3-127) into Eq. (3-112), we find that the aberrated PSF along the x-axis
is given by Eq. (3-117), where
(
B = 2t 2 W5 + 2tW3 - x t 1 2 ) . (3-131)
Figure 3-33 shows the aberrated PSF I(x) for several values of W5 with W3 given by Eq.
(3-128) and 2 = 0.5 . The values of x m , x p , and < x > and the corresponding irradiances
Im , I p , and Ic are given in Table 3-11. Note that x m , x p and < x > are all negative.
Moreover, their magnitude for the values of W5 considered is very large. Therefore, in
Figure 3-33, the horizontal coordinate is chosen to be x - x m .
1.0
W5 = 0
2 = 0.5
W3 ~ – 1.52 W5 5
0.8 xm ~ – 1.52 W5
0.6
10
l
0.4
15
0.2
20
0.0
–3 –2 –1 0 1 2 3
x – xm
Figure 3-33. PSF for aberration given by Eq. (3-127) and 2 = 0.5 . Note that the
horizontal coordinate is x - x m .
Im , I p , and Ic for annular pupils aberrated by a combination of primary and
secondary coma given by Eq. (3-127).
0 0 0 0 1 1 1 1
5.0 –5.60 –5.60 –5.35 0.8832 0.8832 0.0005 0.0039
10.0 –11.19 –11.23 –10.69 0.6101 0.6128 0.0000 0.0014
15.0 –16.79 –16.94 –16.04 0.3353 0.3558 0.0000 0.0000
20.0 –22.38 –22.80 –21.38 0.1493 0.2296 0.0000 0.0008

REFERENCES
1. V. N. Mahajan, “Included power for obscured circular pupils,” Appl. Opt. 17,
964–968 (1978).
2. H. F. Tschunko, “Imaging performance of annular apertures,” Appl. Opt. 18,

1820–1823 (1974).
3. E. L. O’Neill, “Transfer function for an annular aperture,” J. Opt. Soc. Am. 46,
285–288 (1956). Note that a term of - 2 h2 is missing in the second of O’Neill’s
Eq. (26).
4. W. H. Steel, “Étude des effets combines des aberrations et d’une obturation

centrale de la pupille sur le contraste des images optiques.” Rev. Opt. (Paris) 32,
143–178 (1953).
circular and annular pupils,” J. Opt. Soc. Am. 72, 1258–1266 (1982); errata, J.
Opt. Soc. Am. A10, 2092, (1993); also, “Strehl ratio for primary aberrations in
terms of their aberration variance,” J. Opt. Soc. Am. 73 860–861 (1983).

pupils,” J. Opt. Soc. Am. 71, 75–85 (1981); 71, 1408 (1981); 1, 685 (1984);
“Zernike annular polynomials and optical aberrations of systems with annular
pupils,” Appl. Opt. 33, 8125–8127 (1994).

8. V. N. Mahajan, “Axial irradiance and optimum focusing of laser beams,” Appl.

Opt. 22, 3042–3053 (1983).
9. V. N. Mahajan, “Symmetry properties of aberrated point-spread functions,” J.

Opt. Soc. Am. A11, 1993–2003 (1994).
A2, 833–846 (1985).
Problems 331
PROBLEMS
1. Consider an optical imaging system with an annular exit pupil of obscuration ratio
= 0.5 forming the image of a point object with a focal ratio of 8. Let the pupil of
outer radius 2 cm be uniformly illuminated with an irradiance of 0.1 W/cm2 at a
wavelength of 0.5 mm. (a) Determine the radius of the central bright spot of the
image, the amount of power contained in it, and the value of the irradiance at its
center. (b) Give the radius of the first bright ring of the image and the maximum
value of its irradiance distribution. How much power is contained in the bright
ring? (c) Give the diameter of the third dark ring and the power lying outside it. (d)
How much power would a square detector centered on the image collect if its half
width is equal to the radius of the third dark ring. (d) What is the spatial frequency
for which the MTF is equal to 0.5? What is the cutoff frequency of the system?
2. (a) Show that the standard deviation of a primary aberration is given by the
expressions listed in Table 3-4. (b) Also show, using Eq. (3-52), that a balanced
primary aberration is given by the expressions listed in Table 3-5.
3. Determine the focal point irradiance of an annular beam with an outer diameter of
25 cm and an obscuration ratio of 0.3 focused at a distance of 10 m. Let the total
power in the beam be 5 W with a wavelength of 10.6 mm. Determine the depth of
focus for a Strehl ratio of 0.8.
4. If the beam in Problem 3 has 0.5 l of primary spherical aberration, (a) determine
the range of distance across which a Strehl ratio of 0.8 is obtained. (b) Determine
the axial point about which the axial irradiance is symmetric. (c) Give approximate
maximum and minimum values of the power contained in a circular detector whose
radius is equal to that of the central bright spot in the focal plane as it is moved
along the axis in the range calculated in (a).
5. Determine the location of the point about which the axial irradiance of a beam
aberrated by 2 l of secondary spherical aberration W6r6 is symmetric.
6. Determine the location of and irradiance value at the centroid and peak of the focal
plane distribution if the beam in Problem 3 is aberrated by l 4 of primary coma.
CHAPTER 5
RANDOM ABERRATIONS
5.1 Introduction ..........................................................................................................389

5.2 Random Image Motion ........................................................................................389
5.2.1 General Theory ........................................................................................390
5.2.2 Circular Pupils ......................................................................................... 391
5.2.2.1 Theory ....................................................................................... 391
5.2.2.2 Gaussian Approximation........................................................... 392
5.2.2.3 Numerical Results ..................................................................... 393
5.2.3 Annular Pupils ......................................................................................... 393
5.2.3.1 Theory ....................................................................................... 393
5.2.3.2 Numerical results ......................................................................397
5.3 Imaging Through Atmospheric Turbulence......................................................401
5.3.1 Long-Exposure Image ............................................................................. 402
5.3.2 Kolmogrov Turbulence............................................................................407
5.3.3 Circular Pupils ......................................................................................... 413
5.3.4 Annular Pupils ......................................................................................... 417
5.3.5 Phase Aberration in Terms of Zernike Circle Polynomials..................... 421
5.3.6 Short Exposure Image..............................................................................430
5.3.6.1 Near-Field Imaging ................................................................... 430
5.3.6.2 Far-Field Imaging......................................................................436
5.3.6.3 Short-Exposure Images ............................................................. 439
5.3.7 Adaptive Optics ....................................................................................... 441
Appendix: Fourier Transform of Zernike Circle Polynomials ................................443
References ......................................................................................................................445
Problems ......................................................................................................................... 447
387
Chapter 5
Random Aberrations
5.1 INTRODUCTION
So far we have considered deterministic aberrations such as those that are inherent in
the design of an optical imaging system. These aberrations are deterministic in the sense
that they are known or can be calculated, for example, by ray tracing the system. Now we
consider the effects of aberrations that are random in nature on the quality of images. The
aberration is random in the sense that it varies randomly with time for a given system, or
it varies randomly from one sample of a system to another. An example of the first kind is
the aberration introduced by atmospheric turbulence when an optical wave propagates
through it, as in ground-based astronomical observations. An example of the second kind
is the aberration introduced due to polishing errors of the optical elements of the system.
The polishing errors of an element fabricated similarly in large quantities vary randomly
from one sample to another. In either case, we cannot obtain the exact image unless the
instantaneous aberration or the exact polishing errors are known. However, based on the
statistics of the aberrations, we can obtain the time- or ensemble-averaged image.
We discuss the effects of two types of random aberrations: random wavefront tilt
causing random image motion, and random aberrations introduced by atmospheric
turbulence. The time-averaged Strehl ratio, point-spread function (PSF), optical transfer
function (OTF), and encircled power are discussed for the two types of aberrations. A
coherence length of atmospheric turbulence is defined, which limits the resolution of an
imaging system, regardless of how large its aperture is. Both long- and short-exposure
images are discussed, and expressions for the aberration variance are given in both cases.
These expressions can be used to define the requirements of a steering mirror for
corrections of wavefront tilt and a deformable mirror for corrections of wavefront
deformation or the aberrations. It is shown that in a severe turublence, a short-exposure
image breaks up into speckles whose size is determined by the resolution of the system.
Although much of our discussion is on systems with circular pupils, systems with annular
pupils are also considered, and differences between the two types are outlined.
5.2 RANDOM IMAGE MOTION1

In many optical imaging systems, especially those used in space, there is always
some image motion during an exposure interval. The source of image motion may, for
example, be vibration of optical elements and servo dither in the pointing system. In the
case of beam transmitting systems, the beam itself may have some motion associated with
it. Here, we obtain expressions for the time-averaged PSF, Strehl ratio, OTF, and
encircled power for an imaging system with a circular exit pupil undergoing Gaussian
random motion. A simple approximate model based on a Gaussian approximation of its
motion-free PSF is also developed, and numerical results provided by it are compared
with the exact results.
389
390 RANDOM ABERRATIONS
5.2.1 General Theory

r
Let Ii ( ri ) be the irradiance distribution of the motion-free image of a point object.
r
Let the random motion of the image be described by a probability density function p( ri ) .
r r r
Then, the probability that the center of the image lies between r j and rj + d rj is given by
r r r r r r r
( ) (
p rj d rj . The corresponding irradiance at a point ri is given by Ii ri - rj p rj d rj . The
r ) ( )
time-averaged irradiance Ii ( ri ) at this point is obtained by integrating this expression
r
over all possible values of r j , i.e.,
r r r r r
Ii ( ri ) = Ú Ii ( ri - rj ) p ( rj ) d rj , (5-1)
where angular brackets indicate a time average. Thus, the time-averaged irradiance
distribution of the image formed by an imaging system undergoing random motion is
given by the convolution of its motion-free distribution and the probability density
function describing its motion. The corresponding PSF is obtained by dividing both sides
of Eq. (5-1) by Pex , where Pex is the total power in the exit pupil of the system and,
therefore, in the image. Following Eq. (1-56c), we note that the time-averaged PSF can
also be interpreted as the irradiance distribution of the image of an incoherent object
r
whose Gaussian-image irradiance distribution is given by p ( ri ) . Fourier transforming Eq.
(5-1), we obtain
r r r
t ( vi ) = t ( vi ) P ( vi ) , (5-2)
r r
where t ( vi ) and t ( vi ) are the motion-free and time-averaged OTFs, respectively,
r
corresponding to image spatial frequency v i , and
r r r r r
P ( vi ) = Ú p ( ri ) exp (2 p i vi ◊ ri ) d ri (5-3)
describes the image motion in the spatial frequency domain.
Following Eq. (1-76), the time-averaged irradiance distribution can be written in

r
terms of the corresponding time-averaged OTF t ( vi ) according to
r r r r r
Ii ( ri ) = Pex Ú t (vi ) exp ( - 2 p i vi ◊ ri ) d vi . (5-4)
For image motion described by statistically independent Gaussian random processes

of zero mean and equal standard deviation s i along each of the two axes of an image, the
probability density function may be written
( )
p (ri ) = 1 / 2 p s i2 exp - ri2 / 2s i2( ) , (5-5)
r r
where ri = ri . Substituting Eq. (5-5) into Eq. (5-3) and letting vi = vi , Eq. (5-2) for the
time-averaged OTF may be written
r r
(
t ( vi ; s i ) = t ( vi ) exp - 2 p 2 s i2 vi2 ) . (5-6)
5.2 Random Image Motion 391
5.2.2 Circular Pupils

Now we consider how the aberration-free image formed by a system with a circular
pupil degrades with image motion. A Gaussian approximation of the results is also
considered. Numerical results are given for the time-averaged Strehl ratio, PSF, and
encircled power.
5.2.2.1 Theory
Consider systems with circular pupils and radially symmetric motion-free PSF and,
therefore, radially symmetric motion-free OTF. Substituting Eq. (5-6) into Eq. (5-4) and
using Eq. (2-12), we obtain the time-averaged irradiance
Ii (ri ; s i ) = 2 p Pex Ú t (vi ; s i ) J 0 (2 p ri vi ) vi dvi , (5-7)
where
(
t (vi ; s i ) = t (vi ) exp - 2 p 2 s i2 vi2 ) (5-8)
is the radially symmetric time-averaged OTF. The corresponding PSF is obtained by

dividing both sides of Eq. (5-7) by Pex . Equation (5-7) also follows from Eq. (1-85).
Using normalized quantities defined by Eqs. (2-9a), (2-10), and (2-47), and letting
s = s i l F , Eq. (5-7) reduces to
1
I (r; s ) = 8 Ú t (v; s ) J 0 (2 p rv) v dv , (5-9)
0
where
(
t (v; s ) = t (v) exp - 2 p 2 s 2 v 2 ) (5-10)
is the time-averaged OTF. The motion-free OTF t (v) is given by Eq. (2-44).
Defining the Strehl ratio of the image as the ratio of the central irradiances with and
without image motion, the time-averaged Strehl ratio can be written
S (s ) = I (0; s ) I (0)
1
= 8 Ú t (v; s ) v dv .
0
(5-11)
( )
Note that the motion-free central irradiance I(0) is unity in units of Pex Sex l2 R 2 .
The encircled power in terms of the OTF is given by Eq. (1-89). Accordingly, the
time-averaged fractional encircled power is given by
1
P (rc ; s ) = 2 p rc Ú t (v; s ) J1 (2 p rc v) dv . (5-12)
0
This result may also be obtained by substituting Eq. (5-9) into Eq. (2-22b) and using Eq.
(1-88). Numerical results for an aberration-free system undergoing Gaussian random
motion may be obtained by substituting Eq. (2-44) into Eq. (5-10) and substituting the
expression obtained into Eqs. (5-9), (5-11), and (5-12). However, the integration in Eq.
(5-11) thus obtained can be carried out analytically, yielding the results
( ){
S (s ) = 2 p 2 s 2 1 - exp - p 2 s 2 ( ) [ I (p s ) + I (p s )] }
0
2 2
1
2 2
, (5-13)
where I0 (∑) and I1 (∑) are the hyperbolic Bessel functions of zero order and first order,
respectively. In the following, we consider an aberration-free system undergoing random
Gaussian motion as described above.
5.2.2.2 Gaussian Approximation

The Gaussian approximation of the motion-free irradiance distribution having the
same central value and total power as the actual is given by Eq. (2-125), namely,
[
Ig (r ) = exp - ( p r 2)
2
] . (5-14)
The corresponding approximation of the OTF is given by its (slightly modified) zero-
order Hankel transform according to Eq. (2-49a), i.e.,
t g (v) = exp - 4v 2 ( ) . (5-15)
Following Eq. (5-6), the corresponding time-averaged OTF is given by
t g (v; s ) [
= exp - 2 p 2 s 2 + 2 p 2 v 2 ( ) ] . (5-16)
The time-averaged irradiance distribution is also Gaussian with a variance equal to the
sum of the variances of the motion-free distribution and the image motion; i.e.,
[ ( p 2 ) (s )] [ ( )]
-1
I g ( r; s ) = 2 2
+ 2 p2 exp - r 2 2 s 2 + 2 p 2 . (5-17)
The time-averaged Strehl ratio is given by
Sg (s ) = Ig (0; s )
( )
-1
= 1 + p2s2 2 . (5-18)
For large values of s , Eq. (5-18) may be written
Sg (s ) Æ 2 p 2 s 2 . (5-19)
This result may also be obtained from Eq. (5-13), since, for large arguments, the
hyperbolic Bessel functions approach
I m ( z ) Æ e z (2 p z )
12
. (5-20)
Following Eq. (2-22b), the time-averaged encircled power is given by

rc
( )
Pg (rc ; s ) = p 2 2 Ú Ig (r; s ) r dr
0
[ (
= 1 - exp - rc2 2 s 2 + 2 p 2 )] . (5-21)
5.2.2.3 Numerical Results

Now we consider aberration-free systems with circular pupils undergoing Gaussian
random image motion. For such systems, if we let s = 0 , the equations obtained above
for the Strehl ratio, PSF, and OTF would reduce to the corresponding equations given in
Chapter 2. For example, Eq. (5-10) would reduce to Eq. (2-44) for the aberration-free
OTF. Similarly, Eqs. (5-9) and (5-12) for the PSF and encircled power would reduce to
Eqs. (2-15) and (2-24), respectively. Of course, Eqs. (5-11) and (5-13) would give a
Strehl ratio of unity.
Figure 5-1 shows how the exact and approximate time-averaged Strehl ratios given
by Eqs. (5-13) and (5-18), respectively, vary with s . The exact results are shown by the
solid curves and the corresponding approximate results are shown by the dashed curves.
The values of the time-averaged Strehl ratio are given in Table 5-1 for s = 0 (0.05) 1. We
note that the Gaussian approximation given by Eq. (5-18) overestimates the Strehl ratio.
( )
However, the maximum fractional difference S - Sg S is less than 12 percent and
occurs at s = 1.1. The time-averaged irradiance distributions obtained according to Eqs.
(5-9) and (5-17) for typical values of s are shown in Figure 5-2. The corresponding
encircled powers obtained by use of Eqs. (5-12) and (5-21) are shown in Figure 5-3. Once
again, we note that the Gaussian approximation overestimates the irradiance and
encircled power (at least within the Airy disc). However, the differences between the
exact and approximate results are not large. In the case of encircled power, the difference
increases monotonically as the radius of the circle increases.
5.2.3 Annular Pupils

Next, we apply the general equations obtained in Section 5.2.1 to systems with
annular pupils and give numerical results on the time-averaged Strehl ratio, PSF, and
encircled power. We show, for example, that the Strehl ratio for a certain image motion
decreases as the obscuration increases.
5.2.3.1 Theory
Now we consider the degradation of an image formed by a system with an annular
pupil due to its random motion. Following the discussion of the previous section, the
time-averaged irradiance distribution of the image of a point object for Gaussian random
motion may be written
1.0 1.0
0.8 0.8
0.6 0.6
S
S
0.4 0.4
<S> <Sg>
0.2 0.2
<S>
<Sg>
0.0 0.0
0.0 0.2 0.4 0.6 0.8 1.0 0 1 2 3 4 5
s s
Figure 5-1. Time-averaged Strehl ratio for Gaussian random image motion. s is in
units of l F. The dashed curves represent the Gaussian approximation of the exact
results represented by the solid curves.
Table 5-1. Time-averaged Strehl ratio for various values of s .
s S Sg
0 1 1
0.05 0.988 0.988
0.10 0.953 0.953
0.15 0.898 0.900
0.20 0.831 0.835
0.25 0.756 0.764
0.30 0.680 0.693
0.35 0.606 0.623
0.40 0.538 0.559
0.45 0.476 0.500
0.50 0.422 0.448
0.55 0.375 0.401
0.60 0.334 0.360
0.65 0.298 0.324
0.70 0.268 0.293
0.75 0.241 0.265
0.80 0.218 0.241
0.85 0.198 0.219
0.90 0.181 0.200
0.95 0.165 0.183
1 0.152 0.169
1.0 1.0
s= 0 s = 0.25
0.8 0.8
0.6 0.6
I
I
0.4 0.4

0.2 0.2

<Ig>
<Ig>
0.0 0.0
0 1 2 3 0 1 2 3
r r
1.0 1.0
s = 0.5 s = 0.75
0.8 0.8
0.6 0.6
I
<Ig>
0.4 0.4

0.2 <Ig> 0.2

0.0 0.0
0 1 2 3 0 1 2 3
r r
1.0
s= 1
0.8
0.6
I

0.4
0.2
<Ig>
0.0
0 1 2 3
r
Figure 5-2. Time-averaged irradiance distribution for Gaussian random image

motion. The solid and dashed curves represent the exact and approximate results,
respectively.
1.0 1.0
<Pg> <Pg>
0.8 0.8 
0.6 0.6
P(rc)
P(rc)
0.4 0.4
0.2 0.2
s= 0 s = 0.25
0.0 0.0
0 1 2 3 0 1 2 3
rc rc
1.0 1.0
<Pg>
<Pg>
0.8 0.8
 
0.6 0.6
P(rc)
P(rc)
0.4 0.4
0.2 0.2
s = 0.5 s = 0.75
0.0 0.0
0 1 2 3 0 1 2 3
rc rc
1.0
0.8 <Pg>

0.6
P(rc)
0.4
0.2
s= 1
0.0
0 1 2 3
rc
Figure 5-3. Time-averaged encircled power for Gaussian random image motion.
The solid and dashed curves represent the exact and approximate results,
respectively.
Ii (ri ; ; s i ) = 2 p Pex Ú t (vi ; ; s i ) J 0 (2 p ri vi ) vi dvi , (5-22)
where
(
t (vi ; ; s i ) = t (vi ; ) exp - 2 p 2 s i2 vi2 ) (5-23)
is the time-averaged OTF of the system. The corresponding PSF is obtained by dividing
both sides of Eq. (5-22) by Pex . The motion-free OTF t (vi ; ) is discussed in Section
3.2.6. Using normalized quantities defined by Eqs. (3-10a), (3-11), and (2-47), and letting
s = s i l F , Eq. (5-22) becomes
[ ( )] Ú
1
I (r; ; s ) = 8 1 - 2 t (v; ; s ) J 0 (2 p rv) v dv , (5-24)
0
where
(
t (v; ; s ) = t (v; ) exp - 2 p 2 s 2 v 2 ) . (5-25)
The motion-free OTF t (v; ) is given by Eq. (3-40). The Strehl ratio of the image, i.e.,
the ratio of the central irradiances with and without image motion, is given by
S (; s ) = I (0; ; s ) I (0; )
[ ( )] Ú
1
= 8 1 - 2 t (v; ; s ) v dv . (5-26)
0
Note that the motion-free central irradiance I (0;) is unity in units of Pex Sex ( ) l2 R 2 .
The encircled power in terms of the OTF is given by Eq. (1-89). Following this
equation and Eq. (3-20), the time-averaged fractional encircled power may be written
P (rc ; ; s ) = 2 p rc Ú t (v; ; s ) J1 (2 p vrc ) dv . (5-27)
If we let = 0 , Eqs. (5-24) through (5-27) for the time-averaged irradiance, OTF, Strehl
ratio, and encircled power reduce to the corresponding Eqs. (5-9) through (5-12) for a
system with a circular exit pupil.
5.2.3.2 Numerical Results
Figure 5-4 shows how the Strehl ratio varies with s for = 0 (0.25) 0.75 . It
decreases monotonically as s increases. Numerical values of the Strehl ratio for
s = 0 (0.05) 1 and = 0 (0.05) 0.95 are given in Table 5-2. We note that as increases,
the drop in Strehl ratio due to image motion for a given value of s increases. This occurs
because the motion-free PSF (normalized to unity at the origin) for a larger value of is
smaller for small values of r for r £ 1. Figure 5-5 shows how the irradiance distribution is
affected by the image motion. We note that as s increases, the central irradiance
1.0
0.8
0.6 = 0.00
< S(; s)>
0.25
0.4 0.50
0.75
0.2
0.0
0.0 0.2 0.4 0.6 0.8 1.0
s
Figure 5-4. Time-averaged Strehl ratio as a function of s for several typical values
of .
decreases, the minima of the distribution are filled in, and the distribution becomes
smoother. The ring structure of the motion-free image disappears for s ≥ 0.5 , which is
approximately the spacing between adjacent maxima and minima.
Figure 5-5 also shows how the encircled power changes due to the image motion. We
note that the power contained in small circles decreases as s increases, because power
flows out of the central bright disc of the irradiance distribution. However, for large
circles, the power difference becomes alternately positive and negative as more power
flows inward or outward, respectively, from a bright ring of the irradiance distribution.
The range of s values for which this effect occurs increases as increases. Moreover,
whereas the central irradiance drops rapidly as s increases, it is found that the encircled
power for large circles (rc ≥ 3) does not change significantly for the S values considered
here. The time-averaged PSFs may again be approximated by a Gauusian, as in the case
of circular pupils.
Table 5-2. Time-averaged Strehl ratio for Gaussian image motion characterized by s
in units of l F (from Reference 1).
s \ 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45
0 1 1 1 1 1 1 1 1 1 1
0.05 0.988 0.988 0.988 0.987 0.987 0.987 0.987 0.986 0.986 0.985
0.10 0.953 0.952 0.952 0.952 0.951 0.950 0.948 0.947 0.945 0.943
0.15 0.898 0.898 0.898 0.896 0.895 0.892 0.890 0.887 0.883 0.879
0.20 0.831 0.831 0.830 0.828 0.825 0.822 0.817 0.813 0.807 0.801
0.25 0.756 0.756 0.754 0.752 0.748 0.744 0.738 0.732 0.724 0.716
0.30 0.680 0.679 0.678 0.674 0.670 0.665 0.658 0.650 0.642 0.632
0.35 0.606 0.606 0.603 0.600 0.595 0.589 0.582 0.573 0.564 0.554
0.40 0.538 0.537 0.535 0.531 0.526 0.520 0.512 0.503 0.494 0.483
0.45 0.476 0.476 0.473 0.469 0.464 0.458 0.450 0.442 0.433 0.423
0.50 0.422 0.421 0.419 0.415 0.410 0.404 0.397 0.389 0.380 0.371
0.55 0.375 0.374 0.372 0.368 0.364 0.358 0.351 0.244 0.336 0.328
0.60 0.334 0.333 0.331 0.328 0.323 0.318 0.312 0.306 0.299 0.293
0.65 0.298 0.298 0.296 0.293 0.289 0.285 0.279 0.274 0.268 0.263
0.70 0.268 0.267 0.265 0.263 0.260 0.256 0.251 0.247 0.242 0.237
0.75 0.241 0.241 0.239 0.237 0.234 0.231 0.227 0.223 0.220 0.216
0.80 0.218 0.218 0.216 0.215 0.212 0.209 0.206 0.203 0.201 0.198
0.85 0.198 0.198 0.197 0.195 0.193 0.191 0.188 0.186 0.184 0.182
0.90 0.181 0.180 0.179 0.178 0.176 0.175 0.173 0.171 0.169 0.168
0.95 0.165 0.165 0.164 0.163 0.162 0.160 0.159 0.158 0.157 0.156
1.00 0.152 0.152 0.151 0.150 0.149 0.148 0.147 0.146 0.145 0.145
s \ 0.50 0.55 0.60 0.65 0.70 0.75 0.80 0.85 0.90 0.95
0 1 1 1 1 1 1 1 1 1 1
0.05 0.985 0.984 0.983 0.983 0.982 0.981 0.980 0.979 0.979 0.981
0.10 0.941 0.939 0.936 0.933 0.930 0.927 0.924 0.920 0.917 0.916
0.15 0.875 0.870 0.865 0.859 0.853 0.847 0.841 0.834 0.827 0.823
0.20 0.794 0.787 0.779 0.771 0.762 0.752 0.743 0.733 0.723 0.715
0.25 0.708 0.698 0.688 0.677 0.666 0.655 0.643 0.630 0.618 0.608
0.30 0.622 0.611 0.599 0.587 0.575 0.562 0.549 0.536 0.524 0.513
0.35 0.543 0.531 0.519 0.507 0.494 0.481 0.468 0.456 0.443 0.433
0.40 0.472 0.461 0.449 0.437 0.425 0.413 0.402 0.390 0.379 0.370
0.45 0.412 0.402 0.391 0.380 0.369 0.359 0.348 0.338 0.328 0.321
0.50 0.362 0.352 0.343 0.333 0.324 0.315 0.306 0.297 0.289 0.283
0.55 0.320 0.312 0.304 0.296 0.288 0.280 0.273 0.265 0.258 0.253
0.60 0.286 0.279 0.272 0.265 0.259 0.252 0.246 0.240 0.234 0.229
0.65 0.257 0.251 0.246 0.240 0.235 0.229 0.224 0.218 0.212 0.203
0.70 0.233 0.228 0.224 0.219 0.215 0.210 0.206 0.201 0.196 0.187
0.75 0.212 0.209 0.205 0.202 0.198 0.194 0.190 0.186 0.181 0.174
0.80 0.195 0.192 0.189 0.187 0.184 0.181 0.177 0.174 0.169 0.163
0.85 0.180 0.178 0.176 0.173 0.171 0.169 0.166 0.163 0.159 0.153
0.90 0.166 0.165 0.164 0.162 0.160 0.158 0.156 0.153 0.150 0.144
0.95 0.155 0.154 0.153 0.152 0.150 0.149 0.147 0.144 0.141 0.136
1.00 0.144 0.144 0.143 0.143 0.142 0.140 0.139 0.137 0.134 0.131
1.0


0.8
; <P(rc)>
0.6
= 0.25
0.4
s=1
0.75
0.2 0.5
0.25
0
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
r; rc
1.0


0.8
s=1
; <P(rc)>
0.75
0.6
0.5 = 0.5
0.25
0
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
r; rc
1.0
s=1
= 0.75
0.8

0.75
; <P(rc)>
0.6

0.5
0.4
0.25
0
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
r; rc
Figure 5-5. Time-averaged irradiance and encircled power distributions for

annular pupils for several typical values of random image motion s .
5.3 Imaging Through Atmospheric Turbulence 401
5.3 IMAGING THROUGH ATMOSPHERIC TURBULENCE

Now we consider the effect of random aberrations introduced by atmospheric
turbulence on the image quality as in ground-based astronomy. A plane wave of uniform
amplitude and phase representing the light from a star is incident on the atmosphere. As it
propagates through the atmosphere, it undergoes both amplitude and phase variations due
to the random inhomogeneities in its refractive index. The wavefront gets distorted, i.e., it
becomes nonplanar with nonuniform amplitude across it. The amplitude and phase at a
point on the wavefront also vary randomly with time. The distorted wavefront of
nonuniform amplitude is incident on a ground-based imaging system. Unless the system
is aberration free, it also introduces aberrations. These aberrations are deterministic in the
sense that they are either time independent or their variation with time is known. The
system may also introduce deterministic amplitude variations across the wavefront.
First, we derive expressions for the time-averaged PSF and OTF degraded by
atmospheric turbulence. The pupil function of the overall imaging system is written as the
product of the pupil function of the optical system and a complex amplitude factor
introduced by turbulence. As a result, it is shown that the time-averaged OTF of the
overall system is also equal to the product of the OTF of the optical system and an OTF
reduction factor representing the effect of turbulence. The structure functions
representing the mean square difference of the log-amplitude and phase aberration
fluctuations at two points are introduced. The OTF reduction factor is obtained in terms
of a wave structure function that is the sum of the structure functions for log-amplitude
and phase fluctuations. The time-averaged images thus obtained are referred to as long-
exposure images. The exposure time may be 1 to 10 sec.
Next, Kolmogorov turbulence is discussed briefly and structure functions for

refractive index, and phase fluctuations are given. The corresponding power spectral
densities are also given. The concept of atmospheric coherence length r0 is introduced,
showing that it limits the resolution of the image regardless of how large the pupil
diameter of the optical system is. Systems with both circular and annular apertures are
considered; and Strehl ratio, PSF, and encircled power are discussed as a function of the
ratio of the pupil diameter and coherence length of atmosphere. The phase aberration
introduced by Kolmogorov turbulence is expanded in terms of Zernike polynomials, and
auto-correlation and cross-correlation of the expansion coefficients are given. It is shown,
for example, that if the random wavefront tilt introduced by turbulence is corrected in real
time, the variance of the aberration introduced by it is reduced by a factor of 7.7. The
resolution of a tilt-corrected image is equivalent to a short-exposure image; i.e., the
exposure is short enough that the image does not wander. The characteristics of a tilt-
corrected image are discussed and compared with those of a long-exposure image.
Examples of short-exposure images of a point object are shown, where the pupil
diameter D is kept fixed while r0 is varied, or r0 is varied while D is varied. When r0 is
much smaller than D, the image is shown to break up into small spots, called speckles,
whose size depends on the value of D. The overall size of the image in this case is
determined by the value of r0 .
5.3.1 Long-Exposure Image2

When the medium between an object and the optical system imaging it is
homogeneous, a spherical wavefront of uniform amplitude centered on an object point is
incident on the entrance pupil of the system. If Ien is the irradiance at a point on the
entrance pupil of area Sen , then the total object power incident on the pupil is given by
Pen = Ien Sen . (5-28)
If the system is aberration free and has uniform transmittance, a spherical wavefront of
uniform amplitude centered at the Gaussian image point emerges from its exit pupil.
However, if the system is aberrated and has nonuniform transmittance, then a distorted
r r
( )
wavefront of nonuniform amplitude emerges from its exit pupil. Let FL rp and AL rp
r ( )
be the phase and amplitude transmittance corresponding to a point r p on the exit pupil. A
dimensionless pupil function of the system may be written
(r ) (r )
PL rp = AL rp exp i F L rp [ (r )] . (5-29)
The total power in the exit pupil and, therefore, in the image is given by
(r ) r
2
Pex = Ien Ú PL rp d rp
(r )
= Ien Ú AL2 rp d rp
r
. (5-30)
When the medium between the object and the imaging system is inhomogeneous as
in ground-based astronomy, it introduces random phase and amplitude variations across
r r
the wavefront propagating through it. Let FR ( ren ) and l ( ren ) be the random phase and
r
(dimensionless) log amplitude introduced by atmospheric turbulence at a point ren on the
entrance pupil. The instantaneous object power at the entrance pupil is given by
r r
Pen = Ien Ú exp [2l ( ren )] d ren . (5-31)
Hence, the time-averaged total power at the entrance pupil may be written
r r
Pen = Ien Ú [
exp 2l ( ren ) ] d ren . (5-32)
Comparing Eqs. (5-28) and (5-32), conservation of energy yields that we must have
r
[
exp 2 l ( ren ) ] = 1 . (5-33)
As the wave propagates through the imaging system, it undergoes additional phase and
r
amplitude variations. The total phase aberration at a point r p on the exit pupil may be
written
(r )
F rp = F L rp + F R rp(r ) (r ) , (5-34)
r r
where the position vector r p is related to the position vector ren by the pupil
r r
( )
magnification m p i.e., rp = m p ren . The pupil function of the overall system (i.e.,
including the effects of atmospheric turbulence) representing the wavefront at the exit
pupil may be written
(r )
P rp = PL rp PR rp (r ) (r ) , (5-35)
where
(r ) [ (r )
PR rp = exp l rp + i F R rp (r )] (5-36)
(r )
is truncated by the pupil function PL rp and represents the complex amplitude variation
introduced by turbulence. The instantaneous total power at the exit pupil is given by
r r
Ú ( )
2
Pex = Ien P rp d rp
r r r
= Ien Ú AL (rp ) exp [2l (rp )] d rp
2
. (5-37)
The corresponding time-averaged total power is given by
r
Pex = Ien Ú A (r )
2
L p exp 2 l rp[ (r )] r
d rp
= Ien Ú (r )
AL2 rp d rp
r
, (5-38)
where we have made use of Eq. (5-33). Comparing Eqs. (5-30) and (5-38), we note that,
as expected, the average total power at the exit pupil is the same as the corresponding
total power in the absence of atmospheric turbulence. If the transmittance of the optical
r
system is uniform; e.g., if the transmittance is 100 percent so that AL rp = 1 m 2p , then ( )
Pex (
= Ien m 2p Sex ) (5-39a)
= Ien Sen , (5-39b)
where Sex is the area of the exit pupil. It should be evident that in this case, the irradiance
at the exit pupil is uniform, with a value of Iex = Ien m 2p .
From Eq. (1-62), where the pupil function is not dimensionless, the instantaneous
irradiance distribution of the star image formed by the overall system is given by
r Û r Ê 2 pi r rˆ r
2
Ii ( ri ) =
Pex
Sex l2 R 2 ı
( )
Ù P rp exp ÁË - l R rp ◊ ri ˜ d rp
¯
. (5-40)
Substituting Eq. (5-35) into Eq. (5-40), the time-averaged distribution may be written
r r r
ÚÚ PL (rp ) PL (rp¢) (r ) (r )
Pex
Ii ( ri ) = *
PR rp PR* rp¢
Sex l2 R 2
È 2p i r r
¥ exp Í-
Î lR
(
rp - rp¢ ) ◊ rri ˘˙˚ d rrp d rrp¢ . (5-41)
Dividing both sides of Eq. (5-41) by Pex yields the PSF. For simplicity of notation, we
let, for example
(r )
f rp = f1 (5-42a)
and
(r )
f rp¢ = f2 . (5-42b)
We also drop the subscript when convenient. For example, if f1 = f2 , we may simply
write f . Thus, considering Eq. (5-36), we may write the mutual coherence function
(r ) (r )
PR rp PR* rp¢ = P1 P2*
[
= exp (l1 + l 2 ) + i (F1 - F 2 ) ] . (5-43)
If the refractive index fluctuations are statistically stationary, i.e., they are statistically
homogeneous and isotropic, then so are the fluctuations in l and F which they generate.
Therefore,
(l1 + l 2 ) (F1 - F 2 ) = ( l1F1 - l2F 2 ) + ( l 2 F1 - l1F 2 ) (5-44a)
= 0 , (5-44b)
where the two averaged quantities in the first term on the right-hand side of Eq. (5-44a)
cancel each other due to homogeneity of turbulence, and those in the second term cancel
due to its isotropy. Thus, l1 + l 2 and F1 - F 2 are uncorrelated random variables. Hence,
Eq. (5-43) may be written
(r ) (r )
PR rp PR* rp¢ = exp (l1 + l 2 ) [
exp i (F1 - F 2 ) ] . (5-45)
For a Gaussian random variable x with a mean value of x and a standard deviation of
s , it is easy to show that
•
Ú exp (bx ) exp[- ( x - x ) ]

1
exp(bx )
2
= 2s 2 dx
2 ps
-•
= exp ÈÍ b 2 + b x ˘˙ ,
1
(x - x )2 (5-46)
Î2 ˚
where b is an arbitrary constant. Now we assume that l and FR are Gaussian random
variables and FR has a mean value of zero. Hence,
exp (2 l) = exp È2 (l - l )2 +2 l ˘ . (5-47)

ÎÍ ˚˙
Comparings Eqs. (5-33) and (5-47), we find that the argument of the exponent on the
right-hand side of the latter must be zero, or
2
l2 = l - l . (5-48)
Since l1 and l 2 are Gaussian random variables, l1 + l 2 is also a Gaussian random

variable. Moreover, because of homogeneity of atmospheric turbulence,
l1 = l 2 (5-49)
and
l12 = l 22 . (5-50)
Hence,
exp (l1 + l 2 ) = exp ÏÌ

Ó2
1
[ (l + l )
1 2
2
- l1 + l 2
2
]+ l1 + l 2 ¸˝
˛
[
= exp l 2 + l1l 2 - 2 l 2
+2 l ]
= exp ( l l 1 2 - l2 ) , (5-51)
where in the last step we have used Eq. (5-48). Now we define a log-amplitude structure
function l according to
( rr r
) = [l (rr ) - l (rr ¢)]
2
l p - rp¢ p p
= (l1 - l 2 )2
= 2 l 2 - l1l 2 , (5-52)
where isotropy of turbulence is assumed so that the structure function depends on

r r r r
rp - rp¢ and not on rp and rp¢ . In view of Eq. (5-52), Eq. (5-51) may be written
exp (l1 + l 2 ) = exp ÈÍ- l

1
Î 2
( rr - rr ¢ )˘˙˚
p p . (5-53)
Since F1 and F 2 are Gaussian random variables with zero mean values, F1 - F 2 is
also a Gaussian random variable. Hence, following Eq. (5-46), we obtain
[
exp i (F1 - F 2 ) ] = exp ÈÍ-
Î 2
1
( F1 - F 2 ) 2 ˘
˙˚
= exp ÈÍ- F
1
Î 2
( rr - rr ¢ )˘˙˚
p p , (5-54)
where
( rr r
) = [F (rr ) - F (rr ¢ )]
2
F p - rp¢ p p (5-55)
is the phase structure function of turbulence. Substituting Eqs. (5-53) and (5-54) into Eq.
(5-45), we may write
(r ) (r )
PR rp PR* rp¢ = exp ÈÍ- w
1
Î 2
( r
rp - rp¢
r
) ˘˙˚ , (5-56)
where
(r
w rp - rp¢
r
) = ( rr l p
r
- rp¢ ) + ( rr F p
r
- rp¢ ) (5-57)
is called the wave structure function of turbulence.
r ÛÛ P r * r
Ii ( ri ) =
Pex
2
È 1
ÙÙ L rp PL rp¢ exp Í- w
Sex l R ıı 2
Î 2
( ) ( ) ( rr
p
r
)
- rp¢ ˘˙
˚
È 2p i r r
¥ exp Í-
Î lR
rp - rp¢ ( ) ◊ rri ˘˙˚ d rrp d rrp¢ . (5-58)
Because of a Fourier transform relationship between the PSF and the OTF, we identify
r r r
( )
rp - rp¢ l R with a spatial frequency vi . Thus, we let
r r r
rp - rp¢ = l R vi . (5-59)
r
Substituting Eq. (5-59) into Eq. (5-58) and carrying out the integration over r p , we may
write
r r r r r
Ii ( ri ) = Pex Û È 1 ˘
Ù t L ( vi ) exp Í- w (l R vi )˙ exp ( - 2 p i vi ◊ ri ) d vi , (5-60)
ı Î 2 ˚
where
r
t L ( vi ) = Sex
-1
(r ) (r
Ú PL rp PL rp - l R vi d rp
* r
) r
(5-61)
r
is the OTF of the (turbulence-free) optical system and vi = vi . Now, we introduce
normalized quantities
r
I (r ) = Ii ( ri ) I (0) , (5-62a)
r r
r = ri l F , (5-62b)
and
r r
vi = v l F , (5-62c)
where
Pex Sex
I (0) = (5-63)
l2 R 2
is the aberration-free central irradiance for a uniform-amplitude wavefront with a total

power of Pex and 1 l F is the cutoff spatial frequency of the optical system. Here, R is
the radius of curvature of the reference sphere with respect to which the aberration is
defined, and F = R D is the focal ratio of the image-forming light cone. In terms of the
normalized quantities, Eq. (5-60) for the time-averaged PSF may be written
r r r r r
◊
I ( r ) = ( 4 p) Ú t ( v ) exp ( - 2 p i v r ) d v , (5-64)
where
r r
t ( v ) = t L ( v ) exp ÈÍ- w (vD)˘˙
1
(5-65)
Î 2 ˚
is the time-averaged OTF of the overall system. We note that this OTF is equal to the
product of the OTF of the optical system and a reduction factor associated with
atmospheric turbulence, just as the pupil function of the overall system is equal to the
product of the pupil function of the optical system and the complex amplitude variation
introduced by turbulence. Indeed, the former is a consequence of the latter. We also note
that, since vD = vi l R , the modification of the system OTF due to turbulence is
independent of the diameter D of the exit pupil of the system. The exponential factor on
the right-hand side of Eq. (5-65) is sometimes referred to as the MTF associated with
atmospheric turbulence, or simply the atmospheric MTF. However, this is not correct.
Since the OTF of a system is equal to the autocorrelation of the pupil function, the
atmospheric turbulence, which introduces phase errors in a wave propagating through it,
cannot by itself have an MTF associated with it. The fact that the effect of turbulence
appears as a multiplier (and is therefore separable) implies only that turbulence reduces
the sytem OTF by the exponential factor. Hence, a proper term for this factor is an MTF
reduction factor associated with turbulence.
5.3.2 Kolmogorov Turbulence

r r
The refractive index N ( r ) of the turbulent atmosphere at a point r in space
fluctuates due to fluctuations of its temperature. It can be written as the sum of its mean
r r r
value N ( r ) at a point r and a fluctuating part n( r ) at that point, in the form
r r r
N (r ) = N (r ) + n (r ) . (5-66)
r r r
Whereas N ( r ) ~ 1, n( r ) is only on the order of 10 - 6 . It should be evident that n( r )
r rr
has a mean value n ( r ) = 0 . The structure function n ( r1 , r2 ) of the refractive index
fluctuations represents the mean square value of the difference of refractive index at two
r r
points r1 and r2 , i.e.,
rr r r
n (r1, r2 ) = [n (r1) - n (r2 )] 2
r r r r
= n 2 ( r1 ) + n 2 ( r2 ) - 2 n ( r1 ) n( r2 ) . (5-67)
The turbulent atmosphere consists of packets of air called eddies each with a
characteristic value of its refractive index. However, it is reasonable to assume that the
turbulence is locally statistically homogeneous so that the mean square value of the
refractive index is the same at every point, i.e.,
r r
n 2 ( r1 ) = n 2 ( r2 ) . (5-68)
The homogeneity of turbulence also implies that the correlation function of refractive
index fluctuations,
r r r r
Rn ( r1 , r2 ) = n ( r1 ) n ( r2 ) , (5-69)
r r
depends on the difference r2 - r1 of the position vectors of the two points. Thus, letting
r r r
r = r2 - r1 , (5-70)
we may write the correlation function in the form

r r r r
Rn ( r ) = n ( r1 ) n ( r1 + r ) . (5-71)
r
The power spectral density Qn (k ) of the refractive index fluctuations is given by the 3-D
Fourier transform of their correlation function, i.e.,
r r r r
Qn (k ) = Ú Rn ( r ) exp (2 p i k r ) d 3r , ◊ (5-72)
r
where k is a 3-D vector representing a spatial frequency in units of cycles/m. It gives a
r
measure of the relative abundance of eddies with dimensions k -1 , where k = k . For a
spatial period L of the refractive index fluctuations, k = 1 L . Similarly, we may write
r r r r 3
Rn ( r ) = Ú Qn (k ) exp (- 2 p i k ◊ r ) d k . (5-73)
If in addition to being homogeneous, we assume that the turbulence is statistically

r
isotropic as well, then the correlation function depends only on the distance r = r
between two points, and we may write
Rn (r ) = n (0) n (r ) . (5-74)
For statistically homogeneous and isotropic turbulence, the structure function of the
refractive index fluctuations can be written in terms of their corresponding correlation
function in the form
n (r ) = 2 [ Rn (0) - Rn (r )] . (5-75)
For Kolmogorov turbulence, it is given by
n (r ) = Cn2 r 2 3 , l 0 r L0 , (5-76)
where Cn2 (in units of m -2 3 ) is called the refractive index structure parameter. The
quantities l0 and L0 are called the inner and outer scales of turbulence representing the
smallest and the largest eddies, respectively. Typical values of Cn2 vary from 10 -13 m -2 3
for strong turbulence to 10 -17 m -2 3 for weak turbulence. Values of l0 are on the order
of a few millimeters, and those of L0 vary from 1 m to 100 m. Substituting Eq. (5-73)
into Eq. (5-75), letting l0 Æ 0 and L0 Æ • (and noting that d 3k = k 2 dk sin q dq df,
0 £ q £ p , 0 £ f < 2 p , and carrying out the angular integrations), we obtain
•
Û Ê sin 2 p k r ˆ 2
n (r ) = 8p Ù Qn (k ) Á1 - ˜ k dk . (5-77)
ı Ë 2p k r ¯
0
Noting that
•
Û aÊ sin bx ˆ G ( a) sin ( p a 2)
Ù x Ë1 - bx ¯ dx = - , - 3 < a < -1 , (5-78)
ı b a +1
0
it can be shown that the right-hand side of Eq. (5-77) reduces to the right-hand side of Eq.
(5-76) if
Qn (k ) = 9.69 ¥ 10 -3 Cn2 k -11 3 . (5-79)
If we divide r by 2p in Eq. (5-72) [and, therefore, in Eqs. (5-73) and (5-77)], then the
numerical coefficient in Eq. (5-79) is replaced by 9.69 ¥ 10 -3 (2 p)2 3 or 0.033. Now
k = 2 p L is a wavenumber, similar to the optical wavenumber k, which for an optical
wavelength l is k = 2p l . Similarly, if we divide k by 2p , as was done by Fried,2 then
the coefficient becomes 9.69 ¥ 10 -3 (2 p)11 3 or 8.18.
The fluctuations of refractive index of the turbulent atmosphere introduce

fluctuations in phase of an optical wave propagating through it. The structure function of
phase fluctuations introduced by statistically homogeneous and isotropic turbulence may
be written
r r r r
F (r ) = [F (r1 ) - F (r1 + r )]
2
= 2 [ RF (0) - RF (r )] , (5-80)
where the correlation function RF (r ) of phase fluctuations is given by

r r r
RF (r ) = F( r1 ) F ( r1 + r ) = F(0) F(r ) . (5-81)
The wave structure function for a spherical wave propagating through Kolmogorov
turbulence is given by 2
L
w (r ) = 2.914 k 2 r 5 3 Ú Cn2 ( z ) ( z L)
53
dz , (5-82)
0
where k = 2p l and z varies along the atmospheric path of total length L from a value of
zero at the source (or the object plane) to a value of L at the receiver (or the image plane).
The wave structure function can also be written in the form
w (r ) = 6.88 (r r0 )
53
, (5-83)
where
-3 5
ÏÔ L 2 ¸Ô
Ì Ú C n ( z ) (z L) dz ˝
53
r0 = 0.1847l 1.2
(5-84)
ÓÔ 0 ˛Ô
is a characteristic length of Kolmogorov turbulence called its coherence length or

diameter. We call it Fried's coherence length in honor of his pioneering work in this
area.2-4 We note that r0 varies with the optical wavelength as l 6 5 . If the line of sight
makes an angle q with the zenith, then the path length L and the integral in Eqs. (5-83)
and (5-84) are increased by sec q , or r0 decreases by a factor of (sec q) .
35
Since Cn2 decreases with altitude, which is different for different sites, the integral in
Eq. (5-82) has a higher numerical value when observing a space object from ground
(looking upwards) than when observing a ground object from space (looking
downwards). Correspondingly, r0 is smaller when a satellite is observed from ground
compared to when a ground object is observed from a satellite. Consequently, image
degradation is much smaller when a ground object is observed from space than when a
space object is observed from ground. In the first case, the object is near the region of
turbulence and it is observed from far away. In the second case, the object is away from
the region of turbulence, but it is observed from nearby. The fact that the image quality is
superior in the first case is similar to when an object behind a diffuse shower glass is
observed. One can see some detail in the object when it is in contact with the shower
glass. However, as soon as the object is moved slightly away from the shower glass, it
appears only as a halo, illustrating complete loss of image resolution. This does not,
however, mean that reciprocity of wave propagation does not hold. For example, if the
wavefront errors of a wave from a point source in space propagating downwards are
measured on ground and a conjugate correction is introduced in a beam transmitted
upwards with a deformable mirror, the beam focus in space will be diffraction limited
(neglecting any measurement or correction error), illustrating that the atmosphere
introduces the same wavefront errors whether a beam is propagating up or down through
it.
For plane wave propagation, the factor ( z L) under the integral in Eqs. (5-82) and
53
(5-84) reduces to unity, which may be seen as follows. A plane wave can be thought of as
a spherical wave originating at an infinite distance and traveling through a uniform
medium for which Cn2 = 0 , except for the propagation path through the atmosphere. The
value of z L in the region for which Cn2 π 0 is infinitesimally different from unity. Thus,
for example, starlight propagation in ground-based astronomy can be considered as plane
wave propagation with ( z L) replaced by unity, or a spherical wave propagating an
53
infinite distance to reach the earth's atmosphere with nonzero Cn2 value only near the end
of its path for which z L is negligibly different from unity.
Neglecting the variation of Cn2 for horizontal propagation, we obtain
ÏÔ2.91Cn2 L k 2 r 5 3 Plane Wave (5 – 85)

w (r ) = Ì
ÓÔ(3 8) 2.91Cn L k r
2 2 53
Spherical Wave (5 – 86)
and
( )
Ï1.68 C 2 L k 2 - 0.6
Ô n Plane Wave (5 – 87)
r0 = Ì
( )
- 0.6
Ô3.02 Cn2 L k 2 Spherical Wave (5 – 88)
Ó
( )
In the near field L D2 l , the amplitude variations are negligible and, therefore,
( )
F (r ) = w (r ) . In the far field L D2 l , F (r ) = (1 2) w (r ) . For in-between
ranges, the multiplying factor varies smoothly between 1 and 1/2. The correlation
function RF (r ) of phase fluctuations and the corresponding power spectral density
QF (k ) are related to each other by a 2-D Fourier transform, e.g.,
r r
RF (r ) = Ú ◊
QF (k ) exp ( - 2 p i k r ) d 2 k (5-89a)
= 2 p Ú QF (k ) J 0 (2 p k r ) k d k , (5-89b)
where we have used Eq. (1-81) to carry out the angular integration. Substituting Eq. (5-
89b) into Eq. (5-80), we obtain
•
[
F (r ) = 4 p Ú QF (k ) 1 - J 0 (2 p kr ) k d k . ] (5-90)
0
Noting that
•
Û -a p b a -1
Ù x [1 - J 0 (bx )] dx = a 2 ,1< a < 3 , (5-91)
ı [ ] [
2 G ( a + 1) 2 sin p ( a - 1) 2 ]
0
it can be shown that the right-hand side of Eq. (5-90) reduces to the right-hand side of Eq.
(5-83) if
ÏÔ0.023r0-5 3k -11 13 Near Field, L D2 l , (5 – 92a)

QF (k ) = Ì
ÔÓ(0.023 2)r0 k
-5 3 -11 13
Far Field, L D2 l . (5 – 92 b)
Substituting Eq. (5-83) into Eq. (5-65), we may write the time-averaged OTF of the
overall system
r r
t ( v ; D r0 ) = t L ( v ) t a (v; D r0 ) , (5-93)
where
[
t a (v; D r0 ) = exp - 3.44 (vD r0 )
53
] (5-94)
is the long-exposure atmospheric MTF reduction factor. As stated earlier, t a is

independent of the pupil diameter D, as may be seen by replacing the normalized spatial
frequency v by vi l F and noting that F = R D . The exponent in Eq. (5-94) varies with
wavelength as l-1 3 .
Since exp ( - 3.44) ~ 0.03 , atmospheric turbulence reduces the overall system MTF
corresponding to a spatial frequency v = r0 D by a factor of 0.03. From Eqs. (5-56) and
(5-83), we find that r0 represents a correlation length such that the correlation of complex
amplitudes at two points on a wave separated by a distance r0 is 0.03. Moreover, t a (v)
represents the mutual irradiance function of the wave. Therefore, its magnitude describes
the degree of spatial coherence of the wave, and thus the visibility of fringes formed in a
two-pinhole experiment and observed in the vicinity of a point that is equidistant from the
two pinholes. Note that because of the random nature of atmospheric turbulence, the
time-averaged irradiances at the two pinholes are equal to each other. Hence, r0
represents a spatial coherence length of the wave so that its degree of coherence
corresponding to two points on it separated by r0 is 0.03, or that the visibility of the
fringes formed by the secondary waves from these points is 0.03. We may add that there
is no standard for defining a correlation or coherence length. It is evident, however, that a
3 percent coherence or fringe visibility implies that the wave elements at the two points
separated by r0 are spatially practically incoherent. The value of r0 on a mountain site
may vary from 5 to 10 cm in the visible region of the spectrum. As stated earlier, it
increases with wavelength as l1.2 .
5.3.3 Circular Pupils

We now assume that the turbulence-free system is aberration free with uniform
r
transmittance so that t L ( v ) is given by Eq. (2-44). Hence, using Eq. (2-12), Eq. (5-64)
for the time-averaged PSF of the overall system reduces to
1
I (r; D r0 ) = 8 Ú t (v; D r0 ) J 0 (2 p r v) v d v . (5-95)
0
The time-averaged Strehl ratio of the turbulence-degraded image is given by
S ( D r0 ) = I (0; D r0 )
1
= 8 Ú t (v; D r0 ) v d v . (5-96)
0
It gives the average central irradiance relative to its aberration-free value. Similarly, from
Eqs. (1-89) and (2-21b), it follows that the average fractional encircled poweris given by
1
P (rc ; D r0 ) = 2 p rc Ú t (v; D ro ) J1 (2 p rc v) d v , (5-97)
0
where rc is in units of l F . This result may also be obtained by substituting Eq. (5-95)
into Eq. (2-22b) and using Eq. (1-88). Equations (5-95) through (5-97) are similar to Eqs.
(5-9) through (5-12), respectively. Their differences lie in the difference of the average
OTFs given by Eqs. (5-10) and (5-93).
We now consider the unnormalized time-averaged central irradiance
Pex Sex
Ii (0; D r0 ) = S ( D r0 ) (5-98a)
l2 R 2
Pex Sa
= h ( D r0 ) , (5-98b)
l2 R 2
where Sa = p r02 4 is the coherent area of the atmosphere and
h ( D r0 ) = ( D r0 ) S( D r0 )
2
. (5-98c)
Since Sex = p D2 4, the quantity h is propotional to the central irradiance for a fixed
total power Pex . For example, we may have a ground-based laser transmitter whose
total power is fixed, but whose beam diameter can be selected to change the central
irradiance on a target. From Eqs. (5-93) and (5-96), we may write
21
Ê Dˆ
h ( D r0 ) = 8 Á ˜
Ë r0 ¯ Út
0
L [ ]
(v) exp - 3.44(vD r0 ) v d v ,
53
(5-99)
where t L (v) is the aberration-free OTF of the imaging system given by Eq. (2-44).
For small values of D r0 , the exponential factor is approximately equal to unity.

Hence, noting Eq. (2-45), the integral reduces to 1 8, or the Strehl ratio is approximately
equal to unity. Since Sex depends on D as D2 , the central irradiance, for a fixed total
power Pex , increases as D2 , as for an aberration-free system. For very large values of
D r0 , the contribution to the integral in Eq. (5-99) comes from values of v small enough
that the exponential factor is not vanishingly small. In Eq. (2-44), t L (0) = 1. Moreover,
cos -1 v ~ p 2 for small values of v, and for such values it dominates the second term
( )
12
v 1 - v 2 . Thus, L (v) ~ 1 near the origin. Hence, for large values of D r0 , Eq. (5-99)
may be written
1
( )
h Dr0 = 8 ( D r0 ) Ú exp [- 3.44 (vD r ) ]v d v
2 53
0
0
•
= 8 (3.44) - 6 5 (3 5) Ú x ( ) exp ( - x ) dx
6 5 -1
= 1 , (5-100)
where x = 3.44 (vD r0 ) and the integral over x is the gamma function G (6 5) . Hence,
53
the central irradiance given by Eq. (5-98b) approaches a limiting value of Pex Sa l2 R 2 ,
which is equal to the aberration-free value for a system with an exit pupil of diameter r0 .
Since Sa ~ ro2 ~ l2.4 , the limiting value varies with wavelength as l0.4 .
Figure 5-6a shows how turbulence degrades the system MTF for several values of
D r0 . The aberration (or turbulence)-free MTF (which corresponds to D r0 = 0 ) is also
shown for comparison. For example when D r0 = 2 , not only is the MTF at any
frequency reduced, but the cutoff frequency is practically reduced from a value of 1 to
0.5. Similarly, when D r0 = 5 , the cutoff frequency is practically reduced to 0.2. The
OTF at any frequency reduces as D r0 increases.
Figure 5-6b shows how S varies with D r0 . It decreases to zero monotonically as

D r0 increases. Thus, for example, for a given value of D, the Strehl ratio decreases
rapidly as r0 decreases. Even when r0 is as large as D, the Strehl ratio is only 0.445. (The
values of S corresponding to a few values of D r0 are listed later in Table 5-3.) The
value of r0 varies with the wavelength of object radiation as l1.2 . At a visible
wavelength, its approximate value is only 10 cm. Consequently, the performance of a
ground-based telescope making astronomical observations in the visible region is limited
primarily by atmospheric turbulence. The purpose of a large ground-based telescope has
therefore generally not been better resolution (before the advent of adaptive optics) but to
collect more light so that dim objects may be observed.
Figure 5-6c shows how h varies with D r0 . Its aberration-free or diffraction-limited

( )
value varies as ( D r0 ) , as illustrated by the straight line. For small values of Dr0 , its
2 2
aberrated value also varies as ( D r0 ) . As D increases, h increases much more slowly,

2
and for D r0 > 5 , the increase is negligible. As D r0 Æ •, h Æ 1. The two asymptotes

0.8
D/r0 = 0
0.6
<t> 1
2
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1
v
(a)
1.0
<S>
0.1
0.01
0 2 4 6 8 10
D/r0
(b)
10
(D/r0)2
1
h
(D/r0)2 <S>
0.1
0.01
0.1 1 10
D/r0
(c)
Figure 5-6. Effect of atmospheric turbulence on time-averaged system performance.
(a) MTF. The diffraction-limited case is represented by D r0 = 0 . (b) Strehl ratio. (c)
Central irradiance for a fixed total power represented by h . Its aberration-free
value increases as ( D r0 ) , but its aberrated value approaches unity as D r0 Æ • .
2
( )
of h Dr0 intersect at D r0 = 1 ; indeed Fried3 defined r0 in a way so as to yield this
result. He called the quantity h the normalized resolution.
In astronomical observations, Pex increases as D increases. However, if the

observation is made against a uniform background, then the background irradiance in the
image also increases as D2 . Hence, the detectability of a point object is limited by
turbulence to a value corresponding to an exit pupil of diameter r0 , no matter how large
the diameter D of the exit pupil is. In the case of a laser transmitter with a fixed value of
laser power Pex , the central irradiance on a target will again be limited to its aberration-
free value for an exit pupil of diameter r0 , no matter how large the transmitter diameter
is. Similarly, the ratio of signal and noise powers in an optical heterodyne detection of a
turbulence-degraded signal is limited to the aberration-free value corresponding to an exit
pupil of diameter r0 .
Figure 5-7 shows the average irradiance distribution normalized to unity at the center
for several values of D r0 . The diffraction rings disappear and the PSFs may be
approximated by Gaussian functions. The corresponding encircled power is also shown in
this figure. As D r0 increases, a given fraction of the total power is contained in an
increasingly larger circle. As an example, whereas 84 percent of the total power is
contained in a circle of radius rc = 1.22 when there is no turbulence, it is contained in a
circle of radius 1.9 when D r0 = 1 .
1.0
D/r0 = 0
1
0.8 2
3
, <P(rc)>
4
0.6
5
0.4
0.2
0.0
0 1 2 3 4 5 6
r; rc
Figure 5-7. Time-averaged irradiance and encircled-power distributions for

different values of D r0 .
5.3.4 Annular Pupils5

The treatment of random aberrations in systems with circular exit pupils can be
easily extended to those with annular exit pupils. It can be shown, for example, that for
random aberrations introduced by atmospheric turbulence, the time-averaged irradiance
distribution, Strehl ratio, and encircled power of the image of a point object are given by
1
8
I (r; ; D r0 ) = Ú t (v; ; D r0 ) J 0 (2 p rv) v dv , (5-101)
1 - 2 0
1
8
S ( ; D r0 ) = Ú t (v; ; D r0 ) vdv , (5-102)
1 - 2 0
and
1
P (rc ; ; D r0 ) = 2 p rc Ú t (v; ; D r0 ) J1 (2 p rc v) dv , (5-103)
0
where
[
t (v; ; D r0 ) = t (v; ) exp - 3.44 (vD r0 )
53
] (5-104)
and t (v;) is given by Eq. (3-40). The irradiance is in units of Pex Sex ( ) l2 R 2 .
Figure 5-8 shows how a quantity
(
h ( ; D r0 ) = 1 - 2 ) ( D r ) S (; D r )
0
2
0 (5-105)
( )
varies with D r0 . Since Sex () = p 1 - 2 D2 4 , this quantity is proportional to the
aberrated central irradiance and thus shows its variation with D for a given value of and
a fixed total power Pex . Its aberration-free or diffraction-limited value variyng as
( )
1 - 2 ( D r0 ) is illustrated by the straight lines for several values of . For small values
2
of Dr0 , the atmospheric MTF reduction factor is approximately equal to unity.

Accordingly, S ( ; D r0 ) is also approximately equal to unity, and the aberrated value of
h increases with D r0 as its aberration-free counterpart. However, for larger values of
D r0 , it increasees slowly with a negligible increase beyond a certain value of D r0 ,
depending on the value of . The saturation effects of atmospheric turbulence occur at
larger and larger values of D r0 as increases. Since t (v; ) is approximately equal to
unity near the origin irrespective of the value of ,
h ( ; D r0 ) Æ 1 as D r0 Æ • , (5-106)
as in the case of circular pupils. The two asymptotes of h (; D r0 ) for a given value of
( )
-1 2
intersect at the point D r0 = 1 - 2 . The aberrated central irradiance is given by the
product of its aberration-free value Pex Sex l2 R 2 and the Strehl ratio S ( ; D r0 ) or
10.0
0.25
=0
0.50
0.75
=0
1.0
h(D/r0; )
0.25
0.50
0.75
0.1
0.01
0.1 1.0 10.0 100.0
D/r0
Figure 5-8. Variation of h ( ; D r0 ) with D r0 for several values of . Its aberration-

( )
free value given by 1 - 2 ( D r0 ) is represented by the straight lines. Its aberrated
2
value approaches unity as D r0 Æ • , regardless of the value of .
(P )
Sa l2 R 2 h ( ; D r0 ) . Hence, regardless of how large D is, the central irradiance is
ex
less than or equal to the aberration-free central irradiance for a system with an exit pupil
of diameter r0 , equality approaching as D r0 Æ •. The limiting value of the central
irradiance is independent of the value of .
( )
In astronomical observations, Pex = p 1 - 2 D2 I0 , where I0 is the time-
averaged irradiance across the exit pupil, increases as D increases. However, if the
observation is made against a uniform background, then the background irradiance in the
image also increases as D2 . Hence, as in the case of a circular pupil, the detectability of a
point object is limited by turbulence to a value corresponding to an exit pupil of
diameter r0 , no matter how large the diameter D of the exit pupil is. In the case of a laser
transmitter with a fixed value of laser power Pex , the central irradiance on a target will
again be limited by its aberration-free value for an exit pupil of diameter r0 . Similarly,
the ratio of the signal and noise powers in an optical heterodyne detection of a
turbulence-degraded signal is limited to the aberration-free value corresponding to an exit
pupil of diameter r0 .
Figure 5-9 shows how the difference in Strehl ratios for circular and annular ppupils
D S (; D r0 ) = 10 S (0; D r0 ) - S (; D r0 ) [ ] (5-107)
varies with D r0 . It is evident that the Strehl ratio decreases to zero monotonically as
D r0 increases, irrespective of the value of . However, we note that DS > 0 for
D r0 <
~ 3 , i.e., S decreases faster for annular pupils than for circular pupils. The opposite
is true for D r0 >~ 3 . Some typical values of the Strehl ratio are given in Table 5-3 for
several values of D r0 and . Figure 5-10 shows how the irradiance distribution and
encircled power change as D r0 increases. We note that as D r0 increases, a given
fraction of total power is contained in a circle of increasingly larger radius. Figure 5-11
shows how the encircled power in the Airy disc, i.e., for rc = 1.22 , varies with D r0 . It is
evident from this figure that for a given value of D r0 , the relative loss of power from the
Airy disc decreases as increases.
1.2
1.0
DS (0.75)
0.8
<S (D/r0 ; 0)>
<S (D/r0 ; 0)>, DS (e)
0.6
DS (0.50)
0.4
DS (0.25)
0.2
0.0
–0.2
0 2 4 6 8 10
D/r0
Figure 5-9. Variation of D S ( ; D r0 ) = 10 S (0; D r0 ) - S ( D r0 ) . The Strehl [ ]

ratio for = 0 shown in Figure 5-6 on a log scale is shown here on a linear scale.
Table 5-3. Time-averaged Strehl ratio for various values of and D r0 .
\ D/ r0 1 2 3 4 5
0 0.445 0.175 0.089 0.053 0.035

0.25 0.430 0.169 0.088 0.054 0.036
0.50 0.391 0.160 0.090 0.058 0.040
0.75 0.344 0.152 0.095 0.067 0.050
1.0
= 0.5
0.9 D/r0 = 5
0.8
4
0.7
3
; <P(rc)>
0.6
2
1
0.5 0
0.4
0.3
0.2
0.1
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
r; rc
Figure 5-10. Time-averaged PSF and encircled power for several typical values of
D r0 and = 0.5 .
1.0
0.8 =0
<P(D/r0 ; ) >
0.6
0.25 rc = 1.22
0.4 0.50
0.2 0.75
0.0
0 1 2 3 4 5
D/r0
Figure 5-11. Time-averaged encircled power in the Airy disc as a function of D r0

for several typical values of .
5.3.5 Phase Aberration in Terms of Zernike Circle Polynomials6

In Sections 5.3.3 and 5.3.4, we calculated the degradation of system performance by
propagation through atmospheric turbulence without any knowledge of the magnitude of
the phase aberrations introduced by it. In this section, we expand the Kolmogorov
turbulence-degraded aberration function into a complete set of Zernike polynomials and
determine its time-averaged variance as well as the variance of each component
polynomial. We also show that the Zernike expansion coefficients are not statistically
independent and give the values of the correlation coefficients. we show that the total
variance is infinity due to an infinite value of the piston aberration. Since piston
aberration does not affect the image, the variance of the piston-removed aberration ,
which is finite, is the true aberration variance. The Strehl ratio based on this variance
( )
using an approximate expression S = exp - s 2F gives a very poor estimate of the true
value given by Eq. (5-96). Another simple formula is also given, which gives a much
better estimate.
The phase aberration introduced by atmospheric turbulence can be expanded in terms

of orthonormal Zernike circle polynomials considered in Section 2.4 in the form
F(r, q) = Â aj Zj (r, q) , (5-108)

j
where r = r a and a j are the orthonormal expansion coefficients. The polynomials may
be written
Zeven j (r, q) = 2(n + 1) Rnm (r) cos mq, m π 0 , (5-109a)
Zodd j (r, q) = 2(n + 1) Rnm (r) sin mq, m π 0 , (5-109b)
Z j (r, q) = n + 1 Rn0 (r), m = 0 , (5-109c)
where Rnm (r) are the radial polynomials given by Eq. (2-61b). Polynomials varying as
sin mq are also included here, since the turbulent atmosphere does not have an axis of
rotational symmetry. As stated in Section 2.4, n and m are positive integers (including
zero) and n - m ≥ 0 and even. The index n represents the radial degree or the order of
the polynomial since it represents the highest power of r in the polynomial, and m may
be called the azimuthal frequency. The index j is a polynomial-ordering number and is a
function of n and m. The orthonormal polynomials and the relationship among the indices
j, n, and m are given in Table 5-4. The polynomials are ordered such that an even j
corresponds to a symmetric polynomial varying as cosmq, while an odd j corresponds to
an antisymmetric polynomial varying as sinmq. For a given value of n, a polynomial with
a lower value of m is ordered first. The orthonormality of Zernike polynomials implies
that
1 2p 1 2p
Ú Ú Z j (r, q) Z j ¢ (r, q) r dr dq Ú Ú r dr dq = d jj ¢ . (5-110)
0 0 0 0
The orthonormal expansion coefficients, representing the standard deviation of the

corresponding aberration term (with the exception of a1 ), are given by
a j = p -1 Ú F(r, q) Z j (r, q) r dr dq , (5-111)
as may be seen by substituting Eq. (5-108) into Eq. (5-111) and using Eq. (5-110).
Since the phase aberration is a Gaussian random variable with zero mean, so are the
expansion coefficients. The cross correlation of the expansion coefficients is given by
r r r r r r
a*j a j ¢ ( )
= 1 p 2 Ú d r Ú d r ¢ Z *j ( r ) Z j ¢ ( r ¢) F( r ) F( r ¢)
r r r r r r
( )
= 1 p 2 Ú d r Ú d r ¢ Z *j ( r ) Z j ¢ ( r ¢) RF ( r - r¢ ) , (5-112)
where an asterisk denotes a complex conjugate. It is needed, as we shall see in what

follows, even though the expansion coefficients and the Zernike polynomials are real. For
Kolmogorov turbulence, the Fourier transform QF (k ) of the phase correlation function
RF (r ) is given by Eq. (5-92). Hence, we write Eq. (5-112) in Fourier space. We let
r r r r
S j (k ) be the Fourier transform of Z j ( r ) , where k is now a dimensionless vector like r ;
i.e., we let
r r r r r
S j (k ) = Ú Z j ( r ) exp (2 p i k r ) d r◊ (5-113a)
so that
r r r r r
Z j (r) = Ú S j (k ) exp ( - 2 p i k r ) d k .◊ (5-113b)
It is shown in the appendix that
J n +1 (2 pk )
Seven j (k, f) = 2(n + 1) i m ( -1)( n - m ) 2 cos mf , m π 0 , (5-114a)
k
J n +1 (2 pk )
Sodd j (k, f) = 2(n + 1) i m ( -1)( n - m ) 2 sin mf , m π 0 , (5-114b)
k
and
J n +1 (2 pk )
S j (k , f) = n + 1( -1)n 2 , m=0 , (5-114c)
k
r
where (k, f) are the polar coordinates of k . Substituting Eq. (5-113b) into Eq. (5-112),
we obtain
r r r r r r r r
a*j a j ¢ = 1 p2 ( ) Ú d k S (k ) Ú d k ¢ S (k ¢) Ú d r ¢ exp [2p i (k - k ¢) ◊ r ¢]
*
j j¢
r r r r r r
¥ Ú d r RF ( r - r¢ ) exp [2p i k ◊ (r - r ¢) ] . (5-115)
Table 5-4. Orthonormal Zernike circle polynomials Z j (r, q) . The indices j, n, and m
are called the polynomial number, radial degree, and azimuthal frequency,
respectively. The polynomials Z j are ordered such that an even j corresponds to a
symmetric polynomial varying as cos mqq , while an odd j corresponds to an
antisymmetric polynomial varying as sin mqq. For a given n, a polynomial with a
lower value of m is ordered first.
j n m Z j (r, q)
1 0 0 1
2 1 1 2 r cos q
3 1 1 2 r sin q
4 2 0 (
3 2r 2 - 1 )
5 2 2 6 r sin 2q
2
6 2 2 6 r 2 cos 2 q
7 3 1 (
8 3r3 - 2r sin q )
8 3 1 8 (3r 3
- 2r) cos q
9 3 3 8 r 3 sin 3 q
10 3 3 8 r 3 cos 3 q
11 4 0 (
5 6r 4 - 6r2 + 1 )
12 4 2 (
10 4r 4 - 3r2 cos 2q )
13 4 2 10 ( 4r 4
- 3r ) sin 2q
2
14 4 4 10 r 4 cos 4 q
15 4 4 10 r 4 sin 4 q
16 5 1 ( )
12 10r5 - 12r3 + 3r cos q
17 5 1 12 (10r - 12r + 3r) sin q

5 3
18 5 3 12 (5r - 4r ) cos 3q
5 3
19 5 3 12 (5r - 4r ) sin 3q
5 3
20 5 5 12 r 5 cos 5 q
21 5 5 12 r 5 sin 5 q
Table 5-4. Orthonormal Zernike circle polynomials Z j (r, q) . The indices j, n, and m
are called the polynomial number, radial degree, and azimuthal frequency,
respectively. The polynomials Z j are ordered such that even j corresponds to a
symmetric polynomial varying as cos mqq , while odd j corresponds to an
antisymmetric polynomial varying as sin mqq. For a given n, a polynomial with a
lower value of m is ordered first (continued).
j n m Z j (r, q)
22 6 0 (
7 20r6 - 30r 4 + 12r2 - 1 )
23 6 2 ( )
14 15r - 20r + 6r sin 2q
6 4 2
24 6 2 14 (15r - 20r + 6r ) cos 2q

6 4 2
25 6 4 14 (6r - 5r ) sin 4q
6 4
26 6 4 14 (6r - 5r ) cos 4q
6 4
27 6 6 14 r 6 sin 6 q
28 6 6 14 r 6 cos 6 q
29 7 1 ( )
4 35r7 - 60r5 + 30r3 - 4r sin q
30 7 1 4 (35r - 60r + 30r - 4r) cos q

7 5 3
31 7 3 4 (21r - 30r + 10r ) sin 3q

7 5 3
32 7 3 4 (21r - 30r + 10r ) cos 3q

7 5 3
33 7 5 4 (7r - 6r ) sin 5q
7 5
34 7 5 4 (7r - 6r ) cos 5q
7 5
35 7 7 4 r 7 sin 7 q
36 7 7 4 r 7 cos 7 q
37 8 0 (
3 70r8 - 140r6 + 90r 4 - 20r2 + 1 )
r
The integration over r can be carried out using Eqs. (5-89a) and (5-92). Thus, since
r r
r = r a,
r r r r r r
Ú d r RF ( ) [
r - r ¢ exp 2 p i k ◊ (r - r ¢)]
) Ú d (rr - rr ¢) R ( r r r r
r
(
= 1 a2 F ) [
r - r ¢ exp 2 p i (k a) ◊ (r - r ¢)]
= 0.023 ( a r0 ) k -11 3 ,
53
(5-116)
for the near field, or divide the right hand side by a factor of 2 for the far-field
r r r
propagation. The integration over r ¢ yields d(k - k ¢) . Hence, Eq. (5-115) reduces to
r r r
a*j a j ¢ (
= 0.023 p 2 ) (a r )
0
53
Ú d k k -11 3 S *j (k ) S j ¢ (k ) . (5-117)
r r r
Substituting for S j (k ) and S j ¢ (k ) from Eqs. (5-114) into the integral over k , we obtain
r r r
Ú d k k -11 3 S *j (k ) S j ¢ (k )
=
2
1 + d m0
[ 12
]
(n + 1) (n ¢ + 1) i - m + m ¢ (-1)(n + n ¢ - m - m ¢ ) 2 Inn ¢ Imm ¢ , (5-118)
where7
•
Inn ¢ = Ú dk k
-14 3
J n +1 (2 p k ) J n ¢ +1 (2 p k ) (5-119)
0
=
[
p11 3 G (14 3) G (n + n ¢ - 5 3) 2 ] , n π 0 π n ¢ (5-120)
[ ] [
2 G (n - n ¢ + 17 3) 2 G (n ¢ - n + 17 3) 2 G (n + n ¢ + 23 3) 2 ] [ ]
and
Ïcos mf cos m ¢f , j and j ¢ are both even

Ô
2p Ôcos mf sin m ¢f , j is even and j ¢ is odd
Û Ô
Imm¢ = Ù df Ì
ı Ôsin mf cos m ¢f , j is odd and j ¢ is even
0
Ô
Ô
Ósin mf sin m ¢f , j and j ¢ are both odd
Ïp (1 + d m 0 )d mm ¢ , j and j ¢ are both even

Ô
= Ìp d mm ¢ , j and j ¢ are both odd (5-121)
Ô0
Ó , otherwise .
◊
In Eq. (5-120), G ( ) is the gamma function. Thus, Eq. (5-117) may be written
= 0.1534 ( -1) [(n + 1)(n ¢ + 1) ]1 2 (Dr0 ) d mm ¢

( n + n ¢ - 2m) 2 53
a *j a j ¢
G (14 3) G [(n + n ¢ - 5 3) 2] (5-122)
¥ .
G [(n - n ¢ + 17 3) 2] G [(n ¢ - n + 17 3) 2] G [(n + n ¢ + 23 3) 2]
The autocorrelation, i.e., the mean square value of an expansion coefficient, may be
obtained from Eq. (5-122) by letting j = j ¢ , i.e., by letting n = n ¢ and m = m ¢. Thus,
G ( n - 5 6)
53
Ê Dˆ
a 2j = 0.7587 (n + 1) Á ˜ , n, n ¢ π 0 . (5-123)
G (n + 23 6) Ë r0 ¯
When j and j ¢ are both equal to unity so that both n and n ¢ are equal to zero, Eq. (5-117)
reduces to
•
5 3Û J12 (2 p k )
a12 = (0.046 p) ( a r0 ) Ù dk k
-8 3
. (5-124)
ı k2
0
Since 2 J1 ( x ) x Æ 1 as x Æ 0 , the integral in Eq. (5-124) is equal to infinity. Thus, the

piston autocorrelation is infinity. This is an artifact of Kolmogorov spectrum of
turbulence given by Eq. (5-79) with a pole at k = 0 . The mean square value or the
variance of the phase aberration (since its mean value is zero) introduced by such a
spectrum is also equal to infinity. This may be seen from Eqs. (5-89) and (5-92) (and
noting that we are using a dimensionless k like the variable r ) according to which
F2 = RF (0)
•
= 0.046 p ( a r0 ) -8 3
53
Ú k dk . (5-125)
0
However, the variance of the piston-removed aberration given by
D1 = F 2 - a12 = Â a 2j
j =2
•
= 0.046 p ( a r0 )
Ú {1 - [2 J (2p k ) 2p k] }k
53 2 -8 3
1 dk
0
= 1.0299 ( D r0 )
53
(rad ) 2
(5-126)
is finite. Of course, the piston aberration Z1 (r, q) , being a constant, has no effect on the
quality of an image. Since r0 μ l6 5 , the variance in terms of the optical path-length
errors is independent of l , as expected in the absence of atmospheric dispersion. In
obtaining Eq. (5-126), we have used the relation
•
Û ÏÔ È 2 J1 ( x ) ˘ 2 ¸Ô - p pG ( p + 2)
Ù Ì1 - Í ˙ ˝ x dx = 2 p G 2 ( p + 3) 2 G ( p + 5) 2 G 1 + p 2 sin p( p - 1) 2 .
ı
0
ÔÓ Î x ˚ Ô˛ [ ] [
( ) ] [ ]
(5-127)
Numerical values of cross-correlation and autocorrelation of the expansion coefficients

are given in Tables 5-5 and 5-6. 8 Note that, as is evident from Eqs. (5-122) and (5-123),
for a given value of n, these values do not depend on the value of m. The first nonzero
cross-correlations are a2 a8 and a3 a7 , i.e., tilt-coma cross-correlations. Moreover, for
a given value of n, the correlation values decrease rapidly as the order difference n ¢ - n
increases.
If the first J modes of the phase aberration are corrected, then the mean square
residual error is given by
1 2p 2
Û Û È J ˘
ÍF(r, q) - Â a j Z j (r, q)˙
-1
DJ = p Ù Ù r dr dq
ı ı Î j =1 ˚
0 0
J
= D1 - Â a 2j . (5-128)
j =2
Numerical values of D j are given in Table 5-7. For Kolmogorov turbulence, F 2 = •.

However, this infinity is contained in the piston ( j = 1) mode so that D1 is finite. We
note that the phase aberration with a standard deviation of 1 radian is obtained when
D = r0 without any correction (except for the piston mode). However, if the x and y tilts
are corrected, we find from the expression for D 3 , that D can be as large as 3.34 r0 for
one radian of phase aberration.
The approximate expressions of Eq. (1-204) through (1-206) are not suitable for
calculating the time-averaged Strehl ratio for a random aberration, where D1 ∫ s 2F ,
especially when its value is small. For example, the Strehl ratio given by
S1 ( D r0 ) [
~ exp -1.03( D r0 )5 3 ] (5-129)
is illustrated in Figure 5-12 by the dashed curve, which is quite steep owing to the
( D r0 )5 3 dependence of s 2F . Even for a small value of D r0 = 1 , it gives a Strehl ratio of
0.357, compared to a true value of 0.445. For larger values of D r0 , it underestimates the
Strehl ratio by larger factors. A much better approximation is given by
[ ]
-6 5
S2 (D r0 ) ~ 1 + (D r0 )5 3 , (5-130)
as may be seen from Figure 5-12, where it is illustrated by the dotted curve. It slightly
overestimates the true value.
100
<S>
10– 1
<S>
< S1 > < S2 >
10– 2
2 4 6 8 10
D/r0
Figure 5-12. Variation of time-averaged Strehl ratio S with D r0 . The solid curve
represents the exact value S given by Eq. (5-96), which is also shown in Figure 5-
6b. The dashed curve represents the approximate value S1 given by Eq. (5-129) in
terms of the aberration variance. Similarly, the dotted curve represents the
approximate value S2 given by (5-130).
Table 5-5. Correlation of Zernike polynomial expansion coefficients for near-field

propagation in units of ( D r0 ) .
53
Zernike Order Correlation Value
Correlation Coefficient Pairs n n¢ ajaj¢
a 22 , a 32 1 1 4.49 ¥ 10 - 1
a 42 , a 52 , a 62 2 2 2.32 ¥ 10 - 2
a2 a8 , a3 a7 1 3 - 1.42 ¥ 10 - 2
a 72 , a 82 , a 92 , a 10
2
3 3 6.19 ¥ 10 - 3
a 4 a 11 , a 5 a 13 , a 6 a 12 2 4 - 3.88 ¥ 10 - 3
2
a 11 2
, a 12 2
, a 13 2
, a 14 2
, a 15 4 4 2.45 ¥ 10 - 3
a 7 a 17 , a 8 a 16 , a 9 a 19 , a 10 a 18 3 5 - 1.56 ¥ 10 - 3
2
a 16 2
, a 17 2
, a 18 2
, a 19 2
, a 20 2
, a 21 5 5 1.19 ¥ 10 - 3
a 11 a 22 , a 12 a 24 , a 13 a 23 , a 14 a 26 , a 15 a 25 4 6 - 7.60 ¥ 10 - 4
Table 5-6. Correlation of Zernike polynomial expansion coefficients for near-field

propagation in units of ( D r0 ) for n = 1 and n ¢ ≥ 1 .
53
Order Correlation
Zernike Order Difference Value
Correlation Coefficient Pairs n n¢ n¢ - n ajaj¢
a 22 , a 32 1 1 0 4.49 ¥ 10 - 1
a2 a8 , a3 a7 1 3 2 - 1.42 ¥ 10 - 2
a 2 a 16 , a 3 a 17 1 5 4 7.54 ¥ 10 - 4
a 2 a 30 , a 3 a 29 1 7 6 - 9.52 ¥ 10 - 6
a 2 a 46 , a 3 a 47 1 9 8 8.61 ¥ 10 - 7
a 2 a 68 , a 3 a 67 1 11 10 - 1.41 ¥ 10 - 7
a 2 a 92 , a 3 a 93 1 13 12 3.24 ¥ 10 - 8
a 2 a 122 , a 3 a 121 1 15 14 - 9.29 ¥ 10 - 9
a 2 a 154 , a 3 a 155 1 17 16 3.14 ¥ 10 - 9
Table 5-7. Variance of residual phase errors for near-field propagation in units of
( D r0 )5 3
D1 = 1.0299 D12 = 0.0352
D 2 = 0.582 D13 = 0.0328

D 3 = 0.134 D14 = 0.0304
D 4 = 0.111 D15 = 0.0279
D 5 = 0.0880 D16 = 0.0267
D 6 = 0.0648 D17 = 0.0255
D 7 = 0.0587 D18 = 0.0243
D 8 = 0.0525 D19 = 0.0232
D 9 = 0.0463 D 20 = 0.0220
D10 = 0.0401 D 21 = 0.0208
D11 = 0.0377 D J ª 0.2944 J - 3 2

(For large J )
5.3.6 Short-Exposure Image

Atmospheric turbulence introduces aberrations including wavefront tilt. A short-
exposure image is equivalent to a tilt-free image. If the tilt is corrected in real time with a
steering mirror, then a time-averaged short-exposure image can be obtained. Imaging
through atmospherewhen a wave propagates a path length L that is much smaller than
D2 l is referred to as near-field imaging. Similarly, if L >> D 2 l , then it is referred to
as far-field imaging. In near-field imaging, turbulence primarily introduces aberrations,
and a short exposure image provides a better resolution. We discuss the effect of tilt
correction on the overall system OTF, Strehl ratio, and the central irradiance. We show
that the central irradiance increases as D increases, achieves a maximum value when
D r0 = 3.74 , and then decreases asymptotically to its value for the long-exposure image.
However, in far-field imaging, the improvement is not as significant due to the effect of
scintillations or amplitude variations. Examples of short-exposure PSFs are shown
illustrating the speckled image. We show that the size of a speckle is determined by l D
(similar to the radius of the aberration-free Airy disc), but the overall size of the PSF is
determined by l r0 .
5.3.6.1 Near-Field Imaging

It is evident from Tables 5-6 and 5-7 that a large portion of the phase aberration
introduced by turbulence is a random wavefront tilt represented by the coefficients a2
and a3 . If tilt is corrected, the variance of the phase aberration is reduced by a factor of
1.0299 0.134 ~ 7.7 . Hence, unless D r0 is very large, we should expect a significant
improvement in the quality of an image if wavefront tilt introduced by turbulence is
r r
corrected in (near) real time. Let I ( r ) and t ( v ) be the instantaneous PSF and the
corresponding OTF of the overall system consisting of the imaging system and the
turbulent atmosphere. The effect of a wavefront tilt is to displace the (aberrated) image as
a whole. If at a certain instant in time, a wavefront tilt displaces the image to a position
r r r
rt , then the irradiance distribution of a tilt-corrected image would be given by I ( r - rt ) .
This distribution would be similar to that for a short-exposure image, short in the sense
that the image has not wandered around; the only difference between the two would be
their locations. The exposure time may be 0.1 sec or less.
r r r
The tilt-corrected OTF t t ( v ) of the system and the corresponding PSF I ( r - rt ) are
related to each other by a Fourier transform according to
r r r r r r
◊
t t ( v ) = Ú I ( r - rt ) exp (2 p i v r ) dr
r r r r r r r r r
◊ [
= exp (2 p i v rt ) Ú I ( r - rt ) exp 2 p i v ◊ (r - rt )] d (r - rt )
r r r
◊
= exp (2p i v rt ) t ( v ) . (5-129)

r r r r
◊
t ( v ) = exp ( - 2 p i v rt ) t t ( v ) . (5-130)
This result simply represents the fact that a displacement of the image produces a linear
phase factor in the OTF, i.e., a linearly varying PTF. Taking the time-average of both
sides, we obtain
r r r r
◊
t (v ) = exp ( - 2 p i v rt ) t t (v ) . (5-131)
r r
◊
Assuming that the random variations of exp (2p i v rt ) are uncorrelated with those of
t t (v) , we may write Eq. (5-131)
r r r r
◊
t ( v ) = exp ( -2 p i v rt ) t t ( v )
r r ˘ t ( vr )
= exp ÈÍ- 2 p 2
Î
(v ◊ rt )2 ˙˚ t
r r r
[
= exp - p 2 v 2 rt rt ◊ ] t t (v ) , (5-132)
r
where in the middle step we have used Eq. (5-46) and the fact that rt is a Gaussian
random variable with a mean value of zero.
Now, the x component of wavefront tilt aberration in terms of Zernike polynomials is

given by
F tx (r, q) = 2 a2 r cos q . (5-133)
If b is the wavefront tilt angle, also called the angle of arrival at the imaging system, the
phase aberration can be written according to Eq. (3-22) of Part I (and letting ni = 1) in the
form
F tx (r, q) = (2 p l ) abr cos q , (5-134)
where a is the radius of the circular exit pupil. Comparing Eqs. (5-133) and (5-134), we
find that
b = l a2 p a . (5-135)
The x component of the image displacement is given by
xt = R b
= a2 l R p a
= 2 a2 l F p , (5-136)
where R is the radius of curvature of the reference sphere with respect to which the
aberration is defined and F = R 2 a = R D is the focal ratio of the image-forming light
cone. Similarly, its y component is given by
yt = 2 a3l F p . (5-137)
Hence, we obtain
r r
◊
rt rt = xt2 + yt2
= (2 l F p )
2
a22 + a32
= (2 l F p) ¥ 0.896 ( D r0 )
2 53
(5-138a)
(
= 3.584 p 2 ) ( Dr )0
53
, (5-138b)
where we have substituted a22 = a32 = 0.448( D r0 ) for the near-field propagation
53
r
from Table 5-7, and in the last step we have let rt be in units of l F. Since F = R D , it is
evident from Eq. (5-138a) that the mean square value of the image displacement
decreases with the pupil diameter D as D -1 3 . Substituting for r0 from Eq. (5-84), we
find that the variance of the image wander or the angle of arrival for a plane or a spherical
wave is independent of wavelength, as expected in the absence of any atmospheric
dispersion.
Substituting Eq. (5-138b) into Eq. (5-132), we obtain the tilt-corrected MTF of the
overall system
r
[
t t ( v ) = exp 3.584v 2 ( D r0 )
53
] t (vr) . (5-139)
r
Substituting Eq. (5-93) for t ( v ) , we may write
r r
t t ( v ) = t L ( v ) t ta (v) , (5-140a)
where
t ta (v) = exp { [- 3.44 (vD r ) ] (1 - 1.042v ) }

0
53 13
. (5-140b)
is the time-averaged tilt-corrected or short-exposure MTF reduction factor associated

with atmospheric turbulence. Replacing the normalized spatial frequency v by vi l F and
noting that F = R D , we find that t ta (v) does depend on the pupil diameter D, in
contrast to the long-exposure OTF reduction factor given by Eq. (5-94), which is
r
independent of D. As expected, the tilt-corrected MTF t t ( v ) is larger than the long-
r
exposure MTF t ( v ) . The time-averaged Strehl ratio of the tilt-corrected image is
given by
1
St ( D r0 ) = 8 Ú t t (v; D r0 ) v d v . (5-141)
0
Replacing the long-exposure atmospheric MTF reduction factor in Eq. (5-99) by the
short-exposure OTF reduction factor, the short-exposure h function may be written
2 1
ht ( D r0 )
Ê Dˆ
= 8Á ˜ Û Ù t L (v) exp
Ë r0 ¯ ı
0
{ [- 3.44 (vD r ) ] (1 - 1.042v ) } v d v . (5-142)
0
53 13
Figure 5-13a shows how the tilt-corrected MTF reduction factor t ta (v) varies with
the spatial frequency for a few values of D r0 . The figure illustrates what is evident from
Eq. (5-140b) that t ta (v) > 1 for 0.884 < v < 1, regardless of the value of D r0 , i.e., the
reduction factor is greater than unity in the high frequency range. It has the consequence
r r
that the tilt-corrected MTF t t ( v ) is larger than the turbulence-free MTF t L ( v ) .
This is not possible, since aberrations reduce the MTF at every spatial frequency [see Eq.
(1-107)]. Equation (5-140a) and, in turn, Eq. (5-132) were obtained from Eq. (5-131)
under the assumption that the random variations of wavefront tilt and the tilt-corrected
OTF were uncorrelated. This, in turn, implies that the random wavefront tilt and the tilt-
corrected aberrations introduced by turbulence are uncorrelated. And that, of course, is
not true, as we see from Tables 5-5 and 5-7. The tilt aberration coefficients a2 and a3 are
correlated with other aberration coefficients; correlations a2 a8 , a3 a7 , a2 a16 ,
a3 a17 , etc. are not zero. Of course, the MTF reduction factor given by Eq. (5-140b)
gives an incorrect value at all spatial frequencies within the passband of the system, and
not just the high frequency region of 0.884 < v < 1. Fried 3 circumvented this problem by
replacing the number 1.042 in Eq. (5-140b) by unity (Actually his number was 1.026
owing to the use of a slightly different value of the variance of the wavefront tilt
aberration). The MTF reduction factor thus obtained is £ 1, as illustrated in Figure 5-13b.
Figure 5-13c shows how the tilt-corrected time-averaged Strehl ratio varies with
D r0 . The uncorrected Strehl ratio is also shown to illustrate the improvement made by
( )
tilt correction. An approximate value given by exp - s 2F , where s 2F = 0.134( D r0 ) is
53
the tilt-corrected time-averaged phase variance, and shown by the dashed curve
approximates the true value reasonably well for D r0 £ 6 .
Figure 5-13d shows how ht varies with D r0 . For small values of D r0 , it increases
approximately as for a diffraction-limited system, since the aberration is small. However,
for D r0 > 6 , its value is (incorrectly) greater than the corresponding diffraction-limited
value of ( D r0 ) . If we define htc as the h-function obtained by replacing 1.042 by
2
unity, its value, as shown in Figure 5-13d, is approximately equal to that of ht for small
values of D r0 , reaches a maximum of 3.74 for D r0 = 3.5 , and then decreases slowly
but monotonically to unity. Since the image displacement due to wavefront tilt decreases
as D -1 3 , the effect of tilt correction becomes negligible for large values of D r0 due to
the large residual phase errors. The correct analysis, which takes into account the
correlation of wavefront tilt with the residual aberration, yields ht values that are higher
than those given by htc , i.e., the actual performance is better.8 The aberration function
may also be expanded in terms of the Karhunen-Loève functions whose coefficients are
statistically independent of each other. 8, 9 It is found that the effect of correlation of
Zernike coefficients is negligible for D r0 < ~ 4.
1.2
Forbidden Region
1.0
0.8
D/r0 = 2
< tta >
0.6
0.4
2
5
0.2
0
0 0.2 0.4 0.6 0.8 1
v
0.8
D/r0 = 2
0.6
< tta >
0.4
0.2 5
0
0 0.2 0.4 0.6 0.8 1
v
Figure 5-13. (a) Near-field short-exposure atmospheric MTF reduction factor t ta

given by Eq. (5-140b) and shown by the solid curves, illustrating the error resulting
from ignoring the correlation of the tilt aberration coefficient with others. t ta > 1
is the forbidden region, since it makes the overall system MTF greater than the
corresponding diffraction-limited MTF. (b) MTF reduction factor obtained when
1.042 in Eq. (5-140b) is replaced by unity. The dashed curves represent the long-
exposure or the uncorrected MTFs shown earlier in Figure 5-6a.
100
exp(–s 2F )
<S>
10– 1
< St >
<S>
10– 2
2 4 6 8 10
D/r0
Figure 5-13c. Near-field tilt-corrected time-averaged Strehl ratio St . The result

obtained from Eq. (5-141) is shown by the solid curve and its approximate value is
= 0.134( D r0 ) . The corresponding
53
indicated by the dashed curve, where s F 2
uncorrected Strehl ratio S , shown earlier in Figure 5-6b, is shown here for
comparison.
103
102
ht
101
(D/r0)2
h
h tc
100
h
10–1
10–2 –1
10 100 101 102
D/r0
Figure 5-13d. Variation of near-field h with D r0 for long- and short-exposure

images. h : Long exposure or uncorrected; ht : Short exposure or tilt corrected, but
correlation of coefficients neglected; htc : Short-exposure or tilt-corrected with 1.042
replaced by unity; ( D r0 ) : Diffraction limited.
2
5.3.6.2 Far-Field Imaging

So far, we have assumed near-field wave propagation in the atmosphere, i.e., the path
length L in the atmosphere is small so that L D2 l . Accordingly, the amplitude
variations caused by atmospheric turbulence are negligible, and F (r ) = w (r ) . In
principle, a diffraction-limited image can be obtained if the aberrations are corrected
completely. However, if the path length is long so that L D2 l , then significant
amplitude variations are introduced in addition to the phase variations. The far-field long-
exposure image is identical to the corresponding near-field image, since w (r ) is the
same in both cases. However, in the case of far-field propagation through turbulence,
only half of w (r ) is due to the phase errors, i.e., F (r ) = (1 2) w (r ) and, therefore,
even a complete correction of the phase errors does not yield a diffraction-limited image.
The variance and correlation of the coefficients are accordingly smaller by a factor of 2
compared to the corresponding near-field values. Hence, the tilt-corrected atmospheric
MTF reduction factor in this case is given by
t ta (v) = exp { [- 3.44 (vD r ) ] (1 - 0.521v ) }

0
53 13
.
(5-143)
The corresponding ht -function is given by
2 1
Ê Dˆ
ht ( D r0 ) = 8 Á ˜ Û Ù t L (v) exp
Ë r0 ¯ ı
0
{ [- 3.44 (vD r ) ] (1 - 0.521v ) } v d v
0
53 13
,(5-144)
and the Strehl ratio St by ht ( D r0 ) . The number 0.521 in these equations may be
2
replaced by 0.5 for consistency with replacing 1.042 by 1 in Eqs. (5-140b) and (5-142).
When the phase errors are corrected completely, the atmospheric MTF reduction
factor, the h-function, and the Strehl ratio are given by
[
t a (v) = exp - 1.72 (vD r0 )
53
] , (5-145)
21
Ê Dˆ
hc ( D r0 ) = 8 Á ˜
Ë r0 ¯ Út
0
L [ 53
]
(v) exp - 1.72 (vD r0 ) v d v , (5-146)
and Sc = hc (( D r ) ) , respectively.
0
2
Figure 5-14a shows the aberration-corrected MTF t c = t L t a as a function of

the spatial frequency. It is evident that even with full correction, its value at any
frequency is less than the corresponding diffraction-limited value, and decreases as D r0
increases. This is a consequence of t a being less than unity due to scintillation or
amplitude variations produced by turbulence. Perhaps it is worth remembering that, in
principle, the MTF of a system with nonuniform amplitude across the pupil can be higher
at some frequencies, as in the case of a Gaussian pupil (see Figure 4-4). Figure 5-14b
shows how the tilt-corrected Strehl ratio St , aberration corrected Sc , and no correction
0.8
D/r0 = 0
0.6
< tc >
2
0.4
5
0.2
0
0 0.2 0.4 0.6 0.8 1
v
(a)
100
< Sc >
<S>
10– 1
<S>
< St >
10– 2
2 4 6 8 10
D/r0
(b)
Figure 5-14. (a) Far-field aberration-corrected MTF. (b) Far-field Strehl ratio. S :
Long exposure or uncorrected; St : Short exposure or tilt corrected, Sc :
Aberration corrected.
hc
ht
100
(D/r0)2
h
h
10–1
10–2
10–1 100 101 102
D/r0
Figure 5-14c. Variation of far-field h with D r0 for long- and short-exposure

images. h : Long exposure or uncorrected; ht : Short exposure or tilt corrected; hc :
Aberration corrected; ( D r0 ) : Diffraction limited.
2
S vary with D r0 . It is seen that the improvement even with full aberration correction
is not that significant. Figure 5-14c shows the variation of ht with D r0 . It is evident that
the improvement in this case is also much smaller. The performance improves with full
aberration correction, but falls significantly short of its diffraction-limited value.
It should be noted that a wavefront tilt is represented by r (cos q, sin q) . In our

discussion of the short-exposure image, we have corrected the tilt represented by the
Zernike polynomials R11 (r) cos q and R11 (r)sin q , where R11 (r) = r . However, there are
other tilt terms in the expansion of the aberration function in terms of Zernike
polynomials. These tilts are contained in the Zernike polynomials Rn1 (r) cos q and
Rn1 (r)sin q with n = 3, 5, 7 , etc. representing the balanced coma aberrations. It should
also be emphasized that a wavefront tilt does not correspond to the centroid of the image.
Thus, if the wavefront tilt is corrected in real time with a steering mirror, the time-
averaged image thus obtained will not represent addition of the short exposure images
with the same centroid. This is due to the fact that, neglecting amplitude fluctuations, the
(x, y) coordinates of the centroid of an instantaneous image depend on the sum of the
coefficients of all Zernike polynomials Rn1 (r) cos q and Rn1 (r)sin q , respectively [see Eq.
(2-151)].
5.3.6.3 Short-Exposure Images

Now we consider some short-exposure PSFs as examples of star images. Figure 5-
15a shows a possible aberration introduced by atmospheric turbulence as seen by a
ground-based telescope for D r0 = 10 . On the average, the standard deviation of the
[
instantaneous aberration is 0.134( D r0 ) ]
5 3 1/2
, which for D r0 = 10 is 2.494 radians or
approximately 0.4 l . The interferogram for this aberration is shown in Figure 5-15b.
When 25 l of tilt are added to the aberration, the interferogram appears as shown in
Figure 5-15c. (This aberration is used as an example of a random aberration in Part I; see
Figure 3-11.)
Figure 5-16 shows instantaneous short-exposure PSFs corresponding to different

values of D r0 . In each case, three exposures are shown. In Figure 5-16a, D is kept fixed
while r0 decreases from a value equal to D to D 3 and to D 10, e.g., D = 1m and
r0 = 1 m, 33.3 cm, and 10 cm, respectively. We note that each image is broken up into
small spots called speckles, which is a characteristic of random aberrations. The size of a
speckle is determined by D, its angular radius being approximately equal to l D.The size
of the overall image is determined by r0 , its angular radius being approximately equal to
l r0 (varying as l- 0.2 ) The size of a speckle in this figure is approximately constant, but
the overall image size increases as r0 increases. The image becomes progressively worse
as r0 decreases, showing the effects of what astronomers call seeing.
D/ r0 =10
sw = 0.4l
Aberration
(a)
No tilt 25l tilt
(b) (c)
Figure 5-15. Aberration introduced by atmospheric turbulence corresponding to
D r0 = 10 . (a) Aberration shape. (b) Aberration interferogram. The standard
deviation of the tilt-free aberration introduced by turbulence is 0.4 l . (c)
Interferogram with 25 l of tilt.
D r0 = 1
D r0 = 3
D r0 = 10
Figure 5-16a. Short-exposure PSFs aberrated by atmospheric turbulence. D is kept

fixed and r0 is varied. For example, D = 1 m and r0 = 1 m , 33.3 cm, and 10 cm,
giving D r0 = 1 , 3, and 10. The value of D determines the size of a speckle, while r0
determines the overall size of the image.
In Figure 5-16b, the value of r0 is kept fixed while the value of D increases from a
value equal to r0 to 3 r0 and to 10 r0 , e.g., r0 = 10 cm and D = 10 cm , 30 cm, and 1 m,
respectively. Now the size of a speckle decreases as D increases, but the image size is
approximately constant. Thus, an increase in D does not significantly improve the
resolution of the system (as determined by the overall size of the image). For
convenience, the pictures in Figure 5-16b are shown reduced by a factor of 1.5 compared
to those in Figure 5-16a. Thus, for example, the pictures corresponding to D r0 = 10 in
these two figures are otherwise similar.
D r0 = 1
D r0 = 3
D r0 = 10
Figure 5-16b. Short-exposure PSFs aberrated by atmospheric turbulence. r0 is kept

fixed and D is varied. For example, r0 = 10 cm , and D = 10 cm, 30 cm, and 1 m,
giving D r0 = 1 , 3, and 10. The value of D determines the size of a speckle while r0
determines the size of the overall image. For convenience, the PSFs shown here have
been reduced by a factor of 1.5 compared to those in Figure 5-16a.
5.3.7 Adaptive Optics

The correction of wavefront errors in (near) real time is accomplished by using
adaptive optics.10, 11 In practice, a steering mirror with only three actuators is used to
correct the large x and y wavefront tilts (also called tip and tilt). A deformable mirror is
deformed by actuating an array of actuators attached to it. The signals for the actuators
are determined either by sensing the wavefront errors with a wavefront sensor in a closed
loop to minimize the variance of the residual errors, or the actuators are actuated to
produce Zernike modes (e.g., focus, two modes of astigmatism, two modes of coma, etc.)
iteratively until image sharpness is maximized12-14 (see Problem 1.7). The signals are
independent of the optical wavelength provided atmospheric dispersion is negligible. The

two approaches are referred to as zonal and modal approaches, respectively. The zonal
approach has the advantage that the rate of correction is limited only by the rate at which
the wavefront errors can be sensed and the actuators can be actuated. However, the
amount of light that is used by the wavefront sensor is lost from the image. In the modal
approach, there is no loss of light, but the rate or the bandwidth of correction15-17 can be
slow due to its iterative nature, especially when turbulence is severe and a large number
of modes must be corrected. Moreover, for imaging an (isoplanatic) extended object,
wavefront sensing requires a point source in its vicinity, but the modal approach is
applicable to the extended object itself.
Appendix 443
APPENDIX
Fourier Transform of Zernike Circle Polynomials6
r r
The Fourier transform S j (k ) of a Zernike polynomial Z j ( r ) is given by
r r r r r
S j (k ) = Ú ◊
Z j ( r ) exp (2 p i k r ) d r , (5A-1)
r
where k is a 2-D dimensionless spatial frequency variable as before. Using polar
r r
coordinates (k, f) and (r, q) for k and r , respectively, we may write
r
k = k (cos f, sin f) (5A-2a)
and
r
r = r (cos q, sin q) , (5A-2b)
so that
r r
◊
k r = k r cos (q - f) . (5A-3)
Hence, Eq. (5A-1) may be written

1 2p
Û Û
[
S j (k, f) = Ù Ù Z j (r, q) exp 2 p i kr cos (q - f) r dr dq .
ı ı
] (5A-4)
0 0
Let us consider an even value of j. Substituting Eq. (5-109a) into Eq. (5A-4), we obtain
1 2p
Seven j (k, f) = [ ]
2(n + 1) Ú Rnm (r) r dr Ú exp 2 p i k r cos (q - f) cos mq dq . (5A-5)
0 0
Letting
q-f = a , (5A-6)
the integral over q becomes

2p - f
Û
Ù exp (2 p i k r cos a ) cos m (a + f) da
ı
-f
2p -f
Û
= Ù exp(2 p i k r cos a )(cos ma cos mf - sin ma sin mf) da
ı
-f
2p
Û
= cos mf Ù exp(2 p i k r cos a ) cos ma da
ı
0
= 2 p i m J m (2 p k r) cos mf , (5A-7)
where we have used the fact that the integral over 0 to 2p of an odd function of a is zero
and that18
p
Ú exp(i cos x ) cos nxdx = p i Jn () .
n
(5A-8)
0
Substituting Eq. (5A-7) into Eq. (5A-5), we obtain

1
Û
Seven j (k, f) = 2(n + 1) 2 pi cos mf Ù Rnm (r) J m (2 p k r) r dr
m
ı
0
J n +1 ( 2 p k )
= 2(n + 1) i m ( -1)( n - m ) 2 cos mf , m π 0 , (5A-9)
k
where we have used the identity 19

1
Û m ( n - m ) 2 J n +1 ( v ) .
Ù Rn (r) J m (v r) r dr = ( -1) (5A-10)
ı v
0
Similarly, we can show that
J n +1 (2 p k )
Sodd j (k, f) = 2(n + 1) i m ( -1)( n - m ) 2 sin mf , m π 0 , (5A-11)
k
and
J n +1 (2 p k )
S j (k , f) = n + 1( -1)n 2 , m=0 . (5A-12)
k
References 445
REFERENCES
1. V. N. Mahajan, “Degradation of an image due to Gaussian image motion,” Appl.,
Opt. 17, 3329–3334 (1978).
2. D. Fried, “Limiting resolution looking down through the atmosphere,” J. Opt.

Soc. Am. 56, 1380–1384 (1966).
3. D. Fried, “Optical resolution through a randomly inhomogeneous medium for

very long and very short exposures,” J. Opt. Soc. Am. 56, 1372–1379 (1966); also
“Statistics of a geometric representation of wavefront distortion,” J. Opt. Soc. Am.
55, 1427–1435 (1965); errata, J. Opt. Soc. Am. 56, 410 (1966); “Optical
heterodyne detection of an atmospherically distorted signal wavefront,” Proc.
IEEE 55, 57–67 (1967); and “Probability of getting a lucky short-exposure image
through turbulence,” J. Opt. Soc. Am. 68, 1651-1657 (1978).
4. D. Fried, “Evaluation of r0 for propagation down through the atmosphere,” Appl.

Opt. 13, 2620–2622 (1974); errata 1, Appl. Opt. 14, 2567 (1975); errata 2, Appl.
Opt. 16, 549 (1977).
5. V. N. Mahajan and B. K. C. Lum, “Imaging through atmospheric turbulence with

annular pupils,” Appl. Opt. 20, 3233–3237 (1981).
6. R. J. Noll, “Zernike polynomials and atmospheric turbulence,” J. Opt. Soc. Am.

66, 207–211 (1976).
7. M. Abramowitz and I. A. Stegun, Eds., Handbook of Mathematical Functions

with Formulas, Graphs, and Mathematical Tables, Dover, New York (1970), p.
692.
8. J. Y. Wang, “Optical resolution through a turbulent medium with adaptive phase

compensation,” J. Opt. Soc. Am. 67, 383–390 (1977); “Phase-compensated optical
beam propagation through atmospheric turbulence,” Appl. Opt. 17, 2580–2590
(1978); J. Y. Wang and J. K. Markey, “Modal compensation of atmospheric
turbulence phase distortion,” J. Opt. Soc. Am. 68, 78–87 (1978). There are some
minus signs missing in Tables II (a) and (b) in the last paper.
9. Guang-ming Dai, “Modal compensation of atmospheric turbulence with the use of

Zernike polynomials and Karhunen Loève functions,” J. Opt. Soc. Am. A12,
2182–2193 (1995).
10. J. W. Hardy, Adaptive Optics for Astronomical Telescopes, Oxford, New York,
(1998).
11. R. K. Tyson, Introduction to Adaptive Optics, SPIE Press, Bellingham

Washington (1999).
12. R. A. Muller and Buffington, “Real-time wavefront correction of atmospherically

degraded telescopi images through image sharpening,” J. Opt. Soc. Am. 61, 1200–
1210 (1974).
13. A. Buffington, F. S. Crawford, R. A. Miller, A. J. Schwemin, and R. G. Smits,

“Correction of atmospheric distortion with an image-sharpening telescope,” J.
Opt. Soc. Am. 67, 298–305 (1977).
14. V. N. Mahajan, J. Govignon, and R. J. Morgan, “Adaptive optics without

wavefront sensors,” SPIE Proc. 228, 63–69 (1980).
15. J. Y. Wang, “Effect of finite bandwidth on far-field performance of modal

wavefront-compensative systems” J. Opt. Soc. Am. 69, 819–828 (1977).
16. C. B. Hogge and R. R. Butts, “Frequency spectra for the geometric representation
of wavefront distortions due to atmospheric turbulence” IEEE Trans. Antennas
Propagation AP-24 , 144–154 (1976).
17. D. P. Greenwood, “Bandwidth specification of adaptive optics systems,” J. Opt.

Soc. Am. 67, 390–393 (1977).
18. M. Abramowitz and I. A. Stegun, Eds., Handbook of Mathematical Functions

with Formulas, Graphs, and Mathematical Tables, Dover, New York (1970), p.
492.
19. M. Born and E. Wolf, Principles of Optics, 7th ed., Oxford, New York (1999), p.
910.
Problems 447
PROBLEMS
1. Consider a beam of light of wavelength l = 0.5 mm incident parallel to the optical
(or z) axis of a beam expander made up of two confocal (i.e. common focus)
paraboloidal mirrors. Let the diameters and the focal ratios of the mirrors be
D1 = 10 cm, D2 = 50 cm, and F1 = F2 = 2 . Let the exit pupil of the system be at its
secondary mirror.
(a) Determine the change in mirror spacing required to focus the beam at a distance
of 1 km.
(b) Determine the time-averaged Strehl ratio if M1 moves randomly along the x
and y axes with a standard deviation of 0.5 mm.
(c) If the beam has a coma aberration
W (r, q) = (l 4) r3 cos q, 0 £ r £ 1 ,
then based on the centroid of the aberrated irradiance distribution in the focal plane,
determine the beam line of sight in units of l D2 .
(d) Determine the mirror fabrication (figure) specification for a Strehl ratio of 0.9
for the beam expander.
2. Consider an astronomical observation with a ground-based telescope having a

circular exit pupil of diameter D operating at a wavelength l = 0.5 mm. Let the
atmospheric turbulence be characterized by a coherence diameter r0 = 10 cm.
(a) Determine D such that a long-exposure image is obtained with a Strehl ratio of
0.8 without the use of adaptive optics.
(b) Repeat problem (a) for a short-exposure image.
(c) Now assume that the value of D obtained in problem (b) is used to obtain a
long-exposure image of Strehl ratio 0.8 with the aid of a steering mirror. Specify
the angular steering requirements and compare them with the angular radius of the
Airy disc.
(d) Determine the total wavefront error in units of l if D = r0 . How much of the
error is a wavefront tilt (in units of l D) and how much is a wavefront
deformation (in units of l )? How do the total error and tilt angle scale as D is
increased?
(e) Now consider D = 1m. If a deformable mirror is used to correct the wavefront
deformation, specify the range of actuators needed.
Bibliography
M. Born and E. Wolf, Principles of Optics, 7th ed., Cambridge University Press, New
York (1999).
J. D. Gaskill, Linear Systems, Fourier Transforms, and Optics (Wiley, New York, 1978).
J. W. Goodman, Introduction to Fourier Optics (McGraw-Hill, New York, 1968)
M. V. Klein and T. E. Furtak, Optics, Wiley, New York (1988).
E. H. Linfoot, Recent Advances in Optics, Clarendon, Oxford (1955).
V. N. Mahajan, Aberration Theory Made Simple, SPIE Press, Bellingham, Washington

(1991).
V. N. Mahajan, Optical Imaging and Aberrations, Part I: Ray Geometrical Optics, SPIE
Press, Bellingham, Washington (1998).
V. N. Mahajan, ed., Effects of Aberrations in Optical Imaging, MS 74, SPIE Press,

Bellingham, Washington (1993).
D. J. Schroeder, Astronomical Optics, 2nd ed., Academic Press, New York (2000).
R. R. Shannon, The Art and Science of Optical Design, Cambridge University Press, New
York (1997).
C. S. Williams and O. A. Becklund, Introduction to Optical Transfer Function, Wiley,

New York (1989).
A. Walther, The Ray and Wave Theory of Lenses, Cambridge University Press, New York
(1995).
449
References for additional reading
These references are author’s collection as the editor of Milestone Series 74 entitled
Effects of Aberrations in Optical Imaging, published by SPIE Press in 1993.
Section One: Aberration-Free Systems

1. G. B. Airy, “On the diffraction of an object-glass with circular aperture,” Trans.
Cambridge Philos. Soc. 5, 283-291 (1835).
2. Lord Rayleigh, “On images formed without reflection or refraction,” Philos. Mag.
5, 214–218 (1881)
3. L. Beiser, “Perspective rendering of the field intensity diffracted at a circular

aperture,” Appl. Opt. 5, 869–870 (1966).
4. J. E. Harvey, “Fourier treatment of near-field scalar diffraction theory,” Am. J.

Phys. 47, 974–980 (1979).
5. R. Barakat and A. Houston, “Reciprocity relations between the transfer function

and total illuminance. I” J. Opt. Soc. Am. 53, 1244–1249 (1963).
6. V. N. Mahajan, “Asymptotic behavior of diffraction images,” Canadian J. Phys.

57, 1426–1431 (1979).
7. P. P. Clark, J. W. Howard, and E. R. Freniere, “Asymptotic approximation to the

encircled energy function for arbitrary aperture shapes,” Appl. Opt. 23, 353–357
(1983).
8. I. Ogura, “Asymptotic behavior of the response function of optical systems,” J.

Opt. Soc. Am.. 48, 579–580 (1958).
9. B. Tatian, “Asymptotic expansions for correcting truncation error in transfer-

function calculations,” J. Opt. Soc. Am. 61, 1214–1224 (1971).
10. W. S. Kovach, “Energy distribution in the PSF for an arbitrary passband,” Appl.
Opt. 13, 1769–1771 (1974).
11. H. S. Dhadwal and J. Hantgan, “Generalized point spread function for a

diffraction-limited aberration-free imaging system under polychromatic
illumination,” Opt. Eng. 28, 1237–1240 (1989).
Section Two: Defocused Systems
12. E. Wolf, “Light distribution near focus in an error-free diffraction image,” Proc.
Royal Soc. A 204, 533–548 (1951)
451
452 REFERENCES FOR ADDITIONAL READING
13. J. C. Dainty, “The image of a point for an aberration-free lens with a circular
pupil,” Opt. Comm. 1, 176–178 (1969).
14. R. E. Stephens and L. E. Sutton, “Diffraction image of a point in the focal plane
and several out-of-focus planes,” J. Opt. Soc. Am. 58, 1001–1002 (1968).
15. P. A. Stokseth, “Properties of a defocused optical system,” J. Opt. Soc. Am. 59,
1314–1321 (1969).
16. D. S. Burch, “Fresnel diffraction by a circular aperture,” Am. J. Phys. 53, 255–260
(1985).
17. H. Osterberg and L. W. Smith, “Defocusing images to increase resolution,”

Science 134, 1193–1196 (1961).
18. T. S. McKechnie, “The effect of defocus on the resolution of two points,” Optica
Acta 20, 253–262 (1973).
19. T. S. McKechnie, “The effect of condenser obstruction on the two-point

resolution of a microscope,” Optica Acta 19, 729–737 (1972).
20. D. K. Cook and G. D. Mountain, “The effect of phase angle on the resolution of
two coherently illuminated points,” Optical and Quan. Elec. 10, 179–180 (1978).
21 A. Arimoto, “Intensity distribution of aberration-free diffraction patterns due to

circular apertures in large F-number optical systems,” Optica Acta 23, 245–250
(1976).
22. Y. Li and E. Wolf, “Focal shifts in diffracted converging spherical waves,” Opt.
Comm. 39, 211–215 (1981).
23. Y. Li, “Dependence of the focal shift on Fresnel number and f number,” J. Opt.
Soc. Am.. 72, 770–774 (1982).
24. V. N. Mahajan, “Axial irradiance and optimum focusing of laser beams,” Appl.
Opt. 22, 3042–3053 (1983).
25. H. H. Hopkins, “The frequency response of a defocused optical system,” Proc.

Royal Soc. A231, 91–203 (1955).
26. W. H. Steel, “The defocused image of sinusoidal gratings,” Optica Acta 3, 65–74
(1956).
27. L. Levi and R. H. Austing, “Tables of the modulation transfer function of a

defocused perfect lens,” Appl. Opt. 7, 967–974 (1968).
REFERENCES FOR ADDITIONAL READING 453
Section Three: Strehl and Hopkins Ratios
28. Lord Rayleigh, “Investigations in optics, with special reference to the

spectroscope. Sec. 4: Influence of aberrations,” Philos. Mag. 8, 403–411 (1879).
29. K. Strehl, “Ueber Luftschlieren und Zonenfehler,” Zeitschrift fur

instrumentenkunde 22, 213–217 (1902).
30. A. Maréchal, “Etude des effets combines de la diffraction et des aberrations

geometriques sur l'image d'un point lumineux,” Revue d'Optique 26, 257–277
(1947).
31. H. H. Hopkins, “The use of diffraction-based criteria of image quality in

automatic optical design,” Optica Acta 13, 343–369 (1966).
32. W. H. Steel, “The problem of optical tolerances for systems with absorption,”
Appl. Opt. 8, 2297–2299 (1969).
circular and annular pupils,” J. Opt. Soc. Am. 72, 1258–1266 (1982).
34. V. N. Mahajan, “Strehl ratio for primary aberrations in terms of their aberration
variance,” J. Opt. Soc. Am. 73, 860–861 (1983).
35. G. Martial, “Strehl ratio and aberration balancing,” J. Opt. Soc. Am. A8, 164–170
(1991).
36. W. B. King, “Dependence of the Strehl ratio on the magnitude of the variance of
the wave aberration,” J. Opt. Soc. Am. 58, 655–661 (1968).
37. J. J. H. Wang, “Tolerance conditions for aberrations,” J. Opt. Soc. Am. 62, 598–
599 (1972).
38. H. H. Hopkins, “Geometrical-optical treatment of frequency response,” Proc.

Phys. Soc. B70, 449–470 (1957).
B70, 449–470 (1957).
40. W. B. King, “Correlation between the relative modulation function and the
magnitude of the variance of the wave-aberration difference function,” J. Opt.
Soc. Am.. 59, 285–290 (1969).
41. S. Szapiel, “Hopkins variance formula extended to low relative modulations,”

Optica Acta 33, 981–999 (1986).
Section Four: Aberration Balancing

discussion of the geometrical aberrations,” Physica 10, 679–692 (1943).
43. B. R. A. Nijboer, “The diffraction theory of optical aberrations. Part II:

Diffraction pattern in the presence of small aberrations,” Physica 13, 605–620
(1947).
44. K. Nienhuis and B. R. A. Nijboer, “The diffraction theory of optical aberrations.

Part III: General formulae for small aberrations: experimental verification of the
theoretical results,” Physica 14, 590–608 (1949).
45. B. Tatian, “Aberration balancing in rotationally symmetric lenses,” J. Opt. Soc.

Am. 64, 1083–1091 (1974).
pupils,” J. Opt. Soc. Am. 71, 75–85; 1408 (1981).
pupils,” J. Opt. Soc. Am. A1, 685 (1984).
48. S. Szapiel, “Aberration-balancing technique for radially symmetric amplitude

distributions: a generalization of the Maréchal approach,” J. Opt. Soc. Am. 72,
947–956 (1982).
49. K. Pietraszkiewicz, “Determination of the optimal reference sphere,” J. Opt. Soc.

Am. 69, 1045-1046 (1979).
50. A. Magiera, K. Pietraszkiewicz, “Position of the optimal reference sphere for

apodized optical systems,” Optik 58, 85-91 (1981).
Section Five: Zernike Polynomials
51. S. N. Bezdid'ko, “The use of Zernike polynomials in optics,” Soviet J. Opt. Tech.
41, 425–429 (1974).
52. S. N. Bezdid'ko, “Calculation of the Strehl coefficient and determination of the

best-focus plane in the case of polychromatic light,” Soviet J. Opt. Tech. 42, 514-
516 (1975).
53. S. N. Bezdid'ko, “Determination of the Zernike polynomial expansion coefficients

of the wave aberration,” Soviet J. Opt. Tech 42, 426–427 (1975).
54. S. N. Bezdid'ko, “Numerical method of calculating the Strehl coefficient using

Zernike polynomials,” Soviet J. Opt. Tech. 43, 222–225 (1977).
55. S. N. Bezdid'ko, “Use of orthogonal polynomials in the case of optical systems

with annular pupils,” Opt. Spectroscopy 43, 2000–2003 (1977).
56. G. Conforti, “Zernike aberration coefficients from Seidel and higher-order power-
series coefficients,” Opt. Lett. 8, 407–408 (1983).
57. R. K. Tyson, “Conversion of Zernike aberration coefficients to Seidel and higher-

order power-series aberration coefficients,” Opt. Lett. 7, 262–264 (1982).
58. R. J. Noll, “Zernike polynomials and atmospheric turbulence,” J. Opt. Soc. Am.
66, 207–2111 (976).
59. J. Y. Wang and D. E. Silva, “Wave-front interpretation with Zernike

polynomials,” Appl. Opt. 19, 1510–1518 ( 1980).
Section Six: Aberrated Systems
60. R. Barakat, “Total illumination in a diffraction image containing spherical

aberration,” J. Opt. Soc. Am. 51, 152–167 (1961).
61. R. Barakat and A. Houston, “Diffraction effects of coma,” J. Opt. Soc. Am. 54,
1084–1088 (1964).
62. V. N. Mahajan, “Aberrated point-spread functions for rotationally symmetric

aberrations,” Appl. Opt. 22, 3035–3141 (1983).
63. S. Szapiel, “Aberration-variance-based formula for calculating point-spread

functions: rotationally symmetric aberrations,” Appl. Opt. 25, 244–251 (1986).
A2, 833–846 (1985).
65. H. H. Hopkins, “Image shift, phase distortion and the optical transfer function,”
Optica Acta 31, 345–368 (1984).
Section Seven: Annular Apertures
66. G. B. Airy, “On the diffraction of an annular aperture,” Philos. Mag. 18, 1–10,
132-1331 (841).
67. H. F. A. Tschunko, “Imaging performance of annular apertures,” Appl. Opt. 18,

3770-3774 ( 1974).
68. A. T. Young, “Photometric error analysis. X: Encircled energy (total illuminance)

calculations for annular apertures,” Appl. Opt. 9, 1874–1888 (1970).
69. I. L. Goldberg and A. W. McCulloch, “Annular aperture diffracted energy

distribution for an extended source,” Appl. Opt. 8, 1451–1458 (1969).
70. V. N. Mahajan, “Included power for obscured circular pupils,” Appl. Opt. 17,
964–968 (1978).
71. J. J. Stamnes, H. Heier, and S. Ljunggren, “Encircled energy for systems with
centrally obscured circular pupils,” Appl. Opt. 21, 1628–1633 (1982).
72. E. H. Linfoot and E. Wolf, “Diffraction images in systems with an annular

aperture,” Proc. Phys. Soc. B66, 145–149 ( 1953).
73. T. Asakura and H. Mishina, “Irradiance distribution in the diffraction patterns of

an annular aperture with spherical aberration and coma,” Japanese J. Appl. Phys.
7, 751–758 (1968).
74. E. L. O'Neill, “Transfer function for an annular aperture,” J. Opt. Soc. Am. 46,
285–288 (1956).
Section Eight: Gaussian Beams
75. A. L. Buck, “The radiation pattern of a truncated Gaussian aperture distribution,”

Proc. IEEE 55, 448–450 (1967).
76. J. P. Campbell and L. G. DeShazer, “Near fields of truncated-Gaussian apertures,”

J. Opt. Soc. Am. 59, 1427–1429 (1969).
77. G. O. Olaofe, “Diffraction by Gaussian apertures,” J. Opt. Soc. Am 60, 1654–

1657 (1970).
78. R. G. Schell and G. Tyras, “Irradiance from an aperture with truncated-Gaussian

field distribution,” J. Opt. Soc. Am. 61, 31–35 (1971).
79. V. P. Nayyar and N. K. Verma, “Diffraction by truncated-Gaussian annular

apertures,” J. Opt. [Paris] 9, 307–310 (1978).
80. D. A. Holmes, J. E. Korka, P. V. Avizonis, “Parametric study of apertured

focused Gaussian beams,” Appl. Opt. 11, 565–574 (1972).
81. Y. Li and E. Wolf, “Focal shift in focused truncated Gaussian beams,” Opt.
Comm. 42, 151–156 (1982).
82. K. Tanaka, N. Saga, and K. Hauchi, “Focusing of a Gaussian beam through a

finite aperture lens,” Appl. Opt. 24, 1098–1101 (1985).
83. D. D. Lowenthal, “Maréchal intensity criteria modified for Gaussian beams,”

Appl. Opt. 13, 2126–2133, 2774 (1974).
84. D. D. Lowenthal, “Far-field diffraction patterns for gaussian beams in the

presence of small spherical aberrations,” J. Opt. Soc. Am. 65, 853–855 (1975).
85. R. Herloski, “Strehl ratio for untruncated aberrated Gaussian beams,” J. Opt. Soc.
Am. A2, 1027–1030 (1985).
86. V. N. Mahajan, “Uniform versus Gaussian beams: a comparison of the effects of

diffraction, obscuration, and aberrations,” J. Opt. Soc. Am. A3, 470–485 (1986).
87. S. C. Biswas, J.-E. Villeneuve, “Diffraction of a laser beam by a circular aperture

under the combined effect of three primary aberrations,” Appl. Opt. 25, 2221–
2232 (1986).
Section Nine: Random Aberrations
88. D. L. Fried, “Optical resolution through a randomly inhomogeneous medium for

very long and very short exposures,” J. Opt. Soc. Am. 56, 1372–1379 (1966).
89. D. L. Fried, “Optical heterodyne detection of an atmospherically distorted signal

wave front,” Proc. IEEE 55, 57–67 (1967).
90. J. Y. Wang, “Optical resolution through a turbulent medium with adaptive phase
compensations,” J. Opt. Soc. Am. 67, 383–390 (1977).
91. V. N. Mahajan, B. K. C. Lum, “Imaging through atmospheric turbulence with

annular pupils,” Appl. Opt. 20, 3233–3237 (1981).
92. V. N. Mahajan, “Degradation of an image due to Gaussian motion,” Appl. Opt.

17, 3329–3334 (1978).
Section Ten: Coherent Systems
93. W. H. Steel, “Effects of small aberrations on the images of partially coherent

objects,” J. Opt. Soc. Am. 47, 405–413 (1957).
94. R. Barakat, “Diffraction images of coherently illuminated objects in the presence

of aberrations,” Optica Acta 17, 337–347 (1969).
95. R. Barakat,” Partially coherent imagery in the presence of aberrations,” Optica

Acta 16, 205–223 (1970).
96. J. P. Mills and B. J. Thompson, “Effect of aberrations and apodization on the

performance of coherent optical systems. I. The amplitude impulse response,” J.
Opt. Soc. Am. A3, 694–703 (1986).
97. J. P. Mills and B. J. Thompson, “Effect of aberrations and apodization on the

performance of coherent optical systems. II. Imaging,” J. Opt. Soc. Am. A3, 704–
716 (1986).
98. D. B. Allred and J. P. Mills, “Effect of aberrations and apodization on the

performance of coherent optical systems. 3: The near field,” Appl. Opt. 28, 673–
681 (1989)
asymptotic PSF
Index annular ..........................................266
circular ............................................91
The terms annular, circular, Gaussian,
general............................................. 45
and general refer to a pupil.
atmospheric coherence length ..401, 410
A atmospheric MTF reduction factor
without correction ................. 407, 411
2-D PSFs with tilt correction......................... 432
annular atmospheric turbulence ............389, 401
aberrated ........................... 311-321 autocorrelation
aberration free .......................... 274 general....................................... 29, 75
circular expansion coefficients ..................426
aberrated ............................ 150-156 axial irradiance
aberration free ............................ 86 annular ..........................................308
aberration balancing circular ..................................113, 141
annular .................................. 283, 291 Gaussian
circular ................................... 99, 105 annular ....................................... 374
Gaussian circular ......................................349
annular ...................................... 371 azimuthal frequency ......................... 419
circular ...................................... 344
weakly truncated ...................... 363 B
aberration difference function 29, 39, 59
variance .................................. 60, 185 balanced aberrations
aberration of diffracted wave. 19, 75, 78 annular ..................................283, 291
aberration tolerance ......................... 101 circular ....................................99, 105
annular .......................................... 281 Gaussian
circular ........................................... 97 annular ....................................... 371
Strehl ........................................ 101 circular ......................................344
Hopkins .................................... 182 weakly truncated ....................... 363
Struve........................................ 196 best image........................................... 99
Gaussian bibliography......................................449
annular ...................................... 370
circular ...................................... 343 C
adaptive optics ......................... 414, 441
Airy diffraction pattern ...................... 85 central irradiance
annular ..........................................301
Airy disc .......................................... 135
circular ..........................................117
angle of arrival................................. 431 Gaussian
angular spectrum.................... 6, 78, 240 annular ....................................... 376
annular pupil circular ......................................350
uniform......................... 261, 415, 393
Gaussian ....................................... 366 centroid in terms of
OTF slope........................................50
apodization ................................ 53, 335
wavefront perimeter ........................52
astigmatism ...................................... 173 wavefront slope............................... 51
annular .......................... 283, 285, 286 annular ..........................................322
balanced .................................... 296 circular ..........................................159
PSF ................................... 318, 319 Gaussian ........................................379
tolerance ................................... 290
chief ray ........................................14, 21
circular
balanced .................................... 107 closely-spaced objects ......................245
OTF .......................................... 173 coherence length............................... 410
PSF ................... 146, 147, 154, 155 coherent imaging ........................67, 222
tolerance ........................... 101, 102 coherent area ............................412, 417
459
460 Index
coma encircled power

annular annular ..........................265, 269, 271
primary ..... 283-290, 296, 306, 308, circular ............................................87
320, 321, 324-326 Gaussian
secondary ........................... 327-329 annular ....................................... 369
circular circular ......................................340
primary ....... 99-105, 107, 109, 139, ensquared power
140, 144, 148, 149, 162, 164 annular ..................................265, 271
secondary ... 106, 107, 165, 166-168 circular ............................................88
Gaussian
primary ...... 344, 347, 363-365, 381 evanescent waves................................12
secondary .................................. 382 excluded power
turbulence............................. 438, 441 annular ..........................................266
comparison of circular ............................................90
coherent and incoherent imaging . 240
diffraction and geometrical OTFs .. 44 F
diffraction and geometrical PSFs . 157 far-field distance
contrast reversal ....... 169, 171, 173, 241 annular ..........................................303
correlation of phase fluctuations...... 409 circular ..........................................120
cross correlation............................... 420 Fourier lenses....................................241
cutoff frequency........... 29, 94, 278, 407 Fourier transform
definition ......................................... 74
D obtaining ....................................... 238
Zernike polynomials ..................... 443
dark rings
annular .................................. 267, 270 Fraunhofer approximation ..................11
circular ................................... 88, 135 Fraunhofer diffraction ............9, 22, 121
Gaussian ....................................... 341 Fresnel approximation ..................11, 22
defocus Fresnel diffraction ..................9, 22, 121
aberration ..................................... 112 Fresnel integrals................................127
annular Fresnel number 112, 117, 298, 300, 349,
PSF ........................... 298, 312, 313 375
circular Gaussian ........................................356
OTF .................................. 172, 182 Fresnel’s half-wave zones ................120
PSF ................................... 142, 151
two-point resolution... 248-252, 258 Fried’s coherence length ..................410
defocused Fraunhofer distribution... 121
G
defocused pupil function.................... 21
deformable mirror .................... 411, 441 Gauss quadrature formula ................124
degree of spatial coherence.............. 412 Gaussian amplitude ..........................336
depth of focus Gaussian approximation ........... 127, 134
annular .......................................... 299 encircled power............................. 134
circular ......................................... 114 PSF ................................................134
Strehl ratio ............................127, 134
deterministic aberrations.................. 389 random motion ..............................392
diffraction focus Gaussian beam imaging....................359
annular .......................................... 284 Gaussian Fresnel number ................. 356
circular ......................... 100, 101, 119
Gaussian image ..........12, 14, 31, 33, 71
diffraction limited .............................. 84
Gaussian model ................................134
E Gaussian pupil ..................................336
OTF ............................................... 341
eddies ............................................... 408 PSF ................................................338
edge-spread function.................. 64, 198 Gaussian radius ................................336
Index 461
Gaussian random image motion ...... 393 near-field diffraction pattern ............121
Gaussian random processes ............. 390 nonoptimally balanced aberrations... 103
Gaussian random variables .............. 405 normalized quantities..........................84
Gaussian reference sphere ................. 15 normalized spatial frequency..............94
geometrical PSF........... 50, 72, 157, 160
geometrical OTF ........................ 41, 187 O
geometrical Hopkins ratio................ 184 optical transfer function
annular ..........................................272
H circular
Hankel transform ..... 27, 30, 71, 95, 392 aberration free ............................. 94
Hermitian ..................................... 34, 35 primary aberrations ........... 169-181
general....................................... 27-45
Hölder’s inequality ...................... 34, 55
OTF slope ..................................... 31, 35
Hopkins ratio .... 4, 35, 61, 182, 184-186
circular ............................................95
Huygens’ secondary wavelets.............. 8 annular ..........................................278
Huygens-Fresnel principle............... 3, 8
P
I
phase structure function....................406
image of a disc phase transfer function ....................... 33
coherent ........................................ 234
incoherent..................................... 209 pinhole ..............................213, 235, 257
image motion ............ 390, 391, 393-400 pinhole camera..........................218, 255
impulse response of free space ............ 8 piston autocorrelation ....................... 424
included power .................................. 90 piston-removed aberration ................426
inverse Fourier transform ............ 22, 74 point-spread function of free space ......8
polishing errors................................. 389
K polychromatic OTF ..........................208
polychromatic PSF ........................... 205
Kolmogorov turbulence ... 401, 409, 410
polynomial-ordering number ............419
Karhunen Loève functions .............. 433
power spectral density
phase fluctuations..........................410
L refractive index fluctuations ......... 408
primary aberrations
line of sight
see centroid annular ..................................283, 308
circular ......................98, 99, 142, 169
line-spread function Gaussian
coherent ........................................ 224 annular ............................... 373, 374
general ............................................ 61 circular ..............................344, 347
incoherent.............................. 191-202 weakly truncated ............... 363, 365
log amplitude ................................... 402
pupil
log-amplitude structure function...... 405 annular ..........................................262
long-exposure image........................ 402 circular ............................................83
Gaussian ........................................336
M pupil function......................................17
Maréchal formula ...................... 58, 127 defocused pupil function................. 21
relative pupil function ............... 23, 70
MTF reduction factor .............. 407, 412
tilt corrected ................................. 432
mutual irradiance function... 39, 59, 411 Q
quadratic phase factor....... 9, 18, 68, 240
N
462 Index
circular ....................97, 103, 104, 394

R atmospheric turbulence
annular ....................................... 417
radially symmetric pupil function...... 29
circular ......................................413
random motion......................... 390, 393 Gaussian ........................343, 370, 394
random phase ................................... 402 general............................. 4, 53, 55, 57
Rayleigh criterion of resolution ....... 245 image motion
Rayleigh’s h 4 rule ........................ 102 annular ....................... 397, 398, 399
circular ..............................391, 394
Rayleigh range ......................... 359, 362
Rayleigh-Sommerfeld formula ........ 5, 9
structure functions ............................401
Rayleigh-Sommerfeld theory .............. 3 phase ............................................. 406
reciprocity ........................................ 411 wave ......................................406, 410
refractive index fluctuations ............ 408 Struve ratio ................. 62, 193, 196, 197
refractive index structure parameter 409 symmetry properties of PSF ..... 112, 121
relative pupil function .................. 23, 70 annular ..................................305, 378
rotationally symmetric aberrations . 121, circular ..........................136, 257, 365
124 Gaussian ................................365, 378
S T
scales of turbulence.......................... 409 tilt-corrected OTF............................. 429
Schwarz’s inequality.......................... 34 time-averaged encircled power ........391
secondary coma time-averaged OTF ..........397, 407, 411
annular .......................................... 327 time-averaged PSF....................391, 412
circular .......... 105, 164, 166-168, 258 time-averaged Strehl ratio ........391, 412
Gaussian ....................................... 382 transfer function of free space ............12
secondary spherical aberration 105, 107, transmission factor........................13, 83
124
two-point resolution..........................245
seeing ............................................... 439
shift-invariant imaging U
coherent .......................................... 67
incoherent....................................... 22 uniformly illuminated pupil..............188
short-exposure image............... 401, 430
near-field imaging ........................ 430 V
far-field imaging .......................... 436
variance
short-exposure PSFs ........ 439, 440, 441
aberration along an axis ................194
spatial coherence length .................. 412 aberration difference function....... 60,
spatial filtering ................. 238, 241, 242 185, 186
spatial frequency ................................ 27 aberration function
spatial-frequency spectrum ................ 28 annular ............................... 283, 295
speckle ............................. 401, 430, 439 circular ..........................57, 98, 108
spherical aberration... 10, 104, 107, 139, Gaussian ....................................346
173 turbulence ..................................427
standard deviation W
primary aberration
annular ...................................... 284 waist..................................................356
circular ........................................ 99 waist magnification ..........................362
image motion................................ 390 wave aberration ..................................16
Strehl ratio wave structure function ............406, 410
annular .. 281, 282, 288-290, 307, 398, wavefront ............................................15
399
Index 463
weakly truncated Gaussian beam ... 353,

362, 363
Z
Zernike expansion coefficients ........ 322
annular .................................. 291, 371
circular ......................... 105, 419, 344
Gaussian ............................... 344, 371
Zernike annular polynomials .. 291, 293-
295, 330
Zernike circle polynomials ..... 105, 107,
419, 421, 422
Fourier transform ......................... 443
Zernike-Gauss annular polynomial.. 371
Zernike-Gauss circle polynomials ... 344
ABOUT THE AUTHOR
Virendra N. Mahajan was born in Vihari, Pakistan,

and educated in India and the United States. He
received his Ph.D. degree in optical sciences from
the College of Optical Sciences, University of
Arizona. He spent nine years at the Charles Stark
Draper Laboratory in Cambridge, Massachusetts,
where he worked on space optical systems. Since
1983, he has been at The Aerospace Corporation in
El Segundo, California, where he is a distinguished
scientist working on space-based surveillance
systems. The Optical Imaging and Aberrations
textbooks evolved out of a graduate course he taught
as an adjunct professor in the Electrical Engineering-
Electrophysics department at the University of Southern California. Dr. Mahajan
is an adjunct professor in the College of Optical Sciences at the University of
Arizona and the Department of Optics and Photonics at the National Central
University in Taiwan, where he teaches graduate courses on imaging and
aberrations. He also teaches short courses on aberrations at meetings of the
Optical Society of America (OSA) and SPIE. He has published numerous papers
on diffraction, aberrations, adaptive optics, and acousto-optics. He is a fellow of
OSA, SPIE, and the Optical Society of India. He is an associate editor of OSA’s
3rd edition of the Handbook of Optics, and a recipient of SPIE’s Conrady award.
He has served as a topical editor of Optics Letters, chairman of OSA’s
Astronomical, Aeronautical, and Space Optics technical group, and a member of
several committees of both OSA and SPIE. Dr. Mahajan is the author Aberration
Theory Made Simple (1991), editor of Selected Papers on Effects of Aberrations
in Optical Imaging (1994), and author of Optical Imaging and Aberrations, Part
I: Ray Geometrical Optics (1998), and Part II: Wave Diffraction Optics (2001),
all published by SPIE Press.

Virendra N Mahajan - Optical Imaging and Aberrations PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Virendra N Mahajan - Optical Imaging and Aberrations PDF

Uploaded by

Copyright:

Available Formats

Virendra N.

The Aerospace Corporation

Bellingham, Washington USA

Copyright © 2001 Society of Photo-Optical Instrumentation Engineers

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

Printed in the United States of America.

mother Shrimati Sushila Devi

and father Shri Ram Chand

Mahajan includes a detailed discussion of the effect of imaging with a Gaussian

The completeness of the discussions regarding the calculation of image structure is

Tucson, Arizona R. R. Shannon

PART II. WAVE DIFFRACTION OPTICS

CHAPTER 1: IMAGE FORMATION ............................................................

CHAPTER 2: OPTICAL SYSTEMS WITH CIRCULAR PUPILS ...................... 79

CHAPTER 3: OPTICAL SYSTEMS WITH ANNULAR PUPILS ..................... 259

CHAPTER 4: OPTICAL SYSTEMS WITH GAUSSIAN PUPILS.................... 333

CHAPTER 5: RANDOM ABERRATIONS................................................. 387

BIBLIOGRAPHY ......................................................................................................................... 449

In Chapter 1, the diffraction theory of image formation is discussed. Starting with a

I am grateful to Dr. P. Mouroulis and Professor Emeritus D. Schroeder for their

El Segundo, California Virendra N. Mahajan

a pupil radius r0 atmospheric coherence diameter

Kalidasa Kumarasambhava 1.3

1.1 Introduction ..............................................................................................................3

1.12 Shift-Invariant Imaging of a Coherent Object ....................................................67

In this chapter, we first describe the diffraction theory of image formation of

Finally, imaging of a coherently illuminated object is discussed. The complex

1.2 RAYLEIGH-SOMMERFELD THEORY OF DIFFRACTION AND THE

Theorem 1. As a wave propagates, its disturbance according to the Huygens-

Theorem 2. Under certain approximations, the propagation of a wave is described

1.2.1 Rayleigh-Sommerfeld Formula

where 2 is the 3D Laplace operator, i.e.,

Solving Eq. (1-9), we find that

showing how the amplitude of a plane wave component changes as it propagates. We

Substituting Eq. (1-13) into Eq. (1-11), we obtain

Substituting Eq. (1-18) into Eq. (1-15), we obtain

r 1  r 1  z exp (ikl ) drr  ,

where (see Figure 1-1)

1.2.2 Fresnel and Fraunhofer Approximations

for z >> krmax

A better approximation of the Rayleigh-Sommerfeld diffraction integral applicable

1.2.3 Transfer Function of Free Space

The free-space transfer function, namely, exp (ik  z ) may be written

1.3 GAUSSIAN IMAGE

Object Optical Gaussian

Comparing Eqs. (1-32) and (1-33), we obtain

Substituting Eqs. (1-22) into Eq. (1-30), we obtain

1.4 DIFFRACTION IMAGE

Theorem 4. The diffraction PSF of an imaging system is proportional to the modulus

Theorem 5. The diffraction image of an isoplanatic incoherent object is equal to the

1.4.1 Pupil Function

Figure 1-3. Elements of diffraction imaging. The reference sphere S of radius of

1.4.2 Diffracted Wave

Since the quantity l

 ri2 rg2  1 r  r zi r  1  1 1  2 1  ri4 rg 

1 2  ri2 rg2  1  r 2 rg2 

Aberration Axial Value Value for field angle

Defocus /8 /8

Astigmatism ( 8zg ) ( 8F)

Field Curvature ( 16zg ) (3 16) 2

Distortion ( 2 D)2 ( 4)

However, we have demonstrated that such an approximation is quite valid in describing

1.4.3 Incoherent PSF and Shift-Invariant Imaging of an Incoherent Object

CHAPTER 1: IMAGE FORMATION ............................................................

r 1 r 1 z exp (ikl ) drr ,

The free-space transfer function, namely, exp (ik z ) may be written

ri2 rg2 1 r r zi r 1 1 1 2 1 ri4 rg

1 2 ri2 rg2 1 r 2 rg2

1 2 2(U , V ) 2 2(U , V ) 2 2(U , V )

[Re 0 (0, 9)] (n) = [( n