fundamentals for

Infrared Systems
心聞局個簡翩翩
Optical Design

l
』翩翩翩翩翩
究闖 闖闖
翩翩翩翩翩翩翩翩
研臨剛剛”
3
峙肘酬翩翩翩
6
胸
4耐 MZmmm闌
m珊可
4－
技酬酬M
悔也
RmMm翩
m翩m
工翩翩翩
金串帥那叫刷品苔，于將此自卻
如呵。否則，一切合付巾
。～付相印人來叫
-
、
 Tutorial Texts Series
• Optical Design Fundamental.＜戶r Infrared System>, Second Edition. Ma> J. R』edl Vol. Tf48
• Resolution Enhancement Techniques in Optical /Jthograpl扒九 Alfred Kwok-Kit Wong Vol TfH
國 Copper /111erconnect Technology Cluistoph Steinbriichel and B叮叮 L Chin Vol. TI46
• Optical Design for Visual Systems. Bruce H. Walk盯 Vol. TI唱
• Fundamentals of Contamination Comrol. Al 叩 C Tribble. Vol TI抖
• Evolutimwr.< Computation.- Principles and Practice for Signal Pro何ssing Da叫d Fogel. Vol. TI41
• Infrared Optics and Zoom Lens.凹 Allen Mann. Vol. Tf42
• Introduction to Adapth·c Optics. 訣。•bert K. Tyson. Vol. Tf·ll
• Fractal ond Wavelet Image Compression Tec/miques. Stephen Weist阻d. Vol. Tf40
• Analysis of Sampled b即ging Systems. R.H. Vollmerhau甜n and R. G. Driggers Vol. Tf39
• Tise位.e Opti•·s.- Light 俊叫tering Me的ι.•lrn’，d !nstmments for Mnlir付I Diogno.•i.•. Vale可 Tuchin Vol.
Tf38
• Fu岫me附叫e Elect叫司，tica para /11ge11附vs. Glenn D. Boreman tran血ted by Jarier Alda Vol. 甘37
Infrared D叫ign £rumple丸 Wdliam L. Wolfe、 Vol. Tf36
• Se叫or and Data Fu.rim> Concepts and AppIi< ations, Second Edition L. A. Kl凹n Vol. Tf35
• Practical Applicatfrms of Infrared Tltenna/ Sensing and lma.iing Equipm叫t. Second Edition ‘ Herben
k呻恤 Vol. Tf34
Fundomentals of Machine Vision Harley R. MyleιVol. Tf33
• Design and Mm.肌肉，tg of Prisms mtd Stn<tll Mirrors in Optical ln.rtmmems. Paul R. Yoder‘ Jr.. Vol. Tf32
• Basic Electro Optics 戶r Ele口rim/ Engin呵呵， Glenn D. Boremnn Vol Tf31
• Optical Engineering Fundm肘，ttal民 Bruce H 、Walker. Vol. T口。
• !nm叫uctimt 帥 Radiometry, William L. Wolfe Vol. Tf29
。 /Jtltographr Proces.r Colllral. Har可 J Levinson Vol. Tf28
• An bttrrxluctioo to /me中rerarion of Grapltic lmage.r, Sergey Abla甜味。， Vol. Tf'7
• Tltenttal Inf'ared Clwnrcteri::titiott of Ground Targets and BackRrounds. P. Jacobs Vol.π26

• imroducrion ta Imaging •pectrometer.r Willian1 L. Wolfe‘ Vol. TT25

• /111roduaion to Infra阿d System Design Wtlli』am L Wolfe Vol. Tf24
• /mradur 恤，n to Computa·bascd Imaging Sy.rtem.r D. Sinha. E. R Dougheny. Vol. Tf'3
• Optical Communication Receiver De>i伊. Stephen B. Alexander. Vol. Tf'2
。 Mounting Lenses in Oprica/ In.rtnanellls. Paul R. Yoder Jr 、 Vol. Tf21
• Optical Design Futtdamentol.r戶r Infrared Sr.rtems. M阻 j Riedl Vol. Tf10
• .An /111nxlucrioo to Real-Time Imaging. Edw.,-d R. Doughenv Phillip A. Laplante Vol. TI 19
• Introducti。”的 \Va阿'front Senso,,, Joseph M. G閏月， Vol. Tf18
• Integration of Lams and Fiber Optics into Robotic Snt凹的 J. A. Marszalec. E. A. Mors叫且 Vol. TTl7
• An /111rnductirm to N。’，linear Image Processing. E. R Dougheny J. A.'1ola. Vol. Tf16
• lmroducrim’的 Optical Testing. Jo時ph M. Gearv Vol. Tfl5
• Image Fomwtion in Low-Voltage Scmming Electron Micro.rcopv. L. Reimer Vol. Tf!2
• Dia:onophtltoqninone hosed Re.rist.r. Ralph Dammel. Vol. Tfl l
• Infrared Window and Dome Materials Dan時l C. Harris Vol. TflO
• An lntmduction to Morplwlogicol Image Processing, Edward R Dougher恥 Vol. Tf9

• An Introduction to Optics in Computers, Henri II. Arsenault Yunlong Sheng Vol. TIS
• Digital bmtge Compression 盯"eclm呵ues. Majid Rabb個i. Paul W. Jon回 Vol. TI守
• Abmation 訂2叩門 Made Simple, Vi 間ndJ可B N. Mahajan Vol. Tf6

• Single-frequencr Semiconductor Lusa乳 Jens Buus Vol. Tf5

• An lntroducrioo to Biological aod A'1ijkial Neuntl Nerwm·ks for Parten> 爪，cognition Steven K. Rogers.
Matthew Kabrisky Vol. TI4
包 u>Sa Beam Propagation i’, the Anno.旬，here Hugo Weichel. Vol. Tf3
• /nfraccd Fiber Optics Paul Klocek. George II. Sigel. Jr. Vo!. Tfl
• Spectra/Iv Seleι ti"e Su吶ices for Heming and cυoling Applications. C. G. Gr:mq叫st. Vol. Tfl
 Optical Design
fundamentals for
Infrared Systems

Max J. Riedl

Tutorial Texts in Optical Engineering

Volume TT是8

Arthur R. Weeks,} r., Series Editor

Invivo Research Inc. and University of Central Florida

霎
SPIE PRESS
A Publicat岫n of SPIE叫一The International Socieiy [o, Optical Engmecnng
Bellinglir間， W且liington USA
Library of Congress Cataloging-in-Publication Data

Riedl. Mo凹 J
Optic叫 design fundamentals for infrared sys阻n" IM阻 J. Riedl.一 2"' ed.
p. cm. 一－ (Tutorial 阻xt• in optical engin間ring: v. TT48)
Includes b1bliogr叩hlcal references and index.
ISBN 0-8194-40訓，5
I. Optical instrum叩ts呵。esign and cons甘uction 2. Infrared equipment Desi伊 and
construction l Title JI Seri目

TS514.R53 2001
621 36'2一吐c21 前l-068758
CIP

Published by

SPIE-The International Society for Optical Engineering

P.O. Box IO
Bellingham. Washington 98227 0010 USA
Phone: 11) 360 676 3290
F日 (I) 360 叫7 1445
Email: spie 申 spie org
WWW: www spie.org

Copyright© 200 I 1l1e Society of Photo-Optical Instrumentation Engineers

All rights reserved No p缸t of 曲的 publication may be reproduced or distribute吐

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

Printed in the Uni1ed S扭扭S of America.

 Introduction to the Series
The Tutorial Texts senes was imtiated in 1989 as a way to make the ma扭rial presented in
SPIE short courses available to those who couldn't attend and to provide a reference book
for those who could Typically, sho11 course notes a間 developed with the thought m
mind that supporting matenal will be p間sented verbally to complement the notes. which
are generally written in summary form, highlight key technical topics, and are not
intended as stand-alone documents. Additional妙， the figures, tables, and other
graphically fonnatted information included with the notes require further explanation
given in the instructor's lecture As stand-alone documents, short course nuks du not
generally serve the student or reader well

Many of the Tutonal Texts have thus started as short course notes subsequently expanded
111to books The goal of the senes is to provide readers with books that cover focused
technical interest areas m a tutorial fasluon What separates the books in this series from
other technical monographs and 阻xtbooks ts the way in which the material is presented.
Keepmg m mind the tutorial natu甜 of the series, many of the topics presented in these
texts 訂e followed by detailed examples that further explain the concepts presented. Many
pictures and illustrations are mcluded with each text，朋d where appropnate tabular
reference data are also included.

To date. the 阻xts published in this series IL.we encompassed a wide range of topics司白了om
geometrical optics to optical detectors to image processing. Each proposal 1s evaluated to
detennine the relevance of the propose<i topic. This initial 間叫ewing process has been
very helpful lo autho1 s in idenufyi嗯， early in the writing process. the need for additional
material or other changes m approach that serve to strengthen the text. Once a manuscript
is completed. it is peer reviewed to ensure dtat chapters commurucate accurately the
essential ingredients of the processes and technolog1es under discussion.

During the past nine years. my predecessor, Donald C O'Shea, has done an excellent job
in building the Tutorial Texts series‘ which now numbers nearly forty books It has
expanded to include not only texts developed by short course instructors but also those
written by other topic experts. It is my goal to m阻ntain the style and quality of books in
the series. and to further expand the topic 盯·eas to include emerging as well as mature
subjects in optics 、 photom口， and m祖ging

Arthur R. W'"ks, Jr.

Im'il'o f1,0search Inc. and U11i1'ei 到＇rv of Ccntrai F/01 ida
Once more ‘ to Hermu時， Renee, Jim, Bryan, and Stephanie.
CONTENTS
Prefa 臼…. . . .....…........“.....……... " .....…..也＂＂＂‘白..…口
ll悶。γica/ Remarks .... 巴 ι 川、辛辛 色色 ，巴巴‘ ......白色色“…“．．雪白苟，…….. XVH

Chapter 1. Radiometric Considerations

l l Introduct10n . . . . . .司，心. . . .. . ..叫“.......。也....... l

1.2 Basic Optical Relations. .. .. . ..也.. ..... .. . ...... I
l3 Signal-to-Noise Rat10 .. 吟吟..“也 . . . . . ... ’ . . .. 2
I4 Extended Simphfied Radiometric Performance Equation. .. .. ...。”， 4
I5 Thermal Radiation Laws . ... ... ... . . . . . ........ .. . . . . ..... 4
I5 I Blackbody radiation and Planck's law ．。.. . . . .... 5
I 5.2 The Stefan-Boltzmann law. . . . .........................’”,... 6
1.5.3 W1en's displacement law . . . . . . .. . ‘ ’ . . . .. ..... . .. 7
口，4 Kirchhoffs law and emissivity ..... ….............… －．也凹，’。”，… 7
I .6 Transmission through the Allllosphe間a ”．．．．．．．‘...﹒.. . ' . 8
1.6. l Penuanent constituents of 也-y atmospl時間 e .......…‘ ' ..卒， .9
1.6.2 Variable constituents .. ...... . ... . ... .... ...... .9
1.6 3 Approximation (assumption).. . ...…..............‘ 9
I 6‘ 4 Precipitable wat盯（definition ）甸，＇ . . ... . .. . JO
l 6.5 日um』dtty . . . . .. . ........‘. . . ....’、. I l
1.6.6 p肥cipitable water (calculation) .. .. .. ...“ ll
1.6 7 Atmospheric transmission (calculauon). . ...........’。”，…旬， il
1.6.8 Computer models ……舟已島也．。’． ．。….................... . 13
1.7 Typical IR Detectors ‘已司令已 e 巴 巴， ............。 14
I7I Thermal detectors . . . . . . ...........’。.... ..... ..... . .’, 14
I. 7 l Photon or quantum detectors ～．。－司，. . .. 14
1.7.1 Photoconductive detectors. 魁。‘.. ... . . . .. ............. 14
l.7.4 Specific detectivity and noise equivalent bandwidth ..‘.....一 15
l.7.5 Detector configurations ’,… e •• • • ...’,...... . 16
References . . . . .。“－－－心. . ' . . . . . . .. ' ...... 17

Chapter 2. Basic Optics

l I Introduct10n. . . . ..... .. . . . . .........”....。........ 0 戶砂 • • 19

22 Snell's Law 阻d the Pnsm .. . . . . . .. .............“……。”也 20
2.3 The Transition frnm a Prism to a Lens... . .. . . . . . .. ... .. 20
2.4 Image Formation . . . - - - ‘ . . . . . . . . .... ' 21
25 Ob.1ect-Image Rclauons . ’ , . ......”,........”,...... . . . 23
2.6 Stops. Pupils, and Windows.......... .. ...“...... ..... ... .’。. .. 24
2.7 Throughput.. ﹒－恥...........…. ... . ... . .. ... . 26
28 Energy Transfer ........ ,’,…. . . . . . . . . .... ' . . . 28
2 8. I Signal-to-noise calculations..... . . ‘ . .. ... . .. 28

lX
 2.9 Differenual Changes................ . . . . . . . . . . . . . . . . . .。 29
2.10 Optical Gain. . . . . . . . . . . . . . . . . . ......同. . . .......... 30
2.10. l Immersion lenses .......-...,-…........ 30
2 l 0 2 Light pipes ...吟。. . . ...........‘..卒，.... 33
2 10.3 Field lens . . . .................. . . . .................. 34
2.11 Field of View for Staring Arrays ..司也 •.• • • • • . •. .’, . 34 w

References . . 也心 也. . . . . . . . ........‘ ’,..e 35

Chapter 3. Primary Aberrations

3. l lntroduct10n . . . . . . .‘﹒’．．．．。’”，......... . . . . 37
3.2 Pnmary Abe叮allons ..........……·······…－－’，．．．“’”’”，－．．。 37
3 2 1 Spherical aberration .....。－．．．－．．．．“ e • 37
3.2 2 Coma ...’．．．．．……e ﹒…...﹒盼...............”… 38
323 Astigmal!sm . . .........-----……“. .... . . 39
3.2.4 Field curvature ...….....………........’.…”,.. . . . . 40
325 Distortion . . . . . . . . . . .....................色色。－－ .. 41
3.2.6 Axial chromat1c aberrat10n ...............”……-’...’,”,.......... 41
3.2.7 Lateral chromatic aberration ....…......... ……·············… 42 e

3.3 Calculanons of Primary Aberrauons …..”... . . . . . . . . . .. ..…....-….... .42

3.3.1 Spherical aberration . . .......’,. . . . . . . ... .. . .. . .44
3 3.2 Coma. . . . . .. . ....’,……………. . . . . . . .46
3 3.3 Asl!gmallsm. 也……………－－－．．’，. . ........“-.… 47
3.3 4 Field curvature . .. . . . .’，．．．…．．吟，可. . . . . . . . 47
3.3.5 Astigmatism and field curvature combined. ... .. . . .. . ..... 4耳
3.3 6 Axial chromatic abe叮ation ....…．．．．．．．．。…’，…．．… .49
3 3.7 ~、lumencal ex剖nple..... ... ..... . . . ...……’, .50
3 -I General Aberrat10n Correcuon Methods... . . . . ... . . .. . . . ‘ 53
35 Doublets.. . . . . . . . . . . ... . ... ......... . . . . . .. . .也’，. . .53
3 5。 i Two elements, same matenal . . . . . .. ... . .... 哥哥 哥 53
3.5.2 Two elem目us. different materials .......’。。”。旬，。.... 5 月
3.5.3 The achromat. .................…”. . . . . . . .56
3.6 Two Thin Air-spaced Elements... . . ... . . . . .’，.............也可.. 57
3.6 I The Petzval obiecuve ....…’,. .... . . . ..........”，。......... 57
3.6.2 Refractive beam expanders ….............. . .....................… 59
363 Telephotos. ….. . .. . . . . . . . ...... . 62
3.7 Reflective Optics .,...…’,. ... .. .. . .’。..’..... 62
3.7.l The spherical mirror p ’。............‘....平﹒益， 63
372 The Mangin m盯or. ... . . . ... ..... .. . .. 65
3 7.3 Classical two-mirror configurations. . . .... …............ －．．．叮叮 67
3. 7.4 The two-sphere Cassegrain system .........…..……. . ....... 68
3.7.5 The two-sphere G阻耳。可 system ... ............ .. . 71
3.7.6 Schwarzschild, a very special case ...… …. ........ E . ..... 71
3.7.7 Reflective beam expanders ’。......。......…. .... .. ........ 73
3. 8 Diffracl10n Linut. ................‘…。”。.................................‘ 74
3。9 Resolution of Imagmg Systems. . . . . . . .....................。 75
References ........，也 戶已司， . . ....... ．．。屯，．．．…。 76

文
Chapter 4. Wave Aberrations

4 l Introducllon. ....................... ．。 …. .. .,., .....－司令 77 ••

42 Diverging and Converging Waves ...。.............“ 77

43 Opttcal Path Len評h OPL. ......‘ .....................ι ”，........... 78
4.4 Optical Path Di 晶叮ence OPD (Wave Front Abeπation). 。 ........ e 戶也 78
45 Sphencal Abe吭吭ion ...函，．”“ .........’ 。“.... . . ....。” . 78
4 5.1 Nurnencal ex 血nple. . ..’,.. 0 ......♂.... 79
".. •• •

4.5 2 Best focus pos1llon. ..... . . . . ............. ... . ......... . . .......“ 81

4.6 Third-Order Spherical Aberrauon . ... ’ , . - ’ , . .. . . .. . . . . . . 82
47 Depth of Focus ’的’，... . .. . .. . . .平. . . 82
References …. . .. . .. . .’。.... '.. . .吟。….. 82

Chapter 5圖 Special Optical Surfaces and Components

5. I I ntroductlon 旬，．．’，........心。．．。－．．”“. . . ......... 83

52 The Plane-Parallel Plate . . .. . . . . . .. . . ’ ‘ ’ . . . . 83
52I Displacements . ...叫“台．．．．‘......... . ..函，“……。心，. .83
522 Optical micrometer . . . ... . ... . . . . . ......舟。。也已 .....‘ 85
5.2 3 Aberration contrib叫 ions. . . . .... . . .. . .. . .. 86
5.2 4 Application remarks .. .... ... . . . ... . ... . ... .. .. 87
5.2 5 The wedge (thin prisn甘心. . . .. .. . . .. .... . ... 9。
另 1 Domes 心’，也 d 刊，－心. . .. . .. . . 91
5 -f The Ball Lens ’..… h • • • • • •• .’。 93•

5.4. I Spherical abe叮ation e 島 也甸， • • ... 94 • .....

5.4 2 An aspherized ball lens . .. . .. ..…’。’。.. 94

5.5 Gradient Index Lεns. ... ... •••n ’, 也恥 , “ …. 。 95
5.6 Conic Sections and General Aspheres. . . .. . ..一. 97
5.6.1 Mathern;itical expressions.ι －．．。. .. .. ... .. .. 98
5.6 2 Rellectors with conic section surfaces ’”…......色合..... 99
5.6 3 Lenses 吶th conic section surfaces ....... 也E V 吟吟， !00
5.6.4 Connnon two仆niπor configurations using conic section
surfaces …....... . .....苟.....。‘.. a …............ . . . . !03
5.6 5 General aspheres (surfaces of rotation ），，也. .•..•• •. e 合！ OJ
5.6.6 Two conic secllon nurrors v明白 an aspheric corrector ..‘心 一 104
5.6 7 Three-mirror configurations . . . .. . . . . .. ..句，。 105
方7 Diffractive (Bin山γ ） Optics. . . ............. . .. .. ..“...... ＇＂。 107
5.7.l The simple diffrnctive singlet .. .. ..... . ... .. .. .. !07
5 7.2 The hybnd achromat …... ...。 …………”。
u 心平－ !09
5 7.3 Nurnencal exampl帥，…. . . . . ..戶。..…ιe 。一也 111
5 ‘ 7.4 Diffraction efficiency... .. .. . . . .. .. ... .. ... .. .. .. .. 114
5.7.5 "Useful" spectral bandwidth . . . ” , . . ...... .... .. 115
5 7.6 Diffraction efficiency for a particular order ...…。“······ .. 115
5.7.7 The hybrid achrornat. corrected for chromatic and spherical
aberrations ．．．． ψ.. 白. • l 16
• .... • ....

5.7 8 Bmary optics 甸甸 計 ，血，－ 。 ”.....缸” .. . . 117

References .....心... . . . e ………－－心。 a ••• .. • . . • 119

Xl
Chapter 6. Design Examples

6I In甘oducuon ..‘' . . .......................’,..’,......... .121

6.2 Basic Assumptions for the High-and Low Temperature Applications. 121
62 l Optics for !ugh-temperature system (3一5µm）。。－－．．。. !22
6 2.2 Optics for low間temperatu間 system 制的一 12 µm) ............... 122
6.2 3 Optics for low-temperature system #2 的－ 12 但m) ........... 123
6.2 4 Optics for low-temperature system #3 的一12 阻n ），…．．． 123
6.3 The Improved Petzval Objective . . ...﹒－．．．可.........﹒－．．．．’，句.... 125
6 3.1 Numerical example for a LWIR application ...‘............. 126
6.3.2 Manufacturing remarks ．…．．．．．．．．．．。. . 128
64 Instantaneous Field of View .............…．．．．．墟，。．．．．．－．．。 128
References . . . . . ............。...’. ... ....... . ..…….......‘...…e !29

Chapter 7. Thermal Effects

7 I Introduction. . . . . . . ...’,...…. . . . . . . .......... !31

72 Changing P缸ameters……．．．．．．．．．。……. .......... .. ... l3l
73 Defocus 明白 Change of Temperature .... …...,-… -……... ．也 132
7.4 Defocus of Singlet ......…－．．．．。.............也. . .... 132
7.5 Athermalization with a Doublet.............. . .................. .. 111
76 The Athe口nalized Achromat.. ..”…. . . . . ....... . . . .....…﹒ ...φ ’， 135
7 .6 I The all-refractive athem祖i achromal . - - ’ , . . . 136
7 .6.2 The hybrid athermal achromal －．．．’”。旬， p e 心............ 137
77 Cold Stop and Cold Shield. . . . . . .... .. . . .. . . .’,.. 138
77 I Cold stop ....包，．．－．．．．．也....﹒. . 138
7.7 2 Cold shield. .. .. .. ... .. .. ....... ................舟刊 138
References. 。. . . . .. .. . . . . .. ‘ . . .
e !39

Chapter 8. Optical Coatings

8l Inuoducuon ..﹒司，＇ ".. . .. . .. 141

8.2 Effects at a Single Surface . . . . . ............... .. .. .. . .. 141
8.3 Two Plane-Parallel Surfaces －－也’，叫 “‘....... .. . .. . 142
8.4 Antireflection Coatings . ... 司令。．， . ．舟旬，也 "'' . 。 143
85 Reflective Coatings. 也e ............’，平時 e .. 145
8.6 Typical Interference Filters ............…－－．．．‘．．．．可 P 心，．．’，...... 146
8.6.1 Angular sensitivity of filters ...…’，．令， e 148
8 6.2 Thermal sensitivity of filters . … 。－ 心信 149
References 。’”也. . . ... . . . . . .. .字， ψ.... . .. 150

Chapter 9. Image Evaluation

9 l Introduction . . . .. . . . .,…“. . . . ... . ... . 15!

9.2 Blur Spot Measurements ..... .. .... .…..‘-…........... 151
9 2.1 Circular mask... . . . .....…. . ...”。.. . . . . . . 15!
9 2.2 Sin . 司令 . . ... . ... 一－－ .. ’, 152
923 Krufe Edge. …….........同．．．．．．叭，’，－．．．。 152
9.3 Energy Distribution. ‘ . .. . .... .... . ......’。. ...... ........ ......... 154

別3
 9.4 Modulation Transfer Function ．．。….......................‘。 ψ154
9A I Overview . . . . ..
w . ...... "'"…… 155
9.4.2 Contrast and resolving pov.er . ...... . . . ...........…．．叫“ 157
9.4.3 Diffraction MTF. .. ... ....。’， 4 ‘. 160
9.4.4 Geometnc MTF .... ..... .. .. . . .. .. . ... . . ... 16!
9.4 5 Numencal example ...... e 秒，. . . . ..........旬，．“心， 161
References . . . .. ...... .. . . . ．．”，甸甸 ... ’ ... a . . 163

Chapter 10. Diamond Turning

10 1 Introduct10n ....‘…..令，…‘ e 平。甸， 165

10 2 Overview . . . . . . ... . .. . ... ....卒， u 165
JO 3 Surface Fm1sh . . .. e 心 ‘. . .e . ...甸甸 戶， 已 166
10 4 Scattering ..’,. . . . . . .......ψ．．．－’，. . . ..........卒，’，. . . ........ 168
I 0.5 Shape Conection ，‘－－．．’。’．一...心. . . . 169
10.6 Optical Surface Testing. . ........ ... .. . .... . . ... ... ...... 169
10.6 I Surface roughness. . ...... ... . ...平 司也 心。... 169 .....

I 0.6 2 Surface shape －司， φ. ....... • • ..….

• • 17 l
•

JO 7 Maclmung Time . . . . . . . . ....…...’,.-…….. . 171

10 8 Further Progress and Developments... .. ..。…。…．．．．益，－．．．．．．….... 172
References .. .. .. ... . .﹒司令－’，“．．益， ... ... A 平， 172

Appendix

A. I Paraxial Ray Tracing. 也苛 ，令 ........ ψ • 心 。 173 …

A. I.I Surface equat10ns. . .……. . . . ......φ …….. . 173
A 1.2 Power equatmns { ray tracing through thin lenses) .. . .. 175
A.2 Spherical Aberration of a Thin Lens. . ’ . . .. ......心心，。恥、 。 今 176
A 2.1 Derivation of expression ....白，’，－’，’， 176
A立之 Blur spot stze 才沙”，......。”，司令 ’ ，心 。”“ .. . . [78
A.3 Optical and Thermal Data for Some Infrared Materials. .. .....”“ 178
A.3.1 Selected materials for the 3-5-!lDl spectral band 辛辛， 179 •• •

A可 2 Selected materials for the 8-12-µm spectral band .......... 179

References . . . . . .. .... . . . . . ...’。 18!

Index . . . .... . ... .... . . . ...… 181

xiii
PREFACE TO THE SECOND EDITION
The many positive remarks about the first edition and listening to the feedback from
the students over the p阻t five years encouraged me to expand upon the origmal matenai
in this second edition

To that end. the subject of beam expanders was mod1fied‘ and achromats have been
covered in more detail. In a chapter on Special Optical Surfaces and Components, the ball
iens ‘ gradient opucs有 and three-nurror conliguratoons have been added.

New chapters are Wave Aberrations. Thermal Effec恆、 Design Examples. and
Dian1ond Turning. In Wave Aberrations besides the con臼pt. comparison of spherical
aberration with the Seidel coefficient is discussed. The chapter Thennal Effects deals with
methods of designmg atbermats, leuses that compensate for the unde如ed results ca阻ed
by temperature excursions. The chapter Design Example is an application-b晶edsumm缸γ
of the su吋eels covered in the earlier chapters of the book. Because di目nond 抽ming is
especially suitable for producmg 阻；pheres 祖d d1ffractive mfrared elements司 a chapter has
been added to descnbe tl世s manufacturing metl10d

To follow the style of the first edition. the added material con臨別 practical
approaches with approximations aml many numerical ex缸nples

I thank Rick Hermann and Sharon Streams of SPIE for their support and editorial
assist聞自 I also thank my friend and colleague Robert E. l'ischer for rev甜wing this
second edition His suggestions have been most valuable

iii,瓜、 l
Riedl
FebmmT ]001

PREFACE TO THE Fil主ST EDITION

Tlus tutorial is intended to provide a basic approach lo the optical design of infrared
systems. It 1s written for systems engineers whose expertise is outside 血e field of optics.
The material presented can be apphed directly to the 血1tial optical system layout phase to
evaluate trade-offs of vanous configurations It will also be very help ul in conveying
reqmre1uents and 目pcctations to an experienced lens designer.

Over the past decades‘ much emphasis has been placed on the use of computers in
lens design. Powerfol programs developed for the lens designer have ma<le it possible to
explore new and different approaches for findmg better solutions to optical design
challenges Unfortunately. the process of usmg C淘mputers to perform the required
calculations is often refe叮ed to as a11tomat1c lens d由1gn. But to obtain a sensible optical
system‘ one th剖 is practi囚l to manufacture and meets cost-related and other specml
demands司 the starting configuration must have a chance to meet those demands.

xv
Optimization programs only succeed in finding the best available solution from local
conditions. 祖祖refore司 it is important to start in the right neighborhood. 1 o identify this
ne1ghborhood-a promising starting point-IS one of the goals of this tutorial text
Designing lenses for the infrared region is 凶 some ways easier than working in the
V間ible spec討um. since the wavelengths 前e longer. the index of refraction of most lens
matenals is higher. 血d therr relative dispersion is Im間r Tins generally results in smaller
pnmary aberr現ions. Tltlrd-order abe甘ation calculations 缸e often sufficient to predict
me扭扭gful perform叩ce expectations even if the system is simplified to a set of thin
lenses. (lenses 叫th zero thickness!. The fact that the diffraction limit is IO to 20 times
l盯ger in the mfrared th叩 m the visible regrnn adds to the usefulness of applying thtrd-
order abe甘ation theory.

Most of the matenals 間ed for infr田ed lenses and mirrors lend themselves to single-
pomt diamond turning For that 間ason. asphenc and di街active surfaces are routmely
employed since they are no more difficult to generate by 臨時 method than spherical
surfaces 血d offer to the correction of aberrations. As~heric and diffractive surfaces 訂e
discussed in detail 血d are also treated m several numerical 閉目nples

Tlrroughout this tutorial text. much emph扭扭 h阻 been placed on the practical 品pect
of the mater祖l ~resented. This is reflected by the many approximations that yield useful
answers especially welcomed d叮切g the proposal stage when tin扭曲d resources are
閏uallv sc訂ce.

T!us book 阻 organized to follow the flow of radiant energy from the source lo the
detector. This flow is expressed wi血 a simplified radiome制c equation whose components
identify the m句or contributors to overall system performance.

While a great deal of material is covered. det剖Is had to be limited. It is hoped that
what is presented will be of value not only in the predesign stages of infrared systems. but
also as a sltmulus lo dig deeper into the ex1stmg litera如何 of this exciting field.

The material presented is a collection of notes from early m-house engineering

seminars that led eventually to a more fom祖Uy structured p間sentation 品 p訂t of the SPIE
short-course program Much of what I am sharmg m these pages 1s based on my long-time
professional 叩d personal association with Warren J. Smith‘ who along with Donald C.
O’Shea reviewed the 田anusc討pt. 個d Lo、;veil L. Baskins. My special thanks go to the
reviewers for their substanl!al help. thetr comments and suggestions greatly improved the
quality of the textι

MaxJ Riedl
October 1995

、vi
IDSTORICAL REl\生ARKS
It was 200 years ago that Sir William Herschel, the Royal Astronomer to King
George III. made his famous discovery. He u油1ted to find some protection for his eyes
while looking mto the sim. In his expenments he noticed an increased response 阻 he
scanned the thermometer from the blue end of the spectrum toward the red This was not
ne叫 11 had been done before However, when he moved the thennometer mto the dark
portion beyond the red. where his eyes could not perceive any light, he noticed that the
heat effect increased Thal is the regwn we now call the i11什ared 1egion.

Over the years『 the thennometer was replaced by other detectors 1 In 1829‘ the first
thermopile was introduced by Nobili. which was improved fo閉只師 later by Melloni.
Durmg the 1880s. several more sensitive detectors were developed. Most notable was the
L個gley bolometer, which was about 30 times more sensitive than Melloni's them10pile.
In 1917司 Case developed the fi時t photoconductive detector.‘ using tballous sulfide

An interestmg discovery occurred dunng the 1930s at 血e Institute of Physics at

Berlin, when Kutzscher experimented with lead sulfide crγstals he had found in Italy.
Ace沾自Itally he noticed that these crystals we間 infrared sensitive ' This discovery led
eventually to an mlrared search and tracking device

In 1952‘ the U.S Anny built the first scanning thermal imagers, which were called
themwgrnphs.3 With the development of cooled short”time-constant 阻dimn antimonrdc
(lnSb) and mercury doped genna世間n (Ge:Hg) photodetectors in the late 1950s. the first
t血t framing sensors appeared, and in 1956 the first long wavelength FLIR W晶 built at the
University of Chicago.

Now‘ the technology has adv血iced to a point where thennal imagers and other
infrared dev臨的 have found their place 111 many apphcations other than military functions
These applications range from medical句 border control. safety and secunty.
telec01mnm1icatron, forensic investigations‘個1d many more

References

I. R. Hudson, bifrared System E1ψnee’·fng. John Wiley (1969 ）司 p. 6.

2. A. R. Vogler. TheOdysseyof11Sc1entis1‘ California State University‘
Fullerton ( 1986).
3. J. M. Lloyd、 Thermal Imaging System.氏 Plenum Press (1975). p. 5.

xvii
 Optical Design
fundamentals for
Infra『ed Systems
CHAPTERl

Radiometric Considerations
1.1 Introduction

This first part begins with identi秒ing the basic elements that make up an optical
system: so凹·ce or target. aperture stop, field stop司 image pl胡e‘自由'3nce and exit pupils司
en甘甜ce and exit 咐ndows. These elements 阻d ne何ssary radiometnc def1mt10ns are
discussed and applied in analyzing the extended simplified radiometric perfom間1ce
equation 但q. (1.7 )]. This eq間lion is structured 個問lude the major co甜ibutions to the
detected signal in an optical system the thermal radiation laws that describe the
transmission through the atmosphere. the optics of the system. 血d finally the response of
the detector.

These concepts are discussed m sequence‘ except the third one, which will be
addressed last and in more detail. because it embraces the major subject of this tutorial ‘
optics.

1.2 Basic Optical Relations

An axial ray travels from the axial object point through the lens and on to the image
plane The 阻 rnl ray that passes through the edge of the aperture stop 1s called the
marginal axial ray. The p討ncipal ray‘ also referred to as the chief ray‘ is an obhque ray
trom an off-axis object point through the center of the aperture stop TI1e margmal
principal ray begms at the edge of the covered object and travels through the center of the
aperture stop and the edge of the field stop. As will be discussed in Chapter3 ‘ these two
marginal rays are the only ones needed lo calculate the pn血ar弘 thi則，order abeπatio醋， II
and 2年 are the angles formed by the 阻ial ray and the principal ray 間lative to the optical
axis. This is indicated 凶 Fig. 1.1. The aperture stop is tl1e physical ope位ng in tl1e optical
system that limits tl1e size of the axial energy cone from 也e object. The image of the
aperture st叩組出e object space is the en個nce pupil, and the image of tl1e apertu間 stop
in the image space is the exit pupil. In lens systems. 也e object space is to the left of the
first lens surface. The image space is to the right of the last lens surface The opening that
limits angle 的 of the p討ncipal ray is the field stop. Its m祖ge in object space is ihe
entrance window h1 lhe image space, U1e field stop image 1s the 間It 、社ndow In Fig I. L
D is the size of the aperture stop and d’的 the size of the field stop.

Figure J.l also indica阻s that an ob.ject located a distance s to the left of tl1e lens is
imaged at di別扭ice s ’ to the right of the lens
2 OPTICAL DESIGN FUNDAM即7ALS FOR 卸FRARED S目TEMS

In m阻y c晶es. the object (t苗get) can be considered to be located at inf血吵， The
image is then formed in the focal plane of the lens. The axial ray angle u will be zero and
the image distance s' becomes f the focal length. TI1e ratio of 血e focal length and the
en恤nce pupil diameter is called the relative ape甜m orfn叫nber (f的

En甘胡ceWindow En叮叮ice Pupil Exit Pupil Exit Window

EW EP=EP ’ EW
Tar盲目 Lens Detector
(Sou凹e) (Objective) (Image Pl闊的

.. v-uz
aaD
..
d可 UA

d d'

司，

... 司，

Marginal M由-ginal
Optical Axis
Principal Ray Axial Ray
s s’
‘ ’,..

Apertu間 Stop
’,

Field Stop

FIG. I.I B 品ic relations of imaging optics.

l.3 Signal －切－Noise Ratio

There are several means of evaluating the perfonnance of a complete system司 such
as MRT' (minimum resolvable temperature) and MDT (minimum detectable
temperature) However‘ because the objective of this tutorial is to concentrate primarily
on the optical portion of an infrared system咱 we limit ourselves to just one meast叮t the
signal“ to-noise ratio ( S川）

In its simplest form. the signal-noise-ratio 祖 stated by

SIN 一- ＿！＿
NEP ‘
(I I)

where P is the collected radiant power in watts that is received by the de甜ctor. NEP
represents 也e noise equivalent power. a measure of the muum回n signal 血at yields a
tmity signal-t。”nmse ratio

The power can be calculated from

W ET
p 一

pz一

-N W
fL一
F

ELH X
島一
、r 一仰、

- -
：

P …N
M
ω2

J3 ( l.2)
d

where EP and EW are the entr阻ce pupil and entrance window 叮e阻（ cm2) ands is the
separation of the entr剖ice window from the en甘ance pupil (cm). N is called the 悶diance
 CHA肘1'R I. RADIOMET即C CONSIDERATICNS 3

of the so叮閱（ W cm-' ster-1) The primed symbols refer to the image side of the system
’ ’
EP andEW are exit pupil and exit window ands' ts the separation of the two.

In this fund目nental 且pression. N appe缸s equal in bo性E relations. indicating that no

reduction of 間diation has been accmmted for due to loss in transmission or other factors.

The princip叫 point to be made is that Eq. (1.2) is an invariant. It provides a choice
for detem1ining the power transfer from either the object side (target side ）。r the image
side (detector side).

When !he object is located at infinity、 the image is fom1ed in the focal plane. In thi且
一，
case. the 訂ea of the exit pupil is D'rr/4可 ands' becomesf. which modifies the image晴side
expression ofEq (I 2) to read
z
D 一

1,’-
π－
、－－

P N
4 z-
一－

-
rj

( l.3)

’
where d is the linear dimension of the square detector. The detector is the exit window.

TI1e radiated power in watts per square centin1eter from a 目前 diffuse source surface
mto a hemisphere is the radiant emittance W. The relationship betwe叩 the radiant
emitt聞自 and 血e radiance is N = WIπ

With this and the substitution off/# for f ID句 Eq. ( 1.3) becomes
立一悍

d
-fy
，一

pa I UP
一一
一叫“
一（

一）
－可一

( 1.4)
T

The noise equivalent power NEP is a function of the detector sized'. the electrical
bandwidth 4fused in the 甜甜 urement啥 and 也e detector figure of merit D失 which h血 the
somewhat unusual dimenston of cm Hz'" W 1 D* is the signal-to-noise ratio when I W is
而－

incident on a detector havmg a sensitive area of 1 cm', and the notse ts measured With an
electrical bandwidth of I Hz So,
p

( 1.5)

Substituting Eqs. ( 1.4) and ( 1.5) into Eq. ( 1.1) yields

D*d ’ W
SIN 一 ( 1.6)
一 4(J !#)" .p;f

Thts simple expression mdicates the strong influence of the chosen opltcal system.
The S月吋 is inversely proportional 個自e square of//#, the relative aperture. This means
that an IR system with an/ II objective perfom1s fot甘 times better with regard to SIN than
4 OPTICAL DESIGN FUNDAMENI <LS FOR J.附臼REDSr;TEMS

an f 12 Unfortunately‘晶 we will see in Chapter 3 、 the "faster” (low//#) a lens is, the
larger the aberrations.

1.4 Extended Simplified Radiometric Performance Equation

Equation ( 1.6) can be extended to 坦dicate the flow of the radiant energy 仕um the
source (target) to the detector by stating

S1皂nal-to-nmse ratio =[source power - background power)x[ atmospl時間 tr酬lillSsion]

×[optical throughput )x[detector efficiency]‘
or
'L d ’ D*
Er 叫en ﹞﹝几﹞﹝否可﹞﹝有i ( 1.7)

i i
I. Source & background
i
2。 Medium
i
3. Optics
•
4. Detector

In Eq. ( 1. 7). d' is included in the optics bracket. Even though d’的 the linear
dunens助n of the squ叮e detector. it is also the dimension of the field stop 祖d exit
window. The expression assumes that the image of the t訂get is larger than the detector.
i e川 the radiation received from the t缸get “fills叫 the detector.

We shall now address each of the brackets of this simplified expression. Because of
O叮 emphasis on optics‘ the third bracket will be discussed last.

1.5 Thermal Radiation Laws

，..／’ r、＊

州立︱﹝民εT 一作sll [tA ］［古詩情l

I I .,.

In this first bracket of Eq. (I. 7). W is the radiant emittance in watts per square
centimeter within the spec恤l b個dwid崗位 interest.εis called emissiv1恥 εis
dimensionless and its meaning will be discussed shortly. The subscripts r and 8 refer to
個get and background respectively. Background radiation is the radiation coming from
the s回roundmg target The effective or net radiant emi社間ce is the difference between
 CHAl'fER l. RADJOM凹RIC CONSIDERATIONS 5

target and background radiation. This me血1s that there h品 to be a radiallon difference.
positive or negative司 to detect or resolve a target (obj郎。

1.5.l Blackbody radiation and Planck's law

A blackbody is defined as a perfect radiator which absorbs all radiation incident

upon 11

In lus investigation. to find a relation between the radiation emitted by a blackbody

削 a function of temperature and wavelen~tb, M酷的anck (1 的8一 1947) developed tl1e
now famous equation named after him His efforts laid the foundation of the quantum
theory‘ for which he received the Nobel prize in 1918.'
On October 19陶 1900. he first reported his findings、 which were based on his
expenmental work.' Only two months lat前．。n December 14‘ he presented the theoretical
derivation of the equation' that described the blackbody radi阻tion curve:

w c, ( 1.8)
-
A 一司計亡1]

\\i1ere W; ~ spectral radiant cnuttance ( W cm ' i間一＇）， λz wavelength （悍的‘ T=

blackbody tempera！叮e (K). C1 = 37『418. and c, = 14,388‘ when the 缸扭扭 in sq咽re
centm祖ters e = base of the natural log訂ithm (2.718 ... ).

ln Eq. ( 1 8) notice there is a stron且 dependence on the wavelength and that W; goes
’
to zero whenλ ＝ 0 and =. Planck S curve for a 500 K blackb。你 is shown in Fig. 1.2.

叫［W cm-2 帥， I]
j.
入明 ＝ 5.796 間
。.04

。。2

。 抄 λ ［ µmj

2 4 6 8 10 12 14 16

FIG. 1.2 Planck's curve for a blackbody sour閻明自 T=500K.

To determme the radiant emittance over a spectral band‘ we mtegrate under the
Planck curve between the hmitmg wavelengthsλ； and Ao An example of a selected
spectral band is indicated in Fig. 1.3.
6 0/'TICAL D自1 日NF!IND.訓ENTALS FOR INFRAREDSYSTE.間

k 、

W= f~dA (l.9)
). I

w,
‘
A λ＂＂＂＇

．．）、
λ1λ2

FIG. 1.3 Racli曲廿四ittance m血m a selected spectral bandwid血

To perform the integration, one can use published tables or spec凶 shde rules' A
very convenient way is to apply Simpson's rule 叫自由e aid of a calculator. For 也at
pu中 ose we state that
‘
hT中 h
W

W A RA
叫

( )
～一－

仇V
Using a Aλof 0.05 or even 0.1 間n is sullicient for s凹11mat1ons over the mid-
wavelength (MWIR) and long-wavelength (LWIR) infrared regions (3… 5. and & 12 µm).
It 1s also good to remember that if the spec甘al band is relatively na甘ow均 W ~ W入 4λ

1.5.2 The Stefa骨－Boltzmann law

切dependently. Josef Stefan ( 1835… 1893) and Ludwig Boltzmann (184午 1906)
found expe討mentally that the total radiation emitted by a blackbody (the area m1der the
Planck curv的時

(l l I)
w!OIITT ＝ σT＇ 也

，一、→勻… A
where σ ＝ 5.66961 × 10 " (W cm - K 可 and T ~ source tempera細閱（ K）。
 CHAPTER I. RADIOMETRIC CONSIDERATIONS 7

1.5.3 Wien's displacement law

Tb肘e is a simple 個d interesting relationship between the peak 、.vavelength and the
temperature at which a blackbody radiates. Wilhelm Carl Werner Otto Fritz Wi閱
(1864一1928). a Nobel prize recipient 油 1911 咱 discovered 也is behavior of the blackbody
[Eq. ( 1.12)]. The eq閏lion states that the product of the peak wavelength and the so叮ce
扭mperature is cons！血1t. which E扭曲s the peak of the radiation shifts to shorter
wavelengths 扭曲e tempera恥re increases. Fig. 1.4 shov時 the displacement of the peak
W官velength

)'PT= 2897.8 (µm K). (!.12)

Applying Planck's la叭 the spectral rad· ant emittance at the peak wavelen研h is

Wi.rnax = 1.288 × 1。一15 T5 (W cm~2 µm一1 ). (l.13)

航
AA

High-temperature Curve
λpeak

．， λ
FIG. I 4 W1en's displac個1凹.tlaw.

1.5.4 Kirchholl's law and emissivity

’
Gustav Robert Kirchhoff (l 824-1887) stated in 1860 that ··at thermal equilibrium.
the power rad ated by an object m阻t be equal to the power abs凹的d." This leads to the
observation that if an object absorbs I 00 percent of the radiation mcident upon it. it must
reradiate l 00 percent. As already stated. thi咕咕 the definition of a blackbody radiator.

Most radiation sources are not blackbodies. Some of the energy incident upon them
may be reflected or transmitted The ratio of the radrnnt emit!面臨C W’。f such a source and
the radiant emittance W of a blackbody at the same temperature is called the emissivityε
of the source·
8 OmcAL DESIGN FUNDAA館附'41.SFOR/NF制REDSY.ffE.叫

£= wγw. (l.14)

With this relation. different types of radiation sources can be classified as indicated
in Ftg 1.5 where the c叮ve for the blackbody withε ＝ 1 is Pl阻k’s curve. The curve for a
graybody ts proportional to Plank's c前ve for all wavelen研hs. The spec個1 radtant
emittance for a selective radiator vanes not only 前th t回nperature but also wtth
wavelength.
Wλ

‘
A Blackbody (E = 1)

Selective Radiator （ε ＝ f （λ，T）﹜

Graybody （ε ＜ 1)

．． λ
FIG. 1.5 Spectral radiant en甜甜ce of three t)lles of 間tlrntors.

1.6 Transmission through the Atmosphere

／叫.f T 院＜ a﹞曰 t 仇 d'

﹝－－－＂－－；－；－﹞﹝？片﹞
4(/1＃）一叫A J
D*

The second bracket of Eq. ( l. 7) addresses the transtruss1011 ofradiant energy through
!he atmosphere.

To assess all influences that can affect the transmission is a ve可 complex matter.
Much work has been done by many researchers over many year可 to model the attenuation
ca間ed by the 甜nospheric g臨的一' Many vanables‘ such 晶 changes in temperatu扭‘
pressure of the g扭曲. shapes. sizes. and chemical compositions of suspended particles:
and slanted opti間i path for example‘ make it extremely difficult to predict IR
仗ans1mttance of the atmosphere. Our pu叩ose he間 is to pomt out that when the IR sys阻m
under consideration is to be employed over a long dist叩開‘ atmospheric abso叩ti on
elfec阻 may have to be mcluded Dependmg on the application. a cursory look may be
suilic1ent. In some c剖es‘ however. a very detailed analysts 咐II be necess剖y.

We limit our discusston to a horizontal path嗯 near sea level, and develop
approximations that allow 田 to obtain a basic understanding of the primary impact of the
atmosphere on IR systems.
 Clh\l'阻R 1. RADIOMETRIC CONSIDERATIONS 9

'’
1.6.1 I e rmanent constituents of drγ 剖mospbere

The relahve 叩阻四恤twns of E且es present 凹的『 ahnosphe間 are ne：前ly constant
and are therefore called the permanent constituents Table l.l ident1 ies these consti削翩翩
and indicates their abso中tion charac阻ristic in the infra間d spectrum.

Table 1.1 Permanent constituents of dry ahnosphere.6

Constituent Chemical Percent bv Absorbs

(m句 or) formula volume between 2 胡d 15 ~lill

Nitrogen N、 78.084 No
。、ygen o, 20.946 No
Argon A 。 934 No
Carbon dioxide co主 0.032 Yes
Neon Ne l.818xlO~i No
Helium He 5.24xl0-' No
H直ethane CH, 2.0xlO' Yes
Krypton Kr l.14× 10 J No
Nitrous oxide N,O 5.0xlW' Yes
Hydrogen 11, 5.0× 10 , No
Xenon Xe 9.0xlO' No
Carbon monoxide co 7.5× 10←。 Yes

1.6.2 Variable constituents

The two major constituents in th" atmospl時間 that va丹F w!lh tempera仙re and
altitude are ozone 由1d water vapor

Since the maximum concentration of ozone occurs al a high altitude (between I 0

and 30 km above the e盯Ll1's surface )‘ we can ignore its effect 四d state that

wate1 vapm causes most of the absorpt馴的 the JR region 叫 sea /evd

1.6.3 Approxim剖岫n “
(assump on)

’
Near sea level one can approxuna扭 the transmi位ance through 血e earth s atmosphere
by considenng the abso叩tion effects of carbon dioxide and water vapor onl｝ιThis is
expressed by
＝ τ co 、 × t H 、0 (l.15)

Here. H20 represents water vapor whose aensity vanes 吼叫h temperarure and
humidity. Figure 1.6 shows the inf詢問d wavelength regions where the transmittance is
afleeted by the presence of water vapor and carbon dioxide.
10 o，阿／CAL DESIGN Fr!Nn品1ENTALS FOR INFRARED SYSTEMS

100

50

2 4 6 8 10 12 14 µm

’, MWIR ‘ ‘
LWIR
..
FIG. 1.6 Spectral W111dows in the infrared,

l.6.4 Precipitable water (deliniti。”

The amount of water vapor contained in the optical path is called precipitable wat凹
尬1portant to remember 1s the word vap凹， i.e 司 water in g晶eous form Precipitable water
(Fig. l.7) is defmed 品 the depth of the layer of water that would be formed if all the
water vapor along the lme of sight was condensed m a container having the s血ne cross
section 由 the optical bundle.

立
Optical Bundle
Water Vapor
‘ .
Optical Path Length
,. (RangeR)
Precipi個hie Water

·=3 " ""Total Precipitable Water PW••

(”condensed water"}

FIG. 1.7 Precipitable wat叮司 concept

Nollce that it 1s not necess血y to know the cross-sectional area of the optical bundle
(cylinder). If the cross section inc甜甜的.mo間 water is condensed. but tt is spread over a
larger area and its depth remains constant.
 CHA!河ER l. 扎白的關T目C CCNSIDERATIONS 11

1.6.5 Humidity

TI1e calculation of the total runount of precipitable water along the line of sight of
an optical system includes Ute existmg relative humidity Let us state a few 甜latiouships
about humidity in general and then ap~ly them to an ex血nple. To simplify mat阻rs
somewhat. we shall develop some approximat10us.

Absolute humidity I AH (gm~﹜ l

AH(ti is the m阻s of water vapor in unit volume of the atmosphere at tempera恥re
I (°C). AH,,.(t) is the m阻llll間1 amount of water vapor the atmosphere can hold at
tempera個re t This condition is called sa個問組on

Relative humidity (RH)

Relative humidity is sunply the ratio between the m阻s of water vapor per umt
volume pr品ent in the air and the mass of water vapor in saturated arr at the s血ne
tempera缸rre. i.e’ ‘
Al/(t)
RH ＝ 一一一斗一 (1.16)
AH間（ 1)

1.6.6 Precipitable water (calculation)

The an10unt of precipitable water is conveniently expressed in millimeter per met叮

path lengU1 and can be calculated by

/Jll' = I 0 3 AJ-1(1) = l 。一3 AH蚓（ l ) × RH. ( l.17)

For the temperature range from 0° to 35°C. (32° to 95。F). an approximation for the
absolute bumid1tv at saturation L可

AH＂＂ 王 5.071962 eoos9688' (l.l8 }

This approximation is accura扭 to better than 5% throughout the stated tempera仙re

range

For a I-km path length‘ Eq. (l.18) can be restated as

I只 1• =: 5.072e00597' x RH (mm/km). ( 1.19)

1 he nomogram shown in Fig. 1.8 is based on this expression.

12 UPTJCAL DESIGN FUNDAl>館前ALS FOR INFRARED SYSTEMS

Temperature Precipitable Water Relative Humidity

I (°C) pw (mm I 恤） RH(%)

100
80
35

。
60
。 50

30
。 40
司。

25
20
4
15
20
2
t.5 10
8
15
0.8 6
0.6

0.4 4
10

0.2
5 ~

0.15

。1
。
FIG. 1.8 Nomogram for detennuung tl1e amouut of precipitable water from tl1e temperature 祖d
relative humid1tv.

In the ex缸nple shown m Fig. l 8. an atmosphere with 60% relallve hwnidity al

n。C. contains 20 mm/km precipitable 、m阻r.

1.6.7 Atmospheric transmission {calculation}

Equation ( 1.15) states that the to祖l transmission through the atmosphere for spec的c
conditions can be appro之ima阻d by the product of the transmission components of carbon
dioxide and 、~ater vapor心

To 晶sess the effects over different spectral bandwidths‘ vaηing path lengths and
precipitable water contents requires even with our simplified expre<sion. the aid of
lookup tables such as provided in Re[ 6. To delemune the total transmission 111 the
MWIR region (3-5 間叫“r the example refeπed to in Fig. l.8. we find the ave凹ge
 CHA~fER I. RADIOMETRIC CONSIDERATIONS 13

transmittance of CO, from the 祖.bles in Ref. 6 for a 卜km path length 吐。2 ~ 0.848. and
the average transmittance for 20-mm prectpitable water 缸，o = 0.591. Therefore, the total
transrmss1on through the a恤osphere over I km near sea level at 32 ℃ and 60 percent
relative humidity is aP.proximatelyτ個凶＝ 0.8.48 x 0.591 =o 0.50. For the s回ne
temperature and humidity condillor旭、 the total transmission over I km m the L WIR
region (8 to 12 µm) is consrderably better, namely trumo, = 0.995 × 0 723 0 72. =
Figure 1.9 ts a plot of the atinospheric transmissron 品or both infr前ed 、Nllldows
(MWIR. 胡d
”
LWIR) over I km. The above calculated examples are indicated. The two
curves provide information of the impact of the preci table water contents vanation over
the two m句or IR windows

Atmospheric TransmissionτA
l

0.8
0.72 ...

0.6
0.5 ...

0.4

。2

。 1
仇υ

2 5 20 50 100
Precipi個ble Wat拾r(mm)

FIG. 1.9 Tr:msmission over I-km path for 由e MWIR 阻d LWIR spectral b胡也

l.6.8 Computer models

lt is obvious that the sub,1ect of atmospheric transmission is extremely complex. 1 he

increase of available computer power and the dedicated efforts of many specialists in this
field have led to computer models that are very detailed. At the present. some of the
commercially available progran1 models are LOWTRAN. HITRAN. and SENTRAN.
14 a月TCAL DESIGN FUND，晶fEJ'll'ALS FOR 卸FRARED Srsm1s

1.7 Typi阻I IRDet缸”問

D*
SIN ：﹝WrEr 一 rt主E Rj [t Al﹝」~］
I I N 4(J!#)•
N A l可言11
In this section we will briefly address the bracket of the simplified performance
equation [E~. (1.17)]. which relates to the detector of the system. We begin by identifying
the two basic group.的 of detecto間 that are connnonly used for conv叮tmg radiant energy
into 祖 electrical signal.

1.7.1 Thermal detectors

A thermal detector absorbs radiant ener的心 which causes a change of the detector's
electrical charactensllcs. This electncal 間spouse to a ch胡ge m the target temperature
produces a signal 也at can be amplified and displayed.

One of the most attractive characteristics of thermal detectors is the equal response
to all wavelengths. This contributes to the stability of a system that must opera扭 over a
wide temperature range. Another significant factor is that thermal detectors do not require
cooling

However, the response tune of these detectors 1s m milliseconds and therefore

relatively slow. Furthermore. their de阻ctivity is 田 much as one or two orders of
magnitude lower tlrnn that e文penenced 、討自 photon detectors.

The most conunon thermal detectors are thermocouple. thermopile. bolometer. and
pyroelectric.

1.7.2 Photon or quantum detectors

Photon or quantum detectors operate on the quantum or photon elfoct Photons are
absorbed and produce free-charge ca甘iers that change the electncal charactenstic of the
responsive element.

Photon detectors are much faster than thermal detectors their response 1s m
:‘

microsecon也 As already mentioned. their detectivity is considerably lugher. To ob個m

this high detectlvi吵， the detector must be cooled. For moderate temperature reductions.
one- or multistage thermoelectric coolers are employed. To provide coohng to very low
temperatures of 77 K 祖d even lower司 cryogenic cool mg methods must be applied.

1.7.3 Photoconductive detectors

llus 1s the most widely 田ed group of photon detectors Their function is based on
the photoconductive effect. Incident infrared photons are absorbed producmg free-
charge carriers that change the electrical conductivity of the sensitive element
 CmerER l. RADIOMETRIC CoNSIDERATI側5 15

Some of the materials found in this family of dete叫ors are

. lead sulfide PbS

.
• lead selenid。
indium antnnonide
• 1nerc凹y cadnuum tellunde
Pb Se
InSb
HgCdTe

1.7.4 Specific detectivity and noise equivalent bandwidth

One figure of merit 田ed to describe the perform祖ce of a detector is the specific
detectivity D叮D-st的 H is the si伊al－峙的師自tio wh間 I W is incident on a detector
having a sensitive area of I cm' and the noise is measured with an electrical bandwidth of
I Hz. This definition was introduced by Clark Jones in the 1950s. Its din時nsion is a
cumbersome cm Hz112W1; there has been the suggestion to change it 妞”Jones叫 units 7

The denominator of the bracket bemg discussed here refers the electrical bandwidth
of the system's 誼nplifier used for the noise me品盯ement. The most common defmition of
tl1is bandwidth is the frequency interval wi也m which the power ga詛 exceeds one-half of
its m肌imum value. It is also often called the 3-dB bandwidth.8

Figure 1.10 shows the specific detectivity as a function of wave I叩gth for some of
the most commonly used IR photoconducttve detectors.

D*
i
nυ
令－

PbS
195
PbS
I!" 77
PbS
295
I AUO

／話：~＼ ~；
H~~dTe GeCu
4.2
1o•
Thennistor
ilolome曲r
10" 295 λ

2 4 6 8 JO 20 40

FIG. 1.10 Spectral se酷itivity of some commonly used IR photoconductive detectors.

To illustra阻 the effect of cooling on p_erformance. some curves in Fig. 1.10 have
been highlighted. They refor to a lead selenide detector at three different temperattues:
295. 195. and 77 K (or 22.一 78.and 一 l 96°C). It ts mteresting to note that the spectral roll-
o仔 response shifts to longer wavelengths as the temperature of the detector decreases
This is a v的 impo巾nt prope叫y of a sensor As a practical example司 it can be seen
clearly that to measure carbon dto、ide, the sensitivity and stability is much greater if the
16 0阿7CALD.目！GN FUNDAME，盯'ALSFORINF制REDS間站附

detector is cooled to 195 K than it would be for 曲 uncooled de扭ctor. The C02
absorption occurs at 4.25 ﹔冊， which is whe間 the cooled de扭也tor peaks. The importance
is not only in the increased D*; the detector is not operating at 也e slope of the curve,
providing grea阻T S姐bility with t回npera扭扭 changes. It can also be s間n that when the
PbSe detector is cooled to an even lower te~perature, the shift to longer wavelengths
continues but there is no further gain m sensitivity (1志，血creaseofDη－

1.7.5 Detector configurations

Quite frequently, detectors are not smgle elements but 盯e aπanged in 由祖抖， In
thermal imagir嗯， for ex缸nple, a number of detector elements a阻 mounted in either a
single line or a two-dimensional layout. Line缸 a口ays reduce the required scan叫ng
mechanism to one axis, horizon阻l or vertical. (Sc間ning is the proc間S of scene dissection
by sequential!~ detecting the radiation 音。m a field covered by a single element or a linear
M祖y). Two-d1mens10nal 由四ys eliminate scanning comple阻ly This 扭曲gement is
known 晶 S出ng arrays or focal plane 缸祖ys (FPA) One can e由ily imagine how
ad van祖geous such a grouping of many small detector elements in the focal plane can be.

From an op臼cal standpomt, however.，甘1ere is a trade-off, smce such an a叮ang回nent

pre sen扭曲 extended field of view which leads to a higher degree of complexity 喃自
問gard to aberration coπections. This will be discussed in de個il in Chapters 3 and 4.

Fi評rre I.I I shows a multistage thennoelec出cally cooled detector 缸祖y The

individual elements (pixels) are as small as 30 × 30 µm and the total number of elemen甜
in such an array can be more th凹的， 000

FIG. 1.11 Thennoelectrically cooled HgCdTe focal plaoe array. (courtesy of Hughes, Santa
Barbara Research Center）可

With the next chapter we shall begin to discuss the optics bracket of the extendcd
simplified radiometric performance equation [ Eq. (1.7 )}
 CUAfiER l. RADIO!惟＇TRIC CONSIDERATIONS I7

References

L G. Holst. Testing and Emluation qf b阱，·ared Imaging Systems‘ JCD Publishing

Company (1993).
2. E. Hecht. Optics. second ed叮 Addison·司Wesley (1990 ).
3. F. von Krbek. Erlebte Pl(vs1ιDeutscher Verlag ( 194 2)‘ page 119 心
4. b妒。red Radiation Calc11/ator. lnfr宙ed Infonnation Analysis Center‘
Ann Arbor‘ MI. and EG&G Judson. Montgomeryville. PA and others.
5. F. Smith苟 η＇1e ff!什ared and Eleen旬，Optical Systems Handbook ‘ Vol. 2‘ Atmospheric
Propagation qf Radialio比 ER品也＆ SPIE (1993).
岳 R Hudson.In升a月d System Engineering、 John Wiley (1969）可 page l lR.
7 W. Rogatto，刃，e h斤。red and Electr加Optical.勻•stems Handbook司 Vol. 3. Elect’。－
Optical Components. ER必也＆ SPIE ( 1993 ), page 251.
且， R Hudson‘ h阱。，·ed System Engine臼；，唔， John Wiley(l969), page 312.
CHAPTER2

Basic Optics
2.1 Introduclton

With this chapter. we begin to cover 血e thi的 bracket of our simplified radiometric
performance equation [Eq. (1.7)].

SIN ＝ ﹝ WrEr 一 W8£8)(t4]

﹝最精l
where 1(} is the net transmittance after absn叩tion and Fresnel losses of all optical
elements, which includes not only le酷的 and mirrors but wmdows 曲d filters 晶 well. As
mentioned earlier‘ even tho略h d ’ is the I泊ear size of the detector element. it is mcluded
in the 的optics bracket'" because 1t is the dimension for the field stop

Included m this chapter are some numerical exan1ples of S別 calculations for the
reader who is not concerned about optical aberrations 叩d is satisfied 咐th a very
prelimmary per品ormance prediction for a conc句阻al system configuration

But. at some porn!. more detailed 祖alys1s of the optics is 間quued to do the job. For
that pu叩ose. a fundamental understanding of optics is ess自1tial.

’
We begin with the foundation of geomet討c optics司 which is Snell s law, named after
Willibrord Snel van Roy閱（ 1581-1626 ）‘ a Dutch 晶tronomer and mathematician who
worked at the University of Leiden in Holl叩d It slates that the sine盼i11dex product is
equal across an interface when hght passes from one transparent medmm into another.
τhe index of refraction is simply the ratio of the velocity of light in a vacuum to the light
velocity in a medium. With reference to Fig. 2.1 we c祖 write

sini N'
(2. l)
sini' N

19
20 0阿K AL DESIGN FUND.品位問 ALS FOR 卸＇FRARED SYSTEMS

FIG. 2.1 Snell’s law applied to a ray 扭扭ring a denser medium 叭，〉的

2.2 Snell's Law and the Prism

’
ln applying Snell s law to the surfaces of a prism. we notice that a rny p品sing
through the prism ts always bent toward the thicker part. the base of the pnsm 羽tis IS
illus仿制ed by Fig. 2.2.

FIG 2.2 Snell's law applied to a prism.

2.3 The Transition from a Prism to a Lens

If a number of truncated prisms are stacked 品 indicated in Fig. 2.3. rays

emanating from an axial point converge or diverge symmet討cally after p由sing through
the prism a叮ay‘ depending on the orientation of the prisms. By incre品ing the number of
pnsms to infinity. a convex or concave cylindrical lens 、社ll emerge. For a rotationally
symmetrical lens. the sections sho"n in Fi臣， 2.3 are merid10nal cross sections through the
lenses.
 CHAP克R 2. BASIC O!'f!CS 2J

個ngent
normal

~智、
。 0 必訟法

?/
i ：＜作
。 。

FIG. 2.3 1r血silion from pris前IS to lenses.

A tangent drawn where the ray enters the lens 晶晶。、叩 m Fig. 2.3 is equivalent
’
to the face of a ··Jocalized prism·· Snell' s law c叩 be applied to the direchonal change of
the ray‘ where i is formed by the ray with the nom1al to the lens surface.

2.4 Image Formation

For simplicity we lim!l this discussion to a single thin Jens and use so-called
first-order exp閃電 sions.

A thm lens 峙。ne whose tluckness is usually small compared to its focal i叩gth;
in other words it 1s assumed that the thickness is zero. We ,..,11 show later that even
though such a lens does not exist侖山e concept of a thin lens 1s very convement. It
simplifies calculations and still provides meaningful results.

First-order expressions 宙e derived from the ass凹nphon that the sine of an angle
is equal to the tangent and therefore equal to the angle itse旺. The domain for which tlus is
the case is called the paraxial 悶gion. and the assumption is called the para.xial
approximation because some simple approximations can be made in this region. This
concept w1 II be discussed in detail m Chapter 3.

It 1s impo11ant to pay close attention to the sign convention in optical ray tracmg
Recently‘ there has been a ch由1ge in the sign of the angle between a ray and the ophcal
axis Therefore司 it 1s advisable to take a good look at the definition of the sign conventions
used in a particular reference. The sign conventions show吼：n Fig. 2.4 are those most
commonly used today心
22 OPηCALDESIGNF凹m品fENf.<LS FOR INFRARED Srsr卸的

φ呵呵一－－－…呵呵呵－呵， +

+
A其jal JlJIY all
y (+)
U （吋 ’ {
- n LT

u;+ (+)
principal R•Y

L 一 L ＇ 一一一一一一一一，
(-) (+)

FIG, 2.4 Baste 間lattons 聞d si伊 conventions for a 也且 lens in tl1e paraxial 間＇gion

If the object is located at in且nity. angle 11 becomes zero and the distance from the
lens where the image is fom阻d is called the focal length.( of the lens. and the image plane
is the focal plane. The focal length of a converging lens is posi訂閱司 and the focal length of
a diverging lens is negative. This is indicated 區 Fig. 2.5. If the lens is surrounded by air.
the focal leng吐ts are equal for each side of the lensι

...
Convergmg lens (posiuve focal le羽且也） Divergmg lens (negative focal length)

’, ……’, 品已

~
、
F F
。 F'
0. 0 .0
F .，...、

…弓，“ 白白” 4 『 ﹜司
- - ＿.，..平 一一心

.K
‘ f
(+)(÷)
~ f ’
‘ f
（寸（寸
~ F ’
FIG. 2.5 日tgn conventions for converging 祖d diverging lenses.

There are a number of convenient and simple expressions to detennme location‘

magnification. and orientation of 祖 unage for the arrangement shown m Figs. 2.4 and
2 5. The first 匡正文pression ts

1 1 l
(2.2)
L ’- L + f
 C!IAPTER 2. BASIC 0阿JC電 23

Magmfication c由1 be stated in several ways

h' L' 11
(2.3)
m h L u'

The 個gular relation is often not recognized. The magnification is zero when the
image is formed m the focal plane.

2.5 Object而Image Relations

The relationship between object and image is best w1dersiood by observing what
happens to the image as an object. fixed m size. is bemg moved closer toward the lens
(see Fig. 2.6) To find the loc甜 on of the image 祖d !IS SI間， a ray p叮allel to 由e opt1Cal
axis 阻 drawn from the top of object 1 (arrow head) to the right until 也e ray intercepts the
pl個e of the lens. After p晶sing 也rough the 也in lens, the ray travels downward and
intercepts the optical a世s at focal point F’ Ano也er ray is drawn from the 咽me object
point 剖ming at focal point F. This ray leaves the lens parallel 扭曲e optical 甜甜. Where
the two rays intercept lies 出e arrow h個d's in.age pomt of image I ＇.的也e object-
located to the left of the converg泊E lens-is moved closer to the lens (object 2 ). the
image 2' moves to 也e nght, further away from the lens When the o吋ect is placed at the
focal point F of the lens句也e image is formed at infinity. This is the principle of a
collimator, whose function it is to simulate a so叮ce located at 垣fmity. When the object is
placed between focal point F and the lens. the formed image is virtual 3' is the virtual
image of object 3. To determine size and location of the v甘甜al image, the exiting rays
after passmg through the lens are 目tended backward until they cross each other (dotted
lmes m Fig 2.6). Object and image are at 血e s缸ne side of the lens and the image
orientation does not change as m cases l and 2 whe間 real ima甚es are formed A reai
image can be received by a screen. a v廿恤I unage cannot. Good examples of virtual
images are what can be seen in a mirror or through a maguif1er
vim且l image

3•t <!

. ”,

"
t :¢ ~ F
~ 。F'
問al images

.A.
』

‘
’,

L, L ’1
‘ ’‘ ’,

FIG 2.6 Objec• image relations.

24 0P7JCAL DESIGN FUNDAMENTALS FOR ！.抑制REDSfSTEI悶

扭扭 import祖t to understand the concept of real 胡d vim岫l ima惡戰 because as we

will see shortly. entr削ce and exit pupils. winch control the throughput of radiant energy
t尬。ugh an optical syste且 can be real or virhml.

Applying Eqs. (2.2) and (2 月 to a lens with a focal length of/= 30 祖d 祖 object
height of h = IO. the three c晶es shown 詛 Fig. 2.6 are tabulated in Table 2. L

Table 2.1 Ch祖ge of image locatton. magnification. 阻d 叩開tali閻明曲曲曲ge ofobject

postlion

L L' 111 Image

一75 50 一0.667 -6 667 real

2 45 90 -2 一20 real

3 一20 一60 +3 +30 virtual

•
Figure 2. 7 shows the objec image relation [Eq. (2.2 )] for a positive lens‘ plotted in
focal length 凹tits The three ex由nples from Table 2.1 are identified.

L’
;.
間al object,
real image 5/
Objectives 4f
: 3f
2/
一－－一一一一一一一，吭一－ ／－－＿寸土＂＂－＂＂＇ τ

’, L
-7/ -6/ -5/ -4f -3/ -2/ -.( -f

real ob1ect, 3 -2/

virtual image
-3/
Magnifiers
-4/

FIG 2 7 Plot ofEq. (2.2) for a positive Jens.

2.6 Stops, Pupils, and Windows

There are two stops that html tl1e bundles of energy p祖srng through an opllcal
system. One is the ape1t11re stop (ASλand the other is the jield stop (FS).
 C!IAPI主R2. BASICO柯：臼 25

The aperture slop can be an element of its own, such as an iris, or 11 can be the edge
of a lens mount. The apert盟海 stop sets the 個gul宙 limit of an 阻ial ray. Its image !n
object space is called the ent1 ance p叩ii (EPλand its unage in the 油1age space is the 四，，
pl伊ii (EP'). Intermediate in盟ges of血e ape叮叮海 stop wi出扭曲e system 盯e simply referred
to as p1.伊ifs.

For 血1 optical system 、.vith a real object (target), the space extending to the left of the
first optical surface is the object space, and the space to 也e right of the last optic羽i
surfa阻 is the image space if the image is 間al (detector pl甜的’Ifie situation becomes
somewhat complex when the object or the image formed by the system is v廿岫l.12 TI1ese
C晶es will not be considered h叮c.

The openmg that lmnts the size of an o句ect that can be imaged by the sys阻mis
called the field stop. The principal ray traveling from the edge of the field stop through
the center of the aperture stop is the m叮gmal principal ray The field stop 1s 由e always
related (conjugate } to the image and object ~lanes. Frequently, the field stop lies in the
nnage plane. In a c血nera it is the film司 and man 函frared system it can be the detector
itself. In a focal plane ar祖y （押的‘ the full field 阻 limited by the dimensions of the 研勾F
The field c滑vered by an ind1v1dual detector element 1s called the mstan臨icons field. The
images of the field stop in object space and ima_ge s~ace are respectively called entrance
window (EW) and 叫f wmdow (EW'). Intermediate ima耳目 of the field stop are log1cally
referred to as windows

To better understand the int前那tion between the slops. we will discuss the
arrangement shown in F1g 2 8 The ap叫.ure slop 1s placed between two lenses and the
也tector is the field stop. From the layout. it C胡 be seen t加t the size of the aperture stop
AS controls the energy cone angle 211, and the field stop size FS lintlts the angul宙 field
size 211p The hm1tmg rays for !he energy cone 阻d the field angle are the m前gmal axial
and tl1e m盯ginal principal rays‘ respectively. Reducmg the size of the ape叮叮e stop does
not change !he field coverage. nor does changing the field stop size have an impact on the
cone angle 2’,
Ape話間E Stop

AS
Target
Field Stop
FS

..
“
2u

4‘
2u,
司，
。
I(D加et目咿 tor)

Lens I Lens2

FIG. 2.8 Stops Tiie aperture stop controls 211 ，

： 也e field stop 帥ntrols 21年
26 OPTICAL DESIGN FUND.』fENTALS FOR /NF.凹RED S悶悶悶

In applymg the techniques of detennirung locations and sizes of images of the stops.
the pupils 甜d windows can be constructed. For the chosen ex缸nple. both the entrance
pupil and the exit pupil are vir個al images of the aperture. 晶 shown in Fig. 2.9. There is
much more to be learned fro血 Fig. 2.9.

From its start at the target. the marginal 缸ial ray aims at 也c edge of the entrance
pupil. TI1e power of the first leus bends the ray to direct it toward the edge of the ap自tu間
stop After p酷smg throu的 the second lens司 the ray crosses the optical 阻is This JS the
image location Look間E from the image side at the backward extended marginal axial ray
reveals that this exiting ray appe宙S to be coming from the edge of the exit pupil.

A similar observation c血 be made by tracing an oblique ray t加ough the system

from the edge of the object auned at the center of the en叮甜甜e pupil. This ray‘ the
m叮·ginal principal ray. p描ses through the center of the aperture stop and terminates at the
edge of the exit window. It appe前S to be coming from the center of the exit pupil.
EP
.
即

.... AS ...
- ~ ~－＂’卜，
- -=-...:::--
一’，一，
’，亡，四’，
--－~
－－－一， F ’，
M缸ginal Axial Ray •’’
,-,-
一，－一，

研斗
- d

→’，一
’
EW
FM
ww

干了可
‘戶 心
-
l

。制正』－~ -

Marginal Principal Ray

Lens I -.
E
Lens2
E
.- BS

Object Space Image Space .

.--
、，
a

..
FIG. 2.9 Interaclwn among stops. pupils. and windows.

2.7 Throughput

Throughput is one n剖ne for the optical invariant that is used for the product of the
pupil area and the solid angle subtended at this pupil by the window area. This means that
the interaction between entrance pupil and the entrance wmdow 訂e the s田ne as for the
間it pupil and exit window. Stated mathematically‘

- --
W一

x FZ ~A E V
Pa一
P 一

fu
ι －

ra - -S
組

CIM
nva (2.4)
 CHAPTER 2. BASIC 0凹！ CS 27

In this expression. EP 阻d EP ’ as well 田 EW 祖d EW’缸ethe 缸e晶。f the respective

pupils 祖d windows. S is the spacing between EP and EW.。叩dS’ is the separation of EP'
’
fromEW '. Figures 2.10 and 2.11 show the 缸四ngement for o世棚恥lens example.

Throughput has been called by many different m血es‘ optical extent, light ga也ering
’
power. and area-solid-angle product. E endue is the French tenn, 阻d "geometrical light
flux”的 the tJ田station of the German term Geomet1 ischer L1chtst1 om.'-'

What is really import阻t is the choice 也at is available 咐th 也is invariance. One can
calculat.e 也e throughput 1扭扭E information from the object side or from the image side.
When the image is at infmi紗，品 i且也e c晶e of a collun副or‘ dtmensions referring to 也e
entrance pupil 叩d entr祖ice window can be 田ed because S 阻 a fini阻 quantity. while S ’ IS
infinite. Similarly‘ when the object is at infini紗， values for EP', EW＇， 血d S ’ from the
image Side can be used.

EP

S’
s
FIG. 2.10 lbroughput inv叩胡ce.

EP

EW ’ 呵，

EP ’
EW

‘ ... S'
. .,.
s
‘
a

l'IG. 2 II A 伽ee-dimens1onal look at throughput mvariance.

It must be added that pupils and windows are not always crrcular. In fact. for
infrared systems‘ extt windows are mostly rectangular in shape A typical example is the
28 0月1日L DESIGN FtlNDAMENTALS FOR INF叫RED SYSTEMS

staring focal plane 前ray. In a Cassegrain telescope‘ the pupils are doughnut-shaped‘ due
to 也e cen回l obstruction of the secondary m肘。t What counts is not the shape but the
are晶。f the pupils and the windows.

2.8 Energy Transfer

The invariance r can be extended by multiplying it with the so田·ce radiance N to

yield the power P 間ceived by the detector. As stated earlier, this is predicated on an
overfilled detector 晶晶 is the field stop.

EP x EW EP'x EW'
P=fxN ＝ ~N 乏 的一§.1 N (2.5 }

This is the second equation [Eq. (1.2)) that W曲 presented in Chapter I. To a品。間t
for tl1e transnussion losses through the optics 祖d. if applicable, for the a個osphe間. the
second eq阻lity for P has to be multiplied by tl1e appropriate transmission factors.η10se
for the a阻回sphere were discussed in Chapter I. Others will be e"-plained in Chapter 8 on
Optical Coatings.

2.8.1 Signal-to-noise calculations

We Will now analyze a stmple IR system We begin 叫th a ltigh-temperature target

個d observe the impact on the signal－！。”noise ratio (SIN）且也e target temperature drops.
We then switch to another wavelength region to see how that 前ll influence SIN

Let l晶晶S田ne a t叮且et temperature of Tr = 800 K and a backgrotmd temperature of

Tn = 300 K‘ and postulate that the emissivities for the target and the background are the
same, namely εr ＝ ε8 = 0 85 TI1e chosen spectral band is 仕om 3 to 5 間n. the linear lead
selenide (PbSe) detector dimension d ’= 0.05 cm (square). and its specific de阻ctivity D*
= O 5xl09 cm Hz112 W 1 。 The total system transmittance t = 0.6‘ the lens diameter D = 5
叫‘ its focal length/= 20 cm and the noise equivalent electrical band width ti/= 100 Hz.

With Eq. (1.10 )‘ we find for this high target tempera扭扭 that the radiant emittance of
the target Wr = 0. 791 的 W cm' The background contribution W8 = 0.00059 W cm '
The relative ape他rre or／／＃竺f ID= 20/5 = 4.
i 、 D*d’ t
The SIN is therefore S/ N ~ E （既 － Ws ） 一-;---;zτ立三 15.747. This is aveη

comfortable n阻nber

For the next calculation. the only change that 叫Ii be made is to lower the target
temperature from 800 個 320 K. This reduces the Signal to noise ratio to 11 4. a veηr low
number Just 晶 a point of reference, a stgnal-t伽 noise ratio of 100 would indicate that the
noise is l % of the signal.

The 咽。ng dependence between signal-t軒noise ratto and target temperature is

shown in Fig. 2. 泣， The curve is based on the above specified system with a 300 K back-
ground temperatu時
 ι：1JAP祖晨 2. BASIC O何！CS 29

S丹d
.
10,000

、A
l’。。。
I AUAU

＼I
10

’, Source
Temperature (K)
1,000 800 600 400 300

FIG. 2.12 SIN a strong function of the source 祖mperature

What happens if we observe the target at longer wavelengths (8-12 µm）。 Can we
improve that low S刑 of 11.4 for a target 記mpera恥re of 320 K 姐 a background with
300K?

TI1e rad闊別。mt 仗叩ces in this spectral window for the given tempera個res are
W1 = 0.0165 and W8 = 0.0121 W cm…一， . These valu目 result in a signal-to-noise rauo of
87.5. While this in an improvement by a factor of mo間 than 7 司 there may be additional
means for improvement The most effective modification of the system would be to lower
the ／縛， because it en臼rs into the eq閥lion inversely sq阻red In other words. a change
from the assumed f 14 to f 12 would quadruple the S月'1 from 87.5 to 350.

This little exercise indicates how to read 胡d inte叩ret Eq (I 7) It can save much
time in evaluating a given si仙at10n during the imtial layout phase.

2.9 Differential Changes

It is O自en interestmg to know what differential changes m radtant emittance W can

be exp郎ted with differential ch血1ges in temperature T. To refer to a tempera tu晶晶晶1ge
of I K, we simply subtract the integrated Planck functio阻 for two temperatures司 1° apart.
』1athematically stated 、

ow 人，（） JV,(T) "' "'

- = j 一一－ dT=l:W>.(T+l ） 必 － 2'.:WdT)L.lλ (2.6)
oT 人， or ,., ,,

For T= 300 K:
ow
一一一一： 2.1 × 10·' W cm 這 K一＼
iiT, 5µ
30 0阿1CAL DESIGN FUND.品何問ALI FOR 卸F制RED Sl'STEMS

and
ow
一一一一－＝ 197xtQ'Wcm 品 K'

a 兀－12
Of course‘站 the temperature of the 祖rget changes. so does the ratio of the

們一們
differential changes in the two spectral windows. We define this ratio as

n 一TJ
制－iz
(2.7)

It can be clearly seen that there is a crossover point where the sensitlVlty becomes
higher in the shorter wavelength region. That crossover occ叮s at 470 K or about
387°F. This is illus前ated in Fig. 2. 13.

IO
9.4 ..
‘
4

抄 T (K)

300 350 400 450 .. 500

470

FIG. 2.13 Ratio of radiant emitt扭扭s per degree Kelvin in 血e spectral reg10nsι12 and 3-5 µm

2.10 Optical Gain

Immersion lenses. light pipes. and field lenses can be employed to incre描e the S肘
。fan Il主 optical system without changing the field of view. These additional elements
perm必t reduction of the physical size of the detector and have an added integrat扭E effect
over the sens11Ive detector area As expected. there are certain lim1tat10ns to such an
aπangement

2.10.1 lmmers1on lenses

As the name 1mphes‘ the detector 1s immersed at the rear side of an optical element
1.e.. the detector 1s in direct coo阻ct with the lens surface. If the immersion lens is a
hemisphere. axial rays will not be refracted. Oblique rays‘ however. are bent as indicated
in Fig. 2.14. The reduced detector size can be shown to be d ’ = d'IN
 Cl!AP11'R l. BASIC O門res 31

N
‘

v戶
VUB
42d
吋’，伽

σ”Aa
P

..

.. f ’,

FIG. 2.14 Cone四位ichem呻here.

The rad1 山 of the hermsphere rs not critical. because of the concentric aπangement:
however‘ it should be of such a size that the angle between the extreme oblique ray and
the normal at the lens is not too steep. The focal pl祖e position is not affected by the
insertwn of the hemispherical immersion lens

The signal-tc• noise ratw 1s improved by a factor eq阻I to the index of refractwn N
of the immersion lens Therefore‘ the optical gain is N

Another type of 田unersion lens is the hyperhemisphere. The optical gain 、明th a
properly shaped hyperhemisphencal imme自ion lens is even higher.、 n缸nely N2 The
proper shape refe間 to the so-called aplanatic cond1tion. when aberrations 缸e elimmated
For such a condition to exist. the following relations m叫t be observed
川γ 川江

nkH

FLFL ++-
(2.8)
and
R
一

N (2.9 }

With these 甜lations、 there 叫 II be a displacement of the focal plane as can be seen
from Fig. 2.15ι
32 0月JCAL D目前N FUNDAMENTALS HIR [NF，制REDS"7回釘

.. ... . ..
..
．。

amsj mzlj

p一樹
E

Fι
n－
u－間
‘‘ R L.’
.,
l 、，
‘
FIG. 2.15 Hyperhenuspherical immersion l個＆

The optical gain for this aplaoatic case is equal to the sq·聞re of the refractive index.

2時orettcally. a 且也n of 42 ~ 16 cao be achieved with a germanium lens. However.

there is a practical limit (see Fig. 2.16) which relates to the relative ap白白lfe or ／／再. For
the aplaoatic condition of the third k姐d 晶 applied here‘ sin 11' ~ N s區 11. Therefore有 at the
limit when sin 11' = 1 、

sinu"""' = N ＝ 京有可 (2.10)

or

廿九三﹔JN1:i (2.

A compa吐son of four materials that range from low to high refractive index shows
that a careful approach is m order when a hyperbemtspherical immerston lens 1s being
considered to increase the 5月4

Matenal N if！＃）.曲（ Limit)

Glass l.5 。9
ZnSe 2.4 1.3
Si 3.4 l.8
Ge 4 勻
 CHAPTER 2. BASIC O!'f!CS 33

FIG.2.16 Alowfl#conemi扭曲曲e lens.

Because of the steep angles involved al the extreme rays. it is advisable to reduce
the hmit even further. A factor of two may be reasonable.

2.10.2 Light pipes

The 句＇Pe of light pipes discussed here are solid an挂 hollo、v cones that find their
applical1on in reducing the detector area. These cones are not imagmg elements ‘ they
concentrate the received energy from the target onto a smaller cross section In the
pro扭扭 they have an additional effect. the energy is mtegrated. compe由atmg for any
臥isting non-unifom1 問sponse across the detector ar間 Figure 2 17 shows the basic
a叮angement. Some possible variations of the light pipe are indicated in Fig. 2.18. Solid
cones with a flat or curved front surface are also suggested Further configurations are
p訂abolic reflectors and a combination of refractive 胡d reflective elements A more
detailed discusswn of such light pipes or cone conde拙ers can be found in Ref. 5.

FIG. 2.17 Reflective lighl pipe (cone condenser}

It is apparent that the optical gain. which JS the ratio between the entrance area of the
light pipe 回1d the detector area‘ is limited If the cone 血1gle of the pipe 1s too steep. the
entering>rav” will reverse its directwn and never reach the detector. In such a case. the
hght pipe acts as a retroretlector 6
34 （｝月1C4l DESIGN FUNDAMENTAL.I FOR INFRARED Srsr帥的

﹜叫必這

> 一月~
FIG. 2.18 Some possible shap間。flight pipes to 咀crease S別

2.10.3 Field lens

A lens placed at or near the focal plane of the objective lens can also serve to reduce
the needed detector size for a given field coverage. The power of the lens is chosen to
image the aperture stop of the objective 帥to the detector pla肘 This aπ叩gem間t has the
S阻ne integratmg d臼ect 田 the cone already discussed. The lens 也ed for the reduction of
the detector is called the .field lens. If placed exactly at the focus of the objective. it h阻 no
e班ect on the power of the o吋ective. For a high optical gam factor. the field lens consists
of several elements to co何自t for abe叮當io酪. From Fig. 2. 吟吟 it can be concluded that the
optical gain is the ratio between the are晶。f the field lens and 也e detector.

Objective

Field Lens

Detector

FIG. 2 19 日eld lens 1IDag＜目中 erture stop onto detector

2.11 Field of View for Staring Arrays

Frequently the field of view for staring focal plane a叮ays is expressed in honzontal
and vertical dimensions as mdicated m Fig 2.20. It m山t be remembered that from aa
optical standpomt. the total field dimension is the diagonal of the array. The extreme off-
axis dimension is the one that enters into the aberration exp扭扭ions It is also the extent
that needs to be clear of anv obstructi。”。
 CHAJ'fER 2. BASIC 0凹的 35

可－

H
•

...
v
..
、，

f
‘ 2u,
、，

,.

FIG. 2.20 Optical field of view d' ＝尚可？

References

I. R. Kingslake有 Optical Sys/em Design‘ Academic Press (1983 ）、 page 10.

2. Bergm叩n. Schaefer. Lehrb’,ch dei Experimenlalp句苛政， Band m、 F Mantossi‘ Op1站，
vierte Auflage. Walter De Gruyter & Co ( 1966 ）、 page 78.
3. W. Steel · L血ninosity. throughput or etendue？ 弋 Apphed Optics 13. (1974 ), page
704.
4. K. Rautsch, Die Op1ik in der Feinme.δ宮／echnik.可 Carl Hanser Verlag ( 1949）、 page 53.
5。 D E. Willi剖nsson. “Cone channel condenser optics·. J Opt. Soc. Am. 42. (1952 ）令
page712.
6. W. J. Smith, Modern Optical Enginee’”哼， McGraw-Hill‘ Inc. (2000 )‘ page 280.
CHAPTER 3

Primary Aberrations
3.1 Introduction

Jn the previous chapters. comments we甜 made relatmg to aberrations Now 1t is t阻ie
to discuss and explam them and find out how they a何eel the performance of an optical
system

In geneml呵 abermtions are deviations of 胡 image point from its ideal pos1t1on
These aberrations come in various forms. For this fundamental tutorial text we limit
ourselves to the so-called prim由y abe叮ations·

• Spherical abe叮ation
•Coma
• Astigmatism
• Field curvature
• Distortion
• Axial chromatic aberral1011
• Lateral chromatic aberration.

These seven aberral!ons will first be descnbed and then expressions will be
presented to be used for <letenmning the q岫nt1tative impact of each on image quality.

3.2 Primary Aberrations

One can sep目-ate these aberrations by general categones. The first five types listed
deal 叫th monochromatic radiation ; the last two address polychromatic effects
Sometunes the categories are split between on-axis and off-axis aberrations. By 吐iis
defmit1on. sphencal and 缸ial chromal!c aberrations 缸e on-axis abermt10us b間 ause they
refer to object points located on the optical axis (the system's axis of s·戶nmetry ). The rest
are oft:axis abeπ·ations.

3.2.1 Spherical aherration

’
Applying Snell s law across a spherical surface of a lens or a mirror shows that rays
closer to the edge of such an element are bent more strongly than they should to meet with
rays closer to the center of the element at the optical axis crossover point of these rays.
tlo"ever‘ because spherical surfaces are much easier to manufacture th叩 asphenc ones‘
they are by far the more standard surface to be found in optical systems.

Figure 3. 1 indicates that the marginal axial ray‘ after passing through the lens
intercepts the optical axis at point M. The ray offset from the opl!cal 扭扭 by an
infinitesunal amotmt (a ray in the paraxial reg凹的 m值的自ts the axis at P. If the rays enter

37
38 OPT/CAL D目的N FUND品仔NTALS FOR INFRAREDSrsr.品IS

the lens parallel to the ax阻， P is the focal point of the lens. the location of 也e paraxial
image plaoe Deviatmns from the par祖祖l image plaoe are the me酷ure of spherical
aberration. According to this statement the distaoce PM in Fi皂. 3.1 is the longitudinal
sp/1e1 ical abe1rat10n.

Matginal

Para珀al ~ p
σ→

FIG 制 3.1 Spherical aberration ts the variation of focus with aperture.

3.2.2 Coma

Figure 3.2 shows 棚。 oblique rays. A 胡d B ‘ entering the edge of the lens 阻 part of
a bundle of light. parallel to the principal ray PP . Rays A aod B cross the image plane at
a different height thao the principal ray. The di好erence. h'm h'r• is a measure of coma.

p
"f
。
A
• 、，h'
A,B m h',
p Optical axis " "
B

FIG. 3.2 Coma is the variation of magnification wi血 aperture.

The unage pattern formed by 個cing many rays of the entering bundle t尬。u掛 the
lens is shown 詛 Fig. 3.3. TI1e optical axis and the principal ray of the system are
identified by PO aod PP‘間spectively. The construction of the 凶1age blur shape is
p間sented by the sequence （的 thro略h (e) TI1e circle marked with point pairs l一［， 2 2‘
3-3. and ι4 is the exit pupil of the lens. Rays entering 也e lens. parallel to the principal
間y‘ passing through these identified points are focused in the 世祖ge plane 品 po咀個 1 、 2.
3. aod 4. forming the outline of the image blur 也at resembles the shape of a comet. hence
the name coma. The plane that contams pomts I from the e、“ pupil and the principal ray
PP is called the taogential plane The plane perpendicular to the tangential plane.
containing the principal ray and exit pupil points 2 is the sa星ittal plane. The image pattern
[Fi呂 3.3 （η］ identifies the d間taoce P 2 品閥割位al coma and P I as tangential coma. As
noted. the tangential coma 1s three tim間 the size of the sagittal coma.
 CHAPTER3. P恥地ARY ABERRATIONS 39

。p 。p

甜甜d但i my 。2

。1
。 。
4 。。 4 。。
3 3
2 2。
。p Opti回l 阻is 。p
2 。2
3 3
。
4 。 4

（吋 T胡8間tial rays I {b) Saginal rays 2

。p 。p
2 2
3。 3 ，。 4
l
4 。
3 。。 4
3 。。
。
。
2 P, 2 p
。
、 2
3。 3
。
4 。4

(c} Skew rays l (d) Skew rays 4

"i＇。心.、。p 4
p
‘
呵，

3 ·J<
a

4。
。 l。。 ‘ ~祖
om星a回（~~C)
3 A
。
2。
p
。 3
。 2
3。
。 。4 。
嗯，

(c) Coma image 但由rn outline P413 (I) Coma blur spotshape

FIG. 3. 3 Construction of coma bl叮 spot

3.2.3 Astigmatism

Astigmatism is best described with a llu間，dimer泌的nal p時恤re. Figure 3 .4 identifies

two planes that are pe中en也cular to each other. The plane that contains the object point
and the opl!cal 阻扭扭 called the meridional or tangentrnl plane. The rays in 也is plane
fonn the tangenttal fan. The plane perpendicular to 血e tangential plane、 contaming the
object pomt、 is the sag1ttal plane.

On the image side there is a longitudinal separation between the tangential and
sagittal images. The images of an object point 訂e fonned in two different planes. Because
40 0.阿"JCAL DESIGN FUNρ＇AMENTAL' FOR INFRARED Sr;nMS

of this. the sagit阻l and tan~ential ima惡es of the object point 訂e lines. Approximately
halfway between these tw。 image Imes lies 也e so-called circle of least confusion. a blur
with the smallest linear dimension 晶 its diameter.

門︱叫i
image line
Tangential
image line - 'Ci1℃leof
least confusion

Tangential fan

.:cs! ~,5-, -
o~’，’
Sagittal fan

Object point
-
FIG. 3.4 Astigmatism.

3.2.4 Field curvature

Even if sphencal aberrat10n. coma. and as討gmaltsm were absent. the nnage of an
off-a.xis object point would still not lie in a plane 叫thout coπ·ection for field curvature.
Assummg the first 伽ee aberrattons were eliminated. the image surface would be
parabolic in shape. Dependin~ on the power of the lens‘ the parabola is either curved
toward the lens or away from it. Positive lenses { positive-power lenses) have an inward-
curved image surface (toward the lens) Negative lenses (negative-power lenses) have
their image surface bent in the other d1rect10n. Tlus basic 且eld curvature is named after
Josef M阻 Petzval. who discovered the behavior of astigmatism relative to field
curva制間可 Petzval was a Hungarian mathematician who d間d heartbroken a缸er b叮glars
destroyed much of the mannscnpt in which he had recorded. among other findings. his
denvations of aberration coefficients up to the seventh order 1 The nnage surfaces and the
Petzval sur詢問叮e identified in Fig. 3.5. Measured p前allel to the optical axis. for the
single element shown. the tangential image surface is three times farther away from the
Petzval surface than the sagittal tmage surface. If ast1gmat1sm is eliminated. the tangentrnl
and sagittal image surfaces coincide and lie on the Pe包val surface.
 CHAP 距 R3 P虹MARY ABERRATIONS 4!

。L

p缸祖XJal

Tangential .. A ‘ image plane

扭扭 ge surfaces Sagi加I •
Petzval

FJG 3.5 Field curvature 阻d Pelzval surface.

3.2.5 Distortion

Distort10n ts the variat10n in the magrufication of the optical system with field angle.
If the magmficatton increases with 也e field, pincushion distortion results If it decreases
baπel distortion is present. This is depicted in Fig. 3.6. Pincushion d阻tortion is called
positive. barrel d1stort1on ts negattve.

Pincushion (positive) Bar祖l (negative)

‘ •

‘.
Undistorted unage

FIG. 3.6 Disto甜on.

3.2.6 Axial chromatic aberratmn

As state昌 m Chapter 2 under Snell's law‘ the amount of ray bending depends on the
index of refraction. Because the index of refract10n ts not constant but varies 叫th
wavelen~th。 a lens foc阻es rays of different colors at diffe闊別 places alon~ the optical
axis. This dill扯開ce or spread between the two foe阻 pos1ttons ts called longitudmal 眼阻l
chromatic aberrati011 For a smgle positive lens. the short－吼叫•elength focus ts closest to
the lens (see Fig. 3. 7).
42 OPTICAL DESIGN FUNDAMENTALS FOR INF，凡<RED Sf昨日的

Short λfoe田 Longλfoe閏

縛，
。。

’, 4
Longi恥dinal 阻i_al
chromatic aberratton

FIG. 3.7 Longitudmal a.xial chromatic aberration

3.2. 7 Lateral chromatic aberration

The separation of an ima_ged off-axis point. formed by 悶1ys of different colo風 is

called lateral chromatic aberratton or lateral color It is a consequence of the difference in
magnification with wavelength. Figure 3.8 11lustrates thts sin圓hon.
..US 問
尚
ah、
3h
nr
B

Lateral
’•
a內人

color
b間
肘
、

Image
pl個e

FIG. 3.8 Lateral chromatic aberration or la扭ral color.

3.3 Calculations of PrimaηAberrations

To determine 也e abeπations described. we apply expressions developed with the

so-called Seidel aberration th即可 and neglect the e釘ects of the thickness of the lens.'
Seidel's theory is also named third-ord叮 theory because it is based on the truncation of
the series exp個sion for the sine of 祖祖gle after the second term. as indicated in
Eq. (3. l ).
x刊 5
XI Xλ 7
smx=x 一一斗＋一一一 +.. (3.1)
3!1 5! 7!
The power of a lens 1s the reciprocal of its focal length. expres冶ed as

(3.2)

、吋1ere R1 胡卓 R， 叮e the radii of the fir訕訕d second surfaces respectively‘ t is the

thickness of the lens. and N is the refractive index of the lens material as indicated in Fig.
3.9. The radius 阻 positive if its center lies to 也e right of its surface: if it's to the left. it is
negative. Therefore‘ R, in Fig. 3.9 站 positive: R2 is negative.
 cw叮叮"ER 3. PR必1ARY ABERRATIONS -13

l 2

R,
R.
。 。

.. I 司

FIG. 3.9 The thick lens.

By neglecting the thickr時間 t of the lens司 Eq. (3.2) reduces to

中＝ (N 吋 l)(c：一 c,)= (N-l)c (3.3)

where c1 and c2 are the c凹vatures or reciprocal radii of the lens surfaces c is called the
net or to叫 C叮vature of the lens Equa訂閱（ 3.3) states that 晶 long as the net C叮vature is
kept constant. the surface curva個res c個 V由7 咐t】1out changing the power of the lens.
This changing of the lens shape is called lens bending. It is a powerful basic tool in lens
design. because the shape of the lens affects sphencal aber官司ti on 祖d coma but not its
focal length. The depend聞聞。f spherical aberration on the shape for a gennanium lens
(N = 4) ts illustrated m Ftg 3.10‘ where sphencal abeπatton 1s plotted wtth 間spect to the
ratio of the first surface C凹va個re and the net curvature. cifc. This ratio is called the shape
factor K.

..
Spherical abe甘a II on

” 2 -I 2 3 4 5 Shape factor K

’,

R、－ 1.5/ - 3f ∞ 3f 1 旦f f 0.75{ 0.6f

R、 -f -1.5! -3! 由
3/ l.5/ f 可 .75/

/
FIG. 3.10 Change of spherical aberration WI 血 change of lens shape (lens bending) for a
gennaniun> ions 、明th index N = 4. l\咀血mum spherical aberration 尬。btamed in this case wheu
tl1e first radius of the lens is eq阻l to its focal length 祖d 曲e second radius 時 1.5 times the focal
length (K = 3).
44 0阿拉<L DESIGN FUND.必1E間：：＜LS FOR INFRARED Sr. fEMS

3.3.l Spberical aberration

The ang旭lar blur spot size in radians ts

。 B N' - (2N + l)K + [(N + 2) !N]K' …

咖－ f - 32(N 1)2 （／紛）、，，，可

、社1ere N is tile index of refraction of the lens material and K the shape factor discussed
above . .fis 也e focal length. 組d f I# is the relative apertur當

Tite best shape for a singlet is obtained by differentiating Eq. (3.4 ）胡d solving for
K. In doing so‘ one finds

R}_ 2(N + 2)(N- l) f. (3.5)

- N(2N +I)
and
R， 叫 2(N + 2)(N- l) f.
白 7

(3.6)
N(2N - I）一 4 J

A lens shaped 扭扭路的ted by these eq閏tions e油ibi臼 a minimum 祖gular blur spot
size of

p N(4N-I)
(3.7)
帶＂＇＂·＂ - 128(N-1}2(N + 2)(/1#)3

The mmimurn hnear blur spot B is located a distance ii. away from the pa阻xi al
image plane (see Fig. 3.1 l ).

Minimum p缸直到al
blur spot image pl聞e

B
司，

。 一一--0 .,,
.I.
TSC
a巴
.. /j,,
.,
.. SC
’,

FIG. 3.11 Bl凹叩otdue t。可herical aberration.

 CHAPl田 3. PRIMARY ABeRRATJONS 45

B= 0.5 1口百C. (TSC is

the transve目e spherical
aberration contribution)
t.= 0.75 SC, (SC is the
Jongitudmal sphencal aber-
ration contributton).

It is interesung to note
tl1at the relations expressed
in Eqs. (3.5 ）阻d (3.6) we阻
already stated in 1762 by
the Swiss mathematician
Leona叫 Euler.

It is also inte阻sting to
observe that some countries
honor their scientists by
placing portraits of them on
their bank notes with some
remarks 間lating to the
scientist's achievem缸Jt

Leonard Euler is
pict叮ed on the Swiss
SFr JO bank note with a
叮ain of lenses.

Karl Friedrich Gauss

(1777-1855), the German
ma 出emaucian and astrono-
mer who was 也e fi目tto use
the paraxial-ray approx-
irr祖tion, is pictured on tlte
German DM JO bank note
with his distribution c旺ve.
Unfortuna阻ly, 自治 will
change when 也e E叮0，血e
cu何ency of the European
Union’的 introduced.

Equations (3.5 ){3.7)

are applied in Table 3.1 for
a few singlets made from
出e most common infrared
materials and compared to
a glass lens. For simplici旬，
the refractive indices shown
are rounded off.
46 0阿JC4L DESIGN FUND.訓苦悶.ALS FOR INFRARED Sr.盯EJ.!S

Tabl叫.1 M血im阻E 叩＇he話叫 blur 可ot sizes for s祖glets made fro血 d逝erent materials

Material N R, R, F叩... (radi阻s)

Glass 1.5 。 583/ 3.5/ 。 0670 (f ／恥之

Zinc Selenide 2.4 0.885/ 2.40~｛ 0.0187 (f/#) J

Silicon 3.4 。.977/ 1.649/ 0.0108 (f/#f3

Gennamum 4.0 !.O/ qr 0.0087 (f/#f3

It is e圓y to remember that 也e best-shaped thin germaniwn lens‘ ( shaped for

minimwn sphe吋cal aberrat泌的 h且 a front radius eq＇閏i to its fo叫 length 朋d a rear radius
that is 1.5 甜甜es its focal Ieng油（see also Fig. 3.10).

From the denomma個r of Eq (3.6) it ts app前·ent that for a conv臼 piano lens
（他旦開L N(2N I) = 4. This m間ns that the index for such a convex piano lens is
N = 1. 686. That is the index of 目.pp凶m at 3. 7 µm. The front radius of such a lens would
be 0.686/ and the 個gular bl前 spot p恥當眾 0.0436 (f /#) J 個dians. Awarer時ss of these
sunple relations c曲。由：en save 御風 money, and di臼.ppoin恤目前．

We will now determine the rem剖mug Seidel aberrattons for a lens shaped for
minimwn spherical aberration

3.3.2 Coma

As already mentioned and repea阻d in here Fig. 3.12同 the third-order coma patch
resembles the shape of a comet. Even though sagi血l coma CC is only one third in size of
the 旭ugential coma. the triangle covered by the sagittal dimension contains about 55% of
all the en前gy from the imaged object point.

.. .
cc
.
3CC

.
FIG. 3.12 Third-order coma patch

If we use the 血nount of sagtttal coma as a measure of the blur spot司 its angular
extent can be expressed by
 CHA間"ER 3. PRIMARY AIJERRAflONS 47

告司 cc －一＿＿＿＿！！＿丘 (3.8)
曲曲叫－／叮 16( N + 2)(/1#)1

叫1ere 2年 is the angle of the principal ray 伽 the specified off 臨時 po叫 Notice Eq. (3.8)
ts for a thin lens. shaped for minimum spherical aberratton

3.3.3 Astigmatism

The tangential and 閥gittai image 咖Us a間呻ara叫 by 2AC ~ 11~ f. The circle of
least confusion is located halfway between (see Fig. 3.13) and has a diameter of

B=2TAC=ic“.,v. (3.9)

Dividing this by the focal length of the lens. the ang叫叮 blnr spot due to astigmatism
is

。 B u; (3.10)
甜－ f - 2(1 !#)
Sagi位al
Image ,.
;rJC
,,

恤ge ~勻 ..
B
" ...
;JC
.
FIG. 3.13 Astigmatism. minimum blur spot size aod location.

3.3.4 Field curvature

The Petzval surface (Fig. 3.14) ts acrnally parabolic in shape 祖d is expressed as

Petzval contnbution PC hv
勻的

2，白 f
PC=-2f;r 守（ 3.1 I }

fhe verte.、 radius of lhts parabola is called the Petzval Radt叫 p Its measure for a
thin lens is
p= NJ (3.11)
48 0f'I'JCAL DESIGN FUND.肘，E用ALS FOR INFRARED Sr>TEMS

PC

可
N

UPA
p

Petzval I I p甜可oxial
Surface ' I Image Plane

FIG. 3.14 inward-curving Petzval surface for positive sin到前

3.3.5 Astigmatism and Held en鬥ature combined

F1gure 3 15 indicates 也e relative locat10閻部non皂白e astigmaltsm shells. the Petzval

surface and the p缸缸ial i血age plane at some off－阻.is point 恆 the image. Notice that ev叩
if astigmatism is elnnmated. which of co回祖 reqmres more than 1ust a single element.
field curvature will still exist On the other hand. if the Petzval curvature were zero. 1. e .
coincident 吋th the par缸ial 泊age pl個e. astigmatism would not be automatically
corrected It is important to understand the interτelat10n b閻明問 astigmattsm and field
curvature to 扭扭ss their contnbut抽n to 1ma喜e degradation

司，

a‘

A.

FIG. 3.15 Astigmalism 剝ld field curvature.

 cw心，TER) P阻恥、RY ABE阻lATIONS 49

3 3.6 Axial chromatic aberration

As st叫ed in Eq. ( 3.3 ). the power or inverse focal len且th of a thin lens is
t 立（ N-1 Sii 1d岫叫action va丹y 叫川ve

λ，r的pective
li 、＇ I J,
the limits of the spectral window under consideration. then the difference of the two focal
lengths •.；九三只 is the longitudinal 由ml chromatic aberrat1011 LchC 品 identified in Figs 3.7
and 3.1 后， Forming this difference y間Ids

N‘一 N時
LchC = f, - .f. 三于甘亡于fι (3.13)

N -I
is called inverse relative dispersion or Abbe number V, named after
N, N可

Ernst Abbe (1840一 1905 ）‘ chief optical designer and partner of Carl Zeiss司 Germany. For
the visible spectrnm. the chosen wavelengths for the Abbe nu血ber 缸·e 0.5876 ！岫向
0.4861 仙也 and 0.6563 µm. These are identified 品 F raunhofer absorption lines d. F, and
C (yellow, blue-green. and red). At 血e age of27 ye訂s, Josef von Fraunhofer discovered
and cataloged several hundred of these absorpllon lines from the sun spec甘um He
classified the stronger absorbers with capital letters. 3

With llus司 the Abbe number for the visible spectrum is stated by
N

扎一N
一凡
d一－
、

V
可

(3.14)
paL
r

For the infrared regmn. the Abbe number has been modified to read

N .. -1
v ＝一一－－一一一 (3.15)
N5-Ni.

where 孔1. S. and L stand for middle‘ short‘ and long "avelengths.

Gomg back to Eq. (3 13). we c品1 now write

lchC=[_ (3.16 }
v

Figure 3.16 sho、vs the relation between longitudinal a、 ial chromatic abe叮割的n LchC
and the blur spot diameter B‘ wluch is approxnnately equal to the transverse axial
chromatic aberr剖ion TAchC. With 11' = l/[2(/ ／＃）﹞、 the measure of the blur spot diameter
becomes
50 OPTICAL DESIGN FUND.品館前'.ALS FOR INFRARED SrSTEMS

LchC j
B = LchC 11' = ＝一一一一… (3.17)
2(fi#) 2v(Ji#}

The angul缸 blur spot is 也en

戶一 B _ I (3.18)
一叫一一－－－－一一一－
=
clrn
f 2v(Ji#)

• B 司，

u’
、， TAchC
' 。
F血 λ
‘
a
‘
4

LchC 、，
‘
FIG. 3.16 Longi如dinal 曲d transverse 阻1叫 chromatic aberrations { lchC and TAchC
呵呵ecttvely). B 站出e minimum blur spot (circle ofleast confnsion).

For the sake of completeness it must be menttoned that for a thin lens. with the stop
at the lens‘ third” order lateral c滑lor 由id d1stort1on are zero

3.3.7 Numerical example

Let us look at a thin singlet made from silicon that is used in the 3- to 5 －~Lnl region.
The chosen focal length is I 00 mm‘ the lens diameter is 50 mm‘ and the half field to be
covered is 0.1rad(about5.7 deg).

For the chosen spectral range. N = NM = 3 -1254. and "= 235 5. (These data can
vary from one material supplier to another).

Applying the equations presented ‘ beginning with Eq. (3.5 )‘ yields:

Tue radii for the lens with minimum spherical aberratton are R1 = 97 9 mm. and R, =
164.1 mm. i e. a memscus-shaped element. The Petzval rad1間 p = 342.5 mm The
negative S1gn indicates an mward curved field. i e ‘ one concave in shape. The angular
aberrat10n blurs amount to
詩句＂~ = 1 33 mrad
13,.,,, = 0 29
13,,,, = 2 50
戶d=m= 1 06

~13 = 5.18 mrad.

The linear blur spot size B = /rj3 = 0.518 mm.

 CHA凹CR J. PRIMARY ABfoRRATIONS 5]

For comparison and validation of the concept of third-order thin lens c叫culatio阻句
this lens w曲目ven a 也ickness of 5 mm and its shape w晶。pt卸tlzed for the given
specification. The result is summarized 阻 Fig. 3 .17. Rays uniformly distributed across the
e甜甜ce pupil form an image pattern (s~ot diagr珊的臨 they penetrate the iα1age pl胡e.
The 扭扭曲d shape of 血e spot diagram mdicates how well the lens c祖 concen甘ate the
tr缸回mitted energy‘ which is of course a measure of the aberrations present. To obtain the
spot diagrams shown for our simple singlet, several hundred rays were traced for each
object point. The surrounding box size in the di咚rams w祖 chosen to be 0.518 mm. the
size of the third-order blur spot. We can see that the thin lens approximations compare
very well indeed wi也 the full scale ~eome甘ic ray tracing results. To better demonstrate
the presence of field c回vature for a smgle lens element, sphencal abeπatton 、咽sremoved
by aspherizing the front surface. and monochromatic radiation was 甜甜med. Figure 3.18
shows the result; one can see dist泊ctly the formation of the astigmatic image lines
Depending upon the application司 one may W剖1t to "'eq阻l垃晦” the spot size somewhat by
refocusing the system as it w晶 done for Fig. 3.17. The principle is ill凶甜甜ed in Fig. 3.19.

(
0 0'190 OEG

Li ····I -? I
l
'L•t I;
川斗「
,0, I
巴

FW

I ": ,,,,I
阻

,
阻

a
、

q E
四

內句口

ls.7個殉口自 且一 、一
j

也 200 -100 白 100 200

且l!Rffi豆豆午4且旦
τHROυCH FOCUS SPOγOIAGRAH
OPτIHIZEO S 工NGLET 3 17
SPOT S工ZEυNITS ARE 州工CRONS.
FIELD I 2 3
附E RAllJLIS ' 181. 咕3 101 3ll2 127.526
GEO R聞且陷 a 2間， 908 186.377 3呵。咱自
BDX UIDlH 518 經OFERENCE ：由IEF RA1

FIG. 3.17 Optimized smglet with added lhickness.

52 Ol'T/CAL DESIGN FCJND.品巫師＞ LS FOR INFRARED SYS7日佑

...回閉目EG

可 .BeH 臨G .,+
‘“官們 g魚
: l ，也

[J
S. 70MJ os;
︱心（ 姆
JS阿

、，

-2000 -1000 的 1000 2000

SURFACE' IMA
γHR 口UGH FOCUS SPOT 0工AGRA”
SINGLET, FIEL口 CURυRτυRE. 3.18
SP口 T SIZE UNITS ARE M工 CRONS.
吹起LD 2 3
間§他司。I出
哇由。
X 即司DIUS ' 8密岫
79~ m. 呵”呵
68 ~＇咽＇•。1仔
3目
掛E限章，CE
WIDTH 16個： Cl任Ef 間Y

FIG 3.18 Presence offield c凹vature 胡晶 astigmatism for lens .

，心
u,

.,.

A吐justed p缸寇珀al
Image Plane Image Plane

FIG. 3.19 Image position adjustment to compensate for blur spot size variation.
 er恤l'TER 3. PRlMAR Y AB血RATlONS 53

3.4 General Aberration Correction 研fetbods

A very simple yet powerful ~pproach to correc也g ofι削s aberrations 叫也out

affecting spherical and axial chromatic aberrations 1s to place the aperture stop 恥咽y from
the lens. as shown in Fig. 3.20. The benefit of this method is achleved by cor啦。Hing the
si甜甜d location of 也e ray bundles forming the image. Shifting the stop a distance from
the lens does司 however‘ reqmre an increase in the lens diameter可。＞lhenvise‘ VI伊e性旭E
occ阻. Vignetting is the failure of an oblique ray bundle to fill the aperture stop.
Relatively simple expressions have been developed to 甜甜甜 the e叮ects of moving the
aperture stop along the optical 悶ts. These expre咽。瞄缸e called stop sht丹 equations and
are treated in de阻ii in Ref. 4.

一。

一寸

FIG. 3.20 Position 祖位 size of aperture stop control off－阻is bundle.

Be治泌的 lens bending and proper positioning of the apert凹e stop, there 訂e other
general methods to be applied for aberration correction or bal祖cing. Spli社區g a lens
into more elements.、由phenzmg s凹faces‘ carefully selectm皂白e most suitable lens
matenals w1自 regard to mdex and dispersion, combining refractive elements with
rellective or diffractive components-all a扭扭cbniques for improving image quah可
Some configuratmns that employ these techniques will be discussed in a ftmdamental
fashion. in line with the goal of thls 個tonal text

3.5 Doublets

Doublets 肛e defined 盟 two elements 咀 close proxnni紗， For the visible spectrum,
doublets are frequently cemented together. Because optical cement is not transmissive in
the m企ared region, the elements 訂e ah咽ys separated from each other The separation of
the elements makes mountmg somewhat more difficult, but one more degree of freedom is
gained by being able to bend each element independently.

Let us look at some examples.

3.5.1 Two elements, same material

From the man) combmations possible. 、λ·e look m detail at the

a盯angement where each element is bent for rninimun1 s~herical aberration
and made from the same material As before. the object is considered to be
located at infmi紗， There 1s no a'<ial gap between the two thin elements and
each element has half the power of the doublet. Tlus means that the focal
length of each element is twice 田 long as the focal lengthfofthe doublet.
54 OPTICAL DESIGN FrIND.鈴聲NTALSFOR 卸F創REDSl'S1~日，s

With these 晶S阻nptions and applying the thin lens third-order theory'. the radii of the
elements can be stated bv

R 4(N +2l(N 一 1) f
一，

I - N(2N +I) J
(3 ‘ 19)

+2l(N 一 I) -
- - 4(N
"2 - N(2N -1) 一 4 J
' (3.20)

R 4(N +2)(N-1) f
一，
(3.21)
, - N(2N+D+4(N2-i)J

4(N+2)(N 一 I)
個d ’，一 τ
(3.22)
叫一 N(2N 一 1l +4(N2 -2) J

The angular bl叮 spot size for 也is doublet configuration caused by sphe吋cal
aberration is

戶＿ N[5+4N(N -3)]
(3.23)
帶•~• .,, - 512(N-1)2(N+2)(/!#)3

lnspectin甚 Eq. (3.23) reveals that the spherical aberration vanishes if N = 2.5.
Figure 3.21 shows the dependency ofj3柏前 on the material index N.

TI1e angular coma blur spot s阻e IS

13 (3.24)

The reversal at N = 2.5 indicates the change from tmdercorrected to overcorrected

spherical aberration Undercoπ·ected me祖s that the minimum blur spot lies inside the
p由axial focal plane For the overcorrected c晶弘 the location of the blur is outside the
focal plane. TI世s reversal indicate治 also that doublets made from matenal with 叩 index of
refraction greater than 2.5 c祖 be corrected for spherical aberration by properly bending
the elements.
 CI!Al月TIR3 P阻MARY ABERRATION' 55

~ spb叮 PcoIDJI
A A

。。 I 0 (/1#)"3 0.02 u, (/1#)"2

’, 0.0116
0.005

0.0022 ‘
。 ’, N

1.5 2.4 3.4 4.0

Glass Zn Se Si Ge

FIG. 3.21 Spherical 曲enatlon 祖d co血a as a fur叫100 oftbe mdex The example shows 血e
coefficients for a s1hcon doublet. The abe叮ations are expressed in radi 四S

It is an educational exercise to develop expressions for a specific configuration. For

ex剖nple. two elements. equal in shape句 bent 品 a combmation for mmimum spherical
aberration may be of econonuc interest. Another doublet form would be two conv缸”
plano elements This is a partic叫arly mteresting exercise because 1t turns out 也at there is
no gain with regard to spherical aberrat10n. uoless the material chosen has an mdex of
refi百ction lower than 2. For IR applicat10ns. sapphire fits mto this category.

3.5.2 Two elements, different matermls

To find a reasonable starting point for a doublet with two different materials‘ we
assmne that the surface radi啦。f the first element ( elen祖nt A) is eq閏l but opposite in
sign lo the first surface radi阻 The first or front radius of the second element (element B)
is equal to the second radius of element A Finall ）九 the 間ar s叮face of element B is a
piano as shown in Fig. 3.22.
1 2 3 4

NA N"

“ / .’,

FIG. 3.22 Simple starting configuration for doublet with two dil!ercnt materials.
56 0阿1CAL DESIGN F凹的"1>fENTALS FOR 卸F創REDSl且凹的

Exp間ssing the above s個tement mathemaltcally i聞ds to R, = R, = R1 ， 胡d R, ＝目的

Wi曲 that

R, =(2NA -NB 一 l)f (3.25)

The focal length of the front element A is

f R1 (2NA-NB 一 I) r
(3.26)
A - 2(NA -1）一 2(NA 一 1) J

個d for the 間ar element B. the focal length 1s

天－ ＿＿！！＿」一 （2N A- NB -1) (3.27)

一一 一一
(N8 -1) (N8 一 1)

3.5.3 The acbromat

AB
A thin acbromat c叩 be corrected for third-order spherical aberration ‘
coma. and axial chromatic abe叮atton The combination consists of a positive
祖d a negative element. To eliminate axial c尬。matic aberration. a simple
relation between powers and Abbe nwnbers of the two elements must be met
This relat10n 1s

中 VA
4 一一一申 (3.28)
(v ，一 Vn)

And司 since 中＝申A ＋ 申B

中一 VB (3.29)
一－
" (v.-vA)
The process of deriving the complete prescription for such a doublet is
S甘aigbtforward but too cwnbersome to be included here 雪

The radii as a factor of an achromafs focal length arc listed in Table 3.2 for two
objectives: one for the MWIR 祖d one for the LWIR region. For the MWIR le他 silicon
and germanium have been the materials chosen. and the combination Amtir-1/zinc sulfide
was selected for the LWIR objecti珊， Amtir is 祖 a叮Oll"戶n for amorphous material
transmitting infrared radiation. The composition of Amtir-1 is 33% Ge. 12% As. and
55% Se.
 CHM＇阻R 3. PRIMARY ABE則lATIONS 57

Table 3.2 Radii oflens elements for two selected achro血ats,

Spectral region MWffi LWffi

(3 5 µrn) (8 12 µrn)

Front element Silicon Amtrr-1

R1 。.97[ If

R、 3.25.f ~f

Rear element Zinc sulfide

R, 4.f 一6.f

R, 2.f 24f

These chmces 訂c a sound startmg pomt that lead quickly to good sol叫ions m teT!lls
of optimization 咐th the computer after adding thicknesses and spacings.

3.6 Two ’'fbin Air-Spaced Elements

Separating the two elements provides 祖 additional degree of freedom tlmt can be
applied for aberration correction or lo detem1ine the basic configuration of the sys扭m
The Petzval objective、 telephoto lenses. telescopes 司 and eyep阻ces 訂e all examples that
belong 岫 tins catego丹！

Much can be learned and a good starting point can be obtained by modeling these
arran且ementswith separated thm lenses.

With a few judgments and assumptions司 closed-foT!ll solutions for the elements
shape c胡 be derived for some config闊前ions Sunple expressions provide a quick answer
m the layout stages for trade-off evalna”。n 師10ng different system cand1da扭S

3.6.l The Petzval objective

As show羽 in Fig. 3.23. this lens consists in pnnciple of two elements. spaced by
approximately the focal length/of the lens.
58 Of'T/CAL DE;1GN fllNDAME肝，，S FOR INFRARED 品古TEt.1S

,\

.. f 、， f/2 ‘
.. 2/
..
FIG.3.23 B值ic configuration of the Petzval objective.

The focal length of the 仕ont element is t、\ice that of the lens; the focal length of the
rear element is equal to the focal length of the lens. We fur誼ier stipulate that the elements
are individually shaped 伽 minimum s.pherical aberration and made from the s酬e
material. Again. applying the thin lens third-order abe叮副ion equations. the fo叮 radn for
this o吋 ecllve can be denved.

R 4(N+2)(N-ll f.
一于 (3.30)
1一 （ 2N+DN J

R1 4(N +2)(N-])
= N(2N-ll-4 f 有 (3.3 l)

2(N +2HN 一 I)
w
"3 -
一－
N(6N + D 一 4 J• (3.32)
and
R 2(N +2l(N-D f.
一于
(3.33)
'- N(6N 一 D-8 J

The 扭扭lar blur spot due to spherical abe叮ntion is

。一 N ﹝4N(2N 一 7 ）十 11﹞ (3.34)

中出一 2048 ( N -1)2 (N + 2)(//#)'

Axial chromatic aberration amounts to

。 cbmm 一付高 (3.35)

Based on these equat10ns‘ a germanium o均ect1ve with 100 mm focal length and a
relative aperture of/ I I ‘山ed for the 心 to 12-µm re~ion. would yield 祖間，axis blur spot
size diameter of 0 135 mm. A very goo拉到缸tmg point mdeed. if one considers that this is
 CllAP'而民 3 p阻MARY ABERRATIONS 59

only several tunes the size of the diffraction limit In Chapter 6 it will be shown how to
reduce the spherical aben冒犯on by changing one surface from spherical to aspherical. and
by adding a diffractive ph品e profile to one surface to lessen the chromatic aberrations of
thi間 objective.

Closed-form equations can be also developed for the off•阻is aberrations by

apply油E the same general proced叮t that was used to de討ve Eqs. (3.34) and (3.35).
However‘ they become very lengthy and 叮e no longer simple expressions. Therefore.
they will not be included.

3.6.2 Refractive beam expanders

The two basic configurations 阻ed for these afocal sys祖ms are the Galilean and the
Keplerian telescope types. shown in Figs. 3 24 and 3 25.

B
A 車了

~
y‘一’，
’d
..
, 0
D
'
呵，

‘ IA
（語）
.... .,

‘ In
（吋
..
FIG. 3.24 Galilean beam expander

可F
A .ii
.....
y
－~ D
d
。 ’
A 一’，
y

‘
IA
（吋
’ .. !B
（吋
’,

.,
‘
FIG. 3.25 Kepierian beam exp叩der.

Smee each of the two elements is suffering from spherical aberration‘ the exiting
beam ts not precisely collunated The marginal ray converge這 w1der an angle y 旺 each
element 1s shaped individually for minim凹n sphencal aberration, this convergence angle
is
60 0阿1CAL DESIGN FUND.訓E問 ＜LS FOR INFRA甜D SnTEMs

y= N （朴 I ）「 m一 ll (3.36)
64(N 一 1)2(N +2)(J紛Y L m J
N is the index of refracti閱（both elements 前e from the same material）.η1e
magnification of the system m 咱們pressed by the ratio of the ex1出s and entering be咽
diameters D 祖d d. In Eq. ~3.36). m is positive for the Galilean system Beca阻c of the
image reversal, it is negative for 也e Kepleri扭曲nfiguration It is apparent 胸前 the
convergence for the Galilean system is less than for the Keplerian. because the negative
first element is affiicted with over-corrected spbencal abe叮ation司 compensating somewhat
for the larger 由noun! of under” corrected spherical aberration from 血e positive second
element It will be noticed that the factor before the sqnare bracket in Eq. (3 36) is twice
the value of the minimum 血1gular blur spot s凶 for a single thin lens due to spherical
aberration. The relative ape削ef ／＃話 fr;ID ~ IJA IID，阻d the separation of the elements
I= fA ＋舟， For a 皂白manium be棚 expander 叫th 4.f紹說 2. m = 5呵 the convergence N=
angles are Yo划＝一 I 7 mrad. 間d"fK句＝一2.6 mrad

The fact that a negative lens has over-co叮ected spherical aberration. 晶 mentioned
above. p自宜uts us to bend 也e negative entr苗時e element of a Galilean beam expander m
such a way that it will compensa祖 completely for the under-corrected sphencal aberration
of the positive exit lens.

lf the positive element B is shaped for mi血血山n spherical aberration. its radii 叮e

-)
、川斗／－B
N 一－

+- NBm --
7M 一

I
RM－’，

R rfu
miu
一－

w 4 4 Dm (3.37)
i EE1、
H

吋吋－

--+
N

dil 、－

y
叫一
吋旬

4
－﹜吋／
F 一

3
R Jh
(
…
一－

nu
N

(3.38)
B
nυ

To compensate for the spherical aberration of this element. the proper shape of the
negative element A can be extracted from the q間dratic equation

(NA +2) 2 N.(4NB 一 t)(N,-1)211

7于一到一 （2N A + l)K, + N 一 (J.39)
4﹛NB 一 1)2(N8 +2)
’,1 is the expansion 悶t10 or magnificat1011 of the beam expander. It 1s expressed by

fn D
m=-y;=d. (3.40)

where D and d are the free apertu間 diameters of the elements as indicated m Fig 3 24.
 CHflIFR 3. PRn、WlY ABE阻ATIONS 61

KA is the lens shape factor of element A. Rearr祖ging an earlier given expression for the
shape factor yields for the first radius of element A

R (NA 一 I) ,.
A ＝一一一一－一－ rA.an (3A1)
KA ·'

(N,-1)
R 一一一一…－ f (3.42)
2 一 （ KA 一 1) J

For the c品e when both elements 叮e made from the same matenal. Eq.(3.39)
reduces to

(N +2 ）刁 N(4N -l)m
一－ K﹔一（ 2N
+ l)KA ＋昕一一一一－ = 0 (3.43)
4(N + 2)

Table 3.3 lists ti時間dii of four Galilean beam exp個ders using diffe間nt materials
for the elements. The focal length of element B was chosen to be 50 mm for all cases. The
magnification of the systems is 5.

Table 3.3 Radii in ruilhmeters of Galile且 beam exp扭曲也 corrected for sphencal aberration.

Material N KA RA, RA, Ru, Ra,

Glass 1.5 l.81547 一2.750 一6.131 +175.00 29.15

Zmc Selenide 2.4 3. 18146 一－L401 一6.412 一 120 30 44.25

S1hcon 3.4 4.69063 一5 117 一6 503 一82.45 48.85

Germ血ti um 4.0 5.58199 5.374 一6.548 一75 00 50.00

Opltmizi 月 the tlun lens values with added thicknesses of 0 8 胡d I 2 mm results m

the aπangemenl shO\\U m Fig. 3 26. In optimization‘ only the second and forth radii were
modified m additton to the adjusted spacing between the elements While the process was
executed for a single wavelength-namely 10.6 µm…the C02 laser wavelength. this bean1
expander performs well over the entire LWIR band from 8 to 12 阱。
62 OPTICAL DESIGN F UNDJMENTAL.IFOR INF卸REDSrSTEMS

一三江三三令

LA'i6J7

一一…~＂
＇＂＇凡咄帥 ”~棚”慚

FIG. 3. 26 G-J!ilean 5x beam exp扭曲r

3.6.3 Telephotos

Telephotos and reversed telephoto systems c浴血 be analyzed the s副ne way with the
thin lens concept for predesign purposes. The two types 缸e shown 凶 F喀 3.27.

.. BFL .. ‘且＇FL ’，
。F

。
吋。
。
。
。一

.. EFL
., .. EFL ‘
FIG. 3.27 Telephoto and reversed telephoto lenses.

A telephoto lens is compact for its long focal length. A reversed telephoto has a
short focal length comp前·ed to its overall size. Notice the striking difference in the ratios
between the effective focal lengths EFL (system focal length./) and the back fo叫 lengths
BFL (the distance from the last optical surface to the focal plane).

3. 7 Reflective Optics

There 叮e several advantages to usmg reflective optical elements: one is that m盯'OfS
are not wavelength selective. This is very convenient for system alignment. the 間er can
foe田 the system in the visible spectrum and then place the detector at the same image
position Anyone who has had to align 個 IR system "in the dark:· i.e . With a detector
alone‘ knows what a relief it is to be able to observe vis恤lly what is going on. hnaging
mirrors are usually less expensive than lenses made for the infrared because most of the
time the mπor substrate material is of secondary importance The demand ts more of a
mechanical na個re However‘ there are also d1sadv曲tages One of them is that the ima且e
formed by a mirror lies at the same side as the o~ject. This means that the detector is in
the way. resulting in the obstruction of mcoming radrnllon as it is shown 111 Fig. 3.28. As
 Ci!Al'TER 3. p阻馳、RY ABERRATIONS 63

we shall see shortly、 there are methods to avoid the obstruction, but. 品 expecte益， one has
to pay a price.

一＋
4‘
D

闢酬，

司，

..
f .,
FIG. 3.28 Centered imaging mirror showing central obstmction

3.7.1 The spherical mirror

A mirror has only one opllcal surface. If the s叮face 1s a section of a sphere. it is
relatively easy to m甜ufacture 祖d performs rather well. Spherical aberration of a
spherical m肝or with the same (-number as a gl晶s lens. shaped for minimum spherical
aberration‘ is approximately eight times smaller than 誼iat of the lens.

The angular bl叮 spot due to thrrd-order spherical aberration for a spherical mi叮叮－
\\ith an 喇eel located at infinity and the ape巾re stop at the mirror. is

P 咽2位一 128(f絡）＇ (3.44)

The coma blur is larger by a factor of (N+2) th個 that for a best”shaped single lens
、叫th index N It is

~ '0'"" -16(j;#) 2
(3.45)

Tt時個igular astigmatic blur spot si甜 is the same as for the thrn lens, namely

B ~·＇• = 2 (11#) (3.46)

Using a spherical miπor 可恥axis" to avoid obstructmn can be looked upon m

several wa)S. This is mdicated in Figs. 3.29‘ 3 30‘ and 3.3 l. Figure 3 29 shows the
a pert叮e stop decentered lo select a portion of the total mirror. In Ftg 3.30 the spherical
mirror is laterally displaced 在om the 叩lical axis and tilted. The apert叮e stop is the
physical size of the mirror Figure 3 31 has the mirror centered about the opllcal 缸 is and
the mcoming colliniated beam is oblique 阻！alive to the optical axis The performance is
the s抽1e for all three c011缸里urations because a sect10n of a sphere remains a spherical
surface of rotation about anv normal of the surface
64 Omcu DESIGN FUNDAMENTALS FOR INF犯組ED SYSTEMS

Ape即問 Stop

~
~t
'I

白
。
個－呵，帽”－一司、，
、自

_ －』

FIG. 3.29 Spherical mirror with decentered 叩erture stop.

FIG. 3.30 Tilted 血llTOf FIG. 3.31 C間1前-ed m凹or m曲

laterally displaced. 。bhquely euteri時間ergy btutdle.

Fi耳. 3.31 is pres叩led three-dimensionally in Fig. 3.32 to illustrate how the sa且ittal
and tangential foci sep缸ate 、.vith the change of 11P ‘ the angle of the incoming collimated
beam. While the sagi社al 的0cus Fs moves along a line as a function of angle 11,, the
tangential Fr follows the path of a circle whose radiusρ1s one q闊前er of the mirror radius
of curvature
 CHAPTER 3. PRIM咄Y An四RATIONS 65

¥:::n~~7~o~'::"s FT ~:r加
e r~由e世
for凡

•
u, 抄

司，
＇＂（、；~」
。 Fo ~ 0..F,
...
...
...
a‘
。 h I Is
J~ p
4‘

. ’,
’,
’,

FIG. 3.32 Astigmatism of spherical mirror for 翩。bject at infinity.6

From Fi且， 3.32 it can be re晶晶詛咒l that the sagiltal focus can be made to coincide
with the tangential locus if the mirror radius that focuses the sagtttal ray is made shorter.
Ami叮or with such a shape is a toroid For a given focal length. the two radii are

" 2f
tangential ra <lms …一一
叮－ cos t也司
(3.47)

sagrttal radms R, = 2.f cos z九 (3.48)

3.7.2 The Mangm mirror

This element is a thick negative meniscus with two spherical surfaces. The rear
surface 1s the reflector A Mangin、 a French otlicer, invented tt m 1876 to avoid the cost
of generating a parabola. The overcorrected sph巳rical aberration of the negative lens
portion compensates for the undercorrected spherical aberration from the mirror. Smce a
ray p閻明S twice through the memscus (Fig. 3.33 )‘ the Mangin mirror functions like a
triplet.
66 OPTICAL DESIGN FUND.品1EN7AlS FOR INFRARED SYS1EMS

An element that uses refractive 晶 well as refl自tive power‘晶晶 the case for the
Mangin mirror‘ is called a catad10ptric element.

1,3 2 2 3
., 育生 一..
... ‘
。 。
F
‘ F

..

FIG. 3.33 111e M阻.gin mπ·or 血d its equival 間t lens triplet configuration.

Mang姐 elected to make the front surface concentric to the elemenes focus. where he
placed the source for the search light.

η1e power for a thin Mangin mirror can be expressed by

申 l 2(N - !) 2N
－－一一一一 (3.49)
f R, R1

2N
If R1 = -f, R1 ＝一口百亡uf· (3.50)

and the aberrations for such a configuration are

spherical aberration
p 2N- 3
(3.51)
中加自 128N2(f ！＃）＇ 司

con1a P,o＝＝~. (3.52)

M于
astigmatism p 何一」一 … (3.53)
岫g - 2(f!#) . ""

阻ial chromatic aberration

p N- l
(3.54)
呵呵 2Nv (J !#)

Eq. (3.51 ）時veals that for the chosen arrangement 的＝ -/). spherical aberration
vanishes if N = 1.5 Thud-order spherical aberr叫ion can be elimmated for other mater抽ls
by applyi月也e lens bending tee恆叫 ue‘ which leads to solving the cubic eq凶tion7
 CHAPTER 3. PRIMARY Al您別tATJONS 67

(tJ' ＋~（tr 吽旦出j＋~＝ O.

The radius for the reflective rear surface can be calculated by rea叮個gi月祠，（ 3.49)
to read

D 一一三且Z (3.56)
"' - 2(N 一 Df-R， 一

Table 3.4 hsts the radii for Mangm mirrors made from different matenals‘ co汀ected for
third-order spherical abe叮·at ion

Table 3A Surlace radii for Mangin mirrors‘四訂·ected for tllir品order spherical aberration.

Material N R, R,

Glass 1.5 一 ！.Of 1 耳f

Zmc Selemde 2.4 l.i506/ 1.3980/

Sthcon 3.4 一 l 179{!f 1.3409/

Germanium 4.0 一 i 1823/ 1.3169/

3.7.3 Classical two-mirror conlignrations

1 here are four basic two-mirror c均nfigurations v.e shall discuss in detail. They are
named after their mventors. All suffer from central obscuration. The svstems are
ill 山trated in Fig. 3.34.

l fhe N的 l'/017 co凶igurallonconsists of a focusing prim缸y min or and a secondary flat

臼！ding m盯or that deviates 也e light bundle by 90° relative to 也e optical 眼is to locate the
focus of the arrangement outside the mcoming be血泊

2. The Gregmy 盯rangement h品 two concave mirrors, opposin皂 each other. An

intenmed1ate image is formed between the two m盯ors. The final image is right side up‘
and therefore the configuration ts a te叮estrial telescope.

3. Cassegrmn changed G間臣。可＂s setup sorne\\1iat by turning the secondary nuπur around ”
makmg it a conve、 element The system becomes much more compact. It is the most
widely used configuration for infrared applications.τhe fact that the image is upside
dov.n is of not much concern. Such a type ts refeπed to 血 an astronomical telescope.
Nev.ton never credited Cassegrain for inventing this compact system. He claimed that it
明s just a log1cal extens10n of Gregory's arrangement.
68 0月1CAL DESIGN FUNDAME盯"ALS FOR 卸F削REnS！＇了TEMS

4. Schwar=schi/d reversed the basic Cassegrain type. making his configuratmn s1m1lar to a
reversed telephoto This objective 1s most frequently used as a 1mcroscope objective.

F
/ 一。

。F

Newton Gregory

F
叫。
/ F
。

Cassegrain Schwarzscbild

FIG. 3.34 The four classical two-nurτor confieurations

‘ v

3.7.4 The two-sphere Cassegrain sys組m

As stated. because of its compactness‘ the Cassegrain configuration is frequently

used for lR apphcallo間， 1 herefore. it ts wo甘hwlule to mvestigatc what two spherical
m盯ors c叩 do for this configuration. Figure 3.35 shows the physical length of the
Cassegrain objective in relation to its focal len皂白．

Secondary
Mirror ~的i:O'i
可，
。卜 …抄

y 2
~
、，
可

。 斗，

d
d
可‘

hR

ao hphv

... I
FIG. 3.35 Basic C品segrain configuration dimens1ons. f IS 血e effective focal l開車th 司 d is the
spacmg between the 棚。血irrors. bis 曲e dist扭扭甘om the secondary mrrror 2 to 血e image plane.
called the back focal distance y 1s 也e semi 叩自tnre size. and h 間也e image height. All dimens1ans
identified 田e cons1d盯ed to be positive.
 CHAP'在R3. P!mU、RY ABERRATIONS 69

The miπor radii for this a叮angement are

~ 2fd
一一 (3.57)
刊 － b-f ‘
and
2bd
..- - -
, - b+d- r· (3.58)

With the ape1ture stop 剖 the primary mt叮nr‘ the general third-order aberration
contribution equations' y時Id the follo明白E expressim間的or 也e a.~gu!ar blur spots in
radi叩s:

~ £(b-.ff+b(f-d-b)(f+d-b)'
Spherical abe叮·at ton (3.59)
仙一 128）臼3 （！紛）＇

戶 一坐0!__二 ！）＇ + (J …刊）（J + d-b）（川－ b)]

Coma (3.60)
四＝ - 32dγ （JI#)'

Astigmatism
自一呻bf~企立 d-b)(d 且 f- b) l (3.6 l)
叫－ 8f3bd (fl#)'

bfd
The Petzval radius is p P<t~叫＝ (3.62)
fil-(b-f)'

.. j;
.,
司， 1
。 2 司，

v
., ..
h
u, .. ‘
a 。 F。 F

’‘
d=v ＝~主
(jl#)
J

‘ h = 0.5/ ...
f ...
‘
FIG. 3.36 Cassegrain 叮st甜1 layout 祕1th 25% 肘ea obstn』cti<
the prin】叮 minor. 111e syslen也／／＃二／／（勾）. 'md 血e half-field 血gle "" = h(f for an object
localed al infinity.
70 OPT!＜臼L DESIGN FUNDAMENTALS FOR INF. RED Sri叮EMS

With (/ /l ）品 practical choice for the f-nwnbcr of the primary miπor. and a 25%
obstruction for the incoming energy bundle. the general layout from Fig. 3.35 changes to
the one shown in F悟， 3.36,

The surface radn for the arrangement in Fi臣， 336 叮e

R 一－
2f (3.63)
i 一一（／／＃）．

自td
n
w 一－
J (3.64)
"2 一 I （／，惜！

The angular blur spot sizes are

Spheri心rration P 叩··~ = ik［卜（州九 （fl#(+ (fl#(]. (3訂）

Coma P,0岫＝告﹛叫
＋［（fl#

蚓帆ism 丸b惡＝~﹛［（fl＃） 也 1][1- 3(fl#)j'(J1＃）一3 - 8}. (3.67)

and the radi血。fthe Pe包val c叮vature is p ＝站滿 (3.68 }

Spherical aberration reaches a m凶m酬 with/ /3 、泊的 1t is 0 003183 rad. Coma

for that f-nwnber is 0.025463 峙，朋d astigmatism 的。.203704 11,' The Petzval radius is
f﹔ 1ιconcave If we asswne a halιfield 叩gle of 0.026 rad ( 1.5°). and a I 00』mm focal
length.f. the radii become R1 ~ 66.6667 mm. 關d R2 = 50 nun. The total 胡gular blur
spot size adds up to= 4 mrad. The physical diameter is 0.004 x 100 = 0.4 l11111. While this
is relatively large. it may be adeq阻te for the application under consideration.

To demonstrate the usefulness of the thir注明order calculations once more‘ a plot of

the encircled ene喀y distributio阻 ob祖ined by geometrical ray tracing is presented in Fig.
3.3 7 for comparison.
 OW'TER ). PRIMARY ADERRATlONS 71

..岫岫 DEG
1.seee DεG
’I
．心『
〉臼u山Z川

ava765
w 口恤帥白
0

』
Ju芝凶”比口ZDH

”,
．

3218
←UεN泓
uH
m

mumu
100.000
健伺DIUS FROM CHIEF 健開Y IN ” ICRONS
2四a .00

GE。”IET健IC ENCIRCLE口 EN 籃鍵 GY
TWO 間 SPHERE CASSEGRAIN, F = 100例”， F/3 3 3?
WAυELENGTH' POLYCHROM向 TIC
。”γm HAS BEEN SCALED BY OIFF 快開CTIDN LI州工τ

FIG. 3.37 Encircled energy plot obtamed by geome出c ray tracing

3.7.5 The two戶sphere Gregory system

In Fig. 3.38, the Gregory system is compared to the Casse~rain aπangement. For the
same obstruction ratio between the secondary and prim血y m1叮ors a間i the same nnage
pos1t10n. the focal length and therefore the j-nnmber incre品es for the Grego哼哼S脂m、
and so does the overall physical length
的帕

mwhM
松（
Ednα

Cassegrain
mmw
r

Secondary Primary
Mirror Mirror

le國呻臨
‘ ’,

五﹜呵間
‘ ’,

FIG. 3.38 Layout comparison between the Gregory 阻d Cassegrain a叮阻gements.

72 0阿1CAL DESIGN Fl!NDAMEN1 4.LS FOR INFIW<ED Sm日，fS

To calculate the radii of the sphencal m盯0凹 for the Gregorian system and i祖
ab甜甜ions. Eqs. (3.57 ）伽ough (3.62) c祖 be applied by changing the focal length to a
negative value Con甘ary to the Cassegram. \\bere the mirrors have opposi阻 shapes
(concave pnm訂y and convex secondary)‘ both m盯ors for the Gregorian are concave.
which resnlts in a higher total spherical aberration. To improve 也e perfonnance of both
configurations. 晶pheric surfaces c叩 be 閏ed. This will be discussed in Chapter 5.

3.7.6 Scbwarzschild, a very special 個se

H 出e two spherical mirrors are concentric to each other. sep訂a阻d by lwlce the
syste血，s focal length‘ third-order spherical aberration. coma and astigmatism a間
eliminated, provided the apeπure stop is iocated at the common cen阻r of c山vature c as
shown in Fig. 3.39 Adding a thin field flattener lens in the focal plane co叮ects for
distortion and field c叮va恥間描 well.

當
h占vc。
』‘ nv
Y2 ’,- 34
una

" 0
:N

在
令

~ f ....
d
....‘一 .,
b
、，
FIG. 3.39 羽ie concentnc Schwarzschild objective wi也且eld tlatten前，

The specific 間lattons lo achieve the co叮ections of sphe討cal aberration. coma. anti
astigmatism are

d=2f 、

b=(J5+2)f.

R,=(../5 一 I)f. (3.69)

R、 ＝（ ..JS+ I)f.

月＝（而 ＋ 2)y,

The Petzval radius is‘ according to Eq. (3.62). p = :f. To coπ·ect for this effect. we
insert a field lens whose Petzval radius is +l Since the function of this field lens is to
 CHAP'正R 3 P也址、RY ABERRAnONS 73

”llatten” the image in the field‘ it is called a field llattener. By apply姐g Eq. (3.12) to
calculate the focal length of the field llattener‘ we obtain

如－｛－ .β.70)
where N is the index of refraction for the field llattener lens. and f is the focal length of
the Schwarzschild objective. By placing the field flattener in the plane of the image. the
lens can be shaped as stated in Eqs. (3.71 ）血泊（ 3.n) to eliminate distortion.'
N2-I
The fr叫叫帥 R,=f=N市工nf· (3.71)

and the rear radius 几＝（ N ＇一 1)/. (3.72)

With a high-index material, such 晶 N ~ 4 for germamum. 凡 becomes 15 tunes the

focal length Thi臨 sug溫ests substttution of a !lat surface and correction of R3 for power
Not only would tins lower the cost of the lens. but it invites the idea of depositmg the
detector elements dtrectly onto the re缸 S叮face of the field llatten前，

For further study司 including the aberrations for finite cm\jugate config間tions『 see
Re[ JO.

3.7.7 Reflective beam expanders

Reflective beam exp祖ders 缸e modi且ed Cassegrain 祖d Gre且orian m盯or S)

Both the object 缸id the image 缸e located at 血finily‘ v.hich me祖S that the input and
output beams are collimated (afocal system) We shall assess how well that c剖t be
aclneved with two sphencal mirrors Figure 3 40 shows the Cassegram configuratton. To
avoid obstruction. the mirrors are used in an ofι缸is mode. Due to the sphe討cal shape of
the mirrors. the exiting be祖n ts not collimated The m血-ginal ray converges 阻1der an
angle y relative to the optical axis Due to the off-set of the mirrors from the optical axis.
the convergence angle a of the total beam in the meridtonal-plane shown is somewhat
smaller th臨下 The Gregorian beam expander is shown in Fig. 3.41.

B Marginal Ray
.
可：
’, a • 4且

D 且
可
,. γ

.... (A
司，
Ya

‘
4
..
YA d ’,
。
可， 司，

.
.
‘ .. f
’,
} ’,
Optical Axis

:•. j~
”,
（吋

”,.....
FIG. 3.40 Casseg:min beam e文pander.
74 OPT/C<L DESIGN FUND.品fENTALS FOR INF制RED SYSTEMS

Marginal Ray
B
mi ~：r一－ .. • • ‘
a
'"'•"''I
心d抖出， ’, D
..""
Ail 四

、， ..
叫
叫

,,.
、而
ιa y Ya
......
VVdaa
’ . .
•
二、~~： A
司，

A b』
d Optical Axis
』‘

‘. fa
(+) .‘
f,
(+) ’,

t
.. .,
FIG. J.41 Gregorian be胡2 間pander.

’The angular convergence for both configurations is

y (3.73)

where m is the ma甚nification or expansion ratio Did. It is important to remember that m is

pos1t1ve for the Cassegrain exp四der and negative for the G甜gorian.

The relative ape個e (f ／的 ＝ fa 1(2川＝ IJ;, I1<2川 the separation of the mirrors
I= !A + [8 For an[ 12 system with m = 5司 the convergence angle y for the Cassegrain
expander is ＝ 一 1.5 mrad. and for the Gregorian with m = 5. y 三 2.3 mrad. Notice that
the factor 旭 front of the square bracket of Eq. (3.73). is equ叫 to 仇Nice the value of the
mimm阻n angular blur spot size of a single spherical mirror due to sphe討cal abe叮atio詞，

To correct sphencal aberratton with two reflectors. at least one of them has to be
aspheric. For both configurations. the classical Gregorian and the Cassegrain. all mt叮ors
have to be paraboloids. Details will be covered in Chapter 5.

3.8 Diffraction Limit

An ideal optical s﹜叫em would image an object pomt perfectly as a point. However‘
due to the wave nature of radiat10n. dt倍action occ山·s. caused by the liI世tmg edges of the
system‘ s aperture stop. The result is that the image of a pomt ts a blur. no ma虹er how well
the lens is corrected. This is the di知action blur or Airy disk, named in honor of Lord
George Biddel Aiiy. a Bntish mathematician (1801-1892) Its cross section and 恥
appearance are shown in Fi耳 3.42.
 CHAPTER3. P阻叫RY ABERRATIONS 75

Since this blur s阻e is proportional to the 、wvelength. as indicated in Eq. (3.74 ）.也e
dif音action effect can often become the limiting factor for infrared systems. If 也e aperture
of the lens 1s circul缸， approximately 84% of the energy from the energy from an imaged
point is spread over the ce甜al disk surrounded by the first dark ring of the Airy pattern.
The diameter of this central disk is

B"iff = 1.44 入（／鮮） (3.74)

... 83.9% ’ …抄 B岫 4

91% ’ 91% ...

‘ ..

FIG. 3.42 Ai可 disk‘ en叮叮 distribution and appearance.

For a LWIR system operating at a \Wvelengtl1of10 µm and anf-nurnber of 2 ‘ the

diffraction blur with 84% of the energy is 48且間n (approximately 0.002 血） in diameter.
fhe trend to decrease the pixel sizes in focal pl胡e 訂閱ys to mcrease 出c system's
resolution 間quires awareness of tlus lhmt set by nan叮e Including the energy of a
di的action blur twice in di缸neter of the central disk (the 聞自ure of U1e second dark ring)
increases the total encircled by only 7%. The spread of the energy outside the central disk
can cause disturbing 叮osstalk into 吋acent detector elements in an 研ay

3.9 Resolution of Imaging Systems

Two points are resolved when their images are sep盯ated enough from each other so
that they can be clearly recognized 阻 two pom組 A convenient and widely used value for
the minin1un1 sep叮ation is the radius of the first dark ring of the Airy pattern. 品 indicated
in Fig. 3.43. This is Rayleigh's criterion司扭曲：ne丘 after Lord Rayleigh. a British physicist
(1842一 1919).
76 UPT/CAL DESIGN FUND.品巫師Al.S FOR INFRARED SYSTEMS

.. i.22λ(Jii的 4

FIG. 3.43 Lord Rayleigh's resolution en阻rion.

References

I. J. R. Meyer-Arendt. /11troductio11 to Classical and Modern Opfi，且隨心恤.d edition.

Practice-Hall。 Inc. (1984 )‘ page 120.
2. W. T. Welford. Aberrations of Optical Systems. Ad血n Hilger ( 1989 )‘ page 228”
3. K. Miitze. ABC der Opt峙， Verlag Werner Dausien (1961), page 307.
4. W. J. Smith. Modem Optical Engineering. McGraw-Hill Inc. (2000). pa_ge 335.
5. M. J. Riedl.
Pub I』s趾n惡 Co. (September 1981). pa惡es 49-52.
6. M. J. Riedl. "Astigmatism and the spherical mirror.” Electro-Optical Systems Des嗯，’，
Milton S. Kiever Publications‘ Inc. (Sept 1977）。 pages 27…況，
7. M. J. Riedl ‘”The Mangin mirror and its primary aberrations‘.. Applied Opti口， Vol. 13.
No.7. pp. 1690一 1694 (1974).
8. W. J. Smith司 Modem Optical Engineering. McGraw” Hill Inc. (2000). page 328.
9. R. Kingslake. Lens Design Fundamentals. Academic Press ( 1978 )‘ page 208.
I 0. I. A. Artioukov and K. M. Krymski. “ Schwarzschild objective for soft x-rays", Optical
E’,1gineering. (Au學凶 2000). Vol.39. No. 8. pages 2163 2170.
CHAPTER4

Wave Aberrations
4.1 Introduction

So far. it has been assumed that light propagates 晶 ra~s. This is not the case. Rays
only exist m theory 祖d are a convement b值1s for geometric optics In reahty. rays are
lines drawn in the direction of the propagation of light. The~ a扭曲e normals to the
wavefront in 伽 directmn of 也e flow of the radiant energy. Tl別 is indicated in Fig 4.1.
where Iiβt radiates h血 a point source P m spherical waves After exitmg an aberration-
free optical s﹜叫em。 the wavefront conve啥es precisely toward point P'.

If the media 也rough which the light travels are not isotro阱， the 明vefront is no
longer spherical. Adv曲tage is taken of this fact with gradient index material to form the
desired wavefront shape.

To explain the concept of wave 油erration in this fundan1ental text, we limit

ourselves to the discussion of the spherical aberr現ion It is sometimes desirable lo
compare the results obtained with geome肘cal opl!cs. This will be done for tlnrd ·order
spherical aberration For an in-depth stu句， the reader is enco叮咚ed to consult references
I and 2.

4.2 Diverging and Converging Waves

Light emanating from a point source forms a diverging spheric叫 wavefront. A血er
p品sing through a fitting optical system, it is converted into a conve增加g wavefront,
aunmg for the 阻ial image p。如t 甜 mdicated m Fig 4 I 。

• 司，

、 Opl!cal System

p . 一， 事－
P ’

T
4 ‘
FIG. 4.1 Diverging and converging waves m isotropic media.

77
78 0.阿7CAJ. DESIGN FUND.晶1/iNTALS FOR INFRARED SYSTEMS

4.3 Optical Path Length OPL

The total optical path length through an optical sy到em is the sum of the product Nt
along each ray‘ where N is the refractive md侃 of any portion of the ray path and I is the
length of that portio且 i.e.‘

OPL=I.Nt (4.1)

In Fig. 4.2 two rays 缸e shown: the axial and the marginal ray. The optical paths
lengths 叮e. respectively‘ OPLA =tA1 +Nt_.2 +t .. 3and0PLM =tM1 +NtM2 +tM3

t,, t.. ,
p P’

FIG. 4.2 Opti且d pa也 lengths for aual 祖d marginal rays.

If OPLM is equal to OPLA. there 時 no difference between the two optical path
lengths and. therefo間， there is no wavefront defom祖tion The wavefront exiting the lens
is spherical and concentric about the image point P'.

4.4 Optical Path Difference OPD (Wavefront Aberration}

Once agam referring to Fig 4.2. if the時間 a difference between the two optical path
lengths. aberration is present This wavefront abe甘ation is expressed 晶

OPD ＝（ 工 Nt ）， 一 （ l:Nt)M. (4.2)

4.5 Spherical Aberration

If the paraxial focal pomt of the lens or system is used as a reference pomt about
which a reference sphere is constructed. then the distance between the point where the
margmal ray the normal to the ex1位ng wavefront and the paraxial focus is the
longitudinal spherical aberration. This is shown in Fig. 4.3.
 C1L們芷R4. WA而且目RATIONS 79

Reference Paraxial Marginal

sph 目它 f田間 foe阻

Center of
0 cl'
exit pupil

Exiting
wave front ... LA ‘
(l』ngitudinal Spherical Aberration)

FIG. 4.3 Longitudinal 叩＞herical aberration is tl1e dist個ce betwe扭曲e marginal 曲d para也al foci.

4.5』 Numerical example

We analyze a doublet made from germani凹n with mdex N = 4 The chosen focal
length/= 100 mm，也e//# = 2‘四d the design wavelength λ ＝ I 0 µm. After optimizing
the thin lens configuration, the folio吶ng prescnpllon 1s obtained.

Front element R, ~民 232.5 nnn

l hickness t 112 = 4 5 mm
R、＝ 348.5 mm
Spacmg 1,13 = 0.5 mm
Rear element R3 =87 Omm
Thickness 131, =4.5 mm
R,=99 572 mm

To calculate the opllcal path lengths is somewhat cumbersome司 but the exercise
brings with it a better understanding of the subject nnder discuss1011 Figure 4.4 and Table
4 l provide a stm祖1血y of tlus exercise
80 OPTICAL DESIGN FUNDAMENTALS FORfNFRARED SYSTEMS

a:Nt•= 130.7297
M 一一呵呵＂＂＇.－－『』 -· .•

A 閉－：－－回風 可圖旨
I F祖祖i剖伽聞 Marginal foe阻
︱ αNt), = 13且7331

..
94.0157 •LA ‘
• 94.2331 .,

FIG. 4.4 Optical path I叩E血s for rays A 曲d M {Tue 1llustrat1on is not to scale).

Table 4.1 OPLs with respect to marginal focus for doublet shown in Fig. 4.4.

Optical Path Length for Ray M Optical Path Length for Ray A

l.3479940 18.0000000
16.1589969 0.5000000
3. ll55744 18.0000000
15.716241 l 94.2331377
94.3909159

(l:Nt)M = OPLM = 130.7297223 (l:Nt)A = OPLA = 130.7331377

The optical path difference OPD = 130.733 l 一 130. 7297 = 0.0034 mm‘ which is
•.~uivalentto 0.34 waves for the chosen wavelength of l 0 µm. This is sho"n as plots in
Ftg. 4.5

The longitudinal spherical aberration LA= 94.2331 94.0157 = 0.2174 mm. As

already stated. this is the distance between the margmal and the paraxrnl foci
 Ci叫可ER4. WA呃 AllERRATIONS 81

Waves
0.5

＼／岫p~p~e (normalized)

4」~·tr
,,
for marginal focus
0.34 wav＜間 P-V

FIG. 4.5 Wave ab甜甜 on (OPD) witl1 reference to margmal focus.

4.5.2 Best focus position

In Fig. 3.11 we have show羽 that there 扭曲 optmmm location for the thud-order blur
spot due to spherical aberration That location is b血ed on geome甘ic optics Based on
wave aberratmn. the optunum positton lies hal!Way between the m訂ginal and paraxial
image planes. In other words 『 by 間foc 凶ing. the wavefront will be referenced to a sphere
whose center lies now haltway between the m缸ginal and paraxial foci. The e時ect IS
shown in Fig. 4.6 ‘ the plots refer to tlte s田ne doublet discussed above.

Waves
0.5

C土豆』何~ Exit pu!'il

(norrnahzed)

J二三c-r=-:竺？甘于？＇＂＂＇干干r
／三啞體間－費？

for best focus

0.085 wave活

FIG. 4.6 Wave aberration for best focus position.

82 OPTICAL DESIGN F!INDAMEN1ALS FOR INFRARED SnT.凹的

4.6 Third-0吋er Spherical Aberration

To simplify the calculations and to show the relation between geometrical and wave
aberration 阻4 田ing third·“ order aberration theory. the followmg relation exis恆

LA
OPD E王一一一一－－－＝－ (4.3)
間 16(//#)"

where OPD，~ indicates the optical path difference with 間gard 個 the parmdal focus.

Applying Eq. (4.3) to our example‘叫ih LA= 0.2174 mm 叫u..r111=2‘ the ti誼叫，
order wave aberration is
0.2174
OPD. E王一一一－－：：－ = 0.0033969 mm ‘
間， 16· 2'

咐1ich is 0.34 waves peak to valley for a wavelen~th of 1 。”n. This agrees well with ti時
result obtamed 問時也e more tedious way shown m Eq. (4.2).

For the best image position-which lies ‘ as already stated. half way between the
p由那ml and rnargmal focal pomts-the tlurd-order spherical aberration reduces to

OPD
LA
E王一一一一－－－＝一立一一一」~
OPD … (4.4)
山＂ 64(f 紛r 4

τhis translates for our example into 0.34 / 4 = 0 085 waves and coπ·esponds lo the results
illustrated in Fi品 4. 5.

4.6 Depth of Focus

Lord Ray le1gh found that if the optl開l path difference of an optical system is not
more than one qu訂ter of a wavelength. for all practical pu叩OS品， the sys扭曲 can be
considered diffraction limit•臥l. Therefore the permissible defocusing range for such a
svstem is
tJ.foa• '= ±2λ （f 掃）2 (4.5)

This would mean that for our doublet. the allowable defocusmg limit would be 士 80µm

Reference治

’
W. J. Smith. Modern Optical Engi11eeri1 ·g. McGraw-Hill (2000 )‘ pp. 348 352.
一， R. E. Fischer I B. Tadic-Galeb‘ Optical System Design. McGra＇恥Hill (2000）司
Chap. 4.
CHAPTERS

Special Optical Surfaces

and Components
5.l Introduction

An aspheric surface is considered a special optical surface. It is spec拙I because it can

do more th個 a sphencal surface wtth regard to aberration co叮ection. But it is also special
because it 阻 much mo甜 expensive to manufac扭扭可 especially if the surface is produced by
means other than diamond turning. A di鈕active surface wi也 its grating profile engraved
into it to act as a thin lens belongs 凶 this category of special surfaces 品 well. Both "'ill be
discussed in detail.

Because windows ‘ filter substrates, and other plane-parallel elements such as beam
splitters and Dew，訂 windows 訂e often ignored b~ the 凶er and not considered to be optical
elemen妞， we treat them here as elements with special surfaces. These elements in
converging or diverging light contribu阻 to aberrations j叫t as a lens or a curved mirror
does In addition. they slu岳 and displace image locations. Knowmg the behaviors of plane-
parallel plates allows us to correct for their aberratmn contributions or t心e advantage of
that knowledge in balancing the system’s aberrat的肘。

Another element s叮prisingly often ignored m the layout stage is the concentnc
dome. It contributes to the abeπations and has focusing power. The details of these
characteristics will be addressed

Ball lenses 血id gradient mdex lenses are frequently used as coupling elements 、間th
optical fibers. Their 品eatures and performances 訂e mcluded 品、問11. even though their
applications are limited to the visible (VIS) and near-infrared {NIR) spectra.

5.2 The Plane-Parallel Plate

By definition. a plane-parallel plate is flat and parallel. If thlS is not the case‘ it is a
lens with convex or concave spherical or cylrndrical surfaces The surfaces can also be
toroidal or general aspheres on either or both sides The plate can also be flat on both
sides‘叫th the surfaces not parallel to each other. In that c品e the element is a prism.

5.2.1 Displacements

If a plane-parallel plate is inserted into an image心orming optical system‘ with its

surfaces perpendicular to the optical 阻阻‘ the image plane is shilled toward the right.
assuming the light is coming from the le自（see Fig. 5.1 ). The amount of the shill depends

83
84 OPTICAL DESIGN HIND.制'ENT.~LS FOR 卸F創RED Sr.盯卸的

on the index of refraction of the material 田ed. The higher the inde丸 the larger the shift.
For the paraxial region. this longitudinal displacement is

A 一（ N-1)
一－ (5.1)
" N
f is the thickness of the plate with index N. Outside the parmdal region. a more
complex relationship apphes:

-l :--i -z-i;- TZ83802213111L

--
--

l.
『
il
rl『

－ n 一－
4（卅一l

li
uy-
N -- ; nzz
sEIBEfitL

l AO

－－
A -
一－

et 5.2)
-

ettlef
m 2
t
u
4N'(f圳－ -I I

」
where tan 11 ＝」：；；，－ = sin 11 • W凶let凶岫岫ces 祖 error of about 12% for anf ／！叩阻
2lf I#)
the e叮or is already down to 3% for f 12 and less than I% for f 14.

The difference between ~and ~ is the longitudinal spherical aberration contributed

by the plane-parallel plate. It is a positive value. 、吋rich is refeπ·ed to as over-co何ected
spherical abe叮ation. We remember that a s油gle element has under-corrected spherical
aberration. In other words. the combma!ion improves the situat10n slightly.

P’

.. ．枷

FIG. 5.1 Longi仙dinal shi晶。fimage by pl 血e-p訂allel plate.

The second type of displacement occurs when the plate is tilted relative to the optical
axis. rhe tilt causes the image to be shifted laterally. as seen m Ftg 5 2.
「

rlil
l
2ωn
．”“呵。3
ι 一一

lllIll

rn -N 1 tw2
4

l
－－、

: -u
p - edt nzd
一－

--
月

(5.3)
月

Ill -
－－

1
v

-
h
 CIC\!'IER 5. SPECIAL OPTICAL SURFACES AND COMPONENTS 85

Often. a beamsplitter 阻 tilted 衍。 relative to the optical axis. The offset then becomes

品
ra N
(5.4)

This lateral displacement is about one third of the thickness for glass with N = 1.5
1md more th阻 one half of I for germ缸uum with N = 4.

For small tilt angl峙的﹒

r曰四＝一一一一一
（ N 一 I) tu =Li 11 (5.5)
, N r " "

、v
r
2年 .,

’ ..
FIG. 5.2 Lateral displacement of image by a ttlted pl祖e-parallel plate.

5.2.2 Optical micrometer

Frequently、 a plane-parallel plate is employed to measure or set lateral

displacements. For a chosen thickness and material. the displacement is only dependent
on 由e sme of the tilt angle. In alignment telescopes. for example. two such plates are used
o此h。gonal to each other to me由叮c horizontal and vertical shifts (see Fig. 5.3).
Sometimes司 a detector assembly is too heavy and bnlky 、Mth its cooling system and
preamplifier paιkage 岫 be aligned. and the !we• plate tilt method can be used to align the
optical axis of the system to the detector
86 OPη'CAL DESIGN FuND.品（£，問：.4L.~ FO.叫NFRARED SYSTEMS

Vertical

//

FIG. 5.3 Prine中le of 祖 optical micrometer.

For a 20備rmn-thick plate made from BK7 with inde文 1.517‘ the following
relationship exists:

Plate Rotation Angle "r (deg) l.681 3.358 5 日29 6.690 8.339

Lateral Displacement r (mm) 0.2 0.4 0.6 0.8 l.O

5.2 3 Aberration contributions

The linear aberration blurs introduced by a plane-parallel plate. based on third-order

lhem-v are
);
（一口主制

:
RWB

Jd
、“一

ti

I
心－ tv
一切一－

-- -
－－／

“
的）＋
心3

1 tl2
的

spherical aberration (5.6)

一

ma3 AJ3
F！”－

lt

12

! 5
、、
1

I
By
－－AAπ
叫

- - )
－－

coma vJ 1F (5.7)
ts

肉
－
th
F’

astigmatism
門S蠅一
也~
2N3(jl＃） 嗯
(5.8)

B <N 一 I) r
axial chromatic abe甘a ti on (5.9)
曲om 一一
- 2N2 v(jl的 q

(N-Dtu..
and lateral color 冉...ooloc - N2 V " (5.IO)
 CHAI可ER 5. SPE口ALO何·1cA1 SURIAC臼 M叩 COMPONENTS 87

As indicated in Figs 5. l ar泌 5 2. "r is the bit angle of the plate. 血1d N is the
refractive mdex of the plate matenal. The relative aperture(//#)= 11(211).

5.2.4 Applicatmn remarks

Let us look al a beanisplitter application in 甘1e 3- to 5卡m region. The be前nspli世er is

inserted into an f I 1. 5 cone.” the tilt angle is 4夕，個d the 耐心1ess of the beam-splitter is I
mm. We 叫II compare two different materials‘ silicon and sapphire.

The index of refraction at 4 µm is 3.425 for silicon and 1.675 for sapphire 心 Their
Abbe numbers are 235 血d 7.6 respectively. This large difference between the Abbe
numbers in this region should be noted.

Material I Bspher Bcoma B品lig Bchrom B1at.color B1ota1

- A.,
Stltcon

Sapphire
2.5

3.6
11.6

16.8
54.9

78.9
。3
10.5
。7
24.9
70.0 fUTI

134.7 間 / 自

λ2

The spot diagrams shown tn Ftgs. 5 .J. 削d 5 5, obtamed 、Nith geometrical ray traces.
demonstrate at once an 凹cellent agreement with the third-order calculat10ns from above

om· 0 磁胡亂目， 000ll DEG

．．．‘...：﹒且，且，：
．．．．”毛．．．皂，.
·' .
•••',’,. H

’ ,
．．．”一【’，’，
一眉，一 薯l! " ~ , •

’,
’,. •
’,
’,

agaaF 簣 ，龔 自語

, ’ ..
a閥
曾
割

胸，

. ,
會

驗
圖，

..
自前

IHR• 舟，＠恤， 0.129 剛

SP口7 日IAG健AH
SIU由N 歐盟”阻.ITT田
U盯泊的RE HI間帥的
Fii正D • I
R阻路羽IUS• 221 由
GED Rf羽IUS ' 33 ，可38
CIRCl五 0)'.A1: 7當 被于回白地CE ：阿I田LE

FIG. 5.4 Tire spot diagram resulting from the silicon beamsplitter.
88 OPTICAL DESIGN F1mo.刷E盯＇！.S FOR INF.制REDS閉目前

瓜葛U UI擔0, ...磁絢 IE

- ·-
,. Y 、. • • .. . ••
;
•• .‘ : .. 、 I •, :• • ..
• • 司唱
''{.‘
-

•. ":,,.,,.
/-‘
可＇－－
•.. • :,,.·
：~ .，.，可
~、. .,t~
：：：.－：.＇
• • .』

；；.·－~﹔~－－.；； ＇：＇·，：.＿： 于
: .-, • （.互會繭 , ”,. •• • .;• :
也圈，間hn

””
，也 P -
• ..,..

IltR' -0.0帥， 8，胡7 剛

:HA
SPOγOIAGRF司”
S睜開IRE 臨到” SPLITT回
UNITS ARE HIC醫ONS.
FI!且日 I
剛S R間DJS 呵呵 775
GEO R冊DJS 73.390
CI11CLE OIAH' 135 旺F回回CE HIDDLE

FIG. 5.5 The spot diagram resulting from the sapphire beamsphtter.

This little exercise indicates that the contribution of a relatively t趾n plane” parallel
plate c血mot be ignored Moreover可 the comparison of the matenals pom臼 out the danger
m qmckly subst1tutmg one matcnal for 削other without cons1dermg the consequences.

As mentioned earlier, filters 、社th vari‘>us l趾cknesses are t全equently inserted into a
well· lanced optical system without consideration of the e宜ects discussed above. These
e缸ec阻 di叮er鷗 depend函g on where those filters are placed within the optical train. A few
examples are shown in Fig. 5.6. In the upper layout the case of finite co阿ugates is
presented with S and S ’ indicating object 四d image distances. Notice the appa間nt shift il.
of the object when the filter (plane-p前allel plate) is inserted on the object side‘晶 is
mdicated m the center a甘個gement In this c潤趴 the lens ··sees” the object closer than
without the filter 叩d shifts the image to the right by 8. which also changes tl1e
mag位fication of the system. In the bottom sketch. the filter is placed to the right of the
lens. which mtroduces the shift il.. The magni品cation is not affected It should be clear that
a plane-p訂allel plate mserted m a collimated bundle of light does not contribute any
aberrations. However. if the plate is tilted. there is呵。f course. a lateral displacement of the
beam but not 也e 凶iage.
 CHAPTFR 5. S阻ClAL 0l'TJCAL SUllτACES AND Cc MPONENTS 89

s ’
‘ ’‘一 S ’,

秒 2 ‘

FIG. 5.6 Effects of inserting a plru時偷·parallel plate into a converging or diverging beam

As before.

峙，，＇；：，. ＝﹝：rt:,. ＝﹝刮 a (5. i I)

For a system with a focal length off需 JOO mm ‘ S= 400mm‘ f = 3 mm. N = 4. and
=
'1 = 2.25 mm. the shift ii 0.25 mm.

A better 訂rangement for the application 1llustra扭d in Fig. 5. 7 1s shown in Fig. 5.8.
The o吋 ective is split mto two elements‘ both chosen for mfini 扭 conjugate. Several
advantages can be gained from such an arrangement:

I. Each lens can be used for focusing withou!E_ha叩ng the 恥tem magnification.
2 The distance from the ob1ect to the image PP'. which is called the track length ‘
is flexible within re品on. depending on the o~ject 別扭
90 0PTJC.lL DESIGN FUND.品JENTALS FOR INFRARED SI'STEMS

3. The p叮allel beam between the two elements allow for better filter performance.
4. There is no image s凶ft m也 change of index variation of the filter material.
5 Variation of filter 也ickness does not have any influence on image position
or ma伊ification. This is a v前y convenient advantage 、.vhen filters with
different thicknesses are m回eddunngm聞surements.
6. Abe口百lions are reduced by splitting the le扭扭to two elements. Each cao
be shaped for optun盟n performance
7 The filter or beamsplitter can be lllted without introducing ast憑matism.

A B

p I'
0 0

IA 紗， t “ Is
‘ ’,

FIι5.7 A b甜前甜祖皂em開t 也an shown in Fig. 5.6. Magoif1c甜on m ~ :fa/f;、

A similar aπ血gemcnt 1s P凹sented in Fig 5.8 for a reflective system. Herc. the filter
is not only 油 parallel light, its size is the smallest possible.

FIG. 5.8 Retlecttve syst目n wi血 same principle applied as in Fi耳 5.7.

5.2.5 The wedge (thin pri間的

Two flat surfaces. slightly out of parallel to each other‘ form a wedge. Applying
Snell's law in detennming the direction a ray ＇叫ill take after passing through the wedge司
and ass阻ning that the wedge 胡gle is small. we find the deviation from the entering ray
direction to be
o ＝目 （ N … I) (5.12)
The dispersrnn 1s

必＝~v (5.13)
 Cl!APTI來 5. SPEι！AL Orne AL SURF ACES AND COMPONENTS 91

NM- I
where V ＝一一一一. All other S｝咽bols are shown in Fig 5 9.
N,- NL

向A
然心
W叭叭
f扎f扎
wp三VA

FIG. 5.9 Dev抽Hon 血d d呵，ersion of a thin pris盟（wedge).

A useful combmat10n of lens plus 、Nedge‘ as shown m Fig. 5.10, h由 been applied by
the 訕訕。r in a simple twc• color discnminator Two wedges installed behind a conunon
singlet, shaped for minim阻n spherical aberration司 provide a cost-effective method to split
出e aper仙re and create two subsystems司 each working in a specific spectral band. The
wedges serve 晶 filter substrates and redirect 吐間 focused energies for cle缸 sep訂ahon, yet
叮e close enough so that the two detector elements of different chemistry can be installed
into a single herrnet1cally sealed standard detector housing.

Aλ1
”圖

Aλ2 -
FIG. 5.10 A two-color dis口iminator optical system.

5.3 Domes

Domes are mostlv monocentric shells司 which me叩s that the second‘。r inside, radius
is shorter by the amount of the element's thickness than the first司 or outside radt間（see F1呂
‘
5。 11 ). Such an element has a focal length of

, NR,(R, - t)
, 一一 (5.14)
J~~~ - (N - I)t

It is a negative element and contributes an overcorrecteu spherical aberration. Of

course. one can take advantage of tlus fact in balancing the total system abe甘ations.1
92 0PTTCAL DESIGN FUNDAMENT<LS FOR INFRARED SYSTEMS

’,
’,

’ t.. N

FIG. 5.11 Concentric <lome is a negative element

A dome with zero power can be derived by setting Eq. (3.2) to ze悶。 l e ‘

1 .I I l (N 一］） I I
一…一一－~一… (5.15)
j 一、＂ "LR1 R， 『 N R1R1 」－ v.

Solving for R2 yields that R2 is equal to R1 less the 阻ial optical displacement caused
by the thickness of the dome司 or

(N -1)
R、 ＝ R「于有一 1 (5.16)

If thi阻 concept of a zero power lens ts C文tended to a veηthick element. a beam

expander evolves (see Fig. 5.12). The magnification of a beam expander is the ratio
between the exiting and entering beam diameters. With this definition. tlte thickness of a
solid be剖n expander can be stated by

I 立一立一（
(N-ll
1-m)R, I
(5.17)

The second radius is simply

R2 = mR,. (5.18)

i e ‘ the magnificatton is the 阻tio of the two radii. independent of the lens material

for example” a solid 4x germaru山n beam expander (N ~ 4) h師 a first radius of

I 0 mm Therefore. tis thickness according to Eq. ( 5 17) is 40 mm司 and l恆 second surface
radius is -40 mm [Eq. (5.16) or (5.18)]. The spherical surfaces introduce spherical
aberrations The correction for these 、.vtll be discussed in Section 5 6
 CJ!APT玄R 5. SPECIAL 0I'f!CAL SURF A凹S AND COMPONENTS 9J

.. ’,
FIG. 5.12 Solid be祖n exp祖d前 with m ~Did

5.4 The Ball Lens

A sphere or ball perfonns su叩risingly well as a lens. At closer scrutiny司 one finds
that such an eleme叫 can be broken down into two pl曲o-convex lenses可 separate址 by a
plane-parallel plate. The positive lenses have under-corrected spherical aberration and the
plane-parallel plate is over-corrected Therefo時， there ts a compensating effect. A ball lens
finds its main application as a coupling element for optical fibers. It deserves 阻 analysis司
especially since there are certain limitations that need to be understood

Figure 5.13 indicates that the focal length is measured from the center of the sphere.
where the ex扭阻岫ns of the ente吋ng 血id exit泊g rays meet TI1is means that the back focal
distance bf/ is merely the difference between the focal length and the radius R of the
sphere One can see inlmediately that the focus falls on the rear surface of the sphere if
f~R.

""'R ~ 句1 ...
... f ’
FIG 5.13 Ray 甘•Ce tlirough " ball lens.
94 0PTTCAL DESIGN FlJND.<MENTALS FOR INF/WIED SYSTEMS

The focal length of the ball lens is

RN
f 可N=l)· (5.19)

Its hack focal length or back focal distance is，品 already indicated.

(2-N)R
1月 ＝ J-R ＝ 一一一一 (5.20)
(N 一 1)

One c祖 see that 句＇／ = 0 if N = 2 That means a bali lens is noi suitable 心r 帥。到 lR
materials. such 晶 silicon. germaniwn. zinc sulfide. or zinc selenidιwhose mdcx of
refraction are l叮gerth胡 2 Sapphire wifu N = 1.6753 at 4 µmis an exception.

5.4.1 Spherical aberration

The transverse third-order spherical abeπation is expressed by

iN N )- tl 3
- -2
司、》7
（即

自－

一一…

cI
V 一

T
n、d

--
i -- -
N

Z
加

ι ？

(5.21)
－司

TSC
and smce the angular blur spot s泣民間focused for the best position is p 叫 立一一一
r 'P'"' 2j
‘

~， N(N-3)+1
;pf" ＝一一一一一一一一一一一一了 (5 22)
''"'' I」8(N 一 1) (//#)

Tius indicates that for a sapphire ball‘ the angular spherical aberratton. measured 帥

radians. is rll rnpp. : 0.034 / ( fl# I)3 ‘ which is about

、J
18% better than a lens. made 甘om the
S自由 material and shaped for minimum spherical aberration.

5.4.2 An aspherized ball lens

To elirnmate sphencal aberration‘ a t凶l epoxy shell can be attached This is us聞lly
done by a method called optical replication It is a 討這nsfer process from a highly polished
mold that has the opposite shape of the d目ired 品ph叮c. Figure 5.14 shows such an
aspherized sphere. The correcting shell has been placed at the front surface. Because
epoxy does not transmit in the MW玖姐d LWIR regions‘ this method of 出pherizing is
li1nited to the shorter wavelengths‘ including the NIR. Aspherizing by diamond turnin臣
、Nill be discussed in Chapter!O.

Before the ball is 晶pherized司 it should be ground into a cylinder‘ as indicated in

Fig. (5.14). This is necessary to establish a centerline for the rotationally symmetrical
asphere.
 Cl!APTER 5. SPECIAL OPTICAL SU盯A且SAND COMP叫住NTS 95

』h

Epoxy shell
'
.., R .. bf/""
咱 f 一 ..
FIG 5.14 Ball lens wi血 a叩henc epoxy shell added to correct 中h叩開l aberration.

5.5 Gradient Index Lens

Instead of a ball lens. a gradi阻nt index lens is frequenlly used 晶 a fiber coupling
element. Such a lens h田 a gradient profile in which the refractive index vwies in the
direction pe叩endicular to the optical 阻尬， as expressed in Eq. (5.23).

N 早 N0[I-(k /2)r'] (5.23)

where N0 is the base index (at the center of the lens). k is called the Gradient Consla耐，血d
1 is the vanable radi阻（ mm). Figure 5 15 shows a gradient mdex rod (GRIN rod) wtlh the
length of one full smusoidal path. or one ·pitch."'

Object at ’, Image at
front surface 一 自訂 surfi缸e

FIG. 5.15 A GRIN ro止。”e ""pitch” long.

Shortening the rod to a length of 1/4 pitch. as shown m Fig 5 16、 forms a lens wtlh a
lucal length of

(5.24 }
f - N0.fk
96 0阿7CAL DESIGN FuND.品fENTALS FOR INFRARED SYST曰佰

一 114 pitch, object at

in街tity

FIG. 5.16 Tue gradi阻t index lens.

Le阻es of that kind are commercially available under the trade n由ne SELFOC‘ a
derivative 宜。且l ..,.1ιfocusing ...

If the length of such a lens 阻 identified 品 I. then the focal length is expressed by'

／一一一J (5.25)
- N 0 .fk sin( t .fk)

The back focal length is

研＝ f cos(t.fk). (5.26)

Fi皂山·e 5.17 sho、;vs the details of a GR別 lens with a central index of refraction of
N0 ~ 1.5834 and a gradi閥nt const曲t of k = 0.1067. With a length t = .i mm司 we obtain by
using Eqs. (5.25 ）組d (5.26) a focal length off= 2 mm. and a back focal length of bjl =
0.52 rmn.

.. f ...

... bjl ..
...

FIG. 5.17 Details of a GRIN rod lens for 祖 object located at 帥fuuty. Such 間 element is also
known as a }Vood lens.

Figure 5.18 shows the same lens as Fig 5.17 but now the object is located at a finite
distance. To better unde間tand the function of a GRIN lens司曲”eqmvalenf' thin lens has
been superimposed for reference. This clanfies the object and image d阻lances and. with
= =
the fields added. the magnification m s '/ s h '/ h.
 Cl山 TER5 S閉口Al. OPTICAL SURFACES 且租 COMP< 悶悶s 97

Equivalent thin lens

h'
h

... s
....‘ s’
’,

FIG 5.18 A GRIN lens imaging an object localed al a 品ule disl血ce.

5.6 Conic Sections and General Aspberes

C011ic sections are called o叫 m Fig 5.19 They are of special interest to the optical
designer Their dist凶ct characteristtcs allow the ehminatton of spherical aberration for
悶fracttve as well as reflective surfaces. That is why they 叮e 叫despread for IR
applications. Furthermore‘ al血ough it is still costly to produce 晶pheres in glass‘ it is quite
economical to machine them for the MWIR and L W竄， even in very small ~uantities, by
smgle pomt diamond turmng. Materials such as germ曲ium. zinc sulfide. zmc selenide‘
and Amtir lend themselves nicely to thi阻 process. Metal mirrors. mostly nickel” plated
aluminum structures fall mto the same category

P訂泊。la

FIG. 5.19 Conic seclions

For the sake of comple阻ness we include the circle as a comc section in our general
discussion. The others are the ellipse‘ the parabola. and the hyperbola Our intere這t hes. of
co叮se. in surfaces that are generated by rotating these cross sections about an axis of
symme訂了v.
98 OPTIC且 D目1GNFUND.刷ENW.S FOR INFR且即 E站了晶宮S

5.6.1 Mathematical expr晶sioos

The general eq間tion for a conic section going through the origin of the C酬esian
coordinate system is given by

y2 -2Rx+ px2 = 0, (5.27)

WI血 p being the COlllC 叩間臨別司 sometunes refe叮ed to as the aparabollc constantτhe
relation to the 晶pheric constant K. also often called the conic const：曲t.isK=p l.Ris
the radi阻。f the vertex crrcle.

p detennmes the shape of the conic section.

For p>l Ellipse (oblate). y

p=l
l>p>O
Circle.
Ellipse {prolate).
4 ‘ Vertex
p=O Parabola. • Circle
R
p<O Hyperbola. 。.. x
。
Explicit exp甜甜的且s for a comc section are
Come

R-R二歹 (5.28)
p
2
mw

i }+ 4i•p- c;-vw:
一－

一w

X =
由td (5.29)
－

with c = llR. Equation (5.29) is often prefe叮ed because it includes the special case of a
flat surface, for winch c思 0.

Frequently. especially for manufacturing purposes‘ it is desirable to know the 阻di us

of a sphe甜 that contacts the conic section at the center and at the edges as indicated in Fig.
5.20. This sphere is called the b凹，七月t sphere (BFS). Its radius can be calculated from

p BFS = R+ 0.5川x_= R＋~［川正-;;ii']. (5.30)

Also of mterest ts the slope angle at any given point on the conic surface (see F嚕。
5.21 ). This is simply the first derivative of the curve's equation. which is

dy R- px
一于＝ tana ＝…一一一一－ (5.31)
ax v
where R is the vertex radt血。f the come surface

..-vd sJr

」間，／texh
tangent

“’。甜、

時
－
tpv
－

hu
nrvil

eBI
y
‘

MM
AD a

－
a

4
dFhnNA4
?

J
p (<.y)

間間
／、

nuT
Jat

PJ
• I
x
•x
一抄即可：
~9
hF

-Y
、＼揖

叮叮
E

到
、、

、電
?
-
v
nv

FIG. 5.20 Vertex sphe間曲d best-fit sph叮e FIG. 5.21 Slope 個gleα

5.6.2 Reflectors with conic section surfaces

Applied as reflectors. the family of comc sections c個 be described 扭曲 evolutton

from the circle An object placed at the center of the circle wtll be imaged at the san時 spot
lliis is indicated in Fig. 5.22. F1 is imaged at F2, as F1 is displaced some fi泊te dis旭nee to
the left. 也e conic section becomes 祖 ellipse. Next, we let F 1move out to infinity時 and the
shape of the section 恥H盟 into a parabola. Final紗， one can think of F 1 reappearing from
infllllty on the right sight of the illustration, where it becomes a virt間l object imaged by
the 1~.Y!'erbola at F2. The diagram indicates that the second branch of the hyperbola is 叫so
a valid optical surface and can be used to form an aberration-free on” axts image.
 Circle (Sphere}

Ellipse (Ellipsoid)
F2

oo .. F1
Parabola (Paraboloid)

F,

Hyperbola (Hyperboloid)

fl

FIG 5.22 Conic section reflectors.

5.6.3 Lenses with conic section surfaces-'

Demanding eq阻i optical paths from a plane reference surface to the focal pomt or a
lens. for all rays of a light bundle. yields a lens shape that eliminates spherical aberration.
One is sho、叩 m Fig. 5.23. where the front surface is an ellipsoid and the second surface 站
concentric to the focal point of the lens. The spherical aberration is corrected by the front
surface aione. and the lens thickness is immaterial
 W Ith reference to Fig. 5.芯， the simple relationships for the surface radms‘ the come
constant‘ and the ellipse ha！ι缸is dimenswns are

(N -1)
R1 ：一百一 B‘ (5.32)

and

(5.33)

The l叮ge half 缸峙的

(5.34)

and the small half ax時間

b= (5.35)

...
’,

R,
’,
F, .. F,.
。 。 、，
"' 。均γ于F勻~N。
司，

~ a
’,

叫‘ B
’,

FIG 5.23 Singlet 叫tlJ elliptical front surface and 叩herical rear smface

Another singlet. correc阻d with one conic surface‘ JS a pl曲。同convex 可pe with the
curved surface being a hyperboloid This configuratwn 1s 11lustrated m Fig 5.24. Again.
the parameters for the hyperbola are derived by maintaJrung equal path lengths for all rays
across the lens diameter.
 •
.•
• F
。
.,
N

。 B
(+} ’,

FIG. 5.24 Positive singlet with piano front surface 阻d h)perbolic rear surface,

The relations for the second surface radius and the conic constant are

R、 ＝ 0-N)B 、 (5 。 36)

and p=l-N2 (5.37)

These equations are also valid for a negative lens 描 sho＇咽 m Fig 5 25. One JUS! has
to observe that B h品 a negal!ve value in 由is c由e.

‘ r

一，

F。
I N
’,

冷笑

B 。
‘ (-)

FIG. 5.25 Negative smgl閻明自 piano front surface and hwerboltc re缸 surface.

It can easily be seen that combimng 棚。 such lenses mto one and selecting the
appropriate hyperbolic shap間 Will yield a singlet co叮ec阻d for spherical abermtrnn for the
C品e of f田ite conjug揖es. as indicated 包 Fig. 5.26.
 ~..
。 J ’, O'
<>
。

. B, B立

FIG. 5 26 Singlet 、Nitb

’‘
two hyperbohc surfaces coπected
.,
for spheric ii aberration, applied to
finite conjugates.

5.6.4 Common two-mirror configurations using come section surfaces

There are a number of two-mirror systems that have become standards. These are of
special inte闊別 for IR applica區ons. In principle. they 叮e alike because they have a
concave primary mirror and a convex secondary m盯ur. The Gregorian arrangement is
diffe闊別 As was pomted out m Chapter 3. its secondary mt叮ur is also concave

I. The cla叩cal Cassegmin 叮rangement consists of a p叮abohc primary and a

hyperbolic secondary m盯or This configuration 1s corrected for spherical aberration 甜d
is‘。fco叮se. without chromatic abe甘ation

2 Because a convex asphenc miπor is difficult to measure ‘ a system was developed

that uses a sphencal secondary mrrror. TI1e spherical aberration correction is achieved by
shapmg the pr凶1血y mirror 泊to a prolate ellipsoid This configuration is named after its
mventors‘ Dall-Kirkham.

3 A system that 1s corrected for third” order spherical aberration 胡卓 coma 1s

aclueved by using two hyperbolas. It is called a Ritchy-C/11etie11 system

The equations for the basic radii of the mirrors have been presented in Chap扭r3‘ as
Eqs. (3.57) and (3.5的 Expressions for the conic constants and the uncoπected aberrations
of the n由加d configurations can be found in Ref. 4.

5.6 5 General aspheres (surfaces of rotation}

In modifying a basic conic surface. higher欄order deformation coefficients are added

to the equation. This provides an ex信a tool for the correction of some abe叮at1ons

The general rotationally symmetric 田phere is presented by

x ＝~－~ 叫 1月十 A" v6 +.....+A, vj (5.38)

l + 'J 1- pc'y ＇ 『W ~ ,.
 There 由它 two special c晶es: first. if p = 0‘ the base surface is a sphere: second司 tf
也e c叮va扭扭 of the surface is a flat ( c 宜的咐th only the deformation coefficients applied.

5.6.6 Two conic section mirrors with a general aspberic correc個r

This 甜Ta11gement
allows the correction of third-order sphencal aberration. coma. and
血ttgmattsm It 的 depicted in Fig. 5 27 Because the correct前 is a refractive element, some
chro血attc aberration 、吋ll be introduced. 日。、明ver‘ smce the ωπ·ector is 血serted close to
the focal pl祖e and its thickness is small司也e effect is minimal. The given eq凶.tions have
been developed by using the standard third-order surface contribution expressions.'

Se凹且也可 Primary Tertiary ﹛Corre曲，γ）

華
.’。
戶已

藍
。
d ’, e ..
‘ .,

.. b
”,

..
f ..
FIG. 5.27 Two coruc section 血irrors m血 aspheric corrector.

The comc cnnstan組 for the primary and secondary mt討ors can be calculated 甘om

I ., 1bd'f(b+e) I
－ ~I 一（b-j)+
j L (de-ef +bf)J
i (5.39)

for the pnmary nuπ·or. and

p, ＝品了﹝2川2 叫一晰叫一 b)(d 一卜的

+ 1d'e(d - 2/) 1 巴 1
一一
(e-b) 1r• (5.40)
品or the secondatγnurror For the 扭rt1前y element, the co叮·ecto民 the fourth-order
defonnation coefficient is
4; = b（（｛＿一－ 2f) (5.41)
16e'(l 一川 （ e-b)(de-ef +bf)

N 1s the refractive mdex of the corrector matenal

As elements are added to a system. developing closed solutions for the predict10n of
the ahen就ions becomes more complicated 1bat 祖叫1ere the computer com目恆的 the
picture One begi阻明白 a pronusmg fundamental arr祖gement 血d 唱的 optimization
routines to fmd a useful solution. The trick is to know what a promis旭E 血Tangement 時‘
for which there is no substitute for experience.

5.6.7 Three回mirror configurations

mentioned 曲at due to their wavelengths insensit1v1ty. mirror

It has been already
configurations 訂e frequently favonte candidates for infrared systems. Therefore特 some
limited remarks about three-m盯or systems will be included in this fundamental text.

As with refractive syst臼凹， increasing the number of el叩ienls in an all欄reflective

system provides additional opportunities lo eliminate or coπect more aberrations.
However. the simplici~ which still exis個 for most two-mirror systems is quickly lost and
with that司 it becomes in叮easingly more difficult to find an opt1m凹n configuration to
satisfy given req倒閉men ts

Figure 5.28 shows an m1tial tliree-m1rror layout It Is import叩t to notice that all three
mirrors are centered about a common axis of rotation. which is also the optical axis The
three muTors are parabolas and‘ therefore‘ spherical aberration 1s completely corrected.

z14EEast
2 3
－ aμze

Optical Axis
-

• J; ’ ‘ / ,
‘ f、 ’,

.. ；；.如 •
 In a real system. this initial layout may be modified to 凶e only off-axis portions of
the individual elements 田 indicated in Fig. 5.29. This modification shows also how the
prim目y and tertiary mirro問 C血 be combmed into a monolithic stntcture. suitable for
smgle-point diamond turning (The dt甜nond turnmg process will be described in
Chapter. IO).

恤叫叩
叫阿岫
血伊』間
扭曲n也可
muror

opti且i 姐姐

回iagepl阻e

FIG. 5.29 nu自－mirror system wi由 monolithic primary and tertia可血盯·ors.

It is someti血es helpful to pie仙re a mirror svstem 晶 lenses. Figure 5 30 shows such a

lens-eqmvalent layout of Fi皂， 5.29. It becomes clearer what the system ‘ s effective focal
length 1s. Notice also the decentered aperture stop which indicates the partial use of the
lens一－or mirror elemen胎、 centered about the rotational mds.

D回叩ler吋
Ap軒t阻＜ Stop

• ls向 ’,

FIG. 5.30 Lens• equivalent configuration of three咀tlrror system show in Fig 5.29.

Figures 5.31 and 5 32 show two additional three-nu叮or configurations. one as an

objective. the other 扭曲 afocal system Both present a clear demonstration of the
heo”，而d 盯m叫，叫 tv nf cnmc 肥ctinns
 S=nd>ry(h＂＂扭扭）

。開國岫
S俱冊曲y{hyf間bo峙－
... v ～一一 。1，岫JAx;,

一斗三三l
／正向剛剛 T閻明＂＂＇曲。岫

~·1G 5.31 Three nurror objective yste n1‘ FIG 5.32 Afocal three-mirror system

All this systems discussed so far have been only co訂ected for sphencal aberrat10n To
correct for other abe叮ation requires the proper selection of separation. shape. and
品phencity of the elements as we扭扭曲e location of the aper個re stop. One can see
immediately that 也is is not a trivial task司 and certainly beyond any fundamental
introduction to lens design. The reader who is interested in studying the subject in depth is
encouraged to consult references 5 自1d 6

5.7 Diffractive (Binary) Optics

As a prJctical matter‘ one can look at diffraction simply as ano由er method besides
refraction and reflection to change the direction of light rays. Fig. 5.33 sho、制 the three
different ways 。f阻directing rays by an 叩tical element.

。
’,A ’, ... 。 祖圓茍抄
vv
‘
vbs

, 、’ 一一恤， ’,

Reflection Re仕action Di岱action

(miπor) (lens) (grating)

FIG. 5.33 11uee ways to change direction ofhght rays.

Usually‘ one thinks first of a linear gratmg as a diffractive element. because 1t has been
in t的e for a long time. By arrangmg a diffract1ve phase prolile properly as concentric
rings on a flat substrate. the grating becomes a d在rachve lens. Such a lens is sometm1es
referred to as a single-wave Fresnel lens because of the one-wave opucal path difference
that exJSts at the zone transition心

5.7.1 The simple diffractive singlet

A l:IT..伯、j "'"" P﹔”。＇＂＇也c chnnrn ”，官，α 情之.l i"'£n1i;.:1cti;.: 仇f > nbne”，、＂ rollel di<l 1、叫 h "
surface-relief s甘ucture lies in the pe討odic phase delay introduced by me conunuous
change of the physical thickness of the element according to the ph由e eq間ti on

中仲 ￥.（ar2+br＇ 叫＋ ) (5.42)

In this equation. a, b, c, etc ‘ are the phase coefficients司 λ0 is the wavelength and 1 is
血e radial coordinate. Di份制ive zone boundaries 凹cur at each 27! transition. The
maximum zone depth at the transition is

), "
d ：一－＂－一 (5.43)
血 No 吋 l

N0 is the refractive 姐dex of the lens material at the design wavelengthλ。

‘
a

•r,
D A
。
n zones
呵，

d_ • • I .,
FIG. 5.34 Cross secti 個 of a simple dtffractive lens (single-wave Fresnel lens).

The location of the zones m a d1ffractive lens can be detennmed from the equation
for the classical Fresnel zone plate 前th

九三 ~211'/.. 0/ ＋品： (5.44)

If/» llλ）司
戶在京立、 (5.45)

where /1 is 伽 zone number. Therefore‘你 first Z鵬 radius 'i 三 v2兀7 阻

戶

九三 r1 .J 11 ‘ from which 附 derive the to祖i number of zones required for a lens with a
diameter D.

1 至（~）“一 f
) - 8'!..o(fl#f
(5.46)

Equation (5 .46) indicates 自at the 甜quired zone number for a diffractive lens is
m 品兩個 .... -A ....”φ ，，.. f' fha rohri.<'"'" 1>1lhc.'fT呵，戶”、 ~tPn”l
 Ba叫 on the q間dra叫erm of the ph剖e eq咽肌肉（5.42）］‘ r,= F蒜、 which
leads to the relation 2A 0/ = A 0/a 。 or

(5.47)
。一
- 2f

The mmus sign indicates the direcl!on of the ph品e profile

For a lens with a focal length of／~ 80 mm and a d1ame助。f D ~ 16 mm. used Ill the
infrared spectr阻n at a wavelength of A.i = 4 仰n明 we fit哼，

。＝ 0.00625 mm 一＇. 11 =O 8 m恥 and 11,0叫＝ JOO zones.

,.. =J 13lmm
門口 I 386 mm

F 明＝ 7.96日 mm
I' ＂的＝ 8.00mm

The zone spacing at the ed_g_e of the lens h品 diminished 仕om the spacing of 0.8 nun
for the first zone to 0.04 mm. This is an 1mpo叫ant indicator for the manufacturii唱 process
to he chosen.

As a quick reference appro、iination. the minimnm zone spacing is stated by

（ ~r).,;. :=2λ 。（ f !#) (5.48)

If we select sihcon‘ a suitable ma阻nal Jorλ﹜＝ 4 ~m1. where its index N,, = 3.425 the
maximum zone depth d.,"' = 1.65 ~tm.

5.7.2 The hybrid achromat

For many applications. especially for broadband systems in the infrared region. the
hybrid achromat presents a very interestmg alternative to the convent10nal doublet. ’ By
combmii1g re甘active and diffracl!ve powers. one can effectively co叮ect the chromatic
aberration. TI1e basic pr咀ciple for this correction lies m the fact that a diffractive lens
focuses the longer wavelengths closer to the lens than the shorter wavelengths As
indicated in Fig 5 35. llus 1s.1ust the opposite of the behavior ofa refractive lens The total
power of the refractive/diffracl!ve combmation is the smn of the powers of the elemen阻﹒

中 b油rid ＝申＝中 (5.49)

 ,

1
f-_j,正L .
m

一申
玄
or j (5.50)
. J ；－.－甘、

Refractive + D1ffiactive Hybrid Achromat

。自（shortλ） 。 R (longλ）
。 R (longλ） 。 B (shortλ） 。 R&B

FIG. 5.35 Principle of combllllDg refractive 叩d di叮racllve powe凹 for color correction.

By ass1gmng the proper pow＇叮 to each clement. their individual chromatic

contributions c阻eel.

The followmg 叩nditions m阻t eXJst to aclueve the co叮echon:

(
l
)
(f)
wv
--l f
d
p什
而冉冉－
f

wvwv (5.51)
zzd

-vv
J／

and i (5.52)

with
v'
N＂ 一 I
NS 三 NL
and
,,
'
__
人
d 自 k s-
0
le L 咱

where v, 1s the Abbe number for re仕action. and 川 1s the Abbe number for diffraction ~峙
is the center wavelength for which the index of refraction is N0﹔胡d 沁 and A.1, are the
short- and long-wavelength !units of the spectral band Notice that v. is independent of any
material used for the lens.

With these relationships‘ it b郎。mes quite simple to estimate quickly. based on the
first term (the q問dratic term）。f the phase equation 恆q. (5.42)]. what the first zone radius
will be and how many zones 叮e needed for a partic叫ar le肘。

只＝扛立了逸， (5.53)

1,mru =[ £.；）也 (5.54)

5.7.3 Numerical examples

To correct an/ /I, 100圖mm focal length germ削um lens used in the 8-12
for which we select I 0 i.un as the center wavelength司 we find 血at v, = 990‘ v. = 2.5.
/, = 100’ 253 mm司 and ./d = 39. 700 mm. F叮甘1er司 r, = 28.178 mm. 11toml = 4司、Nith r2 =
39 450 mm. and 門＝ 48.806 mm The fourth zone ends at the edge of the lens. where
r=, = 50 mm Tiie maximum depth of the zones at thetr transitions is dm，，也 z 3 33 µm.

From the equations we can see that large Abbe nwnbers and long wavelengths result
in low zone n山nbers. as demonstrated wi也 the 甜甜nple. This ve叮 favorable characteristic
ex阻ts for most common materials used 扭曲e mid- and long-JR wavelength re皂的ns A
phase profile with only a few zones can be machined very econonucally by single-point
diamond turning.7 (See also Chapter 10)

For a qmck reference. Figs 5 36 叩d 5 37 show nomograms covenng aclrromatic

hybnds for the 3- to 于個d 8- to 12-/.Ull wmdows Only materials that can be diamond
turned have been included in these nomogr.甜拈

In Fig. 5.36句 a ZnSe lens with 50-mm focal lengtl1 /is mvestigated The first zone
radi帥，， = 6 ntrn. For a lens diameter of D = 50 nun (f/1 )‘ the total zone nwnber requtred
to co叮ect chromatic aberrat10n is found to be 11,0,,1 ; 17 Table 5.1 hs旭 all nwnerical
values zone radii for ti時 lens discussed.

Tabtc 5 I Zone radii for f/l ZnSe example lens.

Zonen Radons r Zonen Radius r

l 6.000 mm !O !8.974 mm
’, 8.485 11 19.900
3 10.392 12 20.785
~ 12.0﹛）。 !3 21.633
a、 13.416 l~ 22A50
6 14.697 15 23.238
7 15.875 16 24.000
8 16.971 17 24.739
9 18.000 18 25.456 (D/2 且 25 mm)
 Lens 弘fa扭rial 上oc剖 Len旦出 fl問t Zone Kadms Lens lJ1剖neter 10祖i zones
f r, D n,"""
2 4
Si 2
4 6 3
2 8
6 4
ZnSe 8 10 5
10 6
。 3 15
8
Amtir3 20
20 10
30 4 30 14
40 40 。 20
5
6-0 。 。一 60
80 。 6 30
ZnS 100 7
80 40
150 8 100
Ge 60
200 9
10 80
200 100
12 300
14 150
16 400 200
500
~~
FIG. 5.36 Nomogram for hybrid achromat for !he 3- lo 弘阻n region (obj eel localed at I扭曲1ity).

This nomogram quickly 晶sesses whether a proposed hybrid achromat requires a

reasonable number of zones. For economic re醋。ns‘ the tool radius is kept 晶 large as
possible m smgle-pomt diamond turning. However‘ the larger the tool radius. the more
抽血milted energy is being redirected at each zone transition step and turned into stray
radiation Fortunately。 more 曲曲 20 zones a間 seldom needed for IR applications. The
blockage or shadowing effects caused by the cutting tool c扭扭nount to several percent <
(See also Chapter 10).

-”
For the 8- to 12 n r巳gion. germanium is a supenor material with its lugh Abbe
number of almost 1000 It is in a class of 1個 own comp缸·ed to other diamond-turnable
materials. This is indicated in Fig. 5.3 7 by the large sep即可ation from the other materials on
the material scale. An f II germ血tium lens with 100』mm focal length requires only 也ree
steps (four zones）肺oss its diameter. From Pchrom 至 l/[2v(f ／；的］ one can see that for
germanium. the angulaτchromatic blur amounts to only 0.00的／(fl的 rad.
 Lens Material
Geo

Fo臼l Length First Zone Radius Lens Diameter To個l Zones

/ r, D n"""
4 2
6
：』亨 2 8 。 3
4
10 5
4 15 6
6 3 8
8 20 10
10 4 30 。 14
40
20 20
。 60
30 已 30
40 。 7 80 40
60 。 8 。 100
80 9
60
100
Amtir 3 。
150
。 10 200 80
IOO
12
200 300 150
14
16 400 200
18
500
20

。 30
Zn Se

FIG. 5.37 Nomogram for hybrid achromal for the 8-to u昀帥，1 region (object located at mfinity).

In Fig. 5.37‘ example· I shows an//1 Amtir 3 lens with a 50-mm focal length. The
first zone radms is found to be 6 7 mm司 le缸ling to a total of 14 zones across the lens
diameter. The second ex缸nple '.! 1s thef/l germ缸uum lens discus冶ed above 吶th a focal
length of l 00 mm Here. the first radms 時 quite large‘ measuring 28 3 mm Over the l 00-
mm diameter only 4 zones are required to coπect chromatic aberration
5.7.4 Diffraction efficiency

As Eq. (5.53) indicates. the zone radii 叮e calculated with refe間nee to the center
wavelength .10. At any other 、vavelength‘ detuning occurs. This means that the diffraction
efficiency 前II be reduced.

Based on the s℃al前也eory.' the di鉛action effic間ncy for the first di岱action order is
stated bv

(5.55)

Fig. 5.38 shows the diffraction efficiency E1 阻 a function of the wavelengths ratio
λJλTo balance the e旺iciency over a given spec抽l band. betweenλ1 and "J.., the center
、;vavel開展h should be chosen as

). -21. -
!). '
(5.56)
0戶 k ，＋入 2
宮門a

..

.8

.6
pλJλ
.6 自 1.2 1.4

FIG. 5.38 Diffraction efficiency 晶 a function oftl1e ratio A,, and λ

Tiie approximate average efficiency can be stated by

f1 至 l五﹝~r (5.57)

where LI.λ ＝＝ λzλ卜

As an example for "J..1 = 8 and "J.., = 12 [Lm, the center wavelength for eq阻l efficiency
redm;tion at both ends of the spec甘al window would be
k呦芯（ 2x8× 12)/(8+ I 月＝ 9.6 ~間. Wilh that, the average diffiaction effici間CYE1 =95%. If
we would have chosen IO µmas the center wavelength instead of9.6, the efficiency would
have dropped to 81% at 8 µm and to 91% at 12 戶且 The average efficiency would have
stayed about the s宙間. The shifting of tl1e center or blaze wavelengtl1 can be effectively
employed 岫 favor one end of the spectral window relative to the other. This can be of
mterest for applications where signal strengths need to be balanced

5.7.5 “Useful" spectral bandwidth

Rearr祖iging Eq. (557) and solving for the bandwidth .6'λyields

61.. 戶一－
!ii. 至 －；： .JI- f, (5.58)

Assuming 胡 average dif告action efficiency of 95%‘

Aλ ＝ 0.427λ 。 (5.59)

For the LWIR region. 、社thλ0 = I 0 µm. the ..useful” bandwidth is the扭曲re 4.27 ~un.
At~，＝ 4µ肌 the center of MW鼠， .6'λ ＝ 1.71 間且 and for the visible spectrum withλo=
0.55 ~l, .6'λ ＝ 0.23 µm. These numbers reveal 血at over the common bandwidths ‘ 5% or
more of the radiation Wiil be diffracted 盟to other orders. In other words‘ there will be at
least 5% stray radiat10n. which may cause some system problems One can also look at
this from a different perspective. The stray radiation will ra時e the backgrolllld signal level
and therefore lower the contrast llus affects the modulation transfer function {MTF).9
MTF 咐11 be covered in Chapter 9.

The distribution of the energy into other than the first order will be discussed next.
and an example for the MWIR region 叫II be presented.

5.7.6 Diffraction effi 口ency for a particular order

W1tl1 some ass凹options and simpiifications, the diffraction efficiencyεm for a

particular order m can be stated as

心（令叫﹞ (5.60)

戶﹝仆叫r
Figure 5.39 shows the diffraction efficiency of several orders for a d1ffractive element
designed 伽 the 3- to 5－ ~un region. Notice that the drop at the limiting wavelengths is
equal for the first order‘ which results from the chosen design wavelength λ，＝ 3.75 ~lm.
[Sec the remarks on page 113 and Eq. (5.56)].
 Em

i
0.9
0.8
0.7
0.6
0.5
nVAOnυ
AUT為3 弓－

m=2 m=O
0.1 ~
0 '::- _..,. 1且，“一一一一』－~且·.，：.－－＇. -- ..：.一二T了一一兩’， λ （µm﹜
3 3.25 3.5 3.75 4 4.25 4.5 4.75 5

FIG. 5.39 Diffraction effici個cy for the zero‘ first 祖d second orders in 血e MWIR 扭扭on.

5.7.7 The hybrid achromat句 corrected for chromatic and spherical aberrations

Changmg one of the two surfaces of a singlet to an 揖phere will correct the lens for
spherical aberration. To eliminate chromatic aberration we add the diffractive ph由e
profile 描 described above. A good choice is to combine the two ~pproaches by
superimposmg the phase profile onto the 晶phenc surface. For better protection against the
environment. the second lens surface is prefe叮ed.

We remember from Eqs. (3.5) and (3.6) that a thin lens sl祖ped for minimum
spherical aberration has the following radii:
如一
日－

N
(:
); ...
u-
rA 一叫
、1

）已
R 2{ N +2)(N-l) f ‘一A哼
N

DA fIJ
一－

and 2
（

一，
Fm

,- N<2N ÷1) .I
The 晶pheric shape (conic sectton for third-order correction）。f the second surface for
such a basic lens is expressed by d回 comc cot自姐姐t

p, =I ＋圳4N 一 I) (5.61)
' [N(2N-J)-4]

For companson. the conic constants for the four different basic materials mentioned
frequently tlrroughout this tutorial text arc
NI
Material p, Shape
是

--
J4

Glass 一 18275 Hyperbola

Zinc Selenide .. 6.9544 Ellipse
吋 AU

Sihcon 4 1.6431 Ellipse

Gennamum 1.3125 Ellipse
 Applying these eq咽lions to the jll germ血liwn singlet with a focal length of
I 00 m明1. 由ed for the 8- to 12－~llll window. we coπ·ect for spherical and chromatic
aberrations and find:

First radius R1= JOO 敢當n [Eq. (3.5)]

Second radius R勻＝ 150 mm [Eq. (3.6)]
Conic const割草 p, = l.3125 [Eq. (5.61)]
First zone radius r1 = 28 2 mm [Fig. 5.36 or Eq. (5.45 )]
Total zones required 111"'"=314=4 [Fig. 5.36 or Eq. (5.46)].
Usmg these nwnbers as a startmg point. adding a suttable thickness and the refined‘
r開I index of refraction. and optimizing the pa間neters With a computer yields m a matter
of seconds 閻 achr。mat1c hybrid 也at does not v由y too much from this tlun-lens third-
or甘er’abe叮ation· orrected sm草let司 even for the very low relative aperture off 11. This 1s
another verification of the thin-lens approach for the IR.

5. 7.8 Binary optics

The term bi11G1J' optics relates to 血e manufacturing process of a di岱active s凹face

In the photolithographic process the ph阻e profile is not etched into the subs缸訕e material
as a continuous surface 、吋也in a given zone. but approximated 、Nith a nwnber of steps.
Tllis is illustrated in Fig. 5.40. The rel且ionship between the nwnber M of photolitographic
masks needed to produce K steps or levels is

K= 2M (5.62)

r,
‘ ’,

True Phase Profile Bin由yPh出e Profile

y

d=)J(N-1)
.

FIG. 5.40 Bina可 oplics approximates the lme phase profile.

The diffraction efficiency of such a bmarγstruct叮e is determined by

(5.63)
 From this equation it can be s間n that it takes lo steps or levels to acrueve a ~~＇•

Considering 也at the to個I zone depth is only 3.33 ”

di紡-action efficiency. The nwnber of masks required to produce 16 steps is 4 (24 = 16).
n for a german阻n lens applied at a
wavelength of 10 仰n. the individual ste~ for the b迦ary lens profile is 3.33/16 言。.208 µm.
This reqmres precise control in the etching process. Mo間over. the lateral alignment of the
masks has to be accurate 個 well to achieve concentncity of the zones relative to the center
of the lens. It becomes a very demanding m胡ufactur訟E challenge to add the binary phase
profile to a curved surface. In con恤st. it is convenient to just add the di岱active phase
function to the function of the base surface when the lens material can be machined with a
single-point diamond tool.

An interesting 祖d pow前直JI application of binarγoptical elements is their use as

field lenslets or even innnersion lenslets to achieve an optical gain 、社th foc剖 plane aπays.
Figure 5.41 indicates such an appli間tion. The microlenslet images the aperture stop onto
the detector element. This has 祖 integrating effect over the detector element. More
importantly時 by reducing the detector element s間‘ space 阻 gained around the elements
咐thout reducin皂白e effective fill factor of the 由祖1y.τhis space can be 固ed
advantageously for electrical connections of the elements The arrangement is shown as an
assembly of a micro-optics 由ray and a deteclor array. In principle. the two C臨 be
combined into one component by plac切E the binary lenslets on the front side of the
subs恤te and the detector 缸ray. on the back side. The success of such an arrangement is
甜ictly a matter of manufacturing yields. To make the challenge even great叮司 one can
consider depositing spec釘al 且lters directly onto 也e optical side of the array.

J
E－－

Objective
E一寄一
：L

...
－一一

.
實E－
一費回一
BJB一
攝取

..
一
z

vma
寶”，

v
m－）

‘
、， ..
Binary Micro-Optics Array Detector Array Element Size

FIG 5.41 Bmary mic冊，optics 叩plied to a focal pl個e array.

References

1. H. Dubner, "'Optical Design for Infrared Missile-Seekers ，” Procee益mgs of the IRE

(September 1959), pages 1537-1539.
2. W. J. Smith, Modern Optical Engineering, McGraw-Hill. Inc. (2000), page 286.
3. R. Kingslake, Lens Design Fundamentals. Academic Press (1978), pages 112 個d l 13.
4. W. J. Smith. Modem Optical Engineering, McGraw-Hill. Inc. (2000 )‘ pages 479-481.
5. D. Korsch. R~flective Opti臼. Academic Press‘ Inc. (1991 ).
6. L.CG. Cook, ·Tue last three-mirror anastigmat (TMA)?." SPIE Critical Review.‘ Vol.
CR41 (1992 )‘ pages 310-324.
7. M. J. Riedl司”Predesi伊 of diamond turned defractive/diffractive elements for IR
objectives.'' SPIE Critical Review. Vol. CR41 (1992). pages 140一 156.
8. G J. Swanson, B’，叩門、 Optics Technologv: The Theory and D仰伊 qf M’,/ ti/eve/
D!t汁。ctive Elements. Technical Report 854, Lincoln Laboratory. (1989).
9. D. A. Buralli and G M. Moπis、“Effects of di鈕action efficiency on the modulation
.
transfer function of di鈴active lenses." Applied Optics. Vol. 31. No. 22 (Aug 1992).
CHAPTER6

Design Examples
6.I Introduction

After having covered the aberr割ion effects‘ we can go back to the S亦t calculations
we performed in Section 2且 l for high－祖d low-temperature applications. We shall
analyze several opllcal configuratto即 and deternune the impact of the aberrations for
those chosen arrangements. As an additional design exa血ple. 也e Petzval objective from
Chapter 3 will be modified to improve its performance by using an 晶pheric diffractive
element. At the end of the chapt缸， some remarks about U扭扭ntaneous field of view
(IFOV ）前e pr臼ented.

6.2 Basic Assumptions for the High- and Low-Temperature Applications

We assume that the field ts scan阻止世mg a 的－element linear detector a叮叮﹒ Each
detector element me晶ures 0.5 mm square. With a spacing of 0 I mm between each of the
elements‘ the pitch of the aπ·ay is 0.6 n間1. The baste optical system layout with a smgle
leus is shown m fig 6. l.

間
自
..
"' T

• ..
T

·~
可V

.”
.”
0
可V
d'
因
"' 可V

“ ’, "'
’‘ d'
FIG. 6.1 日臨時 optical s抖tern param前ers: ／~ 200 附注 d' ~ 0.5 mm司品本 0.6 口1111‘
mid 11 ~ 45 elements.

With these gmdeline‘ \ \ e can calculate

h ’= 0.5 (11一 lJ 8 = 13.2 mm‘ and "" = h'/f = 0 07 rad

6.2.1 Optics for high-temperature sys脂ms ( 3-5 µ四）

The 200哨on-focal-length singlet ts made from s1licon wi出 N = 3.4 and v= 235. Its
relative aperture is f 14.

Applying the thin-lens third-order expression from Chapter 3. we find that for a lens
shaped for m扭扭mm spherical aberr割ion R1 = 195.48 mm and R2 = 329.77 mm. The
aberrations add up to a total blur spot size of B 三 0273 mm

To demonstrate the validity of the thin-lens third-order aberτation concept‘ a

thickness of 5 mm w田 added and the lens W訕。伊拉山ed with a computer program. The
results a間 shown in Fig. 6.2.

國．闡剛

f + I·••·· ····!··•• ···1· ··I

島寶

9,29 ””
LI十十︱法（
lJ.2盟問咱

︱ ιI I., I 持 I ,. I
-200 -100 0 100 200
SURFACE' TM<>
THROυGH F口CUS SP口T 口 IAGRf哥們

SPOγ 口IRGRA柯s 3 也 5 ”ICRONS 6 2

SP口T S工ZE UNITS ARE MICRONS.
F棚
E冊~1日咽~到到TH的S :
刊iilJ于~四3日刪除
19臨3 ~~自71 ~~·§§~2 1
273 ：口直EF 間Y

FIG. 6.2 Spot diagrams for optimized sthcon singlet.

The scales in the spot diagrams have been matched with the predicted thin-lens
third-order results for direct comp訂ison (Tite box Width is equal to blur spot dimension
0.273 mm.) To balance the energy distribution over the h.tll field. the concept of field
curvature compensation as shown in Fig 3 19 、問s applied

6.2.2 Optics for low

The lens material for tlie smglet used for this apphcatton is germanium with 由1 index
of N = 4 and an mvcrse relative dispersion of v = 990. The relative aperture is f 12.
 The thin喃lens third-order results for thts case 缸e R1~200 mm‘也~ 300 mm. and
the blur spot siz浴 B := 0.55 阻血﹒ As expecte且也e computer-optimized results also agree
here very well We notice 血at the blur spot is larger than the 0.5-mm detector element
size. Therefore. we look at another configuratio且也e Pe加al objective.

6.2.3 Optics for low-temperature systems, design 的（ 8--12 µm)

Applying Eqs. (3.34 ）個d (3.35) for an/ /2 Pe且val objective yields an on－缸is blur
due to spheri阻1 and chromatic aberrations of B 三 0.062 回血﹒ Adding coma. astigmatism.
阻d lateral color for the 4° field angle brings the off－缸is blur to approximately 0.37 m品
TI1e m吋or contributor is asti~atism‘ with 0.2 mm. o_ptimiz泊s the radii and the image
plane position results m addtlional improvem叩ts This is indicated m Fig. 6.3

＠曲曲 DEG

I· ·";· l :>I··﹔品﹔... /.:i\; I

＿；，·：＇.；~｝＜叮
2.80臨盟G

/1'' （手：（
刊 0080 DEG
/ 0!c/ <f I 部 I \"ti I ：注：！
一 200 一 100 自 1 囡囡 2白白

革!.!RE旦旦旦ι4且旦
τHROUGH FOCυS SPOT 口工伺GRAM

PETZυ伺L. 8 12 ”工巳RONS. 6 3
SPOτSIZE UN工TS ARE 阿ICRONS.
rn主D I 2 3
m悠絢司OIUS 77.212 ?0.335 70 207
GEO 信也IIJS 117.501 112. 岫9 159.B濁
冊X WIDTH 273 REFE船到CE , CHIEF 間Y

的G圖 6.3 Spot diagrams for the gennani 山1 Petzval objective.

6.2.4 Optics for low-temperature systems, design 粉（ 8--12 抖m﹜

As an in扭restmg exercise. we shall analyze an optical system that cons1s 恆 of two

aspheric miπ·ors and an aspheric co叮ector. Tl11s configuration ts pie阻red in Fig. 5.27.
First. "e make some reasonable assumptio恥，＂hich provide the basic dimensions e, d.
and e. With a 25% central area obstruction. b becomes o.yor 100 mm (we remember that
f = 200 mm and f I# = 2). Not exceeding/ /I as relative apertu甜 for the primary mirror
focal plane. i.e” e ~ 25 mm Finally‘阻 matenal for the corrector‘ we use germ血uum with
index N = 4. From the basic Cassegra泊 Eqs. (3.57) and (3.58); we find the radii for the
mirrors. They are identical：凡 ＝ R, ＝一200mm.Eq阻lions (5.39) through (5.41 ）訂e 阻cd
to deterrnme the conic constants for the m盯O白血d 由e defom祖tion coefficient for the
coπector plate:

P1 ＝ 一0. 769231, p, = -20.66667 ‘ A., =-9.57265x10 7.

Both mirrors are hyperbolas. It 個ms out that the Petzval contribution. which
determines the field curvature. is zero for the selected dimensions. That me缸阻 the image
S叮face is flat.

To correct for higher-order aberrations ‘ two deformation coefficients were added to

the basic come sections The final result ts shown m Ftg 6 4

0.ee間臨G 有0~
’
昌，．．．．
@
{

︱合︱ , 勻 晶

。
...圓圓圓 DEG t紛3 、但圖
J，一

呵，眉，圓圓 DEG
I .; , I 咒罵 曹“陸 內
情：

-2的自 -100 自 100 200

SURFACE: ™"
THREE ELEMENTS
1
”
THROUGH F口 CUS SPOT OIAGRA刊
8-12 ICRDNS. 6 刊
2 3
SPOT SIZE UNITS ARE MICR口NS.

F閉U間~乳白際彪眩m也羽到w :
<
919 ~~－ 篇3啊有 ~她，．鬥
21 .567
3可 1于臨自國ENCE
273 : CHIEF 閉Y

FIG. 6.4 Spot diagram for the three”eleme叫 system.

As mentioned before‘ for better compariso且 the box size surrounding the blur spots
has been kept the s血ne for all c阻es shown‘ namely 0 273 mm. the size of the blur spot for
the/ /4 singlet. It should be noted that the central obstruction reduces the amount of the
甘ansmitte凡I radiatton by the ratio of the area of the secondary mirror to the area of the
primary m盯肘， One c祖 express that 阻 m effective (//#)by
 （ρ」 (6.1)

where ε1s
the ratio of the secondary and primary n吐rror diameters In the case discussed
ε 詣。5. w凶ch yields 胡 effective (f /#) of 2.3. It is e品y to see that the SIN 、吋II be
reduced by 25%. If this is not acceptable, the l巫師I (fl的 1臨 to be lowered accordingly
and the impact on the aben祖ions h晶 to be r間valuated

6.3 The l阻proved Pe位val Objectiv•

This example will serve several pu中oses. First‘ it will teach how the matenal
presented so far can be eficcllvely used m the predes1gn stages; and second. it 叫II
demons個阻 how such a fundamental approach can provide a better understandmg of the
interaction among the d的erent p缸缸neters. Finally. 1t will also indicate that arnvmg at a
sui祖ble optical co凶ig叮at1on to meet the existing requirements is not a process often
called ” automatic lens design. 叫

We begin \\1th the basic configuration as d閏cnbed in Chapter 3, especially as

shown in Fig. 3.23.

1 o present the entire design process可 we rep閏t the equations for the radii:

4(N+2}(N-D
R, (2N+IlN f . (6.2)

R、 4(N +2l(N 一 l)
= N(2N-ll-4f 情 (6.3)

2(N +2)(N 一！）

R、 N(6N + ll 一 4 f、 (6.4)

2(N +2}(N-ll
and R, N(6N-ll-8 l (6.5)

The angular blur spot ‘ due to sphencal aberration is‘ as has been already stated ‘
:( !i
12
叭 γ一÷

叮叮－rj
川川、一一
中－ N
N 一“

2
「II－

＋一？一
一－N

-
I」－

QUZ
一一
一叫／－
－nυ

h ky ( -/ 3
) (6.6 }
叩

一，
自

4什
i2

i
．、

To co叮ect this residual sphencal abe叮·at10n. we apply the procedure descnbed in

Ref. l and find that if the second surface of the front element is changed to a conic
section. its conic constant should be'

I I 4( N + 2)( N - 1) l'
 The moal chromatic aberration of the arrangement under discussion 1s

。＝~一－
Sv,.chmm (J 叫‘
(6.8)

where v, 1s the refractive Abbe number.

The equation for the d1ffractive phase profile 田 ed to reduce the axtal chromatic
aberration can be derived from the general ph由e eq閏ti on [Eq. (5.42)]. Limiting the
expression to the quadratic term yields for the chosen Petzval config叮·a110n

27
在一。
( )=

可FU
。
AWO
’P (6.9)

叫th the quadratic phase constant

a ＝ 缸瓦戶可7 (6.10)

V.1 is the di岱active Abbe number that is exp自由ed by the ratio of the design wavelength
and the difference betw宙間 the short and long wavelengths

6.3.1. Numerical example for a LWIR application

To demonstrate the usefulness of all these equat10ns. we shall design a Petzval ob1ective
for U1e 8-12 µm region. By assuming an index ofre仕action of 4 叩d an Abbe number of
1000. the calculations become very simple. For an objective with a focal length of
JOO mm 個d a relative ape山間（／／＃）。f I‘ we find R1 = 2f= 200 mm. R, = 3f= 300 mm.
R.1 = (3/8)[= 37.5 mm and R4 = (317)[= 42.857 mm.

The come constant p for the asphenc second surface of the front element becomes
[with Eq. ( 6. 7 )]
u廿一

AO
VR
「。

「
勻，“A－

「Ilia －L
1

11Ill
+-
4

哼，一弓，但
勻，“－

勻4 －ny

叫／”一句d、
寸，－LU

句3 一勻，“
A

-4
1l
斗－ A且可

i
lsaz
1111ll
一%

川一一

p -ny-6
Aυ
守f

勻，’
。。
去叫

”J
一－

一－

一一
一一

一一
hNF

x
－－

A哼
IIJ
哼

--
J
』

winch represen臼 a prolate ellipsoid The phase coeflicienl a is obtained with Eq. (6. 10).
IL is kFPEF
之一的
1二，
λ

”、
叫一－
一－uu

-- Ay ×i 6
d－AU

-(
仇U
句、
’J
。

-- ~
一勻，“

…Aυ

、
J
ωl

hOXu
J

AU

TI1c first radi阻 of the ph品e profile is

 戶仟 z 0.01
一一一一－：－ =32.70 mm
9.35x!O吋

and the total number of zones [Eq. (5.54)] is

沁－z
fJ

﹛二句 3
\

『t
tililll

弓－
riti

吋、
一一

勻
一一
7

3
一×
一句r－

一句J
叫
AU

－l
,/,

－＼
、、<hich means the profile has 2 full zones. The third one does not reach its full step size d‘
because its radius is larger than the half-di甜1eter of the lens. The second and third zone
radii are
r, = r, fi = 46.25 mm, and r3 = r, f3 = 56.64 mm

11us step size is

0.01
d ＝ 一一－ = 000333nun = 3.33µm
( 4-1)

The average ditlfa酬me的ciency amounts to E 三 1 －~I 土 I "'o.956.

36 \ 10 J

Fig. 6.5 shows the spot diagram of this improved design. While the on－缸is blur spot
for the all-sphere aπargement in Section 3.6 I was 135 ;1m 恆 diameter. adding a
dillractive asphere yields a diffraction limited S）叫em. even for the off﹒血is posttions.
Notice that the circle surrounding the spot diagrams 1s the Airy disk. which 1s 24.4 µm
 心
間叮甸，回曲 DEG 由Z J. qg姻自旺E

sa
－翎計
I鈍＂＇ 0. 磁!0 t1l1
曲Z ' 2. 徵!00 QB;
心
D俯＇ 2. 啊b 剛

也取＂＇司r'.F' T揖Q

HYBRID PETZVI在陪讀到回 FR個 T阻N LENS

心Itll1'
SPOT
DESI酬 6.5
3. 呵呵嘲

口工肉GRA鬥

糊口S 冊E ”ICR由3。 zqg

且8..0 I 67 3
呵

呵呵

閉S 即可or間， 3,7q0 1 233

g

GEO R間 IU5 ' 8 65B 11.506

AIRY 前剛 到＇＇ 使野田8位 ' HI研llE

FIG. 6.5 Spot diagr咽怒 for 血e improved Petzval o均帥ve(hyb耐）. with a conic sectmn surface
祖位 a s呻erimpo盟d di般active ph品e profile on tl1e re前 S前face of the front element

6.3.2 Manufactur岫g remarks

Diamond turning is the prefe叮ed machining _process to genera扭曲e 由pheric surface

upon which the di鉛active phase profile is supenmposed. By properly programming the
motion of the single point d悶nond tool‘ bo也 features can be produced siln凶taneously.
However‘ there is 血E aspect to this process which can be limiting to its usefulness、
especially if the spacing between the zones of the ph自e profile is narro肌 This 啊ll be
discussed in detail in Chapter l 0.

6.4 Instantaneous Field of View

For a perfect optical sys阻m‘ the inst由lt祖eous field of view (IFOV) is strictly
detemuned by the size of the detector element d' Its angular 仇tent is expressed as
IFOV ~ 2的思 d'lf. The blur spot is at best the di!Iraction 刮目 of B = 2.4.＼ λ fl約. but
frequently larger‘ depending on the residual abe叮·ations. In either case. one has to agree
with the user on the energy percentage of the blur spot size that is acceptable for the
application as a field cutoff.
 CHA們ER 6. DESIGN EXAMPI ES 129

呵，
~n~~~I
、，
211,
B 4‘’
... d
‘
a ?

.. I .,

FIG. 6. 6 Blur spot 凹 dins臨別叩開us field of view.

References

I. M. J. Riedl. "P問design of diamond 個med refraclive/dift全active elements for IR

objectives." SPIE C吋tical Review‘ Vol. CR41 (1992). pp. 14心156.
、 W. J. Smith. Modern Optical Engineering. McGraw-Hill. Inc. (2000). pp 332-335.
CHAPTER 7

Thermal Effects
7.l Introduction

Frequently司 optical systems must perform over a wide 阻mperature range. Due to the
thermal expansion and the ch叩ge of the index of refraction of the lens mater祖i with
temperature. the performance of the system is affected This is esp間ially pronounced in
the i的ared region. where most materials suffer from a high dN!dT. 也e change of 姐dex
叫出個nperature. The values for the i扭曲之 of refraction and espe心ally the thermal
coefficients vary-in some cases 叫自ly-in the literature and also 泊 theda旭 sheets from
suppliers. It ts therefore necess由γto confirm or recheck these data

To maintain an acceptable image quality, in many cases refoc由mg will be

necessary. This can be accomplished mecha血cally or optically. Specifically‘ the
mechanical adj間tment can be done m個阻lly or by other means, such as feedback servo
systems and others 1 Optically司 the compensation can be achieved by selecting smtable
optical materials and element ~owers. This latter method will be discussed 叫th the
assumphon that the lens syst目n ts stab1hzed at a given temperature. m other words司 there
are no temperature grad1en臼 pr臼ent

To close out the chapter 司 some remarks are made relating to the cold stop and cold
shield.

7.2 Changing Parameters

The changing parameters of a lens 111 a housing 缸e the followmg:

at temp咽 ture I al tempera阻時

t I tH11

I Lens
radius r r(I ＋α， l!J)

也ickn間 d ÷
d(I a,l!J)

index ofrefi間ction N N刊甜I曲曲

2 Housing (mount)
spacing between s s(I ta.,,l!J)
lens 間d dc'tector

1,, = coefficient of lineai exp祖訓。” for housing

<

叫 z coefficient of linear exp胡sion for lens

dN/dt = change of refractive index with tempera叫阿

13 l
132 OPTIC、AL DESIGN FUN/J.WENW.S FOR INFRARED SYSTEMS

The symbols are identified in Fig. 7.1:

r, r r,
...a..
N
αL
α”

.. s~f ...
(for obj甜 at infinity)

FIG. 7.1 Id四惱cation of changing P扭曲1ete凹 for a lens mount吋 m aho山ing

7.3 Defocus with Change of Temperature

To sunplify the procedure‘ thin elements 前e assumed For an objecttve cons1stmg of

j thin elements mounted in a housing. the focal shift is

午 2 卡的叫tJ..t. (7.1)

、吋3叮e

dN一 I dt
T, ~ γι守的， ~ thermal glass constant.
l川，－ 1)
f~ focal length of system （扭扭mecl to be the length of the housmg).
f, ~ focallength of element i,
N, ~ refractive mdex of element i
dN, /dt = index change of element i,
呵，＝ thermal exp胡田on coefficient of element i,
a"= Thermal exp祖sion coefficient of housin草

7.4 Defocus of Singlet

The defocusing distance ti/ 1s the separation of the lens foe田 from the detector 1 his
is 社epicted in F嗯， 7.2. where a temper酒lure incre阻e is assumed. It is indicated that the
focal length actually decreases with an me甜甜e m 個mp era個問‘、叫1ile the ho由mg expands
so that the total defocusing is the sum of both el叮叮ts. Tuts total separation of the lens
focus from the detector posit10n 1s

D.f = -JD.t(T ＋ αH ). (7.2 }

Some of the most frequently used IR lens materials 附th their respective gl阻S
constan臼 are called out m Table 7 I BK7 ‘ the verv common lens material for the visible
spectrum. ts mcluded for reference Notice that i臼 gl 描s constant ts negative. winch means
that the focal length for a lens made from BK7 mcreases with temperature.ηllS IS the
 CHAP'距R 7. T!租恥1ALE何＂ECTS 133

叩posite from IR lenses made from suitable malenals for that region司 such as gennani凶l
silicon. zinc sulfide. and others. Typical housing materials such as aluminwn and steel
wtlh the1r !hernial c間的口ents of expansion are also included in ’Table 7.1.

~
·>I !'if，叫‘
’
Bl叮 Spot

之
mcrease
．色｛，且‘
UCρ

‘
....
…

t +D.t
u念

4
－
b
n念：…﹒

l'ifT

FIG. 7.2 Components of focus shift.

Table 7.1 Glass constants 胡d coefficients of exp叩SIOO for 叩me frequently used materials.

Lens Glass Constant Housing Coeff’。f Expansion

Material T × 10... Material 。＜ uX IO~mmi圓圓1•c
BK7 -1 VIS Aluminwn 24
ZnS 25 IR Steel 1015 12
ZnSe 36 IR lnvar 36 1.3
Si 601R
Ge 1261R

We shall 叫ialyze with Eq. (7.2) the thennai effects fo川 gennani叩1 日nglet with a
focal length of I 00 mm mounted in an al山nm凹n housing. The temperature excursion is
as批une位 to be +20。c.

~r~ ~ 100× 20x(l26+24 )× l O 6 = -0.252 0.048 = -0.30 mm.

t t t
6Ji.m，吼叫時 4Fr叫

TI1is litlle exercise sho附 how devastatmg a change m temperature can be without
any co叮ecuon for the image pos1110扎 if one considers that the tmage blur spot increases
according to 68 = 6Fr叫l(f版）

7.5 Athermahzabon with a Doublet

For two elemenls Eq. (7.1) simplifies to

j
九－
九一戶，心
「lili －－ 1」

\
tiji11IJ
人心

FJ +
斗。

@,,
沌叫

f
L

f
, nu
吋
一

] J (7.3)
134 OPTICAL DESIGN FUIVD.品，＇ENTALS FOR INFRARED SY.且EMS

Element A is the 仕ont element and B is the 阻ar element. To eliminate any focal
shift. we set !if ~ O. For that c甜可 the focal length of the front element is

f 干（
（干九） (7.4)
4 可（九＋ Cl H)
and the focal length of the rear element is

f 一些主l
(T, ÷UH)
B -
f or (7.5)

Applying these expressions for a doublet mounted in an aluminum housmg with a silicon
front element and a germanium rear element leads to a focal length of 44 mm for the
silicon lens. and 78.6 mm for the germ阻阻n lens. The desi伊 wavelength has been
chosen to be 4 間羽 If we select a convex-piano front element and a piano-concave rear
element. the respective radii are r1 = !05 6 mm‘吧：~. r3 ＝ ~.and r.1 = -235 8 mm This
is a good starting pomt for a computer opllmization after addmg 3 血1d 2.5 mm for lens
thicknesses‘ and a spacing of I mm between the elements. II 1s important to remember
that while opt世1izing the radii to reduce spherical abeπation. the focal lengths of the
elements m田t be maintained. otherwise the athennalization effect 叫II be lost. Figure 7.3
shows the expected blur spot sizes at 一切。C and at +60°C for an f 14 doublet. designed
and optimJZed for 20°C. The di般action blur spot sJZe is sho"'n for 悶ference.

Si Ge

Diffraction blur spot size 盟””n

. .
-20 c +20 c 吋0 c

FIG. 7 3 Ath 臼mali且d doublet (Alhermat). TI1e blur spots indicate that the 叮嚀tem remains
di甜action lit血led over a temp 前刮目e exc世sion of± 峙。c
 CHAP昆R 7. nIER必!AL EFFECTS 13 5

7.6 The Athermalized Achromat

The equatio肘 for the doublet analyzed in the pre\1ous secllon 缸e vahd 犯r one
wavelength only. Such a doublet would therefore be suitable for laser applicatious or
other monochromatic fonctions. To achromallze and athermahze 四 objective, three
conditions must be satisfied. Thev are

申＝ I 申‘ (7.6)
'"'
工（肘，）＝吋 H中可 (7.7)

and
寸
J 」叫

AU
(7.8)

一 NM-I
As pre、甘usly st剖凶，川s the Abbe nwnber. For refractive elements‘ v, －一一一一一一俺
N8-Nr
),
and for dilfractive elements. vJ ＝一一」ι－ The roots for Eqs (7.6卡（7.8）訂e
“ λs 一 λ L

l/v1 'li-+: ati

l/v, l/v 電 T,- T;
11r； 三中 2
乖于了中
= 0 • (7.9)
!Iv, …1 Iv}.
l/v3-l/v1 耳一 R

l/v,-1/v 令 I Iv.
I ／人＝中＝一一斗一一一中 一一一ι一申、 (7 10)
; 3 l/v3- llv, Y l/v3 liv, Y

111; ＝中 l ＝ 中－中 2 心中 3ι ( 7. l l)

dN ldt
For refractive elements‘竟說？一一γ － a1_. and for di品active elements、九＝ laL
Ii' - ll
To apply these seemingly cumbersome equations to a practical example. we extend the
design of the doublet from the prev10us section to cover the MWIR region from 3 to
5 抖m For the first design司 three lens elements will be used In the second desi皂n. the use
of a diffractive element !or 叩 athermat will be demonstrated. As before ‘ the elfecllve focal
length is l 00 mm and the relative aperture is 4.
136 0月1CAL DESIGN FUNDAMENTALS FOR INFRARED SrsTEMS

7.6.1 The a ”,refractive athermal achromat

As lens matenal. we elect for

Element I: silicon 咐th v, = 236句祖d r, = 60xlO們（ see Table 7.1)

Element 2: germ血U阻II with V2 = 103‘個d T,= 126xl06
Element 3: zinc sulfide withη ＝ llO‘ and r, = 25 × 10·6.

The housing is made 仕om al祖閥n間n with 曰：1 認 24xl0 6. Eqs (7.1). (7.2). and (7.3)
result in

1/236 60+24
申＇＝ ，！？，！”一H~ ~一ω ×0.01 =2:旦旦旦旦x0.01 = -0.0108630 ‘
＇，.，， ι 冒，， ＂可＂＇ Z 一－3 0130066

I/I IO …1/236 25-60

1/236-1/!03 1/236
申 J ＝一一一一一一－ x ( 0.0108630）也一一一一一一一－ x 001 = 0.0035156 .
1/110-1/236 l/ll0-1/236

中 I= 0.0 卜（”。010863) 0.0035156= 0.0173474.

The respective focal lengths are /1 = 57 .646 m明l. 五＝ 92.056 mm. andfi = 284.45 tmn.
Asswmng plan伽convex shapes for the first 祖d third element and a biconcave shape for
the center element provides a good starting configuration for the compu扭r optimization.
Fig田e 7.4 shows the optimized configuration and the expected blur spots.

Si Ge ZnS

Di世action blur spot size = 39 µm

. @

·20 。c +20"C +60"C

FIG. 7.4 Athennalized achromat (Triplet) for tl1e MWlR region.

 αL＇＂江R 7. THERMAL EFFECTS 13 7

7 .6.2 The hybrid a thermal acbromat

For this design. the zinc sulfide element is elunn祖ted.' A d旺ractive ph由e profile on
the rear surface of the germanium lens takes its place With the reduction from three to
two lenses‘ such an 誼祖且皂白neut is cost effective and especially of mterest for reducing
spherical aberr誼ion when the phase profile is su.perimposed on 甜甜pheric surface. since
the asphere and the diffract1ve structure c血 be diamond 個med al the same time.

As before. the p叮ameters for the fi時t and second element remam as

Element 1: silicon 叫你們＝ 236守 and r, = 60xl0-6 (see Table 7.1)

Element 2: germ抽血m with 阿拉 1的同 and r,= 126× 10 6

For the d1ffract1ve surface‘

VJ= AM/(As-A,)= 4/(J-5)= -2 ‘ T, = 2α L : 6.2 X JO-o

We 問member that the thermal coefficient of expansion for the aluminum housing 1s
αH = 24× I ff 6 Inserting these values into Eqs. (7 .9), ( 7. IO), and (7. I I) yields focal
lengths of.前＝ 43ιIO mm. 昆＝一78.299mm‘ and.Ii= -I 且332mm.

The perfonnance of this athemlalized hybrid triplet is shown in Fig 7 5 The temperature
excursion is again± 40°C.

Si Ge Di飯active phase profile superimposed

on r且r surface of Ge-le血

Di借action blur spot size= 39 µm

電盔 @ 、J

.”
-
-20。c +20°C +60。c

FIG 7 SA曲 em1ah且d achromat (hybrid) for the MWIR T噁ion.

138 OPTICAL DESIGN FUND.品恆NTALS FOR INFRARED Sr.~IEMS

7.7 Cold Stop and Cold Shield

7.7.1 Cold s個p

In Cl阻pter 2. the cold stop was mentioned 品 a spec咽l stop As shown 咀 Fig 7 6.
the cold stop is 也e aperture stop mounted inside 也e D訊問r package. In this arrangement.
the detector sees only rad酬。n from the 喇ect space. Re-radiation effects from the stop
itself are el迦linated or at least co由1derably reduced One of the most important
observations to be made from 也e Fig 7 6 時吐旭t the cold stop is not only the aperture
stop, it is the exit pupil EP’晶 well. The detector is the field stop and the 回it window
EW'. and therefore‘ wi曲曲e aperture stop mounted a dist祖ce 1.αp = S ’ away from the
detector. the energy throughput eq間tion [Eq. ( 2.4)] is already satisfied without any
consideration for the nnaging optics. We can see that With a round ape叮叮e and a squ盯e
detector.

r~~~.
s"
一一－
41;.op 4 (f紛）＇
(7.12)

The importance of these relations is indicated by the fact that the V的 of the
objective is determined by the detector packa惡e geometry and any reduction of the
objective's (//#)is meaningless and only wasteful.

Obj 血tive Dewar Window Cold Stop Detector

‘
4

可 電台
..,
D " d'
直 -c+-=i A.

t詞句，
,,.
.‘

FIG 7 6 The cold stop as system's ap前1山e stop 血side the Dewar package.

We also noticed 廿·om the simple optical layout that the lens diameter becomes qmte
large to cover the enttre field without vigne臨時. To solve this problem‘ re1maging optics
can be applied.

7. 7 .2 Cold shield

If the cold stop is not the ap巳rture stop. it does not control the radi剖開n cone from
object space. Depending on the position within the entire optical train. the portion of the
 CHAY距R7‘ I 阻恥1ALEF甜口s 139

undesired radiation impinging upon the detector varies, and one defines this variation 由
the cold shield efficiency. A 100% efficiency makes the cold shield a cold stop, because
in that case the cold shield has become the aper!叮e stop of the system. 3

References

I. P. J. Rogers '·Athermalization of IR Optical Svstems,” SPIE Vol. CR 38 (1991) pp.

69-94.
2. R. M. Hudyma ” Applications of Diffraclive Optical Elements in Infrared 。”也叫
Design:‘ SPIE Short Course Notes 1994.
3. R. E. Fischer I B. Tadic-Galeb, Optical 句wtem Design, McGraw-Hill. Inc. (2000).
pp. 221-224.
CHAPTERS

Optical Coatings
8.1 Introduction

Optical coatings are essential to control the radiation throughput of an optical

element. Referring to the basic relation

T+p ＋ αt L (8.1)

whereτ~ transmillance‘ p = reflectance‘ and α ＝ absorption, we are pnmanly interested

in the first two contributors. Most of the llme. we like to see either a high transm加ance
for refractive elements (lenses）。r a high reflectance for reflectors (m肘。rs). In the
applications for beamsplitters‘ we like to maximize the efficiency between the transmitted
and the reflected energy. which me由阻 the absorption m阻t be kept low

Smee the subject of optical coatings 祖 complex and extensive，也e coverage will be
limited to some fundan1ental behaviors of 也in film coatings 咽d to application remarks.
Of special interest for the infrared spectrum are antireflecllon coatings and bandpass
filters. Without these, the enl!re field of IR optics would be very much restricted.

An important breakthrough c祖ne in 1935，、社ten Alexander Smakula from Carl Zeiss

111Jena. Genn叩y‘ received his patent for "a method to increase the transm1ss1vity of
optical elements. by reducing the index of refraction on the boundaηr layers of these
optical elements ” 1Afewye缸s later, in 1939, Wal.ter Geffcken received bis patent for
transmission句·pe mterference filters 2

8.2 Effects at a Single Surface

For radiation at nomial incidence‘ the reflectance for a smgle surface is expressed by
the fundamental Fresnel equation

p, ＝（出）－ (8.2)

which rcmams a good approximalton for ltlt angles of up to 45° Neglecting absorption.
the tmnsmiued portion is then

4N
t, =I p, ＝一一－－；－ (8.3)
. (N+I)

141
 CHAP'時R 8. OPTICAL CoAτINGS 142

Figure 8.1 shows the incident ray being separated into a reflected part 血d a
refracted fraction that enters 也e substrate

Incident ray Reflected ray

Substrate
or film

FIG. 8.1 Relleciion 祖d transmission on 祖 uncoated single su巾cc.

8.3 Two Plane-Parallel Surfaces

As shown in Fig. 8.2司也e effects of reflection and transmission for a substrate with
plane-parallel surfaces can be easily detemuned司 even if the mternal reflections are
mcluded At each surface bound田y there is transmission and reflection The existing
internal reflectio阻 can be quite disturbing because these 凹flections will genera扭
secondary or ghost images.

FIG. 8.2 Transmission and 阻llection at a pl咽e-parallel plate.

By including the internal reflections in the equation for the transmitted fraction of an
irnpmging energy bundle‘ we find

2N
τ＇－ TN'可 (8.4)
143 01'11CAL DESIGN FUNDAMENTALS J.OR INFRAREDS~訂且必

If the in阻rnal reflecl!ons are ignored, the equation ch祖ges lo

什品了r (8.5)

In compar泊g these two eiq>ressions for different materials, i.e. ‘ for different indices
of refraction‘ we notice 也at for glass with mdex 1.5句 the two values differ by less than
!%. For germ缸世間丸 W!th I阻 high 阻dex of 4, the numbers contrast by approX11llately
13%. While this is relatively ht叭， it is more important to notice the absolute transmission
differences between a lov叫“ex 個d a high-index material. For example，以甜甜田tsmits
about 92%‘ but gennani阻n tr叩smits only half of that. That is why an antireflection
coating for these high-ind間 materials句、吋1en used for IR applications. is absolutely
requ甘·ed.

8.4 Antirefiect帥n Coatings

111e import缸tee o! antireflection coatings (AR coatings) can be judged quickly by

the number of elements required for a typical zoom lens. Ten elements, or twenty
surfaces. c祖 be found easily m such a l阻S Assuming all-glass optics with 叩 average
index of re告action of I 5 ，、社th a 4% loss per surfa阻， the total transmiss10n would be
reduced to approximately 44%. The loss increases dramatically as the index ofref全action
mcreases. For gennanmm. one of the prinuuγmaterials used m the infrared region. the
index of refraction N 至 4. This high index yields a transmission of 41% for only a single
element.

To coπ·ect for a reduction of such magnitud此 a method was devised by which the
transmitted amo間t of radiation is increased by reducing the reflection losses. For such an
aπangement‘ the two requirements for a single layer coatin品 based on interference. are
material comp剖ibility for equal amplitude

NF ＝ 抓耳﹔ {8.6)
and thin film thickness

N.1.
•. .. =2_4 λ (8.7 }

where N1, is the index of the thm film. N、 is the substrate。S index, and N,, represents the
index of the surrounding medmm, which is usually atr.

The pnnc1ple of the intemct10ns at surface boundaries is that there is a half-wave

phase change 叫阻n light travels through a low-index med間n and 1s reflected from a lugh巴
index medium There is no pl祖se change when hght travels through a l11gh-111dex medmm
and is reflected from a low-mdex medmm. Therefore. as indicated in Fi臣，且3. part of the
incident radiat10n will be reflected at surfaces I (bean1 B) and 2 (beam A). undergoing a
half-wave phase change. If the optical thickness of the thin film layer is made one quarter
wavelength [Eq. (8. 7 ）﹞司 the relative phase shift between the two reflected beams is an
 CHAP'肥R 8. OPTICAL COATINGS 144

additional half 、wve which leads to destructive interference (cancellation of 缸npl刷de

intensities) TI1e two transmitted beams of the mc1dent radiation have also a half-wave
ph晶e shift relative to each other due to the qu前ter－、wve fihn thickness. However‘ only
one heam (be副n A in Fig. 8.3）阻dergoes a ph晶e shift caused b~ reflection at surface 2.
1 he result 1s that when these two transmitted beams meet. they will be in ph甜e and their
間中li扭曲 intensiti闊前II be reinforced.

Cancellat10n
A B
\
／且eflecte
(out of phase)
N﹔日
1 』‘

N, > 1吭 λ
’一
可 4N,
甘世n Film Layer

2
司，

代J~
FIG. 8.3 Smgle﹒layer 祖tire!lection coating. 3 For clarity‘ the rays have been shown 聞自r 扭曲gle
rather than pe叩個dicular to 出e th旭 film 胡d substrate.

An example demonstrates the use ofEqs (8.6) and (8.7). A silicon subs個峙的 to be
AR－叩at•吋‘ optit叫zed for λ ＝ 1.7 仰n At that wavelength the index of re仕action for
silicon is N， 話 3 47。

According to Eq. (8.6). the index for the film layer should be equal to the square
root of the substrate· s index N月 if the substrate is in air. for which the index No is I. The
square root of 3.47 is 1.86; thi自 is the ind侃 of refraction for s1hcon monoxide at I. 7 ~tm.
The e{Teet 1s shown m Fi皂， 8 4.
145 01引C<L DESIGN FUND且JENJ:41.S HJR INFJURED E呵呵MS

τ
4

。5

， λ
1.5 2 2.5 3
1.7

FIG. 8.4 l ] ncoated aod AR-coated sihcon surface

.
Without the coating the transmission for this single surface is 69%. Figure 8.4
indicates that the theoretical 100% transm祖sion is achieved only at the specified
wavelength However司 the c叮ve does not abruptly plunge down to lower values.
lherefore句 a certain useful transition band exists. Hudson4 s姐tes that a I 0% deviation in
optimum index of refraction for the coat阿拉ne tlrat 血扭扭 the square-root relationship of
Eq (8 6)] r臼ul胎 m a decrease of only I% 血也e efficiency of the coatmg. TI世s means
that if NF varies from 1.87 to 2.的‘ for our example. the plate (t、珊珊faces）叫II still have
a total transmittance of at least 98% for 也e design wavelength of 1 7 µm.

Of cour間， most applications 缸·e noi limited to one specific wavelength. but cover a
spectral window that can be rather broad For these cases. a single-layer AR coa血ig is not
sufficient Multilayer coatll唱s have to be used Thelf performance depends on 出e same
basic interference prmc1ple.

8.5 Reflective Coatings

II 、開s only in the 1930s that the chemical process of"'s1lvering" a surface to create a
high-quality optical reflector was replaced by the method of evaporating thin fihns onto
the polished mirror substrate.5 Today‘ the most commonly evaporated metal films for
reflectors are

•Aluminum
•Silver
•Gold
•Copper
•Rhodium.

1 heir rellectance vanalions 、明th" aveiengths are summarized in Fig. 8S

 CHAPTER 8. OPTICAL COATINGS 146

Reflectance p

” .. . .

自
1.0
.. .. ” ... -’ m a

.. . . . . . - - -’ e F

關
0.8

•Alu血血間n
0.6
. Copper
• Silver
0.4
•Rhodium
•Gold
。2

0.4 0.6 0.8 1.0 2.0 4.0 8.0 16.0

Wavelengthλinµm

FIG. 8.5 Reflecl 師﹔e of evaporated metal films

To protect 也e relatively soft reflective surfaces from the envuonment. overcoatings

are applied. Qutte frequently. silicon monoxide or ma且nesium tluo討de is used for that
p叮pose.

Dielectric multilayer coatings 缸e being applied to meet special protective and/or

high-retlective requiremen臼 A typical ex割nple is the correction for the dip at about 0.85
間n iu the aluminum retlectance curve.

8.6 Typical Interference Filters

The _principle of interference fil扭rs was discussed for the special a~plication for
antiretlection coatings. IR systems often requue transmission control w他扭 specific
spectral bands. This can be achieved with multilayer interference filters. These filters vaη
m their complexity according to the perform扭扭e demand In some cases. a few thin film
layers may be sufficient: m others. 50 or even more may be needed w1也 additional
prov1s10ns to block radiation to very low levels from bemg tr宙間血泊ed outside the
spectral regions of inte間st.

Some of the most commonly used fil但r types are sho"n in Figs 8 6-8.9.
l 47 OPTICAL DESIGN FUNDAMENT.4U FOR [NFJIARED SYSTEMS

1.0

FIG. 8.6 Short-pass fi1個‘（ cuto旺的ter). Tue braoch 'to be blocked” requires 血 add1tJonal filter
(blocking filt叫自at absorbs or reflects 也e 阻diation o山erwise trausmitted 阻也at region.

ττ

l.O i 沁叫...以… 計叫一﹔ι－－－：－／＿：；~：，；，；－＞：－，川 1.0

λ I叫 、一一一， λ

FIG. 8.7 Long-pass filter (cut-on filter). FIG. 8.8 Droad b阻dpass filter.
 CHAPT莖R 8. OPTICAL COAT間GS 148

't

nυ

FIG.8.9 N帥。w b祖dpass filter (spike filter).

There 訂e m祖y factors mvolved m producing th間e sophishcated filters. and one can
sense that neither the design nor the fabrication of optical filters is a trivial matter.6

For rugged field spectrometers. where mo間 than just a few selected wavelengths are
of interest. linear variable filters (LVF). or circular variable filte閱（CVF) are employed.
Both 訂e based on the behavior of a wedged-film confi耳聞划。n. In the CVF arrangement
shown in Fig. 8.10‘ the wavelength 個nsmitted by the filter varies 品 a function of Ute
rotation of the filter disk. 7

1.0

FIG. 8.10 Tr阻smission versus wavelength for CVF.7

8.6.1 Angular sensitivity of filters

The performance of a filter changes 、社th the radiation’s angle of incidence. For a
filter on a plane-parallel substrate tilted by an angle i (see Fig. 8.11 )‘ the shilt to a
different center wavelength is approximated by
149 OPTIC Al DESIGN FUND.鉤，EN1:4LS FOR INFRARED S}'STEMS

λ ＝主且5亡高了？
N •. (8.8)

where A; is the center wavelength at i ~ 沁 is tl1e cen阻r wavelength at normal inci de nee
and N, is the effective index. derived from indices of thin film layers.

援學
f
t修
g一
FIG. 8.11 Tilted filter.

From Eq. (8 8) tt can be seen that al normal mc1dence when i ~ 0‘ the cen阻r
wavelength A.; reaches its maxunmn value At any other 叫le there ts a shift to shorter
wavelengths This feature is sometimes used for spec信ally fine-tu血ng a system.

The angle of incidence at a curved lens surface varies over its entire dian闊前‘
resulting m a continuous sl世宜。f the center wavelength While the 胡gular change is most
likely never very steep‘ the fact that a change occ叮S should not be ignored. This is of
course especially important if the lens is not merely coated with an AR film but serves 值
的時 substrate for the bandpass filter.

Frequently. a narrow-bandpass filter is inserted in a convergmg beam near the

detector. There are several reasons for that. The physical size of the filter can be kept
small. and stray radiation bypassing the filter is greatly redu目d Furthermore, the close
proximity to the detector allows for the filter to be included in the cooled Dewar
assembly. The variation of angles across the focusing cone h描 a detunmg effect on the
filter because of the varying shifts to shorter wavelengths for marginal rays. This is
usually considered in the design of the filter. leadmg to 值1 integrating compromise.

8.6.2 Thermal sensitivity of filters

The index of refract10n and the thickness of thin film ma但rials ch凶惡e with
temp前ature. Even though the cliange of the index N with temperature T (dN!dηis
opposite in sign of the ch個ge in thickness I, (dtldη 。 the perform閻明 of the filter is still
affected Because the first mfluence is greater than the second. there is a shift toward the
longer wavelengths as the tempera扭扭 increases.' This is expressed by

λ r ＝ λ 。＋ t:;.Tα ‘ (8.9)

whe甜 λI is il1e center wavelength at temperature t 沁 is the center wave！叩皂白 at

temperature To. :; T ~ F 九 and 且 is the temperature c的ftic1ent （~勵。 C) (Typical value
for a= 5 to 12 E-5).
 CHAPTER 8 OPTICAL COATINGS 150

A very 田eful and extended sunnn血y of terms connnonly 田ed to descnbe the
spectral characteristics of infra間I filters can be found in Ref. 9。

References

I. J. R. Meyer-Arendl Int10d11ction to Cl即s1cal and Modem Optics. second edition

Prentice-Hall. Inc. (1984). page 274.
2. Ibid.. page 27 l.
3. R. K凶gslake. Optical System Design. Academic Press (1983). page 106.
4. R. D. Hudson‘ Jr.. l頭。red System Engineering. John Wiley & Sons ( 1969）司 page2!9.
5. J. Strong. Procedures in Applied Optics. Marcel Dekker. Inc. ( 1989). page 7.
6. Cerac『 Inc.
te‘;hniques for optical coating.” Photonics Design and Application Handbook (1995 )‘
page H” 81.
7. J. D. Ranc油田t Optical Thin Films Users’ H仰dbook. MacMill阻 Company (I 987 )‘
pages 110-114.
8. Ibid .. page 73.
9 Optical Coating Laboratory. Inc J‘h企ared filters· glossary equations. par血neters .＂‘
Photonics Design and Application Handbook ( 1995). page H-244.
CHAPTER 9

Image Evaluations
9.1 Introduction

With the knowledge of how aberrations and di倍action influence the perfonnance of
an optical system, we shall now discuss ways to me品ure and evaluate image q阻lity. For
the mfrare止 the blur spot size and the radml energy distribul!on m the image plane a間
two quantilles of great mterest They 血dieate 也e mmnn田n detector element s血流＇ required
to collect a certain amow1t of energy From this information, one can derive the MTF. a
n甜甜ure of contrast versus resolution.

As With previo田ly discussed sub3ects句 tins chapter encompa田es a very sp間划 and
large field. However‘ tl1e presented approximations 咐II help to quickly obtain some
valuable benchmarks 品or a given sil!Jation.

9.2 Blur Spot Measurements

It was pointed out m Chapter 3 that due to diffraction and uncorrected aberrat1011s‘
the image of an o句eel point is never a point but a d晶晶ed disk‘間岫lly referred to as the
blur circle or blur spot. To confirm the predicted size of a blur spot‘ several me由urement
methods are used.

9.2.1 Circular mask

One way of measuring the size of 血1 image blur is to place the lens wider test m the
exiting beam of a collimator 血1d insert at the image location of the lens a circular mask
Just large enough to encompass the p叮·centage of the energy the system design asks for
The 100% level ts established by recording the signal without any mask. Of course. the
detector h阻 to be l缸ge enough to accept all the energy transmitted through 也e lens wider
扭st. Next司 a mask 1s inserted. properly sized to the specificatio肘。f the desired blur. spot
size. The signal will indicate if the energy p品sing tlrrough the mask mee扭曲e expected
percentage level The principle of such a se個pis shown in Fig 9.1. The use of a field lens
provides the necessary clearance for the pmhole aperttire insertion.

15 l
的2 OP叮臼L DESIGN FUNDAMENTALS FOR INFRARED Srsr日，可

Lens under Test Field Le阻 Detector

一﹒』

Radiation ‘
由
音。m
Collimator
.. ..’
Pinhole Aperture
.,
FIG. 9.1 Setup for blur spot me晶叮em冊t m也 circulor masks.

9.2.2 Slit

Substituting a very narrow slit in place of the pinhole 胡d moving it across the image
blur is another 、wy of measurmg the total blur spot s阻e. To obtam 扭扭curate result. the
吶dtb of the slit should be no 叫der than about 1/20 of the blur spot size to be measured.
The scanning across the blur spot is done in two directions‘ orthogonal to each other‘ if
the shape of the bl叮 is not circul血

9.2.3 Knife edge

A third way of me晶uring the blur spot size is 抽 trave阻e a sha叩 edge (kt世fe edge)
across the blur. The to叫 spot si扭扭d the optim凹n image pos1t10n for the m耐mmn blur
spot s肥e are f01md when 也e travel of the knife edge requ的d to raise the si伊叫齒。m zero
to I 00% is a minlln間n. Tl世s is 油dicated 姐 Fig. 9.2.

The question arisin哩 from Fig. 9.2 is whether one can simply subtract 10% from
both ends of the scan to fmd the 80% blur spot 刮目. Figure 9.3 shows that this is not the
case Based on the reasonable 晶sumption that the bl 叮 spot envelope resembles a
Gauss1叩 C盯ve. a correction factor c胡 be apphed.
 CHAPT芷民 9. IM,GE EVALUATION i 53

Blur Spot Energy Contents (%)

‘
4

100
90

咀＝－
司，
'",) ‘ B.? .
Knife Edge
Travel

Knife Edge ‘ B祖 .
FIG. 9.2 Knife edge 扭曲 a口。扭曲e blur spot.

苟， rE ‘

‘e •
叫 00 ＇匙，
"E "'f
1>r ＂的＝ a ,,,,‘
， Ea叫
ad
hp帥
可
dw

.......
FIG. 9.3 Comparison between measurements with circular masks and 旭ife edge scan.
154 0PI7CAL DESIGN FUND.品fEN犯扎刊 οH INF，制RED. SYSTEMS

The relation between the distance a and the radms r for the same amo凹It of energy
E contained in the exposed po討ion of the blur spot is approxinmted by

r =I .25a E-0·477, (9.1)

For ex血nple, for E = 80%。 r = I 25a 0 80 --0.m = I 39a In other words. if the
dimension a obtained with the knife meas叮·ement was 0.2 mm‘ the radius of the circular
blur spot would be 0.2 x 1.39 = 0.28 mm The mam pomt lo remember is that one has lo
be careful 叫lh the interpretation of data obtained by one measuring method when applied
to anolher.

9.3 Energy Distribution

ro properly 扭曲 ss the energy dis甘ibution of a blur spot. me阻urements 叮e reqmred

that indicate the change of the encircled energy as a function 。f the bl阻 spot radius. The
se仙p shown in Fig. 9.1 is suitable for such a measurement. A nmnber of well-centered
crrcular masks of increasing size are inserted su郎自由vely at the unage locatton and the
energy levels are recorded from 0 to I 00%. The resultant plot looks typically like the one
shown m F1g. 9 .4

Ene砲y Contents

LO
‘
A

0.8

0.6

0.4

。2

几。 Pinhole Radi叫

FIG. 9.4 Radial energy distribution

9.4 Modulation Transfer Function

Removing lhe subjechvtty from evaluatmg lenses by jusl visually JUdgmg their
perform胡ce 、NllS achieved 叫th lhe introduct10n of the modulalion transfer function. This
occurred about a half century ago. 1

While the subject of MTF is very complex. there are a n山nber of approx1mat1011s
that are especially helpful in the layout stages to assess the performance expectations of a
 CI凶唱的恥IAGEEVALUATION 155

lens or le恥 system. Before we discuss these, let l組 look at some b臨時 p泊1ciples related
to the MTF.

9.4.1 Overview

The image q_uality of ao object is a measure of the quality of a lens. With the MTF
the image quality is measured as contrast agamst spatial 仕equenC｝九

Contrast or modulation is expressed by

心Im
IMLW
--+
M
(9.2)

For the object‘ lo m.x and lo""" B扭曲e maxnnum 胡d m扭扭3間1 intensihes of the radiaot
emittaoce aod for the o吋ect. 11 =< and 11 刪 relate to the irradiance of the image. This is
identified for a typical bar pattem in Fig. 9.5.

The modulation transfer function 抖叮"F(v)] is the ratio of the modulation of the
image [M, (v ）﹞ aod 也e modul訕。n of the Ob甘eel [M,,]. v is the spatial frequency in line
pairs (one dark and one light line) per umt length，間閏lly milhmet缸（ Other dimensions.
such 且 cycles per radian司 are used as well ). The nmnber of hne pairs is 姐ere晶祕 until
the contrast m 血e imag_e is too low to be detected and the imaged pattem can no longer be
resolved At 血at 』 'limitmg resolution們 point. the spatial frequency is called the cutoff
frequency 的 For an aberration-free sys阻m呵

(9.3)
Vo ＝可：／＃）

I ig. 9.6 shows how the modulat1011 reduces as the spatial frequency increases.
 156 0P17CAL DESIGN FUNDAMENIAL., FOR INF臼RED SYSIEMS

Object

。句 ect Radiant Emi岫間

~ (Wlcm")
• 、，
4叫 I, """
Image Irradi阻閱（Wlcm')
’ ’•

Image

FIG. 9.5 Imagery of a b缸 p甜em with lom ~ I 血d 10,,;. ~ 0 (Mo~ I)。

MTF
Spatial Object Radiant Emittance •
Frequency

v,
Image Irrad1ance

v, nru毛凡
v，肘凹別注 ~v

、4、、d立、9、
Spatial Frequency (Ip／阻血）

FIG. 9.6 訂閱 modulation transfer function.

MTF analyzers are being offered m a variety of configuratio肥， Whether sl山‘ knife edges‘
sinusoidaL or bar pattern t訂gets are being used. the b由ic principle is always the same:

Tile modu/atio11 oftlte image is compared 圳的 tlte moli11/atio11 of

的e object at 叫rious spatial frequenci四
 C!l'1'TER 9. IMAGE EVALUATION 157

9.4.2 Contrast and resolving power

As mentioned earlier. con甘田t is 也e relationship between two shades of gray which.

at the !in訟， becomes black and white. Resolution is the I扭曲 of disc前nibleness of two
fine struct間·es. such 甜也e 卸祖ged lines of the MfF analyzer's bar pattern 個rget. With
these two definitions. it c阻 be unde自tood 由at nei也er co甜甜 nor resolution alone
describe 也e full perform朋ce of an optical system. 11重s is indicated in Fig. 9. 7 where the
modulation transfer functions of two lenses are compared Lens A h由 higher modulation
(con缸訕。 at lower frequencies than Lens B. On the other h阻止 LensB h晶 a higher cutoff
frequ阻cy (resolution lin誼。你an Lens A. 1『he ~uestion is not which lens is better.‘ it is
which lens is more suitable for the intended application The c凹ves provide the 扭扭wer.

Modulation M
(Con回到）

‘
4

。一 Frequency v VA VB .,
偎晶。lution)

FIG. 9.7 Con紅·ast 血d resolving pow盯. two general cases.

Erich Heynacher and Fritz Kober, both with Carl Zeiss, Inc. have published an
interesting paper on this subject of contr祖t and resolution.2 With their kind permission‘
we show reproductions in Fig. 9.8 of four photographs of the same subject (Emperor
Henry II in a stamed glass 叫ndow at the Strassburg cathe也"lll) taken with fo凹 different
lenses. While there is of course a degradation m the quality of the reproductions presented
here. the relation between con甘ast and resolution is well enough p何時間ed to demons甘甜e
the principle indicated in Fig. 9.7. As the authors of that paper suggest, a very educational
experiment is to observe the pictures from a larger viewing distance. As the distance is
increased, the resolution limit of the eye becomes the limiting factor and confirms again
that eval岫ting or judging the image quality of an optical system is not just a matter of the
resolvmg power of the lens. All of this can be applied to JR imagmg systems where the
eye is replaced by a detector
158 0.月1CALD晶晶'NFUND.晶，即TALS FOR INFRARED SYSTEMS

I •Poor co甜ast 2 •Bet脂r con虹ast

• Reasonable r鳴。lution •Worse 扭曲lution

3 • Bett前 con缸ast thar仆， but worse 出血 2 4 • Best contr那t

Best 問solution • Same resolution 晶 l

FIG. 9.8 Co曲回t 阻d resolving power

 CHAPTER 9. lMAGEEVALUATION 159

Tl甜甜甜ciated MTF curves sh肌油田 Fig 9 9 confinn clearly the characterisllcs

of the fo叮 lenses 阻edto take 且1e pie個res shown in Fig. 9.8.

M (modulation or contrast)

l.O

。8 …一～一－·一→…一→一

。4

.2

v
2 3 4 5
Spatial frequency （缸hitrary units)
FIG. 9.9 Contrast 祖d resolving power司 the transfer luoctions.

TI祖t resolution is only part of the mfonnatwn for judging image quality IS evident
from the shapes of curves I 祖d 4. Both show the s祖師 cutoff t全·equency‘ 1.e
limitmg resoh且hon令 but VaJγstrongly with reg叮d to contr揖t This confir寶ns the remark s
made in Fig 9.8. Another observation substanhal自由e vist抽l renditions of s31llples I and
2. The limiting resolution of s血nple 2 is about half the one of S31llple I; however. i但
contrast is much higher at lower frequencies. All tins indicates that much can be predicted
and confinned with modulation transfer functions

A typical MTF mial)Zer 的 shown in Fig. 9.10. Such a system can be equipped with a
number of convenient a的acltments for evah叫mg a lens al d1虹ferent conjugates and off-
缸is positions. P。lychromatic tests can be conducted 個d defocusing effects C由1 be
analvzed
160 0月7臼L D目的wF凹D訓ENTALS FOR [NFIWIElJ SYSTEMS

FIG 9.10 Comm閻明i MTF 胡說y甜r

(R叩tinted wi曲 p宙間臨ion ofC制k田 Co中0酬。吋

9.4.3 Diffractio間，1TF

e缸ects, the MTF is stated by 川

For a di份通ction-lim1阻d optical system, which 1s one with negligible aberr胡on
叮圳
凡

This equations refers to an optical system with a circular aperture. As stated before,
10 is the cutoff 仕·equency of the system, expr·的sed by

Vo 三 )JF耳﹔ (9.5)

“
This simple relatton provides us with another convenient to remember” fact
Selecting 0.5 戶m as the wavelen皂白 for the visible spectrum, 4 µm for the MWIR, and I 0
µm for the L WIR leads to 也e following limits:

2000 250 . 100

v"υ ＂＇＂早－一一－一. V ＂包””m Z 一－－一－ .and V "’”-- = (9.6)
，，， (fl#). ＂叫阻 (fl#)' U LWOK
(fl#)

These relations refer to lme pairs p叮 millime阻r. It 1s good to remember that even a
di街action-limited ((II) system in the LWIR cannot resolve more than I 00 Ip/mm.
 CHAP亞泉 9. 泌1AGE EVALUATION 161

Within the boundaries of 0 < (1' I 吭，） < 0.6. a good linear approximation for the
diffraction 扎叮F is given by
I V \
MFF,.，研討－ 12181 -1 (9.7)
一.. IV I
\ 0J

Also‘ using Eq. (6.5 )‘

MT凡晶至 1-1.218λ （ f!#)v (9.8)

9.4.4 Geometric I組TF

1be calculation of the geometric MTF assumes a .~erfect optical sys阻m and ju甸es
the performance of a l阻S 品 a function of defocusing. Diffraction effects 甜可＇ ignored. This
a阻umption leads to

2J, (1吼叫～ 2J，［耐ol(f l#}J

MFF,..,,. ( v) ＝一一－＝ L ~ (9.9)
S間 rrvB 附01 (f I#)
With J1 the first-order Bessel function, B the linear blur spot size‘ and " the spatial
fr缸iuency.δis the defocusing necessary to yield blur spot B. i.e刊

S = B(f!#). (9.10)

Expression (6 8) is not too reliable for small amounts of de必cusmg.3 A better result
is obtamed by usmg an approximation that mcludes a defocusing mult1plymg factor. J H
Haines and J. P. King,'' as well as ll R. Sh剖moff; have discussed the approximation
adopted here

MFF = MFF＂ι × DMF.

'PP ＇也＂
(9.11)

The defocusing multiplying factor

DMF 一叫7tvB[1 －（戶的川］｝

一－ (9.12)
7tVB[1-(fl#)1'.v]

9.4.5 Numerical example

Given: Lens with f = I 00 m肌 f/#=2司、,vavelengthλ ＝ 4µm.

Detennine Diffractmn blur spot size『 cutoff frequency司

modulation at 30 Ip/mm for 逝世action-limited system
and for sys阻m m銜。 32 mrad angular bl叮 spot size.
Plot MTI C前ves‘ mcluding appro、imat1ons

l. Diffraction blur B也噩＝ 2.44x0.004x2 = 0.0195mm.

162 OPTICAL D阻＇GNFUND創血TALS FOR INFRARED SYSTE.I 悠

2. Cutoff frequency Vo ＝一＿！＿＿＝ 125lp/mm

0.004x 2

v 3。
3. Nonnalized frequency 一－＝一－＝ 0.24.
V0 125
2 I 「一一一~、
4. Modulation MI'F＂且叩之一 l arc cos0.24- 0.24" I - 0.24' I= 0.70.
自叫一 昆、 '

5. Appro文imation MI'F剖活啊n 30 =1- 1.218 X 0.24 = 0.71

6 The 祖gularbl叮 of ﹛﹜ 32 mrad becomes a linear bl間。t 0.00032× 100 = 0.032 mm
in the focal plane and results in
巾仙

F8

7. Improved approxunation. applymg coπectmg method

2J,l30x 0.032it(l 而 2x 0.004x 30)1

MTF_ = 0.70 ×＂』＝ 0.33
30x 0.032it(l 2 x 0.0個 x 30)
8. MTF plots

M
v=30 Ip/mm
1.0

0.5

。 v
(Ip/mm)

75
。 25 50 100 125

Remark: The geometric MTF curve s_hows negative values. This indicates a 180° ph阻e
shift in the image. The effect is a d缸k image where it should be light 血d vi臼 versa. This
is kno'.Vn as spurious resolutwn
 CHAPTER 9. IMAGE EVALUATION J63

References

I. C. S. Williams and 0. A. Becklund. lnh·od11ctio11 to the Optical Transfer Function,

John Wiley & Sons (1989). page 9.
2. Erich Heinacher 祖d Fritz Koh吼叫Resolving power 咽d contrast。”加ss Information
No5Ll2"'ye缸， pages 29 to 32.
3. W. J. Smith. Modern Optical Engineeril:嗯， McGraw-Hi缸， Inc. ( 2000 ), page 379.
4. J. 泣. Haines and J. P. King,”Approximate Computation of Physical MfF." SPIE
proceedings, vol. 13 (1968), pages 105 to 113.
S. R. R. Shannon, "A useful optical engineering approximatio凰”。'Plies and Photonics
News, (April 1994), page 34.
CHAPTER IO

Diamond Turning
IO. I Introduction

Many infrared components 甜e m曲叫ac阻red by diamond twning. For that reason

some 間m血ks 咐II be included in 出is text. Th間e 阻血缸ks 缸e intended 甜 guidelines for
the systems design ph阻e. Understanding the process provides the designer with the
mea由 to 垣vestigate whether lens elem凹的 C個 be ehmmated or system performance can
be improved. At the same t帥， m awareness of the process limitation c阻 save one from
disappo切tments and undesired surp討ses.

Diamond turning is 個1 ultrap間cision machi血ng process especially suitable for

generatmg aspheric and diffract1ve surfaces on most in晶宮ed lens materials such as
gem1間1阻n. zinc sulfide. zmc selemde、 silicon. gall仙n arsenide. calcium fluonde. arsenic
trisulfide、 Amtrr司 and some chalcogenide gl晶ses. Diamond tummg 1s also the prefe甘ed
method for producmg high-precision reflective surfaces on metals Heat-treated al山nm凹n
allov 6061 1s a favored substrate matenal for such mirrors Smce 叫阻ninum 1s relatively
soft. the preshaped subs甘ate is nickel plated p討or to the final cut. The harder nickel
surface also p前mits postpolish泊g 1f the required surface finish c由mot be achieved by
diamond tun世ng.'

10.2 Overview

Flat and spherical swfaces as well as cylinders have been 皂enerated by singk~point
di剖nond tummg more than a half century ago. The technology was used primarily to
produce precision mech目ucal components. To achieve the accu間cy and preci剖on
街mantled for optical surfaces. better control techniques and h1gh-prec1sion be盯mgs w叫e
required When laser distance-measuring 111terferometers and air bearing spindles were
developed in !he early 1970s‘ diamond turning became an acceptable pr凹的s for
generating a wide variety of rotationally symmetnc surfaces such as conic sections and
other aspheres.'

Over the past decades. progress continued and besides the advancements made m
better controls and improved general machine fea恥res, diamond-turning machmes have
加come more affordable While there were only a few org缸阻ations and mst1tutions 111 the
early days that had such machines, today可 many smaller optical companies consider these
machines standard equipment Figure 10. l shows a moder百 tw← axis contouring diamond-
turning machine.

165
J66 OPTICAL DESIGN FrlNDAMEfffALS FOR INF，制REDSJ"STEMS

FIG. 10.1 Modem diam祖d 個ming machine.

（切關e可 of Prec1tech Joe )

10.3 Surface Finish

To minimize the machining time 胡卓 achieve a reasonable smooth surface. the

radi間。f the diamond-cutting tool hp ts chosen to be relatively large. The relation among
valley fini凶1 (P-V) ‘ the tool radius Rr. and the tool feed per revolution }; is
peak…

：－＇ι (l 1)

仇U
(P-V)
8Rr

TI1egeom甜y of this relauon is shown m Fig I0.2

Cuttting
Tool

P-V
。RT ..
可V

"'
.. !. •

FIG. 10.2 Cuttmg tool radius 胡d surface finish.

 CI 山吋ER 10. DIAMONDTlJRNINn 167

Considering the impact on the surface roughness from such factors 晶 tlte lhmted
machine stiffness、 vibration. and otl時間， a good assumption for most c品es is to use Eq_. (10.1 }
as an estimate for the exp回ted nns司開血er than the P-V surface rou必由自 For mfrared
applicatio醋， a surface Ii凶sh of 5 nm ( 50 A) is sufficient for most c品es, and a tool radius of
0.8 mm is often chosen. This leads to a feed rate of 5 7 µm per 間volut10n w}1ich is a
reasonable number

It has been indicated in Chapter 5 that a large tool radius c祖 have a d阻turbing drawback
in generating a diffractive phase profile. A relative l缸ge tool radius creates a shadowing or
blocking effect at the zone tr宙間tions. This is illus位當ed in Fig. !0.3.

r
’,
.,

VJωaa

Blockage

FIG 圖 10.3 Blockage at the wne transition reduc申廿四smission 加OU且h 也e lens.

The transmission losses due to tlus blockage can be estin祖ted wtlh

L 至 1彷·~fo:. ( 10.2)

where D ts the lens diameter and d is the maxim間n depth of the profile R,. is the tool
radius as indicated in Fig 10.3. The to祖I number of transition steps is 11,'" "' Tlus number
is smaller by I than the number of total zones

We apply these observations lo our example in chapter 5. page 111. The diameter of
the lens 、vas
D ~JOO mm. d~ 0.0日333 n凹， and 阱，叫＝ 3 If we use a tool radius of R，. 泣
。 8 mm. a transmission loss of L = 0. 7 % will result In this case‘ it is a relatively sτnail
most likely negligtble 血noun!.

For applications in the MWI孔 where silicon is a favored lens material. the 別扭曲。n
is not that hannless because the dispersion of silicon in the MWIR region is about four
times stronger than that of germanium in the LWIR spectr1lln. The result is thai 30 steps
168 Om臼L DESIGN FUND.品1ENTALS FOR INFRARED SJ叮酬S

across a diameter of 50 mm 前e quite poss拖le Therefore. with a maximu血 step of

d=I 的 μ且 such a 50-mm di血neter lens would have a transmission loss of

L ： 剝 2·0.0月﹔叫

An 8.1 % loss may be acceptable. If not. one c祖 reduce the tool radius and m也 1t. 也e
feed rate for the tool to ob扭扭曲e same s叮face E凶sh. Of course. since the machining
tune 咐II increase proportionally司 this wdl impact the cost of the lens

As c阻 be seen. the loss changes with the squ前e root of the tool radius. Thal is. if
we reduce the tool radius to 0 4 mm the transmission loss would decrease to
s.1,f(OA疋函＝主7%.

Another approach is to modify 自e contour of the tool nose. A possibility is to 間e a

semicircular shape. This is depicted in Fig. 10.4.

r
' ’

’,
da

FIG. 10.4 A semicirc叫缸 tool nose elurunates shadowing cfiect at the zone tr:msition.

Of course. there 1s a practical lmut to how pointed the tip can be ground and lapped
but a radi間 of few micrometers can be achieved. This semicircular shape has the
advantage of at least drastically reducing-if not completely eliminating these transition
transmission losses and still main姐說ing a favorably large radius for cost-effectively
obtaining an acceptable surface fm1sh

I 0.4 Scattering

The transmission losses due to back scattering caused by surface roughness are usually
negligible. To 扭扭扭 the magnitude of this loss. the relation for the total integrated scatter
(TIS) developed by the Naval Weapons Center' can be applied :.

TIS z﹝守J (10.3)

 C!L\PTER l 0. DlAMOND 1i爪NING 169

o
”
where is RMS roughness and 入。站時1e wavelength of interest. For a roughness of 100 A
and a wavelength of 4 n. TIS is approximately only 0.1 %.

10.5 Shape Correction

Due to the centrifugal forces and other influences present during machining‘ it is
possible that the surface shape ob祖祖ed deviates from the one desrred Today.『 this can be
easily corrected by recording 曲目e deviations 祖d feeding them back to the control
system for the next cut.

10.6 Optical Surface Testing

10.6.1 Surface roughness

One method of meas田mg the roughness of a diamond-turned surface is a mechamcal

stylus profilometer. For higher resolution. optical profilometry is being used. This
concept is based on a phase-measnnng Mtreau micro－咀terferometer.' Figures 10.5‘ 10.6有
and 10. 7 show the roughness profiles of diamond-turned surfaces measured with a
comrnerctally available opllcal profilometer. In addition to the substrate material. the
arithmetic average roughness 恨。）明個d the rrns roughness of ti時間faces 耐 called out
Figures I 0.5 祖d 10.6 refer to crys旭!line 取 materials ( CaFI 個d Ge). Fig叮e 10.7 shows
the roughness of an electroless mckel-plated al凹的n阻n mirror surface.

+10. 4 ＇年

nm
_o
-·
昏，自7
".'::>
'

r
0.00 mm 巴 .S?

Ra 2.1 nm Ca Fl Rms 2.9 nm

FIG. 10 5 Substrate: calci山n n叫>ride.

l 70 Ol'TJCAL DESJGN FUNDAMENTALS FOR INFRARED SYSTEMS

+6.95

nm

-15.25
0.87

國 .ea ... 包 .97

Ra 2.2 nm Ge Rms 1.6 nm

FIG. 10.6 Substrate· gennanium.

句 5. 33

的m

呵5. 7 日
tl.87

以 00 呵＂ 0. 9?

Ra 1.4 nm Aluminum Ni Rms 1.8 nm

FIG. 10.7 Electroless nickel-plated aluminum.

(Ali surface roughness diagrams were supplied by Precite:h Inc. }

 C!C\Pη·R 10. DIAM印制 T凹心也NG 171

10.6.2 Surface shape

Just as for measuring surface roughness. mechanical and optical methods can be
applied. With the mechanical techmque. a stylus is used to trace across several meridional
sections of the surface to be tested. These repeated me甜nrements 叮e done to confinn
rotational symmebγof the p缸t Tlus proced叮e is relatively simple and cost-effecllve.
However. using a mechanical s1?'lus requires physical contact‘ which is acceptable 自or
most cases To avoid the possibility of scratchin星 the optical surface of a so自 material. the
surface shape can be measured optically.

To cover the entire surface ‘ interferometnc meast叮·ement methods are bemg used
For volume production‘ a smtable computer” generated hologram (CGH) is pr<呵，ared and
inserted in the optical train to compensate for the wavefront distortion mtroduced by the
optical surface under test Sim1I訂ly、 so-called ‘回ull-lenses” are also used for the same
purpose.

If possible. it is strongly recommended to install the 描phe討c or hybrid element into

the complete opllcal train and m四sure the image q帥lity oft11e to個l sys扭m ThJS way. 也e
目前 of the optical system serves 晶 a null lens and the test is most realistic. since it
con伽ns not only the surface shape of an element, it verifies the perfonnance of the entire
system without having to mvest in addit10nal a阻diary components and eqmpment

10.7 Machining Time

The total nlllilber of revolullons required to maclune a lens surface with a dian1eter
Dis DI（以）. If P is the number of rev。lutions per minute （叩m）。f the machine spindle ‘
the total macbir血1g time for one cut is

D
T ＝ 一一－；：：：：：＝＝＝＝＝ (10.4)
2P,,)8R,h

A conservallve number of revolutions for the spmdle is P = 1000 中m With tlus

assumptton and a tool radius of Rr = 0.6 mm. the time to for a single cut to achieve a
surface fimsh of h = 50 A (5 run) is

T 三 O.lD. (I 0.5)

叫ien the diameter D is spec1日ed in millimeters. Th站時 only a rough gmdeline for the
time it takes for one cut across the surface. It vanes 、社th such factors as material and
nmchme rigidity‘ among others

From Eq. (IO..!) we see that the machining time is inversely proportional to U1e
square root of the tool radius In oth前 words、 reducing the tool radius by a factor of four
as we did to lessen the shadowing or blockage effect by two『 would double the cutting
time.
J72 0P11CAL DESIGN FUNDAME師丸LSFOR[NF，凡胡EDSYST品的

It must be noted that while the machining t扭扭由 indicated here is relatively sho民
sever這i cuts are required for the necessary material removal to meet 自e general shape and
thickness specifications of the lens. However. r目ent advances made m machine design.
particularly with the addition of lmear motor 街ives‘ high-speed spmdles and hydrostatic
slideways have led to much-reduced cycle times. Many medium-sized IR components are
now being machined at spindle speeds 區 excess of 6000 中血 With this increased speed.
the slide fee心rate per 間volullon is adj凶ted to still produce an acceptable material
volu血e removal to avoid !Illcrofrac恥re.5

The drastic cycle time reduction is the 扭曲on why manufacturers of infrared lenses
叮e now gen前atmg even spherical surfaces 叫th this process.

10.8 Further Progress and Developments

Progress is continuously made with reg缸d to generatmg complex surfaces and bemg
able to diamond turn even gl晶s. which has been a goal for a long llme. Several
companies are offering now so-called 台eeform generators which are fo叮﹒個d five-axis
machines These machines are smtable for producing non-rotallonally symmetric optic叫
surfaces. The p叩開自 is based on shaping the surface 咐th an array of individually
positioned subsurfaces. p血ellike elements in scallop form. Another method of producing
出問e general aspheres uses piezoeletric ac個ators 祖d l詛ear motors for 也e add1t1onal
superimposed tool motion required. 6 This freedom presents an interesting op port間1ity for
many applications‘ such as compact 1mrror 扭扭llgmats and head-up displays.

Another bre泡kthrough is the ultrasonic-assisted d1個nond turnmg of glass and

ceramic. In this tee迦1ology. tl1e tool oscillates ul缸訕。nically during the turning process in
the cutting direction Using this method‘ surface q開lity of an arithmet目前叮咚e
roughness (R,) of< I 0 nm ( IOOA) has been achieved. 7·'

References

l. R. A. Clark · Design specification of dian1011d turned optics. ‘ SP!E Vol. CR 38

(1991)pp. 16•ι183.
2. R. R. Shannon 祖d J. C. Wyant. Applied Opti呵。nd Optical Engil甜甜削·g. Vol. X肉
Academic Press. Inc. ( 1987). p.258.
3. J. M. Bennett and L. Mattsson ..Introduction to Surface Rouglmess and Scattering‘也
Optical Society of America. Washington D.C. ( 1989).
~ R R Shannon and J. C. Wyant. Applied Opti呵。nd Optical Engineering. Vol. X.
Academic Press. Inc. (1987). pp. 353-355.
5 P. N Blacke and R 0 S個ttergood. •‘Ducllle regime machining of gennanium and
silicon.” Journal American Ce1 a1111c Sac1e.句： 73 ﹝4] pp.949-957 ( 1990).
6, J. Hennig "'S叮face Struc個ring with Fast Tool Servo Systems‘ Fraunhofer lnstilul
Produktionstechnologie. Aachen. Data sheet (2000).
7. 0. Riibenach司已 UI仕asoruc-ass1sted diamond l!uning” Fraunhofer lnstitut
Produktionsteclmologie. Aachen. Data sheet (2000).
8. F. Klocke and 0. Riibenach ” Ultraschall-unterstiitztes Dreben von Glas und S個hi
mit Diam阻ten:· lndustrie Diamanten Rund』chau 34‘( 2000). Nr. 2. pp.133 146.
APPENDIX
A. t Paraxial Ray Tracing

A.1.1 Surface equations

The symbols and their s1g11 conventions used in the following equations are
1dent1fied m Fig. A.\.

Opening:

For finite object distance L1

y, = h- L,111 (A. I)
\',
or 11, ＝一 τ (A.2)

For object at infinity‘ 111 = 0 and y1 = h.

Trans跆r N﹔← N.,u.， 一几·，c,(N﹔一 A「.，） with c. 立 (A.3 }

- n

Yn+I =.l ’.+ t;< (A.4)

Repeat Eqs. (A.3) and (A.4) fork snrfaces in the optical train.

Closing:

For oblique ray

ll= y, + l{1(. (A. 幻

間 UK
M

n3o on CAO a
、，－

斗，“鼠
一一

- (A.6)

Magni且cation:
此 ／（
- 1( - h .
Ill 一~一一一 (A.7)

For miπor surfaces. the index N =~ l is to be inserted into these equattons.

173
174 OP口CAL DESIGN FUNDAME•師AlS FOR 卸F缸RED SYSTEMS

(+JR, (-) R,
Ublique Ray

-31R•Y 可 (+)y,
’
AJ<l (+) u, (+) y, •
（寸 u', izvh
』h
• ‘.
N, N',

~ (-) L, ... (÷) I', - (+)L, ..

FIG. A.I Symbols and sign conv開hons for parmoal raytrace eq叫tions.

Numerical Example

Given: R1 話 50. (c, = 1150 = 0.02 )‘

R, = 50‘( c2 = 0.02).

and 1'1 = 10. N l = 1.5可 l1 =’ 200

Trace 叩血ial ray with y1 = 20. and 叩 obhque ray 叫th h = 15 andy1 = 10.

I. Axial Rav

u, ＝一γl IL ， 且一20／（一200) = 0.1

I[ ’l = 0.066667 ; y2 = 19.333333; u', = 0.293333.

l', = ·.nlu'1= 65.909091.

m = u1/u'2 = 0.1/( 0.293333) = 0.340909.

2. Oblique Ray

u, = (h y1)/L1 = (25 IO)/( 200) = -0.075.

u'1 = 0.116667 ; y, = 8.833333 ; u', = 0.263333.

h' = ｝三十 l'2u'2 = 8.522727.

Ill= h'. h = 8.522727/25 = 0.340909.

 APPENDlX I 75

A.1.2 Power equations (ray tracing through thin lens間）

Fig A.2 shows a train of thin lenses where the elements are separated from each
other by d'n·

Opemng for fimte object distance:

v, = h- L,u, (for oblique ray）、 (A.8)

v,
肝 t廿一王 (for ax.ial ray\. (A.9)

For object at infi甜可『 y1 = h. and 111 = 0

Transfer: 11;, = "• r0中n with power 中＝ 7 (A.JO)

Yu+• =y. +d 1(,. ; (A.11 i

Rep且t Eqs. (A.10) and (A. I I) fork elements in the optical train.

Closing: I ＝几十 ι1( (for oblique ray ）司 (A.12)

or 1 - 2.主 (for 制ial ray). (A.13)

14' 一－ 1(
Ma n cuca onMmM
αb

-- -h
u
4
UH
一－

Z
(A.14)

2 k
j已
•
y, u', = u,
‘•u,
’,
u,’
可F

•
呵，

~
L, .... d’ ! ’ .. L; ..

FIG. A 2 Ray trace through separated thin lens目

J76 OPTIC.<£ DESIGN FUNDAME間ALS FOR 卸＇FllARED sr.rr明IS

Numerical example

Given: Ji = 100, （中t = 1/100=0.01).

f, = 137.5, （中，＝一0.007273),

and d'1 = 50, L1 ＝汁，000.

Trace an ax揖i ray for this telephoto lens withy1 = 20.

的＝ -y1/ L, = -20/ （一 1,000）罰。但

z’,t =0.02 20 × 0.01 = -0.18

只＝ 20 + 50 x （一0.18 ）目 1 I.

u'2= 0.18 !Ix （一0.007273 ）賞一0.1.

L', = -}2 lu'2 ＝一 l ！／（一0.1)= 110.

m= 111111 ’,= 0 2 or 1/5

The same equations are 阻ed to trace oblique rays

To mclude m盯ors for these power equations. we remember that the focal length of
a spherical mirror is/m = R/2. Therefore‘ the power 申m = 2/R or 泣，（ c is the curvature
臨 d阻cussed before). With mtJTors m the optical train‘ one has to be careful wi也 the sign
convention: light 阻verses drrection a自er reflection.

A.2 Spherical Aberration of a fhin Lens

A.2.1 Derivation of expression

Tirroughout tlus tutorial text. much reference h酷 been made lo thud-order

aberrations of a thin lens. To provide general guidan間 about the derivation of the stated
exp甜甜自由 based on the third” order theory. Eq. (3.7) will be developed ”

We begin with the expresst0n for the transverse spherical aberration cont吋bution

TSC Z GC GmC + ct
ρt

C i (A.15)
 APPENDLX I 77

This equation is the trw1catcJ version for the c品e where the o ject is located at
infinity. For the general c晶e. three more terms have to be added inside 出e bracket.'
Conrady' presents the derivation of this equation and describes w detail the meaning of
the so called G恥sum. For o叮 P凹pose. it is sufficient to know that

N l
N一
、】向
（一

-
）一
G
一－

I 2 - (A.16)

G 一（ 2N +l)(N 一 I)
(A. l 7)
勻

G. 一（ N÷2)( N -1)
血1d (A.18)
• 一－
2N

N 1s the mdex of refraction of the lens material. Eq (3.3) tells us that the net
curva山re c of a thin lens with power 中間

c=-,-L (A. I 9)
\N 1)

further. 、時間call from Eq. (3.5) that the first surface curvature c, of a thin lens
shaped for minimwn spherical abeπattm】 is

c … N(2N+l）。 (A.20)
I 2(N+2)(N …1) ~

finally. w℃ can state that the slope 11' of the exitmg ray ts the ratio of the half
aperture r and the focal lengthf The間fore司

1( 一
E … f - 2 (JI#)
一一一一一一一… (A立 I)

Plugging all this into Eq. (A.15) yields

N(4N 一 J)j
TSC ＝ 一一一一一一一→－ ' (A.22)
64 ( N - D2(N + 2l(j/#)'
178 OPTJCALD間＇GNFUNDAME前，LS FOR INFRARED SrsTEMS

A.2.2 Blur spot size

The linear bl叮 spot size B is about half of T古C; hence‘

n N(4N-l)j
M 一 (A.23)
u,ph~＂~一 128( N - 1)2( N + 2l(/1#)3

Angul缸ly exp間ssed司也e bl叮 spot size m radians becomes

f3 _B_ N(4N-I)
(A.24)
啊帥一 f-128(N 一！）2(N+2){f結）＇

This is Eq. (3. 7).

The plot in Fig. A.3 shows lhe benefit of a higb-index material for the reduction of
sphencal aberration

戶咽的
A

。06 （仰r'
0.04

0.02

。 1.5 2 2.5 3 3.5 4 ”, N

Glass ZnSe Si Ge

FIG. A 3 Mmmmm 血gular blur spot size of tlli.n lens due to spherical aberration
as a function oflens material.

A.3 Optical and Thermal Data for Some Infrared Materials

The data presented are to be considered as guidelines. Data from vendors and
handbooks vary, in some c血es widely It 1s therefore strongly reconunended to check
with the material supplier for final design commitment H令已
 APPE>叩IX 179

A.3.1 Selected materials for the 耳－ 5-µm spectral band

Optical Re仕·acltve Abbe Index Coefficient of

matenal mdex number change I 。c exp目盟ion I °C
N」um v，＿，，~ dN/dt ( × 10-•) 。（× 10-')
Gennamum Ge 4.025 103 396 5.7
S1hcon Si 3.425 236 150 2.6
Zinc sulfide ZnS 2.253 110 43.3 6.6
Zinc selemde ZnSe 2A33 179 60 7. l
Amtir l 2.514 198 72 12
Magn. fluoride 弘1gF, 1.349 13.5 ’,
12
Calcium fluoride CaF、 l.410 21.7 一75 19
Barium fluoride Baf、 1.457 45. l 一 16 18
Ars. trisulfide As2S3 2.412 181 9.3 21
Sapphire Al203 1.675 7.65 3.4 6

A.3.2 Selected materials for 也e 8-12-µm spectral band

Optical Refractive Abbe Index Coefficient of

ma ten al mdex number ch祖ge I 。c expansion I °C
Nwum Vs-11um dN/dt (xlO 、 。（×1。一的）
Gennantum Ge 4.003 IOOO 396 5.7
Zinc sulfide ZnS 2.200 23 39 6.6
Zinc selenide ZnSe 2.406 58 61 7. 1
Amtir l 2.498 110 72 12
Gallium arsenide GaAs 3.278 110 185 7、
Cad. telluride CaTe 2.672 170 97 有

Sodium chloride NaCl l.495 19 一25 44

Cesium 10d1de Cs! l.740 230 一89 50

References

l. W. J. Smith司 Modem Optical Engineering‘ McGraw-Hill Inc. (2000). p.336.

2. A. E. Conrady‘ Applied Optics and Optical Design. Dover Publications. Inc. ( 1957).
p. 95.
3 S. Ballard. et al Optical Matei ials 戶p b，你1 ed Jnst1 umentation, University
of Michigan. Ann Arbor ( 1959).
4. Handbook qf Optics‘ edited by OSA. Chapter 39 "Thennal Compensation
Techniques＂，如lcGra、恥Hill Inc. ( 1995 ).
5 ZEMAX optical design software. Version 8. Focus Softwa間‘ Tucson. Arizona.
6. W. L Wolfe司 and G. J. Z1ssis‘ 1月he Infrared Handbook ‘ Office of Naval Research ‘
Department of the Navy‘ W品hmgton‘ DC. (3'" Printing司 1989)
INDEX

Abbe. 49 Diamond-tuminι165
Abe訂atio且 L 4. 16 Differential changes. 29
Aberration contributions‘ 69‘ 86 Diffraction efficiency九 l 14. 115
Aberration co叮ection methods, 53 Diffraction 1泊圳、 74
Achromat, 56, 109 Diffraction MTF司 160
Alf-spaced elements, 57 Diffractive optics‘ 107
Airy disk 75 Dispersion咱 49
Angular blur spot size. 44 Distortion‘ 37‘ 41
Angular sensitivi月九 148 Domes. 91
Ant1Teflect10n coatin品 143
Apertnre stop. L 24 Effective focal length. 62
Appendix, 173 Emis日vity角 4. 7
Asph•自es. (see General aspheres} Encircled energy, 154
Astigmatism. 3 7‘ 39司 48 Energy transfer. 28
Athermal achromat司 135 Entrance pupil‘ 1. 2
A them祖lization. 133 Entrance window.]. 2
Atmospheric transmission‘ 12. 13 Etendue. 27
Axial chromatic abeπation. 37. 41. 49 Euler. 45
Exit pupil‘ l 可 2
Dack focal length. 62 Exit window.‘ I. 2、
日ackground. 4
Dall lens. 93 f”number, 2
Ileam expanders. 59‘ 73 Field curvatu時， 37‘ 40
Ile品I fit sphere‘ 98 Field flattener. 72‘ 73
Binary optics. 107. 117 Field lens ‘ 34
13lackbody,5 Field of view. 1ι35
Illur spot size. 151. 178 Focal length. 2

Cassegrain configuration. 67 ‘ 68 Gauss‘ 45

Chromatic aberration‘ (see Axial General asphere, 103
chromatic aben·叫ion﹜ Geometric M叮、 161
Circle ofleasl confosion. 斗。‘ 47. 50 Gcnmmium‘ 43.46‘ 61
Circular masks. 151 Gradient index le也 95
Classical lw<>-mirror Gregory configuration. 71
configurations‘的
Cold shield. 138 Humidttv 11
Cold stop. 138 Hybrid achromat, I 09
Coma. 37、 3且 39
Come sec11ons‘ 97 Image fonnal!on。 21
Contr品L 157 明！ 59 Immersion lenses 司 30
Curvature. 43 ‘ 47 Infrared regions. 9司 49
Instantaneous field of view。 128
D大 3. 15. 16 Interference filters. 146
Design examples. 12 l Internal reflections ‘ J-13
Detector co叫igurations司 16 1πadiance. 155

181
182 OPTICAL DESIGN FUND.品種ENW.S FOR /NFRARED Srsnus

Kir℃hhoff.7 Scattering. 168

Knife edge. 152 Schwarzschild司 6且 72
Seidel. 42
Lateral color. 42 Signal-tc• noise ratio. 2司 4
Light pipes. 33 Silic河on.46司 61
Longitudinal spherical a悅目前io且 38 Slit‘ 152
Snell’s law. ”, 20
Machining time. 171 Spatial frequency. 155
Magnification. 23 Specific detectivity司 15
M祖gin. 65 Spectral radiant emittance.‘ 5
Marginal ray‘ I. 59 Spherical aberration. 37
卸位血mum blur. 50 Spherical m盯or“ 63
Modulation 甘ansfer 品mction‘ 156 Spike filters. 148
Staring arrays. 34
Net curvature‘ 42 Stops. 24
Noise equivalent b祖dwidth‘ 15 Surface finish. 166
No1Se eqmvaient power. 3
Telephoto‘ 62
Optical gain‘ 30‘ l 18 Thermal detectors句 14
Optical materials for the IR. 178 Thermal effects‘ 131
Optical micrometer. 85 Thermal sensitivity‘ 149
Optical path length. 78 ηm lens. 21. 43
Optical path length difference‘ 78 Three’,nirror co1ill串urations. 105
Optical surface testin忌 169 Throughput. 26
Transmission‘ 143
I’訂axial ray tracing‘ 173 Transverse spherical aberration. 44
Performance eq阻tion. 4
l'etzval 40. 41. 57‘ 125 Wave aberrations‘ 77
Photon detector民 14 Wavelength. 5. 6
Planck. 5 司 6 Wien’s displac泡ment law呵 7
Plane-parallel plate. 83. 86
Power equations‘ 175 V廿tual imagι23.24
Precipi 恤ble water. l 0
Prism. 20. 90 Zinc selenide. 46‘ 61
Pupils. 24

Radiance. 2‘ 3
Radiant emitrance司 3同 4
Rayleigh. 75
Refle心tive beam expanders. 73
Reflective coatines. 145
Reflective optics. 62
Refractive beam expanders. 59
Refractive md間， 42
Relative humiditv司 11 、 12
Resolution. 75
 About the Auth。r

Max J. Riedl was born in Kempten,

Germ缸旬， where he was e址ucated at the
Mathematisch Mechanisches lnst1tut. He
al甜 graduated from the Akademie 如r
Angewand阻 Technik, where he studied
prec 凹on mechamcs and optics. He has
worked in the field of mathematical and
optical instruments for many ye虹S 朋d w品
awarded n凹nerous pa祖 nts for mechanical,
optical and electro-optical devices He has
held 扭chmcal as well as business
1間 dership positions during the 45 y聞自 he
ltved in the United States. Riedl has published and presented many pape間 on fundamental
lens design as well 醋。n replicating optical elements and diamond 旭ming di岱active
opltcal componen阻 for applicatio阻 in the in甜甜·ed spec釘um He teaches short courses on
tl1ese su句 ec峙。n a regular baslS and is a Fellow ofSPIE Presently, he resides in Bavaria,
Germany.